Effect of yttria addition on the microstructure and mechanical behavior of ODS ferritic alloys processed by High Energy Milling and Spark Plasma Sintering

2.1. High energy milling

For this investigation, the following reagent-grade commercial powders were acquired from Sigma Aldrich company (particle size and purity, in wt.%, are indicated inside the parentheses): Fe (83 µm, 99%), Cr (45 µm, 99%), W (12 µm, 99.9%), Ti (5-10 µm, 99.9%), Y₂O₃ (5.3 µm, 99.9%), and C₁₈H₃₆O₂. The latter material (stearic acid) was used as a lubricant or process control agent (PCA) during milling, setting 0.2 g per batch. To analyze the effect of the amount of added yttria on both the sintering process and the resulting properties of the ODS ferritic alloys, three compositions were prepared (see Table 2).

Selected compositions were thoroughly milled using a commercial HEM mill (Simoloyer, Zoz GmbH CM01), operated in an air atmosphere with yttria-stabilized zirconia balls of 10 mm in diameter, and with a balls/powder weight ratio of 40:1. The milling cycles were carried out for 15, 30, 45 and 60 min at 900 rpm.

2.2. Spark Plasma Sintering

As for the sintering runs, a commercial SPS device (Sumitomoto Dr. Sinter, SPS-1050) was used, operating at high vacuum conditions under on/off current pulses of 12/2, as well as 3500 Amp, 4 V, 60 Hz, and 6 kN axial load. Using the latter technique, 4.5 g of each of the milled powder series was densified, using graphite cylindrical dies with an inner and external diameter of 10 and 50 mm, respectively. The sintering runs were conducted at a heating rate of 200 ^oC·min^-1 up to 1100 ^oC, with no holding time at the peak temperature, obtaining sintered coin-like specimens 10 mm in diameter and 7 mm high. The total sintering time was 5.5 min per sample.

2.3. Characterization of milled powders

To study the effect of milling time on the particle size, a particle size distribution (PSD) analysis was performed using a Horiba apparatus, model LA950V2, employing ethanol as dispersing medium. To monitor any changes in the crystalline structure of the powders, X-ray diffraction analysis (XRD, Bruker model D8 Advance Echo, CuKα radiation, λ = 1.5406 Å, 2θ = 20-100°, 0.6 mm slit, and step size of 0.02° with a duration of 0.3 s per step) was carried out. The crystallite size and lattice parameter of the α-Fe phase were determined for the 0 ODS powder as well as for the sintered alloys containing 0, 0.4, and 0.8 wt.% Y₂O₃ from the corresponding XRD patterns by Rietveld refinement, using the Panalytical’s HighScore Plus software, and the size-strain analysis routine. Details of the crystal structure of the phases were obtained from the COD and ICDD databases.

2.4. Characterization of Sintered Samples

A SEM (JEOL JSM-7600-F, field emission) microscope was used to analyze the morphological evolution of the milled powder, as well as the microstructure of the sintered samples. The sintered specimens were prepared for SEM analysis following standard metallographic techniques. The relative density of the sintered specimens was measured by the Archimedes’ principle in distilled water. Their Vickers microhardness was determined using a Shimadzu HMV-G 20DT apparatus, under a load of 100 g for 10 sec (HV_0.1). To determine the phase transformations occurring in the studied materials, samples of 9x4x3 mm were prepared by cutting the sintered specimens with diamond discs, and then analyzing using a quenching dilatometer (Linseis Model L78 Rapid Induction Thermal Analysis). The analysis conditions consisted of heating and cooling rates of 10 °C·s^-1, a temperature range of 25-1100 °C, no holding time at the peak temperature, helium atmosphere, and chromel-alumel thermocouples for temperature measurements.

