摘要:

J O U R N A L O F C H E M I S T R Y Materials Feature Article Synthetic materials capable of reporting biomolecular recognition events by chromic transition Patrick Englebienne Department of Nuclear Medicine, Free University of Brussels, Brugmann Hospital, Place Van Gehuchten 4, B-1020 Brussels, Belgium. E-mail: patrick.englebienne@skynet.be Received 19th August 1998, Accepted 8th February 1999 This article presents a survey of synthetic materials capable been primarily explored as new controlled drug-delivery sysof reporting in real time by chromic transition the recog- tems.1 However, the modifications experienced by these matenition event occurring between a specific ligand and a rials from ground to new environmental conditions can be as biomolecule such as a receptor or an antibody.The review varied as a change in volume,4 in color,5 from solid to liquid, focuses on conductive polymers and colloidal gold. Besides from water-soluble to water-insoluble,6–9 from extinct to light their mode of synthesis and coupling to biomolecules, it emitting.10–13 Consequently, many of the modifications in their explores their physicochemical reactivity, emphasizing the properties as induced by changes in their environmental conmechanisms underlying the chromic transition occurring ditions can be detected by common means, which make them upon the biomolecular recognition event.The future pros- suitable candidates for transducer elements for biosensors. The pects offered by composite materials based on these concept is illustrated in Fig. 2. Some of these materials have polymers and colloids are further discussed. been used successfully in various direct biosensing systems in which the binding event induced respectively either a change Introduction in solubility,14 a change in volume,15 or the fusion,16 of the polymeric materials. Among these new materials, those Biomolecular recognition occurs when a biomolecule such as reporting a change in their environmental conditions by a a receptor or an antibody meets its specific ligand.The chromic transition are of major interest for application in encounter leads to the binding of one species by the other. direct biosensors because the change in their optical properties The biomolecular recognition event is usually monitored by occurring upon a biomolecular recognition event can be the introduction of a label into one of the complementary detected by a simple photometer and sometimes even by the binding molecules, which transduces the recognition event into naked eye.Among such materials, conductive polymers and a signal. colloidal solutions of gold particles are very promising and As a result of important advances made these last few years will be considered in this article.in the field of synthetic polymers and colloids usually qualified as ‘intelligent’ in the literature,1 the means for monitoring biomolecular recognition events have evolved quite tremen- Conductive polymers dously. These advances have allowed the design of eYcient biosensors capable of a high level of user-friendliness and Nature and properties automation.2 A biosensor is a device comprising either a ligand or a Conductive polymers are extensively conjugated organic macromolecules made of repetitive sequences of respectively receptor (antibody) species coupled to a signal transducer, which detects the binding of the complementary species.As alkyne, phenyl or heterocyclic monomers. These polymers can gain near-metallic conductivity by oxidation (p-doping) or illustrated by the sketch presented in Fig. 1, an indirect biosensor uses a separate labelled complementary species which reduction (n-doping) from their insulating state.17 The structures of the most extensively studied conductive polymers, is detected after the binding event has occurred, by measuring the activity of the label (i.e.radioactivity, fluorescence, chemi- along with the conductivities they can attain, are displayed in Fig. 3. For comparison, the conductivity of metallic copper is luminescence) associated with the complex. In the field of immuno- and receptor-assay technology, an indirect biosensing 5.8×107 S cm-1.18 Another important feature common to the conductive polymers is their capacity to undergo chromic system is termed heterogeneous because the ligand–receptor complex must be sequestered from the free fractions before transitions upon changes in their oxidation state.19 The conductive polymers are intrinsically semi-conductors in which recording the signal of the label.On the other hand, a direct biosensor (i.e. homogeneous immuno- or receptor-assay) com- the highest occupied (HOMO) and lowest unoccupied molecular orbitals (LUMO) can give rise to valence (p) and conduc- prises a ligand or a receptor respectively coupled to a signal transducer detecting in real time the binding of the complemen- tion (p*) bands.The energy gap (Eg) between the HOMO and LUMO electron bands determines the intrinsic electrical and tary species by a change in potential diVerence, current, resistance, mass, heat, or optical properties.3 optical properties of a given polymer.20 The doping process modifies the electronic band structure of the polymer by Chemical materials able to serve as transducers in direct biosensors are currently hot topics in biomedical research.The producing new electronic states that are localized in the band gap (polarons and bipolarons) and cause its passage from so-called ‘intelligent’ materials are particularly interesting candidates for direct biosensing systems.These are polymers or insulating to conducting as well as its chromic transition.21,22 In cyclic conductive polymers, these transitions are indicative colloids sensitive to external stimuli, which experience respective changes in their structure or in either their chemical or of the passage from aromatic to quinonoid molecular forms.The process of interband electronic transition is schematically physical properties in response to changes in their environmental conditions such as temperature, mass, pH, light, electric illustrated in Fig. 4. In chemical terminology, a polaron is a radical ion (spin 1/2) and a bipolaron is a dication (pair of field, or oxido-reduction.As a result of their physical responsiveness to temperature, many of these new materials have like charges, spinless). In their insulating conformation, the J. Mater. Chem., 1999, 9, 1043–1054 1043Fig. 1 Sketch emphasizing the diVerence between indirect (heterogeneous) and direct (homogeneous) biosensing systems. Fig. 2 Schematic diagram illustrating the concept of applying ‘intelligent’ synthetic materials in direct biosensing systems for biomolecular recognition. Eg of conductive polymers is usually above 1.5 eV. In the case of poly(pyrrole), it levels at 3.2 eV and by increasing p-doping it decreases progressively to 2.7 and 1.0 eV, respectively.20 When compared to other cyclic conductive polymers, poly(aniline) presents some unique structural features which allow for the control of its conductive and optical properties.23 This control results from the possibilities that exist to finetune the level of oxidation and protonation of poly(aniline).These possibilities are the consequence of the many structural variations due to the alternation in the polymer backbone of p-phenylene rings with the nitrogen atoms.The degree of oxidation depends on the fraction of these latter which is imine or amine. This is exemplified by the two extreme structures in Fig. 5A and B, with respectively a 0 and 100% NH n Structure Name Conductivity / S cm-1 Poly(acetylene) 1.5 x 105 Poly(pyrrole) 7.5 x 103 S n n Poly(thiophene) 103 n Poly( p-phenylene) 103 N N n Poly(aniline) 200 degree of oxidation.In between these two extreme states, the Fig. 3 Structures and conductivities of the most extensively studied conductive polymers. degree of oxidation may be varied in order to aVord numerous 1044 J. Mater. Chem., 1999, 9, 1043–1054Fig. 4 Schematic illustration of the energy level of one electron organic molecule in its electronic configuration adopting the equilibrium geometry of respectively (a) its ground state, (b) its first ionized state, (c) its polaron and (d) bipolaron states.Fig. 6 Electronic absorption spectra of colloidal poly(aniline) respectively in its oxidized (dotted line, Eg 2.15 eV) and n-doped state (solid line, Eg 1.45 eV). intermediate structures each displaying diVerent conductivities and electronic absorption spectra upon external protonation.produce self-doped polymers. The process of self-doping is When an anionic substituent capable of protonation is introillustrated in Scheme 1b with poly(thiophene-3-ethanesulfon- duced on the p-phenylene ring, the polymer structure may be ate) as the example. The advent of water-soluble and self- in a partial conductive form when the substituent is protonated doped conductive polymers allowed for their direct use in (Fig. 5C). This latter structure transforms into the fully conbiochemical reactions, particularly as electron transfer means ductive form upon self-doping (Fig. 5D) resulting from the in enzyme sensors.30,35 loss of the proton by the substituent.24 The Eg and chromic transitions of poly(aniline) upon n-doping are illustrated by Synthesis the electronic absorption spectra in Fig. 6. Another class of conjugated polymers worth mentioning is Whether soluble or insoluble in water, most conductive the substituted poly(diacetylenes). As illustrated in Fig. 7, the polymers are synthesized by electrochemical or chemical structure of these polymers is made of ene–yne alternating means.These synthetic procedures have been extensively bonds. The molecules of this polymer class are particularly reviewed, in particular for poly(thiophene).36 Whether chemi- flexible and they undergo important color changes when cal or electrochemical, the most commonly used synthetic subjected to planar–nonplanar conformational transitions.25 processes occur through the oxidation of the monomers.The Up to the mid eighties, all the conductive polymers available most commonly used chemical oxidants are iron(III ) chloride were soluble only in organic solvents,26–28 and their use in and ammonium persulfate. aqueous biological systems was consequently restricted to that When substituted heterocyclic monomers are used as starting of electronic transfer films incorporated in amperometric sen- blocks, oxidative polymerization presents the drawback of a sors.29–31 These water-insoluble polymers can only be doped lack of control of the structure of the oligomers or polymers by the addition of oxidative dopant counter-ions which diVuse formed.As shown in Scheme 2, the resulting substituted in and out of the polymer backbone to balance the charges polymers are structurally inhomogeneous.37 This may be of created.32 This process is called external doping and is illus- importance for the application in biomolecular recognition trated in Scheme 1a with poly(thiophene) as an example.In because there is experimental evidence that the regiochemistry the late eighties, it was observed that the introduction of an of substituted poly(heterocycles) controls their conformational alkanesulfonate sidechain in cyclic or heterocyclic monomers features, which in turn, governs the degree of p–p conjugation yielded water-soluble conductive polymers.33,34 Moreover, between adjacent rings.36 This control has consequently a because the counter-ion is attached to the polymer through a covalent bond, these water-soluble polymers can lose a proton upon oxidation, simultaneously with the loss of an electron to HN HN n A N N n B N N n HN HN n COOH COOH COO- COOC D Fig. 5 Extreme structures of poly(aniline) with an oxidation degree of 0 and 100% (respectively A and B). The introduction of a carboxylic substituent on the p-phenylene ring transforms the polymer into a partially conductive form (C), capable of aVording the fully conductive S n + yMX a.b. S n SO3H S n SO3 - - nH+ + nH+ S n y+ (X-) + yM Oxidation Reduction Scheme 1 form (D) upon self-doping. J. Mater. Chem., 1999, 9, 1043–1054 1045C C C C R R' n R = R' = O n NH O O O x n = 3 and x = 1 or 3 R = R' = OH O Colour Yellow R = R' = O n NH O O O x n = 4 and x = 1 or 3 Yellow Blue R = R' = NH O Purple OH R = R' = OH Orange R = R' = OH O Blue R = R' = O O Red Fig. 7 Examples of substituted poly(diacetylene) structures and their corresponding colour in the coplanar conformation. tremendous influence on the capacity of the polymer to translate a departure from coplanarity by a chromic transition. 38 Poly(thiophene) derivatives are in this respect of particular interest because of their specific photochemical properties.11,12 As discussed below, such departure from coplanarity is an important mechanism of transducing biomolecular recognition.As shown in Scheme 3, control of the polymer architecture can be achieved by coupling 2-bromo-3- alkyl-5-thienylmagnesium bromide in the presence of a nickel catalyst.39 Other strategies using the cross coupling of 2- bromo-3-alkylthiophene with 2-(trimethylstannyl )-3-alkylthiophene monomers have also been reported to provide a similar advantage.40 It has been claimed that water-soluble poly(anilinesulfonate) could not be synthesized by chemical or electrochemical oxidation from o-aminobenzene- or m-aminobenzene-sulfonic acid because the substitution of a hydrogen atom on the phenyl ring by electron withdrawing substituents like -SO3H led to an increase of the steric hindrance and to a strongly deactivating influence likely to limit the polymerization process.Until recently, this problem was addressed by sulfonating the already polymerized aniline with fuming sulfuric acid.24 S R MgBr Br Ni(dppp)Cl2 S R S R n S R FeCl3/(NH4)2S2O8 or e- at Pt S S R R S R S R S R n S R S R n S R S R S R S R n S R S R S R S R n Scheme 2 Scheme 3 1046 J.Mater. Chem., 1999, 9, 1043–1054Fig. 8 DiVerence electronic spectra recorded at various times (3, 5, 120, 192 and 288 h; arrow) during the polymerization of o-anthranilic acid in aqueous phase using ferric chloride as the oxidant. This strategy leads however to ill defined and partially sulfonated polymers. The question is still under debate and is quite controversial. In a recent paper,41 the synthesis of fully sulfonated poly(aniline) from o- and m-aminobenzenesulfonic acid by chemical oxidation in the liquid phase under high pressure has been reported.Poly(aniline) substituted with a carboxylic acid group has been obtained by oxidative polymerization of Fig. 9 Evolution of the A530 nm of a diluted buVered solution of o-anthranilic acid with ammonium persulfate, aVording a poly(o-anthranilic acid) as a function of pH between pH 6.5 and 9.water-soluble polymer.42 However, the UV–vis spectra of the polymerized material showed no absorption in the visible region and the authors concluded that the carboxylic acid protons and hence to pH. A decrease in pH enhances doping substituent restricted the p-conjugation along the polymer and conversely, pH increases progressively undope the poly- chain.However, the successful electropolymerization of substimer. The doping–undoping process induced by pH changes is tuted poly(anilines) from o- and m-aminobenzoic acid as well translated by important modifications in the absorption at the as from m-aminobenzenesulfonic acid has also been reported wavelength corresponding to the p–p* transition of the poly- and the polymers obtained were shown to be self-doped and mer.The pH eVect on a buVered solution of poly(o-anthranilic electrochemically active in aqueous solutions.43 In a more acid) is illustrated in Fig. 9. It is worth noting that when the recent publication,44 we have shown that conjugated and conductive polymers most eVective at translating a biomolecu- water-soluble poly(o-anthranilic acid) could be obtained by lar recognition process by chromic transition are considered, chemical oxidation with ferric chloride.Progressive p–p* transthe highest pH eVect is observed around physiological pH itions at lower energy levels (475 nm, 2.6 eV and 530 nm, values, such as in the example of Fig. 9. In other cases, the 2.3 eV) occur during the polymerization process as shown in range of pH sensitivity may be broad and this property can the diVerence spectra in Fig. 8, which are indicative of the be applied to the fabrication of reversible optical sensors progressive conjugation along the polymer backbone during for pH measurement such as recently reported with synthesis.The same approach allowed us to polymerize also poly(pyrrole).52 various substituted poly(thiophenes).44,45 Similarly, the respective addition of oxidants or of reductants Conductive polymers have also been synthesized in colloidal to the polymers at a fixed pH exerts a profound influence on form, particularly poly(aniline) and poly(pyrrole).46–48 These the absorbance (A) at the p–p* transition wavelength.This materials can only be externally doped since they do not eVect is illustrated with poly(o-anthranilic acid) in Fig. 10. contain any counter ion covalently linked to the polymer The addition of increasing millimolar concentrations of backbone. One exception to this is the colloidal copolymers ammonium persulfate to the polymer solution buVered at of poly(pyrrole) and poly(styrenesulfonate), the sulfonate pH 9 increases the A530 nm and conversely, the addition of the counter-ions being incorporated into the poly(pyrrole) during same increasing concentrations of sodium borohydride to the the polymerization providing cation-exchange sites.49 This polymer solution buVered at pH 7 decreases the A at the same latter copolymerization approach could generate in the near wavelength in a corresponding way.This capacity of conduc- future new composite materials with promising properties for tive polymers to respond spectrophotometrically to oxido- application in the biomolecular recognition field. reduction has been applied to the quantitative detection of Finally, substituted poly(diacetylenes) on the one hand can ascorbic acid using poly(thiophene) films.5 We have adapted be produced by 1,4 addition of the substituted diacetylenic the technique in an automated clinical chemistry analyzer monomers initiated by either c25 or UV50 irradiation. On the using poly(o-aminobenzenesulfonic acid) and poly(o-anthra- other hand, soluble derivatives of poly(acetylene) have been nilic acid) in aqueous solution.In these conditions, the polymer obtained by ring-opening metathesis polymerization of reduction by ascorbic acid monitored by photometry occurs monosubstituted cyclooctatetraenes.51 in less than 10 min. The chromic sensitivity of conductive polymers to both pH Chromic transition of water-soluble conductive polymers upon and oxido-reduction could lead in the near future to the oxido-reduction availability of cheap analytical reagents or instruments, finding application in clinical chemistry, pharmacology, biotechnology As a consequence of their self-doped nature, water-soluble conductive polymers are highly sensitive to the presence of or the food industry.J. Mater. Chem., 1999, 9, 1043–1054 1047H2N N H N N NH R-NH2 O O O O O O NH NH N N NH NH R n + R = biomolecular species n Scheme 4 merization with bromoacetyl bromide, depending on the synthetic conditions.These derivatives can then further be either carboxylated (Scheme 6c) or aminated (Scheme 6d) by being Fig. 10 EVect of increasing concentrations of either ammonium reacted with either mercaptoacetic acid or triethylenetetramine, persulfate (triangles) or sodium borohydride (circles) on the A at respectively.54 The derivatized polymer can then be linked 530 nm of poly(o-anthranilic acid) diluted in 50 mM phosphate buVer adjusted respectively at pH 9 or 7.covalently to ligands, receptors or antibodies. Chromic transition of conductive polymers upon biomolecular Covalent coupling of water-soluble conductive polymers to recognition biomolecules Colloidal conductive polymers have been used as early as in The covalent linkage between one of the biomolecular reactants 1992 by Tarcha’s group at Abbott Laboratories as reporter and the water-soluble conductive polymer label is of prime reagents in solid-phase heterogeneous immunoassays.55 In this importance because it is supposed to secure the close proximity particular application, advantage was taken of the deep colour of the label to the biomolecular recognition site.Consequently, of the colloidal particles so as to detect either by densitometry zero-length linkages or small bridges are usually preferable to or visually any binding of antibody coated particles to the long connecting arms. This prerequisite is of less importance for water-insoluble polymers since the mechanism of transducing the biomolecular recognition event is likely to proceed from diVerent principles which are discussed in the next section.Most molecules used in biomolecular recognition are proteins, peptides or sugars which contain amine, hydroxy, mercapto or carboxylic groups available for coupling to conductive polymer labels.When the molecule to be labelled is a small ligand such as a steroid or a drug which does not contain vacant functionalities, it must be chemically modified prior to coupling. As shown in Scheme 4, poly(aniline) contains terminal amines which can be used in a cross-linking reaction to corresponding amines in proteins or peptides through glutaraldehyde. Except for poly(aniline), most conductive polymers do not have corresponding groups available for coupling.Consequently, substituting the polymer backbone is useful not only for granting self-doping capacity and water solubility, but also for providing vacant functionalities for their covalent coupling to ligands and receptors. For instance, the substitution of poly(thiophene) with carboxylic acid groups allows for its covalent linkage in aqueous solutions to amine residues (i.e.e-amino groups of lysine in proteins) in presence of a water-soluble carbodiimide such as 1-ethyl-3-(3-(dimethylamino) propyl )carbodiimide. The carbodiimide forms a reactive carboxylic anhydride intermediate (Scheme 5) which further reacts with the amine to form an amide bond.53 Alternatively (Scheme 6), conductive polymers which contain a reactive heterocycle such as poly(pyrrole) can be S COOS COOn + N N N S C S COOH n O O N HN N R NH2 R = biomolecular species S C S COOH n HN O R + N N H NH O Scheme 5 either N- (Scheme 6a) or C- (Scheme 6b) acylated after poly- 1048 J.Mater. Chem., 1999, 9, 1043–1054NHn + BrH2C CBr O D N n C CH2Br O N n C CH2Br O + HS C H2 C OH O N n C CH2 O S CH2 COOH NHn + BrH2C CBr O AlCl3 NHn C O CH2Br a.b. c. NHn C O CH2Br d. + H2N HN NH NH2 NHn C O CH 2 NH HN NH NH3 Br- Scheme 6 antigen previously captured by the solid phase. The mechanism receptor binds the ligand, the film changes colour from blue to red. The red colour intensity is directly proportional to the of indicating the biomolecular recognition in this kind of application does not depart from the use of a coloured quantity of virus reacted with the film.The system consequently constitutes a direct colorimetric biosensor. In another poly(styrene) latex and is consequently beyond the scope of the present discussion. article,60 the authors reported the colorimetric detection of the molecular recognition between the Cholera toxin and Gm1 For the sake of clarity, I will arrange the next discussion around the regioregularity of the conductive polymers con- ganglioside incorporated in liposomes made of a substituted poly(diacetylene). In this latter system, the intensity of the sidered.The mechanism of transducing biomolecular recognition by chromic transition is indeed likely to diVer depending color change was directly proportional to the dose of toxin interacting with the ligand and the least detectable dose was on the degree of regioregularity of the polymeric structures.Substituted poly(diacetylene) derivatives are regioregular in the mg l-1 range. Other examples of application have been reported with substituted poly(diacetylenes) derivatized with and undergo dramatic color changes upon planar–nonplanar conformational transitions induced by either a change in toxin- or virus-specific ligands, which undergo a blue–red chromic transition upon ligand recognition by the patho- temperature (thermochromism), solvent composition (solvatochromism) or mechanical stress (mechanochromism).25,50 As gen.61,62 A departure from coplanarity has also been observed when a biotin-functionalized regioregular substituted poly(thi- further shown in Fig. 7, the side-chain structure plays a critical role in determining the original colour of the polymer in the ophene) reacts with the biotin receptor (avidin) in aqueous solution. This eVect results in the polymer solution turning planar conformation of the ene-yne backbone, and hence on the type of chromic transition occurring upon departure from from violet to yellow.63 The transduction of biomolecular recognition by chromic coplanarity.25,50 Similar eVects have been observed with regioregular substituted poly(acetylenes)51 and poly(thiophenes)38,56 transition resulting from specific changes in the conjugated polymer backbone conformation is possible with well which can undergo severe thermochromic, solvatochromic and ionochromic transitions.In poly(thiophenes), the departure organized regioregular polymers. The mechanism of chromic transition by random polymers in the same circumstances is from a coplanar conformation decreases the conductivity of the polymer57 and can be counteracted by the rigidification of less clear. We have reported the synthesis of several watersoluble conductive random polymers or oligomers and their the p-conjugated system.58 Such chromic transition resulting from a decrease in the eVective conjugation length of the use as transducing reagents acting by chromic transition in homogeneous competitive immunoassays for antigens and polymer backbone has also been observed when the weight stress due to a biomolecular species binding its counterpart is haptens.44,45 Currently, these assays allow detection of ligands in the nanomolar range.45 Fig. 11 illustrates the performance applied on a given polymeric structure. Such application of molecular mechanochromism has been reported59 with the use of these reagents by showing the kinetics of the chromic transition occurring when theophylline labelled with poly(thi- of a Langmuir–Blodgett film made of a polymerized diacetylenic lipid matrix functionalized with a sialic acid ligand specific ophene-3-carboxylic acid) associates with a specific antibody and its eventual dissociation from the complex by an excess for the influenza virus hemagglutinin.When the hemagglutinin J. Mater. Chem., 1999, 9, 1043–1054 1049Fig. 12 Schematic drawing illustrating the capacity of water-soluble conductive polymers to translate directly a biomolecular recognition event by a change in their visible absorption spectrum under the influence of the local pH existing around the binding pocket.The conductive polymer is represented by the circles, respectively in its ground-state colour (hatching) and transition-state colour (black). The ligand and receptor are respectively represented by a triangle and Y.Fig. 11 Kinetics of association and dissociation between theophylline suspended in water. Under normal circumstances, the metallic labelled with poly(thiophene-3-carboxylic acid) and a specific anti- gold is surrounded by a predominantly anionic double layer. theophylline antibody transduced by the chromic transition of the The sol is very stable due to the repulsive forces and shows a label as measured in a clinical chemistry automated analyzer at deep red color.However, due to induced changes in the outer 340 nm. The dissociation (arrow) is induced by the addition of a 100 charges, the particles are agglutinated and coagulated by the fold excess of unlabelled theophylline. addition of millimolar concentrations of various salts and this irreversible process changes the colour from red to blue.66 Colloidal gold hydrosol can be protected from agglutination of unlabelled theophylline.The rapidity of the kinetics makes it possible to measure the signal in a random access clinical by salts by coating the particles with a protein layer. Gold hydrosols are characterized by a discrete band in the chemistry analyzer.These instruments, which are widely used in clinical laboratories, are walk-away instruments capable of visible electronic spectrum which is called the surface plasmon resonance (SPR). It is a quantized plasma oscillation occurring mixing and incubating reagents and samples in individual cuvettes according to programmed conditions. They are at the surface of the gold particles, resulting in a characteristic peak location and width at maximal A.67 The SPR wavelength equipped with a flash lamp photometer allowing monitoring of subtle changes of A in each cuvette at a given wavelength and width depend on the refractive index of the particles and hence on their average diameter.Any increase in average and at short regular intervals (usually 25 s).64 The potential application of conductive polymers to homogeneous immuno- particle diameter in the sol induces a red shift of the SPR wavelength along with a decrease in maximal A.This is assays that can be run in such automated instruments constitutes an interesting and quite revolutionary challenge for the illustrated in Fig. 13 which shows the change occurring in the visible electronic spectrum of colloidal gold when the particles next few years.Because of the lack of regularity in the polymeric chains used are coated with a small layer of human serum albumin. We have applied this characteristic of colloidal gold in quantitative in such applications, the molecular mechanism inducing the chromic transition cannot be assigned only to a departure from colorimetric assays for proteins.68,69 coplanarity.As discussed previously, these water-soluble polymers are easily doped or undoped by subtle pH changes, and Synthesis we have suggested that the chromic transition depends on the Isodisperse colloidal gold hydrosols are synthesized by the local pH existing around the binding pocket of the receptor or controlled reduction of an aqueous solution of tetrachloroauric the antibody, which is known to change upon ligand recogacid.The reduction of Au3+ ions produces a supersaturated nition.65 This mechanism is illustrated by the schematic drawing molecular Au0 solution and as the Au0 concentration increases in Fig. 12. Such an explanation is further supported by the fact the gold atoms cluster and form seeds of nuclei.Particle that we were unable to apply the technique to sandwich assays growth occurs by further deposition of metallic gold upon using labelled antibodies, most probably because the conductive the nuclei.70 polymer label attached to an antibody molecule is situated too Strong reductants like white phosphorus or sodium far from the binding pocket to sense and transduce the pH borohydride produce a great number of nuclei and hence change occurring upon ligand recognition.rather small particles with average diameter of 2 to 10 nm depending on the synthetic conditions. Bigger particles (20 to Colloidal gold 100 nm average diameter) can be obtained by using a weaker reductant like sodium citrate advocated by Frens.71 The size Nature and properties of the particles obtained by this method depends on the citrate5tetrachloroauric acid ratio.The lower the ratio, the Gold hydrosols are hydrophilic colloids made of monodisperse, finely divided particles (5–100 nm diameter) of metallic gold greater the diameter of the particles. 1050 J. Mater. Chem., 1999, 9, 1043–1054Fig. 13 Visible electronic spectrum of a colloidal gold hydrosol (solid line) compared to that of the same sol in which the particles are coated with a small layer of human serum albumin (dashed line).The change in refractive index due to the protein coating induces a red shift and a decrease in A at the SPR wavelength. Coating colloidal gold particles with proteins The coating of colloidal gold nanoparticles with proteins is performed by electrostatic charge adsorption.Colloidal gold nanoparticles remain negatively charged over a wide range of pH. When a partially protonated protein is contacted with the Fig. 14 EVect of coating pH on the capacity of the antibody-coated particles at a suitable pH, the positive charges on the protein gold probes to recognize the specific ligand. The decrease in A at 600 nm after salt addition to the coated particles (open circles, left are attracted by the anionic layer surrounding the particles axis) shows stabilization of the colloid between pH 6.5 and 8.The with which it forms ionic bonds. The stability granted to the capacity of the particles to bind a small amount of the ligand, as hydrosol by the protein coating process is usually verified by measured by the increase of A at 600 nm after incubation which is the absence of coagulation upon addition of a concentrated representative of the refractive index change (solid triangles, right salt solution.72 axis) peaks at a narrower coating pH range of 7–7.5, indicating proper When colloidal gold is coated with proteins capable of antibody orientation at this pH.biomolecular recognition such as receptors or antibodies, the selection of the pH for coating is most critical. This parameter mic transition resulting from the selective agglutination of the allows control of the part of the protein which is positively particles has been also applied to the detection of charged and hence determines the correct orientation of the complementary polynucleotide strands.75,82 external binding site of the particle for proper ligand recog- Recently,83 I have shown that a lattice formation was not nition.73,74 This point is illustrated in Fig. 14 with the stability an absolute requisite for observing a measurable SPR shift and reactivity of colloidal gold particles coated by an antibody and that such a shift could be observed when colloidal gold at various pHs. The decrease in A600 nm after salt addition particles coated with a monoclonal antibody specific for a indicates that the sol is stabilized for a large range of pH (6.5 single epitope bind the protein ligand.This is illustrated by to 8), although the peak of ligand recognition by the gold the artist’s view on the cover of this issue. In this case, the probes is much narrower and lies at coating pHs around shift in SPR wavelength reflects very small changes in the 7–7.5.Whilst the optimal pH for coating usually lies around refractive index at the particle surface which are the direct the isoelectric point of the protein, there are no general rules result of mass changes in the approximate medium induced and adsorption interaction isotherms such as that shown in by the biomolecular recognition event.The SPR wavelength Fig. 14 should be conducted in each case. However, once the shift phenomenon is further dose-related as illustrated by the optimal conditions are identified, the coating process is usually visible spectra shown in Fig. 15. These spectra were recorded reproducible and easy to scale-up.after the respective incubation of monoclonal antibody-coated colloidal gold particles with increasing concentrations of the Mechanism of chromic transition upon biomolecular recognition specific ligand, human chorionic gonadotropin (a pregnancy protein). The SPR wavelength shift of colloidal gold is used When colloidal gold probes coated with an antibody bind the specific ligand, a red shift occurs in the SPR wavelength of as transducer of the biomolecular recognition event of protein ligands in an automated clinical chemistry analyzer, the signal the gold sol along with a decrease in A at the maximal wavelength.When the antibody used for coating recognizes being measured at wavelengths in the red part of the spectrum (600 nm), where both the wavelength shift and the increase in various molecular determinants on the ligand, the gold probes progressively form a lattice of agglutinated particles around peak width resulting from changes in the refractive index of the particles enhance the A in a dose-dependent manner.ligand molecules and the SPR shift reflects the apparent increase in diameter of the individual particles in the solution Fig. 16 shows the typical dose–response curve of a homogeneous immunoassay for human ferritin performed in a which turns blue.75 The remarkable deep red colour and capacity of colloidal gold to undergo a chromic transition clinical chemistry analyzer. This new application of the chromic transition of colloidal gold upon biomolecular recognition upon agglutination have made possible its use for many years as a reporter reagent in either heterogeneous enzyme- is likely to have a tremendous potential, not only as a bioanalytical tool, but also as a real-time biosensing system like immunoassays76–78 or in homogeneous agglutination immunoassays,77,79–81 respectively.The same principle of chro- providing kinetic83–85 and even conformational86 information J.Mater. Chem., 1999, 9, 1043–1054 1051on the biomolecular interaction such as those obtained in more sophisticated systems like the BIAcore instrument. This new applicability of the technique confirms its interesting potential in the new drug discovery process, particularly for either the selection of the highest aYnity ligands87 or the detection of natural ligands and of orphan receptors,88 respectively.Colloidal gold has also been incorporated into a classical SPR biosensing system,89 comprising the interaction between an antibody immobilized on a gold film and the antigen coated on colloidal gold particles. The presence of the colloid increased tremendously the SPR sensitivity to protein–protein interactions. Future prospects One of the problems usually pointed out in the use of colloidal gold as a chromic transition reagent is the lack of possible covalent linkage of the biomolecular species to the probes.When they are stored in an insuYciently stabilizing medium, the particles coated by charge adsorption are prone to the progressive release of the adsorbed protein in the surrounding medium, which decreases their reactivity.This problem can be circumvented by the use of thiol-derivatized gold nanoparticles, the synthesis of which has been reported in two-phase90 as well as single-phase91 liquid systems. The attachment of a monolayer of bifunctional organic thiol molecules to the gold nanoparticles provides a vacant functionality for the covalent linkage of proteins.91 In a similar way, gold colloids have been Fig. 15 Visible electronic spectra recorded after incubation of mixtures stabilized by molecules containing amine functional groups,92 containing a fixed amount of colloidal gold particles coated with a specific monoclonal antibody and increasing concentrations (respect- including poly(amidoamine) dendrimers.93 This latter process ively no ligand added, solid spectrum; 33 pmol dm-3, dashed spec- allows for the amine groups to be available at the external trum; 60 pmol dm-3, dotted spectrum) of the specific ligand protein, surface of the particles.human chorionic gonadotropin. The quite recent discovery and progressive understanding of the redox chemical character of alkanethiolate monolayer protected colloidal gold nanoparticles, which exhibit HOMO–LUMO Eg of 0.4 to 0.9 eV94 open new prospects for application in biosensors.In the light of these new perspectives, I have attempted the template polymerization of substituted poly(anilines) on colloidal nanoparticles according to a scheme similar to that already reported for poly(diacetylene).95 As shown in Fig. 17, a suitable selection of the conductive polymer with a p–p* transition absorption wavelength matching that of colloidal gold SPR aVords a composite colloidal material displaying a tremendously enhanced absorption at the SPR wavelength.The use of such a material, which combines the capacities for chromic transition of both gold and watersoluble conductive polymers, could allow for a further Fig. 16 Dose–response curve of an homogeneous immunoassay for human ferritin using antibodies labelled with colloidal gold.The assay Fig. 17 DiVerence visible spectrum of o-anthranilic acid polymerized is run in an automated clinical chemistry analyzer (Cobas-Mira) with an incubation time of 4 minutes. Data are the average ±sd of on colloidal gold nanoparticle templates versus that of the underivatised gold colloid. duplicate measurements. 1052 J. Mater. Chem., 1999, 9, 1043–105441 H. S. O. Chan, A. J. Neuendorf, S.-C. Ng, P. M. L. Wong and improvement in the level of analytical sensitivity attained by D. J. Young, Chem. Commun., 1998, 1327. the current homogeneous liquid-phase biosensors based on 42 M. T. Nguyen and A. F. Diaz, Macromolecules, 1995, 28, 3411. these independent chromic transducers. 43 A.A. Karyakin, A. K. Strakhova and A. K. Yatsimirsky, J. Electroanal. Chem., 1994, 371, 259. 44 P. Englebienne and M. Weiland, J. Immunol. Methods, 1996, 191, Many thanks to my son Gwenn for his excellent design and 159. computer generation of the three-dimensional figure of the 45 P. Englebienne and M.Weiland, Chem. Commun., 1996, 1651. cover. 46 B. Vincent and J. Waterson, J. Chem.Soc., Chem. Commun., 1990, 683. 47 N. Gospodinova, P. Mokreva and L. Terlemezyan, J. Chem. Soc., References Chem. Commun., 1992, 923. 48 S. P. Armes and B. Vincent, J. Chem. Soc., Chem. Commun., 1 I. C. Kwon, Y. H. Bae and S. W. Kim, Nature, 1991, 354, 291. 1987, 288. 2 J. P. Gosling, in Immunotechnology, eds. J. P. Gosling and 49 Z. Qi and P. G. Pickup, Chem. Commun., 1998, 15.D. J. Reen, Portland Press, London, 1993, p. 91. 50 S. Okada, S. Peng, W. Spevak and D. Charych, Acc. Chem. Res., 3 C. L. Morgan, D. J. Newman and C. P. Price, Clin. Chem., 1996, 1998, 31, 229. 42, 193. 51 C. B. Gorman, E. J. Ginsburg and R. H. Grubbs, J. Am. Chem. 4 G. Chen and A. S. HoVman, Nature, 1995, 373, 49. Soc., 1993, 115, 1397. 5 H. S. O. Chan, S. C. Ng and S. H. Seow, Synth. Met., 1994, 52 R.Koncki and O. S. Wolfbeis, Anal. Chem., 1998, 70, 2544. 66, 177. 53 N. Nakajima and Y. Ikada, Bioconjugate Chem., 1995, 6, 123. 6 R. Yoshida, K. Uchida, Y. Kaneko, K. Sakai, A. Kikuchi, 54 M. R. Pope, S. P. Armes and P. J. Tarcha, Bioconjugate Chem., Y. Sakurai and T. Okano, Nature, 1995, 374, 240. 1996, 7, 436. 7 Y. G. Takei, T. Aoki, K. Sanui, N. Ogata, T.Okano and 55 P. J. Tarcha, D. Misun, D. Finley, M. Wong and J. J. Donovan, Y. Sakurai, Bioconjugate Chem., 1993, 4, 42. in Polymer Latexes, ed. E. S. Daniels, E. Sudol and 8 G. Chen and A. S. HoVman, Bioconjugate Chem., 1993, 4, 509. M. S. El-Aasser, American Chemical Society, 1992, p. 347. 9 A. Chilkoti, G. Chen, P. S. Stayton and A. S. HoVman, 56 I. Le�vesque and M. Leclerc, J.Chem. Soc., Chem. Commun., Bioconjugate Chem., 1994, 5, 504. 1995, 2293. 10 M. Berggren, O. Ingana�s, G. Gustafsson, J. Rasmusson, 57 M. Leclerc and G. Daoust, J. Chem. Soc., Chem. Commun., 1990, M. R. Andersson, T. Hjertberg and O. Wennerstro�m, Nature, 273. 1994, 372, 444. 58 J. Roncali, C. Thobie-Gautier, E. H. Elandaloussi and P. Fre` re, 11 B. S. Kang, M. L. Seo, Y. S. Jun, C.K. Lee and S. C. Shin, Chem. J. Chem. Soc., Chem. Commun., 1994, 2249. Commun., 1996, 1167. 59 D. H. Charych, J. O. Nagy, W. Spevak and M. D. Bednarski, 12 B. Xu and S. Holdcroft, J. Am. Chem. Soc., 1993, 115, 8447. Science, 1993, 261, 585. 13 R. H. Friend, R. W. Gymer, A. B. Holmes, J. H. Burroughes, 60 J. J. Pan and D. Charych, Langmuir, 1997, 13, 1365. R. N. Marks, C. Taliani, D.D. C. Bradley, D. A. Dos Santos, 61 A. Reichert, J. O. Nagy, W. Spevak and D. Charych, J. Am. J. L. Bre�das, M. Lo�gdlund and W. R. Salaneck, Nature, 1999, Chem. Soc., 1995, 117, 829. 397, 121. 62 W. Spevak, J. O. Nagy and D. H. Charych, Adv. Mater., 1995, 14 S. M. Barnard and D. R. Walt, Science, 1991, 251, 927. 7, 85. 15 P. J. Tarcha, W. Bindseil and V. P. Chu, J. Immunol.Methods, 63 K. Faý� d and M. Leclerc, Chem. Commun., 1996, 2761. 1989, 125, 243. 64 L. SchoeV, Anal. Chem., 1997, 69, 206R. 16 P. S. Stayton, T. Shimoboji, C. Long, A. Chilkoti, G. Chen, 65 C. S. Lee, P. Y. Huang and D. M. Ayres, Anal. Chem., 1991, J. M. Harris and A. S. HoVman, Nature, 1995, 378, 472. 63, 464. 17 R. L. Greene and G. B. Street, Science, 1984, 226, 651. 66 D. H.Everett, in Basic Principles of Colloid Science, Royal Society 18 D. Clery, Science, 1994, 263, 1700. of Chemistry, London, 1988, p. 30. 19 A. O. Patil, A. J. Heeger and F. Wudl, Chem. Rev., 1988, 88, 183. 67 J. L. CoVer, J. R. Shapley and H. G. Drickamer, J. Am. Chem. 20 J. L. Bre�das and G. B. Street, Acc. Chem. Res., 1985, 18, 309. Soc., 1990, 112, 3736. 21 S. Etemad, A. G. Heeger and A.G. MacDiarmid, Annu. Rev. 68 L. Ellens, G. Ameryckx and P. Englebienne, Meded. Fac. Phys. Chem., 1982, 33, 443. Landbouwwet. Rijksuniv. Gent, 1988, 53, 2035. 22 M. J. Nowak, S. D. D. V. Rughooputh, S. Hotta and A. J. Heeger, 69 L. Ellens, G. Ameryckx and P. Englebienne, Int. Biotech. Lab., Macromolecules, 1987, 20, 965. 1989, 7, 33. 23 W.-C. Chen and S. A. Jenekhe, Polym.Prepr., 1992, 33, 204. 70 J. E. Beesley, in Colloidal Gold: A New Perspective for 24 J. Yue, Z. H. Wang, K. R. Cromack, A. J. Epstein and Cytochemical Marking, Oxford University Press, Oxford, 1989, A. G. MacDiarmid, J. Am. Chem. Soc., 1991, 113, 2665. p. 5. 25 G. N. Patel, R. R. Chance and J. D. Witt, J. Chem. Phys., 1979, 71 J. De Mey, in Immunochemistry: Modern Methods and 70, 4387.Applications, ed. J. M. Polak and. I. M. Varndell, J. Wright and 26 K. Y. Jen, R. Oboodi and R. L. Elsenbaumer, Polym. Mater. Sci. Sons, Bristol, 1986, p. 115. Eng., 1985, 53, 79. 72 W. D. Geoghegan and G. A. Ackerman, J. Histochem. Cytochem., 27 J. E. Frommer, Acc. Chem. Res., 1986, 19, 2. 1977, 25, 1187. 28 S. Hotta, S. D. D. V. Rughooputh, A. J. Heeger and F. Wudl, 73 W.D. Geoghegan, J. Histochem. Cytochem., 1988, 36, 401. Macromolecules, 1987, 20, 212. 74 W. D. Geoghegan, J. Clin. Immunoassay, 1988, 11, 11. 29 N. C. Foulds and C. R. Lowe, Anal. Chem., 1988, 60, 2473. 75 R. Elghanian, J. J. StorhoV, R. C. Mucic, R. L. Letsinger and 30 T. Matsue, M. Nishigawa, T. Sawaguchi and I. Uchida, J. Chem. C. A. Mirkin, Science, 1997, 277, 1078. Soc., Chem. Commun., 1991, 1029. 76 J. H. W. Leuvering, P. J. H. M. Thal, M. van der Waart and 31 B. Fabre, G. Bidan and M. Lapkowski, J. Chem. Soc., Chem. A. H. W. M. Schuurs, J. Immunoassay, 1980, 1, 77. Commun., 1994, 1509. 77 T. J. C. Gribnau, J. H. W. Leuvering and H. van Hell, 32 M. Mastragostino, C. Arbizzani, A. Bongini, G. Barbarella and J. Chromatogr., 1986, 376, 175. M. Zambianchi, Electrochim. Acta, 1993, 38, 135. 78 J. H. W. Leuvering, P. J. H. M. Thal and A. H. W. M. Schuurs, 33 A. O. Patil, Y. Ikenoue, F. Wudl and A. J. Heeger, J. Am. Chem. J. Immunol. Methods, 1983, 62, 175. Soc., 1987, 109, 1858. 79 J. H. W. Leuvering, P. J. H. M. Thal, D. D. White and 34 Y. Ikenoue, Y. Saida, M. Kira, H. Tomozawa, H. Yashima and A. H. W. M. Schuurs, J. Immunol. Methods, 1983, 62, 163.M. Kobayachi, J. Chem. Soc., Chem. Commun., 1990, 1694. 80 J. H. W. Leuvering, B. C. Goverde, P. J. H. M. Thal and 35 T. Tatsuma, K. Ariyama and N. Oyama, Anal. Chem., 1995, 67, A. H. W. M. Schuurs, J. Immunol. Methods, 1983, 60, 9. 283. 81 J. H. W. Leuvering, P. J. H. M. Thal, M. van der Waart and 36 J. Roncali, Chem. Rev., 1992, 92, 711. A. H. W. M. Schuurs, J. Immunol.Methods, 1981, 45, 183. 37 G. Barbarella and M. Zambianchi, Tetrahedron, 1994, 50, 11249. 82 C. A. Mirkin, R. L. Letsinger, R. C. Mucic and J. J. StorhoV, 38 S. D. D. V. Rughooputh, S. Hotta, J. Heeger and F. Wudl, Nature, 1996, 382, 607. J. Polym. Sci., 1987, 25, 1071. 83 P. Englebienne, Analyst, 1998, 123, 1599. 39 R. D. McCullough, R. D. Lowe, M. Jayaraman and 84 M. M. Morelock, R.H. Ingraham, R. Betageri and S. Jakes, D. L. Anderson, J. Org. Chem., 1993, 58, 904. J. Med. Chem., 1995, 38, 1309. 40 G. Barbarella, A. Bongini and M. Zambianchi, Macromolecules, 1994, 27, 3039. 85 R. Nakamura, H. Muguruma, K. Ikebukuro, S. Sasaki, J. Mater. Chem., 1999, 9, 1043–1054 1053R. Nagata, I. Karube and H. Pedersen, Anal. Chem., 1997, 69, 92 D.V.LeV, L. Brandt and J.R. Heath, Langmuir, 1996, 12, 4723. 4649. 86 H. Sota, Y. Hasegawa and M. Iwakura, Anal. Chem., 1998, 70, 93 M. E. Garcia, L. A. Baker and R. M. Crooks, Anal. Chem., 1999, 71, 256. 2019. 87 G. Lowe, Chem. Soc. Rev., 1995, 24, 309. 94 S. Chen, R. S. Ingram, M. J. Hostetler, J. J. Pietron, R. W. Murray, T. J. SchaaV, J. T. Khoury, M. M. Alvarez and 88 G. P. Smith and V. A. Petrenko, Chem.Rev., 1997, 97, 391. 89 L. A. Lyon, M. D. Musick and M. J. Natan, Anal. Chem., 1998, R. L. Whetten, Science, 1998, 280, 2098. 95 H. S. Zhou, T. Wada aJ. Chem. Soc., Chem. 70, 5177. 90 M. Brust, M. Walker, D. Bethell, D. J. SchiVrin and R. Whyman, Commun., 1995, 1525. J. Chem. Soc., Chem. Commun., 1994, 801. 91 M. Brust, J. Fink, D. Bethell, D. J. ShiVrin and C.Kiely, J. Chem. Soc., Chem. Commun., 1995, 1655. Paper 8/06540C 1054 J. Mater. Chem., 1999, 9, 1043–1054 J O U R N A L O F C H E M I S T R Y Materials Feature Article Synthetic materials capable of reporting biomolecular recognition events by chromic transition Patrick Englebienne Department of Nuclear Medicine, Free University of Brussels, Brugmann Hospital, Place Van Gehuchten 4, B-1020 Brussels, Belgium. E-mail: patrick.englebienne@skynet.be Received 19th August 1998, Accepted 8th February 1999 This article presents a survey of synthetic materials capable been primarily explored as new controlled drug-delivery sysof reporting in real time by chromic transition the recog- tems.1 However, the modifications experienced by these matenition event occurring between a specific ligand and a rials from ground to new environmental conditions can be as biomolecule such as a receptor or an antibody.The review varied as a change in volume,4 in color,5 from solid to liquid, focuses on conductive polymers and colloidal gold. Besides from water-soluble to water-insoluble,6–9 from extinct to light their mode of synthesis and coupling to biomolecules, it emitting.10–13 Consequently, many of the modifications in their explores their physicochemical reactivity, emphasizing the properties as induced by changes in their environmental conmechanisms underlying the chromic transition occurring ditions can be detected by common means, which make them upon the biomolecular recognition event.The future pros- suitable candidates for transducer elements for biosensors.The pects offered by composite materials based on these concept is illustrated in Fig. 2. Some of these materials have polymers and colloids are further discussed. been used successfully in various direct biosensing systems in which the binding event induced respectively either a change Introduction in solubility,14 a change in volume,15 or the fusion,16 of the polymeric materials.Among these new materials, those Biomolecular recognition occurs when a biomolecule such as reporting a change in their environmental conditions by a a receptor or an antibody meets its specific ligand. The chromic transition are of major interest for application in encounter leads to the binding of one species by the other. direct biosensors because the change in their optical properties The biomolecular recognition event is usually monitored by occurring upon a biomolecular recognition event can be the introduction of a label into one of the complementary detected by a simple photometer and sometimes even by the binding molecules, which transduces the recognition event into naked eye.Among such materials, conductive polymers and a signal.colloidal solutions of gold particles are very promising and As a result of important advances made these last few years will be considered in this article. in the field of synthetic polymers and colloids usually qualified as ‘intelligent’ in the literature,1 the means for monitoring biomolecular recognition events have evolved quite tremen- Conductive polymers dously.These advances have allowed the design of eYcient biosensors capable of a high level of user-friendliness and Nature and properties automation.2 A biosensor is a device comprising either a ligand or a Conductive polymers are extensively conjugated organic macromolecules made of repetitive sequences of respectively receptor (antibody) species coupled to a signal transducer, which detects the binding of the complementary species.As alkyne, phenyl or heterocyclic monomers. These polymers can gain near-metallic conductivity by oxidation (p-doping) or illustrated by the sketch presented in Fig. 1, an indirect biosensor uses a separate labelled complementary species which reduction (n-doping) from their insulating state.17 The structures of the most extensively studied conductive polymers, is detected after the binding event has occurred, by measuring the activity of the label (i.e.radioactivity, fluorescence, chemi- along with the conductivities they can attain, are displayed in Fig. 3. For comparison, the conductivity of metallic copper is luminescence) associated with the complex. In the field of immuno- and receptor-assay technology, an indirect biosensing 5.8×107 S cm-1.18 Another important feature common to the conductive polymers is their capacity to undergo chromic system is termed heterogeneous because the ligand–receptor complex must be sequestered from the free fractions before transitions upon changes in their oxidation state.19 The conductive polymers are intrinsically semi-conductors in which recording the signal of the label.On the other hand, a direct biosensor (i.e. homogeneous immuno- or receptor-assay) com- the highest occupied (HOMO) and lowest unoccupied molecular orbitals (LUMO) can give rise to valence (p) and conduc- prises a ligand or a receptor respectively coupled to a signal transducer detecting in real time the binding of the complemen- tion (p*) bands.The energy gap (Eg) between the HOMO and LUMO electron bands determines the intrinsic electrical and tary species by a change in potential diVerence, current, resistance, mass, heat, or optical properties.3 optical properties of a given polymer.20 The doping process modifies the electronic band structure of the polymer by Chemical materials able to serve as transducers in direct biosensors are currently hot topics in biomedical research.The producing new electronic states that are localized in the band gap (polarons and bipolarons) and cause its passage from so-called ‘intelligent’ materials are particularly interesting candidates for direct biosensing systems. These are polymers or insulating to conducting as well as its chromic transition.21,22 In cyclic conductive polymers, these transitions are indicative colloids sensitive to external stimuli, which experience respective changes in their structure or in either their chemical or of the passage from aromatic to quinonoid molecular forms.The process of interband electronic transition is schematically physical properties in response to changes in their environmental conditions such as temperature, mass, pH, light, electric illustrated in Fig. 4. In chemical terminology, a polaron is a radical ion (spin 1/2) and a bipolaron is a dication (pair of field, or oxido-reduction. As a result of their physical responsiveness to temperature, many of these new materials have like charges, spinless). In their insulating conformation, the J. Mater.Chem., 1999, 9, 1043–1054 1043Fig. 1 Sketch emphasizing the diVerence between indirect (heterogeneous) and direct (homogeneous) biosensing systems. Fig. 2 Schematic diagram illustrating the concept of applying ‘intelligent’ synthetic materials in direct biosensing systems for biomolecular recognition. Eg of conductive polymers is usually above 1.5 eV. In the case of poly(pyrrole), it levels at 3.2 eV and by increasing p-doping it decreases progressively to 2.7 and 1.0 eV, respectively.20 When compared to other cyclic conductive polymers, poly(aniline) presents some unique structural features which allow for the control of its conductive and optical properties.23 This control results from the possibilities that exist to finetune the level of oxidation and protonation of poly(aniline).These possibilities are the consequence of the many structural variations due to the alternation in the polymer backbone of p-phenylene rings with the nitrogen atoms. The degree of oxidation depends on the fraction of these latter which is imine or amine. This is exemplified by the two extreme structures in Fig. 5A and B, with respectively a 0 and 100% NH n Structure Name Conductivity / S cm-1 Poly(acetylene) 1.5 x 105 Poly(pyrrole) 7.5 x 103 S n n Poly(thiophene) 103 n Poly( p-phenylene) 103 N N n Poly(aniline) 200 degree of oxidation.In between these two extreme states, the Fig. 3 Structures and conductivities of the most extensively studied conductive polymers. degree of oxidation may be varied in order to aVord numerous 1044 J.Mater. Chem., 1999, 9, 1043–1054Fig. 4 Schematic illustration of the energy level of one electron organic molecule in its electronic configuration adopting the equilibrium geometry of respectively (a) its ground state, (b) its first ionized state, (c) its polaron and (d) bipolaron states. Fig. 6 Electronic absorption spectra of colloidal poly(aniline) respectively in its oxidized (dotted line, Eg 2.15 eV) and n-doped state (solid line, Eg 1.45 eV).intermediate structures each displaying diVerent conductivities and electronic absorption spectra upon external protonation. produce self-doped polymers. The process of self-doping is When an anionic substituent capable of protonation is introillustrated in Scheme 1b with poly(thiophene-3-ethanesulfon- duced on the p-phenylene ring, the polymer structure may be ate) as the example.The advent of water-soluble and self- in a partial conductive form when the substituent is protonated doped conductive polymers allowed for their direct use in (Fig. 5C). This latter structure transforms into the fully conbiochemical reactions, particularly as electron transfer means ductive form upon self-doping (Fig. 5D) resulting from the in enzyme sensors.30,35 loss of the proton by the substituent.24 The Eg and chromic transitions of poly(aniline) upon n-doping are illustrated by Synthesis the electronic absorption spectra in Fig. 6. Another class of conjugated polymers worth mentioning is Whether soluble or insoluble in water, most conductive the substituted poly(diacetylenes).As illustrated in Fig. 7, the polymers are synthesized by electrochemical or chemical structure of these polymers is made of ene–yne alternating means. These synthetic procedures have been extensively bonds. The molecules of this polymer class are particularly reviewed, in particular for poly(thiophene).36 Whether chemi- flexible and they undergo important color changes when cal or electrochemical, the most commonly used synthetic subjected to planar–nonplanar conformational transitions.25 processes occur through the oxidation of the monomers.The Up to the mid eighties, all the conductive polymers available most commonly used chemical oxidants are iron(III ) chloride were soluble only in organic solvents,26–28 and their use in and ammonium persulfate.aqueous biological systems was consequently restricted to that When substituted heterocyclic monomers are used as starting of electronic transfer films incorporated in amperometric sen- blocks, oxidative polymerization presents the drawback of a sors.29–31 These water-insoluble polymers can only be doped lack of control of the structure of the oligomers or polymers by the addition of oxidative dopant counter-ions which diVuse formed. As shown in Scheme 2, the resulting substituted in and out of the polymer backbone to balance the charges polymers are structurally inhomogeneous.37 This may be of created.32 This process is called external doping and is illus- importance for the application in biomolecular recognition trated in Scheme 1a with poly(thiophene) as an example.In because there is experimental evidence that the regiochemistry the late eighties, it was observed that the introduction of an of substituted poly(heterocycles) controls their conformational alkanesulfonate sidechain in cyclic or heterocyclic monomers features, which in turn, governs the degree of p–p conjugation yielded water-soluble conductive polymers.33,34 Moreover, between adjacent rings.36 This control has consequently a because the counter-ion is attached to the polymer through a covalent bond, these water-soluble polymers can lose a proton upon oxidation, simultaneously with the loss of an electron to HN HN n A N N n B N N n HN HN n COOH COOH COO- COOC D Fig. 5 Extreme structures of poly(aniline) with an oxidation degree of 0 and 100% (respectively A and B).The introduction of a carboxylic substituent on the p-phenylene ring transforms the polymer into a partially conductive form (C), capable of aVording the fully conductive S n + yMX a. b. S n SO3H S n SO3 - - nH+ + nH+ S n y+ (X-) + yM Oxidation Reduction Scheme 1 form (D) upon self-doping. J. Mater. Chem., 1999, 9, 1043–1054 1045C C C C R R' n R = R' = O n NH O O O x n = 3 and x = 1 or 3 R = R' = OH O Colour Yellow R = R' = O n NH O O O x n = 4 and x = 1 or 3 Yellow Blue R = R' = NH O Purple OH R = R' = OH Orange R = R' = OH O Blue R = R' = O O Red Fig. 7 Examples of substituted poly(diacetylene) structures and their corresponding colour in the coplanar conformation. tremendous influence on the capacity of the polymer to translate a departure from coplanarity by a chromic transition. 38 Poly(thiophene) derivatives are in this respect of particular interest because of their specific photochemical properties.11,12 As discussed below, such departure from coplanarity is an important mechanism of transducing biomolecular recognition. As shown in Scheme 3, control of the polymer architecture can be achieved by coupling 2-bromo-3- alkyl-5-thienylmagnesium bromide in the presence of a nickel catalyst.39 Other strategies using the cross coupling of 2- bromo-3-alkylthiophene with 2-(trimethylstannyl )-3-alkylthiophene monomers have also been reported to provide a similar advantage.40 It has been claimed that water-soluble poly(anilinesulfonate) could not be synthesized by chemical or electrochemical oxidation from o-aminobenzene- or m-aminobenzene-sulfonic acid because the substitution of a hydrogen atom on the phenyl ring by electron withdrawing substituents like -SO3H led to an increase of the steric hindrance and to a strongly deactivating influence likely to limit the polymerization process.Until recently, this problem was addressed by sulfonating the already polymerized aniline with fuming sulfuric acid.24 S R MgBr Br Ni(dppp)Cl2 S R S R n S R FeCl3/(NH4)2S2O8 or e- at Pt S S R R S R S R S R n S R S R n S R S R S R S R n S R S R S R S R n Scheme 2 Scheme 3 1046 J.Mater. Chem., 1999, 9, 1043–1054Fig. 8 DiVerence electronic spectra recorded at various times (3, 5, 120, 192 and 288 h; arrow) during the polymerization of o-anthranilic acid in aqueous phase using ferric chloride as the oxidant.This strategy leads however to ill defined and partially sulfonated polymers. The question is still under debate and is quite controversial. In a recent paper,41 the synthesis of fully sulfonated poly(aniline) from o- and m-aminobenzenesulfonic acid by chemical oxidation in the liquid phase under high pressure has been reported.Poly(aniline) substituted with a carboxylic acid group has been obtained by oxidative polymerization of Fig. 9 Evolution of the A530 nm of a diluted buVered solution of o-anthranilic acid with ammonium persulfate, aVording a poly(o-anthranilic acid) as a function of pH between pH 6.5 and 9. water-soluble polymer.42 However, the UV–vis spectra of the polymerized material showed no absorption in the visible region and the authors concluded that the carboxylic acid protons and hence to pH. A decrease in pH enhances doping substituent restricted the p-conjugation along the polymer and conversely, pH increases progressively undope the poly- chain. However, the successful electropolymerization of substimer.The doping–undoping process induced by pH changes is tuted poly(anilines) from o- and m-aminobenzoic acid as well translated by important modifications in the absorption at the as from m-aminobenzenesulfonic acid has also been reported wavelength corresponding to the p–p* transition of the poly- and the polymers obtained were shown to be self-doped and mer.The pH eVect on a buVered solution of poly(o-anthranilic electrochemically active in aqueous solutions.43 In a more acid) is illustrated in Fig. 9. It is worth noting that when the recent publication,44 we have shown that conjugated and conductive polymers most eVective at translating a biomolecu- water-soluble poly(o-anthranilic acid) could be obtained by lar recognition process by chromic transition are considered, chemical oxidation with ferric chloride. Progressive p–p* transthe highest pH eVect is observed around physiological pH itions at lower energy levels (475 nm, 2.6 eV and 530 nm, values, such as in the example of Fig. 9. In other cases, the 2.3 eV) occur during the polymerization process as shown in range of pH sensitivity may be broad and this property can the diVerence spectra in Fig. 8, which are indicative of the be applied to the fabrication of reversible optical sensors progressive conjugation along the polymer backbone during for pH measurement such as recently reported with synthesis.The same approach allowed us to polymerize also poly(pyrrole).52 various substituted poly(thiophenes).44,45 Similarly, the respective addition of oxidants or of reductants Conductive polymers have also been synthesized in colloidal to the polymers at a fixed pH exerts a profound influence on form, particularly poly(aniline) and poly(pyrrole).46–48 These the absorbance (A) at the p–p* transition wavelength.This materials can only be externally doped since they do not eVect is illustrated with poly(o-anthranilic acid) in Fig. 10. contain any counter ion covalently linked to the polymer The addition of increasing millimolar concentrations of backbone.One exception to this is the colloidal copolymers ammonium persulfate to the polymer solution buVered at of poly(pyrrole) and poly(styrenesulfonate), the sulfonate pH 9 increases the A530 nm and conversely, the addition of the counter-ions being incorporated into the poly(pyrrole) during same increasing concentrations of sodium borohydride to the the polymerization providing cation-exchange sites.49 This polymer solution buVered at pH 7 decreases the A at the same latter copolymerization approach could generate in the near wavelength in a corresponding way.This capacity of conduc- future new composite materials with promising properties for tive polymers to respond spectrophotometrically to oxido- application in the biomolecular recognition field.reduction has been applied to the quantitative detection of Finally, substituted poly(diacetylenes) on the one hand can ascorbic acid using poly(thiophene) films.5 We have adapted be produced by 1,4 addition of the substituted diacetylenic the technique in an automated clinical chemistry analyzer monomers initiated by either c25 or UV50 irradiation.On the using poly(o-aminobenzenesulfonic acid) and poly(o-anthra- other hand, soluble derivatives of poly(acetylene) have been nilic acid) in aqueous solution. In these conditions, the polymer obtained by ring-opening metathesis polymerization of reduction by ascorbic acid monitored by photometry occurs monosubstituted cyclooctatetraenes.51 in less than 10 min. The chromic sensitivity of conductive polymers to both pH Chromic transition of water-soluble conductive polymers upon and oxido-reduction could lead in the near future to the oxido-reduction availability of cheap analytical reagents or instruments, finding application in clinical chemistry, pharmacology, biotechnology As a consequence of their self-doped nature, water-soluble conductive polymers are highly sensitive to the presence of or the food industry.J. Mater. Chem., 1999, 9, 1043–1054 1047H2N N H N N NH R-NH2 O O O O O O NH NH N N NH NH R n + R = biomolecular species n Scheme 4 merization with bromoacetyl bromide, depending on the synthetic conditions. These derivatives can then further be either carboxylated (Scheme 6c) or aminated (Scheme 6d) by being Fig. 10 EVect of increasing concentrations of either ammonium reacted with either mercaptoacetic acid or triethylenetetramine, persulfate (triangles) or sodium borohydride (circles) on the A at respectively.54 The derivatized polymer can then be linked 530 nm of poly(o-anthranilic acid) diluted in 50 mM phosphate buVer adjusted respectively at pH 9 or 7. covalently to ligands, receptors or antibodies. Chromic transition of conductive polymers upon biomolecular Covalent coupling of water-soluble conductive polymers to recognition biomolecules Colloidal conductive polymers have been used as early as in The covalent linkage between one of the biomolecular reactants 1992 by Tarcha’s group at Abbott Laboratories as reporter and the water-soluble conductive polymer label is of prime reagents in solid-phase heterogeneous immunoassays.55 In this importance because it is supposed to secure the close proximity particular application, advantage was taken of the deep colour of the label to the biomolecular recognition site. Consequently, of the colloidal particles so as to detect either by densitometry zero-length linkages or small bridges are usually preferable to or visually any binding of antibody coated particles to the long connecting arms.This prerequisite is of less importance for water-insoluble polymers since the mechanism of transducing the biomolecular recognition event is likely to proceed from diVerent principles which are discussed in the next section. Most molecules used in biomolecular recognition are proteins, peptides or sugars which contain amine, hydroxy, mercapto or carboxylic groups available for coupling to conductive polymer labels.When the molecule to be labelled is a small ligand such as a steroid or a drug which does not contain vacant functionalities, it must be chemically modified prior to coupling. As shown in Scheme 4, poly(aniline) contains terminal amines which can be used in a cross-linking reaction to corresponding amines in proteins or peptides through glutaraldehyde.Except for poly(aniline), most conductive polymers do not have corresponding groups available for coupling. Consequently, substituting the polymer backbone is useful not only for granting self-doping capacity and water solubility, but also for providing vacant functionalities for their covalent coupling to ligands and receptors. For instance, the substitution of poly(thiophene) with carboxylic acid groups allows for its covalent linkage in aqueous solutions to amine residues (i.e.e-amino groups of lysine in proteins) in presence of a water-soluble carbodiimide such as 1-ethyl-3-(3-(dimethylamino) propyl )carbodiimide. The carbodiimide forms a reactive carboxylic anhydride intermediate (Scheme 5) which further reacts with the amine to form an amide bond.53 Alternatively (Scheme 6), conductive polymers which contain a reactive heterocycle such as poly(pyrrole) can be S COOS COOn + N N N S C S COOH n O O N HN N R NH2 R = biomolecular species S C S COOH n HN O R + N N H NH O Scheme 5 either N- (Scheme 6a) or C- (Scheme 6b) acylated after poly- 1048 J.Mater. Chem., 1999, 9, 1043–1054NHn + BrH2C CBr O D N n C CH2Br O N n C CH2Br O + HS C H2 C OH O N n C CH2 O S CH2 COOH NHn + BrH2C CBr O AlCl3 NHn C O CH2Br a. b. c. NHn C O CH2Br d. + H2N HN NH NH2 NHn C O CH 2 NH HN NH NH3 Br- Scheme 6 antigen previously captured by the solid phase. The mechanism receptor binds the ligand, the film changes colour from blue to red.The red colour intensity is directly proportional to the of indicating the biomolecular recognition in this kind of application does not depart from the use of a coloured quantity of virus reacted with the film. The system consequently constitutes a direct colorimetric biosensor. In another poly(styrene) latex and is consequently beyond the scope of the present discussion. article,60 the authors reported the colorimetric detection of the molecular recognition between the Cholera toxin and Gm1 For the sake of clarity, I will arrange the next discussion around the regioregularity of the conductive polymers con- ganglioside incorporated in liposomes made of a substituted poly(diacetylene).In this latter system, the intensity of the sidered.The mechanism of transducing biomolecular recognition by chromic transition is indeed likely to diVer depending color change was directly proportional to the dose of toxin interacting with the ligand and the least detectable dose was on the degree of regioregularity of the polymeric structures. Substituted poly(diacetylene) derivatives are regioregular in the mg l-1 range.Other examples of application have been reported with substituted poly(diacetylenes) derivatized with and undergo dramatic color changes upon planar–nonplanar conformational transitions induced by either a change in toxin- or virus-specific ligands, which undergo a blue–red chromic transition upon ligand recognition by the patho- temperature (thermochromism), solvent composition (solvatochromism) or mechanical stress (mechanochromism).25,50 As gen.61,62 A departure from coplanarity has also been observed when a biotin-functionalized regioregular substituted poly(thi- further shown in Fig. 7, the side-chain structure plays a critical role in determining the original colour of the polymer in the ophene) reacts with the biotin receptor (avidin) in aqueous solution.This eVect results in the polymer solution turning planar conformation of the ene-yne backbone, and hence on the type of chromic transition occurring upon departure from from violet to yellow.63 The transduction of biomolecular recognition by chromic coplanarity.25,50 Similar eVects have been observed with regioregular substituted poly(acetylenes)51 and poly(thiophenes)38,56 transition resulting from specific changes in the conjugated polymer backbone conformation is possible with well which can undergo severe thermochromic, solvatochromic and ionochromic transitions. In poly(thiophenes), the departure organized regioregular polymers.The mechanism of chromic transition by random polymers in the same circumstances is from a coplanar conformation decreases the conductivity of the polymer57 and can be counteracted by the rigidification of less clear.We have reported the synthesis of several watersoluble conductive random polymers or oligomers and their the p-conjugated system.58 Such chromic transition resulting from a decrease in the eVective conjugation length of the use as transducing reagents acting by chromic transition in homogeneous competitive immunoassays for antigens and polymer backbone has also been observed when the weight stress due to a biomolecular species binding its counterpart is haptens.44,45 Currently, these assays allow detection of ligands in the nanomolar range.45 Fig. 11 illustrates the performance applied on a given polymeric structure. Such application of molecular mechanochromism has been reported59 with the use of these reagents by showing the kinetics of the chromic transition occurring when theophylline labelled with poly(thi- of a Langmuir–Blodgett film made of a polymerized diacetylenic lipid matrix functionalized with a sialic acid ligand specific ophene-3-carboxylic acid) associates with a specific antibody and its eventual dissociation from the complex by an excess for the influenza virus hemagglutinin.When the hemagglutinin J. Mater. Chem., 1999, 9, 1043–1054 1049Fig. 12 Schematic drawing illustrating the capacity of water-soluble conductive polymers to translate directly a biomolecular recognition event by a change in their visible absorption spectrum under the influence of the local pH existing around the binding pocket.The conductive polymer is represented by the circles, respectively in its ground-state colour (hatching) and transition-state colour (black). The ligand and receptor are respectively represented by a triangle and Y. Fig. 11 Kinetics of association and dissociation between theophylline suspended in water. Under normal circumstances, the metallic labelled with poly(thiophene-3-carboxylic acid) and a specific anti- gold is surrounded by a predominantly anionic double layer.theophylline antibody transduced by the chromic transition of the The sol is very stable due to the repulsive forces and shows a label as measured in a clinical chemistry automated analyzer at deep red color. However, due to induced changes in the outer 340 nm. The dissociation (arrow) is induced by the addition of a 100 charges, the particles are agglutinated and coagulated by the fold excess of unlabelled theophylline.addition of millimolar concentrations of various salts and this irreversible process changes the colour from red to blue.66 Colloidal gold hydrosol can be protected from agglutination of unlabelled theophylline. The rapidity of the kinetics makes it possible to measure the signal in a random access clinical by salts by coating the particles with a protein layer.Gold hydrosols are characterized by a discrete band in the chemistry analyzer. These instruments, which are widely used in clinical laboratories, are walk-away instruments capable of visible electronic spectrum which is called the surface plasmon resonance (SPR). It is a quantized plasma oscillation occurring mixing and incubating reagents and samples in individual cuvettes according to programmed conditions.They are at the surface of the gold particles, resulting in a characteristic peak location and width at maximal A.67 The SPR wavelength equipped with a flash lamp photometer allowing monitoring of subtle changes of A in each cuvette at a given wavelength and width depend on the refractive index of the particles and hence on their average diameter.Any increase in average and at short regular intervals (usually 25 s).64 The potential application of conductive polymers to homogeneous immuno- particle diameter in the sol induces a red shift of the SPR wavelength along with a decrease in maximal A. This is assays that can be run in such automated instruments constitutes an interesting and quite revolutionary challenge for the illustrated in Fig. 13 which shows the change occurring in the visible electronic spectrum of colloidal gold when the particles next few years. Because of the lack of regularity in the polymeric chains used are coated with a small layer of human serum albumin. We have applied this characteristic of colloidal gold in quantitative in such applications, the molecular mechanism inducing the chromic transition cannot be assigned only to a departure from colorimetric assays for proteins.68,69 coplanarity.As discussed previously, these water-soluble polymers are easily doped or undoped by subtle pH changes, and Synthesis we have suggested that the chromic transition depends on the Isodisperse colloidal gold hydrosols are synthesized by the local pH existing around the binding pocket of the receptor or controlled reduction of an aqueous solution of tetrachloroauric the antibody, which is known to change upon ligand recogacid. The reduction of Au3+ ions produces a supersaturated nition.65 This mechanism is illustrated by the schematic drawing molecular Au0 solution and as the Au0 concentration increases in Fig. 12. Such an explanation is further supported by the fact the gold atoms cluster and form seeds of nuclei. Particle that we were unable to apply the technique to sandwich assays growth occurs by further deposition of metallic gold upon using labelled antibodies, most probably because the conductive the nuclei.70 polymer label attached to an antibody molecule is situated too Strong reductants like white phosphorus or sodium far from the binding pocket to sense and transduce the pH borohydride produce a great number of nuclei and hence change occurring upon ligand recognition.rather small particles with average diameter of 2 to 10 nm depending on the synthetic conditions. Bigger particles (20 to Colloidal gold 100 nm average diameter) can be obtained by using a weaker reductant like sodium citrate advocated by Frens.71 The size Nature and properties of the particles obtained by this method depends on the citrate5tetrachloroauric acid ratio.The lower the ratio, the Gold hydrosols are hydrophilic colloids made of monodisperse, finely divided particles (5–100 nm diameter) of metallic gold greater the diameter of the particles. 1050 J.Mater. Chem., 1999, 9, 1043–1054Fig. 13 Visible electronic spectrum of a colloidal gold hydrosol (solid line) compared to that of the same sol in which the particles are coated with a small layer of human serum albumin (dashed line). The change in refractive index due to the protein coating induces a red shift and a decrease in A at the SPR wavelength.Coating colloidal gold particles with proteins The coating of colloidal gold nanoparticles with proteins is performed by electrostatic charge adsorption. Colloidal gold nanoparticles remain negatively charged over a wide range of pH. When a partially protonated protein is contacted with the Fig. 14 EVect of coating pH on the capacity of the antibody-coated particles at a suitable pH, the positive charges on the protein gold probes to recognize the specific ligand.The decrease in A at 600 nm after salt addition to the coated particles (open circles, left are attracted by the anionic layer surrounding the particles axis) shows stabilization of the colloid between pH 6.5 and 8. The with which it forms ionic bonds. The stability granted to the capacity of the particles to bind a small amount of the ligand, as hydrosol by the protein coating process is usually verified by measured by the increase of A at 600 nm after incubation which is the absence of coagulation upon addition of a concentrated representative of the refractive index change (solid triangles, right salt solution.72 axis) peaks at a narrower coating pH range of 7–7.5, indicating proper When colloidal gold is coated with proteins capable of antibody orientation at this pH.biomolecular recognition such as receptors or antibodies, the selection of the pH for coating is most critical. This parameter mic transition resulting from the selective agglutination of the allows control of the part of the protein which is positively particles has been also applied to the detection of charged and hence determines the correct orientation of the complementary polynucleotide strands.75,82 external binding site of the particle for proper ligand recog- Recently,83 I have shown that a lattice formation was not nition.73,74 This point is illustrated in Fig. 14 with the stability an absolute requisite for observing a measurable SPR shift and reactivity of colloidal gold particles coated by an antibody and that such a shift could be observed when colloidal gold at various pHs.The decrease in A600 nm after salt addition particles coated with a monoclonal antibody specific for a indicates that the sol is stabilized for a large range of pH (6.5 single epitope bind the protein ligand. This is illustrated by to 8), although the peak of ligand recognition by the gold the artist’s view on the cover of this issue.In this case, the probes is much narrower and lies at coating pHs around shift in SPR wavelength reflects very small changes in the 7–7.5. Whilst the optimal pH for coating usually lies around refractive index at the particle surface which are the direct the isoelectric point of the protein, there are no general rules result of mass changes in the approximate medium induced and adsorption interaction isotherms such as that shown in by the biomolecular recognition event.The SPR wavelength Fig. 14 should be conducted in each case. However, once the shift phenomenon is further dose-related as illustrated by the optimal conditions are identified, the coating process is usually visible spectra shown in Fig. 15. These spectra were recorded reproducible and easy to scale-up. after the respective incubation of monoclonal antibody-coated colloidal gold particles with increasing concentrations of the Mechanism of chromic transition upon biomolecular recognition specific ligand, human chorionic gonadotropin (a pregnancy protein).The SPR wavelength shift of colloidal gold is used When colloidal gold probes coated with an antibody bind the specific ligand, a red shift occurs in the SPR wavelength of as transducer of the biomolecular recognition event of protein ligands in an automated clinical chemistry analyzer, the signal the gold sol along with a decrease in A at the maximal wavelength. When the antibody used for coating recognizes being measured at wavelengths in the red part of the spectrum (600 nm), where both the wavelength shift and the increase in various molecular determinants on the ligand, the gold probes progressively form a lattice of agglutinated particles around peak width resulting from changes in the refractive index of the particles enhance the A in a dose-dependent manner.ligand molecules and the SPR shift reflects the apparent increase in diameter of the individual particles in the solution Fig. 16 shows the typical dose–response curve of a homogeneous immunoassay for human ferritin performed in a which turns blue.75 The remarkable deep red colour and capacity of colloidal gold to undergo a chromic transition clinical chemistry analyzer.This new application of the chromic transition of colloidal gold upon biomolecular recognition upon agglutination have made possible its use for many years as a reporter reagent in either heterogeneous enzyme- is likely to have a tremendous potential, not only as a bioanalytical tool, but also as a real-time biosensing system like immunoassays76–78 or in homogeneous agglutination immunoassays,77,79–81 respectively.The same principle of chro- providing kinetic83–85 and even conformational86 information J. Mater. Chem., 1999, 9, 1043–1054 1051on the biomolecular interaction such as those obtained in more sophisticated systems like the BIAcore instrument. This new applicability of the technique confirms its interesting potential in the new drug discovery process, particularly for either the selection of the highest aYnity ligands87 or the detection of natural ligands and of orphan receptors,88 respectively.Colloidal gold has also been incorporated into a classical SPR biosensing system,89 comprising the interaction between an antibody immobilized on a gold film and the antigen coated on colloidal gold particles. The presence of the colloid increased tremendously the SPR sensitivity to protein–protein interactions.Future prospects One of the problems usually pointed out in the use of colloidal gold as a chromic transition reagent is the lack of possible covalent linkage of the biomolecular species to the probes. When they are stored in an insuYciently stabilizing medium, the particles coated by charge adsorption are prone to the progressive release of the adsorbed protein in the surrounding medium, which decreases their reactivity.This problem can be circumvented by the use of thiol-derivatized gold nanoparticles, the synthesis of which has been reported in two-phase90 as well as single-phase91 liquid systems. The attachment of a monolayer of bifunctional organic thiol molecules to the gold nanoparticles provides a vacant functionality for the covalent linkage of proteins.91 In a similar way, gold colloids have been Fig. 15 Visible electronic spectra recorded after incubation of mixtures stabilized by molecules containing amine functional groups,92 containing a fixed amount of colloidal gold particles coated with a specific monoclonal antibody and increasing concentrations (respect- including poly(amidoamine) dendrimers.93 This latter process ively no ligand added, solid spectrum; 33 pmol dm-3, dashed spec- allows for the amine groups to be available at the external trum; 60 pmol dm-3, dotted spectrum) of the specific ligand protein, surface of the particles. human chorionic gonadotropin.The quite recent discovery and progressive understanding of the redox chemical character of alkanethiolate monolayer protected colloidal gold nanoparticles, which exhibit HOMO–LUMO Eg of 0.4 to 0.9 eV94 open new prospects for application in biosensors.In the light of these new perspectives, I have attempted the template polymerization of substituted poly(anilines) on colloidal nanoparticles according to a scheme similar to that already reported for poly(diacetylene).95 As shown in Fig. 17, a suitable selection of the conductive polymer with a p–p* transition absorption wavelength matching that of colloidal gold SPR aVords a composite colloidal material displaying a tremendously enhanced absorption at the SPR wavelength. The use of such a material, which combines the capacities for chromic transition of both gold and watersoluble conductive polymers, could allow for a further Fig. 16 Dose–response curve of an homogeneous immunoassay for human ferritin using antibodies labelled with colloidal gold. The assay Fig. 17 DiVerence visible spectrum of o-anthranilic acid polymerized is run in an automated clinical chemistry analyzer (Cobas-Mira) with an incubation time of 4 minutes. Data are the average ±sd of on colloidal gold nanoparticle templates versus that of the underivatised gold colloid.duplicate measurements. 1052 J. Mater. Chem., 1999, 9, 1043–105441 H. S. O. Chan, A. J. Neuendorf, S.-C. Ng, P. M. L. Wong and improvement in the level of analytical sensitivity attained by D. J. Young, Chem. Commun., 1998, 1327. the current homogeneous liquid-phase biosensors based on 42 M. T.Nguyen and A. F. Diaz, Macromolecules, 1995, 28, 3411. these independent chromic transducers. 43 A. A. Karyakin, A. K. Strakhova and A. K. Yatsimirsky, J. Electroanal. Chem., 1994, 371, 259. 44 P. Englebienne and M. Weiland, J. Immunol. Methods, 1996, 191, Many thanks to my son Gwenn for his excellent design and 159. computer generation of the three-dimensional figure of the 45 P.Englebienne and M.Weiland, Chem. Commun., 1996, 1651. cover. 46 B. Vincent and J. Waterson, J. Chem. Soc., Chem. Commun., 1990, 683. 47 N. Gospodinova, P. Mokreva and L. Terlemezyan, J. Chem. Soc., References Chem. Commun., 1992, 923. 48 S. P. Armes and B. Vincent, J. Chem. Soc., Chem. Commun., 1 I. C. Kwon, Y. H. Bae and S. W. Kim, Nature, 1991, 354, 291. 1987, 288. 2 J. P. Gosling, in Immunotechnology, eds. J. P. Gosling and 49 Z. Qi and P. G. Pickup, Chem. Commun., 1998, 15. D. J. Reen, Portland Press, London, 1993, p. 91. 50 S. Okada, S. Peng, W. Spevak and D. Charych, Acc. Chem. Res., 3 C. L. Morgan, D. J. Newman and C. P. Price, Clin. Chem., 1996, 1998, 31, 229. 42, 193. 51 C. B. Gorman, E. J. Ginsburg and R. H. Grubbs, J. Am. Chem. 4 G.Chen and A. S. HoVman, Nature, 1995, 373, 49. Soc., 1993, 115, 1397. 5 H. S. O. Chan, S. C. Ng and S. H. Seow, Synth. Met., 1994, 52 R. Koncki and O. S. Wolfbeis, Anal. Chem., 1998, 70, 2544. 66, 177. 53 N. Nakajima and Y. Ikada, Bioconjugate Chem., 1995, 6, 123. 6 R. Yoshida, K. Uchida, Y. Kaneko, K. Sakai, A. Kikuchi, 54 M. R. Pope, S. P. Armes and P. J. Tarcha, Bioconjugate Chem., Y.Sakurai and T. Okano, Nature, 1995, 374, 240. 1996, 7, 436. 7 Y. G. Takei, T. Aoki, K. Sanui, N. Ogata, T. Okano and 55 P. J. Tarcha, D. Misun, D. Finley, M. Wong and J. J. Donovan, Y. Sakurai, Bioconjugate Chem., 1993, 4, 42. in Polymer Latexes, ed. E. S. Daniels, E. Sudol and 8 G. Chen and A. S. HoVman, Bioconjugate Chem., 1993, 4, 509. M. S. El-Aasser, American Chemical Society, 1992, p. 347. 9 A. Chilkoti, G. Chen, P. S. Stayton and A. S. HoVman, 56 I. Le�vesque and M. Leclerc, J. Chem. Soc., Chem. Commun., Bioconjugate Chem., 1994, 5, 504. 1995, 2293. 10 M. Berggren, O. Ingana�s, G. Gustafsson, J. Rasmusson, 57 M. Leclerc and G. Daoust, J. Chem. Soc., Chem. Commun., 1990, M. R. Andersson, T. Hjertberg and O. Wennerstro�m, Nature, 273. 1994, 372, 444. 58 J. Roncali, C. Thobie-Gautier, E. H. Elandaloussi and P. Fre` re, 11 B. S. Kang, M. L. Seo, Y. S. Jun, C. K. Lee and S. C. Shin, Chem. J. Chem. Soc., Chem. Commun., 1994, 2249. Commun., 1996, 1167. 59 D. H. Charych, J. O. Nagy, W. Spevak and M. D. Bednarski, 12 B. Xu and S. Holdcroft, J. Am. Chem. Soc., 1993, 115, 8447. Science, 1993, 261, 585. 13 R. H. Friend, R. W. Gymer, A. B.Holmes, J. H. Burroughes, 60 J. J. Pan and D. Charych, Langmuir, 1997, 13, 1365. R. N. Marks, C. Taliani, D. D. C. Bradley, D. A. Dos Santos, 61 A. Reichert, J. O. Nagy, W. Spevak and D. Charych, J. Am. J. L. Bre�das, M. Lo�gdlund and W. R. Salaneck, Nature, 1999, Chem. Soc., 1995, 117, 829. 397, 121. 62 W. Spevak, J. O. Nagy and D. H. Charych, Adv. Mater., 1995, 14 S. M. Barnard and D.R. Walt, Science, 1991, 251, 927. 7, 85. 15 P. J. Tarcha, W. Bindseil and V. P. Chu, J. Immunol. Methods, 63 K. Faý� d and M. Leclerc, Chem. Commun., 1996, 2761. 1989, 125, 243. 64 L. SchoeV, Anal. Chem., 1997, 69, 206R. 16 P. S. Stayton, T. Shimoboji, C. Long, A. Chilkoti, G. Chen, 65 C. S. Lee, P. Y. Huang and D. M. Ayres, Anal. Chem., 1991, J. M. Harris and A. S. HoVman, Nature, 1995, 378, 472. 63, 464. 17 R. L. Greene and G. B. Street, Science, 1984, 226, 651. 66 D. H. Everett, in Basic Principles of Colloid Science, Royal Society 18 D. Clery, Science, 1994, 263, 1700. of Chemistry, London, 1988, p. 30. 19 A. O. Patil, A. J. Heeger and F. Wudl, Chem. Rev., 1988, 88, 183. 67 J. L. CoVer, J. R. Shapley and H. G. Drickamer, J. Am. Chem. 20 J. L. Bre�das and G.B. Street, Acc. Chem. Res., 1985, 18, 309. Soc., 1990, 112, 3736. 21 S. Etemad, A. G. Heeger and A. G. MacDiarmid, Annu. Rev. 68 L. Ellens, G. Ameryckx and P. Englebienne, Meded. Fac. Phys. Chem., 1982, 33, 443. Landbouwwet. Rijksuniv. Gent, 1988, 53, 2035. 22 M. J. Nowak, S. D. D. V. Rughooputh, S. Hotta and A. J. Heeger, 69 L. Ellens, G. Ameryckx and P. Englebienne, Int. Biotech. Lab., Macromolecules, 1987, 20, 965. 1989, 7, 33. 23 W.-C. Chen and S. A. Jenekhe, Polym. Prepr., 1992, 33, 204. 70 J. E. Beesley, in Colloidal Gold: A New Perspective for 24 J. Yue, Z. H. Wang, K. R. Cromack, A. J. Epstein and Cytochemical Marking, Oxford University Press, Oxford, 1989, A. G. MacDiarmid, J. Am. Chem. Soc., 1991, 113, 2665. p. 5. 25 G. N. Patel, R. R. Chance and J. D. Witt, J. Chem. Phys., 1979, 71 J. De Mey, in Immunochemistry: Modern Methods and 70, 4387. Applications, ed. J. M. Polak and. I. M. Varndell, J. Wright and 26 K. Y. Jen, R. Oboodi and R. L. Elsenbaumer, Polym. Mater. Sci. Sons, Bristol, 1986, p. 115. Eng., 1985, 53, 79. 72 W. D. Geoghegan and G. A. Ackerman, J. Histochem. Cytochem., 27 J. E. Frommer, Acc. Chem. Res., 1986, 19, 2. 1977, 25, 1187. 28 S. Hotta, S. D. D. V. Rughooputh, A. J. Heeger and F. Wudl, 73 W. D. Geoghegan, J. Histochem. Cytochem., 1988, 36, 401. Macromolecules, 1987, 20, 212. 74 W. D. Geoghegan, J. Clin. Immunoassay, 1988, 11, 11. 29 N. C. Foulds and C. R. Lowe, Anal. Chem., 1988, 60, 2473. 75 R. Elghanian, J. J. StorhoV, R. C. Mucic, R. L. Letsinger and 30 T. Matsue, M. Nishigawa, T. Sawaguchi and I. Uchida, J. Chem. C. A. Mirkin, Science, 1997, 277, 1078. Soc., Chem. Commun., 1991, 1029. 76 J. H. W. Leuvering, P. J. H. M. Thal, M. van der Waart and 31 B. Fabre, G. Bidan and M. Lapkowski, J. Chem. Soc., Chem. A. H. W. M. Schuurs, J. Immunoassay, 1980, 1, 77. Commun., 1994, 1509. 77 T. J. C. Gribnau, J. H. W. Leuvering and H. van Hell, 32 M. Mastragostino, C. Arbizzani, A. Bongini, G. Barbarella and J. Chromatogr., 1986, 376, 175. M. Zambianchi, Electrochim. Acta, 1993, 38, 135. 78 J. H. W. Leuvering, P. J. H. M. Thal and A. H. W. M. Schuurs, 33 A. O. Patil, Y. Ikenoue, F. Wudl and A. J. Heeger, J. Am. Chem. J. Immunol. Methods, 1983, 62, 175. Soc., 1987, 109, 1858. 79 J. H. W. Leuvering, P. J. H. M. Thal, D. D. White and 34 Y. Ikenoue, Y. Saida, M. Kira, H. Tomozawa, H. Yashima and A. H. W. M. Schuurs, J. Immunol. Methods, 1983, 62, 163. M. Kobayachi, J. Chem. Soc., Chem. Commun., 1990, 1694. 80 J. H. W. Leuvering, B. C. Goverde, P. J. H. M. Thal and 35 T. Tatsuma, K. Ariyama and N. Oyama, Anal. Chem., 1995, 67, A. H. W. M. Schuurs, J. Immunol. Methods, 1983, 60, 9. 283. 81 J. H. W. Leuvering, P. J. H. M. Thal, M. van der Waart and 36 J. Roncali, Chem. Rev., 1992, 92, 711. A. H. W. M. Schuurs, J. Immunol. Methods, 1981, 45, 183. 37 G. Barbarella and M. Zambianchi, Tetrahedron, 1994, 50, 11249. 82 C. A. Mirkin, R. L. Letsinger, R. C. Mucic and J. J. StorhoV, 38 S. D. D. V. Rughooputh, S. Hotta, J. Heeger and F. Wudl, Nature, 1996, 382, 607. J. Polym. Sci., 1987, 25, 1071. 83 P. Englebienne, Analyst, 1998, 123, 1599. 39 R. D. McCullough, R. D. Lowe, M. Jayaraman and 84 M. M. Morelock, R. H. Ingraham, R. Betageri and S. Jakes, D. L. Anderson, J. Org. Chem., 1993, 58, 904. J. Med. Chem., 1995, 38, 1309. 40 G. Barbarella, A. Bongini and M. Zambianchi, Macromolecules, 1994, 27, 3039. 85 R. Nakamura, H. Muguruma, K. Ikebukuro, S. Sasaki, J. Mater. Chem., 1999, 9, 1043–1054 1053R. Nagata, I. Karube and H. Pedersen, Anal. Chem., 1997, 69, 92 D.V.LeV, L. Brandt and J. R. Heath, Langmuir, 1996, 12, 4723. 4649. 86 H. Sota, Y. Hasegawa and M. Iwakura, Anal. Chem., 1998, 70, 93 M. E. Garcia, L. A. Baker and R. M. Crooks, Anal. Chem., 1999, 71, 256. 2019. 87 G. Lowe,v., 1995, 24, 309. 94 S. Chen, R. S. Ingram, M. J. Hostetler, J. J. Pietron, R. W. Murray, T. J. SchaaV, J. T. Khoury, M. M. Alvarez and 88 G. P. Smith and V. A. Petrenko, Chem. Rev., 1997, 97, 391. 89 L. A. Lyon, M. D. Musick and M. J. Natan, Anal. Chem., 1998, R. L. Whetten, Science, 1998, 280, 2098. 95 H. S. Zhou, T. Wada and H. Sasabe, J. Chem. Soc., Chem. 70, 5177. 90 M. Brust, M. Walker, D. Bethell, D. J. SchiVrin and R. Whyman, Commun., 1995, 1525. J. Chem. Soc., Chem. Commun., 1994, 801. 91 M. Brust, J. Fink, D. Bethell, D. J. ShiVrin and C. Kiely, J. Chem. Soc., Chem. Commun., 1995, 1655. Paper 8/06540C 1054 J. Mater. Chem., 1999, 9, 1043–1054

