Effect of Cold Swaging and Heat Treatment on Properties of the P / M 91 W6 NI-3 Co Heavy Alloy

This paper describes the effects of rotary swaging and heat treatment under different conditions on mechanical properties and microstructure of the P/M 91W-6Ni-3Co alloy. The investigation was performed on cold rotary swaging rods with 10, 20 and 25% reduction in area, heat treated at 1130C for 90min under different cooling conditions and strain-aged at temperatures from 300 to 800C for 60, 180 and 360min. Tensile and toughness testing of rotary swaging rods showed that increase in the reduction resulted in intensive strengthening and toughness lowering. The results of tensile and toughness testing of heat treated rods have shown a decrease in strength, ductility and partly toughness in comparison with the mechanical properties of swaged alloys. The results of strain aged rods demonstrate that, for this cold rotary swaging alloy, there is a temperature range from 500 to 600C for 60min in which the maximum ultimate strength and hardness may be attained.


Introduction
The mechanical properties of W-Ni-Co alloys are higher than in the case of W-Ni-Fe and W-Ni-Fe-Co alloys.For extremely demanding applications, even higher mechanical properties are obtainable from the W-Ni-Co system with nickel to cobalt ratios ranging from 2 to 9 [1][2][3][4].The cooling rate from the temperature of liquid-phase sintering influences the formation of brittle intermetallic phases (Co 3 W and Co 7 W 6 , Ni 4 W, NiW, NiW 2 ) and carbides (M 6 C, M 12 C) [1,3,5,6].The high temperature heat treatment of heavy alloys has the following advantages: removal of dissolved hydrogen from the matrix phase [1][2][3][4][5][6], increase of homogeneity in the alloy chemical composition [1,[7][8][9][10][11][12] and change of the γ-phase structure [3,6].The purpose of this treatment is to improve mechanical properties.The heat treatment usually includes rapid cooling in water, oil or argon, thus preventing impurity segregation on the W/W, W/γ and γ/γ interfaces, as well as causing a decrease or a complete removal of composition inhomogenity and the prevention of intermetallic compound precipitation on the interface and grain boundaries [3,4,[6][7][8][9][10][11][12][13][14][15].A more successful approach to achieve higher levels of strength and hardness involves cold worked alloys up to 40% reduction in area and annealing at 350 0 C -550 0 C for 60 to 120 min.This behavior of heavy alloys is caused by strengthening with a W and nickel-based solid solution (γ-phase).Besides, the effect of the W-phase on strain ageing is predominant in comparison with the γ-phase [2][3][4][5]16,17].The gradual softening of these alloys, after ageing over 700 0 C, is caused by a synergistic effect on softening of both structure constituents provoked by recrystallization of the γ-phase and partial softening or recovery of the W-phase [2][3][4][5]16,17].
The aim of this study was to examine the effects of rotary swaging, high temperature heat treatment (annealing and cooling in water or oil) and ageing treatment on the structure and mechanical properties of the 91W-6Ni-3Co tungsten heavy alloy.

Experimental
Tungsten heavy 91W-6Ni-3Co alloy was prepared by mixing elemental powders of W, Ni and Co.The physical properties and chemical composition of powders are presented in Tab.I.

Tab. I. Chemical compositions of examined alloy (in mass%)
W Ni Co Fe Cu Pb Cr Al Mn Ca Mg K 90.03 6.82 2.97 0.064 0.011 0.004 0.017 0.002 0.001 0.07 0.004 0.006 The mixed powders were isostatically cold pressed at 200 MPa and liquid-phase sintered at 1480 o C for 1h as cylindrical rods.After heat treatment at 1155 o C for 2h and water cooling rods were machined to round specimens of 39 mm in diameter.The chemical composition was determined by an atomic absorption spectrophotometer (AAS) type 1100, "Perkin Elmer" (USA).Carbon, sulphur and oxygen contents were determined by a Lecco apparatus.The chemical composition of the alloy is presented in Tab.II.

