Characterisation of microstructure and mechanical

58

Int. J. Materials and Product Technology, Vol. 40, Nos. 1/2, 2011

Characterisation of microstructure and mechanical properties of cermets at micro- and nanoscales Irina Hussainova* and Maksim Antonov Department of Materials Engineering, Tallinn University of Technology, Ehitajate tee, 5, Tallinn, 19086 Estonia E-mail: [email protected] E-mail: [email protected] *Corresponding author

Iwona Jasiuk Department of Mechanical Science and Engineering, University of Illinois, 1206 West Green Street, Urbana, IL 61801-2906, USA E-mail: [email protected]

Xiangdong Du Department of Mechanical Engineering, McGill University, 817 Sherbrooke St. West, Montreal, Quebec H3A 2K6, Canada E-mail: [email protected] Abstract: High resolution measurements of mechanical properties of the constituent phases of multi-phase materials are of immerse importance in design of new composites. In this study, the nanoindentation, X-ray analysis and microstructural SEM investigations have been used to reveal the properties and structural features of ceramic – metal composites involving chromium carbide based cermets with different additives (Mo and Cu) in nickel binder. The additives influence microstructural parameters such as grain size and residual stresses; however, nanohardness and Young’s moduli of constituent phases remain less affected. Phase – specific mechanical properties are measured and correlated with bulk behaviour. Furthermore, hardness of the binder metal is found to be higher in cermet as compared to the bulk hardness of metal. Keywords: cermets; microstructure; nanoindentation; hardness; modulus of elasticity. Reference to this paper should be made as follows: Hussainova, I., Antonov, M., Jasiuk, I. and Du, X. (2011) ‘Characterisation of microstructure and mechanical properties of cermets at micro- and nanoscales’, Int. J. Materials and Product Technology, Vol. 40, Nos. 1/2, pp.58–74.

Copyright © 2011 Inderscience Enterprises Ltd.

Characterisation of microstructure and mechanical properties of cermets

59

Biographical notes: Irina Hussainova graduated from the Leningrad Polytechnic Institute (present: St. Petersburg Polytechnic University), Physical-mechanical faculty, Solid Physics department; she obtained her MS in Natural Science and then in 1999, her PhD in Engineering Science. She is a Senior Researcher at the Institute of Materials Engineering, Tallinn Technical University, Estonia. She is the author or co-author of more than 90 scientific articles published in international journals and the Laureate of the 2005-year State Award of the Republic of Estonia in the field of Technical Science for the development and investigations in the field of nanomaterials. Maksim Antonov is a Senior Researcher at the Institute of Materials Engineering, Tallinn University of Technology, Estonia. He obtained his PhD in Technical Science in 2006 and deals with wear and high temperature properties of multiphase materials for more than eight years. He is the Head of Tribology Laboratory of Tallinn University of Technology and Designer of several test rigs. He is the author or co-author of 15 scientific articles published in Estonia as well as internationally. He has attended several international conferences and was working as a Visiting Researcher in European R&D institutions. Iwona Jasiuk received her PhD in Theoretical and Applied Mechanics from Northwestern University in 1986. Since January 2006, she is a Professor of Mechanical Engineering at the University of Illinois at Urbana-Champaign. Prior to that, she held faculty appointments in the Department of Mechanical and Industrial Engineering at Concordia University in Montreal, Canada (2004–2006), in the School of Mechanical Engineering at Georgia Institute of Technology (1996–2004) and in the Department of Materials Science and Mechanics at Michigan State University (1986–1996). She is an expert in micromechanics modelling of materials, including biological materials and nanocomposite materials. Xiangdong Du received his PhD in Mechanical Engineering from McGill University in Montreal, Canada in 2006. His expertise is in micromechanics modelling of composite materials. Currently, he is a Senior Engineer at Rowan Williams Davies and Irwin Inc. consulting company in Ontario, Canada.

