- Email: [email protected]
Thermodynamic studies on hydrogen adsorption on the zeolites Na-ZSM-5 and K-ZSM-5
Microporous and Mesoporous Materials 80 (2005) 247–252 www.elsevier.com/locate/micromeso
Thermodynamic studies on hydrogen adsorption on the zeolites Na-ZSM-5 and K-ZSM-5 C. Otero Area´n a,*, M. Rodrı´guez Delgado a, G. Turnes Palomino a, M. Toma´s Rubio a, N.M. Tsyganenko b, A.A. Tsyganenko b, E. Garrone
c
a
c
Departamento de Quı´mica, Universidad de las Islas Baleares, E-07122 Palma de Mallorca, Spain b Institute of Physics, St. Petersburg University, 198504 St. Petersburg, Russia Dipartimento di Scienza dei Materiali ed Ingegneria Chimica, Politecnico di Torino, I-10126 Torino, Italy Received 17 September 2004; accepted 6 December 2004 Available online 26 January 2005
Abstract Adsorption of dihydrogen onto the zeolites Na-ZSM-5 and K-ZSM-5 renders the fundamental H–H stretching mode infrared active. The corresponding infrared absorption bands were found at 4101 and 4112 cm1 for H2/Na-ZSM-5 and H2/K-ZSM-5, respectively. Thermodynamic characterization of the adsorbed state was carried out by means of variable-temperature infrared spectroscopy; simultaneously measuring integrated band intensity, temperature and equilibrium pressure of the gas phase. For the H2/ Na-ZSM-5 system, the standard adsorption enthalpy and entropy resulted to be DH = 10.3 (±0.5) kJ mol1 and DS = 121 (±10) J mol1 K1. For H2/K-ZSM-5 corresponding values were 9.1 (±0.5) kJ mol1 and 124 (±10) J mol1 K1, respectively. 2004 Elsevier Inc. All rights reserved. Keywords: Adsorption thermodynamics; Hydrogen; FTIR; Zeolites; ZSM-5
1. Introduction Zeolite molecular sieves are often used for industrial gas separation processes based on differences in adsorption strength among components of a gas mixture. Thus, pressure swing adsorption (employing zeolites) is commercially used, inter alia, for producing oxygen from air and hydrogen from a variety of feedstocks; such as refinery off-gas and hydrogen-rich gas from steam reforming of hydrocarbons. On the other hand, both zeolites and active carbons can be considered [1–3] as being potential candidates for large-scale hydrogen storage in cryogenically cooled vessels; a subject of current interest
*
Corresponding author.
1387-1811/$ - see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.micromeso.2004.12.004
in view of the potential use of this gas as a cost-effective energy vector for buildings and transportation. Regarding hydrogen, it should be clear that precise knowledge of the interaction energy between the gas and the adsorbent would facilitate improvement of the above mentioned applications. However, direct measurement of such an interaction energy by adsorption microcalorimetry is a difficult task to perform. This is so because the low temperature needed for adsorption to take place, and the small amount of heat evolved, demand very stringent requirements on instrument design and sensitivity. We have recently shown [4,5] that variable-temperature infrared (IR) spectroscopy can advantageously be used for studying adsorption thermodynamics of weakly interacting adsorbent–adsorbate systems, and the method is applied here to hydrogen adsorption on both
C. Otero Area´n et al. / Microporous and Mesoporous Materials 80 (2005) 247–252
Na-ZSM-5 and K-ZSM-5. Related work on Li-ZSM-5 was reported elsewhere [6]. Interaction of the dihydrogen molecule with the zeolite adsorbing centres renders the fundamental H–H stretching mode IR active. By measuring the integrated intensity of the corresponding IR absorption band over a range of temperature, and the corresponding gas equilibrium pressure, the desired values of adsorption enthalpy and entropy are obtained. These values fully characterise the adsorbed state, as shown below.
2. Experimental details The Na-ZSM-5 and K-ZSM-5 samples used were prepared from a commercial NH4-ZSM-5 sample (Si/ Al = 25) by ion exchange with the corresponding alkali-metal nitrate solutions. Powder X-ray diffraction showed good crystallinity in both cases, and absence of any diffraction line not assignable to the corresponding MFI structure type. Complete ion exchange was checked by the absence of IR absorption bands corresponding to either the ammonium ion or the Brønstedacid OH group, which would be formed during thermal activation (see below) of the zeolite samples if total exchange of alkali-metal ions for ammonium did not take place. For IR spectroscopic measurements, thin selfsupported wafers of the zeolite samples were prepared and activated (outgassed) in a dynamic vacuum (residual pressure < 107 bar) for 2 h at 700 K inside an IR cell [7], which allowed in situ high-temperature activation, gas dosage and variable-temperature IR spectroscopy to be carried out. For each sample, after running the blank spectrum of the zeolite wafer at 77 K the cell was dosed with hydrogen, it was then closed and spectra were taken at 2–5 K intervals while simultaneously recording sample temperature and hydrogen equilibrium pressure inside the cell. A platinum resistance thermometer and a capacitance pressure gauge were used for this purpose. The precision of these measurements was of about ±2 K and ±2 · 105 bar, respectively. In order to check reproducibility, and also for improving accuracy, three independent sets of measurements were taken on each sample. In these independent runs different initial doses of hydrogen were used, so as to cover a larger pressure range. Note that the high thermal conductivity of hydrogen, together with the static nature of measurements, should minimize (possible) temperature gradients inside both the IR cell and the sample wafer. Since no temperature offset effects were found when plotting results from data obtained under widely different hydrogen pressure (in the different runs for each sample), it seems reasonable to assume that temperature gradients (if any) were of little concern. Transmission FTIR spectra were recorded at 3 cm1 resolution by means of a Bruker IFS66 spectrometer.
