A Study of the Effects of Hf and Sn on the Microstructure, Hardness and Oxidation of Nb-18Si Silicide-Based Alloys-RM(Nb)ICs with Ti Addition and Comparison with Refractory Complex Concentrated Alloys (RCCAs)

In this paper, we present a systematic study of the as-cast and heat-treated microstructures of three refractory metal intermetallic composites based on Nb (i.e., RM(Nb)ICs), namely the alloys EZ2, EZ5, and EZ6, and one RM(Nb)IC/RCCA (refractory complex concentrated alloy), namely the alloy EZ8. We also examine the hardness and phases of these alloys. The nominal compositions (at.%) of the alloys were Nb-24Ti-18Si-5Hf-5Sn (EZ2), Nb-24Ti-18Si-5Al-5Hf-5Sn (EZ5), Nb-24Ti-18Si-5Cr-5Hf-5Sn (EZ6), and Nb-24Ti-18Si-5Al-5Cr-5Hf-5Sn (EZ8). All four alloys had density less than 7.3 g/cm3. The Nbss was stable in EZ2 and EZ6 and the C14-NbCr2 Laves phase in EZ6 and EZ8. In all four alloys, the A15-Nb3X (X = Al,Si,Sn) and the tetragonal and hexagonal Nb5Si3 were stable. Eutectics of Nbss + Nb5Si3 and Nbss + C14-NbCr2 formed in the cast alloys without and with Cr addition, respectively. In all four alloys, Nb3Si was not formed. In the heat-treated alloys EZ5 and EZ8, A15-Nb3X precipitated in the Nb5Si3 grains. The chemical compositions of Nbss + C14-NbCr2 eutectics and some Nb5Si3 silicides and lamellar microstructures corresponded to high-entropy or complex concentrated phases (compositionally complex phases). Microstructures and properties were considered from the perspective of the alloy design methodology NICE. The vol.% Nbss increased with increasing ΔχNbss. The hardness of the alloys respectively increased and decreased with increasing vol.% of A15-Nb3X and Nbss. The hardness of the A15-Nb3X increased with its parameter Δχ, and the hardness of the Nbss increased with its parameters δ and Δχ. The room-temperature-specific strength of the alloys was in the range 271.7 to 416.5 MPa cm3g−1. The effect of the synergy of Hf and Sn, or Hf and B, or Hf and Ge on the macrosegregation of solutes, microstructures, and properties of RM(Nb)ICs/RCCAs from this study and others is compared. Phase transformations involving compositionally complex phases are discussed.


Introduction
Metallic ultra-high-temperature materials (UHTMs) that are currently under development include refractory metal (RM) intermetallic composites (RMICs) based on Nb (i.e., RM(Nb)ICs), refractory metal high-entropy alloys (RHEAs), and refractory metal complex concentrated alloys (RCCAs) [1,2]. The RM(Nb)ICs are also known as Nb-silicidebased alloys, Nb in situ composites, or Nb in situ silicide composites [1,3], and some of them are also RHEAs or RCCAs, which below are referred to as RM(Nb)ICs/RHEAs or RM(Nb)ICs/RCCAs. The RHEAs and RCCAs are a subset of high-entropy alloys (HEAs) [2,4,5], the development of which was motivated by [6,7], whereas the development of RM(Nb)ICs hardness of phases in the alloys were measured using a Mitutoyo microhardness testing machine with a load of 0.1 kg that was applied for 20 s. At least 10 measurements were taken for each phase. A Sartorius Masterpro Series electronic analytical balance, along with a Sartorius YDK density determination kit, was used to calculate the density of the alloys. The Archimedean principle was applied for measuring the density of the alloys.

Results
Data for the densities and hardness of the alloys and the % area of phases in the alloy micro-structures are given in Table 1, which lists the average value, the standard deviation, and the minimum and maximum density and hardness values. The macrosegregation of alloying additions in the cast alloys is given in Table 2. The phases in the microstructures of the as-cast (AC) and heat-treated (HT) alloys are summarized in Table 3. The chemical compositions of different parts of the AC alloys, of the bulk of the HT alloys, and of the phases in all parts of the alloys are given in Tables S1-S4 in the Supplemental Data, which summarize the EPMA data for the whole ingot for the phases that were confirmed using XRD and EPMA. Bottom of the ingot refers to the part of the ingot that was close to the water-cooled Cu crucible during arc melting. These tables give the average value, the standard deviation, and the minimum and maximum analysis values. The XRD data of the alloys for the AC and HT conditions are given in Figures S1-S4 in the Supplemental Data.   Table 3. Summary of phases in the AC and HT alloys EZ2, EZ5, EZ6, and EZ8. Note that α, β, and γ refer to the structure of Nb 5 Si 3 [38]: α and β, tetragonal Nb 5 Si 3 ; γ, hexagonal Nb 5 Si 3 .

As-Cast EZ2
The actual alloy composition (at.%) was 46Nb-24.1Ti-19.6Si-5.3Hf-5Sn. Compared with the nominal composition, the EZ2-AC was richer in Si. There was macrosegregation of Si and Ti ( Table 2). The microstructure consisted of Nb ss , Nb 5 Si 3 , and A15-Nb 3 X (X = Si, Sn) phases (Table 3, Figure 1 and Figure S1 in the Supplemental Data). No hafnia particles were observed. There was eutectic of the Nb ss and Nb 5 Si 3 ( Figure 1) and Ti-rich Nb ss and Hf-rich Nb 5 Si 3 (Table 3 and Table S1 in the Supplemental Data). Note that the latter was also rich in Ti. Eutectic was observed in-between or around faceted Nb 5 Si 3 grains (Figure 1). The Ti-rich Nb ss exhibited darker contrast under BSE imaging, had Nb/Ti < 1, and was also richer in Hf and Sn compared with the "normal" Nb ss (Table S1 in the Supplemental Data). The Hf-rich Nb 5 Si 3 exhibited a brighter contrast than the Nb 5 Si 3 . This can be seen in the Figure 1a, where the darker contrast "core" of the silicide grains had average Ti/Hf = 4.3, compared with 2.8 for the surrounding grey contrast Hf-rich (and Ti-rich) Nb 5 Si 3 . The Nb/(Ti + Hf) ratios for the dark and grey contrast areas of the silicide were 1.8 and 1.97, respectively. According to the XRD data ( Figure S1 in the Supplemental Data), both αNb 5 Si 3 and βNb 5 Si 3 were present. Considering the analyses in all parts of the ingot and the Nb/(Ti + Hf) ratios of the Nb 5 Si 3 and the Hf-rich Nb 5 Si 3 (which were 2.3 and 1.6, respectively), the data would suggest that there was no hexagonal Nb 5 Si 3 in the microstructure, in agreement with the XRD data. The A15-Nb 3 X was confirmed only in the bottom of the ingot using EPMA. It exhibited brighter contrast than the Nb ss (Figure 1c). The microstructure in the bottom of the ingot was similar to that shown in Figure 1b.
The microstructure in the top and bulk of the ingot was the same and comprised of Nb ss and Nb 5 Si 3 . In some areas of the bulk, the microstructure in the vicinity of blocky faceted Nb 5 Si 3 was different from that shown in Figure 1. This microstructure, which is shown in Figure 2, was observed near Nb ss + Nb 5 Si 3 eutectic (see 5 in Figure 2b) and consisted of a lamellar microstructure, which either (i) grew into a blocky faceted Nb 5 Si 3 grain (see the left-hand side of number 17 in Figure 2b) or (ii) connected adjacent Nb 5 Si 3 grains (indicated with the numbers 1, 2, 3, and 4 in Figure 2a). The average composition (at.%) of the lamellar microstructure was 28.9Nb-38.9Ti-17.2Si-8.5Hf-6.5Sn, that is, a complex concentrated or compositionally complex (CC) lamellar microstructure [11], with Si + Sn, Si/Sn, Ti/Hf, Ti + Hf and Nb/(Ti + Hf) of about 23.7 at.%, 2.65, 4.6, 47.4 at.%, and 0.6, respectively.   In (c,d), the Nb 3 Sn is the phase with brighter contrast than the Nb ss (grey contrast phase). In (d), the Hf-rich Nb 5 Si 3 exhibits contrast slightly darker than the Nb ss and less dark than the large faceted Nb 5 Si 3 .
The Nb 5 Si 3 grains in between the lamellar microstructure (e.g., 6 and 7 in Figure 2a) were rich in Hf and Ti, with average composition 30.5Nb-22.4Ti-36.2Si-8.7Hf-2.2Sn (at.%) and Si+Sn, Si/Sn, Ti/Hf, Ti + Hf, and Nb/(Ti + Hf) of about 38.4 at.%, 16.5, 2.6, 31.1 at.%, and 1, respectively, which corresponds to a complex concentrated or compositionally complex (CC) [11] tetragonal Nb 5 Si 3 based on the Nb/(Ti + Hf) ratio. The blocky faceted Nb 5 Si 3 grains near the lamellar microstructure (e.g., 9 in Figure 2a and 8 in Figure 2b) had average composition 43.75Nb-13.5Ti-36.3Si-4.85Hf-1.6Sn (at.%) with Si+Sn, Si/Sn, Ti/Hf, Ti + Hf, and Nb/(Ti + Hf) about 37.9 at.%, 22.7, 2.8, 18.4 at.%, and 2.4, respectively, which (i) is consistent with tetragonal Nb 5 Si 3 in accordance with its Nb/(Ti + Hf) ratio and (ii) was not significantly different from the average composition of Nb 5 Si 3 in the whole of the ingot (see Table S1 in Supplemental Data). Hexagonal Nb 5 Si 3 was not confirmed by XRD. In other words, the Hf (and Ti)-rich Nb 5 Si 3 that was connected (associated) with the lamellar microstructure (or the Hf (and Ti)-rich Nb 5 Si 3 from which the lamellar microstructure formed) was CC tetragonal Nb 5 Si 3 that had higher Ti + Hf sum and lower Nb/(Ti + Hf) ratio than the blocky Nb 5 Si 3 . The solid solution away from the lamellar microstructure (e.g., 10, 13, 14, and 16 in Figure 2b) had average composition 53Nb-31.9Ti-2.65Si-4.45Hf-8Sn (at.%), not significantly different from the average composition of Nb ss in the whole of the ingot (see Table S1 in the Supplemental Data), and it had Si+Sn, Si/Sn, Ti/Hf, Ti + Hf, and Nb/(Ti + Hf) of about 10.65 at.%, 0.33, 7.2, 36.35 at.%, and 1.46, respectively. The Nb ss close to the lamellar microstructure (e.g., 11, 12, 15, 17, and 18 in Figure 2) had average composition 43.2Nb-39.9Ti-2.4Si-5.7Hf-8.8Sn (at.%), which was different from the average composition of the Ti-rich Nb ss in the whole of the ingot (see Table S1 in the Supplemental Data), and it had Si+Sn, Si/Sn, Ti/Hf, Ti + Hf, and Nb/(Ti + Hf) of about 11.2 at.%, 0.27, 7, 45.6 at.%, and 0.95, respectively. In other words, the Nb ss associated with the lamellar microstructure had higher Ti + Hf sum and lower Nb/(Ti + Hf) ratio compared with the Nb ss away from it.
The microstructure near the bottom of the ingot (Figure 1c) consisted of three phases, namely the Nb ss , A15-Nb 3 X, and Nb 5 Si 3 . There was no Ti-rich Nb ss , only Hf (and Ti)-rich Nb 5 Si 3 . The Nb 5 Si 3 and Hf-rich Nb 5 Si 3 had Nb/(Ti + Hf) of, respectively, 2.4 and 1.5, which would suggest tetragonal Nb 5 Si 3 . The Si+Sn sum and the Si/Sn ratios of the Nb ss and A15-Nb 3 X were 10 at.% and 0.3, and 18.9 at.% and 0.4, respectively.

