- Email: [email protected]
Crystal structure and its stability in CeCuAl3 single crystal
Intermetallics 46 (2014) 126e130
Contents lists available at ScienceDirect
Intermetallics journal homepage: www.elsevier.com/locate/intermet
Crystal structure and its stability in CeCuAl3 single crystal a M. Klicpera a, *, P. Javorský a, P. Cermák , A. Rudajevová a, S. Danis a, T. Brunátová a, b I. Císarová a b
Charles University in Prague, Faculty of Mathematics and Physics, Department of Condensed Matter Physics, Ke Karlovu 5, 121 16 Prague 2, Czech Republic Department of Inorganic Chemistry, Faculty of Science, Charles University in Prague, Hlavova 8, 128 43 Prague 2, Czech Republic
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 September 2013 Received in revised form 4 November 2013 Accepted 7 November 2013 Available online 30 November 2013
We present the crystal structure investigation of CeCuAl3 compound. Our X-ray diffraction study confirms unambiguously BaNiSn3-type structure as the crystal structure of CeCuAl3 in whole temperature range. Moreover, phase transition around 300 C was found by our high-temperature powder X-ray diffraction measurement. This transition preserves BaNiSn3-type structure, only structural parameters change during the transition. The study is provided on single crystalline and polycrystalline samples as well as on La counterpart to compare mainly high-temperature behavior of these compounds. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: A. Rare-earth intermetallics B. Crystallography D. Site occupancy F. Analysis, chemical F. Diffraction F. Calorimetry
1. Introduction Intermetallic compounds containing Ce have attracted much attention due to the large variety of physical phenomena they present, including different types of magnetic order, valence fluctuations, heavy-fermion behavior or unconventional superconductivity. The exceptional behavior of these compounds originates in the vicinity of the energy of the cerium 4f shell level to the 5d and 6s levels and the unique physical properties are mainly associated with the instability of the 4f states. The tetragonal compounds CeTX3 or more generally CeTxX4x, where T is a transition dmetal and X is a p-metal, represent a model system where we meet various ground states and phenomena depending on the actual chemical composition or applied external pressure [1e5]. Particularly remarkable is the observation of the pressure induced superconductivity in the non-centrosymmetric BaNiSn3-type crystal structure of antiferromagnetically ordered CeRhSi3 and CeIrSi3 [6,7]. Furthermore, recent inelastic neutron scattering experiment revealed another highly interesting feature: the strong electrone phonon interaction in CeCuAl3 leads to a formation of a vibron quasibound state [8].
* Corresponding author. Tel.: þ420 221911479. E-mail address: [email protected] (M. Klicpera). 0966-9795/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.intermet.2013.11.004
Despite numerous studies on CeCuAl3, some of its basic properties remain unclear, including still some ambiguity concerning its crystal structure, what is a fundamental point when discussing some of its electronic properties. The stoichiometric CeTX3 compounds with X ¼ Si or Ge crystallize in the BaNiSn3-type structure what is the ordered non-centrosymmetric variant of the BaAl4 tetragonal structure. However, for compounds with X ¼ Ga or Al also the disordered variants BaAl4 or ThCr2Si2 structure are often considered (see Fig. 1 for comparison of all three structures). CeCuAl3 was first reported to crystallize in the ThCr2Si2-type structure [9] with Cu and one-third of the Al atoms randomly distributed over the 2a-positions of this structure. This conclusion was adopted and confirmed by several later studies [10,11], some papers mention also the BaAl4 structure without deeper analysis [12]. Contrary to these earlier studies, the ordered BaNiSn3-type structure was concluded from the powder neutron diffraction data [13]. This is rather reliable result as the neutron scattering length for Ce is much smaller than for Cu (and comparable with Al) atoms, which allows for a more accurate Cu/Al site occupation assignment compared to powder X-ray diffraction, where the Ce contribution dominates. Subsequently many further papers adopt the BaNiSn3 structure [5,14,8]. On the other hand, some recent studies stated again the BaAl4 structure [15]. All the previous structural studies are based on polycrystalline data. To resolve the structure type in detail, we have performed a thorough study based on both the
M. Klicpera et al. / Intermetallics 46 (2014) 126e130
127
Fig. 1. The tetragonal disordered BaAl4- and two its derivatives: centrosymmetric ThCr2Si2- (I4/mmm) and ordered non-centrosymmetric BaNiSn3-type structure (I4mm).
powder and mainly the single crystal diffraction. We note that the single crystal diffraction favored the BaAl4 structure in the case of related CeCuGa3 compound [3]. As the electronic properties of CeCuAl3 might be sensitive to the sample thermal treatment [11], we have studied the structure of both, as-cast and annealed samples and include additionally diffraction at elevated temperatures which finally led to some non-trivial results. 2. Experimental methods Single-crystalline CeCuAl3 sample was prepared by Czochralski growing method using the tri-arc furnace under protection of argon atmosphere. Polycrystalline sample of 8 g weight prepared by arcmelting from pure elements (2N8 for Ce, 4N for Cu and 5N for Al) was used as the precursor. Small polycrystalline sample of nonmagnetic analog LaCuAl3 was also prepared, the purity of La was 3N. Prepared polycrystalline samples were annealed for 8 days at 900 C in quartz tubes, the rate of heating the sample from room temperature was 3 C per minute, the rate of cooling down from 900 C was 1 C per minute. The single crystal was first investigated in its as-cast form and subsequently it has been annealed in the same way. The chemical composition of prepared samples was investigated by scanning electron microscope equipped with a backscattered electron (BSE) detector and with an energy dispersive X-ray (EDX) analyzer. The crystal structure investigation of poly- and singlecrystalline samples at room temperature was performed by powder X-ray diffraction using the Cu K-a radiation. High-temperature X-ray diffraction measurements were performed on powdered single crystal sample using the PANanalytical MPD diffractometer with MRI high-temperature chamber in conventional BB symmetric qe2q scan. Tantalum strip heater and platinum radiation heater were used as a heating elements, the heating rate was 5 C per minute. The sample was directly put to tantalum strip heater in order to have the good thermal contact. The pressure in heating chamber was of the order 102 Pa to protect the tantalum strip from corrosion and to ensure the same conditions for measurements at each temperature. The measurement was done in 2q range between 10 and 80 with the step of 0.025 at temperatures from 32 to 500 C. The diffraction peaks belonging to Ta were subtracted from diffraction data. The refinement of the structural parameters from the diffraction patterns was done using the Rietveld analysis employing the Fullprof program [16]. The quality and orientation of the single crystal was verified by X-ray diffraction using the Laue method. Laue neutron diffraction performed on Orient Express equipment in Institute Laue-Langevin (ILL), Grenoble, France, was used in the second step to check the quality in the whole volume of the prepared single crystal. Each measurement was taken in backscattering distance 68.0 mm from the middle of the sample for 20 min with 1.5 mm collimation.
