摘要:

J. Chem. Soc., Faraday Trans. 1, 1982, 78, 1785-1794 Dielectric Properties of Pro tein-Me thylglyoxal Adducts Interfacial and Bulk Effects BY STEPHEN BONE AND RONALD PETHIG* Laboratory of the National Foundation for Cancer Research at the School of Electronic Engineering Science, University College of North Wales, Dean Street, Bangor, Gwynedd LL57 1UT Received 22nd June, 198 1 Steady-state electrical conduction and dielectric measurements over the frequency range 1 0-5- 1 O5 Hz are reported for polycrystalline samples of protein-methylglyoxal adducts. The results are found to be influenced by sample heterogeneity effects associated with a chemical instability that results in the formation of resistive layers at sample surfaces exposed to air. The conductivity data confirm earlier reports regarding the action of methylglyoxal in increasing the electrical conductivity of proteins, and an analysis of possible interfacial effects indicates that the increased conductivity and observed dielectric dispersions are associated with a true bulk phenomenon.Evidence to support the possibility that proteins can form charge-transfer complexes with electron-acceptor or electron-donor molecules was first obtained from conduction and spectroscopic measurements on chloranil-albumin and chlorophyll-albumin complexe~,~-~ and more recently it has been proposed4 that unsaturated ketones and aldehydes (e.g. methylglyoxal) can act as electron-acceptors with respect to protein structures. When proteins such as bovine serum albumin (BSA), casein, collagen and lysozyme are reacted with methylglyoxal, they assume a stable brown colour and exhibit an increased steady-state conductivity and electron spin resonance (e.s.r.) compared to the untreated p r o t e i n ~ .~ - ~ A distinct electrical feature of such protein- methylglyoxal adducts is the existence of a large dielectric loss peak at low frequencies. Methylglyoxal has been founds to occur naturally in beef liver in a form bound to membrane proteins and, although methylglyoxal can react with terminal a-amino groups and the side-chains of arginine, lysine and ~ysteine,~ the primary reaction most relevant to the e.s.r and electrical studies appears to involve a Schiffs base (imine) linkage between methylglyoxal and the &-amino groups of the protein lysine residues.lo Stereochemical investigations'l have shown that such a Schiff s base linkage can bend back to the N-atom of the neighbouring peptide residue in such a way that a charge-transfer interaction is possible, and this concept is supported by ab initio molecular-orbital calculations.ll Szent-Gyorgyi4 has drawn attention to the possible significance of methylglyoxal in the control of cell division and cancer, and he proposes a regulatory role for the ubiquitous enzyme glyoxalase, whose purpose in actively converting methylglyoxal to D-lactic acid is not understood.In brief, it is considered that methylglyoxal can act as an electron acceptor in charge-transfer reactions with membrane-associated proteins and that the electron 'holes' so formed can become delocalised in the electronic ground states of the protein structure.Such a process is envisaged to lend to these proteins a sub-molecular electronic reactivity which is essential for the full maintenance of normal cellular activity. 17851786 P R 0 TE I N-MET H Y LG L Y OX A L ADD U C T S Before the steady-state conductivity and dielectric data can be used as evidence to support the concept of methylglyoxal being able to form a charge-transfer complex with proteins, the data should first be shown to characterise a true bulk effect in the test samples. The studies reported here attempt to distinguish bulk phenomena from those that could occur as a result of interfacial and sample inhomogeneity effects. EXPERIMENTAL BSA (fraction V) and egg-white lysozyme (grade 1) obtained from Sigma Chemicals were separately suspended in the dark at 310 K in 10 volumes of methanol containing 10% neutralised methylglyoxal solution.The methylglyoxal (Aldrich 40 %) had previously been twice distilled under vacuum and stored in 20% aqueous solution acidified with HCI to reduce polymerisation. The suspensions were incubated for periods ranging from 2 to 6 days, after which the proteins were separated on a filter. Before being washed in methanol and dried in a vacuum desiccator, the samples were dialysed at 275 K against two changes of distilled and deionised water to remove any unreacted methylglyoxal. For collagen (bovine Achilles tendon, Sigma) the samples were suspended in 0.1 mol dm-3 dichloroacetic acid with neutralised methylglyoxal, and the addition of methanol resulted in the precipitation of a brown collagen-methylglyoxal adduct.In all preparations pH adjustment was made using NaOH or HCl and no buffer solutions were employed. Scintillation counts using 14C-labelled methyl- glyoxal and weight increase measurements indicated that the protein-methylglyoxal adducts contained on average 8- 10 % methylglyoxal on a dry-weight basis.12 The resultant polycrystalline powders were compressed at 1.3 x lo* Pa into discs of surface area 1.3 x m2 and then retained in the dark between two spring-loaded polished copper electrodes in a temperature-con trolled and vacuum- tight electrical measurement cell. Steady-state conductivity and dielectric measurements for the frequency range 2 x 1 0-5 to 1 O5 Hz were made using techniques and methods of analyses that have been fully described elsewhere for work on organic complexes and protein l4 RESULTS AND DISCUSSION As reported previously5? the brown protein-methylglyoxal adducts exhibit larger steady-state conductivities than do untreated proteins in similar ambient conditions.For the temperature range investigated (200-330 K) the conductivity CJ of the adducts followed an activated law of the form CJ = a,exp(-AE/kT) (1) with the activation energy AE exhibiting a transition from a low to a higher value at ca. 260 K as demonstrated in fig. 1 for a BSA-methylglyoxal sample. A summary of the conductivity data is given in table 1, together with the values of the relaxation time, z, and limiting low- and high-frequency relative permittivity, E,, values that characterised the low-frequency dielectric dispersion exhibited by the adduct0 The dielectric relaxation time, z, followed an activated law z = z,exp( W/kT) and as exemplified in fig.1 the activation energy W remained constant over the temperature ranges investigated. The data in table 1 refer to test samples of nominal thickness 1 x m. The low-frequency dielectric dispersion typically observed for the methylglyoxal- treated proteins is given in fig. 2 for the case of BSA. Changing the surrounding atmosphere from vacuum to dry air did not influence the form of the dielectric loss peak and neither did the insertion of thin discs of brass or aluminium between theS . BONE AND R. PETHIG - - 1787 10: 10 -10 n - I E v) . W e -11 -l2l - 13 ‘ 0 ‘0 ‘ 0 ‘0 I I I , I 30 3.5 4 0 4 5 50 lo3 KIT FIG.1.-Variation of conductivity, 0, a, and dielectric relaxation time, T, 0, with reciprocal temperature for BSA-methylglyoxal. TABLE 1.-CONDUCTIVITY AND DIELECTRIC DATA FOR THE DRY PROTEINS AND DRY PROTEIN- METHYLGLYOXAL ADDUCTS. THE PARAMETERS ARE DEFINED IN EQN (1) AND (2) AND IN THE TEXT. protein-met h ylglyoxal untreated parameter proteins collagen BSA, lysozyme ~ ~~~ ~ o/S m-l, 297 K 10-15-10-18 3 x 10-10- 10-13 3 x 10-11-2 x 10-13 AEIeV, T > 260 K 1.4kO.l 1.15 & 0.15 0.65 f 0.05 T < 260 K not measured 0.30 f 0.05 0.25 f 0.1 W/eV - 1 . 1 f O . l 0.55 f 0.1 E,/ 105 HZ 3.2 k 0.1 4.2 f 0.2 4.5f0.5 z/s, 297 K - 250-500 10-30 E,/ 10-5 HZ 3.3 f 0.2 50-80 1 5-40 test samples and the copper electrodes.The loss peak remained in existence when thin sheets of poly(ethyleneterephtha1ate) or mica were placed between the electrodes and the sample. These results suggested that the dielectric properties were not associated with effects involving the sample-electrode interfaces. Although the change from vacuum to dry atmosphere produced no immediate alteration of the dielectric properties, re-examination of the compressed disc samples after a period removed from the electrical measurement cell has produced lower conductivity and dielectric loss values. Protein-methylglyoxal powders exposed to air assume a pale orange colour, and samples that have been stored for several months in air in closed containers exhibit low conductivities that approach those of untreated proteins.This indicates that the protein-methylglyoxal interaction that produces the increased electrical conductivity is not stable and is possibly influenced by oxygen.1788 10 € I f - P R 0 TE I N-ME TH Y LG L Y OX A L ADD U C TS 8 8" '8 8 8 8 %5 - 4 -3 log,, CfiHZ) FIG. 2.-Frequency dependence of the dielectric loss factor E" at 294 K. x , Untreated BSA in vacuum; 0, BSA-methylglyoxal in vacuum; 0, BSA-methylglyoxal in dry air. / /e 3 I I E -12 - --. v 0 0 - - 2 -13 . 6s +It+ I I 0 2 4 6 8 tlmm FIG. 3.-Variation of 294 K conductivity with sample thickness. 0, BSA-methylglyoxal (5 x lo4 V m-l); 0, collagen-methylglyoxal (1.5 x lo4 V m-l). The inset shows a proposed sample model with bulk conductivity ob and two surface layers of thickness 6s and conductivity os.Recent e.s.r. studies15 of the BSA-methylglyoxal adduct have indicated that in the presence of oxygen the proposed Schiff s base linkage can undergo oxidation to form a semidione. Since the sample preparation procedure and the mounting into the electrical measurement cell led to a finite exposure to air, the possibility existed that the test samples were not homogeneous in their electrical properties and that the surfaces in contact with the electrodes could be of lower conductivity than the bulk region. To test for this, measurements were made as a function of sample thickness, and typical examples of the conductivity and dielectric loss data obtained are shown in fig. 3-5. The results were found to have a significant dependence on sample thickness.S. BONE AND R.PETHIG 80. 60. 1789 I 20 I 0 o o o 0 0 0 0 0 0 0 0 0 x x x x x 8 X X X 0 0 X 0 X X x X 0 log,, CflHZ) FIG. 4.-Frequency variation of the dielectric loss factor E" at 294 K as a function of sample thickness for collagen-methylglyoxal. Sample thicknesses (mm): 0, 6.45; x , 3.61 ; A, 1.67; 0, 1.05; 0, 0.63. - L - 3 - 2 -1 FIG. 5.-Frequency variation of the dielectric loss factor E" at 294 K as a function of sample thickness for BSA-methylglyoxal. Sample thicknesses (mm): 0, 1.87; x , 1.34; A, 1.09; 0 , 0.69; 0, 0.42. Iog,, (flHz) The conductivity results of fig. 3 can be understood in terms of the three-layer model shown in the inset to fig. 3, for which the overall sample conductivity o is given by Os o = [a,ds+o,(t -2Ss)l (3) where o b and os are the conductivities of the bulk and surface regions, respectively, t is the total sample thickness and 6s is the mean thickness of each surface layer.The results of fig. 3 are reasonably described in terms of eqn (3) using the following parameters : BSA-methylglyoxal: o b = 5 x lo-'' s m-I, os = 9 x s m-I, 6s = 2.2 x lop5 m; collagen-methylglyoxal: ob = 3 x S m-I, os = 8 x S m-', 6s = 1.9 x m. (4)1790 PROTEI N--M ETHY LGL Y OX A L A DDU C TS This indicates that the samples were inhomogeneous to a considerable extent, with their outermost surfaces poorly conducting by comparison with their bulk regions. Also, the applied field stress across the sample would not have been uniform, with the fields E, and Eb at the surface and bulk regions, respectively, being given by eqn (5) and (6): 17- ( 5 ) " "b E, = [ 26s0, + as( t - 26s)l where V is the applied voltage across the sample.The electrical data for fig. 3-5 were obtained for constant values of the applied field (1.5 x lo4 and 5 x lo4 V m-l for the collagen and BSA samples, respectively), so that with increasing sample thickness the effective field E, across the resistive outer surfaces increased to a value > lo6 V m-l. An improved fit of eqn (3) to the conductivity data could be made if the surface conductivity a, was assumed to have a field dependence of the form of fig. 6. By contrast the field Eb across the sample bulk was considerably less than the assumed value V / t . Note that the plots of fig. 6 appear to represent the high-field regions of sinh (qEa/2kT) relationships, and as such are consistent with a model involving the field-assisted transport of charges q across potential-energy barriers of width a.From the gradients of fig. 6 values for a of 0.7 and 1.6 nm can be derived for the BSA and collagen samples, respectively, and such barrier widths are reasonable for hopping charge transport in protein -1 3 h - I E 2 -14 vl W 0 M - o-r----o--o -1 5 lLj 0 5 10 15 20 E,/106 V rn-l FIG. 6.-Variation of the conductivity 6, of the sample surface layer of thickness 6s as a function of the effective field stress E, across 6s. Values for E, were calculated using eqn (5). 0, BSA-methylglyoxal; a, collagen-methylglyoxal. Although the conductivity data can be used to show that the interaction with methylglyoxal results in an increase of the electrical conductivity of the reacted protein, the heterogeneous nature of the samples does, however, complicate interpret- ations of the dielectric data of fig.4 and 5. The possibility arises that the observed dielectric loss peaks were associated with interfacial polarisations of the so-called Maxwell-Wagner type. Basically, charge diffusing across the bulk region can accum- ulate at the resistive surface layers and the dynamic behaviour of this interfacialS . BONE AND R. PETHIG 1791 build-up of charge as a function of the applied a.c. field can give rise to a significant dielectric dispersion. Several theoretical analyses of this phenomenon have been given in the literature and, in particular, van Beek16 has shown that the heterogeneous structure of fig.3 will exhibit an interfacial dielectric dispersion which can be described in terms of the standard Debye equations. Following van Beek’s theory and using the data of eqn (4) and fig. 6 the interfacial dispersions expected for the BSA and collagen complexes are as shown in fig. 7. Two important features of fig. 7 are that with increasing sample thickness the dispersion and its characteristic relaxation time are expected to increase in magnitude. The physical basis for this is that with increasing sample thickness the number of charges that can accumulate at the resistive surface layers increases and the time for their diffusion from one sample surface to the other also increases. Although the frequency ranges and dispersion magnitudes of fig.7 are of the same order as the experimental results given in fig. 4 and 5, the change of relaxation time with sample thickness was not observed. log,, CflHZ) FIG. 7.-Interfacial dispersions based on the theory of van Beek16 for the sample model of fig. 3 and the data given in eqn (4). Collagen-methylglyoxal: (a) t = 6.45 mm, (b) t = 0.63 mm; BSA-methylglyoxal: (c) t = 1.87 mm, ( d ) t = 0.42 mm. As detailed elsewhere13 the dielectric loss dr values are derived from the experimental data using formulae that involve the factor (V/t)-l, and as such sample homogeneity is assumed. The field Eb at the sample bulk was considerably less than V / t because of the resistive surfaces, and this should have led to an underestimate of the true bulk values for E”. Furthermore, this underestimate should have decreased as the sample thickness increased, and this could explain the results of fig.4 and 5 . The E” values of fig. 4 and 5 have been recalculated using the relationship together with the data of eqn (4) and fig. 6 to correct for the voltage drop that occurs across the high-resistance surface layers; the results obtained are given in fig. 8 and 9. The use of eqn (7) leads to the dielectric loss being independent of sample thickness and, particularly for the BSA complex, to E” values considerably greater than those obtained assuming the samples to be homogeneous. This in turn gives support to the previous conclusions6* that, because of the large magnitude of the loss peaks, their interpretation should be considered in terms of the hopping of charge carriers over potential-energy barriers rather than as a relaxation of an ensemble of molecular dipoles.We are then led to consider that the protein-methylglyoxal samples can effectively be represented by a model in which conducting protein regions are1792 P R OTE I N-METHY LG LY O X A L ADD U C T S 250r 200 150 El' 100 50 - - - - 8 X X X d X Q X I 1 1 % -3 - 2 -1 log,, (flHz) FIG. 8.-Dielectric loss data of fig. 4 for collagen-methylglyoxal corrected using eqn (7) to take account of voltage drop across the resistive surface layers. Symbols as for fig. 4. 200 - 1 5 0 0 f" 100 I 0 0 X 8 1 I I -3 - 2 - 1 log,, V/Hd FIG. 9.-Dielectric loss data of fig. 5 for BSA-methylglyoxal corrected using eqn (7) to take account of voltage drop across the resistive surface layers.Symbols as for fig. 5. separated by poorly conducting barriers, with the mobile charges hopping from one conducting region to the next. KoopsL7 has used such a model to describe the dielectric properties of ferrites, and it is of interest to note that according to his formulae the field-corrected loss peak for the BSA complex of fig. 8 can be described by assigning to the conducting regions a small spread of conductivity values centred around 3 x S m-l with the poorly conducting regions having an effective conductivity of 4 x S rn-l and an effective thickness 1/140 of that of the conducting regions. For the collagen complex and the results of fig. 9 the corresponding values are S m-l for the conductive and resistive regions, respectively, with the ratio of the thickness of the resistive regions to that of the conducting regions being 1/156.S m-l and 2 x CONCLUSIONS Although the results described here were influenced by effects associated with sample heterogeneity, the conductivity data can be taken to support the earlierS. BONE AND R. PETHIG 1793 findings5 that the interaction with methy iglyoxal increases the electrical conductivity of proteins. In the earlier work the effect of exposure to air in producing samples with outer regions of lower conductivity than the bulk was not appreciated, and thus there would have been a tendency to underestimate the extent to which methylglyoxal enhanced the protein conductivity. Whether or not the action of methylglyoxal is to form a charge-transfer interaction with the proteins cannot at present be precisely determined, but this action does appear to involve a true bulk phenomenon.Investigations of the nature of the air-induced chemical instability should assist a clearer understanding of the protein-methylglyoxal interaction. The extent to which an interpretation of the dielectric loss measurements can be given in terms of bulk- rather than surface-influenced effects depends essentially on the observation that the frequency of the loss peaks did not vary with sample thickness. Although the theoretically derived results of fig. 7 have similar frequency ranges and loss-peak magnitudes to the experimental data of fig. 4 and 5, fig. 7 does emphasise the fact that the relaxation frequency for an interfacial polarisation involving the sample surfaces should decrease with increasing sample thickness as a consequence of the mobile charges in the bulk taking an increasingly longer time to diffuse back and forth across the sample.Small changes in the frequency of the loss-peak maximum were observed for the collagen samples (see fig. 4 and S), but since no consistent trend with change in sampie thickness was observed this effect most likely arose from slight differences in sample temperature and hydration content. The ill-defined loss appearing at frequencies < Hz for the BSA complex of fig. 4 could possibly be associated with sample surface effects. The inference that the experimental data of fig. 4 and 5 represent a bulk effect IS supported by the field-corrected results of fig. 8 and 9, and although the results can be well described in terms of the Koops model of an inhomogeneous dielectric a clearer understanding of the physical mechanisms underlying the hopping charge-transport processes should follow from dielectric theories such as those developed by Lewis.lBj l9 Further work is required to characterise the nature of the dominant charge carriers in the protein-methylglyoxal adducts, and in the light of the effects reported here the extent to which interfacial effects may have influenced the recent dielectric investiga- tions of uncomplexed haemoglobin,20 BSA21 and of multilayer films of haemoglobin22 prepared by the Langmuir-Blodgett technique, should also be considered.We thank Joy Behi and Jane McLaughlin for their valuable preparative work, Prof.T. F. Slater for kindly supplying 14C-labelled methylglyoxal, Dr P. R. C. Gascoyne for valuable discussions and one of the referees for drawing our attention to the sinh relationship of fig. 6. K. M. C. Davies, D. D. Eley and R. S. Snart, Nature (London), 1960, 188, 724. J. B. Birks and M. A. Slifkin, Nature (London), 1963, 197, 42. D. D. Eley and R. S. Snart, Biochim. Biophys. Acta, 1965, 102, 379. A. Szent-Gyorgyi, Int. J . Quantum Chem., 1976, QBS3, 45; 1977, QBS4, 179. R. Pethig and A. Szent-Gyorgyi, Proc. Nut1 Acad. Sci. USA, 1977, 74, 226. S. Bone, T. J. Lewis, R. Pethig and A. Szent-Gyorgyi, Proc. Natl Acad. Sci. USA, 1978, 75, 315. S. Bone and R. Pethig, in Submolecular Biology and Cancer, Ciba Foundation Series 67 (new series), Excerpta Med., 1979, 83. G. Fodor, R. Mujumdar and A. Szent-Gyorgyi, Proc. Natl Acad. Sci. USA, 1978, 75, 4317. K. Takahashi, J . Biochem., 1977, 81, 403. lo J. A. McLaughlin, R. Pethig and A. Szent-Gyorgyi, Proc. Nut1 Acad. Sci. USA, 1980, 77, 949. P. Otto, J . Ladik, K. Laki and A. Szent-Gyorgyi, Proc. Nut1 Acad. Sci. USA, 1978, 75, 3548. l 2 W. M. Arnold, J. Behi and R. Pethig, unpublished work. P. Carnochan and R. Pethig, J. Chem. Soc., Faraday Trans. 1, 1976, 72, 2355.1794 P R 0 T E I N-ME T H Y L G L Y 0 X A L ADD U C T S l4 J. Eden, P. R. C. Gascoyne and R. Pethig, J . Chem. SOC., Faraday Trans. I , 1980, 76, 426. l5 P. R. C . Gascoyne, Int. J. Quantum Chem., 1980, QBS7, 93. l6 L. K. H. van Beek, Prog. Dielectr., 1960, 7, 69. l7 C. G. Koops, Phys. Rev., 1951, 83, 121. la T. J. Lewis, in Dielectric and Related Molecular Processes (The Chemical Society, London, 1977), vol. l9 T. J. Lewis, in Submolecular Biology and Cancer, Ciba Foundation Series 67 (new series), Excerpta 2o D. D. Eley, N. C. Lockhart and C. N. Richardson, J . Chem. Soc., Faraday Trans. I , 1979, 75, 323. 21 J. Eden, P. R. C. Gascoyne and R. Pethig, J. Chem. SOC., Faraday Trans. 1, 1980, 76, 426. 22 J. B. Hasted, H. M. Millany and D. Rosen, J . Chem. SOC., Faraday Trans. 2, 1981, 77, 2289. 3, pp. 186-218. Med., 1979, 65. (PAPER 1/1013)

