Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.issp.ac.ru/lhpp/PapersAntonov/113.pdf Äàòà èçìåíåíèÿ: Fri Feb 10 18:35:19 2012 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 06:30:10 2012 Êîäèðîâêà:

Journal of Alloys and Compounds 264 ( 1998 ) 1 ­ 7

L
x

The crystal structure of Nb 3 AuH
a

a a b a c, a V.E. Antonov , E.L. Bokhenkov , B. Dorner , V.K. Fedotov , G. Grosse *, A.I. Latynin , c c F.E. Wagner , R. Wordel

Institute of Solid State Physics, Russian Academy of Sciences, 142432 Chernogolovka, Moscow district, Russia b Institute Laue-Langevin, 15 X, F-38042 Grenoble Cedex, France c Physics Department E15, Technical U niversity of Munich, D-85747 Garching, Germany Received 21 November 1996

Abstract The A15 alloy Nb 3 Au and its hydrides Nb 3 AuHx with x 52.9 and x 54.2 prepared at high hydrogen pressures were studied by neutron and X-ray diffraction. Their crystal structures were refined using the Rietveld method. Neutron diffraction showed that in both hydrides the hydrogen atoms occupy approximately 2.6 interstitial d sites of the 3 available per formula unit in the A15 metal lattice. The remaining hydrogen fills i sites. The X-ray examination revealed that the metal lattice of the Nb 3 Au alloy and its hydrides was partly disordered, with 861% of the regular Au positions randomly occupied by Nb. This atomic disorder is shown to be the cause of the observed maximum occupancy of d sites in the hydrides. © 1998 Elsevier Science S.A.
Keywords: Metal hydride; Neutron diffraction; X-ray diffraction

1. Introduction The Nb 3 Au alloy belongs to a large and well studied group of A 3 B phases with the A15 structure, space group Pm3n. A number of these phases were shown to form hydrides, the maximum attained hydrogen contents being Ti 3 SbH 2 .1 [1], Ti 3 IrH 3 .9 [2], Ti 3 AuH 2 .8 [3], V3 GaH 1 .9 [4], Nb 3 GeH 0 .86 and Nb 3 AlH 2.18 [5], Nb 3 SnH 2 .2 [6], Nb 3 RhH 6 [7], Nb 3 OsH 4.1 , Nb 3 IrH 4 .7 , Nb 3 PtH 5.1 and Nb 3 AuH 4.3 [8]. All these A 3 BH x hydrides retain the A15 metal lattice, which exhibits the usual expansion with increasing hydrogen concentration. One of the A15 hydrides, Nb 3 SnH x with x ¯1, was studied by neutron diffraction [9]. The hydrogen atoms were found to randomly occupy the 6-fold d positions of the space group Pm3n. Complete occupancy of these positions corresponds to a maximum hydrogen content of x 53. The maximum hydrogen solubility in the A15 alloys of Nb with the transition metals Rh, Os, Ir, Pt and Au, however, exceeds x 54, which implies that other interstitial sites are occupied either alone or in addition to the d sites. Of this family, the Nb 3 AuH x hydrides were chosen for the present structure study because of the low lower limit of their homogeneity range of compositions [8]. This made it possible to prepare hydrides with hydrogen contents both
*Corresponding author. 0925-8388 / 98 / $19.00 © 1998 Elsevier Science S.A. All rights reserved. PII S0925-8388( 97 )00237-5

below (x ¯2.9 ) and substantially above (x ¯4.2 ) the value of x 53 characteristic of the complete occupancy of the d ¨ sites. Moreover, a 197 Au Mossbauer investigation of Nb 3 AuH x [6] has revealed an interesting phenomenon: the spectrum of the Nb 3 Au starting alloy was found to consist of an absorption line with a pronounced shoulder, which was no longer observed in the spectra of the Nb 3 AuH 2 . 8 and Nb 3 AuH 4 .2 hydrides. As the spectrum of the Nb 3 Au alloy should consist only of the single absorption line arising from Au on c positions with m3 symmetry in the ideal A15 structure, a careful investigation of the structure of this alloy and of the metal lattice of its hydrides seemed desirable. The hydrogen positions in the crystal structure of the Nb 3 AuH x hydrides were determined by neutron diffraction, whereas the investigation of the Nb and Au distribution over the sites of the A15 metal lattice required a separate X-ray examination because of the small difference between the neutron scattering lengths of niobium ( 7.054 ° ° A ) and gold ( 7.90 A ) [10].

