摘要:

J. Chem. SOC., Faraday Trans. I , 1989, 85(3), 551-560 Acid-Base Equilibria in Aqueous Micellar Solutions Part 3.-Azine Derivatives Calum J. Drummond,*t Franz Grieser, and Thomas W. Healy Department of Physical Chemistry, The University of Melbourne, Parkville, Victoria, 3052, Australia The acid-base behaviour of two azine derivatives, viz. neutral red and acridine, in aqueous non-ionic Brij-35 and n-octyl 8-D-glucoside micellar solutions, in aqueous anionic sodium dodecyl sulphate micellar solutions and in 1,6dioxane-water mixtures have been investigated. The factors primarily responsible for the difference between the pK, value of an azine derivative in pure water and its apparent pK, value when it is located quantitatively within the micelle-aqueous solution interface have been determined.For the non-ionic micellar systems, the difference can be explained solely in terms of the different intrinsic solvent properties of the two solvating media. For the anionic micellar systems three major factors are considered to be responsible for the difference. These factors are the electrostatic micellar surface potential, the interfacial solvent characteristics, and specific molecular interaction between the cationic protonated moieties of the azine derivatives and the anionic surfactant headgroups. Neutral red has been used extensively as a probe in model biological systems. The interfacial acid-base dissociation of neutral red has been exploited to monitor the kinetics of proton deposition inside thylakoid membranes during the photosynthetic oxidation of The difference between the bulk aqueous pK, value and the apparent pK, value of neutral red adsorbed onto intact bovine rod outer segments has enabled it to be employed to measure the ionic permeability properties of these membrane^.^^^ In addition, neutral red has been utilised in a number of studies aimed at gaining insight into the solvent and/or electrostatic characteristics of the aqueous interfacial microenvironments of micelles,6-8 biological membrane^,^ proteinslO and polyelectrolytes.l1 Therefore, it is perhaps surprising that the factors which contribute to the magnitude of the apparent interfacial pK, values of neutral red have not hitherto been ascertained. In the past, very little attention has been paid to acridine as a probe.12 Nevertheless, there is a vast array of acridine-related indicators, many of which have been employed in aqueous micellar l4 Moreover, acridine has the same prototropic group as neutral red and therefore is ideally suited for comparison purposes. The present investigation was undertaken in order to determine the factors which are principally responsible for the difference between the bulk aqueous pK, value of an azine derivative and the apparent pK, value measured for the same azine derivative when it is located within the aqueous interfacial region of either a non-ionic Brij-35 micelle, a non- ionic n-octyl 8-D-glucoside (OG) micelle or an anionic sodium dodecyl sulphate (SDS) micelle.Two azine derivatives, uiz. acridine and neutral red were examined. The species involved in the acid-base equilibrium of each of the selected azine derivatives are depicted in fig.1. t Present address : CSIRO, Division of Chemicals and Polymers, G.P.O. Box 433 1, Melbourne, Victoria, 3001, Australia. 55 1552 Azine Acid-Base Equilibria in Micelles - H+ 2 ' +H+ I H Fig. 1. Acid-base equilibrium of (a) acridine and (b) neutral red. Experimental Acridine and neutral red were obtained from Sigma Chemical Co. and Tokyo Kasei Kogyo Co., respectively. The purity of both of these azine derivatives was confirmed by recommended IUPAC tests? The water, surfactants, inorganic reagents and organic solvents used in this study were of the same grade and from the same source as those given in Parts 1 and 2 of this series.l'. l7 pH-Titrations were performed with the experimental arrangement detailed in Part 1 of this series.l6 Titrations were conducted at 25 "C.The acid-base equilibria of both acridine and neutral red can be represented by Ka HIN+ e H+ + IN (1) where HIN+ is the singly charged protonated form of the azine derivative, IN is the neutral deprotonated form of the azine derivative, and H+ is the proton. The thermodynamic acid-base equilibrium constant for equilibrium (1) is given by where [H+], [IN] and [HIN+] denote the concentrations of the species involved in the acid-base equilibrium and yH+, yIN and YHIN+ the activity coefficients of the species referred to the particular solvating medium at infinite dilution. It was assumed that the activity coefficients of the singly charged species in pure water (w) and 1,4-dioxane-water mixtures (m) can be given by the mean activity coefficient for HC1, y * , in the same medium at the same total ionic strength.18 The activity coefficients of the HIN+ species residing within the interfacial microenvironments of micelles (i) were unknown.Therefore, the correction for the log (YIN/YHIN+) term could not be included in the determination of the apparent acid-base equilibrium constants in micellar solution, pKZbS. For the titrations carried out in pure water and aqueous micellar solution, the negative logarithm of the hydrogen ion activity (pH) was given by the pH-meter reading. The method of Van Uitert and Haas'' was employed to obtain the hydrogen-ion activity from the pH-meter reading (B) in 174-dioxane-water mixtures, i.e.- log ([H+] y:) - = B+ log VH (3) where log WH is a correction factor discussed in Part 2 of this series.17C. J . Drummond, F. Grieser and T. W. Healy 553 Hence The u.v.-visible absorption spectra of the azine derivatives were utilized to .obtain the magnitude of the [IN/[HIN+] ratio as a function of pH. This was accomplished by means of the relationship [IN] - 1-cc [HIN+] - 7 (7) A - A I N A H I N + -Am with 01= where A is the absorbance at the long-wavelength band maximum of the protonated form of the azine derivative at the particular pH under study, and AHIN+ and A,, are the absorbances at this wavelength when all the molecules are protonated and deprotonated, respectively. Representative examples of the u.v.-visible absorption spectrum of acridine and neutral red as a function of the bulk aqueous pH in micellar solution can be found in ref.(20). Results The theoretical background to the forthcoming analysis of the interfacial acid-base equilibria of the azine derivatives has been given in Part 1 of this series." Providing there is no significant contribution to the apparent acid-base equilibrium constant of an interfacially located azine derivative from specific solute-solvent interactions, the following relationships hold pKg = PK," - log (8) where myH+ is the medium effect on the proton thoroughly discussed in Part 1 of this series,16 ytN and y h I N + denote the activity coefficients of the deprotonated and protonated forms of the azine derivative, respectively, referred to the interfacial phase at infinite dilution, and I;, R, T and yo represent the Faraday constant, the universal gas constant, the absolute temperature and the electrostatic surface potential, respectively.In addition, it is convenient to define Fig. 2 and 3 show the acridine and neutral red ApK," and ApKQ behaviour as a function of the dielectric constant of 1,4-dioxane-water mixtures. The experimental data554 Azine Acid-Base Equilibria in Micelles 10 30 50 70 dielectric constant Fig. 2. Acridine ApK," (0) and ApK: (a) values as a function of the dielectric constant of 1,4- dioxane-water mixtures. 0 * -1.0 - 2.0 I I I I 10 30 50 70 dielectric constant Fig. 3. Neutral red ApK," (0) and ApK: (0) values as a function of the dielectric constant of 1,4- dioxane-water mixtures.C.J . Drummond, F. Grieser and T. W. Healy 555 Table 1. pH-titration results for acridine and neutral red in pure water and aqueous non-ionic micellar solutions, and the corresponding cerr values A,,, k l/nm pK;bS APK: %ff derivative medium" HIN+ (IN) acridine water 5 YO Brij-35 10% Brij-35 5% Brij-35 10% Brij-35 40 mg Brij-58/mle neutral red water 5 % OG 10% OG 353 (353) 353 (355) 353 (355) 530' (447) 541 (456) 540 (456) 539 (454) 540 (454) ~ - 5.26 f 0.03O - 3.87 f 0.03 - 1.39 43 f I 3.69f0.02 - 1.57 39f 1 - - 6. 5d 5.69f0.06 -0.81 41 +2 5.64 f 0.09 - 0.86 40 f 4 5.50 f0.05 - 1 .OO 35 f 2 5.90k0.02 -0.60 50+ 1 5.85 f0.04 -0.65 48 f 1 Per cent refers to weight YO surfactant. pK,". f 4 nm. Estimated pK," for monomeric neutral red; see the text for discussion. Table 2.pH-titration results from acridine and neutral red in aqueous micellar SDS solutions pKzbS from ref. (8). A,,, & 1/nm cy, = - 140 mV (v, = - 195 mV derivative wt% HIN+ (IN) pK:bS PK," PK," acridine 2 354 (356) 7.01 & 0.08 4.64 3.71 5 355 (356) 6.71 k0.16 4.34 3.41 10 354 (355) 6.M+ 0.10 4.07 3.14 neutral red 2 539 (454) 9.17 & 0.02 6.80 5.87 5 540 (453) 9.06 f 0.05 6.69 5.76 10 539 (453) 8.93 k 0.04 6.56 5.63 points refer to 0, 20, 40, 60 and 80 weight O/O 1,4-dioxane in the mixtures. The dielectric constants were obtained for the work of Critchfield et aL21 As in Parts 1 and 2 of this series,l'. l7 the ApKg values were derived from the ApK," values with the aid of eqn (8) and by assuming that the myH+ value can be approximated by the mean medium effect on HCl, my+, in the same 1,4-dioxane-water mixture.Table 1 contains the results of the pH-titrations carried out with acridine and neutral red in pure water and aqueous solutions of non-ionic micelles. The values for the effective interfacial dielectric constants, ceff, are also given in table 1. These were determined by comparing the ApK," values with the ApKg values of fig. 2 and 3. Note that this procedure involves the assumption that the log (yiN/yLIN+) term in eqn (9) is negligibly small. For neutral red in pure water, the spectrum as a function of pH did not exhibit a clear isosbestic point and the pK," was ill-defined with a value of 6.7k0.3. This type of behaviour has also been observed by other researchers.22* 23 Bartels22 proposed that there were two independent overlapping acid-base equilibria with pK," values of 7.38 and 5.89 to account for this observation.However, in the 1,4-dioxane-water mixtures and micellar solutions examined in the present study the pK, values were well-defined and all the spectra exhibited clear isosbestic points. Hence, there is no evidence, based on the pH titrations performed in these media, to support the proposition of two separate pK," values. The 'aberrant' pH-titration behaviour of neutral red in pure water is probably a result of a pH-dependent self-aggregation of neutral The pKr value of neutral red employed in the present study is an estimate obtained556 Azine Acid-Base Equilibria in Micelles by extrapolating the pK," and pKf, curves as a function of solvent dielectric constant to the dielectric constant of water (see fig.3). This procedure yielded a pK; value of 6.5. The pK," value of acridine has been reported as 5.60.27*28 The reason for the discrepancy between this result and the pKr value found in the present work is not known. The results of the pH titrations performed with acridine and neutral red in aqueous solutions of SDS micelles are given in table 2. Also included in table 2 are the pK," values for the azine derivatives in these micellar systems. Eqn (10) gives the relationship between the pKzbS value of an azine derivative residing within a charged micellar interfacial region and its pKZ value. Discussion Non-ionic Micellar Solutions As can be seen in table 1, the pKibS values of acridine and neutral red residing within the interfacial microenvironment of non-ionic micelles comprising surfactant molecules with poly(ethy1ene oxide) headgroups are in accord with an interfacial Eeff of 39 6.This ceff estimate is in excellent agreement with the estimates derived from the acid-base behaviour of a range of ' simple ' weak acids and bases, sulphonephthalein indicators and azo indicators.16. '9 29* 30 The pKihs values for neutral red located within the interfacial region of OG micelles suggest that the interfacial solvent properties of these micelles can be characterized by an eeff of 49 2. A similar interfacial Eeff estimate for these micelles was inferred from the acid-base behaviour of the sulphonephthalein indicators bromocresol green and bromothymol blue." Neutral red has a different prototropic group to that possessed by the sulphonephthaleins, and the ApK," values for the two kinds of molecules are different in both magnitude and sign. Therefore, the fact that a concordant Eeff estimate is obtained from the interfacial acid-base behaviour of these two types of molecules leads one to conclude that, for neutral red and the sulphonephthaleins, the assumptions that had to be made in order to estimate the interfacial Eeff value are valid.Hence, for neutral red and the sulphonephthaleins located within the interfacial microenvironment of OG micelles (a) a 1,4-dioxane-water mixture can mimic the effect of the interfacial solvent properties on the acid-base equilibrium ; (b) specific molecular interactions, if present, do not perturb the acid-base equilibrium; and ( c ) in the absence of added electrolyte the log (yiN/yLIN+) term is negligibly small.Ample coverage has already been given in Parts 1 and 2 of this seriesl6.l7 to the significance of the magnitude of the interfacial Eeff values of the two distinct types of non- ionic micelles, and therefore will not be discussed further here. Charged Micellar Solutions To date the most reliable experimental estimates of the electrostatic surface potential of an SDS micelle in aqueous solution without added electrolyte have been inferred from the acid-base behaviour of 1 -hexadecyl-5-hydroxyquinoline and 1 -hexadecyl-6-hy- droxyquinoline and the lipoidal 4-alkyl derivatives of 7-hydroxycoumarin and 7- a r n i n o c ~ u m a r i n . ~ ~ - ~ ~ These estimates cluster around - 140 mV and refer to reasonably dilute micellar SDS solutions.In order to ensure the complete micellar solubilization of the water-soluble azine derivatives, the SDS solutions employed in the present study were more concentrated. As a result of the comparatively large excluded micellar volume and micellar counter-ion dissociation, one would expect the micellar systems investigated in the present study to have a greater intrinsic ionic strength. Hence - 140 mV is considered to be an overestimate of the micellar y o in the present SDS solutions, and is considered to be a progressively poorer approximation to the actualC. J. Drummond, F. Grieser and T. W. Healy 557 micellar ry, value as the surfactant concentration is increased.Nevertheless, as no better experimental ry0 estimates are available at this time, - 140 mV was initially employed to gain some idea of the magnitude of the pKz values of the azine derivatives when they reside within the interfacial microenvironment of SDS micelles. For acridine and neutral red, the yo = - 140 mV SDS pK," values contained in table 2 are not the same as the Brij-35 pK,DbS values in table 1. Fernandez and F r o m h e r ~ , ~ ~ employing a dual acid-base indicator technique, deduced that the interfacial micro- environment of non-ionic micelles comprising surfactant molecules with poly(ethy1ene oxide) headgroups has equivalent non-electrostatic interfacial solvent properties to those of SDS micelles. Why, then, are the Brij-35 pKibS values of acridine and neutral red not equivalent to their ry, = - 140 mV SDS pKz values? Several possibilities can be con sidered.(a) Under-estimate of the surface potential. If a yo value of the order of - 195 mV is used (instead of - 140 mV) to calculate the pK," values, values closer to the pKibs values found in micellar Brij-35 solutions are obtained. This can be seen by comparing the results contained in table 1 within those in table 2. A surface potential of ca. - 195 mV is the theoretical upper limit, based on the unmodified non-linearized Poisson- Boltzmann equation, for a spherical SDS micelles with 64 charged surface sites in aqueous solution with no added ele~trolyte.~~ However, spectroscopic probe, electrophoretic mobility, conductivity and ion-selective electrode studies all indicate that a certain degree of counter-ion binding to the micelle occurs,32 hence a potential lower than - 195 mV is expected.Furthermore, the more reliable determinations of the surface potential of SDS micelles using spectroscopic probes, as already indicated, yield ry0 values of < - 140 mV. Therefore the deviations of the pKibS values cannot be attributed to a poor choice of ry,. A reanalysis of the results of several other r e s e a r c h e r ~ , ~ ~ ~ ' ~ who carried out pH titrations in micellar solutions with indicators of similar prototropic moieties to those possessed by acridine and neutral red, provides further support for the above contention. This reanalysis is summarized in table 3. It is worthy of note that for 9- heptadecylacridine and stearylchinin there is still a large disparity between the Triton X- 100 pKibs value and the SDS pKz value calculated with ry, = - 195 mV.From this observation it is clear that if the interfacial ceff of SDS and Triton X-100 micelles are roughly equivalent then there must be, in addition to the micellar ry, and low interfacial E ~ ~ ~ , some other factor which influences the interfacial acid-base equilibria of these types of molecules in SDS micelles. (b) Non-equivalence of the interfacial ceff. Recent work suggests that, contrary to the findings of Fernandez and F r o m h e r ~ , ~ ~ the interfacial ceff of SDS micelles and non-ionic micelles comprising surfactant molecules with poly(ethy1ene oxide) headgroups may not be equivalent. 29* 347 35 All previous estimates of the ceff of the interfacial region of SDS micelles do, however, fall within the range 16-67.35 If an ceff value of ca.60 is taken as the interfacial E~~~ value of the SDS micelles then the difference between the Brij-35 pKibS and the ry0 = - 140 mV SDS pK," values for acridine can be rationalized. However, the wo = - 140 mV SDS pK," values for neutral red are greater in magnitude than the pK," value and thus cannot be rationalized in this manner. This is also the case for nile blue and atebrin, table 3. (c) Interfacial 'salt-eflect'. There is only a small salt-effect on the aqueous acid-base equilibrium of both acridine and neutral red.36 Therefore, one would not expect any interfacial ' salt-effect ' to significantly affect the acid-base equilibria of the azine derivatives located within the interfacial region of SDS micelles.( d ) Specific molecular interactions. The most reasonable explanation for the marked difference between the Brij-35 pKzbS values of the azine derivatives and their w0 = - 140 mV SDS pKz values is that there is some form of specific molecular interaction between the cationic protonated nitrogen moieties and the anionic sulphate headgroups.558 Azine Acid-Base Equilibria in Micelles Table 3. Re-analysis of literature pH-titration data for molecules related to the azine derivatives SDS triton X-100 yo = - 140 mV yo = - 195 mV molecule pK," pKZbS PK: PK," pKZbS ref. 13 acridine orange 10.4 12.4 10.0 9.1 13 nile blue 9.9" 12.2" 10.5b 9.9" 13 ate brin 7.95 10.6 8.2 7.3 9-heptadecyl acridine 5.50 7.70 5.33 4.40 2.50 14 ~tearylchinin~ 4.00 6.20 3.83 2.90 1.80 14 __ - - 8.50 10.60 8.23 7.30 6.00 a mA+] maintained at 0.075 mol dmP3 by the addition of NaCI.'yo gauged from ref. (30) and (31). Poisson-Boltzmann equation [see ref. (32)]. atoms. yo = - 100 mV; experimental yo = - 138 mV; theoretical 'yo from non-linearized Molecule has two protonatable ring nitrogen Stabilization of the protonated form of an azine derivative could conceivably occur by the formation of either an ion-pair or a hydrogen bond with a sulphate headgroup. In connection with this concept of specific molecular interaction between the anionic sulphate headgroups and the cationic nitrogen centres of the protonated azine derivatives, it is worthwhile mentioning some work of Yamashita et aZ.379 38 on the effect of SDS micelles on the kinetics of the hydrolysis of primary and secondary amines.The detailed mechanism for the hydrolysis of primary and secondary amines is described in eqn (14): RR'NH2+ + OH- RR'-NH2+. . . OH- $ RR'NH + H 2 0 (14) In pure water the rate-determining step for the hydrolysis reaction is the diffusion- controlled process (I) + (11). However, in aqueous micellar SDS solution the rate- determining step is the intramolecular proton-transfer process (11) + (111). This finding, as noted by Yamashita et aZ., is consistent with the cationic amino group being stabilized through the formation of an ion-pair with an anionic sulphate headgroup. The results of table 3, in conjunction with those of tables 2 and 1, suggest that most acid-base equilibria which involve the protonation of a heterocyclic nitrogen group within the interfacial microenvironment of a SDS micelle are affected by specific molecular interactions between the anionic sulphate headgroups and the cationic protonated nitrogen moieties.Hong and Junge2 have measured the pK:bs values of neutral red residing within the interfacial region of thylakoid membranes under different external solution conditions. Interestingly, when sufficient electrolyte had been added to screen the surface charge density to a level where the yo value was extremely low, a pKEbS value of 6.6 was obtained. Hence, the pK," value of neutral red residing within this interfacial microenvironment is approximately 6.6 and comparable in magnitude to the pK," value calculated for neutral red in SDS micelles utilising a yo of - 140 mV.Estimates of the interfacial .zePf of phospholipid vesicles and membranes at room temperature vary from 4 to 58.34-39-43 Therefore, one expects the .zeff of the interfacial microenvironment of thylakoid membranes to be low. Thus, it is inferred that specific molecular interaction also occurs between interfacial phosphate groups and the cationic protonated nitrogen moiety of neutral red. Indeed within the confines of a lipid-water interfacial microenvironment, (1) (11) (IWC. J. Drummond, F. Grieser and T. W. Healy 559 the cationic conjugate acid forms of neutral red, acridine and acridine-related indicators may ion-pair with a wide variety of negatively charged groups.44 Conclusions The difference between the pK," value of an azine derivative and its pK:bs value when located within the interfacial region of a non-ionic Brij-35 or OG micelle has been solely attributed to the low interfacial e,,,.From the acid-base behaviour of acridine and neutral red, it has been inferred that the interfacial microenvironments of Brij-35 and OG micelles are characterized by eeff values of 39 6 and 49 2, respectively. Three major factors are believed to be responsible for the difference between the pKzbs value of an azine derivative residing within the interfacial microenvironment of an SDS micelle and its pK," value. These factors are the low interfacial E , ~ ~ , the electrostatic surface potential, and some form of specific molecular interaction between the cationic protonated moieties of the azine derivatives and the anionic sulphate headgroups.This work was supported by the Australian Research Grants Scheme. References 1 W. Junge, G. Schonknecht and V. Forster, Biochim. Biophys. Acta, 1986, 852, 93. 2 Y. Q. Hong and W. Junge, Biochim. Biophys. Acta, 1983, 722, 197. 3 A. B. Hope and A. Morland, Aust. J. Plant Physiol., 1979, 6, 289. 4 P. P. M. Schnetkamp, J. Membrane Biol., 1985, 88, 249. 5 P. P. M. Schnetkamp, J. Membrane Biol., 1985, 88, 263. 6 C. E. Williamson and A. H. Corwin, J. Colloid Interface Sci., 1972, 38, 567. 7 M. Gutman, D. Huppert, E. Pines and E. Nachliel, Biochim. Biophys. Acta, 1981, 642, 15. 8 M. Gutman, E. Nachliel, E. Gershon and R. Giniger, Eur. J. Biochem., 1983, 134, 63.9 P. P. M. Schnetkamp, U. B. Kaupp and W. Junge, Biochim. Biophys. Acta, 1981, 642, 213. 10 C. E. Williamson and A. H. Corwin, J. Colloid Interface Sci., 1972, 38, 577. I1 E. Baumgartner, R. Fernandez-Prini and D. Turyn, J. Chem. SOC., Faraday Trans. I , 1974, 70, 1518. 12 T. Wolff, Ber. Bunsenges. Phys. Chem., 1981, 85, 145. 13 A. D. James, B. H. Robinson and N. C. White, J. Colloid Interface Sci., 1977, 59, 328. 14 A. Haase, Doctoral Dissertation (Justus Liebig Universitat, Giessen, 1980). 15 G. Ackermann, L. Sommer and W. I. Stephen, Pure Appl. Chem., 1985, 57, 845. 16 C. J. Drummond, F. Grieser and T. W. Healy, J. Chem. Soc., Faraday Trans. I , 1989, 85, 521. 17 C. J. Drummond, F. Grieser and T. W. Healy, J. Chem. SOC., Faraday Trans. 1, 1989, 85, 537. 18 H.S. Harned and B. B. Owen, The Physical Chemistry of Electrolyte Solutions (Reinhold, New York, 19 L. G. Van Uitert and C. G. Haas, J. Am. Chem. SOC., 1953, 75, 541. 20 C. J. Drummond, Doctoral Dissertation (The University of Melbourne, 1987). 21 F. W. Critchfield, J. A. Gibson and J. L. Hall, J. Am. Chem. Soc., 1953, 75, 1991. 22 P. Bartels, 2. Phys. Chem. Neue Folge, 1956, 9, 74. 23 W. Junge, W. Auslander, A. J. McGeer and T. Runge, Biochim. Biophys. Acta, 1979, 546, 121. 24 P. Bartels, Z. Phys. Chem. Neue Folge, 1956, 9, 95. 25 N. V. Rao and K. L. Narayana, Indian J. Chem., 1982, 21, 995. 26 N. V. Rao and K. L. Narayana, Indian J. Chem., 1983, 22, 887. 27 A. Albert and R. Goldacre, J. Chem. Soc., 1946, 706. 28 S. G. Schulman and A. C. Capomacchia, J. Am. Chem. SOC., 1973, 95, 2763. 29 C. J. Drummond, F. Grieser and T. W. Healy, J. Chem. Soc., Faraday Trans. I , 1989, 85, 561. 30 M. S. Fernandez and P. Fromherz, J. Phys. Chem., 1977, 81, 1755. 31 J. Frahm, S. Diekmann and A. Haase, Ber. Bunsenges. Phys. Chem., 1980, 84, 566. 32 G. V. Hartland, F. Grieser and L. R. White, J. Chem. Soc., Faraday Trans. I , 1987, 83, 591. 33 C. J. Drummond and F. Grieser, Photochem. Photobiol., 1987, 45, 19. 34 C. J. Drummond, F. Grieser and T. W. Healy, J. Phys. Chem., 1988, 92, 2604. 35 F. Grieser and C. J. Drummond, J. Phys. Chem., 1988, 92, 5580. 36 E. Banyai, in Indicators, ed. E. Bishop (Pergamon Press, Oxford, 1972), chap. 3. 37 T. Yamashita, H. Yano, S. Harada and T. Yasunaga, J. Phys. Chem., 1984, 88, 2671. 3rd edn, 1958).560 Azine Acid-Base Equilibria in Micelles 38 T. Yamashita, M. Sumino, H. Yano, S. Harada and T. Yasunaga, Bull. Chem. SOC. Jpn, 1984, 57, 39 P. I. Lelkes and I. R. Miller, J . Membrane Biol., 1980, 52, 1 . 40 K. A. Zachariasse, N. Van Phuc and B. Kozankiewicz, J. Phys. Chem., 1981, 85, 2676. 41 S. Lukac, J . Am. Chem. Soc., 1984, 106, 4386. 42 Y. Kimura and A. Ikegami, J . Membrane Biol., 1985, 85, 225. 43 C. J. Drummond and F. Grieser, Lmgmuir, 1987, 3, 855. 44 P. Dell’Antone, R. Colonna and G. F. Azzone, Eur. J . Biochem., 1972, 24, 566. 2352. Paper 7/00 1 12F ; Received 2 1 st December, 1987

