J . Chem. SOC., Faraday Trans. 1, 1985,81, 311-319 35Cl Nuclear Quadrupole Resonance Studies of Hydrogen Bonding in Solid Complexes of Chlorobenzoic Acids with Amines BY EUGENIUSZ GRECH,~ JERZY KALENIK AND LUCJAN SOBCZYK* Institute of Chemistry, University of Wroclaw, 50-383 Wroclaw, Poland Received 9th February, 1984 35Cl nuclear quadrupole resonance studies of hydrogen-bonded adducts containing o-, rn- and p-chlorobenzoic acid and 2,6-dichlorobenzoic acid have been carried out. Average frequencies are correlated with ApK, values in terms of the proton-transfer model. The results obtained for various proton donors are compared and discussed taking into account both intra- and inter-molecular effects. Numerous spectroscopic studies of the complexes of chlorobenzoic acids with amines carried out in solution1* have indicated the existence of a proton-transfer tautomeric equilibrium.Dipole-moment measurements3 in non-polar solvents have shown that in these systems an inversion region for ApK, [defined by pK,(BH+)- pK,(AH)], corresponding to a stepwise increase in the hydrogen-bond polarity, Ap, occurs and can be interpreted in terms of a shift of the proton-transfer equilibrium The proton-transfer equilibrium constant, KpT, may be related, according to the Huyskens and Zeegers-Huyskens to ApK, via the equation where and C' are constants. The constant ( expresses the coupling of hydrogen-bonded (HB) and proton-transfer (PT) states, whereas the constant C' depends on the environment and is related to the value of ApK, for which KPT = 1 (the inversion point).Infrared investigations of crystalline complexes of benzoic acids with pyridines5 have revealed the presence of such an inversion region in the solid state also. The purpose of this work was to investigate the influence of changes in the charge-density distribution of hydrogen bonds for crystalline complexes of chlorobenzoic acids with nitrogen bases of various strengths using the 35Cl n.q.r. technique. From earlier investigations on solid complexes with trichloroacetic pentachlorophenoP and 2,6-dichloro-4-nitropheno19 it follows that the complexation process is accompanied by a decrease in the average n.q.r. frequencies in relation to the average frequency of the pure proton donor. This means that the n.q.r. frequencies of nuclei in chlorine atoms which do not participate directly in hydrogen-bond formation can be applied successfully as a probe in order to examine the changes in charge-density distribution which take place in hydrogen bonds.t Permanent address : Institute of Fundamental Chemistry, Technical University, 7 1-065 Szczecin, Poland. 31 1312 35c1 N.Q.R. STUDIES OF HYDROGEN BONDING Table 1. 35Cl quadrupole resonance frequencies and pK, values of bases for the 2,6-dichlorobenzoates no. of reson- ance PK," no. amine v*.q. r . /MHz lines v/MHz (amine) a b 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2,6-dichlorobenzoic acid potassium 3-cyanopyridine 4-cyanopyridine 2,6-dichlorobenzoate 4-form ylp yridine aniline quinoline isoquinoline 4-methylquinoline 2-methylp yridine 4-met hylp yridine 3 ,Cdimethylpyridine 4-methylmorpholine morpholine triethylenediamine 4-dimethylaminop yridine tributylamine ethylpiperidine trie t hy lamine piperidine dibutylamineC 35.622; 35.683; 36.01 8 ; 36.067 34.632; 34.812 35.508; 35.682 35.542; 35.662; 35.730; 35.785 35.345; 35.821 35.388; 35.774 34.971 ; 35.788 34.885; 35.093; 35.554; 36.058 35.259; 35.768 35.1 15; 35.331 34.981 ; 35.245 35.075 34.483; 35.401 34.694; 35.266 34.888; 35.004 34.862; 35.305 34.893 34.828; 35.109 34.669; 34.947 35.023 35.218; 35.274 4 2 2 4 2 2 2 4 2 2 2 1 2 2 2 2 1 2 2 1 2 35.85 34.72 35.60 35.68 35.58 35.58 35.38 35.40 35.51 35.22 35.1 1 35.08 34.94 34.98 34.95 35.08 34.89 34.97 34.81 35.02 35.25 1.59b - 1.35 1.86 4.53 4.6 1 4.93 5.40 5.59 5.94 6.03 6.48 7.38 8.49 8.82 9.61 9.93 10.45 10.75 11.20 11.25 Ref.