- Email: [email protected]
Temperature and load influence on in-situ formed layers during high temperature abrasion
Author’s Accepted Manuscript Temperature and load influence on in-situ formed layers during high temperature abrasion M. Varga, E. Badisch
www.elsevier.com/locate/wear
PII: DOI: Reference:
S0043-1648(16)30688-3 http://dx.doi.org/10.1016/j.wear.2017.04.020 WEA102149
To appear in: Wear Received date: 28 November 2016 Revised date: 11 April 2017 Accepted date: 25 April 2017 Cite this article as: M. Varga and E. Badisch, Temperature and load influence on in-situ formed layers during high temperature abrasion, Wear, http://dx.doi.org/10.1016/j.wear.2017.04.020 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.
Temperature and load influence on in-situ formed layers during high temperature abrasion M. Varga*, E. Badisch AC2T research GmbH, Viktor-Kaplan-Straße 2C, 2700 Wiener Neustadt, Austria.
Abstract Abrasive wear at high temperature (HT) applications is a serious issue in industry, limiting the lifetime of core components, e.g. in steel- or cement production. In literature abrasive wear is commonly linked to material’s hardness, but it is known that microstructure and temperature show also major influence. Current studies also found wear protecting effects by the in-situ formation of mechanically mixed layers (MML) with the abrasive, which can be especially beneficial at HT. To investigate the temperature and load influence of this MML formation at high-stress abrasive conditions a modified ASTM G65 test, allowing for temperatures up to 700°C, was utilised. Two metal matrix composite (MMC) materials prone to MML formation were chosen: a Ni-based and Fe-based cast alloy with carbide content of ~15 %. Thereby the MMC matrix influence on MML formation was studied. Load was varied from 10 to 45 and 80 N, and temperature from room temperature to 500°C and 700°C. Abrasive is collected during tests in order to determine the severity of the contact during the wear process. Post-test analysis investigate the MML coverage and depth in order to estimate the wear protecting effect. Further microhardness investigations and nano-scratch experiments should identify fundamental wear mechanisms and load bearing capacity of the hard phase network. It was found that increasing load leads to higher severity of the contact (more abrasive breakage). MML coverage was strongly dependent on load and temperature at the Fe-based material. Despite a pronounced drop of the hardness at 500°C this material features efficient wear protection by MML formation at HT. On the other hand the Ni-based MMC shows minor temperature and load influence on MML formation, which can be put down to its relative temperature stable fcc matrix. Wear rates of the Ni-based MMC are superior to the Fe-based material at highest testing temperature. Strain hardening and dynamic recrystallisation were found to be beneficial especially at 500°C for the Ni-based MMC.
Keywords: High temperature, abrasion, high-stress, MMC, nano-scratch, mechanically mixed layer. *Corresponding author: Markus Varga ([email protected])
1
1. INTRODUCTION High temperature (HT) abrasion is a serious issue in many industrial applications, like in steel, cement or chemical industry. Often hard particles, e.g. ores, need to be processed at HT and optimal material selection for such environments requires abrasive wear testing at application temperature. Commonly used abrasion tests are the ASTM G65 [1] and the micro-scale abrasion test [2]. Both tests operate with a block sample pressed against a rotating wheel and ball, respectively, with applying abrasive in the contact zone. By variation of the counter body and load the severity of the abrasive contact can be changed from low-stress abrasion (e.g. rubber wheel, low load) to high-stress abrasion (e.g. steel wheel, high load) which implies breaking of the abrasive [3,4]. This work concentrates on high-stress abrasion which is present e.g. in crusher applications. Abrasive breakage is rarely investigated in tribology. Nahvi et al. put it into spotlight in their work [3]. They found significant crushing for a soft abrasive (ash) in a dry sand rubber wheel apparatus, while they did not observe significant crushing of the silica abrasive. This can change significantly when using a steel wheel counterbody and high-stress conditions become dominant. Antonov et al. [5] observed a majority of abrasive breakage in such a test configuration especially at high test loads with up to ~70 % of broken abrasive, wherefrom more than half is smaller than half of the original diameter. Dube and Hutchings [6] studied especially the load influence on low- and high-stress abrasion and also investigated the fraction of broken particles (by their measurement method they were limited to the fraction smaller than half of the original diameter). The original particle shape (round vs. angular) showed a small effect, where round particles featured slightly higher breakage. For particles of 425-500 µm already at 20 N load ~ 25 % were broken, which increased to ~45 % at 60 N and ~75 % at 100 N. For low-stress conditions breakage was below 10 %. For such severe wear conditions optimal wear protection is inevitable to maintain sufficient service time of the components. Thereto often metal matrix composites (MMC) are applied, either as structural component or as local wear protection. While in single phase materials mostly the hardness dominates the abrasive wear resistance, in MMCs the amount, size, distribution, etc. of the hard phases also strongly influences the wear behaviour. Further the hardness of the hard phases is of major importance: zum Gahr [7] gives in Fig. 1a possible wear mechanisms for different types of hard phases. Hard abrasive particles with higher hardness than the matrix and hard phases easily cut or break the MMC. On the other hand soft abrasive particles (softer than the hard phases) cannot damage the entire MMC and protect the component sufficiently from abrasive attack. Too small hard phases are insufficient for wear protection, they have to be in the size range of the abrasive particles for good wear protection. When HT is present at the application, the wear regime may change, as all phases in the MMCs suffer from temperature induced softening, as well as the abrasives. Berns gives in [8] an overview of MMC behaviour. In Fig. 1b the hardness decrease of various matrix materials, hard phases and abrasives in dependence of temperature is summarised. Obviously temperature influences the various phases differently. Fe-based bcc matrices soften at >600°C rapidly, Ni-based matrices stay relatively constant up to 800°C. Typical precipitated M7C3 hard phases remain high hardness >1000 HV0.05 up to ~800°C. Flint’s hardness rapidly
2
drops and reaches the level of typical Fe- and Ni-based matrices at 500-600°C, while abrasives like SiC or Al2O3 stay at much higher level in the whole temperature regime investigated.
Figure 1: a) Possible wear mechanisms at abrasive wear of MMCs [7]; b) temperature dependence of micro-hardness of different abrasives, hardphases and matrices [9].
For soft materials or matrices of MMCs also wear reducing effects during abrasion are known. Mechanically mixed layers (MML) can be formed with abrasive particles during tribocontact [10-12]. This layers can offer a good wear protection, as they harden the surface by embedding of the hard abrasive particles. This was found to be especially beneficial at enhanced temperatures, as matrix softening favours the intermixing [11]. But the effect of wear protection is strongly dependent on the materials microstructure, application temperature [11,13,14] and applied load [4-6]. The aim of this work is to study systematically the influence of applied load and temperature on the abrasive wear behaviour. A Fe-based and Ni-based MMC were chosen to quantify the benefits of higher temperature stability of the Ni-based matrix. Differences in abrasive wear mechanisms should be investigated in detail, with special focus on in-situ formed MMLs. Temperature and load influence on changing wear regimes and MML extent should be investigated, in order to propose material concepts and economic solutions for HT abrasive conditions.
2. EXPERIMENTAL 2.1 Materials Two heat resistant metallic materials with hardphase reinforcement were investigated, due to their applicability in HT wear environment. A Fe-base and a Ni-base cast material were chosen to identify differences in the HT wear behaviour by different matrices. Both materials contain a significant amount of C to favour hardphase precipitation. The chemical compositions of the both alloys are given in Table 1. Samples were taken from cast rods in case of the Fe-based MMC and of a cast block for the Ni-based MMC. The microstructures of the two materials are given in Fig. 2 as seen by optical microscopy (OM). The Fe-based MMC shows dendritic structure with Cr-carbide precipitation around the grain boundaries. Carbides are <10 µm at a fraction of ~15 % as measured by QWIN® and of Cr7C3 type as found by XRD. The ferritic matrix in between shows a grain size of 30-40 µm. The resulting compound hardness at RT is ~275 HV10. The Ni-based MMC is given in Fig. 2b, also featuring dendritic structure: the austenitic matrix is surrounded 3
by a fine-structured Cr-carbide network, with occasional WxC precipitations. The grain size of the matrix zones is ~20-30 µm, with some larger areas in the range of 100 µm without carbides. Carbide content is ~13 -% and XRD-measurements identified them of Cr23C6 type. This alloy has similar hardness at RT as the Fe-based with ~245 HV10. Table 1: Chemical composition of the materials investigated. [wt.%] C Si Mn Cr W Fe Ni
Fe-based MMC 1.2-1.4 1.0-2.5 0.5-1.0 27-30 balance -
Ni-based MMC 0.35-0.55 1.0-2.0 <1.5 27-30 4.0-6.0 15 balance
Figure 2: Microstructures of the two alloys as seen by OM: a) Fe-based MMC, b) Ni-based MMC.
2.2 Hot hardness test The hardness is an important parameter in wear of materials. Especially the hardness at application temperature is of great interest, when dealing with HT abrasion. At the Austrian Competence Centre for Tribology (AC²T) a Hot Hardness Test (HHT) was developed [15] allowing for hardness testing up to 1000°C. The hardness test is done according Vickers method with 10 kg test load (HV10) and takes place in vacuum atmosphere (<5 mbar) to avoid oxidation of the surface. Test temperatures were chosen as RT, 100°C, 300°C, 500°C, 600°C, 700°C and 800°C with 3 indents each and the test series was repeated on a second sample for statistic evaluation. The indent diagonals were measured after cooling down of the samples by OM and the Vickers hardness were calculated. Failures due to thermal shrinkage were found to be insignificant for the materials investigated and are therefore not taken into account [15].