Heat-Treated EZ2 (1500 • C/100 h)
The XRD ( Figure S1 in the Supplemental Data) and EPMA (Table S1 in the Supplemental Data) indicated Nb ss , A15-Nb 3 X, and Nb 5 Si 3 in the microstructure. Hf (and Ti)-rich Nb 5 Si 3 was also present. The typical microstructure is shown in Figure 1d. The Si+Sn concentration of the Nb ss was 6.6 at.%, compared with 10.3 at.% in EZ2-AC, with the concentrations of Si and Sn about 1.1 at.% and 2.6 at.% lower, but the Si/Sn ratio was the same (0.3). The A15-Nb 3 X had Si + Sn and Si/Sn of, respectively, 18.7 at.% and 0.3, and the Sn concentration had increased by about 1 at.%. The Nb/(Ti + Hf) ratios for the Nb 5 Si 3 and the Hf-rich Nb 5 Si 3 were, respectively, 2.1 and 0.7; the latter value would suggest hexagonal Nb 5 Si 3 (see Figure S1 in the Supplemental Data).

As-Cast EZ5
The actual composition (at.%) of the alloy was 42.1Nb-23.9Ti-19.3Si-4.7Al-5.1Hf-4.9Sn. Compared with the nominal composition, the EZ5-AC was richer in Si. There was stronger macrosegregation of Si than Ti in the ingot ( Table 2). Study of the microstructure was difficult owing to the partitioning of Hf and the presence of the A15-Nb 3 X (X = Al,Si,Sn), the contrast of which was similar to that of the Nb ss ( Figure 3). According to the XRD data ( Figure S2 in the Supplemental Data), the phases in EZ5-AC were αNb 5 Si 3 , βNb 5 Si 3 , γNb 5 Si 3 , Nb ss , and Nb 3 Sn. There was Ti-and Hf-rich Nb 5 Si 3 (Table 3 and Table S2 in the Supplemental Data). The Nb/(Ti + Hf) ratio of the Hf-rich Nb 5 Si 3 would suggest the presence of γNb 5 Si 3 , in agreement with the XRD results. Table S2 in the Supplemental Data gives the compositions of the phases that were confirmed using XRD and EPMA, namely the Nb 5 Si 3 , Nb ss , and A15-Nb 3 X (Table 3, Figure 3 and Figure S2 in the Supplemental Data). The latter was not observed in all parts of the ingot. The former two formed a eutectic consisting of the Nb ss and Ti-and Hf-rich Nb 5 Si 3 ( Figure 3). This eutectic was not observed in all parts of the ingot. No HfO 2 was observed in EZ5-AC. The Ti-and Hf-rich Nb 5 Si 3 exhibited a brighter contrast than the Nb 5 Si 3 . This can be seen in Figure 3a (see areas indicated with the numbers 1 to 7), in Figure 3b (numbers 1 to 3), in Figure 3d (numbers 1 to 3), and in Figure 3e (numbers 1 and 2). The Ti-and Hf-rich Nb 5 Si 3 in these areas had average Hf+Sn and Ti/(Hf+Sn) of, respectively, about 9.8 at.% and 2.3, compared with 6.2 at.% and 2.8 for the "normal" Nb 5 Si 3 .
The microstructure in the top of the ingot consisted of large Nb 5 Si 3 grains surrounded by Nb ss (Figure 3a), but A15-Nb 3 X was not observed. The Hf-rich areas of Nb 5 Si 3 grains exhibited different contrasts of grey (see 1 to 7 in Figure 3a). The Nb ss had Si/(Sn+Al) of about 0.2. In both the Nb 5 Si 3 and the Hf-rich Nb 5 Si 3 , the Si + Sn + Al concentration was about 37.8 at.%. In the latter, the Nb/(Ti + Hf) ratio was 1.4, indicative of tetragonal Nb 5 Si 3 .   The microstructure in the bulk of the ingot was different than that in the top and consisted of Nb 5 Si 3 , Ti-and Hf-rich Nb 5 Si 3 , Nb ss , and A15-Nb 3 X phases (Figure 3b,c). Large Nb 5 Si 3 grains were surrounded by a eutectic of the Ti-and Hf-rich Nb 5 Si 3 and the Nb ss (see Table S2 in Supplemental Data). In the Figure 3b, the Hf-rich Nb 5 Si 3 in the eutectic is indicated with the numbers 1 to 3. The A15-Nb 3 X compound was adjacent to the eutectic. In more than one case, it grew directly next to Nb 5 Si 3 (see Figure 3b,c). The Nb ss in the eutectic exhibited a lighter grey contrast compared with the Nb 5 Si 3 and Hf-rich Nb 5 Si 3 . The Nb ss had Si/(Sn+Al) of about 0.2. The Si + Sn + Al concentration of the eutectic was 23.2 at.%, with Si/(Sn+Al) of about 1.4, and the corresponding values for the A15-Nb 3 X were about 20.6 at.% and 0.3, respectively. In the Nb 5 Si 3 and the Ti-and Hf-rich Nb 5 Si 3 , the Si + Sn + Al concentration was about 37.6 and 38.5 at.%, respectively. The Ti-and Hf-rich Nb 5 Si 3 had Nb/(Ti + Hf) of about 0.7, indicative of hexagonal Nb 5 Si 3 .
In the bottom of the ingot, the microstructure consisted of Nb 5 Si 3 adjacent to or surrounded by Ti-and Hf-rich Nb 5 Si 3 , which was occasionally surrounded by A15-Nb 3 X, and by Nb ss adjacent to or surrounded by the Ti-and Hf-rich Nb 5 Si 3 (Figure 3d,e). The Tiand Hf-rich Nb 5 Si 3 grew from the Nb 5 Si 3 with a distinct morphology (see Figure 3d). No eutectic was observed in the bottom of the ingot. The Nb ss had Si/(Sn+Al) of about 0.2. The Si + Al + Sn content and the Si/(Sn + Al) ratio of the A15-Nb 3 X were about 20.3 at.% and 0.3, respectively. The Si + Sn + Al concentration in the Nb 5 Si 3 was about 36.9 at.%, whereas in the Ti-and Hf-rich Nb 5 Si 3 , it was about 37.9 at.%. Furthermore, the Nb/(Ti + Hf) ratio of the Ti-and Hf-rich Nb 5 Si 3 was about 0.8, which pointed to hexagonal Nb 5 Si 3 .

Heat-Treated EZ5-HT1 (1500 • C/100 h)
According to the XRD data ( Figure S2 in the Supplemental Data) the microstructure consisted of the αNb 5 Si 3 , βNb 5 Si 3 , γNb 5 Si 3 silicides and the Nb 3 Sn compound (Figure 4a). Nb ss and HfO 2 were not observed. The microstructure consisted of Nb 5 Si 3 (dark contrast) and Hf-rich Nb 5 Si 3 (grey contrast) surrounded by the A15-Nb 3 X (brighter contrast). Submicron particles precipitated in the Nb 5 Si 3 and the Hf-rich Nb 5 Si 3 . The composition of these particles could not be determined owing to their size. The Si + Sn + Al concentration in the A15-Nb 3 X was about 19.7 at.% and the Si/(Sn+Al) ratio was about 0.3. In the Nb 5 Si 3 and the Ti-and Hf-rich Nb 5 Si 3 the Si + Sn + Al concentration was about 38.2 at.% and 39.3 at.%, respectively. The Nb/(Ti + Hf) ratio in the Ti-and Hf-rich Nb 5 Si 3 was about 0.8, indicating hexagonal Nb 5 Si 3 .  The alloy EZ5 was given a second heat treatment at 1500 • C for an additional 100 h, in order to find out whether equilibrium had been achieved. The same specimen that was first heat treated for 100 h was given another 100 h heat treatment. According to the XRD data, the microstructure consisted of the same phases as in EZ5-HT1, namely αNb 5 Si 3 , βNb 5 Si 3 , γNb 5 Si 3 silicides, and the Nb 3 Sn compound ( Figure S2 in the Supplemental Data), and did not change compared with the EZ5-HT1. In some parts of the microstructure, the A15-Nb 3 X exhibited a small variation in contrast (see numbers 1 and 2 in the Figure 4b). Data for the chemical composition of the phases in EZ-HT2 are given in Table S2 in the Supplemental Data. The precipitates that were observed in both the Nb 5 Si 3 and the Ti-and Hf-rich Nb 5 Si 3 in EZ5-HT1 were still present (Figure 4b), but in the EZ5-HT2, the size of these precipitates in some Nb 5 Si 3 grains made possible their chemical analysis. The results confirmed that the precipitates were indeed the A15-Nb 3 X compound. The Si + Sn + Al concentration of the Nb 3 X was about 19.6 at.% with an Si/(Sn + Al) ratio of about 0.3. The S + Sn + Al concentration of the Nb 5 Si 3 and the Ti-and Hf-rich Nb 5 Si 3 was about 37.7 at.% and 38.4 at%, respectively. Both values were lower than the corresponding ones in EZ5-HT1. Similar to the EZ5-HT1, the Nb/(Ti + Hf) ratio of the Ti-and Hf-rich Nb 5 Si 3 phase was about 0.8, which is indicative of hexagonal Nb 5 Si 3 .
The microstructure near the top of the ingot (Figure 6a) was similar to that in the bulk ( Figure 5a) and consisted of bulky, faceted Nb 5 Si 3 (dark contrast) surrounded by Nb ss (grey contrast), A15-Nb 3 X (very bright contrast), and NbCr 2 Laves (very dark contrast). The A15-Nb 3 X was surrounded by the Nb ss or by the Laves phase and Ti-and Hf-rich Nb 5 Si 3 (Figure 6a). The Laves phase was formed in between A15-Nb 3 X grains or Ti-rich Nb ss and Ti-and Hf-rich Nb 5 Si 3 (Figure 6a). The Ti-rich Nb ss exhibited a darker contrast compared with the "normal" Nb ss (numbers 1 and 2 in Figure 6a). The Ti-and Hf-rich Nb 5 Si 3 (numbers 3 and 4 in Figure 6a) was separate from the "normal" Nb 5 Si 3 .
In the top of the ingot, the Si+Sn concentration in the Nb ss was about 8.9 at.%, and the Si/Sn ratio was about 0.3, while the corresponding values for the Ti-rich Nb ss were about 9.0 at.% and 0.2, respectively. In the A15-Nb 3 X, the Si+Sn content and Si/Sn ratio was about 18.6 at.% and 0.4. The Si+Sn concentration in the Nb 5 Si 3 and the Ti-and Hfrich Nb 5 Si 3 was about 37.4 at.% and 38.3 at.%, respectively, and the Nb/(Ti + Hf) ratio of the latter was about 0.7, which corresponds to hexagonal γNb 5 Si 3 . The Si + Sn + Cr concentration of the Laves phase was about 58.9 at.%. The Laves phase formed a very fine eutectic with the Nb ss . In this part of the ingot, the average composition of the eutectic was 23.3Nb-24.7Ti-7.4Si-35.7Cr-6.0Hf-2.9Sn (at.%), with Si+Sn+Cr of about 46.0 at.%.    In the bulk, the microstructure consisted of the same phases as near the top (Figures 5a and 6b-d), but the vol.% of the A15-Nb 3 X was higher. The Si+Sn concentration in both the Nb ss and the Ti-rich Nb ss was slightly reduced compared with the top (about 8.4 at.%), and both had the same Si/Sn ratio (about 0.3). In the A15-Nb 3 X, the Si+Sn content was about 18.7 at.%, and Si/Sn about 0.4. The Si+Sn concentration in the Nb 5 Si 3 and the Tiand Hf-rich Nb 5 Si 3 was the same (about 38.1 at.%), but the Nb/(Ti + Hf) ratio of the latter silicide was about 1.4. The Si + Sn + Cr concentration of the Laves phase was about 51.1 at.%. There was also a Nb ss +NbCr 2 eutectic, with average composition 21.2Nb-32.6Ti-11.6Si-22.6Cr-8.2Hf-3.8Hf (at.%), and Si+Sn+Cr of about 38.0 at.%. The eutectic was formed near Ti-and Hf-rich Nb 5 Si 3 . Some eutectic areas contained chunks of the Laves phase (e.g., see eutectic on the left-hand side of number 1 in Figure 6c).
The microstructure in the bottom of the ingot consisted of the same phases as near the top and bulk (Figures 5b and 6e), but the vol.% of the A15-Nb 3 X was significantly reduced. Furthermore, there was no evidence of the NbCr 2 + Nb ss eutectic. The Si+Sn concentration and Si/Sn ratio in the Nb ss and Ti rich Nb ss were about 8.6 at.% and 0.3, and 8.8 at.% and 0.2, respectively. The Si+Sn content and the Si/Sn ratio of the A15-Nb 3 X were about 18.4 at.% and 0.4, respectively. In the Nb 5 Si 3 and the Ti-and Hf-rich Nb 5 Si 3 , the Si+Sn concentration was essentially the same (about 38.2 at.%), and the Nb/(Ti + Hf) ratio in the latter was about 0.8, which corresponds to hexagonal γNb 5 Si 3 , as was the case near the top of the ingot. The Cr + Si + Sn concentration in the Laves phase was about 56.8 at%. The Laves phase was surrounded by either Ti-rich Nb ss or Ti-and Hf-rich Nb 5 Si 3 .

