摘要:

J. Chem. SOC., Faraday Trans. 1, 1984, 80, 1985-1990 Active Phases in Chromia-Alumina Dehydrogenation Catalysts Electron Spin Resonance and Kinetic Studies BY FATHY M. ASHMAWY Department of Chemistry, Faculty of Science, University of Tanta, Tanta, Egypt AND CHARLES A. MCAULIFFE* Department of Chemistry, University of Manchester Institute of Science and Technology, Manchester M60 1QD Received 14th October, 1983 The rate of propane dehydrogenation on chromia-alumina has been measured in a flow system over the temperature range 800-891 K. Results show that the reaction rate dependence was of a fractional order (0.66) with respect to propane partial pressure, and an apparent activation energy of 138.8 kJ mol-' was obtained. The effect of various pretreatments has been found to influence strongly both the e.s.r.spectra of the chromia phases and the activation energy of the reaction. Previous investigations on chromia-alumina have shown that Cr3+ ions are probably the centres for hydrogenation on these ~atalysts.l-~ However, the role of Cr2+ is still open to ~peculation.~-l~ Recently we have investigated the activity of a series of catalysts containing known amounts of Cr2+ and Cr3+ on alumina where it has been found that Cr2+ plays an active role in the dehydrogenation of ethane.ll In order to investigate this point further it was decided to extend our investigations to practical catalysts of higher chromia contents operating close to industrial conditions. Propane dehydrogenation to propene, a reaction which might become an important industrial process, was chosen because few studies of it have been published.12-15 We report the influence of various pretreatments on the e.s.r.spectra of the chromia phases on alumina and the effect of these pretreatments on the kinetics of the dehydrogenation. EXPERIMENTAL The procedures adopted for the purifications of materials were essentially those described previously.1° In the present work two catalysts were employed. The first was a commercial chromia-alumina catalyst (10% Cr203) supplied by Laporte and which had a surface area of 253 m2 g-l. A second catalyst of a higher chromia content (1 6.8 % Cr203) was prepared by impregnation of alumina with a concentrated solution of chromium nitrate and by subsequent drying at 393 K for 4 h. The potassium-promoted catalyst was obtained by impregnation of the alumina support with potassium nitrate followed by drying at 393 K for 4 h.This catalyst has the following specifications: particle size 1.5 mm, pore volume 0.37 cm3 g-l and surface area of 275 m2 g-l. Propane dehydrogenation was studied in a continuous-flow system at 1 atmt according 'f 1 atm = 101 325 Pa. 65 1985 F A R 11986 DEHYDROGENATION ON CHROMIA-ALUMINA CATALYSTS I I I \ I I 1500 2000 2500 \3000 3500 (4 Fig. 1. Electron spin resonance spectrum of chromia-alumina: (A) 10.0% Cr203 and (B) 16.8 O 0 Cr,0,-2% K,O. (a) Catalyst reduced with 'wet' hydrogen at 823 K (only Cr3+ produced) and (b) catalyst reduced with 'dry' hydrogen at 823 K (Cr2+ produced). to a procedure which has been described previously.14 Analysis of the reactor outlet was carried out by g.1.c.with a 10 m silica gel column (30-60 mesh) at 373 K. Helium was used as carrier gas at a flow rate of 60 cm3 min-I. The reactor contained 8 g of catalyst. which were reduced in situ. The absence of diffusional limitations was verified by measuring reaction rates at different catalyst particle sizes and at fairly high space velocities of between 450 and 850 h-' (volume of feed, at s.t.p.. per bulk volume of catalyst per hour). Reaction rates were determined at partial pressures of propane between 0.15 and 1 .O atm using nitrogen as a diluent. The reaction temperature was varied between 800 and 891 K (k 1 K). In general the conversion of propane to propene varied between 2.5 and 62.59;. The maximum amounts of cracking products such as methane, ethane and ethene were 2.8, 1.6 and 0.8",.respectively. The e.s.r. measurements were carried out on a Decca XI spectrometer with 100 kHz field modulation and phase-sensitive detection. All the e.s.r. measurements were carried out at room temperature. The samples to be investigated were first prepared in a silica reaction vessel attached to a conventional glass system which could be evacuated to Torr. THERMODYNAMICS OF CHROMIA REDUCTION The thermodynamics of the reaction : Cr203 + H, + 2Cr0 + H,O (1) show that at 800 K AG* = 66.94 kJ.16*" This value corresponds to an equilibrium constant K = pHpO/pHe = 2 x Therefore Cr,03 may be reduced to CrO if the water partial pressure is kept below 0.017 Torr.*lH In a previous investigationll we presented evidence supporting the presence of Cr2+ by carrying out the reverse of reaction (1) at high temperatures when hydrogen was evolved.Therefore, in the present work reduction treatments were carried out. as previously described," at 823 K and at 1 atm pressure for 8 h followed by outgassing at 823 K for I h: (a) reduction with 'dry' hydrogen (Cr2+ produced) and (h) reduction with 'wet' hydrogen (only Cr3+ produced), (this was achieved by allowing hydrogen to flow through a water trap maintained at room temperature). Saturation of the hydrogen with water vapour was sufficient to reoxidise Cr2+ to Cr3+ according to reaction (1 ). * I Torr = 101 335/760 Pa.F. M. ASHMAWY AND C. A. MCAULIFFE 1987 RESULTS AND DISCUSSION E.S.R. RESULTS BEHAVIOUR OF THE CATALYSTS AS A FUNCTION OF REDUCTION TREATMENTS The commercial and impregnated catalysts were investigated. Each catalyst was subjected to two different reduction treatments as previously described.The e.s.r. spectra of both catalyst systems are given in fig. l(A) and (B). The following features may be noted. The commercial catalyst, fig. 1 (A, a), shows a broad signal with a maximum at 2500 G, which has its origin in strongly interacting Cr3+ sites ( g = 1.986). This resonance signal is thought to be due to clumped Cr3+ ions in the surface layer and it is referred to as bulk or P-phase re~0nance.l~ The e.s.r. of this sample is also characterized by the presence of a relatively small signal with a maximum at 1700 G and it is referred to as &phase characteristic of Cr3+ ions which are rather is01ated.l~ Fig.1 (A, 6) shows a reduction in the e.s.r. spectrum upon reduction with ‘dry’ hydrogen, which is believed to be due to the formation of Cr2+. Fig. 1 (B, a) shows the e.s.r. spectrum of the impregnated catalyst (16.8% Cr203). The e.s.r. signal is symmetrica120 and shows only the P-phase with a maximum at 2500 G while the &phase disappeared completely. This indicates that the Cr3+ ions are present only as bulk chromia in the P-phase. The P-phase resonance signal was further reduced with ‘dry’ hydrogen, fig. 1(B, 6). Note that the transition between the two spectra is reversible by alternate treatments in ‘wet’ and ‘dry’ hydrogen. This demonstrates the sensitivity of chromium ions to water vapour according to reaction (1).KINETICS OF PROPANE DEHYDROGENATION DEPENDENCE OF REACTION RATE ON PROPANE PARTIAL PRESSURE The commercial catalyst shows appreciable cracking activity and its use in the kinetic measurements was, therefore, limited. In order to avoid poisoning by side reaction the potassium-promoted chromia-alumina (1 6.8 % Cr203-2% K,O) was used, which shows oxidation and reduction behaviour comparable to the chromia- alumina used for e.s.r. measurements. Rates of propane dehydrogenation were calculated from the steady-state balance ea uation : 21 xF rate = ~ A where x is the fraction of propane reacted, F is the feed rate (mol propane h-l) and A is the surface area (m2gs-l). Preliminary runs showed that the activity of the impregnated catalyst was essentially constant over a period of 3 h before significant cracking products are observed.The rate measurements were always started 30 min after the propane was introduced into the catalyst. The effect of the partial pressure of propane on the reaction rates was studied over the temperature range 868-883 K. The results are depicted in fig. 2 and suggest fractional-order kinetics. The dependence of the rate of reaction on the partial pressure of propane can be expressed in the form of the kinetic law: rate = kpEJH8 (3) where k is the rate constant, pc:,jH, is the partial pressure of propane and n is the order of reaction. Typical plots of log(rate) against are shown in fig. 3. At all 65-21988 Fig. 2. Rate of DEHYDROGENATION ON CHROMIA-ALUMINA CATALYSTS 3.6 3.2 2.8 I E i 2.4 2.0 s - Y 5? S 1.6 E 1 1.2 0.8 0.4 , Id1 0 propane partial pressure/atm propane dehydrogenation as a function of propane partial pressure temperatures: (a) 868, (b) 873, (c) 878 and (d) 883 K.at different 0.7 I 1 1 1 1 1 1 0 0.2 0.4 0.6 0.8 1 .O 1.2 log,, (P) + 1 Fig. 3. Log of the rate of propane dehydrogenation at different temperatures against log of partial pressure of propane: (a) 868, (b) 873, (c) 878 and (d) 883 K.F. M. ASHMAWY AND C. A. McAULIFFE 1989 1.6 1.5 1.4 1 . 3 2 1.2 E 2 $ 1 . 1 1 .o 0.9 0 . 0 v, + - v - 0.7 0.6 0 . E I I I I I I I I 11.2 11,4 11.6 11.8 12.0 12.2 12.4 12.6 104 K I T Fig. 4. Arrhenius plots: propane dehydrogenation over 16.8% Cr,0,-2% K,O (solid lines) and on commercial catalyst (broken line); 0 and A, catalyst reduced with ‘dry’ hydrogen; 0 and 0, catalyst reduced with ‘wet’ hydrogen. temperatures these plots were linear and parallel and the value of the kinetic order, n = 0.66, could be obtained from the slopes.DEPENDENCE OF REACTION RATE ON CATALYST PRETREATMENT The difference between ‘dry’ and ‘wet’ hydrogen treatment is also brought out by the kinetic behaviour. This difference is also illustrated in fig. 4, which reports the catalytic activity for propane dehydrogenation as an Arrhenius plot for the samples studied. The apparent activation energy of the catalyst treated with ‘ wet’ hydrogen (only Cr3+ produced) was found to be 138.8 kJ mol-l. However, this value was significantly reduced to 103.7 kJ mol-l as a result of formation of Cr2+ on the surface by ‘dry’ hydrogen. The former value, obtained in the temperature range 801-891 K, is in good agreement with the value of 141.7 kJ mol-1 which was obtained by Zuzuki et a1.;15 however, their investigations were conducted in the temperature range 7 16-778 K, in a static system and on a chromia-alumina catalyst of lower chromia content (7.2%).The results from the commercial catalyst were less reproducible, and although the catalyst shows significant changes in the e.s.r. spectra upon different treatments with hydrogen, the differences in the corresponding rates were less significant and an activation energy of 166.8 kJ mol-1 was obtained.I990 DEHYDROGENATION ON CHROMIA-ALUMINA CATALYSTS It is possible to draw a few general conclusions from the results so far obtained. First, both the e.s.r.and the activation-energy results indicate that the Cr3+ ions are operating as active sites. It is likely that these sites are formed during the following process at high temperatures : OH- HO- Cr3+ OH- -+ HO- Cr3+ 02- + H,O (g) (4) which involves the loss of water by condensation of surface OH- groups and therefore resulted in the formation of coordinatively unsaturated Cr3+ sites.2 Besides Cr3+, Cr2+ probably contributes to the overall activity by facilitating the breakage of a C-H bond, as previously sugge~ted.~ Secondly, the e.s.r. spectra show that most of the chromium is aggregated into clumps forming the /?-phase resonance; it is, therefore, reasonable to believe that the active sites, either as Cr3+ or a pair of Cr3+, are associated with this phase, and since Cr2+ is also present, under certain conditions, one may think in terms of a Cr2+/Cr3+ pair, as has been discussed by Dowden and Wells22 as a potential active site for dehydrogenation of alkanes.Finally, our results suggest that the activity and stability of chromia-alumina catalysts could be improved: (a) by increasing the chromia content, thus increasing the active phase which is associated with the /?-phase, and ( b ) by prolonged reduction with carefully purified and dried hydrogen at high temperatures. R. L. Burwell Jr, A. B. Littlewood, M. Gardew, G. Pass and C. T. H. Stoddart, J. Am. Chem. SOC., 1960,82, 6272. R. L. Burwell Jr, G. L. Haller, K. C. Taylor and J. F. Read, Adv. Catal., 1969, 20, 1 . P. P. M. M. Wittgen, C. Groenenveld, P.J. C. J. M. Zwaans, H. J. B. Morgenstern, A. H. van Heughten, C. J. M. van Heumen and G. C. A. Schiut, J. Catal., 1982, 77, 360. C. A. McAuliffe and F. M. Ashmawy, J . Chem. SOC., Faraday Trans. 1, 1984, 80, 1083. L. L. van Reijen, W. H. Sachtler, P. Cossee and D. Brouwer, Proc. 3rd Int. Conf. Catal. (North Holland, Amsterdam, 1964), vol. 2, p. 829. S. Carra, L. Forni and C. Vintani, J . Catal., 1967, 9, 154. J. Masson and B. Delmon, Proc. 5th Znt. Conf. Catal. (North Holland, Florida, 1972), vol. 1 , p. 207. J. Masson and J. M. Bonnier, J, Chem. SOC., Faraday Trans. I , 1977, 73, 1471. C. M. Marcilly and B. Delmon, J . Catal., 1972, 24, 336. lo F. M. Ashmawy and H. M. Steiner, J . Chem. SOC., Faraday Trans. I , 1977,73, 1646. l 1 F. M. Ashmawy, J. Chem. SOC., Faraday Trans. I , 1980, 76, 2096. l2 G. M. Panchenkov, A. S. Kazamskaya and A. D. Pershin, Pet. Chem. USSR, 1967,7,4. l3 A. Weiss, Refining Petroleum for Chemicals (American Chemical Society, Washington, 1970), p. 153. l4 F. M. Ashmawy, J. Appl. Chem. Biotechnol., 1977, 27, 137. l5 I. Suzuki and Y. Kaneko, J . Catal., 1977, 47, 239. l6 F. D. Richardson and J. H. Jeffes, J. Iron Steel Znst. (London), 1948, 160, 261. l7 C. G. Maier, U.S. Bur. Mines Bull., 1942, 436. la S. Voltz and S. Weller, J. Am. Chem. SOC., 1953, 75, 5227. l9 D. O’Reilly and D. MacIver, J. Phys. Chem., 1962, 66, 276. 2o Y. I. Petcherskaya, V. B. Kazansky and V. V. Voevodsky, Proc. 2nd Int. Conf. Catal. (Technip, Paris, 21 0. A. Hougen and K. M. Watson, Chemical Process Principles (Wiley, New York, 1943), vol. 3. 22 D. A. Dowden and D. Wells, Proc. 2nd Znt. Conf. Catal. (Technip, Paris, 1960), vol. 2, p. 1499. 1960), vol. 2, p. 2121. (PAPER 3/1826)

ISSN:0300-9599    

DOI:10.1039/F19848001985

