Dynamical mean-field theory investigation of specific heat and electronic structure of $\alpha$- and $\delta$-plutonium

We have carried out a comparative study of the electronic specific heat and electronic structure of $\alpha$- and $\delta$-plutonium using dynamical mean-field theory. We use the perturbative $T$-matrix and fluctuating exchange as a quantum impurity solver. We considered two different physical pictures of plutonium. In the first, $5f^5+$, the perturbative treatment of electronic correlations has been carried out around the nonmagnetic [local-density approximation (LDA)] Hamiltonian, which results in an $f$ occupation around a bit above $n_f=5$. In the second, $5f^6-$, plutonium is viewed as being close to a $5f^6$ configuration, and perturbation theory is carried out around the $(\mathrm{LDA}+\mathrm{U})$ starting point bit below $n_f=6$. In the latter case, the electronic specific-heat coefficient $\gamma$ attains a smaller value in $\delta \text{−}\mathrm{Pu}$ than in $\alpha \text{−}\mathrm{Pu}$, in contradiction to experiment, while in the former case, our calculations reproduce the experimentally observed large increase of $\gamma$ in $\delta \text{−}\mathrm{Pu}$ as compared to the $\alpha$ phase. This enhancement of the electronic specific-heat coefficient in $\delta \text{−}\mathrm{Pu}$ is due to strong electronic correlations present in this phase, which cause a substantial increase of the electronic effective mass, and high density of states at $E_F$. The densities of states of $\alpha$- and $\delta$-plutonium obtained starting from the open-shell configuration are also in good agreement with the experimental photoemission spectra.

Figure 1

(Color online) The total (solid line) and partial $f$ state (dashed line) densities of states of $\alpha$ (left panel) and $\delta$ (right panel) plutonium in the LDA, $\mathrm{LDA}+\mathrm{FLEX}$, $\mathrm{LDA}+\mathrm{U}$, and $\mathrm{LDA}+\mathrm{U}+\mathrm{FLEX}$ approaches, respectively.

Figure 2

(Color online) The imaginary part of self-energy ${\Sigma }_{mm}\left(i\omega \right)$ for $m=-3$ on the Matsubara axis, computed for $\alpha$- and $\delta \text{−}\mathrm{Pu}$ within the $\mathrm{LDA}+\mathrm{FLEX}$ and $\mathrm{LDA}+\mathrm{U}+\mathrm{FLEX}$ techniques.

Figure 3

(Color online) The theoretical $\mathrm{LDA}+\mathrm{FLEX}$ and experimental PES (Ref. 14) for $\alpha$- and $\delta \text{−}\mathrm{Pu}$.