J . c'hem. Soc., Furuduy Trans. I , 1985, 81, 2323-2331 Ultraviolet-Visible Spectropho tome tric Study of N-Alkylpyridinium Iodides in Non-aqueous Solvents Evidence for the Formation of Solvent-shared Ion-pairs BY MOHAN PAL AND SANJIB BAGCHI* Department of Chemistry, University of Burdwan, Burdwan 7 13 104, India Received 28th September, 1984 The u.v.-visible absorption spectra of solutions of various alkylpyridinium iodides have been studied at various temperatures and in binary-solvent mixtures. The results are interpreted as due to the existence of an equilibrium between the contact ion pair and solvent-shared ion pair in a solution in non-dissociating solvents. The band at ca. 42000 cm-' has been attributed to a modified charge-transfer-to-solvent transition of the iodide ion within a solvent-shared ion pair.The thermodynamic parameters (AGO, AH" and AS") for the interconversion of two forms of ion pairs have also been determined in various solvents. N-alkylpyridinium iodides (RPy+I-) provide a suitable system for the spectroscopic study of ion-pair formation with respect to the solvation characteristics of the ion pair. These compounds exist as ion pairs in solution, and the longest-wavelength band arises through to a charge-transfer (c.t.) transition within the contact ion The existence of another form of ion pair, viz. the solvent-separated ion pair, in the case of these compounds was invoked by Mackay and Poziomek4 to explain qualitatively the variation of the extinction coefficient of the visible c.t. band with solvent polarity. Recently we obtained indirect evidence for the presence of a solvent-shared ion pair (s.sh.i.p.) where one solvent molecule separates the pyridinium ion and the iodide However, u.v.-visible spectroscopy has yet to provide direct evidence for the existence of s.sh.i.p.in a solution of these salts. In the present work we have studied the u.v.-visible absorption of N-methylpyridinium iodide, N-methylpyrazinium iodide and N-ethyl- 4-cyanopyridinium iodide in various solvents and at various temperatures. The U.V. spectra of these compounds in non-dissociating solvents indicate the presence of a solvent-shared ion pair in equilibrium with a contact ion pair (c.i.p.). Thermodynamic parameters for the process of interconversion of the ion-pair subspecies have also been determined. EXPERIMENTAL MATERIALS The N-alkylpyridinium iodides were prepared by quaternising the corresponding N-substituted heteroaromatics with methyl or ethyl iodide in the dark.4.6 The purities of the compounds were checked by noting the melting points, the longest-wavelength band maxima (Amax) and the corresponding E values, as follows.N-Methylpyridinium iodide: found m.p. 118 "C (lit. 1 1 6 1 17 "C);' found Amax = 375 nm (lit. 374 nm); found E = 1200 dm3 mo1-l cm-' (1200 dm3 mol-l cm-l) in CHCI,. N-Ethyl-4-cyanopyridinium iodide: found m.p. 140 "C (lit. 140-141 "C);4 found Amax = 484 nm (lit. 482 nm);4 found E = 1210 dm3 mol-l cm-l (lit. 1200 dm3 mol-l ~ m - ' ) ~ in CH,CI,. N-Methylpyrazinium iodide: found m.p. 135 "C; A, = 470 nm; E = 2000 dm3 mo1-l cm-l in CH,Cl,. The purified and dried solvents were 2323u.v.-VISIBLE STUDY OF N-ALKYLPYRIDINIUM IODIDES j0 i50 300 '400 450 500 A/nm Fig. 1. Complete spectrum of N-methylpyrazinium iodide in dichloromethane at (I) 25 and (11) 45 "C. (111) N-Methylpyrazinium perchlorate in dichloromethane. ( a ) 0.1 cm cell; (b) 1 cm cell. distilled immediately before the experiment and all solvents were refluxed with calcium hydride immediately prior to use. This ensured the absence of peroxides. SPECTROPHOTOMETRIC MEASUREMENTS The experiments at room temperature were carried out in either a Beckman model 26 or a Cary model 17 spectrophotometer using stoppered cells placed in a thermostatted cell holder. Measurements at lower temperatures were made