«Oxalate-precursor processing for high quality BaZrO3 Corresponding Author Nigel M. Kirby* Curtin University of Technology Department of Applied ...»
Oxalate-precursor processing for high quality BaZrO3
Nigel M. Kirby*
Curtin University of Technology
Department of Applied Physics
G.P.O. Box U1987, Perth 6845, Western Australia, Australia
e-mail: N.Kirby@exchange.curtin.edu.au, fax: 61 (0)8 9266 2377
Nigel M. Kirby*, Arie van Riessen, C. E. Buckley, Vaughan W. Wittorff a
Curtin University of Technology
Department of Applied Physics
G.P.O. Box U1987, Perth 6845, Western Australia, Australia
Curtin University of Technology Department of Electrical and Computer Engineering G.P.O. Box U1987, Perth 6845, Western Australia, Australia e-mail: V.Wittorff@ece.curtin.edu.au fax: 61 (0)8 9266 2584 Abstract BaZrO3 is by far the most inert crucible material that has been used for melt processing of high quality single crystal YBCO superconductors. To overcome the processing difficulties of existing solid-state methods, solution processing methods are increasingly important in powder synthesis.
This study investigates several methods of producing oxalate precursors for subsequent thermal decomposition to BaZrO3 with a view to producing high quality BaZrO3 ceramics. The most favourable system used barium acetate, ammonium oxalate and zirconium oxychloride, which unlike other previously reported oxalate processes allowed near stoichiometric precipitation without requiring a large excess of Ba reagents, elevated precipitation temperatures or slow addition of reagents. Precise control over precipitate stoichiometry was achieved by variation of the solution Ba : [Zr+Hf] mole ratio without requiring accurate control over oxalate addition. XRF, XRD, N2 BET adsorption, DTA/TGA and TEM analysis showed this process to be capable of producing BaZrO3 powders suitable for ceramics applications. The phase purity, particle size and surface areas of BaZrO3 powders produced by calcination of these precursors can be adjusted by variation of stoichiometry and calcination temperature. Crucibles formed from oxalate precursors have been able to contain Y2O3-BaCuO2-CuO melts for up to seven days.
Keywords: Barium zirconate, YBCO, melt processing, oxalate precursor, chemical synthesis.
Introduction Crucible corrosion is an important factor hindering the routine synthesis of large, high purity rareearth barium cuprate superconductor single crystals. These compounds do not melt congruently, and must be grown in a molten flux system, for example YBa2Cu3O7-δ is crystallised from a BaCuO2-CuO eutectic melt. Molten BaCuO2-CuO is highly corrosive to substrate materials, and the corrosion products may lead to flux and crystal contamination, poor control over crystal growth conditions, or perforation and leakage of crucibles. BaZrO3 has been found to be inert to BaCuO2-CuO melts, but must be produced with high phase purity and low porosity to allow controlled and repeatable growth of high quality YBa2Cu3O7-δ single crystals 1. Off-stoichiometric or residual secondary phases lead to leakage of the crucible either through crack formation or percolation through grain boundaries. The corrosion performance of this otherwise highly resistant material is determined by trace phases, hence the key processing requirements are very accurate control over product stoichiometry and the attainment of high phase purity and density 1.
BaZrO3 crucibles for melt YBCO processing have so far been produced from BaCO3 and ZrO2 powders by solid-state synthesis, and considerable care is needed to produce ceramics of adequate quality. To produce powders of high phase purity, repeated grinding and re-firing is necessary to sufficiently complete the solid-state reaction of BaCO3 and ZrO2. BaZrO3 is a highly refractory material, and extended mechanical grinding is typically required to sufficiently reduce the particle size to allow sintering to high density 1. Milling contamination from ZrO2 ball milling media requires compensation by additional BaCO3 and may restrict the final phase purity.
Chemically synthesized ceramic powders have numerous potential benefits over solid-state derived powders including increased phase purity, reduced particle size, reduced milling demand, and improved sintering properties. There are numerous processes reported for the production of alkali-earth zirconates and titanates, including hydrothermal, sol-gel, sol precipitation, hydroxide, peroxide, oxalate, citrate and freeze-drying processes 2-18. Sol-gel and sol-precipitation methods in particular, have high equipment costs due to the sensitivity of reagents to moisture, and very high costs for reagents. Few investigations of solution chemical processes for crucible production for YBCO melt processing have been reported 3. The very high level of phase purity required for melt tightness is very difficult to assess directly from bulk measurements, and hence claims of the suitability of a process to provide melt tight ceramics require verification through melt exposure.