Fig. 2. Neutron Laue patterns of the CeCuAl3 single crystal oriented along the [001] direction taken with Orient Express at ILL. The significant crystallographic directions [100] and [110] of tetragonal BaNiSn3-type structure are shown in picture.
The crystal structure of single crystal was determined by X-ray diffraction using the Bruker APEXII CCD diffractometer equipped with Mo X-ray tube. Structure factors have been extracted from single crystal diffraction pattern and the structure was identified using SHELXS-97 (Sheldrick, 2008). The X-ray diffraction patterns on CeCuAl3 single crystal of the prism shape with (0.080 0.100 0.140 mm) dimension were done at temperatures 296 and 150 K. Differential scanning calorimetry (DSC) was carried out on CeCuAl3 single crystal as well as on LaCuAl3 polycrystalline sample in order to obtain melting temperatures and to find any sign of presence of other transition as well as to verify the quality of the sample. Measurements were performed under protection of He atmosphere on SETSYS Evolution 24 instrument (SETARAM Instrumentation
Table 1 Structure parameters of CeCuAl3 single crystal determined from X-ray diffraction at room temperature and at 150 K. Vf.u. is the volume per fundamental unit. zCe, zCu, zAl and zAlb are fraction coordinates atoms in the fundamental unit (see subsection 3.1). R is the agreement factor for the fit of diffraction patterns (goodness of fit on F2). The thermal displacements of atoms are described by: 2p2 (h2a*2U11 þ . þ 2hka*b*U12), where matrix elements U12, U13 and U23 are equal to zero. SHELXS-97 program employing the Rietveld analysis was used for the determination of structural parameters. T (K)
150
296
a (pm) c (pm) c/a Vf.u. (106 pm3) Atomic positions: 2Ce in 2(a) 2Cu in 2(a) 2Al in 2(a) 4Al in 4(b) Thermal displacement: U11 (Ce) U22 (Ce) U33 (Ce) U11 (Cu) U22 (Cu) U33 (Cu) U11 (Al) U22 (Al) U33 (Al) U11 (Alb) U22 (Alb) U33 (Alb) R (%)
425.27(3) 1066.80(7) 2.5085(2) 96.47(2)
426.20(2) 1068.06(5) 2.5060(2) 97.01(2)
(0, (0, (0, (0,
(0, (0, (0, (0,
0, 0 (fix)) 0, 0.0946(3)) 0, 0.3689(1)) 1/2, 0.2508(2))
0.0032(2) 0.0032(2) 0.0041(2) 0.0068(4) 0.0068(4) 0.0093(7) 0.006(1) 0.006(1) 0.0061(5) 0.0056(8) 0.0056(8) 0.011(2) 2.36
0, 0 (fix)) 0, 0.0942(2)) 0, 0.36909(8)) 1/2, 0.2505(2))
0.00490(7) 0.00490(7) 0.0069(1) 0.0097(2) 0.0097(2) 0.0108(3) 0.0086(9) 0.0095(9) 0.0092(3) 0.0077(4) 0.0077(4) 0.0128(9) 1.12
128
M. Klicpera et al. / Intermetallics 46 (2014) 126e130
Fig. 3. The SEM image (upper left part) of CeCuAl3 single crystal surface. The showed area of surface contains some scratches because of polishing of the surface. Remaining three parts show the distribution of atoms obtained by the EDX analysis in the same surface area.
Company) and data analysis were done using SETARAM software. Heating and cooling scans were taken with the rate of 10 C per minute at temperature range from 28 to 1400 C. 3. Results 3.1. Single crystal characterization Laue X-ray and neutron diffraction performed on the grown CeCuAl3 single crystal verified the quality and tetragonal symmetry of the sample (see Fig. 2). However, there are still several structure types which can be adopted by the studied compound as mentioned in the Introduction. The centrosymmetric BaAl4 (I4/ mmm) and two its derivatives e disordered form of ThCr2Si2 (I4/ mmm) structure and the non-centrosymmetric BaNiSn3 (I4mm) e are displayed in Fig. 1. The only difference between them is a distribution of Cu and Al atoms in the lattice, while the Ce atomic positions are the same in all three structure types. In the BaAl4type, Cu and Al atoms are randomly distributed over the 2a and 4b sites. In the ThCr2Si2-type, the random distribution remains on the 2a-positions, whereas the 4b sites are occupied exclusively by Al. Both structures are clearly centrosymmetric. The noncentrosymmetric BaNiSn3 is completely ordered structure type with following atomic coordinates (see also Table 1): 1; 1; 1 2 2 2
þ zCe
2Cu in 2ðaÞ : ð0; 0; zCu Þ;
1; 1; 1 2 2 2
þ zCu
4Alin4ðbÞ: 0; 12;zAlb ; where zCe z 0 and
1; 1; 1 þ z Al 2 2 2
1;0;z Alb 2
zAlb z14
;
3.2. DSC and high-temperature X-ray diffraction Some properties of CeCuAl3 are slightly influenced by sample thermal treatment [11]. These changes can be ascribed to the annealing process which generally improves the quality of the sample, i.e. reduces the residual resistivity and also improves the thermal properties of the sample. To optimize the sample thermal treatment, the knowledge of the phase diagram and mainly the melting temperature is essential to choose a correct one. We have performed DSC measurement to reveal the melting temperature. Surprisingly, we have observed a further phase transition in course of these measurements.