1

Introduction

Properties of composite materials are often derived from the properties of constituents measured in their bulk form (Fan et al., 1994; Ling and Hou, 2007). However, properties of phases may be modified due to sintering process, presence of impurities and inclusions and thermo-mechanical mismatch between two or more phases. Chemical and physical interactions between different materials result in formation of new multiphase inhomogeneous structures. Measuring the intrinsic properties of each phase separately gives the information on the spatial heterogeneity in local material properties and serves as guidance to process engineering and advanced materials design. The knowledge of the in situ properties is also needed for inputs for modelling so predictive tools can be developed. Cermets composed of ceramic and metallic phases are one of the widely used groups of multiphase materials. A cermet is designed to have the optimal properties of both a ceramic, such as high temperature resistance and hardness and those of a metal, such as the ability to undergo plastic deformation. However, due to generally greater magnitudes

60

I. Hussainova et al.

of coefficient of thermal expansion of the metallic constituent as compared to ceramics, the phases are in the states of tension and compression which influence their hardness (Olivas et al., 2006). Cermets are particle reinforced composites which usually involve microsized particles. Thus, the evaluation of their in situ properties involves measurements over micronsize volumes. Nanoindentation provides a unique capability for measuring the hardness, elastic modulus and other material properties through one basic test. The test requires almost no sample preparation, is non-destructive and can be applied to variety of materials. Nanoindentation characterises the material response by recording the load and displacement during a complete loading-unloading cycle. Elastic properties are extracted from the initial unload portion of the curve and plastic properties such as hardness are calculated directly from curve fitting parameters (Gammage et al., 2004; Hay et al., 1999; Oliver and Pharr, 1992). The literature on indentation of composite materials includes only few very recent publications (Constantinides et al., 2006; Engqvist and Wiklund, 2000; Guicciardi et al., 2008; Ling and Hou, 2007; Qu et al., 2003; Rodriguez et al., 2006) because of the complexity of the mechanical response of a material system when indentation volumes and microstructural volumes are of the same order. This lack of length scale separation makes it difficult to translate indentation data into meaningful mechanical properties. Nanoindentation has been recently used to evaluate in situ properties of some metal matrix composites (Rodriguez et al., 2006) and WC-Co hardmetals (Engqvist and Wiklund, 2000; Qu et al., 2003). However, there is no systematic use of nanoindentation technique for testing of ceramic based and metal bonded composites with ceramic particles content higher than 60% of volume fraction. The aim of the present study is to test cermet materials at micron-level to evaluate the micromechanical properties of the constituent phases of chromium carbide based and nickel or nickel alloys bonded composites. Ceramic-metal composites are a success story from the viewpoint of their many applications (Hussainova, 2007a, 2007b; Stack et al., 2006). Chromium carbide-nickel based composites can be used in many environments involving tribo-corrosion due to their combined ability to resist wear and corrosion (Stack et al., 2006). Hence, they are candidate materials for use either in bulk as surface coatings in crude oil (offshore) applications or in power and marine industries. However, the lack of design criteria and predictive models for cermets poses a significant barrier to their wider application. Because of their high hardness, good strength and oxidation – corrosion resistance at elevated temperatures (Hussainova, 2007a, 2007b; Stack et al., 2006), the Cr3C2-Ni (Mo; Cu) cermets were chosen for the present investigation.

2

Experimental methods and materials

2.1 Materials The chromium carbide based composites were produced at the Laboratory of Powder Metallurgy of Tallinn University of Technology by sinter HIP technique and routine that have been described in detail elsewhere (Pirso et al., 2006): 20% of nickel powder was milled together with Cr3C2 powder, pressed to compacts and sintered at temperatures of 1250–1300°C. Materials tested had a carbide particle content of 80% by weight or 84.2%


61

by volume. Five types of composites with different additives in the binder metallic phase were tested. Small amount of molybdenum and/or copper was added to nickel to improve the wettability of the ceramic phase and inhibit grain growth. Nickel forms a complete solid solution with copper and can dissolve about 35% of chromium and about 20% of molybdenum and tungsten. Thus, the tough, ductile face centred cubic matrix of Ni can dissolve extensive amounts of elements in various combinations to provide solid solution strengthening as well as improved corrosion and oxidation resistance. The degree of solid solution strengthening is related to the atomic size difference between nickel and the alloying element and therefore to the ability of the solute to interfere with dislocation motion. The main difference in the chemical compositions of five grades is the concentration of the additives. Figure 1

(a) SEM image of (Cr3C2)-(Cr7C3)-(CrNi3) composite with no additives: dark grey areas – Cr3C2 grains; light grey areas – Cr7C3 grains; white areas – Ni alloy and black regions – pores (b) elemental composition (%) of the constituent phases of the cermet fabricated from chromium carbide and nickel powders with no additives (see online version for colours)

(a)

(b) Note: The numbers 1, 2, 3 and 4 label the phases subjected to EDS analysis.

