Wear 319 (2014) 160–171
Contents lists available at ScienceDirect
Wear journal homepage: www.elsevier.com/locate/wear
Microstructure-hardness-fretting wear resistance correlation in ultrafine grained Cu–TiB2–Pb composites Amit Siddharth Sharma a, Krishanu Biswas a, Bikramjit Basu a,b,n a b
Department of Materials Science and Engineering, Indian Institute of Technology, Kanpur-208016, UP, India Materials Research Center, Indian Institute of Science, Bangalore-560012, Karnataka, India
art ic l e i nf o
a b s t r a c t
Article history: Received 3 April 2014 Received in revised form 16 July 2014 Accepted 16 July 2014 Available online 1 August 2014
The retention of the desired combination of mechanical/tribological properties in ultrafine grained materials presents important challenges in the field of bulk metallic composites. In order to address this aspect, the present work demonstrates how one can achieve a good combination of hardness and wear resistance in Cu–Pb–TiB2 composites, consolidated by spark plasma sintering at low temperatures ( o500 1C). Transmission electron microscope (TEM) studies reveal ultrafine grains of Cu (100–400 nm) with coarser TiB2 particles (1–2 μm) along with fine scale Pb dispersoid at triple junctions or at the grain boundaries of Cu. Importantly, a high hardness of around 2.2 GPa and relative density of close to 90% relative density (ρtheo) have been achieved for Cu–15 wt% TiB2–10 wt% Pb composite. Such property combination has never been reported for any Cu-based nanocomposite, by conventional processing route. In reference to the tribological performance, fretting wear tests were conducted on the sintered nanocomposites and a good combination of steady state COF (0.6–0.7) and wear rate (10–4 mm3/N m) were measured. An inverse relationship between wear rate and hardness was recorded and this commensurates well with Archard’s relationship of abrasive wear. The formation of a wear–resistant delaminated tribolayer consisting of TiB2 particles and ultrafine oxide debris, (Cu, Fe, Ti)xOy as confirmed from subsurface imaging using focused ion beam microscopy has been identified as the key factors for the low wear rate of these composites. & 2014 Elsevier B.V. All rights reserved.
Keywords: Fretting Metal matrix composite Hardness Bearings Profilometry
1. Introduction The development of materials for bearing applications at higher operating temperatures requires continuous efforts on improving the structural stability and wear resistance. Cu-based metal– matrix composites (MMCs), formed with harder titanium diboride (TiB2) dispersoids is considered as a suitable candidate material for these applications [1–4]. Unlike precipitation-strengthened Cualloys (e.g. Cu–Cr and Cu–Zr), Cu–TiB2 composites are investigated in the present work. The selection of TiB2 from a wide range of available ceramics is also supported by its superior mechanical and physical properties, such as hardness (25 GPa), elastic modulus (565 GPa), high melting point (43000 1C), thermal (96 W m K), electrical ( 7.7 106 (Ω-m) 1) conductivity [5–7] and low density (4.5 g/cm3). In addition to the above mentioned approach for strength enhancement by dispersion strengthening, the Cu-matrix can also n Corresponding author at: Department of Materials Science and Engineering, Indian Institute of Technology, Kanpur-208016, UP, India. Tel.: þ 91 80 2293 3256; fax: þ91 80 2360 7316. E-mail addresses:
[email protected],
[email protected] (B. Basu).
http://dx.doi.org/10.1016/j.wear.2014.07.014 0043-1648/& 2014 Elsevier B.V. All rights reserved.
be strengthened by restricting the matrix grains to nanometric (o100 nm) or ultrafine (100–400 nm) size range. This would facilitate the deformation mechanisms, mediated by grain boundary process such as grain boundary dislocation emission [8–10], grain boundary sliding [11,12] and grain boundary migration [13–15]. The plastic deformation, in the case of nanocrystalline materials, is different from conventional intragranular dislocation mechanisms and is embodied mathematically in the celebrated Hall–Petch expression [16,17]. In addition, the matrix grains can further be strengthened by twins generated within the nanocrystalline or ultrafine grains during deformation. Various attempts to process Cu–TiB2 based composites are summarized in Table 1 and has been described elsewhere also [3,18–25]. Though in the existing open literature, diffusion data and related kinetics of Cu and Pb in TiB2 or vice versa are not available, it is perceived that TiB2 will act as a non-sintering agent in the absence of any suitable binder in the Cu-matrix. Since TiB2 has a melting temperature of higher than 3000 1C, sintering of Cu–TiB2 at lower temperatures will lead to poor densification. On the other hand, sintering at temperatures greater than 450 1C or so will nullify the advantageous effect of nanocrystalline Cu grains as they are reported to grow abnormally beyond 500 1C [26]. Thus, in
A. Siddharth Sharma et al. / Wear 319 (2014) 160–171
161
Table 1 Summary of literature results on Cu–TiB2 alloys, processed via various manufacturing routes, process parameters and comparison with present results. Material composition
Processing route
Processing parameters
Hardness
References
Cu–10 wt% TiB2 Cu–1.6 wt% TiB2 Cu–2.5 wt% TiB2 Cu–30 wt% TiB2 Cu–1 and 2 vol% TiB2
Ball milling, cold compaction and annealing In-situ reaction with rapid solidification
T ¼ 600 1C T ¼ 1300–1400 1C
[18]
In-situ combustion synthesis Mechanical alloying and hot pressing
– Hv ¼ 1.4 GPa Hv ¼ 1.7 GPa HRC ¼ 48 HRb ¼ 78.7
Cu–15 vol% TiB2 Cu–2 wt% TiB2 Cu–48 vol% TiB2
Ball milling, cold compaction and reactive hot pressing Mechanical alloying and hot pressing Cold compaction and hot pressing
T ¼ 1903 1C at 33 MPa pressure Mixing for 12 and 36 h, hot pressing at T ¼ 650 1C at 90 MPa pressure with 2 h soaking T ¼ 850 and 900 1C – T ¼ 800 to 880 1C at 50 MPa pressure – T ¼ 950 1C at 116 MPa pressure applied for HB ¼2187 10 25 min
Cu–57 vol% TiB2
Ball milling, self propagating exothermic reaction and spark plasma sintering
Cu–10 wt% TiB2 Cu–10 wt% TiB2– 10 wt% Pb
Ball milling and spark plasma sintering
order to obtain sufficiently high density and simultaneously retaining nanocrystalline grains, liquid phase sintering can be a viable option. This can be pursued by adding a low melting point constituent such as Pb, Sn etc. In order to achieve better hardness and wear resistance, composites were consolidated at lower temperatures and pressures. The intent was to produce a matrix with ultrafine grains and a uniform distribution of phases. It is in this perspective, an attempt has been made in the present investigation to incorporate 10 and 15 wt% of TiB2 in the nanocrystalline Cu-matrix with simultaneous addition of 10 wt% Pb to aid the sintering process and to improve the achievable density by selecting Spark Plasma Sintering (SPS) as a processing route. The emphasis has been laid out in this study to understand the sintering of Cu–TiB2 composites with and without Pb addition and to look into the wear behavior of these composites by correlating the microstructure with properties, notably friction, wear resistance and hardness. For the sake of brevity, wt% will be omitted in the compositional notations for e.g. Cu–15 wt% TiB2–10 wt% Pb will be denoted as Cu–15TiB2–10Pb and likewise.
