摘要:

J. Chern. Soc., Furuduy Trans I , 1989, 85(6), 1493-1500 Effects of Pressure on the Photoreduction of p-Benzoquinone in Normal and Reversed Micellar Systems Katsuhiro Tamura,* Masatoshi Abe and Masayoshi Terai Department of Chemical Engineering, Faculty of Engineering, The University of Tokushima, Tokushima 770, Japan The photoreduction of p-benzoquinone (p-BQ) in normal and reversed micellar systems has been studied kinetically under high pressures up to 150 MPa. Anionic sodium dodecyl sulphate (SDS) micelles accelerated the reaction, while cationic hexadecyltrimethylammonium bromide (CTAB) micelles retarded it. Pressure promoted the photoreaction in CTAB micellar systems, but not in SDS micelles. The reaction in Aerosol-OT reversed micellar systems (AOT/H,O/n-heptane) was faster than that in its bulk solvent, n-heptane.Micellar effects were found ; however, these occurred at almost the same rates as that in water. The location of p-BQ in the micellar systems was studied using data for their critical micelle concentrations. The mechanisms of the reaction are also discussed. Photochemical reactions in homogeneous systems are little influenced by moderate pressure up to ca. 100 MPa, except that the effects of viscosity increase with pressure, which usually retards chemical reactions such as dimerizati0n.l On the other hand, large pressure effects are often observed in heterogeneous systems such as micellar solutions, because the properties of micelles (shape, critical micelle sol~bilization~ and aggregation number6) are often affected by pressure, and this leads to large pressure effects on the photochemical reactions.', In the present paper we describe the effects of pressure on the photoreduction of p-benzoquinone (p-BQ) in anionic sodium dodecyl sulphate (SDS) micelles, cationic hexadecyltrimethylammonium bromide (CTAB) micelles in aqueous media and di(2-ethylhexy1)sodium sulphosuccinate reversed micelles (AOT/H,O/n-heptane) at high pressures up to 150 MPa.Micelles in aqueous media solubilize non-polar solutes in the non-polar core or adsorb them on the surfaces and give the reactants organized reaction fields. Reversed micelles provide unique fields to solubilize ionic or polar solutes in apolar media, and can control the reaction rate or pathway occurring in the interior core of the micelles. This is due to the assistance of the restricted field produced in the interior core of the micelles.To check the probability of solubilization of p-BQ in micelles, the solubilities of p-BQ in n-dodecane and n-hexadecane were studied. These alkanes correspond to the hydrocarbons of SDS and CTAB. Furthermore, the effects of p-BQ on the critical micelle concentrations (c.m.c.) of SDS and CTAB were also determined in order to give some indication of the location of the reactants in the micelles. Experimental p-Benzoquinone (p-BQ) was obtained from Wako and recrystallized twice from ethanol. Sodium dodecyl sulphate (SDS) and hexadecyltrimethylammonium bromide (CTAB) were purchased from Nakarai and washed with diethyl ether for 12 h using a Soxhlet extractor and then recrystallized twice from ethanol.Di(2-ethylhexy1)sodium sulpho- succinate (Aerosol-OT, AOT) (Tokyo Kasei) was dissolved in methanol, filtered, 14931494 Photoreduction of p- Benzoquinone in Micelles and the solvent was removed under vacuum. n-Dodecane and n-hexadecane were dried with sodium and distilled under reduced pressure. Water was distilled twice, once from potassium permanganate solution. The apparatus used for the high-pressure experiments consisted mainly of three a high-pressure vessel with quartz windows, a 400 W mercury lamp (Toshiba, H 400-P) and a high-pressure generator (Hikari Kohatsukiki Ltd; KP-3 A). The reaction solution was placed in a sliding cell (12@ x 15 mm), which was convenient as no correction of reactant concentration was required.A silicon photocell (Hamamatsu TV S780-5B) was used to monitor the lamp emission and to correct the concentration of reaction products, since the light intensity might vary with time. A pressure up to 150 MPa was used, and the reaction temperature was controlled at 308k0.2 K by circulating water of constant temperature around the high-pressure vessel. mol kg-l p-BQ and 1 .OO x 10-1 mol kg-l surfactant was irradiated for the desired time after being pressurized. The residual p-BQ was determined by U.V. spectrophotometry, and the apparent first-order rate constants were evaluated. The solubilities of p-BQ in n-hexadecane, n-dodecane, n-heptane and water were also determined by U.V. spectrophotometry. The critical micelle concentrations of SDS and CTAB containing p-BQ were determined by an electric-conductivity method using a Yanagimoto MY-8 conductivity instrument. The reaction solution containing 1 .OO x Results The absorption spectra of solutes are strongly dependent on the nature of the media.Accordingly, these spectra provide valuable information on the properties of the environment surrounding the solutes. Table 1 shows the maximum wavelengths (Amax) of p-BQ in micellar and non-micellar systems at atmospheric pressure. A, had two values: 245.7 nm for water and normal micelles in aqueous media and 240.8 nm for n-heptane and reversed micelles (AOT/H,O/n-heptane). Amax did not depend on the nature of the micelle but on the media used. These results suggest the location ofp-BQ molecules in micellar systems, namely that p-BQ is present in the bulk solvent phase and/or on the surfaces of the micelles and surrounded by the media.The photoreduction of p-BQ was markedly affected by the electric