Effect of oxygen on the microstructure and hydrogen storage properties of V–Ti–Cr–Fe quaternary solid solutions

ww.sciencedirect.com

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 2 0 0 0 0e2 0 0 0 8

Available online at w

ScienceDirect

journal homepage: www.elsevier .com/locate/he

Effect of oxygen on the microstructure andhydrogen storage properties of VeTieCreFequaternary solid solutions

Ulrich Ulmer a,*, Kohta Asano b, Thomas Bergfeldt c,Venkata Sai Kiran Chakravadhanula d,e, Roland Dittmeyer f,Hirotoshi Enoki b, Christian Kubel a,d,e, Yumiko Nakamura b,Alexander Pohl a, Maximilian Fichtner a,d

a Karlsruhe Institute of Technology (KIT), Institute of Nanotechnology, P.O. Box 3640, D-76021 Karlsruhe, Germanyb National Institute of Advanced Industrial Science and Technology (AIST), AIST Central-5, 1-1-1 Higashi, Tsukuba,

Ibaraki 305-8565, Japanc Karlsruhe Institute of Technology (KIT), Institute of Applied Materials, P.O. Box 3640, D-76021 Karlsruhe, Germanyd Karlsruhe Institute of Technology (KIT), Helmholtz Institute Ulm for Electrochemical Energy Storage,

Helmhotzstr. 11, 89081 Ulm, Germanye Karlsruhe Nano Micro Facility, Hermann-von-Helmholtz Platz 1, 76344 Eggenstein-Leopoldshafen, Germanyf Karlsruhe Institute of Technology (KIT), Institute of Micro Process Engineering, P.O. Box 3640, D-76021 Karlsruhe,

Germany

a r t i c l e i n f o

Article history:

Received 12 June 2014

Received in revised form

8 August 2014

Accepted 19 August 2014

Available online 22 October 2014

Keywords:

BCC alloys

Hydrogen storage

Oxygen effect

* Corresponding author. Tel.: þ49 (0) 721 608E-mail address: [email protected] (U.

http://dx.doi.org/10.1016/j.ijhydene.2014.08.10360-3199/Copyright © 2014, Hydrogen Energ

a b s t r a c t

The effect of low (<300 ppm O) and high (10,000 ppm O) residual oxygen concentration in

vanadium raw metals on the microstructure and hydrogenation properties of V40Fe8T-

i26Cr26, was investigated by means of XRD, SEM, TEM and pressure-composition isotherms.

A high oxygen concentration in the vanadium raw metal led to the formation of an oxygen-

rich secondary phase isostructural with a-Ti. The lattice parameter of the BCC main phase

of the high-oxygen sample was reduced to 3.0141 (3) A compared to 3.0308 (2) A for the low-

oxygen sample. As a result of the high oxygen content the equilibrium hydrogen pressure

of the material was increased from 1 MPa to 4 MPa. Deoxidization through the addition of

1 at% rare earth metal could be achieved. The lattice constant of the deoxidized sample

was 3.0297 (3) A, and the thermodynamic properties were also the same as in case of the

low-oxygen sample.

Copyright © 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights

reserved.

22680; fax: þ49 (0) 721 608 26368.Ulmer).52y Publications, LLC. Published by Elsevier Ltd. All rights reserved.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 2 0 0 0 0e2 0 0 0 8 20001

Introduction

The utilization of hydrogen storage materials provides a

compact, safe, and energy efficient method of hydrogen

storage for stationary and mobile applications [1]. For sta-

tionary hydrogen storage applications, the gravimetric

hydrogen density is less crucial than for mobile applications,

however, candidate hydridesmust satisfy a number of criteria

for successful utilization such as H-storage capacity >1 mass

%, low cost (both materials and processing), fast kinetics,

durability during pressure/temperature swing cycles, safety

and low toxicity, resistance to contamination and common

impurities and minimal demands for hydride sorbent bed

activation [2].

The volumetric hydrogen density of common intermetallic

AB-, AB2- and AB5-type alloys is high (around 60 kg/m3), while

their gravimetric hydrogen density does not exceed 2 mass%

H [3]. Gravimetric capacities of >2 mass% H at ambient tem-

perature and pressure conditions are achieved only with

vanadium-based alloys with a body centred cubic (BCC) crys-

tal structure [4e6]. However, hydrogenation reactions of BCC

alloys exhibit two plateau regions. In the lower plateau region

up to a hydrogen content of ~1.2e1.5mass%, themonohydride

is formed: M þ H / MH1, where M stands for metal and H

stands for hydrogen. This region typically shows a very low

equilibrium pressure of <1 kPa at ambient temperature. The

second plateau region begins at ~1.5 mass% where the dihy-

dride is formed: MH1 þ H / MH2. A total capacity of up to

4.37 mass% for both pressure plateaus has been reported [7].

At ambient temperature and pressure conditions, only the

formation/dissociation of the dihydride is useful for hydrogen

storage, with reversible capacities typically ranging between

1.7 and 2.6 mass%.

