Diffusion behavior in titanium-chromium hydrides

0360--3199/83/$3.00 + 0.00 Pergamon Press Ltd. (~ 1983 International Association for Hydrogen Energy.

Int. J. Hydrogen Energy, Vol. 8. No. 10. pp. 801-808. 1983. Printed in Great Britain.

DIFFUSION BEHAVIOR IN TITANIUM-CHROMIUM HYDRIDES* R. C. BOWMAN, JR, B. D. CRAFT, A. ATTALLAand J. R. JOHNSON+ Monsanto Research Corporation-Mound, Miamisburg, OH 45342 U.S.A. and +Department of Energy and Environment, Brookhaven National Laboratory, Upton. NY 11973 U.S.A.

(Received for publication 2 February 1983) Abstract--NMR measurements of the proton relaxation times T1, Ttp, and T2 have been performed on the low (i.e. s-phase) and intermediate (i.e. o/-phase) hydrogen concentrations in CO-stabilized TiCrzHx with both the hexagonal (C14) and cubic (C15) Laves structures. The relaxation times indicate rapid proton diffusion above 200 K for all the TiCr:Hx phases; however, large differences in the diffusion activation energies are observed. This behavior is structure sensitive and has been associated with variations between interstitial site occupancies and diffusion pathways for the C15 and C14 structures. The estimated 300 K diffusion constants in the a/-phase of TiCr2Hx are comparable to the literature values for HfV2H2~, LaNisH65 and PdH070 but are orders of magnitude larger than the proton diffusion constant in TiFeH10.

INTRODUCTION Many intermetallic compounds with the nominal stoichiometry AB: possess C15 cubic (i.e. space group Fd3m) or C14 hexagonal (i.e. space group P63/mmc) Laves crystal structures [1]. As described in the recent papers by Shaltiel [2], many alloys with the C14 and C15 Laves structure will reversibly react with gaseous hydrogen to form ternary metal hydrides that often have rather large maximum storage capacities. Hence, several Laves alloys are very promising candidates [3] for various hydrogen storage and energy conversion applications. Based upon relatively limited (but apparently representative) published data [4, 5], many AB2 alloys with either the C14 or C15 structure have good to excellent kinetics for both the absorption and desorption of hydrogen. Although the surface segregation of the AB2 alloys during hydriding produces [5, 6] free B metal particles that may catalytically decompose H2 molecules at the alloy surface, a relatively rapid bulk hydrogen diffusion rate in the AB2Hx phases is also an important factor in the favorable kinetics. In fact, recent nuclear magnetic measurements (NMR) of the proton relaxation times by Shinar et al. [7, 8] for the Laves hydrides ZrMn2Hx(C14), HfV2Hx(C15) and ZrV2Hx(C15) indicate relatively high hydrogen diffusion rates with correspondingly low diffusion activation energies E,. However, there has not been any previous comparative evaluation of the diffusion behavior between the C14 and C15 hydrides while retaining the same nominal ratio of identical metal atoms. The present paper reports on proton NMR studies of the hydrogen diffusion behavior in both the C14 and C15 allotropes of nominal TiCr2Hx. The TiCr2 alloys * Portions of this work were presented at the International Symposium on the Properties and Applications of Metal Hydrides, Colorado Springs (1980) and World Hydrogen Energy Conference IV, Pasadena (1982).

