摘要:

S. Chem. SOC., Faraday Trans. I, 1988, 84(8), 2585-2594 Fourier-transform Infrared Vibrational Circular Dichroism of Simple Carbohydrates Chandra M. Tummalapalli, Darlene M. Back and Prasad L. Polavarapu" Department of Chemistry, Vanderhilt University, Nashville, Tennessee 37235, U.S.A. Vibrational circular dichroism spectra have been measured in [2HH,]DMS0 solutions for simple carbohydrates in the 1650-800 cm-' region. The links between the observed spectral features and configurational and con- formational aspects of carbohydrates were investigated. Some useful correlations and difficulties involved in deciphering the spectral content are pointed out. Vibrational circular dichroism (v.c.d.) is a measure of differential absorption of left versus right circularly polarized i.r.radiation due to the vibrational transitions of chiral molecules. The unique nature of v.c.d. spectroscopy results from the combination of a large number of vibrational transitions with the stereochemical sensitivity inherent in circular dichroism (c.d.) The conventional c.d. studies are carried out in the visible spectral region, probing the electronic transitions, and such accessible electronic transitions are usually very few in number. For simple carbohydrates, in particular, there are no electronic transitions in the visible spectral region, which required a major portion of optical activity studies on simple carbohydrates to be limited to studying the optical rotation.' Thus v.c.d. spectroscopy offers two advantages for simple car- bohydrates. There are 3n - 6 fundamental vibrational transitions where n is the number of atoms in a molecule, and every one of these transitions can, in principle, exhibit c.d.; this provides a very large base of accessible spectral transitions.Secondly, since different vibrations encompass different portions of the molecule, c.d. features associated with different vibrational bands provide three-dimensional structural information in different portions of the molecule. Thus a complete v.c.d. spectrum contains the necessary information about the three-dimensional structure of the entire molecule in the solution phase. In practice, though, the determination of the complete molecular structure from v.c.d. spectra is rather ambitious, with the current state of knowledge, owing to (i) the complicated nature of molecular vibrations, (ii) the overlapping of vibrational bands and (ili) the small v.c.d.signals that are beyond instrumental sensitivity. In this paper we present some of the v.c.d. spectral studies carried out in our laboratory in the last five years in the 1650-800 cm-I region. V.c.d. spectra of lactones, sorbose and some additional simple carbohydrates were also investigated in our laboratory, and they will be reported separately. A preliminary report from our laboratory2 and two other v.c.d. paper^^^^ in the C-H stretching region comprise the previously published v.c.d. studies for carbohydrates. Experimental The v.c.d. spectrometer used for the present studies is based on a commerical Fourier- transform i.r. spectrometer (Nicolet 6000C).The instrumental details for these v.c.d. measurements are given el~ewhere.~ The raw v.c.d. features are usually dominated by the artifacts resulting from the birefingence in the optical path, which makes it difficult to determine the zero line. It is common practice to eliminate such artifacts either by 25852586 Vibrational Circular Dichro ism B 0.0 i I 1 I I I 1575 1475 1375 1275 11 75 1075 w avenum ber/cm -' Fig. 1. Vibrational absorption (A) and v.c.d. (B) spectra in [2H,]DMS0 solution (0.5 mol dmP3) of D-allose-(OD),. (a) and (b) were obtained employing 100 pm and 205 pm pathlengths, respectively. 1 0 Q) C 0.5 9 .n 0.0 A A - I I I 1 I 1575 1475 1375 1275 11 75 1075 wavenumber/cm -' Fig. 2. Vibrational absorption (A) and v.c.d. (B) spectra in [2H,]DMS0 solution of D-fucose- (OD), [(a) pathlength, 65 pm] and D-fucose-(OH), [(b) pathlength, 65 pm].C. M .Tummalapalli et al. 1 .o- a c1 0 5 - s1 -2 0.0 , 2587 A 1 I I 1 I wavenumberlcm-' Fig. 3. Vibrational absorption (A) and v.c.d. (B) spectra in [2H,]DMS0 solution of D-ribose- (OD), [(a) pathlength, 70 pm] and D-ribose-(OH), [(b) pathlength, 60 pm]. The bisignate v.c.d. feature at 1240 cm-' in (b) was not reproducible. B w T 0 . 5 x 1 0 - 4 1 A 0 1575 1475 wavenumber/cm-' Fig. 4. Vibrational absorption (A) and v.c.d. (B) spectra in [2H,]DMS0 solution of D-arabinose- (OD), [(a) pathlength, 50 pm] and D-arabinose-(OH), [(b) pathlength, 75 pm]. At a higher pathlength, a negative v.c.d. band was found at 13 15 cm-l for D-arabinose-(OD),.2588 Vibrational Circular Dichroism I 0.5 x 10-4 1 .Q - A f 1575 147 5 1375 1275 1175 1075 wavenumber/cm-' Fig.5. Vibrational absorption (A) and v.c.d. (B) spectra in [2H,]DMS0 solution of D-xylose- (OD), [(a) pathlength, 45 pm] and D-xylose-(OH), [(b) pathlength, 75 pm]. At a higher pathlength two weak bands of uncertain nature were found at 1375 and 1335 cm-I for D-xylose-(OD),. 0 . 5 ~ 1 0 - ~ (a) 7 %- -a+-- A - 8 E 2 0 . 0, P 0 5 - .O I I 1 I 1 1575 1475 1375 1275 1175 1075 wavenumber/cm-' Fig. 6. Vibrational absorption (A) and v.c.d. (B) spectra in [*H,]DMSO solution of D-galactose- (OD), [(a) pathlength, 50 pm] and D-galactose-(OH), [(b) pathlength, 60 pm].C. M. Tummalapalli et al. 2589 0.0 I' I I I I 1575 1475 1375 1275 1175 wavenumber/cm -' Fig. 7. Vibrational absorption (A) and v.c.d.(B) spectra in [*H,]DMSO solution of D-glucose- (OD), [(a) pathlength, 85 pm] and D-glucose-(OH), [(b) pathlength, 60 pm]. substracting the raw v.c.d, of racemic mixture or by taking one-half of the difference between raw v.c.d. of enantiomers. Commercially available D- and L-enantiomers of carbohydrates do not necessarily contain the same composition of anomers. This required us, in most cases, to equilibrate a given enantiomer in water and lyophilize the equilibrated solution. For studies on hydroxyl-deuterated carbohydrates, the samples were equilibrated in D,O. The lyophilized samples were studied in [2H,]DMS0 solvent at 1 mol dm-3 concentration, unless specified otherwise. In the fig. 1-7, ordinary absorption is displayed at the bottom and v.c.d.is displayed at the top. V.c.d. is represented as AA = A , - A , , where A , and A,, respectively, are the absorbances for left and right circularly polarized light. The scale displayed above the v.c.d. spectra is for AA. From these spectra, it can be seen that the ratio A A / A for the majority of the vibrational bands of carbohydrates is less than Results and Discussion Owing to the nature of vibrations and of overlaping vibrational bands the spectra are divided into two regions, one from 1500 to 1200 cm-' and another from 1200 to 1100 cm-l. Strong absorption due to the [2H,]DMS0 solvent precludes measurements from ca. 1100 to 950 cm-l. The region below 950 cm-I is not discussed in this paper. Exocyclic C-0 Stretching Motions In a given category of pentoses or hexoses, simple carbohydrates differ in the orientation of individual 0-H groups being either in an axial or equatorial position.If a C-0 stretching vibrational band from each chiral centre is identifiable one would expect toTable 1. The structural details and v.c.d. of (OH deuterated) D-sugars ~ _ _ _ _ ___ ~ __ - ~ pyranose individual group contributions cumulative net vibrational compound composition (%)a conformation C,-C, C,-C, C,-C, c4-c.5 contribution effect band position arabinose a P lyxose a P P P P xylose a glucose a galactose a fucose a P 30 60 75 25 80 20 42 58 32 64 38 59 sign 'C4 'C4 'C1 C.W. a.c.w. C.W. 0 C.W. C.W. C.W. a.c.w. C.W. a.c.w. C.W. a.c.w. C.W. a.c.w. a.c.w. a.c.w. 0 a.c.w. a.c.w. C.W. C.W. C.W. C.W. C.W. C.W.C.W. C.W. C.W. a.c.w. a.c.w. C.W. a.c.w. 0 a.c.w. a.c.w. a.c.w. a.c.w. a.c.w. C.W. C.W. C.W. C.W. - - - - - - - - C.W. C.W. a.c.w. a.c.w. a.c.w. a.c.w. a.c.w. a.c.w. 3a.c.w. 2c.w. 2a.c. w. 2c.w. a.c.w. a.c.w. C.W. C.W. a.c.w. C.W. C.W. 0 C.W. C.W. 0 C.W. C.W. 0 I148 1157 1149 1155 1161 1180 + - a Determined from proton n.m.r. spectra. Furanose compositions are not listed.C. M. Tummalapalli et al. 259 1 derive the configurational information at a given chiral centre from the v.c.d. associated with C-0 stretching at that centre. On the other hand, if these C-0 stretching vibrational bands overlap one can only extract the overall orientation of C-0 groups around the carbohydrate ring. Each 0-C-C-0 segment associated with C,-C,, C,-C, and C,-C, with either a clockwise (c.w.) or an anticlockwise (a.c.w.) dihedral angle will have local C, symmetry, resulting in a symmetric C-0 stretching motion and an antisymmetric C-0 stretching motion.For an a.c.w. dihedral angle, the antisymmetric stretching vibration would have6, ' negative v.c.d. and the symmetric stretching vibration would have positive v.c.d.. For a C.W. dihedral angle, the v.c.d. signs would be interchanged. For 0 or 180" dihedral angle, the local plane of symmetry precludes any v.c.d. Thus if we take exocylic 0-C-C-0 segments and add up the individual contributions for symmetric and antisymmetric motions separately, then one can estimate the v.c.d. associated with the C-0 stretching vibrations. When C-0 groups are mechanically and (or) electronically dependent the relative absorption intensities can be7 such that one of these two stretching bands might have very weak absorption intensity.With this theoretical background, v.c.d. associated with a band at ca. 1150 cm-' will be interpreted. The v.c.d. spectra showing this band are displayed in fig. 1-6 for selected carbohydrates. The dihedral angles for all the exocyclic 0-C-C-0 segments in the dominant conformers of carbohydrates are summarized in table 1 under the column labelled ' individual contributions '. For hexoses, the exocyclic CH20H (or CH, for fucose) is taken as equivalent to a OH group since the C-0 stretches are usually coupled to C-C stretches. The net effect is derived by taking the composition-weighted sum of the cumulative contributions from individual conformers. It can be seen that a net a.c.w.effect corresponds to positive v.c.d., and a C.W. effect corresponds to negative v.c.d. experimentally observed for the band at ca. 1150 cm-l. This band can be inferred to originate from a normal vibration with the symmetric C-0 stretching component being dominant. The presence of this band with identical v.c.d. signs for both hydroxyl- deuterated (table 1) and undeuterated carbohydrates2 provides additional support for assigning this band to C-0 stretching motions. For more confidence in the presented correlation, experimental verification of the assignment of the 1150 cm-l band to delocalized C-0 stretches would be important. For this purpose, we have investigated8 the F.t.i.r. absorption of isotopomers of glucose in [2H,]DMS0 and time-dependent F.t.i.r.absorption of a-D-glucose, a-D-galactose and am-fucose in H,O. During the conversion from cc to anomer in water solution, the absorption band under consideration shifts from 1152 to 1167 cm-' for galactose, from 1167 to 1178 cm-l for fucose and from 1148 to 1150 cm-' for glucose. In galactose and fucose this band appears to be of the same intensity relative to the other bands in this region, but in glucose the corresponding relative intensity is much weaker. With regard to the position of the band, there is a ca. 10 cm-' shift from galactose to fucose. These observations suggest that the band under consideration has significant influence from the orientation of exocyclic C,-0, C,-0 and C,-C groups. These points are supportive of the above analysis (table 1) that included the contributions from all exocyclic C-0 (and C-C) groups.In table 1, allose and ribose were not included for one reason. Allose-(OD), exhibits two absorptions bands one at 1161 cm-l with positive v.c.d. and another at ca. 1150 cm-l (overlaped by solvent absorption) with negative v.c.d. It is not obvious as to which one of these two should be used for verifying the correlation in table 1. Similar uncertainty exists with ribose-(OD),, which has absorption bands at ca. 1159 and C~Z. 1125 cm-' with opposite v.c.d. signs.2592 Vibrational Circular Dichroism Table 2. Frequencies and signs of v.c.d. bands for D-sugars in the 1600-1200 cm-l region B-allose-(OH), allose-(OD), glucose-(OH), glucose-(OD), galactose-(OH), galactose-(OD), a-fucose-(OH), fucose-(OH), fucose-(OD), ribose-(OH), ribose-(OD), arabinose-(OH), arabinose-(OD), a-x ylose-(OH), 1367(+) 1405( -) 1375( +) 1383( -) 1455(-) 1400(+) I456( -) 1398( + ) 1456( -) 1400( +) 1378( -) none - - none none none - - 1398( -) 1452( -) 1406( +) 1367( -) xylose-(OH), 1458( - ) 1406( + ) 1363( - ) xylose-(OD), 1 yxose-(OH), lyxose-(OD), 1375( +) none none - - 1335(-) 1304(+) 1356(-) 1313(+) 13 12( + ) 1361(+) 1327(+) 1282( +) 1304(+) 1313(+) - 1344(-) 131 5( -) 1300(+) 1300(+) 133% + ) - 1248( -) 125 I( - ) 1231(+) 1212(+) 1255( -) 1236( +), 1257( -) 1236( +), 1205( -) 1204( -) H-0-C and H-C-C Bending Motions The vibrational bands with major contributions from H-0-C and H-C-C bending motions are expected to be present in the 1600-1200 cm-l region.The v.c.d. spectra in this region for hydroxyl-deuterated as well as undeuterated carbohydrates are presented in fig. 2-7, and the v.c.d. features are summarized in table 2. Some general conclusions are in order. (a) In order to obtain the same level of absorbance for undeuterated and deuterated compounds, at the same concentration, the pathlengths employed for deuterated compounds were approximately twice the corresponding pathlengths employed for undeuterated compounds. Therefore the strong absorption features seen for undeuterated compounds can be attributed to the H-0-C bending vibrations. By the same reasoning most of the absorption features seen for deuterated compounds can be attributed to the H-C-C bending vibrations. (6) While xylose-(OH),, galactose- (OH), and fucose-(OH), have measurable v.c.d.features, the corresponding deuterated compounds have either weak or no v.c.d. features. This indicates that for undeuterated compounds v.c.d. is dominated by the contributions from H-0-C bending vibrations. Some coupling of H-0-C bending motions with H-C-C bending motions cannot be excluded. (c) Two pentoses, ribose and lyxose have no measurable v.c.d. features, both in undeuterated and deuterated forms. Arabinose has some v.c.d. features, but these are definitely much weaker than those seen for xylose. Recalling that lyxose, ribose and arabinose are also the ones which can exist in both C1 and IC conformations, and that allose and galactose, which are similar to ribose and arabinose except at C , and exist only in the C1 conformation, have measurable v.c.d.features, one conclusion becomes transparent, i.e. the additional conformation present for ribose, lyxose and arabinose provides mutually cancelling v.c.d. contributions in this region. (d) Identical v.c.d. features are seen for galactose-(OH), and fucose-(OH),, except that the v.c.d. feature at ca. 1375 cm-l appears to be more negative for fucose than it is for galactose. This observation suggests that the replacement of the CH,OH group by the CH, group has no major influence on v.c.d. ( e ) Comparison of the v.c.d. spectra of glucose-(OH), and xylose-(OH), provides some information on the coupling between H-0-C andC. M . Tummalapalli et al. 2593 H-C-C bending motions. These two differ only in the substituent at C, and exist only in the C1 conformation, yet the v.c.d.spectra look quite different. V.c.d. spectra of the corresponding deuterated compounds are also significantly different, with glucose having larger signals. This indicates that the H-C-C bending motions in glucose have significant v.c.d. contributions and the interaction of' H-0-C and H-C-C bending motions provide cancelling v.c.d. contributions in glucose-(OH)5. Such cancelling contributions make the v.c.d. spectrum of glucose-(OH), look simpler than that of xylose-(OH),. Here one may note that glucose-(OH), has ca. 60 YO /3 conformer, while xylose-(OH), has 80 YO a ; however, this difference in composition may not be completely responsible for the observed differences. This is because the v.c.d.spectrum of a-xylose- (OH), is found to be identical to that of equilibrated xylose-(OH),. (e) The v.c.d. spectra of glucose-(OD), and allose-(OD), are similar, in that a -, + , + pattern is seen from the high-frequency side, with some difference in the spacing between the v.c.d. bands. As mentioned earlier, galactose-(OD), and fucose-(OD), do not have any v.c.d. features. Since allose and glucose differ in the orientation of groups at C,, galactose and glucose differ in the orientation of groups at C,, the observed v.c.d. in glucose and allose can be considered to originate from the interaction of axial H-C-C bending motions at C, and C,. The above discussion indicates that v.c.d. spectra of carbohydrates in the region 1600-1 200 cm-l provides useful information on the nature of internal coordinate contributions to the observed vibrational bands.Therefore v.c.d. data are a useful source of information for inferring the vibrational band assignments. Now we address the feasibility of deriving configurational or conformational information at each of the chiral centres. This is a rather challenging task, as will be evident from the following discussion. As mentioned earlier, xylose and glucose are similar except for the substituent at C,; galactose differs from glucose at C,. From a simplistic viewpoint, the v.c.d. associated with the H-0-C, and H-0-C, bending motions should be the same for these molecules if the vibrations are localized at a given centre. The negative v.c.d. band at ca. 1356 cm-l, the positive v.c.d.band at ca. 1313 cm-l and the negative v.c.d. band at ca. 1248 cm-I seen for glucose have similar counterparts for xylose at 1363, 1300 and 1257 cm-'. The bands corresponding to the latter two are also present for galactose at 1282 and 1251 cm-'. Thus the positive v.c.d. band in the range 1280-1315 cm-I and the negative v.c.d. band at ca. 1250 cm-l may reflect the configuration at C, and C,. Since allose differs from glucose at C,, one of the abovementioned bands can be expected to either change sign or disappear. In the v.c.d. spectrum of allose, the band corresponding to that at ca. 1250 cm-l in glucose and galactose is in fact absent. This leaves the positive v.c.d. band at ca. 1300 cm-', in all the four molecules considered, to represent the configuration at C,.However, considerable uncertainty exists in the above analysis. It is quite possible that the aforementioned common v.c.d. features are present for entirely different reasons. There is no proof that the positive v.c.d. at ca. 1300 cm-l originates from the H-0-C, bending mode. When the axial/equatorial orientations of the groups at C, or C, are changed the H-0-C, bending mode can shift in frequency. Thus the difficulty in making a precise identification of the vibrational origin of a given band, especially where several bands appear together, is the single important reason that curtails the utility of v.c.d. spectral features in deducing the three-dimensional molecular sl ructure. V.c.d. data discussed earlier suggested that the normal modes responsible for the bands in the 1600-1200 cm-l region appear to have mixed contributions from the H-0-C and H-C-C bending internal coordinates.In such cases, v.c.d. features will be determined by the relative orientation of groups at different centres and the phase associated with the vibrational motions at different centres. We are currently concentrating on the ah initio calculation of vibrational properties2594 Vibrational Circular Dichroism of carbohydratesg which will help us to understand the order and nature of the vibrational bands. We hope to interpret the v.c.d. in the 1600-1200 cm-l region more confidently when these calculations are completed. Conclusions The v.c.d. associated with a band at ca. 1150 cm-' indicates that the overall orientation of hydroxyl groups can be deduced from the v.c.d. studies. In the 1600-1200 cm-l region, v.c.d. data indicate that the vibrational bands appearing here have mixed contributions from the H-0-C and H-C-C bending motions. However, the presence of several bands in this region renders the identification of vibrational origin of the bands quite difficult. As v.c.d. interpretations require a reliable knowledge of the internal coordinates participating in a given normal vibration, further progress is required before v.c.d. can be used to its full potential. This work was supported by grants from NIH (GM 29375) and Vanderbilt University. References 1 R. S. Shallenberger, Advanced Sugar Chemistry (AVI Publishing Co, Westport, 1982). 2 D. M. Back and P. L. Polavarapu, Carbohydr. Res., 1984, 133, 163. 3 C. Marcott, H. A. Havel, J. Overend and A. Moscowitz, J . Am. Chem. Soc., 1978, 100, 7088; H. A. 4 M. G. Paterlini, T. B. Freedman and L. A. Nafie, J . Am. Chem. SOC., 1986, 108, 1389. 5 P. L. Polavarapu, Fourier Transform Infrared Spectroscopy, ed. J. R. Ferraro and L. J. Basile 6 G. Holzwarth and I. Chabay, J . Chem. Phys., 1972, 57, 1632. 7 P. L. Polavarapu, J . Chem. Phys., 1986, 85, 6245; 1987, 87, 4419. 8 D. M. Back and P. L. Polavarapu, Carbohydr. Res., 1987, 165, 173; 1983, 121, 308. 9 Preliminary calculations with semi-empirical wavefunctions have been reported: D. M. Back and P. L. Havel, Ph.D. Thesis (University of Minnesota, 198 1). (Academic Press, New York, 1985), vol. 4, p. 61. Polavarapu, J. Comput. Chem., 1987, 8, 772. Paper 712083; Received 23rd November, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402585

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1988, 84(8), 2595-2602 A Nuclear Magnetic Resonance Study of Isomeric Pentitols in Aqueous and Non-aqueous Solutions Felix Franks" Pafra, Biopreservation Diuision, 150 Science Park, Milton Road, Cambridge CB4 4GG Robert L. Kay and Josef Dadok Department of Chemistry, Carnegie-Mellon University, Pittsburgh, PA 15213, USA. 3J coupling constants for ,H exchanged pentitols in D,O, [2H,]pyridine and [2H,]acetone have been determined at 25 "C after refinement by computer simulation of the fully resolved lH n.m.r. spectra obtained at 620 MHz. There are significant differences between the coupling constants for aqueous and non-aqueous solvents but only minor differences between those for pyridine and acetone. A conformational analysis is proposed which is based on the observation that in solid crystals all the H-C-C-H dihedral angles are close to either 60" (gauche) or 180" (trans).From an application of the Karplus equation to the solution 3J values it appears unlikely that single unique conformations of the alditols exist in solution. Instead the results are interpreted in terms of different populations of rotamers that are rapidly equilibrating between dihedral angles of & 60 and 180". Thermo- dynamic properties of the alditols in D,O are shown to be a reflection of the conformation of the specific pentitol in solution rather than that of the crystal structure. Carbohydrates and related polyhydroxy compounds (PHC) are ubiquitous natural materials. They fulfil many and varied biological and technological functions, such as load-bearing structures, sources of metabolic energy, biological recognition, in vivo lubrication, freeze/desiccation resistance, and gel formation.Although chemically quite similar, simple PHC molecules display a variety of molecular structures and shapes, including the pyranoid and furanoid ring structures and their acyclic analogues. One common characteristic of PHCs is their solubility in, or even complete miscibility with water and other polar solvents. Many PHC crystal structures have been determined; they are characterised by complex intra- and inter-molecular hydrogen-bond arrange- ments. In recent years increasing attention has been paid to the shapes of PHC molecules in solution and their interactions with the solvent and with one another.14 Atomic-group (CH,, CH,, CHOH, >C=O etc.) additivity schemes have been proposed as means of predicting the solution properties of organic compounds. 5-7 Whatever might be the practical value of such procedures, they are unable to account for observed differences in the solution behaviour of different PHC epimers.This approach, as applied to solutions of xylitol, ribitol and arabinitol, will presently be examined in more detail. From an experimental viewpoint all physico-chemical studies on sugars are complicated by their chemical heterogeneity, i.e. the coexistence in solution of two or more anomeric and/or tautomeric species. In some cases the equilibrium mixture might contain as many as eight distinct species (e.g. ribose), in proportions which depend on temperature, concentration and the nature of the solvent medium.* One way of dealing with this problem is to block the anomeric OH group.However, such a chemical modification produces changes in the subtle solvation interactions and also in the 25952596 OD I H,-C-H,/ I An N.M.R. Study of Isomeric Pentitols OD I OD I H,-C-H,/ H,-C-H , I I I H,-C--BD H,-c-OD DO-C-H, I I I H,-C-OD I DO-C-€4, I H,-C-OD I H,--6-OD H,-C--OD H,-C-OD I I I H,-C-H,/ I H5-C-Hjf I H,-C-H,. I OD OD OD ribitol xylitol arabinitol Fig. 1. Fischer projections of the three pentitols. interactions between the sugar molecules themselves, the general result being that the modified sugar takes on a more hydrophobic ~ h a r a c t e r . ~ An alternative method of studying solvent effects on PHC shapes, without the complications posed by the different co-existing anomeric and tautomeric species, is to examine non-reducing acyclic molecules which would be expected to be flexible and might be treated as suitably weighted time-averaged structures, depending on the details of their respective free-energy surfaces.An experimentally determined average shape might then be compared, in terms of the torsional angles, with the known molecular conformation as it exists in the crystal or to the shape of the isolated molecule, as obtained from conformationai energy calculations or molecular dynamics simulations. We have adopted this approach to study the effects of water, pyridine and acetone on the molecular geometries of the three epirneric pentitols : arabinitol, ribitol and xylitol; the respective OH positions on the asymmetric carbon atoms, 2, 3 and 4 are shown in fig.1 as Fischer projection formulae (the proton numbering system used is indicated). Crystal-structure determinations have led to the general rule that alditols exist as planar zig-zag carbon chains, unless such a conformation leads to parallel I73-interactions between C-0 bonds (0 )I 0), in which case bond rotation through 120" takes place to produce a non-linear, sickle-shaped conformation.2 Thus, of the three pentitols, only arabinitol adopts the linear conformation in the solid state, while xylitol and ribitol exist in the sickle-shaped carbon chain conformation." It has also been suggested that, at least in aqueous solutions, the presence of an 0 (1 0 interactior, directs the bond about which rotation occurs.Thus, Horton and Wander," who studied the conformations of diethyldithioacetal derivatives of ribose and xylose, found that in ribose the eclipsing of acetoxy groups was avoided by rotation about the C(3)--C(4) bond, whereas in xylose rotation occurs about the C(2)-C(3) bond. It has been suggested that in aqueous solution the alditol molecules retain the conformations that exist in the crystal, i.e. planar zig-zag or sickle-shaped.'2 If this is correct, then ribitol and xylitol would be expected to resemble each other in their solution properties with differences being observed for arabinitol. However, calorimetric measurements on alditol solutions (limiting heat capacities and heats of dilution) show that it is arabinitol and ribitol which are indistinguishable (within experimental error) but different from xylitol."~"" Tt does not seem obvious why the alditol conformations should be identical in the crystal and in hydrogen-bonding solvents.The crystal conformation is dictated by packing constraints and details of the intermolecular hydrogen-bonding networks which, in solution, are presumably replaced by solute- solvent hydrogen bonds.F. Franks, Robert L. Kay and J . Dadok 2597 Table 1. 3J Coupling constants for ribitol, xylitol and arabinitol in D20 and [2H,]pyridine and xylitol in [2H,]a~etonen 4 1 . 2 ) 3.07 3.83 4.29 5.40 5.47 5.04 6.12 7.18 6.19 6.87 5.70 5.59 7.57 6.19 6.27 6.64 4.49 3.65 3.40 2.09 2.08 4 4 . 5 ) 3.03 4.12 41.. 2) 4 3 , 4 ) 8.46 7.93 4 4 , 5 ' ) 6.45 5.95 4 2 .3 ) Units: Hz. Primary geminal protons are identified by a prime to indicate the proton having the larger coupling constant to the adjacent proton. Apart from optical activity, the methods which have led to the most significant advances in studies of carbohydrate conformations, are based on n.m.r. spectroscopy. The advent of 13C n.m.r. spectroscopy led to progress in the elucidation of complex oligosaccharide structures, and more recently the development of high-field 'H n.m.r. spectroscopy has provided a renewed stimulus to more refined investigations. Chemical shifts, coupling constants, nuclear Overhauser effects and nuclear magnetic relaxation rate measurements are currently being applied to studies of the detailed conformations of individual sugar rings15 and the shapes of di- and oligo-saccharides, as they exist, for instance, in the blood-group substances. l6 The present study was performed by 'H n.m.r.spectroscopy at 620 MHz on 2H-exchanged alditols in solution and followed preliminary investigations based on proton-decoupled 13C n.m.r. spectroscopy on solutions and cry~ta1s.l~ The 13C spectrum of arabinitol in solution exhibits five resolved signals of equal intensity, whereas the corresponding spectra of xylitol and ribitol consist of three lines with intensity ratios 2 : 2 : 1, indicating the time-average symmetry of the latter two molecules. This is in contrast to the crystalline states in which these molecules adopt non-linear conformations without symmetry, a fact which was supported by the appearance of the solid-state 13C n.m.r. spectra (five signals).However, because of the symmetry of the molecules, they could be in rapid equilibrium between two equivalent non-linear forms, and still give 3 signals with intensity ratios 2 : 2 : 1 Experimental The experimental details will be published elsewhere. l8 Spectra were recorded using the 620.6 MHz spectr~rneterl~ at the N.M.R. Facility for Biomedical Studies in Carnegie- Mellon University, in the rapid scan correlation mode2' with 8 K data points at a typical spectral sweep width of 300 Hz. Moderate resolution enhancement was applied to all spectra. 1 -D spin-decoupling experiments were performed in the conventional way to establish assignments. The spectra were simulated with the PANIC program integrated in the Bruker software.Ribitol, xylitol and arabinitol were obtained from Sigma, and used without further purification. The D,O was 99.8% D (Aldrich Gold Seal) and the [2H,]pyridine and [2H,]acetone were 99.5% D.2598 An N.M.R. Study of Isomeric Pentitols Table 2. H-H dihedral angles in the crystal" ____~ ~_____ __ __ Qcr,, H,-H, H,,-H, H,-H, H,-H, H,-H, H,-H,, ribitol 53 70 159 71 69 179 xylitol 63 I77 60 172 68 173 arabini to1 53 167 69 171 70 62 ____ __ " Data from ref. (24). Angles in degrees. 0 20 40 60 80 100 120 140 160 18( 9 Fig. 2. Plot of the Karplus equation [eqn (I)] us. the H-H dihedral angles in the crystal. The points give the experimental 3J values: 0, ribitol; 0, xylitol; a, arabinitol. Open symbols, D,O; filled symbols, [2H,]pyridine.3J = 0.8 cos Q + 10.2 cos' Q. Results The 3 J ~ ~ ~ p l i n g constants for ribitol, xylitol and arabinitol in D,O and [2H,]pyridine and for xylitol in [2H,]acetone shown in table 1 were determined by least squares with a standard deviation of 0.02 Hz or less, Since ribitol and xylitol are symmetrical molecules the entries for these pentitols in table 1 reflect this symmetry. On the other hand, the lack of symmetry in arabinitol results in six independent coupling constants although it is not possible to identify the orientation unambiguously. The assignment is consistent with those given for the other two pentitols but it is not excluded that the parameters for arabinitol should be inverted. Our data are in excellent agreement with those of Hawkes and Lewis21 where comparison is possible (i.e.D,O solutions). Discussion The coupling constants for the pentitols in pyridine show significant differences up to 1 Hz when compared to those in D20. The coupling constants for xylitol in acetone, on the other hand, differ from those in pyridine by 0.2 Hz at most, indicating an insensitivity of 'J to the nature of the non-aqueous solvent.F. Franks, Robert L. Kay and J. Dadok 2599 Table 3. Percentage H-H trans-rotamers in the pentitols in solution compared to the conformations in the crystal ribitol xylitol arabini to1 crystal D,O E2H5]Py crystal D,O [,H,]Py crystal D,O [2H5]Py HI-H, g 0% 13% g 20 Yo 38 Yo g 32% 49% H,-H, t 51 57 g 23 10 g -14 -14 H,,-H, g 65 47 t 60 42 t 71 49 Id,-H, g 51 57 t 23 10 t 85 77 K-H5 g 0 13 g 20 38 g 0 18 H,-H,, t 65 47 t 60 42 g 54 46 Table 4.Percentage departure from free rotation ribitol ( O h ) xylitol (YO) arabinitol (YO) It is possible to attempt a detailed conformational analysis using these coupling constants in extensions to the Karplus equation22 that take into account the orientations and electronegativities of the substituents about the coupling but, considering the known fundamental uncertainties associated with the Karplus equation, the value of such an analysis is uncertain. We have taken a less complicated approach to illustrate the conformational differences of the pentitols in solution and in the crystal. We have used the Karplus equation in a form devised for J = 0.8 COS#+ 10.2 C O S ~ # (1) to estimate the proton-proton coupling constants from the dihedral angles #crys (recorded in table 2) found in the known crystal structures of the pentit01s.'~ An inspection of the data in table 2 indicates that the dihedral angles with few exceptions are centred within 10" around 60 and 180" in the crystal.Thus, in the crystal the conformations of the protons are close to being completely gauche or completely trans. According to eqn. (l), these two extreme values of q5,,,, correspond to 3J values of 3.0 and 9.4 Hz, respectively. On the other hand, many of the solution values of 3J lie between these two extremes, suggesting that the measured 3J values reflect rapidly interconverting mixtures of trans and gauche conformers. This can be seen in fig. 2 where the solid line gives the coupling constants calculated from eqn.(1) and the points are the measured solution 3J values plotted us. the corresponding dihedral angle in the solid. If the crystal conformation was the dominant conformation in solution, the points would lie close to the line. Instead, there is considerable vertical displacement indicating significant amounts of rapidly inter- converting mixtures of gauche and trans conformations in solution. Most coupling constants corresponding to the gauche arrangement in the crystal are much larger than predicted, and this demonstrates equilibrium with the trans conformations. Similarily, most points corresponding to trans arrangement in the crystal are well below the line, showing the presence of some gauche conformers.2600 calculated from An N.M.R.Study of Isomeric Pentitols The fraction 4 of each conformer in the trans conformation in the mixture can be The results of such a calculation are shown in table 3. The negative values for the arabinitol H,-H, conformer in both solvents result from small values and indicated the uncertainties in this type of calculation. Included in table 3 are the conformations found for the The results in table 3 clearly show that the pentitols in solution exhibit more flexibility than the rigid structure of the crystal. At a high-enough temperature one would expect the molecules to show free rotation about each bond and the entries in table 3 would each be 33 YO since there are two possible gauche conformations but only one trans. At the lower temperature used here the distribution is more complex because of solvation effects and the avoidance of unfavourable 0 I( 0 interactions. The average departure from the 33 YO trans conformation for each pentitol system is reported in table 4.Thus, for example, the average bond in ribitol in D,O has a conformation that departs by 28 % from the free rotation 33% value. Two effects can be distinguished in the results in table 4. In the non-aqueous solvent the pentitols have more free rotation than in D,O presumably due to decreased solvation effects. Of greater importance is the fact that, in D,O, xylitol shows significantly more free rotation than either ribitol or arabinitol. Thus, the lower partial molar volume and heat capacity for xylitol compared to ribitol and arabinitol is demonstrated to be a reflection of the conformation in solution and is not related to the crystal structure.Plans are underway to measure temperature dependences of the coupling constants reported here. We would like to acknowledge the considerable assistance of A. A. Bothner-By and the other members of the Biomedical Facility at Carnegie-Mellon University. We are also indebted to the Faculty Research Fund at CMU, and to NATO for financial assistance in support of this work. F.F. thanks the Royal Society and the NSERC of Canada for a visiting fellowship to the University of Waterloo. References 1 A. Suggett, in Water - A Comprehensive Treatise, ed. F. Franks (Plenum Press, New York, 1975), vol. 4, 2 R. S. Shallenberger, Advanced Sugar Chemistry (Ellis Horwood, Chichester, 1982).3 A. Cesaro, in Thermodynamic Data .for Biochemistry and Biotechnology, ed. H-J. Hinz (Springer, 4 F. Franks, Pure Appf. Chem., 1987, 59, 1189. 5 J. J. Savage and R. H. Wood, J . Phys. Chem., 1976, 10, 733. 6 I. R. Tasker and R. H. Wood, J . Solution Chem., 1982, 11, 469; 481. 7 Y-N. Lian, A-T. Chen, J . Suurkuusk and I . Wadso, Actu. Chem. Scand., Ser. A, 1982, 36, 735. 8 S. J. Angyal, Adv. Carbohydr. Chem. Biochem., 1985, 42, 1. 9 F. Franks, J. R. Ravenhill and D. S. Reid, J . Solution Chem., 1972, 1 , 3. 10 G. A. Jeffrey and H. S. Kim, Carbohydr. Res., 1970, 14, 207. 1 1 D. Horton and J. D. Wander, Carbohydr. Res., 1969, 10, 279. 12 S. J. Angyal and R. Le Fur. Carbohydr. Res., 1980, 84, 201. 13 G. Di Paola and B. Belleau, Can. J. Chem., 1977, 55, 3825. 14 G. Barone, B. Bove, G. Castronuovo and V. Elia, J . Solution Chem., 1981, 10, 803. 15 K. Bock and R. U. Lemieux, Carbohydr. Res., 1982, 100, 63. 16 C. A. Bush, Z-Y. Yan and B. N. N. Rao, J . Am. Chem. Soc., 1986, 108, 6168. 17 K. Watson, F. Franks and R. L. Lenkinski, unpublished results. 18 F. Franks, K . Watson, J. Dadok and R. L. Kay, to be published. 19 J. Dadok and A. A. Bothner-By, in N.M.R. and Biochemistry, ed S . J . Opella and P. Lu (M. Dekker, p. 519. Berlin, 1986), pp. 177-210. New York, 1979).F. Franks, Robert L. Ka-y and J. Dadok 260 1 20 J. Dadok and R. S . Sprecher, J. Mugn. Reson., 1974, 13, 243. 21 G. E. Hawkes and D. Lewis, J . Chem Soc., Perkin Tram. 2, 1984, 2073. 22 M. Karplus, J. Am. Chem. Soc., 1963, 85, 2870. 23 C. A. G. Haasnoot, F. A. A. M. De Leeuw and C. Altona, Tetrahedron, 1980, 36, 2783. 24 0. Jardetzky and G. C. K. Roberts, Molecular Biology (Academic Press, New York, 1981), p. 201. 25 H. S. Kim, G. A. Jeffrey and R. D. Rosenstein, .4cta Crystullogr., SNI. B, 1969, 25, 2223; H. S. Kim and G. A. Jeffrey. Acta Crystallogr.. Sect. B., 1969,25,2607; F. D. Hunter and R. D. Rosenstein, Acta Crystullogr., Sect. B 1968. 24. 1652. Paper 7/2076; Received 23rd November, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402595

