- Email: [email protected]
Spark plasma consolidation of continuous fiber reinforced titanium matrix composites
Author’s Accepted Manuscript Spark plasma consolidation of continuous fiber reinforced titanium matrix composites A. Muthuchamy, G.D. Janaki Ram, V. Subramanya Sarma www.elsevier.com/locate/msea
PII: DOI: Reference:
S0921-5093(17)30994-2 http://dx.doi.org/10.1016/j.msea.2017.07.085 MSA35333
To appear in: Materials Science & Engineering A Received date: 20 December 2016 Revised date: 10 July 2017 Accepted date: 27 July 2017 Cite this article as: A. Muthuchamy, G.D. Janaki Ram and V. Subramanya Sarma, Spark plasma consolidation of continuous fiber reinforced titanium matrix c o m p o s i t e s , Materials Science & Engineering A, http://dx.doi.org/10.1016/j.msea.2017.07.085 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Spark plasma consolidation of continuous fiber reinforced titanium matrix composites
A. Muthuchamy*, G.D. Janaki Ram, V. Subramanya Sarma B.
Department of Metallurgical and Materials Engineering, Indian Institute of Technology Madras, Chennai 600 036, India
*
Corresponding author. [email protected]
Abstract An attempt was made to produce continuous fiber reinforced titanium matrix composites using spark plasma sintering machine from foils and fibers. The materials used were metastable beta titanium alloy Ti-15V-3Cr-3Al-3Sn foils (125 µm thick) and silicon carbide fibers (100 µm in diameter). The method involved consolidating foils with fibers placed in between them at regular intervals. After a series of trials, the feasibility of the process was established. Spark plasma consolidation enabled satisfactory foil bonding and fiber embedment at a significantly lower temperature and time compared to conventional diffusion bonding or vacuum hot pressing. The composite samples consolidated above the beta-transus temperature of the titanium alloy were found to develop satisfactory strengthening precipitation upon direct aging. The current study shows that spark plasma consolidation is an excellent alternative to conventional diffusion bonding for producing continuous fiber reinforced titanium matrix composites.
Keywords: Titanium matrix composite; Spark plasma sintering; Beta-titanium alloy; Silicon carbide fiber
1. Introduction Metal matrix composites are one of the several classes of advanced materials which are expected to play significant role in future aerospace, automotive and other structural applications. They can have dramatic effect on performance and weight as compared to conventional unreinforced metals. They combine the high strength and stiffness of the ceramic reinforcement material with the damage tolerance, impact resistance and environmental stability of the metallic matrix material. Among the metal matrix composites, continuous fiber reinforced titanium matrix composites are of particular interest. Over the past two decades, SiC fiber reinforced titanium matrix composites (SiCf/Tim) are being developed for use in aeroengine and airframe applications [1, 2]. For use in airframes, the high specific modulus of titanium matrix composites has been the impetus [3, 4], while engine makers have sought to take advantage of their high specific strength, especially for compressor rotor applications [5].
Several methods have been proposed and demonstrated for producing continuous fiber reinforced metal matrix composites in the past. The choice of the process is mainly governed by the type and the nature of the fiber and the matrix materials and the geometry of the parts to be manufactured. For producing SiCf/Tim composites, pressure infiltration and squeeze casting techniques are not very useful as liquid Ti is extremely reactive to SiC. Consequently, solid-state methods such as diffusion bonding or vacuum hot pressing are normally used for making SiCf/Tim composites. Using these processes, composites have been successfully manufactured by various investigators adopting different approaches, such as: (i) foil-fiber-foil, (ii) monotape and (iii) matrix coated fiber [6]. In the foil-fiber-foil approach, foils and fibers are arranged in alternating layers and the sandwich is pressed together at high temperatures to embed the fiber in the matrix material. This approach is relatively simple and
economical, but fiber shifting or misalignment and fiber-fiber contact defects are very difficult to avoid (due to extensive plastic flow of the matrix material).
A critical issue to be addressed in the use of titanium matrix composites is the high reactivity between the fiber and the matrix phases, leading to undesirable reaction products at the fiber/matrix interface and consequent degradation in mechanical properties [7-9]. In order to improve the fiber/matrix interface compatibility and avoid or minimize deleterious fiber/matrix interactions, many different protective coatings such as pyrolytic carbon, TiB2, Al2O3, TiC, TiN, and Mo have been tried on SiC fibers [7-11]. Much of the early work on titanium matrix composites was done using an alpha titanium alloy or an near-alpha titanium alloy or an alpha+beta titanium alloy as the matrix material. For reinforcement, SiC fibers were favored and the tendency was to employ them in relatively high volume fractions. Years of intense research on these lines, however, were not very fruitful. The main problems encountered were [12]: (i) Large variation in properties and poor transverse and throughthickness properties; (ii) marginal benefits in creep and fatigue resistance, especially at lower stresses; (iii) Very high cost of processing. With time, a realization came that the SiC/Ti system is too reactive for any reasonable service life at high temperatures and that the composite suffers more and more from fiber-fiber contact defects as the fiber volume fraction is increased, unless processed via matrix-coated fiber method, which is the most expensive of all the available processing methods. As a result, the goals for fiber reinforced titanium matrix composites were redefined by lowering the expectations, targeting primarily room temperature applications. In subsequent efforts, accordingly, the fiber volume fractions were significantly reduced (to less than 10 vol. %) in order to simplify fiber management and avoid fiber-fiber contact defects. Secondly, high-strength metastable beta titanium alloys were favored as matrix materials to other titanium alloys. The main reasons for this are: (i) because metastable beta titanium alloys are precipitation hardenable, they are much stronger than
near-alpha and alpha+beta titanium alloys at room temperature; (ii) because metastable beta titanium alloys can be readily produced in foil form, they allow for cost-effective foil-fiberfoil methods for producing the composite; (iii) because metastable beta titanium alloys have a relatively lower beta-transus temperature, they can be consolidated at lower temperatures and in shorter times, thereby minimizing undesirable fiber/matrix interfacial reactions.
All the existing methods/approaches of making SiCf/Tim composites involve high temperatures and long holding times. One way to minimize the consolidation temperature and time is to make use of in-situ resistance heating techniques such as spark plasma sintering. Spark plasma sintering is widely used for nanocrystalline powders as it can achieve good densification at significantly lower temperatures and times when compared to conventional sintering processes thereby minimizing undesirable grain growth [13]. In-situ resistance heating and field-assisted mass transport distinguishes spark plasma sintering from conventional sintering processes. The technique can similarly benefit SiCf/Tim composites by significantly lowering the processing temperature and time as compared to conventional diffusion bonding. Employing titanium powders and SiC fiber mats, Izui et al. [14, 15] first demonstrated the feasibility of producing SiCf/Tim composites using spark plasma sintering. The technique will be more efficient, flexible and useful if the matrix material is taken in the form of a foil rather than powder. However, at present, it is not clear whether titanium foils and SiC fibers can be consolidated together using a spark plasma sintering machine. Therefore, in the current work, spark plasma consolidation (SPC) of titanium foils and SiC monofilaments was attempted for producing SiCf/Tim composites.
2. Experimental Procedure
2.1 Materials
The matrix material used in this investigation was a precipitation hardenable metastable beta titanium alloy Ti-15-3 (Ti-15V-3Cr-3Sn-3Al) in the form of cold rolled foils of 125 µm thickness. SiC fibers (Sigma SM1240) of 100 µm diameter produced by chemical vapor deposition were used as reinforcements. The SiC fibers contained a protective pyrolytic carbon coating of ~ 1.5 µm thickness for minimizing undesirable titanium silicide formation at the fiber/matrix interfaces during processing as well as during high temperature service.
2.2 Diffusion bonding experiments
In order to assess the benefits of using spark plasma sintering machine for producing TMCs, a few conventional diffusion bonding experiments were conducted for consolidating alloy Ti-15-3 foils. The foils were first cleaned with acetone and then with diluted hydrofluoric acid. The foils (25 mm ×25 mm × 125 µm, 15-20 foils) were stacked one over the other and placed in a fixture consisting of a screw mechanism for application of pressure. Diffusion bonding experiments were conducted in a vacuum furnace at different temperatures in the range of 800 to 950 C. The applied pressure was in the range of 25 to 40 MPa and the bonding time ranged from 1 to 4 hours.