are particularly interesting since both the C15 and C14 structures have been shown [9, 10] to form unstable hydride phases with rather large storage capacities that make these alloys suitable for applications [3] involving temperatures well below room temperature. The proton relaxation times indicate rapid diffusion rates above 200 K for the low (i.e. o~-phase) and intermediate (i.e. cr'-phase) hydrogen concentrations in both the C14 and C15 structures of TiCr2Hx; however, the diffusion behavior is complex with non-Arrhenius temperature dependences as well as evidence for simultaneous slow and rapid motion processes below 170 K. Above 200 K the Ea values for the C14 and C15 o-phase samples are nearly identical while a large difference exists between the d-phase samples. This behavior can be qualitatively associated with differences in hydrogen site occupancies and more restricted diffusion pathways in the C14 Laves structure. EXPERIMENTAL Extensive descriptions of the preparative procedures for the TiCr2Hx alloys with the C15 and C14 structures as well as thermodynamic quantities and tentative phase diagrams have been previously published [9-11] and thus only a brief summary of sample preparation will be given. The C15 Laves hydrides were prepared from a single-phase alloy of composition TiCrl.s0 while the C14 hydrides were from a single-phase alloy with the composition TiCrl~. Recent unpublished diffraction and density data indicate the nonstoichiometry of the TiCry alloys corresponds to substitutional disorder where some Ti-atoms occupy Cr lattice positions of the ideal C15 and C14 Laves structures. It has not been possible to prepare single phase alloys with the stoichiometric composition TiCr2.0. The NMR samples of both the low hydrogen content (i.e. o~-phase or solid solution) and intermediate hydrogen content (i.e. a/phase) were prepared by direct reaction with H2 gas in 801

802

R. C. BOWMAN et al.

the single-phase regions of the C15 (TiCrl.8-H) and C14 (TiCrLg-H) phase diagrams [9-11]. Because the TiCryH~ samples will spontaneously decompose at room temperature or even much lower temperatures due to the combined effects of high dissociation pressures and rapid desorption kinetics, it was necessary to stabilize [9] the TiCryHx powders by exposure to high-pressure CO at low temperatures before removing the hydride products from the synthesis reactor. The CO forms a surface film on the TiCryHx particles that inhibits the loss of hydrogen from the interior. Portions of the CO-stabilized TiCryH~ were subsequently sealed in evacuated 7 m m o.d. pyrex NMR sample tubes. To prevent accidental decomposition these samples were never heated above 300 K. The reported hydrogen contents of the TiCryH~ samples were determined by volumetric analyses following completion of the NMR experiments. Conventional pulse techniques were used to measure the various proton relaxation times over the temperature range 100-300 K. The inversion-recovery sequence yielded the spin-lattice relaxation times T1 while the phase-shifted rf pulse method with an applied magnetic field of 7.3 G was used to determine T~p, the spin-lattice relaxation in the rotating frame. The simple two-pulse spin-echo sequence gave the T'2 values. The spin-spin relaxation times T2~ were determined using the CarrPurcell-Meiboom-Gill (CPMG) sequence [12] with a 333

250

200

r

r

TEMPERATURE (K) 167 143 125 l

p

o

i

:

111 i

spacing of 200 #s between the ~r-pulses. All the measurements were performed at a proton resonance frequency of 34.5MHz on a previously described [13] NMR spectrometer. RESULTS The proton relaxation times for the four CO-stabilized TiCryH~ samples (where y ~ 2) are summarized in Figs. 1-4. TiCrLsH0.55 (ol-phase) and TiCn.sH25s ( ~ phase) have the cubic C15 structure while TiCn9H0.63 (o~-phase) and TiCrL9H2.s5 (W-phase) have the hexagonal C14 structure. From the general relationships between nuclear relaxation times and the hydrogen diffusion rate that have been throughly described by Cotts [14], the data in Figs. 1-4 indicate substantial hydrogen mobility which is characteristic of both the C14 and C15 TiCryH~ phases. The specific evidence for rapid hydrogen diffusion includes the following: (1) the T~ minima, which correspond [14] to a mean H-atom hopping rate of approx. 2 x 10S/s, occur at temperatures between 240 and 260K; (2) the T~ o minima, which correspond to a hopping rate of about 2 x 106/s, occur in the temperature range 140-170K; and (3) the observed T2 values from both the CPMG and spin-echo techniques exceed the predicted dipolar rigid-lattice values to temperatures below 160 K for the W-phase samples and to nearly 100 K for the ol-phase samples.

100

1000

333

250 [

TEMPERATURE (K) 167 143 125

200 t

9

1

--

[

~ - - r

111 ....

