Manufacture and hydriding characteristics of unidirectionally solidified LaNi5Ni eutectic alloys with disintegration resistance

Manufacture and hydriding characteristics of unidirectionally solidified LaNi5Ni eutectic alloys with disintegration resistance

Journal of the Less-Common 143 Metals, 138 (1988) 143 - 154 MANUFACTURE AND HYDRIDING CHARACTERISTICS OF UNIDIRECTIONALLY SOLIDIFIED LaNis-Ni EUTEC...

1MB Sizes 0 Downloads 1 Views

Journal of the Less-Common

143

Metals, 138 (1988) 143 - 154

MANUFACTURE AND HYDRIDING CHARACTERISTICS OF UNIDIRECTIONALLY SOLIDIFIED LaNis-Ni EUTECTIC ALLOYS WITH DISINTEGRATION RESISTANCE T. OGAWA and K. OHNISHI The Japan Steel Works, Ltd., Muroran (Japan) T. MISAWA Department

of Metallurgical Engineering,

Muroran Institute of Technology,

Muroran

(Japan) (Received March 9,1987;

in revised form July 22, 1987)

Summary Disintegration of metal hydride powder due to repeated hydriding and dehydriding cycles has been a serious problem which has retarded the practical application of metal hydrides. In order to solve this problem, unidirectional solidification was applied to LaNis-Ni eutectic alloys of chemical compositions ranging from hypoeutectic to hypereutectic. Examination of the disintegration resistance and hydriding characteristics revealed that these unidirectionally solidified alloys possess better resistance to disintegration than the as-arc-melted alloys of similar compositions after 30 cycles of hydriding and dehydriding. Disintegration resistance is enhanced as the amount of ductile nickel phase increases because the development of microcracks is arrested and the volume expansion is relaxed by the nickel phase. These LaN&-Ni alloys had higher equilibrium pressures and lower hydrogen concentrations compared with those of LaNis powder. The results obtained are discussed in terms of the effects of lattice strain in hydrided LaNi, and the amount of LaNi,.

1. Introduction In recent years, various chemical, thermal and mechanical systems to which metal hydrides are applied have been proposed and discussed. These systems include hydrogen storage tanks [ 11, hydrogen separators and purifiers [2], heat pumps [3] and actuators [4]. Although these systems have demonstrated the advantages of metal hydrides, it is known that metal hydrides are so brittle that they tend to fall to pieces after hydridingdehydriding cycles. (Henceforth, this phenomenon will be referred to as disintegration although various terms such as decrepitation or atomization 0022-5088/88/$3.50

@ Elsevier Sequoia/Printed in The Netherlands

144

have also been used for the same phenomenon.) Disintegrated powder flows around in the system with the hydrogen gas causing mechanical troubles. Moreover, disintegration decreases the efficiency of the equipment because of the lower thermal conductivity of fine metal hydride powder. Measures to counter disintegration which have been proposed to date, such as addition of small amounts of alloying elements [5], coating with copper over the metal hydride particles [6], encapsulation of the metal hydride powder [ 71, mixing with silicone oil [8] and the adoption of metal hydride thin films [9] have some inevitable disadvantages. For example, coating treatments of metal hydride powder complicate the manufacturing processes of the material, and the use of thin films is, in many cases, limited by the shape and size of the film. One of the ways to eliminate these disadvantages is to develop a microstructure resistant to disintegration. Especially, a fibrous two-phase microstructure, similar to fibrous composite materials, could possess enhanced ductility and hence improved disintegration resistance. In the present work, the unidirectional solidification technique was applied to alloys in the LaNi,-Ni eutectic system in order to produce a fibrous microstructure consisting of brittle LaNis and ductile nickel. This paper will report the microstructural variation, disintegration resistance and hydrogen absorption characteristics of the unidirectionally solidified alloys.

2. Experimental

details

The nickel-rich portion of the La-Ni binary phase diagram obtained by Buschow and van Mal [IO] is shown in Fig. 1. The eutectic between the LaNi,-Ni phases is at 1270 “C and 93 at.% Ni. On the basis of this phase diagram, the chemical compositions of alloys used for the present study were chosen to range from the hypoeutectic to the hypereutectic. The

I600

I

1

I

I

LaNi,

1

6oOe0

I

I

I

I

I

1

84

66

92

96

100

at. % Ni

Fig. 1. The La-Ni

phase diagram

from 80 to 100 at.% Ni (from

ref. 10).

