Microstructure and hydriding characteristics of FeTi alloys containing manganese

Microstructure and hydriding characteristics of FeTi alloys containing manganese

Journal of the Less-Common Metals, 134 (1987) 275 275 - 286 MICRDSTRUCTURE AND HYDRIDING CHARACTERISTICS FeTi ALLOYS CONTAINING MANGANESE HIROSHI ...

1MB Sizes 0 Downloads 7 Views

Journal of the Less-Common

Metals, 134 (1987)

275

275 - 286

MICRDSTRUCTURE AND HYDRIDING CHARACTERISTICS FeTi ALLOYS CONTAINING MANGANESE HIROSHI

NAGAI,

KATSU

KITAGAKI*

and KEIICHIRO

SHOJI?

Department of Materials Science and Engineering, Faculty of Engineering, Un~uersity, 2-l Yamada-oka, Suita, Osaka 565 (Japan) (Received

October

15, 1986; in revised

form February

OF

Osaka

2, 1987)

Summary The effect of the substitution of manganese for iron, titanium or both iron and titanium in an FeTi alloy (i.e. Fei ~ ,TiMn, , FeTil _ XMn, and during the initial Fe1 -(x/2)331 -(x/2) Mn, (X = 0.1 - 0.3)) on the hydriding hydriding process, the pressure-composition isotherms and the microstructure was investigated. The partial substitution of manganese for iron, titanium or both iron and titanium yielded FeTi activation at 303 K after an incubation period without any activation treatment (with the exception of the Fe0_9TiMn0.1 alloy). Fe1 _ (x,2jTil ._(x,2jMn, and FeTi, _ xMnx alloys were composed of an FeTi phase and a second phase, (Fe, _,Mn,)i.sTi. Fe, __,TiMn, alloys were predominantly composed of an FeTi phase with small amounts of second phases, (Fe, _zMn,)1.5Ti and Ti,,,(Fe, _wMnw). However, after annealing, the second phases in the Fe1 _ XTiMn, alloys disappeared and an activation treatment at 673 K with a hydrogen pressure of 4 MPa was then required. The results obtained in this study showed that the presence of a second phase in the FeTi alloy played a crucial role in shortening the incubation time. The incubation time for the initial hydriding was dependent not only on the number of interfaces between the FeTi phase and the second phase, but also on the composition of the second phase, i.e. if the second phase readily reacts with hydrogen to form its hydride, even a small amount of second phase shortens the incubation period. Manganese substituted for iron in the FeTi phase and its composition was approximately represented by (Fe 1 _ yMn, )Ti. The plateau pressure decreased and the lattice parameter of the FeTi phase increased with increasing manganese content.

*Graduate student of Osaka University. Present address: Mitsubishi Tokyo 141, Japan. ‘Present address: Department of Metallurgy, Faculty of Science Kinki University, Higashi-Osaka 577, Japan. @ Elsevier

Sequoia/Printed

Metal Industries, and Technology,

in The Netherlands

276

1. Introduction It is well known that the intermetallic compound FeTi is a promising material for the storage and purification of hydrogen. However, the most significant problem of FeTi is that it is difficult to activate and activation requires a rather complica~d process. It must be subjected to high temperature annealing in vacuum or in hydrogen, or to a high hydrogen pressure over a long period of time [ 1,2]. It has been proposed [3 - 51 that the partial substitution of transition elements for iron in FeTi (i.e. an increase in the Ti/Fe ratio as compared with the ratio for a stoichiometric composition of FeTi) promoted the activation of FeTi during the initial hydriding process. However, the mech~ism by which activation was improved has not been fully elucidated. In our previous studies [6,7], it was found that activation of FeTi was promoted even if niobium and copper were partially substituted for titanium in FeTi (i.e. a decrease in the Ti/Fe ratio). It was also found from microstructural observations and analyses of the phases in the alloys with niobium and copper that the presence of the second phase played a crucial role in promoting the initial hydriding. In addition, it was noted that the copper substitu~d for the iron in the FeTi phase, whereas the niobium substituted for the titanium. In order to obtain an FeTi alloy with better performance and to obtain a better fundamental understanding of the relation between the microstructure, composition of the phases in the alloy and the characteristics of hydriding and hydrogen storage, the effect of the addition of manganese on the hydriding process, the hydrogen storage characteristics, and the microstructure of the FeTi alloy was investigated in this study.