Heat-Treated EZ6-HT1 (1500 • C/100 h)
After the heat treatment at 1500 • C for 100 h, there was liquation in EZ-HT1. The microstructure had significantly coarsened compared with EZ6-AC (compare Figure 7 with Figure 5a) and exhibited features similar to those reported for the heat-treated alloy Nb-24Ti-18Si-8Cr-4Al (alloy KZ2-HT1 (1500 • C/100 h), in [21]), which also had undergone liquation. Data for the microstructure of EZ6-HT1 is given in Table S3 and in Figure S3 in the Supplemental Data, and in Figure 7. According to the XRD data, αNb 5 Si 3 , βNb 5 Si 3 , γNb 5 Si 3 , Nb 3 Sn, Nb ss , HfO 2 , and C14-NbCr 2 Laves phase were present. There were no Ti-rich areas in the Nb ss , but there was still Ti-and Hf-rich Nb 5 Si 3 . Precipitation of the Laves phase had occurred not only at the interfaces of the Nb ss with the A15-Nb 3 X or Hf-rich Nb 5 Si 3 , but also within the Nb ss grains, as shown in Figure 7. The needle-like Laves precipitates that were formed within and around the grains of the Nb ss exhibited similar morphology to those observed in the KZ2-HT1 alloy that had experienced liquation [21]. After this heat treatment, a very small volume fraction of HfO 2 was formed. The Si+Sn content and Si/Sn ratio of the Nb ss were about 7.3 at.% and 0.1, respectively. The corresponding values for the A15-Nb 3 X were about 18.5 at.% and 0.3, respectively. The Si+Sn concentration in the Nb 5 Si 3 and the Ti-and Hf-rich Nb 5 Si 3 was about 38.1 at.% and 38.8 at.%, respectively. The Nb/(Ti + Hf) ratio in the Ti-and Hf-rich Nb 5 Si 3 was about 0.7, indicating hexagonal γNb 5 Si 3 . The Cr+Si+Sn concentration in the Laves phase was about 54.1 at.%. The latter value should be viewed with caution owing to the size of the Laves phase in EZ6-HT1.
The microstructure of EZ6-HT2 was significantly different compared with EZ6-HT1 ( Figure 8a). There was no Ti-rich Nb ss , but Ti-and Hf-rich Nb 5 Si 3 was still present. The vol.% of the Nb ss was significantly reduced compared with EZ6-AC ( Table 1). The Nb 5 Si 3 was still facetted (Figure 8a). The average Si+Sn concentration and the Si/Sn ratio of the Nb ss were about 4.6 at.% and 0.2, and the corresponding values for the A15-Nb 3 X were about 18.2 at.% and 0.2, respectively. There was growth of the Laves phase ( Figure 8), of which the average Cr + Si + Sn concentration was about 59.7 at.%. The Si+Sn concentration in Nb 5 Si 3 was about 37.4 at.%. The Ti-and Hf-rich Nb 5 Si 3 was distinct from the "normal" Nb 5 Si 3 , and its contrast was very close to that of the A15-Nb 3 X (Figure 8b). Figure 8b,c,d shows that adjacent to "normal" Nb 5 Si 3 silicide grains, but not surrounding the whole Nb 5 Si 3 grain, there was a microstructure that had formed after this heat treatment (indicated with A in Figure 8, and referred to below as microstructure A). The average composition of the "normal" Nb 5 Si 3 grains, next to which the microstructure A was formed, was 42.5Nb-13.9Ti-36.7Si-0Cr-5Hf-1.8Sn (at.%), with Si + Sn = 38.5, Si/Sn = 20.8, Ti/Hf = 2.8, Ti + Hf = 19 at.%, and Nb/(Ti + Hf) = 2.25 (i.e., tetragonal silicide). Note that there was no Cr in these silicide grains. The phases in microstructure A exhibited different contrasts, and possibly formed a "lamellar" microstructure ( Figure 8d). The contrast of the phases was similar to that of the phases in the microstructure of EZ6-HT2. Separate analysis of the composition of each of the phases was not possible owing to their size. Instead, large area analysis of microstructure A gave its average composition as 25Nb-27.1Ti-33.8Si-1.2Cr-11.2Hf-1.7Sn (at.%), with Si + Sn = 35.5 at.%, Si/Sn = 19.9, Ti/Hf = 2.4, Ti + Hf = 38.3 at.%, and Nb/(Ti + Hf) = 0.66.
The interface of microstructure A with the A15-Nb 3 X was often decorated with fine particles exhibiting white contrast, that is, a bright contrast phase (BCP) (Figure 8b

As-Cast EZ8
The actual composition (at.%) of the alloy was 36.9Nb-24.6Ti-17.8Si-4.9Al-5.1Cr-5.4Hf-5.3Sn. The macrosegregation of Si was the highest of all the alloys. In addition, there was macrosegregation of Cr and Ti (Table 2). Typical microstructures are shown in Figures 9 and 10. Study of the microstructure of EZ8 was difficult owing to the partitioning of Hf and the formation of A15-Nb 3 X, the contrast of which was similar to that of the Nb ss . According to the XRD data ( Figure S4 in Supplemental Data), the microstructure consisted of αNb 5 Si 3 , βNb 5 Si 3 , γNb 5 Si 3 , Nb 3 Sn, Nb ss , HfO 2 , and C14-NbCr 2 Laves phase. The analysis data (Table S4 in Supplemental Data) supports the presence of γNb 5 Si 3 , as the Nb/(Ti + Hf) ratio of the silicide was less than one. The HfO 2 was observed only in the bulk of the ingot, where it had formed at a very small vol.%.
The microstructure in the top of the ingot was similar to that in the bulk. It consisted of large Nb 5 Si 3 grains that were surrounded by Ti-and Hf-rich Nb 5 Si 3 . There was strong microsegregation of Hf in the Nb 5 Si 3 that exhibited different contrasts (Figure 10a,b). Adjacent to the Ti-and Hf-rich Nb 5 Si 3 , Nb ss was formed. The A15-Nb 3 X was formed adjacent to the Nb ss , and it was often completely surrounded by it. In many parts of the microstructure, a complicated network of Ti-and Hf-rich Nb 5 Si 3 and A15-Nb 3 X had grown adjacent to Nb 5 Si 3 . The microstructure in these areas (denoted as A in Figure 10a,b) seemed to have grown directly from the Nb 5 Si 3 , and even though it was coarser, it resembled a lamellar microstructure. The Laves phase was often observed at the interface of the Nb ss with the A15-Nb 3 X. The average Si + Al + Sn concentration in the Nb ss was about 13.1 at.% with an Si/(Sn + Al) ratio of about 0.2. The corresponding values for the A15-Nb 3 X were about 20.9 at.% and 0.3, respectively. In the Nb 5 Si 3 , the Si + Al + Sn concentration was about 35.8 at.%, and it was about 37.2 at.% for the Ti-and Hf-rich Nb 5 Si 3 . The Nb/(Ti + Hf) ratio of the latter was about 0.7, indicating γNb 5 Si 3 . The Si + Sn + Al + Cr concentration of the Laves phase was about 47.9 at.%. There was a Nb ss +NbCr 2 eutectic in the areas close to the top of the ingot.
The microstructure in the bulk of the ingot was similar to the one observed in the top (see Figures 9a and 10c,d). The areas between the A15-Nb 3 X and the Ti-and Hf-rich Nb 5 Si 3 (denoted as A) that were observed in the top of EZ8-AC were also present in this part of the ingot. The average Si + Al + Sn concentration in the Nb ss was about 12.9 at.%, the Si/(Sn+Al) ratio was about 0.2, and the corresponding values for the A15-Nb 3 X were about 21.1 at.% and 0.3, respectively. In the Nb 5 Si 3 , the Si + Al + Sn concentration was about 38.3 at.%. The respective concentration in the Ti-and Hf-rich Nb 5 Si 3 was about 37.2 at.%, with Nb/(Ti + Hf) of about 0.7. The average Si + Al + Sn + Cr concentration in the Laves phase was about 50 at.%. There was Nb ss + NbCr 2 eutectic in some areas of the bulk of EZ8-AC.
The microstructure in the bottom of the ingot was different compared with the top and bulk. The vol.% of the Nb 5 Si 3 was significantly decreased. The Si concentration in the bottom of the ingot was about 14.7 at.%, that is, significantly lower than the nominal composition of Si in the alloy (Table S4 in Supplemental Data). Furthermore, the microstructure shown in Figures 9b and 10e,f was distinctively finer compared with that observed in the top and bulk, and consisted of Nb 5 Si 3 , Ti-and Hf-rich Nb 5 Si 3 , A15-Nb 3 X, Nb ss , and the C14-NbCr 2 Laves phase. A sharp (flat) interface was formed between the Nb 5 Si 3 and Nb ss . The areas where this morphology was evident are denoted as B in Figure 10f. The average Si + Sn + Al concentration in the Nb ss was about 13.0 at.%, with Si/(Sn + Al) of about 0.1. The corresponding values for the A15-Nb 3 X were about 20.6 at.% and 0.3, respectively. The Si + Sn + Al concentration in the Nb 5 Si 3 and the Ti-and Hf-rich Nb 5 Si 3 was about 37.0 at.% and 38.5 at.%, respectively. As was the case in other areas of EZ8-AC, the Ti-and Hf-rich Nb 5 Si 3 had Nb/(Ti + Hf) of about 0.7, indicating hexagonal γNb 5 Si 3 . The Si + Sn + Al + Cr concentration in the Laves phase was about 59.6 at.%. The Nb ss + NbCr 2 eutectic was also observed.    The Laves phase is indicated with L, and the Nb ss with ss. The A15-Nb 3 X is shown as Nb 3 Sn. In (a), the Nb 3 Sn is shown with the numbers 1, 6, and 7; the Ti-and Hf-rich Nb 5 Si 3 with the numbers 3, 4, 5, 8, and 9; and the Laves phase with 2. In (b), the Nb 3 Sn is shown with the number 1, and the Ti-and Hf-rich Nb 5 Si 3 with 3. In (c), the Nb 3 Sn is shown with the number 1, and the Ti-and Hf-rich Nb 5 Si 3 with 3. In (d), the Nb ss is shown with the number 4; Nb 3 Sn with 3, 5, and 6; Nb 5 Si 3 with 1; Ti-and Hf-rich Nb 5 Si 3 with 7 and 8; the Laves phase with 2; and HfO 2 is the very bright phase on the left-hand side of the number 5. In (e), Nb 3 Sn is shown with the numbers 1, 3, 4, and 5; the Nb ss with 6; and the Nb 5 Si 3 with 2. In (f), the Nb 3 Sn is shown with the numbers 1, 2, 3, and 6; the Laves phase with 4; and the Ti-and Hf-rich Nb 5 Si 3 with 5. For the areas A in (b-d) and the area B in (f), see text.

Heat-Treated EZ8 (EZ8-HT)
After heat treatment at 1300 • C for 100 h, the micro-structure of EZ8-HT consisted of αNb 5 Si 3 , βNb 5 Si 3 , γNb 5 Si 3 (Nb/(Ti + Hf) about 0.7), A15-Nb 3 X, C14-NbCr 2 Laves phase, and HfO 2 ( Figure S4 in the Supplemental Data). The microstructure is shown in Figure 11 and consisted of large Nb 5 Si 3 grains surrounded by a network of interpenetrating Ti-and Hf-rich Nb 5 Si 3 and A15-Nb 3 X. Nb ss was not observed. The C14-NbCr 2 Laves phase had grown significantly larger and formed distinct areas at the boundaries of either the Ti-and Hf-rich Nb 5 Si 3 or the A15-Nb 3 X phase. The average Si + Sn + Al concentration in the A15-Nb 3 X was about 19.7 at%, with Si/(Sn + Al) of about 0.2. In the Nb 5 Si 3 and the Ti-and Hf-rich Nb 5 Si 3 phases, the Si + Sn + Al concentration was about 36.7 at.% and 38.1 at.%, respectively. The Laves phase had average Si + Sn + Al + Cr of about 62 at.%. A15-Nb 3 X had precipitated in the Nb 5 Si 3 .

