The Effect of the Chemical Composition to the End-Properties of Ceramic Dispersed Strengthened 316L/Y2O3 Composites

In this paper the influence of chemical composition to the end-properties of ceramic dispersed strengthened 316L/Y2O3 composites ceramic has been studied. Two various compositions were studied and compared to reference 316L sintered sample. These two compositions are 316L/0.33 wt% Y2O3 and 316L/1 wt% Y2O3. The high-efficient attrition milling has been used for grain size reduction and oxide distribution in the austenitic matrices. Spark Plasma Sintering (SPS) was used as fast compaction method of the milled powders in order to avoid excessive grain growth. In this work it was found that changing the chemical composition by increase of the Y2O3 addition in the composite matrix improves the milling efficiency, increases the hardness of the 316L and reduces significantly the wear rate.


Introduction
Spark Plasma Sintering (SPS) has been successfully used for the densification of a wide variety of materials (ceramic [1,2], metals and alloys [3,4], polymer and composites [5,6]). Balázsi et al. [7] used as first SPS to realize nanostructured steel compacts. Many researchers studied the effect of yttria addition on the structural and mechanical properties of the steel where it was found that ultrafine yttria particles can improve significantly the hardness of steel, also its strength at high temperatures [8].
Composites with small particles exhibit higher strength because of their higher joining zone which is an effective heat source [9]. The interaction between the yttria and the steel matrix increases the diffusion of this last which increases the densification of steel, however, 12wt% or higher content of yttria will not improve the densification due to the agglomeration of yttria [10]. Oxides with Y and Cr are influencing the porosity having a direct effect on the oxidation rate in steel composites [11]. A composite with high content of yttria nano particles results in a delay in the 316L sintering, the yttria nano particles agglomerate together when sintered at lower temperatures [12]. The oxides can enhance the upper temperature limit in mechanical creep strength with 100 C° at least [13]. Lindau et al. [14] studied the effect of 0.3 % and 0.5 % yttria on the mechanical properties of the EUROFER 97 at high temperatures where it was found that the yttria improved all of the yield strength, ductility and the creep strength. Baek et al. [15] studied the hydrogen susceptibility of nanosized Y 2 O 3 dispersed strengthened 316L austenitic steel, they found that hydrogen didn't affect the yield strength and the elastic modulus of the 316L. Hutař et al. [16] studied the small fatigue crack propagation in Y 2 O 3 strengthened steels and they found that yttria content has no significant effect to the crack propagation. Kumar et al. [17] studied the effect of Y 2 O 3 and ZrO 2 on the microstructure and mechanical properties of nano-ODS (oxide dispersion strengthened steel) where it was found that the hot pressed Y 2 O 3 containing ODS steels show higher ultimate tensile strength (UTS) comparing to the ZrO 2 containing ODS and it shows even higher UTS values when compared with ZrO 2 . The yttria content under 0.6wt% improves the oxidation resistance by forming a stable oxide scale and improving the adhesion, while higher amounts of yttria results in segregation of this last, The yttria addition tends to decrease the bending strength and the hardness of the steel [18]. In this work the chemical composition effect on the structural, mechanical and tribological properties of the 316L composite was studied.

Experimental methods
The austenitic 316L stainless steel powder from Höganäs with 16.8Cr-12Ni-2.5Mo-1.5Mn-0.6Si elemental composition and with average grain size ~70 μm and the Y 2 O 3 powder grade C from H.C. Starck with ~700 nm in average particle size have been used to prepare two composites: 316L/0.33 wt% Y 2 O 3 and 316L/1 wt% Y 2 O 3 . The Union Process attritor mill type 01-HD/HDDM has been used in order to reduce the steel grains size and insure a homogeneous distribution of the ceramic particles in the steel matrix. The composite powders have been elaborated by milling in ethanol for 5 hours at 600 rpm using 3 mm stainless steel balls with ball/material weight ratio of 10:1. The composites have been sintered under 50 MPa mechanical pressure at 900 °C for 5 minutes dwelling time in vacuum using the Sinter-SPS-7.40MK-VII apparatus in Istanbul Technical University. The sintered samples were solid disks with 100mm in diameter and ~9 mm thickness. The composites density has been measured using Archimedes method. The tribological properties of sintered samples have been investigated using the setup type CSM-HT-Tribometer. Different grinding papers (grade up to 100 μm) have been used for polishing the samples before measuring the tribological properties. 5 N normal load was applied with the 5 mm diameter Si 3 N 4 ball against the steel sample surface with 1mm shift from the rotation axe of the sample. The samples have been tested for 2161 m at room temperature in air at 47 % atmosphere humidity for the 0.33 wt% Y 2 O 3 and 53 % for the 1wt% Y 2 O 3 . The wear track was investigated by Keyence Microscope for wear volume calculations. The structure and morphology of the milled powders have been investigated using the scanning electron microscopy (SEM, Zeiss-SMT LEO 1540 XB and Jeol JSM-25-SIII) equipped with EDS. The X-ray diffractometer Bruker AXS D8 with CuKα radiation have been used to analyze the phases in the two composites. The Philips CM20 Transmission Electron Microscope (TEM) operated at 200 kV acceleration voltage has been used to investigate the composites microstructures, discs with 3 mm diameter and 20 μm thickness in the centre have been prepared by mechanical thinning, polishing, and dimpling. Technoorg Linda ionmill has been used to reach electron transparency thickness, 10 keV Ar+ ions at an incidence angle of 5° with respect to the surface. In the final period of the milling process, the ion energy was decreased gradually to 3 keV to minimize ion-induced structural changes in the surface layers. The TEM images were taken in plan view. The hardness of the sintered composites has been measured by Vickers method where 5 N was applied for 10 seconds.