2.3 Spark Plasma Consolidation Experiments
2.3.1 Monolithic samples
SPC experiments were conducted on a Dr. Sinter SPS-5000 machine (Sumitomo Metals, Japan). During the process, the bulk temperature of the sample being consolidated was continuously monitored using an optical pyrometer. Based on the feedback from the
pyrometer, the magnitude of current was continuously adjusted to maintain the sample at a constant temperature throughout the process. Initially, SPC experiments were conducted for producing monolithic samples by consolidating a stack of alloy Ti-15-3 foils (25 mm × 25 mm × 125 µm, 15-20 foils placed one over the other). Experiments were conducted to evaluate the effects of process parameters (consolidation temperature, pressure, and time) and to identify the optimum parameter combination. The experimental conditions are summarized in Table 1.
Table 1. SPC parameters used for producing alloy Ti-15-3 monolithic samples. Sample
Consolidation
Pressure
Consolidation time
#
temperature (°C)
(MPa)
(Minutes)
1
650
16
60
2
700
16
60
3
750
8
30
4
800
8
20
Following consolidation, the monolithic samples were subjected to post-process aging heat treatment for strengthening the beta titanium matrix by precipitating fine alpha platelets. The samples were aged at 530 ºC for 10 h followed by argon quenching in a vacuum furnace. Each sample was polished using a series of emery papers followed by cloth polishing with 0.5 µm alumina suspension. The polished samples were etched using Kroll’s reagent (92 ml H2O, 5 ml HNO3, 2 ml HF). Samples were examined under a scanning electron microscope (SEM) to study the bonding between individual foils.
The amount of alpha phase in the monolithic samples before and after the heat treatment was evaluated using x-ray diffraction (XRD). XRD characterization was done using X’Pert Pro (PANalytical, The Netherlands) X-ray diffractometer with Cu-Kα radiation (45
kV, 30 mA). XRD measurements were performed in the 2θ range of 20 to 90˚ using a step size of 0.2˚. Quantitative phase analysis was carried out by Rietveld analysis, approximating the intensity line profiles with a symmetric pseudo-Voight function using FullProf suite software.
Samples for electron back-scattered diffraction (EBSD) studies were prepared by electropolishing after standard metallographic preparation. Scans were performed using an EBSD camera attached to an FEI Quanta 200 scanning electron microscope operating at 30 kV. Scans were carried out using a step size of 0.5 µm and a hexagonal grid with the reference patterns for both alpha and beta phases loaded in the orientation imaging microscopy software (TSL-OIM, version 6.3). Standard clean-up procedure (grain dilation for single iteration) was applied before analyzing the EBSD data. EBSD studies were carried out on monolithic samples consolidated at 750 and 800 °C before and after heat treatment. For each sample, four different scans were conducted and the average fractions of alpha and beta phases were reported.
For transmission electron microscopy (TEM), initially 200 µm thick slices were cut using a low speed diamond saw. The slices were then thinned to about 90 µm thickness by mechanical polishing. Disks of 3 mm diameter were then punched and further thinned using a twin-jet electro polisher in a solution consisting of one part HNO3 and three parts CH3OH (temperature: −30 °C, voltage: 12 V). Prior to loading in TEM, the jet-thinned specimens were ion milled for 0.5 h (beam energy: 3.5 keV, angle of incidence: 2°). The specimens were examined using an FEI Tecnai T20 microscope operated at 200 kV.
Microhardness tests were conducted on monolithic samples using a Leco LM248AT Vickers microhardness tester. Tests were conducted using load of 500 g and a dwell time of
10 s. Ten hardness readings were taken from each sample and the average hardness was reported.
2.3.2 Composite samples
SPC experiments were conducted for producing SiCf/Ti-15-3 composite samples by consolidating a stack of beta titanium foils (25 mm ×25 mm × 125 µm, 15-20 foils) with SiC fibers placed in between foils every alternating layer. SPC experiments were conducted at 750C and 800C. The consolidation pressure ranged from 16 to 40 MPa. A consolidation time of 20 minutes was targeted. The process parameter combinations used are listed in Table 2. After consolidation, the composite samples were aged at 530 ºC for 10 h followed by argon quenching in a vacuum furnace. Cross sections cut from the composite samples were subjected to microstructural examination and were investigated for bonding between individual foils and fiber embedment. Based on microstructural examination, the optimum consolidation temperature, pressure, and time for SiCf/Ti-15-3 composites were found 800 C, 40 Mpa, and 20 minutes, respectively.
Table 2. SPC parameters used for producing SiCf/ Ti-15-3 composite samples. Sample
Consolidation
Pressure
Consolidation time
#
temperature (°C)
(MPa)
(Minutes)
1
750
16
20
2
750
24
20
3
750
32
20
4
750
40
20
5
800
16
20
6
800
24
20
7
800
32
20
8
800
40
20
Using the optimum SPC parameters, SiCf/Ti-15-3 composite samples of larger dimensions (110 mm x 30 mm x 2 mm, unidirectional ply) were prepared for tensile testing. In these samples, the fibers (corresponding to ~ 2.3 vol%) were so arranged that they do not get cut or exposed during subsequent tensile specimen preparation by machining. The composite samples were aged at 530˚C for 10 hours in a vacuum furnace followed by argon quenching to room temperature. Tensile tests were carried out on a computer controlled universal testing machine as per ASTM E8 using a crosshead travel speed of 0.1 mm/min. For comparison, tensile tests were also conducted on monolithic samples, prepared under the same conditions.
3. Results and discussion
3.1 Diffusion bonded monolithic samples Microstructural examination of the diffusion bonded samples at 900˚C/40MPa/3h revealed unsatisfactory foil bonding (Fig. 1). It was found that a bonding temperature of 950 C and bonding time of 3 hours were needed for achieving satisfactory consolidation of alloy Ti-15-3 foils. As can be expected, still higher bonding temperatures and longer times may be needed for producing SiCf/Ti-15-3 composites, which can easily lead to undesirable reaction products at the fiber/matrix interface.
Fig. 1. SEM micrograph of a diffusion bonded Ti-15-3 monolithic sample (900˚C/40MPa/3h).
3.2 Spark plasma consolidated monolithic samples
The effects of processing conditions on foil bonding and microstructure were investigated. An SEM micrograph of the as-received Ti-15-3 foil (long-transverse section) is shown in Fig. 2. As expected, due to extensive cold work, the foils showed severely deformed grains along the rolling direction.
Fig. 2. SEM micrograph of as-received Ti-15-3 foil. Note severely deformed grains. Microstructural examination of the monolithic samples consolidated at 650 °C revealed poor bonding at the foils (Fig. 3). Samples consolidated at temperatures above 650°C, however, showed satisfactory foil bonding (Fig. 4). During SPC, the foils were found to undergo recrystallization and grain growth. The grain size was found to increase with increasing consolidation temperature. Fig. 5 shows the EBSD maps of a monolithic sample consolidated at 800 °C. It can be seen that the grain boundaries at the foil interfaces have the same character as the grain boundaries within the bulk of the foils. The beta-transus temperature of alloy Ti-15-3 is ~ 770 °C. As can be seen from Fig. 4a and Fig. 4b, the monolithic samples consolidated below the beta-transus showed a continuous layer of alpha phase along the original foil interfaces as well as along the grain boundaries within the foils.
In contrast, the samples consolidated at 800°C showed no grain boundary alpha phase (Fig. 4c and Fig. 4d).
Fig. 3. SEM micrograph of a monolithic sample spark plasma consolidated at 650 C. Note poor foil bonding.
Fig. 4. SEM micrographs of alloy Ti-5-3 monolithic samples spark plasma consolidated at 750C (a and b) and 800C (c and d). Samples consolidated at 800C show no grain boundary alpha phase.
Fig.5. EBSD maps of a monolithic sample spark plasma consolidated at 800C: (a) Grain boundary misorientation map, (b) Unique grain color map. Vickers bulk hardness measurements were conducted on Ti-15-3 monolithic samples, before and after the aging treatment, spark plasma consolidated at 750°C and 800°C. The results are summarized in Table 3. The samples consolidated at 800 °C were found to respond better to the aging treatment.