in the Cary model 17 instrument.The cell compartment in the low-temperature experiments consisted of an unsilvered quartz Dewar vessel into which nitrogen or oxygen gas was passed.6 The temperature was controlled to within & 1 "C by regulating the rate of flow of cold vapour. For temperature measurements a copper-constantan thermocouple was used; one junction of the thermocouple was inserted directly into the solution. The cell was stoppered and sealed with wax to prevent any absorption of moisture. This has been verified by checking the reproducibility of the spectrum at room temperature and at the end of each experiment. For spectral measurements in the U.V. region a cell of 1 mm pathlength was used. This was done to reduce the absorbance values and to nullify the effect, if any, of the absorption edge of the solvents.For measurements in the visible region the length of the cell was varied from 0.5 to 4 cm. The concentrations of the solutions were kept at ca. 10-3-10-4 mol dm-3. RESULTS AND DISCUSSION SPECTRAL CHARACTERISTICS PURE SOLVENTS The complete absorption spectra of N-methylpyrazinium iodide and perchlorate in dichloromethane are shown in fig. 1 . The spectral characteristics of the various alkylpyridinium iodides under study are summarised in table 1 . All the bands present in the case of dichloromethane and dichloroethane are found to obey Beer's law. TheM. PAL AND S. BAGCHI 2325 Table 1. Spectral characteristics of N-alkylpyridinium iodides in various solvents compound solvent N-methylpyrazinium iodide N-methylpyrazinium perchlorate N-et hyl-4-cyanopyridinium iodide N-ethyl-4-cyanopyridinium perchlora te N-methylpyridinium iodide N-methylpyridinium perchlorate dichloromethane dichloroethane ace toni trile dichloromethane dichloroethane acetonitrile dichloromethane dichloroethane acetonitrile dichloromethane dichloroethane acetonitrile dichloromet hane dichloroethane acetonitrile dichloromethane band maximum/nm group A group B group C 470 318 267 (273)" 238 475 320 268 (273)" 237 415 293 267 (273)" 246 _ _ 267 (273)" - _ _ 268 (273)" - - _ 267 (273)" - 485 327 275 (284)" 238 (235)" 490 330 275 (284)" 237 (235)" 420 295 276 (283)" 246 (235)" -.- 276 (284)" 234 _ - 276 (284)" 234 _ _ 276 (284)" 234 372 290 (254)" (259)" (264)" 241 375 290 (254)" (259)" (264)" 239 340 - (254)" (259)u (264)" 246 - _ (254)" 259 (264)" - a Shoulders obtained.band maxima have been collected in three groups. Group A corresponds to c.t. transitions within a contact ion 4 9 *, The band in group B is common to both the perchlorates and iodides and is presumably due to a n-n* transition of the cation.lou*b In group C the band at 246 nm in acetonitrile appears for all the alkylpyridinium iodides and is supposedly due to the charge-transfer-to-solvent (c.t.t.s.) transition of free iodide ions.llaTb Inspection of table 1 shows that in a non-dissociating solvent, e.g. dichloromethane or dichloroethane (where free iodides are not pre~ent),~? the iodides show an absorption at ca. 240 nm. The position of the band depends on the solvent and the nature of the cation.In dichloromethane or dichloroethane the intensity of band C increases (a simultaneous decrease in the intensity of the c.t. bands) with a decrease in temperature, although the bands position and shape remain unaltered (fig. 1). The process is reversible. At any temperature the relative heights of band C and the longest-wavelength c.t. band is independent of solute concentration; thus a solute-solute interaction is ruled out. MIXED SOLVENTS Studies in mixed solvents containing dichloromethane and cyclohexane at 300 K (the concentrations of the solute in various mixtures were kept constant) indicate that with a decrease in the mole fraction of dichloromethane the intensity of band C decreases. A simultaneous increase in the c.t. absorption is also noted. Experimental curves show an isosbestic point [fig.2(a) and (b)]. However, we could not vary the mole fraction of the non-polar component in the binary solvent mixtures over a wider range owing to the limited solubilities of the compounds in these solvents.U.V.