J. Chem. SOC., Faraday Trans. I , 1986,82, 3097-3 112 A Nuclear Magnetic Resonance Study of the Solvatochromism of a Pyridinium Betaine J. Graham Dawber* and Richard A. Williams Department of Chemistry and Biology, North Stafordshire Polytechnic, Stoke-on-Trent ST4 2DE The lH and 13C nuclear magnetic resonance spectra of the solvatochromic indicator 2,4,6-triphenylpyridinium-N-3,5-diphenylphenol betaine have been measured in CDCl,, CD,OD, [2H,]DMS0 and [2H,]acetone solvents and 1 : 1 molar mixtures of CDC1,-DMSO, CDC1,-CD,OD, CDC1,- acetone, DMSO-acetone, DMSO-D,O and CD,OH-D,O. An attempt has been made to assign the resonance signals and to measure the changes in resonance positions with change of solvent. The major changes in resonance positions that accompany change of solvent are for those parts in the molecule close to the centres of charge. The effects of the differences in solvation at these centres, which are likely to be associated with the extreme solvatochromic properties of the betaine, are also transmitted to other parts of the molecule.The important effect of solvents upon chemical reactions has long been known and there have been several attempts to correlate empirical and experimental parameters of solvents with their polarity and solvation properties. These include1 the so-called Y values, X values, R values, II* values, 2 values and ET values. Of these, the 2 values and the ET values1 depend upon the solvatochromic behaviour of dyes, the term solvatochromism referring to the shift of an electronic absorption band when varying the polarity of the medium.Such changes result from physical intermolecular solute- solvent interaction forces, which usually involve alterations in the electronic ground state or may also involve the excited state of the absorbing species. The u.v.-visible solvatochromism of such dyes, while providing useful solvent polarity parameters, gives no information regarding the localised changes within the solute concerned when the solvent polarity is changed. The proton-decoupled 13C n.m.r. spectrum, on the other hand, can, in principle, give such information since each different carbon atom gives rise to a single line frequency. Dimroth, Reichardt et aL2 discovered a pyridinium N-phenol betaine (I) (2,4,6- trip hen ylp yridinium-N- 3,5 -dip hen ylphenol bet aine), whose solva toc hr omism was very large (A,,, = 810 nm in diphenyl ether to 450 nm in water) and which was used to establish the solvent polarity ET sca1e.l It was decided, therefore, to undertake a lH and a 13C n.m.r.study of this interesting compound to give possible information concerning the localised intramolecular changes which occur when the solvent in which it is dissolved is changed, but also to study solvent mixtures with a view to detecting possible preferential solvation by one of the solvent components. A similar study has recently been carried out with a merocyanine dye which has also been used as a solvatochromic indicator .3 30973098 Solvatochromism of a Pyridinium Betaine H H H H H H H 0- Experimental 2,4,6-Triphenylpyridnium-N-3,5-diphenylphenol betaine (I) This compound was synthesised by the condensation of 2,4,6-triphenylpyrilium perchlorate4y with 2,6-diphenyl-4-aminopheno16 (11) to give the perchlorate salt which is the precursor of the betaine (I); m.pt.273-274 "C, (30% yield). Treatment of the perchlorate salt with sodium methoxide in methanol as described by Dimroth et aL2 gives, after recrystallisation from methanol-water (1 : 3), the required betaine (I), m.pt. 269-273 "C, which is in fact a dihydrate. The N.M.R. Spectra The n.m.r. spectra were recorded on a Jeol FX90Q Fourier-transform spectrometer (proton resonance at 89.55 MHz, carbon resonance at 22.49 MHz). The lH spectra were obtained using a 90" tip-angle and a spectral width of 900 Hz. The noise-decoupled 13C spectra were obtained with a tip-angle of 30" with a pulse repetition time of 1 s and ca.50 000 scans. Quantitative decoupled 13C spectra without nuclear Overhauser enhance- ment (NNE) for (I) in CDCl, were obtained using a tip-angle of 30" and a pulse