3. Results and discussion IR spectra of dihydrogen adsorbed (at 77 K) on alkali-metal-exchanged faujasites, mordenite, and A-type zeolites were reported by several authors [8–13]; mainly with a view of using dihydrogen as a probe molecule for IR spectroscopic studies of zeolites and related materials. These spectra show a main IR absorption band in the 4070–4100 cm1 range which corresponds to the fundamental H–H stretching mode of adsorbed dihydrogen. Perturbation of the molecule (mainly) by the cationic adsorption centre renders this vibrational mode IR active and causes a bathochromic shift from the gas phase value of 4163 cm1 corresponding to the free molecule (Raman active H–H stretching vibration). Earlier studies [8,14] suggested that the H2 molecule is adsorbed on alkali-metal cations of zeolites forming T-shaped adsorption complexes. However, more recent work [10,12,15] has shown that the adsorbed molecule can also interact with neighbouring anions of the zeolite framework. Hence, the adsorption centre should be taken to represent both the extra-framework alkali-metal cation and the nearby oxygens [16]. Fig. 1 shows some selected variable-temperature IR spectra, in the H–H stretching region, of hydrogen adsorbed on Na-ZSM-5; the H–H stretching band is seen at 4101 cm1. For hydrogen adsorbed on K-ZSM-5, corresponding spectra are depicted in Fig. 2; which shows the H–H stretching band at 4112 cm1. For the
0.18 100
0.16
Absorbance
248
102
0.14
104
0.12
108 113
0.10
115
0.08
119 123
0.06
126
0.04
129
0.02 0.00 4140
4120
4100
4080
Wavenumber / cm-1
4060
Fig. 1. Selected difference IR spectra (zeolite blank subtracted) in the H–H stretching region corresponding to the large hydrogen dose on Na-ZSM-5 (Table 1). Numbers denote temperature in K.
C. Otero Area´n et al. / Microporous and Mesoporous Materials 80 (2005) 247–252
96
Eq. (2), leads to Eq. (3) below, from which the standard enthalpy and entropy changes (involved in the adsorption process) were obtained:
97
kðT Þ ¼ exp½DS =R exp½DH =RT
ð2Þ
ln½h=ð1 hÞp ¼ ðDH =RT Þ þ ðDS =RÞ
ð3Þ
0.12
0.10
99 100
0.08
Values of temperature, pressure and absorbance (integrated band intensity) corresponding to the whole set of IR spectra taken in three independent runs on the H2/Na-ZSM-5 and H2/K-ZSM-5 systems are given in Tables 1 and 2, respectively. In order to convert absorbance into the corresponding h value, knowledge of AM is needed. For each system, an approximate value of AM was obtained by running spectra corresponding
Absorbance
101 104
0.06
249
107 110
0.04
115 119
0.02
Table 1 Numerical data corresponding to three measurement runs on NaZSM-5
0.00
Absorbance (cm1)
Coverage (h)
Small hydrogen dose 89 3.78 91 4.20 95 4.88 107 5.76 121 6.58 132 7.27
2.52 1.92 1.61 0.69 0.31 0.14
0.72 0.55 0.46 0.20 0.09 0.04
Medium 96 99 101 104 107 112 116 120 123 125 128 130 132 134 137 140
2.67 2.42 2.13 1.84 1.58 1.26 0.99 0.83 0.68 0.53 0.47 0.38 0.33 0.26 0.26 0.19
0.76 0.69 0.61 0.53 0.45 0.36 0.28 0.24 0.20 0.15 0.13 0.11 0.09 0.08 0.07 0.05
2.95 2.74 2.49 2.33 2.10 1.86 1.65 1.51 1.41 1.26 1.13 0.94 0.91 0.75 0.66 0.56
0.84 0.78 0.71 0.66 0.60 0.53 0.47 0.43 0.40 0.36 0.32 0.27 0.26 0.21 0.19 0.16
T (K) 4140
4120
4100
4080
4060
Wavenumber / cm-1 Fig. 2. Same as in Fig. 1. Medium hydrogen dose on K-ZSM-5 (Table 2).
case of hydrogen adsorbed on Na-ZSM-5, previously reported values are 4099 [11] and 4110 cm1 [10]; the former coincides (within experimental error) with the present value of 4101 cm1. For K-ZSM-5 no previous data seem to be available. However, a smaller bathochromic shift (referred to the gas-phase value of 4163 cm1) should be expected for K-ZSM-5, as compared to Na-ZSM-5. This is so because of the smaller electric field at the cation site in the case of the potassium-exchanged zeolite [17]. It should be noted that, in some cases, an ortho–para splitting of the H–H band (by about 6 cm1) has been observed for dihydrogen adsorbed on some zeolites [11,18]. No such splitting is clearly seen in the spectra here reported; however, those in Fig. 2 do seem to suggest a composite nature of the IR absorption band. From the integrated intensity of the H–H stretching band for spectra taken at a variable temperature, thermodynamic characterisation of the H2/Na-ZSM-5 and H2/K-ZSM-5 systems was carried out following a method described in detail elsewhere [4,7]. In essence, band intensity, A, temperature, T, and pressure, p, are considered to be interrelated by the Langmuir-type equation, A=AM ¼ h ¼ kðT Þp=½1 þ kðT Þp
ð1Þ
where AM stands for the integrated intensity corresponding to full coverage and h is the fractional coverage of adsorption sites by dihydrogen molecules. Combination of Eq. (1) with the well-known vant Hoff
P (· 0.75, mbar)
hydrogen dose 15.80 16.20 16.60 16.98 17.35 17.70 18.05 18.25 18.45 18.65 18.80 18.90 19.00 19.10 19.15 19.30
Large hydrogen dose 100 40.95 102 41.65 104 42.30 106 42.70 108 43.25 113 43.80 115 44.25 117 44.45 119 44.75 121 44.95 123 45.25 126 45.65 127 45.85 129 46.10 132 46.35 135 46.65
C. Otero Area´n et al. / Microporous and Mesoporous Materials 80 (2005) 247–252
250
Table 2 Numerical data corresponding to three measurement runs on K-ZSM5 P (· 0.75, mbar)
Absorbance (cm1)
-1
Coverage (h) -2
Small hydrogen dose 98 16.60 100 16.95 102 17.25 104 17.45 106 17.65 108 17.85 110 17.95 112 18.00 113 18.10 115 18.15 116 18.30
0.82 0.67 0.53 0.45 0.39 0.32 0.29 0.26 0.22 0.21 0.18
0.29 0.23 0.19 0.16 0.14 0.11 0.10 0.09 0.08 0.07 0.06
Medium hydrogen dose 96 43.35 97 43.80 99 44.25 99 44.70 101 45.10 103 45.40 105 46.10 107 46.40 108 46.65 110 46.90 112 47.20 113 47.45 115 47.70 116 47.95 118 48.15 119 48.35 121 48.55 123 48.78 124 48.95 125 49.15 126 49.30
1.69 1.59 1.48 1.38 1.24 1.16 0.98 0.87 0.82 0.76 0.67 0.61 0.57 0.51 0.45 0.41 0.37 0.32 0.29 0.27 0.24
0.59 0.56 0.52 0.49 0.44 0.41 0.34 0.31 0.29 0.27 0.24 0.21 0.20 0.18 0.16 0.14 0.13 0.11 0.10 0.09 0.08
Large hydrogen dose 97 133.75 99 134.75 100 135.65 103 136.55 105 137.45 106 138.35 107 139.20 111 139.85 113 140.50 115 141.40 117 142.50 119 143.20 121 143.85 123 144.40 126 145.90 129 146.90 130 147.50 132 147.85
2.23 2.07 1.96 1.84 1.71 1.58 1.48 1.35 1.27 1.15 1.02 0.94 0.84 0.77 0.60 0.50 0.47 0.44
0.78 0.73 0.69 0.65 0.60 0.56 0.52 0.47 0.45 0.40 0.36 0.33 0.29 0.27 0.21 0.18 0.17 0.15
to increasing hydrogen doses (equilibrium pressure) at 77 K and extrapolating the resulting Langmuir isotherm. This gave AM = 3.64 for H2/Na-ZSM-5 and AM = 3.03 for H2/K-ZSM-5. These figures were refined by plotting the left-hand side of Eq. (3) against recipro-
ln [θ/(1- θ)p]
T (K)
0
Na-ZSM-5 Large dose Medium dose Small dose K-ZSM-5 Large dose Medium dose Small dose
Na-ZSM-5 2
R = 0.992 SD = 0.126
-3 -4
K-ZSM-5 2
R = 0.989 SD = 0.091
-5 -6 -7 6.5
7.0
7.5
8.0
8.5
9.0
9.5 10.0 10.5 11.0 11.5
1000 / T Fig. 3. Plot of the left-hand side of Eq. (3) versus reciprocal temperature for both H2/Na-ZSM-5 and H2/K-ZSM-5. Different symbols correspond to independent measurement runs for each system (Tables 1 and 2). R, linear regression coefficient; SD, standard deviation.
