7923
J . Phys. Chem. 1992,96,7923-7928
Sorption Propettles of Catlon-Exchanged &Zeolites K. S. N. Reddy, M. J. Eapen, H. S. Soni, and V. P. Shiralkar* Catalysis Division, National Chemical Laboratory, Pune 41 1 008, India (Received: September 4, 1991)
0-Zeolites cation exchanged with H+, Mg2+,Ca2+,Sr2+,and La3+were characterized by equilibrium sorption capacities of different probe molecules like water, n-hexane, cyclohexane, and benzene at 298 K and PIPo= 0.5. Specific surface areas (BET), void volumes, and sorption affinities were obtained by low-temperature (78 K) nitrogen sorption. Ammonia sorption isotherms in these cation-exchanged @-zeoliteshave been measured in the temperature range 303-453 K up to 400 Torr. The amount of ammonia retained irreversibly (on account of chemisorption) during desorption measurements was found to be correlated with intrinsic acidity possessed by the cation-exchanged zeolites. The analysis of these isotherm data in terms of different isotherm equation revealed satisfactory representation by Dubinin, Langmuir, and Sips equations. However both BET and KobleComgan equations failed to represent ammonia sorption data in these zeolites. Sorption models describing localizedlmobile sorption withlwithout interaction also could not represent the sorption data satisfactorily. All the cation-exchanged zeolites were compared in terms of chemical potential of ammonia sorption. Isosteric heats of ammonia sorption calculated from the linear sorption isosteres revealed the heterogeneity of sorbent surface of the cation-exchanged zeolites as compared to the parent zeolite NaHbeta. Electronegativity, polarizability, and size of the exchanged cation were found to influence the sorption energetics.
TABLE I: Unit Cell Composition (on Anhydrous Bmb) of Large pore zeolites (12 MR) have always been the focus of commercial catalytic applications' on account of their most open framework structure. @-Zeoliteis no exception to this, when its excellent performance in the production of cumene2 during isopropylation of benzene is taken into consideration. 8-Zeolite usually crystallizes3either into a single or into a mixture of two or three polymorphs. In spite of its lower thermal stability compared to mordenite, faujasites, and pentad, zeolites, on account of typical framework structure and the acidic sites associated with it, are looked upon as a potential industrial catalyst in hydrocarbon conversion reactions. Usually the acidic properties of aluminosilicate zeolites are probed into by basic molecules like ammonia, n-butylamine (n-BA), etc. The former, being comparatively smaller and more polar, is probably considered to be the best suited for this purpose. With the exception of some scanty reports? literature does not reveal much work on the sorption properties of &zeolite. Prompted by this, we studied sorption kinetics with water, n-hexane, cyclohexane, and benzene (298 K), low temperature (78 K) nitrogen sorption, and sorption isotherms of ammonia in different cationic forms of Bmlite in the temperature range of 303453 K. The results on the sorption isotherm analysis and thermodynamic parameters derived therefrom are reported in the present communication.
Experimental Section Materials. A synthetic sample of 8-zeolite in as-synthesized form was obtained from M/s Konteka (USA), with the product Si02/A1203ratio to be around 26.1. The as-synthesized sample was converted into sodium form by calcining in air at 803 K. The sodium occupancy as a framework charge balancing cation was only 30% of the total aluminum in the sample. These sodium ? ' , cations w m almost totally exchanged by NH,', Mg2+,Ca2+,S and La3+ions by contacting the zeolite powder with 15% aqueous solutions (15 mL/g of mlite) of the chloride salts of the respective cationic species for 4 h at 368 K. Cation exchange treatments were repeated until the residual sodium in the sample was below 200 ppm. After the cation exchange treatments the samples were thoroughly washed with deionized water until the wash filtrate was free from chloride ions. The samples were dried at 393 K for 4 4 h. Prior to the determination of a chemical composition, the samples were moisture-equilibriated at 298 K by keeping them for 24 h over the saturated solution of ammonium chloride. Table I lists the unit cell compositions of all the samples determined by wet chemical methods including atomic absorption spectrometry To whom all correspondence should be addressed.