SPS temp. (1C)
Holding time (min)
SPS pressure (MPa)
Hv (GPa)
700 800 950 950 950 950 950
5 5 0 5 0 5 30
70 70 50 50 70 70 70
2.3 3.3 5.7 – – – 6.6
T ¼ 400, 500, 600 and 700 1C at 50 MPa with 5 min soaking
[3]
[19] [20] [21] [22] [23]
[24]
1.5–2 GPa [25]
uniaxial pressure of 50 MPa at a heating rate of 90 1C/min to the final sintering temperature with a soak time of 5 min. During the entire heating and cooling cycle, inert Ar atmosphere was maintained by continuous purging at a flow rate of 2 l/min. 2.2. Characterization 2.2.1. Density, phase analysis and hardness The density of sintered samples was determined according to Archimedes’ principle using distilled water and calculated following the rule of mixtures, considering the theoretical densities of Cu, TiB2 and Pb as 8.96, 4.50 and 11.68 g/cm3, respectively. The phase identification of the powders and sintered samples were performed by X-ray diffraction (XRD) (Bruker D8 Focus, Germany) using Cu-Kα (λ ¼0.154056 nm) radiation generated at 40 kV and 40 mA. The samples were scanned at a scan rate of 0.51/min in the angular range of 2θ ¼251 to 1001 with a step size of 0.051. Vickers hardness was measured using microhardness tester (Bareiss Prüfgerätebau, Germany) as per ASTM E92-82 standards by applying a load of 200 g with a dwell time of 10 s. At least 10 measurements have been carried out for each sample.
2. Experimental details 2.1. Processing For the composite preparation, high purity Cu (Average Particle Size (APS) o100 nm, 99.7% purity, Sigma-Aldrich, USA), Pb (APS 10 μm, 99.98% purity, Sigma-Aldrich) and TiB2 (APS 3 μm, Japan New Metals Co. Ltd., Japan) powders were used as the starting materials. To remove oxides, Cu powder was reduced in flowing 5% H2/95% He gas mixture at a flow rate of more than 15 cm3/min and pressure of 300 mbar in a tubular furnace (P330, Nabertherm, Germany). The powder mix, taken in appropriate ratio, was milled in a tungsten carbide (WC) planetary micro-mill (Fritsch Pulverisette 5, Germany) for 6 h at 100 rpm with a charge ratio of 10:1. High purity argon was replenished after every 30 min in the sealed jars to avoid oxidation. Spark Plasma Sintering (SPS) of powders was carried out at temperatures 350 and 400 1C with
2.2.2. Fretting wear test The tribological behavior of the composites against spherical steel counterbody (SAE 52100 grade) was studied using a fretting tribometer (TR281M, DUCOM, Bangalore, India) at a normal load of 10 N, with a relative displacement stroke of 100 μm and a frequency of 5 Hz for 60,000 cycles. Out of three fretting modes, mode I has been employed for the present study which is characterized by the linear reciprocatory sliding motion, whereas mode II and mode III are the radial and circumferential modes, respectively. The fretting test was conducted 5 times for each sample and the average value of the measurements is reported with error bars. A complete and detailed description about the fretting methodology, tests and analysis can be found elsewhere [27]. As there is no well established empirical formula or model for making the selection but still the parameters were so chosen to satisfy the conditions in such a manner that the reinforcement
162
A. Siddharth Sharma et al. / Wear 319 (2014) 160–171
phases (both harder (TiB2) and softer (Pb) in the present case) should separately adhere to the intended role with which these were added to the Cu matrix. In case of dispersoid-reinforced metal matrix composites intended for tribological applications, the friction and wear behavior of the composites containing harder or softer particles than the matrix are widely different because of different wear mechanisms. The normal load was selected such that the load should transfer to the harder TiB2 particles but at the same time any preferential formation of Pb film onto the wear scar can be avoided. Sliding amplitude on either side of the mean position was kept at 100 μm to facilitate conditions of ‘stick’ only without any occurrence of ‘slip’. Frequency was kept low at 5 Hz so that the effect of the strain-induced fracture effects mainly at the interfacial region between different phases owing to lattice mismatch should be as minimal as possible. Fretting cycles was maximized after the conditions of steady state coefficient of friction have been achieved without any sharp fluctuations. 2.2.3. Examination of microstructure, worn surfaces and quantification of wear parameters 2.2.3.1. Electron microscopy. The metallographically polished samples and topographical features of the wear scars and debris were investigated using scanning electron microscope (SEM) (Carl Zeiss EVO50, Germany) equipped with energy dispersive X-ray spectroscopy (EDS) (Oxford Instruments). Details about the transmission electron microscopy (TEM) and related sample preparation can be found elsewhere [26]. 2.2.3.2. Ion beam microscopy. Subsurface features underneath the worn surface of the samples were investigated using focused ion beam (FIB) microscopy. The protocol followed for milling and polishing the trench has been reported elsewhere [28]. 2.2.3.3. Profilometry. The geometry i.e. three dimensional shape (extent and maximum depth) of the wear scars was mapped using a laser surface profilometer (Perthometer PGK 120, Mahr, Germany). Details about the profilometry can be found elsewhere [28]. From the calculated wear volume (V), wear rate (WR) can be calculated as per ASTM G40-13 standards as follows: WR ¼
V ðN c 2dÞ
Fig. 1. XRD patterns of (a) Cu–10TiB2 and (b) Cu–10TiB2–10Pb sintered composites. Standard ICDD peaks are also shown (bottom) corresponding to each phase.
