A Study of the Effects of Hf and Sn on the Microstructure, Hardness and Oxidation of Nb-18Si Silicide Based Alloys without Ti Addition

The paper presents the results of an experimental study of large (≈0.6 kg) arc melted buttons of four Ti free Nb-silicide based alloys with Sn addition with nominal compositions (at.%) Nb-18Si-5Hf-5Sn (EZ1), Nb-18Si-5Al-5Sn (EZ7), Nb-18Si-5Cr-5Hf-5Sn (EZ3) and Nb-18Si-5Al-5Hf-5Sn (EZ4). The alloys were studied in the as-cast and heat treated conditions. In all the alloys there was macrosegregation of Si (MACSi). Among the single element additions Hf had the weakest and Sn the strongest effect on MACSi. The simultaneous presence of Cr and Hf in the alloy EZ3 had the strongest effect on MACSi. In all the alloys the βNb5Si3 was the primary phase and was present after the heat treatment(s), the Nb3Si silicide was suppressed and the A15-Nb3Sn intermetallic was stable. The Nbss was not stable in the alloys EZ7 and EZ4 and the C14-NbCr2 Laves phase was stable in the alloy EZ3. Very Hf-rich Nb5Si3 was stable in the alloy EZ4 after prolonged heat treatments. Eutectics were observed in all the alloys. These were binary eutectics in the alloys EZ1 and EZ7, where respectively they consisted of the Nbss and βNb5Si3, and βNb5Si3 and A15-Nb3Sn phases. Most likely ternary eutectics consisting of the Nbss, C14-NbCr2 and βNb5Si3, and Nbss, βNb5Si3 and A15-Nb3Sn phases were observed, respectively in the alloys EZ3 and EZ4. The addition of Al increased the vol% of the Nb5Si3 and A15-Nb3Sn phases, particularly after the heat treatment(s). The lattice parameter of Nb respectively increased and decreased with the addition of Hf, and Al or Cr and the latter element had the stronger negative effect. Pest oxidation was not suppressed in the alloys of this study.

0.0-0.5 0.2±0.1 * Only two analyses were possible owing to the size of this phase.

On the Nb-Si-Sn liquidus projection
The liquidus projection for the Nb-Si-Sn system has areas for βNb5Si3 and αNb5Si3 [1]. At 1974 o C it gives the invariant reaction L1 + βNb5Si3  L2 + αNb5Si3 for the transformation of βNb5Si3 to αNb5Si3, where L1 and L2 result from a miscibility gap in the Si-Sn binary, and at 1874 o C gives the invariant reaction L  (Nb) + A15 + αNb5Si3 [1]. Unfortunately, the paper by Sun et al [1] is inconsistent with the experimental work reported in [2] for the alloy Nb-18Si-5Sn (alloy NV9), the results of which were used by Sun et al. According to Sun et al (i) "the divorced eutectic of Nb5Si3 + A15_Nb3Sn" was reported in the ref.
[7] in their paper and (ii) "the fine eutectic was roughly identified as (Nb) + Nb5Si3 binary eutectic in samples from top and centre of the ingot, while it was identified as (Nb) + Nb5Si3 + A15_Nb3Sn in sample from the bottom of the ingot [7]" (ref. [7] in Sun et al is the paper by Vellios and Tsakiropoulos, which is the reference [2]). Vellios and Tsakiropoulos [2] did not report a divorced eutectic of Nb5Si3 + A15-Nb3Sn and did not identify the eutectic in the bottom of the ingot as (Nb) + Nb5Si3 + A15_Nb3Sn. Instead, they reported that only the Nbss + Nb5Si3 eutectic was observed in all parts of the as-cast alloy. Figure S1 shows the microstructure of the alloy NV9 in the as-cast condition (Figures S1a and S1b), together with analysis data for the indicated phases and areas. Figures S1a and S1b clearly show that the lamellar microstructure consisted of the Nbss and Nb5Si3 phases. Figure S1 is given to highlight how difficult can be the identification of Nbss and Nb3Sn in the microstructures of Nb-silicide based alloys. This difficulty increases further when Hf is present in the alloy owing to the partitioning of the element between the phases.
Referring to ref.
[7] in their paper, Sun et al also stated "since a small fraction of Nb3Sn was found between Nb5Si3 and (Nb) in the 1200 o C, 1500 o C and 1600 o C heat treated samples, it seems from the non-equilibrium as-cast microstructure, this evidence may lend certain support to (Nb) + Nb5Si3 + A15_Nb3Sn ternary eutectic in the as-cast alloy, and it is more reasonable to identify the fine eutectic as (Nb) + Nb5Si3 + A15_Nb3Sn ternary eutectic in the cast microstructures" [1]. Figure S1c shows the microstructure of the alloy NV9 after the heat treatment at 1500 o C for 100 h, together with analysis data for the indicated phases. Figure S1c clearly shows that after the heat treatment the prior lamellar microstructure areas still consisted of the Nbss and Nb5Si3 phases.
The XRD data in [2] for the as-cast alloy NV9 had only one peak that corresponded only to αNb5Si3, 6 peaks that corresponded to other phases and αNb5Si3, and 4 peaks that corresponded to βNb5Si3 and other phases. Peaks were shared between the βNb5Si3 and other phases in the XRD data for the heat treated alloy NV9 at 1200, 1500 and 1600 o C, and one peak corresponded only to βNb5Si3 in the diffractograms for the 1200 and 1600 o C heat treatments. Sun et al [1] stated "for the as-cast and heat treated NV9 alloy, Nb5Si3 including the structures of both αNb5Si3 and βNb5Si3 were indicated by XRD results in Vellios et al.'s work, but no particular peaks of βNb5Si3 were found when we carefully checked the XRD results, so the identification of βNb5Si3 was not considered credible. However, the particular peaks of αNb5Si3 confirmed its presence, and the existence of αNb5Si3 is more reasonable to coincide with Waterstrat et al.'s work and its stability in the Nb-Si binary, especially for the heat treated samples, so the identification of αNb5Si3 is accepted for this modelling work" (the reference to Waterstrat et al is for the partial isothermal section of the Nb-Si-Sn system at 1600 o C which is given in ref. [6] in [1]. Waterstrat and Muller did not give a liquidus projection for Nb-Si-Sn in their paper (ref.6] in [1]).