(10). Ref. (1 1). This complex has not been used in further calculations. EXPERIMENTAL Solid complexes were prepared by crystallization from equimolar chlorobenzoic acid + amine solutions in acetonitrile. The complex compositions were determined from chlorine analysis. Measurements of the resonance frequencies were performed at liquid-nitrogen temperature (77 K) on an ISSh-1-13M pulse spectrometer. Since not all the complexes with a given chlorobenzoic acid exhibited the same number of resonance lines, the average resonance frequencies v ~ . ~ . ~ . , being arithmetic means of the all frequencies measured, were used in the correlations. RESULTS AND DISCUSSION The resonance frequencies of the 1 : 1 complexes with 2,6-dichlorobenzoic acid and with 0-, m- and p-chlorobenzoic acid, together with the frequencies for the pure acids and their salts (mainly those of potassium) are summarized in tables 1-4.Also given are the average frequencies v ~ . ~ . ~ . and the pK, values for amines as components of the hydrogen-bonded adducts investigated.E. GRECH, J. KALENIK AND L. SOBCZYK 313 Table 2. 35Cl quadrupole resonance frequencies and pK, values of bases for the o-chlorobenzoates no. amine no. of reson- ance PKa" vn. 9. r. /MHz lines o/MHz (armne) a b 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 o-chlorobenzoic acid ammonium o-chlorobenzoate" 3-c yanopyridine 4-c y anopyridine 4-acetylpyridine quinoline isoquinoline 4-met hylquinoline 3-rnethylisoquinoline 4-methylpyridine 3,4-dirnethylpyridine 2,6-dimethylpyridine 2-amino-4-methylpyridine triethylenediamine 4-aminop yridine 4-amino-2-met h y lquinolined 4-dimethylaminopyridine piperidine dibutylaminee 36.312 35.05; 35.18; 35.27 35.983; 361060 35.825 35.634; 36.114 35.587; 36.312 35.484 35.639; 35.998 35.876 35.697 35.675 35.505; 35.996 35.453 34.89 1 34.344 35.797 34.579 34.599; 34.719 34.59 1 3 2 1 2 2 1 2 1 1 1 2 1 1 1 1 1 2 1 36.31 35.17 36.02 35.83 35.87 35.95 35.48 35.82 35.88 35.70 35.68 35.75 35.45 34.89 34.34 35.80 34.58 34.66 34.59 2.92b - 1.35 1.86 3.51 4.93 5.40 5.59 5.64 6.03 6.48 6.64 7.38 8.82 9.12 9.42 9.61 11.20 11.25 a Ref.(10). of f 10 kHz. Ref. (12). Ref. (13). This complex has not been used in further caclulations. Since the resonance line was very broad the n.q.r. frequency was measured with an accuracy The dependence of v ~ ~ ~ .~ . on ApK, for 2,6-dichlorobenzoates (fig. 1) and for o-chlorobenzoates (fig. 2) may be interpreted in terms of a formal proton-transfer model (as previously*? g, such that Vn.q.r. = XHB VHB +XPT VPT (2) where xHB and xPT are the mole fractions of the complexes with no proton transfer and with full proton transfer, while vHB and v,, are average resonance frequencies of two boundary forms of the hydrogen bonding. (This means that the observed average resonance frequency v ~ . ~ . ~ . is a linear function of the proton-transfer fraction.) From the v ~ . ~ . ~ . and pK, values gathered in table 1 (for the 2,6-dichlorobenzoates) and in table 2 (for the o-chlorobenzoates), the parameters in eqn (1) and (2) have been estimated using the generalized least-squares method. Thus the values obtained for <, C', ijHB and vPT are 0.92, -3.86, 35.62 MHz and 34.95 MHz for the 2,6- dichlorobenzoates and 0.82, -4.00, 35.83 MHz and 34.95 MHz for the o-chlorobenzoates.The experimental points and the plots of v ~ . ~ . ~ . against ApK, calculated for the estimated values of c, C', vHB and vPT are presented in fig. 1 (2,6-dichlorobenzoates) and 2 (o-chlorobenzoates). The average frequencies of the pure acids and their salts are also shown. The scattering of experimental points visible in fig. 1 and 2 is presumably314 35c1 N.Q.R. STUDIES OF HYDROGEN BONDING Table 3. 