ð1Þ
where N is the normal load, c are the number of cycles and d is the stroke length. A factor of 2 appears with stroke length equalizing the one full fretting cycle. The wear resistance is defined as the inverse of the wear rate.
3. Results 3.1. Phase analysis Fig. 1(a and b) shows the diffractograms obtained from the polished surfaces of Cu–10TiB2 and Cu–10TiB2–10Pb composites, respectively. The peak positions corresponding to each individual phase from the ICDDs (International Centre for Diffraction Data) database and the corresponding data card number are also shown. It is to be noted that the plotted ICDDs peaks are only symbolic positional representation of the phases i.e. without any sign of peak broadening or strain effects. The Cu peaks appear to be broad and the broadness of each peak does not change substantially even after sintering. The peaks corresponding to TiB2 and Pb can be clearly discerned. In case of unleaded samples (Cu–10TiB2), low intensity TiO2 peaks can be detected (at 2θ ¼35.17, 36.47 and 38.921), whereas
Fig. 2. BSE–SEM micrographs of sintered composites: (a) Cu–10TiB2, SPS at 400 1C and (b) Cu–10TiB2–10Pb, SPS at 350 1C.
A. Siddharth Sharma et al. / Wear 319 (2014) 160–171
163
for Pb–containing samples (Cu–10TiB2–10Pb), a low intensity Cu2O peak at 2θ ¼29.121 can only be detected. 3.2. Microstructural characterization 3.2.1. Scanning electron microscopy (SEM) Fig. 2(a and b) shows the representative back scattered electron (BSE–SEM) micrographs of Cu–10TiB2, sintered at 400 1C and Cu– 10TiB2–10Pb, sintered at 350 1C composites, respectively. It can be noted that light gray contrast is for Cu-matrix, dark gray is for TiB2 (white arrows), whereas bright phase is the Pb (black arrows). Homogeneous and uniform distribution of TiB2 and Pb in the Cumatrix is the common feature of both the composites. The inset in Fig. 2b shows high magnification image, revealing a few TiB2 particles (marked with white dotted lines). These lines (drawn offset) indicate the positions of the cracks, which developed in the TiB2 particles during the course of ball-milling (pre-fracturing process). Upon close inspection of Fig. 2b, the distribution of still finer TiB2 particles can also be noticed in the composites, which might have occurred by the fracturing and subsequent detachment of corners in the course of ball-milling. Pb particles (m.p. 327.5 1C), being molten at sintering temperatures (350 and 400 1C) can be traced as contours (Fig. 2b). Pb, being immiscible, delineates the Cu grain boundaries, as indicated by white contours in Fig. 2b (marked by black arrows). The porosities present in the microstructures are also marked as white circle. 3.2.2. Transmission electron microscopy (TEM) As part of finer scale microstructural analysis, some representative results from the transmission electron microscopic (TEM) analysis of Cu–15TiB2–10Pb composite are shown in Figs. 3 and 4. Figs. 3 and 4 present the results of the sample sintered at 350 and 400 1C, respectively. Extensive imaging was performed to find out the grain size distribution and deformation structure of Cu and TiB2 during the processing, whereas selected area diffraction (SAD) patterns have been acquired to confirm the presence of different phases. 3.2.2.1. Morphology and size distribution of Cu grains. Fig. 3c and Fig. 4d present the Cu grain size distribution of the samples, sintered at temperatures of 350 and 400 1C, respectively. Two important inferences can be drawn from these plots. First, the overall distribution is quite broad with the presence of grains smaller than 100 nm as well as relatively coarse grains ranging from 100 to 500 nm, observed at both the sintering temperatures. The wider grain size distribution is more evident for sample sintered at 400 1C. The mean grain size increases from 283 7 140 nm to 364 7187 nm as the sintering temperature is increased from 350 to 400 1C. Added to this, the different morphologies of the Cu grains, equiaxed to elongated are observed. Some of the Cu grains showed anisotropic growth with tabular morphology (see Fig. 3b and Fig. 4b). These grains might have some molten Pb (for e.g. as marked in Fig. 4 by white arrows) in their vicinity and have attained elongated morphology with deformed substructure. 3.2.2.2. Deformation substructure in the Cu and TiB2 grains. The detailed microstructural characterization reveals the signature of deformation in the Cu grains. The deformation substructure is characterized by the presence of dislocation (Fig. 3b) and twins within Cu grains (Fig. 3a). Many Cu grains show the presence of a large number of dislocations and dislocation cell structures (Fig. 4a). Comparatively elongated grains reveal more deformed substructure than the equiaxed grains, indicating extensive deformation of the elongated grains. Even harder TiB2 grains
Fig. 3. TEM micrographs of Cu–15TiB2–10Pb composite processed at sintering temperature of 350 1C and 50 MPa: (a) a representative microstructure; (b) Cu grain with twin formation and elongated morphology and (c) Cu grain-size distribution.
show deformation structure. Fig. 4a shows one of the TiB2 grains having undergone deformation with well developed dislocation cell substructure (marked by white dotted lines). This activity may lead to the fracturing of specific grains and could explain the occurrence of a wide grain size distribution, as shown in Fig. 3c and Fig. 4d. 3.2.2.3. Dispersion of Pb. The observation of distinct Pb phase in the vicinity of Cu grain boundaries is consistent with the immiscibility of Cu and TiB2 with Pb at 350 and 400 1C. However, the encapsulated Pb
164
A. Siddharth Sharma et al. / Wear 319 (2014) 160–171
Fig. 4. TEM micrographs of Cu–15TiB2–10Pb composite processed at sintering temperature of 400 1C and 50 MPa: (a) a typical ensemble of phases with individual phases marked along with corresponding SADs; (b and c) Cu grains with intergranular dislocation structure and (d) Cu grain-size distribution. White arrows mark the Pb position in the matrix. Dotted line shows the cellular substructure in one TiB2 grain. Inset in (b and c) shows the SAD acquired from the grains marked with þsign.