2. Experimental

procedure

FeTi alloys, in which manganese was partially substituted for iron, titanium or both iron and titanium (ie. Fe1 _ ,TiMn, , FeTi, _ nMn, and pure ingot of Fe1 - WT~I - WZ)Mn,) were prepared from a commercially iron (purity, 99.9%), sponge tit~ium (purity, 99.5%) and m~g~ese (purity, 99.9%) by nonconsumable arc melting on a watercooled copper hearth in an argon atmosphere. The argon atmosphere was purified by melting a zirconium button and allowing it to react with residual oxygen and nitrogen. After initial melting, the ingots were reversed several times to ensure homogeneity. Each ingot was then mechanically pulverized in air to a size of -40 to +60 mesh. The effect of annealing at 1273 K for 72 ks on the microstructure, hydriding rate and hydrogen storage characteristics was also examined. Transverse sections of specimens were prepared by standard metallographic techniques and the compositions of the phases in the alloys were determined by energy dispersive X-ray (EDX) analysis.

The hydriding rate and pressure-composition isotherms were measured using a Sievert apparatus. All the measurements were carried out using the volumetric method. The alloy powders were loaded in the reaction tube which was sealed, evacuated and then filled with hydrogen to 4 MPa. The pressure changes were continuously monitored to determine the hydriding rate during the initial hydriding process. The pressure-composition isotherms for hydrogen absorption (desorption) were obtained by adding (withdrawing) a known amount of hydrogen to (from) the reaction system.

3. Results The amount of hydrogen absorbed us. time during the initial hydriding (activation) process at 303 K for Fe1 - .TiMn,, Fe1 _ (,,zjTil _ Cx,zjMn, and FeTi, _xMn, (x = 0.1 - 0.3) are shown in Figs. 1, 2 and 3 respectively. As was not activated at 303 K but recan be seen from Fig. 1, Fe,_,TiMn,.i quired an activation treatment at 673 K with a hydrogen pressure of 4 MPa. and Fe,.,TiMn, s were hydrided at 303 K after an incubation Fe,.sTiMn,., period. The incubation period required decreased with increasing manganese content. The amount of hydrogen absorbed decreased slightly with increasing manganese content. As can be seen from Fig. 2, Fe1 ~ Cx,ZJTil_ (x,ZjMnx (x = 0.1 - 0.3) alloys were hydrided at 303 K and the incubation period for hydriding decreased markedly with increasing manganese content. However, the amount of hydrogen absorbed decreased markedly with increasing manganese content. In the case of FeTi, _xMn, alloys, the incubation time decreased from 4 X lo3 s for x = 0.1 to 3 X lo2 s for x = 0.15, but increased again for higher manganese contents (1 X lo4 s for x = 0.2). It should also be pointed out

0.8

0.7 I 0.6 t z 0.5 = 0.4

-o-

TiFe0.g Mng.1

I

-A-

TiFeg.gMn0.g

%

-a-

0.3 0.2 0.1 -I

TiFe0.7Mn0.3

I

d

i ALO-.4L,-,,104 Time / set

Fig. 1. Amount of hydrogen absorbed 303 K for Fe1 _xTiMn, (x = 0.1 - 0.3).

I 105

1

during

the initial

hydriding

process

us. time

at

278

I CM-

-O-TiO~6FeO~g6MnO,l

OJ-

-~-TiO.g0FeOsMn~2

0.6~

--"-Tb.66Feo.asMnO.3

Time I set

Fig. 2. Amount of hydrogen 303 K for Fel-(,,2~Til_(,lz~Mn,

0.7 0.6

-o-

hydriding

process

us. time

at

process

us. time

at

TiO~gOFeMnO~l

--a-T1 -o-Ti

0

absorbed during the initial (x = 0.1 - 0.3).