charges of the micelles.lo The changes in reaction conversion with time in various heterogeneous and homogeneous systems at atmospheric pressure are shown in fig. 1. The results demonstrate large medium effects ; for example, SDS aqueous solutions promoted the reaction, while CTAB aqueous solutions retarded it when compared to the water system. Kano and Matsuo indicated the formation of a cationic species as an intermediate of this reaction, based on the concentration effects of micelles,lo because anionic surfactants accelerated the reaction and cationic surfactants suppressed it.On the other hand, AOT reversed micelles significantly accelerated the reaction as compared to the bulk solvent (n-heptane), and showed almost the same rates as those in water. From these results, the presence of concentration effects from AOT reversed micelles is predictable. A water content up to R = 30 in the interior core of the reversed micelles hardly affect the reaction rates, where R is the molar ratio of water to AOT (R = [H,O]/[AOT]). Water has the ability to promote the photoreduction of p-BQ rather than n-heptane. Table 2 shows the apparent first-order rate constants of the photoreduction of p-BQ in various systems at high pressures up to 150 MPa. The logarithms of the ratios of the rate constants at high and atmospheric pressures (k,/ko) plotted against pressure are shown in fig.2. The pressure dependence of the rate constants are strongly influenced by the nature of the reaction fields, i.e normal micellar, reversed micellar and homogeneous systems. The reaction was accelerated by pressure in normal micellar and homogeneous systems but was suppressed in the AOT reversed micellar systems. At aK. Tamura, M . Ahe and M. Terai 1495 Table 1. The maximum wavelengths of p- benzoquinone in various systems at atmospheric pressure and 298 K SDS" 245.7 water 245.7 CTAB" 245.7 n-heptane 240.8 AOT" (R = 2)b 240.8 (R = 5) 240.8 (R = 20) 240.8 (R = 30) 240.8 a [surfactant] = 1.00 x 10-1 mol kg-'. b R = [H P I / W T I . 1001 - E G .I E 5 0 > 8 n - 0 20 4 0 reaction time/rnin 60 Fig. 1. Variation with time of the photoreduction conversion of p-benzoquinone in micellar and non-micellar systems at atmospheric pressure and 308 K.a, SDS; a, water; 0 AOT, R = 2 ; A, AOT, R = 5; A, AOT, R = 2 0 ; 0, AOT R = 30; @, n-heptane; W, CTAB. [surfactant] = 1.00 x lo-' mol kg-'. [p-BQ] = 1.00 x mol kg-'. R = [H,O]/[AOT]. water content R = 30 the reaction rates in AOT reversed micelles were greatly reduced at higher pressures. Such drastic changes at R z 30 were also observed in our studies on the fluorescence polarization of AOT reversed micelles at high pressures.1' These results suggest large fluctuations in the properties of the reaction fields. Zulauf and Eicke studied the effect of temperature on the structures of AOT reversed micelles in iso-octane using photon-correlation spectroscopy.l2 They concluded that polydispersity occurred in these water concentration ranges, with pressure seeming to promote the aggregation of the reversed micelles. The apparent activation volumes of this reaction ( A V ) were estimated using the following equation : A V = -RT(alnk/i3p), where In k is the logarithm of the rate constant, and can be shown to be a first-order function of pressure for SDS, water and n-heptane systems, and a second-order function of pressure for CTAB systems. The apparent activation volumes at atmospheric pressure1496 Photoreduction of p- Benzoquinone in Micelles Table 2. Pressure dependence of the apparent first-order rate constant of the photoreduction of p-benzoquinone in micellar and non-micellar systems at 308 K" system 0 MPa 30 MPa 50 MPa 100 MPa 150 MPa SDSb 18.3 18.4 20.6 23.9 water 7.5 8.00 8.48 8.98 CTABb 0.725 0.893 0.993 1.06 n-heptane 1.94 2.03 2.07 2.21 AOTb (R = 2)' 7.47 7.32 6.72 6.93 (R = 5) 7.74 7.35 6.98 7.10 ( R = 20) 7.45 6.70 6.65 6.37 (R = 30) 6.77 6.77 6.57 5.93 ~~~ ~ a [p-BQ] = 1.00 x mol kg-'.[surfactant] = 1.00 x lo-' mol kg-'. R = [H,O]/[AOT]. 0.4 0.2 h * 5 s 0 -0.2 0 50 100 150 p l w a Fig. 2. Plots of In (k,/k,) us. p for the photoreduction of p-benzoquinone in micellar and non- micellar systems at 308 K. @, CTAB; A, n-heptane; (3, SDS; ., water; 0 AOT, R = 2; 0, AOT, R = 5 ; A, AOT, R = 20; 0, AOT, R = 30. are shown in table 3. The activation volumes of the homogeneous systems, water and n-heptane, have small negative values and their absolute values are similar.Such small values are a feature of photochemical reactions in homogeneous systems.* In heterogeneous systems such as SDS or CTAB micellar solutions, large activation volumes can be expected because the micellar structures (shape, size and aggregation number) and the solubilization of reactants are often influenced by pressure, i.e. theK . Tarnura, M. Abe and M. Terai Table 3. Apparent activation volumes of the photoreduction of p-benzoquinone in micellar and non-micellar systems at atmospheric pressure and 308 K" system A Vf/cm3 mol-1 SDS - 4.7 water - 3.0 CTAB - 22 n-heptane - 3.3 AOT (R = 2) 1.5 (R = 5) 1.7 ( R = 20) 2.4 (R = 30) (3.8) " The concentrations of p-BQ and surfactants are the same as those in table 2. Table 4. Solubilities ofp-benzoquinone in various solvents at atmospheric pressure and 308 K solvents [p-BQ]/ 1 O-' rnol kg-' n- hexadecane 5.39 n-dodecane 8.83 n- heptane 7.12 water 15.6" 1497 a 12.5 x mol dmP3 [ref.