EBSD studies confirmed the presence of some amount of alpha phase in Ti-15-3 monolithic samples consolidated at 750 °C (Fig. 6a). The alpha phase was practically absent in the samples consolidated at 800 °C (Fig. 6b). After aging, as expected, both the samples showed considerable amount of alpha phase, as seen in Fig. 7a and Fig. 7b. Using the EBSD data, the amount of alpha phase was quantified in these samples. The results are presented in Table 4. The X-ray diffractograms of these samples are shown in Fig. 8 and Fig. 9. The results are broadly in agreement with EBSD observations. Quantification of the amount of alpha phase in these samples was also attempted using Rietveld analysis (Table 4). As can be seen, the measurements from XRD closely match with the EBSD measurements.
Table 3. Results of Vickers hardness testing on monolithic samples (based on 10 measurements in each case). Hardness (HV)
Sample Consolidated at 750°C (before aging)
280 ± 6
Consolidated at 750°C (after aging)
325 ± 7
Consolidated at 800°C (before aging)
240 ± 6
Consolidated at 800°C (after aging)
410 ± 5
Fig. 6. EBSD phase maps of alloy Ti-5-3 monolithic samples (before aging) spark plasma consolidated at 750C (a) and 800C (b). Table 4. Amount of alpha phase in Ti-15-3 monolithic samples as determined from EBSD and XRD. Amount of alpha phase Sample EBSD
XRD
750°C consolidated, before aging
20
19
750°C consolidated, after aging
32
36
800°C consolidated, before aging
0
0
800°C consolidated, after aging
41
40
Fig. 7. EBSD phase maps of alloy Ti-5-3 monolithic samples (after aging) spark plasma consolidated at 750C (a) and 800C (b).
Fig. 8. X-ray diffractograms of Ti-15-3 monolithic samples (before aging) spark plasma consolidated at 750C and 800C.
Fig. 9. X-ray diffractograms of Ti-15-3 monolithic samples (after aging) spark plasma consolidated at 750C and 800C. TEM studies were also carried out on Ti-15-3 monolithic samples spark plasma consolidated at 750 ºC and 800 ºC in as-processed and aged conditions. Representative TEM micrographs are shown in Fig. 10 and Fig. 11. The presence of alpha and beta phases was confirmed from selected area diffraction studies. In as-processed condition, the samples consolidated at 750 ºC showed a large number of coarse alpha plates (Fig. 10a). The alpha plates were found to coarsen further after aging (Fig. 10b). Importantly, these samples showed hardly any fine alpha platelets. In contrast, the samples consolidated at 800ºC showed no alpha phase in as-processed condition (Fig. 11a), but showed numerous fine alpha platelets after aging (Fig. 11b). These observations corroborate well with EBSD and XRD results.
Based on the above investigations, microstructure evolution in metastable beta titanium alloy Ti-15-3 during SPC at 750 ºC and 800 ºC can be understood as follows. During consolidation at 750 ºC, because the material is heated to a temperature high in the alpha +
beta phase field and held for sufficient time, considerable precipitation of coarse alpha plates takes place. When these samples are subsequently aged at 530 ºC, some additional alpha precipitation occurs. This additional alpha precipitation involves coarsening of existing alpha plates rather than formation of any new fine alpha platelets. This explains why the samples consolidated at 750 ºC did not pick up much hardness after aging. Whereas, when the consolidation temperature is 800 ºC, because the temperature is above the beta-transus, no alpha phase can form during SPC. Thus, in as-processed condition, the microstructure of the material is just supersaturated metastable beta (similar to a properly solution treated material). Upon aging, these samples develop optimum strengthening precipitation (fine alpha platelets) and satisfactory hardness.
Fig. 10. TEM micrographs of Ti-15-3 monolithic samples spark plasma consolidated at 750 °C: (a) before aging, (b) after aging. (c) and (d) show the selected area diffraction patterns obtained from alpha and beta phases, as marked in (a).
Fig. 11. TEM micrographs of Ti-15-3 monolithic samples spark plasma consolidated at 800 °C: (a) before aging, (b) after aging. (c) and (d) show the selected area diffraction patterns obtained from alpha and beta phases, as marked in (b). From the above results, it can be seen that for achieving optimum strengthening precipitation and best possible mechanical properties in spark plasma consolidated metastable beta alloy Ti-15-3 monolithic samples, it is necessary to avoid formation of excessive coarse alpha plates by: (i) processing the samples above the beta-transus temperature, and (ii) cooling the samples rapidly enough from the processing temperature to room temperature. During subsequent aging at 530 ºC, fine alpha platelets precipitate in the beta matrix, leading to significant strengthening. Processing above the beta-transus temperature is also important for avoiding the detrimental grain boundary alpha formation in this alloy. Based on these considerations, a consolidation temperature of 800 C was considered optimum for alloy Ti15-3.
3.3 Spark plasma consolidated composite samples
The SiCf/Ti-15-3 composite samples consolidated under different conditions were examined under optical and scanning electron microscopes. During consolidation, the titanium foils undergo extensive plastic deformation resulting in metal flow around the fiber and is assisted by diffusional creep and/or dislocation creep mechanisms [16-18]. At hightemperatures and low-stress levels diffusional creep dominates. In contrast, at high stress levels, dislocation creep dominates [18, 19]. As densification progresses, grain boundary diffusion facilitates boundary migration and void shrinkage at the fiber/matrix interfaces [17, 19]. The time of consolidation mainly depends on the temperature and the applied pressure. Apart from the effects of heat and pressure, the electric field can assist in foil bonding and fiber embedment by promoting mass transport [20]. The effects of electric field can be evaluated from the electromigration theory. However, in a typical field-activated sintering process like spark plasma sintering, the temperature and current are not independent parameters and hence the thermal effects of the current (Joule heating) cannot be unambiguously separated from the intrinsic field effects [21].
In composite samples consolidated at both 750 ºC and 800 ºC using an applied pressure of less than 40 MPa, the fiber embedment was found to be unsatisfactory (Fig. 12a). While the composite samples consolidated at 750C for 20 minutes using 40 MPa pressure showed satisfactory foil bonding and fiber embedment (Fig. 12b), considering the microstructural advantages in processing alloy Ti-15-3 above its beta-transus temperature, a consolidation temperature of 800 C was identified as the optimum consolidation temperature in combination with 40 MPa pressure and 20 minutes of consolidation time (Fig. 13). In an attempt to reduce the consolidation time further, SPC experiments were conducted by increasing the pressure to 50 MPa, which, however, was found to result in fiber breakage.
Fig. 12. Fiber embedment in spark plasma consolidated SiCf/Ti-15-3 composite samples: (a) 750C/32MPa/20minutes (incomplete fiber embedment), (b) 750C/40MPa/20minutes (satisfactory fiber embedment).
Fig. 13. SEM micrograph of a SiCf/Ti-15-3 composite sample spark plasma consolidated at 800 C. The results of tensile testing on optimally consolidated monolithic and composite samples after aging are summarised in Table 5. Typical stress-strain plots are shown in Fig. 14. The results show that the composite samples are stronger and stiffer as compared to the monolithic samples. The increase in strength and stiffness is commensurate with the fiber volume fraction. As per the rule of mixtures (σc = Vf σf + Vm σm, where Vf and Vm are the
volume fractions of fiber and matrix, respectively, and σc, σf, and σm are the strengths of composite, fiber, and matrix, respectively), the Young’s modulus and the yield strength of a SiCf/Ti-15-3 composite with 2.3 fiber vol.% should be 128 GPa and 1470 MPa, respectively (calculated using the tensile data obtained on alloy Ti-15-3 monolithic samples for the matrix properties). These values compare well with the test results obtained on the composite samples. This indicates that the SiC fibers in the SiCf/Ti-15-3 composite samples produced in the current study are serving to their full potential. As can be expected, the composite samples showed slightly lower tensile elongations compared to the monolithic samples. SEM examination of the fracture surfaces of monolithic and composite samples did not reveal any foil debonding or fiber pull-out (Fig. 15), confirming that the foil bonding and fiber embedment are indeed satisfactory in the SiCf/Ti-15-3 composite samples produced in the current study. The fracture surfaces of the SiC fibers were very much flat, but the titanium matrix showed ductile dimpled rupture features.