Oxalate processing may be more suited for industrial BaZrO3 processing than other chemical
control of the speciation of solution complexes, stability of solution speciation to pH, and control of microstoichiometry. The reaction products of zirconium compounds are known to depend on zirconium solution speciation, in particular its state of hydrolysis and polymerisation 19,21. Control over zirconium speciation adds complexity to zirconium processing and may lead to confusing or conflicting results. For example zirconium salts precipitated from freshly prepared solutions by alkali oxalates, tartrates and citrates are readily soluble in an excess of the precipitating agent, but in aged or previously boiled solutions the precipitate remains insoluble in an excess of the precipitating reagent. The speciation of zirconium is dependent on factors including pH, temperature, complexing agents, solution concentration and time. Because some reactions controlling solution speciation are irreversible, the complete reaction path including the order of mixing of reagents and thermal history may need to be controlled in order to direct the reaction to the desired outcome.
Previous workers have outlined oxalate processes which claim high quality BaZrO3 powders can be readily achieved. Reddy and Mehrotra reported a process for the production of barium zirconyl oxalate hydrate using barium chloride, zirconyl chloride and hot oxalic acid, though full details of temperatures, molar ratios of solutions, and the order and rates of addition were not reported. The precipitate was claimed to closely match BaZrO(C2O4).5H2O and to decompose to BaZrO3 at approximately 1000ºC. The results of attempts to reproduce this experiment are provided below and are not consistent with those previously reported. Potdar et al. reported a process using zirconyl nitrate and sodium oxalate to produce a soluble molecular precursor, which was subsequently reacted with barium nitrate to produce a precipitate of stoichiometric barium zirconyl oxalate. Gangadevi et al. showed the importance of starting reagents and pH in the control of product stoichiometry. These earlier studies required either a significant excess of barium reagents, elevated temperatures or reagents containing alkalis in order to produce a stoichiometric product.
The current research was conducted to develop a simple production process yielding high quality BaZrO3 powder, without requiring elevated temperatures, large excesses of barium reagents, or reagents containing alkalis. We required a powder for processing into a ceramic of sufficiently high quality for the demanding application of molten BaCuO2-CuO containment. Our primary concerns were control of the stoichiometry of precursors, calcination to BaZrO3 powders of high phase purity for sintering into dense ceramics of high phase purity, and verification of the tightness of sintered ceramics to YBCO melts.
1. Materials and Methods
• zirconium oxychloride, Millennium Performance Chemicals, Rockingham, Western Australia, gravimetric assay 36.8wt.% ZrO2+HfO2
• ZrOCl2.8H2O, Riedel-de Haën 99.5%+
• BaCl2.2H2O, AR-grade, Sigma Chemicals, Balcatta, Western Australia
• oxalic acid dihydrate, AR-grade, Sigma Chemicals, Balcatta, Western Australia
• Ba(CH3COO)2, Riedel-de Haën 99%+, gravimetric assay 99.46 wt.% • (NH4)2C2O4.H2O, Riedel-de Haën 99.5%+
1.2 Barium Chloride, Zirconium Oxychloride, Oxalic Acid System Preliminary experiments were performed following the method of Reddy and Mehrohtra in which “equimolar (0.5M each) aqueous solutions of barium chloride and zirconium oxychloride were added to the hot solution of oxalic acid (1.0M) which was 10% in excess” 14. In the current study, solutions of 0.5M BaCl2, 0.5M zirconium oxychloride and 1M oxalic acid were mixed using different orders of addition at temperatures of 80 to 95ºC at solution mole ratios of 1.00:1:2.20 BaCl2:[Zr+Hf]:H2C2O4. The precipitate was stirred at constant temperature for 30 minutes, cooled to ambient temperature, filtered using Whatman #6 filter paper, washed with deionised water and dried at 120ºC. The dried precipitate was calcined in yttria-stabilised ZrO2 crucibles in air at 1150ºC. A second series of experiments was performed to investigate the use of larger excesses of oxalic acid, because preliminary results showed that a severely barium deficient product was produced at solution mole ratios of 1.00:1:2.20 BaCl2:Zr:H2C2O4.