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1984, 80, 1991-1997 Inelastic Neutron Scattering Spectroscopy of Hydrogen Adsorbed on Raney Nickel BY HERVE JOBIC* AND ALBERT RENOUPREZ Institut de Recherches sur la Catalyse, CNRS, 2 Avenue Albert Einstein, 69626 Villeurbanne Cedex, France? Received 18th October, 1983 We have obtained new results by inelastic neutron scattering spectroscopy for the adsorption of hydrogen on Raney nickel. Our interpretation is based on the vibrational frequencies and on the peak intensities which have been simulated for the different samples. We conclude that hydrogen is preferentially adsorbed in sites of C,, symmetry but that a significant occupation of sites having nearly three-fold symmetry can also be found for some samples. Hydrogen adsorption on nickel has been studied by a large number of spectroscopic techniques : optical spectroscopy,l? electron energy-loss spectroscopy (EELS)3-8 and inelastic neutron scattering spectroscopy (i.n.~).~-ll Several binding sites have been proposed from the vibrational frequencies, and comparisons have been made with the frequency values derived from theoretical calculations.l2 The usefulness of i.n.s. in observing vibrational modes involving hydrogen motions is now well established. The basic property of i.n.s. is the lack of selection rules: all the vibrations can be observed. Further, the i.n.s. frequencies and intensities, which can be computed from a normal-coordinate ana1ysis,l3 can be used to test the proposed models. The i.n.s. results for hydrogen adsorption have been obtained using Raney nickel because a high-surface-area substrate is required in this technique.14 Although the experimental results obtained by the different workers are in qualitative agreement, several revisions of the assignments have been reported.l5> l6 The difficulty in interpretating this system has been stressed in a recent review.17 Most authors have interpreted the i.n.s.spectra in terms of multiply bonded protons on the surface, but the presence of singly bound protons is disputed. Note that the stretching vibration of a terminal hydrogen atom occurs at a relatively high frequency (ca. 2000 cm-l) so that it can only be observed on a neutron reactor equipped with a hot source or on a pulsed neutron source. We report in this paper new experimental results for hydrogen adsorbed on Raney nickel.These new data will be used to complete the existing assignments and to give the relative proportions of the different species. EXPERIMENTAL SAMPLES The spectra reported in this paper were obtained under various experimental conditions. Sample 1 corresponds to the work described in ref. (9). The Raney nickel was evacuated to t Experimental work performed at the Institut Laue Langevin, Grenoble, France. 19911992 HYDROGEN ADSORBED ON RANEY NICKEL Elcrn-' 500 1000 1500 I I I I Fig. 1. 1.n.s. spectra of hydrogen adsorbed on Raney nickel, sample (1): (a) e~perimental,~ (b) calculated. loa6 Torr at 473 K and then heated at 573 K for 1 h. Hydrogen adsorption was performed during the neutron experiment. The i.n.s.spectrum of ref. (9) is reproduced in fig. l(a) (the curves corresponding to 8 = 0.42 and 8 = 1 are very similar). Sample 2 was prepared like sample 1, except that hydrogen adsorption, corresponding to 8 = 0.5, was carried out 15 days before the neutron experiment. Sample 3 is Raney nickel which was covered with hydrogen at saturation for 10 days, after which it was pumped at 600 K for 10 h, just before the neutron experiment, but with a relatively low pumping rate in comparison with sample 1. NEUTRON SPECTROSCOPY The neutron spectra were recorded at the Institut Laue Langevin using the old version of the beryllium-filter detector spectrometer IN 1 (this instrument is under reconstruction). The spectrometer has been described elsewhere.ls The experimental spectra given in fig.1 (a), 2(a) and 3(a) are shifted in energy because of the beryllium-filter transmission function, but the frequencies quoted in this paper have been corrected for this. The observed transitions are given with a precision of ca. 20 cm-l. All the i.n.s. spectra were recorded at 80 K.H. JOBIC AND A. RENOUPREZ 4000 I/) - c $ 3500- 2 5 3000- 2500- E: c 1993 - E l m - ' 500 1000 1500 ZOO( I I I I 30000 - m Y K 8 25000- e K 4- 2 20000- c 15000 - 50 100 150 200 250 ElrneV Fig. 2.1.n.s. spectra of sample 2: (a) experimental points ( x ) and (0) obtained with the (200) and (220) copper planes, respectively; (b) spectrum calculated for the main peaks. Elcrn-' 500 lo00 1500 2000 i I I 1 1 h v) Y ._ E: 3 f v .. 0.15 0.10 - - I I 1 50 100 150 200 250 ElmeV Fig.3. 1.n.s. spectra of sample 3: (a) experimental, (b) calculated.1994 HYDROGEN ADSORBED ON RANEY NICKEL The scattering from the hydrogen-nickel system can be analysed only in terms of incoherent scattering because of the large incoherent cross-section (cH) and the low mass (mH) of the hydrogen atom. The differential cross-section, corresponding to our experimental conditions,'* can be written as: In this expression, the neutron momentum transfer fiK is defined as K = k, - k where k, and k are, respectively, the incident and final wavevectors ; exp ( - 2 wd) is the Debye-Waller factor for atom d(this factor can be neglected under certain conditions).lg Therefore, in the one-phonon approximation, one obtains a peak at frequency col corresponding to the normal mode A, its intensity is governed by the vector c d ( n ) , which describes the displacement of the dth hydrogen atom.For a simple case such as hydrogen adsorbed on metal atoms, all the non-degenerate modes will have roughly the same i.n.s intensity. Likewise an E mode will be, to a first approximation, twice as intense as an A mode. A measure of the integrated intensities of the bands can then yield the populations of the various sites. RESULTS The dominant features of the i.n.s spectra of hydrogen adsorbed on Raney nickel are the peaks at 940 and 1100 cm-l. However, the relative intensities of these peaks appear to be different on going from one sample to another. For sample l9* lla the band at 1 100 cm-l appears only as a shoulder [fig. 1 (a)] whereas in ref.(1 1 b) the peak at 1100 cm-l is much more pronounced. This difference cannot be accounted for by a resolution effect. Therefore, the interpretation of the i.n.s. spectrum in terms of only one hydrogen species appears difficult. We believe that the extra intensity observed in some cases at ca. 1 100 cm-l is related to another peak which is situated at ca. 780 cm-l. This peak is clearly visible in fig. 2 (a), which corresponds to sample 2. In this spectrum one can observe three intense bands at 780,940 and 1100 cm-l, one broad band centred at 1900 cm-l and two weak bands at ca. 600 and ca. 1300 cm-l. It has been shown previously that the i.n.s. spectrum of hydrogen on Ni is almost independent of the degree of c ~ v e r a g e , ~ ? l l ~ and indeed we have found in other experiments the same shoulder at 780 cm-l for samples saturated with hydrogen.The same feature was also observed by Cavanagh et al.'lb but was not assigned. Because this peak is rather intense, it corresponds to a hydrogen species whose concentration is not negligible. We have succeeded in isolating this species by pumping at a high temperature (sample 3). The resulting i.n.s. spectrum is shown in fig. 3(a). Compared with fig. 2(a) the profile is much simplified and one obtains only two bands at 780 and 1080 cm-l. DISCUSSION On the basis of the frequency range and the relative intensities of the bands at 940 and 1100 cm-l, the i.n.s. spectrum of fig. 1 (a) can be easily assigned. The highest- frequency mode corresponds to the symmetric stretch (v,, A, symmetry) and the lowest-frequency mode to the antisymmetric stretch (v,,, E symmetry) of a CSfi symmetry species.We have simulated this spectrum by taking the mode at 940 cm-l to be twice as intense as the one at 1100 cm-l. The calculated spectrum, after introducing the experimental resolution, is shown in fig. l ( b ) (the widths of the calculated peaks are fitted to the experimental ones). The agreement with the experi- mental spectrum [fig. 1 (a)] is satisfactory.H. JOBIC AND A. RENOUPREZ 1995 The intensity which was observed around 2000 cm-l was due to multiphonons, since a terminally bonded species would have a bending mode which would contribute to the i.n.s. profile in the frequency range 