Hardness
The hardness of the alloys and the hardness of the A15-Nb 3 X, Nb ss , and Nb 5 Si 3 phases are given in Tables 1 and 4, respectively. The hardness of the alloy EZ2 did not change after the heat treatment, whereas the hardness of the alloys EZ5, EZ6, and EZ8 increased. In the case of the alloy EZ2, there was a slight reduction in the hardness of Nb 5 Si 3 and a more significant reduction in the hardness of Nb ss . The hardness of Nb ss in EZ6 also decreased after heat treatment. Regarding the alloys EZ5, EZ6, and EZ8, the hardness of A15-Nb 3 X increased after heat treatment. However, the hardness of Nb 5 Si 3 increased in EZ5 and decreased in EZ6 and EZ8.

Lattice Parameter of Nb ss
The lattice parameter of the bcc Nb ss in the alloys of this work is given in Table 5, where data for other comparable alloys is included (see discussion).

Macrosegregation
Macrosegregation of solute additions is common in cast RM(Nb)ICs [40]. The macrosegregation of Si (MACSi) in Nb-18Si silicide-based alloys with/without the addition of Al, Cr, Hf, Sn, or Ti is compared in Table 6. The data show the following: In other words, (1) the synergy of 5 at.% Hf with 5 at.% Al, 5 at.% Cr, 5 at.% Sn, and 24 at.% Ti slightly reduced MACSi, compared with the Hf-free alloy ZX8; and (2) the synergy of Hf and Sn with the addition of Ti, Al, and Cr increased MACSi in EZ8, compared with the alloys EZ2, EZ5, and EZ6. However, compared with the Ge addition in equivalent Nb-18Si silicide-based alloys, where the synergy of 5 at.% Hf with 5 at.% Al, 5 at.% Cr, 5 at.% Ge, and 24 at.% Ti reduced MACSi in the alloy ZF9 (MACSi = 3.1 at.%, ZF9 = Nb-24Ti-18Si-5Al-5Cr-5Hf-5Ge, nominal [35]) but increased MACSi in the Hf-free alloy ZF6 (MACSi = 4.3 at.%, ZF6 = Nb-24Ti-18Si-5Al-5Cr-5Ge, nominal [35]), the synergy of Hf and Sn had a similar but weaker effect on MACSi (compare ZX8 and EZ8 (10 versus 7.7 at.%) with ZF6 and ZF9 (4.3 at.% versus 3.1 at.%)). To put it another way, MACSi is less of an issue when Al, Cr, Hf, Si, and Ti are simultaneously in synergy with Ge than with Sn, but given that all the aforementioned elements are key for the oxidation resistance of RM(Nb)ICs, and that the simultaneous presence of Ge and Sn in RM(Nb)ICs suppressed pest oxidation and scale spallation at high temperatures [41] (see Section 4.2.14), it is unlikely that MACSi-free and oxidation-resistant metallic UHTMs can be produced using cold-hearth processing.
Comparison of the alloys EZ2, EZ5, EZ6, and EZ8 with regard to the macrosegregation of Ti (MACTi) ( Table 2) shows (a) that Al, when it was in synergy with Hf and Sn, decreased MACTi (compare the alloys EZ2 and EZ5); (b) that Cr, when it was in synergy with Hf and Sn, had a very strong effect on MACTi (alloys EZ2 and EZ6); and (c) that in the presence of Al, the effect of Cr, Hf, and Sn on MACTi was reduced (alloys EZ6 and EZ8). The macrosegregation of Cr (MACCr) that was observed in the alloy EZ6 (Table 2) was slightly reduced with the addition of Al in the alloy EZ8, which would suggest that the effect of the synergy of Ti, Hf, and Sn on MACCr was not annulled by the addition of Al. Table 6. Effect of alloying addition(s) (nominal composition, at.%) on the macrosegregation of Si (MACSi = C Si max − C Si min , at.% [40]) in as-cast Nb-18Si silicide-based alloys with/without Ti addition.
The Nb 3 Si silicide was destabilized in RM(Nb)ICs (i) by Al (alloy KZ7, [21]); and (ii) by the synergy (a) of Hf and Al (alloy YG2, [25]), (b) of Sn and Al in the absence of Hf (alloy EZ7 [26]), and (c) of Sn with Al and Hf (alloy EZ4 [26], for the nominal compositions of alloys see the Table 6). Thus, in selecting the alloy EZ5 for this study, it was expected that the Nb 3 Si would be suppressed by the synergy of Hf and Sn with Al and Ti. This was confirmed by the experimental results. Furthermore, it has also been shown that in the alloy Nb-18Si-5Hf-5Cr (alloy YG1, [25]), the synergy of Hf and Cr destabilized the Nb 3 Si either via enhancing the transformation of the latter to Nb ss and αNb 5 Si 3 or by rendering the formation of Nb 3 Si sensitive to the cooling rate, so that it could not form during solidification with the high cooling rates prevailing in the bottom of the ingot. In other words, the synergy of Cr with Sn and Hf in EZ3-AC [26] further strengthened the destabilizing effect of Sn on the Nb 3 Si. This effect was not cancelled out by the addition of Ti in the alloy EZ6. Thus, it was expected that in EZ8-AC, the formation of Nb 3 Si would be suppressed by the synergy of Sn and Hf with Ti, Al, and Cr. This also was confirmed by the experimental results.

Nb ss + Nb 5 Si 3 Eutectic
The suppression of Nb 3 Si in Nb-18Si-5Sn (alloy NV9, [39]) was accompanied with the stabilization of the Nb ss +Nb 5 Si 3 eutectic, the formation of which was attributed to the addition of Sn given that such a eutectic does not exist in the equilibrium Nb-Si binary system [38]. The Nb ss +Nb 5 Si 3 eutectic was not destabilized by the synergy of Sn with Hf in EZ1-AC [26].
The vol.% of the Nb ss + Nb 5 Si 3 eutectic was high in the alloy NV9, but with the addition of Ti in the alloy NV6, the vol.% of the eutectic was reduced [39]. Comparison of the microstructures of the alloys YG3-AC [25] and KZ3-AC (Nb-24Ti-18Si [21]) confirmed that the addition of Hf in the former stabilized the Nb 5 Si 3 , but the vol.% of Nb 5 Si 3 was very low, as was the Nb ss +Nb 5 Si 3 eutectic. The synergy of Sn and Hf with Ti in the alloy EZ2 reduced the vol.% of the eutectic compared with the alloy EZ1-AC [26]. Thus, it was concluded that when the alloying elements Ti, Hf, and Sn are in synergy in Nb-18Si based alloys, the vol.% of the Nb ss +Nb 5 Si 3 eutectic is controlled by Sn, as the synergy of Ti and Hf favors Nb 3 Si selection and the Nb ss +Nb 3 Si eutectic (alloy YG3, [25]).
The effect of different alloying elements on the formation of the Nb ss +Nb 5 Si 3 eutectic is summarized in Table 7. The addition of Al in the alloy YG2 [25] stabilized an Nb ss +Nb 5 Si 3 eutectic in all parts of the as-cast ingot (Table 7). However, the addition of Al in the alloy EZ4 [26] made the formation of the Nb ss +Nb 5 Si 3 eutectic susceptible to solidification conditions, as the eutectic was not formed in the bottom of the ingot ( Table 7). Comparison of the alloys EZ1, EZ4, and YG2 would suggest that it is most likely the synergy of Al with Sn in the presence of Hf that makes the formation of the Nb ss +Nb 5 Si 3 eutectic sensitive to cooling rate. This effect was accentuated in the alloy EZ5, in which the synergy of Al with Ti rendered the formation of the Nb ss +Nb 5 Si 3 eutectic sensitive to cooling rate during solidification ( Table 7). Comparison of EZ7-AC [26], in which a eutectic between Nb 3 Sn and Nb 5 Si 3 was formed, with the alloys EZ4-AC [26] and EZ5-AC would suggest that it was Hf that destabilized the above eutectic, and that the addition of Ti in the alloy EZ5 enhanced the role of Hf in the formation of the Nb ss +Nb 5 Si 3 eutectic. The <Si> = Si + Sn + Al content of the Nb ss +Nb 5 Si 3 eutectic was in agreement with the <Si> of eutectics in RM(Nb)ICs and RM(Nb)ICs/RCCAs [13].

Nb ss + C14-NbCr 2 Eutectic
In the alloy EZ3 [26], the synergy of Cr with Sn and Hf destabilized the Nb ss + Nb 5 Si 3 eutectic and instead resulted in the formation of a Nb ss + C14-NbCr 2 Laves phase eutectic (Table 7). In accordance with the results for the alloy EZ3 [26], no Nb ss + Nb 5 Si 3 eutectic was formed in EZ6-AC, in which the addition of Ti did not suppress the Nb ss + C14-NbCr 2 eutectic that had formed in the Ti-free alloy EZ3 [26]. Aluminum stabilizes the C14-NbCr 2 Laves phase in the Nb-Cr-Al ternary system [46]. The addition of Al in Ti-containing RM(Nb)ICs did not suppress the formation of the C14-NbCr 2 Laves phase (e.g., compare the alloy KZ4-AC with KZ5-AC in [21]), but the addition of Hf did (compare JN1-AC in [20] with KZ5-AC), and in both these alloys (i.e., KZ5 and JN1), the Nb ss +Nb 5 Si 3 eutectic formed. The additions of Hf and Sn in EZ6 did not suppress the Laves phase (compare the alloy KZ4-AC [21] with EZ6-AC), in which the Nb ss + NbCr 2 eutectic formed. Similarly, the addition of Al in EZ8 did not suppress the Laves phase and the Nb ss + NbCr 2 eutectic, and shifted its composition closer to the eutectic in EZ3-AC [26] (Table 7). In the alloys EZ6-AC and EZ8-AC, the composition of the Nb ss +NbCr 2 eutectic corresponded to that of a high-entropy eutectic (EZ6) or a complex concentrated eutectic (EZ8) [11,12]. It was concluded that in Ti-containing RM(Nb)ICs and RM(Nb)ICs/RCCAs, the synergy of Cr, Hf, and Sn promotes the stability of the C14-NbCr 2 Laves phase and the formation of Nb ss +NbCr 2 eutectic in the cast alloys, where the latter is a high-entropy eutectic or a complex concentrated eutectic that forms together with "conventional" phases (see [11]).

