J. Chem. SOC., Faraday Trans. I , 1988, 84(1), 41-56 A Polarimetric and llB and 13C Nuclear Magnetic Resonance Study of the Reaction of the Tetrahydroxyborate Ion with Polyols and Carbohydrates J. Graham Dawber," Stuart I. E. Green and (in part) John C. Dawber and Sundus Gabrail Department of Chemistry and Biology, North Staffordshire Polytechnic, Stoke-on-Trent ST4 2DE The reaction of the tetrahydroxyborate ion, B(OH);, with 31 polyols and carbohydrates in aqueous solution has been studied by complementary investigations involving polarimetry and llB and 13C n.m.r. spectroscopy. The calculation of equilibrium constants for the complexation of the carbohydrates with B(0H); from the polarimetry results, using a previously derived equation, was only partially successful owing to the presence of multiple equilibria in a number of cases.The presence of several complexed species in these reactions is demonstrated by llB n.m.r. studies. A calibration method devised to relate peak area in llB n.m.r. spectra with concentration has been used to calculate equilibrium constants for the various equilibria present, and an attempt has been made to rationalise the equilibrium constants with the molecular structures of the substrates. In the majority of cases the 13C n.m.r. spectra confirm the polarimetric and llB n.m.r. studies, and in most cases allow a more specific identification of the reaction sites in the polyols/carbohydrates to be made. The use of polyhydroxy compounds as a means of increasing the strength of boric acid for its titration has been known for many years and the reaction has also been used to characterise carbohydrates.However, the stoichiometry of the complexes formed has often been uncertain.1-6 It is known that the reaction with polyols is much more pronounced when the tetrahydroxyborate ion, B(OH),, is used instead of boric acid itself.'* While the complexation of borate across adjacent hydroxy groups in the polyol is commonly assumed, it is now evident that complexation across alternate hydroxy groups is also possible, as demonstrated by 'H n.m.r. spectros~opy,~ "B n.m.r. spectroscopy, lo circular dichroismll and polarimetry . l1 At high ratios of [polyol]/ [borate] it is possible to observe two molecules of polyol to one molecule of boratelO in the complex. In the case of cyclic myo-(meso) inositol it is thought that three hydroxy groups on alternate carbon atoms can become involved in a tridentate complex with the B(0H); ion.12 While our previous ''B n.m.r.work was in progress, unknown to us, Kieboom and coworker^^^^ l4 were carrying out similar studies, and the two pieces of work agree very well, although their results were published while ours were still in press." The work of Kieboom and c o ~ o r k e r s ~ ~ ~ l4 has since taken a different direction from ours in that they have carried out an in-depth study of polyol carboxylate salts and their reactions with borate. Our work, on the other hand, has included a study of a wider range of non-ionic polyols and also a range of monosaccharides and two disaccharides, since our interest was in the reaction of borate with un-ionised species.The techniques used have been polarimetry, "B n.m.r. and l3C n.m.r., and these results are reported here. We have modified our method of calculating the equilibrium constants of complexation from the 4142 Polarimetry and "B and 13C N.M.R. "B n.m.r. studies and the recalculated values of our previous worklo are also presented. Experimental Materials The materials used were all of GPR quality (except glucose and sucrose which were AR), and were used without further purification. The materials were as follows (structures are given in table 1) : ethane- 1,2-diol (l), propane- 1,2-diol(2), propane- 1,3-diol(3), glycerol (4), butane- 1,2,4-triol (5), meso-erythritol (6), threitol (7), xylitol (8), D-arabitol (9), adonitol (ribitol) (lo), dulcitol (galacticol) (1 l), mannitol (12), sorbitol (13), cis- cyclohexane- 1,2-diol (14), trans-cyclohexane- 1,2-diol (15), cyclohexane- 1,3-diol (16), cyclohexane-r- 1 ,c-3,c-5-triol (17), myo(rneso)-inositol (18), dihydroxydioxane (19), sodium ascorbate (20), a-methyl glucoside (21), D( + )-glucose (22), D( +)-galactose (23), D( +)-mannose (U), L( +)-rhamnose (25), D( +)-xylose (26), D( -)-arabinose (27), L( +)-arabinose (28), fructose (29), sucrose (30) and maltose (31).The NaB(OH), solution (stock solution, 2.5 mol dm-3) was prepared by direct reaction of A.R. NaOH and H,B03, the pH was then adjusted to 12.5, and this solution was diluted as required. Preparation of Borate of Cyclic Polyols (a) cis-Cyclohexane- 1,2-diol Borate Equimolar quantities of the diol and NaB(OH), were mixed in solution and heated to ca.70 "C. Ethanol was then added carefully until the solution was just cloudy. The solution was next heated until it was clear again and then allowed to cool very slowly and left to stand for at least 24 h. White, well formed shiny leaflets of the borate crystallised out (decomp. above 300 "C after loss of water at lower temperatures; analysis gave C, 29.8 YO ; H, 7.6 YO ; the trihydrate C,H,,O,BNa requires C, 30.5 YO ; H, 7.7 %). (b) Cyclohexane-r-1 p3,c-j- triol Borate The borate of this triol was prepared as for (a). Well formed lustrous flakes were obtained (m.p. > 340 " c ; analysis gave c , 30.8 YO ; H, 6.9 % ; the trihydrate C,H,,O,BNa requires C, 30.8 YO ; H, 6.8 %).