Low temperature hydrogen desorption in MgAl thin ...

APPLIED PHYSICS LETTERS 94, 241901 共2009兲

Low temperature hydrogen desorption in MgAl thin films achieved by using a nanoscale Ta/Pd bilayer catalyst Helmut Fritzsche,1,a兲 Colin Ophus,2 Chris T. Harrower,2 Erik Luber,2 and David Mitlin2 1

National Research Council Canada, SIMS, Canadian Neutron Beam Centre, Chalk River Laboratories, Chalk River, Ontario K0J 1J0, Canada 2 Chemical and Materials Engineering, University of Alberta, T6G 2V4, and National Research Council Canada, National Institute for Nanotechnology, T6G 2M9, Edmonton, Alberta, Canada

共Received 10 March 2009; accepted 8 May 2009; published online 15 June 2009兲 We used a nanoscale 共5 nm Ta/5 nm Pd兲 bilayer catalyst to achieve remarkable desorption kinetics for thin films. Full hydrogen desorption occurred at 100 ° C with a noticeable desorption even at room temperature. This is a significant improvement relative to the 175 ° C needed to fully desorb an identical film with a single Pd layer acting as the catalyst. Neutron reflectometry confirmed that the Ta/Pd bilayer remained intact both after hydrogen absorption and following the hydrogen desorption. We used x-ray diffraction analysis to gather complementary information regarding the crystal structure of the as-synthesized, sorbed and desorbed film. © 2009 American Institute of Physics. 关DOI: 10.1063/1.3154550兴 MgAl alloys have attracted interest for their favorable hydrogen storage properties for more than 20 yrs. This relatively simple system, available commercially in ingot form, is attractive due to a combination of the relative low material cost and the environmentally benign nature of the alloy. In general, most studies found MgAl alloys promising,1–10 however, the kinetics was still inadequate for the rapid low temperature desorption required of a commercially viable hydrogen storage material for automotive and portable hydrogen applications.11 Low temperature hydrogen absorption in MgAl thin films has been achieved with a single layer Pd, a bilayer Ti/Pd, and Fe共Ti兲/Pd catalysts.5–7 However, appreciable hydrogen desorption is only possible at temperatures too high for practical applications.4,12,13 In this work we explore the bilayer catalyst system Ta/Pd. Mg0.7Al0.3 films were cosputtered onto a Si 共100兲 substrate with a native oxide layer of about 1 nm thickness.13 The films were absorbed for 24 h at 125 ° C and 40 bar hydrogen. Desorption was performed in a 1 bar Ar atmosphere in a sample cell equipped with a heater.14 When varying the temperature for the desorption experiment, the sample was always kept at the various temperatures for 1 h before starting the neutron measurement. The neutron reflectometry 共NR兲 experiments were performed on the D3 reflectometer at the neutron research reactor National Research Universal 共NRU兲 in Chalk River. For reflectometry the interaction with the film is reduced to a one-dimensional problem and for grazing incidences the reflectivity can be described with an optical potential V j, known as Fermi’s pseudopotential, V j = 2␲ប/mN jb j ,

共1兲

where m is the neutron mass, N j is the number density, b j is the coherent nuclear scattering length, and the product N jb j is the scattering length density 共SLD兲 in layer j. The SLD depends on the elements and their isotopes in the sample.15,16 With Eq. 共1兲 the neutron refractive index and the Fresnel reflectivity arising at interfaces can be calculated based on a兲

Electronic mail: [email protected].

0003-6951/2009/94共24兲/241901/3/$25.00

the Parratt recursion algorithm.17 A layer model is fit to the measured data by varying the SLD, layer thickness, and interface roughness of each individual layer j. More technical details about NR can be found in Ref. 18. Figure 1 shows the neutron reflectivity curves of a Si共100兲 / Ta/ Mg0.7Al0.3 / Ta/ Pd film structure 共a兲 before hydrogen absorption, 共b兲 after hydrogen absorption, measured

FIG. 1. Reflectivity curve of a 27 nm thick Mg0.7Al0.3 film prepared on a Si共100兲 wafer with 10 nm Ta buffer layer and capped with a 共5 nm Ta/5 nm Pd兲 bilayer: 共a兲 as prepared, 共b兲 after hydrogen absorption, measured at 25 ° C, and 共c兲 after annealing of 1 h at 100 ° C. Open circles represent experimental data, the solid lines are fits, and the insets show the corresponding SLD profile.

94, 241901-1

© 2009 American Institute of Physics

Downloaded 16 Jun 2009 to 132.246.151.39. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp

241901-2

Fritzsche et al.