62


Analysis of Cr-C-Ni phase diagram shows the cooperative eutectic crystallisation of two-phase (Cr3C2)-(Ni) and three-phase (Cr3C2)-(Cr7C3)-(Ni) solutions at the sintering temperature. Figure 1(a) represents a microstructure of the cermet with a Ni binder fraction of 20 wt.%. In Figure 1(a), different grey colour levels indicate Ni phase as the lightest area, Cr7C3 particles as the middle grey area and Cr3C2 grains as the darkest area. The carbides in the composite have a variety of sizes and aspect ratios that also vary with the additives content. The irregular shaped carbide grains form a skeleton with increased rigidity. The dihedral angle, that represents the balance between interfacial energies, is large enough and the liquid spread over the solid surfaces is not fully completed. Black regions on the image in Figure 1 indicate pores. The adjacent grains may weld together and the resulting structure consists of connecting carbide particles with fractionated metallic phase. The mean grain size (Feret diameter) was measured with an image analysis system from the digital scanning electron microscope (SEM) micrographs as an average value of five fields to get more reliable results. For each measurement at least 1000 particles were counted. The carbides particle size in cermets with no additives was in the range of 4–15 μm and with Mo or Cu additives was 3–8 μm although, some large grains of around 30 μm were also found. The mean aspect ratio was 4:1 for elongated Cr3C2 and 3:2 for rounded Cr7C3. The porosity measured by Archimedes technique was of about 1 vol.% for all materials.

2.2 Experimental techniques Microstructural examination was conducted in a SEM JEOL 6060LV. Thermo-Electron state-of-the-art energy dispersive spectroscopy (EDS) and wavelength dispersive spectroscopy (WDS) X-ray Microanalysis System with atmospheric thin window EDS detector and high performance WDS parallel beam spectrometer with hybrid X-ray optics were used for qualitative and quantitative phase analysis. The SEM/EDX microanalysis was performed at the acceleration voltage of 15 kV and a Cu source of Ka radiation with possibilities of line or point focus. To identify a free carbon in a material, the Auger electron spectroscopy was applied. The Physical Electronics model PHI 660 scanning Auger microprobe (SAM) has a system of elemental mapping with high spatial resolution and is capable of performing SEM. Mechanical characterisation of the bulk materials was carried out by means of indentation technique, which is the simplest and widely used method for macroproperties evaluation although, for multiphase non-homogeneous materials prone to cracking and the values measured may be highly variable. Vickers hardness was measured in accordance with the ASTM Standard E384. Instrumented indentation was applied to estimate the modulus of elasticity because of difficulties in obtaining true elastic deformation using conventional stress-strain tests due to crack phenomena. Each test point indicates the average value of 12 measured results. The nanoindentation tests were performed on well-grinded and polished samples using a Triboscope indenter system (Hysitron, Minneapolis, MN) with a diamond Berkovich tip. The software of control and data acquisition system was Triscan 6.0. A specially designed Triboindenter transducer (piezo plate) was used. The maximum load


63

for this nanoindenter is 10,300 μN with 1 μN resolution and load noise is 100 nN. The maximum depth is 20 μm and displacement resolution is 0.04 μm. The displacement noise is 0.2 nm. Thermal drift is less than 0.05 nm/sec. Our tests were conducted at room temperature and were load-controlled. The optical monitor has a ×10 telescope for setting up samples and preliminary tests. The Triboindenter can scan the surface by the indenter tip to obtain the topographic image and gradient image. The calibration range covered the indentation depths from 50 nm to 600 nm. In all experiments a trapezoidal load history was prescribed, defined by loading segment duration of 10 s, a holding period at maximal load of 2 s and unloading segment duration of 10 s. During nanoindentation tests the load on specimen and the depth of penetration were recorded and care was taken to properly calibrate it for the range of the penetration depths recorded in both hard and soft phases. Young’s modulus and hardness were estimated from the load-displacement curve using the Oliver-Pharr approximation for a Berkovich indenter (Engqvist and Wiklund, 2000; Gammage et al., 2004). The reduced elastic modulus (Er) and hardness (HD) can be obtained using the following relations: Er =

π 2 A

s

HD =

Pmax ; A

(1)

where Pmax is the maximum load applied; s is stiffness of contact between the indenter and tested material and A is a projected area of contact at maximum load. The stiffness s is given by the slope of the initial portion of the unloading curve (s = dp/dh). The area of the contact A is determined from the indenter tip geometry. For the Berkovich tip it is given by A = 24.5hc2, where hc is the contact depth calculated from hc = hmax – 0.75(Pmax/s). The elastic modulus of the tested material Es can be obtained from the following formula (assuming a value for Poisson’s ratio νs of the tested material): 1 1 − ν i2 1 − ν s2 = + ; Er Ei Es

(2)

where Ei and νi are Young’s modulus and Poisson’s ratio of the indenter; Ei = 1141 GPa and νi = 0.07 for a diamond indenter. Two types of the nanoindentation tests were carried out: 1

manual

2

automated.

