Application of Porous Phosphate Heterostructure Materials for Gas

1260

Langmuir 2006, 22, 1260-1267

Application of Porous Phosphate Heterostructure Materials for Gas Separation G. Aguilar-Armenta,*,† M. E. Patin˜o-Iglesias,† J. Jime´nez-Jime´nez,‡ E. Rodrı´guez-Castello´n,‡ and A. Jime´nez-Lo´pez‡ Centro de InVestigacio´ n, Facultad de Ciencias Quı´micas, Beneme´ rita UniVersidad Auto´ noma de Puebla, 14 Sur y AV. San Claudio, Ciudad UniVersitaria, 72570 Puebla, Pue., Me´ xico, and Departamento de Quı´mica Inorga´ nica, Cristalografı´a y Mineralogı´a, Unidad Asociada al ICP-CSIC, Facultad de Ciencias, UniVersidad de Ma´ laga, 29071 Ma´ laga, Spain ReceiVed September 1, 2005. In Final Form: October 28, 2005 In this study, the adsorbents Cu+-3SiPPH723 and Cu2+-3SiPPH723 were prepared starting from a silica-expanded zirconium phosphate heterostructure, 3SiPPH(0.2), which was subjected to an ion exchange with Cu(I) and Cu(II). These materials were characterized using powder X-ray diffraction, X-ray photoelectron spectroscopy, ammonia thermally programmed desorption, hydrogen temperature-programmed reduction, and N2 adsorption (77 K). The equilibria and kinetics of adsorption of pure propylene (C3H6) and propane (C3H8) were studied using a conventional glass high-vacuum volumetric device, equipped with grease-free valves, in the temperature range of 273-393 K. The starting material, 3SiPPH(0.2), presented a high acidity and irreversible chemisorption of the olefin, which increases with temperature. Unlike the support, the irreversible adsorption of the olefin on the Cu+-3SiPPH723 and Cu2+3SiPPH723 samples decreases with increasing temperature and disappears at 393 K, showing a very high selectivity toward propylene. The C3H8 adsorption in all the samples was always reversible. On the basis of the results of this study, both Cu+-3SiPPH723 and Cu2+-3SiPPH723 samples can be efficiently applied in the separation of a C3H6/ C3H8 mixture at 393 K. Cu+-3SiPPH723 would have the highest efficiency, because its capacity for C3H6 adsorption was higher than that for the Cu2+-3SiPPH723 sample.

Introduction al.1

Ten years ago, Galarneau et reported a novel process of synthesis of a new family of porous materials called porous clay heterostructures (PCHs), where the surfactant molecules are templates of inorganic arrays in the clay interlayer space. In these materials, cationic surfactant molecules are placed in the interlayer space of a clay host by means of a cationic exchange process. For this purpose, clays, with a high cationic exchange capacity (CEC), such as fluorohectorite,1-3 montmorillonite, or saponite have been used as hosts.4-7 In all cases, a supplementary neutral surfactant is necessary as cosurfactant for the correct formation of silica galleries by the hydrolysis and condensation of tetraethylorthosilicate (TEOS), used as the silica source. In addition, extensive research has been devoted to the preparation of mesoporous silica of hexagonal (MCM-41), cubic (MCM48), or lamellar (MCM-50) geometry, by reacting surfactant micelles with different inorganic arrays.8,9 In the case of solids of the MCM-41 or MCM-48 type, porous materials with high specific surface areas and well-defined porous diameters are * To whom correspondence should be addressed. E-mail: geaguila@ siu.buap.mx. † Beneme ´ rita Universidad Auto´noma de Puebla. ‡ Universidad de Ma ´ laga. (1) Galarneau, A.; Barodawalla, A.; Pinnavaia, T. J. Nature 1995, 374, 529. (2) Galarneau, A.; Barodawalla, A.; Pinnavaia, T. J. Chem. Commun. 1997, 1661. (3) Mercier. L.; Pinnavaia, T. J. Microporous Mesoporous Mater. 1998, 20, 101. (4) Polverejan, M.; Pauly, T. R.; Pinnavaia, T. J. Chem. Mater. 2000, 12, 2698. (5) Ahenach, J.; Cool, P.; Vasant, E. F. Phys. Chem. Chem. Phys. 2000, 2, 5750. (6) Benjelloun, M.; Cool, P.; Linssen, T.; Vasant, E. F. Microporous Mesoporous Mater. 2001, 49, 83. (7) Pichowicz, M.; Mokaya, R. Chem. Commun. 2001, 2100. (8) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature 1992, 359, 710. (9) Corma, A. Chem. ReV. 1997, 97, 2373. (b) Ciesla, U.; Schu¨th, F. Microporous Mesoporous Mater. 1999, 27, 131.

