摘要:

J . Chem. SOC., Faraday Trans. 1, 1985, 81, 2467-2474 Photocatalytic Dehydrogenation of Aliphatic Alcohols by Aqueous Suspensions of Platinized Titanium Dioxide BY SEI-ICHI NISHIMOTO, BUNSHO OHTANI AND TSUTOMU KAGIYA* Department of Hydrocarbon Chemistry, Faculty of Engineering, Kyoto University, Sakyo-ku, Kyoto 606, Japan Received 17th December. 1984 Photoirradiation (Aex > 300 nm) of Ar-purged aqueous propan-2-01 solution gave hydrogen and acetone in the presence of platinum- and/or ruthenium dioxide-loaded TiO,. The photocatalytic activity of anatase TiO, depended significantly on the amount of metal or metal oxide present; the effect on the activity increased in the order platinum black + platinum powder > ruthenium dioxide. The photocatalytic activity of rutile TiO, was negligible even when loaded with platinum black.The effective wavelengths for the photocatalytic dehydro- genation of propan-2-01 were below ca. 390 nm, in agreement with the U.V. absorption spectrum of anatase TiO,. In a similar way primary, secondary and tertiary aliphatic alcohols underwent photocatalytic oxidation, accompanied by hydrogen liberation, by the platinized TiO,. The primary and secondary alcohols gave the corresponding carbonyl derivatives, while 2- methylpropan-2-01 and acetone gave dimeric products accompanied by stoichiometric hydrogen evolution. The initial rate of dehydrogenation in these photocatalytic systems was in proportion to the rate constants of hydrogen abstraction by hydroxyl radical in the homogeneous systems. A common feature of semiconductor photocatalysts is that they generate electron- hole pairs when irradiated with light of energy greater than their band gaps.l Although the importance of electrons and holes as the primary active species in photocatalytic reactions has been recognized, the mechanism of their action is still a subject of considerable controversy.Many authors have reported the photocatalytic reactions of alcohol^^-^ by platinized titanium dioxide (TiO,/Pt). They discussed the photocatalytic activity of TiO, and the effect of metal loading. However, few investigations3. lo have been reported on the reactivity of alcohols which are oxidized on the photoirradiated semiconductor particles. In the present work the photocatalytic reaction of a series of primary, secondary and tertiary aliphatic alcohols by platinized TiO, powder is reported.The correlation of the initial oxidation rate in aqueous suspension with the corresponding rate constant of hydrogen abstraction by hydroxyl radicals in a homogeneous system is discussed. EXPERIMENTAL MATERIALS Titanium dioxide powder (anatase > 99 % ; SO:-, 0.05% ; C1-, 0.0 1 % ) was supplied by Merck and used without further activation. Rutile TiO, powder was prepared by calcination of anatase at 1200 "C for 10 h in an electric furnace. This treatment transformed almost all the crystals into rutile (as determined by X-ray diffraction analysis, described elsewhere"). Platinum [Pt black (Nakarai Chemicals) or Pt powder (Higuchi Chemical laboratory), 5 wt %I or ruthenium 24672468 PHOTOCATALYSIS ON TiO, dioxide [RuO, (Nakarai Chemicals), 10 wt % I was mixed with TiO, powder in an agate mortar to prepare the ~atalyst.~? 12-15 The catalysts thus prepared showed reproducibility within 10% for the photocatalytic reaction.Water used in these experiments was purified by distillation after passage through an ion-exchange resin. Acetone and the aliphatic alcohols were used without further purification. PHOTOIRRADIATION The powdered catalyst (50 mg) was suspended in an aqueous solution (5.0cm3) of the substrate (0.50 mmol) under neutral or alkaline (6 mol dmP3 NaOH) conditions in a glass tube (15 mm in diameter and 180 mm in length, transparent to light of wavelength > 300 nm). The magnetically stirred solution under Ar was irradiated at > 300 nm using a merry-go-round apparatus equipped with a 400-W high-pressure mercury arc (Eiko-sha 400). MONOCHROMATIC PHOTOIRRADIATION In the quantum-yield measurements 30 mg of the platinized TiO, powder and 4.0 cm3 of 6 mol dm-3 NaOH solution were placed in a rectangular quartz cell (1 .O cm path-length) with a Pyrex finger (10 mm in diameter and ca.15 cm in length) and purged with Ar. The suspension was irradiated by a Philips SP 500 ultra-high-pressure mercury arc (operated at 750 W) through a grating monochromator (Jasco CT 25N). The number of photons incident upon the suspension through the monochromator and the quartz cell was evaluated by ferric oxalate actinometry.ls In the present system the formal quantum efficiency was evaluated as the ratio of twice the number of hydrogen atoms evolved to the number of incident photons, rather than the number of absorbed photons, because of the difficulty in measuring the exact number of photons absorbed by TiO, particles.Hence the true efficiency should be larger than the efficiency reported in fig. 1. The ultraviolet absorption spectrum of TiO, was recorded on a Hitachi EPS-3T spectro- photometer equipped with an R-1OA integrating sphere. PRODUCT ANALYSIS After irradiation the gas phase of the sealed sample was analysed by gas chromatography (g.c.)." The suspension was centrifuged to filter off the catalyst and the solution was subjected to product analysis by g.c. under conditions described previously.ll The amount of 2,5- dimethylhexane-2,5-diol present was also determined by g.c.with a Tenax GC column (3 mm in diameter and 1 m in length) at 200 "C. Formaldehyde was determined by a modified acetylacetone (pentane-2,4-dione) method. A portion of the aqueous sample containing formaldehyde was mixed with an aqueous solution of pentane-2,4-dione and ammonium chloride." The resulting yellow derivative, 1,4-dihydro-3,5-diacetyl-2,6-lutidine was extracted with chloroform and subjected to U.V. absorption measurements (A,,, z 396 nm) with a Shimadzu UV-200s spectrophotometer. RESULTS AND DISCUSSION PHOTOCATALYTIC ACTIVITY OF PLATINUM- AND/OR RUTHENIUM DIOXIDE-LOADED Table 1 shows the hydrogen yield for the photocatalytic reaction of propan-2-01 in the presence of platinum- and/or ruthenium dioxide-loaded TiO, suspension under deaerated conditions.Although the activities of both anatase and rutile TiO, without loaded metal were negligible in neutral and in alkaline solutions, a small amount of metal on the TiO, surface enhanced the activity of anatase, e.g. only 5 wt% of platinum black loading enhanced the activity by ca. 100-fold under both neutral and alkaline conditions. Pt powder and RuO, were also effective for photocatalytic hydrogen formation, but the effects on the activity enhancement were small compared with platinum black. These modified anatase systems also produced acetone, whose yield was almost equal to that of hydrogen and to the consumption of propan-2-01, showing that dehydrogenation of propan-2-01 proceeds photocatalytically [see also reaction (3) TiO, POWDERS FOR PROPAN-2-OL DEHYDROGENATIONS-I.NISHIMOTO, B. OHTANI AND T. KAGIYA 2469 Table 1. Photocatalytic dehydrogenation of propan-2-01 by aqueous suspensions of anatase and rutile loaded with Pt and/or RuO,~ in 6 mol dm-3 in H,O NaOH run catalystb 1 TiO,(A) 0.7 1.6 2.6 2 Ti02(A) + Pt powder (5 % ) 13 11 38 4 TiO,(A) + RuO, ( 10 % ) 8.9 9.0 30 6 TiO,(R) 0.1 0.6 1.3 7 TiO,(R)+Pt black (5%) 0.4 4.3 2.4 3 TiO,(A)+Pt black (5%) 93 79 200 5 TiO,(A)+Pt black (5%)+Ruo2 (10%) 130 132 210 a Catalyst (50 mg), NaOH aqueous solution (6 mol dm-3, 5.0 cm3) and propan-2-01 (38 mm3, 500 pmol) were irradiated with 500-W high-pressure mercury arc under Ar at room temperature. TiO,(A) and TiO,(R) refer to anatase and rutile. later].5* 14. l5 The enhanced activity of the TiO, loaded with platinum black is accounted for by the facilitated reduction of H+, or H,O under alkaline conditions, on the metal surface,18$19 which increases the efficiency of charge separation of the photogenerated electron (e-) - hole (h+) pair,l as is also the case for RuO,.~O The smaller enhancement by platinum powder may be due to its poor contact with TiO,, since the crystallite size of the platinum powder was of the same order of magnitude (> 200 nm) as that of anatase, while that of platinum black is ca.6 nm as evaluated from X-ray diffraction. On the other hand, the activity of rutile was negligible even when loaded with platinum black. Rutile with a particle size larger than that of anatase has been reported to show similar or even greater activity for the photodeposition of Ag metal compared with anatase.This is a strong indication that the efficiencies of anatase and rutile in photochemically producing electron-hole pairs are of the same order. Therefore, the small activity of rutile for H, formation is attributable to the disadvantageous energy of the photogenerated electron.'** 1 5 7 21 The effect of adding a base was evident in the present systems, as reported previously;22 a 1.5-3-fold increased amount of hydrogen was obtained compared with the neutral solution (see table 1). Although there is no structural evidence at present, the surface modification induced by the alkaline treatment may account for this enhancement [see reaction (l)]. An increase in the specific surface area and in the amount of surface hydroxyl groups accompanying the alkaline treatment has been reported :23 \ Ti4+-- OH- \ - Ti4+ NaOH / / \ \ \ / / / * 02- (1) 02- 02- +H,O Ti4+ - OH- - Ti4+ The surface hydroxyl is expected to be a hole-trapping site, as discussed in the next section.2470 1 .oo 0.75 x c a 0 .d 5 0.50 t l 5 c 5 0.25 0 PHOTOCATALYSIS ON TiO, \ \ \ 250 300 350 400 450 X/nm Fig.1. Variation in quantum efficiency of the photocatalytic dehydrogenation of propan-2-01 in a TiOJPt suspension under Ar as a function of the wavelength of monochromatic light. The broken line represents the absorption spectrum of the TiO, suspension. WAVELENGTH DEPENDENCE OF PROPAN-2-OL DEHYDROGENATION BY SUSPENDED PLATINIZED TiO, CATALYST Fig. 1 shows the dependence of the formal quantum efficiency of the photocatalytic dehydrogenation of propan-2-01 on the irradiation wavelength.The anatase powder loaded with platinum black (TiO,/Pt) was used as a photocatalyst and 6 mol dm-3 NaOH solution as solvent. The efficiency over the wavelength range 260-380 nm was apparently constant (0.6-0.8 % ) and decreased drastically at the longer wavelengths of ca. 380-400 nm. Photoirradiation at wavelengths > 400 nm produced negligible amounts of hydrogen. The observed action spectrum of anatase reasonably corresponds to the absorption spectrum shown in fig. 1 . These facts indicate that propan-2-01 dehydrogenation is initiated by the photoexcitation of an electron from the valence band into the conduction band with light of energy greater than the band gap [ca. 3.2 eV (corresponding to ca. 390 nm) for anatase Ti02].21 INFLUENCE OF PROPAN-2-OL CONCENTRATION The initial rate of H, formation [r(H,)] by the Ti02/Pt photocatalyst was measured as a function of the initial concentration of propan-2-01 (C).A sharp increase in r(H,) was observed on increasing C to 100 mmol dm-3. In the higher-C region r(H,) was practically constant (ca. 3.5 pmol h-l). Plots of these data according to a Langmuir adsorption isotherm, 1/r(H2) =fll/C") (n = 1 or 1/2), are shown in fig. 2. A linear relation was obtained for n = 1 (molecular adsorption) rather thann = 1 /2 (dissociative adsorption). Pichat and coworkers reported the dissociative adsorption of methanol onto the TiO, surface in the photocatalysis by platinized Ti0,,3 on the basis of a linear relation for n = 1 /2.Since the molecular adsorption of propan-2-01~~ is more probable in the present system, the amount of adsorbed propan-2-01 is thought to determine the rate of overall reaction [see reaction (3) later], as discussed in the following section. PHOTOCATALYSIS OF ALIPHATIC ALCOHOLS BY PLATINIZED TiO, Table 2 shows the results of the photocatalytic dehydrogenation of a series of aliphatic alcohols and acetone in the presence of TiO,/Pt under neutral conditions. These reactions could not be observed in the dark, nor in the absence of TiO,/Pt.S-I. NISHIMOTO, B. OHTANI AND T. KAGIYA 247 1 ( 1/C)/dm3 mol-' 50 100 150 I I I 2.5 5.0 7 . 5 10.0 12.5 ( 1 /C) ;/dm 9 mol-f Fig. 2. Linear transform of the Langmuir equation of the rate of H, formation [r(H2)] and the propan-2-olconcentration (C): [l/r(H2)] =A1/Cn) for n = 1 (0) and n = 1/2 (0).Irradiation was performed with a 500-W high-pressure mercury arc for 1.0 h. Table 2. Photocatalytic dehydrogenation of alcohols and acetone by platinized Ti02a substrate H2/pmol organic product/pmol ~~~~~ ~ ~~ CH,COCH, 3.3 (CH,COCH,-), 5.9 (CH,),COH 18 [(CH3)2CH(OH)CH 2-12 18 CH,OH 20 HCHO 14 CH2(OH)CH20H 16 HOCH,CHO 14 CH,CH,OH 20 CH,CHO 17 (CH,),CHOH 54 CH,COCH, 46 CH,COCH, 1.6 CH,CHO 5.3 The substrate (0.5 mmol) in water (5.0 cm3) was irradiated in the presence of platinized TiO, (50 mg) for 10 h under Ar at room temperature. All the alcohols except 2-methylpropan-2-01 gave carbonyl Similarly, ethylene glycol gave 2-hydroxyethanal. In the case of 2-methylpropan-2-01 a dimeric product 2,5-dimethylhe~ane-2,5-dioI,~~ was obtained together with small amounts of acetone and 2-methylbutan-2-01.~~ Acetone was less reactive and gave only a small amount of dimeric product, hexane-2,5-dione.Evolution of H, was observed simultaneously in each system. It is clear from table 2 that the H, yield is virtually equal to that of the carbonyl and dimeric products derived from these alcohols. Thus the overall photocatalytic reaction of alcohols at the initial stage can be represented by RCHzOH - RCHO+H, (2)2472 PHOTOCATALYSIS ON TiO, R' -CH-R2 * R'-CO-R2+H2 I OH OH OH OH I I I 1 _ . - + I (3) (4) I I * R2 - CCH2CH2C - R2 + H2 . I R' R1 2 CH3C -R2 R' I I The reactivity of the alcohols varies in the order secondary > primary > tertiary (table 2).This order agrees with that observed in the conventional oxidation of alcohols. In fig. 3 the initial rate of H, evolution for the alcohol solutions (0.1 mol dm-,) is plotted against the rate constant of hydrogen abstraction by hydroxyl radicals in aqueous The linear correlation in both neutral and alkaline solutions shows that the photocatalytic reactivity of the alcohols is similar to hydrogen abstraction in the dark. A possible mechanism for the dehydrogenation of the alcohols is as follows. In the presence of the Pt-loaded catalyst, reduction of H+ by the photoexcited electron to produce H, proceeds readily because of the small activation energy needed to release H, from the Pt surface. However, in the absence of reductants such as alcohols, recombination of the hole and the electron leads to a negligible liberation of H,.Thus the photocatalytic reaction strongly depends on the oxidation step induced by the photogenerated positive hole. A similar mechanism was also suggested by Dunn and coworker^.^^^ 29 The quantum efficiency can be approximately expressed in terms of the efficiency of the oxidation of the adsorbed alcohol (k,,[ROH]) and that of electron-hole recombination ( kd), ( 5 ) wherefdenotes the efficiency of photoexcitation (0 < f < 1). Under conditions where cf, + 1, i.e. k , B Kox[ROH], @ is approximately proportional to kox[ROH]. Furthermore, on the assumption that the effective concentration of alcohol ([ROH]) is identical (i.e. with alcohol the TiO, surface is almost saturated, see fig. 2), the rate of H, evolution, which is directly proportional to cf,, should be approximately proportional to the rate constant of oxidation of the alcohol (kox).Accordingly, the linear relations shown in fig. 3 strongly suggest that the alcohol oxidation proceeds via hydrogen abstraction from the alcohol molecule. Many researchers suggested the intermediacy of a hydroxyl radical adsorbed on the TiO, surface, which is produced by the reaction of a hole:,. 5 q 7 7 1 5 7 2 6 y 307 31 cf, = Jkox[ROHl/(ki + kox[ROHI) RH + 'OH,,, + 'R + H,O (6) [R = -CH,OH, CH,CHOH, (CH,),COH,-CH,C(CH,),OH, HOCH,CHOH I I I and -CH,COCH,]. In the case of the primary and secondary alcohols the hydroxyalkyl radicals undergo further oxidation via hydrogen abstraction to carbonyl derivatives : R1R2'COH + 'OH,,, -, R1R2C=0 + H,O.(7)S-I. NISHIMOTO, B. OHTANI AND T. KAGIYA 2473 20 10 0 0 5 10 15 rate constant of hydrogen abstraction by OH in aqueous solution/ 1 O8 dm3 mol-'s-' Fig. 3. Hydrogen evolution rate as a function of the rate constant of hydrogen abstraction by hydroxyl radical in aqueous The substrate (500 pmol) and Ti02/Pt (platinum black, 5 wt % , 50 mg) were placed in 5.0 cm3 of water (a) and 6 mol dm-3 NaOH solution (O), and irradiated by a 400-W high-pressure mercury arc for 2.5 and 0.5 h, respectively. ( I ) Acetone, (2) 2-methylpropan-2-01, (3) methanol, (4) ethylene glycol, ( 5 ) ethanol and (6) propan-2-01. Radicals produced from 2-methylpropan-2-01 and acetone, both of which bear no a-hydrogen, predominantly undergo recombination into dimeric derivatives.In the light of the results shown in fig. 3, the steps leading to the final products proceed more efficiently than the primary step, i.e. hydrogen abstraction. ( a ) A. J. Nozik, Annu. Reu. Phys. Chem., 1978, 29, 189; (b) M. S. Wrighton, Acc. Chem. Res., 1979, 12, 303; (c) A. J. Bard, J. Photochem., 1979, 10, 59; Science, 1980, 207, 139; J. Phys. Chem., 1982, 86, 172. T. Kawai and T. Sakata, J . Chem. SOC., Chem. Commun., 1980, 694. P. Pichat, J-M. Herrmann, J. Disdier, H. Courbon and M-N. Mozzanega, Nouu. J. Chim., 1981, 5, 627. S. Teratani, J. Nakamlchi, K. Taya and K. Tanaka, Bull. Chem. SOC. Jpn, 1982, 55, 1688. S. Nishimoto, B. Ohtani, H. Shirai and T. Kagiya, J . Polym. Sci., Polym. Lett. Ed., 1985, 23, 141. S. Nishimoto, B. Ohtani and T.Kagiya, in New Trends in Polymer Photochemistry, ed. N. S. Allen and J. F. Rabek (Applied Science Publishers, London, 1985), in press. F. H. Hussein and R. Rudham, J. Chem. SOC., Faraday Trans. I , 1984,80, 2817. S. Davidson, C. L. Morrison and J. Abraham, J. Photochem., 1984, 24, 27. S. Nishimoto, B. Ohtani, H. Kajiwara and T. Kagiya, J. Chem. SOC., Faraduy Trans. I, 1983,79,2685. ' H. Harada and T. Ueda, Now. J. Chim., 1984, 8, 123. l o T. Kawai, T. Sakata, K. Hashimoto and M. Kawai, Nippon Kagaku Kaishi, 1984, 277. l 2 T. Kawai and T. Sakata, Nature (London), 1980, 286, 474. l 3 S. Nishimoto, B. Ohtani, T. Yoshikawa and T. Kagiya, J. Am. Chem. SOC., 1983, 105, 7180. I * S. Nishimoto, B. Ohtani, A. Sakamoto and T. Kagiya, Nippon Kugaku Kaishi, 1984, 246. S. Nishimoto, B. Ohtani, H. Kajiwara and T. Kagiya, J . Chem. SOC., Furaday Trans. 1, 1985, 81, 61. C. G. Hachard and C. A. Parker, Proc. R. SOC. London A, 1956, 235, 518. ( a ) T. Nash, Biochem. J., 1953, 55,416; (b) Japan Industrial Standard, K 0102, 1974. I e S. Nakabayashi, A. Fujishima and K. Honda, Chem. Phys. Lett., 1983, 102, 464. It) J . Disdier, J-M. Herrmann and P. Pichat, J . Chem. Soc., Faraduy Trans. I , 1983, 79, 651. zo T. Sakata, K. Hashimoto and T. Kawai, J. Phys. Chem., 1984, 88, 5214. 81 FAR 12474 PHOTOCATALYSIS ON TiO, B. Kraeutler and A. J. Bard, J . Am. Chem. SOC., 1978, 100, 2239, 5985. 22 T. Kawai and T. Sakata, Chem. Lett., 1981, 81. 23 S. Okazaki and T. Kanto, Nippon Kagaku Kaishi, 1976, 404. 24 P. R. Harvey, R. Rudham and S. Ward, J . Chem. SOC., Faraday Trans. 1, 1983,79, 2975. 25 Sagami Chemical Research Center, Japan, p. 57 35525; Chem. Abst., 9: 38481~. S. Nishimoto, B. Ohtani, H. Shirai and T. Kagiya, J . Chem. SOC., Perkin Trans. 2, submitted for publication. 27 M. Anbar and P. Neta, Inr. J. Appl. Radial. Isot., 1967, 18, 493. 2R W. W. Dunn, Y. Aikawa and A. J. Bard, J . Am. Chem. Soc., 1981, 103, 3456. W. W. Dunn, Y. Aikawa and A. J. Bard, J. Electrochem. SOC., 1981, 128, 222. 30 R. B. Cundall, R. Rudham and M. S. Salim, J . Chem. SOC., Faraday Trans. I , 1976,72, 1642. 31 P. R. Harvey, R. Rudham and S. Ward, J . Chem. SOC., Faraday Trans. 1 , 1983, 79, 1381. (PAPER 4/21 38)