(c) myo-Inositol Borate This was prepared as for (a), but on cooling, the complex initially separated from the solution as an oil, which slowly solidified to a crystalline solid. This product appears to retain more water of crystallisation than the rather better formed crystals in (a) and (b), but of course contains many more hydrophilic centres (decomp. from 320 "C after losing water at much lower temperatures; analysis gave C, 20.5 % ; H, 6.7 YO ; the heptahydrate C,H,,O,,BNa requires C, 20.4 YO ; H, 6.8 YO). Polarimetric Measurements with the Carbohydrates Solutions for the polarimetric measurements were made up containing 0.25 mol dm-3 of carbohydrate and the concentration of NaB(OH), varied from 0 to 1.25 mol dm-3. The optical rotations of the solutions were measured with a Bellingham and Stanley model A photoelectric polarimeter using a 100 mm pathlength tube.Angular rotations couldJ . G. Dawber, S. I. E. Green, J . C. Dawber and S. Gabrail 43 Table 1. Structures of compounds studied HO HO HO WOH HO Ho* * HO OH HO HO OH OH OH HO (9) Horn &OH HO HO HO OH HO OH OH OH (14) OH (17) OH (15) OH 6.;.;, OHHo 0 Na OH (19) \ OH44 Polarimetry and "I3 and I3C N.M.R. Table 1. Structures of compounds studied (con?.) OCH3 (21) CH,OH HovTj-ti - - bo (OH,CH,OH ) OH (OH , CHZOH) OH (29) a Labelling system for carbohydrate carbon atoms. be estimated to 0.002". The measurements were carried out at a wavelength of 435.8 nm using a low-pressure Hg lamp with the unwanted wavelengths filtered out by a solution of NaNO, and a cobalt blue filter.15 This wavelength was used rather than the sodium D-line since it provided larger differences in optical rotation between successive solutions, thereby giving better discrimination. The temperature was 20 1 "C.J.G. Dawber, S. I. E. Green, J. C. Dawber and S. Gabrail 45 N.M.R. Measurements (a) "B N.M.R. The llB n.m.r. spectra were measured using a Jeol FX-90Q Fourier-transform spectrometer (lH resonance at 89.55 MHz, llB resonance at 28.75 MHz) using a tip angle of 45" and a pulse repetition time of 2 s. The spectral responses over 4500 Hz were acquired into 8 K data points and zero-filled to 16 K data points and an experimental broadening of 0.3 Hz applied prior to Fourier transformation. All solutions were made up in water with the spectrometer locked onto a D20 capillary.Referencing of the "B chemical shifts was made relative to H,BO, in water as 6 = 0.0 ppm. The concentration of NaB(OH), used was 0.25 mol dm-, and the concentrations of polyol/carbohydrate ranged from 0 to 1.0 mol dm-3. When the B(OH), ion complexes with polyols the llB n.m.r. signal appears at characteristic positions for the various types of 11* 1 3 7 l4 The formation of complexes involving vicinal hydroxy groups is accompanied by downfield shifts (for 1 : 1 and 1 : 2 complexes), and accompanied by upfield shifts for complexes involving hydroxy groups on alternate carbon atoms. The calculation of equilibrium constants of formation for the various complexes has previously been based upon the integration of the areas of the various peaks in the llB n.m.r.spectra. Implicit assumptions in this approach are (i) that the area of a given llB peak is directly proportional to the concentration of that boron-containing species, and (ii) that the same relationship holds between peak area and concentration for all the different boron species. The half-widths (A$ of the various llB peaks in this study ranged between 4 and 60 Hz, and, if one approximates the effective quadrupolar relaxation time as (zAv$',l6 the relaxation times of the various boron species should lie within the pulse repetition time of 2 s. Nevertheless, it was decided to check the linearity of the peak area with concentration for the B(0H)i species in two series of experiments in which the n.m.r. spectra of various concentrations of NaB(OH), were measured simultaneously with a fixed concentration of boron trifluoride etherate in a capillary in the n.m.r.tube. In one series neat boron trifluoride etherate was used, whereas in the other series a solution of the etherate in CHC1, (ca. 2 mol drn-,) was placed in the capillary. The area of the B(OH), ion peak (the borate solution containing ca. 10% D20 was placed in the main n.m.r. tube) was measured relative to that of the boron trifluoride etherate. This was done for a wide range of borate concentrations. Fig. 1 shows plots of the ratio of the areas of the two peaks as a function of B(OH), concentration (the etherate being effectively an