2Ce in 2ðaÞ : ð0; 0; zCe Þ;
2Al in 2ðaÞ : ð0; 0; zAl Þ;
The powder X-ray diffraction patterns of CeCuAl3 powderized single crystal recorded at room temperature confirms that this compound crystallizes in the tetragonal structure. The best fit of the diffraction pattern was obtained for the BaNiSn3-type structure (RBragg ¼ 6.9% compared to RBragg ¼ 31.3% for the ThCr2Si2-type). Introducing certain degree of CueAl disorder to the BaNiSn3 structure did not lead to any improvement of the fit. We examined closely also the diffraction patterns of polycrystalline sample with similar result. These results, pointing to the BaNiSn3 structure, are well consistent with the neutron diffraction data [13]. However, we should have in mind somewhat limited sensitivity of powder X-ray diffraction to Cu and Al positions and CueAl disorder. The crystal structure was then uniquely determined by single crystal diffraction (295 independent reflections, Rint ¼ 0.0169) as tetragonal BaNiSn3-type structure (space group I4mm, 107) without any sign of disorder in Cu and Al atomic positions. We should note, the best structure solution were obtained when a racemic-twinning have been introduced. It seems that there exists two equally populated domains which are connected by a mirroring, where the mirror plane is perpendicular to the four-fold axis. The refined structure parameters of studied CeCuAl3 single crystal at room temperature and at 150 K are summarized in Table 1. The thermal displacements of atoms are given in Table 1 as well. The chemical composition investigation of our single and polycrystalline CeCuAl3 samples done using the EDX analysis confirms the stochiometric composition Ce:Cu:Al ¼ 1:1:3 (with 0.01 error of analysis). The distribution of elements was also checked by EDX equipment, homogeneous distribution of elements can be seen in Fig. 3.
1;1; 1 þz Alb 2 2
; 1; 12; 12 þzAlb ;
in the case of BaNiSn3-type structure.
Fig. 4. Differential scanning calorimetry performed on as-cast and annealed CeCuAl3 single crystal as well as on polycrystalline LaCuAl3. Heating/cooling rate was 10 C per minute. The inset shows data for annealed single crystal up to melting temperature.
M. Klicpera et al. / Intermetallics 46 (2014) 126e130
Fig. 5. The lattice parameters of the annealed single crystal CeCuAl3 obtained by hightemperature X-ray diffraction. The error for each lattice parameter is almost constant during the heating/cooling process, for this reason and for better lucidity we show error bars only for the point at 500 C. The agreement factor of the fit of our model to the experimental data was almost the same (RBragg z 6.9%) for all measurements. See text for more details.
The DSC temperature scans revealed the melting temperature of 1275 C (see inset of Fig. 4) what justifies the optimal annealing temperature around 900 C. Additionally, the measured scans, represented in Fig. 4, show a clear l-peak between 279 and 335 C which corresponds to the structural phase transition in the sample. The transition temperature of 320 C was obtained in usual way for this type of transition, i.e. as the temperature where the heating peak has minimum. This anomaly is observed also for cooling and is reproduced in all measurements independently on maximal temperature of the scan (500 Ce900 C), even not by achieving the melting temperature. The change of enthalpy associated with this transition is higher for heating (1.353 J g1) than for cooling (0.959 J g1) what might suggest a certain change of structure after the heatingecooling cycle. Further DSC measurement on polycrystalline sample showed the same behavior. The lanthanum analog LaCuAl3 exhibits signs of similar anomaly during the heating cycle, nevertheless much weaker as shown in Fig. 4. The enthalpy corresponding to this phase transition reaches only 0.153 J g1, nearly 10 times smaller value than for CeCuAl3. No anomaly is observed during the cooling process in LaCuAl3. The same behavior was reproduced by second subsequent cycle on the same LaCuAl3 sample. The high-temperature powder X-ray diffraction was performed on as-cast and annealed CeCuAl3 single crystal, as well as on annealed polycrystalline CeCuAl3 and LaCuAl3 samples to reveal the microscopic origin of the anomaly observed by DSC. Results of the refinement of diffraction patterns using the Fullprof program [16] are represented in Figs. 5 and 6. Linear increase of lattice parameters with increasing temperature is observed up to 250 C, the usual effect of thermal expansion takes place. The atomic fraction coordinates, i.e. zCu and zAl, are unaffected up to 250 C. The transition seen in DSC is then reflected by anomalous development of the c
129
Fig. 6. The atomic fraction coordinates of annealed single crystal CeCuAl3 and LaCuAl3 polycrystal obtained by high-temperature X-ray diffraction. The error of determination of atomic positions is z6$103. The lines are to guide the eye. See text for more details.
parameter (or c/a ratio) around 300 C. The lattice parameters increase again linearly above z350 C which suggests the stable high-temperature phase, in agreement with DSC analysis. The measurement during the cooling to room temperature shows very similar development of lattice parameters but the transition from high-temperature to low-temperature phase does not lead to the original values of lattice parameters, see Fig. 5. The lattice parameter a has higher value and parameter c has lower value compared to initial one before the heating/cooling cycle. The change can be
Fig. 7. The BaNiSn3-type structure adopted by CeCuAl3 compound below 300 C. The atomic fraction coordinates are changed above 300 C as denote the smeared Cu and Al atoms with arrows. For better clarity, the shift is 10 times increased compared to the real change. The illustrative marking of atoms is done for easier understanding the Table 2.
