Ba(OH)2 Equilibria in the System Ba-O-H-F, With Application to the Formation of Ba2YCu3O6.5 + x From BaF2-Precursors

The ex situ process for fabricating Ba2YCu3O6.5 + x superconducting tapes from BaF2- based precursors involves a hydration/oxidation reaction at ≈730 °C to 750 °C generally written as: (2BaF2+Y+3Cu)(amorphous)+(2H2O+2.25O2)(g)→Ba2YCu3O6.5+x(s)+4HF(g). However, microscopic observations of partially processed films suggest the presence of a transient liquid phase during conversion. Alternatively, the conversion reaction can be rewritten as the sum of several intermediate steps, including the formation of a barium hydroxide liquid: (BaF2)(amorphous)+2H2O(g)→Ba(OH)2(liq)+2HF(g). To evaluate the possibility of a hydroxide liquid conversion step, thermodynamic calculations on the stability of Ba(OH)2(liq) have been completed from 500 °C to 900 °C at 0.1 MPa ptotal. Based on currently available data, the calculated phase diagrams suggest that a viable hydroxide reaction path exists in the higher part of this temperature range. The calculations indicate that Ba(OH)2(liq) may be stable at log pH2O (Pa) values from ≈4 to 5, provided log pHF (Pa) values can be maintained below 0 to −1. Limited experimental confirmation is provided by results of an experiment on BaF2(s) at 815 °C, 0.1 MPa pH2O, in which essentially all F at the surface was replaced by O. It is therefore possible that processing routes exist for producing Ba2YCu3O6.5 + x based on the presence of a Ba(OH)2 liquid, which might have an effect on conversion rates and texturing in the superconducting film.

However, microscopic observations of partially processed films suggest the presence of a transient liquid phase during conversion. Alternatively, the conversion reaction can be rewritten as the sum of several intermediate steps, including the formation of a barium hydroxide liquid: (BaF 2 )(amorphous) + 2 H 2 O(g) → Ba(OH) 2 (liq) + 2 HF(g).
To evaluate the possibility of a hydroxide liquid conversion step, thermodynamic calculations on the stability of Ba(OH) 2 (liq) have been completed from 500 ºC to 900 ºC at 0.1 MPa p total . Based on currently available data, the calculated phase diagrams suggest that a viable hydroxide reaction path exists in the higher part of this temperature range. The calculations indicate that Ba(OH) 2 (liq) may be stable at log p H2O (Pa) values from ≈4 to 5, provided log p HF (Pa) values can be maintained below 0 to -1. Limited experimental confirmation is provided by results of an experiment on BaF 2 (s) at 815 ºC, 0.1 MPa p H 2 O , in which essentially all F at the surface was replaced by O. It is therefore possible that processing routes exist for producing Ba 2 YCu 3 O 6.5 + x based on the presence of a Ba(OH) 2 liquid, which might have an effect on conversion rates and texturing in the superconducting film.
promoting textured growth of the overlying oxide superconductor phase. One of the leading processes for producing highly textured second generation "coated conductors" is the "BaF 2 ex situ" method [5][6][7], a process which consists of two principal steps. In the first step, amorphous films having (2 BaF 2 + Y + 3 Cu) stoichiometry are deposited on the buffer layer by physical vapor deposition (PVD) or metal/organic deposition (MOD) techniques. The second step (shown diagrammatically in Fig. 1) involves the conversion of the amorphous film to superconducting material, nominally according to the following overall reaction: (2 BaF 2 + Y + 3 Cu)(amorphous) + (2 H 2 O + 2.25 O 2 )(g) → (1) Ba 2 YCu 3 O 6.5 + x (s) + 4 HF(g).
Typically, the reaction takes place at ≈730°C to 750°C. The principal advantages of the ex situ process are that the first step is a relatively simple, rapid operation completed without the necessity of substrate heating, and that the second, more time-consuming step can be completed separately from the deposition process, without the need for vacuum conditions. The exact nature of the conversion process remains controversial. Indirect evidence has been presented for the formation of an amorphous intermediate phase (possibly a liquid) during the conversion [8].

