**Gradient Distribution of Microstructures and Mechanical Properties in a FeCoCrNiMo High-Entropy Alloy during Spark Plasma Sintering**

**Mingyang Zhang 1, Yingbo Peng 2, Wei Zhang 1,\*, Yong Liu 1, Li Wang 3, Songhao Hu <sup>4</sup> and Yang Hu <sup>5</sup>**


Received: 31 January 2019; Accepted: 15 March 2019; Published: 19 March 2019

**Abstract:** A novel graded material of a high-entropy alloy (HEA) FeCoCrNiMo was fabricated by spark plasma sintering (SPS) processing. After SPS, the HEA specimens consisted of a single face-centred cubic (FCC) phase in the center, but dual FCC and a tetragonal structure σ phase near the surface. Surprisingly, the sintering pressure was sufficient to influence the proportion of phases, and thus the properties of HEA samples. The hardness of the specimens sintered under the pressures of 30, 35, and 40 MPa increased gradually from 210 HV0.2, which is the single FCC phase in the center, to the maximum value near the surface as a result of the gradual increase in the fraction of the transformed σ phase. The σ phase, being a complex hard and brittle intermetallic particle to manipulate the properties of FCC-type HEA systems, which could be influenced by pressure, indicated a major possibility for designing gradient HEA materials.

**Keywords:** high-entropy alloy; spark plasma sintering; pressure; microstructure; mechanical properties

#### **1. Introduction**

Because of the rapid development of modern engineering and manufacturing industries, high-performance alloys urgently need to be developed. High-entropy alloys (HEAs) constitute a unique class of alloys exhibiting high strength and hardness, decent wear and corrosion resistance, and other attractive mechanical properties for both scientific research and practical applications [1–4]. For further studies on HEAs, phase transformations could be crucial for controlling their microstructures to obtain superior properties [5–7]. A significant amount of research on HEAs has been performed to study phase transformation using vacuum arc melting [8]; this procedure was usually restricted to laboratory settings. Using spark plasma sintering (SPS) to consolidate mechanically alloyed HEA powders is a promising method to obtain high-performance bulk HEAs [9–12]. The sintering pressure is one of the most important parameters in the SPS method [13]. Moreover, pressure will influence the equilibrium between the gaseous and liquid phases. The influence of pressure is typically neglected for the equilibrium of two or more solid phases. Standard phase diagrams involve only composition and temperature as the relevant variables [14]. This approach is based on phase transformations involving only solid phases not being associated with significant volume changes, unlike changes from liquid to gas. According to this theory, if the pressure could be

controlled and further influence the phase transformation, it would represent a breakthrough in the field of controlling the structure and performance of HEAs.

In the present study, a FeCoCrNiMo HEA was successfully prepared using the SPS method under different pressures. The gradient distribution of microstructures and mechanical properties in the SPS samples were investigated. The face-centered cubic (FCC) to a tetragonal structure σ phase transformation was also studied. There is the possibility of the σ intermetallic compounds [15]. Consequently, the microstructure of the alloy can be varied by careful control of thermal treatment. Moreover, this phase transformation of FCC to σ phase, which can be influenced by SPS pressure, indicates a major possibility to design a novel gradient material of HEAs used in tools and dies.

#### **2. Experimental**

The investigated alloy with a nominal composition Fe24.1Co24.1Cr24.1Ni24.1Mo3.6 (in at. %) was prepared using powder metallurgy (99.9%, Vilory new materials Co. Ltd, Xuzhou, China). Powders consisting of particles under 200 mesh in size were prepared using gas atomization and were mechanically milled using conventional planetary-milling equipment. Next, the powder with a particle size under 200 mesh was mechanically milled using conventional planetary milling equipment. The weight ratio between the powder and the stainless-steel balls was 1:10 and ethanol was added as the milling medium. The milling time was 20 h and the milling speed was 300 rev min<sup>−</sup>1. The milled powders were then added into a graphite die 40 mm in diameter and consolidated using an HPD 25/3 SPS equipment under reduced pressure (10−<sup>3</sup> Pa). The sintering temperature was 1150 ◦C and the pressures were 30, 35, and 40 MPa. After a holding time of 480 s, the sintered billets were cooled down to room temperature in the furnace. Samples were prepared by mechanical grinding using 1200 to 4000 grit SiC papers followed by a final polishing step (size: <sup>ϕ</sup> <sup>40</sup> × 2 mm3). The transverse fracture strength of the samples (size: 12 × <sup>2</sup> × 30 mm3) was determined by an Instron 3369 mechanical testing facility (Instron, Norwood, MA, USA) using the three-point method (span length: 25 mm, test speed: 2 mm/min), and tested twice for each treatment. A FEI Quanta FEG 250 scanning electron microscope (SEM, FEI, Hillsboro, OR, USA) equipped with an energy-dispersive X-ray (EDX) analyzer was used to investigate the microstructure and chemical compositions of the sintered specimens (20 kV, using spot analysis and backscattering mode). A Philips CM 200 transmission electron microscopy (TEM, Royal Philips, Amsterdam, The Netherlands) operating at 200 kV was used to identify the structure of the precipitates by selected area electron diffraction (SAED) analysis. The TEM specimen was prepared by a crossbeam workstation AURIGA 40 (Zeiss, Oberkochen, Germany) equipped with a focused ion beam (FIB) column and scanning electron microscopy (SEM). The phase constitution of the specimens used a Rigaku Rapid IIR (Rigaku, Tokyo, Japan) micro-area X-ray diffractometer (XRD, 40 kV, from 20◦ to 100◦, a circular region with a diameter of 30 μm, PDF database 2009, using 20 minutes for each position) equipped with a 2D detector (phi: 360◦, omega: −15◦~150◦) utilizing Cu Kα radiation. The phase transition temperature was analyzed by differential scanning calorimeter (DSC) using a NETZSCH STA 449C thermal analyzer (RT~1300 ◦C, 40 K/min, Ar atmosphere; Netzsch, Selb, Germany). The hardness of the alloy was determined using Buehler 5104 hardness tester (Buehler, Lake Bluff, IL, USA) under a 200 g load for 15 s and was averaged from three measurements. The indentation profile was obtained by NanoMap 500 DLS 3D surface profiler (Aep Technology, Santa Clara, CA, USA).

#### **3. Results and Discussion**

#### *3.1. Microstructure*

The morphology and phase composition of the powders before and after ball milling are shown in Figure 1a,b. The previously spherical particles of HEA powders were crushed to form irregular shape particles; the specific surface area of powders increased. As the specific surface area of the particles increases, a higher sintering driving force can be obtained, and the degree of sintering densification is improved. According the XRD pattern in Figure 1c, both powders exhibit an FCC structure. This means that the ball milling process does not lead to phase transition.

**Figure 1.** Microstructure of FeCoCrNiMo high-entropy alloy (HEA) powders before (**a**) and after (**b**) ball milling, (**c**) X-ray diffractometer (XRD) pattern of two powders. FCC—face-centered cubic.

Figure 2a–c shows the longitudinal cross-sections of the HEAs samples sintered at 1150 ◦C under different pressures. Judging from the contrast images, the microstructure did not consist of a single phase, as did those obtained after lower temperature sintering in our previous study [16].

**Figure 2.** Longitudinal cross-sections of the same area in the FeCoCrNiMo HEA samples sintered at 1150 ◦C under different pressures: (**a**) 30, (**b**) 35, and (**c**) 40 MPa, as well as (**d**) the macrostructure of the 30 MPa sintered sample.

The microstructure gradually changed along the radial direction, towards the center of the samples. Obviously, the phase transformation occurred on the upper and lower surfaces of the three samples during SPS. The thickness of the phase transformed region decreased gradually along the radial direction of the cylinder towards both ends, as shown in Figure 2d. The reason for this phenomenon is the inhomogeneous distribution of the temperature field along the radial direction in the cylindrical graphite die during SPS. By simulating the temperature field of graphite die in the SPS process, the sintering temperature at the center is about 1700 ◦C and the border of the sample and the die can be as high as 450 ◦C [17]. Thus, because of the inhomogeneous distribution of temperature field, the volume fraction of phase transition is less than that of the middle part.

Under the sintering pressure of 30 MPa, the microstructure exhibited a distinct gradient distribution, as illustrated in Figure 3. It can be seen that multiple phase structures were successfully synthesized during SPS. The volume fraction of the transformed phase increased gradually from the center to edge of the specimen. As shown in Figure 3b, the volume fraction of the transformed phase decreased as the depth increased, which shows that sintering pressure directly affects the degree of phase transformation and presents a gradient distribution. For example, when the sintering pressure is 30 MPa, the volume fraction of the transformed phase is reduced from about 27% at the edge to 14% at a depth of 550 microns, and the volume fraction is reduced by half. When the sintering pressure is 40 MPa, the volume fraction of the transformed phase is sharply reduced from about 19% at the edge to 1% at a depth of 550 μm, which is about 20 times lower. It can be seen that the sintering pressure has a significant influence on the transformed phase volume fraction and the distribution of the transformed phase along the thickness direction. By controlling the sintering pressure, the gradient distribution of the multiple phase along the pressure direction can be realized and regulated.

**Figure 3.** Scanning electron microscopy (SEM) image (**a**) and volume fraction of the transformed phase (**b**) of the FeCoCrNiMo HEA after sintering at 1150 ◦C and different sintering pressures. The mark shows the approximate position of the energy-dispersive X-ray (EDX) map.

#### *3.2. Phase Identification*

To understand the phase transformation, the phase constitution of the specimens was investigated using a combination of SEM, EDX, and XRD techniques.

As shown in Figure 4, the distribution of the elements Fe, Co, Cr, Ni, and Mo could be clearly identified in the samples. The precipitated phase was a Cr-rich phase, containing relatively low amounts of Co and Ni. No intermetallic compound was formed; therefore, the precipitate was a hard, Cr-rich σ phase. Powder metallurgy with a fast cooling rate was employed to reduce the preferred orientation effects observed in the cast alloys [18], the result of which could be used for accurately measuring the local lattice distortion. The distribution of the FCC and σ phase structures in the specimen can be identified using XRD analysis, as shown in Figure 5. The center of the specimen that underwent SPS under 30 MPa of pressure was a single FCC structure. While gradually transitioning towards the surface of the specimen, the {111} FCC diffraction peak became less intense than that of the center. However, the peaks corresponding to the {110} and {200} planes could be identified, indicating the formation of the σ phase structure. In addition, EDX analyses (Table 1) showed that the FCC matrix was rich in Mo and Ni, whereas the precipitated σ phase was rich in Cr, indicating that the matrix was represented by the FCC phase and the precipitate was the ordered σ phase. The XRD peak intensities are in a good accordance with the polycrystalline powders and the chemical compositions. Previous studies on the phase transformation of HEAs caused by SPS or the vacuum arc melting method reported that the body-centered cubic (BCC) phases or a tetragonal structure σ phase were composed of a spinodally modulated matrix, and precipitates exhibiting a near-equiaxed shape were distributed uniformly throughout the HEA. The FCC phases exhibiting net-like structure were located at the boundaries of the BCC phases [19,20]. However, the distribution of the σ phase and FCC structures observed was significantly different than those described in these previous studies. Interestingly, the microstructure presented a significant gradient distribution as the volume fraction of the σ structure increased. The mixed structures presented neither a net-shaped nor a dendritic form. This suggested that the FCC phase primarily formed during sintering, while the σ phase precipitated. The σ phase is an intermetallic precipitate, which dispersed and distributed in the FCC matrix.

