Journal of Materials Synthesis and Processing, Vol. 7, No. 4, 1999

TiC/Ni3Al Composites Manufactured by Self-Propagating High-Temperature Synthesis and Hot Isostatic Pressing J. Keskinen,1 J. Maunu,1 P. Lintula,1 M. Heinonen,2 and P. Ruuskanen1

TiC-20 wt% Ni3 Al and TiC-40 wt% Ni 3 Al composite materials were produced by self-propagating high-temperature synthesis (SHS) and hot isostatic pressing (HIP). In the SHS method the reacted powders were compacted by uniaxial pressing immediately after the reaction. The microstructure of the materials produced by SHS consisted of spherical carbides embedded in the Ni 3 Al matrix, whereas the microstructure of the materials produced by HIPing was more irregular. A maximum hardness of 2010 HV| was measured for the material produced by HIP and a maximum fracture toughness of 10.5 MPa m 1 / 2 was measured for materials produced by SHS. High-temperature resistance was investigated by exposing the materials to 800oC in air for 110 h. The results obtained showed that the TiC + Ni 3 Al composite materials can be recommended for use in environments consisting of oxidizing atmosphere at temperatures around 800oC where high wear resistance is required. KEY WORDS: Self-propagating high-temperature synthesis; hot isostatic pressing; intermetallic matrix composites; utanium carbide; nickel aluminide.

1. INTRODUCTION

be carried out soon after the reaction when the temperature is still high enough to allow plastic deformation. Titanium carbide can be produced by the SHS method due to the highly exothermal reaction between Ti and C (DH = -184.6 kJ/mol [1]). Titanium carbide exhibits excellent wear resistance and a high hardness (about 3000 HV). Also, its resistance to corrosion by hot concentrated acids is very good. When TiC is heated a protective TiO2 layer forms on the material surface. This layer exhibits good stability in high-temperature oxidizing atmospheres. The density of TiC is 4.91 g/cm3. When combined with low-density aluminides, TiC/aluminide composites can offer material for industrial use which is of a lower weight than conventional cemented carbides. When elemental metal powders, for instance, Ni and Mo, are added to the powder mixture before the SHS reaction, a composite material with a hard carbide phase embedded in a softer matrix can be produced [2]. The SHS method can also be used to achieve composite materials having different intermetallics as a matrix com-

Self-propagating high-temperature synthesis (SHS), also called combustion synthesis, is a materials synthesis method in which both the material and the product can be made simultaneously. Combustion synthesis is based on highly exothermic chemical reactions between starting material powders which are mixed in the desired compositions. The mixture is locally heated to an ignition temperature at which the exothermic reaction starts. After ignition, the reaction propagates through the prepressed powder mixture without any external energy. Generally, the speed of the reaction front is in the range of 0.1 to 10 mm/s. The reacted material is usually highly porous following the reaction, but it can be densified to almost full density by simple uniaxial pressing. Such pressing must

1VTT

Manufacturing Technology, P.O. Box 17031. FIN-33101 Tampere, Finland. 2Tampere University of Technology, Institute of Materials Science. P.O. Box 589, FIN-33101 Tampere, Finland.

253 1064-7562 99/0700-0253$16.00/0 © 1999 Plenum Publishing Corporation

254

Keskinen, Maunu, Lintula, Heinonen, and Ruuskanen

ponent. In recent years the level of interest in intermetallic matrix composites has increased because methods to increase the ductility of intermetallics have been developed. Subratnanian et al [3] have produced TiC/Fe-40 at % Al composites by pressureless melt infiltration. Tiegs et al. [4] have made TiC/Ni3Al and WC/Ni3Al composites by not pressing mixed Ni3Al and carbide powders. Ohriner et al. [5] have produced Al2O3/Ni3Al by a variety of methods. Production techniques and possible combinations of material are discussed by WardClose et al. in a review article [6]. The motivation behind this study was the development of a new type of composite material consisting of TiC and Ni3Al. Nickel aluminide (Ni3Al) is an intermetallic compound of interest for structural applications, especially at elevated temperatures [7]. Its strength behavior is anomalous, the yield strength increasing as the temperature rises up to about 800°C. Nickel aluminide also has good oxidation and carburizing resistance up to 1100°C. Conventional WC/Co cannot be used at high temperatures because its oxidation rate increases rapidly at temperatures above 600°C [8]. TiC/Ni3Al composites can also be expected to be able to withstand wear in aggressive environments because of the excellent wear resistance exhibited by Ni 3 Al at high temperatures [9].