In the manual mode the image scan was obtained and then the indents with 2,000 μN load were applied at selected locations. In the automated mode the indentation tests were made at ten random locations on each sample. At each location, a pattern consisting of six indentation points was set-up. The load was increased from 1,000 μN to 6,000 μN, in 1,000 μN load intervals, for the six points at each location. Then, additional 500 μN loads were applied at ten random locations. We are considering this wide range of applied loads to explore any possible size dependence.

64

3


Results

Chromium carbides have three crystallographic structures: cubic Cr23C6, hexagonal Cr7C3 and orthorhombic Cr3C2, with the last one possessing best mechanical properties. However, the real sintering process can hardly be maintained to avoid any Cr7C3 grain growth. With the help of SEM/EDS technique, the local chemical analysis of the constituent phases was carried out. According to the EDS analysis, the metallic binder phase of the sintered cermets consists of nickel-chromium solid solution with the presence of a free carbon as indicated in Figure 1(b). The traced elements of W and Co are revealed in negligibly small amounts and their presence in the material is due to the process of powder preparation in an attractor mill equipped with WC-Co hardmetal vial. XRD analysis identified three distinct crystalline phases on the basis of their different lattice parameters, namely Cr7C3, Cr3C2 and chromium-nickel alloy CrNi3 (see Figure 2). The elemental distribution studied by SAM indicates the presence of high amount of free carbon throughout the structure, solubility of nickel in hexagonal chromium carbide Cr7C3 and spurious phases formation (Figure 3). Figure 3(a) presents the SEM image of the area studied by means of Auger electron spectroscopy and Figures 3(b), 3(c) and 3(d) show nickel, carbon and chromium distribution maps, respectively. Figure 2

X-ray diffraction phase analysis of chromium carbide based composite (see online version for colours)

Characterisation of microstructure and mechanical properties of cermets Figure 3

65

Elemental distribution map for CN cermet (a) SEM image of the region tested (b) distribution of nickel throughout of the specimen (c) distribution of chrome throughout of the specimen (d) distribution of carbon throughout of specimen (see online version for colours)

(a)

(b)

(c)

(d)

In the present work five cermet grades based on chromium carbides were tested. Their composition and macroscopic mechanical properties at room temperature are summarised in Table 1. The microstructure of the CN cermet is shown in Figure 1 while microstructures of the CNM10 and CNC10 cermets are presented in Figure 4. Table 1 Grade

Materials composition and macroscopic mechanical properties Content of Cr3C2 (% by weight)

Binder composition

Content of additives in Ni binder (% by weight)

Vickers hardness, GPa

Modulus of elasticity, GPa

CN

80

Ni

0

12.1 ± 2

295 ± 12

CNM3

80

Ni + Mo

3

12.2 ± 2.5

300 ± 15

CNM10

80

Ni + Mo

10

12.2 ± 3

302 ± 18

CNC3

80

Ni + Cu

3

10.7 ± 2.7

290 ± 20

CNC10

80

Ni + Cu

10

10.1 ± 3

287 ± 10

66 Figure 4

I. Hussainova et al. SEM images of (a) CNM10 cermet (b) CNC10 cermet

(a)

(b)

Measurements of elastic modulus and hardness of the cermet constituents have been conducted using nanoindentation in manual and automatic modes. It is suggested that the measured Young’s modulus or hardness are representative of the indented particle if the penetration depth (h) is smaller than 1/10 of the radius of an indented particle (D) (Constantinides et al., 2006; Guicciardi et al., 2008). The maximum contact depth recorded in the tests was about 100 nm, i.e., smaller than 1/10 of the grain size. A special care was taken when the binder phase was tested as the contact depth was around 300 nm that is close to 1/10 of alloy islands between hard grains. Figure 5 depicts typical indentation curves for the hard ceramic grains and soft binder alloy as well as scanned images of the indentation sites. The contact stiffness was calculated from the slope of the unloading curve at the maximum depth. Generally, the in-situ properties are evaluated by imaging the indentation marks using, for example, SEM to detect the phase that was indented or properties can be derived from a statistical analysis of nanoindentation results by fitting the experimental data to a proper number of statistical distributions using an indentation grid (Constantinides et al., 2006; Guicciardi et al., 2008). SEM imaging is a time consuming procedure especially if the indentation sites are tiny as it happens in hard phases such as chromium carbides. In principle, in scanning probe microscopy (SPM) mode, the instrument can provide the in-situ images of the sample surface immediately before and after the test is performed, but the quality of the images is often unsatisfactory to draw conclusions about the indented phase. However, the load-displacement curves obtained for carbides exhibit a typical elastic behaviour, as it could be expected for ceramic particles (Figure 5), while the binder alloy shows a plastic behaviour. Therefore, even if the phases are not obviously defined by the imaging, the resulting curve provides us with useful information about the indented phase. SPM images of residual indentations [Figure 5(b)] demonstrate that there is no significant pile-up in the Berkovich indentation of the carbides, suggesting that the Oliver and Pharr method would yield reliable estimates of the contact area and therefore, provide the reliable parameters for hard particles. Much more complicated is the case of binder alloy indentation. The results of the manual test are shown in Figure 6 which demonstrates the scanned image of the CNM3 composite along with the indentation data