**Figure 4.** SEM image and EDX maps of Fe, Co, Cr, Ni, Mo, and C of marked area in Figure 3a of HEA after spark plasma sintering (SPS) at 1150 ◦C and 30 MPa.

**Figure 5.** XRD pattern of different areas in the sample of HEA after SPS at 1150 ◦C and 30 MPa, (**a**) the macrostrucutre of the 30 MPa sintered sample, and (**b**) XRD pattern of the corresponding position in (**a**).

**Table 1.** Chemical composition of FeCoCrNiMo after spark plasma sintering (SPS) at 1150 ◦C and 30 MPa. FCC—face-centered cubic.


In addition, there is no segregation of C in both the FCC and σ phases, which excludes the possibility of carbide formation (C diffusion of graphite die during SPS). There are several places where C element is segregated in the micropore or porosities; it is certain that these micropore or porosities are inevitable in sintering process, but have nothing to do with and are not the result of the phase transition from FCC to σ phase.

In Table 1, the Mo content of the σ phase is only 2.6%, whereas that of the FCC phase is approximately 6.7%. As the Mo content of nominal composition is 4.34 at.%, the FCC phase was enriched in Mo. This was the result of the good inter-solubility of Cr, Fe, Co, and Ni, while the solubility of Mo in the other elements was poor. Therefore, during the dissolution process of the solid solution, Mo, as a solute element, was repelled towards the FCC phase and redistributed along with Cr, while Cr entered the solid solution. The precipitated σ phase became rich in Cr and poor in Mo.

According to the (CoFeNi)–Cr–Mo pseudo ternary phase diagram at 900 ◦C and the (CoFeCrNi)–Mo pseudo-binary phase diagram in the work of [21], the microstructures of the dual phase (FCC and σ phase) structure (marked area in Figure 3) are examined and presented in Figure 6. The P1 area exhibits a single-phase polycrystalline structure, which can be identified in Figure 3. With the proceeding to the center of the specimen, the microstructures evolve to a dendritic structure, and Mo starts to segregate in the interdendritic areas and gradually shows a mixed two-phase structure. The TEM images and the selected-area diffraction (SAED) pattern in Figure 6a,c, showing a large precipitate embodied in the FCC matrix, clearly confirm that the precipitate consists of a mixture of the FCC and the σ phase, which are enriched with Cr and Mo elements (Table 1). It is noted that the FCC phase particles were precipitated out from the σ phase particle during cooling from a high temperature by solid-phase transformation. Thus, this complex structure can be simply referred to as a "double precipitation" during cooling, which can be explained by the pseudo binary diagram.

**Figure 6.** Transmission electron microscopy (TEM) image (**a**), (**b**) of the marked area in Figure 3a of FeCoCrNiMo HEA after sintering at 1150 ◦C and 30 MPa, and the selected-area diffraction (SAED) pattern (**c**) of the marked area in (**b**).

The phases formed in the SPS-processed HEA are essentially metastable—they are indeed formed solid-state phases during SPS, which are then kept at the ambient temperature because of the sluggish diffusion kinetics of HEAs. The TEM images are shown in Figure 6b. It was identified that an FCC phase re-precipitated in the σ phase. It can be concluded that the pseudo phase diagram quite successfully predicts all the structural features in the current alloy system. However, the solubility of Mo in the FCC matrix at SPS temperatures should correspond to the "FCC + σ" zone in the pseudo phase diagram in the work of [21], so there was no Mo-rich μ phase observed. Also, FCC twinning has also been observed in Figure 6b, which may be the result of phase transformation induction. This twinning structure will have a beneficial effect on the mechanical properties of HEA [22]. It is noted that the FCC matrix has intrinsically low stacking fault energy [23] and alloying of Mo could further reduce such energy, all of which promotes both twinning and widely dissociated and reactive dislocations.

According to the above analysis, the precipitates were identified as the σ phases that were transformed from the initial FCC structure during SPS. As can be seen in Figure 2, the thickness of the σ phase precipitates decreased gradually as the SPS pressure increased from 30 to 40 MPa. In addition, the FCC to σ phase transition temperature of FeCoCrNiMo HEA was about 1260 ◦C, which was measured by DSC (Figure 7). The first exothermic peak is at about 960 ◦C, which represents the beginning of the sintering reaction and the formation of the FCC phase. However, phase transition occurs at 1150 ◦C under the SPS condition, which is attributed to the effect of sintering pressure in the SPS process. When pressure existed, the phase transformation became easier, which reduced the transformation temperature and shortened the time of transformation. Thus, the sintering pressure significantly reduced the phase transformation temperature. Moreover, the σ phase is an ordered tetragonal structure as an intermetallic compound. From the point of view of atomic stacking density, the change of phase-volume is sensitive to pressure, and the larger the pressure, the smaller the trend of phase volume grown up. This theory has been proven in TiAl-based alloys, which are also intermetallic compounds. Under HIP (Hot Isostatic Pressing) conditions, the phase volume is sensitive to sintering pressure [14]. This explained why when sintering pressure rose to 35 and 40 MPa, the volume fraction of the transformed σ structure decreased. Accordingly, the gradient distribution of microstructure is caused by the sintering pressure and by adjusting the sintering pressure, the volume fraction of phase transformation can be changed to obtain the required gradient materials.

**Figure 7.** Differential scanning calorimeter (DSC) of the FeCoCrNiMo HEA sample processed by SPS.

#### *3.3. Mechanical Properties*

Bending and microhardness tests were used to analyze the effect of the distribution of the FCC and σ phases on the mechanical properties of HEA.

Figure 8 shows the bending curve of the HEA sample SPS processed under different pressures. According to Figure 8, when the sintering pressure is 40 MPa, the sample has the highest transverse strength of 1004 MPa and the highest fracture strain of 2.3%. With the decrease of sintering pressure, the strength and strain also decrease. When the sintering pressure is 30 MPa, the bending strength is 779 MPa. Obviously, the transverse strength decreases with the increase of the volume fraction of the σ phase. This is because the σ phase is an intermetallic compound with intrinsic brittleness. In contrast, the FCC phase has higher toughness. Comprehensively, when the volume fraction of the σ phase is small, the transverse strength of the FCC phase is more reflected. When sintering pressure is 30 MPa, the σ phase increases significantly, which leads to the decrease of transverse strength.

**Figure 8.** Bending curve of the HEA sample SPS processed under different pressure.

In addition, the results of the bending tests demonstrated that the SPS sample, which contained a larger volume fraction of the σ phase, could resist greater bending stress. This observation agreed with the work of Tsai et al., who found that the hardness of the Al0.3CrFe1.5MnNi0.5 HEA nearly tripled after the σ phase formed [24], as the σ phase is a very hard phase [25]. Although, because of the different composition, it is different compared with the σ phase in FeCoCrNi–Mo HEA, other chemical elements may result in different properties. However, the σ phase is a typical ordered, tetragonal structure. Whatever the elements formed, its essence is a hard phase compared with the FCC phase.

The sample exhibited a gradual microhardness change similar to the gradually distribution of the σ phase. The center of all the three specimens consisted of a single FCC phases structure with a constant microhardness value of approximately 210 HV0.2. As an increasing amount of the new structure precipitated, the hardness gradually increased as the distance from the surface decreased, and the hardness increased to the maximum value at 50 μm from the surface, as shown in Figure 9. The highest hardness could be attributed to the effect of the large volume fraction of hard σ phases. It also illustrates that the sintering pressure directly affects the gradient distribution of the FCC to σ phase transformation in the thickness direction. When the sintering pressure is 30 MPa, the hardness distribution shows a steep trend owing to the transformation of the FCC phase to the σ phase. However, at 40 MPa, the overall hardness distribution tends to be flat owing to the thinner phase transition area.

**Figure 9.** Microhardness distribution with depth from edge of the FeCoCrNiMo HEA after sintering at 1150 ◦C and 30, 35, and 40 MPa.

Moreover, from the indentation morphology of different depths, taking the sintering of 30 MPa as an example, the volume fraction of σ phase at the edge is larger, and the indentation depth is shallower, as shown in Figure 10. As the depth increases to the middle, the volume fraction of the σ phase decreases, the FCC phase increases, and the indentation area and depth increase. This further proves that the σ phase is a hard and brittle intermetallic phase, and the FCC phase is softer than the σ phase.

**Figure 10.** Indentation morphology of different depths from edge (**a**) 50 μm, (**b**) 350 μm, (**c**) 650 μm, and (**d**) indent profile of the FeCoCrNiMo HEA after sintering at 1150 ◦C and 30 MPa.

#### **4. Conclusions**

The mechanical properties of the FeCoCrNiMo HEA exhibited a significantly gradual change because of gradient distribution of the mixed FCC and σ phase structure. This gradient distribution was mainly assumed to be caused by the sintering pressure, which was under the specific sintering temperature of 1150 ◦C. It was found that the alloying of Mo into the CoCrFeNi HEA system generated the precipitation of the hard and brittle σ phase in the FCC matrix. The hard, Cr-rich σ phase was homogeneously distributed throughout the FCC matrix, and the volume fraction of the σ structure increased from the center to the surface of the HEA sample, like it would for gradient materials. The volume fraction of transformed σ structure can easily be adjusted by sintering pressure. The sintering pressure directly affects the gradient distribution of the FCC to the σ phase transformation in the thickness direction of the HEA samples. The implication for control of properties via changing the phase balance in HEAs will provide a strong technical base for the tool and dies of a novel gradient material.

**Author Contributions:** Conceptualization, W.Z. and Y.L.; Methodology, W.Z.; Validation, Y.P.; Formal Analysis, W.Z. and Y.P.; Investigation, M.Z. and L.W.; Resources, Y.L. and S.H.; Data Curation, M.Z.; Writing-Original Draft Preparation, M.Z.; Writing-Review & Editing, W.Z. and Y.P.; Project Administration, W.Z. and Y.L.; Funding Acquisition, Y.L., S.H. and Y.H.

**Funding:** This research was funded by the project of Innovation and Entrepreneur Team Introduced by Guangdong Province, grant number 201301G0105337290, and the Special Funds for Future Industrial Development of Shenzhen, grant number HKHTZD20140702020004.

**Acknowledgments:** The authors wish to acknowledge the financial support of State Key Laboratory of Powder Metallurgy (CSU 621011808), XUCHANG Fellowship Program [grant number: XW2017-40], the project of Innovation and Entrepreneur Team Introduced by Guangdong Province (201301G0105337290), and the Special Funds for Future Industrial Development of Shenzhen (No. HKHTZD20140702020004).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Microstructure and Properties of High-Entropy Al***x***CoCrFe2.7MoNi Alloy Coatings Prepared by Laser Cladding**

#### **Minghong Sha** †**, Chuntang Jia** †**, Jun Qiao, Wenqiang Feng, Xingang Ai, Yu-An Jing, Minggang Shen and Shengli Li \***

School of Materials and Metallurgy, University of Science and Technology Liaoning, Anshan 114051, China; s13591200297@163.com (M.S.); jia19950608@163.com (C.J.); juncourses@163.com (J.Q.);

wqfeng14b@imr.ac.cn (W.F.); aixingang@126.com (X.A.); jyachina@163.com (Y.-A.J.); lnassmg@163.com (M.S.)

**\*** Correspondence: hongsmh\_116@163.com; Tel.: +86-156-4125-9199

† These authors contributed equally to this work.