repetition time of 30 s with gated decoupling during acquisition. Undecoupled 13C spectra were obtained with nuclear Overhauser enhancement (NOE) using gated decoupling with the proton decoupler switched off during acquisition and a pulse repetition time of 3 s. For all the 13C spectra the responses were acquired into 8K and zero-filled to 16K data points, and an experimental broadening of 0.9 Hz was applied prior to Fourier transformation. The lH and 13C chemical shifts were all measured relative to TMS as internal standard at S = 0.0 ppm.Spectra were recorded in CDCl,, CD,OD (also CD,OH), [2H6]DMS0 and [2H6]acetone solvents and 1 : 1 molar mixtures of CDC1,-DMSO, CDC1,-CD,OD, CDC1,-acetone, DMSO-acetone, CD,OD-D,O, and DMSO-D,O, and also with added HCl and NaOD to the latter mixture. TheJ. G. Dawber and R. A . Williams 3099 i j k I /I 4 a b c I 8.0 7.5 7.0 6, (PPm) Fig. 1. (a) lH n.m.r. spectrum of (I) in [2H6] acetone. (b) lH n.m.r. spectrum of (I) in CDC1,.3100 Solvatochromism of a Pyridinium Betaine Table 1. 'H chemical shifts of pyridinium betaine (I) in various solvents 'H chemical shift, 6 solvent abc g ij k n 9 r S ~~ ~~~ CDC1, [ 2H6]ace t o ne CD,OD 7.18 6.72 7.36 7.19 7.01 7.46 (split) (split) 7.19 6.78 7.43 [ 2H ;I DMSO 7.23 [2H6]DMSO-[2H6]acetOne 7.24 (split) CDC1,-CD,OH 7.20 [2H,]DMSO-CDCl, 7.20 CDC1,-[2H6]ace tone 7.17 (split) [2H6]DMSO-D20 7.2 1 (split) CD30H-D20 7.24 6.98 6.99 6.70 6.91 6.80 (split) 6.83 6.82 7.48 7.47 split) 7.47 7.47 7.41 7.46 split) 7.50 8.01 7.83 7.54 7.26 8.48 8.24 7.67 7.33 8.34 8.07 7.60 7.26 8.56 8.30 7.63 7.29 8.55 8.32 7.63 7.35 8.26 7.98 7.65 7.25 8.45 8.18 7.60 7.28 8.27 8.06 7.58 ? 8.40 8.21 7.64 7.31 8.39 8.09 7.69 ? ~ ~~~ -~ deuterated solvents (Aldrich) were used as received and had the following isotopic purities: CDCI, (99.8% D), CD,OD(99.5% D), [2H,JDMS0 (99 +% D) and [2H6]acetone (99.5 % D).The CDC1, was checked for acidity by extraction with water and measurement of the pH of the aqueous extract. 1 cm3 of CDCl, was shaken with 50 cm3 of distilled water, the pH of which changed from 6.4 to 5.8, corresponding to an acidity in the CDCl, of ca.2.5 x mol drn-,; this is insignificant in comparison with the concentration of betaine used, which is three orders of magnitude greater than this. The lH and 13C spectra were measured on the same filtered solution, which normally contained ca. 20 mg of betaine in 1.5 cm3 solvent. Results and Discussion The greatest resolution of the lH resonances was obtained for a solution of (I) in [2H6]acetone (and also in the [2H6]DMSO-[2H6]aCetOne mixture and the CDC1,- [2H6]acetone mixture) and this is shown in fig. 1 (a). The least splitting was observed for CDCl, solvent, shown in fig. 1 (b). From the integration traces of the lH spectra and the numbers of similar protons in different parts of the molecule a tentative assignment of the signals was made.These lH assignments and their chemical shifts in the various solvents are given in table 1. The proton spectra in the various solvents show qualitative similarities and differences. For example, the splittings of the bands due to Habc and Hijk depend quite markedly on solvent, and yet for these protons the chemical shifts do not change greatly. The overall changes in proton chemical shift for the rest of the molecule lie in the range 0.1-0.7 ppm. Several factors can contribute to solvent-induced chemical-shift changes.' These include contributions from the bulk solvent susceptibility, anisotropy, reaction field, hydrogen bonding, specific interactions and van der Waals forces.' In the case of electronic spectra, effects accompanying changes in solvent may also have contributions from differences between electron density distributions and dipole moments in the ground and excited states.However, in the case of n.m.r. spectra the majority of the molecules are in their electronic ground state and the lifetimes of the excited states are very short on the n.m.r. timescale. In any case, the n.m.r. spectra measured on the Jeol FX90Q instrument are carried out with the sample in light-free environment. Conse-J. G. Dawber