ISSN:0959-9428    

DOI:10.1039/a806540c

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Communication Supramolecular architecture of a functionalized hexabenzocoronene and its complex with polyethyleneimine Andreas F. Thu�nemann,*a Dirk Ruppelt,a Shunji Itob and Klaus Mu�llenb aMax Planck Institute of Colloids and Interfaces, Kantstraße 55, 14513 Teltow, Germany. Email: andreas@terra.mpikg-teltow.mpg.de; Tel. 03328–46–271, Fax. 03328–46–272 bMax-Planck-Institut fu�r Polymerforschung, Ackermannweg 10, D-55128 Mainz, Germany Received 5th February 1999, Accepted 8th March 1999 A hexabenzocoronene, functionalized with a carboxylic acid group, was complexed with a hydrophobic modified polyethyleneimine and the thermotropic columnar liquidcrystalline- like complex was investigated by X-ray scattering and dynamic–mechanical analysis.The construction of disc-like molecules with large p-conjugated cores opens a way to new supramolecular aggregates,1,2 which are interesting because of their electronic and optoelectronic properties.3 It has been observed, for instance that pericondensed hexabenzocoronenes show diode-like current–voltage signals for single molecules4 and a rapid charge transport along self-assembled columns with high carrier mobilities.For an oriented layer with a thickness of 1 mm and an applied potential of 1 V the drift time of charges across such a layer was calculated to be less than one microsecond.5 The onedimensional nature of the charge transport makes these molecules promising as nanowires in molecular electronic devices or as building blocks for transport layers in electrocopying or electrophotography.A decisive aspect is to take advantage of their self-organizing capability particularly with respect to the control of the spatial orientation of molecular wires, as formed e.g. by hexakis(tetradecyl )hexa-peri-hexabenzocoronenes.5 The first step in this direction is the functionalization of such sheet-like molecules by an anchor group, which may, in principle, allow the immobilization of supramolecular aggregates.The binding of molecular columns to polyelectrolytes is a possibility to enhance their long-range order and stability. In this article we describe the phase behavior and mechanical properties of a hexabenzocoronene (b) which is functionalized by a carboxylic acid group and that of its complex with hydrophobic modified polyethyleneimine (a).The structures of the compounds are shown in Fig. 1. The synthesis of (b) will be described elsewhere.6 Fig. 1 The molecular structure of an ionically bound complex (c), In the last few years, it has been shown that the complexation formed by a hydrophobic modified polyethyleneimine (a) and a of polyelectrolytes with surfactants in water results in a number hexabenzocoronene (b).(a) and (b) are not drawn. 16% of the amino of new solid materials,7,8 e.g. with interesting optical9 and functions of (a) are alkylated by n-docosyl chains, (b) is functionalized surface-energy lowering properties.10–12 Analogously to poly- by a carboxylic acid function which serves as anchor group. The electrolyte–surfactant complexes we treated (b) as an amphiph- sketch at the bottom illustrates the columnar mesophases formed by ile and used (a) as a polyelectrolyte.Due to the insolubility of (b) and (c). For clarity only the aromatic cores of (a) are drawn. At (b) in water, the complexation with (a) was carried out in room temperature a rectangular centered lattice is found for both. The aromatic cores are tilted with respect to the column axis. At chloroform using an equimolar amount of (a) with respect to higher temperatures the columns are oriented in a hexagonal columnar amino and carboxylate functions.The resulting complex (c) mesophase. was solvent-cast as films and analyzed by IR spectroscopy, diVerential scanning calorimetry (DSC), X-ray scattering and dynamic–mechanical measurements.Comparison of the IR resulting in a stoichiometric complex (c). In the wide-angle Xray diagram of (a) at 20°C, sharp reflections were found at a spectra of (b) and (c) shows that the carboxylic acid vibration found for (a) at 1710 cm-1 is absent for (c) (see Fig. 2). scattering vector of s=2.39 nm-1 and 2.63 nm-1. This is attributed to a side-chain crystallinity of (a), in which the Instead of the sharp carboxylic acid band in the spectrum of (b) a broad diVuse band between 1500 and 1760 cm-1 is docosyl chains crystallize.For determination of the lattice parameters we compare the data with that found for ortho- present in (c) resulting from ionic ammonium carboxylate structures. In the two spectra both the CLC stretch vibrations rhombic polyethylene13 with lattice constants of a=0.7417 nm, b=0.4945 nm and c=0.2547 nm.Assuming the same lattice at 1610 cm-1 and the CH2 bending vibrations at 1466 and 1456 cm-1 are identical. From this we conclude that within for (a), the indexing of the upper reflections is (110) and (200) giving lattice constants a=0.76 nm and b=0.50 nm. In the the experimental error all of the carboxylic acid functions of (b) form acid–base pairs with the amino functions of (a), unit cell the alkyl chains are oriented perpendicular to the a–b J.Mater. Chem., 1999, 9, 1055–1057 1055Fig. 2 FTIR-spectrum of (b) (curve i) and that of complex (c) (curve ii). In the spectrum of (b) an intense carboxylic acid band at a wave number of 1710 cm-1 is found resulting from the CLO stretch Fig. 3 The X-ray diVraction of (b) (curve i) and that of complex (c) vibration. This band is absent in the spectrum of (c) indicating a (curve ii) in the hexagonal mesophase at 110 °C. The (001)-reflection stoichiometric 151 complexation. of (c) is more intense than that of (b) indicating a higher stacking order of aromatic cores for (c) than that of (b). The insert shows the small-angle X-ray scattering of compound (b) at 20°C (curve iii), plane and the average area per chain is 0.19 nm2, which is which is due to a rectangular centered columnar phase (see Table 1).slightly larger than that for polyethylene (0.183 nm2). Neither (b) nor the complex (c) show sharp reflections in the wide- Table 1 Small-angle X-ray data of (b). Reflex positions and Miller angle region of the scattering curves, proving the absence of indices are given for the two-dimensional ordered lattices at 20 and crystallinity in the structures. The disappearance of crystallinity 110 °C.The scattering vector is defined as s=2/lsin h in (c) is also indicative of a stoichiometric complex. The DSC curve of (a) shows a melting peak at 59 °C with an enthalpy (hkl ) sobserved/nm-1 scalculated/nm-1 of 73 J g-1.Under the usual assumptions made for the calcu- Rectangular discotic structure, T=20 °C lation of crystallinity,14 the amount of crystalline side-chains (110) 0.401 0.400 of (a) can be determined to be in the range of 40 to 50%. In (200) 0.495 0.495 the DSC trace of (b) an endothermic transition was found at (020) 0.629 0.629 93 °C for heating and at 62 °C for cooling.For (c) an endo- (220) 0.787 0.800 thermic transition is found at 70 °C for heating and at 52 °C (310) — 0.806 upon cooling. The diVerent values found for the hysteresis, (130) 0.910 0.975 18 °C (c) and 31 °C (b), may be interpreted as a consequence (330) 1.194 1.201 of a faster relaxation to an equilibrium state of (c). Considering (040) 1.262 1.257 the absence of crystalline reflections in the X-ray diagrams of (510) — 1.277 (b) and (c), the exothermic transitions could be assigned to Hexagonal discotic structure, T=110 °C liquid-crystalline phase transitions.This assumption was (100) 0.389 0.389 proven by optical polarization microscopy and by small-angle (110) 0.679 0.674 X-ray scattering. Both compounds show strong birefringence (200) 0.790 0.778 when observed between crossed polarizers and a texture change at transition temperatures found by DSC.Unfortunately, the textures do not allow an unambiguous determination of the than for (b), but the (001) reflection located in the wide-angle region is significantly smaller for the complex than for the free liquid-crystalline phases.Therefore,-angle X-ray scattering diagrams were recorded at 20 and 110 °C. As shown in acid (see Fig. 3). The inter-columnar distances, derived from the (001) reflection for both are the same (dintra=0.354 nm). Fig. 3, at 20 °C seven reflections were found in the small-angle scattering curve of (b). This pattern can be interpreted as a This value is nearly the same as found earlier for nonfunctionalized hexa-peri-benzocoronene1 and indicates an columnar structure with a two-dimensional centered superstructure (see Fig. 1). The columns are aligned parallel, where eVective p–p-overlap of adjacent cores.5 For quantification of the diVerence between the inter-columnar ordering and intra- the aromatic cores of (b) form stacks and are tilted with respect to the column axis.The indices of the reflections and columnar ordering of (b) and (c), the correlation lengths were calculated from the widths of the (100) and the (001) reflec- the observed and calculated reflex positions are listed in Table 1. On the basis of these values the lattice parameters are tions. The correlation lengths of (b) are 45 nm and 3.7 nm for the inter- and intra-columnar long-range ordering, respectively.determined to be 4.04 nm and 3.18 nm, giving an inter-columnar distance of 2.71 nm. A tilt angle between the plane normal For (c) the corresponding values are 33 nm and 5.7 nm. Thus, the inter-columnar order in the pure compound is higher than to the aromatic cores aligned and the column direction is calculated to be about 38°. The SAXS diagram of complex (c) in the complex, whereas the intra-columnar long-range order of the pure substance is lower than in the complex.For the shows a reflection pattern very similar to that of (b), but with broader reflections. We conclude that both materials have the desired rapid charge transport along the columns, the p–poverlap and the intra-columnar long-range ordering are critical same liquid-crystalline rectangular columnar structure at room temperature, although diVering in the lateral packing of the values.Since the p–p-overlap of (b) and (c) is equal, but the intra-columnar long-range order of (c) is better, a higher columns, which appears to be better in (b) compared to (c). At a temperature of 110 °C, in the small angle X-ray region charge transport may be expected for (c) than for (b).In addition to the control of the mesomorphic structure, of (b) and (c) three reflections were found with relative positions of 15Ó352. This can be interpreted as an ordered the mechanical properties play a key role in estimating the potential of (b) and (c) as promising materials. For many columnar structure with a hexagonal superstructure as was already found for alkylated non-functionalized hexa-peri- polyelectrolyte–surfactant complexes it is known that the mechanical properties were significantly improved due to the benzocoronenes.1 From the reflex positions, inter-columnar distances of 3.00 nm (b) and 3.12 nm (c) were determined.polymeric compound.15 Hence we expect such an improvement for the complexation of (b), too.The shear modulus G of Again, the small-angle reflections observed for (c) are broader 1056 J. Mater. Chem., 1999, 9, 1055–1057(HPLC grade), was added slowly to an equimolar amount of (b) dissolved in 25 mL chloroform (HPLC grade). The complex precipitated as a fine yellow powder and was removed by centrifugation, washed three times with 10 mL chloroform and dried under vacuum.The yield was about 95%. The complex is soluble in warm dimethylformamide and tetrahydrofuran. Wide-angle X-ray scattering measurements were carried out with a Nonius PDS 120 powder diVractometer in transmission geometry. A FR590 generator was used as the source of Cu- Ka radiation; monochromatization of the primary beam was achieved by means of a curved Ge crystal.The scattered radiation was measured with a CPS120 position sensitive detector. The resolution of this detector is better than 0.018°. Small-angle X-ray scattering measurements were recorded with an X-ray vacuum camera with pinhole collimation (Anton Paar, Austria; model A-8054) equipped with image plates Fig. 4 Temperature dependent loss-angle determined by dynamic– (type BAS III, Fuji, Japan).The image plates were read out mechanical measurements of (b) (squares) and of (c) (circles). The with a MACScience Dip-Scanner IPR-420 and IP reader shear rate was 0.1 s-1. DIPR-420 (Japan). DiVerential scanning calorimetry (DSC) measurements were performed on a Netsch DSC 200 (Germany). The samples were examined at a scanning rate of films, prepared by solvent casting of (b) and (c), was deter- 10 K min-1 by applying one cooling and two heating scans.mined by temperature-dependent dynamic–mechanical The phase transition temperatures were determined as onset measurements. It was found for the pure substance and the points. Polarized light optical microscopic observations of the complex that the shear modulus of the rectangular columnar films were performed with a Zeiss DMRB microscope phase is one order of magnitude higher than that of the (Germany).A Bohlin CVO-50 rheometer with a plate-plate hexagonal columnar phase. Below the phase transition, values geometry (20 mm diameter) was used for dynamical–mechanof G in the range from 1 MPa to 10 MPa were measured for ical measurements. The substances (b) and (c) were cast from (b) and (c).Above the phase transition, G is about 0.1 MPa. a 0.2% (w/w) dimethylformamide solution to give a homo- Characteristic is the diVerent behavior of loss-angle d as shown geneous film on the lower plate. Measurements were carried in Fig. 4. It can be seen that d increases from about 5 to 40° out at a temperature range from 20 to 130 °C with a cooling in the range from 20 to 130 °C for (b), whereas d increases and heating rate of 2 K min-1 at a shear rate of 0.1 s-1, only from 10 to 20° in the same temperature range for (c).loading with a constant shear stress of 50 Pa. This means that (b) is more elastic at room temperature and viscous at higher temperature. On the other hand (c) shows We wish to thank M. Antonietti for helpful suggestions and an approximately constant elasticity over the whole tempera- C.Burger for critical discussion of the X-ray data ture range. This may be explained by considering (a) as interpretation. a kind of cross-linker, which forms a flexible threedimensional network bound ionically to (b). References In conclusion it was found that the carboxylic acid functionalized hexabenzocoronene (b) forms two columnar 1 P.Herwig, C. W. Kayser, K. Mu� llen and H. W. Spiess, Adv. liquid-crystalline structures. At room temperature a rectangu- Mater., 1996, 8, 510. lar centered symmetry is found, which changes to a hexagonal 2 V. S. Iyer, M. Wehmeier, J. D. Brand, M. A. Keegstra and ordered columnar structure at temperatures higher than 93 °C. K. Mu� llen, Angew. Chem., 1997, 109, 1676.Furthermore, the complexation of (b) in chloroform with a 3 J. S. Moore, Curr. Opin. Solid State Mater. Sci., 1996, 1, 777. hydrophobic modified polyethyleneimine gives a stoichiometric 4 A. Stabel, P. Herwig, K. Mu� llen and J. P. Rabe, Adv. Mater., 1995, 10, 36. complex similar to solid waterborne polyelectrolyte–surfactant 5 A. M. van de Craats, J. M. Warman, K.Mu� llen, Y. Geerts and complexes. For the stoichiometric complex of (a) and (b) J. D. Brand, Adv. Mater., 1998, 10, 36. resulting in (c), the same columnar liquid-crystalline phases 6 J. D. Brand, S. Ito and K. Mu� llen, J. Mater. Sci., to be published. are present, but with a lower phase-transition temperature 7 C. K. Ober and G. Wegner, Adv. Mater., 1997, 9, 17. (70 °C vs. 93 °C) and a lower hysteresis (18 °C vs. 31 °C), 8 M. Antonietti, C. Burger and A. Thu�nemann, Trends Polym. Sci. observed for the phase-transition temperature between heating (Cambridge, UK), 1997, 5, 262. and cooling. Furthermore, the intra-columnar order of the 9 A.F.Thu� nemann, Adv. Mater., 1998, accepted. complex is higher than that of the non-complexed specimen, 10 M. Antonietti, S.Henke and A. F. Thu�nemann, Adv. Maternter-columnar order is lower. From the tempera- 1996, 8, 41. 11 A. F. Thu�nemann and K. H. Lochhaas, Langmuir, 1998, 14, 4898. ture dependence of the loss-angle it was concluded, that the 12 A. F. Thu�nemann, A. Lieske and B.-R. Paulke, Adv. Mater., elasticity of (b) could be improved by complexation, probably 1999, April. without the danger of reducing high carrier mobilities. 13 J. Brandrup and E. H. Immergut, Polymer Handbook,Wiley, New The hydrophobic modification of polyethyleneimine York, 1989, 3rd edn., vol. 16. (Polymin G100A, BASF, Mw=5.000–7000 g mol-1) with 14 E. A. Ponomarenko, A. J. Waddon, D. A. Tirell and docosyl bromide (Aldrich, 97%) was carried out by a procedure W. J. MacKnight, Langmuir, 1996, 12, 2169.described by No�ding and Heitz.16 The degree of alkylation, as 15 M. Antonietti, M. Neese, G. Blum and F. Kremer, Langmuir, determined by 1H-NMR spectroscopy, was 16 mol%. Synthesis 1996, 12, 4436. and characterization of the functionalized hexabenzocoronene 16 G. No�ding and W. Heitz, Macromol. Chem. Phys., 1998, 199, 1637. (b) (see Fig. 1) are described in detail elsewhere.6 For complexation 6.5×10-2 mmol (100 mg) of the hydrophobically modified polyethyleneimine, dissolved in 10 mL chloroform Communication 9/00989B J.Mater. Chem., 1999, 9, 1055–1057 1057 J O U R N A L O F C H E M I S T R Y Materials Communication Supramolecular architecture of a functionalized hexabenzocoronene and its complex with polyethyleneimine Andreas F. Thu�nemann,*a Dirk Ruppelt,a Shunji Itob and Klaus Mu�llenb aMax Planck Institute of Colloids and Interfaces, Kantstraße 55, 14513 Teltow, Germany.Email: andreas@terra.mpikg-teltow.mpg.de; Tel. 03328–46–271, Fax. 03328–46–272 bMax-Planck-Institut fu�r Polymerforschung, Ackermannweg 10, D-55128 Mainz, Germany Received 5th February 1999, Accepted 8th March 1999 A hexabenzocoronene, functionalized with a carboxylic acid group, was complexed with a hydrophobic modified polyethyleneimine and the thermotropic columnar liquidcrystalline- like complex was investigated by X-ray scattering and dynamic–mechanical analysis.The construction of disc-like molecules with large p-conjugated cores opens a way to new supramolecular aggregates,1,2 which are interesting because of their electronic and optoelectronic properties.3 It has been observed, for instance that pericondensed hexabenzocoronenes show diode-like current–voltage signals for single molecules4 and a rapid charge transport along self-assembled columns with high carrier mobilities.For an oriented layer with a thickness of 1 mm and an applied potential of 1 V the drift time of charges across such a layer was calculated to be less than one microsecond.5 The onedimensional nature of the charge transport makes these molecules promising as nanowires in molecular electronic devices or as building blocks for transport layers in electrocopying or electrophotography.A decisive aspect is to take advantage of their self-organizing capability particularly with respect to the control of the spatial orientation of molecular wires, as formed e.g. by hexakis(tetradecyl )hexa-peri-hexabenzocoronenes.5 The first step in this direction is the functionalization of such sheet-like molecules by an anchor group, which may, in principle, allow the immobilization of supramolecular aggregates. The binding of molecular columns to polyelectrolytes is a possibility to enhance their long-range order and stability.In this article we describe the phase behavior and mechanical properties of a hexabenzocoronene (b) which is functionalized by a carboxylic acid group and that of its complex with hydrophobic modified polyethyleneimine (a). The structures of the compounds are shown in Fig. 1. The synthesis of (b) will be described elsewhere.6 Fig. 1 The molecular structure of an ionically bound complex (c), In the last few years, it has been shown that the complexation formed by a hydrophobic modified polyethyleneimine (a) and a of polyelectrolytes with surfactants in water results in a number hexabenzocoronene (b).(a) and (b) are not drawn. 16% of the amino of new solid materials,7,8 e.g. with interesting optical9 and functions of (a) are alkylated by n-docosyl chains, (b) is functionalized surface-energy lowering properties.10–12 Analogously to poly- by a carboxylic acid function which serves as anchor group.The electrolyte–surfactant complexes we treated (b) as an amphiph- sketch at the bottom illustrates the columnar mesophases formed by ile and used (a) as a polyelectrolyte. Due to the insolubility of (b) and (c). For clarity only the aromatic cores of (a) are drawn.At (b) in water, the complexation with (a) was carried out in room temperature a rectangular centered lattice is found for both. The aromatic cores are tilted with respect to the column axis. At chloroform using an equimolar amount of (a) with respect to higher temperatures the columns are oriented in a hexagonal columnar amino and carboxylate functions.The resulting complex (c) mesophase. was solvent-cast as films and analyzed by IR spectroscopy, diVerential scanning calorimetry (DSC), X-ray scattering and dynamic–mechanical measurements. Comparison of the IR resulting in a stoichiometric complex (c). In the wide-angle Xray diagram of (a) at 20°C, sharp reflections were found at a spectra of (b) and (c) shows that the carboxylic acid vibration found for (a) at 1710 cm-1 is absent for (c) (see Fig. 2). scattering vector of s=2.39 nm-1 and 2.63 nm-1. This is attributed to a side-chain crystallinity of (a), in which the Instead of the sharp carboxylic acid band in the spectrum of (b) a broad diVuse band between 1500 and 1760 cm-1 is docosyl chains crystallize. For determination of the lattice parameters we compare the data with that found for ortho- present in (c) resulting from ionic ammonium carboxylate structures.In the two spectra both the CLC stretch vibrations rhombic polyethylene13 with lattice constants of a=0.7417 nm, b=0.4945 nm and c=0.2547 nm. Assuming the same lattice at 1610 cm-1 and the CH2 bending vibrations at 1466 and 1456 cm-1 are identical.From this we conclude that within for (a), the indexing of the upper reflections is (110) and (200) giving lattice constants a=0.76 nm and b=0.50 nm. In the the experimental error all of the carboxylic acid functions of (b) form acid–base pairs with the amino functions of (a), unit cell the alkyl chains are oriented perpendicular to the a–b J. Mater. Chem., 1999, 9, 1055–1057 1055Fig. 2 FTIR-spectrum of (b) (curve i) and that of complex (c) (curve ii). In the spectrum of (b) an intense carboxylic acid band at a wave number of 1710 cm-1 is found resulting from the CLO stretch Fig. 3 The X-ray diVraction of (b) (curve i) and that of complex (c) vibration. This band is absent in the spectrum of (c) indicating a (curve ii) in the hexagonal mesophase at 110 °C.The (001)-reflection stoichiometric 151 complexation. of (c) is more intense than that of (b) indicating a higher stacking order of aromatic cores for (c) than that of (b). The insert shows the small-angle X-ray scattering of compound (b) at 20°C (curve iii), plane and the average area per chain is 0.19 nm2, which is which is due to a rectangular centered columnar phase (see Table 1).slightly larger than that for polyethylene (0.183 nm2). Neither (b) nor the complex (c) show sharp reflections in the wide- Table 1 Small-angle X-ray data of (b). Reflex positions and Miller angle region of the scattering curves, proving the absence of indices are given for the two-dimensional ordered lattices at 20 and crystallinity in the structures. The disappearance of crystallinity 110 °C.The scattering vector is defined as s=2/lsin h in (c) is also indicative of a stoichiometric complex. The DSC curve of (a) shows a melting peak at 59 °C with an enthalpy (hkl ) sobserved/nm-1 scalculated/nm-1 of 73 J g-1. Under the usual assumptions made for the calcu- Rectangular discotic structure, T=20 °C lation of crystallinity,14 the amount of crystalline side-chains (110) 0.401 0.400 of (a) can be determined ange of 40 to 50%.In (200) 0.495 0.495 the DSC trace of (b) an endothermic transition was found at (020) 0.629 0.629 93 °C for heating and at 62 °C for cooling. For (c) an endo- (220) 0.787 0.800 thermic transition is found at 70 °C for heating and at 52 °C (310) — 0.806 upon cooling. The diVerent values found for the hysteresis, (130) 0.910 0.975 18 °C (c) and 31 °C (b), may be interpreted as a consequence (330) 1.194 1.201 of a faster relaxation to an equilibrium state of (c).Considering (040) 1.262 1.257 the absence of crystalline reflections in the X-ray diagrams of (510) — 1.277 (b) and (c), the exothermic transitions could be assigned to Hexagonal discotic structure, T=110 °C liquid-crystalline phase transitions.This assumption was (100) 0.389 0.389 proven by optical polarization microscopy and by small-angle (110) 0.679 0.674 X-ray scattering. Both compounds show strong birefringence (200) 0.790 0.778 when observed between crossed polarizers and a texture change at transition temperatures found by DSC. Unfortunately, the textures do not allow an unambiguous determination of the than for (b), but the (001) reflection located in the wide-angle region is significantly smaller for the complex than for the free liquid-crystalline phases.Therefore, small-angle X-ray scattering diagrams were recorded at 20 and 110 °C. As shown in acid (see Fig. 3). The inter-columnar distances, derived from the (001) reflection for both are the same (dintra=0.354 nm).Fig. 3, at 20 °C seven reflections were found in the small-angle scattering curve of (b). This pattern can be interpreted as a This value is nearly the same as found earlier for nonfunctionalized hexa-peri-benzocoronene1 and indicates an columnar structure with a two-dimensional centered superstructure (see Fig. 1). The columns are aligned parallel, where eVective p–p-overlap of adjacent cores.5 For quantification of the diVerence between the inter-columnar ordering and intra- the aromatic cores of (b) form stacks and are tilted with respect to the column axis.The indices of the reflections and columnar ordering of (b) and (c), the correlation lengths were calculated from the widths of the (100) and the (001) reflec- the observed and calculated reflex positions are listed in Table 1.On the basis of these values the lattice parameters are tions. The correlation lengths of (b) are 45 nm and 3.7 nm for the inter- and intra-columnar long-range ordering, respectively. determined to be 4.04 nm and 3.18 nm, giving an inter-columnar distance of 2.71 nm. A tilt angle between the plane normal For (c) the corresponding values are 33 nm and 5.7 nm.Thus, the inter-columnar order in the pure compound is higher than to the aromatic cores aligned and the column direction is calculated to be about 38°. The SAXS diagram of complex (c) in the complex, whereas the intra-columnar long-range order of the pure substance is lower than in the complex. For the shows a reflection pattern very similar to that of (b), but with broader reflections. We conclude that both materials have the desired rapid charge transport along the columns, the p–poverlap and the intra-columnar long-range ordering are critical same liquid-crystalline rectangular columnar structure at room temperature, although diVering in the lateral packing of the values.Since the p–p-overlap of (b) and (c) is equal, but the intra-columnar long-range order of (c) is better, a higher columns, which appears to be better in (b) compared to (c).At a temperature of 110 °C, in the small angle X-ray region charge transport may be expected for (c) than for (b). In addition to the control of the mesomorphic structure, of (b) and (c) three reflections were found with relative positions of 15Ó352.This can be interpreted as an ordered the mechanical properties play a key role in estimating the potential of (b) and (c) as promising materials. For many columnar structure with a hexagonal superstructure as was already found for alkylated non-functionalized hexa-peri- polyelectrolyte–surfactant complexes it is known that the mechanical properties were significantly improved due to the benzocoronenes.1 From the reflex positions, inter-columnar distances of 3.00 nm (b) and 3.12 nm (c) were determined.polymeric compound.15 Hence we expect such an improvement for the complexation of (b), too. The shear modulus G of Again, the small-angle reflections observed for (c) are broader 1056 J. Mater. Chem., 1999, 9, 1055–1057(HPLC grade), was added slowly to an equimolar amount of (b) dissolved in 25 mL chloroform (HPLC grade). The complex precipitated as a fine yellow powder and was removed by centrifugation, washed three times with 10 mL chloroform and dried under vacuum.The yield was about 95%. The complex is soluble in warm dimethylformamide and tetrahydrofuran. Wide-angle X-ray scattering measurements were carried out with a Nonius PDS 120 powder diVractometer in transmission geometry.A FR590 generator was used as the source of Cu- Ka radiation; monochromatization of the primary beam was achieved by means of a curved Ge crystal. The scattered radiation was measured with a CPS120 position sensitive detector. The resolution of this detector is better than 0.018°. Small-angle X-ray scattering measurements were recorded with an X-ray vacuum camera with pinhole collimation (Anton Paar, Austria; model A-8054) equipped with image plates Fig. 4 Temperature dependent loss-angle determined by dynamic– (type BAS III, Fuji, Japan). The image plates were read out mechanical measurements of (b) (squares) and of (c) (circles). The with a MACScience Dip-Scanner IPR-420 and IP reader shear rate was 0.1 s-1. DIPR-420 (Japan).DiVerential scanning calorimetry (DSC) measurements were performed on a Netsch DSC 200 (Germany). The samples were examined at a scanning rate of films, prepared by solvent casting of (b) and (c), was deter- 10 K min-1 by applying one cooling and two heating scans. mined by temperature-dependent dynamic–mechanical The phase transition temperatures were determined as onset measurements.It was found for the pure substance and the points. Polarized light optical microscopic observations of the complex that the shear modulus of the rectangular columnar films were performed with a Zeiss DMRB microscope phase is one order of magnitude higher than that of the (Germany). A Bohlin CVO-50 rheometer with a plate-plate hexagonal columnar phase.Below the phase transition, values geometry (20 mm diameter) was used for dynamical–mechanof G in the range from 1 MPa to 10 MPa were measured for ical measurements. The substances (b) and (c) were cast from (b) and (c). Above the phase transition, G is about 0.1 MPa. a 0.2% (w/w) dimethylformamide solution to give a homo- Characteristic is the diVerent behavior of loss-angle d as shown geneous film on the lower plate.Measurements were carried in Fig. 4. It can be seen that d increases from about 5 to 40° out at a temperature range from 20 to 130 °C with a cooling in the range from 20 to 130 °C for (b), whereas d increases and heating rate of 2 K min-1 at a shear rate of 0.1 s-1, only from 10 to 20° in the same temperature range for (c). loading with a constant shear stress of 50 Pa. This means that (b) is more elastic at room temperature and viscous at higher temperature.On the other hand (c) shows We wish to thank M. Antonietti for helpful suggestions and an approximately constant elasticity over the whole tempera- C. Burger for critical discussion of the X-ray data ture range. This may be explained by considering (a) as interpretation.a kind of cross-linker, which forms a flexible threedimensional network bound ionically to (b). References In conclusion it was found that the carboxylic acid functionalized hexabenzocoronene (b) forms two columnar 1 P. Herwig, C. W. Kayser, K. Mu� llen and H. W. Spiess, Adv. liquid-crystalline structures. At room temperature a rectangu- Mater., 1996, 8, 510. lar centered symmetry is found, which changes to a hexagonal 2 V.S. Iyer, M. Wehmeier, J. D. Brand, M. A. Keegstra and orderedolumnar structure at temperatures higher than 93 °C. K. Mu� llen, Angew. Chem., 1997, 109, 1676. Furthermore, the complexation of (b) in chloroform with a 3 J. S. Moore, Curr. Opin. Solid State Mater. Sci., 1996, 1, 777. hydrophobic modified polyethyleneimine gives a stoichiometric 4 A.Stabel, P. Herwig, K. Mu� llen and J. P. Rabe, Adv. Mater., 1995, 10, 36. complex similar to solid waterborne polyelectrolyte–surfactant 5 A. M. van de Craats, J. M. Warman, K. Mu� llen, Y. Geerts and complexes. For the stoichiometric complex of (a) and (b) J. D. Brand, Adv. Mater., 1998, 10, 36. resulting in (c), the same columnar liquid-crystalline phases 6 J.D. Brand, S. Ito and K. Mu� llen, J. Mater. Sci., to be published. are present, but with a lower phase-transition temperature 7 C. K. Ober and G. Wegner, Adv. Mater., 1997, 9, 17. (70 °C vs. 93 °C) and a lower hysteresis (18 °C vs. 31 °C), 8 M. Antonietti, C. Burger and A. Thu�nemann, Trends Polym. Sci. observed for the phase-transition temperature between heating (Cambridge, UK), 1997, 5, 262. and cooling. Furthermore, the intra-columnar order of the 9 A.F.Thu� nemann, Adv. Mater., 1998, accepted. complex is higher than that of the non-complexed specimen, 10 M. Antonietti, S. Henke and A. F. Thu�nemann, Adv. Mater., whereas the inter-columnar order is lower. From the tempera- 1996, 8, 41. 11 A. F. Thu�nemann and K. H. Lochhaas, Langmuir, 1998, 14, 4898. ture dependence of the loss-angle it was concluded, that the 12 A. F. Thu�nemann, A. Lieske and B.-R. Paulke, Adv. Mater., elasticity of (b) could be improved by complexation, probably 1999, April. without the danger of reducing high carrier mobilities. 13 J. Brandrup and E. H. Immergut, Polymer Handbook,Wiley, New The hydrophobic modification of polyethyleneimine York, 1989, 3rd edn., vol. 16. (Polymin G100A, BASF, Mw=5.000–7000 g mol-1) with 14 E. A. Ponomarenko, A. J. Waddon, D. A. Tirell and docosyl bromide (Aldrich, 97%) was carried out by a procedure W. J. MacKnight, Langmuir, 1996, 12, 2169. described by No�ding and Heitz.16 The degree of alkylation, as 15 M. Antonietti, M. Neese, G. Blum and F. Kremer, Langmuir, determined by 1H-NMR spectroscopy, was 16 mol%. Synthesis 1996, 12, 4436. and characterization of the functionalized hexabenzocoronene 16 G. No�ding and W. Heitz, Macromol. Chem. Phys., 1998, 199, 1637. (b) (see Fig. 1) are described in detail elsewhere.6 For complexation 6.5×10-2 mmol (100 mg) of the hydrophobically modified polyethyleneimine, dissolved in 10 mL chloroform Communication 9/00989B J. Mater. Chem., 1999, 9