(14)]. Table 5. Effect of p-benzoquinone on the critical micelle concentration of SDS and CTAB at atmospheric pressure and 308 K c.m.c. [p-BQ]/ mol kg-l [SDS]/ lop3 mol dmP3 [CTAB]/ mol dm-3 0 0.5 1 .o 1.5 2.0 8.10 8.07" 8.1b 8.17 8.24 8.3 1 8.39 9.27 9.19" 9.2b 9.61 9.92 10.28 10.60 a At 298 K. At 298 K [(ref. 13)]. properties of the reaction fields change (medium effects). The volume of the CTAB systems had a large negative value ( - 2 2 cm3 mol-l), while that for SDS had a small negative one (-4.7 cm3 mol-l). In contrast, the apparent activation volumes in AOT reversed micelles had small positive values and increased with increasing water content (R) in the interior core (1.5-3.8 cm3 mol-l).Table 4 shows the solubilities of p-BQ in n-alkanes (n-heptane, n-dodecane and1498 Photoreduction of p-Benzoquinone in Micelles 1.2 0 ti 2 ; 1.1 2 1 W Li 1.0 0 1.0 2.0 [p -BQJ/mmol kg-' Fig. 3. Effect of p-benzoquinone on the c.m.c. of SDS (0) and CTAB (0) micellar solutions at atmospheric pressure and 308 K. n-hexadecane) and water at 308 K. These results show that the micelles of SDS and CTAB can solubilize p-BQ in our concentration ranges, and the AOT reversed micelles also can solubilize p-BQ in both the hydrocarbon phase and the water pool. The critical micelle concentrations (c.m.c.) of SDS and CTAB containing p-BQ were determined by the electric-conductivity method and the values obtained are shown in table 5.To check the correctness of these values, the c.m.c. at 298 K was compared with the values in ref. (1 3), and good agreement was obtained. The c.m.c. of both surfactants increased linearly with the concentration of p-BQ. The ratios of the c.m.c. (c.m.c.,,Q/c.m.c.,) are plotted against the concentration of p-BQ in fig. 3. The line for CTAB is much steeper than that for SDS. Discussion Reaction Mechanisms Many mechanisms for the photoreduction of p-BQ under different conditions have been proposed. For example, Kurien and Robins1* showed that the primary photochemical product of aqueous solutions of p-BQ was benzene- 1,2,4-triol, and fast reactions followed which produced hydroquinone and 2-hydroxy- 1,4-benzoquinone. Joschek and Miller'' reported that hydroquinone was the only product in oxygen, but that hydroquinone and benzene- 1,2,4-triol were obtained as main products in a nitrogen atmosphere.One mechanism proposed for the production of hydroquinone in aqueous media is shown in scheme 1. The semiquinone radical is produced in the first step, and disproportionation of the two radicals creates one hydroquinone molecule. Water is considered to be a hydrogen donor in this reaction because the quantum yield does not change with the concentration of p-BQ.16 Generally, the ability of hydrogen donors increases in the following order : aliphatic hydrocarbons, alcohols (water) and metal hydrogenates. The slower reduction of p-BQ in n-heptane solution than in water can beK.Tamura, M . Abe and M. Terai 1499 0 OH P disproportionation 0 + 0 (HZ: hydrogendonor) 0 OH Scheme 1. explained by this order. Although p-BQ molecules can be considered as localized in heptane-like environments according to the A,,, data for p-BQ in AOT reversed micelles (see table l), the reaction rates in AOT reversed micelles are almost the same as those in water. An anionic surfactant (SDS) promoted the photoreduction of p-BQ, while cationic CTAB retarded it (see fig. 1). This suggests that a cationic intermediate is formed and Scheme 2. is concentrated in the anionic micelles or on the anionic surface of the micelles (concentration effects of micelle). Scheme 2 for the carbonyl group may explain the above results. Thus the cationic properties of the carbonyl group promote the reaction in SDS micellar systems.In contrast, the surface of CTAB micelles are cationic, although some parts are neutralized by the Br- counter-ions. In this case the electric repulsion between cations prevents the concentration of p-BQ molecules in the micelles, and this may be one of the causes of the slow reaction rates, If electric repulsion is the only cause of the slower reaction in CTAB micellar systems, the reaction rates in these systems should be faster than that in water, because p-BQ molecules are concentrated in the bulk water phase by the repulsion of p-BQ molecules from the CTAB micelles. However, the reaction in CTAB micellar systems is much slower than that in water. The slower reaction in this system is attributed to Br- ions which generally inhibit the withdrawal of hydrogen radicals from hydrogen donors.Apparent Activation Volumes The features of the apparent activation volumes of the p-BQ photoreduction are as follows : (1) homogeneous solvents and aqueous micellar solutions have small negative values, (2) the CTAB system only has a large negative value and (3) reversed micelles have small positive values. It is known that the activation volumes for thermal reactions are influenced by the nature of the media." In particular, the reactions in which ions are formed or disappear are strongly affected. However, almost all photoreactions we studied in homogeneous media were hardly influenced by moderate pressures up to ca. 150 MPa, and their apparent activation volumes are very small.Pressure affects micellar properties (c.m.c., solubilization, aggregation number etc.), and this often results in large pressure effects on the photoreactions in micellar systems as reaction fields.' Nishikido et al. reported that the solubilization of 2-phenylethanol in CTAB micelles1500 Photoreduction of p- Benzoquinone in Micelles was enhanced with pressure and had a maximum value at ca. 100 MPa.5 It can be seen in fig. 2 that the rate constant of p-BQ photoreduction also has a maximum value at ca. 100 MPa. Accordingly, the large pressure effects on the reaction rates (large