摘要:

IntroductionIn this study, we have measured the equilibrium constants of proton-transfer reactions in the gas phase in order to determine a quantitative scale of thermodynamic stabilities of aryldimethylsilyl anions, which allows us to analyze in detail substituent effects on their stabilities.Despite the extensive use of silyl anions as synthetic reagents to form silicon–carbon bonds,1as polymers, and as precursors in semiconductor manufacturing, the thermochemistry of silyl anions2,3has been studied much less than that of carbanions.4Silyl anions, like carbanions, are solvated in solution and they are associated with alkali metal cations. Gas-phase experiments, in the absence of solvent, allow the determination of the intrinsic properties of the anions through the measurement of their acidities, so that the relative stabilities of the anions can be compared. In the gas phase, silyl anions are much more stable than carbanions,e.g., ΔG°acid(CH4) = 408.5 kcal mol−1(ref.4)vs.ΔG°acid(SiH4) = 363.8 kcal mol−1(ref.4) (1 cal = 4.184 J). This is related to correspondingly higher electron affinities [EA(SiH3) = 32.4 kcal mol−1(ref.2and5)vs.EA(CH3) = 1.8 kcal mol−1(ref.6)]. In addition, the acidity of silane was found to decrease by approximately 3–5 kcal mol−1with successive methyl substitution,2,7,8in contrast to that for methane, where methyl substitution increases the acidity of methane, except for the first methyl substitution.9The substitution of a phenyl group was shown to have essentially no effect on the acidity of silanes,2suggesting that π-delocalization does not play a role in the stabilization of silyl anions. This is also in contrast to the observed π-effects in carbanions. We sought more detailed knowledge about the stabilization mechanism in silyl anions.In this paper we report the gas-phase acidities for a series of ring-substituted dimethylphenylsilanes determined by measuring the equilibrium constants of the proton transfer reaction(1). The substituent effects on the stabilities of silyl anions are compared with those for carbanions.Ab initiocalculations are also used to aid our understanding of the stabilization mechanisms for silyl anions.1

ISSN:1472-779X    

DOI:10.1039/b100488n

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

IntroductionParamagnetic materials have been investigated as MRI contrast agents (CAs) for more than two decades.1These materials enhance the contrast of the image indirectly by lowering the magnetic relaxation time of the water protons in the surrounding tissues.2,3Gd(III) is particularly suited for this purpose because of its favourable magnetic properties (seven unpaired electrons). The aqua ion itself is too toxic for human use. This can be circumvented by chelation by a polydentate ligand. The most frequently used Gd(III) based CAs are stable, hydrophilic poly(aminocarboxylate) derived complexes with rapid extracellular distribution and renal elimination. Depending on the denticity one or more water molecules might be directly coordinated to the paramagnetic centre.Gd complexes with amphiphilic properties have previously been synthesised and evaluated as blood-pool and liver imaging agents. Long chain amides and esters of DTPA (diethylenetriaminepentaacetic acid) are the most common.4More recently amphiphilic complexes based on the DO3A (1,4,7-tris(carboxymethyl)-1,4,7,10-tetraazacyclododecane) structure have been reported.5,6These complexes are able to form supramolecular systems such as micelles, mixed micelles and liposomes in aqueous solutions, in the presence or absence of surfactants and phospholipids. The formation of these systems increases the efficacy (T1-relaxivity) of the contrast agent due to an increase in the rotational correlation time (τR) of the Gd complex.In the present work Gd–DO3A complexes mimicking the phospholipid structure have been prepared in order to achieve CAs able to form stable and rigid supramolecular systems. Macrocyclic chelates were chosen for the studies because of their favourable stability and hence reduced probability for intracellular dechelation.7

ISSN:1472-779X    

DOI:10.1039/b100405k

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

IntroductionRecently, weak interactions such as hydrogen-bonding, π–π interactions, and charge-transfer interactions have attracted much attention from organic, biological and physical chemists,1because they play very important roles in supramolecular aggregates. Small-sized metacyclophanes, which are characterized by a specific transannular interaction, could be unique candidates for the study of such weak inter- and intramolecular interactions.2Böckmann and Vögtle reported preparative methods for various kinds of dithia[3.3]metacyclophanes (MCPs) and related compounds.3These MCPs seem to be very suitable for the investigation of weak interactions since two aromatic rings and the substituents on the MCPs are located in close proximity. Böckmann and Vögtle studied the transannular interactions in these compounds by means of their electronic spectra,4however, a detailed study of the molecular and crystal structure of dithia[3.3]MCP in terms of the weak interactions has not yet been carried out.In this paper, we report the preparation of dithia[3.3]MCPs having electron-releasing and -withdrawing substituents on their internal positions and discuss their structural properties in order to evaluate the weak interactions. Furthermore, their molecular and crystal structures were studied from the viewpoint of intra- and intermolecular π–π interactions or hydrogen-bonding properties.