-VISIBLE STUDY ! lr, 02- 04 0.1 ' 02 -._ I 220 250 300 350 A/nm OF N-ALKYLPYRIDINIUM IODIDES 30 250 280 310 A/nm Fig. 2. U.V. spectra of N-alkylpyridinium iodides in a mixed binary solvent (dichloromethane + cyclohexane) : (a) N-methylpyrazinium iodide and (b) N-methylpyridinium iodide. X, decreases in the order I > I1 > I11 > IV. Since free iodide ions are not present in these non-dissociating solvents the band at ca. 240 nm in these compounds is presumably due to an associated species. However, the high intensity of this band ( E c lo4 dm3 mol-1 cm-l) relative to that of the visible c.t.bands ( E M lo3 dm3 mo1-1 cm-l) excludes the possibility that the band originates through charge transfer within a contact ion pair. The work of Griffith et a1.l29 l3 shows that alkylammonium iodides in such solvents exist predominantly as solvent-shared ion-pairs characterised by similar bands. Thus our observations at different temperatures and in mixed binary solvents indicate that both s.sh.i.p. and c.i.p. exist in equilibrium in solution: where RPy+..-S..-I- is an s.sh.i.p. and its formation is favoured at lower temperatures. From the relative heights of the two bands (band C and the longest-wavelength c.t. band) at different temperatures one may evaluate AHo for the process, assuming the extinction coefficients to be independent of temperature. In the present case a value of AH" M - 2.00 kJ mol-l for N-methylpyrazinium iodide has been obtained.The role of cyclohexane (a non-polar solvent) is only to change the concentration of dichloromethane (the polar component), thus modifying the above equilibrium. The existence of solvent-shared species in equilibrium with the c.i.p. for these compounds in dichloroethane has also been reported by us in a recent comm~nication.~ In our case we have been able to detect spectrophotometrically two distinct forms of ion-pairs, while the N-alkylammonium iodides exist predominantly in one form,12 viz. s.sh.i.p. The ratio of the concentration of the solvent-shared ion pair to the contact ion pair (z.e.K, in dilute solutions) depends on the relative strength of the ion-solventM. PAL AND S. BAGCHI 2327 interaction and the cation-anion interaction. In the case of alkylammonium iodides the bulky alkyl groups probably hinder the approach of the iodide ion to the positively charged nitrogen, thus making the cation-anion interaction relatively weaker, leading to a very large value of K,; i.e. the concentration of c.1.p. in solution is too small. However, in our case the iodide ion may approach the positively charged pyridinium ring more Thus there is a greater possibility of the formation of c.i.p. in solution. Moreover, the existence of vacant n orbitals in the pyridinium ring makes it a good acceptor, so that a c.t. band is observed in the visible region for the c.i.p.This probably explains why we get both forms of ion pairs in detectable amounts in these solutions in non-dissociating solvents. NATURE OF THE TRANSITION Symons and coworker^^^,^^ have explained the origin of similar bands in alkyl- ammonium iodides as being due to a c.t.t.s. transition of the iodide ion within a solvent-shared ion pair. The high value of E and the dependence of the band position on the cation and the solvent (table 1) in our case also indicate the c.t.t.s. character in these transitions. Note that the c.t.t.s. band of the iodide ion in an s.sh.i.p. undergoes a hypsochromic shift with respect to the unperturbed c.t.t.s. band of the iodideion. The hypsochromic shift 0fthec.t.t.s. transition in the case ofalkylammonium