and R. A . Williams 3101 quently the changes observed in the n.m.r. chemical shift produced by the solvent are likely to be caused by the normal solvent effects due to shape and polarity of the solute and solvent rather than excited-state contributions.Proton-proton couplings are fairly insensitive to solvent but the lH chemical shifts are solvent dependent. However, Benson and Murrel* concluded that no simple relation can be derived between the lH chemical shifts and the n-electron densities in aromatic molecules. It has been showng that the 13C and lH n.m.r. chemical shifts of the simpler CHCl, molecule in a range of solvents show that the changes in 6, range from 0 to 4.2 ppm going from cyclohexane to hexamethylphosphorimide, for which the corres- ponding changes in 6, are approximately half those for 6,. The values of AS, for CHCl, going from cyclohexane to methanol and acetone solvents are 1.35 and 1.76 ppm downfield, respectively, with the values being about half of these. Similarly, the addition of benzene to non-polar solutions of the more complex molecules of flavones produces characteristic proton chemical shifts.1° However, if the chemical-shift differences between non-equivalent protons are reduced in a particular solvent then the splittings will be affected.The splitting of the lH resonances of (I) is most apparent in acetone solvent alone or any mixture containing acetone. Thus the chemical-shift differences between adjacent protons are decreased in solvents other than acetone and the coupling constants are insufficient to split the peaks. Poly-halogen compounds, such as CCl, and CHCl,, are known to form weak charge-transfer complexes with aromatic compounds,ll and this must be a distinct possibility, with the pendant phenyl rings attached to the main betaine structure thereby influencing the chemical shifts of Habc and Hijk. In addition, the extent of splittings in the bands from H, and H, are also slightly solvent dependent.The proton resonances which are most affected by solvent are those from H, and H,, which might be expected in view of their positions in the molecule. Also affected, rather surprisingly, is the resonance from H,. The greatest solvent effect was upon the H, resonance, which is no doubt due to these protons being nearest to the cationic centre of the betaine molecule and being favourable to solvation by polar solvent molecules. The H, protons are deshielded by solvation in the order DMSO > acetone > CD,OH > CDCl,. In the case of binary solvent mixtures it is possible in principle to judge the presence of any preferential solvation by comparing the resonances of the solute in the solvent mixtures with the corresponding values in the pure solvents.If a particular line is nearer to its position in one of the pure solvents when the solute is dissolved in a solvent mixture then this could indicate preferential solvation at that particular site. From the H,, H, and H, chemical shifts it can be seen that there is little or no preferential solvation of these protons in the CDC1,-CD,OD mixture and the CDC1,-acetone mixture. It must be borne in mind, however, that the lH spectra will probably only reflect the extent of changes in hydrogen bonding, and to a lesser extent any changes in van der Waals complexing or n-complexing, which are more likely to be observed in the 13C spectra.In the case of the DMSO-CDCl, and DMSO-acetone mixtures the evidence suggests that there is some preferential solvation of the betaine in the region of H, by DMSO. The 13C chemical shifts of (I) in the various solvent systems are listed in table 2. There are a total of 19 distinguishable carbon atoms in the structure of (I), and in CDCl, solvent there are 19 lines in the noise-decoupled 13C n.m.r. spectrum (fig. 2). These lines are numbered 1 to 19 starting from the low-field end. The NNE decoupled 13C n.m.r. spectrum allowed the number of carbon atoms associated with each signal to be established and showed whether or not they were quaternary or proton-bearing. The carbon atoms in the betaine are labelled a to s as shown in structure (I).The 13C n.m.r. spectrum of the precursor of the betaine, namely 3,Sdiphenyl- 4-hydroxyaniline (11) was measured in CDCl, and the resonances designated by means of standard tables of assignments;I2-l4 these, along with other relevant assignments used, are given in structures (11)-(VIII). Using this information and that from the undecoupledTable 2. 