J . Cheni. Soc., Farackij. Truns. I , 1989. 8 3 1). 3 3 4 5 Temperature-programmed Desorption of Ethanol from ZSM-5, ZSM-11 and Theta-1 Chen Li-feng,? Thomas Wacker$ and Lovat V. C. Rees* Department of Chemistry, Impwiirl c'olltyc of Science und Technologj., Lomion SW7 2'4 Y Ethanol was adsorbed on low Si/AI ratio ZSM-5. ZSM-I I and Theta-l at room temperature and relative prcssure p/p(, = 0.5. in situ. in a thermal gravimetric balance. The saturation capacity of ZSM-I 1 was found to be slightly greater than ZSM-5 and cci. twice as large as that of Theta-I. Temperature-programmed desorption profiles of ethanol were determined and analysed by ( u ) a variable heaLing rate and ( h ) single heating rate method using an on-line computer. The effect of cation exchange on the sorption capacities and activation cnergieq of desorption have also been determined. In previous studies the temperature-programmed desorption (t.p.d.1 of non-polar hydrocarbons from low silica-alumina ratio ZSM-5, ZSM-I 1 and Theta-l have been reported.'.' In the present paper the t.p.d.of the polar molecule, ethanol, from the same zeolites is studied. Comparison with the former results shows that the electrostatic interaction between the permanent dipole moment of ethanol and the zeolite channel surface is strong. This electrostatic interaction tends to overcome the steric hindrance of the cations present in these zeolite channels. This hindrance is quite important in the packing of hydrocarbon molecules. '. The t.p.d. of ethanol from the above zeolites in their H' and Na' forms demonstrates that increasing the Na content decreases the amount of ethanol adsorbed slightly and increases the activation energy of desorption. As before,'.' two methods of analysing the t.p.d.profiles have been used, i.e. the variable heating rate (v.h.r.) and the single heating rate (s.h.r.) methods. Experimental Details of the thermogravimetric balance are given in ref. (1) and (3). The unit-cell formulae of the low Si/AI ratio zeolites used in this study are as follows: ZSM-S(§i/Al = 15) : H, ,,Na, 52A16Si900192 ZSM- 1 1 (Si/Al = 15) : H, ,,Na, 64Al,Si,,,01,2 Theta- I (Si/AI = 32) : H,, ,,Na, ,,Al, 91Si93,09019z. The sodium-exchanged forms of these three zeolites were prepared by ion exchange with 0.1 mol dm-3 NaCl solution under a controlled pH of 10.The sodium contents of these samples were determined by neutron activation and an Na/A1 ratio of 1 was found, within experimental error, for all three samples. The hydrogen form of Theta-l was prepared by ion exchange with 0.1 mol dm '' NH,Cl at pH 5 followed by calcination at 500 "C. The unit-cell formula of this sample was found to be Na,,,H,,,Al,,,Si,,,,,O,,,. i- Present address : Chemical Engineering Department. Nanchang Institute of Aeronautical Technology, No. 2 Shanghai Road, Nanchang, Jiangxi. The People's Republic of China. Present address : Weidhoelzliweg 19, 5024 Knettigen. Switzerland. 33 2-234 T.P.D. of Ethanol from Zeolites Table 1. Comparison of the amount of ethanol adsorbed under various saturation conditions experimental conditions ZSM-5 ZSM- 1 1 Theta- 1 p / p , z 0.5; room 13.68 14.12 6.42 temperature ; sorption time, 30 minu temperature ; sorption time, 4 hb 1 h at 50 "C then 3 h at room temperatureb p / p , x 0.5; room 13.62 14.36 6.51 p / p , x 0.5; sorption for 13.73 14.18 6.30 The saturation capacities are given as molecules per unit cell.a Average value from the 10 different heating rate experiments. Determined from 10 K min-' run. The sodium content of this sample was determined after dissolution of the sample by atomic absorption spectroscopy. Absolute alcohol (A. R. quality, > 99.7%) was used. The reproducibility of the t.p.d. profiles was very good. Results and Discussion Various, in situ experimental conditions were tested to ensure complete saturation of the sample with adsorbate prior to starting the t.p.d.runs. The results of these