cal temperature for AM values, changed in successive steps of 0.05 cm1, covering the range ±0.5 cm1 around the corresponding original value. The best fit to the whole set of experimental results was found at AM = 3.50 for H2/Na-ZSM-5 and AM = 2.85 for H2/KZSM-5. Fig. 3 shows that these AM values lead to excellent linear fits of Eq. (3) for both systems studied. From the linear plots in Fig. 3, the standard enthalpy of adsorption results to be DH = 10.3 kJ mol1 for H2/Na-ZSM-5 and DH = 9.1 kJ mol1 for H2/KZSM-5. The corresponding entropy change is 121 and 124 J mol1 K1 for the sodium and potassiumexchanged zeolite, respectively. The estimated error limits are of about ±0.5 kJ mol1 for enthalpy and ±10 J mol1 K1 for entropy. The liquefaction enthalpy of dihydrogen (at 20.45 K) is 0.90 kJ mol1 [19]. By comparing this value with those of DH reported above, it is seen that the energy involved in the adsorption process is significantly larger (for both zeolite samples) than that corresponding to liquefaction, and this is a favourable feature regarding the potential use of the adsorbents for cryogenic hydrogen storage. The interaction between the bare Na+ ion and a hydrogen molecule has been the object of many theoretical calculations [20–23], and there is agreement on the general aspects of the interaction. The dihydrogen molecule approaches the sodium ion side on, to maximize the charge–quadrupole interaction, and the resulting (weak) bond has an energy of about 10 kJ mol1. Actual values range from 8.03 to 10.25 kJ mol1, depending on the relative importance given to electron correlation and basis set superposition error. For the K+/H2 system, Bushnell et al. [23] give an approximate value of 9.2 kJ mol1. These latter authors also report experimental results [23] derived from temperature-dependent
C. Otero Area´n et al. / Microporous and Mesoporous Materials 80 (2005) 247–252
equilibrium measurements of mass-selected (using a double-focusing mass spectrometer) Na+ and K+ ions with hydrogen gas. The values thus obtained for the enthalpy change (DH) in the process M+ + H2 = M+ Æ H2 (M = Na, K) were 10.25 and 6.07 kJ mol1 for Na+ and K+, respectively. These values are close to those found in the present work for the systems H2/NaZSM-5 (DH = 10.3 kJ mol1) and H2/K-ZSM-5 (DH = 9.1 kJ mol1). However, it was already stated that in the zeolite the adsorption site is very likely to involve the cation and a nearby anion [10,16,24]; so that interaction of the adsorbed hydrogen molecule with the anion can compensate (till some extent) for the slightly smaller (net) electric charge at the cation centre. The higher (absolute) value of DH found for the H2/ Na-ZSM-5 system, as compared to H2/K-ZSM-5, can be ascribed to the fact that the (positive) electrostatic field, E, at the cation site is higher in the former case; reported values of E are 6.3 and 4.1 V nm1 for Na-ZSM-5 and K-ZSM-5, respectively [17]. Since charge–quadrupole and charge-induced dipole electrostatic interactions are expected to dominate bonding, a higher electric field should correlate with a higher DH value. It should be noted, however, that this explanation contrasts with the recently reported [6] experimental result of DH = 6.5 (±0.5) kJ mol1 for hydrogen adsorbed on Li-ZSM-5. For this system, the H–H stretching band was found at 4092 cm1; the larger bathochromic shift (referred to the 4163 cm1 value for free H2) seems to reflect the higher electrostatic field at the Li+ adsorbing centre, for which the value of E = 9.5 V nm1 was reported [17,25]. Hence, the adsorption enthalpy for the H2/Li-ZSM-5 system would be expected to be higher than for H2/Na-ZSM-5, which is not the case. While the exact reason for this discrepancy is not known, the comparatively small value of DH for hydrogen adsorption on Li-ZSM-5 could be related to the small size of the Li+ ion, which has an ionic radius of 74 pm, as compared to 102 and 133 pm for Na+ and K+, respectively [26]. Because of its small size, the Li+ ion might remain sunk among neighbouring oxygen anions when the system is not perturbed. Interaction with an adsorbed molecule could bring about a small movement of the cation in order to maximise such an interaction. This step is necessarily endothermic, so that the overall adsorption enthalpy can turn out to be anomalously low. Several reasons lend support to this tentative explanation. In the first place, Bolis et al. [27] found similar phenomena when studying the adsorption of carbon monoxide on calcium containing aluminas. Secondly, a slight movement of extra-framework cations upon adsorption of gases in zeolites was suggested several times in the literature. In particular, single crystal X-ray diffraction work has shown [28] that the Ca2+ ion in calcium-exchanged faujasites moves by about 25 pm (away from the zeolite framework) upon adsorption of benzene. Finally, a
251
similar endothermic step was recently documented [29] for the adsorption of CO on Li-ZSM-5, and also for the adsorption of acetonitrile and pyridine on boronsubstituted silica [30]. Regarding the entropy change, the values here reported are DS = 121 J mol1 K1 and 124 J mol1 K1 for H2/Na-ZSM-5 and H2/K-ZSM-5, respectively. These values can be considered to be both equal, within experimental error. The absolute entropy of dihydrogen (at the standard pressure and temperature of 1.33 mbar and 100 K) is 163 J mol1 K1 [6]. Hence, the values of DS obtained for hydrogen adsorbed on both Na-ZSM-5 and K-ZSM-5 lead to a standard entropy of the adsorbed phase of about 40 J mol1 K1. Qualitatively, this result suggests a substantial residual freedom of the H2 molecule in the adsorbed state, which probably reflects transformation of translational degrees into low-lying vibrational modes with preservation (at least to some extent) of rotational freedom. This is in agreement with a report [15] on dihydrogen adsorbed at 77 K on alkali-metal-exchanged faujasites, in which very weak (and broad) IR absorption bands observed above 4300 cm1 were assigned to a combination mode of the stretching vibration with rotation of the adsorbed molecules.
