The Phase Relations of the Co-Ni-In Ternary System at 673 K and 873 K and Magnetic Properties of Their Compounds

The phase relationships of the ternary Co-Ni-In system at 673 K and 873 K were investigated by means of powder X-ray diffraction, scanning electron microscopy equipped with energy dispersive spectroscopy, and optical microscopy. Though CoIn2 does not exist at 873 K, the ternary solid solution Co1−xNixIn2 exists at both 673 K and 873 K with different composition ranges. The Rietveld refinements were carried out to investigate the crystal structure of Co1−xNixIn2 (x = 0.540, and 0.580) and Ni2−xCoxIn3 (x = 0.200). The magnetization dependence of temperature (MT) curves of Ni2−xCoxIn3 (x = 0.200) and Co1−xNixIn2 (x = 0.540) are similar to those of the ferromagnetic shape memory alloys Ni-Mn-A (A = Ga, Sn, and In), but do not undergo martensitic transformation. The maximum magnetic entropy changes in Ni2−xCoxIn3 (x = 0.200) and Co1−xNixIn2 (x = 0.540) under 3T are 1.25 and 1.475 J kg−1K−1, respectively.


Introduction
Recently, Ni-Mn-In has drawn increasing attention due to its fascinating multifunctional properties including its shape memory effect [1], magnetocaloric effect [2], elastocaloric effect [3], magnetothermal conductivity [4], magnetic superelasticity [5], barocaloric effect [6], and large exchange bias effect [7] associated with the martensitic-type phase transformation. A large number of researches have shown that the properties of the Ni-Mn-In alloys have been highly improved when accompanied with a small amount of Co doping into Ni-Mn-In ternary compounds [8][9][10]. This brings more attention to the Ni-Co-Mn-In quaternary compounds and another upsurge of research of the alloys [11][12][13][14]. The martensitic transformation, which belongs to first-order magnetostructural transformation, led to a large magnetic entropy change (∆S) which makes Ni-Co-Mn-In alloys promising candidates for magnetic refrigeration materials [15][16][17]. A giant magnetocaloric effect driven by structural transitions was found in Ni 45 . 7 Mn 36 . 3 In 13 Co 5 , which resulted in a high adiabatic temperature change (∆T ad ) of −6.2 K under a low field of 2T [2]. In Cheng's work, the temperature-induced martensitic transformation of Ni 42 Co 8 Mn 37 . 7 In 12 . 3 alloy achieved a giant (∆S) of 14.30 J·K −1 ·kg −1 and refrigeration capacity (RC) up to 549 J·K −1 . Additionally, the near-room-temperature working temperature range

Experimental
All samples with a mass of 2 g were prepared in an electric arc furnace under argon atmosphere. The purities of initial metals Co, Ni, and In were 99.9 wt. %, 99.99 wt. % and 99.9 wt. %, respectively. The material source is General Research Institute for Nonferrous Metals (Beijing, China). A proper amount of In was added to each samples due to the loss of In. Titanium was used as the oxygen scavenger during the melting process. The samples were melted three times for the purpose of homogeneity of the alloy samples. All the as-cast samples were sealed in evacuated quartz tubes to anneal at higher temperature for further homogenization. The In-rich (> 50 at.% In) alloys were kept at 923 K/673 K for 30 days, while the other alloys homogenized at 1073 K for 20 days. After homogenization annealing, the samples used for the 873 K section were cooled down to 873 K with a cooling rate of 50 K/day and maintained at 873 K for 20 days to reach equilibrium, while those used for the 673 K section were cooled down to 673 K and maintained at 673 K for 20 days. Finally, all the samples were quenched into ice-water mixture.
The X-ray powder diffraction (XRD) data of the Co-Ni-In alloy samples were obtained by using a Rigaku D/max 2500 V powder diffractometer (Cu Kα1 radiation, λ = 1.54060 Å; Tokyo, Japan). The powder XRD data for phase analysis were collected in a continuous scanning mode, and those for Rietveld refinement were collected in a step scanned mode with a step size of 0.02 • . The high temperature XRD data were collected by using a Bruker D8 advance diffractometer. The Rietveld refinements for some selected samples were carried out by using FullProf programs [34,35]. Scanning electron microscopy (SEM, Hitachi S-3400N or SU8000; Tokyo, Japan) and optical microscopy (Axio Imager A2m, Zeiss, Jena, Germany) were used for microstructure analysis, and energy dispersive analysis (EDS) was applied for the measurements of sample chemical and phase compositions. The samples used for SEM/EDS measurements were corroded with clear water. Magnetic measurements were carried out in the Physical Property Measurement System (PPMS, Quantum Design; San Diego, CA, USA).