Corrosion tests were carried out using a 1 M HNO₃ solution, a 3.8 M Ag/AgCl reference electrode (199 mV vs. Standard Hydrogen Electrode, SHE), and a graphite bar as the counter electrode. The failure analysis of the material due to its susceptibility to corrosion was carried out using the “Tafel technique and polarization curves” (Buchanan and Stansbury, 2012Buchanan, R.A., Stansbury, E.E. (2012). 4 - Electrochemical Corrosion. In Handbook of Environmental Degradation of Materials. Second Edition, William Andrew Publishing, pp. 87-125. https://doi.org/10.1016/B978-1-4377-3455-3.00004-3.). The reason for using HNO₃ to conduct the corrosion tests of the studied alloys is that it is an aggressive oxidant that is typically used in the separation of spent fuel from nuclear fission reactors in the “Purex process” (Ramanujam, 2001Ramanujam, A. (2001). Purex and Thorex Processes (Aqueous Reprocessing). In Encyclopedia of Materials: Science and Technology. Second Edition, Buschow, K.H.J., Cahn, R.W., Flemings, M.C., Ilschner, B., Kramer, E.J., Mahajan, S., Veyssière, T. (Ed.), Elsevier, pp. 7918-7924. https://doi.org/10.1016/B0-08-043152-6/01426-1.). Before the electrochemical analysis, the alloys were mounted in epoxy resin, which was perforated to ensure good electrical contact between the sample, electrolyte, and working electrode. The test area of the samples was 0.785 cm². In conducting the corrosion tests, a commercial ACM Instruments model Gill AC equipment was used, following the electrochemical polarization technique (Buchanan and Stansbury, 2012Buchanan, R.A., Stansbury, E.E. (2012). 4 - Electrochemical Corrosion. In Handbook of Environmental Degradation of Materials. Second Edition, William Andrew Publishing, pp. 87-125. https://doi.org/10.1016/B978-1-4377-3455-3.00004-3.), establishing an electrochemical sweep range of -250 to 250 mV versus the resting potential, at a rate of 10 mV·min^-1.

To determine the Young’s and Poisson’s elastic moduli of the alloys under study, the diametral compression test was employed, for which two strain gauges with 2.11 NOM gauge factor were used, that were placed longitudinally and transversely on each one of the specimens tested. The compression tests were conducted using a Universal 647 Hydraulic Wedge Grip machine, MTS Landmark, coupled to a p3 Strain Indicator - Record equipment, to obtain the values of transverse and longitudinal micro deformations, for each compression tested material.

3.1. Characterization of Milled Powders

From the PSD analysis, reported in Fig. 1a - 1c, it can be confirmed that in general, the particle size distribution of the alloy series prepared in this work is both multi-sized and multi-modal at the beginning of the milling stage, whereas it tends to be monomodal towards the end of the milling time. The mean particle size of all initial admixtures of elemental powders, i.e., without milling, is smaller than that obtained after 15 to 30 min of mechanical alloying, the lapse in which particle agglomeration predominantly occurs. After around 15 min of the high-energy milling process of the studied system, no apparent effect of yttria content on the alloying mechanism that occurs between chemical elements interacting in the solid state is observed, at least not in the particle size distribution reached (agglomerates of 175 microns). Regardless of the amount of yttria present in the powdered ferritic alloy, no difference is observed between the PSD curves (red) of these three cases (Figs. 1a- 1c). The first clear evidence of the PSD change is observed after 30 min of milling (blue lines) when the yttria content is 0.4 wt.%, revealed by a slight left-shift of the curves towards a smaller particle size (Fig. 1a). In addition, after milling for 45 min under the established conditions, the 0.4 wt.% yttria particles appear to have a preferential influence, narrowing the monomodal geometry of the particles and decreasing their size to approximately 25 µm. In general, from Fig. 1 we can conclude that, even if yttria particles have been added to the ferritic alloy or not, the final size of the particles is less than 20 µm, after 60 min milling. However, it should be noted that even when the three series of alloys were successfully refined and homogenized, the yttria content between them is different and the 0.4 wt% series revealed the highest percentage cumulative volume of particles.

It is remarkable that the absence of yttria in the powdered alloy obtained after the mechanical alloying is associated with a mean particle size up to 35% smaller than that obtained with the addition of yttria. Compared with other investigations in which HEM was not used, but rather conventional milling in a planetary ball mill (Hilger et al., 2015Hilger, I., Bergner, F., Weißgärber, T. (2015). Bimodal Grain Size Distribution of Nanostructured Ferritic ODS Fe-Cr Alloys. In Sintering 2014. Wiley Subscription Services, Inc., 98 (11), pp. 3576-3581. https://doi.org/10.1111/jace.13833.), in our work, a 60% smaller particle size, was achieved, in a considerably shorter milling time. Figure 2 shows the mean particle size evolution obtained from the PDS analysis. Previously, from the experimental data used to plot Fig. 1, it was noticed that, even though PSD curves with bimodal behavior were recorded even after milling for 60 min, both a more homogeneous distribution and the finest particle size were achieved. A separation of the overlapped PSD curves was carried out using the Fityk data analysis program, for the three alloys milled for 60 min. It was found that the powder of the yttria-free alloy contains particles whose sizes ranged from 8 to 33 µm, whereas those for the alloy with 0.4 wt.% Y₂O₃ ranged from 20 to 134 µm, and those for the alloy with 0.8 wt.% Y₂O₃ ranged from 40 to 276 µm. From these results, it can be observed that the addition of yttria to the studied alloy entails a substantial increase in its particle size, making it difficult to be refined.