Appl. Phys. Lett. 94, 241901 共2009兲

FIG. 2. Comparison of the desorption characteristics of a Mg0.7Al0.3Hy film capped with a 10 nm Pd single catalyst layer 共open circles, from Ref. 13兲 and a 共5 nm Ta/5 nm Pd兲 catalyst bilayer 共solid dots兲. Both samples were annealed for 1 h at the respective temperature prior to the NR scan, which took 11 h for the sample with the Pd layer and 2 h for the sample with the Ta/Pd bilayer.

at 25 ° C, and 共c兲 fully desorbed, measured at 100 ° C. The fits, displayed as solid lines, were calculated using the software PARRATT32.19 The changes in the film structure can be best visualized by plotting the SLD profile, i.e., the SLD along the surface normal z of the film. The SLD profiles corresponding to the fits are shown in Fig. 1 as insets. In all cases the model consisted of a Si substrate with a native SiO2 layer, a Ta buffer layer, a MgAl layer, and a Ta/Pd bilayer. The effect of hydrogen absorption on the SLD profile can be easily seen by comparing the SLD profile of the sorbed 关inset Fig. 1共b兲兴 to the unsorbed film 关inset Fig. 1共a兲兴. The negative scattering length bH = −3.739 fm of the hydrogen15 causes the SLD to drop from 2.2⫻ 10−6 Å−2 for the unsorbed MgAl film to 7.3⫻ 10−8 Å−2 for the sorbed MgAl film. From this decrease in the SLD we can calculate that 4.7 wt % hydrogen are stored in the MgAl film.13 After annealing the film to 100 ° C for 1 h, the SLD of the MgAl film goes back up to 2.0⫻ 10−6 Å−2 关see Fig. 1共c兲兴 proving that the hydrogen has been released. The SLD of the desorbed film does not reach exactly the SLD value of the unsorbed film because the whole film structure expands by about 15% due to the hydrogen absorption creating cracks and voids that result in a lower SLD of the layers. Furthermore, we can deduce from the SLD profile that the hydrogen is uniformly dispersed within the MgAl layer and no hydrogen is stored in the Pd layer. The SLD of the Ta layer decreases during the desorption process further from 3.6 ⫻ 10−6 Å−2 to 3 ⫻ 10−6 Å−2, which is an indication of a small amount of hydrogen being stored in the Ta layer after the annealing to 100 ° C. The SLD profiles shown in Fig. 1 prove that the Ta/Pd bilayer is still intact at the applied temperatures. In earlier experiments on MgAl films with single Pd catalyst layers we found that Pd diffuses into the MgAl layer.20 Figure 2 shows the total hydrogen content y of the Mg0.7Al0.3Hy film capped with a Ta/Pd bilayer 共solid dots兲 as calculated from the SLD, plotted as a function of temperature. For comparison the earlier data from a Mg0.7Al0.3 film capped with a single Pd layer13 are included as open circles. That demonstrates that the bilayer catalyst lowers the temperature necessary to achieve full hydrogen desorption to 100 from 170 ° C for the single Pd layer.

FIG. 3. XRD pattern of a 27 nm thick Mg0.7Al0.3 film prepared on a Si共100兲 wafer with a 10 nm Ta buffer layer and capped with a 共5 nm Ta/5 nm Pd兲 bilayer: 共a兲 as prepared, 共b兲 measured immediately after hydrogen absorption, 共c兲 after 30 h at 25 ° C, 共d兲 after 30 h at 100 ° C, and 共e兲 XRD scan of the sample that was investigated with NR after 3 h at 125 ° C.

Figure 3 shows the x-ray diffraction 共XRD兲 results for an as-synthesized thin film 关Fig. 3共a兲兴, measured immediately after hydrogen absorption 关Fig. 3共b兲兴, stored at room temperature for 30 h after absorption 关Fig. 3共c兲兴, and annealed at 100 ° C in argon 关Fig. 3共d兲兴. It is the same film structure as investigated with NR but it is not the identical film. The XRD scan of the sample that was investigated with NR is displayed in Fig. 3共e兲, measured after the annealing at 125 ° C for 3 h. Because the films are strongly textured, not all possible reflections appear in the x-ray scan. The as-synthesized microstructure consists of a supersatured solid solution of Al in Mg. In the diffraction pattern we can clearly identify the Mg共002兲 peak at 2␪ = 35.1° 关Fig. 3共a兲兴 which is shifted from the pure Mg共002兲 peak at 2␪ = 34.4° to larger angles due to the slightly smaller lattice constant of the alloy. After sorption 关Fig. 3共b兲兴 the Mg共002兲 peak disappears and at the same time an ␣-MgH2 共110兲 peak occurs at 2␪ = 27.9°. There is no evidence of any ternary hydride formation and no Mg peaks are present, indicating a full transformation to ␣-MgH2. The diffraction pattern after storage at room temperature for 30 h is shown in Fig. 3共c兲.


241901-3

Appl. Phys. Lett. 94, 241901 共2009兲

Fritzsche et al.