iodides in the field of the cations has been explained in terms of an increased electrostatic stabilisation of the ground state of the ion pair.14 For spherical ions the interaction is expected to be a function of the distance between two charge centres; for alkylammonium iodides the shift is strongly dependent on the nature of the cation, shifting to lower energies with increasing cation size.12 If we assume a model for c.t.t.s.in which the transfer of charge does not take place centrosymmetrically, it is expected that the cation would not only interact with the iodide ion, stabilising its ground-state energy, but also interact with the solvent in the Franck-Condon excited state. While the former interaction produces a hypsochromic shift, any stabilising interaction between the solvent and the cation would lead to a bathochromic shift of the iodide c.t.t.s.transition. In the case of alkylammonium iodides the interaction with the excited states is probably constant and the band shows a cation dependence as depicted earlier. On the other hand, for alkylpyridinium iodides the charge distribution in the cation is far from spherical and both interactions are probably dependent on the nature of the cation, so no definite conclusions can be made. We offer here a qualitative explanation in terms of a modification of the energy levels of iodine atom (formed due to c.t.t.s.). The presence of the cation at close proximity will distort the solvent shell around the iodide ion from spherical to axial symmetry. In such a field the 2P3,2 level of the iodine atom will split into two doublets corresponding to mj values of f 3/2 and 1/2.Thus the c.t.t.s. band of the free iodide will split into two bands; the extent of the splitting will depend on the environment of the iodine atom. According to Mulliken's theory of charge transfer' the intensities of the bands will depend upon the overlap between the iodine orbital and the orbital which the electron enters. Nothing definite can be inferred, owing to the lack of a complete theory of c.t.t.s.16 It is possible that the intensities differ appreciably and that we detect only the hypsochromically shifted band experimentally. A similar explanation was invoked to explain the variation of energy difference between the two c.t. band maxima as a function of the cation in these compounds.8 The shift in the band maxima on varying the cation in the present work parallels the difference between the two c.t.band maxima in these compounds (table l), indicating that a similar mechanism may be operative in the two cases. Further work with other alkylpyridinium iodides is in progress.2328 U.V.-VISIBLE STUDY OF N-ALKYLPYRIDINIUM IODIDES THERMODYNAMIC PARAMETERS SPECTROPHOTOMETRIC PROCEDURE FOR THE DETERMINATION OF Ks AND THE EXTINCTION COEFFICIENTS (a) NON-DISSOCIATING SOLVENTS. The spectrophotometric evaluation of K, requires a knowledge ofthe molar extinction coefficients at the band maxima for the two ion-pair subspecies present in solution which are not determinable independently. For this reason we have determined Ks by a procedure similar to that described by us in the case of dissociating solvent^.^ The method consists of monitoring the visible c.t.band as a function of the composition of a binary solvent mixture with dichloromethane as one component. The other component is a non-polar solvent like cyclohexane (D = 2.023 at 20 "C) or tetrachloromethane ( D = 2.22 at 25 "C). These solvents do not take part in the formation of an s.sh.i.p. and only serve to vary the activity (a,) of the solvent taking part in s.sh.i.p. formation. Assuming that the activity coefficients of ion-pair species may be taken as unity and that the s.sh.i.p. does not absorb at the characteristic c.t. band of the c.i.p., we may write for the absorbance of a solution in non-dissociating