13C chemical shifts, 6 (ppm) of the pyridinium betaine (I) in various solvents relative to TMSa peak no. from low CDC13- CDC13- CDCl3- [2H,]DMSO- [2H6]DMSO- field CDC1, [2H,]DMS0 [2H,]acetone CD30D [2H,]DMS0 CD30D [2H,]acetone [2H,]a~etone D2O 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 161.3 156.9 155.7 138.7 134.2 133.7 132.0 130.26 130.19 I 129.81 129.41 129.11 128.78 128.51 128.10 127.80 126.40 126.00 124.23 161.0 156.5 154.0 139.7 134.0 133.5 132.0 129.47 128.96 128.73 128.60 128.49 (sh) (sh) 127.21 128+oo 1 125.50 125.07 124.00 162.9 158.0 156.I 141.0 135.1 135.0 133.0 131.1 (sh) 130.61 130.02 129.89 l g 6 } 129.17 128.25 126.61 126.55 125.20 164.4 158.7 157.2 141.6 135.7 135.5 133.2 132.9 (sh) 131a0 (sh) 1 130.89 I 130.45 130.2 I 129.50 129.39 128.48 127.98 126.82 125.9 1 161.1 156.5 154.6 139.5 133.9 133.7 132.0 129.61 (sh) 129.53 139.20 128.80 128.69 128.36 128.20 127.30 125.71 125.40 124.10 (sh) 164.2 157.7 156.5 140.2 134.2 134.0 132.9 132.3 130.76 130.35 129.78 129.52 129.10 128.91 128.40 127.91 126.51 126.09 123.87 164.0 157.3 155.6 140.4 134.7 134.6 132.4 130.50 130.17 130.07 129.86 129.40 128.90 128.80 128.52 127.70 126.21 125.97 123.26 1 (sh) 163.1 157.1 154.7 140.8 134.9 134.3 132.4 i I 130.04 129.62 129.22 128.90 128.56 127.60 1 125.79 J 123.59 ~- a (sh).shoulder: a brace indicates the merging together of two or more lines. 164.6 156.6 153.9 140.6 134.1 133.6 132.0 129.61 129.41 128.84 128.56 128.29 128.05 127.15 125.13 121.76 CD30H- D2O _____. 158.4 157.3 151.3 148.6 140.5 135.1 134.9 133.42 132.59 131.11 130.90 130.57 130.18 129.50 129.16 128.80 127.53 125.34 - -~ ~J . G. Dawber and R. A . Williams 3103 I , , , I 12 I , I , , , , $ I I , , , , I , , , , , ( I ( , I , , , ) , , , , I 13C spectrum of the betaine in CDCl, (obtained with Overhauser enhancement) (fig. 3) the assignment of the 19 lines was attempted making use of the carbon-hydrogen coupling constants (lJCH, 2JCH, 3JcH).The results of this and relevant comments are given in table 3. A few of the assignments may not be completely unequivocal, but for the few doubtful cases intuitive suggestions are made in table 3. In table 4 are presented the assignments of the 13C resonances of the betaine in CDCl, along with the changes (Ad,,,,,) produced when the betaine is dissolved in [2H,]DMS0, [*H],acetone and CD,OD. Also presented are the data for 50 : 50 mol % of CD30H-D20 and 50: 50 mol % of [2H6]DMSO-D20. In general, all of the 6, values of (I) were found to be influenced by solvent, the effects for the pure solvents being approximately in the order CD,OD > acetone > CDCl, > DMSO. The solvent-induced 13C chemical-shift changes, Adc, fall in the range &3 ppm, and this is comparable with values found for simpler molecules, e.g.CHC1,.g Also, by comparison the values of Adc for benzene in different solvents are between -0.1 and 0.3 ppm, while those for the polar solute of ani- line range from - 2.7 to + 2.8 going from CCl, to DMSO s01vent.l~ Unlike proton chemi- cal shifts,s the changes in 13C chemical shift correlate with charge-density calculations for the various carbon atoms. l5 Solvent-induced chemical-shift changes can cause over- lapping of 13C resonance signals, which in one solvent may be separate, but in another solvent they are partially overlapped, for example the aromatic compound 3-bromo- biphenyl in cyclohexane and DMSO.lG Similar effects were observed for the betaine (I), as shown in table 2. Thus the chemical-shift changes, Adc, for (I) produced by different solvents appear to be similar in magnitude to those for simpler molecular systems, and hence the observed effects are likely to correspond to normal solvent effects arising from polarity changes.Table 4 shows that the greatest solvent-induced chemical-shift changes are for those carbon atoms which we have designated as being nearest the centres of charge in the molecule, and the least effkcts are for those carbon atoms which are3104 Solvatochromism of a Pyridinium Betaine 127.6 129.2 127.7 136.0 123.8 149.9 OH 118.2 132.6 148.6 128.5 0 141.9 0- 121.2 132.6 123.3 (VII) 118.1 cvrrr)J. G. Dawber and R . A . Williams 1 8 1 8 134 132 125 4 h 3105 ! 