Previous Work
Research has previously been conducted in our laboratory to determine the nature of low-meltingtemperature liquids involved in the BaF 2 ex situ conversion [9][10][11]. The relevant phase equilibria can be discussed with reference to the Ba,Y,Cu//O,F quaternary reciprocal system, portrayed in Fig. 2a as a trigonal prism in compositional phase space. Initially, the relative thermodynamic stabilities of Table 1 were used to subdivide the trigonal prism of Fig. 2a into its constituent tetrahedra, shown in Fig. 2b. The phase stabilities defining these tetrahedra have since been confirmed experimentally, and hence the tetrahedra serve as a valid outline for discussion of the conversion process. In traversing along a compositional vector from the oxides at the base of the trigonal prism, to the fluorides at the top, the three constituent tetrahedra encountered are: BaF 2 -BaO-½Y 2 O 3 -CuO x , BaF 2 -YF 3 -½Y 2 O 3 -CuO x , and BaF 2 -YF 3 -CuF 2 -CuO x . The first two tetrahedra share the very stable BaF 2 -½Y 2 O 3 -CuO x compositional plane, where the ideal compositions of PVD precursor films would be plotted. In terms of fluoride/oxide ratio, the progression of such films during ex situ conversion to superconductor would ideally follow the path shown in the lower tetrahedron of Fig. 2b. Yet, melting temperatures below 815°C were not found in this tetrahedron, even in the presence of  In the BaF 2 -YF 3 -CuF 2 -CuO x tetrahedron, fluoride-rich liquids melting as low as 580°C were found; possibly these could be involved in the early stages of MOD

Possible Presence of Ba(OH) 2 During Ex Situ Conversion
Reaction (1) may be rewritten as the sum of several constituent reactions, based in part on high temperature powder x-ray diffraction (HTXRD) observations in our laboratory [10,13] and in other laboratories [14,15]. Moreover, under sufficiently high p H 2 O , the BaO component of Fig. 2 could effectively be replaced by Ba(OH) 2 , resulting in the following conversion steps: Reactions (5) and (6) indicate the presence of Ba(OH) 2 liquid as an intermediate step.

Goal and Approach of Present Investigation
The primary goal of this paper is to evaluate the possible presence of Ba(OH) 2 liquid during the ex situ conversion process, using a combination of experimental and calculative methods. Demonstrating the presence of liquids at high temperature is of course is not possible by HTXRD, and so we must rely on less direct methods. Also, maintaining the high p H 2 O and low p HF thought necessary to form Ba(OH) 2 liquids from BaF 2 at high temperature in a controlled experimental environment is in itself a daunting task. For this purpose, we constructed a special high-flow steam furnace, as described below. We have completed extensive equilibrium calculations on the stability of Ba(OH) 2 liquids using available thermodynamic data, including the effect of CO 2 contamination. Based on the calculated stabilities, we designed experiments using the highflow steam furnace to test for the formation Ba(OH) 2 (liq) according to its predicted stability field. As a background for this work, differential thermal analysis (DTA) experiments were completed on carefully chosen compositions.