analytical internal standard) and it can be seen that there is good linearity, indicating that the peak area is proportional to boron concentration.There is still, however, the question of the relationship between the different peaks for the various borate species, namely, B(OH),, the 1 : 1 complexes (BP-) and the 1 :2 complexes (Bpi). This feature is important since in the normal experimental runs the proportion of each boron-containing species is taken as a fraction of the total boron concentration as calculated from its percentage of the total area of the llB n.m.r. signals. A further two series of experiments were performed using the boron trifluoride etherate capillary as the fixed concentration standard. In one set of experiments various proportions of sorbitol were added to 0.25 mol dm-, NaB(OH), in order to compare the relative sizes of the peaks from B(OH),, BP- and BP, (P = polyol).In a second set of experiments xylitol was used as the polyol. For the sorbitol experiment it was found that the relative peak area sizes for the same boron concentration for the species BP,, BP- and B(0H); were in the ratio 1.00:0.79:0.68, and for the corresponding xylitol experiment the values were 1 .OO : 0.76 : 0.62, which was considered to be in reasonable agreement. The data from the sorbitol experiment were chosen for the purpose of correcting the relative peak areas for the calculations of the equilibrium constants of the various equilibria (see later).46 Polarimetry and "B and 13C N.M.R. [ NaB(OH)4]/mol dm-3 Fig. 1. B(0H); calibration: 0, neat etherate standard; x , diluted etherate standard. (b) 13C N.M.R.The proton-decoupled 13C n.m.r. spectra were measured on a Jeol FX90Q spectrometer (13C resonance at 22.49 MHz) using a tip angle of 30" and a pulse repetition time of 1 s. The spectral responses over 4500 Hz were acquired in 8 K and zero-filled to 16 K data points and an experimental broadening of 0.7 Hz was applied prior to Fourier- transformation. The solutions were all made up in H,O and the instrument was locked onto a D,O capillary in the n.m.r. tube. Referencing was relative to TSP at 6 = 0.0 ppm. The spectrum of each polyol or carbohydrate was measured at 1 mol dm-3 in H20 and 1 mol dmd3 in 2.5 mol dm-3 NaB(OH), solution. Results and Discussion Polarimetry The optical rotation results were converted to molar optical rotation, [Y], by E\y] = a/d, where a is the optical rotation in a polarimeter tube of 1 metres for a carbohydrate concentration of c mol m-3, giving units of O m2 mol-1 for [Y].The values of [Y] at 436 nm for the various carbohydrates in water are given in table 2. For each carbohydrate, when borate was added, the change in molar rotation, A["], from the value in water was calculated and these values are plotted as a function of NaB(OH), concentration in fig. 2-4. It can be seen that in all cases the reaction of the chiral carbohydrates with the B(OH), ion was evident from the changes in molar rotation. In most cases the variation of A w l changes simply with NaB(OH), concentration, but for L( +)-rhamnose and D( +)-xylose the behaviour can be seen to be more complicated In previous work8*l1 a method was devised for calculating the equilibrium constant, K,, for the formation of complexes of polyols using polarimetry results.Changes in the molar optical rotation are dependent upon the extent of complexation, the stoichiometry (fig- 4).J . G. Dawber, S. I. E. Green, J . C. Dawber and S. Gabrail 41 Table 2. Optical rotation results anomeric composition pyranose (YO) a /3 furanose (YO) ["I/' m2 molt1 K,/dm3 mol-I a-methyl glucoside D( + )-glucose D( + )-galactose D( + )-mannose L( + )-rhamnose D( + )-xylose D( - )-arabinose L( + )-arabinose fructose sucrose maltose sodium ascorbate 100 36 27 67 34 33 63 63 3 37 - - 0 64 73 33 66 67 34 34 90 67 - - 0.5954 0.1830 0.2753 0.05 12 0.0273 0.0567 0.2993 0.4379 0.90 18 0.439 1 -0.3046 - 0.3208 1.3 18.9, 2.9 a a a a a a a 0.7, 0.5 6.1, 0.9 4.7 a Unable to be evaluated from the polarimetric results with any great certainty, the results giving curves rather than the required straight lines.1 I I I I 1 2 3 4 (borate]/[ sugar] Fig. 2. Changes in molar optical rotation with added NaB(OH), + , a-methyl glucoside; x , D( +)- glucose ; A, D( + )-mannose ; 0, D( + )-galactose.48 Polarimetry and "B and 13C N.M.R. 300 200 100 4 I I i E O 2 e m ;= Q -100 -200 - 300 X I I I I 1 2 3 4 [borate]/ [ sugar] Fig. 3. Changes in molar optical rotation with added NaB(OH),: 0, D( -)-arabinose; x , L( +)- arabinose ; A, fructose; 0, sodium ascorbate. of the complex, and the molar rotation of the complex (or complexes) compared to that of the original polyol/carbohydrate. The equations derived assumed the formation of a 1 : 1 complex and were successfully applied to systems involving a number of anions with sorbitol and mannitol* and also a chiral diol.ll However, in the case of B(0H); and sorbitol and mannitol the results did suggest the