14024
2006, 110, 14024-14027 Published on Web 06/30/2006
Molecular Hydrogen Occupancy in Binary THF-H2 Clathrate Hydrates by High Resolution Neutron Diffraction Keith C. Hester, Timothy A. Strobel, E. Dendy Sloan, and Carolyn A. Koh* Center for Hydrate Research, Colorado School of Mines, Golden, Colorado 80401
Ashfia Huq and Arthur J. Schultz Intense Pulsed Neutron Source, Argonne National Laboratory, Argonne, Illinois 60439 ReceiVed: May 23, 2006; In Final Form: June 15, 2006
We have determined the time-space average filling of hydrogen molecules in a binary tetrahydrofuran (THF)d8 + D2 sII clathrate hydrate using high resolution neutron diffraction. The filling of hydrogen in the lattice of a THF-d8 clathrate hydrate occurred upon pressurization. The hydrogen molecules were localized in the small dodecahedral cavities at 20 K, with nuclear density from the hydrogen approximately spherically distributed and centered in the small cavity. With a formation pressure of 70 MPa, molecular hydrogen was found to only singly occupy the sII small cavity. This result helps explain discrepancies about the hydrogen occupancy in the THF binary hydrate system.
1. Introduction Clathrate hydrates are a group of inclusion compounds which incorporate small molecules in molecular-sized water cavities.1 The ability of the hydrate to concentrate guest molecules has been explored for application in areas such as gas storage and transportation.2 The work on pure H2 clathrate by Mao et al.3 has stimulated much interest because of its possible application to hydrogen storage. However, the stability of hydrates has a strong dependence on the size and shape of the incorporated guest molecule.4 Due to the small size of H2, high pressures (200 MPa at around 273 K5) are required to effectively stabilize the hydrate lattice of the pure H2 clathrate. The use of binary hydrates containing H2 and another guest provides a method of storing H2 in the hydrate framework under milder conditions. While the reduction in required pressure (∼7 MPa at 280 K6) is remarkable with the introduction of tetrahydrofuran (THF), the storage capacity is reduced with the addition of the second guest. Because THF in its stoichiometric hydrate fills all of the large cavities,7-10 only small cavities are available for the H2. In both the pure H2 and the binary H2-THF hydrate, the occupancy in the small dodecahedral (512) cavity is a very important parameter to assess whether these materials will be a viable hydrogen storage medium. From Mao et al.,3 it was first reported that double occupancy of H2 in the 512 cavity was possible. Following that initial report, a more definitive neutron diffraction study showed that a maximum of one hydrogen molecule was found in the small cavity in the pure hydrogen hydrate.11 Although single H2 occupancy in the 512 cavity was shown for the pure H2 hydrate, there is still debate about the H2 * Corresponding author. E-mail:
[email protected]. Phone: (303) 273 3237. Fax: (303) 273 3730.
10.1021/jp063164w CCC: $33.50
occupancy in the small cavity of the binary THF-H2 hydrate. In the first stoichiometric THF-H2 papers,6,12 single occupancy was inferred from volumetric gas release measurements for a pressure up to 350 MPa.12 However, broadening of the H2 peak using 1H NMR suggested that double occupancy could be present at much lower pressures.6 Following this, Lee et al.13 reported that double occupancy was obtained in the stoichiometric THF-H2 hydrate using Raman band areas at 12 MPa; these workers indicated that gas evolution measurements confirmed this result. In addition to these experimental results, numerous modeling studies for the pure H214-17 and the binary H2-THF hydrate18 have been reported, with much disagreement over the occupancy of H2 in the 512 cavity. The existence or nonexistence of H2 double occupancy in the small cavities will be a key factor in determining the feasibility of binary cubic clathrate hydrates for hydrogen storage, because double occupancy changes the possible hydrogen concentration in the hydrate from about 1 to 2 wt % for the stoichiometric hydrate. Neutron diffraction has been used in this study to directly determine the occupancy of hydrogen in the small cavities of this binary hydrate. 2. Experimental Methods The GPPD neutron powder diffractometer at the Intense Pulsed Neutron Source, Argonne National Laboratory,19 was used in this study. Due to the contribution of the inelastic scattering from H atoms to the background, all H atoms were substituted with D atoms. It is assumed that the deuterated and protonated guest molecules exhibit similar behavior, as in previous hydrate studies.10,11,20,21 All neutron data was modeled using the Rietveld refinement package, GSAS22 with the EXPGUI23 interface. Atom positions and the sII lattice parameter were allowed to vary in the refinements. The background was © 2006 American Chemical Society
Letters
J. Phys. Chem. B, Vol. 110, No. 29, 2006 14025
Figure 1. Pure THF-d8 (1a, wRp ) 0.031, χ2 ) 6.3, collected for 18 h) and D2 + THF-d8 (1b, wRp ) 0.036, χ2 ) 4.4, collected for 8 h) formed at 70 MPa. Both were measured at 20 K and 0.1 MPa. Tick marks indicate phase reflections (top, Al; middle, Ih; bottom, sII). The difference line is also shown between the data and model. The background is subtracted.
