On the acid hydriding and activation of FeTi

On the acid hydriding and activation of FeTi

Journal of the Less-Common Metals, ON THE ACID HYDRI~ING 115 (1986) 235 - 241 AND ACTIVATION 235 OF FeTi P. RAJ and A. SATHYAMOORTHY Chemistr...

522KB Sizes 0 Downloads 6 Views


of the Less-Common




(1986) 235 - 241













(Received February 19,1985)

Summary Mijssbauer and X-ray diffraction studies were carried out on FeTi powder, which had been acid hydrided using different strengths of HCl. It was found that cubic a-FeTi was transformed predominantly into the monoclinic y-hydride phase, skipping the two intermediate orthorhombic @ phases. It was established from simultaneous transmission and conversion electron Mijssbauer studies that the surface was richer in the higher hydride phase. From a comparison of the acid hydriding behaviour of unactivated FeTi powder reported in this paper with the gaseous hydriding of activated FeTi reported in the literature, it is argued: (i) that in the case of unactivated powder the rate of hydrogen diffusion through the P-phase region is slower than the rate of transformation of the /3phase to the y phase, which explains the skipping of the p phase as well as the surface enrichment, and (ii) that the activation introduces additional channels for hydrogen diffusion which accounts for the change in the hydriding kinetics. The rate-controlling step is hydrogen diffusion in the unactivated material and chemisorption in the completely activated material. Thus, the difficulties associated with the activation can be related to the skipping of fl phase. This is corroborated from our studies on other FeTi-based systems.

1. Introduction FeTi is known to be a potential candidate for hydrogen storage provided it can be activated. The problem of activation of FeTi forms one of the fundamental problems which needs to be thoroughly investigated. FeTi forms several hydrides, FeTiH, , which have been the subject of numerous investigations [ 1 - 14 3. Some controversies exist [7 - 9, 121 about the fundamental aspects of the hydriding behaviour, such as the role of surface composition and the activation mechanism, nature and stability of the different phases, hysteresis etc. Most of the work reported in the literature deals with hydrogen loading under high pressure from the gaseous phase, requiring activation. Not much attention has been paid to other methods 0 Elsevier Sequoia/Printed

in The Netherlands


which do not require any activation. Schober [4] has stressed the importance of systematic investigation of acid hydriding in view of the industrial importance of hydrogen charging in acids and electrolytes. In this paper we are primarily concerned with the nature and sequence of formation of various hydride phases following the acid hydriding of FeTi. There is no general agreement [ll] about the composition ranges of various hydride phases, particularly with regard to the P-phase region. One view is that the fl phase exists over a composition range 1.0 < x < 1.4 while the other view is that there are two discrete orthorhombic pi and flZ phases for x = 1.0 and 1.4 respectively, From the P-C-T plots of Wenzl and Lebsanft [6], Westlake [ll] has inferred that the terminal solubilities of /3i and flZ phases in the absorption route are rather large. This is attributed to the mechanical constraints on the precipitation of the & and y phases. The ranges of these pi, & and y phases are reported to be appreciably different in the desorption route. It may be noted that in either route, the phase coexisting with 01 is /3i [ 6, 111. It has, however, been shown by Reilly et al. [S] that the phases exhibited by a given FeTiH, specimen depend not only on the value of x but also on the sample history. They have shown that starting from activated FeTi, the composition of the fl phase coexisting with the CYphase during hydrogen absorption is FeTiH,_4 (&), while during desorption it is FeTiHieO (pi). For x > 1.4, the p2 and y phases coexist in both routes. Further, by using strain-free FeTi obtained by pre-annealing the activated FeTi powder at 800 “C, these authors obtained a single plateau in the equilibrium absorption isotherm which represents a direct cubic cy to monoclinic y phase conversion, skipping almost completely the two intermediate /3 phases. This is also attributed to the effect of lattice strains on the behaviour of the /3i and & phases. It should, therefore, be of interest to see in which sequence the hydride phases are formed, starting from the annealed but unactivated FeTi powder samples. This is not possible in the gas hydriding route. We report here the first systematic investigation to our knowledge of different hydride phases formed following acid hydriding of well-annealed unactivated powder samples. As a result of comparing the sequence of phases formed in FeTi with those obtained by acid hydriding a series of compounds Fe, _ ,Ni,Ti and FeTii + Y under identical conditions, it is proposed that the difficulties associated with the activation of FeTi are related to skipping the /3 phase. By comparing the 57Fe conversion electron and transmission Mijssbauer spectra it is shown that the surface of the partially acid-hydrided FeTi powder samples is richer in the hydrogen rich y phase than the bulk. It is argued that the rate-controlling step in the hydriding kinetics of unactivated FeTi is the diffusion of hydrogen through the hydrided phase and not the chemisorption at the surface or the phase transformation at the boundary between the 0 phase and the unhydrided regions. We also predict that the activated and subsequently annealed (800 “C) samples of FeTi of Reilly et al. [8], wherein they obtained direct o to y conversion, should also exhibit a y-rich surface.