ISSN:0959-9428    

DOI:10.1039/a900989b

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Chain folding in poly(e-caprolactone) studied by small-angle X-ray scattering and Raman spectroscopy. A strategy for blending in the crystalline state Susan A. Berrill,a Frank Heatley,a John H. Collett,a David Attwood,a Colin Booth,*a J. Patrick A. Fairclough,b Anthony J. Ryan,b† Kyriakos Viras,c Amanda J. Duttond and Ross S. Blundelle aDepartment of Chemistry and School of Pharmacy, University of Manchester, Manchester, UK M13 9PL bDepartment of Chemistry, University of SheYeld, SheYeld, UK S3 7UF cNational and Kapodistrian University of Athens, Chemistry Department, Physical Chemistry Laboratory, Panepistimiopolis, Athens 157 71, Greece dSolvay-Interox Ltd, PO Box 7, Baronet Road, Warrington, UKWA4 6HB eSanofi-Winthrop Research, Alnwick, Northumberland, UK NE66 2JH Received 9th November 1998, Accepted 18th February 1999 Samples of poly(e-caprolactone), including narrow fractions obtained by preparative gel permeation chromatography (GPC), were characterised by analytical GPC and 13C NMR spectroscopy.The crystallised samples were investigated by small-angle X-ray scattering to obtain lamellar spacings and by low-frequency Raman spectroscopy to obtain LAM-1 frequencies.Together the two methods gave a reliable estimate of the critical stem length for chain folding in poly(e-caprolactone) crystallised at room temperature or below, i.e. 15 chain units [CL, OCO(CH2)5] equivalent to 105 chain atoms. A necessary requirement for molecular blending of block copolymers of e-caprolactone and ethylene oxide with crystalline high molar mass poly(e-caprolactone) is cocrystallisation. Hence it was concluded that a safe lower limit for successful blending is a poly(e-caprolactone) block of 20 CL units.[poly(E), E=OCH2CH2] block must be sited in the interlamel- Introduction lar region of the stacked crystal lamellae, thus providing Poly(e-caprolactone) [poly(CL), CL=OCO(CH2)5] has pathways for permeation of water into the bulk of the material.potential as a biodegradable polyester for use in solubilisation The rate of permeation could then be controlled by varying and/or encapsulation of drugs.1 Its biodegradation is based the proportion of poly(E) in the copolymer. on hydrolysis of the ester linkages, which in turn requires In considering such a strategy, it was apparent that best permeation of water into the polymer.Since the polymer is results would be expected if the copolymers to be blended highly hydrophobic, a number of strategies have been reported were small. This in itself, by promoting the entropy of mixing, for enhancing water permeation. These include statistical would enhance miscibility. Random incorporation of many copolymerisation with less hydrophobic comonomers (e.g.DL- short poly(CL) blocks into a lamellar crystal would also mean or L-lactide),2 block copolymerisation with a hydrophilic an even distribution of short poly(E) blocks at the lamellar comonomer (e.g. ethylene oxide),3 and blending with a hydro- surfaces. Equally importantly, the preparation of block copolyphilic polymer or copolymer (e.g.with Synperonic-PE 61, an mers by sequential anionic polymerisation is relatively straightoxyethylene/ oxypropylene block copolymer).4 forward for short blocks. This being so, it was desirable to Our immediate interest is in using poly(CL) for encapsul- define the minimum length of a poly(CL) block that would ation, for which purpose blending high molar mass poly(CL) ensure cocrystallisation, since very short poly(CL) blocks with a second polymer or copolymer is attractive. However, would be rejected from the crystal.blending is complicated by two factors. One is the well-known Lamellar spacings in poly(CL) were measured many years thermodynamic immiscibility of polymers in the absence of ago by Perret and Skoulios using small-angle X-ray scattering specific attractive interactions, which means that most blends (SAXS).9 The lamellar spacing for a crystallised sample of a are at best mechanically mixed, and so are heterogeneous on high molar mass poly(CL) was found to be 165 A° , this being a local scale.5 The second is the crystallinity of poly(CL),6–8 the stem length between chain folds.Their investigation of the with the consequent exclusion of other polymers from the spacings for short chains showed that this value was closely crystalline regions.Of course crystallinity in a polymer can be approached when the poly(CL) chain length reached 30 CL an advantage, since it is a source of mechanical strength in units (here denoted CL30). However, in their experiments the a material.approach to the limiting value was gradual, probably as a Recently we have sought to define the requirements for a consequence of the chain length distribution in the fractions. stable system based on crystalline high molar mass poly(CL) It seemed possible that the limit might be reached at a shorter blended with diblock poly[e-caprolactone-block-poly(ethylene poly(CL) chain length.As described in this report, related oxide)]. Compatibility with the crystalline poly(CL) was to be experiments using fractions with very narrow chain length ensured through cocrystallisation of the poly(CL) block of the distributions showed this to be so. Investigation of the longicopolymer. Given cocrystallisation, the poly(oxyethylene) tudinal acoustic mode (LAM) vibration by low-frequency Raman spectroscopy provided useful confirmatory evidence for this conclusion, since the single-node longitudinal vibration †Also CCLRC Daresbury Laboratory, Warrington, UK WA4 4AD.J. Mater. Chem., 1999, 9, 1059–1063 1059(LAM-1) is directly related to the crystal-stem length, and so guards against a false conclusion from SAXS data caused by crystal-stem tilting in the lamella.10 Experimental Polymers: preparation and characterisation Three poly(CL) samples with number-average molar mass in the range Mn=2000–4000 g mol-1 (chain length n=17–35 CL units) were selected from the range produced by Solvay Interox, Widnes, UK.These samples had fairly narrow chain length distributions (Mw/Mn=1.2–1.4). A fourth high molar mass polymer (Mn#80000 g mol-1, Mw/Mn=1.5) was also examined in order to fix an upper limit to the lamellar spacing. The anionic polymerisation of these samples was initiated by butanediol. In what follows we denote polymers as P-CLn, where n is the number-average chain length in CL units.Fig. 1 GPC curves obtained for poly(e-caprolactone) fraction F-CL14 Polymer fractions (see below) are denoted F-CLn.using GPC systems A and B, as indicated. A polymer of low molar mass (Mn#450 g mol-1, P-CL4) was prepared in our laboratory, starting from pentanediol. e- Caprolactone (Solvay Interox) was stirred overnight with 2,4- Table 1 Molecular characteristics of the poly(CL) samples diisocyanato-1-methylbenzene (2 wt%, Fluka AG), distilled Mw/Mn a Mn/g mol-1 under reduced pressure (140 °C, 15 mm Hg) directly onto Sample (GPC) (NMR/GPC) calcium hydride (Fischer Scientific), and then redistilled onto activated type 4 A° molecular sieves for storage.Immediately Polymers before use, pentane-1,5-diol (Lancaster Chemicals) was dried P-CL19 1.26 2200 by warming in vacuo for 24 h before being distilled at reduced P-CL25 1.41 2900 P-CL38 1.41 4300 pressure.The catalyst, stannous 2-ethylhexanoate (Sigma), P-CL700 1.46 80000b was used as provided. For polymerisation, pentanediol (3.5 g, Fractions 0.03 mol) and e-caprolactone (11.6 g, 0.1 mol) were syringed F-CL5 1.02 590 into a dry ampoule containing a magnetic stirrer. After F-CL9 1.02 1060 addition of a drop of the catalyst, the contents of the ampoule F-CL11 1.02 1300 were degassed, and the reaction mixture then stirred in a sand F-CL14 1.02 1630 F-CL16 1.02 1800 bath at 160±5 °C for 7 h.Characterisation by 13C NMR spectroscopy and analytical GPC gave Mn=450 g mol-1 and aMw/Mn to ±0.01, Mn to ±2%. bValue supplied by Solvay Interox. Mw/Mn=1.26. Polymer P-CL4 was separated by preparative gel permeation chromatography (preparative GPC) into fractions with very 13C NMR spectra were obtained using a Varian Associates narrow chain length distributions.A single 500 A° Styragel Unity 500 spectrometer operated at 125.8 MHz. Samples were column (120 cm×57 mm, Waters Assoc.) was used. Polymer dissolved in deuterochloroform, concentration 0.1 g cm-3. (0.5 g) in 20 cm3 toluene was injected into the solvent stream Analysis of the NMR spectra varied slightly depending on the (flow rate 10 cm3 min-1) at room temperature.Emergence of initiator used in preparation of the sample. Both commercial sample was detected by diVerential refractometry, upon which polymers and fractions conformed to the general formula 20×50 cm3 volumes were collected. The fractions of polymer CL*CLxICLxCL*, where were isolated by evacuation of solvent (24 h, 10-3 mm Hg, CL=CO·CH2CH2CH2CH2CH2O room temperature).In all, twenty fractions were collected, many very small, and just five were used in further work. a b c d e f The polymers and fractions were analysed by GPC (for CL*=CO·CH2CH2CH2CH2CH2OH distribution width) and 13C NMR spectroscopy (for numberaverage chain length). Two GPC systems were used. System a b c d g h A comprised three PLgel columns (Polymer Laboratories, I=OCH2CH2CH2CH2O or OCH2CH2CH2CH2CH2O 1×500 A° and 2×mixed B) used with tetrahydrofuran (THF) eluent (room temperature, 1 cm3 min-1, dodecane flow i j j i f e k e f marker).Samples were injected as 2 g dm-3 solutions via a 100 mm3 loop, and detected by diVerential refractometry. The chemical shifts (d) for the carbons as labelled are listed in Table 2: as indicated, those in the spectra of the low molar Universal calibration11 was with polystyrene standards and the Mark–Houwink constants for poly(styrene) (K=0.932×10-4, mass samples varied somewhat because of end eVects. With the exception of the resonance of the CO carbon (a), which a=0.740) and poly(e-caprolactone) (K=1.395×10-4, a= 0.786) in THF at room temperature advocated by Schindler had a low integral because of its longer relaxation time and lower NOE, the integrals of equivalent carbons were identical et al.12 System B comprised four PLgel columns (all mixed E) used with chloroform eluent (room temperature, 0.3 cm3 within experimental error.The sum of integrals of all resonances (except a), appropriately related to the sum of integrals min-1), other features being similar to those of System A.System B provided excellent resolution in the low-molar-mass of resonances from end group and initiator residue carbons, gave the values of Mn listed in Table 1 for the polymers and range. As examples, GPC curves obtained using systems A and B for analysis of fraction F-CL14 are shown in Fig. 1. The for fractions F-CL14 and F-CL5.There was no evidence of ring formation. curve from System A indicates a narrow chain length distribution (polydispersity index Mw/Mn#1.02) while that from Because of the small amounts of material available, only two fractions were characterised in this way. Molar masses of System B shows the sample to consist predominantly of three oligomers. Values of Mw/Mn are listed in Table 1.other fractions were obtained by use of analytical GPC (System 1060 J. Mater. Chem., 1999, 9, 1059–1063Table 2 13C NMR chemical shifts (CDCl3) SAXS detector, and high-density polyethylene and an NBS silicon standard were used to calibrate the WAXS detector. Carbon d/ppm The data acquisition system had a time-frame generator which collected the SAXS/WAXS data in 6 s frames separated by a a 173.5 wait-time of 10 ms.Parallel plate ionisation detectors placed b 34.1 c 25.5 before and after the sample cell recorded the incident and d 24.5 transmitted intensity. The experimental data were corrected e 28.3 for background scattering (subtraction of the scattering from f 64.1 the camera, hot stage and an empty cell ), sample thickness g 32.3 and transmission, and for any departure from positional h 62.5 linearity of the detectors.Since the samples crystallised with i 68.9 j 21.7 small domains randomly arranged the resulting pattern was k 22.2 analogous to that of a powder, and all orientations were adequately sampled in the one-dimensional experiment. Range: ±0.1 ppm for polymers (Mn>1000 g mol-1); ±0.5 ppm for fractions (Mn<1000 g mol-1).Further details of the equipment and methods can be found elsewhere.13–16 A) as follows. (i) Values of molar mass at the peak of the Low-frequency Raman spectroscopy GPC curve (Mpk) were calculated as Mpk=Mn(Mw/Mn)1/2, Raman scattering at 90° to the incident beam was recorded with Mn from NMR spectroscopy and Mw/Mn from GPC. by means of a Spex Ramalog spectrometer fitted with a 1403 This provided a common basis for plotting the data for the double monochromator plus a third (1442U) monochromator polymers and fractions.(ii) Note was taken of the eVect of operated in scanning mode. The light source was a Coherent hydrogen bonding of THF with the OH end groups of the Innova 90 argon-ion laser operated at 514.5 nm and 300 mW.samples. From comparison of the elution of low molar mass Typical operating conditions were bandwidth 0.8 cm-1, scan- polyethylene glycols and their dimethyl ethers, this eVect was ning increment 0.05 cm-1, integration time 6–12 s. The fre- found to decrease the elution volume consistent with an quency scale was calibrated by reference to the 9.4 and approximate correction to Mn of each sample of 140 g mol-1. 14.9 cm-1 bands in the spectrum of L-cystine. Samples in a The resulting calibration curve is shown in Fig. 2. With thin-wall capillary were melted and recrystallised at room acceptable scatter, the three commercial samples and the two temperature (approximately 20 °C). Spectra were recorded fractions fit to a common curve. Accordingly values of Mpk with the samples at various temperatures in the range 22 to for the fractions were obtained from their elution volumes in -100 °C, controlled to±1 °C by means of a Harney-Miller System A and converted to values of Mn by reversing the cell (Spex Industries).procedures (i) and (ii). These values, for fractions F-CL9, FCL11 and F-CL16, are included in the last column of Table 1. Results and discussion X-Ray scattering The WAXS patterns obtained were as expected for crystalline Measurements were made on beamline 8.2 of the SRS at the poly(CL), e.g.strong reflections at Bragg angle h=10.4° and CCLRC Daresbury Laboratory, Warrington, UK. The camera 11.6°, and weaker reflections also in keeping with expectation.6 was equipped with a multiwire quadrant detector (SAXS) located 3.5 m from the sample position and a curved knife- Lamellar spacings from SAXS edge detector (WAXS) that covered 70° of arc at a radius of Representative SAXS patterns are shown in Fig. 3 and 4. 0.3 m. Samples were sealed into TA Instruments DSC pans Fig. 3 shows a time-resolved relief diagram of SAXS data containing a 0.75 mm brass spacer ring and fitted with windows obtained during a melting–recrystallisation cycle (ramp rate= made from 25 mm thick mica.The loaded pans were placed in 10 °Cmin-1) for sample P-CL25. Intensity is plotted against the cell of a Linkam DSC of single-pan design, which could temperature and scattering vector q=(4p/l)sinh, where the be heated and cooled to allow melting and recrystallisation of wavelength l=1.54 A° . In this notation, the lamellar spacing the samples.The scattering pattern from an oriented specimen is given by Bragg’s Law in the form d=2p/q*, where q* is the of wet collagen (rat-tail tendon) was used to calibrate the value of q at the first-order peak. Fig. 4 shows the scattering Fig. 3 Time-resolved relief diagram of SAXS data obtained during a melting–recrystallisation cycle (ramp rate=10 °Cmin-1) for sample P-CL25.Intensity (arbitrary scale) is plotted against temperature and Fig. 2 GPC calibration curve (System A) for poly(e-caprolactone) scattering vector q=(4p/l)sinh, where l=1.54 A° and h is the scattering angle. ($) polymers and (&) fractions. J. Mater. Chem., 1999, 9, 1059–1063 1061the orthorhombic sub-cell dimension c=17.05 A° or c=17.26 A° (for 2 units per repeat) respectively reported by Chatani et al.6 and Hu and Dorset,7 and so consistent with the crystal stems being normal to the lamellar end planes.On the basis of the SAXS data, we conclude that the chain length between folds (crystal stem length) in low molar mass poly(CL) (Mn<4500 g mol-1) crystallised at low temperatures is approximately 15 CL units.Extrapolation of the line established for the lower molar mass fractions to d=150 A° , the value of d measured in this work for the high molar mass sample, would suggest 17 CL units, while extrapolation to d= 165 A° (Perret and Skoulios9) would take this value to 19 CL units. Values of d reported by Khambatta et al.17 for moderate molar mass poly(CL) (Mn=13000 g mol-1) are near to 150 A° .It seems that a safe lower limit to ensure that a short CL chain will cocrystallise with chain-folded lamellar crystals of high molar mass poly(CL) is 20 CL units, equivalent to 140 chain atoms. LAM-1 frequency from Raman spectroscopy Fig. 4 Scattering intensity (arbitrary scale) plotted against the ratio q/q* for fraction F-CL14 at 10 °C. The parameter q* is the value of q The three low molar mass polymers were examined by Raman at the first-order peak: in this case q*=0.055 A° -1.spectroscopy. Examples of low-frequency (<100 cm-1) spectra are shown in Fig. 6. There are three prominent bands in each spectrum, the frequencies of which are labelled (i) to (iii) in intensity plotted against the ratio q/q* for sample F-CL14 at Table 4. The spectra of the samples P-CL25 and P-CL38 are 10 °C, at which temperature q*=0.055 A° -1.essentially identical. The spectrum of P-CL19 diVers at the Values of the lamellar spacing are listed in Table 3 and lowest frequencies, mainly by the reduced intensity of the plotted against chain length (in CL units) in Fig. 5. Also lowest frequency band (iii), but with some evidence of splitting.plotted in Fig. 5 are the results of Perret and Skoulios.9 The Presumably this is a result of end eVects from a,v-hydroxy- two data sets are in substantial agreement. The sharp break ended chains in predominantly unfolded-chain lamellar in the curve in the present results at chain length 15 CL units crystals. and d=129 A° is attributable to the use of very narrow The LAM-1 band was assigned to band (ii) as follows.fractions, Mw/Mn#1.02, in this work. The d-spacing of 129 A° Adapting the model of Minoni and Zerbi,18 the lamellar stack corresponds to 8.6 A° per CL unit, which is in keeping with was represented as a one-dimensional crystal, with both chain stems and stem-end groups (chain ends or chain folds) Table 3 Lamellar spacings for crystalline poly(CL) samples at 20 °C Sample d/A° a P-CL19 133 P-CL25 128 P-CL38 128 P-CL700 149 F-CL9 82 F-CL11 103 F-CL14 117 F-CL16 128 ad to ±3 A° .Fig. 6 Low-frequency Raman spectra of poly(e-caprolactone) samples (as indicated) at 20 °C. The intensity scale and baseline levels are arbitrary. Table 4 Band frequencies (cm-1) from low-frequency Raman spectra of crystalline poly(CL) samples at 20 °C Sample (i) (ii) (iii) P-CL19 58.0 31.8 21.1 P-CL25 57.9 31.6 19.6 Fig. 5 Lamellar spacing from SAXS (d) versus chain length in CL P-CL38 57.9 31.8 19.4 units for poly(e-caprolactone) ($) polymers and (&) fractions. The Estimated uncertainties: (i) to ±0.2 cm-1, (ii) to ±0.1 cm-1, (iii) to open symbols (#) denote the results of Perret and Skoulios.9 The full ±0.5 cm-1. lines are drawn through the data points for the present samples. 1062 J. Mater. Chem., 1999, 9, 1059–1063Raman spectroscopy guards against the possibility that the lamellar spacing from SAXS is aVected by crystal-stem tilting. Concluding remarks We are aware that all the samples examined in the present work have an imperfection at the mid-point of the chain. The choice of pentanediol as initiator does not remove this feature, since (on average) the two arms of those chains are symmetrical about the mid-point.While this may aVect the crystallinity of the present samples, we find no evidence that it aVects the extent of chain folding as quantified by SAXS, nor the LAM- 1 frequency from Raman spectroscopy. We conclude that 20 CL units is a safe lower limit for the CL block length in block copolymers of e-caprolactone and ethylene oxide if they are to cocrystallise with chain-folded high molar mass poly(e-caprolactone) and so form blends suitable for encapsulation purposes.Fig. 7 Temperature dependence of the LAM-1 frequency of poly(ecaprolactone) P-CL19. Acknowledgements We thank Messrs S. K. Nixon and P. Kobryn for help with the experimental work.Sanofi-Winthrop, Alnwick, UK accounted for as point masses with appropriate interactions financed a research studentship for SAB. KV had the benefit between them. The SAXS results indicate a spacing equivalent of a Royal Society of Chemistry Journals Grant for to 15 CL units, equivalent to 105 atoms. In our approximate International Authors. calculation, the crystal stem was assigned 105 point masses each of 1.895×10-25 kg (average value for the CL unit), and References the force between chain ends was equated to that typical of a van der Waals interaction ( fe=5 Nm-1).19 Noting the almost 1 C.G. Pitt in Biodegradable Polymers as Drug Delivery Systems, trans-planar conformation of the poly(CL) chain and the ed. M. Chasin and R. Langer, Marcel Dekker, London, 1990. 2 E. Piskin, J. Biomater. Sci., Polym. Ed., 1994, 6, 775. similarity of its packing in the crystal,6 the force between chain 3 J. Bei, W. Wang, Z. Wang and S. Wang, Polym. Adv. Technol., masses was set equal to the value successfully used to model 1996, 7, 104. the LAM-1 frequencies of substituted n-alkanes ( fc= 4 H. Huatan, J. H. Collett, D. Attwood and C. Booth, Biomaterials, 420 N m-1).The necessary equations have been given else- 1995, 16, 1297. where.18,20 The calculated LAM-1 frequency is 29 cm-1 com- 5 P. J. Flory, Principles of Polymer Chemistry, Cornell U. P., Ithaca, pared with 32 cm-1 found. Because of hydrogen bonding of New York, 1953, p. 554. 6 Y. Chatani, Y. Okita, H. Tadokoro and Y. Yamashita, Polym. J., hydroxy groups (hydroxy–hydroxy or, more likely, hydroxy–- 1970, 1, 555.carbonyloxy), it is likely that the chain-end force has a higher 7 H.-L. Hu and D. L. Dorset, Macromolecules, 1990, 23, 4604. value than 5 N m-1, and this would increase the calculated 8 V. Crescenzi, G. Manzini, G. Calzolari and C. Borri, Eur. Polym. LAM-1 frequency. At the same time, the inertial eVect of a J., 1972, 8, 449. chain fold, which is modelled as an end mass, would decrease 9 R.Perret and A. Skoulios, Makromol. Chem., 1972, 156, 157. the LAM-1 frequency. Sensible refinements of the model to 10 See J. F. Rabolt, CRC Crit. Rev. Solid State Mater. Sci., 1977, 12, 165. account for these eVects, using equations already pub- 11 See J. V. Dawkins in Comprehensive Polymer Science, Vol. 1, lished,20,21 move the calculated LAM-1 frequency within the Polymer Characterisation, ed.C. Booth and C. Price, Pergamon, range 29–32 cm-1 but do not change the assignment. Oxford, 1989, p. 252. Further evidence for the assignment of LAM-1 came from 12 A. Schindler, Y. M. Hibionada and G. C. Pitt, J. Polym. Sci., the temperature dependence of its frequency. The following Polym. Chem. Ed., 1982, 20, 319.results were found for the three bands listed in Table 4: 13 W. Bras, G. E. Derbyshire, A. J. Ryan, G. R. Mant, A. Felton, R. A. Lewis, C. J. Hall and G. N. Greaves, Nucl. Instrum.Methods (i) frequency independent of temperature (within Phys. Res., Sect. A, 1993, 326, 587. experimental error), allowing it to be assigned to an optical 14 A. J. Ryan, J. Therm. Anal., 1993, 40, 887.mode; 15 A. J. Ryan, W. Bras, G. R. Mant and G. E. Derbyshire, Polymer, (ii) frequency dependent on temperature (see Fig. 7), much 1994, 35, 4537. as expected for LAM-1;19,22 16 W. Bras, G. E. Derbyshire, A. Devine, S. Clarke, J. Cooke, (iii) frequency dependent on temperature, allowing it to be B. U. Komanschek and A. J. Ryan, J. Appl. Crystallogr., 1995, 28, 26. tentatively assigned to a bending mode, with the evidence of 17 F.B. Khambatta, F. Warner, T. Russell and R. S. Stein, J. Polym. splitting (see Fig. 6, more obvious at -100 °C) suggesting Sci., Polym. Phys. Ed., 1976, 14, 1391. resolution of orthogonal components. 18 G. Minoni and G. Zerbi, J. Phys. Chem., 1982, 86, 4791. In that the LAM-1 frequency for a crystalline polymer 19 K. Viras, F.Viras, C. Campbell, T. A. King and C. Booth, J. Phys. depends predominantly on crystal stem length, the constant Chem., 1989, 93, 3479. frequency found for the three polymers confirms the conclusion 20 C. Campbell, K. Viras and C. Booth, J. Polym. Sci., Part B, Polym. Phys., 1991, 29, 1613. from SAXS that the stem length in those samples is constant. 21 C. Campbell, K. Viras, A.J. Masters, J. R. Craven, H. Zhang, The substantial agreement of the calculated LAM-1 frequency S. G. Yeates and C. Booth, J. Phys. Chem., 1991, 95, 4647. (based on the crystal stem length from SAXS) with the 22 K. Viras, T. A. King and C. Booth, J. Chem. Soc., Faraday Trans. observed frequency is strong confirmatory evidence that the 2, 1985, 81, 491. chains in the poly(CL) samples are normal to their lamellar end planes.As noted in the Introduction, this result from Paper 8/08718K J. Mater. Chem., 1999, 9, 1059–1063 1063 J O U R N A L O F C H E M I S T R Y Materials Chain folding in poly(e-caprolactone) studied by small-angle X-ray scattering and Raman spectroscopy. A strategy for blending in the crystalline state Susan A. Berrill,a Frank Heatley,a John H.Collett,a David Attwood,a Colin Booth,*a J. Patrick A. Fairclough,b Anthony J. Ryan,b† Kyriakos Viras,c Amanda J. Duttond and Ross S. Blundelle aDepartment of Chemistry and School of Pharmacy, University of Manchester, Manchester, UK M13 9PL bDepartment of Chemistry, University of SheYeld, SheYeld, UK S3 7UF cNational and Kapodistrian University of Athens, Chemistry Department, Physical Chemistry Laboratory, Panepistimiopolis, Athens 157 71, Greece dSolvay-Interox Ltd, PO Box 7, Baronet Road, Warrington, UKWA4 6HB eSanofi-Winthrop Research, Alnwick, Northumberland, UK NE66 2JH Received 9th November 1998, Accepted 18th February 1999 Samples of poly(e-caprolactone), including narrow fractions obtained by preparative gel permeation chromatography (GPC), were characterised by analytical GPC and 13C NMR spectroscopy.The crystallised samples were investigated by small-angle X-ray scattering to obtain lamellar spacings and by low-frequency Raman spectroscopy to obtain LAM-1 frequencies. Together the two methods gave a reliable estimate of the critical stem length for chain folding in poly(e-caprolactone) crystallised at room temperature or below, i.e. 15 chain units [CL, OCO(CH2)5] equivalent to 105 chain atoms. A necessary requirement for molecular blending of block copolymers of e-caprolactone and ethylene oxide with crystalline high molar mass poly(e-caprolactone) is cocrystallisation. Hence it was concluded that a safe lower limit for successful blending is a poly(e-caprolactone) block of 20 CL units.[poly(E), E=OCH2CH2] block must be sited in the interlamel- Introduction lar region of the stacked crystal lamellae, thus providing Poly(e-caprolactone) [poly(CL), CL=OCO(CH2)5] has pathways for permeation of water into the bulk of the material. potential as a biodegradable polyester for use in solubilisation The rate of permeation could then be controlled by varying and/or encapsulation of drugs.1 Its biodegradation is based the proportion of poly(E) in the copolymer.on hydrolysis of the ester linkages, which in turn requires In considering such a strategy, it was apparent that best permeation of water into the polymer. Since the polymer is results would be expected if the copolymers to be blended highly hydrophobic, a number of strategies have been reported were small.This in itself, by promoting the entropy of mixing, for enhancing water permeation. These include statistical would enhance miscibility. Random incorporation of many copolymerisation with less hydrophobic comonomers (e.g. DL- short poly(CL) blocks into a lamellar crystal would also mean or L-lactide),2 block copolymerisation with a hydrophilic an even distribution of short poly(E) blocks at the lamellar comonomer (e.g.ethylene oxide),3 and blending with a hydro- surfaces. Equally importantly, the preparation of block copolyphilic polymer or copolymer (e.g. with Synperonic-PE 61, an mers by sequential anionic polymerisation is relatively straightoxyethylene/ oxypropylene block copolymer).4 forward for short blocks. This being so, it was desirable to Our immediate interest is in using poly(CL) for encapsul- define the minimum length of a poly(CL) block that would ation, for which purpose blending high molar mass poly(CL) ensure cocrystallisation, since very short poly(CL) blocks with a second polymer or copolymer is attractive.However, would be rejected from the crystal. blending is complicated by two factors.One is the well-known Lamellar spacings in poly(CL) were measured many years thermodynamic immiscibility of polymers in the absence of ago by Perret and Skoulios using small-angle X-ray scattering specific attractive interactions, which means that most blends (SAXS).9 The lamellar spacing for a crystallised sample of a are at best mechanically mixed, and so are heterogeneous on high molar mass poly(CL) was found to be 165 A° , this being a local scale.5 The second is the crystallinity of poly(CL),6–8 the stem length between chain folds.Their investigation of the with the consequent exclusion of other polymers from the spacings for short chains showed that this value was closely crystalline regions. Of course crystallinity in a polymer can be approached when the poly(CL) chain length reached 30 CL an advantage, since it is a source of mechanical strength in units (here denoted CL30).However, in their experiments the a material. approach to the limiting value was gradual, probably as a Recently we have sought to define the requirements for a consequence of the chain length distribution in the fractions. stable system based on crystalline high molar mass poly(CL) It seemed possible that the limit might be reached at a shorter blended with diblock poly[e-caprolactone-block-poly(ethylene poly(CL) chain length. As described in this report, related oxide)].Compatibility with the crystalline poly(CL) was to be experiments using fractions with very narrow chain length ensured through cocrystallisation of the poly(CL) block of the distributions showed this to be so.Investigation of the longicopolymer. Given cocrystallisation, the poly(oxyethylene) tudinal acoustic mode (LAM) vibration by low-frequency Raman spectroscopy provided useful confirmatory evidence for this conclusion, since the single-node longitudinal vibration †Also CCLRC Daresbury Laboratory, Warrington, UK WA4 4AD. J. Mater.Chem., 1999, 9, 1059–1063 1059(LAM-1) is directly related to the crystal-stem length, and so guards against a false conclusion from SAXS data caused by crystal-stem tilting in the lamella.10 Experimental Polymers: preparation and characterisation Three poly(CL) samples with number-average molar mass in the range Mn=2000–4000 g mol-1 (chain length n=17–35 CL units) were selected from the range produced by Solvay Interox, Widnes, UK.These samples had fairly narrow chain length distributions (Mw/Mn=1.2–1.4). A fourth high molar mass polymer (Mn#80000 g mol-1, Mw/Mn=1.5) was also examined in order to fix an upper limit to the lamellar spacing. The anionic polymerisation of these samples was initiated by butanediol. In what follows we denote polymers as P-CLn, where n is the number-average chain length in CL units.Fig. 1 GPC curves obtained for poly(e-caprolactone) fraction F-CL14 Polymer fractions (see below) are denoted F-CLn. using GPC systems A and B, as indicated. A polymer of low molar mass (Mn#450 g mol-1, P-CL4) was prepared in our laboratory, starting from pentanediol. e- Caprolactone (Solvay Interox) was stirred overnight with 2,4- Table 1 Molecular characteristics of the poly(CL) samples diisocyanato-1-methylbenzene (2 wt%, Fluka AG), distilled Mw/Mn a Mn/g mol-1 under reduced pressure (140 °C, 15 mm Hg) directly onto Sample (GPC) (NMR/GPC) calcium hydride (Fischer Scientific), and then redistilled onto activated type 4 A° molecular sieves for storage.Immediately Polymers before use, pentane-1,5-diol (Lancaster Chemicals) was dried P-CL19 1.26 2200 by warming in vacuo for 24 h before being distilled at reduced P-CL25 1.41 2900 P-CL38 1.41 4300 pressure.The catalyst, stannous 2-ethylhexanoate (Sigma), P-CL700 1.46 80000b was used as provided. For polymerisation, pentanediol (3.5 g, Fractions 0.03 mol) and e-caprolactone (11.6 g, 0.1 mol) were syringed F-CL5 1.02 590 into a dry ampoule containing a magnetic stirrer.After F-CL9 1.02 1060 addition of a drop of the catalyst, the contents of the ampoule F-CL11 1.02 1300 were degassed, and the reaction mixture then stirred in a sand F-CL14 1.02 1630 F-CL16 1.02 1800 bath at 160±5 °C for 7 h. Characterisation by 13C NMR spectroscopy and analytical GPC gave Mn=450 g mol-1 and aMw/Mn to ±0.01, Mn to ±2%. bValue supplied by Solvay Interox.Mw/Mn=1.26. Polymer P-CL4 was separated by preparative gel permeation chromatography (preparative GPC) into fractions with very 13C NMR spectra were obtained using a Varian Associates narrow chain length distributions. A single 500 A° Styragel Unity 500 spectrometer operated at 125.8 MHz. Samples were column (120 cm×57 mm, Waters Assoc.) was used.Polymer dissolved in deuterochloroform, concentration 0.1 g cm-3. (0.5 g) in 20 cm3 toluene was injected into the solvent stream Analysis of the NMR spectra varied slightly depending on the (flow rate 10 cm3 min-1) at room temperature. Emergence of initiator used in preparation of the sample. Both commercial sample was detected by diVerential refractometry, upon which polymers and fractions conformed to the general formula 20×50 cm3 volumes were collected.The fractions of polymer CL*CLxICLxCL*, where were isolated by evacuation of solvent (24 h, 10-3 mm Hg, CL=CO·CH2CH2CH2CH2CH2O room temperature). In all, twenty fractions were collected, many very small, and just five were used in further work. a b c d e f The polymers and fractions were analysed by GPC (for CL*=CO·CH2CH2CH2CH2CH2OH distribution width) and 13C NMR spectroscopy (for numberaverage chain length).Two GPC systems were used. System a b c d g h A comprised three PLgel columns (Polymer Laboratories, I=OCH2CH2CH2CH2O or OCH2CH2CH2CH2CH2O 1×500 A° and 2×mixed B) used with tetrahydrofuran (THF) eluent (room temperature, 1 cm3 min-1, dodecane flow i j j i f e k e f marker).Samples were injected as 2 g dm-3 solutions via a 100 mm3 loop, and detected by diVerential refractometry. The chemical shifts (d) for the carbons as labelled are listed in Table 2: as indicated, those in the spectra of the low molar Universal calibration11 was with polystyrene standards and the Mark–Houwink constants for poly(styrene) (K=0.932×10-4, mass samples varied somewhat because of end eVects.With the exception of the resonance of the CO carbon (a), which a=0.740) and poly(e-caprolactone) (K=1.395×10-4, a= 0.786) in THF at room temperature advocated by Schindler had a low integral because of its longer relaxation time and lower NOE, the integrals of equivalent carbons were identical et al.12 System B comprised four PLgel columns (all mixed E) used with chloroform eluent (room temperature, 0.3 cm3 within experimental error.The sum of integrals of all resonances (except a), appropriately related to the sum of integrals min-1), other features being similar to those of System A. System B provided excellent resolution in the low-molar-mass of resonances from end group and initiator residue carbons, gave the values of Mn listed in Table 1 for the polymers and range.As examples, GPC curves obtained using systems A and B for analysis of fraction F-CL14 are shown in Fig. 1. The for fractions F-CL14 and F-CL5. There was no evidence of ring formation. curve from System A indicates a narrow chain length distribution (polydispersity index Mw/Mn#1.02) while that from Because of the small amounts of material available, only two fractions were characterised in this way.Molar masses of System B shows the sample to consist predominantly of three oligomers. Values of Mw/Mn are listed in Table 1. other fractions were obtained by use of analytical GPC (System 1060 J. Mater. Chem., 1999, 9, 1059–1063Table 2 13C NMR chemical shifts (CDCl3) SAXS detector, and high-density polyethylene and an NBS silicon standard were used to calibrate the WAXS detector.Carbon d/ppm The data acquisition system had a time-frame generator which collected the SAXS/WAXS data in 6 s frames separated by a a 173.5 wait-time of 10 ms. Parallel plate ionisation detectors placed b 34.1 c 25.5 before and after the sample cell recorded the incident and d 24.5 transmitted intensity. The experimental data were corrected e 28.3 for background scattering (subtraction of the scattering from f 64.1 the camera, hot stage and an empty cell ), sample thickness g 32.3 and transmission, and for any departure from positional h 62.5 linearity of the detectors.Since the samples crystallised with i 68.9 j 21.7 small domains randomly arranged the resulting pattern was k 22.2 analogous to that of a powder, and all orientations were adequately sampled in the one-dimensional experiment.Range: ±0.1 ppm for polymers (Mn>1000 g mol-1); ±0.5 ppm for fractions (Mn<1000 g mol-1). Further details of the equipment and methods can be found elsewhere.13–16 A) as follows. (i) Values of molar mass at the peak of the Low-frequency Raman spectroscopy GPC curve (Mpk) were calculated as Mpk=Mn(Mw/Mn)1/2, Raman scattering at 90° to the incident beam was recorded with Mn from NMR spectroscopy and Mw/Mn from GPC.by means of a Spex Ramalog spectrometer fitted with a 1403 This provided a common basis for plotting the data for the double monochromator plus a third (1442U) monochromator polymers and fractions. (ii) Note was taken of the eVect of operated in scanning mode.The light source was a Coherent hydrogen bonding of THF with the OH end groups of the Innova 90 argon-ion laser operated at 514.5 nm and 300 mW. samples. From comparison of the elution of low molar mass Typical operating conditions were bandwidth 0.8 cm-1, scan- polyethylene glycols and their dimethyl ethers, this eVect was ning increment 0.05 cm-1, integration time 6–12 s.The fre- found to decrease the elution volume consistent with an quency scale was calibrated by reference to the 9.4 and approximate correction to Mn of each sample of 140 g mol-1. 14.9 cm-1 bands in the spectrum of L-cystine. Samples in a The resulting calibration curve is shown in Fig. 2. With thin-wall capillary were melted and recrystallised at room acceptable scatter, the three commercial samples and the two temperature (approximately 20 °C).Spectra were recorded fractions fit to a common curve. Accordingly values of Mpk with the samples at various temperatures in the range 22 to for the fractions were obtained from their elution volumes in -100 °C, controlled to±1 °C by means of a Harney-Miller System A and converted to values of Mn by reversing the cell (Spex Industries).procedures (i) and (ii). These values, for fractions F-CL9, FCL11 and F-CL16, are included in the last column of Table 1. Results and discussion X-Ray scattering The WAXS patterns obtained were as expected for crystalline Measurements were made on beamline 8.2 of the SRS at the poly(CL), e.g. strong reflections at Bragg angle h=10.4° and CCLRC Daresbury Laboratory, Warrington, UK.The camera 11.6°, and weaker reflections also in keeping with expectation.6 was equipped with a multiwire quadrant detector (SAXS) located 3.5 m from the sample position and a curved knife- Lamellar spacings from SAXS edge detector (WAXS) that covered 70° of arc at a radius of Representative SAXS patterns are shown in Fig. 3 and 4. 0.3 m.Samples were sealed into TA Instruments DSC pans Fig. 3 shows a time-resolved relief diagram of SAXS data containing a 0.75 mm brass spacer ring and fitted with windows obtained during a melting–recrystallisation cycle (ramp rate= made from 25 mm thick mica. The loaded pans were placed in 10 °Cmin-1) for sample P-CL25. Intensity is plotted against the cell of a Linkam DSC of single-pan design, which could temperature and scattering vector q=(4p/l)sinh, where the be heated and cooled to allow melting and recrystallisation of wavelength l=1.54 A° .In this notation, the lamellar spacing the samples. The scattering pattern from an oriented specimen is given by Bragg’s Law in the form d=2p/q*, where q* is the of wet collagen (rat-tail tendon) was used to calibrate the value of q at the first-order peak.Fig. 4 shows the scattering Fig. 3 Time-resolved relief diagram of SAXS data obtained during a melting–recrystallisation cycle (ramp rate=10 °Cmin-1) for sample P-CL25. Intensity (arbitrary scale) is plotted against temperature and Fig. 2 GPC calibration curve (System A) for poly(e-caprolactone) scattering vector q=(4p/l)sinh, where l=1.54 A° and h is the scattering angle.($) polymers and (&) fractions. J. Mater. Chem., 1999, 9, 1059–1063 1061the orthorhombic sub-cell dimension c=17.05 A° or c=17.26 A° (for 2 units per repeat) respectively reported by Chatani et al.6 and Hu and Dorset,7 and so consistent with the crystal stems being normal to the lamellar end planes. On the basis of the SAXS data, we conclude that the chain length between folds (crystal stem length) in low molar mass poly(CL) (Mn<4500 g mol-1) crystallised at low temperatures is approximately 15 CL units.Extrapolation of the line established for the lower molar mass fractions to d=150 A° , the value of d measured in this work for the high molar mass sample, would suggest 17 CL units, while extrapolation to d= 165 A° (Perret and Skoulios9) would take this value to 19 CL units.Values of d reported by Khambatta et al.17 for moderate molar mass poly(CL) (Mn=13000 g mol-1) are near to 150 A° . It seems that a safe lower limit to ensure that a short CL chain will cocrystallise with chain-folded lamellar crystals of high molar mass poly(CL) is 20 CL units, equivalent to 140 chain atoms.LAM-1 frequency from Raman spectroscopy Fig. 4 Scattering intensity (arbitrary scale) plotted against the ratio q/q* for fraction F-CL14 at 10 °C. The parameter q* is the value of q The three low molar mass polymers were examined by Raman at the first-order peak: in this case q*=0.055 A° -1. spectroscopy. Examples of low-frequency (<100 cm-1) spectra are shown in Fig. 6. There are three prominent bands in each spectrum, the frequencies of which are labelled (i) to (iii) in intensity plotted against the ratio q/q* for sample F-CL14 at Table 4. The spectra of the samples P-CL25 and P-CL38 are 10 °C, at which temperature q*=0.055 A° -1.essentially identical. The spectrum of P-CL19 diVers at the Values of the lamellar spacing are listed in Table 3 and lowest frequencies, mainly by the reduced intensity of the plotted against chain length (in CL units) in Fig. 5. Also lowest frequency band (iii), but with some evidence of splitting. plotted in Fig. 5 are the results of Perret and Skoulios.9 The Presumably this is a result of end eVects from a,v-hydroxy- two data sets are in substantial agreement. The sharp break ended chains in predominantly unfolded-chain lamellar in the curve in the present results at chain length 15 CL units crystals. and d=129 A° is attributable to the use of very narrow The LAM-1 band was assigned to band (ii) as follows.fractions, Mw/Mn#1.02, in this work. The d-spacing of 129 A° Adapting the model of Minoni and Zerbi,18 the lamellar stack corresponds to 8.6 A° per CL unit, which is in keeping with was represented as a one-dimensional crystal, with both chain stems and stem-end groups (chain ends or chain folds) Table 3 Lamellar spacings for crystalline poly(CL) samples at 20 °C Sample d/A° a P-CL19 133 P-CL25 128 P-CL38 128 P-CL700 149 F-CL9 82 F-CL11 103 F-CL14 117 F-CL16 128 ad to ±3 A° .Fig. 6 Low-frequency Raman spectra of poly(e-caprolactone) samples (as indicated) at 20 °C.The intensity scale and baseline levels are arbitrary. Table 4 Band frequencies (cm-1) from low-frequency Raman spectra of crystalline poly(CL) samples at 20 °C Sample (i) (ii) (iii) P-CL19 58.0 31.8 21.1 P-CL25 57.9 31.6 19.6 Fig. 5 Lamellar spacing from SAXS (d) versus chain length in CL P-CL38 57.9 31.8 19.4 units for poly(e-caprolactone) ($) polymers and (&) fractions.The Estimated uncertainties: (i) to ±0.2 cm-1, (ii) to ±0.1 cm-1, (iii) to open symbols (#) denote the results of Perret and Skoulios.9 The full ±0.5 cm-1. lines are drawn through the data points for the present samples. 1062 J. Mater. Chem., 1999, 9, 1059–1063Raman spectroscopy guards against the possibility that the lamellar spacing from SAXS is aVected by crystal-stem tilting.Concluding remarks We are aware that all the samples examined in the present work have an imperfection at the mid-point of the chain. The choice of pentanediol as initiator does not remove this feature, since (on average) the two arms of those chains are symmetrical about the mid-point. While this may aVect the crystallinity of the present samples, we find no evidence that it aVects the extent of chain folding as quantified by SAXS, nor the LAM- 1 frequency from Raman spectroscopy.We conclude that 20 CL units is a safe lower limit for the CL block length in block copolymers of e-caprolactone and ethylene oxide if they are to cocrystallise with chain-folded high molar mass poly(e-caprolactone) and so form blends suitable for encapsulation purposes.Fig. 7 Temperature dependence of the LAM-1 frequency of poly(ecaprolactone) P-CL19. Acknowledgements We thank Messrs S. K. Nixon and P. Kobryn for help with the experimental work. Sanofi-Winthrop, Alnwick, UK accounted for as point masses with appropriate interactions financed a research studentship for SAB. KV had the benefit between them. The SAXS results indicate a spacing equivalent of a Royal Society of Chemistry Journals Grant for to 15 CL units, equivalent to 105 atoms.In our approximate International Authors. calculation, the crystal stem was assigned 105 point masses each of 1.895×10-25 kg (average value for the CL unit), and References the force between chain ends was equated to that typical of a van der Waals interaction ( fe=5 Nm-1).19 Noting the almost 1 C.G. Pitt in Biodegradable Polymers as Drug Delivery Systems, trans-planar conformation of the poly(CL) chain and the ed. M. Chasin and R. Langer, Marcel Dekker, London, 1990. 2 E. Piskin, J. Biomater. Sci., Polym. Ed., 1994, 6, 775. similarity of its packing in the crystal,6 the force between chain 3 J. Bei, W. Wang, Z. Wang and S. Wang, Polym. Adv. Technol., masses was set equal to the value successfully used to model 1996, 7, 104.the LAM-1 frequencies of substituted n-alkanes ( fc= 4 H. Huatan, J. H. Collett, D. Attwood and C. Booth, Biomaterials, 420 N m-1). The necessary equations have been given else- 1995, 16, 1297. where.18,20 The calculated LAM-1 frequency is 29 cm-1 com- 5 P. J. Flory, Principles of Polymer Chemistry, Cornell U.P., Ithaca, pared with 32 cm-1 found. Because of hydrogen bonding of New York, 1953, p. 554. 6 Y. Chatani, Y. Okita, H. Tadokoro and Y. Yamashita, Polym. J., hydroxy groups (hydroxy–hydroxy or, more likely, hydroxy–- 1970, 1, 555. carbonyloxy), it is likely that the chain-end force has a higher 7 H.-L. Hu and D. L. Dorset, Macromolecules, 1990, 23, 4604. value than 5 N m-1, and this would increase the calculated 8 V.Crescenzi, G. Manzini, G. Calzolari and C. Borri, Eur. Polym. LAM-1 frequency. At the same time, the inertial eVect of a J., 1972, 8, 449. chain fold, which is modelled as an end mass, would decrease 9 R. Perret and A. Skoulios, Makromol. Chem., 1972, 156, 157. the LAM-1 frequency. Sensible refinements of the model to 10 See J. F. Rabolt, CRC Crit.Rev. Solid State Mater. Sci., 1977, 12, 165. account for these eVects, using equations already pub- 11 See J. V. Dawkins in Comprehensive Polymer Science, Vol. 1, lished,20,21 move the calculated LAM-1 frequency within the Polymer Characterisation, ed. C. Booth and C. Price, Pergamon, range 29–32 cm-1 but do not change the assignment. Oxford, 1989, p. 252. Further evidence for the assignment of LAM-1 came from 12 A. Schindler, Y. M. Hibionada and G. C. Pitt, J. Polym. Sci., the temperature dependence of its frequency. The following Polym. Chem. Ed., 1982, 20, 319. results were found for the three bands listed in Table 4: 13 W. Bras, G. E. Derbyshire, A. J. Ryan, G. R. Mant, A. Felton, R. A. Lewis, C. J. Hall and G. N. Greaves, Nucl. Instrum.Methods (i) frequency independent of temperature (within Phys. Res., Sect. A, 1993, 326, 587. experimental error), allowing it to be assigned to an optical 14 A. J. Ryan, J. Therm. Anal., 1993, 40, 887. mode; 15 A. J. Ryan, W. Bras, G. R. Mant and G. E. Derbyshire, Polymer, (ii) frequency dependent on temperature (see Fig. 7), much 1994, 35, 4537. as expected for LAM-1;19,22 16 W. Bras, G. E. Derbyshire, A. Devine, S. Clarke, J. Cooke, (iii) frequency dependent on temperature, allowing it to be B. U. Komanschek and A. J. Ryan, J. Appl. Crystallogr., 1995, 28, 26. tentatively assigned to a bending mode, with the evidence of 17 F. B. Khambatta, F. Warner, T. Russell and R. S. Stein, J. Polym. splitting (see Fig. 6, more obvious at -100 °C) suggesting Sci., Polym. Phys. Ed., 1976, 14, 1391. resolution of orthogonal components. 18 G. Minoni and G. Zerbi, J. Phys. Chem., 1982, 86, 4791. In that the LAM-1 frequency for a crystalline polymer 19 K. Viras, F. Viras, C. Campbell, T. A. King and C. Booth, J. Phys. depends predominantly on crystal stem length, the constant Chem., 1989, 93, 3479. frequency found for the three polymers confirms the conclusion 20 C. Campbell, K. Viras and C. Booth, J. Polym. Sci., Part B, Polym. Phys., 1991, 29, 1613. from SAXS that the stem length in those samples is constant. 21 C. Campbell, K. Viras, A. J. Masters, J. R. Craven, H. Zhang, The substantial agreement of the calculated LAM-1 frequency S. G. Yeates and C. Booth, J. Phys. Chem., 1991, 95, 4647. (based on the crystal stem length from SAXS) with the 22 K. Viras, T. A. King and C. Booth, J. Chem. Soc., Faraday Trans. observed frequency is strong confirmatory evidence that the 2, 1985, 81, 491. chains in the poly(CL) samples are normal to their lamellar end planes. As noted in the Introduction, this result from Paper 8/08718K J. Mater. Chem., 1999, 9, 1059–1063 1063

ISSN:0959-9428    

DOI:10.1039/a808718k

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Electropolymerization of hydrophobic dipyrrolyls in aqueous medium based on inclusion chemistry Kathleen I. Chane-Ching,* Jean-Christophe Lacroix, Mohamed Jouini and Pierre-Camille Lacaze ITODYS, Universite� Paris 7-Denis Diderot, associe� au CNRS, 1 rue Guy de la Brosse, 75005 Paris, France. E-mails: K. I. Chane-Ching: chane@paris7.jussieu.fr, J.C. Lacroix: lacroix@paris7.jussieu.fr, M. Jouini: jouini@paris7.jussieu.fr, P. C. Lacaze: lacaze@paris7.jussieu.fr Received 6th January 1999, Accepted 25th February 1999 The host–guest complexation of hydrophobic dipyrrolyl monomers 1–4 with b-cyclodextrin has been studied in aqueous solution by means of 1H NMR spectroscopy, chemical analysis and electrochemical techniques.In each case, 152 stoichiometry is found, the cyclodextrin cavities being likely located on the extremities of the monomers. Polymeric materials were obtained from such inclusion compounds in water by either electrochemical or chemical oxidation. ESCA and elemental analyses show that the host molecule is absent from these polymers, whose characteristics are quite similar to those synthesized in organic media.A scheme for polymerization in the presence of cyclodextrin is proposed. subsequent behaviour of such complexes when they are Introduction oxidized by electrochemical or chemical means. Organizing the encounter of reagents via molecular self-assembly is an important challenge in chemistry. Within Results this context, some of the simplest assemblies start with cyclodextrins (CD) which are cyclic oligosaccharides formed by six Inclusion complexes (aCD), seven (bCD) or eight (cCD) D-glucopyranose units All the dipyrrolyl monomers studied here (see Scheme 1) are joined together by a-(1,4) bonds.Their ability to form stable poorly soluble in water (10-3, 2×10-4, <10-4 and <10-4M inclusion complexes with hydrophobic molecules in aqueous for 1, 2, 3 and 4, respectively); this can be improved by the solution has been established1 and has led to many appliaddition of 0.1M of HPbCD (1.2×10-2, 8×10-3, 4×10-3 cations.2 Most previous electrochemical studies of host–guest and 8×10-4 M, respectively). This behaviour implies the CD solutions fall into one of two classes: selective electrosynthformation of inclusion complexes between the dipyrrolyl esis based on CD complexation reactions3 and electroanalytical derivatives and the CD. 1H NMR is the most reliable technique techniques for studying the ability of cyclodextrins to hold, available to characterize this phenomenon. Studies on the two orient, conceal or separate guest molecules.4 The electrochemiinner (H-3 and H-5) hydrogen atoms of CDs in the presence cal behaviour of electroactive species in the presence of CD of a guest molecule provide valuable proof of inclusion.9 Fig. 1 has also been studied for the characterization of organized displays the characteristic upfield shifts of the H-3 and H-5 layers consisting of host–guest complexes5 or CD-grafted protons upon complexation with 1. These small but significant polypyrrole films capable of molecular recognition.6 Recently, shifts (0.024 and 0.042 ppm for H-3 and H-5, respectively) are CD complexes have been used to electropolymerize a hydrorelated to an inclusion phenomenon with a low association phobic monomer in water, hydroxypropyl-b-cyclodextrin constant.10 Such shifts are observed in the 6–7 ppm spectral (HPbCD) being the host and bithiophene the guest.7 HPbCD region corresponding to the pyrrolyl proton resonance.was chosen since it is more soluble in water than simple bCD: Conversely, the singlet peak at 3.58 ppm corresponding to the the presence of hydroxypropyl groups on the external crown two central methylene groups of the spacer does not shift of the CD increases considerably the host solubility without significantly upon complexation, indicating that the CDs inter- modification of the hydrophobicity of the cavity.Cyclodextrins act much less with the aliphatic spacer chain. The stoichiometry remain in the polymer film, suggesting that partial encapsulof the complexes was determined by isolating the very slightly ation of the polymer chains occurs. This behaviour can be water-soluble bCD inclusion compounds.In every case 152 explained by a high aYnity of the hydrophobic inner cavity stoichiometry is found by chemical analysis (see Experimental of the CD macrocycles for the growing chains of polybithiosection). These results are confirmed by 1H NMR spectroscopy phene which favours the formation of poly-pseudorotaxane in DMSO-d6 where the host molecules are unthreaded, and structures during electropolymerization.the resulting signals correspond to the free species. For We have shown in a previous paper8 that certain dipyrrolyl monomers can be used to form polymers presenting ladderlike structures, these monomers consisting of more or less rigid spacers, the extremities of which are bonded to the N atom of two pyrrolyl entities. The use of a rigid spacer leads undoubtedly to a ladder-structure whereas with long, flexible spacers intramolecular coupling may occur during polymer formation.Thus, depending on the structure of the complexes, encapsulation of such monomers by CDs could be a good way to make the spacers inflexible, or CD could be expelled into the bulk during the polymerization process. This work is devoted to the study of the interaction of CD Scheme 1 Monomers 1–4.with dipyrrolyls having long, flexible spacers and to the J. Mater. Chem., 1999, 9, 1065–1070 1065Fig. 2 Partial 200 MHz 1H NMR spectrum of 3–bCD inclusion complex in DMSO-d6. Chemical shifts in ppm; 152 stoichiometry is deduced from integration of monomer and CD peaks. Electrochemical oxidative polymerization The most water-soluble monomer complex ‘1–CD’ was first investigated electrochemically at 20 °C in water.The voltammetric behaviour of 1 is characterized by the irreversible oxidation of the pyrrolyl group as in an organic medium.8 Whereas monomer 1 oxidizes at 1.21 V in acetonitrile, its oxidation potential becomes 1.09 V in water in the presence of a 10-fold excess of HPbCD. This potential shift could be explained by a greater stabilization of the cation-radical in aqueous medium.As can be seen in Table 1, addition of bCD causes the peak oxidation current to decrease, indicating that much of the monomer exists as the more slowly diVusing inclusion complex, since no change in the solution viscosity is expected over the range of 0–10 mM bCD.13 Simultaneously, the oxidation peak potential of 1 is shifted substantially to a more positive value, indicating that 1 is more diYcult to oxidize in the presence of bCD.Such eVects are also observed with HPbCD addition. Evans et al. observed the same behav- Fig. 1 Partial 200 MHz 1H NMR spectra of bCD in D2O: (a) no additive; after addition of 1 with [host/guest] ratios of (b) 3.3, (c) 1.1 iour in voltammetric studies of ferrocenecarboxylic acid in the and (d) 0.25 (only host protons are assigned).presence of bCD and conclusively demonstrated the prevalence of a CE (C: chemical; E: electrochemical ) mechanism13 in which the complex first dissociates followed by the oxidation example, Fig. 2 exemplifies the results obtained with the of the free ferrocene derivative. We conclude that the same 3–bCD complex: the peak at 4.83 ppm corresponds to the mechanism applies for the oxidation of the dipyrrolyl seven H1 protons of the cyclodextrin and those at 5.95 and derivative (see Scheme 2). 6.72 ppm correspond to the four HB and HA of the dipyrrolyl Regular film formation at 20 °C by electrochemical oxidation derivative, respectively. The relative integrated intensities of of a 10-2 M solution of monomer 1 in water and in the these peaks give the proton ratio (3.236/7)5(1.05/4), that is absence of HPbCD is impossible for solubility reasons: the nearly 251.The same results are obtained with the other electrode surface is covered with submicronic polymer islets monomers. It should be noted that spacer length does not (Fig. 3), resulting from the perization of droplets of 1 in aVect the stoichiometry of the complex and, conversely, the suspension in water.It becomes possible when HPbCD is longer the spacer, the less the resulting complex is water- added at a ratio of 10 HPbCD per monomer, using the soluble. This, added to the NMR study described above, galvanostatic method or cyclic voltammetry. Fig. 4 shows a indicates that the two pyrrolyl rings of the monomers constitute multicycle voltammogram recorded between 0 and 1.0 V at a the complexation sites, the spacer remaining free.This assump- scan rate of 50 mV s-1. The first anodic sweep reveals a wave tion is in good agreement with a study of the complexation of ferrocene derivatives containing n-alkyl chains, where it was Table 1 Current and potential peaks measured on a carbon graphite demonstrated that bCD interacts much less eVectively than electrode immersed in 0.1 M KNO3 containing 1.0 mM 1 and various aCD with aliphatic chains.11 Any attempt to obtain complexes concentrations of HPbCD.Scan rate: 50 mV s-1; reference electrode: of the monomers with aCD in good yields failed, although SCE. aCD is known to complex alkyl or PEG chains,12 possibly 103 c HPbCD/M i/mA E/V because it is diYcult for the cavity to thread into the pyrrolyl ring and reach the spacer chain. 0 61.5 1.05 In what follows, we will devote our attention to the most 1 53 1.07 soluble complex, being given the necessary concentrations to 10 42 1.09 carry out polymerization reactions under good conditions. 1066 J. Mater.Chem., 1999, 9, 1065–1070Scheme 2 Proposed mechanism of oxidative polymerization of 1 in presence of HPbCD. with an onset at 0.75 V, corresponding to the oxidation of 1; additional anodic and cathodic peaks increase regularly with the number of successive cycles. The electroactivity of the resulting poly-1 films in water is similar to that for films obtained in acetonitrile (Fig. 5). This behaviour is quite unusual for a polypyrrole (PPY) derivative. For instance, when PPY is synthesized in organic medium, the cyclability performance of the film in water is much worse than in the synthesis solvent. The presence here of ether functions, which are well known to increase the ionic conductivity of the polymer,14 in the crosslinking spacers probably induces a permeability suYcient to allow free movements of the counter ions.This could explain the great reversibility of the oxidoreduction of the poly-1 in water independently of the medium used for the electrosynthesis. Fig. 4 Potentiodynamic multisweep electrodeposition and redox Surface analyses of these films in the oxidized state have cycling (50 mV s-1) of poly-1 on Pt in aqueous 0.1 M LiClO4 and been performed by XPS in order to estimate the doping level 0.1 M HPbCD (S=1.5 cm2).Monomer concentration: 10 mM; reference electrode: SCE. Fig. 5 Electroactivity in aqueous 0.1 M LiClO4 at diVerent scan rates of poly-1 films generated in galvanostatic mode using a charge of 132 mC cm-2 from: (a) 10 mM1 aqueous+0.1 MHPbCD; (b) 10 mM 1 in acetonitrile. and to determine whether CD remained in the material.The results are given in Table 2. The comparison with theoretical stoichiometry strongly suggests that no CD remains, and that the doping level is in the order of 15%, this being slightly lower than the value of 20% found previously for such films formed in acetonitrile.8 The observed excess of oxygen atoms in the materials may be provided from an over-oxidation process during the electrosynthesis, as evidenced by IR spectroscopy (see below).The perpendicular conductivity of free standing films, ca. 1 mm thick, was measured: poly-1 films have a conductivity of 6×10-6 and 4×10-7 S cm-1 when Fig. 3 SEM photographs of poly-1 deposited on platinum from a generated in the galvanostatic mode at 0.1 mA cm-2 in aceto- 10 mM 1 solution in water; top: without HPbCD; bottom: with 0.1 M HPbCD.Supporting electrolyte: 0.1 M LiClO4. nitrile and water+HPbCD, respectively. These modest values, J. Mater. Chem., 1999, 9, 1065–1070 1067Table 2 Empirical formulae deduced from ESCA analysis of thin films of poly-1 films in oxidized state and calculated by assuming no CD incorporation and a ratio of 0.15 ClO4- per pyrrolyl ring (i.e.a doping level of 15%). Sample (1) was obtained in galvanostatic mode by cycling between -0.2 and 1.0 V at a scan rate of 50 mV s-1, and sample (2) was obtained in galvanostatic mode using a current of 0.2 mA cm-2 and a total charge of 72 mC cm-2. Monomer concentration: 10-2 M in 0.1 M HPbCD and 0.1 M LiClO4 aqueous solution Sample Empirical formula Calcd.for poly-1+30% ClO4- C14N2O3.2Cl0.3 Found for (1) C14N2.3O4.7Cl0.28 Found for (2) C14N1.9O4.3Cl0.24 when compared with classic poly-N-alkylpyrrole,16 are due to a great steric eVect of the spacer which could highly disturb the ring–ring planarity. However, these results are in accordance with the respective doping levels of the two polymers. Mass measurements were compared with the theoretical masses for 100% yield and without CD incorporation.A doping level of 15% and a consumption of four electrons per monomer are assumed (the two pyrrolyl moieties are assumed Fig. 6 A and B: IR spectra of reduced poly-1 films deposited on to be involved in the polymerization process), as previously platinum in potentiodynamic mode from 10 mM of 1 in: (A) aceto- found in acetonitrile.8 Table 3 gives the masses of some polynitrile; (B) water+0.1 M HPbCD; C: IR spectrum of HPbCD (1% 1 generated in water at diVerent current densities.Although dispersion in KBr). the polymerization yields do not always reach 100%, possibly due to some oligomer dissolution, these new results confirm that cyclodextrin is not incorporated in the films. Table 4 Frequencies (cm-1) and assignments of the principal bands observed on thin films of poly-1 deposited on a platinum electrode in Moreover, the IR spectra of such poly-1 films are very either acetonitrile or aqueous 0.1 M HPbCD similar to those generated in acetonitrile, as can be seen in Fig. 6: the two spectra show a series of peaks characteristic of Band position/cm-1 Assignments the polymer without CD (see Table 4).The dramatic increase in intensity of the IR band near 1700 cm-1 observed in the 1100 n (CH2–O–) film generated in water may be due to the introduction of 1300 n N–C(H2) 1448 d CH2 carbonyl groups as end groups which is favoured in this 1585 n CLC out-of-phase medium15 whereas the weak band at 3360 cm-1 is attributable 1705 n CLO (default) to residual water.Nevertheless, if substantial amounts of 2865 n CH2 HPbCD were incorporated in great quantity in poly-1 films 3115 n (CH ring) obtained in aqueous medium, the bands at ca. 1100 and n=stretching vibration. d=deformation vibration. 3400 cm-1 would be at least as strong as that at 2900 cm-1, as can be seen in the IR spectrum of HPbCD (see Fig. 6). This is not the case here, which confirms the absence of polymerization process and aVords finally a material similar constitutive CD from the material.to that generated in an organic medium where ladder-like structures are formed. Chemical oxidative polymerization In principle, two structures are possible for the CD Monomers 1 and 2 were chemically polymerized in water in complexes: i) an association of the monomer at the outside of the presence of HPbCD.The other monomers were not the cavity, and ii) an inclusion inside the ring. 1H NMR investigated in this way for reasons of solubility. The synthesis measurements show that the inner protons of the CD are conditions are given in the Experimental section. In both aVected upon complexation and that the partners form 152 cases, a black powder was collected and analysed.The results complexes. These results make mechanism ii) more likely. are given in Table 5. When the measured percentage of each Moreover, the materials formed either electrochemically or by element is compared with the theoretical values based on chemical polymerization do not contain constitutive CD. This various assumed compositions, only that with no CD incorpor- behaviour is consistent with a complex in which only the two ation and a doping level of 15% fits well.These new results pyrrolyl entities are encapsulated, the spacer remaining free. are in good agreement with those obtained by electropolymer- We can expect such complexes to be formed according to ization, i.e. although the starting monomer is encapsulated in cyclodextrin in aqueous medium, it is unthreaded during the Table 5 Results of elemental analyses of poly-1 and poly-2 obtained in aqueous HPbCD by chemical polymerization, using Fe3+ as Table 3 Calculated masses (mth) by assuming no CD incorporation oxidising agent.Theoretical percentages are calculated by assuming a and a doping level of 15%, and experimental masses (mexp) of deposited doping level of 15% and total absence of CD films of poly-1 at varying current densities.E is the observed deposition potential. Electropolymerization was performed in aqueous solutions X Elemental analysis (%) of 10-2 M monomer with 0.1 M HPbCD and 0.1 M LiClO4 1 Found: C: 59.8, H: 5.8, N: 9.5, O: 21.0, Cl: 3.2 j/mA cm-2 E/V Q/C cm-2 mth/mg ±0.04 mexp/mg Calcd.: C: 61.3, H: 5.8, N: 10.2, O: 18.7, Cl: 3.9 2 Found: C: 67.4, H: 7.3, N: 10.1, O: 10.5, Cl: 4.0 0.06 0.63 0.443 0.74 0.89 Calcd.: C: 69.5, H: 6.6, N: 11.6, O: 7.9, Cl: 4.4 0.21 0.67 0.625 0.59 0.54 0.545 0.8�1.03 0.227 0.16 0.14 X: starting monomer. 1068 J. Mater. Chem., 1999, 9, 1065–1070Table 6 Conditions for synthesis of poly-1 and poly-2 by chemical equilibria (1) and (2), oxidation S+L=SL (1) Synthesis solution Added dropwise T/ °C Yield (%) SL+L =SL2 (2) 0.1 M HPbCD+0.01 M 1 0.046M Fe(ClO4)3 20 25 where S and L denote Substrate (monomer) and Ligand (CD), 0.4 M HPbCD+0.04 M 2 0.200M Fe(ClO4)3 50 30 respectively.Even if the monomer is made soluble by complexation of the two pyrrolyl rings, the resulting complex is For quantitative analysis, each peak area was corrected by the subject to association/dissociation phenomena.Oxidation of appropriate experimentally determined sensitivity factor. the pyrrolyl rings involves only the free species, and this is The electrochemical studies were carried out in a immediately followed by a coupling step which leads to single-compartment three-electrode cell using an EG&G PAR insoluble oligomers.Since the polypyrrole chains are substi- 362 potentiostat in the potentiodynamic or galvanostatic mode. tuted on the N atoms of each pyrrole ring, leading to a ladder- Glassy carbon (disk area=7.1×10-2 cm2) and platinum like structure, no binding site is available for subsequent coated glass (plate area=2 to 3 cm2 ) were used as anodes. inclusion of the chains by surrounding CD.The counter-electrode was in all cases a platinum grid. All In conclusion, these first experiments show that electropotentials were measured with respect to a saturated calomel polymerization of highly hydrophobic dipyrrolyl derivatives electrode (SCE). The electrolyte solution consisted of aceto- can be performed in aqueous media with the help of host–guest nitrile (CH3CN) (Prolabo Chromanorm HPLC grade, used chemistry. Quite surprisingly, the electrochemical properties without further purification) or water ( purified on a MilliQ of the resulting films are very similar to those of films generated Water System, Millipore Inc.) containing 0.1M lithium per- in acetonitrile.This could be explained by similar interchain chlorate (Acros) as supporting salt.organization during the film growth in water as in organic Polymer mass measurements were performed on platinum- medium, the CD having a minor influence on the structure of coated glass. Polymer samples were rinsed with water, then the film itself. DiVerent characteristics could be observed for dried at 50 °C to constant mass, which was obtained by ex materials formed starting from such monomers in which the situ determination, using a Mettler AE 163 balance with a spacer moieties are also and remain complexed by CD during precision of 10-5 g.Reproducibility was checked by three electropolymerization. Experiments with this aim are under consecutive polymerizations. investigation. Elemental analyses were performed by the CNRS Analysis Centre, Vernaison, France.Conclusion Chemicals The interaction in water of dipyrrolyl monomers with b-cyclodextrin has been studied by 1H NMR spectroscopy and b-Cyclodextrin (bCD) and hydroxypropyl-b-cyclodextrin chemical analyses. It is demonstrated that the two pyrrolyl (HPbCD) were provided by Aldrich and used as received. All rings are encapsulated, the cyclodextrin being positioned at details concerning monomer synthesis are given in ref. 8. the extremities of the monomer. The solubility of the CD The bCD–dipyrrolyl complexes were obtained by complex depends mainly on the length of the spacer. This precipitation under argon by the following general procedure: peculiar structure of the inclusion compound leads to polymers a 16.3 mM bCD and 16.3 mM monomer solution in water free of host molecule, resulting from the polymerization of the was heated to reflux with vigorous stirring for 2 h, then allowed two pyrrolyl rings, both by electrochemical and chemical to cool at room temperature.The inclusion complex separated polymerization. The polymers thus formed present electro- from the solution by precipitation, and was isolated by filchemical characteristics similar to those obtained in organic tration, washed with water then with ether, and dried under medium.Moreover, the presence of ethoxy groups in the vacuum (2×10-2 torr) for 4 h at room temperature. spacer could explain the peculiar electroactivity of the polymer bCD–1: Yield: 33%. Anal. Calcd for 152(HPbCD):5H2O in water, by favouring the formation of a reticulated, non (C98H170N2O77): C, 45.13; H, 6.52; N, 1.07; O, 47.27.Found: compact material in which the movement of the doping ions C, 45.12; H, 6.79; N, 1.27; O, 46.82%. can easily take place during the doping–dedoping process. bCD–2: Yield: 33%. Anal. Calcd for 252(HPbCD):5H2O In this work inclusion chemistry has been proved to be a (C98H170N2O75): C, 45.69; H, 6.60; N, 1.09; O, 46.62. Found: good strategy for forming polymeric materials in aqueous C, 45.29; H, 6.78; N, 1.17; O, 46.76%. media starting from hydrophobic monomers.This demon- bCD–3: Yield: 30%. Anal. Calcd for 352(HPbCD):6H2O strates that the applications of cyclodextrin can still be (C100H176N2O76): C, 45.80; H, 6.72; N, 1.07; O, 46.41. Found: extended, in particular in the field of polymeric film electro- C, 45.95; H, 6.90; N, 1.10; O, 46.04%.synthesis in non-polluting media. bCD–4: Yield: 21%. Anal. Calcd for 452(HPbCD):6H2O (C102H180N2O76): C, 46.22; H, 6.80; N, 1.06; O, 45.92. Found: C, 46.34; H, 6.82; N, 1.08; O, 45.75%. Experimental Two of the monomers presented in this study were polymerized chemically in water (magnetically stirred for two General methods hours and purged of dissolved oxygen by bubbling argon) in 1H NMR spectra were recorded on a BrukerW200 (200 MHz) the presence of Fe3+ under conditions given in Table 6.The spectrometer in D2O and DMSO-d6; chemical shifts (d) are polymers were repeatedly washed with water and then dried given relative to tetramethylsilane as internal standard. under vacuum. IR spectra were recorded on a Nicolet SX 60 Fourier transform spectrometer. Acknowledgements ESCA spectra were recorded on a Vacuum Generator Escalab MK2 Spectrometer equipped with an unmonochrom- The authors thank Dr S.Aeiyach for helpful discussions. ated Al-Ka X-ray source (power applied to the anode 200W) at pressures in the 10-8 mbar range. The analyser was operated References at constant pass energy (20 eV).The spectra were digitized, summed, smoothed and reconstructed using Gaussian-shaped 1 K. A. Connors, Chem. Rev., 1997, 97, 1325, and references therein. 2 G. Wenz, Angew. Chem., Int. Ed. Engl., 1994, 33, 803. components. Binding energies were referenced to C1s 285 eV. J. Mater. Chem., 1999, 9, 1065–1070 10693 (a) A. M. Martre, G. Mousset, P. Pouillen and R. Prime, 9 P. V. Demarco and A.L. Thakkar, J. Chem. Soc., Chem. Commun., 1970, 2. Electrochim. Acta, 1991, 36, 1911; (b) G. Farnia, G. Sandona, 10 D. Salvatierra, C. Jaime, A. Virgili and F. Sanchez-Ferrando, R. Fornasier d F. Marcuzzi, Electrochim. Acta, 1990, 35, 1149; J. Org. Chem., 1996, 61, 9578. (c) T. Matsue, C. Tasaki, M. Fujihira and T. Osa, Bull. Chem. 11 R. Isnin, C. Salam and A. E. Kaifer, J.Org. Chem., 1991, 56, 35. Soc. Jpn., 1983, 56, 1305; (d) C. Z. Smith and J. H. P. Utley, 12 (a) A. Harada and M. Kamachi, Macromolecules, 1990, 23, 2821; J. Chem. Soc., Chem. Commun., 1981, 492; (e) C. Z. Smith and (b) A. Harada, J. Li and M. Kamachi, Macromolecules, 1994, J. H. P. Utley, J. Chem. Soc., Chem. Commun., 1981, 792. 27, 4538. 4 (a) P. M. Bersier, J. Bersier and B.Klingert, Electroanalysis, 1991, 13 T. Matsue, D. H. Evans, T. Osa and N. Kobayashi, J. Am. Chem. 3, 443, and references therein; (b) A. Essalim, E. Saint-Aman and Soc., 1985, 107, 3411. D. Serve, Bull. Soc. Chim. Fr., 1994, 131, 407, and references 14 (a) J. Roncali, R. Garreau, D. Delabouglisse, F. Garnier and therein; (c) C. Retna Raj and R. Ramaraj, J. Electroanal. Chem., M.Lemaire, J. Chem. Soc., Chem. Commun., 1989, 679; 1996, 405, 141; (d) R. Castro, I. Cuadrado, B. Alonso, (b) C. Arbizzani, M. Mastragostino, L. Meneghello, T. Hamaide C. M. Casado, M. Moran and A. E. Kaifer, J. Am. Chem. Soc., and A. Guyot, Electrochim. Acta, 1992, 37, 1631; (c) C. Arbizzani, 1997, 119, 5760; (e) A. Mirzoian and A. E. Kaifer, Chem. Eur. J., A. M. Marinangeli, M.Mastragostino and L. Meneghello, 1997, 3, 1052. J. Power Sources, 1993, 43–44, 453. 5 (a) J. Yan and S. Dong, Electroanalysis, 1997, 9, 1219; (b) J. Yan 15 (a) V. Bocchi, L. Chierici and G. P. Gardini, Tetrahedron, 1967, and S. Dong, Langmuir, 1997, 13, 3251. 23, 737; (b) G. Wegner, W. Wernet, D. T. Glatzhofer, J. Ulanski, 6 J. C. Lepretre, E. Saint-Aman and J. P. Utille, J. Electroanal. C.Kro�hnke and M. Mohammadi, Synth. Met., 1987, 18, 1; Chem., 1993, 347, 465. (c) P. Nova�k, Electrochim. Acta, 1992, 37, 1227. 16 (a) A. Diaz, Chem. Scr., 1981, 17, 145; (b) A. F. Diaz, J. Castillo, 7 C. Lagrost, J. C. Lacroix, S. Aeiyach, K. I. Chane-Ching and P. C. K. K. Kanazawa and J. A. Logan, J. Electroanal. Chem., 1982, Lacaze, J. Chem. Soc., Chem. Commun., 1998, 489. 133, 233. 8 K. I. Chane-Ching, J. C. Lacroix, R. Baudry, M. Jouini, S. Aeiyach, C. Lion and P. C. Lacaze, J. Electroanal. Chem., 1998, 453, 139. Paper 9/00177H 1070 J. Mater. Chem., 1999, 9, 1065–1070 J O U R N A L O F C H E M I S T R Y Materials Electropolymerization of hydrophobic dipyrrolyls in aqueous medium based on inclusion chemistry Kathleen I. Chane-Ching,* Jean-Christophe Lacroix, Mohamed Jouini and Pierre-Camille Lacaze ITODYS, Universite� Paris 7-Denis Diderot, associe� au CNRS, 1 rue Guy de la Brosse, 75005 Paris, France.E-mails: K. I. Chane-Ching: chane@paris7.jussieu.fr, J. C. Lacroix: lacroix@paris7.jussieu.fr, M. Jouini: jouini@paris7.jussieu.fr, P. C. Lacaze: lacaze@paris7.jussieu.fr Received 6th January 1999, Accepted 25th February 1999 The host–guest complexation of hydrophobic dipyrrolyl monomers 1–4 with b-cyclodextrin has been studied in aqueous solution by means of 1H NMR spectroscopy, chemical analysis and electrochemical techniques.In each case, 152 stoichiometry is found, the cyclodextrin cavities being likely located on the extremities of the monomers. Polymeric materials were obtained from such inclusion compounds in water by either electrochemical or chemical oxidation.ESCA and elemental analyses show that the host molecule is absent from these polymers, whose characteristics are quite similar to those synthesized in organic media. A scheme for polymerization in the presence of cyclodextrin is proposed. subsequent behaviour of such complexes when they are Introduction oxidized by electrochemical or chemical means.Organizing the encounter of reagents via molecular self-assembly is an important challenge in chemistry. Within Results this context, some of the simplest assemblies start with cyclodextrins (CD) which are cyclic oligosaccharides formed by six Inclusion complexes (aCD), seven (bCD) or eight (cCD) D-glucopyranose units All the dipyrrolyl monomers studied here (see Scheme 1) are joined together by a-(1,4) bonds.Their ability to form stable poorly soluble in water (10-3, 2×10-4, <10-4 and <10-4M inclusion complexes with hydrophobic molecules in aqueous for 1, 2, 3 and 4, respectively); this can be improved by the solution has been established1 and has led to many appliaddition of 0.1M of HPbCD (1.2×10-2, 8×10-3, 4×10-3 cations.2 Most previous electrochemical studies of host–guest and 8×10-4 M, respectively).This behaviour implies the CD solutions fall into one of two classes: selective electrosynthformation of inclusion complexes between the dipyrrolyl esis based on CD complexation reactions3 and electroanalytical derivatives and the CD. 1H NMR is the most reliable technique techniques for studying the ability of cyclodextrins to hold, available to characterize this phenomenon.Studies on the two orient, conceal or separate guest molecules.4 The electrochemiinner (H-3 and H-5) hydrogen atoms of CDs in the presence cal behaviour of electroactive species in the presence of CD of a guest molecule provide valuable proof of inclusion.9 Fig. 1 has also been studied for the characterization of organized displays the characteristic upfield shifts of the H-3 and H-5 layers consisting of host–guest complexes5 or CD-grafted protons upon complexation with 1.These small but significant polypyrrole films capable of molecular recognition.6 Recently, shifts (0.024 and 0.042 ppm for H-3 and H-5, respectively) are CD complexes have been used to electropolymerize a hydrorelated to an inclusion phenomenon with a low association phobic monomer in water, hydroxypropyl-b-cyclodextrin constant.10 Such shifts are observed in the 6–7 ppm spectral (HPbCD) being the host and bithiophene the guest.7 HPbCD region corresponding to the pyrrolyl proton resonance. was chosen since it is more soluble in water than simple bCD: Conversely, the singlet peak at 3.58 ppm corresponding to the the presence of hydroxypropyl groups on the external crown two central methylene groups of the spacer does not shift of the CD increases considerably the host solubility without significantly upon complexation, indicating that the CDs inter- modification of the hydrophobicity of the cavity.Cyclodextrins act much less with the aliphatic spacer chain.The stoichiometry remain in the polymer film, suggesting that partial encapsulof the complexes was determined by isolating the very slightly ation of the polymer chains occurs. This behaviour can be water-soluble bCD inclusion compounds. In every case 152 explained by a high aYnity of the hydrophobic inner cavity stoichiometry is found by chemical analysis (see Experimental of the CD macrocycles for the growing chains of polybithiosection).These results are confirmed by 1H NMR spectroscopy phene which favours the formation of poly-pseudorotaxane in DMSO-d6 where the host molecules are unthreaded, and structures during electropolymerization. the resulting signals correspond to the free species. For We have shown in a previous paper8 that certain dipyrrolyl monomers can be used to form polymers presenting ladderlike structures, these monomers consisting of more or less rigid spacers, the extremities of which are bonded to the N atom of two pyrrolyl entities.The use of a rigid spacer leads undoubtedly to a ladder-structure whereas with long, flexible spacers intramolecular coupling may occur during polymer formation.Thus, depending on the structure of the complexes, encapsulation of such monomers by CDs could be a good way to make the spacers inflexible, or CD could be expelled into the bulk during the polymerization process. This work is devoted to the study of the interaction of CD Scheme 1 Monomers 1–4. with dipyrrolyls having long, flexible spacers and to the J. Mater. Chem., 1999, 9, 1065–1070 1065Fig. 2 Partial 200 MHz 1H NMR spectrum of 3–bCD inclusion complex in DMSO-d6. Chemical shifts in ppm; 152 stoichiometry is deduced from integration of monomer and CD peaks. Electrochemical oxidative polymerization The most water-soluble monomer complex &lsquh;CD’ was first investigated electrochemically at 20 °C in water. The voltammetric behaviour of 1 is characterized by the irreversible oxidation of the pyrrolyl group as in an organic medium.8 Whereas monomer 1 oxidizes at 1.21 V in acetonitrile, its oxidation potential becomes 1.09 V in water in the presence of a 10-fold excess of HPbCD.This potential shift could be explained by a greater stabilization of the cation-radical in aqueous medium. As can be seen in Table 1, addition of bCD causes the peak oxidation current to decrease, indicating that much of the monomer exists as the more slowly diVusing inclusion complex, since no change in the solution viscosity is expected over the range of 0–10 mM bCD.13 Simultaneously, the oxidation peak potential of 1 is shifted substantially to a more positive value, indicating that 1 is more diYcult to oxidize in the presence of bCD.Such eVects are also observed with HPbCD addition.Evans et al. observed the same behav- Fig. 1 Partial 200 MHz 1H NMR spectra of bCD in D2O: (a) no additive; after addition of 1 with [host/guest] ratios of (b) 3.3, (c) 1.1 iour in voltammetric studies of ferrocenecarboxylic acid in the and (d) 0.25 (only host protons are assigned). presence of bCD and conclusively demonstrated the prevalence of a CE (C: chemical; E: electrochemical ) mechanism13 in which the complex first dissociates followed by the oxidation example, Fig. 2 exemplifies the results obtained with the of the free ferrocene derivative. We conclude that the same 3–bCD complex: the peak at 4.83 ppm corresponds to the mechanism applies for the oxidation of the dipyrrolyl seven H1 protons of the cyclodextrin and those at 5.95 and derivative (see Scheme 2). 6.72 ppm correspond to the four HB and HA of the dipyrrolyl Regular film formation at 20 °C by electrochemical oxidation derivative, respectively. The relative integrated intensities of of a 10-2 M solution of monomer 1 in water and in the these peaks give the proton ratio (3.236/7)5(1.05/4), that is absence of HPbCD is impossible for solubility reasons: the nearly 251.The same results are obtained with the other electrode surface is covered with submicronic polymer islets monomers. It should be noted that spacer length does not (Fig. 3), resulting from the polymerization of droplets of 1 in aVect the stoichiometry of the complex and, conversely, the suspension in water. It becomes possible when HPbCD is longer the spacer, the less the resulting complex is water- added at a ratio of 10 HPbCD per monomer, using the soluble.This, added to the NMR study described above, galvanostatic method or cyclic voltammetry. Fig. 4 shows a indicates that the two pyrrolyl rings of the monomers constitute multicycle voltammogram recorded between 0 and 1.0 V at a the complexation sites, the spacer remaining free.This assump- scan rate of 50 mV s-1. The first anodic sweep reveals a wave tion is in good agreement with a study of the complexation of ferrocene derivatives containing n-alkyl chains, where it was Table 1 Current and potential peaks measured on a carbon graphite demonstrated that bCD interacts much less eVectively than electrode immersed in 0.1 M KNO3 containing 1.0 mM 1 and various aCD with aliphatic chains.11 Any attempt to obtain complexes concentrations of HPbCD.Scan rate: 50 mV s-1; reference electrode: of the monomers with aCD in good yields failed, although SCE. aCD is known to complex alkyl or PEG chains,12 possibly 103 c HPbCD/M i/mA E/V because it is diYcult for the cavity to thread into the pyrrolyl ring and reach the spacer chain. 0 61.5 1.05 In what follows, we will devote our attention to the most 1 53 1.07 soluble complex, being given the necessary concentrations to 10 42 1.09 carry out polymerization reactions under good conditions. 1066 J. Mater. Chem., 1999, 9, 1065–1070Scheme 2 Proposed mechanism of oxidative polymerization of 1 in presence of HPbCD. with an onset at 0.75 V, corresponding to the oxidation of 1; additional anodic and cathodic peaks increase regularly with the number of successive cycles.The electroactivity of the resulting poly-1 films in water is similar to that for films obtained in acetonitrile (Fig. 5). This behaviour is quite unusual for a polypyrrole (PPY) derivative. For instance, when PPY is synthesized in organic medium, the cyclability performance of the film in water is much worse than in the synthesis solvent.The presence here of ether functions, which are well known to increase the ionic conductivity of the polymer,14 in the crosslinking spacers probably induces a permeability suYcient to allow free movements of the counter ions. This could explain the great reversibility of the oxidoreduction of the poly-1 in water independently of the medium used for the electrosynthesis.Fig. 4 Potentiodynamic multisweep electrodeposition and redox Surface analyses of these films in the oxidized state have cycling (50 mV s-1) of poly-1 on Pt in aqueous 0.1 M LiClO4 and been performed by XPS in order to estimate the doping level 0.1 M HPbCD (S=1.5 cm2). Monomer concentration: 10 mM; reference electrode: SCE.Fig. 5 Electroactivity in aqueous 0.1 M LiClO4 at diVerent scan rates of poly-1 films generated in galvanostatic mode using a charge of 132 mC cm-2 from: (a) 10 mM1 aqueous+0.1 MHPbCD; (b) 10 mM 1 in acetonitrile. and to determine whether CD remained in the material. The results are given in Table 2. The comparison with theoretical stoichiometry strongly suggests that no CD remains, and that the doping level is in the order of 15%, this being slightly lower than the value of 20% found previously for such films formed in acetonitrile.8 The observed excess of oxygen atoms in the materials may be provided from an over-oxidation process during the electrosynthesis, as evidenced by IR spectroscopy (see below).The perpendicular conductivity of free standing films, ca. 1 mm thick, was measured: poly-1 films have a conductivity of 6×10-6 and 4×10-7 S cm-1 when Fig. 3 SEM photographs of poly-1 deposited on platinum from a generated in the galvanostatic mode at 0.1 mA cm-2 in aceto- 10 mM 1 solution in water; top: without HPbCD; bottom: with 0.1 M HPbCD. Supporting electrolyte: 0.1 M LiClO4. nitrile and water+HPbCD, respectively. These modest values, J.Mater. Chem., 1999, 9, 1065–1070 1067Table 2 Empirical formulae deduced from ESCA analysis of thin films of poly-1 films in oxidized state and calculated by assuming no CD incorporation and a ratio of 0.15 ClO4- per pyrrolyl ring (i.e. a doping level of 15%). Sample (1) was obtained in galvanostatic mode by cycling between -0.2 and 1.0 V at a scan rate of 50 mV s-1, and sample (2) was obtained in galvanostatic mode using a current of 0.2 mA cm-2 and a total charge of 72 mC cm-2. Monomer concentration: 10-2 M in 0.1 M HPbCD and 0.1 M LiClO4 aqueous solution Sample Empirical formula Calcd.for poly-1+30% ClO4- C14N2O3.2Cl0.3 Found for (1) C14N2.3O4.7Cl0.28 Found for (2) C14N1.9O4.3Cl0.24 when compared with classic poly-N-alkylpyrrole,16 are due to a great steric eVect of the spacer which could highly disturb the ring–ring planarity.However, these results are in accordance with the respective doping levels of the two polymers. Mass measurements were compared with the theoretical masses for 100% yield and without CD incorporation. A doping level of 15% and a consumption of four electrons per monomer are assumed (the two pyrrolyl moieties are assumed Fig. 6 A and B: IR spectra of reduced poly-1 films deposited on to be involved in the polymerization process), as previously platinum in potentiodynamic mode from 10 mM of 1 in: (A) aceto- found in acetonitrile.8 Table 3 gives the masses of some polynitrile; (B) water+0.1 M HPbCD; C: IR spectrum of HPbCD (1% 1 generated in water at diVerent current densities.Although dispersion in KBr). the polymerization yields do not always reach 100%, possibly due to some oligomer dissolution, these new results confirm that cyclodextrin is not incorporated in the films. Table 4 Frequencies (cm-1) and assignments of the principal bands observed on thin films of poly-1 deposited on a platinum electrode in Moreover, the IR spectra of such poly-1 films are very either acetonitrile or aqueous 0.1 M HPbCD similar to those generated in acetonitrile, as can be seen in Fig. 6: the two spectra show a series of peaks characteristic of Band position/cm-1 Assignments the polymer without CD (see Table 4). The dramatic increase in intensity of the IR band near 1700 cm-1 observed in the 1100 n (CH2–O–) film generated in water may be due to the introduction of 1300 n N–C(H2) 1448 d CH2 carbonyl groups as end groups which is favoured in this 1585 n CLC out-of-phase medium15 whereas the weak band at 3360 cm-1 is attributable 1705 n CLO (default) to residual water.Nevertheless, if substantial amounts of 2865 n CH2 HPbCD were incorporated in great quantity in poly-1 films 3115 n (CH ring) obtained in aqueous medium, the bands at ca. 1100 and n=stretching vibration. d=deformation vibration. 3400 cm-1 would be at least as strong as that at 2900 cm-1, as can be seen in the IR spectrum of HPbCD (see Fig. 6). This is not the case here, which confirms the absence of polymerization process and aVords finally a material similar constitutive CD from the material. to that generated in an organic medium where ladder-like structures are formed.Chemical oxidative polymerization In principle, two structures are possible for the CD Monomers 1 and 2 were chemically polymerized in water in complexes: i) an association of the monomer at the outside of the presence of HPbCD. The other monomers were not the cavity, and ii) an inclusion inside the ring. 1H NMR investigated in this way for reasons of solubility.The synthesis measurements show that the inner protons of the CD are conditions are given in the Experimental section. In both aVected upon complexation and that the partners form 152 cases, a black powder was collected and analysed. The results complexes. These results make mechanism ii) more likely. are given in Table 5. When the measured percentage of each Moreover, the materials formed either electrochemically or by element is compared with the theoretical values based on chemical polymerization do not contain constitutive CD.This various assumed compositions, only that with no CD incorpor- behaviour is consistent with a complex in which only the two ation and a doping level of 15% fits well. These new results pyrrolyl entities are encapsulated, the spacer remaining free.are in good agreement with those obtained by electropolymer- We can expect such complexes to be formed according to ization, i.e. although the starting monomer is encapsulated in cyclodextrin in aqueous medium, it is unthreaded during the Table 5 Results of elemental analyses of poly-1 and poly-2 obtained in aqueous HPbCD by chemical polymerization, using Fe3+ as Table 3 Calculated masses (mth) by assuming no CD incorporation oxidising agent. Theoretical percentages are calculated by assuming a and a doping level of 15%, and experimental masses (mexp) of deposited doping level of 15% and total absence of CD films of poly-1 at varying current densities.E is the observed deposition potential. Electropolymerization was performed in aqueous solutions X Elemental analysis (%) of 10-2 M monomer with 0.1 M HPbCD and 0.1 M LiClO4 1 Found: C: 59.8, H: 5.8, N: 9.5, O: 21.0, Cl: 3.2 j/mA cm-2 E/V Q/C cm-2 mth/mg ±0.04 mexp/mg Calcd.: C: 61.3, H: 5.8, N: 10.2, O: 18.7, Cl: 3.9 2 Found: C: 67.4, H: 7.3, N: 10.1, O: 10.5, Cl: 4.0 0.06 0.63 0.443 0.74 0.89 Calcd.: C: 69.5, H: 6.6, N: 11.6, O: 7.9, Cl: 4.4 0.21 0.67 0.625 0.59 0.54 0.545 0.8�1.03 0.227 0.16 0.14 X: starting monomer. 1068 J. Mater. Chem., 1999, 9, 1065–1070Table 6 Conditions for synthesis of poly-1 and poly-2 by chemical equilibria (1) and (2), oxidation S+L=SL (1) Synthesis solution Added dropwise T/ °C Yield (%) SL+L =SL2 (2) 0.1 M HPbCD+0.01 M 1 0.046M Fe(ClO4)3 20 25 where S and L denote Substrate (monomer) and Ligand (CD), 0.4 M HPbCD+0.04 M 2 0.200M Fe(ClO4)3 50 30 respectively.Even if the monomer is made soluble by complexation of the two pyrrolyl rings, the resulting complex is For quantitative analysis, each peak area was corrected by the subject to association/dissociation phenomena. Oxidation of appropriate experimentally determined sensitivity factor. the pyrrolyl rings involves only the free species, and this is The electrochemical studies were carried out in a immediately followed by a coupling step which leads to single-compartment three-electrode cell using an EG&G PAR insoluble oligomers.Since the polypyrrole chains are substi- 362 potentiostat in the potentiodynamic or galvanostatic mode. tuted on the N atoms of each pyrrole ring, leading to a ladder- Glassy carbon (disk area=7.1×10-2 cm2) and platinum like structure, no binding site is available for subsequent coated glass (plate area=2 to 3 cm2 ) were used as anodes.inclusion of the chains by surrounding CD. The counter-electrode was in all cases a platinum grid. All In conclusion, these first experiments show that electropotentials were measured with respect to a saturated calomel polymerization of highly hydrophobic dipyrrolyl derivatives electrode (SCE).The electrolyte solution consisted of aceto- can be performed in aqueous media with the help of host–guest nitrile (CH3CN) (Prolabo Chromanorm HPLC grade, used chemistry. Quite surprisingly, the electrochemical properties without further purification) or water ( purified on a MilliQ of the resulting films are very similar to those of films generated Water System, Millipore Inc.) containing 0.1M lithium per- in acetonitrile.This could be explained by similar interchain chlorate (Acros) as supporting salt. organization during the film growth in water as in organic Polymer mass measurements were performed on platinum- medium, the CD having a minor influence on the structure of coated glass.Polymer samples were rinsed with water, then the film itself. DiVerent characteristics could be observed for dried at 50 °C to constant mass, which was obtained by ex materials formed starting from such monomers in which the situ determination, using a Mettler AE 163 balance with a spacer moieties are also and remain complexed by CD during precision of 10-5 g. Reproducibility was checked by three electropolymerization.Experiments with this aim are under consecutive polymerizations. investigation. Elemental analyses were performed by the CNRS Analysis Centre, Vernaison, France. Conclusion Chemicals The interaction in water of dipyrrolyl monomers with b-cyclodextrin has been studied by 1H NMR spectroscopy and b-Cyclodextrin (bCD) and hydroxypropyl-b-cyclodextrin chemical analyses.It is demonstrated that the two pyrrolyl (HPbCD) were provided by Aldrich and used as received. All rings are encapsulated, the cyclodextrin being positioned at details concerning monomer synthesis are given in ref. 8. the extremities of the monomer. The solubility of the CD The bCD–dipyrrolyl complexes were obtained by complex depends mainly on the length of the spacer.This precipitation under argon by the following general procedure: peculiar structure of the inclusion compound leads to polymers a 16.3 mM bCD and 16.3 mM monomer solution in water free of host molecule, resulting from the polymerization of the was heated to reflux with vigorous stirring for 2 h, then allowed two pyrrolyl rings, both by electrochemical and chemical to cool at room temperature.The inclusion complex separated polymerization. The polymers thus formed present electro- from the solution by precipitation, and was isolated by filchemical characteristics similar to those obtained in organic tration, washed with water then with ether, and dried under medium. Moreover, the presence of ethoxy groups in the vacuum (2×10-2 torr) for 4 h at room temperature.spacer could explain the peculiar electroactivity of the polymer bCD–1: Yield: 33%. Anal. Calcd for 152(HPbCD):5H2O in water, by favouring the formation of a reticulated, non (C98H170N2O77): C, 45.13; H, 6.52; N, 1.07; O, 47.27. Found: compact material in which the movement of the doping ions C, 45.12; H, 6.79; N, 1.27; O, 46.82%. can easily take place during the doping–dedopingrocess.bCD–2: Yield: 33%. Anal. Calcd for 252(HPbCD):5H2O In this work inclusion chemistry has been proved to be a (C98H170N2O75): C, 45.69; H, 6.60; N, 1.09; O, 46.62. Found: good strategy for forming polymeric materials in aqueous C, 45.29; H, 6.78; N, 1.17; O, 46.76%. media starting from hydrophobic monomers. This demon- bCD–3: Yield: 30%. Anal.Calcd for 352(HPbCD):6H2O strates that the applications of cyclodextrin can still be (C100H176N2O76): C, 45.80; H, 6.72; N, 1.07; O, 46.41. Found: extended, in particular in the field of polymeric film electro- C, 45.95; H, 6.90; N, 1.10; O, 46.04%. synthesis in non-polluting media. bCD–4: Yield: 21%. Anal. Calcd for 452(HPbCD):6H2O (C102H180N2O76): C, 46.22; H, 6.80; N, 1.06; O, 45.92.Found: C, 46.34; H, 6.82; N, 1.08; O, 45.75%. Experimental Two of the monomers presented in this study were polymerized chemically in water (magnetically stirred for two General methods hours and purged of dissolved oxygen by bubbling argon) in 1H NMR spectra were recorded on a BrukerW200 (200 MHz) the presence of Fe3+ under conditions given in Table 6. The spectrometer in D2O and DMSO-d6; chemical shifts (d) are polymers were repeatedly washed with water and then dried given relative to tetramethylsilane as internal standard.under vacuum. IR spectra were recorded on a Nicolet SX 60 Fourier transform spectrometer. Acknowledgements ESCA spectra were recorded on a Vacuum Generator Escalab MK2 Spectrometer equipped with an unmonochrom- The authors thank Dr S.Aeiyach for helpful discussions. ated Al-Ka X-ray source (power applied to the anode 200W) at pressures in the 10-8 mbar range. The analyser was operated References at constant pass energy (20 eV). The spectra were digitized, summed, smoothed and reconstructed using Gaussian-shaped 1 K. A. Connors, Chem. Rev., 1997, 97, 1325, and references therein. 2 G. Wenz, Angew. Chem., Int.Ed. Engl., 1994, 33, 803. components. Binding energies were referenced to C1s 285 eV. J. Mater. Chem., 1999, 9, 1065–1070 10693 (a) A. M. Martre, G. Mousset, P. Pouillen and R. Prime, 9 P. V. Demarco and A. L. Thakkar, J. Chem. Soc., Chem. Commun., 1970, 2. Electrochim. Acta, 1991, 36, 1911; (b) G. Farnia, G. Sandona, 10 D. Salvatierra, C. Jaime, A. Virgili and F. Sanchez-Ferrando, R. Fornasier and F. Marcuzzi, Electrochim. Acta, 1990, 35, 1149; J. Org. Chem., 1996, 61, 9578. (c) T. Matsue, C. Tasaki, M. Fujihira and T. Osa, Bull. Chem. 11 R. Isnin, C. Salam and A. E. Kaifer, J. Org. Chem., 1991, 56, 35. Soc. Jpn., 1983, 56, 1305; (d) C. Z. Smith and J. H. P. Utley, 12 (a) A. Harada and M. Kamachi, Macromolecules, 1990, 23, 2821; J. Chem. Soc., Chem. Commun., 1981, 492; (e) C. Z. Smith and (b) A. Harada, J. Li and M. Kamachi, Macromolecules, 1994, J. H. P. Utley, J. Chem. Soc., Chem. Commun., 1981, 792. 27, 4538. 4 (a) P. M. Bersier, J. Bersier and B. Klingert, Electroanalysis, 1991, 13 T. Matsue, D. H. Evans, T. Osa and N. Kobayashi, J. Am. Chem. 3, 443, and references therein; (b) A. Essalim, E. Saint-Aman and Soc., 1985, 107, 3411. D. Serve, Bull. Soc. Chim. Fr., 1994, 131, 407, and references 14 (a) J. Roncali, R. Garreau, D. Delabouglisse, F. Garnier and therein; (c) C. Retna Raj and R. Ramaraj, J. Electroanal. Chem., M. Lemaire, J. Chem. Soc., Chem. Commun., 1989, 679; 1996, 405, 141; (d) R. Castro, I. Cuadrado, B. Alonso, (b) C. Arbizzani, M. Mastragostino, L. Meneghello, T. Hamaide C. M. Casado, M. Moran and A. E. Kaifer, J. Am. Chem. Soc., and A. Guyot, Electrochim. Acta, 1992, 37, 1631; (c) C. Arbizzani, 1997, 119, 5760; (e) A. Mirzoian and A. E. Kaifer, Chem. Eur. J., A. M. Marinangeli, M. Mastragostino and L. Meneghello, 1997, 3, 1052. J. Power Sources, 1993, 43–44, 453. 5 (a) J. Yan and S. Dong, Electroanalysis, 1997, 9, 1219; (b) J. Yan 15 (a) V. Bocchi, L. Chierici and G. P. Gardini, Tetrahedron, 1967, and S. Dong, Langmuir, 1997, 13, 3251. 23, 737; (b) G. Wegner, W. Wernet, D. T. Glatzhofer, J. Ulanski, 6 J. C. Lepretre, E. Saint-Aman and J. P. Utille, J. Electroanal. C. Kro�hnke and M. Mohammadi, Synth. Met., 1987, 18, 1; Chem., 1993, 347, 465. (c) P. Nova�k, Electrochim. Acta, 1992, 37, 1227. 16 (a) A. Diaz, Chem. Scr., 1981, 17, 145; (b) A. F. Diaz, J. Castillo, 7 C. Lagrost, J. C. Lacroix, S. Aeiyach, K. I. Chane-Ching and P. C. K. K. Kanazawa and J. A. Logan, J. Electroanal. Chem., 1982, Lacaze, J. Chem. Soc., Chem. Commun., 1998, 489. 133, 233. 8 K. I. Chane-Ching, J. C. Lacroix, R. Baudry, M. Jouini, S. Aeiyach, C. Lion and P. C. Lacaze, J. Electroanal. Chem., 1998, 453, 139. Paper 9/00177H 1070 J. Mater. Chem., 1999, 9, 1065–