Results and Discussion
The phase analysis was performed on all the XRD data of the equilibrated Co-Ni-In samples with the aid of the Powder Diffraction File 2 (PDF2) database released by International Center for Diffraction Data (ICDD). The diffraction patterns of those compounds absent in the PDF2 database were calculated from the crystallographic data available in references. By carrying out the phase identification on the XRD patterns of each Co-Ni-In sample, the phase components of each sample were obtained. The selected samples were further observed by SEM/EDS and optical microscopy for phase identification and composition measurements. Table S1 shows the XRD and SEM/EDS analysis results of the selected Co-Ni-In samples at 673 K. The analysis on the XRD patterns of all the binary and ternary Co-Ni-In samples shows eight binary compounds, i.e., Ni 3 In, Ni 2 In, Ni 13 In 9 , NiIn, Ni 2 In 3 , Ni 3 In 7 , CoIn 3 , and CoIn 2 , existing at 673 K. The binary compounds Ni 4 In and Ni 13 In 7 were not observed in our experimental conditions. This is in good agreement with the Co-Ni, Co-In, and In-Ni binary phase diagrams. The Backscattered electrons (BSE) micrographs of the alloys No. 1 (Co10Ni75In15) and No. 2 (Co4Ni34In62) are shown in Figure 1a,b. Both alloys contain two phases. EDS analysis on alloy No. 24 (Co10Ni75In15) in Figure 1a revealed that the dark gray phase with composition of Co35.53(5)Ni62.15(4)In2.31(4) was identified as α-Ni Co x (x = 0.355), and the gray phase with composition of Co1.02(4)Ni72.26(5)In26.72 (5) was Ni 3 In. In Figure 1b, for the sample No. 25 (Co4Ni34In62), the dark gray phase with composition of Co 2.36(4)Ni59.12(3)In38.52(5) was verified to be Ni 2 -x Co x In 3 (x = 0.118), and the gray phase with composition of Co0.86(6)Ni32.43(5)In66.72 (6) was Ni 3 In 7 . These results proved the existence of Ni 3 In, Ni 2 In 3 , and Ni 3 In 7 at 673 K, which is in agreement with the literature [26,31]. The binary compounds Ni4In and Ni13In7 were not observed in our experimental conditions. This is in good agreement with the Co-Ni, Co-In, and In-Ni binary phase diagrams. The Backscattered electrons (BSE) micrographs of the alloys No. 1 (Co10Ni75In15) and No. 2 (Co4Ni34In62) are shown in Figure 1a,b. Both alloys contain two phases. EDS analysis on alloy No. 24 (Co10Ni75In15) in Figure  1a revealed that the dark gray phase with composition of Co35.53 (5)  In order to verify the existence of the ternary compound Ni2CoIn, which was investigated by using the first principle calculation as described by Bai et al. [35], a series of samples with compositions near Ni2CoIn was prepared. Figure 2 presents the XRD patterns of sample No. 3 (Co25Ni50In25), No. 4 (Co22Ni52In26), and No. 5 (Co27Ni48In25), and none of the XRD data of these samples correspond to the diffraction pattern calculated from the crystallographic data of Ni2CoIn as shown in the literature [35]. The XRD analysis on all of these alloys showed that these alloys contained the three phases of Ni13 − xCoxIn9 (x = 0.702), Ni1 − xCoxIn (x = 0.125), and α-Ni1 − xCox (x = 0.400) pointing to the absence of Ni2CoIn at 673 K. Figure 3a,b show the SEM micrographs of samples No. 5 and No. 6, respectively, as well as the composition of each phase obtained by EDS. The results also indicated that these two alloys contained the three phases of Ni13 − xCoxIn9 (x = 0.702), Ni1 − xCoxIn (x = 0.125) and α-Ni1 − xCox (x = 0.4), which proved the non-existence of Ni2CoIn at 673 K. In order to verify the existence of the ternary compound Ni 2 CoIn, which was investigated by using the first principle calculation as described by Bai et al. [35], a series of samples with compositions near Ni 2 CoIn was prepared. Figure 2 presents the XRD patterns of sample No. 3 (Co25Ni50In25), No. 4 (Co22Ni52In26), and No. 5 (Co27Ni48In25), and none of the XRD data of these samples correspond to the diffraction pattern calculated from the crystallographic data of Ni 2 CoIn as shown in the literature [35]. The XRD analysis on all of these alloys showed that these alloys contained the three phases of Ni 13 − x Co x In 9 (x = 0.702), Ni 1−x Co x In (x = 0.125), and α-Ni 1−x Co x (x = 0.400) pointing to the absence of Ni 2 CoIn at 673 K. Figure 3a,b show the SEM micrographs of samples No. 5 and No. 6, respectively, as well as the composition of each phase obtained by EDS. The results also indicated that these two alloys contained the three phases of Ni 13 − x Co x In 9 (x = 0.702), Ni 1−x Co x In (x = 0.125) and α-Ni 1−x Co x (x = 0.4), which proved the non-existence of Ni 2 CoIn at 673 K.  3.1.2. Phase Analysis at 873 K Table S2 gives the XRD and SEM/EDS analysis results of the selected Co-Ni-In samples at 873 K. The analysis on the XRD patterns of all the binary and ternary Co-Ni-In samples shows that six binary compounds, i.e., Ni3In, Ni2In, Ni13In9, ξ, NiIn, and Co1 − xNixIn2, exist at 873 K.
According to [12], a peritectic reaction L + (α) Co → CoIn2 occurs at 823 K, and the compound CoIn2 does not exist at 873 K. However, in the present work, a solid solution Co1 − xNixIn2 was found at 873 K. The solid solution of Co1 − xNixIn2 crystallized in the same crystal structure as that of CoIn2. Figure 4a,b show the XRD pattern and SEM micrograph of alloy No. 26 (Co28Ni24In48). The XRD analysis result revealed that the alloy contained the three phases of Co1 − xNixIn2 (x = 0.612), Ni2 − xCoxIn3 (x = 0.450), and α-Co1 − xNix (x = 0.200), as seen in Figure 4a. The composition measurement showed that the grey phase with composition of Co8.69(5)Ni32.99(4)In58.32(4), the light gray phase with composition of Co20.10(5)Ni16.82(6)In63.08 (5), and the dark phase with composition of Co82.21(5)Ni17.16(6)In0.63(5) were identified to be the three phases of Ni2 − xCoxIn3 (x = 0.160), Co1 − xNixIn2 (x = 0.612), and α-Co1 − xNix (x = 0.200), respectively, as seen in Figure 4b. Figure 5 shows the XRD pattern of alloy No. 27 (Co26Ni6In68). The XRD pattern in Figure 5 clearly indicates that the alloy contained the two phases of Co1 − xNixIn2 and In, which confirmed Co1 − xNixIn2 existed at 873 K once again. This suggests that the addition of Ni into CoIn2 stabilized the compound and raised the temperature of the peritectic reaction L + (α) Co → Co1 − xNixIn2.  Table S2 gives the XRD and SEM/EDS analysis results of the selected Co-Ni-In samples at 873 K. The analysis on the XRD patterns of all the binary and ternary Co-Ni-In samples shows that six binary compounds, i.e., Ni 3 In, Ni 2 In, Ni 13 In 9 , ξ, NiIn, and Co 1−x Ni x In 2 , exist at 873 K.
According to [12], a peritectic reaction L + (α) Co → CoIn 2 occurs at 823 K, and the compound CoIn 2 does not exist at 873 K. However, in the present work, a solid solution Co 1−x Ni x In 2 was found at 873 K. The solid solution of Co 1−x Ni x In 2 crystallized in the same crystal structure as that of CoIn 2 . Figure 4a,b show the XRD pattern and SEM micrograph of alloy No. 26 (Co28Ni24In48). The XRD analysis result revealed that the alloy contained the three phases of Co 1−x Ni x In 2 (x = 0.612), Ni 2−x Co x In 3 (x = 0.450), and α-Co 1−x Ni x (x = 0.200), as seen in Figure 4a. The composition measurement showed that the grey phase with composition of Co8.69(5)Ni32.99(4)In58.32(4), the light gray phase with composition of Co20.10(5)Ni16.82(6)In63.08 (5), and the dark phase with composition of Co82.21(5)Ni17.16(6)In0.63 (5) were identified to be the three phases of Ni 2−x Co x In 3 (x = 0.160), Co 1−x Ni x In 2 (x = 0.612), and α-Co 1−x Ni x (x = 0.200), respectively, as seen in Figure 4b. Figure 5 shows the XRD pattern of alloy No. 27 (Co26Ni6In68). The XRD pattern in Figure 5 clearly indicates that the alloy contained the two phases of Co 1−x Ni x In 2 and In, which confirmed Co 1−x Ni x In 2 existed at 873 K once again. This suggests that the addition of Ni into CoIn 2 stabilized the compound and raised the temperature of the peritectic reaction L + (α) Co → Co 1−x Ni x In 2 .