2. Sample preparation and experimental details As in Refs. [6] and [8], the starting material was a Nb 2.96 Au 1.04 alloy ( Nb 3 Au for brevity hereafter) melted from the elements of 99.99 wt % purity in a levitation


2

V.E. Antonov et al. / Journal of Alloys and Compounds 264 ( 1998 ) 1 ­ 7

induction furnace under an argon atmosphere and then annealed in a vacuum of 10 2 4 Pa at 1100 8C for 24 h and cooled together with the furnace. The hydrides, Nb 3 AuH 2 . 8 8 6 0.06 and Nb 3 AuH 4 .20 6 0.08 , were prepared by exposing 0.3 mm thick plates, cut from the ingot with an abrasive wire saw, to hydrogen pressures of 0.85 and 5.1 GPa, respectively, at 325 8C for 24 h with subsequent cooling to 100 K in the high pressure cell. The method of hydrogenation is described in more detail elsewhere [11]. At ambient pressure, the sample with x 5 4.20 began to lose hydrogen on warming to about 2 20 8C. At room temperature, the hydrogen content of this sample decreased to x ¯ 2.8 in a few hours; no release of hydrogen from the sample with x 5 2.88 was observed at room temperature in the course of a week. The hydrogen content was determined by hot extraction into a calibrated volume at temperatures up to 500 8C using about 5 mg of each sample. When not in use, the hydrides were stored in liquid nitrogen. Prior to the diffraction measurements, they were ground in an agate mortar under liquid nitrogen in order to avoid texture effects. The starting Nb 3 Au alloy was ground at room temperature. The neutron diffraction experiments were performed at 120 K, using the D20 instrument at the ILL, Grenoble, with neutrons of a wavelength of l 5 1.295. The powder diffraction patterns were scanned in steps of 0.18 in 2Q. Both fixed-time and monitor counting schemes were employed. About 1 g samples of Nb 3 Au, Nb 3 AuH 2.88 and Nb 3 AuH 4 . 2 0 were used. The X-ray diffraction experiments were performed at room temperature for the Nb 3 Au and Nb 3 AuH 2.88 samples using a SIEMENS D-500 diffractometer and monochromated CuKa 1 radiation. The patterns were scanned in steps of 0.028 in 2Q. The Nb 3 AuH 4 . 2 0 sample was not measured by X-ray diffraction because it would have lost hydrogen at ambient temperature. The neutron and X-ray data were analysed using the Rietveld profile refinement technique implemented in the DBWS-9411 computer program [12], which allows the simultaneous refinement of several phases.

Fig. 1. X-ray diffraction pattern of the Nb 2.96 Au 1.04 alloy at 300 K and results of its Rietveld refinement involving the Nb 2.98 Au 1.02 , Nb 3 Au 2 and Nb 11 Au 9 phases ( Table 1 ). The experimental data are shown by dots, the calculated profile as a solid line. The difference curve is presented in the lower part of Fig. 1(a), the contributions from the Nb 3 Au 2 (short-dashed curve) and Nb 11 Au 9 ( long-dashed curve) in the lower part of Fig. 1( b).