100

~' - T i Cr I . 8H2.58

500 (~-TiCrl .8Ho. 55

"TI

o ,3

©Q

T1

100

00 0

~o °

100

c c~°° o o o o n°

E 5O ~'~v £ 1 o (H1 = 7.3 G)

V-- 10 z

0

:7

O

(CPMG) @

AV

X

r2~

[I 7"

,7 tx

(CPMG)

O FO

z O j.-

v

1 = 7.2G),~

1.0

O

~7

v

rr 13.

~Tlo(H

tr

v

1.0

/

v v

r2'(Spin Echo) •

0.1

0.5 =-T2'(Spin Echo)

0.1 3.0

1

1.

4.0

50

i

,.

60 70 1031T(K -1)

,

,

8.0

9.0

0.01 10.0

Fig. 1. The proton relaxation times T1, T1o, T~, and T2mfor CO-stabilized oL-phase TiCrlsH055 with the C15 Laves structure.

t

3.0

,10

I

,0

I

g0

100

I031T (K -I )

Fig. 2. The proton relaxation times T1, Tip, T'2, and T2m for CO-stabilized W-phase TiCrl.sH2.5s with the CI5 Laves structure.

803

DIFFUSION IN TITANIUM-CHROMIUM HYDRIDES 333 lOOOI

250 I

TEMPERATURE (K) 200 167 143 125 I

t

I

500 ~ a - TiCr1,9H0.63 T1

100

[

100

333 500

I

• •

111 •

TEMPERATURE (K) 200 167 143 125

250 i

i

i

i

111

100

i

a'- TiCr1.9H2.85

•

10(3

ee~eoeeoe'~O • •

Tl......~ed~e • • A

50

5e

o °

E C--G

10

Z

z

0 T<

5.0

T1% ,,

Z

I-

5

@

I

w

5 O

10

V< ×

5

X

~:

o

1.0

7 O

0.5

O

1.0

o-

0.5

°

0.10

a~

0.05 0"13 0.01

t

I

I

4

5

I

I

I

I

6

7

8

9

1

I

I

1

I

I

4

5

6

7

8

9

10

103/T (K-1) 10

103/T (K - 1 )

Fig. 3. The proton relaxation times T~, Tip, T'2, and T2m for CO-stabilized o~-phase TiCrx.gH0.63with the C14 Laves structure. The short vertical lines connect limiting Tip and T'2 values in regions of nonexponential recovery.

Combining these observations provides an unmistakable case of high hydrogen mobility as has been previously been found [7, 8] in other C14 and C15 hydrides. Before proceeding to the quantitative analyses of the proton relaxation times, some unusual aspects of the data in Figs. 1--4 will be noted. First, the relaxation times Tip and T~ for the C14 TiCrl.9Hx sample are nonexponential below 160 K. Similar behavior has been observed in /3-LaNisH7 [15] and /3-LaNis-yAlyI-Ix [16] and is attributed to the simultaneous motion of inequivalent H-atoms with different hopping rates. The unusual minimum in T2 for o~-TiCrlsH0.55 is probably caused by an atomic exchange process [17] between mobile and locally trapped H-atoms. The substitutionally disordered Ti atoms that occupy Cr positions are believed to be the trapping sites. This mechanism is currently under experimental study and will be described in more detail elsewhere. Translational diffusion of H-atoms affects proton relaxation times through a time-dependent modulation of the dipolar interactions [14]. Since the 47Ti, 49Ti, and 53Cr magnetic nuclei have small moments and low isotopic abundances, only the proton-proton dipolar interactions need be considered. However, the relaxation times in Figs. 1-4 need some corrections to remove the

Fig. 4. The proton relaxation times Ta, Tap, and T2mfor COstabilized o/-phase TrCil.9H2.a5 with the C14 Laves structure. The short vertical lines connect limiting Tlo values in a region of nonexponential recovery.