67

(Er and HD). In the case of metal binder indenting, a pronounced pile-ups around the imprint are clearly recognisable. The analysis of the data obtained from manual testing allows us to distinguish three ranges of parameters for all cermets. The first one is grouped around mean values of 22 GPa for HD and 350 GPa for Er; while another group exhibits values of about 18 GPa for HD and 300 GPa for Er. Undoubtedly, these parameters describe the carbides in the cermet materials. When binder metal is indented, hardness lowers down to about 4.5 GPa, however, no certain conclusion can be drawn from those particular measurements about reduced modulus for the binder phase because a lack of reliable information. As low value of HD as 3.27 GPa (Figure 6) was measured in the immediate vicinity of the pore that may dramatically influence a material response to loading. For all cermet grades tested the hardness values of the binder phase ranging from 4 to 6 GPa seem to be much more realistic. Figure 5

(a) Typical experimental load-displacement curve after testing binder metal and ceramic particles in cermets (b) SPM image of the nanoindentation imprints (see online version for colours)

binder carbides

(a)

(b)

68


Figure 6

Results from manual tests on scanned topographical images of CNM3 specimen (see online version for colours) Er= 257.19 GPa HD = 20.10 GPa

Er = 262.17 GPa HD = 19.79 GPa

Er = 260.07 GPa HD = 18.99 GPa

Er = 235.56 GPa HD = 3.27 GPa Er = 316.80 GPa HD = 21.79 GPa Er = 311.26 GPa HD = 5.78 GPa Er = 302.81 GPa HD = 20.83 GPa Er = 310.2 1GPa HD = 18.86 GPa

Er = 296.25 GPa HD = 22.10 GPa

Er = 320.71 GPa HD = 21.17 GPa

Er = 300.83 GPa HD = 21.34 GPa

Er = 304.65 GPa HD = 18.95 GPa Er = 322.02 GPa HD = 21.50 GPa

Er = 248.27 GPa HD = 5.75 GPa

Er = 316.78 GPa HD = 19.50 GPa

Note: The applied force is 2000 µN. The image is 20 µm square.

The indentation (reduced) moduli can be converted to the elastic moduli of the individual phases by using equation (2) and assuming a Poisson’s ratio for each phase. For the chromium carbides the Poisson’s ratio is taken to be equal to 0.2 as for almost all carbides of metals of transition group VIB (Matweb, 2008) and for nickel alloys the Poisson’s ratio of 0.3 is assumed although the exact value is not known for any specific binder composition. However, it is not necessary to know the value of the Poisson’s ratio with great precision to obtain a reasonable estimate of the Young’s modulus. In fact, a Poisson’s ratio of 0.1–0.4, representative of the range including most engineering metals and ceramics, induces an error on E of less than 10% (Constantinides et al., 2006). Based on the results of manual tests, the elastic modules of three phases can be calculated and their values are summarised in Table 2. Mean values of nanohardness of the constituent phases are listed in Table 3. Table 2

Mean modulus of elasticity of the constituent phases within cermets tested CN

CNM3

CNM10

CNC3

CNC10

Ebinder, GPa

Grade

200

213

218

193

190

ECr3C2, GPa

300

305

305

300

295

ECr7C3, GPa

345

350

350

340

340

Characterisation of microstructure and mechanical properties of cermets Table 3

Mean hardness of the constituent phases within cermets tested

Grade

CN

CNM3

CNM10

CNC3

CNC10

Hbinder, GPa

5.2

5.3

5.4

5.2

5.1

HCr3C2, GPa

18.5

18.1

18.1

17.5

17.4

HCr7C3, GPa

21.5

21.7

22

21.1

21.1

Figure 7

Hardness and reduced modulus Er for (a) CN cermet (b) CNC3 cermet (c) CNM3 cermet (see online version for colours)

