Self-Lubricating Composites Enriched in the Surface with Molybdenum and Nickel

In powder metallurgy the addition of alloying elements is a widely used alternative to increase the service life of mechanical components. However, they have an associated cost. One less exploited process is surface enrichment (or surface alloying), whose same result is obtained on the surface, but at a much lower cost. Self-lubricating composites are especially complex to be enriched, precisely because they have to avoid the deposition of alloying elements over the lubricant stocks. This paper reports the characterization and tribological properties of self-lubricating composites enriched with molybdenum or nickel. Sintered samples were produced for comparative purposes. Wear resistance, friction coefficient and wear rate were analyzed for the evaluation of tribological properties. The results showed that the enrichment process did not cover the lubricant stocks. In addition, the Mo-enriched samples presented the best tribological results, with a 250 % increase in scuffing resistance in relation the sintered sample, and lower wear rate.


Introduction
A way of improving mechanical properties of a selflubricating component produced by powder metallurgy (PM) is by adding alloying elements, which increase the properties of the matrix and at the same time maintain the self-lubricating properties of the component 1 . The presence of the alloying elements increase the mechanical strength of the matrix due to the microstructural modification, such as the formation of different phases or precipitates 2,3 .
However, this performance improvement has a high cost since the addition of alloying elements occurs throughout the component while, in most engineering applications, an increase strength is necessary only in the volume near the surface. This occurs because in the mechanical components the wear of a few micrometers is enough for them to lose their function and dimensional tolerance and to be discarded to avoid the failure of the mechanical system in which they are installed 4,5 .
Plasma surface enrichment is an attractive process to increase the mechanical and tribological properties of self-lubricating components. In this process the addition of alloying elements occurs only in the region where it is needed, that is, on the surface and just below it 6 . In addition, as the process is carried out simultaneously the sintering step 7,8,9 , there is not the increase in processing time and still cost reduction relative to the raw material used.
Recently, some authors 10,11,12,13 evaluated the tribological properties of plasma surface alloyed materials. The results showed that the surface alloying process generally reduces the coefficient of friction (COF) and the wear rate of the materials. However, studies about the enrichment of selflubricating components are still a gap. This is because enrich the surface of a self-lubricating composite without cover the lubricant stocks is a challenge, since the enrichment process occurs over the entire surface. Considering this, our research group has investigated the enrichment of selflubricating composites avoiding cover the lubricant stocks on the surface 14 .
In this investigation, the enrichment of a self-lubricating composite with molybdenum and nickel was carried out. The modifications obtained by the enrichment and the tribological and mechanical properties of the obtained self-lubricating composite are shown and discussed.

Materials and Methods
In this paper, samples of self-lubricating composites were submitted to the plasma surface enrichment process with molybdenum or nickel. The composite samples were produced by powder injection molding (PIM) using an Arburg Allrounder model 320s. Sintered iron-based composites were produced by mixing carbonyl iron powder CL-OM (d 50 of 3.9 -5.2 µm and 0.6 wt.% carbon) and silicon carbide (SiC) powder with a particle size of 15 μm. The SiC powder was used as precursor to produce the graphite nodules (solid lubricant) due the dissociation of SiC previously described in 15 17 . Therefore, in these samples a large part of the solid lubricant is generated during the sintering stage. The samples were produced with a chemical composition of Fe + 0.6C + 3SiC. For injection, the feedstock was prepared in a Winkworth mixer (180 °C, 70 rev min −1 , 90 min) using 8 wt.% of an organic binder system containing polypropylene, paraffin wax, ethylene vinyl acetate copolymer (EVA), stearic acid (surfactant) and amide wax. After the extraction of the binders, described in 18 , theses specimens were sintered and enriched (with Mo or Ni) simultaneously in a hybrid plasma reactor described in other works 7,19,20 . The parameters used in the plasma treatments and samples nomenclature are presented in Table 1. Reference samples, just sintered (S1 sample) was also produced for comparatives purposes.
The samples were plasma sintered and enriched onto the grounded electrode (confined anode-cathode configuration). Three holders with different compositions were designed and assembled for placing the samples in the reactor. To perform the Mo-enrichment, a holder was made of TZM alloy (approximately 99.4 wt.% Mo) and to perform the Ni-enrichment, a holder was made of Ni200 alloy (approximately 99.7 wt.% Ni). In order to set up basis for comparison, an additional set of samples (S1) was sintered on an AISI 1020 steel holder.
After plasma treatments, the top and cross section of the samples were analyzed by scanning electron microscopy (SEM) (TESCAN Vega 3), energy dispersive X-ray (EDX) (Oxford x-act) and by microhardness measurements. The microhardness measurements were performed using a Vickers indenter with a 0.01 kg load for 15 seconds in a Future-Tech FM-800 microdurometer. The microhardness tests were performed according to MPIF Standard 51 and 52 21,22 .
To determine which phases were present after plasma treatments, X-ray diffraction (XRD) was performed using a Philips X'Pert diffractometer. The diffractograms were recorded from the specimen surfaces to identify the phases from the positions of the diffraction peaks, data from the JCPDS were used The tribological characterization was evaluated using two experimental routes with a CETR UMT tribometer. First using reciprocating sliding tests with regular increments of 7 N each 10 minutes described in 23 to determine the scuffing resistance. In this paper, the scuffing resistance was defined as the work (N.m) until the COF exceeds the value of 0.2 and stays above this limit for at least 190 seconds 17 , this is important to avoid a premature interruption of the test due fluctuations in COF that may occur during increments of normal load. The second route was also carried out with reciprocating sliding tests, but using a constant normal load of 7 N for 1 hour to obtain the friction coefficient and wear rate of the samples and counter-bodies. For all experimental routes the counter-body was hard AISI 52100 steel ball with 5 mm diameter, the oscillation frequency was 2 Hz and the wear track length was 10 mm. The tests were replicated at least 3 times for each experimental route.
White light interferometry (ZygoNewView 7300) was used to quantify the specimen wear volumes. Topographical data processing was performed using the MountainsMap Universal 7.1 ® software. The wear volume (ΔV) of the counter-body was estimated using equations 1, 2, and 3 for the spherical cap volume, as illustrated in Figure 1. The value of "a" was obtained using optical microscopy of the wear scar. Where "ΔV" is the wear volume, "r" is the radius of the sphere; "a" is the radius of the cap base. "h" is the height of the cap, and "d" is the difference between the radius of the sphere and the radius of the cap base.