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chem. SOC., Faraduy Trans. I, 1988, 84(8), 2603-2608 Conformation of Polyols in Water Molecular-dynamics Simulation of Mannitol and Sorbitol J. Raul Grigera University of La Plata, IFLYSIB, C.C. 565, 1900 La Plata, Argentina Mannitol and sorbitol are steriosomers that differ only in the position of one of their six hydroxy groups. In aqueous solution the properties of both are quite different, and it is generally assumed that the differences are due to specific interactions with water. In order to study the effect of solvent on the conformation of sorbitol and mannitol, a simulation by molecular dynamics of both polyalcohols, in uacm and in water was made. Mannitol was also simulated in an argon-like Lennard-Jones solvent. Simulation in vacuo showed that both polyols are bent, while in water it becomes clear that mannitol adopts a planar zig-zag configuration and sorbitol is bent.The resulting end-to-end distances are 0.635 nm for mannitol and 0.547 nm for sorbitol, with a radius of gyration of 0.417 and 0.372 nm, respectively. When mannitol is simulated in a Lennard-Jones solvent, starting from the extended configuration, it is observed that it quickly bends. It is concluded that the assumption that the configurations of sorbitol and mannitol are determined by solute-solvent interactions is well justified. From an analysis of water residence times a negative hydration is observed in both polyalcohols. The two steroisomeric polyalcohols mannitol and sorbitol differ only in the position of one of their six hydroxy groups, as is shown schematically in fig.1. Unlike what would be expected from their structural similarities, their specific interactions with water are different. It is accepted that in aqueous solution mannitol adopts a planar zig-zag conformation, whereas sorbitol adopts a bent-chain conformation. '9 Sorbitol would be a stronger 'structure breaker' than mannit01.l.~. Moreover, the solubility of sorbitol in water at 25 "C is 3.5 times larger than that of mannit01.~ In order to study the effect of solvent on the conformation of these substances we have undertaken a simulation by molecular dynamics of both polyalcohols in uacuo and in water. Mannitol was also simulated in an argon-like Lennard-Jones solvent. The Model The molecular model involves only atom-atom potentials, i.e.no torsional potentials are explicitly included. The interaction potentials of the OH groups are compatible with the water model used. The van der Waals parameters for CH, and CH, were taken from the literat~re,~ and those for 0 were coincident with those of the water model. Partial charges of both oxygen- and carbon-containing groups were selected to match neutrality when proton charges are those of water protons. The water molecules are taken as the SPC/ELP model [equivalent to the SPC/E1 model of ref. (6)], which is a modification of the well known SPC model7 that improves the radial distribution function and the diffusion coefficient when compared to the original SPC. Table 1 shows the actual parameters used in the simulation. The selected geometry was based on crystallographic information,' with a C-C bond length of 0.152 nm, a C-0 bond length of 0.143 nm, LCCC = 113" LCCO = 110".26032604 Conformation of Mannitol and Sorhitol OH OH Fig. 1. Fisher projection of (1) I CH2 I I CH2 I CH-OH CH-OH I I I CH-OH CH-OH I I I CH-OH HO-CH HO-CH HO-CH Table C 0 H 0,. H, CH2 I OH CH2 OH I ( 1 ) ( 2 ) mannitol and (2) sorbitol. Both polyalcohols differ only in the position of one hydroxy group. 1. Atom-atom interaction parameters for mannitol, sorbitol and water" c 0 0, H H, Q/e C,,/10-9 r n l 2 kJ mol-.' 9944 5322 5322 0 0 0.09 5322 2848 2848 0 0 -0.527 5322 2848 2848 0 0 -0.847 0 0 0 0 0 0.437 0 0 0 0 0 0.437 C6/ m6 kJ mol-I 9096 4879 4879 0 0 4879 2617 2617 0 0 4879 2617 2617 0 0 0 0 0 0 0 0 0 0 0 0 ~ _ _ _ - _ _ _ a Both CH and CH, groups have the same interaction parameters and are indicated by C ; the subscript w denotes water.Computational Procedure Molecular-dynamics simulations were performed in a cubic box of 1.8662 nm average length with periodic boundary conditions. The equations of motion were integrated using a leap-frog algorithm with At = 0.002 ps in such a way that temperature and pressure are kept constant by coupling the system with a heat bath and hydrostatic pressure system. The simulations were carried out with the GROMOS package on the VAX 11/750 computer of the IFLYSIB. All data reported are from runs made after at least 5 ps simulation time for equilibration. Reference temperature was 298.16 K and pressure 1 atm.? t 1 atm = 101 325 Pa.J . R. Grigera 2605 O 7 t 0.4 1 I I I 0 3 12 18 tips Fig.2. Time evolution of the end-to-end distance (after equilibrium) for mannitol(0) and sorbitol (0). 0 7 E . 8 Em u" u" Y .I 0.6 I 0.5 I I I I 10 20 30 tips Fig. 3. Changes in the end-to-end distance of sorbitol in water during simulation starting from the planar extended configuration. Results In the first step mannitol and sorbitol were simulated in vacuo, starting form a planar zig-zag configuration. A 20 ps simulation run, after equilibration, showed that the conformations of both polyalcohols without solvent do not differ significantly. For example, the carbon end-to-end distance (as averaged over 20ps) is 0.53 nm for mannitol and 0.55 nm for sorbitol. Although the end-to-end distance does not fluctuate more than 10% along the run, the fluctuation in the conformations is high, as can be followed by monitoring the dihedral angle. Since the end-to-end distance may not be a representative parameter with which to determine the conformation of a small molecule, we compute the mean radius of gyration, which results in 0.368 for mannitol and 0.381 nm for sorbitol. The final coordinates and a planar zig-zag conformation were used in both cases as starting coordinates for the simulation of the solution, with the addition of 205 SPC/ ELP water molecules for sorbitol and 203 for mannitol.A single solute molecule was simulated in a periodic box, resulting in an approximate equivalent concentration of2606 Conformation of Mannitol and Sorbitol d Fig. 4. Snap-shot of an equilibrium situation of the systems sorbitol-water (a) and mannitol-water (b).One can clearly see the bending conformation of sorbitol and the planar zig-zag conformation of mannitol. Only a few water molecules are shown. 0.256 mol dm-3 referred to the amount of water present. However, since there is only one solute molecule in the box, this is equivalent to an infinitely dilute solution. Note that the inclusion of a cut-off radius avoids the interaction between a particle and its own image. After the equilibration run, the configurations of the following 20 ps were collected for averaging. The resulting end-to-end distances are 0.635 nm for mannitol and 0.547 for sorbitol, with a radii of gyration of 0.41 7 and 0.372 nm, respectively. In fig. 2 the changes in the end-to-end distances along the run are plotted for the two polyols.Although the difference in the average value is only ca. 14% the fluctuation is not as high, so the difference can be considered as significant.J . R. Grigera 2607 0.4 0 3 12 18 tips Fig. 5. Changes in the end-to-end distance of mannitol during simulation. Starting from the bent conformation for both solutes (the one resulting from the simulation in vacuo) very slow changes are observed for mannitol, and no noticeable changes are produced in sorbitol within the relative long simulation time. It may be that the conformation is trapped in a relative minimun and remains there for the simulation period. This situation may induce us to think that sorbitol has a bent conformation in water merely because of a simulation artifact.To check this, we also start from the planar zig-zag conformation in water. For mannitol, as expected, almost no changes were observed, whereas for sorbitol a continuous change in the end-to-end distance was detected until it reached the value obtained previously. The time evolution of the ‘shrinkage’ of sorbitol is shown in fig. 3 . In order to compare the hydration properties of both compounds the coordination numbers and hydration lifetimes were calculated. The coordination number for mannitol is 13.23 water molecules, while for rnannitol it is 11.45. This is an expected result : mannitol is more extended, so that more water molecules will be near the polyol, whereas for sorbitol the bending conformation will diminish its surface area.The operational definition used for the coordination number is simply the average number of molecules that are within 1.2 water molecular diameters from any atom of the solute. Different atoms have their own coordination number that are by no means equal. The hydration lifetime will be more representative of the hydration properties than a mere counting of water molecules around the solute: therefore we have computed the parameter R, defined as R = zJz, ( 1 ) where zi is the average lifetime of a water molecule near the polyalcohol and is z, the average lifetime of a water molecule near another water molecule. Both substances have R < 1, which defines a negative h y d r a t i ~ n , ~ but the values differ : whereas for mannitol R = 0.80, for sorbitol R = 0.39.Fig. 4 shows the final configuration of sorbitol and mannitol, including some of the water molecules in the environment. Note that they are only snapshots, and cannot be considered as conformational average.2608 Conformation of Mannitol and Sorbitol It became clear that it is the solvent that determines the conformation of mannitol in water. We may ask ourselves if any solvent will produce the same effect, and as a test we performed a simulation of mannitol in a solvent of the same diameter as the water molecule and acting with the same Lenard-Jones potential selected for the water oxygen, which constitutes an argon-like molecule. The result can be clearly seen in fig. 5. Mannitol also ‘shrank’ reaching its in vacuo end-to-end distance. Discussion The simulation shows that the assumption that the configuration of sorbitol and mannitol is determined by the solute-solvent interaction is well justified.Specific interactions are the factors that decide between a linear or bent chain. Moreover, it is concluded that both solutes have negative hydration. If we derive, as is commonly (and wrongly) done, structural implications form dynamic results, the ‘ structure breaker ’ classification of both polyalcohols is well justified, although sorbitol produces a larger disruption in water structure. It appears that simulation does not incorporate much new information : however, note that by means of an artificial ‘experiment’ one can confirm indications that cannot be fully checked through actual experiments. It is shown here that the conformations of mannitol and sorbitol in aqueous solution are similar to those observed in the crystalline state,’ and that the conformation in solution is dictated by the characteristics of the solvent.I thank Professor F. Franks for his interest in this work and continuous encouragement. The help of Tomas Grigera in writing some computer programs is also acknowledged. I thank Professor H. J. C. Berendsen and Professor W. van Gunsteren for granting permission to use the GROMOS package. This work was partially supported by the Organization of the American States and the Consejo Nacional de Investigaciones Cientificas y Tecnicas of Argentina (CONICET). I am a member of the Carrera del Investigador of CONICET. References I G. Di Paola and B. Belleau, Can. J . Chem., 1977, 55, 3825. 2 E. M. Arnett and Jeffry, J . Solution Chem., 1973, 2, 114. 3 D. R. Wilson and Wen-Yang Wen, J . Phys. Clzem., 1976, 80,413. 4 F. J. Kelly, R. A. Robinson and R. H. Stokes, J . Phys. Chem., 1961, 65, 1958. 5 J. Hermans, H. J. C. Berendsen, W. F. van Gunsteren and J. P. M. Postma, Biopolymers, 1984, 23, 6 H. J. C. Berendsen, J. R. Grigera and J. Straatsrna, J . Phys. Chem., in press. 7 H. J. C. Berendsen, J. P. M. Postma, W. F. van Gunsteren and J. Hermans, in Intermolecular Forces, 8 G. A. Jeffrey and H. S. Kim, Carbohydr. Res. 1970, 14, 207. 9 0. Ya. Samoilov, Structure of Aqueous Electrolyte Solutions and the Hydration of Ions (Consultants 1513. ed. B. Pullman (Reidel, Dordecht, 1981). Bureau, New York, 1965). Paper 71208 I : Received 23rd November, I987