Tab. II. Chemical compositions of examined alloy (in mass. %)
W Ni Co Fe Cu Pb Cr Al Mn Ca Mg K 90.03 6.82 2.97 0.064 0.011 0.004 0.017 0.002 0.001 0.07 0.004 0.006 Specimens were cold rotary swaged on vertical GFM equipment (Austria) with four hammers.All rods were preheated from 450 to 500 0 C in a heat-resistant "Ipsen" air furnace.The rotary swaging step was 70 mm/min.The hammer stroke speed was 140 strokes/min.The achieved compression strains were 10, 20 and 25%.The swaged rods were heat treated in air in a heat-resistant furnace.The furnace temperature was automatically controlled in a range of ± 5K.
The first groups of forged samples were annealed under the following conditions: heating at 1130 0 C for 90min followed by water cooling and (or) oil.The second groups of forged samples were strain-aged under the following conditions: heating between 300 and 800 0 C (heating time was between 60 and 360 min) that was followed by oil cooling.The heating rate was ≈ 10 0 C/min.The oil temperature for cooling was set at 20 0 C.
The heavy alloy density was measured according to the JUS C.A2.026 Standard.Round specimens with a gage length of 30mm were prepared for the tensile test by machining according to the JUS EN 10002-1 Standard.The tensile tests were carried out on an Instron M8032 tensile machine of 200kN capacity, with a tension rate of 1.2x10 -5 m/s.The impact energy was determined according to the JUS EN 10045-1 Standard, on a Tinius Olsen Charpy pendulum of 348 Nm equipped with a Teletronix 51-13 oscilloscope.The dimensions of the unnotched specimens were 10x10x55mm.The Vickers hardness HV30 was measured on the rod cross-sections by Wolpert-Dia testor 2Rc -type equipment.The microstructure of heavy alloy was analyzed by a Leitz (OM) optical microscope at the magnification of 150 and 300 times and a Philips scanning electron microscope SEM-515 (SEM) with an energy dispersion spectroscope (EDS).
The samples were prepared by mechanical grinding and polishing.The structure was revealed by etching in a solution consisting of 5 cm 3 of FeCl 3 molar solution, 20 cm 3 concentrated HCl and 50 cm 3 distilled water.The microstructual parameters of the average W phase grain size and the volume fraction ratio of the phases were determined by automatic and semiautomatic quantitative area method by a TAS PLUS image analyzer on 100 consecutive measuring areas.The chemical compositions of the phases were analyzed by EDS.

Results and Discussion
Results of density measurements have shown that there was no change in material density after the plastic deformation, i.e. the alloys density after deformation (10, 20 and 25%) remained the same as it was in the sintered condition -17.6 g/cm 3 .
The properties of the heavy 91W-6Ni-3Co alloy for different conditions: sintered and heat treated (Sint/HT), rotary swaged with different deformation degree (RS1, RS2, RS3), heat treated with different cooling rates (HT1 and HT2) and strain-aged with different deformation degrees, are presented in Fig. 1 and 2. Mechanical properties such as tensile strength -Rm, yield strength -Rp 0.2 , elongation -A and area reduction -Z are given in Fig. 1.For the same conditions, the results of the impact energy -K and hardness -HV are presented in Fig. 2. From these results it is evident that strength and hardness increase, while plasticity and impact energy decrease with increasing of the rotary swaging deformation degree (RS1, RS2, RS3).  2 Effects of different conditions of the alloy on the impact energy (K) and hardness (HV); Sint/HT -sintered and heat-treated; RS -rotary swaged; HT1 and HT2 rotary swaged and heat treated; SA -strain age; CO -cooled in oil; CW -cooled in water.
Hardness variations within cross sections of the rods for different deformation degrees are given in Fig. 3.It is evident that hardness increases with the higher degree of deformation (ε) and hardness from the central to the peripheral zone of the rod increases.Whatsoever, hardness in the peripheral zone of the rod is higher than in the central zone (Fig. 3).This behaviour of the swaged rods is probably caused by higher deformation energy accumulation in the peripheral zone than in the central zone rods.The micrographs of sintered-heat treated and swaged (after ε=25%) samples are shown in Fig. 4. Quantitative metallographic investigations of sintered-heated samples have shown that the mean grain size of tungsten was 15-20 μm, and the volume fraction of the W phase was 70-73%, while for the γ phase it was 27-30%.There are no significant changes in the W phase grain size or in the volume fraction ratio of W and γ phases as an effect of heat treatment and plastic deformation.After swaging the mean distance between W-grains was reduced, along with an obvious elongation of W-grains (4b).The nonhomogeneous hardness distribution in the cross section of a swaged rod is due to nonhomogeneous deformation Mechanical properties of the swaged heavy W-Ni-Co alloy depend essentially on hardening in both structural constituents, tungsten grains (W-phase) and γ-phase [2,3].The nonhomogeneous hardness distribution in the cross section of a swaged rod is due to a nonhomogeneous deformation structure.Deformation of individual grains is obviously lower in most parts of the cross section of a swaged rod (Fig. 5b) than in the peripheral zone (Fig. 5a), because the shape of grains was not significantly changed during reduction.A comparison of Figs.4a and 4b clearly shows that the mean distance between grains of the Wphase is reduced during reduction (the average thickness of the γ-phase between W-grains is reduced from 10.1 μm to 5.2 μm), which was also reported earlier [2,3].
Regression analysis of variations of hardness as a function of applied strain was carried out using a polynomial function of the third order (Tab.III and Fig. 3), which showed the best fit with the increase of the deformation degree from 10% to 25%.The mechanical properties of the heat treated alloy at 1130 0 C/90 min.followed by oil or water cooling (HT1 and HT2) in are presented Fig. 1 and 2. It is evident that strength, hardness, plasticity and partly impact energy of swaged samples (ε=20%) decrease as a result of heat treatment.The cooling rate effect on mechanical properties is not the same for these samples.The results show that, after heat treating samples at 1130 0 C/90 min.followed by oil (HT1), higher strength (Rm=1180 MPa) compared with the water cooled (HT2) samples (Rm=1158 MPa) is obtained.On the other hand, heat treated samples at 1130 0 C/90 min.followed by water cooling, showed higher impact energy.The micrographs of these heat treated samples for different cooling conditions are shown in Fig 6 a and b for cooling in oil and in water, respectively.