solvents at the c.t.band maximum A = E,( 1 + K, as)-' Co 1 (2) where Co is the total concentration of the solute and E, is the molar extinction coefficient of the contact ion pair at the c.t. band maximum. For a particular solvent composition as is constant and thus Beer's law appears to apply. However, the extinction coefficient determined experimentally will be given by (3) eapp = E,( 1 + Ks as)-'. Eqn (3) may be rearranged to give 1 1 Ks - --+-aa, - Eapp Ec Ec (4) where as has been replaced by the mole fraction (Xs). A plot of against X , in the range 1 b X , b 0.5 is almost linear (fig. 3). (The limited solubility of these compounds in dichloromethane prevented us from taking readings in solutions where X, < 0.5.) Moreover, the nature of the straight line does not depend on the choice of the non-polar component, which again supports the assumption that the non-polar component does not form an s.sh.i.p. From the slope and the intercept of the straight line we may determine E , and K, separately.By a similar procedure the monitoring of the characteristic U.V. band due to an s.sh.i.p. might give the value of K, and E ~ , the molar extinction coefficient for the c.t.t.s. transition within the s.sh.i.p. However, the reduction of transparency of the solvent and the existence of neighbouring strong bands for the cations prevented us from making any quantitative measurements in the U.V. region. (b) DISSOCIATING SOLVENTS. In order to determine the value of capp in a solvent where free ions are also present, only the intensity of the visible band was monitored as a function of salt concentration.Use was made of the equation and capp = E,( 1 + K, as)-' was determined by a graphical m e t h ~ d . ~M. PAL AND S. BAGCHI 2329 0.0 0 7.00 6.0 0 P 9 2 5.0 C 4.0 0 3.0 C 1.0 0.'5 1.00 XS Fig. 3. Plot of l/capp against X,, the mole fraction of dichloromethane: (a) N-ethyl-4- cyanopyridinium iodide, (b) N-methylpyridinium iodide and (c) N-methylpyrazinium iodide ; 0, dichloromethane; + , cyclohexane; A, dichloromethane + tetrachloromethane and x , pure dichloromethane. Table 2. K , and E, values for various ion pairs in dichloromethane at 25 "C compound K , at 25 "C c,/drn3 mol-l cm-I N-methylpyrazinium iodide 0.5773 3150 N-methylpyridinium iodide 0.9238 2400 N-ethyl-4-cyanop yridinium iodide 1.001 2400 Table 2 gives values of K, and E , for various ion pairs in dichloromethane. The magnitude of K, for a particular ion pair does not depend on the wavelength region studied (within the c.t.band), indicating that the assumption that the s.sh.i.p. does not absorb at the c.t. band is correct. The value of K, is also of the same order of magnitude as that estimated by other procedure~.~9~~ The value of E , at the band2330 U.V.-VISIBLE STUDY OF N-ALKYLPYRIDINIUM IODIDES Table 3. AH" and AS" for process (1) in various solvents Eapp/dm3 AH" AS" compound solvent T/K mol-l cm-l /kJ mo1-l /J K mol-l N-methylpyrazinium iodide N-methylpyridinium iodide N-ethyl-4-cyanopyridinium iodide dichloromethane 269.7 288.0 298.0 acetone 298.0 303.0 313 ethanol 293.0 303.0 3 13.0 acetone 293.0 303.0 310.0 dichloromethane 283.0 288.0 298.0 acetone 288.5 298.0 303.5 309.7 2000 -5.0148 920 - 8.988 660 - 2.898 580 - 5.67 1200 - 7.469 833 877 - 3.032 - 2.1836 - 4.3025 - 0.069 -4.52 - 4.628 maximum is the same as that obtained for other solvent^,^-^^ so that the true molar extinction coefficients for the absorbing species seem to be relatively insensitive to solvent variation.The values of eapp in pure solvents, however, are dependent on temperature (table 3). This is readily understandable from the expression for E , ~ ~ , which contains K, in the denominator. We have not been able to determine Ks at various temperatures using the above procedure. On the other hand the temperature variation of cap4 in pure solvents may provide a