ISSN:1472-779X    

DOI:10.1039/b100285f

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

MethodsAb initiomolecular orbital calculations were carried out using the Gaussian 94 19or Gaussian 98 20program. Geometry optimisations were performed using standard gradient techniques at the SCF, MP2 and B3LYP levels of theory using restricted (RHF, RMP2 and RB3LYP) and unrestricted (UHF, UMP2 and UB3LYP) methods for closed and open shell systems, respectively.21Further single-point QCISD and CCSD(T) calculations were performed on each of the MP2 optimised structures. When correlated methods were used, calculations were performed using the frozen core approximation. Except for MP2/aug-cc-pVDZ, whenever geometry optimisations were performed, vibrational frequencies were calculated to determine the nature of located stationary points. Calculations were performed on all reactants, products and transition states to obtain barriers and energies of reaction. Where appropriate,zero-point vibrational energy (ZPE) corrections have been applied. Values of ⟨s2⟩ never exceeded 0.86 before annihilation of quartet contamination and were mostly below 0.79 at correlated levels of theory.Standard basis sets were used. In addition, the (valence) double-ζpseudopotential basis set of Hay and Wadt 22supplemented with a single set ofd-type polarisation functions was used for the heteroatoms in this study (exponentsd(ζ)Si = 0.284,23d(ζ)Ge = 0.230 23andd(ζ)Sn = 0.200) while the double-ζall-electron basis sets of Dunning and Hay 24with an additional set of polarisation functions (exponentsd(ζ)C = 0.75 andp(ζ)H = 1.00) were used for C and H. We refer to this basis set as DZP throughout this work.13,14,18Calculations were performed on DEC AlphaStation 400 4/233, DEC Personal Workstation 433au or 600au, Compaq DS10 or NEC SX-4 computers.Optimised geometries and energies for all structures in this study (Gaussian Archive entries) are available as Supplementary Material.

ISSN:1472-779X    

DOI:10.1039/b100162k

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

IntroductionCyclodextrins (CDs) are truncated, cone-shaped, cyclic oligosaccharides composed of six or more α-1,4-linked glucose moieties.1–3CDs present a hydrophobic cavity with their primary hydroxy groups on the narrow side of the macrocycle and the secondary hydroxy groups on the opposite side.3–6They can function as chiral hosts, and for this behaviour they have been widely studied as receptors for a large variety of molecules 2,7–9and as molecular carriers.1,2,8Furthermore, CDs have received considerable attention as models mimicking the behaviour of biological macromolecules.2,10CDs possess as functional groups only hydroxy groups. Consequently, the introduction of other functional groups on their skeleton can modify and possibly improve some of their features, such as solubility, stability and selectivity, when forming inclusion complexes.1,2,8Among known CDs, β-CDs (or cyclomaltoheptaoses), have been the object of many investigations. A large number of monofunctionalized β-CDs have been described and characterized in solution 11–16as well as in the solid state 17–21and their applications as metal enzyme models or chiral discriminating agents, as well as their inclusion and catalytic abilities have been investigated.Substitution of more OH groups at desired positions by different chemical functions can lead to multisite recognition systems, in which the hydrophobic cavity and different substituent groups as recognition elements can be involved. Therefore selectivity and enantioselectivity can be greatly increased and modulated.22–25A few 6-difunctionalized cyclodextrins have already been described as molecular receptors 22,23,26or models for the mechanism of action of biomolecules.27–32In particular, the functionalization by metal-ion-complexing groups has led to systems of undiscussed interest.12–14,28,33–36Ferredoxin,30haemoglobin-like,27,29carbonic anydrase,32superoxide dismutase (SOD),34and P-450 cytochrome models 37have been built using difunctionalized CDs. Recently, the crystal structure of 6I,6II-diamino-6I,6II-dideoxy-β-cyclomaltoheptaose and its complex with platinum has been reported by us.38,39They represent the only examples of difunctionalized β-cyclodextrins structurally characterized by single-crystal X-ray diffraction and NMR. The molecule is able to complex a metal ion, and shows chiral recognition ability 33towards amino acids. The functionalized cyclodextrins show different properties with respect to other regioisomers.33This suggests that the improvement in the structural and conformational properties on more sophisticated difunctionalized β-CDs can represent a very important step in the rational design of this class of compounds.Here, we report the synthesis and the first structural characterization by X-ray diffraction analysis of a branched difunctionalized β-cyclodextrin, namely 6I,6II-dideoxy-6I,6II-bis[2-(2-pyridyl)ethylamino]-β-cyclomaltoheptaose (ABAEPY) (Fig. 1).Schematic representation of the 6I,6II-dideoxy-6I,6II-bis[2-(2-pyridyl)ethylamino]-β-cyclomaltoheptaose molecule with numbering of the atoms.