fit with a model that included diffuse scattering. The estimated standard deviation in the hydrate occupancy was calculated by GSAS. A stoichiometric THF-d8/D2O hydrate was crushed to less than 200 µm and added to a pressure cell described elsewhere.24 Diffraction data were measured for 18 h on the THF-d8/D2O hydrate at 20 K (Figure 1a). The system was then heated to 270 K and pressurized with D2 to 70 MPa for greater than 2 h. Under these conditions, rapid adsorption of the D2 into the hydrate lattice is expected. Subsequently, the system was then cooled in steps over 8 h to 70 K, well below the dissociation temperature for even the pure hydrogen hydrate at ambient pressure. Therefore, the system was at 70 MPa for a total of over 10 h. In laboratory gas evolution experiments, THF hydrate pressurized at 12.5 MPa for 2 h gave equivalent filling to the same system pressurized over 10 days, both at 270 K. With the higher driving force at 70 MPa (versus 12.5 MPa in the gas evolution experiments), we would expect an even more rapid hydrogen filling into the lattice. This result gave confidence that the system studied in this work had reached maximum filling. The D2 was then vented to bring the system to ambient pressure and cooled further to 20 K for the diffraction measurement. Neutron diffraction patterns (not shown) were analyzed before and after depressurization, ensuring hydrogen gas did not escape during depressurization. The diffraction pattern for the D2 + THF-d8 system is shown in Figure 1b. Data were collected for 8.5 h.
3. Results and Discussion The data for the pure THF-d8/D2O hydrate (Figure 1a) were fit by Rietveld profile refinement with a starting model for sII hydrate atom positions10 for the water framework. A section of the Fourier map derived from the observed structure factor amplitudes in the 51264 cavity at (3/8,3/8,3/8) is shown in Figure 2. In accordance with previous diffraction studies,10,25 the THFd8 appears in the Fourier map as a hollow sphere of nuclear density centered in the 51264 cavity. Because THF is disordered in the large cavity, creating a shell of density, previous researchers10,25 have used spherical harmonics or a rigid body approach to model a shell of nuclear density. Using a neutron diffraction study of solid THF-d8,26 an accurate Cartesian representation of THF-d8 was created. This THF-d8 molecule was input as a rigid body in the center of the 51264 cavity at (3/8,3/8,3/8) to model the disordered THF-d8. With the bond lengths and angles constrained, only the rotation of the THF-d8 rigid body, thermal parameters, and occupancies were refined. With all of the THF-d8 atoms in the rigid body in general positions, arbitrary rigid body rotation angles were chosen to improve the fit. As the wRp, a measure of the goodness of fit, was minimized (3.5%), the THF-d8 occupancy refined to unity, in agreement with previous analyses of stoichiometric THF hydrate.7-10 The diffraction pattern for the D2 + THF-d8 hydrate system is shown in Figure 1b. Obvious differences in the Bragg peaks
14026 J. Phys. Chem. B, Vol. 110, No. 29, 2006
Figure 2. Observed Fourier map of the THF-d8 system centered in the 51264 cavity at (3/8,3/8,3/8). This 2D slice also contains other density contours which are portions of waters which make up the cavity. The map size is 10 Å, and contours are drawn at 0.1-1.0 fm. Note the ring of density attributed to the THF-d8 molecule in the 51264 cavity. The FOBS map for the D2-THF-d8 system (not shown) is analogous to this map.