The Nb 5 Si 3 Silicide
Compared with the as-cast Nb-18Si-5Sn (alloy NV9 in [39]), alloying with Hf (alloy EZ1) or with Al (alloy EZ7) or with Hf and Cr (alloy EZ3, see Table 6 for the nominal compositions) shifted the composition of the Nb 5 Si 3 away from the Nb-rich corner (also see Table 2 in [34]). This effect was reduced when Hf was in synergy with Ti (alloy EZ2) or with Al (alloy EZ4) or with both Ti and Cr (alloy EZ6) and was further reduced when Hf was in synergy with both Ti and Al (alloy EZ5) or with Ti and Al and Cr (alloy EZ8) (see Table 8). In EZ6-AC, the average concentrations of alloying elements in Nb 5 Si 3 were similar to those of the same elements in the as-cast alloys EZ2, EZ3 [26], YG1, YG3 [25], and KZ4 [21]. The same was the case for Nb 5 Si 3 in EZ6-HT2 compared with the aforementioned alloys. However, in EZ6-HT2, the range of Si concentrations in Nb 5 Si 3 was wider and varied from 30.2 to 36.9 at%. In the Hf-rich Nb 5 Si 3 , the Ti and Hf concentrations were higher than those in the as-cast alloys EZ2, EZ3 [26], and YG1 [25]. After heat treatment, the Ti and Hf concentrations in the Hf-rich Nb 5 Si 3 increased compared with EZ6-AC and were higher compared with the heat-treated alloys EZ2, EZ3 [26], and YG3 [25]. It is suggested that in the presence of Hf, the dominant elements controlling the partitioning of the different solutes between the Nb 5 Si 3 phase and the melt were Ti and Al, with the latter being the most potent. Note that complex concentrated silicide [11] co-existed with "conventional" phases in EZ2-AC (see 6 and 7 in the Figure 2a). Hexagonal Nb 5 Si 3 was stable in all the alloys of this work after the heat treatment(s) ( Table 3). All the data for the Nb 5 Si 3 for all the alloys of this work gives Nb/(Ti + Hf) = 0.92, whereas this ratio is 0.97 and 0.93, respectively, for the "normal" Nb 5 Si 3 and the Hf-rich Nb 5 Si 3 ( Figure 12). This is considered to indicate that the synergy of Sn with Hf and Ti in the absence of other TMs, RMs, and metalloid elements encourages the stability of the hexagonal γNb 5 Si 3 in RM(Nb)ICs and RM(Nb)ICs/RCCAs. The Ti versus Hf and the Nb versus Ti/Hf maps for the Nb 5 Si 3 in the alloys of this work are shown in Figure 13. The Ti/Hf ratio was 1.46 when all the data for Nb 5 Si 3 were taken into account (1.2 for the Nb 5 Si 3 in EZ8), and it was 1.25 for the Hf-rich Nb 5 Si 3 . The maximum value of the Ti/Hf ratio was 3.67 for Nb = 43.65 at.%.

The Nb ss Solid Solution
Compared with EZ1-AC [26], in EZ2-AC the synergy of Ti with Hf and Sn did not affect the solubility of Si in the Nb ss and Ti-rich Nb ss , but it increased the concentration of Sn in the Nb ss by about 1.8 at.%. Thus, the Si+Sn concentration in Nb ss and Ti-rich Nb ss was about 10.3 at.% and 11.7 at.%, respectively, compared with 8.2 at.% in EZ1-AC [26]. In EZ2-HT, the Si concentration in Nb ss was reduced to the same value as in EZ1-HT1 [26].
Compared with the alloy EZ7-AC [26], in which the Nb ss was not stable, the data for the alloys EZ2, EZ4 [26], and EZ5 would suggest that the synergy of Hf with Ti (alloy EZ2), or with Al (alloy EZ4 [26]), or with Ti and Al (alloy EZ5) did not suppress the formation of Nb ss during solidification that was controlled by the Si/Sn or Si/(Sn + Al) ratios (see Table 8). However, compared with the alloys EZ2-AC and EZ4-AC [26]-in which the Si+Sn and Si + Sn + Al concentrations in Nb ss were, respectively, 10.3 at.% and 10.9 at.%-in EZ5-AC, the corresponding concentration was higher (about 15 at.%; see Table S2 in the Supplemental Data).
Compared with the alloy EZ2-AC, the Nb/Ti and Nb/(Ti + Hf) ratios in the Ti-rich Nb ss in EZ6-AC were greater than 1, and in both alloys, the Hf and Sn content increased with the Ti concentration. However, for both solutes (i.e., Hf and Sn), the increase was more significant in EZ2-AC (no Cr present). The Si content of the Nb ss was higher in EZ2-AC than in EZ6-AC, where it was the same as in EZ3-AC [26], KZ4-AC [21], and YG1-AC [25]. Thus, the data for the as-cast alloys YG1, YG3 [25], EZ2, EZ3 [26], and EZ6 would suggest that the synergy of Ti and Hf increased the Si concentration in the Nb ss . After heat treatment, there was no Ti-rich Nb ss , as was the case in the alloys EZ2, KZ4, and YG3. In EZ6-HT2, the Hf content of the Nb ss decreased, as was the case in YG3-HT and EZ3-HT [26]. Furthermore, the Sn content of the Nb ss decreased, as was the case in EZ3-HT. The concentration of Si in the Nb ss was the same as in other RM(Nb)ICs but lower than that in EZ2-HT. Compared with the alloys EZ1 [26], EZ2, and EZ3 [26], the data for the alloy EZ6 would support the conclusion that the Nb ss (and A15-Nb 3 X, see below) formation during solidification was controlled by the Si/Sn ratio (about 0.3) and the Si+Sn concentration (about 18 at.%), respectively, for the two phases (see Table 8).
The data for the Ti-free alloys EZ4 and EZ7 suggested that the stability of Nb ss was controlled by the synergy of 5 at.% Al with 5 at.% Sn. However, the Nb ss was stable in the alloys EZ6, ZX5, ZX7, and ZX8 but not in the alloys ZX6, EZ5, and EZ8. The data would suggest (i) that Hf does indeed play a role in the stability of Nb ss ; (ii) that the concentration of Sn in the alloy is important for the stability of Nb ss in RM(Nb)ICs where Al, Sn, and Ti are in synergy; and (iii) that the effect of the synergy of Al, Hf, and Sn on the stability of Nb ss in the absence/presence of Ti in the alloy is very strong and cannot be annulled with the addition of Cr.
Given that Al, Hf, and Sn are key alloying additions in RM(Nb)ICs and RM(Nb)ICs/ RCCAs to obtain a balance of properties [3], alloy design must aim to optimize the concentrations of these elements to control (a) the vol.% of Nb ss and (b) the stability of Nb ss in the alloy, due to the fact that the Nb ss is important for all three property goals, namely fracture toughness, creep, and oxidation. The alloy design methodology NICE [12] can take care of (a) and (b).
The lattice parameters of the Nb ss in the alloys of this work were given in Table 5, where data is also included for comparable alloys. The lattice parameter of the Nb ss was lower than that of pure Nb (3.303 Å), with the exception of EZ1-HT. Changes of the lattice parameter of Nb ss were attributed to the size effect of solutes (1.429, 1.462, 1.153, 1.578, 1.62, 1.432, and 1.249 Å, respectively, for Nb, Ti, Si, Hf, Sn, Al, and Cr) and to changes in the chemical composition of the Nb ss after heat treatment.
In cast alloys free of Hf and Sn addition, with the addition of Al or Cr, the lattice parameter (α Nbss ) respectively decreased (∆α Nbss < 0) and increased (∆α Nbss > 0), whereas with the simultaneous addition of Al or Cr with Hf and Sn, the lattice parameter increased, most significantly in the case of Cr (Figure 14a). In heat-treated alloys with/without Hf and Sn in synergy with Cr, the lattice parameter increased (Figure 14b). With the addition of Hf, the ∆α Nbss was positive in as-cast and heat-treated alloys with Sn, whereas in Hf-free alloys, the addition of Sn resulted to ∆α Nbss < 0 in as-cast and ∆α Nbss > 0 in heat-treated alloys ( Figure 14). The synergy of Hf and Sn simultaneously with Al or Cr decreased α Nbss in cast alloys (Figure 14a) and increased α Nbss in heat-treated alloys (Figure 14b). The synergy of Ti simultaneously with Hf and Sn in Al-and Cr-free alloys resulted in ∆α Nbss > 0, and the increase was more significant in the heat-treated condition.
It should be noted that the Nb ss of the alloys of this work was Ti-rich (or in other words, the synergy of Hf with Sn and Ti resulted in Ti-rich Nb ss ), with a minimum Ti concentration of about 31.4 at.%, for which the corresponding concentrations of Si, Sn, and Hf, respectively, were 0.8 at.%, 5.6 at.%, and 3.1 at.% (that is to say, the chemical composition of the Nb ss with the minimum Ti content was 59.1Nb-31.4Ti-0.8Si-3.1Hf-5.6Sn, with Si + Sn and Ti + Hf sum, and Si/Sn, Ti/Hf, and Nb/(Ti + Hf) ratio, respectively, of 6.4 at.%, 34.5 at.%, 0.14, 10.1, and 1.71). The data also indicated Nb/(Ti + Hf) = 0.7 for minimum concentrations of Nb and Ti + Hf in the Nb ss of, respectively, 34.6 at.% and 49.2 at.% (Figure 15a). Note that the Nb ss associated with the lamellar microstructure in EZ2-AC had similar Ti + Hf sum and Nb/(Ti + Hf) ratio (respectively 45.6 at.% and 0.95). Increasing Ti concentration in the Nb ss also increased the Hf concentration (Figure 15b), and with increasing Hf content in the Nb ss , the concentrations of Si and Sn increased by, respectively, 0.8 at.%/at.%Hf and 1.6 at.%/at.%Hf, as did the Si/Sn ratio and the Si+Sn sum, the latter by 1.96/at.%Hf (Figure 15c,d).
Considering that Hf, Si, Sn, and Ti are key elements for oxidation resistance and that refractory metals, in particular Mo and W, are key for strength and creep [1,3,11,12,33], the synergy of Hf, Sn, and Ti that promotes Ti rich Nb ss (i) improves the oxidation resistance of the Nb ss but (ii) reduces the strength and creep of the Nb ss in alloys with RM addition, owing to the relationship between the concentration of Ti and RMs in the Nb ss in RM(Nb)ICs and RM(Nb)ICs/RCCAs (e.g., see Figure S4 in the Supplemental Data in [47], Figure 12 in [48], and Figure 12 in [49]). Thus, a challenge for alloy design is to "balance" these effects. This is achievable with the alloy design methodology NICE [3,11,12].

The A15-Nb 3 X Compound
The addition of Sn promoted the formation of Nb 3 Sn in the alloys Nb-18Si-5Sn (NV9) and Nb-24Ti-18Si-5Sn (NV6) [39], where the Nb 3 Sn was present in all parts of the ingot of each alloy. However, the addition of Hf in the alloys EZ1 [26] and EZ2, despite the fact that it did not completely destabilize the A15-Nb 3 X, essentially rendered its formation sensitive to cooling rate, probably due to the fact that Hf affected the Si+Sn and Si/Sn values that control its formation. On the other hand, Al addition not only promoted the formation of the A15-Nb 3 X (EZ1 vs. EZ7 [26]) but actually reversed the effect that Hf had on it, by stabilizing the formation of the A15-Nb 3 X to most parts of the as-cast ingots and also by increasing the vol.% of this phase (EZ1 vs. EZ4 [26], EZ2 vs. EZ5), indicating that its effect on the partitioning of Sn (and thus the Si + Sn + Al and Si/(Sn + Al) values) between the solid phases and the melt was stronger than that of Hf. As a matter of fact, the Si + Sn + Al concentration in the Al-containing alloys EZ4, EZ5, and EZ7 was higher by about 2 at.% (see Tables 8 and 9). A similar effect to that of Al was seen with the addition of Cr in the alloys Nb-18Si-5Cr-5Hf-5Sn (EZ3 vs. EZ1 [26]) and Nb-24Ti-18Si-5Cr-5Hf-5Sn (EZ6 vs. EZ2), the presence of which resulted in the stabilization of the A15-Nb 3 X in all areas of the as-cast ingots (Table 9). Thus, it was expected that when Al and Cr were present simultaneously, as in the case of the alloy EZ8, the A15-Nb 3 X would be stabilized in all parts of the ingot. This was confirmed by the experimental results for EZ8-AC. Furthermore, the increased vol. % of A15-Nb 3 X in EZ8-AC was in agreement with the results for the alloys EZ1 to EZ7 [26]. Table 9. Effect of the different alloying elements on the formation and composition (at.%) of the A15-Nb 3 X compound in various as-cast RM(Nb)ICs.

Alloy Designation and Nominal Composition
It should be noted that for the alloys of this work, the minimum concentration of Ti in the A15-Nb 3 X (about 24.9 at.%) corresponded to Si = 4.9 at.%, Sn = 11.26 at.%, that is, the chemical composition of the A15-Nb 3 X with the minimum Ti content was 58.9Nb-24.9Ti-4.9Si-11.3Sn, with Si+Sn sum and Si/Sn ratio of, respectively, 15.2 at.% and 0.43, in good agreement with the Si, Sn, and Ti concentrations in the cast alloys EZ5, EZ6, and EZ8. Figure 16 shows correlations between Hf, Si, Sn, and Ti concentrations in A15-Nb 3 X compounds in the alloys of this work. Note the similar trends exhibited by the alloys EZ2 and EZ6 (green and red data) and the alloys EZ5 and EZ8 (brown and blue data), and remember the addition of Cr in EZ6, and that Al was present in both EZ5 and EZ8. In all the alloys, as the Hf content in A15-Nb 3 X increased, that of Ti decreased (Figure 16a), but the trends of Hf vs. Sn and Ti vs. Sn were opposite for the alloys EZ5 and EZ8 (as Hf increased, the Sn content increased too (Figure 16b), and as Ti decreased, the Sn content increased (Figure 16c)) and the alloys EZ2 and EZ6 (Figure 16b,c). Note also that the Sn and Si+Sn content of the A15-Nb 3 X in the latter alloys was higher than in the former (Figure 16b-d), but with the Si+Sn concentration in a very narrow range (18.2 to 18.8 at.%, Figure 16d). Furthermore, the Nb concentration of the A15-Nb 3 X decreased significantly with increasing Ti/Hf ratio in the Al-free alloys (Figure 16e).

