摘要:

J. Chem. SOC., Faraday Trans. 1 , 1988, 84(7), 2279-2287 Volumetric Properties of Aqueous Solutions of Polyols between 0.5 and 25 "C Silvana Wurzburger,* Roberto Sartorio, Gennaro Guarino and Maria Nisi Department of Chemistry, University of Naples, Via Mezzocannone 4 , 80134 Naples, Italy The volumetric properties of aqueous solutions containing non-electrolytes have been determined by density measurements in the temperature range 0.5-25 "C. Linear polyols CH,OH(CHOH),CH,OH with n = 2-4 (n = 4 has three diasteroisomers), myo-inositol and t-butylurea have been studied. Vz and the coefficients of the virial expansion of the excess property (V,,, V,,, . . .) have been evaluated and interpreted in terms of solute-solute and solute-solvent interactions with the aid of the enthalpy and free-energy data where known.The theory of group contributions has been used to interpret the V,, values of those linear polyols not exhibiting solute-solute interactions at 25°C. The results confirm the relation between the signs of the enthalpic and volumetric contributions also found for other kinds of non-electrolyte. In the last few years many researchers have focussed their attention on the study of the physico-chemical properties of non-electrolyte aqueous solutions essentially for two reasons ; these solutions provide useful information and predictions on the properties of biological systems1 and furthermore they have a theoretical interest because all the possible interactions in solution are almost of the same order of magnitude.'T3 Most studies have been directed towards investigation of the associative solute-solute interactions existing in these Generally these interactions are studied by means of thermodynamic excess functions, defined for a binary system as: (1) where YE(rn,) is the generic excess property, Y(m,) is the thermodynamic property of a solution containing 1 kg of solvent and m, moles of the solute, Y z is the standard property of 1 kg of solvent and Yz is the partial molal quantity of the solute at infinite dilution.The excess property Y "(m,) can also be represented in terms of a virial expansion :'-' (2) Y"(m,) = Y(m,) - Y z - Yz m, Y"(m,) = K,m:+ Kxzm:+ ... where, according to the McMillan-Mayer theory of so1utions,10-12 the coefficients &,, l',,,, etc. represent the contribution to the thermodynamic excess property of pair, triplet and higher-order solute aggregates. There is a considerable amount of literature concerning the determination of g,,, g,,, and h,,, h,,,,'.13* l4 and a general classification of non-electrolytes according to their g,,, h,, and -Ts,, has also been proposed.15 However, for a more general understanding of aqueous non-electrolyte solutions, due consideration of other properties of these solutions is necessary. In previous papers1'. l7 we have already pointed out that the changes of the diffusion coefficients with concentration can also be interpreted on the basis of associative phenomena and thus can be correlated to the enthalpy and entropy results. Volumetric studies in principle could yield information on these phenomena in a simpler and more direct way, but a systematic study of the volumetric properties V:, 2279 75-22280 Volumetric Properties of Aqueous Polyols K,, c,,, etc.) of aqueous non-electrolytes solutions has not been carried out to date. Among the many papers dealing with the determination of V:18-25 only a few evaluate and discuss the K,, K.,, .. . values (for alcohols, ethylene glycol derivatives and amino acids,' 26, 27) which can supply useful information on solute-solute interactions. In this paper we study the volumetric properties of non-electrolyte solutions. We have examined a series of linear polyols, CH,OH(CHOH) .CH,OH, that manifest a peculiar trend in the variability of the energetic data with n, and in particular have values that change signs from positive to negative for diasteroisomers with n = 4.28 The particular linear polyols we have examined are meso-erythritol with n = 2, adonitol with n = 3, and mannitol, sorbitol and dulcitol with n = 4.We have also examined the volumetric properties of myo-inositol, a cyclic polyol (CHOH),, and of t-butylurea in order to have examples representative of the three classes of non-electrolytes described below. The volumetric analysis has been carried out at five temperatures to study the effect of the water structure on solute-solvent and solute-solute interactions. Experimental Materials The products, meso-erythritol (ERI), adonitol (ADO), mannitol (MAN), sorbitol (SORB), dulcitol (DULC), myo-inositol (INO) and t-butylurea (TBU), obtained from different sources (Sigma, BDH, Serva, Fluka) were purified as described elsewhere.28, 29 The solutions were prepared by mass, using a Sartorius digital balance type 1712, with doubly distilled degassed water. A large amount of solution was prepared in each case in order that the maximum relative error in rn, was always <2 x mol kg-l. Density Measurements Densities were measured with a DMA 602 PAAR model ultrasonic densimeter connected to an AB model SI220 System Teknik digital thermometer as a temperature- control device. Doubly distilled degassed water and dry air were used for the calibration of the densimeter. The estimated error in the density values is < 5 x lo-, g ~ m - ~ . The densities were measured in the concentration ranges reported in table 1 at temperatures of 0.50+0.01, 4.000+0.007, 8.000+0.007, 15.000&0.007 and 25.000f0.007 "C.Treatment of the Data From the density measurements of a binary solution containing 1 kg of water and rn, moles of solute it is possible to obtain V(m,) by the relation (3) V(m,) = (1000 + M , rn,)/d(m,) where d(m,) is the density of the solution and M, is the solute molecular weight. From eqn (1) and (2) it follows that Vm,) = V i + Vzm,+ K,m:+ Vxzxm~+ ... where v:: = 1OOO/d, (4) ( 5 ) and d, is the density of pure water. possible to evaluate the constants r:, V,,, . . . according to the equation Therefore, fitting the values of [V(m,)- V;] us. rn, by least-squares analysis, it is (6) V(m,)- V i = I;;:m,+ K,m:+ c.,m:+ .... For all the systems investigated, only two parameters, V z and K,, were sufficient to represent the data adequately.Table 1.Partial molal volumes at infinite dilution, v: ERI ADO MAN SORB DULC IN0 TBU 0 < rn < 0.25 0 < rn < 0.5 0 < rn < 0.6 T/ "C O<m< 1.6 O<m< 1.6 0 < rn < 0.8 O<m< 1.4 0.5 4.0 8.0 15.0 25.0 84.63 f 0.02 (20) 85.00 f 0.02 (29) 85.43 f 0.02 (20) 85.99 f 0.01 (20) 86.80 _+ 0.02 (18) 100.54 t 0.0 1 100.94I_+0.02 101.44f0.03 102.26+0.01 103.1 7 _+ 0.0 1 (21) (1 5) (16) (20) (16) 115.79f0.13 1 16.70 _+ 0.03 1 17.20 f 0.02 1 18.29 f 0.02 I19.44+0.01 (15) (18) (19) (24) (23) 1 15.23 f 0.03 1 15.82 _+ 0.02 1 16.60 f 0.03 117.82f0.02 119.15f0.02 (13) (16) (22) (20) (30) 11 5.81 f 0.07 (13) 1 16.34 f 0.03 (24) 116.98 f 0.04 (24) 1 17.90 f 0.04 (19) 1 19.07 f 0.03 (21) 97.32 f 0.12 (10) 98.01 + 0.03 (11) 98.86 & 0.04 (23) 100.30 & 0.02 (19) 10 1.94 f 0.0 1 (22) 107.09 f 0.03 107.70 f 0.02 108.29 & 0.03 109.48 f 0.03 110.38 f0.02 (16) (18) (12) (17) (16) V:/cm3 mol-l calculated using eqn (6); m = mol kg-'; numbers of experimental data are in parentheses.Table 2. Contribution to the volume made by solute pairing, V,, T/OC ERI ADO MAN SORB DULC IN0 TBU ~ _ ~ _ _ _ _ ~ 0.5 0.28 f 0.02 0.48 & 0.01 0.99 f 0.20 0.96 & 0.03 -0.37 f 0.13 2.00 & 0.06 - 1.21 & 0.05 0.31 fO.01 0.52 f 0.02 0.83 f 0.04 0.94 f 0.02 0.57 f 0.07 2.03 0.06 - 1.28 k 0.04 8.0 0.26 f 0.02 0.45 k 0.03 0.94 f 0.04 0.79 f 0.03 0.64 0.12 2.15 f 0.09 - 1.37 f 0.07 15.0 0.29+0.01 0.39 kO.01 0.70 k 0.04 0.59 f0.02 1.37 f 0.08 1.8 1 0.05 - 1.58 f 0.08 25.0 0.24 & 0.0 1 0.36f0.01 0.68 f 0.02 0.46 & 0.02 1.44 + 0.13 1 S O & 0.04 - 0.82 f 0.05 4.0 VJcm3 kg molP2 calculated using eqn (6).2282 Volumetric Properties of Aqueous Polyols In tables 1 and 2 are reported the vfs.and I/Z, values and their standard deviations for each temperature in the concentration range explored, as well as the number of experimental data points used in the computation. Results and Discussion In order to discuss our volumetric results it is advantageous to report briefly the interpretation adopted for energetic investigations. As mentioned earlier, the parameters g,,, h,, and - Ts,, are used to classify non-electrolytes as follows:15 class I: hydrophobic structure-making solutes; g,, < 0; Ts,, > h,, > 0 class 11: hydrophilic structure-breaking solutes; g,, < 0; h,, < Ts,, < 0 class 111: hydrophilic structure-making solutes; g,, > 0; h,, > Ts,, > 0.For solutes belonging to the first two classes [class I : alcohols, urea derivatives (TBU); class 11: urea, thiourea, biuret, I N 0 and DULC], the negative g,, values indicate that solute-solute interactions are favoured at increasing concentration ; these interactions are supposed to occur by the overlapping of the hydration cospheres and a subsequent release of water molecules from these to the bulk. Class I solutes have positive h,, values, which are interpreted in terms of a loss of hydrogen bonds when water molecules pass from the overlapping hydration cospheres to the bulk. This fact indicates an overall structuring effect of these solutes on the water structure. Class I1 solutes, on the other hand, have negative h,, values which are interpreted in terms of increased hydrogen bonding due to water molecules going from the hydration cospheres to the bulk.This indicates an overall structure-breaking effect of these solutes on the water structure. Class I11 solutes [polyols (ERI, ADO, MAN, SORB) and monosaccharides] have positive g,, values, showing that solute-solute interactions are not favoured at increasing concentration. The h,, and - Ts,, values are interpreted, at increasing concentration, in terms of a decrease in the number of water molecules participating in the hydration cospheres. The signs of h,, and -Ts,, for class III solutes indicate a decrease of hydrogen bonding when water molecules pass from the cosphere to the bulk, thus suggesting an overall structure-making effect for these solutes in water.In evaluating the volumetric data from eqn (6) two quantities are obtained, vz and 5,: which we will discuss separately. V: represents the partial molal volume at infinite dilution where only solute-solvent interactions exist, therefore it could provide evidence on the structure-making or struc- ture-breaking effects of the solutes on water. Pz is the sum of two contribution^:^' the intrinsic volume of the non-hydrated solute molecule and a term which takes into account the volume change undergone by the water molecules in the hydration process. This is caused by the different water structuring of the bulk as compared to the hydration cospheres.The second contribution is positive for those solutes that manifest a structure- making effect for water, whereas it is negative for structure-breaking solutes. Since the volume of a non-hydrated solute molecule is unknown, it is impossible to obtain information on the nature of solute-solvent interactions from a single vz value. The analysis of v: at different temperatures could be helpful on the assumption that the intrinsic volume of the solute and its hydration cospheres are independent of temperature. Thus the observed change of vz us. Tcould be attributed to the change of the water molecules of the bulk going to the hydration cospheres. Even on these assumptions due consideration of the dependence of the equilibrium nH,O ;t (H,O), on the temperature, of thermal expansion of the solvent and of a possible change of water molecules in the hydration cospheres is necessary.S. Wurzburger et al.Fig. 1. . ( a ) -119 - 117 - 115 114 110 - - 107 % 106- E 102 98 84 0 10 20 T/OC Temperature dependence of the partial molal volume at infinite dilution : DULC (O), SORB (0); (b)TBU; (c) ADO; ( d ) INO; (e) ERI. MAN The data reported in table 1 and fig. 1 show that pz increases with temperature for all the solutes examined whatever their interaction with the solvent. This trend is already known for alcohols and glycol derivatives3' and does not allow us to extract comparative information on the solute-solvent interactions. The Cs values are generally thought to be proportional to the volume change of the hydrated molecules with increasing solute l2, 2 6 y 27 This is assumed to be a consequence of the overlap of the hydration cospheres of the solute pair, or to the redistribution of water among the solute molecules when pairing does not occur.Hence thle c. values could yield information on solute-solute as well as solute-solvent interactions. For solutes belonging to the first two classes it is assumed that associative interactions favoured at increasing concentrations (gsz < 0) are due to an association of the type:s V V Two hydrated molecules of volume V overlap their cospheres by an extent equal to V, and release n water molecules which occupied a volume K in the separate cospheres and a different volume aK in the bulk; a # 1 because of the different structuring in the two domains.The total volume change is where a is < 1 if the n water molecules in the bulk occupy a volume smaller than in the cospheres (if the water molecules are more structured in the cospheres than in the bulk). a is > 1 if the n water molecules in the bulk occupy a volume larger than in the2284 Volumetric Properties of Aqueous Polyols cospheres (if the water molecules are less structured in the cospheres than in the bulk) . The 5, values are obviously proportional to V, (8) and they will be positive if a > 1 and negative if a < 1. In the case of class I11 solutes associative phenomena are not favoured (gZz > 0). Therefore, the water redistribution between the hydration cospheres occurs as follows : <, = KV,(a- 1) V When the concentration is raised a hydrated solute molecule of volume V releases n water molecules, which occupied a volume VR in the cosphere and a different volume aVR in the bulk.The total volume change in this process is A V = V- VR+aVR- V = VR(a- 1) (9) and the related &, vz5 = K&(a - 1). (10) K, is positive if a > 1 (structure-breaking effect of the solute on the water structure) and negative if a < 1 (structure-making effect of the solute on the water structure). Eqn (8) and (10) are similar, the only difference being the significance of and VR. Therefore the sign of the cz values reveals the nature of the solute-solvent interaction, i.e. if the solute has a structure-making or a structure-breaking effect on water.The experimental results for K, for all the solutes we have studied are reported in table 2 at five temperatures. We will discuss first the results at 25 "C, because there are energetic data in the literature relating to this temperature. The experimental values reported in table 2 (T = 25 "C) exhibit a negative 7/2, value for TBU, thus indicating that a < 1, therefore a structure-making effect of the solute on the solvent molecules exists. This result is in complete agreement with the interpretation proposed for the TBU energetic data. Furthermore, yZ. values for other substancesL8* 26 belonging to the first class of solutes are also negative, showing that a general trend for hydrophobic structure-making solutes exists. On the other hand, DULC n = 4 and IN0 cyclic (class 11: g,, < 0, h,, < Ts,, < 0) have positive 5, values at 25 "C.This [eqn (8)] infers that a > 1, therefore suggesting a structure-breaking effect of these solutes on the water structure. Again in this case the interpretation of the volumetric results agrees with the energetic data. This behaviour seems to be generally true for class I1 K, values for solutes belonging to the two first classes seem to confirm the existence of a relationship between h,, and K,.' In fact, in a simplified approach it seems reasonable to expect that negative values of c, (decrease of the volume) correspond to positive values of h,, (breaking of hydrogen bonds) and vice versa; this relationship, however, has only a qualitative meaning, in fact any attempt to find a quantitative dependence of K.with respect to h,, fails, not only for solutes of different classes, but also for solutes of the same class but of different nature. Moreover, in the case of linear class I11 polyols (ERI, ADO, MAN and SORB) K, and h,, (at 25 "C) are both positive; thus the relation proposed above does not seem to hold for all solutes. On the other hand, it is reasonable to assume that the enthalpy change in breaking or forming water-water hydrogen bonds in the hydration cospheres and the following volume variation does depend on the nature of the solute. For hydrophobic structure-makingS. Wurzburger et al. 2285 (a) -0-- - --- - - - - 4,. 10 20 T/'C Fig. 2. Temperature dependence of the contribution to the volume made by the solute pairing of SORB (a), MAN (b), ADO (c)and ERI ( d ) .solutes it is well known that water molecules in the hydration cospheres are bound to each other via hydrogen bonds in an ice-like lattice which does not exhibit appreciable interaction with solute molecules. In this case the nature of the solute does not affect the water interactions in the cospheres. For hydrophilic solutes, the direct interaction existing between solute and water molecules does not allow any prediction of the enthalpy and volume changes connected with the formation or breaking of water-water hydrogen bonds in the hydration cospheres. Because both depend on the nature of the different groups constituting the solute molecules a deeper insight into this matter can be obtained by considering the single group contributions to &, and h,,.For the systems in which a concentration increase yields only a water redistribution among the cospheres, the total q, can be expressed as where Y,,-, represents the contribution to Y,, due to the redistribution of water around the single group I. The use of this equation is limited in our case because it does not take into account the interaction between vicinal groups and the stereochemical effects which play an important role in determining the physical-chemical properties of polyols. Nevertheless, it is possible to extract some qualitative information on the contribution of =CHOH and -CH,OH groups to &, and h,,; looking at the K. and h,, values at 25 "C relative to ethylene glycol (EG),31* 33 ERI, ADO, MAN and SORB (table 3) it is evident that the K, and h,, contributions for =CHOH and -CH,OH are opposite in sign.In fact, as it can be seen from table 3, the addition of each =CHOH to ethylene glycol increases K, while decreasing h,,; this indicates, for a =CHOH group, a positive contribution to &, and a negative contribution to h,, (the differences between MAN and SORB originate from the different orientations of the =CHOH groups). These results confirm our previous hypothesis of a different contribution from each functional group to the volumetric and enthalpic properties and support the view that a qualitative relationship between the signs of &, and h,, may be valid only if applied to a single functional group within the solute molecule. Also, at the other temperatures the c, values are positive and they increase from ERI to SORB (fig.2), indicating the overall structure-breaking effect of this group in the temperature range used. It is also2286 Volumetric Properties of Aqueous Polyols Table 3. Contribution to the volume and enthalpy made by solute pairing at 25°C Y,/cm3 kg mo1-2 h,,/J kg rnoP EG - 0.6" 362b ERI 0.24 0.01 358 a 22c ADO 0.36 f 0.01 295 k 5" MAN 0.68 _+ 0.02 66+ 12' SORB 0.46 f 0.02 -11f5' a Ref. (33). Ref. (31). ' Ref. (28). TI" C Fig. 3. Temperature dependence of the contribution to the volume made by the solute pairing of IN0 (a), DULC (b) and TBU (c). interesting to note the dependence of 5. on temperature for the other systems examined (table 2 and fig. 3). For IN0 (class I1 at 25°C) the positive K.values show a trend similar to that of the solutes belonging to class 111 (ERI, ADO, MAN and SORB); however, the dependence of Cz on DULC is totally different even though DULC belongs to the same class as IN0 at 25 "C. It is very interesting to note that the K. value of DULC changes from positive to negative values at low temperatures; because negative values of yZz are obtained for structure-making solutes (TBU, alcohols etc.), this trend seems to indicate that DULC shows an overall structure-making effect at low temperatures and an overall structure- breaking effect at higher temperatures. 5. values for TBU, the only hydrophobic structure-making solute examined, are negative in the whole temperature range (fig. 3), thus showing a structuring effect on the bulk.Even for TBU K. seems to become smaller at increasing temperature. vs. T results, we can say that the increasing temperature yields a decrease in the effect of solutes on the water structure, independent of solute-solvent interactions. Looking at the entire set ofS. Wurzburger et al. 2287 Conclusions Systematic studies of the volumetric properties of the aqueous solutions containing non- electrolytes are, in our opinion, an important and necessary step in the understanding of the thermodynamic properties of these solutions. The results of our studies point out, in fact, that only by combining energetic and volumetric analysis it is possible to obtain an exhaustive knowledge of the properties of these solutions. Even if, at present, it is not possible to extract much information from the results at different temperatures, a collection of these data might be very helpful in the future.We acknowledge the financial support of the Minister0 della Pubblica Istruzione and of the Italian National Research Council (CNR), Rome. References 1 G. Nemethy and H. Scheraga, Q. Rev. Biophys., 1977, 10, 239. 2 J. S. Rowlinson, Liquids and Liquid Mixtures (Butterworths, London, 1st edn, 1959, 2nd edn, 1969). 3 Water: A Comprehensive Treatise, ed. F. Franks (Plenum Press, New York, 1973). 4 A. E. Guggenheim, Trans. Faraduy Soc., 1960, 56, 1159. 5 R. H. Wood, T. H. Lilley and P. T. Thompson, J . Chem., Soc., Faraday Trans. I , 1978, 74, 1301. 6 T. E. Leslie and T. H. Lilley, Biopolymers, 1985, 24, 695. 7 F. Franks, Proc. R .Soc. Lopdon, Ser. B, 1977, 278, 33. 8 H. L. Friedman and C. V. Krishnan, J . Solufion Chem., 1973, 2, 119. 9 T. H. Lilley and R. D. Scott, J . Chem. Soc., Faraday Trans. I , 1976, 72, 184. 10 W. G. McMillan Jr. and J. E. Mayer, J . Chem. Phys., 1945, 13, 276. 11 J. J. Kozak, W. S. Knight and W. Kauzmann, J . Chem. Phys., 1968, 48, 675. 12 F. Franks, M. Pedley and D. S. Reid, J. Chem. SOC., Faraday Trans. 1, 1976, 72, 359. 13 J. H. Desnoyers, J. Perron, L. Avedikian and J. P. Morel, J. Solution Chem., 1976, 5, 631. 14 A. Cesaro, in Thermodynamic Data for Biochemistry and Biotechnology, ed. H-J. Hinz (Springer-Verlag, 15 G. Barone, P. Cacace, G. Castronuovo and V. Elia, J. Chem. Soc., Faraday Trans. I , 1981, 77, 16 R. Sartorio, V. Vitagliano, L. Costantino and 0. Ortona, J . Solution Chem., 1982, 11, 875. 17 R. Sartorio, S. Wurzburger, G. Guarino and G. Borriello, J . Solution Chem., 1986, 15, 1041. 18 F. Franks and H. T. Smith, J. Chem. Eng. Data, 1968, 13, 538. 19 F. Franks, J. R. Ravenhill and D. S. Reid, J. Solution Chem., 1972, 1, 3. 20 S. Cabani, G. Conti and L. Lepori, J . Phys. Chem., 1974, 78, 1030. 21 H. Hoiland and E. Vikingstad, Acta Chem. Scand., Sect. A , 1976, 30, 182. 22 S. Harada, T. Nakajima, T. Komatsu and T. Nakagawa, J . Solution Chem., 1978, 7, 463. 23 H. Hoiland, J. Solution Chem., 1980, 9, 857. 24 R. V. Jasra and J. C. Ahluwalia, J . Solution Chem., 1982, 11, 325. 25 I. R. Tasker, J. J. Spitzer, S. K. Surl and R. H. Wood, J . Chem. Eng. Data, 1983, 28, 266. 26 C. Jolicoeur and G. Lacroix, Can. J. Chem., 1976, 54, 624. 27 G. Perron and J. E. Desnoyers, J . Chem. Thermodyn., 1981, 13, 1105. 28 G. Barone, B. Bove, G. Castronuovo and V. Elia, J . Solution Chem., 1981, 10, 803. 29 V. Abate, G. Barone, P. Cacace, G. Castronuovo and V. Elia, J . Mol. Liquids, 1983, 27, 59. 30 F. Shahidi, P. G. Farrell and J. T. Edward, J , Solution Chem., 1976, 5, 807. 31 J. J. Savage and R. H. Wood, J . Solution Chem., 1976, 5, 733. 32 A. D'Orsi, Thesis (University of Naples, 1987). 33 G. Di Paola and B. Belleau, Can. J . Chem., 1977, 55, 3825. ' Berlin, 1986), chap. 6. 1569. Paper 711048; Received 5th June, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402279