ISSN:0300-9599    

DOI:10.1039/F19858102467

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. SOC., Faradaj Trans. I , 1985, 81, 2475-2483 Kinetics of the Solvolysis of the trans-Dichloro bis( 1,2-diarninoe thane)cobal ~(III) I on in Water + t-Butyl Alcohol Mixtures BY GRAHAME S. GROVES AND CECIL F. WELLS* Department of Chemistry, University of Birmingham, Edgbaston, P.O. Box 363, Birmingham B15 2TT Receiued 18th December, 1984 The kinetics of the solvolysis of the 1,6-[Coen2C1,]+ ion in mixtures of water with t-butyl alcohol have been investigated for concentrations of the alcohol up to a mole fraction of 0.16 for a range of temperatures. Values for the enthalpy and entropy of activation show an extremum at the same mole fraction where the relative partial molar volume of t-butyl alcohol shows a minimum; both also show an inflection at the concentration of t-butyl alcohol where the mixtures have a maximum in their ultrasonic absorption. In general, plots of log(rate constant) against the reciprocal of dielectric constant are curves for the solvolysis of 1 ,6-[Coen2C1,]+ in water +cosolvent mixtures, and the application of a free-energy cycle to this solvolysis in water + t-butyl alcohol mixtures shows that changes in solvent structure have a greater effect on the pentacoordinated cobalt ion in the transition state than on the hexa- coordinated ion in the initial state.These results are compared with those for the solvolysis of other complex cations in water + cosolvent mixtures, particularly where the cosolvent is t-butyl alcohol. In our investigation of the kinetics of the solvolysis of the 1,6-[Coen2C1,]+ ion in water and in water+propan-2-01 mixtures we found' an extremum in the enthalpy and entropy of activation in the same region of solvent composition where extrema occur in the physical properties associated with critical changes in solvent structure in the mixture.Moreover, a plot of log(rate constant) against the reciprocal of the dielectric constant did not provide1 the linear plot expected from a consideration of the solvolysis on the basis of point charges immersed in a dielectric c ~ n t i n u u m . ~ - ~ For such a solvolysis in water + cosolvent mixtures of a complex Czc with essentially full extension of the MZ~*-.XZx bond in the transition state, the Laidler-Landskroener equation2 can be applied in extended form5 involving the introduction of free energies of transfer of the ionic species from water into the mixture as in the eqn + AG,"(MZ~), + AG,"(X"x), - AG,"(C"c),.(1) k is the first-order rate constant, N is Avogadro's number, e is the electronic charge, D is the dielectric constant, r is the radius and G is related to the dipole moment. AG,O(i), is the free energy of transfer of species i between water (subscript w) and the mixture (subscript s), excluding all the electrostatic effects included in the first term on the right-hand side of eqn (1). For cases like 1 ,6-[Coen,C12]+ in water + propan-2-01 mixtures,' where a plot of log k against 0;' is not linear, the following equation holds: (2) AG,"(C"c) # AG,"(MZ~), + AG,"(XZx),. 2475 81-22476 SOLVOLYSIS OF tranS-DICHLOROBIS( 1,2-DIAMINOETHANE)COBALT(III) This is also the case for the solvolysis of l,6-[Copy4C1,]+ with methanol, propan-2-01, t-butyl alcohol or dimethyl-sulphoxide as cosolvent (py = pyridine)6 for (i) the solvolysis of 1 ,6-[Co(4Mepy),C12]+ with propan-2-01 or t-butyl alcohol as cosolvent (4Mepy = 4-methylpyridine),8 (ii) for the solvolysis of 1,6-[Coen2N,C1]+ and 1,2- [Coen,SCNCl] with propan-2-01 as cosolvent8T and (iii) for the solvolysis of 1,2-[Coen2N,C1]+ with t-butyl alcohol as cosolvent.lo Where linear plots of log k against 0;' are obtained independent of the cosolvent employed5 the equation AG:(C"c), = AG:(MZ~), + AG:(XZx), (3) must apply, and for cases where plots of log k against 0;' are linear for each cosolvent but not coincident5 deviations from eqn (3) towards eqn (2) begin to occur.Examples of the latter are the solvolysis of 1,6-[Coen,N3C1]+ with acetone, ethanol and ethylene glycol as cosolvents5 and the solvolysis of 1,2-[Coen,N3C1]+ with the latter cosolvents and with propan-2-01 as coso1vent.ll The application of eqn (1) or of a thermodynamic cycle5 to free energies, such as AG; [ML,YX]$--,[ML,Y]$ + XZ,x -AGP(ML,YXZC) AC:(ML,Yzr + zx) AG;(XZX) 1 AGZ 1 1 [ML,YX],Zc+[ML,Y]fc + zx + Xfx where AG* is the free energy of activation and AG,"(i) is the total free energy of transfer of i from water into the mixture, is only possible where the extension of the M Z ~ . . .XZx bond is virtually complete in the transition state. The comparison12 of volumes of activation, AV*, with overall volumes of reaction, AVO, and the constancy13 of the ratio of stereochemical forms in the products for varying Xzx show that this assumption is correct for complexes like [Coen,Cl,]+, where M z ~ = Co3+ and the ligands L involve attachment to Co3+ through a total of four nitrogen atoms.The equation 2.303 RTlogk\*-AG:(XZx) = AG,"(ML, YZ~+Z~)-AG:(ML,YXZc) (4) ks can be deduced5 from the above cycle, and in all the cases for which it has been applied involving solvolysis of complexes with M Z ~ = Co3+ the left-hand side has been found to be negati~e.~-'l As AGy(cation) is normally negative for water + cosolvent m i x t u r e ~ , ~ ~ ~ l5 owing to the influence of changes in solvent structure, these negative plots show that the influence of such changes on solvolysis reactions with M z ~ = Co3+ usually dominates on the cation in the transition state rather than on the cation in the initial If it is possible to determine values for AG,O(ML,YXzc), values for AG:(ML,YZc+Zx) can be determined for the pentacoordinated cation in the transition state.'.l6 As additions of t-butyl alcohol to water induce greater changes in the solvent structure at lower mole fractions of cosolvent than for additions of propan-2-01, the kinetics of the solvolysis of 1 ,6-[Coen2C1,]+ has been investigated in water + t-butyl alcohol mixtures. EXPERIMENTAL The materials used were as described earlier. t-Butyl alcohol was fractionally distilled. Rates were measured using spectrophotometry, as described previously.G. S. GROVES AND C. F. WELLS 2477 Table 1. Values for the first-order rate constant k/10-? s-' for the solvolysis of 1,6-[Coen,Cl,]+ in water + t-butyl alcohol mixtures [t-butyl alcohol] T/"C mole wt% fraction 35.0 45.0 50.0 55.0 1.97 3.95 7.99 12.13 16.35 20.68 25.10 29.63 34.27 39.0 1 43.88 0.0049 0.010 0.02 1 0.033 0.045 0.060 0.075 0.093 0.113 0.135 0.160 1.67 1.51 1.22 1.12 1.19 1.07 0.99 1 .oo 1.03 0.97 0.99 6.1 5.7 4.78 4.0 1 4.28 3.87 3.8 1 3.76 3.55 3.71 3.60 12.3 12.5 9.8 9.0 8.7 8.4 8.1 7.4 6.9 6.8 6.8 22.4 20.0 17.7 16.0 17.4 15.2 14.0 13.0 13.3 12.8 10.5 RESULTS AND DISCUSSION VARIATION OF THE RATE CONSTANT WITH COMPOSITION AND WITH TEMPERATURE As with the investigation of the kinetics of the solvolysis in mixtures of water with propan-2-01,~ the increase in optical density at 525 nm was followed with time using an initial concentration of 1 .O x lop2 mol dm-3 of the complex in solutions which were 0.020 mol dm-3 in perchloric acid: the reasons for using these latter conditions have been discussed ear1ier.l The concentrations of t-butyl alcohol varied from 1.97 wt% up to 43.88 wt% for the temperatures 35, 45, 50 and 55 "C.In all cases plots of log (O.D., -O.D.,) against time were linear (where O.D. is the optical density) and values for the first-order rate constant were determined from the slopes. In all cases a minimum of two rate measurements were made for each concentration of t-butyl alcohol at each temperature. The mean of these determinations are given in table 1 : for over 80% of the determinations the deviation of individual values of k from the mean never exceeded ca. 4%. Plots of log k against the reciprocal of the absolute temperature are linear for all solvent compositions.Values for the enthalpy, AH*, and entropy, AS*, of activation were determined by the application of the least-squares procedure to all the individual values of k using a computer, as has been done for all our investigations of the kinetics of solvolysis reactions.lV6-lf These values, together with those for the free energy of activation at 25 "C, all with standard errors, are given in table 2. Fig. 1 shows that plots of log k against D;l at 50 and 25 "C are curves. Values for D, at 25 "C were obtained by interpolation of the data of Broadwater and Kay1' and of Brown and Ives,18 and at 50 "C the values of %ikerloPg were used. Values for k were calculated using the values of AH* and AS* in table 2.Values for the Grunwald-Winstein Y factor20 have been calculated from the available kinetic data for the solvolysis of t-butyl chloride in water + t-butyl alcohol mixtures.21 These values are given in table 3. When log k at 25 "C is plotted against these Y values (fig. 2) a curve is obtained. Such a curve was also obtained for the plot of log k against Y in water + propan-2-01 mixtures' for Y values calculated using data2478 SOLVOLYSIS OF lTaHS-DICHLOROBIS( 1,2-DIAMINOETHANE)COBALT( 111) 1.0 0.8,- - -Y MI d + T Oe6: 0.1 Table 2. Values for the transition-state parameters for the solvolysis of 1 ,6-[Coen2C1,]+ in water + t-butyl alcohol mixtures i, 0 0 1 I I , I ~~~~ ~~~ ~ ~ ~ [t-butyl alcohol] AS* at 25.0 "C AG* at 25 "C wt% mole fraction AH*/kJ mol-' /J K-' mo1-I /kJ mol-l 1.97 3.95 7.99 12.13 16.35 20.68 25.10 29.63 34.27 39.01 43.88 0.0049 0.010 0.02 1 0.033 0.045 0.060 0.075 0.093 0.1 13 0.135 0.160 108+ 1 108 +_ 5 110+1 111f3 110+2 109 f 2 110f6 104f2 105f3 105f2 99k3 31 f 4 32+ 17 38+4 38f9 36+6 32f8 37+ 18 17f7 20+7 19f7 o* 11 9 8 f 3 9828 99+2 99k6 9 9 f 4 99+5 99f11 99+4 99+6 99+4 99&7 from the same source, whereas Y values obtained using the alternative kinetic data available in water + propan-2-01 for the solvolysis of t-butyl chloride gave a linear plot with log k for the solvolysis of 1,6-[Coen2C12]+ in these mixtures.l Non-linear plots are usually found for the solvolysis of similar complexes6q ' 9 lo in water + t-butyl alcohol mixtures when the Y values in table 3 are used.G.S. GROVES AND C. F. WELLS 2479 Table 3. Values for the Grunwald-Winstein Y factor calculated for water + t-butyl alcohol mixtures at 25 "C t-butyl alcohol (mole fraction) Y 0 3.50 0.020 3.34 0.050 2.92 0.10 1.88 0.20 0.95 0.6 -Y 00 - + wl 0.L 0.21 I I I 1 2 3 4 Y Fig. 2. Plot of log k for the solvolysis of 1,6-[Coen,Cl,]+ at 25 "C in water+ t-butyl alcohol mixtures against the Grunwald-Winstein Y factor. INFLUENCE OF SOLVENT STRUCTURE ON THE SOLVOLYSIS It has been shown earlier' that a plot of log k against D;' in water + propan-2-01 mixtures for the solvolysis of the 1,6-[Coen2C1,]+ ion is curved, as found here for the same solvolysis in water + t-butyl alcohol mixtures. Moreover, similar curves are also found when log k against 0;' is plotted5.22 for this solvolysis using the kinetic data available at one temperature over a reasonable range of solvent compositions for cosolvents methan01,,~-,~ ethan01,,~-,~ ethanol- 1,2-di01,,~ a~etone,,~-,~ d i ~ x a n e , ~ ~ 24 and ethanonitrile,' and for the solids sucrose26 and ethylene carbonate28 each individually mixed with water. It must therefore be concluded that this solvolysis does not conform to the linear relationship between log k and D;' on the basis of a consideration of point charges in a dielectric continuum. The change in the discrete molecular structure of the solvent must therefore play a part, with eqn (2) holding as expected from the marked deviation of AG,"(i) (where i is a cation) for a wide range of cations dissolved in mixed solvents formed by adding cosolvents to waterl4.l5 from the positive values expected for the transfer of a cation from water into a mixture of lower dielectric constant.In fig. 3 the enthalpy of activation for the solvolysis is plotted against the mole fraction of t-butyl alcohol. This shows a maximum in the region of x, = 0.025 and a plateau at x, z 0.10 followed by a further decrease in AH* at x, > 0.12. Fig. 4 shows that a similar maximum and plateau occur in the same regions of x, for the variation of the entropy of activation with solvent composition. The maximum at x, z 0.0252480 SOLVOLYSIS OF ?ranS-DICHLOROBIS( 1,2-DIAMINOETHANE)COBALT(III) I I i i 0.1 0 0.1 5 0.2 0 951 0.0 5 x2 Fig. 3. Plot of the enthalpy of activation for the solvolysis of 1 ,6-[Coen2C1,]+ in water + t-butyl alcohol mixtures against mole fraction of t-butyl alcohol, x,.x2 Fig. 4. Plot of the entropy of activation at 25 "C for the solvolysis of 1,6-[Coen,Cl,]+ in water + t-butyl alcohol mixtures against mole fraction of t-butyl alcohol, x,.G. S . GROVES AND C . F. WELLS 248 1 -101 I I 1 I 0.0 5 0.1 0 0.1 5 0.20 x2 Fig. 5. Plot of the left-hand side of eqn (4) for the solvolysis of 1 ,6-[Coen2C1,]+ in water + t-butyl alcohol mixtures against mole fraction of t-butyl alcohol, x2. correlates well with the composition at which the sharp, deep minimum in the relative partial molar volume of t-butyl alcohol, - V i , O C C U ~ S ~ ~ ~ 30 in its mixtures with water, and the plateau at x2 x 0.1 correlates well with the composition where the sharp, high maximum occurs in the ultrasonic absorption of these Although the minimum in the excess enthalpy of mixing, AHE, for water and t-butyl alcohol mixtures occurs at mole fractions x, < 0.025, the minimum in the excess entropy of mixing, ASE, occurs in the region of x, sz 0.025, where the maxima in AH* and AS* occur.As found with the solvolysis of other complexes, notably 1,6-[C0py,Cl,]+~ and 1,2-[C0en,N,Cl]+,~~~ l1 the extrema in AH* and AS* move to lower values of x, when the cosolvent changes from propan-2-01 to t-butyl alcohol, correlating closely with the movement of the minimum in - V; for these c o ~ o l v e n t ~ . ~ ~ ~ 30 The movement and relative size of the extrema in - V20299 30 and in the ultrasonic a b ~ o r p t i o n , ~ ~ together with the values of the excess change in the temperature of the maximum density of water with added cos01vent~~ and the movement in the e ~ t r e m a ~ ~ v 33 in AHE and ASE with varying cosolvent, show that t-butyl alcohol induces more structure at lower values of x, than propan-2-01 when added to water.The extremum in - V i has been associated5 with the situation where the exclusion of the alkyl groups of the cosolvent outside the ‘flickering iceberg’ of water produces maximum strain on the latter including maximum hydrogen-bond formation within the cluster and the extremum in the ultrasonic absorption at higher x, has been associated with the onset of the breakdown of the ice-like clusters by the alkyl groups resulting eventually at much higher x, in the predominance of the structure of the pure cosolvent.As for all the solvolyses of similar complexes of ColI1 in water + cosolvent mixtures,6-11 a plot of AH* against AS* is linear for 1 ,6-[Coen2C1,]+ in water + t-butyl alcohol mixtures. As the comparison of AV* and AVO and the stereochemical ratio of products for2482 SOLVOLYSIS OF ~uuzS-DICHLOROBIS( 1,2-DIAMINOETHANE)COBALT(III) the solvolysis of this type of complex ci?tionl27l3 show that the Co3+.-C1- bond is completely broken in the transition state, the free-energy cycle discussed above can be applied in this case. This involves using eqn (4) at 25 "C with M Z ~ = Co3+ and Xz.u = C1-. Values for k at 25 "C are calculated using the values of AH* and AS* in table 2, and values for AGP(Cl-) are available for water + t-butyl alcohol mixtures14 with all corrections a~p1ied.l~ Values for the left-hand side of eqn (4) are plotted against mole fraction of t-butyl alcohol in fig.5 for x, < 0.16. Under these conditions the left-hand side of eqn (4) is always negative, as found for similar plots of eqn (4) for the solvolysis of a range of complexes in mixtures of water with a cosolvent for MZb1 = C03+.1,5-11 As values of AG:(i) where i is a cation are negative at low x, in water+t-butyl alcohol m i x t u r e ~ , ~ ~ . ~ ~ it can be concluded from fig. 5 that - AG:(Coen,C12+) > - AG:(Coen,Cl:) for x, < 0.16. Therefore, changes in solvent structure in water + t-butyl alcohol have a greater influence on the pentacoordinated cation, [Coen,Cl]*+, in the transition state than on the hexacoordinated cation, [Coen,Cl,]+, in the initial state; this is as found for these complexes in water + propan- 2-01 and is comparable with the conclusion reached for similarly related pairs of complexes having M Z ~ = Co3+ in water+cosolvent mixtures.This conclusion is also in accord with the general observation for simple cations of approximately the same size that AG:(i) is more negative for the higher overall positive charge.'? l4, l5 G. S. Groves and C. F. Wells, J . Chem. SOC., Faraday Trans. I, 1982, 78, 619. K. J. Laidler and H. Eyring, Ann. N.Y. Acad. Sci., 1940, 39, 303; S. Glasstone, K. J. Laidler and H. Eyring, The Theory of Rate Proceses McGraw-Hill, New York, 1941), chap. 8; K. J. Laidler and P. Landskroener, Trans. Faraday Soc., 1956, 52, 200; K.J. Laidler, Suom. Kemistil. A, 1960,33,44; K. J. Laidler, Chemical Kinetics (McGraw-Hill, New York, 2nd edn, 1965), chap. 5. E. A. Moelwyn-Hughes, Proc. R. Soc. London, Ser. A, 1936, 155, 308; 1936, 157, 667; The Kinetics of Reactions in Solution (Oxford Univ. Press, London, 2nd edn, 1947), chap. 4, 5 and 7; Physical Chemistry (Pergamon Press, Oxford, 2nd edn, 1961), chap. 7-9, 24. E. S. Amis, Kinetics of Chemical Change in Solution (Macmillan, New York, 1949), chap. 5 and 9; Solvent EHects on Reaction Rates and Mechanisms (Academic Press, New York, 1966), chap. 1-3; Solvent Eflects on Chemical Phenomena (Academic Press, New York, 1973), vol. 1, chap. 5. C. F. Wells, J . Chem. Soc., Faraday Trans I, 1977, 73, 185. C. N. Elgy and C. F. Wells, J. Chem. Soc., Dalton Trans., 1980, 2405; 1982, 1617; J .Chem. SOC., Faraday Trans. I , 1983, 79, 2367 and in press (paper 4/1981). I. M. Sidahmed and C. F. Wells, J . Chem. SOC., Dalton Trans., 1983, 1035; 1984, 1969. A. E. Eid and C. F. Wells, J . Chem. SOC., Faraday Trans. 1, 1983, 79, 253. A. E. Eid and C. F. Wells, J . Chem. Soc., Faraday Trans. 1, 1985, 81, paper 4/1226. lo A. E. Eid and C. F. Wells, Trans. Metal Chem., in press. l 1 A. E. Eid and C. F. Wells, J. Chem. SOC., Faraday Trans. 1, 1981, 77, 1621. l 2 W. E. Jones, L. R. Carey and J. W. Swaddle, Can. J . Chem., 1972,56, 2739; G. A. Lawrance, Inorg. Chim. Acta, 1980, 45, L275; G. A. Lawrance and S. Suvachittanont, Austr. J . Chem., 1980, 33, 277; D. A. Palmer and H. Kelm, Inorg. Chem., 1977.16,3 139; Coord. Chem. Rec., 1981,36,89; G.Daffner, D. A. Palmer and H. Kelm, Inorg. Chim. Acta, 1982, 61, 57. l 3 W. G. Jackson and A. M. Sargeson, Inorg. Chem., 1978, 17, 1348; W. G. Jackson and C. M. Begbie, Inorg. Chim. Acta, 1982, 60, 1 1 5. l 4 C.F. Wells, J . Chem. Soc.. Faraday Truns. 1, 1973, 69, 984; 1974. 70, 694; 1975, 71, 1868; 1976, 72, 601; 1978,74,636, 1569; 1981,77, 1515; 1984,80,2445; Adz>. Chem. Ser., 1979,177, 53; Thermochim. Acta, 1982, 53, 67; G. S. Groves and C. F. Wells, J . Chem. Soc., Faraday Trans. 1, in press (papers 4/1972; 5/365). G. S. Groves, I. M. Sidahmed and C. F. Wells, unpublished results. C. F. Wells, Ausrr. J . Chem., 1983, 36, 1739. l6 I. M. Sidahmed and C. F. Wells, J . Chem. Soc., Dalton Trans., 1981, 2034. l 7 T. L. Broadwater and R. L. Kay, J . Phys. Chem., 1970, 74, 3802. l 8 A. C. Brown and D. J. G. Ives. J . Chrm. Soc., 1962, 1608. l9 G. Akerlof, J . Am. Chem. Soc., 1932, 54, 4125. ** E. Grunwald and S. Winstein, J . Am. Chem. Soc., 1948, 70, 846; S. Winstein, E. Grunwald and H. W. Jones, J . Am. Chem. SOC., 1951, 73, 2700. * l R. E. Robertson and S. E. Sugamori, J . Am. Chem. SOC., 1969. 91. 7254.G. S. GROVES AND C. F. WELLS 2483 2 2 G. S. Groves, Ph.D. Thesis (University of Birmingham, 1983). 23 V. N. Vasileva and K. B. Yatsimirskii, Russ. J. Inorg. Chem., 1962, 7 , 1307. 24 M. Pribanic, M. BiruS, D. Pavlovid and S. ASperger, J. Chem. Soc., Dalton Trans., 1973, 2518. 25 V. D. Panasyuk and E. R. Falendysh, Dopo. Akad. Nauk. Ukr. RSR, 1964, 741. 26 V. D. Panasyuk and A. V. Arkharov, Ukr. Khim. Zhur., 1966, 32, 716. 27 J. Burgess, J. Chem. Soc. A, 1970, 2351. 28 M. J. Blandamer, J. Burgess, S. J. Hamshere and F. M. Mekhail, Trans. Metal Chem., 1979, 4, 77. 29 J. Kenttamaa, E. Tommila and M. Martti, Ann. Acad. Scient. Fenn., 1959, no. 93. 3o K. Nakanishi, Bull. Chem. SOC. Jpn, 1960, 33, 793. 31 M. J. Blandamer, Introduction 10 Chemical Ultrasonics (Academic Press, London, 1973), chap. 11. 32 G. Wada and S. Umeda, Bull. Chem. Soc. Jpn, 1962, 35, 646. 33 R. F. Lama and B. C-Y. Lu, J. Chem. Eng. Data, 1965, 10, 216. 34 H. S. Frank and M. W. Evans, J. Chem. Phys., 1945, 13, 507; H. S. Frank and W-Y. Wen, Discuss. Faraday Soc., 1957, 24, 133; G. Nemethy and H. A. Sheraga, J. Chem. Phys., 1962,36, 3382, 3401; N. Laiden and G. Nemethy, J. Phys. Chem., 1970,74, 3501. (PAPER 4/2 145)

ISSN:0300-9599    

DOI:10.1039/F19858102475

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. I, 1985,81, 2485-2493 Catalysis by Amorphous Metal Alloys Part 2.-Effects of Oxygen Pretreatment on the Catalytic Activity of Amorphous and Crystallised Ni-P Alloys BY HIROMI YAMASHITA, MASAHITO YOSHIKAWA, TAKUZO FUNABIKI* AND SATOHIRO YOSHIDA Department of Hydrocarbon Chemistry and Division of Molecular Engineering, Faculty of Engineering, Kyoto University, Kyoto, Japan Received 19th December, I984 The effects of pretreatments for activation of amorphous Ni-P alloy, especially the role of treatment with oxygen, have been studied by estimating the numbers of surface nickels by temperature-programmed reduction of the oxidised species and recording the ESCA spectra of the surface layers. Oxygen treatment followed by reduction with hydrogen creates surface nickels that interact with phosphorus and oxidised nickel and phosphorus species.The number of surface nickels and the strength of the interaction increases with the temperature of the oxygen treatment and are correlated with the catalytic activity. The decrease in the catalytic activity caused by over-oxidation and crystallisation is due to the disappearance of the homogeneity of the amorphous alloy and reconstruction of the structure involving the aggregation of nickel species. In studies of the catalytic properties of amorphous m e t a l ~ l - ~ we have found that amorphous Ni-P and Ni-B alloys are more active than crystallised alloys for the hydrogenation of olefins, and that it is necessary to pretreat the alloys with acid, oxygen and hydrogen in order to generate the catalytic activity.We have assumed that the formation of oxidised species of nickel and metalloids forms electron-deficient nickel species and that the catalytic activity is dependent on the surface concentration of nickel species in the different electronic states. In order to study the effects of the pretreatment, especially oxygen treatment, on the catalytic activity it is important to know how the number of surface nickels, the turnover frequency of the catalytic reaction and the electronic and structural states of the surface components change after the pretreatment. We have studied the effects of varying the temperature of the pretreatment by measuring the catalytic activity for the hydrogenation of ethene and buta- 1,3-diene, by measuring the chemisorption of carbon monoxide and hydrogen, by the temperature-programmed reduction (t.p.r.) of the oxidised species and by recording the ESCA spectra. EXPERIMENTAL HYDROGENATION REACTION The Ni-P (Ni81P19) amorphous alloy, in the form of ribbons ca.2 mm wide and 10-20 pm thick, was prepared by the rapid quenching method4 using a single steel roll. The crystallised alloy was prepared by treating the amorphous alloy at 723 K for 2 h under vacuum. The alloys were pretreated with dilute HNO,, oxygen and hydrogen prior to use for hydrogenation as reported previo~sly.l-~ Thus, the Ni-P alloys were treated with 6 mol dm-, HNO, for 10 min, washed thoroughly with distilled water and dried in air at room temperature for 24 h. The alloys 248 52486 CATALYSIS BY AMORPHOUS METAL ALLOYS were then placed in the reaction vessel and treated with 6.7 kPa oxygen for 1 h at different temperatures and with 13.3 kPa hydrogen at 573 K for 2 h.Hydrogen (19.3 kPa), purified by a hydrogen diffusion purifier (Japan Pure Hydrogen Co., LS-O9B), and commercial ethene and buta-l,3-diene (7.3 kPa) were introduced into the reaction vessel (300 cm3), which was connected to a conventional closed-circulation system. The reaction was performed at 373 K and the initial rates were estimated from the initial pressure change. THE CHEMISORPTION OF CARBON MONOXIDE AND HYDROGEN The chemisorptions of carbon monoxide and hydrogen were measured using a corrected Pirani vacuum gauge (ULVAC GP-2T) in the pressure range 13-667 Pa. The number of surface nickels was estimated by assuming an adsorption equilibrium according to the Langmuir is0 therm where N is the number of the sites occupied by an adsorbed carbon monoxide molecule or hydrogen atom at a pressure P Pa, N, is the total number of sites and K is the equilibrium constant for The assumptions of the single-site adsorption of carbon monoxide and dissociative adsorption of hydrogen were supported by good linear Langmuir plots and the absence of substantial dependence of the values on the kind of absorbate.The probability of reactions of the adsorbate molecule at room temperature with the reactive oxygen species on the surface, which might be formed in the oxygen treatment, was regarded as negligible since the species were probably removed in the treatment with 13.3 kPa hydrogen at 573 K and evacuation.P(N,--N) = KN ESCA SPECTRA ESCA spectra were recorded with Shimadzu ESCA-750 using Mg standard radiation (10 kV and 30 rilA) as described previously.2 After pretreatment, the alloys were placed in an ESCA analyser chamber under Ar atmosphere and the spectra were recorded either with or without sputtering the surface with Ar+. All binding-energy values (accuracy & 0.3 eV) were referred to the value of the contaminant carbon [C( 1s) = 285.0 eV] for convenience. TEMPERATURE-PROGRAMMED REDUCTION The t.p.r. studies of the alloys were carried out in a flow system similar to that described The reducing gas, 30 % hydrogen in helium, was passed through molecular sieves at 77 K and at a rate of 43 cm3 min-l and then split between the reactor and reference columns.The alloys in the reactor were pretreated with HNO, and oxygen, cooled in oxygen and evacuated for 20 min before observation of the t.p.r. After the steady state was attained at room temperature the reactor was heated at a constant rate of 17 "C min-l and the reduction of the oxidised species of the alloys was monitored by the thermal-conductivity cell. RESULTS Fig. 1 shows the effect of the temperature of the treatment with oxygen in the successive pretreatments on the catalytic activities for the hydrogenation of ethene and buta- 1,3-diene over amorphous and crystallised Ni-P alloys. Treatment with oxygen at fairly, high temperatures (> 473 K) was necessary for activation of the alloy and the maximum activity was observed at 573 K in the case of the amorphous alloy.We have reported previously2 that the activity for hydrogenation of buta- 1,3-diene decreases at > 513 K, but prolonged evacuation after each reaction resulted in an increase in the activity at > 5 13 K, as shown in fig. 1. In the case of the crystallised alloy, the activity increased with the treatment temperature over the range of temperatures studied. The amorphous alloy was more active than the crystallised alloy for both reactions at < 573 K. Fig. 2 shows the dependence of the number of surface nickels on the temperature of the pretreatment with oxygen. Substantial differences were not observed betweenH. YAMASHITA, M. YOSHIKAWA, T. FUNABIKI AND S. YOSHIDA 2487 1 0 4 5 0 500 550 600 A 4 4 5 0 500 550 600 TIK Fig.1. Effect of oxygen pretreatment of the Ni-P alloy on the initial rate of hydrogenation of ethene and buta-1,3-diene. Initial pressures PH2 = 19.3 kPa and Polefin = 7.3 kPa; reaction temperature 373 K. The alloys were pretreated with dilute HNO,, 6.7 kPa oxygen for 1 h and 13.3 kPa hydrogen at 573 K for 2 h. Hydrogenation of ethene: 0, amorphous and 0, crystallised. Hydrogenation of buta- 1,3-diene: A, amorphous and A, crystallised. - 0 5 0.4 --. .: 0 . 3 z 0 TI K Fig. 2. Effect of the temperature of the oxygen pretreatment on number of surface nickels. Estimated from hydrogen chemisorption : 0, amorphous and A, crystallised and carbon monoxide chemisorption: 0, amorphous and A, crystallised. The values at A correspond to those measured without pretreatment.Pretreatment conditions as shown in fig. 1.2488 CATALYSIS BY AMORPHOUS METAL ALLOYS Table 1. Effect of the temperature of the oxygen pretreatment on the turnover frequencies for the hydrogenation of ethene and buta-l,3-diene over amorphous and crystallised Ni-P alloys" turnover frequency/s-l hydrogenation of hydrogenation of ethene buta-l,3-diene T/K amor. cryst. amor. cryst. 473 1.7 1.9 1 . 6 ~ 10 1.0 513 8.1 x 10 1.7 1.1 x lo2 1.8 543 3.1 x lo2 2 . 0 ~ lo2 1.5 x lo2 4.1 x 10 573 1.1 x lo3 1.5 x lo2 2.3 x lo2 9.3 x 10 4.7 x 102 - 2.2 x 102 623 b b - a Initial pressures pH2 = 19.3 kPa and Polefin = 7.3 kPa; reaction temperature 373 K. Pre- The alloy is not regarded as amorphous at this treatment conditions as shown in fig. 1. treatment temperature.' Fig.3. Temperature-programmed reduction of the amorphous Ni-P alloy after the oxygen treatment. Temperature of the oxygen treatment: 0, 448; A, 503; 0, 563 and T7, 593 K. values estimated from the chemisorption of carbon monoxide and hydrogen. The number of surface nickels increased gradually to a maximum at 5 13 and 573 K with the amorphous and crystallised alloys, respectively. The crystallised alloy required a higher temperature to obtain the same number of nickels as the amorphous alloy. The turnover frequency (TF) for the surface nickel atom in the hydrogenation ofH. YAMASHITA, M. YOSHIKAWA, T. FUNABIKI AND S. YOSHIDA 2489 ! I I 8 6 0 850 I I I I 1 4 0 130 binding energy/eV Fig. 4. ESCA spectra of the amorphous Ni-P alloy pretreated with oxygen at different temperatures for 1 h: (a) 423, (b) 513 and (c) 573 K.Spectra were measured after reduction with hydrogen at 573 K for 2 h. Scale of peak intensity (xcps) for Ni(2p3,,): (a) 1000, (b) 2000 and (c) 500; for P(2p): (a)-(c) 200. olefins was calculated using the values of the catalytic activity in fig. I and the averaged number of specific surface nickels in fig. 2. As shown in table 1, the TF for the hydrogenation of olefins was affected by the oxygen treatment and increased with the temperature of the oxygen treatment. The amorphous alloy showed greater TF values than the crystallised alloy, and the values for ethene was greater than those for buta-l,3-diene. Fig. 3 shows the effect of the temperature of the treatment with oxygen on the t.p.r. curves of the amorphous alloy.When the treatment temperature was low, distinct t.p.r. curves were not observed. When the treatment temperature was 503 K, a curve with a maximum at 503 K was obtained. When the treatment temperatures were 563 and 593 K, two maxima at 503 and 528 K and one at 523 K were observed, respectively. Increasing the treatment temperature brought about not only a shift of the maximum, but also an increase in the peak area, indicating that the oxidised species become more stable at the higher treatment temperature. ESCA spectra provide information about changes in the surface layers brought about by the pretreatment. Fig. 4 shows the spectra of the amorphous alloy treated with oxygen at different temperatures, followed by reduction with hydrogen. The alloy pretreated with oxygen at 423 K, which showed no activity for hydrogenation, gave a strong peak of an oxidised species of phosphorus (abbreviated as phosphorus oxide, 133.9 eV) but a very small peak of an oxidised species of nickel (abbreviated as nickel oxide, 856.9 eV).The relative intensity of the peak of the nickel oxide to that of nickel (853.4 eV) increased with the pretreatment temperature, but that of phosphorus oxide to phosphorus (129.6 eV) was little altered. At 573 K, where the maximum activity was observed, the peaks of the phosphorus species became small, indicating a decrease in the total concentration of the phosphorus species. The Ni(2p3,,) peak of nickel2490 CATALYSIS BY AMORPHOUS METAL ALLOYS Table 2. Effect of the temperature of the oxygen pretreatment on the nickel and phosphorus species in the surface layers estimated from the ESCA spectraa 423 0.26 0.2 1 0.52 0.8 1 0.49 0.52 513 0.3 1 0.23 0.18 0.43 0.52 0.48 573 0.33 0.48 0.13 0.33 0.45 0.40 a I N i , INiPo, Ip and Ip-o denote the intensities of the ESCA peaks of nickel, nickel oxide, phosphorus and phosphorus oxide, respectively.The values were obtained by dividing the peak areas of the nickel and phosphorus species by 13.92 and 1.25, respectively, for normalisation of the different sensitivities. A and B correspond to the results before and after reduction with hydrogen. Other pretreatment conditions as shown in fig. I . I ' I ' I I I I t I 1 1 1 a60 8 5 0 1 4 0 1 3 0 120 binding energy/eV Fig. 5. ESCA spectra of the crystallised alloy (-) in comparison with the amorphous alloy (---).Pretreatment: 0,, 513 K, 1 h; H,, 573 K, 2 h. Sputtering: (a) 0, (b) ca. 5 and (c) ca. 25 nm. Scale of peak intensity kcPs) for Ni(2p3,,): amorphous (a), (b) 2000 and (c) 10000, crystallised (a)-(c) 2000; for P(2p): amorphous and crystallised (a)-(c) 200.H. YAMASHITA, M. YOSHIKAWA, T. FUNABIKI AND S. YOSHIDA 249 1 (853.4 eV) shifted from that of pure nickel (852.2 eV),ll indicating electron transfer from nickel to other species. The P(2p) peak of phosphorus (129.6 eV) shifted from that of red phosphorus (1 30.4 eV),12 indicating electron transfer to phosphorus. Similar negative shifts of P(2p) have been reported with MnP (0.8 eV),13 CrP (1.3 eV),13 Cu,P (0.3 eV),14 Cu3P (0.4 eV),14 and NIP (0.7 eV),12 indicating interaction between the metal and phosphorus.The peak intensities are summarised in table 2. The total concentration of the nickel species (nickel and nickel oxide) in the surface layer increased with the temperature of the oxygen treatment. The reduction of the oxidised alloy with hydrogen increased the concentration of nickel, but had little effect on the peaks of the phosphorus species, indicating the preferential reduction of nickel oxide. Fig. 5 shows the extent of oxidation of the crystallised alloy in comparison with the amorphous alloy. As seen from the observation of the peak of phosphorus oxide at a depth of 25 nm, oxidation of the crystallised alloy proceeds more deeply inside the bulk of the alloy. DISCUSSION In previous papers we described how the rate of the hydrogenation of buta- 1,3-diene over amorphous Ni-P and Ni-B alloys is affected not only by the partial pressure of hydrogen, but also by the amount and electronic states of the active sites, based on a kinetic analysis.lY2 In the present study we have obtained results to show that the amount and electronic states of the active sites are affected by the temperature of the oxygen pretreatment of the alloy.As shown in fig. 2, the number of surface nickels determined by the adsorption of carbon monoxide and hydrogen increased with the temperature of the oxygen treatment to show a maximum value at 5 13 K (amorphous) and 573 K (crystallised). On the other hand, the T F estimated from the activities for the hydrogenation of ethene and buta-l,3-diene shown in fig.1 and the number of surface nickels also increased with the oxygen treatment temperature, as shown in table 1. These results indicate that modifications of the electronic states and the arrangement of the active species are promoted by the treatment with oxygen. The t.p.r. results indicate that increasing the temperature of the oxygen treatment increases the strength of the Ni-0 bonds of the nickel oxides and the amount of nickel oxide to be reduced by hydrogen. The latter increases with temperature to 573 K, unlike the results in fig. 2, which shows the maximum number of nickel at 513 K. It seems probable that the results in fig. 2 correspond only the surface nickels, but those in fig. 3 correspond to nickels in the surface layers.Treatment of the alloys with oxygen at the higher temperatures oxidises nickel more deeply inside the alloy, as indicated by the ESCA spectra, and reduction of nickel oxide in the surface layers may require higher temperatures than those on the surface. Provided that the t.p.r. curve with a maximum near 503 K corresponds to reduction of the surface nickel oxide, the results in fig. 2 are not inconsistent with those in fig. 3. On the other hand it is also probable that the increase in the amount of the oxide species in the surface layers stabilises the nickel oxide on the surface. It has been reported that the reduction of nickel oxide over SiO, is more difficult than the reduction of unsupported nickel oxide,s and electron transfer from nickel to Si0,15 may be ascribed to stabilisation of the nickel oxide.16y17 However, the increase in the peak area with treatment temperature in fig.3 seems to suggest that the reduction of nickel oxide occurs not only on the surface but also in the surface layers. The ESCA spectra give information about the changes in the nickel and phosphorus species in the surface layers. As shown in fig. 4 and table 2, the relative intensities2492 CATALYSIS BY AMORPHOUS METAL ALLOYS of nickel species compared with those of phosphorus species are smaller than expected from the composition of the alloy, Nig1Pl9. The treatment with acid reduces the total concentration of the surface nickel species, but even the virgin alloy without acid treatment shows a low relative intensity of nickel species.This indicates that the total concentration of phosphorus species in the amorphous Ni-P alloy increases from the bulk to the surface. The total concentration of nickel species in the surface layers increases after the treatment with oxygen; this effect is more obvious when the treatment temperature is high. One possible reason for this phenomenon is the elimination of phosphorus species from the surface during the treatment at high temperatures, but it is also possible that formation of nickel oxide in the surface layers facilitates the diffusion of electron-deficient nickels formed in the bulk to the surface layers. The increase in the concentration of nickel species in the surface layers causes an increase in the total concentration of nickel species after reduction with hydrogen, but not necessarily an increase in the number of surface nickels.Since the catalytic activity increases when unreduced nickel oxide is present in the surface layers and the activity does not correspond to the number of surface nickels, the change in the electronic state of the surface nickels is important for the generation of high catalytic activity. This was shown by the dependence of the TF on the temperature of the oxygen treatment. The shifts in the binding energy of the nickel on the formation of the amorphous alloy and on the oxygen treatment suggest that electron transfer from nickel to phosphorus and oxides of phosphorus and nickel plays an important role. Electron transfer from nickel to phosphorus species is seen from the shift of the binding energies, but since the ratio of phosphorus to phosphorus oxide is not greatly affected by the temperature of the oxygen treatment and by the reduction with hydrogen, electron transfer from nickel to nickel oxide seems to be more important.The reason for the nearly constant ratio of phosphorus and phosphorus oxide is not clear at present, and the phenomenon is very different from the case of Ni-B, in which only the boron oxide peak is detected after treatment with oxygen at high temperature.2 Somorjai and coworkersls? l9 observed that the dehydrogenation and hydrogenation activity of platinum was enhanced by the strongly bound oxygen and proposed that the surface reconstruction caused by oxidation enhances the activity of kink sites of the surface. We have reported that the catalytic activity and selectivity for the hydrogenation of olefins and the hydrogenolysis of alkanes over amorphous Ni-B is related to the structural modification toward the metastable ~ t a t e .~ It is reasonable to assume that similar surface reconstruction takes place during the pretreatments, especially during the oxygen treatment, and affects the catalytic activity, as well as the amount and electronic state of the surface nickels. Compared with the amorphous alloy, the catalytic activity of the crystallised alloy is lower and oxygen treatment at a higher temperature is required for activation. Since the effects of the temperature of the oxygen treatment on the number of surface nickels (fig. 2) and the TF (table 1) are similar to those for the amorphous alloy, the above discussion may be applicable to the crystallised alloy. The main difference between the two alloys is related to the greater homogeneity of the amorphous alloy than the crystallised alloy.The discontinuity and aggregation of nickel species in the surface cause the species in the bulk to be oxidised, as shown in fig. 5, but the interaction between nickel and other species becomes weaker than that of the amorphous alloy, as seen from the lower binding energy of the nickel of the crystallised alloy (853.0 eV) compared with that of the amorphous alloy (853.4 eV). The difference in the electronic state of nickel is reflected in the low TF in table 1. A similar effect was reported for Ni-Fe-P-B amorphous alloy, in which the binding energy of nickel was decreasedH.YAMASHITA, M. YOSHIKAWA, T. FUNABIKI AND S. YOSHIDA 2493 on crystallisation. A decrease in the interaction between nickel and iron on crystal- lisation has been proposed as an explanation.20 In conclusion, the oxygen treatment and reduction of nickel oxide of amorphous and crystallised alloys create surface nickels that play a role as active species. The catalytic activity depends not only on the number of surface nickels but also on the electronic state of the nickel and the surface structure. The electronic state of nickel is affected by the interaction of nickel with other species in the surface layers, but electron transfer from nickel to nickel oxide is important. The higher activity of the amorphous alloy compared with the crystallised alloy is related to the greater homogeneity of the former.S. Yoshida, H. Yamashita, T. Funabiki and T. Yonezawa, J. Chem. SOC., Chem. Commun., 1982,964. S . Yoshida, H. Yamashita. T. Funabiki and T. Yonezawa, J. Chem. SOC., Faraday Trans. 1, 1984,80, 1435. H. Yamashita, T. Funabiki and S. Yoshida, J. Chem. SOC., Chem. Commun., 1984, 868. Proc. 3rd Znt. Conf. Rapidly Quenched Metals, ed. B. Canter (The Chameleon Press, London, 1978). C. S. Brooks and G. L. M. Christopher, J. Catal., 1968, 10, 21 1. C. H. Batholomew and R. B. Pannell, J. Catal., 1980, 65, 390. ' C. H. Batholomew and W. L. Sorensen. J. Catal., 1983, 81, 131. S. D. Robertson, B. D. McNicol, J. H. de Bass and S. C. Kloet, J. Catal., 1975, 37, 424. S. J. Gentry, N. W. Hurst and A. Jones, J. Chem. SOC., Faraday Trans. I , 1981, 77, 603. lo S. J. Gentry and P. T. Walsh, J. Chem. SOC., Faraday Trans. I , 1982, 78, 1515. l 1 V. V. Nemoshkalenko, A. I. Kharlamov, T. I. Serebryakova and V. G. Aleshin, Kinet. Katal., 1978, I f Y. Okamoto, Y. Nitta, T. Imanaka and S. Teranishi, J. Chem. SOC., Faraday Trans. I , 1979,75,2027. l 3 M. Pelavin, D. N. Hendrickson, J. M. Hollander and W. L. Jolly, J. Phys. Chem., 1970, 74, 11 16. l 4 V. 1. Nefedov, Y. V. Solyn, E. P. Domashevskaya, Y. A. Ugai and V. A. Terekhov, J. Electron l 5 M. Houalla, C. L. Kibby, L. Petrakis and D. M. Hercules, J. Phys. Chem., 1983, 87, 3689. l6 J. C. Vedrine, G. Hollinger and T. M. Duc, J. Phys. Chem., 1978, 82, 1515. l 7 H. Haberlandt, J. Phys. Chem., 1983, 87, 3244. l 8 S. M. Davis and G. A. Somorjai, SurJ Sci., 1980, 91, 73. 19. C. E. Smith, J. P. Biberian and G. A. Somorjai, J. Catal., 1979, 57, 426. 2o K. Asami, H. M. Kimura, K. Hashimoto, T. Masumoto, A. Yokoyama, H. Komiyama and 19, 1567. Spectrosc. Relat. Phenom., 1975, 6 , 231. H. Inoue, J. Non-Cryst. Solids, 1984, 64, 149. (PAPER 4/2148)