ISSN:0300-9599    

DOI:10.1039/F19827801785

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1982, 78, 1795-1808 Effect of Polyelectrolytes on the Rate of Ligand-Metal-ion Reactions Part 1 .-‘ Catalysis’ of the Complexation of Nickel@) with an Azo-dye BY CHRISTIAN TONDRE Laboratoire de Chimie Physique Organique, ERA CNRS 222, Universite de Nancy I, B.P.239, 54506 Vandoeuvre-les-Nancy Cedex, France Received 7th July, 1981 The effect of different anionic polyelectrolytes on the rate of complexation of Nickel(i1) with the bidentate ligand pyridine-2-azo-p-dimethylaniline (PADA) is studied using the stopped-flow technique. The ‘catalytic’ effects observed aredependent on the structure of the polyelectrolyte considered : two anionic polyelectrolytes characterized by the same linear charge density, but chemically different, will either inhibit (polyphosphate) or accelerate [poly(styrene sulphonate)] the complexation reaction.Intermediary situations are observed with alternated copolymers of maleic acid and with carboxymethylcelluloses. The results are consistent with Manning’s polyelectrolyte theory, but give additional knowledge of the specific interactions occur- ring between counterions and polyelectrolyte charged sites. The rate of reactions involving ionic species is known to be modified by the presence of charged particles like polyelectrolytes or micelles,’ due not only to the existence of a large electrostatic potential, but also to other forces, e.g. hydrophobic forces. Rate enhancements are usually observed when similarly charged ionic species react in the presence of oppositely charged polyions, whereas reactions between oppositely charged ionic species are inhibited by polyions.Less attention has been paid to reactions involving one charged species and one neutral species capable of showing hydrophobic interactions. Such studies are interesting for the following reasons: (a) improvement of the understanding of the mechanisms responsible for polyelectrolyte ‘catalysis ’, which depend on each specific system studied; (b) as suggested by Morawetz and Shafer,la9 the reaction rate changes may provide information on the distribution of counterions and on the extent of polyion+ounterion interactions or site binding;3 (c) polyelectrolyte ‘catalysis’ was recently shown to be of potential interest for inducing charge separation in photoinitiated electron-transfer reactions in solar-energy storage ~ t u d i e s .~ ~ The system chosen for the present study is the complexation reaction between nickel(r1) and pyridine-2-azo-p-dimethylaniline (PADA), for which chemical-relaxation method^^-^ have shown that the rate-limiting step is the loss of water from the nickel(I1) inner hydration shell. Furthermore, it has recently been shown that the reaction rate is enhanced by the presence of sodium dodecylsulphate (SDS)lo or sodium decylsulphatell micelles in the reaction medium. The question as to whether the hydrophobic environment or the charged surface or both are necessary to increase the reaction rate is, as yet, unanswered. It was hoped, by usingdifferent polyelectrolytes, to be able to understand contributions of the electrostatic field and the hydrophobic character to the reaction rate change.In addition, PADA can act as a probe, allowing a study of the specific interactions of Ni2+ ions with different polyelectrolytes. 17951796 LIG A N D-META L-I ON REACTIONS Our experiments involve two anionic polyelectrolytes characterized by the same linear charge density, but chemically different, one of which inhibits whereas the other accelerates the same rea~tion.~ Such behaviour cannot be accounted for only in terms of the condensation theory of polyelectrolytes,12 but also gives information on the state of the counterions in the vicinity of the polyelectrolytes. EXPERIMENTAL MATERIALS PADA (Sigma Chemicals) was used without further purification. All materials used were of analytical grade. The following polyelectrolytes were used : a sodium poly(styrene sulphonate) (PSS-Na), a sodium and a tetramethylammonium polyphosphate (PP-Na and PP-TMA), alternated copolymers of maleic acid with styrene (MA-STYR), ethylene (MA-ET) and methylvinylether (M A-MVE or ' Gan trez '), carboxymet hylcelluloses (CMC) with degrees of substitution (DS) 0.98, 2.1 and 2.9.The origin and purification procedures of most of these polyelectrolytescan be found in preceding papers [see, e.g. ref. (1 3)], with the exception of MA-ET (Monsanto), and MA-STYR (Aldrich Chemicals). METHODS A modified Durrum stopped-flow apparatus1* with a 2 cm optical-path cell was used for the kinetic experiments at 25 O C . The PADA and polyelectrolyte were equilibrated in the same syringe, and the initial pH was adjusted with sodium hydroxide before mixing with the Ni(NO,), solution. Buffers were avoided because of their possible interference with the system studied.The reaction products were collected at the exit of the optical cell so that the final pH could be measured. All concentrations given are concentrations after mixing. The condition [PADA] < [Ni2+] was always fulfilled, giving rise in most instances to an exponential change in the absorbance at 550 nm (i.e. at the position of the absorption band of the metal-PADA complex),1o from which pseudo-first-order rate constants kobs were obtained. The relaxation time z of the exponentials (kobs = l/z) was calculated by a non-linear least-squares procedure using an LSI 2/20 computer (Computer Automation) on-line through a transient recorder (Biomation model 802 analogue-digital converter) to the stopped-flow apparatus.Each point on the figures is an average over at least 4 experiments. The estimated accuracy of each point is f 5 %. In some cases a better fit of the absorption change with time was obtained by introducing a second relaxation process, which was longer and had a smaller amplitude. This was the case with PSS at low polyelectrolyte concentration and at pH > 7, and with MA-ET. Nevertheless, the single exponential fit usually gave a value of z close to that for the main process, and only this relaxation time is considered. The slow process may be due to the release of bound PADA, caused by the dilution, occurring during mixing, of the equilibrated PADA-polyelectrolyte solution.A relaxation time of small amplitude was observed when PADA-PSS was mixed with water instead of the Ni(NO,), solution. RESULTS EFFECT OF DIFFERENT POLYELECTROLYTES O N THE REACTION RATE The reciprocal relaxation time (or kobs) characterizing the complexation of PADA and Ni2+ in the presence of the different polyelectrolytes is shown on fig. 1, as a function of the polyelectrolyte concentration Cp [(mol ionizable group) dm-3]. (This is equivalent to monomer concentration for PSS and PP only.) In order to make a valid comparison between polyelectrolytes, the final pH was as close as possible to 7, the reaction rate reaching a plateau value at this pH in the absence of polyelectrolyte.1° This implies that weak polyacids, like alternated copolymers of MA or CMC, are not fully neutralized.For these compounds the charge parameter c 5 = e2/ekTbC . TONDRE I 7 1797 t 100 - I < 10 D 0 4 1 0 .I 0 0.5 1 C,/ 1 O-’ (mol ionizable group) dm-’ FIG. 1 .-Plot of kobs against polyelectrolyte concentrations C,. [PADA] = 1.25 x lod5 mol dm-3; [Ni2+] = lop3 mol dm-3; pH 7k0.5. (where e is electronic elementary change, E is the dielectric constant of the solvent, b is the projection of the distance between two neighbouring charges on the polyelectrolyte axis, k is Boltzmann’s constant and T is the absolute temperature), which characterizes the polyelectrolyte properties, depends on their degree of neutral- ization. However, PSS and PP, which are strong polyacids, have practically the same charge density, yet fig.1 shows that PSS has a strong accelerating effect on the reaction whereas PP has an inhibiting effect, the difference in reaction rates being three orders of magnitude. Note also that both polymers containing styrene units induce a marked enhancement of the reaction rate. On the other hand, the copolymers MA-ET and MA-MVE have a small effect (almost no effect in the case of MA-MVE). A factor of 3 in the degree of substitution (average number of carboxymethylated sites on one pyran ring) and thus in the charge parameter of CMC has little effect on the reaction rate. For PP the same behaviour is observed when TMA+ is used as the counterion in place of Na+, which interacts much more strongly with PP13 (see fig. 9). The influence of reactants concentration was studied at fixed PSS concentrations and the results are shown in fig.2. As expected, the PADA concentration has only a small effect (possibly no effect at all), confirming that it is still a pseudo-first- order reaction when PSS is present. On the other hand, kobs increases until the Ni2+ concentration reaches a value slightly below equivalence with the number of polyion charged sites, and then stays almost constant. A curve showing the same break was also obtained with a sample of1798 LI G A N &MET A L-ION RE A C T I ONS 0 1 2 3 [PADA] / 1 O-’ mol dm-3 FIG. 2.-Plot of /cobs against concentrations of [Ni2+] (0, upper scale) and [PADA] (+, lower scale) in the presence of PSS at pH 7f0.5. (+) [Ni2+] = mol dm-3; (0) mol dm-3, [PSS] = 1.25 x mol dm-3.[PADA] = 1.25 x mol dm-3, [PSS] = 2.5 x MA-STYR, but in this case the degree of neutralization of the polyion is expected to change with Ni2+ concentration when the pH is ca. 7 (see fig. 6), which is a complicating factor. EFFECT OF pH The rate of PADA-Ni2+ complexation is dependent on the pH because of the possible protonation of the -N(CH,), group of PADA, which has pK, 4.5 in the absence of polyelectrolyte or micelles : an S-shaped curve has been theoretically accounted for by James and Robinsonlo for kobs when the pH varies from 2 to 8. Similar behaviour is observed in the presence of PSS, as shown in fig. 3. The situation is more complicated for the copolymers of MA, whose degree of neutralization changes with pH. The pH dependence of kobs for these copolymers is plotted in fig.4. The corresponding values of a were obtained from the titration curves determined in the presence of Ni2+, shown in fig. 5. A plot of kobs against a can now be obtained (fig. 6): note that the maxima observed for the three copolymers considered are occurring for a z 0.5. Spectrophotometric titration curves were also determined for some of the polyelec- trolytes, in order to see how the pK, of PADA is affected by their presence. The change of the absorption maximum corresponding to the protonated form of PADA (A,,, = 560-565 nm) with pH is plotted in fig. 7, which shows that the pK, is shifted ca. + 2 units by PSS and a smaller amount by MA-STYR. The pK, is unaffected by MA-MVE and only slightly affected by MA-ET.The curve obtained in the presence of PP-TMA does not allow us to determine a pK,, but the protonation state of PADAC . TONDRE 1799 2 3 4 5 6 7 8 9 PH FIG. 3.-Plot of kobs against pH in the presence of PSS. [PADA] = 1.25 x mol dm-3; mi2+] = mol dmP3; [PSS] = 2.5 x mol dm-3. 6 5 4 - rn . 9” e0 3 2 1 0 10 7 m --. 9” Y 5 I 1 I I I I lo 3 4 5 6 7 8 9 PH FIG. 4.-Plot of kobs against pH in the presence of MA-ET (+), MA-MVE (0) and MA-STYR (0). PADA] = 1.25 x (mol ionizable group) dm-3. rnol dm-3; [Ni2+] = 10-3moldm-3; polyelectrolyte concentrations = 5 x1800 LIG AND-MET A L-ION RE ACTIONS FIG. 5.-Potentiometric titrations of MA-MVE, MA-STYR and MA-ET in the absence of Ni2+ (-) and with a ratio [Ni*+]/[polyelectrolyte] = 0.4 (----). at pH > 5 does not seem to be affected by the presence of polyphosphate.The decreasing part of the absorption at low pH (dashed lines in fig. 7) is probably related to the colour change associated with the protonation of the ring nitrogen.15 EFFECT OF TEMPERATURE The activation enthalpy of the complexation reaction was determined from the temperature dependence of kobs in the presence of PSS and MA-STYR, giving in both cases a value of 56 kJ mol-l. As shown in fig. 8, this value compares well with the slope obtained by Cobb and Hague,’ who used 0.3moldm-3NaN03 and no polyelectrolyte, and also with the activation enthalpy measured in the presence of SDS micelles.lO DISCUSSION Many examples of polyelectrolyte ‘catalysis’ exist in the literature, and many different treatments have been applied (the word ‘catalysis’ has been put between quotation marks because it is not always strictly appropriate, as already discussed by Iselb).Some of these results have been interpreted with an approach analogous to that of the Bronsted-Bjerrum primary salt effect,l6>l7 or in terms, at least partly, of the modified local concentration of reactant ions in the polyelectrolyte domains.2T 18-20C . TONDRE 1801 0 I I I 10 - I v, \ e 4 5 0 0.5 1 FIG. 6.-Plot of kobs against degree of neutralization CI for MA-STYR (O), MA-ET (+), MA-MVE (@) 5 x (mol ionizable group) dm-3; [PADA] = 1.25 x mol dm-3; [Ni2+] = mol dmP3. In a recent paper,21 both approxhes were shown to be equivalent when second-order reactions between ions of equal charge sign are involved. None of the preceding models is immediately capable of explaining the present situation, where a reaction is accelerated by the salt of a strong polyacid (PSS-Na) whereas it is inhibited by another salt of a strong polyacid having the same linear charge density, but which is chemically different (PP).In the absence of polyelectrolyte, the observed pseudo-first-order rate constant kobs can be written in the simplified form:'' lo 7-' = kobs = kf[Ni2+],+kd ( 2 ) where k, and k , are respectively the forward and backward overall rate constants for formation of the PADA-Ni complex. There seems to be no doubt that the reaction rate observed is, as in the bulk, related to the loss of water molecules from the nickel(I1) inner hydration shell. The exchange rate of the hydration water molecules of the cobalt ion is knownz2' 23 to be faster than for Ni2+ by more than an order of magnitude: a relaxation time ca.80 times faster was measured when Co2+ was used in place of Ni2+ and the reaction was inhibited by PP-TMA. (In cases where the reaction rate was accelerated, the relaxation time would have been too fast to be measured using the stopped-flow technique.) The fact that the activation enthalpy is not different in the presence of PSS, for1802 0.7 0.6 0.5 $$ 0.4 0.3 0.2 0.1 ‘0 T LI G A N D-MET A L-ION RE A C T I 0 NS 1 I 0 5 10 PH FIG. 7.-Spectrophotometric titrations of PADA alone (+) and in the presence of polyelectrolytes: PSS (O), MA-STYR (A), MA-ET (a), MA-MVE (@), PP-TMA ( x ) . [PADA] = 2.5 x mol dm-3; (mol ionizable group) dm-3.polyelectrolyte concentrations = 2.5 x 3.2 3.3 3.4 3.5 103 K I T FIG. 8.-Arrhenius plots from the temperature dependence of kobs: PSS 2.5 x MA-STYR 5 x and in the absence of polyelectrolyte (@). [PADA] = 1.25 x mol dm-, (O), (mol ionizable group) dm-, (+); from Cobb and Hague’ in 0.3 mol dm-3 NaNO, mol drn-,; pH rnol dmP3; [Ni2+] = 7k0.5.C . TONDRE 1803 example, from its value in the bulk suggests that a similar mechanism is involved. This would favour a simple local concentration effect in the neighbourhood of the polyelectrolyte. Such considerations have successfully explained the catalysis observed in the presence of SDS micelles.lo The number of counterions ‘condensed’ in the vicinity of the polyelectrolytes can be deduced from Manning’s theory,12 assuming with Manning24 that the probability that divalent ions condense is much greater than the corresponding probability for monovalent ions.This does not take into account the specific interactions which may occur between the condensed counterions and the polyelectrolyte charged sites. Such specific interactions can be studied by dilat~metry,~~ ultrasonic absorption,13 density measurements26-28 or n.m.r.289 29 Although the present study is concerned with Ni2+ ions and what follows refers to Co2+ ions, it is striking that the maximum extent of the acceleration or inhibition (maximum or minimum value, respectively, of kobs in fig. 1 for each polyelectrolyte) are in the same order as the average number of water molecules left in the first hydration shell of the divalent ion after it has been added to the corresponding fully neutralized polyelectrolytes with an excess concentration of charged sites: PSS > MA-STYR > CMC N MA-ET > MA-MVE > PP the corresponding average number of water molecules remaining being 6 for PSS,29b 3 to4 for CMC,28 3 for MA-ET30 and 0 for MA-MVE and PP28 [the value for MA-MVE could possibly be between 0 and 1 according to fig.5 in ref. (28)]. From this point of view the rate of the ‘catalysed’ reaction seems dependent on the state of hydration of the divalent ion, depending on its specific interaction with the polyelectrolyte considered (‘catalysis ’ experiments with the Co2+ are in qualitative agreement with the present results). Further evidence of the predominant role of the divalent metal is found in the dependence of kobs on the degree of neutralization a of the copolymers of maleic acid (fig.6): it is known from previous work13 that for such copolymers a strong chelation of counterions occurs at degrees of neutralization > 0.5, when pairs of charged sites start to appear. Therefore it is reasonable to think that the accelerating effect observed for a < 0.5 is due to the increase of the effective local concentration of Ni2+ ions which keep their first hydration shell as long as no strong chelation between neighbouring charged sites can occur. Note that this does not contradict the preceding argument, because the curves of fig. 1 were obtained at pH 7 , i.e. at a > 0.5 for all the MA copolymers in the presence of Ni2+ (see fig.5). We now turn to the role and extent of interaction of PADA with the different polyelectrolytes. The shift of pK, observed in fig. 7 suggests that a stronger interaction exists between PADA and styrene-containing polymers for which an enhancement of the reaction rate is observed. Such changes of apparent acidity constants of indicators in polyelectrolyte solutions have been attributed mainly to the large charge density of polyions, but also to non-electrostatic interaction^,^^ which are supposed to prevail here. The hydrophobic interaction between PADA and PSS has previously been reported by Kunugi and who have given a value for the binding equilibrium constant. Okubo and IS^,^^ for another system, have shown that hydrophobic attractive forces between a dye and a polyion may in some instances predominate over electrostatic repulsive forces.34 Note that the results relative to the three alternated copolymers of maleic acid indicate that the interaction of PADA is of secondary importance compared with the interaction between Ni2+ and these polyelectrolytes.Indeed the position of the maxima observed in fig. 6 does not depend on the extent of interaction with PADA as1804 L I G A N D-ME T A L-ION RE A C TI ON S characterized by the shifts of pK, observed in fig. 7. Only the magnitude of the maxima is related to the hydrophobic character of the polyelectrolyte. Conformational changes in the polyelectrolyte when changing the pH are also possible (particularly in the case of styrene-containing cop01ymers~~~ 36) but they do not seem to be responsible for the observed maxima.The rate acceleration or inhibition observed may arise from different rates of reaction of the substrates in the bulk and in the polyelectrolyte domains and thus may depend on the distribution of the substrates between the two phases (this is probably the situation occurring with PSS). It may also be due to a competitive action between the polyelectrolyte and PADA for binding Ni2+ ions (as in the case of PP). According to this model, the condensed Ni2+ ions are allowed to react with PADA in a restricted space in the vicinity of PSS, whereas with PP they are so strongly dehydrated that they are no longer available for reaction with PADA. Due to the respective concentrations of PADA and polyelectrolytes, the initial condition [PADA] 6 [Ni2+] is obviously also fulfilled in the polyelectrolyte domains whatever the interaction between PADA and the polyelectrolyte, so that formation of the PADA-Ni complex may be expected to be a pseudo-first-order reaction. A general kinetic model is not easy to establish, so we will instead try to consider separately the different cases which can occur: (a) a case of rate acceleration (PSS), (b) a case of rate deceleration (PP) and (c) intermediary cases (weak polyacids).RATE ACCELERATION I N THE PRESENCE OF P s s For the reasons given above, we will consider a mechanism similar to that in the absence of polyelectrolyte [eqn (2)], with the difference that local concentrations inside and outside the polyelectrolyte domains have to be used, and we will assume that in these domains the kinetics are governed by the same rate constants as in the bulk.In the case of catalysis of the same reaction by SDS micelles, the introduction of an effective surface concentration in place of [Ni2+IT took the experimental observed rate into accountlo with the assumption that all the PADA is located close to the micelle surface. We have thus tried to construct a model to fit the increasing part of the kobs against C, curve up to the maximum, by considering simultaneous reactions in the bulk and in the polyelectrolyte domains. If there is no fast coupling by the exchange of reactants between the bulk and the polyelectrolyte domains we would expect two relaxation times to be observed, which is not the case (except in some particular instances, and then probably for another reason, as already discussed).On the contrary, if fast coupling exists, such a model has already been constructed by Diekmann and Frahm,ll leading (after simplification due to the condition [PADA] 6 [Ni2+]) to an expression of the form: kobs = kf([Ni2+lbfb + [Ni2+’1fff) -k kd (3) where the indices stand for free (f) and bound (b) and thefrepresent fractions of bound and free PADA. The limiting cases,fb = 0 andf, = 0, are in agreement with eqn (2). Quite a similar form of equation has also been used by Morawetz and Voge1.l’ If effective local concentrations of Ni2+ are introduced in eqn (3), the maximum kobs should occur when f b = 1. This would be consistent with our results only if this happens when C, is such that all the Ni2+ ions are just condensed.In a preliminary report of the present work3’ a tentative interpretation was proposed using an equation similar to eqn (3) but withf, andf, representing, respectively, the fractions of bound and free Ni2+ ions : the introduction off factors at that time was founded on intuition rather than on a theoretical basis. The reasons were the following: (a)C . TONDRE 1805 when theffactors are not introduced in the expression and as [Ni2+]gff % [Ni2+]feff from the first polyelectrolyte addition, the increase of kobs with C, cannot be taken into account; (b) the introduction of the respective volume fractions in place of theffactors leads to a value of the term between brackets which is also independent of C,; ( c ) the PADA concentration does not seem to play a significant role in the observed results.The present model is consistent with the curve obtained for PSS (fig. 1). It assumes that both reactants are progressively concentrated in the vicinity of the polyelectrolyte until a maximum value of kobs is reached when all the free species are exhausted. Then the decreasing part of the curve is due to the dilution effect which arises when the number of polyelectrolyte domains is increased.l8 If the affinity of PADA for the polyelectrolyte is less than its affinity for Ni2+ ions, the Ni2+ ions being in much higher concentration than PADA, the maximum will occur when the more concentrated species are completely bound.21 The maximum corresponds to the polyelectrolyte concentration C , for which all the Ni2+ ions will be condensed according to Manning’s polyelectrolyte theory.Indeed, when there is equivalence between Ni2+ and polyelec- trolyte charged sites, the fraction [I -(2<)-l] ‘v 0.82 of divalent ions is condensed12 so that all the Ni2+ ions are expected to be condensed when C, is ca. 20% higher than the equivalent Ni2+ concentration, i.e. 2[Ni2+] x 1.2. This corresponds almost exactly to the concentration C, for which the maximum ‘catalysis’ is observed (see fig. 9). The effective local concentration [Ni2+Ieff at the maximum can be calculated from an 80 60 krn 40 4- 20 0 :“ 0 0.2 5 0.5 0.75 1 Cp/10-2 mo1 dm-3 FIG. 9.-Plot of /cobs against calculated from eqn (5). (0) polyelectrolyte concentrations Cp. Dashed line with non-zero slope is PSS-Na; (+) PP-Na; (0) PP-TMA.[PADA] = 1.25 x lop5 mol dm+; [Niz+] = lop3 mol dmp3; pH 7F0.5. equation similar to eqn (2) with the preceding assumption for the values of k , and k,. A value of the ‘condensation volume’ (the volume around a monomer in which counterions are condensed, expressed as cm3 per monomeric residue) can then be estimated from : u1 = {500[ 1 - (2r)-1]>/(N[Ni2+]eff) (4)1806 L I G A N D-M E TA L-I ON RE ACTIONS where Nis Avogadro’s number. From the value obtained, a radius of 38 A* is deduced for the cylinder in which Ni2+ ions are condensed, assuming a totally expanded conformation of the macromolecule. The order of magnitude obtained is consistent with the theoretical calculations of Gueron and Weissbu~h,~~ who found in mol dm-3 salt a ‘condensation radius’ of 43 and 60 A for charge parameters 5 = 2 and 5 , respectively (5 = 2.8 for PSS), when monovalent counterions are condensed on a cylinder of radius 10 A (approximately the radius for PSS39).This could be the first experimental evidence supporting the theoretical predictions, since the value is expected to be lower for divalent ions. RATE DECELERATION I N THE PRESENCE OF PP The case of inhibition by polyphosphate seems easier to understand. If all ‘condensed’ Ni2+ ions are considered not to participate in the reaction with PADA because of their strong chelation with polyphosphate (vide supra), then only the concentration of free Niz+ ions has to be considered, and the expression for kobs takes the form: kobs 21 kf([Ni2+lT - 0.5[ 1 - (25)-’] C,> + kd where the term in the brackets expresses the decrease of [Niz+], with increasing C,.This expression is meaningful only as long as the term in the brackets remains positive. Fig. 9 shows a good agreement between the theoretical curve calculated from eqn ( 5 ) and the initial slope of the experimental decrease of kobs. In addition, the experimental curve tends asymptotically towards the value 0.1 s-l, which corresponds to the value of kd, i.e. the extrapolated value of kobs at zero Ni2+ concentration in the absence of polyelectrolyte. The intercept between this asymptote and the initial slope occurs at the Cp value for which all the Ni2+ ions are expected to be condensed, i.e. at the maximum obtained with PSS. In fact, neglecting the presence of monovalent ions is correct only for low value of C,.Then we are in a mixed salt situation and according to I ~ a s a ~ ~ the effective charge density may take values between 0.5 (as assumed above) and 1 . This is difficult to account for in a quantitative manner, but it could explain the curvature observed in the variation of kobs with the polyphosphate concentration. In addition, only ca. 50% of the counterions (in a stoichiometric situation) have been shown, by n.m.r. chemical shift and density experimentsz8 on polyphosphates, to be totally dehydrated : this is consistent with the fact that kobs continues to decrease with further increase of C, up. to a concentration close to twice the equivalent concentration (see fig. 9). The particular complexation of divalent metal ion with polyphosphates is discussed in more detail in the following paper.31 INTERMEDIARY CASES I N THE PRESENCE OF WEAK POLYACIDS The situation with MA-ET or MA-MVE is not as clear-cut as that for polyphos- phates. The results reported in fig.1 were obtained at pH close to 7, and the charge parameter, which is ( when the polyelectrolyte is fully neutralized, is only a5 when the neutralization degree is a. At pH 7, a is close to 0.75 for these two polyelectrolytes in the presence of Niz+, according to fig. 5. The situation as regards the interaction between Ni2+ and these polyelectrolytes is intermediate between the case of PSS and the case of PP and thus the observed influence on the reaction is very small, as was previously observed for poly(viny1 sulphate)lo [although poly(viny1 sulphate) is a strong polyacid, it behaves very likely as poly(viny1 sulphonate), which interacts more * 1 A 3 10-10 m = 10-1 nm.C .TONDRE 1807 strongly than PSS with divalent ionsz5]. The ‘catalytic’ effect observed with MA-STYR is obviously due to its larger interaction with PADA due to the presence of styrene units. For the three copolymers of MA it is reasonable to assume that condensation has an accelerating effect only up to o! = 0.5; above this value the situation is similar to that in PP and the amount of Ni2+ available for complexation with PADA decreases. CONCLUSION We have studied in the present paper the effect of different anionic polyelectrolytes on the rate of formation of a complex between Ni2+ ions and the bidentate ligand PADA.Depending on the chemical structure of the polyelectrolyte, different behaviour was demonstrated, going from an acceleration by approximately two orders of magnitude for PSS to a deceleration by one order of magnitude for PP. The results are consistent with Manning’s polyelectrolyte theory, but give additional knowledge about the specific interactions occurring between counterions and polyelectrolyte charged sites. In particular, the observed rate constant appears to be dependent on the state of hydration of Niz+ ions when ‘condensed’ in the polyelectrolyte domains. In some experiments where an accelerating rate is observed, the radius of the cylinder in which counterions are condensed may be estimated, the PADA molecule playing the role of a very sensitive probe.See, for example, the following review articles or books: (a) H. Morawetz, Acc. Chem. Res., 1970, 3, 354; (b) N. Ise, J . Polym. Sci., Polym. Symp., 1978, 62, 205; (c) E. Baumgartner and R. Fernandez-Prini, in Polyelectrolytes, ed. K. C . Frisch, D. Lempner and A. Patsis (Technomic Publications, 1976); ( d ) J. H. and E. J. Fendler, Catalysis in Micellar and Macromolecular Systems (Academic Press, New York, 1975). H. Morawetz and J. A. Shafer, J . Phys. Chem., 1963, 67, 1293. C. Tondre, C. R . Acad. Sci., Ser. B, 1981, 292, 713. D. Meisel, J. Rabini, D. Meyerstein and M. S. Matheson, J . Phys. Chem., 1978, 82, 985. D. Meyerstein, J. Rabini, M. S. Matheson and D. Meisel, J. Phys. Chem., 1978, 82. 1879. M. A. Cobb and D. N. Hague, J .Chem. SOC., Faraday Trans. I , 1972, 68, 932. E. F. Caldin and P. Godfrey, J . Chem. SOC., Faraday Trans. I , 1974, 70, 2260. H. P. Bennett0 and Z. Sabet Imani, J . Chem. Soc., Faraday Trans. I , 1975, 71, 1143. lo A. D. James and B. H. Robinson, J . Chem. SOC., Faraday Trans. I , 1978, 74, 10. l 1 S. Diekmann and J. Frahm, J. Chem. SOC., Faraday Trans. I , 1979, 75, 2199. l2 G. S . Manning, J . Chem. Phys., 1969, 51, 924. l 3 C. Tondre and R. Zana, J. Phys. Chem., 1971, 75, 3367. l4 C. Tondre and R. Zana, J . Colloid Interface Sci., 1978, 66, 544. l5 J. M. Klotz and W. C. Loh Ming, J . Am. Chem. Soc., 1953, 75, 4159. l6 K. Mita, S. Kunugi, T. Okubo and N. Ise, J. Chem. Soc., Faraday Trans. I , 1975, 71, 936. l 7 K. Mita, T. Okubo and N. Ise, J . Chem. Soc., Faraday Trans.I , 1976, 72, 1033. l8 H. Morawetz and B. Vogel, J . Am. Chem. Soc., 1969, 91, 563. 2o R. Fernandez-Prini and D. Turyn, J. Chem. Soc., Faraday Trans. I , 1973, 69, 1326. 21 R. Fernandez-Prini, J . Chem. SOC., Faraday Trans. I , 1978, 74, 2460. 22 M. Eigen, Pure Appl. Chem., 1963, 6, 97; M. Eigen and R. G. Wilkins, in Mechanism of Inorganic 23 M. A. Cobb and D. N. Hague, J . Chem. Soc., Faraday Trans. I , 1971, 67, 3069. 24 G. S . Manning, in Charged and Reactive Polymers. Volume I , Polyelectrolytes, ed. E. Selegny (D. 25 U. P. Strauss and Y. Po Leung, J . Am. Chem. Soc., 1965, 87, 1476. 26 N. Ise and T. Okubo, J . Am. Chem. SOC., 1968,90,4527. 27 C . Tondre and R. Zana, J. Phys. Chem., 1972, 76, 3451. 29 ( a ) P. Spegt and G. Weill, C . R. Acad. Sci., Ser. C, 1972,274, 587; (6) P. Karenzi, B. Meurer, P. Spegt ti R. G. Wilkins, Inorg. Chem., 1964, 3, 520. H. Morawetz and G. Gordimer, J . Am. Chem. SOC., 1970, 92, 7532. Reactions, Adv. Chem. Ser., 1965, 49, 55. Reidel, Dordrecht, Holland, 1974). P. Spegt, C. Tondre, G. Weill and R. Zana, Biophys. Chem., 1973, 1, 55.1808 LI G AND-MET A L-ION REACTIONS and G. Weill, in Protons and Ions Involved in Fast Dynamic Phenomena (Elsevier, Amsterdam, 1978), p. 299. 30 G. Weill, personal communication. 31 N. Sbiti and C. Tondre, J . Chem. SOC., Faraday Trans. I , 1982, 78, 1809. 32 E. Baumgartner, R. Fernandez-Prini and D. Turyn, J . Chem. Soc., Faraday Trans. I , 1974,70, 1518. 33 S . Kunugi and N. Ise, Z . Phys. Chem. N . F., 1974, 92, 69. 34 T. Okubo and N. Ise, J . Am. Chem. Soc., 1973, 95, 2293. 35 J. D. Ferry, D. C. Udy, F. C. Wu, G. E. Heckler and D. B. Fordyce, J . Colloid Interface Sci., 1951, 36 C. Conio, E. Patrone, S. Russo and V. Trefiletti, Makromol. Chem., 1976, 177, 49. 37 C. Tondre, in 26th IUPAC International Symposium on Macromolecules (Mainz, West Germany, 1979), Reprint of Communications, ed. I. Liiderwald and R. Weiss, vol. 11, p. 843. 38 M. Gueron and G. Weisbuch, Biopolymers, 1980, 19, 353. 39 K. Iwasa, J . Phys. Chem., 1977, 81, 1829. 6, 429. (PAPER 1/1071)