ISSN:1472-779X    

DOI:10.1039/b100126o

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

IntroductionCharacterization of the excited singlet states of pyrazolotriazole (PT) azomethine dyes is of considerable importance in view of their extensive use as photographic image-forming magenta dyes. The 7H-pyrazolo[5,1-c][1,2,4]triazole skeleton,1(Fig. 1), was first introduced by Bailey to take advantage of the excellent color reproduction properties of the respective dyes in a subtractive color system.1Their particularly noteworthy feature is a negligible unwanted secondary absorption in the blue region. The isomeric skeleton, 7H-pyrazolo[1,5-b][1,2,4]triazole,2, was also reported to bring forth comparable magenta dyes.2Both dye types typically exhibit a high absorption coefficient and an oscillator strength approaching unity. Such absorption spectrum features are likely to resultfrom the planar structure assumed by the PT dyes in the ground state as demonstrated by X-ray crystallography 3andab initioquantum chemical calculations.4Molecular structures and atom numbering of the PT dyes.Whereas the structure and properties of the ground state of PT dyes appear to be relatively well understood, very little conclusive evidence for the excited singlet state structures is available. In terms of their photochemical behavior, PT dyes have proved to be puzzling molecules. Only a few studies dealing with excited singlet state properties of PT dyes have been published. Steady state fluorescence measurements reported by Douglas reveal structureless emission with a quantum yield ofca. 10−4at room temperature in low viscosity solvents.5Low temperature and high viscosity conditions were reported to lead to a considerable increase in the fluorescence quantum yield.6A relatively small Stokes shift of the fluorescence observed for both classes of PT dyes can be taken as an indication of the similarity between the ground and excited state geometries. Accordingly, similar oscillator strengths maybe expected for the emission and absorption. Based on these assumptions, a lifetime ofca. 10 ps for the fluorescent state can be estimated from the fluorescence quantum yield at a room temperature in glycerol–ethanol (96 ∶ 4), which is in good order-of-magnitude agreement with the results of the direct lifetime measurements using the photon-counting method.6However, the lifetime measurements were reported to yield nonmonoexponential fluorescence decays, which required two or three exponentials to attain a satisfactory fit. It should be noted that, although this seemingly implies the involvement of two or three separate fluorescent transients, other interpretations might be equally plausible. Furthermore, it is known that the lifetimes obtained from nonmonoexponential decays may not even have any physical significance.7Even more puzzling is the pronounced mismatch between changes in fluorescencequantum yields and changes in fluorescent lifetimes, which may imply partitioning between several short-lived transients.6Several groups have applied picosecond absorption recovery techniques to study excited singlet states of PT dyes.6,8Unlike the fluorescence lifetime measurements, the absorption recovery techniques permit the detection and lifetime measurement of nonfluorescent transients. Although, in principle, both fluorescent and nonfluorescent transients can be detected, their assignment has proven to be ambiguous for the PT dyes. Interestingly, multiexponential kinetics of the absorption recovery sensitive to the probe wavelength were generally observed, which had been interpreted in terms of several transients of obscure nature formed either sequentially or in parallel.To sum up, despite the considerable experimental efforts by means of a variety of absorption and emission spectroscopic techniques, conclusive evidence for the excited singlet state structures of PT dyes is still lacking. On the other hand, substantial progress in the development of the theoretical methods applicable to excited state calculations, particularlyab initiomethods such as CI-singles, multiconfigurational self-consistent field (MCSCF ), and coupled-cluster has been achieved in the past decade.9These methods have proved to be quite useful in elucidating photochemical reactions.10For the PT dyes, the lowest singlet excited state has been briefly examined using semiempirical PM3-CI 11andab initioCI-singles 12methods. However, such calculations frequently yield several different models describing a given photochemicalreaction. Thus, the assignment of the excited state structures and information about photochemical reaction mechanisms can only be successfully obtained by experimental and theoretical efforts in conjunction. In the present study we have used steady state fluorescence in conjunction with CI-singles and MCSCF computational methods to study the nature and behavior of the excited singlet state of PT dyes.

ISSN:1472-779X    

DOI:10.1039/b010165f

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

IntroductionCompounds containing quinazolinone ring systems exhibit a large variety of biological effects,1such as antiinflammatory,2antimalarial,3anticonvulsive 4and hypotensive 5activities. Once it was recognized that substituents at the 2 and 3 positions modulate the hypertensive and antiinflammatory activities,5,6a large number of 2-alkyl-3-arylquinazolin-4(3H )-ones were synthesized and screened to develop new drugs. However, in many of these studies it was not recognized that the 2-alkyl-3-arylquinazolin-4(3H )-ones were stereochemically analogous to theorthosubstituted biphenyls and therefore isolatable rotational isomers might exist. The first observation of restricted internal rotation in the 3-aryl C–N bond was reported in 1975 based on1H NMR evidence.7Later it was shown that these types of atropisomers 8are resolvable and stable at room temperature but racemize at elevated temperatures.9We report here activation parameters for atropisomerization of (S )-3-(2-chlorophenyl)-2-[2-(6-diethylaminomethylpyridin-2-yl)vinyl]-6-fluoroquinazolin-4(3H )-one1(Scheme 1) and propose a mechanism for the process. It has been shown in the literature that the biological effects of atropisomers can be different 10and because the atropisomers can interconvert during synthesis and formulation development, we determined the activation parameters in order to estimate the atropisomerization rate constants at different temperatures.Atropisomerization of 3-(2-chlorophenyl)-2-[2-(6-diethylaminomethylpyridin-2-yl)vinyl]-6-fluoroquinazolin-4(3H)-one (1).