ISSN:0300-9599    

DOI:10.1039/F19858102485

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

,I. Chem. SOC., Faraday Trans. I, 1985, 81, 2495-2502 Stability Constants and Free Energies of Complexation of Metal-ion Cryptates in Nitromethane Derived Parameters for the Extraction of Cations by Cryptand 222 from Water to Pure Ni tromethane BY ANGELA F. DANIL DE NAMOR,* LILY GHOUSSEINI AND WALTER H. LEE Department of Chemistry, University of Surrey, Guildford, Surrey GU2 5XH Received 31st December, 1984 Stability constants have been measured and free energies of complexation of alkali-metal and silver cations in nitromethane at 298 K have been derived. The results obtained in this solvent are compared with corresponding data in different reaction media. Free energies of transfer of metal-ion cryptates from water to nitromethane are derived. Among the dipolar aprotic solvents the free energies of transfer of cryptand 222 and metal-ion cryptates are in most cases close to 0 kcal mol-'.Ion-pair formation constants and ion-size parameters of cryptate salts in nitromethane as obtained from conductance measurements are reported. The implications of these results on the extraction of cations by cryptand 222 in the water+ nitromethane solvent systems are discussed. The ability of macrocyclic compounds to complex with metal ions has encouraged research into the extraction of cations from aqueous solutions into non-aqueous solvents using natural and synthetic ligands. Numerous publications in this area can be found in the literature.lP8 The stability constants of metal ions and macrocyclic compounds in the various solvents, distribution studies for these ligands and the complexed cation as well as the degree of ion-pair formation in the reaction media are relevant parameters in the analysis of extraction processes.Among the synthetic macrocyclic ligands are those synthesised by Lehng and commonly known as cryptands. A considerable amount of data on the stability constants of metal ions and cryptands in a number of solvents are available. Several papers dealing with free energieslO9 l1 and, more recently, enthalpies and entropies for the transfer of metal-ion cryptates from water to dipolar aprotic solvents12~ l3 have been published. Ion-pair formation constants for the cryptate salts in only a few solvents are known.14-17 As the complexation reaction between the macrobyciclic ligands and metal ion in a given medium involves competition between the ligand and the solvent for the cation, stability constants are higher in a poor cation-solvating medium.ls This is the case for nitromethane, as recently shown by our work on the free energies of transfer of ions from water to this s01vent.l~ In addition, the use of this solvent in extraction processes is of practical significance.In this paper we report the stability constants of metal ions with cryptand 222 in nitromethane and the free energies of transfer for the cryptates from water to this solvent are derived. Values for the partition coefficient of cryptand 222 between water and the appropriate solvent and ion-pair association constants, Ka, of the cryptate salts in nitromethane as obtained from conductance measurements are given.The implications of these results to the extraction of cations from pure water to the pure solvent are discussed. 24952496 STABILITY CONSTANTS OF METAL-ION CRYPTATES EXPERIMENTAL Nitromethane (Aldrich Chemical Co., 98%) was used as purchased. The water content, checked by gas-liquid chromatography and Karl Fisher titration, was 0.005 % . Fractionally distilled solvent was found to be of the same purity (boiling-point range 98-100 "C). The solvent was always kept in a dry box in the dark. It was found that some samples of nitromethane (boiling-point range 100-1 03 "C) presented difficulties during purification. The impurity, as checked by gas chromatography, was found to be not less than ca. 1-2% even after several distillations and therefore unsuitable for experimental use.Cryptand 222 (commercial sample, Merck) was used without further purification, dried in uacuo and stored in a desiccator over CaCl, in the dark. The salts used were LiClO, (B.D.H., AnalaR grade), NaClO, (Koch-Light, cryst. puriss), KClO, (Fisons, SLR), CsClO, (Ventron), RbC10, (Alfa), CsI (Fluka, puriss, p.a.) and RbI (B.D.H., AnalaR grade). The salts were dried under vacuum for several days prior to use. The stability constants of alkali-metal and silver cryptates in nitromethane were measured by the potentiometric titration technique, which has been used extensively by Schneider and coworkers.20~ 21 All potential readings were taken with a 9PHM 82 radiometer of high input impedance (> 10l2 Q). For the partition experiments, the solvents n-tetradecane and nitromethane were mutually saturated before use.A solution of cryptand 222 in nitromethane was prepared and analysed. An accurately measured volume of the cryptand solution in nitromethane was added to a stoppered test tube containing a known volume of n-tetradecane and the mixture was shaken and allowed to equilibrate in a thermostatted water bath at 25 0.01 "C for several hours. The two solvent layers were separated just before analysis. The concentration of cryptand 222 in the nitromethane layer both before and after partition was analysed gravimetrically. Experiments were performed in triplicate. For the conductance measurements, a Hartley-Barrett conductivity cell with bright platinum electrodes was connected to a Wayne-Kerr automatic conductivity bridge of the transformer radio arm type B 642; an input voltage 0.2-1.2 V and a fixed frequency of 1592 Hz were used.All conductances were measured under nitrogen atmosphere. Corrections for the conductance of the solvent were applied. The results are expressed in calories, 1 cal = 4.1840 J. RESULTS AND DISCUSSION Table 1 shows the stability constants, expressed as log K,, and derived free energies of complexation for the alkali-metal and silver cations with cryptand 222 in nitromethane at 298 K. This is the first set of data for the stability of metal ions and cryptand 222 in this solvent. Relevant to this discussion are the stability constants of the corresponding cryptates in different solvents ; for comparison purposes these are shown in table 2.Close examination of these data indicates that the pattern observed for the stability constants of cryptand 222 and alkali metals (K+ > Rb+ > Na+ > Cs+ > Li+) in H20, MeOH, EtOH, DMF and Me,SO is changed in PC and AN (Na+ > Rb+ and Li+ > Cs-'). In spite of this, among these solvents the stability constants are higher for potassium than for any other alkali metal when cryptand 222 is the ligand. This is in accord with the concept of maximum stability observed among cryptands and the cation which best fits into the cavity of the ligand. However, when nitromethane is the reaction medium, this is no longer the case. Among the alkali-metal cations the highest stability constant obtained is that of sodium and cryptand 222 (table 1) in this solvent.Results obtained recently for the heats of complexing of metal ions and cryptand 222 also show that among the alkali-metal cations an exothermic maximum is found for sodium and cryptand 222 in nitromethane.22 Single-ion free energies of transfer of alkali-metal and silver cations from water show that nitromethane is a poorer solvator for these cationslg than any of the solvents so far considered in this paper. Comparison of the data (table 1 and 2) shows that theA. F. DANIL DE NAMOR, L. GHOUSSEINI AND W. H. LEE 2491 Table 1. Stability constants (log K,) and derived free energy of complexing of alkali-metal and silver cations with cryptand 222 in nitromethane at 298 K complex log K," AGzb/kcal mo1-l [Li+222] 1 1.49 f 0.05 - 15.68 [Na+222] 13.56 k 0.09 - 18.50 [K+222] 12.58 f 0.10 - 17.16 [Rb+222] 10.30+0.10 - 14.05 [Cs+222] 5.10+0.10 - 6.96 [Ag+222] 17.7 1 & 0.06 -24.16 a This work.AGO = - RT 2.303 log K,. Table 2. Stability constants (log K,)" of alkali-metal and silver cations with cryptand 222 in various solvents at 298 K solvent complex H,O MeOH DMF Me,SO PC AN ~~ ~ ~ [Li+222] 0.98 2.60 - < 1.0 6.94 6.99 [Na+222] 3.98 7.98 6.05 5.35 10.54 9.63 [K+222] 5.47 10.41 7.95 6.99 11.15 11.01 [Rb+222] 4.24 8.98 6.73 5.78 9.02 9.50 [Cs+222] 1.47 4.40 2.13 1.43 4.10 4.55 [Ag+222] 9.6 12.20 10.05 7.22 16.31 9.07 a Average values from ref. (10); precision in log K, is 20.1 log units. stability constants are greater in nitromethane than in any other solvent so far investigated. This provides further evidence of the important role played by ion-solvent interactions in the formation of metal ion and cryptand complexes.We have recently shown13 that for enthalpies in dipolar aprotic solvents, the experimental data so far obtained led to a straight line represented by of unit slope and almost zero intercept when the differences between AH&, the heat of complexing of a given metal cation M+ and cryptand 222 in a solvent B (DMF, Me,SO and AN), and the corresponding value AH& in solvent A (PC) were plotted against the single-ion enthalpy of transfer, AK[M+],,B, based on the Ph,AsPh,B convention (molar scale) from solvent A to solvent B. As pointed out before, when eqn (1) was tested in terms of free energy using the values for the stability constants of metal ions and cryptand 222 in dipolar aprotic solvents reported by Cox et it was found that ten out of sixteen data points considered followed eqn ( I ) to within 0.6 kcal mol-I.Free energies of complexation of metal ions and cryptand 222 in nitromethane (table 1) are combined with corre- sponding values in PC (reference solvent) and compared with the single-ion AG: values for the metal ions from PC -+ MeNO, based on the Ph,AsPh,B c o n ~ e n t i o n . ~ ~ ~ ~ ~ The results (table 3) indicate that the experimental data obtained for the six cations in the two solvents considered follow eqn (1) to within 0.5 kcal mol-I. Fig. 1 shows2498 STABILITY CONSTANTS OF METAL-ION CRYPTATES Table 3. Differences between free energies of complexing of metal ions and cryptand 222 in nitromethane and propylene carbonate and single-ion AG; values from PC -+ MeNO, (Ph,AsPh,B convention) at 298 K [Li+222] - 15.68 - 9.47 -6.21 [Na+222] - 18.50 - 14.38 -4.12 [K+222] - 17.16 - 15.20 - 1.96 [Rb+222] - 14.05 - 12.31 - 1.74 [Cs+222] -6.96 -5.59 - 1.37 [Ag+222] - 24.16 - 22.25 - 1.91 6.54 4.34 2.38 2.23 1.74 1.89 a Data from table 1 .Data derived from log K, in PC given in table 2. Calculated from data given in ref. (19), (23) and (24). \ AGP [M+] (PC -+ B)/kcal rno1-I Fig. 1. Plot of AGg[M+ +Cry] (B) - AGZ[M+ +Cry] (PC) against single-ion AG;[M+] (PC -+ B) based on the Ph,AsPh,B convention: x , cryptand 21 1 ; 0 , cryptand 221 ; 0, cryptand 222. the correlation between (AG&--AGgA) as AG:[M+],,B for M+ = Li+, Na+, K+, Rb+, Cs+ and Ag+, B =DMF, Me,SO, AN and MeNO, and A = PC.All available data for cryptand 221 and 21 1 are included.1° Comparison of the stability constants of metal-ion cryptates (222) with those reported in the literature for Li+, Na+, Cs+ and Ag+ and cryptand 22 in nitr0methane~~9 26 shows a considerable increase in stability for [M+222] with respect to [M+22] complexes, this being more pronounced for sodium. The stability constantsA. F. DANIL DE NAMOR, L. GHOUSSEINI AND W. H. LEE 2499 (log K , ) reported for this cation and cryptand 22 in nitromethane26 of 3.37 increased by approximately ten units when cryptand 222 is the complexing agent. The free energy of transfer of metal-ion cryptates from water to nitromethane are calculated via a thermodynamic cycle AGF[M+222] = AG~Me,,2-AG~Hp,+AGZ)[M+]+AGF[222]. (2) As there is no data available for the standard free energy of transfer of cryptand 222 from water to this solvent, the AGF values was evaluated using a distribution method with tetradecane + nitromethane and tetradecane + water as pairs of immiscible solvents.A value of 0.3258 for the partition coefficient of cryptand 222 in the tetradecane + nitromethane system was obtained. When combined with a partition coefficient of 0.0187 for the same ligand in the tetradecane+water system,,' a value of 0.0574 was obtained for the partition of cryptand 222 in the nitromethane + water solvent system. Therefore, the free energy of transfer for the process [222] pure water -+ [222] pure MeNO, (3) is calculated to be 1.69 kcal mol-l.This result differs from the value of 2.2 kcal mol-l obtained by us from partition experiments in the water (sat. MeNO,) + nitromethane (sat. H,O) solvent system.,, In considering the differences in the AGE) [222] values for the water + pure nitromethane and nitromethane (sat. H,O)+ water (sat. MeNO,) systems, the solubility of water in nitromethane and vice versa could be of relevance, although the differences found between the AGF for the pure and AG; for the saturated phases are not very significant. If the solubility of water in MeNO, was considerable, the nitromethane (sat. H,O) solvent having a stronger dipolar character was likely to behave as a mild hydrogen-bond donor, as a result, there could be a slight shift in the position of cryptand 222 in favour of the organic phase and consequently AG; 12221 would have been more negative than AGF 12221 for the transfer between two pure solvents.However, the opposite is observed in this case. This could be related to the higher solubility of nitromethane in water (10%) with respect to that of water in nitromethane (2%).,* The free energy of transfer of cryptand 222, combined with free energies of complexing of metal ions and cryptand 222 in nitromethane (table 1) and the corresponding values in water together with single-ion free energies of transfer based on the Ph,AsPh,B convention, yields [eqn (2)] free energies of transfer of metal-ion cryptates from water to nitromethane. Details are given in table 4. A variation of AGF[M+222] (H,O -+ MeNO,) with the central ion M+ is observed.Free-energy values for the transfer of metal-ion cryptates and cryptand 222 to five dipolar aprotic solvents (table 5 ) clearly indicate that among dipolar aprotic solvents the AGE)[M+222] and AG,"[222] are, with a few exceptions (mainly in transfers involving Me,SO), close to 0 kcal mol-l. These exceptions are being examined.29 Cox et al.25 have reported AG,"[M+22] from acetontrile to several dipolar aprotic solvents (DMF, Me,SO, AN, PC, MeNO, and Me,CO) and found that AG,"[M+22] were around zero for transfer to DMF, PC and Me,CO, with the exception of AG,"[M+22] values for the transfer to Me,SO. In table 6 we compare our data for the AG,"[M+222] from AN to MeNO, with those reported by these authors for the transfer of [M+22]. The results suggest that in transfers from acetonitrile to nitromethane the behaviour of [M+22] differs considerably from that observed for the AGE)[M+222] complexes in the same solvent systems, and unlike transfers to DMF, PC and Me,CO (AG,"[M+22] z 0) and to Me,SO (AGf[M+22] z -2 kcal mol-l), free energies of transfer of [M+22] (AN -+ MeNO,) seem to be dependent on the metal ion.2500 STABILITY CONSTANTS OF METAL-ION CRYPTATES Table 4.Free energies of transfer (kcal mol-I) of metal-ion cryptates (Ph,AsPh,B convention) from water to nitromethane in kcal mo1-I at 298 K cation AG~MCx02u AGZHzOb AG:[222]" AG;[M+ld AG:[M+222] Lit - 15.68 - 1.34 1.69 12.05" - 0.60 Na+ - 18.50 - 5.42 1.69 7.55 - 3.84 K+ - 17.16 - 7.45 1.69 3.69 - 4.33 Rb+ - 14.05 - 5.78 1.69 2.64 - 3.94 Cs+ - 6.96 - 1.97 1.69 1.35 - 1.95 Ag+ - 24.16 - 13.10 1.69 6.20 -3.12 a This work, table 1.Ref. (10). This work, see text. Ref. (19). Ref. (24). Table 5. Free energies of transfer (kcal mol-I) of alkali-metal and silver cryptates (Ph,AsPh,B convention) and of cryptand 222 from water to different dipolar aprotic solvents at 298 K cation H,O + DMFa H,O -, Me,SOa H 2 0 -+ PCa H,O -+ ANa H 2 0 -+ MeN02b -0.72 0.10 - 0.60 [Na -I- 22 21 - 3.56 - 3.54 - 3.85 - 3.19 - 3.84 [ Li + 2221 - - [K+222] - 4.26 - 3.67 - 4.54 -4.72 -4.33 [ R b-I-2221 -4.16 - 3.05 -4.22 -4.51 - 3.94 [ c s + 2 2 21 - 1.96 - 1.49 -2.1 1 -2.16 - 1.95 [Ag+222] -3.16 -3.13 - 2.94 - 3.69 - 3.12 cryptand 222 I .56 1.54 1.90 1.10 I .69 a Ref. (13). This work. Table 6. Free energies of transfer (kcal mol-I) of [M+222] and [M+22] from acetonitrile to nitromethane at 298 K ~~ cation [M+] [M+22] [M+222Id [22Ib [222Id Li + 4.95 1.12b -0.70 -0.29 0.59 Na+ 4.25" 3.00" -0.65 -0.29 0.59 Cs+ 0.15" -0.86' 0.21 -0.29 0.59 Ag+ 11.10 2.98' 0.57 -0.29 0.59 a Calculated from ref.( 1 3) and (19). Ref. (25). From ref. (25), adjusted for single-ion data given in footnote (a). Calculated from table 5. EXTRACTION OF CATIONS Separation between water and nitromethane can be achieved and consequently the extraction of cations from pure water to the pure nitromethane phase using cryptand 222 as the extracting agent is discussed. Information regarding ion-pair formation in the organic phase is relevant to this discussion. Data for the process represented by M+222(MeN02) + X-(MeNO,) 2 [M+222]X- (MeNO,) (4)A.F. DANIL DE NAMOR, L. GHOUSSEINI AND W. H. LEE 250 1 Table 7. Ion-pair association constants, K,, and ion-size parameters, ii, of metal-ion cryptates in nitromethane at 298 K concentration range A;C [M+222]X- a /mol dm-3 /W1 cm2 mol-' KJmo1-l dm3 &/A [Lif222]C10; 0.73 x 10-"9.87 x lo-' 89.28 0.59d 10.2 [ N a+ 222IClO; 0.98 x 10-4- 1.29 x 10-3 87.13 0.77d 10.4 [Kf222]C10; 0.77 x 10-4- 1.12 x 10-3 97.10 0.30d 10.3 [Rb+222]1- 1.09 x 10-4- 1.28 x 10-3 88.70 0. lod 10.8 [Cs+222]I- 1.02 x 10-4- 1.28 x 10-3 94.49 0.1 5d 11.0 a Conductance data analysed by Fuoss-Shedlovsky plots for pairwise associated electrolytes. Conductance data analysed by Fuoss-Acascina equation for strong electrolytes and Fuoss- The A,,, values for each electrolyte Shedlovsky plots for pairwise associated electrolytes.concentration are available on request. Electrolytes fully dissociated in MeNO,. Table 8. Extraction of cations by cryptand 222 in the water+nitromethane system at 298 K cation anion Li + Na+ Kf Rb+ c s + c10; I- ~~~~~ ~~~~~~~~~~~~~~~ ~ K,n 9 . 6 ~ 10 9.5 x lo3 2.9 x lo5 1.7 x lo4 2.8 x 10 - - Kt[M +222Ib 2.7 6.5 x 10' 1.5 x lo3 7.7 x 10' 2.7 x 10 - - Kt[X- I' K?xt(CIO;)d KexdI- )" - - - - 1.51 x 10-1 4.95 x 10+ - 3.9 x 10 9 . 3 ~ 1o.i 6 . 6 ~ lo7 2 . 0 ~ lo6 1.1 x 10' - - 1.3 x 10-1 3.1 x 103 2.1 x ioj 6.5 x 103 3.7x 10-1 - - Kt[M++C1O;]f 2.3 x 4.4 x lo-' 3 . 0 ~ lop4 1.8 x lop3 1.5 x lo-' a Stability constants of cryptand 222 and metal ions in water (table 1). Values obtained Eqn ( 5 ) , see text, using from AG:[M+222] (H,O C10, as anion.MeNO,) given in table 4. See ref. (19). Eqn (9, see text, using I- as anion. f Ref. (19). as obtained from conductance measurements of metal-ion cryptates in nitro- methane at 298 K are reported in table 7. These results show that no ion-pair formation is observed for [M+222]X- (X- = C10,- or I-) in nitromethane. Further evidence of the complete dissociation of these electrolytes in nitromethane is also found in the measurements of the stability constants and the heats of complexation of alkali-metal and silver ions with cryptand 222 in this solvent, since no variation is observed in these measurements with changes in the electrolyte concentration. Hence, the extraction constants for the following process: M+(aq) + X-(aq) + 222(aq) -+ M+222(MeNO,) + X-(MeNO,) ( 5 ) are calculated by Kext = K, Kt [M+222] Kt [X-] where K,, Kt [M+222] and KJX-] indicate the stability constants of metal ions and cryptand 222 in water and the partition coefficient of metal-ion cryptate and of the anion from water to nitromethane.Details are given in table 7. Comparison between these values and those previously reported by us19 for the transfer of the 1 : 1 electrolyte M+(aq)+X-(aq)+ M+(MeNO,)+X-(MeNO,) (7)2502 STABILITY CONSTANTS OF METAL-ION CRYPTATES shows that Kext are larger than Kt[M+ + X-] since the electrolytes are transported by cryptand 222 to the organic phase. A quantitative evaluation on the effect of this ligand on the transfer of the electrolyte M++X- from water shows that for the alkali-metal cations the greatest change in selectivity is observed for the Na+/Cs+ pair of ions.A change of selectivity in favour of sodium with respect to caesium by a factor of 2.9 x lo8 in the presence of cryptand 222 is observed [K,(Na+)/K,(Cs+) =2.9 x lop5 and Kext(Na+)/Kext(Cs+) = 8.4 x lo3]. A quantitative evaluation of the influence of the anion on the extraction of cations indicates that the extraction could be enhanced by a factor of ca. 3 x 10, if perchlorate is used instead of iodide. Since there is mutual solubility between water and nitromethane we are now in the process of investigating an actual extraction process which refers to the H,O (sat. MeNO,) and MeNO, (sat. water) solvent systems. Finally, note that if the anion is excluded in eqn (6), the extraction constant is reduced to Kext = Kw Kt[M+222], and considering that there is not an appreciable variation in the AG:[M+222] (from which Kt[M+222] is obtained) from water to the dipolar aprotic solvents (table 5), it becomes quite clear that Kext in the water+ nitromethane solvent system would be practically of the same magnitude for any water + dipolar aprotic solvent so far considered.We thank Dr B. G. Cox, University of Stirling for valuable discussions. G. Eisenman, S. M. Ciani and G. Szabo, J. Membr. Biol., 1969, 1, 294. H. K. Frensdorff, J. Am. Chem. Sac., 1971, 93, 4684. I. M. Kolthoff, Anal. Chem., 1979, 51, 1R. I. Takeda, Bull. Chem. SOC. Jpn, 1983,56, 931. H. Nakamura, M. Takagi and K. Ueno, Anal. Chem., 1980,52, 11. I. Takeda, I. Wada and S. Fujiwara, Bull. Chem.SOC. Jpn, 1981, 54, 3727. 'I I. Takeda, Bull. Chem. SOC. Jpn, 1983, 56, 2589. * S. K. Sahni and J. Reedijk, Coord. Chem. Rev., 1984, 59, 1. J. M. Lehn, Struct. Bonding (Berlin), 1973, 16, 1. lo B. G. Cox, J. Garcia Rosas and H. Schneider, J . Am. Chem. Soc., 1981, 103, 1384. B. G. Cox, N. van Truong and H. Schneider, J. Am. Chem. SOC., 1984,106, 1273. l 2 A. F. Danil de Namor and L. Ghousseini, J. Chem. SOC., Faraday Trans. I, 1984,80, 2349. l 3 A. F. Danil de Namor and L. Ghousseini, J . Chem. SOC., Furaduy Trans. I , 1985, 81 781. l4 S. Boileau, P. Hemery and J-C. Justice, J. Solution Chem., 1975, 4. l5 A. Deffieux, F. Menezes and S. Boileau, Symposium on Ion and Ion Pairs in Non-aqueous Media (University of Leuven Press, Leuven, Belgium, 1976), p. 115. l6 F. C. Thyrion, Symposium on Ion and Ion Pairs in Non-aqueous Media, University of Leuven Press, Leuven, Belgium, 1976), p. 107. M. H. Abraham, A. F. Danil de Namor, W. H. Lee and R. J. Wheaton, Acta Chem. Scand., Ser. A, 1980, 34, 621. B. G. Cox, Annual Reports (Royal Society of Chemistry, London, 1981), part C. l9 A. F. Danil de Namor and L. Ghousseini, J . Chem. SOC., Furuday Trans. I , 1984, 80, 2843. 2o B. G. Cox, H. Schneider and J. Stroka, J. Am. Chem. SOC., 1978, 100,4746. J. Gutknecht, H. Schneider and J. Stroka, J. Inorg. Chem., 1978, 17, 3326. 22 A. F. Danil de Namor and L. Ghousseini, unpublished results. 23 M. H. Abraham, Monatsch. Chem., 1979, 110, 517. 24 Y. Marcus, Rev. Anal. Chem., 1980, 5, 63. 25 B. G. Cox, P. Firman, H. Horst and H. Schneider, Polyhedron, 1983, 2, 543. 26 M. Shamsipur and A. I. Popov, Inorg. Chim. Acta, 1980, 43, 243. 27 H. C. Ling, Ph.D. Thesis (University of Surrey, 1981). 28 A. F. Danil de Namor and H. Berroa de Ponce, work in progress. D. M. Murray, Rust, H. J. Hadow and H. J. Hartley, J . Chem. SOC., 1931, 215. (PAPER 4/2 192)

