Composition dependence of the specific heat of FeSi

We wish to thank the Referee for their effort made to review our paper. They made us aware of an important issue in our analysis that has been corrected in our revised manuscript. We hope that the referee is satisfied with the modifications and recommends the revised manuscript for publication in SciPost.

Regarding the weaknesses pointed out by the Referee, we reply: Following revision of the data analysis (see requested changes), we find no sample dependence of the linear specific heat coefficient γ within experimental accuracy. We want to point out that a quantitative analysis of the precise compositions of samples in the concentration range studied is, to the best of our knowledge, currently not possible. For instance, both energy-dispersive x-ray spectroscopy and lab-based powder x-ray diffraction did not allow us to resolve the iron concentrations in our samples with the required accuracy.

Possible insights may be accessible by virtue of the direct determination of defect concentrations, most notably vacancies and antisite disorder. For instance, we have determined the concentration of point defects in the compositional series Mn1+xSi, an isostructural sibling of FeSi, using positron annihilation spectroscopy in combination with ab-initio calculations [M. Reiner et al., Sci. Rep. 6, 29109 (2016)]. The feasibility of positron-lifetime spectroscopy has been reported on a single composition of FeSi [A. Bharathi et al., Phys. Rev. B 55, R13385 (1997)]. However, a comprehensive determination of point defect concentrations using positrons is well beyond the scope of the work reported here. We have added a corresponding discussion to the manuscript.

Regarding the requested changes by the Referee, we reply: We wish to thank the Referee for this remark. Further measurements of the heat capacity confirmed the presence of a substantial contribution of Apiezon N grease, as suggested by the Referee. In turn, we repeated all of our specific heat measurements, carefully determining the specific heat contribution of the grease prior to the measurements on each sample. We find that for all samples investigated, the specific heat tracks each other above ~20 K. At low temperatures, a shallow maximum is observed, consistent with the literature [S. Paschen et al., Phys. Rev. B 56, 12916 (1997)]. This maximum may be attributed to a Schottky anomaly, is suppressed under magnetic field, and decreases in size with increasing iron content.

In the revised manuscript, we have updated the account of the experimental methods. Moreover, we have exchanged the figures and rewritten the paragraphs describing and discussing our experimental data. As our main conclusion, we identify a magnetic field and composition dependence of the low-temperature specific heat of FeSi that contrasts the expected behavior of simple magnetic impurities.

The minor mistakes pointed out by the Referee have been corrected.