Cation-Excbnged &Zeolites
zeolite NaHbeta Hbeta MgHbeta CaHbeta SrHbeta
unit cell composition
no. of unit cells/e x 10-20
LaHbeta
(Hitachi, Z-8000), flame photometry (Toshniwal), and gravimetric techniques. Unit cell compositions listed in Table I reveal that during cation exchange part of the protons in addition to most of sodium present in the parent zeolite has also been exchanged. Although the total percent cation exchange is around 9996, the percent cation exchange by alkaline earth and lanthanum varies from 45% to 75%. The cylinder ammonia (purity >99%) was further purified by passing it over freshly ignited calcium oxide, potassium hydroxide pellets, and activated molecular sieves. Metbocls An all-glass BET apparatus5 was used for volumetric sorption measurements. A temperature-programmed furnace connected to a Aplab temperature controller was used for controlling the temperature between 333 and 453 K. The isotherm at 303 K was measured with a liquid paraffin thermostat. The temperature accuracy throughout the sorption measurements was within f l K. The zeolite sample ( e 2 5 0 mg) was degassed at 723 K. The temperature of the sample was raised with the heating rate of 2 deg/min with simultaneous evacuation at 1od Torr. The sample was degassed at 723 K for 10 h and the temperature was then lowered down to the isotherm temperature at which it was then allowed to stabilize and maintain at least for 2 h before the commencement of the measurement. The sorption isotherms were measured up to 400 Torr within 303453 K with 30 K intervals. Isotherm measurements were initially carried out at 453 K and then subsequently at lower temperature. In order to check the revmibility of the sorption, desorption measurements were carried out. After each isotherm the sample was evacuated at 1od Torr for 10 h. X-ray diffractograms were recorded for each sample before and after the sorption measurements in order to check the structural stability. Prior to the measurement of sorption uptake, the sample was activated under vacuum ( lod Torr) as mentioned above. The sample was cooled down to 298 K and was then contacted with vapors of liquid sorbates at a relative pressure of 0.5. Equilibrium sorption was measured up to 2 h. A low temperature nitrogen sorption was carried out using Omnisorp, 100 CX.
0022-3654/92/2096-7923$03.00/00 1992 American Chemical Society
Reddy et al.
7924 The Journal of Physical Chemistry, Vol. 96, No. 20, 1992 TABLE II: Sorption Properties of Cation-Exchanged &Zeolites: PIP. = 0.5: Temoenture = 298 K
NaHbcta
34.85 (8.99) Hbeta 41.79 (11.17) MgHbcta 41.46 (10.73) CaHbeta 38.88 (9.95) SrHbeta 42.09 (10.62) LaHbeta 42.14 (10.61) a
8.20 (2.12) 8.66 (2.32) 8.86 (2.29) 7.83
8.58 (2.21) 9.18 (2.45) 9.21 (2.38) 8.15 (2.00) (2.08) 9.80 8.27 (2.09) (2.47) 8.31 9.12 (2.09) (2.30)
11.18 (2.88) 11.42 (3.05) 11.85 (3.07) 11.05 (2.83) 11.85 (2.99) 11.66 (2.94)
759
0.311
708
0.283
684
0.277
690
0.277
U 3
713
0.290
ul W -1 3
700
0.283
\
w V
B
Figures in parentheses indicate sorption capacities in mmol/g.
Results and Discussion I. Equilibrium Sorption hta, Specific Surface Areas, and Sarptioa AflIiaities for Nitrogen, Table I1 summarizes equilibrium sorption capacities obtained from sorption measurements at PIPo = 0.5 and 298 K of different sorbates in cation-exchanged 8zeolites. In general it is observed that the sorption capacities of most of the sorbates increased marginally upon cation exchange of the parent NaHbeta. A replacement of Na+ by smaller protons and a decrease in cation density on account of exchange of monovalent sodium by multivalent cations are expected to bring about an increase in the void volumes of the lattices of the cation-exchanged zeolites. With the exception of CaHbeta the water sorption increases (from 35 to 42 molecules/u.c.) almost by 20%. The sorption capacities of n-hexane, cyclohexane, and benzene were fairly constant around 8.5 f 0.3,9.5 f 0.3, and 11.6 f 0.2 molecules/u.c., respectively. It is really surprising that among the alkaline earth exchanged samples CaHbeta exhibited comparatively lower sorption values. As far as the water sorption is con& among the alkaline earth exchanged samples the degree of &a2+ exchange is the highest and consequently the concentration of protons is the lowest. Unit cell composition of CaHbeta shows some dealuminationleading to lower void volume and hence lower sorption values for n-hexane, benzene, and cyclohexane. Although these sorption capacities increased marginally, no drastic changes are observed in sorption capacities in the lattices of the