ISSN:1472-779X    

DOI:10.1039/b010090k

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

Results and discussionSynthesisThe syntheses of1and8are outlined in ref.6but their detailed experimental procedures are given here in the Experimental. Compounds2–7were synthesised by conventional palladium(0)-catalysed C–C cross coupling reactions of a terminal alkyne (Hagihara coupling) or of trialkylstannylalkynes (Stille coupling) with the appropriate aromatic halide derivative as outlined inScheme 1. Quaternisation of the amine or pyridine nitrogen followed for2,4,5,6and7. For the synthesis of3, the already quaternised (bromophenyl)triphenylphosphonium salt was used and coupled with the tin derivative. This procedure was checked in order to provide a general method for the synthesis of substituted tetraphenylphosphonium salts. However, palladium-induced scrambling of the aryl substituents 22can occur and might be the reason for the low yield. The yields for all the other C–C coupling reactions range betweenca. 50 and 95% per step. In many cases, the overall yield was decreased considerably by the successive incomplete quaternisation and counter ion exchange steps. The choice of more reactive reagents for quaternisation was limited due to the susceptibility of the tolane C&z.tbd;C triple bond to electrophilic attack.The phosphonium ion9was directly prepared by palladium-catalysed coupling of triphenylphosphine with the iodoarene derivative. Again, scrambling of the phosphorus substituents might be the reason for the low yield. In general, the purification of the ion pairs proved to be difficult due to extremely broad and overlapping fractions on the chromatography column.Linear and nonlinear optical propertiesThe UV–Vis spectra of all the compounds1–9were measured in THF and MeCN in order to estimate the trend of solvatochromism. These solvents proved to be very suitable because they readily dissolve most ion-pairs, they are both aprotic and their polarity in terms of their Dimroth–Reichardt parameters are quite different.23InTable 1the data for the CT bands of1–10in THF and MeCN are given. The zwitterions1and8display a pronounced negative solvatochromism, which indicates a large ground state and a small excited state dipole moment. Fromeqn. (1)it is evident that a large dipole moment difference will increase the hyperpolarisability. However, the solvatochromism of the cations2–4is distinctly smaller and even absent in5–7. At first glance one might conclude that the dipole moment difference for2–4and5–7is quite small and, consequently, the quadratic hyperpolarisability is small, too. However, it will be explained later, on the basis of semiempirical calculations, that this conclusion is wrong.Linear and nonlinear optical data forp-nitroaniline (pNA) and1–10λTHF /nmλMeCN /nmϵMeCN /cm−1M−1Δ&z.ngrt;/cm−1 |β 0|/10−30esu M βM relative to pNAEnergy difference between the absorption maxima in THF and MeCN.Static hyperpolarisability in MeCN,βzzzfor1–5,8,9and10,βyyyfor6.Molecular mass used of cations only.In CHCl3.In MeOH, see ref.9.Calculated by tensor addition from theβvalue of5.Value from the two-level model, see footnote ‡.See ref.31a.See ref.6.pNA 36236630013.6138.1&z.tbd;11 34932927100−174019 (18) 417.50.47236736030000−53038547.30.70340639531100−69087666.31.34446345435300−430128558.82.34535435434100041463.60.906454454762000127990.41.317354354145000032 1313.80.25 8 2532377900−26705317.40.169326 (331) 32220000 −38022 501.50.45 10 270 27000 8 90.20.91 Although in MeCN theD3symmetric species6absorbs at the same wavelength as its one-dimensional counterpart,4, the band width at half-height is somewhat smaller (4140 cm−1) than in the 1D species (4330 cm−1). This leads to a long-wavelength tailing of the CT band of4compared to6. The same behaviour has also been observed in the 1D and 2D species of thep-nitrophenyl and thep-tricyanovinylphenyl derivatives (instead of the pyridinio acceptor).24The ratio of oscillator strength of4and6is only 1 ∶ 2.04 while a ratio of 1 ∶ 3 would have been expected in the case of negligible interaction between the chromophore branches of6. Deviation towards a smaller ratio has also been observed for tricyanovinyl-substituted triphenylamines 19fand for triarylphosphonium cations.19cHowever, the ratio of the oscillator strengths of species5and7isca. 1 ∶ 4.3, much closer to the expected ratio of 1 ∶ 4. We conclude that the interaction in the triarylamine branches of6is much stronger than in the tetrahedral methane derivative7, although the conjugation in6is diminished by the propeller-like arrangement of theN-phenyl substituents.25The quadratic hyperpolarisability has been measured in MeCN at 1500 nm. The long incident wavelength has been chosen so as to avoid two- and three-photon induced fluorescence which otherwise might interfere with the HRS signal.21b,cAs the HRS method allows only the square of the hyperpolarisability tensor to be determined, the sign of the hyperpolarisability is unknown but assumed to be negative because of the negative solvatochromism of1–4,8and9. For1–5,8and9only theβzzztensor component out of 18 possible components is significant; for6it isβxxy = −βyyy . For compound7it was impossible to measure theβvalue because of its poor solubility in MeCN.The corresponding static hyperpolarisabilities have been calculated by the frequency dependence term of the two-state model [eqn. (1)] and are given together with the linear optical data inTable 1.InFig. 