Table 5. Result of tensile testing on spark plasma consolidated monolithic and composite samples (based on six tests in each case). Young’s
Yield strength
Ultimate tensile
Elongation
modulus (GPa)
(MPa)
strength (MPa)
(%)
Monolithic
117 ± 5
1270 ± 10
1360 ± 5
7.5
Composite
136 ± 7
1480 ± 8
1510 ± 6
6
Sample
Fig.14. Stress-strain plots of spark plasma consolidated monolithic and composite samples.
400 µm
100 µm
Fig. 15. Fractographs of spark plasma consolidated monolithic (a) and composite (b) samples. Despite their relatively low fiber volume fraction, the tensile properties of the SiC f/Ti15-3 composite samples produced in this work compare favourably with the test data on similar composites reported in earlier investigations. For example, for a Ti-15-3 matrix composite with 21 vol. % SiC fibers produced under optimal conditions, Izui et al. [14, 15] reported a tensile strength of 1570 MPa in as-consolidated condition. Similarly, Gabb et al.
[22] reported a tensile strength of 1461 MPa for a heat-treated Ti-15-3 matrix composite (solutionized at 788 °C for 15 minutes, water quenched, and then aged at 300 °C for 24 hours) consisting of 37 vol. % SiC fibers. These comparisons clearly show that the matrix microstructure plays an important role in determining the composite properties. In the first case, as no heat treatment was attempted to induce alpha precipitation, the matrix is not as strong as that in the present composite samples. In the second case, the titanium matrix is significantly harder (500 HV) than that in the present composite samples (410 HV). It may be noted that the heat treatment employed by Gabb et al. [22] does not provide the optimum combination of strength and ductility in alloy Ti-15-3.
4. Conclusions
Spark plasma consolidation can be advantageously utilized for producing SiC f/Ti-153 composites via foil-fiber-foil method. Compared to conventional diffusion bonding or hot pressing, spark plasma consolidation can achieve satisfactory foil bonding and fiber embedment at significantly lower temperatures and times. Satisfactory consolidation of SiC fibers and Ti-15-3 foils was demonstrated at temperatures as low as 750 C in just 20 minutes. The reduced processing temperature and time with spark plasma consolidation translates to a very healthy fiber/matrix interface as well as significant cost savings.
The present study shows that a metastable beta titanium alloy is best processed above its beta-transus temperature. In the case of alloy Ti-15-3, consolidation at temperatures below its beta-transus temperature (~ 770 C) was found to result in formation of undesirable grain boundary alpha phase as well as intragranular precipitation of a large number of coarse alpha plates. Consequently, these samples could not develop satisfactory strengthening precipitation upon subsequent aging. On
the other hand, when alloy Ti-15-3 was consolidated at 800 C, a fully beta microstructure was obtained as the material was taken into the beta phase field during consolidation and was cooled sufficiently rapidly to room temperature to preclude any alpha precipitation. These samples developed satisfactory strengthening precipitation (in the form of fine alpha platelets) upon subsequent aging. For this reason, the current study identifies 800C as the optimum consolidation temperature for SiCf/Ti15-3 composites although spark plasma consolidation can facilitate satisfactory foil bonding and fiber embedment even at 750 C.
SiCf/Ti-15-3 composites produced in the current study with a fiber volume fraction of 2.3 vol. % using spark plasma consolidation showed satisfactory improvements in Young’s modulus (from 117 GPa to 136 GPa) and yield strength (1270 MPa to 1480 MPa) compared to monolithic Ti-15-3 samples. The composite properties compared well with the expectations from the rule of mixtures, indicating that the SiC fibers are serving to their full potential in the composite.
References [1] T. W.Clyne, P.J. Withers, An Introduction to Metal Matrix Composites, Cambridge University Press, Cambridge, 1993. [2] R. A. Mackay, P.K. Brindley, F.H. Froes, Continuous fiber-reinforced titanium aluminide composites Journal of the Minerals, The journal of the Minerals, Metals & Materials Society 1991, 43 (5), 23-29. [3] S.F. Bauman, P.K. Brindley, S.D. Smith, Reaction Zone Microstructure in a Ti3AI+Nb/SiC Composite, Metall. Trans. A 1990, 21A, 1559-1569. [4] D.B. Gundel, F.E. Wawner, Interfacial Reaction Kinetics of Coated SiC Fibers with Various Titanium Alloys, Scripta Metallurgica et Meterialia 1991, 25, 437-441.
[5] Y.G. Zhou, Y.Q. Yang, Progress in the study of titanium matrix compositesreinforced by SiC fibers, Journal of Acta Met Sinica 2002, 38 (suppl), 461-465. [6] C. Leyens, J. Hausmann, J. Kumpfert, Continuous fiber reinforced titanium matrix composites: Fabrication, properties and applications,, Advanced Engineering Materials 2003, 5(6), 399-410. [7] S. Djanarthany, J.C. Viala, J. Bouix, Development of SiC/TiAl composites: Processing and interfacial phenomena, Materials Science and Engineering A 2001, 300(1-2), 211-218. [8] Y.Q. Yang, Z.J. Ma, X.H. Lü, J. K. Li, Y. Chen, Y.L. Ai, Studies on Interface of SiCf/Ti-6Al-4V Composites, Rare Met Mater Eng. 2006, 35(10), 1516-1521. [9] Y.Q. Yang, X.H. Lü, X. Luo, Z.J. Ma, J.K. Li, Q. Wen, Y. Chen, Kinetics of Interfacial Reaction in SiCf/Ti6Al4V Composites, Materials Science Forum 2007, 546/549, 1627-1632. [10] Z.A. Munir, R.M. German, A Generalized Model for the Prediction of Periodic Trends in the Activation of Sintering of Refractory Metals, High Temperature Science 1977, 9, 275-
283. [11] X. Luo, Y. Yang, Y. Yu, X. Wang, B. Huang, Y. Chen, Effect of Mo coating on the interface and mechanical properties of SiC fiber reinforced Ti6Al4V composites, Materials Science and Engineering A 2012, 550, 286–292. [12] G. Lutjering, J.C. Williams, Titanium, Springer Science & Business Media, 2007. [13] Z.A. Munir, U. Anselmi-Tamburini, M. Ohyanagi, The effect of electric field and pressure on the synthesis and consolidation of materials: A review of the spark plasma sintering method, Journal of Materials Science 2006, 41(3), 763–777.
[14] H. Izui, S. Kinbara, M. Okano, Mechanical Properties of Ti-15-3 Alloy Reinforced With SiC Fibers by Spark Plasma Sintering, Proceedings of the 6th Pacific Rim Conference on Ceramics and Glass Technology 2006, Maui, Hawaii, 289-300 [15] H. Izui, S. Kinbara, M. Okano, Tensile and Fatigue Strengths of SCS-6/Ti-6Al-4V Titanium Matrix Composite by Spark Plasma Sintering, Journal of the Japan Society of Powder and Powder Metallurgy 2005, 52 (10), 722-729 [16] J.C.M. Hwang, R.W. Balluffi, On a possible temperature dependence of the activation energy for grain boundary diffusion in metals, Scr. Metall. 1978, 12, 709-714 [17] B. Derby, The dependence of grain size on stress during dynamic recrystallization, Acta Metallurgica Et Materialia 1991, 39(5), 955-962. [18] C.H Hamilton, N.E Paton, Super plastic forming of structural alloys, TMS, Warrendale, PA 1982. [19] M. Zhou M, F.P.E. Dunne, Mechanism-based constitutive equations for the superplastic behaviour of a titanium alloy, J. Strain Anal Eng Des. 1996, 31, 187-196. [20] Z.A. Munir, H. Schmalzried, The effect of external fields on mass-transport and defect-related phenomena, Journal of Materials Synthesis and processing 1993, Vol.1, 3-16. [21] H. Conrad, Electroplasticity in metals and ceramics, Materials Science and Engineering A 2000, 287(2), 276-287. [22] T.P. Gabb, J. Gayda, B. A. Lerch, G. R. Halford. The Effect of Matrix Mechanical Properties on [0]8 Unidirectional SiC/Ti Composite Fatigue Resistance, Scripta Metallurgica et Materialia, 1991, 25: 2879-2884.