600-1000 cm-l. Therefore, we agree with other authors16*20 that, in this case, hydrogen adsorbs predominantly on sites of three-fold symmetry, as on Ni(ll1).Note that there are two types of three-fold symmetry sites on Ni( 1 1 1) which appear equally occupied by hydrogen;21 both types appear to have the same binding energy as judged from the i.n.s. frequencies. However, we do not assign the small band at ca. 600 cm-l [fig. 2(a)] or ca. 640 cm-l [fig. 1 (a)] to a 'deformation mode' of the same species of C3v symmetry since this mode is expected at a much lower frequency on a metal surface.16 We tentatively assign this band, as have other authors,l19 2o to a symmetric stretching mode of a hydrogen atom bonded in a four-fold site (possibly distorted), since (i) a band has been observed at 600 cm-l in Ni( and (ii) theoretical calculations predict a frequency of 590 cm-l for a four-coordinated site.12 The antisymmetric stretches cannot be found in the i.n.s.spectrum; their frequency range is not known with accuracy and they may well be hidden by the multiphonon features if they are situated at frequencies above 1500 cm-1.20p 22 According to its intensity, this last species represents a very small proportion (< 5%) of the hydrogen adsorbed on Raney nickel, for all the i.n.s. experiments. For the predominant three-fold symmetry sites, the values of vas and vs can be used to calculate the Ni-H bond length.23 We find 1.87 f 0.03 A, in good agreement with the distance derived from LEED, 1.84 A thorough investigation based on the i.n.s. peaks at 940 and 1100 cm-1 gives the same value.llb If we turn to the i.n.s.spectrum of fig. 3(a), we find only two bands which are situated in the frequency range characteristic of three-coordinated sites. We can assign the peak at 1080 cm-l to a symmetric stretch and the peak at 780 cm-l, which is twice as intense, to the antisymmetric stretch. However, the width of the peak at 780 cm-l suggests that vas is not of E symmetry and that a lifting of degeneracy has occurred. Indeed, the instrumental resolution is ca. 10% of E,, the incident energy; the width of a mode at 1080 cm-l is thus expected to be greater than the width of a mode at 780cm-', but it is the reverse order which is observed. We have simulated this spectrum by taking two modes at 750 and 8 15 cm-l ( Ifr 10 cm-l) and a third one at 1080 cm-l (all these modes having the same intensity).The resulting spectrum is shown in fig. 3(b) and it well reproduces the experimental profile of fig. 3(a), except for the bandwidths. The widths of the experimental peaks are noticeably reduced compared with the other samples (this effect is more obvious for the mode at 1080 cm-l). This narrowing of the peaks, which has already been observed at low hydrogen concentrations by i.n.s.,'lb is due to a reduction of dynamical interaction (presumably indirect24) between the adsorbed hydrogen atoms. To obtain an estimated Ni-H bond length for this species one can take an average value of 780 cm-l for v,, and a value of 1080 cm-l for vs; a distance of 2.02 A is then obtained. Possible sites for this species are the nearly three-fold hollow sites on the Ni(ll0) unreconstructed surface (taking into account the second nickel layer) or the corre- sponding sites on the reconstructed A reconstruction on the Ni( 1 10) surface occurs irreversibly above 220 K in the presence of hydrogen.25 In our case, it seems that a reconstruction occurs after several days on the Raney nickel surface.A similar effect appears to be obtained by using a higher temperature of treatment, which provokes sintering with probable surface reconstruction; this could explain the different i.n.s. results reported in ref. (1 1 a ) and (1 1 b), where the temperature of treatment was 473 and 550 K, respectively. A combination of the preceding sites can now be used to interpret the i.n.s. spectrum1996 HYDROGEN ADSORBED ON RANEY NICKEL of sample 2 [fig.2(a)]. For this sample we find that ca. 60% of the hydrogen atoms are in the C,, sites and ca. 40% in the nearly three-fold symmetry sites. The resulting spectrum, which is shown in fig. 2(b), reproduces well the main features of fig. 2(a). Note that, because the instrumental resolution depends on the incident energy,18 the comparison between the experimental and calculated intensities is based on the peaks’s areas. The small peak at 600 cm-l was discussed for sample 1. The broad band centred at 1900cm-l corresponds to the overtones of the intense peaks (780, 940 and 1100 cm-l). Finally, a small shoulder exists at 1300 cm-l; it was also observed in other experiments where background subtraction was performed. Its frequency range fits well with the symmetric stretching mode of a (p2-H) species.12 The very small intensity of this peak indicates that its concentration is negligible.For the majority of samples that we have studied, the spectrum is intermediate between spectra 1 and 2, and thus the relative proportions of the two three-coordinated species vary accordingly. COMPARISON WITH EELS RESULTS If one compares the neutron data with EELS result^,^-^ it appears that many observed frequencies coincide at ca. 620, 800, 940, 1100 and 1300 cm-l. However, a precise assignment of the EELS spectra is difficult because a quantitative prediction of the peak intensities is not yet possible and thus the relative proportions of the different species cannot be determined. Further, a critical examination of the results obtained by EELS, on the same Ni planes, shows discrepancies between the observed intensities. On the Ni( I 10) (1 x 2)-H reconstructed room-temperature phase, Nishijima et al.observed three losses at 641,936 and 1121 cm-l. The peak at 641 cm-l had the lowest intensity and the peak at 936cm-l the strongest.8 DiNardo and Plummer, who published another study for the same (1 x 2) reconstructed phase, report that ‘the losses are very weak and appear to correspond to the high-coverage modes’. At saturation exposure they observe at 130 K a strong peak at 610 cm-l and a very weak peak at 940 cm-l; a third peak at 1130 cm-l, the intensity of which is not stated, was observed in the off-specular direction.’ Both groups interpreted the data in terms of three-coordinated species but the assignments differ : the symmetric stretch (v,) was assigned at 610’ or at 1 1208 cm-l. In our work the symmetric stretches of the two three-coordinated species are found near 1 100 cm-l.On Ni( 11 1) two EELS studies have also been published. Ibach and Bruchmann have found a well pronounced peak at 1130 cm-1,6 whereas Ho et al. observed two broad peaks at 710 and 1120 ~ m - l . ~ Ho et al. assigned the peak at 1120 cm-l to the asymmetric stretch of a three-fold coordinated species, but we prefer to assign it to the symmetric stretch on the basis of the i.n.s. intensities. CONCLUSIONS The i.n.s. spectrum of hydrogen adsorbed on Raney nickel has been reanalysed. It is clear that the spectrum may change depending on the sample preparation.In all cases, the proportion of (pl-H), (p2-H) and (pU4-H) species is small. We find, in agreement with previous studies, that the main species occupies a three-fold symmetry site exactly as on the Ni(ll1) surface. However, the proportion of sites of nearly three-fold symmetry found on Ni( 1 10) can be as high as 40 % .H. JOBIC AND A. RENOUPREZ 1997 T. Nakata, J . Chem. Phys., 1976, 65, 487. W. Krasser and A. Renouprez, J. Raman Spectrosc., 1979, 8, 92. S. Anderson, Chem. Phys. Lett., 1978, 55, 185. S. Lehwald and H. Ibach, Surf. Sci., 1979, 89, 425. W. Ho, N. J. DiNardo and E. W. Plummer, J . Vac. Sci. Technol., 1980, 17, 134. H. Ibach and D. Bruchmann, Phys. Rev. Lett., 1980, 44, 36. N. J. DiNardo and E. W. Plummer, J. Vac. Sci. Technol., 1982, 20, 890.M. Nishijima, S. Masuda, H. Kobayashi and M. Onchi, Rev. Sci. Instrum., 1982, 53, 790. A. Renouprez, P. Fouilloux, G. Coudurier, D. Toccheti and R. Stockmeyer, J. Chem. Soc., Faraday Trans. I , 1977, 73, 1 . R. Stockmeyer, H. Stortnik, I. Natkaniec and J. Mayer, Ber. Bunsenges. Phys. Chem., 1980, 84, 79. I 1 (a) R. D. Kelley, J. J. Rush and T. E. Madey, Chem. Phys. Lett., 1979,66, 159; (b) R. R. Cavanagh, R. D. Kelley and J. J . Rush, J . Chem. Phys., 1982, 77, 1540. lZ T. H. Upton and W. A. Goddard, Phys. Rev. Lett., 1979, 42, 472. l3 H. Jobic and A. Renouprez, Surf. Sci., 1981, 111, 53. l4 R. K. Thomas, Prog. Solid State Chem., 1982, 14, 1 . l5 C . J. Wright, J . Chem. Soc., Faraday Trans. 