130
M. Klicpera et al. / Intermetallics 46 (2014) 126e130
Table 2 Interatomic distances in CeCuAl3 compounds obtained by employment of the Rietveld analysis at temperature below and above structural phase transition. The patterns taken on powder sample were processed with Fullprof program [16]. The atomic positions are labeled as shown in Fig. 7. T ( C)
CeeCu (pm)
CeeAl1 (pm)
CeeAl2 (pm)
CueAl1 (pm)
CueAl2 (pm)
Al1eAl2 (pm)
500 100
329.0(5) 329.7(5)
326.9(6) 322.4(7)
342.7(7) 341.2(7)
250.1(6) 250.8(6)
255.8(7) 249.3(7)
258.9(6) 261.2(6)
parameter a and from 10.672 to 10.720$102 pm for parameter c), without any trace of phase transition. The development of fraction coordinates zCu and zAl of LaCuAl3 is plotted in Fig. 6. Also here we are not able to see any clear hint of phase transition as seen in CeCuAl3. If any such transition exists in LaCuAl3, what might indicate the little anomaly in DCS measurements, it is below the experimental sensitivity of our X-ray diffraction analysis. 4. Conclusions
expressed as y0.05% of their initial values. Although the lattice parameters in CeCuAl3 are different after heating/cooling cycle, the volume of fundamental unit stays almost unaffected because the parameters a and c change in opposite way. The transition around z300 C is much better seen in development of atomic fraction coordinates zCu and zAl plotted in Fig. 6. The difference between these coordinates at temperatures below and above 300 C is z6$103 what represents almost 2% change of the value of these coordinates (see also illustrative picture of changed atomic coordinates in Fig. 7 and changes of interatomic distances given in Table 2). We are aware that the numerical error of determination of these values is also approximately 6$103. Furthermore, the uncertainty of determination of atomic positions from powder diffraction data is higher than only numerical error. In powder diffraction patterns, we face the problem of overlapping of the diffraction lines which is not present when using the single diffraction data. The overlapping can strongly affect, together with diffraction data quality, the accuracy of parameters of the structural model. The atomic positions determined using single crystal diffraction are much more accurate. The values obtained from powder diffraction are bounded with relatively high error and here serve mainly to follow qualitatively their temperature development. Despite relatively high error, the tendency, scatter of the values and mainly reproducibility of the effect on several samples is clear enough to conclude that the phase transition is present in CeCuAl3. The phase transition was observed for both, as-cast and annealed samples, as well as for polycrystalline sample, at the same temperature range, i.e. between 270 and 340 C. The changes in structure parameters corresponding to this transition are very similar for all samples. Especially similar changes in as-cast and annealed samples suggest that each heating/cooling cycle changes the structure parameters independently on temperature, heating/ cooling rate and duration of sample annealing. More detailed comparison of this effect on as-cast and annealed samples will be published separately [17]. The refinement of diffraction patterns of LaCuAl3 polycrystal shows ordinary development of lattice parameters, i.e. linear increase with increasing temperature (from 4.295 to 4.325 .102 pm for
Our study of CeCuAl3 single crystal unambiguously confirmed the ordered non-centrosymmetric BaNiSn3-type structure as the crystal structure of this compound. Moreover, we found the structural phase transition around 300 C. This transition is related to a small change of structural parameters while the crystal structure is preserved. Acknowledgments The work of M.K. was supported by Grant agency of Charles University under projects 76213 and SVV-2013-267303. We acknowledge ILL for the allocation of time and technical services. References [1] Hillier AD, Adroja DT, Manuel P, Anand VK, Taylor JW, McEwen KA, et al. Phys Rev B 2012;85:134405. [2] Paschen S, Felder E, Ott HR. Eur Phys J B 1998;2:169. [3] . Joshi Devang A, Burger P, Adelmann P, Ernst D, Wolf T, Sparta K, et al. Phys Rev B 2012;86:035144. [4] Haen P, Lejay P, Chevalier B, Lloret B, Etourneau J, Sera M. J Less-Common Met 1985;110:321. [5] Mock S, Pfleiderer C, Lohneysen H v. J Low Temp Phys 1999;115:1. [6] Sugitani I, Okuda Y, Shishido H, Yamada T, Thamhavel A, Yamamoto E, et al. J Phys Soc Jpn 2006;75:043703. [7] Kimura N, Ito K, Saitoh K, Umeda Y, Aoki H, Terashima T. Phys Rev Lett 2005;95:247004. [8] Adroja DT, del Moral A, de la Fuente C, Fraile A, Goremychkin EA, Taylor JW, et al. Phys Rev Lett 2012;108:216402. [9] Zarechnyuk OS, Kripyakevich PL, Gladyshevskii EL. Sov Phys Crystallogr 1965;9:706. [10] Mentink SAM, Bos NM, van Rossum FJ, Nieuwenhuys GJ, Mydosh JA, et al. J Appl Phys 1993;73:6625. [11] Kontani M, Ido H, Ando H, Nishioka T, Yamaguchi Y. J Phys Soc Jpn 1994;63:1652. [12] Bauer E, Pillmayr N, Gratz E, Hilscher G, Gignoux D, Schmitt D. Z Phys B 1987;67:205. [13] Moze O, Buschow KHJ. J Alloy Compd 1996;245:112. [14] Nishioka T, Kawamura Y, Kato H, Matsumura M, Kodama K, Sato NK. J Magn Magn Mater 2007;310:e12. [15] Hu XD, Zhou HY, Li JB, Pan SK, Wang T, Yao QR, et al. Intermetallics 2009;17:775. [16] Rodriguez-Carvajal J. Phys B 1993;192:55. [17] Klicpera M, Javorský P, Danis S, Brunátová T. Acta Physica Polonica. [submitted for publication].