quantitative measure of AH" and ASo in the following way.For pure solvents as = 1 and we have In the narrow temperature range employed in the present study E, may be assumed to be independent of temperature and one may calculate the value of K, at each temperature from the relation (7) using the experimental value of E,, assuming it to be independent of temperature. A plot of log K, against l / T (fig. 4) would then give AH" and AS". These values are listed in the table 3. Process (1) is exothermic, a fact also noted by others.18 The value of AH" obtained by this procedure is the same as that obtained from the variation in intensity of the bands with temperature. For a particular ion pair the magnitude of AHo for the formation of a solvent-shared ion pair increases in the order alcohol-shared pair > acetone-shared pair > dichloro- methane-shared pair, indicating that the stability of an s.sh.i.p.runs parallel with the polarity of the solvent. The magnitude of AHo for a particular solvent also depends on the nature of the ion pair. The negative value of ASo means greater ordering when an s.sh.i.p. is formed. According to the primitive model in which the solvent is represented as a continuum, the value of ASo should be zero since the continuum has no structure. However, theM. PAL AND S. BAGCHI 233 1 05 i i 1 1 i 3 3 2 3.4 3.6 3.8 103 K I T Fig. 4. Plot of log Ks against 1 / T for various alkylpyridinium iodides in different solvents: (a), (b) and (c) N-methylpyrazinium iodide in dichloromethane, acetone and ethanol, respectively; (d) N-methylpyridinium iodide in acetone; (e) and (f) N-ethyl-4-cyanopyridinium iodide in dichlorome t hane and ace tone, respectively .finite negative value of ASo observed in these cases may be thought of as mainly being due to a loss of translational and/or other degrees of freedom of the solvent molecule engaged in the formation of an s.sh.i.p. The magnitude of ASo also depends on the solvent and the nature of the ion pair, but no correlation can be made with the structure of the component at present. We are thankful to the referees for their valuable comments and suggestions and we thank Prof. Mihir Chowdhury of I.A.C.S. Calcutta for helpful discussions. M. P. thanks the Indian University Grants Commission, New Delhi, for a scholarship. I 2 3 3 5 6 7 8 9 10 1 1 12 I 3 14 15 16 l i IH R. S. Mulliken and W. B. Person, Molecular Complexes (Wiley Interscience, New York, 1969). E. M. Kosower, An Introduction to Physical Organic Chemistry (Wiley International Edition, New York, 1968). T. R. Griffiths and D. C . Pugh, J . Solution Chem., 1979, 8, 247. R. A. Mackay and E. J. Poziomek, J . Am. Chem. SOC., 1970,92, 2432. M. Pal and S. Bagchi. J . Cfiem. SOC., Farudq Truns. I . 1985, 81. 961. S. Bagchi and M. Chowdhury, J . Phys. Chem., 1976, 80, 21 1 1 . R. A. Mackay, J. R. Landolph and E. J. Poziomek, J . Am. Chem. SOC.. 1971,93, 5026. S . Bagchi and M. Chowdhury, J . Phys. Chem., 1979, 83, 629. J. W. Verhoeven, I. P. Dirkx and Th. J. De Boer, Tetrahedron, 1969, 25, 3395. a E. M. Kosower and J. A. Skorcz, J . Am. Chem. SOC., 1960,82, 2195. b E. M. Kosower, Molecular Biochemistry (McGraw-Hill, New York, 1962). a E. M. Kosower, R. L. Martin and V. W. Meloche, J . Chem. Phys., 1957,26, 1353. b M. Smith and M. C. R. Symons, Trans. Faraday SOC., 1958, 54, 338; 346. T. R. Griffiths and R. H. Wijayanayake, Trans. Furaday SOC.. 1970. 66, 1563. R. G. Anderson and M. C. R. Symons, Truns. Faraduj Soc.. 1969, 65, 2537. T. R. Griffiths and M. C. R. Symons, Mol. Phys., 1960. 3, 90. M. J. Blandamer, T. E. Cough and M. C. R. Symons. Trans. Furadu?* Soc., 1966, 62, 286. M. F. Fox and E. Hayon, J . Chem. Soc., Faruday Truns. I , 1977. 73. 1003. M. Pal and S. Ragchi, lnd. J . Chem.. 1984, 23A, 800. T. F. Hogen-Esch and J . Smid, J . Am. ChPrn. Soc.. 1966. 88, 307. (PAPER 4/1675)