High raw material costs, especially for vanadium, hinder

the commercial application of V-based hydrogen storage

materials with BCC structure. The development of ternary or

quaternary alloys has considerably decreased the material

cost through the partial substitution of expensive rawmetals,

e.g. V (390 US $/kg) or Ti (35 US $/kg), by cheaper raw metals,

such as Mn (4 US $/kg) or Fe (1 US $/kg) [8e10]. In order to

further decrease the cost of V, several researchers have

attempted to replace V by a ferrovanadium master alloy

[11e13]. Another strategy to reduce the price of V is to use low-

purity V containing impurities [14,15].

The effect of impurities on the microstructure and

hydrogen storage properties of BCC alloys has been reported

by various authors [16e22]. Nakamura et al. recently reported

a high oxygen concentration of 10,000 mass-ppm in the V raw

metal to raise the second plateau of a VeTieMn alloy

compared to the alloy prepared with low oxygen V (300 mass-

ppm O) [23]. The thermodynamic characteristics of the lower

plateau of the high oxygen containing alloy were essentially

the same as those of the low oxygen alloy. This behaviour was

attributed to two origins, the primary and the secondary ox-

ygen effect. The primary effect was related to oxygen atoms

dissolving in the BCC main phase and interacting with

hydrogen. The secondary effect was related to the formation

of a secondary, oxygen rich phase. Oxygen reacts with ele-

ments showing a high affinity towards oxygen, e.g. Ti, which

promotes the formation of a secondary phase. The composi-

tion of the BCC main phase is hereby changed, altering the

microstructure and thermodynamic properties of the alloy

[22]. A detailed analysis of the forming phases has not been

provided yet.

In this paper, the effect of oxygen on the microstructure,

hydrogen storage characteristics and thermodynamics of a

quaternary VeTieCreFe alloy is investigated. The negative

effect of oxygen on the structure and hydrogenation proper-

ties can be circumvented through the addition of 1 at% rare

earth metal, which is in agreement with the results published

by Mi et al. [24].

Experimental

Three different types of alloys were prepared by arcmelting in

a water-cooled copper crucible using two V materials of

differing oxygen concentration (low oxygen: <300 ppmO; high

oxygen: <10,000 ppm O), Ti, Fe, Cr and rare earth (La, Y or Ce)

materials (purity > 99.9%). One sample was prepared with low

oxygen V, one sample with high oxygen V and one sample

with high oxygen vanadium and 1 at% of rare earth metal (La,

Y or Ce). Rare earth metal oxides agglomerated mainly at the

ingot's surfaces and could be easily removed by grinding.

Special care was taken during the preparation of the alloys to

avoid contamination with impurities such as oxygen or

nitrogen.

The oxygen content was analysed by carrier gas hot

extraction (CGHE). A commercial oxygen/nitrogen analyser

TC600 (LECO) was used which was calibrated with dried Fe

powder (JK47) and verified with a standard from LECO (502-

201) and a standard from Alpha Resources (AR 640-ZR702B).

The calibration range was close to the concentration of the

samples. The standard oxides and the samples were weighed

with a mass in the range from 50 to 300 mg (weighing

accuracy ±0.002mg) in Sn crucibles (9e10 mm). Together with

a nickel capsule (about 500 mg), which served as fusion aid,

the package was put into an outgassed (outgassing power:

6300 W) high temperature graphite crucible. The measure-

ments were performed at 5800W heating power. The evolving

gases CO2 and CO were swept out by helium as inert carrier

gas and measured by infrared detectors. The oxygen concen-

tration of each material was measured three times, and an

average value was calculated.

The alloy morphology and composition was analysed by

SEM (S3400-N, Hitachi) and EDX spectroscopy (GENESIS 2000H,

EDAX). According to the manufacturer, the accuracy of the

compositions as determined by EDX was ±0.5 at%. Average

compositions were determined at three different spots of the

ingots: at the edge (x1z r), at the centre (x2¼ 0$r) and at a third

spot between those two points (x3 z 1/2$r). They were deter-

mined by collecting EDX spectra of an area of approximately

100 � 100 mm and are referred to in the text as average

composition. In addition, the compositions of individual

points within the different phases were also determined by

collecting EDX point spectra. These are the relative concen-

trations referred to in Table 1.

TEM samples were prepared by focused Gaþ-ion beam

milling using a Strata 400 A (FEI Company) with the final

Fig. 1 e SEM images of V40Fe8Ti26Cr26 with (a) high oxygen concentration, (b) low oxygen concentration and (c) deoxidized

using 1 at% La with 1000£ magnification. SE: secondary electron images (top); BSE: back-scattered electron images (bottom).

Table 1 e Relative concentrations of constituent phases in low-oxygen, high-oxygen and deoxidized V40Fe8Ti26Cr26determined by SEM-EDX. Oxygen is not included.