1the quadratic hyperpolarisabilities of1–5are plotted against the reciprocal square of the CT energy asβshould increase with this term according toeqn. (1)as confirmed by Stiegmanet al. 26for a series of tolanes substituted with conventional donors and acceptors (D/A-tolanes). For comparison these data are included. Our series of tolane derivatives1–5also shows an approximate linear correlationvs.1/ωeg2. This means that theμeg2(μe − μg) term ineqn. (1)is fairly constant and that the two-level model is valid for both series of compounds,i.e.an increase ofβis solely due to an increase in absorption wavelength. However, the slope of series1–5(11.44 × 10−20) is much steeper (factor 2.7) than that of the D/A-tolane series (4.30 × 10−20),i.e.1–5display a much better nonlinearity–transparency trade-off than the D/A-tolane species.Because tolanes substituted with conventional donors and acceptors show a positive solvatochromism, these chromophores will likely have somewhat higherβvalues in dipolar MeCN than in the less polar CHCl3, thus reducing the ratio of the slopes. On the other hand, species1–4show negative solvatochromism and, consequently will show somewhat higherβvalues in CHCl3than in MeCN which would result in a higher ratio of the slopes. Therefore, measurements in solvents in which both series of compounds display their lowest hyperpolarisabilities seem to be a reasonable basis for a fair comparison.Consequently, theμeg2(μe − μg) term must be higher in the1–5series than in the D/A-tolanes for a given CT energy. Becauseμeg2varies approximately in proportion to the molar absorptivity,ϵ, and the average value forϵis 24700 for the D/A-tolanes and 31500 for1–5, this difference can only account for a factor ofca. 1.3 of the slope. Therefore, the major effect is likely to originate from a higher (μe−μg) difference in1–5compared to the D/A-tolanes.Hyperpolarisabilityvs.reciprocal square of the CT energy of1–5(values from Table 1) and a set of tolanes substituted in 4,4′-position with conventional electron donors and acceptors from ref.26(measured in CHCl3).The most important observation is that the intercept of the correlation lines with thexaxis coincides for both series and is at the absorption wavelength of unsubstituted tolane (300 nm), which, owing to its centrosymmetry, has a vanishing quadratic hyperpolarisability. This is because both series can be regarded as tolane, more or less weakly perturbed by substituents or by replacement of an aromatic CH group by anN +-alkyl moiety. Therefore, the absorption energy of the unsubstituted parent chromophore dictates the limit of transparency for a series of chromophores. Though trivial we have not seen this point made clear in the literature before.Analogous plots of our data (seeFig. 2) for species8–10compared to benzene substituted by conventional donors and acceptors (D/A-benzenes) taken from Chenget al. 27also show linear correlations. The intercept with thexaxis is at the energy of the1Lbband of benzene (200 nm). Again, the slope of the correlation line for8–10(3.27 × 10−20) is much steeper (factor 4) than that of the D/A-benzene series (0.80 × 10−20) which proves that the (zwitter)ionic chromophores8–10show a much better nonlinearity–transparency trade-off than their conventional counterparts. However, compared to both tolane series, the benzene-type chromophores are inferior because of the inherently lower polarisability of benzene compared to tolane. A similar conclusion for stilbenes has been drawn before by Chenget al. 27.Hyperpolarisabilityvs.reciprocal square of the CT energy of8–10(values from Table 1) and a set of benzenes substituted in the 1,4-position with conventional electron donors and acceptors from ref.27(measured in dioxane).We stress that the linear correlations observed should be confined to systems where the substituents (be they ionic or conventional) induce a weak perturbation in the parent chromophore. In cases where the response to a perturbation is strong,e.g.in polyenes, and/or with stronger electron donors or acceptors, cyanine structures may originate from mixing donor and acceptor type states which then show vanishing hyperpolarisability at decreasing CT energies (increasing 1/ωeg2term). This type of behaviour expresses itself in a reduced bond order alternation and has thoroughly been examined by Marderet al.28However, the force constant of the C&z.tbd;C triple bond in tolane and the aromaticity of benzene counteract the electron-donating and -accepting influences of the substituents, which results in only a weak perturbation.HRS measurement of the two-dimensional chromophore6gave aβ0yyyvalue of 127 × 10−30esu (Table 1). If one ignores interactions of the tolane subchromophore units in6, tensor addition of theβzzzvalue of three subchromophores in the proper orientation (taken from4: 128 × 10−30esu) results in 0.75 × βzzz (4) = 96 × 10−30esu. Thus, the observed value of 127 × 10−30esu isca.13higher than expected. We and other groups have observed enhancements of hyperpolarisability in higher-dimensional chromophores compared to their one-dimensional counterparts.19c,d,f,hThe increases were interpreted in terms of excited state couplings, which is likely also the case for the triarylamine6in the present study.In terms of their hyperpolarisability to molecular weight ratio (β/M, seeTable 1), which is often used as a figure of merit, chromophores3,5and6, but especially4, perform well compared top-nitroaniline (pNA) as the standard. The value of6could be further improved if thetert-butylphenyl substituents, which have been introduced for solubility reasons, were replaced by smaller groups.It is often stated that higher-dimensional NLO-chromophores such as6and7are inferior to one-dimensional analogues owing to their high molecular mass relative to theirβvalues that are in the range of those of the 1D species. However, this is not true at all: for applications in the crystalline state one has to consider the orientation of the chromophores in the crystal lattice; owing to phase-matching conditions this leads to effective SHG coefficientsdwhich are much smaller per molecule than theβvalue of an isolated chromophore. For example, the maximal possible relative SHG coefficient in the crystal point groups 1 (triclinic), 2,m(monoclinic) andmm2 (orthorhombic) isd = 2/(3√3) ≅ 0.38.29In other point groups the maximaldvalue is smaller,e.g.in 3 (trigonal) it is ¼ and in 4 (tetragonal) it is 1/(3√3) ≅ 0.19.29The latter point groups are those which can be adopted by6and7, respectively. This means—if one presumes additivity for the subchromophore branches in7—that a crystal built from7(in the tetragonal point group 4) can only show ½ of thedvalue of a crystal built from5in the monoclinic point groupm. However, if one takes the molecular mass (seeTable 1) into account, this difference levels completely out becausedrel(5)/(4 × 558.8) ≅ drel(7)/(1313.8). This is because four subchromophores in7share one “donor” which makes the molecular mass of7much smaller than four times the mass of5. The same holds true for4and6. If subchromophore interactions increase the molecular hyperpolarisability compared to its one-dimensional analogue as ine.g.6, the relativedcoefficient of a crystal of a higher-dimensional chromophore might be much higher than that of its analogous 1D chromophore.Semiempirical calculationsIn order to get a closer insight into the polarisation mechanism of the tolane chromophores studied in this paper, we performed semiempirical calculations at the NDDO level using the AM1 parametrisation. The excited state properties of species1–5and of a dianisylamino- and nitro-substituted tolane11were calculated at the CISD level with an active window comprising the four highest occupied and the four lowest unoccupied orbitals. The hyperpolarisability tensor was calculated using the time-dependent Hartree–Fock method (TDHF ).30Both the linear and the nonlinear optical properties of dipolar chromophores are usually quite sensitive to the solvent.31Although this fact has been known for quite a long time, with some exceptions 32most quantum chemical calculations on nonlinear optical properties still refer to the gas phase. We modelled the influence of the medium using the COSMO method introduced by Klamt and Schüürmann 33because this model involves a solvent accessible surface modelled by van der Waals radii rather than a spherical or ellipsoidal cavity. Thus, this model should be more suitable for the rod-like molecules employed in this study.32e,kInTable 2the AM1 CISD computed absorption energies are given which deviate strongly from the observed ones (Table 1) with an unsigned mean deviation of 4750 cm−1. This error reduces significantly to 2850 cm−1when computing the absorption energies in MeCN. However, the error is still large compared to the rather small range of absorption energies of1–5. The quadratic hyperpolarisability was computed by two entirely different methods. The first is based on the time-dependent Hartree–Fock theory,i.e.the time-dependent response to an oscillating electric field is calculated analytically. The second method is based on the two-level approximation [eqn. (1)] where only the first excited singlet state is used in a truncated sum-over-states expansion. For this purpose we calculated the dipole moments of the ground state, the dipole moment and the energy of the first excited singlet (CT) state as well as the transition dipole moment by a CISD expansion in MeCN. This method has been employed by Zhanget al.34with much success. The computed staticβzzzvalues are given inTable 2together with the dipole moment differences. Although no solvatochromism was observed for3–5the computed changes of dipole moment upon excitation are quite substantial (seeTable 2) and no qualitative difference can be seen between zwitterion, ion-pair and neutral chromophores. Thus, the assumption that the absence of solvatochromism indicates a vanishing dipole moment difference is false, at least for the ion-pairs of this study.AM1 computed linear and nonlinear optical properties of1–5and11in MeCN. Values in italics refer to the gas phaseλmax /nm&z.ngrt;max/cm−1fμe − μg / Dβ 0zzz(TDHF ) /10−30esuβ 0zzz(two-level) /10−30esuCalculated by the AM1 TDHF method.Calculated by the two-level approximation [eqn. (1)] using AM1 CISD computed data.Exp. value 413 nm.Exp. value of diphenylamino derivative in CHCl3: 28.2 × 10−30esu, see ref.17a.Exp. value, see footnote ‡.1357280000.768−12.7 (−18) −30.1−23.9 (−18) 503199000.660−23.4−166.0−106.22356281000.862−13.4−37.7−28.1405247000.636−22.0−93.8−50.23338296000.490−23.0−51.2−35.8350286000.612−30.0−95.3−42.54397252000.721−18.5−69.6−44.9520192000.565−22.0−214.5−94.35309324000.863−14.0−20.2−19.1309324000.118−39.2−63.3−7.411353 283000.70421.138.6 35.4343292000.99312.632.727.2Comparison of the AM1 calculatedμe − μgdifference of1(−12.7) with the experimental one (−18 D) The ground state dipole moment of1, which isμg = 38 ± 4 D, has been determined by dielectric relaxation spectroscopy 43in MeCN. The excited state dipole moment of1has been measured by a solvatochromic method. As the absorption energies of1show a very good linear correlation with the Dimroth–ReichardtE NTsolvent parameters 23we used theseE NTvalues for an empirical solvent characterisation instead of more complicated solvent functions based on the permittivity and/or the index of refraction with Δ&z.ngrt; = K  ΔE NT.44The unknown constantKwhich absorbs all theoretical inaccuracies is 1.90 × 104and can be evaluated by usingμg,μe − μganda(the effective solute diameter) of the Dimroth–Reichardt dye (μg = 14.8 D,μe − μg = −8.7 D anda = 6.0 Å).45The ratio Δ&z.ngrt;/ΔE NTrefers to the slope (6.45 × 103) of a linear correlation of the absorption energies of1in 9 different solvents with the appropriateE NTparameters.46Forawe used the AM1 calculated B–N distance of1(12.7 Å). With these values we evaluatedμe − μg = −18 D for1which lies in between the AM1 computed values for the gas phase and the MeCN solution. Using the two-level approximation [eqn. (1)] withμeg = 6.0 D (from the integrated absorbance of the CT band) andμe − μg = −18 D we estimated the quadratic hyperpolarisability for1to be −18 × 10−30esu in MeCN, which is in excellent agreement with the HRS measurement.shows that the computation underestimates the dipole moment difference in MeCN solution but overestimates this difference in the gas phase (−23.4 D). On the other hand, comparison of the experimentalμe − μgdifference of 4-dimethyamino-4′-nitrotolane (5.7 D) 26with those of the AM1 calculated value of11indicates that the calculation overestimates theμe − μgdifference both in MeCN and in the gas phase, with the MeCN value being much higher than the gas phase value. These effects are due to the negative and positive solvatochromism of1and11, respectively and are also due to the inability of the COSMO procedure to model correctly the solvent influences quantitatively. Thus, it seems likely that the zwitterions and ion-pairs actually show a higherμe − μgdifference than conventional D/A-tolanes as anticipated previously.As can be seen fromFig. 