J. Phys. Chem. B 2005, 109, 22181-22185
22181
Decomposition Kinetics of the AlH3 Polymorphs Jason Graetz* and James J. Reilly Department of Energy Sciences and Technology, BrookhaVen National Laboratory, Upton, New York 11973 ReceiVed: August 19, 2005; In Final Form: October 1, 2005
Aluminum hydride polymorphs (R-AlH3, β-AlH3, and γ-AlH3) were prepared by organometallic synthesis. Hydrogen capacities approaching 10 wt % at desorption temperatures less than 100 °C have been demonstrated with freshly prepared AlH3. The temperature-dependent rate constants were determined by measuring the isothermal hydrogen evolution between 60 °C and 140 °C. Fractional decomposition curves showed good fits using both the second and third-order Avrami-Erofeyev equations, indicating that the decomposition kinetics are controlled by nucleation and growth of the aluminum phase in two and three dimensions. The large activation energies measured for the AlH3 polymorphs suggest that the decomposition occurs via an activated complex mechanism with complexes consisting of approximately nine AlH3 molecules (1-2 unit cells for R-AlH3).
I. Introduction Aluminum hydride, AlH3, is a fascinating material that has recently attracted attention for its potential as a hydrogen storage medium for low-temperature fuel cells. It has a volumetric hydrogen capacity (1.48 g/mL) greater than that of liquid hydrogen and a gravimetric hydrogen capacity exceeding 10 wt %. AlH3 is stable at room temperature despite having an equilibrium hydrogen pressure of between 1 and 10 kbar at 298 K.1,9 The stability is generally attributed to a surface oxide layer, which acts as a kinetic barrier to decomposition and protects the AlH3 from the environment. As a result of the high-activation barrier and the overall slow desorption kinetics measured for large cuboids of AlH3 prepared by the Dow Chemical Co.,2-4 this material was long thought to be unsuitable for lowtemperature systems. Recent studies by Sandrock et al. demonstrated that a dopant introduced by ball milling can alter this surface barrier and lead to decomposition kinetics at 100 °C that are adequate for use in H-powered vehicles. The conventional procedure for synthesizing a solvated form of AlH3, developed by Finholt et al., involves an ethereal reaction of LiAlH4 and AlCl3:5
AlCl3 + 3LiAlH4 + n[(C2H5)2O] f 4AlH3‚1.2[(C2H5)2O] + 3LiCl (1) A nonsolvated form of AlH3 was initially prepared by Chizinsky et al.6 and subsequently prepared by Brower et al.7 by heating AlH3 in the presence of a complex metal hydride. Early studies of nonsolvated AlH3 by Brower et al. at the Dow Chemical Co. identified seven polymorphs, namely, R, R′, β, δ, , γ, and ζ. However, subsequent studies by a number of different groups, which included a structural characterization,8 thermodynamic measurements,1,9 and thermal and photolytic kinetic studies,2-4,9-13 were all limited to the Dow material, consisting of 100 µm cuboids of the R polymorph. Consequently, little is known about the other polymorphs and their properties. In this study, three crystalline AlH3 polymorphs were prepared via organometallic * To whom correspondence should be addressed. E-mail address:
[email protected].
synthesis, and their kinetic properties were investigated. We demonstrate that, at a temperature 0.7). The data were plotted as ln[-ln(1 - R)] versus ln(t) (not shown) to measure the constant, n, and determine the most suitable kinetic equation. The average values of n were approximately 2, suggesting the second-order Avrami-Erofeyev equation. The acceleratory period was fit using eq 3 with n ) 3 and n ) 2, which showed similar results. Although both values are suitable, n ) 2 was chosen for this analysis because it provided a slightly better fit.
22184 J. Phys. Chem. B, Vol. 109, No. 47, 2005
Graetz and Reilly
TABLE 1: Isothermal Decomposition Rate Constants (s-1) for r-, β-, and γ-AlH3a
a
polymorph
k(60 °C)
k(80 °C)
k(99 °C)
k(120 °C)
k(138 °C)
R-AlH3 β-AlH3 γ-AlH3
1.35 × 10-6 4.41 × 10-6 3.97 × 10-6
9.14 × 10-6 1.36 × 10-5 1.18 × 10-5
4.21 × 10-5 5.99 × 10-5 3.94 × 10-5
2.80 × 10-4 4.88 × 10-4 2.71 × 10-4
1.39 × 10-3 2.46 × 10-3 7.98 × 10-4
The sample temperatures are typically within ( 1 °C of the value listed.