ISSN:0300-9599    

DOI:10.1039/F19888402603

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1988, 84(8), 2609-2618 Influence of Water on Pure Sorbitol Polymorphism S. Quinquenet, C. Grabielle-Madelmont and M. Ollivon Organisation Moleculaire et Macromoleculaire, CNRSI ER286, BP28, 94320 Thiais, France M. Serpelloni" Roquette Freres, 62136 Lestrem, France The behaviour of very pure sorbitol and sorbitol-water mixtures has been examined in relation to water content by both differential scanning calorimetry (d.s.c.) and a new thermal and dielectric analyser operating at microwave frequencies. A new hydrate has been isolated. Four different crystalline forms have been obtained from this hydrate and characterized by d.s.c. and X-ray diffraction. The influence of both temperature and water content on transitions from hydrate to other polymorphs has also been studied.Different types of water interactions have been evidenced from dielectric measurements for sorbitol and glucose, and comparisons of these, according to the physical states of the mixtures and hydrate properties have been made. For both molecules, the influence of water content on dielectric relaxation maximum temperatures has been studied and compared with the viscosity or the entropy. Non-congruency and possible deviation from stoichiometry are suggested for the new sorbitol hydrate. The dielectric properties of mannitol and sorbitol diastereoisomers have been measured at room temperature and compared. The glass transitions of sorbitol and sorbitol solutions have also been investigated down to 170 K by calorimetry measurements.The lowering of Tp with respect to water content is compared with isoviscosity curves. In the solid state D-sorbitol, which is an acyclic polyol obtained by hydrogenation of D-glucose, exhibits monotropic complex polymorphism. Several polymorphs of sorbitol with different melting points are described in the literature but with neither agreement on number, nomenclature and properties nor purity and origin de~criptionl-~. A vitreous state is also observed.4 Recently, starting from ultrapure sorbitol, we described the evidence for four different crystalline forms characterized by X-ray diffraction, i.r . spectroscopy and d.s.c. meas~rements.~ The present study is concerned with sorbitol polymorphism and its relationship with water. We briefly describe procedures used to obtain a well defined, new hydrate and four different polymorphs.The relationship with water content was examined by d.s.c. and a new thermodielectric method at a microwave frequency of 2.43 G H z . ~ , ~ The behaviour of these various crystalline species, of their metastable melts and aqueous solutions were compared with those observed for glucose and mannitol. Experiment a1 Ultrapure sorbitol (purity 99.9 YO) was obtained by successive chromatography on Neosorb 70/02 (Roquette) on a simulated fluidized bed Duolite C204-type resin (Roquette, French patent 2 454 830) and by recrystallization from purified aqueous solution (70% by dry weight). The purity of sorbitol was checked by h.p.1.c. (Waters apparatus using a Biorad Q15S-type column and detection by refractive index).It crystallized into thin needles (ca. 1 mm long), the atomic composition of which, determined by microanalysis, corresponded to a non-stoichiometric hydrated form described below. 26092610 Influence of Water on Pure Sorbitol Polymorphism Calorimetric measurements were carried out on a DSC4 Perkin-Elmer differential scanning calorimeter equipped with a Cryoson cooling system. The apparatus was calibrated for temperature and enthalpy measurements with lauric and benzoic acids and cyclohexane at rates of 4, 8, 20 and 40 "C mi^--'.^ Transition temperatures (T,) were taken at the onset of the transition. Special thermal treatments including isotherms were achieved either on Perkin-Elmer or Thermanalyse calorimeters. All water-sorbitol compositions are expressed in % water per 100% dry, pure sorbitol (w/w).Dielectric measurements at microwave frequency were measured with a new thermal analyser.' With this apparatus, complex permittivities of sorbitol or glucose were determined as a function of temperature on samples in the form of crystalline powders or melts placed in special Teflon sample-holders and isothermally on aqueous solutions in capillary tubing. With crystalline powders, in which measured values were less precise than with bulk materials since they depend to a small extent on the powder compactness, measured values are given in arbitrary units. With this technique, the dielectric constants E' and E" of water-sorbitol and water- mannitol mixtures, melts and crystalline sorbitol or glucose were measured using the small perturbation method5v6 of TE 0, 1, 3 (for crystals or melts) and TE 0, 1, 11 (for solutions or vitreous states) rectangular cavities excited at 2.43 GHz.Input power ranged from 6 W to a few mW according to the type of experiment, low power being used for isothermal measurements in the larger cavity. In these last experiments, each value is the average of 20 measurements, obtained by comparison with dielectric standards. The standards, measured at 21 "C, were chosen to maintain similar perturbations of the cavities. They were, respectively, decanol (E', 2.67; E", 0.41 3) for the small cavity and methanol ( E ' , 21.3; d', 13.8) and water (E', 78.15; E", 10.17) for the larger. When required, and especially with mannitol, the preheated capillary tubing was filled with hot and premelted mixtures, then sealed and rapidly cooled prior to measurement in order to keep the rapidly crystallizing mixtures in the metastable undercooled state. Obtention and Characterization of Sorbitol Polymorphs A, B and r polymorphic forms? were obtained from the hydrated form of sorbitol by the following thermal treatments.The A-form resulted from 80 days dehydration using phosphorus pentaoxide or from storage at 50 "C for 65 h in a closed sample-holder. B was obtained from the hydrate by dehydration at 50 "C down to 1 YO water, cooling to room temperature followed by complete melting at 120 "C, then a 72 h crystallization at 62.5 "C. The r form was obtained from the A form after 12 h under vacuum at 85 "C. The maximum water loss reached 6.3 '/o in this last case. The solidified melt was obtained by cooling the undercooled liquid resulting from any stable polymorph melting, down to room temperature and storing the cooled material for a few days.However, the cooling of this undercooled liquid leads to a glass (see below). The calorimetric heating scans of polymorphs of pure and hydrated sorbitol are reported in fig. 1. In order to simplify the results and the presentation, and because of the importance of water in sorbitol crystallization, the hydrate will be reported here as a polymorph. Thermal data, with i.r. and powder X-ray diffraction recordings of the 5 polymorphs3 confirm the existence of several polymorphic species and allow their identification. X-Ray patterns of the I-, A and B forms are identical with those described by Park and Jeffrey.' The hydrated form displays a powder pattern which differs from the other p~lymorphs.~ As far as we know, the existence of this species has not been reported in the literature.t Notation according to Park.'S. Quinquenet et al. 261 1 n (1) " I v) c i 2 0.5- 2 3 t 50 60 70 80 90 100 T/"C O40 Figure 1. Calorimetric scans of the sorbitol polymorphs (scan rate: 8 "C min-'), 1, solidified melt (1 ); 2, hydrate (1.3); 3, B-form (3); 4, A-form (3); 5, r-form (3); the numbers given in parentheses are y-scale expansion factors. The structure determination from hydrate monocrystals was underway when this paper was being prepared. Since then, the hydrate structure has been completely determined from X-ray diffraction measurements and will be published elsewhere.' The first results, before any ,refinements,, indicated a, triclinic unit cell, with the following parameters: a = 16.39 A; b = 8.!l A; c = 4.81 A; a = 94.14'; p = 95.78"; y = 100.95' and a unit cell volume of 679.84 A3.The hydrate densities both measured and calculated from the above data are 1.481 and 1.458, respectively. These data, as well as weight losses after thermal treatment (6.3 %), which suggest 2 water molecules are associated with 3 sorbitol molecules per unit cell, are now confirmed.' Water Interactions in Crystalline or Molten Sorbitol Interactions of water with crystalline or molten sorbitol have been examined at low water concentration (0-1 2 Yo) by measuring the complex permittivity dependence on temperature at 2.43 GHz.For simple dipolar relaxation, measurements during heating at fixed frequency are equivalent to frequency-sweeping at constant temperature. Anhydrous crystalline (whichever the polymorph considered) and hydrate sorbitol samples display very low losses and therefore were only examined at room temperature. Thin layers of hydrate samples maintained in 90 O/O RH adsorb 6-10 O h water in a few hours. Water adsorption was controlled both by weight increase and d.s.c. melting measurements. In these conditions, dielectric losses (e.g. E" 0.16) sometimes occur, resulting in self-heating to above the hydrate melting temperature. The dielectric behaviour is dependent on the adsorption rate. Slow adsorption leads to super-hydrated crystals with low dielectric losses (e.g.E" 0.002). Although hydration homogeneity is difficult to establish, it is thought that these samples were equilibrated. Pure hydrate samples kept under a dry atmosphere in order to remove part of their constituent water, whatever their final content also display low dielectric losses (e.g. E" 0.003). Therefore, the hydrate seems to accept positive or negative non-stoichiometry and only unequilibrated adsorbed water is responsible for dielectric losses (e.g. E" 0.98). This dielectric behaviour, which will be compared with dextrose monohydrate below, and the 86 FAR 12612 Influence of Water on Pure Sorbitol Polymorphism sorbitol (%) 30 20 E M (0) 10 0 1 30 20 E l ) (0) 10 0 0 20 LO 60 80 100 mannitol (%) Figure 2.Dielectric constant variations at 22 "C and 2.43 GHz: (a) sorbitol-water mixtures as a function of sorbitol content; (b) mannitol-water mixtures as a function of mannitol content. thermal properties previously recorded3 (fig. 2) indicate that this hydrate is non- congruent and stoichiometric, with possible non-stoichiometry. In the melted state hydrated (6.3%) or anhydrous sorbitol has more important dielectric losses, resulting in dielectric heating and indicating that sorbitol molecules interact relatively more weakly in the molten than in the crystalline state, whatever the water content. The variations in dielectric constant us. temperature are presented in fig. 3. No water losses were observed when the samples were heated. Maximum dielectric losses were observed for anhydrous (E" 0.31) and hydrated (E" 0.27) samples at 85 and 65 "C, respectively.S. Quinquenet et al.- ........................................ f ............................... ... (2)- T - - (3)- - ...... 1.-. .............. - .... ..- ........ E" -... - 1 I I I I I I ' 0 0 - 261 3 140 120 0 100% 80 6o 40 20 5 4 3 E l 2 1 0 t (bl 4 - 0' e' 3 - /---- - Ell 2 - // /' 1 - I - 1 e l 1 0.5 - 0 Figure 3. (a) Kinetics of sorbitol at 2.43 GHz; (b) and hydrated (6.3 heating of anhydrous (-) and hydrated (.....) (6.3 YO water) molten dependence of dielectric constant on temperature for anhydrous (-) YO water content) (----) molten sorbitol samples at 2.43 GHz. Although the temperature dependence of dielectric losses are often considered as more difficult to interpret than frequency plots, both determinations are c~mplementary.~.lo Moreover, for simple relaxations, maxima are observed by both techniques (frequency and temperature scanning) at the same values, provided no thermal event occurs in the explored temperature range. As multiple relaxations are expected for sorbitol," temperature spectrum deconvolution is not possible. However, the relative position of the two maxima can be compared together and with those of the sorbitol-water mixtures (see below). The strong negative temperature dependence of these maxima on the increased water content has already been observed for viscosity and dielectric relaxation of polyols near their glass-transition temperature" and can be related to changes induced by water insertion in the hydrogen bond distribution and to the associated decrease in viscosity.Therefore, water insertion in the undercooled liquid 'structure' can be considered either as plasticization or loss of structure of sorbitol. 86-22614 Influence of' Water on Pure Sorbitol Polymorphism In their careful analysis of carbohydrate dielectric relaxations, Tait et a1.l' separated three relaxation times for glucose solutions which were linearly dependent on solution viscosity. The relaxation times (7) of these substances cs. temperature ( T ) obey simple Vogel-Tammann-Fulcher (VTF) equations where A and B are constants and T, is a temperature close to the temperature at which the liquid entropy curve extrapolated below Tg would intersect the crystal entropy curve.BIT, can be approximated to 12.7.1° The 85 "C value observed for pure sorbitol at 2.43 GHz is ca. 20 "C above the expected value deduced from VTF equation. The 0.6 YO water retained by sorbitol used in ref. (10) cannot account for such a difference since T = 65 "C was found for the relaxation maximum of the 6.3% water sample. Nevertheless, addition of the value obtained in the microwave region to those previously reported at lower frequencies keeps the same good correlation coefficient ( r = 0.986), showing no excessive departure from VTF law at high temperature and frequency. Comparison with Glucose and Glucose-Water Mixtures Since the properties of glucose are well known4, 11-14 and because it forms a hydrate, the dielectrical properties of glucose and glucose-water mixtures were studied.Glucose samples were submitted to the same treatments and examined under the same experimental conditions for comparison with the results observed for sorbitol and its new hydrate. Both pure anhydrous glucose and glucose monohydrate undergo very slight losses at 2.43 GHz as observed for sorbitol and sorbitol hydrate. However, monohydrated or anhydrous glucose in the molten state, as previously observed for sorbitol, are subject to substantial dielectric losses. Fig. 4 (a) and 4(b) represent, variations in dielectric constant during dielectric heating as a function of time and temperature, respectively. This similarity between the properties of the two substances is also observed in the hydration of the monohydrate beyond its stoichiometry as shown in table 1.These results show that water exerts a similar effect on both sorbitol and glucose. At room temperature, these dielectric effects of water on glucose have previously been observed by Abadie et al.14 The very weak losses observed for glucose monohydrate, as well as for the new sorbitol hydrate, show that this water must be considered as a constituent. This type of water is easily distinguished from the sorbed water which displays higher losses. Note also that the maximum relaxation frequency of glucose melts are observed at higher temperatures than for sorbitol. This could be related to stronger hydrogen bonds between glucose molecules. Thus, the thermodielectric analyser used in this study for sorbitol and glucose, although less informative than frequency measurements, can be considered as an alternative method for rapidly obtaining essential information about states of water.Dielectric Behaviour of Sorbitol and Mannitol Solutions Dielectric properties of sorbitol and mannitol solutions have been examined at 22 "C and 2.43 GHz and the measured dielectric constants us. sorbitol and mannitol concentration are reported fig. 2(a) and 2(b). For both, the permittivity decreases as the polyol concentration increases. The corresponding losses initially increase up to ca. 50 % polyol then decrease sharply at ca. 65 %. This behaviour is similar to that of solutions of glucose or sucrose, but different from that of glycerol measured at 1 and 3 GHz.13 The high E" values observed for ca. 1 : 1 mixtures have been explained by water hydrogen- bond stabilization by the polyol hydroxy groups.13 These correspond to the relaxation frequency maxima, since 2.43 GHz is intermediate between pure water (17 GHz) andS .Quinquenet et al. 5-0 L . o - 3.0 E l 2.0 1.0. O20 261 5 * ( b ) * 1.0 //’ 8 . 0.8 . .0.6 //’ f l ) . . 0.4 - 0.2 0 50 100 150 5.0 - 4.0 - 3.0 - Figure 4. (a) Kinetics of heating of anhydrous (----) and monohydrated (-) molten glucose at 2.43 GHz; (6) dependence of dielectric constant on temperature for anhydrous (----) and monohydrated (--) molten glucose at 2.43 GHz. sorbitol (1 KHz) or mannitol relaxations. Moreover, the E” maximum value of ca. 22, corresponds to an average of the free water and sorbitol or mannitol loss maximum values. Finally, despite marked physico-chemical differences, such as in the respective mannitol and sorbitol solubilities (sorbitol is 3.5 times more soluble in water than niannitol), and in the less marked differences, e.g.the hydrodynamic volumes, solution viscosity and hydration,15-16 the solution dielectric behaviour of these diastereoisomers is similar.261 6 Influence of Water on Pure Sorbitol Polymorphism Table 1. Qualitative comparison of the dielectric properties of sorbitol and glucose in the microwave domain (2.43 GHz)." molten hydrate hydrate slow rapid product stoichiometry anhydrous anhydrous hydrated hydration hydration sorbitold + - b glucosed - + + + + a +, dielectric losses are high enough to allow self-heating (see fig. 3 and 4); -, low dielectric losses, self-heating is less than 10 "C.* Depending on the extent of hydration and the presence or absence of sorbed water (see text). 'Weak with evaporation. Dextrose monohydrate is a Roquette Co. industrial product with traces of impurity; sorbitol hydrate is pure with a 3:2 sorbitol : water stoichiometry. Influence of Water on Sorbitol Glass Transition The influence of water on the glass transition temperature (T,) of sorbitol which has been investigated down to 170 K by d.s.c., is shown fig. 5. T, depression as a function of water content was expected since water reduces glass viscosity and acts as a plasticizer of sorbitol17 by increasing the free volume and hydrogen-bond exchange possibilities. Moreover, the decrease in Tp agrees with that expected for binary mixtures,18 assuming the Tg for pure water is probably close to - 140 OC.19 It can be considered here as the 'copolymer ' of a water-sorbitol 'ideal mixture'.At the glass transition point, water-sorbitol mixtures, whatever their water concentrations, reach the same critical vis~osity.~' Thus, it is not surprising that the curve in fig. 5 is similar to the isoviscosity curve observed by Angel1 et aL4 Levine and Slade20 have found, by extrapolation of the water crystallization peak enthalpies to zero, TL and W; values, which are the glass-transition temperature and concentration limits of the glassy domain, respectively, at -43.5 "C for 0.23 g incongealable water per g sorbitol. These values are in good agreement with our data (AT = 12 "C) taking into account the different methods of determination and sorbitol purities.T, determinations for pure sorbitol vary within ca. 10 "C depending on the method and conditions chosen along with, often unknown, variations in the effective water content of the sorbit01.~ Delayed water crystallization has also been observed for sorbitol-water mixtures. This well known phenomenon, which is generally considered as a devitrification process2' and which has already been observed for glucose,21 glycerol22 and other sugar solutions, occurs for sorbitol on heating at a few tens of degrees above the glass-transition temperature with water-rich mixtures, over WL, e.g. for >, 70% water. Over the domain in which water crystallization is possible and below the hydrate water concentration there is a zone in which water is rendered incapable of freezing by the high sorbitol concentration, often referred to as 'unfreezable or unfrozen water '.17 This water is often wrongly called 'bound water',23 despite the lifetimes of the water-substrate bonds being shorter than in bulk water (in the ps range).Conversely, the high water concentration cannot be considered as inhibiting sorbitol crystallization, since sorbitol self-inhibits its own crystallization (it exhibits an ' undercooling ' temperature of ca. 100 "C3 and, by lowering the mixture viscosities, water favours the crystallization of sorbitol hydrate and polymorphs. However, both water and sorbitol act as 'structure breakers' to each other since strong hydrogen bonds inhibit their respective separations and crystallizations and thus do not permit phase separation.Indeed at low temperature, close to q, supersaturated solutions become so viscous thatS. Quinquenet et al. 2617 N J 0 10 20 30 40 50 60 water (YO) Figure 5. Effect of water content on the glass transition temperature Tg of pure sorbitol. N is the average number of water molecules per sorbitol molecules in aqueous mixtures. any crystallization is inhibited. In this domain, the glass transition must correspond to a blocking of sorbitol and water molecules leading to a glassy, metastable structure exhibiting only very short-range order.20 Thus, the relaxation peaks following glass transitions (data not shown) were attributed to the melting of this weakly organized part of the glass.3 Continuous depression observed as a function of increased water content indicates, as well as the dielectric behaviour of water-sorbitol mixtures shown before, that water and sorbitol are co-participating in the glassy state (fig.5 ) . Therefore, both this co-participation of sorbitol and water in the glassy state and in the undercooled liquid mixture, along with high viscosity and the existence of a maximum dielectric relaxation for pure molten sorbitol indicate that a complete analysis of the dielectric relaxations of water-sorbitol mixtures, which is not possible with data determined us. temperature, should not only take into account possible water relaxations but also possible sorbitol relaxations. At high sorbitol concentrations (0-70 % water in sorbitol-water mixtures), three kinds of crystallization are possible depending on the relative proportions, temperature and pressure,24 namely water, hydrate and pure sorbitol polymorph crystallizations.Water crystallization may occur at high water content and low temperature. The polymorphs crystallize only at very low water concentration and high temperature. Hydrate crystallization occurs for intermediate values. Therefore, influence of medium viscosity on crystallization is evident, and in this respect sorbitol behaves exactly as other sugars and polyols. However, the influence of water on viscosity is variable : increase in water concentration leads to decreased viscosity, when its influence on crystallization processes is different and depends on whether water or sorbitol is crystallizing. Other parameters influence crystallizations, such as the purity and thermal properties of the constituents, but these have not been studied to any great extent.From the investigation of solid-state sorbitol using a variety of techniques, the occurrence of four different crystalline forms has been confirmed, including a hydrate a n d at low temperature, a vitreous state which leads to a particularly complex polymorphism. The characterization of the hydrate, its two water and three sorbitol2618 Influence of Water on Pure Sorbitol Polymorphism molecules per the unit cell, complicates the study of its relationship with water. However, the general behaviour observed for sugar and polyol solutions, melts and vitreous states, which is mainly the dependence of properties on the viscosity of the system, is also observed for sorbitol.It has been shown that the use of our new thermal and dielectric method helps in this respect. The examination of water-sorbitol mixtures, by the way of calorimetry and dipolar relaxation in the microwave domain, proves the existence of different types of water-sorbitol interactions depending mainly on the physical and thermodynamic state of the system i.e. whether crystallized, in solution (stable), melted or vitreous (metastable). S. Q. was supported by a joint grant CNRS-Roquette Freres Company. We thank Mrs Gillier and Mrs Neuman (Universite Paris-Nord) for the determination of the hydrate structure, and Mr Comini (Roquette Freres) for i.r. spectra. References 1 Y. J. Park and G. A. Jeffrey, Acta Crystallogr., Sect.B, 1971, 27, 2393. 2 J. Sztatisz, S. Gal, L. Fodor and E. Pungor, J . Therm. Anal., 1977, 12, 351. 3 S. Quinquenet, M. Ollivon, C. Grabielle-Madelmont and M. Serpelloni, Thermochim. Acta, 1988, 125, 4 C. A. Angell, R. C. Stell and Z . Sichina, J . Phys. Chem., 1982, 86, 1540. 5 M. Ollivon, IEEE-Microwaves Theory and Techniques Int. Symp. (St Louis, USA, 1985), p. 645. 6 M. Ollivon, S. Quinquenet, M. Seras, M. Delmotte and C. More, Thermochim. Acta, 1988, 125, 141. 7 M. Ollivon and R. Perron, Thermochim. Acta, 1982, 53, 183. 8 H. Gillier and S. Quinquenet, in preparation. 9 P. Sixou, P. Dansas and D. Gillot, J . Chern. Phys., 1967, 64, 834. 125. 10 C. A. Angell and D. L. Smith, J . Phys. Chern., 1982, 86, 3845. 11 M. J. Tait, A. Suggett, F. Franks, S. Ablett and P. A. Quickenden, J . Solution Chem., 1972, I , 12 J. B. Hasted, in Aqueous Dielectrics (Chapman and Hall, London 1973). 13 B. D. Roebuck and S. A. Goldblith, J . Food Sci., 1972, 37, 199. 14 P. Abadie, R. Charbonniere, A. Gidel, P. Girard and A. Guilbot, J . Chim. Phys., 1953, 50, 46. 15 A. Mayaffre, R. Bury and M. Chemla, J . Chim. Phys., 1976, 83, 637. 16 J. R. Grigera, J . Chern. SOC., Faraday Trans. I , 1988, 84, 2603. 17 F. Franks, Properties of Water in Food.y, ed. D. Simatos and J. L. Multon (NATO Asi Series, 1985), 18 A. J. Kovacs, Fortschr. Hochpolym. Forsch., 1963, 3, 394. 19 D. R. Macfarlane, Cryo-letters, 1986, 7, 136. 20 H. Levine and L. Slade, Water Science Rez;ieu.s, ed. F. Franks (Cambridge University Press, 1987), vol. 3, 21 Water, A Comprehensive Treatise, ed. F. Franks (Plenum Press, New York, 1982), vol. 7. 22 D. Simatos and M. Faure, in Water Relations of Foods, ed. R. B. Duckworth (Academic Press, New York, 1975), p. 193. 23 F. Franks, Cryo-letters, 1983, 4, 73. 24 T. Atake and C. A. Angell, J . Phys. Chem., 1979, 83, 3218. 131. p. 497. pp. 79-185. Paper 712077; Received 23rd November, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402609