possibility of more than one complexed species being present.In the determination of K, for the carbohydrates of the present work, by far the simplest behaviour was found for a-methyl glucoside. For the remainder of the carbohydrates the evaluation of K, proved to be less simple. Instead of a simple linear plot being obtained from the method,8 for some carbohydrates the graphs consisted of two linear portions leading to two values for K,. For a number of other cases (galactose, mannose, arabinose, xylose, rhamnose and fructose) K, could not be evaluated satisfactorily, and the reasons for this are discussed later.The K, values for the carbohydrates for which the above method of evaluation was successful are presented in table 2. Most of the carbohydrates studied in this work exist in solution as a mixture of anomers, e.g. the a- and B-pyranose, and the a- and 8-furanose structures. The initial polarimetry studies showed that whatever the anomeric form of the initial solid carbohydrate, the equilibrium anomeric mixture in solution was achieved very rapidly (< 30 s) on addition of a small amount of NaB(OH), solution. The proportions of theJ . G. Dawber, S. I. E. Green, J. C. Dawber and S. Gabrail 49 I I I I I 1 2 3 I [borate I/[ sugar] Fig. 4. Changes in molar optical rotation with added NaB(OH), 0, D(+)-xylose; A, L(+)- rhamnose; x , sucrose; 0, maltose.a- and 8-pyranose structures make up the majority of the anomeric mixture, with the furanoside structures accounting for only 1-2 % of the total composition1' (table 2). Hence we assume that complexation with the B(0H); ion involves principally the a- and 8-pyranoside structures. The structural difference between glucose itself and that of a-methyl glucoside is the availability for complexation of the OH group on the C , carbon atom in glucose along with its adjacent OH group on C,, hence one would expect a different and enhanced value of K, and this was found to be the case. A similar relationship exists for the disaccharides sucrose and maltose [ i e . a-glucopyranosyl fructose and 4-0-a-(D)- glucopyranosyl D-glucose], with the maltose complexing considerably more with B(OH), than with sucrose, a significant difference between the structures of the two sugars being the a- and 8-OH group on the C , carbon atom of maltose.In situations where complex formation can occur at many sites, including adjacent and alternate carbon atoms (see later), it is not surprising that the optical rotation results in some of the cases do not allow an evaluation of K, from a method based upon a simple model of complexation. * "B N.M.R. For the purposes of comparing the complexation of the B(0H); ion with the various polyols/carbohydrates we have considered three general reaction equilibria. Two of these represent complexation across adjacent carbon atoms and the other represents complexation across alternate carbon atoms (including the case of tridentate coordination of the B atom): K , B- + P f Bpidj.K2 BP,,. + P f B P i ~ j . K3 B- + P f Bpil,.50 Polarimetry and llB and 13C N.M.R. [B- represents B(OH),, P represents polyol/carbohydrate, and adj. and alt. represent reaction across adjacent and alternate carbon atoms, respectively]. With higher-field llB n.m.r. l4 it is possible to observe a more comprehensive range of equilibria involving various conformers, but these were only just discernible in our studies at lower field. In calculating the various equilibrium constants we used the weighting of the peak areas found in the calibration procedure with sorbitol. If the fractional areas of the total peak area fyr "B are a, p, y and 6 for the species B-, Bpidj., Bpiadj.and Bpil,., respectivelyJand the weighted peak area factors from the calibration study are a, b, c and d (d was assumed to be the same as b), and if m, is the total concentration of borate in all its forms and rn, is the total polyol concentration in all its forms, then and [PI = m,-(bp+2cy+d6)ml/D where D = (aa+bp+cy+dd). From the sorbitol calibration experiment a = 0.68, b = 0.79, c = 1.00 and d = 0.79, and hence the equilibrium constants K,, K, and K, can be calculated for the various polyol/carbohydrate - borate systems from the relative integrated peak areas (a, p, y, 6 ) in the "B n.m.r. spectra. For the purposes of comparison the values of K,, Kz and K3 were calculated at a total polyol concentration (m,) of 0.5 mol dm-3 and a total borate concentration (m,) of 0.25 mol dm-3, and the values obtained are presented in table 3.In the small area of duplication of other the trends in our values of the equilibrium constants are similar. For the systems where Kc could be evaluated from the polarimetry results the corresponding values of the equilibrium constants from the llB n.m.r. results agree at the same level of compatibility as previously found.lOvll It can be seen from table 3 that for the various linear polyols the magnitude of the values of the equilibrium constants are