This is a preprint of a paper intended for publication in a journal or proceedings. Since changes may be made before publication, this preprint should not be cited or reproduced without permission of the author. This document was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, or any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for any third party’s use, or the results of such use, of any information, apparatus, product or process disclosed in this report, or represents that its use by such third party would not infringe privately owned rights. The views expressed in this paper are not necessarily those of the United States Government or the sponsoring agency.

INL/CON-10-20068PREPRINT

On the Preparation of TiAl Alloy by Direct Reduction of the Oxide Mixtures in Calcium Chloride Melt

Fray International Symposium

Prabhat K. Tripathy Derek J. Fray

November 2011

1

On the preparation of TiAl alloy by direct reduction of the oxide mixtures in calcium chloride melt

Prabhat K.Tripathy† and Derek J. Fray‡

† Idaho National Laboratory, P.O. Box 1625, Idaho Falls, ID 83415, USA

‡Department of Materials Science and Metallurgy, University of Cambridge, Pembroke Street, Cambridge, CB2 3QZ, UK

Key words : titanium aluminide, intermetallic alloy, molten salt, polarization, in situ

alloy formation

ABSTRACT

Titanium aluminide based intermetallic alloys offer excellent high temperature properties and are being considered, as structural materials, for extensive applications in high speed civil propulsion system and commercial automobile sectors. As an alternative approach, to the traditional process that involves multi-step unit operations, a cost-effective molten salt based process has been attempted to prepare this alloy. The new preparative process describes co-reduction of the oxide mixtures in fused calcium chloride electrolyte which enabled the preparation of the alloy in just one step. The effects of various experimental parameters, on the preparation of the alloys, have been discussed. The alloy powders were characterized and evaluated by powder XRD, SEM-EDX and residual oxygen measurements. 1.INTRODUCTION

Titanium aluminides belong to the class of advanced intermetallic materials that have attractive engineering properties. Because of the low densities (iron and nickel aluminides are nearly as twice as dense1) and better properties at elevated temperatures, these materials find applications as structural components in extremely demanding conditions. Owing to the existence of a long range order, in these alloy phases, these materials can effectively reduce dislocation mobility and diffusion processes at elevated temperatures and hence can exhibit attractive elevated temperature properties, such as higher strength (up to ~ 8000C), stiffness, resistance to environment etc2-3. Precisely because of this reason, these alloys are increasingly being considered for application in areas such as (i) manufacturing of blades in industrial gas-turbine engines3 (ii) automobile/transport sector4 (passenger cars, trucks and ships) and (iii) aerospace industry5 (jet engines and high speed civil transport propulsion system). Since the microstructures of these intermetallic alloys affect, to a significant extent, their ultimate performance, further improvements (by way of alteration/modification of these microstructures), have been the subject matter of intense research investigations. It has now been established that certain alloy additives, such as niobium, tantalum, manganese, boron, chromium, silicon, nickel and yttrium etc. in specific quantities, impart marked improvement to their properties that include fatigue strength, fracture toughness, oxidation resistance and room temperature ductility5.

2

From three possible compounds (TiAl, TiAl2 and TiAl3), it is the TiAl (also known as gamma titanium aluminide) which has the highest melting point and associated properties, required for key engineering applications4. Although for certain automotive engine valves, unalloyed TiAl has been found suitable, small addition of chromium, silicon and niobium are required to improve creep and oxidation resistance as well as ductility for certain other types of engines4. Niobium is the only element that increases the cost of �-TiAl significantly4. In the family of �-TiAl alloys, two specific alloys (Ti-48Al-2Nb-2Cr and Ti-45Al-2Nb-2Mn plus 0.8 vol.% TiB2) are preferentially used for structural applications, in the temperature regime 560-7500C, in advanced aircraft turbine engines5.