69

70


Also, automatic tests have been conducted under load values ranging from 500 μN to 6,000 μN. For each load magnitude, ten random locations on each composite type sample were selected. Representative results are illustrated in Figure 7. Since the materials are randomly heterogeneous, only two groups of values of HD and Er were identified and averaged (carbide and binder metal phases). For each case illustrated in Figure 7, the Er and HD, both are highest under the load 500 μN. The results decrease for 1,000 μN and further for 2,000 μN. When the load is greater or equal to 2,000 μN, the results remain nearly constant, i.e., they do not show anymore the so-called size effect. Following this observation the load of 2,000 μN was selected for manual tests. Table 4 lists the elastic modulus and hardness of the carbides and nickel as bulk materials and as CN composite constituents measured by nanoindentation. Results from manual and automated tests are taken into consideration and averaged. Distinct properties of two carbide modifications can hardly be distinguished from automated tests. A particular attention has to be paid to visualisation of the area investigated. Therefore, further tests are needed to improve our knowledge on the properties of constituent carbide phases when automated mode is applied. Table 4

Modulus of elasticity and hardness of bulk materials and CN composite constituents Bulk Ni (Matweb, 2008)

Ni(Cr) in cermet CN

Bulk Cr3C2 (Matweb, 2008)

Bulk Cr7C3 (Matweb, 2008)

Chromium carbides in cermet (statistical)

E (GPa)

207

200

373–386

340–400

280–360

H (GPa)

0.7

5.5

10.2–18

16–20

16–22

4

Discussion

The Ni-based binder phase is the residual of the liquid phase formation during the sintering process. At the temperature of about 1100°C the dissolution of chromium carbides begins. As a result of carbides dissolution, free carbon and chromium were found at the binder phase. The widened and shifted X-ray diffraction lines indicate a mutual dissolution of the constituents during sintering. In our particular case the composition of the cermets produced by PM technique can be reported as Cr3C2-Cr7C3-CrNi3(Mo; Cu)-C. The nanoindentation method was employed to analyse the micro-scale properties of such multiphase materials. The main assumption of the Oliver and Pharr method is that the shape of the deformed solid outside the area of contact points to the elastic behaviour. This is not true when plastic deformation occurs around the indenter to form a material pile-up. The degree of pile-up depends upon the ratio E/Y of the specimen, where Y is the yield stress of a material. For such materials as nickel alloys with a large value of E/Y ~ 3,000, the effect of piling-up on the contact area may result in errors in determining the true value of HD and Er (Fisher-Cripps, 2002). To minimise such influence, the load should be chosen low enough for testing binder phase. However, the load should not be very low for testing ceramic particles, either. Otherwise, the indentation imprints will be hardly visible under microscope. In Rodriguez et al. (2006) the influence of pile-ups was estimated by finite


71

element modelling and overestimation of the results obtained with nanoindentation was shown. The effect could be corrected by means of the simulated test analysis. The high standard deviation observed for the binder phase should be noted. This may be a consequence of the hard particles effect due to proximity of grains leading to violation of the conditions h/D ≤ 1/10. This experimental obstacle could be overcome by further reducing the maximum load or penetration depth but it is beyond the limits of the applied test capabilities due to system calibration. Recalibration for testing small volumes under low loads or low penetration depth may result in another type of errors in values evaluation because of imperfect geometry at the indenter apex (Fisher-Cripps, 2002). Moreover, if the neighbouring matrix affects the measured values of the indented area, the properties of that area will reflect a strong correlation with the same properties of the neighbouring phases. The moduli of elasticity values are overlapped, which points to a strong influence of the hard phase on the binder alloy, Figure 7. The ‘proximity effect’ can be studied by the decrease in the carbides content. This issue will be discussed in our next paper. Quite a large scattering of data for binder phase may also be explained by structural inhomogeneities of metal alloy, influence of porosity and possibility of a situation when only a shallow layer of binder material overlays the void in material. The hardness measured by nanoindentation in the binder metal is more than 500% higher than that in bulk nickel. In this context it should be mentioned that direct comparison between hardness obtained during nanoindentation and microhardness measurements may draw erroneous conclusions. The difference in hardness can originate from difference in tip geometry, i.e., Berkovich in nanoindentation and Vickers in microhardness. For some reasons, nanoindentation often exhibits somewhat higher values of hardness (Engqvist and Wiklund, 2000; Rodriguez et al., 2006). No tests on single crystals of chromium carbide or bulk nickel-chrome have been performed in this work. Therefore, available data (Matweb, 2008) have been used for illustration. A high hardness of binder metal is supposed to be due to solid solution of Cr and C, to residual stresses present within the specimens as a result of processing (mismatch in coefficients of thermal expansion between two phases) and to surface preparation (polishing). The modulus of elasticity should not depend on a measuring technique so the results can be compared to the reported ones with no risk of serious error. Binder CrNi3 solid solution shows the modulus of elasticity quite similar to that reported for bulk nickel. However, the modulus measured for both modifications of chromium carbides is smaller than it could be expected. It is very possible that the solubility of nickel in carbide and its influence on the carbide properties results in the formation of carbide structures that differ from structures of commonly referred ones. As expected, modulus of elasticity and nanohardness of the carbide phase is not dramatically affected by changing in the binder alloy composition (Tables 2 and 3). However, different cermet grades do not reveal a significant difference in the metallic phase nanohardness, either. This fact points to a significant role of thermal residual stresses that influence the hardness of phases. The results confirm the necessity of the development of new criteria or concepts for the mechanical properties evaluation in such complex systems as ceramic-metal composites taking into consideration phases’ interaction and/or bonding (interfaces) between phases and the fact that material constituents are somewhat different from pure