2.3 Abrasive wear test The HT-Continuous Abrasion Test (HT-CAT) was chosen, in order to test high-stress abrasion, as it occurs in high loaded components like crusher systems. The test is of block-on-ring configuration, based on an ASTM G65 [1] test rig, enhanced for HT testing (Fig. 3a). Thereto an inductive heating system enables the
4
heating of the sample up to ~700°C. Tests were done at RT, 500°C and 700°C. Temperature is controlled by a type K thermocouple, placed in a drilled hole near the tribocontact. The rubber wheel is changed to a Hardox® 400 wheel and has a circumferential speed of 1 m/s. The abrasive was chosen according the ASTM G65 standard: 212-300 µm rounded quartz as displayed in Fig. 3b. The flow rate was reduced to 180 g/min to avoid excessive cooling of the sample by the cold abrasive. To study the load influence, tests were done at 10 N, 45 N and 80 N. Test time/distance was adjusted to acquire wear scars of similar size, to keep contact conditions comparable between the tests (width: 12 mm, length: ~15 mm at the end of the test) [cf. 16]: 10 N-20 min/1200 m, 45 N-5 min/300 m, 80 N-2 min/120 m. No running-in phenomena were observed even at shortest test duration for this high-stress regime, i.e. steady-state conditions were present for all experiments. Starting from an initial line contact according Hertz’s contact theory, the contact pressure decreases with test duration and increasing conformity of the counter bodies. Nevertheless, in abrasive conditions where particle fracture takes place on a regular basis the contact pressure is limited by the crushing strength of the abrasive used [cf. 17]. A similar sized quartz abrasive used in [3] showed fracture at ~7 N load. The failure was catastrophic, i.e. the particle did not just break in two pieces, but shatters. Tests were repeated three times each condition. Wear rates are calculated by measuring the weight loss, calculating the volume loss via the material’s density and dividing by the wear distance. Test parameters are summarised in Table 2 and details on the HT-CAT can be found in [11]. Collection of the abrasive during test allows the evaluation of contact severity. The fraction of abrasive passing the wear track was evaluated similar to Antonov et al. [5]. Thereto on both sides of the wheel a sheet of paper was mounted with its edges at the height of contact. By these sheets of paper the flow of abrasive which did not pass the wear track was separated from the abrasive in contact, and both fractions were weighted afterwards. In our test configuration with a sand flow of 180 g/min (reduced from the G65 standard) a fraction of 37-40 % passes the wear track, while 60-63 % never come in contact with the sample. At one test of each configuration an abrasive sample was taken for 1 min at the end of the test duration and a sieve analysis undertaken. The fractions corresponding to the unbroken abrasive, (91.8 % >250 µm, 8 % 200-250 µm, 0.2 % 150-200 µm), were reduced by the percentage which did not pass the wear track. After sieve analysis the fraction of broken abrasive can be seen, giving a measure of the contact severity [4]. Further, one must keep in mind, that broken abrasive may have higher abrasivity due to the sharp edges compared to the rounded original abrasive [18.19]. Post-tribotest analyses were done on worn surfaces and cross-sections. Surface Scanning Electron Microscopy (SEM, Zeiss® Supra 55 VP) was utilised to calculate the covered surface by abrasive due to tribotesting. Thereto the surface is captured in Back Scattered Electron (BSE) mode, which shows abrasive in dark onto the bright metallic surface of the sample. By quantitative image analysis (QWIN ®) the percentage of the covered surface by abrasive is calculated. As the abrasive get intermixed with the matrix material to a certain limit the depth of maximal penetration of the abrasive is measured on cross-sections. Details of the depth measurement can be found in [11,13] and area measurement in [13].
5
Figure 3: a) HT high-stress abrasion test rig [11] and b) abrasive used for abrasion testing. Table 2: Main parameters used during HT high-stress abrasion testing. Parameter Temperature Normal load – sliding distance Sliding speed Counter body Abrasive Wear quantification
Value RT, 500°C, 700°C 10 N – 1200 m, 45 N – 600 m, 80 N – 120 m 1 m/s Hardox 400, 360 HV10, ø 232 × 12 mm Standard Ottawa silica sand, round, 212-300 µm at 180 g/min flow rate Volume loss / sliding distance → wear rate [mm³/m]
Further measurements were carried out on cross-sections of the two materials. Micro-hardness measurements were performed on different features of the microstructures to identify work hardening and supporting effect of the carbide network. Measurements were done with micro-hardness tester Future-Tech® FM-700 at RT. Nano-scratch experiments were carried out on the two materials to study the bonding and deformation mechanisms at carbide-matrix boundaries, aiming the simulation of abrasive attack by a single particle. The Hysitron® Triboindenter TI900 was utilised for this nano-scratching, which was also performed at RT. A Berkovich-tip was used for nano-scratching.
3. RESULTS 3.1 Hot hardness of the materials The hardness as important parameter in wear of materials was investigated up to 800°C by Vickers method. The results are displayed in Fig. 4a. Both alloys have relative low hardness compared to other wear protective materials for abrasive environment e.g. hypereutectic hardfacings [13]. The Fe-based MMC starts at ~275 HV10 at RT and drops linearly to ~180 HV10 at 500°C. >500°C the hardness loss is more pronounced due to excessive matrix softening, resulting in a minimum of 40 HV10 at 800°C. The Ni-based MMC starts at ~245 HV10 at RT and slowly decreases to ~130 HV10 at highest testing temperature of 800°C. The hardness progress is much shallower compared to the ferritic Fe-based MMC due to the temperature stable fcc-matrix of the Ni-based MMC. Especially the differences in hardness at temperatures >600°C will be interesting regarding HT wear behaviour. 6
Figure 4: a) Hot hardness and b) abrasive wear rates of the materials investigated.
3.2 Abrasive wear results High-stress abrasive wear rates measured at the different loads and temperatures are displayed in Fig. 4b. Wear rates were normalised by the distance, to be able to compare the different distances tested at the various loads. (Additional normalisation by the load was omitted, for better visualisation in the diagram.) Obviously the various loads entail different levels of wear rates. Temperature influence is less pronounced compared to the load dependency. Nevertheless, for the Fe-based MMC a significant increase of wear rate with ascending temperature can be measured for all load levels investigated. At 10 N load the increase for the Fe-based material is 25 % from RT to 500°C and further pronounced (~40 %) from 500°C to 700°C. On the other hand, the Ni-based alloy shows insignificant temperature influence at 10 N test load. Wear rates of the Fe-based MMC at 45 N load increase from 0.034 mm³/m at RT to ~0.039 mm³/m at 500°C and 0.054 mm³/m at 700°C, which stands for a increase of 37 % from RT to 700°. At 80 N the temperature influence also worsens wear loss by ~27 % from RT to 700°C. I.e. the rise of wear rate due to temperature influence decreases with increasing test load at the Fe-based MMC. The temperature dependence of the Ni-based MMC shows very interesting behaviour at higher loads. While at 10 N no temperature influence can be measured (within the accuracy of the applied test- and measurement methods), there manifests a minimum at 500°C for 45 N and 80 N test load. At 45 N, both, the RT and 700°C result are ~0.038 mm³/m and the 500°C wear rate is ~0.035 mm³/m. This effect gets more pronounced at 80 N, where the highest wear rate was measured at RT (0.103 mm³/m), followed by a minimum of 0.075 mm³/m (-37 %) and is increasing to 0.086 mm³/m at 700°C. Possible reasons for such unexpected behaviour will be discussed within the following sections.
3.3 Wear mechanisms at HT abrasion Abrasive wear rates are given in the previous section. Now the wear mechanism should be investigated in detail. As both materials are of similar structure, hard carbides in a soft matrix, wear mechanisms on surface SEM were found to be very similar, as displayed in Fig. 5-6. A ploughing wear mechanisms is observed at
7
the larger matrix zones, as clearly visible in Fig. 5a, with a lot of plastic deformation and abrasive embedding. Only at this lowest stress level a preferred direction of the abrasive, corresponding with the wheel movement, can be seen. Due to the free moving abrasive and random breakage of the abrasive, there is no distinct orientation of the wear grooves at higher stress levels. This is contrary to rubber-wheel testing, where temporary abrasive particles are fixed in the rubber and may groove the sample’s surface [cf. 3]. With increasing temperature and load the abrasive embedding becomes more dominant, covering the wear scars. Therefore cross-sections through the wear scars were performed for all conditions. The 45 N images are given in Fig. 7. For these materials the MML formation is clearly visible at all conditions. For the Fe-based MMC the increasing temperature clearly benefits MML formation and at 700°C a very thick and almost surface covering layer can be found. While at RT for the Ni-based MMC the layer formation is similar, at higher temperatures it is markedly less pronounced than at the Fe-based MMC. Quantitative results on the MML spread will be discussed in the following section 3.4.
Figure 5: Surface SEM-BSE after testing of Fe-based MMC: a) RT, 10 N; b) RT, 80 N; c) 700°C, 10 N; d) 700°C, 80 N.
8
Figure 6: Surface SEM-BSE after testing of Ni-based MMC: a) RT, 10 N; b) RT, 80 N; c) 700°C, 10 N; d) 700°C, 80 N.
Figure 7: Cross sections of wear zones after 45 N abrasion testing at different temperatures as seen by OM: a-c) Febased MMC; d-f) Ni-based MMC after a, d) RT; b, e) 500°C and c, f) 700°C testing.
Carbide regions impede the intermixing with the abrasive. At the Fe-based material this is obvious at RT and 500°C, while at the highest temperature of 700°C no effect of the carbide network is discernible. It seems that the material, or at least the matrix is soft enough, that massive plastic deformation can take place, 9
including displacement and intermixing of the carbides. This is different at the Ni-based material. Here the carbide network hinders a massive intermixing, which does not change even at highest testing temperature of 700°C. This means that the carbide network and matrix compound at the Ni-based material is much more temperature stable than at the Fe-based material. The load bearing ability of the carbide network will be discussed in section 3.5.
3.4 Abrasive interaction during tribocontact In order to study abrasive investigation two data sources were investigated: abrasive breakage during testing as well as post-test analysis of the covered surface by abrasive and abrasive penetration depth on cross sections. The severity of the contact can be assessed by evaluation of the broken abrasive during tribocontact. Thereto the abrasive during test was collected and afterwards a sieve analysis undertaken. After taking out of the fraction which did not pass the wear track [cf. 5] Fig. 8a can be drawn, showing the amount of broken abrasive at the various conditions. Increasing load clearly leads to a higher fraction of broken abrasive, i.e. the contact severity increases with the applied load and high-stress conditions get dominant. Interestingly the two alloys investigated show different temperature dependence. The Fe-based MMC shows a minimum of abrasive breakage at 500°C for high loads (45, 80 N), while the Ni-based material shows a maximum at this temperature for all loads investigated. For explanation the next diagram Fig. 8b and Table 3 can be helpful. There the abrasive coverage after the test, i.e. the amount of embedding during tribocontact, is displayed, as measured on surface SEM and evaluated by quantitative image analysis, and the depth of abrasive penetration measured on cross-sections. For the Fe-based MMC a clear load- and temperature influence was identified. Surface coverage increases with ascending temperature, as well as with increasing load, where the load influence is pronounced, leading to 52-58 % coverage at 80 N. At 10 N and 80 N the temperature influence is less significant, but at 45 N it is very pronounced: the surface coverage changes from ~38 % at RT to 50 % at 700°C. The surface coverage on the Ni-based MMC on the other hand shows just minor influence in the investigated borders. Nevertheless, temperature influence is obvious. At RT it is about 15 % for all loads, at 500°C 20 % were measured. The only differentiation between loads is possible at 700°C, where 23 % coverage were found at 10 N and ~28 % both at 45 and 80 N. Table 3: Abrasive penetration depth as measured by OM on cross-sections of the wear scars. Fe-based MMC
Ni-based MMC
Load
RT
500°C
700°C
RT
500°C
700°C
10 N
19 ± 4 µm
26 ± 4 µm
39 ± 9 µm
21 ± 5 µm
22 ± 7 µm
25 ± 7 µm
45 N
17 ± 4 µm
20 ± 6 µm
28 ± 6 µm
18 ± 2 µm
20 ± 2 µm
21 ± 8 µm
80 N
21 ± 9 µm
25 ± 9 µm
43 ± 14 µm
17 ± 4 µm
24 ± 5 µm
18 ± 5 µm
10
Figure 8: a) Abrasive breakage during high-stress abrasion testing; b) abrasive coverage as measured via surface SEM-BSE.
