Subsolidus phase equilibria of the CuO–SrO–ZnO pseudoternary system in air at 900 °C

The subsolidus phase equilibria of the CuO – SrO – ZnO system were determined at 900 °C in air. The pseudoternary section does not contain ternary oxide phases but is made of 5 three-phase regions and 2 narrow two-phase regions linked to a Sr 14 Cu 24– x Zn x O 41– y solid solution. The maximum solubility of Zn in this phase is limited to x ≈ 0.1, but this low doping level results in a significant decrease of the electrical resistivity by about one order of magnitude compared to the undoped compound. The other binary oxide phases SrZnO 2 , Sr 2 CuO 3 , and SrCuO 2 do not form solid solutions extending into the ternary system. SrZnO 2 was found to decompose upon contact with ambient air.


Introduction
The discovery of high-temperature superconductivity in the Ba-doped La 2 CuO 4 insulator [1] has initiated a wealth of research aiming at finding new compounds with ever higher superconducting critical temperatures, in particular among cuprate compounds.One of the strategies used for this purpose consists in studying the phase equilibria of pseudoternary and higher dimensional systems con taining CuO and an alkaline earth oxide, in particular BaO, SrO or CaO.Nevertheless, whereas ternary sections involving these oxides with rare-earth oxides have been systematically determined (refe rence [2] and references therein), combinations of CuO and (BaO, SrO or CaO) with a transition metal oxide are still poorly investi-gated.In the case of SrO, the exceptions encompass systems in which the transition element is V [3], Nb [4,5], Fe [6], Ti [7,8], Zr [8], Hf [8], Ta [9], W [10] and Co [11].The CuO -SrO -ZnO system does not appear to have been stu died.However, Zn being not magnetic, it represents a favo rable case for high-temperature superconductors.Furthermore, the use of ZnO in association with SrO or CuO for various applications [12][13][14][15][16][17][18][19] as well as that of SrZnO 2 as a catalyst for PET depolymerization under microwave irradiation [20] or as host material for red or blue emitting phosphors [21,22] make the study of this system interesting in view of optimizing the processing parameters of samples aimed at various technological applications.The equilibration conditions for the present study (air atmosphere and 900 °C) correspond to conditions similar to those used for the synthesis of various high temperature superconductors [23,24].Surprisingly, searches in databases such as the ACerS-NIST Phase Equilibria Diagrams database (version 4.0) or Web of Science did not deliver references to CuO -ZnO or SrO -ZnO pseudobinary systems, not even as side-systems in ternary or higher dimensional systems so that even these simple combinations call for investigations.

Experimental
CuO (Alfa Aesar, 99.995% purity), SrCO 3 (Sigma Aldrich, 99.95) and ZnO (Alfa Aesar, 99.9%) powders were used as starting reagents.33 different nominal compositions were selected and studied.After the powders were weighted in proper amounts, they were thoroughly mixed in an agate mortar and calcined at 900 °C for 60 h.After this first treatment and further grinding, 1 g pellets with 12 mm diameter were pressed under a pressure of 1.8 kbar and sintered at least twice at 900 °C for 100 h with intermediate grinding and re-pressing.The pellets were quenched in air at the end of each heat treatment.All manipulations were performed in air.
The phase composition of the samples was determined using X-ray diffraction (XRD) patterns recorded in a Bruker D8 Robot diffractometer using Cu Kα radiation (λ = 1.54 Å).Equilibrium was considered as achieved when no differences were detectable in the XRD patterns after two consecutive heat treatments.Si powder was added as an internal standard after the heat treatments for lattice parameter calculations.
Low temperature electrical resistivity measurements were conducted in a Displex setup with a DC current of 1 mA by the 4-contact method on samples cut from the pellets into parallelepipeds with 1.5 × 1.5 mm 2 cross section and 8 mm length.Gold wires were attached with silver paste cured at 200 °C.