obtained after removal of surfactant molecules, generally by calcination. However, in the case of solids of the MCM50 type, nonporous materials were obtained upon calcination. On the other hand, we synthesized a lamellar zirconium phosphate, of MCM-50 type, expanded with cationic surfactant molecules. This phosphate was used as the starting material for the insertion of gallium oxide into its interlayer space by cationic exchange of the surfactant guest with oligomeric gallium species, obtaining a porous structure as in the case of PLS materials.10 Very recently, we have combined the strategies for obtaining mesoporous materials, such as the formation of pillared layer structures (PLSs) and MCM-41, for the preparation of porous phosphate heterostructure (PPH) solids consisting of silica galleries in the interlayer space of zirconium phosphate with high specific surface area and acidity,11 where the obtained solids present a strong Bro¨nsted acidity as deduced by the formation of iso derivatives in the 1-butene isomerization test. The separation of mixtures of olefin/paraffin is a subject of current interest, and several types of zeolites (NaX, ZSM-5) exchanged with Co(II), Cu(II), and Zn(II) have been tested.12-14 Other porous solids, such as γ-Al2O3 and SiO2, ion-exchanged or impregnated with monovalent cations proved to be efficient in the separation process.15,16 MCM-41 ion-exchanged with Cu(II) and Ag(I),16 mesoporous Cu(II)- and Ag(I)-derivatized (10) Jime´nez-Jime´nez, J.; Maireles-Torres, P.; Olivera-Pastor, P.; Rodrı´guezCastello´n, E.; Jime´nez-Lo´pez, A. Langmuir 1997, 13, 2857. (11) Jime´nez-Jime´nez, J.; Rubio-Alonso, M.; Eliche-Quesada, D.; Rodrı´guezCastello´n, E.; Jime´nez-Lo´pez, A. J. Mater. Chem. 2005, 15, 3466. (12) Khelia, A.; Derriche, Z.; Bengueddach, A. Appl. Catal., A 1999, 178, 61. (13) Corma, A.; Ray, F.; Rius, J.; Sabater, M. J.; Valencia, S. Nature 2004, 431, 287. (14) Ivanov, A. V.; Graham, G. W.; Shelef, M. Appl. Catal., B 1999, 21, 243. (15) Masuda, T.; Okubo, Y.; Mukai, S. R.; Kawase, M.; Hashimoto, K.; Shichi, A.; Satsuma, A.; Hattori, T.; Kiyozumi, Y. Chem. Eng. Sci. 2001, 56, 889. (16) Blas, F. J.; Vega, L. F.; Gubbins, K. E. Fluid Phase Equilib. 1998, 150151, 117.

10.1021/la052390e CCC: $33.50 © 2006 American Chemical Society Published on Web 12/22/2005