For the penetration depth on the other hand no clear trend can be observed for both materials. Load influence is insignificant, temperature influence is detectable at the Fe-based material. Generally it is reasonable, that softening of the Fe-based matrix with increasing temperature also entails increasing penetration depth, as previously reported in [4]. This softening effect is much less pronounced at the Ni-based material due to the fcc-structure and therefore more temperature stability in the investigated temperature range. Hence no significant temperature influence was found for the Ni-based MMC regarding penetration depth. The difficulties with this evaluation lie also in the 2-dimensional data source: only one cross-section throughout the wear scar can be analysed. Further the inhomogeneous microstructure of the MMCs and the carbide networks influence the MML formation, leading to the large deviations. Nevertheless, a clear difference between the Fe-based and the Ni-based MMC becomes obvious: the coverage and penetration depth is strongly depending on temperature and load for the Fe-based MMC, while the Ni-based MMC shows nearly no changes with the load and just minor changes of the MML formation with the temperature in the investigated range.
3.5 Detailed investigations on the micro-scale To study the fundamentals of the abrasive attack, detailed measurements on the micro-scale were conducted. By use of micro-hardness measurements comparing the matrix with the carbide zones the stability of the carbide network and matrix-backup was investigated. Further, the temperature influence on work hardening was measured by micro-hardness below the wear track. Nano-scratch experiments should study the fundamental micro-mechanisms of the abrasive attack. For the investigation of matrix-backup and carbide network stability micro-hardness measurements were done with loads from 98 mN (HV0.01) up to 9.81 N (HV1). Indents in matrix regions should point out the load effect on the Vickers-hardness, as with decreasing loads the elastic component of the indent becomes more significant, the hardness-value increases. Indents in zones of the carbide network on the other hand aim
11
on the identification of the load which the carbide network can carry: if the difference between matrix and carbide network hardness is minor, it is thought, that the carbides have insignificant effect on the abrasive attack by large particles. When the hardness of the carbide network indents becomes higher, it can carry the load and a beneficial effect on the wear resistance is expected. The results of these micro-hardness investigations are displayed in Fig. 9a. The matrix of the Fe-based MMC is slightly harder with ~250 HV than the Ni-based with ~200 HV. A significant influence of the test load on the matrix-hardness is not visible in the investigated load range. The Ni-based matrix shows ~200 HV hardness, which increases at the lowest test load to 230 HV0.01. At the Fe-based MMC the difference between carbide network and matrix is minor at 9.81 N and 2.94 N, but becomes distinct at loads <0.98 N. At the Ni-based MMC already at 9.81 N a pronounced difference between matrix and carbide network can be measured, which increases slightly at 2.94 N and stays constant until low loads. (The difference at 10 mN becomes less, as the matrix hardness appears higher because of the increasing elastic component.) These measurements lead to the conclusion that the Cr23C6 carbide network of the Ni-based MMC is much stronger against abrasive penetration as the Cr7C3 carbide network of the Fe-based MMC.
Figure 9: Micro-hardness measurements: a) Influence of load on matrix-backup of the carbide network; b) work hardening of the matrix during tribological interaction measured on cross-sections through the wear track.
Additionally to the wear reducing effect of the hard phases, work hardening can take place during tribocontact in the matrix regions of these materials. To study this behaviour micro-hardness measurements with 98 mN (HV0.01) were done on cross-sections. Starting at the worn surface hardness lines were done until ~200 µm depth below the wear track. Just the hardness of the matrix regions was measured and carbides avoided. This was done for both materials at 80 N abrasive wear tracks, as there the highest hardening effect is suspected. All three test temperatures were compared and results of representative hardness-curves are given in Fig. 9b. Work hardening at the Ni-based MMC is very pronounced at all temperatures investigated, while at the Fe-based MMC it is almost disappearing at HT. The Ni-based matrix features ~230 HV0.01 as shown in Fig. 9a, which increases to >500 HV0.01 in the first 10 µm under the 12
wear track. The progress is very smooth: the hardness decreases slowly. The original matrix hardness was found in depths of 150-200 µm, depending on the carbide structure. It has to be noted, that carbide lines do not stop the hardening effect in surface near regions, and it could be also measured below carbide clusters. At the Fe-based MMC hardening was distinct in the first <10 µm, and its effect decreases with increasing temperature. At RT hardening from 250 HV0.01 bulk matrix hardness to ~450 HV0.01 was measured, at HT ~350-370 HV0.01 were measured near the worn surface. At RT and 500°C a hardening effect could be measured until 150-200 µm, while at 700°C the original matrix hardness is reached after 40-50 µm. The effect of abrasive attack with even smaller loads as possible by single abrasive particles was investigated by nano-scratch experiments with 10 mN down to 1.5 mN. At attack by these small loads the phases do not act like a “compound” any more, and the properties of the single phases and the interface becomes important [cf. 20]. A comparison of the matrices at 1.5 mN is given in Fig. 10a. A micro-ploughing mechanism dominates the matrix zones, where the material is plastically deformed from the groove to the ridges. Wear scars at the Fe-matrix are less pronounced than at the Ni-matrix, which can be put down to the lower hardness and better deformability of the Ni-matrix. In Fig. 10b a length profile through the scratch is given, where the scratch starts on the left with penetration of the matrix, reaching a carbide at 0-coordinate, passing the carbide and entering the matrix again at the end of the scratch. The damage of the carbide is in the range of 3-8 µm depth, but differences are very small. Cr23C6 are slightly harder than Cr7C3, what might be the reason for the different scratch depth [9].
Figure 10: Profiles of the nano-scratches at 1.5 mN load: a) cross-section at the matrix, b) length-section of the entire scratch.
13
Figure 11: SEM-BSE detail of the 10 mN scratch in the Fe-based MMC: matrix-carbide-matrix boundary.
For good imaging by SEM the scratches at 1.5 mN were too small, thereto additional scratches at 5 and 10 mN were performed. Exemplarily the scratch at the phase interface from matrix to carbide to matrix of the Fe-based MMC is shown in Fig. 11. The micro-ploughing wear mechanisms is clearly visible on the left, with the bright ridges. At the boundary to the carbide the ridges are ploughed over the hard carbide. One can suspect, that material is easily removed in a further abrasive attack. At the carbide no plastic deformation is visible. Here a cutting wear mechanism must dominate, with also some micro-cracking at the end of the carbide. The interface in the wake of the scratch also shows some crack formation, which is not visible on the surging border of the carbide.
4. DISCUSSION High-stress abrasive wear rates of the here investigated materials are on a relative low level, compared with other high sophisticated hardfacings like e.g. hypereutectic FeCrC alloys. Varga et al. compares in [4,9,11,13,14] different high temperature materials regarding their abrasion resistance, also within the test field is a high sophisticated complex-alloyed hardfacing with a high content of hard phases. Wear rates of this hardfacing at enhanced temperatures were much higher than at the Fe-based MMC also studied within this work. The in-situ MML formation in [4,11] offers better wear resistance than a high hardphase content, were brittle wear behaviour was dominant. Aim of this work is to investigate, if even better wear resistance can be reached by Ni-based materials with austenitic matrix and similar hardphase amount. While wear rates at RT are higher for the Ni-based MMC at all loads investigated, this changes at elevated temperatures. At 500°C the Ni-based material is slightly better than the Fe-based MMC, at highest temperature of 700°C a pronounced difference can be measured. The Ni-based MMC is superior to all materials investigated previously [4,11,13]. As investigated by micro-hardness experiments the load carrying capacity of the Cr23C6 carbide network in the Ni-based alloy is much better than the Cr7C3 network of the Fe-based alloy. This may be a decisive factor, especially at enhanced temperatures, when matrix softening exacerbates carbide backup. As found by the nano-scratch experiments the single carbides are an effective measure against abrasive wear loss on the micro-scale. They are well bonded to the matrix and feature very little wear loss, which may be decreased to zero when dealing with softer abrasives than the carbide [7]. The matrix of both alloys becomes
14
easily plastically deformed, which is slightly more pronounced at the Ni-based material. Carbides won’t undergo a significant loss in hardness in the investigated temperature range [8], contrary to the matrices, where especially the Fe-based matrix shows a distinct loss in hardness >500°C as indicated by the macro hot hardness measurements. Despite their hard phase network both materials form MMLs, which were found to reduce wear loss e.g. [10,21,22]. Nevertheless, the MML extension is much smaller at the Ni-based material, but entails superior wear resistance. This leads to the assumption, that thick MMLs do not necessarily entail best wear resistance, but also a thinner MML may lead to the wear reducing effect. Load influence on abrasive coverage was not significant for most conditions at the Ni-based alloy, and increasing coverage is mainly driven by temperature rise. For the Fe-based MMC also the load has significant influence on the coverage. Penetration depth on the other hand was found to be just dependent on temperature and material softening, respectively. An interesting behaviour of the wear rates with temperature increase shows the Ni-based MMC, especially at high loads. The wear rate at RT is significantly higher than at 500°C and increasing again at 700°C. Similar behaviour at HT abrasion was also reported by Berns (ed.) [8]: he found a decreasing wear rate until a critical temperature (dependent on the matrix structure) due to MML formation. Exceeding this temperature, wear rates increase again, because of dynamic recrystallisation and loss of work-hardening ability. The critical temperature is in the range of 0.6× melting temperature, further certain conditions of deformation have to be present according [8]. One must keep in mind, that this regime is limited to very surface near regions in the wear zone.
Figure 12: Deformation of Ni-based MMC after 80 N testing as seen on cross-sections after: a) RT, b) 500°C and c) 700°C testing.