2, 1977, 73, 1497. l6 D. Graham, J. Howard and T. C. Waddington, J. Chem. Soc., Faraday Trans. 1, 1983, 79, 1281. I’ C. J. Wright and C. M. Sayers, Rep. Prog. Phys., 1983, 46, 773. l9 A. Griffin and H . Jobic, J. Chem. Phys., 1981,75, 5940. 2o C. M. Sayers, J. Phys. C, 1983, 16, 2381. 21 K. Christmann, R. J. Behm, G. Ertl, M. A. Van Hove and W. H . Weinberg, J . Chem. Phys., 1979, 22 J. E. Black, P. Bopp, K. Lutzenkirchen and M. Wolfsberg, J. Chem. Phys., 1982, 76, 6431. 23 J. A. Andrews, U. A. Jayasooriya, I. A. Oxton, B. Powell, N. Sheppard, P. F. Jackson, B. F. G. Johnson and J. Lewis, Inorg. Chem., 1980, 19, 3033. 24 C . Nyberg and C. G. Tengstal, Phys. Rec. Lett., 1983, 50, 1680. 25 T. Engel and K. H. Rieder, Surf. Sci., 198 1, 109, 140. 26 K . H. Rieder, Phys. Rev. B, 1983, 27, 7799. 27 J. E. Demuth, J. Colloid Interface Sci., 1977, 58, 184. H. Jobic, R. E. Ghosh and A. Renouprez, J. Chem. Phys., 1981, 75, 4025. 70, 4168. (PAPER 3/ 1852)

ISSN:0300-9599    

DOI:10.1039/F19848001991

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

J. Chem. SOC., Faruduy Trans. I, 1984,80, 1999-2016 Interactions between Calcium Ions and a Range of Monosaccharides Studied by Hydroxy-proton Resonance Spectroscopy? BY MARTYN C. R. SYMONS,* JOHN A. BENBOW AND HEATHER PELMORE Department of Chemistry, The University, Leicester LE 1 7RH Received 19th December. 1983 At low temperatures in a small pH range, well resolved hydroxy-resonance features have been detected for dilute aqueous solutions of sugars. Shifts in these resonances on the addition of calcium chloride have been followed for the major forms of a range of monosaccharides. These trends take three forms: (i) normal, linear, up-field shifts, comparable to that for the water-proton resonance, indicating no preferential interaction, (ii) reduced up-field or weak down-field shifts, which remain almost linear, assigned to weak binding of two OH groups to Ca2+ ions, either adjacent axial +equatorial groups or two cis axial groups separated by one -CHX- group, and (iii) down-field shifts becoming normal up-field shifts at high salt concentrations, assigned to strong binding of Ca2+ ions to three adjacent OH groups in the axiakquatorial-axial configuration. Only in the case of strong complexing (iii) is there clear evidence for cation-induced conformational changes.These assignments apply to the pyranose forms of the sugars. Our more limited results for the furanose forms strongly support the suggestion that three cis OH groups are required for strong bonding. When the anomeric (0,H) proton in a- or P-D-glucose is replaced by a methyl group the weak binding to Ca2+ ions observed for the 0,H and 0,H groups is lost, since the down-field shift for the 0,H-proton resonance changes to a normal up-field shift.This means that the -0Me group fails to coordinate to the Ca2+ ions. Reasons for this result are proposed. Some time ago we established that separate water- and hydroxy-proton resonances of alcohols' and sugar^^-^ can be resolved for dilute aqueous solutions. Normally there is rapid proton transfer between all OH protons, giving a single average resonance, but slow exchange can be achieved by lowering the temperature and carefully controlling the pH. Although of potential importance to carbohydrate chemists, these results have not been widely exploited. We have recently shown that metal-ion binding to sugars can be probed by this technique, using D-ribose with calcium chloride or perchlorate as an e ~ a m p l e .~ The results strongly supported the more indirect work of others on these thereby establishing the validity of the method. In our studies, all solvent and ionic interactions at the OH group in question contribute to the observed shifts, whereas only major ion-induced conformational changes are detected when C-H resonances are studied. These studies have been extended to a range of monosaccharides, and despite the complexity of the system several useful results have been obtained. EXPERIMENTAL The sugars were of the highest grades available and were used as supplied. Aqueous sugar solutions were generally ca. 1 .O mol dm-3 or higher to give strong hydroxy features for the major t Taken as Solvation Spectra, Part 74.19992000 20 If E E T 10 5 INTERACTIONS BETWEEN CALCIUM IONS AND MONOSACCHARIDES I I I I I 6.5 5.0 5.5 6.0 6.5 PH Fig. 1. pH minima plots for D-mannose (1.25 mol dm-3). ., [CaCl,] = 0.84 mol dm-3; 0, [CaCl,] = 0.5 mol dm-3; (i) non-anomeric hydroxy; (ii) 0,H (ap); (iii) 0,H (pp). conformers present in solution. The pH values of the solutions were measured at room temperature (ca. 22 "C) with a Pye Microstem combination glass electrode and a Pye Dynacap HOSE"2 meter modified to take E"7 electrodes. In all cases the pH was controlled by a buffer system consisting of ca. 1 mmol dm-3 of sodium hydrogen malate and maleic acid, and optimum conditions were sought to give the narrowest features.Calcium chloride (AnalaR) was dried in uucuo at ca. 100 "C, weighed and added to stock solutions of the sugar in purified water. In some cases the solutions were warmed to ca. 80 "C to effect complete dissolution. The pH was then adjusted as described above. N.m.r. spectra were measured on a Jeol PS-100 spectrometer at -9 "C. Temperatures were measured with a Comark thermocouple. Measurements were made rapidly after cooling since certain solutions tended to phase-separate on long standing at this temperature. Shifts were measured from the most prominent C-H resonance feature. In all cases, shifts induced in these features were zero within kO.025 ppm from the marked resonance of TSP [3-(trimethylsilyl)propionic acid sodium salt]. These shifts are small compared with those under consideration and no attempt was made to obtain more accurate absolute shifts.RESULTS AND DISCUSSION A typical pH-against-linewidth plot is shown in fig. 1. In all cases such plots were obtained in order to provide minimum widths and hence maximum accuracy. The temperature of - 9 "C was selected as the best compromise between the aim for narrow lines and the need to prevent phase-separation. Significant, systematic trends with [CaCI,] were found for the pH values giving minimum linewidths (fig. 2). We note that as [CaCl,] increases so the minimum shiftsM. C. R. SYMONS, J. A. BENBOW AND M. PELMORE 200 1 i 6.5 x I ~ 0.5 1 .o 1.5 [ CaCl, 1 /mol dm-3 Fig. 2. pH minima for D-mannose (1.25 mol drnp3) against [CaCl,]. x , Non-anomeric hydroxy ; , anomeric hydroxy.to low pH values, the non-anomeric resonance being most affected. It is also noted that the pH shifts for ribose and mannose, which form ‘strong’ complexes, are greater than for the remaining sugars. This suggests that binding to Ca2+ ions increases the acidity of the OH protons so that a lower pH is required to suppress base-catalysed proton exchange. Most of these plots are effectively linear, their slopes being obtained using a standard linear regression analysis (table 1). In some cases small changes in the relative concentrations of a- and p-forms were observed during measurement. However, in no case were they significant enough to interfere with spectral analysis. As we have shown previously,2. features for the anomeric hydroxy-proton resonance were strongly shifted down-field from the water-proton resonance and could be monitored without difficulty. However, the remaining resonances were bunched close to each other in the tail of the water feature and it was not always possible to follow shifts for specifically assigned OH resonances.A typical spectrum is shown in fig. 3 and specific peak assignments follow those given previo~sly.