No abstract

La(Fe, Si)13-based compounds are well-known magnetocaloric materials, which show a pronounced negative thermal expansion (NTE) around the Curie temperature but have not been considered as NTE materials for industrial applications. The NaZn13-type LaFe13-xSix and LaFe11.5-xCoxSi1.5 compounds were synthesized, and their linear NTE properties were investigated. By optimizing the chemical composition, the sharp volume change in La(Fe, Si)13-based compounds was successfully modified into continuous expansion. By increasing the amount of Co dopant in LaFe11.5-xCoxSi1.5, the NTE shifts toward a higher temperature region, and also the NTE operation-temperature window becomes broader. Typically, the linear NTE coefficient identified in the LaFe10.5Co1.0Si1.5 compound reaches as much as -26.1 × 10(-6) K(-1), with an operation-temperature window of 110 K from 240 to 350 K, which includes room temperature. Such control of the specific composition and the NTE properties of La(Fe, Si)13-based compounds suggests their potential application as NTE materials.

No abstract

Zhao et al. Reply In the first sentence of the Comment [1], the authors refer to four systems: GaNð10 10Þ-1H, ZnOð10 10Þ-1H, GaNð10 10Þ-Li, and ZnOð10 10Þ-1Li. In the second sentence, the authors claim to demonstrate that "a charge-density-wave (CDW) phase is due to the artifact of the generalized gradient approximation (GGA), while the antiferromagnetic (AFM) ground state is predicted by the hybrid DFT calculation…" The Comment contains a single table, showing the total energy of the GaNð10 10Þ-1H system only. No numerical results were presented for the other three systems. It is not clear how the authors arrived at their very general conclusion.We have checked the validity of Kim et al.'s claim by performing new calculations using two popular hybrid functionals, HSE06 [2,3] and B3LYP [4], on the four systems mentioned in the Comment. These two hybrid functional methods are generally considered to be at the same level of accuracy. All hybrid functional calculations are performed on a 12-layer slab separated by a 12 Å vacuum. The two-dimensional k grid of the Brillouin zone is 8 × 6 for the pð1 × 1Þ cell and the same density is used for larger cells. The results are listed below.Based on these results, we draw a few observations: (i) For the GaNð10 10Þ-1H system, different hybrid functionals produce different ground states. The conclusions are not consistent. Our Table I shows that the B3LYP hybrid method favors the CDW state by 0.21 eV per pð2 × 2Þ unit cell, and that this preference is reversed with the HSE06 hybrid method, which instead favors the AFM state by 0.06 eV. This is similar to Kim's HSE results, which also favored the AFM state by 0.07 eV. Importantly, Kim et al. also perform RPA calculations, which favor the AFM state by 122 meV per pð2 × 2Þ unit cell in GaN-1H. We have confirmed this value for RPA. For the GaNð10 10Þ-1Li system, however, both hybrid functionals B3LYP and HSE06 favor the CDW ground state, by 0.5 and 0.25 eV, respectively. For GaNð10 10Þ-1Li, we have also carried out RPA calculations [5-8] using a four-layer slab. The RPA results favor the CDW state over the AFM state by 0.16 eV. Thus, for GaNð10 10Þ-1Li, all results using B3LYP, HSE06 and RPA support our findings [9], and disagree with the claim of Kim et al. (ii) For ZnOð10 10Þ with 1H and 1Li, both HSE06 and B3LYP functionals favor the CDW ground state, according to our new results shown