To investigate the deformation mechanisms in more detail the cross-sections of the wear tracks of the Nibased MMC were etched with Ni grain boundary etchant (25 ml distilled water, 50 ml hydrochloric acid 32 %, 15 g iron-III chloride, 3 g ammonium tetrachlorocuprate(II)). Details of the cross-sections in surfacenear regions are given in Fig. 12. At RT a massive formation of dislocations can be found in the first micrometres beneath the wear track. These dislocations are also visible between the carbides, i.e. deformation is not stopped by the carbide network. As found by the micro-hardness lines the work-hardening reaches down to 150-200 µm. At 500°C the etchant reveals a clearly different image. Almost no dislocation lines can be found, but small recrystallized grains, probably by dynamic recrystallisation [8,23,24]. This may also goes conform to a hardening effect in the very surface near zones by grain refinement, entailing the low wear rates measured. At 700°C the recrystallisation is much more pronounced, reaching deep beneath the
15
tribocontact. Recrystallisation starts at ~0.6× melting point, which is ~760°C at the Ni-based and ~810°C at Fe-based material. This temperature may be reached at the Ni-based material within the tribocontact due to additional frictional heat. The combined effect of dynamic recrystallisation and MML formation at the Nibased MMC obviously benefits the wear resistance at 500°C. Future EBSD measurements may substantiate this hypothesis. At higher temperatures the softening of the material and easier plastic deformation seems to entail increasing wear loss again. Nevertheless the HT high-stress abrasive wear rates >500°C of the Nibased material is superior to any other material reported so far, and the application and cost effectiveness of such MMCs should be studied in the future.
5. CONCLUSIONS High temperature abrasive wear at high-stress condition was studied on two metal matrix composites (MMCs) with ~15 % hard phases. A Ni-based matrix was compared with a Fe-based matrix, in order to study the effect of the matrix’ lattice and its influence on wear resistance at increasing load and temperature in abrasive contact. Following conclusions can be drawn:
The test load has significant influence on the wear rate and severity of the contact. A larger fraction of the abrasive is broken at higher loads.
At the Fe-based MMC wear rates increase from RT to 700°C, with a more pronounced ascend from 500°C to 700°C which goes conform to a more distinct softening of the material >500°C.
The Ni-based MMC shows slightly higher wear rates at RT than the Fe-based MMC. At 500°C the Ni-based material is slighly better and at 700°C the difference is pronounced. The higher hot hardness of the Ni-based material is one reason for its superity. Also the Cr23C6 carbide network of the Nibased material can carry higher loads then the Cr7Cr3 network of the Fe-based MMC.
Mechanically mixed layer (MML) formation was found for both alloys. At the Fe-based MMC it is increasing with load and temperature. At the Ni-based alloy a temperature influence could be observed, but much less pronounced than at the Fe-based material.
Pronounced strain hardening was found at the Ni-based material at all temperatures investigated. Further dynamic recrystallisation is probable at higher temperatures leading to decreasing wear rates at 500°C.
Fundamental abrasive wear mechanisms were studied by nano-scratch experiments. Micro-ploughing dominates in matrix regions, while micro-cutting was found at the carbides, accompanied by microcracking at the edges of carbides.
As the Ni-based material shows much better wear resistance at high temperatures, its application in industrial plants and cost effectiveness will be investigated. In the future detailed studies of the wear resistance of different carbide structures, e.g. deposited by welding, may further enhance the wear resistance.
6. ACKNOWLEDGEMENTS This work was funded by the Austrian COMET Program (Project K2 XTribology, no. 849109) and carried out at the “Excellence Centre of Tribology”. Special thanks goes to Harald Rojacz for intense discussions, 16
Anifa MF. Azhaarudeen for evaluating experiments and Christian Tomastik for performing nano-scratch experiments.
7. REFERENCES 1. ASTM G65-04: Standard test method for measuring abrasion using the dry sand/rubber wheel apparatus, American Society for Testing and Materials (2010) 2. RI. Trezona, DN. Allsopp, IM. Hutchings: Transition between two-body and three-body abrasive wear: influence of test conditions in the microscale abrasive wear test, Wear 225-229 (1999) 205214 3. SM. Nahvi, PH. Shipway, DG. McCartney: Effects of particle crushing in abrasion testing of steels with ash frombiomass-fired powerplants, Wear 267 (2009) 34-42 4. M. Varga, AMF. Azhaarudeen, K. Adam, E. Badisch: Influence of load and temperature on abrasion of carbidic cast steel and complex alloyed hardfacing, Key Engineering Materials 674 (2016) 313318 5. M. Antonov, I. Hussainova, R. Veinthal, J. Pirso: Effect of temperature and load on three-body abrasion of cermets and steel, Tribology International 46 (2012) 261-268 6. NB. Dube, IM. Hutchings: Influence of particle fracture in high-stress and low-stress abrasive wear of steel, Wear 233-235(1999) 246-256 7. KH. zum Gahr: Microstructure and Wear of Materials, Elsevier Science Publishers, Tribology Series 10 (1987) 8. H. Berns (ed.): Hartlegierungen und Hartverbundwerkstoffe, Springer Berlin, Heidelberg (1998) 9. M. Varga: High temperature abrasion in sinter plants and their cost efficient wear protection, Dissertation Montanuniversität Leoben (2016) 10. LC. Jones, RJ. Llewellyn: Sliding abrasion resistance assessment of metallic materials for elevated temperature mineral processing conditions, Wear 267 (2009) 2010-2017 11. M. Varga, H. Rojacz, H. Winkelmann, H. Mayer, E. Badisch: Wear reducing effects and temperature dependence of tribolayer formation in harsh environment, Tribology International 65 (2013) 190-199 12. M. Varga, H. Winkelmann, E. Badisch: Impact of microstructure on high temperature wear resistance, Procedia Engineering 10 (2011) 1291-1296 13. M. Varga, AMF. Azhaarudeen, E. Badisch: Influence of in-situ formed tribolayer on abrasive wear reduction, Materials Science Forum 825-826 (2015) 85-92 14. M. Varga, High temperature abrasive wear of metallic materials, accepted for publication in Wear (2017) 15. M. Varga, M. Flasch, E. Badisch: Introduction of a novel tribometer especially designed for scratch, adhesion and hardness investigation up to 1000°C, Proceedings of the Institution of Mechanical Engineers Part J: Journal of Engineering Tribology 231-4 (2017) 469-478 16. IM. Hutchings: Abrasive an erosive wear tests for thin coatings: a unified approach, Tribology International 31/1-3 (1998) 5-15
17
17. H. Rojacz, H. Pahr, S. Baumgartner, M. Varga: High temperature abrasion resistance of differently welded structural steels, Tribology International, in press (2017) 18. DA. Kelly, IM. Hutchings: A new method for measurement of particle abrasivity, Wear 250 (2001) 76-80 19. S. Wirojanupatump, PH. Shipway: Abrasion of mild steel in wet and dry conditions with the rubber and steel wheel abrasion apparatus, Wear 239 (2000) 91-101 20. M. Varga, S. Leroch, H. Rojacz, M. Rodríguez Ripoll: Study of wear mechanisms at high temperature scratch testing, accepted for publication in Wear (2017) 21. H. Berns: Microstructural properties of wear-resistant alloys, Wear 181-183 (1995) 271-279. 22. A. Fischer: Mechanisms of high temperature sliding abrasion of metallic materials, Wear 152 (1992) 151-159. 23. U. Cihak-Bayr, G. Mozdzen, E. Badisch, A. Merstallinger, H. Winkelmann: High plastically deformed sub-surface tribozones in sliding experiments, Wear 309 (2014) 11-20. 24. H. Rojacz, G. Mozdzen, F. Weigel, M. Varga, Microstructural changes and strain hardening effects in abrasive contacts at different relative velocities and temperatures, Mater. Charact. 118 (2016) 370381.
8. FIGURES Figure 1: a) Possible wear mechanisms at abrasive wear of MMCs [7]; b) temperature dependence of micro-hardness of different abrasives, hardphases and matrices [9]. Figure 2: Microstructures of the two alloys as seen by OM: a) Fe-based MMC, b) Ni-based MMC. Figure 3: a) HT high-stress abrasion test rig [11] and b) abrasive used for abrasion testing. Figure 4: a) Hot hardness and b) abrasive wear rates of the materials investigated. Figure 5: Surface SEM-BSE after testing of Fe-based MMC: a) RT, 10 N; b) RT, 80 N; c) 700°C, 10 N; d) 700°C, 80 N. Figure 6: Surface SEM-BSE after testing of Ni-based MMC: a) RT, 10 N; b) RT, 80 N; c) 700°C, 10 N; d) 700°C, 80 N. Figure 7: Cross sections of wear zones after 45 N abrasion testing at different temperatures as seen by OM: a-c) Fe-based MMC; d-f) Ni-based MMC after a, d) RT; b, e) 500°C and c, f) 700°C testing. Figure 8: a) Abrasive breakage during high-stress abrasion testing; b) abrasive coverage as measured via surface SEM-BSE. Figure 9: Micro-hardness measurements: a) Influence of load on matrix-backup of the carbide network; b) work hardening of the matrix during tribological interaction measured on cross-sections through the wear track. Figure 10: Profiles of the nano-scratches at 1.5 mN load: a) cross-section at the matrix, b) length-section of the entire scratch. Figure 11: SEM-BSE detail of the 10 mN scratch in the Fe-based MMC: matrix-carbide-matrix boundary.
18
Figure 12: Deformation of Ni-based MMC after 80 N testing as seen on cross-sections after: a) RT, b) 500°C and c) 700°C testing.
9. TABLES Table 1: Chemical composition of the materials investigated. [wt.%] C Si Mn Cr W Fe Ni
Fe-based MMC 1.2-1.4 1.0-2.5 0.5-1.0 27-30 balance -
Ni-based MMC 0.35-0.55 1.0-2.0 <1.5 27-30 4.0-6.0 15 balance
Table 2: Main parameters used during HT high-stress abrasion testing. Parameter Temperature Normal load – sliding distance Sliding speed Counter body Abrasive Wear quantification
Value RT, 500°C, 700°C 10 N – 1200 m, 45 N – 600 m, 80 N – 120 m 1 m/s Hardox 400, 360 HV10, ø 232 × 12 mm Standard Ottawa silica sand, round, 212-300 µm at 180 g/min flow rate Volume loss / sliding distance → wear rate [mm³/m]
Table 3: Abrasive penetration depth as measured by OM on cross-sections of the wear scars. Fe-based MMC
Ni-based MMC
Load
RT
500°C
700°C
RT
500°C
700°C
10 N
19 ± 4 µm
26 ± 4 µm
39 ± 9 µm
21 ± 5 µm
22 ± 7 µm
25 ± 7 µm
45 N
17 ± 4 µm
20 ± 6 µm
28 ± 6 µm
18 ± 2 µm
20 ± 2 µm
21 ± 8 µm
80 N
21 ± 9 µm
25 ± 9 µm
43 ± 14 µm
17 ± 4 µm
24 ± 5 µm
18 ± 5 µm
19
Highlights
The Ni-based MMC shows better wear resistance at 700°C than the Fe-based MMC
The Cr23C6 carbide network of the Ni-based material is more resistant than the Cr7Cr3 network of the Fe-based MMC.
Mechanically mixed layer formation was found for both alloys.
Pronounced strain hardening was found at the Ni-based material at all temperatures investigated.