Nb as Nb

2.98

Au Au

1.02

. The studied sample can thus be considered ¯ 0.97Nb Au 1 0.02Nb 3 Au

2.96

1.04

2.98

1.02

2

1 0.002Nb 11 Au 9 . The refined parameters for the component and the impurity phases are listed in Table 1. Since variations of the site occupancy numbers, v, and the thermal factors of the atoms, B, influence the calculated spectrum in much the same manner, and since their values are therefore strongly correlated, we used the B values determined from the neutron diffraction data ( Table 2 ), which are insensitive to the v values because of the similar neutron scattering lengths of Nb and Au [10]. As Table 1 shows, the A15 structure of the Nb 2.98 Au 1.02 phase exhibits a substantial degree of chemical disorder: about 8% of the 2-fold a sites normally occupied by gold are randomly occupied by Nb and about 3% of the 6-fold c niobium sites are occupied by Au. The X-ray diffraction pattern of Nb 3 AuH 2 . 8 8 is similar to that of Nb 3 Au. Its Rietveld refinement converged to

3. Results

3.1. X-ray diffraction
The analysis of the X-ray diffraction spectrum of the initial Nb 3 Au alloy ( Fig. 1 ) showed that, in addition to the A15 phase, the sample contained impurities which were identified as the tetragonal Nb 3 Au 2 and cubic b-Mn type Nb 11 Au 9 compounds described in [13]. The relative amount of the impurities given by the Rietveld refinement agree with the notion [14] that the maximum gold content of the A15 phase corresponds to the composition


V.E. Antonov et al. / Journal of Alloys and Compounds 264 ( 1998 ) 1 ­ 7

3

Table 1 Positional parameters (x, y, z) and thermal factors (B ) for the constituent phases of the Nb 2.96 Au 1 .04 alloy according to the Rietveld profile refinement analysis of the X-ray diffraction data collected at 290 K. N is the number of atoms per formula unit, v is the site occupancy, M is the number of formula units per unit cell, R p and R e x are the obtained and expected profile factors. Phase Nb 2.98 Au 1.02 Pm3n, M 5 2 ° a 5 5.201 A Atom Au Nb Nb Au Nb Nb Au Au Nb Au Site 2(a) 6(c) 2(a) 6(c) 2(a) 4(e) 4(e) 8(c) 12(d ) 12(d ) x 0.000 0.250 0.000 0.250 0.000 0.000 0.000 0.061 0.125 0.125 y 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.061 0.206 0.206 z 0.000 0.500 0.000 0.500 0.000 0.400 0.200 0.061 0.456 0.456 °2 B (A ) 0.32 0.42 0.42 0.32 0.42 0.42 0.32 0.32 0.42 0.32

v
0.92 0.97 0.08 0.03 1.00 1.00 1.00 1.00 0.92 0.08

N 0.92 2.90 0.08 0.10 1.00 2.00 2.00 8.00 11.00 1.00

Nb 3 Au 2 I4 / mmm, M 5 2 ° ° a 5 3.391 A, c 5 15.18 A Nb 11 Au 9 P4 1 32, M 5 1 ° a 5 7.092 A R p / R e x 5 7.6 / 9.9.

R p / R e x 5 7.0 / 6.0. The impurity fractions and the occupation numbers of the a and c sites in the A15 phase did not change upon hydrogenation. The lattice parameter of the Nb 11 Au 9 phase did not change either, whereas the lattice parameters of the Nb 2.98 Au 1.02 and Nb 3 Au 2 phases increased from the values given in Table 1 for the non° ° hydrogenated material to a 5 5.415 A and to a 5 3.452 A, ° respectively. c 5 15.30 A, The impurities in the samples were found to practically not affect the accuracy of the Rietveld fit for the A15 phase. Their investigation allows conclusions about the interaction of hydrogen with Nb 3 Au 2 and Nb 1 1 Au 9 which have not yet been studied. The dissolution of hydrogen in metals and alloys with a variety of crystal structures, including the A15 structure [8] of Nb 3 Au, has been found °3 to cause the lattice to expand by about 2 ­ 3 A per

dissolved hydrogen atom [11,15]. Within the experimental error, the lattice parameter of the Nb 11 Au 9 phase did not change upon hydrogenation of the sample. This indicates that its hydrogen content did not exceed few atomic per cent. The lattice expansion of the Nb 3 Au 2 phase in the Nb 3 AuH 2 . 8 8 sample corresponds to the formation of a hydride with the estimated composition Nb 3 Au 2 H ¯ 1.5 . From the equation Nb
2.96

Au

1.04

H

2.88

¯ 0.97Nb

2.98

H x 1 0.02Nb 3 Au 2 H
9

1.5

1 0.002Nb 11 Au

it follows that x ¯ 2.94. The hydrogen content of the A15 phase, x ¯ 2.9, thus did not differ from the average ¯ hydrogen content of the sample, x 5 2.8860.06, within the experimental accuracy.

