6
2005, 109, 6-8 Published on Web 12/17/2004
Hydrogen Desorption Exceeding Ten Weight Percent from the New Quaternary Hydride Li3BN2H8 Frederick E. Pinkerton,*,† Gregory P. Meisner,† Martin S. Meyer,† Michael P. Balogh,§ and and Matthew D. Kundrat‡,# Materials and Processes Laboratory and Chemical and EnVironmental Sciences Laboratory, General Motors Research and DeVelopment Center, 30500 Mound Road, Warren, Michigan 48090-9055 ReceiVed: September 30, 2004; In Final Form: NoVember 17, 2004
Mobile applications of hydrogen power have long demanded new solid hydride materials with large hydrogen storage capacities. We report synthesis of a new quaternary hydride having the approximate composition Li3BN2H8 with 11.9 wt % theoretical hydrogen capacity. It forms by reacting LiNH2 and LiBH4 powders in a 2:1 molar ratio either by ball milling or by heating the mixed powders above 95 °C. This new quaternary hydride melts at ∼190 °C and releases g10 wt % hydrogen above ∼250 °C. A small amount of ammonia (2-3 mol % of the generated gas) is released simultaneously. Preliminary calorimetric measurements suggest that hydrogen release is exothermic and, hence, not easily reversible.
Practical on-board storage of sufficient hydrogen is one of the most challenging issues in implementing fuel cell technology for automotive use.1 This has stimulated intense research to create new hydrogen-rich materials that release hydrogen at practical temperatures and pressures. We report a new quaternary hydride with the approximate composition Li3BN2H8 synthesized by reacting LiNH2 and LiBH4 powders in a 2:1 molar ratio. When heated above ∼250 °C the quaternary composition releases g10 wt % H2 gas. The released gas also contains a small amount of NH3 (2-3 mol %). Li3BN2H8 exemplifies a promising new class of hydride materials whose significant hydrogen storage potential for stationary and mobile applications we are now actively exploring. Known reversible metal hydrides either have insufficient hydrogen capacity (e.g., LaNi5H6, 1.4 wt % H2)2 or have unacceptably high operating temperatures (e.g., Mg alloys and Mg2Ni, 300 °C).3 Considerable progress has been reported recently on light metal compounds for hydrogen storage. For example, sodium alanate (NaAlH4) is known to undergo reversible two-step decomposition through Na3AlH6 to NaH plus Al, producing a theoretical capacity of 5.6 wt %. Bogdanovic´ and Schwickardi4 dramatically improved the kinetics and operating temperature for NaAlH4 by incorporating Ti and Ti alloys in the material. More recently, Chen et al.5 reported that lithium nitride (Li3N) absorbs hydrogen in a two-step reaction to form LiNH2 + 2LiH, with a theoretical hydrogen capacity of 10.4 wt %. However, only the second reaction step, Li2NH + LiH + H2 T LiNH2 + 2LiH, is reversible under practical temperature and pressure conditions, and it releases only 5.2 wt % hydrogen. By eliminating an extra LiH in the second * Corresponding author. E-mail:
[email protected]. † Materials and Processes Laboratory. § Chemical and Environmental Sciences Laboratory. ‡ Aerotek Corporation. # Now with the State University of New York at Buffalo, Buffalo, NY 14260-3000.