Micro-ploughing dominates in matrix regions, while micro-cutting was found at the carbides.
20
www.elsevier.com/locate/wear
PII: DOI: Reference:
S0043-1648(16)30688-3 http://dx.doi.org/10.1016/j.wear.2017.04.020 WEA102149
To appear in: Wear Received date: 28 November 2016 Revised date: 11 April 2017 Accepted date: 25 April 2017 Cite this article as: M. Varga and E. Badisch, Temperature and load influence on in-situ formed layers during high temperature abrasion, Wear, http://dx.doi.org/10.1016/j.wear.2017.04.020 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.
Temperature and load influence on in-situ formed layers during high temperature abrasion M. Varga*, E. Badisch AC2T research GmbH, Viktor-Kaplan-Straße 2C, 2700 Wiener Neustadt, Austria.
Abstract Abrasive wear at high temperature (HT) applications is a serious issue in industry, limiting the lifetime of core components, e.g. in steel- or cement production. In literature abrasive wear is commonly linked to material’s hardness, but it is known that microstructure and temperature show also major influence. Current studies also found wear protecting effects by the in-situ formation of mechanically mixed layers (MML) with the abrasive, which can be especially beneficial at HT. To investigate the temperature and load influence of this MML formation at high-stress abrasive conditions a modified ASTM G65 test, allowing for temperatures up to 700°C, was utilised. Two metal matrix composite (MMC) materials prone to MML formation were chosen: a Ni-based and Fe-based cast alloy with carbide content of ~15 %. Thereby the MMC matrix influence on MML formation was studied. Load was varied from 10 to 45 and 80 N, and temperature from room temperature to 500°C and 700°C. Abrasive is collected during tests in order to determine the severity of the contact during the wear process. Post-test analysis investigate the MML coverage and depth in order to estimate the wear protecting effect. Further microhardness investigations and nano-scratch experiments should identify fundamental wear mechanisms and load bearing capacity of the hard phase network. It was found that increasing load leads to higher severity of the contact (more abrasive breakage). MML coverage was strongly dependent on load and temperature at the Fe-based material. Despite a pronounced drop of the hardness at 500°C this material features efficient wear protection by MML formation at HT. On the other hand the Ni-based MMC shows minor temperature and load influence on MML formation, which can be put down to its relative temperature stable fcc matrix. Wear rates of the Ni-based MMC are superior to the Fe-based material at highest testing temperature. Strain hardening and dynamic recrystallisation were found to be beneficial especially at 500°C for the Ni-based MMC.
Keywords: High temperature, abrasion, high-stress, MMC, nano-scratch, mechanically mixed layer. *Corresponding author: Markus Varga ([email protected])
1
1. INTRODUCTION High temperature (HT) abrasion is a serious issue in many industrial applications, like in steel, cement or chemical industry. Often hard particles, e.g. ores, need to be processed at HT and optimal material selection for such environments requires abrasive wear testing at application temperature. Commonly used abrasion tests are the ASTM G65 [1] and the micro-scale abrasion test [2]. Both tests operate with a block sample pressed against a rotating wheel and ball, respectively, with applying abrasive in the contact zone. By variation of the counter body and load the severity of the abrasive contact can be changed from low-stress abrasion (e.g. rubber wheel, low load) to high-stress abrasion (e.g. steel wheel, high load) which implies breaking of the abrasive [3,4]. This work concentrates on high-stress abrasion which is present e.g. in crusher applications. Abrasive breakage is rarely investigated in tribology. Nahvi et al. put it into spotlight in their work [3]. They found significant crushing for a soft abrasive (ash) in a dry sand rubber wheel apparatus, while they did not observe significant crushing of the silica abrasive. This can change significantly when using a steel wheel counterbody and high-stress conditions become dominant. Antonov et al. [5] observed a majority of abrasive breakage in such a test configuration especially at high test loads with up to ~70 % of broken abrasive, wherefrom more than half is smaller than half of the original diameter. Dube and Hutchings [6] studied especially the load influence on low- and high-stress abrasion and also investigated the fraction of broken particles (by their measurement method they were limited to the fraction smaller than half of the original diameter). The original particle shape (round vs. angular) showed a small effect, where round particles featured slightly higher breakage. For particles of 425-500 µm already at 20 N load ~ 25 % were broken, which increased to ~45 % at 60 N and ~75 % at 100 N. For low-stress conditions breakage was below 10 %. For such severe wear conditions optimal wear protection is inevitable to maintain sufficient service time of the components. Thereto often metal matrix composites (MMC) are applied, either as structural component or as local wear protection. While in single phase materials mostly the hardness dominates the abrasive wear resistance, in MMCs the amount, size, distribution, etc. of the hard phases also strongly influences the wear behaviour. Further the hardness of the hard phases is of major importance: zum Gahr [7] gives in Fig. 1a possible wear mechanisms for different types of hard phases. Hard abrasive particles with higher hardness than the matrix and hard phases easily cut or break the MMC. On the other hand soft abrasive particles (softer than the hard phases) cannot damage the entire MMC and protect the component sufficiently from abrasive attack. Too small hard phases are insufficient for wear protection, they have to be in the size range of the abrasive particles for good wear protection. When HT is present at the application, the wear regime may change, as all phases in the MMCs suffer from temperature induced softening, as well as the abrasives. Berns gives in [8] an overview of MMC behaviour. In Fig. 1b the hardness decrease of various matrix materials, hard phases and abrasives in dependence of temperature is summarised. Obviously temperature influences the various phases differently. Fe-based bcc matrices soften at >600°C rapidly, Ni-based matrices stay relatively constant up to 800°C. Typical precipitated M7C3 hard phases remain high hardness >1000 HV0.05 up to ~800°C. Flint’s hardness rapidly
2
drops and reaches the level of typical Fe- and Ni-based matrices at 500-600°C, while abrasives like SiC or Al2O3 stay at much higher level in the whole temperature regime investigated.
Figure 1: a) Possible wear mechanisms at abrasive wear of MMCs [7]; b) temperature dependence of micro-hardness of different abrasives, hardphases and matrices [9].
For soft materials or matrices of MMCs also wear reducing effects during abrasion are known. Mechanically mixed layers (MML) can be formed with abrasive particles during tribocontact [10-12]. This layers can offer a good wear protection, as they harden the surface by embedding of the hard abrasive particles. This was found to be especially beneficial at enhanced temperatures, as matrix softening favours the intermixing [11]. But the effect of wear protection is strongly dependent on the materials microstructure, application temperature [11,13,14] and applied load [4-6]. The aim of this work is to study systematically the influence of applied load and temperature on the abrasive wear behaviour. A Fe-based and Ni-based MMC were chosen to quantify the benefits of higher temperature stability of the Ni-based matrix. Differences in abrasive wear mechanisms should be investigated in detail, with special focus on in-situ formed MMLs. Temperature and load influence on changing wear regimes and MML extent should be investigated, in order to propose material concepts and economic solutions for HT abrasive conditions.
2. EXPERIMENTAL 2.1 Materials Two heat resistant metallic materials with hardphase reinforcement were investigated, due to their applicability in HT wear environment. A Fe-base and a Ni-base cast material were chosen to identify differences in the HT wear behaviour by different matrices. Both materials contain a significant amount of C to favour hardphase precipitation. The chemical compositions of the both alloys are given in Table 1. Samples were taken from cast rods in case of the Fe-based MMC and of a cast block for the Ni-based MMC. The microstructures of the two materials are given in Fig. 2 as seen by optical microscopy (OM). The Fe-based MMC shows dendritic structure with Cr-carbide precipitation around the grain boundaries. Carbides are <10 µm at a fraction of ~15 % as measured by QWIN® and of Cr7C3 type as found by XRD. The ferritic matrix in between shows a grain size of 30-40 µm. The resulting compound hardness at RT is ~275 HV10. The Ni-based MMC is given in Fig. 2b, also featuring dendritic structure: the austenitic matrix is surrounded 3
by a fine-structured Cr-carbide network, with occasional WxC precipitations. The grain size of the matrix zones is ~20-30 µm, with some larger areas in the range of 100 µm without carbides. Carbide content is ~13 -% and XRD-measurements identified them of Cr23C6 type. This alloy has similar hardness at RT as the Fe-based with ~245 HV10. Table 1: Chemical composition of the materials investigated. [wt.%] C Si Mn Cr W Fe Ni
Fe-based MMC 1.2-1.4 1.0-2.5 0.5-1.0 27-30 balance -
Ni-based MMC 0.35-0.55 1.0-2.0 <1.5 27-30 4.0-6.0 15 balance
Figure 2: Microstructures of the two alloys as seen by OM: a) Fe-based MMC, b) Ni-based MMC.
2.2 Hot hardness test The hardness is an important parameter in wear of materials. Especially the hardness at application temperature is of great interest, when dealing with HT abrasion. At the Austrian Competence Centre for Tribology (AC²T) a Hot Hardness Test (HHT) was developed [15] allowing for hardness testing up to 1000°C. The hardness test is done according Vickers method with 10 kg test load (HV10) and takes place in vacuum atmosphere (<5 mbar) to avoid oxidation of the surface. Test temperatures were chosen as RT, 100°C, 300°C, 500°C, 600°C, 700°C and 800°C with 3 indents each and the test series was repeated on a second sample for statistic evaluation. The indent diagonals were measured after cooling down of the samples by OM and the Vickers hardness were calculated. Failures due to thermal shrinkage were found to be insignificant for the materials investigated and are therefore not taken into account [15].