Table 2 Positional parameters (x, y, z) and thermal factors (B ) for the Nb 2.98 Au 1 .02 compound and its hydrides according to the Rietveld profile refinement analysis of the neutron diffraction data. N is the number of atoms per formula unit, v is the site occupancy, R p and R e x are the obtained and expected profile factors. Cubic structure, space group Pm3n, T 5 120 K. Phase Nb 2.98 Au 1.02 ° a 5 5.192 A R p / R e x 5 2.9 / 4.2 Atom Au Nb Nb Au Au Nb Nb Au Hd Hi Au Nb Nb Au Hd Hi Site 2(a) 6(c) 2(a) 6(c) 2(a) 6(c) 2(a) 6(c) 6(d ) 16(i ) 2(a) 6(c) 2(a) 6(c) 6(d ) 16(i ) x 0.000 0.250 0.000 0.250 0.000 0.250 0.000 0.250 0.250 0.217 0.000 0.250 0.000 0.250 0.250 0.215 y 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.500 0.217 0.000 0.000 0.000 0.000 0.500 0.215 z 0.000 0.500 0.000 0.500 0.000 0.500 0.000 0.500 0.000 0.217 0.000 0.500 0.000 0.500 0.000 0.215 °2 B (A ) 0.32 0.42 0.42 0.32 0.33 0.42 0.42 0.33 1.03 1.03 0.39 0.48 0.48 0.39 1.28 1.85

v
0.92 0.97 0.08 0.03 0.92 0.97 0.08 0.03 0.85 0.02 0.92 0.97 0.08 0.03 0.86 0.18

N 0.92 2.90 0.08 0.10 0.92 2.90 0.08 0.10 2.56 0.15 0.92 2.90 0.08 0.10 2.58 1.42

Nb 2.98 Au 1.02 H 2 .9 ° a 5 5.409 A R p / R e x 5 2.9 / 4.1

Nb 2.98 Au 1.02 H 4 .2 ° a 5 5.465 A R p / R e x 5 5.4 / 4.1


4

V.E. Antonov et al. / Journal of Alloys and Compounds 264 ( 1998 ) 1 ­ 7

Fig. 2. The neutron diffraction pattern of the Nb 2.96 Au 1 .04 H 2.88 sample at 120 K. The background and contributions from Nb 3 Au 2 H 1.5 and Nb 11 Au resulting from the Rietveld refinement have been subtracted. The experimental data are shown by dots, the fitted and difference curves as solid lines.

9

3.2. Neutron diffraction
The neutron diffraction patterns of both studied Nb 2.96 Au 1.04 H x hydrides ( Fig. 2 presents an example) contained no new lines compared to that of the initial Nb 2.98 Au 1.02 phase. The crystal structures of the hydrides therefore belong to the same Pm3n space group as the structure of the initial A15 compound. Applying the Rietveld profile refinement method to the neutron data we found the stable solutions listed in Table 2. Hydrogen occupies d and i positions in the A15 metal lattice of both hydrides, with nearly the same occupancy of the d positions. The calculated total hydrogen content of the hydrides, Nd 1 Ni , is in a satisfactory agreement with the data of hot extraction. The sites occupied by different atoms in the resultant hydride structure are shown in Fig. 3, the nearest interatomic distances are given in Table 3. The results of the Rietveld profile analysis ( Table 2 ) are in agreement with qualitative estimates from the relative line intensities, Ihkl , in the neutron diffraction spectra of Nb 2.98 Au 1.02 and its two hydrides. For example, as is seen