ISSN:0300-9599    

DOI:10.1039/F19827801795

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1982, 78, 1809-1816 Effect of Polyelectrolytes on the Rate of Ligand-Metal-ion Reactions Part 2.-Retardation by Polyphosphates of the Complexation of Cobalt(I1) with an Azo-dye BY NAJIBA SBITI A N D CHRISTIAN TONDRE* Laboratoire de Chimie Physique Organique, ERA CNRS 222, Universite de Nancy I, B.P.239, 54506 Vandoeuvre-les-Nancy Cedex, France Received 7th July, 1981 The effect of long-chain polyphosphates on the rate of complexation of cobalt(I1) with pyridine- 2-azo-p-dimethylaniline (PADA) is studied using the temperature-jump method. PADA is shown to be a very sensitive probe allowing the study of the specific interactions between Co'+ ions and polyphosphates. The retardation effect observed is attributed to the ' site-binding' of CO'+ ions accompanied by complete dehydration.The results confirm, through a totally different approach, previous conclusions from n.m.r. measurements. In the previous paper,' the effect of different polyelectrolytes on the rate of complexation between Ni2+ and the bidentate ligand PADA (pyridine-2-azo-p- dimethylaniline) was studied. The results obtained show that, depending on the structural characteristics of the polyelectrolyte and its electrostatic and hydrophobic character, a range of behaviour is observed, going from a large 'catalytic' effect in the presence of poly(styrene sulphonate) to a retardation effect in the presence of pol yphosphates. An interpretation was proposed, based (a) on the polyelectrolyte condensation theory of Manning2 and (b) on tne existence of specific interactions between divalent metal ions and polyelectrolyte charged sites.The mechanisms required the assumption that Ni2+ ions, when interacting with the different polyelectrolytes, behave like Co2+ ions, for which more relevant work exists in the 1iteratu1-e.~ Taking this into account, the results were consistent with a simple local concentration effect of reactants, the rate constants for complex formation being assumed to be similar to those in the bulk, which are well known from previous work on this system.6 * The purpose of the present work is to determine the precise mechanism responsible for the retardation effect observed in the presence of polyphosphates. Although less spectacular than some 'catalytic' effects shown by polyele~trolytes,~ the retardation of reaction rates may sometimes be of valuable importance.This is the case, for instance, in electron-transfer reactions used experimentally in the production of hydrogen via solar-energy conversion. For these reactions, the problem of charge separation is crucial in order to prevent the back reaction.1° Polyelectrolytes have been used for this,'lT l2 and they act by concentrating one reactant in a restricted space of the solution while the other is repelled. The temperature-jump relaxation technique was used for this study instead of the stopped-flow technique used in the previous paper because the reaction of PADA with Co2+ takes place on a shorter time-scale than the reaction with Ni2+ ions.6y13 59 1809 FAR 11810 L I G AN D-ME T A L-ION RE ACT I 0 N S EXPERIMENTAL CHEMICALS The potassium polyphosphate (PP-K) was of the same origin as in previous papers.14 The molecular weight of the present sample was determined to be 1.86 x lo6 from viscosity measurements in 0.35 rnol dmV3 NaBr at 25 *C using the empirical law given by Strauss and Wineman.15 The sample was dissolved and transformed into PP-Na by stirring with a Na-neutralized cation-exchange resin and by passing the resulting solution through a column of the same Na resin.The concentrations C, were determined as previously indicated,14" and are expressed as mol monomeric residue dmP3. PADA was supplied by Sigma Chemicals and used without further purification. The salts used in this work were of analytical grade. The cobalt@) solutions were checked for water content by complexometry with EDTA.16 TECHNIQUES Temperature-jump apparatus with light absorption detection (from Messanlagen studienges mbH) was used for the relaxation experiments.The apparatus was on line to a Biomation 805 transient recorder interfaced to a Texas Instruments 980 A computer. Relaxation times z given in the figures are usually averages over 6-10 experiments; they are computed from a non-linear least-squares fitting procedure of the experimental curves. The accuracy of each point is estimated to be & 5%. The solution was subjected to a temperature jump of 5.4 OC in order to reach the final temperature indicated. The condition [PADA] < [Co2+] was always fulfilled, giving rise to single-exponential signals when detecting at 546 nm, i.e.close to the absorption maximum of the PADA-Co2+ complex. Buffers were avoided in order to prevent possible interference with the system studied. The pH was brought to 7.0 (sfI0.25) by addition of sodium hydroxide and was checked after the solution had undergone several temperature jumps. RESULTS The influence of polyelectrolyte concentration C, on the rate of complexation of Co2+ with PADA was investigated, as was the influence of parameters such as ionic strength, metal-ion concentration and temperature. INFLUENCE OF POLYPHOSPHATE CONCENTRATION As shown in fig. 1, the presence of increasing concentrations of PP-Na in the reaction medium results in a sharp decrease of the reciprocal relaxation time, i.e. a retarding effect similar to that observed when Ni2+ was used in place of Co2+. From a value of C, slightly above twice the equivalent concentration of divalent metal ion, further addition of polyphosphate does not bring about a new change in the reciprocal relaxation time at the limit of the experimental accuracy. INFLUENCE OF IONIC STRENGTH Fig.2 shows the effect of NaCl addition when the retardation is at maximum (C, = 5 x mol monomeric residue dm-, on fig. 1) or approximately half its maximum extent (C, = 1.65 x mol monomeric residue drn-,). In both cases, when enough salt is added the relaxation time measured in the absence of polyelectrolyte is restored. Also shown on fig. 2 is that the addition of salt has no effect on the rate of reaction in the absence of polyelectrolyte. This allows comparison of the present results with results obtained in 0.3 mol dm-, NaNO, previously published by Cobb and Hague.',N.SBITI A N D C. TONDRE 181 1 100 7 5 - I v) --. n 1 z 50 25 Equivalence a I 2.5 ~ 0 - 3 5 XI 0-3 7.5 FI 0-3 Cp/mol monomeric residue dm-3 FIG. 1 .-Plot of z-l against polyphosphate concentration Cp at 25 OC: [PADA] = 1.25 x 10-5 mol dmP3; [CoCl,] = lop3 mol dmP3; [NaCl] = lop2 mol dmp3. Dashed line: calculated according to eqn (2). INFLUENCE OF c02+ CONCENTRATION AT DIFFERENT TEMPERATURES The effect of Co2+ concentration was studied in the presence of a polyphosphate concentration of 5 x 1 0-3 mol monomeric residue dmP3, i.e. almost at the concentration giving maximum inhibition (see fig. l), with mol dm-3 CoCl,. The curves representing the variation of the reciprocal relaxation time with increasing Co2+ concentration (fig.3) are characterized by a sharp change of slope at a concentration which is approximately half the equivalent concentration of charged sites. The temperature dependence of these curves was also determined. The results indicate that the change of slope of the curve is shifted towards lower concentration of divalent metal ions with increasing temperature. TABLE AP APPARENT RATE CONSTANTS FOR COMPLEX FORMATION IN THE PRESENCE OF PP-Na 12 2.3 1 2.32 1.25 4.7 18 4.57 3.20 2.4 8.3 25 6.25 4.80 6.0 14.4 30 10.80 6.87 9.3 23 34.6 15.15 8.73 14.3 31.6 40 16.5 10.80 29.6 51 59-21812 L I G A ND-MET A L-ION R E A C TIONS 1 I I I I 0 0.5 1 1.5 [NaCl] /mol dm-3 FIG. 2.-Plot of 5-l against salt concentration at 25 O C : [PADA] = 1.25 x mol dmP3; [CoCl,] = lop3 mol dmP3.+, [PP-Na] = 5 x mol dm-3 (bracketing indicate points for which a slight mol dm-3; A, no polyelectrolyte. precipitation occurred); 0, [PP-Na] = 1.65 x DISCUSSION As was recalled in the previous paper, the mechanism of the complexation reac- tion of divalent metal ions with PADA in the absence of polyelectrolyte is well docu- mented.6-7~ l3 These investigations using chemical relaxation methods have shown that the rate-determining step is the loss of water molecules from the inner hydration shell of the divalent metal ion. This allows for the coordination of the metal ion to a first nitrogen atom, followed by fast ring closure17 with the formation of a second coordination bond. The observed pseudo-first-order kinetics is characterized by a relaxation time of the simplified f0rm~9~ (1) where k, and k, are, respectively, the forward and backward overall rate constants for complex formation.A similar mechanism gives a good description of the kinetics observed when the reaction is ' catalysed ' by micelless3 or polyelectrolytesl provided that effective local concentrations are taken into account, the rate constants being similar to that in pure water. As in the case of Ni2+ a retardation of the complexation rate by a factor of 10 is observed in the presence of PP-Na. The linear decrease of l/z can again be theoretically accounted for if we asume that the competition between PP and PADA z-l = kf[Me2+] + k,N. SBITI AND C . TONDRE 1813 200 15C - I m ---- h ---.- W c- 10( 0 1 2 3 [C0C1~1/10-~ mol dm-3 FIG. 3.-Plot of 5-l against CoCl, concentration at different temperatures: [PP-Na] = 5 x [PADA] = 1.25 x mol dmP3; mol dm-3; [NaCl] = lop2 mol dm-3. x , 12; 0, 18; a, 25; +, 30; *, 34.6; A, 40 OC. for binding Co2+ ions is so much in favour of PP that the 'condensed' Co2+ counterions cannot participate to the reaction with PADA (see the previous paper). The dashed line in fig. 1 was calculated from eqn (2), which gives a fairly good approximation of the experimental results up to the PP-Na concentration for which all the Co2+ ions are condensed: (2) T - ~ = kf{[C0~+]T-O.5[1-(2~)-'] Cp} + k , where the term in brackets gives the effective free Co2+ concentration and < is the charge parameter.2 k , and k , are taken to be 6.3 x lo4 dm3 mol-l s-l and 34 s-l, respectively, as determined in this work at 25 O C , lop2 mol dm-, NaCl [these values are in fairly good agreement with values determined in ref.(1 3) in 0.3 mol dm-, NaNO,, 7.6 x lo4 dm3 mol-l s-l and 36 s-l, respectively]. Unlike the case with Ni2+, when using Co2+ the limiting value of the reciprocal relaxation time with increasing PP-Na concentration does not tend towards the value of k , measured in bulk water. This may indicate that the rate constants for formation of the PADA-Me2+ complex are more affected by the presence of the polyelectrolyte in the case Co2+ than in the case of Ni2+. Apart from this the same remarks as those previously made for Ni2+ hold here;l for instance, the intercept between the initial1814 LI G A N E M E T A L-I ON R EAC T I 0 N S slope and the final asymptote occurs for a value of C, such that all the Co2+ ions are expected to be condensed, that is for a charged-sites concentration ca.20% higher than the equivalent concentration (see fig. 1). The ionic strength (fig. 2) has no influence on the reaction rate in the absence of polyelectrolyte, as expected for a reaction involving a neutral species and an ion. The situation is different in the presence of polyelectrolyte: in this case an increase in the ionic strength results in a decrease in the Debye screening length and thus of the range of polyion-counterion electrostatic interactions. Theoretical c a l c ~ l a t i o n s ~ ~ have shown that the radius of the cylinder around a polyelectrolyte in which counterions are condensed decreases with increasing salt concentration. Thus the effect of ionic strength on the rate of PADA-Me2+ complexation should give some information regarding the polyion<ounterion interactions. In fact, two contradictory situations might occur: ( a ) decreasing the dimensions of the ‘condensation cylinder ’ may result in an increase in the retardation effect due to the concentration of more Co2+ ions in a restricted domain, or ( b ) decreasing the range of polyion-ion interactions may result in a release of bound Co2+ ions because they will be less attracted by the PP.The second is obviously true according to fig. 2, which shows that even when the Co2+ ions existing in the solution are not totally bound at the beginning (the intermediate curve) there is no further inhibition.On the contrary, the polyelectrolyte effect is progressively cancelled by the addition of salt, and the reciprocal relaxation time tends asymptotically to its value in the absence of polyelectrolyte. The results obtained seem to indicate that when present in great excess compared with divalent Co2+, Na+ can displace Co2+. Thus the predictionz0 that in a mixed-salt situation the divalent ions will condense first in order to lower the effective charge parameter and then monovalent counterions will condense only if the net charge parameter is still > 1, may not be absolutely right. The curves of fig. 3 show a break at a concentration corresponding to a stoichiometry of ca. 1 Co2+ ion for 4 charged sites (PO; groups) on the polyelectrolyte, i.e.a ratio r = [Co2+]/[PO;] x 0.25. Similar behaviour was previously observed in n.m.r. experiments, by looking at the chemical shift of the water protons when Co2+ ions are added to polyph~sphates.~-~ The observed chemical shift was shown to reflect the change in the average number of water molecules in the first hydration shell of the Co2+ ions. It was concluded from these results that for r c 0.25 the Co2+ ions were completely dehydrated, whereas for r > 0.25 they behave almost as free Co2+ ions. The change of hydration of Co2+ ions was taken as a criteria for their ‘site-binding’ (i.e. strong chelation on specific sites as opposed to ‘atmospheric binding’). The change of slope in the curves of fig. 3 possibly have the same origin.The particular complexation of divalent metal ions with polyphosphates (1 molecule of M2+ for 4 phosphate groups), may be due to the fact that repulsions between divalent ions discourage their binding on consecutive sites.Ig Such a stoichiometry does not necessarily indicate a change of coordination number from 6 to 4, because two water molecules might occupy the remaining places in the coordination sphere. Such a. change of coordination number should be accompanied by a colour change, which was not observed. The water molecules could be partially released and so not observed by n.m.r. In the case of Mg2+, which when free is usually coordinated to six water molecules, as is free Co2+, Glonek21 has proposed, from 31P chemical-shift experiments with chain polyphosphates, the existence of a magnesium polyphosphate complex involving a helix with four phosphates per turn.The change of slope would occur when the Co2+ ion concentration is high enough compared with the polyphosphate concentration so that, apart from the strongly bound Co2+ ions, atmospherically bound or free Co2+ is present. In the first regionN. SBITI AND C . TONDRE 1815 TABLE 2.-cOMPARISON OF ACTIVATION ENTHALPIES IN THE PRESENCE AND IN THE ABSENCE OF PP-Na apparent activation enthalpies in the presence of PP-Na /kJ mol-1 activation enthalpies in the absence of polymer13 /kJ mol-' AH& 50.2 (f4.0) 40.3 (k 1.2) AH: 4 3 . 0 (k2.l) AH,Z,,, 82.0 (f 2.0) AH,f,,, 60.9 (kO.8) AH? 62.7 (f 5.8) AH:,pr, (before the change of slope) the reaction is strongly inhibited by the polyelectrolyte, whereas in the second region (after the change of slope) the kinetics might be similar to that in the absence of polyelectrolyte.In order to check that point we determined the apparent rate constants at the different temperatures, assuming that each piece of curve could be characterized by an equation similar to eqn (1). So, from the slope and intercept of the first part of the curve, apparent rate constants k;,,, and k;,,, are determined; klapp and kZapp are then obtained from the slope of the second part of the curve and the position of the change of slope, respectively. The values obtained are shown in table 1, and from the corresponding Arrhenius plots the apparent activation enthalpies have been calculated and are compared in table 2 with the values in the absence of polyphosphate.These results show that the kinetics observed after the change in slope is not very much different from that in pure water (at 25 OC, klapp = 4.8 x lo4 dm3 mol-1 s-l and kzapp = 14.4 s-l, whereas the values determined above in the absence of polyelectrolyte were 6.3 x lo4 dm3 mol-l s-l and 34 s-l, respectively) and is characterized by very similar activation enthalpies (table 2). Thus the kinetics is not very sensitive to the presence of polyphosphate when Co2+ ions are in excess in the solution. On the other hand, higher apparent activation enthalpies are obtained when the polyelectrolyte affects the kinetics. Note that short-chain polyphosphates like triphosphate (TP) have only a small effect on the rate of complexation:13 the similarity of the kinetic parameters for the complexation of PADA with Coz+ and with CO(TP)~- led Cobb and Hague to conclude that the effect of charge is of secondary importance in complex formation.The effect observed in this paper is thus characteristic of polyelectrolyte behaviour. CONCLUSION Polyelectrolyte 'catalysis ' of ligand-metal reactions provides an interesting way of studying polyion-counterion interactions. In some cases of rate enhancement, information can be obtained concerning the radius of the cylinder in which the counterions are c0ndensed.l On the other hand, the retardation effect observed in the present study provides evidence of the specific interactions which exist between some divalent metal ions and polyphosphates.C. Tondre, J. Chem. SOC., Faraday Trans. 1, 1982, 78, 1795. P. Spegt and G. Weill, C . R. Acad. Sci., Ser. C , 1972, 274, 587. * G. S. Manning, J . Chem. Phys., 1969, 51, 924.1816 LI G A N D-META L-ION R E ACT I ON S .I P. Spegt, C. Tondre, G. Weill and R. Zana, Biophys. Chem., 1973, 1, 55. P. Karenzi, B. Meurer. P. Spegt and G. Weill, in Protons and Ions Involced in Fast Dynamic Phenomena (Elsevier, Amsterdam, 1978), p. 299. R. G. Wilkins, Inorganic Chemistry, 1964, 3, 520. M. A. Cobb and D. N. Hague, J . Chem. Soc., Faraday Trans. I , 1972, 68, 932. A. D. James and B. H. Robinson, J . Chem. Snc., Faraday Trans. 1, 1978, 74, 10. H. Morawetz and B. Vogel, J . Am. Chem. Soc., 1969, 91, 563. lo M. Gratzel, J . Chim. Phys., 1981, 78, 1. l 1 D. Meisel, J. Rabini, D. Meyerstein and M. S. Matheson, j . Phys. Chem., 1978, 82, 985. l 3 M. A. Cobb and D. N. Hague, Trans. Faraday Soc., 1971, 67, 3069. l4 C. Tondre and R. Zana, J . Phys. Chem., (a) 1971, 75, 3367, (b) 1972, 76, 3451 ; C. Tondre and R. Zana, in Charged and Reactive Polymers. Volume I , Polyelectrolytes, ed. E. Selegny (D. Reidel, Dordrecht, Holland, 1974), pp. 323-338; R. Zana and C. Tondre, Biophys. Chem., 1974, 1, 367; R. Zana and C. Tondre, in Chemical and Biological Applications of Relaxation Spectrometry ed. E. Wyn-Jones (D. Reidel, Dordrecht, Holland, 1975), pp. 333-341. D. Meyerstein, J. Rabani, M. S. Matheson and D. Meisel, J. Phys. Chem., 1978, 82, 1879. l5 U. P. Strauss and P. L. Wineman, J . Am. Chem. Soc., 1958, 80, 2366. l 6 Methodes d’analyses complexomttriques par les Titriplex (Merck, Darmstadt, West Germany). l 7 B. H. Robinson and N. C. White, J. Chem. SOC., Faraday Trans. I , 1978, 74, 2625. l8 S. Diekmann and J. Frahm, J. Chem. SOC., Faraday Trans. 7, 1979, 75, 2199. l9 M. Gueron and G. Weisbuch, Biopolymers, 1980, 19, 353. 2o G. S. Manning, in Charged and Reactitle Polymers. Volume I , Polyelectrolytes, ed. E. Selegny (D. Reidel, Dordrecht, Holland, 1974), p. 14. T. Glonek, Phosphorus and Sulphur, 1978, 4, 235. (PAPER 1/1072)

ISSN:0300-9599    

DOI:10.1039/F19827801809

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1 , 1982, 78, 1817-1826 Optical and Thermal Studies of Transitions between Phases 11, I11 and IV of Ammonium Nitrate BY JOHN S. INGMAN, GORDON J. KEARLEY AND SIDNEY F. A. KETTLE* School of Chemical Sciences, University of East Anglia, Norwich NR4 7TJ Received 10th July, 1981 Transitions between phases 11, I11 and IV of ammonium nitrate are studied in some detail using microscopy, dilatometry, electrical conductivity and differential scanning calorimetry (d.s.c.). Clear evidence of a two-step IV-I11 transformation is presented and the competition between phase transitions IV-I1 and IV-111-11 is explained in terms of a kinetic versus thermodynamic control. A similar explanation accounts for the inhomogeneity of typical specimens of pure ammonium nitrate, as demonstrated by many examples of 'mixed transitions'.Ammonium nitrate is reported to exist in at least six different polymorphic forms at atmospheric pressure. The transition temperatures of the various phases are generally accepted to be: ca. -200 OC - 18 OC 32 OC 84 OC 125 OC VI I v---- IV--- I11 ~ I1 I \ I 5 2 OC /.-- This paper is concerned with the transformations between phases 11, I11 and IV. The above diagram of transition temperatures is deceptive. In fact, considerable hysteresis occurs on several of the transitions. Of those with which we are concerned here, there have been reports of the I1 -+ I11 transition occurring anywhere between 48 and 84 OC.'? * In contrast, the I11 + I1 transition occurs between 84 and 89 0C.3*4 The subject is complicated by the possibility of direct IV f+ I11 transitions, the latter also showing considerabk thermal hysteresis (IV -+ I11 occurring between 3 1.5 and 55.0 O C and I11 -+ IV between 0 and 35 oC).5*6 Part of the problem arises from the sensitivity of the transition temperatures to a diverse range of variables, some of which are difficult to quantify or even identify.In particular, water content, impurity content and thermal history are of recognised importance' but the inter-relationships between the effects of these variables is little understood or quantified. Frequently, transition temperatures have been reported in the literature with no mention of any of these variables or evidence to support the chosen transition. The present work was stimulated by optical microscopic studies of the phase transitions.These revealed several phenomena and we now enumerate the most important of these. Videotape recordings were made of the observed phase transitions and appropriate ' stills ' are shown in plate I . First, clear evidence was obtained of a phase intermediate between phases IV and I11 [plate 1 (a) and (b)]. The existence of such a phase has previously been suggested.8 The two photographs shown in plate I ( a ) and (6) were taken ca. 1 min apart with no alteration in microscope settings. The sample temperature increases from right to left in each photograph. In plate l(a) phase IV is white and the intermediate dark; in plate 1 (b) the intermediate is seen together with phase I11 (white). The life-time of 18171818 PHASE TRANSITIONS I N AMMONIUM NITRATE this intermediate increases with increasing hysteresis, and when large hysteresis occurs it may also be detected by electrical conductivity or dilatometric measurements.Secondly, it became clear that different crystals within a microscopic sample may transform at different temperatures. The microscope used was fitted with a heated stage which gave rise to a small thermal gradient across the sample. However, this gradient could not explain our observations. In extreme cases, different crystals within the same sample were observed to have transition temperatures differing by ca. 10 "C for the same phase transition. We believe that crystal defects (both lattice and impurity) are of key importance in the phase transitions of ammonium nitrate and would associate this crystal-dependent behaviour with the varied occurrence of defects.Thirdly, the rate of movement of phase boundaries varies from crystallite to crystallite. Some movements, such as the I11 -+ IV transition, can be very slow (taking days to cross the field of view of the microscope) or the same transition can be complete in less than a second. At the other extreme the IV --+ I1 transition tends to be rather fast. The fastest that we recorded is shown in plate 1 (c) and (d), where the crystallite indicated (ca. 0.2 mm in length) changed phase between frames of the video-recording (50 frames s-l). From these data we calculate that the phase transition passed through each ion in < ca. 3 x s (a typical vibration occurs in Fourthly, the phase transition which occurs may be different in different crystallites.Microscopically, the IV + 111 transition is seen to lead to crazing of the crystal; the beginning of this may be seen in plate 1 (b). The IV --+ I1 transition has no such effect, the crystal retaining its optical clarity and axes. The intermediate phase in the IV -+ 111 transition, which also serves to characterise it, does not lead to crazing; this occurs immediately after the intermediate to phase I11 step. When, on further heating, the I11 -+ I1 transition occurs, the crystal obtained is also crazed. It is thus possible to distinguish IV -+ I11 and IV --+ I1 transitions with considerable confidence and to establish the co-occurrence of these two transitions in different crystallites.An example of this behaviour is shown in plate 2(a) (phase IV) and 2(b) (a mixture of phases 111 and 11). As indicated above, from plate 2(b) alone one cannot establish which regions are phase I1 and which are phase 111. (They are distinguished by their subsequent thermal behaviour.) The regions in plate 2(b), transitions IV -+ 111 and IV --+ 11, are approximately indicated in the associated diagram. Fifthly, when grown as thin plates, between microscope slides, ammonium nitrate undergoes a IV + I1 transition at ca. 50 O C rather than a IV --+ I1 transition (at ca. 35 "C), despite the presence of saturated mother liquor. We take this as clear evidence that although it seems generally accepted that a high moisture content usually promotes the IV --+ 111 transition, other factors can be of greater importance.s). EXPERIMENTAL Small crystals of ammonium nitrate, suitable for microscopic observation, were grown from aqueous solution contained either in an evaporating dish or between microscope slides. A polarising microscope was used which was fitted with a standard hot-stage arrangement with the heating/cooling rates controlled to between 0.5 and 2.5 "C min-l. Volume changes during the phase transitions were followed using a purpose-built dilatometer employing silicone oil as hydraulic fluid. A sample size of ca. 6.0 g was found to be necessary to give adequate precision in the volume-change measurements which were carried out using a heating/cooling rate of 0.1 O C min-'. Electrical conductivity. measurements were recorded using a standard ax.conductivity bridge. The sample was in the form of a pellet which was formed under pressure in a steel braced nylon tube. This method of sample preparation permitted the use of mercury electrodes.J . Chem. SOC., Faraday Trans. I , Vol. 78, part 6 Plate 1 PLATE 1 .+a) and (b) Videotape recordings of crossed-polariser microscopic evidence for a phase intermediate between phases IV and IIL on heating ammonium nitrate at ca. 45 O C . Phases IV and I11 are the white regions in (a) and (b), respectively; the intermediate is the dark region. The field of view in these, and all other plates, is ca. 0.2 mm x 0.15 mm. (c) and (d) Successive frames from a videotape recording of a crossed-polariser microscopic study of the IV --+ I1 transition at ca.52 O C . The phase transition is completed for the majority of the crystal within this interval. J. S. INGMAN, G. J. KEARLEY AND S. F. A. KETTLE (Facing p . 18 18)J. Chern. SOC., Faraday Trans. I , Vol. 78, part 6 Plate 2 PLATE 2.-(a) and (b) Evidence for a mixed IV + I1 and IV --* 111 transition in a single crystal of phase IV of ammonium nitrate observed between crossed polarizers. On further thermal cycling the regions transforming IV -, I11 fragmented whereas those transforming IV --* I1 remained intact. The appropriate boundaries between regions transforming differently are indicated in the sketch below. J. S. INGMAN, G. J. KEARLEY AND S. F. A. KETTLEJ . S . I N G M A N , G. J. KEARLEY A N D S. F. A. K E T T L E 1819 A Perkin-Elmer DSC2 machine was used to record the d.s.c.traces with the samples (ca. 6.00 mg) contained in air-tight stainless-steel pans. Heating/cooiing rates were selected between 0.25 and 2.5 "C min-l and temperature calibration was achieved by recording the melting point of pure ice (0.00 "C) and pure indium (156.6 "C). Analytical grade ammonium nitrate (Fisons Ltd) was used throughout this work; moisture levels were determined by the Karl Fischer method. RESULTS AND DISCUSSION We have found that the enthalpy changes which occur when ammonium nitrate is heated or cooled through its phase transitions frequently include endotherms or exotherms which are not in accord with the normally accepted phase-transition sequences. In the present work we have used differential scanning calorimetry (d.s.c.) as a rapid and convenient technique for further investigations.In many instances this anomalous behaviour can be explained, by reference to our microscopic work, as a consequence of different crystallites undergoing either different phase transitions or the same phase transition at different temperatures. As is evident from our microscopic observations, we must consider the co-existence of IV -111 transitions concurrent with IV --+I1 transitions in any one sample. Repeated thermal cycling of the sample between the regions of stability of phases IV and 11 may lead to a change in the relative importance of these competing concurrent changes. In terms of an overall mechanism two possibilities, not entirely separate, should be recognised.First, that preconditioning of a crystallite, its thermal, chemical and temporal history, may determine whether it transforms IV ft I1 or IV ft 111. Previous workers have asserted the importance of these f a c t o r ~ ~ ~ ~ O and this assertion now appears to be generally accepted. If true, this hypothesis leads to the conclusion that for a sample to behave irreproducibly on thermal cycling it must be modified in some way within each cycle in order to alter its transition characteristics. Secondly, the transition behaviour may be determined by phase metastability, a well recognised phenomenon for ammonium nitrate. Thus, phase IV may be metastable for long periods at temperatures at which phase I11 is, thermodynamically, the more stable. A similar argument may be developed for the metastability of phase I1 on cooling.We distinguish, then, between thermodynamic and kinetic stabilities. A consequence of this metastability is thermal hysteresis which may be such that the IV -+ I11 transition temperature may overlap with the transition temperatures of the IV ft I1 and I1 -+ 111 transitions. In such a situation, if on warming the IV -+ I11 transition has not occurred below ca. 52OC, then a IV-+II transition will occur instead. A similar problem arises, however, for the I1 -+ 111 transition which, whilst showing considerable hysteresis, has never been reported below ca. 48 O C . We note that the I1 -+ IV transition occurs at 52 "C with little hysteresis; in this case it seems entirely likely that if a I1 -+ I11 transition has not occurred on cooling to 52 O C it will be replaced by a I1 --+ IV transition.These arguments contain implications for the thermal hystereses associated with the various phase transitions. In particular, it implies that thermal hystereses of IV -+ I11 and I1 -+ I11 transitions are of approximately equal magnitude. Bringing together the preconditions and kinetic/thermodynamic arguments we see that if a particular crystallite in a sample is preconditioned to show extreme thermal hysteresis then, on heating, it will remain in phase IV until a temperature is reached at which the IV -+ IT transition rapidly occurs. Subsequently, on cooling, it will remain in phase I1 until the rapid TI + IV transition occurs.1820 P H A S E T R A N S I T I O N S I N AMMONIUM N I T R A T E 70t n O O 0 10 r O 0 0 IV-I1 11-IV sample - FIG.1.-Plot of the thermal hysteresis of phase transitions for 1 1 samples all recorded using a heating/cooling rate of 2.5 O C min-'. A, I1 + 111; Z, IV 4 111; 0, I11 + IV. This rationale may be tested by recording the approximate transition temperatures of a variety of samples and then arranging these in order of increasing overall hysteresis. This plot is illustrated in fig. 1, which shows d.s.c. data from a variety of samples of mixed origin. The data shown in fig. 1 were chosen to be representative of a much larger number; the samples were all of high purity material but varied in water content, in previous mechanical and thermal treatment and came from a variety of sources. That, despite these diversities, there should be the clear pattern shown in fig.1 provides, we believe, support for the rationale presented above. The data shown in fig. 1 had one aspect in common: the heating and cooling rates were the same for all samples. As is implied in the thermodynamic/kinetic stability arguments presented above, a statement about transition temperatures for a particular sample may, on its own, be almost meaningless. It is our experience, for example, that a sample which normally undergoes IV f+ I1 transitions, transforms into phase I11 by extended storage at a constant temperature between 33 and 84 O C . MIXED T R A N S I T I O N S It was suggested above that the hystereses of IV -+ I11 and I1 -+ I11 transitions are approximately equal for a given crystallite within a given thermal cycle.Deviations from equality or variations between crystallites can give rise to several phases being encountered on heating and cooling. Fig. 2 shows the d.s.c. trace obtained when a crystallite was cycled 0-100-0 "C. On heating, a single endotherm was recorded atJ . S . I N G M A N , G . J. KEARLEY A N D S. F. A. KETTLE n 1821 I I I 1 I TI" C FIG. 2.-D.s.c. trace from 7.5 mg of ammonium nitrate (0.16% H,O) recorded using a heating/cooling rate of 2.5 O C min-' showing a IV-I1 phase transition on heating but I1 -+ IV and 111 -, IV transitions on cooling. 0 20 40 60 80 "C r i - - - T - - -7-7 k, I I I I I I I 20 40 60 "C TI" C FIG. 3.-D.s.c. trace from 6.2 mg of ammonium nitrate (0.20% H20) recorded using a heating/cooling rate of 2.5 O C min-' showing an apparent IV -+ I1 phase transition on heating but I1 + IV and I11 -+ IV transitions on cooling.1822 PHASE TRANSITIONS IN AMMONIUM NITRATE 53 "C and the absence of an endotherm close to 85 O C , corresponding to a 111 -+ I1 transformation, indicates that the observed feature is due to a IV + I1 transition.On cooling, however, two exotherms were recorded. The first, at 57 OC, must be due to a I1 +I11 transition since this temperature is too high for the alternative I1 +IV transition. The low-temperature exotherm, at 10 OC, can only be interpreted as due to a I11 -+ IV transition. For this sample, the phase sequence on heating clearly differed from that on cooling. An apparently similar, but actually rather different, case is shown in fig.3 . In this case the endotherm on heating occurred at the upper limit for the IV -+ 111 transition, 55 O C . There are two ways in which the presence of phase 111 could have been checked; seeking features corresponding to the 111 + I1 or 111 + IV transitions. We chose the latter and therefore did not heat above 60 OC. On being cooled the sample gave two exotherms, one at 45 OC the other at 19 "C. The former must be a I1 + IV transition and the latter a I11 + IV transition. We therefore conclude that the apparently single endotherm obtained on heating is actually composite, containing both IV -+ I11 and IV -+ I1 features. I I 1 1 -1 43 45 47 OC TI" C FIG. 4.-Multiple endotherms obtained when 6.9 mg of ammonium nitrate (0.18% H,O) was heated slowly (0.25 'C min-l) through the IV -, 111 phase transition.Non-homogeneous transition behaviour in a single specimen is quite common and is exemplified by fig. 4, which shows the d.s.c. trace obtained when a polycrystalline sample was heated slowly (0.25 OC min-l) through the IV -+ 111 transformation. The multitude of small endothems presumably arise from different crystallites in the sample transforming independently. Another type of mixed transition which we have observed involves the formation of one phase from another with the product phase being unstable with respect to a third phase. Fig. 5 shows a d.s.c. trace in which both endotherms and exotherms were obtained on cooling and since production of lower-temperature phases can only yield exotherms, the endotherm must indicate that a higher-temperature phase has been produced on cooling.J .S . I N G M A N , G . J. KEARLEY AND S. F. A. KETTLE 1823 A L - - - - - TI" C 20 30 40 5 0 O C FIG. 5.-Combined exotherm/endotherm recorded when a sample of ammonium nitrate was cooled. This d.s.c. trace illustrates a I1 -+ IV endotherm followed immediately by a IV -+ 111 endotherm. The exotherm at 26 "C corresponds to a 111 -+ IV transformati-n. The exotherm at 45 OC indicates a I1 + IV transition. The phase IV produced in this way is thermodynamically unstable with respect to phase I11 and we attribute the endotherm at 40 OC to a IV + I11 transition. That similar endotherms are not seen in fig. 3 and 4 is an indication of the sample-specific behaviour of ammonium nitrate. The exotherm at 26 OC is undoubtedly due to a I11 -+ IV transition and is consistent with the explanation we have presented.The sample of fig. 5 was initially at a temperature in excess of 90 OC so that it would have contained no phase 111. The absence of an exotherm above 48 OC indicates no I1 -+ I11 transition. Unless the 45 "C exotherm is composite, containing some I1 + 111, the exotherm at 26 OC corresponds to the phase TI1 formed at 40 OC converting into phase IV. THERMAL C Y C L I N G It has been reported1' that thermal cycling increases thermal hysteresis; we endorse this observation. Note that ammonium nitrate loses moisture rapidly on heating; a loss which is believed to cause an increase in hysteresis. Whilst it may be argued that increasing hysteresis on cycling occur as a result of the loss of moisture, this cannot be true in our case because our measurements were made on samples sealed in air-tight containers.Changes in defect concentration seem a more appropriate explanation. Our results are shown in fig. 6. One sample was recycled over 0-100-0 "C for a total of twenty-three cycles. Each cycle was complete in 80 min and was immediately followed by another cycle. Although there are many examples of individual erratic behaviour, a clear trend is evident. On thermal cycling the sample tends to behave in an apparently more simple manner; it is presumably becoming more uniform in terms of impurities, crystal defects and other factors influencing thermal hysteresis. The data in fig. 6 provide an indication of relative hysteresis.Thus, on cooling the I1 -+ IV transition always occurs at the same temperature whereas the I1 -+I11 transition is variable. Similarly, the I11 -+ IV transition is very variable and may not have occurred at 0 OC in some cases (e.g. cycles 14, 16 and 21). The IV + 111 transition is more variable than the IV + I1 transition at a higher temperature; the invariant1824 PHASE TRANSITIONS I N AMMONIUM NITRATE I i I 20 100 T/OC 16 For description see opposite.J . S . I N G M A N , G. J. K E A R L E Y AND S. F. A. KETTLE 1825 \ I I I I I I I I I I FIG. 6.-A record of the effects of thermal cycling, 0-100-0 OC, on 5.8 mg of ammonium nitrate (0.16% H,O) over 23 cycles using a heating/cooling rate of 2.5 OC min-l. For convenience of presentation the recovery step occurring in the change from cooling to heating is omitted and so each cycle in this figure is shown cycling over 20-100-0 O C .1826 PHASE TRANSITIONS I N AMMONIUM NITRATE presence of a I11 -+ I1 peak demonstrates that for cycles 17 and above, the apparently single endotherm at ca.55 O C is in fact composite. Fig. 6 reveals that the 111-11 endotherm, at 87 O C , is of approximately equal integrated area throughout the series of measurements, and it must be concluded that an approximately constant mass of phase I11 exists between 60 and 87 O C on heating for each cycle. It follows that an approximately constant mass of the sample must be transforming via the IV+II +IV pathway and that the IV + 111 + I1 + III(1V) + IV transitions of the remainder of the sample are super- imposed on this constant IV + I1 + IV background.The simplicity of the cycles 18-23 (with the exception of cycle 21) is therefore deceptive. There is a clear, if qualitative, correlation between the structure of the endotherms at ca. 55 O C and the exotherms at around the same temperature. However, the existence of a single endotherm at this temperature for cycles 18-23 (excepting cycle 21) does not prevent approximately equal proportions of the sample transforming IV --+ I11 and IV + TI on heating. The balance between the proportions transforming by the two paths is not constant (cf. cycles 22 and 23) and likely, therefore, to be sensitive to both impurities, defects and to their distribution. CONCLUSIONS The thermal properties of ammonium nitrate are difficult to study.Even when the moisture content is controlled there are wide variations in the behaviour of different samples and even a given sample does not usually exhibit reproducible behaviour, as evidenced by our thermal cycling results. Nevertheless, of the hundreds of d.s.c. determinations we have carried out on ammonium nitrate, we have not recorded a single trace which could not be interpreted by the thermal hysteresis model we have presented in this paper. It is generally accepted that moisture plays a crucial role in determining the phase transitions of ammonium nitrate. However, it is unreasonable to suggest that non-homogeneous sample behaviour is due to differences between the moisture contents of the crystallites, since it is known that this material equilibrates rapidly with atmospheric moisture.12 Furthermore, samples which are heated to a temperature in excess of ca. 100 OC subsequently transform I1 -+ IV --+ I1 for a considerable number of cycles, notwithstanding the presence of moisture.6 Obviously, other factors are implicated and the identification of these factors is crucial if our understanding of the detailed mechanism of the phase transitions of ammonium nitrate is to be increased. We are indebted to Fisons Ltd for supporting this work. K. Kamiyoshi and T. Ymamkami, Sci. Rep. Res. Znst. Tohoku Univ., Ser. A, 1959, 11, 418. K. Sekiguchi, T. Yotsuyanagi and S . Mikami, Chem. Pharm. Bull., 1964, 12, 994. H. Komatsu, Rep. Inst. Sci. Technol. Tokyo Univ., 1951, 5, 15. M. Nagatani and T. Seiyama, Kogyo Kagaku Zasshi, 1964, 67, 1342. E. J. Griffith, J . Chem. Eng. Data, 1963, 8, 22. R. N. Brown and A. C. McLaren, Proc. R . Soc. London, 1962, 266, 329. I. M. De Saenz, J. C. Tessore and R. Leone, Schweiz. Mineral. Petrogr. Mitt., 1970, 50, 209. W. Engel and P. Charbit, J. Therm. Anal., 1978, 13, 275. B. V. Erofeev and N. I. Mitskevich, Zh. Fiz. Khim., 1950, 24, 1235 and 1952, 26, 848. lo Y . Shinnaka, J . Phys. Soc. Jpn, 1956, 11, 393. l1 Q. Joupperi, Ann. Acad. Sci. Fenn., Ser. A6, 1972, 384. l2 S . I. Volfkovich and R. E. Remen, Trans. Sci. Inst. Fertilizers (Moscow), 1927, 46, 5. (PAPER 1 / 1092)