Tab. III
In the oil cooled samples, local composition inhomogeneity in the γphase is presented in Fig. 6c and it was not observed in the water cooled samples (Fig. 6b).The results of the mean EDS analysis of the precipitations on the γ/γ grain boundaries and in the γ grains (from Fig. 6c) are given in Fig. 6d.It can be seen that the precipitation has high tungsten content.
It is already known that heat treatment promotes chemical homogeneity, especially in the γ phase [3,6,9,12,13] and, as a consequence, affects mechanical properties [6,12,14,15].The results of the EDS analysis of the samples cooled in oil (Fig. 6d), reveal the composition inhomogeneity of tungsten and nickel in the γ phase as a result of precipitation [5,6,[9][10][11][12][13][14][15].The cooling effect on impact energy is not the same for these swaged (ε=20%) and heat treated samples (Fig. 2).As a result of cooling conditions after annealing at 1130 0 C for 90 min, the impact energy of the samples cooled in water is higher (116 J) than in oil (37 J).A possible reason for this phenomenon can be the presence of local inhomogeneity in the γ phase in the samples cooled in oil (Fig. 6c and d).This is also implied by higher impact energy in the samples cooled in water, where significant composition inhomogeneity was not found (Fig. 6b) Generally speaking, during heat treatment of the swaged tungsten heavy 91W-6Ni-3Co alloy (degree of deformation ε=20%) strength, hardness, plasticity and partly impact energy decrease, in comparison to the previous swaged state (Fig. 1 and Fig. 2).This could be explained as a result of the partial softening of the W-phase and complete recrystallization of the γ-phase [2][3][4]10].A comparison of Figs.6a and 6b clearly shows that there are no significant changes in the microstructure of the swaged and different heat treated samples.
The mechanical properties of the rotary swaged alloy samples, for different deformation degrees and strain-aged conditions (SA1, SA2, SA3 and SA4) are presented in Fig. 1 and 2. It is evident that the strength and hardness of strain-aged samples increase, while plasticity and the impact energy decrease as an effect of ageing at 550 0 C/60min (SA2, SA3 and SA4) in comparison with the cold rotary swaged samples (RS1, RS2 and RS3).It is also evident that the strength and hardness of the strain-aged samples at 300 0 C/180min.(SA1) slightly increase and its plasticity and the impact energy decrease compared to only rotary swaged sample RS2.Fig. 7 shows the influence of strain-ageing temperature (300 0 C to 800 0 C) for the ageing time of 60min on hardness.It is evident that after ageing of the swaged samples (ε=20% and ε=25%), the hardness of the samples has reached a maximum value, which is in the temperature range from 500 to 600 0 C and that a higher compression strain gave a higher hardness level.On the other hand, for the sintered samples there was no change in hardness after ageing.
Generally speaking, during strain ageing of tungsten heavy 91W-6Ni-3Co alloy strength and hardness increase while the ductility and toughness decreases, in comparison to the previous swaged state (Fig. 1 and 2).The effect of strengthening could be the result of a synergic effect of hardening of both structural constituents, tungsten grains (W-phase) and the γ-phase.The behavior of W and γ-phases is in accordance with the results of Baosheng [4] and Katavic [17], which showed greater influence of W-phase on hardness of the strain aged 90W-7Ni-3Fe and 92.4W-4.9Ni-2.5Fe-0.25Coheavy alloys in comparison with γ-phase.The observed predominant effect of the W-phase on strain aging phenomena could be explained as a result of the carbon and/or nitrogen primary interactions with dislocations in the W-phase as presented in literature [18][19][20][21][22].It could be concluded that strain aging of tungsten heavy 91W-6Ni-3Co alloy is a consequence of the same effect as the diffusion of carbon and/or nitrogen interstitially solute atoms on the dislocations and blocks them with the formation of Cottrell atmospheres [4,16,[18][19][20].Gradual softening of these alloys, after ageing over 600 0 C (700 0 C), could be caused by effects of recovery and further recrystallization of the γ-phase and recovery of the W-phase [1,[3][4][5][6][7].
Regression analysis of hardness change in the strain aged samples was described by the polynomial fit function of third order (Tablе IV), which showed a higher degree of correlation.