Oxalic acid addition in the barium chloride, zirconium oxychloride, oxalic acid system was optimised at a fixed 1.00:1 Ba:[Zr+Hf] solution mole ratio, at both 25º and 95ºC. A freshly prepared equimolar solution of zirconium oxychloride (0.25M) and barium chloride (0.25M) solution was added dropwise to 1M oxalic acid solution maintained at 95ºC under constant stirring. The volume of oxalic acid used was varied to study the above system with solution mole ratios of Ba:[Zr+Hf]:C2O4 over the range 1.00:1:2.50 to 1.00:1:3.00. The slurries were cooled to ambient temperature, filtered using Whatman #6 filter paper, washed twice with deionised water, dried at 100ºC in air, and calcined in air at 1000ºC for two hours.
Precipitation was also studied at 25ºC for Millennium zirconium oxychloride over the range 1.00:1:2.20-2.80 in order to assess the viability of ambient temperature production. The effect of rapid addition of mixed barium zirconium solution to oxalic acid at 95ºC at a solution Ba:[Zr+Hf]:C2O4 mole ratio of 1.00:1:2.60 was also studied.
1.3 Barium Acetate, Zirconium Oxychloride, Ammonium Oxalate System The potential benefit of performing precipitation at higher pH than oxalic acid systems was investigated using barium acetate, zirconium oxychloride and ammonium oxalate over the mole ratio range of 1.00:1:2.00-3.00 respectively. Zirconium oxychloride and ammonium oxalate solutions were mixed until a clear solution was formed at 25ºC (0.075M Zr, ~0.15M (NH4)2C2O4), then 0.25M barium acetate was added rapidly at ambient temperature under vigorous stirring. The slurry was stirred for 90 minutes and the precipitate was filtered using Whatman #6 filter paper, washed twice in de-ionised water, dried at 100ºC then calcined in air between 1000 and 1500ºC.
The same procedure was used for solution mole ratios 1.027:1:2.4 and 1.027:1:3.00.
1.4 Ceramic Fabrication Powders for ceramic fabrication were produced by rapid addition of barium acetate solution to a
Ba:[Zr+Hf]:(NH4)2C2O4 solution mole ratio at ambient temperature. Powders used for crucible fabrication were deliberately made slightly barium rich (Ba: [Zr+Hf] mole ratio 1.005 – 1.015), as our other work based primarily on solid-state derived BaZrO3 has shown corrosion resistance if dramatically reduced by the presence of even trace amounts of residual ZrO2 24,25. After washing, drying and calcination at 1300ºC, 3 wt.% cetyl alcohol was added as a pressing lubricant by ring milling in a solution containing cetyl alcohol dissolved in ethanol, after which the ethanol was evaporated at 80ºC. The lubricated powder was packed into a flexible mould with a stainless steel internal former and cold isostatically pressed at 140 MPa for 60 seconds. Sintering was conducted in a molybdenum silicide resistance furnace in air for 6 hours at 1700ºC. Sintered density was determined by Archimedes method.
Crucibles of 7mL capacity were tested for corrosion resistance to a mixture of Y2O3, BaCO3 and CuO (mole ratio of 1:32:90 respectively) at 1050ºC in air. The rate of leakage was observed visually for up to seven days.
1.5 Analysis Methods Millennium zirconium oxychloride, barium acetate and barium chloride reagents were standardised gravimetrically. Zirconium oxychloride was precipitated with DL-mandelic acid from acidic solution and assayed gravimetrically as ignited Hf/ZrO2. The mole ratio of HfO2:ZrO2 was determined by XRF as described in earlier work. In this study, Hf was assumed to have identical chemical properties as Zr, and XRF analysis was conducted for molarity control from gravimetric assays of Zr+Hf. Barium reagents were standardised by gravimetric assay of BaSO4 precipitated from HCl-acidified solutions with dilute H2SO4 after ignition at 1000 ºC.
BaO : [ZrO2+HfO2] mole ratios for all samples in this study were determined by x-ray fluorescence spectrometry (XRF) to an accuracy of ±0.002 using a procedure described in detail previously. X-ray diffraction (XRD) analysis was conducted using a Siemens D500 diffractometer with a Cu tube, using 1º incidence slits, 0.15º receiving slits, a graphite secondary monochromator, and scan speed of 0.3º 2θ/min with step increment of 0.02º or 0.04º 2θ. Analysis at 40 mA and 0.3º 2θ/min with step increments of 0.04º 2θ provided a detection limit of 0.15 wt.% BaCO3 (3σ counting errors) determined from calibration using experimental standards. Crystallite size was estimated by Voigt function profile fitting using the method of de Keiser et al. 23. Voigt function profiles were fitted to the (100) diffraction peak of BaZrO3 using SHADOW v.4.2 (Materials Data Inc. 1999). LaB6 (NIST CRM 660a) was used to measure instrument broadening.