Figure 5. Temperature and fractional decomposition of R-AlH3 demonstrating that the reaction rate goes to zero as the sample temperature is reduced to ∼23 °C. The inset shows an expanded view of the decomposition curve and illustrates that the rates before and after the temperature change are equivalent.
The fraction reacted plotted for R-AlH3, β-AlH3, and γ-AlH3 as [-ln(1 - R)]1/2 are shown in Figures 2b, 3b, and 4b, respectively. These plots demonstrate good least-squares fits over the range of 0.04 eR e 0.95 with a linearity constant of R ≈ 0.99, with a few exceptions for the γ phase, which are discussed in more detail below. This supports the use of the Avrami-Erofeyev equation (eq 3). The rate constants determined from the slope of the least-squares fit are displayed in Table 1. Analysis of the isothermal decomposition curves gives consistent values of n ≈ 2. This value indicates that the kinetics are limited by random nucleation and growth. The same Avrami-Erofeyev equation applies to decomposition over a wide range in temperature (60-138 °C), suggesting a common decomposition pathway. This consistency is crucial for this analysis and to assign any physical meaning to the constant n and the corresponding kinetic function kt ) [-ln(1 - R)1/2]. The reaction rate can be lowered and even stopped by decreasing the sample temperature, as shown in Figure 5. Decreasing the sample temperature to ∼23 °C during the acceleratory period completely stops the evolution of H2. Upon subsequent heating, no induction period is observed, and the decomposition reaction rate returns directly to the rate prior to reducing T, as shown in the inset of Figure 5. Arrhenius plots for R-AlH3, β-AlH3, and γ-AlH3 over a temperature range of 60 e T e 138 °C are displayed in Figure 6 along with data from Herley et al. for Dow-synthesized R-AlH3.4 The activation energies and preexponential constants (A using eq 4 and σ using eq 6) were determined from Figure 6 and are shown in Table 2. Surface area measurements performed on the decomposed material (Al powder) and the calculated crystallite size based on isolated spherical particles are also shown in Table 2. The molar volume of R-AlH3 is approximately twice that of Al metal. Assuming this volumetric ratio applies to the other polymorphs, the particle diameters of the hydrides are approximately 27% larger than the values listed
Figure 6. Arrhenius plot for large crystallites of R-AlH3 (Dow)4 and small crystallites of R-AlH3, β-AlH3, and γ-AlH3. Reaction rates for the large crystallites of R-AlH3 were measured at 135 °C e T e 160 °C and are extrapolated down to T ∼ 60 °C.
TABLE 2: Morphological and Kinetic Values for the AlH3 Polymorphsa surface particle area diameter (nm) Ea (kJ/mol) polymorph (m2/g) R-AlH3 β-AlH3 γ-AlH3
10.9 14.6 16.4
204 152 136
A
σ
102.2 ( 3.2 1.2 × 1010 1.9 × 10-3 92.3 ( 8.6 8.8 × 108 1.4 × 10-4 79.3 ( 5.1 8.5 × 106 1.4 × 10-6
a Particle diameters were calculated from the surface area using a spherical geometry. The preexponential constants A and σ were determined from eqs 4 and 6, respectively.