cationexchanged zeolites. Specific surface area (BET) listed in Table 11, obtained from low temperature (78 K) nitrogen sorption, decreases from 759 for NaHbeta to 700 f 10 m2/g for cation-exchanged zeolites. Void volumes obtained by applying the Dubinin equation5 also confirm the same trend of decrease from 0.3 11 for NaHbeta to 0.277 mL/g for cation-exchanged zeolite%.The chemical affhities calculated from nitrogen sorption data decreased from 3000 to 100 J/mol with the increase in the amount of nitrogen sorbed. NaHbeta showed highest sorption selectivity over the entire coverage. The drop in chemical affinity was found to be comparatively sharper in CaHbeta and slower in NaHbeta. In the lower coverage region the sorption selectivity sequence follows the order NaHbeta > SrHbeta > LaHbeta > CaHbeta > MgHbeta > Hbeta. In the mid coverage region selectivity sequence changes to exhibit the lowest sorption selectivity by CaHbeta. In the higher coverage region the sequence becomes complicated. The a-bonding of almost nonpolar cylindrical nitrogen molecule seems to be exerting weak repulsive interaction with the extra-framework cations decreasing in the low coverage region and hence the sequence in nitrogen sorption selectivity follows the increasing trend in Sanderson's electronegativities.6 In the mid or high coverage region perhaps the total void volume and interaction between sorbed nitrogen molecules may be dominating over the electronegativity effects and hence the sorption selectivity sequence changes. LI. Ammonia Sorption Isotherms. Families of isotherms of ammonia sorption in typical cation exchanged beta zeolites are
P(Torr)
Figure 1. Isotherms for ammonia sorption in (A) Hbeta,(B) LaHbeta, (C) NaHbeta, and (D) MgHbeta at (1) 303 K, (2) 333 K, (3) 363 K, (4) 393 K, (5) 423 K, and (6) 453 K.
depicted in Figure 1. The shape of the isotherms approximates to the type I (Langmuir type) a d i to Kiselev's classification? Almost 75% of the total sorption takes place over a very narrow range of relative pressure. The equilibrium sorption uptake at 400 Torr and especially at lower temperature follows in the sequence of Hbeta > LaHbeta > MgHbeta > NaHbeta > CaHbeta > SrHbeta. The factors like Sanderson's electronegativity and polarizability ( e / r ) of the extra-framework cationic species are expected to influence the sorption capacities and energetics of a polar basic molecule like ammonia. Being a highly polar basic molecule with an ease of a donation of a lone electron pair, ammonia interacts strongly with extra-framework cations including protons of increasing electronegativity. The increasing polarizability (e/r) generates more protons on the hydrolytic dissociation of occluded water molecules on dehydrating zeolite up to 773 K. The observed sequence for ammonia capacity is then in accordance with the combined effect of Sanderson's electronegativities and polarizabilities of the cation. At lower pressures ( Y 100 Torr) and at higher temperatures the sequence in ammonia sorption capacities become rather complicated on account of varying contribution of sorption irreversibility as discussed later. Cation density in zeolites usually decreases with increase in cationic charge. If these extra-framework cations and protons, produced upon dissociation of water molecules associated with them, act as sorption centers, then the equilibrium amount sorbed of polar molecule like ammonia is expected to decrease with the decrease in cation density in the zeolite framework. Accordingly, Figure 1 confirms this general expectation at higher pressure (400Torr); however, at lower pressure (100 Torr) the amount sorbed in zeolites with decreasing cation density follows the sequence NaHbeta < MgHbeta < LaHbeta. This shows that at lower pressure (lower coverage) sorption capacity secms to be dependent on the cationic charge rather than on the cation density. Similar observation was reported* earlier for the carbon dioxide sorption in cation-exchanged Y zeolites. In the case of cationic species with the same charge (Mg2+,Ca2+,and S$+)the sorption uptake seems to increase with the decrease in ionic radius. Even though only 30% sodium occupancy is there in the parent zeolite NaHbcta, upon replacement of those sodium cations by protons (via deam-
Sorption Properties of Cation-Exchanged fl-Zeolites
The Journal of Physical Chemistry, Vol. 96, No. 20, 1992 7925 1.2
0.4
T(K)
Figure 2. Amount of ammonia retained irreversably as a function of temperature in (1) Hbeta, (2) LaHbeta, (3)MgHbeta, (4) CaHbcta, ( 5 ) SrHbcta, and (6)NaHbcta.