Compared to the phases at 673 K, the binary compounds, i.e., Ni3In, Ni2In, Ni13In9, NiIn, Ni2In3, and ξ, exist at 873 K, while the binary compounds Ni3In7, CoIn3, and CoIn2 disappeared at 873 K. Although CoIn2 does not exist at 873 K, the ternary solid solution Co1 − xNixIn2 exists at both of 673 K and 873 K with different composition ranges. No new binary and ternary compounds were found at 673 K and 873 K.    and ξ, exist at 873 K, while the binary compounds Ni3In7, CoIn3, and CoIn2 disappeared at 873 K. Although CoIn2 does not exist at 873 K, the ternary solid solution Co1 − xNixIn2 exists at both of 673 K and 873 K with different composition ranges. No new binary and ternary compounds were found at 673 K and 873 K.    The XRD pattern and SEM micrograph of alloy No. 28 (Co8Ni22In70) are shown in Figure 6a,b, respectively. The XRD analysis of the alloy indicated that the three phases of Co 1 -x Ni x In 2 (x = 0.612), Ni 2 -x Co x In 3 (x = 0.450) and In (Liquid) coexisted in the alloy, and no diffraction patterns of Ni 3 In 7 and/or CoIn3 were observed. The SEM/EDS analysis on alloy No.28 also gave the same results. This indicates that Ni 3 In 7 and CoIn 3 do not exist at 873 K. Figure 7 shows the SEM micrograph of No. 29 (Co18Ni52In30) alloy. It is clearly seen that the alloy is composed of three phases. EDS measurements showed that the gray phase with composition of Co10.91(5)Ni46.53(4)In42.56(4) was identified as Ni 13 − x Co x In 9 (x = 2.634), the dark gray phase with composition of Co5.94(6)Ni55.83(5)In38.28(5) was verified to be ξ, and the dark phase with composition of Co48.69(4)Ni49.12(4)In2. 19(3) was α-Ni 1−x Co x (x = 0.500). This result suggests that the ξ phase existed at 873 K, which is similar to Schmetterer's investigation [36].
Compared to the phases at 673 K, the binary compounds, i.e., Ni 3 In, Ni 2 In, Ni 13 In 9 , NiIn, Ni 2 In 3 , and ξ, exist at 873 K, while the binary compounds Ni 3 In 7 , CoIn 3 , and CoIn 2 disappeared at 873 K. Although CoIn 2 does not exist at 873 K, the ternary solid solution Co 1−x Ni x In 2 exists at both of 673 K and 873 K with different composition ranges. No new binary and ternary compounds were found at 673 K and 873 K.

Solid Solubility
The solid solubilities of Co In Ni2In, NiIn, N13In9, Ni2In3, and Ni, as well as Ni in CoIn2 and Co at 673 K and 873 K were determined by XRD using the phase-disappearing and lattice parameter method combined with the SEM (EDS). The rough maximum solid solubility of above compounds was estimated by comparing the movement of the diffraction patterns of the single phases to the disappearance of the phases. A few series samples such as Ni1 − xCoxIn and Co1 − xNixIn2 were prepared for the purpose of the solid solubility determination in the Co-Ni-In ternary system. The computer software Jade 5.0 was used to calculate and refine the lattice parameters of the samples Ni1 − xCoxIn and Co1 − xNixIn2 from the XRD patterns.
3.2.1. Solid Solubility at 673 K Figure 8 presents the XRD patterns of the samples Ni1 − xCoxIn (x = 0.04, 0.08, 0.12, 0.14) at 673 K. It can be clearly seen that these samples (except that of x = 0.14) contained the single phase of Ni1 − xCoxIn, pointing to the maximum solid solubility of Co in NiIn being between x = 0.12 and x = 0.14. Figure 9a,b show the variation in the lattice parameter a and the lattice parameter c of Ni1 − xCoxIn with the content of Co, which were calculated from the XRD patterns by Jade 5.0. It can be seen from Figure 9a,b that the maximum solid solubility of Co in Ni1 − xCoxIn is x = 0.125 (6.25 at.% Co). Further

Solid Solubility
The solid solubilities of Co In Ni2In, NiIn, N13In9, Ni2In3, and Ni, as well as Ni in CoIn2 and Co at 673 K and 873 K were determined by XRD using the phase-disappearing and lattice parameter method combined with the SEM (EDS). The rough maximum solid solubility of above compounds was estimated by comparing the movement of the diffraction patterns of the single phases to the disappearance of the phases. A few series samples such as Ni1 − xCoxIn and Co1 − xNixIn2 were prepared for the purpose of the solid solubility determination in the Co-Ni-In ternary system. The computer software Jade 5.0 was used to calculate and refine the lattice parameters of the samples Ni1 − xCoxIn and Co1 − xNixIn2 from the XRD patterns.
3.2.1. Solid Solubility at 673 K Figure 8 presents the XRD patterns of the samples Ni1 − xCoxIn (x = 0.04, 0.08, 0.12, 0.14) at 673 K. It can be clearly seen that these samples (except that of x = 0.14) contained the single phase of Ni1 − xCoxIn, pointing to the maximum solid solubility of Co in NiIn being between x = 0.12 and x = 0.14. Figure 9a,b show the variation in the lattice parameter a and the lattice parameter c of Ni1 − xCoxIn with the content of Co, which were calculated from the XRD patterns by Jade 5.0. It can be seen from Figure 9a,b that the maximum solid solubility of Co in Ni1 − xCoxIn is x = 0.125 (6.25 at.% Co). Further

Solid Solubility
The solid solubilities of Co In Ni 2 In, NiIn, N 13 In 9 , Ni 2 In 3 , and Ni, as well as Ni in CoIn 2 and Co at 673 K and 873 K were determined by XRD using the phase-disappearing and lattice parameter method combined with the SEM (EDS). The rough maximum solid solubility of above compounds was estimated by comparing the movement of the diffraction patterns of the single phases to the disappearance of the phases. A few series samples such as Ni 1−x Co x In and Co 1−x Ni x In 2 were prepared for the purpose of the solid solubility determination in the Co-Ni-In ternary system. The computer software Jade 5.0 was used to calculate and refine the lattice parameters of the samples Ni 1−x Co x In and Co 1−x Ni x In 2 from the XRD patterns.
