计算声子谱计算考虑lo-to分裂,bron怎么设置

[转载]声子计算方法
已有 1069 次阅读
|个人分类:|系统分类:|关键词:声子方法|文章来源:转载
&有三个方法:Linear Response TheoryFrozen Phonon methodFinite Displacement methodLinear response 方法(或者称为 density perturbation functionaltheory,DFPT),直接计算出原子的移动而导致的势场变化,再进一步构造出动力学矩阵。这种方法在计算谱时,Borneffective charge(对极性的材料)和声子谱都能计算出。现在很流行,CASTEP,Quantum-Expresso,Abinit等等都在用,后面两个原理差不多;Linear Response优点:晶体晶胞没有大小限制,即使用包含一个原子的Primitive cell,计算得到的DynamicalMatrix也是很准确的,主要原理:Hellmann-Feymann theorem and Perturbation theory,原子施加很小的位移,计算波函数,电子密度对位移的响应函数,主要方法见Gonze1997年的两个PRB文章。计算速度一般,特别是采用NormalConserving PPs的时候,单原子晶胞RAM占用量在3-4G之间,CASTEP里面只支持NCPP的Linearresponse 计算,USPPs不支持。另外一个是Finite displacement (直接的方法): 构造超原胞,把原子移动一下,计算原胞中所有原子所受的力(这个根据体系的周期性,要多移动几个原子),然后根据这个力构造力常数矩阵。而且一般情况下对LO-TO的split不能计算出(只有在计算了Born effective charge之后,进一步考虑了non-analyticityterm,才能计算出)。phonon,phonony等就是结合vasp或者其他计算软件如wien2k等计算声子谱的。RAM占用量和计算量在Cell一样的情况下,可节约2倍的RAM和CPU时间,但这个方法最大的缺点是需要生成一个Supercell来获得比较可靠的力常数,虽然鉴于力常数是短程作用,在最邻近原子以外衰减很快,但所需要的Supercell大小也很大,一般截止半径大小是4A以上,对于金属这个半径可能会小一下,因为金属的电子Coulomb屏蔽很显著,但对于其他的晶体结构,以及晶体结构较复杂的体系,这种方法自动生成的Cell一般都包含100原子以上,基本上没有人能采用单机计算Phonon,如BCC, Ba元素,Primitive cell只包含一个原子,%BLOCK LATTICE_CART-2.672 2.674 2.6732.672 -2.673 2.6732.672 2.672 -2.672%ENDBLOCK LATTICE_CART%BLOCK POSITIONS_FRACBa 0.0 0.0000在采用有限位移方法计算声子时晶胞是:%BLOCK PHONON_SUPERCELL_MATRIX2 0 00 2 00 0 2即2*2*2的supercell,里面包含8个Ba原子,采用AMD Dual Core,2GRAM,计算需要2h左右即可完成。另外一个最大的优点是可以用Ultra soft Pseudo potentials, 这个可以极大的节约时间,减小kineticenergy curoff数值。Finite Displacement方法只计算Brillouin Zone G点的NormalModes,其他k点的Dynamical Matrix利用Fourier Transofrmation得到,C(k)=SumC(R)*Exp(-ikR),只要Cell足够大,可以获得和Linear Response 一样可靠的DynamicalMatrix。Linear Response计算有带隙的晶体最好,也是最省事的方法,但计算金属,采用Finiteelement方法最好,Linear response对金属体系基本上失效的。可能原因:Perturbation theory本身对于金属不成立(金属能隙太小);Fermi面Smearing方法本身对计算力常数不利;(目前有几个常用的Smeaing方法,Gaussan函数,或者有限温度下的FermiDirac函数)后者基本上可以排除,采用Linear response,同事采用NCPP+fixoccupation的方法计算得到的Phonon和NCPP+ Smearing方法是一样子的,因此可以推断是LinearResponse theory对metal不适用,CASTEP小组在其网页上也指出Lnear Responsetheory对Magnetic和Metal不适用。因此CASTEP不支持金属体系NCPP+LinearResponse计算,也是有原因的。采用Quantum-Expresso计算PHONON,可以完全得到与CASTEP一致的结论,即LinearResponse不适用于金属体系。下面给出Na的例子,有实验数据,势函数计算结果,NCPP+ Linearresponse,USPP+Finite element结果,可以看到Linearresponse精度连势函数都不如,数值完全是错误的&&声子计算的几种方法:转自:Practicalschemes for phonon calculations (见castep说明)A good review of the existing schemes can be found in Baroni et al.(2001). The theoretical study of phonon properties has to rely onone of the three available methods for determining the forceconstants matrix: analytical calculations, supercell calculationsor linear response calculations. The analytical approach is onlyviable when the energy model is sufficiently simple to allow adirect evaluation of the second derivatives of the energy withrespect to atomic displacements (e.g., empirical pair potentialmodels). Therefore, it is unsuitable for first principlescalculations. Further alternatives such as extracting vibrationalproperties from molecular dynamics runs (Arias et al. 1992) areless transparent and noticeably more expensive.The supercell method involves perturbing the positions of the atomsslightly and calculating the reaction forces (Ackland et al. 1997).It is necessary to