moniation of ammonium form) there is at least as much as a one and half times increase in the uptake of ammonia molecules. This clearly shows that protons act as stronger sorption centers than other cationic species such as Na+, Mg2+,Ca2+,Sr2+,and La3+. III. IrrevemibMty of Ammonia Sorption aod Acidity. During the desorption experiments under vacuum it was found that all the cation-exchanged fl-zeolite samples retained a certain amount of ammonia. This clearly indicates irreversible sorption of ammonia. Irreversibility in ammonia sorption has also been reported earlier in MnY? RUM, CrM, and FeM'oJ' zeolites at 303 K, supporting our observation. The irreversibility of ammonia sorption in zeolites is attributedg to (a) formation of ammonia complexes with intrmlitic cations and to (b) chemisorption. In the case of the complex formation?J0color changes were indicated during outgassing and sorption of ammonia in transition-metal zeolites. In the present studies irreversibly sorbed ammonia was found to be completely desorbed on heating the zeolite at 733 K. As the temperature of desorption was increased, the number of arminonia molecules per cation decreased from 1.5 to nearly 1.O in the temperature range of 453-623 K. The approaching of ammonia to cation ratio to unity and taking into consideration a possibility of a formation of a chemisorptive NH4+species by the interaction of ammonia with a proton suggests the chemisorption to be a possible cause of a sorption irreversibility. To arrive at a definite conclusion regatdingthe cause of irrcvmibility, a detailed and a separate investigation needs to be done. Figure 2 shows the variation in the chemisorbed ammonia with temperature in d i f f e r t cation-exchanged &zeolites. The number of ammonia molecules held on account of chemisorption especially at higher temperatures (>453 K) may be taken as a measure of acidity. Strictly considered physically sorbed ammonia molecules must be desorkd by lowering pressure to OTorr without inmasing the temperature. At temperatures lower than 453 K, some ammonia molecules (not necessarily physically sorbed) are sorbed less strongly on extra-framework cations other than protons and on surface silanol groups and they approximately correspond to a and B peaks on ammonia TPD.12 As the deaorption temperature increases the NH4+/H+ratio approaches unity corresponding to the y peak on ammonia TPD. At 733 K all the ammonia molecules were found to be desorbed. The amount of chemisorbed ammonia is highest for Hbeta followed by LaHbeta, MgHbeta, CaHbcta, SrHbeta, and NaHbeta, and this clearly is regarded as the sequence of total acidity in the zeolites. It is seen from F i 2 that the dccrasc in the amount of chemisorbed ammonia with the increasein the sorption temperature is more sharp in the sample retaining more ammonia. The total acidity of the exchanged zeolites, with the exception of MgHbeta, follows the sequence of the e / r ratio and Sanderson's electronegativity of the cation.
I 0
1 0.6
12
18
24
3.0
(LO0 po/P12
Figure 3. Dubinin plots for ammonia sorption in (A) LaHbeta and (B) MgHbeta at (1) 303 K,(2) 333 K,(3) 363 K,(4)393 K,(5) 423 K,and (6)453 K. W is the amount of sorbed molecules/u.c.
IV. Application of Isotherm EqurtToa. (a) D m W Equrtioa Polanyi potential theory when modified by Dubinin and Radu~hkevich'~ takes the form log (w)= log (W,) - B/2.303fl2[TlOg (P0/P)l2
(1)
where Wis the amount sorbed at equilibrium pressure P, W,is the total sorption capacity, B is a constant independent of temperature and is characteristic of a sorbent pore structure, and 6 is the affinity coefficient. Dubinin plots were constructed by plotting (log against log Wfor all the cation+xchange zeoIites in the temperaturerange of 3 0 3 4 5 3 K. In all the samples excellent linear plots were obtained and typical Dubinin plots are shown in Figure 3. The saturation capacities obtained from the intercept of their linear plots were slightly higher than those " d e d at 400 Torr experimentally. This seems to be reasonable because the values obtained from the Dubinin plots are extrapolated over higher pressures. The magnitude of the affinity coefficient fl is inversely proportional to the square root of the factor B/f12obtained from the slope. An increase or a decrease in B is usually looked upon as an increase or a dccmsc in sorption affinity of a sorbate molecules toward the sorbent surface. From the linear plots in Figure 3 it is seen that the affinity coefficient /3 dccreasts with the increase in temperature for all the zadites except for MgHbeta and Hbeta. Although there d a s not seem to be any spa% trend in the value of fl with cationic charge or with e / r ratio of the extra-framework cationic species;the affinity d c i e n t for NaHbeta is lower than that for other cation-exchanged zeolites. In general the excellent linearity of these plots suggests that the ammonia sorption in cation-exchanged &zaolitcs could be rapresented satisfactorily by Polanyi's p o t m t i a l y of volume filling modified by Dubinin and Radushkevich. C02 and NH314sorption data in cationexchanged Y zeolites were also satisfactorily reprmented by Dubinin equation. (b) L "&Eqprtiaa Langmuir sorption isotherm equation has been derived on the assumption of localized monolayer sorption on the sorption centers of equal energy and 1:l correspondence between the sorption centers and sorbate molecules. We tried
Reddy et al.