ISSN:0300-9599    

DOI:10.1039/F19898500551

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. I, 1989, 85(3), 561-578 Acid-Base Equilibria in Aqueous Micellar Solutions Part 4.-Azo Indicators Calum J. Drummond,*t Franz Grieser and Thomas W. Healy Department of Physical Chemistry, The University of Melbourne, Parkville, Victoria, 3052, Australia. The acid-base equilibria of three azo indicators, viz. dimethyl yellow, methyl orange and methyl red, have been examined in organic solvent-water mixtures, in aqueous non-ionic micellar solutions of Brij-35 and n-octyl B-D-glucoside and in aqueous charged micellar solutions of n-dodecyl- trimethylammonium bromide and/or sodium dodecyl sulphate. The factors primarily responsible for the disparity between the apparent pK, value of an azo indicator residing within the interfacial microenvironment of a micelle and its pK, value in pure water have been isolated.For the non-ionic micellar systems the disparity can be attributed to the different intrinsic solvent characteristics of the two solvating media. For the charged micellar systems the electrostatic micellar surface potential is an additional contributing factor. Since Hartley and Roe1 first tried to utilize an azo indicator to ascertain the mean field electrostatic potential at the surface of a charged micelle, many researchers have attempted to exploit the acid-base properties of azo indicators to gain information about the solvent and/or electrostatic characteristics of lipid-water interfaces.2-10 As yet, however, no one has unequivocally established which factors contribute to the magnitude of the apparent pKa value of an azo indicator residing within a lipid-water interfacial microenvironment.The purpose of the present work was to ascertain quantitatively the factors primarily responsible for the difference between the pKa value of an azo indicator in pure water and its apparent pK, when it is located within the interfacial microenvironment of a micelle. The acid-base equilibria of three azo indicators, viz. dimethyl yellow, methyl orange and methyl red, were examined in organic solvent-water mixtures and in aqueous micellar solutions of Brij-35, n-octyl /h-glucoside (OG), n-dodecyl- trimethylammonium bromide (DTAB) and/or sodium dodecyl sulphate (SDS). In general, for each particular azo indicator the charged micellar systems were chosen such that the charge on the micelles was not of the same sign as the overall charge of either the conjugate acid or conjugate base form of the indicator.This was done in order to ensure that both forms of the water-soluble azo indicators partitioned quantitatively into the micellar phase. The visible absorption spectra of the azo indicators were studied because they can provide information about the solvent properties of the interfacial microenvironments of the various micelles. It is generally a~cepted'l-'~ that the aqueous acid-base equilibria of dimethyl yellow and methyl orange involve the species depicted in fig. 1. Fig. 2 shows the assignment of the species involved in the aqueous acid-base equilibria of methyl red which was initially proposed by Ross and Warwick.13 A consensus has yet to be reached on which species t Present address: CSIRO, Division of Chemicals and Polymers, G.P.O.Box 4331, Melbourne, Victoria, 3001, Australia. 56 1562 Azo Acid-Base Equilibria in Micelles .*+\I+.+ L Fig. 1. Species considered to be involved in the aqueous acid-base equilibrium of dimethyl yellow (R = H) and methyl orange (R = SO,-). are involved in the aqueous acid-base equilibria of methyl red.13-20 Nevertheless, the spectroscopic properties of methyl red in pure aqueous solution as a function of pH do seem to support the assignment given by Ross and Warwick (vide infra). Experimental Dimethyl yellow and methyl orange were guaranteed reagents from Tokyo Kasei Kogyo Co. Methyl red was an analytical grade reagent supplied by B.D.H. Chemicals Pty.Ltd. The purity of these azo indicators was verified by using IUPAC recommended tests.21 The quality and source of each of the surfactants, inorganic reagents, organic solvents and water employed in the present investigation has already been given in Parts 1 and 2 of this 23 An identical experimental arrangement to the one described in Part 1 of this series22 was used for the pH titrations. All pH titrations were carried out at 25 "C. The hydrogen-ion activity in the various media was determined as detailed in Parts 2 and 3 of this 24 In order to determine the various acid-base equilibrium constants, the acid-base equilibrium of an azo indicator was treated as being of the form: where HI"") represents the conjugate acid form of the azo indicator which has an overall charge of z, H' represents the hydrogen ion and IN("-I) the conjugate base form of the indicator which has an overall charge of z - 1.The thermodynamic equilibrium constant for equilibrium (1) is given byC. J. Drummond, F. Grieser and T. W. Healy 563 0 5 + 3: + Q564 Azo Acid-Base Equilibria in Micelles where aH+ is the hydrogen ion activity, [IN(z-l)] and [HIN")] are the concentrations of the conjugate base and conjugate acid forms of the azo indicator, and yIN and yHrN are the activity coefficients of the conjugate base and conjugate acid forms, referred to the particular solvating medium at infinite dilution. Note that in the remainder of this work the quantity -log aH+ is referred to as the pH irrespective of the particular solvating medium.The visible absorption spectra of the azo indicators as a function of pH were utilized to obtain their p& values. Use was made of the relationship with A - A , , a = where A is the absorbance at the long-wavelength band maximum of the conjugate acid form of the azo indicator at the particular pH being examined, and A , , and A,,, are the absorbances at this same wavelength when the pH is such that there is a 100% population of the conjugate base form and conjugate acid form, respectively. In a number of the systems investigated the pH required to obtain a 100 % population of the conjugate acid form of the azo indicator is extremely low (in some cases negative). In these systems the absorbance value of 100 O/O population of the conjugate acid form was estimated.Representative examples of the visible absorption spectra as a function of pH are given later. As yet the values of the activity coefficients of species residing within the interfacial microenvironment of micelles have not been quantified. Also, it is not clear how one approximates the activity coefficients of complex resonance hybrid ions such as those involved in the acid-base equilibria of the azo indicator^.^^ Consequently the activity coefficient term in eqn (2) was neglected in the calculation of the p K values. Results and Discussion Visible Absorption Spectra Dimethyl yellow and methyl orange. The visible absorption spectrum of dimethyl yellow as a function of the bulk aqueous pH in pure water, a 5 wt YO Brij-35 solution and a 10 wt O/O SDS solution is shown in fig.3. The spectral profile of dimethyl yellow in the 5 wt YO Brij-35 solution is also representative of that obtained in a 10 wt YO Brij-35 solution while the spectral profile of dimethyl yellow in the 10 wt O/O SDS solution is representative of those obtained in 2 and 5 wt O/O SDS solutions. The positions of the long-wavelength absorption band maximum, L,,,, for the protonated and deprotonated forms of dimethyl yellow in each of the media investigated are given in table 1. Fig. 4 shows the visible absorption spectrum of methyl orange as a function of the bulk aqueous pH in pure water and a 5 wt YO DTAB4 mol dmP3 NaBr solution. The spectral profile of methyl orange in the 5 wt YO DTAB-4 mol dmP3 NaBr solution is representative of its spectral profile in all the aqueous micellar solutions examined.Table 2 contains the Amax values for the protonated and deprotonated forms of methyl orange in the media studied. As shown in fig. 3 and reported in table 1, micellar solubilization causes the A,,, of the protonated form of dimethyl yellow to undergo a small bathochromic shift. This can be ascribed to the effect that the different solvent properties of the micellar interfacial microenvironment has on the energy of the x-n* electronic transition. In contrast to this, the A,,, values for the protonated form of methyl orange, fig. 4 and table 2, indicateC. J . Drummond, F. Grieser and T. W. Healy 565 0.08 0.04 0 0.16 B 0.08 4 0 1 .o 0.5 400 500 600 Alnm Fig. 3. Visible absorption spectra of dimethyl yellow (DMY) as a function of pH in (a) pure water, [DMY] = 2.3 x lop6 mol dm-3; (6) a 5 wt % Brij-35 solution, [DMY] = 8.3 x mol dmp3 (the value in parentheses denotes the degree of protonation of DMY at the lowest pH investigated); and (c) a 10 wt YO SDS solution, [DMY] = 4.3 x mol dm-3.Table 1. pH-Titration results for dimethyl yellow in pure water and aqueous micellar solutions with the corresponding E , ~ ~ estimates mediuma IN HINb p KaObS APK,. Eeff water 445 509 3.21 +O.OY 0 - 5% Brij-35 41 5 516 (0.55) 1.01 kO.05 - 2.20 38k 1 10% Brij-35 414 51 6 (0.42) 0.85 k 0.02 - 2.36 36k 1 2% SDS 41 5 517 4.73 f 0.07 -0.85 63f I d 5% SDS 412 517 4.61 f0.05 -0.97 61 I d 10% SDS 410 516 4.38 k 0.03 - 1.20 57k I d a % refers to wt YO surfactant. Values in parentheses denote the degree of protonation of dimethyl yellow in the systems where the absorption spectrum of the fully protonated form was not obtained.p e . See text.566 Azo Acid-Base Equilibria in Micelles 0.4 0. I s 8 1 0 0.4 c I I I ' . I I I 400 500 600 Unm Fig. 4. Visible absorption spectra of methyl orange (MO) as a function of pH in (a) pure water, [MO] = 1.3 x lop5 mol dm-3; and (6) a 5 wt YO DTAB-4 mol dm-3 NaBr solution, [MO] = 1.5 x mol dm-3 (the value in parentheses denotes the degree of protonation of MO at the lowest pH investigated). Table 2. pH-Titration results for methyl orange in pure water and aqueous micellar solutions with the corresponding E , ~ ~ estimates medium" IN HINb PK,"~" APK," Eeff ~~~ ~ - water 462 507 3.42 & 0.02" 0 5 % Brij-35 423 507 (0.86) 1.83 f 0.02 - 1.59 495 I d 10% Brij-35 423 507 (0.78) 1.63 & 0.01 - 1.79 46_+ I d 5 % OG 420 507 (0.74) 1.90 5 0.03 - 1.52 51+1 10% OG 419 507 (0.53) 1.63 & 0.0 1 - 1.79 46+ I 5 YO DTAB4 mol 423 510 (0.67) 0.88f0.11 - 2.24 38f2 dm-3 NaBr a O/O refers to wt YO surfactant.Values in parentheses denote the degree of protonation of methyl orange in the systems where the absorption spectrum of the fully protonated form was not obtained. p e . See text.C. J. Drummond, F. Grieser and T. W. Healy 567 that there is little or no difference between the influence of pure water and that of the micellar interfacial microenvironment on the energy of this n-n* electronic transition. For the deprotonated form of both dimethyl yellow and methyl orange micellar solubilization results in a large hypsochromic shift in the A, tables 1 and 2, and a change in the overall shape of the absorption band, fig.3 and 4. have previously established that the long-wavelength absorption band of the deprotonated form of p(dimethy1amino)azobenzenes like methyl orange and dimethyl yellow in aqueous solutions consists of two overlapping bands. Moreover, it has been suggested by Reeves et a1.26 and a number of earlier ~ o r k e r s ~ ~ ~ ~ ~ that the two overlapping bands belong to the n-n* transitions of the species involved in the equilibrium Reeves et with the species which is hydrogen-bonded to water being the long-wavelength component and the other species being the short-wavelength component to the composite absorption band.Reeves et aZ.26 contended that upon micellar solubilization, the above equilibrium is shifted to the left with an attendant hypsochromic shift of Amax because of the change in the relative intensities of the two component absorption bands. We see no reason not to invoke the same explanation to account for our spectral findings for dimethyl yellow and methyl orange. Note also that inherent in this explanation is the concept that the micellar interfacial microenvironment possesses low water activity. Interestingly, for the deprotonated form of dimethyl yellow, the absorption band is more highly skewed towards the short wavelength component in micellar Brij-35 solution than it is in micellar SDS solution, fig. 3. It is tempting to attribute this to the interfacial microenvironment of the SDS micelles being more ‘ aqueous-like ’ in nature than that of the Brij-35 micelles. It is difficult to reconcile this deduction with the findings of Fernandez and Fromher~.~’ In their work, Fernandez and Fromherz examined the acid-base behaviour of a lipoidal 7-hydroxycoumarin and a lipoidal 7-aminocoumarin in dioxane-water mixtures and aqueous micellar solutions.With the information gained from these systems they built an excellent case in support of the equivalent nature of the mean interfacial solvent properties of SDS micelles and non-ionic micelles such as Brij-35 micelles. Nevertheless, it should also be mentioned that there has been several other r e p o r t ~ ~ ’ - ~ ~ of the interfacial microenvironment of SDS micelles being more ‘ aqueous-like ’ than those of non-ionic and cationic micelles.Intriguingly, the interfacial microenvironment of OG micelles is known to be more ‘aqueous-like ’ than the interfacial microenvironments of Brij-35 and DTAB micelles (vide infra) yet there is no discernible difference in the shape of the absorbance band of the deprotonated form of methyl orange in these three systems. In connection with this, however, it can be argued that the solvent properties of the interfacial microenvironments of these three types of micelles, as sensed by methyl orange, are not significantly different (cf. the ceff values given in table 2). Methyl red. As depicted in fig. 5 and 6, the spectral behaviour of methyl red as a function of pH is markedly dependent upon the wt % of 1,4-dioxane in the 1,4-dioxane-water mixtures.The unique spectral behaviour of methyl red as a function of pH in each of the 1,4-dioxane-water mixtures is considered to be a result of the unique effect that each 1,4-dioxane-water mixture has on the analogous equilibrium to equation (4) and the energy associated with the z-z* electronic transitions of each of the different acid-base forms of methyl red. In addition to these contributing factors some of the568 Azo Acid-Base Equilibria in Micelles 0. 0. 400 500 600 400 500 600 Alnm Fig. 5. Visible absorption spectra of methyl red (MR) as a function of pH in (a) pure water, [MR] = 8.6 x lop6 mol dm-3; (a') pure water, [MR] = 1.1 x mol dm-3; and (b) a 20 wt YO 1,4-dioxane-water mixture, [MR] = 1.0 x mol dm-3.The values in parentheses denote the degree of protonation of MR at the lowest pH investigated. spectral differences are considered to be due to the fact that the mode of protonation of methyl red is not the same in all of the 1,4-dioxane-water mixtures (vide infra). Methyl red in micellar SDS solution has spectral properties which are intermediate between those observed in the 20% (e = 61.9) and 40% ( E = 44.5) 1,4-dioxane-water mixtures, fig. 5-7. The spectral properties of methyl red in micellar Brij-35 and DTAB solutions are intermediate between those observed in the 40% ( E = 44.5) and 60% ( E = 27.2) 1,4-dioxane-water mixtures, fig. 6 and 7. Thus the spectral behaviour of methyl red, like that of dimethyl yellow, seems to indicate that the interfacial microenvironment of SDS micelles is more 'aqueous-like' than that of Brij-35 micelles.Acid-Base Equilibria In the forthcoming discussion we make use of a number of relationships, uiz. PK: = PC-lOgmY, ( 5 ) Y t X p K = pKi, + log - Y',," where p c denotes the p& of an azo indicator in an organic solvent-water mixture; the mean medium effect on HCI; ytN and yLIN the activity coefficients of the deprotonated and protonated forms of an azo indicator, respectively, referred to theC. J . Drummond, F. Grieser and T. W. Healy 569 0.2 0 400 500 600400 500 600 Fig. 6. Visible absorption spectra of methyl red, [MR] = 1.0 x mol dm-3, as a function of pH in 1,4-dioxane-water mixtures. (a) 40 wt YO dioxane; (6) 60 wt % dioxane; and (c) 80 wt YO dioxane. The values in parentheses denote the degree of protonation of MR at the lowest pH investigated.interfacial micellar phase at infinite dilution; pKbs the apparent p& of an azo indicator in a particular micellar solution ; F the Faraday constant; R the universal gas constant ; T the absolute temperature and ry, the electrostatic surface potential of a micelle. The implicit assumptions involved in each of these relationships has been thoroughly discussed in Part 1 of this series22 and therefore will not be repeated herein. For ease of discussion we also define where p c denotes the p& of an azo indicator in pure water. Dimethyl Yellow and Methyl Orange Organic solvent-water mixtures. The p q value of dimethyl yellow has previously been reported as 3.2536,37 and 3.35 f 0.05,1° and that of methyl orange as 3.45,36 3.41 kO.04l5570 0.2 0 0.2 0 Azo Acid-Base Equilibria in Micelles 400 500 600400 500 600 Alnm Fig. 7.Visible absorption spectra of methyl red, [MR] = 1.1 x mol dm-3, as a function of pH in (a) a 10 wt % SDS solution and (b) a 5 wt % DTAB4 mol dm-3 NaBr solution. The value in parentheses denotes the degree of protonation of MR at the lowest pH investigated. and 3.38 & 0.02.38 Hence the p c values of dimethyl yellow, table 1, and methyl orange, table 2, found in this work are in good accord with the literature values. Fig. 8 depicts the ApFa behaviour of dimethyl yellow in ethanol-water mixtures and methyl orange in methanol-water mixtures as a function of the solvent dielectric constant. The p e data for dimethyl yellow and methyl orange in these media (not shown) were obtained from the work of Gutbezahl and G r ~ n w a l d ~ ~ and de Ligny et al.,38 respectively. These p e values were converted into p& values with the aid of eqn ( 5 ) and the relevant my+ values, which were interpolated from fig.4 in Part 1 of this series.22 The dielectric constants of the methanol-water and ethanol-water mixtures were interpolated from the work of Akerl~f.~’ It is apparent from fig. 8 that the ApPa behaviour of dimethyl yellow as a function of the dielectric constant of ethanol-water mixtures is equivalent to the ApPa behaviour of methyl orange as a function of the dielectric cons tan t of met hanol-wa ter mixtures. Micellar solutions. The pH-titration results for dimethyl yellow in the aqueous micellar solutions investigated are summarized in table 1.Also included in table 1 are estimates of the effective interfacial dielectric constants, e,,,, of the micelles. These e,,, estimates were obtained by comparing the ApC values with the ApFa values as a function of solvent dielectric constant, fig. 8. It is emphasized that these Eeff estimates, and the other ceff estimates in this work (vide infra), are only valid if both the protonated and deprotonated forms of the indicator have partitioned quantitatively into the interfacial micellar phase and if the micellar acid-base equilibria are not influenced by specific molecular interactions or interfacial ‘ salt-effects.’ As indicated in table 1, the p K values of dimethyl yellow in micellar Brij-35 solutions yield an interfacial Eeff value of 37 2.This estimate is in agreement with previous EeffC. J . Drummond, F. Grieser and T. W. Healy 57 1 - 1.0 - a - 2.0 40 60 80 dielectric constant Fig. 8. Dimethyl yellow (0) and methyl orange (0) ApPa values as a function of the dielectric constant of ethanol-water and methanol-water mixtures, respectively. estimates gauged from the acid-base behaviour of many other types of indicators residing within the interfacial microenvironment of similar non-ionic micelles. 22-247 29 Hence, it is deduced that the disparity between the p c value of dimethyl yellow and its p g value in micellar Brij-35 solutions is primarily due to the interfacial microenvironment of the Brij-35 micelles possessing a low ceff. The p g values in the micellar SDS solutions were ascertained by using eqn (7) and a yo value of - 140 mV for the SDS micelles.This choice of - 140 mV for the yo value has been discussed in Part 3 of this series.24 The calculated p c values for dimethyl yellow in the micellar SDS solutions are less than the p c value but greater than the p K values in the micellar Brij-35 solutions. The spectral properties of dimethyl yellow are consistent with it experiencing a more ' aqueous-like ' interfacial microenvironment in SDS micelles than in Brij-35 micelles (uide supra). Therefore it is not clear whether the inequality of the p c values of dimethyl yellow in Brij-35 and SDS micellar solutions stems solely from a difference in the interfacial ceff values, as given in table 1, or whether other factors also contribute.Table 2 contains the results of the pH titrations performed with methyl orange in aqueous micellar solutions. The pK," value of methyl orange in the micellar DTAB-4 mol dm-3 NaBr solution was ascertained by employing eqn (7) and a y o of + 18 mV.33,40 ceff estimates were determined by comparing the ApK values with the ApKg behaviour of methyl orange as a function of solvent dielectric constant, fig. 8. The interfacial ceff estimates for the Brij-35 micelles derived from the ApK values of methyl orange are larger than those obtained with dimethyl yellow, methyl red (uide infra) and other acid-base indicator^.'^-^^ With the data presently available one can572 Azo Acid-Base Equilibria in Micelles Table 3. pH-titration results for the first protonation of methyl red in pure water and aqueous micellar solutions with the corresponding ceff estimates water 5% Brij-35 10% Brij-35 5% DTAB 5% D T A W mol 2% DTAB 10 Yo DTAB dm-3 NaBr 2% SDS 5% SDS 10% SDS 423 409 409 413 41 7 416 41 5 416 41 5 412 522 508 508 514 515 513 511 522 522 522 5.01 f 0.03b 6.20 f 0.03 6.30 f 0.01 4.27 & 0.09 4.36 + 0.02 4.58 0.02 5.87f0.12 7.06 +_O.12 7.40 + 0.20 7.58 f0.05 0 + 1.19 + 1.29 + 1.19 + 1.01 + 1 . 1 1 + 1.16 40+ 1 38-L 1 40-L 1 44+ 1 42+ 1 40+3 a Yo refers to wt% surfactant. p q ( 1 ) . See text. only speculate on the reasons for this difference. Since there is nothing ‘anornolous’ about the acid-base behaviour of dimethyl yellow and methyl red in micellar Brij-35 solutions, it is not thought to be a consequence of any kind of specific molecular interaction between the prototropic groups of methyl orange and micellar poly(ethy1ene oxide) headgroups.It may be the result of incomplete partitioning of the highly water soluble deprotonated form of methyl orange into the interfacial micellar phase. Alternatively, but far less likely, it may be the result of the extremely hydrophilic sulphonate moiety ‘anchoring’ methyl orange in a more ‘aqueous’ part of the interfacial poly(ethy1ene oxide) headgroup region of the Brij-35 micelles than is generally the case for the other acid-base indicators.*l The interfacial Eeff estimate of 48 f 3 determined from the p g values of methyl orange in the micellar OG solutions agrees with the interfacial ceff estimate for OG micelles derived from the acid-base behaviour of bromocresol green, bromothymol blue and neutral red,23v24 Similarly, the interfacial Eeff estimate for the DTAB micelles, table 2, is in accord with Eeff estimates based on the acid-base behaviour of molecules with different prototropic groups to those of methyl orange22 (vide infra).It is concluded, therefore, that the difference between the p e value and the p c values of methyl orange in the aqueous OG and DTAB micellar solutions can primarily be attributed to the low interfacial ceff of the micelles. Methyl Red 1,4-Dioxane-water mixtures. The p K ( 1) value of methyl red has previously been reported as 5.0,36 4.920 and 4.95,” and the p e ( 2 ) value as 2.336 and 2.4.20 Hence the p e ( 1 ) and pK(2) values found in this work (tables 3 and 4) are in reasonable agreement with literature values.Fig. 9 shows the A p e and ApKi behaviour of methyl red as a function of the dielectric constant of 1,4-dioxane-water mixtures. The data points refer to 0, 20, 40, 60 and 80 wt O/O 1,4-dioxane-water mixtures. The dielectric constants of these mixtures were acquired from the work of Critchfield et aZ.42 The acid-base behaviour of methyl red in 1,4-dioxane-water mixtures is not straightforward and hitherto has not been properly characterized. Some insight into theC. J . Drummond, F. Grieser and T. W. Healy 573 Table 4. pH-titration results for the second protonation of methyl red in pure water and aqueous micellar solutions with the corresponding E , ~ ~ estimates ~~ - water 5 13 (0.90) 2.09 & 0.09" 0 5% Brij-35 513 (0.55) 1.08 & 0.09 - 1.01 31 + 2 10% Brij-35 514 (0.53) 1.10 f 0.06 - 0.99 31 +_2 5% DTAB-4 mol 514 (0.55) 0.85 f 0.24 - 0.94 32+5 dm-3 NaBr 2% SDS 517 4.36 & 0.06 -0.10 67 & 7d 5% SDS 517 4.22 & 0.10 -0.24 55 f 9d 10% SDS 517 4.02 + 0.07 -0.44 4 4 + 3 d YO refers to wt % surfactant.Values in parentheses denote the degree of second protonation of methyl red in the systems where the absorption spectrum of the fully protonated form was not obtained. pKZ(2). See text. dielectric constant Fig. 9. Methyl red ApK," and ApKi values as a function of the dielectric constant 1,4-dioxane-water mixtures: 0, A p c ( 1 ) ; 0, ApPa(l); a, A p e ( 2 ) ; H, ApKi(2). of FAR I 20574 Azo Acid-Base Equilibria in Micelles acid-base behaviour of methyl red in 1,4-dioxane-water mixtures may be gained if one examines each pro totropic group in isolation.One would expect an isolated aromatic carboxylic acid group to have a p c value in pure water of ca. 4.2. This expectation is based on the p& value of benzoic acid in pure water.22 Similarly, based on dimethyl yellow, one would expect the p& value of an isolated azo group in pure water to be ca. 3.2. Methyl red, which possesses both an aromatic carboxylic acid group and an azo group, has p& values in pure water of 5.01 0.03 and 2.09k0.09 (tables 3 and 4). Clearly the prototropic groups in methyl red are not isolated. At high pH the carboxylic acid group of methyl red is deprotonated and thus negatively charged. The effect of a charged substituent on the p& value of a neighbouring acid-base group in a molecule may be formalized in the expression23.34~43-45 z , e2 '& = pc-2.303 Ddk,T where p c is the intrinsic pK, of the acid-base group, z , is the charge of the substituent, D is the effective dielectric constant separating the substituent and the acid-base group, d is the distance between the substituent and the acid-base group, and e, k , and Tare the electronic charge, the Boltzmann constant and the absolute temperature, respectively. On the basis of eqn (1 I), it is not unreasonable to infer that the p e value of the azo group in methyl red will be shifted to a greater value than it is in dimethyl yellow.Moreover, it is apparent, from a comparison of the effect of charged substituents on the p c values of other acid-base groups, that a value of 5.01 kO.03 is quite conceivable for the p c value of the azo group of methyl red.43 This p e assignment is given additional credence by the observation that the spectrum of the singly protonated form of methyl red in pure water closely resembles the spectrum of protonated dimethyl yellow, cJ: fig.3 and 5. Consequently, it is deduced that the p c value of 2.09k0.09 refers to the acid-base equilibrium of the carboxylic acid group of methyl red. Ross and Warwick13 have previously derived the same sequence of protonation for methyl red from an analysis of spectral data. The acid-base equilibria scheme proposed by Ross and Warwick is shown in fig. 2. Three factors combine to make a quantitative interpretation of the acid-base behaviour of methyl red in 1,4-dioxane-water mixtures in terms of eqn (1 1) extremely difficult, First, the fact that one of the species involved in each acid-base equilibrium of methyl red in pure water is a resonance hybrid of a hydrogen-bonded structure probably means that the p K value of the azo and carboxylic acid groups of methyl red in pure water are not 3.2 and 4.2, respectively.It is not clear how one quantitatively accounts for the effect that hydrogen bonding has on the pKf: values. Secondly, since the value of D in eqn (1 1) is not equivalent to the solvent dielectric ~ o n s t a n t , ' ~ - ~ ~ there is no way, at present, of assessing how the various 1,4-dioxane-water mixtures influence the value of D. Thirdly, in the hydrogen-bonded resonance hybrid which participates in the acid-base equilibrium scheme in pure water the positive charge is delocalized.Thus, for the pl(, relating to the second protonation of methyl red in pure water d in eqn (1 1) is not a well-defined quantity. In spite of the aforementioned problems associated with the quantitative use of eqn (1 l), its form can still be utilized to provide a qualitative guide to the acid-base behaviour of methyl red in 1,4-dioxane-water mixtures. For example, although the azo group of methyl red is protonated before the carboxylate group in pure water, it is possible, with the aid of eqn (1 l), to envisage a reversal i n this order of protonation for the two acid-base groups in some 1,4-dioxane-water mixtures.From the acid-base behaviour of benzoic acid22 and dimethyl yellow in organic solvent-water mixtures, it can be inferred that the pKf: value of the carboxylic acid group will progressively increase as the amount of 1,4-dioxane in the 1,4-dioxane-water mixtures is increased while theC. J . Drummond, F. Grieser and T. W. Healy 575 r C-OH II \c-OH Q Fig. 10. Species considered to be involved in the acid-base equilibria of methyl red in 40, 60 and 80 wt ?LO 1,4-dioxane-water mixtures and aqueous micellar Brij-35 and DTAB solutions. pKHf value of the azo group will progressively decrease. If in any 1,4-dioxane-water mixture the electrostatic component to the pK, value of the azo group, eqn (1 l), is not greater than the difference between the p e values of the carboxylic acid and azo groups then the carboxylate group will be protonated in preference to the azo group. It is considered that a reversal in the order of protonation of the two acid-base groups is principally responsible for the spectral properties of methyl red in pure water and in the 20 wt O h 1,4-dioxane-water mixture as a function of pH (fig.5) being markedly different to those of methyl red in the 40, 60 and 80 wt YO 1,4-dioxane-water mixtures (fig. 6). In other words, the contention is that in pure water and in the 20 wt % 1,4-dioxane-water mixture the acid-base equilibria scheme for methyl red is as shown in fig. 2 while in the 40,60 and 80 wt YO 1,4-dioxane-water mixtures it is as shown in fig. 10. In order to indicate that a reversal in the order of protonation of the carboxylate and azo groups is believed to occur, the region between the A p C and ApKa values in the 2WO wt O/O 1,4-dioxane-water mixtures is dashed in fig.9. 20-1576 Azo Acid-Base Equilibria in Micelles Micellar Solutions Tables 3 and 4 contain the pH-titration results for the first and second protonation of methyl red in the aqueous micellar solutions investigated, respectively. Eqn (7), in conjunction with the known yo values of the DTAB m i c e l l e ~ ~ ~ , ~ ” and a ry, of - 140 mV for the SDS m i c e l l e ~ , ~ ~ was utilized to determine the p c values in the charged micellar solutions. The Eeff estimates were obtained by comparing the A p g values with the appropriate ApK; curve of fig. 9. The Apc(1) and ApC(2) values of methyl red in micellar Brij-35 solutions yield similar E~~~ estimates.As well, the ApK(1) and ApK(2) values of methyl red in micellar DTAB solutions yield reasonably concordant Eeff estimates. The Eefl estimate for the Brij-35 micelles agrees with that derived from the acid-base behaviour of dimethyl yellow (table I), while the Eeff estimate for the DTAB micelles concurs with that determined with methyl orange (table 2). Therefore it is inferred that the primary cause of the difference between the p c values of methyl red and its p g values in micellar Brij-35 and DTAB solutions is the low Eeff which characterizes the solvent properties of the interfacial microenvironment of Brij-35 and DTAB micelles. The close correspondence between (a) the p e b s value of 4-octadecyloxy- 1 -naphthoic acid22 and the p e b S ( l ) value of methyl red in micellar Brij-35 solutions, ( b ) the pK!”” value of dimethyl yellow and the pebS(2) value of methyl red in micellar Brij-35 solution, and (c) the p e b S value of methyl orange and the pKhS(2) value of methyl red in micellar DTAB-4 mol dm-3 NaBr solution confirms that the acid-base equilibria scheme depicted in fig.10 is appropriate for methyl red residing within the interfacial microenvironment of Brij-35 and DTAB micelles. The increase in the p c b s ( l ) value of methyl red as the wt % of SDS is increased (table 3) indicates that at least one of the acid-base forms of methyl red has not partitioned completely into the micellar phase. In view of the fact that the high pH form of methyl red and the SDS micelles are both negatively charged this is hardly a surprising finding.Since the deprotonated form of methyl red is evidently not solubilized fully in the interfacial phase of the SDS micelles, it was considered pointless to analyse these pEbs( I) results. As was the case with dimethyl yellow, the spectral properties of methyl red in micellar SDS solutions suggest that the chromophore may reside in an interfacial micro- environment which has a lower Eeff than the dielectric constant of water but a greater qff than that of Brij-35 micelles (uide supra). The Eeff estimates derived from the ApK(2) values of methyl red in the micellar SDS solutions (table 4) are entirely consistent with this viewpoint. In addition, the Eeff estimates for the SDS micelles are not far removed from the Eeff estimates based on the acid-base behaviour of dimethyl yellow (table 1).In spite of all this consistency there is still some doubt as to whether or not the difference between the p c ( 2 ) values of methyl red in micellar Brij-35 solutions and those in micellar SDS solutions can be ascribed to a difference in the interfacial of these two types of micelles. As mentioned earlier, the uncertainty arises because the assertion that Brij-35 and SDS micelles possess a different interfacial E~~~ conflicts with the findings of Fernandez and Fromher~.~’ Since the mean field potential rapidly decays with distance from the plane of surface charge,46 it is considered that this difference in the E~~~ estimate cannot be ascribed to the prototropic moieties of methyl red and dimethyl yellow residing further out from the low-dielectric hydrocarbon core than those of the lipoidal 7-aminocoumarin and 7-hydroxycoumarin.Recent 47* 48 suggests that SDS micellar properties are particularly susceptible to pertubation by the presence of foreign solubilizates. This may explain the disparity between the eeff estimates.C. J. Drummond, F. Grieser and T. W, Healy 577 Conclusions The difference between the apparent pK, values of an azo indicator residing within the interfacial microenvironment of a non-ionic Brij-35 or OG micelle and the pKr value has been ascribed to the low interfacial Eeff of the non-ionic micelle. An analysis of the acid-base behaviour of the selected azo indicators indicated that the solvent properties of the interfacial microenvironments of Brij-35 and OG micelles are well characterized by E~~~ values of 35 4 6 and 48 4, respectively. These values are in excellent agreement with previous Eeff estimates.Two effects are considered to be primarily responsible for the difference between the p c value and the apparent pK, value of an azo indicator residing within the interfacial microenvironment of a cationic DTAB micelle, namely the electrosatic surface potential and the low interfacial ceff of the micelle. The interfacial Eeff of DTAB micelles has been deduced to be 36+9. This value is in accord with other independent ceff estimates. If the interfacial microenvironment of the SDS micelles possesses an E~~~ of ca. 60 and the y o of a SDS micelle is approximately - 140 mV then the difference between the p K value and the apparent pK, of an azo indicator located within the interfacial region of a SDS micelle can be explained solely in terms of the influence that the micellar electrostatic surface potential and low interfacial Eeff have on the acid-base equilibrium.Rough estimates of the mean interfacial solvent properties of the micelles gauged from the visible absorption spectra of the azo indicators concurred with the more exact E~~~ estimates derived from the acid-base behaviour of the azo indicators. The contention that the interfacial Eeff of SDS micelles is not equivalent to that of Brij-35 micelles is at odds with the earlier deduction by Fernandez and F r ~ r n h e r z ~ ~ which was also based on interfacial acid-base behaviour.This research was supported by the Australian Research Grants Scheme. References 1 G. S. Hartley and J. W. Roe, Trans. Faraday Soc., 1940, 36, 101. 2 I. M. Klotz and J. Ayers, J. Am. Chem. Soc., 1957, 79, 4078. 3 I. M. Klotz, Science, 1958, 128, 815. 4 I. M. Klotz, Brookhaven Symp. Biol., 1960, 13, 25. 5 I. M. Klotz, E. C. Stellwagen and V. H. Stryker, Biochim. Biophys. Acta, 1964, 86, 122. 6 C. E. Williamson and A. H. Corwin, J. Colloid Interface Sci., 1972, 38, 567. 7 C. E. Williamson and A. H. Corwin, J. Colloid Interface Sci., 1972, 38, 577. 8 M. Montal and C. Gitler, J. Bioenerg., 1973, 4, 363. 9 N. Funasaki, Nippon Kagaku Kaishi, 1976, 5, 722. 10 A. D. James and B. H. Robinson, J . Chem. SOC., Faraday Trans. I , 1978, 74, 10.11 G. M. Badger, R. G. Buttery and G. E. Lewis, J . Chem. Soc., 1954, 1888. 12 E. Sawicki, J . Org. Chem., 1956, 21, 605. 13 W. C. J. Ross and G. P. Warwick, J. Chem. Soc., 1956, 1719. 14 J. C. Merlin, J. L. Lorriaux, A. Dupaix and E. W. Thomas, J. Raman Specfrosc., 1981, 11, 131. 15 D. Jannakoudakis, E. Theodoridou and L. Moumtzis, Chimika Chronika, New Series, 1981, 10, 143. 16 E. Stahl and E. Dumont, J . Chromatogr. Sci., 1969, 7, 517. 17 T. R. Griffiths and P. J. Potts, Anal. Chim. Acta, 1974, 71, 1. 18 K. Machida, B. K. Kim, Y. Saito, K. Igarashi and T. Uno, Bull. Chem. SOC. Jpn, 1974, 47, 78. 19 A. T. Terpko, R. J. Serafin and M. L. Bucholtz, J. Colloid Interface Sci., 1981, 84, 202. 20 V. C. Reinsborough and J. F. Holzworth, Can. J . Chem., 1986, 64, 955.21 G. Ackerman, L. Sommer and W. I. Stephen, Pure Appl. Chem., 1985, 57, 845. 22 C. J. Drummond, F. Grieser and T. W. Healy, J . Chem. Soc., Faraday Trans. I , 1989, 85, 521. 23 C. J. Drummond, F. Grieser and T. W. Healy, J . Chem. Soc., Faraday Trans. 1, 1989, 85, 537. 24 C. J. Drummond, F. Grieser and T. W. Healy, J . Chem. Soc., Faraday Trans. I , 1989, 85, 551. 25 R. G. Bates, Determination of pH: Theory and Practice, (Wiley, New York, 2nd edn, 1973), p. 145. 26 R. L. Reeves, R. S. Kaiser, M. S. Maggio, E. A. Sylvestre and W. H. Lawton, Can. J. Chem., 1973, 27 W. R. Brode, I. L. Seldin, P. E. Spoerri and G. M. Wyman, J. Am. Chem. SOC., 1955, 77, 2762. 51, 628.578 Azo Acid-Base Equilibria in Micelles 28 W. F. Forbes and B. Milligan, Aust. J. Chem., 1962, 15, 841. 29 M. S. Fernandez and P. Fromherz, J. Phys. Chem., 1977, 81, 1755. 30 K. A. Zacchariasse, N. Van Phuc and B. Kozankiewicz, J. Phys. Chem., 1981, 85, 2676. 31 C. Ramachandran, R. A. Pyter and P. Mukerjee, J. Phys. Chem., 1982, 86, 3198. 32 K. Kalyanasundaram and J. K. Thomas, J. Phys. Chem., 1977, 81, 2176. 33 C. J. Drummond, F. Grieser and T. W. Healy, Faraday Discuss. Chem. SOC., 1986, 81, 95. 34 C. J. Drummond, F. Grieser and T. W. Healy, J. Phys Chem., 1988, 92, 2604. 35 F. Grieser and C. J. Drummond, J. Phys. Chem., 1988, 92, 5580. 36 L. S. Guss and I. M. Kolthoff, J. Am. Chem. SOC., 1940, 62, 249. 37 B. Gutbezahl and E. Grunwald, J. Am. Chem. SOC., 1953, 75, 559. 38 C. L. de Ligny, H. Loriaux and A. Ruiter, Recueil, 1961, 80, 725. 39 G. Akerlof, J. Am. Chem. SOC., 1932, 54, 4125. 40 C. J. Drummond, F. Grieser and T. W. Healy, Chem. Phys. Lett., 1987, 140, 493. 41 M. Nakagaki, M. Sakai and T. Handa, Chem. Pharm. Bull., 1984, 32,4241. 42 F. W. Critchfield, J. A. Gibson and J. L. Hall, J . Am. Chem. SOC., 1953, 75, 1991. 43 J. T. Edsall and J. Wyman, Biophysical Chemistry (Academic Press, New York, 1958), vol. 1, 44 S. H. Yalkowsky and G. Zografi, J . Pharm. Sci., 1970, 59, 798. 45 E. L. Mehler and G. Eichele, Biochemistry, 1984, 23, 3887. 46 G. V. Hartland, F. Grieser and L. R. White, J. Chem. SOC., Faraday Trans. I . , 1987, 83, 591. 47 D. Fornasiero, F. Grieser and W. H. Sawyer, J. Phys. Chem., 1988, 92, 2301. 48 G. G. Warr and F. Grieser, Chem. Phys. Lett., 1985, 116, 505. chap. 8. Paper 7/00186J; Received 21st December, 1987