Results and discussion
As expected, the "ZnO" sample annealed at 900 °C consists of the ZnO (zincite) phase.Its hexagonal lattice parameters were calculated as a = 3.250(1) Å and c = 5.206(1) Å, in close agreement with the values reported by McMurdie et al. [25].Including small amounts (1 at.%) of either Sr or Cu in the nominal composition resulted in two-phase mixtures, as shown in Fig. 1a, with the appearance of SrZnO 2 or CuO peaks, respectively.The lattice parameters of these samples are compared with those of the pure ZnO composition in Table 1.For the Zn 0.99 Sr 0.01 O 2 nominal composition, the lattice parameters are unchanged within the standard deviation, whereas for Zn 0.99 Cu 0.01 O 2 , a slight decrease of the c-axis length seems to take place.The possibility of introducing Sr into the zincite structure has been explored within various contexts such as modification of the optical and electrical properties of ZnO [12,13], improving its photocatalytic properties [14,15], or using SrO as a sintering aid [15,16].
Conclusions on the actual doping of Sr into ZnO are contradictory, probably due to the widely different preparation routes used for these various applications.Güntürkun and Toplan [16] (solid state reaction at 1000 °C to 1300 °C) found a two-phase equilibrium between ZnO and SrZnO 2 even for 1 wt.%SrO in ZnO.Harish et al. [14] (hydrothermal synthesis) reached a similar conclusion.In contrast, Vijaran et al. [13] (thin films made by chemical bath deposition) as well as Srivastava et al. [12] (thin films made by solution spin coating and annealing at 400 °C) claim that Sr enters the ZnO structure.However, the lattice constants of ZnO show a nonmonotonic behavior versus doping level, with the c-axis parameter first decreasing before increasing.On the other hand, Das et al. [15] (solid-state reaction at 1100 °C) observed a broadening of some ZnO XRD peaks at low Sr doping level and concluded that two distinct ZnO phases with different lattice constants coexist.Besides the fact that processing condicitons may induce different metastable situations, it is worth mentioning that the relatively low signal-to-noise ratio in the XRD patterns of the thin film samples [12,13] makes the detection of low intensity impurity peaks difficult, while some high intensity XRD reflections of SrZnO 2 have a near overlap with several ZnO reflections.The latter fact may partly be responsible for the reported behavior of the ZnO c-axis parameter upon Sr doping [12,13,15].In any case, there is a clear need for more systematic investigations in order to assess the actual solubility of Sr in ZnO and its dependence on processing parameters.The present study allows concluding that for equilibration at 900 °C in air, the solubility of Sr is clearly lower than 1 at.%.
Concerning Cu substitution in ZnO, some recent publications, in which 5 at.%Cu doped samples systematically show a clear CuO XRD impurity peaks [17,18,26], claim that a non-negligible amount of Cu actually enters the zincite structure.Again, the very different processing parameters, including hydrothermal systhesis [17], xerogel route [18] or sol-gel [26], all with relatively brief heat treatments at 600 °C or lower temperatures, certainly have an influence on the conclusions.Moreover, no precise studies based on systematic lattice parameters variation were presented.The results of the present equilibration study at 900 °C are in fact more in line with other studies, including TEM characterization showing that CuO nanoparticles segregate on the surface of ZnO crystallites even for 1 at.%Cu doping [19].The slight decrease of the ZnO c-axis length could be due to very low level Cu doping (ppm range) as observed in ZnO samples exhibiting green luminescence [27].
The presence of CuO in the XRD pattern of the sample with Zn 0.99 Cu 0.01 O 2 nominal stoichiometry indicates that there are no binary oxide compounds between ZnO and CuO.This is further confirmed by the fact that ZnO is detected in the XRD pattern of a sample with Zn 0.01 Cu 0.99 O 2 overall composition (Fig. 1b).In spite of this two-phase equilibrium, it is clear, as shown in Table 1, that the presence of Zn has resulted in changes in the lattice parameters of CuO, with, in particular, a significant decrease of the b-axis length and increase of the a-axis parameter (as well as an increase of the β-angle from 99.5° to 99.7°).
The unambiguous detection of low-intensity ZnO reflections in the XRD pattern of the Cu 0.99 Zn 0.01 O 2 sample, however, indicates that at 900 °C the solubility limit of Zn into CuO is lower than 1 at.%.Similar lattice parameter variations upon Zn doping in CuO had beed reported previously [28].
Prabhakaran and Boothroyd [29] report up to 5 at.% Zn doping, but the pro cessing conditions in that case (single crystals grown by the floating-zone method under 4 to 8 atm pressure) mostly provide a hint that the solubility limit of Zn in CuO might in-crease with temperature and may also show some dependence on the oxygen pressure.
In the SrO -ZnO binary oxide side system, SrZnO 2 was the only binary oxide phase detected in the XRD patterns.The lattice parameters of samples with Sr 0.98 Zn 1.02 O 2 , Sr 1.00 Zn 1.00 O 2 and Sr 1.02 Zn 0.98 O 2 (Table 1) show little variation that is well inside the confidence interval.This observation, coupled with evidence for two-phase equilibria in the two samples with nominal composition deviating from SrZnO 2 (Fig. 2), indicates that the latter phase has a very narrow solid-solution range, if any.It is worth mentioning that the formation on SrO in the Sr-rich This assumption is supported by the fact that a sample with a composition closer to pure SrO (Sr 0.99 Zn 0.01 O -not shown), still contained some identifiable SrO reflections.The Sr(OH) 2 •H 2 O compound appears to form quickly upon exposure to the atmosphere.The XRD pattern of the Sr 1.02 Zn 0.98 O 2 sample (Fig. 2b) was recorded within less than 1 hour after quenching from 900 °C.In view of potential applications of SrZnO 2 , it is also important to note that this compound degrades at a moderate rate when exposed to ambient air.After two weeks, formation of SrCO 3 was ascertained in the originally single-phase SrZnO 2 sample.
In the CuO -SrO pseudobinary side system, the three binary oxide phases Sr 2 CuO 3 , SrCuO 2 and Sr 14 Cu 24 O 41-y were found as expected [31][32][33][34][35][36].Replacing 5 at.% of Cu by Zn in the respective compositions of these compounds resulted in samples with equilibrium between three phases.The lattice parameters of Sr 2 CuO 3 and SrCuO 2 (Table 1) were not affected by the presence of Zn within the confidence interval.In contrast, while the a-and b-axis lengths of Sr 14 Cu 24 O 41-y were not influenced, it appears that the c-axis lattice parameter became significantly shorter at a Zn substitution level for Cu of 0.4 at.% only.The XRD patterns of samples with Sr 14 Cu 24-x Zn x O 41-y (x = 0.00; 0.10; 0.50 and 1.00) nominal compositions are shown in Fig. 3, where it is observed that ZnO and SrCuO 2 reflections appear beyond x = 0.10 in agreement with the saturation of the c-axis parameter value (Table 1 and inset in Fig. 4).
The effect of Zn substitution on the Sr 14 Cu 24 O 41-y phase was studied further by means of electrical conductivity measurements.As demonstrated by Fig. 4, the room-temperature resistivity of Sr 14 Cu 23.9 Zn 0.1 O 41-y is lower than that of Sr 14 Cu 24 O 41-y , the difference increasing at lower temperatures to reach one order of magnitude at 175 K.In the Sr 14 Cu 23.75 Zn 0.25 O 41-y sample, the resistivity is even lower.Although the lattice parameters are not changed as compared with the Sr 14 Cu 23.9 Zn 0.1 O 41-y composition, it is likely that this further decrease in resistivity is related to a little more Zn doping into the Sr 14 Cu 24 O 41-y lattice (albeit not all the Zn has entered the structure), instead of being due to minute amounts of impu-  rity phases.This hypothesis is based on the fact that the samples with higher nominal Zn contents, i.e.Sr 14 Cu 23.5 Zn 0.5 O 41-y and Sr 14 Cu 23 Zn 1 O 41-y , exhibit higher resistivities (not shown in Fig. 4).Nominal compositions situated further within the pseudoternary section resulted in two-phase or three-phase equilibria, without any evidence for the formation of ternary oxide phases.Based on all these observations, the phase equilibria and tieline compatibilities of phases in the CuO -SrO -ZnO system at 900 °C in air can be summarized as shown in Fig. 5.The section contains 5 three-phase regions and 2 narrow two-phase fields associated to the Sr 14 Cu 24-x Zn x O 41-y solid solution.This system is dominated by ZnO, which is in equilibrium with all other binary oxide phases.