AA

0.MFeM"0.15 0.80 FeMn0.2

,A' '1 10=

;

loTime

t set Fig. 3. Amount of hydrogen absorbed 303 K for FeTil_,Mn, (x = 0.1 - 0.3).

during

the initial

hydriding

that the amount of hydrogen absorbed decreased markedly with increasing manganese content. Comparing the effect of the addition of manganese on the shortening of the incubation period at the same manganese content in all the alloys, showed that the effect was more pronounced in the order FeTi, _XMn,, Fe, _ t,izjTi, _ c,,z,Mn, and Fe, _ ,TiMn, , with the exception of the Fe~~.sMn~.~ and Fe~~.,Mn~,~ alloys. The pressure-composition isotherms of Fe, - ,TiMn, , Fe1 _ (X,21Ti, _ (X,2jMn, and FeTi, _%Mn, alloys at 303 K are shown in Figs. 4, 5 and 6 respectively. The plateau pressure of Fe1 _,TiMn, decreased markedly

279

q

50

Cc)

-

10

1

,d

p”

-,M .rJ

Bt o. 0’ /’ I’

[email protected] /“;

i/

0

0.2

0.4 HIM

0.6

0.8

Fig. ‘4. Pre~ure-composition 0.2; (c) x = 0.3.

0

0.2

0.4 HIM

isotherms

0.6

0.8

0.1 0

for Fe, _,TiMn,

0.2

0.4 HIM

0.6

0.8

at 303 K: (a) x = 0.1; (b) x =

50

Cc) /i

10 _.Qd 1

fy ’

0.1 7 0 HIM

0.2

Fig. 5. Pressure-composition 0.1; (b) 1: = 0.2; (c)x = 0.3.

isotherms

Q4

0.6

for Fe;-fxi2~Til-(,,2fMn,

50

at 303 K: (a) x =

Cc) gB

10 +

~

1

0

0.2

0.4 HIM

0.6

0.0

Fig, 6. Pressure-composition 0.2; (c)x = 0.3.

0

0.2

0.4 HIM

isotherms

0.0

H/M

HIM

0.6

0.i 1

i 0.1 t 0

for FeTi,_,Mn,

0.2

0.4 HIM

0.6

I

0.i 3

at 303 K: (a) x = 0.1; (b) x =

with increasing manganese content, but the hydrogen storage capacity did not change significantly. In contrast, for the Fe1 _ Cx,ZjTil_ Cx,2jMnwand FeTii __xMn, alloys, the plateau pressure did not change significantly with manganese content, but the hydrogen storage capacity decreased markedly with increasing manganese content. In order to unde~t~d the cause of such large variations in hyd~ding rates (Figs. 1 - 3) and in the P-C-T curves (Figs. 4 - 6), microstructural

Fig. 7. SEM photographs of cross-sections of the arc-melted alloys: (a) Feo.sTiMno.2; phase; B: (Fel_,Mn,)1.STi (b) Feo.9Tio.sMnc.z; (cl FeTb.BMno.z. A: (FeI_,Mn,)Ti phase; C: Til.S(FeI_,Mn,) phase.

observations of the alloys and the analysis of the phases in the alloys were undertaken using scanning electron microscopy (SEM) and electron probe microanalysis (EPMA). As typical examples, SEM photographs of crosssections of the arc-melted FeasTiMn,.,, Fe,~,Ti,,Mn,~2 and FeTi0.sMn0.2 alloys are shown in Figs. 7(a), ‘7(b) and 7(c) respectively. All the alloys consisted of two or three phases. The compositions of these phases were determined by EDX and are plotted in Fig. 8. It is very interesting to note that the compositions of the phases change along the line of constant titanium content, i.e. the compositions of these phases can approximately be represented by Ti(Fe, _ YMn,), Ti(Fe, _tMnz)l.S and Ti,.,(Fe, _ wMn,). This means that manganese replaced iron in these phases.

281

Ti

0

FeagTi Mno.,

O

F~.8TiMno.z

*

Feo.7TiMno3

A

FeTio.9~0,,

A

FeTio 8 Mno2

A

FeTios7Mno,3

at %

Fig. 8. The compositions

of the phases in the alloys.