The conventional fabrication process of these alloys includes several unit

operations such as (i) mixing of the individual metallic components (ii) preparation of the alloy button by arc casting/melting (iii) forging (or extrusion) of the alloy (iv) homogenization of the alloy through heat-treatment at temperatures up to 12000C followed by (iv) annealing of the alloy (in the temperature regime 700-11000C). An electrochemical process, as an alternative fabrication route, can offer significant economic advantages. The new process, popularly known as FFC Cambridge process6, after the names of the discoverers, Fray, Farthing and Chen) can, in principle, be used to prepare these alloys directly from their oxides in just one step. This process has been successfully used to reduce the individual metal oxides, viz. TiO2

7, Al2O38, Nb2O5

9 and Cr2O3

10, to their respective metallic components, in fused calcium chloride based electrolytes. The present investigation describes the experimental results pertaining to the preparation of Ti-47Al-2Nb-2Cr alloy from the mixed oxides. The effect of experimental parameters on the quality of the alloy powder has been described. 2.EXPERIMENTAL

2.1. Apparatus

The polarization experiments were carried out in a one end closed vertical inconel 600 reactor vessel (125 mm inner diameter, 625 mm long with a 6mm thick lower flange, welded to its open end). A matching circular stainless steel top lid (upper flange), with 5-6 holes drilled on its center, was used for sealing the reactor. The electrodes were placed through those holes, with the help of the silicone rubber bungs. Both the silicone rubber bungs and viton “O” rings were used to secure the reactor leak/gas-tight. Gas inlet and outlet arrangements were made by welding 9 mm i.d. stainless tubes on the outer surface of the top lid. Continuous (high purity) argon flow (in the range 100-150 mL/min) was maintained throughout the entire duration of a particular experimental cycle. The argon gas, drawn from a commercial cylinder, was rendered moisture-free by connecting the gas inlet tube to a self-indicating calcium sulphate (Drierite, 8 mm) column and allowing the gas to pass through the column before it entered into the reactor. The molten electrolyte was contained in a re-crystallized alumina crucible (internal diameter : 8.5 cm ; height : 10 cm). The schematic diagram of the experimental set up, used in the present investigation, is described in one of the previous publications11.

3

2.2. Electrolyte

Both nominally anhydrous (97% pure) and di-hydrate calcium chloride (CaCl2.2H2O) were used as the electrolytes. The salts were subjected to a programmed heating up to a temperature of 1800C, under dynamic vacuum (1Pa), for several hours prior to use. 3.3. Cathode material

Calculated quantities of high purity TiO2, Al2O3, Nb2O5 and Cr2O3 (99.5% pure, particle size ~ 1-2μm) powders were thoroughly mixed and blended with a mixture of 0.1 wt.% (by mass) of poly(vinyl butyral-co-vinyl alcohol–co vinyl acetate-80% by weight and 1ml of poly ethylene glycol (average molecular mass 200, Sigma Aldrich make). The blended powder was converted to a slurry with isopropyl alcohol. The partially-dried slurry was then ball-milled, at room temperature, for a duration of ~18h. The blended powder was pressed into cylindrical pellets by uniaxial pressing in a laboratory hydraulic pressing unit with an applied pressure in the range 5-20MPa. The green pellets were sintered in air, for a duration of 2h, in the temperature range 950-11000C in order to be able to maintain their mechanical integrity during the subsequent polarization experiment. Each pellet weighed about 4 g. The diameter and thickness of the pellets were 25 mm and 5mm respectively. 3.4. Anode material

Standard graphite round bars (6-8mm dia. and 300 mm long), procured from Tokai Carbon, UK Ltd., were used as the anode material.

3.5. Preparation of the electrolytic bath

Calculated quantities of the pre-dried CaCl2 was kept in the alumina crucible, which, in turn, was placed inside the reactor. Through the top-lid, two graphite electrodes, one oxide electrode and one thermocouple were positioned properly and then kept in suspended positions just above the alumina crucible. The top-lid was tightened properly with six metallic clamps. Both argon gas and cooling water inlet and outlet connections, to the cell, were made. Argon gas was passed continuously till completion of the polarization experiment. 3.6. Procedure

The salt was slowly heated to a temperature of 5000C, for a period of 12h, under continuous argon flow. The temperature was then gradually raised to the operating temperature (9500C). The thermocouple and graphite electrodes were positioned properly in the melt by pushing them gently through the top lid. The melt was then pre-electrolyzed at different applied voltages for various durations. After the pre-electrolysis, one of the graphite electrodes was pulled out of the molten electrolyte and then the oxide cathode was lowered into the molten CaCl2. The polarization experiments were carried out, at a given (constant) potential, with the help of a DC power supply (Thurlby Thandar PL154 (15V and 4A). The voltage measurements were made with an Agilent 34970A

4

data acquisition/switch unit. The cell current was measured by monitoring the potential drop across a1� resistor.