ISSN:0300-9599    

DOI:10.1039/F19898500561

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1989, 85(3), 579-583 Surface Oxygen Species involved in the Oxidation of Carbon Monoxide over Chromium(1n) Oxide Masayoshi Kobayashi" and Tohru Kanno Department of Industrial Chemistry, Kitami Institute of Technology, Kitami, Ho k ka id0 090, Japan The reduction exceeding a certain amount of surface oxygen species on a-Cr,O, by CO causes a change in surface structure, with the conversion of more active to less active sites, which is visualized by measuring the distribution of the oxidation power of the surface oxygen species. There are many kinds of surface oxygen species on Cr,03 with different degrees of coordinated un~aturationl-~ and variously ionized forms which are affected by surface chromium ions.*-' Surface heterogenity should introduce a variation in the oxidation power of the oxygen species to a depth of a few atomic layers.e In the present study the distribution of the oxidation power of surface oxygen species on Cr,O, has been carefully measured using the KI meth~d,~~'O and the results obtained have shown that the progress of surface reduction with CO during CO oxidation causes the reconstruction of surface structure with a change in the relative rate of CO oxidation.Experimental The chromium sesquioxide used in this study was prepared by decomposing anhydrous chromium trioxide in an air flow at 450 "C for 5 h. The catalyst was confirmed to be a-Cr,O, by X-ray diffraction analysis, and the B.E.T. surface area of the sample was 21 m2 g-l. A weight of 5-55.2 g catalyst (24-32 mesh) was used, depending on the reaction temperature (105-172 "C), in order to keep below 10% conversion for a differential reactor.Commercially prepared CO, CO,, O,, He and N, were purified by passing them through a dry-ice-methanol trap to remove water vapour. Three gas chromatographs (thermal conductivity detector) were simultaneously used to follow the transient response curves at the outlet of the reactor. An ordinary tubular flow reactor (Pyrex glass) was employed under atmospheric pressure at a total gas flow rate of (10-160)+ 2 cm3 (s.t.p.) min-l. Intraparticle and external mass-transfer effects were confirmed to be negligible prior to carrying out the experiment. Our previous papers'' have described the details of the experimental procedure. The distribution of the oxidation power of the surface oxygen species was measured by reducing the catalyst surface with potassium iodide solutions at various pH values from 7.1 to 12.2 (the KI method), in which a higher pH produces a higher oxidation power.After the catalyst had been treated under reaction steady-state conditions it was cooled to room temperature. The catalyst was then carefully exposed to the KI solutions with no exposure to air. A more detailed explanation of the KI method can be found el~ewhere.~' lo* l2 579580 Oxidation of CO on Cr,O, 2.5 / i 0.l 0 0.0 5 0 P%o /am Fig. 1. Distribution of the oxidation power of surface oxygen species on the fresh catalyst and on the catalyst used for CO oxidation. Results and Discussion Distribution of the Oxidation Power of Surface Oxygen Species during CO Oxidation The amounts of surface oxygen species ( q o ) on a fresh catalyst and the catalysts used for the reaction at various pressures of CO:pco, with constant po, were measured.The curves obtained by plotting qo, vs. the oxidation power were graphically differentiated to show the distribution curve of the oxidation power of the surface oxygen species, and the results are shown in fig. 1. The curve for the fresh catalyst clearly demonstrates the existence of two kinds of oxygen species with different oxidation powers (lower and higher than pH 9.0; designated 0; and O!, respectively). The amounts of 0; and 0: measured at 163 "C and various pco values are presented in fig. 2(a). The amounts of 0; are evaluated to be 1.1 x mol g-' at two different regions below and above pco = 0.06 atm,? respectively, with no dependence on pco within the same region.Here, one may designate the specified partial pressure of CO to be pco(c), with region 1 below pco(c) and region 2 above pc,(c), as shown in the figure. The value of pco(c) involves some fluctuation, depending on the history of the catalyst resulting from the degree of surface reduction, which is the reason why the curve for 163 "C in fig. 2(b) shows hysteresis. and 0.6 x t 1 atm = 101 325 Pa.M . Kobayashi and T, Kanno 58 1 0 region1 / 7 region 2 0.0 5 0.1 0 0.1 5 PCO /am Fig. 2. The amounts of O:, 0; and r as a function of pco. 0," and 0; are surface oxygen species with oxidation powers higher and lower than pH 9.0, respectively; r is the reaction rate at a steady state.po, = 0.20 atm. (a) T = 163 "C; (b) T = c), 105; 0, 131, 0, 163 and 0, 172 "C. The constancy of 0; in regions 1 and 2 may be attributed to the fact that the rate of regeneration of the oxygen species from gaseous 0, is much faster than the rate of consumption due to the reaction with CO and the fast desorption of CO, formed. In addition, when the catalyst is highly oxidized, the value of 0; and the steady-state reaction rate are independent of pOQ, and the former is proportional to the degree of catalytic activity of the catalysts prepared with various procedures. From these results one may recognize that 0; is one of the oxygen species working at the reaction steady state. The amount of O:, on the other hand, decreases with increasing pco within regions 1 or 2, as can be seen in fig.2(a). This diminution will result from both the slow regeneration of the surface oxygen species from gaseous 0, and the accumulation of CO, formed on the active sites, which are different from those for 0;. The anomalous stepwise increase in the amount of 0: at pco(c) z 0.075 atm may be attributed to the reconstruction of the surface layer, even though no change in the X-ray diffraction pattern of the catalysts between regions 1 and 2 is detected. Zecchina et aZ.2 demonstrated that the surface structure was easily rearranged to form the complex CrOz- under ambient conditions at temperatures > 200 "C. The colour of the catalyst used in the present study was varied from dark to light green by the reduction with CO, suggesting a change in the surface structure even at temperatures < 200 "C.In region 2 one may speculate that the reconstructed surface causes a higher degree of unsaturation582 Oxidation of CO on Cr,O, which will give a higher affinity with oxygen than that in region 1, which is the reason why the anomalous increase of 0; appears at p,,(c). In our previous papers on MnO,,* similar behaviour was observed for the rate of CO oxidation, in which the surface crystal structure was easily changed from MnO, to Mn2O,.l2 Rate Analysis at the Steady State and the Transient State Fig. 2(b) illustrates the steady-state reaction rate as a function ofp,, at 105-172 "C. An anomalous decrease in the rate is observed at temperatures > 131 "C and at a specified p,,(c).One may again designate regions 1 and 2 as shown in the figure. The slow rate of regeneration of 0; causes a non-linear relationship between the apparent reaction rate and pco. The reaction is unstable near p,,(c), and thereby the rate data obtained fluctuate between the two points presented atp,,(c) = 0.06 atmand 163 "C.p,,(c) shifts to lowerp,, with increasing temperature; p,,(c) = 0.08 at 131 "C, 0.07 at 163 "C and 0.06 atm at 172 "C. The reason for this shift may be that the change in surface structure is accelerated with an increase in temperature, producing mobile surface species. Note the change in the total amounts of 0; and 0; at pco > p,,(c) (1 63 "C) in fig, 2(a); the former shows a decrease of 40 YO and the latter an increase of 29 YO at pc0(c) in region 2.Taking into account both the arithmetic balance of the two amounts and the selectivity to reaction path I (~OYO), which will be mentioned later, one can estimate a reduction of ca. 12 % [( - 40 x 0.6) + (29 x 0.4)] in the overall reaction rate in region 2 at p,,(c). This is close to a 10 YO diminution of the actual rate, which is estimated graphically from the stepwise reduction of the curve for - 163 "C at p,,(c) = 0.075 atm in fig. 2(b). At the reaction steady-state in regions 1 and 2 the activation energy and the rate equation are commonly evaluated as 57 kJ mol-1 and r = k&' p;,, respectively, suggesting no differences in reaction mechanism between the two regions. The reduction of the apparent reaction rate at p,,(c) may therefore be attributed to a decrease in the amount of 0;.The mode of the transient response curves of CO, obtained in region 2 is, on the other hand, characterized by two parts : one is an instantaneous change at the initial stage of the response (corresponding to reaction path I) and the other is followed by a slow-overshoot response (corresponding to reaction path 11), indicating similar behaviour to that in region 1 [see ref. (11) and (13)J. From these similarities one may recognize that a two-reaction-path mechanism is acceptable in both regions 1 and 2. In our previous papers1'. l3 the two-reaction-path mechanism was demonstrated in region 1: for reaction path I cog+o;*s,-+co,~s, (1) 02,g + 2 c 0 , s, -+ 2CO,,, + 2 0 3 , (2) for reaction path I1 (3) cog + 0; s,, -+ CO, s,, The two paths progress simultaneously during the reaction, in which S, and S,, are active sites for the adsorption of 0; and Oi, respectively. A kinetic analysis of the steady-state rate data and a graphical analysis of the transient response curves in region 1 roughly estimate the kinetic parameters at 163 "C: k, = 3.4 x k, = 1.3 x k, = 3.5 x k, = 7.2 x k, = 2.1 x mol atm-l g-' min-l mol atm-l g-' min-l mol atm-' g-' min-' mol atm-l g-l min-' mol g-l min-IM .Kobayashi and T. Kanno 583 Using the estimated surface-reaction rate constants k, and k,, since the total numbers of active sites S, and S,, are previously evaluated as 6.6 x and 2.34 x mol g-', re~pectively,'~ the turnover frequencies are evaluated to be 4.3 x lo-, and 2.25 x s-l, respectively. Since the surface reaction (1) is the rate-determining step in path I, it is 1.9 times faster than reaction (3) in path 11, whereas the rate of path I1 is controlled by all the steps (3H5).The selectivity to path I, which varies with temperature and pco, is evaluated to be 60 % at p,,(c) = 0.075 atm and 163 O C , indicating the predominance of path I rather than path 11. Conclusions The oxidation of carbon monoxide is sensitively affected by the distribution profile of the oxidation power of the surface oxygen species, which results in a complex reaction mechanism. A mobile surface model and the two-reaction-path model are proposed to interpret the drastic change in the steady-state rate atp,,(c) and the characteristic transient response curves. References 1 D. A. Dowden and W. E. Garner, J. Chem. SOC., 1939, 893. 2 A. Zecchina, S. Coluccia, L. Cerruti and E. Borello, J. Phys. Chem. 1971, 75, 2783. 3 M. P. McDaniel and R. L. Burwell Jr, J. Catal., 1975, 36, 394; 404. 4 H. Marczewska and S. Benbenek, Bull. Acad. Pol. Sci., 1979, XXVII, 35. 5 H. Marczewska and S. Benbenek, Heterogeneous Catal., 1979, 1, 1205. 6 T. A. Egerton, F. S. Stone and J. C. Vickerman, J. Catal., 1974, 33, 307. 7 A. A. Davidov, Yu. M. Shehekochikhin and N. P. Keier, Kinet. Katal., 1969, 10, 1341. 8 W. Weller and S. E. Voltz, J. Am. Chem. SOC., 1954, 76, 4695. 9 T. Uchijima, M. Takahashi and Y. Yoneda, J. Catal., 1967, 9, 402. 10 T. Uchijima, M. Takahashi and Y. Yoneda. Bull. Chem. SOC. Jpn, 1972,40, 2767. 11 M. Kobayashi and H. Kobayashi. Bull. Chem. SOC. Jpn., 1976, 49, 3009; 3014; 3018. 12 M. Kobayashi and H. Kobayashi. J. Catal., 1972, 27, 100; 108; 114. 13 M. Kobayashi, J. Chem. Tech. Biotechnol., 1979, 29, 552. Paper 8/00056E; Received 4th January, 1988

ISSN:0300-9599    

DOI:10.1039/F19898500579

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1989, 85(3), 585-599 Stepwise Adsorption at the Same Site A Thermodynamic Treatment Edoardo Garrone* and Piero Ugliengo Dipartimento di Chimica Inorganica, Chimica Fisica e Chimica dei Materiali, Universita di Torino, via P . Giuria 7, 10125 Torino, Italy Either elementary statistical thermodynamics or the use of chemical equilibria permits the treatment of the adsorption of L ligands on S centres with more than one coordination vacancy. The case of two coordinative unsaturations, i.e. the formation of surface complexes up to SL,, is treated in an exact, analytical way. The properties computed are: (i) the partition of L ligands into S, SL and SL, species, (ii) the overall adsorption isotherm and (iii) the molar and differential energies of adsorption.The case of an unstable SL complex is also considered. With regard to the case of three coordinative unsaturations, i.e. the formation of surface complexes up to SL,, the problem can only be solved numerically, although some overall features can be calculated analytically. The case of unstable complexes, either SL or SL,, is also treated. In the absence of interactions within the same site, the problem is analytically soluble: the adsorption isotherm is Langmuirian, and the partition of the adsorbate into different SL, surface complexes is governed by a statistical binomial distribution. In surface studies it is often observed that adsorption centres may coordinate more than one ligand. Typically this occurs on ions or atoms of transition metals dispersed on inert surfaces.For example, such behaviour is exhibited by the reduced Phillips catalyst for ethylene polymerization, a system that has been studied in detail in our laboratory' and which consists of Cr" dispersed on amorphous silica. With regard to NO, the vast majority of surface ions coordinates at room temperature two ligands, without the formation of mononitrosyl intermediate.' The adsorption of CO is complex, but is known to involve up to three ligands: various adsorption schemes have been and all envisage the stepwise addition of CO ligands. Many other examples could be given. A recent one is the formation of a stable Rh' dicarbonyl on highly dealuminated zeolite Y :5 the monocarbonyl is not observed. The question arises as to how different must the energies of mono- and di-ligand complexes be in order that the former is absent.Note that entropic reasons would in any case favour its formation to some extent. To the best of our knowledge, no thermodynamic treatment is available concerning the relatively simple problem of the adsorption of more than one ligand per site. This is proposed in the present paper on the basis, on the one hand, of elementary statistical thermodynamics, and, on the other hand, of a chemical equilibrium formalism. Two cases are considered : the formation of up to two-ligand complexes and up to three-ligand complexes. In the following, the surface phase is considered ideal, i.e. all sites, as well as the surface complexes, are regarded as equivalent (when appropriate) and non-interacting.Obviously this limitation renders the model adopted a 'zeroth-order ' approximation to more sophisticated treatments. 585586 Adsorption Thermodynamics Results and Discussion Formation of up to Two-ligand Complexes Let us consider M sites designated as S, which can form, upon adsorption, both SL and SL,. Let - E (E > 0) be the adsorption energy of L (the energy of formation of SL), and - 2 ~ - 2 w be the energy of formation of SL,; w is thus the interaction energy between the two L ligand molecules adsorbed on the same site per L molecule: w > 0 indicates attraction, w < 0 repulsion. Let N molecules be adsorbed, of which Nl are as SL and Nz are as SL,: obviously N = Nl + N,. The numbers of occupied sites are Nl and N2/2, respectively. By definition we have: el = N,/M 8, = N,/(2M) e = ~ p w = e , + 2 e , .6, = [M- (Nl + N,/2)] / M = 1 --01 - 6, St atistical- thermodynamical Treatment For given values of E, w and 8, it is possible to calculate the equilibrium value of 8, (or 0,) by minimizing the free energy of the system with respect to the same parameter 6, or 8,. The internal energy of the system is u = - N, E - ~,/2(2E + 2 4 = - i q E o + w(e- el)]. (2) The configurational entropy Sconr is S,,,,,/k, = In (M!/[N,!(M-N, -N,/2)! ( 4 / 2 ) ! ] } . By recursion to the usual Stirling approximation and the above definitions we readily obtain The non-configurational entropy is Xc = Nl sy + N, si, where sy and si denote, respectively, the molar entropy of the ligand L in SL and SL,. We obtain ~ , , , , / k , = - ~ ( ( 1 - e l 2 - 8,121 In (I - e l 2 - e ~ 2 ) - el In 6, - (e- el) In [(o- el)]>.(3) sc = we, 3; + 28, s;] = Wes; + (sy - s;) ell = w8.g + ssoe,] with 6s" = si-sy. The total entropy is S = qonf + &. The free energy of the system is A = U - TS. By substituting the above formulae, an explicit expression for A is obtained. The equilibrium is attained for the minimum value of A , i.e. by taking By simple algebraic manipulation eqn (4) yields (dA/se,), = 0. (4) ( 5 ) - TSSO + w = k, ~ / 2 In [2(1- e l 2 - e ~ 2 ) (8- e,)/e;]. We define a = exp (- 6s0/k,) exp ( w / k , T ) the equation defining 0, then becomes (a2-i)e;+2e1-e(2-e) = 0.E. Garrone and P. Ugliengo 587 Before giving the explicit solution of eqn (6), we show that the same formula can be reached by a different and simpler pathway.For the sake of simplicity, from here onwards the assumption will be made that s; = si; i.e. the non-configurational entropy per mole of L ligand will be assumed to be the same in both SL and SL,. This means 6s" = 0 and a = exp (w/k, T ) . Chemical Equilibrium Formalism Let us write the equilibrium between the surface species S, SL, and SL, and the gaseous species L: s + q g ) e SL K~ = e,/(e,p) ( 7 ) SL+L(~)=SL, K, = e,/(e,p). (8) Owing to the absence of interaction among sites, the thermodynamic activities for S, SL and SL, are simply O,, 8, and O,, respectively. The activity for the gaseous species L is its pressure p . By means of eqn (1) it is possible to express 8, and 8, as functions of 8 and 0,. It turns out that e, = (8- e , y 2 O, = 1 -812 - e,p.(9) The zero-point energy change associated with reaction (7) is AU: = - E ; that associated with reaction (8) is AU; = - ~ - 2 w . K, may be written as Kl = ql/(qOqL)exp(E/k, T ) , where q,, q, and qL are the partition functions of S, SL and gaseous L, respectively. stands Because two equivalent configurations (scheme 1) are possible for SL (where Scheme 1. for coordinative unsaturation), we get where O indicates the partition function over the internal degrees of freedom. Obviously 41 = 24; 40 = q L = qoL- q,/(q, q J exp [ ( E + W / k , TI. In a similar way K, may be written as Again q1 = qy and q2 = q i ; qL = q:. Dividing eqn (8) by ( 7 ) gives K,/K, = e2 e,p; = [(e- e,)/2(1- 012 - O,/~)]/O:. (10) Also K 2 / K 1 = qi qi/(2qy)2 exp (2w/kB T).By analogy with results obtained previously, we assume that the partition function for the internal degrees of freedom of L is the same both in SL and SL,. This means we may assume that Recalling that exp(w/k, T ) = a, we get (43, = 4; qi. a' 4 K,/K, = - .588 Adsorption Thermodynamics Eqn (10) transforms into (a2 - 1) 8; + 28, - 8(2 - 8) = 0 which coincides with eqn (6). Eqn (10) describes the equilibrium among the surface species: 2SL * s + SL, which must hold when equilibria (7) and (8) take place. Note that the statistical treatment yields this equilibrium constant as a consequence of minimizing the free energy. Coverage of S, SL and SL, as a Function of 8 Eqn (6) is easily solved to yield 8, = 8,(8). The other two functions are readily obtained from 8, and 8 according to eqn (9).For a = 1, i.e. w = 0 (no interaction between the ligand molecules in SL,) we obtain el = e(2-e)/2 8, = (2 - 4214 8, = 8214. (1 1 ) Note that Oi simply follow, in this case, a statistical binomial distribution, as 8 / 2 and (2 - 8)/2 are the probabilities of finding a coordinative position full or empty. For 8, a factor 2 takes account of the two equivalent configurations (see scheme I). Fig. l ( a ) illustrates the behaviour of the three coverages Bi as a function of 8, for w = 0. From eqn (1 1 ) the following symmetry relationships hold : e,(e) = el(2-e) e,(s) = e,(2 - el. These symmetry eqn (6) is relationships hold for any 8, = {-1+[1+(a2- value of w. In fact, the 1 ) e(2 - 8)l+>/(a2 - 1 ) general solution ( for both w > 0 and w < 0.Eqn (12) is invariant for the substitution 8+ 2-8; thus 8,(8) = 8,(2-8). Eqn (10) may also be written as 0, = (2-8-8,)/2, so that we get 8,(8) has a maximum at 8 = 1, whatever the value of w. The maximum value is e,(e) = e2(2-e). O r = 8,(l) = 1/(1 +a) = [I +exp(w/k, T)]? Fig. 1 (b), (c) and ( d ) illustrate the sets of Oi curves obtained for the following w/kB T values: f 1, f 2 , + 3 . For increasing w > 0 (attraction) 8, becomes less and less important: e.g. for w/k, T = 3 [fig. 1 (d)] 8, is rather small and nearly negligible. As a consequence, 8, and 8, approach, as a function of 8, the straight lines 8, = 1 -8/2 and 8, = 812, respectively. In order that the coverage 8, of the SL complex is negligible with respect to 8,, it is sufficient to assume Or to be below, e.g., 0.01.This corresponds to w/k, T 2 4.6. At room temperature this means that, if w > ca. 12 kJ mol-', the concentration of the monoligand species SL will be below the detectability threshold. For decreasing w < 0 (repulsion), 8, becomes more and more important. For w/kB T = - 3 , SL, formation practically does not take place until 8, = 1; then each SL is converted into SL,. As a consequence, in the range 0 < 8 < 1 we have 8, z 1-8; 8, z 8 ; 8, = 0; in the range 1 < 8 < 2 we have 8, z 0; 8, = 2-8; 8, = 8-1. Adsorption Isotherm Since 4 = O,/(O,p), from eqn (9) and (12) it is readily deduced that 2&p = (8-8,)/8, = { - I +o+[i +(a2-l)e(2-e)lt~/(2-e).E. Garrone and P . Ugliengo 589 0 1 2 0 i 2 e Fig.1. Partition of adsorbed L molecules into S, SL and SL, species as a function of the total coverage. 8, is the coverage of empty sites; 0, the coverage of SL species; 0, the coverage of SL, species; 8 is the total coverage. Note that 0 6 Oi 6 1 ; 0 d 0 d 2. The value of w/kll T to which the curves refer is reported in parentheses as O,(wi/k, T ) . (a) w / k , , T = 0 (no interaction between L molecules adsorbed on the same site); (b) w / k , T = f 1 ; (c) w / k , , T = & 2; ( d ) w / k , , T = & 3.590 Adsorption Thermodynamics O{ 0.0 215 510 I I- 0 50 100 K1 P I 2 Fig. 2. Adsorption isotherm for the formation of species up to SL,. The value of w / k , T to which the curves refer is reported above each curve. (a) Dimensionless pressure in the range [0,5]; (6) dimensionless pressure in the range [0,100].Note that K2 still contains w, being given by & = 42/(41 qL) exp [('+ 2w)/kB whereas Kl = ql/(qOqL)exp(e/k, T ) does not. It is convenient to make explicit the dependence of K2 on w (or a) by taking into account that K2/K1 = a2/4. Thus eqn (13) becomes (14) Fig. 2 reports some curves for different values of a. The abscissa scale is the dimensionless pressure K1p/2. For the sake of clarity, fig. 2 is divided into sections (a) and (b), differing by the ranges of dimensionless pressure studied. For w = 0 [no interaction, fig. 1 (a)] the isotherm reduces to the Langmuir expression: K1p/2 = { - 1 + 8 + [ 1 + (a2 - 1) 8(2 - 6)]i)/[a2(2- 8)]. K1p/2 = 8/(2 - 8). For w > 0 (a > 1, attraction) the curves lie above the w = 0 curve and tend rapidly to 8 = 2.For w < 0 (repulsion) the curves lie below the w = 0 case and tend to 9 = I for pressures not too large. The pressure p at which half the possible coverage is attained (0 = 1) is calculated to be iKlpt = a-' from which lnp; = 1n(2qOqL/q,)-(E+ w)/kB T* In the Henry region 6 z 0, by series expansion of eqn (14) we obtain 8 z Klp, and in fact all curves in fig. 2 have the same initial slope of 2. For large values of w > 0, the parameter a increases rapidly and SL becomes negligible. For very large a, eqn (1 4) becomes iKl up = [8/(2 - S)];E. Garrone and P. Ugliengo 59 1 0 i e 2 Fig. 3. Variation of the molar energy of adsorption with total coverage, as expressed by the function r defined in the text.The values of w/k, T to which the various curves refer are reported above each curve. which gives ( 14’) where Again, Langmuir behaviour is obtained: eqn (14’) can also be reached by posing the following chemical equilibrium conditions : S+2L,SL2 K = 42/(40 q2L) exp E2(E + w)/k, TI whence 8 = 26,, K‘ = 6,/(O0p2) and so forth through straightforward manipulations. Molar and Differential Energies of Adsorption The tractability of the process under study allows the explicit calculation, besides the isotherm, of the basic adsorption energies. The internal energy of the adsorbed system is, according to eqn (2), By definition the molar energy is u = U / N , for which the following formulae hold: u= -M(&6+w6-w61). u = - & - w + we,/o = -&-w+w{-i+[i +(a2-i)o(2-e)lt>/[(a2-~)o].(1 5 ) The limit of the molar energy for vanishing 6 is - E ; when 8 = 2 the molar energy is A convenient way to represent in a normalized fashion the behaviour of the molar - & - W . energy of adsorption is as follows. Eqn (15) may be rewritten as r = (u+E)/w = - 1 + { - I +[I +(a2- 1>0(2--8)1+/[(a~- 1)0] r = (6, -6)/6. Fig. 3 shows the function r(8) for some relevant w values. For w > 0 (attraction) r592 Adsorption Thermodynamics 0 1 e 2 Fig. 4. Variation of the differential energy of adsorption with total coverage, as expressed by the function r defined in the text. The values of w / k , T to which the various curves refer are reported above each curve. declines rapidly as a function of 9 : for large w / k , T > 0 the molar energy rapidly approaches - E - 2w, as SL, species are immediately formed.For w < 0 the decline of r is less pronounced: for w 4 0 (strong repulsion) the molar energy is constant at --E for 0 < 9 < 1, because only SL species are formed, and then tends to - E - w as the formation of SL, takes place. Fig. 3 seems to suggest a linear decline in molar energy of adsorption for w = 0. This is clearly an artifact: for w = 0 the value of u is constant at -6, so that r is an undefined ratio of the kind O/O. As r = (0,-9)/9, in this case 9, = 9(2 - 9)/2 so we obtain r = - 9 / 2 , as observed. case The differential energy of adsorption u is, by definition, u = 6U/6N. In the present u = ( i / ~ ) ( m / ~ ) = -E-w++We,pe. In conclusion u = - & - w + w ( ~ -e)/[i +(a2- i)e(2-e)lf.(16) For 9 = 0, u = --e; for 9 = 2 the differential energy of adsorption is equal to - - E - ~ w . As before, we define Clearly, Y is defined between zero and -2. Fig. 4 reports various plots of Y for some relevant w values. The comments to fig. 4 are similar to those to fig. 3. For strong attractive interaction (w >> 0) the adsorption is pairwise, to yieId SL,, and the differential heat is close to - E - w over the whole range 0 < 9 < 2. For strong repulsion (w 4 0) SL complexes are formed at first, with aE. Garrone and P. Ugliengo 593 constant differential heat - E , then insertion of a second ligand occurs starting at 6 = 1 with a differential heat - ~ - 2 w : an abrupt jump is therefore present at 0 = 1 . Note that the linear plot for w = 0 is again an artifact, for the same reasons discussed above: r is an undefined ratio of the kind O / O ; because it is equal to - 1 +S6,/69, in the present case we get r = -6.Note that each curve has an inversion centre at the point (1, - 1). Furthermore, it is easily shown that eqn (16') is invariant under the substitution w + - w, r --+ - 6 and 6 + - r ; i.e. the r = - 6 line acts as a mirror axis for two curves with opposite w values. The behaviour of both the molar and differential energies of the adsorbed phase is similar to that encountered in an apparently different case, the quasi-chemical treatment of an assembly of interacting particles adsorbed on a lattice, given by Wang.6 (The use of the functions r and r has been inspired by that paper.) The reason is that in the present case we are dealing with the actual surface equilibrium 2SL s S + SL, : in the quasi-chemical approximate treatment of the surface Ising problem7 a similar pseudo-equilibrium is assumed between interacting pairs, empty sites and non-interacting adsorbates, to allow the calculation of the partition function of the system: hence the similarity of the results.Formation of up to Three-ligand Complexes Let us define as O,, O,, 6, and O3 the surface coverage of the various complexes having 0, 1, 2 and 3 ligands, respectively. Obviously 8,+Ol+6,+O3 = 1. (17) (18) note that: 0 < Oi < 1, i = 0, I , 2 ; O < 6 < 3. Clearly, for a given 9, only two Oi are independent, say 6, and 6,. In principle, it is possible to carry out a statistical treatment similar to that given for the SL, problem.Once the free energy of the surface phase has been calculated, the equilibrium conditions are those ensuring that A has a minimum : The surface coverage with respect to L is 9 : 6 = 6, + 26, + 36, (6A/S81)o = 0 (SA/S8,), = 0. For the sake of brevity these calculations are not reported, as the same results are arrived at in a much simpler way by the chemical-equilibria formalism. Let us assume that S + L(g) e SL Kl = 01/9,,p = 3q,",/q; qL exp (- A U,"/k, T ) SL -k L(g) * SL, 4 = 92/81 P = qsOLz/qgL qL exP ( - A U,O/k, T ) (19) As before, the ideality of the system allows the use of coverages Oi as thermodynamic activities; the equilibrium constants are expressed in terms of the zero-point energies of adsorption, AU:, AU; and AU;, and the partition function of the various surface complexes : qi stands for the partition function calculated over the internal degrees of freedom ; numbers take account of the configurational degeneracy.Although con- ceptually simple, the illustration of the energy levels of adsorbed systems may give rise to some confusion. Scheme 2 describes the meaning of the symbols adopted. sL2+L(g)*sL3 K3 = '3/'2P = q~L,/3q~L2qLexp(-Au~/k, From eqn (19) we obtain K2/& = qgL2 qs0/3(q,",l2 exp [ ( A G - A W k , TI = 8, 9,/8: K3/K2 = 4,"J3(q,",J2 exp [ ( A G - AU,")/k, TI = 63 9,/0;* By assuming, as before, that the partition function over the internal degree of freedom594 Adsorption Thermodynamics S SL SL SL3 -~ zero-point energy changes in successive adsorption energy of the system SL&i=O, ... 3) 0 no interaction , .. . . . . . . . . . . . . . . , , . . . . . - € - 2E - 3 E - & L .-.-. -.-.- Scheme 2. of L is more or less the same in all surface complexes, it results that the various qo cancel, and the following much simpler expressions are obtained : (20) c0; = 3008, where c = exp [(AU," - AU,")/k, TI bOZ, = 38,8, where b = exp [(AU," - AU,")/kB TI. Eqn (20) are the equilibrium constants for the processes s + SL, $2SL SL + SL, e 2SL, which express the thermodynamic equilibrium among the surface species. These are the expressions also arrived at by minimizing the free energy of the adsorbed phase. Eqn (17) and (18) may be manipulated as follows: (21) eo = 3 3 - o - ( 2 4 + e,)] 8, = ;[e - (el + 2e,)l.By means of these relationships, eqn (20) become These two equations form a system whole solution yields 8, = O,(O) and 8, = O,(O). It can be shown that system (22) is equivalent to a fourth-degree equation in 8, or O,, which in principle has an analytical solution, but which in practice must be handled numerically. The non-availability of a simple analytical solution for the whole problem of multistep adsorption at the same site for more than two ligands prevents theE. Garrone and P. Ugliengo 595 development of explicit formulae for the adsorption isotherm and the molar and differential energies of adsorption, as in the previous case. Some general considerations may, however, be made, as follows. In the case of non-interaction between adsorbed ligands, w2 = w, = 0, i.e.AU: = AU; = AG; then c = b = 1. Subtraction between eqn (22) in this case simply yields The use of this relationship is twofold: on the one hand, from the second equilibrium constant of eqn (19) : 9, = (3-9)/99, K2 = 92/(91 P ) 8(3 - 9) = K,p we readily obtain the isotherm for the overall process, which, as expected, has a Langmuirian form. calculate that On the other hand, insertion of this relationship into part of eqn (22) permits us to 0, = 9(3 - 9)’/9. The other coverages result : 8, = 02(3 - 9)/9 9, = 9,127. In conclusion, the overall process obeys a Langmuir isotherm and the coverages are given by a statistical binomial distribution, just as in the corresponding SL, case. Fig. 5 ( a ) shows the various ei (i = 0, 1,2,3) as a function of 9 in the case of non- interaction.9, has a maximum at 8 = 1, and 9, at 8 = 2. Also, the following symmetry relationships hold : 9,(9) = 9,-,(3 - 9) i = 0 , l . These symmetry properties are still valid when b = c # 1. In fact, as far as eqn (2 1) are concerned, the replacement 9+ 3-8 transform 9, into 9,, i.e. el(@ = 8,(3-9). It is readily checked from eqn (1 7) and (1 8) that, if this condition holds, go(@ = 9,(3 - 9). The whole set of functions ei(9) thus obey the above symmetry properties. Fig. 5(b) shows the various gi (i = 0,1,2,3) as functions of 9 in one case when b = c = 1/10. These symmetric properties are lost when c or b differ from each other. c # 1 means attraction or repulsion between the two ligands in SL, with respect to SL; b # 1 means attraction or repulsion between the three ligands in SL, with respect to SL,.As illustrated in scheme 2, AU:, AU,O and AU; can be written as AU; = - e , AU; = -c--2w2 and AU; = - E - 3w, + 2w,, where w2 and w, represent the molar interaction energies of L ligands in SL, and SL,, respectively; c is defined as exp[(AU:-AUi)/kB TI = exp(2w2/k, T ) and is thus equal to a2 defined above. b is equal to exp [(A Ui + A U,”)/k, TI = exp [(3 w, - 4w,)/k, TI and c is equal to b when w, = 2w,. Even in the general case a # b # 1 it can be shown that el(@ and 0,(9) reach their maximum at 9 = 1 and 9 = 2, respectively. The proof is as follows. Differentiation of formulae (1 7), (1 8) and (20) yields the following system of linear equations in doi : d9, + d9, + d9, +do, = 0 (23) d9, + 2d9, + 3d0, = d9 38, d9, - 2c9, d9, + 39, d9, = 0 39, d9, - 2b02 do, + 38, d8, = 0596 Adsorption Thermodynamics -c-- - ------------ _cc--- --- 0 1 2 3 Fig. 5.Partition of adsorbed L molecules into S, SL, SL, and SL, species as a function of the total energy. 8, coverage of empty sites; 01, coverage of SL species; 8,, coverage of SL, species; O,, coverage of SL, species; 8, total coverage. Note that 0 < 8, < 1; 0 < 8 < 3. The difference in stability between SL and SL,, SL, and SL,, respectively, is expressed by the two parameters b and c (see text). (a) b = c = 1 (no interaction between L molecules adsorbed on the same site); (b) b = c = 1/10; (c) b = 1/10; c = 5. which can be solved to give doi = D,/D,de, where Di and Do are the appropriate determinants as defined by Cramer’s rule.The maximum value of 8, is found when D , = 0. From eqn (23), D, is calculated to be D, = - 2be; - 3 4 e, +- 30, e, D, = - 681 e, - 3 4 0, +- 38,e, = - 3e,(2e, +- e, - OJ. orE. Garrone and P. Ugliengo 597 D, = 0 when 8, = 8, +20,. From eqn (17) and (18) it is seen that this condition corresponds to 8 = 1. Similarly, the maximum in O,(8) is found for D, = 0: D, = - 3( 8, e3 - 2ce; - 8,8,) = - 3(8,8, - 28,0, - 8,8,) = - 38,(8, - 28, - 8,). D, = 0 when 83 = 28,+8,. Again from eqn (17) and (18) it is seen that this condition corresponds to 8 = 2. Fig. 5 (c) illustrates the case b = 1 / 10, c = 5. It is evident that 8, still has a maximum at 8 = 1, and 8 2 at 8 = 2, but all symmetry properties are lost.Two particular cases appear to be of interest. Like the SL, case, it is possible that one (or both) intermediates are unstable. SL Unstable If a 9 1,8, becomes negligible and can be set equal to zero. Thus formulae (1 7) and (1 8) can be rewritten as 1 = 8 0 + 8 2 + 8 3 0 = 28, + 30,. (24) The equilibria between the gas phase and the surface complexes are SL, K; = 8 2 / 8 0 p 2 = q2/(qO qi) exp (2~’/k, T ) S + 2L sL2+ * sL3 Ki = ‘ 3 / ’ 2 P = q 3 / ( 9 2 41,) exp I(&’+ w’)/k13 E’ is the molar adsorption energy of L ligands in the SL, complex, and is also equal to E + w,. w’ is the interaction energy among the three L ligands in SL,. w’ < 0 means repulsion, w’ > 0 attraction. We thus obtain (25) with c’ = exp(w’/k, T ) . The above expression is the equilibrium constant of the following surface reaction : Because 8, = (0-28,)/3 and 8, = (3-8-8,)/3, eqn (25) becomes K12/K; = 8:8,/8: = qEq0/qiexp(2w’) = d2/27 3SL, e s + 2SL3.~’~0~-(0-282)~(3-8-82) =f(8,82,~‘) = 0. (26) This third-degree equation allows the calculation of 6, as a function of 8 for any given value of c’. The 8 value for which 8, has a maximum is given by (Sf/S8),, = 0. It is readily confirmed that this condition corresponds to 8 = 2, as in the general case. In the present case the value of the maximum 0; can be calculated explicitly, because eqn (26) becomes and thus ~ ” ( 8 : ) ~ = 4( 1 - 8r)3 8: = [ 1 + (c’2/4)p. SL, Unstable This case is quite similar to the preceding one. We assume that e2 = 0; thus 1 = 8,+8,+8, 8 = 8,+38,.598 Adsorption Thermodynam ics The equilibria between the gas phase and the adsorbed species are + * sL K;' = el/(90 P) = qI/(qO qL) exp (&lkB T , SL + 2L $ SL, K l = B3/(9,pZq,)/(q, q i ) exp [ ( 2 ~ + 3w3)/kB TI where w, is the interaction energy among the three L ligands per mole of L ligand; as before, w3 > 0 means attraction, w, < 0 repulsion.We thus obtain K;,/Ki = q:/(qi 4,) eXp ( - 3 W/kB T ) = 2 7 / f 3 (27) where C" = exp(w3/kB T ) . Because e3 = (9-9,) and 9, = (3-9-29,)/3, eqn (27) becomes ~ " ~ 9 ~ - ( 9 - 9 , ) ( 3 - 9 - 2 9 , ) ~ =f(9,,9, c') = 0. This third-degree equation allows the calculation of 9, as a function of 9 for any given value of c". By putting it is found that 8 = 1 is a maximum, as in the general case. (df/d9)&, = 0 gy = (1 + 2-5c7-1.is computed to be Conclusions The use of either elementary statistical thermodynamics or chemical equilibria allows us to treat adsorption on centres with two coordinative vacancies, i.e. capable of forming surface complexes up to SL,, in an exact analytical way. The surface coverages of each species S, SL and SL, may be calculated as a function of the overall coverage 9, the only parameter being w, the molar interaction energy between ligands in SL,; the adsorption isotherm 9 = 9@) is also computable explicitly, as are the molar and differential energies of adsorption. The case of unstable SL may also be considered. SL becomes negligible if the molar energy of attraction w is larger than some 4k, T. The characterization of this system may be considered complete. As far as the SL, case is concerned, in the general case of SL, SL, and SL, all stable, a system of two non-linear equations is obtained, which are readily solvable numerically but not analytically. This fact prevents detailed formulae to be given, in particular as far as the adsorption isotherm or adsorption energies are concerned. Some symmetry properties of the gi = 9,(9) functions may, however, be established, as may the conditions for the maxima in el(@ and 9,(9), which are always 9 = 1 and 9 = 2, respectively. Also numerically tractable are the cases where either SL or SL, is unstable. The case in which there are no interactions between particles on the same site is, instead, always analytically soluble. By generalizing the results found for the SL, and SL, cases, it turns out that, for the adsorption of n ligands per site, the isotherm is always Langmuirian : and the partition of L ligands into the various SL, species is simply given by a binomial law: Pi" being the classical binomial coefficient. 9/(n-9) = Kp 0 < 9 < n Oi = P; Oi(n - 9)n-i n-n 0 < i < n We thank the CSI Piemonte computing centre for free allowance of computer time and for use of SAS-GRAPH and IMSL packages.E. Garrone and P . Ugliengo 599 References 1 E. Garrone, G. Ghiotti, C. Morterra and A. Zecchina, Z . Naturforsch., Ted B, 1987, 42, 728. 2 G. Ghiotti, E. Garrone, G. Della Gatta, B. Fubini and E. Giamello, J. Catal., 1983, 80, 249. 3 G. Ghiotti, E. Garrone and A. Zecchina, J . Mol. Catal., 1988, 46, 61. 4 B. Rebenstorf and R. Larson, J . Mol. Catal., 1982, 11, 247. 5 I. Burckhardt, D. Gutschick, U. Lohse and H. Miessner, J . Chem. SOC., Chem. Commun., 1987, 291. 6 J. S . Wang, Proc. R . SOC. London, Ser. A , 1937, 161, 127. 7 R. H. Fowler and E. A. Guggenheim, Statistical Thermodynamics (University Press, Cambridge, 1960), p. 441. Paper 8/008 19A ; Received 29th February, 1988