ISSN:0300-9599    

DOI:10.1039/F19827801817

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. Sor., Faraday Trans. 1, 1982, 78, 1827-1833 Standard Enthalpies of Transfer of Ionic Surfactants from Water to Water + t-Butyl Alcohol Mixtures at 298.16 K BY RAYMOND BURY A N D CLAUDE TREINER* Laboratoire d’Electrochimie, Universite P. et M. Curie ERA 3 10, 4 Place Jussieu, 75005 Paris, Fraiice Received 15th July, 1981 The standard enthalpy of solution of two ionic surfactants, sodium decylsulphate and trimethyldecyl- ammonium bromide, as well as those of tetramethylammonium bromide and methylsodium sulphate, have been determined in water+ t-butyl alcohol (ButOH) mixtures at 25 O C . The maximum endothermic effect observed for the surfactants at 0.05 mole fraction of ButOH is the largest ever measured in these mixtures at 25 O C [AHp(max) = 76.6 kJ molF1] for sodium decylsulphate.It is observed that in water+ ButOH mixtures the effect is larger for a rod-like ion than for the bulky tetrabutylammonium ion. The reverse is true in water + acetone mixtures. These results are discussed in terms of the structures of the respective binary solvents. The relevance of these results to the opposite variations in the critical micelle concentration of the surfactants in water on the addition of acetone (increasing) or ButOH (decreasing) is also considered. Finally the contribution of the n-nonyl hydrocarbon chain of each ionic surfactant to the standard enthalpy of transfer is calculated. The hydrocarbon chain attached to the cationic head group is more endothermic than the attached to the anionic head group over the whole range of water+ButOH and water+acetone mixtures.There is a growing interest in an understanding of the thermodynamics of ionic surfactants in presence of various additives,l-s which can be thought of as an introduction to the difficult problem of micellar solubilization. Because of the complexity of the micellar system, these studies are confined to the region of the critical micelle concentration (c.m.c.) of the surfactant where the interaction between micelles is assumed to be negligible. The addition of hydrophilic or slightly hydrophobic molecules such as urea, acetone or methanol to an aqueous surfactant solution increases the c.m.c.; these molecules are believed to destroy the structure of water. However, strongly hydrophobic molecules such as tetrahydrofuran or the higher homologue alcohol molecules, which are thought to enhance the water structure in dilute solutions, decrease the c.m.c.We have thus formulated the following question : to what extent is the change in c.m.c. related to the influence of the additives on the structure of the medium? In order to investigate this problem one must determine the thermodynamic properties of the monomer ions of the surfactant below the c.m.c. We were able to show in a previous studysqg that the change in c.m.c. of ionic surfactants on the addition of slightly hydrophobic molecules such as acetone can be attributed to a large extent to the change in Gibbs free energy of the monomer ions on addition of the cosolvent. In the present study we consider the interaction in aqueous solutions of ionic surfactants with t-butyl alcohol (ButOH), which is known to decrease the c.m.c. in dilute solution.It was of interest to consider the enthalpy of solution, which is more sensitive to the structural characteristics of the medium than the free energy function, in the case of sodium decylsulphate (SDS) and trimethyldecylammonium bromide (n-DTMABr) and compare the results obtained with those published previously for the same electrolytes in water + acetone mixtures.1° 18271828 SURFACTANT TRANSFER ENTHALPIES I N H,O+ButOH In addition to the two ionic surfactants we have studied the enthalpy of solution in water + ButOH mixtures of sodium methylsulphate (MeOS0,Na) and of tetra- methylammonium bromide (Me,NBr), which may be considered as the lowest homo- logues of the anionic and cationic surface-active agents.EXPERIMENTAL Sodium decylsulphate (SDS) from Merck was used without purification, the c.m.c. (m = 0.0345 mol kg-') being in good agreement with literature values. n-DTMABr from Eastman Kodak was recrystallized three times from pure acetone; the c.m.c. (m = 0.0655 mol kg-l) was also used as a purity criterion. Both c.m.c. values were deduced from conductance measurements. Me,N Br, MeOS0,Na and ButOH (Merck) were used without purification. All salts were dried under vacuum before use. The submarine-type microcalorimeter used has been described previously.ll Briefly summar- ized, the experimental procedure was as follows. A 300 cm3 volume Dewar was immersed in a constant-temperature bath maintained at 25.000 & 0.001 "C.The solution was stirred with a Teflon agitator; a cup containing the salt was kept above the solvent surface and when the equilibrium temperature was reached it was pushed mechanically into the solvent. Before and after each experiment a calibration was performed using the Joule effect through a platinum resistance thermometer. RESULTS The lowest concentrations attained were ca. (1-2) x mol kg-l. These concen- trations are much lower than the c.m.c. of the two surfactants. Furthermore, above 12% by weight of ButOH in water, the c.m.c. could hardly be detected.12 All enthalpies of solution were corrected for the Debye-Hiickel ion-ion interaction effect using the necessary physical parameters of the water + ButOH mixtures.13-15 At high water concentrations (below 50 wt% or 0.2 mole fraction), the ion-pair association constants KA are small enough for the degree of association of the ions to be assumed to be unity.At higher ButOH concentrations the association constants become large and a correction term should be introduced to the Debye-Hiickel law. This was not done in the present study and a few comments are necessary. A recent study on the enthalpy of solution of copper sulphate in water16 has suggested that it might not be appropriate to correct the Debye-Hiickel term for ion association using an association constant deduced from another technique in order to reproduce the experimental relative partial molar enthalpy values. Furthermore, in ButOH-rich mixtures we were essentially interested in the thennochemical properties of the hydrocarbon chain attached to the ionic group on the surfactant ion.The corresponding quantity may be obtained by obtaining the experimental AHF values for the lower homologues from the data for the ionic surfactants. Under these conditions we have benefitted from the experimental observation that in such a series of homologues the association constants, as determined from the conductance method, are almost independent of the chain length of the longest hydrocarbon chain.l77 la This is due to the fact that the approach by the counter-ion to the head group of the ionic surfactant is hampered to the same extent whatever the hydrocarbon chain length. Therefore the A(AHF) values do not require a correction for ion-pairing. As an example, in pure propan- 1-01 at 25 OC the KA values of Me,NBr,lg n-BuMe,NBr,la n-HexMe,NBrls and n-OctMe3NBrls are equal to 638 5, 555 f 5, 551 f 5 and 558 +4, respectively.Even the small difference between the result for Me,NBr and those for the other electrolytes may be caused by the different conductance equations used to analyse the data. Table 1 presents the experimental results for the four electrolytes studied. The valuesR. BURY AND C. TREINER 1829 for Me,NBr may be compared to some recently published data in the same water + ButOH mixtures up to 40 wt%.20 The agreement is very good (ca. 1 %). The experiments for this electrolyte had to be restricted to a narrower range of ButOH concentration than for the other electrolytes because of solubility limits.It would have been possible to obtain rough AH? values by studying the enthalpy of dissolution TABLE HEATS OF SOLUTION IN WATER-I-t-BUTYL ALCOHOL MIXTURES AT 298.16 Ka SDS MeOSO, Na n-DTMABr Me,NBr wt % AV 103m AT 103m AT 103m AHS 103m 0 27.91 10 59.45 15 100.16 20 104.39 30 80.83 40 59.62 50 44.35 75 21.76 90 19.33 25 - 2.20 8.49 1.44 20.50 1.50 25.02 1.31 - 1.69 23.97 1.39 21.13 1.63 17.07 1.74 3.89 1.57 4.06 4.02 1.79 1.99 2.15 3.00 3.05 1.70 2.57 - - 3 1.80 59.16 88.70 99.58 99.04 93.39 67.36 49.75 34.73 32.64 2.36 24.48 1.30 1.83 27.28 1.52 2.01 - 1.44 29.37 1.52 2.29 - 1.87 29.41 1.14 2.01 - 2.14 - 1.92 17.28 1.15 - - - - - 2.17 - a A T is the heat of solution (kJ mol-l) at final molality m (mol kg-l). of higher homologues such as n-Pen,NBr, n-Bu,NBr and n-Pr,NBr, which are sufficiently soluble in ButOH-rich mixtures, and extrapolate the AH? values as a function of the molecular weight of the salts to that corresponding to Me,NBr.This is the basis of an extrathermodynamic approach used to obtain single-ion standard enthalpy values.21 However, the correction to the degree of association would not have been adequate, as discussed earlier, although the principle of the procedure might be useful in less difficult cases. DISCUSSION COMPARISON O F A% VALUES FOR ROD-LIKE AND BULKY IONS I N AQUEOUS BINARY SOLVENTS Fig. 1 presents the essential characteristics of our experimental results. There is a large endothermic effect for the two ionic surfactants with a maximum around 0.05 mole fraction of ButOH, as observed by many authors.22 One striking aspect of these data is the height of the endothermic effect, especially for SDS.It is the largest observed in water + ButOH mixtures [Ae(max) = 76.6 kJ moll. A value of 70.7 kJ mol-1 had been obtained previously by Arnett and McKelvey for the standard enthalpy of transfer from water to a 20 w t x solution of ButOH for sodium tetraphenylb~rate.~~ These results are more interesting when they are compared with those of a symmetrical electrolyte such as n-Bu4NBr2, (fig. 2). In water + ButOH mixtures the maximum A% value is much larger for n-DTMABr than for n-Bu,NBr. More generally, in these solvents the endothermic effect is larger the larger the size of the ions in a given series.2o However, in the water + acetone mixtures studied previously the reverse was observed: the endothermic effect was larger for n-Bu,NBr than for1830 SURFACTANT TRANSFER ENTHALPIES I N H,O+BUtOH 60 - b I I 1 I I I 1 1 1 50 wt % FIG.1.-Standard enthalpy of transfer of +, n-DTMABr; 0, SDS; 0, Me,NBr and 0, MeOS0,Na from water to water + ButOH mixtures: Aliz) z AW - d e . - P 'A I ' 4 \ I \ 1 4 I \ i \ wt 7c FIG. 2.-Standard enthalpy of transfer of n-DTMABr (dotted line) and n-Bu,NBr (full line) from water to water + ButOH (0, A) and water + acetone (0, A) mixtures.R. B U R Y AND C. TREINER 1831 n-DTMABr. The volume of the rod-like ionic surfactant being larger than that of n-Bu4N+, the result in water + ButOH mixtures is in agreement with the general rule, and it seems that it is the water+acetone mixtures which exhibit unusual behaviour.In order to rule out an interpretation of these phenomenon in terms of a difference in specific interactions it is useful to recall that theoretical models such as the scaled-particle have convincingly shown that in water-rich binary mixtures the thermodynamics of dissolution of non-electrolytes are governed by a cavity effect which overcomes a specific interaction effect. (Note that the respective contributions of n-DTMABr and Me,NBr to A* indicate that the electric charges do not change the magnitude of the observed effects; see last section and fig. 3.) According to the scaled-particle theory, the enthalpy function depends essentially on the hard-sphere diameter of the particles, and the density and expansibility of the solvent.Calculations based on this theory agree, in the case of water + ButOH, with the general rule stated above. However, the ionic surfactants are evidently far from having a spherical shape. One can thus argue that in the case of strongly structured solvents such as the water-rich ButOH mixtures, because of the difficulty in accommodating a solute above a certain size in the lattice of hydrogen-bonded solvent molecules, the gross properties of the medium adequately describe the solvent with respect to the rather large ions of the alkylammonium series. In the case of less structured mixtures such as water + acetone, a large but rod-like molecule may have the possibility of entering loose structures without much disruption of the less strongly hydrogen-bonded solvents, contrary to a bulky n-Bu,N+ ion.The situation is reminiscent of that of the mobility of small ions in solvents of high viscosity, such as in mixtures of an organic solvent with a soluble polymer. The mobility of the small ions is much larger in these media than one would expect from the macroscopic viscosity. This may be interpreted by assuming that the ions have the possibility of moving in the solvent through the tangle of solvent molecules by displacing only segments of the macromolecules without sensing the macroscopic hydrodynamic characteristics of the 27 Note that the initial slopes of the variation of AX with ButOH concentration do not present the complication just outlined. If the trends are similar the differences between the initial slopes corresponding to n-DTMABr and n-Bu,NBr are very small in both binary solvents.It was not the purpose of the present study to investigate very water-rich mixtures; however, fig. 2 seems to indicate that in this region (which corresponds to the Henry’s law region for the free-energy function) thc: structure of water changes smoothly in comparison with the region of maximum endothermic effect, thus permitting theoretical approaches to the enthalpy function in terms of pair and triplet interaction effects.28 We have pointed out the specificity of the surfactant ions in terms of structural features in the case of water + acetone and water + ButOH mixtures. However, the two organic molecules induce an opposite change of c.m.c. when added to an aqueous surfactant solution. It may thus be interesting to discuss the possible relevance of structural considerations to the micellar solubilisation phenomenon.It is generally accepted that the c.m.c. decrease on the addition of a hydrophobic organic molecule is due to some type of interaction between the additive and the micelle. However, these molecules are also thought to be responsible for the reinforcement of the water structure through an increase in the strength of the water hydrogen bonds. Thus these molecules (e.g. ButOH) might enhance the very structure of the medium which favours micelle formation. That the composition of the water + ButOH mixtures for which the maximum endothermic effect is observed also corresponds to the composition of a solid lathr rate^^ confirms this point of view.1832 SURFACTANT TRANSFER ENTHALPIES I N H,O+ButOH The very large difference observed between the maximum A% value for n-DTMABr in water+ButOH and the much less structured mixtures of water+acetone (ca.54 kJ mol-l) may be taken as further evidence of the importance of the strength of the hydrogen bonds in an aqueous surfactant solution to the variation of the c.m.c. on the addition of a particular cosolvent. The addition of acetone to water is known essentially to destroy the water structure, hence creating a medium that is increasingly less favourable to micelle formation, which will result in an increase in the c.m.c. The same phenomenon might explain the increase in the c.m.c. after an initial decrease in the case of some water+alcohol mixtures.It then appears that the effects of structure and of hydrophobic interactions (which are responsible for the micelle- additive interactions) are closely related phenomena. 1 I I I I I I 1 1 50 wt 7c FIG. 3.-Contribution to A% of 0, (CH,): and 0, (CH,), from water to water+ButOH mixtures. CONTRIBUTION OF NONYL HYDROCARBON CHAINS TO THE STANDARD ENTHALPY OF TRANSFER We now consider the change in standard enthalpy of transfer from water to water + ButOH mixtures for the n-nonyl hydrocarbon chain attached to the cationic A% (CH,); and to the anionic A% (CH2); head groups (fig. 3). This was obtained by simply subtracting the A e values of Me,NBr and MeOS0,Na from those of their higher homologues. We obtain the same overall results as with the acetone + water mixtures.1° A% (CH,); is more positive below the maximum than A% (CH,);, and this difference, ca.4 kJ mol-l, increases to ca. 12 kJ mol-l beyond the maximum. In so far as it is accepted that the maximum corresponds to the most stabilized structure of the m e d i ~ m , ~ ~ . ~ ~ we may consider that the difference observed is essentially due to the relaxation of the water molecules around the methyl groupsR. B U R Y A N D C. TREINER 1833 nearest to the head group. The results show that the water molecules near the anionic group are more subject to its electrostatic influence than those near to the cationic group. Furthermore, this effect is much larger with ButOH than with acetone. One may conclude that in organic-rich mixtures the water-acetone interactions are stronger than water-ButOH interactions.This is in agreement with what would be expected if the relative values of the dipole moments of acetone ( p = 2.66 D) and ButOH ( p = 1.66 D) were taken as being indicative of the strength of the solvent-solvent interactions in organic-rich binary mixtures. Finally, in the case of water+acetone mixtures it was possible to show that the contributions of A% (CH,); and A% (CH,), to the standard enthalpy of transfer were equal in pure acetone, confirming our interpretation. Such evidence cannot be presented in the present case because of the solubility limits of Me,NBr discussed previously. This would imply the presence of a shallow maximum in the AW, profile for A% (CH,): at high ButOH concentrations. I M. Manabe and M. Koda, J.Colloid Interface Sci., 1980, 77, 189. K. Hayase and S. Hayano, J. Colloid. Interface Sci., 1978, 63, 446. :' H. N. Singh, S. Swarup and S. M. Saleem, J . Colloid Interface Sci., 1979, 68, 128. H. N. Singh and S. Swarup, Bull. Chem. SOC. Jpn, 1978, 51, 1534. K. Shirahama and T. Kashiwabara, J. Colloid Interfuce Sci., 1971, 36, 65. S . Miyagishi, Bull. Chem. Soc. Jpn, 1974, 2972, C. Treiner, A. Le Besnerais and C. Micheletti, Adv. Chem. Ser., 1979, 177, 109. C. Treiner and A. Le Besnerais, J. Chem. Soc., Farudq Trans. I , 1977, 73, 44. fi S . Miyagshi, Bull. Chem. SOC. Jpn, 1975, 48, 2349. lo R. Bury and C. Treiner, Can. J. Chem., 1978, 56, 2940. ' I R. Bury, A. Mayaffre and M. Chemla, J. Chim. Phys., 1977, 78, 745. l 2 M. F. Emerson and A. Holtzer, J. Phys. Chem., 1967, 71, 3320. l 3 Y. Pointud, J. Juillard, L. Avedikian, J. P. Morel and M. Ducros. Thermochim. Acta, 1974, 8, 423. l 4 J. Kenttamaa, E. Tommila and M. Martti, Ann. Acad. Sci. Fenn., Ser. A, 1959, 93, 3. Plenum Press, New York, 1973), vol. l 5 J. B. Hasted, in Water, a Comprehensive Treatise, ed. F. Franks l6 R. Bury, A. Mayaffre and M. Chemla, C.R. Acad. Sci., Ser. C l 7 B. Sesta, Rev. Roum. Chim., 1975, 20, 473. B. Sesta and G. Goratti, Rev. Roum. Chim., 1975, 20, 37. l9 D. F. Evans and P. Gardam, J. Phys. Chem., 1968, 72, 3281. 2o J. Juillard, J. Chem. SOC., Faraday Trans. I , 1982, 78, 37. 11. 1980, 291, 259. 21 H. L. Friedman and C. V. Krishnan, in Water, a Comprehensiue Treatise, ed. F. Franks (Plenum 22 J. B. F. N. Engberts, in Water, a Comprehensive Treatise, ed. F. Franks (Plenum Press, New York, 23 E. M. Arnett and D. R. McKelvey, J. Am. Chem. Soc., 1968, 87, 1393. 24 R. K. Mohanty, T. S. Sarma, S. Subramanian and J. C. Ahluwalia, Trans. Faraday Soc., 1971, 67, 25 R. A. Pierotti, Chem. Rev., 1976, 76, 717. 26 D. J. Mead and R. M. Fuoss, J. Am. Chem. SOC., 1945, 67, 1566. 27 C. Treiner and R. M. Fuoss, J. Phys. Chem., 1965, 69, 2576. 2R J. E. Desnoyers, D. Joly, S. Leger, G. Perron and J. P. Morel, J. Solution Chem., 1976, 5, 681. 29 K . Iwasaki and T. Fujiyama, J. Phys. Chem., 1977, 81, 1907. Press, New York, 1973), vol. 111. 1979), vol. VI. 305. (PAPER 1 / 1 127)

