关注：

1)  U参数的加入

2） 自旋极化的考虑

3） 铁磁、反铁磁的考虑

来自文章的计算方法介绍

The similar MSUs of Pu di- and trihydride provide aframework within which intermediate compositions can be exploredcomputationally.

  The calculationspresented here are limited to unit cells with four Pu sites and between eightand twelve H atoms.

     Initial unit cells correspond to the two MSUsshown in Fig. 1(a) and in Fig. 1(b) (where the chevron shape is made into anactual unit cell) with added and removed H atoms, respectively.

     These initialcon_gurations often relax surprisingly slow with false plateaus, which requirerather small convergence criteria to overcome (see below).

   These unitcells limit the current study to structures with AB and ABC stacking. A systematic study of whether the Pu hydrides favor otherstacking sequences will likely shed more light on the system but also require signifficantlymore computational effort.

   Thecalculations discussed here neglect thermal effects. Contributions to the free energy from thephonons in particular are important, given the small mass of the H atoms. 【氢原子的热振动对能量影响比较大】

     The currentresults do reveal many small energy differences betweenmagnetic states and various conffigurations of the H atoms, which eventhe zero-point energy will affect.

   However, in additionto being beyond the scope of this work, the calculation of phonons and including their thermal effectswill not likely overcome the inaccuracy inherent in the density functional theory(DFT) method.

   Connectionsmade between the results and experimental data do involve thermal effects inthe form of comparisons with the energy scale set by room temperature, but themain conclusions involve large enough energy differences to be immune to neglectedcontributions from the phonons.【忽略声子贡献，可以？】

   The resultspresented here originate in DFT calculations using the VASP package.17,18

    The calculations make use of the generalized gradientapproximation (GGA) of Perdew, Burke, and Ernzerhof.19 The Pu(5f⁶, 6d^8?,7s²) and H(1s) electrons are treated in the valence withprojector-augmented wave potentials.     20

    The calculations employ the linear tetrahedron methodwith Blochl corrections,21 ak-point mesh of density 60 _A􀀀1, and an energy cuto_ of 500eV.  

     Theself-consistent cycles are converged to within 10^-5meV to enable the 10^-4meV ionic stopping criteria needed to overcome the falseplateaus mentioned above. The calculations allow spin polarization and compare twotypes of magnetic structure, ferromagnetic (FM) and antiferromagnetic (AFM). 【如何开展此项计算？】

   Because DFT inthe GGA fails to describe PuH3 as a semiconductor,6 the effects of including either strong electroncorrelation and spin-orbit coupling are investigated.

    Some results reported here stem from calculations thattreat the on-site Coulomb repulsion between 5felectrons with aHubbard parameter U (GGA+U) in the rotationally invariant form of Dudarev et al.22

    In this form the Hubbard parameter U and theexchange parameter J appear only in the difference U -J, throughoutthis report the difference is simply referred to as U.

   A single valuefor U (4 eV) for all stoichiometries allows a comparison of thecalculated energies and the evaluation of formation energies.

     While Ai et al. suggestspin-orbit coupling (SOC) can be neglected,6 the results presented hereshow some energy differences small enough for SOC to matter.

网络摘录：

1 http://210.75.240.149/home.php?mod=space&uid=574584&do=blog&id=471831

发信人: valenhou (AAA), 信区: Gaussian
 标 题: Re: 有人用VASP算过磁性体系吗？
 发信站: BBS 大话西游站 (Tue Jun 3 09:43:22 2003), 转信

 你说的是对的。计算体系的磁性，第一步就是要让ISPIN=2。如果要进一步考虑具体的
 磁序，在设置MAGMON等等参数。
 如果只指定了ISPIN=2,而没有去设置MAGMON，则这种情况是计算体系的铁磁性，此时
 每个原子的MAGMON都相等，默认值为1。

 【 在 hsg221 (升仙啦) 的大作中提到: 】
 是不是就是说，如果我算一个体系，不是反铁磁也不是亚铁磁，就直接
 ISPIN=2就可以了，不用管MAGMON？

 【 在 valenhou (AAA) 的大作中提到: 】
 : 你想哪几个原子指定有初始磁矩，值可正可负，单位为波尔磁矩ub。
 : 一般设置MAGMON是在考虑反铁磁或亚铁磁情况下。

网络摘录2

SPIN  =      2  
   MAGMOM = 0 0 4 0 0 4 0 0 4 0 0 4 0 0 4
   SAXIS = 0 0 1
这是五个原子的磁矩都向z轴的设置   你照着弄下

网络摘录3  http://blog.sciencenet.cn/blog-72871-255382.html

如何用VASP计算铁磁、反铁磁和顺磁

来自：VASP手册

MAGMOM= [real array]

Specifies the initial magnetic moment for each atom, if and only if  ICHARG=2,  or if the CHGCAR file contains no magnetisation density (ICHARG=1).

   If one is searching for a spin polarised (magnetic or antiferromagnetic) solution,  it is usually safest to start from larger local magnetic moments, because in  some cases, the default values might not be sufficiently big. A save default is usually the experimental magnetic moment multiplied by 1.2 or 1.5. It is important to emphasize that the MAGMOM tag is  used only, if the CHGCAR file holds no information on the magnetisation density, and if the initial charge density is not calculated from the orbitals supplied in  the WAVECAR file.

This means that the MAGMOM tag is useful for two kind of calculations

• Calculations starting from scratch with no WAVECAR and CHGCAR file.

• Calculations starting from  a non magnetic WAVECAR and CHGCAR file (ICHARG=1). Often such calculations converge more reliably to the desired magnetic configuration than calculations of the first kind. Hence, if you have problems to converge to a desired magnetic solution, try to calculate first the non magnetic groundstate, and continue from the generated WAVECAR and CHGCAR file. For the continuation job, you need to set   ISPIN=2 ICHARG=1 in the INCAR file.

   

Starting from VASP.4.4.4, VASP also determines whether the magnetic moments  supplied in the MAGMOM line break the symmetry. If they do, the corresponding symmetry operations are removed  and not applied during the symmetrization of charges and forces. This means that antiferromagnetic calculations can be performed by  specifying anti-parallel magnetic moments for the atoms in the cell MAGMOM = 1 -1 As an example consider AF bcc Cr with the POSCAR file: Cr: AF 2.80000 1.00000 .00000 .00000 .00000 1.00000 .00000 .00000 .00000 1.00000 2 Kartesisch .00000 .00000 .00000 .50000 .50000 .50000 With the MAGMOM line specified above, VASP should converge to the proper groundstate. In this example, the total net magnetisation  is matter of factly zero, but it is possible to determine the local magnetic moments by using the RWIGS or LORBIT tags (see sections 6.346.33).