ISSN:0300-9599    

DOI:10.1039/F19898500585

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1989, 85(3), 601-607 Electron Spin Resonance Investigation of the Copper@)- P-Glucosidase Interaction in Aqueous Solution Franco Laschi" and Claudio Rossi Department of Chemistry, University of Siena, Pian dei Mantellini 44, Siena, Italy Electron spin resonance spectroscopy has been used to investigate the coordination behaviour of /?-glucosidase aqueous solutions in the presence of divalent metal ions such as copper(I1). The strong affinity of the paramagnetic metal ions towards the enzyme was verified; two metal- enzyme interaction sites were characterized as a function of pH and temperature and a NJO, copper(I1) atom donor set proposed. In the presence of cellobiose solutions the Cu"-/?-glucosidase electron spin resonance parameters were unaffected in the range of pH and molar ratios investigated.The enzymatic hydrolysis of the P-glucoside linkage is due to the presence of /3-glucosidase. A full understanding of the mechanism of P-glucoside bond cleavage by cellulolytic enzymes is crucial in determining the complete pattern of the hydrolysis cellulose to glucose. An analysis of the metal-ion-/3-glucosidase interaction and its possible role in enzymatic activity provides a key to understanding the role of metal ions as enzymatic modulators. For the investigation of biomolecular metal complexes, magnetic resonance is the most suitable technique because of its non-invasive nature.2 In particular, electron spin resonance (e.s.r.) is a very specific and widely applied technique for studying the interaction of paramagnetic metal ions such as Cu" with macro- m o l e c u l e ~ .~ ~ ~ The aim of this paper is to contribute to a deeper understanding of the nature and the mechanisms of the Cu"-P-glucosidase interaction, as well as the extent and the limits of the copper-nzyme chemical equilibrium, in an attempt to clarify the hydrolytic process of cellulose. Experimental /3-Glucosidase (P-D-glucoside glucohydrolase ; FC 3.2.1 .21) enzyme and cellobiose [D( + )cellobiose, 4-O-~-~-glucopyranosyl-~-glucopyranose] were obtained from Sigma ; both were used without further purification. Cu(C10,);6H20 (Aldrich) was used as a source of Cu" ions in order to minimize anionic complexation. The pH was varied by adding KOH to the copper(1r) aqueous solutions, and the pH values were measured with a Metrohm model E-388 potentiometer.E.s.r. spectra were recorded with a Bruker ER 200D-SRC spectrometer operating in the X-band (w, = 9,78 GHz). The temperature was controlled with a Bruker ER 41 11 VT variable-temperature assembly (accuracy 1 K). The magnetic parameters were measured by direct field calibration with the diphenylpicrylhydrazyl (DPPH) free radical ( g = 2.0036) as field marker. In order to have quantitative reproducibility, the solutions were introduced into a calibrated quartz capillary permanently positioned in the resonance cavity. 60 1602 E.S. R . Study of the Cu"-b-Glucoside Interaction Results and Discussion General Remarks on the E.S.R. Copper(r1) Spectra In this paper the Cu"-B-glucosidase interaction is analysed in terms of the X-band e.s.r.hyperfine structure of the aqueous-solution copper(1r) spectra. The 3d9 copper( 11) ion can be suitably treated as a 'd" unpaired system and its e.s.r. spectrum interpreted in terms of the corresponding general Hamiltonian :5 In eqn ( I ) the first term refers to the Zeeman interaction between the applied external magnetic field ( H ) and the unpaired electron spin (S). The second term refers to the nuclear hyperfine energy, taking into account the magnetic interaction between the copper(rr) ion nucleus (4" = $) and the unpaired electron ( S = i). In eqn ( I ) H, I and S are expressed in vector notation and g and a in tensor notation.',' E.s.r. theory forecasts for the Cu" hyperfine interaction a pattern of four lines of equal intensity ( N , the number of hyperfine lines, is given by 21+ 1 = 4) equally spaced by a, the so-called hyperfine coupling constant, which is assumed to be a measure of the entity of the isotropic interaction between I and S.The rapid random tumbling of the copper species in solution rules out the original anisotropic nature of g and a in eqn (I), so that room-temperature copper(1r) solutions show characteristic isotropic e.s.r. spectra with single values of both the g and a terms. On the other hand, liquid-nitrogen-temperature (100 K) spectra reveal the anisotropic nature of the g and a tensors as a consequence of the inhibited motion of the paramagnetic metal-ligand complex. In the case of axially symmetric structures the explicit form of the Hamiltonian can be written as6 The first term in square brackets is the Zeeman interaction with anisotropy in g referred to the parallel component (11 = z axis) and the perpendicular component (I = x y plane). The second and third terms represent the nuclear hyperfine interaction with anisotropy in a referred to the z axis and the x y plane, respectively.E.s.r. theory states that there are changes in the magnitude of the a hyperfine splitting as well as in the value of the g-factor, depending on the nature and the number of the ligand atoms in the first coordination sphere of the copper(I1) ion;'" the analysis of such variations can be suitably carried out both at room temperature and in frozen solutions in order to obtain information about the structure and the stability of the copper(I1) complexes.This approach is particularly useful in tentatively proposing the nature and number of binding atoms lying in the first coordination sphere of the copper(1r) The Cu"-/l-Glucosidase System Fig, 1 shows the observed changes in the X-band e.s.r. lineshape of the CU"- a-glucosidase binary system in aqueous solutions at room temperature as a function of the pH ([Cu"] = 8 x mol dm-3; ~-glucosidase] = 20 g dmP3). A lineshape analysis suggests the presence of two different Cu"-P-glucosidase interaction sites in the pH range 4.0-6.8 [fig. 1 (b)-(d)]. The corresponding isotropic e.s.r. parameters are reported in table 1 ; they support the existence of two different enzyme-atom donor sets directly interacting with the metal ion. Here the A and B notation refers to the two copper(I1) binding pathways.F. Laschi and C.Rossi H I DPPH - 603 Fig. 1. X-Band e.s.r. spectra of Cu"-P-glucosidase solution for various pH values. T = 300 K ; [Cu2+] = 8 x mol dm-3; @glucosidase] = 20 g dmP3. Table 1. X-Band e.s.r. isotropic u and g para- meters of the Cu"-a-glucosidase A- and B-type interactions in a Cu"-a-glucosidase solution ( T = 300 K) complex QisolG giso PH [Cu"-b-Glu], 61.3 2.150 5.5 [Cu1'-/3-Glu], 73.3 2.121 6.8 [Cu2+] = 8 x mol dmP3; [P-glucosidase] = 20 g dm-3; T = 300 K. The presence of two Cu"-P-glucosidase coordination sites can be explained by taking into account the following chemical equilibrium : +OH- Cu1'-H20 +P-Glu [CU"-P-G~U], + H 2 0 $ [Cu"-P-Glu],. (3) K A K , An analysis of the spectra in fig.1 (b) and (c) shows that the A-type complex is difficult to characterize as a pure form owing to the presence of large amounts of the Cu"-H,O604 E.S.R. Study of the Cu"-P-Glucoside Interaction H - \DPPH 6.8 Fig. 2. X-Band e.s.r. spectra of a Cu"-8-glucosidase frozen solution for various pH values. T = 100 K ; [Cu"] = 8 x mol dm-3 ; [P-glucosidase] = 20 g dm-3. complex which produce a signal that overlaps that of the A-type complex; in contrast, on increasing the pH the [Cu"-~-gluc~sidase]~ complex becomes the major species, and is largely predominant in the pH range 6.0-7.2. High P-glucosidase concentrations cause pH changes: for 18-glucosidase] = 70 g dm-3 the corresponding pH is 7.2. Under these experimental conditions and in the presence of copper(r1) ions there is no evidence of the Cu"+OH), light-blue precipitate : this means that the copper(1r) is totally involved in the direct interaction with the enzyme at acid as well as at neutral pH values.Moreover, on decreasing the pH the lineshape varies from that shown in fig. 1 (c) to that in fig. 1 ( a ) : this demonstrates that the equilibrium (3) is reversible with pH. These findings can be explained assuming that the P-glucosidase enzyme acts as an effective weak base, making previously protonated coordination sites available to the copper ions ; the competing precipitation of copper(r1) hydroxide is therefore ineffective. The lineshape analysis of fig. 1 (b)-(d) does not show superhyperfine Cu"-N splittings at room temperature: it is reasonable to suppose that the relatively large linewidth cancels possible superhyperfine contributions. Further information about the extent of equilibrium (3) and some structural evidence of the different characteristics of the two previously identified metal-enzyme interaction sites can be obtained by recording e.s.r.spectra at liquid-nitrogen temperatures (100 K). Assuming axial symmetry for both the copper(r1)-8-glucosidase A and B sites as the lineshape analysis suggests, several hypotheses concerning the composition of the ' atom donor set' in the xy equatorial plane can be verified by taking account of the so-called Peisach-Blumberg appr~ach.~.' In this approach, depending on the nature of the fourI;. Laschi and C. Rossi 605 Table 2. X-Band e.s.r. anisotropic a and g parameters of the Curr-a-glucosidase B-type interaction in a Cu'I-lJ-glucosidase frozen solution ( T = 100 K) complex QlG g [ C u '-lJ- GI u] R( ex p t 1) g,, = 2.236 [CuT'-j?-Glu]R(calcd) U, = 13.0 g, = 2.063 a,, = 198.2 [CU ' '-P-GI U] uiso = 73.8 giso = 2.121 [Cu"] = 8 x mol dm-3; [P-glucosidase] = 50 g drnp3; pH 6.8; T = 100 and 300 K.The a, and g , parameters are obtained by the following formulae: 'is" = gall + 2'1) ; giso = Skll + 'g,). coordinating atoms in the xy equatorial plane around the metal ion, a close relationship between the experimental g and a spectral parameters can be inferred, particularly in the case of N and 0 coordinating atoms. Our experimental data in frozen solutions permit an evaluation of the g and a parameters for the [C~'~-/3-glucosidase] B-type interaction site [fig.2(c)], while the A-type site has not been characterized as a pure form [fig. 2(a)]. Table 2 reports the B-type frozen solution parameters. The well resolved spectrum of fig. 2(c) can be interpreted, in agreement with the previous discussion, on the basis of a total shift to the right equilibrium (3). A lineshape analysis of the B-type spectrum does not show the parallel low-field component (MI = -$) resolution of the single absorption lines of the copper isotopes:lo'" this means that the width of both the MI = -: parallel absorption lines is larger than the main separation between the two corresponding Cu63 and C d 5 absorption lines ( C U ~ ~ = 69.1 %, Cu65 = 30.9 % in natural abundance). In this situation the two lines overlap, giving rise to a single absorption.It is known that the largest copper isotopic separation is 6-10 G at room temperature and 1 1-1 5 G at liquid-nitrogen temperature;12 in our discussion this value can be assumed as the lower limit for the linewidth of the parallel components : where B~us3-cus5 is the separation between the MI = - $ C U ~ ~ and MI = - - $ C U ~ ~ parallel absorption lines. On the other hand, from table 2 it is evident that in the perpendicular region of the e.s.r. spectrum the value of the calculated hyperfine constant (a, = 13.0 G) is less than the true value of the corresponding [Cu"-P-Glu], linewidth : a,([Cu"-/3-glucosidase],) < AH,([Cu"-/3-glucosidase],). ( 5 ) In such a situation, neither a hyperfine or superhyperfine splitting is evident in the perpendicular region of the spectrum.Recently a large number of copper(u) biomolecular species have been analysed on the basis of the Blumberg-Peisach approach' in order to verify the actual atom-donor pattern. Taking into account the hypothesis for 0 and N donor atoms, the Cu" ion could experience the interaction of four binding atoms from the N,/O, to N,/O, coordinating set in the equatorial plane. Assuming that for the [C~~'-/3-glucosidase]~ form the coordination site is characterized by a neutral or single negative total charge (pH 6.5), the ai,/gil ratio (88.7 G) can be explained by a N3/Ol equatorial atom-donor set. 21 FAR I606 E.S.R. Study of the Cull-P-Glucoside Interaction Table 3. X-Band e.s.r. a and g parameters of Cu"-p-glucosidase<ellobiose systems at room and liquid-nitrogen temperatures complex g,,, a,s,lG PH ~ -.-____.__ T = 298 K Cu'I-B-Glu + Cell 2.121 73.3 6.8 Cull-Cell +p-Glu 2.120 73.8 6.5 Cu" + Cell-P-Glu 2.120 73.2 6.8 T = 100 K Cu1'-/3-Glu + Cell 2.235 198.2 6.8 CuII-Cell +P-Glu 2.238 195.0 6.5 Cull +Cell-p-Glu 2.236 195.0 6.5 ~ [Cu2+] = 8 x mol dm-3; [P-glucosidase] = 50 g dm-3. Cu"-Cellobiose Addition of cellobiose to copper(1r) aqueous solutions at room temperature in the pH range 4.0-6.8 does not alter the e.s.r. parameters characteristic of the Cu''-H,O system. On the other hand, pH values > 6.2 cause loss of the absorption signal due to the massive precipitation of Cu''-(OH),. The analysis of a number of e.s.r. measurements performed in the pH range 4.0-6.8 at various ligand metal molar ratios and at liquid-nitrogen temperature does not show resolution of the broad copper(r1) glassy spectrum.This fact excludes any direct binding interaction between the metal ion and the sugar. Cu"-fl-Glucosidase-Cellobiose System In order to characterize the influence of the cellobiose molecule on the Cull- /?-glucosidase binary system, e.s.r. measurements have been carried out by adding increasing amounts of cellobiose to the Cur'-/?-glucosidase system (i) and by adding /?-glucosidase to the Cu"-cellobiose system (ii). The aim of this double treatment was to reveal a possible competing interaction between the metal ion and the sugar molecule towards the enzyme. The two series of experiments (i) and (ii) were performed at room and liquid-nitrogen temperatures and at three different pH values (4.5, 5.5 and 6.8).Under the same experimental conditions (T, pH and [Cu2']) the e.s.r. analysis of the corresponding solutions (i) and (ii) showed that (a) the e.s.r. features of solutions (i) and (ii) are almost identical both at room and liquid-nitrogen temperatures and (h) pH variations in the range 4.0-6.8 do not induce changes in the e.s.r. parameters of (i) and (ii) with respect to those of the Cu"-P-glucosidase binary system at the same metal-enzyme molar ratios. Further e.s.r. experiments were carried out by adding Cu" solutions to the P-glucosidase-cellobiose system at pH 4.5, 5.5 and 6.8 (iii). The experimental data of system (iii) strongly agree with those of solutions (i) and (ii); table 3 summarizes the e.s.r.parameters of the Cu"-/?-glucosidase + cellobiose system (i), the CuII-cellobiose + P-glucosidase system (ii) and the Cu" + cellobiose-P-gluco- sidase system (iii), both at room and liquid-nitrogen temperatures and at pH 6.8.F. Laschi and C. Rossi 607 Conclusions The e.s.r. analysis of the experimental results carried out on the Cu"-P-glucosidase binary system and on the two Cu"-P-glucosidase-cellobiose ternary systems (i) and (ii) underlines the strong affinity between the enzyme and the copper(r1) ion and allows one to identify structurally different copper(n) coordination sites on the enzyme molecule. The marked pH dependence of the binding activities in the pH range 4.0-6.8 is characterized in terms of e.s.r. lineshape analysis and both isotropic and anisotropic g and a spectral parameters. The e.s.r.technique provides experimental verification of the ineffectiveness of the cellobiose interaction on the [Cu"-~-glu~osidase]~ and [Cu"-/?-gl~cosidase]~ systems, even at high sugar concentrations; moreover, our e.s.r. analysis excludes any direct interaction between the copper(i1) and the cellobiose molecule in the pH range considered. On the basis of these findings we conclude that in the pH range 4.G6.8 two different copper(u)-P-glucosidase interaztion sites are active, each characterized by significantly different e.s.r. parameters, i.e. by different complexation pathways ; in particular, for the [Cu"-~-gl~cosidase]~ form a N,O atom-donor set can be proposed for the equatorial coordination. It is clear that these two metal-ion binding sites are not related to the active site between the substrate and the enzyme.Our results also suggest that metal complexation in the limits of our experimental conditions does not affect the hydrolytic activity of the enzyme. This study was supported by CNR ' Progetto Finalizzato Energetica ', grant no. 87.02295.59. References 1 G. C. Argerinos and D. I. C. Wang, in Annual Reports on Fermentation Processes, ed. G . T. Tseo, 2 (a) L. D. Campbell and R. A. Dwek, Biological Spectroscopy (Benjamin, New York, 1984); (6) Metal 3 T. VLnngard, in Biological applications of E.S.R., ed. H. M. Swartz and D. C. Borg (Wiley Interscience, 4 G. Rotilio, in E.S.R. and N.M.R. of Purumagneric Species in Biological and Related Systems, ed. R. S . 5 A. H. Maki and B. R. Mac Garvey, J . Chem. Phys., 1958, 29, 31. 6 A. Abragam and M. H. L. Pryce, Proc. R. Soc. London, Ser. A, 1951, 205, 135; 1957, 206, 164. 7 J. Peisach and W. E. Blumberg, Archiv. Biochem. Biophys., 1974, 165, 691. 8 N. N. Tikhomirova and K . I. Zamaraev, J . Struct. Chem., 1983, 4, 200. 9 F. Laschi, M. P. Picchi, C. Rossi and R. Cini, Inorg. Chim. Acta, 1987, 135, 215. M. C. Fliekeringer and R. K. Flinn (Academic Press, New York, 1980), vol. 4, p. 165. Ions in Biological Systems, ed. H. Siege1 (Dekker, New York, 1987), vol. 21 and 22. New York, 1972). Drago and I. Bertini (Reidel, Dordrecht, 1979). 10 T. Vannegard and S. Akestrom, Nature (London), 1959, 184, 183. 1 1 B. A. Goodman, D. B. McPhail and H. K. J. Powell, J . Chem. Soc., Dalton Trans., 1980, 822. 12 ((1) D. R. Flentge, J. H. Lunsford, P. A. Jacobs and J. B. Vytterhoeven, J . Phys. Chem., 1975,79, 354; ( h ) Y . Y . H. Chao and D. R. Kearns, J. Phys. Chem., 1977, 81, 666. Paper 8/00957K; Receitjed 9th March, 1988 21-2