Morphological and Structural investigation
The investigation of the atomised 316L stainless steel starting powder by SEM (Fig. 1A) shows that it is consisted of ~70 μm globular shape grains with the presence of satellites on their surface. The Y 2 O 3 powder (Fig. 1B) is consisting of flake-like shaped grains with ~700 nm particle size in average. The globular shape grains of 316L starting powder has been transformed into 4 types/shapes (as in Fig. 2, 1-slightly damaged, 2-flattened, 3-flake-like shape and 4-small broken grains) due to the high impact forces implemented by the steel balls during the 5 hours milling. The impact of steel balls is higher in the bottom part of the milling jar comparing the top part, that is why we have different grains morphologies.
The small grains ( Fig. 2 spot 4) are broken flakelike shape steel grains, we assume that more 1-2 hours in addition of attrition milling would make a significant grain size reduction. The ratio of flake-like shape grains to the flattened and slightly damaged grains is bigger in the case of the Y 2 O 3 additions ( Fig. 3A and Fig. 4A) comparing to milled 316L powder (Fig. 2).
The presence of the Y 2 O 3 in the case of the 316L/0.33wt% Y 2 O 3 was not clearly detectable, therefore, we investigated the surfaces of grains at high magnifications (Fig. 3B), the EDS spectras (Fig. 3C) shows the presence of yttria in all of the selected spots. The Aluminum presence is a contamination trace in the starting 316L powder (as received from Höganäs company). In the case of the 316L/1 wt% Y 2 O 3 , the EDS spectra clearly show the yttria presence in the milled composite powder (Fig. 4B). The yttria particles were embedded into the surface of the 316L steel grains making it harder, this increased the impact of milling balls which in turn increased the evolution of flake-like grains.
The milling efficiency was slightly decreased in the case of the 316L/1 wt% Y 2 O 3 composite comparing to the 316L/0.33 wt% Y 2 O 3 because the yttria particles agglomerated together which influenced (decreased) the milling efficiency.
After the sintering process, the ferritic α-Fe phase has been transformed to the austenitic ɣ-Fe 3 Ni 2 phase as it was expected. The Y 2 O 3 peaks disappeared after sintering. In the case of the reference 316L sample we noticed a reorientation in the direction [200] of the grains after sintering.
The EDS spectra of the sintered 316L/0.33 wt% Y 2 O 3 (Fig. 6) show the presence of small amount of Y 2 O 3 in the composite. This finding is indicating some kind of diffusion or agglomeration of the yttria particles in the steel matrix, resulting in the decrease of the Y peak intensity in the sintered composite comparing the milled powders.
The Y peak was not detectable by EDS in the case of the sintered 316L/1 wt% Y 2 O 3 (Fig. 7B). The Y peak is  not observed even with its higher content in this composite as a result of the non-homogeneous distribution of the yttria in the steel matrix. Presumably the Y peak can not be observed on EDS because the agglomeration of yttria particles during the sintering process (Fig. 7B).
The Fig. 8 and Fig. 9 show the TEM images of the sintered composites. In the case of the 316L/0.33 wt% Y 2 O 3 (Fig. 8) we observed that the material is composed of relatively small steel grains of ~ 5-20μm with a good and homogeneous distribution of the Y 2 O 3 in the grain boundaries, unlike the 316L/1 wt% Y 2 O 3 (Fig. 9) where the yttria particles are agglomerated together. Micrometer    After tribology investigation both Si 3 N 4 ball and the samples surfaces have been damaged. The damage on the ball was insignificant. The wear track width and depth have been measured for wear volume calculation. The wear rate of the 316L reference, 316L/ 0.33 wt% Y 2 O 3 and the 316L/ 1 wt% Y 2 O 3 have been calculated ( Table 1).
The wear rate has been reduced significantly with the addition of yttria to the matrix. The investigation of the damaged surfaces shows the formation of tribo films on the samples surfaces ( Fig. 10A and Fig. 11A). No yttrium content was detected by EDS on the wear track of the two composites ( Fig. 10B and Fig. 11B). High intensity peaks of oxygen and silicon were measured in the same spot of the tribo-film in both composites, which can be explained by the formation of silicon oxide during the tribology test.
Silicon oxide formation is a result of temperature increase in the contact zone as a result of a relatively high sliding speed. The lower intensity peaks of the 316L components in zone 2 (Fig 10B and 11B) is due the coverage of 316L composite by the tribo-film. The friction coefficients are represented in Fig. 12.
The 316L/Y 2 O 3 composites are showing higher density comparing with the 316L/Si 3 N 4 composite elaborated using the same parameters [19]. The 316L/Y 2 O 3 composites are harder than the 316L reference sample as it is represented in Fig 12. The HV values: 1.75±0.05, 2.63 ± 0.32 and 2.33 ± 0.165 GPa have been measured for the 316L (reference), 316L/0.33wt% Y 2 O 3 and the 316L/1 wt% Y 2 O 3 respectively. The 316L/ 0.33wt% Y 2 O 3 composite was showing somewhat higher hardness values than the 316L/1 wt% Si 3 N 4 composite from our previous work [20]. The friction coefficient dropped with addition