72


metals or carbides. For example, the definite conclusion on Vickers hardness of bulk material with application of 5 kg load is hardly drawn because of microcracking around impression on the surface and subsurface. Since microfracture occurs mainly during the loading, a portion of energy, used to create the deformation, is dissipated by the crack formation. As it was shown in Hussainova, (2007a, 2007b), the cracks in chromium carbide based cermets mostly propagate through the interfaces between two hard grains or, at larger loading, through the grains, that point to the need of fracture toughness determination at micro/nanoscale and looking for the ways of enhancing the interface properties. In conventional continuum mechanics, the yield behaviour of a material is size independent. However, in nanoindentation, plasticity size effects have been observed by many researches (Fisher-Cripps, 2002; Chiu and Ferguson, 2007), where a higher hardness is measured for a smaller indentation size. For yield by densification (fused silica), there was no size effect in the nanoindentation regime. For phase transition (silicon), the length scale was of the order tens of nanometres. Many reports of indentation size effects are actually due to artifacts: surface layers that were not accounted for, poor tip shape calibration, etc. In all probability, in the case of cermets the observed indentation size effect is the error associated with the area function of the indenter, particularly under very low load. To clear up this phenomenon, recalibration of the system was conducted. No pronounced size effect was revealed during the tests using the device recalibrated for low loads.

5

Concluding remarks

In the study, five grades of chromium carbide based cermets with different composition of binder alloys were tested using SEM, EDS and nanoindentation techniques. It was shown that binder phase represents some kind of solid solution of nickel and chromium with the presence of free carbon and other traced elements. Mutual dissolution and diffusion of elements during sintering results in formation of complicated multiphase composites comprised of at least three phases and some amount of free carbon: Cr3C2-Cr7C3-CrNi3(Mo; Cu)-C. Solubility of Ni in Cr7C3 is quite high and the phase may be described as Cr7C3(Ni) phase possessing properties different from those commonly reported for Cr7C3. In such particular case the method that allows probing the phases of composite despite the small characteristic scales of constituents is of great importance. In the present work, the nanoindentation technique and Oliver-Pharr analysis were employed to conduct a microcharacterisation of five grades of chromium carbide based cermets with high percentage (84 vol.%) of the hard ceramic grains. The mechanical properties (modulus of elasticity and hardness) of the ceramic grains and binder alloys were found to be different from those reported elsewhere (Hussainova, 2007a). Modulus of elasticity was slightly lower while hardness was larger than those of bulk materials. However, it should be noted that the formation of spurious phases may provide an inefficient comparison between properties of the constituent phases in cermet and pure bulk materials. Indentation to a constant maximum depth chosen separately for hard and soft phases rather than a constant maximum load would give some advantages for modulus and hardness evaluation in multiphase systems of high percentage one of the phases.


73

Adding a small amount of molybdenum or/and copper influences insignificantly the micro-properties of cermet constituents. Incorporation of molybdenum is important from metallurgical point of view for improvement in phase bonding and inhibition of grain growth. These parameters contribute to macro-mechanical properties preventing cracking under loading. More tests are needed to establish the relationships between macro- and microscale measurements on the highly inhomogeneous multi-phase materials. However, the issue of reasonableness in comparison between results obtained for such a complicated system as ceramic-metal composites with the help of different techniques at different scales is still under debate.