ISSN:0300-9599    

DOI:10.1039/F19898500601

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1989, 85(3), 609-620 Photocatalysed Isomerization of Butenes on MgO Powders with Coordinatively Unsaturated Surface Ions Masakazu Anpo* and Yoshiaki Yamada Department of Applied Chemistry, College of Engineering, University of Osaka Prefecture, Sakai, Osaka 591, Japan Salvatore Coluccia and Adriano Zecchina Dipartimento di Chimica Inorganica, Chimica Fisica e Chimica dei Materiali, Universita ’ di Torino, Corso Massimo d’Azeglio 48, 10125 Torino, Italy Michel Che Laboratoire de Rkactivite‘ de Surface et Structure, Universite‘ P . et M. Curie, U.A. 1106 - CNRS, 4 Place Jussieu, Tour 54, 75252 Paris Cedex 05, France The photocatalysed isomerization of but-2-enes has been studied on a standard MgO-I catalyst which exhibits photoluminescence at ca.340- 450 nm associated with coordinatively unsaturated surface ions with 4-coordination. A geometrical isomerization reaction, i.e. trans + cis iso- merization, occurs predominantly with high efficiency under U.V. irradiation of the MgO catalyst. The excitation wavelength that is effective on the rate of photocatalysed isomerization was found to be in accordance with the photoluminescence excitation spectrum. A complete parallel between the photoluminescence intensity and the rate of photocatalysed isomerization of but-2-enes was also observed. The addition of CO molecules led to the inhibition of the photocatalysed isomerization, its extent increasing with the CO pressure, as well as quenching of the photoluminescence. Stern-Volmer plots for the rate of the photocatalysed isomerization and for the yield of the photoluminescence showed complete agreement.These results clearly indicate that 4-coordinated surface ions play a significant role in the photocatalytic activity of degassed MgO powder catalysts. Photocatalysis with powdered semiconductors such as TiO,, Pt/TiO, and CdS etc. has received much attention from the standpoint of photochemistry of solid surfaces as well as potential utilization of solar energy.’ Recent studies focus on problems with extremely small size colloidal semiconductors where size quantization effects are expected to be observable,2 as well as on semiconductors supported on inert supports, where carrier effects are e~pected,~ and on various binary metal oxides and sulphides, where complex effects are ~perating.~ However, most studies of photocatalysis have mainly been undertaken with reactions associated with the photodecomposition of water over semiconductors in the solid-liquid reaction system, in which it is relatively easy to set up an experimental system and to achieve higher quantum yields compared with those in the solid-gas reaction systems.Consequently, very few studies have been concerned with surface-active sites, excited states of catalysts and primary processes in photo~atalysis.~ On the other hand, it is well established that surface ions at the position of coordinative unsaturation play a significant role in heterogeneous catalysis.6 Such coordinatively unsaturated surface ions are also well known to play a significant role in the appearance of abnormal absorption and photoluminescence in insulating materials with a high surface area, such as powdered MgO and SrO.’ However, their role is still unclear in photocatalysis.To understand the true nature of photocatalysis, it is 609610 Photocatalysed Isomerization of Butenes on MgO important to clarify the role of such surface ions in photocatalytic reactions, especially in primary processes. Recently, Anpo et al.8 have reinvestigated the photoluminescence spectrum of powdered MgO degassed at high temperatures using a standard JRC-MgO-I catalyst because the contribution of low-coordination surface sites to the observed photo- luminescence of the degassed MgO samples has been questioned by Shvets et al.9 Anpo et a1.8 have clarified that the observed short-lifetime ( 10-6-10-3 s) photoluminescence spectra at ca.340-450 nm are directly associated with the following charge-transfer process in reaction (I) (LC means low coordination) on the 4-coordinated ions located on the surfaces of the well degassed MgO powders, reconfirming the charge-transfer mechanism proposed by Tench and Pott," Stone'' and Cunningham et al:12 hv (Mg~-O2,-,)~(MgtC-O,,)*. hv' Although little attention was paid on the photocatalysed isomerization reactions, recently, various photocatalysed isomerization reactions have been studied by a number of worker~.'~-'~ More recently, Anpo et al. have reported that the double-bond shift isomerization as well as cis + trans isomerization of but-2-ene is markedly enhanced on metal oxides, such as TiO,," binary oxides, such as Ti-Si" or Ti-Al,'' and the highly dispersed supported metal oxides such as V205,20 under U.V.irradiation. For these photocatalysed isomerization reactions, it has been suggested that the pair state of photoproduced electron and hole, (M('+l)+- O-)*, plays a significant role in weakening the C=C double bond of butene molecules to result in their isomerization. It has been also found that such a cis'trans isomerization of but-2-ene is photocatalysed on degassed MgO 22 Therefore, it is of particular interest to investigate the isomerization of but-2-ene on a standard JRC-MgO-I catalyst under U.V. irradiation in order to clarify the role of coordinately unsaturated surface ions in the photocatalysed isomerization reaction, studying the relationship between the intensity of photoluminescence and the rate of pho t oca tal ysed isomerization.Experimental Samples of MgO microcrystals (JRC-MgO-I) were supplied from the Catalysis Society of Japan as a standard catalyst (MgO purity, 99.02 '/O ; major impurities, Ca, Si and Fe; B.E.T. surface area, ca. 40 m2 8-l; bulk specific density 0.42 g ~m-~).', 21, 22 Detailed information about this standard MgO-I catalyst is available from the Catalysis Society of Japan.23 MgO catalysts were degassed for 2 h at the desired temperature. The rate of increase of the degassing temperature was ca. 1 K min-' and ultimate pressures of ca. 1 0-6 Torr ( I O-* NmP2) were attainable. The photoluminescence spectra were recorded at 77 K or at 295-298 K using a Shimadzu RF-501 spectrofluorophotometer (equipped with 500 W Xe lamp as an excitation source) with a resolution of 0.3 nm, equipped with colour filters to eliminate scattered light.The quartz cell with window and furnace sections, having a volume of ca. 50 cm3, was connected to a vacuum system. A 0.05-0.1 g sample was degassed for 2 h at the desired temperature. The rate of increase of the degassing temperature was also ca. 1 K min-l. Then the MgO powder was spread on the quartz window, having a surface area of ca. 20 cm2. The reactant butene was introduced onto the catalyst, its temperature being adjusted to 273 K. Then, U.V. irradiation of the catalyst in the presence of trans-but-2- ene (6 Torr) was also carried out at 273 K, to minimize the dark reaction, using a high- pressure mercury lamp (Toshiba, SHL- lOOUV) through various colour filters and a water filter.The reaction products were collected at 273-293 K at definite time intervals, and analysed by a gas chromatography using a 2,4-dimethylsulpholane column. InM . Anpo et al. 61 1 I I 200 300. 300 400 500 w avelengthhm Fig. 1. (a) Photoluminescence spectrum and (6) its corresponding excitation spectrum at 298 K of MgO sample degassed at 1073 K for 2 h and effect of the addition of trans-but-2-ene of the photoluminescence spectrum. Excitation, 240+ 10 nm; trans-but-2-ene added at 298 K ; amount of trans-but-2-ene, lop6 mol g-l, ( 1 ) under vacuum, (2) 1.16, (3) 2.32, (4) 7.87, (5) 22.2. addition to the photoformed cis-but-2-ene and but- 1-ene, small amounts of CH,, C,H, and C2H6 were also observed as minor products.At 273 K, thermal isomerization (dark reaction) was negligible. Experimental details were described elsewhere.8* 17-22 Results and Discussion Quenching of the Photoluminescence with Added But-2-ene As shown in fig. 1, the degassed MgO sample exhibits a photoluminescence spectrum at ca. 340-450 nm when it is excited with U.V. light with 240-280 nm wavelength, owing to the charge-transfer process described in reaction (I).8 Fig. 1 also shows that the addition of trans-but-2-ene at 293 K onto the MgO sample that had been degassed at 1073 K leads to efficient quenching of the photoluminescence intensity without any change of shape, its extent increasing with the pressure of trans-but-2-ene. For example, the addition of only 1.16 x mol g-' trans-but-2-ene quenches the photoluminescence by ca.66%, the addition of ca. 7.9 x mol g-' of trans-but-2-ene leading to 92 YO quenching of the emission. As shown in fig. 1, the evacuation of trans-but-2-ene at 298 K for 30 min, after complete quenching of the emission, leads to a recovery of most of the emission. As described 22 such quenching behaviour suggests that dynamic quenching, whereby quencher molecules interact with the emitting sites in their metastable excited state to give a non-radiative deactivation pathway, mainly operates for quenching with butenes. Therefore, the quenching efficiency is dependent on the amount of quencher adsorbed on the surfaces, and in turn the equilibrium pressure of the added butene molecules. Photocatalysed Isomerization of trans-But-2-ene As shown in fig.2 and 3, the U.V. irradiation of a degassed MgO sample in the presence of trans-but-2-enes at 273 K is found to lead to a remarkable enhancement of the formation of cis-but-2-enes, as well as the formation of a small amount of but-1-enes.612 Photocatalysed Isornerization of Butenes on MgO 10 20 30 40 U.V. irradiation time/min Fig. 2. Photocatalytic (solid lines) and thermal (dotted lines) isomerization of trans-but-2-enes on MgO catalyst degassed at 1173 K for 2 h. (a) cis-But-2-ene, (b) but-1-ene. MgO used, 0.05 g; initial pressure of trans-but-2-enes7 6.0 Torr ; reaction temperature, 273 K. 0 10 20 30 40 U.V. irradiation time/min Fig. 3. Photocatalytic (solid lines) and thermal (dotted lines) isomerization of trans-but-2-enes on an MgO catalyst degassed at 1273 K for 2 h.(a) cis-But-2-ene; (b) but-I-ene. MgO used, 0.05 g; initial pressure of trans-but-2-enes, 6.0 Torr ; reaction temperature, 273 K.M. Anpo et al. 613 273 473 673 073 1073 1273 degassing temperatureIK Fig. 4. Effect of degassing temperature of MgO catalysts upon the initial rate of photocatalysed isomerization of trans-but-2-enes at 273 K. 0, cis-But-2-ene; 0, but- I -ene. It was found that the extent of enhancement is strongly dependent on the degassing temperature of the MgO samples. The yields of photoisomerization products increase with U.V. irradiation time. With the MgO samples degassed at higher temperatures, the yields of photoisomerization products increase linearly with U.V.irradiation time, indicating that the reactions proceed catalytically under U.V. irradiation at 273 K. Fig. 2 and 3 also show that at this temperature, with these MgO samples degassed at higher temperatures, thermal isomerization (dotted line) is very small, and is negligible in comparison with photoisomerization, especially for geometrical isomerization. 21 Thus the major photocatalysed isomerization product is the geometrical isomer, cis-but-2-ene. The formation of but- 1-ene, i.e. double-bond shift isomerization, is a minor reaction. CH, and C,H, were also observed as minor products. Small amounts of but-2-enes were irreversibly adsorbed onto the catalyst, which may correspond to the unrecovered fraction of the photoluminescence measured after the evacuation of added but-2-ene molecules, as described above.These features were common for all MgO samples which were degassed at various temperatures. Relationship between Photocatalysed Isomerization and Photoluminescence The initial rates of the photocatalysed isomerization determined from the slopes of the plots of the yields of products us. U.V. irradiation time are markedly dependent on the degassing temperature of the MgO samples. The results are shown in fig. 4. The initial rate of geometrical isomerization increases, passing through a maximum at 1 173 K, and then decreases on increasing the degassing temperature. On the other hand, the rate of double-bond shift isomerization increases, passing through a maximum at 773 K, and then decreases on increasing the degassing temperature up to 1073 K, again increasing a little at the highest degassing temperature.As described in a previous paper (fig. 9,' the trend of the effects of the degassing614 Photocatalysed Isomerization of Butenes on MgO 273 473 673 073 1073 1273 degassing temperature/K Fig. 5. Effect of degassing temperature on the photoluminescence spectrum of Mg0.8 *, MgO was degassed for 3 h rather than 2 h. temperature upon the intensity and wavelength of the photoluminescence of the MgO samples is very similar to that seen in fig. 4. From fig. 4 and 5 it is clear that a good parallel exists between the initial rate of photocatalysed geometrical isomerization and the photoluminescence intensity. Although details have been described in the previous paper,' the results in fig.6 show that the concentration of the low-coordinated surface ions produced increases with increasing degassing temperature up to 1173 K, and then decreases with increasing temperature. Simultaneously, with increasing degassing temperature the emitting environment associated with the 4-coordinated surface sites becomes more uniform, the uniformity changes being complete at ca. 1073 K. Therefore, it is indicated that the isomerization of butenes proceed catalytically on the low coordinated surface sites with uniform environment, under U.V. irradiation at 273 K. On the other hand, it is also seen that a good agreement between (i) the photoluminescence intensity at ca. 420 nm, having its excitation spectrum at ca. 280 nm (linked with the existence of surface OH- ions, which was shown in fig.6 of our previous paper)8 and (ii) the rate of photocatalysed double-bond shift isomerization was also observed (vide infra). Active Sites for Photocatalysed Isomerization The effect of excitation wavelength upon the rate of the photocatalysed isomerization - - - - - A ! - _ _ _ __.__ I__-_ A f - - L - j t-. ---:-- ---a:---- a - 1 -__- C I A - - - l-?z- -t A t - - m _ _ A -c reactions was invesugaieo DY using va~ious wiuur iiiicrs. rig. o snows inc ciiec;~ 01 excitation wavelength upon the rate of the major photocatalysed isomerization, i.e. geometrical isomerization, taking place on an MgO catalyst which had been degassed at 1173 K, by using various colour filters. The spectroscopic features of the colour filters used, together with the excitation spectrum (i.e. the absorption spectrum) of the MgOM.Anpo et al. 61 5 10 I 5 10 15 20 u. v. irradiation time/& Fig. 6. Effect of excitation wavelength on the rate of photocatalysed isomerization of trans-but- 2-enes on an MgO catalyst degassed at 1 173 K for 2 h. High-pressure mercury lamp omitting light with A> 237 nm; reaction temperature, 273 K ; filters, UV-25 with 50% cut-off at 250 nm wavelength, UV-29 with 50 % cut-off at 290 nm wavelength, UV-35 with 50 YO cut-off at 350 nm wavelength. samples degassed at higher temperatures clearly showed that the excitation wavelength that is effective on the photocatalysed isomerization is < 270 nm, in accordance with the excitation spectrum (fig. 1) for the photoluminescence of the samples due to charge- transfer processes on the 4-coordinated surface sites [reaction (I)].As described in our previous paper,8 the addition of 0, or CO molecules led to the quenching of the photoluminescence of MgO samples. Therefore, the effects of adding CO molecules upon the photocatalysed isomerization reaction were investigated. Fig. 7 shows the effect of the addition of CO upon the photocatalysed geometrical isomerization on the MgO sample which had been degassed at 1073 and I173 K. The rates of the photocatalysed isomerization reaction to produce cis-but-2-ene from trans- but-2-ene are easily inhibited by the addition of CO molecules, the extent increasing with CO pressure. The addition of CO molecules was also found to inhibit the minor photocatalysed isomerization to produce but- 1 -ene, although in this case a quantitative analysis was difficult owing to lower yields of but-I-enes.These results suggest that photocatalysed isomerization reactions on the MgO catalyst proceed via the same excited state as with photoluminescence. Thus the photophysical processes on MgO surfaces in the presence of but-2-ene and CO molecules can be described as follows to extend the processes reported in our previous paper:' pho toca tal ysed reaction (k,) (Mg2+-02-) 5 (Mg+-0-)* radiationless decay (k,) deactivation by added CO ( k J .616 Photocatalysed Isomerization of Butenes on MgO 5 10 15 20 U.V. irradiation time/min Fig. 7. Effect of the addition of CO molecules on the photocatalysed isornerization of trans-but- 2-enes on an MgO catalyst degassed at 1173 K.Reaction temperature, 273 K ; initial pressure of trans-but-2-ene, 6.0 Torr; pressure of added CO molecules (Torr): (1) 0.0, (2) 0.041, (3) 0.105, (4) 0.636, ( 5 ) 3.8 1. As a result, the following Stern-Volmer equation is obtained for the rates of photocatalysed isomerization on the MgO sample using the steady-state treatment :23 Q,/Q = 1 +tk,C where Q, and Q are the yield of the photocatalysed isomerization of but-2-ene on the MgO in the absence and presence of CO molecules, respectively, and t , k, and C are lifetime of the excited active sites on the MgO surface, the quenching rate constant and concentration of added CO molecules on the surface, respectively. In fact, as shown in fig. 10, Q,/Q in the presence of CO molecules is a linear function of CO pressure in the low-pressure region, although there is some deviation from linearity at higher pressure. In other words, in the low-pressure region, dynamic quenching of the excited active sites by CO molecules mainly operates, in which the quenching efficiency is dependent on the amount of CO molecules on the surface and in turn the equilibrium pressure of the added CO molecules.8 Deviation from linearity in the Stern-Volmer plots in the higher- pressure region might be attributed to the fact that the number of CO molecules on the surface did not increase linearly with the CO pressure in the pressure range 1-3 Torr.The addition of CO at a pressure of 150 Torr onto thermally activated MgO which had been washed with boiling water prior to activation, leads to a change in the colour of the MgO sample from white to peach and/or yellow, and even to a very slow growth of radical species, although this takes a long time.24*25 Recently, Garrone and Stone26 have studied the U.V.- visible spectra developed after adsorption of H, and/or CO and showed that the adsorption of H, occurs on the MgO sample, on which the adsorption of CO proceeds to yield coloured products, and suggested that 3-coordinated oxygen ions (0;;) are involved in this adsorption. As described in our previous paper,' with theM. Anpo et al. 617 CO pressure/Torr Fig. 8. Experimental Stern-Volmer plots, Q,/Q us. pressure of CO for the rate of photocatalysed isomerization of trans-but-2-enes and the photoluminescence yield of an MgO catalyst degassed at 1173 K for 2 h.Experimental conditions as for fig. 1, 2 and 7. 0, Photoluminescence; 0, photoreaction. present JRC-MgO-I sample, such 3-coordinated surface sites are not involved, and in fact no effect of the addition of H, upon the photoluminescence due to the presence of 4-coordinated surface sites was observed. In agreement with the previous results, after the addition of CO at a pressure of 1-3 Torr onto the degassed standard MgO-I sample, no change in the colour of the white sample was observed. Fig. 8 also shows Stern-Volmer plots for the quenching of the photoluminescence intensity. Good agreement is seen between the Stern-Volmer plots for the photo- luminescence intensity and those of the yields of the photocatalysed isomerization reactions.This agreement supports the assumption mentioned above, i.e. both the photocatalysed isomerization reaction and the photoluminescence observed at ca. 340-450 nm might proceed through the same excited state of the MgO catalyst. Taking into account the fact that added CO molecules easily interact with the charge-transfer excited state of the 4-coordinated surface sites to give a non-radiative deactivation pathway,’ the results obtained in the present work clearly show that the photocatalysed isomerization reactions proceed through the charge-transfer excited state on the 4-coordinated surface sites, i.e. (Mg&-Oic)* complexes. Characteristics of the Photocatalysed Isomerization on MgO It has been reported by a number of workers that on an MgO catalyst degassed at ca.573-873 K, the thermal isomerization of butene proceeds uia n-ally1 carbanion intermediate^.^',^^ In this mechanism the key step is abstraction of hydrogen from butene molecules, which is easier with but-1-ene than with but-2-ene. This feature has been established as the characteristic of the thermal isomerization on oxide catalysts such as MgO. As described previou~ly,~’~ 18* 29 however, in the photocatalysed isomerization of butenes the reactivity of but-2-enes is much higher than that of but-1-enes. This was also confirmed in the present work with the MgO catalyst.618 Photocatalysed Isomerization of Butenes on MgO These results clearly indicate that the active species associated with photocatalysed isomerization is quite different from that in the thermal isomerization reactions.Recently, Anpo et al. reported that the charge- transfer excited complex, W-l)+- 0-)*,17929*30 on various metal oxides easily reacts with butene molecule to produce a radical intermediate species (A) resulting from the opening of the C=C double bond of the butene molecule. It is likely that the photocatalytic isomerization of trans-but-2-ene to cis-but-2-ene on an MgO catalyst proceeds via a mechanism similar to that proposed for metal oxides such as Ti0,17 and supported V,0,.207 30 trans-But-2-ene interacts with the charge-transfer excited complex, (Mgi:-O;c)*, resulting in the opening of the C=C double bond to produce species (A), which participates in geometrical isomerization to cis-but-2-ene. The adsorption site is not specified at the present time between Mg2+ or 0'- ions.In this mechanism, therefore, if added quencher molecules such as CO interacted with the charge-transfer excited complex to give a deactivation pathway, the photocatalysed isomerization would be suppressed as a result, its extent increasing with the pressure of CO molecules. CH,-CH-CH-CH, (A) I As described previously,20*29*30 on the other hand, the presence of H atoms appears to be a prerequisite for the occurrence of the double-bond shift isomerization. Although the details of the mechanism are not clear, this requirement is intuitively linked with surface OH- groups with a high acidity. Accordingly, the small increase in the rate of photocatalysed double-bond shift isomerization observed on the MgO sample degassed at ca.773 K, as seen in fig. 4, suggests the existence and contribution of the acidic surface OH- groups on the MgO surfaces. Taking into account the previous results of the photoluminescence due to surface OH- ions obtained with an MgO sample,8 it is likely that surface OH- ions in specific low-coordination sites act as acidic groups which contribute to the photocatalytic double-bond shift isomerization of but-2-enes. The selectivity of the photocatalysed double-bond shift isomerization to the photocatalysed geometrical isomerization on MgO catalysts was much lower than that of photocatalytic isomerization on a TiO, catalyst, although such selectivity more or less depends on the catalyst pretreatment.17 This feature seems to reflect the lower acidity of the OH- groups on an MgO catalyst as compared with that on a TiO, ~atalyst,~' supporting the concept mentioned above.Conclusions The present results clearly indicate not only that degassed MgO powders exhibit photocatalytic activity for the isomerization of but-2-ene molecules, but also that the rate of photocatalysed isomerization and the intensity of the observed photo- luminescence of the catalyst are closely associated with each other. The yield of the observed photoluminescence at ca. 340-450 nm, which is due to radiative decay from the charge-transfer excited state in reaction (I), scarcely changed in intensity whether it was recorded at 77 or 298 K. This suggests that the non-radiative pathway is a minor process ; otherwise the photoluminescence yield would drastically change, because non- radiative pathways are temperature-sensitive processes.23, 32 Consequently, with the present JRC-MgO-I sample, non-radiative decay processes do not appear to play a significant role in the deactivation of the photon energy absorbed by the MgO catalyst.In other words, a radiative decay process might be a major deactivation pathway to control the fate of the photon energy injected into the present MgO catalyst.M. Anpo et al. 619 From these results, together with the rexlt obtained by Coluccia33 with degassed MgO, CaO and SrO, the following generalization concerning the fate of the photon energies absorbed by the alkaline-earth oxide catalysts emerges. Ions in a high coordination have a larger number of bonds to the oxide and couple more strongly with the phonon transitions of the lattice, providing a high probability of non-radiative decay, while on the lower-coordinated surface ions non-radiative decay is less efficient and the radiative-decay pathway is more predominant.Therefore, on an oxide catalyst with lower-coordinated surface sites it is expected that higher yields for photocatalytic reactions are achieved, if reactants are present on them, as well as photoluminescence. The photoluminescence associated with the charge- transfer process, like those expressed in reaction (I), is also easily observed with highly dispersed supported metal oxides in which ions are expected to be located in coordinatively unsaturated surface sites but not in the bulk o x i d e ~ .~ , ~ ~ Taking into account these facts, a proposed generalization concerning the fate of the photon energies absorbed by the oxides would be more widely applicable for various types of oxides.'2*20 Thus the present work provides useful information not only concerning the role of low- coordination unsaturated surface ions in photocatalytic reactions on oxide catalysts, but also concerning the fate of the photon energy absorbed by the oxides. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Photoelectrochemistry, Photocatalysis and Photoreactions, ed. M. Schiavello (Reidel, Dordecht, 1984), and references therein; M. Anpo and Y. Kubokawa, in Hikari-Shokuhai (Photocutalysis), ed. Y. Kubokawa, K. Honda and Y. Saito (Asakura Shotenn, Tokyo, 1988), and references therein.M. Anpo, T. Shima, S. Kodama and Y. Kubokawa, J. Phys. Chem., 1987, 91, 4305, and references therein. M. Anpo, N. Aikawa, Y. Kubokawa, M. Che, C. Louis and E. Giamello, J. Phys. Chem., 1986, 89, 50 17 ; 5689, and references therein. M. Anpo, H. Nakaya, S. Kodama, Y. Kubokawa, K . Domen and T. Onishi, J. Phys. Chem., 1986,90, 1633, and references therein. J. Cunningham, in Comprehensive Chemical Kinetics, ed. C . H . Bamford and R. G. Compton (Elsevier, Amsterdam, 1984), chap. 3, and references therein. M. Che and A. J. Tench, Adv. Catal., 1982,31, 78, and references therein; Adsorption and Catalysis on Oxide Surfaces, ed. M. Che and G. C. Bond (Elsevier, Amsterdam, 1985), and references therein. A. Zecchina, M. G. Lofthouse and F.S. Stone, J. Chem. Soc., Faraday Trans. I , 1975, 71, 1476; F. S. Stone and A. Zecchina, Proc. 6th Int. Congr. Catal. (The Chemical Society, London, 1986), A8; A. Zecchina and F. S . Stone, J. Chem. Soc., Faraday Trans. I , 1976, 72, 2364; 1978, 74, 2278. M. Anpo, Y. Yamada, Y. Kubokawa, S. Coluccia, A. Zecchina and M. Che, J . Chem. Soc., Faraday Trans. I , 1988, 84, 751. V. A. Shvets, A. V. Kuznetsov, V. A. Fenin and V. B. Kazansky, J. Chem. SOC., Faraday Trans. I , 1985, 81, 2913. A. J. Tench and G. T. Pott, Chem. Phys. Lett., 1974, 26,590; S. Coluccia, M. Deane and A. J. Tench, J. Chem. SOC., Faraday, Trans. I , 1978, 74, 2913; S. Coluccia and A. J . Tench, Proc. 7th Int. Congr. Catal. (Kodansha, Tokyo, 1980, p. B 1154. F. S. Stone, in Surface Properties and Catalysis by Non-metals, ed.J. P. Bonnelle, B. Delmon and E. Derouane (Reidel, Dordrecht, 1983), p. 237. J. Nuran, J. Cunningham, A. M. Deane, E. A. Colbourn and W. C. Mackrodt, in Adsorption and Catalysis on Oxide Surfaces, ed. M. Che and G. C. Bond (Elsevier, Amsterdam, 1985), p. 83. H. Al-Ekabi and P. de Mayo, J. Phys. Chem., 1985, 89, 5815. S. Yanagida, K. Mizumoto and C. Pac, J . Am. Chem. Soc., 1986, 108, 647. S. Kodama, M. Yabuta and Y. Kubokawa, Chem. Lett., 1982, 1671. A. Morikawa, M. Hattori, Y. Yagi and K. Otsuka, Z . Phys. Chem. (Frankfurt am Main), 1977, 104, 309. M. Anpo, M. Yabuta, S. Kodama and Y. Kubokawa, Bull. Chem. SOC. Jpn, 1986, 59, 259. S. Kodama, H. Nakaya, M. Anpo and Y. Kubokawa, Bull. Chem. SOC. Jpn, 1985, 58, 3645. M. Anpo, T. Kawamura, S.Kodama, K. Maruya and T. Onishi, J. Phys. Chem., 1988, 92, 438. M. Anpo and Y. Kubokawa, Reo. Chem. Intermed., 1987, 8, 105; M. Anpo, T. Suzuki, Y. Yamada, M. Che, Proc. 9th Int. Congr. Catal., Calgary (1988), vol. 4, p. 1513. M. Anpo, Y. Yamada and Y. Kubokawa, J . Chem. Soc., Chem. Commun., 1986, 714620 Photocatalysed Isomerization of Butenes on MgO 22 M. Anpo and Y. Yamada, in Advances of Basic Solid Materials, ed. K. Tanabe (Elsevier, Sequoia, 23 N. J. Turro, in Modern Molecular Photochemistry, (Benjamin/Cummings, Menlo Park, 1978). 24 Data Book of 9th Reference Meeting of Catalysis Society of Japan, Toyama, 1985 (Catalysis Society of Japan, Tokyo 1986). 25 A. Zecchina and F. S. Stone, J. Chem. Soc., Faraday Trans. 1, 1978, 74, 2278; E. Gugglielminotti, S. Coluccia, E. Garrone, L. Cerruti and A. Zecchina, J. Chem. Soc., Faraday Trans. I , 1979, 75, 98; R. M. Morris and K. J. Klabunde, J. Am. Chem. Soc., 1983, 105, 2633, and their earlier series. 1988), vol. 18, p. 465. 26 E. Garrone and F. S. Stone, J. Chem. Soc., Faraday Trans. 1, 1987, 83, 1237, and earlier papers. 27 H. Hattori, Shokubai (Catalysis), 1984, 26, 250 and references therein. 28 H. Hattori, in Adrorption and Catalysis on Oxide Surfaces. ed M . Che and G. C. Bond (Elsevier, 29 S. Kodama, M. Yabuta, M. Anpo and Y. Kubokawa, Bull. Chem. SOC. Jpn, 1985, 58, 2307. 30 Y. Kubokawa and M. Anpo, in Adrorption and Catalysis on Oxide Surfaces, ed. M . Che and G. C. 31 K. Tanabe, Solid Acids and Bases (Kodansha, Tokyo, 1970); T. Seiyama, Metal Oxides and Their 32 D. M. Hercules, Fluorescence and Phosphorescence Analysis, (Wiley, New York, 1966). 33 S. Coluccia, in Adrorption and Catalysis on Oxide Surfaces, ed. M . Che and G. C. Bond (Elsevier, 34 M. Anpo, T. Shima, T. Fuji, S. Suzuki and M. Che, Chem. LRtt., 1987, 1997; M . Anpo, M. Kondo, Amsterdam, 1985), p. 319 and references therein. Bond (Elsevier, Amsterdam, 1985), p. 127. Catalysis, (Kodanshya, Tokyo, 1977), and references therein. Amsterdam, 1985), p. 59. M. Che and C. Louis, J . Lumin., 1988, 40 & 41, 829. Paper 8/01293H; Received 31st March, 1988