Application of PPH Materials for Gas Separation

aluminosilicate materials,17 mesoporous Ag(I)-impregnated SBA15,18 and SBA-15 ion-exchanged with Ag(I) and K+ natural erionite19 have also been used, where the idea is the separation of propylene/propane mixtures via π-complexation. With the same mechanism, very recently, mesoporous γ-Al2O3 and microporous SiO2 membranes impregnated with Ag(I) have been successfully tested in this separation process at 313 K.20 In this study we propose the use of a silica-expanded zirconium phosphate heterostructure, which will be ion-exchanged with Cu(II) and Cu(I) in the separation of propylene (C3H6)/propane (C3H8) mixtures. Experimental Section Materials. A cetyltrimethylammonium (CTMA)-expanded zirconium phosphate was prepared from a solution of (CTMA)Br (Aldrich) in 1-propanol (Rectapur), to which H3PO4 (85%, BDH) and zirconium(IV) propoxide (70%, Aldrich) were loaded according to previously reported procedures.11 The obtained solid (CTMAZrP) was suspended in water (10 g L-1), and a solution of hexadecylamine (Aldrich) in 1-propanol (35 g L-1) was added as cosurfactant. After 1 day under stirring, a solution (50%, v/v) of TEOS (Aldrich) in 1-propanol was added, and this suspension was stirred at room temperature for 3 days. The solid obtained was then centrifuged, washed with ethanol, and dried at 333 K in air. This precursor material was calcined in air at 823 K for 5 h (1.5 K min-1 heating rate). The obtained solid 3SiPPH(0.2), with a d001 ) 40 Å, a specific surface area (BET) of 619 m2 g-1, and a total pore volume of 0.552 cm3 g-1, was used as the support for this study.11 The ionic exchange capacity of the support (1.73 mequiv g-1) was determined by CNH elemental analysis of a sample exposed under an atmosphere of NH3 for 5 min and then kept for 24 h in a desiccator with concentrated H3PO4. This value was similar to that obtained by ammonia thermally programmed desorption (NH3-TPD) (1.66 mequiv g-1). Cu(II) was exchanged by adding an amount equivalent to 5 times the cationic exchange capacity using a 0.1 M (1:1, v/v, ethanol/water) solution of copper(II) acetate monohydrate to the support, 3SiPPH(0.2), and stirring for 1 day at 298 K. The solid was centrifuged, washed with ethanol, and dried at 333 K (sample Cu2+-3SiPPH). This sample then was calcined at 723 K for 5 h with a heating rate of 1.5 K min-1 (sample Cu2+-3SiPPH723). The sample Cu+-3SiPPH723, ion-exchanged with Cu(I), was prepared in a similar way using an ethanolic solution of copper(I) chloride, but adding an amount equivalent to the cationic exchange capacity. Characterization of Samples. X-ray diffraction (XRD) powder patterns were recorded on a Siemens D501 diffractometer (Cu KR radiation) provided with a graphite monochromator. X-ray photoelectron spectroscopy (XPS) analyses were recorded using a Physical Electronics PHI 5700 spectrometer with nonmonochromatic Mg KR radiation (300 W, 15 kV, 1253.6 eV) as the excitation source. High-resolution spectra were recorded at a 45° takeoff angle by using a concentric hemispherical analyzer operating in the constant pass energy mode at 29.35 eV and an analysis area 720 µm in diameter. Under these conditions, the Au 4f7/2 line was recorded with 1.16 eV fwhm at a binding energy of 84.0 eV. The spectrometer energy scale was calibrated by using the Cu 2p3/2, Ag 3d5/2, and Au 4f7/2 photoelectron lines at 932.7, 368.3, and 84.0 eV, respectively. Cu 2p core level spectra were first recorded by irradiation for 10 min to avoid the photoreduction of Cu(II) to Cu(I). N2 adsorptiondesorption isotherms at 77 K were obtained using a conventional volumetric device, after outgassing of samples at 413 K for 24 h, or using a Micromeritics ASAP2020 apparatus. NH3-TPD was used (17) Kargol, M.; Zajac, J.; Jones, D. J.; Steriotis, Th.; Rozie`re, J.; Vitse, P. Chem. Mater. 2004, 16, 3911. (18) Grande, C.; Araujo, J. D. P.; Cavenati, S.; Firpo, F.; Basaldella, E.; Rodrigues, A. E. Langmuir 2004, 20, 5291. (19) Aguilar-Armenta, G.; Patin˜o-Iglesias M. E. Langmuir 2002, 18, 7456. (20) Stoitsas, K. A.; Gotzias, A.; Kikkinides, E. S.; Steriotis, Th. A.; Kanellopoulos, N. K.; Stoukides, M.; Zaspalis, V. T. Microporous Mesoporous Mater. 2005, 78, 235.