Received: 25 October 2019; Accepted: 17 November 2019; Published: 20 November 2019

**Abstract:** High-entropy Al*x*CoCrFe2.7MoNi (*x* = 0, 0.5, 1.0, 1.5, 2.0) alloy coatings were prepared on pure iron by laser cladding. The effects of Al content on the microstructure, hardness, wear resistance and corrosion resistance of the coatings were studied. The results showed that the crystal phases of the Al*x*CoCrFe2.7MoNi coatings changed from Mo-rich BCC1 + FCC to (Al, Ni)-rich BCC2 + Mo-rich BCC1 when *x* increased from 0 to 0.5, and the phase changed to an (Al, Ni)-rich BCC2 + (Mo, Cr)-rich σ phase as *x* increased further. The hardness of the coatings increased as the Al content increased. The Al2.0CoCrFe2.7MoNi coating exhibit best wear resistance. Addition of Al increased the corrosion potential in a 3.5 wt.% NaCl solution, and the coating with *x* = 1.0 exhibited the highest corrosion resistance.

**Keywords:** high-entropy alloys; laser cladding; corrosion resistance; wear resistance; microstructure

#### **1. Introduction**

Yeh et al. [1,2] put forward the concept of a high-entropy alloy (HEA) in 2004, and changed the traditional concept of alloy design. An HEA is defined as an alloy with a configurational entropy larger than 1.5R in the random solution state [3,4], where R is the gas constant. Owing to their special composition and structure, an HEA exhibits high phase stability, wear resistance, and corrosion resistance [5–12]. Zhang et al. [13] prepared an HEA coating of FeCoCrAlNi on 304 stainless steel by laser cladding. The results showed that the coating exhibits better corrosion resistance and pitting resistance than uncoated 304 stainless steel in a 3.5 wt.% NaCl solution. Niu et al. [14] studied the effect of Al content in an Al*x*FeCoCrNiCu (*x* = 0.25, 0.5, 1.0) HEA on its corrosion resistance in a 1 mol/L H2SO4 solution and a 1 mol/L HCl solution, respectively. The corrosion resistance and pitting resistance in the 1 mol/L H2SO4 solution increased when the Al content was less than 0.5, while they decreased when the Al content reached 1.0. Kao et al. [15] studied the corrosion resistance of an Al*x*CoCrFeNi HEA and found that the corrosion potential (*E*corr) and corrosion current (*I*corr) are independent of Al content. Therefore, the effects of Al content on corrosion resistance of HEAs are still not fully understood. An AlCoCrFeNi HEA has been extensively studied for its uncomplicated FCC and BCC phases [16–19]. Some researchers added Ti, Nb, and other elements to the alloy to obtain the desired microstructures, hardness and wear resistance [20–22]. Mo has small thermal expansion coefficient, high strength at high temperatures, high hardness, strong corrosion resistance and high thermal conductivity [23]. It is shown that the addition of Mo increases the strength of AlCrFeNiMo*x* and CoCrFeNiMo*<sup>x</sup>* HEAs due to the formation of the sigma (σ) phase [24,25]. The σ phase is a hard, brittle phase commonly found in superalloys and can significantly change the mechanical properties of the alloy [26–28]. Its effect on corrosion resistance has not been reported. The effects of Mo content on the structure and properties of AlCoCrFeNiMo*<sup>x</sup>* HEAs are being investigated in another of our studies. Due to a higher Fe content in the coating, it is helpful to improve the coating's bond with a steel substrate; Al*x*CoCrFe2.7NiMo HEAs were determined as the coating materials to be studied in our current work.

As a new technology, laser cladding has many advantages over traditional cladding technologies, providing coatings with minimum dilution, minimum deformation, and high surface quality. The effects of Al content on microstructure, hardness, wear resistance and corrosion resistance of Al*x*CoCrFe2.7MoNi coatings prepared by laser cladding were evaluated in this study.

#### **2. Materials and Methods**

Pure iron was selected as the base material in order to eliminate the effects of other elements. Its high purity allows accurate analysis and characterization of the structure and properties of HEAs. Table 1 shows the chemical composition of the base material measured by chemical analysis.

The pure metal powders used in the experiments are the atomized powders produced by BGRIMM (Beijing, China). Pure powders of Al, Co, Cr, Fe, Ni, and Mo (>99.9%, wt.%) with an average particle size of 75 μm were used as raw cladding materials. The powders were weighed and mixed according to the proportions listed in Table 2 (Fe2.7 was achieved by appropriate laser parameters determined through multiple test attempts to control the dilution ratio of the coating and the base material). Then, the mixed alloy powder was put into a stainless-steel tank and thoroughly dry blended for 5 h in a planetary ball mill with a rotating speed of 300 r/min. After sieving, placed in a vacuum dries oven to prevent oxidation. A mixed powder layer with 1 mm thickness was placed on the base material and radiated by the laser in an argon atmosphere. Single-pass laser cladding was used to deposit coatings of Al*x*CoCrFe2.7MoNi at 1350 W, 980 nm wavelength, 20 mm/s scanning rate, and 3 mm spot diameter (Laserline LDF 4000, Laserline GmbH, Mülheim-Kärlich, Germany). The radiated samples were then annealed at 900 ◦C for 5 h to relieve thermal stress and prevent microcrack formation. The structures of the samples were analyzed using X-ray diffraction (XRD, PANalytical X-pert Power, Malvern Panalytical Ltd., Worcestershire, UK) with a line detector (X'Celerator) at 2θ ranging from 15◦ to 90◦ in 0.065◦ increments with Cu Kα radiation. High Score Plus software and PDF-2004 database (JCPDS, Newtown Square, PA, USA) were used to analyze the diffraction pattern. The specimens were eroded in aqua regia for 5–10 s, and the morphologies and compositions of the coatings were analyzed using a scanning electron microscope (SEM, Hitachi S-3400N, Hitachi, Ltd., Tokyo, Japan) with an energy dispersive spectrometer (EDS, TEAM PEGASUS2040), and with a transmission electron microscope (TEM, FEI Talos F200, FEI Co. Ltd, Hillsboro, OR, USA) with an EDS (FEI Super X). The size of the investigated area used for the measurements of the overall compositions of coatings in Table 3 is 1300 μm × 265 μm. Microhardness was measured from the bond zone to the coating surface using a microhardness tester (Qness Q10A, Qness GmbH, Golling, Austria) with a 9.81 N loading force and 15 s loading time. The wear resistance was tested with friction and wear test equipment (UMT TriboLab, Bruker Corporation, Billerica, MA, USA) with a pair of ceramic balls. A 13 N normal load, 100 mm/s reciprocating speed, a 10 mm reciprocating straight line distance and 1800 mm total wear distance were used in the wear tests. The weight of the samples before and after wear tests was weighed with a balance (0.01 mg precision). The *E*corr and *I*corr were measured with an electrochemical workstation (Autolab PGSTAT302N) from −1.2 to 1.2 V and 1.0 mV/s scanning speed in a 3.5 wt.% NaCl aqueous solution. The *E*corr and *I*corr of the coatings were obtained by Tafel linear extrapolation. A platinum electrode, saturated AgCl electrode and the specimen were used as the auxiliary, reference, and working, respectively. The chemical valence states of metal elements in passive films formed on the surfaces of the Al1.0 and Al2.0 coatings were measured using X-ray photoelectron spectroscopy (XPS, Escalab 250Xi, Thermo Fisher Scientific, Waltham, MA, USA) with monochromatic Al Kα excitation.

**Table 1.** Pure iron content for matrix (wt.%).


**Table 2.** Composition of the mixed powder (wt.%).


#### **3. Experimental Results**

#### *3.1. Crystal Structure*

X-ray diffraction patterns from the Al*x*CoCrFe2.7NiMo coatings are shown in Figure 1. The coatings are mainly composed of simple solid solutions and intermetallic compounds. The phase structure is composed of both BCC and FCC solid solutions when *x* = 0, while BCC1 and BCC2 solid solutions appear and the peak value is more intense when *x* = 0.5. Tiny Bragg peaks corresponding to the σ and BCC phases are visible in the XRD pattern when *x* = 1.0, and no new phase appears in the XRD pattern as the Mo content increases (*x* = 1.5 and 2.0).

**Figure 1.** XRD pattern from Al*x*CoCrFe2.7MoNi coatings.

#### *3.2. Microstructure*

SEM images of the microstructure of the Al*x*CoCrFe2.7MoNi (*x* = 0, 0.5, 1.0, 1.5, 2.0) coatings are shown in Figure 2a–c, e, and g. In view of the fineness of microstructure and the limitations of the SEM, TEM images of the microstructures of Al1.0, Al1.5, and Al2.0 HEAs are presented in Figure 2d, f and h. The target composition and actual composition of the coatings measured by EDS are listed in Table 3. The chemical compositions in different micro-regions of Al*x*CoCrFe2.7MoNi are shown in Table 4. A few precipitates containing Fe and Cr appear in region A of the coating without Al, as shown in Figure 2a. Figure 2b shows that the Al0.5 coating consists of dendrites. Figure 2c,e,g shows that Al1.0, Al2.0, and Al3.0 alloys have fine microstructures, respectively. The light and dark phases appear in Al1.0, indicated by D and C, respectively, as shown in Figure 2d. The EDS results and analysis of the diffraction spots show that the C phase is a Mo-rich σ phase, and the D phase is an (Al, Ni)-rich BCC2 phase. Figure 2f shows that two kinds of dark areas (granules and sheets) and one bright area can be seen in the Al1.5 coating, indicated by C, C1, and D, respectively. An analysis of the diffraction spots show that the C and C1 phases belong to the (Mo, Cr)-rich σ phase, while the D phase is an (Al, Ni)-rich BCC2 phase. The dark strip disappears as Mo content increases, as shown in Figure 2h, and the microstructure is composed of a granular (Mo, Cr)-rich σ phase and (Al, Ni)-rich BCC2 phase.

**Figure 2.** *Cont*.

**Figure 2.** Microstructures of the Al*x*CoCrFe2.7MoNi coatings. (**a**–**c**,**e**,**g**) show SEM images; (**d**,**f**,**h**) show TEM images.


**Table 3.** Composition of coatings measured by EDS (at.%).

**Table 4.** Composition of coatings micro-regions (at.%).


#### *3.3. Microhardness*

The microhardness measurements from different positions in the Al*x*CoCrFe2.7MoNi coatings are shown in Figure 3. The CoCrFe2.7NiMo alloy has the lowest average hardness (272 HV), which can be attributed to generation of the FCC phase. The microhardness increases as Al content increases, and Al2.0CoCrFe2.7NiMo has the highest average hardness (1142 HV). Hardness test results show that the formation of the BCC2 phase increases the hardness of the coating.

**Figure 3.** Microhardness of Al*x*CoCrFe2.7MoNi coatings.

#### *3.4. Wear Resistance*

Wear in a material is related to its structure and external environment. The wear resistance of samples Al1.0, Al1.5, and Al2.0 was analyzed in this paper. The morphology of worn Al*x*CoCrFe2.7MoNi (*x* = 1.0, 1.5, 2.0) coatings is shown in Figure 4a1–c1 are enlarged partial details of Figure 4a–c, respectively. There is a convex scaly plastic deformation layer on the friction surface of the Al1.0 sample, as shown in Figure 4a,a1, which resulted from repeated grinding during the wear test. An oxide developed at the junction of the flaky furrow and the scaly deformed layer because of severe friction and high temperature during the wear test. The wear mechanism is mainly adhesive wear and oxidative wear. The wear of the Al1.5 sample is with a few flake furrows, in which oxides were found, as shown in Figure 4b,b1, indicating the wear occurs via oxidation, slight adhesion wear and slight abrasive wear. Sample Al2.0 also exhibits a scaly plastic deformation layer with oxides and a flaky furrow in Figure 4c,c1. The wear mechanism in sample Al2.0 is adhesive wear and oxidative wear. The measured weight losses from the coatings due to wear are listed in Table 5; sample Al2.0 exhibited the least wear of 0.1 mg.