ISSN:0959-9428    

DOI:10.1039/a900177h

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials An NMR study of chain transfer to diols containing both primary and secondary hydroxy groups in the polymerization of e-caprolactone Alexander Kavros, Thomas N. Huckerby and Stephen Rimmer* The Polymer Centre, School and Chemistry, Lancaster University, Bailrigg, Lancaster, Lancs, UK LA1 4YA. E-mail: s.rimmer@lancaster.ac.uk Received 24th December 1998, Accepted 16th February 1999 e-Caprolactone has been polymerized in the presence of various diols, viz.propane-1,2-diol (PD), butane-1,3-diol (BD) and hexane-1,5-diol (HD). 1H and 13C NMR spectroscopy were used to evaluate the structures. The final products were found to be a function of the diol used. It was shown that reactions incorporating PD gave low conversions and/or low degrees of polymerizations when compared with those involving BD or HD.In polymerizations involving BD two 13C resonances could be seen in the carbonyl region, assignable to the ester carbonyls adjacent to the oxymethine and oxymethylene groups derived from the residues of the diol units. Thus, both primary and secondary hydroxy groups were shown to be active in the transfer reaction. Selective decoupling was used to assign the down-field resonance to the carbonyl adjacent to oxymethylene and the up-field resonance to the carbonyl adjacent to oxymethine.However, in the case of the polymerization incorporating BD, approximately 10% of end groups were shown to be secondary alcohols, which are derived from the secondary hydroxy group of BD that does not transfer.In polymerizations involving HD only one carbonyl resonance, which could be assigned to an ester adjacent to the diol residue, was observed. From COSY spectra it was possible to assign a peak due to the ester of the secondary hydroxy. The fraction of secondary chain ends was thus observed to be greater, at ca. 13%, than in the BD polymerizations. Introduction The polymerization of e-caprolactone (CL) by ring-opening insertion polymerization, mediated by metal alkoxides, is a well established synthetic tool.The method is used widely in the preparation of both high molecular weight and oligomeric poly(e-caprolactone) (PCL). Much of the work in elucidating the mechanism and applying ring-opening insertion polymerization can attributed to Kricheldorf et al.,1–3 Jerome et al.4 and Penczek et al.5–8 These authors showed that the polymerization does in fact proceed by an insertion mechanism.Metal alkoxides that work well include those based on tin, titanium, aluminium and rare-earth metals (for examples of rare-earth catalyzed polymerizations see references 9–11). In the case of the synthesis of telechelic oligomers the chainend functionality can be added in one of two ways.In the first Fig. 1 Transfer of an alcohol (R¾OH) to a tin catalysed ring-opening insertion of Cl. method a metal alkoxide is prepared and is then used as an initiator. The molecular weight of the resultant oligomer is chain transfer agent, 3-mercaptopropane-1,2-diol. Thus the inversely proportional to the metal alkoxide concentration.end group contained both primary and secondary hydroxy Chain-end functionality, usually hydroxy, is then formed by groups. In order to show that both hydroxy groups were active hydrolysis of the metal alkoxide bond that is successively in the transfer process, we determined the concentration of transferred to the chain end as the polymerization ensues. secondary hydroxy end groups in the resulting poly(e- Thus, if one uses a metal alkoxide prepared from a diol the caprolactone-g-methyl methacrylate). Since this was found to product is a dihydroxy functional oligomer.In the second, be vanishingly low we concluded that in this case both primary commercially significant, variant of this methodology the metal and secondary hydroxy groups were transferred.To our knowl- alkoxide is employed in catalytic amounts in the presence of edge no previous work on the transfer to similar diols has the hydroxy species,12,13 the concentration of which determines been reported. We have therefore embarked on a study using the final degree of polymerization of the polymer. In this high field NMR spectroscopy, the results of which are reported method the hydroxy species acts as a transfer agent.That is, here, of the transfer to diols containing both primary the hydroxylic proton is transferred to the alkoxy chain end and secondary hydroxy groups in the polymerization of e- with scission of the metal alkoxide bond and formation of a caprolactone. new metal alkoxide bond. The transfer step is shown in Fig. 1. We have used this latter technique to prepare poly(ecaprolactone- g-methyl methacrylate) materials.14 In that work Results and discussion a poly(methyl methacrylate) with diol functionality at one Low molecular weight models for assignment of carbonyl chain end was prepared.This was then used as a transfer resonances agent in the dibutyl tin oxide (DBTO) catalyzed polymerization of CL. The diol functionality was derived from a radical The carbonyl regions of the 13C spectra of CL synthesized under conditions in which chain transfer to hydroxy group polymerization of methyl methacrylate in the presence of the J.Mater. Chem., 1999, 9, 1071–1076 1071Fig. 2 A PCL with the carbonyl groups giving rise to resolved resonances in 13C NMR spectra labelled (a, b, c).The polymerization was carried out in the presence of the alcohol, ROH. dominates chain termination, contain three resonances that are derived from three diVerent ester carbonyls.15 The dominant resonance is due to the ester carbonyl of repeat units that lie within the bulk of the chain. The other two resonances are derived from ester carbonyls of an ultimate repeat unit together with those that are formed from esterification of the transferred hydroxy group.The carbonyls giving rise to this Fig. 4 Fully decoupled (a); C–H coupled (b); selectively C–H pattern of resonances are illustrated in Fig. 2. decoupled (irradiation at d 5.15) (c) and selectively C–H decoupled The resonances of carbonyls b and c have been reported to (irradiation at d 4.125) (d) 13C NMR spectra of compound 1, in the occur at 173.9 and 174.1 ppm respectively.15 The chemical carbonyl region.shift of b is also a function of molecular weight.15 The chemical shift of a is a function of R. Thus if transfer to diVerent thus observed down-field of the carbonyl ester resonance alcohol groups in the same transfer agent is possible then resulting from transfer to the secondary hydroxy.Similar separate resonances for each carbonyl may be observed in experiments with compounds 2 and 3 gave the assignments medium field NMR spectra. In this work two diVerent alcohol given in Table 1. groups are present in the transferring diol. No data were The other models that were required were those for the available to enable the prediction of the position of the two secondary hydroxy end groups, which are formed as a result chemical shifts to be expected from ester carbonyls derived of transfer to only the primary hydroxy group.Spectral data from transfer to one or other of the two, primary or secondary, for the diols themselves were suYcient for these purposes. The hydroxy groups in the diols. Therefore the model compounds observed 1H and 13C resonances for hydroxymethine sites and shown in Fig. 3 were synthesized. Experiments to unambiguhydroxymethylene sites are given in Table 2. ously assign the carbonyl resonances were then carried out. In compound 1, COSY unambiguously assigned the complex NMR of PCLs: assignment of resonances and treatment of data multiplet at 4.05–4.21 ppm to the oxymethylene protons and the multiplet at 5.15 ppm to the oxymethine protons. The The main aim of this work was to examine the eVect of proton-noise-decoupled 13C spectrum showed carbonyl ester changing the distance between primary and secondary hydroxy resonances at 173.75 and 174.02 ppm.The peaks were of equal groups of diols involved in transfer during the polymerization intensity indicating that the NOES were the same for both of CL.The first aspect to be examined was the eVect on the carbonyl resonances. The observed spectrum for this region is final conversion. The final % conversion of CL was determined shown in Fig. 4(a). The completely C–H coupled spectrum is from both the 1H and 13C NMR spectra. The 1H NMR values shown in Fig. 4(b); as expected the spectrum is highly complex were determined by comparison of the 1H resonances due to owing to multiple C–Hcouplings.Fig. 4(c) shows the spectrum the oxymethylene protons of CL (d 4.17 (t)) and the converted with irradiation at d 5.15, which corresponds to the resonance oxymethylene 1H resonances of the PCL (d~4.05 (t) and position for the oxymethine protons. Clearly, the up-field ~3.60 (t)). An example of the signals in this region and multiplet became less complex in this spectrum, while the assignment15 of the resonances from the PCL polymer is down-field multiplet was unaVected.Irradiation at d 4.125, shown in Fig. 5. Alternatively, the carbonyl resonances in the the resonance position for the oxymethylene protons, produced 13C spectra may be used. The resonances of interest here are the trace in Fig. 4(d). In this case the down-field multiplet the carbonyl carbons from CL (176.1 ppm) and the four became simplified. It was therefore possible to assign the peak carbonyls of PCL ( labelled as C1–C4 in Tables 4 and 6). at d 173.75 to an ester carbonyl adjacent to the oxymethine While C–H decoupled 13C NMR must not be assumed to be group and that at d 174.02 to an ester carbonyl adjacent to generally quantitative, these carbonyls are adjacent to similar the oxymethylene group.The resonance from an ester carbonyl resulting from transfer to the primary hydroxy of PD was Table 1 Assignments of 13C resonances in the ester carbonyl region of the model diesters. Calculated values are in parentheses Oxymethine carbonyl Oxymethylene carbonyl Compound (d/ppm) (d/ppm) 1 174.02 173.75 2 174.34 173.88 3 174.44 174.04 Table 2 13C and 1H NMR resonances associated with the primary and secondary hydroxy groups of the diols –CHOH (d/ppm) –CH2OH (d/ppm) Diol 13C 1H 13C 1H PD 67.96 3.861 67.44 3.369/3.555 BD 66.43 3.991 60.04 3.730/3.780 HD 67.36 ~3.78 61.86 ~3.60 Fig. 3 Diesters used as models for the diol residues. 1072 J. Mater. Chem., 1999, 9, 1071–1076Fig. 5 1H NMR spectrum for the product of the TBO catalysed polymerization carried out at 80 °C in the presence of hexane-1,5-diol. An example of the resonances used to determine % conversion of CL. Fig. 7 COSY-45 spectrum for polymer formed in the presence of BD. groups containing identical numbers of hydrogens so that NOES are expected to be equal. This assumption is supported by the fact that carbonyl resonances observed for the model spectra of the free BD and HD diols showed that such diester compounds (for example see Fig. 2) were indeed of oxymethine resonances should be expected at d 3.99 and equal intensity and were derived from groups of equal 3.78 ppm respectively (see Table 2). Peaks in this region could concentration. be observed in the relevant PCL spectra.However, these A typical spectrum for the carbonyl region, showing all five resonances, postulated as arising from end groups at relatively resonances is shown in Fig. 6. The PCL resonances are labelled low concentration, cannot be unambiguously assigned in this C1–C4 on progressing from high to low frequency. The two manner. Therefore COSY measurements were also employed, resonances common to all spectra are those at 173.6 ppm (C1) so as to ensure correct identifications for these important and 173.4 ppm (C2). C1 arises from carbonyl carbons on the peaks.The COSY spectra of PCLs prepared in the presence ultimate repeat unit while the source of C2 is the in-chain of BD and HD are shown in Fig. 7 and 8. caproate ester carbonyl.15 The two ester carbonyls derived The important cross-peak connections to be made in these from transfer to each of the alcohol groups are observed at spectra involve couplings between methyl of the diol residue varying chemical shifts depending upon the diol used.By adjacent to the oxymethine protons of either a secondary reference to the model diesters we were able to assign the up- alcohol (secondary alcohol end group) or esterified secondary field resonance (C4) to the carbonyl attached to an oxygen of alcohol (i.e.the residue of a diol that has transferred). Two the oxymethine group and the down-field resonance (C3) to couplings to methyl can be observed in both spectra. These the carbonyl attached to an oxygen of the oxymethylene are labelled a and b for the connections from ester methine group.The observed resonance positions and their signal and hydroxymethine respectively. Clearly the resonances at d assignments are given in Table 3. 3.84 (Fig. 6) and d 3.73 (Fig. 7) in BD and HD respectively It was also necessary to identify the 1H resonances that are coupled to the methyl protons at d 1.19 (Fig. 7) and d arise from the hydroxymethine protons of secondary end 1.15 (Fig. 8). The identity of these methine protons is congroups. Such groups will arise if transfer does not occur to firmed by the presence of further coupling to adjacent methylthe secondary hydroxy group of the diol. Examination of the ene protons at ~d 1.71 (BD) and ~d 1.425 (HD) respectively. The ester oxymethine protons derived from diol residues that have undergone transfer show cross-peaks at d 4.89 (Fig. 6) and d 4.995 (Fig. 8). It can, therefore, be stated with confidence that the resonances at d 3.84 and d 3.73 are indeed due to secondary hydroxy groups that are derived from either BD or HD diol residues, which are located at chain ends. Fig. 6 The carbonyl region of a 13C NMR spectrum of a PCL prepared at 80 °C with TBO catalyst.Table 3 Observed resonances derived from ester carbonyls adjacent to the transferring diols C3 C4 Diol (d/ppm) (d/ppm) PD 173.0 172.8 BD 173.2 172.8 HD n.a. 173.0 Fig. 8 COSY-45 spectrum for polymer formed in the presence of HD. J. Mater. Chem., 1999, 9, 1071–1076 1073Table 4 Mn (SEC), DP (NMR) and % conversion data for polymerizations catalyzed by TBO % Secondary Reaction Mn (SEC)/ DP % Conversion % Conversion hydroxy Diol temp./ °C gmol-1 (NMR) 1H 13C end groups PD 120 477 4.7 34 56 n.a.BD 120 1690 12.1 100 100 10.0 HD 120 1660 11.0 100 100 12.7 PD 80 642 5.8 29 50 n.a. BD 80 1390 10.8 80 93 8.6 HD 80 1028 11.2 100 100 14.0 Tetrabutoxy titanium (TBO) catalysis integral when compared with that of the primary oxymethylene peak at d 3.60 indicated that only 12.7% of chain ends could Tables 4 and 5 contain the data derived from polymer systems be accounted for in this manner.Table 4 indicates the fraction formed via catalysis using TBO. The first observation to be of secondary hydroxy end-groups formed in the polymerizmade from these data is that none of the polymerizations ations incorporating BD and HD. It can be seen from this attain degrees of polymerization (DPs) that are close to the table that, in BD polymerizations, ca. 10% of the end groups theoretical value of 23. Secondly, both the DP and the final were formed from secondary residual BD units. Also, conversion of CL appear to be a function of the choice of consistently higher secondary hydroxy contents were found in diol. PD in particular is a rather poor choice of diol for these polymerizations involving HD than in those with BD.polymerizations. Both the degree of polymerization and final conversion of CL are significantly reduced at both reaction DBTO catalysis temperatures. No significant diVerences in DPs or in rates of polymerization were observed when the polymerizations were The results from DBTO catalysis are shown in Tables 6 and conducted in the presence of either BD or HD.The data imply 7. DBTO catalysis data echoed the results with TBO. Thus, that the PD–Ti complex, which forms as a result of transfer in polymerizations involving PD, at a polymerization temperafrom the propagating PCL chain, is a less eYcient initiator of ture of 120 °C, although the final conversion was high (in fact CL polymerization than the similar complexes formed from significantly higher than in the TBO catalyzed system), the transfer to either BD or HD.PD in these systems should in degree of polymerization and Mn, as measured by SEC, was fact be regarded as a degradative chain transfer agent. low. This again indicates that the PD–DBTO complex may Table 5 shows the relative peak sizes of components in the not be an eYcient reinitiator of polymerization. The high carbonyl region from the PCLs polymerized by catalysis with conversion and low degree of polymerization in this system TBO.The degradative behaviour of PD in these systems is may indicate the existence of a chain-breaking side reaction in again noteworthy. At neither of the polymerization tempera- this system.Significantly, when the reactions were carried out tures, 80 °C or 120 °C, are resonances due to esters adjacent at the lower temperature of 80 °C, but for longer reaction to the diol unit observed. This again implies that initiation, times, much improved results were obtained. Thus the degrees following transfer to PD, by the PD–Ti complex is ineYcient.of polymerization (as measured by NMR) and Mns (SEC) The data imply that the majority of the PCL chains present were much higher. The DPs were still, however, well removed are initiated by some process other than insertion into a PD from the theoretical value. Examination of the 13C spectrum alkoxide–Ti bond. The second point of note is that only one showed that both of the expected oxymethylene and oxymeresonance due to carbonyls attached to HD units is observed.thine ester carbonyls arising from esterification of both the The spectra of all the polymerizations that include BD in the secondary and primary hydroxy groups were present. reaction mixtures show two peaks in this region of equal Interestingly, the 1H spectrum of the product from the PD intensity.From reference to the spectra of the model diesters polymerization carried out at 120 °C also showed resonances it is clear that two resonances should be present. It can thus that could be attributed to free PD. The polymerizations be concluded that both the primary and secondary hydroxy involving BD and HD proceeded to high conversion and gave groups of BD are active in transfer and in reinitiation.Two polymers of degrees of polymerization between 9 and 13, explanations are possible for the absence of the second peak which are again below the expected value. With this catalyst, in the polymerizations involving HD. Firstly, the down-field as with TBO, polymerizations involving HD appear to give a resonance due to a carbonyl attached to oxymethylene, derived significantly higher fraction of secondary hydroxy.Also, the from transfer to primary hydroxy, may be hidden within the concentration of secondary hydroxy sites found in these polydominant main chain resonance (C3). Alternatively, transfer merizations appears to be lower than in equivalent TBO may occur only to primary hydroxy sites leaving the secondary polymerizations. No resonances that could be assigned to hydroxy group as a chain end.In the latter case it should be secondary hydroxy sites from the PD residue were observed. possible to observe the secondary alcohol end group. For Table 7 shows the results of 13C spectroscopy in the carbonyl example, in the product from a reaction conducted at 120 °C, region. In the case of the PD (carried out at 80 °C) and BD a peak which could be assigned to the secondary alcohol containing polymerizations the carbonyls attached to both oxymethylene protons was observed at d 3.76.However, its oxymethine and oxymethylene groups of the residual diol were observed. In each case both resonances were observed to have Table 5 C1–C4 integrations for TBO catalysis equal intensity. In the case of the HD containing polymerization only one resonance was observed.However, as discussed Reaction above for the TBO catalyzed polymerizations, this resonance Diol temp./ °C C4 C3 C2 C1 is derived from an oxymethine carbonyl and therefore the second resonance must lie underneath the main chain reson- PD 120 n.a. n.a. 0.58 0.42 BD 120 0.06 0.04 0.70 0.20 ance, C3. Thus, as with the TBO catalyzed polymerizations, HD 120 0.08 n.a. 0.78 0.14 both primary and secondary hydroxy groups are active in PD 80 n.a. n.a. 0.71 0.29 transfer and reinitiation. However, reference to the 1H spectra, BD 80 0.08 0.08 0.68 0.16 discussed above, indicates that, as in the TBO catalyzed HD 80 0.05 n.a. 0.67 0.19 system, the primary alcohol is more reactive in the transfer 1074 J. Mater.Chem., 1999, 9, 1071–1076Table 6 Mn (SEC), DP (NMR) and % conversion data for polymerizations catalyzed by DBTO % Secondary Reaction Mn (SEC)/ DP % Conversion % Conversion hydroxy Diol temp./ °C gmol-1 (NMR) 1H 13C end groups PD 120 788 6.9 97 100 n.a. BD 120 1474 12 89 95 10 HD 120 1445 11 78 90 13 PD 80 1782 14 100 100 0 BD 80 2109 11 100 100 9 HD 80 1782 14 96 100 14 Table 7 C1–C4 integrations for DBTO catalysis (GC)) and hexane-1,5-diol (HD) (Aldrich, purity=99.9+% (GC)) were used as supplied.Since all of the reactants gave Reaction single peaks on analysis by GC further purification was deemed Diol temp./ °C C4 C3 C2 C1 unnecessary. PD 120 0.08 0.08 0.61 0.23 BD 120 0.05 0.05 0.57 0.22 HD 120 0.05 n.a. 0.80 0.15 Analysis PD 80 0.05 0.05 0.80 0.10 BD 80 0.06 0.06 0.74 0.14 NMR spectra were recorded at 399.78 MHz (1H) and HD 80 0.04 n.a. 0.80 0.16 100.54 MHz (13C) using a JEOL GSX400 instrument. Magnitude-mode 2D-COSY-45 spectra of polymers were measured using a spectral width of 2718.9 Hz and 56 acquisitions reaction and this is reflected by the presence of secondary for each of 1024 increments were sampled into 1024 complex hydroxy end groups.points. Data were processed for presentation and analysis using the software suite nmrPipe.16 The arrays were zero-filled to 2048×2048 complex points and transformed in each dimen- Conclusions sion after application of a sinebell window function. Molecular The purpose of this work was to further investigate, following weights and molecular weight distributions of the PCLs were on from our initial reports on the use of PMMA diols, the measured by size exclusion chromatography (SEC) (calibrated eVect of using diols with both primary and secondary hydroxy against polystyrene standards) with Styragel@ 5 mm mixed gel groups as transfer agents for the preparation of PCL diols.As columns (Polymer Laboratories) and a UV detector system. in our previous report14 conditions were chosen that model Tetrahydrofuran was used as the eluent at a flow rate of those in operation on an industrial scale.Therefore water was 1.0 cm3 min-1. Sample concentrations were approximately removed from the system by passing a dried stream of nitrogen 2 g dm-3. through the reaction mixtures at the reaction temperature. Water, in these insertion polymerizations, acts as another type of transfer agent.The occurrence of significant transfer to Synthesis of model compounds water would result in the formation of carboxylic acid end In order to correctly assign the 13C resonances derived from groups. Since these end groups were not observed in the NMR the carbonyl carbon of the diester unit, model compounds spectra we assume that the eVect of residual water under these were prepared.Models for each diester unit were prepared by conditions is negligible. The presence of small amounts of preparing the dipropionyl ester of each diol. Thus each diol water also has no eVect on the final conversion of CL. (50 mmol) (PD, BD or HD), in turn, was mixed with propionic We have shown that under these conditions both 1,3 BD acid (100 mmol) and toluene-p-sulfonic acid (0.35 mmol).The and 1,5 HD act as transfer agents in which both hydroxy mixture was heated using Dean–Stark conditions at 175 °C for groups are able to transfer the propagating chain end. approximately 24 h. The products were then distilled under However, the secondary hydroxy has been shown to be less vacuum. The 1H NMR (400 MHz) and mass spectrometric reactive and this results in the presence of an appreciable (EI+) data for each diester were as follows.number of secondary hydroxy end groups. A higher fraction Propane-1,2-diyl dipropionate (1): d 1.14 (2×t, 6H), 1.25 of secondary end groups was found in the polymerizations (d, 3H), 2.33–2.35 (2×q, 4H), 4.07 and 4.18 (2×dd, 2H), involving HD than in those with BD.PD showed anomalous 5.15 (m, 1H); m/z 115 (1 minus –O–COCH2CH3). behaviour under some conditions. The results appear to indi- Butane-1,3-diyl dipropionate (2): d 1.14 (t, 6H), 1.26 (d, cate that PD may be involved in degradative chain transfer 3H), 1.89 (m, 2H), 2.40–2.42 (2×q, 4H), 4.12 (dd, 2H), 5.03 that may be a result of side reactions. However, at a polymeriz- (m, 1H); m/z 133 (2 minus –O–COCH2CH3).ation temperature of 80 °C, PD gave polymer of a similar Hexane-1,5-diyl dipropionate (3): d 1.13, 1.14 (2×t, 6H), quality to that from polymerizations containing BD or HD. 1.21 (d, 3H), 1.4–1.65 (3×m, 6H), 2.30–2.32 (2×q, 4H), Thus there are structural diVerences in the final PCL product 4.07 (t, 2H), 4.92 (m, 1H); m/z 157 (3 minus –O–COCH2CH3).that are a function of the distance between the two hydroxy groups in the transferred diol. These observations strongly suggest that the structure of the catalytic site undergoing transfer is cyclic. Polymerizations Polymerization mixtures were prepared from CL (6.3 g, Experimental 55 mmol), catalyst (0.1 mmol) (DBTO or TBO), xylene (12 cm3) and the required diol (4.7 mmol).These mixtures Materials were then heated under a stream of dry nitrogen at either 120 °C for 6 h or 80 °C for 7 days. This procedure would give CL (Aldrich, purity=99.9+% (GC)), dibutyl tin oxide (DBTO) (Aldrich), titanium tetrabutoxide (TBO) (Aldrich), a theoretical degree of polymerization (DP) of 23. The whole reaction mixture, including the xylene solvent, was then ana- xylene (Aldrich), propane-1,2-diol (PD) (Aldrich, purity= 99.9+% (GC)), butane-1,3-diol (BD) (Aldrich, purity=99.9% lysed by 1H and 13C NMR and SEC.J. Mater. Chem., 1999, 9, 1071–1076 10752 H. R. Kricheldorf, M. V. Sumbal and I. Kreiser-Saunders, Assignment of carbonyl resonances in models Macromolecules, 1991, 24, 1944. In order to assign the two resonances derived from the ester 3 H.R. Kricheldorf and S.-R. Lee, Macromolecules, 1996, 29, 8689. 4 D. Tian, Ph. Dubois and R. Jerome, Macromolecules, 1996, 30, carbonyls of the primary and secondary esters the following 1947. procedure was used. A fully decoupled spectrum was first run; 5 A. Duda and S. Penczek, Macromolecules, 1995, 28, 5981. this located the two carbonyl resonances from each diester. 6 S. Penczek and A. Duda, Macromol. Symp., 1996, 107, 1. Fully C–H coupled spectra were then obtained, displaying 7 J. Baran, A. Duda, A. Kowalski, R. Szymamski and S. Penczek, complex couplings for both sites, involving interactions with Macromol. Symp., 199, 123, 93. several protons. From 1H COSY spectra ester methylene 8 J. Baran, A. Duda, A. Kowalski, R. Szymamski and S. Penczek, Macromol.Rapid Commun., 1997, 18, 325. (–CH2O–) and methine (–CHO–) proton resonance positions 9 W. M. Stevels, M. J. K. Ankone, P. J. Dijkstra and J. Feijen, were identified and frequency values determined. The 13C Macromolecules, 1996, 29, 8296. coupled experiment was repeated but with selective irradiation 10 Y. Shen, Z. Shen, Y. Zhang and K. Yao, Macromolecules, 1996, at the resonance frequency for the methylene protons of the 29, 8289.primary ester. This caused the primary carbonyl response to 11 Y. Shen, Z. Shen, Y. Zhang and K. Yao, Macromolecules, 1996, be partially decoupled with simplification of the multiplicity. 29, 3441. 12 A. Duda, Macromolecules, 1994, 27, 576. This procedure was repeated with irradiation at the methine 13 Examples of patents include: UK 859 640; UK 859 642; UK frequency; in this case the secondary ester carbonyl response 859 643; UK 859 644; UK 859 645 (all 1961).showed partial collapse. 14 S. Rimmer and M. H. George, Eur. Polym. J., 1993, 29, 205. 15 R. F. Storey and A. E. Taylor, J. Macromol. Sci., 1996, A33, 77. 16 F. Delaglio, S. Grzesiek, G. Vuister, G. Zhu, J. Pfeifer and A. Bax, References J.Biomol. NMR, 1995, 6, 277. 1 H. R. Kricheldorf, M. Berl and N. Scharnagl, Macromolecules, 1988, 21, 286. Paper 8/10010A 1076 J. Mater. Chem., 1999, 9, 1071–1076 J O U R N A L O F C H E M I S T R Y Materials An NMR study of chain transfer to diols containing both primary and secondary hydroxy groups in the polymerization of e-caprolactone Alexander Kavros, Thomas N. Huckerby and Stephen Rimmer* The Polymer Centre, School and Chemistry, Lancaster University, Bailrigg, Lancaster, Lancs, UK LA1 4YA.E-mail: s.rimmer@lancaster.ac.uk Received 24th December 1998, Accepted 16th February 1999 e-Caprolactone has been polymerized in the presence of various diols, viz. propane-1,2-diol (PD), butane-1,3-diol (BD) and hexane-1,5-diol (HD). 1H and 13C NMR spectroscopy were used to evaluate the structures.The final products were found to be a function of the diol used. It was shown that reactions incorporating PD gave low conversions and/or low degrees of polymerizations when compared with those involving BD or HD. In polymerizations involving BD two 13C resonances could be seen in the carbonyl region, assignable to the ester carbonyls adjacent to the oxymethine and oxymethylene groups derived from the residues of the diol units.Thus, both primary and secondary hydroxy groups were shown to be active in the transfer reaction. Selective decoupling was used to assign the down-field resonance to the carbonyl adjacent to oxymethylene and the up-field resonance to the carbonyl adjacent to oxymethine. However, in the case of the polymerization incorporating BD, approximately 10% of end groups were shown to be secondary alcohols, which are derived from the secondary hydroxy group of BD that does not transfer.In polymerizations involving HD only one carbonyl resonance, which could be assigned to an ester adjacent to the diol residue, was observed. From COSY spectra it was possible to assign a peak due to the ester of the secondary hydroxy.The fraction of secondary chain ends was thus observed to be greater, at ca. 13%, than in the BD polymerizations. Introduction The polymerization of e-caprolactone (CL) by ring-opening insertion polymerization, mediated by metal alkoxides, is a well established synthetic tool. The method is used widely in the preparation of both high molecular weight and oligomeric poly(e-caprolactone) (PCL).Much of the work in elucidating the mechanism and applying ring-opening insertion polymerization can attributed to Kricheldorf et al.,1–3 Jerome et al.4 and Penczek et al.5–8 These authors showed that the polymerization does in fact proceed by an insertion mechanism. Metal alkoxides that work well include those based on tin, titanium, aluminium and rare-earth metals (for examples of rare-earth catalyzed polymerizations see references 9–11).In the case of the synthesis of telechelic oligomers the chainend functionality can be added in one of two ways. In the first Fig. 1 Transfer of an alcohol (R¾OH) to a tin catalysed ring-opening insertion of Cl. method a metal alkoxide is prepared and is then used as an initiator. The molecular weight of the resultant oligomer is chain transfer agent, 3-mercaptopropane-1,2-diol.Thus the inversely proportional to the metal alkoxide concentration. end group contained both primary and secondary hydroxy Chain-end functionality, usually hydroxy, is then formed by groups. In order to show that both hydroxy groups were active hydrolysis of the metal alkoxide bond that is successively in the transfer process, we determined the concentration of transferred to the chain end as the polymerization ensues.secondary hydroxy end groups in the resulting poly(e- Thus, if one uses a metal alkoxide prepared from a diol the caprolactone-g-methyl methacrylate). Since this was found to product is a dihydroxy functional oligomer. In the second, be vanishingly low we concluded that in this case both primary commercially significant, variant of this methodology the metal and secondary hydroxy groups were transferred. To our knowl- alkoxide is employed in catalytic amounts in the presence of edge no previous work on the transfer to similar diols has the hydroxy species,12,13 the concentration of which determines been reported.We have therefore embarked on a study using the final degree of polymerization of the polymer. In this high field NMR spectroscopy, the results of which are reported method the hydroxy species acts as a transfer agent. That is, here, of the transfer to diols containing both primary the hydroxylic proton is transferred to the alkoxy chain end and secondary hydroxy groups in the polymerization of e- with scission of the metal alkoxide bond and formation of a caprolactone.new metal alkoxide bond. The transfer step is shown in Fig. 1. We have used this latter technique to prepare poly(ecaprolactone- g-methyl methacrylate) materials.14 In that work Results and discussion a poly(methyl methacrylate) with diol functionality at one Low molecular weight models for assignment of carbonyl chain end was prepared.This was then used as a transfer resonances agent in the dibutyl tin oxide (DBTO) catalyzed polymerization of CL. The diol functionality was derived from a radical The carbonyl regions of the 13C spectra of CL synthesized under conditions in which chain transfer to hydroxy group polymerization of methyl methacrylate in the presence of the J.Mater. Chem., 1999, 9, 1071–1076 1071Fig. 2 A PCL with the carbonyl groups giving rise to resolved resonances in 13C NMR spectra labelled (a, b, c). The polymerization was carried out in the presence of the alcohol, ROH. dominates chain termination, contain three resonances that are derived from three diVerent ester carbonyls.15 The dominant resonance is due to the ester carbonyl of repeat units that lie within the bulk of the chain.The other two resonances are derived from ester carbonyls of an ultimate repeat unit together with those that are formed from esterification of the transferred hydroxy group. The carbonyls giving rise to this Fig. 4 Fully decoupled (a); C–H coupled (b); selectively C–H pattern of resonances are illustrated in Fig. 2. decoupled (irradiation at d 5.15) (c) and selectively C–H decoupled The resonances of carbonyls b and c have been reported to (irradiation at d 4.125) (d) 13C NMR spectra of compound 1, in the occur at 173.9 and 174.1 ppm respectively.15 The chemical carbonyl region. shift of b is also a function of molecular weight.15 The chemical shift of a is a function of R. Thus if transfer to diVerent thus observed down-field of the carbonyl ester resonance alcohol groups in the same transfer agent is possible then resulting from transfer to the secondary hydroxy.Similar separate resonances for each carbonyl may be observed in experiments with compounds 2 and 3 gave the assignments medium field NMR spectra. In this work two diVerent alcohol given in Table 1.groups are present in the transferring diol. No data were The other models that were required were those for the available to enable the prediction of the position of the two secondary hydroxy end groups, which are formed as a result chemical shifts to be expected from ester carbonyls derived of transfer to only the primary hydroxy group. Spectral data from transfer to one or other of the two, primary or secondary, for the diols themselves were suYcient for these purposes.The hydroxy groups in the diols. Therefore the model compounds observed 1H and 13C resonances for hydroxymethine sites and shown in Fig. 3 were synthesized. Experiments to unambiguhydroxymethylene sites are given in Table 2. ously assign the carbonyl resonances were then carried out.In compound 1, COSY unambiguously assigned the complex NMR of PCLs: assignment of resonances and treatment of data multiplet at 4.05–4.21 ppm to the oxymethylene protons and the multiplet at 5.15 ppm to the oxymethine protons. The The main aim of this work was to examine the eVect of proton-noise-decoupled 13C spectrum showed carbonyl ester changing the distance between primary and secondary hydroxy resonances at 173.75 and 174.02 ppm.The peaks were of equal groups of diols involved in transfer during the polymerization intensity indicating that the NOES were the same for both of CL. The first aspect to be examined was the eVect on the carbonyl resonances. The observed spectrum for this region is final conversion. The final % conversion of CL was determined shown in Fig. 4(a). The completely C–H coupled spectrum is from both the 1H and 13C NMR spectra. The 1H NMR values shown in Fig. 4(b); as expected the spectrum is highly complex were determined by comparison of the 1H resonances due to owing to multiple C–Hcouplings. Fig. 4(c) shows the spectrum the oxymethylene protons of CL (d 4.17 (t)) and the converted with irradiation at d 5.15, which corresponds to the resonance oxymethylene 1H resonances of the PCL (d~4.05 (t) and position for the oxymethine protons.Clearly, the up-field ~3.60 (t)). An example of the signals in this region and multiplet became less complex in this spectrum, while the assignment15 of the resonances from the PCL polymer is down-field multiplet was unaVected. Irradiation at d 4.125, shown in Fig. 5. Alternatively, the carbonyl resonances in the the resonance position for the oxymethylene protons, produced 13C spectra may be used. The resonances of interest here are the trace in Fig. 4(d). In this case the down-field multiplet the carbonyl carbons from CL (176.1 ppm) and the four became simplified. It was therefore possible to assign the peak carbonyls of PCL ( labelled as C1–C4 in Tables 4 and 6). at d 173.75 to an ester carbonyl adjacent to the oxymethine While C–H decoupled 13C NMR must not be assumed to be group and that at d 174.02 to an ester carbonyl adjacent to generally quantitative, these carbonyls are adjacent to similar the oxymethylene group.The resonance from an ester carbonyl resulting from transfer to the primary hydroxy of PD was Table 1 Assignments of 13C resonances in the ester carbonyl region of the model diesters.Calculated values are in parentheses Oxymethine carbonyl Oxymethylene carbonyl Compound (d/ppm) (d/ppm) 1 174.02 173.75 2 174.34 173.88 3 174.44 174.04 Table 2 13C and 1H NMR resonances associated with the primary and secondary hydroxy groups of the diols –CHOH (d/ppm) –CH2OH (d/ppm) Diol 13C 1H 13C 1H PD 67.96 3.861 67.44 3.369/3.555 BD 66.43 3.991 60.04 3.730/3.780 HD 67.36 ~3.78 61.86 ~3.60 Fig. 3 Diesters used as models for the diol residues. 1072 J. Mater. Chem., 1999, 9, 1071–1076Fig. 5 1H NMR spectrum for the product of the TBO catalysed polymerization carried out at 80 °C in the presence of hexane-1,5-diol. An example of the resonances used to determine % conversion of CL.Fig. 7 COSY-45 spectrum for polymer formed in the presence of BD. groups containing identical numbers of hydrogens so that NOES are expected to be equal. This assumption is supported by the fact that carbonyl resonances observed for the model spectra of the free BD and HD diols showed that such diester compounds (for example see Fig. 2) were indeed of oxymethine resonances should be expected at d 3.99 and equal intensity and were derived from groups of equal 3.78 ppm respectively (see Table 2).Peaks in this region could concentration. be observed in the relevant PCL spectra. However, these A typical spectrum for the carbonyl region, showing all five resonances, postulated as arising from end groups at relatively resonances is shown in Fig. 6. The PCL resonances are labelled low concentration, cannot be unambiguously assigned in this C1–C4 on progressing from high to low frequency. The two manner. Therefore COSY measurements were also employed, resonances common to all spectra are those at 173.6 ppm (C1) so as to ensure correct identifications for these important and 173.4 ppm (C2). C1 arises from carbonyl carbons on the peaks.The COSY spectra of PCLs prepared in the presence ultimate repeat unit while the source of C2 is the in-chain of BD and HD are shown in Fig. 7 and 8. caproate ester carbonyl.15 The two ester carbonyls derived The important cross-peak connections to be made in these from transfer to each of the alcohol groups are observed at spectra involve couplings between methyl of the diol residue varying chemical shifts depending upon the diol used. By adjacent to the oxymethine protons of either a secondary reference to the model diesters we were able to assign the up- alcohol (secondary alcohol end group) or esterified secondary field resonance (C4) to the carbonyl attached to an oxygen of alcohol (i.e.the residue of a diol that has transferred). Two the oxymethine group and the down-field resonance (C3) to couplings to methyl can be observed in both spectra.These the carbonyl attached to an oxygen of the oxymethylene are labelled a and b for the connections from ester methine group. The observed resonance positions and their signal and hydroxymethine respectively. Clearly the resonances at d assignments are given in Table 3. 3.84 (Fig. 6) and d 3.73 (Fig. 7) in BD and HD respectively It was also necessary to identify the 1H resonances that are coupled to the methyl protons at d 1.19 (Fig. 7) and d arise from the hydroxymethine protons of secondary end 1.15 (Fig. 8). The identity of these methine protons is congroups. Such groups will arise if transfer does not occur to firmed by the presence of further coupling to adjacent methylthe secondary hydroxy group of the diol.Examination of the ene protons at ~d 1.71 (BD) and ~d 1.425 (HD) respectively. The ester oxymethine protons derived from diol residues that have undergone transfer show cross-peaks at d 4.89 (Fig. 6) and d 4.995 (Fig. 8). It can, therefore, be stated with confidence that the resonances at d 3.84 and d 3.73 are indeed due to secondary hydroxy groups that are derived from either BD or HD diol residues, which are located at chain ends.Fig. 6 The carbonyl region of a 13C NMR spectrum of a PCL prepared at 80 °C with TBO catalyst. Table 3 Observed resonances derived from ester carbonyls adjacent to the transferring diols C3 C4 Diol (d/ppm) (d/ppm) PD 173.0 172.8 BD 173.2 172.8 HD n.a. 173.0 Fig. 8 COSY-45 spectrum for polymer formed in the presence of HD.J. Mater. Chem., 1999, 9, 1071–1076 1073Table 4 Mn (SEC), DP (NMR) and % conversion data for polymerizations catalyzed by TBO % Secondary Reaction Mn (SEC)/ DP % Conversion % Conversion hydroxy Diol temp./ °C gmol-1 (NMR) 1H 13C end groups PD 120 477 4.7 34 56 n.a. BD 120 1690 12.1 100 100 10.0 HD 120 1660 11.0 100 100 12.7 PD 80 642 5.8 29 50 n.a.BD 80 1390 10.8 80 93 8.6 HD 80 1028 11.2 100 100 14.0 Tetrabutoxy titanium (TBO) catalysis integral when compared with that of the primary oxymethylene peak at d 3.60 indicated that only 12.7% of chain ends could Tables 4 and 5 contain the data derived from polymer systems be accounted for in this manner. Table 4 indicates the fraction formed via catalysis using TBO.The first observation to be of secondary hydroxy end-groups formed in the polymerizmade from these data is that none of the polymerizations ations incorporating BD and HD. It can be seen from this attain degrees of polymerization (DPs) that are close to the table that, in BD polymerizations, ca. 10% of the end groups theoretical value of 23. Secondly, both the DP and the final were formed from secondary residual BD units.Also, conversion of CL appear to be a function of the choice of consistently higher secondary hydroxy contents were found in diol. PD in particular is a rather poor choice of diol for these polymerizations involving HD than in those with BD. polymerizations. Both the degree of polymerization and final conversion of CL are significantly reduced at both reaction DBTO catalysis temperatures.No significant diVerences in DPs or in rates of polymerization were observed when the polymerizations were The results from DBTO catalysis are shown in Tables 6 and conducted in the presence of either BD or HD. The data imply 7. DBTO catalysis data echoed the results with TBO. Thus, that the PD–Ti complex, which forms as a result of transfer in polymerizations involving PD, at a polymerization temperafrom the propagating PCL chain, is a less eYcient initiator of ture of 120 °C, although the final conversion was high (in fact CL polymerization than the similar complexes formed from significantly higher than in the TBO catalyzed system), the transfer to either BD or HD.PD in these systems should in degree of polymerization and Mn, as measured by SEC, was fact be regarded as a degradative chain transfer agent.low. This again indicates that the PD–DBTO complex may Table 5 shows the relative peak sizes of components in the not be an eYcient reinitiator of polymerization. The high carbonyl region from the PCLs polymerized by catalysis with conversion and low degree of polymerization in this system TBO.The degradative behaviour of PD in these systems is may indicate the existence of a chain-breaking side reaction in again noteworthy. At neither of the polymerization tempera- this system. Significantly, when the reactions were carried out tures, 80 °C or 120 °C, are resonances due to esters adjacent at the lower temperature of 80 °C, but for longer reaction to the diol unit observed.This again implies that initiation, times, much improved results were obtained. Thus the degrees following transfer to PD, by the PD–Ti complex is ineYcient. of polymerization (as measured by NMR) and Mns (SEC) The data imply that the majority of the PCL chains present were much higher. The DPs were still, however, well removed are initiated by some process other than insertion into a PD from the theoretical value.Examination of the 13C spectrum alkoxide–Ti bond. The second point of note is that only one showed that both of the expected oxymethylene and oxymeresonance due to carbonyls attached to HD units is observed. thine ester carbonyls arising from esterification of both the The spectra of all the polymerizations that include BD in the secondary and primary hydroxy groups were present.reaction mixtures show two peaks in this region of equal Interestingly, the 1H spectrum of the product from the PD intensity. From reference to the spectra of the model diesters polymerization carried out at 120 °C also showed resonances it is clear that two resonances should be present. It can thus that could be attributed to free PD.The polymerizations be concluded that both the primary and secondary hydroxy involving BD and HD proceeded to high conversion and gave groups of BD are active in transfer and in reinitiation. Two polymers of degrees of polymerization between 9 and 13, explanations are possible for the absence of the second peak which are again below the expected value. With this catalyst, in the polymerizations involving HD.Firstly, the down-field as with TBO, polymerizations involving HD appear to give a resonance due to a carbonyl attached to oxymethylene, derived significantly higher fraction of secondary hydroxy. Also, the from transfer to primary hydroxy, may be hidden within the concentration of secondary hydroxy sites found in these polydominant main chain resonance (C3).Alternatively, transfer merizations appears to be lower than in equivalent TBO may occur only to primary hydroxy sites leaving the secondary polymerizations. No resonances that could be assigned to hydroxy group as a chain end. In the latter case it should be secondary hydroxy sites from the PD residue were observed. possible to observe the secondary alcohol end group.For Table 7 shows the results of 13C spectroscopy in the carbonyl example, in the product from a reaction conducted at 120 °C, region. In the case of the PD (carried out at 80 °C) and BD a peak which could be assigned to the secondary alcohol containing polymerizations the carbonyls attached to both oxymethylene protons was observed at d 3.76. However, its oxymethine and oxymethylene groups of the residual diol were observed.In each case both resonances were observed to have Table 5 C1–C4 integrations for TBO catalysis equal intensity. In the case of the HD containing polymerization only one resonance was observed. However, as discussed Reaction above for the TBO catalyzed polymerizations, this resonance Diol temp./ °C C4 C3 C2 C1 is derived from an oxymethine carbonyl and therefore the second resonance must lie underneath the main chain reson- PD 120 n.a. n.a. 0.58 0.42 BD 120 0.06 0.04 0.70 0.20 ance, C3. Thus, as with the TBO catalyzed polymerizations, HD 120 0.08 n.a. 0.78 0.14 both primary and secondary hydroxy groups are active in PD 80 n.a. n.a. 0.71 0.29 transfer and reinitiation. However, reference to the 1H spectra, BD 80 0.08 0.08 0.68 0.16 discussed above, indicates that, as in the TBO catalyzed HD 80 0.05 n.a. 0.67 0.19 system, the primary alcohol is more reactive in the transfer 1074 J. Mater. Chem., 1999, 9, 1071–1076Table 6 Mn (SEC), DP (NMR) and % conversion data for polymerizations catalyzed by DBTO % Secondary Reaction Mn (SEC)/ DP % Conversion % Conversion hydroxy Diol temp./ °C gmol-1 (NMR) 1H 13C end groups PD 120 788 6.9 97 100 n.a.BD 120 1474 12 89 95 10 HD 120 1445 11 78 90 13 PD 80 1782 14 100 100 0 BD 80 2109 11 100 100 9 HD 80 1782 14 96 100 14 Table 7 C1–C4 integrations for DBTO catalysis (GC)) and hexane-1,5-diol (HD) (Aldrich, purity=99.9+% (GC)) were used as supplied. Since all of the reactants gave Reaction single peaks on analysis by GC further purification was deemed Diol temp./ °C C4 C3 C2 C1 unnecessary.PD 120 0.08 0.08 0.61 0.23 BD 120 0.05 0.05 0.57 0.22 HD 120 0.05 n.a. 0.80 0.15 Analysis PD 80 0.05 0.05 0.80 0.10 BD 80 0.06 0.06 0.74 0.14 NMR spectra were recorded at 399.78 MHz (1H) and HD 80 0.04 n.a. 0.80 0.16 100.54 MHz (13C) using a JEOL GSX400 instrument. Magnitude-mode 2D-COSY-45 spectra of polymers were measured using a spectral width of 2718.9 Hz and 56 acquisitions reaction and this is reflected by the presence of secondary for each of 1024 increments were sampled into 1024 complex hydroxy end groups. points.Data were processed for presentation and analysis using the software suite nmrPipe.16 The arrays were zero-filled to 2048×2048 complex points and transformed in each dimen- Conclusions sion after application of a sinebell window function.Molecular The purpose of this work was to further investigate, following weights and molecular weight distributions of the PCLs were on from our initial reports on the use of PMMA diols, the measured by size exclusion chromatography (SEC) (calibrated eVect of using diols with both primary and secondary hydroxy against polystyrene standards) with Styragel@ 5 mm mixed gel groups as transfer agents for the preparation of PCL diols.As columns (Polymer Laboratories) and a UV detector system. in our previous report14 conditions were chosen that model Tetrahydrofuran was used as the eluent at a flow rate of those in operation on an industrial scale. Therefore water was 1.0 cm3 min-1. Sample concentrations were approximately removed from the system by passing a dried stream of nitrogen 2 g dm-3.through the reaction mixtures at the reaction temperature. Water, in these insertion polymerizations, acts as another type of transfer agent. The occurrence of significant transfer to Synthesis of model compounds water would result in the formation of carboxylic acid end In order to correctly assign the 13C resonances derived from groups.Since these end groups were not observed in the NMR the carbonyl carbon of the diester unit, model compounds spectra we assume that the eVect of residual water under these were prepared. Models for each diester unit were prepared by conditions is negligible. The presence of small amounts of preparing the dipropionyl ester of each diol.Thus each diol water also has no eVect on the final conversion of CL. (50 mmol) (PD, BD or HD), in turn, was mixed with propionic We have shown that under these conditions both 1,3 BD acid (100 mmol) and toluene-p-sulfonic acid (0.35 mmol). The and 1,5 HD act as transfer agents in which both hydroxy mixture was heated using Dean–Stark conditions at 175 °C for groups are able to transfer the propagating chain end.approximately 24 h. The products were then distilled under However, the secondary hydroxy has been shown to be less vacuum. The 1H NMR (400 MHz) and mass spectrometric reactive and this results in the presence of an appreciable (EI+) data for each diester were as follows. number of secondary hydroxy end groups. A higher fraction Propane-1,2-diyl dipropionate (1): d 1.14 (2×t, 6H), 1.25 of secondary end groups was found in the polymerizations (d, 3H), 2.33–2.35 (2×q, 4H), 4.07 and 4.18 (2×dd, 2H), involving HD than in those with BD.PD showed anomalous 5.15 (m, 1H); m/z 115 (1 minus –O–COCH2CH3). behaviour under some conditions. The results appear to indi- Butane-1,3-diyl dipropionate (2): d 1.14 (t, 6H), 1.26 (d, cate that PD may be involved in degradative chain transfer 3H), 1.89 (m, 2H), 2.40–2.42 (2×q, 4H), 4.12 (dd, 2H), 5.03 that may be a result of side reactions.However, at a polymeriz- (m, 1H); m/z 133 (2 minus –O–COCH2CH3). ation temperature of 80 °C, PD gave polymer of a similar Hexane-1,5-diyl dipropionate (3): d 1.13, 1.14 (2×t, 6H), quality to that from polymerizations containing BD or HD. 1.21 (d, 3H), 1.4–1.65 (3×m, 6H), 2.30–2.32 (2×q, 4H), Thus there are structural diVerences in the final PCL product 4.07 (t, 2H), 4.92 (m, 1H); m/z 157 (3 minus –O–COCH2CH3). that are a function of the distance between the two hydroxy groups in the transferred diol. These observations strongly suggest that the structure of the catalytic site undergoing transfer is cyclic.Polymerizations Polymerization mixtures were prepared from CL (6.3 g, Experimental 55 mmol), catalyst (0.1 mmol) (DBTO or TBO), xylene (12 cm3) and the required diol (4.7 mmol). These mixtures Materials were then heated under a stream of dry nitrogen at either 120 °C for 6 h or 80 °C for 7 days. This procedure would give CL (Aldrich, purity=99.9+% (GC)), dibutyl tin oxide (DBTO) (Aldrich), titanium tetrabutoxide (TBO) (Aldrich), a theoretical degree of polymerization (DP) of 23. The whole reaction mixture, including the xylene solvent, was then ana- xylene (Aldrich), propane-1,2-diol (PD) (Aldrich, purity= 99.9+% (GC)), butane-1,3-diol (BD) (Aldrich, purity=99.9% lysed by 1H and 13C NMR and SEC. J. Mater. Chem., 1999, 9, 1071–1076 10752 H. R. Kricheldorf, M. V. Sumbal and I. Kreiser-Saunders, Assignment of carbonyl resonances in models Macromolecules, 1991, 24, 1944. In order to assign the two resonances derived from the ester 3 H. R. Kricheldorf and S.-R. Lee, Macromolecules, 1996, 29, 8689. 4 D. Tian, Ph. Dubois and R. Jerome, Macromolecules, 1996, 30, carbonyls of the primary and secondary esters the following 1947. procedure was used. A fully decoupled spectrum was first run; 5 A. Duda and S. Penczek, Macromolecules, 1995, 28, 5981. this located the two carbonyl resonances from each diester. 6 S. Penczek and A. Duda, Macromol. Symp., 1996, 107, 1. Fully C–H coupled spectra were then obtained, displaying 7 J. Baran, A. Duda, A. Kowalski, R. Szymamski and S. Penczek, complex couplings for both sites, involving interactions with Macromol. Symp., 199, 123, 93. several protons. From 1H COSY spectra ester methylene 8 J. Baran, A. Duda, A. Kowalski, R. Szymamski and S. Penczek, Macromol. Rapid Commun., 1997, 18, 325. (–CH2O–) and methine (–CHO–) proton resonance positions 9 W. M. Stevels, M. J. K. Ankone, P. J. Dijkstra and J. Feijen, were identified and frequency values determined. The 13C Macromolecules, 1996, 29, 8296. coupled experiment was repeated but with selective irradiation 10 Y. Shen, Z. Shen, Y. Zhang and K. Yao, Macromolecules, 1996, at the resonance frequency for the methylene protons of the 29, 8289. primary ester. This caused the primary carbonyl response to 11 Y. Shen, Z. Shen, Y. Zhang and K. Yao, Macromolecules, 1996, be partially decoupled with simplification of the multiplicity. 29, 3441. 12 A. Duda, Macromolecules, 1994, 27, 576. This procedure was repeated with irradiation at the methine 13 Examples of patents include: UK 859 640; UK 859 642; UK frequency; in this case the secondary ester carbonyl response 859 643; UK 859 644; UK 859 645 (all 1961). showed partial collapse. 14 S. Rimmer and M. H. George, Eur. Polym. J., 1993, 29, 205. 15 R. F. Storey and A. E. Taylor, J. Macromol. Sci., 1996, A33, 77. 16 F. Delaglio, S. Grzesiek, G. Vuister, G. Zhu, J. Pfeifer and A. Bax, References J. Biomol. NMR, 1995, 6, 277. 1 H. R. Kricheldorf, M. Berl and N. Scharnagl, Macromolecules, 1988, 21, 286. Paper 8/10010A 1076 J. Mater. Chem., 1999, 9, 1071–1076