A. Siddharth Sharma et al. / Wear 319 (2014) 160–171
165
Table 2 Density, hardness and wear rate data for all the composites viz. Cu–10TiB2, Cu–10TiB2–10Pb, Cu–15TiB2 and Cu–15TiB2–10Pb. Sample composition (wt%)
Sintering temperature (1C)
Relative density (% theoretical density)
Vickers hardness (Hv0.2, GPa)
Mean COF
Wear Rate ( 10 4 mm3/N m)
Cu–10TiB2
350 400 350 400 350 400 350 400
76.1 70.7 83.6 70.9 84.5 71.6 88.7 70.2 77.2 70.6 79.5 70.9 85.3 72.5 89.5 74.4
0.89 7 0.08 0.92 7 0.14 1.65 7 0.12 1.88 7 0.08 1.85 7 0.06 2.03 7 0.10 1.90 7 0.05 2.20 7 0.09
0.687 0.04 0.647 0.04 0.647 0.01 0.60 7 0.03 0.647 0.06 0.62 7 0.03 0.667 0.04 0.65 7 0.03
2.147 0.38 1.69 7 0.34 0.96 7 0.17 0.75 7 0.016 1.20 7 0.27 0.36 7 0.09 0.65 7 0.08 0.447 0.07
Cu–10TiB2–10Pb Cu–15TiB2 Cu–15TiB2–10Pb
Fig. 5. Variation of wear rate with hardness for the sintered composites.
particle in the vicinity of triple junction or at the grain boundary (GB) region of Cu grains has been the most common morphology of Pb, as marked by white arrows in Fig. 4(a–c). Similar microstructural features have been observed for the other Cu–xTiB2 and Cu–xTiB2–10Pb (x¼10 and 15 wt%) composites. 3.3. Hardness properties Table 2 shows the hardness variation of all the samples, sintered at a pressure of 50 MPa and at temperatures of 350 and 400 1C. Overall, the hardness varies in the range of 0.8–2.2 GPa for all samples sintered at temperatures of 350 and 400 1C and also for each composition, the composites sintered at 400 1C show higher hardness than those sintered at 350 1C. The composites with 15 wt% TiB2 yield higher hardness than those with 10 wt% TiB2. Moreover, the trend in hardness variation for each category of composite complements the corresponding variation in relative density for the sintered composites. For example, hardness and relative density values for Cu–10TiB2 composite at 350 1C are higher in comparison to the corresponding values for the same composite sintered at 400 1C. 3.4. Tribological behavior 3.4.1. Wear data Fig. 5 presents the plot of wear rate data vs. hardness for all the composites viz. Cu–10TiB2, Cu–10TiB2–10Pb, Cu–15TiB2 and Cu– 15TiB2–10Pb. The overall wear rates for all the composites are found to vary in the range of 0.3–2.1 10 4 mm3/N m. A comparison of wear rates between these four set of composites on the basis of Pb as well as TiB2 content leads to the conclusion that the addition of 10 wt% Pb and an increase of TiB2 content from 10 to 15 wt% assists in lowering the wear rates. The overall trend of data
points, shown by black dotted line in Fig. 5, is in accordance with the Archard’s law (Q¼kW/H), stating that wear rate (Q) is inversely proportional to the hardness (H) of the softer constituent (composite flat) in a tribosystem. Here, Q is the volume removed per unit distance per unit load, W the applied normal load and k the wear coefficient. As shown in Fig. 5, a systematic decrease in wear rate with an increase in hardness in the range of 0.8–2.2 GPa clearly signifies the predominance of abrasive wear. Table 2 provides the relative densities, hardness and wear rate data for all the composites. A closer look at Table 2 does not reveal any difference in steady state coefficient of friction. Therefore, the addition of Pb does not have any marked influence on the frictional behavior under the operating conditions. However, the Pb containing composites distinctly show low wear rate as compared to the unleaded composites.
3.4.2. Morphology of worn surfaces Detailed microstructural investigation of the wear scars for the selected composites was carried out to understand the mechanisms governing the wear process. For all the composites, the worn scar is encircled by white circle along its periphery to clearly distinguish it from the unaffected region. The direction of the fretting wear is also indicated by a double headed black arrow on the figure. The different elemental peaks in the EDS spectra were identified and labeled on the basis of corresponding photon energy (in keV) as follows: Cu K-8.630, Cu L-1.012, Pb L-10.550, Pb M-2.342, Fe K-6.398, Fe L-0.705, Ti K-4.508, Ti L-0.452 and O K0.525. Fig. 6(a and b) presents the worn scar morphology of Cu– 10TiB2–10Pb composite, sintered at 350 1C. Fig. 6a provides the extent of the worn scar with high magnification micrographs of the regions 1 and 2 marked on the worn scar. Elemental analysis of region 2 performed by EDS is also shown in the inset. Apart from elemental Cu, Pb and Ti peaks, the strong peaks due to Fe and O have also been detected. Fig. 6b is the micrographs of the worn debris produced after the wear process. From the low magnification image (left), it is clear that the debris resemble that of fine chips with each chip as an aggregate of ultrafine particles, evident from the high magnification micrograph (Fig. 6b, right). Fig. 7(a and b) presents the micrographs of the worn surface and the debris of Cu–15TiB2 composite, sintered at 350 1C. The marked regions 1 and 2 from the centre of the worn scar region have been further magnified and they show appreciable formation of porosity beneath the worn scar region. Because of the inherent porosity and loose interfacial bonding between Cu and TiB2 grains, the TiB2 particles are more prone to be knocked out from their original microstructural sites, leaving behind soft Cu matrix against the harder steel counterbody (Hv 8 GPa). Further, the now dislodged TiB2 particles, acting as third-body abrasive
166
A. Siddharth Sharma et al. / Wear 319 (2014) 160–171
Fig. 6. BSE–SEM micrographs of Cu–10TiB2–10Pb composite, SPS at 350 1C showing (a) worn scar region (encircled white) with regions 1 and 2 further imaged at high magnification and (b) wear debris.
particles will start abrading the flat surface as well as softer steel counterbody. This hypothesis about the sequence of wear mechanism is confirmed by the micrographs and EDS analysis of the debris (Fig. 7b), revealing peaks of elemental Ti, Fe and O. Fig. 8(a and b) shows BSE–SEM micrographs of the worn surface on Cu–15TiB2–10Pb composite, sintered at 350 1C. Unlike previous samples, it appears that the wear induced damage is not accompanied by any extensive cracking and/or knocking out of the TiB2 particles. One region with delaminated tribolayer is highlighted (Fig. 8a, left view) and is further magnified (Fig. 8a, right view). The ultrafine particles as wear debris are also observed (Fig. 8b).