Langmuir, Vol. 22, No. 3, 2006 1261 to determine the total acidity of the samples. Before the adsorption of ammonia at 373 K, the samples were heated at 773 K in a He flow. NH3-TPD was performed between 373 and 773 K, at a heating rate of 10 K min-1, using a thermal conductivity detector (TCD) for the analysis. Hydrogen temperature-programmed reduction (H2-TPR) experiments were carried out between 323 and 823 K, using a flow of 10% H2/Ar (48 mL min-1) and a heating rate of 14 K min-1. Water produced in the reduction reaction was eliminated by passing the gas flow through a coldfinger (193 K). The H2 consumption was controlled by an on-line gas chromatograph (Shimadzu GC-14A) provided with a TCD. CHN elemental analysis was used to determine the surfactant content and was carried out with a LECO instrument; ion-exchanged metals were determined by atomic absorption (AA) with a Perkin-Elmer AA420 spectrometer, while phosphorus was determined colorimetrically. The C3H6 and C3H8 adsorption capacity of the samples was measured in a conventional glass high-vacuum volumetric device, equipped with grease-free valves. Pressures were registered with two types of pressure transducers (Balzers) of different ranges: TPR 017 (0.01333-666.61 Pa) and APR 011 (0.13332-133.322 kPa). Before the measurements, the samples were dehydrated in situ at 693 K in an oven with a residual pressure of 0.01333 Pa and kept in these conditions for a period of 3 h. After sample dehydration, the temperature was lowered to the desired point, and the sample was allowed to stabilize for at least 1.5 h before the measurements were started. The adsorbed amount of the gases was referred to 1 g of dehydrated adsorbent. Weight loss of the adsorbent was previously assessed by heating the sample at 623 K under atmospheric pressure in a conventional oven. The adsorption uptake of gases as a function of time, t, was obtained on the basis of the difference between the initial amount of gas introduced into the cell and the amount of gas remaining in the dead space of the cell at any given time t, from t ) 0 up to teq (equilibrium). During the kinetic measurements, the decrease of pressure in the system was measured automatically with a custom acquisition data card, which allows simultaneous monitoring and recording of time and pressure. In the period from 0 to 3 min, pressure was monitored five times per second. Afterward, in the period from 3 to 13 min, pressure was registered once per second, and pressure was recorded once every 10 s in the period from 13 min to teq. In all experiments, the initial pressure was 59.995 kPa. Since the adsorbents usually work continuously in a given adsorption process, as for instance in pressure-swing adsorption (PSA) cycles, it is necessary to evaluate the extent of irreversible adsorption, i.e., the amount of adsorbate which cannot be desorbed in a vacuum at the given experimental temperature. The experimental procedure to assess irreversible adsorption was described in detail in a previous paper.21

Results and Discussion Structural Properties of Samples. The formation of silica galleries in the interlayer space of the layered zirconium phosphate is the key to generate high surface areas, and in Scheme 1 the mechanism of formation of the used support is described. Table 1 shows the structural and textural properties of the studied solids.11 The basal spacing is modified upon impregnation with copper, but this modification is very ambiguous due to the broadness of the d001 reflection line. As an example, Figure 1 shows the powder patterns of sample Cu2+-3SiPPH723, where a maximum centered at 48 Å close to several shoulders at a higher angle is observed. The N2 adsorption-desorption isotherms at 77 K of the copperimpregnated solids are type IV isotherms corresponding to mesoporous materials and similar to that observed in the case of the support, 3SiPPH(0.2).11 However, the surface BET values are smaller due to the formation of a surface of copper species (21) Patin˜o-Iglesias, M. E.; Aguilar-Armenta, G.; Jime´nez-Lo´pez, A.; Rodrı´guez-Castello´n, E. Colloids Surf., A 2004, 237, 73.

1262 Langmuir, Vol. 22, No. 3, 2006

Aguilar-Armenta et al. Scheme 1. Mechanism of Support Formation

Table 1. Structural and Textural Properties of the Support and Exchanged Samples sample

d001 (Å)

SBET (m2 g-1)

Vp (cm3 g-1)

rp (Å)

3SiPPH(0.2) Cu2+-3SiPPH723 Cu+-3SiPPH723

41 48 43

619 409 325

0.552 0.397 0.316

16.3 17.2 17.2

Table 2. XPS Data of the Samples sample 3SiPPH(0.2) Cu2+-3SiPPH Cu2+-3SiPPH723 Cu+-3SiPPH723

Cu 2p3/2 935.8 935.2 933.0 (64%) 935.4 (36%)