Acknowledgments This work was supported by the Spanish MCyT (Project No. MAT2002-03603) and FEDER funds.
References [1] A.A. Strub, G. Imarisio, Hydrogen as an Energy Vector, D. Reidel, Dordrecht, 1980. [2] J. Weitkamp, M. Fritz, S. Ernst, Int. J. Hydrogen Energy 20 (1995) 967. [3] A.W.C. van den Berg, S.T. Bromley, E. Flikkema, J. Wojdel, Th. Maschmeyer, J.C. Jansen, J. Chem. Phys. 120 (2004) 10285. [4] C. Otero Area´n, O.V. Manoilova, G. Turnes Palomino, M. Rodrı´guez Delgado, A.A. Tsyganenko, B. Bonelli, E. Garrone, Phys. Chem. Chem. Phys. 4 (2002) 5713. [5] E. Garrone, B. Bonelli, A.A. Tsyganenko, M. Rodrı´guez Delgado, G. Turnes Palomino, O.V. Manoilova, C. Otero Area´n, J. Phys. Chem. B 107 (2003) 2537. [6] C. Otero Area´n, O.V. Manoilova, B. Bonelli, M. Rodrı´guez Delgado, G. Turnes Palomino, E. Garrone, Chem. Phys. Lett. 370 (2003) 631. [7] C. Otero Area´n, O.V. Manoilova, A.A. Tsyganenko, G. Turnes Palomino, M. Pen˜arroya Mentruit, F. Geobaldo, E. Garrone, Eur. J. Inorg. Chem. 12 (2001) 1739. [8] H. Fo¨rster, M. Schuldt, J. Mol. Struct. 61 (1980) 361. [9] H. Bo¨se, H. Fo¨rster, W. Frede, Chem. Phys. Lett. 138 (1987) 401. [10] L.M. Kustov, V.B. Kazansky, J. Chem. Soc. Faraday Trans. 87 (1991) 2675. [11] K. Beck, H. Pfeifer, B. Staudte, J. Chem. Soc. Faraday Trans. 89 (1993) 3995.
252
C. Otero Area´n et al. / Microporous and Mesoporous Materials 80 (2005) 247–252
[12] S. Bordiga, E. Garrone, C. Lamberti, A. Zecchina, C. Otero Area´n, V.B. Kazansky, L.M. Kustov, J. Chem. Soc. Faraday Trans. 90 (1994) 3367. [13] V.B. Kazansky, V.Yu. Borovkov, H.G. Karge, J. Chem. Soc. Faraday Trans. 93 (1997) 1843. [14] H. Fo¨rster, W. Freude, G. Peters, in: H.G. Karge, J. Weitkamp (Eds.), Zeolites as Catalysts, Sorbents and Detergent Builders, Elsevier, Amsterdam, 1989. [15] V.B. Kazansky, J. Mol. Catal. A 141 (1999) 83. [16] E. Garrone, V.B. Kazansky, L.M. Kustov, J. Sauer, I.N. Senchenya, P. Ugliengo, J. Phys. Chem. 96 (1992) 1040. [17] C. Otero Area´n, Comments Inorg. Chem. 22 (2000) 241. [18] A. Zecchina, C. Otero Area´n, G. Turnes Palomino, F. Geobaldo, C. Lamberti, G. Spoto, S. Bordiga, Phys. Chem. Chem. Phys. 1 (1999) 1649. [19] R.H. Perry, D.W. Green, Perrys Chemical Engineers Handbook, McGraw-Hill, Sydney, 1998. [20] L.A. Curtis, J.A. Pople, J. Phys. Chem. 92 (1988) 894. [21] D.A. Dixon, J.L. Gole, A. Kormornicki, J. Phys. Chem. 92 (1988) 1378.
[22] M.F. Falcetta, J.L. Pazun, M.J. Dorko, D. Kitchen, P.E. Siska, J. Phys. Chem. 97 (1993) 1011. [23] J.E. Bushnell, P.R. Kemper, M.T. Bowers, J. Phys. Chem. 98 (1994) 2044. [24] A. Zecchina, C. Otero Area´n, Chem. Soc. Rev. 25 (1996) 187. [25] C. Otero Area´n, O.V. Manoilova, M. Rodrı´guez Delgado, A.A. Tsyganenko, E. Garrone, Phys. Chem. Chem. Phys. 3 (2001) 4187. [26] R.D. Shannon, C.T. Prewitt, Acta Cryst. B 25 (1969) 925. [27] V. Bolis, G. Magnacca, C. Morterra, Res. Chem. Intermed. 25 (1999) 25. [28] Y.H. Yeon, A.N. Kim, Y. Kim, S.H. Song, K. Seff, J. Phys. Chem. B 102 (1998) 6071. [29] B. Bonelli, E. Garrone, B. Fubini, B. Onida, M. Rodrı´guez Delgado, C. Otero Area´n, Phys. Chem. Chem. Phys. 5 (2003) 2900. [30] A. Travert, A. Vimont, J.-C. Lavalley, V. Montouillout, M. Rodrı´guez Delgado, J.J. Cuart Pascual, C. Otero Area´n, J. Phys. Chem. B 108 (2004) 16499.