**Figure 4.** *Cont*.

**Figure 4.** Wear morphology of the Al*x*CoCrFe2.7MoNi (*x* = 1.0, 1.5, 2.0) coatings; (**a**,**a1**) *x* = 1.0; (**b**,**b1**) *x* = 1.5; (**c**,**c1**) *x* = 2.0.



#### *3.5. Corrosion Resistance*

Potentiodynamic polarization curves of the Al*x*CoCrFe2.7MoNi coatings in the 3.5 wt.% NaCl solution are shown in Figure 5. The *Ecorr* and *Icorr* of the coatings were obtained by Tafel linear extrapolation, as shown in Table 6. These results show that, except for sample Al1.0, the self-corrosion potential of the other coatings increases as the Al content increases. The self-corrosion current density with sample Al1.0 is the least, while the self-corrosion current density with samples Al1.5 and Al2.0 are slightly larger. Sample Al0.0 exhibits the lowest self-corrosion potential and higher self-corrosion current density, indicating that it has the greatest corrosion tendency, highest corrosion rate and worst corrosion resistance. Figure 6 shows XPS results from passive films on the Al*x*CoCrFe2.7NiMo (*x* = 1.0, 2.0) coatings after the corrosion experiments in the 3.5 wt.% NaCl solution. The composition of the passive film is Al2O3, CoO, Co2O3, Cr2O3, Fe2O3, MoO3, and NiO when *x* = 1.0, while the passive film is primarily composed of Al2O3, CoO, Cr3O4, Fe2O3, MoO3, and NiO when *x* = 2.0. Al2O3, Cr2O3, CoO, MoO3, and NiO were detected on the surfaces of all coatings, which can provide certain protection in a corrosive environment. The compositions (in relative at.%) of the passive films determined from XPS measurements are summarized in Figure 7. This shows that the content of Al2O3 is higher than that of other metal oxides, and the relative contents of Ni, Co and Fe oxides decrease as the Al content increases. This is primarily due to the fact that Al is active and oxidizes easier than the other elements listed here.

**Figure 5.** Polarization curves of Al*x*CoCrFe2.7MoNi coatings in a 3.5 wt.% NaCl solution.

**Figure 6.** *Cont*.

**Figure 6.** XPS spectra from the passive film formed on the surfaces of the coatings. (**a**,**c**,**e**,**g**,**i**,**k**) *x* = 1.0; (**b**,**d**,**f**,**h**,**j**,**l**) *x* = 2.0.

**Figure 7.** Composition (in relative at.%) of the surface of Al1.0 and Al2.0 coatings determined from XPS measurements. (**a**) *x* = 1.0; (**b**) *x* = 2.0.

**Table 6.** Measured electrical behavior during corrosion of the Al*x*CoCrFe2.7MoNi coatings in a 3.5 wt.% NaCl solution.


#### **4. Discussion**

#### *4.1. Microstructure and Phase*

Results in many studies have shown that the CoCrFeNi alloy consists of a single FCC phase [26,29,30]. Our XRD analysis results show that the CoCrFe2.7NiMo alloy consists of FCC and BCC phases (as shown in Figure 1), which indicates that adding Mo causes formation of the BCC1 phase. This conclusion is consistent with Wu's research [29]. Table 7 shows the enthalpy of mixing between elements. The enthalpy of mixing between Fe and Cr is relatively higher, and it is difficult to form a stable solid solution. Therefore, it is considered that the high melting point of the mixed powders leads to incomplete melting while depositing an Al0 alloy. The enthalpy of mixing between Al and other elements is lower than that between Mo and other elements, and results from many studies show that adding Al can cause the FCC phase to change to the BCC phase in HEAs [30–32]. Therefore, the FCC phase disappears in the Al1.0 alloy and the peak in the XRD pattern becomes more intense due to the increased BCC phase content (BCC1 phase and (Al, Ni)-rich BCC2). Elements such as Co, Ni and Al enrich to form a disordered (Al, Ni)-rich BCC2 phase (random solid solution) as the Al content increases further, as shown in Figure 8a. Meanwhile, Mo dissolves with other elements, e.g., Ni, to form an ordered σ phase (intermetallic), as shown in Figure 8b. A random solid solution tends to have a large configuration entropy due to random mixing of its various elements [1,2]. According to the Gibbs free energy formula *G* = *H* − *TS* (*G* is the Gibbs free energy, *H* is the enthalpy of mixing, *T* is the temperature, and *S* is the configuration entropy). *G* is negative when *S* > (*H*/*T*) and the solid solution phase forms easily, and the enthalpy of mixing between Al and other elements is lower than that between Mo and other elements, indicating that other elements dissolve more easily in Al than Mo, forming a solid solution, as shown in Table 7 (the data are derived from the literature [33]). Therefore, the addition of Al generates a large amount of the (Al, Ni)-rich BCC phase, which forces the Mo-containing BCC1 phase to transform into a (Mo, Cr)-rich σ phase. In conclusion, for the alloys studied herein, Mo allows a BCC structure to form more easily when Al is absent in the coating, while

both Al and Mo easily form a BCC structure when a small amount of Al is added. Al is the primary driver of BCC formation, and addition of Mo tends to cause formation of the σ phase when more Al is added.


**Table 7.** Mixing enthalpy values for different element pairs, data from [33].

**Figure 8.** High resolution TEM images of the σ and BCC phases in the AlCoCrFe2.7NiMo coating. (**a**) shows σ phase; (**b**) shows BCC phase.

#### *4.2. Hardness and Wear Resistance*

Hardness and wear resistance of alloys are closely related to their microstructures. Figure 9 shows the relationship between the volume fraction of the phases in the coating, obtained by XRD and the hardness of the coating. The presence of the FCC phase minimizes the hardness of the Al0 alloy, and formation of (Al, Ni)-rich BCC2 causes the hardness of the Al0.5 alloy to increase. The volume fraction of BCC2 phase has a greater influence on the hardness and wear resistance of the coatings than does that of σ phase. Greater content of the (Al, Ni)-rich BCC2 phase correlates with higher hardness and wear resistance as the Al content increases. However, the presence of the flaky σ phase causes large flaky exfoliation on the Al1.0 coating in wear resistance experiments, which can be attributed to the brittleness of the σ phase. It is worth noting that the σ phase changes from sheet-like to granular as the Al content increases further, which plays a role of dispersion strengthening in the alloy. Furthermore, plastic deformation can be observed in the wear microstructure diagram of the Al1.0 coating, while no plastic deformation can be observed in the Al1.5 and Al2.0 coatings. Oxides appear, which indicates that oxidation wear is the primary wear mechanism, and dispersion strengthening can significantly increase the coating's wear resistance. In summary, introduction of the (Al, Ni)-rich BCC2 phase increases the hardness and wear resistance of the coating. Adding Al reduces the size of the σ phase, which also increases the hardness and wear resistance of the coating.

**Figure 9.** Relationship between the volume fraction of the phases and the average hardness of the coatings.

#### *4.3. Corrosion Resistance*

Figure 10 shows a comparison of *Ecorr* and *Icorr* in our Al*x*CoCrFe2.7NiMo coatings and Al*x*CoCrFeNi alloys from the literature [34]. One can see that the addition Mo leads to increased corrosion resistance. This may be attributed to the fact that Mo is prone to produce dense passivation films. In addition, the formation of the σ phase increases the corrosion resistance of the coating, but its dispersion distribution reduces the corrosion resistance of the coating. It is generally believed that Al2O3 can effectively resist chloride ion corrosion because of its compact structure. However, increasing the Al content further increases the differences in the content of different metal oxides in the passivation film, which will reduce the coating's corrosion resistance. To date, a definitive understanding of the effect of oxide interaction in passive film on the corrosion resistance of HEAs is yet to emerge. However, the data presented in this paper can provide more important information for other researchers in this emerging field.

**Figure 10.** Comparison of the corrosion properties of Al*x*CoCrFe2.7NiMo HEAs and Al*x*CoCrFeNi HEAs in a 3.5 wt.% NaCl solution.

#### **5. Conclusions**

Al*x*CoCrFe2.7MoNi coatings were prepared on pure iron via laser cladding, and their microstructure, hardness, wear resistance and corrosion resistance were studied.

The increase of Al content promotes the releasing of Mo from Mo-rich BCC1 phase, and the formation of the (Mo, Cr)-rich σ phase. The increase of Al content causes the increase in volume fraction of BCC2 phase, and correspondingly the increase of hardness and wear resistance. The formation of the strip-shaped σ phase contributes to the improvement of the corrosion resistance of the coating, but the dispersed distribution of the σ phase deteriorates corrosion resistance.

**Author Contributions:** Conceptualization, M.S. (Minghong Sha), C.J. and S.L.; Funding acquisition, M.S. (Minghong Sha) and S.L.; Formal analysis, C.J.; Investigation, C.J.; Resources, W.F., X.A., Y.-A.J. and M.S. (Minggang Shen); Writing - original draft, C.J.; Writing - review & editing, C.J., M.S. (Minghong Sha) and J.Q.

**Funding:** This work was financially supported by the National Key Research and Development Program of China (NO.2017YFB0304201), National Natural Science Foundation of China (NO.51774179), Natural Science Foundation of Liaoning Province (NO.20180550546, NO.20170540460), and the Innovation Team Project of Liaoning Education Department (NO.LT2016003).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Slurry Erosion Behavior of Al***x***CoCrFeNiTi0.5 High-Entropy Alloy Coatings Fabricated by Laser Cladding**

#### **Jianhua Zhao 1,2, Aibin Ma 1,\*, Xiulin Ji 2,\*, Jinghua Jiang <sup>1</sup> and Yayun Bao <sup>2</sup>**


Received: 7 January 2018; Accepted: 5 February 2018; Published: 11 February 2018

**Abstract:** High-entropy alloys (HEAs) have gained extensive attention due to their excellent properties and the related scientific value in the last decade. In this work, Al*x*CoCrFeNiTi0.5 HEA coatings (*x*: molar ratio, *x* = 1.0, 1.5, 2.0, and 2.5) were fabricated on Q345 steel substrate by laser-cladding process to develop a practical protection technology for fluid machines. The effect of Al content on their phase evolution, microstructure, and slurry erosion performance of the HEA coatings was studied. The Al*x*CoCrFeNiTi0.5 HEA coatings are composed of simple face-centered cubic (FCC), body-centered cubic (BCC) and their mixture phase. Slurry erosion tests were conducted on the HEA coatings with a constant velocity of 10.08 m/s and 16–40 meshs and particles at impingement angles of 15, 30, 45, 60 and 90 degrees. The effect of three parameters, namely impingement angle, sand concentration and erosion time, on the slurry erosion behavior of Al*x*CoCrFeNiTi0.5 HEA coatings was investigated. Experimental results show AlCoCrFeNiTi0.5 HEA coating follows a ductile erosion mode and a mixed mode (neither ductile nor brittle) for Al1.5CoCrFeNiTi0.5 HEA coating, while Al2.0CoCrFeNiTi0.5 and Al2.5CoCrFeNiTi0.5 HEA coatings mainly exhibit brittle erosion mode. AlCoCrFeNiTi0.5 HEA coating has good erosion resistance at all investigated impingement angles due to its high hardness, good plasticity, and low stacking fault energy (SFE).