10.1021/jp0455475 CCC: $30.25
reaction step, the hydrogen capacity increases to 6.5 wt % for the reaction LiNH2 + LiH T Li2NH + H2.6,7 Both the NaAlH4 and Li-N-H systems retain a large portion of their hydrogen content in the dehydrided state as NaH and Li2NH, respectively. We therefore sought to identify chemistries that could release all of their available hydrogen. The ternary compound Li3BN2, for example, is known to exist,8 and several polymorphs of Li3BN2 have been reported,9-11 including high temperature12 and high-pressure10,13 phases. Assuming Li3BN2 to be the hypothetical hydrogen-free endpoint of a dehydriding reaction, the mixture 2LiNH2 + LiBH4 has the correct LiB-N stoichiometry to form the Li3BN2 phase if the full hydrogen content of 11.9 wt % is released. We discovered that this mixture not only releases all of its hydrogen at elevated temperatures, but it reacts to form a new, previously unknown quaternary hydride at lower temperatures. Commercially available LiNH2 and LiBH4 powders were mixed in a 2:1 molar ratio and ball milled for times up to 960 min. General Area Detector Diffractometer System (GADDS) X-ray diffraction (XRD) results in Figure 1 show the effects of milling time on a mixture of 2LiNH2 + LiBH4 powders, where the starting mixture is labeled “0 min”. As the milling time increases, the LiNH2 and LiBH4 diffraction peaks weaken and a new set of peaks emerges. After 40 min, the sample is substantially converted to the new phase, with only a small remnant of LiNH2 visible in the XRD pattern. When the mixture is milled for 300 min, the conversion is complete, and continued milling up to 960 min produces no further change. The final XRD pattern appears to be single phase, except for a small amount of Li2O impurity. Furthermore, the XRD background intensity is essentially unchanged with milling time; thus ball milling evidently does not produce an amorphous phase. We therefore assign the approximate composition of the new phase to be Li3BN2H8; its density is 0.96 g/cm3 as measured by a He gas pycnometer. Preliminary analysis indicates that all of the © 2005 American Chemical Society
Letters
Figure 1. GADDS X-ray diffraction scans for a mixture of 2LiNH2 + LiBH4 ball-milled for the indicated times. The initial mixture (0 min) is composed of the LiNH2 and LiBH4 starting materials, which convert with increasing milling time to the Li3BN2H8 phase. The vertical lines indicate the 2θ values of the Li3BN2H8 diffraction peaks determined using the high resolution Siemens D5000 powder diffractometer. The scans have been displaced vertically for clarity.
observed XRD peaks are consistent with a single phase having a body-centered cubic structure with a lattice constant of a ) 10.76 Å. More precise 2θ values for the Li3BN2H8 XRD peaks were obtained from a Siemens D5000 powder diffractometer; these are shown as vertical lines in Figure 1. Further work to determine the exact stoichiometry and the crystal structure is in progress. Heating a mixture of 2LiNH2 + LiBH4 powders also produces the Li3BN2H8 phase, as shown by Figure 2. First, LiNH2 and LiBH4 were milled separately for 10 min each, and then the milled powders were combined and thoroughly mixed by hand. The two GADDS XRD patterns shown in Figure 2 are taken from a series of in situ scans14 at ∼4 °C intervals as the mixed sample was heated at 2 °C/min to 168 °C. The XRD pattern at 71 °C (lower trace) is essentially identical to the pattern obtained at ambient temperature before heating. It is interesting to note that when this sample was originally prepared, its XRD pattern showed only the expected superposition of LiNH2 and LiBH4 peaks. The sample was stored in a sealed bottle in a glovebox at ambient temperature for 12 days prior to the experiment shown in Figure 2. During that time, the sample spontaneously reacted and formed a substantial quantity of the Li3BN2H8 phase, providing a dramatic illustration of the quaternary phase’s roomtemperature stability. Above 71 °C the sample rapidly transforms to Li3BN2H8, and by 109 °C (upper trace) the transformation is complete. The midpoint of the transformation occurs at about 95 °C. Gravimetric experiments show that there is no weight loss during the thermal transformation, thus the full hydrogen content of the starting materials is retained in forming Li3BN2H8. Figure 3 shows a thermogravimetric analyzer (TGA) measurement15 of desorption from 2LiNH2 + LiBH4 synthesized using the second method (powders milled individually and then
J. Phys. Chem. B, Vol. 109, No. 1, 2005 7
Figure 2. Conversion of 2LiNH2 + LiBH4 to Li3BN2H8 by heating. The X-ray diffraction pattern at 71 °C is identical to the scan obtained at room temperature at the start of the experiment. Note that this sample, stored 12 days at room temperature after originally being mixed, has already spontaneously formed a substantial quantity of Li3BN2H8. Heating above ∼95 °C completes the conversion to Li3BN2H8, as shown by the pattern at 109 °C. The squares show the Li3BN2H8 peak positions determined from higher resolution data obtained using the D5000 powder diffractometer. The scans have been displaced vertically for clarity.
