A Study on the E ﬀ ect of Ultraﬁne SiC Additions on Corrosion and Wear Performance of Alumina-Silicon Carbide Composite Material Produced by SPS Sintering

: Alumina-silicon carbide (Al 2 O 3 –SiC) composites of varying compositions (15, 20, 25 and 30 vol.%)–SiC were produced by the ball milling of Al 2 O 3 and SiC powders, followed by spark plasma sintering. The samples were sintered at a temperature and pressure of 1600 ◦ C and 50 MPa, respectively, thermally etched at 1400 ◦ C and mechanically fractured by hammer impact. The e ﬀ ect of SiC additions to monolithic Al 2 O 3 on the densiﬁcation response, microstructural and phase evolutions, and fracture morphologies were evaluated. The wear performance of the composites using a ball-on-sample conﬁguration was evaluated and compared to that of monolithic Al 2 O 3 . In addition, the corrosion performance of the composites in a 3.5% NaCl solution was examined using open circuit potential and potentiodynamic polarization assessments. SiC additions to monolithic Al 2 O 3 delayed densiﬁcation due to the powder agglomeration resulting from the powder processing. SiC particles were observed to be located inside Al 2 O 3 grains and some at grain boundaries. Intergranular and transgranular fracture modes were observed on the fractured composite surfaces. The study has shown that the Al 2 O 3 –SiC composite is a promising material for improved wear resistance with SiC content increments higher than 15 vol.%. Moreover, the increase in SiC content displayed no improvement in corrosion performance.


Introduction
Currently, there has been increasing attention for 'ceramic nanocomposites' [1]. These consist of a ceramic matrix such as alumina (Al 2 O 3 ) reinforced by dispersing nanoparticles of another ceramic material like silicon carbide (SiC). Besides SiC, mechanical properties can be improved by introducing other secondary hardened phases to the Al 2 O 3 matrix, including TiB 2 , TiC, Ti(C,N) and ZrO 2 particles. A study conducted by Kim and Lee showed that adding TiC particles to the Al 2 O 3 matrix improved flexural strength and fracture toughness [2]. Properties such as flexural strength, fracture toughness, and hardness were later reported to be improved with an increment in TiB 2 [3]. Furthermore, the resistance to oxidation of the Al 2 O 3 matrix can be improved by adding TiN particles [4]. Literature reported that SiC addition to Al 2 O 3 as reinforcement enhances the mechanical properties significantly as related to other reinforcing additives [5][6][7][8][9].
The addition of ultrafine SiC particles to monolithic Al 2 O 3 was reported to significantly improve the hardness [1,7,8], fracture strength [10,11], and toughness [12]. Creep resistance at elevated temperatures was also reported to be enhanced significantly by SiC additions [13]. Ceramics can be manufactured using traditional methods such as sintering, hot or cold pressing, and slip casting.

Powder Production
The Al 2 O 3 -SiC nanocomposites were produced using γ-Al 2 O 3 (Sigma-Aldrich Pty Ltd., Johannesburg, South Africa) and β-SiC (Industrial Analytical Pty Ltd., Johannesburg, South Africa) powders with mean particle sizes of 0.1 and 44 µm, respectively. Oxygen contamination of the powders was prevented by handling the powders in an argon gas-sealed glove box. Al 2 O 3 powder was milled using 2 mm alumina balls in a Reeves ball mill Szegvari attritor system type B ® for 8 h using hexane as a dispersant. The same was done for the SiC powder using distilled water as a dispersant for 16.5 h. Four mixtures (15,20,25, and 30 vol.% SiC) were milled for 8 h inside a Fritsch Pulverisette 6 planetary mono mill ® using hexane as a dispersant and oleic acid ((Monitoring & Control Laboratories Pty Ltd., Johannesburg, South Africa) as a process control agent. The slurries were dried using a digital HeidolphLaborota 4010 rotary evaporator (Heidolph Instruments GmbH & CO. KG, Schwabach, Germany), sieved, and the dispersant burnt-off inside an Elite TSH17/75/150 ® tube furnace (Elite Thermal Systems Ltd., Leicestershire, UK) to aid consolidation during SPS.