**Keywords:** high-entropy alloy; laser cladding; microstructure; slurry erosion

#### **1. Introduction**

Slurry erosion (SE) is a serious concern for hydraulic turbines and other fluid machines due to silt entrained in water flow, especially in the Yellow River regions of China. Slurry erosion results in the surface degradation of flow components of hydroturbine equipment and reduces all efficiency [1,2]. Hydraulic turbine equipment generally made of mild steel, white cast iron or stainless steel. However, these materials are considerably less resistant to erosive wear. It is important to develop new erosion-resistant materials. Recently, high-entropy alloys (HEAs) have attracted extensive attention due to their versatile combinations including high strength and hardness, good thermal stability, excellent corrosion and wear resistance [3–6]. These characteristics make them suitable candidates for structural and functional materials. The main method of preparing high-entropy alloys is vacuum arc melting and then casting [7,8]. As the formation of simple solid solution phase requires high cooling rate, the shape and size of bulk ingots prepared by arc-melting technique are limited. Meanwhile, this preparation method causes high production cost due to many expensive metals such as Ni, Co, and Cr being contained in HEAs. Therefore, some researchers have been turning to explore the effective and economical HEA coating on the low-cost metallic substrate.

Compared to the other processing techniques such as magnetron sputtering, electrochemical deposition, and plasma arc cladding, laser cladding can be used to deposit coatings with thickness more than 1 mm, which is more beneficial for engineering applications. In addition, the coatings can be deposited in a few steps, which eliminates the influence of the substrate and allows gradual composition and property changes through the coating thickness [9]. These favorable advantages have made the laser-cladding alloy attractive among surface modification technologies. Huang et al. [10] prepared TiVCrAlSi HEA coatings on Ti-6Al-4V substrate by laser cladding and investigated the dry sliding wear behavior. The combination of the hard (Ti,V)5Si3 phase and relatively ductile and tough BCC matrix improved the sliding wear resistance. Yue et al. [11] studied the solidification behavior in laser cladding of AlCoCrCuFeNi high-entropy alloy on magnesium substrates using the Kurz-Giovanola-Trivedi and the Gaümann models. Except for some Cu rejected into the Mg melt, no serious dilution of the HEA composition occurred in the top layer of the coating. This is considered to be important because any dilution of the HEA composition with Mg would likely decrease the corrosion resistance of the HEA. Kunce et al. [12] produced the AlCoCrFeNi high-entropy thin-walled samples using the laser engineered net shaping (LENS) technology. The effect of the cooling rate during solidification on the microstructure of the alloy was studied through different laser-scanning rates. It was found that with an increasing in the laser-scanning rate during the solidification process, the average grain size of the alloy decreased. Vickers microhardness increases with the decrease of the average grain size. AlCoCrFeNiTi alloy system has been investigated in bulk state. Zhou et al. [13] studied the microstructure and strengthening mechanism of theAlCoCrFeNiTi0.5 alloy. For AlCoCrFeNiTi0.5 alloy, its super-high strength and good plasticity were attributed to its microstructure of intrinsic strong body-centered cubic solid solution, and effective multiple strengthening mechanisms such as solid solution strengthening, precipitation strengthening, and nano-composite strengthening effects, etc. Jiao et al. [14] studied the superior mechanical properties of AlCoCrFeNiTi*<sup>x</sup>* High-Entropy Alloys upon dynamic loading. They found that the ultimate strength and fracture strain of AlCoCrFeNiTi*<sup>x</sup>* alloys are superior to most of bulk metallic glasses and in situ metallic glass matrix composites. However, the slurry erosion properties of Al*x*CoCrFeNiTi0.5HEA coatings have been rarely studied. In this article, Al*x*CoCrFeNiTi0.5 HEA coatings with different Al content were fabricated by laser cladding. The effects of Al addition on the microstructure and slurry erosion wear behavior were investigated. It is necessary for practical industrial applications.

#### **2. Experimental Procedure**

#### *2.1. Material*

As-received Q345 steel plate with dimensions of 25 × <sup>40</sup> × 10 mm3 was used as the substrate material. The substrate was sandblasted to remove surface contaminants and increased the absorption of laser energy. The Al*x*CoCrFeNiTi0.5 HEA coatings(*x*: molar ratio, denoted as Al1.0, Al1.5, Al2.0 and Al2.5 alloy, respectively) were prepared in this study by laser cladding. The HEA powder used in the experiment had a high purity (more than 99.5%) with a mesh size of 200–300. The mixed powders with the aid of a high-energy ball milling equipment were pre-placed on the steel specimens with a thickness of approximate 300–400 μm using PVA (Shanghai Zengye Industrial Co., Ltd., Shanghai, China) as a binder. The samples were dried in a vacuum oven (Nanjing Huanke Testing Equipment Co., Ltd., Nanjing, China) at 100 ◦C for 1h prior to laser cladding. Laser cladding was carried out using an EFW-300 type YAG pulsed laser (Guangda Laser Technology Co., Ltd., Shenzhen, China), which was equipped with a four-axis numerical control working table. With a series of optimization trial runs, the optimized process parameters were obtained: laser power 2.5 kW, laser beam spot diameter 1.2 mm, scanning velocity 1.5 mm·s<sup>−</sup>1, pulse frequency 20 Hz, pulse width 2.5 ms. High-purity argon gas at a flow rate 5 L·min−<sup>1</sup> was used as ash ielding gas to prevent oxidation during the cladding experiment. A 50% overlap condition for multi-tracking was employed. Three layers high-entropy alloys were deposited under the same processing parameter.

After laser cladding, metallographic and erosion samples with dimensions of 10 × <sup>10</sup> × 10 mm3 were cut by electrical sparkle machining. All samples were ground and polished using abrasive papers down to 1200 grit size to obtain a smooth surface. Samples of microstructural observation were etched with alcohol dilute aqua regia. The top surface microstructure was investigated by scanning electron microscopy (SEM, JSM-6360, JEOL Ltd., Tokyo, Japan). The phase composition of HEA coatings was identified by X-ray diffraction (XRD) with a D/max-2550 diffractometer (Rigaku Corporation, Tokyo, Japan) using Cu Ka radiation. The top surface and cross-section morphology of the HEA coatings was examined by SEM. The microhardness of the polished surface of the HEA coatings was performed by a Vickers hardness tester (HXD-1000TC, Shanghai Optical Instrument Factory, Shanghai, China) at a load of 200 g and 15 s loading time. The average of five points was reported for each sample.

#### *2.2. Slurry Erosion Test*

Slurry erosion test was performed using man-made jet type rig shown in Figure 1. The test rig provides the flexibility to regulate experiment parameters such as impingement angle, sand concentration, working media and impact velocity. The velocity of the slurry jet is controlled by changing the frequency of the motor converter used for driving the pump. The sand concentration is adjusted by changing the rotation speed of driving motor. The test parameters used for the slurry erosion experiment are shown in Table 1. Irregular sand particles in the size range of 16–40 mesh were used for slurry erosion studies. Slurry with a concentration of 10 kg/m<sup>3</sup> and 30 kg/m<sup>3</sup> was prepared using sand obtained from the Yangtze River Delta. The main composition of river sand is SiO2. Each sample was tested for 30 min with a cycle time of 5 min. In this study, the distance is 6 cm between the tested specimen and the ejector nozzle. The erosion samples were cleaned thoroughly with industrial acetone solution to remove contaminants and dried. A precision balance to an accuracy level of 0.1 mg was used to measure the mass loss before and after the test at regular intervals. The erosive wear rate is calculated based on the cumulative mass loss of sample with time, i.e., mg·min<sup>−</sup>1. The eroded surface characterization was examined by SEM. For comparison, 00Cr16Ni5Mo alloy (denoted as Cr16 alloy), widely used to fabricate various hydraulic turbine components, was tested under the same erosion condition.

**Figure 1.** Schematic view of the slurry erosion test rig.


**Table 1.** Parameters employed for slurry erosion testing.

#### **3. Results**

#### *3.1. Microstructure and XRD Analysis*

The laser-cladding process parameters have great influence on the quality, microstructure, and properties of the HEA coatings. With the aforementioned optimized parameter, AlCoCrFeNiTi0.5 HEA coating with few pores could be formed on Q345 substrate as is shown in Figure 2a. It is obvious from Figure 2a that the HEA coating exhibits a typical fish scale lap structure. Figure 2b shows the cross-section SEM image of AlCoCrFeNiTi0.5 single-track coating. The bonding line shows a curved shape, rather than a straight line, indicating a good metallurgical bond between the cladding layer and the substrate, which is favorable for the mechanical performance of the coating.

**Figure 2.** Micro-morphologies of the AlCoCrFeNiTi0.5 HEA (High-entropy alloy) coating for (**a**) surface and (**b**) cross-section single-track coating.

The XRD patterns of Al*x*CoCrFeNiTi0.5 HEA coatings with different Al content are shown in Figure 3. As can be seen, the Al*x*CoCrFeNiTi0.5 HEA coatings exhibit only simple solid solution structure, specifically face-centered cubic (FCC), body-centered cubic (BCC) and their mixture due to the effect of high mixing entropy [15]. A mixture of FCC + BCC crystal structure is observed in Al1.0HEA alloy. The relative intensity of (110)B peak increases and FCC peak disappears in Al1.5HEA alloy. The reflection shift can be partially attributed to the difference of local topologies between FCC and BCC structures [16]. Al has a larger metallic radius, compared with several transition cluster elements such as Co, Cr, Fe, Ni. The increase in the lattice constant with increasing the Al content indicates a corresponding larger lattice-strain effect. To relax the lattice distortion, the metastable FCC phase prefers to transform to a relatively stabilized BCC structure as the Al content in the alloy is increased. Only two BCC phases can be detected in the XRD pattern of the Al2.0 and Al2.5 HEA alloys. Compared with the XRD pattern of the Al1.5 HEA alloy, a minor order BCC peak appears in the Al2.0 and Al2.5 HEA alloys. The order BCC phase in Al*x*CoCrFeNiTi0.5 alloy system has been confirmed as NiAl-based intermetallic (IM) phase [17]. The XRD results show that Al addition exhibits a remarkable influence on the phase composition of the AlxCoCrFeNiTi0.5 HEA alloy.

**Figure 3.** X-ray diffraction patterns of Al*x*CoCrFeNiTi0.5 HEA coatings.

Figure 4 shows Vickers hardness as a function of Al content. The hardness of the Al*x*CoCrFeNiTi0.5 HEA coatings exhibited a strong correlation with their aluminum content and phase structure. This suggests that the formation of a BCC type structure is a dominant factor of hardening, and the increase of the relative amount of BCC phase leads to a large increase in hardness. The larger atomic radius, the transformation of the crystal structure and dispersion of nanocrystallite could be responsible for the increased hardness of the alloys [18]. The microhardness of the HEA coatings in this work can reach 667.3 to 801.1 HV, which is at least 1.8 times that of 00Cr16Ni5Mo alloy (370.5 HV).

**Figure 4.** Vickers hardness of Al*x*CoCrFeNiTi0.5 HEA coatings with different aluminum contents.

Figure 5 presents the SEM images of the Al*x*CoCrFeNiTi0.5HEA top surface layers. The typical microstructure consists of dendritic (DR) and interdendritic (ID) as a result of faster nucleation and solidification. With the addition of Al content, the solidification structure varies from columnar dendrite to non-equiaxed dendrite grain, and finally to equiaxed dendrite grain. The Al1.0 alloy shows the morphology of columnar dendrite and minor non-equiaxed dendrite. When Al content reaches *x* = 1.5, the columnar phase dissolves and single non-equiaxed dendrite appears, which is consistent with the XRD result. The Al2.0 and Al2.5HEA alloys exhibit the same equiaxed dendrite structure. The morphology features indicate the Al2.0 and Al2.5HEA alloys should have similar phase composition and solidification behavior.