ISSN:0300-9599    

DOI:10.1039/F19898500609

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I , 1989, 85(3), 621-632 Stability and Structure of Formamide and Urea Dimers in Aqueous Solution A Theoretical Study Pierluigi Cristinziano and Francesco Lelj Istituto Chimico, Universita della Basilicata, Via Nazario Sauro 85, 1-85100 Potenza, Italy Pietro Amodeo, Guido Barone and Vincenzo Barone' Dipartimento di Chimica, Universita di Napoli, Via Mezzocannone 4, I-80134 Napoli, Italy The microscopic characteristics of concentrated aqueous solutions of formamide and urea have been investigated through energy minimization of clusters consisting of one or two solute molecules surrounded by up to 19 water molecules. The computations performed for single solute molecules correctly reproduce the pattern of solvent molecules in the first hydration shell found by molecular-dynamics simulations and lead to reasonable solvation enthalpies.The computations performed for two solute molecules indicate that direct CO---NH bridges are not to be expected in aqueous solution. Solute molecules are, instead, linked by water chains, leading to compact structures which are also compatible with the local tetrahedral environment of the solvent molecules. The microscopic description of solute-solute-solvent interactions in aqueous solutions of non-ionic, but polar species, such as amides or ureas, is of particular interest, especially in the medium and high concentration range, owing to the role played by urea molecules on the conformational stability of biopolymers and to the fact that amides represent a suitable model for studying both hydrophylic and hydrophobic interactions in peptides and pr0teins.l A large number of thermodynamic2-22 and s ~ ~ c ~ ~ o s c o ~ ~ c ~ ~ - ~ ~ investigations have been performed in an attempt to collect information about the physicochemical characteristics of aqueous solutions of amides or ureas and to elucidate the role of different interactions in determining their behaviour. The interpretation of experimental results has so far been based on two quali- tative models.The former model, originally proposed by S~hellman,~~ extended by K a ~ z r n a n n ~ ~ and Krescheck and S ~ h e r a g a , ~ ~ and further refined by and by Gill and F a r q ~ h a r , ~ ~ postulates the formation of solute dimers (held together by single or double hydrogen bonds) at moderate concentrations and (by analogy with the behaviour of amides in non-polar the formation both of linear and cyclic oligomers at higher concentrations.The second model, due to Frank and frank^,^' is based on the contemporary presence in water of a low-density ice-like structural domain and of a less organized, but more dense domain. Urea would then be able to modify the equilibrium between these domains because of its inability to match the regular structure of the ice-like domains. The new equilibrium condition would then be responsible for the modification of the thermodynamic properties of the solution with respect to pure water. From a theoretical point of view, a microscopic description of systems with 62 1622 Formamide and Urea Dimers in Aqueous Solution comparable solute-solute, solute-solvent and solvent-solvent interactions involves a number of problems, so that several quantum-mechanical studie~~~-~O have been devoted to the analysis of the relative strengths of different hydrogen bridges in dimers.The results strongly depend on the quality of the basis set; and basis sets that are too small are not able to provide the correct trends, and overestimate the strengths of the hydrogen bridges. The most recent computations performed for formamide dimer~'~, 50 and adducts with water48 agtee in forecasting very similar strengths (25 3 kJ mol-l) and X- - -Y distances (ca. 3 A) for the different X- - -HY bridges (H,O- - -HO, H,O- - -HN, CO---HO and CO---HN). The close similarity between the strengths of the different hydrogen bonds suggests that, as already proposed for acetamide on the basis of i.r. ~ p e c t r a , ~ ~ .~ ~ the zero enthalpy of formation of amide-amide hydrogen bonds in water is due to the equivalence of a CO---water plus an NH---water bond from one side, and a CO---NH plus a water---water bond from the other side. Incidentally we point out that other enthalpy values currently reported in the literature strongly depend on crude models used for the interpretation of the thermodynamic data. On the other hand, a comparison with experiment a1 ;esul t s30. lPs3 suggests that theoretical X- - -Y distances are overestimated by ca. 0.005 A. Small clusters containing up to five water molecules around a single molecule of f~rmamide,~' a ~ e t a m i d e ~ ~ or urea4, have also been investigated in order to analyse the structure of the first hydration shells, but the level of the computations and the procedure adopted to build the clusters (vide infra) do not allow one to draw confidently quantitative conclusions. The same remarks apply to studies devoted to the more interesting problem of the effect of methylation on the strengths of amide-amide and amide-water hydrogen b r i d g e ~ .~ ~ - ~ l The only reliable com- parisons performed so far concern formamide- - -water us. formamide- - 48 and indicate that, contrary to the results obtained using small basis sets, the strengths and geometries of the bridges remain essentially unmodified. More recently, increased computer power has permitted the investigation of the microscopic behaviour of low-concentration aqueous solutions of amides and urea by molecular dynamics (MD) and Monte Carlo (MC) numerical s i m u l a t i o n ~ .~ ~ ~ ~ ~ * Although the solvation enthalpies computed by these methods are not completely satisfactory, structural characteristics appear well defined and are only scarcely affected by different choices of the empirical potential functions used in the simulations. The situation is much more involved in the case of concentrated solutions. The results of a first simulations7 at large urea concentration (0.08 mole fraction) seem to indicate appreciable self-association of the urea molecules. However, very recent MD simu- l a t i o n ~ ~ ~ . 58 have shown that the use of over-strong solute-solute attractive interactions (based on unreliable ab initio computations), the choice of an inappropriate statistical ensemble (canonical instead of isobaric isothermal), the limited time range covered by the simulation and the initial conditions chosen in the previous MD simu- l a t i ~ n ~ ~ could have biased the results in favour of the Schellman model.Moreover, the new MD results suggest that urea dimers in solution are unstable and that bridges formed by water molecules can hold solute molecules together at quite short distances, but without significant direct interactions. These findings are further supported by experimental results concerning solutions of thiourea, which cannot dimerize via hydrogen-bond f o r m a t i ~ n . ~ ~ - ~ ~ This situation prompted us to reinvestigate the choice of potential functions in the microscopic description of amide and urea solutions and to attempt an analysis of general trends by means of a simple computational model based on geometry optimization of relatively small clusters containing one or two solute molecules and a relatively low number of water molecules.This simple approach should lead to physically significant results whenever the observables to be studied are essentially determined by the local environment of the solute molecules. The model is part of a general attempt at an analysis of general trends in solution chemistry in terms ofP. Cristinziano et al. 623 structural and energetic modifications involving only a limited number of molecules, even at the expense of some formal rigour and quantitative aspects.Computational Model The systems studied in this work include : (a) insulated dimers of formamide and urea, (b) clusters consisting of a single urea (U) or formamide (FA) molecule surrounded by an increasing (from 1 to 1 1) number of water molecules and (c) clusters consisting of two molecules of formamide or urea surrounded by up to 19 water molecules. We have described the intermolecular energy (V) in terms of pairwise potentials V(rij), which are analytical functions of the distance rii between pairs of atoms belonging to different molecules : (1) The potential functions V(r) have the form V(r) = - BrP6 + qi qj/Er. (2) v = - x V(rJ. The parameters A and B depend only on the chemical nature of atoms i and j , qi and qj are net charges, and E is an effective dielectric constant.One solute molecule was always frozen in the origin of the reference frame with the orientation shown in fig. I and 2. The motion of each additional molecule was described by the Cartesian coordinates ( X , Y , Z ) of its centre of mass and by the Euler angles ($,O,t) describing the orientation of the local reference frame with respect to the previously defined laboratory frame.62 Minimum energy configurations of the whole set of molecules were searched by the OPTIM package,63 which employs Davidon- Flet~her-Powell~~~ 65 and Newton-Raphson a1gorithmP with analytical first and (eventually) second derivatives of the energy with respect to the above coordinates. Further minimization of the gradient norm and diagonalization of the second derivative matrix has been performed in order to check that the geometrical parameters obtained for the system correspond to true minima.A preliminary aspect of any microscopic simulation concerns the choice of the atom-atom interaction parameters. These are usually determined by obtaining the best fit to either experimental data (e.g. heat of sublimation, spectroscopic properties etc.) or ab initio computations. In order to retain some advantages of both strategies, we first selected a number of different force fields able to reproduce experimental data. Then refined ab initio computations, recently performed for the systems FA-W48 and FA-FA4’9 50 (see next section), were used for a final choice. Standard bond lengths and valence angles are used for amides and correspond to those obtained by Benedetti67 from an extensive analysis of crystal structures.No rotation has been allowed around the C-N bond in view of the high torsional potential connected with this movement. Results and Discussion Insulated Pairs and Choice of the Potentials Table 1 collects the structural and energetic parameters obtained by different force fields for the structures of FA-W and (FA), systems shown in fig. 1. It is apparent that the EPEN/268 and QPEN/255 potentials do not give reliable results. In particular, the QPEN/2 potential was fitted to minimal-basis-set ah initio computations, which do not properly describe the relative stabilities of different hydrogen bonds. The other four potentials generally give satisfactory results, the only general shortcoming being an624 Formamide and Urea Dimers in Aqueous Solution (4 (b) Fig.1. Schematic drawing of the structures of the pairs FA-FA (below) and FA--W (above) considered in geometry optimizations. Fig. 2. Energy map for a representative cross-section of the U-U interaction energy hypersurface. The interactions are computed for fixed positions (A', Y ) of the mass centre of the second U molecule in the plane of the first U molecule. The distance (2) from this plane and the Euler angles of the second U molecule are optimized for each point and shown in the right-hand side of the figure. The contour lines are spaced 5 kJ mol-' up to 40 kJ mol-' above the two minima corresponding to equal interaction energies of 43.1 kJ mol-'.underestimation of hydrogen-bond distances of ca. 0.2 A with respect to refined ab i n i t i ~ ~ * - ~ ' or experimental3'* 51-53 data. It is particularly significant that the empirical potentials are able to reproduce the correct stability order between different FA-W adducts and to confirm that the cyclic (FA), dimers are bound by less than twice the energy of two acyclic hydrogen bonds. From an energetic point of view these potentials can be divided into two sets, including the very similar OPLS46~69 and AMBER7' models on one side and the nearly equivalent HHL7'* 72 and GROMOS73s 74 models on the other. The ab initio stabilization energies of different structures of the (FA), dimer are well reproduced by the HHLo potentials, whereas the OPLS functions lead to optimum hydrogen bonds ca.0.15 A shorter and 10-1 5 kJ mol-l stronger. These differences lead to a better description of the energy and density of liquid formamide by the OPLS potential^,^' but to a better description of the geometric and energetic characteristics of amide crystals by the HHL potentials.P. Cristinziano et al. 625 Table 1. Structures (distances in A) and interaction energies (in kJ mol-') for the systems (FA),, FA-W and W, (labelling as fig. 1) structure parameter ref.a QPEN/2 EPEN/2 OPLS HHL GROMOS AMBER (FA), (4 V (FA), (4 V Ro, R O N Roo Roo R N O %O FA-W (a) V FA-W (b) V FA-W (c) V w-w V 45.6 91.2 3.09 - 23.8 3.08 - 28.0 43.9 2.99 - 39.7 54.0 3.01 - 25.1 3.08 - 22.6 38.1 2.96 - - - 25.5 2.86 8.81 3.08 17.2 29.7 - 27.2 59.0 45.2 45.6 33.5 23.8 23.8 27.2 22.2 21.3 36.0 33.1 36.0 28.5 25.5 25.9 25.9 25.1 27.6 2.8 1 2.94 2.83 2.78 2.9 1 2.84 2.73 2.70 2.76 2.73 2.75 2.80 2.84 2.89 2.85 2.75 2.70 2.70 59.8 31.4 26.8 42.3 28.9 21.3 2.83 2.83 2.77 2.76 2.84 2.79 a Reference data correspond to SCF computations for (FA), [ref.(49)], spectroscopic data and CI computations for FA-W [ref. (30) and (48)] and to spectroscopic data for W, [ref. (51) and (52)l- W L L L L C L L L L < L L C < < ( ( < < < A / < < < \ - < < c 'I? r r r 7 7 - - - > 7 7 1 - ) ) 7 7 - J ) > > > J J > > > Fig. 3. Isoenergy map for a representative cross section of the U-W interaction energy hypersurface. The interactions are computed for fixed positions (X, Y ) of the centre of mass of the water molecule in the plane of the urea molecule.The distance ( Z ) from this plane and the Euler angles of the water molecule are optimized for each point and shown in the right-hand side of the figure. Contour lines are spaced 5 kJ mo1-' up to 40 kJ mol-' above the absolute minimum corresponding to an interaction energy of 33.9 kJ mol-l. On the other hand, the energetics of FA-W and W-W dimers provided by the two sets of potentials are similar and in close agreement with ab initio results. It therefore appears that HHL potentials provide a more balanced treatment of solute-solute and solute-solvent interactions, which is the key point in an analysis of the stability of solute dimers in solution. Since they are furthermore very similar to the GROMOS functions used in the most recent MD ~ i m u l a t i o n s , ~ ~ * ~ ~ we have finally decided to use only this set in all successive computations.Fig. 2 and 3 summarize the results reIative to the geometric and energetic behaviour of the pairs U, and U-W obtained by the HHL force field. Each point on the map has been obtained by minimizing the total energy with respect to four rigid-body parameters of the second molecule ( Z and the Euler angles) and keeping constant the remaining two (Xand Y ) . A comparison of the (FA), and U, systems shows that the second system has626 Formamide and Urea Dimers in Aqueous Solution Fig. 4. Schematic drawing of the most stable structures of the pairs U-U (A) and U-W (B). b b A 8 Fig. 5. Structures of the first hydration shells of urea (a) and formamide (b).a supplementary minimum corresponding to the oxygen atom of one molecule being placed between the two NH, groups of the other molecule [structure A(b) of fig. 41. The interaction energies of this last structure and of the cyclic dimer are identical (43.1 kJ mol-’) and close to the values found from the GROMOS parameters used in our recent MD ~imulation.~~ First Hydration Shells The hydration of amides and urea has been studied several times by ab initio computation^,^^^ 42-44 building the first hydration shell by sequential addition of water molecules. The same general procedure has been followed in the present study. However, since the addition of the Nth molecule could modify the positions of the other N - 1 water molecules already added, we have carefully mapped the whole space around the cluster, allowing the N - 1 solvent molecules to readjust their orientation, while the Nth molecule was used to span the space around the cluster.This last molecule was then left free to reorient in the region around the cluster where we located the deepest energy minimum. At this point a complete energy minimization was performed, allowing all N solvent molecules to move freely. Fig. 5(a) shows the structure of the cluster U-W,, which represents the first hydration shell of this solute. Note that the final shape of the cluster is significantly different from the structure one would have obtained by adding one molecule at a time without relaxing the whole geometry. It is also of interest that the positions where we found the five water molecules are very close to those obtained by MD simulation at infinite This confirms that the first hydration shell is tightly linked to the urea molecule.P.Cristinziano et al. 627 Fig. 5(b) shows the first hydration shell found for FA following the procedure outlined above. The number of water molecules is reduced to four and no water molecules are found near the aliphatic H atom. This is in partial disagreement with the quoted ab initio computations. However, it is known that the basis sets used in these computations overestimate the polarity of the C-H bond, thus leading to spurious electrostatic attractions between the H atom and the 0 atom of water. Our results are coherent with MC radial distribution functions which suggest that only two water molecules are linked to the carbonyl oxygen and that the hydrogen bonds are not bifurcated.With regard to the hydration of the NH, group, we find one water molecule for each hydrogen, in agreement with the results of the integration of the peaks of the MC radial distribution function, which indicate 0.9 water molecules for each H atom. We have further increased the number of water molecules around the amides, but in any case the water molecules are expelled from the first hydration shell and give rise to interactions with the water molecules of the first hydration shell. The energies obtained on increasing the number of water molecules around each cluster are shown in table 2. In order to evaluate the solvation energy we have also computed the total energy for a cluster of the same number of water molecules (see table 5 later) with the same procedure we used for the U-W, and FA-W, clusters.The ratio between the total energy and the total number of water molecules in the cluster reaches a stable value (ca. 37 kJ mol-') for cluster sizes of ca. 20 water molecules. Following the SCSSD mode175,76 the solvation enthalpy A 4 0 1 can be computed according to the equation = V(R, r,) - Vbulk(~!u'k) + [(A V( R ) ) - (A VbUlk)] (3) where R and Y represent solute and solvent coordinates, respectively; Y, is the minimum- energy configuration of the solvent molecules (at a given R ) ; is the minimum-energy configuration of the solvent molecules in the pure solvent; A V is the difference between V(R,r) and V(R,r,), and ( ) designates a Boltzmann average.The quantity in square brackets is neglected in energy minimizations, but its value, according to MC simulations of model The computed values of A e O l for FA and U are ca. - 50 kJ mol-', and must be compared with experimental value^'*^^^-^^ of - 58.1 & 0.6 and - 79 2 kJ mol-1 at 25 "C, and with MC values for FA of look 10 and ca. 180 kJ mol-', using OPLS46 and QPEN/255 potentials, respectively. In this connection note that the dimensions of our clusters should be sufficient to obtain converged results for solute-solvent interactions, whereas they are probably not large enough for the computation of reliable differences between the solvent-solvent interactions in pure water and in dilute solutions of FA and U.The solute-solvent contribution obtained for our FA-W,, cluster is not far from the corresponding term obtained by MC computations in ref. (46). These results suggest that the cluster approach, although very rough, is able to recover the essential geometrical and energetical features of short-range interactions in solution. is only ca. 2 kJ mol-'. Dimers in Solution In order to analyse the stability of amide dimers in aqueous solution we have performed a series of energy minimizations starting from different relative positions of the two solute molecules. Each group of minimizations was performed using two clusters containing one solute molecule with its first solvation shell (four water molecules in the case of FA and five water molecules in th% case of U).As a general procedure we started with two clusters at a distance of ca. 12 A and the two solute molecules oriented so as to form (i) a double hydrogen bridge, (ii) a single hydrogen bridge and (iii) a stacked dimer. Then the energy was minimized by freezing the distance R between the centres of mass of the two solute molecules, while relaxing all the other degrees of freedom. The628 Formamide and Urea Dimers in Aqueous Solution Table 2. Energetic contributions (in kJ mol-') for the systems FA-W, and U-W,iu ~~ ~ solu te-solven t solven t-solven t total solvation, n V V / n V V/n V V/n - 4 0 1 4 6 9 1 1 5 6 8 10 10 80.8 20.2 97.9 16.3 96.7 10.7 92.9 8.4 85.5 17.1 105.8 17.6 121.1 15.1 125.9 12.6 151.2 15.1 FA-W, 52.4 13.1 109.6 18.3 234.7 25.9 322.6 29.3 109.5 21.9 128.3 21.4 174.6 21.8 249.3 24.9 217.1 21.7 u-w, 133.2 207.5 33 1.4 41 5.5 195.0 234.1 295.8 375.2 368.3 33.3 34.6 36.8 37.8 39.0 39.0 37.0 37.5 36.8 20.5 28.0 42.3 48.5 44.4 54.6 42.6 47.2 40.1 The energies of the clusters W, are 25.1 (2)' 11 12.7 (4), 150.5 (6), 179.5 (7), 253.2 (8), 289.1 (9), 328.0 (lo), 367 (1 I), where the values of n are given in parentheses.Table 3. Energetic contributions (in kJ mol-') for the system (FA),W,u distance/A solute-solute solute-solvent solvent-solvent total R = W 0.0 161.5 104.2 265.7 cyclic dimer Rcc, = 3.6 44.8 100.8 173.2 R,,. = 3.6 4.6 197.5 138.9 RCN. = 4.2 25.9 118.0 171.2 RcK, = 6.5 4.6 151.5 167.8 Rcc, = 3.9 12.6 165.3 142.3 RcC. = 5.1 3.8 172.4 164.8 single hydrogen bond stacked 318.8 341 .O 315.5 323.8 320.1 341 .O a For each structure the first line corresponds to constrained minimization and the second one to unconstrained minimization (see text).distance was then reduced by ca. 1 A and the minimization repeated. The minimum of the 'effective potential' curve obtained by the above procedure was finally refined by relaxing all the degrees of freedom. The results obtained by this procedure for FA are shown in table 3. In the case of a constrained minimization of the cyclic diqer structure [fig. 6(a)] an FA-FA energy of 44.8 kJ mol-' was obtained at &c, = 3.6 A. Complete geometry optimization starting from this structure reduces the solute-solute energy to 4.6 kJ mol-1 and the solvent-solvent energy from 173.2 to 138.9 kJ mo1-l. These two effects are, however, overcompensated by the strong improvement of solute-solvent energy from 100.8 to 197.5 kJ mol-l, and the structure obtained [fig.6(b)] corresponds to the absolute energy minimum. The same effect is found for a single hydrogen bond [fig. 6(c)], whereas for the stacked dimer [fig. 6(d)] both solute-solvent and solvent-solvent interactions are improved by full geometry optimization at the expense of worse solute-solute interactions. The situation is essentially the same in the case of urea dimers (see table 4) for which we have considered only cases (i) and (ii) (see fig. 7).P. Cristinziano et al. 629 Fig. 6. Structures of different (FA),-W, clusters. (a) cyclic dimer, (b) absolute minimum, (c) single hydrogen bond and ( d ) stacked dimer. Table 4.Energetic contributions (in kJ mol-’) for the system U,W,, for different arrangements of the two urea molecules solute-solute solute-solvent solvent-solvent total 0.0 45.2 17.5 0.5 separate monomers 170.8 2 19.4 cyclic dimer 173.2 207.5 single hydrogen bond 237.8 179.7 minimum 299.2 136.9 390.2 425.9 435.0 436.6 The results of these computations clearly show that also in very concentrated solutions (> 10 mol dm-3) amide dimers do not involve direct solute-solute interactions, but instead a compact structure stabilized by water bridges. In fact, although amide-amide, water-water and amide-water hydrogen bridges have comparable strengths, contemporary optimization of all these interactions is not possible and the630 Formamide and Urea Dimers in Aqueous Solution Fig.7. Starting (a) and fully optimized (b) structures of the cluster U,-W,,. Table 5. Energetic contributions (in kJ mol-') for the system (FA), W," n solute-solute solute-solvent solvent-solvent solvationb 8 9 13 15 17 19 ~~ 3.8 172.4 164.8 43.9 5.3 185.4 194.1 48.1 5.9 155.2 387.0 45.2 3.2 174.1 455.2 39.7 2.7 177.8 526.8 43.1 5.1 207.1 587.9 41.8 a The energies of the clusters W, are 253.1 (8), 289.1 (9), 366.9 (1 I), 457.7 (13), 553.1 (15), 618.4 (17) and 716.7 (19), where the value of n is given in parentheses. Per FA molecule. Fig. 8. Optimized structures of the clusters (FA),-W,, (a) and (FA),-W,, (b). best compromise is obtained at the expense of solute-solute interactions. In order to confirm further this result we have studied a number of clusters (FA),-W, with n going from 8 t o 19 (table 5).Amide-amide hydrogen bridges are never formed, and the optimized structures of the clusters (see fig. 8) closely resemble those obtained including only the first solvation shell of FA. Furthermore, the solvation enthalpy for each FA molecule is essentially the same as that previously found for the clusters FA-W,, thusP. Cristinziano et al. 63 1 confirming that amide ‘dimerization’ in water would be essentially athermal. The absence of true dimers is strongly supported by the fact that the potentials used in the present study favour solute-solute and solvent-solvent hydrogen bridges with respect to solute-solvent ones since NH- - -OC, W- - -W and NH- --OH, interactions are well reproduced, whereas CO- - -W interactions are underestimated (see table 1).Concluding Remarks In this study we have performed a series of energy minimizations on model clusters designed to simulate concentrated solutions of urea and formamide. The results show that this procedure is able to reproduce correctly general trends and to provide physically significant results in a fraction of the time required by sophisticated numerical simulations. In particular the structure of the first solvation shell of amides is in remarkable agreement with that obtained by molecular-dynamics simulations, and the solvation enthalpy of FA and U are of the right order of magnitude. A thorough study of dimers in solution gives strong evidence against the formation of amide-amide hydrogen bridges. Amide-amide interactions are rather mediated by water molecules which form chains between the solute molecules that are compatible with the tetrahedral environment of each solvent molecule.These results are in good agreement with recent MD ~ i m u l a t i o n s ~ ~ . ~ ~ and extend the range of concentrations in which theoretical studies tend to exclude the formation of CO- - -HN hydrogen bridges in aqueous solution. Work is in progress to investigate the generality of this model, which strengthens the essential mediation role of solvent molecules, for the description of interactions between polar molecules in water. References 1 G. Nemethy, W. J. Peer and H. A. Scheraga, Annu. Reu. Biophys. Bioeng., 1981, 10, 459. 2 G. C. Kresheck, J. Phys. Chem., 1969, 73, 2441. 3 J. Konicek and 1.Wadso, Acta Chem. Scand., 1971, 25, 1541. 4 J. J,.Savage and R. H. Wood, J . Soh Chem., 1976, 5, 733. 5 G. Ojelund, R. Skold and J. Wadso, J . Chem. Thermodw., 1976, 8, 45. 6 C. de Visser, P. Pel and G. Somsen, J . Soh Chem., 1977, 6, 571. 7 C. de Visser, G . Perron, J. E. Desnoyers, W. J. M. Heuvelsland and G. Somsen, J . Chem. Eng. Dutu, 8 C. de Visser, W. J. M. Heuvelsland, L. A. Dunn and G. Somsen, J . Chem. Soc., Faraday Trans. 1. 1978, 9 A. Lo Surdo, C. Shin and F. J. Millero. J . Chem. Eng. Data, 1978, 23. 197. 1977, 22, 74. 74, 1159. 10 T. H. Lilley and R. H. Wood, J . Chem. Soc., Faraday Trans. I , 1980, 76, 901. I 1 R. H. Wood and L. H. Hiltzik, J. Soln Chem., 1980, 9, 45. 12 J. N. Spencer, S. K. Berger, C. R. Powell, B. H. Henning, G. S.Furman, W. M. Loffredo, E. M. Rydberg, R. A. Neubert, Ch. E. Shoop and D. N. Blauch, J . Phys. Chem., 1981, 85, 1236. 13 I. R. Tasker and R. H. Wood, J . Soh Chem., 1982, 11, 729. 14 G . Barone, P. Cacace, V. Elia and A. Cesaro, J . Chem. SOC., Faraday Trans. I , 1984, 80, 2073. 15 G. Barone, P. Cacace, G. Castronuovo, A. Cesaro and V. Elia, Fluid Phase Equil., 1985, 20, 169. 16 G. Barone, G. Castronuovo, G. Della Gatta, V. Elia and A. Iannone, Fluid Phase Equil., 1985, 21, 17 T. E. Leslie and T. H. Lilley, Biopolymers, 1985, 24, 695. 18 J. P. E. Grolier, J. J. Spitzer, R. H. Wood and I. R . Tasker, J . Soh Chem., 1985, 14. 393. 19 G. Della Gatta, G. Barone and V. Elia, J . Soh Chem., 1986, 15, 157 and references therein. 20 G. M. Blackburn, T. H. Lilley and P. J.Milburn, J . Chem. Soc., Faraday Trans. I , 1986, 82, 2965 and references therein. 21 G. Barone, G. Castronuovo, P. Del Vecchio, V. Elia and C . Giancola, Thermochim. Acta, 1987, 122, 105. 22 G. Barone, G. Castronuovo, P. Del Vecchio and V. Elia, J . Chem. Soc., Faruday Trans. 1. 1989. 85, in press. 23 I. M. Klotz and J. S. Franzen, J. Am. Chem. Soc., 1962, 84, 3461. 157.632 Formamide and Urea Dimers in Aqueous Solution 24 G. E. Walrafen, J. Chem. Phys., 1966, 44, 3726. 25 G. Kresheck and I. M. Klotz, Biochemistry, 1969, 8, 8. 26 G. Barone, E. Rizzo and V. Vitagliano, J . Phys. Chem., 1970, 74, 2230. 27 E. G. Finer, F. Franks and M. J. Tait, J . Am. Chem. SOC., 1972, 94, 4424. 28 G. Barone, G. Castronuovo, C. Della Volpe and V. Elia, J . Soh Chem., 1977, 6, 117.29 M. Kobayashi and K. Nishioka, J . Phys. Chem., 1987, 91, 1247. 30 F. J. Lovas, R. D. Suenram, G. T. Fraser, G. W. Gillies and J. Zozom, J. Chem. Phys., 1988, 88, 31 J. A. Schellrnan, C.R. Trav. Lab. Carlsberg, Ser. Chim., 1955, 29, 223. 32 W. Kauzmann, Adv. Protein Chem., 1959, 14, 1. 33 G. C. Krescheck and H. A. Scheraga, J. Phys. Chem., 1965, 69, 1704. 34 R. H. Stokes, Aust. J . Chem., 1967, 20, 2087. 35 S. J. Gill and E. L. Farquhar, J . Am. Chem. SOC., 1968, 90, 3039. 36 H. S. Frank and F. Franks, J . Chem. Phys., 1968, 48, 4746. 37 M. Dreyfus, B. Maigret and A. Pullman, Theor. Chim. Acta, 1970, 17, 109. 38 A. Johansson and P. Kollman, J . Am. Chem. SOC., 1972, 94, 6196. 39 G. Alagona, A. Pullman, E. Scrocco and J. Tomasi, Znt. J. Peptide Protein Res., 1973, 5, 251.40 A. Johansson, P. Kollman, S. Rothenberg and J. McKelvey, J. Am. Chem. SOC., 1974, 96, 3794. 41 J. E. Del Bene, J . Chem. Phys, 1975, 62, 1961. 42 Y. Orita and A. Pullman, Theor. Chim. Acta, 1977, 45, 257. 43 Y. Orita and A. Pullman, Theor. Chim. Acta, 1977, 46, 251. 44 J. F. Hinton and R. D. Harpool, J . Am. Chem. SOC., 1977, 99, 349. 45 J. E. Del Bene, J . Am. Chem. SOC., 1978, 100, 1387. 46 W. L. Jorgensen and C. J. Swenson, J . Am. Chem. SOC., 1985, 107, 1489. 47 H. Tanaka, K. Nakanishi and H. Touhara, J . Chem. Phys., 1985, 82, 5184. 48 P. G. Jaisen and W. J. Stevens, J. Chem. Phys., 1986, 84, 3271. 49 K. Sagarik and R. Alrichs, J . Chem. Phys., 1987, 86, 51 17. 50 M. Dory, J. Delhalle, J. G . Fripiat and J. M. Andre, Int. J. Quantum Chem., Quantum Biol.Symp., 51 T. R. Dyke and J. S . Muenter, J . Chem. Phys., 1974, 60, 2929. 52 J. A. Odutola and T. R. Dyke, J. Chem. Phys., 1980, 72, 5062. 53 E. D. Stevens, Acta Crystallogr., Sect. B, 1978, 34, 544. 54 R. A. Kuharski and P. J. Rossky, J. Am. Chem. SOC., 1984, 106, 5796. 55 F. T. Marchese, P. K. Mehrota and D. L. Beveridge, J . Phys. Chem., 1984, 88, 5692. 56 P. L. Cristinziano, F. Lelj, P. Amodeo, V. Barone, J. Phys. Chem., submitted for publication. 57 H. Tanaka, H. Touhara, K. Nakanishi and N. Watanabe, J . Chem. Phys., 1984, 80, 5170. 58 P. L. Cristinziano, F. Lelj, P. Amodeo and V. Barone, Chem. Phys. Lett., 1987, 140, 401. 59 G. Barone, P. Cacace, G . Castronuovo and V. Elia, Gazz. Chim. Ztal., 1980, 110, 215. 60 G. Barone, Znt. J . Quantum Chem., 1981, 19, 1137. 61 G. Barone, P. Cacace, G. Castronuovo and V. Elia, J . Chem. SOC., Faraday Trans. I, 1981, 77, 62 H. Goldstein, CIassical Mechanics (Addison Wesley, New York, 1955). 63 P. L. Cristinziano, Thesis (University of Naples, 1981). 64 R. Fletcher and M. J. D. Powell, Comput. J . , 1963, 6, 163. 65 W. C. Davidon, Cornput. J., 1968, 10, 406. 66 Routine, N. VAIIA, Harwell Library, A.E.R.E. 67 E. Benedetti, in Proc. Fifth American Peptide Symposium, ed. M. Goodman and J. Meienhofer (Wiley, 68 R. A. Nemenhoff, J. Snir and H. A. Scheraga, J . Phys. Chem., 1978, 82, 2591. 69 W. L. Jorgensen and C. J. Swenson, J . Am. Chem. SOC., 1985, 107, 569. 70 S. J. Weiner, P. A. Kollman, D. T. Nguyen and D. A. Case, J . Comput. Chem., 1986, 7, 230. 71 A. T. Hagler, E. Huler and S . Lifson, J . Am. Chem. SOC., 1974, 96, 5319. 72 A. T. Hagler and S . Lifson, J . Am. Chem. SOC., 1974, 96, 5327. 73 H. C. Berendsen, J. P. M. Postma, W. F. van Gusteren and J. Hermans, in Intermolecular Forces, 74 J. Hermans, H. J. C. Berendsen, W. F. Van Gusteren and J. P. M. Postma, Biopolymers, 1984, 23, 75 A. Warshel, Chem. Phys. Lett., 1978, 55, 454. 76 A. Warshel, J . Phys. Chem., 1979, 83, 1640. 77 D. Ferro, G. Barone, G. Della Gatta and V. Piacente, J . Chem. Thermodyn., 1987, 19, 915. 78 E. P. Egan Jr and B. B. Luff, J . Chem. Eng. Data, 1962, 11, 192. 79 C. de Visser, H. J. M. Grunbauer and G. Somsen, Z . Phys. Chem., Neue Folge, 1975, 97, 69. 722. 1987, 14, 85. 1569. New York, 1977). ed. B. Pullman (Reidel, Dordrecht, 1981), p. 331. 1513. Paper 8/01 3821 ; Received 7th April, 1988