ISSN:0300-9599    

DOI:10.1039/F19858102495

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1985, 81, 2503-2512 The Thermal Decomposition in the Solid Phase (Crystolysist ) of Silver Malonate BY ANDREW K. GALWY AND MOHAMED A. MOHAMEDS Department of Pure and Applied Chemistry, The Queen’s University of Belfast, Belfast BT9 5AG, Northern Ireland Received 31sr December, 1984 From the complementary consideration of microscopic and kinetic evidence we conclude that the thermal decomposition of silver malonate, CH,(COOAg),, in the temperature range 453-485 K, occurs in the solid state by a nucleation-and-growth reaction. During this decomposition the small crystallites of reactant undergo no detectable textural change, except for some surface roughening. Isothermal fractional reaction (a) against time curves are sigmoid, and after subtraction of a subsidiary initial process the main decomposition obeys the Avrami-Erofe’ev equation where n = 2 (and in a few instances 3).This kinetic fit is ascribed to the growth of coherent but elongated nuclei (composed of Ag metal embedded in a carbonaceous material) in the rectangular but asymmetric (lath-shaped) reactant crystallites. Appreciable differences in the shapes of the a against time curves for decompositions of the five differently pretreated reactant samples studied are ascribed to variations in mean reaction geometry. However, it is concluded that the chemical changes occurring at the interface are identical for all preparations, since the Arrhenius parameters agree within experimental error. Mean values of the activation energy ( E ) and pre-exponential factor ( A ) were 170 & 10 kJ mo1-I and 15.2 f 1.2 s-l, respectively. Comparisons with the limited information available concerning the decomposition of silver acetate lead us to conclude that decomposition here is controlled by the breakdown of adsorbed radicals, probably including -CH2C02- and related species, on the surface of the metallic silver product.The chemical steps probably involve heterogeneous catalytic-type interactions between several dissociatively adsorbed intermediates on the active metal at the reactant-product interface. [-log (1 - = kt This kinetic and mechanistic study of the thermal decomposition of silver malonate was undertaken to extend a systematic investigation of the pyrolyses of metal salts of malonic acid, related work2 having been concerned with n i ~ k e l , ~ cobalt4 and copper5 malonates.Many contributions towards our understanding of the thermal reactions of solids have been based on observations of the pyrolyses of metal carboxylates,2 notably oxalates and formates. The present reactant was selected to permit comparisons with the behaviour of the other malonates, particularly the copper salt5 (the subject of a parallel study in these laboratories) and the intensively studied silver oxalate.2* 6? A fortunate consequence of this selection of reactant was that silver malonate is anhydrous, and the reaction occurs in the crystallites as prepared, rather than after the recrystallization and retexturing processes which is inevitable when dehydration precedes the anion-breakdown reaction.t For reasons given in detail in a recent review’ we suggest the use of this term as an aid to indexing t. Permanent address: Assiut University, Department of Chemistry, Faculty of Science, Qena, Egypt. studies in this field. 25032504 CRYSTOLYSIS OF SILVER MALONATE The present study demonstrates once again2 that decompositions of solids are individual processes, and silver malonate exhibited no obvious close relationships with the similar reactions of the other malonates mentioned or silver oxalate. More detailed comparisons with the results for copper(r1) malonate5 only served to emphasize the differences in behaviour of these reactants; this is discussed further below. The limited information that could be found concerning the pyrolysis of silver acetateR3 and the silver-metal-catalysedlO decomposition of acetic acid was, however, of greater value in formulating a mechanism for the decomposition of silver malonate. We provide evidence here, from complementary kinetic, microscopic and analytical measurements, that this is a solid-state nucleation and growth rate process. We could find no literature reference to any study concerned with the decomposition of silver malonate apart from our short initial report.ll The present article includes additional measurements, and these are combined with some reanalyses of the earlier data together with a more comprehensive consideration of the reaction mechanism.EXPERIMENTAL APPARATUS The isothermal decomposition of silver malonate was studied kinetically in a conventional constant-volume vacuum apparatus.The pressure of gas evolved was measured at known times by a McLeod gauge with a 250 or 178 K trap interposed between it and the heated (& 1 K) reactant. Pressures corresponding to completion of reaction were normally measured on completion of every experiment. MICROSCOPY Electron-microscopic examinations were made of the salt as prepared, after decomposition to various known fractional reaction values (a) and after completion of the reaction (a = 1 .O). Scanning electron photomicrographs (SEM) were obtained using a Jeol 35CF instrument, operated at low voltage (10 or 15 kV) to minimize specimen damage. Before examination, samples were precoated with Au/Pd. Such samples were examined in duplicate, as prepared and after internal structures had been exposed by gentle crushing following reaction.Since this method did not permit characterization of the detailed structures that were of interest here, higher magnifications were obtained from replicas examined by transmission (TEM) in a Philips EM 400 electron microscope. Surface textures were observed as replicas from impressions of partly (a = 0.3) or completely (a = 1.0) decomposed crystallites on acetone-softened cellulose acetate films, using a two-stage preshadowed carbon replication technique.12 REACTANT SALT Silver malonate was prepared by the slow addition of a dilute aqueous solution of sodium malonate to dilute aqueous silver nitrate. The insoluble white precipitate obtained was composed of small crystallites and was separated by filtration, washed and dried. Although later work showed silver malonate to be insensitive to light, the preparations of the five salts (A-E) studied were completed under conditions of minimum illumination and the salts stored in the dark.In an attempt to increase the sizes of the crystallites prepared in this way, four of the salts (B-E) were maintained for various times in water; details of such digestions are specified below in the context of the kinetic data. The elemental compositions (C and H) of all five salts (from combustion analyses) were close to the theoretical requirements (1 1.3 % C and 0.63 % H) of the anhydrous salt, CH,(COOAg),. The silver contents, however, were less (by ca. 3 % ) than the theoretical proportion (67.9% Ag) but values tended to increase towards this value after longer times of digestion.This is perhaps indicative of the retention of a small proportion of the acid salt in the freshly precipitated material.A. K. GALWEY AND M. A. MOHAMED 2505 RESULTS AND DISCUSSION STOICHIOMETRY The overall decomposition of silver malonate can be satisfactorily expressed by CH,(CO,Ag), -+ 1 .45C02 + 0.04CO + 0.60(CH,C02H/CH,CO) + [2Ag + C0.*, H0.2] and there was no detectable variation with temperature between 455 and 480 K. Volatile products, CO, and CO, were measured from the pressures of gas evolved in the calibrated volume of the apparatus using appropriate (i.e. 78 and 178 K) refrigerant traps. These products were also shown to predominate by mass- spectrometric analyses. The residual product is shown in square brackets: carbon and hydrogen were determined by combustion analyses and a small amount of oxygen may also have been retained.The evolution of acetic acid as a product was confirmed by n.m.r. measurements on D20 solutions of the condensed volatile materials. However, the quantity detected was greater than that allowed by the hydrogen content of the reactant. Mass- spectrometric measurements gave direct evidence that both acetic acid and ketene were formed; these products were also formed during the decomposition of acetate adsorbed on silver (110) in a slightly higher temperature range.lO There was no evidence that acetic anhydride was produced, and the elemental balance in the above equation is satisfied, within experimental error, if the reaction yields approximately equimolar quantities of acetic acid and ketene.N.m.r. measurements for solutions in D 2 0 of samples of the salt previously decomposed to various known extents showed that these contained no detectable quantity of acetate. Therefore silver acetate does not accumulate as a reaction intermediate [in contrast to the reaction of copper(1r) malonateI5 but is expecteds? to decompose above 453 K under reaction conditions (yielding products similar to those obtained from the present decomposition). ELECTRON MICROSCOPY Significant textural characteristics of the reactant (salt D) are illustrated in Plates 1 (SEM) and 2 (replicas, TEM). At a = 0.85 [plate 1 (a) (SEM)] there is no indication of melting, disintegration or cracking of the lath-shaped reactant crystallites; the sizes and shapes of the particles here are closely similar to those of salt D as prepared.The only perceptible change is the textural modification of the crystal surfaces, which were smooth before reaction. When a = 1 .O [plate 1 (b) (SEM)] the crystal shape is preserved intact and the section exposed by fracture on crushing reveals the coherent structure of the product occupying the interior. The residual particle is pseudomorphic with the reactant crystallite. Plate 2(a) (a = 1.0, TEM) shows the texture of a thin product layer adhering to the replica. This residual material is composed of numerous spherical particles, diameters largely between 0.04 and 0.16 pm, opaque to the electron beam and identified as metallic silver. These are embedded in a structural connective matrix, less dense to the electron beam and identified as the carbonaceous component of the residue.Plate 2(6) (a = 0.3, TEM) is a replica of an apparently undamaged crystal surface. The textured zone, part of which has adhered to the replica, is identified as an asymmetric (i.e. non-spherical) nucleus that has developed preferentially along the crystal edge. Adjoining areas have apparently remained unreacted. This, together with 82 F A R 12506 CRYSTOLYSIS OF SILVER MALONATE t/min Fig. 1. Comparison of the sigmoid a against time curves for the evolution of gaseous products (CO, and a small proportion of CO) during the isothermal (466.5 & 1 K) decomposition of silver malonate. Appreciable differences in behaviour are shown for the five differential pretreated reactant salts: A-A; A-B; 0-C; e-D; x-E.similar observations for other crystals, provides important direct evidence that decomposition proceeds by a nucleation-and-growth reaction mechanism. REACTION KINETICS The small sizes of the reactant crystallites made it difficult to characterize by microscopy the systematic changes in reaction geometry13 that occurred during the progress of the decomposition. Accordingly we digested four preparations of salt (B-E) in water at a range of temperatures in an attempt to increase reactant crystallite sizes. This was only partially successful and resulted in relatively limited crystal growth. However, such pretreatment caused the appreciably different patterns of reactivity and rate characteristics illustrated in fig.1 for the decompositions of salts A-E at 466.5 K. Despite these evident variations, the principal features of kinetic behaviour (i.e. rate equation obeyed and Arrhenius parameters) for all five salts were closely similar. Results of the kinetic analyses are summarized in table 1, which includes the times and temperatures of reactant digestion and approximate estimates of reactant crystal li te sizes. Isothermal fractional reaction (a) against time curves for the decompositions of all five reactants were sigmoid (fig. 1). Typical behaviour of salt A for reactions at several temperatures between 465 and 480 K are shown in fig. 2. Induction periods to the onset of salt breakdown were short, and after completion of an initial rate process at or, the main reaction obeyed the Avrami-Erofe’ev equation2 [modified by the subtraction of the contribution of the initial process, a’ = (a-oc,)/(l - a , ) ] : [-log (1 - = k t .(1) The obedience of the data in fig. 2 (salt A) to this equation, with a, = 0.04, is shown in fig. 3. Individual accounts of the behaviour of each salt are given in the followingA. K. GALWEY AND M. A. MOHAMED 2507 1 .c 01 0.5 0 . . 0 . . . 0 . 0 . . . . . . . . . . . . . (a) . 50 100 t/min 150 Fig. 2. Comparative fractional reaction (a) against time curves for the decomposition of salt A at the following temperatures: (a) 465, (b) 469, (c) 474, ( d ) 477 and (e) 480 K. Induction periods to onset of reaction are short. 101 . . . . I I 1 25 50 75 1 t/min '0 Fig.3. Avrami-Erofe'ev obedience test from plots of [-log (1 - a')f against time for data shown in fig. 2 (salt A) with = 0.04. Eqn (1) is obeyed with n = 2 in the range 0.05 < a < 0.54; subsequently the reaction rate diminished and the first-order equation (P = 1) was obeyed in the range 0.57 < a < 0.88, (a-e) as in fig. 2. paragraphs. These were the same salt preparations as those referred to in our previous publication. l1 Salt A . This preparation was dried without digestion in water, unlike the other reactants studied, and it had the smallest particle size (table 1). It was the most reactive salt, exhibiting the most rapid onset of decomposition (fig. 1). Fifteen isothermal rate 82-22508 CRYSTOLYSIS OF SILVER MALONATE Table 1. Summary of kinetic results for the thermal decomposition of silver malonate salt parameter A B C D E digestion of 0 (0) salt in water during preparation : t/h ( T / K ) parameters in eqn (1): n 2 aia 0.04 obeyedb Arrhenius parameters : log ( A / s - l ) 14.87 E/kJ mol-l 164 & 4 reaction rate at 466.5 K, a' range 0.01-0.50 maximum 2.5 (da/dt)rn,, /lo4 S-1 approximate ranges of crystallite dimensions : length/pm 7-10 thicknesslpm 1 width/pm 1-2 14 (290) 2 0.00 0.00-0.60 15.39 171 + 4 1.5 20-40 5-10 2 4 0.5 (313) 2 0.03 0.01-0.50 14.95 168+8 0.95 10-20 3-5 1-2 1.5 (333) 2 0.17 0.09-0.8 5 17.16 185+_9 1.6 30-50 5-10 2-4 3.0 (338) 3 0.05 0.02-0.80 14.33 161 & 12 2.8 20-40 5-10 1-2 a a, values were determined from the subtracted values of a necessary to obtain obedience of data to eqn (1).See text also. measurements were made (fig. 2 and 3) between 450 and 482 K and data obeyed (1) with n = 2 and a, = 0.04 between 0.05 < a < 0.52. Thereafter the reaction rate diminished and the first-order equation was obeyed 0.57 < a < 0.88 with virtually unchanged Arrhenius parameters [log (A/s-l) = 15.12 and E = 165 Decomposition rates were insensitive to the effects of preirradiation or mechanical pretreatment. Irradiation in the visible o r near-ultraviolet regions for 30 min resulted in no perceptible change in the rate of the subsequent reaction. The effects of reactant crushing or pelleting or variations in sample weight were very small or negligible. Ageing for two years in the dark at ambient temperatures resulted in a slight reduction in the rate of decomposition at 455 K and the diminution of a, from 0.04 to 0.0 I .Salt B. Digestion for 14 h in water at 290 K resulted in a significant increase in crystallite sizes and a relative reduction in reaction rate, (daldt),,, (see table 1). Eqn (I), with n = 2 and cc, = 0.00, was again obeyed, but at a > 0.60 there was a positive deviation and eqn (1) with n = 3 provided a satisfactory fit between 0.55 < a < 0.85. 5 kJ mol-l].A. K . GALWEY AND M. A. MOHAMED 2509 Salt C. This salt was digested 0.5 h at 313 K and underwent the least rapid rate of subsequent decomposition (table 1). Following the initial process (a, = 0.03) eqn (1) was obeyed with n = 2 up to a = 0.6. Behaviour thereafter resembled that of salt B (above) showing a positive deviation and a range of fit to eqn (1) with n = 3.Salt D . Digestion for 1.5 h at 333 K yielded the largest reactant crystals of the present series of salts, and these exhibited the slowest onset of reaction, although the rate subsequently attained was greater than those of salts C and B. Kinetic analyses were less straightforward because obediences to eqn (1) with n = 2 and 3 were almost equally satisfactory, and both fits extended over a wider range of a values than those of salts A-C. At lower temperatures the n = 3 obedience was preferred and the value of a, was lower (0.09) than that recorded in table 1. Salt E. The relatively most intense digestion conditions used here, 3.0 h at 338 K (table l), did not result in any notable increase in crystallite sizes or perfection.The decomposition rate was similar to that of the most reactive salt (A) but here the delay during initiation was relatively greater. This salt differed from the others of the series in that eqn (1) with n = 3 was obeyed over the greater part of the reaction. INFLUENCE OF ADDITIVES The mixtures used in these studies were prepared by crushing in a pestle and mortar known weights of silver malonate and the additive. Silver acetate. Admixture with silver acetate (10%) resulted in little, if any, modification of silver malonate decomposition at 466 K (near the median temperature of the present kinetic studies). This is consistent with the conclusion, based on n.m.r. analyses of partially decomposed salt, that silver acetate does not accumulate as an intermediate during the present reaction [unlike the copper(I1) salt].5 Silver oxalate.Mixtures containing this additive underwent a rapid first reaction6$ at 466 K, and on subtraction of this contribution the process remaining was identical with expectation for silver malonate decomposition. The components of the mixture therefore react independently, and the presence of the more reactive additive did not influence the course of breakdown of the present salt. Copper(@ malonate. This salt was of interest because recent5 evidence indicated the participation of a molten acetate phase as intermediate. However, kinetic comparisons showed that the components of silver and copper(r1) malonate mixtures decomposed independently. Gold. The incorporation of gold (50%) resulted in the virtual elimination of the induction period and the acceleratory process from silver malonate decomposition at 466 K, but the rate thereafter was little influenced by the additive.The noble metal is therefore identified as an effective nucleation promotor, probably behaving as the initial metallic particle, but it does not influence subsequent interface development where metallic silver is the active metallic participant. Potassium bromide. The decompositions of pelleted mixtures of salt A with 10, 15 or 20% KBr were ca. 80% complete in 60 min at 455 K, whereas under identical conditions decomposition of the pure salt A had proceeded only to a < 0.1. Component drying did not influence behaviour, and the addition of AgBr did not result in a comparable increase in decomposition rate.Potassium malonate, however, reacted rapidly at 455 K, reaching completion (a = 1 .O) after ca. 20 min. We therefore conclude that disproportionation in the mixture 2KBr + CH,(CO,Ag), + 2AgBr + CH,(CO,K), precedes breakdown of the CH,(CO,K), thus formed.2510 CRYSTOLYSIS OF SILVER MALONATE DISCUSSION REACTION GEOMETRY13 The observations from electron microscopy give direct evidence (plate 2) that the product is a coherent assemblage of opaque (silver metal) particles dispersed in a less dense (carbonaceous) matrix that maintains, throughout the reaction, the shapes and sizes of the original silver malonate reactant crystallites. We obtained no indication, from detailed examination of the micrographs, that reaction was accompanied by melting, and there was no intercrystailine aggregation, sintering or evidence of intracrystalline froth formation.l4 There was some surface roughening during reaction (plate 1). The local development of crystal surface areas having textural features of the product, while other zones remain unchanged, in partly decomposed salt [plate 2 (b)] provides direct confirmation that reaction proceeds by a nucleation-and-growth mechanism. Such growth of compact product assemblages without melting can therefore be regarded as an example of functional nuclei,l3 wherein the product silver actively catalyses salt breakdown. The microscopic evidence also shows that the nuclei are not spherical, but elongated, although the small sizes and irregular shapes of the opaque reactant crystallites precluded the possibility of directly characterizing the systematic changes in interface geometry during the growth process.Evidence supporting this conclusion is the ability of gold to eliminate the induction and acceleratory periods to silver malonate reaction. It is known that gold is a catalyst for formic acid breakdownf5 at 455 K, and it may therefore be expected that the metal surface promotes the breakdown of the malonate ion. Initial onset of decomposition on the added metallic gold may then be expected to yield product silver and subsequent development of normal growth nuclei. The nucleation-and-growth reaction mechanism is also entirely consistent with the observed kinetic obedience to the Avrami-Erofe'ev equation,2 based on this theoretical reaction model.This kinetic description was not, however, simple and generally applicable, since behaviour varied between the different reactants, and the linear plots were acceptable over various and incomplete a ranges. Plots of log [-log (1 -a')] against log t , using data from the present series of salts, tended to give fractional values to the exponent, n z 2.5. If it is assumed that the initial nucleation process is completed rapidly, perhaps during the initial (linear) reaction, a < 4 , then the main reaction may be ascribed to the growth of three-dimensional coherent but asymmetric (elongated) nuclei in the rectangular, lath-shaped crystals. For most preparations the rate data could be acceptably described as approximating to two-dimensional growth (n = 2, table 1) during the first part of reaction.For the more intensively digested preparations, expected to improve crystal perfection, three-dimensional growth (n = 3) provided the more acceptable representation of the kinetic observations. Kinetic behaviour is therefore perceptibly different between the alternatively pretreated reactants; it depends on crystallite sizes, shapes and perfection but usually can be regarded as approximating to two-dimensional growth. Thus both the microscopic evidence concerning the geometry of interface development and the pattern of kinetic obediences together confirm that the reaction proceeds by a nucleation-and-growth mechanism. The observed obedience to the first-order equation during the later stages of decomposition of salt A is ascribed to the continued advance of interface in the remaining fragments of this ill-crystallized material.Deceleratory behaviour is expected2 as the total area of the active reaction zone diminishes and progressively fewer crystallites remain undecomposed.J. Chem. SOC., Faraday Trans. 1, Vol. 81, part 10 Plate 1 Plate 1. Scanning electron micrographs of extensively decomposed crystals of silver malonate : there has been little if any melting, sintering, disintegration or cracking of the lath-shaped reactant crystallites, which remain intact throughout reaction. The only perceptible textural change is the roughening of surfaces, smooth before the onset of reaction. (a) (a = 0.85). Typical corners of the largely decomposed crystals.Scale bar 2.0 pm. (b) (a = 1.00). A section of a decomposed crystal, exposed on fracture during gentle crushing, showing a coherent product structure occupying the interior of the particle which remains pseudomorphic with the original reactant crystallites. Scale bar 2.0 pm. A. K. GALWEY AND M. A. MOHAMED (Facing p . 25 10)J . Chem. SOC., Faraday Trans. I , Vol. 81, part 10 Plate 2 Plate 2. Transmission electron micrographs of replicas of decomposed silver malonate. (a) (a = 1 .OO). Micrograph of decomposed, and subsequently crushed, salt remaining adhering to the replica. Two components can be distinguished : approximately spherical particles (most diameters 0.04-0.16 pm) opaque to the electron beam, identified as metallic silver, embedded in a matrix of less dense connective material, identified as the carbonaceous component of the residual product.Scale bar 0.2 pm. (b) (a = 0.30). Replica of an undamaged crystal. The textured zone, part of which had adhered to the replica, is identified as a nucleus, composed of a coherent and asymmetric region of product, extending along the crystal edge. Adjoining areas are apparently unreacted. This, taken with other similar micrographs, is direct evidence that decomposition proceeds by a nucleation-and-growth mechanism. Scale bar 1 .O pm. A. K . GALWEY AND M. A. MOHAMEDA. K. GALWEY AND M. A. MOHAMED 251 1 REACTION INTERFACE13 The constancy of the Arrhenius parameters calculated for all five reactant prepar- ations (salts A-E, table 1) is strong evidence that the decompositions are controlled by the same chemical steps at the reaction interfaces.The differences in rates illustrated in fig. 1 may therefore be ascribed entirely to differences in reaction geometry and reactant crystallite sizes. We conclude, from comparisons of the above results with the limited information available in the literature, that the decompositions of the present reactant and of silver acetate (443 K9 and 453 Ks) proceed in comparable temperature intervals and yield generally similar product mixtures (CO,, CH,CO,H, residual C and Ag). We identify the controlling steps at the reaction interface during silver malonate breakdown as those participating in the reactions of acetate ions adsorbed on metallic silver.l0 This comparison provides a mechanism that satisfactorily explains the present observations, regarding the surface bonded intermediates as present in low concentrations only and not detectable by n.m.r.analyses of solutions of partly decomposed salt. No discussion of the mechanism of decomposition of silver acetate could be found in the relevant literature. Barteau et al.,l0 working in a higher temperature interval than that of the present work, mention the complexity of the surface processes that contribute to the breakdown of acetate radicals chemisorbed on Ag( 1 10). We regard this as the most relevant information available upon which we can formulate a mechanism for the present reaction. Both rate processes yield a similar range of products, and are therefore expected to involve the same surface equilibria involving an identical range of adsorbed radicals and probably the same controlling parameters in product desorption.Such dynamic surface interactions, involving dissociated species, are familiar from mechanisms proposed in the field of heterogeneous catalysis but less frequently discussed with reference to interfacial processes in solid-state decompositions. Surface phenomena may be schematically represented as follows. malonate reactions similar t o those adsorption dynamic surface equilibria ion in acetate decomposition on Ag( 1 lo); 'The equilibrium established during silver malonate decomposition is expected to differ from that involving acetate because the quantity of hydrogen involved (2H per molecule) is less. This explains the absence of hydrogen and methane in the products.The proposed mechanism does not permit the identification of a single rate- controlling factor. However, it may be that the relevant promotional property of the metallic silver is that hydrogen transfer becomes appreciable at reaction temperature,16 thus permitting reorganization of the chemisorbed -CH2C02- groups. This provides a chemical explanation of the similarities of reactivity of silver malonate and of acetate on metallic silver yielding similar products. COMPARISON WITH OTHER SALTS Silver oxalute. The decomposition of silver oxalate, also a nucleation-and-growth rea~tion,~ proceeds in a lower temperature range (378-393 K)6 than that of the present salt (above ca. 450 K). This difference may be due to one (or more) of the following2512 CRYSTOLYSIS OF SILVER MALONATE effects.No hydrogen-transfer step is required during breakdown of the surface intermediate species. Reaction does not yield residual carbon, expected to deactivate the metal surface. The methylene group prevents interaction between the carboxyl groups in the malonate, whereas cooperative electronic interactions may participate in the relatively more rapid breakdown of the oxalate. This latter effect may also be the reason for the relatively lower sensitivity of silver malonate to photolysis. Other metal malonates. There were significant differences between the present reaction and those of the other malonates for which mechanistic studies have been reported. Copper(r1) malonate decomposes in a melt with acetate accum~lation.~ Decomposition of nickel malonate3 is a nucleation-and-growth process that occurs in a higher temperature interval (543-6 13 K) than the present reaction to yield mainly metallic carbide and a relatively smaller proportion of acid. The decomposition of cobalt malonate4 (572-600 K) was sensitive to the gases present, and although nucleation-and-growth behaviour was indicated, the kinetic characteristics were complicated.THE INITIAL REACTION The different extents of the initial reaction (q) could not be correlated with any feature of reactant composition, texture or behaviour. It seems most probable, therefore, that this rate process represents the decomposition of a superficial layer of the reactant crystals. There was some indication, from microscopic examinations, that this boundary layer possessed a texture that was slightly different from that within the decomposed crystallites.Such an initial process may then represent the precursor to nucleation, and the metal product constitutes the germ nuclei for the main decomposition. This is suggested to explain the relatively rapid inception of reaction following reactant heating in kinetic studies. We thank Mr J. McCrae and Mr R. Reed for help and advice in obtaining the electron micrographs. M. A. M. thanks the Egyptian Government and the O.R.S. Award Scheme for Scholarships held during the period of this work. N. J. Carr and A. K. Galwey, Thermochim. Acta, 1984,79, 323. M. E. Brown, D. Dollimore and A. K. Galwey, Comprehensive Chemical Kinetics, Vol. 22: Reactions in the Solid State (Elsevier, Amsterdam, 1980). K. A. Jones, R. J. Acheson, B. R. Wheeler and A. K. Galwey, Trans. Faraday Soc., 1968, 64, 1887. A. K. Galwey, D. M. Jamieson, M. Le Van and C. Berro, Thermochim. Acta, 1974, 10, 161. N. J. Carr and A. K. Galwey, Proc. 10th Int. Symp. Reactivity of Solids, Dijon (Elsevier, Amsterdam, 1984), p. 697. A. Finch, P. W. M. Jacobs and F. C. Tompkins, J. Chem. Soc., 1954, 2053. ' A. G. Leiga, J. Phys. Chem., 1966,70, 3254. M. D. Judd, B. A. Plunkett and M. I. Pope, J . Thermal Anal., 1974,6, 555. P. Baraldi, Spectrochim. Acta, 1982, %A, 51. lo M. A. Barteau, M. Bowker and R. J. Madix, J . Catal., 1981, 67, 118. 11 A. K. Galwey and M. A. Mohamed, Proc. 10th Int. Symp. Reactivity of Solids, Dijon (Elsevier, l2 M. J. McGinn, B. R. Wheeler and A. K. Galwey, Trans. Faraday SOC., 1971, 67, 1480. l3 A. K. Galwey, Proc. 7th Int. Conf. Thermal Analysis, Kingston, Ontario (Wiley, New York, 1982), l4 A. K. Galwey, L. Poppl and S. Rajam, J. Chem. Soc., Faraday Trans. I , 1983, 79, 2143. l5 J. Fahrenfort, L. L. Van Reyen and W. M. H. Sachtler, The Mechanism of Heterogeneous Cutalysis l6 C . Kemball and C. T. H. Stoddart, Proc. R. Soc. London, Ser. A, 1957, 241, 208. Amsterdam, 1984), p. 699. p. 38. (Elsevier, Amsterdam, 1960), p. 23. (PAPER 4/2 193)