ISSN:0300-9599    

DOI:10.1039/F19827801827

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. SOL'., Furaday Trans. 1, 1982, 78, 1835-1839 Photolysis of Solid KC10, at 185 nm Studied by Chemical Analysis, Electron Spin Resonance and Optical Spectroscopy B Y U L R I K K. KLANING, NIS BJERRE AND J 0 R G E N R. BYBERG* Department of Chemistry, Aarhus University, DK-8000 Aarhus C, Denmark Received 16th July, 1981 0, and ClO; have been established as primary products in the photolysis of KClO, at 185 nm, whereas C1- is the dominant product of secondary photolysis. The primary photolysis resembles the radiolysis of KClO,, with the exception that C1- is a primary product in radiolysis but not in photolysis. It is proposed that the primary photolytic process is photoionzation, in which an electron is transferred between adjacent anions. The resulting charge-transfer complex may account for the observed transient optical absorptions.Photolysis of solid alkali-metal perhalogenates inevitably leads to reduction of the perhalogenate ions, implying that oxygen is a photolytic product. It has been shown' that molecular oxygen is formed in perhalate crystals during photolysis at 26 K, which suggests that 0, arises in the primary photolytic process. In the present investigation of the photolysis of crystalline KClO, at 185 nm, the so far unknown photolytic products complementing 0, are determined, and the primary photolytic process is discussed in the framework of earlier results for the radiolysis of KC10,.2-4 EXPERIMENTAL KClO, was Merck p.a., recrystallized three times from triply distilled water. The final solution was filtered through a Whatman GF/F filter with pore size 0.7 pm.Single crystals were grown by evaporation at ca. 30 OC. H,SO, was Merck Suprapur. Other chemicals were analytical grade. Ultraviolet irradiations were made with filtered and unfiltered light from a Thermal Syndicate TM5/524C no. 2 low-pressure mercury lamp fitted with a pair of lenses made from Spectrosil. The filters were an Acton Research Corp. 185N interference filter (transmissions 21 % at 185 nm and 0.2% at 253.7 nm) and a Schott 1 mm UG 5 filter. 0.1 g samples of KClO, powder of grain diameter (1-3) x cm were irradiated in an aluminium tray lined with platinum foil. The tray was placed in a Dewar containing liquid nitrogen on a vertical brass rod which was partly submerged in the liquid nitrogen.The tube enclosing lamp, lenses and filters, placed above the tray, was flushed with dry nitrogen gas. The wet analysis of the irradiated KClO, powders was made as de~cribed.~ Optical spectra of single crystals were recorded with a Cary 2 19 spectrophotometer equipped with an AC Heli Tran LT-3-110 cryostat operated with liquid nitrogen. The crystals were plates approximately 0.1 cm thick with faces parallel to (001). A selected area measuring ca. 0.3 x 0.8 cm was bounded by a mask made from aluminium foil, and the crystal was glued to the copper sample holder of the cryostat with epoxy resin. The analysing beam was incident on (001). A Glan prism polarizer was inserted between the sample crystal and the phototube. The spectra were corrected for the absorbance of the unirradiated crystal and the polarizer.U.V. irradiations were carried out in the sample compartment of the spectrophotometer 18351836 PHOTOLYSIS OF SOLID KCIO, without moving the cryostat. During recording of spectra in the region 185-220 nm and during irradiations, the spectrometer was flushed with dry nitrogen. E.s.r. spectra of single crystals were recorded with a Varian E-15 spectrometer operated at 9.3 GHz, usingat 77 K the Heli Tran LT-3 cryostat with liquid nitrogen and at 26 K an AC-2-110 Cry0 Tip. Irradiations with X-rays were carried out in the V 453 1 microwave cavity as described previously. RESULTS WET ANALYSIS C1-, C10-, C10; and Clog were detected in aqueous solutions of KClO, powder irradiated with the low-pressure Hg lamp at 77 K. Fig.1 shows the relative yields of these products against the degree of decomposition of KCIO,. The curves show that 0 h 20 - p " I OO 0.2 0.2 0.6 3 decomposition of KClO, (mole %) FIG. 1.-Relative yields of the photoproducts C10; (e), C10; ( x ) , C10- (A) and C1- (0) against the decomposition of KC10, resulting from photolysis at 77 K. Squares c] indicate measurements on samples irradiated through a 185 N interference filter. The remaining samples were irradiated with unfiltered light. The sample having 0.75% decomposition was irradiated for 7 h. C10; is the major primary product and that C1- is a secondary product. Due to the very low yields, the data allow no conclusions regarding ClO; and C10-. After irradiation with unfiltered light the powder was yellow, whereas it acquired a violet tinge following irradiation through the 185 N interference filter.However, the relative yields of photoproducts were the same in the two cases, as indicated in fig. 1 . No photoproducts were detected in powders irradiated through the UG 5 filter. OPTICAL SPECTRA A KClO, crystal was irradiated through the 185 N filter for 3 h. The resulting optical absorption in the range 220-800 nm, recorded with the electric vector E parallel to the 4- and b-axes, is shown in fig. 2, curve (a). The absorbance in the range 190-220U. K. KLANING, N. BJERRE AND J. R. BYBERG I o . o ~ , , l , l l l , l l , , , 200 300 400 500 600 700 801 1837 h/nm FIG. 2.-Absorbance of a KClO, crystal after irradiation through a 185 N inteference filter for 3 h at 77 K.Curve (a), absorbance immediately after the irradiation. Curve (b), absorbance after bleaching for 20 min at 254-400 nm. Curve (c), curve (b) subtracted from curve (a). Curve (d), absorbance after warming to room temperature. The electric vector, E, of the analysing beam is parallel to the crystallographic axes a (-) and b (- - -) in curves (a)-(c), whereas unpolarized light is used in curve (d). The absorbance of the unirradiated crystal has been subtracted. nm, which could be measured with unpolarized light only, increased steadily with decreasing wavelength. The crystal was subsequently bleached by irradiation for 20 min through the UG 5 filter. As shown in fig. 2, curves (b) and (c), the bleaching has removed the band at 530 nm for Ella, a very broad absorption for Ellb ranging from 300 to 800 nm, and an almost unpolarized band at 250 nm.The absorbance below 210 nm increased slightly during bleaching. The change in colour of the irradiated powders from violet to yellow apparently stems from bleaching of the 530 nm band. Only the strong absorption below 220 nm and a weak band around 300 nm survive annealing at room temperature. These bands correspond to absorptions of the stable photoproducts : C10- and C10; have bands around 300 nm and C10-, C10; and C10; all absorb below 220 nm. ELECTRON SPIN RESONANCE SPECTRA The e.s.r. spectrum of a KC10, crystal, which has been irradiated through the 185 N filter for a few minutes at 26 K, contains several sets of the strongly anisotropic transitions characteristic of O,.l No e.s.r.signals from paramagnetic species in spin1838 PHOTOLYSIS OF SOLID KC10, doublet states have been detected. The e.s.r. spectra indicate that 0, is trapped in several inequivalent configurations. Moreover, the growth of the signals during irradiation shows that some configurations correspond to primary defects, whereas others are formed by secondary photolysis of primary defects. No e.s.r. signal from ClO, was detected in KClO, crystals after photolysis at 26 or 77 K and subsequent irradiation with X-rays. This indicates that the amount of ‘free’ Cloy was < I ppm.6 On the other hand, if the crystals were annealed at room temperature between photolysis and X-ray irradiation, C10, in two inequivalent sites was observed, indicating the presence of C10; in two different environments.DISCUSSION Since the electronic transitions leading to photolysis are also induced by secondary electrons during irradiation with X-rays,’ we expect that the radiolytic processes include the primary photolytic process and, hence, that the primary photolytic products are a subset of the radiolytic products. Indeed, C10, is the principal radiolytic product in KC10,,2 and e.s.r. spectra of 0, as well as transient optical absorptions at 430 and 530 nm are also observed in radiolysed KC104.8 The latter finding in turn suggests that these absorptions belong to primary products in photolysed KCIO,. The photolytic process corresponding to the primary products C10; and 0, clearly involves two Cloy, so that we may write the net process hv 2c10, + 2c10, + 0,.However, process (1) does not account for the transient optical absorptions. As described earlier, radiolysis of KClO, leads to formation of spatially separated electron-excess and hole defect^.^ The separation between the electron and the corresponding hole indicates that the excitation energy exceeds the ionization potential. At the threshold for ionization of CIO,, electron transfer between two adjacent anions may occur. * We propose that this limiting case of photoionization is the primary photolytic process in KClO,. We assume, moreover, that the relaxa- tion mechanisms of the electron-excess and hole components of the charge-transfer complexes are similar to those of separated electrons and holes, thus leading to defect structures of the type [Cloy, 0-, O,, C10,].3* Optical absorptions of complex defects have been attributed to charge transfer between the components of the defects.8 In the present case, the observed absorptions may similarly arise from charge transfer within the complexes.The decay of these absorptions then indicates a thermally activated collapse of the complexes to 2C1O;+O2 and a subsequent escape of 0,, in accordance with the detection of ‘free’ C10; after annealing. The absence of e.s.r. signals from 0- and C10, is probably due to a coupling between the electron spins, analogous to that observed between the spins of C10, and 02,3 resulting either in a singlet ground state or in an extremely efficient spin-lattice relaxation which broadens the e.s.r. spectra beyond recognition. If the coupling also involves the spins on O,, the observed e.s.r. spectra of 0, must arise from complexes that have already collapsed to 2C10,+O2. The high-quality crystals of KClO, used for optical measurements were grown by Mr 0. Lillelund. * Electron transfer between adjacent anions was proposed earlier to account for the formation of defects involving two chlorine nuclei in the radiolysis of KClO, (R. S . Eachus and M. C. R. Symons, J . Chem. Soc. A , 1968, 2433).u. K. KLANING, N. BJERRE AND J. R. BYBERG J. R. Byberg, Chem. Phys. Lett., 1978, 57, 579. J. R. Byberg and J. Linderberg, Chem. Phys. Lett., 1975, 33, 612. J. R. Byberg, J . Chem. Phys., 1981, 75, 2663. T. Chen, Anal. Chenz., 1967, 39, 804. J. R. Byberg, Chem. Phys. Lett., 1978, 56, 563. F. Williams, Radiat. Chem. Macromol., 1972, 1, 723. N. Bjerre and J. R. Byberg, J . Chem. Phys., 1981, 75, 4776. N. Bjerre and J. R. Byberg, Phys. Rev. B, 1979, 20, 3597. * L. A. Prince and E. R. Johnson, J. Phys. Chem., 1965, 69, 359; 1965, 69, 377. 1839 (PAPER 1 / 1 132)

ISSN:0300-9599    

DOI:10.1039/F19827801835

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1 , 1982, 78, 1841-1846 Spectroscopic Evidence for Hydrogen-bond Formation between Methanol and Halogenoalkanes BY MARTYN C. R. SYMONS,* NICK G. M. PAY AND GRAHAM EATON Department of Chemistry, The University, Leicester LE 1 7RH Received 20th July, 1981 Addition of mono chloro-, bromo- or iodo-alkanes to methanol in tetrachloromethane results in the loss of the OH stretching bands characteristic of monomeric solvent and the growth of distinct bands due to Rhal- - -HOMe hydrogen-bonded complexes. Equilibrium constants ( K ) , estimated from the spectra, show that the hydrogen-bond strength increases in the series RCl, RBr, RI. Shifts and equilibrium data for nitromethane and cyanomethane have been obtained for comparative purposes. The band maxima increase monotonically and almost linearly with K for this series of weak bases.Concomitant decreases in the carbon-halogen stretching bands have also been monitored. The significance of these results to the mechanism of SN1 reactions is discussed. Solvolysis of alkyl halides has been one of the mosL thoroughly studied reactions,l and is still receiving wide attention.2 One of us6 has recently expressed the view that S,1 type solvolyses can best be understood in terms of thi initial formation of hydrogen bonds to the halide entity, this model being favoured over the more usual but less specific concept of transition-state solvation. For aqueous systems in particular, reaction with OHf groups6 was invoked, the most favoured sequence being Rhal + OH, $ Rhal- - -HO f Rhal- - -HO c + OH, Rhal- - -(HO r ), + R+ + hali0,, (1) (2) (3) Rhal-- -(HO r ), step (3) being rate-determining.This process provides a semiquantitative explanation of the effect of basic cosolvents on rate,6 and possibly also of the marked non-linearity of the Arrhenius plots.5 Furthermore, eqn (1)-(3) can be taken as a possible alternative to the ion-pair concept which has recently been invoked to explain the curvature observed for Arrhenius plots for these reaction^.^ It therefore seemed important to obtain independent evidence for the formation of hydrogen bonds as required by equilibrium (1). In our experience, the best method for studying hydrogen-bond formation between protic solvents and bases is to measure the O-H stretching frequency for ternary systems using inert solvents.Tetrachloramethane was selected since its basicity is less than those expected for monohalogenoalkanes, and it is easier to use than more inert solvents such as alkanes. Furthermore, it has been widely used as a solvent for studies of this t ~ p e . ~ - ~ Nevertheless, the shifts and equilibrium constants reported herein must be viewed as being relative to tetrachloromethane rather than being absolute values. The behaviour of nitromethane and cyanomethane has been studied for comparative purposes since these are accepted as typical weakly basic solvents capable of forming hydrogen bonds to protic solvents such as water or methanol. We selected methanol as a typical protic solvent because, for the monomers, its infrared spectrum in the O-H stretch region, being simpler than that of water, is easier 60 1841 t A R 11842 METHANOL - HALOGENOALKANE HYDROGEN BONDING to monitor quantitatively.Also, it is clear that the hydrogen-bonding ability of monomeric methanol in tetrachloromethane is comparable with that of monomeric water.* We have, however, established that, qualitatively, the two solvents, methanol and water, are comparable in their interactions with halogenoalkanes. We have studied the fundamental and first overtone of the OH stretch in order to check the accuracy of our results. Because both the halogenoalkanes and the solvent tetrachloromethane are weakly basic, we decided to express the equilibria as competitions for the hydroxy protons: (4) MeOH- - -ClCCl, + Rhal + MeOH-- -halR + CCl,.The equilibrium constants given in table 1 are defined by [MeOH- - -halR] [CCl,] [MeOH- - -ClCCl,] [Rhal] K = the narrow ‘monomer’ band being used to estimate [MeOH- - -ClCCl,]. This treatment gave far less scatter in our estimates of Kthan the more conventional expression, which ignores the r6le of CCl,. This competition is probably only of significance in studies of weak bases. The absorbance for the monomer 0-H band is well-established,1° and in all experiments Beer’s law was obeyed. All other concentrations in eqn ( 5 ) were deduced from direct measurements. Results for the systems CCl,+MeOH+Rhal are illustrated in fig. 1 and 2, and are summarised in table 1 together with results for the CCl,+MeOH+MeCN and CC1, + MeOH + MeNO, systems. The correlations between shift [Av = V(OH),,,~ - v(OH),,,,] and Kare shown in fig.3. In all studies, reasonably good isosbestic points were obtained (cf. fig. 1). This suggests that we are dealing with two well-defined states for methanol, rather than with non-specific solvent shifts. The correlation between shift and equilibrium constant for the basic solvents MeNO, and MeCN confirms that this interaction may reasonably be described by eqn (4). The results suggest that RBr molecules have about the same basicity as MeNO, molecules, and that the basicity increases in the sequence RC1, RBr, RI. AG values estimated from the overtone data are generally lower than those from the fundamental infrared shifts, although the errors are large. We have no explanation for these differences other than to suggest that they do indeed stem from experimental errors.The sequence of hydrogen-bond strengths, RCl < RBr < RI is at first sight surprising, since for the halide ions, C1- > Br- > I-.l0 Two factors probably contribute to this change. Clearly the presence of the carbon-halogen bond has greatly reduced the halogen basicity. Since the bond strengths fall in the series C-Cl > C-Br > C-I, this reduction may also fall in this order. Furthermore, since the hydrogen bonds are now very weak, the ability of the halogen atom to polarise so as to give a partial negative charge at the bonding site may be important, and this increases in the series R-C1 < R-Br < R-I. These two factors are apparently just sufficient to invert the order observed for the ions.Similar spectral changes were observed for aqueous systems. Note that these solutions are stable, there being no observable tendency for loss of halide ions at room temperature. This may be taken as some support for the concept that a second hydrogen bond is needed for chemical reaction [eqn (2)]. However, we stress that the hydrogen bonds formed by monomeric methanol or water are only approximately half as strong as those formed by bulk methanol or water.ll Thus the interaction depicted in eqn (1) will be stronger than that studied herein. It will therefore not be possible to utilise the present data in any attempt to quantify kinetic data for S,1 processes in water or methanol.0.20 0.18 0.16 0.14 ; 0.12 -e 2 2 0.10 0.08 0.06 0.04 0.02 0 - M.C. R. SYMONS, N. G. M. PAY AND G. EATON 9 1343 3700 3600 3500 wavelength/nm FTC 1 -FiindamPntal nu ctretrhinu hand fnr methannl in tPtrirhlnrnmPthine q c Q fiinrtinn nf o A A J I &.,_ & . - U I I Y Y . I I - . I I C . . . -a- Y . * - . C I I . I . I "..*... ."I ...-....~*. "I '11 C V C I - I l l l V l V l l l r C l l U l l r u.3 u ILLIIGLIWII W l UUUGU iodomethane. Mole fraction of MeI: (1) 0.00, (2) 0.13, (3) 0.24, (4) 0.44, (5) 0.61, (6) 0.76, (7) 0.89, (8) 0.94, (9) 1.00. We have also studied the C-C1 stretching band for a range of alkyl chlorides and find that there is a well-defined change on hydrogen bonding, with a shift of ca. 10 cm-l to low frequencies. This is important since it suggests that there is a small but significant weakening of the C-Cl bond on forming one hydrogen bond.Although 10 cm-l is small, it is nevertheless a ca. 2% change, which is approximately twice as big as the relative change in the 0-H stretching frequency. Thus the system is indeed moving towards the heterolytic cleavage required for the S , 1 mechanism. EXPERIMENTAL N E AR-I N FR A RED Samples were measured on a Perkin-Elmer 340 u.v.-visible-near-infrared spectrometer with Hellma cells (10 mm) thermostatted to kO.2 OC. All work was carried out at 25 OC (298 K). The cell compartment was flushed continually with N,. Samples were made up accurately by volume using Eppendorf pipettes and Hamilton syringes. All chemicals were purified by distillation and kept over the appropriate drying agent in a dry box flushed with nitrogen.The results were calculated using a computer program written in BASIC running under nos. 1.4 on a CYC CYBER 73. 60-21844 METHANOL - HALOGENOALKANE HYDROGEN BONDING 1 .o 0.8 0.6 cd -e z P 0.4 0.2 1400 1440 wavelength/nm FIG. 2.-As fig. 1, in the first overtone regon. Mole fraction of MeI: (1) 0.0, (2) 0.44, (3) 0.61, (4) 0.89, ( 5 ) 1.00. 20 '40 60 80 100 Av/cm-' FIG. 3.-Correlation between the shift of the OH stretching frequency (Av) and AG for equilibrium (4); (a) from the fundamental region, (6) from the overtone region. (A) 2-Chloropropane, (B) 2-bromopropane, (C) 2-iodopropane, (D) cyanomethane, (E) nitromethane.M. C . R. SYMONS, N. G. M. PAY AND G. EATON 1845 TABLE 1 .-INFRARED SHIFTS AND EQUILIBRIUM CONSTANTS FOR THE FORMATION OF METHANOL-HALOGENOALKANE ADDUCTS fundamental overtone shift, shift, Rhal Av1cm-l K Av1cm-l K 1.66 f 0.7 33 0.95 & 0.6 1.87 f 0.5 2PrC1 24 1.25 f 0.24 2.04 f 0.4 1.21 f 0 .2 2PrBr 36 3.73 f 0.8 3.10k0.6 3.09f0.5 2PrI 47 4.46 f 0.9 4.02 k0.7 3.91 f0.6 MeCN 76 6.83 f 0.9 7.22 & 0.9 7.53 f 0.9 53.5 2.2 * 0.2 1.9f0.3 1.9 0.3 74 3.73 f 0.3 3.61 f0.5 3.5f0.3 104.5 6.5 & 0.3 7.1 f 0.5 6.86 f 0.5 MeNO, 36 2.0 f 0.7 51 3.08 k 0.4 2.8 f 0.8 3.08 f 0.4 2.9 f 0.9 3.00 f 0.3 138 140 142 144 146 148 wavenumberlcm-' FIG. 4.-Absorption spectrum for methanol (OH overtone) in tetrachloromethane containing iodomethane. The thick full line is the experimental curve plus the envelope curve computed by the addition of the two GaussianjLorentzian curves shown by dashed lines.(a) (OH)free oscillators, (b) OH ---I-CH, oscillators. FUNDAMENTAL INFRARED Spectra were recorded using a Perkin-Elmer 580 spectrophotometer, with Infrasil cells (1 mm). Samples were made up by volume (as above). SPECTRAL ANALYSIS Deconvolutions of the spectra were achieved using an interactive graphics program ; an example is shown in fig. 4. This provides for the linear combination of bands defined by several variable parameters, i.e. height, position, width at half height and profile (Lorentzian to1846 METHANOL - HALOGENOALKANE HYDROGEN BONDING Gaussian, defined by percentage Gaussian). The ensuing envelopes were compared graphically with the previously digitised spectra to ensure an optimum fit. Band widths and profiles were found to be constant so that peak heights could be used directly for the determination of K values. We thank N. J. Fletcher and M. J. Blandamer for helpful discussions. C. K. Ingold, Structure and Mechanism in Organic Chemistry (G. Bell & Sons, London, 1953). D. J. Raber, J. M. Harris, R. E. Hall and R. von Schleyer, J. Am. Chem. SOC., 1971, 93, 4821. J. M. Harris, A. Becker, F. A. Fagan and F. A. Walden, J. Am. Chem. Soc., 1974, 96, 4484. T. W. Bentley and R. von Schleyer, Advances in Physical Organic Chemistry, ed. V. Gold and D. Bethall (Academic Press, London, 1977). M. J. Blandamer, J. Burgess, P. P. Duce, R. E. Robertson and J. W. M. Scott, J. Chem. SOC., Chem. Commun., 1981, 13. M. C. R. Symons, J . Chem. Soc., Chem. Commun., 1978, 418; Ace. Chem. Res., 1981, 14, 179. R. Mecke, Discuss. Faraday SOC., 1950, 9, 161. M. C. R. Symons, T. A. Shippey and P. P. Rastogi, J. Chem. SOC., Faraday Trans. I , 1980,76, 2251. S. C. Mohr, W. D. Wilk and G. M. Barrow, J. Am. Chem. Soc., 1965, 87, 3048. lo R. R. Ryall, H. A. Strobe1 and M. C. R. Symons, J. Phys. Chem., 1977,81, 253. l 1 M. C. R. Symons, Philos. Trans. R. SOC. London, Ser. B, 1975, 272, 13. (PAPER 1 / 1 147)