~$~ In several cases addition of salt resulted in line-narrowing and consequent resolution into multiplets because of coupling with C-H resonances. This extra resolution can be confusing when shifts in several adjacent features are monitored. However, the plots given in the following figures show good consistency and we are satisfied that no serious confusion occurred in our analysis.2002 INTERACTIONS BETWEEN CALCIUM IONS AND MONOSACCHARIDES & s 0 n 0 rA % 2' x a x W El a x c 0 L .r( g 2 0 P 0 [I h ? z 2 0 + d 2 + 3 Z 0" I I r= 0 + c ID-galactosec 1.46 ap O,HB Pp +0.32 o*12] O,Ha -0.30 { 03,4Ha,B :::$] O,Ha)B +0.34 +0.24 C,HB -0.09 cL2.0 D-fuCOSeC 1.05 (6-deoxy- D-galactose) - +0’43] O,Ha -0.27d ( O,HB 03,4Hm,B] +0.78 - Pp +0.65 +0.70 - - cL2.0 L-sorboseC 1.38 ap -0.09 0 3 H a -0.27 04,,Ha +0.37 O,Ha +0.24 +0.24 - - G1.7 - - - - - - - - - -0.01 - -O.lOd -0.22- f i D-fructosec 1.60 Pp -0.15 0,,4HB { O,HB +0.15 O,HB f0.31 +0.29 C,H +0.03 cL1.5 - - - - - - - - pf -0.17 - - a Linear regressions were calculated using a TI 57 programmable calculator.The standard linear regression equation applies : m20x2 slope = rn = ( i!l xi yi -xy)/ox2 and correlation coefficient = ~ m I oy2 I.Slopes are quoted to kO.005 for correlation coefficients 3 0.98 unless specified. of the alkyl resonances and plotted against [CaCl,]. was added to follow the shift of the alkyl reference peak. All hydroxy chemical shifts are measured w.r.t. the tallest feature Tentative assignments of non-anomeric hydroxys based on the ax-eq model proposed. TSP reference Correlation coefficient < 0.9. Correlation coefficient 3 0.9. Chemical shift measured from internal TSP (ca. 0.2 mol dm-3). Only one resonance was clearly defined in the spectra.2004 INTERACTIONS BETWEEN CALCIUM IONS AND MONOSACCHARIDES ( i ) (ii) (iii) (iv) (v)) (vii) 6 5 4 3 2 1 6 (PPm) Fig. 3. N.m.r. spectrum for a solution of D-mannose (mol dm-3) in water at -9 “C and buffered at pH 5.95. (i) 0,H up); (ii) 0,H (ap); (iii) O,,,,,H (a, pp); (iv) 0,H (a,pp); (v) sideband; (vi) OH water; (vii) ref.C,H. GENERAL ASPECTS OF HYDROXY -PROTON RESONANCE SHIFTS The shift for the water-proton resonance on adding CaCl, is shown in fig. 4 (slope 0.14). This up-field shift has been analysed in terms of a down-field component from Ca2+ ions and an up-field contribution from chloride ions, which are present at twice the concentration.8 Although we now have some reservations regarding the precise numerical values previously proposed for these individual shift^,^ nevertheless they are probably correct qualitatively. As stressed in our study of D-ribose, the limiting shift obtained for strong binding to Ca2+ ions is close to that expected for these ions in water.This gives us a rough measure of the extent of interaction of OH groups with Ca2+ ions. Nevertheless, we stress that changes in the concentration of ‘free’ hydroxy and lone-pair groups in water, thought by one of us to be of major importance in the chemistry of water and alcohols,1° must also play a part in governing the observed shifts. These are, in effect, taken together under the general heading of ‘anion’ shifts in both our previous work and here. These general ‘anion’ shifts for the OH resonances are expected to be greater than that for pure water (slope ca. 0.26). As can be seen, this is true for several of the shifts given in table 1 and shown in the figures. However, in certain cases the positive (up-field) shifts are greater than average.These are found when strong binding to Ca2+ ions is observed. We argue that preferential solvation of Caz+ by the sugar reduces the number of Ca2+. * .OH interactions with water and the remaining sugar hydroxy groups. Hence these experience larger nett up-field shifts.2005 I I I I 0 0.5 1.0 1.5 [CaC12 1 /mol dm-3 Fig. 4. Shifts (in ppm) in the hydroxy-proton resonance features for D-xylose in water as a function of the concentration of calcium chloride. (i) + (ii) CaC1, water shift, (i) with D-xylose (1.4 mol dm-3) present, slope 0.29, and (ii) without D-xylose present, slope 0.14; (iii), (iv) and (v) O,,,,,H pyranose, not specifically assigned; (iii) slope 0.45, (iv) slope 0.39 and (v) slope - 0.13; (vi) 0,H (ap), slope 0.27; (vii) 0,H (ap), slope 0.43.(See scheme 1 for additional information.) w4C Cu-lC4 (-, ---) Conformational preference; (- --) possible weak complexing with Ca? Scheme 1.2006 INTERACTIONS BETWEEN CALCIUM IONS AND MONOSACCHARIDES When weak complexing with Ca2+ ions occurs, there is an extra down-field contribution to the shifts of the OH protons involved. This is diagnosed for slopes well below the ‘normal’ range (i.e. < ca. 0.3). In some cases, zero or small negative slopes were obtained, but the trends are all linear in the 0-2 mol dmP3 range. Such changes are taken as evidence for ‘ weak ’ complexing. In all cases, two hydroxy-proton resonances are involved, the most common being for adjacent OH groups in pyranose rings, one being axial and the other equatorial [see structure (I)].However, in certain cases, two cis axial groups separated by one -CH(OH)- unit gave comparable weak complexes. This seems to be the first clear evidence for such weak complexing, or ‘preferential solvation ’. Strong complexing for the pyranoses was only observed when there was an axial-equatorial-axial arrangement of three adjacent OH groups, the two axial units being cis to each other (ribose and mannose). For the furanoses our more limited data show that only a cis-cis-cis arrangement gives rise to ‘ strong’ interactions. As we established for r i b ~ s e , ~ when ‘ strong’ complexes ,are formed there are marked down-field shifts for the hydroxy-proton resonances of the groups involved, which reach minima and then shift up-field again as [CaCl,] is increased.[See, for example, fig. 2 and 3 of ref. (5) and fig. 8 below.] We postulate that complexing is effectively complete close to these minima, the ultimate up-field shifts arising from normal ‘anion’ effects on the OH protons of the complexes. In our previous work we endeavoured to obtain accurate estimates of equilibrium constants for ~omplexing.~ The results for the present complexes are quite comparable. In addition to solvent- and ion-induced shifts, we need to consider possible conformation shifts. Besides the slowly established equilibria between a- and /?