Calorimetric studies of C14 and C15 YMn2 and YMn2(H,D)6

(1)

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling or

licensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of the

article (e.g. in Word or Tex form) to their personal website or

institutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies are

encouraged to visit:

(2)

Calorimetric properties of C14 and C15 YMn

2

and YMn

2

(H,D)

6 R.S. Liu

a,

*

, I. Baginskiy

a

, H.T. Kuo

a

, S.M. Filipek

b

, R. Wierzbicki

b

, R. Sato

b

,

A.V. Tsvyashchenko

c

, L. Fomicheva

c

, H.H. Wu

d

, C.B. Tsai

d

, C.C. Yang

d

, R. Asmatulu

e

,

J.C. Ho

e

, Y.Y. Chen

d,

**

a_{Department of Chemistry, National Taiwan University, Taipei 106, Taiwan}

b_{Institute of Physical Chemistry, Polish Academy of Science, 44/52, Warsaw 01-224, Poland} c

LF Vereshchagin Institute of High Pressure Physics, Moscow 142092, Russia

d

Institute of Physics, Academia Sinica, Taipei 115, Taiwan

e_{Department of Physics, Wichita State University, Wichita, KS 67260, USA}

a r t i c l e i n f o

Article history:

Received 14 August 2010 Received in revised form 25 October 2010

Accepted 10 November 2010 Available online 24 December 2010

Keywords: YMn2 Hydrogen Deuterization Latent heat Calorimetry

a b s t r a c t

The capacity of YMn2 to undergo reversible hydrogenization makes this compound

a potential candidate for hydrogen storage. This report describes the successful synthesis of YMn2H6from hexagonal C14 YMn2as well as calorimetric studies of hexagonal C14

YMn2, cubic C15 YMn2and their isostructural derivatives, YMn2H6and YMn2D6.

Comple-mentary structural and physical property characterizations include X-ray diffraction, electrical resistivity, and magnetic susceptibility measurements. Of particular interest is a hysteretic first-order phase transition in C15 YMn2near 100 K, which does not occur in

C14 YMn2. Analysis of a specific heat anomaly that is associated with this transition reveals

a substantial latent heat of 330 J/mol. No such anomalous specific heat exists in YMn2D6,

suggesting that the transition is suppressed upon deuterization.

1. Introduction

Intermetallic compositions related to RM2 Laves phases

(R¼ rare earth and M ¼ transition metal Mn, Fe, Co) have large hydrogen absorption capacity at room temperature owing to the formation of RM2Hx(0 x 6) hydrides. Consequently,

they are considered as promising materials for hydrogen storage[1,2]. Intermetallic YMn2crystallizes in either a cubic

C15 (Fd3m) or a hexagonal C14 (P63/mmc) Laves phase. Both

phases can potentially be used to store hydrogen by forming YMn2Hx. Most of the relevant literature discusses C15 YMn2,

which undergoes first-order structural and antiferromagnetic (mMn ¼ 2.7 mB) transitions near 100 K, accompanied by

a significant volume change of 5% [2,3]. No such phase transformation occurs in C14 YMn2. The conventional

low-pressure hydrogenation process of C15 YMn2at relatively low

temperatures yields YMn2Hx with an x value of less than

around 4.5 [4,5]. Only recently, YMn2D6 (Fm-3m) was

successfully synthesized from C15 YMn2under high-pressure

deuterium at elevated temperature[6]. Deuterium was used rather than hydrogen to facilitate neutron diffraction studies. This effort increased x to six [6]. Just as important is the reversibility of the deuterization process, releasing the six deuterium atoms at high temperatures, enhancing the viability of YMn2 for hydrogen storage. Interestingly,

iso-structural Fm-3m RMn2H6hydrides (deuterides), where R¼ Y,

* Corresponding author. ** Corresponding author.

E-mail addresses:rsliu@ntu.edu.tw(R.S. Liu),cheny2@phys.sinica.edu.tw(Y.Y. Chen). A v a i l a b l e a t w w w . s c i e n c e d i r e c t . c o m

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / h e

(3)

Dy, Ho, Gd or Er, have been also synthesized from several C14 RMn2Laves [7], but not yet from C14 YMn2, which is

a high-pressure modification. Therefore, one related task was to verify whether YMn2H6hydride can be formed not only

from C15 but also from hexagonal C14 YMn2modification.

