TSFZ Growth of Eu-Substituted Large-Size LSCO Crystals

Abstract

The travelling solvent floating zone (TSFZ) growth of Eu-substituted LSCO (La_1.81−xEu_xSr_0.19CuO₄, with nominal x = 0 ÷ 0.4) single crystals was systematically explored for the first time. The substitution of La with Eu considerably decreased the decomposition temperature. Optimal growth parameters were found to be: oxygen pressure 9.0–9.5 bars; Eu-free CuO-poor solvent (66 mol% CuO) with a molar ratio of La₂O₃:SrCO₃:CuO = 4:4.5:16.5 and growth rate 0.6 mm/hour. The obtained single crystals were characterized with optical polarized microscopy, X-ray diffraction and energy-dispersive X-ray spectroscopy analysis. The solubility of Eu in LSCO appeared to be limited to x~0.36–0.38 under the used conditions. The substitution of La³⁺ with smaller Eu³⁺ ions led to a structural transition from tetragonal with space group I4/mmm for La_1.81Sr_0.19CuO₄ (x = 0) to orthorhombic with space group Fmmm for La_1.81−xSr_0.19Eu_xCuO₄ (x = 0.2, 0.3, 0.4), and to a substantial shrinking of the c-axis from 13.2446 Å (x = 0.0) to 13.1257 Å (x = 0.4). Such structural changes were accompanied by a dramatic decrease in the superconducting critical temperature, Tc, from 29.5 K for x = 0 to 13.8 K for 0.2. For x ≥ 0.3, no superconductivity was detected down to 4 K.

1. Introduction

Despite over three decades of intense research, high-Tc superconducting cuprates remain a focus of investigation. The nature of the so-called pseudogap phase existing at temperatures above Tc and its suppression at a critical doping p* close to where Tc reaches its maximal value is still being actively studied in Nd- and Eu-substituted La_2−ySr_yCuO₄ (LSCO) [1,2,3]. The substitution of La³⁺ with Sr²⁺ allows to systematically vary the doping, while the substitution of a fixed fraction of La³⁺ with smaller Nd³⁺ or Eu³⁺ lowers Tc and increases p*, increasing the range of observation of the pseudogap phase. In contrast, the evolution of the physical properties at fixed doping upon the systematic substitution of La³⁺ with isovalent smaller cations, akin to a chemical pressure, has received less attention. Growth conditions for Nd- or Eu-substituted LSCO in the literature are rarely given, mainly referring to those of rare-earth-free LSCO crystals [4,5,6]. An exception is a report on the crystal growth of La_1.6−yNd_0.4Sr_yCuO₄ in air using a 78 mol% CuO, which claims that growth speeds higher than 1 mm/h dramatically affect the crystal quality, 0.65 mm/h yielding optimal results [7]. Recently, we showed that the growth of La_1.81−xNd_xSr_0.19CuO₄ with the TSFZ technique using a copper-rich solvent (78–80 CuO mol%), as routinely employed for x = 0 but with the appropriate ratio of Nd, La and Sr, for x ≥ 0.2 led to a considerable amount of bubbles in the molten zone (produced by the reduction in CuO with the release of oxygen gas) and introduced large voids in the grown ingots, which presented a poor crystallinity [8]. We could considerably suppress the bubbling and improve the quality of the Nd-substituted crystals by reducing the copper content of the solvent to 65–66 mol%, increasing the oxygen pressure to 5.5–6.5 bar and decreasing the growth rate to 0.6 mm/h.

The Eu³⁺ ion possesses a smaller ionic radius (0.947 Å) than Nd³⁺ (0.983 Å), potentially allowing for a larger chemical pressure effect at similar substitution levels. Changing the nature of the substituting ion, however, can influence not only the properties of the solid, but also the liquidus temperature, the solubility, the viscosity and the reactivity of the melt, requiring a further optimization of the growth parameters. We show that the crystals grown using the optimal growth parameters found for Nd-substituted LSCO, although intact on the border, present a considerable amount of decomposition at the core. This effect becomes more marked upon increasing the Eu content, but could be suppressed by using an Eu-free solvent and increasing the oxygen pressure to 9–9.5 bar. Eu-substituted crystals present an orthorhombic (Cmce) structure instead of a tetragonal (I4/mmm) one of LSCO, and the volume per formula unit decreases with x, accompanied by a rapid decrease in Tc.