3.2.1. Solid Solubility at 673 K Figure 8 presents the XRD patterns of the samples Ni 1−x Co x In (x = 0.04, 0.08, 0.12, 0.14) at 673 K. It can be clearly seen that these samples (except that of x = 0.14) contained the single phase of Ni 1−x Co x In, pointing to the maximum solid solubility of Co in NiIn being between x = 0.12 and x = 0.14. Figure 9a,b show the variation in the lattice parameter a and the lattice parameter c of Ni 1−x Co x In with the content of Co, which were calculated from the XRD patterns by Jade 5.0. It can be seen from Figure 9a,b that the maximum solid solubility of Co in Ni 1−x Co x In is x = 0.125 (6.25 at.% Co). Further analysis on the sample Ni 1−x Co x In (x = 0.14) by the SEM (EDS) also showed that the alloy contains the three phases of Ni 1−x Co x In (x = 0.125), Ni 2−x Co x In 3 (x = 0.400), and ε-Co 1−x Ni x (x = 0.280), as seen in Figure 10, which is felt in a three-phase region. The composition measurement shows that the dark grey phase with composition of Co5.75(5)Ni45.02(6)In49.23 (6) is Ni 1−x Co x In (x = 0.125), the grey phase with composition of Co7.82(4)Ni33.11(5)In58.07 (5) is Ni 2−x Co x In 3 (x = 0.400) and the dark phase with composition of Co72.81 (7)Ni26.18(6)In1.01 (5) is ε-Co 1−x Ni x (x = 0.280). This caused the maximum solid solubility of Co in Ni 1−x Co x In to be 5.82 at.% Co at 673 K, and this value is similar to that obtained by the lattice parameter method, i.e., x = 0.125 (6.25 at.% Co). This further supports that the maximum solid solubility of Co in Ni 1−x Co x In is 6.25 at.% Co at 673 K. analysis on the sample Ni1 − xCoxIn (x = 0.14) by the SEM (EDS) also showed that the alloy contains the three phases of Ni1 − xCoxIn (x = 0.125), Ni2 − xCoxIn3 (x = 0.400), and ε-Co1 − xNix (x = 0.280), as seen in Figure 10, which is felt in a three-phase region. The composition measurement shows that the dark grey phase with composition of Co5.75(5)Ni45.02 (6)In49.23 (6) is Ni1 − xCoxIn (x = 0.125), the grey phase with composition of Co7.82(4)Ni33.11(5)In58.07 (5) is Ni2 − xCoxIn3 (x = 0.400) and the dark phase with composition of Co72.81 (7)Ni26.18(6)In1.01 (5) is ε-Co1 − xNix (x = 0.280). This caused the maximum solid solubility of Co in Ni1 − xCoxIn to be 5.82 at.% Co at 673 K, and this value is similar to that obtained by the lattice parameter method, i.e., x = 0.125 (6.25 at.% Co). This further supports that the maximum solid solubility of Co in Ni1 − xCoxIn is 6.25 at.% Co at 673 K. Similarly, the maximum solid solubilities of Ni in Co1 − xNixIn2 and ε-Co1 − xNix were determined to be 18.64 and 28 at.% Ni at 673 K, respectively. The maximum solid solubilities of Co in Ni2 − xCoxIn, Ni13 − xCoxIn9, Ni2 − xCoxIn3, and α-Ni1 − xCox were about 3, 3.2, 8, and 60 at.% Co at 673 K.   analysis on the sample Ni1 − xCoxIn (x = 0.14) by the SEM (EDS) also showed that the alloy contains the three phases of Ni1 − xCoxIn (x = 0.125), Ni2 − xCoxIn3 (x = 0.400), and ε-Co1 − xNix (x = 0.280), as seen in Figure 10, which is felt in a three-phase region. The composition measurement shows that the dark grey phase with composition of Co5.75(5)Ni45.02(6)In49.23 (6) is Ni1 − xCoxIn (x = 0.125), the grey phase with composition of Co7.82(4)Ni33.11(5)In58.07 (5) is Ni2 − xCoxIn3 (x = 0.400) and the dark phase with composition of Co72.81(7)Ni26.18(6)In1.01 (5) is ε-Co1 − xNix (x = 0.280). This caused the maximum solid solubility of Co in Ni1 − xCoxIn to be 5.82 at.% Co at 673 K, and this value is similar to that obtained by the lattice parameter method, i.e., x = 0.125 (6.25 at.% Co). This further supports that the maximum solid solubility of Co in Ni1 − xCoxIn is 6.25 at.% Co at 673 K.
Similarly, the maximum solid solubilities of Ni in Co1 − xNixIn2 and ε-Co1 − xNix were determined to be 18.64 and 28 at.% Ni at 673 K, respectively. The maximum solid solubilities of Co in Ni2 − xCoxIn, Ni13 − xCoxIn9, Ni2 − xCoxIn3, and α-Ni1 − xCox were about 3, 3.2, 8, and 60 at.% Co at 673 K.   Similarly, the maximum solid solubilities of Ni in Co 1−x Ni x In 2 and ε-Co 1−x Ni x were determined to be 18.64 and 28 at.% Ni at 673 K, respectively. The maximum solid solubilities of Co in Ni 2−x Co x In, Ni 13 − x Co x In 9 , Ni 2−x Co x In 3 , and α-Ni 1−x Co x were about 3, 3.2, 8, and 60 at.% Co at 673 K.