2. EXPERIMENTAL In the case of producing TiC/Ni3Al composites by SHS, elemental titanium (particle size, <50 um), carbon, and commercial gas atomized nickel aluminide powder (20-40 um) were used as starting materials. The composition of the Ni 3 Al powder used was IC-50 (Ni, 88.3 wt%; Al, 11.1 wt%; Zr, 0.6 wt%; and B, 0.02 wt%). The

starting materials were mixed in a ball mill. The powder mixture was packed into a mold (diameter 70 mm) using a pressure of 5 MPa. The mold was filled with about 400 g of the powder. Immediately following the SHS reaction between titanium and carbon, the pellet was densified using a pressure of 25 MPa. In HIPing experiments the starting materials were TiC powder (sieved to be below 63 um) produced by the SHS method at VTT and the already described commercial IC-50 Ni 3 Al powder. In the HIPing process the temperature was kept at 1200°C and the pressure at 170 MPa for 3 h. Some experiments were also performed with nickel aluminide that had been produced at VTT by mechanical alloying followed by heat treatment to produce the Ni3Al phase. In this case the process parameters were 1180°C, 100 MPa, and 3 h. The compositions and processing methods are summarized in Table I. The microstructure of the materials was investigated by scanning electron microscopy and energydispersive spectroscopy (EDS) and the crystal structure of the materials was defined by X-ray diffraction using MoKa, radiation. Vickers hardnesses were measured using a weight of 1 kg. The results reported are the averages of 20 measurements. Fracture toughness values (KIC) were obtained by the Vickers indentation fracture toughness method. Earlier studies carried out in our laboratory have established that Vickers indentation fracture toughness values correspond well with SENB (single edge notch beam) for TiC-NiMo with nickel and molybdenum contents of up to 25% [10], especially when the equation proposed by Anstis et al. [11] was applied. High-temperature resistance was investigated by exposing the materials in a furnace to an air atmosphere at 800°C for 110 h. The TiC/Ni3 Al samples were compared to commercially available WC/Co (6 wt% cobalt). The WC grain size was measured to be 2-4 um.

Table I. Compositions and Coding of TiC/Ni 3 Al Samples

Composition 1 TiC-20 wt% Ni3Al

2 3 4 5 6 7 8

TiC0.8-20 wt% Ni 3 Al TiC-40 wt% Ni3Al TiC 0.8 -40 wt% Ni3Al TiC-20 wt% Ni3Al TiC-40 wt% Ni3Al TiC-20 wt% Ni3Al TiC-40 wt% Ni3Al

Process SHS, gas atomized Ni 3 Al SHS, gas atomized Ni 3 Al SHS, gas atomized Ni3Al SHS, gas atomized Ni 3 Al HIP, gas atomized Ni 3 Al HIP, gas atomized Ni 3 Al HIP, mechanically alloyed Ni 3 Al HIP, mechanically alloyed Ni 3 Al

Sample code SHS-TiC-20Ni3Al SHS-TiC0.8-20Ni3Al SHS-TiC-40Ni3Al SHS-TiC0.8-40Ni3Al HIP, GA-TiC-20Ni3Al HIP, GA-TiC-40Ni 3 Al HIP, MA-TiC-20Ni3Al HIP, MA-TiC-40Ni3Al