Sample Phase V [at%] Fe [at%] Ti [at%] Cr [at%]

Target composition 40 8 26 26

Low-oxygen Main phase 41.4e43.8 6.4e7.1 22.5e24.1 26.8e27.3

Ti-rich sec. phase 28.4e32.7 11.2e13.6 34.0e38.2 19.8e22.1

High-oxygen Main phase 41.2e42.9 8.1e8.9 21.9e24.1 26.2e27.4

Ti-rich sec. phase 2.8e13.4 0.1e2.9 78.3e95.9 0.7e6.5

Deoxidized Main phase 38.1e44.0 7.1e8.5 22.9e24.3 23.9e27.9

Ti-rich sec. phase 27.6e34.1 6.9e11.9 35.5e46.1 15.3e21.5

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 2 0 0 0 0e2 0 0 0 820002

polishing performed at 5 kV acceleration voltage. TEM char-

acterization was performed using an aberration corrected

Titan 80-300 (FEI Company) operated at 300 kV equipped with

an HAADF-STEM detector, a GIF 863 (Gatan) and an S-UTW

EDX detector (EDAX).

Powder XRD data of the samples were obtained using a

Rigaku RINT-2500 V X-ray diffractometer equipped with a

rotating anode X-ray source. CueKa radiation was used for

diffraction. The crystalline phases were quantified by the

Rietveld method using the TOPAS software package [25].

For homogenization, samples were annealed at 1673 K for

25 h in argon atmosphere. Formeasurement of P-C-isotherms,

the sample ingots were crushed into particles of a few milli-

metres. To activate the material, approximately 2 g of sample

were sealed in a stainless steel container and evacuated at

323 K for 2 h using a turbo molecular pump. After cooling to

room temperature, the container was pressurized with 5 MPa

of hydrogen for 30 min. An absorption reaction was noted

after a short incubation time through a temperature rise of the

container. Subsequently, the samples were evacuated for

30 min at room temperature. This procedure was repeated

three times. Before each absorption measurement, the sam-

ples were evacuated at 773 K for 2 h to ensure a complete

dehydrogenation.

Results and discussion

Microstructural analysis

Scanning electron microscopyIn Fig. 1 SEM images of three V40Fe8Ti26Cr26 alloys prepared

with high-oxygen V, low-oxygen V and high-oxygen V and 1 at

% La for deoxidization (denoted below as “high-oxygen sam-

ple”, “low-oxygen sample” and “deoxidized sample”) are pre-

sented. The compositions of the phases determined by EDX

are listed in Table 1.

The main phase of the low-oxygen and of the deoxidized

samples contained an average of 23.8 and 23.5 at% Ti,

respectively.

The high-oxygen sample contained a large fraction of Ti-

rich secondary phase. The secondary phase was clearly

Fig. 2 e XRD patterns of V40Fe8Ti26Cr26 (A) prepared with

low-oxygen V, (B) prepared with high-oxygen V and (C)

prepared with high-oxygen V and 1 at% La in the annealed

state.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 2 0 0 0 0e2 0 0 0 8 20003

visible in the form of dark areas in the secondary electron or

back-scattered electron images, respectively. This phase

consisted of 78.3e95.9 at% Ti. Consequently, the fraction of Ti

in the main phase was reduced to an average of 22.3 at%,

which was slightly lower than in the low-oxygen and deoxi-

dized samples. EDX spectra of the secondary phase also

showed considerable oxygen amounts (not shown). This

observation suggests a reaction between oxygen and titanium

proceeded during the preparation of the alloy, corresponding

to the secondary oxygen effect described above. However, due

to the low sensitivity of SEM-EDX for light-weight elements,

an accurate quantification of the oxygen content was not

possible. The acceleration voltage applied in the investigation

was 20 kV. We estimated the electron beam to penetrate the

sample to a depth of ~2 mm. A cross section of the alloys was

taken for the investigation, and it cannot be ruled out that the

EDX spectra also contain information from different phases

underneath the surface. The Ti-content would be lowered and

the V-, Cr- and Fe-content would be increased for some parts

of the secondary phase.

Both the low-oxygen and the deoxidized sample only

showed a small fraction of secondary phase. For the low-

oxygen sample, the secondary phase consisted of

34.0e38.2 at% Ti and for the deoxidized sample, the fraction

of Ti was found to be 35.5e46.1 at%. The secondary phase is

most likely a Laves phase or a related one. Its volume frac-

tions were too low to be detected by XRD. However, a phase

of similar composition was formed during the preparation of

alloys containing varying Ti/Cr ratios, as shown in Fig. 1s in

the supporting information, and here the phase could be

identified as C14 Laves phase by XRD. The fractions of sec-

ondary phase in the low-oxygen and deoxidized samples

were much lower than in the case of the high-oxygen sample.

In addition, the secondary phases in the low-oxygen and

deoxidized samples contained only about 10% more Ti than

the BCC main phase. Consequently, the fraction of Ti in the

BCC main phase was almost unaffected by the formation of

secondary phase. For the high-oxygen sample, the Ti content

of the secondary phase was 50e70% higher than the Ti con-

tent of the BCC main phase. This resulted in a reduction of

the Ti content of the BCC main phase as compared to the

target composition.

For the deoxidized sample, small inclusions of La oxide

were visible as white spots distributed within the alloy. The

solubility of La or La oxides in V, which is themain constituent

of the alloy, is below 0.1% [26].

XRDFig. 2 shows the X-ray diffraction patterns of the high-oxygen,

low-oxygen and deoxidized samples.