non-diffusion contributions. For example, T1 in metallic systems will generally obey the relation T i -1 = T i "1 + Ti-e1

(1)

where Tld is the diffusion contribution [14] and Tx~arises from hyperfine interactions with the conduction electrons. The Tie term is easily determined [8, 13, 16] from the temperature dependence of 7"1 at low temperatures where the Txd contribution becomes insignificant [14]. Similar corrections are usually required [16] for the T2m and Tip parameters as well. The temperature dependences of the corrected proton relaxation times for the four TiCryHx samples are presented in Figs. 5-8. For rather general conditions [14], these corrected proton relaxation times can be related to the diffusion correlation time ~ through the expressions [13, 16]: T? 1 = Clr7'

(below the T, minimum)

(2)

TF~ = C2r~/(1 + 4~o~r~)

(3)

T~ = C3rc

(4)

where Ci (i = 1, 2, 3) are model-dependent constants [14] and (Ol is the angular frequency of the spin-locking field HI. The rc value is inversely proportional [14] to

804

R.C. B O W M A N et al.

100

312 i

250

208

i

i

TEMPERATURE (K) 179 156 139 i

r

I

125

312

114

100

i

i

i~

250

208

i

i

60

60

TEMPERATURE (K) 179 156 139 i

i

i

125

114

t

, (Ea = 0.25(2)eV)

lo(Ea "-- O.19(2)eV) 3O

3O

g

E m

O3 I.U

510 z O

10 z O

X

5 ~3 z O O ¢r

z 0 I-(3

Q.

Q.

o

1.0

"'

0.6

uJ )-rr"

~ 1.0

(Ea = 0.20(2)eV)

~: 0.6

(Ea = 0.26(2)eV)

rr

I1;

0 0

0 0

0.3

0.1

\

3.2

~ 4.0

~. 48

0+3

z~

i ~ ~ 56 6.4 103/T (K "1)

= 0.10(1)eV)

0.11 ~

7.2

'

S.0

3.2

8.S

iX ~,

410

418

5.6 614 10.31T(K-1)

712

8J.0

8.8

Fig. 5. Corrected proton relaxation times for mTiCrl.sH0.55 (C15) from data in Fig. 1.

Fig. 6. Corrected proton relaxation times for d-TiCrl.sH2.ss (C15) from data in Fig. 2.

the hydrogen diffusion constant D through the relation [18] D =/a/12rc. (5)

role of structure on this complex and currently incompletely developed diffusion mechanism will be briefly considered for the C14 and C15 TiCry hydrides in the next section.

The Ea values may be derived from the temperature dependence of the relaxation times given in equations (2)-(4) by assuming ~ obeys the Arrhenius relation rc = r0

exp(Eo/kBr)

(6)

where v0 is the inverse of the attempt frequency and kB is the Boltzmann's factor. The E, values that were obtained from the corrected proton relaxation times in Figs. 5-8 are summarized in Table 1. It is immediately obvious that in each TiCryHx sample unique E, values were not obtained from the different relaxation times. Although the r 2 d and Tip data for temperatures above the Tip minima give nearly identical Ea values; the T~d, T~, and low-temperature Txo data yield significantly smaller Ea values. Similar behavior has recently been seen in several other metal hydrides including fl-LaNisH7 [15] and fl-LaNis-yAlyHx [16]. It has been suggested [15, 16] that the various NMR relaxation times can respond differently to two types of motion. The smaller E, values represents local hopping among closely spaced interstitial H-sites while the larger E= values corresponds to jump processes involved in long-range diffusion. The