Results and Discussions
Considering that for a self-lubricating component to maintain the lubricating properties, it is necessary that the solid lubricant is also available on the surface after the plasma enrichment process. To evaluate the surface microstructure of the samples after the enrichment processes, SEM and EDX investigation was performed. The results are shown in Figure 2. It is possible observe three different regions for the Mo2 sample (as indicated in Figure 2(a)), which have different Mo concentrations. The region 1 is representative of the lubricant stocks. The enrichment with nickel was homogeneous, reaching up to 11.2 wt.% on the Ni2.5 sample (Figure 2(b)). In this sample, some lubricant stocks regions was enriched reaching up to 2.2 wt.% of nickel. This behavior was not observed for the Mo2 sample. It is assumed that the non-significant coverage of the solid lubricant by nickel or molybdenum is due to the process of generation of lubricant reservoirs used in this work. In this process, the solid lubricant reservoirs (stocks) are generated in situ during the sintering by dissociation of SiC precursor 15,16 . The graphite nodules were produced at the same time to the enrichment processes, which is performed simultaneously with the sintering. This allowed the formation of new graphite nodules during all enrichment, which resulted in a non-significant coverage of the lubricant stocks at the end of the process. For comparative purposes, sample S1 shows the surface of a specimen without enrichment (Figure 2(c)). No significant differences were observed in the morphology of the solid lubricant between the samples.
The X-ray diffractograms recorded from the surface of samples are shown in Figure 3. The diffractogram of the S1 sample presented the ferrite and graphite phases. For the Mo2 sample, the martensite phase was also present. The presence of martensite is justified by the high concentration of molybdenum in some regions, (until 7.8 wt. % showed at Figure 2(a)) which increased the hardenability of these material 24,25,26 .
Images obtained by SEM of the cross-sectioned samples are presented in Figure 4. The micrograph of sintered sample (Figure 4(a)) corroborates the results obtained by XRD, where the graphite peak corresponds to the graphite nodules generated during the dissociation of SiC 15,16 . Ferrite and perlite were also observed, which is in accordance with the microhardness profile obtained for this sample shown in Figure 5.
In the micrograph of Mo2 sample it is possible to observe the Mo-enriched layer (region between the arrows) ( Figure  4(b)). This layer has a thickness of 15.82 ± 4.02 μm and an almost constant molybdenum profile, between 1.4 and 1.0 wt.% in the first 10 μm as observed in the EDX analysis presented in Figure 6. According to micrograph and the microhardness profile ( Figure 5), this layer is composed predominantly of bainite. Despite this, some regions confirms the presence of martensite-austenite islands 31 (indicated by circles in the Figure 4 (b)) previously demonstrated by XRD and reported by 32 in an Astaloy Mo + 0.6% C alloy. Figure 4 (c) shows the SEM micrograph of the Ni2.5 sample. The enriched layer is not clear, but it was possible to observe the presence of two regions. The first region (surface region), present a thickness of 3.65 ± 0.71 μm and according to the chemical analysis (Figure 6), has a higher nickel content (between 17.8 and 9.9 wt.%). In this region, the surface microhardness measurements showed a value of 441.9 HV + 72.0 HV, that is in accordance with the taenite microhardness reported in 28,29,30 . Therefore, it is considered that this region is formed by the taenite phase. Below the taenite phase is the second region, this region is poorer in nickel (between 0.8 -8.0 wt.%) and has a lower microhardness (~200 -250HV), as shown in the microhardness profile ( Figure 5).  The diffractogram of the Ni-enriched sample indicated ferrite, graphite, kamacite and taenite phases. The kamacite is a solid solution of nickel-poor iron, with CCC structure and microhardness around 181 HV 27,28 . On the other hand, the taenite is a solid solution of nickel-rich iron, with CFC structure, and microhardness depending on the nickel content, ranging from 350 -500 HV 27,28,29,30 .  This result is indicative that the area below the taenite phase can be considered composed by kamacite phase.