related to the number of OH groups available for complexation. In general the magnitudes are in the order K, > K, > K3, showing that complexation across adjacent OH groups is easier than that across alternate OH groups.We are of the opinion that for the linear polyols there is relatively ease of rotation of the CH,OH groups at the end of the chains, thus accounting for the low K, values for the lower members of the series. When one reaches the C, members, meso-erythritol and threitol (and to some extent the trio1 glycerol), there is a considerable increase in the value of K,, and it is for these compounds that there is relatively more restriction to movement of the OH groups in the 'inner' CHOH groups, and it is these which are likely to be involved in the complexation reaction with the B(OH), ion. For the C, polyols, xylitol and D-arabitol, there is a further increase in K,. In the case of xylitol there is some competition for complexation across alternate OH groups (K3 = 17) and inspection of its structure (table 1) shows that the OH groups on C, and C , are favourably situated for complexation. This is not the case for D-arabitol, which has no value for K3 but a larger value of K l .In the case of adonitol, K, (at 32) is the lowest for the C, polyols, but K3 is considerable (20). This enhanced value for K3 may be due to the fact that adonitol has a number of alternate OH groups favourably disposed with respect to each other, namely Cl:C3, C,:C4 and C3:C,. In these three polyols it is thought that the most favourable positions for complexation involve adjacent OH groups which are disposed to each other at an angle of 60". Of course the conformations shown in table 1 are not fixed and rotation about single C-C bonds willJ.G. Dawber, S. I. E. Green, J. C. Dawber and S. Gabrail Table 3. llB N.m.r. results for determination of equilibrium constants of complexation (dm3 mol-l) 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 ethane- 1,2-diol propane- 1,2-diol propane- 1,3-diol glycerol butane- 1,2,4-triol meso-erythri to1 threi to1 xylitol D-arabitol adonitol dulcitol mannitol so rbi to1 cyclohexane- 1,2-diol trans-cyclohexane- 1,2-diol cyclohexane- 1,3-diol cyclohexane-r- 1 ,c-3,c-5-triol myo-inositol 2,3-dihydroxydioxane sodium ascorbate a-methyl glucoside D( + )-glucose D( + )-galactose D( + )-mannose L( + )-rhamnose D( + )-xylose D( - )-arabinose L( + )-arabinose fructose sucrose maltose 3 .O 4.7 - 37 70 78 126 540 32 1000 232 540 6.2 1.4 - - - 4.4 3.7 2.3 525 17 113 I99 44 347 246 246 235 - 12.3 0.4 0.8 3.9 1.7 - 12 12 15 68 5 32 68 - 1.8 - - - 1.9 0.4 8 I5 6 3 31 32 26 111 58 - - 2.2 occur, but the structures are likely to represent the most energetically favoured conformations.For the C, linear polyols the magnitude of Kl is in the order mannitol < sorbitol < dulcitol. In the case of mannitol there is a favourable pair of OH groups on C, : C , ; in the case of sorbitol there are two pairs of potential OH groups (but one OH group is common), i.e. C, : C , and C,: C,; whereas in the case of dulcitol there are two separate pairs of OH groups, i.e. C, : C , and C, : C, (there was some evidence from the 13C n.m.r. data that C,:C, OH groups might also be involved).Hence the values of Kl are approximately compatible with the molecular structures. cis-Cyclohexane- 1,2-diol was found to form a complex with B(OH),, but for the trans isomer we found no evidence of reaction. The trans-1,2-diol in the neat liquid can form intramolecular hydrogen bonds between the adjacent OH groups when both are in equatorial positions. However, we feel that in aqueous solution the opposite chair configuration could be favoured, since in this conformation the two OH groups become axial, thus allowing more access for solvation by water molecules. In this conformation complexation of the trans isomer with B(OH), would be impossible. The cyclohexane-1,3-diol (mixture of cis and trans) showed no evidence of complexation with borate.By contrast, however, cyclohexane-r- 1 ,c-3,c-5-triol showed52 Polarimetry and "B and 13C N.M.R. (a 1 -14.1 -17.4 -18.5 -17.4 -13.6 -19.2 Fig. 5. llB n.m.r. spectra of polyol borate complexes in water: (a) cis-cyclohexane-l,2-diol, (b) cyclohexane-r- 1 ,c-3,c-5-triol, (c) myo-inositol. considerable reaction with B(0H); (K3 = 39), but in this case the reaction is thought to involve a tridentate complex with the b0r0n.l~ In the case of myo-inositol the largest value of equilibrium constant is for K3, and again this is likely to form a tridentate complex12 similar to cyclohexane-r-1 ,c-3,c-5-triol. Nevertheless, our "B n.m.r. results showed that some reaction also occurred across adjacent OH groups of the inositol. The complexes of cis-cyclohexane- 1,2-diol, cyclohexane-r- 1 ,c-3,c-5-triol and myo-inositol were isolated (see Experimental) and the llB n.m.r.spectra of the compounds dissolved in water are shown in fig. 5. Here it can be seen that the complexes are extensively