in Table 2. XRD after isothermal analyses confirmed that, in each case, the final products were fcc aluminum powder. VI. Discussion The sigmoid shape of the fractional decomposition curve is indicative of an autocatalytic reaction and is typical of solidstate decomposition. The induction period is attributed to nucleation and “slow” growth.3 Since little decomposition accompanies the formation of a nucleation site, this region exhibits slow H2 evolution. This model is supported by decomposition experiments in which the reaction was stopped during the acceleratory period by reducing T (Figure 5). At this stage, many nucleation sites are present, but the growth is inhibited by the low sample temperature. Upon subsequent heating, the growth continues, and the reaction rate returns directly to the acceleratory period (no induction period). During the acceleratory stage of the reaction, the kinetics are controlled by random nucleation and rapid growth. The decomposition curves demonstrate the best fits with a geometric constant of n ) 2 or 3, suggesting that the growth of the Al phase occurs in two or three-dimensions. The average value of n is insensitive to the polymorph structure, indicating that a common decomposition mechanism exists for the three phases. It is interesting
Decomposition Kinetics of AlH3 Polymorphs to note that the kinetics are not limited by diffusion through a surface oxide as previously expected. Diffusion-controlled decomposition reactions typically yield values of 0.54 e n e 0.6227 and can be easily differentiated from nucleation and growth reactions. Finally, the decay period is attributed to the disappearance of the unreacted AlH3 phase. In most reactions, the value of the activation energy is independent of the enthalpy. However, in certain cases, such as evaporation and thermal decomposition, the activation energy can be related to the enthalpy through activated complex theory.26,27 The decomposition kinetics of AlH3 clearly do not obey the Polanyi-Wigner relation (equation 5), since the preexponential factors for the three polymorphs are more than 2 orders of magnitude lower than kBT/h at 298 K (kBT/h ) 6.2 × 1012), and the activation energies (Table 2) are much greater than the decomposition enthalpy measured for R-AlH3 (∆H ) 7.6 kJ/mol H21). However, Shannon27 and others28 demonstrated that activated complex theory can be used to predict the thermal decomposition rates of solids and other reactions that do not obey the Polanyi-Wigner relation. Therefore, the large activation energies measured for AlH3 may be attributed to a decomposition process involving activated complexes, rather than individual molecules. Upon the basis of the measured activation energy and the known dissociation enthalpy for R-AlH3, these complexes consist of approximately nine AlH3 molecules, or 1-2 unit cells. Although this is only one possible decomposition mechanism, it is reasonable to suggest that the conversion of R-AlH3 to Al occurs in increments of whole unit cells. Values of σ from eq 6 are listed in Table 2 (in which σ ) hA/kBT) and are in agreement with the expected values for a solid complex with only a few degrees of freedom (low mobility). The preexponential and activation energy for R-AlH3 (Table 2) are considerably smaller than the values measured by Herley et al. (A ) 3.5 × 1016 and Ea ) 150.3 ( 10.0 kJ/mol).4 The hydrides synthesized for this study demonstrate reaction rates that are an order of magnitude greater at 60 °C than those measured for the Dow material. This is attributed to a smaller particle size and a reduced surface oxide in the freshly prepared material. In addition to particle size and surface coatings, the rate constants are also sensitive to the polymorph structure as shown in Table 2 (the small variations in particle size are not believed to have a large affect on the kinetics). The decrease in activation energy of the β and γ phases with respect to that of the R phase is attributed, in part, to a smaller decomposition enthalpy. Preliminary calorimetry results suggest that R is the most stable, followed by β and ultimately the γ phase. This is also supported by the order in which these phases appear during the synthesis. γ- and β-AlH3 precipitate early and quickly decompose to the more-stable R phase. The transition to the R phase is exothermic, supplying an energetic boost to the decomposition. The total energy required is reduced, and therefore there is a larger driving force toward decomposition. This may also explain the odd shape of the decomposition curve for γ-AlH3 at 60 °C, which has an acceleratory region that exhibits rapid H2 evolution at short times (t < 80 × 103 s). It is likely that, during the region of rapid hydrogen evolution, the material is decomposing while simultaneously transforming to the R phase. When the phase transition is complete, the decomposition continues at a slightly slower rate because of the greater stability of R-AlH3.