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chem. SOC., Furuduy Trans. I, 1988, 84(7), 2289-2295 Phase Transfer of the 12-Tungstosilicate Anion across the Water/Ni t ro benzene Interface Erkang Wang* and Yuqing Liu Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun, Jilin 130021, People's Republic of China The transfer behaviour of the heteropoly anion [SiW,204,]4- across the water/nitrobenzene interface has been investigated using cyclic voltammetry and chronopotentiometry with cyclic linear current scanning. It was found that the heteropoly anion with a negative charge of 4 can transfer from the aqueous phase to the organic phase electrochemically. The transfer process is diffusion-controlled at a lower scan rate, but it is partially kinetically controlled at a higher scan rate. The effects of some higher-valent cations such as Sr2+b, Co2+, Ni2+, Zn2+, Fe2+ and Fe3+, as well as some anions such as NO;, C1-, SO:- and PO:- etc.on the phase transfer have been systematically observed from cyclic voltammograms. The formal transfer potential and the transfer free energy for the heteropoly anion across the water/nitrobenzene interface have been calculated according to the voltammetric data. Recently, the simple and facilitated transfer of some ~ationsl-~ and anions6-' at the interface between two immisible electrolyte solutions (ITIES) has been successfully investigated. However, these studies were mainly restricted to the transfer of ions with low charge numbers because of the stronger solvation and higher Gibbs transfer energy of' higher-valent ions. Homolka and Wendt' studied the complex-assisted transfer of trivalent iron using the neutral ligand o-phenanthroline, whereas the phase transfer of anions was mainly studied using monovalent anions such as halides, 0x0-acid anions arid other polyatomic anions,* with the exception of Cr2O;-.lo Here we report for the first time that the phase transfer of the heteropoly anion [SiW,20,0]4- with more than three negative charges can occur at the water/nitrobenzene interface.The investigations of cyclic voltammetry and chronopotentiometry with cyclic linear current-scanning show that the transfer species is the tetravalent heteropoly anion [SiW12040]4- and the transfer process is diffusion controlled at the lower scan rate. Additionally, the effects of counter cations such as Sr2+, Co2+, Ni2+, Zn2+, Fe2+ and Fe3+ as well as the medium anions NO;, C1-, Ac-, SO:- and PO:- on the transfer process of the heteropoly anion halve been systematically observed using cyclic voltammetry.It is well known that heteropoly acids possess unique properties which are extremely valuable for catalysis. These acids are polynuclear complexes which are strong acids and are highly soluble and fairly stable in water and oxygen-containing organic solvents. However, the form the heteropoly acid takes in solution is difficult to determine, although some methods were used previously to analyse heteropoly acids, e.g. U.V. spectroscopy, elemental analysis, polarography etc.l1 In particular, the question of the charge numbers of the heteropoly anion in the solution remains unsolved and the changes of molecular structure cannot be observed directly when the conditions are modified.For these reasons some new techniques should be developed. Electrochemistry on the liquid/liquid interface is suitable for the investigation of heteropoly acids. 22892290 Phase Transfer of 1 2- Tungstosilicate Anion Hg/Hg,Cl, RE 1 Experiment a1 Materials Analytical-grade nitrobenzene used as the organic phase was further purified by washing with 0.1 rnol dmP3 HCl and doubly distilled water until its colour became pale yellow. 5 x l O-, mol dm-3 tetraphenylarsonium tetraphenylborate (TPAsTPB) prepared by the precipitation of sodium tetraphenylborate (Fluka, Puriss, p.a.) and tetraphenylarsonium chloride (Fluka, p.a.) in a water+thanol mixture (1 : 1) was employed as the supporting electrolyte of the organic phase.A solution of 0.1 mol dmP3 HAc-NaAC buffer or dilute HC1 at different pH values was used as the aqueous phase. All other reagents such as SrCl,, NiCl,, CoCl,, ZnCl,, FeCI, and FeC1, as well as acetic acid, nitric acid, sulphuric acid etc. are analytical grade. Commercial 12-tungstosilicic acid (H,SiW,,O,, * xH,O analytical reagent) was standardized alkalimetrically. HAc-NaAc buffer x mmol dmP3 W 5 x lop3 mol dm-, 5 x lop3 mol dmp3 or dilute HCl TPAsTPB TPAsCl Hg,Cl,/Hg H4Siw 12'4, NB W' RE2 The potential difference, E, measured is the Galvani potential difference A,"$ = #(W)-$(O) between the aqueous phase and nitrobenzene phase related to the formal potential difference for the tetraphenylarsonium cation, i.e.E = A: d - A," (A: = -0.372 V). Here the potentials of the Hg/Hg,Cl, reference electrodes RE1 and RE2 practically compensate each other. All experimentals were carried out at room temperature. Results and Discussion Voltammetric Behaviour of the Heteropoly Anion [SiW ,20,)4- across the Water/ Nitrobenzene Interface 12-Tungstosilicic acid is a strong acid regardless of whether it is in aqueous or non-aqueous ~ o l u t i o n . ~ " ~ ~ Four protons of 12-tungstosilicic acid are completely dissociated in the aqueous phase. Its acidity is much stronger than the ordinary mineral acids composed of its central atoms, e.g. H,P04 and H,SiO,, owing to its large size, better symmetry and lower charge density. These characteristics make the phase transfer of higher-valent heteropoly anions possible.Fig. 1 shows the voltammogram for [SiW,,0,,]4- across the water/nitrobenzeneE. Wang and Y. Liu 229 1 Fig. 1. Cyclic voltammogram for the transfer of the heteropoly anion [SiW12040]4- across the water/nitrobenzene interface. Aqueous phase: dilute HC1 (pH 2.0) + 0 (---), 4.1 x mol dm-3 H,SiWl,O,, (-). Organic phase 5 x lop3 mol dm-3 TPAsTPB, scan rate 5 mV s-'. 30 50 70 90 AEp/mV Fig. 2. Effects of scan rate on AEp, 6 and 4. Aqueous phase: dilute HCl (pH 2) + 4.1 x d m 3 H,SiW,,O,,. Organic phase: 5 x mol mol dm-3 TPAsTPB. (a) AEp us. v, (b) l$ us. v (V) and E; us. v (0) curves. interface. The dashed line is the background. The negative current corresponds to the transfer of the heteropoly anion from the aqueous phase to the organic phase, while the positive current is induced by the transfer of the heteropoly anion back from the organic phase.The voltammetric behaviour is described as follows : (a) the positive and negative !current peaks are proportional to the square root of the scan rate (2-250 mV s-'); (b) the peak currents are proportional to the concentration of 1Ztungstosilicic acid in the aqueous phase (10-5-10-3 mol dmP3); (c) the potential separation between two peaks AE, is 15 mV at a slower scan rate (v < 10 mV s-'); (6) the peak potentials, E,' and &, shift in the positive and negative directions with increasing sweep rate, respectively, but the peak potentials do not change at v < 10 mV s-'. The dependences of the peak separation, LIE,, and the positive and negative peak potentials, The transfer of the heteropoly anion [SiW,20,,]4- across the water/nitrobenzene interface is diffusion-controlled at lower v, while it is partially kinetically controlled at a larger v.The 15 mV separation shows that the transfer species is the tetravalent heteropoly anion. and E;, on the scan rate, v, are shown in fig. 2.2292 Phase Transfer of 12- Tungstosilicate Anion Fig. 3. Chronopotentiogram for [SiWl,0,0]4- transfer across the water/nitrobenzene interface. Aqueous phase: dilute HCI (pH 2)+0 (---), 4.1 x mol dm-3 H4SiWl,0,0 (-). Organic phase: 5 x lop3 mol dm-3 TPAsTPB. Current scan rate 2 pA s-l. Chronopotentiometry with Cyclic Linear Current Scanning of 12-Tungstosilicic Acid The chronopotentiogram for the heteropoly anion transfer across the water/ nitrobenzene interface is shown in fig.3. The dashed line was obtained when dilute HCl (pH 2) was used as the aqueous phase without 12-tungstosilicic acid. The transfer wave is proportional to the 2/3 power of the concentration of 12-tungstosilicic acid in the aqueous phase and 1/3 power of current scan rate. At lower current scan rates (v < 10 ,uA s-'), the curve of log [13/2/(1p/2 - I3l2)] us. E is linear with a slope of 15 mV. This further demonstrates that the charge number of the transfer species is four. The positive wave (fig. 3) obtained on the first scan corresponds to the transfer of the heteropoly anion distributed into the organic phase back from nitrobenzene. The first negative scan wave corresponds to the transfer of the heteropoly anion from the aqueous phase to the organic phase.On the second positive current-scanning stage, the anions transferred into the organic phase return to the aqueous phase, resulting in the second positive wave. The Effect of pH of the Aqueous Phase on the Transfer Process If the acidity of the solution is adjusted within the proper range, the shift of the transfer potentials with pH cannot be observed. At pH 1-5.7, the well developed peaks for the heteropoly anion across the interface appear, and the peak currents are almost independent of pH of the aqueous solution, indicating that the heteropoly anion is stable over the range pH 1-5.7. At pH > 5.7, the peak current begins to decrease and the shape of the transfer peak changes obviously, demonstrating that the heteropoly anion is partly dissociated.13-15 From voltammetric data, the formal transfer potential A: #' and the Gibbs energy for the heteropoly anion [SiW,,0,,]4- across the interface can be obtained. The value of the transfer potential is 0.423 V, and the Gibbs energy is calculated by the following equation: A:$" = AGY;"'"/ZF where 2 is the charge number and F is the Faraday constant. AGY;"+' is calculated to be 19.7 kJ mol-'. It can be concluded from this work that the lower transfer potential and free energyE. Wang and Y. Liu 2293 Table 1. The effect of medium anions on the transfer" Ei/mV &/mV AEJmV E,,,/mV 0.01 rnol dm-, HC1 438 423 15 43 1 0.01 rnol dmP3 HAc 431 416 15 424 0.01 mol dm-, HNO, 430 415 15 423 0.01 mol dm-, H,SO, 423 403 20 41 3 0.01 mol dm-, H,PO, 423 408 15 416 ~~~~ ~~~ ~ ~ ~ a The concentration of 12-tungstosilicic acid in the aqueous phase was 4.1 x rnol dm-, scan rate 5 mV s-I.40 20 0 --. d 4 -20 Fig. 4. Voltammograms of the heteropoly anion [SiW,,040]4- at the water/nitrobenzene interface. Aqueous phase: 0.04 m FeC13+4. 1 x mol dm-, H,SiWl,04,. Organic phase: 5 x lo-, mol dm-3 TPAsPB. Scan rate: (1) 10, (2) 5, (3) 100 and (4) 200 mV s-l. Table 2. The effect of counter anions on the transfer" q / m V EJmV AEJmV E,,,/mV 0.04 mol dmW3 Sr2+ 418 413 5 416 0.04 mol dm-3 Co2+ 412 413 8 417 0.04 mol dm-, Zn2+ 423 417 10 418 0.04 mol dm-, Ni2+ 42 1 41 3 8 41 7 0.04 mol dm-, Fe2+ 443 425 8 439 0.04 mol dm-, Fe3+ 428 42 1 7 424 " The concentration of H,SiW,,O,,: 4.1 x rate 5 pA s-l.mol dm-,; scan2294 Phase Transfer of 12- Tungstosilicate Anion are mainly subject to the characteristics of the heteropoly anion itself such as the larger molecular size, the lower surface charge density and the better symmetry. Hence the heteropoly anion can transfer from the aqueous phase to the organic phase. The Effect of Medium Anions on the Phase Transfer If medium anions such as NO;, C1-, Ac-, SO:- and PO:- are used in the aqueous phase, the anion effect can be conveniently observed from voltammograms. Table 1 shows the effect of different anions on the transfer potentials, the peak separation and the half- wave potentials for the transfer of the heteropolyanion. From table 1 we can see that the monovalent anions, NO;, C1- and Ac- have no effect on the transfer, while sulphate and phosphate anions effect the phase transfer slightly.The Effect of Counter Cations on the Phase Transfer Some divalent and trivalent cations such as Sr2+, Co2+, Ni2+, Zn2+, Fe2+ Fe3+ etc. effect the voltammetric behaviour of the transfer significantly. When 0.04 mol dm-, FeC1, was used as the aqueous phase the voltammograms shown in fig. 4 were obtained. At a low v, voltammograms do not obviously change although the separation AEp is much less than I5 mV. Here, Ei and E; are a function of scan rate, shifting in a positive and negative direction by an amount of 5-10 mV and more than 15 mV for each tenfold increase in v, respectively. Additionally, the positive and negative currents vary with increasing scan rate, but severe distortion of the positive peak occurs and its peak value is not proportional to vl/', which is different from the negative peak.Another phenomenon is that the baseline increases rapidly with scan rate and the peak becomes wider. The effects of high-valent cations on the transfer of the heteropoly anion are summarized in table 2. It is useful in discussing the cation effect to compare the result with medium anions. As no obvious anion effect on the phase transfer occurs, the heteropoly anion is stable in these media. However, the heteropoly anion forms easily the corresponding ion pairs with high-valent cations, and these ion pairs may interact at the interface during the electrolytic transfer. On the other hand, high-valent counter cations may influence more obviously the structure of the electrical double layer than monovalent cations at the polarized interface.Perhaps the latter is the main factor governing the voltammetric behaviour. From table 2, the effect of iron(r1j is the largest and results in a positive shift of the transfer potential. This may be due to the stronger interaction between the heteropoly anion and iron(1Ij. Other counter cations shift the transfer potential only slightly to more negative values. Conclusions The 12-tungstosilicate anion, with a negative charge of 4, can transfer electrochemically from the aqueous phase to the organic phase because of its large size, better symmetry and lower surface charge density. The diffusion-controlled process can be expected at a small v.The effect of pH on the transfer potentials cannot be observed when the acidity of the aqueous solution is adjusted, i.e. the transfer potential is independent of the pH of the aqueous phase, demonstrating that the hydrogen ion does not participate in the transfer process. The heteropoly anion [SiWI2O4,J4- is very stable in aqueous solution at pH 1-5.7. Counter cations considerably affect the transfer of the heteropoly anion, resulting in the decrease of the positive peak and the peak separation of less than 15 mV. For medium anions, the effect of sulphate anion on the transfer is more considerable than that of acetate, chloride, nitrate and phosphate anions, but medium anions do not influence the voltammetric behaviour. Further investigations in this field are under way in our laboratory.E. Wang and Y. Liu 2295 ‘This research was supported by the National Natural Science Foundation of China. References 1 J. Koryta, Electrochim. Acta, 1979, 24, 293. 2 E. Wang and Z. Pang, J . Electroanal. Chem., 1985, 189, 1. 3 T. Kakutani, Y. Nishiwaki, T. Osakai and M. Senda, Bull. Chem. SOC. Jpn, 1986, 59, 781. 4 Z. Yoshida and H. Freiser. J . Electroanal. Chem., 1984, 179, 31. 5 Z . Samec, D. Homolka and V. MareEek, J. Electroanal. Chem., 1982, 135, 265. 6 D. Homolka, Le Quoc Hung, A. Hofmanova, M. W. Khalil, J. Koryta, V. Maretek, Z. Samec, S. K . 7 E. Wang and Z. Sun, J. Electroanal. Chem., 1987, 220, 235. 8 S. Kihara, M. Suzuki, K. Maeda, K. Ogura and M. Matsui, J . Electroanal. Chem., 1986, 210, 147. 9 D. Homolka and H. Wendt, Ber. Bunsenges. Phys. Chem., 1985, 89, 1075. 10 E. Wang and Z. Sun, Estended Abstract no. 590, 169th Meeting, The Electrochemical Society, Boston, May 4 9 , 1986, p. 857. 1 1 R. Massart, Ann. Chem., 1968, 3, 507. 12 D. Homolka, V. MareCek, Z. Samec, K. Base and H. Wendt, J. Electroanal. Chem., 1984, 163, 159. 13 G. A. Tsigdinos, Top. Curr. Chem., 1978, 76, 1. 14 L. I. Lebedeva, Vestrz. Leningr. Univ. Fiz. Khim., 1976, 4, 128. 15 A. Tize and G. Herve, J. Inorg. Nucl. Chem., 1977, 39, 999. Sen, P. Bvanqsek, J. Weber and M. Biezina, Anal. Chem., 1980, 52, 1606. Puper 7/1127; Received 23rd June, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402289

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1988, 84(7), 2297-2303 Volume and Compressibility Changes in Aqueous Mixed-salt Solutions at 25 O C Kesharsingh Patil* and Girish Mehta Department of Chemistry, Institute of Science, Nagpur-440 001, India The excess molar volumes of mixing (Am Vex) and compression of mixing [A,(PY)""] have been calculated from the measured density and sound velocity parameters for aqueous binary and ternary solutions of NaC1-KC1, Li B r-KB r, Rb Br-K B r , CsB r-K B r, Me, NBr-K Br , Et ,NB r-Me,NB r and Bu,NBr-Me,NBr at 25 "C (in the concentration range 0.1-1 .O mol kg-' and at constant ionic strength of I z 0.2, 0.5 and 1.0 mol kg-' in ternary solutions). Am Vex and A,@V)ex exhibit parabolic behaviour as a function of y [the fraction of ionic strength due to an electrolyte (AX) in a mixture of two electrolytes (AX + BX)].Application of the Friedman equation to the Am Vex and A,(PV)'" values indicated the presence of cation-cation interactions and specificity of the cosphere overlapping effects, which in turn depend upon the nature of the ion pairs, the water structural effects and hydrophobic interactions in solutions. The results are discussed in terms of the combination of water structure-breaking and structure-making pro- perties of ions. The compressional results show that in alkali-metal halide mixtures, electrostatic interactions affect more water molecules in mutual volumes derived from the ionic cosphere overlapping or interact over greater distances than tetra-alkylammonium salts. The structural effect of water on aqueous ionic solutions is best interpreted by the concept of ion-solvent cosphere overlapping developed by Gurney' and Franks.Friedman3 has advanced a theory which includes the overlapping effect, termed the 'Gurney potential', by superimposing it on the coulomb, core and cavity potentials. It has been found useful to measure excess function of mixing for properties like volume (Am Vex), enthalpy (AmHex) and free energy (A,Gex) in ternary aqueous solutions at different ionic strengths so that information about structural changes can be In our previous studies7, * we reported the excess volume and excess compressibility parameters in some mixed-salt solutions where the excess compression values of solution and their variation with ionic strength indicate the effect of ion-ion interactions on the cosphere water in a subtle way.Here we report the excess molar volumes of mixing (Am Vex) and compression of mixing [A,@ V)ex] for aqueous solutions of NaC1-KC1, LiBr-KBr, RbBr-KBr, CsBr-KBr, Me,NBr-KBr, Et,NBr-Me,NBr and Bu,NBr-Me,NBr at 25 "C and at overall ionic strengths of I z 0.2, 0.5 and 1.0 mol kg-l. The results are compared with data of the systems reported earlier and discussed on the basis of the order-producing or order-reducing nature of the ion and the secondary effect arising from the cosphere overlapping. Experiment a1 KBr, NaCI, KCl, RbBr, CsBr, Me,NBr, Et,NBr and Bu,NBr were AR grade (B.D.H.), arid were dried in a vacuum desiccator for several days and then used directly, while LiBr (owing to its hygroscopic nature) was used straight from the bottle.All the solutions were freshly prepared in doubly distilled water. Binary solutions of electrolytes were 22972298 Volume and Compressibility Changes prepared on a molality basis (0.1-1.0 mol kg-'). The mixed aqueous solutions of electrolytes (i.e. ternary solutions) were prepared by mixing solutes directly with a weighed amount of water such that the ionic strength ( I z 0.2, 0.5 and 1.0) was kept constant. The mole fraction of the solute ( y ) , i.e. the fraction of ionic strength due to an electrolyte (AX) in a mixture of two electrolytes (AX-BX) was varied in the range 0.1-1 .O ionic mole fraction. All weights were measured on a Mettler balance (kO.05 mg). The densities of the solutions were measured directly using a calibrated density bottle (having a standard joint thermometer and side-limb with a ground cap for volume adjustment) of volume 25.145 k0.002 cm3 at 25 "C.The necessary buoyancy correction was applied. The density values were reproducible within 5 x lop5 g ~ m - ~ . The density data for aqueous binary solutions agreed well with the literature data for the systems studied.' The compressibilities were determined from sound velocity and density data using the Laplace equation. Sound velocity measurements were carried out using an ultrasonic interferometer (2 MHz) obtained from Mittal Enterprises (New Delhi) at 25 0.05 "C. The instrument was calibrated using water, methanol, carbon tetrachloride and benzene. The tem- perature of the cell was kept constant (50.05 "C) by circulating water through an ultrathermostat with temperature constancy of 0.02 "C.The frequency of the oscillator was also checked periodically using a frequency meter. The sound velocities obtained for water, methanol and carbon tetrachloride were 1497.5, 1104 and 922.5 m s-l, respectively, at 25 "C, agreeing well with the literature data."' l1 The velocity values were reproducible within f0.5 m s-' or better. The reliability of the measurements was checked by obtaining velocity data for aqueous sodium chloride solutions at 25 "C and comparing them with literature data. The details of these measurements are given elsewhere. l2 Results The excess molar volume of mixing (Am Vex) and adiabatic compression of mixing [Am(PV)ex] = - Am K, were calculated using the equations where V, V, and V, are the molar volumes of the ternary solutions and binary solutions of electrolytes AX and BX of appropriate concentrations, respectively, in 1 kg of water.Similarly p, pz and p3 are the compressibilities of the ternary solutions and binary solutions of electrolytes AX and BX at appropriate concentrations. The variation of Am Vex and Am K as a function of y and the variation of Am Vex (0.5, I ) and Am K (0.5, I ) as a function of I are shown in fig. 1-4. The isentropic compressibilities obtained can be converted into isothermal compressibility values if the specific heat and expansivity parameters for the studied solutions are a~ai1able.l~. l4 In the absence of these data, we have assumed that the isothermal and adiabatic compressibility values are the same because the correction term is always small in moderately concentrated sol- utions. Using Am Vex (0.5, I ) and A,,, K (0.5, I ) and following the Friedman equation,15 calculations of the parameters 2 [ = Am VeX/12y( 1 - y ) ] and K [ = Am K/12y( 1 -y)] were made.For the sake of comparison, Am Vex (0.5, I ) and Am K (0.5, I ) for the systems reported here and those reported earlier are collected in table 1.K . Patil and G. Mehta 2299 0 0.2 0.4 0.6 0.8 1.0 Fig. 1. Excess molar volumes of mixing (A,,, Vex) for H,O-NaCl-KCl (a), H,O-LiBr-KBr (A), H,,O-CsBr-KBr (0) and H,O-Me,NBr-KBr (V) plotted against the mole fraction y of NaCl (a), LiBr (A), CsBr (0) and Me,NBr (V) at constant ionic strength I = 1.02 at 25 "C.Fig. 2. Excess molar compression of mixing (A,,, K ) for H,O-NaCl-KCl (O), H,O-LiBr-KBr (A), H,O-CsBr-KBr (0) and H,O-Me,NBr-KBr (V) plotted against the mole fraction y of NaCl (a), LiBr (A), CsBr (0) and Me,NBr (V) at constant ionic strength I = 1.02 at 25 "C.2300 Volume and Compressibility Changes Fig. 3. Plots of A, Vex us. I for mixed-salt solutions at y = 0.5 NaCl (O), LiBr (a), RbBr (O), CsBr (V), Me,NBr (m), Et,NBR (0) and Bu,NBr (A) at 25 "C. 0, NaCI-XCI; A, LiBr-KBr; 0, RbBr-KBr ; V, CsBr-KBr ; H , Me,NBr-KBr ; 0, Et,NBr-Me,NBr ; A, Bu,NBr- Me,NBr. Fig. 4. Plots of A, K us. I for mixed-salt solutions. Symbols as for fig. 3.K . Patil and G. Mehta 2301 Table 1. Compilation of Am Vex (0.5, I ) and Am K (0.5, I ) values for different systems A,,, Vex A,,, K/106 cm3 system /cm3 mol-' atm-' mol-la NaC1-KC1 NaBr-KBrb LiBr-KBr R bB r-K B r CsBr-KBr Me,NBr-KBr Et,NBr-KBrb Bu,NBr-KBrb Et,NBr-Me,NBr Bu,NBr-Me,NBr Bu,NBr-Et,NBrb 0.055 0.1 0.1 I 0.18 0.23 0.52 0.76 1.29 1 .1 1.63 1.9 39 37 34 30 27 18 18 14 4 7 2 ~~ a 1 atm = 101 325 Pa. Ref. (8). Discussion Fig. 1 and 2 show that A, Vex and A, K exhibit parabolic behaviour as a function of y in the systems studied. Also A, Vex increases with increasing ionic strength (fig. 3), while A, K increases in most of the systems, but their variation in the Bu,NBr-Me,NBr and Et,NBr-Me,NBr systems is comparatively small (fig. 4). These observations reveal the presence of cation<ation interactions in solutions involving tetra-alkylammonium ions. The variation of A, Vex and A, K with ionic strength at 0.5 y further indicates that at higher ionic strength ( I = 1.0) for mixtures of two alkali halides the A, Vex values are small and the A, K values are comparatively large, while for mixtures of two tetra- alkylammonium halides and also for mixtures of alkali-metal halides with tetra- alkylammonium halides, A, Vex values are large but A, K values are very small.Thus, A, Vex and A, K can be correlated with each other (see table 1). Recently Miller0 and Lampreia reported a correlation between A, V and A, K parameters for the mixing of NaCl with electrolytes containing a common ion (i.e. Na+ or C1-).16 Fig. 3 of this paper and fig. 3(a) of our previous paper8 indicate that the molar excess volumes of mixing (A, Vex) for mixtures of tetra-alkylammonium salts are almost additive at low ionic strengths ( I = 0.2), with the excess volumes of mixing of tetra- alkylammonium salts with KBr, while no such comparison can be made for the molar compression of mixing (Am K ) .Fig. 2 and 4 show that with increasing ionic strength the molar compression values for mixtures of alkali-metal halides (KCl-NaCl etc.) increase to a large extent compared to the values for mixtures of alkali-metal halides with tetra- alkylammonium halides or even for mixtures of two tetra-alkylammonium halides (where they remain more or less constant). These findings can be rationalized in terms of following explanation. It is known that Li+ and Na+ are order-producing and that K t , Rb+ and Cs+ are order-destroying in aqueous solutions. Seen in this light we can classify our systems into three groups as ( 1) order-producing-order-destroying (NaC1-KC1, NaBr-KBr, LiBr-KBr, Et,NBr-KBr and Bu,NBr-KBr, where we assume that Et,N+ is slightly order-producing ; (2) order-destroying-order-destroying (RbBr- KBr, CsBr-KBr, Me,NBr-KBr) and (3) mixtures of ions having a large number of hydrophobic groups (Et,NBr-Me,NBr, Bu,NBr-Me,NBr and Bu,NBr-Et,NBr).If we study the variation of A, Vex (0.5, I ) and A, K (0.5, I ) with the ionic radius of the cations with KBr in aqueous solution, we find that as the ionic radius of the cation increases, A, Vex increases and Am K decreases. This shows that the volume of cosphere2302 Volume and Compressibility Changes Table 2. The parameters Z and K derived by applying Friedman's equation Z/cm3 m01-~ K / lo6 cm3 atm-I m ~ l - ~ - system U b C U b C NaCI-KCl 2.2 0.6 0.4 840 380 150 LiBr-KBr 2.0 1.0 0.4 730 230 135 R b B r-K B r 3.6 1.6 1.0 720 270 120 CsBr-KBr 4.0 1.6 1.0 565 185 110 Me,NBr-KBr 10.0 4.0 2.0 380 195 70 Et,NBr- 22.0 9.0 4.4 450 45 15 Bu,NBr- 33.6 13.2 6.2 180 95 55 Me,NBr Me,NBr = 0.2 I.* = 0.5 I. = 1.0 I. water displaced by ion-ion interaction increases, but the strength of the charge-dipole (water) interaction decreases. Thus for Li+, Na+, Rbf and Cs+ the coulombic interactions are predominant and they affect more water molecules in the B zone of the Frank and Wen1' model of electrolytic solutions, while the tetra-alkylammonium ions, because of the small charge densities and the hydrophobic attractions of the non-polar groups, although the volume displaced by cosphere overlapping is greater because of the structure-making effect, the loss in compressibility of the water (and hence of the cosphere water) surrounding the ions is very small.This interpretation is in agreement with the observed Am Vex and Am K values for the group 3 ions of our scheme given above. The 2 values (table 2) are in the order NaCl-KCl M LiBr-KBr < RbBr-KBr < CsBr-KBr < Me,NBr-KBr < Et,NBr-Me,NBr < Bu,NBr-Me,NBr. From these values we observe that for our group 1 systems the 2 values and their variation with ionic strength is small, while for group 3 systems high 2 values occur as I -+ 0 and their variation with I is also very large. Thus, in case of group 3 systems, because of the hydrophobic hydration, a large change in 2 occurs which can be attributed to the presence of cationsation as well as triplet interactions or changes in the activity coefficient of cation<ation pairs in solution, as suggested previously." The parameter K describes the compressional effect on water surrounding the ions (as the ions themselves are assumed to be incompressible).The ionic strength dependence of K for our group 1 systems supports the notion that at least for alkali-metal halide solutions, the long-range coulombic effects are important and strongly affect the cosphere water, while at higher concentrations and in the case of group 3 systems the cosphere overlapping effects are predominant and have their origin in water structural effects, which in turn depend on the degree of dissimilarity of order-disorder characters of overlapping ions.We thank Prof. R. B. Kharat and Prof. (Mrs) Sapkal for facilities and encouragement during the course of this work. References I R. W. Guney, Ionic Process in Solutions (Dover, New York, 1962). 2 H. S. Franks, Chemical Physics of Ionic Solutions, ed. B. E. Conway and R. G. Bawadas (John Wiley, New York, 1965).K. Patil and G. Mehta 2303 3 H. L. Friedman and C. V. Krishnan, in Water: a Comprehensive Treatise, ed. F. Franks (Plenum Press, 4 R. A. Robinson, R. H. Wood and P. J. Reilly, J . Chem. Thermodyn., 1971, 3, 461. 5 Y. Ch. Wu, in Structure of Water and Aqueous Solutions, ed. W. A. P. Luck (Verlag Chemie, Weinheim, 6 H. L. Anderson and R. H. Wood, in Water: a Comprehensive Treatise, ed. F. Franks (Plenum Press, 7 K. J. Patil and G. R. Mehta, Ind. J . Chem., 1986, 7, 635. 8 K. J. Patil and G. R. Mehta, J . Chem. Soc., Faraday Trans. I , 1987, 83, 2467. 9 J. L. Fortier, P. A. Leduc and J. E. Desnoyers, J . Solution Chem., 1974, 3, 323; W. Y. Wen and S. New York, 1973), vol. 3, chap. 1. 1974). New York, 1973), vol. 3, chap. 2. Saito, J. Phys. Chem., 1964, 68, 2639. 10 0. Kiohara, P. J. Darey and G. C. Benson, Can. J . Chem., 1979, 57, 1006. I 1 W. Schaaffs, Molekularakustic (Springer Verlag, Berlin, 1963). 12 A. B. Wazalwar, Ph.D. Thesis (Nagpur University, 1985). 13 H. S. Harned and B. B. Owen, The Physical Chemisfry ofElectrolytic Solutions (Reinhold, New York, 14 K. J. Patil and D. N. Raout, Ind. J . Pure Appl. Phys., 1980, 18, 499. I5 H . L. Friedman, J . Chem. Phys., 1960, 32, 1 134. 16 F. J. Miller0 and M. I. Lampreia, J . Solution Chem., 1985, 14, 853. 17 H. S. Franks and W. Y. Wen, Discuss. Faradq Soc., 1957, 24, 133. 18 W. Y. Wen and K. Nara, J . Phys. Chem., 1967, 71, 3907; W. Y. Wen, K. Miyajima and A. Otsuka, 1958). J . Phys. Chem., 1971, 75, 2148. Paper 7/ 1 153 ; Received 26th June, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402297