Powder Characterisation
A Malvern Mastersizer 2000 ® (Malvern Panalytical Ltd., Worcestershire, UK) was used to confirm the particle size distribution of the powders: as-received and after milling. A Carl Zeiss Sigma ® Field Emission Scanning Electron Microscope (FE-SEM) (Carl Zeiss Microscopy GmbH, Jena, Germany) was used to carry out particle morphology studies. Shimadzu X-ray Diffraction (XRD) machine (Shimadzu, Kyoto, Japan) was used to characterize the phases present in the powders. The average crystallite size was calculated by means of the Williamson-Hall (W-H) method (Equation (1)). Plots were drawn with 4 sin θ on the x-axis and β hkl cos θ along the y-axis for all the starting powders. By fitting the data, the crystallite size D was extracted from the slope of the best fit line.

Spark Plasma Sintering (SPS)
The prepared blends were sintered using an FCT Systeme GmbH HP D 5/2 ® Spark Plasma Sintering furnace (FCT Systeme GmbH, Effelder-Rauenstein, Germany) in a graphite mold using a graphite foil and hexagonal boron nitride (hBN) as a coating. A heating rate of 200 • C/min was applied to heat up the samples to the necessary sintering temperature. The samples were sintered at 1600 • C using a dwell time of 10 min in vacuum under a uniaxial pressure of 50 MPa. After sintering, the ram pressure was released, and the specimens cooled at 100 • C/min. After that, the samples were sandblasted and wiped with isopropanol (Monitoring & Control Laboratories Pty Ltd., Johannesburg, South Africa) to remove excess sand and graphite foil.

Sintered Samples Characterisation
Archimedes' method was used to measure the densities of the sintered samples using distilled water as a medium. The sintered samples were metallographically prepared down to a 1 µm diamond finish and etched thermally at 1400 • C. Microstructure and phases present in the sintered samples were analyzed using Energy-Dispersive X-ray Spectroscopy (EDS) (Oxford x-act, Abingdon, UK) and SEM. The average grain size of the sintered composites was obtained by employing the linear intercept method.
The wear behavior of the composites produced was tested using an Anton Paar GmbH Standard Tribometer Version 7.3.13 (Anton Paar GmbH, Graz, Germany). A ball-on-flat geometry was employed with a 6 mm diameter alumina static partner. A load of 10 N was used for 1814 s and the coefficient of friction with time for monolithic Al 2 O 3 and each composite was recorded. The wear scars were examined using a JEOL JSM-IT500 SEM (JEOL Ltd., Akishima, Japan).
A Digi-Ivy DY2300 potentiostat (Digi-Ivy, Inc., Austin, TX, USA) coupled to a standard three electrochemical electrode cell set-up was employed to study the corrosion behavior of the samples in a 3.5 wt.% NaCl (Alfa Aesar by Thermo Fisher Scientific GmbH, Kandel, Germany) solution at 25 • C. The electrochemical cell used consisted of a platinum counter electrode, silver/silver chloride reference electrode, and the working electrode as the sample. The samples were prepared for testing by connecting the sample face to an insulating wire with aluminum tape. Open-circuit corrosion potential measurements were carried out for 60 min, while potentiodynamic polarization measurements were performed using a scan rate of 5 mV/s at a potential initiated at −400 to −950 mV. A fresh electrolyte solution was used with each sample.  Figure 1 gives the particle size distribution of the powders. The statistical parameters for the size distribution of powders are given in Table 1. The particle size distribution of the Al 2 O 3 and SiC powder as-received was unimodal, with 50% of the particles above 6.37 and 48.0 µm in size and 10% above~2.17 and 26.75 µm in size, respectively. The milling was sufficient for both the Al 2 O 3 and SiC powders, with about 90% of the particles below 7.71 and 0.25 µm, respectively.