1926 The Journal of Physical Chemistry, Vol. 96, No. 20, 1992
~~~
200
400
200
400
P(Torr)
Figure 4. Langmuir plots for ammonia sorption in (A) Hbcta,(B) LaHbcta, (C) MgHbeta, and (D)NaHbcta at (1) 453 K, (2) 423 K, (3) 393 K, (4) 363 K, (5) 333 K, and (6) 303 K. Vis the amount sorbed moleculcs/u.c.
to analyze ammonia sorption data in cation-exchanged 8-zeolites and excellent linear plots (typical of which are shown in Figure 4) were obtained for all the cation-exchanged zeolites over the entire temperature range of 303-453 K. The excellent linearity of these plots from Figure 4, undoubtedly, indicates the validity of the Langmuir equation to describe ammonia sorption in cation-exchanged @-zeolites. It seems, therefore, that ammonia sorption in these zeolites is localized with some element of chanisorption indicated by i " i b l y retained ammonia discussed earlier. Similarly n-butylamine sorption in EU-l15 and Fe3+ exchanged Y zeolites16 did yield linear plots when the Langmuir equation was applied. On the contrary, the Langmuir equation failed to yield linear plots for C o t and NH314sorption in a series of Y zeolites exchanged with La3+, Ca2+,and H+.Monolayer capacity (V,) obtained from the reciprocal of the slope of these linear plots was rather higher than those recorded experimentally at 400 Torr. This seems to be reasonable that the monolayer capacity (V,) may not necessarily match the sorption capacity at 400 Torr. The reciprocal of the intercept made by these plots on the ordinate does often represent the magnitude of a constant "C" in the Langmuir equation. The constant "C"may qualitatively be looked upon, in one way or other, related to a heat of sorption during monolayer formation. The higher the value of 'C" (the lower the reciprocal), the stronger is the interaction between the sorption center and sorbate molecule. (c) BET Equation. The BET isotherm equation is based on the assumption of multilayer formation with the special and much higher value of heat of sorption for the formation of the f i t layer than that for the sucassive layers. The analysis of the ammonia sorption data in terms of the BET equation yielded linear plots, and typical examples are shown in Figure 5 . The plots show considerable intercept on the ordinate indicating a not much higher value of a constant 'C" in the BET equation. In spite of excellent linearity of these BET plots; the values of monolayer capacities obtained from the slopes and intercept, with the exception of NaHbeta and Hbeta, deviated considerably from the experimental value8 in magnitude and order. This shows that a BET approach, based on the multilayer formation, has only a limited applicability in the present studies of ammonia sorption in cation-exchanged 8-zeolites. However n-BA sorption data in EU-I zeolites15were
25
75
50 P/P,
lo3
Figure 5. BET plots for ammonia sorption in (A) NaHbeta and (e) Hbcta at (1) 453 K, (2) 423 K, (3) 393 K, (4) 363 K, (5) 333 K, and (6) 303 K. Vis the amount sorbed moleculcs/u.c.