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1988, 84(8), 2619-2633 Thermomechanical Properties of Small-carbohydrate-Water Glasses and ' Rubbers ' Kinetically Metastable Systems at Sub-zero Temperatures Harry Levine" t and Louise Sladet General Foods Corporation, Technical Center, Tarrytown, New York, U.S.A. Aqueous solutions of 24 small carbohydrates have been analysed by low- temperature differential scanning calorimetry (d.s.c.). The method, based on analogue derivative thermograms, measured two thermomechanical pro- perties characteristic of each non-crystallizing solute : Tg, the sub-zero glass transition temperature of a maximally freeze-concentrated solution and Wg, the amount of unfrozen water (UFW) kinetically immobilized in the glass at q. For 84 low-molecular-weight polyhydroxy compounds (PHCs) analysed to date, Tg ranged from -85 "C for ethylene glycol to - 13.5 "C for maltoheptaose, and increased monotonically with increasing M,.Tg plotted us. Mi1 showed a linear correlation characteristic of an homologous family of glass-forming linear oligomers. Wg ranged from 1.9 g UFW g-' for ethylene glycol to 0.2-0.3 g UFW g-I for several sugars and polyols, including maltoheptaose, and decreased with increasing q, showing fair linear correlations for several series of homologous solutes. We describe here the use of and Wi, as invariant physico-chemical properties of glass-forming solutions at sub-zero temperatures, to interpret thermo- mechanical behaviour of small-carbohydrate-water systems in non- equilibrium glassy and ' rubbery ' states, define structure-activity relation- ships and explain and often predict functional behaviour of such PHCs in various applications.Thermal and mechanical properties of concentrated aqueous solutions of small carbohydrates at sub-zero temperatures have been studied for decades. Nearly 50 years ago, Luyetl reported microscopically observed temperatures of irruptive ice re- crystallization (7J for frozen solutions of sugars and polyhydric alcohols. Today, it is recognized by some2-6 that the T, values of such glass-forming aqueous systems'. 7-14 coincide with a characteristic glass-transition temperature, q, measurable calori- metrically. This fact was demonstrated conclusively by our previous study5 of commercial starch hydrolysis products (SHPs), and is illustrated by the results reported here for low-molecular-weight PHCs.It is also recognized that ice re- crystallization is but one of many possible manifestations (referred to as ' collapse ' of the dynamically controlled behaviour of small-carbohydrate-water glasses and ' rubbers ' that exist, at sub-zero temperatures, in kinetically metastable ' states ' rather than equilibrium thermodynamic phases.6* 23-26 [As used here, 'rubbery' refers to the high-viscosity fluid state which exists at < T < crystalline melting temperature (T,) for any glass-forming liquid. In contrast, a true rubber corresponds to a ,viscoelastic network of entangled polymeric chains for a partially crystalline polymer at Tg < T < T,."] Collapse phenomena are diffusion-controlled structural relaxation processes which occur in response to changes in moisture content, temperature and f Present address: Nabisco Brands Inc., Corporate Technology Group, P.O. Box 1943, East Hanover, New Jersey 07936-1943, U.S.A.26192620 Small-carbohydrate- Water Glasses and ' Rubbers' Table 1. Tg and WL values for sugars, glycosides and polyhydric alcohols ~- solute ethylene glycol propylene glycol butane- 1,3-dioi glycerol erythrose threose erythritol thyminose (deoxyribose) ribulose xylose 1 arabinose lyxose ribose arabitol ribitol xylitol methyl riboside methyl xyloside quinovose (deoxyglucose) fucose (deoxygalactose) rhamnose (deoxymannose) talose idose psicose altrose glucose gulose fructose galactose allose sorbose mannose tagatose inositol manni to1 galactitol sorbitol 2-0-methyl fructoside p- 1 -0-methyl glucoside 3-0-methyl glucoside 6-0-methyl galactoside a-1 -0-methyl glucoside 1 -0-methyl galactoside 1 -0-methyl mannoside 1-0-ethyl glucoside 2-0-ethyl fructoside 1-0-ethyl galactoside 1 -0-ethyl mannoside glucoheptose manno heptulose glucoheptulose persei to1 (mannohepti tol) 1 -0-propyl glucoside 1 -0-propyl galactoside Mw ______ 62.1 76.1 90.1 92.1 120.1 120.1 122.1 134.1 150.1 150.I 150.1 150.1 150.1 152.1 152.1 152.1 164.2 164.2 164.2 164.2 164.2 180.2 180.2 180.2 180.2 180.2 180.2 180.2 180.2 180.2 180.2 180.2 180.2 180.2 182.2 182.2 182.2 194.2 194.2 194.2 194.2 194.2 194.2 194.2 208.2 208.2 208.2 208.2 210.2 210.2 210.2 212.2 222.2 222.2 7",/"C Wg/g UFW g-' - 85 -67.5 -63.5 - 65 - 50 -45.5 - 53.5 - 52 - 50 - 48 -47.5 -47.5 - 47 - 47 - 47 -46.5 - 53 - 49 -43.5 - 43 - 43 -44 - 44 - 44 -43.5 - 43 -42.5 - 42 -41.5 -41.5 -41 - 41 -40.5 - 35.5 - 40 - 39 -43.5 -51.5 - 47 -45.5 -45.5 -44.5 -44.5 -43.5 -46.5 -46.5 - 45 -43.5 - 37.5 - 36.5 - 36.5 - 32.5 - 43 - 42 1.90 I .28 1.41 0.85 1.39 (eutectic) 1.32 0.45 1.23 0.49 0.89 0.82 0.75 0.96 1.01 1.11 1.1 1 0.90 0.41 0.96 0.77 0.56 0.45 0.35 1.33 0.30 (eu tec tic) (eutectic) 0.23 1.61 1.29 1.34 0.98 1.32 0.86 1.43 1.35 1.15 1.26 1.21 0.77 1.22 1.05 (eutectic)H.Levine and L. Slade Table 1. (cont.) 262 1 1 -0-propyl mannoside 2,3,4,6-0-methyl glucoside isomaltulose (palatinose) nigerose cellobiulose isomaltose sucrose gentiobiose laminaribiose turanose mannobiose melibiose Iactulose maltose maltulose trehalose cellobiose lactose maltitol isomaltotriose panose raffinose maltotriose nystose stachy ose maltotetraose maltopentaose a-cyclodextrin maltohexaose maltoheptaose -~--___-~- 222.2 236.2 342.3 342.3 342.3 342.3 342.3 342.3 342.3 342.3 342.3 342.3 342.3 342.3 342.3 342.3 342.3 342.3 344.3 504.5 504.5 504.5 504.5 666.6 666.6 666.6 828.9 972.9 990.9 1 153.0 -40.5 -45.5 - 35.5 - 35.5 - 32.5 - 32.5 - 32 -31.5 -31.5 -31 - 30.5 - 30.5 - 30 - 29.5 - 29.5 - 29.5 - 29 - 28 - 34.5 - 30.5 - 28 - 26.5 -23.5 -26.5 -23.5 - 19.5 - 16.5 - 9 - 14.5 - 13.5 0.95 1.41 0.70 0.56 0.26 0.64 0.91 0.72 0.25 0.20 0.69 0.59 0.50 0.59 0.70 0.45 1.12 0.55 0.47 0.50 0.27 time.6 Their kinetics are governed by the free volume, mobility, and viscosity of a water- plasticized glass or ' rubber '.5 In the present study, aqueous solutions of 24 small carbohydrates were analysed by low- temperature d.s .c. When added to 60 other low-molecular- weight compounds analysed previously,6 these PHCs represent a comprehensive series of mono-, di-, and small oligo-saccharides and their derivatives, including many common sugars, polyols arid glycosides, covering an M , range of 62-1 153 dalton. The d.s.c. r n e t h ~ d , ~ based on analogue derivative thermograms, was used to measure two thermomechanical properties characteristic of each non-crystallizing solute : Tg, the particular sub-zero of the maximally freeze-concentrated solute-UFW amorphous matrix surrounding the ice crystals in a frozen solution;3,12.28 and Wk, the amount of water rendered 'unfreezable' (expressed as g UFW g-l solute) by immobilization with the solute in this kinetically metastable, dynamically constrained solid which forms on slow cooling to T < < .3 ' 5 We will demonstrate here how and Wk, invariant physico-chemical properties of glass-forming systems at sub-zero temperatures, can be used to describe the thermomechanical behaviour of PHC-water solutions in non-equilibrium glassy and 'rubbery' states. We will also show how these results can be used to define structure-activity relationships, which can be used in turn to explain and often predict functional behaviour of such carbohydrates, many of which are widely used in food applications.2622 Small-carbohydrate- Water Glasses and ' Rubbers ' Experimental The small PHCs characterized in this study are included in table 1 [updated from ref.(6)]. All the sugars and polyols were reagent-grade chemicals, most of them from Sigma. These materials were analysed as received. Most of the glycosides were synthesized and purified by our colleague, Dr Allan Bradbury. All solutions for T, and Wi determinations were 20 wt YO solids (i.e. 20.0 g solute per 100.0 g solution), prepared in distilled, deionized water. Samples were produced by mechanical stirring, with gentle heating when necessary, to yield clear solutions. D.s.c. measurements were performed with a DuPont 990 thermal analyser coupled to a DuPont 910 d.s.c. equipped with a liquid-nitrogen quench-cooling accessory capable of sample cooling at ca. 50 "C min-'. The analogue derivative function on the DuPont 990 allowed precise determination of transition temperatures, with a reproducibility (for duplicate samples) of & 0.5 "C for 7';.In practice, 20-30 mg of solution were hermetically sealed in an aluminium sample pan and scanned (against an empty reference pan) at a heating rate of 5 "C min-l, from a temperature at least 10 "C below 7*, to 25 "C. In all cases, initial cooling to well below ensured maximal freeze concentration and thus maximally frozen samples. Illustrative thermograms were shown, and the method for determining W; was described previously. Results and Discussion Low-temperature Thermal Properties of Aqueous Solutions of Small Carbohydrates The theoretical basis for the thermal properties manifested by water-soluble solids in solution at sub-zero temperatures has come to be well underst~od.~, 7-12* 16, 28 Assignments of characteristic transitions and temperatures have been reconciled definitively with state diagrams previously reported for various material^.^^ l2 In such diagrams [e.g.fig. 6 of ref. ( 5 ) for poly(viny1 pyrrolidone)] the different cooling-heating paths that can be followed by solutions of monomeric cs. polymeric solutes are revealed. However, in either general case, rewarming forces the system through a glass transition at 7';. In many earlier d.s.c. studies,". 16, 29 performed without benefit of derivative thermograms, a pair of transition temperatures (each independent of initial concentration), called T (antemelting), Tit, and T(incipient melting), Tim, were reported in place of a single 7';.In fact, in many cases, reported values of Tam3' and Ti, bracket that of 7*,. This led us5 to suggest that Tam and Tim actually represent the temperatures of onset and completion of the single thermal event (a glass transition) that must occur at Ti, as defined by the state diagram. Even today, many workers in this field do not recognize the sub-zero T p < T devitrification (7J < 7", sequence of transitions characteristic of frozen solutions of PHCs. Instead, they refer to 'anomalous double glass transitions ' manifested by aqueous solutions of, e.g. propylene glycol and glycerol.31 33 Far from being anomalous, for each solute, the higher Tp of the doublet coincides with Ti.6 Similarly, q, manifested as the onset temperature for opacity during warming of vitrified aqueous solutions and known to be independent of initial concentration, is still a topic of current interest and discussion as to its rigi in,^^-^* but it is not yet universally recognized to coincide with Ti for water-soluble carbohydrates, including several much- studied polyols.Our previous studies5. 6 , 2 3 9 24 have helped explain why Tg is the most noteworthy feature of low-temperature d.s.c. thermograms. The matrix surrounding the ice crystals in a maximally frozen solution is a supersaturated solution of all the solute in the fraction of water remaining unfrozen. This matrix exists as a kinetically metastable amorphous solid (a glass of constant composition) at any T < TL, but as a viscous liquid (a 'rubbery' fluid) at Tg < T < ice c,. Again with regard to a state diagram for a typicalH.Levine and L. Slade 2623 0 20 40 60 80 100 120 140 160 180 105 IM, Fig. 1. Variation of the glass transition temperature, q, for maximally frozen 20 wt % solutions against Mi1 for the sugars, glycosides and polyols in table 1. r = -0.934. solute that does not readily undergo eutectic crystallization, Tg corresponds to the intersection of an extension of the thermodynamically defined equilibrium liquidus curve and the kinetically determined supersaturated glass curve. As such, Franks12 described T, as representing a quasi-invariant point in the state diagram, invariant in both its characteristic temperature (Tg) and composition (i.e. Ci, expressed as wt YO solute, or Wk, expressed as g UFW g-l solute) for any particular solute.This glass, which forms, e.g. on slow cooling to Tg, serves as a kinetic barrier (of high activation energy) to further ice formation (within the experimental timeframe), despite the continued presence of UFW at all temperatures < T,, as well as to any other diffusion-controlled pro~ess.~ Recognizing this, one can begin to appreciate why the temperature of this glass transition is so important to aspects of frozen-food technology involving freezer-storage stability, freeze concentration and free~e-drying,~? all of which are subject to various recrystallization and collapse phenomena. Measured Tg values for 84 small carbohydrates are listed in table 1. As noted earlier, T, values for various solutes fall between previous literature values for cm and Tim,8 and within a few degrees of values reported for and the collapse temperature, T, (as shown later in table 3).T, values for this non-homologous collection of monodisperse sugars, polyols and glycosides ranged from -85 "C for ethylene glycol ( M , = 62) to - 13.5 "C for maltoheptaose (M, = 1 153). These results demonstrated a monotonic relationship between increasing T, and M,, that yielded a linear correlation (Y = -0.93) between T, and Mi', as shown in fig. 1. The major cause of scatter in this plot was the series of chemically different glycosides. In contrast, a plot of T, us. M, for glucose and malto- oligosaccharides of degree of polymerization (D,) 2-7 gave a smooth curve, with Y = -0.99 for the corresponding plot of T, us.The latter results, especially the theoretically predicted linear dependence of Tg on Mil, exemplified the glass-forming behaviour characteristic of a homologous family of n~n-entangling,~~ linear, mono- disperse 01igomers.~~* 35 In contrast, for a polydisperse mixture of solutes, the observed Tg is actually a weight-averaged 23, 26 For example, previous Tg results' for heterogeneous commercial SHPs showed a range from -43 "C for glucose [M, = 180, dextrose equivalent value (DE) = 1001 to -4 "C for 0.5 DE maltodextrin (number- averaged M,, @, x 3.6 x lo'), with T, = - 13.5 "C for maltodextrins of 15-20 DE (A%, = 900-1200). T, values in table 1 illustrate that glucose is representative of many other monosaccharides, while maltoheptaose is comparable to 15-20 DE maltodextrins.depends rigorously on linear, weight-averaged D, (DPw) for such polydisperse solutes, so that linear polymer Our previous study of polymeric SHPs demonstrated that2624 Small-carbohydrate- Water Glasses and ' Rubbers' Table 2. Dependence of on linear b,, for sugars sugar N-mer T,/C structure maltose cellobiose isomaltose gentiobiose nigerose laminari biose maltulose cello biulose isomaltulose lactose melibiose ma1 to triose panose isomaltotriose raffinose ma1 to tetraose stachyose ny stose 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 4 4 4 -29.5 - 29 - 32.5 -31.5 - 35.5 -31.5 - 29.5 - 32.5 -35.5 - 28 - 30.5 -23.5 - 28 - 30.5 - 26.5 - 19.5 -23.5 -26.5 glucose 1 -+ 4 glucose glucose 1 -+ 4 glucose glucose 1 + 6 glucose glucose 1 -+ 6 glucose glucose 1 -+ 3 glucose glucose 1 -+ 3 glucose glucose 1 -+ 4 fructose glucose 1 -4 fructose glucose 1 -, 6 fructose glucose 4 t 1 galactose glucose 6 +- 1 galactose glucose 1 -+ 4 glucose 1 -+ 4 glucose glucose 1 --+ 4 glucose 1 + 6 glucose glucose 1 -+ 6 glucose 1 -+ 6 glucose galactose 1 -+ 6 glucose 1 -+ 2 fructose glucose 1 -+ 4 glucose 1 -+ 4 glucose 1 -+ 4 glucose galactose 1 -+ 6 galactose 1 -+ 6 glucose 1 -, 2 fructose glucose I -, 2 fructose 1 -+ 2 fructose 1 -+ 2 fructose chains give rise to higher values than branched chains of equal Mw.6 As illustrated in table 2, from comparisons of the significant T, differences between maltose (1 -+ 4-linked glucose dimer), gentiobiose (1 -+ 6-linked) and isomaltose (1 -+ 6-linked), and between maltotriose (1 -+ 4-linked trimer), panose (1 -+ 4, 1 -+ 6-linked) and isomaltotriose (1 -+ 6, 1 -+ 6-linked), current results suggest that 1 -+ 4-linked (linear amylose-like) glucose oligomers manifest greater 'effective ' linear chain lengths in aqueous solution (and, consequently, greater hydrodynamic volumes) than oligomers of the same molecular weight which contain 1 -+ 6 (branched amylopectin-like) links.Table 2 also illustrates the intrinsic sensitivity of the Tg parameter to molecular configuration, in terms of linear chain length, as influenced by the nature of the glycosidic linkages in various non- homologous saccharide oligomers (not limited to glucose units) and the resultant effect on solution conformation. Another interesting comparison is between T, values for the linear and cyclic a-(1 -+ 4)-linked glucose hexamers, maltohexaose (- 14.5 "C) and a- cyclodextrin (- 9 "C).In this case, the higher Tg of the cyclic oligomer leads us to suggest that the ring of a-cyclodextrin apparently has a much larger hydrodynamic volume in solution (owing to its relative rigidity3') than does the linear chain of maltohexaose, which is relatively flexible and can assume a more compact conformation in solution. the composition of the glass at T, (as a weight of UFW per unit weight of solute), is one of several methods routinely employed in the food industry to determine the so-called 'water-binding capacity' of a solute. Franks39 12,289 37 h as reviewed this subject in depth, and pointed out repeatedly that so-called 'bound' water is not truly bound in any energetic sense.It is subject to rapid exchange,38 has thermally labile hydrogen 40 shows cooperative molecular mobility,41 has a heat capacity approximately equal to that of liquid water rather than ice40-42 and has some capability to dissolve Furthermore, it has been firmly e~tablished,~~ for water-soluble polymers and monomers alike, that such ' unfreezability ' is not due to tight equilibrium binding by but to purely kinetic retardation of diffusion of water and solute molecules at the low temperatures approaching the vitrification We showed previ~usly,~ for an homologous series of SHP solutes, that Wi decreases with increasing Tg, with a linear regression coefficient of -0.91. In other words, as The procedure used to calculate of a solute-UFW r n i x t ~ r e .~ ~ ~ 40H. Levine and L. Slade 2625 -10 - -20 -30 - - -40 - - 9 0 I / 1 0 10 20 30 40 50 60 70 80 90 100 CL (wt% solute in the glass at T i ) Fig. 2. Variation of the glass transition temperature, Tg, for maximally frozen 20 wt YO solutions against Ck, the composition of the glass at q, in wt YO solute, for homologous series of polyhydric alcohols (r = 0.83) (O), glucose-only solutes (Y = 0.73) (*), fructose-containing (Y = 0.96) ( x ) and galactose-containing saccharides (Y = 0.94) (0) in table 1. [Reproduced, with permission, from ref. (23).] solute(s) A?n increased, the fraction of total water unfrozen in the glass at Ti decreased. Again in the context of a typical state diagram, these results showed that as fin of a solute (or mixture of homologous solutes) in an aqueous system increases, the Tg- CI, point generally moves up the temperature axis towards 0 "C and to the right along the composition axis towards 100 wt o/o solute.The critical importance of this fact became clear when we described5 SHP functional behaviour v i s - h i s T,, and the capacity of inhibiting collapse processes in frozen solutions by formulating a system with polymeric SHPs in order to elevate Tg. Measured WI, values shown in table 1 ranged from 1.90 g UFW g-' for ethylene glycol to 0.20-0.30 g UFW g-l for several sugars and polyols, including maltoheptaose. In apparent contrast to the above-mentioned results for an homologous SHP series of mixed glucose monomer and oligomers, the 7';, us. Wk results for the diverse PHSc in table 1 yielded a regression coefficient of only -0.64.6 Thus, when Franks3 noted that, among ;he (non-homologous) sugars and polyols most widely used as 'water binders' in fabricated foods, 'the amount of unfreezable water does not show a simple dependence on M , of the solute', he was sounding a necessary caution.[In fact, when the Wi data in table 1 were plotted against Mi', Y = 0.47.1 We concluded that such plots could not be used for predictive purposes, so that the prudent approach would be to rely on measured WI, values for each potential ' water-binding ' candidate. However, the situation was not quite as nebulous as it first appeared. When some of the same data were plotted (actually T, us. CL, as shown in fig. 2"), and compounds were grouped by chemical classification into specific homologous series (e.g.polyols, glucose-only solutes, and fructose- or galactose-containing saccharides), better linear correlations became evident. The plots in fig. 2 illustrate the same linear dependence of Tg on the composition of the glass at Tg (i.e. as the amount of unfrozen, plasticizing water in the glass decreases,2626 Small-carbohydrate- Water Glasses and ' Rubbers ' Tg increases) as did the data for the homologous SHP solutes. Still, Franks' suggestion45 that future investigations of T, and viscosity as functions of solute concentration, and the liquidus curve as a function of solute structure would be particularly worthwhile is a good one. We noted with interest the Wb, results for the series of monomeric glycosides listed in table 1, in terms of a possible relationship between glycoside structure (e.g.position and size of the hydrophobic aglycone, which is absent in the parent sugar) and the function reflected by Wb. Wb values for all methyl, ethyl and propyl derivatives were much greater than those for the corresponding parent monosaccharides. However, Wb appeared consistently to be maximized for methyl or ethyl derivatives, but somewhat decreased for the n-propyl derivative. We suggest that these results could indicate that increasing the hydrophobicity (of the aglycone) leads to both decreasing Wi and the demonstrated tendency toward increasing insolubility of propyl and larger glycosides in water. As described elsewhere, 26 r, of aqueous solutions of low-molecular-weight PHCs does not depend solely on free volume and thus correlates more closely with the weight- averaged M , (M,) of the solute-UFW mixture in the glass of composition defined by Wb than with the corresponding an, or M , of the monodisperse dry solute.For those PHCs in table 1 with measured values of T, and Wi, a plot of T, us. M;' showed a better linear correlation (r = -0.95) than the plot of Tg us. Mi' in fig. 1 (r = -0.93). In contrast, T, plotted us. Mi' gave an r value of only - 0.71. For the homologous series of malto-oligosaccharides up to DP 7, 7", plotted us. Mi', Mil, and a;' gave r values of -0.99, -0.99, and -0.20, respectively. In other words, 7", appears to reflect the hydrodynamic volume, in the glass, of the mobile 'cluster entity'46 represented by a solute molecule and its complement of unfrozen water molecules, rather than a property of the isolated solute.26 One might ask whether it would be preferable to correlate r, with partial molar volume ( Y o ) rather than M , of dry solute, since V", like intrinsic viscosity, gives an indication of the effective solute size in solution.Just as free volume is related to inverse M , of monodisperse solutes (or inverse an for polydisperse solutes) in the limit of zero dilution,2i free volume should be related to inverse V" for comparison of conformational homologues of different M , in the limit of infinite dilution. However, it is for comparison of different conformers of the same M , that we look for an advantage in the use of V" to replace M,. Unfortunately, while M , values are exact, an approach based on V" is not straightforward Despite a relative wealth of I/" data,4i differences in the values for isomeric sugars are sufficiently small to discourage their use to interpret the influence of conformation on hydration4s and may lie within the range of values reported by different research groups for a single sugar, due in part to differences in anomeric ratios.49 If this complication is removed by comparison of methyl pyranosides, larger values of Vo are observed for conformers with equatorial rather than axial OCH, sub~tituents.~' Similarly, for comparison of conformers at C4, contributions to apparent molar volumes are said to be greater for equatorial than for axial hydroxyls.50 However, the general observation, for PHCs compared to their apolar structural analogues, is that hydroxy groups are effectively invisible to limiting density measurement~.~'. 49 The greater contribution of certain equatorial hydroxyl groups to V" is attributed to greater spatial and orientational compatibility with the pre-existing liquid water i.e.greater effective ' specific hydration '. Data for a few of the same sugars show that intrinsic viscosity increases with both contributions (M, and ' specific hydration ') to increasing V0.51 We might expect, from this slim evidence, that both contributions to increased V" would also lead to increased Tg. However, there still remain the questions of temperature and concentration dependence of apparent molar volumes of PHCs. Systematic extrapolation of V" data to sub-zero temperatures of interest for correlation with Tg is hindered by the paucity of data for mean limiting apparent molar expansibilities. Based on two relevant cases for which data are a~ailable,~' suchH. Levine and L.Slade 2627 Table 3. Comparison of Tg values and literature values for other ' collapse ' transition temperatures substance ly0ca ?/"Cb q / " c ethylene glycol propylene glycol glycerol ribose glucose fructose sucrose maltose lactose raffinose inosi to1 sorbitol - 7 0 , ~ " -81f - 62" -58, -6.5'~~ - 43 -41, -38 - 48 -32, -30.5 -27, -25.4 - 40 - 48 -32, - - 32 - 32 - 26 - 27 - 45 - 85 - 67.5 - 65 - 47 - 43 - 42 .34d -32 - 29.5 - 28 -26.5 - 35.5 -43.5 a Recrystallization temperatures, from Franks,12 p. 297. Collapse temperatures during freeze-drying, from Franks, l2 p.313. From Luyet,' p. 564. Antemelting temperatures, from Virtis CO.,~O p. 4. 'Completion of opacity' temperatures, from Forsyth and MacFarlane. l3 f Recrystallization temperature, from Thom and Matthes.14 extrapolations would magnify differences in behaviour predicted from measurements made near room temperature. The concentration dependence of apparent molar volumes is more questionable. One of the most important, but often overlooked, aspects of the glass transition is its cooperative Upon slow cooling, the glass at c, with solute-specific composition WL, represents the greatest dilution that retains this maximally cooperative behaviour. Cooperativity is maximum at the glass transition (where arrestation of large-scale molecular mobility occurs without change in structure), but decreases with increasing temperature or dilution above Tp (where retardation of mobility occurs and shows a WLF-type temperature dependence).Of the two extremes, behaviour in the limit of zero dilution is less remote from that of the cooperative system than is behaviour in the limit of infinite dilution. Apparent molar volumes of PHCs in aqueous solution have been shown to be characteristically (in contrast to apolar solutes) independent of c~ncentration,~' yet reported differences between apparent molar volumes for 3 and 10 wt % solutions of a single sugar approach the greatest differences seen between equatorial us. axial conformers at a single con~entration.~~ There exists the possibility that a decrease in apparent molar volume upon extrapolation towards T, and an increase upon extrapolation towards CL might counterbalance.Despite these issues the subject is of sufficient interest to merit further exploration elsewhere. 26 Structure-Property Relationships for Small Carbohydrates in Aqueous Solution T, and W,, as physico-chemically invariant but structure-dependent properties of glass- forming solutions at sub-zero temperatures, can be used to interpret thermomechanical behaviour of small-carbohydrate-water systems in non-equilibrium glassy and ' rubbery ' states. For example, as illustrated in table 3, previous established the fundamental identity of T, with transition onset temperatures observed for structural collapse during freeze-drying (T,) and recrystallization during frozen storage (q), for both model solutions and real systems (i.e.foods, pharmaceuticals, biologicals). These collapse phenomena are translational diffusion-controlled processes, for which a2620 Small-carbohydrate Water Glasses nnd ' Rubbers' physico-chemical mechanism was described,'. involving viscous flow in the ' rubbery ' state, at T > and viscosity, q < qg = 1011-1014 Pa s."' 27,46,53 An extensive list of collapse processes, all of which are governed by 7", of frozen systems (or of low- moisture systems), and involve potentially detrimental plasticization by water. has been identified and illustrated.". 'Q z3* 24 The non-Arrhenius kinetics of collapse and/or recrystallization in a high-viscosity fluid, which are governed by the free volume and mobility of the water-plasticized matrix, were shown5.6 . ':' to depend on the magnitude of AT above 7",, as defined by an exponential relationship derived from Williams-- Landel-Ferry (W'LF) free volume theory for glass-forming liquidsz7 Our previous study5 demonstrated the classical behaviour of commercial SHPs as an homologous family of amorphous glucose oligomers and polymers, and revealed a chain ' entanglement coupling 'z79 34, 54 capability for SHPs of DE < 6 and T, 3 - 8 C . For such polymeric SHPs, a lower limit of A?, z 3000 (calculated from DE : MI, = 1801 6DE-l) for entanglement leading to viscoelastic network formation"* 56 was identified. This an value was within the characteristic range of 1250--19 000 for the minimum entanglement M , of many typical synthetic linear high polymers.34 This entanglement capability was shown to correlate well with various functional attributes of low-DE SHPs, including a predictable ability' to form thermo-reversible, partially crystalline gels from aqueous ~ o l u t i o n .~ ' - ~ ~ Gelation occurred by a mechanism involving crystallization plus entanglement in concentrated solutions undercooled to 12" < 64 In contrast, the PHCs reported here, including malto-oligosaccharides up to DP 7, were found to fall below the Tg limit defined by SHPs for entanglement and the onset of viscoelastic rheological properties, and to be incapable of gelling from solution. It was clear, from the linearity of the 7", zw. Mi1 plot in fig. 1, that these monodisperse sugars, polyols and glycosides did not show evidence of entanglement coupling.For these small carbohydrates, none of which is larger than a heptamer of M, 11 53, the entanglement plateau region (in which remains constant with increasing M,27*35) has not been reached, a finding in accord with the M , range for entanglement limits cited above. Predicted Functional Behaviour of Small Carbohydrates in Solution We demonstrated previously5 how insights into structure-function relationships may be gleaned by treating a plot of T, 21s. Mn as a predictive map of regions of functional behaviour for SHP samples. For polymeric SHPs, the entanglement plateau region defines the useful range of gelation, encapsulation, cryostabilization, thermomechanical stabilization and facilitation of drying processes.For high-DE, low- T'L SHPs, low Mn corresponds to the region of sweetness, hygroscopicity, humectancy, browning reactions and cryoprotection. We described' how such a plot can be used predictively to choose individual SHPs or mixtures of SHPs and other small carbohydrates (targetted to a particular T, value) to achieve desired complex functional behaviour for specific product applications. Especially for applications involving such mixtures, similar use can be made of the corresponding results for the PHCs represented in fig. 1. For low-molecular- weight sugars and polyols, as for high-DE SHPs, low Tk values correlate with functional properties of sweetness, hygroscopicity, humectancy, browning and cryoprotection. One major area of practical significance of the current results on small carbohydrates can be explained in the context of 'cryostabilization', a concept we introduced" to describe a new industrial technology for the storage stabilization of frozen, freezer- stored and freeze-dried foods.This technology emerged from a basic research programme in food polymer sciencez3 and developed from a fundamental understanding of the critical physico-chemical and thermomechanical structure--property principles which underlie the behaviour of water in non-equilibrium food systems at sub-zero temperature^.^, l2 Cryostabilization is a means of protecting products, stored for long periods at typical freezer temperature ( q = - 18 "C), from the deleterious changes inH . Levine and L. Slade 2629 'O0I Liquid solution region 0 T f Z V Fo \ Tf 1 -135 I/ x c; y region I I 0 100 wt% solute Fig.3. Idealized state diagram of temperature us. wt 94 solute for an aqueous solution of a hypothetical small carbohydrate (representing a model frozen-food system), illustrating the critical relationship between Tg and freezer temperature (q), and the resulting impact on the physical state of the freeze-concentrated amorphous matrix. texture (e.g. ' grain growth ' of ice, solute crystallization), structure (e.g. shrinkage, collapse) and chemical composition (e.g. enzymatic activity, oxidative reactions, flavour or colour degradation) usually encountered. Such collapse-related changes are exacerbated in typical fabricated foods whose formulae are dominated by small carbohydrates with low T, and high Wi values.The key to protection and resulting improvement in storage stability lies in controlling the physico-chemical and thermomechanical properties, by controlling the structural state, of the freeze- concentrated amorphous matrix surrounding the ice crystals in a frozen system. This insight is illustrated by the idealized state diagram (modelled after one for fructose- water23) shown in fig. 3. If the matrix is maintained as a kinetically metastable mechanical solid (i.e. if TI < Q, then diffusion-controlled processes that typically result in reduced quality and stability can be virtually prevented or greatly inhibited. If, on the other hand, a natural material is improperly stored at too high a q, or a fabricated product is improperly formulated, so that the matrix is allowed to exist at T2 (> T,) as a 'rubbery' fluid in which translational diffusion is free to occur, then storage stability is reduced.The optimum Tr value for a natural material or optimum formula for a fabricated product is dictated by the characteristic Ti of the specific amorphous solute(s)-UFW matrix, which is governed in turn by Hw of the particular combination of water-soluble solids in a complex food system. Moreover, the dynamic behaviour of 'rubbery' frozen- food products during storage is kinetically controlled, and rates of diffusion-controlled deterioration processes are quantitatively determined by AT between and Tg. These rates increase exponentially with increasing AT above 7",, as defined by WLF, rather than Arrhenius, kinetics.For the cryostabilization of products such as ice cream (with smooth, creamy texture) against ice crystal growth during storage, inclusion of low-DE SHPs elevates the composite T, of a mix of soluble solids, which is usually dominated by low-molecular- weight sugars and other PHCS.~ In practice, a retarded rate of migratory ice recrystallization ('grain growth'7) at results and an increase in observed T,. In such2630 Small-carbohydrate- Water Glasses and ' Rubbers ' 0 CRY0 (STABILIZER + PROTECTANT 1 S i~ PPG STABILIZERS - 10 - 20 w hydroxypropylinulln * #hydroxypropylinulin stachyose -30 - 40 -50 - 60 - 70 -801 - 90' I I I I I I I I I I ene 01 - - 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 W i / g UFW g-' Fig. 4. Variation of the glass transition temperature, Tg, for maximally frozen 20 wt % solutions against Wi, the composition of the glass at q, in g UFW g-' solute, for a series of water- compatible compounds, including many small carbohydrates listed in table 1, illustrating the cryostabilization technology 'spectrum ' from monomeric cryoprotectants to polymeric cryostabilizers.products, ice recrystallization is known to involve a diffusion-controlled maturation process with a mechanism analogous to ' Ostwald ripening ', whereby larger crystals grow with time at the expense of smaller ones which eventually di~appear.".~'-~~ The rate at q, and thus also AT (T- Q, is reduced by formulating with low-DE SHPs of high 7';. We showed23 that ice recrystallization exemplifies a collapse process whose kinetics can be described quantitatively by the WLF equation." Because typical ice- cream products have 7'; values in the range - 28 to -43 OC, they exist as 'rubbery' fluids (with embedded ice and fat crystals) under typical freezer storage at - 18 OC, and WLF (rather than Arrhenius) kinetics describe the rate of ice crystal growth.In practice, the technological utility of Th and Wk results for sugars, polyols (table 1) and SHPs,' in combination with corresponding relative sweetness data, was demonstrated by the successful formulation of ice-cream products68' 69 with an optimum combination of stability and softness in a - 18 "C home freezer. The T, results reported here for PHCs and previously' for SHPs led to an interesting suggestion, which, in the present context, emphasizes the distinction in functional behaviour between non-entangling small carbohydrates and polymeric saccharides capable of entanglement.A freeze-concentrated glass at Tk of an SHP cryostabilizer (of DE < 6) would contain entangled polymer chains, while in a glass at Tg of a low- molecular-weight carbohydrate cryoprotectant, solute molecules could not be entangled. By analogy, various high-molecular-weight polysaccharide gums are believed to be capable of improving freezer-storage stability of ice-containing fabricated foods, in some poorly understood way. In the absence of direct evidence of any effect on ice crystal size, and despite recent work which showed definitively that such 'stabilizers' have no significant effect on either ice nucleation or propagation rates,'"," the effect has been attributed to increased viscosity.66* 7 2 Such gums may owe their limited success not onlyH.Leuine and L. Slade 263 1 to their ability to increase microscopic viscosity (a property shared by all glass-formers, regardless of M,, in the sense that ylg at Tg is independent of MW2'), but to their capability to undergo entanglement and increase macroscopic viscosity in the freeze-concentrated, amorphous matrix of a frozen food. Entanglement might provide an enhanced mouthfeel which masks the limited ability of gums to inhibit diffusion-controlled processes. Tg and WL results for a broad spectrum of water-soluble solids (fig. 4) led to the identification of a class of common food ingredients (including low-DE SHPs) called 'polymeric cryostabilizers '.' Our investigations of these solutes led to a theoretically based understanding of their stabilizing functionality (via their influence on the structural state of a complex amorphous matrix).This functionality derives from their high molecular weight, and the resulting elevating effect of such materials on a complex food system's characteristic T,. Increased TL leads to decreased AT, which in turn results in decreased rates of change during storage, and so increased stability of hard-frozen products. In contrast, the present results on small, carbohydrates (for which low molecular weight translates to low 7", and high W;) illustrate the predicted utility of sugars, glycosides and polyols as 'monomeric cryoprotectants ' (see fig.4) in freezer- stored foods with desirable soft-frozen texture (but undesirably poor stability). For many of the 10w-T~ reducing sugars in table 1, sweetness, hygroscopicity, humectancy and browning reactions are salient functional properties. 5, 7 3 A less familiar one involves the potential for cryoprotection of biological materials, for which the utility of various other low-molecular-weight, glass-forming sugars and polyols is well known.12-14,289 31 By analogy with earlier results for SHPs, the r", results in fig. 1 led us to predict, and d.s.c. experiments confirmed, that such small PHCs, in sufficiently concentrated solution, can be quench-cooled to a completely vitrified state, so that all the water is immobilized in the solute-UFW glass.' Such vitrification has also been suggested as a natural intracellular cryoprotective mechanism in winter-hardened poplar tree^,^' and demonstrated as a means of cryoprotecting whole body organs and ernbryo~.'~ The essence of this cryoprotective activity, indefinite avoidance of ice formation and solute crystallization in concentrated, undercooled solutions of small carbohydrates which have high Wk values, also has an apparent relationship to food applications involving soft, spoonable, or pourable-from-the-freezer products.One example is Rich Corporation's patented ' Freeze-Flo ' beverage concentrate formulated with high fructose corn syrup.76 Part of the underlying physico-chemical basis of such products involves colIigative freezing-point depression. This is illustrated by the idealized state diagram in fig.3, which reveals a hatched liquid solution zone, bordered by the Tm and Tp curves and extending below q2, for wt O h solute concentrations between x and y , which bracket CL. However, contrary to first appearances, functionality of specific PHC solutes in practical products depends more on the kinetically controlled properties of T, and WL than on simple colligative depression of equilibrium freezing point. This fact was demonstrated by model-system experiments with 60 wt % solutions of fructose and mannose (3.3 mol dmP3), methyl fructoside (3.1 mol dm-3), ethyl fructoside, ethyl mannoside and ethyl glucoside (2.9 mol dm-3), in which these samples have remained pourable fluids (completely ice- and solute-crystal free) under kinetic control during - 18 "C freezer storage (i.e.at T < the theoretical freezing point due to colligative depression) for over 4 years to date.23 We thank our former colleagues at General Foods: Drs Timothy Schenz and Allan Bradbury for their contributions to the work reported here; Dr William Eisenhardt for his encouragement of our research programme and Dr William Hall for his support of our research partnership. We especially thank our consultant and mentor, Prof. Felix Franks (University of Cambridge), for invaluable suggestions, discussions, encouragement and support over the years.2632 Small-carbohydrate- Water Glasses and ' Rubbers ' References 1 B. Luyet, J. Phys. Chem., 1939, 43, 881. 2 T. W. Schenz, M. A. Rosolen, H. I. Levine and L.Slade, Proc. 13th NATAS Con$ (NATAS, Philadelphia, 1984). pp. 57-62. 3 F. Franks, Properties of Water in Fo0d.s (Martinus Nijhoff, Dordrecht, 1985), pp. 497-509. 4 D. S. Reid, Cryo-Lett., 1985, 6, 181. 5 H. Levine and L. Slade, Carbohydr. Polym., 1986, 6, 213. 6 H. Levine and L. Slade, Food Structure: Its Creation and Evaluation (Butterworths, London, 1987), 7 B. Luyet. Ann. N. Y. Acad. Sci., 1960, 85, 549. 8 D. Ramussen and B. Luyet, Biodynamica, 1969, 10, 319. 9 A. P. MacKenzie and D. H. Rasmussen, Water Structure at the Water-Polymer Interface (Plenum Press, chap. 9. New York, 1972), pp. 146-71. 10 A. P. MacKenzie, Philos. Trans. R. SOC. London, Ser. B., 1977, 278, 167. 11 F. Franks, M. H. Asquith, C. C. Hammond, H. B. Skaer and P. Echlin, J. Microsc., 1977, 110, 223.12 F. Franks, Water: A Comprehensiae Treatise (Plenum Press, New York, 1982), vol. 7, pp. 215-338. 13 M. Forsyth and D. R. MacFarlane, Cryo-Lett., 1986, 7, 367. 14 F. Thom and G. Matthes, Cryo-Lett., 1986, 7, 31 1. 15 A. P. MacKenzie, Freeze Drying and Advanced Food Technology (Academic Press, New York, 1975). 16 A. P. MacKenzie, Microprobe Analysis of Biological Systems (Academic Press, New York, 1981), 17 S. Tsourouflis, J. M. Flink and M. Karel, J. Sci. Food Agric., 1976, 27, 509. 18 E. C. To and J. M. Flink, J. Food Technol., 1978, 13, 551. 19 J. M. Flink, Physical Properties of Foods (AVI, Westport, 1983), pp. 473-521. 20 M. Karel and J. M. Flink, Advances in Drying (Hemisphere, Washington, 1983), vol. 2, pp. 103-53. 21 M. Karel, Properties of Water in Foods (Martinus Nijhoff, Dordrecht, 1985), pp.153-169. 22 M. Karel, Concentration and Drying of Foods (Elsevier Applied Science, London, 1986), pp. 37-51. 23 H. Levine and L. Slade, Water Science Reviews (Cambridge University Press, Cambridge, 1987), vol. 3, 24 H. Levine and L. Slade, Water and Food Quality (Elsevier Applied Science, London, 1988), in press. 25 L. Slade and H. Levine, Food Structure - Its Creation and Evaluation (Butterworths, London, 1987), 26 L. Slade and H. Levine, Bure Appl. Chem., 1988, to be published. 27 J. D. Ferry, Viscoelastic Properties of Polymers (John Wiley, New York, 3rd edn, 1980). 28 F. Franks, Biophysics and Biochemistry at Low Temperatures (Cambridge University Press, Cambridge, 29 E. Maltini, 1.I.F.-1.I.R.-Karlsruhe, 1977, 1, 1.30 Virtis SRC Sublimators Manual (Virtis Company, Gardiner, New York, 1983). 31 R. Vassoille, A. El Hachadi and G. Vigier, Cr"vo-Lett., 1986, 7, 305. 32 D. R. MacFarlane, Cryo-Lett., 1985, 6, 313. 33 P. Boutron and A. Kaufmann, Cryobiol., 1979, 16, 557. 34 W. W. Graessley, Physical Properties of Polymers (ACS, Washington, 1984), pp. 97-153. 35 F. W. Billmeyer, Textbook of Polymer Science (Wiley-Interscience, New York, 3rd edn, 1984). 36 T. E. Beesley, Am. Lab., 1985, May, 78. 37 F. Franks, Cryo-Lett., 1986, 7, 207. 38 X. Jin, T. S. Ellis and F. E. Karasz, J. Polym. Sci., Polym. Phys. Ed., 1984, 22, 1701. 39 J. Biros, R. L. Madan and J. Pouchly, Collect. Czech. Chem. Commun., 1979, 44, 3566. 40 J. Pouchley, J. Biros and S. Benes, Makromol. Chem., 1979, 180, 745.41 C. A. J. Hoeve, Water in Polymers, ACS Symp. Ser. 127 (ACS Washington, 1980). pp. 135-46. 42 C. A. J. Hoeve and M. B. J. A. Hoeve, Org. Coat. Plast. Chem., 1978, 39, 441. 43 H. G. Burghoff and W. Pusch, Polym. Eng. Sci., 1980, 20, 305. 44 H. W. Starkweather, Water in Polymers, ACS Symp. Ser. 127 (ACS Washington, 1980), pp. 433-40. 45 F. Franks, personal communication. 46 T. Soesanto and M. C. Williams, J. Phys. Chem., 1981, 85, 3338. 47 A. Cesaro, Thermodynamic Data for Biochemistry and Biotechnology (Springer-Verlag, Berlin, 1986), 48 F. Franks, Pure Appl. Chem., 1987, in press. 49 F. Franks, J. R. Ravenhill and D. S. Ried, J. Solution Chem., 1972, 1, 3. 50 G. G. Birch and S. Shamil, J. Chem. Soc.. Faraday Trans. 1, 1988, 84, 2635. 51 M. Mathlouthi and A. M. Seuvre, J. Chem. SOC., Faraday Trans. I , 1988, 84, 2641. 52 B. Wunderlich, Thermal Characterization of Polymeric Materials (Academic Press, Orlando, 198 I), pp. 277-307. pp. 397421. pp. 79-185. chap. 8. 1985). p. 177. pp. 91-234.H. Levine and L. Slade 2633 53 G. E. Downton, J. L. Flores-Luna and C. J. King, Zndust. Eng. Chem. Fund., 1982, 21, 447. 54 J. R. Mitchell, J . Text. Stud., 1980, 11, 315. 55 P. J. Flory, Principles of Polymer Chemistry (Cornell University Press, Ithaca, 1953). 56 P. J. Flory, Faraduy Discuss. Chem. Soc., 1974, 57, 7. 57 E. E. Braudo, E. M. Belavtseva, E. F. Titova, I. G. Plashchina, V. L. Krylov, V. B. Tolstoguzov, 58 E. E. Braudo, I. G. Plashchina and V. B. Tolstoguzov, Carbohydr. Polym., 1984, 4, 23. 59 P. V. Bulpin, A. N. Cutler and I. C. M. Dea, Gums and Srubilizersfor the Food Industry 2 (Pergamon Press, Oxford, 1984), pp. 475-484. 60 F. Reuther. G. Damaschun, C. Gernat. F. Schierbaum, B. Kettlitz, S. Radosta and A. Nothnagel, Colloid Polym. Sci., 1984, 262, 643. 6 1 B. Wunderlich, Macromoleculur Physics. Voi. 2 - Crystul Nucleation, Growth, Annealing (Academic Press, New York, 1976). 62 B. W-underlich. Macromoleculur P/iysics, Vol. 3 -. Cr)wul Melting (Academic Press, New York, 1980). 63 R. F. Boyer, E. Baer and Hiltner, Mucromolecule,r, 1985, 18, 427. 64 R. C . Domszy, R. Alamo, C. 0. Edwards and L. Mandelkern, Mucromolecuies, 1986, 19, 310. 65 A. E. Bevilacqua and N. E. Zaritzky, J. Food Sci., 1982, 47, 1410. 66 E. K. Harper and C. F. Shoemaker, J. Food Sci., 1983, 48, 1801. 67 F. Franks, J. Darlington, T. Schenz, S. F. Mathias, L. Slade and H. Levine, Nature (London), 1987, 68 B. A. Cole, H. 1. Levine. M. T. McGuire, K. J. Nelson and k. Slade, U.S. Paten2 4374154, 1983. 69 B. A. Cole, H. I . Levine, M. T. McGuire, K. J. Nelson and L. Slade, U.S. Patent 4452824, 1984. 70 A. H. Muhr, J. M. V. Blanshard and S. J. Sheard, J . Food Terhnol., 1986, 21, 587. 71 A. H. Muhr and J. M. V. Blanshard. J . Food Technol., 1986, 21, 683. 72 P. G. Keeney and M. Kroger, Fundamentals of Dairy Chemistry (AVI, Westport, 2nd edn. 1974), p. 890. 73 S. Z. Dziedzic and M. W. Kearsley. Cfiicose Syrups: Science and Technology (Elsevier Applied Science, London. 1984), pp. 137-168. 74 A. G. Hirsh, R. J. Williams and H. T. Meryman, Planr Physiol., 1985, 79, 41. 75 W. F. Rall and G. M. Fahy, Nutiire (London), 1985. 313, 573. 76 M. L. Kahn and K. E. Eapen. U.S. Putent 4332824, 1982. F. R. Schierbaum and M. Richter, Starke, 1979, 31, 188. 325, 146. Paper 712079 : Receitled 23rd November, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402619