Figure 3. Observed Fourier map of the D2 + THF-d8 system centered in the 512 cavity at (0,0,0). This 2D slice also contains other density contours which are portions of waters which make up the cavity. The map size is 8 Å, and contours are drawn at 0.2-2.0 fm.
between the diffraction patterns (shown in Figure 1) occurred with the addition of the D2, indicating it had entered the lattice. Observed Fourier maps of the small cage (Figure 3) showed that a high amount of nuclear density was centered in the 512 cavity after the addition of the D2. A preliminary analysis using a combined Rietveld/maximum entropy method27,28 gave similar results when comparing nuclear density maps of D2 in the 512 cage. The rigid body THF model (using the rigid body rotation
Letters obtained for the pure THF-d8 hydrate) was input into the model for the D2-THF-d8 hydrate. As with the pure THF-d8 case, the background was modeled on the basis of diffuse scattering. To model the D2 density, 96 equivalent positions from a single D atom were offset from the center of the cage a distance equivalent to half the D-D nuclear separation in fractional coordinates. This approach has been used to model other diatomic molecules.21 The D occupancy was allowed to vary in the refinement, thereby determining the molecular deuterium occupancy in the 512. The occupancy of the D2 was determined to be one D2 molecule per 512 cavity, for the D2 model used. The D2 occupancy refined to 1.005 ( 0.016 for the disordered D atom approach. The occupancy of THF-d8 in the 51264 refined to unity as in the pure THF-d8 hydrate. The addition of the D2 model in the GSAS Rietveld refinement reduced wRp from around 9% to 3%, significantly improving the fit to the data. We obtained equivalent results using a hydrogen model based on a scattering form factor for a spherical shell (zero-order Bessel function) with a D-D radius of 0.37 Å, similar to the approach of Lokshin et al.11 Various other attempts to model the D2 small cage occupancy were performed, including offsetting the D2, similar to the conceptual model proposed originally by Mao et al.3 where D2 molecules would doubly occupy the cavity. These other attempts to model the density in the small cage led to physically unrealistic results (THF occupancies refined to values much greater than unity). 4. Conclusions In summary, this work provides evidence from neutron diffraction data that is consistent with only one hydrogen molecule in the 512 cage for the experimental conditions described above. This is evidence that only single occupancy of hydrogen in the 512 cavity can be expected in the THF binary hydrate system, at least up to 70 MPa. With only single occupancy of the small cavities, it raises doubts as to whether the THF/H2 sII hydrate system will lead to effective H2 storage based on the current required storage criterion. In order for clathrate hydrates to be practical hydrogen storage, binary hydrate structures which enable multiple occupancy of hydrogen will likely be required, both for the decrease in required formation conditions and the increase in stored hydrogen density. An understanding of the occupancy and orientation of hydrogen in the host water lattice will be key in future development and studies of these binary hydrate materials. Acknowledgment. K. Hester is supported by NURP grant UAF03-0098. T. Strobel is funded by the U.S. Department of Energy, under contract DE-FG02-05ER46242. Work at Argonne was supported by the U.S. Department of Energy, Basic Energy Sciences-Materials Sciences, under contract W-31-109-Eng-38. J. Richardson is thanked for his advice on these experiments using GPPD. We thank A. Gupta and P. Rensing for experimental assistance. E. Maxey, K. Volin, and J. Fieramosca are thanked for their tireless efforts with all of the experimental apparatus on GPPD. References and Notes (1) Sloan, E. D., Jr. Clathrate Hydrates of Natural Gases, 2nd ed.; Marcel Dekker: New York, 1998. (2) Koh, C. A. Chem. Soc. ReV. 2002, 31, 157. (3) Mao, W. L.; Mao, H. K.; Goncharov, A. F.; Struzhkin, V. V.; Guo, Q. Z.; Hu, J. Z.; Shu, J. F.; Hemley, R. J.; Somayazulu, M.; Zhao, Y. S. Science 2002, 297, 2247.