satisfactorily represented by linear BET plots passing through the origin and giving monolayer capacities in close agreement with those obtained experimentally. (a) Sips Quatiom The Sips equation'' based on the localized sorption with sorbate-sorbate interaction usually takes care of any deviation from the Langmuir approach. If the sorption is assumed to be a chemical reaction between the sorption centers and the sorbate molecules, the Langmuir equation results with 1:1 correspondence between them, and if some tolerance is made for the complicating factors, the Sips equation results. Mathematically the Sips equation is expressed as1' log [e(i - e)] = log A c log P (2)
+
where A and c are constants and P is an equilibrium pressure at coverage 8. 0 was calculated by using the monolayer (saturation) capacities obtained by Langmuir equation. In order to check the applicability of the Sips equation, the plots of log [e/(l - e)] against log P for the zeolites were constructed. Figure 6 shows the typical Sips plots for ammonia sorption in cation-exchanged 8-zeolites at 303 K. The linearity of these plots shows the a p plicability of the Sips equation to the sorption of ammonia in cation-exchanged fl-mlites. Similarly C o t and NH314sorption data in a series of La3+,Ca2+,and H+ ion exchanged Y zeolites showed the limited applicability of the Sips equation. However, the Sips equation failed to represent n-BA sorption data in EU-I15 zeolites. As the constant "c" evaluated from linear plots approaches unity, eq 2 reduces to the Langmuir isotherm and the deviation of "c" from unit may be taken as a measure of a deviation from the Langmuir approach. In the pnsent case 'c" always has a value near to unity (1 f 0.1). irmpective of the nature of the exchanged cation and temperature of the sorption. Therefore it seem likely that no complicating factors are involved in ammonia sorption to cause deviation from the Langmuir approach. Although the value of A was always lower for the parent zeolite NaHbeta than that for the cation exchanged forms, it did not show any specific trend with cationic species and with the isotherm temperature. (e) The Kobldorrigan Equation. The KobltCorrigan equation1*was based on the exact solution for dissociative sorption
The Journal of Physical Chemistry, Vol. 96, No. 20, 1992 7927
Sorption Properties of Cation-Exchanged &Zeolites
Log P
Figure 6. Ammonia sorption plots for the Sips equation at 453 K in (1) MgHbeta, (2) SrHbeta, (3) CaHbeta, (4)Hbeta, (5) LaHbeta, and (6)
NaHbeta.
of sorbate molecules on two active centers. Although it rcpme.nted satisfactorily the COz8sorption data in La3+-, Ca*+-, and H+exchanged Y zeolites, the same equation failed to yield linear plots in the present studies of ammonia sorption in cation-exchanged &zeolites. When the data were analyzed in terms of the Kobldorrigan equation, instead of a linear plot, two intersecting linear portions with different slopcs were obtained. It is clear, therefore, that the ammonia sorption in fl-zeolites does not allow the assumption of dissociative sorption of sorbate molecules on two sorption centers, during the derivation of the KobltCorrigan equation. ( f ) Applicrtioa of Statistical Models of Langmuir rad Vdmer. The statistical model equations of Langmuir and V ~ l m e derived r~~ initially for ideal systems are often appliedgJ"to the real systems to yield information on the extent of deviation occurring in such systems due to surface heterogeneity, multilayer formation, and mutual interactions between sorbed molecules. In the present studies, plots of KL, In KL, K,, and In Kv as a function of coverage did not yield linear plots, and hence no meaningful deductions could be drawn from these plots. v. cbemicrrlAffinity md tbe selectivity of the sorbed Phrses. When a gas is transferred reversibly and isothermally from the gas phase at a standard pressure Po into an infinite amount of sorbent-sorbate mixture over which an equilibrium pressure is P, a decrease in chemical potential takes place. Neglecting the nonideality of the sorbate, the chemical affinity may be expressedS.l4+1920as Ap = RT In ( P / P o )
(3)
The value of Ap may be taken as the quantitative measure of the chemical affinity of the sorbate for the sorbent. The plots of -Ap against the amount sorbed also serve as useful criteria for the comparison of sorption affinities of a probe molecule in the lattices of various cation-exchanged zeolites. The plots of -Ap against the amount sorbed are shown in Figure 7. In general it is seen from Figure 7 that a decrease in chemical potential is comparatively sharper at higher temperature as compared to that at lower temperature. In the lower coverage region of the entire temperature range, NaHbeta showed the lowest chemical potential while both LaHbeta and Hbeta showed higher chemical potentials over the entire coverage. At the higher coverage the sequence
NH~,MOLECULE / U~C
Figure 7. Chemical affinity for ammonia sorption at (A) 453 K and (B) 303 K in (0) Hbeta, (0) LaHbeta, (A)MgHbeta, (0)CaHbeta, (A)
SrHbcta, and (X) NaHbeta.