2. Materials and Methods

Powders of La₂O₃ (ChemPUR Feinchemikalien und Forschungsbedarf GmbH, Karlsruhe, Germany), Eu₂O₃, CuO and SrCO₃ (Alfa Aesar by Thermo Fisher Scientific GmbH, Kandel, Germany) with a purity of no less than 4N were used as starting materials in the preparation of feed-rods and solvents. The procedure for raw material synthesis as well as for the production of feed rods was similar to that described in [8], using the same precursor ratios with Eu₂O₃ instead of Nd₂O₃. In particular, a Cu-rich solvent (78–80 mol% CuO) was prepared from a molar ratio of La₂O₃:Eu₂O₃:SrCO₃:CuO of 3:1:3.3:26, while for the Cu-poor solvent (65–66 mol% CuO), we used a 3.5:1:4:16.5 ratio. A small amount (3 wt%) of boron oxide was added in order to suppress the penetration of the solvent into the feed rod and the “swollen effect”, as well as to increase the melt viscosity [9].

We observed that the substitution of La with Eu in LSCO led to a decrease in its decomposition temperature, to a more dramatic extent than the substitution with Nd. Therefore, the sintering temperature for Eu-LSCO was reduced to 1100 °C, compared to 1200 °C used for Nd-LSCO [8]. Unfortunately, decreasing the sintering temperature also led to a substantial decrease in the feed-rod density, which tended to favor the penetration effect and was detrimental for the stability of the TSFZ growth.

Crystals were grown on a four-mirror-type image floating zone furnace (CSC, Hokuto, Yamanashi, Japan) equipped with 300 W halogen lamps as the radiation source. The following conditions were used during the growth experiments: (i) feed and seed shafts rotated oppositely at 20 rpm; (ii) growth rates of 0.6–1.0 mm/h; (iii) oxygen atmosphere with pressure up to 9.5 bars.

Powder X-ray diffraction (XRD) measurements on crushed crystals were carried out at room temperature using an automated powder STOE Stadi P Dual Source diffractometer with Cu K_α1 radiation and a MYTHEN 2K detector. Data were collected in the 2θ range from 5° to 100.25° in 0.015° steps. The Rietveld analysis [10] was performed with a FullProf program package [11]. The space groups I4/mmm and Cmce, as well as the atomic positions published for La_1.85Sr_0.15CuO₄ [12] and La_1.94Sr_0.06CuO₄ [13], were used in the refinement.

Photographs for cross-sections of the crystals were obtained on a Zeiss Axiovert 25 Invert microscope in polarized light. The chemical composition was analyzed by means of energy-dispersive X-ray spectroscopy in a Zeiss EVOMA15 scanning electron microscope. The operating voltage was 30 kV. The La and Eu contents were determined using M- and L-lines, and for Cu and Sr contents, L- and K-lines were used.

Magnetization measurements were performed in a SQUID-VSM (Quantum design Germany, Darmstadt, Germany). A field of 50 Oe was applied in arbitrary directions for all the crystals.

3. Results and Discussion

3.1. Crystal Growth

In analogy to our previous findings for Nd-LSCO [8], growth experiments on Eu-LSCO crystals for x = 0.2 using a Cu-rich solvent and oxygen pressure in the range of 1–2 bar, as routinely used for LSCO, yielded poor results. As mentioned above, the relatively low sintering temperature resulted in a lower feed-rod density for Eu-substituted LSCO (compared to that of LSCO or even Nd-substituted LSCO). This in turn caused intense liquid penetration from the molten zone (MZ) into the feed-rod (“swallowing”), which led to a complete interruption of the growth.

As we had previously shown, using a CuO-poor solvent with 65–66 mol% CuO in the growth of Nd-LSCO efficiently suppressed the penetration of the solvent from the MZ into the feed rod [8]; therefore, we used an analogous solvent with a La₂O₃:Eu₂O₃:SrCO₃:CuO ratio of 3.5:1:4:16.5 to counteract the influence of the low density of feed rods on the growth stability of Eu-LSCO. This resulted in a very stable growth run for x = 0.2 at a pulling rate of 0.6 mm/h and 5.5–6.5 bar oxygen pressure, see Figure 1. The crystallization front exhibited a sinusoidal shape and facets could be observed in the grown crystal.

In spite of a mirror-like surface with a metallic luster of the ingot, an image of its cross section in Figure 2 revealed a large number of voids and impurity phases in the central part of the crystal. Long straight streaks were observed, which the EDX analysis showed to consist of a mixture of La₂O₃ and Eu₂O₃, without Cu. Enclosed between such streaks were areas of composition consistent with the nominal one, as well as Cu-containing areas enriched in La or Eu. The composition in the border regions of the ingot was consistent with La_1.61Sr_0.19Eu_0.2CuO₄. This contrast between the border and the core of the ingot suggested that although the phase crystallized from the melt, decomposition occurred upon cooling after solidification.