ISSN:0300-9599    

DOI:10.1039/F19858102503

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I, 1985,81, 2513-2523 A Thermodynamic Study of the Preferential Interaction of the Polyoxometallate Electrolyte Na,SiW,,O,, 14H20 with Ethers in Aqueous Solution BY MICHEL FROMON, CLAUDE TREINER* AND RAYMOND BURY Laboratoire d'Electrochimie, ERA 3 10, Universite Pierre et Marie Curie, 4 Place Jussieu, Paris 75005, France AND MICHEL FOURNIER Laboratoire de Physico-chimie Inorganique, ERA 608, Universite Pierre et Marie Curie, 4 Place Jussieu, Paris 75005, France Received 3 1st December, 1984 Free-energy and enthalpy changes associated with the interaction of a sodium silico-tungstate salt (Na,SiW,,O,, * 14H,O) with diethylether and tetrahydrofuran in dilute aqueous solutions ( < 1 mol kg-l ether) have been investigated using precise vapour-pressure and microcalorimetry techniques.The standard partial molar volume of the oxometallate salt in pure water at a temperature of 298.15 K has been determined from density measurements in dilute solutions: 5 = 660.5 cm3 mol-I. It is shown that the solvation process is enthalpy driven in very dilute salt or ether solutions and entropy driven at higher ether concentrations; the enthalpy/entropy compensation leads to a decreasing standard function of transfer of the electrolyte from water to aqueous ether solutions. The available data are compatible with a two-step solvation process: first ether molecules solvate the oxometallate anion at given polarized sites and secondly, with increasing ether concentration, a favourable hydrophobic type of interaction takes place between the solvated anion and free ether molecules.An analysis of the salting constant of tetrahydrofuran in the aqueous sodium silico-tungstate solution using the scaled-particle theory is also presented. Early transition metals such as tungsten, molybdenum and vanadium may form polynuclear oxometallates complexes (Keggin structures).l? There has been much progress in the elucidation of their crystallographic structures in recent years and they are finding various applications, e.g. in biology as non-specific electron-dense strains for electron microscopy and other problems of top~chemistry,~ as a precipitant for polymer solutions4 and as a Hill oxidant5 for various catalytic reactions. However, little is known about the physico-chemical properties of these polyanions in solution.For example it was long ago recognized that the acids interact favourably with ethers,2 and this property has been used in extraction and separations processes, but the origin of this interaction is unkn0wn.l Thus we decided to investigate some thermodynamic properties associated with the interaction of ethers with a recently synthesized polyelectrolyte: sodium polytungstate, Na,SiW,,O,,, * 14H20 (NaSiW). This salt is fairly soluble in water and its crystallographic parameters have recently been discussed.6 We have restricted our study to the dilute ether concentration range in order to be able to compare the properties of these solutions to those of more classical electrolyte solutions, which have been the subject of an extensive study' with a cyclic ether, tetrahydrofuran (THF).In particular, we suggested the conditions under which the statistical-mechanical theory known as the scaled-particle theory (SPT)8 could be 25132514 INTERACTION OF POLYOXOMETALLATE WITH ETHERS used to calculate the salting constant, k,, of neutral polar molecules in aqueous solutions of inorganic and organic electrolytes with various electric charges. The heteroanion, which possesses spherical symmetry and is relatively large, could be a good model electrolyte for theoretical calculations. In effect, SPT calculations have made use of organic ions such as those of the tetra-alkylammonium series for testing the influence of size on salting constants; however, these entities can hardly be considered as hard spheres and moreover their hydrophobic properties introduce additional complications to the analysis of the SPT approach to ion +molecule interactions in aqueous sol~tions.~t 9 3 lo The salting constants of THF and diethyloxide (ET20) have been measured in the presence of NaSiW using a precise vapour-pressure technique; from these data the standard free energy of transfer of NaSiW from water to the binary aqueous solution can be calculated.The partial molar volume of NaSiW in pure water has been determined, as it is required for the SPT calculation, by density measurements. Finally, the enthalpy of solution of THF in dilute aqueous electrolyte solutions has been determined by microcalorimetry, as has the enthalpy of solution of the electrolyte in aqueous THF solutions.The combined free-energy and enthalpy data enable the entropy function to be calculated. The essential observations of this preliminary investigation are as follows. (i) The inorganic electrolyte behaves with THF and ET20 as expected, viz. there is a preferential interaction between the heteroanion and the ethers; however, the variation of k , with ether concentration is the opposite to that obtained for a classical complexation effect. (ii) The standard free energy of solution of NaSiW is enthalpy controlled in the very dilute ether concentration range and becomes rapidly entropy controlled above 0.1 mol dm-3 of salt. This result suggests a two-step solubilization process. (iii) The SPT calculation is not applicable to the case of the inorganic heteroanion studied.EXPERIMENTAL MATERIALS The method of preparation was as follows. 400 g of Na,WO,. 2H,O was dissolved in 600 cm3 of hot water (80 "C) and the solution was then heated to its boiling point. 22 g of solid sodium metasilicate (Prolabo) were poured into the solution, 180 cm3 of HCI (12 mol kg-l) were then added drop by drop and the solution was boiled for 2 h. From the clear, purplish-blue solution, 250 g of colourless crystals crystallized (yield 60%). The degree of hydration, h = 14, was checked by thermogravimetric analysis. The purity of the salt (molecular weight 3218) was controlled by polarography (three reduction waves of, respectively, one, one and two Faradays and reduction potentials of -0.260, -0.520 and -0.950 V 1's. ECS in a 1 mol kg-' buffer solution of acetic acid and sodium acetate at pH 4.65).The compound was easily identified by infrared spectrometry using KBr pellets.6 THF and ET,O were reagent-grade compounds (Merck and Prolabo, respectively) and were used as received. METHODS The free-energy data for THF and ET,O were obtained from a precise vapour-pressure technique, which has been described in detail previously.', l 1 A vapour-pressure run may be summarized as follows. A water + ether solution was thoroughly degassed by four cycles of freezing with liquid nitrogen, pumping and melting. Five Pyrex cups containing a known amount of salt were then added under vacuum into the degassed solution and the total pressure measured with a Texas Instruments pressure gauge. The enthalpy data were determined using a ' bioflux' type of calorimeter (Thermanalyse). The experiments were performed only for THF as the high vapour pressure of ET,O introduced relatively large uncertainties.The heat of solution of THF was obtained by injecting the solute, using a thermostatted and calibrated microsyringe (Hamilton), into I5 cm3 of the aqueous electrolyte solution. (Note that the highM. FROMON, C. TREINER, R . BURY AND M. FOURNIER 2515 Table 1. Standard free-energy data for the systems ether + NaSiW + H20 at 298.15 K" x2 aG",/dX, AGF(W + N) (mole fraction) k , /kJ mol-' /kJ mol-l 0.005 08 0.007 26 0.009 98 0.015 44 0.0 17 04 0.001 25 0.004 07 0.007 01 0.009 53 THF - 0.92 85+ 1 - 1.76 -0.97 88+3 -2.51 -1.04 91&2 - 3.56 -1.23 l l l f 2 - 5.82 -1.41 119&2 -6.61 Et,O - 0.9 1 - 1.05 -1.11 - 1.31 " Index 2 refers to the organic component and 1 to water (XI + X, = 1).vapour pressure of diethylether was an advantage in the vapour-pressure work as it enabled determination of the salting constant of the solute down to lower solute concentrations than for THF). The standard heat of dissolution of NaSiW was obtained using another microcalorimeter by adding the solid salt in Pyrex cups to a 250 cm3 Dewar at various THF concentrations. The precision on the enthalpy of dissolution of the salt as judged by the reproducibility of the results is of the order of k0.20 kJ mol-l. The reproducibility was better for the enthalpy of solution experiments with THF added to the electrolyte solution. All experiments were performed at 298.15 K.The density measurements were obtained using a seven digit Van Paar instrument with a specially designed thermostat (Setaram) for the oscillatory density cell. RESULTS From the variation of the total pressure with the salt concentration for a given aqueous ether solution, three related quantities can be obtained. (i) The salting constant is directly deduced from the vapour-pressure measurements using the definition (1) where Pf and PN are the partial pressures of the volatile cosolute in the absence and in the presence, respectively, of electrolyte of molality mE. The mode of calculation of these partial pressures has been described previ0usly.~9 l1 (ii) The standard free energy of transfer, AG:(W --+ E), of the non-electrolyte, N, from water to the aqueous electrolyte solution, E, may be obtained' from the relation k, = l/mE log ( P f l f " ) AG:(W + E) = 2.303RTkS mE.( 2 ) (iii) Conversely one may calculate from the same vapour-pressure data the rate of change of the standard chemical potential of the electrolyte, FEE, with respect to ether concentration, dc",/dX,, where X , is the mole fraction of water,12 and obtain by integration13 the standard free energy of transfer of the electrolyte from water to the different aqueous ether solutions, AG,"(W -+ N). Table 1 presents the three thermodynamic quantities calculated from the vapour- pressure data for THF and ET,O and tables 2 and 3 present the enthalpy data for2516 INTERACTION OF POLYOXOMETALLATE WITH ETHERS Table 2. Standard enthalpies of dissolution of NaSiW in aqueous THF solutions at 298.15 K and the standard enthalpy of transfer of the salt from water to the aqueous THF solutions ( X 2 ) x2 A@, A e ( W + N ) (mole fraction) /kJ mol-l /kJ mol-l 0.0 0.000 45 0.001 50 0.002 97 0.003 42 0.003 69 0.004 50 0.006 22 0.012 43 59.65 58.80 52.70 49.60 54.70 59.10 62.65 66.00 67.30 0.0 - 0.85 - 6.95 - 10.05 - 4.95 - 0.55 3.00 7.20 7.65 Table 3.Standard enthalpies of solution of tetrahydrofuran in aqueous solutions of NaSiW at 298.15 K and the standard enthalpy of transfer of THF from water to the aqueous salt solutions ( M ) M AX AH,(W -+ E) /mol kg-l /kJ mo1-I /kJ mol-l 0.0 0.0025 0.0050 0.0075 0.0 100 0.0 125 0.0150 0.0200 0.0248 0.0500 - 15.03 - 15.30 - 16.52 - 16.98 - 17.50 - 16.02 - 15.60 - 15.01 - 14.72 - 13.70 0.0 - 0.27 - 1.49 - 1.95 - 2.47 - 1.00 -0.57 - 0.02 0.3 1 1.33 THF.Note that for the enthalpy of solution of THF, the final concentration was of the order of 0.0024 mol kg-l. At higher THF concentrations the minimum shown on fig. 1 might not be found, which means that the conditions of infinite dilution necessary for the determination of standard quantities are not met. In the case of the heat of dissolution of the salt, the final concentration was of the order of 0.0012 mol kg-l. The partial molal volume of NaSiW in pure water was obtained from the classical definition and the standard quantity deduced from the relation V = V,+S,C1J2+bC (3) where S, is the Debye-Hiickel term (equal to 59.07 1 cm3 m ~ l - ~ ' ~ at 298.15 K in water). The experiments were performed in the dilute concentration range (see table 4) and were not of enough precision to determine accurately the square-root term of eqn (2), possibly because of some slight hydrolysis of the salt.A least-squares fit of the data gives b = -80.8 cm3 mo1-2 dm3 and = 660.5 with a standard fit of 0.9. Knowing the value for the sodium ion in waterI4 one can deduce the value relative to the oxyanion in water: ro = 688.9 cm3 mol-l.M. FROMON, C. TREINER, R. BURY AND M. FOURNIER 2517 'H F I I I 0 0.5 1.0 rn / m 01 kg- Fig. 1. Plot of the variation of the salting constant with cosolute molality. 0, THF and Q Y Et,O- Table 4. Density of aqueous solutions of NaSiW at 298.15 K M d /kg rnol-1 /cm3 rnol-1 0.0 0.997 070 0.003 73 1.006 549 0.006 79 1 .O 14 284 0.01 1 67 1.026 566 0.016 44 1.038 586 0.01 8 65 1.043 985 0.020 73 1.049 105 0.023 83 1.056 748 0.032 00 1.076 808 0.041 76 1.100 351 0.047 85 1 .1 14 902 0.063 74 1.152 508 DISCUSSION SALTING EFFECT OF OXYANIONS ON ETHERS IN WATER The salting constant of a neutral molecule is usually defined at the limit of infinite dilution with respect to the concentration of that compound. The apparent k, values of fig. 2 may be easily extrapolated to zero ether molality. A common value of - 0.85 f 0.03 is obtained for both ethers. It has been shown15 that the salting constant is proportional to a free-energy pairwise interaction parameter; hence one may deduce from the present results that at the limit of zero concentration the strengths of the interaction force between the oxyanion and either ethers are alike, but that at non-zero2518 INTERACTION OF POLYOXOMETALLATE WITH ETHERS m,/mol kg-' 0.0 5 I 2.0 1.0 0 .1.0 .2.0 - I - 2 - Y \ 5 I 0.5 m /mol kg-' Fig.2. Plot of the variation of the standard enthalpy of transfer of (8) THF from water to aqueous NaSiW solutions and of (0) NaSiW from water to aqueous THF solutions. concentrations triplet and higher-order aggregates form rapidly and are much stronger with diethylether than with the cyclic THF molecule. Salting-in constants are usually associated with non-specific or weak interactions between non-electrolyte and electrolyte solutes in water. There is no need for such a restriction. We have recently determined the limiting k, value of acetonitrile in aqueous silver nitrate solutions.l1 This is a well documented case of weak complexation involving the silver ion and acetonitrile.We obtained a value of - 0.90 0.02, very close to the valuecalculated for the oxyanion +ether interaction; thisexample provides information on the order of magnitude of the strength of the interaction. A fundamental difference is, however, observed : k, increases (becomes less negative) with acetonitrile concen-M. FROMON, C. TREINER, R. BURY AND M. FOURNIER 2519 tration in the presence of silver nitrate, as expected for a complexation phenomenon, whereas with the present system the salting constant becomes more negative with increasing ether concentration. No simple model of ligand + ion equilibrium can explain this behaviour. We shall come back to this point when we consider the enthalpy data.Consideration of the prediction of the SPT calculation is interesting. We recall that the SPT enables calculation of the salt effect associated with the formation of a cavity the size of the solute molecule in a medium characterized by the size of the solvent molecules and ions and the density of the solvent. An interaction term16 is added to the cavity term in order to take partly into account the soft part of the interaction potential (essentially as a Lennard-Jones potential). This is the only theory which enables us to evaluate the non-specific contribution to the standard free energy of interaction and hence it provides an estimate of the specific contribution in terms of solvation or complexation effects. This problem has been the subject of much debate in recent years.17 Some authors have pointed out the inadequacy of the theory'? 8* lo in the case of large organic ions such as those of the tetra-alkylammonium series; however, these ions are responsible for a so-called hydrophobic effect, which means that the entropy of interaction is an important parameter for these entities in water; this effect is not taken into account in the interaction term associated with the cavity term of the SPT.Also, the organic ions of the tetra-alkylammonium series possess an ill-defined shape, which makes them poor model compounds for theoretical calculations. These problems are avoided with the present salt. It has spherical symmetry and the diameter of the oxyanion, as determined by X-ray crystallography, is well characterized : d = 10.42 _+ 0.04 A.6 Using the equations provided in the literature18 with the hard-sphere diameter of THF taken as 5.1 1 A and using the value determined above, one gets a cavity term of - 1.77 (mole-fraction scale); this value is much more negative than the experimental value of -0.85 obtained by extrapolation from fig.2, which means that the interaction term (or rather its derivative with respect to salt concentration) obtained by difference would be positive. The same qualitative result was obtained previously with organic ions.7$ lo As this result seems questionable on physical grounds we suggest that there is a difficulty associated with the use of the SPT when large entities are involved regardless of their chemical nature. However, this conclusion also implies that the specific contribution to k, arising from ion + ligand interactions is not very important in free-energy terms.This conclusion is further supported by the following considerations. Most theories of the salt effect predict a dependence of k, upon the standard partial molar volume of the salt; hence using all the available data on THF in aqueous electrolyte solution7 and making use of the tetraphenylboron a~sumption,'~ which permits the evaluation of the single-ion contribution to the total salt effect, we have constructed fig. 3. The values have been taken from the classical review by Mi1ler0.l~ A few electrolytes have been designated but one should refer to the original paper for completeness. A line has been drawn through the cation series and another one through the anion series (with the notable exception of the tetraphenylarsenium ion).The heteroanion seems to be well situated on the latter line. This observation suggests again that the contribution of specificeffects is not detectable when based only on standard free-energy data; it also means that although the negative charge seems to be insulated from the surrounding medium, its effect is still sensitive enough for the oxyanion behaviour to be non-specifically but distinctly different from that of a cation of like size. Enthalpy data allow us to shed some light on the interaction between the anion and the ether molecules .2520 + 0.5 -1.0 INTERACTION OF POLYOXOMETALLATE WITH ETHERS I I I I '0 0 500 800 V,,o/cmS mol-* Fig. 3.Plot of the dependence of the salting constant of THF in single-ion aqueous solutions on the standard partial molar volume of these ions in water: a, organic ions and 0, inorganic ions. ENTHALPY AND ENTROPY DATA We have indicated that experimental problems related to the high volatility of ET20 prevented its use for very precise microcalorimetry work. However, the similarity in the salting constant profiles of fig. 2 for ET,O and THF indicates that any conclusion attained for THF should be qualitatively applicable to ET20. Addition of small quantities of either THF to the aqueous salt solution or of salt to the THF solution displays the same qualitative behaviour (in a more ideal case the derivative of the two functions with respect to either solute should coincide at vanishingly small con~entrationsl~).A decrease of the enthalpy of solution is first observed, which means that the standard enthalpy of transfer of either solute, A%, is negative from water to the aqueous solution and becomes positive at larger solute concentrations (ca. 0.02 mol dm-3 salt or 0.2 mol dm-3 THF). Clearly a specific interaction occurs between the oxyanion and ether molecules in a very small concentration range. This interaction is not detectable on the free-energy curve, as shown by the smooth variation obtained especially for ET,O in the whole range of composition studied overlapping the concentration range covered by the enthalpy data The rate of change of the standard free energy of transfer of the electrolyte from water to the aqueous binary solution with respect to THF concentration (fig.4) is the result of a compensation effect between enthalpy and entropy, which arises particularly in aqueous solutions. In the very low THF concentration range, the (fig. 2).10.0 - 5.0 - M. FROMON, C. TREINER, R. BURY AND M. FOURNIER 252 1 0.0 / 0.005 0.010 0.015 X, (mole fraction) Fig. 4. Plot of the standard thermodynamic functions of transfer of NaSiW from water to aqueous THF solutions against the mole fraction of THF. solvation of the salt is enthalpy driven but becomes entropy driven at higher THF concentrations. This latter observation is typical of what is observed in aqueous binary solvents for large organic and is interpreted as evidence of hydrophobic behaviour. Note, however, that the standard free energy of transfer of urea from water to dilute aqueous THF solutions is negative:24 it is also an entropy-driven process although urea is evidently not a hydrophobic molecule.Structural considerations of the heteroanion studied may help to suggest a model of interaction between the anion and the medium. It has been suggested6* 25 that the most appropriate representation of the oxyanion would be to consider the structure2522 INTERACTION OF POLYOXOMETALLATE WITH ETHERS (SO:-) (W12036), in which a mononuclear oxyanion, SiOi-, is surrounded and protected by a neutral metal oxide cage, W,,O,,. Thus the anion has a very low negative surface charge density. However, this model does not exclude the possibility of strong polarization effects between the tungstate metal and the exterior oxygen atoms, as advocated by Baker et which could cause the heteroanion to accept lone pair electrons.The simplest chemical model which takes into account the new information is then as follows. A rather weak specific ion-dipole interaction occurs in dilute solutions between a small number of ether molecules and the anion surface at some given polarized sites; with further addition of cosolvent the solvating ether molecules begin to interact with free ether molecules through hydrophobic interactions giving rise to the entropy-driven process. This model interprets the enthalpy results. In free-energy terms it implies that the ion-dipole interactions between the anion and the ether molecules are weaker than the hydrophobic forces between the large solvated anion and the free THF molecules in order to explain the accelerating rate of change of the decreasing standard free energy of transfer of the electrolyte with addition of ether.This model should apply to those molecules which are both polar and hydrophobic, but should not be entirely valid for hydrophilic molecules. For example, the question has been raised as to the origin of the protective effect of some organic solvents such as dimethylsulphoxide and glycerol on the photoreduction of silicomolybdate ions in the case of tobacco chl~roplates.~ It has been noted2’ that most polar organic molecules interact preferentially with positive ions (or in the present case with positively charged sites) because of their more readily available negatively charged dipole end.The present study suggests that the protective effect is the consequence of the solvation of the oxyanion by these organic molecules. However, hydrophilic molecules such as dimethylsulphoxide or glycerol interact strongly with water molecules through hydrogen bonds and are not able to participate in hydrophobic interactions. Thus in these cases only the enthalpy-driven process should be observed. M. T. Pope, Heteropoly and Isopolyoxometales (Springer Verlag, Berlin, 1983). P. Souchay, in Ions mintraux condensts (Masson, Paris, 1969). J. E. Scott, J. Hisrochem. Cyrochern., 1971,19,689; S. Tsuji and H. S. Alameddine, Hisrochem., 1981, 73, 33; S. Tsuji and M. Fournier, Hisrochem., 1984, 80, 19. T. Nemetchek, H. Riedl and R. Jonak, J.Mol. Biol., 1979, 133, 67. J. P. Berg and S. Isawa, Biochim. Biophys. Acta, 1977,460, 206. C. Rocchiccioli-Deltcheff, M. Fournier, R. Franck and R. Thouvenot, Inorg. Chem., 1983, 24, 207. M. Fromon and C. Treiner, J. Chem. Soc., Faraday Trans. I , 1979, 75, 1837. R. A. Pierotti, Chem. Rev., 1976, 76, 717 and references therein. A. Ben-Naim and R. Tenne, J. Chem. Phys., 1977,67, 627. lo I. Sanemasa, K. Haragushi and H. Nagai, Bull. Chem. Soc. Jpn, 1981, 54, 1040. l 1 C. Treiner and M. Fromon, J. Chem. SOC., Faraday Trans. I , 1980,76, 1062. l 2 E. Grunwald and A. L. Bacarella, J. Am. Chem. SOC., 1958, 80, 3840. l 3 C. Treiner, J. Chim. Phys., 1973, 70, 1183. l4 F. J. Millero, Chem. Reg., 1971, 71, 147. l5 J. E. Desnoyers, M. Billon, S. Leger, G . Perron and J. P. Morel, J. Solution Chem., 1976, 5, 681. l 7 S. D. Christian and E. H. Lane, in Solutions and Solubilities, ed. M. J. Dack (Wiley, New York, 1975), S. K. Shoor and K. E. Gubbins, J. Phys. Chem., 1969, 73, 498. vol. 7. W. L. Masterton and T. P. Lee. J. Phys. Chem., 1970, 74, 1776. l9 E. Grunwald, G. Baugham and G. Kohnstam, J . Am. Chem. SOC., 1980, 102, 5801. 2o E. M. Arnett, in Physico-chemical Processes in Mixed Aqueous Solvents, ed. E. D. Covington and 21 R. K. Mohanty, T. S. Sarma, S. Subramanian and J. C. Ahluwalia, Trans. Faraday Soc., 1971, 67, P. Jones (Heineman, London, 1967). 305.M. FROMON, C. TREINER, R. BURY AND M. FOURNIER 2523 22 C. De Visser and G. Gomsen, J . Chem. Thermodyn., 1972, 4, 313. 23 R. Bury and C. Treiner, J . Chem. Soc., Faraday Trans. I , 1982, 78, 1827. 24 C. Treiner and P. Tzias, J . Solution Chem., 1975, 4, 471. 25 K. C. R. Ho, Ph.D. Thesis (Columbia University, 1979). 26 L. C. Baker, L. Lebioda. J. Grochowski and H. G. Mukherjee, J . Am. Chem. Soc., 1980, 102, 3274. 27 C. Treiner, Can. J. Chem., 1981, 59, 2518. (PAPER 4/2195)