Contents lists available at ScienceDirect
Intermetallics journal homepage: www.elsevier.com/locate/intermet
Crystal structure and its stability in CeCuAl3 single crystal a M. Klicpera a, *, P. Javorský a, P. Cermák , A. Rudajevová a, S. Danis a, T. Brunátová a, b I. Císarová a b
Charles University in Prague, Faculty of Mathematics and Physics, Department of Condensed Matter Physics, Ke Karlovu 5, 121 16 Prague 2, Czech Republic Department of Inorganic Chemistry, Faculty of Science, Charles University in Prague, Hlavova 8, 128 43 Prague 2, Czech Republic
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 September 2013 Received in revised form 4 November 2013 Accepted 7 November 2013 Available online 30 November 2013
We present the crystal structure investigation of CeCuAl3 compound. Our X-ray diffraction study confirms unambiguously BaNiSn3-type structure as the crystal structure of CeCuAl3 in whole temperature range. Moreover, phase transition around 300 C was found by our high-temperature powder X-ray diffraction measurement. This transition preserves BaNiSn3-type structure, only structural parameters change during the transition. The study is provided on single crystalline and polycrystalline samples as well as on La counterpart to compare mainly high-temperature behavior of these compounds. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: A. Rare-earth intermetallics B. Crystallography D. Site occupancy F. Analysis, chemical F. Diffraction F. Calorimetry
1. Introduction Intermetallic compounds containing Ce have attracted much attention due to the large variety of physical phenomena they present, including different types of magnetic order, valence fluctuations, heavy-fermion behavior or unconventional superconductivity. The exceptional behavior of these compounds originates in the vicinity of the energy of the cerium 4f shell level to the 5d and 6s levels and the unique physical properties are mainly associated with the instability of the 4f states. The tetragonal compounds CeTX3 or more generally CeTxX4x, where T is a transition dmetal and X is a p-metal, represent a model system where we meet various ground states and phenomena depending on the actual chemical composition or applied external pressure [1e5]. Particularly remarkable is the observation of the pressure induced superconductivity in the non-centrosymmetric BaNiSn3-type crystal structure of antiferromagnetically ordered CeRhSi3 and CeIrSi3 [6,7]. Furthermore, recent inelastic neutron scattering experiment revealed another highly interesting feature: the strong electrone phonon interaction in CeCuAl3 leads to a formation of a vibron quasibound state [8].
* Corresponding author. Tel.: þ420 221911479. E-mail address: [email protected] (M. Klicpera). 0966-9795/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.intermet.2013.11.004
Despite numerous studies on CeCuAl3, some of its basic properties remain unclear, including still some ambiguity concerning its crystal structure, what is a fundamental point when discussing some of its electronic properties. The stoichiometric CeTX3 compounds with X ¼ Si or Ge crystallize in the BaNiSn3-type structure what is the ordered non-centrosymmetric variant of the BaAl4 tetragonal structure. However, for compounds with X ¼ Ga or Al also the disordered variants BaAl4 or ThCr2Si2 structure are often considered (see Fig. 1 for comparison of all three structures). CeCuAl3 was first reported to crystallize in the ThCr2Si2-type structure [9] with Cu and one-third of the Al atoms randomly distributed over the 2a-positions of this structure. This conclusion was adopted and confirmed by several later studies [10,11], some papers mention also the BaAl4 structure without deeper analysis [12]. Contrary to these earlier studies, the ordered BaNiSn3-type structure was concluded from the powder neutron diffraction data [13]. This is rather reliable result as the neutron scattering length for Ce is much smaller than for Cu (and comparable with Al) atoms, which allows for a more accurate Cu/Al site occupation assignment compared to powder X-ray diffraction, where the Ce contribution dominates. Subsequently many further papers adopt the BaNiSn3 structure [5,14,8]. On the other hand, some recent studies stated again the BaAl4 structure [15]. All the previous structural studies are based on polycrystalline data. To resolve the structure type in detail, we have performed a thorough study based on both the
M. Klicpera et al. / Intermetallics 46 (2014) 126e130
127
Fig. 1. The tetragonal disordered BaAl4- and two its derivatives: centrosymmetric ThCr2Si2- (I4/mmm) and ordered non-centrosymmetric BaNiSn3-type structure (I4mm).
powder and mainly the single crystal diffraction. We note that the single crystal diffraction favored the BaAl4 structure in the case of related CeCuGa3 compound [3]. As the electronic properties of CeCuAl3 might be sensitive to the sample thermal treatment [11], we have studied the structure of both, as-cast and annealed samples and include additionally diffraction at elevated temperatures which finally led to some non-trivial results. 2. Experimental methods Single-crystalline CeCuAl3 sample was prepared by Czochralski growing method using the tri-arc furnace under protection of argon atmosphere. Polycrystalline sample of 8 g weight prepared by arcmelting from pure elements (2N8 for Ce, 4N for Cu and 5N for Al) was used as the precursor. Small polycrystalline sample of nonmagnetic analog LaCuAl3 was also prepared, the purity of La was 3N. Prepared polycrystalline samples were annealed for 8 days at 900 C in quartz tubes, the rate of heating the sample from room temperature was 3 C per minute, the rate of cooling down from 900 C was 1 C per minute. The single crystal was first investigated in its as-cast form and subsequently it has been annealed in the same way. The chemical composition of prepared samples was investigated by scanning electron microscope equipped with a backscattered electron (BSE) detector and with an energy dispersive X-ray (EDX) analyzer. The crystal structure investigation of poly- and singlecrystalline samples at room temperature was performed by powder X-ray diffraction using the Cu K-a radiation. High-temperature X-ray diffraction measurements were performed on powdered single crystal sample using the PANanalytical MPD diffractometer with MRI high-temperature chamber in conventional BB symmetric qe2q scan. Tantalum strip heater and platinum radiation heater were used as a heating elements, the heating rate was 5 C per minute. The sample was directly put to tantalum strip heater in order to have the good thermal contact. The pressure in heating chamber was of the order 102 Pa to protect the tantalum strip from corrosion and to ensure the same conditions for measurements at each temperature. The measurement was done in 2q range between 10 and 80 with the step of 0.025 at temperatures from 32 to 500 C. The diffraction peaks belonging to Ta were subtracted from diffraction data. The refinement of the structural parameters from the diffraction patterns was done using the Rietveld analysis employing the Fullprof program [16]. The quality and orientation of the single crystal was verified by X-ray diffraction using the Laue method. Laue neutron diffraction performed on Orient Express equipment in Institute Laue-Langevin (ILL), Grenoble, France, was used in the second step to check the quality in the whole volume of the prepared single crystal. Each measurement was taken in backscattering distance 68.0 mm from the middle of the sample for 20 min with 1.5 mm collimation.