in chemical potential becomes complicated and NaHbeta shows a higher chemical potential than alkaline earth cation exchanged &zeolites. In the lower coverage region, an ammonia molecule probably interacts with acidic protons, whereas in the higher coverage region it interacts with extra-frameworkcations. Perhaps in the casc of CaHbcta and SrHbeta, Ca2+and S F impart more basic character rather than the acidic character compared to Na+ to the zeolite lattice. In the higher coverage region the chemical potentials for LaHbeta and Hbeta compete with each other at 303 K, and the latter exhibits higher chemical potential over the entire coverage at 453 K. These trends in chemical potential arc consistent with the amount of irreversibly retained ammonia and these results also indicate the measure of the total acidity of the zeolite catalyst. Another salient feature of Figure 7 is that the decrease in -Ap with the merage is slower in Hbeta and LaHbeta than that in NaHbeta and alkaline metal exchanged fl-ztolites. This means at any fmed coverage and temperature both LaHbeta and Hbeta show higher chemical affinity than NaHbeta toward the ammonia molecule. VI. Isosteric Heats (QJ of Ammoalp Sorptioa. Isosteric heats of ammonia sorption, Q,,,are computed from the shift of sorption equilibrium with temperature at constant sorbate loading in accordance with the Claussius-Clapeyron relationship as follows2'" -AH Qst = R [ ( T ~ T ~ ) / ( ~-' TI)] z In ( p 2 / P 1 ) (4) If Q,,is temperature independent, the plots of In P against 1 / T are expected to be linear within the experimental error. These isosteres were found to be almost linear in the present studies. Figure 8 shows the variation in Q,,(obtained from the slope of the linear isosteres) with amount of ammonia sorbed in cation exchanged &zeolites. Figure 8 shows that LaHbeta exhibits the highest Q,,value (41 kJ/mol) and lowest Q,, (20 kJ/mol) for SrHbeta at a coverage of 0.8 molecules/u.c. The Q,*value at a coverage of 0.8 molecule/u.c. follows a sequence SrHbeta C CaHbeta C NaHbeta C Hbeta C MgHbeta C LaHbeta. The
7928 The Journal of Physical Chemistry, Vol. 96, No. 20, 1992
1 30
-
20
'"t 0
1
1
2
3
4
5
6
7
N H 3 , MOLECULES/ U C
Figure 8. Isosteric heats (Q,,)for ammonia sorption in (0)Hbeta, (0) LaHbeta, (A) MgHbeta, (0)CaHbeta, (A) SrHbeta,and (X) NaHbeta.
lower Q,, value of SrHbeta and CaHbeta than NaHbeta in this case also implies reduced acidic character (in other words increased basic character) of the former zeolites than the latter one. These results indicate the combined influence of Sanderson's electronegativity and polarizability ( e l f )of the extra-framework cations exchanged on the sorption energetics. Martinez and Dumesic has also reported23similar influence on Sanderson's electronegativity of the added oxide to the silica support on the initial differential heat of adsorption of pyridine by microcalorimetric measurements. All the samples, except NaHbeta, show a wide variation in the Q,,value with amount of ammonia sorbed indicating increased surfae heterogeneity in cation-exchanged 8-zeolites as compared to NaHbeta, Similar surface heterogeneity was also reported22 by a series of La3+-, Ca2+-, and H+-exchanged Y zeolites as compared to NaY zeolites. The variation in Qatwith the amount sorbed in LaHbeta is from 41 kJ/mole at 0.8 molecule/u.c to 27 kJ/mol at 6.5 moleculcs/u.c. On the contrary, the variation in Qa in NaHbeta at identical sorption capacity is only 24-18 kJ/mol. Following a decrease in Qat in the low coverage region, most of the samples show almost a steady value with humps in the higher coverage region. The humps may be due to an experimental artifact. Alternatively the sorption in the low coverage region involves20 predominantly sorbatesorbent interactions, and a e r a g e indicates the presence of sorption duxcase in Qa with the m centers with decreasing strength. Mid coverage region involves both sorbate-sorbate and sorbatesorbent types of interactions, and depending upon their individual contribution toward sorption energetics, the humps in the Q,,curves are involved. A steady decrease or almost constant Q,, value in the higher coverage involves only sorbatesorbate interactions which are rather weak in magnitude. Equilibrium sorption capacities for water, n-hexane, cyclohexane, and benzene increased marginally in cation-exchanged 8-zeolites. Specific surface area (BET), however, decreased slightly in cation-exchanged mlites. NaHbeta exhibited highest sorption selectivity over the entire coverage for nitrogen sorption. Ammonia sorption isotherms in cation-exchanged &zeolites up to 400 Torr in the temperature range of 303-453 K were found to be of the Langmuir type. In the lower pressure region the sorption capacity was found to be influenced by cationic charge rather than cation density. In the case of cationic species with the same charge (Mg2+,CaZ+,and SrZ+)the sorption capacity seemed to increase with the decrease in cationic radius. Protons