dissociated into polyol and B(0H); (6 = - 17.4 ppm). In the case of myo-inositol the small amount of bidentate complex involving adjacent OH groups is also evident at a peak position of 6 = - 13.6 ppm. The extent of complexation of 2,3-dihydroxydioxane with B(OH), is very extensive compared to that of cis-cyclohexane- 1,2-diol. It is possible that the two OH groups are much more acidic in the dioxane compound, owing to the inductive effects of the ring oxygen atoms rather than any stereochemical/configurational influence. Sodium ascorbate, by contrast, is rather more like cyclohexane-1,2-diol in its extent of complexation with B(OH),, The K values for the carbohydrates (compounds 21-31) are shown in table 3 and it can be seen that complexation with borate is extensive in many cases.For four of the carbohydrates (glucose, xylose, mannose and maltose) the manner of complexation is comprehensive, involving the equilibria for K,, K, and K3. In the particular case of xylose its value of K3 is considerable (in addition to K , and K,) and this coincides with the complicated optical rotation behaviour observed on addition of borate. The values of K for the pairs methyl glucoside and glucose, and sucrose and maltose have a similar relationship, and this similarity was also observed in the optical rotation measurements.As discussed earlier, this relationship can be rationalised with theJ. G. Dawber, S. I. E. Green, J. C. Dawber and S. Gabrail 53 availability of an OH group on the C, carbon atom, giving rise to enhanced complexation. For mannose and glucose the configurations about C,, C,, C,, C, are the same, but that for galactose is different. It may be significant that for mannose and glucose the reaction with B(OH), can also involve alternate OH groups to a small extent, whereas in the case of galactose the reaction involves only the equilibria for Kl and K2. The cyclic pyranose structures of the pentoses xylose and the two isomeric arabinoses have similarities to the pyranoside form of fructose, and it is for these four compounds that the highest K values were observed amongst the carbohydrates as a group.Unfortunately there does not seem to be a prima facie case of a simple correlation between the individual molecular structures and their corresponding values of the equilibrium constants, K. 13C N.M.R. The results of the 13C n.m.r. spectra are given in table 4 [the complete table, including chemical shift assignments, has been deposited as supplementary publication no. SUP 56699 ( 5 pp)].t The spectral assignments of the compounds were made by utilising information from several sources. Our chemical shift values are slightly different from those in the literature since they are all relative to TSP at 6 = 0.0 ppm, but all the assignments are consistent with the other published data for the uncomplexed polyols/ carbohydrates. The 13C spectra of the compounds dissolved in 2.5 mol dm-, NaB(OH), (a borate : carbohydrate ratio of 2.5) invariably showed broadening of all the resonances, with the carbon atoms involved in the complexation being affected most. Detailed comments relating to the spectra and their interpretation are given in table 4.For the linear polyols (compounds 1-13) the earlier members (ethane diol, propane- 1,2-diol, propane- 1,3-diol, glycerol, butane- 1,2,4-triol) all show evidence of the terminal CH20H group(s) being involved in complexation. However, when the chain length is more extended (compounds 6-13), complexation with B(0H); favours the inner CHOH groups, which are likely to have less freedom of rotation than the terminal CH20H groups. This would account for the large increase in complexation as the chain-length of the polyol is increased (table 3).Adonitol was found by the llB n.m.r. studies to be involved considerably in complexation with OH groups on alternate C atoms, in addition to the reaction across adjacent carbon atoms, and this extra feature shows itself in the very complicated 13C n.m.r. spectrum of this compound in 2.5 rnoldm-, NaB(OH),. The cyclic polyols cis-cyclohexane- 1,2-diol and cyclohexane-r-1 ,c-3,c-5-triol have relatively simple 13C n.m.r. spectra in water and also as complexes with B(0H);. The results for the diol show simple complexation, while those for the trio1 show unequivocal evidence for the formation of a tridentate complex involving all three OH groups. For myo-inositol, however, the 13C n.m.r. spectrum in borate showed evidence of bidentate as well as tridentate complexation, in agreement with the "B results.The 13C n.m.r. data for the carbohydrates (21-31) are also given in table 4. As might be expected, the simplest behaviour was observed for a-methyl glucoside, where the C, carbon atom has no available OH group and the conformation is also simplified by the absence of other anomers. The spectra for the other carbohydrates confirmed our suspicions that several of the available OH groups can compete for complexation with the borate, and the particular findings are summarised in table 4. In a number of cases the spectra of the carbohydrates in NaB(OH), were so complex that the individual resonances could not be assigned. This strongly suggests in these cases that a variety of complexed species were present together in solution and these cases invariably were t See Notice to Authors, J.Chem. SOC., Faraday Trans. 