Figures 3 to 5Figures 3, 4, 5 show respectively the XRD patterns of the powdered alloys containing 0.4, 0.8 and 0 wt.% Y₂O₃. All these figures show the same trend, as the mechanical milling time advances, the broadening of the peaks (FWHM, full width at half-maximum intensity) increases, which indicates a decrease in the crystallite size of the ferritic matrix (Dash et al., 2018Dash, M.K., Mythili, R., Ravi, R., Sakthivel, T., Dasgupta, A., Soroja, S., Bakshi, S.R. (2018). Microstructure and mechanical properties of oxide dispersion strengthened 18Cr-ferritic steel consolidated by spark plasma sintering. Mater. Sci. Eng. A 736, 137-147. https://doi.org/10.1016/j.msea.2018.08.093.). In these XRD patterns, overlapping of the diffraction peaks occurred due to the presence of chromium traces and α-Fe in the alloy’s matrix, whose diffraction angles are very close to each other. The presence of yttria is not observed in the XRD patterns of Figs. 3 and 5, since it was added in such a proportion that it was below the detection limit of this technique (Sun et al., 2012Sun, Q.X., Zhang, T., Wang, X.P., Fang, Q.F., Hao, T., Liu, C.S. (2012). Microstructure and mechanical properties of oxide dispersion strengthened ferritic steel prepared by a novel route. J. Nucl. Mater. 424 (1-3), 279-284. https://doi.org/10.1016/j.jnucmat.2011.12.020.).

As for the lattice parameter and crystallite size of the α-Fe phase present in the ODS alloys, as can be seen in Fig. 6, in the case of the powdered alloy without Y₂O₃ and no sintering (0 NS), the lattice parameter and crystallite size of the α-Fe phase was 0.286 nm and 500 nm, respectively. After sintering, this alloy showed an increase in the lattice parameter to 0.289 nm, indicating dissolution of the alloying elements into its ferritic matrix with sintering.

Now, for the alloy containing 0.4 wt.% Y₂O₃, both the lattice parameter and the crystallite size decreased to 0.2872 and 250 nm, respectively, indicating a refinement of the grain size of the ferritic matrix of the alloy (Cayron et al., 2004Cayron, C., Rath, E., Chu, I., Launois, S. (2004). Microstructural evolution of Y₂O₃ and MgAl₂O₄ ODS EUROFER steels during their elaboration by mechanical milling and hot isostatic pressing. J. Nucl. Mater. 335 (1), 83-102. https://doi.org/10.1016/j.jnucmat.2004.06.010.). For the alloy containing 0.8 wt.% Y₂O₃, there was an increment in both the lattice parameter and crystallite size to 0.288 and 498 nm, respectively. The abrupt change in the magnitudes of those parameters requires further investigation.

Figure 7 shows a sequence of SEM micrographs taken from the studied powder as a function of the milling time. The morphology changes followed by the admixture of precursor powders are evident. As it was mentioned in the previous section on the PSD analysis, between 15 and 30 min of milling, the particle´s growth is clear, which is nothing but the agglomeration of the particles. As for the alloy having no yttria (series 1), it can be observed that after 30 min of milling the particles, they become flakier, unlike the two yttria-containing alloys (series 2 and 3). From this observation, it can be suggested that, for the case of the yttria-containing alloys, after 45 min of milling the materials experienced cold deformation hardening, eventually fragmenting into smaller particles. For the alloy lacking yttria (series 1), it can be noted that the particles processed at 45 (c1) and 60 min (d1) are similar in shape and size to each other, which indicates that the fracture process between particles tends to stabilize after 45 min. Other authors (Fu et al., 2019Fu, J., Brouwer, J.C., Richardson, I.M., Hermans, M.J.M. (2019). Effect of mechanical alloying and spark plasma sintering on the microstructure and mechanical properties of ODS Eurofer. Mater. Des. 177, 107849. https://doi.org/10.1016/j.matdes.2019.107849.), who did not use high-energy milling, reported that the stable process for a ferritic alloy with oxide dispersion is reached after 13 h, meanwhile, in this work using HEM, such stability is achieved in 45 min, which significantly reduces energy and thus cost.