Experimental Procedure 1
BaO was synthesized from 99.99 % purity (metals, by mass) BaCO 3 by vacuum-calcining at 1300°C for 10 h, followed by transfer to a glovebox. Complete decarbonation and conversion to BaO were verified by x-ray powder diffraction (XRD) in a sealed x-ray mount [16]. Powder XRD was completed on a Philips 2θ diffractometer 1 using Cu K α radiation and a graphite monochromator. Diffractometer control and data acquisition were achieved using the JADE software system.
For melting studies, a Ba(OH) 2 /BaO mixture was prepared by controlled hydration of BaO. Melting and annealing experiments with simultaneous differential thermal analysis/thermogravimetric analysis (DTA/TGA) were completed in a Mettler TA1 thermoanalyzer outfitted with Anatech digital control and data acquisition electronics. DTA was completed at a ramp rate of 10°C/min. The DTA apparatus was calibrated against the α/β quartz transition (571°C) and the NaCl melting point (801°C); DTA temperatures are estimated to have < ± 3°C standard uncertainty. DTA crucibles were of dense slip-cast MgO.
To enable experiments at high p H 2 O , a furnace capable of operation up to p H 2 O = (0.1 MPa) at temperatures above 800°C was constructed, as shown schematically in Fig. 3. The special features of this furnace are: 1. a 3 kW steam generator provides an approximately constant-pressure source of steam for the furnace; 2. the steam generator uses a water feedstock with reduced CO 2 (discussed below); 3. all areas of the furnace and supply lines are maintained at >100°C; 4. a type S thermocouple is positioned in the base of Volume 110, Number 2, March-April 2005 Journal of Research of the National Institute of Standards and Technology the sample crucible within 2 mm of the sample (TC-2); 5. the steam inlet is positioned in the crucible directly above the sample such that the steam flow must circulate around the sample before exiting the crucible; 6. steam flow can be reproducibly metered via a heated control valve; 7. a rapid flow of steam directly over the sample is possible. Steam furnace experimental results were examined optically using a Leica Wild M10 stereomicroscope and SPOT digital image capture software. Results were also characterized with an AMRAY 1400 scanning electron microscope (SEM) operated at ≈ 20 kV. Elemental compositions were determined with a Gresham energy dispersive x-ray spectrometer (EDS) and a 4Pi Analysis data acquisition interface controlled by Revolution software.

Calculative Procedure
Solid-liquid-gas equilibria in the system Ba(OH) 2 -BaO-BaF 2 -H 2 O-HF (and also with CO 2 ) were calculated as a function of temperature using the Janaf thermochemical database [17] and the FactSage software suite [18]. The latter uses a Gibbs energy minimization algorithm to calculate equilibrium concentrations of species in the gas phase, and to calculate activities of liquid and solid phases. The calculative procedure involves minimizing the total stoichiometric summation of terms: (7) for the collection of phases being considered, where G(i,T) = Gibbs energy function for phase i, ∆ f H°m(i, 298) = standard molar enthalpy of formation for phase i, S°(i, 298) = absolute molar entropy for phase i, C p (i) = molar heat capacity of phase i at constant pressure (0.1 MPa) , and T = temperature kelvin. For solution phases, appropriate mixing terms must be added to Eq. (7), including terms of the form RTln(p j ), where R = gas constant and p j = activity or partial pressure of species j. Species considered in the calculations are shown in Table 2. For the present calculations, gaseous species were considered to mix ideally. As described below, for purposes of calculation we assume there is negligible solubility of BaO and BaF 2 in Ba(OH) 2 liquids at the eutectic. Therefore to a first approximation the liquids are simple liquids, essentially pure. With regard to solid phases, there is evidence for solid solution of BaO in BaF 2 [19], although a quantitative determination has not been made, and the extent is not known. Also, existence of a compound intermediate between BaF 2 and Ba(OH) 2 has been proposed [15].