145

alloys were prepared by means of arc melting under a purified argon atmosphere using 99.9 mass% lanthanum and 99.8 mass% nickel as raw materials. (Hereafter these alloys will be called “primary melts”.) Unidirectionally solidified alloys were manufactured from rectangular specimens of about 1 cm X 1 cm X 10 cm which were cut from the primary melts. Unidirectional solidification was carried out using the apparatus shown in Fig. 2. The specimen was loaded in an AlzOs crucible, which was sealed in a quartz tube under an atmospheric pressure of purified argon. The maximum temperature at the centre of the furnace was maintained at 1400 “C throughout the melting; the temperature gradient of the specimen was held at 40 “C cm-’ and was obtained by cooling the bottom surface of the specimen by contact with a water-cooled copper pipe. The specimen was lowered at a rate between 1.3 X 10P4 and 1.3 X lop3 cm s-l. The chemical compositions of the LaNis-Ni alloys before and after unidirectional solidification are summarized in Table 1. The lanthanum and nickel concentrations of each unidirectionally solidified piece of ahoy showed little variation in comparison with the corresponding primary melts. However, the unidirectionally solidified alloys had a higher concentration of oxygen, caused by contamination from the AlzOs crucible in the remelting. The LaNi, alloy, prepared as a reference material, had an almost stoichiometric composition as shown in Table 1 and was confirmed by X-ray diffraction to be composed of a single LaNiS phase with lattice parameters a = 5,022 and 3.990 a. Hydriding-dehydriding characteristics of the alloys were measured using hydrogen gas of purity 99.999 99 mol.% with a dew point of -70 “C. To temperature

+Ar *water

controler

gas

outlet

outlet

W0tbr Inlet Fig. 2. A schematic illustration of the unidirectional solidification apparatus.

146 TABLE 1 Chemical compositions

of the LaNiS-Ni

Preparation

System

Primary melts

(As arc melted)

Unidirectional

alloys investigated

solidification

Alloy

Chemical composition (at.%) LU

Ni

0

Single phase

LaNis

16.7

83.2

0.W

HypUeuteftic

A B C

11.2 9.5

a.8

83.5 90.2 91.0

0.09 0.0?3 0.08

Eutectic

D E

7.6 6.5

92.0 93.2

0.04 0.07

Hypereutectic

F G

5.8

94.0

4.9

95.0

0.05 0.06

H

IO.6

88.9

I

9.2

90.4

0.36 0.41

J

8.5

91.2

0.33

Eutectic

K L

7.3 6.3

92.4 93.5

0.48 0.43

Hypereutectic

M N

5.5 5.1

94.4 95.0

0,37 0.40

Hypoeutectic

The specimens with diameter 1 cm and thickness 05 cm were taken from both primary melts and unidirectionally solidified alloys. The surfaces of each specimen were finished with 1200 emery paper. The disc-shaped specimens were cleaned with acetone and placed in a reactor cell. The alloy was then activated by vacuum heating for 1.8 X 10’s at 80 “C. Then, the specimen was repeatedly exposed to hydrogen under an initial pressure of 40 atm at 20 “C for 7,2 X 104 s for hydriding, f&owed by evacuation at 80 “C for 1.8 X IO3 s for dehydriding. Ten cycles of this hyd~d~g-dehyd~ding were needed in order to activate the specimens of primary melts, whereas all the specimens taken from the unidirectionally solidified alloys were easily activated after only four cycles. Absorption-dissociation isotherms were determined at 20 *C with automatic volumetric equipment for specimens that had undergone 30 hydriding-dehydriding cycles. Each hydriding or dehydriding reaction was regarded as reaching equilibrium when 1.2 X 10’ s had elapsed after inducting or extracting hydrogen respectively. The changes in hydrogen concentration were calculated by taking into account the deviation from ideal gas behaviour. The disintegration resistance was assessed by observing whether the original shape of the disc was maintained,