Powder Characteristics
Metals 2020, 10, x FOR PEER REVIEW 4 of 13 Figure 1 gives the particle size distribution of the powders. The statistical parameters for the size distribution of powders are given in Table 1. The particle size distribution of the Al2O3 and SiC powder as-received was unimodal, with 50% of the particles above 6.37 and 48.0 µm in size and 10% above ~2.17 and 26.75 µm in size, respectively. The milling was sufficient for both the Al2O3 and SiC powders, with about 90% of the particles below 7.71 and 0.25 µm, respectively.  The morphology of the powder after mixing is shown in Figure 2. The blended powder seemed to be made of small particles evenly dispersed amongst much bigger particles. In Figure 2b, the same powder at high magnification reveals that bigger particles are agglomerates of fine particles, with voids between the agglomerated particles. Figure 3 shows the XRD diffractogram for blended and alloyed powders. It shows a shift of the Al2O3 peaks indicative of the alteration of the Al2O3 matrix due to the presence of SiC. Furthermore, peak broadening confirms the presence of ultrafine powders observed also in Table 2.  The morphology of the powder after mixing is shown in Figure 2. The blended powder seemed to be made of small particles evenly dispersed amongst much bigger particles. In Figure 2b, the same powder at high magnification reveals that bigger particles are agglomerates of fine particles, with voids between the agglomerated particles. Figure 3 shows the XRD diffractogram for blended and alloyed powders. It shows a shift of the Al 2 O 3 peaks indicative of the alteration of the Al 2 O 3 matrix due to the presence of SiC. Furthermore, peak broadening confirms the presence of ultrafine powders observed also in Table 2.

Composite Response to SPS
The shrinkage behavior of the Al2O3-SiC composites during the SPS process is shown in Figure  4. The displacement represents the shrinkage profile directly during the densification of the powders in real time. The pure Al2O3 started to densify at about 850 and finished at about 1300 °C. However, the temperatures for the onset of densification were raised by the addition of SiC. The Al2O3-15 vol.% SiC composite and Al2O3-30 vol.% SiC composites started to densify at about 975 and 1550 °C,

Composite Response to SPS
The shrinkage behavior of the Al2O3-SiC composites during the SPS process is shown in Figure  4. The displacement represents the shrinkage profile directly during the densification of the powders in real time. The pure Al2O3 started to densify at about 850 and finished at about 1300 °C. However, the temperatures for the onset of densification were raised by the addition of SiC. The Al2O3-15 vol.% SiC composite and Al2O3-30 vol.% SiC composites started to densify at about 975 and 1550 °C,