In order to suppress this decomposition, we increased the growth rate from 0.6 mm/h to 0.8 mm/h and 1.0 mm/h. We expected that a faster growth rate would favor a rapid cooling of the crystal, avoiding decomposition. Figure 3 shows cross-sections of the grown ingot for different growth rates. Although a slight improvement was observed, even at 1 mm/h, such long streaks of secondary phases were still present.

In order to further favor a rapid cooling of the crystals, we applied a high oxygen pressure (9–9.5 bar instead of 5.5–6 bar, used for the growth of Nd-LSCO crystals [8]). As shown in Figure 4, such an increase in the oxygen pressure significantly reduced the crystal decomposition, but unfortunately did not avoid it completely.

Since in the Eu-free LSCO such a inhomogeneity at the core of the ingots was not observed, it could be related to an excessive concentration of Eu₂O₃. As a next step, we removed all Eu₂O₃ from the solvent, using a “copper-poor” solvent with a molar ratio of La₂O₃:SrCO₃:CuO = 4:4.5:16.5. This led to a drastic improvement in the quality of the grown Eu-LSCO crystals (see Figure 5). In addition to possessing a surface with a metallic luster, cross-section cuts of x = 0.3 ingots as that in Figure 5a showed a homogeneous core, with only tiny occasional areas enriched in Sr or Eu present at grain boundaries only. The decomposition of the crystal after solidification was completely suppressed. The composition of the main phase corresponded to the nominal one: La_1.51Sr_0.19Eu_0.3CuO₄. This indicated that after an initial stage, when the MZ incorporated Eu from the feed, the MZ reached Eu-saturation and the composition stabilized, crystallizing the desired phase.

Using the determined optimal growth parameters, (i) oxygen pressure 9–9.5 bar and (ii) a “CuO-poor” Eu-free solvent with a molar ratio of La₂O₃:SrCO₃:CuO = 4:4.5:16.5, single crystals for Eu-substituted LSCO were obtained for different values of x = 0.2 ÷ 0.4. As shown in Figure 6, the obtained crystals (length up to 70 mm and 5–5.5 mm in diameter) exhibited a metal luster on the surface, while the single-crystal part of the ingots was up to 30 mm in length and up to 4.5 mm in diameter. The polarized microscopy of cross-sections of the end portions of the grown ingots for x = 0.2 and 0.3 showed a large main singe-crystal grain with some secondary grains close to the ingot border. The analysis with EDX confirmed the composition of the crystals to be La_1.61Sr_0.19Eu_0.2CuO₄ and La_1.51Sr_0.19Eu_0.3CuO₄, respectively.

For x = 0.4, however, decomposition in the central part of the crystal was still observed. This could be an indication that the solubility limit for Eu in LSCO lies in between x = 0.3 and x = 0.4. The EDX analysis performed on the surface in Figure 6f showed that the long whitish lines in the central part of the crystal consisted of lanthanum and europium oxides (La₂O₃ and Eu₂O₃), similarly to those observed in Figure 2. The EDX analysis on pieces taken from the border of the ingot proved that the actual Eu content of the crystal lies in the range x = 0.36 ÷ 0.38, which probably represents the Eu solubility limit in LSCO.

3.2. Structural Characterization

Powder XRD patterns of crushed crystals exhibited the presence of the expected La_1.81−xEu_xSr_0.19CuO₄ phase, without any detectable traces of impurities.

The Rietveld analysis of the powder pattern for x = 0 confirmed that this compound crystallized in the K₂NiF₄-type structure (space group I4/mmm), often denoted as HTT (high-temperature tetragonal), in agreement with previous reports [12]. Two changes were observed upon increasing the Eu content: an overall shift of the peaks to higher diffraction angles, and a splitting of certain diffraction peaks (see Figure 7).

Peak splitting is indicative of a symmetry reduction. All reflections of the Eu-substituted LSCO with nominal x = 0.4 were successfully indexed in the orthorhombic system and appeared to be similar to several structural modifications of the LSCO compound with orthorhombic space groups Cmce [13] and Fmmm [14]. The Rietveld refinement for all the Eu-substituted LSCO compounds was performed for both space groups, and showed that the space group Fmmm was slightly favorable compared to Cmce, resulting in lower R_Bragg and especially R_F residuals, and significantly reduced standard deviations in atomic coordinates of the O1 atoms. In our XRD experiment, we did not detect additional reflections that corresponded to the C-centering of the crystal lattice, due to their very low intensity. Moreover, a direct transition from the I4/mmm space group to Cmce was impossible and required an intermediate transfer through the space group Fmmm. This transition from I4/mmm to Fmmm was performed with the transformation of the a and b lattice vectors, resulting in the doubled volume of the unit cell (for more details, see Figure 8).