The low-oxygen sample showed a BCC structure with a

lattice parameter of 3.0308 (2) A. No additional phases were

detected by XRD.

The diffraction pattern of the high-oxygen sample exhibi-

ted peaks of two distinct phases, the BCC main phase and a

secondary phase isostructural with a-Ti. The refined lattice

parameter of the BCC main phase is 3.0141 (3) A, corre-

sponding to a reduction of 0.016 A (0.53%) compared to the

low-oxygen sample. The lattice parameters of the a-Ti-like

secondary phase were determined to be a ¼ 2.9707 (73) A and

c ¼ 4.7843 (61) A. Pure a-Ti has a hexagonal close-packed

structure with lattice parameters of a ¼ 2.951 A and

c ¼ 4.686 A [27]. Thus, the unit cell of the a-Ti-like phase

expanded by ~0.01 A in a-direction and ~0.1 A in c-direction. It

has been reported that the lattice parameter of a-Ti is a

function of the oxygen content [28e30]. With an increased O-

content, the lattice of a-Ti will expand. Based on these reports

we estimated the oxygen content of the secondary, a-Ti-like

phase to be 25 ± 5 at% O. Thus, titanium acts as a deoxidizing

agent forming a TieO phase. Rietveld refinement of the high-

oxygen pattern revealed the fraction of a-Ti-like phase to be

3% and the BCCmain phase 97%. Alloys of various Ti/Cr ratios

were also prepared for comparison, and a similar result was

obtained for Ti/Cr between 26/26 (as presented in this paper)

and 36/16, suggesting that, for this type of multiphase alloy, a

high oxygen content leads to the formation of about 3e5%

secondary phase with a-Ti-like structure. XRD patterns of the

alloys prepared with low- and high-oxygen vanadium with

the respective lattice parameters are presented in Fig. 1s and

Table 1s in the supporting information.

For the deoxidized sample, the BCCmain phase and minor

peaks corresponding to the C14 Laves phase were detected.

Thus, the addition of 1 at% lanthanum inhibited the formation

of a-Ti-like phase that occurred during the preparation of the

alloys with high-oxygen V. The lattice parameter of the

deoxidized sample was found to be 3.0297 (3) A, close to the

value of the low-oxygen sample. The free energy of rare earth

oxide formation is the largest of all known metals [31].

Therefore, these metals show the highest affinity towards

oxygen and can act as oxygen getters. Liu et al. reported the

formation of a TiFeOx-phase in a Ce-free V-based BCC solid

solution. The addition of Ce suppressed the formation of this

phase. However, no detailed analysis of lattice parameters

was provided [32].

TEMIn order to further elucidate on the structure of the a-Ti-like

phase of the high-oxygen sample it was investigated by TEM.

Fig. 3 e (a) STEM-HAADF image of the a-Ti-like and BCC main phase of the high-oxygen sample, (b) SAED pattern on the a-

Ti-like phase in the ⟨223⟩ zone axis with the corresponding d values of 2:57ð110Þa�Ti, 1:77ð102Þa�Ti. (c) & (d) EEL spectra of the

areas 1 & 2 indicated in (a).

Fig. 4 e (a) HRTEM image of the a-Ti phase, (b) and (c) FFT

patterns of the areas indicated in (a).

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 2 0 0 0 0e2 0 0 0 820004

The purpose of the investigation was 1) to confirm the a-Ti-

like structure of the secondary phase, 2) to determine its

composition, and 3) to clarify whether oxygen is found in the

BCC main phase.

Fig. 3 shows a STEM-HAADF image of the a-Ti-like and BCC

main phase, EELS spectra of the indicated areas and SAED

pattern of the a-Ti-like phase.

The a-Ti and the BCC main phase were separated from

each other by a well-defined phase boundary. The EEL spec-

trumof the BCCmain phase showed characteristic energy loss

signals corresponding to V, Ti, Cr and Fe. No oxygen was

detected by EELS, indicating an oxygen concentration below

the detection limit of around 1 at% in the BCCmain phase. The

EEL spectrum of the a-Ti-like phase only showed Ti, O and V

edges. EEL spectra were collected at various positions and

oxygen was found to be distributed homogeneously within

the secondary phase. The presence of V in the secondary

phase was not noticed by SEM or XRD analysis, but is in good

agreement with the TieV and TieO phase diagrams showing a

solubility of V and O in a-Ti at room temperature [30,33]. The

lattice of a-Ti is known to expand in c-direction by Dc ¼ 0.02 A

when increasing the V concentration from 0 to 9.5 at% V [33].

In earlier reports, an expansion of the a-Ti lattice by

Dc ¼ 0.07 A was found with increasing oxygen concentration

from 0 to 9.5 at% O [28e30]. Thus, the lattice expansion effect

of oxygen is more pronounced than that of vanadium,

justifying the assumption in Section XRD that the lattice

expansion of a-Ti was caused mainly by oxygen.

The SAED pattern of the a-Ti-like phase is in agreement

with the ⟨223⟩ zone axis of an HCP structure (a-Ti).