Thermodynamic studies on hydrogen adsorption on the zeolites Na-ZSM-5 and K-ZSM-5 C. Otero Area´n a,*, M. Rodrı´guez Delgado a, G. Turnes Palomino a, M. Toma´s Rubio a, N.M. Tsyganenko b, A.A. Tsyganenko b, E. Garrone
c
a
c
Departamento de Quı´mica, Universidad de las Islas Baleares, E-07122 Palma de Mallorca, Spain b Institute of Physics, St. Petersburg University, 198504 St. Petersburg, Russia Dipartimento di Scienza dei Materiali ed Ingegneria Chimica, Politecnico di Torino, I-10126 Torino, Italy Received 17 September 2004; accepted 6 December 2004 Available online 26 January 2005
Abstract Adsorption of dihydrogen onto the zeolites Na-ZSM-5 and K-ZSM-5 renders the fundamental H–H stretching mode infrared active. The corresponding infrared absorption bands were found at 4101 and 4112 cm1 for H2/Na-ZSM-5 and H2/K-ZSM-5, respectively. Thermodynamic characterization of the adsorbed state was carried out by means of variable-temperature infrared spectroscopy; simultaneously measuring integrated band intensity, temperature and equilibrium pressure of the gas phase. For the H2/ Na-ZSM-5 system, the standard adsorption enthalpy and entropy resulted to be DH = 10.3 (±0.5) kJ mol1 and DS = 121 (±10) J mol1 K1. For H2/K-ZSM-5 corresponding values were 9.1 (±0.5) kJ mol1 and 124 (±10) J mol1 K1, respectively. 2004 Elsevier Inc. All rights reserved. Keywords: Adsorption thermodynamics; Hydrogen; FTIR; Zeolites; ZSM-5
1. Introduction Zeolite molecular sieves are often used for industrial gas separation processes based on differences in adsorption strength among components of a gas mixture. Thus, pressure swing adsorption (employing zeolites) is commercially used, inter alia, for producing oxygen from air and hydrogen from a variety of feedstocks; such as refinery off-gas and hydrogen-rich gas from steam reforming of hydrocarbons. On the other hand, both zeolites and active carbons can be considered [1–3] as being potential candidates for large-scale hydrogen storage in cryogenically cooled vessels; a subject of current interest
*
Corresponding author.
1387-1811/$ - see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.micromeso.2004.12.004
in view of the potential use of this gas as a cost-effective energy vector for buildings and transportation. Regarding hydrogen, it should be clear that precise knowledge of the interaction energy between the gas and the adsorbent would facilitate improvement of the above mentioned applications. However, direct measurement of such an interaction energy by adsorption microcalorimetry is a difficult task to perform. This is so because the low temperature needed for adsorption to take place, and the small amount of heat evolved, demand very stringent requirements on instrument design and sensitivity. We have recently shown [4,5] that variable-temperature infrared (IR) spectroscopy can advantageously be used for studying adsorption thermodynamics of weakly interacting adsorbent–adsorbate systems, and the method is applied here to hydrogen adsorption on both
C. Otero Area´n et al. / Microporous and Mesoporous Materials 80 (2005) 247–252
Na-ZSM-5 and K-ZSM-5. Related work on Li-ZSM-5 was reported elsewhere [6]. Interaction of the dihydrogen molecule with the zeolite adsorbing centres renders the fundamental H–H stretching mode IR active. By measuring the integrated intensity of the corresponding IR absorption band over a range of temperature, and the corresponding gas equilibrium pressure, the desired values of adsorption enthalpy and entropy are obtained. These values fully characterise the adsorbed state, as shown below.
2. Experimental details The Na-ZSM-5 and K-ZSM-5 samples used were prepared from a commercial NH4-ZSM-5 sample (Si/ Al = 25) by ion exchange with the corresponding alkali-metal nitrate solutions. Powder X-ray diffraction showed good crystallinity in both cases, and absence of any diffraction line not assignable to the corresponding MFI structure type. Complete ion exchange was checked by the absence of IR absorption bands corresponding to either the ammonium ion or the Brønstedacid OH group, which would be formed during thermal activation (see below) of the zeolite samples if total exchange of alkali-metal ions for ammonium did not take place. For IR spectroscopic measurements, thin selfsupported wafers of the zeolite samples were prepared and activated (outgassed) in a dynamic vacuum (residual pressure < 107 bar) for 2 h at 700 K inside an IR cell [7], which allowed in situ high-temperature activation, gas dosage and variable-temperature IR spectroscopy to be carried out. For each sample, after running the blank spectrum of the zeolite wafer at 77 K the cell was dosed with hydrogen, it was then closed and spectra were taken at 2–5 K intervals while simultaneously recording sample temperature and hydrogen equilibrium pressure inside the cell. A platinum resistance thermometer and a capacitance pressure gauge were used for this purpose. The precision of these measurements was of about ±2 K and ±2 · 105 bar, respectively. In order to check reproducibility, and also for improving accuracy, three independent sets of measurements were taken on each sample. In these independent runs different initial doses of hydrogen were used, so as to cover a larger pressure range. Note that the high thermal conductivity of hydrogen, together with the static nature of measurements, should minimize (possible) temperature gradients inside both the IR cell and the sample wafer. Since no temperature offset effects were found when plotting results from data obtained under widely different hydrogen pressure (in the different runs for each sample), it seems reasonable to assume that temperature gradients (if any) were of little concern. Transmission FTIR spectra were recorded at 3 cm1 resolution by means of a Bruker IFS66 spectrometer.