2. Experimental


FeTi was prepared by arc-melting pure iron (purity, 99.99%) and titanium sponge (purity, 99.9%) under a pure argon (Iolar Grade) atmosphere. The FeTi buttons so produced were re-melted a few times to homogenize them before being annealed at 900 “C in vacuum for 3 days. These buttons were then crushed, powdered and sieved to pass through 300 mesh. Acid hydriding was carried out at room temperature using different concentrations of HCl ranging from 5 to 25% by volume, A known volume of the acid was added to about 1 to 3 g of powder sample and the mixture was constantly stirred. Aliquots of the hydrided material were periodically removed from the bulk, washed several times with distilled water and finally with methanol and dried at room temperature. The hydrogen content of the samples was estimated by measuring the pressure change due to thermal decomposition at 600 “C in vacuum. X-ray diffraction (XRD) patterns were recorded using Cu Ka radiation with a monochromator. “7Fe conversion electron Mijssbauer spectra (CEMS) as well as transmission Mijssbauer spectra (TMS) were simultaneously recorded using a 57Co(Rh) Miissbauer source.

3. Results and discussion FeTi was found to charge readily with hydrogen under the experimental conditions employed without activation as is required for gas hydriding. A maximum of x = 1.8 could be easily achieved. The hydrides formed were found to be very stable and the hydrogen content did not decrease appreciably over a period of several months, owing to the formation of a strong protective layer on the surface. This made the acid-hydrided material suitable for further experimental investigations. Differential scanning calorimetric (DSC) studies also indicated the presence of a strongly bound sealant film on the surface. The calculation of the total amount of heat absorbed from a broad endothermic peak observed around 520 K revealed that nH was higher by an order of magnitude than the known values for FeTi hydrides, thus suggesting that the breaking of the surface film during desorption accounted for a large fraction of the heat absorbed. The identification of the hydride phases formed was made by XRD. It has been shown by Reidinger et al. [7] that pertinent information concerning hydride phase identification can be obtained by scanning a region of 28 values from 38 to 44” using Cu Ka! radiation. The (x phase is identified by its most intense (110) line at about 42.9”; the y phase has its most intense (200) line at about 38.5”; the pi and & phases can be distinguished by the positions of the (111) reflections which occur at about 42” and 41”, respectively. All our XRD patterns showed a predominant cv to y conversion. In the initial stage of hydridmg (3~< 0.3) a small amount of fll and flZ was also


formed, but these phases did not build up as the hydrogen content increased. For values of 3c = 0.5 only trace amounts of pZ could be seen and no pi was discernible. We would like to mention that for 3c> 1.4 XRD patterns showed almost only y-phase lines. A typical set of 57Fe CEMS and TMS recorded simultaneously on samples hydrided by using 10% HCl are shown in Fig. 1. Very similar Mijssbauer spectra were obtained in other cases where different concentrations of HCl were used. The hydrogen concentration corresponding to each of these samples is given in the figure. The Miissbauer spectra could be satisfactorily fitted to a doublet and a singlet and the Mossbauer parameters obtained are listed in Table 1, along with the relative proportions of the two phases. It is seen from this table that the parameters of the doublet are centred around IS = 0.26 and QS = 0.20 mm s-l (within 0.02) and that for the singlet around IS = -0.11 mm s-l, Here IS is given with respect to o-Fe and the zero velocity in Fig. 1 refers to the centre of gravity of the iron spectrum. These values are in fair agreement with the reported values [ 13,141 for the y and (Y phases respectively. This confirms that the amount of p phase formed is very small compared with the other two phases. Comparing the CEMS and TMS results in Table 1, it is observed that the ratio of the y to (11phases is appreciably higher in CEMS than in TMS and both are considerably higher than the expected average value, showing that the surface is richer in the hydrogen-rich phase. The expected average value is calculated assuming that the material consists of a mixture of Q and y phases only and that the JC values for the two phases are 0 and 2

Fig. 1. (a) CEMS and (b) TMS, recorded simultaneously, for various values of x. The zero velocity corresponds to the centre of gravity of the iron metal.