The lattice parameter of the FeTi phase and the plateau pressure are plotted against the manganese content in the FeTi phase in Fig. 9. It is apparent from this figure that there is a close relationship between them. The lattice parameter increased and the plateau pressure decreased with increasing manganese content in the FeTi phase.

2961 -0

2

4

6 Mn content

8 10 I at %

Fig. 9. Dissociation plateau pressure at 303 us. manganese content in the FeTi phase.

12

K and lattice

parameter

of the FeTi

phase

282

4. Discussion The effect of the substitution of manganese for iron, titanium or both iron and titanium in FeTi (i.e. Fe, _,TiMn,, Fe, -~x,Z~Til -C,..2jMn, and FeTii _ xMnx) on the hydriding process, the hydrogen storage ch~cte~sti~s and the microst~cture of the alloys was investigated in this study. One of the most significant results was that all the alloys with the exception of Feoe9TiMn, 1 were activated at 303 K after an incubation period without any special heat treatment. However, the incubation period and the hydrogen storage characteristics (i.e. plateau pressure and storage capacity) were very dependent on the manganese content and on the element (iron, titanium or both iron and titanium) for which manganese was substituted. In order to clarify the cause of such large variations, the amounts of phases in the alloys are plotted against the addition of manganese in Figs. 10(a), 10(b) and 10(c). In the case of Fel_,TiMn, alloys, the FeTi phase was predominant and slightly decreased with the addition of manganese. The amount of the second phases increased up to 10% for x = 0.3. In contrast, the amount of FeTi phase in the Fe, _Cxf21Ti1_ o..zjMn, alloys decreased markedly and the amount of the second phase ((Fe1 _yMn,)l.5Ti phase) si~~~c~tly increased with the addition of manganese. The decrease in the amount of the FeTi phase with the addition of manganese was more pronounced in the case of FeTi, _xMn, than for Fe1 _ (x,ZjTil - (xlzjMnx, i.e. as can be seen in Fig. 10(c), for FeTi,_.Mn, alloys, the amount of the second phase (Fe, -,Mn,),.,Ti markedly increased with the addition of manganese and the alloy almost became single phase (Fe, _yMny)1.5Ti at x = 0.3. The change in the amount of the FeTi and Fel.sTi phases with the addition of manganese was u~de~tandable when one considers eqns. Mn, and FeTir _ %Mnx alloys respectively. (1) and (2) for Fe1 - (X/2)Tii -it

x

X

Fig. 10. Incubation period and amounts of FeTi and second phases in the alloys vs. the addition of manganese: -O-, amount of FeTi phase (measured);--+-, amount of second phase (measured); - - -, amount of FeTi phase (calculated); -O-, incubation time.

283

Fe I tx/2)Ti 1.. ~12) + xMn = { 1 - (5x/2))(Fei

- ,Mn,)Ti

+ Bx(Fei _ ,Mn,),.,Ti

(1)

y = 2x/(x + 1) FeTii --x + xMn = (1 - 5x)(Fe,

_,,Mn,)Ti

+ 4x(Fe,

_ ,Mn,)l.sTi

(2)

y = x/(x + 1) The manganese concentration (y value) in the FeTi and FeiSSTi phases were assumed to be equal. The amounts of FeTi phase calculated from eqns. (1) and (2) are plotted in Fig. 10 as broken lines. The calculated values are very close to the amounts of FeTi phase measured from microstructural observation. In the case of the Fe, _ ,TiMn, alloys, they are expected to be single phase FeTi because manganese substitutes for iron in the FeTi phase. However, small amounts of (Fe,_,Mn,)i.sTi and Ti,.5(Fe1_,Mn,) phases were detected in the alloys. This was considered to be due to the fact that the arc-melted alloys were not at equilibrium and the (Fe, - ZMn,)i.gTi and Ti,.,(Fer _ wMnw) phases were retained in the alloys. This was confirmed by the almost total disappearance of these second phases when the alloys were annealed at 1273 K for ‘72 ks. After this treatment alloys of almost single phase were obtained. The reaction should be written as eqn. (3), (Fe, _ZMnz),,,Ti