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1988, 84(7), 2305-2309 Electron Spin Resonance of a y-Irradiated Single Crystal of Carbamylcholine Chloride Fevzi Koksal* and Fahri Celik Faculty of Arts and Sciences, Ondokuz Mayis University, Samsun, Turkey Electron spin resonance spectra of a y-irradiated carbamylcholine chloride, H,NCOOCH,CH,N(CH,),Cl, single crystal has been investigated at room temperature, and the spectra were found to be independent of temperature down to 130 K. The radical was proved to be -CHCH,-, and the g factor and the hyperfine coupling constant of P-protons were found to be isotropic, with the values of 2.0029 and 27.6 G, respectively. The principal values of the hyperfine tensor of the a-proton were found to be A,, = 12.0 G, A,, = 27.6 G and A,, = 3.9 G. The results were found to be in agreement with the a-and P-proton coupling and are discussed.The radiation sensitivity of choline chloride [OHCH,CH,N(CH,),]Cl has long been recognized,', and it was observed that crystalline choline chloride rapidly decomposes trimethylamine hydrochloride and acetaldehyde. Selective deuteration of choline chloride has been employed to establish the localization of the unpaired electron in the radical using e.s.r. techniques., Although the spectra were not well resolved it was concluded that the observed radical is 'CH,CH,OH. Furthermore, the e.s.r. spectra of radiation-damage centres in choline chloride were interpreted* as a biradical, (CH,),N+. . .CH,CH,OH. In addition, e.s.r. spectra of choline iodide, choline sulphate, [(CH,),NCH,CH,Cl]Cl, [(CH,),NCH,CH,CH,OH]Cl and [(C,H,),NCH,CH,OH]Cl, were examined, and it was found that the radicals formed in all anologues except choline chloride and choline bromide were not ethanol radicals ; however, detailed inter- pretations were not attem~ted.~ To our knowledge no further e.s.r.studies have been carried out on choline derivatives, and it is the purpose of this study to investigate an analogue of the abovementioned compounds, carbamylcholine chloride, in the hope of determining its radical structure after y-irradiation. Experiment a1 The carbamylcholine chloride single crystals were grown in the laboratory from concentrated ethanol solutions. In its single-crystal form carbamylcholine chloride, H,NCOOCH,CH,N(CH,),~1, is orthorhopbic6 with spaceogroup P,,,, and with unit- cell dimensions a = 10.248 A, b = 12.850 A and c = 6.809 A. The crystals were irradiated at room temperature by a 6oCo y-ray source of 0.3 Mrad h-' for 5 h.The e.s.r. spectra were recorded with a Varian model E-109 C e.s.r. spectrometer using 1 mW microwave power. The low-temperature measurements were carried out using a Varian temperature-control unit. The crystals were rotated on a Lucite pillar about the three axes shown in fig. 1, and the angle of rotation were read on a scale graduated in degrees. Because the crystals were hygroscopic they were coated with a thin film of collodion to protect them against moisture during the observations. The g factor was found by comparison with a DPPH sample ( g = 2.0036). 23052306 I E.S.R. of y-lrradiated Carbamylcholine Chloride Fig. 1. Crystal habit and the axes of carbamylcholine chloride. 20 G &----+ Fig. 2. E.s.r. spectrum of y-irradiated carbamylcholine chloride obtained along the H , l a and H, I1 b axes. Results and Discussion The e.s.r. spectra of carbamylcholine chloride exhibit mainly three different patterns as shown in fig. 2-4. The spectra shown in fig. 2,3 and 4 appear when the magnetic field is parallel to the b, c and a axes, respectively. The spectra are independent of temperature between 300 and 130 K. The g value does not change when the crystal is rotated about the a, b and c axes, and therefore is very nearly isotropic within experimental error; its measured value is g = 2.0029+0.0005. Thespectrashowninfig. 2, 3and4exhibit 1:3:3:1,nearly 1:2:1 and 1:1:2:2:1:1 patterns, respectively, and therefore indicate the hyp$rfine interaction of three protons with the electron spin; the radical must thus be -CHCH,-.Site splitting does notF. Koksal and F. celik 2307 20 G - Fig. 3. E.s.r. spectrum of y-irradiated carbamylcholine chloride obtained along the H , l a and H,, 11 c axes. 20 G - n V Fig. 4. E.s.r. spectrum of y-irradiated carbamylcholine chloride obtained along the H,, I b and H,, 11 a axes. occur, and therefore the molecules in the unit cell of the crystal are magnetically equivalent. The spectra can be fitted to the spin Hamiltonian Z = g/3Ho S, + S. ZiAi* Ii (1) where /3 is the Bohr magneton and H,, is the applied magnetic field. The first term represents the interaction of the electron spin with the proton spins (Ii = :).The2308 E.S.R. of y-Irradiated Carbamylcholine Chloride I I I 1 I I 180 < - 160 - 140 ' - 120 - ?. 100 - 80 - s 60 - 40 - 20 - 0 - M 5 10 15 2 0 2 5 30 A /G Fig. 5. Variation of the a-proton coupling constant in -CHCHz- radical when the crystal is rotated about the a(O), b(O) and c(e) axes. Table 1. Principal values of the a-proton coupling in the -CHCHz- radical produced in carbamylcholine chloride by y-irradiation principal values anisotropic values direction /G /G cosines Am 12.0 - 2.5 1 0 0 A cc 3.9 - 10.6 0 0 1 27.6 13.1 0 1 0 variations of the a-proton coupling about the a, b and c axes are shown in fig. 5. The principal values of the hyperfine tensor of the a-proton are given in table 1. The isotropic value of the hyperfine coupling of the a-proton is a, = - 14.5 G, and the anisotropic values are given in table 1.McConnell and Strathdee' have calculated the pure dipole-dipole coupling of a CH proton with an electron entirely in a p orbital on the carbon atom. They obtain the values 15.6 G along the CH bond, - 1.7 G along the p orbital and - 13.9 G for the direction perpendicular to both the CH bond and the p orbital. The anisotropic values in table 1 are in agreement with these values and the a- proton coupling constants given in the literature.8 The variations in the spectra indicate that the two a-protons are equivalent, and their hyperfine coupling constant is isotropic with a value of 27.6 G. The /?-proton (2) coupling constant is given by9 where 0 is the angle between the C,-H, bond and the p orbital direction, projected perpendicular to the C,-C, bond.Bo is a constant and includes the contributions from spin density which arise from conformation-independent mechanisms, in particular spin A, = Bo+ B , C O S ~ ~F. Koksal and F. celik 2309 polarisation, and B, includes the hyperconjugative contributions. In the case of rapid free rotation about the C,-C, bond the average value of A, becomes a, = B,+iB,. (3) Replacing the very pronounced value^,^.^* B, = (r3.5 G and B, = 50 G in eqn (3), we obtain a, = 28.5 Gs; and this is in good agreement with our experimental result, 27.6 G. Hence we conclude that the /?-protons in the -CHCH2- radical rotate about the C,-C, bond. Since we observed the equivalence of the P-protons down to 130 K, we can state that this rotation of 8-protons exists at these temperatures.As as a result this study shows that the ethanol radical in choline chloride and choline bIomide4y5 does not occur in carbamylcholine choloride; this is in agreement with the conclusion by Nath et aL5 for choline iodide, choline sulphate, [(CH,),NCH,CH,CI]Cl, [(CH,),NCH,CH,OH]Cl and [(C,H,),NCH,CH,OH]Cl. Furthermore, at moderate microwave power, 1-20 mW, there is no trace of the biradical or (CH,),N+ as in choline ~hloride.~ However, when the microwave power is raised above 50 mW the spectra are less clear, probably because of the unknown species. Therefore, further e.s.r. work on y-irradiated choline derivatives will be carried out at various microwave powers and temperatures in this laboratory. This work was supported partly by the Research Fund of Ondokuz Mayis University. References 1 B. M. Tolbert, J . Am. Chem. Soc., 1953, 75, 1867. 2 R. M. Lemmon, M. A. Parsons and D. M. Chin, J. Am. Chem. SOC., 1955, 77,4139. 3 R. 0. Lindblom, R. M. Lemmon and M. Calvin, J . Am. Chem. SOC., 1961, 83, 2484. 4 Y. Tomkiewicz, R. Agarwal and R. M. Lemmon, J. Am. Chem. SOC., 1973, 95, 3144. 5 A. Nath, R. Agarwal and R. M. Lemmon, J. Chem. Phys., 1974, 61, 1542. 6 B. Jensev, Acta Chem. Scand., Part B, 1975, 29, 891. 7 H. M. McConnell and J. Strathdee, Mol. Phys., 1959, 2, 129. 8 D. V. G. L. Narasimha Rao and W. Gordy, J. Chem. Phys., 1961, 35, 362. 9 J. R. Morton, Chem. Rev., 1964, 64, 453. 10 N. M. Atherton, Electron Spin Resonance (John Wiley, New York, 1973). Paper 711260; Received 14th July, 1987 16 FAR I

ISSN:0300-9599    

DOI:10.1039/F19888402305

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J . Cheln. Soc., Faraday Trans. I, 1988, 84(7), 2311-2315 Surface Reactions of Goethite with Phosphate Ralph G. Jonasson,? Ronald R. Martin," Marco E. Giuliacci and Kazui Tazaki Department of Chemistry, The University of Western Ontario, London, Ontario N6A 5B7, Canada Reactions carried out between synthetic goethite (a-FeOOH) and aqueous phosphoric acid (1 mmol dm-3) at pH 3 at 25 "C for 18 days yield mixtures of goethite and tinticite [Fe,(PO,),(OH), .7H,O], as has been shown by scanning transmission electron microscopy (STEM) and selected-area electron diffraction. X-Ray photoelectron spectroscopy (X.P.S.) indicated the presence of small, variable quantities of phosphate (ca. 2%). The presence of phosphate was indicated by X.P.S. only after mild grinding of the product, indicating that sorption of phosphate at the goethite surface is relatively unimportant.(X.P.S. is particularly sensitive to the first few atomic layers in a solid.) The results are important in resolving the difficult question of phosphate availability in soils, particularly in the presence of iron oxides. Two fundamentally different hypotheses regarding the uptake of phosphate by soils, and mineral surfaces in general, have been presented in the literature. The model of Parfitt et al.,', for example, involves the formation of inner-sphere surface complexes between phosphate and such minerals as goethite and other iron hydroxyoxides. The models of Kittrick and J a c k s ~ n , ~ Taylor and Gurney,* and Nriagu and Dell5 describe the immobilization in terms of the replacement of the substrate by a phosphate mineral.It is clear that given enough time, mineral replacement will occur.' The crucial question is whether the short-term surface reactions involve surface complexation or some form of incipient mineral replacement. Certainly, it is possible that both processes may operate simultaneously or sequentially. Madrid and De Arambarri7 showed that the sorption us. time curve has two components, which may correspond to these two processes, but their relative importance needs to be determined for a variety of systems if realistic models of soil evolution and phosphate immobilization are to be developed.8 The development of surface analytical techniques has allowed an increasingly sensitive microscopic view of the surfaces involved in these reactions.STEM has been used3 specifically to look at the surfaces of minerals for evidence of replacement by phosphate minerals. Given sufficiently extreme temperatures (90 "C) and reaction times (e.g. days), newly formed crystalline phosphate phases were observed when the minerals greenalite, (Fe2, Fe3),-,Si20,(OH), and kaolinite reacted with phosphate. The authors speculated whether the use of a more powerful microscope would allow the detection of new phases at earlier reaction times, or for lower reaction temperatures. This work was originally intended to extend X-ray photoelectron spectroscopy (X .p.s.) studies of phosphate adsorption on goethite to other anionic systems. However, the study failed t o reproduce previous X.P.S. work on the phosphate ~ y s t e m .~ Since the previous results had been interpreted using an adsorption mechanism, we sought to extend the study to include scanning transmission electron microscopy (STEM), electron diffraction, X-ray powder diffraction (X.r.d.) and atomic absorption spectroscopy. t' Present address: 62 Juanita Drive, Hamilton, Ontario, Canada L9C 2G3. 231 1 76-22312 Surface Reactions of Goethite with Phosphate Experiment a1 Goethite samples were prepared by the method of Atkinson et aZ."*" using reagent- grade chemicals. This procedure yielded a goethite slurry with approximate con- centration and surface area of 0.32 g and 85 m2 g-l, respectively. These samples were characterized by powder X.r.d. The first set of samples (pH 3, 5, 7, 9, 11) was prepared by adding sufficient concentrated H,PO, to 10 cm3 of goethite slurry to produce a solution 0.01 mol dmP3 in phosphate, and adjusting the pH by addition of KOH.The samples were allowed to equilibrate overnight, the solids were separated by centrifugation (2100 r.p.m.), washed and dried. The surface atom YO phosphorus was then measured using X.P.S. mol dm3 in H3P04 (pH 3) and shaken with an automatic shaker for 18 days. This process resulted in distinct layers of solid sediment. These materials were subjected to X.P.S., X.r.d., STEM and electron diffraction analysis. mol dmP3 in H,P04 at pH 3, 5, 7, 9 and 11, and allowed to equilibrate for 10 days (with occasional manual shaking). The supernatant solution obtained from these samples after centrifugation at 2 100 r.p.m.was subjected to further centrifugation at 20000 r.p.m. to remove any colloidal iron from solution. The resulting liquid was subjected to atomic absorption analysis. The instruments used were : X.P.S., Surface Science Laboratories SSX-100 ESCA spectrometer; X.r.d., Rigaku X-ray diffractometer ; STEM, JEOL JEM lOOC instrument at 200 kV; atomic absorption, I.L. 551 spectrophotometer. A second single sample was prepared, 1 x Thirdly, a range of samples was prepared, 1 x Results and Discussion The surface atom percent phosphorus in four separate determinations on each sample is shown in table 1 ; a typical X.P.S. broadscan is shown in fig. 1. The variation is well beyond the accepted accuracy of the instrument (k 15 O/O>, in contrast to results previously reported.' The difference can be readily explained by the different areas of material analysed.The Surface Science instrument analysed an oval spot having a maximum diameter of 1000 pm, while the previous results' were obtained with a Perkin- Elmer Physical Electronics Division (PHI) model 560 ESCA/SAM/SIMS System from 6 x 6 mm area. Thus the PHI system results are more likely to produce an averaged result if local concentrations of phosphate-rich material are randomly distributed throughout the sample. Since this latter result suggests that the occurrence of phosphate is non-uniform we investigated the possibility of the precipitation of a new phosphate phase. Accordingly, we examined the material resulting from exposure of goethite to phosphate for 18 days with shaking.The resulting sample yielded a layered sediment. The bottom layer gave X.r.d. patterns identical to that of goethite; however, the top layer indicated the presence of new material consisting of either very small crystals or poorly crystallized solids as evidenced by very broad X.r.d. peaks (fig. 2). An initial STEM study of this layer showed masses of crystallites (plate I), which electron diffraction patterns showed to be dominated by goethite with some haematite and magnetite, the latter being an artifact of electron-beam damage (plates 24). In addition, X.P.S. analysis showed little or no phosphorus to be present. The sample was ground with an agate mortar and pestle after which X.P.S. showed 3.58 surface atom YO phosphorus. This material was sonicated and examined again with STEM and electron diffraction, which clearly showed the presence of tinticite, Fe,(SO,),(OH), - 7H20 (plate 5).Scanning electron micrographs obtained by Stringham12 showed samples of natural tinticite to consist of rectangular plates 1.6 ,urn long and 0.3 ym thick. This crystal habit was not preserved in the grain displayed in plate 5. It is difficult to escape the conclusionJ . Chern. SOC., Faraday Trans. 1, Vol. 84, part 7 Plate 1 Plate 1. Scanning transmission electron micrograph of crystal masses obtained after exposure of goethite to phosphate for 18 days. R. G. Jonasson et al. (Facing p . 23 12)J . Chem. Soc., Furuduy Trans. I , Vol. 84, part 7 Plate 2 Plate 2. Scanning transmission electron micrograph and electron dicraction pattern from individual goethite crystals from selected regions in plate 1.d-spacings (A): 5.1 1 1 , 4.182, 2.771, 2.556, 2.2931, 1.9167, 1.6788, 1.5033. R. G. Jonasson et ul.J . Chem. SOC., Faraday Trans. I , Vol. 84, part 7 Plate 3 Plate 3. Scanning transmission electron micrograph and electron diffrtction pattern from individual haematite crystals from selected regions in plate 1. d-spacings (A): 3.67, 2.69, 2.204, 1.838, 1.692, 1.484. R. G. Jonasson et al.J . Chem. SOC., Faraday Trans. I , Vol. 84, part 7 Plate 4 Plate 4. Scanning transmission electron micrograph and electron dicraction pattern from individual magnetite crystals from selected regions in plate 1. d-spacings (A) : 4.842, 2.968, 2.473, 2.284, 1.608, 1.484, 1.211. R.G. Jonasson et al.J . Chem. SOC., Faraday Trans. 1, Vol. 84, part 7 Plate 5 Plate 5. Tincticite crystal and its electron diffraction poattern obtained by sonication of crystal masses similar to the one shown in plate 1. d-spacings (A) : this study : 6.59, 4.6 1, 4.01 , 3.20, 2.09, 1.563; ref. (11): 6.70, 4.56, 3.91, 3.28, 2.09, 1.551. R. G. Jonasson el al.R. G . Jonasson et al. 2313 1000 0 binding energy/eV Fig. 1. X-Ray photoelectron spectrum of a goethite surface exposed to phosphate solution at pH 3. Table 1. PO:- adsorption on goethite as a function of pH run 1 run 2 run 3 run 4 pH P (atom %) P (atom %) P (atom %) P (atom %) 3 1 .so 1.78 1.14 1.90 5 1.21 1.58 1.17 1.59 7 1.54 1.30 1.39 1.69 9 1.11 1.46 1.22 1.71 11 1.29 I .73 0.95 1.93 I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 26 24 22 20 18 16 14 12 10 8 6 angle/ O Fig. 2.X-Ray diffraction pattern obtained from material from goethite after exposure to phosphate solution for 18 days at pH 3. that precipitation of new crystal phases is a major mechanism in the uptake of phosphate by goethite. The next series of samples was prepared exclusively to measure the concentration of iron in solution, both as a function of pH and phosphate concentration, since the formation of Fe3+ complexes with phosphate in solution is well known. The sequestration of Fe3+ by phosphate should increase the solubility of the goethite and hence promote formation of new insoluble iron phosphate phases. The results shown in fig. 3 support this suggestion. The results indicate that the tinticite formed in the reaction between goethite and dilute phosphoric acid at pH 3, constitutes only a small proportion of the final solid mixture.Nriagu and Dell5 constructed a number of Eh us. pH diagrams for various iron phosphates and basic iron(Ir1) phosphates. These diagrams demonstrate that tinticite23 14 Surface Reactions of Goethite with Phosphate PH Fig. 3. Concentration of Fe3+ in pure goethite (a) and goethite phosphate (b) slurries. and other basic iron(m) phosphates are stable over a considerable pH range (3-12) under oxidizing, aqueous conditions. This is consistent with the electron diffraction results obtained for the reaction products obtained at pH 3. It has been suggested that the solubility of goethite is much too low for replacement reactions with phosphate to take ~ 1 a c e .l ~ The fact that various iron(m) iron precipitates slowly interconvert via the solution phase14 suggest that this is not true. Complexation of dissolved iron(II1) should enhance this recrystallization by raising the total dissolved iron concentration. Analyses of the supernatant in these systems by atomic absorption spectroscopy indicated total iron concentrations of ca. 2.5 ppm in phosphate-containing systems, versus 1.5 ppm in supernatants without phosphate. The surface complexation model is not consistent with the X.P.S. data, since such a model predicts that the phosphate is concentrated at the surfaces of the goethite grains. Such a surface saturated with phosphate ought to give a strong X.P.S. signal. In fact, the X.P.S.shows negligible P concentrations in the unground sample. Conclusion Our results are inconsistent with a surface adsorption model for phosphate uptake by goethite and present unequivocal evidence for the formation of at least one new mineral phase. These results, obtained after a longer equilibration period than other similar studies, are closer to true thermodynamic equilibrium and suggest that the classical adsorption models for this system should be re-examined. We acknowledge the support of the Department of Chemistry at The University of Western Ontario and the assistance of Margaret Hyland and Richard Mycroft in the acquisition of X.P.S. data. Finally, we thank Mr J. Young who operated the atomic absorption spectrometer.R. G. Jonasson et al. References 2315 1 R. L. Parfitt, R. J. Atkinson and R. St. C. Smart, Soil Sci. Soc. Am. Proc., 1975, 39, 837. 2 R. L. Parfitt, J. D. Russell and V. C. Farmer, J. Chem. Soc., Faraday Trans. I , 1976, 72, 1082. 3 J. A. Kittrick and M. L. Jackson, J. Soil Sci., 1956, 7, 81. 4 A. W. Taylor and E. L. Gurney, Soil Sci. SOC. Proc., 1965, 29, 18. 5 J. 0. Nriagu and C. I. Dell, Am. Mineral., 1974, 59, 934. 6 N. D. Warry, Thesis (McMaster University, Hamilton, Ontario, 1972). 7 L. Madrid and P. De Arambarri, J. Soil Sci., 1985, 36, 523. 8 G. Sposito, The Surface Chemistry of Soils (Oxford University Press, Oxford, 1984). 9 R. R. Martin and R. St. C. Smart, Soil Sci. SOC. Am. J., 1987, 51, 54. 10 R. J. Atkinson, A. M. Posner and J. P. Quirk, J. Inorg. Nucl. Chem., 1968, 30, 2371. 11 R. J. Atkinson, A. M. Posner and J. P. Quirk, Clays and Clay Mineral., 1977, 25, 49. 12 B. Stringham, Am Mineral., 1946, 31, 395. 13 R. J. Atkinson, A. M. Posner and J. P. Quirk, J. Inorg. Chern., 1972, 34, 2201. 14 G. W. van Oosterhout, J. Inorg. Nucl. Chem., 1967, 29, 1235. Paper 7/ 1378 ; Received 29th July, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402311