147

3. Results and discussion 3.1. Morphology and disin tegm tion resistance The morphological changes in the unidirectionally solidified LaNis-Ni alloys manufactured in this work were plotted in Fig. 3. The morphologies of the unidirectionally solidified LaNi,-Ni alloys could be classified into three groups: columnar, mixed and irregular. Mallard and Flemings [ll] have proposed that the critical condition for dendritic growth of off-eutectic alloys is expressed by the following equation: G _= -m(C, -C,) D R

(1)

where R is the growth rate, G is the temperature gradient, m is the slope of liquidus line at the eutectic point, C, and C,, are the compositions of the eutectic and of the starting alloy respectively, while D is the liquid diffusion coefficient. Although the liquid diffusion coefficient of lanthanum is unknown, the values of the selfdiffusion constants of liquid metals are in the range from 2 X 10e5 to 8 X 10V4 cm2 6’ [12,13]. Taking a value of 2 X 10e5 cm2 s-r, which would result in the widest region of columnar microstructure, the boundaries of the columnar-dendritic transition were estimated as indicated by the broken lines in Fig. 3. Although the region of columnar growth estimated by eqn. (1) was very narrow in the vicinity of the eutectic composition C, = 93 at.% nickel, the columnar structure was actually obtained over a wider range of nickel content and G/R ratio. A set of typical microstructures is shown in Fig. 4 for unidirectionally solidified LaNi,-Ni alloys manufactured at a growth rate R of 3.9 X 10e4 cm

ot . % N i

Fig. 3. The morphology of the solidified microstructures of unidirectionally solidified LaNi5-Ni alloys as a function of G/R and Ce (i.e. nickel content): 0, columnar structure; A, mixed structure of columnar matrix with primary phase; 0, irregular structure. Broken lines show the composite-dendritic boundary predicted by simple constitutional supercooling criteria.

,2OQm,

Eutectic (Alloy K)

50 flrn

(Alloy

Hypereutectic N)

200flm

Fig. 4. Typical microstructures of the unidirectionally solidified LaNirNi alloys formed with a growth rate of R = 3.9 x 10V4 cm 6-l. The upper row shows transverse cross-sections and the lower row shows longitudinal cross-sections along the growth direction.

Hypoeutectic (Alloy HI

E

149 s-l.

The eutectic alloy of 92.4 at.% nickel has an almost perfectly columnar eutectic structure; the greyish LaNi, phase and the white nickel phase can be seen in the optical micrograph. The hypoeutectic alloy of 88.9 at.% nickel has a columnar LaNi,-Ni eutectic matrix with large spherical regions of the primary LaNi5 phase. The structure of the hypereutectic alloy of 95 at.% nickel, however, consists of a LaNi,-Ni eutectic matrix and large spherical regions of the primary nickel phase. All the disc-shaped specimens maintained their bulk shape after 30 cycles of hydriding-dehydriding, proving good disintegration resistance. Figure 5 demonstrates the superiority in disintegration resistance of the unidirectionally solidified alloys as compared with the primary melts. Experience has indicated [14] that the probability of disintegration is mostly determined by the first few hydriding-dehydriding cycles. Therefore, the fact that these unidirectionally solidified alloys can maintain their original shapes after 30 hydriding-dehydriding cycles confirms the applicability of these alloys for practical purposes which require greater numbers of hydriding cycles. Among the unidirectionally solidified specimens, those with a nickel content ranging from eutectic to hypereutectic compositions had good disintegration resistance in spite of the fact that their volume expansion ranged from about 30% in the eutectic ahoy to 15% in the hypereutectic alloy with 95 at.% nickel. The hypoeutectic specimen in Fig. 5 decomposed into flakes at the disc surface, showing slightly inferior disintegration resistance. This is considered to be due to a smaller amount of the nickel phase. Hypoeutectic

Eutectic

Hypereutectic (Alloy 6)

(Alloy Ii)

IOmm Fig. 5. The appearance of the LaNirNi alloys after 30 hydriding-dehydriding cycles. The upper row shows the primary melts and the lower row shows the unidirectionally solidified alloys.