3.2.2. Solid Solubility at 873 K Figure 11a,b show the variation of the lattice parameter a and the lattice parameter c of Co 1−x Ni x In 2 with the content of Ni, which indicated that the maximum solid solubility of Ni in Co 1−x Ni x In 2 is x = 0.612 (about 20.19 at.% Co). The SEM micrograph of No. 26 is given in Figure 4b. The composition measurement of the light grey phase with composition of Co20.10(5)Ni16.82(5)In63.08 (5), which was Co 1−x Ni x In 2 (x = 0.612), suggested that the maximum solid solubility of Co in Co 1−x Ni x In 2 was 20.1 at.% Co at 873 K. These two values are close. Although the binary CoIn 2 is absent at 873 K, the addition of Ni in CoIn 2 stabilized the compounds and kept the solid solution Co 1−x Ni x In 2 with a wide range appeared at 873 K. The solid solubility range of Co 1−x Ni x In 2 is 3-20.1 at.% Co at 873 K.  The SEM micrograph of alloy No. 30 (Co25Ni45In35) in Figure 12 clearly shows that the alloy contains three phases. Further composition measurements indicated that the grey phase with composition of Co7.43(5)Ni43.56(6)In49.01(6) was confirmed as Ni1 − xCoxIn (x = 0.160), the light grey phase with composition of Co12.03(6)Ni47.28 (7)In40.69(6) was verified to be Ni13 − xCoxIn9 (x = 2.634) and the dark phase with composition of Co61.21(5)Ni37.51(5)In1.28(5) was α-Ni1 − xCox (x = 0.600). This further suggests that the maximum solid solubility of Co in Ni13 − xCoxIn9 is about 12.03 at. % Co at 873 K. The SEM micrograph of alloy No. 30 (Co25Ni45In35) in Figure 12 clearly shows that the alloy contains three phases. Further composition measurements indicated that the grey phase with composition of Co7.43(5)Ni43.56(6)In49.01(6) was confirmed as Ni 1−x Co x In (x = 0.160), the light grey phase with composition of Co12.03(6)Ni47.28 (7)In40.69(6) was verified to be Ni 13 − x Co x In 9 (x = 2.634) and the dark phase with composition of Co61.21(5)Ni37.51(5)In1.28(5) was α-Ni 1−x Co x (x = 0.600). This further suggests that the maximum solid solubility of Co in Ni 13 − x Co x In 9 is about 12.03 at. % Co at 873 K. Similarly, the maximum solid solubilities of Co In Ni2 − xCoxIn, Ni13 − xCoxIn9, Ni1 − xCoxIn, Ni2 − xCoxIn3, and α-Ni1 − xCox were found to be about 6, 12.03, 8,9, and 60 at.% Co at 873 K, respectively. Both of the maximum solid solubilities of In in ε-Co1 − xNix and α-Ni1 − xCox were observed to be less than 3 at.% In.
Clearly, temperature has a great effect on the solid solubility of the third element in the binary compounds of the Co-Ni-In ternary system. Normally, the maximum solid solubilities of the third Similarly, the maximum solid solubilities of Co In Ni 2−x Co x In, Ni 13 − x Co x In 9 , Ni 1−x Co x In, Ni 2−x Co x In 3 , and α-Ni 1−x Co x were found to be about 6, 12.03, 8,9, and 60 at.% Co at 873 K, respectively. Both of the maximum solid solubilities of In in ε-Co 1−x Ni x and α-Ni 1−x Co x were observed to be less than 3 at.% In.
Clearly, temperature has a great effect on the solid solubility of the third element in the binary compounds of the Co-Ni-In ternary system. Normally, the maximum solid solubilities of the third element increase with the increasing temperature. For example, the maximum solid solubilities of Co in Ni 13 − x Co x In 9 increased from 3.2 at.% Co at 673 K to 12.03 at.% Co at 873 K. However, the solid solubility range of Co 1−x Ni x In 2 was found to be 0-18.64% at. % Ni 673 K, while it shifted to the range of 3-20.1% at. % Ni at 873 K due to the absence of CoIn 2 at 873 K.

Isothermal Sections of the Co-Ni-In Ternary System at 673 K and 873 K
By comparing and analyzing more than 33 alloy samples of the Co-Ni-In ternary system and identifying the phases presented in each sample by XRD, optical microscopy, and SEM/EDS, the isothermal sections of the phase diagrams of the Co-Ni-In ternary system at 673 K and 873 K were determined. As shown in Figure 13, the isothermal section at 673 K consists of 11 single-phase regions, 21 two-phase regions, and 9 three-phase regions. The typical alloys and the details of the three-phase regions of the isothermal section of the Co-Ni-In ternary system are given in Table 2. By comparing the isothermal sections at 673 K and 873 K of phase diagram of the Co-Ni-In ternary system, the differences between the two sections can be found. The three three-phase regions, i.e., Co1 − xNixIn2 (x = 0.565) + Ni2 − xCoxIn3 (x = 0.4) + Ni3In7, Co1 − xNixIn2 (x = 0.565) + CoIn3 + Ni3In7 and CoIn3 + Ni3In7 + In (Liquid) disappear due to the absence of the binary compounds Ni3In7 and CoIn3 at 873 K [12,13], as seen Figure 4. The three-phase region Ni13 − xCoxIn9 (x = 0.702) + Ni1 − xCoxIn2 (x = 0.091) + α-Ni1 − xCox (x = 0.400) breaks into two three-phase regions, i.e., Ni2 − xCoxIn (x = 0.181) + ξ + α-Ni1 − xCox (x = 0.500), and ξ + Ni13 − xCoxIn9 (x = 2.634) + α-Ni1 − xCox (x = 0.500) since the ξ phase exist from 746 K to 1223 K [13]. According to the Co-Ni binary phase diagram, the crystal structures of Co1 − xNix alloys belong to the hexagonal structure when the concentration of Ni is less than 10 at.% Ni at 673 K, while its structure starts to change from a hexagonal structure into a cubic structure when Ni content exceed 10 at.% Ni. However, the Co1 − xNix alloy only crystallizes in a cubic structure at 873 K. Therefore, the three-phase region Ni1 − xCoxIn (x = 0.125) + α-Co1 − xNix (x = 0.600) + ε-Co1 − xNix (x = 0.280) at 673 K becomes a two-phase region, Ni1 − xCoxIn (x = 0.160) + α-Ni1 − xCox at 873 K. Although the binary compound Coln2 does not exist at 873 K, a narrow three-phase region, Co1 − xNixIn2 (x = 0.091) + ɑ-Co + In (Liquid), presents at 873 K due to the solid solution of Co1 − xNixln2 (x = 0.091-0.612) appearing at 873 K.   As shown in Figure 14, the isothermal section at 873 K contains 8 single-phase regions, 16 two-phase regions, and 8 three-phase regions. The typical alloys and the details of the three-phase regions of the Co-Ni-In isothermal section at 873 K are presented in Table 3.  The XRD and SEM/EDS data for the sample Ni2 − xCoxIn3 (x = 0.200) were collected in order to investigate its crystal structure. The XRD phase analysis points out that this sample is a single phase without any detectable impurity or additional phases. The SEM/EDS testing result shows that the composition of the sample is Ni34.62(3)Co5.23(4)In60.15(3), which reveals that 5.23(4)at.