Tab. IV
Regression equations of hardness variations of the strain aged alloy 91W-6Ni-3Co Hardness, HV30 The ageing time effect on hardness at constant temperatures of 300 0 C and 500 0 C was examined for the swaged samples and is presented in Fig. 8 and 9.

RS, ε=20%
Hardness, HV30 Distance from rod axis, mm center surface Fig. 8. Hardness distribution in the cross-section of the swaged after ε=20% and strain-aged rod at the constant temperature of 300 0 C for different times (60 min., 180 min.and 360 min.).

RS, ε=20%
Hardness, HV30 Distance from rod axis, mm surface center Fig. 9. Hardness distribution in the cross-section of the swaged after ε=20% and strain-aged rod at the constant temperature of 500 0 C for different times (60 min., 180 min.and 360 min.).
The results of this examination show the non-homogenous hardness distribution in a sample cross section, with always harder surface layers and that the hardness of samples increases as an effect of ageing time only at 300 0 C.Besides that, hardness in the peripheral zone of the rod is higher than in the central zone at 300 0 C/180min., 300 0 C/360min and 500 0 C/60min.The ageing time effects on hardness at constant temperatures of 300 and 500 0 C could be explained in a similar manner, according to the theory from Cottrell and Bilby [4,[16][17][18][19][20]. Namely, the rate of the number of interstitial atoms moving towards dislocations in the material volume unit at a constant temperature depends on the aging time.At a constant aging temperature the increase in concentration of carbon and nitrogen interstitial atoms around dislocations is a consequence of increased aging time.Owing to heterogeneity of the deformation structure, the interstitial solute atoms diffuse to the region with a higher concentration of dislocations during longer annealing periods.A higher saturation of Cottrell atmospheres with solute atoms causes a blocking of a considerable number of dislocations.
Regression analysis of hardness distribution of the alloy as a function of the ageing time at constant temperatures of 300 0 C and 500 0 C were carried out using a polynomial function of the third order (Tab.V), which showed a higher degree of correlation.
Tab. V Regression equations of hardness distribution in a rod cross section of the alloy 91W-6Ni-3Co swaged (ε=20%) and strain aged at constant temperatures of 300 0 C and 500 0 C τ a , min.

Standard deviation SD
Correlation coefficient r 2 HV 300 = 446.492+4.821x-0.744x 2 +0.039 x 3  1.987 0.953 60 HV 500 = 495.640+11.884x-1.606 x 2 +0.078 x 3 2.213 0.984 HV 300 = 437.943+13.224x-1.922 x 2 +0.093 x 3  2.786 0.975 180 HV 500 = 510.852+6.610x-0.843 x 2 +0.041 x 3  1.572 0.981 HV 300 = 442.382+13.365x-1.895 x 2 +0.091 x 3  2.442 0.979 360 HV 500 = 506.162+9.989x-1.355 x 2 +0.063 x 3  1.731 0.979 * τ a -the ageing time; HV 300 and HV 500 -hardness distribution in rod cross section rotary swaged and strain aged at 300 0 C and 500 0 C. The ageing temperature effect (300 0 C to 800 0 C) on hardness distribution in a rod cross section was examined for the swaged samples (20% and 25%) at constant time of 60min.and is presented in Fig. 10 and 11.The results of this examination showed that the most homogenous hardness distribution and hardness maximum in the sample swaged after a 20% deformation degree after strain aged at 600 0 C (Fig. 10).Hardness of the samples swaged after a 25% deformation degree is higher after it was strain aged in the temperature range of 600 0 C/60min.to 800 0 C/60min.(Fig. 11), than in the samples swaged after 20%.Besides that, there was no expressive peak in the tested temperature range and it can be assumed that the hardness decreases at higher temperatures.The ageing temperature effects on hardness at a constant time could be explained in a similar manner.Namely, in the case of a high compression strain (25%), the concentrations of the interstitial solute atoms in the temperature range of 600 0 C/60min.to 800 0 C/60min.are sufficient and they could cause a considerable effect of dislocation blocking.
Tab. VI Regression equations of hardness distribution of the alloy 91W-6Ni-3 Co swaged (ε=20% and 25%) and strain aged in the temperature range from 300 0 C to 800 0 C at constant time of 60min.