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chem. SOC., Faraduy Trans. I, 1988, 84(7), 2317-2326 Conductivity of Polypyrrole Films doped with Aromatic Sulphonate Derivatives Susumu Kuwabata, Koichi Okamoto and Hiroshi Yoneyama* Department of Applied Chemistry, Faculty of Engineering, Osaka University, Yamadaoka 2-1, Suita, Osaka 565, Japan The conductivity of polypyrrole films doped with various kinds of aromatic sulphonate derivatives has been studied. The conductivity of polypyrrole films could be varied by more than an order of magnitude by the current density chosen in the film preparation, and the highest conductivity was achieved at moderate current densities such as 0.2 rnAcmw2. The film conductivity was also influenced greatly by the number of sulphonate groups of a dopant molecule in such a manner that an increase in the number caused the conductivity to decrease.This effect was especially noticeable for polypyrrole films prepared at relatively low current densities such as 0.005 mA cm-2. It is well known that polypyrrole films prepared by electrolytic polymerization of pyrrole show high electric conductivities. The mechanism of the appearance of the high conductivity has been widely investigated, and it has been clarified from optical and e.s.r. that polarons and bipolarons in the polymer films serve as positive charge carriers. In addition, much attention has been paid to a variety of factors which control the conductivity of polypyrrole films. The conductivity of a polypyrrole film is primarily controlled by adjusting its potential in an electrolytic but it is also influenced by another factor such as the concentration of a supporting electrolyte,' and the kinds of electrolyte anions in the preparation bath.1°-12 Furthermore, it was reported that stretching the polymer films increased their conductivities.l3 The effect of dopant anions on properties of the polypyrrole film has recently become of importance, since the use of aqueous electrolytes allows the preparation of polypyrrole films doped with a variety of anions having special functions such as electrocatalysis and electrochromic behaviour, as demonstrated, for example, in the polypyrrole films doped with phthalocyanines and porphyrins. 14-16 However, there are few systematic investigations with regard to the effect of the kinds of dopant anions on the conductivity of polypyrrole films.In a previous paper17 we briefly reported that the conductivity of polypyrrole films doped with organic sulphonate derivatives decreased with increasing sulphonate groups of a dopant molecule. Later, it was found that the current density used for the film preparation strongly influences the conductivity of the resulting polymer film. Furthermore, the effect of the number of sulphonate groups of a dopant was found also to be influenced greatly by the current densities of the film preparation. In this paper these findings will be reported in detail and discussion will be focused on how such events appear, based on X-ray diffraction patterns, visible absorption spectra and the electrochemical behaviour of the prepared films. Experiment a1 Pyrrole was purified by distillation under N,.Sodium salts of naphthalene-2-sulphonate (N- l), naphthalene- 1,5-disulphonate (N-2), and naphthalene- 1,3,6-trisulphonate (N-3) were purified first by dissolving in hot water, followed by filtration, extracting the filtrate 23172318 Conductivity of Polypyrrole Films with benzene several times and finally evaporating water. Sodium salts of anthraquinone- 2-sulphonate (A- l), anthraquinone-2, 6-disulphonate (A-2) and copper phthalocyanine- 3,4',4",4"'-tetrasulphonate (P-4) were of reagent grade and were used without further purification. Polypyrrole (PPy) films were prepared by electropolymerization on indium tin oxide- coated glasses (ITO) for measurements of conductivities, visible absorption spectra and elemental analyses, and on Pt plates for electrochemical measurements and X-ray diffraction analyses.These films were prepared by anodic polymerization of 0.1 mol dm-3 pyrrole dissolved in water containing one of the abovementioned aromatic sulphonate derivatives (0.01 mol dm-3) as a supporting electrolyte which served as a dopant. The quantity of electricity chosen for the polymerizations was 1.2 C cm-, for X-ray diffraction analyses and 200 mC cm-, for the other measurements. The conductivity of the PPy films was determined by the four-probe method. For this purpose, the PPy film was torn off from the electrode substrate and its thickness was determined by observations with a scanning electron microscope (SEM) (Hitachi S-450). Electrolyte ions incorporated into the films were analysed using an energy-dispersive electron probe X-ray microanalyser (EPMA) (Horiba EMAX- 1500E) connected to the SEM.X-Ray diffraction analyses of the PPy films were carried out using a Shimadzu XD-3A X-ray diffractometer, and visible absorption spectra were measured using a Shimadzu MPS-5000 u.v.-visible spectrophotometer. Cyclic voltammetry was carried out in an aqueous solution containing 1 mol dm-3 KCl under N, using a potentiostat (Hokuto Denko HA-104), a potential sweeper (Nikko Keisoku NSP-2A) and an X-Y recorder (Yokogawa Type 3086). A saturated calomel electrode (SCE) served as a reference electrode. Results Concentration of Dopant Anions and Conductivity of PPy Films The molar concentrations of sulphonate anions doped in the PPy films prepared at 0.005, 0.2 and 10 mA cm-2 were determined by the elemental analyses.The results are listed in table 1 as the molar ratio of sulphonate groups to pyrrole rings. Except for the case of P-4, the molar concentrations of sulphonate groups of the polymer films are almost the same as those reported for univalent inorganic dopants such as ClO, and C1-,6 and are independent of the number of sulphonate groups of a dopant molecule and the preparation current density. These results show that the degree of ionization of PPy films prepared by the electropolymerization is the same even if multivalent anions are used as a dopant material. In the case of the PPy films doped with P-4, the molar concentration was lower than that of the other polymers. This is attributable to the molecular size of P-4 which may be too large to be doped completely in the polymer films.Plots of the conductivity of all kinds of PPy films prepared in the present study against the current density used in the film preparation are shown in fig. 1 , and those of the conductivity against the number of sulphonate groups of naphthalene sulphonate derivatives are shown in fig. 2. Note that even excluding the PPy films doped with P-4 the conductivity of PPy films is different by more than an order of magnitude between A-1 and N-3 dopants when the films were prepared at 0.005 mA ern-,. Rearrangement of fig. 1 for the cases of naphthalene sulphonate dopants gave fig. 2. ( a ) and (b) in fig. 2, which depict film preparation at 0.005 and 0.02 mA crn-', respectively, show that an increase in the number of sulphonate groups of a dopant molecule reduces the conductivity of the polymer films when the film is prepared at relatively low current densities.However, the difference in the conductivity due to theS. Kuwabata, K. Okamoto and H. Yoneyama Table 1. The molar concentration of sulphonate anions doped in PPy films prepared at various current densities concen tra tionb dopant 0.005" 0.2' 10' A- 1 0.29 0.34 0.32 A-2 0.32 0.33 0.33 N- 1 0.33 0.32 0.3 1 N-2 0.32 0.3 1 0.33 N-3 0.33 0.33 0.34 P-4 0.24 0.26 0.26 a See text for definitions. Molar ratio of suiphonate anions to pyrrole rings. Current density for film preparation (mA cm-2). I 1 I I 2319 0.001 0.01 0.1 1 10 current for PPy preparation/mA cm-* Fig. 1. The relationship between the current density used for polymer film preparation and the Conductivity of PPy films doped with A-1 (O), A-2 (a), N-1 (A), N-2 (A), N-3 (A) and p-4 (0).number of sulphonate groups of a dopant becomes small with an increase in the current density chosen in the preparation of PPy films, as shown by curves (c) and (4. Although not shown here, the same was true in the case of anthraquinone sulphonate derivatives as dopants. Furthermore, it is noted from fig. 1 that the conductivity reaches its maximum at an optimum current density of the film preparation (0.2 mA ern-') for all2320 Conductivity of Polypyrrole Films k I I I 1 2 3 “-1) (N-2) “-3) no. of sulphonate groups Fig. 2. The relationship between the conductivity of PPy films prepared at 0.005 (a), 0.02 (b), 0.2 (c) and 10 ( d ) mA cm-2 and the number of sulphonate groups of a dopant molecule.I I I I I 2 3 4 5 2elo Fig. 3. X-Ray diffraction patterns of PPy films doped with A-1 (a), N-1 (b), and N-3 (c). The polymer film preparation was made at 0.005 mA cmP2.S . Kuwabata, K. Okamoto and H . Yoneyama 232 1 \ \ i., 1 I l l l l l l l l l l 0 0.5 1 .o crystallinity Fig. 4. The relationship between the conductivity and the crystallinity of PPy films prepared at 0.005 (0) and 0.2 (0) mA cm-2. 420 440 460 48 0 500 wavelengthlnm Fig. 5. Visible absorption of PPy films doped with N-1 (a), N-2 (b) and N-3 (c). The polymer films were prepared at 0.005 mA crnp2.2322 Conductivity of Polypyrrole Films Table 2. The peak position of visible absorption bands for PPy films doped with N-1 , N-2 and N-3 preparation peak position/nm current /mA cm-2 N- 1 N-2 N-3 0.005 472 464 457 0.2 483 478 477 10 48 5 48 1 479 - - ~- -__ 1.0 0 - - 6 Q -3 -1.0 - -2.0 - -0.8 - 0.4 0 0.4 EIV us.SCE Fig. 6. Cyclic voltammograms of N-2-doped PPy films prepared at 0.005 (a), 0.2 (b), 10 (c) mA cm-2 in aqueous 1 mol dmP3 KCI. dE/dt = 50 mV s-'. the PPy films with a variety of dopants. The extremely low conductivities of the PPy films doped with P-4 as shown in fig. 1 may be caused partly by the large number of sulphonate groups of the dopant molecule and partly by the low anion concentration in the polymer films as described above. X-Ray Diffraction Analyses of PPy Films It has been reported that PPy films doped with n-alkylsulphates or n-alkylsulphonates show an intense X-ray diffraction peak below 20 = 5" l8 which is indicative of a layered structure of the PPy film doped with the organic anions.The PPy films doped with aromatic sulphonate derivatives used in the present study also showed similar X-ray diffraction peaks. Typical examples are shown in fig. 3. The intensity of the peak observed around 28 = 3" varied depending on the kind of dopants and the current density of the polymer film preparation. It is well known that the intensity of an X-ray diffraction peak is a measure of the crystallinity of The crystallinities of the PPy films prepared in various electrolyte solutions at a variety of current densities were evaluated by comparing the intensity of the diffraction peaks of each sample. The PPy films prepared at 0.005 mA cmP2 in the A-1 electrolyte solution gave the highest diffraction peak, and then the ratio of the peak height of each sample to that of this sample was adopted as a relative measure of the crystallinity. The relationship of theS.Kuwabata, K. Okamoto and H . Yoneyarna Table 3. Cathodic peak potentials of cyclic voltammograms of PPy films doped with N-1, N-2 and N-3 preparation peak position/V us. SCE current /mA cmP2 N- 1 N-2 N-3 0.005 -0.80 -0.82 -0.80 0.2 -0.77 -0.67 -0.76 10 -0.67 -0.56 -0.60 S -\ 1 I K I I 2.0 3.0 4.0 energy/keV 2323 Fig. 7. EPMA spectra of oxidized (-) and reduced (.....-) PPy films doped with N-1. crystallinity and the conductivity of PPy films prepared at 0.005 and 0.2 mA cm-2 are shown in fig.4. As for the polymer films prepared at 0.005 mA cm-2, the film having a high crystallinity shows a high conductivity, as already suggested by other investigator^.^^^^^ In contrast, the polymer films prepared at 0.2 mA cm-2 have lower crystallinities than those prepared at 0.005 mA cm-2, but they possess higher conductivities. These results suggest that the conductivity of PPy films is not governed by the crystallinity of the polymer films alone. Furthermore, the polymer films prepared at 0.2 mA cm-2 have conductivities of almost the same order of magnitude regardless of the kinds of dopants except for the case of P-4. It is suggested that the effects of dopant anions on the film conductivity which are observed in the film prepared at 0.005 mA cmP2 (see fig.1) are obscured in those prepared at 0.2 mA cm-2 by disordering of the polymer structure. Visible Absorption Spectra of PPy Films The visible absorption spectra of PPy films could be measured without dopant interference for the films doped with naphthalene sulphonate derivatives, since these dopants have no absorption bands in the visible region. The as-grown films exhibit absorption bands due to the transition between bonding and antibonding bipolaron bands.lV2 According to the results shown in fig. 5 for the polymers prepared at 0.005 mA cm-2, a decrease in the number of sulphonate groups of the dopant causes a bathochromic shift and broadening of the absorption band. This result may reflect a situation where the overlaps between the bipolaron states'? are weakened by increasing2324 Conductivity of Polypyrrole Films the number of sulphonate groups of dopants, thereby narrowing their resonant area in the polymer chains.Similar bathochromic shifts of the absorption maximum were observed with increasing the current density of the film preparation, as shown in table 2. As far as the results on the polymer films prepared at 0.005 and 0.2 mA cm-2 are concerned, the bathochromic shift in the absorption maximum is accompanied by an increase in the conductivity. The Electrochemical Behaviour of PPy Films Cyclic voltammograms of N-2-doped PPy films prepared at 0.005, 0.2 and 10 mA cm-2 are shown in fig. 6. The redox behaviour of these polymers is different from that of PPy films doped with inorganic anions, which show well defined anodic and cathodic waves in KCl solution.6 Concerning the N-2-doped films, a cathodic current peak shifts anodically with increasing current density used for the film preparation. Similar shifts of the cathodic wave are observed for N-1 and N-3-doped films, as shown in table 3.It has been reported21 that when PPy films doped with large anions such as polyvinylsulphates were reduced, the doped anions were not eliminated but electrolyte cations were incorporated into the films to maintain their electrical neutrality. Similar situations were also observed with the present polymer films doped with the aromatic sulphonate derivatives; these dopants were not eliminated in the polymer films by the reduction, as judged from the results by elemental analyses of the reduced polymers, and furthermore potassium was detected by EPMA only in the reduced polymer films as shown in fig. 7.The cathodic waves in fig. 6 are then judged to be due to the insertion of potassium cations into the films. In this respect, the finding shown in fig. 6 that the cathodic current peak of the voltammograms shifts anodically with increasing current density .of the film preparation may be said to reflect that the polymer films prepared at higher current densities must be so much porous as to allow easier penetration of K+ into the film on its reduction. Discussion PPy films prepared at low current densities such as 0.005mAcm-2 show high crystallinities, reflecting a general rule observed, for example, in metal platings that a lower deposition rate of a crystalline materials results in higher crystallinities.However, in the PPy films prepared in the present study the conductivity and the crystallinity of the films are affected greatly by the number of sulphonate groups of a dopant as fig. 4 shows. The decrease in the conductivity with increasing number of sulphonate groups of a dopant may be explained in terms of the degree of charge localization in the doped PPy films. The larger the number of negative charges of a dopant, the stronger the attraction of positive charges in the polymer chain. The decrease in the crystallinity with increasing number of sulphonate groups of a dopant may also be explained by the interaction between the concentrated negative charges of a dopant with the positive charges in the polymer chains, as schematically illustrated in fig.8. The X-ray diffraction analyses show that an increase in the current density for film preparation causes a decrease in the crystallinity of the resulting films, whereas the results shown in table 2 on the visible absorption spectra suggest that resonance of bipolarons in the polymer chains becomes spread as the crystallinity is lost. These phenomena can be explained by assuming that with increasing current density for the film preparation, cross-linking of the polymer chains becomes more enhanced, thereby the interaction between the positive charge carriers and the dopant anions decreases as illustrated in fig. 9. The increase in the density of cross-linking makes the conductivity of polymer films increase, and the decrease in the interaction between the positive charge carriers and the dopant anions should weaken the charge localization in the doped PPyS.Kuwabata, K. Okamoto and H. Yoneyama 2325 (a) 0 0 0 + + + + + + 0 8 0 8 0 0 + + + + + + 8 8 0 0 0 0 + + + + + + 8 0 0 Fig. 8. Hypothetical illustration of polymer structures of PPy films doped with monovalent (a) and multivalent (b) anions. Fig. 9. As for fig. 8, but at a higher current density. chains, and then the film conductivity is not affected by the number of sulphonate groups of dopants. Although the X-ray diffraction patterns and the visible absorption spectra are not remarkably different for the polymer films prepared at current densities above 0.2 mA cm-2, the conductivities decrease with increasing current density (fig.1). The principal cause for bringing about such a phenomenon may be related to the increase in the porosity of the polymer films with increasing current density for film preparation, as already described based on the voltammetric behaviour of PPy films prepared at different current densities. The degree of contact of the cross-linked polymer chains is low for a polymer film of high porosity and hence the conductivity is low for such polymer films.2326 Conductivity of Polypyrrole Films This research was supported by the Asahi Glass Foundation For Industrial Technology. References 1 J. L. Bredas, J. C. Scott, K. Yakushi and G. B. Street, Phys. Rev. B, 1984, 30, 1023. 2 J. L. Bredas and G. B. Street, Acc. Chem. Res., 1985, 18, 3099.3 J. C. Scott, P. Pflunger, M. T. Krounbi and G. B. Street, Phys. Rev, B, 1983, 28, 2140. 4 A. F. Diaz and J. I. Castillo, J . Chem. SOC., Chem. Commun., 1980, 397. 5 G. B. Street, T. C. Clarke, M. Krounbi, K. K. Kanazawa, V. Lee, P. Pflunger, J. C. Scott and G. 6 S. Kuwabata, H. Yoneyama and H. Tamura, Bull. Chem. Soc. Jpn, 1984, 57, 2247. 7 M. Satoh, K. Kaneto and K. Yoshino, Synth. Met., 1986, 14, 2899. 8 H. S. White, G. P. Kittlesen and M. S. Wrighton, J. Am. Chem. SOC., 1984, 106, 5375. 9 G. P. Kittlesen, H. S. White and M. S. Wrighton, J . Am. Chem. SOC., 1984, 106, 7389. Weiser, Mol. Cryst. Liq. Cryst., 1982, 83, 253. 10 M. Salmon, A. F. Diaz, A. J. Logan, M. Krounbi and J. Bargon, Mol. Cryst. Liq. Cryst., 1982, 83, 11 J. P. Travers, P. Audebert and G. Bidan, Mol. Cryst. Liq. Cryst., 1985, 118, 149. 12 T. Skotheim, M. V. Rosenthal and C. A. Linkous, J . Chem. SOC., Chem. Commun., 1985, 612. 13 M. Ogasawara, K. Funahashi, T. Demura, T. Hagiwara and K. Iwata, Synth. Met., 1986, 14, 61. 14 R. A. Bull, F. R. Fan and A. J. Bard, J . Electrochem. Soc., 1983, 130, 1636. 15 F. Bedioui, C. Bongars and J. Devynck, J . Electroanal. Chem., 1986, 207, 87. 16 M. V. Rosenthal, T. A. Skotheim and C. A. Linkous, Synth. Met., 1986, 15, 219. 17 S. Kuwabata, K. Okamoto, 0. Ikeda and H. Yoneyama, Synth. Met., 1987, 18, 101. 18 W. Werenet, M. Monkenbusch and G. Wegner, Mukromol. Chem., Rapid. Commun., 1984, 5, 157. 19 L. E. Alexander, X-Ray Diffraction Methods in Polymer Science (Wiley-Interscience, New York, 1969), 20 L. F. Warren and D. P. Anderson, J . Electrochem. SOC., 1987, 134, 101. 21 T. Shimidzu, A. Ohtani, T. Iyoda and K. Honda, J. Chem. Soc., Chem. Commun., 1986, 1415. 265. p. 137. Paper 7/1426; Received 3rd August, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402317