4 . , . , , , , , , , , , , , ~ , , l ' " ' ~ " ' ' 1 " ' ~ ' ' 160 140 120 6, (PPm) Fig.3. Undecoupled 13C n.m.r. spectrum of (I) in CDCI,. furthest away from the charge centres; this adds confidence to our assignments for the 13C n.m.r. signals. The carbon atoms which were influenced most by solvent were Cf, C,, C,, C,, C,, C, and C,, which correspond to the quaternary carbon atoms nearest to the centres of charge within the betaine. The C, and C, carbon atoms are less affected, probably owing to their being hydrogen-bearing. The downfield shift (higher 6) accompanying increased polarity of the solvent is greater for the C, and C , carbons, near the 0- centre, than for the C, and C, atoms adjacent to the N+ centre. A It is of interest to compare our results for the betaine (I) with those of another solvatochromic substance, the merocyanine dye (IX), the I3C n.m.r.spectra of which were measured in CDCl,, DMSO and CD,0H,3 and for which values of AdCDCl3 were calculated for comparison with our results. In the case of the change from CDCl, to DMSO, the largest negative AdCDCl3 change is for C,, and this corresponds to the same carbon atom in (I), i.e. Ch, while the largest positive AdCDCl3 change is for C,, at the negative 0- centre. In the case of (I) the largest positive Ad was for the adjacent carbon atom, i.e. C,. In the case of the change from CDCl, to CD,OH for (IX), all the solvent-induced chemical-shift changes were positive as for (I). The largest positive changes were for carbon atoms C,, C,, C, and C,, and these correspond to C1, C, and C, in (I). Thus the two systems are similar.The 13C chemical shifts of (IX) have been3106 Solvatochromism of a Pyridinium Betaine Table 3. 13C assignments of betaine (I) line no. of 6 (ppm) assign- H- comments on undecoupled spectrum no. C atoms in CDC1, ment bearing and assignment 1 1 2 2 3 1 4 2 5 1 6 2 7 1 8 2 9 2 11 4 12 4 13 4 14 2 161.3 156.9 155.7 138.7 134.2 133.7 132.0 130.26 130.19 129.40 129.11 128.78 128.51 f m h e P 1 S d 9 k C j n Likely to be C, since this will be the most deshielded C atom, [cf. structures (VII) and (VIII)]. Undecoupled gives "JCH with protons on C, (J z 8 Hz) Undecoupled gives singlet which is slightly broadened which could be due to unre- solved 2JcH splitting. Should be C, since 2JcH could occur with H,. Likely to be C , rather than C, since orlho to N+ at lowerfield (higher 6) than ortho to 0- [cf.structures (VI) and (VIII)] Undecoupled splits into narrow triplet with J z 4 Hz which is probably 'JCH coupling, i.e. C , split by the two H, atoms Undecoupled gives broadened singlet which could be C , since ortho to -C-O- at higherfield (lower 6) than ortho to N+ Probably C, since this will be shielded relative to C,. In principle there should be "JCH triplet splitting with H,, but the undecoupled spectrum shows shoulders on this peak Undecoupled broadened which may be a narrow triplet, i.e. unresolved 2JcH split- ting. Assigned as C, rather than C, since by comparison C, is at lower field (higher 6) than C, Undecoupled gives 2 triplets, lJCH = 162 Hz and "JCH = 8 Hz. Thus, this must be a proton-bearing carbon with possi- bility of 3JCH coupling, i.e.C, since only one C atom involved; 2JcH not apparent Undecoupled slightly broadened : it may be a narrow triplet (cf. line 6) i.e. unresolved 'JCF. Assigned as C, rather than C, since C, is at higher field (lower 6) than C , by comparison Undecoupled gives two peaks (lJCH = 162 Hz with some messy splitting from overlapping. Likely to be C,, and C,. Assign peak 9 as C, and 10 as C since ortho slightly higher field (lower j) than meta [cf. structure (VIII)] Undecoupled gives quarter, 'JCH and ortho 2JcH [cf. phenoxide anion (VIII)]; may be C, or C,: designate as C, As above (line 11) Probably Cj; see line 16 Undecoupled gives doublet ('JCH = 162 Hz) may be C, or C,; assign as C, since meta of X lowerfield (higher 6) than meta of VI.Coupling with H, and H, probably weak due to possible non- planarity of adjacent phenyl ringsJ . G. Dawber and R. A . Williams Table 3. (cont.) 3107 line no. of 6 (ppm) assign- H- comments on undecoupled spectrum no. C atoms in CDC1, ment bearing and assignment 17 18 19 15 2 128.10 g yes Undecoupled gives doublet (lJCH = 162 Hz); see line 14 16 4 127.80 b yes Undecoupled gives 2 large peaks (lJCH = 162 Hz) with slight broadening due to 2JCH coupling; probably cb or cj : assign as cb, [cf. (VIII) and (VI)] 2 126.40 a yes Undecoupled gives 2 peaks with triplet structure: probably C, or Ci 2 126.00 1 yes As for line 17 1 124.23 0 no Undecoupled gves very narrow triplet, i.e. 2JCH only. Possible for C, in pyridinium ring. Any 3JcH would involve another ring which is probably-not coplanar _ _ ~ ~ _ _ shown to give a linear correlation with the electron densities in the ground state.Although there will be a change in the permanent dipole accompanying an electronic transition, the calculation of the ground-state dipole moment of (IX) shows significant changes when the solvent polarity is changed. Such changes in the ground-state dipole moment induced by the surrounding solvent cage are thought to be particularly important in the case of solvatochromic l7 Thus this is likely to be the case not only for the merocyanine (IX) but also for the betaine (I). The 6, values for (I) in [2H,]DMS0 seem to be out of line with the relative polarity of this solvent as judged by its dielectric constant, E . This is illustrated in fig.4 for carbon atoms nearest the centres of charge, Cf, C,, and C,, where the points for DMSO at E = 50 are very different from the rest. This may be due to a difference in the type of bonding in the solvation of the betaine with DMSO. The relationship between E and ET is not linear, and for DMSO the value of ET seems low for its value of 6.l The changes in the 13C resonances for each carbon atom produced by the addition of D20 to DMSO and CQOH (50: 50 mol%) are plotted in fig. 5. Large changes do occur for the carbon atoms near the centres of charge. The resonances for C, move in the same direction for (I) in both solvents on addition of D20, but move in opposition for Cf. For Ch and C,, i.e. atoms near to the N+ centre, the resonances are affected by the addition of D20 to the CD,OH solvent, but not significantly in the case of DMSO, which may mean that there is strong interaction of the DMSO near the N+ centre of the betaine molecule.For reasons which are not clear, the C, resonance (for DMSO solvent) and C, (for CD30H solvent) are considerably affected by the addition of D20 to the organic solvent. Since D20 in high concentrations (50 mol%) does produce large changes for some carbon atoms of (I) it is perhaps pertinent to question the influence of small amounts of water in the supposedly pure organic solvents. Water is known to affect the 6, values of dipolar solutes by hydrogen bonding. For example, the 6, values for quinoline and isoquinolinel8 [slightly related to (I)] when dissolved in acetone change by up to ca.2 ppm on addition of water, but 50% by volume of water is required to do this. For a few percent (e.g. 2.5%) of water the changes are of the order of 0.2 ppm. Although solvatochromic indicators such as (I) have been suggested as a means of determining the water content of organic solvents, the estimation is based upon an empirical equation obtained for certain mixtures, and presumably with the indicator in the concentration3108 Solvatochromism of a Pyridinium Betaine 1 152 . \ ‘- X 10 20 30 40 50 E Fig. 4. The variation of 13C chemical shift (6,) with solvent polarity (dielectric constant E ) for atoms f, m and h: 0, CDC1,; A, [2H,]acetone; 0, CD,OD; x , [‘%,]DMSO. Fig. 5. Changes in chemical shifts (Ad) for (I) in [2H,]DMS0 ( x ) and CD,OD (0) produced by addition of D,O (1 : 1).range 10-3-10-4 mol dm-3 for visible spectroph~tometry.