activation volumes) in CTAB micellar systems may be explained as follows: the solubilization of p-BQ increases with pressure up to ca. 100 MPa, and p-BQ molecules are concentrated in the micelles.This results in the enhancement of the disproportionation (bimolecular reaction) shown in scheme 1. Critical Micelle Concentration Higher alcohols and n-alkanes'' generally depress the c.m.c., and this is due to the solubilization of these non-electrolytes into the micelle core. In contrast, water-soluble substances such as acetone, ethylene glycol, dioxane or the lower alcohols1s-21 increase the c.m.c. In these cases additives greatly change the nature of the water phase; therefore, the partition equilibrium of the additives between solvents and micelles and changes in micellar properties become important. These additives depress the dielectric constants of the aqueous solutions and also break the water structure around the hydrocarbons of the surfactants.Both effects should raise the c.m.c. p-Benzoquinone linearly and effectively increases the c.m.c of CTAB and SDS in its lower concentration range (fig. 3). The concentration ranges ofp-BQ used in our study are much lower than those of the additives previously described. Therefore, large changes in the nature of the water phase cannot be expected. p-Benzoquinone effectively breaks the water structure around the hydrocarbons of the surfactants. CTAB molecules have longer hydrocarbons than SDS, and these differences affect the concentration effects on the increase in c.m.c. Furthermore, three methyl groups on a nitrogen atom promote water structure, so that the effects of p-BQ are more significant. A change in the effect of pressure on the c.m.c.with p-BQ concentration is not clear; therefore, these c.m.c. data cannot be used to predict the effects of pressure on reaction rates. References 1 K. Tamura and M. Aida, J. Chem. Soc., Faraday Trans. I , 1986, 82, 1619. 2 D. Hamann, J. Phys. Chem., 1962,66, 1359. 3 R. F. Tudden and A. E. Alexander, J. Phys. Chem., 1962, 66, 1839. 4 S. Kaneshina, M. Tanaka, T. Tomida and R. Matuura, J. Colloid Interface Sci., 1974, 48, 450. 5 N. Nishikido, M. Kishi and M. Tanaka, J. Colloid Interface Sci., 1983, 94, 348. 6 N. Nishikido, M. Shinozaki, G. Sugihara, M. Tanaka and S. Kaneshina, J. Colloid Interface Sci., 1980, 7 K. Tamura and M. Suminaka, J. Chem. Soc., Faraday Trans. 1 , 1985, 81, 2287. 8 K. Tamura and S. Sugiyama, Bull. Fac. Eng. Univ. Tokushima, 1984, 21, 33. 9 K. Tamura and S. Sugiyama and W. J. le Noble, J. Org. Chem., 1984, 49, 3836. 74, 474. 10 K. Kano and T. Matsuo, Chem. Lett., 1973, 1127. 11 K. Tamura and N. Nii, J. Phys. Chem., in press. 12 M. Zulauf and H. F. Eicke, J. Phys. Chem., 1979, 83, 480. 13 J. H. Fendler and E. J. Fendler, Catalysis in Micellur and Macromolecular Systems (Academic Press, 14 K. C. Kurien and P. A. Robins, J. Chem. SOC. B, 1970, 855. 15 H. 1. Joschek and S. J. Miller, J. Am. Chem. Soc., 1966, 88, 3273. 16 S. Hashimoto, K. Kano and H. Okamato, Bull. Chem. Soc. Jpn, 1972, 45, 966. 17 K. E. Weale, Chemical Reactions at High Pressures (Spon, London, 1967), p. 137. 18 W. D. Harkins, R. Mittelman and M. L. Corrin, J. Phys. Colloid Chem., 1949, 53, 1350. 19 B. D. Flockhart, J. Colloid Sci., 1957, 12, 557. 20 K. Shirahama and R. Matuura, Bull. Chem. SOC. Jpn, 1965, 38, 373. 21 A. W. Ralston and D. N. Eggenberger, J. Phys. Colloid Chem., 1948, 52, 1494. New York, 1975), p. 20. Paper 8/03056A; Received 28th September, 1988

ISSN:0300-9599    

DOI:10.1039/F19898501493

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1989, 85(6), 1501-1509 Diffusion of Ethane in Silicalite- 1 by Frequency Response, Sorption Uptake and Nuclear Magnetic Resonance Techniques Neil Van-Den-Begin and Lovat V. C. Rees* Physical Chemistry Laboratories, Imperial College of Science and Technology, London SW7 2AY Jurgen Car0 and Martin Bulow Zentralinstitut f u r Physikalische Chemie, Akademie der Wissenschaften der DDR, Rudower Chaussee 5, Berlin - 1199, G.D.R. Michael Hunger and Jorg Karger Karl-Marx- Universitat Leipzig, Sektion Physik, Linnkstr. 5 Leipzig - 701, G.D.R. The diffusion of ethane in ZSM-5 has been studied using four different experimental methods : a frequency-response technique, a single-step frequency-response method (an improved sorption uptake-desorption technique), an n.m.r.pulsed field gradient spin+xho technique and, finally, an n.m.r. tracer desorption technique. Large and uniform silicalite- 1 crystals were prepared and used in all four techniques. Different concentrations of tetrapropylammonium ions in the synthesis batch alter systematically both the crystal habit and the relative amounts of internal silanol groups in the silicalite- 1 framework, indicating different concentrations of defect Si-0-Si bonds. For the genesis of the defective siloxane bonds, a model has been presented. Comparing the translational mobility of ethane on a selected sample, the diffusion coefficients measured by the two n.m.r. methods were found to be ca. 150 times larger than the corresponding coefficients derived from the frequency response and sorption uptake- desorption methods.This experimental finding is explained by crystal bed depth influences on the molecular uptake in the latter sorption experiments. In recent years, a series of experimental and theoretical papers have reported various factors which affect rates of sorption, e.g. external and intercrystalline mass transport, sorption heat and various types of zeolite crystal surface barriers. 