The phase transitions in C15 YMn2, which involve an

interplay between magnetism and lattice stability, - and particularly the associated latent heat and deuterization effect - remain to be further elucidated. This study presents relevant results of a calorimetric study. For comparison, the investi-gation was extended to cover C14 YMn2and its derivative,

hydrogenated YMn2H6. Complementary measurements of

magnetization and electrical resistivity were also made.

2. Material and methods

A C15 YMn2 sample was prepared by induction-melting

a mixture of high-purity yttrium and manganese in a stoi-chiometric ratio. X-ray diffraction patterns inFig. 1(a) confirm the cubic Laves phase. The General Structure Analysis System program[8]yielded the lattice constant a and coordi-nates of Y and Mn, following Rietveld refinement, that are listed inTable 1.

YMn2D6was prepared following a previously documented

scheme [6].Table 1 presents the X-ray diffraction data and

Fig. 2(a) displays the corresponding structure with the tabu-lated fitting parameters.

A C14 YMn2sample was also prepared from high-purity

yttrium and manganese in a stoichiometric ratio, but by high-pressure quenching from the melt, followed with annealing. The X-ray diffraction patterns inFig. 1(b) confirm the hexagonal Laves phase. Under hydrogen at 1 GPa and 100 0C for 24 h, a hydride sample of formula YMn2H6was formed.

Interestingly, the X-ray diffraction data and the deduced structure inFig. 2(b) are the same as those of YMn2D6inFig. 2

(a), even though the parent compounds, cubic C15 and hexag-onal C14 RMn2, have different crystal structures. This result

confirms that unique Fm-3m RMn2H6hydrides can be formed

from either C14 or C15 parent RMn2compounds, regardless of

their different properties (structural, magnetic and other). Thermal relaxation-type calorimetric measurements were made between 0.5 and 200 K for C15 YMn2and YMn2D6, but only

above 10 K for C14 YMn2and YMn2H6. In each experiment, an

mg-sized sample was thermally anchored with N-grease to a small sapphire disk. Thin films of RuO2/Al2O3[9]and NieCr

were deposited on the sapphire to serve as a thermometer and heater, respectively. The relatively low heat capacity of this sample holder was separately measured for addenda correc-tions. The sapphire disk was suspended using four AueCu wires from a copper block, which was heated or cooled in steps, allowing the specific heat of the sample to be measured in both directions, revealing any thermal hysteresis. Standard adia-batic calorimetry, which involves only a heating path, does not

Fig. 1e Observed (crosses) and fitted (solid lines) X-ray powder diffraction patterns of (a) C14 and (b) C15 YMn2.

Short dashes represent calculated Bragg reflections of proposed structure. Bottom lines represent deviation between observed and calculated data.

Table 1e List of the fitting parameters of the crystalline structures.

YMn2, Cubic F de3m, a ¼ 7.6806(8) A˚

Atoms x y z Multi Uiso(A˚2) Y 5/8 5/8 5/8 8a 0.89

Mn 0 0 0 16d 0.03

c2_{¼ 1.758 R}

p¼ 3.69% RWP¼ 4.32%

YMn2, Hexagonal P 63/mmc, a¼ 5.4086(6) A˚, c ¼ 8.8341(13) A˚ Atoms x y Z Site Uiso(A˚2) Y 1/3 2/3 0.063787 4f 0.14 Mn(1) 0 0 0 2a 0.24 Mn(2) 0.169 0.339 ¼ 6h 0.41 c2_{¼ 1.762 R}

p¼ 6.28% RWP¼ 9.78% YMn2D6, Cubic F de3m, a ¼ 6.7096(6) A˚

Atoms x y Z Site Uiso(A˚2) Y 1/4 1/4 ¼ 8c 2.98 Mn(1) 1/4 1/4 ¼ 8c 2.98 Mn(2) 0 0 0 4a 2.95 D 0.284(3) 0 0 24e 5.98 c2_{¼ 4.145 R} p¼ 7.30% RWP¼ 9.59% YMn2H6, Cubic F de3m, a ¼ 6.7132(7) A˚

Atoms x y Z Site Uiso(A˚2) Y 1/4 1/4 ¼ 8c 2.42 Mn(1) 1/4 1/4 ¼ 8c 2.42 Mn(2) 0 0 0 4a 2.96 H 0.287(4) 0 0 24e 5.66 c2_{¼ 1.026 R} p¼ 1.87% RWP¼ 2.31% 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 e n e r g y 3 6 ( 2 0 1 1 ) 2 2 8 5e2 2 9 0

2286

(4)

offer this capability. Before each measurement was made, the temperature of the copper block was stabilized. A small amount of Joule heat was introduced; this process was followed by rapid thermal relaxation between the sample and its holder, during which practically no heat was lost to the copper block. The temperature of the sample was fitted to an exponential dependence on the relaxation time, resulting in a time-constants. The heat capacity was then calculated using the equation c¼ ks, where k represents the thermal conductance of AueCu wires as thermal links. The heat capacity of the sample holder was measured separately for addenda correction. The specific heat of sample is thus given by C¼ (cecaddenda) (M/m)

where m and M denote the mass and molar mass of the sample, respectively.