~~ The samples of pure deuterated solvents used in this n.m.r. study were freshly opened samples, and although they were not dried any further we are confident that by comparison with other n.m.r. work18 any traces of water played only a small role compared with the bulk solvent effect in the observed solvent-induced chemical-shift changes. We also measured the effect upon the 13C spectrum of the betaine dissolved in 50 : 50J. G. Dawber and R. A . Williams 3109 Table 4. Changes in 13C chemical shifts in various media change, Ad, from CDC1, line position 100% 100% 100% atom no. in CDC1, (PPm) (PPm (PPm) carbon line (PPm) [2H,]DMS0 [2H,]acetone CD,OD a b d e f g h i j k 1 m n P 9 r C 0 S 17 16 12 8 4 1 15 3 18 13 11 6 2 14 19 5 9 10 7 126.40 127.80 129.11 130.26 138.7 161.3 128.10 155.70 126.00 128.78 129.40 133.70 156.90 128.51 124.23 134.20 130.19 129.8 1 132.0 - 0.90 -0.59 -0.38 - 0.79 1 .oo -0.30 -0.10 - 1.70 - 0.93 -0.18 - 0.45 -- 0.20 - 0.40 - 0.02 - 0.23 - 0.20 - 0.72 - 0.34 0.00 0.2 1 0.45 0.78 0.84 2.30 1.60 0.15 0.40 0.55 0.48 0.6 1 1.30 1.10 0.66 0.97 0.90 0.42 0.8 1 1 .oo 1.58 0.68 1.10 2.64 2.90 3.10 1.29 1.50 0.82 0.72 1.04 1.80 1.80 0.99 1.68 1 S O 0.8 1 1.08 1.20 carbon line atom no. change from change from excess chemical shift, d,, in CD,OH in DMSO in 50 : 50 m o l x mixtures 50: 50 mixture 50: 50 mixture a 17 b 16 c 12 d 8 e 4 f 1 g 15 h 3 i 18 j 13 k 11 1 6 m 2 n 14 o 19 P 5 q 9 r 10 S 7 with D,O Ad (PPm> -0.45 0.48 0.36 0.52 7.00 - 6.00 - 0.23 - 5.90 - 0.22 0.68 0.45 - 0.40 - 1.40 0.00 - 0.57 4.80 1.59 0.22 1.70 with D,O Ad (PPO CDC1,- DMSO (PPm) - 0.37 - 0.06 -0.36 0.14 0.90 3.60 0.05 -0.10 0.06 - 0.04 -0.12 0.10 0.10 - 0.20 - 2.24 0.10 - 0.06 - 0.06 0.00 - 0.24 -0.21 -0.12 - 0.27 0.30 0.15 - 0.05 - 0.25 -0.14 0.00 0.02 0.10 -0.20 -0.14 - 0.02 -0.20 -0.30 -1.11 0.00 CDC1,- CD,OD (PPm) - 0.68 - 0.23 -0.14 0.72 0.05 1.35 0.05 -0.35 - 0.32 - 0.04 -0.15 - 0.60 -0.10 -0.10 - 1.20 -0.75 0.17 0.00 0.30 CDCI,- ace tone (PPm) DMSO- acetone (PPm) -0.30 -0.33 -0.10 -0.18 0.35 1.90 0.35 - 0.30 -0.31 - 0.09 0.15 0.25 -0.15 - 0.04 - 1.46 - 0.05 - 0.23 -0.14 -0.10 0.23 -0.13 - 0.09 - 0.25 0.45 1.15 0.44 -0.35 - 0.02 0.32 0.13 0.05 -0.15 0.07 - 1.01 0.35 0.00 0.00 -0.1031 10 Solvatochromism of a Pyridiniurn Betaine @ CDC13 [ 2H6]DMS0 Fig.6. Preferential solvation profile in the CDCl,-[2H,]DMS0 mixture. @ CDC13 CD30D Fig. 7. Preferential solvation profile in the CDC1,-CD,OD mixture. [2H,]DMSO-D,0 caused by the addition of H+ ion and OD- ion (by small additions of concentrated HCl and concentrated NaOD). In the case of added acid there were, in general, downfield shifts of 1-2 ppm. However, atom C , experiences a large downfield Adc of 9.7 ppm, and Cf and C, experienced upfield changes of shifts of - 1.2 and - 0.5 ppm. Addition of NaOD to the betaine produced large upfield changes for C, and C , (- 10.1 and - 7.8 ppm) which might be expected in view of their close proximity to the N+ centre of the betaine. Thus the effects of acid and base are considerable when appreciable amounts are added.The infinitesimally small amount of acidity present in the CDCl, (see Experimental section) is unlikely to have affected the 13C chemical shifts in this solvent.J. G. Dawber and R. A . Williams CDClj ['H6 ]acetone 3111 Fig. 8. Preferential solvation profile in the CDC1,-[2H,]acetone mixture. @ [ *H6 lacetone @ [ 2 ~ , ~ ~ ~ ~ ~ Fig. 9. Preferential solvation profile in the [2H,]acetone-[2H,]DMS0 mixture. For the 50: 50 mol % mixtures of organic solvents it is possible to test for preferential solvation by calculating an excess chemical shift, 6,, defined as: 6 E = 6,-0.5(6,+6,) (1) where 6, and 6, are the resonances of a given carbon atom in the pure solvents and 6, is the corresponding resonance in the 50: 50 solvent mixture.The values of A are given in table 4 where, for the systems involving CDCI,, a negative value for dE corresponds to a given line in the solvent mixture being closer to its value in pure CDCI,, (i.e. preferential solvation by CDCI,), whereas a positive value of 6, corresponds to31 