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J . Chem. SOC., Furuduy Trans. I, 1988, 84(8), 2635-2640 Structure, Sweetness and Solution Properties of Small Carbohydrate Molecules Gordon G. Birch” and Syed Shamil Department of Food Science and Technology, Food Studies Building, Unitiersity of Reading, Whiteknights, PO Box 226, Reading RG6 2AP The sweetness of small carbohydrate molecules in relation to their structure has been under investigation for several years. Tentative glucophores can now be assigned to some structures and the sweetness effect is thought to originate in a hydrogen bond between stimulus molecule and sweet receptor. A recent study of solution properties of small carbohydrates in relation to sweetness aims to examine the hydrogen bonding between water molecules and sweet solutes as a preliminary stage of taste chemoreception.Hydrated sugar molecules are presumed to accede to receptors before binding with the receptors is able to occur. Apparent molar and apparent specific volumes give indications of the effective size of sapid solute and its compatibility with water structure. Axial and equatorial hydroxy groups differ in their effects according to their location in the hexopyranose structure. Analogous structures exhibit analogous apparent molar volumes; these effects may be related to the differences of sweetness intensity between some sugar structures and the bitterness of othe;s. Although there is yet no simple relationship between sweetness intensity and apparent molar volume (4 V ) , it is possible that (OV) does govern the taste dominance of one molecular species in binary mixtures.Moreover, apparent specific volumes of a range of sapid solutes seem to correlate with taste quality. _ _ -___ ___ .~ Evidence exists’, that a-glycol systems in sugar molecules are responsible for their sweet tastes. The a-glycol group is thought to hydrogen-bond3 to the sweet receptor and axiakquatorial or diequatorial arrangements of the hydroxy groups are ideally spaced for this type of receptor binding. Sweet receptors appear to have steric specificity. Therefore, in addition to possessing suitable a-glycol systems, sugar molecules must be conformationally commensurate with the sweet receptor in order to elicit the sweet response. This in turn means that the sugar molecule is orientated in a particular way on the sweet receptor and therefore only one particular a-glycol system hydrogen-bonds with it.Earlier work4 to locate this particular a-glycol group or ‘glucophore’ was based on the stepwise elimination of oxygen atoms around the sugar ring and sensory evaluation of the resulting deoxy sugars. By such methods it can be shown that the 3,4- a-glycol system of glucopyranose types of structure appears to constitute the ‘ glucophore ’. Although glucophores may be identified in all sweet-tasting molecules, not all molecules with apparent ‘glucophores’ are sweet. It is therefore clear that for sweetness to be elicited, a molecule must possess a ‘glucophore’ as well as some additional characteristics. For example, the molecule must accede to the appropriate region of the sweet receptor before binding and activation of the ion-channel mechanism occurs.The accession efficiency of the molecule is mediated by water, as probably no molecule can be tasted without first being dissolved in water. Therefore a study of solution properties’ of small carbohydrate molecules, particularly chlorinated sugars, has been proposed as an explanation of the taste chemoreception mechanism. 26352636 Sweetness of Small Carbohydrates Experimental Substances investigated in this work were reagent-grade chemicals obtained from BDH Chemicals, Poole, Dorset, The Sigma Chemical Co., Poole, Dorset. Where possible, physical constants (e.g. m.p. and optical rotation) were checked against literature reports. Water used for solution studies as well as taste work was 'Hyper Solv' water for h.p.1.c. from BDH.Reducing sugars were allowed 30 h in the refrigerator ( 5 "C) after dissolution to reach mutarotational equilibrium. Apparent molar volumes were determined in weight per weight solutions with an Anton-Parr Precision Density Meter (DMA 60) and Density Measuring Cell (DMA 602) (Stanton Redcroft, London) equipped with an automatic sampler (SP2) and Anadex printer. Temperature control was achieved with a Hetofrig bath (Heto Birkerod, Denmark) coupled to density measuring cell. The density meter was calibrated with air and water and the method was as previously described.6 All measurements were carried out at 20 "C (kO.1 "C). Uncertainty in the density determinations was 3.0 x g cm -3. Results and Discussion All taste phenomena probably result from weak reversible adsorption of stimulus molecules on the receptor s ~ r f a c e .~ Since sweet taste is known to be affected by both the concentration and temperature', of solution, a study of solution properties is a logical method of approaching structure/activity relationships in sugar molecules. An important property is apparent molar volume ($V), which is an apparent measure of solute size and reflects displacement or disturbance of water structure. Table 1 gives some examples of increases of ( O V ) with increasing concentration and temperature. Measurements of ($I/) at different concentrations raises a complication in that surface tension (a measure of molecular cohesion or comprespion) is likely to vary. Table 1 therefore lists some apparent solution parachors (OVp; where y = surface tension of solution) which represent apparent molar volumes at 20°C if the surface tension were to remain at unity.Although the monosaccharides generally show an upward trend of 6Vwith increasing concentration in accordance with previous observations,6 the differences between sugars of similar molecular weight are very small. Moreover, despite opposite trends in surface tension values, the solution parachors show no significant differences between sugars of similar molecular weight. D-Fructose exhibits an increase in surface tension as the concentration increases. Accordingly it is the only monosaccharide in table 1 to show an increase in apparent solution parachor. Increases of 4 V with temperature within the range 20-40 "C are never more than 1-2 YO, i.e. of the same order of magnitude as the effects of concentration (table 1).It has been shown that apparent molar volumes of sugars are affected by the axial and equatorial arrangements of their hydroxy groups.8 Equatorial hydroxy substituents are more easily hydrated than axial. Hence P-D-glucopyranose, with all equatorial substituents, forms the most heavily hydrated structure. Heavily hydrated structures, however, are more compatible with surrounding water structure and hence exhibit low apparent molar volumes. These differences may contribute to differences in accession efficiencies of sapid molecules and thus explain differences in taste. D-Galactopyranose (axial OH substituent at C-4) is less sweet than D-glucopyranose (equatorial OH substituent at C-4).However, a-D-mannopyranose (axial OH at C-2) has the same apparent molar volume as D-glucopyranose and a similar sweetness. Also a- and p-D-glucopyranose have similar apparent molar volumes and sweetness values. lo Therefore the effect of configuration on apparent molar volume depends on the positionG. G. Birch and S. Shamil 2637 Table 1. Apparent molar volumes (#V/cm3 mol-l) of sugar solutions at 20, 30 and 40 "C with apparent solution parachors at 20 "C, given in brackets concentration (%, w/w) compound T/"C 10 20 30 D-xylose 20 30 40 D-fructose 20 30 40 30 40 D-glucose 20 30 40 D-galactose 20 95.1 1 (265) 96.96 97.50 110.1 (320) 112.1 112.9 110.7 (318) 111.1 1 1 1.8 (323) 113.0 113.8 95.58 (260) 96.24 96.70 110.8 (323) 112.2 113.0 111.0 (315) 11 1.8 112.6 112.2 (323) 114.0 113.8 96.09 (258) 96.66 97.02 11 1.4 (325) 112.3 113.1 111.7 (311) 112.7 114.4 112.8 (332) 113.9 114.8 Table 2.Differences in apparent molar volumes (QV) of axial and equatorial isomers of pyranoses ~ ~~ ~~~ Q V/cm3 mol-' a position of OH group axial equatorial V differencesb c- 1 110.8 c-2 110.8 (a-D-glucose) (D-mannose) 93.80 c-3 94.78 c-4 109.0 (D-I yxose) (D-ri bopyranose) (D-galactose) 91.84 (L-arabinose) 110.8 110.8 (p-D-glucose) (D-glucose) (D-xylose) (D-xylose) (D-glucose) (D-xylose) 93.80 93.80 110.8 93.80 0 0 0 0.98 1.80 1.96 At 3 % w/w concentration. * Differences in QV values of molecules with axial and equatorial configurations of the OH group. within the molecule and sugar molecules may be thought of as 'polarised' with respect to water structure just as they are orientated with respect to the taste receptors.I n other words, water molecules probably cluster preferentially around one end (hydrophilic) of the sugar molecule and a clue to this effect is the difference between V vales of axial and equatorial substituents around the pyranose ring (table 2). It is clear that equatorial-axial changes at C-3 or C-4 give rise to the greatest disturbance of water structure as evidenced by the difference in d V values and these are the same positions in the sugar molecule that are responsible for the sweet taste.2638 Sweetness of Small Carhoh-vdrates Fig. 1. Postulated clustering of water molecules around the sugar molecule. Fig. 2. /?-D-Fructopyranose. Alteration of the 3,4-diequatorial system of glucose not only disturbs water structure but also depresses sweetness intensity.This same region of the molecule may therefore also be heavily hydrated and fit well with water structure. Fig. I depicts the possible clustering of water molecules around a sugar molecule, which could in turn be related to the evidence that the anomeric centre of D-glucopyranose plays no part in the sweetness response.'O This constitutes the 'hydrophobic' end and is related to the finding P-D-mannopyranose is ' very hydrophobic ' . I 1 The fact that P-D-mannopyranose is bitter whereas P-D-glucopyranose and the inositols are not supports the general observation that hydrophobicity is related to the bitter taste quality.12 The ring oxygen atom, though polar, is well shieldedI3.and therefore contributes to the hydrophobicity and bitter taste. Further evidence to suggest this is provided by D-glucono- 1'4-lactone which is sweet-bitter in taste and has a higher apparent specific volume than the corresponding free acid.g, l5 Calculations of hydrophobic areas of the sugars" also support the finding that P-D-mannopyranose is very hydrophobic. Although the anomeric centres of D-glucopyranose, D-galactopyranose and D- mannopyranose may constitute the 'hydrophobic' ends of the molecules, the same cannot be said of D-fructopyranose. This is because D-fructopyranose exists in the alternative 5C2 conformation, however ; when drawn analogously to the aldopyranoses, the 'hydrophilic ends' may be viewed as analogous (fig.2). 2-Deoxy-~-fructopyranose is identical to 1 -deoxy-D-mannopyranose and is sweet. Thus is it possible that the 3,4-a-glycol system is within the hydrophilic region and responsible for the sweetness of ail pyranoses. D-Fructopyranose, is an analogue of D-arabinose while D-galactose is an analogue of L-arabinose and D-glucose is a structural analogue of D-xylose (fig. 3). The above structural analogues give analogous ratios of 3 V values and exhibit analogous intensities of taste17 (table 3). This result is consistent with those of table 2 in that clustering of water molecules is governed by analogous structural features. It is not clear, from the above results, how apparent molar volumes can be used to predict intensity or durationG. G. Birch and S.Shamil 2639 HO OH (1 1 I OH OH ( 5 ) (6 1 Fig. 3. Structural analogues of pyranoses : (1) D-glucose ; (2) D-XY~OSC ; (3) D-galactose ; (4) L-arabinose; ( 5 ) D-fructose; (6) D-arabinose. Table 3. Ratios of apparent molar volumes (6V) of hexopyranoses to corresponding pentopyranoses sweetness m.wt. hexose fiV hexose sugar intensity m.wt. dVV/cm3 mo1-l" m.wt. pentose fiV pentose D-xylose sweet 150.13 95.1 1 D-galactose less sweet 180.16 110.7 1.20 I .20 p-D-fructopyranose very sweet 180.16 110.6 1.20 1.20 D-glucose sweet 180.16 111.8 1.20 1.18 L-arabinose less sweet 150.13 92.32 D-arabinose sweet 150.13 92.18 At 10 % w/w concentration. of sweet response. Only in mixtures of two sugar^^'^'^ is it clear that the dominant member of a pair is the one with a lower apparent molar volume. From studies of apparent specific volumes,19 it appears that most sapid substances fit into a general pattern related to taste as shown in table 4.Substances with apparent specific volumes greater than ca. 0.93 cm3 g-' are probably so hydrophobic that they are beyond the range of human taste perception. They may, however, be volatile and thus elicit an olfactory response. The explanation of the results of table 4 may be related to disturbance of water structure. Evidently the different values of apparent specific volume are related to differences in taste receptor signals. For example the receptor cells might be sensitive to fine disturbances of water structure or, alternatively, they might be located at different layers of the taste epithelium.The latter hypothesis is already supported by psychophysical, biochemical and anatomical evidence,20-22 and thus it is possible that apparent specific volume reflects accession efficiency of solute stimulus to receptor. Protons and ionisable salts are most compatible with water structure (i.e. small2640 Sweetness of Small Carbohydrates Table 4. Discrimination of basic tastes by apparent specific volume (Shamil et al., 1987) apparent specific volume (6 V/m.wt.)/cm3 g-' taste region 04.33 0.334.52 0.524.7 1 0.714.93 salt sour sweet bitter apparent specific volumes) and they may be conveyed quicker by water to deeper layers of the taste epithelium than are sugar molecules. For a molecule to elicit a particular basic taste it must first accede to an appropriate region and, secondly, must possess an appropriate structural ' sapophore '.l9 Some molecules possess more than one basic taste. For example, many sugar derivatives are bitter-sweet and their apparent specific volumes fall neatly in the border region between the basic tastes, at ca. 0.7 cm3 8-I. Apparent specific volumes may therefore contribute to the prediction of taste quality. Conclusions It seems likely that all taste effects are mediated by water and the solution properties of small carbohydrate molecules therefore give clues to the possible mechanisms of taste chemoreception. The dynamic interplay of hydrogen bonds in aqueous solutions of small carbohydrate molecules governs their solution properties and probably also governs their taste-eliciting capacity.Apparent molar volumes of sugar molecules reflect fine differences in their structures which are in turn related to their taste differences. These and related solution measurements are already providing useful data for explaining the stereochemical basis of taste and signify the direction of this field of research in the immediate future. References 1 G. G. Birch, Crit. Rev. Food Sci. Nutr., 1976, 8, 57. 2 G. G. Birch, Sweetness, ed. J. Dobbing (Springer-Verlag, Berlin, 1987). 3 R. S. Shallenberger and T. E. Acree, Nature (London), 1967, 216, 480. 4 G. G. Birch and C. K. Lee, J. Food Sci., 1974, 39, 947. 5 M. Mathlouthi, A. M. Seuvre and G. G. Birch, Carbohydr. Res., 1986, 152, 47. 6 G. G. Birch and S. Catsoulis, Chem. Senses, 1985, 10, 325. 7 L. M. Beidler, J. Gen. Physiol., 1954, 38, 133. 8 F. Franks, J. R. Ravenhill and D. S. Reid, J. Solution Chern., 1972, 1, 3. 9 S. Shamil, Ph.D. Thesis (Reading University, 1987). 10 G. G. Birch, S. Shamil and Z. Shepherd, Experimentia, 1986, 42, 1232. 11 M. Janado and Y. Yano, J . Solution Chem., 1985, 14, 891. 12 H. D. Belitz and H. Wieser, Food Rev. Int.. 1985, 1, 271. 13 J. W. Brady, J . Am. Chem. Soc., 1986, 108, 8153. 14 J. W. Brady, Carbohydr. Res., 1987, in press. 15 S. Shamil, G. G. Birch, M. Dinovi and R. Rafka, unpublished work. 16 K. Miyajima, K. Machida and M. Nakagaki, Bull. Chem. Soc. Jpn, 1985, 58, 2595. 17 G. G. Birch and S. Shamil, Food Chern., 1986, 21, 245. 18 S. L. Munton and G. G. Birch J . Theor. Biol., 1985, 112, 539 . 19 S. Shamil, G. G. Birch, M. Mathlouthi and M. N. Clifford, Chem. Senses, 1987, 12, 397. 20 Y. Hiji and J. Ito, Comp. Biochem. Physiol., Part A, 1977, 58, 109. 21 R. A. Frank and D. L. Korchmar, Physiol. Behav., 1985, 35, 239. 22 B. G. Green and S . P. Frankmann, Chem. Senses, 1987, 12, 609. Paper 712078 ; Received 23rd Notlernber, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402635

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1988, 84(8), 2641-2650 Solution Properties and the Sweet Taste of Small Carbohydrates Mohamed Mathlouthi* Laboratoire de Chimie Physique Industrielle, De'partement de Chimie, Faculte' des Sciences, UniuersitP de Reims-Champagne- Ardenne, B. P. 347, 51062 Reims Ce'dex. Frunce Anne-Marie Seuvre Institut Universitaire de Technologie, Universitk de Dijon, B.P. 510, 21014 Dijon Cc'dex, France In addition to a knowledge of their structure, especially the geometry of the (AH-B, 7) glucophore, it is necessary to obtain information on solution properties of small carbohydrates if one wishes to elucidate the mechanism of their sweet taste. Macroscopic (viscosity, calorimetry) and microscopic (vibrational spectroscopy) methods have been used to investigate the solution properties of mono- and di-saccharides and a very sweet chloroderivative of sucrose (TGS).The intrinsic viscosity, [q], and Huggins constant, k', permitted a description of the overall solute-solvent inter- actions. It was not easy to correlate calorimetric results with sweetness. Fourier-transform infrared studies of solid sugars permitted proposition of glucophores in the light of the observed OH vibrations. An analysis of Raman spectra in the OH stretching region provided information on the effect of traces of the sugars on the structure of water and helped in defining the role of water in the sweet taste mechanism. The relationship between molecular properties and the sweetness of carbohydrates has been studied for many years. The model which is usually accepted is based on the hypothesis of Shallenberger and Acree' that the functional groups for sweet taste are two vicinal OH groups, one acting as a 'proton donor', AH, and the other as a 'proton acceptor', B, in hydrogen bonding with the receptor site.A third component of the molecular theory of sweetness is a hydrophobic group called a y - ~ e n t r e , ~ ' ~ generally situated in the opposite side of the couple AH-B. The dimensions of this tripartite glucophore (AH-B, y ) have been proposed4 and the dominant role of hydrogen bonding in sweetness ~nderlined.~ The specificity of the stereochemical fitting of the sweet molecule with the receptor site led some authors to propose the existence of barriers' on the taste bud or an enzyme-substrate scheme7 for the arrangement of the sweetener in the receptor.Although the role of water in the sweet taste mechanism is evoked in the ' hydrophobic ' function of the y-centre, no systematic study of the relationships between solution properties of sugars and their sweetness was undertaken. It is only recently that solute-solvent interactions8 and physical propertiesg of aqueous solutions of small carbohydrates were correlated with taste. However, electrophysiology results'' showed that the role of water in taste chemoreception is far from negligible. Indeed, the adsorption of the sweet molecule on the receptor site induces a potential in the taste cells which is probably due to ion transfer across the membrane. Such an ion transfer is facilitated when water mobility is increased, as proposed8 to explain the relatively high sweetness of D-fructose compared to that of D-glucose or sucrose.Information on both the geometry of the molecule and solution properties is needed. To fit with the receptor 264 12642 Solution Properties and Sweetness site, the sweetener should obey the conformational requirements: this fitting is at the basis of the trigger action in the sweetness mechanism. The persistence of the sensation is probably due to the mobility of water induced by solute-solvent interactions in aqueous solutions of small carbohydrates. In order to interpret the differences in sweetness of the monosaccharides (D-fructose, D-glucose and D-galactose), the disaccharides (sucrose, lactose and maltose) and a chlorinated derivative of sucrose (TGS : 4,1',6'-trichloro-4,1',6'-trideoxygalactosucrose) we determined their viscometric constants {[q], the intrinsic viscosity and k', the Huggins constant) and their heats of dilution and recorded their Fourier-transform infrared spectra and the Raman spectra of their dilute solutions in the OH-stretching region. Experimental The sugars studied were commercially available or synthesised (for the chlorodeoxy- derivative) according to methods previously" reported.Doubly distilled water was used for preparation of the solutions. Viscosity results were derived from the time necessary for a given volume to flow through a capillary at a constant temperature of 25 +O. 1 "C in a semiautomatic Schott AVS 400 viscometer. The intrinsic viscosity [q] was obtained from the triple extrapolation of the reduced specific viscosity bs,/c = (q -qo)/qo c], the inherent viscosity [In ( q / q J / c ] and the reduced differential viscosity [qdiff/c = (q - qo)/qc] towards c = 0, where q and qo are, respectively, the viscosities of the solution and solvent and is c the concentration in g cm-".The Huggins constant k' was derived from Huggins' relation :12 The heats of dilution were determined with an LKB- 10700 batch-microcalorimeter at 25k0.1 "C. Fourier-transform infrared (F.t.i.r.) and laser Raman spectra were recorded according to methods previously'3* l4 described. ?Ispic = full + k' [ulI2c + * * * - Results Experimental values of [q] and k' are given in table 1 for the sugars investigated. An example of the triple extrapolation procedure is given in fig. 1.This method of determining [q] involves a test of verification of the results. The experimental values are in agreement with the Meffroy-Biget relation :15 ( ~ s p / c + tldiff/c)/2 = 1' ( q / q o ) / ~ * Virial coefficients (hii) of the excess enthalpies of interactions in aqueous sugar solutions given in table 2 are derived from the heats of dilution (AHdi1) according to the relation where mi and m, are, respectively, initial and final molar concentrations. Calculated values of hii are based on an additivity model proposed by Savage and Wood" involving the contributions of each of the CH,, COH and CH groups. Our experimental results (table 1) show good agreement with those of Barone et a1.l' F.t.i.r. spectra of solid P-D-fructopyranose, sucrose and TGS in the region 2700- 3700cm-l are reported in fig.2. A sharp absorption between 3450 and 3600cm-', generally assigned to free OH vibration, is observed in these spectra. Such a well defined band is not seen in the i.r. spectra of the other sugars. Laser Raman spectra of aqueous solutions of sugars are the most informative on the role of water in the sweetness mechanism. The effect of low concentrations g g-') of D-glucose, D-fructose and sucrose on the Raman spectrum of water in the AH"" = hii mf(mr- mi) + hiii m,(m,2 - mi') + . . .M. Mathlouthi and A-M. Seuvre 20 2643 - I I 1 I I Table 1. Viscometric constants of small carbohydrates sugar [q]/ cm3 g-’ k’ D-glucose 20.15 1 S O D-galactose 19.80 1.04 D-fructose 19.00 1.14 sucrose 20.35 1.20 maltose 22.50 1.01 lactose 22.45 1.16 TGS 18.00 2.12 - Fig.1. Determination of intrinsic viscosity [q] of D-glucose by extrapolation of q,Jc (@); (1’ v / v ~ ) / c (A) and Vdiff/C (HI. w avenumber/cm-1 Fig. 2. F.t.i.r. spectra of solid P-D-fructopyranose (a), sucrose (b) and TGS (c) in the OH stretching region. 87 F A R 12644 Solution Properties and Sweetness 3000 3200 3400 3600 3000 3200 3400 3600 3000 3200 3400 3600 wavenumber/cm-' Fig. 3. Laser Raman spectra of dilute [(-+-) lop2, (----) lo-' and (----) 10-6gg-1] aqueous solutions of D-fructose (a), D-glucose (b) and sucrose (c) and water (-) in the OH stretching region. wavenum ber/cm-' Fig. 4. Experimental plot (-) and calculated gaussian components (. . . - ) of the Raman spectrum of water.wavenumber/crn-' Fig. 5. Experimental plot (-) and calculated gaussian components ( . . . - ) of the Raman spectrum of 1 O/O aqueous D-fructose solution.M. Mathlouthi and A-M. Seuvre 0.3 p 0.2- 0.1 2645 - ' 0.3 0.2 0.1 0.3 p 0.2 0.1 0.3 0.2 0.1 1 . . a . = 3000 3200 3400 3600 wavenumberlcm-' Fig. 6. Variation of the depolarization ratio ( p ) for water (-) and dilute [(-+-) lo-', (----) lo-" and (----) lo-' g g-'1 solutions of D-fructose (a), D-glucose (b) and sucrose (c). 3000 3200 3600 w avenum berlcm-' Fig. 7. Variation of the depolarization ratio ( p ) of water at (-.-) 4, (-+-) 35. (-) 100 (----) and 120 "C. 2700-3700 cm-' region is observed in fig. 3. The overall effect of traces of D-fructose seems to provoke a significant perturbation in the structure of water.A semiempirical method'' of deconvolution of the v(0H) Raman band of water was applied : the results are shown in fig. 4 and 5 for pure water and 1 % D-fructose solution, respectively. The depolarization ratios p@ = I , , /IL) were calculated and are represented (see fig. 6) as a function of wavenumber in the OH stretching region for each of the spectra reported in fig. 3. The evolution of p is found to be comparable to that found" for the perturbation of water structure produced by modifying the temperature (see fig. 7). 87-22646 Soh tion Proper ties and Sweetness Discussion Macroscopic Results Intrinsic viscosities [ r] of monosaccharides were obtained after mutarotational equilibrium was attained. [y~] is generally considered as a shape factor that accounts for the hydrodynamic radius of the solvated molecule. The values of [q] are comparable for monosaccharides [( 19.0-20.15) x lop3 cm3 g-'1 and disaccharides [(20.35-22.50) cm3 gl], which is probably due to the hydrophilic nature of the hydration of sugars and the quasi-spherical shape of their hydrated molecules.This is not the case for TGS because of the marked difference between its hydrophobic and hydrophilic sides and its higher flexibility around the glycosidic linkage as compared to sucrose. l9 The important difference between the effects of the hydrophilic and hydrophobic sides of the molecule is probably the origin of the relatively low value of [r] for D-fructose (see table 1). The Huggins constant k' is taken as an interaction factor accounting for the mobility of water around the solute.The observed differences in the values of k' (table 1) may be due to differences in the compatibility of the hydration of the solute with the water structure. k' is related to the specific viscosity by a linear relation:2" k' = k + k2vsr,. The variation of k' with ylSp was usedz1 as an additional element of information on the behaviour of sugars in aqueous solutions, showing that water molecules are more mobile around D-fructose than around other monosaccharides. Viscometric constants [r] and k' may be used as a first approach to the interpretation of the sweetness of small carbohydrates. A marked opposition between the sweet and bitter ends, or hydrophilic and hydrophobic of the molecule, together with an increased mobility of water around it, is manifested by a low intrinsic viscosity and a high Huggins constant, as seen for TGS (table 1).An analysis of calorimetric data (table 2) shows that the differences in hii values cannot be simply interpreted in terms of differences in sweetness. A relatively high value of hii for g glucose as compared to other monosaccharides may be due to the preponderance in aqueous solution of the p-anomer, which induces long-range order in the water molecules. However, the origins of hii values are n u m e r ~ u s ' ~ and not easy to quantify. As an example, note that the furanose form in D-fructose solution contributes to a l ~ w e r i n g ' ~ of hii, although the P-pyranose form is known to be the sweetest.' The enthalpic data do not seem to account for the small changes in the dynamics of water that are involved in the sweet-taste mechanism.Vibrational Spectroscopy The interpretation of F.t.i.r. spectra of small carbohydrates and their chlorinated derivatives may help in defining their glucophores (AH-B, 7). The sharp vibration (at 3528 cm-' for P-D-fructopyranose, 3565 cm-' for sucrose and 3460 cm-' for TGS) (see fig. 2) corresponds to hydroxy groups that do not take part in the hydrogen bonding of the crystal. Difference in sweetnessz2 between D-glucose and D-fructose (see table 3) may be explained by the difference in their i.r. spectra. While only one tripartite glucophore is generally accepted for P-D-fructopyranose, two AH-B couples of hydrogen bonds are possible in the case of D-glucose (see fig. 8).The uniqueness of AH-B groups is due to the presence in the i.r. spectrum of p-D-fructopyranose of a sharp absorption assigned to free OH and to the fact that the p-pyranose isomer is preponderant in aqueous solution. The role of the hydrophobic y-centre in enhancing sweetness is clear, since in the case of D-fructose this function is occupied8 by a methylene group, which is more hydrophobic than CH,OH. The slightly higher sweetness of D-glucose as compared to D-galactose (see table 3) is also due to the ' y ' centre, and has its origin in the differenceM . Mathlouthi and A-M. Seuvre 2647 Table 2. hit coefficients of the excess enthalpies of aqueous solutions of small carbohydrates [in J mol-' (mol kg-l)-'] -~ our results sugar exptl calcd Barone et al.l7 D-glucose 290f20 279 343+ 10 D-galactose 150f10 279 133+8 D-fructose 214f22 279 2642 18 sucrose 596k17 521 577 +_ 6 maltose 439k22 521 - lactose 507 k 6 52 1 506 k 32 Table 3. Sweetness of the investigated sugarszz sugar sweetness D-glucose 0.7 D-galactose 0.4 D-fructose 1.2 sucrose 1 ma1 to se 0.5 lactose 0.4 TGS 2000 ~ _ _ _ B AH ' Fig. 8. Representation of (AH-/I, y ) glucophores proposed for /I-D-fructopyranose (a) and D-glucose (b) molecules.2648 Solution Properties and Sweetness AH Fig. 9. Representation of the mechanism of sweetness of TGS-2(AH-P, y ) glucophores and opposed hydrophobic and hydrophilic sides of the molecule. Water molecules (v) are repelled by the hydrophobic side. in isomerism of the CH,OH group in the two monosaccharide^.^^ The extremely high sweetness of TGS is probably due to (i) the spatial disposition of chlorine substituents, in such a way that the hydrophobic end (CH,Cl groups) is opposed to the hydrophilic end of the molecule (see fig.9) and (ii) the unshared hydrogen bonding with the site manifested by the presence in the F.t.i.r. spectrum of the solid form of a very intense OH vibration at 3460 cm-l (see fig. 2). Although the synergistic effect of a two-glucophore system was proposed2’ for TGS, it is likely that the extreme opposition of hydrophobic and hydrophilic effects in the two ends of the molecule is at the origin of its enhanced sweetness. The role of water in the sweet taste mechanism of small carbohydrates may be derived from the laser Raman spectra of their aqueous solutions.In order to minimize the effect of the OH stretching from the sugars on the Raman spectrum of water, only the laser Raman spectra of traces (1 0-‘--1 OP2 g g-’) of small carbohydrates in water were analysed. Fig. 3 shows that low concentrations of D-glucose and sucrose do not provoke a noticeable change in the Raman spectrum of water, whereas an important modification of the Raman intensity is shown for the spectra of D-fructose solutions. A 1 YO solution of freshly dissolved D-fructose induces a relatively high decrease in intensity, and a lop6 g g-’ concentration leads to an increase in the overall vibrational energy of water. Such an increase in the OH stretching intensity may be due to an increase in the number free OH groups resulting from a ‘structure breaking’ effect of traces of D-fructose on water.The marked difference in behaviour in an aqueous medium of D-fructose as compared with D-glucose or sucrose has been thoroughly studied.” The chaotropic effect of D-fructose on the structure of water has been deduced from an interpretation of results obtained using different techniques. In order to quantify the perturbation of water structure by D-fructose, a semiempirical method’& of deconvolution of the experimental Raman band of liquid water was applied. Deconvolution of the observed spectrum into four gaussian components was achieved, as may be observed in fig. 4 and 5 for pure water and a 1 % D-fructose solution, respectively. Computed values of the intensities, widths and areas of the four components are given in tables 4 and 5.Analysis of these results shows that despite a contribution from the 0-H stretch of the sugar, the overall vibrational energy (total area) shows a decrease of 4.2 %. Detailed variations of frequencies and integrated intensities (areas) are summarized in table 5. The chaotropic effect of D-fructose is manifested in the shift of frequencies of components ( a ) and (b) towards higher frequencies, indicating a weakening of hydrogen bonding in theseM. Mathlouthi and A-M. Seuvre 2649 Table 4. Position of maximum (v), intensity ( I ) width, area and assignments of the gaussian components of the Raman band of water _ ~ _ __ total component vlcm-' I width area area (YO) assignments (4 3223 268 188 50 384 48.18 quasi-crystalline 3413 248 159 39432 37.70 solid-like amorphous 3547 94 140 13 160 12.58 liquid-like amorphous 3638 27 59 1593 1.52 unassociated H,O and (4 (4 (4 total area 104 569 free OH Table 5.Position of maximum (v), intensity ( I ) , width, area shifts in frequencies (Av) and variation of area [A(area)] for the gaussian components of the Raman band of water with 1% D-fructose total component v/cm-l I width area area (%) Avlcm-l A(area) (YO) 3228.3 259 193 49987 49.9 5.3 -0.78 4.4 - 6.30 3417.4 228 162 36936 36.87 - 11.90 3547 3638 28 59 1652 1.64 - 3.70 total area 100 167 - 4.20 (4 (4 (4 (4 84 138 11592 11.57 - species. The most affected fraction, which shows a decrease of ca. 12 % in area, is the less rigidly organized one [component (c)]. An increase in the mobility of water molecules is shown from the increase in total vibrational energy of unassociated molecules [component (41.Depolarization ratios ( p = I , , / I J were calculated and are plotted as a function of Raman frequency for H,O and dilute solutions (10-6-10-2 g g-') of the sugars (see fig. 6). The sigmoid shape of p and the effect of increasing the sugar concentration from to lo-' w/w is comparable to the depolarization ratio spectra as a function of temperature. '3 23 Indeed, except for D-fructose, the depolarization ratio is an increasing function of sugar concentration and a decreasing function of temperaturels3 23 between 3200 and 3600cm-l. Observation of Raman band shapes (fig. 3) and depolarization ratio spectra (fig. 6) and a comparison of the variation of p as a function of sugar concentration with the temperature effect'@$ 23 leads to the conclusion that D-glucose and sucrose have a structure-making effect on water, whereas D-fructose acts as a structure- breaker.These structural features should be considered when properties of these sugars such as cryoprotection, water activity (A,) depression or sweet taste are interpreted. The behaviour of D-fructose with regard to these properties is clearly differentiated. It is well known that D-fructose is a bad cryoprotector for the freezing of living cells, probably because, although hydration water around D-fructose molecules is not frozen, its mobility is high enough to provoke ion migration and an alteration of the isotonic equilibrium of the cryoprotected cell. The same reason may be evoked to explain that despite a lowering of A , to values below the limit of mould growth, these micro- organisms may develop in the presence of D-fructose and not at the same A , obtained with another sugar. As for the sweet taste reception, the role of the mobility of water molecules around the sweetener molecule seems to be essential in increasing the sweetness. Indeed, the relatively high sweetness of D-fructose as compared t o other natural sugars may be due to the higher mobility of water around it.This leads to2650 Solution Properties and Sweetness rupture of the iso-osmotic equilibrium between saliva and blood serum on both sides of the receptor membrane, which results in a more active Na+/K+ transfer and therefore a more intense sweet taste.However, to allow the process of ion transport across the membrane to start, a membrane potential is needed. Such a potential is induced by the stereochemical fitting of the sweet molecule with the receptor. Moreover, as observed for TGS and to a lesser extent for D-fructose, the trigger action of the stimulus, i.e. the membrane potential, depends on the sharp opposition of the hydrophilic and hydrophobic sides of the sweetener. Solute-solvent interactions of small carbohydrates together with a knowledge of their structure, help in explaining the differences between their sweet tastes. The effect of these molecules on the mobility of water seems to be a parameter of particular interest in this field. We thank Dr D. V. Luu (U.S.T.L. Montpellier) for recording the Raman spectra.References 1 R. S. Shallenberger and T. E . Acree, Nature (London), 1967, 216, 480. 2 L. B. Kier, J. Pharm. Sci., 1972, 61, 1394. 3 R. S. Shallenberger and M. G. Lindley, J . Sci. Food Agric., 1975, 26, 117. 4 R. S. Shallenberger and M. G. Lindley, Food Chem., 1977, 2, 145. 5 G. G. Birch, S. Z. Dziedzic, R. S. Shallenberger and M. G. Lindley, J. Pharm. Sci., 1981, 70, 277. 6 F. Pautet and C. Nofre, 2. Lebensm.-Unters.-Forsch., 1978, 166, 167. 7 F. Lelj, T. Tancredi and C. Toniolo, J. Am. Chem. Soc., 1976, 98, 6669. 8 M. Mathlouthi, Food Chem., 1984, 13, 1. 9 S. Shamil, G. G. Birch, M. Mathlouthi and M. N. Clifford, Chem. Senses, 1987, 12, 397. 10 Y. Kobatake, K. Kurihara and N. Kamo, in Surface Electrochemistry, ed. T . Takamura and 11 L. Hough, S. P. Phadnis and E. Tarelli, Carbohydr. Res., 1975 44, 37. 12 M. Huggins, J . Am. Chem. Soc., 1942, 64, 2716. 13 M. Mathlouthi, A. M. Seuvre and G. G. Birch, Carbohydr. Res., 1986, 152, 47. 14 M. Mathlouthi and D. V. Luu, Carbohydr. Res., 1980, 78, 225. 15 A. M. Meffroy-Biget and A. Unanue, C. R. Acad. Sci., Ser. C, 1977, 284, 57. 16 J. J. Savage and R. H. Wood, J. Solution Chem., 1976, 5, 733. 17 G. Barone, P. Cacace, G. Castruonuovo and V. Elia, Carbohydr. Res., 1981, 91, 101. 18 C. Luu, D. V. Luu, F. Rull and F. Sopron, J . Mol. Struct., 1982, 81, 1. 19 J. A. Kanters, R. L. Scherrenberg, B. R. Leeflang, J. Kroon and M. Mathlouthi, Carbohydr. Res., 20 A. M. Meffroy-Biget, J. Chem. Res. ( M ) , 1978, 4, 2650. 21 M . Mathlouthi, D.Sci. Thesis (Dijon University, 1980). 22 R. Khan and also W. M. McNicol, in Sugar: Science and Technology. G. G. Birch and K. J. Parker 23 K. Cunningham and P. A. Lyons, J. Chem. Phjvs., 1973, 59, 2132. A. Kozawa (Japan Scientific SOC. Press, Tokyo, 1978), pp. 1-63. 1987, submitted. (Applied Science Publishers, London 1979). Paper 7/1085; Received 23rd November, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402641