ISSN:0300-9599    

DOI:10.1039/F19898500621

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

J. Chern. Soc., Faraday Trans. I , 1989, 85(3), 633-644 Reactivation of Zeolite and Oxide Catalysts using Nitrous Oxide Graham J. Hutchings,"? Helen Comninos, Richard G. Copperthwaite, Lawrence Jansen van Rensburg, Roger Hunter and Themistoclis Themistocleous Catalysis Research Programme, Department of Chemistry, University of the Witwatersrand, PO WITS, Johannesburg 2050, South Africa A novel procedure for the removal of carbonaceous deposits from catalysts using nitrous oxide is described. The general applicability of this method is demonstrated for zeolites H-Y and H-ZSM-5, acid-treated clinoptilolite and supported tungsten oxide as catalysts for hex- 1 -ene cracking and methanol conversion reactions. The method is particularly effective for acid-treated clinoptilolite as this material cannot be successfully reactivated using a standard oxygen method.Additionally with zeolite H-Y the procedure permits total removal of the carbonaceous deposits, which cannot be readily achieved using oxygen. For the catalysts investigated, nitrous oxide exhibited a marked variation in its reactivation activity (acid treated clinoptilolite > H-Y > H-ZSM-5 > 10% WO,/AI,O,), a feature which is not observed with the standard oxygen reactivation procedure. Studies of N,O decomposition of both coked and uncoked catalysts demonstrate that the main process occurring during N,O reactivation is the direct oxidation of the coke deposit by molecular N,O. The reactivation efficacy observed is expressed in terms of (a) the amount of coke deposited on the catalyst and (b) the structure of coke deposit which has been studied using solid-state I3C magic-angle spinning nuclear magnetic resonance spectroscopy. The successful operation of a heterogeneous catalyst on an industrial basis depends on a number of interlinked parameters,l and one which is of crucial economic importance is the useful catalyst lifetime.This is particularly important, since many industrial catalysts cannot be successfully reactivated, e.g. promoted fused Fe catalysts* for Fischer-Tropsch or ammonia synthesis fall into this category. Catalyst reactivation may not be possible owing, for example, to the loss of the active component during use (e.g. loss of Hg from acetylene hydrochlorination catalyst^,^ or the modifications of the structure of the catalysts caused by the reactivation procedure (e.g.sintering or loss of promoters). There is therefore an important requirement to understand fully this aspect of catalyst operation if future improvements are to be made. However, one class of catalyst, namely zeolites, can only be successfully operated industrially because they can be reactivated. Zeolites have found widescale operation in the petrochemical and fuels industries since their introduction as catalysts for fluid catalytic cracking4 and more recently for the transformation and synthesis of hydrocarbon^.^ During use, deposition of carbonaceous materials, referred to as 'coke', occurs within the channels of the zeolite, and this is the primary cause of catalyst deactivation.6 Zeolites are known to demonstrate shape selectivity, and this feature is thought to minimise the rate of coke deposition for some specific zeolites, e.g.ZSM-5,' but even then the effective working lifetime is only ca. 2-3 weeks for the methanol conversion reaction.8 The rates and mechanisms of zeolite t Present address : Leverhulme Centre for Innovative Catalysis, Department of I.P.I. Chemistry, University of Liverpool, PO Box 147, Liverpool L69 3BX. 633634 Reactivation of Catalysts using N20 deactivation have received considerable attention in recent years, but the important process of reactivation has, in contrast, been little studied. For zeolites, reactivation is normally achieved by oxidising the deposit with air to carbon oxides and water, using a catalyst bed temperature of 400-500 "C.We have previously demonstrated that use of ozone' as an alternative oxidising agent can be highly effective at lower temperatures (150-190 "C), and we have further shown" that ozone preferentially reacts with the aromatic coke structures. Consequently this procedure could be useful for catalytic materials that are not stable at higher temperatures. However, for large-scale reactivation of catalytic pellets in a fixed bed the diffusion of ozone is found to be a singificant factor, and it is anticipated that this reactivation procedure will only be effective for fluidised catalyst beds. Additionally, ozone only removes some of the coke, ca. 40-50 %, and this is not sufficient for some applications. In a recent communication12 we reported our initial findings for a reactivation procedure based on N 2 0 as oxidant which was shown to be more effective for total coke removal when compared with oxygen or ozone.In this paper we exemplify and extend the general applicability of this method, and discuss these results with respect to the rates of catalytic N 2 0 decomposition on both fresh and coked catalysts. Experimental Catalyst Preparation The sodium form of the pentasil zeolite ZSM-5 was prepared according to the method of Howden', with a SiO,/Al,O, mole ratio of 35. The Na-ZSM-5 was converted into the protonated form (H-ZSM-5) using a method previously described' which involves a total of three separate ammonium-ion exchange and calcination treatments. Samples of partially exchanged ZSM-5 were also retained for catalytic testing.Clinoptilolite was obtained as a natural zeolite from Zululand, South Africa, and contained 75-80 mass YO clinoptilolite with minor amounts of quartz, cristobalite and potassium feldspar (composition : Si/Al ratio = 5.0, Fe 0.83 YO, Ca 0.98 YO, Na 0.88 %, K 3.03 %, surface area = 21 m2 g-'). Natural clinoptilolite (100 g) was treated with aqueous HCl (2 mol dmP3, 1 dm3) for 8 h at 75 "C to convert the zeolite into the acid form. This procedure was repeated three times, following which the zeolite was recovered by filtration, washed with distilled water, dried in air (120 "C, 3 h) and calcined (400 "C, 16 h). The composition of the acid-treated clinoptilolite was as follows: Si/Al ratio = 9.5, Fe 0.67%, Ca 0.30%, Na 0.05 %, K 1.37%, surface area = 85 m2 g-'.Zeolite H-Y (Union Carbide, LZY82, H form) was calcined (500 OC, 2 h) prior to use (Si/Al ratio = 1.46, surface area = 675 m2 g-'). Tungsten oxide (10 O/O) supported on y-A120, was prepared from ammonium metatungstate according to the method of Maitra et al.14 using adsorption from an aqueous solution (pH 6.5, 25 "C, 8 h). After this procedure the sample was dried (120 "C, 8 h, air) and calcined for 16 h. Following calcination the catalyst was found to be amorphous by X-ray diffractometry, in agreement with previous studies.'* The calcined 10 YO W0,/A120,, acid-treated clinoptilolite and H-ZSM-5 were pelleted, without addition of binder, and sieved to give particles (0.5-1.0 mm) for use in the catalytic reactor. Catalyst Testing and Reactivation Catalysts were tested for methanol conversion and hex- 1 -ene cracking using procedures and equipment previously described,' except that reagents were fed to a heated vaporiser using calibrated peristaltic or syringe pumps which afforded improved control ofG .J . Hutchings et al. 63 5 reactant feedrate. Catalyst reactivation was carried out by substituting the methanol (or hex-1-ene)-N, stream for oxygen or nitrous oxide (flow rate 1 cm3 s-l). N,O decomposition was studied over both the fresh and deactivated catalysts in a fixed-bed Pyrex glass reactor. The extent of decomposition was determined by on line gas chromatography using a thermal-conductivity detector. Structure of Coke Deposits Solid-state 13C (c.P. - M.A.S.) n.m.r. spectra of coked and partially reactivated zeolite H-Y were obtained using a Bruker AM300 spectrometer with a contact time of 1 ms and a recycle time of 4 s.The rotor (alumina) was spun at 5.0 kHz and 900-5900 spectra were accumulated prior to Fourier transformation. The rotors containing uncoked zeolite under the conditions specified gave no detectable resonances in the range 6 - 150 to 340 ppm. Results and Discussion Nitrous Oxide Reactivation of H-Y Catalyst for Hex-1-ene Cracking Zeolite H-Y was reacted with hex-1-ene at 500 "C and liquid hourly space velocity 0.38 h-l, and the product distributions are given in table 1 (runs 1 A-I E). Reaction with hex-1-ene was selected as we have previously shown that this reactant leads to rapid catalyst deactivation. Deactivation of the catalyst was indicated by a decrease in hex-1-ene conversion together with an increase in mass due to the deposition of coke.After reaction for 6 h typically 18.0% by mass of carbon had been deposited and the activity had decreased by ca. 30%. Samples of the deactivated zeolite were then reactivated by substituting either 0, or N,O for the N,-hex-1-ene feed at 450 "C for 6 h and 500 "C for 2.5 h. The relative rates of coke removal for these procedures are shown in fig. 1. Initial carbon removal is far more rapid for oxygen than for N,O. However, it is clear that N,O removes all the carbonaceous deposit, whereas the equivalent treatment with 0, always leaves some residual coke within the zeolite which is particularly resistant to oxygen. Hence to obtain virtually complete coke removal with oxygen requires long reaction periods.Moreover, increases in temperature are not particularly beneficial, e.g. 0, treatment at 500 "C for 2.5 h gives a residual carbon level of 0.45 % by mass. Following reactivation the catalytic activity for hex- 1 -ene cracking was evaluated and compared to that of the fresh catalyst (table 1 and fig. 2). The data show that zeolite H-Y is rapidly deactivated by reaction with hex-1-ene and after ca. 100 min a steady conversion level of ca. 75 % is achieved. Analysis of coke deposition with time-on-line indicates that the coke is mainly formed during this period of rapid decline in conversion, leading to a decrease in active surface area, presumably by pore blocking. It is clear that both the N,O and 0, reactivation procedures studied in this work can restore the initial high hex-1-ene conversion, with N,O reactivation (2.5 h, 500 "C) being more effective than an equivalent 0, treatment; however, for both oxidants, the reaction time for which the initial high conversion is maintained is significantly shortened.This effect was enhanced by repeated reaction and reactivation cycles when, although the high initial activity was always restored, subsequent deactivation became more rapid. Since the reactivated zeolite contains very similar amounts of coke following either N,O or 0, reactivation, this difference in initial decay rate is probably not due primarily to residual coke deposition. It is more likely that the use of high concentrations of oxidant during reactivation may cause hydrothermal reactions owing to the rapid initial coke oxidation, leading to a change in the nature and number of the active acid sites.Hence, whilst this study confirms that N,O is a possible alternative regeneration oxidant it is clear that the reactivation procedure requires considerable optimisation to ensure that the integrity and numbers of the acid sites are maintained.Table 1. H-Y catalyst reactivation data for hex-1-ene cracking E catalyst condition : new catalyst after N,O reactivationa after 0, reactivationb o\ ~~~ run number: 1A 1B 1C I D 1E 2A 2B 2C 2 D 2E 3A 3B 3C 3D 3E time on linec/min conversiond( %) selectivity (YO by mass) CH4 c3 c4 c5 c,+ 'ZH4 'ZH6 50 98 142 94.7 80.3 77.5 0.03 0.08 0.08 0.2 0.6 0.7 0.1 0.3 0.3 17.1 21.4 21.5 53.2 42.0 42.6 29.8 32.9 32.5 2.6 2.4 - 268 75.7 0.09 1.2 0.3 19.5 35.6 26.5 18.4 360 20 100 182 78.8 96.5 76.7 75.2 0.06 0.1 0.1 0.1 1.1 1.6 1.1 1.8 0.1 0.4 0.4 0.5 14.8 25.3 22.0 20.8 33.1 44.1 40.6 37.3 26.7 26.7 31.8 21.4 23.9 1.8 4.0 18.3 237 75.2 0.1 1.2 0.3 16.0 33.4 26.6 22.6 360 77.9 0.4 0.5 0.1 15.6 39.8 35.5 8.5 40 90 147 240 93.5 79.1 72.8 76.1 0.07 0.07 0.08 0.08 0.5 0.8 0.8 0.9 0.2 0.3 0.2 0.2 20.8 17.0 12.2 12.5 54.1 46.9 45.4 41.0 22.5 33.3 37.9 36.3 0.8 1.6 3.4 9.1 360 79.0 0.06 0.8 0.1 16.1 40.2 34.5 cb 8.2 U a N,O reactivation 500 "C, 2.5 h, 1 cm3 s-l.0, reactivation 500 "C, 2.5 h, 1 cm3 s-l. Cumulative time on line to end of run. Conversion excludes g- isomerisation to other C6 products. 31 % Table 2. H-ZSM-5 catalyst reactivation data for methanol conversion 2 ~- catalyst condition : new catalyst run number: 1A 1B 1C 1D 1E after N,O reactivationa _ _ _ _ _ _ 2A 2B 2C 2D 2E 3A time on linec/h 4 8 11 23 26 conversion (YO) 100 100 99.8 94.9 86.2 selectivity (% by mass) 0.5 0.4 0.7 0.7 0.3 2.4 3.1 3.6 2.9 3.8 0.3 0.1 0.3 0.3 0.1 C3H6 3.8 4.7 3.9 4.1 5.1 11.1 4.3 8.2 6.9 0.8 19.4 13.1 16.0 12.4 5.0 c5 9.5 7.4 9.0 7.4 4.6 6.8 6.3 8.5 6.6 5.6 46.4 59.2 49.0 58.6 74.6 CH4 'ZH4 'ZH6 C3H8 c4 C6 c,+ 4 8 11 23 26 100 98.3 94.3 59.8 48.3 0.8 0.6 0.4 1.1 1.3 8.1 8.1 8.9 15.1 14.7 0.3 0.2 0.1 0.1 0.1 9.4 8.5 8.5 19.6 19.1 10.5 5.0 2.7 1.9 1.3 25.4 17.3 13.3 13.8 12.0 13.3 10.3 9.1 8.7 7.6 8.5 7.7 8.0 8.5 9.1 23.7 42.4 49.1 31.3 34.9 4 99.0 0.8 5.5 0.3 8.7 8.8 25.0 12.0 7.7 31.2 after 0, reactivationb -~ 2 3B 3C 3D 3E 3' 8 I 1 23 26 @Q 97.7 95.2 80.3 57.1 2: -~ E "0 0.9 0.7 0.9 1.4 8.3 9.2 12.3 16.7 0.2 0.1 0.1 0.1 12.6 10.2 14.2 22.9 6.5 3.4 2.3 1.2 28.6 21.7 15.7 14.1 16.5 13.0 9.2 8.5 4.1 10.9 8.6 8.4 31.9 31.9 31.3 27.0 a N,O reactivation 450 OC, 6 h, 1 cm3 s-l.0, reactivation 450 "C, 6 h, 1 cm3 s-l. Cumulative time on line to end of run.G. J . Hutchings et al. 90 h 5 c 'g 80 8 > 637 - - t l h Fig. 1. Rate of carbon mass loss on reactivation: x , 0,, 500 "C; 0, 0,, 450 "C; 0, N,O, 500 "C; A, N,O, 450 "C. 100 r X 70 t 60 1 I I I I 1 1 0 50 100 150 200 250 300 time on line/min Fig. 2. Catalytic activity of H-Y for hex-1-ene cracking for: A, fresh H-Y (calcined 500 "C, air, 2 h); 0, N,O-reactivated H-Y (500 "C, 2.5 h), after one reaction cycle; 0, N,O-reactivated H-Y (500 "C, 2.5 h), after five reaction/reactivation cycles; x , 0,-reactivated H-Y (500 "C, 2.5 h);- +, 0,-reactivated H-Y (450 "C, 6 h).Only minor differences in product selectivity are observed for catalyst reactivation using equivalent 0, or N,O treatments, the most notable being that after 0, reactivation, formation of C,, hydrocarbons (i.e. oligomerisation products) are decreased relative to those for fresh catalysts; however, this is not observed for N,O. Also, for N,O reactivation an increase in the CH,, C,H, and C , yields are observed at the expense of C, and C , hydrocarbons. Since the zeolites following reactivation do not contain significant residual coke, it is unlikely that these small selectivity differences can be attributed to residual pore blocking, as has been observed for ozone reactivated 22 FAR I638 Reactivation of Catalysts using N,O I'" I Fig. 3.13C c.p. m.a.s. n.m.r. spectra: (A) coked zeolite Y (C 13.2%) 900 scans; (B) sample a following 0, reactivation 22 min, 500 "C (C = 1.2 YO) 5900 scans with TOSS suppression of spinning side bands; (C) sample A following N,O reactivation 32 min, 500 "C (C = 4.3 YO) 13663 scans. zeolite^.^ The observed differences in selectivity probably reflect slight differences in the nature of the active sites for the various materials. The structure of the carbonaceous deposits of coked and partially reactivated H-Y were investigated using solid-state m.a.s. n.m.r. A coked sample was obtained by reaction with hex-1-ene (non-13C-enriched) at 400 "C and liquid hourly space velocity 0.38 h-' for 4 h (13 % by mass carbon).The 13C n.m.r. spectrum of coked zeolite H-Y [fig. 3(A)] shows two distinct and prominent resonances in the range 6 = 20-230 ppm, which can be assigned to aliphatic carbon (6 = 0-50) and aromatic carbon (6 = 100-150) environments, denoted (a) and (b), respectively. l5 This coke structure is significantly different from that produced by methanol conversion over H-ZSM-5,1° when five distinct resonances were observed, and our findings are therefore in agreement with previous studies1' showing that the compostion of the coke is dependent on the structure of the feed reactant molecule. 13C M.a.s. n.m.r. spectra were also obtained for partially reactivated H-Y using N,O and 0, as oxidants [fig. 3 (B) and 3 (C)].It is apparent that both N,O and 0, preferentially remove aliphatic carbon environments. Hence N,O does not demonstrate any selectivity difference during reactivation when compared with 0,,G. J . Hutchings et al. 639 whereas we have previously described such differences with OJO, reactivation in a related publication.'" From this study it can be seen that with both N,O and 0,, the aromatic carbon environments are the most resistant to oxidation ; however, N,O removes all carbonaceous deposits, whereas oxygen is not so effective, and this indicates that, at least in part, different mechanisms are operating for these two oxidants. Nitrous Oxide Reactivation of H-ZSM-5 Catalyst for Methanol Conversion Methanol conversion into hydrocarbons was carried out using the synthetic pentad zeolite H-ZSM-5, at 400 "C and methanol weight hourly space velocity (w.h.s.v.) 1.7 h-', and the results are given in table 2 (runs 1 A-1 E).Following deactivation (i.e. after run 1 E) the catalyst was reacted with N,O or 0, at 450 "C and the reactivated catalysts were again assessed for the methanol conversion reaction (table 2 ) . N,O reactivation restores complete catalyst activity and gives a slightly higher initial methanol conversion when compared with oxygen reactivated H-ZSM-5. However, it is also apparent that N,O reactivated H-ZSM-5 exhibits a shorter useful lifetime when com- pared with 0, reactivated H-ZSM-5. Reactivation at 500 "C, compared with 450 "C, with N,O did not give any significant improvement and N,O reactivated H-ZSM-5 always gave a decreased catalyst lifetime when compared with fresh H-ZSM-5.Nitrous Oxide Reactivation of an Acid-treated Clinoptilolite Catalyst for Methanol Conversion Acid-treated clinoptilolite was used as a catalyst for methanol conversion at 400 "C and methanol w.h.s.v. = 0.1 h-', and the hydrocarbon product distributions are given in table 3. Under these reaction conditions the methanol conversion decreased with reaction time owing to the deposition of carbonaceous material, which was found to occur at a rate of 3 mg, g&,,,,,, h-'. The deactivated clinoptilolite was reactivated with either N,O or 0, and then re-examined as a catalyst for the methanol conversion reaction (table 3). N,O reactivation was found to be significantly more effective than the use of 0, reactivation under comparable conditions, which is apparent from a comparison of the effective catalyst lifetimes for methanol conversion (fig.4). 0, reactivation does not restore the catalytic performance to that of the fresh H- clinoptilolite, whereas N,O reactivation gives improved catalytic performance. Neither N,O nor 0, reactivation treatment significantly affects the product selectivity observed for methanol conversion. Nitrous Oxide Reactivation of a 10 YO W0,/yA120, Catalyst for Dimethyl Ether Conversion To demonstrate the applicability of nitrous oxide reactivation to catalysts other than zeolites, the procedure was investigated with supported tungsten oxide which has previously been cited as a catalyst for methanol or dimethyl ether conversion." Results for the reaction of dimethyl ether over 10 O/O WO,/AI,O, (w.h.s.v. = 0.6 h-I, T = 400 "C, 4 h) are given in table 4.Deactivation of the catalyst was rapid, as shown by a decrease in dimethyl ether conversion from 99 to I8 O h in 190 min time on line, and, following reaction, the catalyst was then found to contain 13 YO by mass carbon. The deactivation catalyst was then reactivated using either N,O (500 O C , w.h.s.v. = 0.4 h-', 10 h) or 0, (400 O C , w.h.s.v. = 0.4 h-', 4 h). These procedures fully restored catalyst activity, and no significant differences in product selectivities were observed (table 4). Use of shorter reactivation times with N,O, e.g. 4 h, proved to be ineffective, and to remove ca. 99% of the carbon deposit, reactivation for at least 10 h was required.Table 3.Acid-treated clinoptilolite reactivation data for methanol conversion __- - c,atalyst condition : after after new catalyst N,O reactivation' 0, reactivationb time on line"/min 120 240 360 120 240 360 120 240 360 conversion (YO) 96.5 95.1 63.7 99.4 98.4 66.9 96.3 83.9 35.4 selectivity (YO by mass) CH. 9.5 10.1 10.8 8.8 6.7 10.0 7.0 8.6 15.3 21.0 20.0 19.8 2.5 2.1 1.7 27.4 24.3 28.0 5.2 4.4 3.6 32.5 28.7 31.6 20.5 22.0 19.4 9.9 13.0 10.0 2.7 3.4 5.5 1.2 0.6 1.1 21.4 17.0 21.0 2.2 1.2 2.0 25.1 27.5 29.4 5.2 3.9 3.9 30.3 31.4 33.3 22.8 24.7 18.3 12.6 14.2 11.2 1.3 4.0 3.1 0.6 0.7 0.9 19.1 18.5 21.4 1.6 1.7 2.1 29.5 30.2 27.6 4.7 4.0 2.4 34.2 34.2 30.0 21.6 19.7 15.5 13.9 13.3 9.7 2.2 3.2 4.6 0.4 0.6 1.3 Q\ P 0 N,O reactivation 500 "C, 1 cm3 s-l, 3 h.0, reactivation 500 "C, 1 cm3 s-l, 3 h. Cumulative time on line to the end of the run. 2 %- 3 x Table 4. 10 Yo W03AI,03 catalyst reactivation data for methanol conversion catalyst condition : after after new catalyst N,O reactivation' 0, reactivationb U 4 time on line"/min conversion (YO) selectivity (YO by mass) CH4 C2H4 c3 c4 c5 C2H6 '6, 40 99.2 41.5 14.3 8.3 13.8 6.2 3.3 12.6 105 41.3 63.9 8.0 6.3 9.3 5.7 2.7 4.1 190 17.9 77.3 5.0 4.8 6.9 1.2 2.5 2.3 50 120 84.9 28.2 36.8 53.1 14.2 9.7 10.5 8.0 15.4 11.4 12.2 12.7 5.1 4.1 5.8 1.0 200 28.0 61.9 6.9 6.6 8.4 10.6 3.3 2.3 40 130 180 88.5 20.4 25.3 36.2 67.6 71.3 17.4 6.7 6.7 11.0 5.3 6.1 17.1 8.2 8.9 7.6 3.6 1.2 4.2 2.4 2.9 6.5 6.2 2.9 a 0, reactivation 400 OC, w.h.s.v. = 0.4 h-l, 4 h * Cumulative time on line to end of run.N,O reactivation 500 O C , w.h.s.v. = 0.4 h-l, 10 h.G. J . Hutchings et al. 64 1 2ol 10 01 I 1 I I I I I 1 0 60 120 180 240 300 360 420 480 time on line/min Fig. 4. Catalytic activity of acid-treated clinoptilolite for methanol conversion : A, fresh acid- treated clinoptilolite; 0, following 0, reactivation ; 0, following N,O reactivation. Nitrous Oxide Decomposition Nitrous oxide decomposition has been widely studied as a test catalytic reaction for many years using oxide catalysts.18 While the thermal decomposition of nitrous oxide is negligible in the temperature ranges used for reactivation (N,O decomposition 0.22-0.23% in temperature range 400-530 "C), a large number of oxides show significant activity in the temperature range of the reactivation studies reported in this paper.It is therefore possible that the role of N,O when used as an oxidant for catalyst reactivation is to provide a controlled supply of dioxygen or a monatomic oxygen species which then oxidises the coke. However, since N,O reactivation has been shown to effect total coke removal (which is not observed with 0, reactivation) it is considered that, at least in part, N,O oxidises coke deposits via a mechanistic pathway not observed with 0, oxidation. To investigate this aspect further N,O decomposition was studied over samples of the unreacted catalysts and the results are shown in table 5. It is clear that the acid form of ZSM-5 was a highly active catalyst for N,O decomposition, whereas the acid form of zeolites Y and clinoptilolite were not particularly active.At 450 "C only 0.43 % decomposition of N,O was observed with zeolite Y, and yet with this zeolite a 5 h reactivation at this temperature completely removed all carbonaceous material from the zeolite (fig. 1). Such a degree of N,O decomposition is insufficient to provide sufficient oxygen if the reactivation process were solely to be based on oxygen as oxidant. For example, at 450 "C N,O decomposition can provide only 0.7% of the 0, required for total coke removal from zeolite Y in 6 h treatment, and 2.5 h at 500 "C provides only 1.5 YO of the 0, required. Further inspection of the data indicates that at the reactivation conditions utilised for all the catalysts N,O decomposition cannot account for the degree of coke removal observed.It is therefore concluded that N,O reactivation must operate via a mechanism involving the interaction of coke with molecular N,O. From the data of fig. 1 it is apparent that such a process occurs at a slower rate than the oxidation of coke using dioxygen, but that this process effects complete coke removal. The acid form of ZSM-5 demonstrates significantly higher N,O decomposition when compared with the other catalyst samples, particularly at temperatures > 480 "C. This642 Reactivation of Catalysts using N,O Table 5. N,O decomposition over fresh catalysts" ZSM-5 catalyst T/"C b C d e zeolite Y clinoptilolite WO,/AI,O, y-Al,O, 400 0.17 0.24 0.25 0.35 0.10 0.2 1 0.27 0.23 450 0.18 1.1 1.6 2.5 0.43 0.2 0.87 0.52 480 0.23 4.2 6.0 8.4 0.75 0.53 1.9 500 0.31 9.2 13.0 18.2 1.7 1 .o 3.7 2.3 5 10 0.35 13.6 21.0 24.6 2.3 1.3 4.9 3.2 520 0.39 23.4 32.1 32.3 3.6 1.7 6.2 4.8 530 0.48 33.5 44.5 40.8 2.2 8.9 6.7 - - " YO N,O decomposition, N,O g.h.s.v.= 3600 h-', blank thermal decomposition when no catalyst present = 0.2 YO over whole temperature range. ZSM-5 after one (NH,),SO, treatment, 1.02 YO Na. ZSM-5 after two (NH,),SO, treatments, 0.98 YO Na. ZSM-5 as prepared, 7.00 YO Na. ZSM-5 after three (NH,),SO, treatments, 0.61 % Na as used in reactivation studies. Table 6. N,O decomposition over coked catalysts" catalyst : T/"C ZSM-5b zeolite Yc clinoptilolited WO,/Al,O, - 0.5 350 1 .o 400 I .9 2.3 0.7 1.7 450 3.2 8.9 1.4 4.5 15.4 480 14.8 20.4 500 32.4 42.4 34.4 47.1 - - _ _ _ _ _ _ ~ ~ ~ _ _ _ _ a YO N,O decomposition, N,O g.h.s.v.= 3600 h-l. Following reaction with methanol, w.h.s.v. = 1.74 h-l, 26 h, 370 "C, 5.77 YO C . Following reaction with hex-1-ene, w.h.s.v. = 1 .O h-l, 6 h, 300 "C. 14.0% C. Reaction with methanol, w.h.s.v. = 0.1 h-l, 6 h, 400 "C, 2.2 YO C. Following reaction with dimethyl ether, w.h.s.v. = 0.6 h-', 4 h, 400 "C, 13.0 YO C. was investigated further and a range of samples containing increasing concentrations of Na were prepared. In general, catalytic activity at any temperature increased with decreasing Na concentration and, in particular, the as-prepared ZSM- 5 was particularly inactive. It is therefore apparent that the strong Bronsted-acid sites present in ZSM-5 are responsible for the high activity in N,O decomposition. This effect of Na concentration on N,O decomposition has been recently observed with y-Al,0,,19 and the present study demonstrates that this could be of more general significance.To study the interaction of N,O with coke deposits a range of N,O decomposition experiments over coked catalysts samples were carried out (table 6). The decomposition of N,O is enhanced for coked catalysts when compared with unreacted catalyst samples. This is particularly marked for low reaction temperatures (< 450 "C) when very little decomposition is noted for the unreacted catalysts. However, the effect is observed at all temperatures investigated and for coked W03/A1,03 and zeolite Y, substantial decomposition occurs which is not observed for the unreacted catalysts. Analysis of exit gases during reactivation, particularly the initial 10 min, confirmed that the oxygen from N,O was reacted to form carbon oxides.This therefore confirms that reactivation with N,O is effected via direct interaction of the coke deposits with N,O and not via a mechanism involving decomposition at a catalytic site followed by reaction of the cokeG . J . Hutchings et al. 643 with the oxygen produced. Further supporting evidence comes from experiments in which N,O decomposition was studied in the absence of catalyst but with co-fed 'model reagents' for coke structures (e.g. anthracene or methyl-a-naphthyl ketone). Enhanced N,O decomposition was always noted in the presence of co-fed organic model compounds. For example, in the presence of low concentrations of gas-phase methyl-a- naphthyl ketone, N,O decomposition was increased to 0.284.33 % in the temperature range 40&530 "C; and in the presence of gas phase anthracene N,O decomposition was increased to 0.36-0.39 O/O in the temperature range 400-450 "C.The oxidation of organic molecules by N,O has been well studied,2U-22 and it is known to oxidise alkenes and alkynes with a wide range of substitution to corresponding carbonyl components. The reaction between alkenes and N,O is considered23 to occur via a concerted 1,3-dipolar cycloaddition. Based on the comparison of N,O decomposition using coked and unreacted catalysts and the model reagent studies, it is clear that direct reaction between N,O and alkenoid groups or other unsaturated structures plays a significant role in the reactivation process observed.From the limited range of catalysts investigated for hex- 1 -ene cracking and methanol conversion we have nevertheless demonstrated that N,O exhibits a marked variation in its reactivation efficacy, and that total coke removal can only be effectively achieved with N,O. For the catalysts studied, the regeneration efficacy is in the order clinoptilolite > ZSM-5 z H-Y > 10 YO WO,/Al,O,. Such a marked variation is not observed for oxygen reactivation, for which typically 4-6 h at 450-500 "C is required for effective coke removal for all catalysts studied. As detailed in the previous discussion the primary process occurring during reactivation can be considered to be the interaction of N,O with unsaturated centres in the coke deposits.Hence the reactivation efficacy can be related to two parameters: ( a ) the amount of coke and (6) the 'chemical structure' of the coke. There exists a broad correlation between the amount of coke and reactivation efficacy, but this does not explain the observation that zeolite Y is readily activated by N,O, whereas 10% WO,/Al,O, requires a very long reactivation treatment, but both catalysts contain similar amounts of carbon deposited as coke. It is therefore considered that the structure of the coke, which is dependent on both the nature of the active sites and the three-dimensional pore structure of the catalyst, is an important factor for N,O reactivation. Zeolites have been shown2* to restrict the deposition of coke owing to shape selectivity, and it can therefore be expected that the average molecular mass of the coke deposits will be much lower for the zeolites compared to the non-shape-selective 10 O/O WO,/A1,0, catalyst.The more complex chemical nature of the coke deposits expected for 10 YO WO,/AI,O, may account for the poor reactivation efficacy observed. Previous studies on the oxidation of organic compounds using N,O as oxidant have shown that aromatic hydrocarbons are resistant to oxidation. Hence the reactivation efficacy observed with the three zeolite catalysts for the two reactions studied may reflect the amount of aromatic structures present in the coke deposits. The studies using solid- state 13C m.a.s. n.m.r. demonstrate that coke derived from methanol as a reactant contains proportionately less aromatic carbon environments1' when compared with coke derived from hex-1-ene (fig. 3).Hence on this basis for zeolite Y used for hex-1-ene cracking N,O would be expected to have a poor reactivation efficacy, which is observed. The observation that clinoptilolite is readily reactivated by N,O is also mainly due to the two factors previously discussed. However, two further factors may also be important. First, clinoptilolite is a small-pore zeolite compared with ZSM-5, and the laydown of carbonaceous deposits in the smaller pores causes rapid deactivation owing to pore blocking. In this case the coke deposits may be more accessible to N,O for reactivation. Secondly, clinoptilolite contains significant levels of iron as an impurity (0.84% Fe by mass).This impurity may be catalytically active for the production of radical species from N,O, and hence may improve the reactivation efficacy of this oxidant. The results of this study demonstrate the general applicability of the use of nitrous644 Reactivation of Catalysts using N,O oxide as a reactivation reagent for the removal of carbonaceous residues from catalytic structures. When compared with the standard oxygen reactivation procedure it is apparent that for zeolite Y and acid-treated clinoptilolite, N,O may offer considerable advantages either with respect to more effective coke removal or improved catalytic behaviour. We thank the the University of Witwatersrand and the FRD, CSIR for financial support. We also thank Kim Pratley for supplying samples of natural clinoptilolite, Dr W.Pick1 for useful discussions and Dr A. A. Chalmers for obtaining the n.m.r. spectra. References 1 G. C. Bond, Heterogeneous Catalysis, Principles and Applications (Clarendon Press, Oxford, 1987), 2 M. E. Dry, in Catalysis, Science and Technology, ed. J. R. Anderson and M. Boudart) (Springer-Verlag, 3 B. Nkosi, N. J. Coville and G. J. Hutchings, J . Chem. SOC., Chem. Commun., 1988, 71. 4 B. C. Gates, J. R. Katzer and G. C. A. Schuit, Chemistry of Catalytic Processes (McGraw Hill, 5 C. D. Chang and A. J. Silvestri, J. Catal., 1977,47,249; L. B. Young, S. A. Butler and W. W. Keading, 6 P. Dejaifve, A. Auroux, P. C. Gravelle, J. C. Vedrine, Z. Gabelica and E. G. Derouane, J. Catal., 1981, 7 D. E. Walsh and L. D. Rollman, J. Catal., 1979, 56, 195. 8 G. J. Hutchings, Chem. Br., 1987, 23, 762. 9 R. G. Copperthwaite, G. J. Hutchings, P. Johnson and S. W. Orchard, J. Chem. SOC., Faruday Trans. 10 L. Carlton, R. G. Copperthwaite, G. J. Hutchings and E. Reynhardt, J . Chem. SOC., Chem. Commun., 11 G. J. Hutchings, R. G. Copperthwaite, T. Themistocleous, G. A. Foulds, A. S. Bielovitch, B. J. Loots, 12 R. G. Copperthwaite, G. Foulds, T. Themistocleous and G. J. Hutchings, J. Chem. Soc., Chem. 13 M. G. Howden, CSIR Report C. Eng. 413 (CSIR, Pretoria, South Africa, 1982). 14 A. M. Maitra, N. W. Cant and D. L. Trimm, Appl. Catal., 1986, 27, 9. 15 E. Breitmaier and W. Boelter, 13C NMR Spectroscopy (Verlag Chemie, Weinheim, 1974); F. W. Wehrli 16 E. G. Derouane, J. P. Gilson and J. B. Nagy, Zeolites, 1982, 2, 42. 17 G. A. Olah, H. Doggweiler, J. B. Felberg, S. Frohlich, M. J. Grdina, K. Karpeles, T. Keumi, S. Inaba, 18 J. C. Vickerman, in Catalysis (Specialist Periodical Report, The Chemical Society, London, 1978), 19 C. Kordulis, L. Vordonis, A. Lycourghiotis and P. Pomonis, J . Chem. SOC., Faraday Trans. 1, 1987,83, 20 F. S. Bridson-Jones, G. D. Buckley, L. H. Cross and A. P. Driver, J . Chem. Soc., Part 1 , 1951, 2999. 21 F. S. Bridson-Jones and G. D. Buckley, J. Chem. SOC., 1951, 3009 (Part 11). 22 G. D. Buckley and W. J. Levy, J. Chem. Soc., 1951, 3016 (Part 111). 23 R. Huisgen, J. Org. Chem., 1976, 41, 403. 24 P. Dejaifve, A. Auroux, P. C. Gravelle, J. C. Vedrine, Z. Gabrelica and E. G. Derouane, J . Catal., p. 70. Berlin, 1981), vol. 1, p. 159. New York, 1979), p. 1. J . Catal., 1982, 76, 418. 70, 123. 1, 1986, 82, 1007. 1987, 1008. G. Nowitz and P. van Eck, Appl. Catal., 1987, 34, 153. Commun., 1987, 748. and T. Wirthlin, Interpretation of Carbon-13 NMR Spectra (Heydon, London, 1978). W. H. Ip, K. Lammertsma, G. Salem and D. C. Taylor, J . Am. Chem. SOC., 1984, 106, 2143. vol. 2, p. 107. 627. 1981, 70, 123. Paper 8/01459K; Received 14th April, 1988