1-6 Severely discrepant diffusion data, however, are still appearing in the literature and a more complete elucidation of the causes of these discrepancies is needed. Besides theoretical investigations' the application of different experimental techniques3. 5, should give some indication of the factors which produce these differences in the diffusion coefficients. Owing to the highly interesting sorptive and catalytic properties of ZSM-S/silicalite molecular sieves,*-1o diffusion measurements of ethane in silicalite- 1 will be reported in this paper using the frequency-response 11-16 adsorption-desorption kinetic^^'-^' and n.m.r.self-diffusion measurements. 20-23 Experiment a1 (a) Frequency-response (f.r.) Technique The f.r. technique operates under near-equilibrium sorption conditions. In this method, the volume of the chamber containing the gas-zeolite system at sorption equilibrium is 15011502 Diflusion of Ethane in Silicalite- 1 2.0 1.5 s“ 1.0 0.5 ethane/silicalite- 8 Torr, 50 “C 0 I I I , 0.01 0.10 1.00 lo. 00 frequency /Hz Fig. 1. as the characteristic functions (a), {[( V / P ) x cos $1 - l}, and (b), (V/P) x sin $, derived from the experimental data in fig.3. The continuous lines represent the best fit of the experimental f.r. data assuming a diffusion coefficient of ethane in silicalite of 9.5 x 10-l2 m2 s-l. modulated by a square-wave perturbation over a range of frequencies. These expansionxompression cycles lead to alternating desorption and adsorption processes, respectively, in the sorbed phase. The phase angle and amplitude of the pressure change induced by the volume variation are measured as a function of the applied frequency. So called ‘characteristic functions ’ 6 y 11-16 (cf. fig. 1) can be calculated assuming Fick’s diffusion law. In the f.r. apparatus used at the Imperial College (IC), London, the amplitude of the volume variation and, thus, the pressure change is limited to <2%.The oscillation frequency is varied over 3 orders of magnitude from 0.01 to 10 Hz. (b) Single-step f.r. Method In the f.r. experiments, during each square-wave, there is a very rapid but small change (f 1 %) in volume of the gas phase. Following such perturbations, the pressure is measured 64 times during each cycle and recorded on an on-line computer. By analysing this pressure change as a function of time an unambiguous determination of the rates of adsorption-or des~rptionl’-~~ can be realised. The response time of the IC f.r. system is 25-35 ms which is ca. 10 times faster than the piezometric meth~d~’-~’ used at the Central Institute of Physical Chemistry, Berlin. By fitting appropriate solutions of Fick’s diffusion law to the sorption-desorption curves obtained in these single steps at the lower frequencies of the f.r.measurements, diffusion coefficients can be calculated. However, due to the changing pressure with adsorption-desorption during a single step, non- fixed boundary conditions have to be considered. Therefore, the sorption curves were analysed using the equation for the diffusion controlled ‘uptake by a sphere from a stirred solution of limitedJ . Chem. SOC., Faraday Trans. I , Vol. 85, Part 6 N. Van-Den-Begin and others Plate I 43 0 (Facing p . 1503)N . Van-Den-Begin et al. 1503 Table 1. Batch compositions and the concentrations of internal silanol groups as determined by 'H-MAS n.m.r. sample TPAOH per 96 SiO, no. of internal SiOH groups crystal no.in the starting gel per g of silicalite- 1 / 1 OZo size/pm 1 2 3 2.36 5.50 7.88 3.3 4.7 5.4 2.9 x 7.9 x 17.1 20.1 14.4 (c) N.M.R. Self-diffusion Measurements For the determination of the intracrystalline self-diffusion coefficient, Di = ( r2(A))/6A, the r.m.s. molecular displacements ( r2(A)>'j2 of the sorbed molecules during an observation time A have been measured by the n.m.r. pulsed field gradient technique.20-22 The measurements were performed using the n.m.r. pulse spectrometer FEGRIS at the Department of Physics at the Karl Marx University of Leipzig. (d) N.M.R. Tracer Desorption Technique N.m.r. tracer desorption experiments allow the determination of the fraction of molecules which leaves the intracrystalline space of a zeolite crystal during a fixed observation time.The latter can be varied over a few milliseconds, under the condition of macroscopic sorption equilibrium. 21-23 By fitting the corresponding solution of Fick's law for diffusion-limited desorption for fixed boundary conditions to the n.m.r. tracer desorption curves thus obtained, a diffusion coefficient D, can be determined. These measurements were also carried out on the above-mentioned n.m.r. spectrometer. Results (a) Influence of TPA Ions on Crystal Morphology and the Concentration of Internal Silanol Groups The silicalite- 1 used in these studies was synthesised by established hydrothermal p r o c e d ~ r e s . ~ ~ The starting gel was prepared by adding a TPA-water solution (20 wt. % TPAOH) to a Ludox solution (40 wt. % colloidal SO,) while stirring.The resultant solution, which was translucent white, was stirred for 1 h at room temperature. Crystallisation was carried out in rotating autoclaves lined with PTFE at 175 "C. The change of the crystal morphology with varying TPAOH concentration in the synthesis batch is displayed in plate 1. Whereas the crystals of sample 1 in plate 1 may be regarded as mono-crystals, it is clearly seen, that with rising TPAOH concentration the crystals become increasingly twinned in the (010) face. This finding is in accordance with similar results reported in the literature.26. 