Complementary magnetic measurements, made using a superconducting quantum interference device, yielded susceptibility values x(T ) from 2 to 300 K. Electrical resistivity r(T ) values were obtained using the standard four-probe method.

3. Results and discussion

Exposing the hexagonal C14 YMn2 compound to hydrogen

under high-pressure and high-temperature conditions caused the formation of cubic Fm-3m YMn2H6- identical to cubic C15

YMn2modification[6]. In this hydride, half of the Mn atoms

occupy 4c sites and are surrounded by six deuterium atoms, while the other half are randomly distributed with the Y atoms on 8c sites. YMn2H6and generally RMn2H6are isostructural

with complex metal hydrides, such as Mg2FeH6, but with the

Mg sites replaced by randomly substituted R and Mn atoms[7]. In the present study, the mechanism by which this radical structural change can affect other properties of the two allo-tropic (C14 and C15) forms of YMn2is of particular interest.

InFig. 3, the magnetic susceptibility of C15 YMn2exhibits

thermal hysteresis. It is clearly associated with the afore-mentioned structural and magnetic transitions near 100 K. As the temperature is lowered to around 85 K, an abrupt decline in x reflects antiferromagnetic ordering. Upon heating, the anomaly extends to almost 120 K. These observations are consistent with those made by Wang et al. [6]. However, Nakamura et al.[10]reported similar thermal hysteresis near 100 K, but lower susceptibility values by almost a factor of two. Moreover, they found a positive dependence of susceptibility on temperature above TN, which differs from the negative

temperature dependence inFig. 3. These discrepancies may reflect variations in sample quality, one aspect of which is the Y/Mn off-stoichiometry, given the rather complex mag-netic interactions and fluctuations of Mn moments in this compound.Fig. 3also plots magnetic susceptibility data for YMn2D6. The values are almost four times those for YMn2,

and are consistent with the results of Wang et al.[6]. Most importantly, no phase transition occurs under 300 K.

Sample cracks that resulted from the large change in volume close to 100 K did not influence magnetic suscepti-bility, but caused serious difficulty in determining electrical resistivity. The making of the electrical resistivity measure-ments inFig. 4began with cooling of the sample from 300 to 4 K, and continued with heating back to 300 K. As the temperature declined, the resistivity remained almost constant down to 75 K, at which temperature the dramatic increase in resistivity reflected micro-cracks in the sample Fig. 2e Refined XRD patterns of (a) hydrogenated C14 and

(b) deuterized C15 YMn2with proposed crystal structures.

Fig. 3e Thermal hysteresis in magnetic susceptibility of C15 YMn2reveals phase transition close to 100 K, which is

(5)

that were produced by the large change in volume. Although the above mentioned anomaly was extended to close to 30 K, the absolute values of the non-reversible and difficult-to-reproduce transport property data should not be concerned for further interpretation. Indeed, another anomalous signif-icant increase in apparent resistivity in the heating cycle began only after the temperature of the sample reached 103 K, and ended at around 125 K. Accordingly, a large difference between the resistivity values at 300 K at the beginning and the end of the thermal cycle is not surprising.

Accurately determining specific heat, one of the most basic thermodynamic quantities, is typically not difficult, especially if a bulk sample is available. At the beginning of this investi-gation, however, a bulk C15 YMn2sample was literally

shat-tered as a result of the large changes in volume during the phase transitions. When such an event occurred, some unknown portions of the sample lost thermal contact, and Joule heating could not reach them. Calculations of the specific heat based on the total sample mass then become meaning-less. To solve this problem, a bulk sample was crushed into powder, which was then pressed using N-grease to form a pellet for measurement. The heat capacity addenda correc-tion included that for the added grease.Fig. 5presents the results.