3.2. Characterization of sintered alloys

The phase changes undergone by the ODS alloys heated up to 1100 ^oC were recorded by dilatometry. In Fig. 8, the T₁, T_3, and T₅ points show the temperature at which the phase transformation from ferrite to austenite (α → γ) begins. Points T₂, T_4, and T₆ correspond to the final temperature of such transformation (Cayron et al., 2004Cayron, C., Rath, E., Chu, I., Launois, S. (2004). Microstructural evolution of Y₂O₃ and MgAl₂O₄ ODS EUROFER steels during their elaboration by mechanical milling and hot isostatic pressing. J. Nucl. Mater. 335 (1), 83-102. https://doi.org/10.1016/j.jnucmat.2004.06.010.), for the ODS ferritic alloys with 0.4, 0.8, and 0 wt.% Y₂O₃, respectively. Points T_M1, T_M2, and T_M3 correspond to the temperature at which the martensitic transformation occurs for the three alloys, respectively. Figure 8 allows verifying that the ODS ferritic alloys with 0 and 0.8 wt.% Y₂O₃ have similar behavior to each other, regarding their linear thermal expansion (dL/L _o ). However, the initial α → γ transformation temperature is lower for the alloy without yttria (T₅ = 760 ^oC). Comparatively, in the case of the alloy with 0.4 wt.% Y₂O₃, such a phase transition begins at 820 ^oC (T₁ = T₃), extending up to 860 ^oC (T₂), which is associated with a linear thermal expansion ∆L approximately 12.5% higher than that for its counterparts with 0 and 0.8 wt.% Y₂O₃. In the case of these last two alloys, the ferrite structure is more stable and therefore it does not fully transform into austenite.

Yamamoto et al. (2010)Yamamoto, M., Ukai, S., Hayashi, S., Kaito, T., Ohtsuka, S. (2010). Formation of residual ferrite in 9Cr-ODS ferritic steels. Mater. Sci. Eng. A 527 (16-17), 4418-4423. https://doi.org/10.1016/j.msea.2010.03.079. added 0.35 and 0.7 wt.% Y₂O₃ to an alloy Fe-0.13C-2W-0.2Ti, finding by mathematical simulation that the speed of movement of the ferrite-austenite interface decreases by the dragging of oxide particles. On the other hand, Kumar et al. (2018)Kumar, D., Prakash, U., Dabhade, V.V., Laha, K., Sakthivel, T. (2018). Influence of Yttria on Oxide Dispersion Strengthened (ODS) Ferritic Steel. Mater. Today: Proc. 5 (2, Part 1), 3909-3913. https://doi.org/10.1016/j.matpr.2017.11.646. concluded that contents of Y₂O₃ higher than 0.5 wt.% lead to the production of materials with high hardness and high elastic modulus due to work hardening, generating microcracks that cause fractures, and, therefore, a lower ductility in the material. The probable occurrence of all these phenomena in our series of ferritic alloys analyzed is high, particularly in the one with the highest percentage (0.8 wt.%) of Y₂O₃. This is explained by the fact that this oxide acts as an element that not only anchors microstructural defects, with the consequent hardening of the alloy but also prevents the linear thermal expansion of the material. Considering that yttria is a ceramic with predominantly ionic-covalent and non-metallic bonds, it is expected that its presence in the metallic matrix reduces the longitudinal elongation of the entire specimen.

Regarding the martensitic phase transformation (γ - α´) (Zhou et al., 2017Zhou, X., Liu, Y., Qiao, Z., Guo, Q., Liu, Ch., Yu, L., Li, H. (2017). Effects of cooling rates on δ-ferrite/γ-austenite formation and martensitic transformation in modified ferritic heat resistant steel. Fusion Eng. Des. 125, 354-360. https://doi.org/10.1016/j.fusengdes.2017.05.095.), it first took place in the 0.4 wt.% Y₂O₃ alloy during the cooling stage (T_M1 = 188 ^oC), while in the 0 and 0.8 wt.% Y₂O₃ alloys (blue and red curves in Fig. 8, respectively) it happened at the same lower temperature (T_M2 = T_M3 = 165 ^oC). Since the latter two alloys have a similar dilatometric behavior they disclosed a 10% difference concerning their starting temperature, for the α → γ transformation. Since, according to Fig. 8, the largest area under the heating and cooling curves corresponds to the ODS alloy with 0.4 wt.% Y₂O₃, it follows that this material is the one undergoing the greatest thermal expansion among all the alloys studied.

In general, it was expected to find a marked influence between the yttria content and the formation of martensite, and therefore on its mechanical properties. In practice, however, the obtained alloys showed a positive effect of yttria regarding their performance against corrosion and adequate mechanical resistance.

In this way, the temperature that indicates the start of the martensitic transformation (T_M) is practically the same for two of the alloys studied (T_M2 = T_M3), presenting only a negligible difference of 23 °C concerning the alloy with 0.4 wt.% of Y₂O₃, being greater for the latter alloy. The value of T_Mis rather related to the grain size of the austenite, as well as the amount of available ferrite (Zhou et al., 2017Zhou, X., Liu, Y., Qiao, Z., Guo, Q., Liu, Ch., Yu, L., Li, H. (2017). Effects of cooling rates on δ-ferrite/γ-austenite formation and martensitic transformation in modified ferritic heat resistant steel. Fusion Eng. Des. 125, 354-360. https://doi.org/10.1016/j.fusengdes.2017.05.095.). As it has been mentioned, the alloy with 0.4 wt.% Y₂O₃ presents the highest amount of austenite, therefore, it exhibits the highest dimensional change during cooling.