Letters (4) Davidson, D. W. Clathrate Hydrates. In Water: A ComprehensiVe Treatise; Franks, F., Ed.; Plenum Press: New York, 1979; Vol. 3. p 115. (5) Dyadin, Y.; Larionov, E.; Manakov, A.; Zhurko, F.; Aladko, E.; Mikina. MendeleeV Commun. 1999, 209. (6) Florusse, L. J.; Peters, C. J.; Schoonman, J.; Hester, K. C.; Koh, C. A.; Dec, S. F.; Marsh, K. N.; Sloan, E. D. Science 2004, 306, 469. (7) Davidson, D. W. Can. J. Phys. 1971, 49, 1224. (8) Gough, S. R.; Davidson, D. W. Can. J. Chem. 1971, 49, 2691. (9) Rosso, J. C.; Carbonnel, L. C. R. Acad. Sci. 1971, C273, 15. (10) Jones, C. Y.; Marshell, S. L.; Chakoumakos, B. C.; Rawn, C. J.; Ishii, Y. J. Phys. Chem. B 2003, 107, 6026. (11) Lokshin, K. A.; Zhao, Y.; He, D.; Mao, W. L.; Mao, H.; Hemley, R. J.; Lobanov, M. V.; Greenblatt, M. Phys. ReV. Lett. 2004, 93, 125503. (12) Udachin, K.; Lipkowski, J.; Tzacz, M. Supramol. Chem. 1993, 3, 181. (13) Lee, H.; Lee, J.; Kim, D. Y.; Park, J.; Seo, Y.; Zeng, H.; Moudrakovski, I. L.; Ratcliffe, C. I.; Ripmeester, J. A. Nature 2005, 434, 743. (14) Alavi, S.; Ripmeester, J. A.; Klug, D. D. J. Chem. Phys. 2005, 123, 1. (15) Patchkovskii, S.; Tse, J. S. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 14645. (16) Sluiter, M. H. F.; Adachi, H.; Belosludov, R. V.; Belosludov, V. R.; Kawazoe, Y. Mater. Trans. 2004, 45, 1452.
J. Phys. Chem. B, Vol. 110, No. 29, 2006 14027 (17) Sluiter, M. H. F.; Belosludov, R. V.; Jain, A.; Belosludov, V. R.; Adachi, H.; Kawazoe, Y.; Higuchi, K.; Otani, T. Lect. Notes Comput. Sci. 2003, 2858, 330. (18) Alavi, S.; Ripmeester, J. A.; Klug, D. D. J. Chem. Phys. 2006, 124, 014704. (19) Jorgensen, J. D.; Faber, J.; Carpenter, J. M.; Crawford, R. K.; Haumann, J. R.; Hitterman, R. L.; Kleb, R.; Ostrowski, G. E.; Rotella, F. J.; Worlton, T. G. J. Appl. Crystallogr. 1989, 22, 321. (20) Gutt, C.; Asmussen, B.; Press: W.; Johnson, M.; Handa, Y.; Tse, J. J. Chem. Phys. 2000, 113, 4713. (21) Chazallon, B.; Kuhs, W. F. J. Chem. Phys. 2002, 117, 308. (22) Larson, A. C.; Dreele, R. B. V. General Structure Analysis System (GSAS); Los Alamos National Laboratory Report LAUR 86-748; 2000. (23) Toby, B. H. J. Appl. Crystallogr. 2001, 34, 210. (24) Henning, R.; Schultz, A.; Thieu, V.; Halpern, Y. J. Phys. Chem. A 2000, 104, 5066. (25) Mak, C. W.; McMullan, R. K. J. Chem. Phys. 1965, 42. (26) David, W. I. F.; Ibberson, R. M. Acta Crystallogr. 1992, C48, 301. (27) Baumert, J.; Gutt, C.; Johnson, M. R.; Tse, J. S.; Klug, D. D.; Press, W. J. Chem. Phys. 2004, 120, 10163. (28) Burger, K. Ph.D. Thesis, Universitat Tubingen, Germany, 1997.