2.3 Abrasive wear test The HT-Continuous Abrasion Test (HT-CAT) was chosen, in order to test high-stress abrasion, as it occurs in high loaded components like crusher systems. The test is of block-on-ring configuration, based on an ASTM G65 [1] test rig, enhanced for HT testing (Fig. 3a). Thereto an inductive heating system enables the
4
heating of the sample up to ~700°C. Tests were done at RT, 500°C and 700°C. Temperature is controlled by a type K thermocouple, placed in a drilled hole near the tribocontact. The rubber wheel is changed to a Hardox® 400 wheel and has a circumferential speed of 1 m/s. The abrasive was chosen according the ASTM G65 standard: 212-300 µm rounded quartz as displayed in Fig. 3b. The flow rate was reduced to 180 g/min to avoid excessive cooling of the sample by the cold abrasive. To study the load influence, tests were done at 10 N, 45 N and 80 N. Test time/distance was adjusted to acquire wear scars of similar size, to keep contact conditions comparable between the tests (width: 12 mm, length: ~15 mm at the end of the test) [cf. 16]: 10 N-20 min/1200 m, 45 N-5 min/300 m, 80 N-2 min/120 m. No running-in phenomena were observed even at shortest test duration for this high-stress regime, i.e. steady-state conditions were present for all experiments. Starting from an initial line contact according Hertz’s contact theory, the contact pressure decreases with test duration and increasing conformity of the counter bodies. Nevertheless, in abrasive conditions where particle fracture takes place on a regular basis the contact pressure is limited by the crushing strength of the abrasive used [cf. 17]. A similar sized quartz abrasive used in [3] showed fracture at ~7 N load. The failure was catastrophic, i.e. the particle did not just break in two pieces, but shatters. Tests were repeated three times each condition. Wear rates are calculated by measuring the weight loss, calculating the volume loss via the material’s density and dividing by the wear distance. Test parameters are summarised in Table 2 and details on the HT-CAT can be found in [11]. Collection of the abrasive during test allows the evaluation of contact severity. The fraction of abrasive passing the wear track was evaluated similar to Antonov et al. [5]. Thereto on both sides of the wheel a sheet of paper was mounted with its edges at the height of contact. By these sheets of paper the flow of abrasive which did not pass the wear track was separated from the abrasive in contact, and both fractions were weighted afterwards. In our test configuration with a sand flow of 180 g/min (reduced from the G65 standard) a fraction of 37-40 % passes the wear track, while 60-63 % never come in contact with the sample. At one test of each configuration an abrasive sample was taken for 1 min at the end of the test duration and a sieve analysis undertaken. The fractions corresponding to the unbroken abrasive, (91.8 % >250 µm, 8 % 200-250 µm, 0.2 % 150-200 µm), were reduced by the percentage which did not pass the wear track. After sieve analysis the fraction of broken abrasive can be seen, giving a measure of the contact severity [4]. Further, one must keep in mind, that broken abrasive may have higher abrasivity due to the sharp edges compared to the rounded original abrasive [18.19]. Post-tribotest analyses were done on worn surfaces and cross-sections. Surface Scanning Electron Microscopy (SEM, Zeiss® Supra 55 VP) was utilised to calculate the covered surface by abrasive due to tribotesting. Thereto the surface is captured in Back Scattered Electron (BSE) mode, which shows abrasive in dark onto the bright metallic surface of the sample. By quantitative image analysis (QWIN ®) the percentage of the covered surface by abrasive is calculated. As the abrasive get intermixed with the matrix material to a certain limit the depth of maximal penetration of the abrasive is measured on cross-sections. Details of the depth measurement can be found in [11,13] and area measurement in [13].
5
Figure 3: a) HT high-stress abrasion test rig [11] and b) abrasive used for abrasion testing. Table 2: Main parameters used during HT high-stress abrasion testing. Parameter Temperature Normal load – sliding distance Sliding speed Counter body Abrasive Wear quantification
Value RT, 500°C, 700°C 10 N – 1200 m, 45 N – 600 m, 80 N – 120 m 1 m/s Hardox 400, 360 HV10, ø 232 × 12 mm Standard Ottawa silica sand, round, 212-300 µm at 180 g/min flow rate Volume loss / sliding distance → wear rate [mm³/m]
Further measurements were carried out on cross-sections of the two materials. Micro-hardness measurements were performed on different features of the microstructures to identify work hardening and supporting effect of the carbide network. Measurements were done with micro-hardness tester Future-Tech® FM-700 at RT. Nano-scratch experiments were carried out on the two materials to study the bonding and deformation mechanisms at carbide-matrix boundaries, aiming the simulation of abrasive attack by a single particle. The Hysitron® Triboindenter TI900 was utilised for this nano-scratching, which was also performed at RT. A Berkovich-tip was used for nano-scratching.
3. RESULTS 3.1 Hot hardness of the materials The hardness as important parameter in wear of materials was investigated up to 800°C by Vickers method. The results are displayed in Fig. 4a. Both alloys have relative low hardness compared to other wear protective materials for abrasive environment e.g. hypereutectic hardfacings [13]. The Fe-based MMC starts at ~275 HV10 at RT and drops linearly to ~180 HV10 at 500°C. >500°C the hardness loss is more pronounced due to excessive matrix softening, resulting in a minimum of 40 HV10 at 800°C. The Ni-based MMC starts at ~245 HV10 at RT and slowly decreases to ~130 HV10 at highest testing temperature of 800°C. The hardness progress is much shallower compared to the ferritic Fe-based MMC due to the temperature stable fcc-matrix of the Ni-based MMC. Especially the differences in hardness at temperatures >600°C will be interesting regarding HT wear behaviour. 6
Figure 4: a) Hot hardness and b) abrasive wear rates of the materials investigated.
3.2 Abrasive wear results High-stress abrasive wear rates measured at the different loads and temperatures are displayed in Fig. 4b. Wear rates were normalised by the distance, to be able to compare the different distances tested at the various loads. (Additional normalisation by the load was omitted, for better visualisation in the diagram.) Obviously the various loads entail different levels of wear rates. Temperature influence is less pronounced compared to the load dependency. Nevertheless, for the Fe-based MMC a significant increase of wear rate with ascending temperature can be measured for all load levels investigated. At 10 N load the increase for the Fe-based material is 25 % from RT to 500°C and further pronounced (~40 %) from 500°C to 700°C. On the other hand, the Ni-based alloy shows insignificant temperature influence at 10 N test load. Wear rates of the Fe-based MMC at 45 N load increase from 0.034 mm³/m at RT to ~0.039 mm³/m at 500°C and 0.054 mm³/m at 700°C, which stands for a increase of 37 % from RT to 700°. At 80 N the temperature influence also worsens wear loss by ~27 % from RT to 700°C. I.e. the rise of wear rate due to temperature influence decreases with increasing test load at the Fe-based MMC. The temperature dependence of the Ni-based MMC shows very interesting behaviour at higher loads. While at 10 N no temperature influence can be measured (within the accuracy of the applied test- and measurement methods), there manifests a minimum at 500°C for 45 N and 80 N test load. At 45 N, both, the RT and 700°C result are ~0.038 mm³/m and the 500°C wear rate is ~0.035 mm³/m. This effect gets more pronounced at 80 N, where the highest wear rate was measured at RT (0.103 mm³/m), followed by a minimum of 0.075 mm³/m (-37 %) and is increasing to 0.086 mm³/m at 700°C. Possible reasons for such unexpected behaviour will be discussed within the following sections.
3.3 Wear mechanisms at HT abrasion Abrasive wear rates are given in the previous section. Now the wear mechanism should be investigated in detail. As both materials are of similar structure, hard carbides in a soft matrix, wear mechanisms on surface SEM were found to be very similar, as displayed in Fig. 5-6. A ploughing wear mechanisms is observed at
7
the larger matrix zones, as clearly visible in Fig. 5a, with a lot of plastic deformation and abrasive embedding. Only at this lowest stress level a preferred direction of the abrasive, corresponding with the wheel movement, can be seen. Due to the free moving abrasive and random breakage of the abrasive, there is no distinct orientation of the wear grooves at higher stress levels. This is contrary to rubber-wheel testing, where temporary abrasive particles are fixed in the rubber and may groove the sample’s surface [cf. 3]. With increasing temperature and load the abrasive embedding becomes more dominant, covering the wear scars. Therefore cross-sections through the wear scars were performed for all conditions. The 45 N images are given in Fig. 7. For these materials the MML formation is clearly visible at all conditions. For the Fe-based MMC the increasing temperature clearly benefits MML formation and at 700°C a very thick and almost surface covering layer can be found. While at RT for the Ni-based MMC the layer formation is similar, at higher temperatures it is markedly less pronounced than at the Fe-based MMC. Quantitative results on the MML spread will be discussed in the following section 3.4.
Figure 5: Surface SEM-BSE after testing of Fe-based MMC: a) RT, 10 N; b) RT, 80 N; c) 700°C, 10 N; d) 700°C, 80 N.
8
Figure 6: Surface SEM-BSE after testing of Ni-based MMC: a) RT, 10 N; b) RT, 80 N; c) 700°C, 10 N; d) 700°C, 80 N.
Figure 7: Cross sections of wear zones after 45 N abrasion testing at different temperatures as seen by OM: a-c) Febased MMC; d-f) Ni-based MMC after a, d) RT; b, e) 500°C and c, f) 700°C testing.
Carbide regions impede the intermixing with the abrasive. At the Fe-based material this is obvious at RT and 500°C, while at the highest temperature of 700°C no effect of the carbide network is discernible. It seems that the material, or at least the matrix is soft enough, that massive plastic deformation can take place, 9
including displacement and intermixing of the carbides. This is different at the Ni-based material. Here the carbide network hinders a massive intermixing, which does not change even at highest testing temperature of 700°C. This means that the carbide network and matrix compound at the Ni-based material is much more temperature stable than at the Fe-based material. The load bearing ability of the carbide network will be discussed in section 3.5.
3.4 Abrasive interaction during tribocontact In order to study abrasive investigation two data sources were investigated: abrasive breakage during testing as well as post-test analysis of the covered surface by abrasive and abrasive penetration depth on cross sections. The severity of the contact can be assessed by evaluation of the broken abrasive during tribocontact. Thereto the abrasive during test was collected and afterwards a sieve analysis undertaken. After taking out of the fraction which did not pass the wear track [cf. 5] Fig. 8a can be drawn, showing the amount of broken abrasive at the various conditions. Increasing load clearly leads to a higher fraction of broken abrasive, i.e. the contact severity increases with the applied load and high-stress conditions get dominant. Interestingly the two alloys investigated show different temperature dependence. The Fe-based MMC shows a minimum of abrasive breakage at 500°C for high loads (45, 80 N), while the Ni-based material shows a maximum at this temperature for all loads investigated. For explanation the next diagram Fig. 8b and Table 3 can be helpful. There the abrasive coverage after the test, i.e. the amount of embedding during tribocontact, is displayed, as measured on surface SEM and evaluated by quantitative image analysis, and the depth of abrasive penetration measured on cross-sections. For the Fe-based MMC a clear load- and temperature influence was identified. Surface coverage increases with ascending temperature, as well as with increasing load, where the load influence is pronounced, leading to 52-58 % coverage at 80 N. At 10 N and 80 N the temperature influence is less significant, but at 45 N it is very pronounced: the surface coverage changes from ~38 % at RT to 50 % at 700°C. The surface coverage on the Ni-based MMC on the other hand shows just minor influence in the investigated borders. Nevertheless, temperature influence is obvious. At RT it is about 15 % for all loads, at 500°C 20 % were measured. The only differentiation between loads is possible at 700°C, where 23 % coverage were found at 10 N and ~28 % both at 45 and 80 N. Table 3: Abrasive penetration depth as measured by OM on cross-sections of the wear scars. Fe-based MMC
Ni-based MMC
Load
RT
500°C
700°C
RT
500°C
700°C
10 N
19 ± 4 µm
26 ± 4 µm
39 ± 9 µm
21 ± 5 µm
22 ± 7 µm
25 ± 7 µm
45 N
17 ± 4 µm
20 ± 6 µm
28 ± 6 µm
18 ± 2 µm
20 ± 2 µm
21 ± 8 µm
80 N
21 ± 9 µm
25 ± 9 µm
43 ± 14 µm
17 ± 4 µm
24 ± 5 µm
18 ± 5 µm
10
Figure 8: a) Abrasive breakage during high-stress abrasion testing; b) abrasive coverage as measured via surface SEM-BSE.