PII: DOI: Reference:
S0921-5093(17)30994-2 http://dx.doi.org/10.1016/j.msea.2017.07.085 MSA35333
To appear in: Materials Science & Engineering A Received date: 20 December 2016 Revised date: 10 July 2017 Accepted date: 27 July 2017 Cite this article as: A. Muthuchamy, G.D. Janaki Ram and V. Subramanya Sarma, Spark plasma consolidation of continuous fiber reinforced titanium matrix c o m p o s i t e s , Materials Science & Engineering A, http://dx.doi.org/10.1016/j.msea.2017.07.085 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Spark plasma consolidation of continuous fiber reinforced titanium matrix composites
A. Muthuchamy*, G.D. Janaki Ram, V. Subramanya Sarma B.
Department of Metallurgical and Materials Engineering, Indian Institute of Technology Madras, Chennai 600 036, India
*
Corresponding author. [email protected]
Abstract An attempt was made to produce continuous fiber reinforced titanium matrix composites using spark plasma sintering machine from foils and fibers. The materials used were metastable beta titanium alloy Ti-15V-3Cr-3Al-3Sn foils (125 µm thick) and silicon carbide fibers (100 µm in diameter). The method involved consolidating foils with fibers placed in between them at regular intervals. After a series of trials, the feasibility of the process was established. Spark plasma consolidation enabled satisfactory foil bonding and fiber embedment at a significantly lower temperature and time compared to conventional diffusion bonding or vacuum hot pressing. The composite samples consolidated above the beta-transus temperature of the titanium alloy were found to develop satisfactory strengthening precipitation upon direct aging. The current study shows that spark plasma consolidation is an excellent alternative to conventional diffusion bonding for producing continuous fiber reinforced titanium matrix composites.
Keywords: Titanium matrix composite; Spark plasma sintering; Beta-titanium alloy; Silicon carbide fiber
1. Introduction Metal matrix composites are one of the several classes of advanced materials which are expected to play significant role in future aerospace, automotive and other structural applications. They can have dramatic effect on performance and weight as compared to conventional unreinforced metals. They combine the high strength and stiffness of the ceramic reinforcement material with the damage tolerance, impact resistance and environmental stability of the metallic matrix material. Among the metal matrix composites, continuous fiber reinforced titanium matrix composites are of particular interest. Over the past two decades, SiC fiber reinforced titanium matrix composites (SiCf/Tim) are being developed for use in aeroengine and airframe applications [1, 2]. For use in airframes, the high specific modulus of titanium matrix composites has been the impetus [3, 4], while engine makers have sought to take advantage of their high specific strength, especially for compressor rotor applications [5].
Several methods have been proposed and demonstrated for producing continuous fiber reinforced metal matrix composites in the past. The choice of the process is mainly governed by the type and the nature of the fiber and the matrix materials and the geometry of the parts to be manufactured. For producing SiCf/Tim composites, pressure infiltration and squeeze casting techniques are not very useful as liquid Ti is extremely reactive to SiC. Consequently, solid-state methods such as diffusion bonding or vacuum hot pressing are normally used for making SiCf/Tim composites. Using these processes, composites have been successfully manufactured by various investigators adopting different approaches, such as: (i) foil-fiber-foil, (ii) monotape and (iii) matrix coated fiber [6]. In the foil-fiber-foil approach, foils and fibers are arranged in alternating layers and the sandwich is pressed together at high temperatures to embed the fiber in the matrix material. This approach is relatively simple and
economical, but fiber shifting or misalignment and fiber-fiber contact defects are very difficult to avoid (due to extensive plastic flow of the matrix material).
A critical issue to be addressed in the use of titanium matrix composites is the high reactivity between the fiber and the matrix phases, leading to undesirable reaction products at the fiber/matrix interface and consequent degradation in mechanical properties [7-9]. In order to improve the fiber/matrix interface compatibility and avoid or minimize deleterious fiber/matrix interactions, many different protective coatings such as pyrolytic carbon, TiB2, Al2O3, TiC, TiN, and Mo have been tried on SiC fibers [7-11]. Much of the early work on titanium matrix composites was done using an alpha titanium alloy or an near-alpha titanium alloy or an alpha+beta titanium alloy as the matrix material. For reinforcement, SiC fibers were favored and the tendency was to employ them in relatively high volume fractions. Years of intense research on these lines, however, were not very fruitful. The main problems encountered were [12]: (i) Large variation in properties and poor transverse and throughthickness properties; (ii) marginal benefits in creep and fatigue resistance, especially at lower stresses; (iii) Very high cost of processing. With time, a realization came that the SiC/Ti system is too reactive for any reasonable service life at high temperatures and that the composite suffers more and more from fiber-fiber contact defects as the fiber volume fraction is increased, unless processed via matrix-coated fiber method, which is the most expensive of all the available processing methods. As a result, the goals for fiber reinforced titanium matrix composites were redefined by lowering the expectations, targeting primarily room temperature applications. In subsequent efforts, accordingly, the fiber volume fractions were significantly reduced (to less than 10 vol. %) in order to simplify fiber management and avoid fiber-fiber contact defects. Secondly, high-strength metastable beta titanium alloys were favored as matrix materials to other titanium alloys. The main reasons for this are: (i) because metastable beta titanium alloys are precipitation hardenable, they are much stronger than
near-alpha and alpha+beta titanium alloys at room temperature; (ii) because metastable beta titanium alloys can be readily produced in foil form, they allow for cost-effective foil-fiberfoil methods for producing the composite; (iii) because metastable beta titanium alloys have a relatively lower beta-transus temperature, they can be consolidated at lower temperatures and in shorter times, thereby minimizing undesirable fiber/matrix interfacial reactions.
All the existing methods/approaches of making SiCf/Tim composites involve high temperatures and long holding times. One way to minimize the consolidation temperature and time is to make use of in-situ resistance heating techniques such as spark plasma sintering. Spark plasma sintering is widely used for nanocrystalline powders as it can achieve good densification at significantly lower temperatures and times when compared to conventional sintering processes thereby minimizing undesirable grain growth [13]. In-situ resistance heating and field-assisted mass transport distinguishes spark plasma sintering from conventional sintering processes. The technique can similarly benefit SiCf/Tim composites by significantly lowering the processing temperature and time as compared to conventional diffusion bonding. Employing titanium powders and SiC fiber mats, Izui et al. [14, 15] first demonstrated the feasibility of producing SiCf/Tim composites using spark plasma sintering. The technique will be more efficient, flexible and useful if the matrix material is taken in the form of a foil rather than powder. However, at present, it is not clear whether titanium foils and SiC fibers can be consolidated together using a spark plasma sintering machine. Therefore, in the current work, spark plasma consolidation (SPC) of titanium foils and SiC monofilaments was attempted for producing SiCf/Tim composites.
2. Experimental Procedure
2.1 Materials
The matrix material used in this investigation was a precipitation hardenable metastable beta titanium alloy Ti-15-3 (Ti-15V-3Cr-3Sn-3Al) in the form of cold rolled foils of 125 µm thickness. SiC fibers (Sigma SM1240) of 100 µm diameter produced by chemical vapor deposition were used as reinforcements. The SiC fibers contained a protective pyrolytic carbon coating of ~ 1.5 µm thickness for minimizing undesirable titanium silicide formation at the fiber/matrix interfaces during processing as well as during high temperature service.
2.2 Diffusion bonding experiments
In order to assess the benefits of using spark plasma sintering machine for producing TMCs, a few conventional diffusion bonding experiments were conducted for consolidating alloy Ti-15-3 foils. The foils were first cleaned with acetone and then with diluted hydrofluoric acid. The foils (25 mm ×25 mm × 125 µm, 15-20 foils) were stacked one over the other and placed in a fixture consisting of a screw mechanism for application of pressure. Diffusion bonding experiments were conducted in a vacuum furnace at different temperatures in the range of 800 to 950 C. The applied pressure was in the range of 25 to 40 MPa and the bonding time ranged from 1 to 4 hours.