150

The micros~~tures are shown in Fig. 6 of the hy~d~-dehydrided specimens for the unidirectionally solidified alloys shown in Fig. 5. In the hypoeutectic specimen, many microcracks were observed not only in the columnar LaNi, phase of the LaNis-Ni eutectic matrix but also in the spherical regions of the primary LaNi, phase. In the eutectic specimen, the number of microcracks was smaller, and their propagation was confined within the LaNi, phase of the columnar eutectic matrix. Moreover, it was observed in the hypereutectic specimen that a few microcracks propagated in the LaNis phase of the columnar matrix were stopped around the spherical primary nickel phase. It can be concluded that the ductile nickel phase in the LaNi,-Ni alloys acts to arrest microcracks originating in the LaNiS phase and relaxes the strain induced by the volume expansion of the LaNis phase caused by hydrogen absorption. 3.2. A bsorp tion-dissocia tion characteristics The absorption-dissociation isotherms at 20 “C for both the primary melts and the unidirectionally solidified alloys after thirty cycles of hyd~d~g-dehyd~d~g are shown in Fig. 7, The dotted lines show the results for LaNi, powder for comparison. The absorption equilibrium pressure at the middle of plateau in the LaNi,-Ni alloy isotherm rises in the order hypoeutectic, eutectic, then hypereutectic. The absorbed hydrogen concentration (expressed by mass%) in the alloys decreases with decreasing amount of LaNi,, i.e. the hydride former. However, the hydrogen concentration expressed by [H] /[ La] in the alloy is equal to that ([HI/[ La] = 6) of LaNiS powder for the unidirectionally solidified alloys and is lower for the primary melts. The absorption pressure at the middle of plateau and the hydrogen concentration at 40 atm are plotted in Fig. 8 as a function of nickel content for all the alloys studied, The absorption equilibrium pressure of the unidirectionally solidified alloys increases linearly with increasing nickel content. Ex~apolation of the line to the low nickel content side corresponds to the pressure of the LaNis alloy. The absorption pressure of the primary melts is higher than that of the unidirectionally solidified alloys. Moreover, it seems that the absorption pressure of the primary melts increases up to the eutectic composition but then remains constant in the hypereutectic region. The dependence of absorption pressure on nickel content is explained in terms of the effects of lattice strain and non-stoichiometric LaNi,. As hydriding proceeds, volume expansion occurs in the LaNiS phase but not in the nickel phase. This provides a constraint against volume expansion, resulting in lattice strain. A scanning electron micrograph of the cross-section of unidirectionally solidified hypereutectic alloys (95 at.% Ni) is shown in Fig. 9, along with profiles of the surface roughness before and afjer the first hydriding at 20 “C. The surface of the matrix with a eutectic microstructure rises by about 5 X lop4 cm compared with the remaining primary nickel phase because of the volume expansion of the hydrided LaNi, phase in the matrix. The ex-

152

2

3

& IO

i! i! j/

5 E a _______-_._a-.----_._______._._._._._._.-._L L LaNi,

g 5

I

F b f 0.5

I’

0.11

0

Primary

/‘I

!; !,I ______._._~-.-_-.-_

1

melts

Unidirectionally

0.

(Alloy

A)

A)

(Alloy

D)

q m (Alloy

C)

i

0. I 0

5

0.4 Hydrogen

(a)

solidlfled

o. (Alloy AA (Alloy

H) K)

q m (Alloy

N)

0.8 concentmtion

alloys

1.2

I.1

boss%)

(b)

Fig. 7. Hydrogen absorption (open symbols) and dissociation (filled symbols) isotherms of the LaNis-Ni alloys at 20 “C: (a) primary melts (as-arc-melted alloys); (b) unidirectionally solidified alloys.

pansion of the LaNi, phase is allowed only at free surfaces, and a significant amount of lattice strain would occur within the matrix. A higher absorption pressure is thus required in order to overcome this lattice strain. This explains the monotonic increase in absorption pressure with nickel content for the unidirectionally solidified alloys. In the case of primary melts, the effect of non-stoichiometric LaNi, appears to be more dominant. Because of the high cooling rate after arc melting, the primary melts are considered to contain the non-stoichiometric LaNi, phase, where x is greater than 5, as reported by Buschow and van Mal [lo]. They found that the equilibrium pressure of the LaNi, powder rose sharply with increasing X. The region of the non-stoichiometric LaNi, is shown schematically by hatching in Fig. 1. In the hypoeutectic region, the value of x for the primary LaNi, phase increases with nickel content. Therefore, the absorption pressure increases in this region. However, the LaNi, phase exists only in the matrix at and above the eutectic composition, therefore, x remains constant at 5.5 in this region. This rationalizes the invariant absorption equilibrium pressure above the eutectic composition. The absorbed hydrogen concentration of both the primary melts and the unidirectionally solidified alloys decreases linearly from that of LaNi, powder with increasing nickel content as shown in Fig. 8. This decrease in hydrogen concentration with nickel content in the LaNi,-Ni system depends primarily on the decrease in the amount of the hydride-forming LaNis phase. The columnar structure of the unidirectional alloys could make it easier for hydrogen to diffuse compared with the three-dimensional composite structure, which would disrupt the long-range diffusion of hydrogen in the primary melts. This seems to be the reason that the absorbed hydrogen