TiC/Ni3Al by SHS and HIP

255

3. RESULTS AND DISCUSSION Table I shows all the TiC/Ni3Al compositions and production techniques used in this work. In the case of SHS, as well as changing the amount of Ni 3 Al, the Tito-C ratio was also varied. The quantity of the Ni 3 Al matrix was 20 and 40 wt% and the carbide stoichiometries were TiC and TiC0.8. This was chosen to use all the carbon in the reaction and prevent any reaction between carbon and the matrix. Basically TiC and TiC0.8 are of the same phase, which is the case for relatively wide range of carbon concentrations. Thus, the X-ray diffraction curves measured for TiC and TiC0.8 have the same peaks, although the peaks are slightly shifted when the carbon concentration is varied. 3.1. Microstructure Figure 1 shows the typical microstructure of materials produced by the SHS technique. As is usually the case when SHS is used, carbide particles are spherical, which improves the fracture toughness compared to that of materials having sharp, edgy carbide particles. An irregular particle shape can induce cracks from the sharp particle edges because of the increased stress in these points. Figure la presents the microstructure of sample SHS-TiC0.8-20Ni3Al, and Fig. 1b that of sample SHS-TiC0.8-40Ni3Al. The most significant difference between these two samples is the size of the TiC particles. In the case of 20 wt% binder material, the size of the TiC particles was of the order of 10-20 um; with the 40 wt% binder material the size decreased to below 5 um. This type of particle size dependence on the quantity of matrix material is typical of SHS materials because lower reaction temperatures are reached when the ratio of Ti-to-C binder material is lower. As the micrographs show, there are some pores in the samples. Samples SHS-TiC-20Ni3Al and SHS-TiC-40Ni3Al were somewhat more porous than samples SHS-TiC0.8-20Ni3Al and SHS-TiC0.8-40Ni3Al. Thus, from a porosity point of view, using a Ti-to-C ratio of 1:0.8 is better than using a ratio of 1:1. Similar structures have been obtained by Terry et al. and Wood et al. for TiC-Fe and (WTi)C-Fe composites [12-14]. The TiC/Ni3Al samples produced by hot isostatic pressing were completely dense. Figure 2 presents the microstructure of the HIP sample, MA-TiC-20Ni3Al (Table I). The powders were mixed in a ball mill before HIPing. According to the EDS investigations, the lightest areas contain primarily Ni and Al and are thus the

Fig. 1. Scanning electron micrographs of SHS samples: (a) SHSTiC0.8-20Ni3Al); (b) SHS-TiC0.8-40Ni3Al.

intermetallic material Ni3Al; the darkest areas consist of TiC. In the gray areas some nickel and aluminium among the titanium can be detected. The microstructure of the gray areas is thus very fine. The larger Ni3 Al grains (light areas) are also relatively small, 10-20 um, and irregular in shape. The maximum TiC grain size is about 5 um. The microstructure of the other HIPed samples was similar to that of this sample, the only significant difference being the relative surface area covered by TiC and Ni3 Al. Based on the X-ray diffraction studies, the main phases after the SHS reaction were TiC and Ni3Al in all cases. Figure 3 presents the XRD curve of sample SHS-TiC-40Ni3Al. The X-ray diffraction peaks belonging to TiC were clear in all samples and dominant in samples having only 20 wt% Ni 3 Al. In the materials SHS-TiC-20Ni3Al, SHS-TiC0.8-20Ni3Al, and SHS-TiC-40Ni3Al, the Ni 3 Al diffraction peak intensities were found to change when the measurement direction was changed. The reason for this is that there the

256

Keskinen, Maunu, Lintula, Heinonen, and Ruuskanen

Fig. 2. Scanning electron micrograph of HIPed sample HIP, MATiC-20Ni3Al. Fig. 4. XRD curve of sample HIP, MA-TiC-20Ni3Al.

Ni3Al matrix has a texture. The peaks belonging to Ni3Al are clearly broader than those belonging to TiC. This can be the result of strain or small crystal size. Also, the peaks of Ni3(Al, Ti)C and Ni3AlC0.5 overlap with those of Ni3Al, which makes the process of interpretation more difficult. The texture makes it impossible to define the quantity of the Ni3Al phase compared to that of Ni3Al-based carbides. Because of the differences in the quantities of these carbides, it was not possible to determine what kind of texture the samples actually possessed. The TiC phase did not exhibit any texture.

Figure 4 shows an example of the X-ray diffraction pattern measured from a HIPed sample. The only visible peaks are those of TiC and Ni3Al; there are no traces of any other phases or broadening of the peaks. This was the case in all the HIP experiments. This also shows that significant reactions between TiC and Ni 3 Al have not occurred during the compaction process. An important fact is that there appears not to be any Ni3AlC0.5 or other Ni3 Al-based carbides. The peak intensity ratios of Ni3 Al were the same as those measured for the intermetallic Ni3Al powder before it was mixed with TiC. The curves obtained did not show any dependence on the direction of measurement. This confirms that there is no texture as in the samples produced using the SHS technique. 3.2. Mechanical Properties

Fig. 3. XRD curve of sample SHS-TiC-40Ni3Al. The Ka2 peaks are not stripped.