+ Ti,_,(Fe,

_ WMnw) = (5/‘2)(Fe, _ ,Mn,)Ti

(3)

.z = y/2, w = 2y The annealed Fe, _,TiMn, alloys (x = 0.1 - 0.3) were not activated at 303 K but required an activation treatment at 673 K with a hydrogen pressure of 4 MPa. This result strongly suggests that the presence of a second phase plays a crucial role in promoting the initial hydriding process. The incubation time decreased with increasing amounts of a second phase in the alloys with the exception of FeTi, _ xMnx (x = 0.2,0.3), as can be seen in Fig. 10(c). This result also indicates that the second phase and/or the interface between the FeTi phase and the second phase act as entrance sites for hydrogen diffusion into the alloy and/or active sites for hyd~ding as previously proposed [7]. In the case of FeTi, xMnr alfoys, as seen in Fig, 10(c), the incubation period increased again at x values higher than 0.15. This may be due to a decrease in the number of interfaces between the FeTi phase and the second phase because of the large increase in the amount of the second phase. However, it is clear from Figs. 10(a), 10(b) and 10(c) that the incubation time is not solely dependent on the number of interfaces. In fact, the incubation time for Fe, _.%TiMn, (x = 0.3, 4 X lo2 s) was shorter than those for Fe, _ Cx,2jTil ~ (r,2tMn, (x = 0.2, 3 X lo3 s) and FeTii _ ,Mn, (x = 0.1, 4 X lo3 s) in spite of the fact that the amount of the second phase in Fe1 _,TiMn, (x = 0.3, approximately 10%) was far less than the amounts in Fe, _ o..2jTil _ fx,2jMn, (x = 0.2, approxima~ly 50%) and FeTii _ xMnx (x = 0.1, approximately 45%). This may mean that

284

composition of the second phase and/or FeTi phase also affects the initial hydriding rate (incubation period). The annealed Fe, _,TiMn, alloys (LX= 0.1 - 0.3), which were mostly composed of single phase FeTi, did not yield activation at 303 K but an activation treatment at 673 K with a hydrogen pressure of 4 MPa was needed. However, the compositions of the FeTi phase in the annealed Fe1 _ xTiMn, alloys were almost the same as those in the arc-melted alloys. This result means that the incubation time depends on the composition of the second phase and not on the composition of the FeTi phase. Moreover, as can be seen from Fig. 8, the manganese content in the (Fe1 ~YMn,)I.sTi phase is higher in the Fe1 _,TiMn, alloys than in of the the Fe1 - (,/z)Til - (x/2)Mn, and FeTi, _ xMnx alloys. The composition (Fe,_,Mn,),.sTi phase approached TiMn,., in each alloy with increasing addition of manganese. The TiMnr.s alloy was reported [8] to react readily with hydrogen to form TiMn,.,H,.,, at room temperature without any activation treatment. According to the results, it may be concluded that the presence and composition of the second phase plays a crucial role in the shortening of the incubation time for the initial hydriding of the ahoy. Even if the second phase does not react with hydrogen, the interfaces between the FeTi phase and the second phase act as entrance sites for hydrogen diffusion into the alloy and/or as active sites for hydriding. An increase in the number of interfaces promotes the initial hydriding. However, if the second phase reacts readily with hydrogen, the presence of even small amounts of second phase will significantly promote hydriding during the initial hydriding process, as is the case with Fe, _,TiMn, alloys (see Figs. 1 and 10(a)). The amount of hydrogen storage capacity shown in Figs. 4, 5 and 6 corresponds to the amount of FeTi phase in the alloys shown in Fig. 10. When manganese is added to stoichiometric FeTi and iron-rich FeTi alloys, and FeTi, _xMn, alloys, very large amounts of Mn, i.e. Fe 1- ~,12~Til - (xl2) the second phase form and the amount of FeTi phase in the alloys decreases markedly according to eqns. (1) and (2). The resulting hydrogen storage capacity of the alloys is also decreased significantly.