**Figure 5.** SEM (scanning electron microscopy) images of the as-laser cladding Al*x*CoCrFeNiTi0.5 HEA coatings: (**a**) Al1.0; (**b**) Al1.5; (**c**) Al2.0; and (**d**) Al2.5.

#### *3.2. Results of Slurry Erosion Tests*

Figure 6 displays the erosion rate of Al*x*CoCrFeNiTi0.5 HEA coatings and Cr16 alloy as a function of impingement angle after erosion for 30 min. The Al1.0 HEA coating and Cr16 alloy showed the maximum erosion rate at 45◦, then the erosion rate decreasing gradually with the impingement angle, which is consistent with the theory of the ductile mode of erosion behavior [19,20]. The erosion rate of the Al1.5 HEA coating approached a maximum at 60◦ and then decreased at 90◦. It showed a mixture mode of erosion behavior. In case of the Al2.0 and Al2.5 HEA coatings, the erosion rate increased monotonically with the impingement angle and arrived at its maximum value at 90◦ impingement angle, which exhibits the brittle mode of erosion behavior [21]. It is evident from Figure 6 that the Al1.0 HEA coating showed smaller erosion rate in comparison with Cr16 alloy at all the investigated impingement angles. The erosion rate of Al1.0 HEA coating is 1.78 times lower than Cr16 alloy at 45◦ impingement angle and 1.68 times lower at 90◦ impingement angle. The reason behind the high erosion resistances of the Al1.0 HEA coating at low impingement angles (from 15◦ to 45◦) is thought to be its higher hardness compared to theCr16 alloy. In slurry erosion wear, the effect of slurry scouring makes the deformation lips or convex bodies easy to be washed away, which results in the more important role of the hardness of target material at a low angle of impingement. At normal impingement angle (90◦), the Al1.0 and Al1.5 HEA coatings still showed significantly lower erosion rate than Cr16 alloy. However, the erosion rate for the Al2.0 and Al2.5 HEA coatings was similar to theCr16 alloy. Closed to normal impingement, ductility and toughness play a more dominant role [22]. Al1.0 HEA alloy, whose yield stress, fracture strength, and plastic strain are as high as 2.26 GPa, 3.14 GPa, and 23.3%, respectively, has the super comprehensive mechanical properties even superior to most of the high-strength alloys such as bulk metallic glasses [23]. It would result in less erosion rate at normal impingement angle. The second possible reason is that Al1.0 and Al1.5 HEA alloys have lower stacking fault energy (SFE) compared to theCr16 alloy. A decrease of SFE results in an increase of work hardening capability of the material, thereby lowering the material removal rate [24]. Using a combination of discrete Fourier transform (DFT) calculation and XRD analysis, Zaddach et al. [25] reported the SFE of the equiatomic NiFeCoCr alloy to be approximate 20 mJ/m2, whereas SFE is of the order of 70–80 mJ/m2 for Cr16 stainless steel [26]. With the Al content increasing, the AlxCoCrFeNiTi0.5HEA coatings showed a transition from the ductile erosion mode to the brittle erosion mode. Limited plasticity likely affects the erosion wear resistance, so the Al2.0 and Al2.5 HEA coatings with higher hardness as well as bigger brittleness can reduce the erosion wear resistance at normal impingement.

**Figure 6.** Variation in erosion rate of Al*x*CoCrFeNiTi0.5 HEA coatings and Cr16 alloy with impact angle at a velocity of 10.08 m/s.

The slurry erosion behavior of Al*x*CoCrFeNiTi0.5 HEA coatings and Cr16 alloy with sand concentration at 45◦ and 90◦ impingement angle is presented in Figure 7. It can be observed that the erosion rates of the test material increased nonlinearly with the increase in the sand concentration, although the percent of the increase is not same for different alloys. Higher sand concentration allows a larger number of sand particles to impact on the surface of the wear specimen, which leads to increase the erosion rate of the material. It is also clear that although sand concentration was increased to 3 times, the erosion rate did not show a similar response. This could be explained by the fact that the shield effect caused by the collisions between the incoming and rebounding particles [27]. Only a portion of particles actually impacts on the target surface while the others lose their way to target surface owing to the interaction between the incoming and rebounding particles. These findings are studied in detail by many researchers [28–30]. For pot-and centrifugal-type impingement experimental rigs, the correlation between erosion rates and sand concentration is highly nonlinear in nature. Some of the investigators have observed the adverse effect of concentration on erosion rate [31,32]. Moreover, as the impingement angle was higher (α = 90◦), the shield zone was higher, and therefore the percent increase of erosion rate for the same material at 45◦ impingement angle is higher than that at 90◦ impingement angle.

**Figure 7.** Variation in erosion rate of Al*x*CoCrFeNiTi0.5 HEA coatings and Cr16 alloy with sand concentration at a velocity of 10.08 m/s and impingement angle of (**a**) 45◦ and (**b**) 90◦.

The slurry erosion behavior of Al*x*CoCrFeNiTi0.5 HEA coatings and Cr16 alloy with time is shown in Figure 8. The results clearly indicate that the erosion rate of specimens was relatively high during the early cycles with some exceptions, and attained a steady state after 15th or 20th min of testing. Comparison with solid particle erosion, slurry erosion has not obvious incubation period. Similar trends were reported by other researchers [2,33]. The relatively stable erosion rate shows erosion time has no effect on the wear mechanism noticeably during the impingement process. The high erosion rate in the initial stage may be attributed to the initial rough surface of the specimens. The existence of micro-peaks and valleys on the surface of the specimens might have resulted in higher material removal rate during the initial period [34]. The second reason may be work hardening owing to the ductility of sample surface subjected to the repeated impact of sand particles, which results in less material removal and thus reduces the erosion rate.

**Figure 8.** Variation in erosion rate of Al*x*CoCrFeNiTi0.5 HEA coatings and Cr16 alloy with erosion time at a velocity of 10.08 m/s and impingement angle of (**a**) 45◦ and (**b**) 90◦.

#### *3.3. Observation of Eroded Surfaces*

The SEM images showing the eroded surfaces of Al*x*CoCrFeNiTi0.5 HEA coatings are given in Figures 9 and 10. Detailed examination of the images indicates that microcuting and mixed cutting and ploughing by irregularly shaped erodent particles are responsible for the material removal at 45◦ impingement angle. Grewal et al. [35] have proposed that with the impact of irregular particles, the primary mode of material removal was a mixture of cutting and ploughing at low impingement angle. Microcutting and mixed cutting-ploughing marks can be seen in Figure 9. At low impingement angle, the normal component of the impact force is too small compared to the tangential one. Material's hardness is a predominant role against these deformation mechanisms. Compared to the hardness of sand particle, which is about 1100 HV, the highest hardness of Al*x*CoCrFeNiTi0.5 HEA coatings is only 801.3 HV.

**Figure 9.** The eroded surface SEM micrographs of Al*x*CoCrFeNiTi0.5 HEA coatings after 30-min slurry erosion: (**a**) Al1.0; (**b**) Al1.5; (**c**) Al2.0; and (**d**) Al2.5 at 10.08 m/s, 1.0 wt. % sand concentration and 45◦ impingement angle.

Figure 10a,b show the SEM images of eroded surfaces of Al1.0 and Al1.5HEA coatings. It can be observed that the eroded surfaces showed the presence of many deformed platelets and indentions. The formation of platelets was mainly through indention of impact sand particles. The material extruded from the craters tends to flow outward and accumulates around the periphery in form of platelets, which are removed by a number of subsequent normal impacts of the sand particles. The highly deformed surface of Al1.0 and Al1.5HEA coatings at 90◦ impingement indicates significant strain hardening. The major erosion mechanism for Al1.0 and Al1.5 HEA coatings is the formation and removal of material in the form of platelets at normal impingement angle. Figure 10c,d show the SEM micrographs of the eroded surface of Al2.0 and Al2.5 HEA coatings at 90◦ impingement angle. The presence of flattened lips indicates the limited ductility of coatings. Some cracks were also observed, which indicates that the coatings also were removed by brittle fracture. For brittle

materials, the energy transfers as a result repeated particle impact results in a fatigue process at or close to normal incidence [36]. Due to high hardness and limited ductility, Al2.0 and Al2.5HEA coatings were undergone a rapid embrittlement during the continuous impact of erodent particles and fractured easily. In the case, erosion of Al2.0 and Al2.5 HEA coatings at 90◦ impingement is carried out by repetitive plastic deformation and brittle fracture. These observations of eroded surfaces appear to be back up the trend in erosion rates as discussed in Figure 6.

**Figure 10.** The eroded surface SEM micrographs of Al*x*CoCrFeNiTi0.5 HEA coatings after 30-min slurry erosion: (**a**) Al1.0; (**b**) Al1.5; (**c**) Al2.0; and (**d**) Al2.5 at 10.08 m/s, 1.0 wt. % sand concentration and 90◦ impingement angle.

#### **4. Conclusions**

The phase composition and microstructure of as-laser cladding Al*x*CoCrFeNiTi0.5 HEA coatings have been studied. The effect of impingement angle, sand concentration and erosion time on the erosion behavior and mechanism of Al*x*CoCrFeNiTi0.5 HEA coatings were investigated by slurry erosion test. The following conclusions could be made.


the Al2.0CoCrFeNiTi0.5 and Al2.5CoCrFeNiTi0.5HEA coatings exhibit the brittle erosion mode. AlCoCrFeNiTi0.5 HEA coating showed good slurry erosion resistance at all the investigated impingement angles due to its high hardness, good plasticity, and low stacking fault energy. The erosion rate of Al1.0 HEA coating is 1.78 times lower than Cr16 alloy at 45◦ impingement angle and 1.68 times lower at 90◦ impingement angle. The erosion rates of the test materials increase nonlinearly with the increase in the sand concentration at 45◦ and 90◦ impingement angles. The erosion time has no effect on the wear mechanism noticeably.

(4) SEM observation confirms the dominant erosion mechanism for all HEA coatings was microcuting and mixed cutting and ploughing at low impingement angle. Platelets were observed to be the primary erosion mechanism for Al1.0 and Al1.5 HEA coatings at normal impingement angle, compared to repetitive plastic deformation and fatigue fracture being the prevailing material removal phenomenon for Al2.0 and Al2.5 HEA coatings.

**Acknowledgments:** The authors thankfully acknowledge to the financial support provided by the National Natural Science Foundation of China (No. 51475140), and the National Natural Science Foundation of China (No. 51774109).

**Author Contributions:** Aibin Ma and Xiulin Ji conceived and designed the experiments; Jianhua Zhao and Yayun Bao performed the experiments; Jinhua Jiang supervised experimental work and data analysis; Jianhua Zhao wrote the paper.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Preparation and Performance Analysis of Nb Matrix Composites Reinforced by Reactants of Nb and SiC**

#### **Zhen Lu \*, Chaoqi Lan, Shaosong Jiang, Zhenhan Huang and Kaifeng Zhang**

National Key Laboratory for Precision Heat Processing of Metals, Key laboratory of Micro-Systems and Micro-Structures Manufacturing Ministry of Education, Department of Materials Science and Engineering, Harbin Institute of Technology, Harbin 150001, China; xiayilcq@163.com (C.L.); jiangshaosong@hit.edu.cn (S.J.); huangfengzhenhan@163.com (Z.H.); kfzhang@hit.edu.cn (K.Z.)

