Raman and weak ferromagnetism in Eu2CuO4

Abstract We show that there is a subtle instability of the T′ structure for the R2CuO4 (R = rare earth) compounds at the center of the R series with the boundary at R = Eu. Crystals grown in Pt crucibles and PbO flux show weak ferromagnetism (WF) and two temperature-dependent forbidden Raman peaks. Crystals grown in alumina crucibles and CuO flux do not show WF and the forbidden Raman peaks are much weaker. The WF and forbidden Raman peaks in Eu2CuO4 suggests that the instability of the T′ structure is associated with O(1) displacement in the CuO2 planes.


I. Introduction
The R2-~M~CuO4 compounds of T'-type structure (R = rate-earths and M = Ce, Th) have been intensively studied since their discovery. For R = Pr, Nd, Sm, Eu and x ~ 0.15, n-type superconductivity is achieved after appropriate thermal treatments in reducing atmospheres [1], but compounds with R = Gd to Tm are not superconductors for any doping concentration [2]. In R2CuO4, the Cu moments order antiferromagnetically (AF) below TN ~ 240-280 K [3]. For heavier R these compounds show weak ferromagnetism (WF), with a boundary at R = Eu 4. For R = Tb, Dy, Ho, Er, Tm and Y the T' structure can be synthesized only under high pressure, with again a boundary for structural stability at the center of the R series [5]. The WF is due to a canting in the ab-plane of the Cu moments away from perfect AF alignment. Although WF is forbidden in the T' (14/mmm) structure, lattice distortions in the CuO2 planes may allow WF. Lattice distortions in Gd2CuO4 and TmzCuO4 were invoked to explain X-ray and * Corresponding author.
Mrssbauer data [6]. It was argued that a lateral displacement of the oxygen ions O(1) away from their symmetric positions in the CuO2 planes gives rise to an antisymmetric Dzyaloshinsky-Moriya exchange interaction between the Cu moments [7,8]. This distortion may also be responsible for the extra Raman lines seen in Nd2-xGdxCuO4 [9] and Gd2CuO 4 [10,11].
Here we give results of Raman and magnetization measurements in single crystals of Eu2CuO4, grown in Pt crucibles from PbO flux (hereafter, Pt/PbO) and in alumina crucibles from CuO flux (hereafter, A1203/CuO).

Experimental details
The crystals were of plate-like shape, with the c-axis perpendicular to the large face. They were grown from stoichiometric mixtures of the oxides, using PbO and CuO fluxes in Pt and alumina crucibles, respectively. The Pb content was less than 1% of the Cu content [12]. In Raman experiments, we used the cold finger of a closedcycle Displex He refrigerator, the 514.5 nm line of an 0921-4526/96/$15.00 ~) 1996 Elsevier Science B.V. All rights reserved PII S092 1-4526(96)00 1 62-7 argon laser, and a Jobin Yvon T6400 triple spectrometer with a CCD camera. A backscattering geometry was used throughout. Magnetic measurements were made with a Quantum Design DC SQUID magnetometer.

Results and analysis
A group-theoretical analysis predicts four Raman active modes in the tetragonal T' structure: A~g + Bag + 2E,. Denoting by z the direction parallel to the crystal c-axis, the modes appear at the configurations: Y(ZZ)Y (A~g), Y(ZX)Y (Eg) and Z(XX)Z (B~g). In the Raman spectrum of the Pt/PbO sample, we identified Raman active modes at 229 cm-~ (Atg), 499 cm-~ (Eg) and 324cm -1 (Big) (Fig. 1). As in Nd2CuO4 [13] and Pr2CuO4 [14], the low-frequency Eg mode was not seen. The peaks at 413cm -1 for XX polarization and 398 cm-1 for XY polarization, which we label B*g and B'g, respectively [10,11], do not correspond to any mode allowed in the T' structure. We tentatively attribute them to local modes associated with oxygen displacements in the CuO2 planes [7,8]. The intensity of the B*g peak, relative to the Big mode, is greater for samples grown in Pt/PbO than for those grown in A1203/CuO. The B*g peak was not seen in A1203/CuO samples (inset of Fig. 1). The most striking result is the temperature dependence of the intensity of the anomalous peaks (Fig. 2). The intensity of the Big mode was independent of temperature, so we used it to normalize the intensity of the B*g peak. In another set of measurements we observed the Big, B~'g, and B~g peaks simultaneously, by rotating the incoming polarization about 22 ' away from the X-axis without using analyzer. The intensities of the B*g and B]'g had almost the same temperature dependence.
Magnetization in the ab-plane was measured at 100 K, after field cooling (FC) and zero-field cooling (ZFC) (Fig. 3). The sample grown in Pt/PbO showed hysteresis  and WF after FC (Fig. 3(a)). The remnant magnetization Mr and the coercive field Hc depended on temperature and cooling field. We obtained the saturation values M~ ~ 22(5) emu/FU and H~ ~ 50-70 Oe, at T = 20 K after FC in 50 kOe. However, ZFC magnetization was reversible and (at T ~-100 K) approached the FC magnetization above ~ 10kOe ( Fig. 3(a)). For our applied fields, the anisotropy within the abplane was negligible and no hysteresis or WF were detected perpendicular to the ab-plane. We found no hysteresis or WF in samples grown in Al2Oa/fuO (see Fig. 3(b)).
the WF and stronger anomalous Raman peaks found in samples grown in Pt/PbO. Owing to their large ionic radius, Pb atoms would probably substitute Eu atoms, and hence not contribute directly to the O(1) displacements.
Our Raman and WF results, along with other work [15], show that a subtle instability in the T' structure occurs at a value of about ao ~ 3.905(5) A for the abplane lattice parameter. Compounds with a < a0 show WF, whereas those with a > ao may become superconductors when properly doped with Ce [15].
Since the anomalous Raman B*g peak (Fig. 2) is still seen above the Nrel temperature (TN = 241 K, determined by the appearance of WF), the distortions responsible for the B** and B~g peaks may be associated with short-range magnetic ordering. Spin-dependent phonon Raman scattering [16] may be responsible for the greater intensity of the B*g and B*g peaks at lower temperatures. However, lattice contractions, resulting in a larger number of these distortions, also could increase the intensity of these Raman peaks at low temperatures.

Conclusions
Our results suggest that the anomalous Raman peaks in Eu2CuO4 are related to the WF. That the anomalous peaks appear at about the same energy for all the R2CuO, compounds showing WF, regardless of crucible and flux used, supports the assumption that these peaks are associated with local vibrations due to O(1) displacement. However, lattice dynamics calculations are needed to see if the anomalous Raman peaks can actually be associated with O(1) displacement in the CuO2 planes.