Numerical values obtained are presented in Table II. Hardness values obtained from samples produced by HIP were higher than those obtained from samples produced by SHS. The maximum HV 1 value, 2010, was obtained from sample HIP,MA-TiC-20Ni3Al, containing 20 wt% Ni3Al. The hardness of sample HIP,MA-TiC-40Ni3Al was 1590 HV 1 . In these samples the TiC and Ni3Al powder mixtures were ball milled for 5 h before HIPing. The maximum hardness value for the SHS material, 1580 HV1, was obtained from sample SHS-TiC-20Ni3Al. Various Ti-to-C ratios do not appear to have any clear influence on the hardness values obtained. The lowest hardness value, 1160

257

TiC/Ni3Al by SHS and HIP Table II. Mechanical Properties of the Samples

Sample code SHS-TiC-20Ni3Al SHS-TiC0.8-20Ni3Al SHS-TiC-40Ni3Al SHS-TiC0.8-40Ni3Al HIP, GA-TiC-20Ni3Al HIP, GA-TiC-40Ni 3 Al HIP, MA-TiC-20Ni3Al HIP, MA-TiC-40Ni3Al

Hardness (HV 1 )

1580 1550 1160 1140 1890 1220 2010 1590

Fracture toughness (MPa m 1 / 2 )

6.2 10.5

8.3 4.4 6.8

MPa m 1/2 . Sample HIP,MA-TiC-20Ni3Al had a fracture toughness of 4.4 MPa m 1 / 2 . The better fracture toughness obtained using SHS was probably due to the more spherical shape of the TiC particles. Another reason is that, in samples made using HIP, there are relatively large areas of Ni3Al, which correspondingly results in lower Ni3Al concentrations elsewhere in the material. In material made using SHS the TiC particles are separated equally by the Ni3Al. 3.3. High-Temperature Properties

and 1140 HV 1 , were obtained for the SHS samples SHS-TiC-40Ni3Al and SHS-TiC0.8-40Ni3Al, respectively. Obviously increasing the binder content decreased the hardness of the samples. Thus, samples manufactured by ball milling and HIPing showed higher hardness values than those made by SHS. The reason for this is most probably the difference in microstructure. In the HIPed material, the matrix phase (Ni3Al) is not evenly distributed, and this means that most of the TiC particles are closer to each other than in samples made by SHS. Another reason for the better hardness value of the HIPed samples is that HIPing made the material completely dense which increased the hardness. KIc values shown in Table II were calculated according to an equation of Anstis et al. [11]:

where E is Young's modulus, H is the hardness, P is the load used, and c is the length from the middle of the indentation to the end of the crack. Young's modulus was calculated using the rule of mixtures. Fracture toughness values obtained from SHS materials were higher than those obtained from HIPed materials. As expected, the harder the samples, the lower their fracture toughness. Fracture toughness values were not defined for SHS samples SHS-TiC-20Ni3Al and SHS-TiCMONi3Al because their porosity made it difficult to measure the exact length of the propagated cracks. Sample SHS-TiC0.8-40Ni3Al had the highest fracture toughness, 10.5 MPa m 1 / 2 . When the Ni3Al concentration was decreased to 20 wt% (sample SHS-TiC0.8-20Ni3Al), the fracture toughness fell to 6.2 MPa m1/2. The effect of the amount of matrix material was similar in all the HIPed samples. The fracture toughness of sample HIP,GA-TiC-40Ni3Al was 8.3

The high-temperature endurance of the SHSed and HIPed materials was tested by exposing the samples to air at 800°C for 110 h. The weight gain of samples SHS-TiC0.8-20Ni3Al, SHS-TiC0.8-40Ni3Al, HIP,MA-TiC-20Ni3Al, and commercial WC/Co were measured as a function of time. Figure 5 shows the increase in weight. As can be seen, the oxidation rate of the TiC/Ni3Al composites is essentially lower than that of WC/Co. According to Basu and Sarin [8] the oxidation rate of WC/Co is very low at 600° C but increases very rapidly with temperature. They have investigated the behavior of the material at 600-800°C at various oxygen concentrations and Co contents of the samples. In all cases WO3 and CoWO4 were formed. The relative amount of these oxides was dependent on the Co content of the starting material. According to Voitovich [15], titanium carbide is oxidized slightly up to 800°C but the weight increase is negligible. At 900°C some weight gain was recorded and at 1000°C the oxidation rate increased sharply. At 600-800°C the identified oxide phases were monoclinic modification of TiO, oxycarbide TiCo.42Oo.58, and TiO2 in both the rutile and the anatase modifications. Ni 3 Al is known to produce a dense protective alumina scale when exposed to air [16, 17]. Guo el al. [16] reported that below 950°C, because of slow diffusion, the protective action of the adherent Al2O3 subscale is the primary factor affecting the oxidation of Ni3Al. At higher temperatures NiO and NiAl2O4 would form because of diffusion through alumina. Practically no differences can be seen among the various TiC/Ni3Al composites. The similarity between SHS-TiC0.8-20Ni3Al and HIP,MA-TiC-20Ni3Al is logical because, in principle, their relative areas of carbide and aluminide are almost-identical. As SHS-TiCo.840Ni3Al and SHS-TiC0.8-20Ni3Al had the same rate of weight gain, it can be concluded that the oxidation rates of TiC and Ni3Al are of the same order.