On average, the relative density (RD) attained by the alloys prepared in this work is above 93%, which is higher than the results reported by other authors (Mihalache et al., 2019Mihalache, V., Mercioniu, I., Velea, A., Palade, P. (2019). Effect of the process control agent in the ball-milled powders and SPS-consolidation temperature on the grain refinement, density and Vickers hardness of Fe14Cr ODS ferritic alloys. Powder Technol. 347, 103-113. https://doi.org/10.1016/j.powtec.2019.02.006.), who also used SPS sintering (for a Fe-14Cr-0.4Ti-0.25Y₂O₃ alloy). The RD values obtained for the 0, 0.4, and 0.8 wt.% Y₂O₃ alloys of this study are 7.8, 7.5, and 8.5%, respectively. Fu et al. (2019)Fu, J., Brouwer, J.C., Richardson, I.M., Hermans, M.J.M. (2019). Effect of mechanical alloying and spark plasma sintering on the microstructure and mechanical properties of ODS Eurofer. Mater. Des. 177, 107849. https://doi.org/10.1016/j.matdes.2019.107849. reported lower RD values (5.2, 5.5, and 4.5%) for a Fe-9Cr-1.1W-0.4Mn-0.2V-0.12Ta-0.3Y₂O₃ alloy. Other research works (Abenojar et al., 2006Abenojar, J., Velasco, F., Martínez, M.A. (2006). Manufacturing of Porous Boron Steels Potentially Useful as Nuclear Materials. J. Nucl. Sci. Technol. 43 (8), 866-873. https://www.tandfonline.com/doi/abs/10.1080/18811248.2006.9711171.), in which the 90Fe/10B₄C alloy was sintered using conventional techniques, without oxide dispersion, and also considering potential applications in nuclear reactors, report RD values which are 13.0, 12.6, and 13.8% lower than those here reported. The density values obtained for the alloys in this work are similar to those reported by other authors (for a Fe-14Cr-0.4Ti-0.25Y₂O₃ alloy) (Zhang et al., 2015Zhang, H., Huang, Y., Ning, H., Williams, C.A., London, A.J., Dawson, K., Hong, Z., Gorley, M.J., Grovenor, R.M., Tatlock, G.J., Roberts, S.G., Reece, M.J., Yan, H., Grant, P.S. (2015). Processing and microstructure characterisation of oxide dispersion strengthened Fe-14Cr-0.4Ti-0.25Y₂O₃ ferritic steels fabricated by spark plasma sintering. J. Nucl. Mater. 464, 61-68. https://doi.org/10.1016/j.jnucmat.2015.04.029.), who also used SPS and who, unlike the present study, did consider a holding time at the peak sintering temperature.

In the practice of this work, both the morphology and the particle size distribution directly affect the degree of densification achievable during the sintering process. The addition of 0.8 wt.% Y₂O₃ to the ferritic alloy leads to an increase in its relative density (up to a value of 94.5%), whereas with 0.4 wt.% of agglomerated Y₂O₃ particles, the RD value is drastically decreased. In the case of the yttria-free alloy, the density value is only 0.7% lower than that of the 0.8 wt.% Y₂O₃ alloy. Therefore, these results are not only attributed to the morphology of the processed particles, but also to their fluidity. As it was already mentioned in the microscopy section, the alloy with 0.8 wt.% Y₂O₃ has fewer particle agglomerates, compared to its counterpart with 0.4 wt.% Y₂O₃, while the alloy without yttria is constituted by flaky particles. That is, the ODS alloy with the highest yttria content has a wider particle size distribution, which results in better packing of the particles when they are subjected to the sintering process. The low density reached by the 0.4 wt.% Y₂O₃ alloy series is attributed to the agglomerated state and clogging effect of the powder particles, which prevents adequate compaction during sintering.