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1988, 84(8), 2651-2665 Thermodynamic Studies of Transfer of Some Amino Acids and Peptides from Water to Aqueous Glucose and Sucrose Solutions at 298.15 K Rajiv Bhat, Nand Kishore and Jagdish C. Ahluwalia* Department of Chemistry, Indian Institute of Technology, New Delhi I10 016, India Partial molar heat capacities and volumes for a homologous series of amino acids and dipeptides have been measured in aqueous I mol kg-I glucose and sucrose solutions at 298.15 K using flow microcalorimetry and densimetry, respectively. These data have been utilized, in conjunction with the data obtained for them in water earlier, to deduce the partial molar heat capacities and volumes of transfer from water to 1 mol kg-I aqueous glucose and sucrose solutions. A comparison of these transfer parameters with similar ones in aqueous sodium chloride and calcium chloride solutions is made.The results are explained using the cosphere overlap model and the factors governing the stability of proteins in these sugar solutions have been discussed. Sugars and polyols help in stabilizing the native conformation of globular proteins. ’-’ The effect of mono- and poly-hydric alcohols on the reversible denaturation of enzymes has also been studied extensively.1°-14 Recently Uedaira and Uedaira’’, Back et a1.l’ and Fujita et aE.17 have studied the effect of a variety of sugars and polyols on the thermal transition of lysozyme and other proteins and enzymes and tried to correlate the stabilizing effect of sugars and polyols to the number and configuration of the OH groups present in them.More recently Busby and Inghaml* demonstrated the ability of various sugars and their derivatives to stabilize the plasma proteins ‘ antithrombin 111’ against heat. It has been concluded that the number and position of the OH groups of sugars has no effect on the normal stability of the protein, whereas the incorporation of carboxylate groups into the sugar molecules enhances their ability to stabilize the protein. In addition to studies on the effect of sugars and polyols on proteins and enzymes, the effect of various carbohydrates in preserving the structural and functional integrity of membranes has been studied by a number of investigator^.^^'^^ Although some trends correlating the stabilising potency of sugars and polyols with the number or configuration of the hydroxy groups have been noted,l’ there are numerous exceptions,ls*lg and all proteins and enzymes do not respond equally to a given compound.Thus our understanding of the mechanism of stabilization of proteins and enzymes by these additives is still incomplete. In order to understand the nature of interactions of sugars with proteins in aqueous solutions, we have studied the partial molar heat capacities and the partial molar volumes of some amino acids and peptides in aqueous solutions of glucose and sucrose. Among various physical parameters, these thermodynamic parameters have been recognised as being sensitive to structural changes occurring in solutions. Moreover, model compound studies have been necessitated owing to the complex structural organisation of the biological macromolecules.265 I2652 Transfer of Amino Acids from Water to Sugar Solutions Experiment a1 A Picker dynamic flow microcalorimeter was employed to measure the apparent molar heat capacities of the solutions. The experimental set-up of the microcalorimeter, including its operational procedure, has been described elsewhere.2' The micro- calorimeter has a precision of 0.5 % and a limit of detectability of 7 x J K-' g-l. A Systronics digital multimeter was used to measure the current and voltage values from the calorimeter, and the recordings were made on a Bryans-28000 potentiometric strip- chart recorder. The chemical calibration was performed by measuring the heat capacities of aqueous sodium chloride solutions at 298.15 K.A correction of 4% in the power output of the calorimeter was required after comparing our data with those of Picker et a1.21 Solution densities were measured with the help of a vibrating-tube digital density meter (model DMA 60/602, Anton Paar, Austria) which has been described elsewhere.22 The temperature around the density-meter cell was maintained by circulating water from a constant-temperature bath through the metal block of the cell at a flow rate of 3 dm3 min-I. The bath temperature was controlled to within K with the help of a Tronac PTC-40 proportional temperature controller and an MK-70 ultra-cryostat. The density meter was calibrated with dry air and water each day, and the density of water at 298.15 K was taken as 0.997047 g cm-3 from the data of Kell.23 The reproducibility in the density measurements was better than 3 x g ~ m ; ~ .The chemical calibration of the density meter was performed by measuring densities of aqueous sodium chloride solutions. The agreement with the literature values was excellent. All the measurements were done at 298.15 K and the solutions were made up by weight , The amino acids and peptides used in the present studies were as follows : glycine, D,L- a-alanine, D,L-a-amino-n-butyric acid, valine, leucine, diglycine, glycylalanine, glycyl- D,L-a-amino-n-butyric acid, glycylvaline, glycyl-leucine and triglycine. All were from Sigma, except glycine and valine, which were from E. Merck, and diglycine, which was from B.D.H. Analytical reagent grade glucose and sucrose, from Sarabhai Chemicals and B.D.H., respectively, were used as obtained.All the amino acids and peptides, together with glucose and sucrose, were dried over P20, in a vacuum desiccator. The moisture content of the samples was checked using a Karl Fisher Aquatest-IV trace-water analyser. All the samples were found to be completely dry. Water used for making the solutions was distilled and deionized by passing it through a Barnstead mixed-bed ion-exchange resin column. Deionized water was further distilled with alkaline KMO, to remove any organic matter and then degassed. 1 x Results The apparent molar heat capacities ($J and the apparent molar volumes ($v) were evaluated from the specific heat capacity and the density data using the following equations : q5c = MC,-(C~-C,) 103/m $v = ~ / n - (n- do) 103/mnn,, (2) where M is the molar mass of the solute, m is the molality and Ci and C , are the specific heat capacities of the reference solvent and the solution, respectively.Since 1 mol kg-l aqueous glucose and sucrose solutions were used as the solvents for studying the transfer parameters of the amino acids and peptides, C; and C, are therefore the specific heat capacity for the sugar solutions (which were measured in each experiment separately) and that of the ternary amino acid (or peptide) - sugar - water system,R. Bhat, N . Kishore and J . C. Ahluwalia 2653 Table 1. Apparent molar heat capacities ($J and volumes (&) of some amino acids and peptides in 1 mol kg-I aqueous glucose solutions at 298.15 K m/mol kg-' 0.101 353 0.106 632 0.1 12 874 0.197 010 0.231 450 0.248 021 0.305 141 0.316 236 0.421 790 0.445 281 0.044 499 0.057 887 0.099 873 0.103 933 0.162 533 0.201 590 0.230 21 1 0.250 898 0.391 459 0.603 805 0.107 884 0.1 14 984 0.132 661 0.147 464 0.152 295 0.067 669 0.077 331 0.084 387 0.086 971 0.099 753 0.117 953 0.157 069 0.160 468 0.055 164 0.070 336 0.087 061 0.090 341 0.091 906 0.098 370 0.124 185 glycine 1.057 493 - 1.057 493 - 1.060 804 - 1.063 802 - 1.064 845 65.6 1.064 861 65.5 1.066 939 65.3 1.067 193 - 1.069 742 65.0 1.070 796 65.7 D,L-a-alanine 1.058 586 - 1.059 414 156.0 1.059 992 - 1.060 588 158.3 1.062 104 158.7 1.063 138 154.3 1.063 188 - 1.064 524 156.2 1.067 097 - 1.072 035 - D,L-a-amino-n-butyric acid 1.060 326 1.060 760 1.060 903 1.061 244 1.061 618 L-valine 1.058 967 1.059 320 1.059 381 1.060 452 1.060 042 1.060 937 - - L-leucine 1.058 606 1.058 866 1.059 533 1.059 204 1.059 685 1.059 699 1.059 783 238.4 236.6 242.1 239.5 - 311.5 312.2 - - 31 1.1 31 1.7 31 1.8 31 1.0 403.8 403.2 399.7 399.1 402.3 402.3 - Qv/crn3 mol-' 44.77 44.58 44.62 44.34 44.46 44.70 44.56 44.85 44.7 1 - 61.47 61.76 61.40 - - 61.56 61.65 61.66 61.80 - 76.39 76.29 76.36 76.37 76.33 91.66 91.54 91.57 91.69 91.49 91.46 - - 108.32 108.35 108.42 - - - 108.402654 Transfer of Amino Acids from Water to Sugar Solutions Table 1.(con?.) WJ K-' rnlmol kg-' d/g mol-' &/cm3 mol-' ~ dig1 ycine 0.051 981 1.060 219 - 0.055 025 1.060 327 0.076 033 1.061 442 - 0.118 396 1.063 580 - 0.132 195 1.064 812 143.9 0.213 640 1.068 916 146.4 0.238 877 1.070 106 142.4 0.250 924 1.070 213 - 0.263 770 1.070 785 0.288 122 1.072 524 145.7 0.321 730 1.074 224 143.6 0.041 256 1.059 452 241.9 0.045 290 1.059 767 - 0.046 568 1.059 720 240.1 0.049 632 1.060 224 242.0 0.078 274 1.061 602 242.1 0.083 562 1.062 240 - 0.084 901 1.061 553 244.8 0.103 687 1.062 475 243.8 0.256 222 1.070 567 0.366 827 1.075 735 - - - glycyl-D,L-ol-alanine - glycyl-D,L-ol-amino-n-butyric acid 0.035 324 0.057 514 0.060 049 0.083 253 0.091 916 0.102 690 0.052 049 0.066 093 0.074 514 0.088 733 0.051 835 0.060 988 0.066 942 0.078 428 0.085 539 0.038 938 0.061 736 0.076 692 0.082 370 0.086 802 0.087 9 18 0.1 10 818 1.059 094 1.060 858 314.5 1.060 308 3 15.8 1.062 050 3 15.9 1.062 418 - 1.062 908 318.7 - glycyl-L-valine 1.060 456 366.0 1.061 107 366.0 1.061 489 369.6 1.062 110 370.5 1.060 152 - 1.060 555 472.8 1.060 750 479.6 1.061 222 474.9 1.061 531 479.0 1.060 284 - 1.061 791 - 1.063 642 229.0 1.063 314 - 1.064 316 234.0 1.064 398 23 1.8 1.065 21 2 - glycyl-L-leucine trig1 ycine 77.75 77.97 77.91 77.81 77.62 77.91 78.08 78.19 78.00 77.85 93.67 93.84 93.56 93.90 93.55 94.06 93.61 93.67 - - - 108.98 - - 108.74 109.1 1 109.18 123.36 123.20 123.22 123.02 140.87 140.42 141.10 140.99 140.75 114.32 114.49 114.46 - - - 114.51R.Bhat, N . Kishore and J . C. Ahluwalia 2655 Table 2. Apparent molar heat capacities ($J and volumes (q&) of some amino acids and peptides in 1 mol kg-' aqueous sucrose solutions at 298.15 K rn/mol kg-' &/J K-' d/g mol-' &/cm3 mol-l 0.121 856 0.161 674 0.282 943 0.294 832 0.338 197 0.485 492 0.160 767 0.193 048 0.227 165 0.251 558 0.231 461 0.080 386 0.124 653 0.147 183 0.204 309 0.235 139 0.312 733 0.085 331 0.099 753 0.121 392 0.157 069 0.194 897 0.031 348 0.037 833 0.037 129 0.040 925 0.050 369 0.053 814 0.071 225 0.078 766 0.094 633 0.054 124 0.090 927 0.128 152 0.045 382 0.061 735 0.072 061 0.099 870 0.1 14 776 glycine 1.106 501 75.2 1.107 601 74.5 1.110 886 75.5 1.111 271 - 1.1 13 838 74.3 1.1 16 31 1 - 1.108 390 156.9 1.108 390 157.4 1.109 958 156.0 1.109 895 - 1.112 180 158.8 D,L-a-alanine D,L-a-amino-n-butyric acid 1.106 312 1.107 370 1.107 827 1.108 988 1.109 576 1.1 10 966 1.106 309 1.106 639 1.107 031 1.107 597 1.108 235 1.103 548 1.105 193 1.103 625 1.103 659 1.105 289 1.103 800 1.104 025 1.105 725 1.105 841 1.106 338 1.108 073 1.1 10 009 L-valine L-leucine dig1 ycine 243.5 242.0 240.7 242.2 241.8 244.6 310.7 311.1 311.8 312.5 - - 403.0 - - 402.3 405.8 398.5 157.1 161.6 161.6 - - glycyl-D,L-a-alanine 1.105 271 250.9 1.106 076 250.4 1.106 564 249.8 1.107 742 254.8 1.108 437 253.9 45.02 44.96 44.99 44.8 1 45.05 44.95 61.58 61.69 61.61 61.65 61.62 - 76.43 76.57 76.57 76.59 76.59 - 91.70 91.55 91.81 91.80 108.64 108.82 108.44 108.54 108.54 108.67 108.37 - - 78.13 78.07 78.01 - 93.86 93.73 93.92 93.802656 Transfer of Amino Acids from Water to Sugar Solutions Table 2.(cont.) $JJ K-l rn/mol kg-’ d/g mol-’ &/crn3 mol-I glycyl-D,L-a-amino-n-butyric acid 0.046 036 1.105 311 - 0.049 069 1.105 481 0.050 728 1.105 507 - 0.057 025 1.105 791 0.082 113 320.3 0.099 731 317.9 0.1 15 791 - 321.4 - - - - glycyl-L-valine - 0.051 956 1.105 400 0.073 725 1.106 300 0.075 083 366.4 0.081 871 1.106 610 - 0.084 993 - 364.3 367.1 0.090 465 0.1 12 310 1.107 827 - - - - 0.043 754 0.051 381 0.058 852 0.065 295 0.065 872 0.071 025 0.079 966 0.110 137 gl ycy 1- L-leucine 1.104 908 - 1.105 163 - 1.105 439 - 1.105 639 - 1.105 852 - 483.7 - 482.2 480.1 - - triglycine 0.047 21 1 1.106 393 - 0.049 500 1.106 499 - 0.096 889 1.109 684 - 0.128 174 1.111 750 - 109.56 108.89 109.37 109.18 - - - 123.15 123.21 123.48 - - - 123.66 140.90 141.16 140.70 141.33 141.17 - - __ 114.29 114.31 114.29 114.37 respectively. Similarly, d and do are the densities of the ternary system and the density of the reference solvent (1 mol kg-l aqueous glucose or sucrose), respectively, and m is the molality of the solute (amino acid or peptide) in sugar-water mixtures.q5c and q5v data in 1 mol kg-l aqueous glucose and sucrose solutions are presented in tables 1 and 2, respectively. Since the concentration dependence of $c and $v, were found to be either negligible or having no definite trend and the concentration effect, if any, was found to be masked by the overall uncertainty, $: and & were simply evaluated by taking an average of all the data points, and the standard deviations were evaluated from the mean. Hereafter the apparent molar terms $: and q5; will be replaced by the more commonly used partial molar terms C”,z and c, since at infinite dilution they have the same meaning. data for all the amino acids and peptides in 1 mol kg-l aqueous glucose and sucrose solutions are listed in tables 3 and 4, respectively, along with their respective standard deviations.The partial molar heat capacities of transfer (C;,, ,,) and the partial molar volumes of transfer ( c,t,.) for the amino acids and peptides from water to 1 mol kg-l aqueous glucose and sucrose solutions are also presented in tables 3 and 4, C”,z andTable 3. Infinite-dilution partial molar heat capacities and volumes of some amino acids and peptides in water and 1 mol kg-' aqueous glucose solutions, together with their corresponding transfer parameters at 298. 15 K" c - _ - _ ________ -__- _ _ _ ~ _ . -- .- - _ _ _ _ _ _ ~ ~ - _____ ~ compound d",,/J K-' mo1-' c / c m 3 mo1-' C.,,, J.J K-1 moI-1 q, ~ c m 3 mol-1 rQ ( I mol kg-' (water+ 1 mol kg-' (water+ 1 mol kg-I CZ,,/J K-' mo1-1 r;/cmi3 mo1-l (1 mol kg-l c2 hc (water) (water) glucose) glucose) glucose) glucose) glycine 39.2 (0.4)b D,L-a-alanine 141.4 (0.2)b D,L-a-amino-n-butyric acid 234.1 (l.O)d L-valine 307.5 (l.O)e L-leucine 399.0 (l.O)d dig1 ycine 105.0 (0.4)b glycyl-D,r,-z-alanine 221.5 (6.5) glycyl-D,L-a-amino-n-butyric acid 306.2 (3.6) glycyl-L-valine 356.4 (5.3) gl ycyl-L-leucine 473.3 (0.8) triglycine 185.9 (0.9)b 43.19 (0.02)c 60.42 (0.02)" 75.50 (0.02)" 90.79 (O.O1)f 107.83 (0.03)' 76.23 (0.03)' 92.37 (0.02)' 107.8 1 (0.05)' 139.70 (0.07)" 1 12.1 1 (0.03)" 121.99 (0.02)' 65.4 (0.3) 156.7 ( I .8) 239.1 (2.3) 311.5 (1.0) 401.7 (2.2) 144.4 (1.6) 242.4 (1.6) 316.2 (1.8) 368.0 (3.7) 476.6 (3.3) 231.6 (2.5) 44.62 (0.16) 61.61 (0.15) 76.35 (0.04) 91.57 (0.09) 108.37 (0.05) 77.91 (0.16) 93.73 (0.18) 109.00 (0.20) 123.20 (0.14) 140.83 (0.26) 114.44 (0.15) 26.2 (0.5) 15.3 (1.8) 5.0 (2.5) 4.0 (1.1) 2.7 (2.4) 39.4 (1.6) 20.9 (6.7) 10.0 (4.0) 11.6 (6.5) 3.3 (3.5) 45.7 (2.6) 1.43 (0.16) *-- 1.19 (0.15) 2 0.85 (0.05) 0.78 (0.09) 2 0.54 (0.06) 1.68 (0.16) % 1.36 (0.20) C 1.19 (0.21) 9 1.21 (0.14) 1.13 (0.27) !$ 2.33 (0.16) a Entries in parentheses are standard deviations.' Ref. (40). ' Ref. (41). Ref. (42). Ref. (27). Ref. (29). 5:2 Table 4. Infinite-dilution partial molar heat capacities and volumes of some amino acids and peptides in water and 1 rnol kg-I aqueous sucrose solutions, $ together with their corresponding transfer parameters at 298.15 K" .a compound glycine D,L-a-alanine D,L-a-amino-n-butyric acid L-valine L-leucine dig1 ycine glycyl-D,L-a-alanine glycyl-D,L-a-amino-n-butyric acid glycyl-L-valine gl ycyl- L- leuci ne trig1 ycine ~ ~ ~~ - - - _ _ _ _ _ _ _ _ ~ ~ ~ ~- - C",,/J K-' mo1-I c / c m 3 mol-I Gz, JJ K-I mol-' c, Jcm3 mo1-' C",,/J K-' mol-I c / c m 3 mol-' (1 rnol kg-I (1 rnol kg-' (water-+ 1 rnol kg-' (water+ I rnol kg-I (water) (water) sucrose) sucrose) sucrose) sucrose) 39.2 (0.4)b 141.4 (0.2)b 234.1 (l.O)d 307.5 (l.O)e 399.0 (l.O)d 105.0 (0.4)b 221.5 (6.5) 306.2 (3.6) 356.4 (5.3) 473.3 (0.8) 185.9 (0.9)b 43.19 (0.02)" 60.42 (0.01)" 75.50 (0.02)' 90.79 (O.O1)f 107.83 (0.03)" 76.23 (0.03)r 92.37 (0.02)" 107.8 1 (0.05)" 121.99 (0.02)C 139.70 (0.07)" 1 12.1 1 (0.03)" 74.9 (0.6) 157.3 (1.1) 242.5 (1.4) 311.5 (1.0) 402.4 (3.0) 160.1 (2.6) 252.0 (2.4) 319.9 (1.8) 365.9 (1.5) 482.0 (1.8) - 44.96 (0.08) 61.63 (0.04) 76.55 (0.07) 91.71 (0.12) 108.57 (0.21) 78.07 (0.06) 93.82 (0.09) 109.25 (0.28) 123.38 (0.24) 141.05 (0.25) 114.31 (0.04) 35.7 (0.7) 15.9 (1.2) 8.4 (1.7) 4.0 (1.5) 55.1 (2.6) 30.5 (6.9) 13.7 (4.0) 8.7 (2.0) 3.4 (3.3) 9.5 (5.5) - 1.77 (0.08) 1.21 (0.05) 1.05 (0.09) 0.92 (0.12) 0.74 (0.22) 1.84 (0.07) I .45 (0.10) 1.44 (0.30) 1.39 (0.25) 1.35 (0.30) 2.20 (0.05) z a Entries in parentheses are standard deviations. Ref.(40). " Ref. (41). Ref. (42). Ref. (27). Ref. (29). 6- 3 bR . Bhat, N . Kishore and J. C. Ahluwalia 2659 Table 5. Amino-acid side-chain contributions towards partial molar heat capacities and volumes of transfer from water to 1 mol kg-' aqueous glucose and sucrose solutions at 298.15 K G2JJ K-' mol-' c, Jcm3 mo1-l __ ___-_ __ water -, 1 water + 1 water -+ 1 water 1 amino-acid mol kg-' mol kg-' mol kg-' mol kg-' side chain glucose sucrose glucose sucrose method CH,CH3 -21.2 - 27.3 -0.58 - 0.72 Aba-Gly CH(CH3)2 - 22.2 -31.7 - 0.65 -0.85 Val-G1 y CH2CH(CH3)2 -23.5 - 32.3 - 0.89 - 1.03 Leu-G1 y - 29.4 -41.4 - 0.49 - 0.40 Gl yAba-Gly, - 27.8 -45.6 - 0.47 - 0.45 Gl yVal-Gly , - 36.1 - 46.4 -0.55 - 0.49 Gl y Leu-Gly , Table 6.Peptide backbone unit contributions toward partial molar heat capacities and volumes of transfer from water to 1 mol kg-' aqueous glucose and sucrose solutions at 298.15 K water -, water -+ water + water --+ 1 mol kg-' 1 rnol kg-' 1 rnol kg-' 1 mol kg-' glucose sucrose glucose sucrose 13.2" 19.4" 0.25" 0.07" 5.6' 14.6' 0.1 7' 0.24' 5.0" 5.3" 0.34" 0.39" 7.6d 5 S d 0.43d 0.47d 0.6" 5.3" 0.59e 0.61' 6.3f -_ 0.65f 0.36f " Gly,-Gly.' GlyAla-Ala. " GlyAba-Aba. GlyVal-Val. " GlyLeu-Leu. f Gly3-Gly2. respectively. The Ci2 and c values for the amino acids and peptides in water reported earlier are included in tables 3 and 4 to facilitate the visualization of the transfer parameters. The transfer parameters were evaluated as follows : C",, tr or c, tr [water to glucose or sucrose (aq)] Ci2 or [in glucose or sucrose (aq)]-Ci2 or c (in water) (5) Amino-acid side-chain contributions, derived from the homologous series of amino acids and peptides, to cp2, tr and c, tr are presented in table 5. Contributions from the peptide backbone unit CH2CONH to these transfer parameters are given in table 6.In fig. 1 and 2 we have plotted C;2,tr and c,,, for the homologous series of amino acids from water to 1 mol kg-l glucose and sucrose solutions as a function of the number of carbon atoms in their alkyl chains, respectively. The cp2,tr and c,,, values for the glycine dipeptides from water to 1 mol kg-' glucose and sucrose have also been plotted as a function of the number of carbon atoms in their alkyl chains and are shown in fig. 3 and 4, respectively, taking diglycine as having one carbon atom in its alkyl chain.2660 Transfer of Amino Acids from Water to Sugar Solutions 40 50 t Fig. 1. Partial molar heat capacities of transfer of a-amino acids from water to 1 mol kg-I glucose (0) and 1 mol kg-' sucrose (A) solutions at 298.15 K as a function of the number of carbon atoms in the alkyl chain, n,.3 I I 1 1 2 3 5 nC I Fig. 2. Partial molar volumes of transfer of a-amino acids from water to 1 mol kg-' glucose (0) and 1 mol kg--' sucrose (A) solutions at 298.15 K as a function of the number of carbon atoms in the alkyl chain, n,.R. Bhat, N . Kishore and J . C . Ahluwalia I 266 1 I I I I I 1 2 3 1, 5 n C -5 I Fig. 3. Partial molar heat capacities of transfer of glycine peptides from water to 1 mol kg-l glucose (0) and 1 mol kg-' sucrose (A) solutions at 298.15 K as a function of the number of carbon atoms in the alkyl chain, n,. I 3 'i 't z 1 2 3 r, 5 n C Fig. 4. Partial molar volumes of transfer of glycine peptides from water to 1 mol kg-' glucose (0) and 1 mol kg-' sucrose (A) solutions at 298.15 K as a function of the number of carbon atoms in the alkyl chain, n,.2662 Transfer of Amino Acids from Water to Sugar Solutions Discussion The C;,, ,,and c, ,, values for transfer of a-amino acids and glycine peptides from water to 1 mol kg-’ aqueous glucose and sucrose solutions are found to be positive and decrease with increasing alkyl chain length.These results can be explained by the cosphere overlap model, as developed by Gurney24 and Frank and Evans.25 Properties of the water molecules in the hydration cosphere depend on the nature of the solute species.26 The types of interactions occurring between the sugar molecules and the amino acids or peptides can be classified as : (i) hydrophilic-ionic group interactions between the OH groups of the sugars and the Zwitterionic centres of the amino acids or peptides, (ii) hydrophilic-hydrophilic group interactions between the OH groups of the sugars and the -CONH group of the peptides mediated through hydrogen bonding and (iii) hydrophilic-hydrophobic group interactions between the OH groups of the sugars and non-polar (-CH,) groups of the amino acids and peptides.Interactions of type (i) lead to a positive contribution to C”,, tr and c, tr, since owing to the overlap of the hydration cosphere of an ion (in this case the NH; and COO- groups) and a hydrophilic OH group, the structure-breaking tendency of the ion and the electrostriction of the solvent caused by these ions are reduced. Moreover, direct interactions between the Zwitterionic and hydroxy groups mediated via the hydrogen-bonded network of water also contributes in a positive way to the heat capacity.Interactions of type (ii) are also expected to make positive contributions to C;,,,, and c,tr, since the overlap of the hydration cospheres of OH and CONH groups should lead to an increase in the magnitude of hydrogen-bonding interactions. On the contrary, interactions between the OH and CH, groups should lead to a decrease in Cop,, tr and c, tr values because of the reduction of water structure that is formed around these groups as a result of their cosphere overlap. The relatively large positive values of Cop,, ,, for lower homologues of amino acids, particularly for glycine and D,L-a-alanine, indicate the predominance of hydrophilic-ionic and hydrophilic-hydrophilic interactions over hydrophilic-hydro- phobic interactions.On the other hand, the smaller values of C”,,,, and c,,, for transfer of the higher homologous of a-amino acids (valine and leucine) indicate a balance of hydrophilic-ionic and hydrophilic-hydrophilic interactions by hydro- philic-hydrophobic interactions owing to the increase in hydrophobicity. In the literature we could find no data available on C”,z,tr for the transfer of amino acids or peptides to sugar solution. However, Di Paola and Bellea~,~ measured C;, and values for some amino acids in sorbitol and mannitol. Our Ci, values for amino acids in glucose and sucrose are in close agreement to theirs in sorbitol and mannitol, showing there that these polyols behave in a similar way to glucose and sucrose as far as their interactions with amino acids are concerned.Regarding the effect of the concentration of glucose and sucrose on C”,, tr and c- tr, owing to the viscous nature of higher-concentration solutions (> 1 mol kg-’) of glucose and sucrose, systematic and reliable studies on C82 and could not be carried out in these solutions. However, for some typical amino acids representing lower homologues (glycine, D,L-a-alanine) and a higher homologue (valine), and for a dipeptide (diglycine), Cop,, tr and c, tll were determined in 0.5 and 1 mol kg-’ glucose and sucrose solutions. The values of C;,,,, and c,tr increased with an increase in the concentration of glucose and sucrose from 0.5 to 1 mol kg-’. The effect was significantly large for glycine and diglycine, less pronounced for D,L-a-alanine and negligible for valine.It may be inferred from this that an increased concentration of glucose and sucrose leads to greater ionic-hydrophilic and hydrophilic-hydrophilic interactions, which are not compensated by a hydrophilic-hydrophobic interaction in the case of glycine and D,L-a-alanine. The results given in tables 3,4 and 7 indicate that the Cz2, tr and c. tr values in sucrose solutions are, in general, larger than those in glucose solutions of the same concentration.R. Bhat, N . Kishore and J . C. Ahluwalia 2663 Table 7. Effect of glucose and sucrose concentration on partial molar heat capacities and volumes of some amino acids in aqueous glucose and sucrose solutions C,/J K-' mol-1 c/cm3 mol-' glucose, sucrose, glucose, sucrose, m/mol kg-' m/mol kg-' m/mol kg-' m/mol kg-I amino acid 0.5 I 0.5 1 0.5 1 0.5 1 ~ _ _ _ glycine 54.15 65.4 62.5 74.9 44.38 44.62 44.58 44.96 r),L-a-alanine 152.2 156.7 154.3 157.3 61.10 61.61 61.12 61.63 r),L-a-anino-n-butyric 239.0 239.1 240.7 242.5 76.04 76.35 76.11 76.55 r,-valine 308.2 311.5 309.8 311.5 90.99 91.57 91.46 91.71 dig1 ycine 122.3 144.4 140.8 160.1 77.40 77.91 77.72 78.07 acid This may be due to the fact that for solutions of the same concentration aqueous sucrose solutions should contain about twice the number of OH groups compared to aqueous glucose solutions.The relative order of C;,,,, and q,,, of amino acids and peptides in aqueous solutions of urea,28* 29 NaC1,30 CaC1,,31 t-butyl glucose and sucrose is found to be: urea > CaC1, > NaCl > sucrose > glucose > t-butyl alcohol.The larger values of C;,, tr and c, tr for amino acids in aqueous solutions of urea, CaCl, and NaCl may be attributed to greater hydrophilic-ionic group interactions between urea and the zwitterions, and stronger ion-ion interactions in the ternary amino acid (or pep tide)-sal t-wa ter sys tem. The Cop,, tr and c, ,, values decrease with an increase in the length of the alkyl chain of the amino acid or peptide in aqueous solutions of sugars, as in solutions of urea, CaCl, and NaCl. This may be explained on the basis that, with an increase in the hydrophobicity of the amino acids and peptides, the larger interactions between the hydrophilic group and the alkyl groups lead to a decrease in C;,,,, and c,,, values because of the decrease in the water structure that is formed around these groups as a result of the overlap of their cospheres.The average decrease in C",z,!, values per CH, group for amino acids and peptides in sugar solutions observed by us is ca. 7 and 10 J K-' mol-', respectively (see fig. 1 and 3), whereas this decrease is larger in aqueous urea solutions (32 J K-l mo1-')28 and close to that in aqueous CaCl, (8 J K-l m ~ l - l ) . ~ l On the other hand, Mishra and Ahluwalia3' observed an increase of ca. 10 J K-l mol-' in the cp2, tr of amino acids in aqueous t-butyl alcohol per CH, group at ca. 1 mol kg-l t-butyl alcohol. From the difference in C",,,, of the glycine peptides and the corresponding amino acids we have evaluated the contribution to C",, tr and c, ,, of the peptide backbone unit (-CH,CONH) in glucose and sucrose solutions (table 6).Except for glycine, the average values of c,, ,, and c, ,, for a peptide backbone unit are ca. 6 J K-' mol-' and 0.4 cm mol-'. The positive values of C;,, ,, and c, ,, for a peptide backbone unit indicates that the interactions of the OH groups of sugars with the CONH part dominate over the OH-CH, group interactions. The magnitude of CEO, tr for the -CONH group (ca. 16 J K-l mol-l) from water to the sugar solutions is much smaller as compared to that in the neutral-salt (NaCl and CaCl,)30~3' solutions. Since CaCl, denatures some proteins, and if we consider NaCl to be on the borderline in its effectiveness to denature proteins [as is evident from their being no change in the transition temperature (T,) of the enzyme ribonuclease on the addition of sodium chloride, even at higher concentrations (1 .5-2.5)],33 we conclude that glucose and sucrose owe their stabilizing effect on proteins to their having relatively2664 Transfer of Amino Acids from Water to Sugar Solutions smaller interactions with the peptide groups.The higher volumes of transfer for the amino acids and peptides from water to aqueous solutions of NaCl and CaCl, support the heat-capacity results in that there are large positive volume contributions due to ion-ion interactions. Another factor that has been shown to effect protein stability in sugar solutions is their relative water-structure-promoting tendencies.Lakshmi and Nandi3* and U e d a i ~ a ~ ~ have obtained positive AGtr values for amino acids and peptides in sugar solutions relative to water. This has been explained as being due to the difficulty of placing non-polar groups in sugar solutions relative to water owing to stronger sugar-water hydrogen bonds. 36 Oakenfull and Frenwick3’ and Bull and B r e e ~ e ~ ~ have also suggested that those sugars which have a maximum water-structure-making potency are more effective stabilizers. These results, coupled with ours, indicate that sugar solutions form a solvent system in which exposed peptide groups attached to non-polar groups have a tendency to enter the protein interior because of the unfavourable environment produced by the sugar molecules; since the interactions of the peptide groups with the sugar molecules are not appreciable, the former should have no tendency to move out of the protein interior.References 1 A. Beilinson, Biochem. Z., 1929, 213, 399. 2 C. D. Ball, D. T. Hardt and W. J. Duddles, J. Biol. Chem., 1943, 151, 163. 3 C. R. Hardt, 1. F. Huddelson and C. D. Ball, J. Biol. Chem., 1946, 163, 211. 4 M. Kerues, Biochim. (Rumania), 1963, 6, 245. 5 As cited in T. S. Lakshmi and P. K. Nandi, J. Phys. Chem., 1976, 80, 249. 6 R. B. Simpson and W. Kauzmann, J. Am. Chem. SOC., 1953, 75, 5139. 7 J. C. Lee and S. N. Timasheff, J. Biol. Chem., 1981, 256, 7193. 8 R. P. Erigon and J. C. Lee, Arch. Biochem. Biophys., 1972, 153, 587. 9 N. L. Simons and R. J. Naftalin, Biochim. Biophys., Acta, 1976, 419, 493. 10 S.Y. Gerlsma and E. R. Stuur, Int. J. Peptide Protein Res., 1972, 4, 377. 11 G. Velicelebi and J. M. Sturtevant, Biochemistry, 1979, 18, 1180. 12 K. Gekko, T. Morikawa and H. Noguchi, 17th Annual Meeting of the Biophys. SOC. Japan, Nagoya, Oct. 1979, abstr. p. 305. 13 B. S. Harrap, Int. J. Peptide Protein Res., 1969, 1, 253. 14 S. Shifrin and C. L. Parrott, Arch. Biochem. Biophys., 1975, 166, 426. 15 H. Uedaira and H. Uedaira, Bull. Chem. Soc. Jpn, 1980, 53, 2451. 16 J. F. Back, D. Oakenfull and M. B. Smith, Biochemistry, 1979, 18, 5191. 17 Y. Fujita, Y. Iwasa and Y. Noda, Bull. Chem. Soc. Jpn, 1982, 55, 1896. 18 T. F. Busby and K. C. Ingham, Biochim. Biophys. Acta, 1984, 799, 80. 19 L. M. Crowe, R. Mouradian, J. H. Crowe, S. A. Jackson and C. Womersley, Biochim. Biophys. Acta, 20 F. Sreter, N. Ikemoto and J. Gergeley, Biochem. Biophys. Acta, 1970, 203, 351. 21 P. Picker, P. A. Leduc, P. R. Philip and J. E. Desnoyers, J. Chem. Thermodyn., 1971, 3, 631. 22 P. Picker, E. Tremblay and C. Jolicoeur, J. Solution Chem., 1974, 3, 377. 23 G. S. Kell, J. Chem. Eng. Data, 1970, 15, 119. 24 R. W. Gurney, in Ionic Processes in Solution (McGraw Hill, New York, 1953). 25 H. S. Frank and M. W. Evans, J. Chem. Phys., 1945, 13, 507. 26 H. L. Friedman and C. V. Krishnan, in Water-A Comprehensive Treatise, ed. F. Franks (Plenum, 27 G. Di Paola and B. Belleau, Can. J. Chem., 1978, 56, 1827. 28 K. P. Prasad and J. C. Ahluwalia, Biopolymers, 1980, 19, 273. 29 C. Jolicoeur, B. Riedl, D. Desrochers, L. L. Lemelin, R. Zamojska and 0. Enea, J. Solution Chem., 1986, 15, 109. 30 R. Bhat and J. C. Ahluwalia, J. Phys. Chem., 1985, 89, 1099. 31 R. Bhat and J. C. Ahluwalia, Znt. J. Peptide Protein Res., 1987, 30, 145. 32 A. K. Mishra and J. C. Ahluwalia, J. Chem. SOC., Faraday Trans. I , 1981, 77, 1469. 33 P. H. Von Hippel and T. Schleich, in Structure and Stability of Biologial Macromolecules, ed. S . N. Timasheff and G. D. Fasman (Marcel Dekker, New York, 1969) , vol. 2, chap. 6. 34 T. S. Lakshmi and P. K. Nandi, J. Phys. Chem., 1976,80, 249. 35 H. Uedaira, Bull, Chem. SOC. Jpn, 1977, 50, 1298. 36 J. B. Taylor and J. S. Rowlinson, Trans. Faraday SOC., 1955, 51, 1183. 1984, 769, 141. New York, 1973), vol. 3, chap. 1.R. Bhat, N . Kishore and J. C. Ahluwalia 2665 37 D. G. Oakenfull and D. E. Fenwick, J . Chem. Soc., Faraday Truns. 1, 1979, 75, 636. 38 H. B. Bull and K. Breese, Biopolymers, 1978, 17, 2121. 39 C. H. Spink and I. Wadso, J. Chem. Thermodyn., 1975, 7, 561. 40 C. Jolicoeur and J. Boileau, Can. J . Chem., 1978, 56, 2707. 41 A. K. Mishra and J. C. Ahluwalia, J. Phys. Chem., 1984, 88, 86. 42 J. C. Ahluwalia. C. Ostiguy, G. Perron and J. E. Desnoyers, Can. J . Chem., 1977, 55, 3364. Paper 612070; Received 23rd October, 1986