Table 3 ° Nearest interatomic distances ( in A ) for Nb 3 AuH x hydrides with the a and x i values from Table 2. Atom-Atom Au-Nb, Au-H Au-H i Nb-H d Nb-H i , H d -H i
d

Formula ] aOE5 / 4 OE]x i a3 ] aOE]]]] 2/4 aoe3( 1 / 4 2 x i )2 1 1 / 8

Nb 3 AuH 3.024 2.033 1.912 1.937

2.9

Nb 3 AuH 3.055 2.035 1.932 1.960

4.2

from the formulae in Fig. 4, I012 should increase whereas I002 and I112 should decrease with increasing occupancy of the d sites. Increasing occupancy of the i sites leads to an increase in I002 , but does not affect I012 and I112 . The experiment shows that the formation of Nb 2.98 Au 1.02 H 2.9 is accompanied by an increase in I012 and a decrease in I112 , which is possible only if the d positions are filled. Both I012 and I112 are virtually the same for Nb 2.98 Au 1.02 H 2.9 and Nb 2.98 Au 1.02 H 4.2 , indicating the same occupancy of the d sites. I002 , however, is higher for Nb 2.98 Au 1.02 H 4.2 , which can be attributed to a higher occupancy of the i sites.

Fig. 3. Crystal structure of Nb 3 AuH x . Au atoms are shown as dark spheres, Nb atoms as light spheres and the H interstices as small black spheres. In the left and right unit cell the interstitial d and i positions, respectively, are shown.


V.E. Antonov et al. / Journal of Alloys and Compounds 264 ( 1998 ) 1 ­ 7

5

which involved only two phases, Nb 2.98 Au 1.02 H x and Nb 11 Au 9 , turned out to be worse than for the two other samples ( Table 2 ).

4. Discussion The arrangement of hydrogen atoms in the crystal structure of the Nb 2.98 Au 1.02 H 2.9 and Nb 2.98 Au 1.02 H 4.2 hydrides suggested by the Rietveld refinement of their neutron diffraction patterns ( Table 2 and Fig. 3 ) appears sound on both chemical and structural grounds. As to the chemical grounds, one should take into account the affinity of hydrogen to niobium, the repulsive interaction of hydrogen with Au atoms dissolved in transition metals [16] and the blocking effect requiring that the distance between hydrogen atoms in a metal should not be much less than 2 ° A [17]. The structural restrictions are that hydrogen should occupy only interstitial sites allowing accommodation of ° atoms with the effective radius of 0.56 ­ 0.60 A [17] and that the hydrogen positions should conform to the Pm3n space group. As is seen from Table 2 and Fig. 3, in the Nb 3 AuH x hydrides the gold atoms at the a positions form a bcc ° lattice with a ¯ 5.4 A. The niobium atoms occupy half of the tetrahedral interstitial sites (c positions) in this bcc lattice. The other half of the tetrahedral sites (d positions) lie at the centres of tetrahedrons formed by nearest-neigh° bour Nb atoms at a distance of about 1.91 ­ 1.93 A, which makes the d sites similar to the tetrahedral positions occupied by hydrogen in niobium hydride [18]. The ° nearest Au atoms being about 3 A away, the d positions therefore look most favourable for hydrogen occupancy. This agrees with the results of the Rietveld analysis showing that d sites in the Nb 2.98 Au 1.02 H x hydrides are the first to be filled with hydrogen. The distance between the neighbouring d sites in the ° Nb 2.98 Au 1.02 H x hydrides is more than 2.7 A, therefore no blocking effects are expected. The complete filling of these sites would correspond to a hydride composition of Nb 2.98 Au 1.02 H 3 . The experiments, however, give a value of Nd 5 2.56 ­ 2.58 for both studied hydrides ( Table 2 ). The observed maximum occupancy of the d sites can be explained by the atomic disorder in the metal sublattice of the hydrides if one assumes that d sites with one or more Au atoms replacing Nb on a nearest-neighbour c site cannot be occupied by hydrogen atoms. Assuming a random occupation of c sites by Au atoms, the number of such Au neighbours of a d site is binomially distributed. The number of d sites that can be occupied by hydrogen atoms is thus N
calc d