3.5. Subsurface observation of the worn surface The FIB analysis to investigate the subsurface damage was carried out on the selected samples. Fig. 9 presents the FIB analysis of the worn surface of Cu–15TiB2–10Pb sample, sintered at 400 1C and fretted at a load of 10 N. Fig. 9a shows an overview of the milled trench along with delaminated tribolayer. Some discrete debris particles are also observed, which might have formed as a result of fragmentation from mechanically mixed tribolayer. The wall marked by the white-dotted line was polished for the final analysis. Fig. 9(b–d) shows the subsurface region, viewed at an angle of 301 with respect to the milling
A. Siddharth Sharma et al. / Wear 319 (2014) 160–171
167
Fig. 7. BSE–SEM micrographs of Cu–15TiB2 composite, SPS at 350 1C showing (a) worn scar region (encircled white) and (b) wear debris. In (a), regions 1 and 2 in the worn scar have been further imaged at high magnification.
direction of FIB. Fig. 9(a and b) has been imaged using SE mode to reveal the crack propagation, whereas Fig. 9(c and d) is obtained in the BSE mode for differentiating the morphological changes of constituent phases. In Fig. 9b, the highlighted area in the immediate vicinity of the worn surface shows subsurface region, exhibiting a layered structure extending to not more than 1 μm from the surface. This plastic deformation might presumably have occurred due to the longitudinal relative motion between counterbody and sample, resulting in the formation of laminated structure as a consequence of thermal softening. From Fig. 9c, one can clearly observe the presence of the elongated Pb particles (continuous white arrows), which are found to
have undergone extensive deformation by the wear process. The presence of some microcracks (dashed white arrows) in the subsurface zone can also be observed. Fig. 9d presents a collate micrograph depicting the behavior of TiB2 particles in the dynamics of the wear phenomena, which is mainly governed by the events occurring at the interfacial region of the constituent phases. It is anticipated that TiB2, being the hardest among the phases present, will not undergo plastic deformation due to subsurface tribological events and therefore, its effect on the wear process will primarily be dictated by the dynamics of crack nucleation and propagation. An appraisal will, therefore, be presented in the discussion section on the wear mechanisms of these composites.
168
A. Siddharth Sharma et al. / Wear 319 (2014) 160–171
Fig. 8. BSE–SEM micrographs of Cu–15TiB2–10Pb composite, SPS at 350 1C showing (a) worn scar region (encircled white) and (b) wear debris. In (a), squared region (left image) is further high magnified (right image).
In Fig. 9d, the origin of the in situ developed lateral cracks and the cavities generated beneath the worn surface, marked by dotted arrows and circles, respectively will be rationalized from the point of view of the thermal and mechanical stimulus imparted by the wear process. Their consequence on the wear resistance of these composites will further be correlated with the tribological properties.
4. Discussion As preclude to this study, the present authors were able to achieve densities close to 90% of ρtheo for Cu–10 wt% Pb nanocomposite without any Pb segregation [26] as well as for Cu–10 wt % TiB2 composites with uniform distribution of the phases [25]. This type of microstructure is desirable from tribological application point of view. It is however to be noted that although extensive scientific studies have been carried out on the sliding wear behavior of TiB2 coatings on different substrates, the studies on monolithic Cu–TiB2 composites is very limited (see Table 3). Though a direct correlation or analogy in wear behavior cannot be drawn between TiB2 deposited as coatings and TiB2-containing bulk composites, but the advantages offered by a processing route on the basis of properties achieved can be made. Nevertheless, only those studies on TiB2 coatings have been cited in which the coating thickness is greater than the depth created by the fretting wear in the present study, which is typically o10 μm (refer to Table 3). Investigations concerning TiB2 deposited as coatings on steel or aluminum substrates have reported the hardness values close to 12 GPa (with 100% TiB2) [29], 9 GPa (with 70% TiB2) and 2.5 GPa (with 30% TiB2) [30]. While Darabara et al. [29] reported a lower wear rate of 0.1 10 3 mm3/N m for TiB2-coated steel substrate, Xu [30]
obtained comparatively higher wear rate values of 19 and 230 ( 10 3 mm3/N m) corresponding to 70 and 30 wt% TiB2 contents, respectively. Kirk et al. [31] electrodeposited TiB2 coatings on Ni and Mo substrates and reported the lowest wear rates of 1–3 10 6 mm3/N m at a load of 4.9 N amongst all the cited studies. In another study, Agarwal et al. [32] refined the microstructure of the TiB2 coatings via pulse engineering and laser treatment and measured wear rates of 1159 and 463 ( 10 3 mm3/N m), respectively. Yust et al. [33] did the surface modification via ion implantation of the sintered monolithic TiB2 and reported the lowest COF of 0.02 against diamond. Tu et al. [34] investigated the sliding wear behavior of bulk Cu–TiB2 alloys with varying content (0, 0.5, 1.5 and 2.5 wt%) of TiB2 nanodispersoids in the Cu-matrix. They found an increase in hardness with TiB2 content and values achieved in their study (0.75–1.22 GPa) are inferior to that achieved in the present case (0.89–2.2 GPa). Importantly, a maximum hardness of 1.2 GPa and corresponding wear rate of 7.5 10 3 mm3/N m were measured with Cu–2.5 wt% of TiB2 composites. In the above perspective, the wear rate of the presently investigated Cu–TiB2–Pb nanocomposites is found to be at least one order of magnitude lower (10 4 mm3/N m) than majority of the earlier study and this can be correlated with the ultrafine-grained Cu matrix with uniformly distributed TiB2 and Pb phases. In order to construct an in-depth understanding of the microstructure-property correlation, an appraisal of the rapid densification kinetics, high hardness, subsurface deformation, role of TiB2 and Pb on wear mechanisms and low wear rate will be discussed in this section in reference to the following aspects: a) Achieved high hardness: Ultrafine grained Cu-matrix, effect of TiB2 addition.