Figure 9 reveals the type of microstructure developed by the powder containing 0 (a - d), 0.4 (b - e), and 0.8 wt.% Y₂O₃ (c - f), after milling for 60 min followed by sintering by SPS at 1100 ^oC. Micrographs d), e), and f) are magnifications of a), b), and c), respectively, to reveal the presence of black pores with semicircular morphology (indicated by the black arrows and letter P), as well as the presence of γ-Fe precipitates, which are identified by the letter G. In Fig. 9 (a, b, c), scattered narrow, and elongated white areas are distinguished, corresponding to W debris (red arrows). In micrographs b) and c), mainly, the presence of such traces of elemental W can be seen, elemental tungsten traces are part of the original powder precursor mix. It was observed that at short milling times of up to 15 min, W goes into a solid solution in the ferritic matrix, but at longer times it dissolves, up to a certain solubility limit. At longer milling times, using XRD, traces of W were detected in its elemental form (Figs. 3, 4 and 5), which is explained as W in stoichiometric excess, which remains mechanically dispersed in the matrix. From these micrographs, not only fine pores of less than 10 μm but also α- and γ-Fe precipitates are evenly distributed throughout the surface of the sintered samples.

The XRD patterns obtained for the three alloys whose powders were milled for 60 min and subsequently sintered by SPS are shown in Fig. 10. The sintered alloys show oxidation of chromium, since this element has a higher affinity for oxygen when compared with Fe and other alloying elements (Yaskiv and Fedirko, 2014Yaskiv, O.I., Fedirko, V.M (2014). Oxidation/Corrosion Behaviour of ODS Ferritic/Martensitic Steels in Pb Melt at Elevated Temperature. Int. J. Nucl. Energy ID 657689, 1-8. https://doi.org/10.1155/2014/657689.) so when these alloys are sintered at 1100 ^°C, a passive surface layer of Cr₂O₃ is formed. The analysis also revealed the presence of elemental traces of Cr and W, as well as of the α-Fe phase. Even though the austenite γ-Fe phase precipitated in the three alloys, the presence of this phase is only clear in the 0.4 wt.% Y₂O₃ alloyed. Under the processing and sintering conditions employed in this work, in these alloys, Fe grains have a fcc lattice structure (γ-Fe phase). Judging from the SEM micrographs shown in Fig. 10 and corresponding XRD patterns, it is evident that, upon sintering, the α- to γ-Fe phase transformation occurred through the formation of submicron-sized grains.

Some of the fundamental mechanical properties of the studied alloys are plotted in Fig. 11, as a function of their yttria content. These parameters are Young’s modulus, Poisson’s ratio, and Vickers microhardness. In general, except for the Poisson’s ratio , the largest values of these properties are observed for the alloy containing 0.8 wt.% Y₂O₃.

For both Young’s modulus and Poisson’s ratio, the magnitude of the values obtained is related to the anisotropy of the microstructure as well as to the direction in which the measurement was made (Sánchez-Gutiérrez et al., 2017Sánchez-Gutiérrez, J., Chao, J., Vivas, J., Galvez, F., Capdevila, C. (2017). Influence of texture on impact toughness of ferritic Fe-20Cr-5Al oxide dispersion strengthened steel. Materials 10 (7), 745. https://doi.org/10.3390/ma10070745.). On the other hand, the low values of the Poisson ratio were due to the difference between the transverse deformation and the longitudinal deformation, the former being much lower than the latter.

The ODS alloy without yttria is the one that showed the lowest value of the elastic modulus, which is congruent with its highest hardness value. The ferritic alloy with 0.8 wt.% Y₂O₃ disclosed the least dimensional change as a function of temperature, whereas its 0.4 wt.% Y₂O₃ counterpart suffered both the greatest linear expansion and the major microstructural changes, which influences its elastic properties. The 0.4 wt.% Y₂O₃ alloy also disclosed the lowest hardness value and the smaller crystallite size (Fig. 6). As it can be seen in Fig. 11, the Young module rises as a function of the Y₂O₃ content in the alloy. Thus, the 0.4 wt.% Y₂O₃ alloy was more ductile than the 0.8 wt.% Y₂O₃ alloy.

These results agreed with Kumar et al. (2018)Kumar, D., Prakash, U., Dabhade, V.V., Laha, K., Sakthivel, T. (2018). Influence of Yttria on Oxide Dispersion Strengthened (ODS) Ferritic Steel. Mater. Today: Proc. 5 (2, Part 1), 3909-3913. https://doi.org/10.1016/j.matpr.2017.11.646., who conclude that contents of Y₂O₃ higher than 0.5 wt.% result in alloys with high hardness and high elastic modulus due to work hardening, generating microcracks that cause fractures, and, therefore, a lower ductility in the material.

When compared with other works (Dash et al., 2018Dash, M.K., Mythili, R., Ravi, R., Sakthivel, T., Dasgupta, A., Soroja, S., Bakshi, S.R. (2018). Microstructure and mechanical properties of oxide dispersion strengthened 18Cr-ferritic steel consolidated by spark plasma sintering. Mater. Sci. Eng. A 736, 137-147. https://doi.org/10.1016/j.msea.2018.08.093.), where ODS ferritic alloys have been synthesized, and have also been densified using the SPS technique, it is found that the values obtained in the present study for the elastic modulus are 30% lower, particularly in the case of the 0.4 wt.% Y₂O₃ alloy. From the value of the Poisson’s ratio reached by the latter alloy, it is verified that this is the one undergoing the most deformation, particularly in some of its axes, when the compressive load is applied (Hook’s law).