ISSN:0300-9599    

DOI:10.1039/F19888402651

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1988, 84(8), 2667-2676 Effect of Micelle Formation on the Absorption Spectra of a Functionalized Detergent with the Anthraquinone Moiety Katsuyoshi Hoshino, Tetsuo Saji, Kosaku Suga and Masamichi Fujihira* Department of Chemical Engineering, Tokyo Institute of Technology, Ohokayama Meguro-ku, Tokyo 152, Japan Measurements of the absorption spectra of aqueous solutions of (2-anthraquinonylmethyl)dodecyldimethylammonium (AQD+) bromide have been performed. From the concentration dependence of the molar absorption coefficient at I,,, = 330 nm ( E , ~ , ) , the c.m.c. and the molar absorption coefficient of AQD+ participating in its micelle formation were calculated by assuming a micellar pseudo-phase model. The change in E , , ~ above the c.m.c. was found to be due to interactions of the anthraquinone moieties, by studying the effect of the addition of hexadecyltrimethyl- ammonium bromide (CTAB) to the AQD+ micellar solution.The effect of the addition of CTAB to the AQD+ monomer solution was also examined, but the spectra of the AQD' monomer were little affected by the addition of CTAB. Measurements of the absorption spectra of aqueous solutions of the monomeric AQD+ and sodium dodecylsulphate (SDS) point to the formation of a stable ion-pair complex between them. On further addition of SDS, the complex disappeared above the c.m.c. of SDS. This amphiphatic anthraquinone derivative serves as a probe for determining the c.m.c. of SDS in the absence and presence of electrolyte because the molar absorption coefficient at 290 nm changes by a factor of 7 above and below the c.m.c.of the host surfac t ant. The effect of micelle formation on absorption spectra was originally investigated by Harkins et a1.l in aqueous solution. They made a study of the absorption spectra of alkylpyridinium iodide (API) and showed that the c.m.c. was determined from the concentration dependence of the molar absorption coefficient at 290 nm. Kitahara2 studied the ultraviolet absorption spectra of alkylammoniumbenzoates (AAB) in cyclohexane and showed that a plot of the absorbance of the absorption peaks us. the concentration of AAB breaks at the point of initiation of reversed-micelle formation, and calculated the molar absorption coefficient of surfactants participating in reversed micelles.Turro and co-workers3~* examined the U.V. spectra of 1 1-(3-hexyl- I-indoly1)un- decyltrimethylammonium bromide (6-In-1 1). Although they did not remark on it, the plot of the absorbance of 6-In- 11 at 320 nm vs. the concentration of 6-In-1 1 deviated from Beer's law above a certain concentration. Gratzel and co-workers5 investigated the U.V. absorption spectra of 1 -alkyl- 1 '-methyl- 4,4'-bipyridinium dibromide and showed that the c.m.c.s determined from the plot of the logarithm of absorbance at 370 nm us. the logarithm of the concentrations were almost equal to those obtained from surface tension measurements. Only a few of these studies were directed to change in u.v.-visible spectra accompanied by the monomer-micelle equilibrium. This paper deals with the U.V.absorption spectra of AQD+, and shows that (1) the value of the molar absorption coefficient at Amax = 330 nrn, E ~ ~ ~ , is quantitatively related to the micelle formation equilibrium of AQD+ itself and (2) the c.m.c. of another anionic surfactant, i.e. SDS, in water can be determined accurately in the presence of a very small amount of AQD+ (0.063 mmol dm-3). 26672668 Efect of Micelle Formation on Absorption Spectra Experimental SDS (Wako Pure Chemicals) was recrystallized from ethanol and washed with ether.6 The synthesis of (2-anthraquinonylmethyl)dodecyldimethylammonium bromide was reported previously. All the micellar solutions were prepared using doubly distilled water and by stirring vigorously for 1 day at 25 "C until the equilibria of dissolution and solubilization were reached.All the organic solvents used (Kanto Chemicals) were of 'spectrograde' or ultra-pure grade and used as supplied. The aqueous solutions of AQD+ homomicellar system contained 0.2 mmol dm-3 LiOH because (1) when the AQD+ micellar system was used in energy-storing photo reaction^,^ AQDO and AQD-, the reduced forms of AQD', were stable under alkaline conditions, and (2) LiOH served as a supporting electrolyte in electrochemistry of AQD+.' The micellar weight of the AQD+ micelle in 0.2 mol dmP3 LiOH aqueous solution was determined to be 14000' by quasielastic light scattering.' Absorption spectra were recorded on a Hitachi UV 220 spect rop ho t ome ter. Results and Discussion Concentration Dependence of Absorption Intensity (A,,, = 330 nm) of AQD+ Fig.1 shows a plot of cap, (the apparent molar absorption coefficient of the peak at 330 nm) of AQD+ us. the logarithm of the AQD+ concentration in aqueous 0.2 mol dm-3 LiOH. The value of E , ~ , is nearly constant at 5.1 x lo3 dm3 mmol-' cm-l below ca. 0.07 mmol dmP3, beyond which concentration it gradually decreases with increasing concentration of AQD'. This gradual change in cap, indicates the gradual increase in the concentration of AQD+ micelles. The change in E~~~ is drastic and well defined compared with that of API, AAB and 6-In- 1 1. The value of capD above ca. 0.07 mmol dm-3 is given (1) by assuming that (1) the pseudo-phase model" holds for the AQD+ homomicellar system, i.e. the concentration of the AQD+ monomer is kept constant at the value of the c.m.c.in the concentration range above the c.m.c., (2) the molar absorption coefficient of the AQD+ monomer, E,, is equal to that below ca. 0.07 mmol dm-3 and (3) the molar absorption coefficient of the AQD+ surfactant in the micellar form, E,, is constant. c is the total concentration of AQD+. Eqn (1) can be rewritten as CE,,, = c.m.c. E, + (c- c.m.c.) E, E,,, = (c", - E ~ ) c.m.c. c-' + E ~ ~ . (2) Fig. 2 shows a plot of E,,,, us. the inverse of the total concentration of AQD' above cu. 0.07 mmol dmW3, The c.m.c. and eM values of 0.072 mmol dm-3 and 4.0 x lo3 dm3 mol-' cm-' were determined from the slope and the intercept, respectively, of the plot in fig. 2 using the experimental value of E,,, 5.1 x lo3 dm3 mmol-' cm-'. To the best of our knowledge, this is the first report in which the U.V.spectra of a functionalized detergent has been related to the micelle-forming equilibrium in water. The reason for the change in eapp might be deduced by comparing the micro- environment of the anthraquinone (AQ) moiety of AQD+ in the monomeric form (below the c.m.c.) with that of AQD+ in the micellar form (above the c.m.c.). The AQ moiety of the AQD+ monomer is thought to be surrounded by water, while that of the AQD+ micelle might be situated in the micelle-water interface (Stern layer6). According to our electrochemical studies of AQD+ solubilized in hexadecyltrimethylammonium bromide (or cetyltrimethylammonium bromide, CTAB) micelles,' the microscopic polarity of the Stern layer of the CTAB micelle, where the AQ moiety of AQD+ might be situated, corresponds to ethanol or methanol.Assuming that the microscopic polarity of the AQ moiety of AQD+ in the AQD+ homomicelle is nearly equal to that in the CTAB micelle, the change in shown in fig. 1 might arise from the change in the microscopic polarity around the AQ moiety.K. Hoshino et al. 2669 5000 ,-I I E r( - I 0 E rn E -0 '0.4500 0 4000 I I 0 I I -5 -4 -3 log ([ AQD'I /mol d ~ n - ~ ) Fig. 1. Plot of the apparent molar absorption coefficient at 330 nm, capp, us. the logarithm of the concentration of AQD'. 0 2 4 6 8 1 0 c-'/dm3 mol-' Fig. 2. Plot of capp us. the inverse of the total concentration of AQD', c. The straight line is drawn by the least-squares method. In order to elucidate the relationship between the value of capp and the polarity of the environment around the AQ moiety, the value of E~~~ was measured in several organic solvents with different polarity.Contrary to our expectations, the value of capp was constant at (5.1 kO.1) x lo3 dm3 mol-l cm-l independently of the polarity of solvents, i.e. E~~~ depended little on the Kosower's 2 value'' for the following solvents: CH2C12 (2 == 64.2), CH3CN (2 = 71.3), C,H,OH (2 = 78.3), C,H,OH (2 = 79.6), CH,OH2670 Efect of MicelIe Formation on Absorption Spectra 0 0.2 0.4 0.6 08 1 X Fig. 3. Plot of E~~~ us. the mole fraction of AQD’ (x) in the AQD’XTAB mixed micellar solution at 25 “C. The concentration of AQD+ is constant at 0.06 mmol dm-3. (2 = 83.6), H 2 0 (2 = 94.6). These results indicate that the change in the capp is not due to the change in the polarity around the AQ moiety.According to the work of Matsuo and co-workers12 dealing with the U.V. spectra of 1-ethyl- l’-hexadecyl-4,4’-bipyridinium (Cl,C2V2+), the molar absorption coefficient at absorption tail (350nm) of c16c2v2+ was found to change as the concentration of c16c2v2+ exceeds the c.m.c. of C&2v2+. They thought that it might be attributable to the n-z interactions between the adjacent bipyridinium groups of C16C,V2+ on the C,,C, V2+ micelles. In order to see if such an explanation could apply to our AQD+ micellar system the shift in capp was followed while varying the mole fraction of AQD+ from 0 to 1 in mixed micelles with CTAB. Fig. 3 shows the plot of capp us. the mole fraction of AQD+, x, in an AQD+-CTAB mixed micellar solution containing 0.2 mol dm-3 LiOH.The concentration of AQD+ was held constant at 0.60 mmol dm-3, where AQD+ itself forms micelles. The value of E~~~ gradually increased with decreasing x, and E~~~ was nearly constant at x < 0.050. The constant value of eapp, 5.1 x lo3 dm3 mol-1 cm-l, was equal to the molar absorption coefficient of the AQD+ monomer, E,. These results suggest that CTAB resides between the AQD+ molecules in the mixed micelles and hinders the interactions between the AQ moieties of AQD+, resulting in an increase in capp. If we assume an ideal mixing13 of AQD+ and CTAB, the concentrations of AQD+, C,, and CTAB, C;, in the monomeric form in equilibrium with the mixed micelles are given by eqn (3) and (4) : C, = Y(c.m.c.)O (3) Ci = (1 - Y ) (c.m.c.’)O (4) with Y = C,/(C, + CM) ( 5 ) where (c.m.c.)O and (c.m.c.’)* are the critical micelle concentrations of AQD+ (0.072 mmol dm-3) and CTAB (0.094 mmol dm-3) in 0.2 mmol dm-3 LiOH aqueous solution, respectively, and C, and C& are the concentrations of AQD+ and CTAB in the mixedK .Hoshino et al. 267 1 t I I I I I I1 I11 Fig. 4. Plot of cmix us. (1 / Y - 1). micelles, respectively. The value of (c.m.c.’)O was determined by dye-solubilization method. Y shows the mole fraction of AQD+ in the mixed micelles. These equations are analogous to Rault’s equation in the form. From eqn (3)-(5), the total concentrations of AQD+, C,, and CTAB, C;, are C, = C, + C, = Y(c.m.c.)O + C, Ck = Cf’+ CL = (1 - Y) (c.m.c.’)O + CL. (6) (7) Introducing the values of C, and C,, calculated from eqn (5)-(7), into eqn (S), we obtain the value of the molar absorption coefficient of AQD+ in the mixed micelle, &mix = ( ~ a p p CT -Ern C,)/CM.(8) A knowledge of gives the number of molecules of CTAB, N, required for setting one molecule of AQD+ free of interactions with the other AQD+ molecules. If we assume that the molar absorption coefficient of one molecule of AQD+ changes from E , to E , when it is surrounded by N molecules of CTAB, the apparent absorbance of AQD+ in the mixed micelle, cmixCM, is given by &mix cM = cMf+EM(cM-cMf) (9) with C,, = CL/N (10) where C,, represents the concentration of AQD+ isolated in the mixed micelle. By introducing the values of C, and C,, from eqn (5) and (10) in eqn (9) and rearranging terms, one obtains the relationship &mix = (Em-&E,)(1/y-l)/N+EM. (1 1) Fig.4 shows a plot of cmix us. ( I / Y- I). The plot gave a straight line at high mole fractions of AQD+ (region I). N was determined to be 4.4 from the slope of the line. This indicates that one molecule of AQD+ in the mixed micelle may be set free from interactions with the other AQD+ molecules when it is surrounded by an average of 4.4 molecules of CTAB. In region I1 in fig. 4, the plot deviates from the curve expected for our model based on ideal mixing (dotted line). This may be due either to invalidity of our model [eqn ( 9 t ( 1 l)] or to deviation of the system from ideality [eqn (3)-(7)]. At very low mole fractions of AQD+ (region I11 in fig. 4), the values of cmix are almost constant and equal to the value of E,.This indicates that there may be no interactions among the AQD+ molecules under such circumstances.2672 Efect of Micelle Formation on Absorption Spectra From both the results of spectroscopic behaviour of AQD+ in organic solvents and the effect of addition of CTAB, it may be concluded that the interactions between the AQ moieties on the AQD+ micellar surface bring about the change in capp. Absorption Spectra of 0.060 mmol dmP3 AQD+ in Aqueous CTAB Solution The absorption spectra of the AQD+ monomer were investigated in varying concentration of cationic aqueous CTAB solution. The shape of the spectra of AQD+ showed no change over the CTAB concentration range from'0.05 to 50 mmol dm-3. This result is in contrast with the behaviour of AQD+ in the SDS solution described below.In addition, the molar absorption coefficient at Amax = 330 nm, capp, was constant at (5.1 0.1) x lo3 dm3 mol-1 cm-l over the whole CTAB concentration range studied. This E~~~ value is equal to the foregoing cm value, i.e. the molar absorption coefficient of the AQD+ monomer in water, indicating that the U.V. spectrum of the AQD+ monomer is not influenced by solubilization into the CTAB micelles. Absorption Spectra of AQD+ in Aqueous SDS Solution The absorption spectra of AQD' were investigated in varying concentration of SDS aqueous solution. The concentration of AQD+, [AQD+], was kept constant at 0.063 mmol dm-3. At this concentration AQD+ exists in the monomeric form in pure AQD+ solution, as described in the previous section (see fig.1). The shape of the absorption spectra of AQD+ depends on the concentration of SDS as shown in fig. 5. Sudden changes of the absorption intensities at 290 and ca. 250 nm were observed at a concentration of ca. 8 mmol dmd3 SDS. Fig. 6 shows plots of the molar absorption coefficients, E , of AQD+ at 290 nm (0) and ca. 250 nm (A) us. the concentration of SDS. Note that the absorption intensity at 290 nm changes by a factor of 7 above and below the c.m.c. of SDS. The spectral change mentioned above can be interpreted by the formation and break- up of an ion-pair complex of hydrophobic solutes of opposite charge, as reported by Roland and Smid.14 Below the c.m.c. of SDS, AQD+ is bound to DS- (monomeric dodecylsulphate anion) as a result of both electrostatic and hydrophobic attractions, so that the spectroactive species is DS--bound AQD+.The new absorption band at 290 nm may be attributed to this ion-pair complex. As the concentration of SDS is increased beyond the c.m.c., the DS--bound AQD+ complex is dispersed over the SDS micelles and this ion-pair complex breaks up, so that the spectroactive species is AQD+. This is supported by the experimental fact that the shape of the spectrum of monomeric AQD+ in the absence of SDS is almost the same as that of AQD+ completely solubilized in the SDS micelles. The solubility of DS--bound AQD+ depends on the ratio [AQD+]/[SDS]. When [SDS] is much higher or much lower than [AQD+] in the absence of SDS micelles, solutions were homogeneous for at least 10 days.When [SDS] is near to [AQD+], a white precipitate was observed. These experimental findings imply that a water-insoluble 1 : 1 complex between AQD" and DS- is formed when [SDS] is close to [AQD'], and that a water-soluble complex in which the ratio [AQD+]/[DS--] is not unity may be formed when [SDS] is much lower or higher than [AQD+]. The formation of such a non- stoichiometric ion-pair complex has already been reported. 15-17 We assume the following chemical equilibrium based on the pseudo-phase model'" s, + s, where S, represents the non-stoichiometric ion-pair complex described above and S, denotes AQD+ solubilized in the SDS micelle. The appearance of the isosbestic pointsK . Hoshino et al. 2673 20G 280 360 wavelength/nm Fig.5. U.V. spectral results upon addition of SDS to 0.063 mmol dm-3 AQD': 1, [SDS] = 6.87 mmol dmF3; 2, [SDS] = 7.51 mmol dm-3; 3, [SDS] = 7.91 mmol dm-3; 4, [SDS] = 8.55 mmol dm-3; 5, [SDS] = 13.7 mmol drnp3. Fig. 6. Plots of the molar absorption coefficients, E , of AQD' at 290 nm (0) and ca. 250 nm (A) us. the concentration of SDS, at 25 "C.2614 Eflect of Micelle Formation on Absorption Spectra 2 n E l cu n W 1 1 Y 3 v 0 [ SDS]/mol dmA3 Fig. 7. Plots of ( E ~ - E ) / ( E - E ~ ) measured at 290 nm (0) and ca. 250 nm (e) us. the concentration of SDS in (a) water and (b) aqueous 0.05 mol dm-3 solution. Table 1. The c.m.c. of SDS in the absence and presence of electrolyte c.m.c. /mmol dm-3 ca. 290 250 lit. c.m.c. solvent nm nm /mmoldm-3 water 7.3 7.4 8.1“ 0.05 mmol dm-3 - 2.2 2.25b NaCl a Ref.(24). Ref. (25). at 202, 210, 230 and 274 nm in fig. 5 supports the existence of the simple partition equilibrium of eqn (12). The partition coefficient of eqn (12), K, is given by = ( n M / ‘M)/(nW/ ‘W) = (‘W/ ‘M) (‘M/ v)/(nW/ ‘) = (‘W/ ‘MI [sMl/[sWl (l 3, where n, and nw are the numbers of moles of S, and S,, respectively, V, and Vw are the volumes of the micellar phase and the water phase, respectively, and V is the total volume of the micellar solution. If both S, and S, obey Beer’s law at a given wavelength, the total absorbance per cm of path, A , is given by with A = E [ S ] = Ew[Sw] +E,[S,] [SI = P,l+ [ S M l (14) (15) where E , and E , are the molar absorption coefficients of S, and S,, respectively, E is the apparent molar absorption coefficient and [S] is the total concentration of AQD+.TheK. Hoshino et al. 2675 volume of the micellar phase, VM, is proportional to ([SDS] -c.m.c.), which is equal to the concentration of SDS taking part in micelle formation :Is (16) (17) VM = cc([SDS] - c.m.c.) ( E ~ - E ) / ( E - E ~ ) = aK([SDS] - c.m.c.)/ V,. where a is the proportionality constant. From eqn (13)-( 16), Plotting (E, - E ) / ( E - E ~ ) against [SDS] close to the c.m.c., we obtain a straight line running through the point at [SDS] = c.m.c. on the abscissa. In fig. 7 ( a ) we have plotted ( E , - E ) / ( E - E ~ ) against [SDS] at 290 nm (0) and ca. 250 nm (a) using the values of E, and E, obtained from fig. 6: E, = 1.40 x lo4 dm3 mol-' cm-l and E, = 2.00 x lo3 dm3 mol-' cm-l at 290 nm; E, = 1.66 x lo4 dm3 mol-' cm-' and E, = 4.24 x lo4 dm3 mol-l cm-' at ca.250 nm. The c.m.c. values obtained by least-squares analysis are listed in table 1, which also gives the c.m.c. in the presence of 0.05 mol dm-2 NaCl, determined from the (E, - E ) / ( E - cM) us. [SDS] plot [fig. 7 (b)], since the SDS aqueous solution is widely used as a medium in electr~chemistry.'~-~' Table 1 lacks the data measured at 290 nm in 0.05 mol dm-3 NaC1. This is because the molar absorption coefficient of DS--bound AQD+ at 290 nm ( E ~ ) was not constant. This may be due to the interaction between the AQD+-DS- and Na+ or C1-, but it cannot be interpreted in detail at the moment. It can be seen from table 1 that the c.m.c. values are in good agreement with the literature values.Evans and Bolton22 measured the c.m.c. of SDS spectrophotometrically using the interaction between 5-methylphenazinium cation and the SDS micelle; however, the determined c.m.c. value (5.6 mmol dm-3) was much lower than the literature value (8.1 mmol dm-3). When compared with 5-methyl- phenazinium cation, the cationic surfactant AQD+ may be a better probe for determining the c.m.c. of SDS in the absence and presence of electrolyte because (1) the absorption intensity at 290 nm changes drastically at the c.m.c., (2) the concentration of AQD+ used is so small that the host micellar system may be only slightly perturbed, (3) the host-like structure of AQD' is expected to perturb the structure of host micelles to a minimal degree and (4) the presence of the large cation and the large ion-pair complex may lower the c.m.c.of the host surfactant;23 however, the c.m.c. of SDS was little influenced by AQD+ and AQD+-DS-. Conclusions The c.m.c. of (anthraquinonylmethyl)dodecyldimethylammonium bromide, AQD+, in aqueous 0.2 mol dm-3 LiOH solution was determined by studying the variation of the molar absorption coefficient of the lowest absorption band at 330 nm, E ~ ~ ~ , of AQD+. It has been found that the change in E~~~ with the concentration of AQD+ could be related to a micelle-formation equilibrium and that it resulted from the interactions among anthraquinone moieties. The latter was confirmed by studying the variation of E~~~ in the mixed micelles with CTAB. The function of AQD+ as a probe for determining the c.m.c.s of normal anionic surfactants was also investigated.The c.m.c. of SDS could be determined using AQD+ as a probe. This function may arise from the formation of the ion-pair complex, AQD+-DS-, below the c.m.c. of SDS and the disruption of the complex by its solubilization into the SDS micelles above the c.m.c. of SDS. References 1 W. D. iiarkins, H. Krizek and M. L. Corrin, J . Colloid Sci., 1951, 6, 576. 2 A. Kitahara, Bull. Chem. SOC. Jpn, 1957, 30, 586. 3 N. E. Schore and N. J. Turro, J . Am. Chem. SOC., 1975, 97, 2488. 4 N. J. Turro, Y. Tanimoto and G. Gabor, Photochem. Photobiol., 1980, 31, 527. 88 FAR 12616 Efect of Micelle Formation on Absorption Spectra 5 M. Krieg, M. P. Pileni, A. M. Braun and M. Gratzel, J. Colloid interface Sci., 1981, 81, 209.6 J. H. Fendler and E. J . Fendler, Catalysis in Micellar and Macromolecular Systems (Academic Press, 7 M. Fujihira, H. Yoneyama and S. Aoyagui, to be published. 8 M. Fujihira, K. Hoshino, T. Saji and S . Aoyagui, Chem. Lett., 1985, 1419. 9 K. Shinoda, T. Nagakawa, B. Tamamushi and T. Isemura, Colloidal Surfactants (Academic Press, New York, 1975). New York, 1963). 10 K. Shinoda and E. Hutchison, J. Phys. Chem., 1962, 66, 577. 11 E. M. Kosower, J . Am. Chem. Soc., 1958, 80, 3253. 12 K. Takuma, T. Sakamoto, T. Nagamura and T. Matsuo, J . Phys. Chem., 1981, 85, 619. 13 For example: K. J. Mysels and R. J . Otter, J . Colloid Sci., 1961, 16, 474; F. Tokiwa, K. Ohki and 14 B. Roland and J . Smid, J. Am. Chem. SOC., 1983 105, 5269. 15 S. S. Atik and A. Singer, J. Am. Chem. Sac., 1979, 101, 6759. 16 G. A. Davis, J. Chem. Soc., Chem. Commun., 1973, 728. 17 Y. Miyashita and S . Hayano, Bull. Chem. Soc. Jpn, 1981, 54, 3249. 18 J. C . Russell and D. G. Whitten, J. Am. Chem. Soc., 1982, 104, 5937. 19 N. Shinozuka and S . Hayano, Solution Chemistry of Surfactants, ed. K. L. Mittal (Plenum Press, New 20 Y. Ohsawa, Y. Shimazaki and S. Aoyagui, J. Electroanal. Chem., 1980, 114, 235. 21 A. E. Kaifer and A. J . Bard, J. Phys. Chem., 1985, 89, 4876. 22 A. Evans and J . R. Bolton, Photochem. Photohiol., 1979, 30, 697. 23 G. 1. Mukhayer and S . S. Davis, J. Colloid interface Sci., 1975, 53, 224. 24 R. J. Williams, J. N. Phillips and K. J . Mysels, Trans. Faraday SOC., 1955, 51, 728. 25 D. Britz and J. Mortensen, J. Electroanal. Chem., 1981, 127, 231. 1. Kokubo, Bull. Chem. SOC. Jpn, 1968, 41, 2845. York, 1979), vol. 2, pp. 599-623. Paper 71062; Received 13th January, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402667