C14-NbCr 2 Laves Phase
The <Cr> = Cr + Al + Si + Sn content of the C14-NbCr 2 Laves phase in the cast and heat-treated alloys of this work was in agreement with other work on RM(Nb)ICs [13,50]. Figure 17 shows the correlation between VEC and Cr content of the C14-NbCr 2 Laves phase and the average atomic size <R>, with the ratio of the average atomic sizes of elements that substitute Nb or Cr in the Laves. The two correlations are in agreement with [13].

Vol.% of Phases
Correlations of the vol.% of phases in RM(Nb)ICs with/without Ti addition with VEC alloy are shown in the Figure 18. Notice (i) that Al, Cr, Hf, Si, Sn, and Ti are key alloying additions in RM(Nb)ICs for improving oxidation and for controlling strength, fracture toughness, and creep (e.g., [32,33,[51][52][53][54]); (ii) that the vol.% of Nb ss and Nb 5 Si 3 is key for fracture toughness, creep, and oxidation resistance, of which the former and the latter two have a tendency to increase and decrease with increasing vol.% Nb ss , respectively, and to decrease and increase with increasing vol.% of Nb 5 Si 3 (e.g., [12,32,55]); (iii) that the C14-NbCr 2 Laves phase can improve the oxidation of RM(Nb)ICs but at the expense of their fracture toughness (e.g., [32,56]); (iv) that the A15-Nb 3 X is key to improving the oxidation of RM(Nb)ICs in the range of temperatures of pest oxidation and at high temperatures [22,23,27]; and (v) that according to NICE, in order to meet the oxidation or creep goal, VEC alloy should decrease and increase, respectively [3,12]. In addition, note that the black arrow in the Figure 18a-c shows that the Ti effect is consistent with the improvement of oxidation resistance of RM(Nb)ICs owing to the corresponding reduction of VEC alloy [1,3,11,12,50].  Figure 18. Correlation of the volume fraction of the phases in RM(Nb)ICs with/without Ti addition with the parameter VEC of the alloys that were studied in [26] and this work: (a) Nb ss , (b) Nb 5 Si 3 , (c) A15-Nb 3 X, and (d) C14-NbCr 2 Laves phase. Colors and symbols as follows: red for as-cast alloys, blue for HT alloys, circles for Ti-free alloys, and triangles for Ti-containing alloys. In (a-c), the black arrow indicates the effect of Ti addition. In (a,d), the dotted lines are given to "guide the eye".
Notice (a) that the vol.% of Nb ss , Nb 5 Si 3 , A15-Nb 3 X, and C14-NbCr 2 Laves decreased and increased after heat treatment for the former two and the latter two phases, respectively (Figure 18a-d); (b) that similar trends were exhibited by the data for Nb 5 Si 3 and A15-Nb 3 X in the alloys with/without Ti addition; (c) that up to about (i) 30% Nb ss could be stable without Al and Cr addition in the alloy EZ2 (Figure 18a), and (ii) 40% Nb 5 Si 3 and 7% C14-NbCr 2 Laves could be stable in the alloy EZ8 with simultaneous Al and Cr addition (Figure 18b,d); (d) that the vol% of A15-Nb 3 X significantly increased after heat treatment in Cr-free alloys with Al addition and with/without Ti addition (Figure 18c), and that the aforementioned volume fractions correlate with low VEC alloy values, which is an essential requirement for improved oxidation resistance according to NICE [1,3,12]. In other words, the data would suggest that with Ti, Hf, Si, Sn, and Al or Cr additions, the oxidation of RM(Nb)ICs and RM(Nb)ICs/RCCAs (note that the alloy EZ8 is also a RM(Nb)IC/RCCA alloy) can be improved owing to the decrease of VEC alloy . It should be noted (i) that improved oxidation resistance in the pest temperature range was confirmed for all the heattreated alloys of this work, which did not pest and did not suffer from scale spallation; and (ii) that in oxidation at 1200 • C of the alloys with Al and/or Cr addition, their scales spalled off, similarly to other Sn-containing RM(Nb)ICs [22,23] (oxidation data is not included in this paper).

Partitioning of Solutes and Solidification of the Alloys
The synergy of Hf and Sn in RM(Nb)ICs with/without Ti addition not only made the characterization of the microstructures difficult (this work and [26]) but also promoted (i) different types of Nb 5 Si 3 in the microstructure (meaning tetragonal α or βNb 5 Si 3 or hexagonal γNb 5 Si 3 , Table 3) and/or (ii) microstructures that arise from phase transformations that involve the two key phases, namely Nb 5 Si 3 and Nb ss . The melt solidified following different solidification paths in different parts of the ingot.
Three different microstructures were observed in EZ5-AC owing to the sensitivity of the solidification of this alloy to the partitioning of solutes in different parts of the solidifying melt. The βNb 5 Si 3 was the primary phase. The microstructure in the bulk of EZ5-AC consisted of primary βNb 5 Si 3 surrounded by Nb ss +Nb 5 Si 3 eutectic and the A15-Nb 3 X. As the primary βNb 5 Si 3 formed, the melt became poorer in Si and richer in Ti, and when the melt concentration reached Si + Al + Sn of about 23 at.%, the eutectic formed. As the melt became poorer in Si, the Si + Al + Sn concentration reached about 20.6 at.%, and the A15-Nb 3 X formed. It is suggested that the solidification path in the bulk of EZ5-AC was L → L + βNb 5 Si 3 → L + βNb 5 Si 3 + γNb 5 Si 3 + [Nb ss + Nb 5 Si 3 ] eutectic → βNb 5 Si 3 + γNb 5 Si 3 + [Nb ss + Nb 5 Si 3 ] eutectic + A15-Nb 3 X + αNb 5 Si 3 . The Si+Al+Sn concentration of the Nb 5 Si 3 and the Hf-rich Nb 5 Si 3 was about 37.6 at% and 38.5 at.%, respectively. These values were very close to the ones observed in the bulk of the alloy EZ4-AC [26]. The Ti addition in EZ5 had a significant effect on the partitioning of Hf, and the concentrations of Ti and Hf increased by, respectively, about 8.1 at.% and 6.3 at.% in the Hf-rich Nb 5 Si 3 , leading to Nb/(Ti + Hf) about 0.7, which indicated hexagonal γNb 5 Si 3 [24]. The Ti addition in EZ5 also affected the composition of the Nb ss , in which the average Si + Al + Sn concentration was about 15.9 at.%, that is, about 5 at.% higher than in the Nb ss in the alloy EZ4-AC [26], and Si/(Al + Sn) about 0.2 (the same as in the alloy EZ4-AC, see Table 8). The Sn content in A15-Nb 3 X was also affected by the addition of Ti, and in the bulk of EZ5-AC, it was about 4 at.% higher than in the alloy EZ4-AC.
In the top of EZ5-AC, the microstructure consisted of primary Nb 5 Si 3 surrounded by the Nb ss . There were Hf-rich areas in the βNb 5 Si 3 , with the core exhibiting a darker contrast owing to the lower Hf concentration. As the primary βNb 5 Si 3 formed, the melt became poorer in Si + Al + Sn, and when the latter concentration reached about 16 at.% and Si/(Al + Sn) about 0.2, the Nb ss formed. It is suggested that the solidification path in the top of EZ5-AC was L → L + βNb 5 Si 3 → βNb 5 Si 3 + Nb ss → βNb 5 Si 3 + Nb ss + αNb 5 Si 3 . In the bottom of EZ5-AC, the microstructure consisted of the Nb 5 Si 3 , Nb ss , and A15-Nb 3 X phases. The βNb 5 Si 3 was the primary phase. It is suggested that the solidification path in this part of the ingot was L → L + βNb 5 Si 3 + γNb 5 Si 3 → L + βNb 5 Si 3 + γNb 5 Si 3 + A15-Nb 3 X + Nb ss → βNb 5 Si 3 + γNb 5 Si 3 + αNb 5 Si 3 + A15-Nb 3 X + Nb ss .
Two different microstructures were observed in EZ8-AC owing to the sensitivity of the solidification of this alloy to the partitioning of the solutes in the different areas of the ingot. The alloy EZ8 was hypereutectic in the top and bulk of the ingot and solidified with the βNb 5 Si 3 as its primary phase. As the primary βNb 5 Si 3 formed, the melt became richer in Ti and Hf, and Hf-rich Nb 5 Si 3 formed. The solubility of Cr and Sn in particular, and of Al in the Hf-rich Nb 5 Si 3 , was small (Table S4 in the Supplemental Data). Thus, the melt surrounding the βNb 5 Si 3 became richer in Al, Cr, and Sn and leaner in Si and Hf. When the composition of the melt reached Si + Al + Sn about 20.9 at.%, the A15-Nb 3 X formed. The Hf and Si were rejected into the melt, which became rich in these elements. When the Si/(Sn+Al) ratio in the melt reached about 0.2, the Nb ss formed. Then, the melt became richer in Si, and when its composition reached Cr + Si + Al + Sn about 49 at.%, a eutectic between the NbCr 2 Laves phase and the Nb ss grew. The aforementioned eutectic formed in between these phases as the solutes partitioned between the solidifying intermetallics and the solid solution. The presence of Nb ss , αNb 5 Si 3 , γNb 5 Si 3 , and Hf-rich Nb 5 Si 3 in EZ8-AC is in agreement with [16]. The C14-NbCr 2 Laves phase was not observed in [16] owing to the low Cr concentration in the Nb-25Ti-16Si-8Hf-2Al-2Cr-xSn (x = 2 to 8 at.%) alloys. The A15-Nb 3 X was not observed in the alloy with x = 5 at.% Sn, but was stable after the heat treatment [16].
Owing to the partitioning of solutes, a lamellar microstructure was formed in the bulk of EZ2-AC and microstructure A in EZ6-HT2. The chemical composition of the former corresponded to high-entropy or complex concentrated lamellar microstructure (see Section 3.1) and coexisted with "conventional" phases [11]. The chemical composition of microstructure A and the bright contrast phase (BCP, see Section 3.8) also corresponded to complex concentrated microstructure [11]. These microstructures are considered further in the next two sections.