Upon termination of the electrolysis, the heating was switched off and the furnace was allowed to cool to room temperature under continuous argon flow. The reduced pellets were taken out of the furnace and washed successively with (i) distilled water (ii) 1N HCl (iii) distilled water and (iv) acetone several hours before finally drying them at 800C under vacuum. The reduced pellets were evaluated and characterized by powder XRD, SEM-EDX and residual oxygen measurements techniques respectively. 4. RESULTS & DISCUSSION Some of the experimental parameters that were found to significantly influence the electroreduction reaction and hence the overall product purity were applied voltage, oxide porosity, duration of polarization and electrolyte composition. The overall chemical reaction in the cell

C/CaCl2,CaO/MO2 (i) can be described as follows

MO2(s) + C(s) = M(s) + CO2(g) (ii)

Although the anode gas composition can be a mixture of both carbon monoxide and carbon dioxide, the major component of the gas phase, however, was determined to be carbon dioxide at the operating temperature. The theoretical reaction potentials for different oxides at 9500C12, assuming that the reactants and products are in their standard states, have been provided in Table1.

Table 1

Standard free-energy change and theoretical reaction potential values for carbothermic reduction of the (alloy) constituent metal oxides at 9500 C

Reaction �G0

9500

C Theoretical Reaction (kJ) Potential (V) Cr2O3(s) + 1.5C(s) = 2Cr(s) + 1.5CO2(g) 226.405 -0.39 Nb2O5(s) + 2.5C(s) = 2Nb(s) + 2.5CO2(g) 379.63 -0.39 TiO2(s) + C(s) = Ti(s) + CO2(g) 327.082 -0.85 Al2O3(s) + 1.5C(s) = 2Al(s) + 1.5CO2(g) 693.505 -1.2 2CaO(s) + C(s) = 2Ca(s) + CO2(g) 616.138 -1.6 CaCl2(s) = Ca(s) + Cl2(g) 617.5 -3.2

The applied cell voltage, during the electroreduction reaction, was maintained in

the range 2.8-3.1V in order to take care of IR drop and the overvoltage associated with the evolution of oxides of carbon at the graphite anode surface. The theoretical reaction potential (Table 1) suggests that the four oxides perhaps underwent a sequential reduction

5

step, both chromium oxide and niobium oxides being the first to be reduced followed by the reduction(s) of titanium and aluminium oxides. Among the four oxides, the free energy of formation of both niobium pentoxide and alumina are highly negative (�G0

9500

C, Nb2O5 = -1369.838 kJ/mol ; Al2O3 = -1287.63 kJ/mol respectively12). Although Nb2O5 could be successfully converted to niobium metal and its alloys8, 13-15, complete conversion of aluminium oxide to aluminium metal was somewhat difficult16. However, in the present set of experiments, formation of different intermediate phases, involving chromium, titanium and niobium during simultaneous reduction of oxides, perhaps contributed to the enhanced reduction reaction of alumina to some extent. The current vs. time profile, during a typical electroreduction experiment (Fig. 1), initially peaked (peak current was dependent upon the mass of the pellet) before registering a gradual declining trend. Unlike, in case of pure TiO2 reduction, the initial rise in current was rather sluggish. The decline in electrolysis current took place through several humps/spikes (Fig. 1). Formation of those spikes/humps were indicative of the progress of the reduction reaction through the formation of several multiphasic reaction intermediates. The onset of the reduction reaction was observed to take place at an applied voltage greater than 2.5V, for a two- electrode configuration. It was also observed that after about 90,000s, the background/residual current, after attaining a minimum value, started rising very slowly, which perhaps indicated the occurrence of another electrode kinetics in addition to the oxygen removal.

Fig. 1 : Current vs. time profile, obtained during the electroreduction of TiO2-Al2O3-

Nb2O5-Cr2O3 mixed oxide pellet, Electrolyte : vacuum dried (nominally anhydrous) and pre-electrolyzed CaCl2-0.1 mol% CaO, Temperature : 9500C, Applied voltage : 3V ; mass of the pellet : 4.4g, pellet diameter : 2.5 cm, pellet thickness: 0.5 cm, sintering

temperature of the pellet : 10750C for a duration of 2.5h, open porosity : 45%

The color of the fully reduced mixed oxide pellet changed from somewhat dull yellow, prior to the reduction (Fig. 2) to complete grey/black (Fig. 3). These pellets were also found to contain somewhat dark bluish mass in addition to grey/black materials.