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1988, 84(8), 2677-2682 Synthesis of a Viologen-Tetratitanate Intercalation Compound and its Photochemical Behaviour Hirokatsu Miyata, Yoshiyuki Sugahara, Kazuyuki Kuroda and Chuzo Kato* Department of Applied Chemistry, Waseda University , 0 h kubo- 3, Sh injuku- ku, Tokyo 160, Japan Methyl viologen has been intercalated into the interlayer space of layered tetratitanic acid by a method involving the displacement of guest molecules using an n-propylammonium tetratitanate intercalation compound. The viologen tetratitanate intercalation compound changed colour to blue when irradiated by a mercury lamp under vacuum conditions or a nitrogen atmosphere. The formation of radical cations was confirmed by visible and e.s.r. spectroscopy. The electron donor for the photoreduction is thought to originate from the tetratitanate layers.The colour was stable as long as the vacuum conditions were maintained, and fading did not occur rapidly even after the introduction of air, requiring > 1 h for the colour to fade completely. These results indicate remarkable stability of the viologen radical cations in this intercalation compound. Viologens (1,l '-disubstituted-4,4'-bipyridinium salts) are photoreduced reversibly in the presence of an electron donor, forming blue radical cations. Several investigations of the photoreduction of viologens have been conducted in relation to the reaction mechanism or their interaction with various matrices, although these are mostly restricted to organic materials. lP5 If various inorganic layered materials could be used as host materials for the inclusion of viologens, it would be expected that physical or chemical properties could be greatly changed by the intercalation.Therefore, such intercalation compounds have attracted much attention from the viewpoint of developing new materials. Potassium tetratitanate (K,Ti,O,) is a layered transition-metal oxide whose structure has been Ti,O, units which have four connected TiO, octahedra are stacked in the a direction, and potassium ions are present in the interlayer space. The potassium ions can be removed by acid treatment and the tetratitanic acid formed (ideal formula ; H,Ti,O, - H,O) retains a layer structure. Since tetratitanic acid can also be recognized as a layered form of TiO,, one might expect that the acid would show similar properties to TiO,.TiO, is an n-type semiconductor having a bandgap of 3.0 eV, and the electrons in the valence band are excited by irradiation by u.v.-visible light of wavelength < 41 5 nm. Charge transfer from TiO, to methyl viologen has been reported recently.'? Therefore, we believe that a study of the interaction of layered titanic acids and viologens should provide interesting information on basic intercalation chemistry, as well as potential applications such as optical memories. However, there have been no studies on viologen-layered-titanate intercalation compounds. As for the intercalation of organic cations into the interlayer space of layered tetratitanate, no reports have been published except for the intercalation of alklyamines and alkylammonium ions.'O* l1 The organic materials which can be directly intercalated into the interlamellar region of layered transition-metal oxides such as tetratitanate are rather restricted.l2 In our preliminary experiments the direct intercalation of methyl viologen into tetratitanic acid or potassium tetratitanate was unsuccessful. Thus the so- 2677 88-22678 Synthesis of a Viologen-Tetratitanate Intercalation Compound called 'guest displacement method ' was employed, in which a preformed intercalation compound was prepared in advance using n-alkylamines and then the guest exchange reaction was performed. The process is promising for the synthesis of intercalation compounds which cannot be obtained directly. The intercalation of viologens into inorganic layered materials has been investigated using the following host materials : clay minerals (montmorillonite and vermic~lite),l~-'~ MPS,l8 and uranyl ph~sphate.'~ However, these studies did not treat with the host-guest interactions. The stability of radical cations reduced chemically was only noted in MPS, crystals, except for our recent investigation on the photochromism of viologen-montmorilloni te-poly(viny1pyrrolidone) intercalation compounds.2o Therefore, the intercalation of methyl viologen into the interlayer region of tetratitanic acid was attempted. The photochemical behaviour of the intercalation compound was also investigated because charge transfer from the tetratitanate host to the intercalated methyl viologen by light irradiation was thought to be probable.Experiment a1 Materials Potassium tetratitanate prepared by a flux method21 was obtained from the Ohtuka Chemical Co. Crystalline impurities could not be detected by X-ray powder diffraction. Methyl viologen and n-propylamine (Tokyo Kasei Co.,) were used as received. Synthesis of Intercalation Compounds Potassium tetratitanate was treated with 1 mol dm-, HC1 for 24 h three times at room temperature (100 cm3 HCl per gram titanate). The tetratitanic acid formed was washed with water until free of C1- and was then air-dried. The amount of remaining potassium ions in the interlayer space was measured by flame photometry. The tetratitanic acid formed was sealed in an ampoule containing 50 O h n-propylamine aqueous solution (4 cm3 per gram titanic acid) and was allowed to stand for one week at 60 "C.The product was washed with acetone thoroughly and dried in air. The n-propylammonium tetratitanate was again sealed with 1 mol dm-3 methyl viologen aqueous solution and allowed to stand for two weeks at 60 "C. The amount of viologen was ten times the value calculated from the ion-exchange capacity of the tetratitanate. The resultant product was washed with methanol until the viologen could not be detected in the washings by U.V. spectrometry. Characterization The products were characterized by the use of X-ray powder diffraction, i.r. spectroscopy and elemental analysis. X-Ray diffraction was conducted using a Rigaku RADI-B diffractometer (Cu Ka radiation, Ni filter). 1.r spectra were recorded on a Shimadzu IR- 400 instrument using KBr disc.The viologen-tetratitanate intercalation compound dispersed in methanol was coated onto a glass plate and was allowed to react with n- butyl lithium. The visible spectral changes were also recorded on a Hitachi u-3200 spectrometer by a reflection method using an integrating sphere. Photochemical Behaviour The viologen-tetratitanate intercalation compound dispersed in methanol was coated onto the inside of a conventional quartz cell and the cell was purged with nitrogen gas. The visible spectra of the product were also recorded on the same spectrometer before and after irradiation by a 100 W mercury lamp (Rikoh-Kagaku-Sangyo).H. Miyata, Y. Sugahara, K. Kuroda and C. Kato 2679 Irradiation was performed without using filters and the distance from the light source was ca.10 cm, the irradiation time being ca. 10 min. E.s.r. spectral changes before and after irradiation of the sample under a vacuum < 0.1 Pa were recorded on a JEOL JES-FE2XG spectrometer at room temperature. In addition, a reference sample in which methyl viologen was absorbed on the external surface of tetratitanic acid was also prepared and measured in a similar manner to provide a comparison with the intercalation compound. Results and Discussion Synthesis of the Viologen-Tetratitanate Intercalation Compound n-Propylammonium-Tetratitanate Intercalation Compound Since alkylammonium tetratitanate intercalation compounds could not be prepared by direct intercalation," layered tetratitanic acid was prepared by acid treatment of potassium tetratitanate. Ca.91 O h of the potassium ions were removed by this treatment. In previous papers, thc dzo0 value corresponding to the average interlayer spacing was reported to be ca. 9.0 A in tetratitanic acid after the complete rFmoval of the interlayer potassium ions.22*23 However, in the present study, dzo0 = 8.6 A. The X-ray powder diffraction pattern of the acid-treated product chapged drastically by the intercalation of n-propylamines.The dzo0 value increased to 15.2 A. According to previous investigations on alkylammonium-tetratitanate compounds, th: dzo0 value of n-propylammonium-tetratitanate was reported to be ca. 16.2 AIO or 15.6 A." Therefore, the interlayer spacings of both tetratitanic acid and the n-propylammonium intercalation compound synthesized in the present study were smaller than those reported previously.The i.r. spectrum of the n-propylammonium-tetratitanate intercalation compound which had been washed thoroughly still exhibited absorption bands due to n- propylammonium ions : 1570 cm-l, N-H bend ; 1460 cm-l, C-H bend ; 1 180 cm-', C-N stretch. From these X-ray and i.r. results, the formation of the n-propyl- ammonium-tetratitanate intercalation compound was confirmed. Methyl Viologen-Tetratitanate Intercalation Compound The X-ray powder diffraction pattern of the product obtained from the reaction of the n-propy!ammonium<tetratitanate with methyl viologen showed a decreased dzoO value of 12.2 A from 15.2 A of the n-propylammFnium-tetratitanate compound.The value is larger than that of tetratitanic acid by 3.6 A, which is similar to the increase in the basal spacing in othe; viologen-inorganic layered host intercclation compounds [mont- morillonite, 3.1 A;'' MPS,, (M = Mn, Cd and Fe), 3.3 A18]. In these intercalation compounds, the intercalated methyl viologen lies flat in the interlayer space. The i.r. spectrum of the product also showed the presence of characteristic bands assigned to methyl viologen, e.g. 1640 and 1560 cm-l ring vibrations of the pyridine ring. On the other hand, the absorption bands due to n-propylammonium ions almost disappeared in the product. Therefore, the variation of the X-ray patterns and i.r. results indicated that n-propylammonium ions in the interlayer region were replaced by methyl viologens.The organic contents of the n-propylammonium-tetratitanate viologen-tetratitanate intercalation compounds are shown in table 1. The ratios of carbon to nitrogen in the intercalation compounds were in good agreement with those calculated for the corresponding guest cations, which also indicated the absence of n-propylammonium ions after the intercalation of the viologen. The amounts of organic component per Ti,O, unit are also shown in table 1, calculated on the assumption that the composition of the inorganic component was 0.09K20+4Ti02. Thus ca. one third of the total2680 Synthesis of a Viologen-Tetratitanate Intercalation Compound Table 1. C and N contents of the intercalation compounds organic cation sample Ti409 C (YO) N (YO) C:N ratio (molar ratio) n-PrNH ,-te trati tana te MV-tetratitanate 6.05 2.16 2.80 (2.57)" 0.67 10.06 I .98 5.08 (5.15)" 0.29 a The values in parentheses are theoretical values of the guest cations.exchange capacity was occupied by the organic cations. The fact that the amount of viologen was slightly smaller than half of the amount of n-propylammonium ions indicated that the latter were partly exchanged with protons (or oxonium ions) when methyl viologen was intercalated. From these results it can be concluded that methyl viologen-tetratitanate intercalation compound was synthesized. Reduction of the Methyl Viologen-Tetratitanate Intercalation Compound by n-Butyl- lithium The viologen-tetratitanate intercalation compound exhibited a blue colour when it was allowed to react with n-butyl-lithium in n-hexane.Fig. 1 (a) shows the visible spectrum of the reduced product. The absorption spectrum, having maxima at ca. 630 and 405 nm, is characteristic of viologen radical cations. Although the reaction of tetra- titanic acid with n-butyl-lithium yielded a dark blue product, the visible spectrum was totally different from that of the coloured intercalation compound. The different colour may be ascribed to the formation of Ti3+. Consequently, fig. l(a) indicated that the intercalated methyl-viologen was reduced by n-butyl-lithium. The charge balance between the titanate layers and the guest cations was probably compensated by the additional insertion of lithium ions. These results showed that the structure of the viologen did not change after intercalation. The blue colour was comparatively stable and did not fade, even after the product was exposed to air for several hours.A similar experiment was carried out on a sample of tetratitanic acid on whose external surface methyl viologen was absorbed. The molar ratio of the viologen per unit of Ti,O, determined by elemental analysis was 0.40, which was comparable to that of the intercalation compound. The sample also showed a blue colour after reaction with n- butyl-lithium. However, the colour was unstable and disappeared within several seconds when it was removed from the hexane solution. Therefore, the stability of the radical cations in the intercalation compound is thought to result from a shielding between the radical cations formed and an oxidizing agent, which is probably oxygen in the present experiment.This is also clear evidence of the presence of the viologen in the interlayer space. Photochemical Behaviour When the viologen-tetratitanate intercalation compound was irradiated by a mercury lamp under an N, atmosphere, the colour changed to blue. The visible spectral change is shown in fig. l(6) and ( c ) . The spectrum after illumination showed a similar pattern to that of the chemically reduced product. Thus it was clarified that the radical cations were formed by the irradiation of this intercalation compound. The electron donors were thought to be tetratitanate layers, and electron transfer from the host to the guest seems to occur by light irradiation. In the tetratitanic acid sample on whose external surfaceH.Miyata, Y. Sugahara, K. Kuroda and C. Kato 268 1 1 .a Q) 9 e 8 % I 1 I I wavelength/nm Fig. 1. Visible spectra of the methyl viologen-tetratitanate intercalation compound (a) after the reaction with n-butyl-lithium, (b) after irradiation and (c) before irradiation. methyl viologen was adsorbed, colour development after illumination was also detected, although the intensity was much weaker than that of the intercalation compound. When the intercalation compound was placed in a vacuum of < 0.1 Pa, its colour changed to blue even before irradiation. The e.s.r. spectrum of the sample showed a definite signal at g = 2.003, indicating the formation of viologen radical cations. When the sample was illuminated in a vacuum, the colour development became much deeper, and accordingly the intensity of the e.s.r.signal became stronger, showing that the amount of radicals being formed was more than ten times that of the sample before irradiation. The e.s.r. spectrum of the tetratitanic acid with adsorbed viologen was also measured under the same conditions. In this case the formation of radical cations could not be detected under vacuum conditions alone. When the sample further irradiated by a mercury lamp, it showed a slight blue colour and the corresponding e.s.r. signal appeared. However, the signal was very weak in comparison with that of the intercalation compound. The colour developed under vacuum conditions was stable as long as the conditions were maintained after irradiation had stopped.Therefore, electron transfer in the opposite direction, from the viologen radical cations to the tetratitanate layers, did not occur spontaneously. The reasons for the formation of radical cations under vacuum conditions without illumination and the inhibition of the reverse reaction were not clarified. Further studies, such as determining the precise configuration of the viologens in the interlayer space before and after irradiation and the electronic structure of the layered tetratitanic acid would provide useful information for an explanation. When the coloured sample which had been irradiated under vacuum conditions was exposed to air, the colour faded gradually. The rate of fading was slow, requiring > 1 h to be complete. This result indicates that electron transfer in the reverse direction, from the methyl viologen radical cations to the tetratitanate, could be caused by the introduction of oxygen.The low rate of the fading process and the unexpected stability2682 Synthesis of a Viologen- Te trat itanate Intercalation Compound of the radical cations formed can be explained by interference from an oxidizing agent (presumably oxygen) between the contact of the viologen in the interlayer space and the inorganic host layer. When the tetratitanic acid with externally adsorbed methyl viologen was exposed to air after light irradiation in uacuo, the colour instantly disappeared. Conclusions A methyl viologen-tetratitanate intercalation compound has been synthesized by the guest-exchange method of the n-propylammonium-tetratitanate intercalation com- pound with methyl viologen di-cations.The colour of the intercalation compound formed changed to blue when it was irradiated by a mercury lamp in the absence of oxygen. The source of the electron donor for the reduction process was thought to be tetratitanate host layers. The reverse electron transfer from the methyl viologen radical cations to the tetratitanate layers did not occur spontaneously without the mediation of an oxidizing agent. We thank Mr M. Suzue (Ohtsuka Chemical Co.) for supplying the potassium tetratitanate. The present work was partially supported by a Grant-in-Aid for Special Project Research in Molecular Assemblies (no. 582 18 033,60 1 040 04) from the Japanese Ministry of Education, Science and Culture. K.K.thanks the Nippon Sheet Glass Foundation for Materials Science for financial support. Financial support from a Waseda University Grant for Special Research Project is also acknowledged. References 1 M. Kaneko, H. Araki and A. Yamada, Sci. Pap. Inst. Phys. Chem. Res., 1979, 73, 67. 2 H. Kamogawa, T. Masui and M. Nanasawa, Chem. Lett., 1980, 1145. 3 H. Kamogawa, T. Masui and S. Amemiya, J. Polym. Sci., Polym. Chem. Ed., 1984, 22, 383. 4 M. Kaneko, Y. Imamura, K. Hayashi and A. Yamada, Kobunshi Ronbunshu, 1982, 39, 665; Chem. 5 M. Kaneko and A. Yamada, Makromol. Chem., 198 1, 182, 1 1 1 1. 6 A. Verbaere and M. Tournoux, Bull. Soc. Chim. Fr., 1973, 1237. 7 M. Dion, Y. Piffard and M. Tournoux, J. Inorg. Nucl. Chem., 1978, 40, 917. 8 D. Duonghong, J. Ramsden and M. Gratzel, J. Am. Chem. SOC., 1982, 104, 2977. 9 K. Chandrasekaran and J. K. Thomas, J . Chem. Soc., Faraday Trans. I , 1984, 80, 1163. Abstr. 98: 63201k. 10 H. Izawa, S. Kikkawa and M. Koizumi, Polyhedron, 1983, 2, 741. 11 P. Clement and R. Marchand, C.R. Acad. Sci. Paris, Ser. II, 1983, 296, 1161. 12 S. Kikkawa and M. Koizumi, Mater. Res. Bull., 1980, 15, 533. 13 0. D. Philen Jr, S. B. Weed and J. B. Weber, Clays Clay Miner., 1971, 19, 295. 14 Y. Soma and M. Soma, Kokuritsu Kogai Kenkyusho Kenkyuu Hokoku, 1982,36,227; Chem. Abstr.: 97, 15 M. Raupach, W. W. Emerson and P. G. Slade, J. Colloid Interface Sci., 1979, 69, 398. 16 M. H. B. Hayes, M. E. Pick and B. A. Toms, J. Colloid Interface Sci., 1978, 65, 254. 17 M. H. B. Hayes, M. E. Pick and B. A. Toms, J. Colloid Interface Sci., 1978, 65, 266. 18 0. Poizat, C. Sourisseau and Y. Mathey, J . Chem. Soc., Faraday Trans. 1, 1984, 80, 3257. 19 Ph. Colomban and M. P. Thi, Rev. Chim. Miner., 1985, 22, 143. 20 H. Miyata, Y. Sugahara, K. Kuroda and C. Kato, J. Chem. Soc., Faraday Trans. 1, 1987, 83, 1851. 21 N. Ohta and Y. Fujiki, Yogyo-Kyokai-Shi, 1980, 88, 1. 22 H. Izawa, S. Kikkawa and M. Koizumi, J. Phys. Chem., 1982, 86, 5023. 23 T. Sasaki, M. Watanabe, Y. Komatsu and Y. Fujiki, Znorg. Chem., 1985, 24, 2265. 168632j. Paper 7/1201; Received 6th July, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402677

出版商:RSC    

年代:1988

数据来源: RSC