钚多相物性的DFT+Gutzwiller方法研究

Due to the strongly correlated effect of 5f electrons, plutonium has been a special material with many novel properties not well understood, which had been classical puzzles for condensed matter physics and strongly correlated field all along. Dynamical mean field theory (DFT+DMFT) got the best results of plutonium studies currently, but it is computationally expensive, so it can't be employed to study the low-symmetry phases of plutonium. DFT+Gutzwiller is a newly developed method for correlated systems, it's quality for ground state is as high as that by DMFT, and yet it is computationally much cheaper, so DFT+G is a suitable method for study the properties of the low-symmetry phases of plutonium. In addition, DFT+G method is fully variational, the charge-density self-consistency can be naturally achieved, and the quantities, such as total energy, can be accurately obtained. The basic idea of our project is that we based on a general interactant model improve DFT+G method first, and then employed it to study the physical properties of multi-phase plutonium as well as phase transition. The mainly research work is as follows: (1)DFT+G method based on a general interactant model; (2) Fix local electronic number model; (3)Physical properties of multi-phase plutonium; (4) Phase transition of plutonium.

钚由于5f电子的关联性，呈现出众多奇异的性质，目前的理论研究还无法全面解释这些特性，这一直是凝聚态物理和强关联领域的经典难题。当前，动力学平均场方法（DFT+DMFT）获得了钚研究的最好结果，但该方法计算量大，在对称性较低的钚晶相结构中无法应用。DFT+Gutzwiller方法（DFT+G）是新兴的关联体系方法，对基态的计算与DMFT方法精度相当而计算量小很多，因此适用于钚的低对称性相结构的模拟。并且DFT+G方法是完全变分法，可以很自然的实现电子密度自洽，从而精确的获得总能等物理量。本项目研究思路是，首先基于广义相互作用模型改进现有的DFT+G方法，然后运用该方法研究钚的多相物性及相变。主要研究内容如下：（1）基于广义相互作用模型的DFT+G方法；（2）固定局域电子数计算模式；（3）钚的多相物性；（4）钚的相变。

结合一种平均场势方法，把LDA+Gutzwiller (LDA+G)方法推广到有限温度情形。发展了LDA+G方法的一种固定局域电子数计算模式，增强了程序的收敛性。采用LDA+G方法研究赝α-Pu、γ-Pu、δ-Pu、δ´-Pu和ε-Pu的总能随体积变化曲线，其中赝α-Pu能量最低，是热力学稳定相，该结果与实验符合；这些相能量相对变化的顺序与常压下钚随温度升高的相变顺序一致，说明了结果的合理性。计算了δ-Pu和ε-Pu的平衡体积和体弹模量，与实验符合，优于LDA/GGA等方法的计算结果。研究δ-Pu到ε-Pu的含温相变，首次获得与实验符合的相变温度、压力，以及两相相图。研究钚的一种相似材料铈的α-γ同构相变，发现电子熵和声子熵都对相变有重要贡献，并得到了与实验符合的相图。