Effect of p CO 2 on Ba(OH) 2 Equilibria
One of the reasons for the relative success of the BaF 2 ex situ method is the elimination of carbonate (usually present as BaCO 3 ) from the processing route. Carbonate not only affects the kinetics of Ba 2 YCu 3 O 6.5 + x (s) formation due to the relative stability of BaCO 3 (s), but also affects the equilibrium phase assemblage by favoring formation of oxycarbonate phases such as Ba 4 Y 2 O 7 · xCO 2 and Ba 2 Y 2 O 5 · xCO 2 [21,22]. Formation of these phases interferes with Ba 2 YCu 3 O 6.5 + x (s) formation and complicates the decarbonation reactions which must take place for formation of Ba 2 YCu 3 O 6.5 + x (s) from BaCO 3 -containing precursors. Furthermore, presence of carbon in the superconductor phase has been shown to adversely affect properties [23].
Calculations have been performed to estimate the effect of CO 2 on Ba(OH) 2 equilibria, by adding CO 2 (g), three polymorphs of BaCO 3 (s), and BaCO 3 (liq) to the list of species considered in Table 2. These calculations have been completed to determine the values of p H 2 O and p CO 2 which are in equilibrium with coexisting Ba(OH) 2 and BaCO 3 , as shown in Fig. 4. From the curves in Fig. 4, at 500°C, a p CO 2 of < 10 -4.6 Pa is required to avoid formation of BaCO 3 , while a p H 2 O of > 10 1.8 Pa) is necessary to stabilize Ba(OH) 2 . At 900°C, p CO 2 must be < 10 1.4 Pa) to avoid formation of BaCO 3 , and p H 2 O of > 10 4.5 Pa is required to stabilize Ba(OH) 2 . For comparison, at 25°C, water in equilibrium with air (p CO 2 = 10 1.5 Pa) contains ≈ 10 -7 mol fraction dissolved CO 2 . Water of this composition, if vaporized, would be within the BaCO 3 stability field up to ≈ 625°C. At 99°C, water in equilibrium with air contains ≈ 10 -9 mole fraction CO 2 , and if vaporized, would fall below the BaCO 3 stability line above ≈ 525°C. From these calculations it is clear that water used to generate p H 2 O must be heated to reduce CO 2 content if formation of BaCO 3 is to be avoided. This is especially critical in experiments involving the formation of Ba(OH) 2 , such as those discussed below.

Ba(OH) 2 /BaO
The melting of a Ba(OH) 2 /BaO mixture was investigated by DTA to estimate the effect of BaO solubility on the melting temperature. A lowering of ≈ 3°C was found, which lies within the estimated uncertainty of the DTA measurements. It is therefore concluded that the solubility of BaO in Ba(OH) 2 liquids is negligible, and that Ba(OH) 2 liquid can be modeled for calculational purposes as essentially pure Ba(OH) 2 .
Accordingly, equilibria between solid and liquid Ba(OH) 2   The effect of HF(g) ex situ conversion product on Ba(OH) 2 (liq)/BaO(s) equilibria must also be considered. As shown in Fig. 6, the log p H 2 O values for the Ba(OH) 2 (liq)/BaO(s) equilibrium remain constant and independent of log p HF for any given temperature. However the maximum value of log p HF to which the Ba(OH) 2 (liq)/BaO(s) equilibrium is stable increases with increasing temperature. Above this value, indicated by the dashed parts of the equilibrium lines at any given temperature, the Ba(OH) 2 (liq)/BaO(s) equilibrium is metastable. Gas phase compositions for selected points near the centers of the equilibrium boundaries in Fig. 6 are given in Table 4. Comparison of Table 3 and Table 4 reaffirms the constancy of the equilibrium p H 2 O in the presence of HF.