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 2 0 0 0 0e2 0 0 0 8 20005

Fig. 4a shows a HRTEM image with the corresponding FFT

patterns of the two indicated areas (4b and 4c). EDX spectra of

the respective areas indicated in the HRTEM image are shown

in the supporting information (Fig. 2s). The image was ob-

tained in the a-Ti-like phase and shows that a third phase is

present in the high-oxygen alloy, which could not be detected

by XRD due to the low volume fraction.

A full crystallographic analysis of this new nanosized

phase was not possible, but EDX analysis indicates that in

addition to V, Ti and O traces of Cr, Fe and N are present. This

results in a significant expansion of the unit cell with peri-

odicities of 11.4 and 1.87 A in the observed zone axis. No

matching phases for VFeTiCr alloyswith O or N could be found

in the literature.

Oxygen analysis

The “path” of oxygen during the synthesis of the alloys was

further elucidated bymeasuring oxygen concentrations of the

raw metals and of the as-cast alloys. Based on the oxygen

analysis of the rawmetals, a theoretical oxygen concentration

of the alloys was calculated. These procedures provide valu-

able information of two types. Firstly, the main sources for

oxygen contamination are identified (aside from high-oxygen

vanadium). Secondly, comparing the theoretical with the

measured values provides additional information about

changes of oxygen concentration during the synthesis.

Several compositions of varying Ti/Cr ratios and constant

amounts of V (40 at%) and Fe (8 at%) were tested. Results are

presented in Table 2.

Both high-oxygen (HO) and low-oxygen (LO) vanadium

were found to contain higher oxygen concentrations than

indicated by the supplier, 1.46 mass% (HO) and 0.0078 mass%

(LO) compared to <1 mass% (HO) and <0.003 mass% (LO) ac-

cording to the supplier. Pure Ti contained 0.107 mass% O,

while Fe and Cr showed only minor oxygen concentrations of

0.0097 mass% and 0.0137 mass% O, respectively.

Ti contains a higher fraction of oxygen than Cr or Fe.

Consequently, the theoretical oxygen concentration of the

alloys increased from 0.615 mass% (HO) or 0.0339 mass% (LO)

for a Ti/Cr ratio as 26/26 to 0.627 mass% (HO) or 0.0428 mass%

(LO) for a Ti/Cr ratio as 36/16.

The measured concentrations of the as-casted high-oxy-

gen alloys were between ~0.06 and ~0.11 mass% lower than

Table 2 e Oxygen concentrations of raw metals and four differ36/16 with low oxygen concentration (LO), high oxygen concenanalysis.

Raw materials

V (high-oxygen) V (low-o

Measured concentration [mass%] 1.4600 ± 0.04 0.0078 ±

Alloys

Ti/Cr ratio Ti/Cr ¼ 26/26

Theoretical oxygen concentration [mass%] 0.034 ± 0.0045 (LO)

0.615 ± 0.024 (HO)

Measured oxygen concentration [mass%] 0.045 ± 0.0037 (LO)

0.515 ± 0.035 (HO)

0.0690 ± 0.0122 (deox)

the theoretical concentrations. In contrast, the oxygen con-

centrations of the low-oxygen alloys were found to slightly

increase compared to the calculated values in three cases

(Ti/Cr¼ 26/26, 32/20 and 36/16). For the Ti/Cr¼ 30/22 alloy, the

measured oxygen concentration was lower than theoretically

anticipated, however, the values are almost identical within

the experimental error.

These results suggest that for higher oxygen concentra-

tions in the rawmetals, a part of the oxygen is released during

the arc melting process. Consequently, the final oxygen con-

tent is reduced. For low oxygen concentrations in the raw

metals, the amount of impurity oxygen atoms is much lower

than for the high-oxygen samples. If the oxygen concentration

is low, only a small amount of oxygen atoms will be released

during arc melting. At the same time, some of the molten

metals may act as oxygen getters. Therefore, it is likely that

below a certain “threshold value”, the oxygen concentration is

not reduced by arc melting.

Hydrogen storage properties

Pressure-composition isotherms of the low-oxygen, high-ox-

ygen and deoxidized sample are presented in Fig. 5.

For all materials tested in the present work, the equilib-

rium pressure of monohydride formation was below the

detection limit of the apparatus of 0.001 MPa. The pressure

transducer recorded a pressure value of 0 MPa. These points

are not shown on the logarithmic scale. The low-oxygen

sample showed a pressure of dihydride formation (pf) and

dissociation (pd) of 1.15 MPa and 0.24 MPa at 298 K, respec-

tively. In hydrogen-metal systems the pressure of hydride

formation is typically larger than the pressure of hydride

dissociation. This effect is known as pressure hysteresis, and

it can be quantified by the hysteresis factor H ¼ pf/pd. For the

low-oxygen sample, the hysteresis factor corresponded to pf/

pd ¼ 4.5. For the high-oxygen sample, the pressure of hydride

formation and dissociation was raised to 3.5 MPa and 0.8 MPa,

respectively, corresponding to a hysteresis factor of pf/pd¼ 4.4.

Thus, both samples exhibited a comparable hysteresis factor.