DISCUSSION In both the C14 and C15 Laves structures for AB2 alloys the only interstitial lattice sites that are available [19] for occupancy by the hydrogen atoms have distorted tetrahedral symmetry with AzB2, AB3 or B, as the nearest neighbor coordination. The recent calculations by Magee et al. [14] have demonstrated that A2B2 sites have the largest radii for both C15 and C14 structures. On the premise that H-atoms preferentially occupy the largest available sites amenable with the formation of reasonable metal-hydrogen bonds, it is expected that the AzB2 sites would be more readily occupied. Furthermore, since chromium can only form a hydride phase with difficulty while Till2 is a prototype stable transition metal hydride, the TiCr3 and Cr4 sites are unlikely to be significantly occupied. This view is in complete agreement with neutron diffraction studies of the C15 deuterides ZrCrzDx (x = 2.89 and 3.08) [20], ZrCr2D3.5 [21], and TiCrl.sDx (x = 0.85 and 2.20) [21] where the best fits to all the powder patterns indicated only the Zr2Cr2 or Ti2Cr2 sites contained deuterium.

DIFFUSION IN TITANIUM-CHROMIUM HYDRIDES 333

1°°° I 50O

250 I

TEMPERATURE (K) 2OO 167 143 125 i

j

i

i

i

• Ea=O.088e V A

100

333

250

200

i

i

i

i

r

^.,

A T1 p A

5o

Z

z 0

Ea=0.40eV

V- 10 x 5 5

10

x

5

7 0

,.~

• Tld D Tlp

50

O

o

100

I

:E l--

5

111

/x T2d ~' T 2'

~1oo

~7 T 2'

I

a ' - TiCrl.9H2.85

Ea=0.28eV

I-

F-

I

I

500 ~ t

Tld

12 T2d

loo

TEMPERATURE (K) 143 125 167

111

a - TiCrl. 9H0.63

805

:7 0

~ 1.0 Q: 0"0.5 a

1 0.5

Ea=0.O47eV

w I.(.) w e~

Fo u.i fie

~ 0,1

~ 0.1 Ea=O.050eV

0.05

0.05

k

0.01

3

4

5

6 7 103/-I. ( K - l )

8

9

10

3

4

5

•

6 7 103/T (K-1 )

8

9

10

Fig. 7. Corrected proton relaxation times for c~-TiCr19H0.63 Fig. 8. Corrected proton relaxation times for ot'-TiCrLgH2.s5 (C14) from data in Fig. 3. (C14) from data in Fig. 4.

Although n e u t r o n diffraction data is not currently available for ZrCrEDx or TiCrEDx with the C14 structure, Didisheim et al. [23] found the D-atoms to occupy only Zr2Mn2 sites in the C14 phase ZrMnED3. Although all

the A2B2 sites are equivalent in the C15 structure (i.e. subgroup 96g), there are four distinct AEB2 sites existing for the C14 structure (i.e. subgroups 6h~, 6h2, 12k and 241). From a consideration of the arrangements [19, 23]

Table 1. Comparison of activation energies Eo from proton NMR measurements for TiCr2Hx samples with the C14 and C15 Laves crystal structures Sample composition

Laves structure type

TiCrl.sH0.~5

C15 (oc)

TiCn.sHzss

C15 (a ~)

TiCrl.gH0.63

C14 (o:)

TiCrl.9H2.s5

C14 ( ~ )

NMR parameter

Temperature range (K)

Ea (eV)

T1o Tip Tip T2d Tld Tip Tip T2d Tld Tip T2d T~ T]d Tip Tip T2d

180-250 190-280 115-170 180-260 180-240 180-280 120-170 180-250 170-250 180-290 180-260 105-170 200-250 200-270 125-175 200-270

0.13 ± 0.01 0.19 ± 0.02 0.03 -+ 0.01 0.20 + 0.01 0.135 + 0.01 0.25 -+ 0.02 0.10 ± 0.01 0.26 - 0.01 0.088 -- 0.010 0.218 - 0.010 0.205 --+0.010 0.050 -+ 0.005 0.28 +- 0.01 0.40 ± 0.02 0.18 ± 0.02 0.40 - 0.02