Fig. 2. Neutron Laue patterns of the CeCuAl3 single crystal oriented along the [001] direction taken with Orient Express at ILL. The significant crystallographic directions [100] and [110] of tetragonal BaNiSn3-type structure are shown in picture.
The crystal structure of single crystal was determined by X-ray diffraction using the Bruker APEXII CCD diffractometer equipped with Mo X-ray tube. Structure factors have been extracted from single crystal diffraction pattern and the structure was identified using SHELXS-97 (Sheldrick, 2008). The X-ray diffraction patterns on CeCuAl3 single crystal of the prism shape with (0.080 0.100 0.140 mm) dimension were done at temperatures 296 and 150 K. Differential scanning calorimetry (DSC) was carried out on CeCuAl3 single crystal as well as on LaCuAl3 polycrystalline sample in order to obtain melting temperatures and to find any sign of presence of other transition as well as to verify the quality of the sample. Measurements were performed under protection of He atmosphere on SETSYS Evolution 24 instrument (SETARAM Instrumentation
Table 1 Structure parameters of CeCuAl3 single crystal determined from X-ray diffraction at room temperature and at 150 K. Vf.u. is the volume per fundamental unit. zCe, zCu, zAl and zAlb are fraction coordinates atoms in the fundamental unit (see subsection 3.1). R is the agreement factor for the fit of diffraction patterns (goodness of fit on F2). The thermal displacements of atoms are described by: 2p2 (h2a*2U11 þ . þ 2hka*b*U12), where matrix elements U12, U13 and U23 are equal to zero. SHELXS-97 program employing the Rietveld analysis was used for the determination of structural parameters. T (K)
150
296
a (pm) c (pm) c/a Vf.u. (106 pm3) Atomic positions: 2Ce in 2(a) 2Cu in 2(a) 2Al in 2(a) 4Al in 4(b) Thermal displacement: U11 (Ce) U22 (Ce) U33 (Ce) U11 (Cu) U22 (Cu) U33 (Cu) U11 (Al) U22 (Al) U33 (Al) U11 (Alb) U22 (Alb) U33 (Alb) R (%)
425.27(3) 1066.80(7) 2.5085(2) 96.47(2)
426.20(2) 1068.06(5) 2.5060(2) 97.01(2)
(0, (0, (0, (0,
(0, (0, (0, (0,
0, 0 (fix)) 0, 0.0946(3)) 0, 0.3689(1)) 1/2, 0.2508(2))
0.0032(2) 0.0032(2) 0.0041(2) 0.0068(4) 0.0068(4) 0.0093(7) 0.006(1) 0.006(1) 0.0061(5) 0.0056(8) 0.0056(8) 0.011(2) 2.36
0, 0 (fix)) 0, 0.0942(2)) 0, 0.36909(8)) 1/2, 0.2505(2))
0.00490(7) 0.00490(7) 0.0069(1) 0.0097(2) 0.0097(2) 0.0108(3) 0.0086(9) 0.0095(9) 0.0092(3) 0.0077(4) 0.0077(4) 0.0128(9) 1.12
128
M. Klicpera et al. / Intermetallics 46 (2014) 126e130
Fig. 3. The SEM image (upper left part) of CeCuAl3 single crystal surface. The showed area of surface contains some scratches because of polishing of the surface. Remaining three parts show the distribution of atoms obtained by the EDX analysis in the same surface area.
Company) and data analysis were done using SETARAM software. Heating and cooling scans were taken with the rate of 10 C per minute at temperature range from 28 to 1400 C. 3. Results 3.1. Single crystal characterization Laue X-ray and neutron diffraction performed on the grown CeCuAl3 single crystal verified the quality and tetragonal symmetry of the sample (see Fig. 2). However, there are still several structure types which can be adopted by the studied compound as mentioned in the Introduction. The centrosymmetric BaAl4 (I4/ mmm) and two its derivatives e disordered form of ThCr2Si2 (I4/ mmm) structure and the non-centrosymmetric BaNiSn3 (I4mm) e are displayed in Fig. 1. The only difference between them is a distribution of Cu and Al atoms in the lattice, while the Ce atomic positions are the same in all three structure types. In the BaAl4type, Cu and Al atoms are randomly distributed over the 2a and 4b sites. In the ThCr2Si2-type, the random distribution remains on the 2a-positions, whereas the 4b sites are occupied exclusively by Al. Both structures are clearly centrosymmetric. The noncentrosymmetric BaNiSn3 is completely ordered structure type with following atomic coordinates (see also Table 1): 1; 1; 1 2 2 2
þ zCe
2Cu in 2ðaÞ : ð0; 0; zCu Þ;
1; 1; 1 2 2 2
þ zCu
4Alin4ðbÞ: 0; 12;zAlb ; where zCe z 0 and
1; 1; 1 þ z Al 2 2 2
1;0;z Alb 2
zAlb z14
;
3.2. DSC and high-temperature X-ray diffraction Some properties of CeCuAl3 are slightly influenced by sample thermal treatment [11]. These changes can be ascribed to the annealing process which generally improves the quality of the sample, i.e. reduces the residual resistivity and also improves the thermal properties of the sample. To optimize the sample thermal treatment, the knowledge of the phase diagram and mainly the melting temperature is essential to choose a correct one. We have performed DSC measurement to reveal the melting temperature. Surprisingly, we have observed a further phase transition in course of these measurements.