35Cl quadrupole resonance frequencies and pKa values of bases for the rn-chlorobenzoates no. no. of reson- ance PKaa amine vn .q .r . /MHz lines Y/MHz (amine) a b rn-chlorobenzoic acid potassium rn-chlorobenzoate" 4-cyanopyridine quinoline isoquinoline 4-methy lpyridine 3,4-dimethylpyridine 2-aminopyridine 2-amino-4-methylpyridine morpholine piperidine 35.232 34.74 34.710 34.242 34.607 34.767 34.985 34.589 35.116 34.61 7 34.51 3; 1 1 1 1 1 1 1 1 1 1 34.823 2 35.23 34.74 34.71 34.24 34.61 34.77 34.99 34.59 35.12 34.62 34.67 3.83b - 1.86 4.93 5.40 6.03 6.48 6.7 1 7.38 8.49 11.20 a Ref. (10). Ref. (12). " Ref. (13). Table 4. 35Cl quadrupole resonance frequencies and pK, values of bases for the p-chlorobenzoates no. amine no. of reson- ance PKaa vn. q. r. /MHz lines v/MHz (amine) a p-chlorobenzoic acid 34.673 b potassium 35.48 1 2-amino-4-methylpyridine 34.758 2 4-amino-2-met hylquinoline 34.302 3 piperidine 34.571 p-chlorobenzoate" 1 34.67 3.98b 1 35.48 - 1 34.76 7.38 1 34.30 9.42 1 34.57 11.20 a Ref.(10). Ref. (12). Ref. (13). due to contributions from lattice effects to the effective electric-field gradient on the quadrupole nuclei, which are not the same for all compounds. These contributions are dependent on the arrangement of molecules in the unit cell, which unfortunately is not known. The procedure of averaging the resonance frequencies allows us to eliminate some of the lattice effects. Some influence on the scattering of the experimental points may also be caused by differences in the conformations of the complexed chlorobenzoic acid molecules, these being characterized by a dihedral angle a between the planes of the carboxylic group and aromatic ring.The magnitude of the dihedral angle a is affected by two factors: (i) the interaction between the carboxylic group and substituents in ortho positions and (ii) the interaction between the neighbouring molecules (especially via hydrogen36.0 t 3 4.5- E. GRECH, J. KALENIK AND L. SOBCZYK -a 15 17 0 13 315 -b 0 2 4 6 8 10 A PK, 3 4 . 5 1 : : : : : : : : : : : : : ~ Fig. 1. Plot of pnn.s.r. against ApK, for the 2,6-dichlorobenzoates; for notation see table 1. bonds). The first of these factors determines the minimal value of a in the case when no specific interaction with the groups in ortho positions takes place. It can be easily calculated using the van der Waals radii.14 In the case of o-chlorobenzoic and 2,6-dichlorobenzoic acids the calculated values of a are 10 and 45", re~pective1y.l~ The second factor, connected with the arrangement of molecules in the crystalline lattice, leads to an additional enhancement of this angle.Thus in pure o-chlorobenzoic acid a is equal to 13.70,16 while e.g. in 2-chloro-5-nitrobenzoic acid it reaches 2 3 . O O . l ' Moreover, note that when there is only one substituent ortho to the carboxylic group the C(0)-OH group is arranged trans to that substituent.l6? l7 The phenomenon of a differentiation in the values of the dihedral angle a is particularly well illustrated for 2,6-dichlorobenzoates, where in some complexes both316 35c1 N.Q.R. STUDIES OF HYDROGEN BONDING chlorine atoms are equivalent (one resonance line is observed), which means that the carboxylic group is perpendicular to the plane of the aromatic ring.However, in other 2,6-dichlorobenzoic acid complexes both chlorine atoms are not equivalent and two or four resonance lines are visible. This clearly suggests that the carboxylic group is not perpendicular to the plane of the aromatic ring. The particularly large deviations visible for point (1 9) in fig. 1 and for point (14) in fig. 2 are probably due to either steric effects (4-amino-2-methylquinoline-o- chlorobenzoate) or the formation of additional hydrogen bonds (dibutylamine- 2,6-dichlorobenzoate). Note that the calculated values of the parameters 5 and C' are similar for both acids (this has also been observed for phenol^).^ The position of the inversion point determined from eqn (1) (where the degree of proton transfer is 50 % ) in the ApK, scale is 4.2 for the 2,6-dichlorobenzoates, while for the o-chlorobenzoates it is 4.9.The above values are close to the value of 3.75 estimated from the i.r. investigations of crystalline complexes of benzoic acids with pyridine~.