Composite Response to SPS
The shrinkage behavior of the Al 2 O 3 -SiC composites during the SPS process is shown in Figure 4.  [21]. The added SiC hindered the densification process.
According to Hsueh and colleagues [35], SiC particles act as tensile stress concentrations influenced by the different characteristics of the composite phases present (Al 2 O 3 and SiC).
Metals 2020, 10, x FOR PEER REVIEW 6 of 13 respectively. Densification for these samples finishes at about 1530 and 1580 °C, respectively, similar to that reported by Chae and colleagues [21]. The added SiC hindered the densification process. According to Hsueh and colleagues [35], SiC particles act as tensile stress concentrations influenced by the different characteristics of the composite phases present (Al2O3 and SiC).  Table 3 shows the relative densities obtained. The samples were almost fully densified with all relative densities above 97% using the SPS conditions stated above. SPS can produce almost fully dense Al2O3-SiC composites at temperatures of ~1600°C at a soaking time of 10 min compared to 1650 °C of hot pressing and an hour of soaking [33,34,36]. The microstructures of the pure Al2O3 and of the Al2O3-SiC composites are shown in Figure 5a,c,e-g. The relations between the Al2O3 grain size and the content of SiC are summarized in Table 3. The addition of 15 and 20 vol.% significantly decreased the grain size of the composites. Further SiC additions to monolithic Al2O3 increased the grain size. The microstructure of the monolithic Al2O3 (Figure 5a) was relatively coarser than the microstructure of composites with 15 and 20 vol.% in Figure 5f,g. The micrographs showed that the SiC particles were mainly located inside the grains and some at the Al2O3 grain boundaries. Figure 5b shows the SiC particles located both intergranularly and transgranularly. Related research described that SiC particles (~200 nm) were located within the Al2O3 grains, while larger SiC particles tended to be at the Al2O3 boundaries [37]. SEM micrographs of the fractured surfaces of pure Al2O3 and the Al2O3-SiC composites showed features resembling a transgranular fracture mode.   Table 3 shows the relative densities obtained. The samples were almost fully densified with all relative densities above 97% using the SPS conditions stated above. SPS can produce almost fully dense Al 2 O 3 -SiC composites at temperatures of~1600 • C at a soaking time of 10 min compared to 1650 • C of hot pressing and an hour of soaking [33,34,36]. The microstructures of the pure Al 2 O 3 and of the Al 2 O 3 -SiC composites are shown in Figure 5a,c,e-g. The relations between the Al 2 O 3 grain size and the content of SiC are summarized in Table 3. The addition of 15 and 20 vol.% significantly decreased the grain size of the composites. Further SiC additions to monolithic Al 2 O 3 increased the grain size. The microstructure of the monolithic Al 2 O 3 (Figure 5a) was relatively coarser than the microstructure of composites with 15 and 20 vol.% in Figure 5f,g. The micrographs showed that the SiC particles were mainly located inside the grains and some at the Al 2 O 3 grain boundaries. Figure 5b shows the SiC particles located both intergranularly and transgranularly. Related research described that SiC particles (~200 nm) were located within the Al 2 O 3 grains, while larger SiC particles tended to be at the Al 2 O 3 boundaries [37]. SEM micrographs of the fractured surfaces of pure Al 2 O 3 and the Al 2 O 3 -SiC composites showed features resembling a transgranular fracture mode.     of friction as compared to pure Al 2 O 3 . In contrast, higher vol.% additions (20, 25, and 30) decreased the mean coefficient of friction values substantially. The previously stated phenomenon is made even more apparent when considering the distance covered to achieve the same coefficient of friction. For example, the same coefficient of friction was achieved after 22 m for monolithic Al 2 O 3 , while it took 47 m for the Al 2 O 3 -15SiC composite. In addition, as seen in Figure 7 and Table 4, the wear track size seems to decrease with the addition of 15 vol.% of SiC. This could be attributed to SiC pull-out, which has been previously reported in other investigations [38][39][40].

Characteristics of the Sintered Composites
the mean coefficient of friction values substantially. The previously stated phenomenon is made even more apparent when considering the distance covered to achieve the same coefficient of friction. For example, the same coefficient of friction was achieved after 22 m for monolithic Al2O3, while it took 47 m for the Al2O3-15SiC composite. In addition, as seen in Figure 7 and Table 4, the wear track size seems to decrease with the addition of 15 vol.% of SiC. This could be attributed to SiC pull-out, which has been previously reported in other investigations [38][39][40].
Further SiC additions of up to 30 vol.%, increased the track sizes to ~1242 µm, also resulting in a substantial reduction in wear coefficient compared to the composite with 15 vol.% additions and the monolith. Finally, the wear surfaces of the composites presented the presence of localized flaking and microcracking (Figure 8), similar to findings made by Guicciardi et al. [41]. The flaking and microcracking behavior were severe for Al2O3-15SiC composite and less severe with higher SiC additions of up to 30 vol.%. The brighter regions were previously reported [41][42][43] as Al2O3 and the darker regions could be a hydroxide phase.   Further SiC additions of up to 30 vol.%, increased the track sizes to~1242 µm, also resulting in a substantial reduction in wear coefficient compared to the composite with 15 vol.% additions and the monolith. Finally, the wear surfaces of the composites presented the presence of localized flaking and microcracking (Figure 8), similar to findings made by Guicciardi et al. [41]. The flaking and microcracking behavior were severe for Al 2 O 3 -15SiC composite and less severe with higher SiC additions of up to 30 vol.%. The brighter regions were previously reported [41][42][43] as Al 2 O 3 and the darker regions could be a hydroxide phase.   The Tafel plots for the composites in 3.5% NaCl solution ( Figure 9) were utilized in carefully studying the corrosion behavior of the composites. The composites of vol.% 15, 25 and 30 generally displayed similar polarization curves and passivity characteristics. The Al2O3-20SiC composite displayed slightly different behavior from the other three variants. Table 5 lists the corrosion potential (Ecorr) and current (Icorr) of the composites. The corrosion potentials (Ecorr) of the composites were distinct and defined in the ranges of −0.731 to −0.499 V. The addition of 15 or 20vol.% of SiC increased the Ecorr, very similar to the observations made by Maahn and Roepstorff [44]. A slight change in Icorr was observed with additions of 15 and 20 vol.% of SiC (0.16 difference in magnitude). Furthermore, the composites containing 25 and 30 vol.% displayed the same Icorr. Conclusively, there was no observed improvement in the corrosion performance of the composites. Table 5. Corrosion potential (Ecorr) and current (Icorr) of the composites in 3.5% NaCl.