ISSN:0959-9428    

DOI:10.1039/a810010a

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Properties of a new pyrazoline derivative and its application in electroluminescence Xi-Cun Gao,a Hong Cao,a Lian-Qi Zhang,b Bao-Wen Zhang,b Yi Cao b and Chun-Hui Huang*a aState Key Laboratory of Rare Earth Materials Chemistry and Applications, Peking University, Beijing 100871, China. E-mail: hch@chemms.chem.pku.edu.cn bInstitute of Photographic Chemistry, Chinese Academy of Sciences, Beijing 100101, China Received 8th January 1999, Accepted 17th February 1999 As evaporated thin film and in solution, the absorption spectra, photoluminescence, electrochemistry and electroluminescence of the newly synthesized 1,3-diphenyl-5-pyren-2-yl-4,5-dihydro-1H-pyrazole (DPP) were investigated.The absorption spectra cover the whole near-ultraviolet region.At lower concentrations, the fluorescence emission is at 415 nm; at higher concentrations, a new face to face excimer emission appears on longer wavelengths; in the thin film state, the fluorescence emission becomes a single band peaking at 470 nm. Cyclic voltammetry of DPP and the hole transport material as evaporated thin films on ITO (indium tin oxide) was compared with that in solution and was used to determine the energy levels.The electric field strength dependent electroluminescent behavior of DPP was explained according to the energy levels by a tunneling mechanism, ruling out the possible cause by an excimer or exciplex formation. At a drive voltage of 18 V, the blue electroluminescence reached 2400 cd m-2. for the synthesis of DPP is as follows: (i) synthesis of DPP Introduction precursor: 0.6 g (5 mmol) acetophenone and 1.15 g (5 mmol) The past decade has seen extensive interest in electroluminesc- pyrenecarbaldehyde in 6 ml ethanol were mixed with a solution ence (EL) from organic light-emitting devices, in view of of 0.36 g (6.4 mmol) potassium hydroxide in 3 ml water.The possible application in low-cost, full-color flat-panel dis- reaction mixture was stirred for 24 h at room temperature.plays.1–3 Usually, devices comprise charge injecting electrodes The crystalline precipitate was filtered, washed with ethanol sandwiching thin film organic hole or electron transport layers and recrystallized from ethanol to give 1 g (60%) yellow and a light-emitting layer. Under high electric field strength crystals (DPP precursor); (ii) 1 g (3 mmol) DPP precursor (~108 V m-1), holes from the anode and electrons from the and 0.7 g (6.5 mmol) of phenylhydrazine in 5 ml ethanol were cathode enter the occupied states of a hole transport layer and refluxed for 6 h.The product was filtered and washed with the unoccupied states of an electron transport layer respect- ethanol to give 1.1 g yellow solid product (crude DPP).It was ively. They will then form negatively and positively charged then purified by silica gel chromatography (eluent chloroform– polarons in the organic layers. These polarons migrate under petroleum ether (154)) and recrystallized from THF to give the influence of the applied electric field, forming a polaron the colorless DPP crystals, mp 254–256 °C. 1H NMR (CDCl3): exciton with an oppositely charged species and subsequently 6.75–8.45 (19H), 6.25–6.43 (1H), 4.17–4.28 (1H), 3.25–3.35 undergoing radiative recombination.4,5 (1H); MS: DPP+, 422; Anal., calc. (found) for C31H22N2: C Testing new materials is of great importance for under- 88.15 (87.76), H 5.21 (5.37), N 6.64 (6.64). The synthesis standing device physics and improving device performance. procedure is shown in Scheme 1.Pyrazolines are known hole transport materials.6,7 Sano et al. have reported pyrazoline dimers as the hole transport materials in organic EL devices.8 In this article, a derivative of pyrazoline, DPP (1,3-diphenyl-5-pyren-2-yl-4,5-dihydro-1H-pyrazole), was synthesized and used as the light-emitting layer of the triple layered EL device based on its strong blue-colored fluorescence and its aligned ionization potential with the hole transport layer.For a thorough understanding of its properties, its absorption, photoluminescence (PL) and electrochemistry in solution and as an evaporated thin film were compared. A revised cyclic voltammetric (CV) method, with evaporated organic thin film on ITO as the working electrode, was used for the determination of energy levels.Based on the energy levels, the electric field strength-dependent electroluminescence behavior was explained. Experimental Materials N,N¾-Bis(3-methylphenyl )-N,N¾-diphenylbenzidine (TPD) was purchased from Aldrich. Aluminium tri(quinolin-8-olate) (ALQ) was synthesised by the reaction of AlCl3 with 8- C CH3 O C H O C H C H C O N N NH NH2 N N CH3 H3C N O N O N O Al + KOH, rt C2H5OH, reflux TPD ALQ DPP hydroxyquinoline in alcohol and was recrystallized from Scheme 1 Synthetic route of DPP and molecular structure of TPD and ALQ.CHCl3 then purified by vacuum sublimation. The procedure J. Mater. Chem., 1999, 9, 1077–1080 1077Preparation of the EL device and the electrode Preparation of the EL devices: The substrate is an indium–tin oxide (ITO) coated glass with a sheet resistance of 30 V%-1.With one-third of the ITO coating stripped oV by hydrochloric acid, the substrate was cleaned by ultrasonication first in a detergent solution then in a mixture of isopropyl alcohol and water (151) followed by toluene degreasing. After drying under an infrared lamp, the substrates were immediately loaded into the vacuum chamber. The organics were then successively thermally evaporated onto the ITO from molybdenum crucibles with rates in the range of 0.1–0.3 nm s-1 below a pressure of 1×10-5 Torr.The aluminum cathode was evaporated from a tungsten wire basket at higher rates (0.8–1.2 nm s-1) in a single vacuum run. The single layer and double layer working electrodes for cyclic voltammetric measurements were prepared by thermally evaporating the tested materials onto ITO by the same procedure as that for preparing the EL device.Instrumentation The absorption measurements were obtained using a Shimadzu Fig. 2 Photoluminescence spectra of DPP, E, 1×10-6M UV-3100 spectrophotometer. The photoluminescence and elec- acetonitrile solution;A, 1×10-4M acetonitrile solution; —, thin troluminescence were measured with a Hitachi model 850 film state.All were excited at 350 nm. fluorescence spectrophotometer. The brightness of the EL device was measured with a ST-86LA spot photometer and a covers the entire 200–400 nm region. As Fig. 1 shows, in closeup lens providing a focal spot of 5 mm.The layer solution, the absorption bands are sharp lines. As evaporated thicknesses were controlled in vacuo by an IL-1000 quartz thin films on quartz substrate, the absorption bands become crystal monitor, and were also measured with a Dektak3 ‘flat’ and shift toward the longer wavelength region. This is surface profile measuring system. Cyclic voltammetric experibecause the intermolecular distances in the film state are ments were performed using a CH Instruments Voltammetric reduced and the repulsion of the occupied p orbitals raises the Analyzer model 600. For solution CV, the working, auxiliary energy level of the ground state.and reference electrodes were a 0.5 cm2 glassy carbon disk, a Pt wire and a Ag/AgCl wire, respectively; 0.1 M Bu4NClO4 in Photoluminescence N,N-dimethylformamide (DMF) was the supporting electrolyte.Film CVs were recorded by using the evaporated thin A similar eVect of intermolecular interaction is found in the films on ITO substrate as the working electrode, saturated PL spectra. As Fig. 2 shows, at a concentration of 1×10-6 M, calomel (SCE) as the reference and a Pt wire as the auxiliary the fluorescence spectrum is composed of pure DPP monomer electrode in 0.1 M KCl aqueous solution.emission. As the concentration of DPP increases, the monomer emission decreases in intensity and a new fluorescent emission appears on the longer wavelength side of the monomer emis- Results and discussion sion and increases in intensity. This new emission is due to Absorption spectra the pyrene excimer,9 a result of strong p electron repulsion as the separation distance is reduced for the two DPP molecules Fig. 1 shows the absorption spectra of DPP in solution and in at higher concentration. However, when the intermolecular the thin film state. The absorption is a result of the combination distance is further reduced as in the thin film state, the of two chromophores (the pyrene and pyrazoline groups) and fluorescence emission becomes a single band peaking at 470 nm.The ‘red shift’ of the fluorescence emission is due to the ground state energy increase by the increased intermolecular interaction as a result of closer packing of the molecules in the film state. The reason for the disappearance of the excimer emission is that in the film state, the pyrenyl substituents cannot easily align face to face to favor the maximal overlap of the p orbitals.Electrochemistry Fig. 3(a) shows the CV of DPP in DMF solution and as a single layer film on ITO substrate. In solution, DPP reveals two oxidation peaks at 0.75 V and 1.24 V. The ill-defined reduction peak indicates the reduction process is incomplete. The two oxidation processes may originate from the electron loss of the two nitrogen atoms.However, in the thin film state, only one oxidation process occurs and the oxidation peak is broader than that in solution. This comparison shows molecules in the thin film state are less electroactive than in solution. Several possible factors related to the solid state nature of the film may complicate the interpretation of this Fig. 1 Absorption spectra of DPP in ca. 1×10-4 M acetonitrile reduced electroactivity. For example, electron transfer through solution and in the thin film state which was measured by evaporating DPP on a quartz substrate. molecules in the film state is less eVective than in solution; a 1078 J. Mater. Chem., 1999, 9, 1077–1080Fig. 4 Electroluminescence spectra of the ITO/TPD/DPP/ALQ/Al device (each layer is 35 nm) under diVerent drive voltages, 8 V (A), 12 V (E) and 16 V (—).When the ALQ layer is increased to 50 nm, the emission stays at 520 nm (C) under drive voltages 8–16 V. Fig. 3 (a), CV of DPP, —, in DMF solution, …, ITO/DPP single layer film. (b), CV of —, ITO/TPD/DPP electrode, …, ITO/TPD single layer. solid/solid interface between ITO and organic materials introduces an additional barrier prohibiting the transport of electrons.Fig. 3(b) shows the CV of the double layered ITO/ TPD/DPP electrode as compared with a single layer TPD on ITO. For the ITO/TPD/DPP electrode, the broad wave at ca. 0.75 V originated in the oxidation process of DPP, identical in peak position to but smaller in peak intensity than that of the single layered electrode; the sharp negative current peak at ca. 1.0 V and the positive current peak at ca. 0.78 V correspond to the oxidation of TPD neutral molecule and the reduction of TPD+ cation. The oxidation of TPD buried by one layer of DPP (1.0 V) needs a higher potential than single layer TPD (0.95 V). Ionization potential (IP) is defined as the minimum energy necessary to bring an electron from the material into vacuum.It is the diVerence between orbital energy at infinite distances and the energy of the highest occupied orbital. Use of solution electrochemical processes to describe electron-transfer reactions in the film state is predicted upon the assumption that diVerences in ionization potentials of these molecules in solution are equal to or smaller than those same energy diVerences in the film state.For electroluminescent or other optoelectronic applications, devices are usually prepared by vacuum depositing the materials onto ITO. In cyclic voltammetric measurements, using ITO with pre-evaporated thin films on it as working electrodes, the driving force for the oxidation of TPD and DPP would be more like that in an electroluminescent device, because electron gain and loss occurs in the thin film state.From the CV of TPD and DPP as precast films on ITO, the IPs for DPP and TPD are 5.50 eV and 5.61 eV (the Fig. 5 Photoluminescence and absorption spectra of single layer film medium of the redox peaks for the single layer TPD) respect- quartz/ALQ and quartz/DPP and double layer films quartz/ DPP/ALQ.ively. These values are higher than those obtained from J. Mater. Chem., 1999, 9, 1077–1080 1079lower than that of TPD so holes would enter from TPD to the DPP layer freely after being injected from the ITO. The energy barrier (0.90 eV) for electrons being injected from the ALQ layer into the DPP layer is bigger than the energy barrier (0.48 eV) for holes being injected from the DPP layer into the ALQ layer.So at low EFS electrons would be blocked in the ALQ layer. When the EFS is suYciently high, electrons would overcome this barrier height by a tunneling eVect13–16 and move into the DPP layer, then recombine with holes there and decay to emit light at 470 nm of DPP. At a drive voltage of 18 V, the luminance obtained reached 2400 cd m-2. At a current density of 1.27 mA cm-2, the eYciency was 0.23 lm W-1.Conclusion Fig. 6 The energy level diagram of the ITO/TPD/DPP/ALQ/Al device. A new pyrazoline derivative molecule was synthesized and The IPs of TPD and DPP were determined by film state CV. The EA of ALQ was determined as evaporated film on ITO by photoemission fabricated into an EL device. Its absorption, photoluminspectroscopy with a laser beam line 4B9B, which is consistent with escence and electrochemistry both in solution and as thin film that in references 17 and 18.The EAs of TPD and DPP and the IP were studied. The absorption spectra cover the whole nearof ALQ then were deduced from the first bands of their absorption ultraviolet region. At higher concentrations, DPP exhibits a spectra in the film state.new excimer emission which derives from the face to face electron interaction of the pyrenyl group. This excimer or a solution CV. We did not observe an apparent reductive CV possible exciplex formation is proved to be not the cause for peak in the negative potential region by using thin film ALQ the new emissions in EL at medium electric strength. These on ITO as the working electrode.So, we use the value new EL emissions were explained by a tunneling eVect based determined by photoemission spectroscopy for ALQ as evapor- on the energy levels which were determined by thin film cyclic ated thin film on ITO. voltammetry. Electroluminescence References EL of the triple layer device, ITO/TPD/DPP/ALQ/Al was 1 C.W. Tang and S. A. Vanslyke, Appl.Phys. Lett., 1987, 51, 913. electric field strength (EFS) dependent. That is, a device with 2 C. Adachi, S. Tokito, J. Tsutsui and S. Saito, Jpn. J. Appl. Phys., thinner organic layers exhibited blue emission from the DPP Part 2, 1988, 27, 713. layer independent of the drive voltage; a device with thicker 3 J. H. Burroughs, D. D. C. Bradley, A. R. Brown, R. N. Marks, organic layers exhibited blue emission at high drive voltages K.Mackay, R. H. Friend, P. L. Burn and A. B. Holmes, Nature but green emission at low drive voltages. Fig. 4 shows the EL (London), 1990, 347, 539. spectra of the triple layer device under diVerent drive voltages 4 I. D. Parker, J. Appl. Phys., 1994, 75, 1659. 5 P. S. Davids, I. H. Campbell and D. L. Smith, J. Appl. Phys., (the thickness of each of the organic layers was kept constant 1997, 82, 6319.so the EFS would be proportional to the drive voltage). With 6 P. M. Borsenberger and L. B. Schein, J. Phys. Chem., 1994, 98, increasing drive voltage, the EL emission gradually shifted 235. from green to blue. At medium electric strength, the emission 7 R. Young and J. Fitzgerald, J. Phys. Chem., 1995, 99, 4230. is at any wavelength from 520 to 470 nm.An excimer forma- 8 T. Sano, T. Fujii, Y. Nishio, Y. Hamada, K. Siubata and tion as mentioned before could possibly be the cause. However, K. Kuroki, Jpn. J. Appl. Phys., Part 1, 1995, 34, 3124. 9 N. J. Turro, Modern Molecular Photochemistry, Benjamin/ the PL of thin film DPP did not show any new excimer Cummings, Menlo Park, 1978.emission except the emission at 470 nm. The possibility of an 10 D. D. Gebler, Y. Z. Wang, J. W. Blatchford, S. W. Jessen, exciplex10–12 formation at the interface between DPP and ALQ D.-K. Fu, T. M. Swager, A. G. Macdiarmid and A. J. Epstein, is also denied by the PL and absorption of the quartz/ Appl. Phys. Lett., 1997, 70, 1644. DPP/ALQ (the same as quartz/ALQ/DPP) double layer films. 11 T. Granlund, L. A. A. Pettersson, M. R. Anderson and Fig. 5 shows the PL and absorption spectra of single layer O. Inganas, J. Appl. Phys., 1997, 81, 8097. 12 K. Itano, H. Ogawa and Y. Shirota, Appl. Phys. Lett., 1998, quartz/DPP, quartz/ALQ and double layer quartz/DPP/ALQ. 72, 636. It is apparent that the wide PL emission peak in quartz/ 13 L. S. Hung, C. W. Tang and M. G.Mason, Appl. Phys. Lett., DPP/ALQ is the result of the overlap of DPP and ALQ 1997, 70, 152. emission, without any new exciplex emission. The absorption 14 F. Li, H. Tang, J. Anderegg and J. Shinar, Appl. Phys. Lett., 1997, spectrum of the double layer films is also the overlap of the 70, 1233. two independent single layers. Therefore an exciplex formation 15 G. E. Jabbour, Y. Kawabe, S.E. Shaheen, J. F. Wang, M. M. Morrell, B. Kippelen and N. Peyghambarian, Appl. Phys. at the DPP/ALQ interface would also be impossible. Lett., 1997, 71, 1762. The reason for this kind of electric field strength dependent 16 H. Cao, X. C. Gao, Jin Zhai, C. H. Huang, B. W. Zhang and EL emission would be clear if the energy levels of each layer Y. Cao, Synth. Met., 1998, 96, 191.were compared. Fig. 6 shows the energy level diagram obtained 17 E. Aminaka, T. Tsutsui and S. Saito, J. Appl. Phys., 1996, 79, by cyclic voltammetry (TPD and DPP) and photoemission 8808. spectra. In this diagram, the ionization potential of DPP is 18 H. Suzuki and S. Hoshino, J. Appl. Phys., 1996, 79, 8816. Paper 9/00276F 1080 J. Mater. Chem., 1999, 9, 1077–1080 J O U R N A L O F C H E M I S T R Y Materials Properties of a new pyrazoline derivative and its application in electroluminescence Xi-Cun Gao,a Hong Cao,a Lian-Qi Zhang,b Bao-Wen Zhang,b Yi Cao b and Chun-Hui Huang*a aState Key Laboratory of Rare Earth Materials Chemistry and Applications, Peking University, Beijing 100871, China.E-mail: hch@chemms.chem.pku.edu.cn bInstitute of Photographic Chemistry, Chinese Academy of Sciences, Beijing 100101, China Received 8th January 1999, Accepted 17th February 1999 As evaporated thin film and in solution, the absorption spectra, photoluminescence, electrochemistry and electroluminescence of the newly synthesized 1,3-diphenyl-5-pyren-2-yl-4,5-dihydro-1H-pyrazole (DPP) were investigated.The absorption spectra cover the whole near-ultraviolet region.At lower concentrations, the fluorescence emission is at 415 nm; at higher concentrations, a new face to face excimer emission appears on longer wavelengths; in the thin film state, the fluorescence emission becomes a single band peaking at 470 nm. Cyclic voltammetry of DPP and the hole transport material as evaporated thin films on ITO (indium tin oxide) was compared with that in solution and was used to determine the energy levels.The electric field strength dependent electroluminescent behavior of DPP was explained according to the energy levels by a tunneling mechanism, ruling out the possible cause by an excimer or exciplex formation. At a drive voltage of 18 V, the blue electroluminescence reached 2400 cd m-2. for the synthesis of DPP is as follows: (i) synthesis of DPP Introduction precursor: 0.6 g (5 mmol) acetophenone and 1.15 g (5 mmol) The past decade has seen extensive interest in electroluminesc- pyrenecarbaldehyde in 6 ml ethanol were mixed with a solution ence (EL) from organic light-emitting devices, in view of of 0.36 g (6.4 mmol) potassium hydroxide in 3 ml water.The possible application in low-cost, full-color flat-panel dis- reaction mixture was stirred for 24 h at room temperature.plays.1–3 Usually, devices comprise charge injecting electrodes The crystalline precipitate was filtered, washed with ethanol sandwiching thin film organic hole or electron transport layers and recrystallized from ethanol to give 1 g (60%) yellow and a light-emitting layer. Under high electric field strength crystals (DPP precursor); (ii) 1 g (3 mmol) DPP precursor (~108 V m-1), holes from the anode and electrons from the and 0.7 g (6.5 mmol) of phenylhydrazine in 5 ml ethanol were cathode enter the occupied states of a hole transport layer and refluxed for 6 h.The product was filtered and washed with the unoccupied states of an electron transport layer respect- ethanol to give 1.1 g yellow solid product (crude DPP).It was ively. They will then form negatively and positively charged then purified by silica gel chromatography (eluent chloroform– polarons in the organic layers. These polarons migrate under petroleum ether (154)) and recrystallized from THF to give the influence of the applied electric field, forming a polaron the colorless DPP crystals, mp 254–256 °C. 1H NMR (CDCl3): exciton with an oppositely charged species and subsequently 6.75–8.45 (19H), 6.25–6.43 (1H), 4.17–4.28 (1H), 3.25–3.35 undergoing radiative recombination.4,5 (1H); MS: DPP+, 422; Anal., calc. (found) for C31H22N2: C Testing new materials is of great importance for under- 88.15 (87.76), H 5.21 (5.37), N 6.64 (6.64). The synthesis standing device physics and improving device performance.procedure is shown in Scheme 1. Pyrazolines are known hole transport materials.6,7 Sano et al. have reported pyrazoline dimers as the hole transport materials in organic EL devices.8 In this article, a derivative of pyrazoline, DPP (1,3-diphenyl-5-pyren-2-yl-4,5-dihydro-1H-pyrazole), was synthesized and used as the light-emitting layer of the triple layered EL device based on its strong blue-colored fluorescence and its aligned ionization potential with the hole transport layer.For a thorough understanding of its properties, its absorption, photoluminescence (PL) and electrochemistry in solution and as an evaporated thin film were compared. A revised cyclic voltammetric (CV) method, with evaporated organic thin film on ITO as the working electrode, was used for the determination of energy levels.Based on the energy levels, the electric field strength-dependent electroluminescence behavior was explained. Experimental Materials N,N¾-Bis(3-methylphenyl )-N,N¾-diphenylbenzidine (TPD) was purchased from Aldrich. Aluminium tri(quinolin-8-olate) (ALQ) was synthesised by the reaction of AlCl3 with 8- C CH3 O C H O C H C H C O N N NH NH2 N N CH3 H3C N O N O N O Al + KOH, rt C2H5OH, reflux TPD ALQ DPP hydroxyquinoline in alcohol and was recrystallized from Scheme 1 Synthetic route of DPP and molecular structure of TPD and ALQ.CHCl3 then purified by vacuum sublimation. The procedure J. Mater. Chem., 1999, 9, 1077–1080 1077Preparation of the EL device and the electrode Preparation of the EL devices: The substrate is an indium–tin oxide (ITO) coated glass with a sheet resistance of 30 V%-1.With one-third of the ITO coating stripped oV by hydrochloric acid, the substrate was cleaned by ultrasonication first in a detergent solution then in a mixture of isopropyl alcohol and water (151) followed by toluene degreasing. After drying under an infrared lamp, the substrates were immediately loaded into the vacuum chamber.The organics were then successively thermally evaporated onto the ITO from molybdenum crucibles with rates in the range of 0.1–0.3 nm s-1 below a pressure of 1×10-5 Torr. The aluminum cathode was evaporated from a tungsten wire basket at higher rates (0.8–1.2 nm s-1) in a single vacuum run. The single layer and double layer working electrodes for cyclic voltammetric measurements were prepared by thermally evaporating the tested materials onto ITO by the same procedure as that for preparing the EL device.Instrumentation The absorption measurements were obtained using a Shimadzu Fig. 2 Photoluminescence spectra of DPP, E, 1×10-6M UV-3100 spectrophotometer. The photoluminescence and elec- acetonitrile solution;A, 1×10-4M acetonitrile solution; —, thin troluminescence were measured with a Hitachi model 850 film state.All were excited at 350 nm. fluorescence spectrophotometer. The brightness of the EL device was measured with a ST-86LA spot photometer and a covers the entire 200–400 nm region. As Fig. 1 shows, in closeup lens providing a focal spot of 5 mm.The layer solution, the absorption bands are sharp lines. As evaporated thicknesses were controlled in vacuo by an IL-1000 quartz thin films on quartz substrate, the absorption bands become crystal monitor, and were also measured with a Dektak3 ‘flat’ and shift toward the longer wavelength region. This is surface profile measuring system. Cyclic voltammetric experibecause the intermolecular distances in the film state are ments were performed using a CH Instruments Voltammetric reduced and the repulsion of the occupied p orbitals raises the Analyzer model 600.For solution CV, the working, auxiliary energy level of the ground state. and reference electrodes were a 0.5 cm2 glassy carbon disk, a Pt wire and a Ag/AgCl wire, respectively; 0.1 M Bu4NClO4 in Photoluminescence N,N-dimethylformamide (DMF) was the supporting electrolyte.Film CVs were recorded by using the evaporated thin A similar eVect of intermolecular interaction is found in the films on ITO substrate as the working electrode, saturated PL spectra. As Fig. 2 shows, at a concentration of 1×10-6 M, calomel (SCE) as the reference and a Pt wire as the auxiliary the fluorescence spectrum is composed of pure DPP monomer electrode in 0.1 M KCl aqueous solution.emission. As the concentration of DPP increases, the monomer emission decreases in intensity and a new fluorescent emission appears on the longer wavelength side of the monomer emis- Results and discussion sion and increases in intensity. This new emission is due to Absorption spectra the pyrene excimer,9 a result of strong p electron repulsion as the separation distance is reduced for the two DPP molecules Fig. 1 shows the absorption spectra of DPP in solution and in at higher concentration. However, when the intermolecular the thin film state. The absorption is a result of the combination distance is further reduced as in the thin film state, the of two chromophores (the pyrene and pyrazoline groups) and fluorescence emission becomes a single band peaking at 470 nm.The ‘red shift’ of the fluorescence emission is due to the ground state energy increase by the increased intermolecular interaction as a result of closer packing of the molecules in the film state. The reason for the disappearance of the excimer emission is that in the film state, the pyrenyl substituents cannot easily align face to face to favor the maximal overlap of the p orbitals.Electrochemistry Fig. 3(a) shows the CV of DPP in DMF solution and as a single layer film on ITO substrate. In solution, DPP reveals two oxidation peaks at 0.75 V and 1.24 V. The ill-defined reduction peak indicates the reduction process is incomplete. The two oxidation processes may originate from the electron loss of the two nitrogen atoms.However, in the thin film state, only one oxidation process occurs and the oxidation peak is broader than that in solution. This comparison shows molecules in the thin film state are less electroactive than in solution. Several possible factors related to the solid state nature of the film may complicate the interpretation of this Fig. 1 Absorption spectra of DPP in ca. 1×10-4 M acetonitrile reduced electroactivity. For example, electron transfer through solution and in the thin film state which was measured by evaporating DPP on a quartz substrate. molecules in the film state is less eVective than in solution; a 1078 J. Mater. Chem., 1999, 9, 1077–1080Fig. 4 Electroluminescence spectra of the ITO/TPD/DPP/ALQ/Al device (each layer is 35 nm) under diVerent drive voltages, 8 V (A), 12 V (E) and 16 V (—).When the ALQ layer is increased to 50 nm, the emission stays at 520 nm (C) under drive voltages 8–16 V. Fig. 3 (a), CV of DPP, —, in DMF solution, …, ITO/DPP single layer film. (b), CV of —, ITO/TPD/DPP electrode, …, ITO/TPD single layer. solid/solid interface between ITO and organic materials introduces an additional barrier prohibiting the transport of electrons.Fig. 3(b) shows the CV of the double layered ITO/ TPD/DPP electrode as compared with a single layer TPD on ITO. For the ITO/TPD/DPP electrode, the broad wave at ca. 0.75 V originated in the oxidation process of DPP, identical in peak position to but smaller in peak intensity than that of the single layered electrode; the sharp negative current peak at ca. 1.0 V and the positive current peak at ca. 0.78 V correspond to the oxidation of TPD neutral molecule and the reduction of TPD+ cation. The oxidation of TPD buried by one layer of DPP (1.0 V) needs a higher potential than single layer TPD (0.95 V). Ionization potential (IP) is defined as the minimum energy necessary to bring an electron from the material into vacuum.It is the diVerence between orbital energy at infinite distances and the energy of the highest occupied orbital. Use of solution electrochemical processes to describe electron-transfer reactions in the film state is predicted upon the assumption that diVerences in ionization potentials of these molecules in solution are equal to or smaller than those same energy diVerences in the film state.For electroluminescent or other optoelectronic applications, devices are usually prepared by vacuum depositing the materials onto ITO. In cyclic voltammetric measurements, using ITO with pre-evaporated thin films on it as working electrodes, the driving force for the oxidation of TPD and DPP would be more like that in an electroluminescent device, because electron gain and loss occurs in the thin film state.From the CV of TPD and DPP as precast films on ITO, the IPs for DPP and TPD are 5.50 eV and 5.61 eV (the Fig. 5 Photoluminescence and absorption spectra of single layer film medium of the redox peaks for the single layer TPD) respect- quartz/ALQ and quartz/DPP and double layer films quartz/ DPP/ALQ. ively.These values are higher than those obtained from J. Mater. Chem., 1999, 9, 1077–1080 1079lower than that of TPD so holes would enter from TPD to the DPP layer freely after being injected from the ITO. The energy barrier (0.90 eV) for electrons being injected from the ALQ layer into the DPP layer is bigger than the energy barrier (0.48 eV) for holes being injected from the DPP layer into the ALQ layer.So at low EFS electrons would be blocked in the ALQ layer. When the EFS is suYciently high, electrons would overcome this barrier height by a tunneling eVect13–16 and move into the DPP layer, then recombine with holes there and decay to emit light at 470 nm of DPP. At a drive voltage of 18 V, the luminance obtained reached 2400 cd m-2.At a current density of 1.27 mA cm-2, the eYciency was 0.23 lm W-1. Conclusion Fig. 6 The energy level diagram of the ITO/TPD/DPP/ALQ/Al device. A new pyrazoline derivative molecule was synthesized and The IPs of TPD and DPP were determined by film state CV. The EA of ALQ was determined as evaporated film on ITO by photoemission fabricated into an EL device. Its absorption, photoluminspectroscopy with a laser beam line 4B9B, which is consistent with escence and electrochemistry both in solution and as thin film that in references 17 and 18.The EAs of TPD and DPP and the IP were studied. The absorption spectra cover the whole nearof ALQ then were deduced from the first bands of their absorption ultraviolet region. At higher concentrations, DPP exhibits a spectra in the film state.new excimer emission which derives from the face to face electron interaction of the pyrenyl group. This excimer or a solution CV. We did not observe an apparent reductive CV possible exciplex formation is proved to be not the cause for peak in the negative potential region by using thin film ALQ the new emissions in EL at medium electric strength. These on ITO as the working electrode.So, we use the value new EL emissions were explained by a tunneling eVect based determined by photoemission spectroscopy for ALQ as evapor- on the energy levels which were determined by thin film cyclic ated thin film on ITO. voltammetry. Electroluminescence References EL of the triple layer device, ITO/TPD/DPP/ALQ/Al was 1 C.W. Tang and S. A. Vanslyke, Appl.Phys. Lett., 1987, 51, 913. electric field strength (EFS) dependent. That is, a device with 2 C. Adachi, S. Tokito, J. Tsutsui and S. Saito, Jpn. J. Appl. Phys., thinner organic layers exhibited blue emission from the DPP Part 2, 1988, 27, 713. layer independent of the drive voltage; a device with thicker 3 J. H. Burroughs, D. D. C. Bradley, A. R. Brown, R. N. Marks, organic layers exhibited blue emission at high drive voltages K.Mackay, R. H. Friend, P. L. Burn and A. B. Holmes, Nature but green emission at low drive voltages. Fig. 4 shows the EL (London), 1990, 347, 539. spectra of the triple layer device under diVerent drive voltages 4 I. D. Parker, J. Appl. Phys., 1994, 75, 1659. 5 P. S. Davids, I. H. Campbell and D. L. Smith, J. Appl. Phys., (the thickness of each of the organic layers was kept constant 1997, 82, 6319.so the EFS would be proportional to the drive voltage). With 6 P. M. Borsenberger and L. B. Schein, J. Phys. Chem., 1994, 98, increasing drive voltage, the EL emission gradually shifted 235. from green to blue. At medium electric strength, the emission 7 R. Young and J. Fitzgerald, J. Phys. Chem., 1995, 99, 4230.is at any wavelength from 520 to 470 nm. An excimer forma- 8 T. Sano, T. Fujii, Y. Nishio, Y. Hamada, K. Siubata and tion as mentioned before could possibly be the cause. However, K. Kuroki, Jpn. J. Appl. Phys., Part 1, 1995, 34, 3124. 9 N. J. Turro, Modern Molecular Photochemistry, Benjamin/ the PL of thin film DPP did not show any new excimer Cummings, Menlo Park, 1978. emission except the emission at 470 nm. The possibility of an 10 D. D. Gebler, Y. Z. Wang, J. W. Blatchford, S. W. Jessen, exciplex10–12 formation at the interface between DPP and ALQ D.-K. Fu, T. M. Swager, A. G. Macdiarmid and A. J. Epstein, is also denied by the PL and absorption of the quartz/ Appl. Phys. Lett., 1997, 70, 1644. DPP/ALQ (the same as quartz/ALQ/DPP) double layer films. 11 T. Granlund, L. A. A. Pettersson, M. R. Anderson and Fig. 5 shows the PL and absorption spectra of single layer O. Inganas, J. Appl. Phys., 1997, 81, 8097. 12 K. Itano, H. Ogawa and Y. Shirota, Appl. Phys. Lett., 1998, quartz/DPP, quartz/ALQ and double layer quartz/DPP/ALQ. 72, 636. It is apparent that the wide PL emission peak in quartz/ 13 L. S. Hung, C. W. Tang and M. G. Mason, Appl. Phys. Lett., DPP/ALQ is the result of the overlap of DPP and ALQ 1997, 70, 152. emission, without any new exciplex emission. The absorption 14 F. Li, H. Tang, J. Anderegg and J. Shinar, Appl. Phys. Lett., 1997, spectrum of the double layer films is also the overlap of the 70, 1233. two independent single layers. Therefore an exciplex formation 15 G. E. Jabbour, Y. Kawabe, S. E. Shaheen, J. F. Wang, M. M. Morrell, B. Kippelen and N. Peyghambarian, Appl. Phys. at the DPP/ALQ interface would also be impossible. Lett., 1997, 71, 1762. The reason for this kind of electric field strength dependent 16 H. Cao, X. C. Gao, Jin Zhai, C. H. Huang, B. W. Zhang and EL emission would be clear if the energy levels of each layer Y. Cao, Synth. Met., 1998, 96, 191. were compared. Fig. 6 shows the energy level diagram obtained 17 E. Aminaka, T. Tsutsui and S. Saito, J. Appl. Phys., 1996, 79, by cyclic voltammetry (TPD and DPP) and photoemission 8808. spectra. In this diagram, the ionization potential of DPP is 18 H. Suzuki and S. Hoshino, J. Appl. Phys., 1996, 79, 8816. Paper 9/00276F 1080 J. Mater. Chem., 1999, 9, 1077–1080

ISSN:0959-9428    

DOI:10.1039/a900276f

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Ultrafast kinetics of 9-decylanthracene photodimers and their application to 3D optical storage Alexander S. Dvornikov,a Henri Bouas-Laurent,b Jean-Pierre Desvergneb and Peter M. Rentzepis*a aDepartment of Chemistry, University of California, Irvine, CA 92697 and Call/Recall Inc., Irvine, CA 92714, USA bPhotochimie Organique, CNRS UMR 5802, Universite� Bordeaux 1, F-33405, Talence Cedex, France Received 26th October 1998, Accepted 24th February 1999 The intermediate states and final product of the photoreaction of 9-decylanthracene monomer and photodimer have been identified and their formation and relaxation pathways and rates determined.In addition, these molecules have been used as media for high density 3D optical storage devices.Introduction The extremely large amounts of data that need to be stored and accessed at very fast rates are being limited by the memory capabilities as much as by interconnects and processors. Because of the huge data storage requirements, which continuously increase, and the rather mandatory need for parallel accessing, the need for compact, very high capacity parallel accessing, low cost memory devices is becoming very acute.One means for satisfying most of these requirements is provided by three-dimensional storage. The main research methods, which are pursued now, that may lead to practical 3D-storage memory devices, are phase holograms1,2 and two photon optical 3D memories. These devices utilize inorganic photorefractive crystals and photopolymers in the case of holography1,2 and organic3–7 or biological molecules such as bacteriorhodopsin8–11 dispersed in polymer matrices.In this paper we will be concerned with the photoreaction kinetics of 9-decylanthracene photodimers and the application of this data to 3D memory devices by means of two photon absorption. In previous papers we have presented the basic theory of two photon excitation and the means for utilizing such nonlinear processes for storing information within the volume of a 3D memory device.3–5 Briefly, the two photon absorption process shown in Fig. 1(a) causes a ground-state molecule, in the unwritten form, to be promoted to the first excited electronic state by the simultaneous absorption of two photons. The energy required to reach the excited state is greater than the energy of either photon alone, therefore each Fig. 1 a) Energy level diagram; b) writing of information in 3D by two-photon absorption. beam propagates through the memory volume without being absorbed. When the sum of the energies of these photons is equal to or greater than the energy gap between the ground shown schematically in Fig. 1(b).For fast parallel information and first electronic excited states of the molecule (see Fig. 1(a)) transfer of an entire 2D plane, located within the volume of at the point of the intersection of the beams, within the the 3D device, a single photon process is used to illuminate volume, the two photons may be absorbed simultaneously. this plane, see Fig. 1. To achieve a high density storage and The excited molecule decays and is transformed into a diVerent be suitable for use in 3D storage devices the photochromic ground-state molecular structure, which becomes the written materials used should fulfil several requirements, including: form of the 3D molecular memory device.This new molecular high nonlinear absorption cross-section; stability in both write structure absorbs at longer wavelengths than the original and read states; high solubility in the polymer matrix; the ground-state molecules, and therefore it may be excited by written form should have a high fluorescence quantum one or two photons of lower energy than those absorbed by eYciency and fluorescence in an easily detected region.the original, unwritten form molecules. The written excited We have shown previously5 that photo-induced reversible state molecules fluoresce and detection of the induced fluores- photodimerization of anthracene and 9-methylanthracene fulfil cence is used for accessing the stored information.It is, quite well most of these requirements except that they do not therefore, possible to access any bit of information within the dissolve suYciently well in PMMA [poly(methyl methacrylate] volume by simply intersecting two beams at the place where and other polymer matrices.This diYculty has been eliminated because we have been able to synthesize 9-n-decylanthracene the information bit to be retrieved is located. This process is J. Mater. Chem., 1999, 9, 1081–1084 1081which, as we anticipated, exhibits photochromism and in addition is quite soluble in polymer matrices, thus being a promising material for optical switching and 3D memory devices.Here we present the ultrafast kinetics of this new photochromic material, which is based on the reversible photodimerization of 9-n-decylanthracene dispersed in PMMA. In addition to the basic science information regarding the mechanism of the photo-transformation of this material, we found that the photo-induced dimers of this substituted anthracene fulfil most of the requirements for use as a 3D computer memory material.Experimental Fig. 2 Absorption spectra of a) photodimer, b) long wavelength A double beam Shimadzu UV160U spectrophotometer and a contribution of monomer and c) fluorescence spectrum of 9-decyl- Shimadzu RF 5000U spectrofluorophotometer recorded the anthracene (monomer) in 1,2-dichloroethane solution, at room ground state absorption and fluorescence spectra respectively.temperature, conc. #10-4 M. A ‘Quantel’ Nd/YAG laser generated 30 ps, 1064 nm pulses, which were converted to 532 nm pulses and used for twosolubility in MMA makes possible the fabrication of solid photon excitation.The ultrafast kinetics, transient absorption PMMA blocks with homogeneously dispersed dimer molecules spectra and mechanism, presented here, were studied by the at the concentrations which are required to store high density ultrafast experimental system described previously.5 of information in 3D volume. The materials used for the synthesis of the photodimer and The absorption spectra of the monomer and photodimer all solvents were Aldrich spectroscopic purity grade. 9- are shown in Fig. 2. The monomer has its long wavelength Decylanthracene (white crystals with a blue fluorescence, absorption band in the 300–400 nm region, while the photod- mp 44 °C) was synthesized using the method described in imer is blue shifted and has practically no absorption at ref. 12 and purified by column chromatography (silica gel; wavelengths longer than 300 nm.The photodissociation of this eluent: pentane). 1H NMR d (CCl4): 0.85–0.95 (t, 3H, Aproduct results in the regeneration of the monomer, which is (CH2)9-CH3), 1.25–1.5 (m, 2H), 1.5–1.65 (m, 2H), 1.7–1.85 a conjugated double bond system and therefore exhibits a red (m, 14H), 3.5–3.6 (m, 2H, benzylic: A-CH2-(CH2)8-CH3), shifted absorption band.The monomer of 9-decylanthracene 7.35–7.45 (m, 4HAr), 7.85–7.95 (m, 2HAr), 8.15–8.25 (m, was found to fluoresce in the 380 to 450 nm region, see Fig. 2. 3HAr). 13C NMR d (CDCl3): 14.3(CH3), 22.9, 28.3, 29.6, 29.8, The fluorescence quantum yields of the monomer in ethyl 29.9, 30.6, 31.6, 32.1 (-(CH2)9-), 124.6, 124.9, 125.4, 125.6, acetate and 1,2-dichloroethane solutions were measured to be 129.4 (CAr-H), 129.7, 131.8, 135.6 (quaternary C).HRMS: wF=0.35 and wF=0.70 respectively; the standard used for the (AutoSpec EQ FAB+) C24H30 calc: 318.234751, found: calibration of the quantum yield was a solution of dimethyl 318.234305. POPOP in cyclohexane, which is known to emit with a The anthracene photodimer was generated by irradiation quantum eYciency of 0.93.16 The photodimer does not with l>320 nm light emitted by a 150W arc Xenon lamp, for fluoresce at room temperature. 48 h, of a deaerated saturated hexane solution of 9-decyl- The formation and photodissociation mechanism of anthracene (#25 mg cm-3). The white solid precipitate was anthracene photodimers and related compounds has been collected, washed with fresh lvent and dried.It was shown extensively investigated previously.14,17–26 Anthracene and its by NMR that the head-to-tail photodimer was the sole derivatives may form the corresponding photodimers after photoproduct formed.13,14 excitation of the monomer to its first singlet excited state14 Thin polymer films of 9-decylanthracene photodimers and the subsequent interaction of two monomers.When the dispersed in PMMA were prepared by pouring a solution of photodimers are excited to the first allowed electronic state it the photodimers and PMMA in 1,2-dichloroethane on the may be shown that some of them may dissociate adiabatically surface of a microscope slide. The slide was then spin-coated, via intermediate excimer formation.5,17–19,22 It has also been resulting in a 20 mm thickness uniformly distributed polymer reported that at low temperature (77 K) the first excited triplet film.state becomes the dominant channel for the photodimer dissociation.22 Results and discussion Studies of 9-decylanthracene–1,2-dichloroethane solutions, by means of ultrafast absorption spectroscopy, show that after 1.Ultrafast kinetics excitation with a 355 nm, 30 ps laser pulse, short lived intermediates are formed which are characterized by the absorption The process of reversible photodimerization and photodissociation of polycyclic aromatic hydrocarbons such as spectra, shown in Fig. 3. Three new absorption bands with maxima at 560, 600 and 700 nm appear immediately after anthracenes may be used for developing photochromic materials, 13–15 which may also prove to be suitable for optical excitation.Comparison of this transient spectrum with the spectrum observed for anthracene and 9-methylanthracene5,27 switching and as 3D optical storage memory media.3,5 The photodimers were formed by excitation of the corresponding strongly suggests that this transient spectrum detected may well be the first excited singlet state of 9-decylanthracene. This monomer with 355 nm laser light.The reverse process occurs when the photodimers are transient singlet state decays with a lifetime of about 7 ns, while the new absorption band at l=430 nm is formed with exposed to 266 nm UV radiation. Reversibility is a very desirable property, because it makes this molecular system the same rate; this process is shown in Fig. 3. It is known28 that the triplet–triplet absorption spectrum of anthracene has suitable for application to re-writable memory disks and photonic switching. We have also studied the dissociation a band at 430 nm and it seems reasonable therefore to assume, that the observed absorption band belongs to the triplet excited kinetics and intermediates of this 9-anthracene derivative dimer photoreaction. The solubility of the photodimer in MMA was state, formed by relaxation of the excited singlet state via the intersystem crossing.A transient with absorption at 430 nm measured to be about 30 mg cm-3 (~5×10-2M). This high 1082 J. Mater. Chem., 1999, 9, 1081–1084Scheme 1 for the photodissociation process of 9-decylanthracene photodimers.Fig. 3 Transient absorption spectra of intermediates formed by 355 nm, 30 ps excitation of the monomer, measured at 100 ps and 16 ns after excitation pulse; the lifetime of the shorter transient 2. Application to 3D storage devices (5–10 ns) was determined by measuring the signal intensity vs. time delay (not shown here). The longer transient (430 nm) was assigned It has been shown previously3,29 that by means of two photon to the monomer triplet excited state.virtual absorption it is possible to write several greater than 100 Mbit disks inside a 2 cm3 volume, with a bit size of ~5 mm in diameter. It is possible, however, to store spots less than was also observed for anthracene and 9-methylanthracene,5 0.5 mm. For 0.5 mm spots more than 1 Tbit can be stored in which was also formed with the same rate as the rate of 9- 1 cm3.Practical devices using the two photon method for decylanthracene.We propose, therefore, that the photoreaction writing and one photon for reading have been demonstrated.29 mechanism of 9-decylanthracene in 1,2-dichloroethane solu- We have demonstrated that we can write and read tions follows the S0+hn�S1�S0+T1, T1�S0 relaxation, information, in 3D space, using 9-decylanthracene dimer dis- which is very similar to the mechanism for anthracene and 9-methylanthracene.persed in a PMMA matrix. To achieve this we have utilized Excitation of the photodimers in 1,2-dichloroethane solution two SHG 532 nm, 30 ps laser pulses, generated by a Nd/YAG with a 266 nm, 30 ps pulse leads to the formation of a transient picosecond laser, propagating in the optical path shown in with a broad absorption spectrum shown in Fig. 4. This Fig. 1. One of the 532 nm beams is passed through a glass spectrum is also similar to the one observed for anthracene slide which contained the information image that was to be and 9-methylanthracene and assigned to excimer formation.5 stored inside the bulk of the optical memory device, in this This excimer decays with a lifetime of about 7 ns, while the case a PMMA cube into which the photodimer is dispersed characteristic absorption band of the monomer triplet excited uniformly. The image on the glass slide was focused into a state at lmax=430 nm is formed with the same rate, Fig. 4. 5 mm×5 mm 2D plane inside the 9-decylanthracene The appearance of the monomer triplet–triplet absorption dimer–PMMA cube. The 532 nm beam carrying this inforband at 430 nm suggests that the excimer dissociates adia- mation is not absorbed because the 9-decylanthracene photodbatically with formation of a monomer molecule in the excited imer does not absorb at this wavelength.However, when it state.It has been shown previously,17–22 that the dissociation intersects the other 532 nm beam, which propagates orthogonal reaction of the photodimer at room temperature proceeds via to the first beam, inside the bulk of the cube, absorption by the S1 excited state. It was also observed, by means of time the 9-decylanthracene photodimer molecules takes place, resolved fluorescence,17 that monomer molecules were formed resulting in the formation of monomers.In computer termiin the S1 state during the photodissociation of the dimer. nology the zero (dimer) is converted to one (monomer) by two Based on the very strong similarity of the transient spectra photon absorption. Consequently, the image is stored in the and ultrafast kinetics found in this study for 9-decylanthracene area where the two beams intersected. The entire image to those of anthracene and 9-methylanthracene studied earlier,5 contained on the glass slide was written simultaneously.In we feel confident to propose the mechanism shown in Scheme 1 Fig. 5 we show the image of the USAF resolution target written by two-photon absorption of two 532 nm, 30 ps, 3 mJ cm-2 beams.Reading the stored information was achieved by one photon excitation of the monomers of a single 2D plane, within the cube, with a thin plane of light, which illuminates only one written plane, and subsequently detecting the fluorescence of the monomer with a 2D charge coupled device. The readout is not destructive because it is based on monomer fluorescence. A rather high concentration is desirable in order to increase the two photon absorption. Also a high concentration of the written molecules will result in higher emission intensity and therefore will be more easily detected.We were unable to detect a similar image using 9-methylanthracene photodimers under the same writing conditions, owing to the low solubility, hence concentration, of this molecule in the polymer host.The advantage of the 9-decylanthracene photochromic material is Fig. 4 Transient absorption spectra of intermediates formed by that a high density of information can be stored because of 266 nm, 30 ps excitation of the photodimer, measured at 100 ps and the high concentration of 9-decylanthracene that can be dis- 16 ns after excitation pulse. The excimer lifetime (100 ps delay specsolved in the polymer matrix.In addition the information can trum) was found to be 5;10 ns by measuring the signal intensity vs. be stored indefinitely because of the high stability of both time delay (not shown here). The signal at 430 nm was assigned to the monomer triplet excited state. photodimer and monomer forms at room temperature. J. Mater. Chem., 1999, 9, 1081–1084 1083References 1 J.H. Hong, I. McMichael, T. Y. Chang, W. Christian and E. G. Paek, Opt. Eng., 1995, 34, 2193. 2 X. An and D. Psaltis, Opt. Lett., 1995, 20, 1913. 3 A. S. Dvornikov, S. Esener and P. M. Rentzepis, in ‘Optical Computing Hardware’, eds. J. Jahns and S. H. Lee, Academic Press Inc., 1993, pp. 287–325. 4 A. S. Dvornikov, J. Malkin and P. M. Rentzepis, J. Phys.Chem., 1994, 98, 6746. 5 A. S. Dvornikov and P. M. Rentzepis, Res. Chem. Intermed., 1996, 22, 115. 6 A. S.Dvornikov and P. M. Rentzepis, Opt. Commun., 1997, 136, 1. 7 J. H. Stickler and W. W. Ebb, Opt. Lett., 1991,16, 1780. 8 R. R. Birge, Annu. Rev. Phys. Chem., 1990, 9, 683. 9 R. R. Birge, P. A. Fleitz, R. A. Gross, J. C. Izgi, A. F. Laurence, J. A. Stuart and J.R. Tallert, IEEE Trans. FMBS, 1990, 12, 1788. 10 C. Brauchle, N. Hampp and D. Oesterhelt, Adv. Mater., 1991, 3, 420. 11 R. Thoma, N. Hampp, C. Brauchle and D. Oesterhelt, Opt. Lett., 1990, 16, 651. 12 F. KrollpfeiVer and J. Branscheid, Ber., 1923, 56, 1617. 13 H. Bouas-Laurent, A. Castellan and J. P. Desvergne, Pure Appl. Chem., 1980, 52, 2633. 14 H. Bouas-Laurent and J.-P. Desvergne, in Photochromism: Molecules and Systems, eds.H. Durr and H. Bouas-Laurent, Elsevier, New York, 1990, ch. 14, pp. 561–630. 15 W. J. Tomlinson, E. A. Chandross, R. L. Fork, C. A. Pryde and A. A. Lamola, Appl. Opt., 1972, 11, 533. 16 I. B. Berlman, Handbook of Fluorescence Spectra of Aromatic Molecules, Academic Press, New York, 1971, p. 302. 17 S. Yamamoto, K. H. Grellman and A.Weller, Chem. Phys. Lett., 1980, 70, 241. 18 S. Yamamoto and K. H. Grellman, Chem. Phys. Lett., 1982, 92, 533. 19 J. Ferguson and M. Puza, Chem. Phys. Lett., 1978, 53, 215. Fig. 5 USAF Resolution target image written in 9-decylanthracene 20 W. R. Bergmark, G. Jones, T. E. Reinhardt and A. M. Halpern, photodimer–PMMA. Stored by two-photon excitation using 532 nm J. Am. Chem.Soc., 1978, 100, 6665. pulses. Accessing by a single, 370 nm, photon process. 21 M. A. Iannone and G. W. Scott, Mol. Cryst. Liq. Cryst., 1992, 211, 375. 22 S. Yamamoto and K. H. Grellman, Chem. Phys. Lett., 1982, 85, Conclusion 73. 23 L. E. Manring, K. S. Peters, G. Jones and W. R. Bergmark, J. Am. The ultrafast kinetics of the photoreaction of 9-decylanthra- Chem. Soc., 1985, 107, 1485.cene monomer and photodimer have been measured including 24 A. Castellan, R. Lapouyade and H. Bouas-Laurent, Bull. Soc. the rates of decay and formation of the various states and Chim. Fr., 1976, 201. relaxation pathways. The transient spectra of the intermediate 25 F. Fages, J.-P. Desvergne and H. Bouas-Laurent, Bull. Soc. Chim. states have been recorded and assigned. In addition, these Fr., 1985, 959.molecules have been used as optical storage media for 26 J. Ferguson, Chem. Phys. Lett., 1981, 78, 466. 27 N. Nakashima, M. Murakawa and N. Mataga, Bull. Chem. Soc. recording and accessing information in 3D storage devices. Jpn., 1976, 49, 854. 28 I. Carmichael, W. P. Helman and G. L. Hug, J. Phys. Chem. Acknowledgements Ref. Data, 1987, 16, 239. 29 A.S. Dvornikov, I. Cokgor, M. Wang, F. B. McCormick, The work of PMR and ASD was supported in part by USAF S. C. Esener and P. M. Rentzepis, IEEE Trans. CPMT-A, 1997, grant F30602–97-C-0029. We also thank Gilles Clavier for 20, 203. assistance in the purification and Vincent Darcos and Dr D. Bassani in the NMR spectra recording of 9-decylanthracene. Paper 8/08272C Travel support by NATO grant CRG 971516 is gratefully acknowledged. 1084 J. Mater. Chem., 1999, 9, 1081–1084 J O U R N A L O F C H E M I S T R Y Materials Ultrafast kinetics of 9-decylanthracene photodimers and their application to 3D optical storage Alexander S. Dvornikov,a Henri Bouas-Laurent,b Jean-Pierre Desvergneb and Peter M. Rentzepis*a aDepartment of Chemistry, University of California, Irvine, CA 92697 and Call/Recall Inc., Irvine, CA 92714, USA bPhotochimie Organique, CNRS UMR 5802, Universite� Bordeaux 1, F-33405, Talence Cedex, France Received 26th October 1998, Accepted 24th February 1999 The intermediate states and final product of the photoreaction of 9-decylanthracene monomer and photodimer have been identified and their formation and relaxation pathways and rates determined.In addition, these molecules have been used as media for high density 3D optical storage devices. Introduction The extremely large amounts of data that need to be stored and accessed at very fast rates are being limited by the memory capabilities as much as by interconnects and processors. Because of the huge data storage requirements, which continuously increase, and the rather mandatory need for parallel accessing, the need for compact, very high capacity parallel accessing, low cost memory devices is becoming very acute.One means for satisfying most of these requirements is provided by three-dimensional storage. The main research methods, which are pursued now, that may lead to practical 3D-storage memory devices, are phase holograms1,2 and two photon optical 3D memories.These devices utilize inorganic photorefractive crystals and photopolymers in the case of holography1,2 and organic3–7 or biological molecules such as bacteriorhodopsin8–11 dispersed in polymer matrices. In this paper we will be concerned with the photoreaction kinetics of 9-decylanthracene photodimers and the application of this data to 3D memory devices by means of two photon absorption.In previous papers we have presented the basic theory of two photon excitation and the means for utilizing such nonlinear processes for storing information within the volume of a 3D memory device.3–5 Briefly, the two photon absorption process shown in Fig. 1(a) causes a ground-state molecule, in the unwritten form, to be promoted to the first excited electronic state by the simultaneous absorption of two photons.The energy required to reach the excited state is greater than the energy of either photon alone, therefore each Fig. 1 a) Energy level diagram; b) writing of information in 3D by two-photon absorption. beam propagates through the memory volume without being absorbed. When the sum of the energies of these photons is equal to or greater than the energy gap between the ground shown schematically in Fig. 1(b).For fast parallel information and first electronic excited states of the molecule (see Fig. 1(a)) transfer of an entire 2D plane, located within the volume of at the point of the intersection of the beams, within the the 3D device, a single photon process is used to illuminate volume, the two photons may be absorbed simultaneously.this plane, see Fig. 1. To achieve a high density storage and The excited molecule decays and is transformed into a diVerent be suitable for use in 3D storage devices the photochromic ground-state molecular structure, which becomes the written materials used should fulfil several requirements, including: form of the 3D molecular memory device.This new molecular high nonlinear absorption cross-section; stability in both write structure absorbs at longer wavelengths than the original and read states; high solubility in the polymer matrix; the ground-state molecules, and therefore it may be excited by written form should have a high fluorescence quantum one or two photons of lower energy than those absorbed by eYciency and fluorescence in an easily detected region.the original, unwritten form molecules. The written excited We have shown previously5 that photo-induced reversible state molecules fluoresce and detection of the induced fluores- photodimerization of anthracene and 9-methylanthracene fulfil cence is used for accessing the stored information. It is, quite well most of these requirements except that they do not therefore, possible to access any bit of information within the dissolve suYciently well in PMMA [poly(methyl methacrylate] volume by simply intersecting two bea at the place where and other polymer matrices.This diYculty has been eliminated because we have been able to synthesize 9-n-decylanthracene the information bit to be retrieved is located.This process is J. Mater. Chem., 1999, 9, 1081–1084 1081which, as we anticipated, exhibits photochromism and in addition is quite soluble in polymer matrices, thus being a promising material for optical switching and 3D memory devices. Here we present the ultrafast kinetics of this new photochromic material, which is based on the reversible photodimerization of 9-n-decylanthracene dispersed in PMMA.In addition to the basic science information regarding the mechanism of the photo-transformation of this material, we found that the photo-induced dimers of this substituted anthracene fulfil most of the requirements for use as a 3D computer memory material. Experimental Fig. 2 Absorption spectra of a) photodimer, b) long wavelength A double beam Shimadzu UV160U spectrophotometer and a contribution of monomer and c) fluorescence spectrum of 9-decyl- Shimadzu RF 5000U spectrofluorophotometer recorded the anthracene (monomer) in 1,2-dichloroethane solution, at room ground state absorption and fluorescence spectra respectively.temperature, conc. #10-4 M. A ‘Quantel’ Nd/YAG laser generated 30 ps, 1064 nm pulses, which were converted to 532 nm pulses and used for twosolubility in MMA makes possible the fabrication of solid photon excitation.The ultrafast kinetics, transient absorption PMMA blocks with homogeneously dispersed dimer molecules spectra and mechanism, presented here, were studied by the at the concentrations which are required to store high density ultrafast experimental system described previously.5 of information in 3D volume.The materials used for the synthesis of the photodimer and The absorption spectra of the monomer and photodimer all solvents were Aldrich spectroscopic purity grade. 9- are shown in Fig. 2. The monomer has its long wavelength Decylanthracene (white crystals with a blue fluorescence, absorption band in the 300–400 nm region, while the photod- mp 44 °C) was synthesized using the method described in imer is blue shifted and has practically no absorption at ref. 12 and purified by column chromatography (silica gel; wavelengths longer than 300 nm. The photodissociation of this eluent: pentane). 1H NMR d (CCl4): 0.85–0.95 (t, 3H, Aproduct results in the regeneration of the monomer, which is (CH2)9-CH3), 1.25–1.5 (m, 2H), 1.5–1.65 (m, 2H), 1.7–1.85 a conjugated double bond system and therefore exhibits a red (m, 14H), 3.5–3.6 (m, 2H, benzylic: A-CH2-(CH2)8-CH3), shifted absorption band. The monomer of 9-decylanthracene 7.35–7.45 (m, 4HAr), 7.85–7.95 (m, 2HAr), 8.15–8.25 (m, was found to fluoresce in the 380 to 450 nm region, see Fig. 2. 3HAr). 13C NMR d (CDCl3): 14.3(CH3), 22.9, 28.3, 29.6, 29.8, The fluorescence quantum yields of the monomer in ethyl 29.9, 30.6, 31.6, 32.1 (-(CH2)9-), 124.6, 124.9, 125.4, 125.6, acetate and 1,2-dichloroethane solutions were measured to be 129.4 (CAr-H), 129.7, 131.8, 135.6 (quaternary C).HRMS: wF=0.35 and wF=0.70 respectively; the standard used for the (AutoSpec EQ FAB+) C24H30 calc: 318.234751, found: calibration of the quantum yield was a solution of dimethyl 318.234305. POPOP in cyclohexane, which is known to emit with a The anthracene photodimer was generated by irradiation quantum eYciency of 0.93.16 The photodimer does not with l>320 nm light emitted by a 150W arc Xenon lamp, for fluoresce at room temperature. 48 h, of a deaerated saturated hexane solution of 9-decyl- The formation and photodissociation mechanism of anthracene (#25 mg cm-3). The white solid precipitate was anthracene photodimers and related compounds has been collected, washed with fresh solvent and dried.It was shown extensively investigated previously.14,17–26 Anthracene and its by NMR that the head-to-tail photodimer was the sole derivatives may form the corresponding photodimers after photoproduct formed.13,14 excitation of the monomer to its first singlet excited state14 Thin polymer films of 9-decylanthracene photodimers and the subsequent interaction of two monomers.When the dispersed in PMMA were prepared by pouring a solution of photodimers are excited to the first allowed electronic state it the photodimers and PMMA in 1,2-dichloroethane on the may be shown that some of them may dissociate adiabatically surface of a microscope slide.The slide was then spin-coated, via intermediate excimer formation.5,17–19,22 It has also been resulting in a 20 mm thickness uniformly distributed polymer reported that at low temperature (77 K) the first excited triplet film. state becomes the dominant channel for the photodimer dissociation.22 Results and discussion Studies of 9-decylanthracene–1,2-dichloroethane solutions, by means of ultrafast absorption spectroscopy, show that after 1.Ultrafast kinetics excitation with a 355 nm, 30 ps laser pulse, short lived intermediates are formed which are characterized by the absorption The process of reversible photodimerization and photodissociation of polycyclic aromatic hydrocarbons such as spectra, shown in Fig. 3. Three new absorption bands with maxima at 560, 600 and 700 nm appear immediately after anthracenes may be used for developing photochromic materials, 13–15 which may also prove to be suitable for optical excitation. Comparison of this transient spectrum with the spectrum observed for anthracene and 9-methylanthracene5,27 switching and as 3D optical storage memory media.3,5 The photodimers were formed by excitation of the corresponding strongly suggests that this transient spectrum detected may well be the first excited singlet state of 9-decylanthracene.This monomer with 355 nm laser light. The reverse process occurs when the photodimers are transient singlet state decays with a lifetime of about 7 ns, while the new absorption band at l=430 nm is formed with exposed to 266 nm UV radiation.Reversibility is a very desirable property, because it makes this molecular system the same rate; this process is shown in Fig. 3. It is known28 that the triplet–triplet absorption spectrum of anthracene has suitable for application to re-writable memory disks and photonic switching. We have also studied the dissociation a band at 430 nm and it seems reasonable therefore to assume, that the observed absorption band belongs to the triplet excited kinetics and intermediates of this 9-anthracene derivative dimer photoreaction.The solubility of the photodimer in MMA was state, formed by relaxation of the excited singlet state via the intersystem crossing. A transient with absorption at 430 nm measured to be about 30 mg cm-3 (~5×10-2M).This high 1082 J. Mater. Chem., 1999, 9, 1081–1084Scheme 1 for the photodissociation process of 9-decylanthracene photodimers. Fig. 3 Transient absorption spectra of intermediates formed by 355 nm, 30 ps excitation of the monomer, measured at 100 ps and 16 ns after excitation pulse; the lifetime of the shorter transient 2. Application to 3D storage devices (5–10 ns) was determined by measuring the signal intensity vs.time delay (not shown here). The longer transient (430 nm) was assigned It has been shown previously3,29 that by means of two photon to the monomer triplet excited state. virtual absorption it is possible to write several greater than 100 Mbit disks inside a 2 cm3 volume, with a bit size of ~5 mm in diameter. It is possible, however, to store spots less than was also observed for anthracene and 9-methylanthracene,5 0.5 mm.For 0.5 mm spots more than 1 Tbit can be stored in which was also formed with the same rate as the rate of 9- 1 cm3. Practical devices using the two photon method for decylanthracene.We propose, therefore, that the photoreaction writing and one photon for reading have been demonstrated.29 mechanism of 9-decylanthracene in 1,2-dichloroethane solu- We have demonstrated that we can write and read tions follows the S0+hn�S1�S0+T1, T1�S0 relaxation, information, in 3D space, using 9-decylanthracene dimer dis- whivery similar to the mechanism for anthracene and 9-methylanthracene.persed in a PMMA matrix. To achieve this we have utilized Excitation of the photodimers in 1,2-dichloroethane solution two SHG 532 nm, 30 ps laser pulses, generated by a Nd/YAG with a 266 nm, 30 ps pulse leads to the formation of a transient picosecond laser, propagating in the optical path shown in with a broad absorption spectrum shown in Fig. 4. This Fig. 1. One of the 532 nm beams is passed through a glass spectrum is also similar to the one observed for anthracene slide which contained the information image that was to be and 9-methylanthracene and assigned to excimer formation.5 stored inside the bulk of the optical memory device, in this This excimer decays with a lifetime of about 7 ns, while the case a PMMA cube into which the photodimer is dispersed characteristic absorption band of the monomer triplet excited uniformly.The image on the glass slide was focused into a state at lmax=430 nm is formed with the same rate, Fig. 4. 5 mm×5 mm 2D plane inside the 9-decylanthracene The appearance of the monomer triplet–triplet absorption dimer–PMMA cube. The 532 nm beam carrying this inforband at 430 nm suggests that the excimer dissociates adia- mation is not absorbed because the 9-decylanthracene photodbatically with formation of a monomer molecule in the excited imer does not absorb at this wavelength.However, when it state. It has been shown previously,17–22 that the dissociation intersects the other 532 nm beam, which propagates orthogonal reaction of the photodimer at room temperature proceeds via to the first beam, inside the bulk of the cube, absorption by the S1 excited state. It was also observed, by means of time the 9-decylanthracene photodimer molecules takes place, resolved fluorescence,17 that monomer molecules were formed resulting in the formation of monomers.In computer termiin the S1 state during the photodissociation of the dimer. nology the zero (dimer) is converted to one (monomer) by two Based on the very strong similarity of the transient spectra photon absorption.Consequently, the image is stored in the and ultrafast kinetics found in this study for 9-decylanthracene area where the two beams intersected. The entire image to those of anthracene and 9-methylanthracene studied earlier,5 contained on the glass slide was written simultaneously. In we feel confident to propose the mechanism shown in Scheme 1 Fig. 5 we show the image of the USAF resolution target written by two-photon absorption of two 532 nm, 30 ps, 3 mJ cm-2 beams.Reading the stored information was achieved by one photon excitation of the monomers of a single 2D plane, within the cube, with a thin plane of light, which illuminates only one written plane, and subsequently detecting the fluorescence of the monomer with a 2D charge coupled device.The readout is not destructive because it is based on monomer fluorescence. A rather high concentration is desirable in order to increase the two photon absorption. Also a high concentration of the written molecules will result in higher emission intensity and therefore will be more easily detected. We were unable to detect a similar image using 9-methylanthracene photodimers under the same writing conditions, owing to the low solubility, hence concentration, of this molecule in the polymer host.The advantage of the 9-decylanthracene photochromic material is Fig. 4 Transient absorption spectra of intermediates formed by that a high density of information can be stored because of 266 nm, 30 ps excitation of the photodimer, measured at 100 ps and the high concentration of 9-decylanthracene that can be dis- 16 ns after excitation pulse.The excimer lifetime (100 ps delay specsolved in the polymer matrix. In addition the information can trum) was found to be 5–10 ns by measuring the signal intensity vs. be stored indefinitely because of the high stability of both time delay (not shown here). The signal at 430 nm was assigned to the monomer triplet excited state.photodimer and monomer forms at room temperature. J. Mater. Chem., 1999, 9, 1081–1084 1083References 1 J. H. Hong, I. McMichael, T. Y. Chang, W. Christian and E. G. Paek, Opt. Eng., 1995, 34, 2193. 2 X. An and D. Psaltis, Opt. Lett., 1995, 20, 1913. 3 A. S. Dvornikov, S. Esener and P. M. Rentzepis, in ‘Optical Computing Hardware’, eds.J. Jahns and S. H. Lee, Academic Press Inc., 1993, pp. 287–325. 4 A. S. Dvornikov, J. Malkin and P. M. Rentzepis, J. Phys. Chem., 1994, 98, 6746. 5 A. S. Dvornikov and P. M. Rentzepis, Res. Chem. Intermed., 1996, 22, 115. 6 A. S.Dvornikov and P. M. Rentzepis, Opt. Commun., 1997, 136, 1. 7 J. H. Stickler and W. W. Ebb, Opt. Lett., 1991,16, 1780. 8 R. R. Birge, Annu. Rev. Phys.Chem., 1990, 9, 683. 9 R. R. Birge, P. A. Fleitz, R. A. Gross, J. C. Izgi, A. F. Laurence, J. A. Stuart and J. R. Tallert, IEEE Trans. FMBS, 1990, 12, 1788. 10 C. Brauchle, N. Hampp and D. Oesterhelt, Adv. Mater., 1991, 3, 420. 11 R. Thoma, N. Hampp, C. Brauchle and D. Oesterhelt, Opt. Lett., 1990, 16, 651. 12 F. KrollpfeiVer and J. Branscheid, Ber., 1923, 56, 1617. 13 H. Bouas-Laurent, A.Castellan and J. P. Desvergne, Pure Appl. Chem., 1980, 52, 2633. 14 H. Bouas-Laurent and J.-P. Desvergne, in Photochromism: Molecules and Systems, eds. H. Durr and H. Bouas-Laurent, Elsevier, New York, 1990, ch. 14, pp. 561–630. 15 W. J. Tomlinson, E. A. Chandross, R. L. Fork, C. A. Pryde and A. A. Lamola, Appl. Opt., 1972, 11, 533. 16 I. B. Berlman, Handbook of Fluorescence Spectra of Aromatic Molecules, Academic Press, New York, 1971, p. 302. 17 S. Yamamoto, K. H. Grellman and A. Weller, Chem. Phys. Lett., 1980, 70, 241. 18 S. Yamamoto and K. H. Grellman, Chem. Phys. Lett., 1982, 92, 533. 19 J. Ferguson and M. Puza, Chem. Phys. Lett., 1978, 53, 215. Fig. 5 USAF Resolution target image written in 9-decylanthracene 20 W. R. Bergmark, G. Jones, T. E. Reinhardt and A. M. Halpern, photodimer–PMMA. Stored by two-photon excitation using 532 nm J. Am. Chem. Soc., 1978, 100, 6665. pulses. Accessing by a single, 370 nm, photon process. 21 M. A. Iannone and G. W. Scott, Mol. Cryst. Liq. Cryst., 1992, 211, 375. 22 S. Yamamoto and K. H. Grellman, Chem. Phys. Lett., 1982, 85, Conclusion 73. 23 L. E. Manring, K. S. Peters, G. Jones and W. R. Bergmark, J. Am. The ultrafast kinetics of the photoreaction of 9-decylanthra- Chem. Soc., 1985, 107, 1485. cene monomer and photodimer have been measured including 24 A. Castellan, R. Lapouyade and H. Bouas-Laurent, Bull. Soc. the rates of decay and formation of the various states and Chim. Fr., 1976, 201. relaxation pathways. The transient spectra of the intermediate 25 F. Fages, J.-P. Desvergne and H. Bouas-Laurent, Bull. Soc. Chim. states have been recorded and assigned. In addition, these Fr., 1985, 959. molecules have been used as optical storage media for 26 J. Ferguson, Chem. Phys. Lett., 1981, 78, 466. 27 N. Nakashima, M. Murakawa and N. Mataga, Bull. Chem. Soc. recording and accessing information in 3D storage devices. Jpn., 1976, 49, 854. 28 I. Carmichael, W. P. Helman and G. L. Hug, J. Phys. Chem. Acknowledgements Ref. Data, 1987, 16, 239. 29 A. S. Dvornikov, I. Cokgor, M. Wang, F. B. McCormick, The work of PMR and ASD was supported in part by USAF S. C. Esener and P. M. Rentzepis, IEEE Trans. CPMT-A, 1997, grant F30602–97-C-0029. We also thank Gilles Clavier for 20, 203. assistance in the purification and Vincent Darcos and Dr D. Bassani in the NMR spectra recording of 9-decylanthracene. Paper 8/08272C Travel support by NATO grant CRG 971516 is gratefully acknowledged. 1084 J. Mater. Chem., 1999, 9, 1081–1084