respectively. The CEMS and TMS scan appreciably different depths from the surface, a small fraction (about 0.1) of a micrometre and around 15 pm, respectively. Cu Ka radiation scans around 6 I.tm deep. Owing to a gradual gradation in the hydrogen content along the depth we find some difference in the relative amounts of hydride phases as seen by XRD and TMS. This is particularly appreciable for x > 1.4. As mentioned earlier, in such cases, whereas XRD shows almost single y phase, TMS results do indicate the presence of an appreciable amount of o phase also, see Table 1. In order to understand this difference in hydrogen content between the surface and bulk one should consider the various steps involved in the hydriding process. These are the chemisorption of hydrogen on the surface, its diffusion through the hydrided region and the phase transformation at the boundary between the unhydrided and hydrided regions. The direct (x to y conversion and the fact that the surface is richer in the higher hydride phase suggest that the rates of the above mentioned subprocesses are not equal in the overall hydriding behaviour. This can be understood in the following way. To start with a small region of (I! is converted to /3 phase. In order that the P-phase region grows to engulf the remaining unhydrided a-phase region, it is required that hydrogen should diffuse through the hydrided region fairly fast. Should the rate of diffusion of hydrogen through the P-phase region be slower than the rate of transformation of the p phase itself to the y phase, it should result in a direct a! to y conversion. This kind of behaviour can be expected in the absence of channels in the lattice that facilitate better communication through hydrogen diffusion between the hydrided and the unhydrided regions. The activation of FeTi creates such channels. Presence of impurities and defects that cause strains in the lattice and microcracks enhance diffusion by reducing fracture toughness or by acting as preferential centres for nucleation and growth of the hydride phase and by increasing the surface area. For such a case a smooth cy to fl to y conversion can take place.’ Therefore, the rate-controlling step may change from hydrogen diffusion through the hydrided region in an unactivated material to the chemisorption of the hydrogen molecule on the sample surface when the material has already been activated. Park and Lee [lo] find that the rate-controlling step is chemisorption in their completely activated material. It may be pointed out that, based on the above arguments, it is possible to explain the results of Reilly et al. [8] who also obtained direct 01 to y conversion following the gaseous hydriding of activated and subsequently annealed (800 “C) FeTi powder. This, they attributed to the strainrelieved state of the lattice. They did not elaborate further on this aspect. We believe that on annealing the activated powder some of the channels open for hydrogen diffusion are blocked and hence j3 to y conversion becomes faster than hydrogen diffusion through the P-phase region to the interior. We predict, therefore, that in their samples also similar differences in hydrogen content between the surface and the bulk should be found.


From a comparison of the hydriding behaviour of activated and unactivated FeTi powder, it can be inferred that the difficulties associated with activation during gas hydriding are related to the difficulties of hydrogen diffusion through the P-phase region and consequent skipping of the /3 phase. This is further reinforced by our results on Fe( 1~ Y,Ni,Ti and excess titanium substituted FeTi, wherein no skipping of the p phase was observed. These two systems could be gas hydrided without elaborate activation procedures. More about the formation of the /3 phase and its Mijssbauer parameters in the above two systems will be published later.

Acknowledgments The authors are grateful to Dr. R. M. Iyer, Associate Director, Group, for constant encouragement and to Mr. P. Suryanarayana paring the FeTi sample.

Chemical for pre-

References 1 J. J. Reilly and R. H. Wiswall, Znorg. Chem., 13 (1974) 218. 2 P. Thompson, F. Reidinger, J. J. Reilly, L. M. Corliss and J. M. Hastings,J. Phys. F, 10 (1980) L57. 3 G. Busch, L. Schlapbach, F. Stucki, P. Fischer and A. F. Andresen, Znt. J. Hydrogen Energy, 4 (1979) 29. 4 T. Schober, Rep. Jul-spez-22, KFA Julich (1978). 5 T. Schober, J. Less-Common Met., 89 (1983) 63. 6 H. Wenzl and E. Lebsanft, J. Phys. F, I2 (1982) L49. 7 F. Reidinger, J. F. Lynch and J. J. Reilly, J. Phys. F, 12 (1982) L49. 8 J. J. Reilly, J. R. Johnson, J. F. Lynch and F. Reidinger, J. Less-Common Met., 89 (1983) 505. 9 L. Schlapbach and T. Riesterer, Appf. Fhys., A, 32 (1983) 169. 10 C. N. Park and J. Y. Lee,J. Less-Common Met., 91 (1983) 189. 11 D. G. Westlake,J. Mater. Sci., 19 (1984) 316. 12 H. Zuchner, U. Bilitewski and G. Kirch, J. Less-Common Met., 101 (1984) 441. 13 J: Schwartzendruber, L. H. Bennett and R. E. Watson, J. Phys. F, 6 (1976) L331. 14 G. K. Shenoy, B. D. Dunlap, P. J. Viccaro and D. Niarchos, Adu. Chem. Ser., 194 (1981) 501.