806

R. C. BOWMAN et al.

of the A2B2 sites in the C14 and C15 structures, it is clear that H-atom hopping paths involving only equivalent sites in the C15 structure will include two or more different sites for the C14 phase. A room temperature neutron diffraction study [23] of ZrMn2D3 with the C14 structure has indicated nonuniform deuterium occupancies among the four types of Zr2Mn2 sites. Since all the A2B2 sites in the C14 structure should have identical dimensions [19], this variation in D-atom population may be reflecting differences in the diffusion characteristics for the inequivalent Zr2Mn2 sites. An examination of the Ea parameters in Table 1, which summarize the hydrogen diffusion behavior in TiCryHx 0' = 1.80 and 1.90) as represented by the proton relaxation time data given in Figs. 5-8, illustrates that host metal structure plays an important role for the diffusion processes in the C15 and C14 Laves phases. This assertion is most exemplified from a comparison of the Ea parameters above 180K for o:'TiCr1.gH~.85(C14) with the Ea values for crTiCrl.gH0.63(C14) and a/-TiCr1.sH2.s8(C15). In the C15 structure E, increases by 20% in going from o: to o,/ phase. This increase may correspond to correlation effects [24] in the H-atom jumps at the larger hydrogen concentration. However, E~ increases by a factor of two between the cr-TiCr1.9Ho.63and a"-TiCrl.9H2.85 samples. Furthermore, the E, parameter is significantly smaller in the C15 a/-TiCra.sHzs8 compared to Ea in the C14 o:'-sample. Since the E~ values above 180 K are similar for o~-phase TiCrzH~ with either the C14 or C15 structures, the reduction in symmetry cannot be entirely responsible for the large E~ value in o/-TiCrL9H2.85. Although hydrogen trapping at the defect Ti3Cr sites could give rise to a larger diffusion activation energy, this process should be much less important for the C14 cr'-TiCrL9Hz.85 sample since it is more nearly stoichiometric than C15 a/-TiCr~.sH2.58 (i.e. there are only half as many Ti3Cr sites in the C14 sample). Hence, if hydrogen trapping at these defects was entirely responsible for increasing E,, cr'-TiCrl.9H2.s5 should actually have the smaller activation energy, which is completely opposite to the experimental results in Table 1. It is possible that above a minimum hydrogen concentration preferential ordering of H-atoms on subsets of the inequivalent Ti2Cr2 sites [19] in the C14 structure can occur. In fact, deuterium ordering has been recently observed via neutron diffraction techniques in several C15 deuterides including HfVD4 [25], ZrV2D3.6 [26] and ZrCrzD3.8 [27]. However, neither X-ray or neutron diffraction [11, 22] have indicated formations of ordered superstructures in TiCr~.sH~ or TiCrl.8D~ samples (x ~< 3.0) with the C15 structure. The possibility of superlattice formation in cr'-TiCra.9H(D)~ with the C14 structure has not yet been experimentally investigated. If some of the presumably ordered H-atoms occupy sites corresponding to deeply bound potential well minima that also lie in the more accessible diffusion pathways [23, 26], the diffusion process would require larger activation energies or alternative pathways. This is speculated to be the situation for a~'-TiCrl.9H2.85. How-

ever, analyses [7] of the proton 7"2' relaxation times for ZrMn2Hx (2.0 ~
DIFFUSION IN TITANIUM-CHROMIUM HYDRIDES

807

Table 2. Comparison of Ea and room temperature diffusion constant D(300 K) for several transition metal hydrides. All results deduced from proton NMR measurements Sample o/-TiCrt.sH2 s~ oc'-TiCrl9H2s5 HfV2H21 /J-TiFeHL03 TiCuHo94 ;/-Till17~ fl-PdH070 /~-LaNi~H6.5

E~ (eV)

D(300 K) (cm2/s)

Reference

0.26 _+0.01 0.40 -+ 0.02 0.19 +- 0.01 0.33 -+ 0.07 0.79 _+0.03 0.543 0.23 -+ 0.01 0.42 _+0.04

3.4 x 10-a* 1.5 × 10 - 7 . 8 x 10-s* 1.2 x 10-~z 9.9 x 10-~ 4.6 × 10-13 1.2 x 10-; 1.4 × 10-~

Present work Present work Shinar et al. [7] Bowman and Tadlock [18] Bowman et al. [13] Bustard et al. [28] Sevmour et al. [29] Karlicek et al. [15]

* Assuming l = 2.5 × 10 8cm.