The volume per formula unit decreased with the increase in the Eu content, as expected, in view of the smaller ionic radius of Eu³⁺ with respect to that of La³⁺. However, a linear dependence was observed only in the range 0.2 ≤ x ≤ 0.4. The distance between the Cu–O planes changed linearly (decreasing by 0.9% between x = 0 and x = 0.4), while the average in-plane Cu–Cu distance remained basically unchanged (see

Table 1. Summary of structural and superconducting information for Eu-LSCO.

Compound	La1.81Sr0.19CuO4	La1.61Eu0.2Sr0.19CuO4	La1.51Eu0.3Sr0.19CuO4	La1.41Eu0.4Sr0.19CuO4
Refinement Data
Space group, prototype	I4/mmm (#139), K2NiF4-type	Fmmm (#69), La2CuO4-type
a, (b) c (Å)	3.7729(7), 13.245(1)	5.3256(4) 5.3339(0) 13.1820(1)	5.3244(3) 5.3393(8) 13.1483(3)	5.3197(3) 5.3435(7) 13.1257(9)
V (Å3)	188.548(1)	374.454(4)	373.796(3)	373.118(3)
Reflections measured	45	68	68	68
RF = Σ|Fo-Fc|/ΣFo	0.0296	0.0735	0.0557	0.0486
RI = Σ|Io-Ic|/ΣIo	0.0453	0.0649	0.0627	0.0419
χ2 = (RwP/Re)2	1.86	1.77	1.89	2.18
Interatomic distances
Cu-O1	0.2444(8)	0.2419(12)	0.2398(12)	0.2418(11)
Cu-O2	0.18864(1)	0.18844(1)	0.18851(1)	0.18850(1)
La/Sr/Eu-O1	0.27345(18)	0.2725(3)	0.2721(3)	0.2722(3)
La/Sr/Eu-O1	0.27345(18)	0.2729(3)	0.2729(3)	0.2734(3)
La/Sr/Eu-O1	2.336(9)	2.332(12)	2.342(12)	2.306(11)
La/Sr/Eu-O2	2.6374(7)	2.6338(11)	2.6302(10)	2.6334(9)
Superconducting temperature
Tc, K	29.5	13.8	None (or lower than 4 K)	None (or lower than 4 K)

). The substitution of La with Eu increased the mismatch between the rocksalt and the perovskite layers, favoring the orthorhombic structural distortion.

An analysis of the interatomic distances in the structures of Eu-substituted LSCO revealed, that among six oxygen atoms that formed an (O12O24) octahedron around each Cu atom, four O2 atoms (which formed a square in the pure LSCO in the ab plane) practically did not change their distances to Cu (d~0.1885 nm), despite a tiny deformation in this square for x ≥ 0.2 due to the symmetry change. Two other O1 atoms somewhat decreased their distance to the Cu atom along the c-direction from 0.2444(8) nm in the pure LSCO to 0.2418(11) nm for x = 0.4. The distances from O2 to the La/Sr/Eu site also remained practically unchanged, while all contact distances between La/Sr/Eu and the O1 atoms decreased. This allowed us to conclude that the Cu–O2 layers were characterized by the strongest bonding, while the La/Sr/Eu-O1 framework was responsible for the lattice deformation while substituting.

These structural changes influenced the superconducting properties of Eu-substituted LSCO crystals. As it can be seen from Table 1, the superconducting Tc was detected for La_1.61Sr_0.19Eu_0.2CuO₄ only. For crystals with x = 0.3 and 0.4, no signature of superconductivity was detected down to 4 K. It is worth noting that the decrease in Tc between x = 0 and x = 0.2 was of roughly 50%, while for the Nd-substitution, the same decrease in Tc was obtained for x = 0.4.

4. Conclusions

The growth of Eu-substituted LSCO (La_1.81−xEu_xSr_0.19CuO₄, x = 0 ÷ 0.4) large-size crystals with the TSFZ method was investigated. Optimal growth conditions were determined: a Eu-free “CuO-poor” solvent with a molar ratio of La₂O₃:SrCO₃:CuO = 4:4.5:16.5, a pulling rate of 0.6 mm/h and 9.0–9.5 bar oxygen pressure. XRD and EDX analyses confirmed the phase purity of the crystals. The solubility limit of Eu in LSCO was found to be x~0.36–0.38. The substitution of the La³⁺ ion with smaller Eu³⁺ induced a structural transition from tetragonal (space group I4/mmm) for Eu-free LSCO to orthorhombic (space group Fmmm). The main effect of the substitution was to reduce the interplanar distance between the Cu–O planes, while the in-plane distances were only weakly affected. The superconducting Tc strongly decreased upon Eu substitution, with no transition detected down to 4 K for x = 0.3 and 0.4. The influence of the Eu substitution was much stronger than that of the Nd substitution, owing to the much smaller ionic size of Eu³⁺.