Acknowledgements The authors would like to thank Senior Researcher J. Pirso for advices and help during fabrication of the materials. Also, Estonian Science Foundation under Grants Nos. G6163 and G6660, and Council for International Exchange of Scholars, CIES, USA are acknowledged for their financial support in this research. This study was carried out at the McGill University in M. Ostoja-Starzewski Laboratory and in the F. Seitz Materials Research Laboratory Central facilities, University of Illinois, which are partially supported by the US Department of Energy under Grants DE-FG02-07-ER46453 and DE-FG02-07-ER46471.

References Chiu, Y.L. and Ferguson, W.G. (2007) ‘Characterization of indentation size effect of hardness using a loading curve from crystalline materials’, Advanced Mat Research, Vol. 29–30, pp.55–58. Constantinides, G., Chandran, K.S.R., Ulm, F.J. and Van Viet, K.J. (2006) ‘Grid indentation analysis of composite microstructure and mechanics: principles and validation’, Mater Sci Eng A Struct Mater, Vol. 430, pp.189–202. Engqvist, H. and Wiklund, U. (2000) ‘Mapping of mechanical properties of WC-Co using nanoindentation’, Tribology Letters, Vol. 8, pp.147–152. Fan, Z., Tsakiropoulos, P. and Miodownik, A.P. (1994) ‘A generalized law of mixtures’, Journal of Materials Science, Vol. 29, pp.141–150. Fisher-Cripps, A.C. (2002) Nanoindentation, Springer Verlag, New York. Gammage, J., Wilkinson, D.S., Brechet, Y. and Embury, J.D. (2004) ‘Model for damage coalescence in heterogeneous multi-phase materials’, Acta Mater., Vol. 52, pp.5255–5263. Guicciardi, S., Melandri, C., Silvestroni, L. and Sciti, D. (2008) ‘Indentation grid analysis of nanoindentation bulk and in situ properties of ceramic phases’, J. Mater. Sci. Vol. 43, pp.4348–4352. Hay, J.C., Bolshakov, A. and Pharr, G.M. (1999) ‘A critical examination of the fundamental relations used in the analysis of nanoindentation data’, J. Mater. Res., Vol. 14, pp.2296–2308. Hussainova, I., Pirso, J., Juhani, K., Antonov, M. and Letunovits, S. (2007a) ‘Erosion and abrasion of chromium carbide based cermets produced by different methods’, Wear, Vol. 263, pp.905–911. Hussainova, I. and Antonov, M. (2007b) ‘Assessment of cermets performance in erosive media’, Int J Mat Prod Tech., Vol. 28, Nos. 3–4, pp.361–376. Ling, Z. and Hou, J. (2007) ‘A nanoindentation analysis of the effects of microstructure on elastic properties of Al2O3/SiC composites’, Compos Sci Technol, Vol. 67, pp.3121–3129.

74


Matweb (2008) ‘Searchable database of material data sheets’, accessed on 10/05/2008, available at http://www.matweb.com. Olivas, E.R., Swadener, J.G. and Shen, Y. (2006) ‘Nanoindentation measurement of surface residual stresses in particle reinforced metal-matrix composites’, Scripta Materialia, Vol. 54, pp.263–268. Oliver, W.C. and Pharr, G.M. (1992) ‘An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments’, J. Mater. Res., Vol. 7, No. 4, pp.1564–1583. Pirso, J., Viljus, M., Letinovits, S. and Juhani, K. (2006) ‘Reactive carburizing sintering – a novel production method for high quality chromium carbide – nickel cermets’, Int J Refractory Metals & Hard Metals, Vol. 24, pp.263–270. Qu, J., Riester, L., Shin, A., Scattergood, R., Lara-Curzio, E. and Watkins, T. (2003) ‘Nanoindentation characterization of surface layers of electrical discharge machined WC-Co’, Mater. Sci. Eng A, Vol. 344, pp.125–131. Rodriguez, J., Garrido-Maneiro, M.A., Poza, P. and Gomez-del Rio, M.T. (2006) ‘Determination of mechanical properties of aluminium matrix composites constituents’, Mater. Sci. Eng. A, Vol. 437, pp.406–412. Stack, M.M., Antonov, M. and Hussainova, I. (2006) ‘Some views on the erosion-corrosion response of bulk chromium carbide based cermets’, J. Physics D: Applied Physics, Vol. 39, pp.3165–3174.