~ The values of the parameters vHB and v,, calculated for the 2,6-dichlorobenzoates and the o-chlorobenzoates differ from those corresponding to average resonance frequencies of the pure acids and their salts. The differences between vHB and the average resonance frequencies of the pure acids are probably caused by the fact that molecules of the pure acids in an associated form (most frequently dimersls) participate in 0-H- - -0 hydrogen bonding, differing greatly from the formation of O-H...N bonds in complexes without proton transfer.On the other hand, the discrepancies between v,, and the average resonance frequencies of the salts mainly result from the fact that, unlike the hydrogen-bonded proton-transfer complexes 0-. --H-N+, in salts one must deal with electric fields arising from ions which are a source of additional resonance-frequency shifts.lg The magnitudes and directions of these shifts depend on the arrangement of ions in the unit cell. Therefore the method of determining the degree of proton transfer in crystalline hydrogen-bonded complexes suggested by Chihara and NakamuraZ0 and based on the assumption that vHB is equal to the average resonance frequency of the pure acid and v,, is equal to the average resonance frequency of its salt, could lead to incorrect results.In the case where the Cl atoms are sufficiently far from the hydrogen bond the shifts in the average resonance frequencies caused by changes in the hydrogen-bond polarity are small in comparison with the shifts caused by the lattice effects, and no correlation between v ~ . ~ . ~ . and ApK, is observed. This is clearly visible in the m-chlorobenzoates, the results for which are shown in fig. 3. The scatter of the experimental points in fig. 3, more distinct than those in fig. 1 and 2, may be caused additionally by two possible conformations of the carboxylic group in relation to the chlorine atom in the meta position. Similarly no correlation can be expected forp-chlorobenzoic acid as a proton donor (see table 4). The resonance frequency of purep-chlorobenzoic acid is even lower than that of its potassium salt, which can be explained by the presence of electric fields arising from the ions19 which, when the difference between vHB and vPT is not high, can increase the resonance frequency of the salt to a value above that of the pure acid.Lynch et aZ.13 have tried to explain this anomaly in terms of the mesomeric effect in p-chlorobenzoic acid. If this were the case, the average resonance frequencies of complexes containing the ionic form of the hydrogen bond, where the carboxylic group is almost entirely ionized, should be closer to the resonance frequency of potassium p-chlorobenzoate than to that of the pure acid. It thus seems that the role of the mesomeric effect in p-chlorobenzoic acid has been overestimated by Lynch et aZ.13 Let us now try to compare the results of n.q.r.investigations for five series of theE. GRECH, J. KALENIK AND L. SOBCZYK 35.25 35.00- 5 5 34.75- I> 3 4.50 34.25.- 317 .- .- 0 1 7 0 5 0 0 4 0 0 8 3 6 2 0 -a -b 0 9 -2 0 2 4 6 8 34.00 A PKa Fig. 3. Plot of v ~ . ~ . ~ . against ApK, for the rn-chlorobenzoates; for notation see table 3. A PKa Fig. 4. Calculated xpT values for various series of hydrogen-bonded adducts plotted against ApK,; for notation see table 5. complexes containing trichloroacetic acid,’ pentachlorophenol,g 2,6-dichloro-4- nitr~phenol,~ 2,6-dichlorobenzoic acid (this work) and o-chlorobenzoic acid (this work). We describe the behaviour of all the complexes using a generalized scheme of the proton-transfer equilibrium model, although we have already stressed’ that a quantitative description of the resonance-frequency changes can also be expressed in terms of the Mulliken theory by using a and b coefficients, which are in turn connected to the contribution of non-polar and dative states.21 In fig.4 are