preliminary experiments are given in table 1. Table 1 shows that the amount of ethanol sorbed in ZSM-11 is slightly greater than that in ZSM-5 and much greater than that in Theta-1. The sorption of various hydrocarbons was found previously' to follow the same pattern. From the saturation amounts sorbed and a density of 0.7893 g cm-3 for liquid ethanol, micropore volumes, W,, of 0.138, 0.142 and 0.064 cm3 g-l for ZSM-5, ZSM-11 and Theta-1, respectively, were calculated. These sorption volumes are much lower than the theoretical volumes of the channels of these zeolites. The micropore volumes calculated from the saturation capacities of five different adsorbate molecules are compared in table 2.The W, value of ethanol in ZSM-5 is close to the corresponding value in silicalite. The low saturation sorption capacity of ethanol in silicalite was ascribed to the hydrophobic character of the channel s ~ r f a c e . ~ The surface of the low Si/A1 ratio ZSM-5 sample used in the present study should be less hydrophobic than that of silicalite and the low sorption capacity of ethanol in this sample must reflect, therefore, some steric hindrance of the cations present in the channels. However, the larger sorption capacities of n- hexane and n-octane over ethanol, in table 2 for the ZSM-5 sample indicates that the surface of these low Si/Al ratio zeolites is still dominantly hydrophobic in character. Ethanol T.P.D. Studies The t.p.d. profiles of ethanol from ZSM-5 (Si/Al = 15), ZSM-11 (Si/Al = 15) and Theta-1 (Si/Al = 32) are shown in fig.1-3 along with the differential of these profiles.C/tm LiTfeng, T. Wuckc~ cind L. P'. C. Reps 35 n-hexaneb n-octaneb p-x ylene' benzenec ethanol 10 % > 0.5 E a Table 2. y, (cm3 g') obtained for different adsorbate molecules ~~~ ~~ - ~~ ZSM-5 ZSM- 1 1 Theta- 1 (SI/AI = 15) (SI/AI = 15) (Si/AI = 32) sillcall tea -~ ~ 0. I50 0 156 0.060 0.173 0.143 0.147 0.055 0.149 0.128 0.133 0.042 0 181 0.108 0.1 10 0.042 - 0.138 0.143 0.064 0.136 I' Ref. (3). ', Kcf. ( I ). Ref. (2). I 0 I 3 0 II] U 14.0 10.5 ..a .- c b a 7.0 2 - ; 8 3.5 0 300 350 400 450 500 550 600 650 TIK Fig. 1. T.p.d. profile and differential of profile of ethanol in (NaH)ZSM-5. /3 = 10 K min-'. 1.0 6 2 05 a 0 0 -0 \ 0 0 0 a U T- !3 14.0 I - 10.5 .C c b a 7.0 3 3 8 I 2 3.5 0 300 350 400 450 500 550 600 650 T/K Fig.2. T.p.d. profile and differential of profile of ethanol in (NaH)ZSM- 11. j3 = 10 K min-'.36 T.P.D. of Ethanol from Zeolites 1.0 9 3 .. Q5 0 0 -0 I 0 0 0 0 0 0 n o D U . I I 300 350 400 450 500 550 600 650 T/K Fig. 3. T.p.d. profile and differential of profile of ethanol in (NaH)Theta-1. /? = 10 K min-'. Table 3. Peak temperatures and peak widths (B = 10 K min-') peak temperature/K peak width/# major peak I peak I1 peak peak I peak I1 diffuse 425490 ZSM-5 diffuse 465 peak I1 ZSM- 1 1 320 470 peak I1 300-380 430-490 Theta- 1 320 510 peak I 30&380 48&550 The peak temperatures and peak widths are listed in table 3. Note that the low- temperature peak at 320 K found with ZSM-11 and Theta-1 is absent or very diffuse with ZSM-5.This result suggests that the sorption sites in ZSM-5 from which the first seven molecules of ethanol are desorbed have gradually increasing energies. ZSM- 1 1 and Theta-1 both have sets of low-energy sites from which the first seven and two molecules, respectively, are easily desorbed at low temperature (i.e. ca. 320 K). Comparison of the peak temperature of ZSM-5 of 465 K with that of silicalite3 of 340 K indicates that the electrostatic interaction between the permanent dipole moment of ethanol and the sites in ZSM-5 (Si/Al = 15) with large electric fields is obviously strong. A small high-temperature peak was found above 625 K with ZSM-5 when heating rates of 2 or 4 K min-' were used. This small peak must be the result of some catalytic reaction of ethanol molecules strongly associated with framework aluminiums or acid sites. At higher heating rates