12 Solvatochromism of a Pyridinium Retaine preferential solvation at that position of the molecule by the other solvent component. For the DMSO-acetone solvent mixture a negative value of B E [from eqn (l)] corresponds to preferential solvation at a given position by the DMSO. From these data graphs similar to fig. 5 were plotted, namely, 6, against carbon atom position, and from such graphs were constructed the solvation profiles of the betaine in the solvent mixtures; these are shown in fig.6-9. Clearly this approach is a simplification, but it does illustrate the possibility of preferential solvation, For the organic solvent mixtures involving CDC1, (fig. 4-8) it appears that preferential solvation by the other more polar component occurs within the vicinity of the CO- centre, whereas the N+ centre seems to prefer solvation by the CDCl,. For the DMSO-acetone mixture (fig. 9) the CO- centre appears to favour acetone, whereas the N+ centre prefers solvation by DMSO in the solvent mixture. The latter finding is similar to the result for the DMSO-D,O solvent mixture. Conclusion From the lH and 13C n.m.r. studies of (I) in various solvent systems it can be seen that, owing to solvation phenomena, the polarity of the medium influences the positions of the n.m.r.signals. The sites in the molecule most influenced by the solvent are those centres nearest to the positive and negative charges within the molecule, and this is seen in the n.m.r. chemical shifts. The resulting differences in electron density near the ionic sites, produced by solvation differences, are transmitted to more distant sites of the molecule by the extensively delocalised n-electron system of this remarkable solvato- chromic compound. By means of an excess chemical shift, B E , it is possible to obtain preferential solvation profiles in mixed solvent systems, although this approach is likely to be a simplified concept. The helpful comments of a referee and also Dr J. W. Akitt, University of Leeds, are gratefully acknowledged.References 1 K. Dimroth and C. Reichardt, Palette No. ZI (Sandoz AG, Basel, Switzerland); C . Reichardt, Angew. Chem.. Int. Ed. Engl. 1965, 4, 29; C . Reichardt, Solvent Efsects in Organic Chemistry (Verlag Chemie, Weinheim, 1979). 2 K. Dimroth, C. Reichardt, T. Siepmann and F. Bohlmann, Liebigs Ann. Chem., 1963, 661, 1. 3 A. Botrel, A. Le Benze, P. Jacques and H. Strub, J . Chem. Soc., Faraday Trans. 2, 1984, 80, 1235. 4 K. Dimroth, G. Arnoldy, S. von Eichen and G. Schiffler, Liebigs Ann. Chem., 1957, 604, 221. 5 K. Dimroth, Angew. Chem., 1960,72, 331. 6 H. B. Bull, C. A. Soch and G . Oenslager, J . Am. Chem. Sac., 1900, 24, 1. 7 J. W. Akitt, N.M.R. and Chemistry (Chapman and Hall, London, 2nd edn, 1983). 8 H. G. Benson and J. N. Murrel, J. Chem. Soc., Faraday Trans. 2, 1972, 68, 137. 9 R. L. Lichter and J. D. Roberts, J . Chem. Phys., 1970, 74, 912. 10 R. G. Wilson, J. H. Bowie and D. H. Williams, Tetrahedron, 1968, 24, 1407. 1 1 See for example J. G. Dawber, J . Chem. Soc., Faraday Trans. I , 1984,80,2133; 1978,74, 1702; 1709; 1979, 75, 370, and references therein. I2 H. Gunther, N.M.R. Spectroscopy (Wiley, Chichester, 1980). 13 F. W. Wehrli and T. Wirthlin, Interpretation of Carbon-I3 N.M.R. Spectra (Wiley, Chichester, 1983). 14 J. D. Memory and N. K. Wilson, N.M.R. of Aromatic Compounds (Wiley, Chichester, 1982). 15 G. L. Nelson, G. C . Levy and J. D. Cargioli, J. Am. Chem. SOC., 1972, 94, 3090; G. C. Levy, R. L. Lichter and G. L. Nelson, Carbon-I3 Nuclear Magnetic Resonance Spectroscopy (Wiley, New York, 2nd edn, 1980). 16 G. C. Levy, J. D. Cargioli and F. A. L. Anet, J . Am. Chem. SOC., 1973,95, 1527. 17 R. Radeglia, W. Sieffert, G . Hohlweicher, C. Jutz and H. L. Springer, Tetrahedron Lert., 1966,41,5053, R. Radegli, G. Engelhardt, E. Lipmaa, T. Pelk, K. D. Nolte and S . Dahne, Org. M a p . Reson., 1972, 4, 571 ; S. Dahne and K. D. Nolte, J . Chem. Soc., Chem. Commun., 1972, 1056. 18 E. Breitmaier and K. H. Spohn, Tetrahedron, 1973, 29, 1145. 19 H. Langhals, Angew. Chem., Int. Ed. Engl., 1982, 21, 724. Paper 512030; Receiced 18th Nocember, 1985