presented plots expressing the dependence of xpT on ApK, for the series of complexes under consideration. The parameters of eqn (1) describing these curves318 35c1 N.Q.R. STUDIES OF HYDROGEN BONDING Table 5. Values of < and C' in eqn ( 1 ) and ApK, values for which KPT = 1 for a series of complexes with various proton donors no. proton donor 1 trichloroacetic acid" 2 2,6-dichloro-4-nitrophenolb 3 pentachlorophenolc 4 2,6-dichlorobenzoic acidd 5 o-chlorobenzoic acidd 0.12 0.04 -0.4 0.75 - 0.62 0.8 0.76 -0.91 1.2 0.92 - 3.86 4.2 0.82 - 4.00 4.9 a Values of the parameters calculated from data taken from ref.(7). This work. Ref. (9). Ref. (8). are listed in table 5. The ApK, values at which 50% proton transfer takes place are also included. We note an unusual broadening (a small value of r) of the curve for the complexes of trichloroacetic acid. A similar broadening has been found when describing the dipole-moment dependences for complexes of carboxylic acids with trieth~lamine.~ Undoubtedly only the chemical properties of aliphatic carboxylic acids are of importance here since complexes of aromatic acids behave normally, a fact confirmed by the results obtained in this work. Attempts to explain this anomaly may seem premature since we have no information about the configuration of the complexes formed in the crystalline state.On the other hand, it seems easier to explain the large difference in the position of the inversion point (at which xPT z 1/2) for the phenol derivatives and benzoic acids. The essential factor here is the additional contribution of the mesomeric effect in phenol complexes, which does not play a prominent role in benzoic acid complexes. In conclusion, measurements of the n.q.r. frequencies of hydrogen-bonded complexes can serve as an useful indicator of the inversion region of donor-acceptor properties, where a stepwise charge rearrangement and proton-transfer equilibrium can be anticipated. Knowledge of this region seems to be important since hydrogen-bonded complexes from this region are characterized by an unusual potential for proton motion.We thank a referee for helpful remarks. 1 2 3 4 5 6 7 8 9 10 11 12 G. M. Barrow, J. Am. Chem. Soc., 1956, 78, 5802. J. Nasielski and E. Vander Donckt, Spectrochim. Acta, 1963, 19, 1989. L. Sobczyk and Z. Pawelka, J . Chem. Soc., Faraday Trans. I , 1974,70, 832. P. Huyskens and Th. Zeegers-Huyskens, J . Chim. Phys., 1964,61, 81. S. L. Johnson and K. A. Rumon, J . Phys. Chem., 1965,69, 74. 0. Kh. Poleshchuk, Yu. K. Maksyutin, 0. F. Sychev, K. K. Koshelev and I. G. Orlov, Zzuest. Akad. Nauk SSSR, Ser. Khim., 1975, 6, 143 1 . J. Pietrzak, B. Nogaj, Z. Dega-Szafran and M. Szafran, Acta Phys. Polon. A, 1977, 52, 779. E. Grech, J. Kalenik and L. Sobczyk, J. Chem. SOC., Faraday Trans. I , 1979, 75, 1587. E. Grech, J. Kalenik, Z. Malarski and L. Sobczyk, J. Chern. Soc., Faraday Trans. I , 1983, 79, 2005. D. D. Perrin, Dissociation Constants of Organic Bases in Aqueous Solution (Butterworths, London, 1965; supplement, 1972). J. F. Dippy, S. R. C. Hughes and L. G. Bray, J . Chem. SOC., 1959, 1717. G. Kortiim, W. Vogel and K. Andrussow, Dissociation Constants of Organic Acids in Aqueous Solution (Butterworths, London, 1961).E. GRECH, J. KALENIK AND L. SOBCZYK 319 l 3 R. J. Lynch, T. C. Waddington, T. A. OShea and J. A. S. Smith, J. Chem. SOC., Faraday Trans. 14' H. A. Staab, Einfiihrung in die theoretische organische Chemie (Verlag Chemie, Weinheim, 1959). l5 M. K. Chantooni Jr and I. M. Kolthoff, J . Phys. Chem., 1974,78, 839. l 6 G. Ferguson and G. A. Sim, Acta Crystallogr., 1961, 14, 1262. l7 G. Ferguson and G. A. Sim, J. Chem. SOC., 1962, 1767. L. Leiserowitz, Acta Crystallogr., Sect. B, 1976, 32, 775. lo C. H. Townes and B. P. Dailey, J. Chem. Phys.,. 1952, 20, 35. 2o H. Chihara and N. Nakamura, Bull. Chem. Soc. Jpn, 1971, 44, 1980. 21 R. S. Mulliken, J. Am. Chem. SOC., 1952, 74, 81 1. 1976,72, 1980. 2, (PAPER 4/235)