the small peak is not observed, probably because the catalytic reaction is spread out over a larger temperature The second peak observed with Theta-1 occurs at a higher temperature than that found with ZSM-5 and ZSM-11, probably due to the higher Na content of Theta-1 and the lack of cavities in the one-dimensional channel structure.The electrostatic interaction between the permanent dipole moment of ethanol and the walls of the channels of Theta-1 must be stronger than that with ZSM-5 and ZSM-11.1LO 120 - 100 d 8 c1 \ 4" 80 60 40 37 0 2 i 6 a 10 12 14 molecules per unit cell Fig. 4. Activation energy of desorption of ethanol as a function of coverage.0. ZSM-5: 0, ZSM-I I : 0. Theta-1 in their (haH)-forms. V.h.r. method of analysis. Analjtsis of T.P. D. Profiles by the V.H. R. Mcthod Heating rates at 2 K intervals between 2 and 20 K min-' were used as before.' ' The activation energies, E(I, and entropies, -AS:, of desorption as a function of coverage derived from the v.h.r. method of analysis:' are shown in fig. 4 and 5 , respectively. The activation energy, averaged over the whole coverage range, Eli, for the three zeolites is given in table 4. All three zeolites show high E,, values at low coverages, consistent with the presence of sites with high electric fields in these low Si/Al samples. The variation of El, with coverage is similar for both ZSM-5 and ZSM- 1 1.as would be expected for two zeolites with very similar channel structures and the same Si/AI ratio. The variation of E,, with coverage for ZSM-5 is very similar in form to that found for the differential molar heats of sorption, (2, of ethanol in silicalite - 1 ;' but Et1 is cu. 30 kJ molpl greater than Q at all corresponding sorbate loadings. In both samples 12 ethanol molecules per unit cell out of the total sorption capacity of 14 molecules are desorbed with near-constant Eli values of ca. 8&90 kJ mol-'. Although ethanol has been reported to be dehydrated in H-forms of ZSM-5 at quite low temperatures, there is no clear evidence in these desorption energy distributions of any catalytic activity. Fig. 1 shows that the desorption of these 12 ethanol molecules from ZSM-5 cover a temperature span of 25-180 "C.i.0. the last of these ethanol molecules is being desorbed at 180 "C where some catalytic activity would be expected. However, Ed remains sensibly constant over this coverage range and similar in form to the variation of Q in silicalite- I , where no catalytic activity would be expected.38 T.P.D. of Ethanol from Zeolites 200 160 - k 120 - 2 h \ t m I a 80 LO 0 0 2 4 6 8 10 12 1L molecules per unit cell Fig. 5. Activation entropy of desorption of ethanol as a function of coverage. 0, ZSM-5; 0, ZSM-11; 0, Theta-1 in their (NaH)-forms. V.h.r. method of analysis. Table 4. Average activation energies of desorption (in kJ mol-') of ethanol by the v.h.r. and s.h.r. methods of analysis ZSM-5 ZSM- 1 1 Theta- 1 Ed (v.h.r.) 89 88 72 Ed (s.h.r.) 85 84 96 It would seem, therefore, that these Ed energies do represent desorption energies and not activation energies for the dehydration reaction.These conclusions should be confirmed by analysis of the effluent gas from the t.g. system but, unfortunately, we do not have a mass-spectrometer attachment on the back-end of the system and could not, therefore, analyse the effluent gases. Although the first seven ethanol molecules to be desorbed from ZSM-5 require slightly higher activation energies of desorption compared with the desorption of these molecules from ZSM- 1 1, there is no obvious reason for these differences observed in the desorption profiles shown in fig. 1 and 2 and table 3 for these two samples in this sorption range. The desorption of the first two ethanol molecules from Theta- 1 involves quite a distinct maximum in Ed at 5 molecules per unit cell which can account for the sharp low temperature peak in fig.3. The smaller E d value for Theta-1 in table 4 and the much lower Ed values in fig. 4 at120 - 100 L 8 2 y" 80 60 Chen Li-feng, T. Wacker and L. V. C. Rees 39 i 0 2 1: 6 8 10 12 14 molecules per unit cell Fig. 6. Activation energy of desorption of ethanol as a function of coverage. 