Nakamura et al. investigated a Ti1.0V1.1Mn0.9 (corresponding to

Ti33.3V36.7Mn30.0 in at%) alloy and reported formation and

dissociation pressures of 0.18 MPa and 0.004 MPa for the low-

oxygen sample and 0.8 and 0.007 MPa for the high-oxygen

sample, respectively [23]. This corresponds to hysteresis

ent alloys with Ti/Cr ratios in the range between 26/26 andtration (HO) or deoxidized (deox), obtained by oxygen

xygen) Fe Ti Cr

0.0009 0.0097 ± 0.0004 0.1070 ± 0.008 0.0137 ± 0.0004

Ti/Cr ¼ 30/22 Ti/Cr ¼ 32/20 Ti/Cr ¼ 36/16

0.038 ± 0.0044 (LO) 0.0389 ± 0.0045 (LO) 0.0428 ± 0.002 (LO)

0.622 ± 0.024 (HO) 0.623 ± 0.024 (HO) 0.627 ± 0.023 (HO)

0.035 ± 0.003 (LO) 0.043 ± 0.002 (LO) 0.053 ± 0.002 (LO)

0.565 ± 0.008 (HO) 0.521 ± 0.003 (HO) 0.053 ± 0.002 (HO)

Fig. 5 e Pressure-composition isotherms of V40Fe8Ti26Cr26at 298 K prepared with high-oxygen V, low-oxygen V and

deoxidized using 1 at% La.

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 2 0 0 0 0e2 0 0 0 820006

factors of 45 (low-oxygen) and 115 (high-oxygen). Apparently,

for the Ti33.3V36.7Mn30.0 alloy, the hysteresis factor of the high-

oxygen sample was increased as compared to the low-oxygen

sample. In the present work, the hysteresis factor of the high-

oxygen V40Fe8Ti26Cr26 sample decreased slightly, suggesting

that the hysteresis increase with oxygen concentration de-

pends strongly on the composition of the alloy.

Nakamura et al. also noted the raise in equilibrium pres-

sure to be more pronounced for hydride formation than for

dissociation [23]. For our VeFeeTieCr alloy, the formation

pressure was increased by a factor ofpf ;HOpf ;LO

¼ 3 and the disso-

ciation pressure was increased by a factor ofpf ;HOpf ;LO

¼ 3:3 (with

HO and LO standing for high-oxygen and low-oxygen and f

and d standing for formation and dissociation, respectively).

For comparison, these factors amounted to 4.4 for formation

and to 1.8 for dissociation for the VeTieMn alloy studied by

Nakamura et al. For the material studied in the present work,

both the hydride formation and dissociation pressures were

increased by a factor of ~3. This suggests that the increase of

hydride formation and dissociation pressure also depends

strongly on the composition of the alloy. The origin of this

effect has not been studied yet, and further work is necessary

in that field.

Compared to the high-oxygen sample without addition of

La, the addition of 1 at% La effectively reduced the equilibrium

pressure to 1.18 MPa for hydride formation and 0.26 MPa for

hydride dissociation. Thermodynamic properties of the

deoxidized sample were similar to those of the low-oxygen

sample. This is in good agreement with the XRD results of

the alloys, as the low-oxygen and deoxidized sample exhibi-

ted similar lattice parameters. Thermodynamic properties are

known to depend strongly on the alloy composition and lattice

parameter [34]. The fact that the two samples showed similar

thermodynamic properties and lattice parameters further-

more suggests that the oxygen content in the BCC phase of

both samples is also similar. If the oxygen content in the BCC

phase of the sample with La addition had been higher than in

the BCC phase of the low-oxygen sample, the equilibrium

pressure of the former sample also should have been higher

than that of the latter sample.

The low-oxygen sample showed a total capacity of

3.5 mass%, the deoxidized sample showed a capacity of

3.4 mass%. This slightly reduced capacity is possibly due to

the inclusions of La oxide found within the alloy. The atomic

weight of V, Fe, Ti and Cr is in the range between 47.88 g/mol

and 55.85 g/mol, while La has an atomic weight of 138.91 g/

mol. Thus, it is more than twice as heavy as the other metals

used in the alloy. Since La oxide does not absorb any

hydrogen, it is an inert material, making the alloy heavier and

thereby reducing the hydrogen content. The high-oxygen

sample exhibited a total capacity of 2.5 mass%. However, as

the maximum pressure of the PCT apparatus was 6 MPa, the

sample could possibly absorb more hydrogen at higher pres-

sure. To investigate that, the experimental temperature was

reduced to 255 K. This procedure lowered the equilibrium

pressure, and the sample could be hydrogenated completely.

The sample temperature of the first cycle was 298 K, and the

monohydride is thermodynamically too stable to desorb

hydrogen at this temperature. Therefore, only the second

plateau (formation/dissociation of the dihydride) was

observed.

Pressure-composition isotherms of the high-oxygen sam-

ple during the first two cycles at 298 K and 255 K are presented

in the supporting information in Fig. 3s. At 255 K, 1.5 mass% H

were stored. Pressure-composition isotherms of the low-

oxygen sample at 298 K, 275 K and 255 K are also presented

in Fig. 4s of the supporting information. At 255 K, the low-

oxygen sample absorbed/desorbed 2.5 mass% H between

0.01 and 6 MPa (formation/dissociation of the dihydride).