0

0.5

1

1.5

2

0 20000 40000 60000 80000 100000

Time(s)

Cur

rent

(A)

6

Partially reduced pellets were found to contain somewhat whitish layer in the bulk/core of the pellet. A quick room temperature measurement of the conductivity revealed that better reduced pellets were conducting all across the entire cross-section of the reduced surface which indicated that samples were predominantly metallic. As expected, fully reduced pellets were found to have compositional gradients, from surface to bulk, because of the formation of different phases across the entire reduction cross section. The existence of a compositional gradient was visible to the naked eye, which was also confirmed by the subsequent SEM-EDX analysis.

Fig. 2 : Photograph of the oxide precursors at different stages : top pellet (sintered at 11000C for 2.5h), middle pellet (sintered at 9500C for 2.5h) and the bottom pellet (before

sintering)

Fig. 3 : Reduced alloy pellet : Polarization temperature : 9500C ; Electrolyte : pre-electrolyzed vacuum-dried (nominally anhydrous) CaCl2 (0.1 mol% CaO) ; duration of

polarization : 30h ; Applied voltage : 3V

7

The pellets were also found to have undergone shrinkage both in diameter and thickness. While the shrinkage in diameter was uniform (about 40%) across the entire cross section of the reduced pellet, the reduction in thickness (in some pellets) was uneven and varied between 40-60%. Yet in some other case, the reduction in thickness was as low as 12-14%. Relatively lower levels of shrinkages (both in diameter and thickness) were observed in better reduced samples.

The reduced pellets were also analyzed by powder XRD. The major phase was

detected to be TiAl with the presence of traces of Ti3Al and Ti2AlNb phases. SEM-EDX Characterization of the reduced alloy pellets

The alloy pellets, upon examination under a scanning electron microscope, revealed different microstructure of the fractured surfaces [Figs. 4(a-d)]. These microstructures represented the distribution of different types of phases.

(a)

8

(b)

(c)

(d) Fig. 4 : SEM photograph of the fractured surface of the alloy pellet, prepared after 70h of

polarization in a pre-electrolyzed melt. The applied voltage, during reduction, was gradually raised from 2.7 to a final value of 3.1V, (a) : shows two types growth pattern

(nodular and plate-type) ; (b) : nodular growth at a higher magnification ; (c) : white dots represent the presence of metallic niobium and (d) : only plate-type growth pattern

Compositional (EDX) Analysis The elemental analysis of the porous alloy powders was carried out by EDX. The distribution of elements was observed to be inhomogeneous across the thickness of the reduced surface (Tables 2 and 3). Table 3 describes the required vales of different

9

elements in a typical alloy compositions and the range of values obtained in the reduced pellets. Table 4 describes the elemental compositions across the specific sections of a reduced pellet.

Table 3

Typical range of elements (EDX), present in different sections of a reduced pellet, prepared after 70h of polarization at applied voltages in the range 2.7-3.1V

Element Required amount

(mass.%) Range of values, obtained at different sections of

the pellets (mass.%) Ti 60.11 48.2- 66.5 Al 32.48 30.43- 39.95 Nb 04.76 0.34- 2.62 Cr 02.66 0.91- 2.49 Fe Trace quantity 0- 1.21 Si Trace quantity 0- 0.45

Oxygen 0.06-0.07 0- 1.67 Table 4

EDX (bulk) analysis of different sections of a particular reduced pellet (Fig. 4a

and 4b)

Figure/SEM Photograph No. Compositional analysis (mass%) Figure 4a Ti = 66.5, Al = 39.95, Nb = 2.57, Cr =

0.91, Si = 0.18, Fe = 0.07 and Oxygen = 1.23

Figure 4b Ti = 59.87, Al = 35.37, Nb = 0.44, Cr = 1.49, Si = 0.01, Fe = 0.52 and Oxygen = 1.13