27 Table 1 shows the starting gel compositions and the concentration of internal silanol groups for different silicalite-1 crystals after calcining at 550 "C for 12 h. Table 1 also shows that the concentration of internal silanol groups in the silicalite- 1 crystals increases with increasing TPAOH concentration. Since the number of non-acidic silanol groups as measured by the 'H-MAS n.m.r. is ca.lo3 larger than the maximum number of terminal silanol groups on the external crystal surface, this experimental finding can be considered as unambiguous proof of the occurrence of structural defects within the bulk phase of the silicalite- 1 crystals. Other- wise the measured concentrations of silanol groups should be proportional to the outer1504 I 10 - a d 5 - Digision of Ethane in Silicalite- 1 Q01 0.1 1 10 frquency/Hz 3.01 0.01 0.1 1 10 frequency /Hz Fig. 2. (a), Phase lag A#, and (b), amplitude ratio ( V / P ) , from f.r. analysis of the sorption system ethane-silicalite- 1 (sample 3 of table 1) at a total pressure of 8 Torr at 50 "C.The continuous lines represent the best fit of the experimental data using the appropriate solutions of Fick's diffusion laws and a diffusion coefficient of 9.5 x 10-l2 m2 s-l. surface area of the individual crystals, i.e. inversely proportional to the crystal size, whereas table 1 shows the opposite trend. These structural defects are due, therefore, to defect Si-0-Si bonds. In agreement with previous finding^,^*-^' the presence of TPA ions during pentasil-type synthesis has been found to induce high concentrations of structural defects in the form of broken Si-0-Si bonds. The role of TPA ions in such syntheses can be explained in terms of a specific crystallisation mechanism assuming the condensation of partially hydrolysed double five ring 32 Especially in silicate solutions containing tetraalkylammonium ions, increased concentrations of double ring silicates are present.In the case of ZSM-5 synthesis, the condensation of double five ring silicates leads to a MFI lattice with unoccupied T7 and T12 tetrahedral ~ i t e s . ~ ~ ? ~ ~ Hence, up to 16 vacancies may be present per unit cell, each of which may involve a nest of four, non-acidic, hydroxyl groups. However, as shown in ref. (28), such structural defects only slightly modify the sorption and diffusion properties of hydrocarbons in ZSM-5. In agreement with this finding, in the present study, the intracrystalline self-diffusion coefficients of ethane in the samples 2 and 3 have been found to be comparable.Due to the relatively small crystal size of sample 1, the intracrystalline self-diffusion coefficient could not be determined since the r.m.s. displacements of the sorbed ethane molecules were comparable to theN . Van-Den-Begin et al. 1505 10.1 10.0 t c & \ 2 9.9 v1 9.8 9.7 timels - Fig. 3. Pressure recording in the single step f.r. method following a fast volume perturbation which in the case of compression (a) causes adsorption to occur, and in the case of expansion (b) desorption. The sorption system was ethane-silicalite (sample 3 of table l), at an equilibrium pressure of ca. 10 Torr and temperature of 0 "C. crystal diameter and, hence, too large for quantitative estimations. Therefore, sample 3 was chosen for the comparative diffusion measurements presented in the following section. (b) Diffusion Measurements Using Several Experimental Techniques As an example of the f.r.measurements, fig. 2(a) and (b) show the experimentally determined values of the phase lag, A$, and the amplitude ratio (V/P), respectively, as a function of the frequency of the volume perturbations for the ethane-silicalite- 1 system (sample 3 of table 1) at 8 Torr and + 50 "C (sorbate loading ca. 0.05 mmol g-'). Fig. 1 contains the characteristic functions {[( V / P ) x cos $1 - 1) and ( V / P ) sin $ representing the in-phase and out-of-phase components of the pressure variation. The continuous lines in fig. 1 and 2 represent the best computer fit of the experimental f.r. data. This fitting procedure leads to a diffusion coefficient of D = 9.5 x lo-'' m' s-l.In fig. 3, using the single step f.r. method, the adsorption and desorption rates of ethane on the sample at a pressure of 10 Torr and 0 "C are presented (loading x 0.5 mmol g-'). By fitting appropriate solutions of Fick's law for 'uptake from a stirred solution of limited volume"* to the experimental points in fig. 3, a diffusion coefficient of ethane in silicalite-1 of D = 4.5 x lo-'' m2 s-' at 0 "C was obtained. From the corresponding measurements at -23 "C a value of D = 1.5 x lo-'' m' s-' was determined (loading = 1.05 mmol g-'). Table 2 shows the intracrystalline n.m.r. self-diffusion coefficients, Di, of ethane in silicalite-1 directly measured by the n.m.r. pulsed field gradient technique and the self- diffusion coefficient, D,, determined experimentally by the n.m.r.tracer desorption technique at sorption equilibrium. In the tracer desorption experiments, a self-diffusion coefficient D, was calculated from the percentage of the ethane molecules which undergo intra-/inter-crystalline1506 Diflusion of Ethane in Silicalite- 1 Table 2. Intracrystalline self-diffusion coefficient, Di, and the tracer desorption coefficient, D,, at different ethane loadings. The latter coefficient was calculated on the basis of Fick’s law using the amounts of fractional desorption in silicalite-1 (sample 3 of table 1) as a function of time at room temperature time taken for 50% exchange of molecules sorbed in a single sample loading /mmol g-l Di/m2 s-l crys tal/ms Dd/m2 s-l 1.4 1.9 x 10-9 1.2 1.3 x 10-9 1.7 1.4 x 10-9 1.4 1.1 x 10-9 2.1 0.9 x 10-9 1.9 0.8 x 10-9 exchange, i.e.leave the intracrystalline