2Ce in 2ðaÞ : ð0; 0; zCe Þ;
2Al in 2ðaÞ : ð0; 0; zAl Þ;
The powder X-ray diffraction patterns of CeCuAl3 powderized single crystal recorded at room temperature confirms that this compound crystallizes in the tetragonal structure. The best fit of the diffraction pattern was obtained for the BaNiSn3-type structure (RBragg ¼ 6.9% compared to RBragg ¼ 31.3% for the ThCr2Si2-type). Introducing certain degree of CueAl disorder to the BaNiSn3 structure did not lead to any improvement of the fit. We examined closely also the diffraction patterns of polycrystalline sample with similar result. These results, pointing to the BaNiSn3 structure, are well consistent with the neutron diffraction data [13]. However, we should have in mind somewhat limited sensitivity of powder X-ray diffraction to Cu and Al positions and CueAl disorder. The crystal structure was then uniquely determined by single crystal diffraction (295 independent reflections, Rint ¼ 0.0169) as tetragonal BaNiSn3-type structure (space group I4mm, 107) without any sign of disorder in Cu and Al atomic positions. We should note, the best structure solution were obtained when a racemic-twinning have been introduced. It seems that there exists two equally populated domains which are connected by a mirroring, where the mirror plane is perpendicular to the four-fold axis. The refined structure parameters of studied CeCuAl3 single crystal at room temperature and at 150 K are summarized in Table 1. The thermal displacements of atoms are given in Table 1 as well. The chemical composition investigation of our single and polycrystalline CeCuAl3 samples done using the EDX analysis confirms the stochiometric composition Ce:Cu:Al ¼ 1:1:3 (with 0.01 error of analysis). The distribution of elements was also checked by EDX equipment, homogeneous distribution of elements can be seen in Fig. 3.
1;1; 1 þz Alb 2 2
; 1; 12; 12 þzAlb ;
in the case of BaNiSn3-type structure.
Fig. 4. Differential scanning calorimetry performed on as-cast and annealed CeCuAl3 single crystal as well as on polycrystalline LaCuAl3. Heating/cooling rate was 10 C per minute. The inset shows data for annealed single crystal up to melting temperature.
M. Klicpera et al. / Intermetallics 46 (2014) 126e130
Fig. 5. The lattice parameters of the annealed single crystal CeCuAl3 obtained by hightemperature X-ray diffraction. The error for each lattice parameter is almost constant during the heating/cooling process, for this reason and for better lucidity we show error bars only for the point at 500 C. The agreement factor of the fit of our model to the experimental data was almost the same (RBragg z 6.9%) for all measurements. See text for more details.
The DSC temperature scans revealed the melting temperature of 1275 C (see inset of Fig. 4) what justifies the optimal annealing temperature around 900 C. Additionally, the measured scans, represented in Fig. 4, show a clear l-peak between 279 and 335 C which corresponds to the structural phase transition in the sample. The transition temperature of 320 C was obtained in usual way for this type of transition, i.e. as the temperature where the heating peak has minimum. This anomaly is observed also for cooling and is reproduced in all measurements independently on maximal temperature of the scan (500 Ce900 C), even not by achieving the melting temperature. The change of enthalpy associated with this transition is higher for heating (1.353 J g1) than for cooling (0.959 J g1) what might suggest a certain change of structure after the heatingecooling cycle. Further DSC measurement on polycrystalline sample showed the same behavior. The lanthanum analog LaCuAl3 exhibits signs of similar anomaly during the heating cycle, nevertheless much weaker as shown in Fig. 4. The enthalpy corresponding to this phase transition reaches only 0.153 J g1, nearly 10 times smaller value than for CeCuAl3. No anomaly is observed during the cooling process in LaCuAl3. The same behavior was reproduced by second subsequent cycle on the same LaCuAl3 sample. The high-temperature powder X-ray diffraction was performed on as-cast and annealed CeCuAl3 single crystal, as well as on annealed polycrystalline CeCuAl3 and LaCuAl3 samples to reveal the microscopic origin of the anomaly observed by DSC. Results of the refinement of diffraction patterns using the Fullprof program [16] are represented in Figs. 5 and 6. Linear increase of lattice parameters with increasing temperature is observed up to 250 C, the usual effect of thermal expansion takes place. The atomic fraction coordinates, i.e. zCu and zAl, are unaffected up to 250 C. The transition seen in DSC is then reflected by anomalous development of the c
129
Fig. 6. The atomic fraction coordinates of annealed single crystal CeCuAl3 and LaCuAl3 polycrystal obtained by high-temperature X-ray diffraction. The error of determination of atomic positions is z6$103. The lines are to guide the eye. See text for more details.
parameter (or c/a ratio) around 300 C. The lattice parameters increase again linearly above z350 C which suggests the stable high-temperature phase, in agreement with DSC analysis. The measurement during the cooling to room temperature shows very similar development of lattice parameters but the transition from high-temperature to low-temperature phase does not lead to the original values of lattice parameters, see Fig. 5. The lattice parameter a has higher value and parameter c has lower value compared to initial one before the heating/cooling cycle. The change can be
Fig. 7. The BaNiSn3-type structure adopted by CeCuAl3 compound below 300 C. The atomic fraction coordinates are changed above 300 C as denote the smeared Cu and Al atoms with arrows. For better clarity, the shift is 10 times increased compared to the real change. The illustrative marking of atoms is done for easier understanding the Table 2.