J. Phys. Chem. B, Vol. 109, No. 47, 2005 22185 VII. Conclusion The kinetics of the aluminum hydride polymorphs (R-AlH3, β-AlH3, and γ-AlH3) are controlled by nucleation and growth in two and three dimensions and are not limited by H2 diffusion through a surface oxide. The decomposition of AlH3 occurs in complexes of approximately nine molecules, or 1-2 unit cells for R-AlH3. Decomposition of the R phase was slower than that of the γ and β phases because of its greater stability. In general, the rapid low-temperature kinetics and high-energy density make AlH3 an unusual and promising hydrogen storage medium for a number of applications. However, the conventional organometallic synthesis is a costly procedure, and AlH3 is not a reversible hydride at moderate H2 pressures. Incorporating dopants or catalytic additives is not likely to produce the large thermodynamic changes required to substantially reduce the equilibrium pressure. Therefore, the utility of this material will depend on the development of new techniques to regenerate AlH3 from the spent Al powder in a cost-effective and energetically efficient manner. Acknowledgment. The authors gratefully acknowledge Gary Sandrock for his insight on hydride kinetics and his encouragement to investigate aluminum hydride. This work was supported by the Department of Energy’s Office of Energy Efficiency and Renewable Energy. This manuscript was authored by Brookhaven Science Associates, LLC under Contract No. DE-AC02-98CH1886 with the U.S. Department of Energy. References and Notes (1) Sinke, G. C.; Walker, L. C.; Oetting, F. L.; Stull, D. R. J. Chem. Phys. 1967, 47, 2759. (2) Herley, P. J.; Irwin, R. H. J. Phys. Chem. Solids 1978, 39, 1013. (3) Herley, P. J.; Christofferson, O.; Irwin R. J. Phys. Chem. 1981, 85, 1874. (4) Herley, P. J.; Christofferson, O. J. Phys. Chem. 1981, 85, 1887. (5) Finholt, A. E.; Bond, A. C., Jr.; Schlesinger H. I. J. Am. Chem. Soc. 1947, 69, 1199. (6) Chizinsky, G.; Evans, G. G.; Gibb, T. R. P., Jr.; Rice, M. J., Jr. J. Am. Chem. Soc. 1955, 77, 3164. (7) Brower, F. M.; Matzek, N. E.; Reigler, P. F.; Rinn, H. W.; Roberts, C. B.; Schmidt, D. L.; Snover, J. A.; Terada K. J. Am. Chem. Soc. 1976, 98, 2450. (8) Turley, J. W.; Rinn, H. W. Inorg. Chem. 1969, 8, 18. (9) Baranowski, B.; Tkacz, M. Z. Phys. Chem. 1983, 135, 27. (10) Herley, P. J.; Christofferson, O.; Todd, J. A. J. Solid State Chem. 1980, 35, 391. (11) Herley, P. J.; Christofferson, O. J. Phys. Chem. 1981 85, 1882. (12) Sandrock, G.; Reilly, J.; Graetz, J.; Zhou, W. M.; Johnson, J.; Wegrzyn, J. Appl. Phys. A 2005, 80, 687. (13) Sandrock, G.; Reilly, J.; Graetz, J.; Zhou, W. M.; Johnson, J.; Wegrzyn, J. J. Alloys Compd., submitted for publication, 2005. (14) Schmidt, D. L.; Diesen, R. W. U.S. Patent 3 840 654, 1974. (15) Matzek, N. E.; Musinski, D. F. U.S. Patent 3 819 819, 1974. (16) Johnson, W. A.; Mehl, R. F. Trans. AIME 1939, 135, 416. (17) Erofeyev, B. V. C. R. Acad. Sci. U. S. S. R. 1946, 52, 511. (18) Avrami, M. J. Chem Phys. 1939, 7, 1103. (19) Avrami, M. J. Chem Phys. 1940, 8, 212. (20) Avrami, M. J. Chem Phys. 1941, 9, 177. (21) Tang, B. T.; Chaudhri, M. M. J. Therm. Anal. 1979, 17, 359. (22) Hancock, J. D.; Sharp, J. H. J. Am. Ceram. Soc. 1972, 55, 74. (23) Constantinou, C. P. Int. J. Chem. Kinet. 1994, 26, 1151. (24) Hutchinson, C. D.; Krishnan Mohan, V.; Millar, R. W. Propellants, Explos., Pyrotech. 1984, 9, 161. (25) Polanyi, M.; Wigner, E. Z. Phys. Chem. A 1928, 139, 439. (26) ComprehensiVe Chemical Kinetics; Bamford, C. H., Tipper, C. F. H., Eds.; Elsevier: New York, 1980; Vol. 22, pp 87-95. (27) Shannon, R. D. Trans. Faraday Soc. 1964, 60, 1902. (28) Gavra, Z.; Johnson, J. R.; Reilly, J. J. J. Less-Common Met. 1991, 172-174, 107.