At low temperatures of 0.5e12 K, the total specific heat could be fitted to the sum of an electronic (Ce) and a lattice (Cl)

contribution that depended on T and T3, respectively:

C¼ Ceþ Cl¼ gT þ bT3; (1)

Or

C=T ¼ g þ bT2 (2)

The linear fit of C/T vs. T2 in the inset ofFig. 5yields coef-ficients g¼ 4.44 mJ/mol K2 and b ¼ 0.667 mJ/mol K4. The latter corresponds to a Debye temperature qD of 210 K. These

parameters are somewhat lower than those reported by

Okamoto et al.[11]and Imai et al.[12]. The differences are understandable, considering the problem regarding thermal contact that was caused by micro-cracks arises in all samples to different degrees.

Anomalous specific heat is clearly observed herein as in earlier studies [11,12]. However, as aforementioned, relaxa-tion-type calorimetry allows measurements to be made along both heating and cooling paths, yielding a broad thermal hysteresis between 70 and 120 K. The specific heat anomaly during heating is expected to reveal a positive latent heat L, superimposed on the background contributions (Clþ Ce),

corresponding to endothermic volume expansion. Lower specific heat values upon cooling are expected, reversing the effect that is associated with the exothermic volume contraction. A latent heat value of 330 J/mol was obtained by simple integration. L¼ Z ðClþ CeÞdTÞ þ L Z ðClþ CeÞdTÞ L 2 ¼ Z ðCdTÞheating Z ðCdTÞcooling 2 (3)

In general, low-temperature structural transformations in solids are not associated with a latent heat that is as high as this value. For example, L from cubic to tetragonal RbCaF3is

only 36 J/mol at 198 K [13]. Given the excessive thermal shrinkage with spontaneous magnetostriction [3], the extra contribution could be a magnetic term TDSm, whereDSmis the

change in entropy that is associated with antiferromagnetic ordering.DSmis typically calculated from Rln(2Sþ 1) ¼ 23 J/

mol K for Mn moments with S ¼ 3/2. However, a similar calculation does not apply here because of the itinerant nature of the Mn moments. One calorimetric study, which involved only heating, yielded DSm¼ 3.9 J/mol K[11]. At an

average transition temperature of 100 K, the magnetic-heat term would thus be 390 J/mol, which is reasonably consistent with the result in this study. This result may also explain the Fig. 4e Temperature-dependence of electrical resistance of

C15 YMn2close to 100 K reveals thermal hysteresis.

Thermal contraction/expansion-induced micro-cracks in the sample resulted in uncertainty in absolute electrical resistivity.

Fig. 5e Hysteretic temperature-dependence of specific heat of C15 YMn2provides a measure of latent heat of

phase transition close to 100 K. This transition is clearly suppressed in YMn2D6. Inset: low-temperature data fitted

to Eq.(2), yielding g and b values.

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 e n e r g y 3 6 ( 2 0 1 1 ) 2 2 8 5e2 2 9 0

(6)

broad transition in the specific heat data. Generally, low-temperature structural changes in solids occur via a lattice shear process, rather than a nucleate-and-growth mechanism that is more strongly related to high-temperature structural transformations. Therefore, low-temperature structural changes are associated with a relatively narrow range of temperatures. For example, the sharp spike anomaly in the specific heat of RbCaF3has a half-width of only 0.5 K[13].

However, magnetic ordering typically extends over a much broader observable range of temperatures, as in the case herein.

Fig. 5 presents specific heat data for the deuteride compound YMn2D6. It reveals a smooth dependence of

specific heat on temperature within measured range, ruling out any magnetic or structural transformation. Such a very large difference from the parent compound C15 YMn2

presumably reflects the fact that YMn2H6 hydride can be

regarded not as an interstitial metal hydride but rather as a coordination compound[7]. This claim is further supported by the magnetic susceptibility data inFig. 3, as well as by the absence of any sample cracking in the cooling/heating cycle during which the calorimetric measurements were made.

C14 YMn2 has an MneMn distance that is close to the

critical value for the onset of a magnetic moment, but stron-gly temperature-enhanced spin fluctuations cause para-magnetism even at 4.2 K [14]. C14 YMn2 has, having

a thermal expansion coefficient that exceeds those of other C14 compounds, and so does not undergo structural trans-formation in the same manner as C15 RMn2[15]. Indeed, the

specific heat data for C14 YMn2inFig. 6, are consistent with

those of Imai et al.[16], increasing smoothly with tempera-ture, as expected.

The temperature-dependence of the specific heat of YMn2H6 inFig. 6 is almost identical to that of C14 YMn2.

Apparently, hydrogenation only slightly affects the thermal or magnetic behavior of C14 YMn2, which can be related to

the paramagnetic behavior of both C14 YMn2 and

YMn2H6throughout the studied range of temperatures.