3. Results and discussion IR spectra of dihydrogen adsorbed (at 77 K) on alkali-metal-exchanged faujasites, mordenite, and A-type zeolites were reported by several authors [8–13]; mainly with a view of using dihydrogen as a probe molecule for IR spectroscopic studies of zeolites and related materials. These spectra show a main IR absorption band in the 4070–4100 cm1 range which corresponds to the fundamental H–H stretching mode of adsorbed dihydrogen. Perturbation of the molecule (mainly) by the cationic adsorption centre renders this vibrational mode IR active and causes a bathochromic shift from the gas phase value of 4163 cm1 corresponding to the free molecule (Raman active H–H stretching vibration). Earlier studies [8,14] suggested that the H2 molecule is adsorbed on alkali-metal cations of zeolites forming T-shaped adsorption complexes. However, more recent work [10,12,15] has shown that the adsorbed molecule can also interact with neighbouring anions of the zeolite framework. Hence, the adsorption centre should be taken to represent both the extra-framework alkali-metal cation and the nearby oxygens [16]. Fig. 1 shows some selected variable-temperature IR spectra, in the H–H stretching region, of hydrogen adsorbed on Na-ZSM-5; the H–H stretching band is seen at 4101 cm1. For hydrogen adsorbed on K-ZSM-5, corresponding spectra are depicted in Fig. 2; which shows the H–H stretching band at 4112 cm1. For the
0.18 100
0.16
Absorbance
248
102
0.14
104
0.12
108 113
0.10
115
0.08
119 123
0.06
126
0.04
129
0.02 0.00 4140
4120
4100
4080
Wavenumber / cm-1
4060
Fig. 1. Selected difference IR spectra (zeolite blank subtracted) in the H–H stretching region corresponding to the large hydrogen dose on Na-ZSM-5 (Table 1). Numbers denote temperature in K.
C. Otero Area´n et al. / Microporous and Mesoporous Materials 80 (2005) 247–252
96
Eq. (2), leads to Eq. (3) below, from which the standard enthalpy and entropy changes (involved in the adsorption process) were obtained:
97
kðT Þ ¼ exp½DS =R exp½DH =RT
ð2Þ
ln½h=ð1 hÞp ¼ ðDH =RT Þ þ ðDS =RÞ
ð3Þ
0.12
0.10
99 100
0.08
Values of temperature, pressure and absorbance (integrated band intensity) corresponding to the whole set of IR spectra taken in three independent runs on the H2/Na-ZSM-5 and H2/K-ZSM-5 systems are given in Tables 1 and 2, respectively. In order to convert absorbance into the corresponding h value, knowledge of AM is needed. For each system, an approximate value of AM was obtained by running spectra corresponding
Absorbance
101 104
0.06
249
107 110
0.04
115 119
0.02
Table 1 Numerical data corresponding to three measurement runs on NaZSM-5
0.00
Absorbance (cm1)
Coverage (h)
Small hydrogen dose 89 3.78 91 4.20 95 4.88 107 5.76 121 6.58 132 7.27
2.52 1.92 1.61 0.69 0.31 0.14
0.72 0.55 0.46 0.20 0.09 0.04
Medium 96 99 101 104 107 112 116 120 123 125 128 130 132 134 137 140
2.67 2.42 2.13 1.84 1.58 1.26 0.99 0.83 0.68 0.53 0.47 0.38 0.33 0.26 0.26 0.19
0.76 0.69 0.61 0.53 0.45 0.36 0.28 0.24 0.20 0.15 0.13 0.11 0.09 0.08 0.07 0.05
2.95 2.74 2.49 2.33 2.10 1.86 1.65 1.51 1.41 1.26 1.13 0.94 0.91 0.75 0.66 0.56
0.84 0.78 0.71 0.66 0.60 0.53 0.47 0.43 0.40 0.36 0.32 0.27 0.26 0.21 0.19 0.16
T (K) 4140
4120
4100
4080
4060
Wavenumber / cm-1 Fig. 2. Same as in Fig. 1. Medium hydrogen dose on K-ZSM-5 (Table 2).
case of hydrogen adsorbed on Na-ZSM-5, previously reported values are 4099 [11] and 4110 cm1 [10]; the former coincides (within experimental error) with the present value of 4101 cm1. For K-ZSM-5 no previous data seem to be available. However, a smaller bathochromic shift (referred to the gas-phase value of 4163 cm1) should be expected for K-ZSM-5, as compared to Na-ZSM-5. This is so because of the smaller electric field at the cation site in the case of the potassium-exchanged zeolite [17]. It should be noted that, in some cases, an ortho–para splitting of the H–H band (by about 6 cm1) has been observed for dihydrogen adsorbed on some zeolites [11,18]. No such splitting is clearly seen in the spectra here reported; however, those in Fig. 2 do seem to suggest a composite nature of the IR absorption band. From the integrated intensity of the H–H stretching band for spectra taken at a variable temperature, thermodynamic characterisation of the H2/Na-ZSM-5 and H2/K-ZSM-5 systems was carried out following a method described in detail elsewhere [4,7]. In essence, band intensity, A, temperature, T, and pressure, p, are considered to be interrelated by the Langmuir-type equation, A=AM ¼ h ¼ kðT Þp=½1 þ kðT Þp
ð1Þ
where AM stands for the integrated intensity corresponding to full coverage and h is the fractional coverage of adsorption sites by dihydrogen molecules. Combination of Eq. (1) with the well-known vant Hoff
P (· 0.75, mbar)
hydrogen dose 15.80 16.20 16.60 16.98 17.35 17.70 18.05 18.25 18.45 18.65 18.80 18.90 19.00 19.10 19.15 19.30
Large hydrogen dose 100 40.95 102 41.65 104 42.30 106 42.70 108 43.25 113 43.80 115 44.25 117 44.45 119 44.75 121 44.95 123 45.25 126 45.65 127 45.85 129 46.10 132 46.35 135 46.65
C. Otero Area´n et al. / Microporous and Mesoporous Materials 80 (2005) 247–252
250
Table 2 Numerical data corresponding to three measurement runs on K-ZSM5 P (· 0.75, mbar)
Absorbance (cm1)
-1
Coverage (h) -2
Small hydrogen dose 98 16.60 100 16.95 102 17.25 104 17.45 106 17.65 108 17.85 110 17.95 112 18.00 113 18.10 115 18.15 116 18.30
0.82 0.67 0.53 0.45 0.39 0.32 0.29 0.26 0.22 0.21 0.18
0.29 0.23 0.19 0.16 0.14 0.11 0.10 0.09 0.08 0.07 0.06
Medium hydrogen dose 96 43.35 97 43.80 99 44.25 99 44.70 101 45.10 103 45.40 105 46.10 107 46.40 108 46.65 110 46.90 112 47.20 113 47.45 115 47.70 116 47.95 118 48.15 119 48.35 121 48.55 123 48.78 124 48.95 125 49.15 126 49.30
1.69 1.59 1.48 1.38 1.24 1.16 0.98 0.87 0.82 0.76 0.67 0.61 0.57 0.51 0.45 0.41 0.37 0.32 0.29 0.27 0.24
0.59 0.56 0.52 0.49 0.44 0.41 0.34 0.31 0.29 0.27 0.24 0.21 0.20 0.18 0.16 0.14 0.13 0.11 0.10 0.09 0.08
Large hydrogen dose 97 133.75 99 134.75 100 135.65 103 136.55 105 137.45 106 138.35 107 139.20 111 139.85 113 140.50 115 141.40 117 142.50 119 143.20 121 143.85 123 144.40 126 145.90 129 146.90 130 147.50 132 147.85
2.23 2.07 1.96 1.84 1.71 1.58 1.48 1.35 1.27 1.15 1.02 0.94 0.84 0.77 0.60 0.50 0.47 0.44
0.78 0.73 0.69 0.65 0.60 0.56 0.52 0.47 0.45 0.40 0.36 0.33 0.29 0.27 0.21 0.18 0.17 0.15
to increasing hydrogen doses (equilibrium pressure) at 77 K and extrapolating the resulting Langmuir isotherm. This gave AM = 3.64 for H2/Na-ZSM-5 and AM = 3.03 for H2/K-ZSM-5. These figures were refined by plotting the left-hand side of Eq. (3) against recipro-
ln [θ/(1- θ)p]
T (K)
0
Na-ZSM-5 Large dose Medium dose Small dose K-ZSM-5 Large dose Medium dose Small dose
Na-ZSM-5 2
R = 0.992 SD = 0.126
-3 -4
K-ZSM-5 2
R = 0.989 SD = 0.091
-5 -6 -7 6.5
7.0
7.5
8.0
8.5
9.0
9.5 10.0 10.5 11.0 11.5
1000 / T Fig. 3. Plot of the left-hand side of Eq. (3) versus reciprocal temperature for both H2/Na-ZSM-5 and H2/K-ZSM-5. Different symbols correspond to independent measurement runs for each system (Tables 1 and 2). R, linear regression coefficient; SD, standard deviation.