130
M. Klicpera et al. / Intermetallics 46 (2014) 126e130
Table 2 Interatomic distances in CeCuAl3 compounds obtained by employment of the Rietveld analysis at temperature below and above structural phase transition. The patterns taken on powder sample were processed with Fullprof program [16]. The atomic positions are labeled as shown in Fig. 7. T ( C)
CeeCu (pm)
CeeAl1 (pm)
CeeAl2 (pm)
CueAl1 (pm)
CueAl2 (pm)
Al1eAl2 (pm)
500 100
329.0(5) 329.7(5)
326.9(6) 322.4(7)
342.7(7) 341.2(7)
250.1(6) 250.8(6)
255.8(7) 249.3(7)
258.9(6) 261.2(6)
parameter a and from 10.672 to 10.720$102 pm for parameter c), without any trace of phase transition. The development of fraction coordinates zCu and zAl of LaCuAl3 is plotted in Fig. 6. Also here we are not able to see any clear hint of phase transition as seen in CeCuAl3. If any such transition exists in LaCuAl3, what might indicate the little anomaly in DCS measurements, it is below the experimental sensitivity of our X-ray diffraction analysis. 4. Conclusions
expressed as y0.05% of their initial values. Although the lattice parameters in CeCuAl3 are different after heating/cooling cycle, the volume of fundamental unit stays almost unaffected because the parameters a and c change in opposite way. The transition around z300 C is much better seen in development of atomic fraction coordinates zCu and zAl plotted in Fig. 6. The difference between these coordinates at temperatures below and above 300 C is z6$103 what represents almost 2% change of the value of these coordinates (see also illustrative picture of changed atomic coordinates in Fig. 7 and changes of interatomic distances given in Table 2). We are aware that the numerical error of determination of these values is also approximately 6$103. Furthermore, the uncertainty of determination of atomic positions from powder diffraction data is higher than only numerical error. In powder diffraction patterns, we face the problem of overlapping of the diffraction lines which is not present when using the single diffraction data. The overlapping can strongly affect, together with diffraction data quality, the accuracy of parameters of the structural model. The atomic positions determined using single crystal diffraction are much more accurate. The values obtained from powder diffraction are bounded with relatively high error and here serve mainly to follow qualitatively their temperature development. Despite relatively high error, the tendency, scatter of the values and mainly reproducibility of the effect on several samples is clear enough to conclude that the phase transition is present in CeCuAl3. The phase transition was observed for both, as-cast and annealed samples, as well as for polycrystalline sample, at the same temperature range, i.e. between 270 and 340 C. The changes in structure parameters corresponding to this transition are very similar for all samples. Especially similar changes in as-cast and annealed samples suggest that each heating/cooling cycle changes the structure parameters independently on temperature, heating/ cooling rate and duration of sample annealing. More detailed comparison of this effect on as-cast and annealed samples will be published separately [17]. The refinement of diffraction patterns of LaCuAl3 polycrystal shows ordinary development of lattice parameters, i.e. linear increase with increasing temperature (from 4.295 to 4.325 .102 pm for
Our study of CeCuAl3 single crystal unambiguously confirmed the ordered non-centrosymmetric BaNiSn3-type structure as the crystal structure of this compound. Moreover, we found the structural phase transition around 300 C. This transition is related to a small change of structural parameters while the crystal structure is preserved. Acknowledgments The work of M.K. was supported by Grant agency of Charles University under projects 76213 and SVV-2013-267303. We acknowledge ILL for the allocation of time and technical services. References [1] Hillier AD, Adroja DT, Manuel P, Anand VK, Taylor JW, McEwen KA, et al. Phys Rev B 2012;85:134405. [2] Paschen S, Felder E, Ott HR. Eur Phys J B 1998;2:169. [3] . Joshi Devang A, Burger P, Adelmann P, Ernst D, Wolf T, Sparta K, et al. Phys Rev B 2012;86:035144. [4] Haen P, Lejay P, Chevalier B, Lloret B, Etourneau J, Sera M. J Less-Common Met 1985;110:321. [5] Mock S, Pfleiderer C, Lohneysen H v. J Low Temp Phys 1999;115:1. [6] Sugitani I, Okuda Y, Shishido H, Yamada T, Thamhavel A, Yamamoto E, et al. J Phys Soc Jpn 2006;75:043703. [7] Kimura N, Ito K, Saitoh K, Umeda Y, Aoki H, Terashima T. Phys Rev Lett 2005;95:247004. [8] Adroja DT, del Moral A, de la Fuente C, Fraile A, Goremychkin EA, Taylor JW, et al. Phys Rev Lett 2012;108:216402. [9] Zarechnyuk OS, Kripyakevich PL, Gladyshevskii EL. Sov Phys Crystallogr 1965;9:706. [10] Mentink SAM, Bos NM, van Rossum FJ, Nieuwenhuys GJ, Mydosh JA, et al. J Appl Phys 1993;73:6625. [11] Kontani M, Ido H, Ando H, Nishioka T, Yamaguchi Y. J Phys Soc Jpn 1994;63:1652. [12] Bauer E, Pillmayr N, Gratz E, Hilscher G, Gignoux D, Schmitt D. Z Phys B 1987;67:205. [13] Moze O, Buschow KHJ. J Alloy Compd 1996;245:112. [14] Nishioka T, Kawamura Y, Kato H, Matsumura M, Kodama K, Sato NK. J Magn Magn Mater 2007;310:e12. [15] Hu XD, Zhou HY, Li JB, Pan SK, Wang T, Yao QR, et al. Intermetallics 2009;17:775. [16] Rodriguez-Carvajal J. Phys B 1993;192:55. [17] Klicpera M, Javorský P, Danis S, Brunátová T. Acta Physica Polonica. [submitted for publication].