Ba(OH) 2 /BaF 2
Due to the formation of Ba(OH) 2 -hydrates, it was not possible to prepare pure Ba(OH) 2 from BaO during this investigation, and therefore the melting point lowering of Ba(OH) 2 /BaF 2 mixtures was not investigated experimentally. For calculational purposes it is therefore     assumed that the solubility of BaF 2 in Ba(OH) 2 (liq) at the eutectic is small, a reasonable assumption, given the much higher melting point of BaF 2 (> 1300°C).
To describe the equilibrium between Ba(OH) 2 (liq) and BaF 2 (s), it is necessary to consider p HF as well as p H 2 O . Figure 7 shows calculated curves for the Ba(OH) 2 (liq)/ BaF 2 (s) equilibrium at several temperatures. As the temperature increases, the maximum p HF value to which the Ba(OH) 2 (liq) stability field extends increases. As temperature increases, the minimum p H 2 O value to which the Ba(OH) 2 (liq) stability field extends also increases. Below this value, the Ba(OH) 2 (liq)/BaF 2 (s) equilibrium is metastable. Gas phase compositions at selected points on the curves in Fig. 7 are given in Table 5. At higher temperatures, as for the Ba(OH) 2 (liq)/BaO(s) equilibrium, the values of p Ba(OH) 2 become significant. From the slopes of Fig. 7 and the data in Table 5, a given increase in p H 2 O results in a proportionately much larger increase in the equilibrium p HF . For example, an order-of-magnitude increase in p H 2 O produces a two orders-of-magnitude increase in p HF . The requirements for HF(g) removal during the conversion of BaF 2 (s) to Ba(OH) 2 (liq) are significantly reduced by maintaining increased p H 2 O , as well as by maintaining higher temperatures.

Ba(OH) 2 /BaO/BaF 2
Treatment of equilibria involving the three phases Ba(OH) 2 , BaO, and BaF 2 is facilitated by the use of the 3-D plot shown in Fig. 8, where the three-phase Ba(OH) 2 /BaO/BaF 2 equilibrium is represented as a function of p H 2 O , p HF , and temperature. The three-phase equilibrium is actually a curve in log p H 2 O -log p HFtemperature parameter space, as illustrated by its projection onto the base of the plot. Table 6 gives gas phase compositions at selected points along the threephase equilibrium.
The data in Fig. 8 can also be conveniently represented through the use of isothermal sections, as shown in Fig. 9. Here the three-phase Ba(OH) 2 /BaO/BaF 2 equilibrium plots as a point at the juncture of the areas   corresponding to the Ba(OH) 2 , BaO, and BaF 2 stability fields. With increasing temperature, the stability field of Ba(OH) 2 shrinks to higher p H 2 O , but simultaneously expands to higher p HF . On the basis of Fig. 9, a steam furnace experiment was designed to test for the formation of Ba(OH) 2 (liq) from BaF 2 (s). Using the steam furnace, p H 2 O of 0.1 MPa at temperatures of 800°C or above can be maintained for extended periods of time. These are the optimum conditions for formation of Ba(OH) 2 (liq) according to reaction (5), provided p HF can be maintained at < 1 Pa. In practice, the latter requirement can be met by rapid-ly flowing steam over the sample to remove product HF, thereby reducing p HF to low levels.
Results of a steam furnace experiment in which a single crystal fragment of optical quality BaF 2 was held at 815°C, for 2h, with p H 2 O = 0.1 MPa, with a steam flow over the sample estimated at > 0.2 L/s, are shown in Fig. 10a. Fluorine on the surface of the BaF 2 has been uniformly replaced by oxygen to a depth of at least 1 µm, as estimated by the lack of a fluorine EDS signal from the underlying BaF 2 (Fig. 10b). The full EDS spectrum (not presented) shows Ba and O as the main constituents on the surface of the reacted crystal, with a relatively small C K α peak (no method was available for detection of hydrogen). The smooth, dense nature of the reacted surface is consistent with the formation of Ba(OH) 2 liquid. As discussed above, CO 2 in the water feedstock for this experiment was reduced to low levels such that the formation of BaCO 3 was minimized. The EDS spectrum of Fig. 10b indicates that BaCO 3 was not a major reaction product, as the C K α intensity is similarly low for both reacted and unreacted crystals. The presence of C in both spectra is an indication of minor hydrocarbon surface contamination. Smaller peaks present in the reacted sample are due to trace contaminants from the steam boiler and transport line (Si, possibly Co), and are not likely to have had a significant effect on the F/O reaction. We conclude that formation of Ba(OH) 2 (liq) from BaF 2 (s) according to reaction (5) is the most probable explanation of the results in Fig. 10.
The high-flow experiment at 815°C, p H 2 O = 0.1 MPa may be near the upper limit of conditions useful for practical processing of second-generation coated conductors, due to the thermal limitations of currently available substrate/buffer combinations. From Fig. 9, production of Ba(OH) 2 liquids at lower temperatures requires more complete removal of HF from the reaction site. An estimate of the requirements for Ba(OH) 2 (liq) formation according to reaction (5) at 700°C can be made as follows. First, it must be noted that, at 700°C, p HF must be ≈< 10 -1 Pa for Ba(OH) 2 (liq) to form. At a steam flow rate of 0.2 L/s over the sample, this gives a maximum HF removal of 10 -3.8 mL/s from the reaction site, assuming equilibrium. The resulting rate of formation of Ba(OH)(liq) is 10 -8.1 mol/s. If reaction (5) is the rate-limiting step, then Ba 2 YCu 3 O 6.5 + x would be formed by reaction (6) at a rate of ≈ 10 -8.4 mol/s. For a 1 cm 2 area, this corresponds to a thickness conversion rate of ≈ 10 -2.3 µm/s at 700°C. Thus, formation of Ba 2 YCu 3 O 6.5 + x superconducting films of 1 µm thickness over an area of 1 cm 2 could conceivably be achieved in 200 s to 250 s, with potential for scale-up   to much larger production rates. High-performance films with 1 µm or greater superconductor thickness could find immediate application in second-generation high T c technology, especially if production costs approach the anticipated $10/kA-m target [24].