cal temperature for AM values, changed in successive steps of 0.05 cm1, covering the range ±0.5 cm1 around the corresponding original value. The best fit to the whole set of experimental results was found at AM = 3.50 for H2/Na-ZSM-5 and AM = 2.85 for H2/KZSM-5. Fig. 3 shows that these AM values lead to excellent linear fits of Eq. (3) for both systems studied. From the linear plots in Fig. 3, the standard enthalpy of adsorption results to be DH = 10.3 kJ mol1 for H2/Na-ZSM-5 and DH = 9.1 kJ mol1 for H2/KZSM-5. The corresponding entropy change is 121 and 124 J mol1 K1 for the sodium and potassiumexchanged zeolite, respectively. The estimated error limits are of about ±0.5 kJ mol1 for enthalpy and ±10 J mol1 K1 for entropy. The liquefaction enthalpy of dihydrogen (at 20.45 K) is 0.90 kJ mol1 [19]. By comparing this value with those of DH reported above, it is seen that the energy involved in the adsorption process is significantly larger (for both zeolite samples) than that corresponding to liquefaction, and this is a favourable feature regarding the potential use of the adsorbents for cryogenic hydrogen storage. The interaction between the bare Na+ ion and a hydrogen molecule has been the object of many theoretical calculations [20–23], and there is agreement on the general aspects of the interaction. The dihydrogen molecule approaches the sodium ion side on, to maximize the charge–quadrupole interaction, and the resulting (weak) bond has an energy of about 10 kJ mol1. Actual values range from 8.03 to 10.25 kJ mol1, depending on the relative importance given to electron correlation and basis set superposition error. For the K+/H2 system, Bushnell et al. [23] give an approximate value of 9.2 kJ mol1. These latter authors also report experimental results [23] derived from temperature-dependent
C. Otero Area´n et al. / Microporous and Mesoporous Materials 80 (2005) 247–252
equilibrium measurements of mass-selected (using a double-focusing mass spectrometer) Na+ and K+ ions with hydrogen gas. The values thus obtained for the enthalpy change (DH) in the process M+ + H2 = M+ Æ H2 (M = Na, K) were 10.25 and 6.07 kJ mol1 for Na+ and K+, respectively. These values are close to those found in the present work for the systems H2/NaZSM-5 (DH = 10.3 kJ mol1) and H2/K-ZSM-5 (DH = 9.1 kJ mol1). However, it was already stated that in the zeolite the adsorption site is very likely to involve the cation and a nearby anion [10,16,24]; so that interaction of the adsorbed hydrogen molecule with the anion can compensate (till some extent) for the slightly smaller (net) electric charge at the cation centre. The higher (absolute) value of DH found for the H2/ Na-ZSM-5 system, as compared to H2/K-ZSM-5, can be ascribed to the fact that the (positive) electrostatic field, E, at the cation site is higher in the former case; reported values of E are 6.3 and 4.1 V nm1 for Na-ZSM-5 and K-ZSM-5, respectively [17]. Since charge–quadrupole and charge-induced dipole electrostatic interactions are expected to dominate bonding, a higher electric field should correlate with a higher DH value. It should be noted, however, that this explanation contrasts with the recently reported [6] experimental result of DH = 6.5 (±0.5) kJ mol1 for hydrogen adsorbed on Li-ZSM-5. For this system, the H–H stretching band was found at 4092 cm1; the larger bathochromic shift (referred to the 4163 cm1 value for free H2) seems to reflect the higher electrostatic field at the Li+ adsorbing centre, for which the value of E = 9.5 V nm1 was reported [17,25]. Hence, the adsorption enthalpy for the H2/Li-ZSM-5 system would be expected to be higher than for H2/Na-ZSM-5, which is not the case. While the exact reason for this discrepancy is not known, the comparatively small value of DH for hydrogen adsorption on Li-ZSM-5 could be related to the small size of the Li+ ion, which has an ionic radius of 74 pm, as compared to 102 and 133 pm for Na+ and K+, respectively [26]. Because of its small size, the Li+ ion might remain sunk among neighbouring oxygen anions when the system is not perturbed. Interaction with an adsorbed molecule could bring about a small movement of the cation in order to maximise such an interaction. This step is necessarily endothermic, so that the overall adsorption enthalpy can turn out to be anomalously low. Several reasons lend support to this tentative explanation. In the first place, Bolis et al. [27] found similar phenomena when studying the adsorption of carbon monoxide on calcium containing aluminas. Secondly, a slight movement of extra-framework cations upon adsorption of gases in zeolites was suggested several times in the literature. In particular, single crystal X-ray diffraction work has shown [28] that the Ca2+ ion in calcium-exchanged faujasites moves by about 25 pm (away from the zeolite framework) upon adsorption of benzene. Finally, a
251
similar endothermic step was recently documented [29] for the adsorption of CO on Li-ZSM-5, and also for the adsorption of acetonitrile and pyridine on boronsubstituted silica [30]. Regarding the entropy change, the values here reported are DS = 121 J mol1 K1 and 124 J mol1 K1 for H2/Na-ZSM-5 and H2/K-ZSM-5, respectively. These values can be considered to be both equal, within experimental error. The absolute entropy of dihydrogen (at the standard pressure and temperature of 1.33 mbar and 100 K) is 163 J mol1 K1 [6]. Hence, the values of DS obtained for hydrogen adsorbed on both Na-ZSM-5 and K-ZSM-5 lead to a standard entropy of the adsorbed phase of about 40 J mol1 K1. Qualitatively, this result suggests a substantial residual freedom of the H2 molecule in the adsorbed state, which probably reflects transformation of translational degrees into low-lying vibrational modes with preservation (at least to some extent) of rotational freedom. This is in agreement with a report [15] on dihydrogen adsorbed at 77 K on alkali-metal-exchanged faujasites, in which very weak (and broad) IR absorption bands observed above 4300 cm1 were assigned to a combination mode of the stretching vibration with rotation of the adsorbed molecules.