In Fig. 11 it is observed that the 0.4 wt.% Y₂O₃ alloy was the one showing the lowest Vickers microhardness value (865 HV), as well as the highest degree of porosity and the lowest relative density RD, among all the alloys studied. For comparison purposes, the experimental results of this work are included at the end of Table 1, along with those reported in the literature.

The addition of yttria particles into the ODS ferritic alloys leads to both a smaller and more homogeneous grain size. Moreover, the addition of Y₂O₃ particles hinders defects in the alloys, since it might occupy the room left for defects (interstitials and vacancies), occasionally avoiding the movement of dislocations in the crystalline structure (Kumar et al., 2018Kumar, D., Prakash, U., Dabhade, V.V., Laha, K., Sakthivel, T. (2018). Influence of Yttria on Oxide Dispersion Strengthened (ODS) Ferritic Steel. Mater. Today: Proc. 5 (2, Part 1), 3909-3913. https://doi.org/10.1016/j.matpr.2017.11.646.).

The 0.4 wt.% Y₂O₃ alloy was found to be more ductile, which is consistent with what was mentioned by Kumar et al. (2018)Kumar, D., Prakash, U., Dabhade, V.V., Laha, K., Sakthivel, T. (2018). Influence of Yttria on Oxide Dispersion Strengthened (ODS) Ferritic Steel. Mater. Today: Proc. 5 (2, Part 1), 3909-3913. https://doi.org/10.1016/j.matpr.2017.11.646., who stated that a Y₂O₃ content greater than 0.5 wt.% reduces ductility. On the other hand, a greater presence of the austenite phase generates a volumetric increase when the phase change from ferrite to austenite occurs, modifying the microstructure and thus, affecting the resulting properties of the alloys. The high porosity remaining in the samples plays an important role in both ductility and hardness. The 0.4 wt.% Y₂O₃ alloy has a lower relative density (1.01%) than the 0.8 wt.% Y₂O₃ alloy, which is also a factor that decreases hardness.

The obtained alloys show a heterogeneous microstructure, thereby an anisotropic behavior would be expected, in such a way that, depending on the orientation or direction in which the specimens are placed for the measurement to be made, the value of the elastic modulus might change. In this respect, Sánchez-Gutiérrez et al. (2017)Sánchez-Gutiérrez, J., Chao, J., Vivas, J., Galvez, F., Capdevila, C. (2017). Influence of texture on impact toughness of ferritic Fe-20Cr-5Al oxide dispersion strengthened steel. Materials 10 (7), 745. https://doi.org/10.3390/ma10070745. carried out a mathematical simulation study of an alloy with Y₂O₃ dispersion (Fe-18.6Cr-5.2Al-0.54Ti-0.04C-0.006N-0.5Y₂O₃). The calculations were made in the crystallographic directions x [100], y[110], and z[111], finding Young’s modulus values of 131, 219 and 238 GPa, very close to those found (234 GPa) for the 0.8 wt.% Y₂O₃ alloy of this work.

A representative graph of the corrosion behavior recorded for the studied alloys, when tested in 1 M HNO₃, is presented in Fig. 12. Unlike both the 0.4 and 0.8 wt.% Y₂O₃ alloys, the anodic region of the potentiodynamic polarization curve (blue) obtained for the yttria-free alloy revealed the occurrence of two passivation events (P₁ and P₂), which are attributed to the formation of a Cr₂O₃ layer which momentarily protects the base alloy, until the P₃ trans passivation zone appears, indicating the breakdown of such protective layer. Meanwhile, the 0.4 and 0.8 wt.% Y₂O₃ alloys, whose current density is higher, showed similar behavior to each other (red and black curves), both showing a long-lasting trans passivation zone (TP), which protects the alloys during longer times, compared to the yttria-free alloy. That is, even though the latter alloy exhibits a lower current density, in such a way that it also shows greater corrosion resistance, the fact that its protective layer breaks down before that formed at the surface of its yttria-containing counterparts makes it less effective