ISSN:0300-9599    

DOI:10.1039/F19858102513

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. Soc., Faraday Trans. I , 1985, 81, 2525-2539 Adsorption of Methylamines on Dehydrated NaX, NaY and KY Zeolites BY KUNIMITSU MORISHIGE,* SHIGEHARU KITTAKA AND SHINICHIRO IHARA Department of Chemistry, Okayama University of Science, 1-1 Ridai-cho, Okayama 700, Japan Received 2nd January, 1985 The isosteric heats of adsorption (qst) of methylamines over dehydrated NaX, NaY and KY zeolites have been determined by measuring the adsorption isotherms. The qst curves for NaX differ remarkably in shape from those for Nay, although the gross features of the qst curves for NaY resemble those for KY. For NaX, site 111’ cations act as the most selective and energetically heterogeneous sites for the adsorption of methylamines, while for NaY and KY methylamines are preferentially adsorbed on site I1 cations.The basicity of methylamines increases in the order NH, < MeNH, < Me,NH < Me,N for NaY and NH, < MeNH, x Me,N < Me,NH for KY. The potential energy of amine molecules on site I1 cations of NaY and KY was calculated on the basis of an electrostatic-interaction model in which the repulsion term results from a configuration in which the dipole axis of the molecule coincides with the line joining the site I1 cation and the centre of the supercage. Despite the smaller size of the Na+ ion, the magnitudes of the electrostatic term for both zeolites are nearly the same, whereas the dispersion and repulsion interactions of an adsorbed molecule with the oxygen ions surrounding the site I1 cation are larger for Nay. It is concluded that the different orders of the methylamines’ basicity over NaY and KY are due to the different environments of the site I1 cations.The systematic dependences of qst on the degree of adsorption have also been observed and analysed by comparison with the calculated energies of the lateral interaction between adsorbed molecules. The order of the basicities of methyl-substituted amines over solid acids is of considerable interest because it contains information on the bond structure and environment of the acid-base interactions of amines with acid sites. In a previous paper’ we reported the thermal desorption and infrared spectra of methylamines adsorbed on dehydrated alkaline-earth-metal X zeolites. The dependence of the desorption-peak temperature on the radius of the exchanged cation was explained by the conclusion that methylamines are selectively adsorbed on site-I1 cations in the supercages through electrostatic interaction. NaX and NaY zeolites are very popular as adsorbents and as starting materials for the other ion-exchanged forms.In this report we examine the order of the basicities of methylamines over dehydrated NaX, NaY and KY zeolites by measuring adsorption isotherms. The isosteric heats of adsorption obtained are discussed by comparing them with the adsorption-potential energies calculated on the basis of the electrostatic- interaction model and the lateral interaction energies calculated using the Stockmayer potential for the interaction between polar molecules. 25252526 ADSORPTION OF METHYLAMINES ON ZEOLITES EXPERIMENTAL MATERIALS Linde NaY(SK-40) and NaX(SK-20) were purified by several cycles of immersion in a 1 mol dmP3 solution of CH,COONa and subsequent decantation. KY was prepared from NaY according to the procedure of Mortier and Bosmans.* Chemical analysis showed that the compositions of anhydrous NaX and NaY were Naa6Ala6Silo603a4 and Na55A155Si13,0384, respectively.The methylamines were purified according to procedures described previously.' ADSORPTION ISOTHERMS Adsorption isotherms were measured using a Sartorius vacuum microbalance equipped with a capacitance manometer (Datametrics, 100 mbar full scale). A 70 mg sample was evacuated for 2 h under a vacuum of 2 x mmHgt at 30Q "C. After the dehydrated sample had been exposed to amine at 2&30mmHg pressure at the adsorption temperature for 1 h, the adsorption isotherms were measured by decreasing the equilibrium vapour pressure stepwise (on the desorption side).Desorption equilibrium was usually attained within 10 min. The temperature range of the measurements depended on the combination of zeolites and amines in question: e.g. 70-1 10 "C for KY-NH, and 14@-180 "C for Nay-Me3N. Both the balance tubes (sample and reference) were immersed in an oil bath whose temperature was regulated within f0.2 "C throughout each run. RESULTS AND DISCUSSION ISOSTERIC HEAT OF ADSORPTION The adsorption isotherms of ammonia (NH,), methylamine (MeNH,), dimethyl- amine (Me,NH) and trimethylamine (Me,N) over dehydrated NaX, NaY and KY zeolites were measured. Fig. 1 shows the adsorption isotherms of NH, over NaX, NaY and KY as representative examples.These were reversible and of type I in Brunauer's classification. However, the curves do not follow the Langmuir isotherm quantitatively in any of the systems investigated. The isosteric heat of adsorption (qst) was determined by applying the Clausius- Clapeyron equation to the isotherms. Fig. 2, 3 and 4 show the qst curves of the methylamines over NaX, NaY and KY, respectively. The curves for NaX differ remarkably in shape from those for Nay, while in terms of gross features the qst curves for NaY resemble those for KY. This suggests that the type of adsorption sites for methylamines over alkali-metal X zeolites differs from that over alkali-metal Y zeolites. The qst curve of NH, over NaX, which is in good agreement with the result of Barrer and Gibbons,, is typical of heterogeneous surfaces, where qst decreases steeply with increasing degree of adsorption.A plateau due to adsorbate-adsorbate interaction appeared for other adsorbates over NaX, in addition to an initial decrease of qst. The height of the plateau increased in the order MeNH, < Me,NH < Me,N. In contrast, qst for methylamines over NaY and KY did not change greatly until a given degree of adsorption (the saturation adsorption) was approached, and then decreased steeply. As the degree of adsorption increased, qst of NH, decreased gradually, while qst of Me,NH increased. The qst of MeNH, and Me,N were almost independent of the degree of adsorption up to saturation. These facts strongly suggest that NaY and KY have energetically homogeneous sites for the selective adsorption of amines and that the molecules adsorbed interact with each other.The saturation degrees of adsorption over NaY were ca. 3 molecules per supercage, independent of the choice of methylamine. Over KY, the saturation degree of adsorption of Me,N T 1 mmHg = 13.5951 x 980.665 x lo-* Pa.K. MORISHIGE, S. KITTAKA AND S. IHARA 2527 - c - I M M 2F I - 0 I I 1 \\\L4 I I 1 I Fig. 1. Adsorption isotherms of NH, on (a) NaX, (b) NaY and (c) KY. Adsorption temperatures are (1) 159.3; (2) 170.4; ( 3 ) 180.5 and (4) 190.3 "C for (a), (1) 67.9; (2) 78.6; (3) 88.8; (4) 100.2 and (5) 109.3 "C for (b) and ( 1 ) 67.6; (2) 77.5; (3) 88.4; (4) 99.6 and (5) 108.4 "C for (c).(ca. 2 molecules per supercage) was surprisingly low compared with those of the other amines (ca. 5 molecules per supercage). The value of qst over NaY was larger than that over KY for a given methylamine. The standard energy of adsorption (qg), which is obtained by extrapolating a plot of qst against degree of adsorption to the zero degree of adsorption, increased in the order NH, < MeNH, < Me,NH < Me,N over NaY and NH, < MeNH, = Me,N -= Me,NH over KY. These orders of the basicity of the methylamines were confirmed by thermal-desorption experiment^.^ Type Y zeolites have essentially the same framework structure as that of type X zeolites, but the former are richer in silicon content than the latter.2q596 The exchangeable cations in alkali-metal Y zeolites are located on three different kinds of sites denoted as I, I' and I12p6 in fig, 5.However, since the total number of exchangeable cations in type X zeolites is greater than that in type Y , a certain number of the exchangeable cations in alkali-metal X zeolites cannot be accommodated in these sites and must be distributed over various sites on the walls of the supercage, collectively referred to as site III'., A far-infrared study' of alkali-metal X zeolites has clearly revealed that site-111' cations are more loosely bound to the walls of the2528 ADSORPTION OF METHYLAMINES ON ZEOLITES degree of adsorption (molecules per cage) I I 0 I 2 3 4 5 I I 20 I4 e re I 8 \ e” 4L I6 I 1 2 0 I 3 1 degree of adsorption/mmol g-’ Fig. 2. Isosteric heats of adsorption of the methylamines on NaX: 0, NH,; A, MeNH,, 0, Me,NH and 0, Me,N.degree of adsorption (molecules per cage) 0 I 2 3 4 I I I 5 - 16- - E“ - Q - Y > e” H- degree of adsorption/mmol g-l Fig. 3. Isosteric heats of adsorption of the methylamines on Nay. Symbols as fig. 2.K. MORISHIGE, S. KITTAKA AND S. IHARA degree of adsorption (molecules per cage) 0 I 2 3 4 5 6 I I I I I I I2t 1 0 L 1 I I 2 3 degree of adsorption/mmol g-I % Fig. 4. Isosteric heats of adsorption of the methylamines on KY. Symbols as fig. 2. 2529 Fig. 5. Schematic diagram of 10 cubo-octahedra surrounding the site-I1 cation (SII) and an indication of the cation sites. supercage than site I1 cations and that polar molecules such as pyridine interact with site-111’ cations more strongly than with site I1 cations.Therefore, the site-111’ cations are expected to be the most selective and energetically heterogeneous sites for the adsorption of polar molecules. As fig. 2 shows, the qst curves of the methylamines over NaX are entirely consistent with this expectation.2530 ADSORPTION OF METHYLAMINES ON ZEOLITES The site-I1 occupancy of dehydrated NaY has been reported to be ca. 3.8 cations per cage from X-ray diffractometry.8 A comparison of this value with the saturation degree of adsorption of the methylamines over NaY (ca. 3 molecules per cage) lead to the conclusion that amine molecules are preferentially adsorbed on site-I1 cations. The saturation degrees of adsorption of the methylamines over KY (ca. 5 molecules per cage), except for Me3N, were greater by ca.1 molecule per cage than the value expected from the site-I1 occupancy of dehydrated KY (3.4-3.8 cations per cage).5* An X-ray-diffraction study5 of dehydrated KY has shown that a small number of ions exist near site 111' that are not in a specific location. These 'unlocated' ions act as selective sites for the adsorption of methylamines, as do the site-I1 cations. If the diversityg? lo of aluminium distribution around site I1 is ignored, the environment of the site-I1 cations may be regarded as independent of position. Therefore the site-I1 cations are expected to act as energetically homogeneous sites for the adsorption of polar molecules. The qst curves observed over NaY and KY support this idea. The specific dependence of qst on the degree of adsorption reveals another interesting point: the amine molecules adsorbed on site-I1 cations, which are separated in a supercage, interact laterally.The coordinates of site-111' cations are uncertain, so we will discuss the configurations and energetics of methylamines adsorbed on site-I1 cations of NaY and KY in the following three sections. CALCULATION OF POTENTIAL ENERGY The potential energy of an amine molecule on site-I1 cations of NaY and KY was calculated on the basis of the electrostatic-interaction model. Ten cubo-octahedra surrounding the chosen site-I1 cation were taken into account as a model of the zeolite according to the results of Kiselev et al."? l2 (see fig. 5). Based on the X-ray-diffraction experiments on dehydrated KY zeolites by Mortier et al.,5 the occupancies of sites I, I' and I1 were assumed to be 0.5'0.5 and 1, respectively. The resulting zeolite model has a unit-cell composition of M+,,&$i1360384, where M is the alkali metal.In the potential-energy calculations only interactions with the oxygen ions and the exchangeable cations were taken into l 1 7 l2 The coordinates of the oxygen ions and the exchangeable cations of NaY and KY were taken from the work of Eulenberger et a1.8 Zero charge was ascribed to the Si and A1 atoms and a charge of + 1 to all the exchangeable cations. The net negative charge was distributed equally among all the oxygen atoms, the numerical value of the charge on oxygen being deter- mined by the electroneutrality of the lattice as a whole. For simplicity of calculations the amine molecules were assumed to be rigid and to have point dipoles at the positions of their nitrogen atoms.The atomic coordinates of the amine molecules were taken from microwave experiments.13--15 The electrostatic interaction energy of a polar molecule with a zeolite lattice is composed of four types of contribution: the field-dipole interaction (Edip), the field-induced dipole interaction (Eind), the dispersion interaction (Edjs) and the repulsion interaction (Erep). Edip and Eind were calculated using the following equations :I6 k where ,u is the dipole moment of the amine molecule, @(r) is the electrostatic potential at a point in the supercage of the zeolite at a distance Y from the centre due to allK. MORISHIGE, S. KITTAKA AND S. IHARA 253 1 the zeolite ions, t is the coordinate parallel to the dipole axis of the adsorbate molecule and a k is the polarizability of the kth atom of the molecule.The electrostatic field -a@(r)/aq in the q direction at the position of the kth atom of the molecule is given by the following lattice surnmation:l1 Here r i j k is the distance between the centre of the kth atom and thejth ion of type i having a charge Qi, qij and q k are the coordinates of thejth ion of type i and the kth atom of the molecule in the q direction, respectively. The Slater-Kirkwood formula I 7 7 l 8 was used for the evaluation of Edis: where ai is the polarizability of an ion of type i, Ni and Nk are the effective numbers of the outer subshell electrons for the ith type ion of the zeolite lattice and the kth atom of the molecule, respectively, and the other symbols have their usual meanings.The repulsion term contributes to the total energy to a great extent; its value is sensitive to the method of calculation as well as to the value of the energy parameter used. At present, however, an established method for this does not exist. Thus the following approach was adopted. Erep was divided into two parts: Ei',,,, the repulsion interaction between the whole of an amine molecule and the site-I1 cation and E$t\, that between each atom of an amine molecule and the oxygen ions of the zeolite. First Ek\ (in kcal mol-l) was calculated according to the method of Gowda and Benson:19 E&L = 1 04 240 exp ( - RN / p ) . ( 5 ) Here R, is the distance between the nitrogen of the amine molecule and the site-I1 cation and p is a range parameter for the repulsion interaction of alkali-metal cations with the whole amine molecule.The value of p can be determined by using the experimental binding energyzo* 21 of alkali-metal cations with amine molecules in the gas phase, in which the assumed configuration brings about an optimization of the cation-dipole interaction. Table 1 summarizes the range parameters selected in this way and the corresponding energy components. The evaluation of EEL was less accurate than that of EyJp and was made using the following equation:24 where A,, and B k , are the repulsion constants between the kth atom of the molecule and the oxygen ions of the zeolite, and r k o j is the distance between the kth atom of the molecule and thejth oxygen ion of the zeolite.As Ne is isoelectronic atom with atomic oxygen, its repulsion constant was used as that for the oxygen ions of the zeolite lattice. The coefficients A,, and B,, were evaluated by the combination rule.24 All the constants required to calculate the adsorption potential energy are listed in tables 2 and 3. ADSORPTION POTENTIAL ENERGY The potential energy was calculated as a function of the rotation angle (0) about the dipole axis and the distance between the nitrogen atom and the site-I1 cation ( R N ) , in such a configuration that the dipole axis of the molecule coincides with the line joining the site-I1 cation and the centre of the supercage (the optimization of the2532 ADSORPTION OF METHYLAMINES ON ZEOLITES Table 1.Range parameters, p, selected on the basis of the experimental binding energies of alkali cation with methylamines and the corresponding energy components NH3 MeNH, Me,NH Me3N . I NH3 &NH, Me,NH Me,N 0.2338 0.2326 0.2325 0.2245 0.2620 0.2612 0.2623 0.2546 2.10 2.07 2.07 1.97 2.43 2.40 2.39 2.29 Na+ 22.73 20.65 16.46 11.40 16.98 15.36 12.34 8.44 K+ 13.65 17.58 21.50 27.36 7.96 10.65 13.69 17.34 3.73 4.50 5.12 6.85 4.62 5.76 7.00 9.13 13.10 14.23 14.17 16.1 1 9.77 10.66 11.50 12.94 27.0 1 28.50b 28.9 1 29.51b 19.78 21.1 1 21.52 2 1.98 a These values coincide with the experimental binding energies2O3 21 corrected by the effect Na+ binding energies wereevaluated by assuming of thezero-point vibration (ca. 2 kcal mol-l). the linear relationship of the log-log plotsz2, 23 between binding energy and cation radius.Table 2. Values of dipole moments, p, and polarizabilities, a, of methylamines [from ref. (25)] molecule p/10-l8 e.s.u. cm3 ~~ NH3 1.45 2.34 MeNH, I .28 4.19 Me,NH 1.02 5.93 Me3N 0.64 7.93 Table 3. Values of the polarizabilities, a, effective number of outer subshell electrons, N , the repulsion constants, A and B, of N, C and H atoms of the methylamines and of the zeolite lattice ions ion or atom a/ cm3 N A/A-1 B/ 1 O4 kcal mo1-I - - Na+ 0.44326 8.818 K+ 1 .37OZ6 1 7.318 Od- 1 .2S2 8.818 -4.7824 28.OZ4 N 1.1 19a 5.518 -2.75328 0. 732928 C 1 .02727 4.718 - 2.753b 0.7329b H 0. 407, 0.918 - 3.Olz4 0.07 1 524 - a This value was estimated on the basis of the additive of molecular polarizability. The repulsion constant for the carbon atom was assumed to be the same as that of the nitrogen atom.K .MORISHIGE, S. KITTAKA AND S. IHARA 2533 Fig. 6. Geometry of the rotation of amine molecules adsorbed on site-I1 cations around the dipole axes: (a) NH,, (b) MeNH,, (c) Me,NH and ( d ) Me3N. Table 4. Interaction energies calculated for the adsorption of methylamines on site-I1 cations of Y zeolites NH3 2.10 MeNH, 2.11 Me,NH 2.21 Me3N 2.09 NH3 2.49 MeNH, 2.45 Me,NH 2.45 Me,N 2.34 15 19.16 0 16.72 0 11.85 0 8.56 60 10.74 0 9.94 0 7.92 0 5.67 NaY 9.04 9.46 10.32 16.02 9.97 22.19 13.51 29.12 KY 2.87 5.63 3.86 8.25 4.84 10.92 6.44 14.50 13.10 1.38 23.19 14.0 11.98 5.47 25.60 15.5 7.76 11.01 25.23 17.6 9.44 12.97 28.79 18.8 7.77 0.02 11.44 11.9 8.80 0.05 13.19 14.5 9.15 0.14 14.39 15.5 10.63 0.16 15.82 14.8 electrostatic-field-dipole interaction23* 29).Here the rotation angle 0 is defined in fig. 6. The resulting minimum energies are compared with the experimental adsorption energies in table 4. Strictly speaking, the calculated potential energy and the experimental energy of adsorption must be corrected for the zero-point energy and the RT term, respectively, for the purposes of comparison. However, they contribute by only a small amount to the total energy. The agreement between -Et and qg is excellent for KY within the limit usually permitted for such a calculation, while for NaY Et is greater by ca. 10 kcal mol-l than qg. The quantum-chemical study of2534 ADSORPTION OF METHYLAMINES ON ZEOLITES Table 5.Interaction energies calculated for the adsorptions of methylamines on the site-I1 cation of NaY on the basis of the revised charge (Na+O.’) of the exchangeable cation ~~ NH3 2.26 0 11.13 3.09 7.33 6.61 0.89 14.05 14.0 MeNH, 2.26 0 9.83 3.67 12.78 6.29 3.46 16.53 15.5 Me,NH 2.35 0 7.07 3.69 18.12 4.25 7.11 17.52 17.6 Me3N 2.23 0 5.09 4.98 23.76 5.06 8.38 20.40 18.8 Fig. 7. Possible configurations of trimethylamine molecules adsorbed on site-I1 cations : (a) NaY and (b) KY. Beran30 revealed that the charge on the exchangeable cations in faujasite zeolites decreases in the order K > Na > Li. The energy parameter concerning the cations will be affected by the decrease in charge. If we ignore this effect, however, we can calculate the potential energy using the decreased charge. When the charge of the Na ion was taken as + 0.7, a good agreement between the calculated and observed values was achieved as is shown in table 5.This effective charge on the Na ion is nearly the same as that estimated by Neddenriep31 from the heats of adsorption for CH,, Xe and Kr. An inspection of tables 1, 4 and 5 discloses several interesting points. First, the adsorption potential energies are compared with the interaction energies of alkali- metal cations with methylamines in gas phase. The electrostatic potential due to the exchangeable cations in the supercages is considerably cancelled out by the negative charge of the zeolite framework. This resulted in a remarkable reduction of the electrostatic interaction energy between the site I1 cation and a polar molecule, especially of the field-induced dipole term, compared with that between an alkali-metal cation and a polar molecule in gas phase.On the contrary, the dispersion energy was increased by the addition of the interaction with the lattice oxygen ions surrounding the site-I1 cation. It is, nevertheless, noteworthy that the content of the interaction energies of a polar molecule with the K,, ion resembles that with K ion in gas phase. Secondly, the potential energies for NaY and KY are compared in order to clarifyK. MORISHIGE, S. KITTAKA AND S. IHARA 2535 01° Fig. 8. Change in potential energy of methylamine molecules adsorbed on site-I1 cations as a function of rotation angle (0) around the dipole axes: (a) NH,, (b) MeNH,, (c) Me,NH and ( d ) Me,N.The full and dashed curves represent the results for NaY and KY, respectively. the reason for the different orders of the basicities. Despite the smaller size of the Na ion, the magnitudes of the electrostatic terms for both zeolites are nearly the same, whereas the dispersion and repulsion interaction with the oxygen ions surrounding the site-I1 cations are larger for Nay. This means that the Na,, ion is screened to a considerable extent by the oxygen ions surrounding it from interacting directly with adsorbed molecules in the supercage. Fig. 7 illustrates possible configurations of trimethylamine molecules adsorbed on site-I1 cations of NaY and KY. The KII ion is held ca. 1 A above the plane of the six-membered ring consisting of the 0, and 0, oxygens while the Na,, ion resides almost in the plane.8 This difference seems to be responsible for the different orders of the basicity of methylamines over NaY and KY.Finally, it is instructive to mention that the contribution due to electrostatic interaction is rather small in the case of the adsorption of a bulky and polar molecule into the supercage of alkali-metal faujasite zeolites. Fig. 8 shows how the potential energy changes when an amine molecule adsorbed on the site-I1 cation is rotated about its dipole axis. On the K , , ion the energy barrier to rotation is extremely low compared with the thermal energy, while the energy barriers for MeNH,, Me,NH and Me,N molecules on the NarI ion are high because of their interaction with the surrounding oxygen ions.These results show that the molecules of NH,, MeNH,, Me,NH and Me,N on both the K,, ion and the NH, molecule on the Na,, ion are able to rotate freely about their dipole axes while the molecules of MeNH,, Me,NH and Me,N on the NaII ion are highly restricted from free rotation.2536 ADSORPTION OF METHYLAMINES ON ZEOLITES I I I I I I ! ! ! ! I 1 I I I I ! ' No I I I I I I I : ! ! I I I Fig. 9. Configurational geometry of four polar molecules adsorbed on site-I1 cations in a supercage. The arrows denote the dipoles of the amine molecules. Table 6. Potential-energy parameters of the amine-amine force field CH3 795034 240034 N 227 1 33 1 23033 H(N) 033 033 ABSORBATE-ADSORBATE INTERACTIONS The dipole of each molecule adsorbed on site-I1 cations coincides with the line joining the site-I1 cation and the centre of the supercage and is in a tetrahedral configuration in a given supercage (see fig.9). In such a configuration dipole-dipole repulsion occurs between two adjacent molecules within the supercage. The adsorbate- adsorbate interaction energy between two adjacent molecules localized on site 11-cations, the configuration of which was determined in the preceding section, was calculated using the Stockmayer potential for the interaction between polar molecules. In this potential the dipole-dipole energy ( E d d ) is given by Edd = - pa p b [ - 2 cos pa cos p b + sin pa sin pb cos (Yb - y,)] (7) r t b where the angles p and y refer to the orientation of the dipole axes with respect to the xyz coordinate frame with the z axis along the line of the centres a and b of the two dipoles.The dispersion and repulsion interactions are represented by the Lennard-Jones potential : 3 3 9 34 E L , = c/t$i - D/& (8) where C and D are the potential-energy parameters, which are listed in table 6. Fig. 10 shows the energy change accompanying the interaction between two molecules in a KY supercage when a molecule is fixed at 8 = 0" about its dipole axis and anotherK. MORISHIGE, S. KITTAKA AND S. IHARA -I 2537 c - \ - L I 1 1 I I 1 I 0 100 200 300 81" Fig. 10. Change in potential energy between two methylamine molecules adsorbed on site-I1 cations in a supercage as a function of the rotation angle (0) of one molecule around its own dipole axis at a fixed angle (0 = 0') of another molecule: (a) NH,, (b) MeNH,, ( c ) Me,NH and ( d ) Me,N.Table 7. Calculated values of the interaction energy between methylamines adsorbed on site-I1 cations of Y-type zeolites molecule Edd/kcal mol-' E,,/kcal mol-' total/kcal mo1-I NH3 MeNH, Me,NH Me,N NH3 MeNH, Me,NH Me3N 0.15 0.12 0.08 0.03 0.49 0.37 0.23 0.08 NaY -0.01 - 0.26 - 0.42 -0.54 -0.10 - 0.60 - 1.20 1.07 KY 0.14 -0.14 -0.34 -0.51 0.39 - 0.23 - 0.97 1.15 molecule is rotated about its own axis. The constancy of the NH,-NH, interaction energies originates from the nature of the simplified model of the point dipole. A strong repulsion observed for MeNH, and Me,N molecules occurs when the CH, groups are brought so close together that their electron clouds begin to overlap. On the other hand, the NaII ions are so far apart in a supercage, as compared to the K,, ions, that there is no repulsion between the CH, groups of two adjacent molecules adsorbed on site-I1 cations of Nay.Table 7 summarizes the minimum energies of the absorbate- 83 FAR 12538 ADSORPTION OF METHYLAMINES ON ZEOLITES absorbate interaction obtained when two adjacent molecules adsorbed on site-I1 cations were independently rotated about their own dipole axes. The interaction between two NH, molecules is repulsive in total because of its relatively large dipole-dipole repulsion. Replacement of a hydrogen atom by a methyl group decreases the dipole-dipole repulsion interaction and increases the dispersion inter- action. Consequently, the interaction between two adjacent molecules becomes more attractive with the substitution of an Me group, except for Me,N on KY.In the case of Me,N on KY there is a repulsion between two adjacent molecules, because this molecule is bulky (owing to its three Me substituents), although there is no steric hindrance between the two adjacent Me,N molecules adsorbed on the site-I1 cations of Nay. This difference is ascribable to the longer distances between the two adjacent site-I1 cations in the supercage of NaY as compared with KY. The observed dependences of qst on the degree of adsorption in fig. 2-4 agree well with these results, supporting the assumed geometries of methylamines adsorbed on the site-I1 cations of NaY and KY zeolites. CONCLUSIONS On dehydrated NaX, site-111’ cations act as the most selective and energetically heterogeneous sites for the adsorption of methylamines, while over dehydrated NaY and KY the methylamines are preferentially adsorbed on site-I1 cations. The isosteric heats of adsorption of methylamines over NaY and KY were successfully analysed by detailed comparisons with the calculated energies of interaction between a molecule and the zeolite lattice, as well as those of the lateral interaction between adsorbed molecules.The main conclusions to be drawn from these comparisons are as follows. (1) Methylamines are selectively adsorbed on the site-I1 cations of NaY and KY in such a configuration that the dipole axis of the molecule coincides with the line connecting the site-I1 cation and the centre of the supercage, causing the optimization of the electrostatic-fielddipole interaction.(2) The electrostatic potential due to site-I1 cations is largely cancelled out by the negative charge of the zeolite framework, resulting in a remarkably small electrostatic interaction energy for a polar molecule with a site-I1 cation as compared with an alkali-metal cation in the gas phase. (3) In the case of the adsorption of a bulky polar molecule on site-I1 cations of alkali-metal faujasite zeolites, the contribution of the non-electrostatic interaction to the total energy is large. (4) The different orders of the basicities of the methylamines on NaY and KY arise from the different environments of the site-I1 cations: the K,, ion is held ca. 1 A above the plane of the six-membered ring consisting of the 0, and 0, oxygens, while the Na,, ion resides almost in the plane.( 5 ) The dipoles of polar molecules adsorbed on site-I1 cations are in a tetrahedral configuration in a given supercage. This geometry explains the dependence of qst on the degree of adsorption. K. Morishige, S. Kittaka, S. Takao and T. Morimoto, J . Chem. SOC., Faraday Trans. I , 1984,80,993. W. J. Mortier and H. J. Bosmans, J . Phys. Chem., 1971,75, 3327. R. M. Barrer and R. M. Gibbons, Trans. Faraday SOC., 1963, 59, 2569. K. Morishige et al., unpublished results. W. J. Mortier, H. J. Bosmans and J. B. Uytterhoeven, J . Phys. Chem., 1972, 76, 650. J. V. Smith, Ado. Chem. Ser., 1971, 101, 171. G. R. Eulenberger, D. P. Shoemaker and J. G. Keil, J . Phys. Chem., 1967, 71, 1812. E. Dempsey, G. H. Kuhl and D. H. Olson, J .Phys. Chem., 1969, 73, 387. ’ W. M. Butler, C. L. Angell, W. McAllister and W. M. Risen Jr, J. Phys. Chem., 1977, 81, 2061.K . MORISHIGE, S. KITTAKA AND S. IHARA 2539 l o B. Beagley, J. Dwyer, F. R. Fitch, R. Mann and J. Walters, J . Phys. Chem., 1984, 88, 1744. A. G. Bezus, A. V. Kiselev, A. A. Lopatkin and P. Q. Du, J . Chem. Soc., Faraday Trans. 2, 1978,74, 367. l 2 A. V. Kiselev and P. Q: Du, J. Chem. Soc., Faraday Trans. 2, 1981, 77, 1. l 3 W. H. Fink and L. C. Allen, J . Chem. Phys., 1967, 46, 2276. (The eclipsed form of methylamine was l 4 J. E. Wollrab and V. W. Laurie, J . Chem. Phys., 1968, 48, 5058. l 5 J. E. Wollrab and V. W. Laurie, J . Chem. Phys., 1969, 51, 1580. l6 W. A. Steele, The Interaction of Gases with Solid Surfaces (Pergamon, Oxford, 1974), p. 59. assumed in the calculation.) J. C. Slater and J. G. Kirkwood, Phys. Reu., 1931, 37, 682. R. A. Scott and H. A. Sheraga, J . Chem. Phys., 1965, 42, 2209. Is B. T. Gowda and S. W. Benson, J . Chem. Phys., 1983,79, 1235. ‘O W. R. Davidson and P. Kkbarle, J . Am. Chem. SOC., 1976,98,6133. 21 R. L. Woodin and J. L. Beauchamp, J . Am. Chem. Soc., 1978, 100, 501. 2 2 B. T. Gowda and S. W. Benson, J. Phys. Chem., 1982,86, 1544. 23 S. F. Smith, J. Chandrasekhar and W. L. Jorgensen, J . Am. Chem. Soc., 1982, 86, 3308. 24 I. Eliezer and P. Krindel, J . Chem. Phys., 1972, 57, 1884. 2b R. J. W. Le Fevre and P. Russel, Trans. Faraday Soc., 1947, 43, 374. 2G V. G. Solomonik, J . Struct. Chem., 1978, 19, 860. 27 J. Applequist, J. R. Carl and K. Fung, J . Am. Chem. Soc., 1972, 94, 2952. 2H J. W. Vanderslice, E. A. Mason and E. R. Lippincott, J . Chem. Phys., 1959, 30, 129. 29 J. E. Del Bene, M. J . Frisch, K. Raghavachari, J. A. Pople and P. von R. Schleyer, J. Phys. Chem., 30 S. Beran, J . Phys. Chem. Solids, 1982, 43, 221. 31 R. J. Neddenriep, J . Colloid Interface Sci., 1968, 28, 293. 32 J. 0. Hirschfelder, C . F. Curtiss and R. B. Bird, Molecular Theory of Gases and Liquids (Wiley, New 33 S. Lifson, A. T. Hagler and P. Dauber, J . Am. Chem. Soc., 1979, 101, 51 11. 34 W. L. Jorgensen, J . Am. Chem. SOC., 1981, 103, 335. 1983, 87, 73. York, 1964), p. 848. (PAPER 5/023) 8 3 - 2