use supercells of the original cell wheninteratomic interaction in the system is long ranged. The mainadvantage of this method (and of the closely related frozen phonontechnique) is that there is no need any totalenergy scheme like CASTEP can be used to evaluate the forces at anumber of carefully selected distorted configurations. The originalfrozen phonon scheme requires a displacement with the given wavevector and has been successfully used since the early 1980s (Yinand Cohen 1982, Ho et al. 1984). The force constants matrixevaluation in this formalism has been used to calculate interplanarforce constants (Wei and Chou 1994) and thus phonon dispersionalong high symmetry directions. More recent applications are basedon the full reconstruction of the force constants matrix (Acklandet al. 1997, Parlinski et al. 1997, and references in Baroni et al.2001).Linear response calculations seek to evaluate the dynamical matrixdirectly for a set of q vectors. The starting point of the linearresponse approach is evaluation of the second-order change in thetotal energy induced by atomic displacements. The main advantage ofthe scheme is that there is no need to artificially increase thecell size in order to accommodate small values of the q vectors, asin the frozen phonon method, or to overcome the long rangeinteraction problem (force constants matrix from supercellcalculations). A more detailed description of the linear responsemethod can be found in Baroni et al. 2001. The CASTEPimplementation is described in the Linear Response topic.第一性原理计算声子方法及常见程序:一,直接法:直接法,或称frozen-phonon方法,是通过在优化后的平衡结构中引入原子位移,计算作用在原子上的Hellmann-Feynman力,进而由动力学矩阵算出声子色散曲线。用该方法计算声子色散曲线最早开始于80年代初。由于计算简便,不需要特别编写的计算程序,很多小组都采用直接法计算材料性质。直接法的缺陷在于它要求声子波矢与原胞边界(supersize)正交,或者原胞足够大使得Hellmann-Feynman力在原胞外可以忽略不计。这使得对于复杂系统,如对称性高的晶体、合金、超晶格等材料需要采用超原胞。超原胞的采用使计算量急剧增加,极大的限制了该方法的使用。这种方法不能很好的预言LO-TOsplitting,只有在计算了Born effective charge和dielectric constant之后,进一步考虑了non-analyticity term,才能计算出;但Direct Method本身并不能给出Born effectivecharge和dielectricconstant.所以这也是它的一个缺陷.目前,vasp+phonon用的就是这种方法.vasp+phonon(或者PHON或者fropho)VASP能计算声子谱的都是采用一种直接的方法:构造超原胞,把原子移动一下,计算原胞中所有原子所受的力(这个根据体系的周期性,要多移动几个原子),然后根据这个力构造力常数矩阵。1,PHONON Softwareby Krzysztof PARLINSKIPhonon is a software (see list of Publications) for calculatingphonon dispersion curves, and phonon density spectra of crystals,crystals with defects, surfaces, adsorbed atoms on surfaces, etc.from either a set of force constants, or from a set ofHellmann-Feynman forces calculated within an ab initio program (notincluded). One can use VASP, Wien2k, MedeA of Materials Design ,Siesta, or other ab initio code which is able to optimize asupercell and calculate the Hellmann-Feynman forces. Phonon buildsa crystal structure, using one of the 230 crystallographic spacegroups, finds the force constant from the Hellmann-Feynman forces,builds the dynamical matrix, diagonalizes it, and calculates thephonon dispersion relations, and their intensities. Phonon findsthe polarization vectors, and the irreducible representations(Gamma point) of phonon modes, and calculates the total and partialphonon density of states. It plots the internal energy, freeenergy, entropy, heat capacity and tensor of mean squaredisplacements (Debey-Waller factor). Phonon finds the dynamicalstructure factor for the coherent inelastic neutron scattering andthe incoherent doubly differential scattering cross section for asingle crystal and polycrystal. For polar cystals the LO/TO modesplitting can be included.Homepage:2,PHONA program to calculate phonons using the small displacementmethodThis program calculates force constant matrices and phononfrequencies in crystals. From the frequencies it also calculatesvarious thermodynamic quantities, like Helmholtz free energy,entropy, specific heat and internal energy of the harmonic crystal.The procedure similar to the one described in Ref. [1], i.e. isbased on the small displacement method. It needs a code capable tocalculate forces on the atoms of the crystal.Homepage:E-mail: Telephone: +44 (0)20 Fax: +44 (0)20 3,frophois the open source implementation of the frozen phononmethod.Function:Phonon band structurePhonon DOS (Vibrational spectra)Thermal propertiesMulliken notation assignment of vibration modefropho is the frozen phonon analyzer mainly for first principles(ab initio) calculation. Periodic boundary condition is assumed.fropho gives good combinations with VASP code or another codeswhich can derive Hellmann-Feynman forces.Homepage:Download: Contact: Authour: Atsushi Togo二,DFPT方法:1987年,Baroni、Giannozzi和Testa提出了一种新的晶格动力学性质计算方法--微扰密度泛函方法(DensityFunction PerturbationTheory)。DFPT通过计算系统能量对外场微扰的响应来求出晶格动力学性质。该方法最大的优势在于它不限定微扰的波矢与原胞边界(supersize)正交,不需要超原胞也可以对任意波矢求解。因此可以应用到复杂材料性质的计算上。此外,能量对外场微扰的响应不仅可以推导出声子的晶体性质,还能求出弹性系数、声子展宽、拉曼散射截面等性质,这种方法本身就能算出Borneffective charge dielectric constant,可以很好的预言LO-TO splitting甚至Kohnanomalies。这些优势使得DFPT一经提出就被广泛应用到了半导体、金属和合金、超导体等材料的计算上。比较常用的程序是pwscf和abinit,castep等采用的是一种linearresponse theory 的方法(或者称为 density perturbation functionaltheory,DFPT),直接计算出原子的移动而导致 的势场变化,再进一步构造出动力学矩阵。来自:.cn/s/blog_65c112ca0101cb2m.html
本文引用地址:&此文来自科学网巩朋来博客,转载请注明出处。
上一篇:下一篇:
当前推荐数:0
评论 ( 个评论)
作者的其他最新博文
热门博文导读
Powered by
Copyright &您所在位置: &
&nbsp&&nbsp&nbsp&&nbsp
pt掺杂to&,2&光吸收性能的第一性原理研究.pdf56页
本文档一共被下载:
次 ,您可全文免费在线阅读后下载本文档。
文档加载中...广告还剩秒
需要金币:200 &&
你可能关注的文档:
··········
··········
摘 要 Ti02作为一种环境材料,由于具有无毒、光催化活性高、物理化学性能稳定、
价廉等特点,在环境污染治理、太阳能综合利用等方面正发挥着越来越重要的作
用。但Ti02是一种宽带隙半导体材料,它的光催化活性通常需要由紫外光线来
激发,由此造成其不能直接利用太阳光中的可见光部分,使其应用过程出现不便,
且提高了使用成本。尽管Ti02可见光化研究已经取得了一定的进展,对Ti02的
各种改性方法或多或少都提高了太阳能的利用率,但从目前的研究成果来看,可
见光催化或能量转换效率还普遍偏低,且对各种改性方法的可见光化机理认识尚
未统一,还存在争议。 针对以上问题,本研究尝试用一种有效方法:利用金属Pt离子掺杂改性,
改变Ti02的能带结构,使掺杂后Ti02的光吸收阈值最大限度地拓宽到可见光区, 即掺杂后Ti02能在可见光范围内发生电子跃迁,使其能吸收太阳能中的可见光
部分。本文以密度泛函理论为基础,利用第一性原理计算软件CASTEP软件,
对本征Ti02及掺杂Ti02进行模拟计算,从理论上探讨掺杂前后Ti02的电子结构
和光学性能,并从能带结构、态密度、分波态密度、介电函数、反射光谱、吸收
光谱等方面仔细分析Ti02的掺杂机理,为实验提供理论依据以及对实验制备高
质量的Ti02材料都具有重要的指导意义。 关键词:Ti02;第一性原理;电子结构;光吸收性能;掺杂 Abstract Duethe chemical to
outstandingstability,highphotocatalytic activity,low―cost
and is an roleasaenvironmental non―toxic,Ti02playingincreasinglyimportant
materialin control,solar isa pollution energy
wide semiconductorits materials, photocatalyticactivity bandgap normally
caused theultravioletofthesolar thevisibleofthe onlyby part ir
正在加载中,请稍后...小木虫 --- 500万硕博科研人员喜爱的学术科研平台
&&查看话题
fropho+VASP方法计算声子谱可以考虑born有效电荷(LO-TO split)吗
如题,可以的话,请大侠指点如何考虑
研究生必备与500万研究生在线互动!
扫描下载送金币

我要回帖

更多关于 castep计算声子谱过程 的文章

 

随机推荐