phase of an individual silicalite- 1 crystal with a mean diameter of 14.4pm (cf. sample 3 in table 1) during observation times in the millisecond range. Fick’s law for diffusion-controlled desorption under constant boundary conditions was used. Discussion The following points can be concluded from a comparison of the diffusion coefficients obtained by the four methods. (i) The diffusion coefficients derived from the f.r. technique and from sorption uptake-desorption measurements by means of the single step f.r. method are comparable. This consistency lends support to the correctness of the relatively complicated mathematical procedure used to evaluate the f.r.data. However, this result is not unexpected since the f.r. technique is based on the analysis of consecutive single-step adsorption-desorption cycles. (ii) The self-diffusion coefficients Di measured directly from the r.m.s. displacement data by means of the n.m.r. pulsed field gradient technique and D, derived from the mean intracrystalline ‘life time’ of sorbed molecules as obtained from the n.m.r. desorption technique are, also, of the same magnitude. This finding indicates that for the case of sorption equilibrium (self-diffusion conditions) molecular exchange between the individual silicalite- 1 crystals of the sample under study is governed by intracrystalline self-diffusion, i.e. no structural surface barriers exist.3* 5 9 22* 23 The data obtained agree well with earlier published self-diffusion coefficients of paraffin-pentasil (iii) The diffusion coefficients derived by means of sorption techniques such as f.r.and single step f.r. experiments are ca. 150 fold smaller than the corresponding self-diffusion coefficients measured by the two n.m.r. techniques. These coefficients are also smaller than those obtained for the propane-ZSM-5 system by means of piezometric investigation^,^^ at which only a lower limit of intracrystalline mobility coefficients could be indicated. (iv) Following the formalism of irreversible thermodynamics, the diffusion coefficient determined by analysing diffusion flux densities under non-equilibrium sorption conditions should be expected to be either equal [in the case of a linear sorption isotherm (Henry’s law)] or larger (for convex shaped isotherms such as those of the Langmuir type) than the self-diffusion coefficient measured at sorption eq~ilibrium.~~ 35 No such behaviour is reflected by the above results.It is necessary therefore, to ascertain whether the diffusion coefficients derived from sorption-desorption experiments are affected by intercrystalline diffusion (bed depth) influences and/or sorption heat-release processes. Whereas in gravimetric or piezometric experiments sorption uptake-desorption rates are measured using minimal amounts of zeolite crystals (e.g. 20 mg and ca. 5 mg,N . Van-Den-Begin et al. 1507 respectively), the present f.r. experiments have used 0.8-2.5 g of zeolite resulting in bed depths of 1-5 mm.Thus both bed depth and sorption heat release influences on the mass transport must be considered. The inter- and intra-crystalline mass transport and the corresponding heat dissipation processes will be dependent on their respective time constants. These time constants will be affected by cross-coefficient terms.35 However, in the system considered in this study, the time constants for the mass transport processes will be much greater (ca. 100 x ) than those for heat dissipation. The influence of the sorption heat dissipation on the sorption kinetics may be estimated from the quantity35 a’ = (R;/D)/(C,/ha) which is proportional to the ratio between the time constants of intracrystalline adsorption4esorption diffusion in the individual crystal and the heat dissipation from the assemblage of zeolite crystals, with the quantities R,, Cs, h and a denoting, respectively, the crystal radius, the sorbent heat capacity, the heat transfer coefficient and the outer surface area of the zeolite assemblage.Obviously, the limiting cases of isothermal and heat dissipation controlled adsorption4esorption correspond to the values a’ = co and a’ = 0, respectively. If the crystal bed is, simply, assumed to be in the shape of a sphere with radius R,, one obtains with c and p denoting the specific heat capacity of the zeolite crystals and the effective assemblage density, respectively. Inserting representative values from the literature (h = 20 J m-2 s-l K-l, c = 1 J g-l K-l, p = 1 g ~ m - ~ ) with R, = 7 ,urn, R, = 1 mm and D = 3 x lo-‘’ m2 s-l as obtained from the n.m.r.experiments as a lower limit of intracrystalline mobility, the parameter a’, is found to be of the order of lop2. Since the intracrystalline diffusion proceeds comparatively quickly,33 this rough estimate indicates that under the given experimental conditions both the adsorption and desorption rate of ethane on silicalite-1 could be influenced by heat-release effects. To distinguish between the two effects (intercrystalline diffusion and superimposed heat transfer) and to eliminate these intercrystalline influences on the intracrystalline rate behaviour, it would be necessary to change the experimental conditions, i.e. the bed depth of crystals would need to be minimized. Despite the discrepancy in the diffusion coefficients obtained by the above four methods, both the f.r.technique and the single step f.r. method have been shown to be powerful tools for observing fast sorption processes. Whereas in conventional gravimetrically controlled sorption kinetic experiments the half-time of sorption kinetics is greater than 2 s, and in piezometric