Conclusions
The subsolidus relationship and phase formation of compounds in the CuO -SrO -ZnO system were determined at 900 °C in air.The pseudoternary phase diagram is divided into 5 three-phase fields and two narrow two-phase fields are induced by the narrow Sr 14 Cu 24 Zn x O 41-y (0 ≤ x ≈ 0.1) solid solution.Some Zn substitution in CuO can also be inferred from the variation of the lattice parameters of CuO, but the solubility limit of Zn in CuO under the present equilibration conditions is lower than 1 at.%.Substitution of Zn for Cu in the Sr 14 Cu 24 O 41-y structure has a very strong effect on its electrical resistivity.The fact that ZnO is in equilibrium with all the other phases save SrO represents a favorable situation for processing and testing various properties of ZnO-based multiphase nanoparticles and films.

Fig. 4 .Fig. 5 .
Fig. 4. Low-temperature electrical resistivity of samples with nominal compositions belonging to the Sr 14 Cu 24-x Zn x O 41-y solid solution.Inset: c-axis lattice parameter of the Sr 14 Cu 24 O 41-y phase in samples with nominal Sr 14 Cu 24-x Zn x O 41-y (0.0 ≤ x ≤ 1.0) compositions
. The solubility limit of Co in the Sr 14 Cu 24 O 41-y structure (up to Sr 14 Cu 19 Co 5 O 41-y ) is significantly larger than that of Zn.In CuO -SrO -RE 2 O 3 systems (RE = rare earth element), RE substitution in Sr 14 Cu 24 O 41-y also takes place to various extents [37-51], however on the Sr sites rather than on the Cu sites, and the pseudoternary sections are dominated by SrRE 2 O 4 compounds.