0, ZSM-5; 0, ZSM-11; 0, Theta-1 in their (NaH)-forms. S.h.r. method of analysis. higher coverages compared with the corresponding values for ZSM-5 and ZSM-11 are consistent with the higher Si/A1 ratio of Theta-1 and are comparable to the differential molar heats, Q, found for ethanol in silicalite-1.5 Analysis of T.P.D.Profiles by the S.H.R. Method A heating rate of 6 K min-l was chosen. The variations of Ed and -AS$ with coverage derived from the s.h.r. method of analysis are presented in fig. 6 and 7, respectively. The average activation energies, Ed, for the three zeolites are given in table 4. The Ed values for all three zeolites at low coverages are quite large at ca. 110 kJ mol-1 for ZSM-5 and ZSM-11 and ca. 130 kJ mo1-l for Theta-1. These low-coverage E d values are similar to those in fig. 4. However, the s.h.r. method of analysis produces a h e a r decrease in Ed with increasing coverage to a final value of ca. 60 kJ mo1-l at saturation coverage for all three zeolites. There is no obvious strong interaction of ethanol with the limited number of sites of high electric field in these samples.The larger Ed value for Theta-1 in table 4 than that found for ZSM-5 and ZSM-11 is the opposite of the behaviour found using the v.h.r. method of analysis in table 4. This high E d value for Theta-1 is not consistent with the higher Si/Al ratio of this sample. The high E d value in Theta-1 could arise because of the one-dimensional channel network and the Naf ions acting as barriers in these channels. However, this explanation introduces the concept of a diffusion-controlled desorption process, which is unlikely since the Ed values are much greater than the activation energies of diffusion normally found in 10- ring channels. The average activation energy of desorption, Ed, was calculated using the s.h.r.method of analysis on each of the 10 t.p.d. profiles needed for the v.h.r. method of analysis. Each profile produced very consistent E d values which varied by < f 2 kJ mol-' for ZSM-5 and ZSM-11 and < f 3 kJ mol-' for Theta-1 from the Ed values listed in40 T.P.D. of Ethanol from Zeolites 160 - k - b 120 h \ t w I a 80 Table 5. Comparison of Ed values (kJ mol-') for ZSM-5 and silicalite adsorbate adsorbent n-hexane n-octane p-xylene benzene ethanol ZSM-5 85 92" 85b 8gb 85 silicalite ca. 73" ca. 86" ca. 77" ca. 70d ca. 53" (Si/A1 = 15) a Ref. (1). Ref. (2). " Ref. (3). Ref. (6). table 4 for the 6Kmin-l profile. Thus the s.h.r. method of analysis seems to give consistent Ed values reasonably independent of the heating rate employed at least over a heating range of 2-20 K min-'.The average desorption energies of ethanol from ZSM-5 and silicalite are compared in table 5. The large difference in Ed between these two samples is the result of the additional electrostatic energy introduced by the sites with high electric fields which exist in the ZSM-5 sample. The value of Ed of ca. 53 kJ mol-' for silicalite is.in very good agreement with the differential molar heat of sorption of ethanol in silicalite5 of ca. 57 kJ mo1-l. The difference in & for the non-polar hydrocarbon molecules listed in table 5 for these two samples is much smaller than that for ethanol, demonstrating that the high electric fields in ZSM-5 have a smaller effect on these energies when the desorbing molecule has no dipole moment.The desorption entropies, ASS, in fig. 4 and 7 for both ZSM-5 and ZSM-11 are ca. 80-90 J mol-' K-'. The v.h.r. and s.h.r. methods of analysis, therefore, give similar desorption entropies, but there are subtle differences in these entropies as a function of coverage between the two methods. The corresponding curves for Theta-1 in fig. 4 and 7 are quite different.Chen Li-feng, T. Wacker and L. V. C. Rees 41 1.0 k 3 .? 