ISSN:0959-9428    

DOI:10.1039/a808272c

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials First ferrocene-containing low molar mass organosiloxane liquidcrystalline materials Harry J. Coles,a* Sebastien Meyer,a Petra Lehmann,a Robert Deschenauxb and Isabelle Jauslinb aSouthampton Liquid Crystal Institute, Department of Physics and Astronomy, University of Southampton, Southampton, UK SO17 1BJ. E-mail: hjc@lc.phys.soton.ac.uk; http://www.slci.soton.ac.uk bInstitut de Chimie, Universite� de Neucha�tel, Av.de Bellevaux 51, 2000 Neucha�tel, Switzerland Received 2nd February 1999, Accepted 16th February 1999 Two ferrocene-containing low molar mass organosiloxane liquid-crystalline materials have been synthesised and their phase-transition behaviour investigated. The v-unsaturated ferrocene precursor was hydrosilylated by addition of pentamethyldisiloxane or heptamethyltrisiloxane in the presence of platinum divinyltetramethyldisiloxane complex.The ferrocene precursor presents smectic A and smectic C phases; the disiloxane compound exhibits a smectic C phase; the trisiloxane compound shows a smectic C phase and two higher order smectic phases. Tilt angle measurements were performed on samples oriented on polytetrafluoroethylene (PTFE) friction deposited layers since rubbed polyimide (PI ) only gave very poor alignment.For the precursor the tilt angle was found to be very small, close to 2°, while for the two organosiloxane ferrocene compounds it was close to 28° and almost independent of temperature. These compounds exhibited ferroelectric electro-optic switching properties when doped with 1–2% w/w of chiral mesogens.(structures 2 and 3), which represent a novel family of metal- Introduction based anisotropic materials (see Synthesis and Fig. 1). We will There is a growing interest in metal-containing liquid- present data for two homologues, containing 2 and 3 silicon crystalline materials1,2 that combine some of the properties of atoms, in comparison with the v-unsaturated ferrocene precurmetals with those of mesogenic moieties since this could lead sor.This will allow us to demonstrate that, despite its relatively to processable materials with interesting anisotropic optical, small size compared to the bulky ferrocene mesogenic group, electronic and magnetic properties. Owing to its unique redox the siloxane moiety plays an important role in the phase characteristics, ferrocene is a valuable unit for building up stability, phase sequence, electro-optic properties and quality switchable systems3 and recently, electron transfer was used of the surface alignment.to generate mesomorphism in the ferrocene–ferrocenium redox system.4 Ferrocene-containing side-chain liquid-crystal polymers Experimental (SCLCPs) have been reported in the literature.1,2,5 The SCLCPs oVer several advantages over the low molar mass Synthesis liquid crystals in that they have better mechanical character- The ferrocene precursor 1 was synthesised as previously istics, a broader mesomorphic range and a reduced or supdescribed. 2 1,1,1,3,3-Pentamethyldisiloxane, 1,1,1,3,3,5,5-hep- pressed tendency to form crystalline phases.However, because tamethyltrisiloxane and the platinum catalyst (platinum– of the rigidity of the polymer backbone, the SCLCPs are divinyltetramethyldisiloxane complex, 3–3.5% platinum con- usually highly viscous and therefore have long switching centration in vinyl terminated polydimethylsiloxane, neutral ) response times.6 The SCLCPs based on flexible polysiloxane were purchased from Fluorochem.Toluene was freed of thio- backbones have shorter response times than those based on phene according to standard procedures and then dried over more rigid polymers such as the polymethacrylates. sodium in the presence of benzophenone. Furthermore, the polydispersity of the commercially available 1H and 13C NMR spectra were recorded on a Bruker AC materials is a serious problem for the production of materials 300 spectrometer.Mass spectra were recorded on a Micromass with reproducible characteristics.6 Recently, several low molar Platform quadrupole mass analyser with an electrospray ion mass liquid-crystalline organosiloxanes have been developed6–9 source. that show fast electro-optic responses whilst retaining some of Compounds 2 and 3 were synthesised following the the ruggedness of the polymeric systems.6 They display mainly procedure outlined in Fig. 1. smectic phases due to the micro-segregation of the mesogenic, 0.1 mmol of the precursor 1 and the platinum catalyst were paraYnic and siloxane moieties into distinct sublayers within dissolved in 1 ml of toluene under dry argon such that a the lamellar phase7,8 and a resultant agglomeration of the Pt5alkene ratio in the range 15104–106 was obtained.The siloxane units into a ‘virtual’ backbone.9 The smectic layers, solution was stirred at 60 °C for one hour. 0.11 mmol of the for fractionated monodisperse siloxane moieties, are particuhybrid functionalised siloxane was added. The reaction mixture larly well defined and the isotropic to smectic phase transition was stirred at 90 to 100 °C for three days.Fresh catalyst was is normally first order. For polydisperse siloxane groups the added every 24 h. On cooling, the solvent was removed under phase sequence may be altered and the materials tend to reduced pressure. The crude product was passed through a favour lower order smectic phases.10 This is important in the silica gel column (Merck, Si60, 40–63 mm) using dichloro- present work since we will present the synthesis and thermal methane as the eluent, and re-precipitated in methanol and properties of the first ferrocene-containing low molar mass and monodisperse organosiloxane liquid-crystalline materials hexane until no amount of precursor could be detected.J. Mater. Chem., 1999, 9, 1085–1090 1085O C O O O C O C O O C O OC18H37 O O C O O O C O C O O C O OC18H37 O Si O Si H Si O Si Fe Fe n Pt#, toluene n 1 2 n = 1 3 n = 2 Fig. 1 Synthetic route to compounds 2 and 3. Freeze-drying from benzene gave the pure product in high ES MS m/z: 1342.5 (M+NH4)+, 1347.3 (M+Na)+, 1363.5 (M+K)+. yield. 1-[4-(4-(11-(1,1,1,3,3-Pentamethyldisiloxyl )undecyloxy)- Physical characterisation benzoyloxy)phenyl]-1¾-[4-(4-octadecyloxyphenyloxycarbonyl)- The mesomorphic properties of the compounds synthesised phenyl]ferrocenedicarboxylate (2).Yield: 89%; 1H NMR (1–3) were studied by thermal optical microscopy and (300 MHz, CDCl3) d: 0.0–0.2 (m, 15 H, Si-CH3), 0.56 (m, 2 diVerential scanning calorimetry (DSC). The DSC measure- H, Si-CH2), 0.90 (t, 3 H, CH3). 1.2–1.9 (m, 50 H, CH2), 3.97 ments were carried out on a Perkin-Elmer DSC7 instrument (t, 2 H, CH2-O), 4.05 (t, 2 H, CH2-O), 4.64 (m, 4 H, Cp), on samples weighing between 2 and 4 mg and at scanning 5.10 (m, 4 H, Cp), 6.95–7.40 (m, 12 H, aromatic), 8.11–8.26 rates of 5 °Cmin-1 (1 and 2) and 2 °Cmin-1 (3) (Fig. 2).(m, 4 H, aromatic); 13C NMR (300 MHz, CDCl3) d: 0.23, Phase characterisation by polarised light microscopy was 1.31, 1.82, 14.15, 18.30, 22.70, 26.00, 29.73, 31.94, 33.46, carried out using an Olympus BH-2 microscope equipped with 114.27, 115.09, 121.35, 122.37, 127.10, 131.75, 132.30, 144.18, a TMS91 Linkam hot stage stable to 0.1 °C over a temperature 147.96, 148.45, 154.79, 156.88, 163.59, 164.82, 168.82; ES MS range from -196 °C to 600 °C.The enthalpies and tempera- m/z: 1269.5 (M+NH4)+, 1276.5 (M+Na)+. tures of the phase transitions are reported in Table 1. Preliminary powder X-ray diVraction was carried out using 1-[4-(4-(11-(1,1,1,3,3,3,5,5-Heptamethyltrisiloxyl)- an apparatus previously described,7 to confirm the basic phase undecyloxy)benzoyloxy)phenyl]-1¾-[4-(4-octadecyloxyphenylstructure. oxycarbonyl)phenyl]ferrocenedicarboxylate (3).Yield: 85%; 1H NMR (300 MHz, CDCl3) d: 0.0–0.2 (m, 21 H, Si-CH3), 0.57 (m, 2 H, Si-CH2), 0.92 (t, 3 H, CH3), 1.2–1.9 (m, 50 H, Results and discussion CH2), 3.95 (t, 2 H, CH2-O), 4.06 (t, 2 H, CH2-O), 4.66 (m, 4 H, Cp), 5.11 (m, 4 H, Cp), 6.95–7.38 (m, 12 H, aromatic), The precursor (compound 1) gives smectic A and sme phases over temperature ranges of 12.5 °C and 12 °C respect- 8.12–8.25 (m, 4 H, aromatic); 13C NMR (300 MHz, CDCl3) d: 0.22, 1.29, 1.83, 14.14, 18.29, 22.71, 26.01, 29.72, 31.93, ively, with a clearing temperature at 146.5 °C.Attaching the pentamethyldisiloxane or heptamethyltrisiloxane units, for 33.47, 114.29, 115.08, 121.33, 122.39, 127.11, 131.76, 132.31, 144.19, 147.98, 148.47, 154.80, 156.90, 163.58, 164.81, 168.80; compounds 2 and 3 respectively, suppresses the smectic A Table 1 Enthalpies and temperatures of phase transitions Compound Phase transitions, T/°C (DH/J g-1) 1 K–SC 122 (39.1) SC–SA 134a SA–I 146.5 (10.9) 2 K–SC 125.3 (18.6) SC–I 142.8 (8.4) 3 K–S2 122 (16.8) S2–S1 123.6a (0.3b) S1–SC 124.5a (2.5b) SC–I 135.6 (6.7) aDetermined by polarized optical microscopy.bDetermined from second DSC cooling run. 1086 J. Mater. Chem., 1999, 9, 1085–1090We were not able to achieve a satisfactory alignment of the specimens on conventional rubbed polyimide (PI ). Instead, the materials were aligned on friction deposited polytetra- fluoroethylene (PTFE) using a technique developed by Wittmann and Smith.13 Using the friction deposition apparatus described by Hanmer14 and under optimal experimental conditions (temperature of 300 °C, pressure of 106 Pa and deposition rate of 0.25 mm s-1)15 a very thin, around 20 to 30 nm thick, quasi-monocrystalline film with the PTFE chains aligned in the direction of friction (called the alignment direction hereafter), can be deposited on a hard counterface such as a glass slide.A wide range of crystalline and liquid crystalline materials have been successfully aligned on friction deposited PTFE layers.13–18 Fig. 4 allows a comparison between the quality of alignment obtained with commercial rubbed PI (Fig. 4a) and friction deposited PTFE layers (Fig. 4b). Sample thicknesses are 10 mm. While only small mono-oriented domains could be grown on PI, large uniformly aligned films could be made on PTFE layers (notice the diVerence in scales in Fig. 4). As far as we are aware this is the first time that such good alignment has been demonstrated for ferrocene based liquid crystals. The optical tilt angle h (i.e. the angle between the director and the layer normal, c.f. Fig. 7) and the angle a between the PTFE alignment direction and the layer normal were measured using a method, described by Dierking et al.,19 developed for Fig. 2 DSC Thermograms (second cooling and third heating) of smectic C* phases. To induce a smectic C* phase in our compound 3. materials we prepared mixtures of the three compounds with a compatible (i.e. miscible) chiral dopant added in very low concentration (i.e. 1–2% w/w). We assume that such a low phase and both materials exhibit a direct isotropic to smectic C phase on cooling (Fig. 2). The clearing temperature concentration will not aVect the value of the optical tilt angle. For compound 1 we used SCE2 (Merck, UK) as the chiral decreased systemically, by a few degrees, with increasing siloxane content. For compound 2 the smectic C phase range additive, whilst for compounds 2 and 3 we used the chlorosubstituted ferroelectric organosiloxanes, with 2 or 3 silicon is broadened to 17.5 °C and this was the only mesophase observed.However while compound 2 shows no para- atoms respectively, described elsewhere.20 Besides providing a method for measuring the optical tilt angle the observed morphotic textures, the DSC thermograms of compound 3 (with the longer siloxane chain) revealed the existence of two switching demonstrates that these organosiloxane grafted ferrocene materials are capable of exhibiting ferroelectric proper- higher order smectic modifications (denoted 1 and 2) below the smectic C phase which is itself 11.1 °C wide.Of these two ties (Fig. 5). Cells made of PTFE covered ITO glass were filled with each mixture and we measured the light transmitted phase transitions only the smectic C�smectic 1 transition is unambiguously observed by thermal microscopy.This trans- through the microscope with crossed polarisers when a square wave voltage is applied to the cell. With a field of 3 V mm-1 ition is marked by fluctuations of the schlieren brushes moving wave-like across the preparation. Compared to the schlieren we found that this tilt angle was invariant with increasing field.The intensity of the transmitted light in the positive texture displayed by the smectic C phase (Fig. 3a) the texture in the S1 phase (Fig. 3b) appears to be ‘frozen’, with shadowed (dpos) and negative (dneg) switched states was recorded and plotted as a function of the rotation angle Q of the sample. areas bordered by optical discontinuities. Comparing the observed textures with the available literature,11 the corre- Q=0 was defined when the PTFE alignment direction was parallel to one of the crossed polarisers.The two intensity sponding phase could be assigned as the hexatic tilted smectic F or I. As these phases only diVer in the tilt direction relative curves can be fitted by20 the following equations: to the local hexagonal lattice, their textures are similar (c.f.Ipos=sin2(2(Q+dpos)) and Ineg=sin2(2(Q-dneg)) Plates 85 and 86 and page 131 in ref. 11). Unless the two phases occur in sequence, which is not the case here, it is The optical tilt angle h is then given by almost impossible to diVerentiate the phases (I and F) by optical microscopy alone. On further cooling the fans develop h= dpos+dneg 2 a striated texture, while the schlieren brushes start to disappear giving way to a shadowed mosaic texture (Fig. 3c). However, assigning these observed changes to the smectic 1�smectic 2 At the position of the cross-over of the intensity curves, the two states appear the same and the corresponding value of Q transition is not obvious as this texture does not seem to be thermodynamically very stable.Crystallisation takes place is the angle a between the PTFE alignment direction and the layer normal. The values of the optical tilt angles are in almost instantly. This is marked by the formation of bands across the focal-conic fans and the development of uniform agreement with those indicated by the preliminary X-ray measurements.21 The layer spacing of the precursor (compound domains consisting of overlapping platelets (Fig. 3d) and these are reminiscent of the texture displayed by the crystal E phase. 1) in the smectic A and C phases is #60.2 A° whilst in compounds 2 and 3 it is 54.8 A° and 55.6 A° , respectively. These To unambiguously define the smectic 1 and 2 phases we would need to carry out detailed X-ray analysis on aligned samples.12 latter values are independent of temperature, except very close to the clearing temperature. Miscibility studies with known smectic I and F materials are not an obvious experimental technique to use with the present The temperature dependences of h and a are shown in Fig. 6(a) and (b) respectively. For compound 1, the precursor, materials since there are currently no known chemically compatible organosiloxane ferrocenes available that exhibit such the optical tilt angle is very small, i.e.close to 2°, while a decreases slightly from 30° to 24° on reaching the smectic A phases. J. Mater. Chem., 1999, 9, 1085–1090 1087Fig. 3 Optical textures given on cooling of compound 3 using untreated glass substrates to promote both schlieren and focal-conic textures.Observation is between crossed polarisers. a) 127 °C. b) 124 °C. c) 123.4 °C. d) 120.4 °C. Scale bar corresponds to 100 mm. 1088 J. Mater. Chem., 1999, 9, 1085–1090Fig. 4 Microphotography between crossed polarisers of compound 1 aligned on diVerent substrates: a) rubbed polyimide, b) friction deposited PTFE. Scale bars correspond to 100 mm. Fig. 6 (a) Temperature dependence of the optical tilt angle, #: compound 1, +: compound 2, 6: compound 3. (b) Temperature Fig. 5 Ferroelectric optical switch of compound 3 doped with the dependence of the angle a between the PTFE alignment direction and chiral additive. The upper trace shows the applied electric field whilhe layer normal, #: compound 1, +: compound 2, 6: compound 3. the lower shows the latched ferroelectric switching. J.Mater. Chem., 1999, 9, 1085–1090 1089C to isotropic phase transition. Thus the organosiloxane based ferrocenes show liquid crystalline properties remarkably diVerent from those of the vinyl precursor. Further addition of chiral dopants has allowed ferroelectric properties to be demonstrated, for the first time, in these compounds.We are currently researching into the origins of this ferroelectric behaviour and its implications for the other electromagnetic properties of these new organometallic materials.Acknowledgements HJC thanks the EPSRC for a research grant GR/K/70908 and Merck UK Ltd for studentship support for PL. We thank Dr Carboni for useful discussions and Dr Guillon (IPCMS) Strasbourg for providing the X-ray facilities.Fig. 7 Relative orientation of the layers (thick lines), director n, layer normal k, and PTFE orienting direction for the precursor 1 and References compound 2. The fine lines indicate the directions of n and k in each case. Typical values (c.f. Fig. 6) of h and a are 2° and 26° for 1 and 1 (a) R. Deschenaux and J. W. Goodby, in Ferrocenes, ed. A.Togni 28° and 24° for 2 respectively (data at T-Ttransition#-5°). and T. Hayashi, VCH, Weinheim, 1995, ch. 9; (b) P. Zanello, in Ferrocenes, ed. A. Togni and T. Hayashi, VCH, Weinheim, 1995, phase. The introduction of a siloxane moiety increases ch. 7. 2 (a) R. Deschenaux, I. Kosztics, U. Scholten, D. Guillon and dramatically the value of the tilt angles. For compounds 2 M. Ibn-Elhaj, J.Mater. Chem., 1994, 4, 1351; (b) R. Deschenaux, and 3, the optical tilt angles are close to 28° (equivalent to a I. Jauslin, U. Scholten, F. Turpin, D. Guillon and B. Heinrich, ferroelectric cone angle of 56°) over a wide temperature range Macromolecules, 1998, 31, 5647. while a decreases from 24° to 17° with increasing temperature. 3 J. C. Medina, I. Gay, Z. Chen, L. Echegoyen and G.W. Gokel, The diVerence between the optical tilt angles of the siloxane J. Am. Chem. Soc., 1991, 113, 365. compounds and the precursor is extremely high, i.e. 26°, and 4 R. Deschenaux, M. Schweissguth and A.-M. Levelut, Chem. Commun., 1996, 1275. this is much higher than in a similar comparison for other 5 R. Deschenaux, V. Izvolenski, F. Turpin, D. Guillon and low molar mass organosiloxane liquid-crystalline materials,22 B.Heinrich, Chem. Commun., 1996, 439. where the increase due to hydrosilylation was typically 3–4°. 6 J. Newton, H. J. Coles, P. Hodge and J. Hannington, J. Mater. Since the values of the angle a (i.e. between the layer normal Chem., 1994, 4, 869. and the PTFE alignment direction), for the precursor and 7 M. Ibn-Elhaj, H. J.Coles, D. Guillon and A. Skoulios, J. Phys. II the siloxane compounds, only diVer by around 6° the (France), 1993, 3, 1807. 8 E. Corsellis, D. Guillon, P. Kloess and H. J. Coles, Liq. Cryst., increased diVerence in the optical tilt angles is not due to a 1997, 23, 235. tilting of the layer but is caused by a tilt of the molecules 9 H. J. Coles, H. Owen, J. Newton and P. Hodge, Liq.Cryst., 1993, within the layers. Fig. 7 shows the relative orientations of the 15, 739. layers and the molecules with respect to the PTFE alignment 10 M. Redmond, H. J. Coles, E. WischerhoV and R. Zentel, direction. Thus addition of the siloxane units has an extremely Ferroelectrics, 1993, 148, 323. marked eVect on the orientation of the mesogens in the 11 G. W. Gray and J. W.Goodby, Smectic Liquid Crystals— Textures and Structures, Leonard-Hill, London, 1984. smectic layer. This is all the more remarkable when the size 12 R. Date, G. R. Luckhurst, M. Shuman and J. M. Seddon, J. Phys. ratio of the ferrocene mesogenic and siloxane moieties is II (France), 1995, 5, 587. taken into account. 13 J.-C.Wittmann and P. Smith, Nature, 1991, 352, 414. 14 J. Hanmer, PhD Thesis, University of Manchester, 1995. 15 S. Meyer, Doctorate es Sciences Thesis, University Louis Pasteur Conclusion Strasbourg, 1995. 16 J.-C. Wittmann, S. Meyer, P. Damman, M. Dosiere and We have successfully synthesised the first ferrocene containing H.-W. Schmidt, Polymer, 1998, 39, 3545. low molar mass organosiloxane liquid-crystalline materials. 17 M. Suzuki, A. Ferenez, S.Iida, V. Enkelmann and G. Wegner, By using PTFE friction-deposited films as alignment substrates Adv. Mater., 1993, 5, 359. we have managed to obtain large planar oriented domains in 18 G. Lester, J. Hanmer and H. Coles, Mol. Cryst. Liq. Cryst., 1995, 10 mm thick cells. DSC measurements and optical microscopy 262, 149. observations have shown that the grafting of a siloxane moiety 19 I.Dierking, F. Gießelmann, J. Schacht and P. Zugenmaier, Liq. Cryst., 1995, 19, 179. leads to a broadening of the smectic C phase as well as 20 W. K. Robinson, C. Carboni, P. Kloess, S. P. Perkins and suppression of the smectic A phase observed in the precursor H. J. Coles, Liq. Cryst., 1998, 25, 301. compound. This grafting leads to a marked increase of the tilt 21 H. Allouchi, unpublished preliminary X-ray data taken at IPCMS, angle for the organosiloxane based compounds in comparison Strasbourg.with the precursor. Additionally, the longer heptamethyltrisi- 22 P. Kloess, Ph.D. Thesis, University of Southampton, 1997. loxane unit induces two, as yet unidentified, higher order smectic phases. The tilt angles are independent of the temperature Paper 9/00867E apart from within a 3° wide pretransitional region at the smectic 1090 J.Mater. Chem., 1999, 9, 1085–1090 J O U R N A L O F C H E M I S T R Y Materials First ferrocene-containing low molar mass organosiloxane liquidcrystalline materials Harry J. Coles,a* Sebastien Meyer,a Petra Lehmann,a Robert Deschenauxb and Isabelle Jauslinb aSouthampton Liquid Crystal Institute, Department of Physics and Astronomy, University of Southampton, Southampton, UK SO17 1BJ.E-mail: hjc@lc.phys.soton.ac.uk; http://www.slci.soton.ac.uk bInstitut de Chimie, Universite� de Neucha�tel, Av. de Bellevaux 51, 2000 Neucha�tel, Switzerland Received 2nd February 1999, Accepted 16th February 1999 Two ferrocene-containing low molar mass organosiloxane liquid-crystalline materials have been synthesised and their phase-transition behaviour investigated. The v-unsaturated ferrocene precursor was hydrosilylated by addition of pentamethyldisiloxane or heptamethyltrisiloxane in the presence of platinum divinyltetramethyldisiloxane complex.The ferrocene precursor presents smectic A and smectic C phases; the disiloxane compound exhibits a smectic C phase; the trisiloxane compound shows a smectic C phase and two higher order smectic phases.Tilt angle measurements were performed on samples oriented on polytetrafluoroethylene (PTFE) friction deposited layers since rubbed polyimide (PI ) only gave very poor alignment. For the precursor the tilt angle was found to be very small, close to 2°, while for the two organosiloxane ferrocene compounds it was close to 28° and almost independent of temperature.These compounds exhibited ferroelectric electro-optic switching properties when doped with 1–2% w/w of chiral mesogens. (structures 2 and 3), which represent a novel family of metal- Introduction based anisotropic materials (see Synthesis and Fig. 1). We will There is a growing interest in metal-containing liquid- present data for two homologues, containing 2 and 3 silicon crystalline materials1,2 that combine some of the properties of atoms, in comparison with the v-unsaturated ferrocene precurmetals with those of mesogenic moieties since this could lead sor.This will allow us to demonstrate that, despite its relatively to processable materials with interesting anisotropic optical, small size compared to the bulky ferrocene mesogenic group, electronic and magnetic properties.Owing to its unique redox the siloxane moiety plays an important role in the phase characteristics, ferrocene is a valuable unit for building up stability, phase sequence, electro-optic properties and quality switchable systems3 and recently, electron transfer was used of the surface alignmenenerate mesomorphism in the ferrocene–ferrocenium redox system.4 Ferrocene-containing side-chain liquid-crystal polymers Experimental (SCLCPs) have been reported in the literature.1,2,5 The SCLCPs oVer several advantages over the low molar mass Synthesis liquid crystals in that they have better mechanical character- The ferrocene precursor 1 was synthesised as previously istics, a broader mesomorphic range and a reduced or supdescribed. 2 1,1,1,3,3-Pentamethyldisiloxane, 1,1,1,3,3,5,5-hep- pressed tendency to form crystalline phases. However, because tamethyltrisiloxane and the platinum catalyst (platinum– of the rigidity of the polymer backbone, the SCLCPs are divinyltetramethyldisiloxane complex, 3–3.5% platinum con- usually highly viscous and therefore have long switching centration in vinyl terminated polydimethylsiloxane, neutral ) response times.6 The SCLCPs based on flexible polysiloxane were purchased from Fluorochem.Toluene was freed of thio- backbones have shorter response times than those based on phene according to standard procedures and then dried over more rigid polymers such as the polymethacrylates. sodium in the presence of benzophenone.Furthermore, the polydispersity of the commercially available 1H and 13C NMR spectra were recorded on a Bruker AC materials is a serious problem for the production of materials 300 spectrometer. Mass spectra were recorded on a Micromass with reproducible characteristics.6 Recently, several low molar Platform quadrupole mass analyser with an electrospray ion mass liquid-crystalline organosiloxanes have been developed6–9 source.that show fast electro-optic responses whilst retaining some of Compounds 2 and 3 were synthesised following the the ruggedness of the polymeric systems.6 They display mainly procedure outlined in Fig. 1. smectic phases due to the micro-segregation of the mesogenic, 0.1 mmol of the precursor 1 and the platinum catalyst were paraYnic and siloxane moieties into distinct sublayers within dissolved in 1 ml of toluene under dry argon such that a the lamellar phase7,8 and a resultant agglomeration of the Pt5alkene ratio in the range 15104–106 was obtained.The siloxane units into a ‘virtual’ backbone.9 The smectic layers, solution was stirred at 60 °C for one hour. 0.11 mmol of the for fractionated monodisperse siloxane moieties, are particuhybrid functionalised siloxane was added.The reaction mixture larly well defined and the isotropic to smectic phase transition was stirred at 90 to 100 °C for three days. Fresh catalyst was is normally first order. For polydisperse siloxane groups the added every 24 h. On cooling, the solvent was removed under phase sequence may be altered and the materials tend to reduced pressure.The crude product was passed through a favour lower order smectic phases.10 This is important in the silica gel column (Merck, Si60, 40–63 mm) using dichloro- present work since we will present the synthesis and thermal methane as the eluent, and re-precipitated in methanol and properties of the first ferrocene-containing low molar mass and monodisperse organosiloxane liquid-crystalline materials hexane until no amount of precursor could be detected.J. Mater. Chem., 1999, 9, 1085–1090 1085O C O O O C O C O O C O OC18H37 O O C O O O C O C O O C O OC18H37 O Si O Si H Si O Si Fe Fe n Pt#, toluene n 1 2 n = 1 3 n = 2 Fig. 1 Synthetic route to compounds 2 and 3. Freeze-drying from benzene gave the pure product in high ES MS m/z: 1342.5 (M+NH4)+, 1347.3 (M+Na)+, 1363.5 (M+K)+.yield. 1-[4-(4-(11-(1,1,1,3,3-Pentamethyldisiloxyl )undecyloxy)- Physical characterisation benzoyloxy)phenyl]-1¾-[4-(4-octadecyloxyphenyloxycarbonyl)- The mesomorphic properties of the compounds synthesised phenyl]ferrocenedicarboxylate (2). Yield: 89%; 1H NMR (1–3) were studied by thermal optical microscopy and (300 MHz, CDCl3) d: 0.0–0.2 (m, 15 H, Si-CH3), 0.56 (m, 2 diVerential scanning calorimetry (DSC).The DSC measure- H, Si-CH2), 0.90 (t, 3 H, CH3). 1.2–1.9 (m, 50 H, CH2), 3.97 ments were carried out on a Perkin-Elmer DSC7 instrument (t, 2 H, CH2-O), 4.05 (t, 2 H, CH2-O), 4.64 (m, 4 H, Cp), on samples weighing between 2 and 4 mg and at scanning 5.10 (m, 4 H, Cp), 6.95–7.40 (m, 12 H, aromatic), 8.11–8.26 rates of 5 °Cmin-1 (1 and 2) and 2 °Cmin-1 (3) (Fig. 2). (m, 4 H, aromatic); 13C NMR (300 MHz, CDCl3) d: 0.23, Phase characterisation by polarised light microscopy was 1.31, 1.82, 14.15, 18.30, 22.70, 26.00, 29.73, 31.94, 33.46, carried out using an Olympus BH-2 microscope equipped with 114.27, 115.09, 121.35, 122.37, 127.10, 131.75, 132.30, 144.18, a TMS91 Linkam hot stage stable to 0.1 °C over a temperature 147.96, 148.45, 154.79, 156.88, 163.59, 164.82, 168.82; ES MS range from -196 °C to 600 °C.The enthalpies and tempera- m/z: 1269.5 (M+NH4)+, 1276.5 (M+Na)+. tures of the phase transitions are reported in Table 1. Preliminary powder X-ray diVraction was carried out using 1-[4-(4-(11-(1,1,1,3,3,3,5,5-Heptamethyltrisiloxyl)- an apparatus previously described,7 to confirm the basic phase undecyloxy)benzoyloxy)phenyl]-1¾-[4-(4-octadecyloxyphenylstructure. oxycarbonyl)phenyl]ferrocenedicarboxylate (3).Yield: 85%; 1H NMR (300 MHz, CDCl3) d: 0.0–0.2 (m, 21 H, Si-CH3), 0.57 (m, 2 H, Si-CH2), 0.92 (t, 3 H, CH3), 1.2–1.9 (m, 50 H, Results and discussion CH2), 3.95 (t, 2 H, CH2-O), 4.06 (t, 2 H, CH2-O), 4.66 (m, 4 H, Cp), 5.11 (m, 4 H, Cp), 6.95–7.38 (m, 12 H, aromatic), The precursor (compound 1) gives smectic A and smectic C phases over temperature ranges of 12.5 °C and 12 °C respect- 8.12–8.25 (m, 4 H, aromatic); 13C NMR (300 MHz, CDCl3) d: 0.22, 1.29, 1.83, 14.14, 18.29, 22.71, 26.01, 29.72, 31.93, ively, with a clearing temperature at 146.5 °C.Attaching the pentamethyldisiloxane or heptamethyltrisiloxane units, for 33.47, 114.29, 115.08, 121.33, 122.39, 127.11, 131.76, 132.31, 144.19, 147.98, 148.47, 154.80, 156.90, 163.58, 164.81, 168.80; compounds 2 and 3 respectively, suppresses the smectic A Table 1 Enthalpies and temperatures of phase transitions Compound Phase transitions, T/°C (DH/J g-1) 1 K–SC 122 (39.1) SC–SA 134a SA–I 146.5 (10.9) 2 K–SC 125.3 (18.6) SC–I 142.8 (8.4) 3 K–S2 122 (16.8) S2–S1 123.6a (0.3b) S1–SC 124.5a (2.5b) SC–I 135.6 (6.7) aDetermined by polarized optical microscopy.bDetermined from second DSC cooling run. 1086 J. Mater. Chem., 1999, 9, 1085–1090We were not able to achieve a satisfactory alignment of the specimens on conventional rubbed polyimide (PI ). Instead, the materials were aligned on friction deposited polytetra- fluoroethylene (PTFE) using a technique developed by Wittmann and Smith.13 Using the friction deposition apparatus described by Hanmer14 and under optimal experimental conditions (temperature of 300 °C, pressure of 106 Pa and deposition rate of 0.25 mm s-1)15 a very thin, around 20 to 30 nm thick, quasi-monocrystalline film with the PTFE chains aligned in the direction of friction (called the alignment direction hereafter), can be deposited on a hard counterface such as a glass slide.A wide range of crystalline and liquid crystalline materials have been successfully aligned on friction deposited PTFE layers.13–18 Fig. 4 allows a comparison between the quality of alignment obtained with commercial rubbed PI (Fig. 4a) and friction deposited PTFE layers (Fig. 4b). Sample thicknesses are 10 mm. While only small mono-oriented domains could be grown on PI, large uniformly aligned films could be made on PTFE layers (notice the diVerence in scales in Fig. 4). As far as we are aware this is the first time that such good alignment has been demonstrated for ferrocene based liquid crystals. The optical tilt angle h (i.e.the angle between the director and the layer normal, c.f. Fig. 7) and the angle a between the PTFE alignment direction and the layer normal were measured using a method, described by Dierking et al.,19 developed for Fig. 2 DSC Thermograms (second cooling and third heating) of smectic C* phases. To induce a smectic C* phase in our compound 3. materials we prepared mixtures of the three compounds with a compatible (i.e.miscible) chiral dopant added in very low concentration (i.e. 1–2% w/w). We assume that such a low phase and both materials exhibit a direct isotropic to smectic C phase on cooling (Fig. 2). The clearing temperature concentration will not aVect the value of the optical tilt angle. For compound 1 we used SCE2 (Merck, UK) as the chiral decreased systemically, by a few degrees, with increasing siloxane content.For compound 2 the smectic C phase range additive, whilst for compounds 2 and 3 we used the chlorosubstituted ferroelectric organosiloxanes, with 2 or 3 silicon is broadened to 17.5 °C and this was the only mesophase observed. However while compound 2 shows no para- atoms respectively, described elsewhere.20 Besides providing a method for measuring the optical tilt angle the observed morphotic textures, the DSC thermograms of compound 3 (with the longer siloxane chain) revealed the existence of two switching demonstrates that these organosiloxane grafted ferrocene materials are capable of exhibiting ferroelectric proper- higher order smectic modifications (denoted 1 and 2) below the smectic C phase which is itself 11.1 °C wide.Of these two ties (Fig. 5). Cells made of PTFE covered ITO glass were filled with each mixture and we measured the light transmitted phase transitions only the smectic C�smectic 1 transition is unambiguously observed by thermal microscopy. This trans- through the microscope with crossed polarisers when a square wave voltage is applied to the cell.With a field of 3 V mm-1 ition is marked by fluctuations of the schlieren brushes moving wave-like across the preparation. Compared to the schlieren we found that this tilt angle was invariant with increasing field. The intensity of the transmitted light in the positive texture displayed by the smectic C phase (Fig. 3a) the texture in the S1 phase (Fig. 3b) appears to be ‘frozen’, with shadowed (dpos) and negative (dneg) switched states was recorded and plotted as a function of the rotation angle Q of the sample.areas bordered by optical discontinuities. Comparing the observed textures with the available literature,11 the corre- Q=0 was defined when the PTFE alignment direction was parallel to one of the crossed polarisers. The two intensity sponding phase could be assigned as the hexatic tilted smectic F or I.As these phases only diVer in the tilt direction relative curves can be fitted by20 the following equations: to the local hexagonal lattice, their textures are similar (c.f. Ipos=sin2(2(Q+dpos)) and Ineg=sin2(2(Q-dneg)) Plates 85 and 86 and page 131 in ref. 11). Unless the two phases occur in sequence, which is not the case here, it is The optical tilt angle h is then given by almost impossible to diVerentiate the phases (I and F) by optical microscopy alone.On further cooling the fans develop h= dpos+dneg 2 a striated texture, while the schlieren brushes start to disappear giving way to a shadowed mosaic texture (Fig. 3c). However, assigning these observed changes to the smectic 1�smectic 2 At the position of the cross-over of the intensity curves, the two states appear the same and the corresponding value of Q transition is not obvious as this texture does not seem to be thermodynamically very stable.Crystallisation takes place is the angle a between the PTFE alignment direction and the layer normal. The values of the optical tilt angles are in almost instantly. This is marked by the formation of bands across the focal-conic fans and the development of uniform agreement with those indicated by the preliminary X-ray measurements.21 The layer spacing of the precursor (compound domains consisting of overlapping platelets (Fig. 3d) and these are reminiscent of the texture displayed by the crystal E phase. 1) in the smectic A and C phases is #60.2 A° whilst in compounds 2 and 3 it is 54.8 A° and 55.6 A° , respectively. These To unambiguously define the smectic 1 and 2 phases we would need to carry out detailed X-ray analysis on aligned samples.12 latter values are independent of temperature, except very close to the clearing temperature.Miscibility studies with known smectic I and F materials are not an obvious experimental technique to use with the present The temperature dependences of h and a are shown in Fig. 6(a) and (b) respectively. For compound 1, the precursor, materials since there are currently no known chemically compatible organosiloxane ferrocenes available that exhibit such the optical tilt angle is very small, i.e. close to 2°, while a decreases slightly from 30° to 24° on reaching the smectic A phases.J. Mater. Chem., 1999, 9, 1085–1090 1087Fig. 3 Optical textures given on cooling of compound 3 using untreated glass substrates to promote both schlieren and focal-conic textures. Observation is between crossed polarisers. a) 127 °C. b) 124 °C. c) 123.4 °C. d) 120.4 °C. Scale bar corresponds to 100 mm. 1088 J. Mater. Chem., 1999, 9, 1085–1090Fig. 4 Microphotography between crossed polarisers of compound 1 aligned on diVerent substrates: a) rubbed polyimide, b) friction deposited PTFE.Scale bars correspond to 100 mm. Fig. 6 (a) Temperature dependence of the optical tilt angle, #: compound 1, +: compound 2, 6: compound 3. (b) Temperature Fig. 5 Ferroelectric optical switch of compound 3 doped with the dependence of the angle a between the PTFE alignment direction and chiral additive.The upper trace shows the applied electric field whilst the layer normal, #: compound 1, +: compound 2, 6: compound 3. the lower shows the latched ferroelectric switching. J. Mater. Chem., 1999, 9, 1085–1090 1089C to isotropic phase transition. Thus the organosiloxane based ferrocenes show liquid crystalline properties remarkably diVerent from those of the vinyl precursor.Further addition of chiral dopants has allowed ferroelectric properties to be demonstrated, for the first time, in these compounds.We are currently researching into the origins of this ferroelectric behaviour and its implications for the other electromagnetic properties of these new organometallic materials. Acknowledgements HJC thanks the EPSRC for a research grant GR/K/70908 and Merck UK Ltd for studentship support for PL. We thank Dr Carboni for useful discussions and Dr Guillon (IPCMS) Strasbourg for providing the X-ray facilities.Fig. 7 Relative orientation of the layers (thick lines), director n, layer normal k, and PTFE orienting direction for the precursor 1 and References compound 2. The fine lines indicate the directions of n and k in each case.Typical values (c.f. Fig. 6) of h and a are 2° and 26° for 1 and 1 (a) R. Deschenaux and J. W. Goodby, in Ferrocenes, ed. A. Togni 28° and 24° for 2 respectively (data at T-Ttransition#-5°). and T. Hayashi, VCH, Weinheim, 1995, ch. 9; (b) P. Zanello, in Ferrocenes, ed. A. Togni and T. Hayashi, VCH, Weinheim, 1995, phase. The introduction of a siloxane moiety increases ch. 7. 2 (a) R. Deschenaux, I. Kosztics, U. Scholten, D. Guillon and dramatically the value of the tilt angles. For compounds 2 M. Ibn-Elhaj, J. Mater. Chem., 1994, 4, 1351; (b) R. Deschenaux, and 3, the optical tilt angles are close to 28° (equivalent to a I. Jauslin, U. Scholten, F. Turpin, D. Guillon and B. Heinrich, ferroelectric cone angle of 56°) over a wide temperature range Macromolecules, 1998, 31, 5647.while a decreases from 24° to 17° with increasing temperature. 3 J. C. Medina, I. Gay, Z. Chen, L. Echegoyen and G. W. Gokel, The diVerence between the optical tilt angles of the siloxane J. Am. Chem. Soc., 1991, 113, 365. compounds and the precursor is extremely high, i.e. 26°, and 4 R. Deschenaux, M. Schweissguth and A.-M.Levelut, Chem. Commun., 1996, 1275. this is much higher than in a similar comparison for other 5 R. Deschenaux, V. Izvolenski, F. Turpin, D. Guillon and low molar mass organosiloxane liquid-crystalline materials,22 B. Heinrich, Chem. Commun., 1996, 439. where the increase due to hydrosilylation was typically 3–4°. 6 J. NewtH. J. Coles, P. Hodge and J. Hannington, J. Mater.Since the values of the angle a (i.e. between the layer normal Chem., 1994, 4, 869. and the PTFE alignment direction), for the precursor and 7 M. Ibn-Elhaj, H. J. Coles, D. Guillon and A. Skoulios, J. Phys. II the siloxane compounds, only diVer by around 6° the (France), 1993, 3, 1807. 8 E. Corsellis, D. Guillon, P. Kloess and H. J. Coles, Liq. Cryst., increased diVerence in the optical tilt angles is not due to a 1997, 23, 235.tilting of the layer but is caused by a tilt of the molecules 9 H. J. Coles, H. Owen, J. Newton and P. Hodge, Liq. Cryst., 1993, within the layers. Fig. 7 shows the relative orientations of the 15, 739. layers and the molecules with respect to the PTFE alignment 10 M. Redmond, H. J. Coles, E. WischerhoV and R. Zentel, direction.Thus addition of the siloxane units has an extremely Ferroelectrics, 1993, 148, 323. marked eVect on the orientation of the mesogens in the 11 G. W. Gray and J. W. Goodby, Smectic Liquid Crystals— Textures and Structures, Leonard-Hill, London, 1984. smectic layer. This is all the more remarkable when the size 12 R. Date, G. R. Luckhurst, M. Shuman and J. M. Seddon, J. Phys. ratio of the ferrocene mesogenic and siloxane moieties is II (France), 1995, 5, 587. taken into account. 13 J.-C.Wittmann and P. Smith, Nature, 1991, 352, 414. 14 J. Hanmer, PhD Thesis, University of Manchester, 1995. 15 S. Meyer, Doctorate es Sciences Thesis, University Louis Pasteur Conclusion Strasbourg, 1995. 16 J.-C. Wittmann, S. Meyer, P. Damman, M. Dosiere and We have successfully synthesised the first ferrocene containing H.-W. Schmidt, Polymer, 1998, 39, 3545. low molar mass organosiloxane liquid-crystalline materials. 17 M. Suzuki, A. Ferenez, S. Iida, V. Enkelmann and G. Wegner, By using PTFE friction-deposited films as alignment substrates Adv. Mater., 1993, 5, 359. we have managed to obtain large planar oriented domains in 18 G. Lester, J. Hanmer and H. Coles, Mol. Cryst. Liq. Cryst., 1995, 10 mm thick cells. DSC measurements and optical microscopy 262, 149. observations have shown that the grafting of a siloxane moiety 19 I. Dierking, F. Gießelmann, J. Schacht and P. Zugenmaier, Liq. Cryst., 1995, 19, 179. leads to a broadening of the smectic C phase as well as 20 W. K. Robinson, C. Carboni, P. Kloess, S. P. Perkins and suppression of the smectic A phase observed in the precursor H. J. Coles, Liq. Cryst., 1998, 25, 301. compound. This grafting leads to a marked increase of the tilt 21 H. Allouchi, unpublished preliminary X-ray data taken at IPCMS, angle for the organosiloxane based compounds in comparison Strasbourg. with the precursor. Additionally, the longer heptamethyltrisi- 22 P. Kloess, Ph.D. Thesis, University of Southampton, 1997. loxane unit induces two, as yet unidentified, higher order smectic phases. The tilt angles are independent of the temperature Paper 9/00867E apart from within a 3° wide pretransitional region at the smectic 1090 J. Mater. Chem., 1999, 9, 1085–1090