% of Co replaces the Ni position in Ni2 − xCoxIn3 (x = 0.200). To determine the crystal structures of the Ni2 − xCoxIn3 (x = 0.200) alloy, Rietveld refinement was performed from the XRD data by using the FullProf program. The powder X-ray diffraction pattern for the Ni2 − xCoxIn3 (x = 0.200) alloy is shown in Figure  15. The Rietveld refinement results of the alloys are listed in Table 4. The low values of the Rp and Rwp factors suggest that the fitted pattern is in good agreement with the experimental data and that the Rietveld refinement is reliable. The Rietveld refinement results support the case that the structure of sample remains unchanged at room temperature when cobalt is doped into the Ni2 − xCoxIn3 (x = 0.200) compound, and Ni2 − xCoxIn3 (x = 0.200) crystallizes in the Al2Ni3-type structure (space group P 3 m1). The lattice parameters are a = 0.43959(5) nm, c = 0.53121(1) nm, and Z = 1. All positions are fully occupied in the compound. The Wyckoff 1a (0, 0, 0) site and 2d (1/3, 2/3, 0.3534 (3)) site are all occupied by In atoms, while the 2d (1/3, 2/3, 0.1381 (2)) site is occupied by Co and Ni atoms (0.1Co + 0.9 Ni).  By comparing the isothermal sections at 673 K and 873 K of phase diagram of the Co-Ni-In ternary system, the differences between the two sections can be found. The three three-phase regions, i.e., Co 1−x Ni x In 2 (x = 0.565) + Ni 2−x Co x In 3 (x = 0.4) + Ni 3 In 7 , Co 1−x Ni x In 2 (x = 0.565) + CoIn 3 + Ni 3 In 7 and CoIn 3 + Ni 3 In 7 + In (Liquid) disappear due to the absence of the binary compounds Ni 3 In 7 and CoIn 3 at 873 K [12,13], as seen Figure 4. The three-phase region Ni 13 − x Co x In 9 (x = 0.702) + Ni 1−x Co x In 2 (x = 0.091) + α-Ni 1−x Co x (x = 0.400) breaks into two three-phase regions, i.e., Ni 2−x Co x In (x = 0.181) + ξ + α-Ni 1−x Co x (x = 0.500), and ξ + Ni 13 − x Co x In 9 (x = 2.634) + α-Ni 1−x Co x (x = 0.500) since the ξ phase exist from 746 K to 1223 K [13]. According to the Co-Ni binary phase diagram, the crystal structures of Co 1−x Ni x alloys belong to the hexagonal structure when the concentration of Ni is less than 10 at.% Ni at 673 K, while its structure starts to change from a hexagonal structure into a cubic structure when Ni content exceed 10 at.% Ni. However, the Co 1−x Ni x alloy only crystallizes in a cubic structure at 873 K. Therefore, the three-phase region Ni 1−x Co x In (x = 0.125) + α-Co 1−x Ni x (x = 0.600) + ε-Co 1−x Ni x (x = 0.280) at 673 K becomes a two-phase region, Ni 1−x Co x In (x = 0.160) + α-Ni 1−x Co x at 873 K. Although the binary compound Coln 2 does not exist at 873 K, a narrow three-phase region, Co 1−x Ni x In 2 (x = 0.091) + α-Co + In (Liquid), presents at 873 K due to the solid solution of Co 1−x Ni x ln 2 (x = 0.091-0.612) appearing at 873 K.

Crystal Structure of Ni 2−x Co x In 3 (x = 0.200)
The XRD and SEM/EDS data for the sample Ni 2−x Co x In 3 (x = 0.200) were collected in order to investigate its crystal structure. The XRD phase analysis points out that this sample is a single phase without any detectable impurity or additional phases. The SEM/EDS testing result shows that the composition of the sample is Ni34.62(3)Co5.23(4)In60.15 (3), which reveals that 5.23(4)at.% of Co replaces the Ni position in Ni 2−x Co x In 3 (x = 0.200). To determine the crystal structures of the Ni 2−x Co x In 3 (x = 0.200) alloy, Rietveld refinement was performed from the XRD data by using the FullProf program. The powder X-ray diffraction pattern for the Ni 2−x Co x In 3 (x = 0.200) alloy is shown in Figure 15. The Rietveld refinement results of the alloys are listed in Table 4. The low values of the R p and R wp factors suggest that the fitted pattern is in good agreement with the experimental data and that the Rietveld refinement is reliable. The Rietveld refinement results support the case that the structure of sample remains unchanged at room temperature when cobalt is doped into the Ni 2−x Co x In 3 (5); see Table S1. The refinement result shows that the Co 1−x Ni x In 2 (x = 0.540) compound, in which a large number of Co atoms are replaced by Ni, remains in the single phase and maintains the Cu 2 Mg-type structure with the space group Fddd (No. 70). The lattice parameters of Co 1−x Ni x In 2 (x = 0.540) are refined to be a = 0.9424(3) nm, b = 0.5288(4) nm, and c = 1.7742(5) nm. The Co 1−x Ni x In 2 (x = 0.580) alloy contains the Co 1−x Ni x In 2 (x = 0.565) (Cu 2 Mg-type structure) phase with a small amount of the Ni 2−x Co x In 3 (x = 0.400) (Al 3 Ni 2 -type structure). The EDS results shows that the compositions of the Co 1−x Ni x In 2 (x = 0.565) and Ni 2−x Co x In 3 (x = 0.400) phases were about Co14.11(4)Ni18.64(3)In67.25(4) and Co7.41(3)Ni35.27(4)In57.3(4)2, which were identified as Co 1−x Ni x In 2 (x = 0.565) and Ni 2−x Co x In 3 (x = 0.400), respectively. This also proves that the solubility of Ni in Co 1−x Ni x In 2 is 18.64 at.% Ni in the Ni 2−x Co x In 3 (x = 0.400) alloy. The Rietveld refinement of the Co 1−x Ni x In 2 (x = 0.580) alloy shows that the mass fractions of these two phases in the sample were 2.7% and 97.3 %, respectively. The Rietveld refinement also indicated that the amount of Ni in the phase of Co 1−x Ni x In 2 was x = 0.565. This is in agreement with the results of solid solubility determination including the SEM/EDS measurements. The refined lattice parameters of Co 1−x Ni x In 2 (x = 0.565) are a = 0. 9421(2) nm, b = 0.5282(3) nm, and c = 1.7739(3) nm, which are slightly smaller than those of Co 1−x Ni x In 2 (x = 0.540). This phenomenon is mainly due to the similar crystal structure of Co and Ni, and the radius of Ni (R Ni = 0.124 nm) atom is slightly smaller than that of Co (R Co = 0.125 nm). The replacement of Ni atoms for Co atoms causes the volume of Co 1−x Ni x In 2 to shrink. This results in smaller sizes. The lattice parameters of Ni 2−x Co x In 3 (x = 0.400) are refined to be a = 0.4397(1) nm and 0.5319(3) nm. In the structure of Co 1−x Ni x In 2 , the In atoms occupy the 16e and 16g sites, while Co atoms (including the substituting Ni atoms) exist at the 16g sites.  [37,38] causes difficulty in obtaining good single-phase Co1 − xNixIn2 samples. Thus, the crystal structure of Co1 − xNixIn2 was investigated with a few selected good single-phase samples. The crystal structure of Co1 − xNixIn2 (x = 0.540, 0.580) at 673 K was investigated via XRD and SEM/EDS. Figures 16 and 17 present the XRD patterns of Co1 − xNixIn2. Table 4 lists the Rietveld refinement results of Co1 − xNixIn2.