Eutectic and Lamellar Microstructures in EZ2-AC
The case of the Nb ss + Nb 5 Si 3 eutectic and the lamellar microstructures observed only in the bulk of EZ2-AC (Figures 1 and 2) requires attention. The chemical composition of said phases will be taken into account. In EZ2-AC, there was Ti-rich Nb ss , similarly to the alloy YG3-AC (see Table 6 for nominal composition), but in EZ2-AC, the concentration of Ti in the Nb ss was higher than that of Nb (i.e., the Ti-rich Nb ss in EZ2 had Nb/Ti < 1), and the concentrations of Sn and Hf in the Nb ss increased with the Ti concentration. It should be noted that Ti-rich Nb ss was not observed in the alloy NV6-AC (Nb-24Ti-18Si-5Sn [39]), which would suggest that in the presence of Sn, the synergy of Ti and Hf had a strong effect on the partitioning of Ti to the Nb ss .
In EZ2-AC, the Ti and Hf concentrations in the Nb 5 Si 3 were similar to those in YG3-AC [25]. In the latter alloy, the Hf-rich Nb 5 Si 3 corresponded to hexagonal γNb 5 Si 3 according to its Nb/(Ti + Hf) ratio, whereas in EZ2-HT, the Hf-and Ti-rich tetragonal Nb 5 Si 3 that was already present in EZ2-AC became richer in both Ti and Hf after the heat treatment, and according to its Nb/(Ti + Hf) ratio (0.7), it was hexagonal γNb 5 Si 3 [24]. In EZ2-AC, the average Si+Sn concentration in Nb 5 Si 3 was 38.0 at% and did not change significantly in EZ2-HT (38.5 at.%), similarly with EZ1-HT [26]. The Ti addition in EZ2 did not change the solubility of Sn in the silicide, which was similar to that in EZ1-AC, but increased the Hf concentration. In EZ2-AC, in the Hf-rich Nb 5 Si 3 , the Ti concentration increased by about 4.6 at.% compared with the "normal" Nb 5 Si 3 , but the Si+Sn, Si, and Sn concentrations did not change.
It was suggested [25] that the Hf-rich Nb 5 Si 3 that formed in YG3-HT was the product of the eutectoid transformation tP32 Nb 3 Si → Nb ss + Hf-rich (hP16) Nb 5 Si 3 and that the synergy of Ti and Hf in YG3 led to the replacement of the eutectoid transformation tP32 Nb 3 Si → Nb ss + (tI32) αNb 5 Si 3 with the alternative eutectoid phase transformation given above, in which the Nb 5 Si 3 had the hexagonal (hP16) structure instead of the tetragonal (tI32) one [24]. As the Nb 3 Si was not formed in EZ2, the above phase transformations cannot account for the presence of γNb 5 Si 3 and αNb 5 Si 3 .
In EZ2-AC, there was tetragonal βNb 5 Si 3 and αNb 5 Si 3 according to the XRD data ( Figure S1 in the Supplemental Data). The synergy of Ti and Sn in the alloy NV6 (Nb-24Ti-18Si-5Sn) enhanced the transformation βNb 5 Si 3 → αNb 5 Si 3 [39]. The formation of αNb 5 Si 3 in EZ2-AC was attributed to the presence of both Ti and Sn in the alloy. Furthermore, the addition of Ti in EZ2 decreased the liquidus of the alloy, and thus, as the homologous temperature increased, the solute diffusivities increased during solid-state cooling. In other words, in EZ2-AC, the βNb 5 Si 3 was the primary phase and the βNb 5 Si 3 → αNb 5 Si 3 phase transformation had started during the cooling of the ingot. Figure 2 shows (a) a lamellar microstructure that grew into a blocky faceted Nb 5 Si 3 grain (see microstructure on the left of number 17 in Figure 2b, and note that outer parts of some Nb 5 Si 3 were richer in Hf with lower Ti/Hf ratio compared with the inner darker contrast parts, Figure 1a,b), with a lamellar microstructure (numbers 1, 2, 3, and 4 in Figure 2a) that connected with (was adjacent to) Nb 5 Si 3 grains (numbers 6 and 7 in Figure 2a).
The lamellar microstructure (i) was significantly richer in Ti and Hf than the Nb ss +Nb 5 Si 3 eutectic in EZ2-AC and (ii) had higher and lower Ti + Hf sum and Nb/(Ti + Hf) ratio, respectively, than the eutectic. Furthermore, the Nb 5 Si 3 associated with the lamellar microstructure (e.g., numbers 6 and 7 in Figure 2a) was (iii) richer in Ti and Hf and had tetragonal structure like the tetragonal, and poorer in Ti and Hf, "normal" Nb 5 Si 3 . In addition, the lamellar microstructure and the Nb 5 Si 3 associated with it had chemical composition consistent with high-entropy or complex concentrated phases [11]. Moreover, the Nb ss near the lamellar microstructure (e.g., numbers 11, 12, 15, 17, and 18 in Figure 2) was (iv) richer in Ti and Hf compared with the Nb ss away from the lamellar microstructure, and (v) had a Nb/(Ti + Hf) ratio close to that corresponding to the minimum Ti + Hf and Nb concentrations in the Nb ss (Figure 15a). Chemical analysis of the individual phases in the lamellar microstructure was not possible. Figure 19a shows the average solute concentrations from Nb 5 Si 3 to the lamellar microstructure to the Nb ss and Figure 20a shows the ratios or sums of solutes from Nb 5 Si 3 to the lamellar microstructure to the Nb ss . From the tetragonal Nb 5 Si 3 to the lamellar microstructure, (a) the Ti concentration increased, the Hf essentially did not change, and thus the Ti + Hf and Ti/Hf increased; (b) the Nb decreased slightly, and thus the Nb/(Ti + Hf) essentially did not change; and (c) the Si content decreased significantly, and the Sn increased slightly, and thus the Si+Sn and Si/Sn decreased. From the lamellar microstructure to the Nb ss , (d) the Ti and Hf increased and decreased slightly, respectively, while the Ti + Hf decreased slightly but the Ti/Hf continued to increase; (e) the Nb increased significantly, but the change of Nb/(Ti + Hf) was marginal; and (f) the increase of Sn and decrease of Si continued, and thus the Si + Sn sum and the Si/Sn ratio continued to decrease, the latter less than the former.  In other words, owing to the partitioning of solutes between the Nb ss and Nb 5 Si 3 , in particular the partitioning of Ti and Hf, a phase transformation started from the Ti-and Hf-rich complex concentrated tetragonal Nb 5 Si 3 that resulted in a complex concentrated lamellar microstructure, namely the eutectoid transformation tetragonal Nb 5 Si 3 → Nb ss + [Nb 5 Si 3 ] Hf and Ti rich . The key solutes in this transformation were Hf and Ti. In the lamellar microstructure (i) the contrast of the Nb ss was similar to that of the nearby solid solution, meaning the solid solution was Ti and Hf rich, and (ii) the contrast of the Nb 5 Si 3 was similar or slightly brighter than that of the silicide adjacent to the lamellar microstructure. Brighter contrast Nb 5 Si 3 means that it is richer in Hf silicide, which could be hexagonal γNb 5 Si 3 depending on its Nb/(Ti + Hf) ratio [24]. The synergy of Hf and Ti with Sn in the alloys of this work promoted the γNb 5 Si 3 ( Figure 12). Thus, it is suggested that the Hf and Ti silicide in the lamellar microstructure was hexagonal γNb 5 Si 3 , that is, the aforementioned eutectoid phase transformation was Hf-and Ti-rich tetragonal αNb 5 Si 3 → Nb ss + γNb 5 Si 3 with orientation relationships between Nb and αNb 5 Si 3 and Nb and γNb 5 Si 3 [29,57]. Now consider the lamellar microstructure on the left of number 17 in Figure 2b. It is suggested that the transformation nucleated at the interface between the Hf-and Ti-rich Nb 5 Si 3 and the Ti-and Hf-rich Nb ss (note that the Ti-rich Nb ss was also richer in Hf than the "normal" Nb ss ). The transformation front moved in a direction from the top right-hand corner to the bottom left-hand corner, and it would have stopped when the chemical composition of the Nb 5 Si 3 grain was similar to that of the dark blocky Nb 5 Si 3 .
The lamellar microstructures were not stable in EZ2-HT, which would suggest that the proposed phase transformation had been completed after 100 h at 1500 • C and the hexagonal Hf-rich Nb 5 Si 3 was a stable phase in EZ2, in agreement with the XRD data ( Figure S1 in the Supplemental Data).
Eutectic and lamellar microstructure similar to those in EZ2-AC were also observed in the alloy NV1-AC (Nb-23Ti-5Si-5Al-5Hf-5V-2Cr-2Sn, [31]). In NV1-AC, the Nb 5 Si 3 that was associated with the lamellar Nb ss + Nb 5 Si 3 had Ti + Hf = 35 at.% and Nb/(Ti + Hf) = 0.67, and was CC hexagonal Nb 5 Si 3 , whereas the lamellar microstructure, the composition of which corresponded to a CC lamellar microstructure [11], had Ti + Hf about 30 at.%, and Nb/(Ti + Hf) about 1.3. In EZ2-AC, the Nb 5 Si 3 with Nb/(Ti + Hf) = 1 and Ti + Hf = 31.1 at.% (CC tetragonal Nb 5 Si 3 ) was associated with a lamellar microstructure that was very Ti + Hf-rich (= 47.4 at.%) and had Nb/(Ti + Hf) = 0.6. Thus, comparison of the cast alloys NV1 and EZ2 shows that the Nb 5 Si 3 silicides associated with the lamellar microstructure were different, CC hexagonal in NV1-AC and CC tetragonal in EZ2-AC, and the chemical compositions of the CC lamellar microstructures were also different. The eutectic in EZ2-AC has Ti + Hf = 31.9 at.% and Nb/(Ti + Hf) = 1.4, essentially the same as the lamellar in NV1-AC. The lamellar microstructure in NV1 was the product of a eutectic reaction plus some eutectoid transformation [31].