The fact of processing the precursor metallic powders, via high-energy ball milling (HEM) allows introducing microstructural defects into the ODS alloy matrix, leading the powder blend mixture into a highly mechano-activated state. The metastable state of the SPS-densified samples is somehow preserved after the rapid sintering, which in turn causes the ODS alloy to react eagerly. EDS analyses (Fig. 13) indicate that, after conducting the corrosion test, a Cr₂O₃ passive layer is formed on the surface of the alloy having no Y₂O₃, whereas in alloys with 0.4 and 0.8 wt.% Y₂O₃ oxides layers of Fe and Cr formed. The Cr layer preferentially expands the passivation zone, as revealed from the potentiodynamic polarization curve, thus associated with Y₂O₃. Thermodynamically, it is easier for some chemical states of iron to be reduced than oxidized, so comparatively in this case, chromium is energetically easier to oxidize, thus forming the protective chrome layer. The chemical affinity between Cr and Y was reported by Wang C. Zhen through the formation of YCrO₃ (Chang-Zhen et al., 1985Chang-Zhen, W., Shu-Qhing, Y., Xin, Z. (1985). A Study On Thermodynamic Properties Of Y₂O₃·Cr₂O₃ Compound. Acta Phys. Sin. 34 (8), 1017-1026. https://doi.org/10.7498/aps.34.1017.) . Although this compound is expected to form at high temperatures, through the reaction Y₂O₃ + Cr₂O₃ → 2YCrO₃, the vigorous mechano-chemical action might eventually allow its formation. Owing to the limited proportion of yttrium oxide added to the ferritic alloy, the XRD pattern showing YCrO₃ is not obvious, but shows Cr₂O₃. Ningshen et al. (2014)Ningshen, S., Sakairi, M., Suzuki, K., Ukai, S. (2014). The corrosion resistance and passive film compositions of 12% Cr and 15% Cr oxide dispersion strengthened steels in nitric acid media. Corros. Sci. 78, 322-334. https://doi.org/10.1016/j.corsci.2013.10.015., claimed that Y₂O₃ promoted the formation of complex Y₂Hf₂O₇ oxide in 15% Cr ODS steel, and this oxide may have a role in suppressing intergranular corrosion attack. The addition of Y₂O₃ has also been reported to increase corrosion resistance in Al 6061 aluminum alloy (Ahmadi and Nouri, 2010Ahmadi, H., Nouri, M. (2010). Beneficial effects of yttrium on mechanical failure and chemical stability of the passive film in 6061 aluminum alloy. J. Mater. Sci. 45 (13), 3426-3432. https://doi.org/10.1007/s10853-010-4368-9.). On the other hand, Y₂O₃ decreases the grain size (Fig. 14) and thus increases the grain boundary area. However, due to the nature of the more stable passive layer, this oxide increases the resistance to intergranular attack, which agrees with other results (Ningshen et al., 2014Ningshen, S., Sakairi, M., Suzuki, K., Ukai, S. (2014). The corrosion resistance and passive film compositions of 12% Cr and 15% Cr oxide dispersion strengthened steels in nitric acid media. Corros. Sci. 78, 322-334. https://doi.org/10.1016/j.corsci.2013.10.015.).

Figure 12 shows three zones. The first one goes from the equilibrium potential to point P₁ and is related to the formation of a first passivation zone. The second one, corresponding to the so-called “active-passive zone”, is related to the breakdown of the first passivation zone and extends from point P₁ to point P₂, and the third one, corresponding to the second passivation zone, extends from point P₂ to point P₃. Table 3 shows the values of current densities and potentials recorded at points P₁, P₂ and P₃.

For the Yttria-free alloy, the breakdown point is located at 10^-1.30 mA·cm^-2 and 96.9 mV vs E Ag/AgCl electrode. In comparison, the breakdown points of the Yttria-containing alloys are located at lower current density values (10^0.73 mA·cm^-2 for the 0.4 wt.% Y₂O₃ alloy, and 10^0.42 mA·cm^-2 for the 0.8 wt.% Y₂O₃ alloy).

In the cathodic branch of the curve, the current limit density for oxygen is -10^1.19 mA·cm^-2 for the alloy without yttria, and - 10^2.057 mA·cm^-2 and -10^2.11 mA·cm^-2 for 0.4 wt.% and 0.8 wt.% Y₂O₃ alloys, respectively. Thus, the Yttria-containing alloys have more current density in the cathodic branch to balance the corrosion current density.

From Fig. 15 it is observed that the dimensional shrinkage/expansion, it is also possible to verify the effect of the longitudinal contraction suffered by the three alloys during the heat treatment, while the α → γ phase change takes place, as a function of the yttria content. It is observed that the 0.8 wt.% Y₂O₃ alloy is the one exhibiting the least dimensional change, while the 0.4 wt.% Y₂O₃ alloy is the one undergoing the greatest expansion due to the influence of the phase transformation on its elastic properties. A higher content and therefore better dispersion of yttria particles in these alloys reduce the atomic movement in their crystal lattice, which leads to the generation of fewer micro-crystalline defects.