- structures and 5- and 6-membered rings, the pyranoses undergo chain-chain inver- sions which interchange the (OH) groups between axial and equatorial sites. For aldoses these inversions are summarised as ‘C, =$ 4C1 and for ketoses as 2C5 5C2.These inversions are rapid on the n.m.r. timescale even at -9 “C and lead to shifts which are the weighted average of the two extreme values. We previously deduced that the protons of axial OH groups generally resonate at higher fields than corresponding equatorial OH g r o ~ p s , ~ the differences being probably up to ca. 0.5 ppm. Since these rapid conformational equilibria induce the change any constraint on the equilibrium may produce a considerable shift in the resonance, to high or low fields. Major conformational changes are only expected when one conformation can form a ‘strong’ complex (ax-q-ax). The induced shift must be added to the negative Ca2+ shift, so this may be augmented or reduced.Nevertheless, the characteristic minimum in the shift plots should be observable in either situation. In the following, details of the results for specific sugars are outlined and discussed.2007 [CaC12 1 /mol dm-3 Fig. 5. Shifts (ppm) in the hydroxy-proton resonance features for D-glucose in water as a function of the concentration of calcium chloride. (i) + (ii) CaCl, water shift, (i) with D-glucose (1.4 mol dm-3) present, slope 0.282 and (ii) without D-glucose present, slope 0.14; (iii) 0,H ap, pp, slope 0.427; (iv) 0 , H Pp, O,,,H ap, slope 0.494; (v) and (vi) O,,,H (Bp), (v) slope 0.548 and (vi) slope 0.44; (vii) 0,H (ap), slope -0.108; (viii) 0 , H (ap), slope 0.221 ; (ix) 0 , H Up), slope 0.398. (See scheme 2 for additional information.) P4C1 w4C a-'C4 Scheme 2.2008 INTERACTIONS BETWEEN CALCIUM IONS AND MONOSACCHARIDES D-XYLOSE This, like ribose, is a C, sugar.The a- and P-pyranose forms were detected, but no features for the low-abundant furanose forms were resolved. In contrast with ribose, none of the forms shown in fig. 4 have the ax-eq-ax conformation thought to be required for strong bonding, and indeed, the shifts, also shown in fig. 4, are all linear, showing that ‘strong’ bonding does not occur. However, the slope for the water-proton shift is large (0.29), suggesting that there is some depletion of Ca2+ ions, and the shift for one of the non-anomeric OH protons is negative. This is taken as strong evidence for ‘weak’ complexing. We note from the structures given in fig. 4 that the lC, form of the /3-sugar has two cis axial OH groups whilst for the a-structure the T, has an ax-eq pair and the lC, has a cis ax-ax pair.However, the/?-C, structure dominates.,, Thus, unless Ca2+ ions can strongly shift the + lC, equilibrium, the 8-sugar will not complex significantly. (Some evidence that this constraint can be important is given below.) For the a-sugar the ,C, conformer dominates, so we postulate that it is this eq-ax interaction that is responsible for the shift. This involves the anomeric OH group, and it is noteworthy that the slope for this line (0.27) is greatly reduced relative to that for the /3-sugar (0.43). Why should the resonance for the 0,H proton be so much more affected than that for the 0,H proton? A shift in the conformational equilibrium from ‘C, to T, changes the 0,H group from equatorial to axial, giving an up-field contribution, and the 0,H group from axial to equatorial, giving a down-field contribution.This nicely accommodates the results, but requires that there be a significant population of the lC, structure, that the ax-eq complex be stronger than the cis ax-ax complex (this is in good accord with all our present results) and that the strength of Ca2+ ion bonding in this ax-eq complex be strong enough to shift the equilibrium (an alternative explanation is given below). D-GLUCOSE Again there are no ax-eq-ax conformers and the furanose structures were not detected. As with D-xylose, however, the 0,H resonance for the a form shifts less than that for the P form and one of the non-anomeric OH resonances (0,H) has a slightly negative shift (- 0.1) (fig.5). All other shifts are slightly more positive than average. All this points to weak Ca2+ complexing. For the jl structure is the major conformer (probably because all the OH groups are equatorial, which permits stronger hydration3g4), and weak complexing is only possible for T,. (This is for cis ax-ax which is not thought to be significant enough to influence the equilibrium.) For the a isomer both ‘C, and conformers can form ax-eq complexes with Ca2+ ions. Once again the form dominates, and, in our view, this is responsible for forming the complex and causing the resonance shifts. Once again, however, the effect is more marked for the 0,H proton than the 0,H proton. In this instance a change in the conformational equilibrium, tentatively suggested to explain this effect for D-xylose, is less satisfactory, since both structures can form ax-eq complexes.We have therefore sought an alternative explanation. We suggest that the basicity of the anomeric oxygen is less than that for the non-anomeric oxygen. This accords with the electron-withdrawing effect of the ring oxygen, which reduces the basicity and increases the acidity of the 0,H groups. In that case, the Ca2+-oxygen bond will be weaker and the change in shift less for the 0,H-proton resonance than that for the 0,H resonance, as observed. This may also pertain to the shifts for D-xylose.M. C. R. SYMONS, J. A. BENBOW AND M. PELMORE 2009 4 .O 0 a/* I,*, 0.5 1.0 1.5 , I I I 0 0.5 1.0 1.5 [CaCl, I /mol dm-3 Fig.6. Shifts (in ppm) in hydroxy-proton resonance features for 2-deoxy-~-glucose in water as a function of the concentration of calcium chloride. (i) and (ii) CaCl, water shift, (i) with 2-deoxy-~-glucose (1.4 mol dm-3) present, slope 0.1 1, and (ii) without 2-deoxy-~-ghcose present, slope 0.14; (iii) 0,H (a, pp), slope 0.15; (iv) and (v) O,,,,,H pyranose not specifically assigned, (iv) slope 0.19 and (v) slope 0.21; (vi) 0,H (ap), slope 0.21; (vii) 0,H (pp), slope 0.13. 2-DEOXY -D-GLUCOSE On removal of the 0 , H group, such a complex for the a-sugar can no longer form, so we predict that no complexing should occur. Our results are in good accord with this postulate (fig. 6). The shifts are all linear, with slopes similar to that for calcium chloride and water containing no sugar.This therefore indicates no binding of Ca2+ ions. METHYL-, a- AND B- D-GLUCOPYRANOSE These methyl esters are of interest since they probe the ability of the methoxy group to complex with Ca2+ ions. The results show quite conclusively that there is no2010 INTERACTIONS BETWEEN CALCIUM IONS AND MONOSACCHARIDES ( v i i ) (viiil 3.01 , I I , [CaCI,] /mol dm-3 0 3 1.0 1.5 2.0 Fig. 7. Shifts (in ppm) in hydroxy-proton resonance features for methyl, a- and p-D- glucopyranoside as a function of the concentration of calcium chloride. (i) CaCI, water shift with methyl, a- and fl-D-glucopyranoside, slope 0.30; (ii) 0,H a, slope 0.48; (iii) CaCl, water shift without methyl, a- and p-D-ghcopyranoside, slope 0.14; (iv) and (v) 02,3,4,5H a, slopes between 0.38 and 0.18; (vi) 0,H p, slope 0.50; (vii) and (viii) O,,,,,,,H p, slopes between 0.41 and 0.34.significant complexing with Ca2+ ions (fig. 7). Thus replacement of H by CH, has either reduced the effective basicity of the anomeric oxygen to such an extent that it no longer coordinates to Ca2+, or there is a steric barrier to complex formation provided by the methyl group. Probably both factors contribute. We stress that for R-OH groups, hydrogen bonding to the proton (R-OH. * - O < ) is always strong,l? and this markedly enhances the basicity of the oxygen atom. This effect is largely responsible for the enhanced basicity of the -0,H group relative to -0,Me.1.0r OH present, slope 0.14; (v) unresolved slopes containing (eq-ax-eq strong complex) (ax-eq-ax strong complex) I (Y-~C, OH HOCH HO 4; HOCH, HO+H2012 INTERACTIONS BETWEEN CALCIUM IONS AND MONOSACCHARIDES This result is important since it provides an explanation of the fact that there is no evidence for complexing via the ring oxygen, and it suggests that the bridging oxygens in polysaccharides are not likely to contribute to complex formation.D-MANNOSE As with ribose, both the furanose and /?