ISSN:0959-9428    

DOI:10.1039/a900867e

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Crystal growth of a stable nonlinear optical organic material: 2-amino-5-nitropyridinium monohydrogen L-tartrate Julien Zaccaro,a Fre�de�ric Lorutb and Alain Ibaneza* aLaboratoire de Cristallographie, C.N.R.S., associe� a` l’Universite� J. Fourier et a` l’Institut National Polytechnique de Grenoble, BP 166, F-38042, Grenoble Cedex 09, France bEuropean Synchrotron Radiation Facility, BP 220, F-38043, Grenoble Cedex, France Received 24th December 1998, Accepted 18th February 1999 We report on the optimization of synthesis, solubility and crystal growth from solution of 2-amino-5- nitropyridinium L-tartrate ( labelled 2A5NPLT).This newly engineered guest–host hydrogen-bonding salt exhibits a quasi-perfect polar alignment of the nonlinear chromophores in the crystalline lattice which is promising for optimized Pockels electrooptic properties.In addition, the dense and 3D-intermolecular framework built up by short hydrogen bonds leads to improved thermal, mechanical and chemical stability compared with that of organic molecular crystals. The solvent selection allowed the elaboration of bulky crystal by the temperature lowering method.Synchrotron X-ray diVraction topography experiments carried out on uncut crystals revealed a high crystalline perfection. 2A5NP+ cations are nearly aligned along the b-axis. In Introduction addition, their H-bonding anchoring onto the polyanionic In the last two decades, extensive research has shown that helical tartrate chains is very dense leading to a 3D framework organic crystals can exhibit nonlinear optical (NLO) (Fig. 1). A first NLO characterization showed a large susceptie Yciencies which are two orders of magnitude higher than bility coeYcient d33=41 pm V-1 at l=1.06 mm.19 Moreover, those of inorganic materials.1,2 This is due to highly polarizable the alignment of the dipole moments in the 2A5NPLT crystalmolecules involving conjugated systems of p electrons such as line lattice is promising for optimized Pockels electrooptic polyenes or aromatic compounds.These organic phases possess properties. Thus, in order to specify these NLO potentialities, other advantages such as unlimited molecular engineering and we have undertaken the elaboration of high quality 2A5NPLT often high laser damage thresholds.In addition, large single crystals. Our first results obtained on the synthesis optimizcrystals can be grown from solution close to room tempera- ation, solubility and crystal growth from solution are displayed ture.3,4 Nevertheless, due to their poor chemical stability and here. Then, a morphological study of 2A5NPLT is introduced. low thermal and mechanical resistance, the use of organic Finally, we have determined their transparency bandwidth, molecular crystals is currently limited in their NLO industrial the thermal and mechanical stability while the crystalline applications.An added drawback is that the polar nature of quality was defined using synchrotron X-ray topography. these phases with one- or two-dimensional structures leads to important problems in crystal growth and processing of Synthesis and solubility optical devices.In order to overcome these diYculties, a strategy has recently The 2A5NP molecule is a weak Brønsted base which can be been developed which aims to build very cohesive acentric protonated in a strong acidic medium (pH <2). This induces crystalline structures based on host–guest chemistry.5 The the dissolution of this molecule in aqueous acidic solutions by anchorage of highly polarizable chromophores onto various formation of the 2-amino-5-nitropyridinium cation (2A5NP+) inorganic or organic anionic matrices leads to short and and leads to the synthesis of hydrogen-bonded salts with the multiple hydrogen-bonded networks.Several polarizable molconjugated bases of strong or medium acids (pKa <3.3).In ecules were selected in this crystal engineering and particularly the case of tartaric acid, the 2A5NPLT phase can be directly 2-amino-5-nitropyridine (2A5NP). A great number of salts obtained by reaction (1). having a 3D acentric structure based on herringbone motifs were obtained with this push–pull molecule such as the dihy- C5H5N3O2+L-C4H6O6CAC5H6N3O2+ drogen phosphate,6 dihydrogen arsenate,7 chloride, bromide,8 +L-C4H5O6-CAC5H6N3O2+·C4H5O6- chloroacetate9 and acetophosphonate of 2-amino-5-nitropyridinium. 10 These salts possess enhanced stability (chemical, (2A5NP) (2A5NPLT) (1) thermal, mechanical ) compared to the corresponding molecu- Nevertheless, tartaric acid is a too weak Brønsted acid that lar organic crystals.Moreover, they exhibit a wider transdoes not completely dissolve the 2A5NP powder in aqueous parency range and bulky crystal morphology.11,12 The first solutions due to partial 2A5NP protonation. This explains the qualitative evaluations of the nonlinear response (powder tests low synthesis yields, around 50%. We have optimized the of second harmonic generation13) showed high eYciencies 2A5NPLT synthesis by using acetic acid as an intermediate for all these materials.These results are confirmed by the solvent. Indeed, this acid (pKa=4.76) does not react with first nonlinear characterizations carried out on single crys- 2A5NP but makes it possible, at high concentrations tals.14–16 Furthermore, optical parametric oscillation has just (13–16M), to reach low pH values (pH<1) and to completely been demonstrated with the 2-amino-5-nitropyridinium dissolve the amount of 2A5NP in reaction (1).Thus, the dihydrogen phosphate (2A5NPDP) compound.17 2A5NPLT salt is obtained by first dissolving the 2A5NP in an From this crystal engineering, another type of salt structure acetic acid solution (16M) at 60 °C. Then, a saturated tartaric was evidenced.Indeed, in the crystal structure of 2-amino-5- aqueous solution (4 M) is added. The molar ratios are 1515555 nitropyridinium L-tartrate (2A5NPLT), space group P21 with a=8.248 A° , b=9.199 A° , c=7.611 A° and b=96°.5,18 the for 2A5NP, C4H6O6, CH3COOH and H2O respectively. The J. Mater. Chem., 1999, 9, 1091–1095 1091solutions higher 2A5NPLT solubilities and solubility–temperature gradients than in tartaric solvent (Fig. 2). Despite these increases, S and DS/DT remain rather low for the use of the typical temperature lowering (TL) method. Thus, in order to adjust the 2A5NPLT solubility, another solvent was used: formic acid (pKa=3.75) which does not react with 2A5NP. Indeed, as for acetic acid, low pH solutions can be achieved at high concentrations (HCOOH 10–20 M).In these formic acid solutions we have obtained an important increase of S and DS/DT compared to tartaric and acetic solvents (Fig. 2). These significant evolutions allow a great flexibility for selecting the growth technique and optimizing crystal growth conditions. Crystal growth and morphology The first growths of 2A5NPLT crystals were carried out in tartaric acid solutions (3M).The low DS/DT value for this solvent prompted us to select thermal gradient methods. A horizontal temperature-gradient technique (HTG) was used to obtain high-quality seeds by spontaneous nucleation. Nevertheless, this growth reactor, described elsewhere,12 does not allow the growth of large single crystals. In addition, the resulting crystals exhibit a lot of growth bands evidenced by Fig. 1 Projection of the polar structure of 2A5NPLT in the ab plane X-ray diVraction topography12 because the seeds and the showing a quasi-perfect alignment of the 2-amino-5-nitropyridinium cations. growth solution are not stirred. Thus, in order to obtain large crystals of high quality we then used a vertical temperaturegradient method (VTG).20 The nutrient comprises millimetric salt crystallizes at room temperature as a white microcrystalline 2A5NPLT crystals placed in the upper zone of a container of powder on addition of acetone. Thus, the synthesis yield is controlled porosity.The thermal configuration, with a higher close to 100%. In order to increase the crystal quality, we take temperature in the upper zone (DT#3–4 °C), reduces the partare in ensuring the purity of the starting materials.natural convection. The seeds, previously grown by the HTG The commercial 2A5NP is first purified by double sublimation method, are stacked to silica suspensions in the lower zone and then dissolved in high purity acidic solutions. Finally, the and are rotated at a rate of 20 rpm with inversion of the salts are purified again from CH3COOH solutions, by rotation direction every 40 s.Thus, after 3–4 months of recrystallization in acetone. growth, large and optically clear crystals are obtained Due to the thermal decomposition of 2A5NPLT (see [Fig. 3(a)]. Unfortunately, these 2A5NPLT crystals have weak Stability section), we have undertaken the crystal growth of growth rates, around 0.1–0.2 mm day-1, and a plate mor- this phase from acidic solutions.We first selected as solvent phology. This crystal morphology was determined from X-ray the acid corresponding to the salt.We determined the solubility diVraction (Lau�e diagrams) and interfacial angle measure- curves between 20 and 60 °C for various molar concentrations ments using a two-circle optical goniometer (Nedinsco) of tartaric acid (2–4 M) which satisfy the pH conditions (pH [Fig. 4(a)]. The crystal morphology has been drawn with the <2). Unfortunately, the magnitude of the 2A5NPLT solubility program SHAPE,21,22 using the experimental growth rates as (S) and its variation with temperature (solubility–temperature the face-crystal center in the WulV plot. The 2A5NPLT plate gradient DS/DT) in these concentrated aqueous solutions are habit is mainly due to a very low growth rate of the {101} always very low (Fig. 2).form. In contrast to other salts of this family,11,12 the On the other hand, we have seen for other salts of this 2A5NPLT phase grows in a number of well faceted forms of family that the solubility curve is directly related to the pH of low Miller’s indexes [Fig. 4(a)]. the growth solution; the lower the pH value, the higher are Then, in order to obtain bulky crystals, we used another the solubility and the solubility–temperature gradient.11,12 As growth solvent. As we also obtain platelet crystals with acetic previously mentioned, acetic acid aqueous solutions (13–16M) acid, we selected formic acid solutions which lead to bulky allow low pH values to be achieved without combining with crystals. Moreover, the higher solubility and solubility– the 2A5NP molecule.Thus, we registered in these acetic temperature gradient values registered for this solvent allow us to carry out the 2A5NPLT crystal growth by the typical temperature lowering (TL) method. For a good control of crystal growth, the temperature lowering rates have to be less than 0.2 °C per day at low relative supersaturations around 0.01.This requires a high temperature precision (±0.002 °C) to avoid undesirable supersaturation changes leading to secondary nucleation, macro steps and multiple solvent inclusions on the growing crystal faces. These growth problems are favored in this case by a narrow metastable zone. The growth solution (around 200 cm3) is seeded with 2A5NPLT crystal about 1 mm3 in size, obtained by spontaneous nucleation in the same TL crystallizer previously described.11 The growing crystals are rotated (20–30 rpm) and are reversed every 45 s.The temperature is slowly decreased in the range 45–35 °C. This second set of growth experiments allowed us to obtain in 2–3 months bulky and optically clear crystals [Fig. 3(b)]. Fig. 2 Solubility curves of the 2A5NPLT salt in tartaric (3 M), acetic (16 M) and formic acid (13 M) aqueous solutions. This bulky morphology is essentially achieved through a 1092 J. Mater. Chem., 1999, 9, 1091–1095Fig. 3 2A5NPLT crystals grown (a) by the HTG method in tartaric acid solutions (3 M), (b) by the LT technique in formic acid (13 M) aqueous solutions.significant increase of the {101} growth rate [Fig. 4(b)]. other hand, hardness measurements were carried out using a Leitz Vickers hardness tester. Indentations were made on Unfortunately, the average growth rates remain very weak ( less than 0.2 mm day-1). This is certainly due to the adsorp- (101), (001), (111), (1-11) and (1-1-1) growth faces with a dwelling time kept constant at 15 s and weights ranging from tion of polar solvent molecules on the crystal faces of this highly polar material. 15–25 g. In agreement with the 3D crystal structure, all these experiments gave the same average value of 105±5 kgmm-2. The 2A5NPLT hardness is the highest of the 2A5NP salt Crystal characterizations family and is rather close to that of KDP (135 kg mm-2).These good mechanical properties make slicing and polishing Stability easier for the optical device processing of 2A5NPLT crystals. A diVerential scanning calorimetry study associated with Finally, this tartrate possesses the lowest moisture sensitivity thermogravimetry experiments showed that the 2A5NPLT of the 2A5NP salt family, the surfaces of the crystal plates phase is decomposed at temperatures over 195 °C following being perfectly stabilized under a dry atmosphere. the chemical reaction: Crystalline perfection C5H6N3O2+·C4H5O6-CAC5H5N3O2+C4H6O6 (2) This thermal stability is the best of all the H-bonded salts Crystal quality assessment can be performed using X-ray diVraction techniques such as X-ray diVraction topography obtained with the 2A5NP chromophore.11,12 In addition, the 2A5NPLT crystals have no mechanical cleavage plane.On the which provides a map of defects and strains in crystals. J. Mater. Chem., 1999, 9, 1091–1095 1093Fig. 5 Schematic representation of the experimental set up of the synchrotron section topography on the ID19 beamline (ESRF Grenoble). The 2A5NPLT crystal position is also specified (the c axis is vertical ).Fig. 4 Typical morphologies of 2A5NPLT crystals grown in (a) tartaric acid or (b) in formic acid aqueous solutions. Moreover, the new possibilities associated with the thirdgeneration synchrotron sources23 allow the investigation of thick and absorbing materials with short exposure times (10-2–102 s) and with high spatial resolution for diVracted images (#1 mm).Section topography consists of intersecting a narrow collimating slit (20 mm in this study) before the crystal in order to reduce the white X-ray beam extension in one direction. Thus, the image recorded onto the film located behind the sample (Lau�e transmission geometry) can be considered as a projection of the virtual slice of the crystal traversed by the beam (Fig. 5). As a small part of the sample is illuminated, the defect images are more easily distinguished than in conventional transmission topography.The section topographs were recorded at the ESRF, on the ID19 beamline. We characterized non-destructively the crystalline quality of the as-grown 2A5NPLT crystals using a beam absorber (6 mm of Al ) to avoid heat load on the samples. In this configuration, we do not observe any crystal damage or image contrast evolution.At these low X-ray wavelengths (0.1 A° <l<1 A° ) the product of the mass absorption coeYcient, m, and the crystal thickness, t, was mt<1. These section topographs allow simultaneous characterization of the growth process and the crystalline quality at a given moment of the growth. Fig. 6 shows section topographs of a 2A5NPLT crystal grown by the TL method in formic acid aqueous solutions Fig. 6 Section topographs of a 2A5NPLT crystal recorded perpendicu- (13M). On the right hand side of Fig. 6 we have given larly to the c axis. The main features of these topographs are schematic representations of the positions of the main features represented in their corresponding outline diagrams.of the corresponding topographs. These three topographs were recorded perpendicularly to the c-axis with a sample–film distance of 20 cm (Fig. 5). Between these exposures the crystal of 2A5NPLT crystals without any large defect. Nevertheless, typical defects, which are likely to be associated with solution- was moved up by 3 mm along the c-axis. Since the section topographs were recorfrom a crystal with fully developed grown crystals, namely, seed position (S), growth sector boundaries (SB) and growth bands (GB), can be observed.In morphology, the diVracted images show rather well defined crystal edges corresponding to the (1 -1 1), (1 1 1), (-1 1 Fig. 6(c) the central seed (S) can be easily detected through the growth restart interface (R) while sections 6b and 6a are -1) and (-2 -1 1) faces.The weak image contrast of all these section topographs reveals a high crystalline perfection those away from the seed. The interface between the seed 1094 J. Mater. Chem., 1999, 9, 1091–1095stability. These properties associated with a high optical eYciency and a transparency bandwidth from 410 to 1500 nm give to 2A5NPLT interesting potentiality for electooptic devices.We have optimized the synthesis of this salt and have selected the growth solution (formic acid 13M) which leads to bulky crystals. The choice of the growth method was based on the temperature-dependent solubility. The best growth results were obtained in aqueous formic acid solution (13 M) using the TL method. Synchrotron X-ray diVraction topography experiments carried out on large uncut crystals revealed a high crystalline perfection.The 2A5NPLT crystals are easily sliced and polished without any cleavage plane and are stable and suitable for optical characterizations. The unique remaining drawback is the weak growth rate (around 0.1–0.2 mm day-1) obtained in all these 2A5NPLT growths Fig. 7 Optical transmission spectrum of 2A5NPLT crystal measured in solution.For this reason we are working now on the perpendicularly to the (001) plate of 3 mm thickness. adjustment of rapid growth conditions. Acknowledgements and the crystal is not very strained, probably due to careful control of the initial growth onto the seed. The seed exhibits The authors thank R. Masse (Lab. Cristallographie, CNRSa low defect density due to the adjustment of the homogeneous Grenoble) and J.Baruchel (ESRF-Grenoble) for the critical nucleation conditions involved for the elaboration of reading of this manuscript, and P. L. Baldeck (Lab. millimetric seeds by the HTG or LT methods. Spectrome�trie Physique, UJF) for optical transmission On the other hand, the position and the nature of the measurements.growth sector boundaries display a clear insight into the growth history of the crystal. Thus, we can see in Fig. 6(c) References the disappearance of the (–111) face due to its rapid growth rate which is associated with growth bands (GB). These 1 J. Zyss, J. F. Nicoud and M. Coquillay, J. Chem. Phys., 1984, growth bands must arise from the incorporation of either 81, 4160.solvent or other impurities into the growing crystal face. We 2 I. Ledoux, J. Badan, J. Zyss, A. Migus, D. Hulin, J. Etchepare, find also GB contrasts in the (1 -1 1), (1 1 1) and (-2 -1 G. Grillon and A. Antonetti, J. Opt. Soc. Am. B, 1987, 4, 987. 3 R. Hierle, J. Badan and J. Zyss, J. Cryst. Growth, 1984, 69, 545. 1) sectors which reveal low varying growth conditions.This 4 B. Y. Shekkunov, E. A. Shepherd, J. N. Sherwood and explains the irregular growth sector boundary (SB, Fig. 6(b)) G. S. Simpson, J. Phys. Chem., 1995, 99, 7130. due to fluctuations in the relative growth rates between the (1 5 R. Masse, M. Bagieu-Beucher, J. Pe�caut, J. P. Le�vy and J. Zyss, -1 1) and (1 1 1) faces during the growth (Fig. 6(c), (b) and Nonlinear Optics, 1993, 5, 413.(a)). The other growth sector boundary, between (1 -1 1) 6 R. Masse and J. Zyss, Mol. Eng., 1991, 1, 141. and (-2 -1 1), is almost perfectly straight and weakly visible, 7 J. Pe� caut, Y. Le Fur and R. Masse, Acta Crystallogr., Sect. B, 1993, 49, 535. indicating that the relative growth rates of the two sectors are 8 J.Pe� caut, J. P. Le�vy and R. Masse, J. Mater. Chem., 1993, 3, 999.uniform. The lack of a (-1 1 -1) sector reveals a very low 9 Y. Le Fur, M. Bagieu-Beucher, R.Masse, J. F. Nicoud and growth rate of this face in constrast to the opposite one (1 J. P. Le�vy, Chem. Mater., 1995, 8, 68. -1 1) in agreement with the highly polar structure of this 10 J. Pe�caut and R. Masse, J. Mater. Chem., 1994, 4, 1851. organic salt. During the whole growth process, and starting 11 A.Ibanez, J. P. Levy, C. Mouget and E. Prieur, J. Solid State from a high quality seed, we observe a high crystalline Chem., 1997, 129, 22. 12 J. Zaccaro, B. Capelle and A. Ibanez, J. Cryst. Growth, 1997, perfection in the diVerent growth sectors. 180, 229. 13 S. K. Kurtz and T. T. Perry, J. Appl. Phys., 1968, 39, 3798. Optical transmission 14 Z.Kotler, R. Hierle, D. Josse, J. Zyss and R. Masse, J. Opt. Soc. Am. B, 1992, 9, 534. The sample was a 3 mm-thick (001) plate of a 2A5NPLT 15 N. Horiuchi, F. Lefaucheux, M. C. Robert, D. Josse and J. Zyss, crystal grown by the TL method in formic acid solution. The J. Cryst. Growth, 1995, 147, 361. absorption spectrum (Fig. 7), recorded with a Perkin-Elmer 16 J. P. Feve, B. Boulanger, I.Rousseau, G. Marnier, J. Zaccaro and l9 spectrometer, shows a lower cut-oV (defined at 50% trans- A. Ibanez, IEEE J. Quantum Electron., 1999, 35, 403. mittance) at 410–420 nm. As previously discussed for other 17 S. Khodja, D. Josse and J. Zyss, J. Opt. Soc. Am. B, 1998, 15, 751. 18 J. Zyss, R. Masse, M. Bagieu-Beucher and J. P. Le�vy, Adv. Mater., salts obtained with the 2A5NP molecule,14,24 this cut-oV is 1993, 5, 120.due to an electronic transition of the 2A5NP chromophore. 19 O.Watanabe, T. Noritake, Y. Hirose, A. Okada and T. Kurauchi, On the other hand, the absorption band in the near IR region J. Mater. Chem., 1993, 3, 1053. (1650 nm) can be attributed to the overtone of C–H vibration 20 J. Zaccaro, M. Bagieu-Beucher, J. Espeso and A. Ibanez, J.Cryst. followed by a large absorption band starting near 2000 nm Growth, 1998, 186, 224. which is typically assigned to aromatic rings (2A5NP). 21 E. Dowty, SHAPE, 521 Hidden Valley Road, Kingsport, TN 37663; Copyright 1989. 22 E. Dowty, Am. Mineral., 1980, 65, 465. Conclusions 23 R. Barret, J. Baruchel, J. Ha�rtwig and F. Zontone, J. Phys. D. Appl. Phys., 1995, 28, A250. The 2A5NPLT salt exhibits a quasi-perfect polar alignment of 24 J.Zaccaro, D. Block, M. Chamel and A. Ibanez, submitted to nonlinear chromophores in a dense and 3D-intermolecular J. Opt. Soc. Am. B. framework built up by short hydrogen bonds. This crystal structure leads to good thermal, mechanical and chemical Paper 8/10005E J. Mater. Chem., 1999, 9, 1091–1095 1095 J O U R N A L O F C H E M I S T R Y Materials Crystal growth of a stable nonlinear optical organic material: 2-amino-5-nitropyridinium monohydrogen L-tartrate Julien Zaccaro,a Fre�de�ric Lorutb and Alain Ibaneza* aLaboratoire de Cristallographie, C.N.R.S., associe� a` l’Universite� J.Fourier et a` l’Institut National Polytechnique de Grenoble, BP 166, F-38042, Grenoble Cedex 09, France bEuropean Synchrotron Radiation Facility, BP 220, F-38043, Grenoble Cedex, France Received 24th December 1998, Accepted 18th February 1999 We report on the optimization of synthesis, solubility and crystal growth from solution of 2-amino-5- nitropyridinium L-tartrate ( labelled 2A5NPLT).This newly engineered guest–host hydrogen-bonding salt exhibits a quasi-perfect polar alignment of the nonlinear chromophores in the crystalline lattice which is promising for optimized Pockels electrooptic properties. In addition, the dense and 3D-intermolecular framework built up by short hydrogen bonds leads to improved thermal, mechanical and chemical stability compared with that of organic molecular crystals.The solvent selection allowed the elaboration of bulky crystal by the temperature lowering method.Synchrotron X-ray diVraction topography experiments carried out on uncut crystals revealed a high crystalline perfection. 2A5NP+ cations are nearly aligned along the b-axis. In Introduction addition, their H-bonding anchoring onto the polyanionic In the last two decades, extensive research has shown that helical tartrate chains is very dense leading to a 3D framework organic crystals can exhibit nonlinear optical (NLO) (Fig. 1). A first NLO characterizatiociencies which are two orders of magnitude higher than bility coeYcient d33=41 pm V-1 at l=1.06 mm.19 Moreover, those of inorganic materials.1,2 This is due to highly polarizable the alignment of the dipole moments in the 2A5NPLT crystalmolecules involving conjugated systems of p electrons such as line lattice is promising for optimized Pockels electrooptic polyenes or aromatic compounds.These organic phases possess properties. Thus, in order to specify these NLO potentialities, other advantages such as unlimited molecular engineering and we have undertaken the elaboration of high quality 2A5NPLT often high laser damage thresholds.In addition, large single crystals. Our first results obtained on the synthesis optimizcrystals can be grown from solution close to room tempera- ation, solubility and crystal growth from solution are displayed ture.3,4 Nevertheless, due to their poor chemical stability and here. Then, a morphological study of 2A5NPLT is introduced. low thermal and mechanical resistance, the use of organic Finally, we have determined their transparency bandwidth, molecular crystals is currently limited in their NLO industrial the thermal and mechanical stability while the crystalline applications.An added drawback is that the polar nature of quality was defined using synchrotron X-ray topography. these phases with one- or two-dimensional structures leads to important problems in crystal growth and processing of Synthesis and solubility optical devices.In order to overcome these diYculties, a strategy has recently The 2A5NP molecule is a weak Brønsted base which can be been developed which aims to build very cohesive acentric protonated in a strong acidic medium (pH <2). This induces crystalline structures based on host–guest chemistry.5 The the dissolution of this molecule in aqueous acidic solutions by anchorage of highly polarizable chromophores onto various formation of the 2-amino-5-nitropyridinium cation (2A5NP+) inorganic or organic anionic matrices leads to short and and leads to the synthesis of hydrogen-bonded salts with the multiple hydrogen-bonded networks.Several polarizable molconjugated bases of strong or medium acids (pKa <3.3).In ecules were selected in this crystal engineering and particularly the case of tartaric acid, the 2A5NPLT phase can be directly 2-amino-5-nitropyridine (2A5NP). A great number of salts obtained by reaction (1). having a 3D acentric structure based on herringbone motifs were obtained with this push–pull molecule such as the dihy- C5H5N3O2+L-C4H6O6CAC5H6N3O2+ drogen phosphate,6 dihydrogen arsenate,7 chloride, bromide,8 +L-C4H5O6-CAC5H6N3O2+·C4H5O6- chloroacetate9 and acetophosphonate of 2-amino-5-nitropyridinium. 10 These salts possess enhanced stability (chemical, (2A5NP) (2A5NPLT) (1) thermal, mechanical ) compared to the corresponding molecu- Nevertheless, tartaric acid is a too weak Brønsted acid that lar organic crystals.Moreover, they exhibit a wider transdoes not completely dissolve the 2A5NP powder in aqueous parency range and bulky crystal morphology.11,12 The first solutions due to partial 2A5NP protonation. This explains the qualitative evaluations of the nonlinear response (powder tests low synthesis yields, around 50%. We have optimized the of second harmonic generation13) showed high eYciencies 2A5NPLT synthesis by using acetic acid as an intermediate for all these materials.These results are confirmed by the solvent. Indeed, this acid (pKa=4.76) does not react with first nonlinear characterizations carried out on single crys- 2A5NP but makes it possible, at high concentrations tals.14–16 Furthermore, optical parametric oscillation has just (13–16M), to reach low pH values (pH<1) and to completely been demonstrated with the 2-amino-5-nitropyridinium dissolve the amount of 2A5NP in reaction (1).Thus, the dihydrogen phosphate (2A5NPDP) compound.17 2A5NPLT salt is obtained by first dissolving the 2A5NP in an From this crystal engineering, another type of salt structure acetic acid solution (16M) at 60 °C. Then, a saturated tartaric was evidenced.Indeed, in the crystal structure of 2-amino-5- aqueous solution (4 M) is added. The molar ratios are 1515555 nitropyridinium L-tartrate (2A5NPLT), space group P21 with a=8.248 A° , b=9.199 A° , c=7.611 A° and b=96°.5,18 the for 2A5NP, C4H6O6, CH3COOH and H2O respectively. The J. Mater. Chem., 1999, 9, 1091–1095 1091solutions higher 2A5NPLT solubilities and solubility–temperature gradients than in tartaric solvent (Fig. 2). Despite these increases, S and DS/DT remain rather low for the use of the typical temperature lowering (TL) method. Thus, in order to adjust the 2A5NPLT solubility, another solvent was used: formic acid (pKa=3.75) which does not react with 2A5NP. Indeed, as for acetic acid, low pH solutions can be achieved at high concentrations (HCOOH 10–20 M).In these formic acid solutions we have obtained an important increase of S and DS/DT compared to tartaric and acetic solvents (Fig. 2). These significant evolutions allow a great flexibility for selecting the growth technique and optimizing crystal growth conditions. Crystal growth and morphology The first growths of 2A5NPLT crystals were carried out in tartaric acid solutions (3M).The low DS/DT value for this solvent prompted us to select thermal gradient methods. A horizontal temperature-gradient technique (HTG) was used to obtain high-quality seeds by spontaneous nucleation. Nevertheless, this growth reactor, described elsewhere,12 does not allow the growth of large single crystals. In addition, the resulting crystals exhibit a lot of growth bands evidenced by Fig. 1 Projection of the polar structure of 2A5NPLT in the ab plane X-ray diVraction topography12 because the seeds and the showing a quasi-perfect alignment of the 2-amino-5-nitropyridinium cations. growth solution are not stirred. Thus, in order to obtain large crystals of high quality we then used a vertical temperaturegradient method (VTG).20 The nutrient comprises millimetric salt crystallizes at room temperature as a white microcrystalline 2A5NPLT crystals placed in the upper zone of a container of powder on addition of acetone.Thus, the synthesis yield is controlled porosity. The thermal configuration, with a higher close to 100%. In order to increase the crystal quality, we take temperature in the upper zone (DT#3–4 °C), reduces the particular care in ensuring the purity of the starting materials.natural convection. The seeds, previously grown by the HTG The commercial 2A5NP is first purified by double sublimation method, are stacked to silica suspensions in the lower zone and then dissolved in high purity acidic solutions. Finally, the and are rotated at a rate of 20 rpm with inversion of the salts are purified again from CH3COOH solutions, by rotation direction every 40 s.Thus, after 3–4 months of recrystallization in acetone. growth, large and optically clear crystals are obtained Due to the thermal decomposition of 2A5NPLT (see [Fig. 3(a)]. Unfortunately, these 2A5NPLT crystals have weak Stability section), we have undertaken the crystal growth of growth rates, around 0.1–0.2 mm day-1, and a plate mor- this phase from acidic solutions.We first selected as solvent phology. This crystal morphology was determined from X-ray the acid corresponding to the salt.We determined the solubility diVraction (Lau�e diagrams) and interfacial angle measure- curves between 20 and 60 °C for various molar concentrations ments using a two-circle optical goniometer (Nedinsco) of tartaric acid (2–4 M) which satisfy the pH conditions (pH [Fig. 4(a)]. The crystal morphology has been drawn with the <2). Unfortunately, the magnitude of the 2A5NPLT solubility program SHAPE,21,22 using the experimental growth rates as (S) and its variation with temperature (solubility–temperature the face-crystal center in the WulV plot. The 2A5NPLT plate gradient DS/DT) in these concentrated aqueous solutions are habit is mainly due to a very low growth rate of the {101} always very low (Fig. 2). form. In contrast to her salts of this family,11,12 the On the other hand, we have seen for other salts of this 2A5NPLT phase grows in a number of well faceted forms of family that the solubility curve is directly related to the pH of low Miller’s indexes [Fig. 4(a)].the growth solution; the lower the pH value, the higher are Then, in order to obtain bulky crystals, we used another the solubility and the solubility–temperature gradient.11,12 As growth solvent. As we also obtain platelet crystals with acetic previously mentioned, acetic acid aqueous solutions (13–16M) acid, we selected formic acid solutions which lead to bulky allow low pH values to be achieved without combining with crystals.Moreover, the higher solubility and solubility– the 2A5NP molecule. Thus, we registered in these acetic temperature gradient values registered for this solvent allow us to carry out the 2A5NPLT crystal growth by the typical temperature lowering (TL) method. For a good control of crystal growth, the temperature lowering rates have to be less than 0.2 °C per day at low relative supersaturations around 0.01.This requires a high temperature precision (±0.002 °C) to avoid undesirable supersaturation changes leading to secondary nucleation, macro steps and multiple solvent inclusions on the growing crystal faces. These growth problems are favored in this case by a narrow metastable zone. The growth solution (around 200 cm3) is seeded with 2A5NPLT crystal about 1 mm3 in size, obtained by spontaneous nucleation in the same TL crystallizer previously described.11 The growing crystals are rotated (20–30 rpm) and are reversed every 45 s.The temperature is slowly decreased in the range 45–35 °C. This second set of growth experiments allowed us to obtain in 2–3 months bulky and optically clear crystals [Fig. 3(b)]. Fig. 2 Solubility curves of the 2A5NPLT salt in tartaric (3 M), acetic (16 M) and formic acid (13 M) aqueous solutions. This bulky morphology is essentially achieved through a 1092 J. Mater. Chem., 1999, 9, 1091–1095Fig. 3 2A5NPLT crystals grown (a) by the HTG method in tartaric acid solutions (3 M), (b) by the LT technique in formic acid (13 M) aqueous solutions.significant increase of the {101} growth rate [Fig. 4(b)]. other hand, hardness measurements were carried out using a Leitz Vickers hardness tester. Indentations were made on Unfortunately, the average growth rates remain very weak ( less than 0.2 mm day-1). This is certainly due to the adsorp- (101), (001), (111), (1-11) and (1-1-1) growth faces with a dwelling time kept constant at 15 s and weights ranging from tion of polar solvent molecules on the crystal faces of this highly polar material. 15–25 g. In agreement with the 3D crystal structure, all these experiments gave the same average value of 105±5 kgmm-2. The 2A5NPLT hardness is the highest of the 2A5NP salt Crystal characterizations family and is rather close to that of KDP (135 kg mm-2).These good mechanical properties make slicing and polishing Stability easier for the optical device processing of 2A5NPLT crystals. A diVerential scanning calorimetry study associated with Finally, this tartrate possesses the lowest moisture sensitivity thermogravimetry experiments showed that the 2A5NPLT of the 2A5NP salt family, the surfaces of the crystal plates phase is decomposed at temperatures over 195 °C following being perfectly stabilized under a dry atmosphere.the chemical reaction: Crystalline perfection C5H6N3O2+·C4H5O6-CAC5H5N3O2+C4H6O6 (2) This thermal stability is the best of all the H-bonded salts Crystal quality assessment can be performed using X-ray diVraction techniques such as X-ray diVraction topography obtained with the 2A5NP chromophore.11,12 In addition, the 2A5NPLT crystals have no mechanical cleavage plane.On the which provides a map of defects and strains in crystals. J. Mater. Chem., 1999, 9, 1091–1095 1093Fig. 5 Schematic representation of the experimental set up of the synchrotron section topography on the ID19 beamline (ESRF Grenoble). The 2A5NPLT crystal position is also specified (the c axis is vertical ).Fig. 4 Typical morphologies of 2A5NPLT crystals grown in (a) tartaric acid or (b) in formic acid aqueous solutions. Moreover, the new possibilities associated with the thirdgeneration synchrotron sources23 allow the investigation of thick and absorbing materials with short exposure times (10-2–102 s) and with high spatial resolution for diVracted images (#1 mm).Section topography consists of intersecting a narrow collimating slit (20 mm in this study) before the crystal in order to reduce the white X-ray beam extension in one direction. Thus, the image recorded onto the film located behind the sample (Lau�e transmission geometry) can be considered as a projection of the virtual slice of the crystal traversed by the beam (Fig. 5). As a small part of the sample is illuminated, the defect images are more easily distinguished than in conventional transmission topography.The section topographs were recorded at the ESRF, on the ID19 beamline. We characterized non-destructively the crystalline quality of the as-grown 2A5NPLT crystals using a beam absorber (6 mm of Al ) to avoid heat load on the samples. In this configuration, we do not observe any crystal damage or image contrast evolution.At these low X-ray wavelengths (0.1 A° <l<1 A° ) the product of the mass absorption coeYcient, m, and the crystal thickness, t, was mt<1. These section topographs allow simultaneous characterization of the growth process and the crystalline quality at a given moment of the growth. Fig. 6 shows section topographs of a 2A5NPLT crystal grown by the TL method in formic acid aqueous solutions Fig. 6 Section topographs of a 2A5NPLT crystal recorded perpendicu- (13M). On the right hand side of Fig. 6 we have given larly to the c axis. The main features of these topographs are schematic representations of the positions of the main features represented in their corresponding outline diagrams.of the corresponding topographs. These three topographs were recorded perpendicularly to the c-axis with a sample–film distance of 20 cm (Fig. 5). Between these exposures the crystal of 2A5NPLT crystals without any large defect. Nevertheless, typical defects, which are likely to be associated with solution- was moved up by 3 mm along the c-axis. Since the section topographs were recorded from a crystal with fully developed grown crystals, namely, seed position (S), growth sector boundaries (SB) and growth bands (GB), can be observed.In morphology, the diVracted images show rather well defined crystal edges corresponding to the (1 -1 1), (1 1 1), (-1 1 Fig. 6(c) the central seed (S) can be easily detected through the growth restart interface (R) while sections 6b and 6a are -1) and (-2 -1 1) faces.The weak image contrast of all these section topographs reveals a high crystalline perfection those away from the seed. The interface between the seed 1094 J. Mater. Chem., 1999, 9, 1091–1095stability. These properties associated with a high optical eYciency and a transparency bandwidth from 410 to 1500 nm give to 2A5NPLT interesting potentiality for electooptic devices.We have optimized the synthesis of this salt and have selected the growth solution (formic acid 13M) which leads to bulky crystals. The choice of the growth method was based on the temperature-dependent solubility. The best growth results were obtained in aqueous formic acid solution (13 M) using the TL method. Synchrotron X-ray diVraction topography experiments carried out on large uncut crystals revealed a high crystalline perfection.The 2A5NPLT crystals are easily sliced and polished without any cleavage plane and are stable and suitable for optical characterizations. The unique remaining drawback is the weak growth rate (around 0.1–0.2 mm day-1) obtained in all these 2A5NPLT growths Fig. 7 Optical transmission spectrum of 2A5NPLT crystal measured in solution.For this reason we are working now on the perpendicularly to the (001) plate of 3 mm thickness. adjustment of rapid growth nditions. Acknowledgements and the crystal is not very strained, probably due to careful control of the initial growth onto the seed. The seed exhibits The authors thank R. Masse (Lab. Cristallographie, CNRSa low defect density due to the adjustment of the homogeneous Grenoble) and J.Baruchel (ESRF-Grenoble) for the critical nucleation conditions involved for the elaboration of reading of this manuscript, and P. L. Baldeck (Lab. millimetric seeds by the HTG or LT methods. Spectrome�trie Physique, UJF) for optical transmission On the other hand, the position and the nature of the measurements.growth sector boundaries display a clear insight into the growth history of the crystal. Thus, we can see in Fig. 6(c) References the disappearance of the (–111) face due to its rapid growth rate which is associated with growth bands (GB). These 1 J. Zyss, J. F. Nicoud and M. Coquillay, J. Chem. Phys., 1984, growth bands must arise from the incorporation of either 81, 4160.solvent or other impurities into the growing crystal face. We 2 I. Ledoux, J. Badan, J. Zyss, A. Migus, D. Hulin, J. Etchepare, find also GB contrasts in the (1 -1 1), (1 1 1) and (-2 -1 G. Grillon and A. Antonetti, J. Opt. Soc. Am. B, 1987, 4, 987. 3 R. Hierle, J. Badan and J. Zyss, J. Cryst. Growth, 1984, 69, 545. 1) sectors which reveal low varying growth conditions. This 4 B.Y. Shekkunov, E. A. Shepherd, J. N. Sherwood and explains the irregular growth sector boundary (SB, Fig. 6(b)) G. S. Simpson, J. Phys. Chem., 1995, 99, 7130. due to fluctuations in the relative growth rates between the (1 5 R. Masse, M. Bagieu-Beucher, J. Pe�caut, J. P. Le�vy and J. Zyss, -1 1) and (1 1 1) faces during the growth (Fig. 6(c), (b) and Nonlinear Optics, 1993, 5, 413.(a)). The other growth sector boundary, between (1 -1 1) 6 R. Masse and J. Zyss, Mol. Eng., 1991, 1, 141. and (-2 -1 1), is almost perfectly straight and weakly visible, 7 J. Pe� caut, Y. Le Fur and R. Masse, Acta Crystallogr., Sect. B, 1993, 49, 535. indicating that the relative growth rates of the two sectors are 8 J.Pe� caut, J. P. Le�vy and R. Masse, J. Mater. Chem., 1993, 3, 999.uniform. The lack of a (-1 1 -1) sector reveals a very low 9 Y. Le Fur, M. Bagieu-Beucher, R.Masse, J. F. Nicoud and growth rate of this face in constrast to the opposite one (1 J. P. Le�vy, Chem. Mater., 1995, 8, 68. -1 1) in agreement with the highly polar structure of this 10 J. Pe�caut and R. Masse, J. Mater. Chem., 1994, 4, 1851. organic salt. During the whole growth process, and starting 11 A. Ibanez, J. P. Levy, C. Mouget and E. Prieur, J. Solid State from a high quality seed, we observe a high crystalline Chem., 1997, 129, 22. 12 J. Zaccaro, B. Capelle and A. Ibanez, J. Cryst. Growth, 1997, perfection in the diVerent growth sectors. 180, 229. 13 S. K. Kurtz and T. T. Perry, J. Appl. Phys., 1968, 39, 3798. Optical transmission 14 Z. Kotler, R. Hierle, D. Josse, J. Zyss and R. Masse, J. Opt. Soc. Am. B, 1992, 9, 534. The sample was a 3 mm-thick (001) plate of a 2A5NPLT 15 N. Horiuchi, F. Lefaucheux, M. C. Robert, D. Josse and J. Zyss, crystal grown by the TL method in formic acid solution. The J. Cryst. Growth, 1995, 147, 361. absorption spectrum (Fig. 7), recorded with a Perkin-Elmer 16 J. P. Feve, B. Boulanger, I. Rousseau, G. Marnier, J. Zaccaro and l9 spectrometer, shows a lower cut-oV (defined at 50% trans- A. Ibanez, IEEE J. Quantum Electron., 1999, 35, 403. mittance) at 410–420 nm. As previously discussed for other 17 S. Khodja, D. Josse and J. Zyss, J. Opt. Soc. Am. B, 1998, 15, 751. 18 J. Zyss, R. Masse, M. Bagieu-Beucher and J. P. Le�vy, Adv. Mater., salts obtained with the 2A5NP molecule,14,24 this cut-oV is 1993, 5, 120. due to an electronic transition of the 2A5NP chromophore. 19 O.Watanabe, T. Noritake, Y. Hirose, A. Okada and T. Kurauchi, On the other hand, the absorption band in the near IR region J. Mater. Chem., 1993, 3, 1053. (1650 nm) can be attributed to the overtone of C–H vibration 20 J. Zaccaro, M. Bagieu-Beucher, J. Espeso and A. Ibanez, J. Cryst. followed by a large absorption band starting near 2000 nm Growth, 1998, 186, 224. which is typically assigned to aromatic rings (2A5NP). 21 E. Dowty, SHAPE, 521 Hidden Valley Road, Kingsport, TN 37663; Copyright 1989. 22 E. Dowty, Am. Mineral., 1980, 65, 465. Conclusions 23 R. Barret, J. Baruchel, J. Ha�rtwig and F. Zontone, J. Phys. D. Appl. Phys., 1995, 28, A250. The 2A5NPLT salt exhibits a quasi-perfect polar alignment of 24 J. Zaccaro, D. Block, M. Chamel and A. Ibanez, submitted to nonlinear chromophores in a dense and 3D-intermolecular J. Opt. Soc. Am. B. framework built up by short hydrogen bonds. This crystal structure leads to good thermal, mechanical and chemical Paper 8/10005E J. Mater. Chem., 1999, 9,