1, 1988, 84, January issue.54 Polarimetry and and I3C N.M.R. Table 4. Complexation trends inferred from n.m.r. data compound comments on complexation with B(0H); 1 ethane- 1,2-diol 2 propane- 1,2-diol 3 propane- 1,3-diol 4 glycerol 5 butane- 1,2,4-triol 6 meso-erythritol 7 threitol 8 xylitol 9 D-arabitol 10 adonitol 11 dulcitol 12 mannitol 13 sorbitol 14 cis-cyclohexane- 1,2-diol 17 cyclohexane-r- 1 ,c-3,c-5-triol 18 myo-inositol 19 2,3-dihydroxydioxan 21 methyl glucoside 22 glucose Simple complexation across C, : C,. Simple complexation across C, : C,. Complexation across C, : C,.Upfield shift of C, resonance. From the 'lB n.m.r. results the greatest complexation involves adjacent C atoms. The three 13C signals indicate C, : C, and C, : C,. From the llB n.m.r. results the greatest complexation involves adjacent C atoms. The 13C indicates C,:C,. Complexation mainly across C,:C,. If C,:C, or C,:C, involved then there would be 4 main peaks. Complexation mainly across C,:C,. If C,:C, or C,:C, involved then there would be 4 main peaks. The inner CHOH groups affected most. Complexation across C, : C, or C, : C,. The C, and C, signals are slightly different. Complexation across C,:C, and C,:C, would produce more lines than for xylitol. The llB n.m.r. suggests considerable reaction across alternate C atoms as well as adjacent positions. If C,:C,, C, : C,, C, : C,, C, : C, and C, : C, positions were involved then this would give possibility of 18 resonances. Principally 3 lines convert into 3 lines indicating C,:C, complexation.The weaker shoulders, however, suggest some C, : C, and C, : C, involvement. Reaction mainly across C, : C,, these two OH groups have the closest disposition. Reaction across C, : C,, C, : C,, C, : C,. Some suggestion that C, may be involved also. Simple complexation across C, : C,, with some different conformations about C, and C,. Formation of the complex produces just two lines and thus must involve a tridentate complex, any other complex would produce more than two resonances. The central 13C signal goes upfield, cJ: propane- 1,3-diol. In excess borate there is probably considerable bidentate complex present in addition to the majority tridentate.The prepared complex when dissolved in water gave (from the llB n.m.r.) 50.3 YO tridentate, 1 1.4 YO bidentate (adjacent OH groups, C,:C, or C,:C,) and 28.3% uncomplexed polyol. Such a mixture should give rise to 14 lines and the 13C n.m.r. spectrum of the prepared complex in water did in fact show 14 signals, at 77.9, 77.7. 77.3, 76.2, 75.4, 75.2, 74.1, 73.7, 71.7, 70.6, 69.0, 66.7, 66.4, 65.3, but of course many of these were overlapping and this influenced their positions. Reaction at C,:C,, although additional small lines at 107.1, 101.9, 99.6, 95.2, 92.7 and at 71.5, 64.6, 63.9, 63.5 suggest that several conformations may be involved. Intensities of C, and C, affected most :. adjacent complexation across C, : C,.Possibly C, has axial conformation. Spectrum suggests complexation mainly across C, : C, and C,:C, and possibly C,:C,. Possibly C, has axial conformation, and molecule in &conformation.J . G. Dawber, S . I. E. Green, J. C. Dawber and S. Gabrail 55 Table 4. Complexation trends inferred from n.m.r. data (cont.) _ _ compound comments on complexation with B(0H); 23 galactose 24 mannose 25 rhamnose 26 xylose 27, 28 arabinose 29 fructose 30 sucrose 31 maltose Spectrum suggests complexation across C, : C, mainly with molecule in a-conformation. In borate all the lines are affected and all broadened. Although the complexation will mainly be across adjacent C atoms the IIB n.m.r. indicated considerable alternate C-atom involvement.Lines at 105.2, 100. I , 97.4, 96.8, 85.4, 84.1, 78.6, 77.2, 73.6, 72.8, 70.7, 67.2, 62.2. Polarimetry showed complex behaviour. 21 new lines produced, many of which occur as pairs: 187.9, 105.4, (99.5, 99.0), (97.1, 96.6), 95.9, (85.1, 84.7), 81.3, (79.3, 78.3, 77.3), 69.5, 68.6, 34.5, (22.7, 22.4), (21.3, 20.9), 15.4. Thus there must be a variety of complexes present, including the possibility of open-chain material (n.b. the line at 187.9 which is likely to be the C=O group). The CH, signal at high field appears at several positions, suggesting a number of different species. Although the llB n.m.r. indicates principally complexation across adjacent C atoms, the optical rotation behaviour was unusual. Spectrum suggests C,, C,, C,, C, involved. The 'lB n.m.r.suggests complexation across alternate as well as adjacent C atoms. Polarimetry gave unusual behaviour. Spectrum suggests