TiCrl.gH285(C14), and HfVzH2A(C15) are quite large and are in the same range as /3-PdH0,70 [29] and /3LaNisH6.5 [15], which are among the highest hydrogen mobilities known for metal hydrides [30]. In contrast, D(300 K) for/3-TiFeHx [18], which has been the prototype hydride for many energy storage applications [3], is more than four orders of magnitude smaller where the values for ]/-TiH1.71 [28] and TiCuHo.94 [13] are still smaller. The large D(300 K) for the o/-TiCr~Hx samples are due both to relatively small Ea as well as large numbers of vacant Ti2Cr2 sites to facilitate [18] the diffusion process. The large proton diffusion coefficients in Table 2 for C14 and C15 o/-TiCryHx are entirely consistent with the rapid hydrogen absorption-desorption kinetics for other Laves alloys [4, 5]. However, the dissociation of hydrogen molecules at the segregated metal surfaces [6, 7] probably remains a dominant factor in controlling the kinetics under most experimental situations. An important advantage of the TiCr2Hx alloys over TiFeH~ for low-temperature storage applications [3] is the much larger bulk hydrogen diffusion which permits relatively rapid attainment of equilibrium conditions at temperatures below 200 K for the TiCr2Hx alloys [9, 10]. Since the C15 and C14 TiCr2Hx systems have high dissociation pressures, large total hydrogen storage capacities, rapid hydrogen diffusion, and essentially no hysteresis effects [9, 10]; the TiCr2 alloys should in these respects be better candidates than LaNi5 (or related AB5 alloys) for many hydrogen storage applications that involve temperatures below room temperature. A high resistance of the chosen alloy to various surface poisoning effects [31] from gaseous impurities (e.g. 02, H20, CO, etc.) is also a desirable characteristic for most applications [3]. For example, LaNi5 is much less sensitive than TiFe to 02 and H20 impurities [31]. Although no systematic studies of the surface poisoning behavior of the TiCr2 alloys have yet been conducted, the experiences during the preparation of numerous CO-stabilized TiCr2Hx samples for a variety of experiments [9-11, 22] imply that TiCr2 alloys are probably rather insensitive to surface poisoning by the common impurities.

Acknowledgements--This work was supported by Division of

Chemical Sciences, Office of Basic Energy Sciences, U.S. Department of Energy. MRC-Mound is operated by Monsanto Research Corporation for the U.S. Department of Energy under Contract No. DE-AC04-76-DP00053. Brookhaven National Laboratory is operated for U.S. Department of Energy under Contract No. DE-AC02-76-CH00016.