For the penetration depth on the other hand no clear trend can be observed for both materials. Load influence is insignificant, temperature influence is detectable at the Fe-based material. Generally it is reasonable, that softening of the Fe-based matrix with increasing temperature also entails increasing penetration depth, as previously reported in [4]. This softening effect is much less pronounced at the Ni-based material due to the fcc-structure and therefore more temperature stability in the investigated temperature range. Hence no significant temperature influence was found for the Ni-based MMC regarding penetration depth. The difficulties with this evaluation lie also in the 2-dimensional data source: only one cross-section throughout the wear scar can be analysed. Further the inhomogeneous microstructure of the MMCs and the carbide networks influence the MML formation, leading to the large deviations. Nevertheless, a clear difference between the Fe-based and the Ni-based MMC becomes obvious: the coverage and penetration depth is strongly depending on temperature and load for the Fe-based MMC, while the Ni-based MMC shows nearly no changes with the load and just minor changes of the MML formation with the temperature in the investigated range.
3.5 Detailed investigations on the micro-scale To study the fundamentals of the abrasive attack, detailed measurements on the micro-scale were conducted. By use of micro-hardness measurements comparing the matrix with the carbide zones the stability of the carbide network and matrix-backup was investigated. Further, the temperature influence on work hardening was measured by micro-hardness below the wear track. Nano-scratch experiments should study the fundamental micro-mechanisms of the abrasive attack. For the investigation of matrix-backup and carbide network stability micro-hardness measurements were done with loads from 98 mN (HV0.01) up to 9.81 N (HV1). Indents in matrix regions should point out the load effect on the Vickers-hardness, as with decreasing loads the elastic component of the indent becomes more significant, the hardness-value increases. Indents in zones of the carbide network on the other hand aim
11
on the identification of the load which the carbide network can carry: if the difference between matrix and carbide network hardness is minor, it is thought, that the carbides have insignificant effect on the abrasive attack by large particles. When the hardness of the carbide network indents becomes higher, it can carry the load and a beneficial effect on the wear resistance is expected. The results of these micro-hardness investigations are displayed in Fig. 9a. The matrix of the Fe-based MMC is slightly harder with ~250 HV than the Ni-based with ~200 HV. A significant influence of the test load on the matrix-hardness is not visible in the investigated load range. The Ni-based matrix shows ~200 HV hardness, which increases at the lowest test load to 230 HV0.01. At the Fe-based MMC the difference between carbide network and matrix is minor at 9.81 N and 2.94 N, but becomes distinct at loads <0.98 N. At the Ni-based MMC already at 9.81 N a pronounced difference between matrix and carbide network can be measured, which increases slightly at 2.94 N and stays constant until low loads. (The difference at 10 mN becomes less, as the matrix hardness appears higher because of the increasing elastic component.) These measurements lead to the conclusion that the Cr23C6 carbide network of the Ni-based MMC is much stronger against abrasive penetration as the Cr7C3 carbide network of the Fe-based MMC.
Figure 9: Micro-hardness measurements: a) Influence of load on matrix-backup of the carbide network; b) work hardening of the matrix during tribological interaction measured on cross-sections through the wear track.
Additionally to the wear reducing effect of the hard phases, work hardening can take place during tribocontact in the matrix regions of these materials. To study this behaviour micro-hardness measurements with 98 mN (HV0.01) were done on cross-sections. Starting at the worn surface hardness lines were done until ~200 µm depth below the wear track. Just the hardness of the matrix regions was measured and carbides avoided. This was done for both materials at 80 N abrasive wear tracks, as there the highest hardening effect is suspected. All three test temperatures were compared and results of representative hardness-curves are given in Fig. 9b. Work hardening at the Ni-based MMC is very pronounced at all temperatures investigated, while at the Fe-based MMC it is almost disappearing at HT. The Ni-based matrix features ~230 HV0.01 as shown in Fig. 9a, which increases to >500 HV0.01 in the first 10 µm under the 12
wear track. The progress is very smooth: the hardness decreases slowly. The original matrix hardness was found in depths of 150-200 µm, depending on the carbide structure. It has to be noted, that carbide lines do not stop the hardening effect in surface near regions, and it could be also measured below carbide clusters. At the Fe-based MMC hardening was distinct in the first <10 µm, and its effect decreases with increasing temperature. At RT hardening from 250 HV0.01 bulk matrix hardness to ~450 HV0.01 was measured, at HT ~350-370 HV0.01 were measured near the worn surface. At RT and 500°C a hardening effect could be measured until 150-200 µm, while at 700°C the original matrix hardness is reached after 40-50 µm. The effect of abrasive attack with even smaller loads as possible by single abrasive particles was investigated by nano-scratch experiments with 10 mN down to 1.5 mN. At attack by these small loads the phases do not act like a “compound” any more, and the properties of the single phases and the interface becomes important [cf. 20]. A comparison of the matrices at 1.5 mN is given in Fig. 10a. A micro-ploughing mechanism dominates the matrix zones, where the material is plastically deformed from the groove to the ridges. Wear scars at the Fe-matrix are less pronounced than at the Ni-matrix, which can be put down to the lower hardness and better deformability of the Ni-matrix. In Fig. 10b a length profile through the scratch is given, where the scratch starts on the left with penetration of the matrix, reaching a carbide at 0-coordinate, passing the carbide and entering the matrix again at the end of the scratch. The damage of the carbide is in the range of 3-8 µm depth, but differences are very small. Cr23C6 are slightly harder than Cr7C3, what might be the reason for the different scratch depth [9].
Figure 10: Profiles of the nano-scratches at 1.5 mN load: a) cross-section at the matrix, b) length-section of the entire scratch.
13
Figure 11: SEM-BSE detail of the 10 mN scratch in the Fe-based MMC: matrix-carbide-matrix boundary.
For good imaging by SEM the scratches at 1.5 mN were too small, thereto additional scratches at 5 and 10 mN were performed. Exemplarily the scratch at the phase interface from matrix to carbide to matrix of the Fe-based MMC is shown in Fig. 11. The micro-ploughing wear mechanisms is clearly visible on the left, with the bright ridges. At the boundary to the carbide the ridges are ploughed over the hard carbide. One can suspect, that material is easily removed in a further abrasive attack. At the carbide no plastic deformation is visible. Here a cutting wear mechanism must dominate, with also some micro-cracking at the end of the carbide. The interface in the wake of the scratch also shows some crack formation, which is not visible on the surging border of the carbide.
4. DISCUSSION High-stress abrasive wear rates of the here investigated materials are on a relative low level, compared with other high sophisticated hardfacings like e.g. hypereutectic FeCrC alloys. Varga et al. compares in [4,9,11,13,14] different high temperature materials regarding their abrasion resistance, also within the test field is a high sophisticated complex-alloyed hardfacing with a high content of hard phases. Wear rates of this hardfacing at enhanced temperatures were much higher than at the Fe-based MMC also studied within this work. The in-situ MML formation in [4,11] offers better wear resistance than a high hardphase content, were brittle wear behaviour was dominant. Aim of this work is to investigate, if even better wear resistance can be reached by Ni-based materials with austenitic matrix and similar hardphase amount. While wear rates at RT are higher for the Ni-based MMC at all loads investigated, this changes at elevated temperatures. At 500°C the Ni-based material is slightly better than the Fe-based MMC, at highest temperature of 700°C a pronounced difference can be measured. The Ni-based MMC is superior to all materials investigated previously [4,11,13]. As investigated by micro-hardness experiments the load carrying capacity of the Cr23C6 carbide network in the Ni-based alloy is much better than the Cr7C3 network of the Fe-based alloy. This may be a decisive factor, especially at enhanced temperatures, when matrix softening exacerbates carbide backup. As found by the nano-scratch experiments the single carbides are an effective measure against abrasive wear loss on the micro-scale. They are well bonded to the matrix and feature very little wear loss, which may be decreased to zero when dealing with softer abrasives than the carbide [7]. The matrix of both alloys becomes
14
easily plastically deformed, which is slightly more pronounced at the Ni-based material. Carbides won’t undergo a significant loss in hardness in the investigated temperature range [8], contrary to the matrices, where especially the Fe-based matrix shows a distinct loss in hardness >500°C as indicated by the macro hot hardness measurements. Despite their hard phase network both materials form MMLs, which were found to reduce wear loss e.g. [10,21,22]. Nevertheless, the MML extension is much smaller at the Ni-based material, but entails superior wear resistance. This leads to the assumption, that thick MMLs do not necessarily entail best wear resistance, but also a thinner MML may lead to the wear reducing effect. Load influence on abrasive coverage was not significant for most conditions at the Ni-based alloy, and increasing coverage is mainly driven by temperature rise. For the Fe-based MMC also the load has significant influence on the coverage. Penetration depth on the other hand was found to be just dependent on temperature and material softening, respectively. An interesting behaviour of the wear rates with temperature increase shows the Ni-based MMC, especially at high loads. The wear rate at RT is significantly higher than at 500°C and increasing again at 700°C. Similar behaviour at HT abrasion was also reported by Berns (ed.) [8]: he found a decreasing wear rate until a critical temperature (dependent on the matrix structure) due to MML formation. Exceeding this temperature, wear rates increase again, because of dynamic recrystallisation and loss of work-hardening ability. The critical temperature is in the range of 0.6× melting temperature, further certain conditions of deformation have to be present according [8]. One must keep in mind, that this regime is limited to very surface near regions in the wear zone.
Figure 12: Deformation of Ni-based MMC after 80 N testing as seen on cross-sections after: a) RT, b) 500°C and c) 700°C testing.