3there is a fairly good linear correlation of the hyperpolarisability calculated by the TDHFvs.the two-level method for1–5and11, the slope being 0.74. Thus, the hyperpolarisability estimated by the two-level model is about14smaller than the TDHF values, which indicates that the first excited state does indeed play the dominant role, but higher excited states are necessary for a complete description of the hyperpolarisability. Similar conclusions have been drawn by Markset al. 35from careful analysis of SOS expansions.Hyperpolarisabilities of1–5and11calculated by the two-level approximationvs.TDHF calculated values in MeCN. The solid line is a linear correlation through the origin.However, comparison of the TDHF-computed hyperpolarisabilities in the gas phase and in MeCN with the experimental ones in MeCN shows no satisfactory correlation (seeFig. 4). While the theoretical MeCN values (signed mean error −21 × 10−30esu, unsigned 25 × 10−30esu) are close to the experimental ones for1and2, they appear to be much too small for3–5. In contrast, the TDHF gas phase values (signed = unsigned error = +68 × 10−30esu) are much too high for1,2,4and5but in good agreement for3. In general, the gas phase values are higher than those in MeCN. The good correlation ofβ(TDHF )vs.β(two-level) indicates that the main problem of correctβcomputations is not the TDHF procedure nor the CISD expansion nor the AM1 parametrisation but the model for the solvent calculation. One way out of this problem might be to parametrise the COSMO model for calculating accurateβvalues by adjusting the relative permittivity empirically. If, for example,ϵr = 3.0 is used, the signed error reduces to 6 × 10−30esu but the unsigned error still is 20 × 10−30esu. Of course, for a reasonable parametrisation, a much larger data set is necessary, which will be the topic of future investigations.AM1-TDHF computed hyperpolarisabilities of1–5in the gas phase, MeCN and in a solvent with with the permittivityϵr = 3.0vs.experimental values in MeCN. The straight line has a slope = 1.In order to investigate the process of charge transfer upon excitation we plotted the AM1 CISD Coulson charge differences between ground and excited state for1–5and11inFig. 5. On going from column 2 through column 11, it is apparent that the charge transfer is alternate (negative, positive) upon excitation. The donor substituents (separated into a central atom and its ligands R) as well as the C atoms in columns 4 and 6 lose negative charge upon excitation whereas those at the C atoms in columns 9, 11 and 12 gain charge. The acetylene bridge also shows a strong change of charge density. The change of charge density is marginal at column 10 and, especially, at the “acceptor” side: the triorganoammonio and phosphonio substituents in1–3stabilise the negative charge density in the adjacent benzene ring but do not accept charge density themselves. In this way, these substituents are electronically very similar to theN-alkylpyridinio group in4and5. In the pyridinio derivatives the electron acceptor of course is confined to the pyridine moiety as theN-alkyl group lies in the σ-plane. This permits theN-alkylpyridinio and trialkyl-ammonio and -phosphonio derivatives to be regarded asone classof tolanes. Much in contrast, the nitro group of11gains negative charge upon excitation and, thus, increases the size of the “active” π-electron system, which, consequently, reduces the CT energy leading to a worse nonlinearity–transparency ratio as demonstrated inFig. 1for the D/A-substituted tolanes.AM1 CISD computed Coulson charge differences upon excitation for1–5and11. The charge differences refer to the sum of all atoms drawn in each column. A positive charge difference indicates a loss of negative charge upon excitation. X = C for1–3and X = N for4and5. The entries in each column are from left to right:1,5,3,4,2,11.In1and5, the ligands R lose most of the charge density of the donor side, which is due to hyperconjugation whereas in2,3and4, it is the amine nitrogen of the dianisylamino group which loses most of the electron density, which in turn reflects the resonance effect. Although2,3and4possess the same type of donor substituent, the charge density change is quite different at the C-atom next to the donor in column 3. This demonstrates that the acceptor substituent has a strong effect on the change of charge density even far away from its position in the vicinity of the donor!It is startling to see the triphenylmethyl group in5acting as a donor substituent. Its Hammett constantσp = 0.02 36indicates almost neutral behaviour. We performed AM1 CISD model calculations on three species for comparison: in compound12, the triphenylmethyl group in5has been replaced by a hydrogen atom, in compound13theN-methylpyridinium moiety in5has been replaced by benzene, and, finally, unsubstituted tolane14. The calculated linear and nonlinear optical data are collected together with those for5inTable 3. From the data inTable 3one can see that substitution of a hydrogen in tolane14by a triphenylmethyl group only results in a quite smallβvalue of the tolane derivative13. The dipole moment difference upon excitation also is marginal (2.0 D). In contrast, substituting a hydrogen in the pyridinium compound12by a triphenylmethyl group to yield5results in a large increase ofβand in a quite substantial dipole moment difference (14 D). This shows that the triphenylmethyl group itself is a weak donor, as indicated by its almost vanishing Hammett constant, but readily releases negative charge in combination with a strong acceptor like a pyridinium group. Thus, in5, the triphenylmethyl moiety serves as as an electronsourcerather than an electrondonor.AM1 computed linear and nonlinear optical properties of5and12–14in MeCNλmax /nm&z.ngrt;max/cm−1fμe − μg / Dβ 0zzz(TDHF ) /10−30esuβ 0zzz (two-level) /10−30esuCalculated by the AM1 TDHF method.Calculated by the two-level approximation [eqn. (1)] using AM1 CISD computed data.5309324000.863−14.0−20.2−19.112297337000.611−4.3−12.4−3.713285351001.1192.01.23.014261383000.6130.00.00.0

