| 查看: 556 | 回复: 6 | |||
| 当前主题已经存档。 | |||
[交流]
【求助】LDA+U
|
|||
| 有哪为高手可以指点一下LDA+U的精髓啊？？或者有什么文献可以看看的啊？？推荐几篇吧~~~（注：本人只听说过LDA+U，但是从未深入了解过，所以最好可以有些基础介绍） |
回复此楼» 猜你喜欢
浙江师范大学是怎么坑我的
已经有10人回复
-
浙江师范大学是怎么坑我的
已经有3人回复
-
求教各位大神：2026教育部人文社科青年基金项目何时公示呀？
已经有6人回复
-
面上有专家说实验设备不是我们单位的
已经有8人回复
-
科研求助
已经有4人回复
-
博士去二本高校当辅导员，如何调整心态？
已经有11人回复
-
江苏省自然基金 什么时候出结果
已经有6人回复
-
闲聊
已经有3人回复
xiaojie7783
金虫 (小有名气)
- 应助: 0 (幼儿园)
- 金币: 957.4
- 散金: 76
- 帖子: 234
- 在线: 32.6小时
- 虫号: 702761
- 注册: 2009-02-17
- 性别: GG
- 专业: 金属材料的力学行为
2楼2009-11-27 08:46:45
3楼2009-11-27 10:58:39
4楼2009-11-27 21:44:53
5楼2009-12-05 16:17:06
★ ★ ★ ★
小木虫(金币+0.5):给个红包，谢谢回帖交流
qasd(金币+3,VIP+0):xiexie~ 12-7 19:55
小木虫(金币+0.5):给个红包，谢谢回帖交流
qasd(金币+3,VIP+0):xiexie~ 12-7 19:55
|
关于为什么要引入+U的计算，ABINIT网站上有说明 The standard Local Density Approximation (LDA), where the exchange and correlation energy is fit to homogenous electron gas results, is a functional that works well for a vast number of compounds. But, for some crystals, the interactions between electrons are so important that they cannot be represented by the LDA alone. Generally, these highly correlated materials contain rare-earth metals or transition metals, which have partially filled d or f bands and thus localized electrons. The LDA tends to delocalize electrons over the crystal, and each electron feels an average of the Coulombic potential. For highly correlated materials, the large Coulombic repulsion between localized electrons might not be well represented by a functional such as the LDA. A way to avoid this problem is to add a Hubbard-like, localised term, to the LDA density functional. This approach is known as LDA+U. In the actual implementation, we separate localized d or f electrons, on which the Hubbard term will act, from the delocalized ones (s and p electrons). The latter are correctly described by the usual LDA calculation. In order to avoid the double counting of the correlation part for localized electrons (already included in the LDA, although in an average manner), another term - called the double-counting correction - is subtracted from the Hamiltonian. In Abinit, two double-counting corrections are currently implemented : -The Full localized limit (FLL) (see A. Lichtenstein et al PRB 52, 5467 (1995)) -The Around Mean Field (AMF) (see M. T. Czyzyk and G. A. Sawatzky PRB 49, 14211 (1994)) For some systems, the result might depend on the choice of the double-counting method. However, the two methods generally give similar results 网址：http://www.abinit.org/documentat ... html/?searchterm=+U |
6楼2009-12-06 11:43:46
7楼2009-12-06 15:31:03