Discussions 4.1 Morphological properties
The milling efficiency has been improved with the addition of the yttria as it observed clearly in Fig. 3A and Fig. 4A where the ratio of the flake-like shape grains to the slightly damaged and flattened grains is higher comparing the milled reference 316L powder (Fig. 2). This improvement in the milling efficiency is attributed to the yttria addition which increased impact effect by hardening surface of the steel grains. The EDS spectra of the milled powder with high intensity yttrium peak (Fig. 4B) and the milled powder X-ray diffractograms (Fig. 5) show the presence of Y 2 O 3 in the 316L/1 wt% Y 2 O 3 composite. In the case of sintered samples no yttria peaks in the XRD and EDS may be observed ( Fig. 5 and Fig. 7B respectively). This finding proves that the agglomeration of the yttria particles in the 316L/1 wt% Y 2 O 3 composite ( Fig. 9) took place during the sintering process.

Structural properties
The presence of the ferrite α-Fe phase in the milled composites might be a result of an austenitic-martensitic/ferritic transformation during the milling process due to the sever deformation under the high impact of the milling balls or might be related to contamination from the milling setup. The ferrite phase was transformed to the ɣ-Fe 3 Ni 2 as it is shown in Fig. 5. Two main lines of yttria has been clearly identified in the case of the 316L/1 wt% Y 2 O 3 (Fig. 5) unlike in the case of the 316L/ 0.33 wt% Y 2 O 3 where the yttria content was under the detection limit. The distribution of the yttria particles in the 316L/1 wt% Y 2 O 3 composite is not homogeneous as we can observe in Fig. 9.
The agglomeration of yttria shown in TEM results is in correlation with the XRD results (Fig. 5) and it explains the disappearance of the yttria peak in diffractogram of the 316L/1 wt% Y 2 O 3 sintered composite.

Mechanical properties
The presence of yttria slightly dropped the density, however, the 316L/Y 2 O 3 composites are showing higher densities comparing to similar composites made by the 3-Dimensional Fiber Deposition (3DFD) technique even when higher sintering temperatures were applied [12]. The prepared 316L/ Y 2 O 3 composites are showing higher hardness values comparing to the 316L/0.4 wt% Y 2 O 3 prepared by electron beam selective melting (EBSM) and Spark Plasma Sintering even  after hot rolling [21]. The lower HV value of the 316L/1 wt% Y 2 O 3 comparing to the 316L/0.33 wt% Y 2 O 3 (Fig. 12) is due to the non-homogeneous distribution of the yttria particles (Fig. 9). The lower friction coefficient in the case of the 316L/Y 2 O 3 composites comparing to the 316L reference sample is due to their higher hardness.

Conclusions
The elaboration of yttria dispersed strengthened steel using attrition milling and Spark Plasma Sintering has been demonstrated. The effect of changing chemical composition by addition of yttria to the 316L matrix on the structural and mechanical properties has been studied.
The addition of yttria improved the milling efficiency. Agglomeration of the yttria particles took place during the sintering process in the case of the 316L/1wt% Y 2 O 3 .
The 316L hardness and tribological properties have been improved with the addition of yttria. The 316L/0.33wt% Y 2 O 3 show better mechanical properties comparing to the 316L/1wt% Y 2 O 3 and the 316L reference samples.