P 2p

Zr 3d5/2

Si 2p

O 1s

133.8 133.7 133.3 133.5

183.1 183.1 182.8 183.1

102.9 102.9 102.9 102.9

532.3 532.1 531.9 532.0

which block some pores, avoiding free access of N2 molecules, whereas their average pore radius (rp) is similar to that of the support (Table 1). That is, the structure of the support is preserved upon impregnation with copper and calcination at 723 K. XPS data show (Table 2) that (i) the binding energy values of the P 2p, Zr 3d5/2, Si 2p, and O 1s core level peaks of the Cu-impregnated samples are similar to those of the support, evidencing again the preservation of the structure upon impregnation and calcination, (ii) the Cu 2p3/2 photoemission of the samples Cu2+-3SiPPH and Cu2+-3SiPPH723 correspond to the exclusive existence of Cu(II) species (Figure 2), but the intensity of the Cu 2p3/2 photoemission is much higher in the case of the calcined sample (Cu2+-3SiPPH723), probably due to the formation of CuO on the surface upon calcination in addition to Cu(II) located in exchange positions, (iii) the Cu 2p core level spectrum (Figure 3) of the Cu+-3SiPPH723 sample shows two types of copper species, Cu(I) in a higher proportion at a lower binding energy of 933.0 eV and Cu(II) with a peak at 935.4 eV,

Table 3. XPS Data for the Support and the Copper-Exchanged Samples (atom %) sample

Cu

P

Zr

3SiPPH(0.2) 7.97 3.77 Cu2+-3SiPPH 2.82 9.08 5.06 Cu2+-3SiPPH723 5.71 8.93 4.36 Cu+-3SiPPH723 2.27 9.82 4.84

Si

O

18.20 14.48 12.57 12.44

70.05 68.56 68.44 70.62

P/Zr Cu/Si 2.11 1.79 2.05 2.03

0.19 0.45 0.18

and (iv) the chemical composition of the samples, expressed in atomic concentration percentages (Table 3), also supports some of the above asseverations since the P:Zr molar ratio is close to the theoretical value of 2.00 and the Cu:Si molar ratio of sample Cu2+-3SiPPH723 is higher than that of sample Cu2+-3SiPPH due to the surface segregation of CuO. The acidity of the samples (Figure 4) shows that the total acidity of the support was higher than that for the other samples (1659 µmol of NH3 g-1) and that upon impregnation with Cu(I) and Cu(II) the acidity was slightly (1623 µmol of NH3 g-1) and noticeably (956 µmol of NH3 g-1) reduced, respectively. The main feature of these results is that an important proportion of ammonia is desorbed at higher temperatures (>673 K) in the case of the support, as compared with the copper-impregnated samples, indicating a higher proportion of strong Bro¨nsted acid sites in the support in comparison with the copper-impregnated samples. H2-TPR is helpful in obtaining more information about copper species present in the copper-impregnated structures. The H2 consumption curve for sample Cu2+-3SiPPH723 (Figure 5a) shows a sharp maximum centered at 510 K assigned to the

Figure 3. Cu 2p core level spectrum of the Cu+-3SiPPH723 sample.

Figure 1. XRD patterns of sample Cu2+-3SiPPH723.

Figure 2. XPS patterns of the Cu2+-3SiPPH723 and Cu2+-3SiPPH samples.

Figure 4. Acidity of the samples (NH3-TPD).

Application of PPH Materials for Gas Separation

Figure 5. H2-TPR curves for samples with Cu: (a) Cu2+3SiPPH723 and (b) Cu+-3SiPPH723.