ISSN:0300-9599    

DOI:10.1039/F19898500633

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1989, 85(3), 645454 A Study of the Exchange Sites per Unit Area of External Surface of Zeolite Al-ZSM-5 G. Paul Handreck and Thomas D. Smith" Chemistry Department, Monash University, Clayton, Victoria, 3168, Australia The adsorption of the cationic dye methylene blue by the sodium-exchanged form of zeolite Al-ZSM-5 results in the release of a sodium ion from the external surface of the zeolite. The measurement of such sodium replacement, together with determinations of external surface area, has been used to assess the occurrence of sodium exchange sites per unit area of external surface of the zeolite. The external surface areas of the zeolite samples have a mean value of 106 m2 g-' and the surface accommodates six molecules of dinitrogen per channel intersection.The sodium form of the Al-ZSM-5 zeolite has 15-25 % of the total number of exchange sites on the external surface. The intracrystalline catalytic activity of zeolite Al-ZSM-5 is dominated by shape- selective reactions which occur within its channel structure through the external surface, which is affected by the adsorption of ba~es,l-~ the removal of aluminium ions,4 ~ilylation,~ coking6-12 and shows shape selectivity in alkylation reactions. l3 The contribution of external acid sites to catalytic activity has been estimated by infrared spectros~opyl~ and X-ray photoelectron spectr~scopy.'~ Measurements of the external surface areas have been described,16-'* while a decline in the intensity of the X-ray diffraction (X.r.d.) powder signal with decrease in micropore volume has been noted." In the present study the measurement of sodium-ion replacement which occurs as a result of the adsorption of methylene blue on the external surface of the sodium- exchanged form of the zeolite has been combined with measurements of the external surface area to assess the existence of exchange sites per unit area of the external surface of the zeolite.Experiment a1 The X.r.d. powder diffraction pattern of each sample of zeolite was recorded on a Rigaku Geigerflex apparatus using nickel-filtered copper Ka radiation. The morphology and crystallite size were observed using a Reichert-Jung Polyvar Met optical microscope and a Hitachi S45 1 LB scanning electron microscope. The chemical composition of the zeolites was established by a gravimetric and spectrophotometric determination of the silica content, while aluminium and sodium were determined by atomic absorption spectroscopy.20 The total water content was measured by weight loss after heating (1 173 K) in a muffle furnace.The cation-exchange capacities of the zeolites were determined by passing an ammonium nitrate (1 .O mol dm-3) solution down a column containing the zeolite and determining the sodium content of the filtrate by atomic absorption spectroscopy. The adsorption of methylene blue from aqueous solution by the zeolitic material was carried out using a batch technique at room temperature (293 K) after the zeolites had been equilibrated over a saturated calcium nitrate solution (relative humidity 56 70) for 7 days.The amount of cationic dye taken up by zeolite was determined spectro- photometrically while the sodium ion replacement was measured by atomic absorption 645646 The External Surface of Zeolite A1-ZSM-5 spectroscopy using potassium nitrate as an ionization suppressant. Using the same procedure, but omitting the dye, the amount of sodium ion released from the zeolite to water was determined. The nitrogen isotherms (77 K) of the zeolite samples were measured using a Carlo- Erba Sorptomatic (model 1800) apparatus after outgassing (623 K for 16 h) to residual pressure of 1 x Torr.7 Samples of zeolite A1-ZSM-5 containing various amounts of aluminium were prepared largely as described in the literature.2' Results The zeolite materials possessed X.r.d. powder diffraction patterns consistent with those described previously.22* 23 Scanning electron microscopy showed that the crystal morphology of the zeolitic material ranged from intergrown monolithic particles (1-3.5 pm in size) to highly intergrown crystallites which formed secondary aggregates (0.5-1.5 pm in size) as the aluminium content of the zeolitic material i n c r e a ~ e d .~ ~ The cation-exchange capacity of the zeolite samples was in close agreement with their sodium ion content, as shown in table 1. Exceptional behaviour was shown by silicalite 1 (composition number l), where the cation-exchange capacity was 15% of the analytical sodium ion content. However, this refers to the small impurity content of this material, with a large uncertainty (25 %) involved in the measurements.Adsorption of methylene blue (Md) from aqueous solution by the sodium-ion- exchanged form of zeolite Al-ZSM-5 is accompanied by a release of sodium ions (Sd) into the aqueous solution, as illustrated for a sample of zeolite with composition number 4 in fig. 1. Over the concentration range used, the amount of sodium ion released by the zeolite was consistently greater (by ca. 0.75 x mmol g-l) than the amount of dye adsorbed. Using identical conditions to the dye-adsorption experiments, but omitting the dye, contact of the zeolite with water caused a small but reproducible release of sodium ion (&) from the zeolite, which is presumed to originate from within the channel structure. showed that silicalite 1 and related microporous silica polymorphs do not adsorb methylene blue from aqueous solution.However, prolonged (72 h) exposure of silicalite 1 to an aqueous solution containing sodium chloride followed by wash treatment resulted in a small uptake of the dye (1.34 x mmol g-') (M,) as indicated by fig. 1. Optical-microscopic observations of such a sample of silicalite 1 exposed to an aqueous solution of methylene blue showed that 90 YO of the crystals were uniformly pale blue in colour. Sodium ion was absent from the supernatant solution, indicating that the dye was not adsorbed at sodium exchange sites. Making the assumption that ik(, is the same for all other compositions and then subtracting this amount from Md gives the amount of methylene blue adsorbed by a sodium ion exchange mechanism (M,) as indicated by line (c) of fig.1. Similarly, the sodium ion released during this process (S,) was determined by subtracting So from Sd and is also indicated by line (c) in fig. 1. Thus, taking into account the quantity & determined for each zeolite, values for M, and S, were obtained for the zeolites and are shown in fig. 2. A previous The methylene blue adsorption isotherms were fitted to the Langmuir equation X = &,., K C / ( l + KC) where & is the monolayer capacity and K the adsorption equilibrium constant. Values of Iyn and K were obtained from the plots shown in fig. 3 and are summarized in table 2. As shown in fig. 4, reasonable correlation exists between the amount of methylene t 1 Torr = 101 325/760 Pa.G. P. Handreck and T. D.Smith 647 Table 1. Properties of the Al-ZSM-5 zeolites composition no. composition ion-exchange ion-exchange capacity capacity /sodium ion mmol g-' con tent 1 Na0.08A10.01S~95.99'1928*5H20 0.2 0.15 2 Na0.72A10.67s195.33'1921 H2° 11.3 0.96 3 Nal. 34A1 1. ,asi 94.62'192 H2' 21.2 0.99 4 Na1.~1A11.77Si94.23'19222H20 27.6 0.96 5 Na'2.54A12.64Si93.36'19226H20 40.0 1.01 6 Na4.08A14.3~Si91.63'19236H2' 62.2 1.01 I I 1 I 10 20 30 methylene blue originally in solution/10-2 m o l g-' Fig. 1. (a) The release of sodium ions (S,) and (6) the adsorption of methylene blue (M,) for zeolite Al-ZSM-5, composition number 4, plotted against the amount of methylene blue originally in solution. (c) The filled squares represent the values obtained for S, after .Sg ( d ) was subtracted from S, and the filled circles represent the values for M , after Mo (e) was subtracted from 4,.blue adsorbed as measured by the Langmuir monolayer and the aluminium content of the zeolites. To determine the number of exchange sites per unit area of external surface of the zeolites the nitrogen adsorption isotherm for each sodium-exchanged form of the AI-ZSM-5 zeolites was measured at 77 K as shown in fig. 5, and the results were analysed by the a-S method using non-porous hydroxylated silica as standard.26 All the isotherms are characterized by high adsorption at very low relative pressure caused by the enhanced adsorption potential of the channel system of the zeolite. Evidence for capillary condensation in slit-like pores in the form of either H3- or H4- type hysteresis loops occurs in all samples.A further hysteresis loop occurs in648 The External Surface of Zeolite Al-ZSM-5 5 I 1 I I 10 20 30 methylene blue originally in solution/lO-' mmol g-' Fig. 2. S, (open circles) and M, (closed circles) plotted against the quantity of methylene blue initially in solution. conjunction with a step in the isotherm ranging from 0.1 1 to 0.25 relative pressure and is particularly prominent in the isotherm for silicalite 1. As the aluminium content of the zeolitic material increases, this low-pressure hysteresis loop becomes more shallow and broad in range, extending from 0.1 1 to 0.36 relative pressure, to close finally, so that it does not occur in the isotherm for the zeolite of composition number 4. The step in the isotherm for samples with lower aluminium contents produced a corresponding step or broad curve in the a-S plots, requiring care in the selection of a region where a-S plots (fig.6) gave consistent linear regions, i.e. between 0.88 and 1.25 a-S values, which is just above the region of greatest variability. In this region, lines of best fit were calculated, and from these the external surface area and the pore volume were determined from the slope and intercept, respectively, as shown in table 3. The surface a!ea per methylene blue molecule was calculated using the expression S'/(& NA) A2, where S' is the external surface area (m2 g-'), 4 is the methylene blue monolayer capacity equivalent for each sample dried at 623 K for 16 h ( mmol g-') (table 2) and NA is Avogadro's number (mol-l).The number of sodium exchange sites per square metre of the external surface was calculated using the following expression : [(& - &) NA1/ O5 '% where & is the monolayer capacity for composition number 1 (1 OP2 mmol g-l).G. P. Handreck and T. D. Smith 649 1.0 2.0 3.0 4.0 c/10-~ mi b - 3 Fig. 3. Langmuir adsorption isotherms of methylene blue on sodium-ion-exchanged Al-ZSM-5 zeolite. Table 2. Parameters for the adsorption of methylene blue by zeolite Al-ZSM-5 monolayer external adsorption capacity surface area equilibrium mmol g-' per methylene composition constant ; mol yule no. K x m 4 /A2 1 467 000 1.34 1.38 1023 2 540 000 3.00 3.13 63 1 3 410000 6.80 7.11 283 4 541 000 8.17 8.93 210 5 499 000 8.83 9.65 163 6 560000 13.2 14.73 116 Discussion The correlation which exists between the amounts of cationic dye adsorbed by the zeolitic materials from aqueous solution and the sodium cation released indicates clearly that methylene blue is adsorbed by a cation-exchange mechanism for the A1-ZSM-5 samples of composition number 2-6.The amount of sodium ion released by zeolite of650 13.0 - 11.0- The External Surface of Zeolite Al-ZSM-5 Al per unit cell Fig. 4. Langmuir monolayer dye adsorption capacities (q) for sodium-exchanged Al-ZSM-5 zeolites (m) and the same zeolites in protonic form (0)25 plotted against aluminium per unit cell for each sample. The difference is due to dealumination by the acid washing process. A line of best fit is plotted for the points. composition number 6 was consistently below the level of dye adsorption.The assumption that the sodium ion measured in the blank experiment originates entirely from the channels does not hold in this case. Silicalite 1 (composition number 1) showed an unexpectedly high level of dye adsorption (1.34 x lop2 mmol g-l) bearing in mind its low level of aluminium content in the bulk analysis (0.01 x mol g-'). The prolonged exposure of silicalite 1 to aqueous solutions containing sodium cations creates new adsorption sites on the surface of the silica polymorph. The Langmuir equation gives a satisfactory fit to the methylene blue adsorption isotherms, while a reasonable correlation exists between the Langmuir adsorption monolayer for methylene blue and the aluminium content of each Al-ZSM-5 zeolite sample.The adsorption equilibrium constants for the sodium exchanged forms of the zeolites are higher than those found for the protonic forms of the zeolite of the same aluminium content, indicating that methylene blue exchange is more favoured by the sodium ion over the protonic form of the zeolite; this is in keeping with the ion-exchange properties of the Al-ZSM-5 zeolite, which favours large weakly hydrated ions.27* 28G . P. Handreck and T. D. Smith 65 1 150 100 200 150 100 - 150 - I M Y 4 “E y 100 150 s? v) v 100. 150’ 100 150 100 1 I I 1 I 1 0.2 0-4 0.6 0.8 1.0 Pa IPO Fig. 5. Nitrogen adsorption isotherms for the sodium-exchanged AI-ZSM-5 zeolite samples. The number of sodium exchange sites per square metre of external surface increases systematically with the aluminium content of the samples (table 4).In the sodium-ion form of the Al-ZSM-5 zeolite there are 15-25 YO of the total number of sodium-ion-exchangable sites, which may be taken as an indication of the number of Bronsted-acid sites, on the external surface (table 4). An earlier study indicated a lower number (5-10%) of the total Bronsted acid sites are situated on the external surface of an AI-ZSM-5 ze01ite.l~ The sample of zeolite used had a low aluminium ion content similar to composition number 2 used here and which showed 15% of its sodium exchange sites to be on the external surface. The methyleneoblue molecule occupjes a rectangular vo!ume 17.0 x 7.6 x 3.25 i3;29 it could cover 129 A’ placed flat or 55 A2 if edge-on or 25 A2 if end-on to the adsorbing surface.The external surface area available for the adsorption of methylene blue, with652 The External Surface of Zeolite A1-ZSM-5 100 2 100 W a 100 150 100 - s- 150 - 100 - t f I I I 1 0 . 5 1.0 1.5 2.0 2.5 a - S values Fig. 6. a-S plots for the sodium-exchanged Al-ZSM-5 zeolite samples. Table 3. Physical characteristics of the Al-ZSM-5 zeolites composition no. se vp N, 1 85 0.15 25.4 2 119 0.13 22.1 3 121 0.13 22.6 4 113 0.14 23.2 5 95 0.14 23.8 6 103 0.14 23.2G. P. Handreck and T. D. Smith 653 Table 4. Measurements of the number of exchange sites on the external surface no. of sodium no. of sodium exchange sites per exchange sites on com- unit area of the external . position external surface/ surface/total ion- no. x 1016 m-2 exchange capacity - - 1 2 8.9 0.15 3 28.5 0.26 4 40.2 0.25 5 52.5 0.19 6 78.1 0.19 the exception of zeolite Al-ZSM-5 of composition number 6, is sufficient to allow the dye to be orientated flat on the adsorbing surface.In the case of zeolite Al-ZSM-5 of composition number 6, some or all of the dye molecules would be required to be edge- on in order that all the adsorbed dye molecules have access to an exchange site. The low-relative-pressure hysteresis and accompanying step in the nitrogen isotherm have been observed earlier for samples with Si02/A1203 > 4000.18 In this study the loop was also evident in samples with SiO2/A1,O3 ranging up to 137. The external surface area does not vary systematically with the aluminium content of the zeolite, while taking a mean value of 106 m2 g-l.Similar values of the external surface area have been observed previously,18 while other samples of zeolite Al-ZSM-5 show much smaller external surface areas, ranging from 4.8 to 40.8 m2 g-1.16*17,19 The micropore volumes of the Al-ZSM-5 zeolites show little variation with change in aluminium ion content, giving a mean value of 0.14 cm3 g-l. The number of dinitrogen molecules per unit cell, N,, was calculated using the expression where 5 is the micropore volume (cm3 g-’), 4, is the density of dinitrogen (0,808 g cm-3 at 77 K),30 Ml is the molecular weight of dinitrogen and M, is the molecular weight of the participate zeolite sample. The values for the number of nitrogen molecules per unit cell (table 4) are centred around a mean [23.4 within experimental error (k lo%)] of 24 molecules per unit cell, with 6 molecules per channel intersection, in agreement with previous ~ t u d i e s ~ ~ .~ ~ in which it was concluded that each dinitrogen molecule contacts the surface of the zeolite pores.31 The methylene- blue adsorption technique has been developed from earlier studies describing the adsorption of dyes on alumino~ilicates,~~ sodium zeolite X,33 natural leading to exchange of sodium ions,35 acidity studies using Hammett indicator^,^^ titration of adsorbed dyes using pyridine3’ and the use of thin-layer chr~matography~~ and photoacoustic spectroscopy to follow dye ad~orption.~’ The present study combines determinations of the sodium-ion replacement from the Al-ZSM-5 zeolite as a result of dye adsorption on the external crystal surface with measurements of the external surface area to establish the number of exchange sites per unit area of external surface of the zeolite.The new knowledge of the number of exchange sites on the external surface may be of assistance in the consideration of the turnover number~,~’-~~ molecules per minute per active aluminium site, particularly for catalytic reactions such as the conversion of 1,3,5- trimethylbenzene or 2,2-dimethylbutane which occur on the external surface of the zeolite.654 The External Surface of Zeolite Al-ZSM-5 We thank the staff of the CSIRO Materials Science and Technology Division, namely Linda Bruce, Tom Mole, Keith Wilshier and Karl Foger, for useful discussions, and Peter De Mun k for assistance with the nitrogen- adsorp t ion measurements.References 1 M. Sugimot, H. Katsuno, K. Takatsu and N. Kawata, Zeolites, 1987, 7, 503. 2 N. R. Meshran, J. Chem. Technol. Biotechnol., 1987, 37, 111. 3 J. R. Anderson, K. Foger, T. Mole, R. A. Rajadhyarsha and J. V. Sanders, J. Catal., 1979, 88, 114. 4 S. Namba, A. Inaka and T. Yasaiwa, Zeolites, 1986, 6, 107. 5 K. G. Wilshier, P. Sart, R. Western, T. Mole and T. Bearsing, Appf. Catal., 1987, 31, 339. 6 F. X. Cormerals, G. Perot and M. Guisnet, Zeolites, 1981, 1, 141. 7 P. Dejaifve, A. Auroux, P. C. Gravelce, J. C. Vedrine, Z. Gabelica and E. G. Derouane, J. Catal., 8 G. D. McLellan, R. F. Howe, L. M. Parker and D. M. Bibby, J. Catal., 1986, 99, 486. 9 D. M. Bibby, N. B. Milestone, J. E. Patterson and L. P. Aldridge, J.Catal., 1986, 97, 493. 1981, 70, 123. 10 M. Guisnet, P. Magnoux and C. Canaff, Stud. Surf. Sci. Catal., 1986, 28, 701. 11 J. Karger, H. Heifer, J. Caro, M. Biilow, H. Schlodder, R. Mostowicz and J. Volter, Appl. Catal., 12 B. A. Sexton, A. E. Hughes and D. M. Bibby, J. Catal., 1988, 109, 126. 13 D. Fraenkel, M. Cherniavsky, B. Ittaa and M. Levy, J. Catal., 1986, 101, 273. 14 J. Take, T. Yamaguchi, K. Miyamoto, H. Ohyame and M. Misoho, Stud. Surf. Sci. Catal., 1986, 28, 15 J. P. Gilson and E. G. Derouane, J. Catal., 1984, 88, 538. 16 I. Suzuki, S. Namba and T. Yashima, J. Catal., 1983, 81, 485. 17 M. Inowata, M. Yamada, S. Okada, M. Niwa and Y. Murakami, J. Caful., 1986, 100, 264. 18 P. J. M. Carrot and K. S. W. Sing, Chem. Znd., 1986, 786. 19 P. Hudec, J. Novansky, S.Silhar, T. N. Trung, M. Zubek and J. Madar, Adsorption Sci. Technol., 20 D. R. Corbin, B. F. Burgess Jr, A. J. Vega and R. D. Farlee, Anal. Chem., 1987, 59, 2722. 21 N. Y. Chen, J. N. Miale and N. Y. Reagen, U S . Patent, 4, 112, 056, 1978. 22 R. von Ballmoos, Collection of Simulated XRD Powder Patterns for Zeolites (Butterworths, London, 23 E. L. Wu, S. L. Lawton, D. A. Olson, A. C. Rohrman and G. T. Kokotailo, J. Phys. Chem., 1979, 83, 24 V. N. Ramannikov, V. M. Mastikhin, S. Hocevar and B. Drzaj, Zeolites, 1983, 3, 31 1. 25 G. P. Handreck and T. D. Smith, J. Chem. SOC., Faraday Trans. I, 1988, 84, 4191. 26 S. J. Gregg and K. S. W. Sing, Adsorption, Surface Area and Porosity (Academic Press, London, 2nd 27 P. Chu and F. G. Dwyer, A.C.S. Symp. Ser., 1983, 218, 59. 28 D. P. Matthews and L. V. S. Rees, in Advances in Catalysis: Science and Technology (Wiley, New York, 29 P. T. Hang and G. W. Brindley, Clays Clay Miner., 1970, 18, 203. 30 CRC Handbook, ed. R. C. Weast (C.R.C. Press, Boca Raton, 61st edn, 1981, p. B-27.’ 31 P. A. Jacobs, H. K. Beyer and J. Valyon, Zeolites, 1981, I, 161. 32 M. J. Schwuger and H. G. Smolka, Colloid Polym. Sci., 1976, 254, 1062. 33 M. Susic, N. Petranovic and B. Miocinovic, J. Inorg. Nucl. Chem., 1972, 34, 2349. 34 A. Takasaka, S. Yoshino, H. Kono and Y. Matsuda, Zairyo, 1987, 36, 1181. 35 S. A. Levina, L. N. Malashevich and N. F. Ermolenko, Kolloidn. Zh., 1963, 25, 567. 36 M. W. Anderson and J. Klinowski, Zeolites, 1986, 6, 150. 37 P. Lemetas, J. Chim. Phys. Chim. Biol., 1982, 79, 277. 38 B. M. Lowe and H. L. Cook, Zeolites, 1982, 2, 29. 39 T. Somasundaram, P. Ganguly and C. N. R. Rao, Zeolites, 1987, 7,404. 40 W. 0. Haag, R. M. Lago and P. B. Weisz, Nature (London), 1984, 309, 589. 41 P. B. Weisz, Znd. Eng. Chem. (Fundam.), 1986, 25, 53. 42 P. B. Weisz, Chem. Technol., 1987, 368. 1987, 29, 21 495. 1986, 3, 159. 1984), p. 74. 2777. edn, 1982). 1985), p. 493. Paper 8/01519H; Received 18th April, 1988

ISSN:0300-9599    

DOI:10.1039/F19898500645

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I , 1989, 85(3), 655-662 Reversible Volume Change of Microparticles in an Electric Field Ryoichi Kishi and Yoshihito Osada* Department of Chemistry, Ibaraki University, Mito 310, Japan Microparticles of the cross-linked sodium salt of poly(acry1ate) with a diameter of 150-300 pm undergo reversible shrinkage when a d.c. electric field is applied. The shrinkage is rapid, and 90% volume change is reached within 50 s by applying an electric field of 46 V cm-' (0.3 mA). The rate of volume change is proportional to the d.c. current. From the contractile behaviour and pH measurements the contraction is explained in terms of ion transport of counter-microions in the electric field (electrokinetic phenomena). Recently, a novel electromechanical transducing system was demonstrated using synthetic polyelectrolyte gels, in which the conversion of electrochemical energy into mechanical energy (a mechanochemical system)' was stimulated by an electric field.Shoenfeld and Grodzinsky measured the tensile forces when a sinusoidal electric field was applied across a collagen membrane.2 The change in intramembrane ionic strength due to the electric field was calculated and an electrodiffusion model, coupled with convection, diffusion and migration of ions, has been proposed. More recently, De Rossi et aL3 studied the contractile behaviour of a partially hydrolysed poly(acry1amide) gel or thermally cross-linked poly(viny1 alcohol )-poly(acry1ic acid) composite membranes. These membranes underwent shrinkage in dilute NaCl solution when a d.c.current was applied through a pair of Pt-plate electrodes at a distance of 4.5 cm from the gel. The authors associated the contraction with a change in the pH of the medium, i.e. the electric field induced pH changes in the fluid, thereby changing the ionic state of the polyelectrolyte membrane. We have reported4.' that a water-swollen cross-linked poly(methacry1ic acid) gel inserted between a pair of electrodes undergoes contraction and concomitant fluid (water) exudation in aerobic conditions. The applied electric field was shown to induce the migration of hydrated microions in the gel, thereby transporting water to the e l e c t r ~ d e . ~ , ~ Thus the hydrated polyelectrolyte gel could contract or bend in the electric field, and some chemomechanical (mechanochemical) devices functioning as models of gel actuators were demonstrated.8 In the course of this study we also found non-linear transport behaviour of these ions in the gel,' which allowed us to make a semiquantitative analysis of current and pH oscillations'" using fast Fourier transport.However, mechanism of contraction of the polyelectrolyte gel is not fully understood. The objective of this paper is to investigate the kinetics of the electroactivated mechanochemical reaction of the gel immersed in water and to demonstrate that shrinkage of the polyelectrolyte gel occurs by an ion-transport process (an electrokinetic phenomenon). For this purpose, microparticles of the sodium salt of a poly(acry1ic acid) gel 15&300 pm in diameter were synthesized and the change in size of the particles with time in various electric fields were measured.Little attention has been paid to the kinetics of shape changes of polyelectrolyte microparticles. Tanaka and Fillmore presented a theory and experimental data for the kinetics of the spontaneous swelling of an uncharged poly(acry1amide) gel in water." Hirotsu studied the influence of a direct 655656 Behaviour of Microparticles in an Electric Field current on the phase-transition temperature of these particles. l2 With regard to applications, mechanochemical devices using hydrated gels can potentially be used as electroactivated actuators. Experiments using microparticles are desirable to establish a system with a minimized response time. Experimental Materials We chose to use microparticles of a mildly cross-linked sodium salt of poly(acry1ic acid) (Na-PAA).The Na-PAA microparticles were prepared by the inverse emulsion polymerization of a solution of 136 mmol of the sodium salt of acrylic acid and 0.065 mmol of N,N-methylenebis(acry1amide) in 68.5 cm3 of cyclohexane in the presence of 0.48 g of Span 80 (sorbitan mono-oleate). The suspension was stirred with a velocity of 180-200 rev. min-l under a nitrogen atmosphere and the polymerization was carried out at 60 "C for 6 h. Potassium persulphate (0.03 g) was used as a radical initiator. Following polymerization the microparticles were washed repeatedly in MeOH and equilibrated in distilled water for 1 day. The microparticles were spherical and the diameter of the swollen samples in distilled water was between 150 and 300 pm.The particles were then dyed by immersing them in a mol dm-3 aqueous solution of methylene blue, followed by washing. This allowed us to observe shape changes more easily under a microscope. Apparatus Two types of electrode cell were used, and their configurations are shown in fig. 1. Cell (A) was used to measure the shrinking of the microparticles, and consists of an SnO, glass electrode (10 R/ 0) a plastic-plate spacer and an SnO, glass electrode. A microparticle was placed between the two SnO, glass electrodes and 0.13 cm3 of distilled water was poured in (fig. 1). The electrode structure was 10 x 10 mm and was mounted on a polymeric frame with a spacing of 1.3 mm. A direct-current density of 0-0.3 mA cm-, was applied via the SnO, electrodes, and the volume change of the microparticles was observed with an optical microscope.In order to eliminate the effect of changing the pH of the fluid, a salt bridge was used instead of the glass electrode and the contraction behaviour was compared. Electrophoretic migration of the microparticles was measured in cell (B) shown in fig. 1. A microparticle was placed between a pair of parallel Pt-wire electrodes 11 mm apart and the space was filled with 0.7 cm3 of distilled water. A d.c. voltage ranging from 0 to 4 V was applied across the electrodes. The migration velocity of the microparticle was measured under an optical microscope. The pH of the gel was measured using a microcombination pH probe (Microelectrodes Inc., model M 1410).Results and Discussion A microparticle 173 pm in diameter was placed between the pair of platinum-wire electrodes of cell (B) and a 6.5pA direct current was applied. The particle started moving quickly towards the anode owing to electrophoresis. The rate of migration increased with increasing d.c. voltage. No changes in the shape and size occurred while the microparticle underwent electrophoretic migration. However, a sudden and quick contraction took place when the particle reached the anode and migration stopped. Plate 1 shows photographic time profiles of the size change when the microparticle was placed on the anodic (bottom) glass electrode in cell (A) and a 4.5 V direct current was applied. As is seen, shrinkage of the particle took place rapidly, and within 5 min the volume of the particle wasJ.Chem. SOC., Faraday Trans. I , Vol. 8 5 , part 3 Plate 1 Plate 1. Photographs of changes in the shape of the microparticle in an electric field : (a) before applying the electric field, (6) after 3 min, (c) after 5 min, ( d ) 12 min after turning off the electric field; voltage, 4.5 V (34.6 V cm-l); current, 0.22 mA; particle diameter, 173 pm. R. Kishi and Y. Osada (Facing p . 656)R . Kishi and Y. Osada 657 water pool \ glass electrode Pt ( ( b ) water pool \ plastic plate \ (11 mml I spacer I Pt wire electrode electrode anode 1 ., glass plate cat ion- exchange p" membrane anion - exc h an ge membrane I I Pt ( C electrode athode 1 85 mm Fig. 1. Configuration of the cells: (a) cell (A), used for measurement of shrinking and swelling processes of the microparticles ; (6) cell (B), used for measurement of the electrophoretic migration of the microparticles; (c) cell (C), used for measurement of volume changes of the rod-like gel.reduced to 1/9, after which no further significant shrinkage occurred. Shrinkage of the particle was usually accompanied by the preservation of the original spherical shape, but sometimes a shape change was observed; an example is shown in plate l(c). The particle started swelling gradually when the electric field was turned off, and the original size and shape were recovered after ca. 15 min [plate 1 (41. On contraction, a bleaching of the blue colour produced by methylene blue (MB) took place. This suggests that the MB was removed from the particle by electrophoretic force and was subjected to electrochemical reduction.No contraction or shape change occurred if the electrode polarity was658 Behaviour of Microparticles in an Electric Field Fig. 2. Time profiles of volume changes of the microparticles; voltage: 0, 2.4 V ( I 8.5 V cm-); @, 3.0 V (23.1 V cm-'); 0 , 3.6 V (27.7 V cm-'); (>, 4.2 V (32.3 V cm-'); 0, 4.8 V (36.9 V cm-'); @, 6 V (46.2 V cm-l); microparticle diameter, 180 pm. changed and the microparticle was placed on the cathode (bottom electrode). Instead, the particle floated up and become detached from the cathode. Fig. 2 shows time profiles of the contraction of a particle (1 80 pm in diameter) in cell (A) when a direct current ranging from 0 to 0.3 mA cm-2 was applied.Below 2.4 V (18.5 V cm-') no contraction or detectable electric current was observed. Shrinkage started when the d.c. voltage reached 3 V or the electric current reached 0.02pA. The rate of shrinkage increased with increasing electric field. Note that an increase in voltage increased not only the rate of contraction but the final size ( Vmin) that the particle could achieve. For example, a microparticle 180 pm in diameter was completely contracted within 50 s when 6 V dc was applied, and the diameter was reduced to 60 pm. No detectable shrinkage occurred after this, even if the electric field was applied for longer. Time profiles of the swelling process after the electric field was turned off are shown in fig. 3. All microparticles, except for that at 4.2V, recovered their original size and shape within 5-10 min and showed good reproducibility.The dependence of the rate of volume change on the applied voltage is illustrated in fig. 4. The current observed in the medium is also shown in this figure. Shrinkage of the microparticle occurs when the electric field exceeds the threshold overpotential (2.4 V for the Pt electrode). A significant direct current starts to appear in the medium at the same time. The rate of volume change increases in proportion to the current, and this fact evidently indicates that shrinkage of the particle is developed by the ion-transport process, accompanied by an electrochemical reaction. In a previous paper3 we have demonstrated that a water-swollen polyelectrolyte gel sandwiched between a pair of electrodes undergoes shrinkage under aerobic conditions and that the rate of shrinkage is proportional to the electric current.* The contraction was associated with the electrohydrodynamic transport of hydrated ions and concomitant water exudation.The shrinkage of the particles in the present case can be explained by the same reasoning; i.e. the negatively charged microparticle moves to the anode, and sodium ions move to the cathode owing to electrophoretic migration. An electric-field gradient will produce a steady diffusion of mobile cations and carry away sodium ions from carboxylates. Carboxylate anions, however, undergo substitution with Hf in the fluid and become largely undissociated. The carboxyl groups in this state areR. Kishi and Y .Osada 659 0 5 10 15 20 25 30 35 40 tlmin Fig. 3. Time profiles of the swelling process of the microparticles after the electric field was turned off. Symbols are the same as in fig. 2. VIV Fig. 4. Dependences of the rate of contraction of the microparticles (a), the current (b) and the final volume of the microparticles (c) on the voltage applied; microparticle diameter, 180 pm. much less hydrated, and the gel can consequently contract. In fact, a decrease in pH was observed in the gel near the anode when a direct current was applied to a crosslinked rod-like Na-PAA gel, as will be described later. Thus ionic migration of these microions induces a local pH change. However, note that the pH change is induced not only by a change in the local concentration of mobile cations, but also by electrode reactions (hydrolysis of water), since the platinum electrode can induce electrochemical reactions under these voltages and give rise to a shinkage of the particles due to a change in the pH of the medium.Therefore, an experiment using a salt bridge instead of the Pt electrode was carried out. In this case the pH of the medium did not change. Nevertheless, the particle underwent shrinkage, although the rate of contraction was slower: if a Pt electrode was used, a 90 % volume change was achieved within 1 min of commencing the experiment, whereas only a 47 O h volume change was attained under the660 - - 0 4 ) Behaviour of Microparticles in an Electric Field 1 . 4 1 . 2 Fig. 5. Time profiles of the volume change of an Na-PAA gel (0) and the electric current (0).Details of the experimental conditions are given in the text and fig. I . 9 F A 0 w- 0 100 200 300 t/min Fig. 6. Variation of pH with time in the Na-PAA gel or in water: (a) in the gel (5 mm deep) near the anode, (b) in the gel (5 mm deep) near the cathode, (c) in water near the anion-exchange membrane and ( d ) in water near the cation-exchange membrane; initial pH of water, 5.30. same conditions in the case of the salt bridge. Similar results were obtained when a half- cylindrical Na-PAA gel, 40 mm long and 15 mm in diameter, was placed in water and isolated from the Pt electrodes (10 x 10 mm) using ion-exchange membranes [fig. 1 , cell (C)]. The gel started to contract from the side near the anode, and after 300 min the size decreased to 51 % of its initial value when a 10.5 V cm-’ field was applied (fig.5). Meanwhile the pH of the fluid near the gel remained unchanged [fig. 6 ( c ) ] . In contrast, the pH value inside the gel near the region close to the anode decreased rapidly withR. Kishi and Y. Osada 66 1 (diameter>2 /mm2 Fig. 7. Dependence of the time for 96% volume change on the square of the microparticle size; d.c. voltage, 3.5 V (26.9 V cm-I). time, probably because sodium ions diffused away from the gel to the anode [fig. 6(a)]. The pH value inside the gel close to the cathode remained unchanged, remaining at 7 throughout the course of the experiment [fig. 6(b)]. A gradual decrease in the direct current (fig. 5 ) may be related to an outflow of sodium ions from the gel and a concomitant influx of hydroxonium ions from the fluid.In fact, an increase in the pH of the fluid near the cation-exchange membrane was observed with time [fig. 6(d)]. The rate of volume change should be a function of the initial size of the particle, and the relative rate of volume change must decrease with increasing size of the particle. Experiments using particles with different sizes were carried out, and the time needed for a 96% volume change is shown in fig. 7. The experimental results are inversely proportional to the square of the particle size. From fig. 7 it was calculated that the time taken for an 96% volume change for a particle of 1 pm diameter is as low as 0.23 ms under these conditions. This result coincides with Tanaka’s equation, showing a proportionality to the square of the characteristic length of the gel.” The reversible shrinking and swelling of the microparticles stimulated by an electric field that we have described have a variety of applications, such as actuators,13 drug- delivery ~ystems’~ etc.and others. Experiments focusing on these subjects are now in progress. We are indebted to Dr Aizo Yamauchi (Research Institute for Polymers and Textiles) for helpful discussions and encouragement. References Y. Osada, Advances in Polymer Science, Conversion of Chemical into Mechanical Energy by Synthetic Polymers (Chemomechanical System) (Springer-Verlag, Berlin, 1987), vol. 82, p. 1, N. A. Shoenfeld and A. J. Gradzinsky, Biopolymer, 1980, 19, 241. D. E. DeRossi, P. Chiarell, G. Buzzigoli and C. Domenci, Trans. Am. Soc. Artif. Intern. Organs, 1986, 32, 157. Y. Osada and M. Hasebe, Preprints of the Ann. Meeting of the Electrochem. Soc. Jpn., 1984, Y. Osada and M. Hasebe, Chem. LRtt., 1985, 1285. Y. Osada, M. Hasebe, R. Kishi and K. Umezawa, 31st IUPAC Macromolecular Symp., Microsymp. 1987, vol. VII, p. 225. Y. Osada, M. Hasebe, R. Kishi and K. Umezawa, 6th CHEMRAWN Preprints, 1987, p. 622. Y. Osada and R. Kishi, J . Polym. Sci., Part C, 1987, 25, 481. Y. Osada, K. Umezawa and A. Yamauchi, Makromol. Chem., 1987, 189, 597. pp. 12-14.662 Behauiour of Microparticles in an Electric Field 10 K. Umezawa and Y. Osada, Chem. Lett., 1987, 1795. I 1 T. Tanaka and D. J. Fillmore, J. Chem. Phys., 1979, 70, 1214. 12 S. Hirotsu, Jpn. J. Appl. Phys., 1985, 24, 396. 13 M. Suzuki, T. Tateishi, T. Ushida and S. Fujishige, Biorheology, 1986, 274. 14 S. R. Eisenberg, A. J. Grodzinsky, J. Membr. Sci., 1984, 19, 173. Paper 8/01533C; Receiced 19th April, 1988

ISSN:0300-9599    

DOI:10.1039/F19898500655

出版商:RSC    

年代:1989

数据来源: RSC