REFERENCES 1. J. H. Wernick, in lntermetallic Compounds (J. H. Westbrook, ed.), p. 197. Krieger, Huntington, New York (1977). 2. D. Shaltiel, J. less-common Metals 62,407 (1978); I. Jacob and D. Shaltiel, J. less-common Metals 65, 117 (1979). 3. H. Buchner, in Hydrogen Energy Progress 1V, (T. N. Veziro~lu, W. D. Van Vorst and J. H. Kelley, eds.), p. 3. Pergamon, Oxford (1982). 4. M. Arita, N. Takashima, Y. Ichinose and M. Somena, Z. Metallk. 72,239 (1981). 5. F. Pourarian, H. Fujii, W. E. Wallace, V. K. Sinha and H. K. Smith, J. phys. Chem. 85, 3105 (1981); H Fujii, F. Pourarian, V. K. Sinha and W. E. Wallace, J. phys. Chem. 85, 3112 (1981). 6. I. Jacob and M. Polak, Mater. Res. Bull. 16, 1311 (1981). 7. J. Shinar, D. Davidov, D. Shaltiel and N. Kaplan, Z. phys. Chem. N.F. 117, 69 (1979). 8. J. Shinar, D. Davidov and R. D. Hogg, Solid State Commun. 37, 175 (1981). 9. J. R. Johnson and J. J. Reilly, lnorg. Chem. 17, 3103 (1978). 10. J. R. Johnson, J. less-common Metals 73,345 (1980). 11. J. R. Johnson, J. J. Reilly, F. Reidinger, L. M. Corliss and J. M. Hastings, J. less-common Metals 88, 107 (1982). 12. H. Y. Carr and E. M. Purcell, Phys. Rev. 94, 630 (1954); S. Meiboom and D. Gill, Rev. Sci. lnstrum. 29,688 (1958). 13. R. C. Bowman, Jr., A. Attalla and A. J. Maeland, Solid State Commun. 27,501 (1978). 14. R. M. Cotts, Ber. Bunsenges. Physik. Chem. 76, 760 (1972); R. M. Cotts, in Hydrogen in Metals l--Basic Properties (G. Alefeld and J. V61kl, eds. ), p. 227. Springer, Berlin (1978). 15. R. F. Karlicek, Jr. and I. J. Lowe, J. less-common Metals 73, 219 (1980); H. Chang, I. J. Lowe and R. J. Karlicek, Jr., in Nuclear and Electron Resonance Spectroscopies Applied to Material Science (E. N. Kaufmann and G. K. Shenoy, eds.), p. 331. Elsevier, Amsterdam (1981).

808

R. C. BOWMAN et al.

16. R. C. Bowman, Jr., D. M. Gruen and M. H. Mendelsohn, Solid State Commun. 32, 501 (1979); R. C. Bowman, Jr., B. D. Craft, A. Attalla, M. H. Mendelsohn and D. M. Gruen, J. less-common Metals 73, 227 (1980). 17. M. Goldman and L. Shen, Phys. Rev. 144, 321 (1966). 18. R. C. Bowman, Jr. and W. E. Tadlock, Solid State Commun. 32,313 (1979). 19. C. B. Magee, J. Liu and C. E. Lundin, J. less-common Metals 78, 119 (1981). 20. D. Fruchart, A. Roualt, C. B. Shoemaker and D. P. Shoemaker, J. less-common Metals 73, 363 (1980). 21. V. A. Yartys, V. V. Brunasheva, N. V. Fadeeva, S. P. Solov'ev and K. N. Semenenko, Soviet Phys. Dokl. 25, 888 (1980). 22. F. Reidinger, J. R. Johnson, P. Thompson. L. M. Corliss and J. M. Hastings, unpublished results. 23. J.-J. Didisheim, K. Yvon, D. Shaltiel and P. Fischer, Solid State Commun. 31, 47 (1979).

24. P. E. Mauger, W. D. Williams and R. M. Cotts, J. Phys. Chem. Solids 42, 821 (1981). 25. A. V. Irodova, V. P. Glaskov, V. A. Somenkov and S. Sh. Shil'sttein, Soviet Phys. solid St. 22, 45 (1980). 26. J.-J. Didisheim, K. Yvon, P. Fischer and P. Tissot, Solid State Commun. 38,637 (1981). 27. A. V. Irodova, O. A, Lavrova, G. V. Laskova and L. N. Padurets, Soviet Phys. solid St. 24, 22 (1982). 28. L. D. Bustard, R. M. Cotts and E. F. W. Seymour, Phys. Rev. B 22, 12 (1980), 29. E. F. W. Seymour, R. M. Cotts and W. D. Williams, Phys. Rev. Lett. 35, 165 (1975). 30. J. V61kl and G. Alefeld, in Hydrogen in Metals 1--Basic Properties (G. Alefeld and J. V61kl, eds.), p. 227. Springer, Berlin (1978). 31. G. D. Sandrock and P. D. Goodell, J. less-common Metals 73, 161 (1980).