ISSN:0300-9599    

DOI:10.1039/F19827801841

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. Soc., Faraday Trans. I , 1982, 78, 1847-1866 Squarilium Dyes, their Infrared and Resonance Raman Spectra and Possible Use in Redox Reactions for Solar-energy Conversion? BY MARTIN FORSTER~ AND RONALD E. HESTER* Department of Chemistry, University of York, Heslington, York YO1 5DD Received 20th July, 1981 Infrared and resonance Raman spectra of squarilium dyes I-VI of the type 0 - HO + --N(C6H5), (IV), -NHC,H, (V), -NHC,H4SO;H2N(CH3), (VI) are presented and some vibrational modes of the central 4-membered ring are assigned. Methyl viologen dication (MV2+) is reduced photocatalytically by IV in DMSO+H,O (95/5. v/v) with a quantum yield 4 of (1.8 k0.4) x at 406.7 nm, if EDTA is present. Oxidation products of MV2+ behave similarly with 4 = (4.7f0.6) x at 457.9 nm in DMF+H,O (2/1, v/v).IV'+ seems to oxidize H,O nearly reversibly. The search for photochemical systems for the conversion and storage of solar energy has become intense in the past few years. One is the photolytic splitting of water into H, and O,, a problem which has been tackled in various ways. Restricting ourselves to redox reactions in solution, two general reaction schemes have been proposed,l where the electronically excited state S* of a sensitizer S acts either as a reducing agent (scheme 1) or as an oxidizing agent (scheme 2): scheme 1: hv s + s* S*+A -+ S'++A'- A'-+H20 + A+iH,+OH- S'++$H20 + S+;O,+H+ cat 2 cat 1 t A preliminary account of this work has been presented previously, cf. M. Forster and R. E. Hester, Proc. 3rd Int. Con$ Photochem. Conversion Storage Solar Energy, Boulder 1980 (SERI, Golden, Colorado, 1980).3 Present address: Schweizerisches Institut fur Nuklearforschung, CH-5234 Villigen, Switzerland. 18471848 SQU A R I LI U M D YES FOR SOL A R-ENER GY CONVERSION scheme 2: hv s+s* S'-+H,O -+ S+;H,+OH- cat 1 A and D are electron acceptors and donors, respectively, and cat 1 and cat 2 are catalysts for the promotion of water reduction or oxidation, respectively. In order to make the reactions in schemes 1 and 2 efficient, S, A, D, cat 1 and cat 2 have to fulfil a large number of condition~,l-~ the most important of which are: (i) A,,, of S is in the visible region, (ii) the lifetime of S* is as long as possible, (iii) the redox potentials of S, S*, A and D match those of water at pH 7, viz. E0(02/H,0) = +0.82 V, E,(H,O/H,) = -0.41 V, (iv) there is efficient separation of the ion pairs S'++A'- or S'-+D'+ and (v) the specific activity of cat 1 (e.g.colloidal noble metals1v3) and cat 2 (e.g. RuO,*) must be right for the desired reactions. In the search for new organic sensitizers S we have investigated squarilium dyes of the type shown in fig. 1. Squarilium dyes have the following properties: (a) they can easily be prepared by the condensation of squaric acid with the appropriate RH,5 (b) they are stable even at temperatures T > 200 O C and in concentrated H2S04,5 (c) they have extinction coefficients up to 4 x lo5 dm3 mol-l cm-16-11 and ( d ) their Amax can be shifted over the whole range of 400-800 nm by varying R. For dye IV in fig. 1 [R = -N(C,H,),] the redox potentials of the excited state S* can be calculated using E,(S'+/S) = + 1.37 V, E,(S/S'-) = - 1.23 V12 and A,,, = 571 nm for emission (instead of the 0-0 transition) by adopting a formalism originally proposed for metal complexes.2 Fig.2 shows the redox levels for S = IV, which lie very close to those of Ru(bpy),2+ (bpy = 2,2'-bipyridine)13-15 and which fulfil criterion (iii). Ru(bpy)i+ has already been established as an efficient sensitizer for the photolytic decomposition of water.lg 4, 16-18 Some vibrational modes belonging mainly to the central part of dyes I-VI have been derived for infrared (i.r.) and resonance Raman (r.R.) spectra. The behaviour of dyes IV and VI in photoredox reactions according to scheme 1 and fig. 2(a) will be reported.Oxidation products of MV2+ dichloride (MV2+ = methyl viologen or 1 ,l'-dimethyl- 4,4'-bipyridinium dication) are shown to be even better photosensitizers than IV. The chemical oxidation of dye IV in aqueous solution was also investigated. EXPERIMENTAL EDTA (ethylenediamine tetra-acetic acid disodium salt), Ce(SO,), . 4H,O and PtO, (all from B.D.H.) were analytical grade and used as received. NaLS (sodium laurylsulphate) (Aldrich) was twice recrystallized from eihanol. MV2+ dichloride from I.C.I. Ltd was recrystallized from methanol + acetone five times.lg All solvents were analyticalgrade; DMSO (dimethylsulphoxide) was twice vacuum distilled and H,O was doubly distilled. Prepurified N, gas was further 0,-scrubbed by bubbling through'a solution consisting of 8 g pyrogallol and 24 g KOH in 200 cm3 H,O before being dried at 77 K.Dye I (Aldrich) was used as received. Dyes 11-V were prepared according to ref. (20) by refluxing one equivalent of squaric acid and two equivalents of the appropriate compound RH in benzene n-butyl alcohol (1 : 2.5, v/v) with concomittant azeotropic removal of the H,O formed. For dye VI a mixture of benzene+n-butylM. FORSTER AND R. E. HESTER 1849 0- 0- I R = H-0, R = R = p 0 -N -N R = $, + H~N(C H&, cl 0- I 0- 0- FIG. 1 .-Investigated squarilium dyes I-VI; ( a ) and (6) are possible mesomeric structures.1850 SQU A R I L I U M DYES FOR SOL A R-E N E R G Y CON V E R S I 0 N V +1.37 + 0.82 - 0 . 4 1 - 0 . 4 4 - 0.8 0 V \ 1--- products/ E DTA = D -1.23 - s_"/ s" 02/H20 = D' Y (0) b) FIG.2.-Redox levels of dye IV as sensitizer S in reactions according to scheme 1 or 2. (a) S* as reducing agent; (b) S* as oxidizing agent. MV2+ = methyl viologen = l,l'-dimethy1-4,4'-bipyridinium dication; EDTA = ethylenediaminetetra-acetic acid disodium salt; A = electron acceptor; D,D' = electron donors. Normal potentials of S calculated from ref. (12). alcohol+dimethylformamide (DMF) was used as solvent, yielding the salt VI, due to the decomposition of DMF into (CH,),NH and CO at elevated temperatures.21 Dye IV was recrystallized once from DMF + DMSO to remove an impurity with Amax = 375 nm in DMSO (probably the 1,2-isomer of dye IV). Analytically pure dyes 11-VI were obtained following these procedures. Table 1 gives the analytical and spectroscopic data for dyes I-VI.Elemental analysis data of dye VI were: C/H/N = 46.68/5.09/10.89 (theor.), 45.96/5.02/10.79 (found). A colloidal dispersion of dye IV was produced by adding an aqueous solution of 5 x mol dm-3 Na LS, buffered at pH 6.2, to dye IV dissolved in concentrated H,SO,. Mixtures of yellow oxidation products of MV2+ dichloride (named X for convenience) could be obtained reproducibly by stirring for 6 h at 80 OC a solution of 70 mg MV2+ dichloride and 1 g K,CO, in 200 cm3 DMF + H,O (1 / 1, v/v) with admission of air. Ca. 23 of the solvent was then removed with a vacuum rotary evaporator. After neutralization with H,SO, the solution was evaporated to dryness in uacuo and the residues extracted with methanol. The methanolic solution was concentrated to 10cm3, 40cm3 acetone were added and the filtered solution then contained X for further use.X absorbs at 408 and 422 nm in DMF + H20 (2/ 1, v/v) and emits intense yellow light upon irradiation with A,,, < 500 nm. Thin layer chromatography of X on cellulose with n-butyl alcohol +acetic acid + H,O (4/ 1 /2, v / v / v ) ~ ~ showed that X contained at least 14 components. The main component with intense yellow emission when irradiated with A = 366 nm had an rf value of 0.63, which agrees well with rf = 0.64 for the monopyridinone VII :22 0 Resonance Raman spectra were taken using the spinning cell 24 on a Jobin Yvon Ramanor spectrometer, linked to a PDP 11/10 computer and a Nicolet 1074 data handling A Cassegrain light collecting optic was used26 and for solid samples plasma lines were rejected by a focusing-lens-pinhole Excitation at 406.7, 457.9, 472.8, 488.0, 5 14.5, 632.8 and 647.1 nrn was provided by Spectra Physics models 170 Kr+ and 125 HeNe andM.FORSTER A N D R. E. HESTER 1851 TABLE 1 .-ANALYTICAL AND SPECTROSCOPIC DATA FOR DYES I-VI dye yield (%) m.p. /"C Amax/nm, (&/dm3 mol-l cm-l), solvent I - > 300 (> 320C) 670, 615, CHC1,; - 720, 655, 420, NaLSd I1 52 (60a) > 335 (276a) 626, CHC1, 111 93 263 647, CHCl,; 647, NaLSd IV 71 (66b) > 320 (360b) 408, ( 4 . 0 ~ lo4), CHCl,; 387, V 41 (87b) > 330 (> 360b) 401, (4.7 x lo4), DMF VI 72 > 330 385, (3.8 x lo4), H,O (3.5 x lo4), concentrated H,S04 a Ref. (20); ref. (5); ref. (7); [NaLS] = 5 x lo-, mol drn-,. Coherent Radiation model CR4 Ar+ lasers. Emission spectra, corrected for spectrometer sensitivity, were recorded using the same instrumentation.Emission lifetimes were measured using a homebuilt fluoroscopez8 with a mechanically chopped excitation beam. The decay times were calculated from the phase shift between exciting and emitted light.29 1.r. spectra of samples in KBr pellets were taken with a Perkin-Elmer PE-580 i.r. spectrometer. The backgrounds of the digitized spectra were artificially flattened and the peaks were determined numerically after the spectra had been smoothed by a seven point parabolic functi01-1.~~ U.v.-visible spectra were taken on a Unicam SP8000 u.v.-visible spectrometer. For continuous irradiations and quantum-yield determinations solutions in 1 cm U.V. cells had N, bubbled through them for 15 min followed by irradiation with a laser beam expanded to ca. 0.5 cm diameter.The solutions were stirred and held under slight N, overpressure during irradiation. Aqueous solutions of MV2+ dichloride containing DMF or DMSO were prepared by mixing aqueous MV2+ dichloride with DMF + H,O or DMSO + HzO solutions, which had been cooled to 0 "C. The solutions were then used immediately for the irradiation experiments. This procedure prevented the formation of the oxidation products X of MV2+ dichloride. By placing a silica lens in the laser beam before the U.V. cell, the photon flux density could be locally enhanced by a factor of 160, calculated according to ref. (31) for a laser beam with a Gaussian intensity distribution. Laser powers were measured with a calibrated Spectra Physics power meter, model 401 C.RESULTS INFRARED, RESONANCE RAMAN AND VISIBLE ABSORPTION AND EMISSION SPECTRA Table 2 lists the vibrational frequencies of all the dyes I-VI. R.R. spectra of dyes IV-VI could be made in solution. Since dyes 1-111 fluoresce strongly in solution, r.R. spectra from dyes 1-111 could be obtained only as 1% (w/w) mixtures in KBr. For dye I11 the excitation had to be to the blue side of the absorption band in order to escape its high emission. Representative of the other dyes, fig. 3 shows the r.R. spectrum of dye IV in CHCl,, taken with A,,, = 406.7 nm. Intense r.R. bands, which are common to dyes I-VI, are collected in table 3. The assignments were made by comparing table 2 with i.r. data from other squarilium dyes6? 3 2 7 33 and using the fact that the r.R.bands of dye IV shift when going from CHCl, to concentrated H2S0, as solvent. The measured depolarization ratios p of the r.R. bands of dye IV all were ca. i. Dyes 1-111 absorb in the range 61 5-670 nm in CHCl, and dyes IV-VI in the range 385-408 nm. Only dyes I and I11 were weakly soluble in micellar NaLS solution. Dye1852 SQU AR I L IU M DYES FOR SOL AR-ENERGY CONVERSION TABLE 2.-INFRARED AND RESONANCE RAMAN DATA FOR DYES I-vI dye I i.r. v"/cm-l i.r. v"/cm-' 3077 vvw 3060 vw 2982 vw 2965 vw 2930 w 2860 vw 1869 vvw 1825 vvw 1790 vvw 1758 vw 1620 m 1578 m 1564 s 1512 w 1481 vs 1457 m, sh 1448 s 1436 vs 1409 w 1388 w 1367 w - - - 1338 vw 1323 vs 1277 w 1260 vs 1247 vw, sh 1230 w 1189 m 1722 m 1623 s 1569 m 1515 m - - 1441 w 1415 m - 1367 s 1350 vs 1271 w 1245 m 1231 s - 1172 s 11 53 w, sh 1140 w 1112 w, sh 1090 vs 1082 s 993 w 978 vw 882 vw 845 w 820 m 805 vvw 780 m 753 m 740 vvw 702 vw 659 w 599 vw 567 w 546 vw 526 vw 483 vvw 470 vw 439 vw 410 vvw 380 vvw - - - - - - 872 w - 680 w 651 m - - - 541 m 492 w - - - 457 w - 157 w 105 m dye I1 i.r.ij/cm-l i.r. ijlcm-' 2975 vvw 2930 vw 2860 vvw 1700 vvw 1608 s 1588 vs - - - - 1736 s - - 1602 m 1570 vs - 1500 m 1460 w, br 1395 m 1384 vw 1350 vs 1318 vw 1250 m - - - 1440 m 1392 m 1356 w - - - 1202 mM. FORSTER AND R . E. HESTER 1853 TABLE 2.-(cont.) 1.r. v'lcm-l 1.r. v"/cm-l - 1182 m 1160 vs 1128 s - - 1154 m 1062 m - 1036 m - 992 s - - - 948 m 832 m 799 vvw - - 782 s 619 vw 501 w 402 w - - - - - - 752 vs 572 vs 486 s 176 vs 90 vs - - - dye I11 i.r.r.R.C v'/cm-l v"/cm-l 1.r. v"/cm-l 3078 vw 2980 w 2940 w 2900 w 2703 w 2640 vw 2587 vw 2240 vw 2178 w 2090 vw - 1609 vs 1561 vw 1533 w 1493 w 1446 vw 1421 vw 1390 vs 1336 vs 1298 w 1276 m 1246 vs 1213 vs 1187 s, sh 1174 vs 1153 s 1140 w, sh 1121 m - - - - 1730 m 1628 s 1609 m 1559 m 1528 vs 1497 w 1421 vs - 1355 s 1300 w 1285 m 1252 m 1229 m 1218 m 1188 m 1153 w - - - - 1077 s - - 1011 m 960 s 928 vw 862 m 816 m 786 w 746 s 723 vw 674 m 652 vw 555 w 522 vvw 511 vw 500 vw 484 vw 464 vw 452 vw 410 m 334 vw - - - - - 1115 vs 1069 w 1038 vs 1009 w 962 m 862 w 814 w 793 m 747 w 699 vs 673 w 654 vw 584 vs 559 w 521 w 502 vw 482 vw 462 w - - - - - - - 341 m 281 m 261 m 178 vs 158 w, sh -1854 SQU AR I L I U M D YES FOR SOL AR-EN ERGY C O N VERSION TABLE 2.-(cont.) dye IV r.R.d C/cm-' C/cm-' P v'lcrn-' r.R.e i.r.3092 vvw 3062 vw 3038 vvw 3017 vvw 2927 vvw 2858 vvw 1952 vvw - 1954 w - 1908 w 1812 w 1692 w - 0.28 0.40 1790 vvw 1790 w 1704 w 1631 vs 1614 w 1598 m 1570 s 1494 m - 0.39 0.36 0.33 - 1596 s 1522 s - 1497 w, sh 1490 s 1462 s 1492 m 1432 vw - 1418 s, sh 1406 vs 1398 s, sh 1362 w, sh 1337 w 1326 vw 1318 m 1290 vw 1270 w 1214 w 1182 vw 1174 vw 1162 vw 1154 vw 1140 w 1110 vw 1082 w 1073 w 1044 w 1032 w 1023 w 1002 vw 990 vvw 968 vw 923 vw 902 vw 886 w 853 vvw - - 1420 w - 1360 w 1334 w 0.36 1262 w 1198 vs 1178 m - - 1196 s 1184 s - 0.34 0.33 - 996 m - - - 0.40 - - 1006 m - - 0.38 0.39 - - - 928 w 888 m - -M. FORSTER AND R. E. HESTER 1855 TABLE 2.-(cont.) r.R.d i.r. v"1cm-l v"/crn-l P 826 vvw 780 w 778 w 766 m 759 m 718 w 701 w, sh 697 m 690 m - 826 w - 832 m - 762 w 714 s - - - 696 w 650 w - - 694 w 624 vw - - - 554 m 494 m - 0.33 609 vw 574 m-s 533 w 459 w 447 w 413 vw - - - 504 s - - 414 w - 0.36 - - 408 m 374 w 315 vw 244w 88 w - - 0.40 - - 218 w - dye V r.R.f v'lcm-l i.r.v"/cm-l r.R.f v"/cm-l i.r. v"/cm-l 3245 w 3180 m 3130 m 3090 w 3060 w 2998 vw 2970 s 2937 vs 2890 vw 2780 vw 2750 vw 2680 vw 2650 vw 2590 vw 2545 vw 2470 vw 2390 vw 1937 vw 1868 vw 1795 vw 1740 vw 1680 vw 1615 vs 1604 s 1587 vs 1550 vs 1499 s 1448 vs 1421 vs 1409 vs 1337 w 1293 vw 1245 m - - - 1792 w - 1660 m 1602 vs 1584 s 1506 m - - - 1232 vs1856 SQU A RI LI U M DYES FOR SOL AR-ENERGY CONVERSION TABLE 2.-(conz.) i.r. v"/cm-l r. R.f v'/cm-l i.r. v"/cm-] r.R.f v'/cm-I 1183 m 1164 w 1145 w 1079 w 1037 w 1023 w - 998 vw 974 vvw 968 w 899 m 853 m 824 s 794 m 763 m 750 s - 1188 m 1164 w 1094 w - - - - 1000 m - - - - 842 s - - 768 w - - 688 s 621 m 601 m 570 w 510 m 463 s 412 vw 322 w 305 vw - - - - - - 720 w 694 w 656 s 614 w 584 w 556 w - - - - - 408 w 358 w 318 w 264 w - dye VI i.r.v"/crn-l r.R.g v"/crn-l i.r. v"/cm-] r. R.g v"/cm-I 2983 s 2900 w, sh 2844 w 2800 w, sh 2780 s 2510 vw 2422 m 1912 w 1790 w 1615 vs 1590 vs 1547 s 1503 m 1470 m 1460 w 1425 vs 1410 vs 1340 vw - - - - - 1806 w 1654 w 1602 s 1578 m 1502 w-m - - - - - - 1315 w 1316 w - 1298 vw - 1245 w, sh 1246 s - 1192 vs 1192 w - 1154 w 125 vs 1128 m 039 vs - 012 vs 972 vw 950 vw 890 vw 866 w - 3240 vw 3195 w 3122 w 3062 w - 1215 s, sh 3025 vw - - - - - 865 vw, sh - - 852 m - 837 s 810 w, sh 764 w 744 vw 740 m 722 m 696 vs 637 s 612 s 586 vw 569 s 570 w 523 w - - - - 622 m - - -M.FORSTER A N D R. E. HESTER 1857 TABLE 2.-(cont.) 1.r. v"/cm-l r.R.g ?/crn-l i.r. F/cm-' r. R.g v"/cm-l 497 m 448 vw 407 vw 378 w 333 m 298 vw - - - 260 w a 1 % (w/w) in KBr, A,,, = 632.8 nm; 1 % (w/w) in KBr, A,,, = 647.1 nm; 1 % (w/w) in KBr, A,,, = 514.5 nm; 4 x mol dm-, in CHCl,, A,,, = 406.7 nm; 1.1 x mol dm-, in concentrated H,S04, A,,, = 406.7 nm; f 2.2 x lo-* mol dmP3 in DMF, Aexc = 406.7 nm; s = strong, m = medium, w = weak, v = very, sh = shoulder, p = depolarization ratio. 1 x mol dmP3 in H,O, A,,, = 406.7 nm; wavenumber/cm-' FIG. 3.-Resonance Raman spectrum of 4 x mol dm-, dye I Y in CHCl,, Aexc = 406.7 nm. Bands marked with an asterisk are due to CHC1,.IV could be obtained in colloidal form in aqueous NaLS solution at pH ca. 6.0. Solutions of dyes 1-111 in CHC1, fluoresce so strongly that they appear red in room light due to their red emission. Interestingly, the emission from dyes 1-111 in CHC1, could be efficiently quenched by adding solvents with OH groups, e.g. H,O or EtOH (direct observation by eye). Dye IV emits only very weakly in the vicinity of its absorption band, as shown in fig. 3 and 4. Dyes V and VI must behave similarly, since r.R. spectra of these compounds could also be obtained in solution. A high Stokes shift of ca. 6700 cm-l occurs for the emission of dye IV (band maxima taken instead of 0-0 transition^,^). The decay time of the emission at 571 nm was z < 3.6 p s (lower time resolution limit of our fluoroscope).1858 SQU A R ILIUM D YES FOR SOL AR-E NER GY CON VERSION 6 5 U 2 3 2 1 0 4 08 476 57 1 inner: wavenumber/ 1 O 3 cm-' outer: wavelength/nm FIG.4.-Absorption and emission spectrum (i,,, = 406.7 nmj of dye IV in CHCl,: (a) absorption, 1.6 x mol drnP3; (bj emission, 1.0 x mol drnP3, corrected for spectrometer sensitivity. CONTINUOUS IRRADIATION EXPERIMENTS In order to test the ability of dye IV to act in photoreactions as shown in schemes 1 and 2 and fig. 2, N,-purged solutions containing 4 . 4 ~ mol dmP3 dye IV + 5 x mol dm-3 MV2+ dichloride + 1.3 x 1 Ob3 mol dmW3 EDTA +phosphate buffer (pH 6.5) in DMF+H,O (2/1, v/v) were irradiated with light of il = 457.9 nm. Similar solutions in DMSO+H,O (95/5, v/v)* were irradiated with il = 406.7 nm.Using laser powers of 30-140 mW the solutions were turned dark blue within minutes of starting the irradiation. U.v.-visible spectra revealed the formation of MV'+ (methyl viologen radical cation).lg In an extensive first series of experiments with systematic omission of one or more of the constituents of the solutions, and varying the way ia which the solutions were prepared, it was discovered that prolonged heating of the solutions prior to irradiation was mainly responsible for efficient MV'+ formation upon irradiation. In a second series, the formation of oxidation products X of MV2+ dichloride in DMF + H,O or DMSO + H,O upon heating could then be established. A third series of experiments, carried out with ice-cooled solutions, yielded reproducible results which are collected in table 4 (average of 3 or 4 experiments, 15 min irradiation time, laser power 30-140 mW).Direct excitation of MV2+ dichloride with subsequent MV'+ formation is obviously possible, as can be seen from table 4. For experiments with dye IV, the amount of MV+' formed by this direct excitation process has been subtracted according to the method described in the Appendix. * We thank Dr T. Tanno, Riken Co. Ltd, Tokyo, Japan, for suggesting this solvent mixture.TABLE 3.-PROMINENT INFRARED AND RESONANCE RAMAN BANDS (Cm-l) OF DYES I-VI ASSIGNED TO THE CENTRAL PART -0 +o- I I1 I11 IV V v1 i.r.a r.R." i.r." r.R." i.r." r.R.a i.r." r . R . b r.R." i.r." r.R.d i.r.a r.R.e assignment ~ -~ -~ ~~ ~ ~ ~~~ ~ - - - - - 1722 1736 - 1730 1790 1812 1792 1806 v,(C=C) - - - v,,(C-O) - 1569 - 1570 - 1559 - 1570 1522 - 1584 - 1578 v,(C-0) 1620 - 1630 1609 - 1631 - 1615 - 1615 - 54 1 - 572 - 584 - 650 624 -_ 656 - 622 benzene? - - 260 ring modes? - 157 - 176 - 178 - 218 244 264 " In KBr; in CHC1,; in concentrated H,SO,; in DMF; in H,O.v = stretching mode; s = symmetric; as = asymmetric. TABLE 4.-QUANTUM YIELDS, 4, AND NORMALIZED RATES OF MV" FORMATION, F, AND OPTICAL ABSORBANCES, A OF NON-IRRADIATED SOLUTIONS AT A,,,? A,,, = 457.9 nm A,,,. = 406.7 nm DMF+H,O (2/1 v/v) DMSO + H,O (95/5, v/v) HZO with or without EDTA with EDTA without EDTA with EDTA ~~ ~- ~~ ~ _ _ _ ~ ________ -~ with SQ without S without S without S with 9 without S with Sa - _ _ _ ~ ~ _ _ _ _ _ _ 0 3 . 8 ~ 1 0 - 2 f 4 . 5 ~ lo-* 7 .9 ~ 10-4+5.4x 1 . 1 x 10-3+0.5x 1 . 0 ~ 1 0 - ' f 0 . 7 ~ lo-' 4 . 6 ~ 1 0 - 2 + 4 . 0 ~ 1 . 8 ~ 1 0 - 3 f 0 . 4 ~ Fc 0 5.7 x 10-'0& 1.3 x 10-1° 5.5 x I O - ' O + 1.3 x 1.6 x 10-10+0.4 x IO-'O 3 . 4 ~ 10-'0+0.6 x 10 5.8 x IO-'fO.I x lo-' 5.9 x 10-9+0.3 x lo-' A d 2 . 0 ~ 10-3fl.4x10-3 1 . 8 ~ 10-3+2.9x 9 . 2 ~ 1 0 - 2 ~ 5 . 0 x 1 . 9 ~ 1 0 - 2 + 3 . 6 ~ 2.68+ 1 x 1.8 x 10-'f4.2 x 2.70f 1 x lo-, 4b " Contribution from direct MVZ+ excitation subtracted; r.m.s. errors with 95% statistical signifi~ance;~~ mol MV'+ produced/[W(light)s(irradiation time)] in 3.1 cm3 solution with r.m.s. errors; I cm path length, with r.m.s. errors.1860 SQU A R I L I U M DYES F O R SOL A R-EN ER GY CON V ER S I 0 N TABLE 5.-QUANTUM YIELDS 4 FOR MV" FORMATION USING OXIDATION PRODUCTS OF MV2+ DICHLORIDE AS PHOTOSENSITIZERS, IRRADIATED WITH Aexc = 457.9 nma solvent v/v viscosity/cP 4 CH3CN + H,O glycerol + H,O ethanediol+ H,O DMF + H,O DMF+H,O DMF+H,O DMF + H,O H2O < 1.0 1 .o 5.0 6.3 2.3 2.3 2.3 2.3 3.0 x 10-4 3.2 x 10-3 3.7 x 10-3 1.8 x lo-, 4 .7 L 0 . 6 ~ 4.7 x 10-2c 8.0 f 1.2 x 10-3d 3.5 x 10-3e a 0.18 g dmP3 X + 1 x mol dm-3 MV2+ dichloride + 1 x lo-, mol dm-3 EDTA at pH focused laser beam, local photon flux density using compound with rf = 0.63 only as 6.5; laser beam diameter x 5 mm; enhanced by a factor of 160 compared with (b); sensitizer, see text; as (b) but without EDTA. as (a); TABLE 6.-OXIDATION OF DYE I v WITH Ce4+" equivalents IV: oxidantb solvent oxidant IV reduced decomposed (%I (%I oxidant ~ 1: 1.1 1 mol dmP3 H,SO, in H,O 100 3.6 f 1.7 C 1: 1.1 CH3CNe 100 4.1 d 1:20 1 mol dm-3 H,SO, in H,O 55 100 C a E, (Ce4+/Ce3+) = + 1.44 V in 1 mol dm-3 H,SO,; 2 x lo-, mol dmP3 dye IV, max.solubility is 5 x mol dm-3 in 1 mol dmP3 H,S04; Ce(SO,),; Ce(NO,), (NH,),; contains traces of H,O. For solutions containing MV2+ dichloride + EDTA and colloidal dye IV in 5 x 1 0-2 mol dmP3 NaLS or dye VI in H20, no MV' + was formed even after prolonged irradiation with ;I.,,, = 406.7 nm. Table 5 contains the quantum yields 4 for MV'+ formation when the yellow oxidation products of MV2+ dichloride are used instead of dye IV in different solvents. Results for different photon flux densities and for reactions using only the compound with rf = 0.63 instead of the whole mixture X are also presented in table 5.When PtO, was added to similar solutions containing X, MV2+ dichloride and an excess of EDTA at pH 6.5, irradiation with A,,, = 457.9 nm produced only a low steady-state concentration of MV'+ (due to reduction of H,O by MV'+ via Pt0216). No depletion ( f 2%) of X could be observed after 5 h irradiation time, corresponding to ca. 20 h AM2 solar irradiation for the given experimental condi ti on^.^^ OXIDATION OF DYE IV BY Ce4+ In the reactions of fig. 2(a) the intermediate S'+ = IV'+ should occur. In order to detect IV' + in such photoreactions, we tried to generate this species by oxidizing dye IV in different ways, as shown in table 6. When solutions of dye IV were mixed with Ce4+ solutions, the colour changed instantaneously to pale orange and then to blue.M.FORSTER A N D R. E. HESTER 1861 By rapidly scanning the u.v.-visible spectra during the oxidation an intermediate absorption at 475 nm could be established. To our surprise the ca. 1: 1 oxidised solutions were still found to contain dye IV (A, = 408 nm), apparently decomposed only by 3.6% (table 6), although all the Ce4+ was reduced. Furthermore, the oxi- dized solution showed a broad absorption with A, = 570 nm, E = 1.1 x lo5 dm3 mol-1 cm-l (calculated with the assumption that all decomposed dye IV has been transformed into the substance with A, at 570 nm). In 1 mol dm-3 H2S04 in CH,CN+H,O (8/2, v/v), the half-life of L,,, = 570 nm was 2 h. DISCUSSION SOLUBILITY, ABSORPTION A N D EMISSION Squarilium dyes I-V are insoluble in water. For the purpose of solar-energy conversion, water solubility would be a great advantage.We tried, therefore, to dissolve dyes I-V in 0.1 mol dmP3 aqueous micellar NaLS solution, which is known to dissolve aromatic substances. But only dyes I and 111, which contain C2H5 groups, could be dissolved in this medium. Obviously, squarilium dyes with their unusual charged central part do not behave like normal aromatic systems with respect to solubility. Only those dyes containing aliphatic chains, which can immerse in the aliphatic core of a micelle, could be solubilized. The red shift of the absorption of dye I upon micellization indicates strong interactions of dye I with either water itself or with the charged surface of the micelle. Dye VI represents the first known water-soluble squarilium dye.The yellow colour of the aqueous solution, however, faded within some weeks. Obviously dye VI decomposes in H,O, probably via OH- attack on the central 4-membered ring.5 Dyes 1-111 absorb in the red, whereas dyes IV-VI absorb in the blue. This blue shift indicates a much shorter chromophore in the latter cases, which can also be seen from fig. l(a) and (b); in dyes IV-VI the conjugated system extends at most over one R group and the central part, whereas in dyes 1-111 both R groups are included for the different mesomeric structures. The emission spectrum of dye IV given in fig. 4 shows a shoulder at 476 nm and a peak at 571 nm. No mirror image relation between absorption and emission exists. At first sight one might assign the 476 nm band to monomer and the 571 nm band to excimer emission.However, for the low concentrations used here (1 x lov5 mol dm-3) excimer formation is unlikely.37 Another interpretation would be that dye IV behaves as a donor-aromatic-acceptor (DArA) molecule with the N(C,H5), substituents as electron donors and the central ring, which is doubly positively charged, as the aromatic system and as electron acceptor. DArA molecules with an NR, group as electron donor are known to relax from the first excited singlet state S, into a twisted internal charge-transfer (t.i.c. t.) ~ t a t e , ~ ~ - ~ ~ which then emits with a large red shift compared with S,. The formation of such t.i.c.t. states is strongly solvent dependent. Since dye IV is soluble only in a small number of similar solvents, no detailed study was possible and the explanation for the 571 nm emissions has to be taken as a mere hypothesis.1862 SQ U A R I L I U M DYES FOR SOL A R-E NERG Y CONVERS I 0 N INFRARED AND RESONANCE RAMAN SPECTRA The main purpose of the investigation of the vibrational spectra of dyes I-VI was to identify the vibrational modes of the central part I With the substituents R assumed to be point masses, the central part has local symmetry Dzh. Four totally symmetric r.R.and seven i.r. bands should be observable for dyes I-VI. However, since the chromophores extend also over the substituents R (fig. l), r.R. bands of the R groups also are expected to show up. For the r.R. band in the range 1720-2812 cm-l no combination or overtone could be assigned and this is believed to be the fundamental symmetric stretching mode v,(C=C). For rectangular cyclobutadiene (CB), the same mode has been calculated at 178541 or 1744 cm-l 42 and at 1632 cm-l for perdeuterated CB.41 Although the masses of the oxygens and the substituents R in the dyes I-VI are much higher than H or D in CB, a very high v(C=C) can be expected in the former since the central part is doubly positively charged and is presumably highly strained.This assignment, however, supposes a central part with partially single and partially double bonds. When dye IV is dissolved in concentrated H2S04 the 0- are likely to be protonated and v(C-0) is expected to decrease. A frequency shift of -48 cm-l is observed for the band of dye I1 at 1570 cm-l when CHC1, is replaced by concentrated H,S04.This frequency is therefore assigned to the v,(C-0), and the corresponding asymmetric mode is observed in the i.r. at ca. 1609-1630 cm-l. Likewise, the band at 650 cm-l moves to 624 cm-' and the band at 218 cm-l moves to 244 cm-l. Both are r.R. bands which occur in the same frequency region in all dyes I-VI, but their assignment is not yet clear. Since a protonation of the phenyl rings in dye IV is unlikely, these bands also appear to belong to the central part. Common to most i.r. and r.R. spectra, bands around 1600 and 1200 cm-' are believed to originate from phenyl ring vibrations. The depolarization ratios p of all the r.R. bands of dye IV indicate totally symmetric modes. The value p 4 indicates that the excited state of dye IV is n~n-degenerate.~~ CONTINUOUS IRRADIATION EXPERIMENTS SQUARILIUM DYES Iv A N D VI Table 4 shows that MV'+ can be formed via direct excitation of MVz+ even if no EDTA is present (in DMSO + H20).According to the oxidation potential of DMSO (E, 5 1.0 V44) water would be the electron donor [E,(O,/H,O) = 0.82 VJ. However, the exact oxidation potential of DMSO does not seem to be known.45 Furthermore, with H,O as the reducing agent for electronically excited MV2+, 0, would also be formed. If no efficient catalyst for H,O reduction is present,, 0, would oxidize MV*+ back to MV2+, thereby resulting in no net formation of MV'+. No reduction took place either with dye IV as a colloid or with dye VI, both in water. For the former case the excited-state lifetime is probably too short to allow efficient quenching by MV2+.That the latter shows no photoactivity might be due either to unfavourable redox potentials or to the solvent water: for analogous photoredox reactions with the sensitizer X or with phthal~cyanines~~ in mixtures of DMF + H20 or DMSO + H,O,M. FORSTER A N D R . E. HESTER 1863 much higher quantum yields were obtained for MV'+ formation than for reactions in H20 alone. According to table 4 dye IV can act in photoredox reactions as shown in fig. 2. Although the direct excitation of MV2+ exhibits a higher quantum yield q5 for MV'+ formation than by excitation via dye IV, the MV'+ concentrations after irradiation were higher when dye IV was present. For solutions with EDTA the quantum yield for MV'+ formation is 18 times higher than for solutions without EDTA.Obviously EDTA is the electron source in this case. Similar sacrificial systems have been proposed for practical use.47 However, the quantum yield for the MV'+ formation found here is too low for a practical application, although it demonstrates that the excited state of dye IV, with a lifetime z probably much shorter than 3 . 6 ~ ~ (zsl of dye I is 0.28 ns49, is able to reduce MV2+. Since squarilium dyes can be prepared with almost any desired substituents R and any colour, the chances for finding squarilium dyes better suited for solar-energy conversion in solution are high. MIXTURE x Although no complete analysis of all the components of X is available, the presence of mono- and bi-pyridinones in this mixture seems to be obvious: the reaction conditions for formation of X are very similar to those for pyridinone formation22 and the t.1.c.results agree with ref. (22) for the monopyridinone VII. On the one hand, VII does not appear to be the substance with the highest photocatalytic activity for formation of MV' + upon irradiation, as table 5 shows. On the other hand, for practical purposes this is not important since the solutions designed for solar-energy conversion could be produced by a single vessel reaction: MV2+ dichloride is heated in basic DMF + H20, then neutralized and more MV2+ dichloride is added. This solution then could be used for reactions as described by K r a ~ n a . ~ ~ The preliminary stability test of X looks promising and the experiments with variable photon flux densities show that a monophotonic reaction drives the reduction of MV2+.Table 5 indicates a drastic solvent effect on the quantum yield for formation of MV+'. Viscosity alone can not explain the results and the highly dipolar character of DMF+ H 2 0 may also have an influence. The quantum yield 4 = 4.7 x lo-, for the formation of MV*+ with X as photosen- sitizer in DMF + H 2 0 (2/ 1, v/v) compares favourably with Ru(bpy):+ as sensitizer. The latter system, also with EDTA as electron source, had 4 = 2 x in H20.49 X promotes the photoreduction of MV2+ even in the absence of EDTA in DMF + H,O (2/ 1, v/v). Probably (CH,),NH, which is always present in small quantities in DMF,21 acts as the electron source (E,[(CH,),NH'+/(CH,),NH] = 1.03 V us. SCE at pH 1 l.950), or otherwise H,O, with the same uncertainty as outlined in above.Finally, the photoactivity of X indicates that photoredox reactions of MV2+ dichloride carried out in DMF + H 2 0 or DMSO + H 2 0 can lead to spurious results (uia X formation) if no precautions such as cooling and rapid use of the solutions are taken. OXIDATION OF DYE IV WITH Ce4+ The intermediate absorption at 475 nm indicates the occurrence of a short-lived transient in the oxidation of dye IV with Ce4+ We assume this transient to be IV'+. The structure of the compound with A, = 570 nm is not yet clear. R.R. spectra of the dark blue solutions, obtained by oxidation of dye IV by a %fold excess of Ce(SO,), in 1 mol dm-, H2S04 in CH,CN+ H 2 0 (8/2, v/v) yielded intense bands at 1585, 1573,1864 SQU A R I L I U M DYES FOR SOL A R-EN E R G Y C 0 N VE R S I 0 N 1451, 1422, 1233, 1208 and 1180 cm-l and weaker bands at 878, 816, 726, 707,689, 538, 513,431 and 41 1 cm-l.No bands below 300 cm-l, which would indicate radical dimer~,~ -56 where observed. The decomposition of Ce4+ by dye IV seems to be partially catalytic: only 3.6% of dye IV was decomposed in reducing 100% of the Ce4+. With a 19-fold excess of Ce4+, 55% of the Ce4+ was reduced and 100% of dye IV decomposed. Although Ce4+ can oxidize water when catalysts such as RuO, are pre~ent,~' we do not believe that a reaction with no chemical change of the catalyst can explain the results obtained here. With RuO, as the catalyst for 0, evolution, the oxidation of H,O by Ce4+ has a half-reaction time of 30 min.57 By contrast, the reduction of Ce4+ in 1 mol dmP3 H,S04 with dye IV present was complete within seconds.We interpret this as indicating that H,O is oxidized by Ce4+ via the transient species IV'+ to O,, although no 0, has been detected so far due to the low/high solubility of dye IV/O, in 1 mol dm-3 H,S04. Provided that the photoreduction of MV2+ proceeds according to fig. 2(a) and that the back reaction between MV'+ and IV'+ can be inhibited, a complete photolytic water splitting cycle using dye IV, MV2+ and a noble-metal catalyst such as colloidal Pt4 (for water reduction via MV'+) can be envisaged. To our knowledge this would be the first example of such a cycle based solely upon a purely organic dye. As shown in scheme 3 below, dye IV could function as a reducing agent in its excited state IV* and as an oxidant in its oxidized state We+.The expensive RuO, catalyst for the oxidation of H,04 could, therefore, be omitted: scheme 3 hv IV + IV* IV*+MV2+ -+ IV'++MV'+ IV'++$H,O -+ IV+iO,+H+ MV'++H,O -+ MV2++$H,+OH- However, the occurrence of IV'+ in the reactions of scheme 3 has first to be proved. Further experiments are under way. Pt Financial support by the S.R.C. and the Swiss National Foundation (M.F.) are gratefully acknowledged. We also would like to thank Mr R. B. Girling for technical assistance. APPENDIX DETERMINATION OF QUANTUM YIELDS OF MV" FORMATION When solutions of dye IV and MV2+ are irradiated with Aexc = 406.7 or 457.9 nm, two processes lead to MV'+ formation: (i) via direct excitation of MV2+ and (ii) via excitation of dye IV.In order to determine the quantum yield for process (ii), the numbers of photons absorbed by dye IV and MV2+ have to be known, as well as the amount of MV'+ produced by each of the two processes. For the situation here A , < A,, A = ECI (A 1) where A is the absorbance, E is the molar decadic absorption coefficient, c is the concentration, 1 is the cell path length and subscripts M and S denote methyl viologen dichloride and sensitizer IV, respectively. One then obtaips AI,(M) = +Io(l - 10-A~) (A 2) and AI,(M+S) x -I,,(lo-AM+S- 10-A~) (A 3)M. FORSTER AND R. E. HESTER 1865 where I,, is the radiation (einstein cm-2 s-l) incident on the sample cell and AIM, s(M, M + S) are the radiations absorbed by M or S in solutions containing only M or both M and S, respectively.Since the average rate F(M) of MV'+ formation for solutions containing only MV2+ is known (table 4), we obtain the amount of MV'+ in solutions containing both MV2+ and dye IV but produced solely by direct MV2+ excitation where P is the laser power in W and t is the irradiation time in s. This value MVG (M + S) is then subtracted from the observed amount of MV+' in solutions containing M + S. Example for the solutions used, with A,,, = 406.7 nm: A , = 0.019, A , = 2.666 1 2 3 4 5 6 7 R 9 10 11 I:! 13 13 15 16 17 18 1 Y 20 21 22 23 24 25 '26 27 28 29 30 31 32 33 34 35 36 3 i 38 39 M. Kirch, J. M. Lehn and J. P. Sauvage, Helu. Chim. Acta, 1979, 62, 1345 and references therein. V. Balzani, F. Bolletta, M. T. Gandolfi and M. Maestri, Topics in Current Chemistry, 1978,75, 1 and references therein.J. Kiwi and M. Gratzel, J . Am. Chem. Soc., 1979, 101, 7214. K. Kalyanasundaram and M. Gratzel, Angew. Chem., Int. Ed. Engl., 1979, 18, 701. H. E. Sprenger and W. Ziegenbein, Angew. Chem., Int. Ed. Engl., 1968, 7, 530. A. Treibs and K. Jacob, Liebigs Ann. Chem., 1966, 699, 153. H. E. Sprenger and W. Ziegenbein, Angew. Chem., Int. Ed. Engl., 1967, 6, 553. A. Treibs and K. Jacob, Liebigs Ann. Chem., 1968, 712, 123. A. Treibs, K. Jacob and R. Tribollet, Liebigs Ann. Chem., 1970, 741, 101. L. A. Wendling, S . K. Koster, J. E. Murray and R. West, J . Org. chem., 1977, 42, 1126. E. Ehrhardt, S. Hiinig and H. Putter, Chem. Ber., 1977, 110, 2506. S. Hunig and H. Putter, Chem. Ber., 1977, 110, 2524.C. R. Bock, T. J. Meyer and D. G . Whitten, J. Am. Chem. Soc., 1975, 97, 2909. C. Creutz and N. Sutin, Inorg. Chem., 1976, 15, 496. C. T. Lin and N. Sutin, J. Phyx Chem., 1976, 80, 97. K. Kalyanasundaram, J. Kiwi and M. Gratzel, Helc. Chim. Acta, 1978, 61, 2720. A. Moradpour, E. Amouyal, P. Keller and H. Kagan, Now. J. Chim., 1978, 2, 547. P. J. DeLaive, B. P. Sullivan, T. J. Meyer and D. G. Whitten, J . Am. Chem. Soc., 1979, 101, 4007. E. M. Kosower and J. L. Cotter, J . Am. Chem. Soc., 1964, 86, 5524. H. E. Sprenger and W. Ziegenbein, Angew. Chem., Int. Ed. Engl., 1966, 5, 894. A. B. Thomas and E. G . Rochow, J . Am. Chem. Soc., 1957, 79, 1843. A. Calderbank, D. F. Charlton, J. A. Farrington and R. James, J . Chem. Soc., Perkin Trans. 2, 1972, 138. W.Kiefer and H. J. Bernstein, Appl. Spectrosc., 1971, 25, 500. W. Kiefer and H. J. Bernstein, Appl. Spectrosc., 1971, 25, 609. E. Ernstbrunner, R. B. Girling, W. E. L. Grossman and R. E. Hester, J . Chem. Soc., Perkin Trans. 2, 1978, 177. R. E. Hester, J. Raman Spectrosc., 1978, 7, 74. M. Forster and R. E. Hester, J . Chem. Soc., Faraday Trans. 2, 1982, in press. D. S. Smith, unpublished results. W. R. Ware, in Creation and Defection of the Excited State, ed. A. A. Lamola (Marcel Dekker, New York, 1971), vol. 1, part A, p. 213. J. P. Porchet and Hs. H. Gunthard, J . Phys. E., Sci. Instrum., 1970, 3, 261. A. E. Siegman, An Introduction to Lasers and Masers (McGraw-Hill, New York, 1971), chap. 8. J. Gauger and G. Manecke, Chem. Ber., 1970, 103, 2696. J. Gauger and G. Manecke, Chem. Ber., 1970, 103, 3553. R. B. Cundall and A. Gilbert, Photochemistry (Nelson, London, 1970), p. 84. P. H. Muller, P. Neumann and R. Storm, Tafeln der mathematischen Statistik (Fachbuchverlag, Leipzig, 1973), p. 135. Handbook of Geophysics, U S . Air Force (McMillan, New York, 1960), chap. 16-19. J. B. Birks, Photophysics of Aromatic Molecules (Wiley, New York, 1970), p. 301. K . Rotkiewicz, K. H. Grellmann and Z . R. Grabowski, Chem. Phys. Lett., 1973, 19, 315. K. Rotkiewicz, K. H. Grellmann and Z. R. Grabowski, Chem. Phys. Lett., erratum, 1973, 21, 212.1866 SQUARILIUM DYES FOR SOLAR-ENERGY CONVERSION 40 K. Rotkiewicz, Z. R. Grabowski, A. Krowczynski and W. Kuhnle, J. Lumin., 1976, 12/13, 877. 41 L. J. Schaad, B. A. Hess Jr and C. S. Ewig, J. Am. Chem. SOC., 1979, 101, 2281. 4 2 M. J. S. Dewar and A. Komornicki, J. Am. Chem. SOC., 1977, 99, 6174. 43 0. S. Mortensen, Chem. Phys. Lett., 1969, 3, 4. 44 K. Stelmach, Chem. Anal. (Warsaw), 1966, 11, 628. 45 B. Kratochvil, CRC Critical Reviews in Analytical Chemistry (Chemical Rubber Company, Cleveland, 46 T. Tanno, D. Wohrle, M. Kaneko and A. Yamada, Proc. 3rd In&. Con5 Photochem. Conversion Storage 4 7 A. I . Krasna, Photochem. Photobiol., 1979, 29, 267. 48 S. Dahne, F. Fink, E. Klose and K. Teuchner, Laser 77 Opto Electronics Con$ Proc., ed. W. Waidelich, 49 K. Takuma, Y. Shuto and T. Matsuo, Chem. Lett., 1978, 983. 50 M. Masui, H. Say0 and Y. Tsuda, J. Chem. SOC. B, 1968, 973. 51 K. Yokoyama and S. Maeda, Chem. Phys. Lett., 1977,48, 59. 52 S. Yamaguchi, K. Yokoyama and S. Maeda, Bull. Chem. SOC. Jpn, 1978,51, 3193. 53 S. Matsuzaki, T. Mitsuishi, C. Etoh and K. Toyoda, Chem. Lett., 1979, 1417. 54 K. Yokoyama, S. Maeda, C. Etoh, S. Matsuzaki and K. Toyoda, Bull. Chem. SOC. Jpn, 1980, 53, 55 K. Yokoyama, Y. Tajima, M. Tahara and S. Maeda, Bull. Chem. Soc. Jpn, 1980, 53, 2489. 56 E. E. Ernstbrunner, R. B. Girling, W. E. L. Grossman, E. Mayer, K. P. J. Williams and R. E. Hester, 57 K. Kalyanasundaram, 0. MiEiE, E. Pramauro and M. Gratzel, Helv. Chim. Acta, 1979, 62, 2432. Ohio, 1971), p. 415. Solar Energy (SERI, Golden, Colorado, 1981), p. 161. p. 151. 36. J. Raman Spectrosc., 1981, 11, 161. (PAPER 1/1151)