The effect of enrichment treatments (Mo or Ni) were evaluated in relation to the tribological behaviour and to confirm whether the self-lubricating characteristic of the material was maintained after the plasma enrichment treatments. Figure 7 shows the wear rates of samples S1, Mo2, Ni2.5 and the respective counter-bodies. The Mo2 sample had the lowest wear rate. The wear rate value determined was 11.37 x 10 -6 mm 3 N -1 m -1 , in opposite the wear rate of the reference composite (S1 sample) was 17.35 x 10 -6 mm 3 N -1 m -1 . This correspond to a 34.5% reduction. The Ni2.5 sample had a wear rate of 14.32 x 10 -6 mm 3 N -1 m -1 ± 3.62 x 10 -6 mm 3 N -1 m -1 that is statistically equal to that of the sintered sample (S1) of 17.35 x 10 -6 mm 3 N -1 m -1 ± 1.85 x 10 -6 mm 3 N -1 m -1 . This indicates that Ni enrichment did not improve the wear rate  of the samples. Figure 8 presents the effect of the enrichment treatments on the friction coefficient (lubricious regime) and the scuffing resistance. The Mo2 sample presented an increase around of 250 % on scuffing resistance in relation to S1 and Ni2.5 samples. These results, were attributed to the increased mechanical strength of the matrix due to the molybdenum enrichment, that changed the metallurgical microstructure from ferrite/perlite to bainite/martensite. This increase in mechanical strength was corroborated by microhardness analyses that showed a hardening depth of 45 μm to Mo2 sample.
Although having a surface microhardness similar to the Mo2 sample, the Ni2.5 sample showed similar scuffing resistance (459,95 Nm) and wear rate (14,32 x 10 -6 mm 3 N -1 m -1 ) as the sample without enrichment (463,04 Nm and 17,35 x 10 -6 mm 3 N -1 m -1 ) (S1 sample). This can be explained by the lower hardness of the phase formed just below the surface (kamacite), which has a lower mechanical strength. The wear rate of the counter-bodies was higher for the samples Mo2 and Ni2.5, this behavior is due to the higher values of surface microhardness of these samples in relation to the reference sample.
The S1 and Mo2 samples presented similar COF values (0.17 x 0.18) indicating that the surface enrichment with molybdenum did not interfere with the lubricity effect. The Ni2.5 sample showed a COF that exceeded the lubricity limit (0.21) and a high standard deviation ( Figure 8). However, a statistical analysis of the COF values showed that there was no significant difference between the COF values in the three conditions studied.

Conclusions
This study evaluated the influence of surface enrichment, with molybdenum or nickel, on tribological properties of selflubricating composites (Fe + 0.6C + 3SiC). The solid lubricant used here were graphite nodules produced by the dissociation of SiC. The following conclusions can be established: a) The surface enrichment treatments does not significantly cover the lubricant stocks.
b) The Mo2 sample presents the best tribological results, with an increase of 250 % of the scuffing resistance, and reduction of 34.5% of the wear rate in relation to the sintered sample (S1).
The obtaining of solid lubricant stocks without significant coverage by the alloying elements, allows advances in the area of self-lubricating composites. This study showed the possibility of producing self-lubricating composites with improved tribological properties, without the conventional use of alloying elements in the form of dispersed powders in the matrix. This allows cost reduction in the production of these components, and avoids potential segregation problems during powder mixing. In addition, the surface enrichment can be realized several times with the same cathode being an environmentally friendly process.