Thus, the second plateau of the high-oxygen sample showed a

reduced capacity as compared to the low-oxygen sample.

Tsukahara et al. reported the hydrogen content of a

V3TiNi0.56Co0.14Nb0.047Ta0.047 to depend strongly on the oxygen

concentration. Below 5000 ppm O, the capacity was largely

unaffected. For oxygen concentrations >5000 ppm, the ca-

pacity was reduced [19]. In contrast, the total hydrogen ca-

pacity of the low- and high-oxygen VeTieMn alloys studied by

Nakamura et al. was not affected by oxygen. Oxygen is known

to reduce the hydrogen capacity of pure V [35]. The alloy

studied by Tsukahara et al. contained a high vanadium frac-

tion of 63 at%. In the present work, the V content was 40 at%,

and the material studied by Nakamura et al. contained 36.7 at

% V. It is interesting to note that for the alloy studied in the

present work, the overall hydrogen capacity was reduced,

while for the VeTieMn alloy investigated by Nakamura et al.,

the overall capacity was not affected by oxygen, despite the

fact that the vanadium content of our alloy was only 3.7 at%

higher. Aside from the different vanadiumconcentrations, the

composition of the alloys studied in the various works was

also different.

Finally, it should be noted that a deoxidization of the high-

oxygen alloys could also be achieved by using other rare earth

elements such as yttrium or cerium.

Conclusion

The effect of the oxygen content of a VeFeeTieCr alloy was

investigated. Two vanadium raw metals containing 300 ppm

and 10,000 ppm O were used for the synthesis of the alloys. A

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 2 0 0 0 0e2 0 0 0 8 20007

significant effect was observed for the sample containing

10,000 ppm O where a secondary isostructural with a-Ti, was

formed and the Ti content of the BCC main phase was

reduced. Apart from Ti and O, the secondary phase also con-

tained V. Furthermore, a third so far unknown phase con-

sisting of Ti, V, Cr, Fe and traces of O and N was confirmed.

Thermodynamic properties of the high-oxygen sample were

altered with respect to its hydrogen absorption properties: a

high oxygen concentration led to an increased pressure of

hydride formation and dissociation. The total hydrogen ca-

pacity of the high-oxygen samplewas reduced as compared to

the low-oxygen sample. The addition of 1 at% rare earthmetal

(La, Y or Ce) effectively deoxidized the alloys, as confirmed by

XRD and oxygen analysis. This resulted in comparable ther-

modynamic properties and a similar hydrogen storage ca-

pacity as the low-oxygen sample.

Acknowledgements

This work was funded partly by the Japanese Society for the

Promotion of Science (JSPS) and the German Helmholtz

Association.

TEM, FIB and oxygen analysis was performed at the

Karlsruhe Nano Micro Facility (KNMF, www.knmf.kit.edu).

The authors thank Torsten Scherer and Robby Prang (KIT-INT)

for TEM sample preparation.

Appendix A. Supplementary data

Supplementary data related to this article can be found at

http://dx.doi.org/10.1016/j.ijhydene.2014.08.152.

r e f e r e n c e s

[1] Schlapbach L, Zuttel A. Hydrogen-storage materialsfor mobile applications. Nature Nov. 2001;414(6861):353e8.

[2] Wilson PR, Bowman RC, Mora JL, Reiter JW. Operation of aPEM fuel cell with LaNi4.8Sn0.2 hydride beds. J Alloys CompdOct. 2007;446e447:676e80.

[3] Sakintuna B, Lamaridarkrim F, Hirscher M. Metal hydridematerials for solid hydrogen storage: a review. Int JHydrogen. Energy Jun. 2007;32(9):1121e40.

[4] Akiba E, Iba H. Hydrogen absorption by laves phaserelated BCC solid solution. Intermetallics Jan.1998;6(6):461e70.

[5] Akiba E, Okada M. Metallic hydrides III: cubic solid-solutionalloys. MRS Bull 2002;(September):699e703.

[6] Cho S-W, Han C-S, Park C-N, Akiba E. The hydrogen storagecharacteristics of TieCreV alloys. J Alloys Compd Jun.1999;288(1e2):294e8.

[7] Kumar A, Shashikala K, Banerjee S, Nuwad J, Das P,Pillai CGS. Effect of cycling on hydrogen storageproperties of Ti2CrV alloy. Int J Hydrogen Energy Feb.2012;37(4):3677e82.

[8] Okada M, Kuriiwa T, Tamura T, Takamura H, Kamegawa A.Ti�V�Cr BCC alloys system with high protium content.Metals Mater Int Feb. 2001;7(1):67e72.

[9] Yan Y, Chen Y, Liang H, Wu C, Tao M. Hydrogen storageproperties of V30eTieCreFe alloys. J Alloys Compd Jan.2007;427(1e2):110e4.

[10] Seo C, Kim J, Lee PS, Lee J. Hydrogen storage properties ofvanadium-based B.C.C. solid solution metal hydrides. JAlloys Compd Jan. 2003;348(1e2):252e7.