C,, C,, C, (C, possibly) involved. The llB n.m.r. indicates predominantly adjacent C atoms involved. Polarimetry could not give value of K,. Complexation probably across C, : C, and C, : C,, mainly in the pyranoside p-form. Main interaction, which is slight, appears to involve C, and C,. The IlB n.m.r. shows little interaction and com- plexation involves alternate C atoms. llB n.m.r. suggests reaction across adjacent and alternate C atoms (more than in sucrose). The 13C n.m.r. spectrum suggests reaction at C,, C,, Ci, Ch, but it is difficult to decide which of the adjacent C atoms are involved. The line at 188.8 ppm suggests that some of the material may be in the open-chain configuration.those for which a value of K, could not be evaluated from the polarimetric results and which also exhibited complex optical rotation behaviour. Of the two disaccharides studied, the 13C n.m.r. data for sucrose were the easier to interpret and suggested (as with the "B n.m.r. data) that the low extent of interaction with B(OH), ion involves complexation across alternate carbon atoms. The spectrum of maltose in NaB(OH), was less easy to interpret, except that it was possible to observe alternate OH-group involvement in addition to that of adjacent groups, and the presence of an n.m.r. line in the C=O region suggests that there may be a small amount of complexation involving the open-chain material. Conclusions Complementary studies of polarimetry, "B n.m.r.and 13C n.m.r. show that the B(OH), ion has almost universal ease of complexation with polyols and carbohydrates. The results show conclusively that in many cases the reaction involves various56 Polarimetry and 'I% and 13C N.M.R. competing equilibria in which several of the OH groups in the polyol/carbohydrate may participate, and that the interpretation of the behaviour is greatly facilitated by llB n.m.r. spectroscopy. The previously developed method of evaluating equilibrium constants from polarimetric results was of limited use for the complex equilibria in these systems. We thank Mr P. Doughty of the Mining Engineering Department, N.S.P., for carrying out the C and H analyses, and the S.E.R.C. for funds towards the cost of the polarimeter.References 1 J. Boiseken, Adv. Carbohydr. Chem., 1949,4,189; 1949,12,81; J. P. Sickels and H. P. Schultz, J. Chem. 2 R. F. Nickerson, J. Inorg. Nucl. Chem., 1968, 30, 1447; 1970, 32, 1400. 3 G. W. Campbell, J. Inorg. Nucl. Chem., 1969, 31, 2626. 4 H. B. Davis and C. J. B. Mott, J. Chem. Soc., Faraday Trans. 1, 1980, 76, 1991. 5 R. Larsson and G. Nunziata, Acta Chem. Scand., 1970, 24, 2145. 6 M. Mazurek and A. S. Perlin, Can. J. Chem., 1963, 41, 2403. 7 J. G. Dawber and D. H. Matusin, J. Chem. Soc., Faraday Trans. I , 1982, 78, 2521. 8 J. G. Dawber and G. E. Hardy, J. Chem. Soc., Faraday Trans. 1, 1984, 80, 2467. 9 R. E. Moore, J. J. Barchi and G. Bartolini, J. Org. Chem., 1985, 50, 374. 10 J. G. Dawber and S. I. E. Green, J. Chem. Soc., Faraday Trans. I , 1986, 82, 3407. 11 J. G. Dawber, J. Chem. SOC., Faraday Trans. 1 , 1987, 83, 771. 12 S. J. Angyal and D. J. McHugh, J. Chem. Soc., 1957, 1423; S . J. Angyal, J. E. Klavins and J. A. Mills, Aust. J. Chem., 1974,27, 1075; P. J. Garegg and K. L. Lindstrom, Acta Chem. Scand., 1971,25, 1559; R. M. Williams and R. H. Attala, in Solution Properties of Polysaccharides, ed. D. A. Brant (American Chemical Society, Washington D.C., 1981). 13 M. Van Duin, J. A. Peters, A. P. G. Kieboom and H. Van Bekkum, Tetrahedron, 1984, 40, 2901; M. Makkee, A. P. G. Kieboom and H. Van Bekkum, Reel. Trav. Chim. Pays-Bas, 1985, 104, 230; M. Van Duin, J. A. Peters, A. P. G. Kieboom and H. Van Bekkum, Tetrahedron, 1985, 41, 3421 ; M. Van Duin, J. A. Peters, A. P. G. Kieboom and H. Van Bekkum, Reel. Trav. Chim. Pays-Bas, 1986, 105, 1986. Educ., 1964, 41, 343; J. Boiseken and N. Vermaes, J. Phys. Chem., 1931, 35, 1477. 14 M. Van Duin, Doctoral Thesis (Technological University of Delft, Delft University Press, 1986). 15 J. G. Dawber, J. Chem. SOC., Faraday Trans. 1, 1978, 74,960. 16 R. K. Harris, Nuclear Magnetic Resonance Spectroscopy (Pitman, London, 1983), p. 71. 17 R. J. Ferrier and P. M. Collins, in Monosaccharide Chemistry, Penguin Library of Physical Sciences: Chemistry (Penguin Books, Harmondsworth, 1972) ; J. F. Stoddart, in Stereochemistry of Carbo- hydrates (Wiley, New York, 1971). 18 W. Voelter, E. Breitmaier, G. Jung, T. Keller and D. Hiss, Angew. Chem. Int. Ed. Engl., 1970, 9, 803; K. Bock and H. Thogersen, Ann. Rep. NMR Spectrosc., 1982, 13, 1; M. Voelter, V. Bilik and E. Breitmaier, Collect. Czech. Chem. Commun., 1973,38, 2054; T. A. W. Koerner, R. J. Voll, L. W. Cary and E. S. Younathan, Biochem. Biophys. Res. Commun., 1978, 82, 1273; L. J. Johnson and W. C. Jankowski, in Carbon-I3 NMR Spectra (John Wiley, New York, 1972, p. 195; E. Breitmaier and W. Voelter, 13C NMR Spectroscopy (Verlag Chemie, Weinheim, 2nd edn, 1978). Paper 6/1998 ; Received 10th October, 1986