ISSN:0300-9599    

DOI:10.1039/F19827801847

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1982, 78, 1867-1872 Hydrogen Bonding Studies with a Stable Free Radical BY ADURTHY S. N. MURTHY* AND ANNADI RAM REDDY Department of Chemistry, Indian Institute of Technology, New Delhi 110 016, India Received 21st July, 1981 The hydrogen bonding of 1,l -diphenyl-2-picrylhydrazyl (DPPH) with chloroform, bromoform, di- chloromethane and 1,2-dichloroethane has been examined in carbon tetrachloride by electronic absorption spectroscopy. The equilibrium constants and enthalpies (AH*) of hydrogen-bond formation vary in the order chloroform > bromoform, dichloromethane > 1,2-dichloroethane. Numerous studies have been reported on the electron donor-acceptor complexes consisting of molecules with closed-shell ground states.'* However, little is known about the interaction of those having unpaired spins in the electronic ground state.The only literature report that has come to our attention concerns the donor properties of the di-t-butyl-N-oxide (DTBNO) free radical. Murata and Mataga3 have studied the interaction of DTBNO with a few electron acceptors by electronic absorption spectroscopy and have detected new transitions due to the formation of either a radical anion or a charge-transfer complex, depending on the nature of the acceptor. With phenol as an acceptor (proton donor), however, the n --, n* transition of DTBNO in hexane undergoes a blue shift giving rise to isosbestic points, thereby indicating that there is an equilibrium between the free and hydrogen-bonded DTBNO. The equilibrium constant for a 1 : 1 hydrogen-bonded complex has been determined to be 6.4 dm3 mol-l in toluene solvent.The concentrations of phenol employed by Murata and Mataga3 were rather large. Thus, the self-association of phenol and non-conformity to Person's4 criterion make the value of the equilibrium constant less reliable. l,l-Diphenyl-2-picrylhydrazyl (DPPH) is a stable free radical. The interaction of DPPH with strong proton donors such as alcohols, phenols and thiols has been investigated5 recently and it has been shown that DPPH abstracts hydrogen from these proton donors to form radicals. Singh et a1.6 have examined the kinetics and energetics of some of these hydrogen-abstraction reactions in detail by electronic absorption and e.s.r. spectroscopy and have proposed a model in which the initial hydrogen-bond formation between DPPH and hydroxylic compounds plays a major role.The nature of the hydrogen bond formed between the hydroxylic compounds and DPPH could not, however, be examined, as the hydrogen-bonded species immediately undergoes further reaction. It would be of interest to obtain experimental evidence in favour of such initial hydrogen-bond formation. We considered that it might be possible to obtain such evidence using weaker proton donors which do not take part in hydrogen-abstraction reactions. Therefore, we have chosen to investigate the hydrogen- bonding interaction of DPPH with chloroform, bromoform, dichloromethane and I ,2-dichloroethane in an inert solvent, carbon tetrachloride, by electronic absorption spectroscopy.The proton donors selected for this purpose undergo negligible self-association at the concentrations employed. 18671868 H BONDING WITH A STABLE FREE RADICAL EXPERIMENTAL MATERIALS DPPH was obtained from E-Merck and used without further purification. The proton donors were purified by standard methods. Care was taken to ensure that they did not contain traces of alcohols as impurities. SPECTRA Electronic absorption spectra were recorded with a Pye-Unicam SP 700 spectrophotometer. Measurements in the temperature range 30-50°C were carried out using a Pye-Unicam SP 770 constant-temperature cell-holder and SP 775 electrical controller. The temperature was kept constant within LO.5 O C . The stock solutions were freshly prepared on the day of the measurements.Matched stoppered quartz cells of 1 cm path length were used. The equilibrium constants were obtained by using the Baba-Suzuki' relationship: where E, is the molar absorption coefficient of DPPH, cb is that of the complexed DPPH and is that of a solution in which the initial concentration of proton donor is B mol dm-3. A plot of ~ ( E ~ - E , ) against 1/B gives a straight line from which K can be calculated. The value of AH* was calculated by measuring K at different temperatures. In all the systems studied the concentration of proton donor is varied, keeping the concentration of DPPH fixed. This is valid since all the proton donors undergo negligible self-association in the concentration range employed. RESULTS AND DISCUSSION The reliability of the electronic absorption spectral method for the study of hydrogen-bonding equilibria with a weak proton donor like chloroform was first checked by studying the latter's hydrogen-bonding interaction with acetone in carbon tetrachloride solution and comparing the equilibrium constant obtained with that obtained by other methods.The spectra showed isosbestic points, and the equilibrium constant for 1 : 1 complex formation was determined by the Baba-Suzuki' method. The equilibrium constant at 30 O C is 1.04 dm3 mol-l, which is in good agreement with the value 1.2 dm3 mol-1 (30 O C ) obtained by Whetsel and Kagarise* using infrared spectroscopy. Murthy et aL9 have in fact demonstrated that reliable thermodynamic data for several hydrogen-bonded systems can be obtained by electronic absorption spectroscopy . The electronic absorption spectrum of DPPH shows two high-intensity bands at ca.330 and 520 nm, respectively. The effect of the solvents listed in table 1 on the 520 nm band were first examined. The position of the band (V) is plotted in fig. 1 against [(E - 1 ) / ( ~ + 2) - (n2 - l)/(n2 + 2)], where E and n are the static dielectric constant and refractive index of the solvent, respectively. Such a plot is expected to be linear in the absence of molecular complex formation, since the solvent shift is a function of E and n only. This is in accordance with McRae's theory of solvent effects.1° Thus in hexane, benzene, toluene and carbon tetrachloride, where there is no specific molecular interaction with DPPH, the two quantities are well-correlated.For the hydrogen- bond-forming solvents, such as chloroform, 1,2-dichloroethane, methanol and propan- 2-01, the values deviate from those of the first four solvents, but they have their own distinct correlation. Acetonitrile, which cannot form a hydrogen bond with DPPH, falls in line with these solvents, possibly owing to a non-specific interaction resulting from its high dielectric constant.A. S. N. M U R T H Y A N D A. R. R E D D Y 1869 TABLE SOLVENT SHIFTS OF THE 520 nm BAND MAXIMA OF DPPH E - 1 n2-lU ___- solvent no. solvent vm,,/ 1 O3 cm-' ~ + 2 n2+2 1 2 3 4 5 6 7 8 9 10 11 12 hexane benzene carbon tetrachloride toluene dioxan chloroform di-isopropylether 1,2-dichloroethane propan-2-01 acetone methanol acetoni trile 19.57 19.38 19.34 19.31 19.42 19.16 19.42 19.16 19.23 19.42 19.23 19.20 0.000 0.005 0.015 0.023 0.050 0.294 0.370 0.480 0.620 0.620 0.710 0.710 The refractive index and dielectric constant values were taken from Handbook of Physics and Chemistry, ed.R. Weast (C.R.C. Press, Cleveland, Ohio, 59th edn, 1978-79). o7 0 6 0 8 10 0 09 O1 01* , 1 .-Typical plot of the solvent shifts of the 520 nm band maxima of DPPH in (1) hexane, (2) benzene, (3) carbon tetrachloride, (4) toluene, ( 5 ) dioxan, (6) chloroform, (7) di-isopropylether, (8) 1,2-dichIoroethane, (9) propan-2-01, (10) acetone, (1 1) methanol and (12) acetonitrile. Since alcohols take part in hydrogen-abstraction reactions with DPPH, our studies have been confined to hydrogen-bonding interaction of DPPH with chloroform, bromoform, dichloromethane and 1,2-dichloroethane.The two bands at 330 and 520 nm in the electronic absorption spectrum ofDPPH (2.23 x lop5 mol dmp3) undergo a decrease in intensity on the addition of increasing amounts (1 .O-5.0 mol dm-3) of chloroform (fig. 2). The spectra also show an isosbestic point at ca. 519 nm, thereby1870 H BONDING WITH A STABLE FREE RADICAL I 1 wavelengt h/n m FIG. 2.-Absorption spectra of the DPPH +chloroform system in carbon tetrachloride: ( 1 ) DPPH (2.32 x mol dm-3), (2)-(5) DPPH (2.32 x mol dm-3)+chloroform (0.52, 1.45, 2.05 and 2.62 mol drnp3, respectively). TABLE 2.-THERMODYNAMIC DATA FOR HYDROGEN BONDING BETWEEN DPPH AND PROTON DONORS IN CARBON TETRACHLORIDE solvent K (30 "C) A H 9 no. proton donor /dm3 mol-l /kJ mol-l 1 chloroform 0.144 -11.5 2 bromoform 0.085 - 7.5 3 dichloromethane 0.08 - 10.9 4 1,2-dichloroethane 0.055 - 9.0 indicating that there is an equilibrium between free and hydrogen-bonded DPPH.Similar spectra were obtained in case of bromofonn, dichloromethane and 1,2-dichloroethane. The variation in intensity of the 520 nm band of DPPH with concentration of proton donors has been used to calculate the equilibrium constant and enthalpy (AH*) of hydrogen-bond formation. The data are shown in table 2. A typical plot for the calculation of the equilibrium constant for the systems chloroform + DPPH and 1,2-dichloroethane + DPPH in carbon tetrachloride is shown in fig. 3.A. S. N. MURTHY AND A. R. REDDY 1871 2 I I I I I ( 1/B>/dm3 mol-' 1 0.2 0.6 0.6 0.8 1.0 FIG.3.-Typical plot for the calculation of equilibrium constant for (1) the DPPH + 1,2-dichloroethane system and (2) the DPPH +chloroform system in carbon tetrachloride by the Baba-Suzuki' method. As can be seen from table 2, the equilibrium constants at 30 OC are in the range 0.055-0.144 dm3 mol-1 and vary in the order chloroform > bromoform > dichloro- methane > 1,2-dichloroethane. In case of the halogenoforms the order is as expected, since chloroform is a better proton donor than bromoform. The difference in equilibrium constants between dichloromethane and 1,2-dichloroethane is not very marked but the small difference is as expected: dichloromethane is a better proton donor than 1,2-dichloroethane. The AH* values for the interaction of DPPH with these proton donors vary in the range 7.5-1 1.5 kJ mol-1 (table 2).The variation in AH* is in the same direction as the equilibrium constants. In the model proposed by Singh et a1.6 it was considered that RXH (X = 0, N or S) initially forms a hydrogen-bonded complex X-H * * N with DPPH, eventually forming an activated complex similar to a symmetrical hydrogen bond (X - - - H - - N). This implies that only strong hydrogen-bond formation (large enthalpies) favours hydrogen transfer. Since the enthalpies for the interaction of DPPH with proton donors reported in this investigation are low, it is clear that hydrogen abstraction from these proton donors is not favourable. A.R. is grateful to the authorities of the Indian Institute of Technology, Delhi, for the award of a research fellowship. A. S. N . Murthy and C. N . R. Rao, Appl. Spectrosc. Rev., 1968, 2, 69. M. D. Joesten and L. J. Schaad, The Hydrogen Bond (Marcel Dekker, New York, 1974). Y. Murata and N. Mataga, Bull. Chem. SOC. Jpn, 1971, 44, 354. W. B. Person, J. Am. Chem. SOC., 1965, 87, 167. A. R. Forrester, J. M. Hay and R. H. Thomson, Organic Chemistry of Stable Free Radicals (Academic Press, New York, 1968), chap. IV, pp. 137-180.1872 H BONDING WITH A STABLE FREE RADICAL S. Singh, K. R. Bhaskar and C. N. R. Rao, Can. J . Chem., 1966,44, 2657. K. B. Whetsel and R. E. Kagarise, Spectrochim. Acta, 1962, 18, 329. A. S. N. Murthy, S. Singh and C. N. R. Rao, Trans. Faraday Soc., 1966, 62, 1056. ’ H. Baba and S. Suzuki, J. Chem. Phys., 1961, 35, 1 1 18. lo E. G. McRae, J . Phys. Chem., 1957, 61, 562. (PAPER 1/1161)