[11] Nomura K, Akiba E. H2 absorbingedesorbing characterizationof the TieVeFe alloy system. J Alloys Compd Dec.1995;231(1e2):513e7.

[12] Yan Y, Chen Y, Wu C, Tao M, Liang H. A low-cost BCCalloy prepared from a FeV80 alloy with a highhydrogen storage capacity. J Power Sources Feb.2007;164(2):799e802.

[13] Wu C, Borgschulte A, Frischknecht U, Yan Y, Yang F, Luo L,et al. Surface properties of V40(TiCr)51Fe8Mn alloy duringhydrogenation/dehydrogenation cycles. J Alloys Compd Feb.2013;40:30e2.

[14] Moskalyk R, Alfantazi A. Processing of vanadium: a review.Miner Eng Sep. 2003;16(9):793e805.

[15] Nakagawa S. Personal communication. Taiyo Koko Ltd; 2013.[16] Yan Y, Chen Y, Liang H, Wu C, Tao M. The effect of Si on

V30Ti35Cr25Fe10 BCC hydrogen storage alloy. J Alloys CompdAug. 2007;441(1e2):297e300.

[17] Yan Y, Chen Y, Liang H, Wu C, Tao M, Mingjing T. Effect of Alon hydrogen storage properties of V30Ti35Cr25Fe10 alloy.J Alloys Compd Dec. 2006;426(1e2):253e5.

[18] Reilly JJ, Wiswall RH. Higher hydrides of vanadium andniobium. Inorg Chem Jul. 1970;9(7):1678e82.

[19] Tsukahara M, Takahashi K, Isomura A, Sakai T. Influence ofoxygen on hydrogen storage and electrode properties formicro-designed V-based battery alloys. J Alloys Compd Jan.1998;265(1e2):257e63.

[20] Tsukahara M, Takahashi K, Mishima T, Isomura A, Sakai T.Influence of various additives in vanadium-based alloysV3TiNi0.56 on secondary phase formation, hydrogen storageproperties and electrode properties. J Alloys Compd Nov.1996;245(1e2):59e65.

[21] Shi J, Sakai T, Takeshita HT, Kuriyama N, Tsukahara M.Influence of carbon impurity on microstructures andelectrode properties for V-based battery alloys. J AlloysCompd Aug. 1999;290(1e2):267e72.

[22] Shen C-C, Chou JC-P, Li H-C, Wu Y-P, Perng T-P. Effect ofinterstitial boron and carbon on the hydrogenationproperties of Ti25V35Cr40 alloy. Int J Hydrogen Energy Nov.2010;35(21):11975e80.

[23] Nakamura Y, Nakamura J, Sakaki K, Asano K, Akiba E.Hydrogenation properties of TieVeMn alloys with a BCCstructure containing high and low oxygen concentrations.J Alloys Compd Feb. 2011;509(5):1841e7.

[24] Mi J, Lu F, Liu X, Jiang L, Li Z, Wang S. Enhancement of ceriumand hydrogen storage property of a low-cost TieV based BCCalloy prepared by commercial ferrovanadium. J Rare EarthsOct. 2010;28(5):781e4.

[25] Coelho AA. TOPAS academic version 4.1 technical reference.Brisbane, Australia. 2007.

[26] Kornilov II, Matveeva NM. The metal chemistry of vanadium.Russ Chem Rev Sep. 1962;31(9):512e28.

[27] Wood RM. The lattice constants of high purity alphatitanium. Proc Phys Soc Sep. 1962;80(3):783e6.

[28] Yamaguchi S, Hiraga K, Hirabayashi M. Interstitialorderedisorder transformation in the TieO solid solution. IV.A neutron diffraction study. J Phys Soc Jpn Apr.1970;28(4):1014e23.

[29] Yamaguchi S. Interstitial orderedisorder transformation inthe TieO solid solution. I. Ordered arrangement of oxygen.J Phys Soc Jpn Jan. 1969;27(1):155e63.

[30] Murray JL, Wriedt HA. The O�Ti (OxygeneTitanium) system.J Phase Equilib Apr. 1987;8(2):148e65.

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 9 ( 2 0 1 4 ) 2 0 0 0 0e2 0 0 0 820008

[31] Karen P, Kjekshus A. High-temp. superconductors e I, vol. 30.Elsevier; 2000. p. 229e373.

[32] Liu XP, Cuevas F, Jiang LJ, Latroche M, Li ZN, Wang SM.Improvement of the hydrogen storage properties ofTieCreVeFe BCC alloy by Ce addition. J Alloys Compd May2009;476(1e2):403e7.

[33] Murray JL. The Ti�V (titaniumevanadium) system. Bull AlloyPhase Diagrams Jun. 1981;2(1):48e55.

[34] Nakamura Y, Akiba E. Hydriding properties and crystalstructure ofNaCl-typemono-hydrides formed fromTieVeMnBCCsolid solutions. J AlloysCompdOct. 2002;345(1e2):175e82.

[35] Asano K, Havashi S, Mimura K, Enoki H, Nakamura Y.Hydrogen diffusion in V studied by 1H-NMR.pdf. In:Collected book of abstracts, metal hydrides conference; 2012.p. 375.