RS,ε=25%
Hardness, HV30 Distance from rod axis, mm surface center Regression analysis of the hardness distribution in the strain aged samples' cross section was described by polynomial fit functions of third order (Table VI), which showed a high degree of correlation.
Microphotographs of the strain-aged samples are shown in Fig. 12. Quantitative metallographic investigations of the strain aged samples have shown that the average grain size of tungsten was 16-20 μm, and the volume fraction of the W-phase was 71-73% while for the γ-phase it was 27-29%.There are no significant changes in the W-phase grain size in comparison with the swaged samples (Fig. 4b).

Conclusions
Effects of cold swaging and heat treatment on properties of the P/M 91W-6Ni-3Co alloy, have been presented.
During rotary swaging, strength intensively increases (Rm of 20-36 %, Rp 0.2 of 40-57 %) while ductility and toughness were found to be quite reduced (A of 27-86 %, Z of 3-92 % and K of 20-70%) compared to the initial sintered condition.A high gradient of hardness in the rod cross section can be noticed as well.The mean distance between grains of W-phase is reduced from 10.1 μm in the sintered alloy to 5.2 μm after deformation.
During high temperature heat treatment of tungsten heavy 91W-6Ni-3Co alloy strength, hardness and plasticity decrease (Rm of 5-7 %, Rp 0.2 of 17-18%, A of 22-26%, Z of 51-59 %), in comparison to the swaged alloy.This could be explained as a result of partial softening of the W-phase and complete recrystallization of the γ-phase.
During strain ageing of tungsten 91W-6Ni-3Co heavy alloy, strength and hardness increase (Rm of 8-15 %, Rp 0.2 to 21 %, HV of 9-24 %) while the ductility and toughness decrease (A of 18-55%, Z of 43-66 % and K of 30-80 %) in comparison with the forged condition.There is a strain-ageing region from 500 to 600 0 C for 60min and at 300 0 C for 180min when the alloy has maximum strength and hardness.It is assumed that the strengthening effect could be a result of the synergic effect of hardening of both structural constituents, tungsten grains (W-phase) and the γ-phase.

Fig. 1
Fig.1Effects of different conditions of the alloy on the tensile strength (R m ), yield strength (R p0.2 ), elongation (A) and reduction (Z); Sint/HT -sintered and heat-treated; RS -rotary swaged; HT1 and HT2 rotary swaged and heat treated; SA -strain age; CO -cooled in oil; CW -cooled in water.

Fig. 3
Fig.3 Hardness distribution in the rod cross section of the sintered-heat treated and swaged rods for different deformation degrees; Sint/HT -sintered and heat-treated; RS -rotary swaged; ε -degree of deformation; CW -cooled in water.

Fig. 6
Fig. 6 Axial section microstructure of a swaged rod after ε=20% and heat treatment: a) heat treated at 1130 0 C/90 min/cooled in oil; b) heat treated at 1130 0 C/90 min/cooled in water; c) precipitates on the γ/γ grain boundaries and in the γ grains of oil cooled samples; d) EDS analysis of mean chemical compositions of precipitates

Fig. 7 .
Fig. 7. Hardness change with deformation degree and ageing temperature at a constant aging time of 60 min.(3.6 ks).

Fig. 10 .Fig. 11 .
Fig.10.Hardness distribution in the cross-section of the sinteredheat treated, swaged after ε=20% and strain-aged rod at the different temperatures (300 0 C to 800 0 C) for the constant time of 60 min.