reduction of Cu(II) ions to Cu(0). This reduction temperature is higher than that of the bulk CuO,22 and near but higher than those observed in some Cu(II)-exchanged zeolites.23 This is indicative of the strong interaction of Cu(II) with the strong acid sites of the support. The presence of the surface CuO was not confirmed, perhaps because this species is only located on the external surface of the support. The H2-TPR curve for sample Cu+-3SiPPH723 (Figure 5b) shows two maxima. The first sharply centered at 507 K is very similar to that observed for sample Cu2+-3SiPPH723 and also assigned to the reduction of Cu(II) species, and the second one has a maximum at 603 K with a shoulder at 564 K. This peak is assigned to the reduction of Cu(I) to Cu(0).24 Kinetics and Equilibrium of Adsorption. Support 3SiPPH(0.2). The kinetic curves of the total (reversible + irreversible) and reversible adsorption of C3H6 on the support at 273, 293, 313, and 333 K for short, t < 160 s, and long, t < 15000 s, gas-adsorbent contact times, respectively, are shown in parts a and b of Figure 6. Clearly, the irreversible adsorption (total - reversible) uptake tends to increase with increasing temperature. The total adsorbed amount at equilibrium (neq) decreases with increasing temperature in the range between 273 and 313 K, but when the temperature was increased to 333 K, neq was higher (1.2192 mmol g-1) than that at 313 K (1.1250 mmol g-1). It was established that the total adsorption capacity for C3H6 at 273, 293, and 313 K occurred rapidly during the first 20 s (Figure 7a), achieving a fractional uptake nt/neq ) 0.8, where nt is the adsorbed amount at time t, and at 333 K this fractional uptake was reached for a contact time of 7500 s; that is, a slow adsorption for temperatures higher than 313 K was observed. The adsorption equilibrium time (teq), i.e., the time required to achieve the total adsorption capacity nt/neq ) 1, increased in the order of 900, 1750, 7900, and 21500 s for temperatures of 273, 293, 313, and 333 K, respectively (Figure 7b). These results (22) Torre-Abreu, C.; Henriques, C.; Ribeiro, F. R.; Delahay, G.; Ribeiro, M. F. Catal. Today 1999, 54, 407. (23) Moreno-Tost, R.; Santamarı´a-Gonza´lez, J.; Rodrı´guez-Castello´n, E.; Jime´nez-Lo´pez, A.; Autie´, M. A.; Gonza´lez, E.; Carreras Glacial, M.; Delas Pozas, C. Appl. Catal., B 2004, 50, 279. (24) Coq, B.; Tachon, F.; Figueras, F.; Mabillon, G.; Prigent, N. Appl. Catal., B 1995, 6, 271.

Langmuir, Vol. 22, No. 3, 2006 1263

Figure 6. Kinetic curves of total (full symbols) and reversible (empty symbols) adsorption of propylene on 3SiPPH(0.2) at various temperatures: (b, O) 273 K; (2, 4) 293 K; (9, 0) 313 K; ([, ]) 333 K.

Figure 7. Fractional adsorption uptake for propylene on 3SiPPH(0.2) at various temperatures.

indicate that the adsorption process of C3H6 exhibits some general features (activated, slow, and irreversible) of chemisorption which involves the formation of a chemical bond between the olefin molecule and the surface of the sorbent. In view of these results, it was interesting to carry out adsorption kinetic measurements at a higher temperature (423 K). The kinetic curves of the total adsorption of C3H6 are shown in parts a and b of Figure 8 for short and long t, respectively, at 273, 293, 313, 333, and 423 K. The adsorption uptake at 423 K for short t was too small and increased continuously with time, reaching not only the adsorption uptake at 333 K, but even the one at 273 K. Apparently, the propylene is chemisorbed, mainly for temperatures higher than 313 K, on the Bro¨nsted acidic centers, leading probably to a cationic polymerization. To support this supposition, the chemical composition of the surface after C3H6 adsorption at 423 K was assessed. After adsorption at 423 K, part of the organic phase of the solid was extracted with n-heptane and then analyzed by GC-MS. Many hydrocarbons were detected, but two intense peaks with molecular weights of 168

1264 Langmuir, Vol. 22, No. 3, 2006

Aguilar-Armenta et al.

Figure 10. Isosteric heat of adsorption as a function of the adsorbed amount. QL ) latent heat of condensation.

Figure 8. Adsorption kinetic curves of propylene on 3SiPPH(0.2) at various temperatures.

Figure 11. Kinetic curves of total (full symbols) and reversible (empty symbols) adsorption of C3H6 on Cu+-3SiPPH723 at various temperatures (K). (b, O) 293; (2, 4) 333; ([, ]) 393, (0) 393, C3H8.

Figure 9. Adsorption equilibrium isotherms of C3H6 (full symbols) and C3H8 (empty symbols) on 3SiPPH(0.2) at 273 K (b, O) and 293 K (2, 4). Lines are the Freundlich model fits.

and 210 were observed, which correspond to C12H24 and C15H30 olefins. Unlike the C3H6 adsorption, the C3H8 adsorption on the 3SiPPH(0.2) sample was reversible at all the studied temperatures. The adsorption equilibrium isotherms of both hydrocarbons at 273 and 293 K (Figure 9) are nonrectangular in shape and were described fairly well by the Freundlich empirical equation with a correlation coefficient R2 > 0.995. To assess the strength of the bonding between the adsorbate and the surface, the isosteric heat of adsorption (Qst) with gas loading was computed from equilibrium data (Figure 9) using the Clausius-Clapeyron equation, which at constant adsorbate loading results in