5. Conclusions The effect of partial substitution of manganese for iron, titanium or both iron and titanium in FeTi (i.e. Fe1 _,TiMn,, FeTir _xMn, and Mn,) on the hydriding rate during the initial hydriding Fe1 - (x,2,Til - (x/2) process, pressure-composition isotherms and microstructure of the alloys was investigated. The following results were obtained. (1) The arc-melted Fe1 _,TiMn, (x = 0.2, 0.3) alloys were activated at 303 K after an incubation period without any activation treatment, whereas the Fe,.9TiMn,. , alloy was not activated at 303 K but required an activation treatment at 673 K and a hydrogen pressure of 4 MPa. Fe1 _,TiMn, alloys were predominantly composed of an FeTi phase with

small amounts of second phases of (Fe, ~,Mn,),.,Ti and Ti,.,(Fe, zMnz). The amount of the second phases increased and the incubation period decreased with increasing addition of manganese. After prolonged annealing, most of the second phases in the alloys disappeared and the activation at 303 K became impossible for all Fe1 _,TiMn, alloys. Activation treatment at 673 K with a hydrogen pressure of 4 MPa was then required. This result means that the presence of a second phase in the alloy played a crucial role in promoting the initial hydriding process. (2) Fe1 Cx/2)Til ~2) Mn, (3~= 0.1 - 0.3) alloys were also activated at 303 K after an incubation period. The incubation time decreased markedly with increasing addition of manganese. The alloys were composed of FeTi and (Fe, ),Mn,) ,.5Ti phases and the amount of the latter phase increased markedly with increasing addition of manganese. The hydrogen storage capacity decreased significantly with an increasing amount of the latter phase. (3) FeTi, _ wMnx (x = 0.1 - 0.3) alloys were also activated at 303 K after an incubation period. The incubation time was shortest at x = 0.15 and increased again at x values higher than 0.2. The alloys were also composed of FeTi and (Fe, yMny),.sTi phases and the amount of FeTi phase decreased significantly with increasing addition of manganese. The decrease in the amount of the FeTi phase was more pronounced in these alloys than Mn, alloys. The hydrogen storage capacity dein the Fe1 - CXlzjTir (Xj2) creased significantly with increasing addition of manganese. (4) The composition of the FeTi phase in the alloys changed along the constant titanium content line and was approximately represented by (Fe, ~~,Mn,)Ti. The manganese concentration (y value) in this phase increased with addition of manganese. This means that manganese substitutes for iron in the FeTi phase. The plateau pressure decreased and the lattice parameter of the FeTi phase increased with increasing manganese content in the FeTi phase. (5) The results obtained in this study show that the presence of a second phase in FeTi alloys plays a crucial role in the shortening of the incubation time. The incubation time for the initial hydriding seems to be dependent on not only the number of interfaces between the FeTi phase and the second phase, but also on the composition of the second phase, i.e. if the second phase readily reacts with hydrogen to form its hydride, even a small amount of this phase decreases the incubation time. References J. J. Reilly and R. H. Wiswall, Jr., Inorg. Chew, 13 (1974) 218. M. Amano, Y. Sasaki and T. Matsumoto, J. Jpn. Inst. Met., 43 (1979) 809. M. Amano, Y. Sasaki and T. Yoshioka, J. Jpn. Inst. Met., 45 (1981) 957. J. J. Reilly and J. R. Johnson, Proc. 1st World Hydrogen Energy Conf., Miami Beach, FL, Vol. 2, Pergamon, Oxford, 1976, p. 3. 5 Y. Sasaki and M. Amano, Proc. 3rd World Hydrogen Energy Conf., Tokyo, Japan, Vol. 4, Pergamon, Oxford, 1980, p. 891.

1 2 3 4

286 6 H. Nagai, M. Nakatsu, K. Shoji and H. Tamura, Technical Rep., 35, No. f 786, Osaka University, 1985, p. 37. 7 H. Nagai, M. Nakatsu, K. Shoji and H. Tamura, J. Less-Common Met., 119 (1986) 131. 8 ‘I’. Yamashita, T. Gamo, Y. Moriwaki and M. Fukuda, J. Jpn. Inst. Met., 41 (1977) 148.