Magnetic Properties
3.5.1. Magnetic Properties of Ni2 − xCoxIn3 (x = 0.200) Figure 18 shows the zero-field-cooled (ZFC) and zero-field-heated (ZFH) temperature dependence of the magnetization under a static magnetic field of 50 mT for the Ni2 − xCoxIn3 (x = 0.200) alloy. During the heating process, the magnetization of the alloy remains constant at 0.044 emµ /g in the range of 5-436 K, and then increases sharply at 436 K (THS) and rises to 1.05 emμ/g at 451 K (THF). After that, the magnetization increases slowly and reaches a maximum of 1.11 emμ/g at 482 K. When the temperature increases from 482 to 528 K, the magnetization slowly drops to 1.05 emµ /g. The magnetization drops rapidly to a minimum value of 0.046 emµ /g with a further temperature increase to 560 K. The cooling curve almost completely coincides with the heating one, except in the temperature range of 339-451 K. In the cooling curve, the magnetization starts to decrease at 446 K (TCS) and stops to decrease at 399 K (TCF). The temperature cooling curve shifts to the low-temperature direction, indicating that the sample has thermal hysteresis in this temperature range; Δ = 21 K (ΔThys= (THS + THF -TCS − TCF)/2). This is very similar to the martensitic transformation phenomenon occurring in the Ni-Mn-A (A = Ga, Sn, In) [39][40][41] systems. As can be seen from the inset of Figure 18, the differential of magnetization to temperature varies with temperature and shows that the Curie temperature of the compound is 550 K.  Figure 18 shows the zero-field-cooled (ZFC) and zero-field-heated (ZFH) temperature dependence of the magnetization under a static magnetic field of 50 mT for the Ni 2−x Co x In 3 (x = 0.200) alloy. During the heating process, the magnetization of the alloy remains constant at 0.044 emµ/g in the range of 5-436 K, and then increases sharply at 436 K (T HS ) and rises to 1.05 emµ/g at 451 K (T HF ). After that, the magnetization increases slowly and reaches a maximum of 1.11 emµ/g at 482 K. When the temperature increases from 482 to 528 K, the magnetization slowly drops to 1.05 emµ/g. The magnetization drops rapidly to a minimum value of 0.046 emµ/g with a further temperature increase to 560 K. The cooling curve almost completely coincides with the heating one, except in the temperature range of 339-451 K. In the cooling curve, the magnetization starts to decrease at 446 K (T CS ) and stops to decrease at 399 K (T CF ). The temperature cooling curve shifts to the low-temperature direction, indicating that the sample has thermal hysteresis in this temperature range; ∆ = 21 K (∆T hys = (T HS + T HF -T CS − T CF )/2). This is very similar to the martensitic transformation phenomenon occurring in the Ni-Mn-A (A = Ga, Sn, In) [39][40][41] systems. As can be seen from the inset of Figure 18, the differential of magnetization to temperature varies with temperature and shows that the Curie temperature of the compound is 550 K.
Magnetization isotherms, measured during heating cycles at different temperatures, are shown in Figure 19. The samples show ferromagnetism at all measuring temperatures. Clearly, the magnetization increases with the increase in temperature. The magnetization increases from 0.178 emµ/g to 3.84 emµ/g when temperature is heated from 430 K to 454 K under the 3 T magnetic field. The sample has weak magnetism at 600 K-the magnetization is only 0.147 at 3 T. Magnetic entropy change (∆S M ) in the system, due to the application of external magnetic field, can be determined from Maxwell's relation (Equation (1)) by integrating the magnetization isotherms over the magnetic field. Magnetization isotherms, measured during heating cycles at different temperatures, are shown in Figure 19. The samples show ferromagnetism at all measuring temperatures. Clearly, the magnetization increases with the increase in temperature. The magnetization increases from 0.178 emμ/g to 3.84 emμ/g when temperature is heated from 430 K to 454 K under the 3 T magnetic field. The sample has weak magnetism at 600 K-the magnetization is only 0.147 at 3 T. Magnetic entropy change (ΔSM) in the system, due to the application of external magnetic field, can be determined from Maxwell's relation (Equation (1)) by integrating the magnetization isotherms over the magnetic field.