ISSN:0300-9599    

DOI:10.1039/F19827801867

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. SOC., Furuduy Trans. 1, 1982, 78, 1873-1885 Influence of Electrostatic Forces upon the Efficiency of Charge Separation for the Zinc Porphyrin/Methyl Viologen System BY MARIE-CLAUDE RICHOUX* AND ANTHONY HARRIMAN Davy Faraday Research Laboratory of The Royal Institution, 2 1 Albemarle Street, London W 1 X 4BS Received 31st July, 1981 The bimolecular rate constant for quenching the triplet excited state of a metalloporphyrin by methyl viologen depends upon the electronic charge on the porphyrin ring. The rate constant increases with increased electrostatic attraction between the reactants and this is due to efficient formation of an encounter complex and a high thermodynamic driving force for formation of an ion-pair. However, dissociation of the ion-pair to separate ion products is favoured by strong electrostatic repulsion between the products due to electrostatic factors.In addition the yield of redox products depends upon the amount of electronic energy that must be dissipated in the form of heat, and the rate of internal conversion of the ion-pair decreases significantly with increased energy. The photochemistry of porphyrins has been studied intensively during the last decade, partly because of the possibility of using such compounds as photosensitisers for transformation of solar energy into a storable form of energy. In particular, the photosensitised dissociation of water into H, and 0, has attracted considerable attention, and previous work has established that zinc porphyrins can function as efficient sensiti~ers~-~ for the photochemical production of hydrogen in aqueous solution.Indeed, three separate laboratories have found that positively charged water-soluble zinc porphyrins are particularly effective photosensitisers when methyl viologen (MV2+) is used as electron re la^,^-^ whilst negatively charged compounds are very much less prorni~ing.~3 The overall reaction mechanism for the former system involved electron transfer from the triplet excited state of the Zn porphyrin (*ZnTMPyP4+) to MV2+, followed by the reduction of the oxidised form of the porphyrin by an irreversible redox couple such as EDTA: *ZnP + MV2+ + ZnP+ + MV+ ZnP+ + EDTA -+ ZnP + EDTA+ EDTA+ --.) products MV+ + H,O + MV2+ + $H2 +OH-. The high quantum yield for H, production obtained with ZnTMPyP4+ (41H2 = 0.60) was attributed to the strong electrostatic repulsion between the primary redox products (ZnP+ and MV+), which favoured dissociation of the ion-pair into redox ion products.Similar proposals have been made for other system^,^^ lo but there has not been a systematic study made to confirm or refute this hypothesis. In this paper we report the results of a study aimed at evaluating the importance of electrostatic factors in determining &tHI for the zinc porphyrin sensitised reduction of H,O to H, 61 1873 F A R 11874 QUENCHING OF Zn PORPHYRIN B Y METHYL VIOLOGEN using EDTA as sacrificial donor, and in so doing we attempt to optimise the reaction for production of H,. Unfortunately, the study was complicated by some synthetic problems, since symmetrically substituted porphyrins with four identical groups (either N-methyl, carboxylic or sulphoxylic) are prepared readily, but it is very difficult to separate and purify the compounds having only two or three charges.Consequently, we have been able to study only zinc porphyrins possessing four identical charges or a single positive or negative charge. Abbreviations for the various porphyrins mentioned are shown below. ABBREVIATIONS ZnTSPP4- ZnT M PyP4' ZnT C PP 4- ZnCPP- Zn M Py P' Rl=R2=R3=R4= SOY Rl=R,= R,= R,= a+- CH, R,=R2 = R 3 =R,= *COO- Rl = *coo- EXPERIMENTAL EDTA (B.D.H.) and methyl viologen (Sigma) were used as received. ZnTCPP4- and ZnTSPP4- were prepared from the metal-free compounds and a large excess of Zn acetate by stirring in aqueous solution at 343 K overnight. The excess of Zn acetate was removed on an ion-exchange column.ZnTMPyP4+ was prepared from meso-tetra(6pyridyl)porphine (TPyP) as described b e f ~ r e . ~ ZnCPP- was prepared as indicated by Anton and Loachll by mixed condensation. 4-Carboxybenzaldehyde (1 .O g) and 4-tolualdehyde (4.0 g) were heated in propionic acid (170 cm3) until the solvent began to reflux. Pyrrole (2.7 g) was added and the mixture refluxed for 30 min. After cooling overnight the solid material was filtered and washedM-C. RICHOUX AND A. H A R R I M A N 1875 with methanol and hot water and chromatographed on alumina. After elution of TPP the required compound passed down the column and was collected. The solvent was removed under reduced pressure and the solid dried at 373 K under vacuum.Samples of the above product were analysed by electrophoresis in aqueous medium (pH 7.0, NaCl 1 mol dm-3), and it was found that the isolated product was a mixture of which the predominant component possessed a single negative charge but that there was also a significant quantity (ca. 20%) of the dicarboxy compound. Zinc was inserted as for ZnTCPP4-. ZnMPyP+ was prepared by the method of Little et a1.12 Pyridyl tritolylporphyrin was prepared by refluxing for 1 h 1.5 g of pyridinecarboxyaldehyde with 5.06 g of 4-tolualdehyde in 187.5 cm3 of propionic acid containing pyrrole (3.78 g). After cooling overnight the solid was filtered and washed with methanol and left to dry at 100 "C for 1 day. This compound was dissolved in chloroform and chromatographed on alumina to remove TTP.The residue was redissolved in chloroform and refluxed with a saturated solution of zinc acetate in methanol until the absorption band situated at 650 nm disappeared (ca. 2 h). After cooling, the solvent was removed under reduced pressure and the residue stirred with water (60 cm3) at 373 K 1 h. Undissolved solid was removed and washed with warm water (100 cm3). The solid was dried for 3 days at ca. 373 K, then rechromatographed on alumina with CHC1, as eluent. Three fractions were collected. The first fraction contained the required product, the second fraction was a zinc chlorin and the third fraction contained an unidentified product. After evaporation of the solvent methylation was achieved by stirring the compound with a large excess of methyl iodide at room temperature for 3 days.The final product was isolated by filtration and dried under vacuum. It was sparingly soluble in water but dissolved in ethanol and this solvent was used for all studies with this compound. Solutions for steady-state irradiation experiments contained a zinc porphyrin at a concen- tration such that the optical density at the irradiation wavelength (1 = 565 & 5 nm) was 0.5 for a 1 cm pathlength. The solutions were prepared from triply distilled water, buffered to pH 6.9 with 0.01 mol dm-3 phosphate, and contained MV2* (10 mmol dm-3) and EDTA (10 mmol dmP3). Irradiations were performed with an Applied Photophysics model U.V.30 quantum-yield reactor equipped with a high-radiance monochromator.The solutions were purged thoroughly with N, prior to irradiation and the light intensity was calibrated with a standard thermopile. The concentration of reduced MV+ was monitored by absorption spectroscopy and the concentration of evolved H, was measured by gas chromatography. Absorption spectra were recorded with a Perkin Elmer model Hitachi 200 spectrophotometer and all luminescence measurements were made with a Perkin Elmer MPF4 spectrofluorimeter. Fluorescence quantum yields were measured by the optically dilute method, as described previously, using ZnTPP as standard. Phosphorescence spectra and lifetimes were recorded in frozen glasses (glycerol/water, 1/3) at 77 K, as described in a previous paper.13 Room- temperature flash-photolysis studies were made with conventional microsecond equipment using KCrO, filter to remove excitation light of A < 470 nm.Triplet quenching rate constants were determined by measuring the triplet lifetime of the Zn porphyrin in the presence of a known concentration of quencher; at least four different concentrations were used for each study and the observed rate constants had a reproducibility of f 20%. All solutions for luminescence and flash-photolysis studies were outgassed by the freeze-pumpthaw method. RESULTS AND DISCUSSION SPECTROSCOPIC PROPERTIES As reported previously5 the zinc porphyrins are good candidates for use in solar-energy storage devices since they are capable of harvesting a reasonable fraction of the solar spectrum (ca. 30%). All the zinc porphyrins studied here absorb strongly in the visible region between 500 and 650 nm with two main Q bands, one at ca.560 nm and the other at ca. 600 nm as shown in fig. 1. The Soret band is found be- tween 420 and 430 nm. Although the absorption spectra of ZnTMPyP4+, ZnCPP-, ZnTCPP4- and ZnTSPP4- are similar we notice a shift of the absorption peaks to 61-21876 QUENCHING OF Zn PORPHYRIN BY METHYL VIOLOGEN the red as the porphyrin becomes more positively charged. Because of the low solu- bility of the compound in water, we were obliged to study ZnMPyP+ in ethanol solu- tion instead of water, but it is well-known that n -+ n* transitions are shifted to the red when polar solvents are used, so that we would expect an absorption shift of ca. 4 nm to the red if the compound were dissolved in water.The fluorescence emission maxima of the five zinc porphyrins depend also on the nature of the water solubilising groups, and the spectra are represented in fig. 2. From these fluorescence maxima we have calculated that the singlet energies are 195 k 2 kJ mol-I. The phosphorescence emission measured at 77 K did not seem to be affected by the nature of the water-solubilising group in such a consistent manner. However, the accuracy of the phosphorescence emission measurements is not so high as for the fluorescence measurements, due to the need to use wide bandwidths so that minor variations in the emission maxima may be lost. However, the triplet energies have been calculated to be 156k2 kJ mol-l. Therefore, the type of substituent used to induce solubility in water appears to have very little effect upon the energy of the excited states of Zn porphyrins.QUENCHING Previously, we reported that the lifetime of the excited singlet state of ZnTMPyP4+ determined in outgassed aqueous solution was very short (zs = 1.4 ns), so that intermolecular reaction involving the excited singlet manifold was unlikely except at very high concentration of q ~ e n c h e r . ~ As shown in table 1 the other water-soluble zinc porphyrins also have very short singlet lifetimes, and concentrations of MV2+ and EDTA as high as 0.05 mol dm-3 exhibited no quenching effect upon the observed lifetime of lowest-energy excited singlet state of the metalloporphyrins. With the negatively charged porphyrins, some static quenching occurred at high concentration of MV2+ ( 2 5 x mmol dm-3) so that, under such conditions, it was possible to observe a decrease of the fluorescence intensity of the porphyrins.This static quenching most probably reflects some degree of ion-pair formation as described for reaction between negatively charged zinc, phthalocyanines and MV2+.’9 In contrast to the singlet excited states, the triplet excited states of all the zinc porphyrins were long lived and the triplet lifetimes (z,) are collected in table 1. Addition of MV2+ resulted in a decrease in the triplet lifetimes as monitored by conventional flash photolysis, and the observed bimolecular quench- ing rate constants [k(MV2+)] are given in table 1. The magnitude of these quenching rate constants is dependent upon the nature of the water-solubilising groups; there is a thousand-fold increase in k(MV2+) upon changing from a positively charged periphery group to a negatively charged one.Since the general properties of the triplet excited state of the zinc porphyrins are not affected by the type of water-solubilising groups used, nor is there any spectral evidence to suggest static quenching at the low concentrations of MV2+ used for these studies, the differences observed for k(MV2+) must be attributed to differences in the efficiency of the quenching mechanism. This quenching mechanism is believed to be predomi- nantly electron transfer. The scheme given below gives a simple representation of intermolecular electron transfer from the triplet excited state of a zinc porphyrin (*P) to MV2+.The scheme is a general one and involves formation of an encounter complex within which there is virtually no binding energy between the reactants. The formation of such an encounter complex will be a diffusion-controlled process ( k , , F F ) and, since the complex will be a very weak one, it will be reversible (kDIss). Once formed, the complexM-C. RICHOUX AND A. HARRIMAN 1877 500 550 600 650 Xinm FIG. 1 .-Absorption spectra: (-) ZnTMPyP4+, (----) ZnTSPP4-, (0 0 0 0) ZnTCPP4-, ( x x x x ) ZnCPP-, ( . . . . ) ZnMPyP+. 550 600 6 50 700 750 FIG. 2.-Fluorescence spectra: (--) ZnTMPyP4+, (----) ZnTSPP“, (00 0 0) ZnTCPP4-, ( x x x x ) ZnCPP-, ( . . . . ) ZnMPyP+. X/nm may undergo an electron-transfer process to form an ion-pair (kA) wherein the redox products are caged by the solvent. The ion-pair may decay either to reform the ground- state reactants (kR) or to form separated ion products (kc).An important feature of this scheme is that formation of the ion-pair ( k A ) is considered as an irreversible process. It could be argued that this step should be reversible, but in order to evaluate1878 QUENCHING OF Zn PORPHYRIN BY METHYL VIOLOGEN TABLE 1 .-PHOTOPHYSICAL PROPERTIES OF SOME ZINC PORPHYRINS AND THE BIMOLECULAR RATE CALCULATED FOR ZERO IONIC STRENGTH [k(MV2+)0] CONSTANTS FOR QUENCHING THE TRIPLET EXCITED STATE WITH MV2+ ; OBSERVED [k(MV2+)] AND Es ET ~T(300 K) r~(77 K) tS(300 K) k(MV2+) k(MV2+)0 compound /kJ mol-l /kJ mol-l / p s /ms /ns /dm3 mol-I s-' /dm3 mol-l s-l ZnTMPyP4+ 190 157 655 3.3 1.4 1.80 x 10' 2.7 x 105 2 .1 4 ~ 107 2.1 x 107 ZnMPyP+ a 197 154 570 2.5 - 1.04 x 109 3.0 x 109 ZnCPP- 196 158 1470 - ZnTCPP4- 197 154 1300 1.4 1.6 1.3 x 10lob 2.7 x 1Olo ZnTSPP4- 198 156 1500 2.5 1.7 1.4 x 2.9 x 1O1O - a In ethanol solution; ref. (4). the individual rate constants involved in the overall scheme we have assumed that kA is very much greater than the reverse step. k m , , P*+MV2+ e (P*,MV2+) kmss 1 k A A B k b i (P+, MV+) P + MV2+ P+ + MV+ SCHEME The rate of formation of the encounter complex (kDIPF) can be calculated from the Debye expression for diffusional encounter between two ions of charges zA and zB14 kDIFF = 4nr.4€3DABNA6/(exp6- l) 6 = ZA Z g e2/&k TrAB where and rAB = 1.0 nm, DAB = Now, as we have shown previously,15 the standard free-energy change (AG*) associated with the formation of a very weakly bound complex between charged ions can be calculated from the expression cm2 s-l.NA is Avogadro's constant. where E is the permittivity of the solvent. This free-energy change can be used to calculate the equilibrium constant for formation of the encounter complex ( K ) = kDIFF/kDISS so that knIss can be obtained (table 2). However, note that the reactant molecules are not spheres, as assumed in the above model, and specific solvation terms are far from negligible. In fact, we might expect to find significant differences in the overall structure of the encounter complex (and the ion-pair) as the local charges on the porphyrin periphery are changed, and consequently all rate constants that have been derived from this treatment are subject to considerable error.Hopefully, the derived values will still reflect the overall trends inherent in the 'true system'. With the exception of ZnMPyP+, which was studied in ethanol solutions, there isM-C. RICHOUX A N D A. HARRIMAN 1879 TABLE 2.-RATE CONSTANTS DESCRIBING FORMATION AND DECAY OF THE ENCOUNTER COMPLEX kDIFF compound /dm3 Inol-' s-l kDISS/S-l P k,/s-' ZnTMPyP4+ 1.3 x lo8 2.95 x 1 O 1 O 2.08 x 6.1 x 107 ZnCPP- 1.0 x 1010 2.06 x 109 - - ZnTSPP4 - 3 . 0 ~ 10l0 1 . 3 ~ Ion 0.97 4.2x 109 ZnMPyPfn 2.6 x 109 2.08 x 10" 3.1 x 10-3 1.7 x lo8 ZnTCPP4- 3.0 x 1 O 1 O 1.3 x loR 0.90 11.7 x lo8 a In ethanol. a steady decrease in k,,,, as the degree of electrostatic attraction between the reactants is increased. Thus, kDIss for ZnTMPyP4+ is some hundred times greater than that found for ZnTSPP4- or ZnTCPP4-.Now, the bimolecular triplet quenching rate constants [k(MV2+)] can be expressed in the form k(MV2+) = kDIFFkA kA + k D I S S so that k , can be evaluated. However, the measured k(MV2+) values are strongly dependent upon the ionic strength ( p ) of the medium, and the calculated k,IFF and k,,,, values refer to zero ionic strength. Consequently, the measured k(MV2+) values have been corrected to zero ionic strength using the relationship log k(MV2+) = logk(MV2+)0 + I .02 ZA zg d p and the calculated values k(MV2+)0 are given in table 1. The use of these values, together with the calculated k , , F F and k,,,, values, allows calculation of k,, and the derived values are given in table 2.The derived k , values refer to the rate of electron transfer from the triplet excited state of the zinc porphyrins to MV2+. The probability of electron transfer (P) can be expressed in the form- kA - k(MV2+)" P = - k , + ~ D I S S k,IF, and it can be seen from table 2 that the nature of the water-solubilising groups has a marked effect upon P. In fact, P varies from 0.2% for ZnTMPyP4+ to 90% for ZnTSPP4-, and inspection of the data provided in table 2 shows that this variation is due to the differences in kDIss. The derived k, values, although different for each metalloporphyrin, fall within a fairly narrow range, but the kDIss values vary over a much wider range and this variation can be attributed to electrostatic factors. Where there is strong electrostatic attraction between the reactants, k,,,, is low and the probability of quenching is high.Thus the electrostatic factors that determine the stability of the encounter complex are of extreme importance in controlling the overall triplet quenching efficiency. Since k, refers to the rate of electron transfer, there should be a correlation between the derived k , values and the overall standard free-energy change associated with the electron-transfer step (AGg-). This latter term can be calculated from the relevant redox potentials MV2+ +e- + MV+ E*(MV2+/MV+) P+ +e- + P* Ee(P+/P*)1880 QUENCHING OF Zn PORPHYRIN BY METHYL VIOLOGEN and the redox potential of the triplet excited state of the zinc porphyrin [E*(P+/P*)] can be obtained by subtracting the triplet energy from the ground-state redox potential E*(P+/P*) = E*(P+/P)- ET.Ground-state redox potentials have been measured for some of the zinc porphyrins used in this study, but data are not available for the singly charged compounds. In general, positively charged substituents have little effect upon the redox potentials but negatively charged substituents lower E*(P+/P) by ca. 60 mV per substituent. Using these approximations, we have estimated E*(P+/P*) values for ZnCPP- and ZnMPyP+ (table 3). Using the above redox potentials, we have calculated AG& for the process P* + MV2+ + P+ + MV+ and we have included a correction to account for any changes in electrostatic forces associated with the electron-transfer step. The data (table 3) are most interesting in that they show that as the overall standard free-energy changes become more negative, the rate of electron transfer becomes faster, which infers that the high probability for electron-transfer quenching (P) observed with the highly negatively charged zinc porphyrins is due to the high thermodynamic driving force available for the reaction.In contrast, with ZnTMPyP4+ the overall AGf& is positive, so that electron transfer should be prohibited on thermodynamic grounds. That electron transfer does take place for this compound, albeit at very slow rate, shows the limitation of such calculations. TABLE 3.-REDOX POTENTIALS FOR THE TRIPLET STATES, STANDARD FREE ENERGY CHANGES FOR FORMATION OF THE ION-PAIR AND CALCULATED RATE CONSTANTS FOR IONIC SEPARATION AND RECOMBINATION OF THE ION-PAIR compound E*(P+/P*)/eV AG@,/eV k-,/dm3 mol-' s-l kc/s-l ZnTMPyP4+ - 0.44 0.12 6.4 x los 1.9 x 1Olo ZnMPyP+ -0.52 - 0.04 2 .6 ~ 109 1.0 x 1O'O ZnCPP- - 0.62 -0.17 5.5 x 109 5.5 x 109 ZnTSPP4- - 0.75 -0.35 1.3 x 1Olo 1.7 x 107 ZnTCPP4- - 0.82 - 0.42 1.3 x 1Olo 1.7 x 107 The ion-pair formed by the electron-transfer process (kA) can dissociate to ground-state reactants (kB), which may be a spin-forbidden step, or to separated ion products (k,), and the partition between these two decay routes controls the yield of redox products. Once formed, the redox products will recombine via a diffusion- controlled reverse electron-transfer step k - , . k-c k , P+ + MV+ + (P+, MV+) + (P, MV2+). kc The procedure outlined above for calculation of k D I F F and kDIss can be used to estimate k - , and k , since k-, is diffusion-controlled, and the equilibrium constant ( K ) for formation of an ion-pair from the separated ion products can be expressed in the form K'= k-,/k,.The derived values are collected in table 3, and again it can be seen that the natureM-C. R I C H O U X A N D A. HARRIMAN 1881 of the water-solubilising groups has some effect upon the magnitude of kc. As expected, strong electrostatic repulsion between the reactants within the ion-pair leads to a higher rate of dissociation. Now, the yield of redox products depends upon the partition fraction of the ion-pair kC 4s = k,+k, and, as we have shown previ~usly,~ #s can be related to the quantum yield for formation of separated products (bions) #ions = 4T x 4~ x 4s- In this expression, 4T refers to the quantum yield for formation of the triplet excited state of the zinc porphyrin (assumed to be 0.90 for all the porphyrins studied here) and dQ refers to the probability of quenching the triplet excited state of the porphyrin at a given concentration of quencher: k(MV2+) [MV2+] 4Q = k(MV2+) [MV2+] + zT-l ' The values were measured by flash-photolysis techniques and are collected in table 4, together with the derived 4s terms.These #s terms, together with the calculated kc values, allow estimation of kB (table 4). TABLE 4.-EFFICIENCY OF CHARGE SEPARATION AND RATES OF NON-RADIATIVE DECAY FOR THE ION-PAIR ZnTMPyP4+ 0.75 0.83 1 . m 109 - 1.58 ZnMPyP+ 0.10 0.1 1 6.59 x 10l0 - 1.55 ZnCPP- 0.037 0.04 1.58 x 10" - 1.48 ZnTSPP4- < 0.01 < 0.01 > 2.67 x 10" - 1.41 ZnTCPP4- < 0.01 < 0.01 > 2.67 x 10" - 1.34 The nature of the water-solubilising group has a marked effect upon the magnitude of &.In fact, 4s varies from 0.83 for ZnTMPyP4+ to < 0.01 for ZnTSPP4- and ZnTCPP4-. Thus, the use of a highly positively charged zinc porphyrin appears to favour formation of redox ions from the photoreduction of MV2+. Deactivation of the ion-pair to ground-state reactants (kB) is a type of non-radiative decay and, as such, the rate of the process should depend inversely upon the energy gap separating the initial and final states. The energy gap refers to the amount of electronic energy that must be converted to vibrational degrees of freedom and can be calculated from the redox potentials of the reactants. Thus, the standard free-energy change associated with decay of the ion-pair to ground-state reactants (AGP) was calculated from the E q values for the processes P++e- -+ P MV+ -+ MV2+ + e- and was corrected for any change in the electrostatic forces during the electron-transfer step.The calculated values are given in table 4, and it is seen that in all cases this process is highly exothermic. In fact, the calculated AGP values show that there is1882 QUENCHING OF Zn PORPHYRIN BY METHYL VIOLOGEN a substantial amount of electronic energy that must be dissipated in the form of heat, and if the system is subject to the Franck-Condon principle then this dissipation step may be relatively slow. It appears that quite small increases in AGP result in substantial decreases in k,.Thus, the rate of deactivation of the ion-pair to ground-state reactants is restricted by the ability of the reactants to dissipate large amounts of vibrational energy. The zinc porphyrins are large planar molecules that are not readily amenable to geometry changes, so that the majority of the excess energy must be removed cia the MV2+ molecule. Overall, the above work has highlighted that a combination of electrostatic factors and free-energy driving forces can lead to favourable conditions for formation of separated ion products. With ZnTMPyP4+ and MV2+, the strong electrostatic repulsion between the reactants results in an unstable encounter complex but assists dissociation of the ion-pair, whilst the high redox potential for oxidation of ground-state ZnTMPyP4+ ensures that there is virtually no thermodynamic driving force for formation of the ion-pair; however, once it is formed, the ion-pair cannot easily dissipate its large amount of electronic energy into vibrational degrees of freedom. When added together these factors favour poor triplet quenching rate irradiation times/min 0 0.5 1 .o 1 .5 2 -0 2.5 3.0 4.0 5.0 6.0 8 .O t ' l I I I I I ] 350 360 370 380 390 400 410 420 430 h/nm FIG. 3.-Absorption spectral profile showing the appearance of MV+ as a function of the irradiation time for the ZnTMPyP4+ system.M-C. RICHOUX A N D A. HARRIMAN 1883 constants but high yields of redox products. With highly negatively charged metallo- porphyrins such as ZnTSPP4- and ZnTCPP4- the converse is true, and although these systems give high triplet quenching rate constants, the yields of redox products are very low.PHOTOGENERATION OF HYDROGEN Previously it was demonstrated that a weak reductant such as EDTA could intercept reverse electron transfer between P+ and MV+ by reducing P+. Since the oxidised form of EDTA undergoes irreversible decomposition, this has the effect of stabilising MV+. As such, irradiation of ZnTMPyP4+ in aqueous solution at pH 6.9 containing MV2+ ( mol dm-3) and EDTA (1 W2 mol dm-3) resulted in a build-up in the concentration of MV+, as shown in fig. 3. For this system, the rate of appearance of MV+ was very fast, as shown by fig. 4, confirming the high efficiency for formation of redox products 0 5 10 15 irradiation time/min FIG.4.-Rate of formation of MV+ as a function of the irradiation time for the ZnTMPyP4+ system. OD(ZnTMPyP4+, 1 cm) = 0.5. expected from a highly positively charged chromophore. In identical experiments both ZnMPyP+ and ZnCPP- gave rise to formation of MV+, although at much slower rates than found for ZnTMPyP4+ (fig. 5 ) . In contrast, irradiation of the highly negatively charged chromophores, ZnTSPP4- and ZnTCPP4-, did not lead to reduction of MV2+. Thus, the yields of MV+ obtained from steady-state irradiations in the presence of EDTA as a sacrificial electron donor follow the order expected on the basis of the calculations outlined earlier. Quantum yields for the formation of MV+ dMV+ from the above systems were measured for irradiation at 567 Ifi 2 nm and are collected in table 4.The observed dMV+ values range from 0.75 for ZnTMPyP4+ to < 0.01 for ZnTSPP4- and ZnTCPP4-, and provide further substantiation for the hypothesis that strongly positively charged chromophores favour the photoreduction of MV2+ to MV+. It is well-established that colloidal dispersions of noble metals such as Pt or Pd are able to catalyse the exchange reaction between MV+ and H20. Consequently,1884 QUENCHING OF Zn PORPHYRIN BY METHYL VIOLOGEN 0.6 --I 10 20 30 40 50 60 70 80 0 irradiation time/min FIG. 5.-Rate of formation of MV+ as a function of the irradiation time for ZnMPyP+ (--@-) and ZnCPP- (-O-). OD(ZnMPyP+, 1 cm) = 0.05, OD(ZnCPP-, 1 cm) = 0.5. irradiation of the above systems in the presence of colloidal Pt allowed the reduction equivalence stored on MV+ to be converted into H, Pt 2MV++2H2O + H,+2MV2++OH-.With ZnTmPyP4+ the quantum efficiency for formation of H,, #iHz, was 0.60 at pH 5 and for short irradiation periods, so that almost all of the reduced MV+ could be used for hydrogen production. Hydrogen was also obtained with ZnMPyP+ and ZnCPP- chromophores, but no H, could be observed, under our conditions, with ZnTCPP4- and ZnTSPP4- as chromophore. CONCLUSIONS The above work has shown that the nature of the water-solubilising groups can have a drastic effect upon the efficiency of the metalloporphyrin photosensitised reduction of water to H,. Using MV2+ as electron relay, optimum results were obtained with a highly positively charged metalloporphyrin, and the effects of the water-solubilising groups can be explained in terms of electrostatic factors and ground-state redox potentials.In fact, the high redox potential reported for ZnTMPyP5+/ZnTMPyP4+ couple suggests that the oxidised form should be capable of the oxidation of water to oxygen, in the presence of a suitable catalyst. Thus, with the ZnTMPyP4+/MV2+ system it may be possible to achieve complete cleavage of water into H, and 0,. Obviously, as regards a practical solar-energy storage device, it is not desirable to produce a mixture of hydrogen and oxygen since such mixtures are explosive and expensive to separate; instead, the gaseous products must be produced at remote sitesM-C. R I C H O U X A N D A. H A R R I M A N 1885 so that the hydrogen can be collected. Before this can be achieved, however, it is necessary to use the oxidised form of ZnTMPyP4+ to liberate 0, from water and the feasibility of this process will be outlined in a later paper. We thank the S.R.C. and the European Economic Community for financial support, and we are greatly indebted to Sir George Porter, F.R.S. for helpful discussions. G. R. Seely, Photochem. Photobiol., 1978, 27, 639. A. Harriman and M-C. Richoux, J . Photochem., 1981, 15, 335. K. Kalyanasundaram and M. Gratzel, Heh. Chim. Acta, 1980, 63, 478. A. Harriman, G. Porter and M-C. Richoux, J. Chem. Soc., Faraday Trans. 2, 1981, 77, 833. G. McLedon and D. S. Miller, J . Chem. SOC., Chem. Commun., 1980, 533. A. Harriman and M-C. Richoux, J . Photochem., 1980, 14, 253. A. Harriman and M-C. Richoux, J . Chem. SOC., Faraday Trans. 2, 1980, 76, 1618. M. Gouterman and D. Holten, Photochem. Photobiol., 1977, 25, 85. J. A. Anton and P. A. Loach, J . Heterocycl. Chem., 1976, 13, 717. * R. H. Felton, in The Porphyrins (Academic Press, I ondon, 1978), vol. V, part C, p. 53. l o A. Holten, M. W. Windsor, W. W. Parson and M. Gouterman, Photochem. Photobiol., 1978,28,951. l 2 R. G. Little, J. A, Anton, P. A. Loach and J. A. Ibers, J . Heterocycl. Chem., 1975, 12, 343. l 3 A. Harriman, J . Chem. Soc., Faraday Trans. I , 1980, 76, 1972. l4 P. Debye, Trans. Electrochem. Soc., 1942, 82, 265. l 5 A. Harriman, G . Porter and M-C. Richoux, J. Chem. SOC., Faraday Trans. 2, 1981, 77, 1175. (PAPER 1/1211)

ISSN:0300-9599    

DOI:10.1039/F19827801873

出版商:RSC    

年代:1982

数据来源: RSC