Acknowledgments This work was supported by the Spanish MCyT (Project No. MAT2002-03603) and FEDER funds.
References [1] A.A. Strub, G. Imarisio, Hydrogen as an Energy Vector, D. Reidel, Dordrecht, 1980. [2] J. Weitkamp, M. Fritz, S. Ernst, Int. J. Hydrogen Energy 20 (1995) 967. [3] A.W.C. van den Berg, S.T. Bromley, E. Flikkema, J. Wojdel, Th. Maschmeyer, J.C. Jansen, J. Chem. Phys. 120 (2004) 10285. [4] C. Otero Area´n, O.V. Manoilova, G. Turnes Palomino, M. Rodrı´guez Delgado, A.A. Tsyganenko, B. Bonelli, E. Garrone, Phys. Chem. Chem. Phys. 4 (2002) 5713. [5] E. Garrone, B. Bonelli, A.A. Tsyganenko, M. Rodrı´guez Delgado, G. Turnes Palomino, O.V. Manoilova, C. Otero Area´n, J. Phys. Chem. B 107 (2003) 2537. [6] C. Otero Area´n, O.V. Manoilova, B. Bonelli, M. Rodrı´guez Delgado, G. Turnes Palomino, E. Garrone, Chem. Phys. Lett. 370 (2003) 631. [7] C. Otero Area´n, O.V. Manoilova, A.A. Tsyganenko, G. Turnes Palomino, M. Pen˜arroya Mentruit, F. Geobaldo, E. Garrone, Eur. J. Inorg. Chem. 12 (2001) 1739. [8] H. Fo¨rster, M. Schuldt, J. Mol. Struct. 61 (1980) 361. [9] H. Bo¨se, H. Fo¨rster, W. Frede, Chem. Phys. Lett. 138 (1987) 401. [10] L.M. Kustov, V.B. Kazansky, J. Chem. Soc. Faraday Trans. 87 (1991) 2675. [11] K. Beck, H. Pfeifer, B. Staudte, J. Chem. Soc. Faraday Trans. 89 (1993) 3995.
252
C. Otero Area´n et al. / Microporous and Mesoporous Materials 80 (2005) 247–252
[12] S. Bordiga, E. Garrone, C. Lamberti, A. Zecchina, C. Otero Area´n, V.B. Kazansky, L.M. Kustov, J. Chem. Soc. Faraday Trans. 90 (1994) 3367. [13] V.B. Kazansky, V.Yu. Borovkov, H.G. Karge, J. Chem. Soc. Faraday Trans. 93 (1997) 1843. [14] H. Fo¨rster, W. Freude, G. Peters, in: H.G. Karge, J. Weitkamp (Eds.), Zeolites as Catalysts, Sorbents and Detergent Builders, Elsevier, Amsterdam, 1989. [15] V.B. Kazansky, J. Mol. Catal. A 141 (1999) 83. [16] E. Garrone, V.B. Kazansky, L.M. Kustov, J. Sauer, I.N. Senchenya, P. Ugliengo, J. Phys. Chem. 96 (1992) 1040. [17] C. Otero Area´n, Comments Inorg. Chem. 22 (2000) 241. [18] A. Zecchina, C. Otero Area´n, G. Turnes Palomino, F. Geobaldo, C. Lamberti, G. Spoto, S. Bordiga, Phys. Chem. Chem. Phys. 1 (1999) 1649. [19] R.H. Perry, D.W. Green, Perrys Chemical Engineers Handbook, McGraw-Hill, Sydney, 1998. [20] L.A. Curtis, J.A. Pople, J. Phys. Chem. 92 (1988) 894. [21] D.A. Dixon, J.L. Gole, A. Kormornicki, J. Phys. Chem. 92 (1988) 1378.
[22] M.F. Falcetta, J.L. Pazun, M.J. Dorko, D. Kitchen, P.E. Siska, J. Phys. Chem. 97 (1993) 1011. [23] J.E. Bushnell, P.R. Kemper, M.T. Bowers, J. Phys. Chem. 98 (1994) 2044. [24] A. Zecchina, C. Otero Area´n, Chem. Soc. Rev. 25 (1996) 187. [25] C. Otero Area´n, O.V. Manoilova, M. Rodrı´guez Delgado, A.A. Tsyganenko, E. Garrone, Phys. Chem. Chem. Phys. 3 (2001) 4187. [26] R.D. Shannon, C.T. Prewitt, Acta Cryst. B 25 (1969) 925. [27] V. Bolis, G. Magnacca, C. Morterra, Res. Chem. Intermed. 25 (1999) 25. [28] Y.H. Yeon, A.N. Kim, Y. Kim, S.H. Song, K. Seff, J. Phys. Chem. B 102 (1998) 6071. [29] B. Bonelli, E. Garrone, B. Fubini, B. Onida, M. Rodrı´guez Delgado, C. Otero Area´n, Phys. Chem. Chem. Phys. 5 (2003) 2900. [30] A. Travert, A. Vimont, J.-C. Lavalley, V. Montouillout, M. Rodrı´guez Delgado, J.J. Cuart Pascual, C. Otero Area´n, J. Phys. Chem. B 108 (2004) 16499.