ISSN:0959-9428    

DOI:10.1039/a810005e

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Self-assembling monolayers of helical oligopeptides on gold with applications in molecular electronics† Andrew E. Strong and Barry D. Moore* Department of Pure and Applied Chemistry, University of Strathclyde, Thomas Graham Building, 295 Cathedral Street, Glasgow, UK G1 1XL. E-mail: b.d.moore@strath.ac.uk Received 3rd September 1998, Accepted 1st March 1999 A generic strategy is outlined for preparing a chiral functionalised surface using self-assembled monolayers (SAMs) of helical oligopeptides for intended application in molecular electronics, sensors and catalysis.Helical peptides have been designed and synthesised with 3 Met or 2 Cys residues per peptide, placed in the sequence so that in a helical conformation their side-chains are on the same side of the helix and organised to bind to gold by selfassembly of the thioether or thiol functions.Circular dichroism (CD) of the Met-containing peptides showed that they are mainly helical in organic solvents and FT-RAIRS (reflection–absorption infrared spectroscopy) of their SAMs on gold confirmed that they are aligned with the helix axes parallel to the surface. However complete coverage of the surface is dependent critically on the preparation of the gold.CD of an oligopeptide containing 2 Cys residues suggested that the conformation was a mixture of helix and random coil. Films of this peptide adsorbed from diVerent solvents were characterised by ellipsometry and cyclic voltammetry. Ethanol and methanol led to multilayer formation indicating protic solvents destabilise the helical structure, allowing formation of intermolecular disulfide bonds.By contrast the aprotic solvent acetonitrile led to formation of a close-packed selfassembled monolayer suitable for elaboration into a functionalised supramolecular architecture. As the size of electronic components decreases, approaches nents it will be necessary to specifically link the peptides in have been made to design and prepare molecules that possess solution prior to the self-assembly step.Methods are available properties that are useful in electronic circuits.1 These eVorts to link the peptides through either covalent or non-covalent may be split broadly into two areas,2 molecular materials for interactions. Possibilities include forming a peptide bond electronics, which use the macroscopic properties of the mate- between the side-chains bearing amine and acid groups (e.g.rials, and electronics at the molecular level which exploits the Lys and Glu) or using amino acids bearing complementary microscopic properties of individual molecules.3 With respect hydrogen bonding arrays such as those found in nucleic acids.to the second area, various ingenious ideas have been proposed The supramolecular structure will ultimately be confined to a for using molecules as a rectifier4 or a shift register5 and some surface by sulfur-mediated self-assembly on gold. The overall molecules have been demonstrated to act as switches6–13 and strategy is summarised in Fig. 1. storage devices.14,15 However most of these have been in Our approach relies on the fundamentals of self-organisation solution.In a functioning computer it is essential that the prevalent in Nature, which uses information ‘encoded’ in the memory, input, output and processing units are at well-defined structures of simple molecules to direct the formation of much positions to enable them to communicate with each other.larger architectures with known conformation, e.g. proteins, This indicates a solid state device. In this paper we present a nucleic acids or cells. Thus choice of peptide sequence ‘selfplan to use self-assembled monolayers (SAMs) of helical organises’ the molecules into a helix; specific linking of helices oligopeptides to produce a surface on which various functional in solution may be performed using complementary molecular groups may be positioned with nanoscale resolution.We recognition, and the final step is self-assembly of the supramolenvisage that this may also be of utility in fields such as ecular architecture onto a surface using the specificity of the sensors, biomaterials and catalysis. sulfur–gold interaction. Our strategy is to prepare helical oligopeptides that contain Self-assembling monolayers of sulfur-containing molecules both attachment sites for molecular computing elements and on gold are attractive systems for tailoring surface properties sulfur containing side-chains to promote self-assembly on gold. and the study of interfacial phenomena, e.g.wetting and By using solid-phase peptide synthesis these diVerent groups adhesion,16–19 and fundamentals of electron transfer.20–22 Most can be introduced into the sequence at predefined sites relative eVorts have concentrated on the formation, structure and to each other.In an ideal a-helix one turn around the helix properties of monolayers of long-chain alkanethiols.23,24 These axis encompasses 3.6 residues. Thus, placement of three S- systems have the advantages16 that a strong, specific interaction containing residues at positions i, i+4 and i+7 will ensure between the adsorbate and the substrate drives spontaneous their side-chains all extend from the same side of the helix, formation of a monolayer, the films are densely packed and preorganised to bind to a gold surface.Including the residue physically robust and a variety of other functional groups may modified with the molecular computing element at i+2 will be accommodated within the monolayer. Examples of other result in this group lying on the opposite side of the helix so functional groups include alkanethiols with terminal ferrothat following self-assembly it will be at the ambient/film cenylazobenzenes,25 tetrathiafulvalene,26 viologens,27 biotin28 interface.To achieve lateral positioning of diVerent compo- and thiol-terminated oligoimides29 and oligo(phenyleneethynylene) s.30 The potential of exploiting these structures in †Peptide purification and characterisation details are available as sensors31–34 and biomaterials18 has been recognised. supplementary data (SUPPL. NO. 57510, pp. 9) from the British Terminally-modified alkanethiols and other molecules where Library. For details of the Supplementary Publications Scheme, see the long molecular axis extends out from the metal surface ‘Information for Authors’, J. Mater. Chem., available via the RSC still present a homogeneous surface to the organic/ambient web page (http://www.rsc.org/authors). For direct electronic access see http://www.rsc.org/suppdata/jm/1999/1097/. interface.Although the mole fraction composition of mixed J. Mater. Chem., 1999, 9, 1097–1105 1097of the helix and formed stable monolayers on gold with the helix axis parallel to the surface as characterised by electrochemistry and reflection–absorption infrared spectroscopy (RAIRS). In this full paper we describe the design and preparation of SAMs of other helical oligopeptides containing several Met or Cys residues and their characterisation by cyclic voltammetry (CV), RAIRS and ellipsometry. The conditions required for monolayer formation of each type of peptide through the side-chain sulfur functions are discussed.Results and discussion General design principles Helical peptides are attractive building blocks with which to construct a functionalised surface as their synthesis and structure have been well studied.The more common a-helix is stabilised by hydrogen bonds between the main chain N–H of the (i+4)th and the main chain CLO of the ith amino acid in the sequence and one turn around the helix axis encompasses 3.6 residues. An alternative known helical geometry is the 310- helix44,45 so called because 3 residues form one complete turn around the helix axis and there are 10 atoms within the intramolecular ring formed by the hydrogen bond between N–H of residue i+3 and the CLO of residue i.The diVerences in pitch and hydrogen bonding mean that a 310-helix is longer and thinner than an a-helix of the same number of residues. The conformation of a peptide is determined by its amino acid sequence.In particular, peptides consisting entirely of amino acids with hydrocarbon side-chains are known to adopt helical Fig. 1 Proposed preparation of an ordered, functionalised surface structures in organic solvents.44,45 The 20 a-amino acids with nanometre scale features using helical oligopeptides. The present in mammalian proteins oVer a variety of side-chain schematic is viewed along the helix axes of peptides 1, 2 and 3 which functions, some of which (e.g.Lys, which contains the NH2 have side-chains modified to carry functional elements A, B and C group, and the CO2H of Glu) may be readily modified to respectively on the opposite sides of the helices to thiol or thioether introduce other desired functions.Two amino acids, Cys and groups. The peptides are specifically linked in solution through complementary binding groups to form discrete supramolecular Met contain sulfur functions in their side-chains (-CH2SH assemblies. These are then locked together and immobilised via and -CH2CH2SCH3 respectively) which are expected to form formation of a self-assembled monolayer on gold using the thiols or strong bonds with a gold surface via self-assembly.Of these thioethers. The resultant surface confined molecular architecture will two Met was initially the more attractive. The greater separa- contain functional elements A, B and C placed at known relative tion between the peptide backbone and the sulfur atom in positions a few nm apart. This plan could be extended to include Met compared with Cys should allow the sulfurs to find more functional elements by increasing the number of diVerent peptides or by including several elements on each peptide chain.optimal positions on the gold surface. It should also reduce unfavourable steric interactions between the metal surface and other side-chains on that side of the helix.Furthermore, monolayers of alkanethiols co-adsorbed with terminally funcalthough the thiol of Cys binds more strongly to gold than tionalised thiols may be controlled empirically, the relative the thioether of Met, it was anticipated that the biomimetic positioning of the two components is still random.35 route of multiple weak interactions might be a more eYcient In the few previous examples of SAMs of peptides that have route to well ordered systems.In this paper we compare the recently been reported, the peptides generally had only one formation and structure of SAMs of peptides containing both attachment point to the gold surface, either through the sidetypes of residue. chain of a terminal Cys residue34,36–38 or via a thioalkyl carboxylic acid coupled to the N-terminal.33,39–41 Outside our Met-containing peptides own work, one notable exception was the recent report from the Mutter and Vogel laboratories,42 which described the Our prototype sequence was that of peptide 1, Fcb-13Met3.Our aim of adsorbing the peptides to gold in a helical formation and characterisation of a SAM of a cyclic peptide template with 2 or 4 pendant alkylthiolates covalently bound conformation via the thioethers placed certain restrictions on the primary structure.In particular, alignment of the surface- to side-chains of the template. The solution side of the template was derivatised with metal-binding residues to which the binding side-chains along the same side of the a-helix such that they could bind cooperatively to a gold surface (Fig. 2) binding of metals and subsequent ligands was demonstrated. Although it was not discussed in their paper this system could required Met to be included at positions 4, 8 and 11. The designed formation of three sulfur–gold bonds was expected also be extended to produce a functionalised surface similar to that we have proposed. to introduce an element of selectivity as helical conformations should be bound more strongly than other conformations As our primary objective we have concentrated on investigating the formation and structure of SAMs of monomeric where only one or two peptide–surface interactions are possible.The remaining amino acids used, Ala and Aib (a- helical oligopeptides aligned with the helix axis parallel to the surface. We have previously described the formation and aminoisobutyric acid) are known strong helix formers in organic solvents44,45 and have small methyl side-chains to characterisation of SAMs of the helical oligopeptide Fcb- 13Met3 (1), Fcb-Ala-Aib-Ala-Met-Aib-Ala-Ala-Met-Ala-Aib- minimise unfavourable steric interactions with the surface.The N-terminus was acylated with ferrocenylbutyric acid to pro- Met-Ala-Ala-NH2 (where Fcb is ferrocenylbutyric acid).43 This 13 residue peptide contained three Met residues (hence vide a convenient electrochemical label.Peptide 2, Fcp-12Leu2, Fcp-Ala-Aib-Ala-Ala-Leu-Aib-Ala-Ala-Ala-Aib-Leu-Ala- 13Met3) positioned so their side-chains aligned along one side 1098 J. Mater. Chem., 1999, 9, 1097–1105parison of the adsorption of peptide 3 without the goldbinding thioether groups.Assembly of each sequence on the solid phase proceeded smoothly and small portions of each Fmoc-peptide-resin (Fmoc=fluorenylmethoxycarbonyl ) were deprotected and acylated as needed to provide the N-acetyl (Ac), N-(4-ferrocenylbutanoyl ) (Fcb) or N-p-nitrobenzoyl (NB) derivatives. As with peptides 1 and 2 cleavage of the ferrocene derivatives was carried out under oxygen-free conditions to minimise the oxidation of the ferrocene to ferrocenium.The Met-containing peptide contained small quantities of the sulfoxide. After purification by reversed-phase HPLC, derivatives of both peptides were only sparingly soluble in a range of solvents. Amphipathic helical peptides are known to aggregate in water to form bundles with the hydrophobic surfaces shielding each other from the solvent.47–49 These completely non-polar sequences may have aggregated in a similar manner even in these less polar solvents (methanol and ethanol; acetonitrile and ethyl acetate were also tried but no dissolution could be detected). Formation of b-sheets, which are often believed to aggregate and cause precipitation, was considered unlikely in peptides containing several Aib residues.44 The circular dichroic (CD) spectra were measured using the Fig. 2 Space-filling model of 1, 13Met3 in an a-helical conformation showing the alignment of the 3 S atoms along the underside of the NB derivatives, since here the peptide concentration could be helix. The main axis of the helix is normal to the plane of the page.determined accurately from the UV-absorbance of the NB The N-terminal group here, acetyl, is shown instead of the actual one group. Each spectrum (Fig. 3) showed the double-minima at (4-ferrocenylbutyric acid) for clarity. The atom colours are C: green, 208 and 222 nm typical of a-helices.50 The value of [h]222 of H: white, N: blue, O: red, S: yellow. -23 300 for NB-16Met3 demonstrated the expected increase in helicity over Fcb-13Met3 (for which [h]222 was -18 300).Films of Ac-16Met3, Ac-3, and Ac-16Nle3, Ac-4, were prepared on evaporated gold slides by immersion in saturated NH2 (where Fcp is ferrocenylpropionic acid), available from previous studies (J. Reid, D. Nonhebel and B. D. Moore, peptide solutions for 2 days. ‘Cast’ films of each peptide were also prepared by pipetting their solutions onto slides and unpublished results) provided a convenient, non-sulfurcontaining control of similar length, composition and second- allowing the solvent to evaporate.The reflectance spectra of the cast films were taken as being representative of disordered ary structure43 to 1. As described previously43 both peptides 1 and 2 were films (Fig. 4). The amide I bands at 1663 cm-1 and 1667 cm-1 for Ac-3 and Ac-4 respectively are in the region more typical synthesised eYciently by conventional solid phase methods. The circular dichroic spectra of Fcb-13Met3 confirmed that it of 310-helices51 but may also indicate a-helices. The latter are characterised by an amide I maximum at 1650–1658 cm-1, was mainly helical in both methanol and acetonitrile. SAMs of 1 were prepared on a gold disk electrode and were found but this may be shifted to higher frequency by ca. 10 cm-1 after adsorption on gold.38 As expected no bands were detected to be remarkably stable to repetitive cycling in ethanol–electrolyte. In contrast the initial surface coverage of peptide 2 was in the reflectance spectrum of the slide immersed in Ac-4.Gratifyingly in the surface spectrum of Ac-3 an amide I band typically less than 15% of peptide 1 and rapidly decreased during cycling under the same conditions. In the RAIRS spectrum of a SAM of the Met-containing peptide on evaporated gold the amide I band was at 1664 cm-1, indicative of a 310-helix, and the amide II and amide III bands were at 1545 cm-1 and 1264 cm-1 respectively.Comparison of the reflectance with the transmission spectrum measured of 1 dispersed in KBr confirmed that the helix axis was parallel to the surface, the orientation expected for a self-assembled layer. In contrast the RAIRS of Fcp-12Leu2 suggested no particular orientation of the peptide at the surface. Following these promising results we sought to design completely a-helical Met-containing peptides and to prepare SAMs with greater coverage of the surface.To increase the ahelicity the length of peptide 3, 16Met3, Ala-Ala-Aib-Ala- Met-Ala-Phe-Ala-Met-Aib-Phe-Met-Aib-Ala-Ala-Ala-NH2, was increased to 16 residues as longer peptides favour the aover the 310-helical conformation.44 Two Phe residues (sidechain CH2Ph) were placed at positions 7 and 11 in the sequence where they could potentially stabilise the a-helical form by solvophobic packing of the side-chain in polar solvents such as acetonitrile and methanol.Although in proteins it is more usually found in b-sheets Phe has been used in organic-soluble helical peptides.44,46 The sequence of peptide 4, 16Nle3, Ala- Ala-Aib-Ala-Nle-Ala-Phe-Ala-Nle-Aib-Phe-Nle-Aib-Ala-Ala- Ala-NH2, diVered from that of 3 only by replacement of the Fig. 3 Circular dichroic (CD) spectra of peptides NB-3 (continuous three Met residues with the isosteric Nle (side-chain line, 130 mM) and NB-4 (dashed line, 180 mM) in methanol at room temperature. -CH2CH2CH2CH3) and was chosen to provide a close com- J. Mater. Chem., 1999, 9, 1097–1105 1099what is found for the SAM of Ac-3 and supports the assignment of the amide I band at 1667 cm-1 to a helical conformation.Similar changes were observed by Boncheva and Vogel who prepared monolayers of helical peptides by the Langmuir–Blodgett and self-assembly techniques, employing a Cys at either the N- or C-terminus.38 Their more detailed theoretical analysis reached the same conclusions as ours and was supported by surface plasmon resonance measurements. The surface coverage and stability of the adsorbed peptide films were assessed by cyclic voltammetry of the ferrocene derivatives.In addition we investigated the ability of a SAM of Ac-16Met3 to block electron transport to a couple in solution. The working electrode used in this experiment was an evaporated gold slide which had been immersed in peptide solution (0.04 mM in methanol for 17 h) and rinsed.Voltammetry of the modified slide in a solution of Ru(NH3)6Cl2 showed no significant diVerence to that recorded at bare gold. Although electron transfer to this complex has previously been shown to be blocked by alkanethiol SAMs these present a much denser, thicker barrier than a closepacked monolayer of helical peptide (C21H43SH forms a monolayer approx. 30 A° thick compared to the peptide diameter of 12 A° ).54 The voltammograms of the much larger couple 5 [Cr(Cl4C6O2)3]3- (chromium tris(tetrachlorocatecholate) 33-) at a bare gold slide and at the peptide-modified electrode are shown in Fig. 5. Clearly the peptide monolayer strongly hindered electron transfer to and from this bulky Cr complex indicating that a coherent monolayer had adsorbed over the whole electrode.Following this encouraging observation we investigated the binding of 3, Fcb-16Met3, to a gold disk electrode. However despite intensive eVorts using a wide variety of surface preparations, solvents and immersion times the measured surface coverages obtained were no better than that of peptide 1. Fig. 4 FT-RAIRS (displayed as % transmission spectra) of (a) Ac- 16Met3 self-assembled from a 1 mM solution in methanol; (b) cast film of Ac-16Met3; (c) cast film of Ac-16Nle3. The x-axes units are cm-1. appeared at 1667 cm-1. Furthermore its intensity was significantly less than that of the amide II band. The RAIRS technique allows the orientation of the peptides to be determined as only vibrations with a component normal to the surface are enhanced.52 The main component of the amide I band is stretching of the carbonyl bonds and in a helical peptide these will be aligned along the helix axis.The main contributions to the amide II band are in-plane bending of NKH and CLO, and CKC and CKN stretching.53 For each of these vibrations the transition dipole moment subtends a range of angles to the helical axis.In the self-assembled peptide monolayers binding via the methionine side-chains is expected to align the helical axis parallel to the surface. This is expected Fig. 5 Cyclic voltammograms (CVs) of Cr(Cl4C6O2)33- at (dashed to lead to a dramatic reduction in the intensity of the RAIRS line) a bare gold electrode and (continuous line) gold covered with a amide I band relative to the amide II compared to the spectrum SAM of Ac-16Met3.CVs were recorded at 50 mV s-1 in saturated KCl in 951 methanol–water. The reference electrode was Ag/AgCl. of the cast film. It can be seen in Fig. 4 that this is exactly 1100 J. Mater. Chem., 1999, 9, 1097–1105Other groups have previously found that thioether the peptide’s orientation on the surface.The four other residues on the side of the helix intended to bind to gold were Ala, monolayers failed to completely displace contaminants from which should oVer the least unfavourable steric interactions the gold surface55 and that the preparation on gold of SAMs with the surface. of resorcin[4]arenes through 4 pendant dialkyl thioether arms Using Cys complicates the synthesis as the strongly has been reported to be of low reproducibility due to the nucleophilic thiol must be protected.Other potential diYculties critical conditions required for the electrode preparation and may occur during the subsequent removal of the protecting for self-assembly.31 In contrast the more frequently used group and manipulation of the peptide-dithiol, which may thiolates have been reported to bind strongly to gold and are oxidise to the disulfide.The shorter spacer group between the able to displace contaminants from the surface.55 In model peptide backbone and the sulfur (one methylene group rather experiments with Fcb-derivatives of Met and cystine esters we than two in Met) may also introduce diYculties during the too observed significantly stronger binding of the thiolate over self-assembly step.Unfavourable steric interactions between the thioether.56 We thus decided to investigate the possibility the peptide and surface could prevent the two sulfur atoms of generating SAMs from an oligopeptide containing two Cys from simultaneously accessing ideal binding positions on the residues as surface binding groups, the expectation being that gold surface.we might more readily prepare films with high stability and The peptide was prepared on the solid phase. Both cysteine surface coverage than those prepared from peptides 1 and 3. residues were protected with the trityl group (Trt, triphenylmethyl )57 which may be cleaved either by TFA (tri- Cys-containing peptide fluoroacetic acid) to aVord the thiol or by iodine to form the The sequence of oligopeptide 6, 16Cys2, Ala-Ala-Aib-Ala- disulfide.These methods avoid the use of toxic reagents (HgII, Phe-Ala-Cys-Phe-Leu-Cys-Aib-(NO2)Phe-Ala-Aib-Leu-Ala- TlIII) or thiols; we were particularly keen to avoid the latter OH, where (NO2)Phe is p-nitrophenylalanine, was based on as even small quantities could compete with the peptide for a peptide, PRM1 (Ac-Ala-Ala-Aib-Ala-Phe-Ala-Acc-Leu-Aib- binding to gold.The peptide-resin was stored Fmoc- and Ala-Don-Ala-Aib-Leu-Ala-NH2, where Don and Acc are Glu ditrityl-protected (Fmoc-16Cys2{Trt}2-OR) and subsequent residues modified with electron donating and accepting groups, modifications were carried out as required. From herein we respectively) which was characterised as forming a 310-helix in will refer to the peptide dithiol as 16Cys2(SH)2 and the acetonitrile.10 In contrast to peptides 1 and 3, with peptide 6 peptide disulfide as 16Cys2(S2).molecular modelling (Fig. 6) indicated that binding of the two Cleavage was attempted first with 95% aq. TFA without Cys side-chains to gold would be more favourable in a 310- protection from the atmosphere with triisopropylsilane (TIPS) helix rather than an a-helix.The two Cys residues were thus preferred to ethanedithiol as the scavenger.58 The peptideplaced at positions i, i+3 where we envisaged that formation dithiol was obtained as a yellow and green solid indicating of an intramolecular disulfide bond between them could be that some oxidation of the ferrocene had occurred.ES-MS of used to both stabilise the 310-helical conformation and protect the crude revealed that as well as the desired product peptides the thiol function (e.g. from intermolecular disulfide formation) lacking Aib and Fcb were also present in minor quantities. Attempts to purify the mixture by HPLC in methanol–water until contact with gold cleaved the SKS bond.The end and by crystallisation were unsuccessful and a sample of the sequences of PRM1 were conserved and two modified Glu crude material was adsorbed to the disk electrode and analysed residues with large electron donor and acceptor substituents by cyclic voltammetry, as described below. were replaced with Phe and (NO2)Phe. These were chosen as The cyclisation by disulfide formation of short b-turn- their side-chains are relatively bulky (which might disrupt the forming peptides has been carried out on the resin using iodine hydrophobic aggregation responsible for the low solubility of by Albericio and co-workers, the ‘pseudo-dilution’ conditions 16Met3) but did not require protecting groups, thus keeping favouring the intramolecular process.59 The similarities of the synthesis as simple as possible.The p-nitrophenylalanine these researchers’ sequence to ours (containing 2 Cys at i, residue served the dual purpose of enabling us to measure i+3) and the excellent yields of disulfide-peptide that they accurately the concentrations of the peptide solutions due to obtained after acidolytic cleavage made this an attractive its strong UV absorption and to act as an infrared probe of starting point.Thus Fmoc-16Cys2(Trt)2-O-resin was treated with a solution of iodine (2 equiv. per Cys) in acetonitrile at 0 °C and subsequently deprotected and acylated with Fcb-OH. After cleavage with 9555 TFA–water analysis of the crude material by Electrospray Ionisation Mass Spectrometry (ESIMS) confirmed that the desired peptide had been obtained. It is notable that eVecting the disulfide formation, deprotection and acylation steps in dimethylformamide (DMF) rather than acetonitrile produced undesired intermolecular disulfide bonds.This was established by the presence of 3+ and 4+ ions of the dimeric peptide in the ESI-MS. Other insoluble material may have been higher oligomers. In retrospect this is not too surprising as DMF hydrogen bonds to main chain N–H groups, destabilising the helix and allowing the peptide to access more extended conformations in which the S–S bonds may form between diVerent molecules.Alternatively as a weakly basic solvent DMF may have ‘scrambled’ the initially formed intramolecular bonds. The CD spectra of 6, Ac-16Cys2(S2) in methanol, ethanol and acetonitrile (Fig. 7) indicated that the peptide conformation was a mixture of helix and random coil. The intensities of the minima at 222 nm (-11 000 to -12 000) showed that Fig. 6 Space-filling model of 6, Ac-16Cys2 in a 310-helical confor- the helicity was significantly less than the Met-containing mation showing the alignment of the 2 S atoms on the underside of peptides. However, the unstructured sections of mixed heli- the helix.The main axis of the helix is normal to the plane of the cal–random coil peptides tend to be near the termini,60 and it page. The atom colours are C: green, H: white, N: blue, O: red, S: yellow. was anticipated that since in peptide 6 the dicysteine gold- J. Mater. Chem., 1999, 9, 1097–1105 1101Fig. 9 Cyclic voltammograms of self-assembled films of Fcb-16Cys2X on evaporated gold slides.Dot–dash line: peptide-disulfide (X=S2) Fig. 7 Circular dichroic spectra of peptide 6, Ac-16Cys2(S2) in self-assembled from acetonitrile; dotted line: disulfide (X=S2) selfmethanol (long dashes, 190 mM), ethanol (short dashes, 160 mM) and assembled from methanol; dashed line: dithiol (X=(SH)2) selfacetonitrile (continuous line, 160 mM) at room temperature.assembled from ethanol; continuous line: dithiol (X=(SH)2) selfassembled from methanol. All CVs were recorded at 80 mV s-1 in 0.1 M aq. NaClO4 vs. a Ag wire (Ag/Ag+) reference. binding site lies in the middle it may be unperturbed. Despite our aim of designing a 310-helix the ratio of the intensities of the minima at 208 nm and 222 nm of ca. 1 is more typical of this and the measured anodic charges, which were now signifi- an a-helix; in 310-helices it is often 0.5 to 0.8.61,62 It is probable cantly greater than calculated for a monolayer of 6, we that at 16 amino acids, 16Cys2 is suYciently long that the a- concluded that the adsorbed film was several layers thick. A helix is the favoured conformation. reasonable explanation for this solvent-dependent voltammetry A self-assembled film of Fcb-16Cys2(SH)2 on a gold disk is that the hydrophobic peptide was solvated better by the electrode was prepared using a saturated solution of freshly organic than by the aqueous solvent, allowing greater percleaved peptide in methanol.CV of the unrinsed electrode meation of the (smaller) electrolyte counter-ion into the after 2 days did not initially contain any faradaic waves but adsorbed film and lowering the activation energy of the Fc0/+1 after repeated cycling the oxidation and reduction waves of oxidation.the ferrocene became discernible (Fig. 8). After 10 days storage Films of the ferrocene derivative of 6, disulfide and dithiol in a sealed jar the CV was measured in organic solution (n- were also prepared on evaporated gold that was immersed in Bu4NBF4 in acetonitrile, Fig. 8). As before several potential dilute peptide solutions for 5 days. The adsorbed layer was sweeps were required to condition the film and reveal the characterised by ellipsometry to measure the thickness and ferrocene peaks, which were still unexpectedly broad. From then by CV to measure the surface coverage.The voltammograms, which clearly feature the ferrocene wave, are shown in Fig. 9 and the calculated surface coverage and changes in the ellipsometric angles are summarised in Table 1. As can be seen from the Table the films prepared from ethanol or methanol Table 1 Thickness and coverage of self-assembled films of Fcb-16Cys2 on evaporated gold slides, calculated from ellipsometry and cyclic voltammetry Sample dDa dYa Surface coverage (%)b Dithiol 230 mM 28.4 -2.3 294 MeOH Dithiol 80 mM 9.5 -1.0 101 EtOH Disulfide 210 mM 17.5 -1.1 216 MeOH Disulfide 90 mM 1.7 -0.1 90 MeCN adD=D(bare gold)-D(coated slide), dY was calculated similarly.bCalculated from CV assuming that each peptide molecule covered Fig. 8 Cyclic voltammograms of Fcb-16Cys2(SH)2 film on a gold 43×12 A° of the surface with a surface density of 1.9×1013 disk electrode vs.Ag wire reference electrode. Continuous line: 20 min molecules cm-2. A monolayer of peptide 6 would be 12 A° thick and after rinsing, measured at 50 mV s-1 in 0.1 Maq. NaClO4. Dashed compared to bare gold would decrease D by 1.5 and result in line: 10 days after rinsing, 10th scan at 500 mV s-1 in 0.1 M tetra-n- approximately no change in Y.butylammonium tetrafluoroborate in acetonitrile. 1102 J. Mater. Chem., 1999, 9, 1097–1105were characterised by ellipsometry as being several molecular column. A column temperature of 30 °C was maintained. Each residue was double coupled with a recirculation time of 40 min layers thick, even though the measured charge of the peptide layer that self-assembled from ethanol was that expected for using a 2.5-fold excess of Fmoc-amino acid (Fmoc=fluorenylmethoxycarbonyl ) for 12Leu2 and a 3-fold excess for all a monolayer.This apparent contradiction is reminiscent of the ‘false’ surface coverage measured of the film prepared on other peptides. Acylations were monitored at 348 nm. Fmoc deprotection by 20% piperidine in DMF (dimethylformamide) the gold disk electrode (described above) and probably has the same cause.This film was also characterised by RAIRS. for 15 min was followed at 300 nm. Deprotection of the resinbound Fmoc (PR500 amide resin) or pre-coupled Fmoc-Ala The intensity of the bands in the surface spectrum (not shown) were consistent with a film of multilayer thickness, in (PA500 acid resin) was carried out twice.HPLC was carried out on analytical (250×4.6 mm) and agreement with ellipsometry. In view of these results we concluded that even on a gold semi-prep. (250×10 mm) columns each containing 5 mm C8 Spherisorb stationary phase. Flow rates were 1 ml min-1 for surface protic solvents could disrupt the peptide helix conformation suYciently that one of the thiol groups became avail- the analytical and 5 ml min-1 for the semi-prep.column and detection was at 230 nm unless stated otherwise. Elution was able to form a disulfide bond with a molecule in solution. The aggregation process could continue and lead to the build up by gradients of water with either methanol or DMF (see supplementary information). of a thick film of oligodisulfide on the metal as observed. This hypothesis was given greater credence by the much more Electrospray ionisation mass spectra (ES-MS) were acquired on a Fisons VG Platform.The carrier solvent was 151 successful experiments carried out in non-protic acetonitrile. The changes in the ellipsometric angles and charge passed acetonitrile–water. Circular dichroic spectra were measured using a Jasco J600 spectropolarimeter with a 1 cm path-length for oligopeptide 6, Fcb-16Cys2(S2)-OH self-assembled from acetonitrile solution are shown in Table 1.Strikingly both sets cell. Molecular models were constructed on a Silicon Graphics Iris workstation using the M. S. I. software packages Builder of values are now in good agreement and correspond very well with those calculated for a close-packed SAM of helical and Biopolymer.peptide 6. The successful preparation of monolayers of 16Cys2, therefore appears to require use of solvents which do not Materials. All commercial samples used in this study were used as received. All resins, Fmoc-amino acids and coupling disrupt the hydrogen-bonding of the backbone peptide and thus stabilise the helix conformation.Interestingly this is reagents were of peptide synthesis grade and were supplied by Novabiochem. Acetic anhydride (98%), acetonitrile (for CD, exactly what was found in the earlier formation of the intramolecular S–S bond. EYcient preorganisation of binding sites and spectrophotometric grade), tert-amyl alcohol (99%), carbon tetrachloride (99%), DBU (98%), dichloromethane (reagent eYcient self-assembly onto gold are therefore both key requirements for successful preparation of helical peptide monolayers. grade), DIPEA (99%), lithium bromide (99+%), methanol (for CD, spectrophotometric grade), p-nitrobenzoyl chloride (NB-Cl, 98%) and triisopropylsilane (TIPS, 99%) were pur- Summary and prospects chased from Aldrich.N,N-Dimethylformamide (DMF), diethyl ether, piperidine and trifluoroacetic acid (TFA) were We have outlined a generic strategy for preparing chiral all of peptide synthesis grade and were supplied by Rathburn.functionalised surfaces using SAMs of helical oligopeptides. Acetonitrile and methanol used for HPLC were HPLC grade Those peptides containing Met are mainly helical in organic and were supplied by Merck.Glacial acetic acid and iodine solvents and RAIRS of their SAMs on gold confirmed that (99.5+%) were supplied by Fisons. 4-Ferrocenylbutyric acid they are aligned with the helix axes parallel to the surface. and 3-ferrocenylpropionic acid were gifts from Dr P. However these peptides are of low solubility which leads to Whittaker, University of Strathclyde.handling problems. Their coverage and strength of binding to the surface are also variable probably because the weak Resin washing protocol. The protected peptide-resin was sulfur–gold bond of thioethers means surface preparation is washed sequentially with DMF, tert-amyl alcohol, glacial of utmost importance. Peptides with the thiol-containing Cys acetic acid, tert-amyl alcohol, dichloromethane and diethyl as the gold-binding residue also face potential problems.In ether, dried under high vacuum for at least 5 h and stored particular there is a strong risk of oligomerisation occurring under nitrogen at -20 °C. through formation of intermolecular disulfide bonds, either during S–S bond formation or during self-assembly to the Post-synthesiser deprotection.Protected peptide-resin gold. This is particularly prevalent in protic solvents and can (22 mmol of peptide) was swelled in DMF (4 ml ) for 1 h. lead to formation of thick films. However, if an aprotic solvent Piperidine was added to a ratio of 25% and the mixture was such as acetonitrile is used, the increased stabilisation of the agitated gently. After 5 min standing the resin was filtered helix and resultant preorganisation of the thiols allows eYcient under gravity, washed twice with DMF and stored with 25% self-assembly of close-packed monomeric peptide monolayers. piperidine in DMF for 10 min, swirling occasionally.The resin These monolayers oVer a viable route for preparing funcwas filtered under gravity, washed three times with DMF and tionalised surfaces with nanoscale spatial resolution. As such acylated immediately.they are likely to find applications in fields such as molecular electronics, biosensors, biomaterials and catalysis. Post-synthesiser acetylation. Acetic anhydride (30 ml, 0.3 mmol) and DIPEA (40 ml, 0.2 mmol) were added to the Experimental deprotected peptide-resin (40 mmol of peptide) in DMF (4 ml ).After occasional gentle swirling for 30–45 min the resin was Peptide synthesis filtered and washed twice with DMF. Fresh DMF, acetic anhydride and DIPEA were added to the resin and the General. Oligopeptides were synthesised by a NovasynA Crystal Continuous Flow Peptide Synthesiser. Fmoc amino acylation was repeated. The Ac-peptide-resin was filtered, washed by the same protocol as the post-synthesis washing acids were activated with PyBOPA (benzotriazol-1-yloxytripyrrolidinophosphonium hexafluorophosphate) and N-hydroxy- then either dried under high vacuum and stored under nitrogen at -20 °C or the peptide was cleaved immediately.benzotriazole (HOBt, one equivalent each) and diisopropylethylamine (DIPEA) (2 equivalents for Ala, Met, Fcb, Phe, (pNO2)Phe, Leu and (Trt)Cys, 3 equivalents for Post-synthesiser acylation with 4-ferrocenylbutyric acid.Deprotected peptide-resin (22 mmol of peptide) was treated Aib and Nle) immediately prior to injection onto the reaction J. Mater. Chem., 1999, 9, 1097–1105 1103twice with a solution of 4-ferrocenylbutyric acid (18 mg, and sodium perchlorate (98%) were supplied by Merck. Acetonitrile (HPLC grade) was puchased from Rathburn. 66 mmol), PyBOPA (34 mg, 66 mmol), HOBt (10 mg, 66 mmol ) and DIPEA (35 ml, 200 mmol ) in DMF (5 ml ) for 30–45 min. Chromium tris(tetrachlorocatecholate) was prepared by the method of Pierpont et al.63 The resin was filtered and washed twice with DMF between couplings. After the second acylation the resin was washed and dried as described in the acetylation procedure.Substrate preparation. Gold disk electrodes (2 or 3 mm diameter, supplied by Oxford Electrodes) were manually pol- Post-synthesiser acylation with p-nitrobenzoyl chloride. ished with alumina paste, sonicated in water and dried gently Swelled, deprotected peptide-resin (3 mmol of peptide) was in a stream of nitrogen. The electrode was immersed in 50% acylated by double-coupling for 45 min with p-nitrobenzoyl aqua regia (45351 water–conc. HCl–conc.HNO3) for 60 s, chloride (3 mg, 16 mmol ) and DIPEA (5 ml, 25 mmol ) in DMF water for 10 s, the contacting solvent for 10 s and then the (0.5 ml ). The peptide-resin was washed twice between couppeptide solution. When the contacting solvent was dichloro- lings with DMF and after the second wash was treated in a methane the electrode was dipped in methanol for 10 s between similar manner to the acetylated peptide-resin.the water and dichloromethane rinses. Electrodes prepared in this manner were measured to have surface roughness of 2, Standard cleavage procedure for amide resin. Ac-peptidefrom the anodic oxidation of adsorbed iodine to aqueous resin (26 mg, 9 mmol of peptide) was swelled by washing on a iodate.64 frit with glacial acetic acid, dichloromethane and methanol Evaporated gold substrates were prepared on glass (4 ml each). The resin was transferred to a flask and a solution microscope slides (from Merck) for RAIRS of peptides 1 and of 10% TFA in dichloromethane (total 10 ml ) was added.The 2, silicon wafers (100 orientation, from Micro-Image beads turned red immediately.The mixture was swirled gently Technology Ltd.) for peptides 3 and 4 and on polyacrylamide every 10–15 min for 45 min, the resin was filtered and the gel slabs (from Pharmacia) for 6. The glass microscope slides beads were stored in fresh cleavage mixture for a further were cut to squares of approx. 2.5 cm, cleaned in piranha 45 min. The resin was filtered again and washed with 10% solution (151 H2O2–conc.H2SO4 CAUTION! THESE TFA in dichloromethane (2×10 ml ) and once with methanol SOLUTIONS ARE HIGHLY OXIDISING AND SHOULD (10 ml ). The combined filtrates were diluted with a volume of BE HANDLED WITH EXTREME CARE!) for 1 h, washed acetonitrile equal to that of TFA and concentrated under thoroughly with distilled water and sonicated for 20 min each reduced pressure to an oil.This was triturated with diethyl in acetone and absolute ethanol. The slides remained immersed ether (16 ml ) and the resulting white precipitate was centriin ethanol until they were placed in the evaporating chamber, fuged (3×20 min). The collected solid was either dried under where 3–5 nm of chromium and 40–50 nm of gold were high vacuum overnight and stored under nitrogen at -20 °C evaporated at a base pressure of 2×10-6 Torr and or used immediately to prepare SAMs.were immediately transferred to the contacting solutions. The other gold slides were prepared under semi-conductor Standard cleavage procedure for acid resin. Triisopropylsilane technology clean room conditions. Polyacrylamide gel slabs (TIPS, 13 ml, 66 mmol ), methanol (0.25 ml ) and TFA (4.75 ml ) were cut into strips 20×70 mm and placed on an Al mask so were added to peptide-resin (82 mg, 16 mmol of peptide).The that the coated area was of 2 rectangles 18×16 mm connected mixture was shaken occasionally and after 1–2 h was filtered. by a strip 2 mm wide and 28 mm long. Si wafers were either The resin was washed with 9555 TFA–methanol (2×5 ml) used whole or cut into strips 8 mm wide.The substrates were and 151 TFA–methanol (5 ml ). The combined filtrates were cleaned by sonication in either IPA or water–detergent then concentrated, precipitated with diethyl ether, centrifuged and in water and dried in a stream of high-purity N2. Cr and Au dried according to the same protocol as used for the amide were evaporated at a rate of 1 A° s-1 to thicknesses of 100 and resin. 1000 A° . The slides prepared on silicon wafers were immersed in piranha solution for 45 min, rinsed thoroughly with distilled Protocol for oxygen-free cleavage. Peptide-resin was water and dried in a stream of N2. All other slides were used transferred to a fritted column equipped with a tap at the without further treatment.bottom and a nitrogen inlet at the top. The apparatus was charged with nitrogen and the resin was washed sequentially with acetic acid, dichloromethane and methanol. Degassed Preparation of peptide and amino acid films. All solutions of TFA and dichloromethane or methanol were added and the peptides 3, 4 and 6 were degassed except for the SAM of 6 on column was sealed under a nitrogen atmosphere, gently agitat- a gold disk electrode.After immersion the substrates were ing the column every 10–15 min. After 45 min the solvents rinsed to remove excess peptide, unless stated otherwise in were filtered, freshly degassed cleavage solution was added the text. and the mixture allowed to stand for a further 45 min. The solvents were filtered and the resin was washed with cleavage mixture and with methanol.The solution of cleaved peptide Electrochemistry was removed via syringe and concentrated under reduced CV was controlled by an AutolabA potentiostat from Eco pressure to an oil. This was subsequently treated in the manner Chemie via General Purpose Electrochemical System 3 on a described in the standard cleavage procedure. PC.A one-compartment cell was used. Oxygen was not excluded. The counter electrode was a coiled Pt wire, supplied Peptide purification and characterisation. Available as by Goodfellow. The reference electrodes used were Ag/AgCl supplementary material. (Oxford electrodes) for voltammetry of peptides 1 and 2, a Ag wire coated in AgCl (a gift from C. Agra-Gutierrez) for CV Preparation and characterisation of self-assembled monolayers of Ru(NH3)6Cl2 and K3[Cr(Cl4C6O2)3] and a silver wire (Ag/Ag+) for all other experiments.The electrolytes were Materials. Absolute alcohol (reagent grade), hexaamineruth- 10 or 100 mM tetra-n-butylammonium tetrafluoroborate in enium(II) chloride, sodium tetrafluoroborate (98%) and tetraorganic solvents, saturated KCl in 951 methanol–water for n-butylammonium tetrafluoroborate (99%) were supplied by K3[Cr(Cl4C6O2)3] and aqueous solutions of 0.3 M KCl for Aldrich.Acetone, hydrogen peroxide (30 vol.%) and conc. electrochemistry of Ru(NH3)6Cl2, 0.1M NaClO4 or 0.1M hydrochloric, nitric and sulfuric acids were all reagent grade and were supplied by Fisons. Potassium chloride (AnalaRA) NaBF4. 1104 J. Mater. Chem., 1999, 9, 1097–1105L. S. Curtin, S. R. Peck and R. W. Murray, Electrochim. Acta, Reflection–absorption infrared spectroscopy (RAIRS) 1995, 40, 1331. Infrared spectra were recorded on a Unicam Galaxy Series 23 L. H. Dubois and R. G. Nuzzo, Annu. Rev. Phys. Chem., 1992, 43, 437. FTIR 7000 equipped with a liquid nitrogen-cooled Hg–Cd–Te 24 A. Ulman, Ultrathin Organic Films; From Langmuir-Blodgett to detector and an FT85 specular reflector from Spectra Tech Self-Assembly, Academic Press, Boston, 1991.Inc. The sample compartment was purged with dry air. The 25 W. B. Caldwell, K. Chen, B. R. Herr, C. A. Mirkin, J. C. Hulteen angle of incidence was at 85° to the surface normal and the and R. P. Van Duyne, Langmuir, 1994, 10, 4109. incident radiation was plane-polarised parallel to the direction 26 C.M. Yip and M. D. Ward, Langmuir, 1994, 10, 549. of propagation. Reflection–absorption spectra were accumu- 27 X. Tang, T. Schneider and D. A. Buttry, Langmuir, 1994, 10, 2235. 28 L. Haussling, H. Ringsdorf, F.-J. Schmitt and W. Knoll, lated over 200 scans at 4 cm-1 resolution and were ratioed Langmuir, 1991, 7, 1837. against an unused gold slide or a control slide that had been 29 W.S. V. Kwan, L. Atanasoska and L. L. Miller, Langmuir, 1991, immersed in the same solvent. 7, 1419. 30 J. M. Tour, L. Jones II, D. L. Pearson, J. J. S. Lamba, Ellipsometry T. P. Burgin, G. M. Whitesides, D. L. Allara, A. N. Parikh and S. V. Atre, J. Am. Chem. Soc., 1995, 117, 9529. Ellipsometry measurements were made on a PC-controlled 31 E.U. Thoden van Velzen, J. F. J. Engbersen, P. J. de Lange, Auto Gain Ellipsometer L116B from the Gaertner Scientific J. W. G. Mahy and D. N. Reinhoudt, J. Am. Chem. Soc., 1995, Corporation, Chicago. The light incident at 70° to the normal 117, 6853. 32 H. Adams, F. Davis and C. J. M. Stirling, J. Chem. Soc., Chem. was at a wavelength of 632.8 nm from a He–Ne laser.The Commun., 1994, 2527. optical properties of the substrate and of the adsorbed layer 33 R. P. H. Kooyman, D. J. van den Heuvel, J. W. Drijfhout and were calculated using the classical 2- and 3-phase parallel layer G. W.Welling, Thin Solid Films, 1994, 244, 913. models. The real component of the refractive index of the 34 C. Duschl, M. Liley, G. Corradin and H. Vogel, Biophys.J., 1994, films, nF, was assumed to be 1.500. The average of four 67, 1229. recordings were made at each of three positions per slide. The 35 C. D. Bain, J. Evall and G. M. Whitesides, J. Am. Chem. Soc., 1989, 111, 7155. surface area sampled by the laser was 2 mm2. The optical 36 M. Knichel, P. Heiduschka, W. Beck, G. Jung and W. Gopel, characteristics of the bare gold were calculated from the Sens.Actuators B, 1995, 28, 85. measurements made on the same slide before immersion 37 C. Duschl, A. F. Sevinlandais and H. Vogel, Biophys. J., 1996, in solution. 70, 1985. 38 M. Boncheva and H. Vogel, Biophys. J., 1997, 73, 1056. 39 J. K. Whitesell and H. K. Chang, Science, 1993, 261, 73. Acknowledgements 40 S. Sakamoto, H. Aoyagi, N. Nakashima and H. Mihara, J.Chem. Soc., Perkin Trans. 2, 1996, 2319. We thank the EPSRC for funding, Novabiochem, Dr S. Kelly 41 R. Naumann, A. Jonczyk, R. Kopp, J. van Esch, H. Ringsdorf, at the BBSRC Circular Dichroism Facility, University of W. Knoll and P. Graber, Angew. Chem., Int. Ed. Engl., 1995, Stirling, Dr R. Raval, E. Cooper and L. Shorthouse at 34, 2056. University of Liverpool for assistance with RAIRS measure- 42 L.Scheibler, P. Dumy, D. Stamou, C. Duschl, H. Vogel and M. Mutter, Tetrahedron, 1998, 54, 3725. ments and Dr L. Berlouis at University of Strathclyde for 43 A. E. Strong and B. D. Moore, Chem. Commun., 1998, 473. advice on ellipsometry. 44 I. L. Karle and P. Balaram, Biochemistry, 1990, 29, 6747. 45 C. Toniolo and E. Benedetti, Trends Biochem. Sci., 1991, 16, 350. 46 F. Donald, G. Hungerford, D. J. S. Birch and B. D. Moore, References J. Chem. Soc., Chem. Commun., 1995, 313. 47 G. P. Dado and S. H. Gellman, J. Am. Chem. Soc., 1993, 115, 1 C. A. Mirkin and M. A. Ratner, Annu. Rev. Phys. Chem., 1992, 12609. 43, 719. 48 S. Marqusee, V. H. Robbins and R. L. Baldwin, Proc. Natl. Acad. 2 R. W.Munn, Biosystems, 1992, 27, 207. Sci. USA, 1989, 86, 5286. 3 F. L. Carter, Molecular Electronic Devices II, Marcel Dekker, 49 S. P. Ho and W. F. DeGrado, J. Am. Chem. Soc., 1987, 109, 6751. New York, 1987. 50 R. W. Woody, in The Peptides: Analysis, synthesis, biology, 1985, 4 A. Aviram and M. A. Ratner, Chem. Phys. Lett., 1974, 29, 277. Vol. 7, p. 15. 5 J. J. Hopfield, J. N. Onuchic and D. N. Beratan, J. Phys. Chem., 51 P. I. Haris and D.Chapman, Biopolymers, 1995, 37, 251. 1989, 93, 6350. 52 R. J. Greenler, J. Chem. Phys., 1966, 44, 310. 6 H. Tachibani, T. Nakamura, M. Matsumoto, H. Komizu, 53 S. Krimm and J. Bandekar, Adv. Protein Chem., 1986, 38, 181. E. Manda, H. Niino, A. Yabe and Y. Kawabata, J. Am. Chem. 54 M. D. Porter, T. B. Bright, D. L. Allara and C. E. D. Chidsey, Soc., 1989, 111, 3080. J. Am. Chem. Soc., 1987, 109, 3559. 7 C.Joachim and J. P. Launay, J. Mol. Electronics, 1990, 6, 37. 55 E. B. Troughton, C. D. Bain, G. M. Whitesides, R. G. Nuzzo, 8 M. D. Ward, Chem. Soc. Rev., 1995, 24, 121. D. L. Allara and M. D. Porter, Langmuir, 1988, 4, 365. 9 M. Mutter and R. Hersperberger, Angew. Chem., Int. Ed. Engl., 56 A. E. Strong, Ph. D. Thesis, University of Strathclyde, 1997. 1990, 29, 185. 57 G. B. Fields and R. L. Noble, Int. J. Pept. Protein Res., 1990, 10 G. Hungerford, M. Martinez-Insua, D. J. S. Birch and 35, 161. B. D. Moore, Angew. Chem., Int. Ed. Engl., 1996, 35, 326. 58 D. A. Pearson, M. Blanchette, M. L. Baker and C. A. Guindon, 11 R. A. Bissell, E. Cordova, A. E. Kaifer and J. F. Stoddart, Nature, Tetrahedron Lett., 1989, 30, 2739. 1994, 369, 133. 59 F.Albericio, R. P. Hammer, C. Garcý�a-Echeverrý�a, M. A. Molins, 12 L. Zelikovich, J. Libman and A. Shanzer, Nature, 1995, 374, 790. J. L. Chang, M. Pons, E. Giralt and G. Barany, Int. J. Pept. 13 M. P. Debreczeny, W. A. Svec, E. M. Marsh and R. Wasielewski, Protein Res., 1991, 37, 402. J. Am. Chem. Soc., 1996, 118, 8174. 60 S. M. Miick, K. Casteel and G. L. Millhauser, Biochemistry, 1993, 14 R.R. Birge, Sci. Am., 1995, March, p. 66. 32, 8014. 15 U. P. Wild, S. Bernet, B. Kohler and A. Renn, Pure Appl. Chem., 61 T. Iwata, S. Lee, O. Oishi, H. Aoyagi, M. Ohno, K. Anzai, 1992, 64, 1335. Y. Kirino and G. Sugihara, J. Biol. Chem., 1994, 269, 4928. 16 C. D. Bain and G. M. Whitesides, Angew. Chem., Int. Ed. Engl., 62 W. L. Fiori, S. M. Miick and G. L. Millhauser, Biochemistry, 1989, 28, 506. 1993, 32, 11957. 17 N. L. Abbott and G. M. Whitesides, Langmuir, 1994, 10, 1493. 63 C. G. Pierpont, H. H. Downs and T. G. Rukavina, J. Am. Chem. 18 K. L. Prime and G. M. Whitesides, Science, 1991, 252, 1164. Soc., 1974, 96, 5573. 19 G. M. Whitesides and P. E. Laibinis, Langmuir, 1990, 6, 87. 64 J. F. Rodriguez, T. Mebrahtu and M. P. Soriaga, J. Electroanal. 20 L.-H. Guo, J. S. Facci and G. McLendon, J. Phys. Chem., 1995, Chem., 1987, 233, 283. 99, 8458. 21 L. A. Hockett and S. E. Creager, Langmuir, 1995, 11, 2318. 22 J. N. Richardson, G. K. Rowe, M. T. Carter, L. M. Tender, Paper 8/06860G J. Mater. Chem., 1999, 9, 1097–1105 1105 J O U R N A L O F C H E M I S T R Y Materials Self-assembling monolayers of helical oligopeptides on gold with applications in molecular electronics† Andrew E.Strong and Barry D. Moore* Department of Pure and Applied Chemistry, University of Strathclyde, Thomas Graham Building, 295 Cathedral Street, Glasgow, UK G1 1XL. E-mail: b.d.moore@strath.ac.uk Received 3rd September 1998, Accepted 1st March 1999 A generic strategy is outlined for preparing a chiral functionalised surface using self-assembled monolayers (SAMs) of helical oligopeptides for intended application in molecular electronics, sensors and catalysis.Helical peptides have been designed and synthesised with 3 Met or 2 Cys residues per peptide, placed in the sequence so that in a helical conformation their side-chains are on the same side of the helix and organised to bind to gold by selfassembly of the thioether or thiol functions.Circular dichroism (CD) of the Met-containing peptides showed that they are mainly helical in organic solvents and FT-RAIRS (reflection–absorption infrared spectroscopy) of their SAMs on gold confirmed that they are aligned with the helix axes parallel to the surface. However complete coverage of the surface is dependent critically on the preparation of the gold.CD of an oligopeptide containing 2 Cys residues suggested that the conformation was a mixture of helix and random coil. Films of this peptide adsorbed from diVerent solvents were characterised by ellipsometry and cyclic voltammetry. Ethanol and methanol led to multilayer formation indicating protic solvents destabilise the helical structure, allowing formation of intermolecular disulfide bonds.By contrast the aprotic solvent acetonitrile led to formation of a close-packed selfassembled monolayer suitable for elaboration into a functionalised supramolecular architecture. As the size of electronic components decreases, approaches nents it will be necessary to specifically link the peptides in have been made to design and prepare molecules that possess solution prior to the self-assembly step.Methods are available properties that are useful in electronic circuits.1 These eVorts to link the peptides through either covalent or non-covalent may be split broadly into two areas,2 molecular materials for interactions. Possibilities include forming a peptide bond electronics, which use the macroscopic properties of the mate- between the side-chains bearing amine and acid groups (e.g.rials, and electronics at the molecular level which exploits the Lys and Glu) or using amino acids bearing complementary microscopic properties of individual molecules.3 With respect hydrogen bonding arrays such as those found in nucleic acids. to the second area, various ingenious ideas have been proposed The supramolecular structure will ultimately be confined to a for using molecules as a rectr4 or a shift register5 and some surface by sulfur-mediated self-assembly on gold.The overall molecules have been demonstrated to act as switches6–13 and strategy is summarised in Fig. 1. storage devices.14,15 However most of these have been in Our approach relies on the fundamentals of self-organisation solution. In a functioning computer it is essential that the prevalent in Nature, which uses information ‘encoded’ in the memory, input, output and processing units are at well-defined structures of simple molecules to direct the formation of much positions to enable them to communicate with each other.larger architectures with known conformation, e.g. proteins, This indicates a solid state device.In this paper we present a nucleic acids or cells. Thus choice of peptide sequence ‘selfplan to use self-assembled monolayers (SAMs) of helical organises’ the molecules into a helix; specific linking of helices oligopeptides to produce a surface on which various functional in solution may be performed using complementary molecular groups may be positioned with nanoscale resolution. We recognition, and the final step is self-assembly of the supramolenvisage that this may also be of utility in fields such as ecular architecture onto a surface using the specificity of the sensors, biomaterials and catalysis.sulfur–gold interaction. Our strategy is to prepare helical oligopeptides that contain Self-assembling monolayers of sulfur-containing molecules both attachment sites for molecular computing elements and on gold are attractive systems for tailoring surface properties sulfur containing side-chains to promote self-assembly on gold.and the study of interfacial phenomena, e.g. wetting and By using solid-phase peptide synthesis these diVerent groups adhesion,16–19 and fundamentals of electron transfer.20–22 Most can be introduced into the sequence at predefined sites relative eVorts have concentrated on the formation, structure and to each other.In an ideal a-helix one turn around the helix properties of monolayers of long-chain alkanethiols.23,24 These axis encompasses 3.6 residues. Thus, placement of three S- systems have the advantages16 that a strong, specific interaction containing residues at positions i, i+4 and i+7 will ensure between the adsorbate and the substrate drives spontaneous their side-chains all extend from the same side of the helix, formation of a monolayer, the films are densely packed and preorganised to bind to a gold surface.Including the residue physically robust and a variety of other functional groups may modified with the molecular computing element at i+2 will be accommodated within the monolayer.Examples of other result in this group lying on the opposite side of the helix so functional groups include alkanethiols with terminal ferrothat following self-assembly it will be at the ambient/film cenylazobenzenes,25 tetrathiafulvalene,26 viologens,27 biotin28 interface. To achieve lateral positioning of diVerent compo- and thiol-terminated oligoimides29 and oligo(phenyleneethynylene) s.30 The potential of exploiting these structures in †Peptide purification and characterisation details are available as sensors31–34 and biomaterials18 has been recognised.supplementary data (SUPPL. NO. 57510, pp. 9) from the British Terminally-modified alkanethiols and other molecules where Library.For details of the Supplementary Publications Scheme, see the long molecular axis extends out from the metal surface ‘Information for Authors’, J. Mater. Chem., available via the RSC still present a homogeneous surface to the organic/ambient web page (http://www.rsc.org/authors). For direct electronic access see http://www.rsc.org/suppdata/jm/1999/1097/. interface. Although the mole fraction composition of mixed J.Mater. Chem., 1999, 9, 1097–1105 1097of the helix and formed stable monolayers on gold with the helix axis parallel to the surface as characterised by electrochemistry and reflection–absorption infrared spectroscopy (RAIRS). In this full paper we describe the design and preparation of SAMs of other helical oligopeptides containing several Met or Cys residues and their characterisation by cyclic voltammetry (CV), RAIRS and ellipsometry.The conditions required for monolayer formation of each type of peptide through the side-chain sulfur functions are discussed. Results and discussion General design principles Helical peptides are attractive building blocks with which to construct a functionalised surface as their synthesis and structure have been well studied.The more common a-helix is stabilised by hydrogen bonds between the main chain N–H of the (i+4)th and the main chain CLO of the ith amino acid in the sequence and one turn around the helix axis encompasses 3.6 residues. An alternative known helical geometry is the 310- helix44,45 so called because 3 residues form one complete turn around the helix axis and there are 10 atoms within the intramolecular ring formed by the hydrogen bond between N–H of residue i+3 and the CLO of residue i.The diVerences in pitch and hydrogen bonding mean that a 310-helix is longer and thinner than an a-helix of the same number of residues. The conformation of a peptide is determined by its amino acid sequence. In particular, peptides consisting entirely of amino acids with hydrocarbon side-chains are known to adopt helical Fig. 1 Proposed preparation of an ordered, functionalised surface structures in organic solvents.44,45 The 20 a-amino acids with nanometre scale features using helical oligopeptides. The present in mammalian proteins oVer a variety of side-chain schematic is viewed along the helix axes of peptides 1, 2 and 3 which functions, some of which (e.g.Lys, which contains the NH2 have side-chains modified to carry functional elements A, B and C group, and the CO2H of Glu) may be readily modified to respectively on the opposite sides of the helices to thiol or thioether introduce other desired functions. Two amino acids, Cys and groups. The peptides are specifically linked in solution through complementary binding groups to form discrete supramolecular Met contain sulfur functions in their side-chains (-CH2SH assemblies.These are then locked together and immobilised via and -CH2CH2SCH3 respectively) which are expected to form formation of a self-assembled monolayer on gold using the thiols or strong bonds with a gold surface via self-assembly. Of these thioethers.The resultant surface confined molecular architecture will two Met was initially the more attractive. The greater separa- contain functional elements A, B and C placed at known relative tion between the peptide backbone and the sulfur atom in positions a few nm apart. This plan could be extended to include Met compared with Cys should allow the sulfurs to find more functional elements by increasing the number of diVerent peptides or by including several elements on each peptide chain.optimal positions on the gold surface. It should also reduce unfavourable steric interactions between the metal surface and other side-chains on that side of the helix. Furthermore, monolayers of alkanethiols co-adsorbed with terminally funcalthough the thiol of Cys binds more strongly to gold than tionalised thiols may be controlled empirically, the relative the thioether of Met, it was anticipated that the biomimetic positioning of the two components is still random.35 route of multiple weak interactions might be a more eYcient In the few previous examples of SAMs of peptides that have route to well ordered systems.In this paper we compare the recently been reported, the peptides generally had only one formation and structure of SAMs of peptides containing both attachment point to the gold surface, either through the sidetypes of residue.chain of a terminal Cys residue34,36–38 or via a thioalkyl carboxylic acid coupled to the N-terminal.33,39–41 Outside our Met-containing peptides own work, one notable exception was the recent report from the Mutter and Vogel laboratories,42 which described the Our prototype sequence was that of peptide 1, Fcb-13Met3. Our aim of adsorbing the peptides to gold in a helical formation and characterisation of a SAM of a cyclic peptide template with 2 or 4 pendant alkylthiolates covalently bound conformation via the thioethers placed certain restrictions on the primary structure.In particular, alignment of the surface- to side-chains of the template. The solution side of the template was derivatised with metal-binding residues to which the binding side-chains along the same side of the a-helix such that they could bind cooperatively to a gold surface (Fig. 2) binding of metals and subsequent ligands was demonstrated. Although it was not discussed in their paper this system could required Met to be included at positions 4, 8 and 11.The designed formation of three sulfur–gold bonds was expected also be extended to produce a functionalised surface similar to that we have proposed. to introduce an element of selectivity as helical conformations should be bound more strongly than other conformations As our primary objective we have concentrated on investigating the formation and structure of SAMs of monomeric where only one or two peptide–surface interactions are possible. The remaining amino acids used, Ala and Aib (a- helical oligopeptides aligned with the helix axis parallel to the surface.We have previously described the formation and aminoisobutyric acid) are known strong helix formers in organic solvents44,45 and have small methyl side-chains to characterisation of SAMs of the helical oligopeptide Fcb- 13Met3 (1), Fcb-Ala-Aib-Ala-Met-Aib-Ala-Ala-Met-Ala-Aib- minimise unfavourable steric interactions with the surface.The N-terminus was acylated with ferrocenylbutyric acid to pro- Met-Ala-Ala-NH2 (where Fcb is ferrocenylbutyric acid).43 This 13 residue peptide contained three Met residues (hence vide a convenient electrochemical label.Peptide 2, Fcp-12Leu2, Fcp-Ala-Aib-Ala-Ala-Leu-Aib-Ala-Ala-Ala-Aib-Leu-Ala- 13Met3) positioned so their side-chains aligned along one side 1098 J. Mater. Chem., 1999, 9, 1097–1105parison of the adsorption of peptide 3 without the goldbinding thioether groups. Assembly of each sequence on the solid phase proceeded smoothly and small portions of each Fmoc-peptide-resin (Fmoc=fluorenylmethoxycarbonyl ) were deprotected and acylated as needed to provide the N-acetyl (Ac), N-(4-ferrocenylbutanoyl ) (Fcb) or N-p-nitrobenzoyl (NB) derivatives.As with peptides 1 and 2 cleavage of the ferrocene derivatives was carried out under oxygen-free conditions to minimise the oxidation of the ferrocene to ferrocenium.The Met-containing peptide contained small quantities of the sulfoxide. After purification by reversed-phase HPLC, derivatives of both peptides were only sparingly soluble in a range of solvents. Amphipathic helical peptides are known to aggregate in water to form bundles with the hydrophobic surfaces shielding each other from the solvent.47–49 These completely non-polar sequences may have aggregated in a similar manner even in these less polar solvents (methanol and ethanol; acetonitrile and ethyl acetate were also tried but no dissolution could be detected).Formation of b-sheets, which are often believed to aggregate and cause precipitation, was considered unlikely in peptides containing several Aib residues.44 The circular dichroic (CD) spectra were measured using the Fig. 2 Space-filling model of 1, 13Met3 in an a-helical conformation showing the alignment of the 3 S atoms along the underside of the NB derivatives, since here the peptide concentration could be helix.The main axis of the helix is normal to the plane of the page. determined accurately from the UV-absorbance of the NB The N-terminal group here, acetyl, is shown instead of the actual one group. Each spectrum (Fig. 3) showed the double-minima at (4-ferrocenylbutyric acid) for clarity. The atom colours are C: green, 208 and 222 nm typical of a-helices.50 The value of [h]222 of H: white, N: blue, O: red, S: yellow. -23 300 for NB-16Met3 demonstrated the expected increase in helicity over Fcb-13Met3 (for which [h]222 was -18 300).Films of Ac-16Met3, Ac-3, and Ac-16Nle3, Ac-4, were prepared on evaporated gold slides by immersion in saturated NH2 (where Fcp is ferrocenylpropionic acid), available from previous studies (J. Reid, D. Nonhebel and B. D. Moore, peptide solutions for 2 days. ‘Cast’ films of each peptide were also prepared by pipetting their solutions onto slides and unpublished results) provided a convenient, non-sulfurcontaining control of similar length, composition and second- allowing the solvent to evaporate.The reflectance spectra of the cast films were taken as being representative of disordered ary structure43 to 1. As described previously43 both peptides 1 and 2 were films (Fig. 4). The amide I bands at 1663 cm-1 and 1667 cm-1 for Ac-3 and Ac-4 respectively are in the region more typical synthesised eYciently by conventional solid phase methods.The circular dichroic spectra of Fcb-13Met3 confirmed that it of 310-helices51 but may also indicate a-helices. The latter are characterised by an amide I maximum at 1650–1658 cm-1, was mainly helical in both methanol and acetonitrile. SAMs of 1 were prepared on a gold disk electrode and were found but this may be shifted to higher frequency by ca. 10 cm-1 after adsorption on gold.38 As expected no bands were detected to be remarkably stable to repetitive cycling in ethanol–electrolyte. In contrast the initial surface coverage of peptide 2 was in the reflectance spectrum of the slide immersed in Ac-4. Gratifyingly in the surface spectrum of Ac-3 an amide I band typically less than 15% of peptide 1 and rapidly decreased during cycling under the same conditions.In the RAIRS spectrum of a SAM of the Met-containing peptide on evaporated gold the amide I band was at 1664 cm-1, indicative of a 310-helix, and the amide II and amide III bands were at 1545 cm-1 and 1264 cm-1 respectively. Comparison of the reflectance with the transmission spectrum measured of 1 dispersed in KBr confirmed that the helix axis was parallel to the surface, the orientation expected for a self-assembled layer.In contrast the RAIRS of Fcp-12Leu2 suggested no particular orientation of the peptide at the surface. Following these promising results we sought to design completely a-helical Met-containing peptides and to prepare SAMs with greater coverage of the surface. To increase the ahelicity the length of peptide 3, 16Met3, Ala-Ala-Aib-Ala- Met-Ala-Phe-Ala-Met-Aib-Phe-Met-Aib-Ala-Ala-Ala-NH2, was increased to 16 residues as longer peptides favour the aover the 310-helical conformation.44 Two Phe residues (sidechain CH2Ph) were placed at positions 7 and 11 in the sequence where they could potentially stabilise the a-helical form by solvophobic packing of the side-chain in polar solvents such as acetonitrile and methanol.Although in proteins it is more usually found in b-sheets Phe has been used in organic-soluble helical peptides.44,46 The sequence of peptide 4, 16Nle3, Ala- Ala-Aib-Ala-Nle-Ala-Phe-Ala-Nle-Aib-Phe-Nle-Aib-Ala-Ala- Ala-NH2, diVered from that of 3 only by replacement of the Fig. 3 Circular dichroic (CD) spectra of peptides NB-3 (continuous three Met residues with the isosteric Nle (side-chain line, 130 mM) and NB-4 (dashed line, 180 mM) in methanol at room temperature. -CH2CH2CH2CH3) and was chosen to provide a close com- J.Mater. Chem., 1999, 9, 1097–1105 1099what is found for the SAM of Ac-3 and supports the assignment of the amide I band at 1667 cm-1 to a helical conformation.Similar changes were observed by Boncheva and Vogel who prepared monolayers of helical peptides by the Langmuir–Blodgett and self-assembly techniques, employing a Cys at either the N- or C-terminus.38 Their more detailed theoretical analysis reached the same conclusions as ours and was supported by surface plasmon resonance measurements. The surface coverage and stability of the adsorbed peptide films were assessed by cyclic voltammetry of the ferrocene derivatives.In addition we investigated the ability of a SAM of Ac-16Met3 to block electron transport to a couple in solution. The working electrode used in this experiment was an evaporated gold slide which had been immersed in peptide solution (0.04 mM in methanol for 17 h) and rinsed.Voltammetry of the modified slide in a solution of Ru(NH3)6Cl2 showed no significant diVerence to that recorded at bare gold. Although electron transfer to this complex has previously been shown to be blocked by alkanethiol SAMs these present a much denser, thicker barrier than a closepacked monolayer of helical peptide (C21H43SH forms a monolayer approx. 30 A° thick compared to the peptide diameter of 12 A° ).54 The voltammograms of the much larger couple 5 [Cr(Cl4C6O2)3]3- (chromium tris(tetrachlorocatecholate) 33-) at a bare gold slide and at the peptide-modified electrode are shown in Fig. 5. Clearly the peptide monolayer strongly hindered electron transfer to and from this bulky Cr complex indicating that a coherent monolayer had adsorbed over the whole electrode.Following this encouraging observation we investigated the binding of 3, Fcb-16Met3, to a gold disk electrode. However despite intensive eVorts using a wide variety of surface preparations, solvents and immersion times the measured surface coverages obtained were no better than that of peptide 1. Fig. 4 FT-RAIRS (displayed as % transmission spectra) of (a) Ac- 16Met3 self-assembled from a 1 mM solution in methanol; (b) cast film of Ac-16Met3; (c) cast film of Ac-16Nle3.The x-axes units are cm-1. appeared at 1667 cm-1. Furthermore its intensity was significantly less than that of the amide II band. The RAIRS technique allows the orientation of the peptides to be determined as only vibrations with a component normal to the surface are enhanced.52 The main component of the amide I band is stretching of the carbonyl bonds and in a helical peptide these will be aligned along the helix axis. The main contributions to the amide II band are in-plane bending of NKH and CLO, and CKC and CKN stretching.53 For each of these vibrations the transition dipole moment subtends a range of angles to the helical axis.In the self-assembled peptide monolayers binding via the methionine side-chains is expected to align the helical axis parallel to the surface.This is expected Fig. 5 Cyclic voltammograms (CVs) of Cr(Cl4C6O2)33- at (dashed to lead to a dramatic reduction in the intensity of the RAIRS line) a bare gold electrode and (continuous line) gold covered with a amide I band relative to the amide II compared to the spectrum SAM of Ac-16Met3.CVs were recorded at 50 mV s-1 in saturated KCl in 951 methanol–water. The reference electrode was Ag/AgCl. of the cast film. It can be seen in Fig. 4 that this is exactly 1100 J. Mater. Chem., 1999, 9, 1097–1105Other groups have previously found that thioether the peptide’s orientation on the surface. The four other residues on the side of the helix intended to bind to gold were Ala, monolayers failed to completely displace contaminants from which should oVer the least unfavourable steric interactions the gold surface55 and that the preparation on gold of SAMs with the surface.of resorcin[4]arenes through 4 pendant dialkyl thioether arms Using Cys complicates the synthesis as the strongly has been reported to be of low reproducibility due to the nucleophilic thiol must be protected.Other potential diYculties critical conditions required for the electrode preparation and may occur during the subsequent removal of the protecting for self-assembly.31 In contrast the more frequently used group and manipulation of the peptide-dithiol, which may thiolates have been reported to bind strongly to gold and are oxidise to the disulfide.The shorter spacer group between the able to displace contaminants from the surface.55 In model peptide backbone and the sulfur (one methylene group rather experiments with Fcb-derivatives of Met and cystine esters we than two in Met) may also introduce diYculties during the too observed significantly stronger binding of the thiolate over self-assembly step.Unfavourable steric interactions between the thioether.56 We thus decided to investigate the possibility the peptide and surface could prevent the two sulfur atoms of generating SAMs from an oligopeptide containing two Cys from simultaneously accessing ideal binding positions on the residues as surface binding groups, the expectation being that gold surface. we might more readily prepare films with high stability and The peptide was prepared on the solid phase.Both cysteine surface coverage than those prepared from peptides 1 and 3. residues were protected with the trityl group (Trt, triphenylmethyl )57 which may be cleaved either by TFA (tri- Cys-containing peptide fluoroacetic acid) to aVord the thiol or by iodine to form the The sequence of oligopeptide 6, 16Cys2, Ala-Ala-Aib-Ala- disulfide.These methods avoid the use of toxic reagents (HgII, Phe-Ala-Cys-Phe-Leu-Cys-Aib-(NO2)Phe-Ala-Aib-Leu-Ala- TlIII) or thiols; we were particularly keen to avoid the latter OH, where (NO2)Phe is p-nitrophenylalanine, was based on as even small quantities could compete with the peptide for a peptide, PRM1 (Ac-Ala-Ala-Aib-Ala-Phe-Ala-Acc-Leu-Aib- binding to gold.The peptide-resin was stored Fmoc- and Ala-Don-Ala-Aib-Leu-Ala-NH2, where Don and Acc are Glu ditrityl-protected (Fmoc-16Cys2{Trt}2-OR) and subsequent residues modified with electron donating and accepting groups, modifications were carried out as required. From herein we respectively) which was characterised as forming a 310-helix in will refer to the peptide dithiol as 16Cys2(SH)2 and the acetonitrile.10 In contrast to peptides 1 and 3, with peptide 6 peptide disulfide as 16Cys2(S2). molecular modelling (Fig. 6) indicated that binding of the two Cleavage was attempted first with 95% aq. TFA without Cys side-chains to gold would be more favourable in a 310- protection from the atmosphere with triisopropylsilane (TIPS) helix rather than an a-helix.The two Cys residues were thus preferred to ethanedithiol as the scavenger.58 The peptideplaced at positions i, i+3 where we envisaged that formation dithiol was obtained as a yellow and green solid indicating of an intramolecular disulfide bond between them could be that some oxidation of the ferrocene had occurred. ES-MS of used to both stabilise the 310-helical conformation and protect the crude revealed that as well as the desired product peptides the thiol function (e.g.from intermolecular disulfide formation) lacking Aib and Fcb were also present in minor quantities. Attempts to purify the mixture by HPLC in methanol–water until contact with gold cleaved the SKS bond. The end and by crystallisation were unsuccessful and a sample of the sequences of PRM1 were conserved and two modified Glu crude material was adsorbed to the disk electrode and analysed residues with large electron donor and acceptor substituents by cyclic voltammetry, as described below. were replaced with Phe and (NO2)Phe.These were chosen as The cyclisation by disulfide formation of short b-turn- their side-chains are relatively bulky (which might disrupt the forming peptides has been carried out on the resin using iodine hydrophobic aggregation responsible for the low solubility of by Albericio and co-workers, the ‘pseudo-dilution’ conditions 16Met3) but did not require protecting groups, thus keeping favouring the intramolecular process.59 The similarities of the synthesis as simple as possible.The p-nitrophenylalanine these researchers’ sequence to ours (containing 2 Cys at i, residue served the dual purpose of enabling us to measure i+3) and the excellent yields of disulfide-peptide that they accurately the concentrations of the peptide solutions due to obtained after acidolytic cleavage made this an attractive its strong UV absorption and to act as an infrared probe of starting point.Thus Fmoc-16Cys2(Trt)2-O-resin was treated with a solution of iodine (2 equiv.per Cys) in acetonitrile at 0 °C and subsequently deprotected and acylated with Fcb-OH. After cleavage with 9555 TFA–water analysis of the crude material by Electrospray Ionisation Mass Spectrometry (ESIMS) confirmed that the desired peptide had been obtained. It is notable that eVecting the disulfide formation, deprotection and acylation steps in dimethylformamide (DMF) rather than acetonitrile produced undesired intermolecular disulfide bonds.This was established by the presence of 3+ and 4+ ions of the dimeric peptide in the ESI-MS. Other insoluble material may have been higher oligomers. In retrospect this is not too surprising as DMF hydrogen bonds to main chain N–H groups, destabilising the helix and allowing the peptide to access more extended conformations in which the S–S bonds may form between diVerent molecules.Alternatively as a weakly basic solvent DMF may have ‘scrambled’ the initially formed intramolecular bonds. The CD spectra of 6, Ac-16Cys2(S2) in methanol, ethanol and acetonitrile (Fig. 7) indicated that the peptide conformation was a mixture of helix and random coil. The intensities of the minima at 222 nm (-11 000 to -12 000) showed that Fig. 6 Space-filling model of 6, Ac-16Cys2 in a 310-helical confor- the helicity was significantly less than the Met-containing mation showing the alignment of the 2 S atoms on the underside of peptides.However, the unstructured sections of mixed heli- the helix. The main axis of the helix is normal to the plane of the cal–random coil peptides tend to be near the termini,60 and it page. The atom colours are C: green, H: white, N: blue, O: red, S: yellow.was anticipated that since in peptide 6 the dicysteine gold- J. Mater. Chem., 1999, 9, 1097–1105 1101Fig. 9 Cyclic voltammograms of self-assembled films of Fcb-16Cys2X on evaporated gold slides.Dot–dash line: peptide-disulfide (X=S2) Fig. 7 Circular dichroic spectra of peptide 6, Ac-16Cys2(S2) in self-assembled from acetonitrile; dotted line: disulfide (X=S2) selfmethanol (long dashes, 190 mM), ethanol (short dashes, 160 mM) and assembled from methanol; dashed line: dithiol (X=(SH)2) selfacetonitrile (continuous line, 160 mM) at room temperature. assembled from ethanol; continuous line: dithiol (X=(SH)2) selfassembled from methanol.All CVs were recorded at 80 mV s-1 in 0.1 M aq. NaClO4 vs. a Ag wire (Ag/Ag+) reference. binding site lies in the middle it may be unperturbed. Despite our aim of designing a 310-helix the ratio of the intensities of the minima at 208 nm and 222 nm of ca. 1 is more typical of this and the measured anodic charges, which were now signifi- an a-helix; in 310-helices it is often 0.5 to 0.8.61,62 It is probable cantly greater than calculated for a monolayer of 6, we that at 16 amino acids, 16Cys2 is suYciently long that the a- concluded that the adsorbed film was several layers thick.A helix is the favoured conformation. reasonable explanation for this solvent-dependent voltammetry A self-assembled film of Fcb-16Cys2(SH)2 on a gold disk is that the hydrophobic peptide was solvated better by the electrode was prepared using a saturated solution of freshly organic than by the aqueous solvent, allowing greater percleaved peptide in methanol.CV of the unrinsed electrode meation of the (smaller) electrolyte counter-ion into the after 2 days did not initially contain any faradaic waves but adsorbed film and lowering the activation energy of the Fc0/+1 after repeated cycling the oxidation and reduction waves of oxidation.the ferrocene became discernible (Fig. 8). After 10 days storage Films of the ferrocene derivative of 6, disulfide and dithiol in a sealed jar the CV was measured in organic solution (n- were also prepared on evaporated gold that was immersed in Bu4NBF4 in acetonitrile, Fig. 8). As before several potential dilute peptide solutions for 5 days. The adsorbed layer was sweeps were required to condition the film and reveal the characterised by ellipsometry to measure the thickness and ferrocene peaks, which were still unexpectedly broad. From then by CV to measure the surface coverage. The voltammograms, which clearly feature the ferrocene wave, are shown in Fig. 9 and the calculated surface coverage and changes in the ellipsometric angles are summarised in Table 1. As can be seen from the Table the films prepared from ethanol or methanol Table 1 Thickness and coverage of self-assembled films of Fcb-16Cys2 on evaporated gold slides, calculated from ellipsometry and cyclic voltammetry Sample dDa dYa Surface coverage (%)b Dithiol 230 mM 28.4 -2.3 294 MeOH Dithiol 80 mM 9.5 -1.0 101 EtOH Disulfide 210 mM 17.5 -1.1 216 MeOH Disulfide 90 mM 1.7 -0.1 90 MeCN adD=D(bare gold)-D(coated slide), dY was calculated similarly.bCalculated from CV assuming that each peptide molecule covered Fig. 8 Cyclic voltammograms of Fcb-16Cys2(SH)2 film on a gold 43×12 A° of the surface with a surface density of 1.9×1013 disk electrode vs.Ag wire reference electrode. Continuous line: 20 min molecules cm-2. A monolayer of peptide 6 would be 12 A° thick and after rinsing, measured at 50 mV s-1 in 0.1 Maq. NaClO4. Dashed compared to bare gold would decrease D by 1.5 and result in line: 10 days after rinsing, 10th scan at 500 mV s-1 in 0.1 M tetra-n- approximately no change in Y.butylammonium tetrafluoroborate in acetonitrile. 1102 J. Mater. Chem., 1999, 9, 1097–1105were characterised by ellipsometry as being several molecular column. A column temperature of 30 °C was maintained. Each residue was double coupled with a recirculation time of 40 min layers thick, even though the measured charge of the peptide layer that self-assembled from ethanol was that expected for using a 2.5-fold excess of Fmoc-amino acid (Fmoc=fluorenylmethoxycarbonyl ) for 12Leu2 and a 3-fold excess for all a monolayer.This apparent contradiction is reminiscent of the ‘false’ surface coverage measured of the film prepared on other peptides. Acylations were monitored at 348 nm. Fmoc deprotection by 20% piperidine in DMF (dimethylformamide) the gold disk electrode (described above) and probably has the same cause.This film was also characterised by RAIRS. for 15 min was followed at 300 nm. Deprotection of the resinbound Fmoc (PR500 amide resin) or pre-coupled Fmoc-Ala The intensity of the bands in the surface spectrum (not shown) were consistent with a film of multilayer thickness, in (PA500 acid resin) was carried out twice. HPLC was carried out on analytical (250×4.6 mm) and agreement with ellipsometry.In view of these results we concluded that even on a gold semi-prep. (250×10 mm) columns each containing 5 mm C8 Spherisorb stationary phase. Flow rates were 1 ml min-1 for surface protic solvents could disrupt the peptide helix conformation suYciently that one of the thiol groups became avail- the analytical and 5 ml min-1 for the semi-prep.column and detection was at 230 nm unless stated otherwise. Elution was able to form a disulfide bond with a molecule in solution. The aggregation process could continue and lead to the build up by gradients of water with either methanol or DMF (see supplementary information). of a thick film of oligodisulfide on the metal as observed. This hypothesis was given greater credence by the much more Electrospray ionisation mass spectra (ES-MS) were acquired on a Fisons VG Platform. The carrier solvent was 151 successful experiments carried out in non-protic acetonitrile.The changes in the ellipsometric angles and charge passed acetonitrile–water. Circular dichroic spectra were measured using a Jasco J600 spectropolarimeter with a 1 cm path-length for oligopeptide 6, Fcb-16Cys2(S2)-OH self-assembled from acetonitrile solution are shown in Table 1.Strikingly both sets cell. Molecular models were constructed on a Silicon Graphics Iris workstation using the M. S. I. software packages Builder of values are now in good agreement and correspond very well with those calculated for a close-packed SAM of helical and Biopolymer.peptide 6. The successful preparation of monolayers of 16Cys2, therefore appears to require use of solvents which do not Materials. All commercial samples used in this study were used as received. All resins, Fmoc-amino acids and coupling disrupt the hydrogen-bonding of the backbone peptide and thus stabilise the helix conformation. Interestingly this is reagents were of peptide synthesis grade and were supplied by Novabiochem. Acetic anhydride (98%), acetonitrile (for CD, exactly what was found in the earlier formation of the intramolecular S–S bond.EYcient preorganisation of binding sites and spectrophotometric grade), tert-amyl alcohol (99%), carbon tetrachloride (99%), DBU (98%), dichloromethane (reagent eYcient self-assembly onto gold are therefore both key requirements for successful preparation of helical peptide monolayers.grade), DIPEA (99%), lithium bromide (99+%), methanol (for CD, spectrophotometric grade), p-nitrobenzoyl chloride (NB-Cl, 98%) and triisopropylsilane (TIPS, 99%) were pur- Summary and prospects chased from Aldrich. N,N-Dimethylformamide (DMF), diethyl ether, piperidine and trifluoroacetic acid (TFA) were We have outlined a generic strategy for preparing chiral all of peptide synthesis grade and were supplied by Rathburn.functionalised surfaces using SAMs of helical oligopeptides. Acetonitrile and methanol used for HPLC were HPLC grade Those peptides containing Met are mainly helical in organic and were supplied by Merck. Glacial acetic acid and iodine solvents and RAIRS of their SAMs on gold confirmed that (99.5+%) were supplied by Fisons. 4-Ferrocenylbutyric acid they are aligned with the helix axes parallel to the surface. and 3-ferrocenylpropionic acid were gifts from Dr P. However these peptides are of low solubility which leads to Whittaker, University of Strathclyde. handling problems. Their coverage and strength of binding to the surface are also variable probably because the weak Resin washing protocol.The protected peptide-resin was sulfur–gold bond of thioethers means surface preparation is washed sequentially with DMF, tert-amyl alcohol, glacial of utmost importance. Peptides with the thiol-containing Cys acetic acid, tert-amyl alcohol, dichloromethane and diethyl as the gold-binding residue also face potential problems.In ether, dried under high vacuum for at least 5 h and stored particular there is a strong risk of oligomerisation occurring under nitrogen at -20 °C. through formation of intermolecular disulfide bonds, either during S–S bond formation or during self-assembly to the Post-synthesiser deprotection. Protected peptide-resin gold. This is particularly prevalent in protic solvents and can (22 mmol of peptide) was swelled in DMF (4 ml ) for 1 h.lead to formation of thick films. However, if an aprotic solvent Piperidine was added to a ratio of 25% and the mixture was such as acetonitrile is used, the increased stabilisation of the agitated gently. After 5 min standing the resin was filtered helix and resultant preorganisation of the thiols allows eYcient under gravity, washed twice with DMF and stored with 25% self-assembly of close-packed monomeric peptide monolayers.piperidine in DMF for 10 min, swirling occasionally. The resin These monolayers oVer a viable route for preparing funcwas filtered under gravity, washed three times with DMF and tionalised surfaces with nanoscale spatial resolution. As such acylated immediately.they are likely to find applications in fields such as molecular electronics, biosensors, biomaterials and catalysis. Post-synthesiser acetylation. Acetic anhydride (30 ml, 0.3 mmol) and DIPEA (40 ml, 0.2 mmol) were added to the Experimental deprotected peptide-resin (40 mmol of peptide) in DMF (4 ml ). After occasional gentle swirling for 30–45 min the resin was Peptide synthesis filtered and washed twice with DMF.Fresh DMF, acetic anhydride and DIPEA were added to the resin and the General. Oligopeptides were synthesised by a NovasynA Crystal Continuous Flow Peptide Synthesiser. Fmoc amino acylation was repeated. The Ac-peptide-resin was filtered, washed by the same protocol as the post-synthesis washing acids were activated with PyBOPA (benzotriazol-1-yloxytripyrrolidinophosphonium hexafluorophosphate) and N-hydroxy- then either dried under high vacuum and stored under nitrogen at -20 °C or the peptide was cleaved immediately.benzotriazole (HOBt, one equivalent each) and diisopropylethylamine (DIPEA) (2 equivalents for Ala, Met, Fcb, Phe, (pNO2)Phe, Leu and (Trt)Cys, 3 equivalents for Post-synthesiser acylation with 4-ferrocenylbutyric acid.Deprotected peptide-resin (22 mmol of peptide) was treated Aib and Nle) immediately prior to injection onto the reaction J. Mater. Chem., 1999, 9, 1097–1105 1103twice with a solution of 4-ferrocenylbutyric acid (18 mg, and sodium perchlorate (98%) were supplied by Merck. Acetonitrile (HPLC grade) was puchased from Rathburn. 66 mmol), PyBOPA (34 mg, 66 mmol), HOBt (10 mg, 66 mmol ) and DIPEA (35 ml, 200 mmol ) in DMF (5 ml ) for 30–45 min.Chromium tris(tetrachlorocatecholate) was prepared by the method of Pierpont et al.63 The resin was filtered and washed twice with DMF between couplings. After the second acylation the resin was washed and dried as described in the acetylation procedure. Substrate preparation. Gold disk electrodes (2 or 3 mm diameter, supplied by Oxford Electrodes) were manually pol- Post-synthesiser acylation with p-nitrobenzoyl chloride.ished with alumina paste, sonicated in water and dried gently Swelled, deprotected peptide-resin (3 mmol of peptide) was in a stream of nitrogen. The electrode was immersed in 50% acylated by double-coupling for 45 min with p-nitrobenzoyl aqua regia (45351 water–conc.HCl–conc. HNO3) for 60 s, chloride (3 mg, 16 mmol ) and DIPEA (5 ml, 25 mmol ) in DMF water for 10 s, the contacting solvent for 10 s and then the (0.5 ml ). The peptide-resin was washed twice between couppeptide solution. When the contacting solvent was dichloro- lings with DMF and after the second wash was treated in a methane the electrode was dipped in methanol for 10 s between similar manner to the acetylated peptide-resin.the water and dichloromethane rinses. Electrodes prepared in this manner were measured to have surface roughness of 2, Standard cleavage procedure for amide resin. Ac-peptidefrom the anodic oxidation of adsorbed iodine to aqueous resin (26 mg, 9 mmol of peptide) was swelled by washing on a iodate.64 frit with glacial acetic acid, dichloromethane and methanol Evaporated gold substrates were prepared on glass (4 ml each).The resin was transferred to a flask and a solution microscope slides (from Merck) for RAIRS of peptides 1 and of 10% TFA in dichloromethane (total 10 ml ) was added. The 2, silicon wafers (100 orientation, from Micro-Image beads turned red immediately. The mixture was swirled gently Technology Ltd.) for peptides 3 and 4 and on polyacrylamide every 10–15 min for 45 min, the resin was filtered and the gel slabs (from Pharmacia) for 6.The glass microscope slides beads were stored in fresh cleavage mixture for a further were cut to squares of approx. 2.5 cm, cleaned in piranha 45 min. The resin was filtered again and washed with 10% solution (151 H2O2–conc.H2SO4 CAUTION! THESE TFA in dichloromethane (2×10 ml ) and once with methanol SOLUTIONS ARE HIGHLY OXIDISING AND SHOULD (10 ml ). The combined filtrates were diluted with a volume of BE HANDLED WITH EXTREME CARE!) for 1 h, washed acetonitrile equal to that of TFA and concentrated under thoroughly with distilled water and sonicated for 20 min each reduced pressure to an oil.This was triturated with diethyl in acetone and absolute ethanol. The slides remained immersed ether (16 ml ) and the resulting white precipitate was centriin ethanol until they were placed in the evaporating chamber, fuged (3×20 min). The collected solid was either dried under where 3–5 nm of chromium and 40–50 nm of gold were high vacuum overnight and stored under nitrogen at -20 °C evaporated at a base pressure of 2×10-6 Torr and or used immediately to prepare SAMs.were immediately transferred to the contacting solutions. The other gold slides were prepared under semi-conductor Standard cleavage procedure for acid resin. Triisopropylsilane technology clean room conditions. Polyacrylamide gel slabs (TIPS, 13 ml, 66 mmol ), methanol (0.25 ml ) and TFA (4.75 ml ) were cut into strips 20×70 mm and placed on an Al mask so were added to peptide-resin (82 mg, 16 mmol of peptide).The that the coated area was of 2 rectangles 18×16 mm connected mixture was shaken occasionally and after 1–2 h was filtered. by a strip 2 mm wide and 28 mm long. Si wafers were either The resin was washed with 9555 TFA–methanol (2×5 ml) used whole or cut into strips 8 mm wide.The substrates were and 151 TFA–methanol (5 ml ). The combined filtrates were cleaned by sonication in either IPA or water–detergent then concentrated, precipitated with diethyl ether, centrifuged and in water and dried in a stream of high-purity N2. Cr and Au dried according to the same protocol as used for the amide were evaporated at a rate of 1 A° s-1 to thicknesses of 100 and resin. 1000 A° . The slides prepared on silicon wafers were immersed in piranha solution for 45 min, rinsed thoroughly with distilled Protocol for oxygen-free cleavage. Peptide-resin was water and dried in a stream of N2. All other slides were used transferred to a fritted column equipped with a tap at the without further treatment. bottom and a nitrogen inlet at the top.The apparatus was charged with nitrogen and the resin was washed sequentially with acetic acid, dichloromethane and methanol. Degassed Preparation of peptide and amino acid films. All solutions of TFA and dichloromethane or methanol were added and the peptides 3, 4 and 6 were degassed except for the SAM of 6 on column was sealed under a nitrogen atmosphere, gently agitat- a gold disk electrode.After immersion the substrates were ing the column every 10–15 min. After 45 min the solvents rinsed to remove excess peptide, unless stated otherwise in were filtered, freshly degassed cleavage solution was added the text. and the mixture allowed to stand for a further 45 min. The solvents were filtered and the resin was washed with cleavage mixture and with methanol.The solution of cleaved peptide Electrochemistry was removed via syringe and concentrated under reduced CV was controlled by an AutolabA potentiostat from Eco pressure to an oil. This was subsequently treated in the manner Chemie via General Purpose Electrochemical System 3 on a described in the standard cleavage procedure. PC. A one-compartment cell was used. Oxygen was not excluded. The counter electrode was a coiled Pt wire, supplied Peptide purification and characterisation.Available as by Goodfellow. The reference electrodes used were Ag/AgCl supplementary material. (Oxford electrodes) for voltammetry of peptides 1 and 2, a Ag wire coated in AgCl (a gift from C. Agra-Gutierrez) for CV Preparation and characterisation of self-assembled monolayers of Ru(NH3)6Cl2 and K3[Cr(Cl4C6O2)3] and a silver wire (Ag/Ag+) for all other experiments. The electrolytes were Materials.Absolute alcohol (reagent grade), hexaamineruth- 10 or 100 mM tetra-n-butylammonium tetrafluoroborate in enium(II) chloride, sodium tetrafluoroborate (98%) and tetraorganic solvents, saturated KCl in 951 methanol–water for n-butylammonium tetrafluoroborate (99%) were supplied by K3[Cr(Cl4C6O2)3] and aqueous solutions of 0.3 M KCl for Aldrich.Acetone, hydrogen peroxide (30 vol.%) and conc. electrochemistry of Ru(NH3)6Cl2, 0.1M NaClO4 or 0.1M hydrochloric, nitric and sulfuric acids were all reagent grade and were supplied by Fisons. Potassium chloride (AnalaRA) NaBF4. 1104 J. Mater. Chem., 1999, 9, 1097–1105L. S. Curtin, S.R. Peck and R. W. Murray, Electrochim. Acta, Reflection–absorption infrared spectroscopy (RAIRS) 1995, 40, 1331. Infrared spectra were recorded on a Unicam Galaxy Series 23 L. H. Dubois and R. G. Nuzzo, Annu. Rev. Phys. Chem., 1992, 43, 437. FTIR 7000 equipped with a liquid nitrogen-cooled Hg–Cd–Te 24 A. Ulman, Ultrathin Organic Films; From Langmuir-Blodgett to detector and an FT85 specular reflector from Spectra Tech Self-Assembly, Academic Press, Boston, 1991.Inc. The sample compartment was purged with dry air. The 25 W. B. Caldwell, K. Chen, B. R. Herr, C. A. Mirkin, J. C. Hulteen angle of incidence was at 85° to the surface normal and the and R. P. Van Duyne, Langmuir, 1994, 10, 4109. incident radiation was plane-polarised parallel to the direction 26 C.M. Yip and M. D. Ward, Langmuir, 1994, 10, 549. of propagation. Reflection–absorption spectra were accumu- 27 X. Tang, T. Schneider and D. A. Buttry, Langmuir, 1994, 10, 2235. 28 L. Haussling, H. Ringsdorf, F.-J. Schmitt and W. Knoll, lated over 200 scans at 4 cm-1 resolution and were ratioed Langmuir, 1991, 7, 1837. against an unused gold slide or a control slide that had been 29 W.S. V. Kwan, L. Atanasoska and L. L. Miller, Langmuir, 1991, immersed in the same solvent. 7, 1419. 30 J. M. Tour, L. Jones II, D. L. Pearson, J. J. S. Lamba, Ellipsometry T. P. Burgin, G. M. Whitesides, D. L. Allara, A. N. Parikh and S. V. Atre, J. Am. Chem. Soc., 1995, 117, 9529. Ellipsometry measurements were made on a PC-controlled 31 E. U. Thoden van Velzen, J. F. J.Engbersen, P. J. de Lange, Auto Gain Ellipsometer L116B from the Gaertner Scientific J. W. G. Mahy and D. N. Reinhoudt, J. Am. Chem. Soc., 1995, Corporation, Chicago. The light incident at 70° to the normal 117, 6853. 32 H. Adams, F. Davis and C. J. M. Stirling, J. Chem. Soc., Chem. was at a wavelength of 632.8 nm from a He–Ne laser. The Commun., 1994, 2527. optical properties of the substrate and of the adsorbed layer 33 R.P. H. Kooyman, D. J. van den Heuvel, J. W. Drijfhout and were calculated using the classical 2- and 3-phase parallel layer G. W.Welling, Thin Solid Films, 1994, 244, 913. models. The real component of the refractive index of the 34 C. Duschl, M. Liley, G. Corradin and H. Vogel, Biophys. J., 1994, films, nF, was assumed to be 1.500. The average of four 67, 1229.recordings were made at each of three positions per slide. The 35 C. D. Bain, J. Evall and G. M. Whitesides, J. Am. Chem. Soc., 1989, 111, 7155. surface area sampled by the laser was 2 mm2. The optical 36 M. Knichel, P. Heiduschka, W. Beck, G. Jung and W. Gopel, characteristics of the bare gold were calculated from the Sens. Actuators B, 1995, 28, 85. measurements made on the same slide before immersion 37 C. Duschl, A. F. Sevinlandais and H. Vogel, Biophys. J., 1996, in solution. 70, 1985. 38 M. Boncheva and H. Vogel, Biophys. J., 1997, 73, 1056. 39 J. K. Whitesell and H. K. Chang, Science, 1993, 261, 73. Acknowledgements 40 S. Sakamoto, H. Aoyagi, N. Nakashima and H. Mihara, J. Chem. Soc., Perkin Trans. 2, 1996, 2319. We thank the EPSRC for funding, Novabiochem, Dr S. Kelly 41 R. Naumann, A. Jonczyk, R. Kopp, J. van Esch, H. Ringsdorf, at the BBSRC Circular Dichroism Facility, University of W. Knoll and P. Graber, Angew. Chem., Int. Ed. Engl., 1995, Stirling, Dr R. Raval, E. Cooper and L. Shorthouse at 34, 2056. University of Liverpool for assistance with RAIRS measure- 42 L. Scheibler, P. Dumy, D. Stamou, C. Duschl, H. Vogel and M. Mutter, Tetrahedron, 1998, 54, 3725. ments and Dr L. Berlouis at University of Strathclyde for 43 A. E. Strong and B. D. Moore, Chem. Commun., 1998, 473. advice on ellipsometry. 44 I. L. Karle and P. Balaram, Biochemistry, 1990, 29, 6747. 45 C. Toniolo and E. Benedetti, Trends Biochem. Sci., 1991, 16, 350. 46 F. Donald, G. Hungerford, D. J. S. Birch and B. D. Moore, References J. Chem. Soc., Chem. Commun., 1995, 313. 47 G. P. Dado and S. H. Gellman, J. Am. Chem. Soc., 1993, 115, 1 C. A. Mirkin and M. A. Ratner, Annu. Rev. Phys. Chem., 1992, 12609. 43, 719. 48 S. Marqusee, V. H. Robbins and R. L. Baldwin, Proc. Natl. Acad. 2 R. W.Munn, Biosystems, 1992, 27, 207. Sci. USA, 1989, 86, 5286. 3 F. L. Carter, Molecular Electronic Devices II, Marcel Dekker, 49 S. P. Ho and W. F. DeGrado, J. Am. Chem. Soc., 1987, 109, 6751. New York, 1987. 50 R. W. Woody, in The Peptides: Analysis, synthesis, biology, 1985, 4 A. Aviram and M. A. Ratner, Chem. Phys. Lett., 1974, 29, 277. Vol. 7, p. 15. 5 J. J. Hopfield, J. N. Onuchic and D. N. Beratan, J. Phys. Chem., 51 P. I. Haris and D. Chapman, Biopolymers, 1995, 37, 251. 1989, 93, 6350. 52 R. J. Greenler, J. Chem. Phys., 1966, 44, 310. 6 H. Tachibani, T. Nakamura, M. Matsumoto, H. Komizu, 53 S. Krimm and J. Bandekar, Adv. Protein Chem., 1986, 38, 181. E. Manda, H. Niino, A. Yabe and Y. Kawabata, J. Am. Chem. 54 M. D. Porter, T. B. Bright, D. L. Allara and C. E. D. Chidsey, Soc., 1989, 111, 3080. J. Am. Chem. Soc., 1987, 109, 3559. 7 C. Joachim and J. P. Launay, J. Mol. Electronics, 1990, 6, 37. 55 E. B. Troughton, C. D. Bain, G. M. Whitesides, R. G. Nuzzo, 8 M. D. Ward, Chem. Soc. Rev., 1995, 24, 121. D. L. Allara and M. D. Porter, Langmuir, 1988, 4, 365. 9 M. Mutter and R. Hersperberger, Angew. Chem., Int. Ed. Engl., 56 A. E. Strong, Ph. D. Thesis, University of Strathclyde, 1997. 1990, 29, 185. 57 G. B. Fields and R. L. Noble, Int. J. Pept. Protein Res., 1990, 10 G. Hungerford, M. Martinez-Insua, D. J. S. Birch and 35, 161. B. D. Moore, Angew. Chem., Int. Ed. Engl., 1996, 35, 326. 58 D. A. Pearson, M. Blanchette, M. L. Baker and C. A. Guindon, 11 R. A. Bissell, E. Cordova, A. E. Kaifer and J. F. Stoddart, Nature, Tetrahedron Lett., 1989, 30, 2739. 1994, 369, 133. 59 F. Albericio, R. P. Hammer, C. Garcý�a-Echeverrý�a, M. A. Molins, 12 L. Zelikovich, J. Libman and A. Shanzer, Nature, 1995, 374, 790. J. L. Chang, M. Pons, E. Giralt and G. Barany, Int. J. Pept. 13 M. P. Debreczeny, W. A. Svec, E. M. Marsh and R. Wasielewski, Protein Res., 1991, 37, 402. J. Am. Chem. Soc., 1996, 118, 8174. 60 S. M. Miick, K. Casteel and G. L. Millhauser, Biochemistry, 1993, 14 R. R. Birge, Sci. Am., 1995, March, p. 66. 32, 8014. 15 U. P. Wild, S. Bernet, B. Kohler and A. Renn, Pure Appl. Chem., 61 T. Iwata, S. Lee, O. Oishi, H. Aoyagi, M. Ohno, K. Anzai, 1992, 64, 1335. Y. Kirino and G. Sugihara, J. Biol. Chem., 1994, 269, 4928. 16 C. D. Bain and G. M. Whitesides, Angew. Chem., Int. Ed. Engl., 62 W. L. Fiori, S. M. Miick and G. L. Millhauser, Biochemistry, 1989, 28, 506. 1993, 32, 11957. 17 N. L. Abbott and G. M. Whitesides, Langmuir, 1994, 10, 1493. 63 C. G. Pierpont, H. H. Downs and T. G. Rukavina, J. Am. Chem. 18 K. L. Prime and G. M. Whitesides, Science, 1991, 252, 1164. Soc., 1974, 96, 5573. 19 G. M. Whitesides and P. E. Laibinis, Langmuir, 1990, 6, 87. 64 J. F. Rodriguez, T. Mebrahtu and M. P. Soriaga, J. Electroanal. 20 L.-H. Guo, J. S. Facci and G. McLendon, J. Phys. Chem., 1995, Chem., 1987, 233, 283. 99, 8458. 21 L. A. Hockett and S. E. Creager, Langmuir, 1995, 11, 2318. 22 J. N. Richardson, G. K. Rowe, M. T. Carter, L. M. Tender, Paper 8/06860G J. Mater. Chem., 1999, 9, 1097–1105

ISSN:0959-9428    

DOI:10.1039/a806860g

出版商:RSC    

年代:1999

数据来源: RSC