Magnetic entropy changes in the alloy have been derived from the isothermal magnetization curves. Figure 20 shows the magnetic entropy change in the Ni2 − xCoxIn3 (x = 0.200) sample as a function of temperature near 445 K under the 3 T, 2 T, and 1 T magnetic fields. Clearly, the greater the magnetic entropy, the stronger the magnetic field. The peak value under a field change of 3 T is 1.25 J kg −1 K −1 at T = 449.5 K. The relative cooling power (RCP = ΔSM × ΔTFWHM), which is the measure of the amount of heat transfer between the cold and hot reservoirs in an ideal refrigeration cycle, is evaluated to be 14.125 J kg −1 K −1 in the vicinity of 430-454 K for 3 T magnetic field change, where ΔTFWHM = 11.3 K. Magnetic entropy changes in the alloy have been derived from the isothermal magnetization curves. Figure 20 shows the magnetic entropy change in the Ni 2−x Co x In 3 (x = 0.200) sample as a function of temperature near 445 K under the 3 T, 2 T, and 1 T magnetic fields. Clearly, the greater the magnetic entropy, the stronger the magnetic field. The peak value under a field change of 3 T is 1.25 J kg −1 K −1 at T = 449.5 K. The relative cooling power (R CP = ∆S M × ∆T FWHM ), which is the measure of the amount of heat transfer between the cold and hot reservoirs in an ideal refrigeration cycle, is evaluated to be 14.125 J kg −1 K −1 in the vicinity of 430-454 K for 3 T magnetic field change, where ∆T FWHM = 11.3 K.   Figure 21 shows the magnetization behavior of the Co1 − xNixIn2 (x = 0.540) alloy during thermal cycling with a heating/cooling rate of 5 K/min at ZFC and ZFH. The tendency of magnetization changes with temperature of the Co1 − xNixIn2 (x = 0.540) compound is very similar to that of Ni2 − xCoxIn3 (x = 0.200). However, there are two hysteresis taking place in the Co1 − xNixIn2 (x = 0.540) alloy. The ΔThys in low temperature is 8.5 K, which is smaller than Ni2 -xCoxIn3 (x = 0.200). As can be seen from the inset of Figure 21, due to the thermal hysteresis, two different Curie temperatures (544 K and 553 K) can be obtained from the heating curve and the cooling curve. Representative isothermal magnetization loops measured around 462.5 K for Co1 -xNixIn2 (x = 0.540) are presented in Figure 22. The sample shows a low magnetization below 459 K, which rises from 0.21 emµ /g to 4.46 emµ /g with a temperature increase from 453 K to 471 K at the magnetic field of 3T. Associated with this magnetostructural transition, a large magnetic entropy change occurs. Figure 23 shows the magnetic entropy change as a function of temperature under the 3 T, 2 T, and 1 T magnetic fields for the alloy. It is clearly seen that as the magnetic field increases, the magnetic entropy of the Co1 − xNixIn2 (x = 0.540) compound becomes larger. Under a magnetic field of 3 T, the magnetic entropy of the alloy  Figure 21 shows the magnetization behavior of the Co 1−x Ni x In 2 (x = 0.540) alloy during thermal cycling with a heating/cooling rate of 5 K/min at ZFC and ZFH. The tendency of magnetization changes with temperature of the Co 1−x Ni x In 2 (x = 0.540) compound is very similar to that of Ni 2−x Co x In 3 (x = 0.200). However, there are two hysteresis taking place in the Co 1−x Ni x In 2 (x = 0.540) alloy. The ∆T hys in low temperature is 8.5 K, which is smaller than Ni 2 -x Co x In 3 (x = 0.200). As can be seen from the inset of Figure 21, due to the thermal hysteresis, two different Curie temperatures (544 K and 553 K) can be obtained from the heating curve and the cooling curve. Representative isothermal magnetization loops measured around 462.5 K for Co 1 -x Ni x In 2 (x = 0.540) are presented in Figure 22. The sample shows a low magnetization below 459 K, which rises from 0.21 emµ/g to 4.46 emµ/g with a temperature increase from 453 K to 471 K at the magnetic field of 3T. Associated with this magnetostructural transition, a large magnetic entropy change occurs. Figure 23 shows the magnetic entropy change as a function of temperature under the 3 T, 2 T, and 1 T magnetic fields for the alloy. It is clearly seen that as the magnetic field increases, the magnetic entropy of the Co 1−x Ni x In 2 (x = 0.540) compound becomes larger. Under a magnetic field of 3 T, the magnetic entropy of the alloy reaches a maximum of 1.475 J kg −1 K −1 at about 463.5 K. The corresponding half-maximum width is ∆T FWHM = 8.9 K for the compound. The direct R CP is evaluated as 13.128 J kg −1 K −1 .
Since the magnetization dependence of temperature (MT) curves of Ni 2−x Co x In 3 (x = 0.200) and Co 1−x Ni x In 2 (x = 0.540) are similar to those of ferromagnetic shape memory alloys with martensitic transformation, temperature-dependent powder XRD measurements were performed in order to obtain further insights into whether the compound undergoes martensite transformation behavior. Figure 24 shows the XRD patterns for the Ni 2−x Co x In 3 (x = 0.200) alloy measured at 490 K. According to XRD patterns, the main phase of the alloy at 490 K still was a Ni 2 In 3 phase with an Al 2 Ni 3 -type structure. Furthermore, a small amount of Ni 10 In 27 (space group Im-3 m) and Ni 2 In (space group P63/mmc) appears in alloy. The results indicate that a small amount of Ni 10 In 27 and Ni 2 In is generated during heating. Indexing of the Ni 2 In 3 phase showed that a = 0.43905 (5) nm, c = 0.52930 (3) nm, and V = 0.08836 nm 3 . All these results show that a structure phase transformation but not a martensitic transformation occurred in Ni 2−x Co x In 3 (x = 0.200) during heating. Figure 25 shows the XRD patterns of the Co 1−x Ni x In 2 (x = 0.540) alloy at 490 K, which still retains the Cu 2 Mg-type structure type at 490 K. These results point to the notion that no martensitic transformation occurred in the Co 1−x Ni x In 2 (x = 0.540) alloys during heating. reaches a maximum of 1.475 J kg −1 K −1 at about 463.5 K. The corresponding half-maximum width is ΔTFWHM = 8.9 K for the compound. The direct RCP is evaluated as 13.128 J kg −1 K −1 .   Since the magnetization dependence of temperature (MT) curves of Ni2 − xCoxIn3 (x = 0.200) and Co1 − xNixIn2 (x = 0.540) are similar to those of ferromagnetic shape memory alloys with martensitic transformation, temperature-dependent powder XRD measurements were performed in order to obtain further insights into whether the compound undergoes martensite transformation behavior. Figure 24 shows the XRD patterns for the Ni2 − xCoxIn3 (x = 0.200) alloy measured at 490 K. According to XRD patterns, the main phase of the alloy at 490 K still was a Ni2In3 phase with an Al2Ni3-type during heating. Indexing of the Ni2In3 phase showed that a = 0.43905 (5) nm, c = 0.52930 (3) nm, and V = 0.08836 nm 3 . All these results show that a structure phase transformation but not a martensitic transformation occurred in Ni2 − xCoxIn3 (x = 0.200) during heating. Figure 25 shows the XRD patterns of the Co1 − xNixIn2 (x = 0.540) alloy at 490 K, which still retains the Cu2Mg-type structure type at 490 K. These results point to the notion that no martensitic transformation occurred in the Co1 − xNixIn2 (x = 0.540) alloys during heating.

Conclusions
More than 130 samples were prepared and investigated by experimental methods to establish the phase equilibrium of the ternary Co-Ni-In system at 673 K and 873 K. The peculiar MT curves of Ni2 − xCoxIn3 (x = 0.200) and Co1 − xNixIn2 (x = 0.540) allow them to have potential to become magnetic functional materials.
Three-phase regions disappear due to the absence of the binary compounds Ni 3 In 7 and CoIn 3 at 873 K. The three-phase region Ni 13 − x Co x In 9 (x = 0.702) + Ni 1−x Co x In 2 (x = 0.091) + α-Ni 1−x Co x (x = 0.400) breaks into two three-phase regions due to the ξ phase existing at 873 K.

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