Fig. 4. A portion of the neutron diffraction spectra of the Nb 2.96 Au 1.04 (solid line), Nb 2.96 Au 1.04 H 2.9 (short-dashed line) and Nb 2.96 Au 1.04 H 4.2 ( long-dashed line) phases at 120 K with subtracted background. The spectra are normalised to the number of formula units and to the intensity of the incident neutron beam. The formulae under the line indices give approximate contributions from different atoms to the corresponding structure amplitudes.

The lattice parameter of the Nb 11 Au 9 impurity is the same in the initial and in the two hydrided samples. This suggests a low hydrogen solubility in Nb 11 Au 9 at 325 8C and pressures up to 5.1 GPa. It is highly improbable that the samples prepared under these conditions could have lost hydrogen before the neutron diffraction experiments, because they were quenched to 100 K in the high pressure cell and never heated above 120 K. No other hydrides studied so far lost hydrogen at temperatures that low [11]. The neutron diffraction values of the lattice parameters of the Nb 3 Au 2 H x impurity in the Nb 3 AuH 2 . 8 8 sample agree with those determined from the X-ray data, if they are corrected for thermal expansion due to the different temperatures of measurement ( 120 and 300 K ). This confirms that Nb 3 Au 2 forms a hydride of the approximate composition Nb 3 Au 2 H 1 . 5 at 325 8C and a hydrogen pressure of 0.85 GPa, and that this hydride is thermally stable at ambient pressure and room temperature. When calculating the contribution from the Nb 3 Au 2 H 1 . 5 impurity to the neutron diffraction spectrum of the Nb 3 AuH 2 . 8 8 sample, we assumed hydrogen to occupy the 8-fold g positions of the space group I4 / mmm. The g positions are tetrahedral interstitial sites with only Nb atoms as the nearest neighbours. The fraction of the Nb 3 Au 2 H 1 . 5 phase in the sample was, however, too small for a reliable test whether this assumption is correct. No lines attributable to a Nb 3 Au 2 H x impurity were observed in the diffraction pattern of the Nb 3 AuH 4.20 sample. Most probably, the Nb 3 Au 2 phase in this sample formed a higher hydride with a different crystal structure, whose strongest diffraction lines are hidden under the peaks of the main Nb 2.98 Au 1.02 H ¯ 4.2 phase. The presence of such a phase can also explain the fact that for the Nb 3 AuH 4 .20 sample the R p / R ex ratio of the Rietveld fit,

53?

4 SD(1 0

2 c)4 5 2.6260.03,

where c 5 0.03360.003 is the concentration of Au atoms on c sites as derived from the X-ray study. The calculated