**\*** Correspondence: luzhen-hit@163.com

Received: 5 February 2018; Accepted: 27 March 2018; Published: 3 April 2018

**Abstract:** In this paper, one kind of new composite material formed with Nb and SiC was prepared by hot pressing sintering. The influence of the addition of SiC particles on the mechanical properties at room and high temperature was analyzed. The composite material consists of three phases: Nb2C, Nb3Si, and Nb solid solution (Nbss). The fraction of SiC particles added in the Nb matrix was 3%, 5%, and 7%, respectively. Flexural strength, Vickers hardness, and compressive strength at room temperature were improved with the increasing of SiC content. Among them, compressive strength and fracture toughness were higher than those of Nb/Nb5Si3 composites. The compressive strength at high temperature of the new composites was higher than that of Nb-Si alloys, which improved with the increasing of SiC content.

**Keywords:** Nb/SiC composite material; hot pressing sintering; microstructure; mechanical property

#### **1. Introduction**

In recent decades, greater demands have been placed on high-temperature structural materials with the rapid development of the aerospace field. At present, there are many high temperature metal materials that can be used under 1100 ◦C, including superalloy, TiAl intermetallic compound, titanium alloy, and others. Among them, the maximum service temperature of single crystal Ni-based superalloy is 1100 ◦C. On the other hand, there are some materials (W, Mo, Ta, Zr, Nb alloys) that can be used between 1100 ◦C and 2000 ◦C. The superhigh temperature materials still have a high strength at this temperature range. Ni-based alloys are most widely used in high-temperature components, but their operating temperature is close to the limiting temperature due to their melting point [1,2]. Refractory metals have received great attention because of their high melting point. Among the refractory metals, niobium alloy has the lowest density, and has attracted more attention from the public. In addition, Nb has stable physical and chemical properties, good corrosion resistance, and great ductility and machinability. Niobium alloy is expected to become a promising candidate for high-temperature applications [3].

Niobium alloy of Nb matrix composites are widely researched by scientists. The laminated Nb/Nb5Si3 composites had been synthesized in situ by Yu et al. [4]. Nb5Si3 phase was the reinforcement and Nbss had great plasticity, which could enhance the fracture toughness. In addition, Yu et al. [5] synthesized in situ Nb-silicide composite by powder metallurgy. Its nominal composition was Nb-16Si-10Ti-10Mo-5Hf. Solid solution strengthening and silicide hardening were the important strengthening mechanisms at high temperature. The reinforcements were Nb5Si3 and a small amount of Nb3Si. Zhang et al. [6] studied the microstructure and properties of the Nb-Ti-C-Si system. The results indicated that borides and carbides effectively increased the strength and hardness of the composites, while Nbss toughened the composites. Higher boron content increased the hardness and strength, but decreased the toughness. These studies show that the addition of reinforcements can significantly enhance the Nbss. The hardness of SiC particles is much higher than that of Si particles. Comparing to Si particles, the SiC particles are more easily surrounded by Nb particles. So, the mixed powders after ball milling are more uniform, the particles are finer and the structure of sintered material is more homogeneous. In addition, a study by Sha et al. [7] showed that Nb-based composites containing silicide and carbide phases had higher high-temperature yield strength and Vickers hardness than the composites with a single-silicide reinforced phase. There is little research on Nb-based materials reinforced by a mixture of carbide and silicide. Therefore, it is expected to produce a new type of composite with higher strength and better properties by adding SiC to the Nb solid solution. The vacuum melting method is the most common method to produce Nb-based composites [8,9]. However, composite materials produced by the vacuum melting method have shrinkage, poor density, and other defects. Beyond that, the reinforcements are prone to uneven distribution, which reduces the service performance of these materials. Hot pressing sintered materials have higher relative density and uniform structure [10]. However, research on preparing Nb-based composites with reinforcements of silicide and carbide phases is rare.

In the present study, Nb matrix composites reinforced by reactants of Nb and SiC (with SiC fractions of 3%, 5%, and 7%) were produced by hot pressing sintering technology. The effects of the amount of additional SiC on the microstructure and mechanical properties of composites were studied.

#### **2. Experimental Procedure**

Figure 1 shows the powder morphology of the original Nb and SiC powders, which was detected by scanning electron microscopy (SEM, FEI corporation, Hillsboro, OR, USA). It can be seen that the original Nb powders had irregular shape. The average particle size of Nb was 60 μm, while the particle size of SiC powders was 2 μm. The Nb-xSiC powders (*x* = 3%, 5%, 7% molar ratio) were ball milled in a QM-BP planetary ball mill under Ar protective environment. The powders, weighing 30 g, were put together with steel balls. The ball-to-powder ratio was 15:1. In order to prevent the powder from welding during the ball milling, 20 mL ethanol was added to the pot. The ball milling time was 25 h and the speed was 250 r/min. The powders, after milling, were sintered at 1550 ◦C and 30 MPa for 1 h in the vacuum hot press sintering furnace. The sintered materials were marked as Nb-3SiC, Nb-5SiC, and Nb-7SiC.

**Figure 1.** Morphology of original powders: (**a**) original Nb powders; (**b**) original SiC powders.

Samples were taken from the sintered materials for property testing and microstructure observation. Actual density was measured according to the Archimedean drainage method. Flexural strength and fracture toughness at room temperature were determined on the Instron-1186 universal testing machine. Flexural strength at room temperature was defined by a three-point bending test on specimens with 16 mm span length (total length = 20 mm), 4 mm width, and 3 mm thickness. Fracture toughness was determined by the single-edge notched beam three-point bending test on specimens with 16 mm span length (total length = 20 mm), 4 mm width, 2 mm thickness,

and a notch (2 mm depth). In the above two experiments, the loading speed of the pressing head was 0.5 mm/min, and the final result was the average value from five testing samples. Vickers hardness was measured on a MICRO-586 Vickers hardness tester with a load of 4 kg and a pressure of 15 s. The average hardness value was obtained from five indents on each sample.

High-temperature compression tests were performed on a Gleeble-1500D thermal simulation tester (DSI corporation, Poestenkill, NY, USA). The sample for testing was a cylinder with 4 mm diameter and 6 mm height, and the axis direction was the same as the hot pressing direction. Compression experiments were carried out at 1050 ◦C, 1100 ◦C, and 1150 ◦C. The strain rate was <sup>1</sup> × 10–3·s<sup>−</sup>1. The deformation was 20% (true strain was 0.223), the heating rate was 20 ◦C/s, and the holding time was 20 s.

Scanning electron microscopy was used to observe the morphology of the powders during the different ball milling stages. Backscattered electron (BSE) images were produced to reveal the microstructure of the hot-pressed materials and high-temperature compressed samples. The samples were wire-cut from the material after hot-pressing and high-temperature compressing, then sanded with sandpaper. The final specimens were obtained by ion beam polishing. X-ray diffraction (XRD, Panalytical corporation, Almelo, Holland) was used to analyze the phase composition of the composite powders after milling and the sintered materials. Different stages of mechanically alloyed powders were placed on glass slides with square grooves, flatting these powders and determining the phase compositions. For sintered materials, after grinding and polishing the sintered block samples, the phase compositions were measured on the polished surface.

#### **3. Results and Discussion**

#### *3.1. Preparation of Nb/SiC Composite Powders*

Ball milling has the benefit of obtaining a dense and uniform structure. By studying the micromorphology of composite powders, we can observe the refinement of powders and obtain the optimum milling time. Figure 2 shows the morphologies of composite powders with different ball milling times at a speed of 250 r/min. The morphology of the pure Nb powders after 20 h milling is shown in Figure 2a. The Nb particles were obviously refined into slices and sticks in comparison with the initial powders (10–20 μm). As shown by red arrows in Figure 2a, cracks formed in the Nb powders. After adding SiC and ball milling for 25 h, the powders were obviously refined. Most of the powders were in the form of crumbs (3–4 μm), and the particles were evenly distributed. Continuing milling to 30 h, Nb powders were not further refined.

**Figure 2.** Scanning electron microscope micrographs of Nb/SiC composite powders ball milled for (**a**) 20 h; (**b**) 25 h; (**c**) 30 h.

The results of XRD analysis of Nb and SiC powders milled for different times are shown in Figure 3. Compared with the XRD patterns of the original Nb powders, it can be seen that the diffraction peaks of Nb powders became broader and the intensities decreased when the Nb powders were milled for 25 h. This is because Nb powders undergo strong plastic deformation during the milling process, which causes lattice distortions and increased dislocation, resulting in an increased amount of subgrains and an area of subgrain boundaries [11]. When the number of subgrains increases to a certain extent, the subgrains will transform into grains so that the grains will be refined. Grain refinement and lattice distortion will lead to broadening diffraction peaks. Meanwhile, grain refinement can also lead to reducing diffraction peak intensity [12].

When powders were milled for 30 h, the secondary peak showed in the left of the main peak of Nb powders, indicating that the main diffraction peaks of Nb powders began to move toward the direction of low angle and the lattice constant of Nb powders increased [13]. This is because the iron element inevitably enters the composite powders during ball milling because of the stainless-steel balls and cans used. The atomic radius of Fe is 1.72 Å and the atomic radius of Nb is 2.08 Å. If an Fe atom enters the lattice interstices of Nb to form an interstitial solid solution, the lattice constant of Nb will increase and the diffraction peak will shift toward the lower angle. As shown in Figure 3, the diffraction peak width of Nb powders does not increase significantly from 25 to 30 h, indicating that the particle size of Nb powders does not change significantly during this process. In addition, the longer the milling time, the more impurities will be introduced. Based on the above analysis, 25 h is the optimal milling time.

**Figure 3.** X-ray diffraction (XRD) patterns of composite powders after different milling times.

Figure 4 shows the XRD patterns of three composite powders after ball milling for 25 h. It can be seen from the figure that no diffraction peak of SiC was detected. This is because at the initial stage of ball milling, Nb particles turn into sheets due to plastic deformation. The brittle SiC powders not only have little content, but also have small particle size (generally no more than 2 μm). The SiC powder particles are easily broken into particles that are less than 1 μm under the high-speed impact of the ball. Only a few SiC powders embed into Nb powders. Pressing composite powders and making

samples for the XRD test at this time, the flaky Nb particles will cover small SiC particles, which makes it difficult to detect the diffraction peak of SiC.

**Figure 4.** XRD patterns of composite powders after milling for 25 h.

#### *3.2. Microstructure and Mechanical Properties at Room Temperature*

#### 3.2.1. XRD Patterns and Microstructure of Sintered Material

XRD patterns of the sintered materials are shown in Figure 5. Compared with the XRD pattern of the milled composite powders, a new phase was formed in the sintered materials. The diffraction peaks of Nb3Si, Nbss, and Nb2C can be detected in the XRD patterns of the three kinds of sintered materials. The diffraction peaks of Nb-Si intermetallic compound were not detected in the XRD pattern after ball milling for 25 h (Figure 3). Therefore, the Nb-Si intermetallic compound found in the sintered materials was generated by the reaction of Nb and SiC in the hot pressing process. Sintered materials consisted of three phases: Nbss, Nb3Si, and Nb2C. Comparing the XRD patterns of the three composites, the diffraction peak of Nbss was strongest and the area the peak covered was the largest, indicating that there was the most content of Nbss in three phases, and the three composite materials were based on Nbss.