Although elemental analysis suggests that it was possible to obtain them in the

required range, a more representative elemental analysis requires prior homogenization and consolidation of the alloy powders under inert environment/high vacuum melting conditions. The oxygen values, as determined by ELTRA oxygen analyzer, were relatively higher in the reduced pellets (~1.9 mass%). As compared to the surface layer/skin of the reduced pellets, the inner core/bulk (of the reduced pellets) was found to have relatively lower amounts of oxygen. However, experimental conditions can be modified to achieve lower values of residual oxygen contents in the reduced alloy pellets17. 5. CONCLUSIONS

Direct electrochemical reduction of the mixed oxides in calcium chloride melt has shown the possibility of developing an alternative and cost-effective production method for �-TiAl aluminides with desired additives such as niobium and chromium. More

10

number of experiments are required to optimize the process conditions with a view to fabricating these alloys with required compositions. REFERENCES

1. F.H.Froes, C. Suryanarayana and D. Eliezer, “Synthesis, properties and applications of titanium aluminides”, J. Mater. Science, 27 (1992) 5113-40.

2. B.H.Kear, C.T.Sims, N.S.Stoloff and J.H.Westbrook (Eds.), “Ordered Alloys – Structural Applications and Physical Metallurgy”, Claitor, Baton Rouge, Louisiana, 1970.

3. Guy Molénat, Marc Thomas, Jean Galy and Alain Couret, “Application of Spark Plasma Sintering to Titanium Aluminide Alloys”, Advanced Engineering Materials, 9, No. 8 (2007) 667-9.

4. Dixon Chandley, “Use of gamma titanium aluminide for automotive engine valves”, Metallurgical Science and Technology, 18, 1 (2000) 8-11.

5. Paul A. Bartolotta and David L. Krause, “Titanium Aluminide Applications in the High Speed Civil Transport”, in “Gamma Titanium Aluminides”, Int. Symp., San Diego, California, Feb. 28-March 04, 1999, NASA/TM-1999-209021, pp. 15.

6. Z. Chen, D.J. Fray and T.W. Farthing, "Direct electrochemical reduction of titanium dioxide to titanium in molten calcium chloride", 2000, Nature 407, 361-4.

7. C. Schwandt, D.J. Fray, “Determination of the kinetic pathway in the electrochemical reduction of titanium dioxide in molten calcium chloride”, Electrochemica Acta, 51 (2005) 66-76.

8. X. Y. Yan and D J Fray, "Production of Niobium Powder by Direct Electrochemical Reduction of Solid Nb2O5 in a Eutectic Melt", Metallurgical and Materials Transactions 2000, 33B, 685-693.

9. A. Cox and D.J Fray, "Production of Aluminium, Magnesium and Aluminium-Magnesium Alloys by direct electrochemical reduction of their solid oxides", Molten Salts XIII, The Electrochemical Society, 2002, 19, 745-757.

10. G.Z.Chen, E.Gordo and D.J.Fray, “Direct Electrolytic preparation of chromium powder”, Metallurgical and Materials Transactions, 35B, 2 (2004) 223-33.

11. P.K.Tripathy, M.Gauthier and D.J.Fray, “Electrochemical de-oxidation of titanium foam in molten calcium chloride”, Metallurgical and Materials Transactions, 38B (2007) 893-900.

12. HSC Chemistry software, version 7.00, Outokumpu Research, Oy, Pori, Finland, 2009.

13. X.Y. Yan and D.J. Fray, "Electrosynthesis of NbTi and Nb3Sn superconductors from oxide precursors in CaCl2- based melts", Advanced Functional Materials, 15(11), 1757-1761 (2005).

14. X.Y. Yan and D.J. Fray, "Electrochemical studies on reduction of solid Nb2O5 in molten CaCl2-NaCl eutectic - I. Factors affecting electrodeoxidation of solid Nb2O5 to niobium" Journal of The Electrochemical Society, 152 (1): D12-D21 (2005)

11

15. X.Y.Yan and Derek J Fray, "Electrochemical studies on reduction of solid Nb2O5 in molten CaCl2-NaCl eutectic - II. Cathodic processes in electrodeoxidation of solid Nb2O5" Journal of The Electrochemical Society, 152 (10): E308-E318 (2005)

16. A. Cox and D.J Fray, "Production of Aluminium, Magnesium and Aluminium-Magnesium Alloys by direct electrochemical reduction of their solid oxides", Molten Salts XIII, The Electrochemical Society, PV2002-19, 745-757, (2002)

17. P.K.Tripathy and D.J.Fray, Unpublished results.