A. Siddharth Sharma et al. / Wear 319 (2014) 160–171
169
Fig. 9. FIB and SEM micrographs of worn subsurface region of Cu–15TiB2–10Pb composite, SPS at 400 1C. (a) SE–SEM micrograph of the milled trench (z-axis perpendicular to the surface and white dashed line marks the boundary between surface and trench wall), (b) SE–SEM micrograph of the immediate subsurface layers, (c) BSE–SEM micrograph showing formation of subsurface cracks (dotted arrows) and Pb particles (continuous arrows) and (d) BSE–SEM micrograph showing debonding and fracturing of TiB2 particles.
b) Lower wear rate: Crack nucleation, debonding of TiB2 and mutual solubility between elemental constituents. a) During real life applications, bearings are always in closed contact with the moving machinery. In order to enhance wear resistance, the mechanical property, in particular, the overall bulk hardness of the material is important. As far as the Cu-based bearing alloys with metallic reinforcement are concerned, achieving hardness values of more than 1 GPa has always been the limiting factor [37]. In an attempt to circumvent this aspect, an addition of 10 and 15 wt% TiB2 to the ultrafine-grained Cu-matrix was made in the present work and the relative density and hardness values for Cu–10TiB2, Cu–10TiB2–10Pb, Cu–15TiB2 and Cu–15TiB2–10Pb composites sintered at temperatures of 350 or 400 1C and 50 MPa pressure are reported in Table 2.
Overall, the sintered density varies in the range of 75–90% ρtheo for all samples. The composites sintered at 400 1C possess higher density than their corresponding composites, sintered at 350 1C and thus, densification can significantly be improved by an increase of temperature of 50 1C from 350 to 400 1C. Also, contrary to the sintering of Cu–Pb nanocomposites at a pressure of 100 MPa and in temperature range 300–600 1C as reported in our earlier work [26], it was found out that the Pb has oozed out of the sintered compact due to change in wetting angle between Cu and Pb. Therefore, the application of pressure well and above 50 MPa will have a significant effect on the squeezing out of ‘liquid Pb’ out of network of porous channels between adjacent Cu grains thereby, voiding the original notion of incorporating Pb as solid lubricant. Hence to keep the Pb still ‘inside the sintered compact’ instead of ‘moving out on the surface’,
170
A. Siddharth Sharma et al. / Wear 319 (2014) 160–171
Table 3 Summary of literature results on friction and wear rate of TiB2-based alloys, processed via various manufacturing routes. Material composition (wt%)
Processing route/technique
Hardness (Hv, GPa) Test conditions
COF
Wear rate Wear ( 10 3 mm3/N m) mechanisms
Ref.
TiB2 coating on steel
Plasma transferred arc
11.8
0.5–0.65
0.1
Plastic deformation and oxidation
[29]
TiB2 coating on Al
Laser cladding
2.6 (30 wt% TiB2) 8.9 (70 wt% TiB2)
Sliding speed—0.15 to 0.6 m/s, load— 4.9, 9.8 N, slid distance—450 m against tool steel, coating thickness—850 to 1050 μm Sliding speed—0.24 m/s, load—8.9 N, slid distance—150 m against steel, coating thickness—430 μm Sliding speed—30.7 m/s, load—1.96 N, coating thickness— 70 μm Sliding speed—0.046 m/s, load—4.9 N, coating thickness— 100 μm Sliding speed—4.4 m/s, load—19.6 N, slid distance—88 m against steel, coating thickness—435–40 μm
–
230 19
[30]
–
0.001–0.0014
Ploughing, adhesve and delamination Spalling and microchipping of TiB2
[32]
Sliding speed—0.0025 m/s, sliding distance—216 m, load—0.98 N, against diamond Sliding speed—0.089 to 0.445 m/s, slid distance—667.5 m and load—60 N
Brittle failure and attrition Mixed adhesive– abrasive Fracturing and oxidation Plastic deformation with groove formation
[34]
TiB2 coating on Ni and Electrodeposition – Mo
TiB2 coating on steel
Monolithic TiB2
Cu–0 TiB2 Cu–0.5 TiB2 Cu–1.5 TiB2 Cu–2.5TiB2
Pulse electrode surfacing Laser surface engineering
–
Cold pressing, – sintering and ion implantation Induction 0.6 heating 0.8 0.9 1.2
it was therefore deemed suitable to restrict the pressure to 50 MPa. This reduction in pressure has its obvious ramifications on Cu-grain size in increasing the mean grain size from the earlier reported nanometric range to ultrafine size range as obtained here [26]. Another widely researched approach to obtain nanometric grains is to resort to multistage SPS cycle. But the application of multistage SPS schedule to obtain nanograined composites in case of highly electrically conducting Cu has proved futile and is attributed to the very nature of the SPS process which involves the simultaneous application of high pressure and electrical current. It can be further noted that the momentary temperatures between Cu grains can be as high as 10,000 1C due to formation of spark discharges. Due to these unavoidable events, it is not possible or rather difficult to apply the multistage SPS schedule in the present case. Instead, asreceived nanocrystalline Cu powder was taken and ball-milled to break agglomerates and single stage SPS cycle was restricted to lower temperatures to restrict abnormal grain growth. The hardness values of the composites span a wide range with the lower limit at 0.8 GPa to 2.2 GPa at its higher end. For Cu– 10TiB2, the hardness value does not go beyond 1 GPa at both the sintering temperatures of 350 and 400 1C, whereas for Cu–10TiB2– 10Pb composites, the hardness spans over a range of 1.6–1.9 GPa. For Cu–15TiB2, the hardness values are higher for composites with 10 wt% TiB2 and extend in the range of 1.85–2.0 GPa. In case of Cu– 15TiB2–10Pb composites, the hardness achieved is the highest (1.9–2.2 GPa). In view of the increment in hardness with the dispersion of 10 or 15 wt% TiB2 together with 10 wt% Pb content in nanocrystalline/ultrafine Cu-matrix, it can be viewed as a significant improvement as far as parametric processing window (350–400 1C and 50 MPa) is concerned. b) In order to explain the experimentally obtained lower wear rate of the order of 10 4 mm3/N m, an appraisal of the crucial role played by the crack nucleation and propagation in the subsurface region can be made here. Due to the very nature of the fretting wear, the excursions imposed by the stress state will be fluctuating and can be thought of as analogous to a fatigue phenomenon. This process will have a severe localized effect at the interfaces between different phases, in particular. Once these interfaces are