Summary and Conclusions
Thermodynamic calculations have outlined a stability field for Ba(OH) 2 (liq) as a function of p H 2 O , p HF , and temperature, based on presently available data. An experiment at 815°C, p H 2 O = 0.1 MPa has provided evidence for the formation of Ba(OH) 2 (liq) by defluorination of BaF 2 (s), as predicted by the calculations. It is possible that under conditions of high p H 2 O and rapid gas flow, the formation of Ba(OH) 2 (liq) may occur as an intermediate step in the formation of superconducting Ba 2 YCu 3 O 6.5 + x (s) from amorphous (BaF 2 , Y, Cu) precursors. The presence of a Ba(OH) 2 liquid could be important for Ba 2 YCu 3 O 6.5 + x (s) processing for several reasons. It is well known that a liquid phase enhances mobility and can aid in local mass transport and the development of oriented microstructures. The presence of a liquid would be expected to improve the kinetics of Ba 2 YCu 3 O 6.5 + x (s) phase formation, although it may be necessary to limit the presence of the liquid phase at some stages during processing to prevent random growth, as opposed to oriented growth. Clearly, it is essential to  control the amount of liquid in order to fully optimize all aspects of Ba 2 YCu 3 O 6.5 + x (s) formation using the BaF 2 ex situ process. With sufficient data on liquid formation, p H 2 O provides an additional parameter, along with precursor F/O composition, gas flow, temperature, time, and p O 2 , with which to reproducibly control the processing of Ba 2 YCu 3 O 6.5 + x (s).
The calculated phase diagrams require further experimental verification to establish the role of BaF 2 solubility in the liquid, and the precise boundaries of the Ba(OH) 2 stability field. Preliminary experiments have indicated that Ba 2 YCu 3 O 6.5 + x (s) is stable in the presence of Ba(OH) 2 (liq) [25]; it is essential to know the range of conditions under which the Ba 2 YCu 3 O 6.5 + x (s) and the Ba(OH) 2 stability fields overlap. Based on the extent to which Ba(OH) 2 -based liquids extend into the phase space of Fig. 2, it may prove possible to design processing routes to control intersection of the PVD and MOD processing paths with the hydroxide liquid phase field. Then the full range of p H 2 O -temperature processing space can be explored to determine if there are new processing routes which might lead to further optimization of superconductor formation and film properties.