2.3 Spark Plasma Consolidation Experiments
2.3.1 Monolithic samples
SPC experiments were conducted on a Dr. Sinter SPS-5000 machine (Sumitomo Metals, Japan). During the process, the bulk temperature of the sample being consolidated was continuously monitored using an optical pyrometer. Based on the feedback from the
pyrometer, the magnitude of current was continuously adjusted to maintain the sample at a constant temperature throughout the process. Initially, SPC experiments were conducted for producing monolithic samples by consolidating a stack of alloy Ti-15-3 foils (25 mm × 25 mm × 125 µm, 15-20 foils placed one over the other). Experiments were conducted to evaluate the effects of process parameters (consolidation temperature, pressure, and time) and to identify the optimum parameter combination. The experimental conditions are summarized in Table 1.
Table 1. SPC parameters used for producing alloy Ti-15-3 monolithic samples. Sample
Consolidation
Pressure
Consolidation time
#
temperature (°C)
(MPa)
(Minutes)
1
650
16
60
2
700
16
60
3
750
8
30
4
800
8
20
Following consolidation, the monolithic samples were subjected to post-process aging heat treatment for strengthening the beta titanium matrix by precipitating fine alpha platelets. The samples were aged at 530 ºC for 10 h followed by argon quenching in a vacuum furnace. Each sample was polished using a series of emery papers followed by cloth polishing with 0.5 µm alumina suspension. The polished samples were etched using Kroll’s reagent (92 ml H2O, 5 ml HNO3, 2 ml HF). Samples were examined under a scanning electron microscope (SEM) to study the bonding between individual foils.
The amount of alpha phase in the monolithic samples before and after the heat treatment was evaluated using x-ray diffraction (XRD). XRD characterization was done using X’Pert Pro (PANalytical, The Netherlands) X-ray diffractometer with Cu-Kα radiation (45
kV, 30 mA). XRD measurements were performed in the 2θ range of 20 to 90˚ using a step size of 0.2˚. Quantitative phase analysis was carried out by Rietveld analysis, approximating the intensity line profiles with a symmetric pseudo-Voight function using FullProf suite software.
Samples for electron back-scattered diffraction (EBSD) studies were prepared by electropolishing after standard metallographic preparation. Scans were performed using an EBSD camera attached to an FEI Quanta 200 scanning electron microscope operating at 30 kV. Scans were carried out using a step size of 0.5 µm and a hexagonal grid with the reference patterns for both alpha and beta phases loaded in the orientation imaging microscopy software (TSL-OIM, version 6.3). Standard clean-up procedure (grain dilation for single iteration) was applied before analyzing the EBSD data. EBSD studies were carried out on monolithic samples consolidated at 750 and 800 °C before and after heat treatment. For each sample, four different scans were conducted and the average fractions of alpha and beta phases were reported.
For transmission electron microscopy (TEM), initially 200 µm thick slices were cut using a low speed diamond saw. The slices were then thinned to about 90 µm thickness by mechanical polishing. Disks of 3 mm diameter were then punched and further thinned using a twin-jet electro polisher in a solution consisting of one part HNO3 and three parts CH3OH (temperature: −30 °C, voltage: 12 V). Prior to loading in TEM, the jet-thinned specimens were ion milled for 0.5 h (beam energy: 3.5 keV, angle of incidence: 2°). The specimens were examined using an FEI Tecnai T20 microscope operated at 200 kV.
Microhardness tests were conducted on monolithic samples using a Leco LM248AT Vickers microhardness tester. Tests were conducted using load of 500 g and a dwell time of
10 s. Ten hardness readings were taken from each sample and the average hardness was reported.
2.3.2 Composite samples
SPC experiments were conducted for producing SiCf/Ti-15-3 composite samples by consolidating a stack of beta titanium foils (25 mm ×25 mm × 125 µm, 15-20 foils) with SiC fibers placed in between foils every alternating layer. SPC experiments were conducted at 750C and 800C. The consolidation pressure ranged from 16 to 40 MPa. A consolidation time of 20 minutes was targeted. The process parameter combinations used are listed in Table 2. After consolidation, the composite samples were aged at 530 ºC for 10 h followed by argon quenching in a vacuum furnace. Cross sections cut from the composite samples were subjected to microstructural examination and were investigated for bonding between individual foils and fiber embedment. Based on microstructural examination, the optimum consolidation temperature, pressure, and time for SiCf/Ti-15-3 composites were found 800 C, 40 Mpa, and 20 minutes, respectively.
Table 2. SPC parameters used for producing SiCf/ Ti-15-3 composite samples. Sample
Consolidation
Pressure
Consolidation time
#
temperature (°C)
(MPa)
(Minutes)
1
750
16
20
2
750
24
20
3
750
32
20
4
750
40
20
5
800
16
20
6
800
24
20
7
800
32
20
8
800
40
20
Using the optimum SPC parameters, SiCf/Ti-15-3 composite samples of larger dimensions (110 mm x 30 mm x 2 mm, unidirectional ply) were prepared for tensile testing. In these samples, the fibers (corresponding to ~ 2.3 vol%) were so arranged that they do not get cut or exposed during subsequent tensile specimen preparation by machining. The composite samples were aged at 530˚C for 10 hours in a vacuum furnace followed by argon quenching to room temperature. Tensile tests were carried out on a computer controlled universal testing machine as per ASTM E8 using a crosshead travel speed of 0.1 mm/min. For comparison, tensile tests were also conducted on monolithic samples, prepared under the same conditions.
3. Results and discussion
3.1 Diffusion bonded monolithic samples Microstructural examination of the diffusion bonded samples at 900˚C/40MPa/3h revealed unsatisfactory foil bonding (Fig. 1). It was found that a bonding temperature of 950 C and bonding time of 3 hours were needed for achieving satisfactory consolidation of alloy Ti-15-3 foils. As can be expected, still higher bonding temperatures and longer times may be needed for producing SiCf/Ti-15-3 composites, which can easily lead to undesirable reaction products at the fiber/matrix interface.
Fig. 1. SEM micrograph of a diffusion bonded Ti-15-3 monolithic sample (900˚C/40MPa/3h).
3.2 Spark plasma consolidated monolithic samples
The effects of processing conditions on foil bonding and microstructure were investigated. An SEM micrograph of the as-received Ti-15-3 foil (long-transverse section) is shown in Fig. 2. As expected, due to extensive cold work, the foils showed severely deformed grains along the rolling direction.
Fig. 2. SEM micrograph of as-received Ti-15-3 foil. Note severely deformed grains. Microstructural examination of the monolithic samples consolidated at 650 °C revealed poor bonding at the foils (Fig. 3). Samples consolidated at temperatures above 650°C, however, showed satisfactory foil bonding (Fig. 4). During SPC, the foils were found to undergo recrystallization and grain growth. The grain size was found to increase with increasing consolidation temperature. Fig. 5 shows the EBSD maps of a monolithic sample consolidated at 800 °C. It can be seen that the grain boundaries at the foil interfaces have the same character as the grain boundaries within the bulk of the foils. The beta-transus temperature of alloy Ti-15-3 is ~ 770 °C. As can be seen from Fig. 4a and Fig. 4b, the monolithic samples consolidated below the beta-transus showed a continuous layer of alpha phase along the original foil interfaces as well as along the grain boundaries within the foils.
In contrast, the samples consolidated at 800°C showed no grain boundary alpha phase (Fig. 4c and Fig. 4d).
Fig. 3. SEM micrograph of a monolithic sample spark plasma consolidated at 650 C. Note poor foil bonding.
Fig. 4. SEM micrographs of alloy Ti-5-3 monolithic samples spark plasma consolidated at 750C (a and b) and 800C (c and d). Samples consolidated at 800C show no grain boundary alpha phase.
Fig.5. EBSD maps of a monolithic sample spark plasma consolidated at 800C: (a) Grain boundary misorientation map, (b) Unique grain color map. Vickers bulk hardness measurements were conducted on Ti-15-3 monolithic samples, before and after the aging treatment, spark plasma consolidated at 750°C and 800°C. The results are summarized in Table 3. The samples consolidated at 800 °C were found to respond better to the aging treatment.