[31]
0.001–0.003
0.7
1159
0.6
463
0.02
–
–
22–52 18–41 15–37 7.5–30
[33]
ruptured, that will result in the nucleation of cracks (white dotted arrows, Fig. 9). Another possibility for interfacial crack nucleation can be related to the mismatch due to the difference in the thermal expansion at the interface between different phases. But as confirmed in an earlier related study [36] on Cu-based nanocomposites, substantial interface temperature rise to kick start the crack nucleation process and hence this viewpoint can be discarded in the present case [35]. As a further step, the instability endowed due to the wear process will have its implication on the structure of the grain boundaries. The morphology of Pb at the grain boundary region between Cu/Pb and TiB2/Pb biphasic couples will be more complaint as compared to the phase boundary of the unmelted Cu and TiB2 phases as Pb has undergone melting during sintering at 350 and 400 1C. Also, it should be recalled here that there is no new phase formation between any of the phases and hence any possibility of the grain boundary sliding will invariably be accommodated by the cracking at the interface. As TiB2 is in a plastically unconstrained matrix and also due to crystallographic mismatch between Cu (cubic structure, a ¼ 3.6147 Å) and TiB2 (hcp structure, a ¼3.029 and c ¼3.229 Å), the strain accumulation at the grain boundaries with TiB2 will lead to crack formation and subsequent propagation in the softer phases (Cu and Pb). Added to this crack formation and propagation, the subsequent cascading effect of this stress accumulation at the interface will lead to the microcavity formation in the vicinity of harder unyielding TiB2 (dotted circles, Fig. 9d). Due to the combined influence of these factors, the possibility of TiB2 particles getting dislodged from their original sites beneath the wear subsurface will increase and thereby leading to the formation of the harder debris particles. Such particles will also be responsible for the abrasion of the steel counterbody. The presence of Fe-rich delaminated tribolayer on the worn scar (Fig. 6a and Fig. 7b, EDS in inset) can thus be explained. The EDS analysis of the debris reveals the presence of Cu, Fe, Ti, Pb and O, indicating the formation of complex oxide in the form of mechanically mixed layer (Cu, Fe, Ti, Pb)xOy. The adhesion of dense (FexOy)-rich tribolayer indicates the entrapment or filling of pores by discrete wear debris particles. The sacrificial role of finer TiB2 particles (mean size 3 μm)
A. Siddharth Sharma et al. / Wear 319 (2014) 160–171
during wear will create voids on the worn scar of similar mean size. After the dislodging of TiB2 particles, the subsequent wear process will be primarily governed by the ultrafine grained Cumatrix and mechanically mixed hardened tribolayer. During wear, the extent of the debris generation as well as mutual transfer between mating surfaces depends upon the mutual solubility between different elements of the mating materials [37]. The solubility limits of the possible binary couples (Cu–Pb, Cu–Ti, Fe–Ti or Fe–Pb) in the present case is negligible enough in one another even in the temperature limit of 200– 300 1C, which is the typical temperature rise due to frictional heating for metal/metal tribocouple though it has not been supported by findings in an earlier study [36]. Summarizing, the present work establishes that low temperature spark plasma sintering can be effectively utilized to obtain around 90% dense Cu–TiB2 based nanocomposites with Pb dispersion. Despite the presence of 10 wt% Pb addition, the frictional properties cannot be reduced to any significant extent compared to Pb-free composition. However, the dominance of abrasive wear mechanism is reflected in the establishment of the inverse relationship between wear resistance and hardness for all the presently investigated TiB2containing nanocomposites. The unique combination of nanosized Cu-grains of equiaxed/elongated morphology along with homogeneous distribution of TiB2 enables us to obtain a good combination of hardness ( 2 GPa) and lower wear rate (10 4 mm3/N m) than other competing Cu-based nanocomposites. 5. Conclusions The major conclusions that be drawn from the present study are as follows: (a) An attractive combination of hardness (2.2 GPa) and fretting wear resistance ( 10-4 mm3/Nm) was achieved for Cu– 15TiB2–10Pb composites, spark plasma sintered at comparably low sintering temperatures vis-a-vis conventionally processed Cu-based alloys. (b) The better fretting wear resistance has been explained on the basis of TiB2 particles dominating the wear process as compared to softer Pb phase by debonding from the subsurface region, as analyzed using Focused Ion Beam microscopy. In addition, the formation of mechanically mixed tribolayer consisting of (Cu,Fe,Ti)xOy formed out of ultrafine debris consisting of harder TiB2 particles inhibits the continuous contact of the steel counterbody and composite surface. (c) A systematic and linear decrease in wear rate with an increase in hardness clearly indicates the predominance of abrasive wear and such a trend is consistent with the well known Archard’s law. Acknowledgements The authors thank the financial support administered by Department of Science and Technology and CARE grant, IIT Kanpur towards the procurement of SPS facility at IIT Kanpur and also AFMM, IISc Bangalore for facilitating FIB characterization. References [1] B. Basu, G.B. Raju, A.K. Suri, Processing and properties of monolithic TiB2 based materials, Int. Mater. Rev. 51 (2006) 352–374. [2] S.C. Tjong, Z.Y. Ma, Microstructural and mechanical characteristics of in situ metal matrix composites, Mater. Sci. Eng. Rep. 29 (2000) 49–113.