ISSN:1472-779X    

DOI:10.1039/b009664b

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

IntroductionThe wavelengths calculated by the Pariser–Parr–Pople molecular orbital (PPP MO) method have been shown to be shorter than the observed values for molecules possessing a large π-conjugated system. A novel two-centre electron repulsion integral, new-γ, was proposed to improve the accuracy of the PPP calculations. New-γhas a spectrochemical softness (SCS) parameterkrs.2For the calculations of thep-bands (nomenclature by Clar:3corresponds to Platt's1Laband 4) of polycyclic aromatic hydrocarbons (PAHs), we have established the proper methods to evaluate the SCS parameterkrsin new-γ, as shown in our previous works.1,5–8In these works the π-conjugated system of linear polyacenes was defined to bean aromatic sextet resonance system (ASRS), and the SCS parameter was set to increase with the degree of annelation.5We showed that PAHs possessed a spectroactive portion (SP), which was defined to be the portion that contributes mainly to the electronic spectra of PAHs,6,7and the SPs of thecata-condensed PAHs were the longest acene-like portions in the molecular frameworks. The SCS parameters evaluated from the length of the SP reproduce well the values obtained from PPP calculations of the excitation energies of thep-band ofcata-condensed PAHs 6,7and non-alternant PAHs, such as a fluoranthene.1Thep-band is identified as a π–π* transition described by a mainly HOMO→LUMO transition, so it appears in the longest wavelength region except in the case of relatively small PAHs.3Clar named the absorption bands in the long wavelength region as the α-band (Platt's1Lbband 4) and the β-band (Platt's1Bbband 4), as well as thep-band.3The α-and the β-bands absorb shorter wavelength energy than thep-band except for in the case of small PAHs.3We have applied the single SCS parameterkrsto every C–C interaction of given PAHs for the calculation of excitation energies using new-γ.1,5–8The adoptedkrsvalues are suitable for the calculation of excitation energies of thep-band, but thekrsvalues are not necessarily the most suitable for the calculation of excitation energies of the α- and/or the β-bands.In this paper, we investigate the use of the SCS parameterkrsin new-γfor the simultaneous calculation of the excitation energies of the α- and the β-bands, as well as of thep-band for typicalcata-condensed PAHs, namely, benzo-annelated polyacenes (Fig. 1).General formula of the benzo-annelated polyacenes. The benzo-annelated positions are indicated by the letters A–D.

ISSN:1472-779X    

DOI:10.1039/b009661j

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

IntroductionDuring the last decade the importance of computational chemistry and chemometrics has increased significantly. These techniques have had a substantial impact on the developments in several fields of chemistry. In particular, the field of medicinal chemistry has benefited immensely from the predictive power of techniques such as docking and quantitative structure activity relationships (QSAR).Docking is especially useful for fitting molecules into cavities, such as active sites of biological targets. An essential requirement to perform a docking search is knowledge of the molecular structure of the binding site. Docking has been successfully employed for studying the interaction energies in protein–ligand complexes,1and also plays an important role inde novodrug design.2Whereas docking makes use of the molecular structure of the binding site, QSAR ignores the binding site and focuses entirely on the molecular structure of the ligands. In contrast to docking QSAR only requires knowledge of the molecular structure of a series of active and non-active ligands for a given binding site. Starting with a set of known ligands QSAR utilizes various statistical methods to derive relationships between the molecular structure of the ligands and their affinity for the binding site. QSAR is based on the assumption that similar molecules exhibit similar properties.This paper deals with the problem of identifying new molecules that form clathrates with the cephalosporin antibiotic cephradine. Clathration of cephalosporins is a valuable methodology for the isolation of these important antibiotics from aqueous solutions.3,4A drawback of the currently known complexants, which are all naphthalene derivatives, is their toxicity and the inherent environmental image problem associated with these compounds.5,6Hence, it is highly relevant from an industrial point of view to identify novel complexing agents having environmentally more acceptable properties. Moreover, complexants will be used in enzymatic cephalosporin synthesis. Due to possible enzyme inhibition by complexants, access to a variety of clathrating agents is desirable. Thus, a predictive model for identifying new complexants will be very helpful in addressing the above-mentioned problem. In these clathrates, the cephradinemolecules form the hosting framework in which the aromatic compounds are hosted. In addition, a number of water molecules are accommodated in these complexes as well. Based on the crystal structure of the clathrate type complex of cephradine and 2-naphthol,7which is depicted inFig. 1, a large series of novel complexing agents for the cephalosporin antibiotics could be identified, using chemical intuition as the main lead. However, due to the capricious behaviour of molecules regarding complex formation with cephradine, the successful selection of a new molecule for a complexation experiment remains a matter of trial and error. Therefore, it is desirable to replace this empirical method by a more rational, preferably unbiased, design and prediction of new complexing agents.The complex of cephradine and 2-naphthol.At first sight docking seems to be an appropriate technique for fitting a guest molecule in the cavity formed by cephradine molecules. However, the hosting framework formed by cephradine is rather flexible and exhibits a remarkable adaptability.6In addition, the available space for the guest molecule is dependent on the number of water molecules incorporated in the complex.6As a consequence, an exact cavity for conceivable guest molecules cannot be defined, which is an essential requirement for the docking method. As QSAR ignores the structure of the binding site and only makes use of the structure of the guest molecules, this method is suitable to derive a model that can predict the complexing behaviour of a molecule. Application of QSAR for the prediction of clathration has no precedent except our preliminary report on this subject.8

ISSN:1472-779X    

DOI:10.1039/b009629f

出版商:RSC    

年代:2000

数据来源: RSC