-pyranose structures should form strong complexes since the former has a cis-cis-cis structure and the latter an ax-eq-ax structure involving O,H, 0,H and 03H groups. The results are in complete accord with this (fig. 8). The minimum for the furanose complex occurs earlier than that for the /?-pyranose complex, showing that the former interacts more strongly. Approximate equilibrium constants for D-mannose were calculated as KBp = 5.0f2 and Kpf = 4.0+ 1.5, using the same procedure as in our previous These results are close to those for ~-ribose,’ which gave Kocp = 4.6 and Kaf = 2.6.(These results are unfortunately inverted in our original paper.5) In contrast with the results for ribose, which showed shifts to low-field of ca. 0.9 ppm, the low-field shift for the /?-pyranose form of mannose is only ca. - 0.1 ppm for the anomeric proton resonance and ca. zero for the non-anomeric proton (fig. 8). However, in this case it is the lC4 conformation that forms a strong complex, whereas the normal structure is 4C,. Hence complexing causes a complete switch from 4C, to lC4, which should give an up-field contribution to the anomeric proton resonance and hence account for the small nett shift.Unfortunately this theory could not be tested for the non-anomeric peaks as the latter were shifting too rapidly for proper identification. The anomeric OH group for the a-pyranose sugar is not able to participate in complex formation, and the strong positive shift is in agreement with this. It is interesting that the water resonance shift is actually slightly reduced, rather than showing the predicted enhancement caused by loss of Ca2+ ions. For ribose, there was almost no change for CaCl, but the expected strong enhancement was observed for Ca(ClO,),. We suggest that the chloride ions, which are strongly solvated, tend to form complexes with sugar molecules, possibly with those complexed with Ca2+ ions, giving a sort of ion pairing.This would remove both Ca2+ and C1- from the water, thus reducing or even reversing the expected trend. D-GALACTOSE Only the a- and P-pyranoses were detected. For the former both the 4C1 and the lC, conformers are expected to give weak ax-eq complexes (fig. 9). Indeed, the latter could bind two Ca2+ ions, but this is considered to be most unlikely. The low values of the linear shifts for the a-0,H and a-0,H resonances accord with this satisfactorily. For the P-sugar, the 4C, structure is favoured and 0 3 H and 04H form an ax-eq pair. In accordance with this, the shifts for the O,H, 0,H and 0,H resonances are normal, the other two being greatly reduced. D-FUCOSE This is 6-deoxy-~-galactose, so the 0,H group is lost.Complexing should therefore be similar to that for galactose, and the results accord well with this. For this reason we have omitted to show these trends. L-SORBOSE This sugar and fructose were the only two ketoses studied. The a structure exists largely as ,C5, which can give a weak ax-eq complex with 0,H (anomeric) and 03H (fig. 10). The expected linear negative shifts were observed, that for the 0,H proton-OH 3.51 - (viii) (-----) Possible weak complexing with Ca” Scheme 4. Fig. 9. Shifts (in ppm) in the hydroxy-proton resonance features for D-galactose in water as a function of the concentration of calcium chloride. (i) CaC1, water shift with D-galactose (1.4 mol dm-3) present, slope 0.242; (ii) 0,H (a$), slope 0.337; (iii) CaCl, water shift without D-galactose present, slope 0.14, 0 0.5 1 .o 1 .s slope-0.302; (vii) 0,H a, slope 0.123; (viii) 0 , H p, slope /. /./. 4.0 ,.A I I (iv) 0,H /3, slope 0.3 18; (v) O,.,H (a, B), slope 0.02; (vi) 0,H a, [CaCl, I /mol dm-3 0.3 17. (See scheme 4 for additional information.) ./. I ,P 2cs (38%) no complex CH,OH cr-5c2 ( Y - T 5 I I I I (none) (100%) 0 0.5 1.0 1.5 [CaC12] /mol dmd3 Scheme 5. Fig. 10. Shifts (in ppm) in the hydroxy-proton resonance features for L-sorbose in water as a function of the concentration of calcium chloride. (i) CaCl, water shift with L-sorbose (1.38 mol dm-3) present, slope 0.238; (ii) O,H, slope 0.236; (iii) O,.,H, slope 0.367; (iv) weak furanose, slope 0.37; (v) 0,H (ap), slope -0.268; (vi) 0,H (ap), slope -0.093; (vii) 0 , H vp), slope -0.05.(See scheme 5 for additional information.)1.50 / (ii' / * I '(v) 3.50 1 (vii) I I I 1 0.5 1.0 1.5 ?H HCHzoH OH pp-5c2 (0%) pp-2c, (1 00%) [78%1 \ -OH OH ap-ZC, (ca. 2 1%) arp-5cz (ca. 79%) [none] I OH p-furanose [18%1 a-furanose [cu. 4%1 [CaCl,] /mol dm-3 Scheme 6. Fig. 11. Shifts (in ppm) in hydroxy-proton resonance features for D-fructose in water as a function of the concentration of calcium chloride. (i) water, slope 0.279; (ii) 0,H (pp), slope 0.309; (iii) 03,4,5H (pp), slope 0.148; (iv) O,,,,,H (pp), slope -0.22; (v) O,,,,,H (pp), slope -0.282; (vi) 0,H (pp), slope -0.145; (vii) pf, slope -0.165. (See scheme 6 for additional information.)2016 INTERACTIONS BETWEEN CALCIUM IONS AND MONOSACCHARIDES being surprisingly steep (fig. 10). We are not clear why this OH proton exhibits such a large negative slope. Both conformers are present for the j3-sugar, but only 5C2 can form a complex (two sets of cis ax-ax groups). The shifts for these resonances are close to zero, showing that weak complexing does occur. Any shift in the conformational equilibrium to favour 5C2 would contribute an up-field shift, so this cannot be of major importance. D-FRUCTOSE The furanose structure has two cis OH groups (fig. 11,O,H and 03H) so we expect weak complexing. This is evidenced by the small, linear, down-field shifts for these protons. The a-pyranose form was not detected because of its very low abundance. The p form is primarily 2C5, which provides two ax-eq sites for weak binding of Ca2+ ions (0,H + 03H and 0,H + 05H). In accordance with this, all resonances except 0,H (the CH,OH proton) show reduced positive or negative shifts, in very satisfactory accordance with our predictions. J. M. Harvey, S. E. Jackson and M. C. R. Symons, Chem. Phys. Lett., 1977,47,440. J. M. Harvey, R. J. Naftalin and M. C. R. Symons, Nature (London), 1976, 261, 435. J. M. Harvey and M. C. R. Symons, J. Solution Chem., 1978,7, 571. M. C. R. Symons, J. A. Benbow and J. M. Harvey, Carbohydr. Res., 1980,83, 9. M. C. R. Symons, J. A. Benbow and H. Pelmore, J. Chem. SOC., Faraday Trans. 1, 1982,78, 3671. S. J. Angyal, Aust. J , Chem., 1972, 25, 1957; Pure Appl. Chem., 1973, 35, 131. R. E. Lenkinski and J. Reuben, J. Am. Chem. SOC., 1976,98, 3089. M. C. R. Symons, J . Chem. SOC., Faraday Trans. 1, 1983,79, 1273. * J. Davies, S. Ormondroyd and M. C. R. Symons, Trans. Faraday SOC., 1971,67, 3465. lo M. C. R. Symons, Acc. Chem. Res., 1981, 14, 179. l1 S. J. Angyal, Angew. Chem., Znt. Ed. Engl., 1969, 8, 157. (PAPER 3/2235)

ISSN:0300-9599    

DOI:10.1039/F19848001999

出版商:RSC    

年代:1984

数据来源: RSC