EBSD studies confirmed the presence of some amount of alpha phase in Ti-15-3 monolithic samples consolidated at 750 °C (Fig. 6a). The alpha phase was practically absent in the samples consolidated at 800 °C (Fig. 6b). After aging, as expected, both the samples showed considerable amount of alpha phase, as seen in Fig. 7a and Fig. 7b. Using the EBSD data, the amount of alpha phase was quantified in these samples. The results are presented in Table 4. The X-ray diffractograms of these samples are shown in Fig. 8 and Fig. 9. The results are broadly in agreement with EBSD observations. Quantification of the amount of alpha phase in these samples was also attempted using Rietveld analysis (Table 4). As can be seen, the measurements from XRD closely match with the EBSD measurements.
Table 3. Results of Vickers hardness testing on monolithic samples (based on 10 measurements in each case). Hardness (HV)
Sample Consolidated at 750°C (before aging)
280 ± 6
Consolidated at 750°C (after aging)
325 ± 7
Consolidated at 800°C (before aging)
240 ± 6
Consolidated at 800°C (after aging)
410 ± 5
Fig. 6. EBSD phase maps of alloy Ti-5-3 monolithic samples (before aging) spark plasma consolidated at 750C (a) and 800C (b). Table 4. Amount of alpha phase in Ti-15-3 monolithic samples as determined from EBSD and XRD. Amount of alpha phase Sample EBSD
XRD
750°C consolidated, before aging
20
19
750°C consolidated, after aging
32
36
800°C consolidated, before aging
0
0
800°C consolidated, after aging
41
40
Fig. 7. EBSD phase maps of alloy Ti-5-3 monolithic samples (after aging) spark plasma consolidated at 750C (a) and 800C (b).
Fig. 8. X-ray diffractograms of Ti-15-3 monolithic samples (before aging) spark plasma consolidated at 750C and 800C.
Fig. 9. X-ray diffractograms of Ti-15-3 monolithic samples (after aging) spark plasma consolidated at 750C and 800C. TEM studies were also carried out on Ti-15-3 monolithic samples spark plasma consolidated at 750 ºC and 800 ºC in as-processed and aged conditions. Representative TEM micrographs are shown in Fig. 10 and Fig. 11. The presence of alpha and beta phases was confirmed from selected area diffraction studies. In as-processed condition, the samples consolidated at 750 ºC showed a large number of coarse alpha plates (Fig. 10a). The alpha plates were found to coarsen further after aging (Fig. 10b). Importantly, these samples showed hardly any fine alpha platelets. In contrast, the samples consolidated at 800ºC showed no alpha phase in as-processed condition (Fig. 11a), but showed numerous fine alpha platelets after aging (Fig. 11b). These observations corroborate well with EBSD and XRD results.
Based on the above investigations, microstructure evolution in metastable beta titanium alloy Ti-15-3 during SPC at 750 ºC and 800 ºC can be understood as follows. During consolidation at 750 ºC, because the material is heated to a temperature high in the alpha +
beta phase field and held for sufficient time, considerable precipitation of coarse alpha plates takes place. When these samples are subsequently aged at 530 ºC, some additional alpha precipitation occurs. This additional alpha precipitation involves coarsening of existing alpha plates rather than formation of any new fine alpha platelets. This explains why the samples consolidated at 750 ºC did not pick up much hardness after aging. Whereas, when the consolidation temperature is 800 ºC, because the temperature is above the beta-transus, no alpha phase can form during SPC. Thus, in as-processed condition, the microstructure of the material is just supersaturated metastable beta (similar to a properly solution treated material). Upon aging, these samples develop optimum strengthening precipitation (fine alpha platelets) and satisfactory hardness.
Fig. 10. TEM micrographs of Ti-15-3 monolithic samples spark plasma consolidated at 750 °C: (a) before aging, (b) after aging. (c) and (d) show the selected area diffraction patterns obtained from alpha and beta phases, as marked in (a).
Fig. 11. TEM micrographs of Ti-15-3 monolithic samples spark plasma consolidated at 800 °C: (a) before aging, (b) after aging. (c) and (d) show the selected area diffraction patterns obtained from alpha and beta phases, as marked in (b). From the above results, it can be seen that for achieving optimum strengthening precipitation and best possible mechanical properties in spark plasma consolidated metastable beta alloy Ti-15-3 monolithic samples, it is necessary to avoid formation of excessive coarse alpha plates by: (i) processing the samples above the beta-transus temperature, and (ii) cooling the samples rapidly enough from the processing temperature to room temperature. During subsequent aging at 530 ºC, fine alpha platelets precipitate in the beta matrix, leading to significant strengthening. Processing above the beta-transus temperature is also important for avoiding the detrimental grain boundary alpha formation in this alloy. Based on these considerations, a consolidation temperature of 800 C was considered optimum for alloy Ti15-3.
3.3 Spark plasma consolidated composite samples
The SiCf/Ti-15-3 composite samples consolidated under different conditions were examined under optical and scanning electron microscopes. During consolidation, the titanium foils undergo extensive plastic deformation resulting in metal flow around the fiber and is assisted by diffusional creep and/or dislocation creep mechanisms [16-18]. At hightemperatures and low-stress levels diffusional creep dominates. In contrast, at high stress levels, dislocation creep dominates [18, 19]. As densification progresses, grain boundary diffusion facilitates boundary migration and void shrinkage at the fiber/matrix interfaces [17, 19]. The time of consolidation mainly depends on the temperature and the applied pressure. Apart from the effects of heat and pressure, the electric field can assist in foil bonding and fiber embedment by promoting mass transport [20]. The effects of electric field can be evaluated from the electromigration theory. However, in a typical field-activated sintering process like spark plasma sintering, the temperature and current are not independent parameters and hence the thermal effects of the current (Joule heating) cannot be unambiguously separated from the intrinsic field effects [21].
In composite samples consolidated at both 750 ºC and 800 ºC using an applied pressure of less than 40 MPa, the fiber embedment was found to be unsatisfactory (Fig. 12a). While the composite samples consolidated at 750C for 20 minutes using 40 MPa pressure showed satisfactory foil bonding and fiber embedment (Fig. 12b), considering the microstructural advantages in processing alloy Ti-15-3 above its beta-transus temperature, a consolidation temperature of 800 C was identified as the optimum consolidation temperature in combination with 40 MPa pressure and 20 minutes of consolidation time (Fig. 13). In an attempt to reduce the consolidation time further, SPC experiments were conducted by increasing the pressure to 50 MPa, which, however, was found to result in fiber breakage.
Fig. 12. Fiber embedment in spark plasma consolidated SiCf/Ti-15-3 composite samples: (a) 750C/32MPa/20minutes (incomplete fiber embedment), (b) 750C/40MPa/20minutes (satisfactory fiber embedment).
Fig. 13. SEM micrograph of a SiCf/Ti-15-3 composite sample spark plasma consolidated at 800 C. The results of tensile testing on optimally consolidated monolithic and composite samples after aging are summarised in Table 5. Typical stress-strain plots are shown in Fig. 14. The results show that the composite samples are stronger and stiffer as compared to the monolithic samples. The increase in strength and stiffness is commensurate with the fiber volume fraction. As per the rule of mixtures (σc = Vf σf + Vm σm, where Vf and Vm are the
volume fractions of fiber and matrix, respectively, and σc, σf, and σm are the strengths of composite, fiber, and matrix, respectively), the Young’s modulus and the yield strength of a SiCf/Ti-15-3 composite with 2.3 fiber vol.% should be 128 GPa and 1470 MPa, respectively (calculated using the tensile data obtained on alloy Ti-15-3 monolithic samples for the matrix properties). These values compare well with the test results obtained on the composite samples. This indicates that the SiC fibers in the SiCf/Ti-15-3 composite samples produced in the current study are serving to their full potential. As can be expected, the composite samples showed slightly lower tensile elongations compared to the monolithic samples. SEM examination of the fracture surfaces of monolithic and composite samples did not reveal any foil debonding or fiber pull-out (Fig. 15), confirming that the foil bonding and fiber embedment are indeed satisfactory in the SiCf/Ti-15-3 composite samples produced in the current study. The fracture surfaces of the SiC fibers were very much flat, but the titanium matrix showed ductile dimpled rupture features.