rate experiments the statistical moment of the blank curve of the gas expansion into the sorption vessel, (without sorbent) has a characteristic time resolution of ca. 0.1 2 ’ 7 l8 the single step mode of the f.r. technique allows these sorption kinetics to be measured with time constants of 25-35 ms. The complete f.r. analyses of these sorption kinetics can shorten even these latter times. The only problem to be overcome is the avoidance of bed-depth influences.The latest developments in the f.r. method have managed to eliminate these bed-depth Conclusions The amount of structural defects, present as defect Si-0-Si bonds in the silicalite-1 framework, was found to be proportional to the TPA concentration of the synthesis gel. The defects seem to have little effect on the diffusion of ethane in silicalite-1 channels. The concentration of TPA ions in the starting gel also determines the morphology of the silicalite- 1 crystals.1508 Difusion of Ethane in Silicalite- 1 Diffusion experiments performed by the frequency response and the single-step frequency-response method (an improved sorption uptake-desorption technique) provide similar values of the diffusion coefficient of ethane in silicalite-1.The self- diffusion coefficients of ethane in the same silicalite-1 sample measured by the n.m.r. pulsed field gradient and the n.m.r. tracer desorption techniques, were found to be the same and to be ca. 150 times larger than the diffusion coefficients derived from frequency-response measurements. These experimental findings indicate that the adsorption-desorption rates obtained in the frequency-response and sorption kinetic experiments are influenced by a combination of both intercrystalline diffusion and sorption heat-release processes with the former being the dominant influence. By minimising these bed effects, we hope to obtain agreement in the diffusion coefficients obtained from all four techniques. N. Van-Den-Begin and L.V. C . Rees thank Dr S. A. I. Barri (B.P. Sunbury-on- Thames) for his advice in the preparation of the zeolite samples. The authors acknowledge the support given by the Royal Society (London) and the Academy of Sciences (Berlin, GDR) under the Joint Research Project ‘Adsorption on Zeolites ’ which was started in 1986. References 1 D. Prinz and L. Riekert, Ber. Bunsenges. Phys. Chem., 1986, 90, 413. 2 L. K. Lee and D. M. Ruthven, J. Chem. SOC., Faraday Trans. 1, 1979, 75, 2406. 3 M. Bulow, J. Karger, M. KoEiiik and A. M. VolosEuk, Z . Chem., 1981, 21, 175. 4 R. Haul and H. Stremming, J . Colloid Interface Sci., 1984, 97, 348. 5 M. Bulow, Z . Chem., 1985, 25, 81. 6 M. Bulow, H. Schlodder, R. E. Richards and L. V. C. Rees, New Developments in Zeolite Science and Technology, ed.Y. Murakami, A. Iijima and J. W. Ward (Kodansha Elsevier, Tokyo, 1986), pp. 579-586. 7 M. KoEifik, P. Struve, K. Fiedler and M. Bulow, J. Chem. SOC., Faraday Trans., in press. 8 D. H. Olson, G. T. Kokotailo, S. L. Lawton and W. M. Meier, J . Phys. Chem., 1981, 85, 2238. 9 E. M. Flanigen, J. M. Bennett, R. W. Grose, J. P. Cohen, R. L. Patton, R. M. Kirchner and J. V. 10 Innovation in Zeolite Materials Science, ed. P. J. Grobet, W. J. Mortier, E. F. Vansant and G. Schulz- 11 L. V. C. Rees, in Structure and Reactivity of ModiJied Zeolites, ed. P. A. Jacobs et al., (Elsevier, 12 Y. Yasuda, J . Phys. Chem., 1976, 80, 1867; 1870. 13 M. Betemps, Proc. 5th Intern. Conf. Zeolites, Naples, ed. L. V. C. Rees, (Heyden, London, 1980), pp. 14 Y. Yasuda and M.Saeki, J. Phys. Chem., 1978, 82, 74. 15 Y. Yasuda, J . Phys. Chem., 1982, 86, 1913. 16 Y. Yasuda and A. Yamamoto, J . Catal., 1985, 93, 176. 17 P. Struve, M. KoEiiik, M. Bulow, A. Zikanova and A. G. Bezus, Z . Phys. Chem. (Leipzig), 1983, 264, 18 M. Bulow, W. Mietk, P. Struve and M. KoEiEik, J. Chem. SOC., Faraday Trans. I , 1984, 80, 813. 19 M. Bulow, P. Struve and W. Mietk, Z . Phys. Chem. (Leipzig), 1986, 267, 613. 20 J. Karger and W. Heink, J. Magn. Reson. 1983, 51, 1. 21 J. Karger and H. Pfeifer, Zeolites, 1987, 7, 90. 22 J. Karger, H. Pfeifer and W. Heink, Adv. Magn. Reson., 1988, 12, in press. 23 J. Karger, AIChE J., 1982, 28, 417. 24 J. Crank, The Mathematics of Diffusion (Clarendon Press, Oxford, 1956). 25 R. T. Argauer and G. R. Landolt, US Patent 3 702 886, (1972). 26 V. N. Romannikov, V. M. Mastikhin, S. Hocevar and B. Drzaj, Zeolites, 1983, 3, 311. 27 R. Mostowicz and L. B. Sand, Zeolites, 1985, 3, 219. 28 M. Hunger, J. Karger, H. Pfeifer, J. Caro, B. Zibrowius, M. Bulow and R. Mostowicz, J . Chem. SOC., 29 H. Pfeifer, D. Freude and M. Hunger, Zeolites, 1985, 5, 273. 30 M. Hunger, D. Freude, T. Frohlich, H. Pfeifer and W. Schwieger, Zeolites, 1987, 7, 108. 31 E. J. J. Groenen, A. G. T. G. Kortbeek, M. Mackay and 0. Sudmeijer, Zeolites, 1986, 6, 403. Smith, Nature, 1978, 271, 512. Ekloff, Stud. Surf. Sci. Catal. (Elsevier, Amsterdam, 1988), vol. 37. Amsterdam, 1984), pp. 1-1 1. 526-534. 49. Faraday Trans. I , 1987, 83, 3459.N . Van-Den-Begin et al. 1509 32 R. A. van Santen, J. Keijsper, G. Ooms and A. G. T. G. Kortbeek, Stud. Surf. Sci. Catal., 1986, 28, 33 H. Pfeifer and S. P. Zhdanov, J. Chem. Soc., Faraday Trans. I , 1985, 81, 2541. 34 M. Bulow, H. Schlodder and P. Struve, Ads. Sci. Technol. 1986, 3, 229. 35 D. M. Ruthven, Principles of Adsorption and Adsorption Processes (Wiley-Interscience Publ., New 36 L. V. C. Rees and N. Van-Den-Begin, contribution to the 8th Int. Conf. Zeolites, Amsterdam, 1989, in Paper 8/0333A; Received 16th August, 1988 169. York, 1984). preparation.

ISSN:0300-9599    

DOI:10.1039/F19898501501

出版商:RSC    

年代:1989

数据来源: RSC