Figure 6 shows BSE images of the three kinds of composite. The material was dense, and the grains were more refined with increasing SiC content. The average grain size of Nb-3SiC was 10 μm, the average crystal size of Nb-5SiC was 8 μm, and the average grain size of Nb-7SiC was 4 μm. The composite materials consisted of three contrasting phases of black, gray, and white. It was known from the XRD patterns that the material consisted of Nb2C, Nb3Si, and Nbss phases. Generally it is well known that phases containing heavier elements exhibit brighter contrast in BSE images, thus the black grains are Nb2C, the white grains are Nbss, and the gray grains are Nb3Si. From the image it can be observed that the black grains were always adjacent to the gray grains, indicating that SiC particles react with the surrounding Nbss during hot press sintering. C and Si atoms diffuse locally: C atoms diffuse into the black grains, and Si atoms diffuse into the gray grains. The original Si and C atoms are adjacent, so the Nb2C and Nb3Si grains produced by the reaction are also adjacent, which proves that the black grains are also Nb2C. In addition, certain areas in Nb2C were darker in color, indicating that C was locally enriched in these regions and was unevenly distributed in the Nb2C grains.

**Figure 5.** XRD patterns of Nb-xSiC composites.

**Figure 6.** Backscattered electron images of the three composites: (**a**) Nb-3SiC; (**b**) Nb-5SiC; (**c**) Nb-7SiC.

#### 3.2.2. Mechanical Properties of Nb Matrix Composites at Room Temperature

Adding SiC to Nbss, mixed reinforcement consisting of carbide and silicide would be produced, which made it possible to give the Nbss better mechanical properties at room temperature. Figure 7a shows the room temperature flexural strength and fracture toughness of the new Nb matrix composites obtained with different amount of SiC. It can be seen from the figure that flexural strength increased with the addition of SiC. Flexural strength of the three components was 754.52 MPa, 793.63 Mpa, and 814.38 Mpa, respectively. The reason is that the content of Nb2C and Nb3Si formed by the reaction

increased with the addition of SiC. Nb2C and Nb3Si, which are both hard and brittle phases, embed into Nbss, hindering the movement of dislocation during deformation and having the function of dispersion hardening. The fracture toughness of the new Nb matrix composites decreased with the addition of SiC. The fracture toughness of the three materials was 13.82 MPa1/2, 12.74 MPa1/2, and 11.37 MPa1/2, respectively. Nb/Nb5Si3 composites have been widely studied by researchers. The room temperature fracture toughness of Nb was 3MPa1/2 [14], and it could go up to 8 MPa1/2 of Nb/Nb5Si3 composites. The room temperature fracture toughness of Nb and Nb/Nb5Si3 composites were both lower than that of the new Nb matrix composites. Figure 7b shows the room temperature compressive strength and Vickers hardness of composites with different amounts of SiC. Compressive strength and Vickers hardness both increased with the amount of SiC. The compressive strength of the three materials was 1659 MPa, 1780 MPa, and 1915 MPa, respectively. The Vickers hardness was 567 HV, 601 HV, and 623 HV, respectively. The compressive strength of Nb/Nb5Si3 composites was 1400 MPa. It can be seen that the compressive strength of the new Nb matrix composites improve greatly compared with Nb/Nb5Si3 composites.

Figure 8a shows the Vickers indentations of the new Nb matrix composites with different amounts of SiC. It can be seen from the figure that cracks existed only in the Nb2C and Nb3Si grains and ended at the grain boundary of Nbss, also known as crack arrest (CA). Cracks passed through the hard and brittle phases in the manner of transgranular fractures (TF) [15] and deflected when the cracks propagated. The Nbss phase is ductile. When a crack propagates in the Nbss, it undergoes plastic deformation, which consumes the energy required for crack propagation during deformation and applies compressive stress to the tip of a crack to close it. Finally, it offsets tensile stress at the tip and slows the propagation of the crack. With the increasing of SiC content, the content of the Nbss phase reduces relatively and the fracture toughness decreases. Besides, there are a greater number of grains (smaller grain size) in the same volume with the increasing of SiC content. So during the deformation process, the deformation that disperses to each grain is smaller. Thus, it is difficult to form cracks in grains. Macroscopically, the material can withstand large deformation without breaking quickly. In addition, with the increasing of SiC content, the grain boundary surface area increases. It hinders the movement of dislocations during deformation. Meanwhile, the amounts of Nb2C and Nb3Si increase and strengthen the Nbss, so the compressive strength increases with the increasing of SiC content. Furthermore, the increase of Nb2C and Nb3Si also augments the number of hard and brittle phases in the indentation area, resulting in increasing hardness.

**Figure 7.** *Cont.*

**Figure 7.** Room temperature mechanical properties of the new Nb matrix composites: (**a**) flexural strength and fracture toughness; (**b**) compressive strength and Vickers hardness.

Figure 8b shows the fracture morphology of an Nb-5SiC specimen after the three-point bending test. As the figure shows, Nb2C exhibited a transgranular cleavage fracture mode, inlaid into the Nbss like skeletons. In the process of flexural fracture, Nb2C gradually bends when it suffers bending load. When deformation reaches a certain extent, the cracks begin to germinate, and they propagate inside the N2C grains, which is known as transgranular fracture. The more SiC is added, the more Nb2C forms, and the greater the load suffered during the deformation process. Besides, in the process of propagation, cracks continually change direction, which virtually slows the crack growth rate, delaying fracture. Therefore, the room temperature flexural strength of composites increased with the amount of SiC.

**Figure 8.** Vickers hardness indentation and fracture morphology of composites: (**a**) Vickers hardness indentation; (**b**) fracture morphology.

#### *3.3. Behavior of Nb Matrix Composites at High Temperature*

#### 3.3.1. Stress-Strain Curve at High Temperature

Figure 9 shows the true stress-strain curves of the three materials at 1050 ◦C, 1100 ◦C, and 1150 ◦C. When compression started, there was an increase in flow stress as dislocations interacted and multiplied, and the true stress grew extremely fast. When it reached the peak, the true stress declined gradually. When the dislocation density increased to a certain degree, dynamic restoration occurred. The climb of edge dislocations, cross-slip of screw dislocations and counteraction of unlike dislocations are basic to the softening mechanism of the dynamic recovery, which contributes to lowering the dislocation density and opening dislocation tangle. Meanwhile, the rate of dislocation migration accelerates with increasing temperature, which decreases the deformation resistance [16]. Thus, the slope of the flow stress curve slowed down gradually.

**Figure 9.** True stress-strain curves of the three materials at different temperatures: (**a**) 1050 ◦C; (**b**) 1100 ◦C; (**c**) 1150 ◦C.

Figure 10 shows the compressive strength of the three materials at 1050 ◦C, 1100 ◦C, and 1150 ◦C. It can be observed that with the increasing of SiC, compressive strength gradually increases. Compressive strength of the three materials at 1050 ◦C was 480 MPa, 513 Mpa, and 548 MPa. Compressive strength at 1100 ◦C was 381 MPa, 417 Mpa, and 470 MPa, respectively. When the temperature was 1150 ◦C, compressive strength was 301 MPa, 348 Mpa, and 387 MPa for the three materials, respectively. Introducing Hf, Cr, and other elements into the Nb-Si alloys could improve the mechanical properties of the Nb-Si alloys. Compressive strength of the 0Hf-2B-3Cr-54Nb-22Si–based alloy and the 2Hf-2B-3Cr-52Nb-22Si-based alloy at 1250 ◦C was 194 MPa and 291 Mpa [17], respectively; compressive strength of Nb-8Si-20Ti-6Hf-(6,10,14)Cr at 1150 ◦C was 268 MPa, 278 Mpa, and 313 MPa [18]. The compressive strength of ordinary Nb-Si alloys was lower than that of the new Nb matrix composites, which shows that the addition of SiC could significantly improve the high-temperature strength of the material.

**Figure 10.** Compressive strength and peak stress at different temperatures.

With the increase of added SiC, deformation resistance and compressive strength of the material increased and the Nbss was strengthened. This is because, as the content of SiC increases, Nb2C and Nb3Si increase, and the bonding areas with Nbss also increase, hindering the movement of dislocation during deformation. In addition, the hard and brittle phases are of great help for second-phase strengthening, which also improves the high-temperature strength of the material.

#### 3.3.2. Microstructure of Material after High-Temperature Deformation

During the compression process, the specimen was divided into three parts. As shown in Figure 11a, zone 1 is the dead zone, with the least deformation, zone 2 is the pressure zone, with the largest deformation, and zone 3 is the tension stress zone. The corresponding areas on the actual compression specimen are shown in Figure 11b, a longitudinal cross-sectional view of the compressed specimen. The three regions indicated by the arrows labeled A, B, and C in Figure 11b correspond to the regions labeled 1, 2, and 3 shown in Figure 11a.

Figure 12a–c show the BSE images corresponding to the three regions A, B, and C, respectively, in Nb-5SiC (Figure 11b). The area shown in Figure 12a is close to the pressing head and belongs to the undeformed area, retaining the equiaxed grain structure of original tissue. Figure 12b is the deformation area of the compression sample. The middle part of the equiaxed grain structure was squashed, showing a strip-shaped structure. Compared with the A region, the morphology of the C region did not have significant changes, because the total compression deformation was small and the deformation in the C region was smaller, resulting in insignificant changes of the tissue.

**Figure 11.** Stress states of different parts in the compressive sample: (**a**) schematic diagram; (**b**) actual sample.

Figure 13 shows BSE images of the deformed area B in the middle of Nb-5Si and area C after deformation at different temperatures. There was a common feature among these diagrams: cracks existed in grain boundaries between Nb2C and Nb3Si, but there were no cracks in the grain boundaries between hard, brittle phases and Nbss. This is because with increased deformation, there is inconsistent deformation between Nb2C and Nb3Si, so that cracks occur in the grain boundaries. However, the deformation between the hard brittle phases and the ductile Nbss is coordinate, and thus it is not easy to produce cracks. This indicates that the material is about to undergo intergranular fracture, while the fracture mode of the material is transgranular cleavage fracture at room temperature. The maximum load of material at room temperature was greater than at high temperature. This is because at room temperature, the grain boundary strength is high and the intragranular strength is low, so the fracture mode is mainly transgranular fracture. Meanwhile, because of the small grain size of the material, the grain boundary not only impedes dislocation, but also withstands large deformation, so that it is difficult to crack. Therefore, the material has good plasticity and high strength at room temperature. As the temperature increases, the grain boundary strength and intragranular strength both decrease, but the strength of the grain boundary decreases more, resulting in the transformation of transgranular fracture at room temperature to intergranular fracture at high temperature. Because of

the low grain boundary strength at high temperature, the material will crack with small deformation. However, the ductile Nbss can absorb the energy of crack growth and delay crack propagation, playing a significant role in improving deformation of the material and preventing premature fracture.

**Figure 12.** Backscattered electron images of different parts of Nb-5SiC at 1050 ◦C: (**a**) zone A; (**b**) zone B; (**c**) zone C.

**Figure 13.** *Cont.*

**Figure 13.** Backscattered electron images of Nb-5Si at different temperatures: (**a**) zone B at 1050 ◦C; (**b**) zone C at 1050 ◦C; (**c**) zone B at 1100 ◦C; (**d**) zone C at 1100 ◦C; (**e**) zone B at 1150 ◦C; (**f**) zone C at 1150 ◦C.

#### **4. Conclusions**


**Acknowledgments:** The authors gratefully acknowledge financial support from the National Natural Science Foundation of China (grant No. 51675126) and the Natural Science Foundation of Heilongjiang Province (grant No. DC2013C048).

**Author Contributions:** Zhen Lu designed this experiment and revised the manuscript. Chaoqi Lan analyzed the experimental data and completed this paper. Zhenhan Huang performed the nanoindentation experiments. Shaosong Jiang and Kaifeng Zhang helped to complete the microstructure analysis.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

#### *Review*