ISSN:0300-9599    

DOI:10.1039/F19858102525

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. Soc., Faraday Trans. I , 1985,81, 2541-2550 Microdynamics of Methane, Ethane and Propane in ZSM-5 Type Zeolites BY JURGEN CARO, MARTIN BULOW AND WOLFGANG SCHIRMER Zentralinstitut fur physikalische Chemie der AdW der DDR, Rudower Chaussee 5, 1199 Berlin, German Democratic Republic AND JORG KARGER,* WILFRIED HEINK AND HARRY PFEIFER Sektion Physik der Karl-Marx-Universitat Leipzig, LinnestraDe 5, 70 10 Leipzig, German Democratic Republic AND SERCEJ PETROWITSCH ZDANOV GrebenSEicov Institute of the Academy of Sciences of the U.S.S.R., Nab. Makarova 2, 199164 Leningrad, U.S.S.R. Received 15th January, 1985 The n.m.r. pulsed field-gradient technique has been used to study systematically the intracrystalline self-diffusion of methane, ethane and propane in ZSM-5. In conjunction with the information obtained from nuclear magnetic relaxation studies the elementary steps of diffusion are found to be activated jumps between the channel intersections.Only for sorbate concentrations > 2.5 molecules per intersection is a decrease in the jump lengths observed. The results are compared with alkane self-diffusion measurements in zeolites A and X, as well as with the self-diffusion of water in ZSM-5 and zeolites A and X. Pentasil zeolites exhibit a number of remarkable properties which have made them an interesting topic of both fundamental research and industrial application. The molecular shape-selectivity in catalytic reactions (para-selective interconversion and alkylation of aromatics,l dewaxing of distillates2 and the conversion of methanol into alkenes and hydrocarbons in the petroleum range3) is the most interesting structure- related property of ZSM-5.In the case of the high-silica form, i.e. Silicalite, its hydrophobic and organophilic character makes it a promising substance for selective In such cases the intracrystalline mobility of molecules is the limiting step in the overall In the present paper, we have used the n.m.r. pulsed field-gradient technique to measure directly the intracrystalline self-diffusion coeffi- cients of methane, ethane, propane and water in ZSM-5. Combining these data with the results of nuclear magnetic relaxation studies and with self-diffusion measurements in NaX- and NaCaA-type zeolites a model for the molecular migration of short-chain alkanes in ZSM-5 is proposed. EXPERIMENTAL The self-diffusion coefficients were measured by the n.m.r. pulsed field-gradient technique using a home-built n.m.r.pulse spectrometer at a proton resonance frequency of 60 MHz. Since the true intracrystalline self-diffusion coefficient D can only be determined if the mean molecular displacement ( r 2 ( A ) ) i of the diffusing molecule during the observation time A is smaller than the crystal radius, laboratory-synthesized single crystals of ZSM-5 (with mean diameters up 25412542 MICRODYNAMICS OF ALKANES ON ZEOLITES to 30 pm, cf. fig. 2 later), NaX (19 pm) and NaCaA (ca. 70% Ca) (42 pm) have been used in this investigation. The crystal size and morphology were checked by scanning electron microscopy and the crystallinity by X-ray techniques.As described previously8 lo the self- diffusion coefficient was determined from the Einstein equation D = (r2(A))/6A by monitoring the n.m.r. signal as a function of the width and magnitude of the field gradients. The proton magnetic relaxation measurements were carried out using a home-built pulse spectrometer at a resonance frequency of 90 MHz. The mean experimental error in the self-diffusion coefficients depends strongly on the systems under study. For methane in ZSM-5 it is ca. 50% and for propane in ZSM-5 even 100% (cf. the significant scattering of the experimental values shown later in fig. 3). The error in the relaxation-time measurements is < ca. 25%. RESULTS Fig. I gives a summary of the concentration dependence of the self-diffusion coefficients of alkanes in the zeolites Silicalite, NaCaA and NaX.In fig. 2 the self-diffusion coefficients of methane and of water in ZSM-5 are plotted as a function of the SiO,/Al,O, ratio and compared with the corresponding values in the A- and X-type zeolites. Fig. 3 shows Arrhenius plots of the self-diffusion coefficients of methane, ethane and propane for different loadings of the Silicalite crystals. In order to guarantee a complete saturation of the intracrystalline pore system, for each system studied a sample with a loading exceeding the saturation capacity was prepared. In these cases the self-diffusion coefficient was determined only from that part of the signal which corresponds to molecules within the crystals [cf. ref. (1 l)]. In addition to the previously determined valuesll of the intracrystalline self-diffusion coefficients of methane and ethane in NaX, fig.4 shows the corresponding data for propane. Results of the relaxation measurements are shown in fig. 5-7 and the temperatures of the c minima determined from these plots are collected in table 1. The transverse relaxation time, T2, has been found to decrease monotonically with decreasing temperature. The relatively large value of ca. 15 for the q/ T2 ratio at the minimum in can be taken as a hint [cf. ref. (9), p. 3151 that T, should be controlled by a small number of molecules strongly adsorbed at structural defects. Hence T, is not characteristic of the mobility of the larger number of adsorbed molecules and will not be discussed in this paper.DISCUSSION The ZSM-5 structure consists of straight channels in the (010) direction and sinusoidal channels parallel to (100); diffusion in the (001) direction proceeds by exchange between these two channels.169 l7 Therefore, in contrast to the cubic A and X zeolites, intracrystalline diffusion in the orthorhombic ZSM-5 is determined by a diffusion tensor rather than by a single diffusion coefficient.l89 l9 Owing to the random distribution of crystal orientations in a powder sample the n.m.r. pulsed field-gradient technique provides an average value of the diffusion coefficients in the directions of the principal axes. In agreement with a preliminary study18 the values of the intracrystalline self-diffusion coefficients of the short-chain alkanes in ZSM-5 are between those in zeolites NaCaA and NaX (cf.fig. I). This result can be understood by comparing the mean diameters of the intracrystalline windows, which are ca. 0.45, 0.55 and 0.75 nm for 5A-, ZSM-5- and X-type zeolites, respectively. The mobility of water in ZSM-5 considerably exceeds that in zeolites NaA and NaX (cf. fig. 2), whichL 2 4 2 2 1 4 2 1 o-c 4 2 ? to-" s 4 E 2 1 o - ~ 4 2 1 o-'O 4 2 lo-" 4 2 lCt2 4 J. CARO et al. 1 2 3 4 5 6 7 8 9 10 I I I I I I I I 1 ~ 1 2 3 - 4 ~ 6 7 0 9 1 0 I I I I I I I 1 2 3 4 5 6 7 I I I I I I I 1 2 3 4 5 6 7 2543 number of molecules per 24 (Si + Al) atoms Fig. 1. Dependence of the intracrystalline self-diffusion coefficients of (a) methane, (b) ethane and ( c ) propane at 27 "C on sorbate concentration expressed by the number of molecules per 24(Si+Al) atoms (representing one large cavity in the case of X and A type zeolites or one channel intersection in the case of Silicalite): 0, zeolite X: NallAlllSi1304a; A, Silicalite: Si2404a; 0, zeolite A : Na4Ca4Al12Si1204,.The values for methane-NaX and propane-NaCaA were obtained by extrapolation of the data in ref. (1 1) and (12), respectively, to 27 "C.2544 MICRODYNAMICS OF ALKANES ON ZEOLITES I0:i 4 ICH4- NaX 0 0 A 2 - lo-*. 4 , - IH20-NaX 1 1 I 1 10 102 lo3 SiO,/AI,O, Fig. 2. Self-diffusion coefficients of (a) methane and (b) water in ZSM-5 at 23 "C for a loading of ca. 3 molecules per channel intersection as a function of the SiO,/Al,O, ratio. The following zeolites have been used: SiO,/Al,O, symbol type provided by ratio size/pm3 ref.a 0 + 0 v 0 v A 0 Silicalite Zdanov, Leningrad > 103 Na,HZSM-5 Rees, London 40 Na,HZSM-5 MostowiEz, Warszawa 185 80 42 HZSM-5 Riekert, Karlsruhe 50 HZSM-5 H oEevar, Ljubljana 540 216 127 72 3 0 ~ 2 5 x 1 5 - 3 5 ~ 1 5 x 1 0 - 5 103 13 1 6 ~ 1 2 x 8 - 5 103 - 5 103 6,14 5 103 15 5 103 - 5 103 - 5 103 - is correlated with the 'hydrophobic' nature of ZSM-5. This finding is in accordance with the results published in ref. (1 8) and (20). A change in the SiO,/Al,O, ratio, and hence in the cation content does not lead to a perceptible shift in the self-diffusion coefficient of methane. This result is in agreement with previous results for the self-diffusion coefficients of n-alkanes in zeolite NaX, where the cation-adsorbate interaction was found to be of no influence on the diffusion behaviour." On the contrary, for water in ZSM-5 a slight increase in the self-diffusion coefficient with increasing SiO,/Al,O, ratio appears to be indicated, which might be a consequence of the water-cation interaction.With increasing alkane concentration both in ZSM-5 and zeolite NaX a monotonous decrease in the self-diffusion coefficients is observed (cf. fig. 1). By combining those results with nuclear magnetic relaxation measurements, as in the work of Karger er a1.,l1 the reduced translational mobility of alkanes in NaX with increasing sorbate concentration could be directly attributed to the diminished molecular free volume.J. CARO et al. 2545 temperature/OC 100 0 -50 -1 00 4 - 2- lo4 - 0 0 0 ooo 0 O 0 0 0 0 0 ( b ) 0 0 0 0 10- 4 8 8 8 0 O O 0 6 4l 2 O - O OoO? I 1 1 I I 3 4 5 6 1 0 3 KIT Fig.3. Intracrystalline self-diffusion coefficients of (a) methane, (b) ethane and (c) propane in Silicalite for 0.5 (a), 1 (A), 1.5 (A), 2 (O), 2.5 (m), 3 (0) and 4 (0) molecules per channel intersection. If the loading exceeded the saturation capacity (0) the contribution of the molecules outside the crystals to the signal was separated. By a modification of the free-volume theory2’ the molecular mean jump length was estimated to be a2 (12) = y vf2/3 exp (-F) du, Vf v* - ( 3 2 ’ 3 r - - ( 5 P*) 3 ’ Of where 0, and v* denote, respectively, the mean free volume and the minimum free2546 10-l0 L 4 - 2 5 2 - lo-"- - MICRODYNAMICS OF ALKANES ON ZEOLITES - tempera ture/OC 100 0 -50 -100 -125 I 1 I 1 I I a 4 - 2 - 2 lo2- R 4 - O o A 0 0 0 0 0 ; 0 0 0 O "0 0 0 0 00 V V V 2 3 4 5 6 7 8 1 0 3 ~ / ~ Fig.4. Intracrystalline self-diffusion coefficients of propane in zeolite NaX for 1.5 (A), 2 (n), 4 (O), 5.5 (0) and 6.5 (V) molecules per large cavity. temperature/OC 75 0 -50 -100 -1 50 I l l I I A A 0 " . 0, A 1 0' 27 2 3 4 5 6 7 8 9 103 KIT Fig. 5. Longitudinal relaxation times ( T I ) of propane in NaX for 1.5 (A), 2 (a), 4 (O), 5.5 (0) and 6.5 (V) molecules per large cavity.J. CARO et al. temperature/'C 75 0 -50 -100 -1 50 loZ- 1 I I I I la) 0 0 2547 2 3 4 5 6 7 8 9 Fig. 6. Longitudinal relaxation times (c) of (a) methane, (b) ethane and (c) propane in Silicalite for 1 (A), 2 (0) and 3 (0) molecules per channel intersection. 103 K I T temperature/ "C 75 0 -50 -100 -1 50 2 3 4 5 6 7 8 9 103 K I T Fig.7. Longitudinal relaxation times (7'J of propane in Silicalite for 1 (A), 1.5 (A), 2 (o), 2.5 (m), 3 (0) and 3.3 (0) molecules per channel intersection. volume still allowing a diffusion step. y is a factor correcting the overlap of the molecular free volumes and T(m, X) = tm-l exp (- t ) dt s: denotes the incomplete gamma function.22 Eqn (1) has been used to estimate the mean jump lengths of the propane molecules (table 1) using the values v* = vmolecule = 0.092 nm3 and y = 0.5 utilized by the authors of ref. (1 1) and with the molecular free volume directly determined from the cavity volume V and from the number of molecules per cavity, N . (2) uf = V/N-Umolecule2548 MICRODYNAMICS OF ALKANES ON ZEOLITES Table 1.Estimated values for the minimum mean jump lengths in Silicalite and NaX-type zeolite at the temperature of the TI minimum (the values in parentheses were obtained by extrapolation of diffusivities to the temperature of the minimum) ~~ concentration temperature [molecules per of the TI Dexptl U 2 ) L p t , Q2)!heor 24(Si+Al) atoms] minimum/K /lo-" m2 s-la /lO-l nmb /lop1 nmc 2 3 1 .o 1.5 2.0 2.5 3.0 3.3 1.5 2.0 4.0 5.5 6.5 130 150 150 175 210 270 315 360 135 135 135 135 135 ethane-Silicalite (13) (10) (13) (12) propane-Silicalite 13 14 6 5 5.8 4.0 2.5 0.6 0.1 propane-NaX 11.7 - 10.3 - 11.7 - 11.3 - 11.7 - 12.2 - - 8 .O 7.3 - 7.8 7.3 6.5 5.5 5.1 3.1 2.5 2.4 1 .o 1.5 a At the temperature of the T, minimum. According to eqn (3). According to eqn (1).The molecular mean jump length {12)$,pt, can be directly determined using the (3) relation D = { l 2 ) / 6 z between the self-diffusion coefficient D and the mean residence time z between two succeeding jumps if the latter quantity is known. On the other hand, the correlation time z, of the longitudinal proton magnetic relaxation can be approximated [cf. ref. (8) and (9)] by the reciprocal value of the proton magnetic resonance frequency (z, z a;&) at the temperature of the minimum and is hence a known quantity. Depending on the dominating interaction there are two cases: (i) z, = z in the case of the magnetic proton interaction with paramagnetic impurities fixed in the zeolite lattice and modulated through the translational motion of the molecules; (ii) 7, < z in the case of the magnetic proton-proton interaction with protons of the same molecule modulated through molecular reorientation together with or without a simultaneous motion.Therefore, by eqn (3) from the self-diffusion and relaxation data a lower limit of the mean jump lengths can be found. Table 1 shows that the mean jump lengths determined experimentally on the basis of eqn (3) are in good agreement with the values of the propane jump lengths in zeolite NaX derived from eqn ( i ) according to the modified free-volume theory. The concentration dependence of the alkane self-diffusion coefficients in ZSM-5 is similar to that in NaX-type zeolites, but the relaxation-time measurements indicate a decisive difference in their transport characteristics. For NaX the temperature of the & minimum (and associated with it the value of the correlation time) remains unchanged with increasing sorbate concentration (cf.fig. 5). However, in ZSM-5 theJ. CARO et al. 2549 minimum is drastically shifted to higher temperatures (cf. fig. 6 and 7). It is remarkable that for small and medium pore filling factors the mean jump lengths determined on the basis of eqn (3) remain constant being of the order of the distances between adjacent channel intersections (ca. 1 nm for a straight channel and ca. 1.2 nm for a sinusoidal channel). In zeolite NaX the reduction of the molecular translational mobility with increasing sorbate concentration is due to the reduced jump length (with a mean lifetime practically unaffected by concentration), but in ZSM-5 a significant decrease in the jump rate is observed.The concentration dependence of the self-diffusion coefficients may be interpreted in terms of the model of successive pore-filling as proposed by Beyer and According to this model the first two alkane molecules of short chain length are assumed to be located within the intersections of the two types of channel. A further increase in sorbate pressure would fill the whole pore system, with the alkane molecules in an ‘end-to-end ’ configuration. The observed jump lengths determined experimentally by eqn (3) are in the order of the intersection distances up to a loading of ca. 2.5 molecules per intersection. The jump lengths decrease with further increasing sorbate concentration, which is in accordance with this model.A determination of the activation energies of self-diffusion in ZSM-5 is complicated by the significant scattering in the experimental data. Just as with zeolite NaX [cf. ref. (1 1) and fig. 41 the activationenergies of self-diffusion seem to be independent of sorbate concentration. For methane diffusion in Silicalite the activation energies of the self-diffusion coefficient D (4.3 1 .O kJ mol-l) and proton magnetic correlation time z, (3.6k0.5 kJ mol-l) are found to be in satisfactory agreement. Hence, via eqn (3) the molecular jump length is found to be independent of the temperature, as it should be on the basis of the model discussed. In the present investigation the following features have been shown to be in accordance with the random-walk model of Theodorou and Wei24 for describing molecular transport in a ZSM-5 grid: (i) the decrease in the intracrystalline mobility with increasing sorbate concentration due to mutual molecular interferences and (ii) the constant jump lengths being ca.I nm. In contrast to the jump mechanism for short-chain n-alkanes, Weisz et aI.l9 have proposed a ‘ supermobile’ surface-diffusion model for the large intracrystalline mobility of longer n-alkanes. For such molecules the sorption sizes should not be separated from each other by distances greater than the length of the molecule, leading to averaged van der Waals forces and to a low temperature dependence of the diffusivity. Further experimental work will decide whether it is in any case necessary to assume the Weisz model for describing the diffusion behaviour of longer n-alkanes in ZSM-5.We thank Prof. Riekert (Karlsruhe) and Drs Rees (London), Mostowitz (Warszawa) and Hotevar (Ljubljana) for providing us with large-diameter zeolite crystals as well as for stimulating discussions. Mr Richter-Mendau is thanked for the scanning electron micrographs. L. B. Young, S. A. Butter and W. W. Kaeding, J . Catal., 1982, 76, 418. N. Y. Chen, R. L. Gorring, H. R. Ireland and T. R. Stein, Oil Gas J., 1977, 75, 165. S. L. Meisel, J. P. McCullough, C. H. Lechthaler and P. B. Weisz, Chem. Technol., 1976, 6, 86. E. M. Flanigen, J. M. Benett, R. W. Grose, J. P. Cohen, R. L. Patton, R. M. Kirchner and J. V. Smith, Nature (London), 1978, 271, 512. R. M. Dessau, ACSSymp. Ser., 1980, 135, 123.H-J. Doelle, J. Heering, L. Riekert and L. Marosi, J . Catal., 1981, 71, 27. ’ J. Wei, J . Catal., 1982, 76, 433.2550 MICRODYNAMICS OF ALKANES ON ZEOLITES a H. Pfeifer, in NMR-Basic Principles and Progress, ed. P. Diehl, E. Fluck and R. Kosfeld (Springer-Verlag, Berlin, 1972), vol. 7, p. 53. H. Pfeifer, Phys. Rep. C, 1976, 26, 293. lo J. Karger and J. Caro, J . Chem. SOC., Faraday Trans. I , 1977, 73, 1363. l1 J. Karger, H. Pfeifer, M. Rauscher and A. Walter, J . Chem. SOC., Faraday Trans. 1, 1980, 76, 717. l 2 J. Karger and D. M. Ruthven, J. Chem. SOC., Faraday Trans. I , 1980, 77, 1485. l 3 J. M. Berak, B. Kanik, J. Mejsner and B. Kontnik-Matecka, React. Kinet. Catal. Lett., 1982,20,431. l4 J. Heering, Thesis (Technische Hochschule Karlsruhe, 1982). l5 V. N. Romannikov, V. M. Mastikhin, S. HoEevar and B. Driaj, Zeolites, 1983, 3, 31 I. l6 D. H. Olson, G. T. Kokotailo and S. L. Lawton, J. Phys. Chem., 1981,85, 2238. l7 G. T. Kokotailo, S . L. Lawton, D. H. Olson and W. M. Meier, Nature (London), 1978, 272, 437. l8 J. Karger, W. Krause and H. Pfeifer, 2. Phys. Chem. (Leipzig), 1982, 264, 838. l9 W. 0. Haag, R. M. Lago and P. B. Weisz, Faraday Discuss. Chem. SOC., 1981, 72, 317. 2o W. Maiwald, W. D. Basler and H. T. Lechert, in Proc. 5th Int. Conf. Zeolites, ed. L. V. C. Rees 21 M. H. Cohen and D. Turnbull, J . Chem. Phys., 1959, 31, 1164. 22 W. I. Pagurowa, Tables of the Incomplete Gamma Function (Computer Centre of the Academy of 23 J. Valyon, J. Mihalyfi, H. K. Beyer and P. A. Jacobs, in Proc. Workshop: Adsorption of Hydrocarbons 24 D. Theodorou and J. Wei, J . Catal., 1983, 83, 205. (Heyden, London, 1980), p. 562. Sciences of the USSR, Moscow, 1963). in Zeolites (Academy of Sciences of the G.D.R., Berlin, 1979), p. 134. (PAPER 5/083)

ISSN:0300-9599    

DOI:10.1039/F19858102541

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1985, 81, 2551-2559 Ultrasonic Attenuation in Aqueous Solutions of a-, p- and y-Cyclodextrins BY SABINE RAUH Max-Planck-Institut fur biophysikalische Chemie, D-3400 Gottingen, Federal Republic of Germany AND WILHELM KNOCHE* Fakultat fur Chemie, Universitat Bielefeld, Postfach 8640, D-4800 Bielefeld 1 , Federal Republic of Germany Received 10th December, 1984 Relaxation effects have been observed in aqueous solutions of a-, p- and y-cyclodextrin and are attributed to a change in the solvation of the cyclodextrin cavity. For y-cyclodextrin, the effects can be described using a narrow distribution of relaxation times. a-Cyclodextrin also shows another faster relaxation due to the rotation of a glucose structure. a-, p- and y-cyclodextrins are cyclic oligosaccharides built up from 6 , 7 and 8 glucopyranose units, respectively. Over the last twenty years the physicochemical properties of these compounds have been studied extensively because inclusion complexes without covalent binding are formed between cyclodextrins and many ligands.9 This paper reports a study of the kinetic behaviour of aqueous solutions of a-, b- and y-cyclodextrins and presents information about conformational changes and the solvation of their cavities in water. Ultrasonic relaxation techniques are the most effective methods for this study and ultrasonic absorption data are reported in the frequency range from 0.3 to 105 MHz. To the authors' knowledge, only one investigation about ultrasonic absorption in aqueous a- and p-cyclodextrin solutions has been reported so far;3 this investigation covered, however, only a restricted frequency range (1 5-145 MHz).The kinetics of aqueous y-cyclodextrin solutions have not been described before. EXPERIMENTAL The ultrasonic absorption was measured between 0.3 and 105 MHz and from 5 to 50 "C using the methods developed by Eggers and F u n ~ k . ~ The concentration ranges were restricted by the solubilities of the cyclodextrins (CD) in aqueous solution: 145 g dm-3 for a-CD, 18.5 g dm-3 for p-CD and 232 g dm-3 for y-CD at 25 O C 5 A pulse method operating with a fixed transducer distance (ca. 10 mm.) was used in the range from 15 to 105 MHz. Using this techniqu:, the attenuation of an ultrasound pulse was measured both in the solution studied and in a reference solution: both solutions had the same physical properties but the reference solution showed no chemical relaxation effect.The difference between the absorptions of the two solutions was defined as the 'excess absorption'. In this study, water was used as the reference solution and the sound absorption coefficient is given by asolvent = B,P, with B, = (44.1 & 0.3), (29.6 _+ 0.3), (21.3 f 0.1) and (14.6 k 0.1) x The range from 0.3 to 25 MHz was covered using the resonator technique with 4 and 5 MHz s2 m-' at 5, 15, 25 and 40 "C, respectively.*, 255 12552 ULTRASONIC STUDY OF CYCLODEXTRINS IN WATER quartz crystals (20 mm diameter) between 1.5 and 25 MHz and 2 MHz quartz crystals (60 mm diameter) between 0.3 and 1.0 MHz. Below 1.0 MHz a hydrostatic pressure of ca. lo5 Pa was applied to the resonator cell in order to increase its mechanical quality.At these frequencies the half-power bandwidth Af was measured by means of the decay time rd of the resonator (2mdAf = l).' An on-line data handling systema was used to evaluate zd. The resonator cell was calibrated with a reference solution and again excess values of the sound adsorption were obtained. For the study of the chemical relaxation of a solution, both techniques had to be employed. In order to check their correct overlap, we measured the weak relaxation effect of a 0.01 mol dm-3 aqueous solution of MnSO,. The frequency dependence of the adsorption coefficient was described by a single relaxation effectg and the deviations were statistical and < 8%.The cyclodextrins were obtained from EGA and purified according to ref. (5). a-CD was recrystallized once from 60% aqueous propanol, twice from water and dried for 12 h in uucuo at 80 "C. p-CD was recrystallized from water and dried for 12 h in uucuo at 110 "C. y-CD was used without further purification, but dried in U ~ C U O at 70 "C for 12 h. The CD solutions were freshly prepared before use. Methyl Orange was obtained from Merck and used without further purification. RESULTS Excess sound absorption is observed in aqueous solutions of a-CD. An example is shown in fig. 1 together with results from the 1iteratu1-e.~ The data cannot be fitted to a single relaxation effect and therefore a superposition of two effects is assumed: where a is the sound absorption coefficient defined as the sum of a,,,,,, and %olvent, f is the frequency, o is the angular frequency, zi are relaxation times, Ai are relaxation amplitudes and B is a constant.All parameters can be evaluated, since the relaxation times differ by more than a factor of 58 and since a/f2 attains both the low-frequency limit ( A , + A 2 + B) and the high-frequency limit (B) in the frequency range of the equipment employed. A fit of the data to a distribution of relaxation times leads to systematic deviations between calculated and experimentally obtained values for the absorption coefficient. For 0.1 and 0.15 mol dm-3 solutions, the relaxation parameters are summarized in table 1. An excess absorption is also observed in 0.05 mol dm-3 solution at 25 "C and in 0.1 mol dm-3 solution at higher temperatures (up to 50 "C).However, in these experiments the effects are very weak and no reliable values of the parameters can be obtained. A 0.016 mol dm-3 solution ofp-CD (nearly saturated) also shows a weak excess absorption, which fits to eqn (1) with A = 0.25 x lod6 sm-l, z = 70 x loA9 s and B = 24 x s2 m-l. However, the experimental error is so great that these values cannot be relied upon. Ultrasound is absorbed strongly in solutions of y-CD, as is demonstrated in fig. 2(a). The best fit of the data to a single relaxation effect yields the deviation plot in fig. 2(6). The deviations are systematic and extend beyond experimental error. Therefore, a fit to two relaxations [eqn (I)] is carried out and statistically distributed deviations are obtained [fig.2(c)]. However, since the relaxation times z, and 7, do not differ greatly, the absorption may also be described using a distribution of relaxation effects. Within experimental error, the data can be fitted to a symmetric Cole-Cole distribution of relaxation times :loS. RAUH AND W. KNOCHE 90 f 2553 \ -- 8ol--------- 70 =. I \ \ '\ + + + ++ YI + - + + + + 0 .5 -" -8 Fig. 1. Ultrasonic attenuation of a 0.1 mol dm-3 solution of a-CD. The curve is calculated using eqn (1) with the parameters given in table 1. The circles and the dotted line are taken from ref. (3). Table 1. Relaxation parameters for aqueous a-CD from eqn (1) ~~~~~~ ~ standard c t B/10-15 A,/lO-s T , / ~ O - ~ A2/10-'j T , / ~ O - ~ deviation /mol dm-3 /"C s2 m-' s m-l S s m-l S (% ) 0.10 25 26.8+0.5 1.6f0.2 6 f 1 0.28f0.04 48k6 2.5 0.15 25 31.1k0.4 2.6k0.2 7.5+1 0.50+0.07 46f5 2.0 0.10 5 55.3k0.6 2.7k0.2 11k1 0.81+0.08 64k5 2.0 0.10 15 36.7k0.6 2.lk0.2 8 f l 0.43f0.07 51+6 2.5 with 0 < b < 1 [see fig.2(d)] and in addition to an asymmetric Cole-Davidson distribution :11 (3) - a - 4 c o s 4Y sin (4B) + B f"- w with 0 < B < 1 and tan 4 = o z R [see fig. 2(e)]. zR is a characteristic relaxation time and b and /3 describe the width of the distribution (with b = 1 and B = 1, eqn (2) and (3) correspond to a single relaxation effect]. A Cole-Davidson distribution is chosen to summarize the measurements in table 2. In order to attribute the observed excess absorption to chemical reactions, we added Methyl Orange to solutions of y-cyclodextrin, since y-CD forms a strong 1 : 2 inclusion2554 ULTRASONIC STUDY OF CYCLODEXTRINS IN WATER 1000 4 7 E 300- N m t + J -&I + Fig.2. (a) Ultrasonic absorption of a 0.015 mol dmP3 solution of y-CD at 5 "C. The curve is calculated using eqn (1) with z1 = 260 x loP9 s, A, = 2.4 x 10+ s m-l, t, = 97 x lop9 s, A , = 1.9 x s m-l and B = 47 x s2 m-l. (6) Deviation plot if the data are fitted to a single relaxation effect (z = 170 x lop9 s, A = 4.1 x s2 m-l). The curve corresponds to the theoretical deviation if a function according to eqn (1) is fitted to a single relaxation effect. (c) Deviation plot assuming two relaxation effects; parameters as in (a). ( d ) Deviation plot assuming a Cole-Cole distribution [eqn (2)] with zR = 170 x s, b = 0.9, A = 4.5 x s2 m-l.(e) Deviation plot assuming a Cole-Davidson distribution [eqn (3)] with zR = 233 x s, p = 0.71, A = 4.63 x s m-l and s m-l and B = 47 x s m-l and B = 46 x B = 45.3 x s2 mP1.S . RAUH AND W. KNOCHE 2555 Table 2. Relaxation parameters for aqueous y-CD from eqn (3) standard c / t l ~/10-15 A / I O - ~ z,/10-9 deviation mol dmP3 "C s2 m-l s m-l S P (% 1 0.0 15 0.01 5 0.015 0.0 15 0.005 0.0 1 0.02 0.03 0.04 5 15 25 40 25 25 25 25 25 45.3 k0.3 30.5 & 0.2 21.9 f0.3 14.9 f 0.3 22.0 f 0.2 21.7f0.3 21.4 k0.4 22.9 & 0.4 23.7 10.5 4.63 f 0.05 4.13 f 0.06 3.7 f 0.1 3.2 f 0.2 1.21 k0.07 2.5 f 0.2 5.5 f 0.2 7.2f0.1 9.9 f 0.2 233 k 20 107+ 10 51f5 20f2 59f6 60f6 57f6 42f4 44f4 0.7 1 & 0.02 0.71 f 0.02 0.72 f 0.05 0.78 f 0.07 0.69 f 0.08 0.64 f 0.07 0.6 & 0.04 0.90 & 0.04 0.86 f 0.04 2.0 2.0 3.0 2.0 2.5 2.5 3.5 3 .O 4.0 f 1 10 100 f/MHz Fig.3. Ultrasonic attenuation at 25 "C: +, 0.01 5 mol dm-3 y-CD; 0,O.OlS mol dm-3 y-CD and 0.01 5 mol dm-3 Methyl Orange; A, 0.01 5 mol dm-3 y-CD and 0.03 mol dm-3 Methyl Orange. The curves are calculated with the constants given in tables 2 and 3. complex with Methyl Orange ( K = 6.2 x mo12 dm+J12 This caused a reduction of the relaxation amplitude and generated a second relaxation effect at higher frequencies. Some measurements are shown in fig. 3 and table 3 comprises the absorption parameters obtained by a fit of the measurements to eqn (1). DISCUSSION Saenger and c o ~ o r k e r s ~ ~ ~ l4 have investigated the structure of crystalline a-CD hydrate complexes, They described three different a-CD hydrate modifications, two hexahydrates (forms I and 11) and a '2.57' water complex (form 111).In form I, the a-CD cavity encloses two water molecules. The a-CD annulus is2556 ULTRASONIC STUDY OF CYCLODEXTRINS IN WATER Table 3. Relaxation parameters for aqueous y-CD-Methyl Orange at 25 "C from eqn (1)' standard CMethyl Orange B/10-15 A1/10-6 ~ ~ / 1 0 - ~ A2/10-6 Z , / ~ O - ~ deviation /mol dm-3 s2 m-l s m-l S s m-l S ( X ) 0.015 25.2k0.3 0.66k0.03 84+4 2.5k0.1 9.6+1 2.5 0.030 24.9f0.8 0.21+0.03 95+14 1.9k0.3 6.3+ 1 4.5 c ; , . ~ ~ ) = 0.015 mol dm-3 and t = 25 "C. distorted and one glucose structure is rotated out of alignment with the other units. The a-CD molecule in form I1 has an almost identical conformation, but only one water molecule and a primary hydroxyl group of an adjacent a-CD are in the cavity.Form 111 is an almost symmetrical macrocyclic conformation. On average, 2.57 water molecules are enclosed in the cavity. The water molecules are disordered and statistically distributed over four positions. The structures attributed to the crystalline a-CD molecule are probably preserved in aqueous solution. Therefore, in solution the a-CD molecule exists in 'tense' (forms I and 11) and 'relaxed' (form 111) states. The fast relaxation process observed in solutions of a-CD is attributed to the transition between these states, as already proposed by Rohrbach et al.3 The relaxation time of the slower process is similar to that of y-CD, and so the slower relaxation may be caused by the reorientation of water molecules included in the cavity, possibly accompanied by the addition of a water molecule, as discussed later for y-CD.The relaxation processes can be observed in a very restricted concentration range only (see table 1). Thus it is not possible to evaluate thermodynamic or kinetic parameters for the reactions involved. The room-temperature crystal structure of y-CD hydrate has been recently reported upon by Harata,15 whereas in an earlier publication Maclennan and StezowskilG described the structure at 120 K. Some differences between the results of these investigations should be mentioned: at 120 K, a y-CD- 17H,O complex has been found. 12 water molecules are characterized as being in the torus and one glucose residue is statistically disordered. Harata, however, has proposed a y-CD 1 3.3H20 complex at ca.22 "C. On average, the y-CD torus includes 5.3 water molecules, which are statistically disordered and may occupy 13 different positions. According to Harata, all glucose residues are well ordered. It is assumed that the structure of 7-CD in aqueous solution resembles the structure in the solid state. In aqueous solutions of y-CD a large ultrasonic absorption effect is observed. The results given in table 2 show that at a constant temperature the relaxation amplitude A is proportional to the concentration of y-CD and the relaxation time zR is independent of the concentration, within experimental error. Both facts indicate that the reaction observed is monomolecular or pseudomonomolecular, as for a reaction k k' A e B with equilibrium constant The relaxation time is given by T--' = k + k' = k'( 1 + K )S .RAUH AND W. KNOCHE 2557 and the relaxation amplitude by 2n2up(A V)2Y R T A = (7) where u is the sound velocity, p is the density and A V is the reaction volume (the contribution of the reaction enthalpy AH may be neglected in aqueous solutions) is the transfer function and the concentration c is given as c = [A] + [B]. The temperature dependence of the relaxation time is given by i3 lnz-l Ei KAH ~ = --- tu-1 R ( l + K ) R (9) where Ei is the activation energy. By differentiating eqn (7) we obtain from the temperature dependence of the amplitude AT ' In(=) (K-1)AH assuming that A V is independent of the temperature. From the experiment we may obtain the four measurable quantities given by eqn (6), (7), (10) and (1 l), but the equations contain five unknown parameters k', K, A V, AH and Ei.Therefore, for a monomolecular reaction it is impossible to obtain either thermodynamic constants (K, AH and A V ) or kinetic constants (k' and Ei) by evaluating chemical relaxation measurements only. However, a lower limit of A V can be calculated from the relaxation amplitude [eqn (7)] since the transfer function has a maximum value r = 0.25~ for K = 1. Eqn (6) and (7) describe the relaxation behaviour of a discrete reaction, but they are also approximately valid for a narrow distribution of relaxation times, as observed for aqueous solutions of y-CD. Therefore, the relaxation amplitude yields for the reaction volume the value AV 3 9 cm3 mol-l. According to the structure of the 7-CD molecule, the relaxation effect may be explained either by the rotation of a glucose residue or by a change of solvation. Obviously, the reaction volume is too large for the rotation of a glucose structure and it can be concluded that the ultrasonic absorption is due to a change of solvation of the cyclodextrin cavity.Not more than 12 water molecules may be in the torus, according to the crystal structure of y-CD. The van der Waals radius of a water molecule is 0.19 nm. Using this value, a simple calculation shows that 12 water molecules occupy ca. 90% of the cavity. As previously mentioned, in the crystal structure at room temperature Harata has found 5.3 water molecules in the cavity, which occupy only 40% of the available space. Thus several y-CD molecules of varying solvation will exist in solution with different amounts and orientations of water molecules.This agrees with the fact that a distribution of relaxation times is observed. The water molecules are disordered and not fully hydrogen-bonded, although they are apparently in an activated state.17 Since the solvating water molecules cause the relaxation effect, it should be reduced when the cavity is occupied by molecules other than water. This was shown by adding different amounts of Methyl Orange to a 0.01 5 mol dm-3 solution of y-CD. The results2558 ULTRASONIC STUDY OF CYCLODEXTRINS IN WATER are shown in fig. 3 and table 3. Methyl Orange forms strong inclusion complexes with y-CD and, as expected, the amplitude of the relaxation effect is reduced by the addition of Methyl Orange.Simultaneously, a second fast relaxation effect is generated, which is therefore due to the existence of the y-CD-Methyl Orange complex. This reaction has not been studied further. The relaxation time in solutions of y-CD agrees with that in solutions of p-CD and with the long relaxation time in solutions of a-CD. Therefore, it is assumed that these relaxation effects are also due to a change in the solvation of the cyclodextrin cavity. A distribution of relaxation times should also be observed for p-CD. However, the effect is so small that it is not possible to differentiate between a single relaxation and a distribution of relaxation effects. The results may be summarized as follows.In aqueous solutions of a-, p- and y-cyclodextrins a relaxation effect is observed with a relaxation time of ca. 5 x lo-* s at 25 “C. The maximum effect is found to be in y-CD solutions, where it can be described using a narrow distribution of relaxation times and the effect is attributed to a change in the solvation of the cyclodextrin cavity. This explanation is supported by the fact that the relaxation amplitude is reduced when the water is partially expelled from the cavity by adding a strongly binding ligand. In solutions of a-CD a second faster relaxation is observed, which is attributed to the rotation of a glucose structure. Note added in proof: As one of the referees has pointed out, similar measurements have been presented by Nomura et a1.18 In solutions of a-CD, they observed two relaxation effects with relaxation frequencies of ca.1 and 20 MHz. The higher frequency agrees well with our value of 23 MHz (corresponding to z = 7 x s), whereas we obtain 3.4 MHz as the lower relaxation frequency (7 = 47 x s). The value of 1 MHz obtained by Nomura et al. is very close to the low-frequency limit of their equipment (0.7 MHz), and therefore an unambiguous evaluation of the ultrasonic data is not possible. We think that our value can be relied upon, since we extended the measurements by more than a factor of two down to 0.3 MHz. Nomura et al. attribute the relaxation effects to a ‘local motion of the ring’ and to the ‘amount of inclused water’,I8 which do not contradict our explanation. Nomura et af.describe the data obtained for solutions of y-CD by two relaxation effects, ‘where the relaxation frequency of the lower-frequency side is shifted to the higher-frequency side in comparison with the results for a-CD, while the opposite behaviour is seen for the relaxation frequency of y-CD of the higher frequency side’.l* That means that the relaxation frequencies approach each other and the data can be fitted as well by a narrow distribution of relaxation frequencies, as preferred by us. The relaxation in solutions of y-CD is explained by the same two effects as for a-CD. We believe, however, that a local motion of the ring, e.g. rotation of a glucose structure, cannot be the origin of the relaxation since (i) all glucose residues of y-CD are well ~rdered,’~ (ii) the volume of reaction is much larger than expected for the rotation of a residue and (iii) the relaxation amplitude is reduced when the water in the cavity of y-CD is partially replaced by Methyl Orange.Therefore, the relaxation effects have to be caused by the reorientation of the water molecules in the cavity. A narrow distribution of relaxation frequencies is expected for this process, since these water molecules are not well ordered. We thank Dr F. Eggers for valuable discussions and his assistance with improvements to the equipment. W. Saenger, Angew. Chem., 1980,92, 343. Cycfodextrins, ed. J. Szejtli (Akademi Kiado, Budapest, 1982). R. P. Rohrbach, L. J. Rodriguez and E. M. Eyring, J . Phys. Chem., 1977, 81, 944. F. Eggers and Th. Funck, Rev. Sci. Instrum., 1973, 44, 969. D. French, M. L. Levine, J. H. Pasur and E. Norberg, J . Am. Chem. Soc., 1949, 71, 353. J. M. M. Pinkerton, Nature (London), 1947, 160, 128. ’ F. Eggers and K. H. Richmann, Rev. Sci. Instrum., 1976, 47, 378. A H. Strehlow, Adv. Chem. Relax. Proc., 1978, 12, 29.S . RAUH AND W. KNOCHE 2559 G. Jackopin and E. Yeager, J . Phys. Chem., 1970, 74, 3766. lo K. S. Cole and R. H. Cole, J. Chem. Phys., 1941, 9, 341. l 1 D. W. Davidson and R. H. Cole, J . Chem. Phys., 1951, 19, 1484. H. Hirai, N. Toshima and S. Uenoyama, Polym. J . , 1981, 13, 607. l 3 K. Lindner and W. Saenger, Acta Crystallogr., Sect. B, 1982, 38, 203. l4 K. K. Chako and W. Saenger, J. Am. Chem. Soc., 1981, 103, 1708. l5 K. Harata, Chem. Lett., 1984, 641. l6 J. Maclennan and J. Stezowski, Biochem. Biophys. Res. Commun., 1980, 92, 926. D. W. Griffiths and M. L. Bender, Adv. Catal., 1975, 23, 209. H. Nomura, S. Kato and Y . Miyahara, 6th Int. Symp. on Solute-Solute-Solvent Interactions, Minoo, Osaka, 1982, abstract 1P-09. (PAPER 4/2088)