258

Keskinen, Maunu, Lintula, Heinonen, and Ruuskanen

Fig. 5. Weight gain in samples SHS-TiC0.8-20Ni3Al, SHS-TiC0.8-40Ni3Al, HIP,MA-TiC-20Ni3Al, and commercial WC/Co as a function of time.

4. CONCLUSIONS TiC/Ni3Al composite materials were produced by self-propagating high-temperature synthesis (SHS) and hot isostatic pressing (HIP). The concentration of Ni 3 Al was varied between 20 and 40 wt%. The microstructures of the materials made by HIP and SHS were different. In the case of SHS the shape of the TiC particles was smoother and the Ni 3 Al matrix was more evenly distributed. The nonsimilarity of the microstructures meant that the SHS material exhibited better fracture toughness, while higher hardness values were obtained for composites made using HIP. Tests on high-temperature endurance showed that these composites have potential when oxidation resistance at 800°C is required and that they can be used at considerably higher temperatures than WC/Co hard metals.

3. 4.

5.

6. 7. 8. 9. 10. 11. 12.

REFERENCES

13. 14.

1. K. K. Udwadia and J. A. Puszynski, in In Situ Reactions for Synthesis of Composites, Ceramics, and Intermetallics, E. V. Barrera, F. D. S. Marquis, W. E. Frazier, S. G. Fishman, N. N. Thadhani, and Z. A. Munir, eds. (TMS, Warrendale, PA, 1995), p. 59. 2. P. Lintula and P. Ruuskanen, in Proceedings of the 5th European Conference on Advanced Materials and Processes and Applica-

15. 16. 17.

tions. EUROMAT 97, Materials, Functionality & Design, Vol. 1. Metals and Composites, L. A. J. L. Sarton and H. B. Zeedijk, eds. (Netherlands Society for Materials Science, Zwijndrecht, 1997), p. 347. R. Subramanian, J. H. Schneibel, K. B. Alexander, and K. P. Plucknett, Scripta Mat. 35, 583 (1996). T. N. Tiegs. P. A. Menchhofer, K. P. Plucknett, K. B. Alexander, P. F. Becher, and S. B. Waters, in Proceedings, 4th International Conference on Powder Metallurgy in Aerospace, Defense and Demanding Applications, F. H. Froes, ed. (MPIF, Princeton, NJ, 1995), p. 211. E. K. Ohriner, V. K. Sikka, and S. C. Deevi, in Proceedings, 4th International Conference on Powder Metallurgy in Aerospace, Defense and Demanding Applications, F. H. Fores, ed. (MPIF, Princeton, NJ, 1995), p. 197. C. M. Ward-Close, R. Minor, and P. J. Doorbar, Intermelallics 4, 217 (1996). C. T. Liu and K. S. Kumar, JOM 45(5), 38 (1993). S. N. Basu and V. K. Sarin, Mater. Sci. Eng. A209, 206 (1996). S. C. Deevi and V. K. Sikka, Intermetallics 4, 357 (1996). M. Heinonen, M.Sc. dissertation (UMIST, 1996). G. R. Anstis, P. Chanticul, B. R. Lawn, and D. B. Marshall, J. Am. Ceram. Soc. 64, 533 (1981). B. S. Terry and O. S. Chinyamakobvu, J. Mater. Sci. Lett. 10, 628 (1991). B. S. Terry and O. S. Chinyamakobvu, Mater. Sci. Tech. 8, 399 (1992). J. V. Wood, K. Dinsdale, P. Davies, and J. L. F. Kellie, Mater. Sci. Tech. 11, 1315 (1995). V. B. Voitovich, High Temp. Mater. Proc. 16, 243 (1997). J. Guo, C. Sun, H. Li, and H. Guan, Mater. Res. Soc. Symp. Proc: 133,591(1989). D. J. Duquette, Mater. Sci. Eng. A198, 205 (1995).