mixed). The conversion to Li3BN2H8 at ∼95 °C is unaccompanied by any gravimetric event, proving that no material is lost during the reaction. The onset of significant weight loss occurs at about 250 °C, and the total weight loss is 13.1 wt %. This is somewhat in excess of the theoretical hydrogen capacity of Li3BN2H8 (11.9 wt %) even before taking into account the presence of inert impurities such as Li2O. The lower panel in Figure 3 represents a simultaneous measurement of the gases evolved during desorption, obtained by sampling the TGA exhaust gas using a mass spectrometer operated as a residual gas analyzer (RGA). Only the relevant signals from the H2, NH3, and the NH2+ ion at 16 amu16 are shown; no significant changes above RGA background levels were seen in other monitored gas species (H2O, N2/CO, O2, CO2). In particular, the sample did not evolve water vapor to within our detection limit. Hydrogen release above 250 °C is predominantly responsible for the weight loss. However, concurrent production of ammonia is also a significant contribution. Semiquantitative analysis of the RGA signals indicates that the mass ratio of NH3 to H2 in the exhaust gas is about NH3:H2 ) 1:7; thus the weight loss due to H2 is about 11.5 wt %, and the contribution from NH3 is about 1.6 wt %. The evolved gas is more than 98 mol % H2. We observed melting of the Li3BN2H8 phase starting at about 190 °C, thus hydrogen evolution actually occurs from the molten state. It dehydrides to a solid product composed predominantly of a new polymorph of Li3BN2, most closely resembling the quenched high-pressure Li3BN2 polymorph,10 together with a small quantity of the monoclinic high-temperature Li3BN2 polymorph and the ubiquitous Li2O impurity. In some cases there is evidence for minor amounts of one or more as yet unidentified additional phases in the dehydrided product.
8 J. Phys. Chem. B, Vol. 109, No. 1, 2005
Letters In summary, we have synthesized a new quaternary hydride with the approximate composition Li3BN2H8 by combining LiNH2 and LiBH4 powders in a 2:1 molar ratio. The new phase is obtained by reacting mixed powders either by ball milling for 300 min or by heating above ∼95 °C, and it is stable at room temperature. Powder XRD data analysis is consistent with a single quaternary phase having a body-centered cubic crystal structure with a lattice constant of a ) 10.76 Å. The quaternary phase melts at ∼190 °C and releases g10 wt % H2 gas when heated above ∼250 °C, forming a mixture of solid Li3BN2 polymorphs. The released gas also contains 2-3 mol % ammonia. Volumetric and gravimetric desorption measurements are in excellent agreement regarding the quantities of hydrogen and ammonia released from Li3BN2H8. Preliminary calorimetric studies indicate that the dehydriding reaction is exothermic, implying that the reverse reaction is not thermodynamically favored. Our attempts to directly rehydride the decomposition product by heating under H2 gas at pressures of up to 8 MPa have so far not achieved significant hydrogen reabsorption. This new material and its derivatives nonetheless represent promising research candidates in the search for practical on-board hydrogen storage materials.
Figure 3. TGA measurement of desorption from Li3BN2H8 during heating (upper panel) and simultaneous RGA measurement of the evolved gas in the TGA exhaust stream (lower panel). The sample was heated at 5 °C/min to 350 °C, and then soaked. RGA signals at times less than 54 min reflect background RGA readings; changes above these levels represent gases emitted by the sample. The RGA signals are delayed by approximately 3 min relative to the TGA weight loss.