6

V.E. Antonov et al. / Journal of Alloys and Compounds 264 ( 1998 ) 1 ­ 7

limiting value, N calc 5 2.62, is in excellent agreement with d the Nd values resulting from the Rietveld fit of the neutron diffraction data ( Table 2 ). The incomplete filling of the d sites also agrees with the presence of a wide plateau with x ¯ 2.8 ­ 2.9 in the pressure dependence of the hydrogen solubility in Nb 3 Au at 325 8C [8]. This plateau extends from about 0.2 to 1.5 GPa and indicates a change in the type of the interstitial sites predominantly occupied by the hydrogen added at higher pressures. According to the Rietveld analysis ( Table 2 ), these sites are the 16-fold tetrahedral i sites with the positional parameter x i ¯ 0.22. These sites lie in pairs along the , 111 . space diagonals ( Fig. 3 ) inside asymmetrical ° tetrahedra of three Nb atoms at a distance of 1.94 ­ 1.96 A ° and a Au atom at 2.03 ­ 2.04 A ( Table 3 ). The distance ] between the two i positions within a pair is a OE3 ( 1/2 2 ° ° 2x i ) ¯ 0.62 ­ 0.66 A, which is much less than 2 A. Therefore, the two i sites of a pair cannot be occupied by hydrogen at the same time due to the blocking effect [17]. Occupancy of half of the i sites corresponds to N max 5 4. i The maximum value of Ni ¯ 1.4 obtained experimentally in Nb 2.98 Au 1.02 H 4.2 ( Table 2 ) is much smaller. One therefore has to assume that an additional blocking mechanism prevents that more i sites become occupied. A conceivable possibility is that the occupation of i sites causes a displacement of the neighbouring Au atom, which prohibits the occupation of other neighbouring i sites by hydrogen. If the local distortions caused by hydrogen atoms occupying asymmetric i sites are random in character, the cubic symmetry of the mean lattice as observed by diffraction methods remains unchanged, but random static displacements of the atoms contribute to the Debye ­Waller factor. This idea is therefore confirmed by the nearly equal B values of Au and Nb in Nb 2.98 Au 1.02 H 2.9 and Nb 2.98 Au 1.02 and the considerable increase in the B values of all atoms in Nb 2.98 Au 1.02 H 4.2 ( Table 2 ). There are still other interstitial sites in the A15 metal lattice, namely the 24-fold k positions situated inside asymmetric tetrahedra formed by three Nb atoms and one Au atom. The k positions are occupied by hydrogen in b-UH 3 , which has the A15 metal lattice with uranium on both a and c sites [19]. The volume of k sites in Nb 3 AuH x is, however, approximately two times smaller than in UH 3 and these sites are therefore unlikely candidates for hydrogen uptake.

approximately 85%, which corresponds to Nd ¯ 2.6 hydrogen atoms per formula unit. The occupancy of the i sites increases from about 2% for the hydride with x 5 2.9 to about 18% for that with x 5 4.2, the latter corresponding to Ni ¯ 1.4. The A15 metal lattice of the studied Nb 3 Au compound and its hydrides was partly disordered. About 8% of the regular Au positions were occupied by Nb and about 3% of the regular Nb positions by Au. This atomic disorder explains the observed maximum concentration, Nd ¯ 2.6, of hydrogen on d sites under the assumption that hydrogen does not occupy d sites with one or more nearest-neighbour Nb atoms replaced by Au atoms. The atomic disorder can also account for the complex 197 ¨ Au Mossbauer spectrum of the initial Nb 3 Au sample reported earlier [6]. Gold atoms occupying c sites expectedly have different hyperfine parameters than those on regular a sites. Furthermore, even Au atoms on regular a sites have different environments arising from the replacement of Nb with Au atoms on neighbouring c sites. The ¨ Mossbauer spectrum can therefore be expected to be complex and can be interpreted correctly only by taking the atomic disorder into account. The results of its analysis will be published later [20] in the framework of a more ¨ general Mossbauer investigation of the Nb-Au-H system. An investigation of impurities in the Nb 3 AuH x samples showed that the hydrogen solubility in the Nb 11 Au 9 compound does not exceed few atomic per cent at 325 8C and pressures up to 5.1 GPa, while the Nb 3 Au 2 compound forms a hydride with an approximate composition of Nb 3 Au 2 H 1 . 5 at 325 8C and 0.85 GPa. This hydride is thermally stable at ambient conditions.

Acknowledgements The authors are thankful to Dr. S.S. Khasanov for the X-ray measurements. The work was supported by Grant No. 96-02-17522 from the Russian Foundation for Basic Research, by the NATO Linkage Grant No. 921403 and by a grant from the Deutsche Forschungsgemeinschaft.