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I , 1988, 84(7), 2327-2333 Infrared Spectra of CO adsorbed on Prismatic Faces of a-Fe,O, A. Zecchina* and D. Scarano Istituto di Chimica-Fisica Universita di Torino, Cso M.D 'Azeglio 48, 10125 Torino, Italy A. Reller Anorg.-Chem. Institut Universitat Zurich, Winterthurerstrasse 190, 8057 Zurich, Switzerland Crystallites of cc-Fe,O, prepared by thermal decomposition of a-FeO(0H) have regular needle-like shape with statistically predominant prismatic faces (as shown by HRTEM and SAED patterns). The i.r. spectrum of CO adsorbed at 77 K on these microcrystals is characterized by a single peak at 2165.5 cm-' which is assigned to the stretching mode of CO a-bonded on Fe3+ ions, with negligible contribution of metal-<: back-donation. The frequency of the band shifts with the coverage : this is due to the building up of lateral interactions (dynamic and static) between the oscillators adsorbed on the prismatic faces.In previous papers concerning the adsorption of CO at 77 K on alkali-metal halide~,l-~ MgOl and a-Cr,03576 it is reported that CO is adsorbed onto the cations and that the CO stretching frequency undergoes a continuous shift with increased coverage of the cation. This shift arises because of the build-up of adsorbate-adsorbate interactions of the dynamic (dipole4ipole) and static (electrostatic and chemical) type occurring in ordered bi-dimensional adlayers of CO oscillators with their axes parallel. The dynamic and static effects are largely influenced by the nature of the surface CO bonds: in particular, on the transition-metal oxides examined the dynamic and static effects are large, while on systems without d electrons (where ~t back-donation effects are absent), the lateral interactions are small.With the aid of HRTEM observations, a detailed knowledge of the surface morphology can be achieved. In particular it is possible to determine which crystal faces are predominant and which are the preferential growth directions of the microcrystals. On this basis, i.r. investigation of the spectrum of the absorbed CO and of the frequency shifts us. coverage (0) combined with the detailed knowledge of the particle morphology obtained by HRTEM, gives information regarding (i) the chemical nature of the molecule-surface interaction, (ii) the nature of the through-substrate chemical forces, (iii) the spatial distribution and the local coordination of the adsorbing centres, (iv) the orientation of the adsorbed oscillators, and (v) the faces responsible for CO adsorption.In the light of these results we have investigated a family of isomorphic crystal systems such as a-A1,03,7 c ~ - c r , O ~ ~ and a-Fe,O,, with a corundum-type structure, in order to study the adsorption of CO on solids having the same crystallographic structure but different cationic electron configurations. In this paper the results obtained for cc-Fe203 are illustrated in detail. Experimental The a-Fe203 samples were prepared by thermal decomposition at 473 K in vacuo from an [x-FeO(OH) (goethite) pellet. The sample was then sequentially dehydroxylated and 23272328 I.R.Spectra of Fe,O,-adsorbed CO sintered by heating it in vacuo (15 min) and then in an 0, atmosphere (15 min) at 573, 673,773, 873 and 973 K. After the last thermal treatment in 0, at 973 K the sample was then evacuated at the same temperature under vacuum ( Torr). The final red-brown product was highly transparent in the i.r., as expected for a nearly stoichiometric oxide. After the final outgassing procedure, which cleans the surface of adsorbed water (with total elimination of the i.r. signals of surface OH) and other impurities, the sample was transferred under high vacuum into the cell compartment through which the i.r. spectra were measured at liquid N, temperatures. Owing to the heating effect of the i.r. beam, the sample temperature during the spectroscopic measurements was 87-97 K.The spectra were recorded on a Perkin-Elmer 580B spectrometer (resolution 4 5 cm-l), equipped with a data station. The frequencies are given to an accuracy of f0.5 cm-l. For electron microscopy measurements, the samples, suspended in isopropyl alcohol, were treated with ultrasonic waves and then deposited onto a copper grid coated with a perforated carbon film. The electron micrographs were obtained with a Jeol 200 CX HRTEM. For the SEM observations (Philips 515 microscope) the a-Fe,O, powder was deposited onto a metal support, then plated with gold film. Results and Discussion The Morphology of the a-Fe,O, Microcrystals The shape of the starting a-FeO(0H) (goethite) crystallites [plate 1 (a)] is needle-like.Both isolated and clustered needles are elongated along a preferential direction. The same morphology is shown by the scanning electron micrographs [plate 1 (b)]. The shape and the average length (ca. 5000 A) of the a-Fe,O, crystallites are illustrated in plate 2(a) (TEM micrograph) and plate 2(b) (SEM micrograph). In both micrographs we again observe needle-like crystallites. Therefore, we can conclude that there is a remarkable analogy between the morphology of the starting goethite and that of the resulting a-Fe,O, crystals, although the latter are shorter. The well known topotactic characteristic of the a-FeO(0H) -, a-Fe,O, transition is thus confirmed.* It is concluded that the thermal decomposition of goethite yields a-Fe,O, crystallites with an elongated crystalline habit and with a high predominance of prismatic faces.In plates 2(c) and 2(d), microcrystals at higher magnification are shown. It is evident that the morphology of the single crystal appears significantly well defined. One of the edges is regular’and geometric defects like steps and/or corners are few or absent. In these plates, inside the microcrystal image, we observe areas with different thicknesses, contrasts and quite regular contours, (nearly hexagonal or octagonal). These features are difficult to account for. However, the hypothesis that they are associated with the different thicknesses of the microcrystals, as result of void formation during the dehydration processg~ lo is feasible. Moir6 fringes are also observed from areas where different crystals overlap.Selected area diffraction (SAD) of the single crystals have been carried out. Quite often the diffraction pattern shows a rectangular symmetry (typical of hexagonal crystal system), resulting from rotation around the twofold axes, i.e. those which cross the prism edges or that are perpendicular to the prismatic faces. From this it can be concluded that the diffraction patterns have symmetry corresponding to prismatic faces. A statistical investigation of different crystalline zones, located in single crystals, shows that the prismatic (1 10) faces are statistically predominant, even if other prismatic faces are also present. From careful study of plate 2(a) we see that many microcrystals terminate with irregular shapes. In reality we can say that the termination of theJ .Chem. SOC., Faraday Trans. I , Vol. 84, part 7 Plate 1 Plate l.(a) TEM micrograph of FeOOH crystallites ( x 91875); (b) SEM micrograph of FeOOH crystallites. A. Zecchina, D. Scarano and A. Reller (Facing p . 2328)J . Chem. SOC., Faraday Trans. 1, Vol. 84, part 7 Plate 2 Plate 2.(a) TEM micrograph of a-Fe,O, crystallites ( x 61790); (h) SEM micrograph of a-Fe,O, crystallites. A. Zecchina, D. Scarano and A. RellerJ . Chem. SOC., Furaday Trans. I , Vol. 84, part 7 Plate 2 Plate 2.(c) HRTEM micrograph of a-Fe,O, ( x 630000); ( d ) HRTEM micrograph of a-FE,O, ( x 1900000; magnification of the selected area on the previous micrograph). A. Zecchina, D. Scarano and A. RellerA . Zecchina, D . Scarano and A . Reller 2329 prismatic microcrystals involves crystallographic planes, which .intersect the c principal axis and which are characterized by Miller indices (hkl), where 1 # 0.The extension of these planes is, however, much smaller than that of the prismatic planes and should not contribute very much to the adsorptive properties of a-Fe,O, from goethite. High-resolution Transmission Electron Microscopy In some of the high-resolution images we observe the presence of clear fringes, in correspondence with finer and hence more transparent areas. In plate 2(d) fringes forming 57 and 37" angles are actually observed. In other plates (not reported for sake of brevity) fringes parallel to the straight edges are also observed. Fringe formation can be briefly explained : the planes parallel to the electron beam, are in a suitable orientation for interference between the maxima of the zeroth- and first- order diffracting beams (i.e.between the direct and the first-order diffracted beams). The fringes, therefore, reproduce the crystal planes intersecting the prismatic face. The distance between the fringes is obtained from the simple relation 1 = Md where I is the distance between the fringes, M is thg magnification and d is the interplanar distance. Since the fringes are 2.7 and 3.68 A apart, we can therefore say that the planes are indexed as (104) and (012), respectively. As these diffraction fringes appear only in perfect crystals, it can be also concluded that the observed microcrystals have a highly ordered structure. The I.R. Spectrum of CO adsorbed on a-Fe,O, Microcrystals As shown in fig.1 the adsorption of CO at 77 K on an a-Fe,O, microcrystalline sample, which was activated under high vacuum at 973 K, gives a strong i.r. peak at 2165.5 cm-' (at maximum coverage). The stretching frequency of CO adsorbed on the a-Fe,O, surface at 77 K is very similar to that of CO adsorption on alkali-metal on a-Al,O,' and on prismatic faces of a-Cr203.6 In agreement with other literature data,ll we explain the high frequency value of the CO stretching mode in terms of interaction with the trivalent coordinately unsaturated Fe3+ metal cations. According to the high-resolution electron microscopy observations these surface ions are mainly situated on the (1 10) prismatic planes. A model of a neutral prismatic face, obtained by cutting the crystal at the level of the 0,- ions plane (where Fe3+ ions have only one coordination vacancy) has been discussed at length in a pjevious paper5 and is illustrated in fig.2. On this face the Fe3+-Fe3+ distance is 5.43 A. The frequency shift with respect to the gas phase (Av z +20 cm-l) is lower than that observed on hexagonal faces of a-Cr,O,. This is in agreement with the highly unsaturated state of the Cr3+ ions located on basal (0001) planes.6 Besides the band at 2165.5 cm-', a shoulder at 2143 cm-l, assigned to physically adsorbed CO, is also clearly observed. The bandshape shown in fig. 1 is slightly asymmetric on the low-frequency side. We think that this asymmetry is due to the presence of faces other than the predominant (hkO) ones.We have also observed several crystallographic projections corresponding to facelets which intersect the c axis, i.e. faces with indices (hkl) such as (441) or (124) etc. These faces are those responsible for the irregular termination of the needle-like microcrystals. As these are not very abundant; the CO adsorbed on them do not make an important contribution to the overall i.r. spectrum. In fig. 1 the spectra corresponding to different coverages at 77 K are also reported (in the inset the 'optical' isotherm is also shown). On passing from coverage 6' z 0 to Om,, we observe a weak downward shift in the frequency; this shift is due to the building up of' lateral interactions between the oscillators adsorbed on the prismatic faces and it is2330 I.R.Spectra of Fe,O,-adsorbed CO 21 50 2100 2050 vlcm-’ Fig. l.I.r. spectra (taken at different CO pressure) of T O + 13C0 mixtures on a-Fe,O,: (-), 99: 1 mixture at (- - -), 8 : 92 mixture at 0 = O,,,,. The inset shows the ‘optical ’ isotherm. Fig. 2. Model of a parallel c-axis prismatic face the result of both dynamic and static effects.12-15 These dynamic effects are essentially of two types : (i) parallel vibrating molecules interact ‘ through space ’ via a dipoleaipole mechanism ; (ii) vibrating molecules interact ‘ through substrate ’ via a vibrational coupling mechanism involving common bonding electrons (which allow dynamic ‘communication ’ between adsorbed molecules). The effect (ii) is negligible owing to the small adsorption enthalpy for this system.As far as the static effects are concerned, weA . Zecchina, D. Scarano and A . Reller 233 1 have to consider both through-space and through-substrate 'chemical effect ' inter- actions. The electrostatic (solvent or matrix) through-space contribution should be small or negligible; in fact, even at Om,, the CO molecules are not densely packed [the CO-CO distance is larger than the miniomurn van der Waals distance between CO molecules with their axes parallel ( i e . 3.3 A)]; the electrostatic shift should be therefore much smaller in absolute magnitude than that observed on passing from CO gas to CO solid (4 ~ m - ' ) . ~ From these considerations it is possible to conclude that in the a-Fe, 0,-CO system the frequency shifts are mainly due to dipole-dipole and chemical effects.It is possible to distinguish the two effects by means of the limiting dilution method using l2C0 : 13C0 (99 : 1) and l2CO : 13C0 (8 : 92) isotopic mixtures. The quantitative estimation of the dynamic effect at Om,, can be made by comparing the frequency of the l2C0 in the (99: 1) and (8: 92) mixtures. The dynamic contribution is barely observable and is estimated to be AVdyn z 1.5 cm-' k0.5. By difference from the total shift we obtain a Av,,,, "N - 14 cm-'. Such a negative (mainly chemical) shift can be explained if the through-substrate effect is of the inductive type. In fact, as the CO molecule acts as a donor species, via a very weak o bond with the Fe3+ ion and as the donated charge is dissipated through the substrate via an inductive mechanism, it turns out that the accepting ability per Fe3+ ions decreases with increasing 8, (with a consequent downward frequency shift).In the absence of vibrational through-substrate coupling, the weak dynamic shift is solely due to dipole-dipole coupling. The weakness of this effect is associated with the low values of vibrational polarizability in the adsorbed state. By replacing the w and w, frequencies in the monolayer and of the dynamically decoupled oscillator (singleton), respectively, in the modified Hammaker equation,14 the value of the vibrational polarizability a, of the adsorbed molecule can be obtained: (w/w,)' = 1 +a, t / ( l +ae t ) where w = 2165.5 cm-' and w, = 2164 cm-l. For CO adsorbed on prismatic faces the resulting value is a, = 0.027 A3 k0.009.It is most remarkable that the a, value is an order of magnitude lower than that of CO adsorbed on a-Cr,O, (0001)6 or on Ni0.16 By analogy with the results of the CO-MgO,' CO-Al,03,7 CO-NaCl, CO-KCl29 systems, we can interpret the low vibrational polarizability of CO on Fe3+ ion in terms of a weak adsorbent-adsorbate interaction consisting primarily of a weak a-donation from the 5 0 orbital of CO to the empty orbitals of Fe3+ ion, leading to subsequent polarization of the CO. The lower polarizing power of Fe3+ on (1 10) faces with respect to Cr3+ on (000 1) faces is likely to be associated with (i) the higher coordination state of Fe3+ on (1 TO) faces and (ii) the absence of crystal field stabilization energy (Fe3+ is a d5 system). These considerations allows us to place a-Fe,O, along with the systems whose cations are weak o-acceptors like oxides of non-transition metals'? l7 or the alkali-metal halides.,? It is most noticeable that the a, value fits very well the Seanor and Amberg curve (fig.3) in the region of the minimum, thus confirming that the adsorbed CO molecule is not heavily perturbed and that 71 back-bonding is substantially absent. We can also mention that the experimental frequency of CO adsorbed on Fe3+ ions is in agreement with the value calculated on the basis of the purely electrostatic Rebenstorf model," which of course neglects ;n back-donation effects. Conclusions The microcrystals of a-Fe,O,, prepared by thermal decomposition under high vacuum from a-FeOOH, are regular needles, in which (1 10) prismatic faces are predominant (as2332 a" I I.R.Spectra of Fe,O,-adsorbed CO 0 I I I I I .o P I I I I 4 I I I I I I I 2 m - 21 50 v1crn-l Fig. 3. Dependence of a, upon V : 4, our data; 0, data from ref. (18); A, data from ref. (19); m, W g ) ; 9, a++,. revealed by high-resolution transmission electron microscopy observations and by selected-area diffraction patterns). The i.r. spectroscopic behaviour of CO adsorbed on the prismatic faces shows that (i) the CO is bonded to the Fe3+ ions via a weak a-bond without substantial contribution of n back-donation ; (ii) the dynamic dipole4ipole interaction and the vibrational polarizability of adsorbed CO are both very low ; (iii) the through-substrate (chemical) interaction between the CO molecules adsorbed on (1 10) faces is responsible for the continuous downward shift of the stretching frequency, as d -+ em,,.References 1 E. Escalona Platero, D. Scarano, G. Spoto and A. Zecchina, Faraday Discuss. Chem. Soc., 1985, 80, 2 A. Zecchina, D. Scarano and E. Garrone, Surf. Sci., 1985, 160, 492. 3 D. Scarano and A. Zecchina, J. Chem. SOC. Faraday Trans. I , 1986, 82, 361 1. 4 G. E. Ewing and J. C. Pimentel, J. Chem. Phys., 1961, 35, 325. 5 D. Scarano and A. Zecchina, Spectrochim. Acta., Part A , 1967, 43, in press. 6 D. Scarano, A. Zecchina and A. Reller, Surf. Sci., submitted. 7 D. Scarano and A. Zecchina, unpublished results. 8 J. R. Giinter and H. R. Ostwald, Bull. Inst. Chem. Res. Kyoto Univ., 1975, 53, 249. 9 F. Watari, J. Van Landuyt, P. Delavignette and S. Amelinckx, J. Solid State Chem., 1979, 29, 137. 183. I0 F. Watari, P. Delavignette, J. Van Landuyt and S. Amelinckx, J. Solid State Chem., 1983, 48, 49. 11 R. Larsson, R. Lykvist and Rebenstorf, 2. Phys. Chem. (Leipzig), 1982, 263, 1089. 12 R. F. Willis, A. A. Lucas and G. D. Mahan, Chem. Phys. Solid Surf. Heterogeneous Catal., 1983, 2, 13 P. Hollins and J. Pritchard, Vibrational Spectroscopy of Adsorbates, ed. R. F. Willis (Springer-Verlag, 14 R. M. Hammaker, S. A. Francis and P. Eischens, Spectrochim. Acta. 1965, 21, 1295. 59. Berlin 1980), p. 125.A . Zecchina, D. Scarano and A . Reller 2333 15 B. M. Persson and R. Ryberg, Phys. Rev. B, 1981, 24, 6954. 16 E. Escalona Platero, S. Coluccia and A. Zecchina, SurJ Sci., 1986, 171, 465. 17 E. A. Paukshits, R. I. Soltanov and E. Yurchenko, React. Kinet. Cataf. Lett., 1981, 16, 93. 18 D. A. Seanor and C. H. Amberg, J. Chem. Phys., 1965, 42, 2967. 19 L. A. Denisenko, A. A. Tsyganenko and V. N. Filimonov, React. Kinet. Cataf. Lett., 1984, 25, 23. Paper 711479; Received 10th August, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402327