ISSN:0300-9599    

DOI:10.1039/F19858102551

出版商:RSC    

年代:1985

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1985, 2561-2568 Reviews of Books Aggregation Processes in Solution. Ed. by E. WYN-JONES and J. GORMALLY. (Elsevier, Amsterdam, 1983). Pp. x + 632. Price Dfl 325. The field of aggregation processes in solution is broad and has stimulated many researches during this century. Such processes are relevant in many areas, including pure and applied chemistry, pharmacy and biology, and it is perhaps not surprising that the various strands of research have proceeded independently. For example, the study of dye aggregation and of surfactant aggregation have largely gone their own ways, and the large body of literature on the association of hydrogen-bonding solutes in non-polar solvents contains scant reference to work on micelle formation in surfactant solutions.Obviously, however, the various fields have both experimental and theoretical aspects in common, and the editors of this book feel that cross-fertilisation of ideas between areas should be encouraged; this is one of the aims of the book. The editors give a brief introductory survey pointing to some of the early seminal researches on soap solutions, dye aggregation and hydrogen bonding in solution. Then follow 19 chapters varying in length between CQ. 10 and 60 pages. There are 6 chapters concerned in whole or in part with micellar systems and a chapter each on lyotropic liquid crystals and lipid bilayers. Two chapters are devoted to drugs, one of these on colloidal properties and the other on drug-protein binding. Some 4 chapters are concerned mainly with hydrogen bonding, dyes are the subject of a further 2 chapters and there are 3 chapters on polymer systems.Ferrofluids and base stacking are the subjects of the other 2 chapters. Some contributions are (mainly) theoretical, the longest being that on the thermodynamics of micellar solutions by D. G. Hall. Kinetic theory of micelle formation is covered in a chapter by E. A. G. Aniansson. The theory of aggregation on linear polymers is the subject of the contribution by G. Schwarz, and S. Scheiner writes on the molecular orbital treatment of hydrogen-bonded systems. These chapters are really for the specialist in the given field and not suitable for browsing. Four further chapters are concerned mainly with the information that can be obtained by use of a particular experimental technique.P. Hemmes writes on ultrasonic absorption and velocity studies (of base stacking) and the scope of ultrasonic relaxation spectrometry in the kinetics of hydrogen-bonded systems is treated by J. Rassing. There is a contribution from A. Lassman and N. Rietbrock on the use of stopped-flow measurements for investigation of drug-protein binding and H-H. Limbach considers the application of n.m.r. spectroscopy to hydrogen bonding in solution. The remaining chapters are more general review articles, some covering rather limited areas (e.g. partial molal volumes and compressibilities of micellar aggregates by E. Vikingstad) and others (the majority) with broader fields such as lyotropic liquid crystals (G. J. T. Tiddy and M. Walsh), colloidal properties of drugs (D.Attwood), aggregation of dyes (B. C . Burdett) and polymer-polymer interactions (R. Pethrick). Other authors include D. G. Oakenfull (hydrogen bonding in the formation of bile acid micelles), Z . A. Schelly (equilibrium and dynamic studies of reversed micelles), R. Bisby (lipid bilayer membranes), V. Vitigliano (aggregation of dyes on polyelectrolytes), A. Martinet (ferrofluids) and E. Morris and I. T. Norton (polysaccharide aggregation in solutions and gels). The book is a collection of, in the main, well written individual chapters by established authors. However, there is no cross-reference between chapters (even within a given area such as micelle formation) and no attempt is made to present an integrated view of aggregation processes.This, of course, would have been a tall order, and it may be that having a selection of contributions on the various areas, collected between the same covers, will promote some cross-fertilisation of ideas. All the chapters are well-referenced, there being some 1700 references in all. There is a subject index. 256 12562 REVIEWS OF BOOKS This is the sort of book that is extremely useful to have in the library and it can be strongly recommended for this purpose, but it is too broad in its coverage to be an attractive buy for individual researchers. R. AVEYARD Received 19th September, 1984 International Tables for Crystallography, Volume A. Space-group Symmetry. Ed. by T. HAHN (D. Reidal, Dordrecht, 1983). Pp. xv + 854. Price Individuals $90, Dfl 21 5 , Institutions $165, Dfl 385.This is the first volume in the third series of the International Tables, and is much more than a revision of the second series. Volume A concerns itself with space-group symmetry up to three dimensions, although not including partially periodic groups or antisymmetry groups. It therefore supersedes volume 1 of the previous series as the volume which will be in the most demand by practising crystallographers. Undeniably this work will be a success as no serious crystallographic group will be properly equipped without it. Nevertheless, the old volume 1 will for some time still be the cherished symbol of the established laboratory, with X-ray film developer enriching the monoclinic pages and with covers worn at the corners through countless journeys to students’ lodgings.The new volume is rather too cumbersome for these journeys and may be thought to be too expensive to be left at the ready outside the darkroom. The size is a severe drawback to the new volume, twice the bulk of the old, and may result in the younger generation not getting the ready access to it that we old folk have enjoyed with its predecessor. The few journeys that my copy has suffered have already put a more noticeable strain on the less robust binding than shown by any of our old well used volumes; this is a sure disadvantage for a book which has been made to be used. Why does the new volume need to be so large one may ask. It has been in the planning for over ten years, during which time a pilot issue has been widely distributed throughout the crystallographic community, seeking suggestions.Perhaps too many of the suggestions have led to additional material and insufficient compromises have been made to keep the volume within bounds. One third of the material of the old volume 1, the structure factor tables, has been completely removed. This decision is justifiable by the progress made in computers over the last three decades, and anyone with the specialised need for these tables will remember where they can still be found, It is impossible to catalogue the other deletions as most of the remaining material has been reformulated in some fashion. The new volume has been boosted by the introduction of new diagrams for the triclinic, monoclinic, orthorhombic and cubic groups. The old straightforward diagrams for the low-symmetry groups have been replaced by expertly produced diagrams which are, however, more demanding of the reader.This uses up much more space ( P 2 , / c takes five times the space required in the old volume) and relieves the crystallographer of the task of redrawing his space group in the orientation that he wishes - a task that could well have been argued to be beneficial. Fascinating diagrams for the cubic groups have been added, these must have been great fun for their creators and should keep a keen crystallographer happy for many hours, but only time will tell whether these diagrams deserve the extra space. The diagrams for the cubic groups, symmetry elements are remarkably clear despite their complexity, but I regret that the stereographic projection has been rejected in favour of an ad-hoc picture for the portrayal of inclined symmetry elements.The general positions are represented by stereoscopic diagrams cleverly constructed so that two different angles of view can be seen simultaneously. Much of the production of this volume has been aided by the computer, as explained in one of the prefaces. This must have led to a very high level of reliability throughout, a reliability which is vital for this work. On occasions the tables produced differ from those of the old volume 1, but all the discrepancies I noted were due to different but equally valid arbitrary choices. Each space group is assigned a set of generators (tabulated) from which the general and special positions are calculated, and they are further used in a compilation of the symmetryREVIEWS OF BOOKS 2563 of special projections and the tabulation of the various subgroups and supergroups.This tabulation will interest all those concerned with phase transitions, though it is incomplete for these purposes. The descriptive sections are more extensive than their counterpart in the previous volume 1, but suffer from being short chapters by different authors. Thus some unnecessary repetitions occur, and there is quite a lot of material that one would expect to find in a text book. This is the dilemma of any editor of a work like this: should the work be simply definitive or should it attempt to be explanatory as well? The decision has clearly been to move more towards the explanatory, though covering the same topics as in the old volume 1.Badly needed in that volume was an index, and as this is now rectified the new volume A stands as a complete reference for the field it sets out to cover. In a nutshell it is competent, reliable, fascinating but overweight. G. S. PAWLEY Received 6th August, 1984 ACS Symp. Ser. no. 237. Chemical and Catalytic Reactor Modelling. Ed. by M. P. DUDUKOVIC and P. L. MILLS. (American Chemical Society, Washington D.C., 1984). Pp. ix+426. Price $71.95. This volume is based upon papers presented at sessions organised by the Division of Industrial and Engineering Chemistry at the Annual Meeting of the American Chemical Society in Seattle in 1983. The common feature of the 21 papers is that they deal with some aspect of modelling; that is, descriptions of complex phenomena with simplified mathematical representations which, to some extent, replace the detailed mechanistic processes by ‘ lumped ’ approximations.Therein lies both the strength and weakness of the modelling procedure : properly done, insight can be gained about the overall behaviour of systems without the unnecessary complexity which often accompanies detailed descriptions. However, one can often be misled; when crucial details are omitted from the models, they can yield deceptive results which would never be observed in reality. In the case of reactor modelling, the two ‘details’ which are most often simplified (because they are the most complex and difficult to describe adequately) are the kinetics and the fluid mechanics of the processes in question.The authors of this collection of papers not only deal with a variety of systems to be modelled, but also have different approaches to the treatment of such detail. The volume is divided into five chapters but, in fact, the subjects covered are more diverse than that classification would indicate. Seven of the 21 papers deal with modelling of three-phase reaction systems: gas-liquid reactions in contact with a solid catalyst, in either a trickle-bed reactor or a bubble/slurry reactor (that is, gas-phase or liquid-phase continuous). In four of these papers the emphasis is on finding a suitable way of dealing with the complex fluid mechanics of these reactor configurations : regimes of flow transition, pressure drops, irrigation and wetting in trickle-bed reactors and solid and bubble distributions and contact areas in slurry reactors.This is usually done by correlation of experimental data but in some cases it is combined with a theoretical treatment of the simplified fluid mechanics. The other papers in this set have chosen simple flow representation (plug flow or perfect mixing) and, using available correlations of fluid mechanical properties, have studied aspects of reactor behaviour under these simplified conditions : conversion against space-time or velocity, and stability limits for model reaction systems. Seven other papers nominally deal with models of packed-bed reactors (although one is really restricted to empty tubular reactors, another is more concerned with overall plant economics and a third treats a moving-bed rather than a fixed-bed of catalyst).Two of the seven examine and compare models of varying complexity to predict the performance of non-adiabatic fixed-bed reactors with and without catalyst deactivation. An objective is to ascertain when the simpler models are adequate, although it is not at all clear how valid the results are since comparable values of parameters may not have been used in the simplified models. There are also two papers dealing with design conditions for stability and parametric sensitivity in fixed-bed reactors; one2564 REVIEWS OF BOOKS is concerned with the case of multiple exothermic reactions and the other with adiabatic reactors which can ‘ take-off’ if the conversion exceeds a critical level.This type of reactor often has a large diameter (to minimize pressure drop) and it appears that the effect of non-ideal flow patterns can be characterised simply in terms of the dispersion of the residence-time distribution and independent of the mixing model used to generate that distribution. The remaining papers in the volume are mostly unrelated: three of these have little to do with reactor modelling but are concerned with kinetics of gas-liquid reactions, diffusion with reaction andcatalyst deactivation, respectively. Two others deal with novel reactor configurations which are analogous but not identical to fixed-bed reactors, one does a simple analysis of a stirred-tank reactor in which one of the reactants is a volatile liquid and one paper reports experimental measurements of catalyst deactivation in a plug-flow reactor operating iso- thermally at stepped temperatures.The papers contained in the volume are, for the most part, interesting, informative and representative of current research. Despite the title, the book is not a detailed treatment of a specialist subject, but rather a collection covering a wide variety of related topics. Although it is desirable to get recent research papers into print quickly and efficiently, one must nevertheless question the wisdom of publishing the volume in a hard-backed edition and selling it at what seems a very high price. Individuals could hardly justify its purchase and even libraries would think carefully before buying a volume of miscellaneous research papers, however interesting.L. QRSHENBAUM Received 20th June, 1984 Diffusion in Liquids. A Theoretical and Experimental Study. By H. J. V. TYRELL and K.R. HARRIS. (Butterworths, Sevenoaks, 1984). Pp. xvi + 448. Price E44. This book is a revision and expansion of the diffusion sections of Tyrell’s book Difusion and Heat Flow in Liquids, which was published in 1961. It deals with diffusion and rotational diffusion in liquids in four main sections of theory, experimental methods, methods of interpreting results and a review of currently available data. It is billed in the preface as ‘an account of the theory and practice sufficiently complete to give a non-specialist an adequate base for the understanding of the original literature, and. . . a critical review of the current state of knowledge in the field’.These aims are certainly fulfilled and are perhaps too modest. The book will, in fact, provide research workers in the field and related fields with a detailed account of much of the material that they will need in their studies. The first four chapters deal with the theory of diffusion in liquids. Within this section, chapter 2 develops the principles of non-equilibrium thermodynamics. My first feeling was that this material was out of place here and should be obtained from another source. However, the subject is well explained in a way relevant to the later material and is worth the 30 pages devoted to it. The long chapter 5 (1 50 pages) is an excellent review of experimental techniques with the theory of each method.Twenty-four pages of this are devoted to rotational diffusion measurements. Chapter 6 describes the treatment of experimental results by various methods, some based on the theory of the early chapters, some more empirical. The remaining two chapters are a review and analysis of the experimental results presently available and contain a mass of information which is easy to access. The book is readable, accurate and provides a comprehensive and authoritative account of the subject. Its production is an important academic achievement and it will be a major source book in the chemical physics of fluids for many years to come. A. A. CLIFFORD Received 23rd July, 1984REVIEWS OF BOOKS 2565 Topics in Current Chemistry, Volume 118. Oscillations in Chemical Reactions.By D. GUREL and The investigation of oscillations in chemically reacting systems has not shown the physico- chemical enterprise at its most imaginative. Even now model systems are being explored with virtuoso mathematics and/or computer simulation, and a cavalier disregard for whether or not the theory fits the facts. After decades of dismissing oscillating reactions as an anomaly, fashion has swung into a search for the chimera of the purely chemical oscillator. On the credit side, there has been virtually no mention of catastrophe theory, despite the verification of multiple steady states in several systems operated continuously. To understand the scope for an advanced text or research monograph at this time, in terms of both the state of the art and the currently outstanding problems, a little history is essential.Although Morgan had, in 1916, observed a periodic fluctuation in the evolution of carbon monoxide from formic acid decomposed by concentrated sulphuric acid and Lotka had schematised oscillatory systems (1 9 10 and 1920), Bray’s (1 92 1) description of the behaviour of the system hydrogen peroxide +potassium iodate + sulphuric acid is generally accepted as the first account of an oscillating reaction, because each surge of oxygen evolution coincides with a temporary fading of the iodine colouration (the blue complex with starch does not form). Ten years later Liebhafsky formulated reaction schemes involving oxidation-reduction loops, including Bray’s originally postulated pair (a) peroxide + iodine -+ iodate and (b) peroxide + iodate -+ iodine + oxygen, but failed to kindle wider interest.Various authors refer to attempts to discount Bray’s results, eventually replicated by Degn in 1967. Others saw this as one more phenomenon special to the reactions of hydrogen peroxide. Whatever the reason, the Bray-Liebhafsky system disappeared from the Journals after 1933, without gaining a mention in undergraduate physical chemistry textbooks. At the same time, biologists had become aware of the mathematical analogy between periodicities in animal or insect populations and the oscillatory behaviour of various electronic circuits, the study of which was already (in 1930) crystallising into modern control theory. It is, therefore, not surprising that the next mention of oscillating reactions (1952) occurs in Philos.Trans. R. Soc. London, Ser. B, nor that the author is A. M. Turing, this being one of his last publications. The contemporary era began, without a doubt, with the description of the system potassium bromate + malonic acid + ceric sulphate + sulphuric acid by Belousov in 1959 and the subsequent explorations by Zhabotinsky. The Belousov-Zhabotinsky reaction is similar to the Bray-Liebhafsky reaction in that a redox couple (CelI1/CeIv and iodine/iodate, respectively) catalyses a reaction yielding a gaseous product (carbon dioxide and oxygen, respectively). Unfortunately the similarity was for long obscured by confusion about the differing roles of the iodate and the bromate in the two cases, and it was the dramatic nature of the colour change (with ferroin indicator) and the occurrence of spatial periodicity that established the Belousov-Zhabotinsky system as the archetypical chemical oscillator. Within 15 years it had begun to appear in textbooks of general physical chemistry.The earliest attempt to model oscillatory behaviour (in glycolytic systems) in terms of feedback mechanisms (forward activation etc.) was published in 1967, and the following year saw the first appearance of the ‘ Brusselator ’ reaction scheme, giving spatio-temporal oscillations and limit-cycle behaviour (unlike the Lotka model) with only two reaction intermediates. Nicolis in 197 1 showed how the Brusselator generates a second-order, non-linear differential equation of the Lienard type, d2X/dt2 +.fix, dX/dt) (dX/dt) + g ( X ) = 0 which, incidentally, includes the van der Pol equation for the diode oscillator. Then Field, Koros and Noyes produced the ‘ Oregonator’ (three reaction intermediates) to illustrate the essentials of their analysis of the Belousov-Zhabotinsky reaction.In December 1974 the Faraday Division held a Symposium entitled The Physical Chemistry of Oscillatory Phenomena. It could not have been more timely. Indeed, discussion of the papers on the Bray-Liebhafsky and Belousov-Zhabotinsky reactions occupies 40 % of the space devoted to inorganic oscillators. The Symposium had been planned to explore the common ground between chemical and other oscillatory phenomena, but little communion developed 0. GUREL. (Springer-Verlag, Berlin, 1983). Pp. 121.Price DM 62, $24.10.2566 REVIEWS OF BOOKS even with the cool-flame people. U. F. Franck did attempt the wider view, demonstrating the parallel between oscillation and propagation using electronic and other analogues, but the discussion on this section concentrated on a critique of the ‘termolecular’ step of the Brusselator. Only after the section on membrane processes had re-introduced the modes of thought of control theory did the discussion return to Franck’s paper. A lone contribution presented analyses of simple reaction sequences with feedback, as a contribution towards defining necessary conditions for spatial instability. Finally Texter described the use of purely mathematical tools to establish whether or not a postulated reaction network will exhibit limit-cycle behaviour.Despite the spate of activity which followed the Faraday Symposium, the proceedings of the Bunsengesellschaft Discussion of September 1979 seem to follow on without a break. Considered together, the two meetings evidence a growing scepticism about the purely chemical oscillator. Liebhafsky (who had returned to this field of research in 1969 after an interregnum of 36 years, surely some kind of record) and his collaborators demonstrated, in their Faraday paper, that a positive feedback step of the form 2X+Y +A -, 3X+P, rather than merely X+Y +A -+ 2X+P, is essential for sustained oscillation. Identifying X as iodous acid and Y as hypoiodous acid they hypothesised that reaction rate with hydrogen peroxide increases along the series H I 0 < H,I,O, < H,I,O,.The reaction cited above is, of course, neither more nor less than the much disputed ‘ termolecular’ step of the Brusselator. Liebhafsky and coworkers claimed that their scheme accords with all of the experimental facts, but without reference to the surges of oxygen evolution, or to the effects of the large enthalpies of reaction, while Noyes, at the Bunsengesellschaft Discussion, emphasised the significance of gas-phase nucleation phenomena, even for the Belousov-Zhabotinsky reaction. To complete the paradox (and to bring the history full cycle), Zhabotinsky reported that the best phenomenological model requires unrealistic values of at least one rate constant to reproduce observed results. A final twist (to the story up to 1983) is provided by the system arsenite+iodate+chlorite, described by de Kepper and coworkers in 198 1.It yields no gaseous product, but it does appear to involve the catalysis by a redox couple of an oxidative process. De Kepper claims it as the first chemical oscillator not to have been discovered serendipitously. More important is the suggestion that one role of the redox couple is to confer bistability: two distinct steady states separated by an ‘unstable’ region. Either a feedback or a ‘trigger effect’ might then suffice to produce oscillation. The theoretical position when this book was being compiled was thus as follows. If the five known chemical oscillators (I include the glycolysis systems) are actuated exclusively by reaction kinetics, diffusion, nucleation and enthalpic effects being of no more than marginal significance, then the rate law for an intermediate must be of the Lienard type instanced above.Numerical solutions and/or computer simulations of detailed reaction schemes give qualitative account of experimental observations, including steady-state concentrations, although congruence of predicted and observed limit cycles is not exact. Both the Oregonator and the Brusselator seem capable of adaptation to various oscillators, including the glycolytic ones, but both have been strongly criticised. However the ‘ termolecular’ step of the latter begins to look plausible in terms of the known or conjectured chemistry of the halogen oxyacids. That there is as yet no known non-biochemical oscillator not involving iodate or bromate supports this approach ; but the glycolytic oscillators then remain problematic.The contrary case is simply that nucleation and enthalpic effects are not peripheral but central. Although the work of D’Alba and Di Lorenzo, attributing all chemical oscillation to nucleation effects, was yet to appear, any chemist familiar with control theory would have been aware that the behaviour of ‘ on-off’ or ‘flip-flop’ systems can be superficially indistinguishable from that of continuous feedback systems, as I myself noted more than 20 years ago (Philos. Trans. R. Soc. London, Ser. A , 1961). But surely, if nucleation alone were the key, would there not be many more chemical oscillators? So much for the context, now for the book. At 12 1 pages, is it to be a brief history, an analysis of what the known chemical oscillators have verifiably in common, a layman’s guide to the proceedings of the Faraday and Bunsengesellschaft meetings, or simply a very short Specialist Periodical Report? In fact it comprises two articles by the same two authors.The first is entitled ‘Types of Oscillations in Chemical Reactions’ (pp. 3-73) and is described (in the second) as an introduction to the field. It is divided into eight sections, of which Section VI promises typologies of oscillating reactions, of their elements and of mathematical solutions ; disap-REVIEWS OF BOOKS 2567 pointingly it is 2.5 pages of simple enumeration. Section VII asserts badly that the main achievements to date are, and that the future developments will be, primarily mathematical.My own view is that the major achievements have been physico-chemical (monitoring rapid changes in electrode potential and single-species concentration) and that significant future developments will focus on microscopic explanations of ‘chemical feedback ’. The article properly begins with Section I11 : Reactions and Models Exhibiting Oscillations (pp. 5-55). Two pages on each of the Bray-Liebhafsky and Briggs-Rauscher reactions are followed by an opaque summary of the Field-Koros-Noyes mechanism for the Belousov-Zhabotinsky reaction and an account of the Oregonator incomprehensible to anyone who does not know it already. Oscillations in glycolysis is discussed in terms of sequences with direct positive feedback and one or more loops but without any comment on bistability or details as to how the feedback is generated.Over 14 pages the authors appear more interested in the mathematical results than in the biochemistry, although without enlightening the reader as to how the differential equations are arrived at, or how they are solved to yield the limit cycles shown. At the end the reader will still await an account of how bistability becomes (with changing parameter values) oscillation, of how a steady state (singular point) explodes into a limit cycle and of how one might recognise the preconditions for either phenomenon. Then the Brusselator appears in the guise of the earliest paper rather than the clearest, followed by 14 pages of Abstract (sic) Reaction Systems, mainly from the work of Rossler, of almost no interest to the non-mathematician.Section IV: Characteristics of Oscillating Systems, introduces the ‘phase space’ representation of limit cycles (dimensionality = number of reaction intermediates) which has been used without explanation thus far. The classification of non-linearity is unclear and there is no explanation of how the type of non-linearity determines the nature of the solution (steady state, limit cycle, or multiple). Section VIII : References, lists ca. 130 : a carefully selected list of 40 would be more appropriate to an introduction. The second article, entitled ‘ Recent Developments in Chemical Oscillations’ (pp. 77-1 18), has the same format as the first part of the first one, but focuses on the literature for 1981 and 1982. A major feature is a careful account of the extensive work on the Belousov-Zhabotinsky reaction during this period (pp.81-91), but again one has to go to the literature cited to be able to draw even tentative conclusions: a critical winnowing would have been of real value here. Many of the sub-sections are a few lines only and were not worth inclusion. The discussion of de Kepper’s chlorite oscillators is again too brief to be informative, but useful for the references. I learnt little from this book, though I am sure that the authors have much to teach me. Their failure stems, as so often, from not defining either objectives or readership at the outset; and the blame for that rests as much on the editorial board as on the authors. Springer’s usual standard of production is marred by all too frequent misprints, mistakes and infelicities of English.One’s money would, even now, be better spent on the proceedings of the Faraday and Bunsengesellschaft meetings. A. J. B. CRUICKSHANK Received 1 1 th September, 1984 Specialist Periodical Reports on Electrochemistry, Volume 9. Senior Reporter D. PLETCHER. (Royal Society of Chemistry, London, 1984). Pp. xi + 290. Price L63 (RSC members f39). The current editorial policy concerning the Specialist Periodical Reports on Electrochemistry is to include in each volume reviews of two types: regular reports and occasional timely surveys. In recent years intense research activity in the electrochemistry of transition-metal complexes and in the electrosynthesis of organic substances justifies the treatments of those areas by means of (approximately) annual reports.The authors of these two chapters are active contributors to the respective fields and are effective reviewers. J. Grimshaw’s account of organic electro- chemistry is more selective than the survey of transition-metal complexes by C. J. Pickett, but in each case the fact that these instalments are regular features provides the reader of the series with a comprehensive picture of significant developments in the subject. The four remaining chapters of this volume are devoted to topics which, although requiring2568 REVIEWS OF BOOKS only occasional review, are nonetheless each of importance in modern electrochemistry. The enthusiasm of J. Robinson is evident from his refreshing and readable presentation of spectroelectrochemistry, a survey of the in situ non-electrochemical methods increasingly used to study electrochemical reactions.The author and several of his colleagues in Southampton have been responsible for some important recent contributions in this area. Electrochemical systems are, in some respects, less well suited to the application of spectroscopic and related techniques than are surfaces exposed to the gas phase. However, the ability to modulate the applied electrochemical signal permits enhancement of the measured signal and this can compensate for the inherent disadvantages. Spectroelectrochemistry will undoubtedly acquire many new practitioners over the next few years, some attracted by reading this stimulating chapter. Principles, experimental requirements, applicability and examples of application are described for reflectance spectroscopy, e.s.r., ellipsometry and the use of optically transparent electrodes as well as for several more esoteric techniques. Semiconductor electrochemistry is the unavoidable concern of most electrochemists; whilst relatively few choose specifically to investigate the topic, it intrudes into many studies where it cannot be ignored. L. M. Peter provides an authoritative account of recent developments in the theory of illuminated semiconductor/electrolyte interfaces and in the application of semiconductor electrodes, especially in photoelectrochemical devices. N. A. Hampson and A. J. S. McNeil conclude their two-part review of porous electrodes with a chapter devoted to two aspects: porous electrodes under flow conditions and three-phase systems. The latter topic is treated at length, more than half of the article being an account of Soviet work in this area. The final chapter by D. E. Williams and P. McGeehin examines electrochemical gas sensors which use ceramic membranes. Sensors based on potentiometry (concentration cells), on amperometry, on conductance changes and on surface effects are all discussed mainly from a practical viewpoint. Applications in combustion control are described. Each of these occasional reviews provides a good entry point to the literature. Dr Pletcher and his authors have produced a volume which deserves success. 0. R. BROWN Received 17th September, 1984

ISSN:0300-9599    

DOI:10.1039/F19858102561

出版商:RSC    

年代:1985

数据来源: RSC