Microstructure A in EZ6-HT
In EZ6-AC, the Nb 5 Si 3 was surrounded by Nb ss . The contrast of the Hf-rich Nb 5 Si 3 and the A15-Nb 3 X in EZ6-HT2 was very similar. The microstructure A formed adjacent to some Nb 5 Si 3 grains but did not surround the whole grain (Figure 8b-d). The Nb 5 Si 3 adjacent to microstructure A had Ti + Hf = 19 at.% and Nb/(Ti + Hf) = 2.25 (i.e., was tetragonal Nb 5 Si 3 ), and microstructure A had Ti + Hf = 38.3 at.% and Nb/(Ti + Hf) = 0.66. A very thin bright contrast phase (BCP) was formed at the interface of microstructure A with A15-Nb 3 X (Figure 8c). This was actually bright contrast hexagonal silicide (BCHS). After the BCHS was the A15-Nb 3 X, and then the C14-NbCr 2 Laves and the Nb ss (Figure 8b-d). In the EZ6-HT2, the Hf-rich Nb 5 Si 3 formed separate (distinct) grains from the "normal" Nb 5 Si 3 (Figure 8b). The Hf-rich Nb 5 Si 3 had Ti + Hf = 39.1 at.%, and Nb/(Ti + Hf) = 0.56, which were similar to those (a) of the BCHS of microstructure A (40.4 at.% and 0.46, respectively) and (b) the microstructure A (38.3 at.% and 0.66, respectively). The chemical compositions of microstructure A and the BCHS corresponded to complex concentrated phases. Figure 19b shows the average solute concentrations, and Figure 20b shows the ratios or sums of solutes from Nb 5 Si 3 to the microstructure A, to the BCHS, to the Nb 3 Sn, to the NbCr 2 , and to the Nb ss . From the tetragonal Nb 5 Si 3 , (a) the concentration of Hf increased to A, and then to the BCHS, whereas the concentration of Ti increased to A and then decreased slightly, and the concentration of Nb decreased to A and then to the BCHS, an thus the Nb/(Ti + Hf) and Ti/Hf ratios decreased to A and then to the BCHS, whereas the Ti + Hf sum increased markedly to A and then slightly to the BCHS; (b) the Si concentration decreased to A and then increased to the BCHS, that is, it exhibited the opposite trend compared with Ti, and the Sn concentration decreased after A to the BCHS, and thus both the Si+Sn sum and the Si/Sn ratio decreased to A and increased to the BCHS, with a remarkable increase of the Si/Sn ratio. The tetragonal Nb 5 Si 3 was Cr free, and the Cr concentration increased slightly to A and then decreased to the BCHS. Thus, as microstructure A formed and the solutes partitioned between the Nb ss and Nb 5 Si 3 , the microstructure became rich in Hf, Si, and Ti, and the BCHS formed.
From the BCHS to the A15-Nb 3 X, (c) the concentrations of Nb, Ti, Sn, and Cr increased, and those of Hf and Si decreased, and thus the Ti + Hf and Si + Sn sums and the Si/Sn ratio decreased, and the Ti/Hf and Nb/(Ti + Hf) ratios increased. On the other hand, (d) from the A15-Nb 3 X to the C14-NbCr 2 Laves phase, the concentrations of Cr, Si, and Hf increased, and those of Nb, Ti, and Sn decreased, and thus the Ti + Hf and Si + Sn sums and the Ti/Hf ratio decreased, the Si/Sn ratio increased, but the Nb/(Ti + Hf) essentially did not change. In addition, (e) from the Laves phase to the Nb ss , the concentrations of Nb, Ti, and Sn increased, and those of Cr, Si and Hf decreased, and thus the Si/Sn ratio and the Si + Sn sum decreased, but the Ti + Hf sum and the Ti/Hf ratio increased, and the Nb/(Ti + Hf) ratio increased slightly. In other words, where the A15-Nb 3 X was stable, the microstructure was poorer in Hf and Si and richer in Nb, Sn, and Ti, but still poor in Cr compared with the BCHS. In the Cr-, Hf-, and Si-rich areas of the microstructure, the Laves phase was stable, and in the Nb-, Sn-, and Ti-rich areas, the Nb ss was stable.
The microstructure A formed in areas of EZ6-HT2 where the Nb ss was adjacent to the tetragonal Nb 5 Si 3 prior to the heat treatment. The average composition of this Nb ss in EZ6-AC was 45.6Nb-32.4Ti-1.8Si-8.7Cr-4.6Hf-6.9Sn, with Si + Sn = 8.7 at.%, Si/Sn = 0.26, Ti/Hf = 7, Ti + Hf = 37 at.%, and Nb/(Ti + Hf) = 1.23, compared with Si + Sn = 8.6 at.%, Si/Sn = 0.28, Ti/Hf = 7.6, Ti + Hf = 33.7 at.%, and Nb/(Ti + Hf) = 1.5 for the Nb ss far away from the Nb 5 Si 3 in EZ6-AC. In other words, the Nb ss around tetragonal Nb 5 Si 3 grains on some parts of which the microstructure A had formed was rich in Ti + Hf at the start of HT. In EZ6-HT2, the average chemical composition of the Nb ss away from microstructure A was 55.1Nb-31.1Ti-0.5Si-7.1Cr-2.3Hf-3.9Sn, that is, the solid solution was poor in Si, and its Hf concentration was half that in the cast alloy, and it had Si + Sn = 4.4 at.%, significantly lower than the Nb ss in the cast alloy; Si/Sn = 0.28, essentially the same with the cast alloy; Ti/Hf = 13.3, significantly higher compared with the cast alloy; Ti + Hf = 33.4 at.%, essentially the same as in the cast alloy; and Nb/(Ti + Hf) = 1.65, higher than the cast alloy. The data would suggest that partitioning of Hf, Si, Sn, and Ti was essential in the transformation that occurred in EZ6-HT2 and led to the formation of microstructure A. It is suggested (a) that the microstructure A was the product of the phase transformation [Nb ss ] Ti+Hf rich, Si+Sn rich → [Nb ss ] Si+Sn poor, Hf poor + [Nb 5 Si 3 ] Ti and Hf rich and (b) that owing to the promotion of the γNb 5 Si 3 by the synergy of Hf and Ti with Sn in the alloys of this work (Figure 12), the [Nb 5 Si 3 ] Ti and Hf rich in the above phase transformation was γNb 5 Si 3 .
Comparison of the solute concentrations from the tetragonal Nb 5 Si 3 to the lamellar microstructure in EZ2-AC and microstructure A in EZ6-HT2 shows similar trends for Nb, Si, and Ti and opposite trends for Hf and Sn. The lamellar microstructure in EZ2-AC and microstructure A in EZ6-HT2 were rich in Ti + Hf (47.4 at.% in EZ2-AC and 38.3 at.% in EZ6-HT2) and essentially had the same Nb/(Ti + Hf) ratio (0.6 in EZ2-AC and 0.66 in EZ6-HT2) (note also that the Ti + Hf sum and the Nb/(Ti + Hf) ratio of Hf-rich Nb 5 Si 3 in NV1-AC and EZ6-HT2 were similar (35 at.% and 0.67 in NV1-AC [31] and 39.1 and 0.56 in EZ6-HT2)).
In EZ6-HT2, the growth of microstructure A was away from the silicide, and was "stopped" by the very Hf-rich BCHS (18. In the heat-treated alloys EZ5 and EZ8, in which the Nb ss was not a stable phase (Table 3), the A15-Nb 3 X precipitated in Nb 5 Si 3 grains (Figures 4 and 11). Both alloys contained Al and Ti. In the Ti-free and Al-containing alloys EZ4 and EZ7 [26], where the Nb ss also was not a stable phase, no precipitates were observed in Nb 5 Si 3 . Thus, the precipitation of A15-Nb 3 X in EZ5-HT and EZ8-HT was attributed to the synergy of Al and Ti in said alloys. Precipitation of Nb ss in Nb 5 Si 3 has been reported by our research group in the Al-and Ti-containing but Sn-free heat-treated RM(Nb)ICs KZ2 (Nb-24Ti-18Si-8Cr-4Al), KZ5, and KZ7 [21], where it was attributed to the phase transformation βNb 5 Si 3 → Nb ss + αNb 5 Si 3 .
Modelling of the properties of metallic UHTMs, for example their toughness and creep (e.g., [56,58,59]), should not ignore the presence of microstructure A and/or precipitates of a second phase that can form respectively around or inside the Nb 5 Si 3 after exposure to high temperature(s), as well as the change of properties at the interface between Nb ss and Nb 5 Si 3 [31].

Hardness
The hardness of alloys and their phases as a function of alloy or phase parameters and volume fraction of phases is considered in Figures 21 and 22. The hardness and roomtemperature-specific strength calculated from hardness (σ y = HV/(3ρ), where HV is alloy hardness and ρ is alloy density, see Table 1, [60]) of the alloys of this work increased with decreasing VEC alloy , in agreement with [1]. Figure 21c shows (1) that there is a correlation between the alloy parameter VEC and the hardness of Nb 5 Si 3 (note that also there are relationships between VEC alloy and VEC Nb5Si3 and VEC alloy and ∆χ Nb5Si3 for RM(Nb)ICs and RM(Nb)ICs/RCCAs, see [3]) and (2) that the hardness of the Nb 5 Si 3 increased with the parameter VEC alloy for the alloys of this work.    In (e), the R 2 value is for the data for the alloys EZ2 and EZ6, where the Nb ss is a stable phase. Colors as follows: blue alloy EZ2, red EZ5, brown EZ6, and green EZ8. In (a-f), the dotted line is to "guide the eye".
The hardness of the A15-Nb 3 X increased with its parameter ∆χ and was higher for the alloys EZ5 and EZ8 (Figure 21d). The hardness of the Nb ss increased with its parameters δ and ∆χ (Figure 21e,f). The trend in Figure 21f is in agreement with previous work for RM(Nb)ICs and RM(Nb)ICs/RCCAs [3]. Compared with the alloy EZ2, the addition of Al and/or Cr in EZ5, EZ6, and EZ8 increased the hardness significantly. The hardness was highest for the as-cast alloys EZ5, EZ6, and EZ8.   In (d), R 2 = 0.5292 for the as-cast alloys and R 2 = 0.9874 for the heat-treated alloys. In (b), the R 2 value is for the alloys where the Nb ss is a stable phase. Colors as follows: blue alloy EZ2, red EZ5, brown EZ6, and green EZ8. In (a,c), the dotted line is given to "guide the eye". Note the following: (i) the remarkable fit of data in Figure 21d, where the trend is in agreement with [13]; (ii) for the alloys where the Nb ss was not stable, namely the alloys EZ5 and EZ8, the hardness of the Nb ss was essentially similar, even though the δ Nbss and ∆χ Nbss changed (Figure 21e,f); and (iii) the δ Nbss was highest for the Nb ss in EZ8-AC. The specific strength of the alloys was in the range 271.7 to 416.5 MPa cm 3 g −1 . The specific strength of the alloys EZ5 and EZ8 (i.e., the two Al-containing alloys of this work), was higher than the room-temperature-specific strength of RCCAs and RHEAs reviewed in [2]. Furthermore, the specific strength of EZ8 was higher than that of B containing RM(Nb)ICs and RM(Nb)ICs/RCCAs [1,36,61].
The hardness of the alloys increased and decreased with increasing vol.% of A15-Nb 3 X and Nb ss , respectively ( Figure 22). Note that the vol.% Nb ss increased with increasing ∆χ Nbss . The trend of the hardness of the AC and HT alloys versus their vol.% Nb 5 Si 3 exhibited, respectively, a minimum and maximum, which correspond to about 45% and 40% Nb 5 Si 3 , respectively. The alloy EZ8-HT, in which the Nb ss was not stable and the Nb 5 Si 3 had the lowest hardness of the alloys of this work (Table 4), had the highest hardness and room-temperature-specific strength with vol.% Nb 5 Si 3 = 40%, vol.% A15-Nb 3 X = 53%, and with VEC alloy = 4.435 and ∆χ A15-Nb3X = 0.905. 4.2.14. Comparison of the Synergy of Hf with B, Ge, or Sn in RM(Nb)ICs/RCCAs In this section, we compare the effects of the synergy of solute additions, namely Al, Cr, Hf, Si, and Ti (group A elements), with each of the metal/metalloid elements B, Ge, or Sn (group B elements), which the alloy designer can use to design/select metallic UHTMs with a balance of properties. The comparison uses data for the alloys EZ8, ZF9 [35], and TT7 [36] (see Table 10). Note that all three alloys are RM(Nb)ICs/RCCAs and are based on the RM(Nb)IC alloy KZ5 (Nb-24Ti-18Si-5Al-5Cr [21]), with Hf plus one of the group B elements, namely Hf plus B or Hf plus Ge or Hf plus Sn addition, respectively, in TT7, ZF9, and EZ8. Table 10. Comparison of the RM(Nb)ICs/RCCAs EZ8, ZF9 (38Nb-24Ti-18Si-5Al-5Ge-5Hf-5Cr [35,63]), and TT7 (38Nb-24Ti-17Si-5Al-6B-5Cr-5Hf [36]). ness of the Nb ss increased with its parameters δ and ∆χ. The room-temperature-specific strength of the alloys was in the range 271.7 to 416.5 MPa cm 3 g −1 , and for EZ5 and EZ8, was higher than that of RCCAs and RHEAs. The effect of the synergy of Hf and Sn, or Hf and B or Hf and Ge on the macrosegregation of solutes, microstructures, and properties of RM(Nb)ICs/RCCAs was compared.
Comparison of the microstructures of the alloys of this work with those of alloys studied previously would suggest (i) that in RM(Nb)ICs and RM(Nb)ICs/RCCAs with Al, Cr, Hf, Si, Sn, and Ti addition, phase transformations that employ the Nb 3 Si silicide to engineer the microstructure cannot be used; (ii) that in Nb-18Si based RM(Nb)ICs where Hf, Sn, and Ti were in synergy, the vol.% of the Nb ss + Nb 5 Si 3 eutectic was controlled by Sn; (iii) that the synergy of Al with Sn in the presence of Hf made the formation of the Nb ss +Nb 5 Si 3 eutectic sensitive to solidification conditions; (iv) that in the presence of Hf, the dominant elements that controlled the partitioning of solutes between the Nb 5 Si 3 and the melt were Al and Ti, of which the former was the most potent; (v) that the synergy of Sn with Hf and Ti promoted the stability of the hexagonal Nb 5 Si 3 ; and (vi) that regarding the stability of Nb ss , (a) Hf was part of the cause, (b) the concentration of Sn in the alloy was important in alloys where Al, Sn, and Ti were in synergy, and (c) the effect of the synergy of Al, Hf, and Sn with/without Ti on the stability of the Nb ss could not be reversed with the addition of Cr.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ma15134596/s1. Table S1: EPMA data (at%) of the as-cast and heat-treated alloy EZ2, Table S2: EPMA data (at%) of the as-cast and heat-treated alloy EZ5, Table  S3: EPMA data (at%) of the as-cast and heat-treated alloy EZ6, Table S4: EPMA data (at%) of the as-cast and heat-treated alloy EZ8, Figure S1: X-ray diffractograms of the as-cast and heat-treated alloy EZ2, Figure S2: X-ray diffractograms of the as-cast and heat-treated alloy EZ5, Figure S3: X-ray diffractograms of the as-cast and heat-treated alloy EZ6, Figure S4: X-ray diffractograms of the as-cast and heat-treated alloy EZ8. Funding: This work was supported by the University of Sheffield, Rolls-Royce plc, and EPSRC (EP/H500405/1, EP/L026678/1).

Informed Consent Statement: Not applicable.
Data Availability Statement: All the data for this work is given in the paper, other data cannot be made available to the public.