153 I

I

Nonstoichiometric

4 _ Strain

1

I

LaNix effect

I ,I I I Eutectic

effect

*-

I

I

I

I .6’k

l*2ri\

i

LaNie 0

0.8 -

%.. \

\

%. 0

0.4\ ’

84

.\

0

I

88

92 Content

\

00

I

Ni

l

0

0

\,

96

100

(at.KNi)

8. Changes in absorption pressure and hydrogen concentration as a function of nickel content in the LaNi5-Ni alloys at 20 “C: 0, primary melts (as-arc-melted alloys); 0, unidirectionally solidified alloys; 0, LaNis powder.

unhydrided J

(a)

(b)

Fig. 9. Surface topography of the hydrided LaNis-Ni hypereutectic alloy: (a) topographical micrograph; (b) change in surface roughness after the first hydriding at 20 “C.

154

concentration of the unidirectionally solidified alloy is much larger than that of the primary melt for all the chemical compositions examined. 4. Conclusion Unidirectionally solidified LaNis-Ni eutectic alloys were successfully prepared. Their manufacture was based on the idea of having a fibrous composite consisting of a brittle hydride-fo~~g LaNi, phase and a ductile nickel phase. It was found that these LaNi,-Ni alloys had improved disintegration resistance after 30 cycles of hydriding and dehydriding, with hydrogen concentrations higher than those for primary melts of the same chemical compositions.

Acknowledgments The authors are grateful to Mr. H. Kakihara of The Japan Steel Works, Ltd. for his helpful cooperation with the specimen preparation and measurements. We would also like to thank Dr. T. Momono of Muroran Institute of Technology for his valuable discussions.

References 1 G. Strickland and J. J. Reilly, BNL Rep. 50421, Brookhaven National Laboratory, Upton, NY, 1974. 2 J. J. Reilly, 2. Phys. Chem., 117 (1979) 655. 3 P. P. Turillon, J. Less-Common Met., 74 (1980) 332. 4 T. Sasaki, M. Kawashima, T. Ifukube, H. Aoyama and T. Ogawa, J. Robotics SOC. Jpn., 4 (1980) 119. 5 J. R. Johnson and J. J, Reilly, Miami Int. Conf. on Alternative Energy Sources, 8 (1978) 3739. 6 H. Ishikawa, K. Oguro, A. Kato, H. Suzuki and E. Ishii, J. Less-Common Met., 107

(1985) 105. Report: Znco Tee. Paper 11 I1 -T-OP, Into, 7 G. D. Sandrock, Hydrogen Technology Suffern, NY, 1981. 8 Y. Ohsumi, H. Suzuki, A. Katou, K. Oguro and M. Nakane, J. Chem. SOC. Jpn.,

(1981) 1493. K. Aoki and K. Masumoto, Bull. Jpn. Inst. Met., 23 (1984) 8bB K. H. J. Buschow and H. H. van Mal, J. Less-Common iUet., 29 (1972) 203. F. R. Mollard and M. C. Flemings, Trans. Metall. Sot. AZME, 239 (1967) 1526. J. Szekely and N. J. Themelis, Rate Phenomena in Process Metallurgy, Wiley, New York, 1971, p. 368. 13 G. H. Geiger and D. R. Poier, Tmnsport Phenomena in Metallurgy, Addison-Wesley, London, 1973, p. 455. 14 H. Uchida, H. Uehida and Y. Huang, J. Less-Common Met., 101 (1984) 459.

9 10 11 12