Table 5. Result of tensile testing on spark plasma consolidated monolithic and composite samples (based on six tests in each case). Young’s
Yield strength
Ultimate tensile
Elongation
modulus (GPa)
(MPa)
strength (MPa)
(%)
Monolithic
117 ± 5
1270 ± 10
1360 ± 5
7.5
Composite
136 ± 7
1480 ± 8
1510 ± 6
6
Sample
Fig.14. Stress-strain plots of spark plasma consolidated monolithic and composite samples.
400 µm
100 µm
Fig. 15. Fractographs of spark plasma consolidated monolithic (a) and composite (b) samples. Despite their relatively low fiber volume fraction, the tensile properties of the SiC f/Ti15-3 composite samples produced in this work compare favourably with the test data on similar composites reported in earlier investigations. For example, for a Ti-15-3 matrix composite with 21 vol. % SiC fibers produced under optimal conditions, Izui et al. [14, 15] reported a tensile strength of 1570 MPa in as-consolidated condition. Similarly, Gabb et al.
[22] reported a tensile strength of 1461 MPa for a heat-treated Ti-15-3 matrix composite (solutionized at 788 °C for 15 minutes, water quenched, and then aged at 300 °C for 24 hours) consisting of 37 vol. % SiC fibers. These comparisons clearly show that the matrix microstructure plays an important role in determining the composite properties. In the first case, as no heat treatment was attempted to induce alpha precipitation, the matrix is not as strong as that in the present composite samples. In the second case, the titanium matrix is significantly harder (500 HV) than that in the present composite samples (410 HV). It may be noted that the heat treatment employed by Gabb et al. [22] does not provide the optimum combination of strength and ductility in alloy Ti-15-3.
4. Conclusions
Spark plasma consolidation can be advantageously utilized for producing SiC f/Ti-153 composites via foil-fiber-foil method. Compared to conventional diffusion bonding or hot pressing, spark plasma consolidation can achieve satisfactory foil bonding and fiber embedment at significantly lower temperatures and times. Satisfactory consolidation of SiC fibers and Ti-15-3 foils was demonstrated at temperatures as low as 750 C in just 20 minutes. The reduced processing temperature and time with spark plasma consolidation translates to a very healthy fiber/matrix interface as well as significant cost savings.
The present study shows that a metastable beta titanium alloy is best processed above its beta-transus temperature. In the case of alloy Ti-15-3, consolidation at temperatures below its beta-transus temperature (~ 770 C) was found to result in formation of undesirable grain boundary alpha phase as well as intragranular precipitation of a large number of coarse alpha plates. Consequently, these samples could not develop satisfactory strengthening precipitation upon subsequent aging. On
the other hand, when alloy Ti-15-3 was consolidated at 800 C, a fully beta microstructure was obtained as the material was taken into the beta phase field during consolidation and was cooled sufficiently rapidly to room temperature to preclude any alpha precipitation. These samples developed satisfactory strengthening precipitation (in the form of fine alpha platelets) upon subsequent aging. For this reason, the current study identifies 800C as the optimum consolidation temperature for SiCf/Ti15-3 composites although spark plasma consolidation can facilitate satisfactory foil bonding and fiber embedment even at 750 C.
SiCf/Ti-15-3 composites produced in the current study with a fiber volume fraction of 2.3 vol. % using spark plasma consolidation showed satisfactory improvements in Young’s modulus (from 117 GPa to 136 GPa) and yield strength (1270 MPa to 1480 MPa) compared to monolithic Ti-15-3 samples. The composite properties compared well with the expectations from the rule of mixtures, indicating that the SiC fibers are serving to their full potential in the composite.
References [1] T. W.Clyne, P.J. Withers, An Introduction to Metal Matrix Composites, Cambridge University Press, Cambridge, 1993. [2] R. A. Mackay, P.K. Brindley, F.H. Froes, Continuous fiber-reinforced titanium aluminide composites Journal of the Minerals, The journal of the Minerals, Metals & Materials Society 1991, 43 (5), 23-29. [3] S.F. Bauman, P.K. Brindley, S.D. Smith, Reaction Zone Microstructure in a Ti3AI+Nb/SiC Composite, Metall. Trans. A 1990, 21A, 1559-1569. [4] D.B. Gundel, F.E. Wawner, Interfacial Reaction Kinetics of Coated SiC Fibers with Various Titanium Alloys, Scripta Metallurgica et Meterialia 1991, 25, 437-441.
[5] Y.G. Zhou, Y.Q. Yang, Progress in the study of titanium matrix compositesreinforced by SiC fibers, Journal of Acta Met Sinica 2002, 38 (suppl), 461-465. [6] C. Leyens, J. Hausmann, J. Kumpfert, Continuous fiber reinforced titanium matrix composites: Fabrication, properties and applications,, Advanced Engineering Materials 2003, 5(6), 399-410. [7] S. Djanarthany, J.C. Viala, J. Bouix, Development of SiC/TiAl composites: Processing and interfacial phenomena, Materials Science and Engineering A 2001, 300(1-2), 211-218. [8] Y.Q. Yang, Z.J. Ma, X.H. Lü, J. K. Li, Y. Chen, Y.L. Ai, Studies on Interface of SiCf/Ti-6Al-4V Composites, Rare Met Mater Eng. 2006, 35(10), 1516-1521. [9] Y.Q. Yang, X.H. Lü, X. Luo, Z.J. Ma, J.K. Li, Q. Wen, Y. Chen, Kinetics of Interfacial Reaction in SiCf/Ti6Al4V Composites, Materials Science Forum 2007, 546/549, 1627-1632. [10] Z.A. Munir, R.M. German, A Generalized Model for the Prediction of Periodic Trends in the Activation of Sintering of Refractory Metals, High Temperature Science 1977, 9, 275-
283. [11] X. Luo, Y. Yang, Y. Yu, X. Wang, B. Huang, Y. Chen, Effect of Mo coating on the interface and mechanical properties of SiC fiber reinforced Ti6Al4V composites, Materials Science and Engineering A 2012, 550, 286–292. [12] G. Lutjering, J.C. Williams, Titanium, Springer Science & Business Media, 2007. [13] Z.A. Munir, U. Anselmi-Tamburini, M. Ohyanagi, The effect of electric field and pressure on the synthesis and consolidation of materials: A review of the spark plasma sintering method, Journal of Materials Science 2006, 41(3), 763–777.
[14] H. Izui, S. Kinbara, M. Okano, Mechanical Properties of Ti-15-3 Alloy Reinforced With SiC Fibers by Spark Plasma Sintering, Proceedings of the 6th Pacific Rim Conference on Ceramics and Glass Technology 2006, Maui, Hawaii, 289-300 [15] H. Izui, S. Kinbara, M. Okano, Tensile and Fatigue Strengths of SCS-6/Ti-6Al-4V Titanium Matrix Composite by Spark Plasma Sintering, Journal of the Japan Society of Powder and Powder Metallurgy 2005, 52 (10), 722-729 [16] J.C.M. Hwang, R.W. Balluffi, On a possible temperature dependence of the activation energy for grain boundary diffusion in metals, Scr. Metall. 1978, 12, 709-714 [17] B. Derby, The dependence of grain size on stress during dynamic recrystallization, Acta Metallurgica Et Materialia 1991, 39(5), 955-962. [18] C.H Hamilton, N.E Paton, Super plastic forming of structural alloys, TMS, Warrendale, PA 1982. [19] M. Zhou M, F.P.E. Dunne, Mechanism-based constitutive equations for the superplastic behaviour of a titanium alloy, J. Strain Anal Eng Des. 1996, 31, 187-196. [20] Z.A. Munir, H. Schmalzried, The effect of external fields on mass-transport and defect-related phenomena, Journal of Materials Synthesis and processing 1993, Vol.1, 3-16. [21] H. Conrad, Electroplasticity in metals and ceramics, Materials Science and Engineering A 2000, 287(2), 276-287. [22] T.P. Gabb, J. Gayda, B. A. Lerch, G. R. Halford. The Effect of Matrix Mechanical Properties on [0]8 Unidirectional SiC/Ti Composite Fatigue Resistance, Scripta Metallurgica et Materialia, 1991, 25: 2879-2884.