4. Conclusions

YMn2H6or YMn2D6can be obtained not only from cubic C15

but also from the hexagonal C14 phase of YMn2. A latent heat

of 330 J/mol was determined for the hysteretic phase transi-tion in cubic C15 YMn2 near 100 K. This transition is

sup-pressed in YMn2D6, and is shown calorimetrically to be absent

in hexagonal C14 YMn2and YMn2H6.

Acknowledgment

The authors would like to thank the National Science Council of the Republic of China, Taiwan, for financially supporting this research under Contract No. 97-2113-M-002-012-MY3. The MNiSW Grant No. N N204 527939 is also gratefully acknowledged.

r e f e r e n c e s

[1] Leblond T, Paul-Boncour V, Cuevas F, Isnard O, Ferna´ndezca JF. Study of the multipeak deuterium thermodesorption in YFe2Dx(1.3 x 4.2) by DSC, TD and

in situ neutron diffraction. Int J Hydrogen Energy 2009;34: 2278e87.

[2] Tarnawski Z, Kolwicz-Chodak L, Figiel H, Kim-Ngan N-TH, Koz1owski A, Dawid T, et al. Specific heat of YMn2Hx

and HoMn2Hx. J Alloys Compd 2007;446e447:415e8.

[3] Shiga M, Wada H, Nakamura Y. Magnetism and thermal expansion anomaly of RMn2(R¼ Y, Gd, Tb, Ho and Er).

J Magn Magn Mater 1983;31e34:119e20.

[4] Fujii H, Saga M, Okamoto T. Magnetic, crystallographic and hydrogen absorption properties of YMn2and ZrMn2

hydrides. J Less-Common Met 1987;130:25e31. [5] Przewoznik J, Paul-Boncour V, Latroche M,

Percheron-Guegan A. Structural study of YMn2hydrides. J Alloys Compd

1995;225:436e9.

[6] Wang CY, Paul-Boncour V, Kang CC, Liu RS, Filipek SM, Dorogova M, et al. The novel YMn2D6deuteride synthesized

under high pressure of gaseous deuterium. Solid State Commun 2004;130:815e20.

[7] Filipek SM, Paul-Boncour V, Kuriyama N, Takeichi N, Tanaka H, Liu R-S, et al. Hydrides of Laves phases

intermetallic compounds synthesized under high hydrogen pressure. Solid State Ionics 2010;181:306e10.

[8] Larson AC, Von Dreele RB. General structure analysis system (GSAS). Report LAUR 86. Los Alamos National Laboratory; 2004.

[9] Chen YY. Temperature: its measurement and control in science and industry, vol. 7. New York: American Institute of Physics; 2003. 387.

[10] Nakamura Y, Shiga M, Kawano S. Antiferromagnetism of YMn2intermetallic compound. Physica B 1983;120:

212e5.

[11] Okamoto T, Nagata H, Fujii H, Makihara Y. Heat capacities of Laves-phase compounds RMn2R¼Y, Gd and Er. J Magn Magn

Mater 1987;70:139e41. Fig. 6e Temperature-dependencies of specific heat of C14

YMn2and YMnH6are similar. Both compounds are

paramagnetic and do not exhibit phase transition-type thermal hysteresis.

(7)

[12] Imai H, Wada H, Shiga M. Effects of spin fluctuations on the specific heat in YMn2and Y0.97Sc0.03Mn2. J Phys Soc Jpn 1995;

64:2198e202.

[13] Ho JC, Unruh WP. Specific heat of RbCaF3below 300 K. Phys

Rev B 1976;13:447e50.

[14] Inoue K, Nakamura Y, Tsvyashchenko AV, Fomicheva L. Magnetism of C14-RMn2(R¼ Y, Gd, Tb, Dy and Yb). J Magn

Magn Mater 1995;140-144:797e8.

[15] Inoue K, Nakamura Y, Ikeda Y, Bando Y,

Tsvyashchenko AV, Fomicheva L. Thermal expansion of C14-Laves phase RMn2(R¼Y, Gd, Dy and Yb). J Phys Soc Jpn

1995;64:4901e5.

[16] Imai H, Wada H, Nakamura H, Shiga M, Inoue K,

Nakamura Y, et al. Low temperature specific heat of RMn2

compounds with the hexagonal C14 structure (R¼Y, Yb and Th). J Phys Soc Jpn 1997;66:2175e7.

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 e n e r g y 3 6 ( 2 0 1 1 ) 2 2 8 5e2 2 9 0