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. 1, 1988, 84(7), 2335-2346 A Correlated X-Ray Photoelectron and Electron Spin Resonance Spectroscopic Study of Rhodium-exchanged X and Y Zeolites Daniella Goldfarb, Salvatore Contarini and Larry Kevan" Department of Chemistry, University of Houston, Houston, Texas 77004, U.S.A. X-Ray photoelectron (X.P.) and electron spin resonance (e.s.r.) spectro- scopies have been employed to study the effect of the location and the Si/A1 ratio on the distribution of the various Rh oxidation states in faujasite- type zeolites after various activation processes. The zeolites studied were RhNa-X, RhCa-X, RhNa-Y and RhCa-Y. The samples were activated in flowing oxygen to 200,400 and 500 "C, activated in vacuum to 500 "C and reduced with H, at 500 "C after vacuum activation.Strong e.s.r. signals were observed only in the samples activated to 400-500 "C in flowing oxygen, indicating the presence of Rh". The amount of Rh' increased with the activation temperature, whereas the amount of Rh"' was substantial in the samples activated in flowing oxygen. While no significant effects were attributed to the Si/Al ratio the co-cation had a considerable effect on the distribution of oxidation states, mainly in samples activated in flowing oxygen. The relative amounts of Rh,,, in RhCa-X and RhCa-Y were larger than in RhNa-X and RhNa-Y; accordingly, the relative amounts of Rh' were larger in the latter zeolites. In recent studies of catalytically important metal species in zeolites we have reported on the formation of paramagnetic Rh" species in X and Y zeolites following activation, oxidation, reduction and adsorption of various adsorbates.1-4 The particular zeolites studied were RhNa-X,1,2 RhCa-X,3 RhNa-Y4 and RhCa-Y4 with Rh loadings of 1-3 wt O/O. Comparison of results obtained from these four zeolites provided information about the effect of the co-cation, i.e. Na+ vs. Ca+,, and the Si/Al ratio (ca. 1.2 in X zeolite and ca. 2.5 in Y zeolite) on the formation and the location of the various Rh" species within the zeolite framework. In general, we found that following activation in flowing oxygen up to 400-500 "C the concentration of Rh" in 1 wt O h RhNa-Y4* and RhNa-X1 was significant, whereas in 1 wt % RhCa-X3 and RhC-Y4 the RhI' concentration was negligible. However, when RhCa-X and RhCa-Y were further exposed to water, methanol, ammonia, oxygen or carbon monoxide a considerable amount of para- magnetic Rh" species was generated.Similar increases in the Rh" concentration upon adsorption of these molecules appeared in RhNa-Y as well. However, activated RhNa-X showed an increase in Rh" concentration only upon adsorption of water, methanol and ammonia; exposure to 0, and CO at room temperature did not alter the electron spin resonance (e.s.r.) spectrum.'. In all cases the increase of the Rh" concentration was attributed to the existence of diamagnetic Rh" dimers which dissociate into Rh" monomers upon interaction with adsorbates. Furthermore, when the Si/Al ratio is small, as in X-zeolite, the exchange of Na+ by Ca2+ has a substantial effect on the Rh" species f ~ r m e d , ~ whereas when the Si/Al ratio is larger, as in Y-zeolite, this exchange affects the Rh'I concentration only after activation.However, both X- and Y-zeolites respond similarly to adsorbate^.^ The observed differences in the generation of the Rh'I species could be accounted for by considering the preferred location of the Na+ cations versus the Ca2+ cations. In 23352336 E.S.R. Study of Rh-exchanged Zeolites activated RhNa-X it was found that site I in the hexagonal prism of the zeolite structure' is the preferred location for Rh".' In RhCa-X, however, the greater affinity of Ca2' in activated samples for site I6 prevents Rh" from stabilizing there, so it remains in the larger cages where it can either form Rh" dimers and/or be further oxidized or reduced to Rh"' or Rh'.In all the above studies the methods of investigation were e.s.r. and electron spin-echo modulation (ESEM) spectroscopies, in which only paramagnetic, i.e. Rh" and in some cases Rho, could be detected. Accordingly the effect of the location and the Si/A1 ratio on the distribution of all the Rh oxidation states, which in turn affects the Rh" species, was not studied. X-Ray photoelectron spectroscopy (X.P.S.) is one technique to study total oxidation state distributions in solids, and it has been used to study Rh-exchanged zeolites. Okamoto et al.7 found that after activation in vacuum Rhrrr, Rh' and Rho were present in RhNa-Y. The relative amounts of these oxidation states depended on the Rh-loading (50 and 18 wt O h ) and on the activation temperature. Kuznicki et aL8 reported that activation of fully exchanged Rh-Y to 300 "C in vacuum resulted in a mixture of Rh"' and Rho.Lars et al.' studied various Rh-containing supports using X.P.S. to understand their catalytic activities in the carbonylation of methanol to acetic acid. Among these supports were Na-X zeolites exchanged with RhCl, or [Rh(NH,),Cl]Cl, with a final Rh-loading of 0.6 wt %. They found that the Rh binding energies were sensitive to the activation process. Treatment with 0, or N, at 400 "C for 0.5-23 h resulted in mainly RhIII, whereas hydrogen treatment at 340 "C indicated the presence of Rh' or highly dispersed Rho. The hydrodesulphurization of thiophene using Rh-exchanged Na-X was also studied using X.p.s.l0 In this case both RhI'INa-X and Rh'INa-X zeolites were prepared by exchange with RhCl, and Rh, (CO,CH,),, respectively. In this work binding energies were assigned to Rh"' and Rh"; however, no definite identification of Rh' and Rho could be made, since in many cases the binding energies for these two oxidation states overlap.Another study of Lars et a1.l' on RhNa-X (0.38 and 4 wt O/O) exchanged with RhCl, and [Rh(NH,),Cl]Cl, reports the presence of mainly Rh' but also Rh"' and Rho after activation in vacuum at 400 "C. Reduction of RhNa-Y (2 wt O/O exchanged with [Rh(NH,),CI]Cl,} activated in flowing oxygen at 300-350 "C followed by H, at 300 "C generated highly dispersed Rho particles.12 In the present study we report X.P.S. results on RhNa-X, RhCa-X, RhNa-Y and RhCa-Y.The objectives were to extract the effect of the co-cation to Rh and the %/A1 ratio on the Rh oxidation states distribution after various activation processes and to correlate the results with our previous e.s.r. and ESEM studies. Correlation of X.p.s. and e.s.r. results have been found to be helpful by assisting in the assignment of the metal ion valence state in X and Y zeo1ites.13-17 It must be kept in mind, however, that e.s.r. is bulk-sensitive, while X.P.S. is surface-sensitive down to a few atomic layers. Experimental The X zeolite used was Linde 13X (Na-X) and the Y zeolite was Linde YZ-52(Na-Y). The zeolites were washed with 0.1 mol dm-, sodium acetate. Ca-X and Ca-Y were obtained by exchanging the corresponding Na zeolites with 0.1 mol dmP3 calcium chloride solution three times at 70 "C.Rh cations were exchanged into the zeolites using [Rh(NH,),Cl]Cl, from Strem Chemical Inc. The exchange was carried out at room temperature by dropwise addition of 25 cm3 of ca. 20 mmol dm-, solution of [Rh(NH,),Cl]Cl, in triply distilled water to a zeolite slurry of 0.5 g in 450 cm3 of water. The mixture was stirred for ca. 24 h, filtered and dried at room temperature. The Rh loadings were 4.47, 4.55, 3.97 and 3.90 wt% for RhNa-X, RhNa-Y, RhCa-X and RhCa-Y, respectively, as determined by commercial atomic absorption. We chose toD. Goldfarb, S. Contarini and L. Kevan 2337 work with Rh loadings higher than in our previous e.s.r. studies (1-2 wt "/o) owing to the lower sensitivity of the X.P.S. measurements. The samples (50 mg of the exchanged zeolites) were treated in three different ways.(a) Samples were heated under flowing oxygen at a rate of ca. 30 "C per 30 min to ca. 200, 400 and 500 "C. The samples were then left for 30 min at the activation temperature, evacuated for 16 h to a residual pressure of ca. Torrt and then cooled to room temperature. ( b ) Samples were heated similarly under vacuum to 45&500 "C. (c) Reduced samples were obtained by heating under vacuum to 500 "C followed by 16 h of evacuation, and exposure to 350 Torr H, for 2 h at 500 "C. Then the samples were outgassed at 500 "C for 2 h followed by cooling to room temperature. The samples were sealed at 77 K in 3 mm 0.d. by 2 mm i.d. Suprasil quartz e.s.r. tubes. E.s.r. measurements were performed at 77 K on a modified Varian E-4 spectrometer interfaced with a Tracor signal averager.After the e.s.r. experiments were performed the zeolites were taken out of the e.s.r. tubes in a nitrogen-controlled atmosphere and were pressed in a hydraulic die at ca. 200 kg to form a thin, smooth, round (1 cm diameter) pellet. Transfer of the pellets to the X.P.S. evacuation chamber was done in ca. 10 s. All samples were evacuated at ca. Torr overnight (1 5 &- 1 h) at room temperature. The base pressure during X.P.S. measurements was in the low lo-' Torr range. The photoelectron spectra were measured in a Perkin-Elmer PHI model 500 ESCA/ SAM spectrometer using Mg Ka X-rays at 1253.6 eV. Binding-energy values were referenced to the Si 2p line at 102.4 eV and to the A1 2p line at 74.2 eV.By doing so a value of 53 1.6 eV for the 0 1s line was found in all samples. The data-acquisition time was 90 min at room temperature, with no specific control of the sample temperature. The regions of interest were ca. 10 eV on each side of the peak maxima of Si 2s, A1 2p, Na 2s, Ca 2p, C12p, 0 Is, Rh 3d, C 1s and N 1s lines. Data smoothing, subtraction of inelastic scattering and deconvolution of the spectra were performed with the PHI software available with the spectrometer. Results E.S.R. Results The fresh Rh-exchanged zeolites do not show any e.s.r. signals, since Rh"' is diamagnetic, however following activation in flowing oxygen e.s.r. signals do appear. Fig. 1 shows the e.s.r. spectra obtained from RhNa-X after various activation processes. Activation to 200 "C in flowing oxygen yields a very weak signal with g,, = 2.054, g3,y = 2.090 and g,, = 1.973, which was attributed to Rh(NH3),C1+.5 Further activation to 400 "C produced strong e.s.r.signals corresponding to two Rh" species denoted as A (gxz = 2.542, gyy = 2.510 and g,, = 1.881) and C (gl = 2.582 and g,, = 2.031). These species have g values very close to those obtained previously in RhNa-X at low Rh loadings.' Upon activation to 500 "C in flowing oxygen only species C appeared. Activation to 450 "C in vacuum showed the same signal, shifted, with a considerable decrease in intensity. Reduced RhNa-X samples did not show any e.s.r. signal. The same trend was observed in the other three zeolites. The e.s.r. spectra obtained for RhCa-X are shown in fig.2. The strong signals at low field for the sample activated in vacuum [fig. 2(b)] correspond to a Rh.0, adduct3v4 which is generated upon introduction of traces of oxygen. This signal, however, disappears upon room temperature evacuation. Note that the relative concentrations of Rh" in RhCa-X and RhCa-Y are lower than in the corresponding RhNa-X and RhNa-Y. Furthermore, the signals in RhCa-X and RhCa-Y after activation to 400-500 "C in flowing 0, are stronger than expected by comparison with 1 wt YO RhCa-X and RhCa-Y.3.4 The Rh" species with their corresponding g values and the relative total Rh" i- 1 Torr = 101 325/760 Pa.2338 E.S.R. Study of Rh-exchanged Zeolites (a) 1 Fig. 1. E.s.r. spectra recorded at 77 K of RhNa-X (a) reduced with 350 Torr H, at 500 "C, (b) activated in vacuum at 450 "C, (c) activated in flowing oxygen at 500 "C, ( d ) activated in flowing oxygen at 400 "C and (e) activated in flowing oxygen at 200 "C.The gains are 5 x lo3, 1.25 x lo3, 1.25 x lo2, 1.5 x lo2 and 2.5 x lo3, respectively. concentration for some samples are listed in table 1. Note that the same RhII species appear in all four zeolites, but there are slight differences in the g values, especially when comparing X and Y zeolites. In the Y zeolites the g , values are shifted towards lower fields. The nature and location of these Rh species were discussed previo~sly.'-~ Species C was assigned to site I in a hexagonal prism of the zeolite structure, whereas species A was suggested to reside within the /3-cage in site 11' or 11.X.P.S. Results As previously reported by other investigators8-12 we also find that in these four zeolites the Rh binding energy (E,) depends on the activation process. Fig. 3 shows Rh 3d5,, band maxima as a function of sample treatment. The Rh binding energy obtained from a fresh RhNa-X sample evacuated in the X.P.S. chamber for 15 h was 309.6 eV. Evacuation for 36 h introduced a shift to 309.0 eV which is indicative of some decomposition of the Rh penta-amino complex. This is supported by a decrease in the relative intensities of the N 1s and C1 2p lines. The Rh 3d5,, binding energy for all four zeolites after 15 h of evacuation were in the 309.2 eV range. This range is lower than the value of 310.2 eV reported by Lars et aZ.,9 even though our A1 2p and Si 2p reference values are the same as theirs.The difference could be attributed to the prolonged evacuation, the heating of the sample during the measurement or to radiation-induced reduction, which may not be negligible in our case owing to the relatively long time ofD. Goldfarb, S . Contarini and L. Kevan 2.6 I 14 x0.05 species C ' k-J- I I s p e c i e s ~ 3 nnac 2339 Fig. 2. E.s.r. spectra recorded at 77 K of RhCa-X (a) reduced with 350 Torr H, at 500 "C, (b) activated in vacuum at 500 "C (note the decrease in signal gain at higher field), (c) activated in flowing oxygen at 500 "C, (d) activated in flowing oxygen at 400 "C and (e) activated in flowing oxygen at 220 "C. The gains are 5 x lo3, 1.25 x lo3, 4 x lo2, 5 x 10' and 2.0 x lo3, respectively.X-ray exposure (90min). Since in our study we are mostly concerned with the comparison of all four zeolites, once all measurements are performed under the same conditions the radiation-induced reduction will affect all samples similarly and should not be an obstacle. The four graphs in fig. 3 show the same general trend. The binding energies of the band maxima decreased upon activation in flowing 0, up to 400 "C, then at 500 "C a slight increase is observed. Vacuum activation to 500 "C yields even lower binding energies which are further lowered when reduction is carried out. Note that RhCa-Y behaves slightly different. The following trend for the Rh 3d5,2 binding energy values could be observed for each of the treatments investigated: NaX z NaY < Ca-X < Ca-Y.Fig. 4 shows the X.p. spectra of RhCa-X as a function of sample treatment. The bands of the reduced samples and the sample activated in vacuum were narrower than the others, especially those of samples activated in flowing oxygen at 400 and 500 "C. This indicates that the bands observed are actually a superposition of several peaks belonging to different Rh oxidation states. Curve fitting was then carried out on all RhCa-X spectra, and on RhNa-Y and RhCa-Y activated in flowing 0, to 400 "C. The peak parameters were restricted to those normally used for Rh compounds : a spin-orbit splitting of 4.6k0.15 eV,t a spin-orbit intensity ratio of 1.5 kO.15 and a peak width of 2.25 f 0.15 eV. The data are reported in tables 2 and 3. Such a procedure shows that all six spectra shown in fig.4 could be reasonably deconvoluted using a total of four t 1 eV = 1.602 18 x lo-" J.2340 E.S.R. Study of Rh-exchanged Zeolites Table 1. E.s.r. parameters and some relative concentrations in Rh-exchanged zeolites zeolite 500 "C/O, 400 "C/O, 200 "C/O, 500 "C/vacuum ~~ --____- ~~~ ~ RhNa-X 4.47 wt Yo C : g, = 2.582 g,, = 2.031 (1 3 Yo)" RhNa-Y 4.55 wt YO C : g, = 2.642 g,, = 2.03 (20 %)" RhCa-X 3.97 wt YO C : g, = 2.602 g,, = 2.046 (4 Yo)" C : g, = 2.652 g,! = 2.043 (5 %)" RhCa-Y 3.90 wt YO A: g,, = 2.510 g,, = 2.542 g,, = 1.88Ib C:g, = 2.582 gll = 2.031 g,, = 2.054 g,, = 2.090 g,, = 1.973 (10 Yo)" A:g, z 2.5 - g,, = 1.887b C:g, = 2.642 g,, = 2.03 g,, = 2.533 g,, = 1.898' A: g,, = 2.496 - A:g, = 2.56 - gll = '? g,, z 2.0 C:g, = 2.642 C: g , = 2.626 g,, = 2.037 C: g, = 2.652 g,, = 2.0 C: g, = 2.652 g,, z 2.0 C : g, = 2.652 g,, = 2.046 __- " The numbers in parentheses denote the amounts of Rh" relative to total Rh; the estimated error is _+ 50 YO based on e.s.r.' The g,, or g,, feature is split by ca. 30 G due to hyperfine interaction with lo2Rh ( I = i). 307 30gz 1 I I I 1 vac. O2 O2 O2 vac. H2 r.t. 200OC 400°C 50o0c50o"c 5 0 0 " ~ Fig. 3. Rh 3d,,, X-ray photoelectron band maximum as a function of sample treatment for (a) RhNa-X, (b) RhNa-Y, (c) RhCa-X and ( d ) RhCa-Y.D. Goldfarb, S. Contarini and L. Kevan 234 1 5/2 binding energy/eV Fig. 4. X-Ray photoelectron spectra of Rh 3d5,, and Rh 3d3/,. regions in RhCa-X: (a) fresh samples, (b) activated in flowing oxygen to 200 "C, (c) activated in flowing oxygen to 400 "C, (d) activated in flowing oxygen to 500 "C, (e) activated in vacuum to 500 "C and (f> activated in vacuum to 500 "C and reduced with 350 Torr H, at 500 "C for 2 h.Table 2. X.P.S. 3d5/, and 3d3,,. peak (A-D) binding energies and total (3d5/,+3d3/,) relative intensities in parentheses for different Rh species in variously treated (A-D) RhCa-X samples binding energy/eV sample A B C D fresh 310.3 315.0" (0.54) (0.32) (0.34) (0.37) (0.16) (0.09) 200 "C/O, 310.2 314.9 400 "C/O, 310.2 314.9 500 "C/O, 310.1 314.9 500 "C/vac 310.5 314.8 500 "C/H, __ 308.8 313.4 0.46) 308.6 313.4 (0.58) 308.6 313.4 (0.36) 308.6 313.5 (0.36) 308.8 313.5 (0.37) 308.8 313.6 (0.36) 307.6 312.2 - 307.6 312.2 - 307.7 312.1 - 307.6 312.2 - 307.7 312.4 306.8 311.4 (0.10) (0.30) (0.27) (0.47) (0.55) 77 FAR 12342 E.S.R.Study of Rh-exchanged Zeolites Table 3. X.P.S. 3d5/, and 3d3/, peak (A-D) binding energies in eV and total ( 3 4 , + 3d3,,) relative intensities in paranthesis of zeolite samples activated in flowing oxygen to 400 "C Rh species RhNa-Y RhCa-X RhCa-Y A 310.2 314.9 310.2 314.9 310.5 315.1 B 308.5 313.2 308.6 313.4 308.9 313.5 C 307.9 312.5 307.6 312.2 307.8 312.1 (0.22) (0.34) (0.37) (0.42) (0.36) (0.40) (0.36) (0.39) (0.23) binding energy/eV Fig. 5. Deconvolution of the Rh 3d5/, and 3d3,, X.P.S. transitions in (a) RhCa-X and (b) RhCa-Y zeolite samples activated in flowing oxygen at 400 "C. Rh 3d5/, and four Rh 3d5/, peaks. The average binding energies of the four (A-D) 3d,,, peaks are 310.25 f0.15, 308.7 kO.1, 307.6f0.1 and 306.8 kO.2 eV.Fig. 5 shows an example of deconvolution for the case of RhCa-X and RhCa-Y activated to 400 "C in flowing oxygen. Although the general trends in all four Rh-exchanged zeolites were similar, significant differences could be observed between the Ca2+- and Na+-exchanged zeolites. The peaks in the RhCa-Y and RhCa-X samples activated in flowing 0, are broader then in the corresponding RhNa-Y and RhNa-X zeolites. This was best demonstrated in samples activated at higher temperature as shown in fig. 6 and 7 for samples activated at 200 and 400 OC, respectively. The shoulder at higher binding energy for the Ca2+ zeolitesD. Goldfarb, S. Contarini and L. Kevan 2343 318 314 310 306 binding energy/eV Fig. 6. X-Ray photoelectron spectra of Rh 3d transitions in (a) Na-X, (b) Na-Y, (c) Ca-X and (d) Ca-Y zeolite samples activated at 200 "C in flowing oxygen.I 512 318 314 310 306 binding energy/eV Fig. 7. X-Ray photoelectron spectra of Rh 3d transitions in (a) Na-X, (b) Na-Y, (c) Ca-X and (d) Ca-Y zeolite samples activated at 400 "C in flowing oxygen. indicates a larger amount of Rh in a higher oxidation state. Deconvolution of the X.p. spectra after activation in flowing oxygen to 400 "C indeed shows a different distribution of the oxidation states. The RhNa-X spectrum was not deconvoluted due to its relatively poor signal-to-noise ratio and to a substantial overlap with the Na KLL Auger line. Since its spectrum is very similar to the RhNa-Y spectrum, the deconvolution is not critical. As seen in table 3, while there is no significant change in the B peaks (ca.308.6 eV) there are definite changes in the relative intensities of the A and C peaks. The intensity of C increases as follows: RhNaY z RhNaX > RhCa-X > RhCa-Y, whereas the intensity of A decreases similarly. 77-22344 E.S.R. Study of Rh-exchanged Zeolites Discussion Okamoto et aL7 studied X.p. spectra of Rh-Y zeolite with 50 and 18 wt YO Rh loading as a function of activation temperature with activation in vacuum. Deconvolution of the spectra showed three peaks at 310.2, 308.3 and 307.5 eV which were assigned to Rh''', Rh' and Rho, respectively. The relative amounts of the Rh oxidation states in the sample depended on the total amount of Rh. For 18 wt % Rh-Y activated at 450 "C they reported relative intensities of 46 : 34 : 20 for Rho, Rh' and RhI'I, respectively.This is rather similar to what we obtained for RhCa-X activated to 500 "C in vacuum for the peaks at 307.6, 308.8 and 310.5 eV, with relative intensities of 47:36: 16. Lars et al.ll found that activation of Rh(NH,),Cl-NaX in vacuum at 400 "C for 2 h gave a peak at 308.2 eV which was assigned to Rh'. However, they also stated that from the peak shape the presence of Rh"' and Rho could not be excluded. Their measurements were performed at - 100 "C with the stated objective to eliminate reduction due to heating by X-ray irradiation. The assignment of peak A 310.25 f0.15 eV to Rh"' is unambiguous and is in agreement with other literature values.8-10 However, the assignment of the other bands seems somewhat ambiguous. One possibility is to adopt the assignment of the values of Okamoto et al.,' which attributed no peak to Rh" based on the absence of e.s.r.signals in their samples. Accordingly band B (308.7 0.1 eV) would correspond to Rh', band C (307.6 0.1 eV) would correspond to isolated Rho clusters while band D (306.8 0.2 eV) would correspond to bulky crystalline Rh metal. These assignments are at variance with our e.s.r. results, since a substantial amount of Rh" is observed after activation of RhNa-X and RhNa-Y in flowing oxygen at 400-500 "C (see table 1). Thus, to account for the e.s.r. signals one would have to introduce a Rh" peak which overlaps with the Rh' peak. Another objection for the assignment of the 307.6 peak to Rho may be its appearance in samples activated in flowing oxygen.However, since all samples activated in flowing oxygen were evacuated at the activation temperature it is possible that further reduction took place during this evacuation. A different possible assignment of the X.P.S. peaks is: Rh" at 308.7f0.1 eV, Rh' at 307.6 and Rho at 306.8 0.2 eV. This assignment is in good agreement with values for RhT1 reported for Rh,(CO,CH,),-exchanged 13X zeo1ite.l' The Rho value is close to the value reported for Rh metal, 307.0-307.4 eV., However, this assignment is also at variance with the e.s.r. results. As indicated by table 2, all samples contained a substantial amount of Rh" if Rh" is assigned to 308.7 eV, but only the samples activated in flowing oxygen to 400-500 "C showed strong e.s.r.signals. As mentioned above, the samples were exposed to the X-rays for a relatively long time (90 min), which caused an unknown amount of radiation-induced reduction. This could account for the larger amounts of Rh" based on X.P.S. data relative to the amount based on e.s.r. results. Furthermore, if a significant amount of Rh'I was present as diamagnetic Rh'I dimers, Rh" species could be seen by X.P.S. but not by e.s.r.1T3*4 The second possibility of the Rh peaks assignment seems more plausible. For samples activated in flowing 0, the Ca2+-exchanged zeolites showed a larger amount of Rh"' by X.P.S. than the corresponding Na+-exchanged zeolite. Furthermore, the RhI" amount in RhCa-Y activated in flowing oxygen was larger than that in the corresponding RhCa-X sample.No significant differences were observed between RhNa-X and RhNa-Y. From this one can deduce that (a) either in the Ca2+-exchanged zeolites Rh"' species, which might have been reduced during the heating process, are readily reoxidized, or (b) that Rh"' is more readily reduced in the Na+-exchanged zeolites. Since all samples activated to 450-500 "C in vacuum showed similar X.P.S. results, with little Rh"' present it seems that the first possibility is more plausible. Similar studies performed in Ni-exchanged Na-X and Ca-Y zeolites17 showed that the co-cationD. Goldfarb, S. Contarini and L. Kevan 2345 to nickel affects the reducibility of Ni" to NiO. In X zeolites Ni2+ is reduced easier in Na-X than in Ca-X, but the opposite trend is observed in the analogous Y zeolite." The oxidation of Rh is easier in Ca2+exchanged X and Y zeolites than in the corresponding Na+-exchanged zeolites. This probably related to the Rh location within the zeolite structure, which is affected by the co-cation type (Na+ or Ca2+).If during the activation process of the Ca2+ zeolites most of the Rh cations reside in the larger a- and ,&cages, then reduction may be inhibited by the presence of the oxygen present during activation. And, if the Rh cations reside within the hexagonal prisms, which are not accessible to oxygen, then the oxygen will not be as effective in inhibiting reduction. This model is supported by our previous e.s.r. studies, which showed that while Rh" in Na-X was mostly located in the hexagonal prisms and did not interact with adsorbates other than water, in Ca-X Rh" is located in the a- and 8-cages and readily interacts with various adsorbate^.^ However, in the X.P.S.experiments the RhNa-Y results are very similar to the RhNa-X data, while in the e.s.r. experiments the RhNa-Y data showed many of the characteristics found in the RhCa-X and RhCa-Y * systems. While the Rh X:p.s. patterns are affected by the various activation processes, their low resolution prevents differentiation between Rh species within the same oxidation state, such as Rh" dimers and monomers. Furthermore, the assignment of the binding energies to specific Rh oxidation states remains ambiguous. However, the additional information from e.s.r. data has added new constraints on the interpretation of the X.P.S.data. Conclusions The distribution of Rh oxidation states in RhNa-X, RhCa-X, RhNa-Y and RhCa-Y depends on the activation process and the activation temperature. Activation in vacuum showed more Rh' than activation in flowing oxygen, while activation in flowing oxygen showed more Rh"'. In samples activated in flowing oxygen the Ca2+ zeolites contained a smaller amount of lower Rh oxidation states than the corresponding Na+ zeolites. E.s.r. spectra showed a substantial amount of Rh" in samples activated in flowing oxygen, especially in Na-Y and Na-X, and no Rh" in samples activated in vacuum. In contrast, the X.P.S. results did not show significant changes in the relative intensity of the 308.8 eV X.P.S. peak assigned to RhII between flowing oxygen and vacuum treatments. This research was supported by the U.S. National Science Foundation and the Texas Advanced Technology Research Program. References 1 D. Goldfarb and L. Kevan, J . Phys. Chem., 1986, 90, 264. 2 D. Goldfarb and L. Kevan, J . Phys. Chem., 1986, 90, 2137. 3 D. Goldfarb and L. Kevan, J . Phys. Chem., 1986, 90, 5787. 4 D. Goldfarb and L. Kevan, J . Am. Chem. SOC., 1987, 109, 2303. 5 C. Naccache, Y. Ben Taarit and M. Boudart, Am. Chem. SOC. Symp. Ser., 1977, 40, 155. 6 D. W. Breck, in Zeolite Molecular Sieves (Wiley, New York, 1974), pp. 12-16. 7 Y. Okamoto, N. Ishida, T. Imankka and S. Teranishi, J . Catal., 1979, 58, 82. 8 S. M. Kuzniki and E. M. Eyring, J . Catal., 1980, 65, 227. 9 S. Lars, T. Anderson and M. S. Scurrell, J . Catal., 1981, 71, 233. 10 K. E. Givens and J. G. Dillard, J . Catal.,, 1984, 86, 108. 1 1 S. Lars, T. Anderson and M. S. Scurrel, Zeolites, 1983, 4, 261. 12 R. D. Shannon, J. C. Verdine, C. Naccache and L. Lefevbre, J . Catal., 1984, 88, 431. 13 M. Narayana, J. Michalik, S. Contarini and L. Kevan, J . Phys. Chem., 1985, 89, 3895. 14 M. Narayana, S. Contarini and L. Kevan, J . Catal., 1985, 94, 370. 15 S. Contarini and L. Kevan, J . Phys. Chem., 1986, 90, 1630.2346 E.S.R. Study of Rh-exchanged Zeolites 16 Kh. M. Minachev, G. V. Antoshin, Yu. A. Yusifov and E. S. Shpiro, React. Kinet. Catal. Lett., 1976, 17 S . Contarini, J. Michalik, M. Narayana and L. Kevan, J. Phys. Chem., 1986, 90, 4586. 18 D. Oliver, L. Bonneviot and M. Che, in Magnetic Resonance in Colloids and Interface Science, ed. 4, 137. J. P. Fraissard and H. A. Resing (Reidel, Dordrecht, 1980), p. 483. Paper 711559; Received 25th August, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402335