-∆H ) Qst ) R

[

]()

T1T2 P2 ln T1 - T 2 P1

(1)

where P2 and P1 are the equilibrium adsorption pressures at temperatures T2 and T1, respectively, at a specific adsorbate loading. According to the obtained results (Figure 10), the isosteric heat of adsorption decreases and tends to the latent heat of condensation (QL) with gas loading for both hydrocarbons, which is a characteristic of highly heterogeneous adsorbents with a wide distribution of gas-solid interaction energies.25,26 The Qst values at very low loading (0.1 mmol g-1) for propylene and propane were 2.4 and 1.6 times higher than the latent heat of evaporation, respectively. By virtue of the adsorption of nonpolar (25) Dunne, J. A.; Mariwala, R.; Rao, M.; Sircar, S.; Gorte, R. J.; Myers, A. L. Langmuir 1996, 12, 5888. (26) Dunne, J. A.; Rao, M.; Sircar, S.; Gorte, R. J.; Myers, A. L. Langmuir 1996, 12, 5896.

molecules such as propane only presenting nonspecific interactions, which are largely determined by the polarizability of the adsorbate, the contribution of π electrons to the total adsorbateadsorbent interactions can be assessed by the difference ∆Qst ) Qst(C3H6) - Qst(C3H8) as a function of loading. The small irreversible adsorption observed at 273 and 293 K (Figure 6a), which increases with increasing temperature, can be due to the chemisorption of propylene beginning to be apparent in the Bro¨nsted acidic adsorption sites on the surface. Cu+-3SiPPH723. The kinetic curves of the total and reversible adsorption of C3H6 were measured at 293, 313, 333, 373, and 393 K. For clarity, Figure 11 presents only the kinetic curves of C3H6 at 293, 333, and 393 K. In contrast to that on the support, irreversible adsorption of C3H6 decreases with increasing temperature and disappears at 393 K. The irreversible adsorption observed at T < 333 K is not due to irreversible polymerization of the olefin because the adsorbed gas could be entirely removed with the aid of heat at vacuum. Figure 11 includes the adsorption kinetic curve for C3H8 at 393 K for comparison. The adsorption of propane, as on the support (393 K), was always reversible. Because of the reversible adsorption of propylene and the insignificant adsorption of propane observed at 393 K, this solid could be used for separating mixtures of these hydrocarbons at this temperature. To study the adsorption behavior at equilibrium, the adsorption isotherms of C3H6 and C3H8 were obtained at 293, 313, 333, and 373 K (Figure 12). For clarity, in this figure only the C3H8 adsorption isotherms measured at 293 and 313 K are included. The adsorption isotherms for propylene show a sharp increase at low pressures ( 2.666 kPa, it can be assumed that, once the copper cations Cu(I) are saturated by C3H6 molecules at low equilibrium pressures (Peq < 2.666 kPa), the adsorption process of this hydrocarbon takes place by a mechanism of nonspecific interactions (attraction forces + repulsion + polarization), that is, following the usual mechanism in the adsorption of C3H8. (27) Duong, D. D. Adsorption Analysis: Equilibria and Kinetics; Series on Chemical Engineering, Vol. 2; Imperial College Press: London, 1998.

Figure 14. Adsorption equilibrium isotherms of propylene on Cu2+SiPPH723 at various temperatures (K): (b) 293; (2) 313; ([) 333; (9) 373; (O) 423; (4) 443.

Cu2+-3SiPPH723. Figure 13 depicts the total and reversible adsorption kinetic curves of propylene at 293, 313, 333, 393, and 423 K for low contact times ( Cu2+-3SiPPH723. These results may be explained assuming that the C3H6 molecules are adsorbed in two different adsorption sites with different adsorption potentials, following a monolayer mechanism at low pressures, where Cu+ and Cu2+ cations are the specific adsorption sites for propylene, and when the pressure is increased, the adsorption process follows a multilayer mechanism. Figure 16 shows that the irreversible adsorbed amount of propylene at 293 K increases in the order 3SiPPH(0.2) < Cu2+3SiPPH723 < Cu+-3SiPPH723. This means that, in the samples with copper, the irreversible adsorption was mainly due to the π-complexation bonding28 between the adsorbate and Cu+. (28) Yang, R. T. Adsorbents, Fundamentals and Applications; John Wiley and Sons: Hoboken, NJ, 2003.