Reddy et al. were found to be the stronger sorption centers for ammonia sorption than the other cationic species. The amount of irreversibly retained ammonia was found to be dependent on the total ammonia sorbed and on the intrinsic acidity associated with the zeolitic lattice. The sorption uptakes and energetics are found to be influenced by Sanderson's electronegativity and polarizability ( e / r ) of the extra-framework cationic species. The saturation capacities obtained from the linear Dubinin plots and monolayer capacities obtained from linear Langmuir plots were in close agreement with the sorption capacities recorded experimentally. Affinity coefficient 8 of the Dubinin equation was lower for NaHbeta than that for the cation-exchanged zeolites. Similarly constant "C" of the Langmuir equation was higher for the cation-exchanged zeolite8 than the parent NaHbeta zeolite. The BET equation showed limited applicability to ammonia sorption in these zeolites, while Sips' equation yielded linear plots. The KobleCorrigan equation and statistical models of Langmuir (KL)and Volmer (Kv) failed to represent ammonia sorption and failed to give useful information on the state of the sorbed phase. Hbeta and LaHbeta exhibited higher chemical potential than other cation-exchanged zeolites over the entire merage and temperature range. The drop in chemical potential with the amount sorbed was, however, sharper in NaHbeta and SrHbeta as compared to that in other zeolites. LaHbeta showed the highest Q,, value and SrHbeta showed the lowest value over the entire merage. The initial isosteric heats seem to be influenced by Sanderson's electronegativity and e / r ratio of the exchanging cation. Acknowledgment. We are grateful to Dr. P. Ratnasamy for his constant encouragement and helpful discussions. K.S.N.R. thanks UGC for the award of a fellowship. The work was partly funded by UNDP.
References and Notes (1) Venuto, P. B.; Landis, P. S. Adu. Caral. 1968, 18, 259. (2) Pradhan, A. R.; Rao, B. S.; Shiralkar, V. P. Proc. I d . Zeol. Conf.
'ZEOCAT-90w,Leipzig, GDR. Srud. Suf. Sci. Coral.;Ohlmann, G., Heifer, H., Fricke, R., Eds.; Elsevier Science: Amsterdam, 1991; Vol. 65, p 347. (3) Newsam, J. M.; Tracy, M. M. J.; Koetsier, W. T.; de Gruyer, C. B. Proc. R. Soc. hndon, Ser. A 1988,420, 375. (4) Joshi, P. N.; Kotasthane, A. N.; Shiralkar, V. P. Presented at Natl. Symp. Catal. Hyderabad, 1990. (5) Shiralkar, V. P.; Kulkami, S. B. Z . Phys. Chem. (Leiprig) 1983,265, 313. (6) Sanderson, R. T. J . Am. Chem. Soc. 1983,105, 2259. (7) Kisclev, A. V. Discuss Faraday Soc. 1%5,40,205. ( 8 ) Shiralkar, V. P.; Kulkami, S.B. Zeolites 1984, 4, 330.
(9) Coughlan, B.; McEntee, J. J. Proc. R. Irish Acad., Sect. B 1976, 76, 473. (10) Coughlan, B.; Mdann, W. A. J. Chem.Sm., Faraday Trans. 1 1979, 75, 1969. (11) Coughlan, B.; Larkin, P. M. Chem. Ind. 1976,275. (12) Topsoe, N.; Pederson, K.; Derouane, E. G. J . Caral. 1981, 70, 41. (1 3) Dubinin, M. M.; Radushkevich, L. V. Proc. Acad. Sci., USSR 1974, 55, 321. (14) Shiralkar,V. P.; Kulkarni, S. B. J . Colloid Interface Sci. 1985, 108, 1.
(15) Rao, G.N.; Joshi, P. N.; Kotasthane, A. N.; Shiralkar, V. P. J . Phys. Chem. 1990, 94. 8589. (16) Kulkarni, S.J.; Kulkarni, S. B. Indiun J. Chem. 1989, 28A, 6. (17) Sips, R. J . Chem. Phys. 1948,16,491. (18) Koble, R. A.; Corrigan, T. E. Ind. Eng. Chem. 1952,44, 383. (19) Barrer, R. M.; Coughlan, B. Moltcular Sieves. Soc. Chem. Ind. London 1968, 141, 233, 241. (20) Coughlan, B.; Kilmartin, S. J. Chem. Soc., Faraday Tram. I 1975, 71,1809, 1818. (21) Shiralkar, V. P.; Kulkarni, S.B. Zeolites 1985, 5, 37. (22) Shiralkar,V. P.; Kulkarni, S.B. J . Colloid Interface Sci. 1986,109, 115.
(23) Martinez, N. C.; Dumesic, J. A. J . Caral. 1991, 127, 706.