Acknowledgment. We thank Matt Scullin for technical assistance and Dr. Jan Herbst for valuable discussions. We also thank Stephen Swarin, James Spearot, and Mark Verbrugge for their support of this research. Supporting Information Available: Chemical analysis of the starting materials; complete synthesis procedures; details of X-ray diffraction, gravimetric, and volumetric apparatus; X-ray diffraction pattern obtained after desorption; calorimetric data. This information is available free of charge via the Internet at http://pubs.acs.org. References and Notes
Figure 4. Volumetric measurement of thermal desorption from Li3BN2H8, heated at 0.5 °C/min to 364 °C, and soaked for 680 min, calculated by assuming that all of the desorbed gas is H2.
Figure 4 shows a volumetric measurement17,18 of hydrogen release from Li3BN2H8 synthesized by ball milling a 2LiNH2 + LiBH4 mixture for 960 min. A 302.2 mg sample was heated at 0.5 °C/min to 364 °C in the evacuated sample chamber, and then soaked for 680 min. Gas release began at about 250 °C and the final equilibrium gas pressure was 550.5 kPa at 364 °C. If all of the desorbed gas were H2, this would represent 10.2 wt % H2 release. A portion of the evolved gas, however, is expected to be NH3 on the basis of the TGA and RGA results. By combining the volumetric measurement of the number of moles of evolved gas with a TGA measurement of 12.4 wt % mass loss on this same material, we infer that the released gas is about 10.0 wt % H2 and 2.4 wt % NH3. Thus, for this sample the evolved gas is approximately 97 mol % H2.
(1) Pinkerton, F. E.; Wicke, B. G. The Industrial Physicist 2004, 10 (February/March), 20. (2) Buschow, K. H. J. In Handbook on the Physics and Chemistry of Rare Earths; Gschneidner, K. A., Eyring, L., Eds.; North-Holland, New York, 1984; Vol. 6, Chapter 47. (3) Schwarz, R. B. MRS Bull. 1999, 24, 40. (4) Bogdanovic´, B.; Schwickardi, M. J. J. Alloys Compd. 1997, 253, 1. (5) Chen, P.; Xiong, Z.; Luo, J.; Lin, J.; Tan, K. L. Nature 2002, 420, 302. (6) Meisner, G. P.; Pinkerton, F. E.; Meyer, M. S.; Balogh, M. P. Proceedings of the International Symposium on Metal-Hydrogen Systems 2004, Cracow, Poland, Sep. 5, 2004; J. Alloys Compd., in press. (7) Ichikawa, T.; Isobe, S.; Hanada, N.; Fujii, H. J. Alloys Compd. 2004, 365, 271. (8) Villars, P. Pearson’s Handbook Desk Edition; ASM International: Materials Park, OH, 1997; p 771. (9) Goubeau, J.; Anselment, W. Z. Anorg. Allg. Chem. 1961, 310, 248. (10) DeVries, R. C.; Fleischer, J. F. Mater. Res. Bull. 1969, 4, 433. (11) Yamane, H.; Kikkawa, S.; Koizumi, M. J. Solid State Chem. 1987, 71, 1. (12) Yamane, H.; Kikkawa, S.; Horiuchi, H.; Koizumi, M. J. Solid State Chem. 1986, 65, 6. (13) Wentorf, R. H., Jr. J. Chem. Phys. 1961, 34, 809. (14) Balogh, M. P.; Tibbetts, G. G.; Pinkerton, F. E.; Meisner, G. P.; Olk, C. H. J. Alloys Compd. 2003, 350, 136. (15) Pinkerton, F. E.; Meyer, M. S.; Tibbetts, G. G.; Chahine, R. Proceedings of the 11th Canadian Hydrogen Conference, Victoria, BC, Canada; Canadian Hydrogen Association: Montreal, 2001; pp 633-642. (16) Though in principle mass 16 amu could also be due to CH4, the absence of significant carbon in the sample argues that the observed mass 16 signal represents the ammonia gas fragmentation ion NH2+. (17) Tibbetts, G. G.; Meisner, G. P.; Olk, C. H. Carbon 2001, 39, 2291. (18) Meisner, G. P.; Tibbetts, G. G.; Pinkerton, F. E.; Olk, C. H.; Balogh, M. P. J. Alloys Compd. 2002, 337, 254.