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1988, 84(7), 2347-2356 Deuterium Nuclear Magnetic Resonance Studies of the Molecular Dynamics of Benzene in Zeolites Bod0 Zibrowius* and Jurgen Caro Zentralinstitut f u r physikalische Chemie der AdW der DDR, Rudower Chaussee 5, 1199 Berlin, German Democratic Republic Harry Pfeifer Sektion Physik der Karl- Marx- Universitat Leipzig, Linnkstrape 5, 7010 Leipzig, German Democratic Republic Deuterium n.m.r. spectroscopy has been used to investigate the micro- dynamics of benzene molecules sorbed on HZSM-5, NaX and NaY zeolites. As in the liquid and solid benzene phases, the sorbed benzene molecules perform fast reorientations about their hexad axes. Superimposed on this C, reorientation are jumps which the benzene molecules perform between the sorption sites.The mean residence time between two succeeding jumps depends significantly on the zeolite framework, the temperature and the loading. From this finding, diffusivities of sorbed benzene molecules can be estimated. Although 2H n.m.r. spectroscopy has been widely and successfully used in polymer science'g2 few papers have dealt with the dynamics of guest molecules in zeolite structure~.~-~ This may be due to the comparably poor signal-to-noise ratio that results from the low concentrations of molecules sorbed on zeolites. On the other hand, the lineshape of a 2H n.m.r. spectrum is mainly determined by the electric quadrupolar interaction. This allows a more direct identification of several kinds of molecular reorientation than an analysis of the corresponding proton magnetic resonance signals : certain molecular motions give rise to characteristic 2H n.m.r.lineshape patterns. In this paper 2H n.m.r. spectra of perdeuterated benzene sorbed on HZSM-5, NaX and NaY zeolites have been studied. In combination with the results of other methods, a model for the microdynamical state of the sorbed benzene molecules is developed. Experimental Zeolites and N.M.R. Sample Preparation The ZSM-5 zeolite was synthesized using TPA as a template.8 After calcination (870 K, 8 h) and direct acidification (350 K, 6 h, 0.1 mol dm-3 HCl) the silicon-to-aluminium ratio of the zeolite was determined independently by 27Al and 'H m.a.s.n.m.r., on the assumption that every bridging OH group is connected with one A1 atom in the f r a m e ~ o r k .~ The %/A1 ratio obtained was 50+5. As plate I shows, the crystals used exhibit a shape characteristic of ZSM-5 with a mean crystal size of 50 x 12 x 10 pm. The faujasites (NaX with %/A1 z 1.8 and NaY with %/A1 z 2.6) were of commercial origin (VEB Chemiekombinat, Bitterfeld) with crystal sizes of ca. 2 4 pm. The X- and Y-type zeolites were investigated in the Na form. For the n.m.r. measurements all zeolite samples (ca. 0.38 g) were prepared by a deep- bed treatment'' at an activation temperature of 670 K and a pressure below 0.05 Pa. The loading was performed by vacuum distillation of perdeuterated benzene and controlled 23472348 2H N.M.R. of Benzene in Zeolites gravimetrically . All sorbate concentrations are given in benzene molecules sorbed per 24 (Si + Al) atoms.In the case of HZSM-5 this loading represents the number of benzene molecules per channel intersection (equal to + of a unit cell), for faujasites it corresponds to the number of benzene per large cavity (i u.c.). The experimental error of the concentration determination does not exceed & 10 %. N.M.R. Measurements The deuterium n.m.r. experiments were performed on a home-made Fourier-transform n.m.r. spectrometer (UDRIS) at a frequency of 13.82 MHz. The solid-like spectra were obtained with the quadrupolar echo pulse sequence [n/2, ( - y ) - z - n/2, x] with z 3 60 ,us." The width of the 7r/2 pulse was 3.5 ,us. In order to avoid spectra distortions, the phase of the second pulse was shifted by n for half of the scans. The spectra are presented as half-spectra resulting from Fourier transformation of the accumulated non- quadrature echo decay starting at the maximum of the echo.The repetition time was between 0.4 and 2 s. Up to 16 x lo3 scans were accumulated. Lineshape of *H N.M.R. Signals The nuclear magnetic resonance frequency w of deuterium ( I = 1) in a magnetic field of intensity B, is given by12 W w = ~ , + ~ ( 3 c o s ~ 8 - 1 - ~ s i n ~ 8 c o s 2 $ ) . 2 (1) Where W, is the Larmor frequency, which is the product of the magnetogyric ratio y of the deuterium and the magnetic field intensity Bo. uQ denotes the quadrupole frequency where Q is the electric quadrupole moment of deuterium and eq = Kz represents the zz component of the electric field gradient tensor. 8 and $ describe the orientation of the magnetic field B, in the principal-axes system of the electric field gradient tensor (x,y,z) and oQ = $e2qQ/-ti (2) is the so-called asymmetry parameter (0 < q < 1): For deuterium bonded to carbon the asymmetry parameter is approximately zero, the z-axis is along the C-D bond and hence eqn (1) simplifies to The corresponding lineshape for a powder follows from an average of eqn (4) over 8.The result for the rigid case (0 does not depend on time) is shown in the first row of table 1. Additional interactions (e.g. the magnetic dipolar coupling) lead to a slight broadening of the spectra and hence to a smearing over of the sharp edges and singularities. If the molecules are not fixed in space, 8 becomes a function of time. As long as the characteristic time constant z, of such a motion (correlation time) is sufficiently large ZcWQ 9 1 ( 5 ) the lineshape does not change.For a more rapid motion, however, characteristic deviations occur, since depending on the type of motion at least a certain part of the quadrupolar interaction is averaged to zero. Some theoretical lineshapes forJ . Chem. SOC., Faraday Trans. 1, Vol. 84, part 7 Plate I 50 p m I Plate 1. SEM of the ZSM-5 crystals (Si/AI = 50k5) used in this paper. B. Zibrowius, J. Car0 and H. Pfeifer (Fucing p . 2348)B. Zibrowius, J. Caro and H. Pfeifer 2349 Table 1. Theoretical 2H n.m.r. lineshapes of deuterated benzene molecules undergoing different types of motion type of motion lineshape and quadrupolar splitting A /kHz rigid molecule I fast C, reorientation fast C, reorientation fast C, reorientation I 1 : 69 fast isotropic reorientation I ‘Fast’ corresponds to a correlation time which is much shorter than me1.perdeuterated benzene molecules undergoing different types of motion are collected in table 1. Values for the quadrupole frequency of deuterium in benzene can be found in the literature or wQ/21c = 145 f 2 kHz13 wQ/2x = 138 1 kHz.14 The latter value has been used in the present work since it corresponds to results of our measurements for benzene sorbed on silica at temperatures below 200 K.15 Benzene sorbed on HZSM-5 A typical 2H n.m.r. spectrum of benzene for high loadings and at low temperatures is given in fig. 1. From a comparison of this spectrum with the theoretical lineshapes (cf.table 1) we must conclude that the sorbed benzene molecules undergo fast reorientations about the Cs axes. This kind of molecular motion is also present in both the liquid16, l72350 2H N.M.R. of Benzene in Zeolites 0 35 70 105 140 [(a -ao)/2nl/kHz Fig. 1. 2H n.m.r. spectrum of 1.9 perdeuterated benzene molecules sorbed per channel intersection on HZSM-5 at 125 K. According to the used technique only the 'right half' (w 2 wo) of the spectrum is shown. 2048 accumulations, repetition time 0.5 s. and the solid13* l8 state of benzene. ' Fast' does mean that according to eqn (5) and (7) even at 125 K the correlation time z, of the c6 reorientation is much shorter than 1 ps. This experimental result is in accordance with the finding of our previous 13C n.m.r.lineshape study.1° The dependence of the 2H n.m.r. lineshape on temperature and loading is given in fig. 2. With increasing temperature the lineshape which is characteristic of an exclusive C, reorientation (cf. table 1) disappears, and above 170-180 K a rectangular curve is found. This change in lineshape is accompanied by a much stronger decrease in the signal-to-noise ratio than would be expected on the basis of Curie's law. Finally, above 270 K no signal (quadrupolar echo) is observed for a pulse interval of 60 ps. Fig. 2 shows that a reduced loading causes a similar effect as a rise in temperature. These experimental findings can be interpreted in terms of the following model for the molecular motion of benzene in the ZSM-5 structure derived from 13C n.m.r.investigations." (i) The benzene molecules sorbed on HZSM-5 undergo a fast motion about their C, axes. From the I3C n.m.r. lineshape a correlation time z, + 100 ps (even at 125 K) may be estimated. However, from the 2H n.m.r. spectra given in this study it follows that z, < 1 p s . (ii) Superimposed on this C6 reorientation there are jumps performed by the benzene molecules between a limited number of sorption sites; these allow only distinct orientations of the hexad axis of the benzene molecule. The mean residence time zj between two succeeding jumps becomes shorter with increasing temperature and decreasing loading. The value of zj at 200 K and for a loading of 1.5 molecules per channel intersection is ca. 10-100 ps. The assumption of a limited number of preferred sorption sites which permit only distinct orientations of the hexad axis of a benzene molecule seems to be justified: as semi-empirical calculations of the interaction potential between a sorbed benzene molecule and the zeolite framew~rkl~-~' have shown, such preferred sorption sites exist in the channel intersections as well as in the pore segments between them.B.Zibrowius, J. Caro and H. Pfeifer 235 1 Fig. 2. 2H n.m.r. 0 35 70 L 0 35 70 0 35 70 [ ( ~ - - ~ 0 ) / 2 ~ l / k H z spectra of perdeuterated benzene sorbed on HZSM-5 for various loadings [(a) 0.4, (b) 1.0, (c) 1.48nd ( d ) 1.9 molecules per channel intersection] and temperatures (in K). 4096 accumulations. Provided that the jumps are accompanied by a translational motion, we must conclude that the intracrystalline diffusion coefficient of benzene sorbed on HZSM-5 should decrease with increasing loading, since the mean residence time zj between succeeding jumps becomes longer.Such a concentration dependence of the diffusivity has been found in sorption-uptake studies. 22 A finite value of the mean residence time zj between two successive jumps of the benzene molecule has two important consequences for the 2H n.m.r. spectra obtained from quadrupolar echos. (i) The lineshape depends on the time interval z between the two n/2 if z z zj. The variation of the lineshape is especially pronounced if coQTj z 1-20. According to eqn (7) this should occur for values of zj of 1-20 ps. Since for the n.m.r. spectrometer used the shortest value of z is ca. 45 ,us, this effect cannot be investigated in this paper.(ii) The intensity of the 2H n.m.r. signal depends on the time interval z between the two n/2 pulses if z z zj. As a consequence of the motion, the 2H n.m.r. frequency according to eqn (1) or (4) changes during the time interval 22 between the first pulse and the echo. Therefore the refocussing of the n.m.r. signal caused by the second pulse becomes i n c ~ m p l e t e . ~ ~ Assuming that the difference 6 between the 2H n.m.r. frequency before and after the molecular jump is much larger than z;', it may be that the signal intensity I decreases exponentially according to I = I, exp (- 22/2,). (9) For powder samples, however, I generally decreases non-exponentially, since the decrease in I depends on the relative value of 6 compared with r;' and T - ' .~ ~ Fig. 3 shows the 2H n.m.r. signal intensity I (the amplitude of the quadrupolar echo) as a function of the pulse interval z for a highly loaded sample between 167 and 267 K. At low temperatures (T < 167 K) the intensity decreases exponentially, with a time constant of ca. 700,us. The corresponding relaxation rate is mainly caused by the magnetic dipole-dipole intera~tion,'~ since molecular jumps are too rare to be of significant influence (zi %- 7). As fig. 3 shows, at medium temperatures (T >, 170-180 K)2352 2H N.M.R. of Benzene in Zeolites the decrease becomes non-exponential, and above 270 K no echo can be obtained. From the latter finding it follows that zj < CCL. 100 ps at this temperature. Owing to the decrease in zj with decreasing loading, the model above explains why for low sorbate concentrations a quadrupole echo can be obtained only at very low temperatures (cf.fig. 2). Both the changes in lineshape and the reduction of the signal intensity could be used for a more detailed and quantitative analysis. However, such a procedure26 would necessitate the calculation of the interaction potential of a sorbed benzene molecule with the zeolite framework, including the interaction between the sorbed molecules. In a recently published paper,5 a 2H n.m.r. spectrum measured at 143 K for a loading of two benzene molecules per channel intersection was presented which is quite similar to the spectra obtained in the present paper for temperatures < 170 K (cf.fig. 2). However, with increasing temperature, Eckman and Vega5 observe an additional ‘narrow peak’ at co = coo. Therefore, they suggest the presence of a second, highly mobile benzene phase in the HZSM-5 framework. The simultaneous existence of two phases for benzene molecules sorbed in the intracrystalline space seems to be questionable if one compares this system with the behaviour of hydrated zeolites A, X and Y: only for a relatively large number (5-10) of interacting molecules per cavity does a (diffuse) phase transition occur.27 In contrast to ref. ( 5 ) we ascribe the ‘ narrow peak ’ to benzene molecules which are not sorbed inside the ZSM-5 channel network, i.e. to molecules on the outer surface of the individual crystals or in a secondary pore system if the individual crystals are stacked together forming polycrystalline grains.These benzene molecules outside the primary pore structure may move isotropically even far below the melting point, as it could be shown using silica as ad~0rbent.l~ Eckman and Vega5 have performed their measurements on ZSM-5 crystallites of a comparably small size (2-4 pm). As we have shown above (cf. fig. 3), the intensity of the molecules sorbed on HZSM-5 is strongly reduced at high temperatures owing to their enhanced mobility (a reduced value of rj). Therefore the relative intensity of the ‘narrow peak’ increases and appears ‘to be too great to be attributed completely to molecules on the external crystal s~rface’.~ Using a HZSM-5 specimen consisting of very small crystals (diameter <0.3 pm) which are clustered into spherulitic polycrystalline grains (diameter ca.12 we have also observed the narrow peak described in ref. (5). Furthermore, the occurrence of narrow peaks for p-xylene and toluene sorbed on ZSM-55 formed from small crystallites may be taken as additional support for our interpretation. Following a 13C n.m.r. the molecular mobility of p-xylene sorbed on HZSM-5 is strongly restricted. Even at 310 K the molecules do not perform fast isotropic reorientations, as in the liquid state, which are necessary to produce such narrow peaks in the 2H n.m.r. spectra. The molecular motion of benzene sorbed on ZSM-5 has also been investigated by quasielastic neutron ~cattering.~, The experimental results for a loading of 1.25 benzene molecules per channel intersection, i.e.5 molecules per unit cell, and at a temperature of 300 K are explained by: (i) three benzene molecules performing isotropic rotations and (ii) two benzene molecules ‘immobile’ on the timescale of the e~periment.~’ However, the benzece molecules performing isotropic rotations on the timescale of the neutron scattering experiment (i.e. 10-11-10-12 s) should cause liquid-like 2H n.m.r. lineshapes which are not observed in our experiment. Therefore, the second possible explanation given in ref. (30) seems to be more probable: not freely, isotropically rotating molecules, but molecules tumbling on the above timescale.B. Zibrowius, J . Caro and H. Pfeifer 2353 1 I I I I I I 0 60 100 200 30 0 400 500 TIPS Fig. 3. Dependence of the 2H n.m.r.signal intensity on the pulse interval for various temperatures. The loading of the HZSM-5 corresponds to 1.9 benzene molecules per channel intersection. The values for z = 0 were calculated on the basis of Curie’s law using the extrapolated (z = 0) intensity at T = 167 K: 0, 167; 0, 182; x , 200; 0, 222; +, 235; a, 250; ., 267 K. Benzene sorbed on Faujasites The temperature influence on the 2H n.m.r. lineshape of benzene sorbed on NaX and NaY type zeolites is shown in fig. 4. As in the case of benzene sorbed on HZSM-5, there is clear evidence (cf. table 1) for a fast C, reorientation of the sorbed molecules with a correlation time z, < 1 p s . On the other hand, our spectra show that on the timescale of the n.m.r. experiment the benzene molecules are strongly fixed at their sorption sites.According to eqn (9, all motions which change the orientation of the hexad axis of the molecule (e.g. jumps between sorption sites) have, even at T = 200 K, correlation times zj much larger than the reciprocal quadrupole frequency oQ, i.e. zj 9 1 p s . This may be caused by the interaction of benzene molecules with sodium ions via their 7z electrons as well as by a stronger mutual interaction between the sorbed molecules in the large cavities of the faujasite-type zeolites. Therefore, a smaller translational mobility of benzene sorbed on faujasites compared with HZSM-5 can be predicted for temperatures below room temperature and medium loading. In quasielastic neutron scattering inve~tigations~l the benzene molecules were found to interact in Na-mordenite predominantly with the sodium ions resulting in an uniaxial C, reorientation, with a correlation time z, x 2 ps at 300 K.The molecular mobility and sorption state of benzene sorbed on faujasites have previously been studied by proton n.m.r.32-34 From the minima in the longitudinal relaxation time as a function of the reciprocal temperature, values for the correlation time zj of the translational motion could be determined. A value of zj x 10ns was obtained for benzene sorbed on NaY at 440 K for a loading of 4 molecules per cavity and at 310 K for 1 molecule per cavity. Furthermore, it was shown that there is no distribution of correlation times, and Eyring’s equation for localized adsorption can be applied.33 The assumed sorption sites (Na+ in the SII position) represent sharp minima in the interaction potential surface of the benzene molecules with the zeolite framework.In the case of NaX, a higher mobility and a reduced mean jump length of the molecules were found.33- 34 These findings were attributed to the non-localizable sodium cations2354 2H N.M.R. of Benzene in Zeolites 250 222 200 182 167 12 5 0 35 70 0 35 70 [(a - ~ 0 > / 2 r I / k H z Fig. 4. Temperature dependence of the 2H n.m.r. lineshape for benzene molecules sorbed on (a) NaX and (b) Nay. The loading corresponds to 4 molecules per large cavity. 2048 accumulations, temperatures in K. (Na+ in S3) in NaX which increase the number of sorption sites and reduce the height of the potential barrier between them. Our 2H n.m.r. results agree very well with the experimental findings and the given interpretation of the above proton n.m.r.studies. (i) The 2H n.m.r. spectra reveal also a higher molecular mobility for benzene sorbed on NaX in comparison with Nay. Deviations from the powder spectrum of an axially symmetric field gradient tensor caused by motional processes occur for benzene on NaX at considerable lower temperatures compared with NaY (cf. fig. 4). (ii) Molecular jumps with a correlation time zj of the order of 10 ns should lead to narrow Lorentzian lineshapes also in the deuterium spectra. This was indeed observed for benzene in NaY at high temperatures. At 440 K the linewidths at half height amount to 560 Hz for a loading of 4 molecules per cavity and 190 Hz for 2 molecules per cavity. Consequently, at temperatures above room temperature the mobility of benzene in faujasites is significantly higher than in HZSM-5.A difference in the translational mobility of benzene has been observed also in diffusion measurements. While the intracrystalline diffusion coefficient for benzene on NaX is ca. 2 x m2 s-' at 353 K,35936 under similar experimental conditions a value of the order of m2 s-' was obtained for benzene on (Na, H)ZSM-5.22B. Zibrowius, J . Caro and H . Pfeifer 2355 Our experimental 2H n.m.r. spectra for benzene sorbed on NaX (cf. fig. 4) agree well with those published by Hasha et al.4 Nevertheless, there are differences in the interpretation of the spectra. First, for certain regions of temperature and loading, the spectra obtained in ref.(4) have been interpreted as a superposition of spectra for molecules in two different states of molecular motion. Secondly, the deviation of the signal intensity from Curie's law caused the authors of ref. (4) to postulate a third phase consisting of ' unobserved molecules '. In our opinion, however, both experimental findings can be explained simply by molecular motion, and hence the existence of different states of molecular motions or phases for benzene sorbed on NaX must be abandoned. (i) As is known from the molecular jumps with a mean residence time zj of the order of the reciprocal quadrupole frequency oQ give rise to spectra similar to those obtained in ref. (4). (ii) The reduction of the signal intensity due to molecular jumps has already been demonstrated above.A similar dependence of the echo amplitude on the pulse interval as given in fig. 3 for benzene sorbed on HZSM-5, was also observed in the case of faujasites. (iii) Furthermore, our interpretation is supported by the observed variation4 of the lineshape in dependence on the pulse interval z (30-80 ps). Such variations are characteristic of a molecular jump mechanism. 23 Conclusions For benzene sorbed on HZSM-5, NaX and NaY zeolites two kinds of molecular motion could be unambiguously distinguished : the uniaxial C, reorientation of the molecules and molecular jumps between the sorption sites. Even at low temperatures (T = 125 K) and high loadings, the C, reorientation is fast on the n.m.r. timescale. The correlation time of this motion must be much shorter than 1 ps.Superimposed on the C, reorientation there are jumps of the benzene molecules between adjacent sorption sites which lead to a reorientation of the c6 axes. Both increasing the temperature and decreasing the loading reduce significantly the mean residence time between two succeeding jumps. Assuming that these reorientations of the C, axes of the molecules are accompanied by a translational motion, this finding corresponds to the intracrystalline diffusivities as obtained by means of sorption uptake experiments. At temperatures above room temperature a higher mobility has been found for benzene molecules sorbed on Na-faujasites compared with that of benzene in HZSM-5. This finding agrees with sorption uptake results. At temperatures below room temperature, the mobility of the sorbed benzene molecules is in HZSM-5 higher than in Na-faujasi tes. Molecular motions with correlation times much shorter than the reciprocal quadrupole frequency oQ give finger-print-like 2H n.m.r.lineshape patterns. For motions with correlation times of the order of oil, the peculiarities of the quadrupolar echo technique have to be taken into account. A more detailed and quantitative analysis of the spectra for zj z WQ' would necessitate the calculation of the interaction potential of a benzene molecule with the zeolite framework including the interaction between the sorbed molecules. We thank Dr W. Oehme (Leipzig) for many stimulating discussions as well as experimental assistance. Prof. Dr M. Biilow (Berlin) is thanked for his continuous interest in this work and valuable comments.We are indebted to Prof. L. V. C . Rees (London) for providing the ZSM-5 zeolite of high quality and Mr J. Richter-Mendau (Berlin) for the scanning electron micrographs. We also thank Dr H. Jobic (Villeurbanne) for helpful comments on the manuscript of this paper.2356 2H N.M.R. of Benzene in Zeolites References 1 H. W. Spiess, Colloid Polym. Sci., 1983, 261, 193. 2 F. A. Bovey and L. W. Jelinski, J. Phys. Chem., 1985, 89, 571. 3 R. Eckman and A. J. Vega, J. Am. Chem. Soc., 1983, 105,4841. 4 D. L. Hasha, V. W. Miner, J. M. Garces and S. C. Rocke, A.C.S. Symp. Ser., 1985, 288, 485. 5 R. Eckman and A. J. Vega, J. Phys. Chem., 1986, 90, 4679. 6 Z. Luz and A. J. Vega, J. Phys. Chem., 1986, 90, 4903.7 Z. Luz and A. J. Vega, J. Phys. Chem., 1987,91, 374. 8 R. J. Argauer and G. R. Landolt, US. Patent, 3.702.886, 1972. 9 M. Hunger, D. Freude, T. Frohlich, H. Pfeifer and W. Schwieger, Zeolites, 1987, 7, 108. 10 B. Zibrowius, M. Bulow and H. Pfeifer, Chem. Phys. Lett., 1985, 120, 420. 11 J. Davis, K. Jeffrey, M. Bloom, M. Valic and T. Higgs, Chem. Phys. Lett., 1976, 42, 390. 12 H. W. Spiess, in NMR: Basic Principles and Progress (Springer, Berlin, 1978), vol. 15. 13 J. Rowell, W. Phillips, L. Melby and M. Panar, J. Chem. Phys., 1965, 43, 3442. 14 N. Boden, S. M. Hanlon, Y. K. Levine and M. Mortimer, Mol. Phys., 1978, 36, 519. 15 B. Zibrowius and W. Oehme, unpublished results. 16 D. J. Winfield and D. K. Ross, Mol. Phys., 1972, 24, 753. 17 P. Linse, S.Engstrom and B. Jonsson, Chem. Phys. Lett., 1985, 115, 95. 18 E. R. Andrew and R. G. Eades, Proc. R. SOC. London, Ser. A, 1953, 218, 537. 19 S. Ramdas, J. M. Thomas, P. W. Betteridge, A. K. Cheetham and E. K. Davies, Angew. Chem., Int. 20 H. Stach, R. Wendt, K. Fiedler, B. Grauert, J. Janchen and H. Spindler, Proc. IUPAC-Symposium on 21 A. K. Nowak, A. K. Cheetham, S. D. Pickett and S. Ramdas, to be published. 22 A. Zikanova, M. Bulow and H. Schlodder, Zeolites, 1987, 7, 115. 23 H. W. Spiess and H. Sillescu, J. Magn. Reson., 1981, 42, 381. 24 H. S. Gutowsky, R. L. Vold and E. J. Wells, J. Chem. Phys., 1965, 43, 4107. 25 A. J. Vega, J. Magn. Reson., 1985, 65, 252. 26 M. S. Greenfield, A. D. Ronemus, R. L. Vold, R. R. Vold, P. D. Ellis and T. E. Raidy, J. Magn. 27 H. Pfeifer, W. Oehme and H. Siegel, Ann. Phys., 1985, 42, 496. 28 J. Karger, H. Pfeifer, J. Caro, M. Bulow, J. Richter-Mendau, B. Fahlke and L. V. C. Rees, Appl. 29 J. B. Nagy, E. G. Derouane, H. A. Resing and G. R. Miller, J. Phys. Chem., 1983, 87, 833. 30 H. Jobic, A. Renouprez, F. Vignt-Maeder, M. Bee and C. Poinsignon, in Dynamics of Molecular Crystals, ed. J. Lascombe (Elsevier, Amsterdam, 1987), p. 573. 31 H. Jobic, M. Bee and A. J. Renouprez, Surf Sci., 1984, 140, 307. 32 M. Nagel, H. Pfeifer and H. Winkler, Z . Phys. Chem. (Leipzig), 1974, 255, 283. 33 H. Pfeifer, ACS Symp. Ser., 1976, 34, 36. 34 H. Pfeifer and H. Winkler, Proc. 4th Spec. CON. Ampere, Leipzig, 1979, p. 31. 35 M. Bulow, W. Mietk, P. Struve and P. Lorenz, J. Chem. SOC., Faraday Trans. I , 1983, 79, 2457. 36 M. Bulow, W. Mietk, P. Struve, W. Schirmer, M. KoEiiik and J. Karger, Proc. 6th Int. Con$ Zeolites, 37 0. Pschorn and H. W. Spiess, J. Magn. Reson., 1980, 39, 217. Ed. Engl., 1984, 23, 671. Characterization of Porous Solids, Bad Soden, FRG, April 1987, to be published. Reson., 1987, 72, 89. Catal., 1986, 24, 187. Reno, USA, 1983, ed. D. Olson and A. Bisio (Butterworths, Guildford, 1983), p. 242. Paper 711599; Received 2nd September, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402347