0.5 0 0 a 300 350 400 450 500 550 600 650 T / K Fig. 8. T.p.d. differential profile of ethanol in NaTheta- 1 . b = 10 K min-’ 1.0 h ? 0.5 3 0 0 0 a a O R 0 0 0 s o n I 0 0 0 300 350 400 450 500 550 600 650 T/K Fig. 9. T.p.d. differential profile of ethanol in HTheta-I. /? = 10 K min-’. T.P. D. of Ethanol from Diflerent Cationic Forms of High-silica Zeolites The t.p.d.differential profiles of ethanol from the sodium and hydrogen forms of Theta- 1 (Si/Al = 32) and the sodium forms of ZSM-5 and ZSM-11 are shown in fig. 8-1 1, respectively. These are the same zeolites which were used in the first part of this study in their mixed (NaH)-forms. The peak temperatures and widths and the maximum rates of desorption of ethanol from these samples are listed in table 6 . The differential dm’/dTcurves in fig. 8, 3 and 9 for ethanol desorbing from the Na-, (NaH)- and H-forms of Theta- 1, respectively, are very similar iq nature. Table 6 shows that the Na- and (NaH)-forms have similar temperatures for peaks I and 11, with the H- form having slightly lower peak temperatures. The maximum rate of desorption from the H-form is greater than the rate of the other two cationic forms. These results suggest that the desorption of ethanol from Theta- 1 involves larger interaction energies when Na ions are present in the one-dimensional channel system.42 T.P.D.of Ethanol from Zeolites Table 6. Peak temperature, peak width and maximum rate of ethanol desorption (,9 = 10 K min-') peak temperature/K ~~ peak I peak I1 larger peak max. desorption rate/ lo-" mg s-' NaTheta- 1 (NaH)Theta- 1 HTheta- I NaZSM-5 (NaH)ZSM-5 NaZSM- 1 1 (NaH)ZSM- 1 1 320 510 320 510 315 500 320 475 diffuse 465 320 475 320 470 I I I I1 I1 I1 I1 1.685 1.809 2.235 2.279 4.13 2.828 3.774 peak width/K peak I peak I1 300-380 465-550 300-380 480-550 _ _ _ _ ~ . . - 300-375 475-525 300400 435-500 diffuse 425490 300-380 430-490 300400 435-540 Table 7.Ethanol saturation capacities and Ed values obtained by the s.h.r. method @? = 10 K min-') Ed amount adsorbed /molecules per unit cell /kJ mol-' NaTheta- 1 (NaH)Theta- I HTheta- 1 NaZSM-5 (NaH)ZSM- 5 NaZSM- I 1 (NaH)ZSM- 1 1 5.98 99 6.25 96 6.50 89 13.21 98 13.60 88 13.93 94 14.27 84 1.0 .$ 0.5 0 a 0 a 0 0 O k 0 U 0 0 a 0 0 a 300 350 400 450 500 550 600 650 TIK Fig. 10. T.p.d. differential profile of ethanol in NaZSM-5. /I = 10 K min-'. The differential curves in fig. 1 and 10 show the effect on the ethanol desorption of additional Na+ ions in the ZSM-5 channel system. The low-temperature peak I is now quite distinct in the pure Na ZSM-5, while peak I1 occurs at a higher temperature with a much broader peak width than that found with the (NaH)-form.Similar behaviour forChen Li-feng, T. Wacker and L. V. C . Rees 1.0 0 8 z OS5 0 300 350 400 450 500 5% 600 650 TIK Fig. 11. T.p.d. differential profile of ethanol in NaZSM-I 1. p = 10 K min-'. 140 120 - L 8 2 100 4" \ 80 60 43 0 1 2 3 4 5 6 7 molecules per unit cell Fig. 12. Activation energy of desorption of ethanol as a function of coverage. 0, (NaH)Theta-1 ; 0, HTheta-1. S.h.r. method of analysis. 0, NaTheta- 1 ; peak I1 is found with the Na- and (NaH)-forms of ZSM-11. Finally, the rate of desorption of ethanol is lower in the pure Na-forms of these two zeolites. The amount of ethanol adsorbed at saturation and the average activation energy of desorption, Ed, calculated using the s.h.r. method of analysis of the profiles in fig. 8-1 1 are listed in table 7.In all of these samples increasing concentrations of Na ions decrease the saturation sorption capacity. The decrease is not very large and may only represent the additional volume of the Na+ ions over that of the replaced H+ ions.44 T.P.D. of Ethanol from Zeolites 140 120 @ k 2 4 100 80 60 0 2 4 6 8 10 12 14 molecules per unit cell Fig. 13. Activation energy of desorption of ethanol as a function of coverage. 