To investigate the deformation mechanisms in more detail the cross-sections of the wear tracks of the Nibased MMC were etched with Ni grain boundary etchant (25 ml distilled water, 50 ml hydrochloric acid 32 %, 15 g iron-III chloride, 3 g ammonium tetrachlorocuprate(II)). Details of the cross-sections in surfacenear regions are given in Fig. 12. At RT a massive formation of dislocations can be found in the first micrometres beneath the wear track. These dislocations are also visible between the carbides, i.e. deformation is not stopped by the carbide network. As found by the micro-hardness lines the work-hardening reaches down to 150-200 µm. At 500°C the etchant reveals a clearly different image. Almost no dislocation lines can be found, but small recrystallized grains, probably by dynamic recrystallisation [8,23,24]. This may also goes conform to a hardening effect in the very surface near zones by grain refinement, entailing the low wear rates measured. At 700°C the recrystallisation is much more pronounced, reaching deep beneath the
15
tribocontact. Recrystallisation starts at ~0.6× melting point, which is ~760°C at the Ni-based and ~810°C at Fe-based material. This temperature may be reached at the Ni-based material within the tribocontact due to additional frictional heat. The combined effect of dynamic recrystallisation and MML formation at the Nibased MMC obviously benefits the wear resistance at 500°C. Future EBSD measurements may substantiate this hypothesis. At higher temperatures the softening of the material and easier plastic deformation seems to entail increasing wear loss again. Nevertheless the HT high-stress abrasive wear rates >500°C of the Nibased material is superior to any other material reported so far, and the application and cost effectiveness of such MMCs should be studied in the future.
5. CONCLUSIONS High temperature abrasive wear at high-stress condition was studied on two metal matrix composites (MMCs) with ~15 % hard phases. A Ni-based matrix was compared with a Fe-based matrix, in order to study the effect of the matrix’ lattice and its influence on wear resistance at increasing load and temperature in abrasive contact. Following conclusions can be drawn:
The test load has significant influence on the wear rate and severity of the contact. A larger fraction of the abrasive is broken at higher loads.
At the Fe-based MMC wear rates increase from RT to 700°C, with a more pronounced ascend from 500°C to 700°C which goes conform to a more distinct softening of the material >500°C.
The Ni-based MMC shows slightly higher wear rates at RT than the Fe-based MMC. At 500°C the Ni-based material is slighly better and at 700°C the difference is pronounced. The higher hot hardness of the Ni-based material is one reason for its superity. Also the Cr23C6 carbide network of the Nibased material can carry higher loads then the Cr7Cr3 network of the Fe-based MMC.
Mechanically mixed layer (MML) formation was found for both alloys. At the Fe-based MMC it is increasing with load and temperature. At the Ni-based alloy a temperature influence could be observed, but much less pronounced than at the Fe-based material.
Pronounced strain hardening was found at the Ni-based material at all temperatures investigated. Further dynamic recrystallisation is probable at higher temperatures leading to decreasing wear rates at 500°C.
Fundamental abrasive wear mechanisms were studied by nano-scratch experiments. Micro-ploughing dominates in matrix regions, while micro-cutting was found at the carbides, accompanied by microcracking at the edges of carbides.
As the Ni-based material shows much better wear resistance at high temperatures, its application in industrial plants and cost effectiveness will be investigated. In the future detailed studies of the wear resistance of different carbide structures, e.g. deposited by welding, may further enhance the wear resistance.
6. ACKNOWLEDGEMENTS This work was funded by the Austrian COMET Program (Project K2 XTribology, no. 849109) and carried out at the “Excellence Centre of Tribology”. Special thanks goes to Harald Rojacz for intense discussions, 16
Anifa MF. Azhaarudeen for evaluating experiments and Christian Tomastik for performing nano-scratch experiments.
7. REFERENCES 1. ASTM G65-04: Standard test method for measuring abrasion using the dry sand/rubber wheel apparatus, American Society for Testing and Materials (2010) 2. RI. Trezona, DN. Allsopp, IM. Hutchings: Transition between two-body and three-body abrasive wear: influence of test conditions in the microscale abrasive wear test, Wear 225-229 (1999) 205214 3. SM. Nahvi, PH. Shipway, DG. McCartney: Effects of particle crushing in abrasion testing of steels with ash frombiomass-fired powerplants, Wear 267 (2009) 34-42 4. M. Varga, AMF. Azhaarudeen, K. Adam, E. Badisch: Influence of load and temperature on abrasion of carbidic cast steel and complex alloyed hardfacing, Key Engineering Materials 674 (2016) 313318 5. M. Antonov, I. Hussainova, R. Veinthal, J. Pirso: Effect of temperature and load on three-body abrasion of cermets and steel, Tribology International 46 (2012) 261-268 6. NB. Dube, IM. Hutchings: Influence of particle fracture in high-stress and low-stress abrasive wear of steel, Wear 233-235(1999) 246-256 7. KH. zum Gahr: Microstructure and Wear of Materials, Elsevier Science Publishers, Tribology Series 10 (1987) 8. H. Berns (ed.): Hartlegierungen und Hartverbundwerkstoffe, Springer Berlin, Heidelberg (1998) 9. M. Varga: High temperature abrasion in sinter plants and their cost efficient wear protection, Dissertation Montanuniversität Leoben (2016) 10. LC. Jones, RJ. Llewellyn: Sliding abrasion resistance assessment of metallic materials for elevated temperature mineral processing conditions, Wear 267 (2009) 2010-2017 11. M. Varga, H. Rojacz, H. Winkelmann, H. Mayer, E. Badisch: Wear reducing effects and temperature dependence of tribolayer formation in harsh environment, Tribology International 65 (2013) 190-199 12. M. Varga, H. Winkelmann, E. Badisch: Impact of microstructure on high temperature wear resistance, Procedia Engineering 10 (2011) 1291-1296 13. M. Varga, AMF. Azhaarudeen, E. Badisch: Influence of in-situ formed tribolayer on abrasive wear reduction, Materials Science Forum 825-826 (2015) 85-92 14. M. Varga, High temperature abrasive wear of metallic materials, accepted for publication in Wear (2017) 15. M. Varga, M. Flasch, E. Badisch: Introduction of a novel tribometer especially designed for scratch, adhesion and hardness investigation up to 1000°C, Proceedings of the Institution of Mechanical Engineers Part J: Journal of Engineering Tribology 231-4 (2017) 469-478 16. IM. Hutchings: Abrasive an erosive wear tests for thin coatings: a unified approach, Tribology International 31/1-3 (1998) 5-15
17
17. H. Rojacz, H. Pahr, S. Baumgartner, M. Varga: High temperature abrasion resistance of differently welded structural steels, Tribology International, in press (2017) 18. DA. Kelly, IM. Hutchings: A new method for measurement of particle abrasivity, Wear 250 (2001) 76-80 19. S. Wirojanupatump, PH. Shipway: Abrasion of mild steel in wet and dry conditions with the rubber and steel wheel abrasion apparatus, Wear 239 (2000) 91-101 20. M. Varga, S. Leroch, H. Rojacz, M. Rodríguez Ripoll: Study of wear mechanisms at high temperature scratch testing, accepted for publication in Wear (2017) 21. H. Berns: Microstructural properties of wear-resistant alloys, Wear 181-183 (1995) 271-279. 22. A. Fischer: Mechanisms of high temperature sliding abrasion of metallic materials, Wear 152 (1992) 151-159. 23. U. Cihak-Bayr, G. Mozdzen, E. Badisch, A. Merstallinger, H. Winkelmann: High plastically deformed sub-surface tribozones in sliding experiments, Wear 309 (2014) 11-20. 24. H. Rojacz, G. Mozdzen, F. Weigel, M. Varga, Microstructural changes and strain hardening effects in abrasive contacts at different relative velocities and temperatures, Mater. Charact. 118 (2016) 370381.
8. FIGURES Figure 1: a) Possible wear mechanisms at abrasive wear of MMCs [7]; b) temperature dependence of micro-hardness of different abrasives, hardphases and matrices [9]. Figure 2: Microstructures of the two alloys as seen by OM: a) Fe-based MMC, b) Ni-based MMC. Figure 3: a) HT high-stress abrasion test rig [11] and b) abrasive used for abrasion testing. Figure 4: a) Hot hardness and b) abrasive wear rates of the materials investigated. Figure 5: Surface SEM-BSE after testing of Fe-based MMC: a) RT, 10 N; b) RT, 80 N; c) 700°C, 10 N; d) 700°C, 80 N. Figure 6: Surface SEM-BSE after testing of Ni-based MMC: a) RT, 10 N; b) RT, 80 N; c) 700°C, 10 N; d) 700°C, 80 N. Figure 7: Cross sections of wear zones after 45 N abrasion testing at different temperatures as seen by OM: a-c) Fe-based MMC; d-f) Ni-based MMC after a, d) RT; b, e) 500°C and c, f) 700°C testing. Figure 8: a) Abrasive breakage during high-stress abrasion testing; b) abrasive coverage as measured via surface SEM-BSE. Figure 9: Micro-hardness measurements: a) Influence of load on matrix-backup of the carbide network; b) work hardening of the matrix during tribological interaction measured on cross-sections through the wear track. Figure 10: Profiles of the nano-scratches at 1.5 mN load: a) cross-section at the matrix, b) length-section of the entire scratch. Figure 11: SEM-BSE detail of the 10 mN scratch in the Fe-based MMC: matrix-carbide-matrix boundary.
18
Figure 12: Deformation of Ni-based MMC after 80 N testing as seen on cross-sections after: a) RT, b) 500°C and c) 700°C testing.
9. TABLES Table 1: Chemical composition of the materials investigated. [wt.%] C Si Mn Cr W Fe Ni
Fe-based MMC 1.2-1.4 1.0-2.5 0.5-1.0 27-30 balance -
Ni-based MMC 0.35-0.55 1.0-2.0 <1.5 27-30 4.0-6.0 15 balance
Table 2: Main parameters used during HT high-stress abrasion testing. Parameter Temperature Normal load – sliding distance Sliding speed Counter body Abrasive Wear quantification
Value RT, 500°C, 700°C 10 N – 1200 m, 45 N – 600 m, 80 N – 120 m 1 m/s Hardox 400, 360 HV10, ø 232 × 12 mm Standard Ottawa silica sand, round, 212-300 µm at 180 g/min flow rate Volume loss / sliding distance → wear rate [mm³/m]
Table 3: Abrasive penetration depth as measured by OM on cross-sections of the wear scars. Fe-based MMC
Ni-based MMC
Load
RT
500°C
700°C
RT
500°C
700°C
10 N
19 ± 4 µm
26 ± 4 µm
39 ± 9 µm
21 ± 5 µm
22 ± 7 µm
25 ± 7 µm
45 N
17 ± 4 µm
20 ± 6 µm
28 ± 6 µm
18 ± 2 µm
20 ± 2 µm
21 ± 8 µm
80 N
21 ± 9 µm
25 ± 9 µm
43 ± 14 µm
17 ± 4 µm
24 ± 5 µm
18 ± 5 µm
19
Highlights
The Ni-based MMC shows better wear resistance at 700°C than the Fe-based MMC
The Cr23C6 carbide network of the Ni-based material is more resistant than the Cr7Cr3 network of the Fe-based MMC.
Mechanically mixed layer formation was found for both alloys.
Pronounced strain hardening was found at the Ni-based material at all temperatures investigated.
Micro-ploughing dominates in matrix regions, while micro-cutting was found at the carbides.
20