Sample
Ecorr (V) Icorr (A/cm 2 ) Al2O3-15SiC −0.731 2.433 × 10 −8 Al2O3-20SiC −0.499 3.841 × 10 −9 Al2O3-25SiC −0.556 9.577 × 10 −9 Al2O3-30SiC −0.694 9.577 × 10 −9 Figure 9. Tafel plots for Al2O3-SiC composites. The Tafel plots for the composites in 3.5% NaCl solution ( Figure 9) were utilized in carefully studying the corrosion behavior of the composites. The composites of vol.% 15, 25 and 30 generally displayed similar polarization curves and passivity characteristics. The Al 2 O 3 -20SiC composite displayed slightly different behavior from the other three variants. Table 5 lists the corrosion potential (E corr ) and current (I corr ) of the composites. The corrosion potentials (E corr ) of the composites were distinct and defined in the ranges of −0.731 to −0.499 V. The addition of 15 or 20 vol.% of SiC increased the E corr , very similar to the observations made by Maahn and Roepstorff [44]. A slight change in I corr was observed with additions of 15 and 20 vol.% of SiC (0.16 difference in magnitude). Furthermore, the composites containing 25 and 30 vol.% displayed the same I corr . Conclusively, there was no observed improvement in the corrosion performance of the composites. The Tafel plots for the composites in 3.5% NaCl solution ( Figure 9) were utilized in carefully studying the corrosion behavior of the composites. The composites of vol.% 15, 25 and 30 generally displayed similar polarization curves and passivity characteristics. The Al2O3-20SiC composite displayed slightly different behavior from the other three variants. Table 5 lists the corrosion potential (Ecorr) and current (Icorr) of the composites. The corrosion potentials (Ecorr) of the composites were distinct and defined in the ranges of −0.731 to −0.499 V. The addition of 15 or 20vol.% of SiC increased the Ecorr, very similar to the observations made by Maahn and Roepstorff [44]. A slight change in Icorr was observed with additions of 15 and 20 vol.% of SiC (0.16 difference in magnitude). Furthermore, the composites containing 25 and 30 vol.% displayed the same Icorr. Conclusively, there was no observed improvement in the corrosion performance of the composites. Table 5. Corrosion potential (Ecorr) and current (Icorr) of the composites in 3.5% NaCl.

Conclusions
In this work, a characterization study was conducted on pure Al 2 O 3 and Al 2 O 3 -SiC powders obtained by milling, as well as the composites consolidated by SPS. The conclusions are:

1.
Ultrafine particles were produced after milling, evident from the peak broadening observed during XRD analyses. The morphology of the mixed powders revealed Al 2 O 3 matrix with small SiC particles evenly dispersed amongst much larger and agglomerated particles. A shift of the Al 2 O 3 peaks during XRD showed that the presence of SiC particles altered the Al 2 O 3 matrix after mechanical alloying.

2.
The addition of SiC increased temperature from the onset of densification. Furthermore, the addition of SiC delayed densification. In addition, the degree of agglomeration observed in the initial milling stage of the powders might have prolonged densification. 3.
The resistance to wear decreased with the addition of 15 vol.% of SiC. Additions of higher vol.% (20, 25, and 30) of SiC significantly improved the resistance to wear of Al 2 O 3 .

4.
The study's composites showed no improvement in corrosion performance with increments in SiC additions.