References
[1] K.V.S. Rama Rao, M. Mrowietz, A. Wess, Ber. Bunsenges. Phys. Chem. 86 ( 1982 ) 1135. [2] M. Schlereth, H. Wipf, Scripta Met. Mat. 24 ( 1990 ) 1159. [3] J.B. Vetrano, G.L. Guthrie, H.E. Kissinger, Phys. Lett. 26A ( 1967 ) 45. [4] M. Schlereth, D. Steinbinder, H. Wipf, Z. Phys. Chem. N.F. 164 ( 1989 ) 1065. [5] V.F. Shamrai, L.N. Padurec, Dokl. Akad. Nauk SSSR 246 ( 1979 ) 1182 ( in Russian). [6] M. Baier, R. Wordel, F.E. Wagner, V.E. Antonov, T.E. Antonova, J. Less-Common Metals 172-174 ( 1991 ) 358. [7] V.E. Antonov, T.E. Antonova, I.T. Belash, O.V. Zharikov, A.V. Pal'nichenko, E.G. Ponyatovskii, V.I. Rashupkin, Fiz. Tverd. Tela,

5. Conclusions Within the homogeneity range of the Nb 3 AuH x hydrides, i.e., from x ¯ 2.8 to x ¯ 4.3, hydrogen occupies two types of interstitial positions, d and i, in the A15 metal lattice. The occupancy of the d positions is nearly independent of the total hydrogen content and equal to


V.E. Antonov et al. / Journal of Alloys and Compounds 264 ( 1998 ) 1 ­ 7 30 ( 1988 ) 2152 ( Engl. translation Sov. Phys. Solid State, 30 ( 1988 ) 1240 ). V.E. Antonov, T.E. Antonova, I.T. Belash, O.V. Zharikov, A.V. Pal'nichenko, E.G. Ponyatovskii, and V.I. Rashupkin, Fiz. Tverd. Tela, 31 ( 1989 ) 12 ( Engl. translation Sov. Phys. Solid State, 31 ( 1989 ) 1659 ). L.J. Vieland, A.W. Wicklund, J.C. White, Phys. Rev. B 11 ( 1975 ) 3311. L. Koester, H. Rauch, E. Seymann, Atomic Data and Nuclear Data Tables 49 ( 1991 ) 65. E.G. Ponyatovsky, V.E. Antonov, I.T. Belash, in: A.M. Prokhorov, A.S. Prokhorov ( Eds.), Problems in Solid-State Physics, Mir, Moscow, 1984, p. 109. R.A. Young, A. Sakthivel, T.S. Moss, C.O. Paiva-Santos, DBWS 9411 User's Guide, Georgia Institute of Technology, Atlanta, 1995. K. Schubert, T.R. Anantharaman, H.O.K. Ata, H.G. Meissner, M. ¨ Potzschke, W. Rossteutscher, E. Stolz, Naturwiss. 47 ( 1960 ) 512.

7

[8]

[9] [10] [11]

[12] [13]

[14] E.M. Savitsky, Yu.V. Efimov, N.D. Kozlova, B.P. Mikhajlov, Yu.A. Uspensky, Superconducting Compounds of Transition Metals, Nauka, Moscow, 1976 ( in Russian). ¨ [15] H. Peisl, in: G. Alefeld, J. Volkl ( Eds.), Hydrogen in Metals II, Springer-Verlag, Berlin, 1978, p. 243. ´ [16] M. Baier, M. Karger, R. Ostermayer, F.E. Wagner, I. Dugandzic, H.J. Bauer, V.E. Antonov, T.E. Antonova, V.I. Rashupkin, S.M. Filipek, A. Stroka, Z. Phys. Chem. N.F. 179 ( 1993 ) 309. [17] V.A. Somenkov, S.Sh. Shil'stein, Z. Phys. Chem. N.F. 117 ( 1979 ) 125. ¨ [18] T. Schober, H. Wenzl, in: G. Alefeld, J. Volkl ( Eds.), Hydrogen in Metals II, Springer-Verlag, Berlin, 1978, p. 11. [19] E. Wicke, Kl. Otto, Z. Phys. Chem. N.F. 31 ( 1962 ) 222. [20] G. Grosse, M. Baier, R. Wordel, F.E. Wagner, V.K. Fedotov, V.E. Antonov, T.E. Antonova, to be published.