出版商:RSC    

年代:1988

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. I , 1988, 84(7), 2357-2367 Photoelectrochemical Electron Spin Resonance Part 2.-The Reduction of Crystal Violet in Acetonitrile Richard G. Compton,* Barry A. Coles, Geoffrey M. Stearn and Andrew M. Waller Physical Chemistry Laboratory, Oxford University, South Parks Road, Oxford OX1 3QZ The reduction of the dye crystal violet in acetonitrile solution at platinum electrodes has been investigated using in situ electrochemical e.s.r. and a channel electrode flow cell. In the dark, in the rigorous absence of oxygen, a one-electron reversible reduction takes place, producing the stable crystal violet radical, the e.s.r. spectrum of which was recorded. The behaviour in the presence of oxygen is also discussed. On irradiation with light of wavelength ca.400 nm the radical is thought to undergo a chemical reaction with the solvent, and the electrode reduction process has the characteristics of an EC mechanism. The nature of this reaction is deduced from the e.s.r. signal strength/current/electrolyte flow rate data together with pure electrochemical measurements, and kinetic parameters are reported. In a previous paper we have described' a flow cell suitable for simultaneous photoelectrochemical electron spin resonance studies. In particular it was demonstrated that a channel electrode made in silica could be located near the centre of the cavity of an e.s.r. spectrometer without seriously damaging its sensitivity, thus allowing the identification of radical species formed during pho toelectrochemical reactions taking place at the electrode.Moreover, the known pattern of flow in the channel electrode cell enabled the calculation of the concentration profiles of electrogenerated radicals which either were stable,' or else decomposed with first-order kinetics.2 In this way the steady- state e.s.r. signal was related to the generating current, solution flow rate and electrode geometry. The cell was also shown to be well suited to the study of dark electrode reaction mechanisms and this was exploited in the study of ECE2* and DISP4 processes, as well as more complicated reactions.5* The role of e.s.r. in the investigation of electrode reaction mechanisms in general has been reviewed recently.' In this paper we employ the flow cell in the study of the reduction at a platinum electrode of the dye crystal violet (CV') in acetonitrile as solvent.This has been shown to be a reversible one-electron process8 resulting in the formation of the free radical (C V.) : NMel I 23572358 Reduction of Crystal Violet in CH,CN In this paper we use in situ electrochemical e.s.r. to show that the free radical product is stable on the timescale of our experiment and to investigate its behaviour in the presence of 406 nm light, which is shown to induce its decomposition. The nature of this photochemical reaction is deduced and kinetic parameters are derived. Experimental The basic apparatus and techniques have been described previously.' The silica channel electrode unit was 30 mm long and had approximate cross-sectional dimensions of 0.4 mm x 6.0 mm.Platinum foils (4.0 mm x 4.0 mm), used as electrodes in this unit, were cemented on to the silica cover plates and were carefully polished flat (with a succession of progressively finer diamond lapping compounds down to 0.25 pm) before use. The e.s.r. spectrometers used were as reported before.'gg Irradiation was provided by a Wotan XBO 900 W/2 xenon arc lamp via a Jarrell-Ash 82-410 grating monochromator. The radiation power incident at the electrode position was typically of the order 40 mW cm-'. Power was varied by attenuating with wire gauzes af differing mesh size. Radiation intensity was measured using an International Light IL700 photometer. A reference electrode was located in the flow system upstream of the cavity. This was either a saturated calomel electrode or a silver-wire pseudo-reference electrode.All experiments were carried out at 25 "C. Acetonitrile (Fisons, dried distilled) was refluxed with calcium hydride and then fractionally distilled. Tetrabutylammonium perchlorate, TBAP (Fluka, purum) was recrystallised once from a mixture of ethanol and petroleum ether. Either this, or lithium perchlorate (Fluka), was used as supporting electrolyte. Crystal violet (as the chloride) was used as received from Aldrich. Electrolysis was performed in acetonitrile solution containing 0.1 mol dmP3 of supporting electrolyte. Oxygen was removed by outgassing with nitrogen (pre-purified of trace oxygen and dried) prior to electrolysis. Electro- chemical measurements were carried out using an Oxford Electrodes potentiostat, modified to boost the counter-electrode v01tage.~ Complementary rotating-disc experiments utilised an Oxford Electrodes rotating-disc assembly and motor controller.U.v.-visible absorption spectra were recorded on a Unicam SP 800 spectrophotometer. Supporting theory (see Appendix) was generated from programs written in FORTRAN on a Norsk Data 520 computer. NAG 11 library routines and GHOST 80 graphics were used. Graphical output utilised a Calcomp 81 plotting device. Results and Discussion The reduction, in the dark, of a 5 x lo-* mol dm-, acetonitrile solution of crystal violet was studied, first, using both rotating-disc and channel platinum electrodes. With rigorous deoxygenation, the half-wave potential was found to be at -590 mV with respect to the standard calomel electrode.Mass-transport-corrected Tafel analysis'O of the reduction wave yielded a straight line of slope 59 mV per decade, indicating that the reduction is electrochemically reversible. In agreement with this the half-wave potential was found to be independent of rotation speed. Transport-limited currenp, ILIM, were analysed by Levich plots: fo'; the rotating disc, ILIM us. (rotation speed)z;1'*12 for the channel, ILrM us. (flow rate)3.l3 These plots were found to be linear, allowing the deduction of the diffusion coefficient of crystal violet as (1.3 f 0.1) x Thorough deoxygenation of the solution was found to be crucial for the generation of reproducible dark electrochemistry. In the presence of trace oxygen the initial part of the reduction wave was displaced to more positive potential (fig.1). This is indicative of a chemical reaction between CV' and 0,, giving rise to an EC process, further evidence for which was gleaned from the fact that the overall limiting current for both waves (the EC wave and the normal CV' reduction wave) was unchanged from that observed on cm2 s-'.R. G. Compton et al. 100 0- 2359 - - - - rigorous deoxygenation, as can be seen from fig. 1. Under the conditions of partial deoxygenation, such as those prevailing during the recording of the curve shown in fig. 1, the concentration of oxygen in solution is less than that of crystal violet, and in the vicinity of the electrode becomes exhausted by reaction with the radical CV'. Thus the reaction causes only part of the wave to be displaced.The following two possibilities were considered for the nature of the reaction of the crystal violet radical with oxygen. CV' + 0, -+ cv+ + 0, NMe2 I CV' + 0, -+ I NMe2 The former, which is a simple electron transfer, has been suggested by Fischer et aI.l4 However, under the conditions of our experiment that reaction would be expected to give rise to an increased limiting current as, overall, two electrons would be transferred for each crystal violet cation reduced: CV+ + 0, + 2e- + CV' + 0;. Such behaviour is not observed (fig. 1) and so we prefer the second scheme. This reaction is well established, e.g. ref. (15), for the structurally similar Ph,C' radical and, indeed, far many other hydrocarbon radicals. We consider next electrochemical e.s.r.measurements made in the absence of light. A radical species was detected, using the channel electrode cell, at potentials corresponding to, the reduction of CV+ (fig. 2). This was attributed to CV', although the hyperfine structure was insufficiently resolved for unambiguous identification. This poor resolution has been attributed14 to spin exchange between the radical and the parent cation. It was observed that when the solution was incompletely deoxygenated no signal was seen until the oxygen-induced EC pre-wave was surmounted (fig. 3), i.e. until the dissolved oxygen had been exhausted by 'titration' with CV'. This observation is consistent with the2360 Reduction of Crystal Violet in CH,CN Fig. 2. The e.s.r. spectrum of the crystal violet radical.0 40 80 12 0 current/ 1 0-6 A Fig. 3. The effect of oxygen on the e.s.r. signal intensity measured at constant flow rate. 2.0 0 -3.0 - 2.0 - 1 -0 log (U/cm3 s-') Fig. 4. The e.s.r. signal-current-flow rate variation, in the dark (0) and under irradiation with 406 nm light of intensity 35 mW cmT2 ( x ). The dashed line has the gradient of -$ expected for a stable radical electrogenerated at the channel electrode. Units : S , arbitrary ; I, PA.R. G. Compton et al. 236 1 2ot 0.1 v Fig. 5. A voltammogram for the reduction of crystal violet in the presence (a) and absence (b) of light, showing the shift in half-wave potential to more positive potentials on illumination. mechanism proposed above, indicating that, in the presence of oxygen, there is insufficient build-up of CV' for the detection of the radical by e.s.r.Once the oxygen has been removed the e.s.r. signal intensity increases linearly, as would be expected for a stable radical.' The intercept of this linear portion with the current axis gives a value of ca. 44 x A, in excellent agreement with the height of the EC pre-wave measured under the corresponding conditions. The e.s.r. signal (S)-current (I)-flow rate ( U ) behaviour was then examined under conditions of rigorous deoxygenation. It was found (fig. 4) that a plot of log ( S / I ) us. log U was linear, with a slope close to --$ This indicates that the CV' radical is 'stable'l on the timescale of flow through the cavity at the flow rates used (10-4-10-1 cm3 s-l), corresponding to a lifetime of, say, 3 10 s.We consider next experiments carried out in the presence of light. U.v.-visible absorption experiments showed the parent material (CV') to have an absorption peak at 560nm. On irradiation no photocurrents were observed at this wavelength at any potential. However, photocurrents were detected on the crystal violet reduction wave, on illumination at ca. 400 nm. The effect was examined over the whole reduction wave by periodically switching the radiation on and off during a slow potential scan. A typical voltammogram so produced is shown in fig. 5. The effect of the irradiation is to shift the half-wave potential to a more positive potential, without any significant change in the transport-limited current. As indicated above, such behaviour is typical of an EC-type mechanism.Since the radical is known to absorb at ca. 400 nm,16 the process must be initiated by its excitation. Thus the following scheme describes the process : CV+ + e- --+ CV' CV' + hv (ca. 400 nm) --+ CV'* --+ products. The maximum photoeffect was obtained by measuring photocurrents at a fixed potential as a function of wavelength. This was found to occur at 406 nm (fig. 6), and this wavelength was employed for all subsequent irradiation. The shift in half-wave potential, AE;, was recorded as a function of both flow rate ( U/cm3 s-') and the incident radiation intensity (I,). Since U.v.-visible absorption measurements showed CV+ itself to be transparent at 406 nm, the radiation intensity at the electrode is essentially that incident on the channel cell.Results were analysed using the 'working curve' for an EC reaction (see Appendix) to give a valu? of the normalised rate constant P (also see Appendix). P was plotted against U-3 (fig. 7) and good straight-line plots were obtained, supporting the choice of mechanism. The slopes enabled the deduction of the first-order rate constant, k, for the following reaction, at each intensity, I,. The following rate constants were deduced : I,/mW cmP2 35.0 16.3 7.3 2.8 kls-' 2.08 1.13 0.72 0.30.2362 Reduction of Crystal Violet in CH,CN ;I g 2.0 c a 1 .o 01 I I 0 I I I I I I 360 400 4 4 3 480 X/nm Fig. 6. The wavelength dependence of the photocurrent. 30 20 K * 10 0 /‘ / / x 0 / 0 50 10 0 150 200 ( ~ / c m 3 s-1>-% Fig. 7. Plots of the normalised rate constant, K, for an EC process deduced from the shift in half- wave potential with the flow rate ( U ) and the working curve shown in the Appendix.Different light intensities (mW cm-2) are shown: x , 35; 0, 16.3; ., 7.3; A, 2.8. In each case the dashed line shows the least-squares fit through the data. The variation of AE; with log U was then computed (see Appendix) for each rate constant using the measured cell geometry and flow rate, and for each radiation intensity. The experimental E; values were in good agreement with the computed values and are depicted in fig. 8 for the 35 mW radiation. The photoreaction was studied fu5ther by e.s.r. spectroscopy. Whereas in the absence of light the relationship, S / I K U-i, held, showing the radical to be stable on the timescale of the experiment, this was not so on irradiation of the electrode with 406 nm light, confirming that the radical was no longer stable.This can be seen from fig. 4. The theoretical variation of the e.s.r. signal strength with flow rate was computed, as described in the Appendix, using the rate constant obtained from the AE; measurements. The experimental results fit reasonably well with the chosen rate constant (see fig. 9),R. G. Compton et al. 2363 1.5 L 5 f 1.0 0.5 1 I I I I 1 -3.5 -2.5 -1.5 -0.5 log (U/cm3 s-l) Fig. 8. The computed (-) and experimental (0) variation of AE; with flow rate for the case where the cell is irradiated with 406 nm light of intensity 35 mW cm-2. 0.8 7 0.6 Y 'yl 0.4 9 -3 2 0.2 00 N I1 0 -3.0 -2.0 -1.0 log(U/cm3 s-l) Fig. 9. Calculated (-) and experimental (0) variation of the e.s.r.signal strength, S, with flow rate. The data were obtained with the electrode being irradiated with 406 nm light of intensity 35 mW cmP2. The cell dimensions are as defined in the text. particularly since the theory assumes that the channel cell is uniformly irradiated. These observations may thus be regarded as giving an independent check on both the mechanism and the value of k. We next consider the fate of the excited radical, CV'*. The negligible increase in the limiting current on irradiation (fig. 5 ) would rule out electron loss giving CV+. Moreover, since the results were unaffected by changing the supporting electrolyte to lithium perchlorate, the involvement of the Bu,N+ ion is excluded. It thus appears likely that the solvent is implicated, and we suggest the possible formation of the leuco-form of the dye, CV-H, by hydrogen abstraction from the solvent.The mechanism may then be written as, CV+ + e- g CV' CV' + hv --+ CV'*; rate = I,E[CV'] rate = k,[CV'*] rate = kd[CV'*] CV'* -+ CV'; CV'* + CH,CN -+ products ;2364 Reduction of Crystal Violet in CH,CN ..o/ / / / / / 0 10 2 0 30 4 0 lo/m W cm -2 Fig. 10. The dependence of the first-order rate constant, k , for the following reaction, on the intensity of the irradiation. where k, is the pseudo-first-order rate constant for the reaction with acetonitrile. Steady- state kinetics on CV'* leads to k = IO&/(l + k,/k,). Thus a linear dependence of k on I. is expected. The results given earlier are plotted in this way in fig.10 and are thus consistent with the above general mechanism. In conclusion, we can say that the reduction of crystal violet in the dark leads to the formation of a stable radical, CV', but that on irradiation at a frequency corresponding to the absorption band of CV' a first-order decay of the radical occurs, leading to an EC mechanism for the electroreduction. The rate constant for this decay depends linearly on the intensity of the incident radiation and may be reasonably ascribed to hydrogen abstraction from the solvent. We thank the S.E.R.C. for a studentship for A.M. W. Appendix Here we outline the theory required for the analysis of the experiments reported above. In the first instance we consider the behaviour of a simple EC process at a channel electrode, i.e.A + e - e B k B -+ products where k is the first-order rate constant of the 'following' reaction. In the case of a reduction the consequence of the chemical step is the shifting of the voltammetric wave to more positive potentials. The pertinent steady-state convective-diffusion equations are andR. G. Compton et al. 2365 L Fig. 11. The coordinate system for the channel flow cell. 1.2 h 5 0.8 Q k c 0*4 0 I I I I I I I I 1 1 .o 0.5 0 0.5 1 .o log K * Fig. 12. Working curves for the EC mechanism. The dashed line represents that generated via analytical methods" with the Levich approximation. The solid lines give the behaviour calculated with the BI method for three flow rates:18 (a) 0.001, (6) 0.01 and ( c ) 0.1 cm3 s-l. where the coordinates x and y are defined in fig.11. The quantity v, is the (parabolic) solution velocity flow profile in the channel, V , = v,( 1 -y'2/h2) where v, (= 3U/2hd) is the solution velocity in the centre of the channel, h is the half- height of the channel (fig. 11) and y' = h-y. An analytical solution of the above equations has been obtained17 with a 'Levich' approach which invokes the LevEque approximation, (A 4) which linearises the flow profile. This is a good approximation provided the concentration changes induced by the electrode are confined close to the electrode surface. It was predicted" that the half-wave potential shift would depend on the following normalised rate constant, (A 5 ) V , x 2v0(l - y ' / h ) = 2 ~ , y / h K* = k(xe2 h2/4vO2D)f2366 Reduction of Crystal Violet in CH3CN c a v i t y length I- 1 y = 2h flow- [ I I I 1 y = o ;r I I ! I : I ! I I 1 1 I I I I -xu 0 xc xe Xd Fig.13. The coordinate system defining the location of the channel electrode in the e.s.r. cavity. where x, is the length of the electrode and D is the diffusion coefficient (assumed to be equal for A and B). The calculated 'working curve' is shown by the dashed line in As suggested above, this curve will only work when the perturbations in concentration are near to the electrode surface, i.e. for deep channels and fast flow rates. For greater generality we have numerically solved eqn (A 1 ) and (A 2), retaining eqn (A 3) rather than eqn (A 4) for the solution flow profile.'* To do this the backwards implicit (BI) method18-2' was employed.In this method we cover the xy plane with a two-dimensional finite difference net such that the difference between grid points in the x direction is x,/K and in the y direction is 2h/J. Concentrations at points on the network are given by cj, k, where the subscripts j and k refer to the variables y and x, respectively, fig. 12. The BI method expressions, and Moldoveanu xk = ( k / K ) x , ; k = 0,1,2 ... K (A 6) y j = ( j / J ) ( 2 h ) ; j = 0, 1,2 ... J . (A 7) involves replacing the derivatives in eqn (A 1) and (A 2) by the following and Anderson2' have shown how the concentrations c ~ . ~ can be computed from these equations, and we have adopted their general method without any important modifications. The specific changes required for the computation of the EC process at the channel electrode have been detailed in ref.(18). Grid sizes defined by J = 1000 and K = 500 were found to be sufficiently large for the results to become independent of J and K, i.e. for satisfactory convergence. Fig. 12 shows the results of the BI calculations for a channel cell of the same geometry as used in the experimental work described in this paper (x, = 0.4 cm, 2h = 0.04 cm, d = 0.6 em, w = 0.4 cm) and for D = 1.3 x cm2 s-l. At the fastest flow depicted ( 1 0-1 cm3 s-l) the behaviour is almost indistinguishable from the ' Levich ' curve, but at slower flow rates small but significant deviations arise. The same theory is readily used to generate a plot of AEi us. U for a given rate constant, diffusion coefficient and cell geometry, such as depicted in fig.8. Since the BI method generates the concentration profiles throughout the channel electrode, the method is readily extended to the calculation of the e.s.r. signal strength due to an electrogenerated radical. Fig. 13 shows the coordinate system defining the position of the electrode relative to the e.s.r. cavity. x, denotes the centre of the cavity. The cavity length, I = x,+u,, where -xu and x, denote the x coordinates of theR . G . Cornpton et al. 2367 upstream and downstream edges of the cavity, respectively. Thus x, = x,-1. In the experimental work undertaken, 1 = 2.4 cm and x, = 0.15 cm. The e.s.r. cavity has a cos2 sensitivity profile, maximum sensitivity being at the centre of the cavity, at x = x,. The sensitivity at the cavity edges (-xu and x,) is effectively zero.The e.s.r. signal strength is given by S/S* = cos2 [(x-xx,) n/l] (1; cB dy) dx (A 10) where S* is the signal due to 1 mol of the e.s.r.-active species, B, located at the centre of the cavity. Since the radical species is generated at the electrode the integral can be shortened to S/S* = rd cos2 [(x - x,) n/ 11 (1: c” dy ) dx. 0 For numerical summation C O S ~ [(k AX - x,) n/ 11 where Ax = x,/K and Ay = 2 h / J . j is summed from 0 to J , and k from 1 to K‘, where K‘ = x,/Ax. The summation in eqn (A 12) was used to calculate the e.s.r. signal for the crystal violet radical, decaying via an EC process in a flow cell corresponding to the geometry employed experimentally. J = 1000 and K = 500 were found to give satisfactory convergence.The dependence of S on the flow rate U is shown in fig. 9. References 1 B. A. Coles and R. G. Compton, J . Electroanal. Chem., 1983, 144, 87. 2 R. G. Compton, D. J. Page and G. R. Sealy, J. Electroanal. Chem., 1984, 161, 129. 3 R. G. Compton, D. J. Page and G. R. Sealy, J. Electroanal. Chem., 1984, 163, 65. 4 R. G. Compton, P. J. Daly, P. R. Unwin and A. M. Waller, J. Electroanal. Chem., 1985, 191, 15. 5 R. G. Compton, B. A. Coles and M. J. Day, J . Electroanal. Chem., 1986, 200, 205. 6 R. G. Compton, B. A. Coles and M. J. Day, J . Chem. Res., 1986, (3260. 7 A. M. Waller and R. G. Compton, Comprehensive Chemical Kinetics, vol. 29, in press. 8 I. NEmcova and I. NEmec, Chem. Zvesti, 1972, 26, 1464. 9 R. G. Compton and A. M. Waller, J . Electroanal. Chem., 1985, 195, 289. 10 W. J. Albery, Electrode Kinetics (Clarendon Press, Oxford, 1975), p. 79. I 1 V. G. Levich, Acta Phys. Chem. URSS, 1942, 17, 257. 12 V. G. Levich, Zh. Fiz. Khim., 1944, 18, 335. 13 V. G. Levich, Physicochemical Hydrodynamics (Prentice-Hall, Englewood Cliffs, NJ, 1962). 14 V. Fischer, W. G. Harrelson, C. F. Chignell and R. P. Mason, Photobiochem. Photobiophys., 1984, 7, 15 H. Singh and J. M. Tedder, J . Chem. SOC., Chem. Commun., 1981, 70. 16 M. Shimura, K. Shakushiro and Y. Shimura, J . Appl. Electrochem., 1986, 16, 683. 17 B. A. Coles and R. G. Compton, J . Electroanal. Chem., 1981, 127, 37. 18 R. G. Compton, M. B. G. Pilkington and G. M. Stearn, J . Chem. Soc., Faraday Trans. I, 1987, 84, 19 R. G. Compton, M. B. G. Pilkington, G. M. Stearn and P. R. Unwin, J . Electroanal. Chem., 1987, 20 S . Moldoveanu and J. L. Anderson, J . Electroanal. Chem., 1984, 178, 45. 21 J. L. Anderson and S. Moldoveanu, J . Electroanal. Chem., 1984, 179, 107; 119. 111. 21 55. 238,43. Paper 711629; Received 7th September, 1987

ISSN:0300-9599    

DOI:10.1039/F19888402357

出版商:RSC    

年代:1988

数据来源: RSC