The relative intensity of ␣-MgH2 to Mg is decreased. This clearly indicates that some hydrogen desorption has occurred at room temperature. After 30 h at 100 ° C, shown in Fig. 3共d兲, the microstructure consists of Mg phase with a small amount of ␣-MgH2. The x-ray pattern shows no evidence of Mg17Al12 or other binary intermetallic formation. There is a small MgH2 peak still visible in Fig. 3共d兲, whereas the XRD scan of the sample that was investigated with NR after 3 h annealing at 125 ° C 关displayed in Fig. 3共e兲兴 shows no MgH2 peaks. In this study we used NR and XRD to investigate the low temperature hydrogen sorption properties of Mg0.7Al0.3 thin films catalyzed with a nanoscale 共5 nm Ta/5 nm Pd兲 bilayer. We have demonstrated that a bilayer catalyst is much more effective than a single Pd layer catalyst, lowering the temperature necessary to achieve full hydrogen desorption to 100 versus 170 ° C. Our experimental findings are in agreement with calculations of properties of alloy catalysts where the Ta/Pd alloy surface has the lowest hydrogen binding energy among the investigated Pd alloys.21 Improving the catalyst by alloying or using bilayers is a very promising technique to reduce the desorption temperatures for hydrogen storing metal hydrides. The research presented herein is made possible by a reflectometer jointly funded by Canada Foundation for Innovation 共CFI兲, Ontario Innovation Trust 共OIT兲, Ontario Research Fund 共ORF兲, and the National Research Council Canada 共NRC兲. We acknowledge the great dedication of Z. Tun 共CNBC兲, M. King 共AECL兲, P. Adams 共AECL兲, D. West 共NRC兲, J. Fox 共CNBC兲, and the whole CNBC technical group to design, build, and commission D3. We also ac-

knowledge the funding by the National Science and Engineering Research Council of Canada and Alberta Ingenuity. D. L. Douglass, Metall. Trans. A 6, 2179 共1975兲. M. Dornheim, N. Eigen, G. Barkhordarian, T. Klassen, and R. Bormann, Adv. Eng. Mater. 8, 377 共2006兲. 3 A. Zaluska, L. Zaluski, and J. O. Ström-Olsen, Appl. Phys. A: Mater. Sci. Process. 72, 157 共2001兲. 4 R. Gremaud, A. Borgschulte, C. Chacon, J. L. M. Van Mechelen, H. Schreuders, A. Züttel, B. Hjörvarsson, B. Dam, and R. Griessen, Appl. Phys. A: Mater. Sci. Process. 84, 77 共2006兲. 5 R. Gremaud, A. Borgschulte, W. Lohstroh, H. Schreuders, A. Züttel, B. Dam, and R. Griessen, J. Alloys Compd. 404, 775 共2005兲. 6 G. Garcia, R. Domènech-Ferrer, F. Pi, J. Santiso, and J. Rodriguez-Viejo, J. Comb. Chem. 9, 230 共2007兲. 7 R. Domènech-Ferrer, M. G. Sridharan, G. Garcia, F. Pi, and J. RodriguezViejo, J. Power Sources 169, 117 共2007兲. 8 Q. A. Zhang and H. Y. Wu, Mater. Chem. Phys. 94, 69 共2005兲. 9 A. Andreasen, Int. J. Hydrogen Energy 33, 7489 共2008兲. 10 R. A. Varin, C. Chiu, T. Czujko, and Z. Wronski, Nanotechnology 16, 2261 共2005兲. 11 DOE roadmap; http://www.hydrogen.energy.gov/roadmaps_vision.html. 12 L. L. Pranevicius and D. Milcius, Thin Solid Films 485, 135 共2005兲. 13 H. Fritzsche, M. Saoudi, J. Haagsma, C. Ophus, E. Luber, C. T. Harrower, and D. Mitlin, Appl. Phys. Lett. 92, 121917 共2008兲. 14 K. G. Yager, O. M. Tanchak, C. J. Barrett, M. J. Watson, and H. Fritzsche, Rev. Sci. Instrum. 77, 045106 共2006兲. 15 V. F. Sears, Neutron News 3, 26 共1992兲. 16 Neutron scattering lengths available online from the NIST Center for Neutron Research at www.ncnr.nist.gov/resources/n-lengths. 17 L. G. Parratt, Phys. Rev. 95, 359 共1954兲. 18 T. P. Russell, Mater. Sci. Rep. 5, 171 共1990兲. 19 C. Braun, Software PARRATT32 developed for BENSC, HMI, Berlin, Germany. 20 H. Fritzsche, M. Saoudi, J. Haagsma, C. Ophus, C. T. Harrower, and D. Mitlin, Nucl. Instrum. Methods Phys. Res. A 600, 301 共2009兲. 21 J. Greeley and M. Mavrikakis, Nature Mater. 3, 810 共2004兲. 1 2