171
[3] C. Biselli, D.G. Morris, N. Randall, Mechanical alloying of high strength copper alloys containing TiB2 and Al2O3 dispersoid particles, Scr. Metall. Mater. 30 (1994) 1327–1332. [4] J. Lee, J.Y. Jung, E.S. Lee, W.J. Park, S. Ahn, N.J. Kim, Microstructure and properties of titanium boride dispersed Cu alloys fabricated by spray forming, Mater. Sci. Eng., A 277 (2000) 274–283. [5] J.R. Groza, J.C. Gibeling, Principles of particle selection for dispersion-strengthened copper, Mater. Sci. Eng., A 171 (1993) 115–125. [6] R.G. Munro, Material properties of titanium diboride, J. Res. Nat. Inst. Stand. Technol. 105 (2000) 709–720. [7] G.B. Raju, K. Biswas, B. Basu, Microstructural characterization and isothermal oxidation behavior of hot-pressed TiB2-10 wt% TiSi2 composite, Scripta Mater. 61 (2009) 104–107. [8] Z. Budrovic, H.V. Swygenhoven, P.M. Derlet, S.V. Petegem, B. Schmitt, Plastic deformation with reversible peak broadening in nanocrystalline nickel, Science 304 (2004) 273–276. [9] M.W. Chen, E. Ma, K.J. Hemker, H.W. Sheng, Y.M. Wang, X.M. Cheng, Deformation twinning in nanocrystalline aluminum, Science 300 (2003) 1275–1277. [10] S. Cheng, J.A. Spencer, W.W. Milligan, Strength and tension/compression asymmetry in nanostructured and ultrafine-grain metals, Acta Mater. 51 (2003) 4505–4518. [11] J. Schiotz, F.D. Di Tolla, K.W. Jacobsen, Softening of nanocrystalline metals at very small grain sizes, Nature 391 (1998) 561–563. [12] H.V. Swygenhoven, P.A. Derlet, Grain-boundary sliding in nanocrystalline fcc metals, Phys Rev B 64 (2001) 224105. [13] J.W. Cahn, Y. Mishin, A. Suzuki, Coupling grain boundary motion to shear deformation, Acta Mater. 54 (2006) 4953–4975. [14] J.W. Cahn, J.E. Taylor, A unified approach to motion of grain boundaries, relative tangential translation along grain boundaries and grain rotation, Acta Mater. 52 (2004) 4887–4898. [15] M. Legros, D.S. Gianola, K.J. Hemker, In situ TEM observations of fast grain boundary motion in stressed nanocrystalline aluminum films, Acta Mater. 56 (2008) 3380–3393. [16] E.O. Hall, The deformation and ageing of mild steel: III discussion of results, Proc. Phys. Soc. London, Sect. B 64 (1951) 747–753. [17] N.J. Petch, The cleavage strength of polycrystals, J. Iron Steel Inst. 174 (1953) 25–28. [18] M. Guo, K. Shen, M. Wang, Relationship between microstructure, properties and reaction conditions for Cu–TiB2 alloys prepared by in situ reaction, Acta Mater. 57 (2009) 4568–4579. [19] X.H. Zhang, C. Yan, Z.Z. Yu, In situ combustion synthesis of ultrafine TiB2 particles reinforced Cu matrix composite, J. Mater. Sci. 39 (2004) 4683–4685. [20] M. Lopez, D. Corredor, C. Camurri, V. Vergara, J. Jimenez, Performance and characterization of dispersion strengthened Cu–TiB2 composite for electrical use, Mater. Charact. 55 (2005) 252–262. [21] Z.Y. Ma, S.C. Tjong, High temperature creep behavior of in-situ TiB2 particulate reinforced copper-based composite, Mater. Sci. Eng., A 284 (2000) 70–76. [22] S.J. Dong, Y. Zhou, Y.W. Shi, B.H. Chang, Formation of a TiB2-reinforced copperbased composite by mechanical alloying and hot pressing, Metall. Mater. Trans. A 33A (2002) 1275–1280. [23] P. Yih, D.D.L. Chung, Titanium diboride copper-matrix composites, J. Mater. Sci. 32 (1997) 1703–1709. [24] Y.S. Kwon, D.V. Dudina, M.A. Korchagin, O.I. Lomovsky, Microstructure changes in TiB2–Cu nanocomposite during sintering, J. Mater. Sci. 39 (2004) 5325–5331. [25] A.S. Sharma, N. Mishra, K. Biswas, B. Basu, Densification kinetics, phase assemblage and hardness of spark plasma sintered Cu–10 wt% TiB2 and Cu– 10 wt% TiB2–10 wt% Pb composites, J. Mater. Res. 28 (2013) 1517–1528. [26] A.S. Sharma, K. Biswas, B. Basu, D. Chakravarty, Spark plasma sintering of nanocrystalline Cu and Cu–10 wt% Pb alloy, Metall. Mater. Trans. A 42A (2011) 2072–2084. [27] B. Bhushan, Principles and Applications of Tribology, first ed., John Wiley and Sons, US, 1999. [28] A.S. Sharma, K. Biswas, B. Basu, Microstructure-wear resistance correlation and wear mechanisms of spark plasma sintered Cu–Pb nanocomposites, Metall. Mater. Trans. A 45A (2014) 482 (-450). [29] M. Darabara, L. Bourithis, S. Diplas, G.D. Papadimitriou, A. TiB2, metal matrix composite coating enriched with nitrogen—microstructure and wear properties, Appl. Surf. Sci. 254 (2008) 4144–4149. [30] J. Xu, W. Liu, Wear characteristic of in situ synthetic TiB2 particulate-reinforced Al matrix composite formed by laser cladding, Wear 260 (2006) 486–492. [31] J.A. Kirk, D.R. Flinn, M.J. Lynch, Wear of TiB2 coatings, Wear 72 (1981) 315–323. [32] A. Agarwal, N.B. Dahotre, Comparative wear in titanium diboride coatings on steel using high energy density processes, Wear 240 (2000) 144–151. [33] C.S. Yust, C.J. Mchargue, L.A. Harris, Friction and wear of ion-implanted TiB2, Mater. Sci. Eng., A 105/106 (1988) 489–496. [34] J.P. Tu, W. Rong, S.Y. Guo, Y.Z. Yang, Dry sliding wear behavior of in situ Cu– TiB2 nanocomposites against medium carbon steel, Wear 255 (2003) 832–835. [35] Friction, Wear and Lubrication Technology, Eighteenth vol. Materials Park Ohio: ASM Handbook, 1992. [36] A.S. Sharma, K. Biswas, B. Basu, Fine scale characterization of surface/subsurface and nanosized debris particles on worn Cu–10% Pb nanocomposites, J. Nanopart. Res. 15 (2013) 1–12. [37] B.K. Prasad, A.K. Patwardhan, A.H. Yegneswaran, Factors controlling dry sliding wear behavior of a leaded tin bronze, Mater. Sci. Technol. 12 (1996) 427–435.