Although the Cu2+-3SiPPH723 sample has a higher amount of copper (Table 3), the fact that the irreversible adsorbed amount of propylene was higher for the Cu+-3SiPPH723 sample can be due to the π-back-donation of d electrons of Cu+ to the antibonding π-orbitals of the olefin being more easily achieved. The maximum of irreversible adsorption of C3H6 (Figure 16), observed for the Cu2+-3SiPPH723 sample, can be due to chemisorption of some molecules on weak acidic sites, since irreversible adsorption begins to decrease starting at 313 K and does not lead to polymerization reaction of the olefin. Unlike that on the support, the fact that the irreversible adsorbed amount on the samples Cu2+-3SiPPH723 and Cu+-3SiPPH723 decreases with temperature, disappearing at 393 K, indicates that the bond between the olefin and the adsorbent is not strong enough to cause a chemical reaction on the surface. The variation of the isosteric heat of adsorption (Qiso), as a function of the adsorbed amount of C3H6 and C3H8, is shown in Figure 17. Unlike that for the support, the heat of adsorption profiles of C3H6 for the samples with copper show a maximum, indicating that the gas-solid interaction energy is fairly constant at low coverage and the contributions of the collateral interactions (C3H6 T C3H6) on the surface are dominant while the monolayer has not been completed. These results indicate that the copper cations occupied the stronger acidic sites over the support surface, converting it into an energetically more homogeneous adsorbent. When the monolayer has been completed, that is, when the copper cations are saturated, the gas-adsorbent interactions are predominant and the adsorption process occurs on an energetically heterogeneous surface. As compared to the support, the samples with copper show a similar slight increase of the isosteric heat of adsorption for C3H8, because this molecule is not able to adsorb specifically. Figure 18 compares the adsorption kinetic behavior for propylene and propane in the Cu+-3SiPPH723 and Cu2+3SiPPH723 samples at 393 K. Taking into account the considerable and insignificant adsorption of propylene and propane, respectively, at 393 K, these samples present a high selectivity toward olefin. Thus, if the C3H6/C3H8 mixtures were contacted with any of the two studied samples, C3H6 would be adsorbed

Application of PPH Materials for Gas Separation

preferentially, leading to a separation of these gases. Figure 18 shows that Cu+-3SiPPH723 would have the highest efficiency, because its ability to adsorb C3H6 was higher than that for the Cu2+-3SiPPH723 sample.

Conclusions In this study, a silica-expanded zirconium phosphate heterostructure has been used as a support, 3SiPPH(0.2), to prepare two mesoporous adsorbents, Cu+-3SiPPH723 and Cu2+3SiPPH723, by ion exchange with Cu(I) and Cu(II) for separation of propylene/propane mixtures. Adsorption equilibrium and kinetic measurements of pure C3H6 and C3H8 in the three samples have been performed at different temperatures in the range between 273 and 393 K. A marked influence of Cu cations on the adsorption behavior of the olefin was established. It was also established that the adsorption process of propylene in the support occurs by a mechanism similar to that of chemisorption, leading to a polymerization reaction at 423 K. This is attributable to the

Langmuir, Vol. 22, No. 3, 2006 1267

high acidity of the adsorbent. Unlike that on the support, 3SiPPH(0.2), the irreversible adsorption of the olefin on the samples with Cu decreases with increasing temperature and disappears at 393 K, showing a high selectivity toward propylene. The C3H8 adsorption was always reversible. On the basis of the results of this study, the Cu+-3SiPPH723 and Cu2+-3SiPPH723 samples can be applied in the separation of a C3H6/C3H8 mixture at 393 K. Acknowledgment. We gratefully acknowledge the Ministerio de Ciencia y Tecnologı´a (Spain) (Project MAT2003-02986) and CYTED Project V.8, “Clean Technology for the Separation of Light Olefins”, for funding this work. We also thank the Consejo Nacional de Ciencia y Tecnologı´a (CONACYT; Me´xico) for financial support via a scholarship (Grant No. 144883) for M.E.P.-I. LA052390E