0, NaZSM-5; 0, (NaH)ZSM-5. S.h.r. method of analysis. 1 LO 120 @ \ 2 100 4 80 60 0 2 L 6 8 10 12 l h molecules per unit cell Fig. 14. Activation energy of desorption of ethanol as a function of coverage. 0, NaZSM-11; 0, (NaH)ZSM- 1 1. S.h.r. method of analysis.Chen Li-feng, T. Wacker and L. V . C . Rees 45 The average desorption energies also clearly demonstrate the increasing sorb- ate-sorbent interaction energies introduced by the Na+ ions in the channels of these three zeolites.In fig. 12 the activation energy of desorption of ethanol determined by the s.h.r. method of analysis for the three cationic forms of Theta-1 is presented as a function of coverage. The three cationic forms all show very similar behaviour, with the H-form demonstrating slightly smaller E d values than the other two forms throughout the coverage range. In fig. 13 and 14 the corresponding Ed us. coverage curves are given for the two cationic forms of ZSM-5 and ZSM-1 I , respectively. Both of these figures clearly demonstrate the higher interaction energies between ethanol molecules and the channel surfaces at low loadings when additional Na+ ions are present in the channels.The desorption energies, however, are very similar for the respective cationic forms of ZSM- 5 and ZSM-11. At saturation loadings the desorption energy for both zeolites and both cationic forms is ca. 60 kJ mo1-l and not dissimilar from the E d value given in table 5 for silicalite. However, this desorption energy in the case of silicalite occurs over a much wider coverage range.3 These comparable desorption energies must represent the energy of desorption of ethanol from a silica site far removed from sites with large electric fields. Conclusions The t.p.d. of a polar molecule such as ethanol on low Si/Al ratio ZSM-5. ZSM-1 I and Theta-1 and different Na-, (NaH)- and H-forms of these zeolites shows that as the Na cation content of the zeolite increases the amount of ethanol adsorbed at saturation decreases and the activation energy for desorption increases.Both the saturation capacity and desorption energy demonstrate a subtle interplay between the electrostatic interaction of the dipole moment of the polar ethanol molecule and the electric field in the zeolite channels around an Al centre and the steric inhibitions created by the cations associated with such A1 centres. The variation in Ed with coverage for ethanol desorbing from ZSM-5 is very similar in form to the variation of the differential molar heat of adsorption with coverage found with silicalite. However, a difference of ca. 30 kJ mol-1 exists at all coverages between these two curves. These results are consistent with Ed being a measure of the desorption energies for these various systems and there is no clear evidence of any catalytic activity in the H- forms of these samples at the lower desorption temperatures. However, the broader high-temperature peaks in the Na-ZSM-5 and Na-ZSM-11 in fig. 10 and 11, respectively, compared with the corresponding narrower, somewhat lower temperature peaks in fig. 1 and 2 could be due to catalysed reaction increasing the ease of desorption of ethanol molecules from the H-forms of these zeolite samples. The authors are indebted to Dr R. E. Richards and Dr E. A. Dima for the development of the computer programs used in this study. We also wish to thank British Petroleum Research Centre, Sunbury-on-Thames for the three zeolite samples and Mr A. M. McAleer for help in the characterisation of these samples. References 1 L. F. Chen and L. V. C. Rees, Zeolites, 1988, 8, 310. 2 L. F. Chen and L. V. C. Rees, Zeofites, in press. 3 R. E. Richards and L. V. C. Rees, Zeolites, 1986, 6, 17. 4 Y. Tokoro, M. Misono, T. Uchijima and Y. Yoneda, BUN. Chem. Soc. Jpn, 1978. 51, 85. 5 H. Thamm, J . Chem. Soc., Faraday Trans. I , 1989, 85, 1. 6 H. Thamm, J . Phys. Chem., 1987, 91, 8. Paper 8/00230D; Received 20th January, 1988