|
symthresh不是越大越好,可以看到手册里可能最多也就设到1e-2。首先,如果symthresh设得太粗糙,会导致把点群判断成D18h;其次,ORCA的代码未必是在symthresh大到不合理的量级的时候仍然能正常工作的,所以你去看symthresh设得特别大的计算(比如设为1.0)的输出,会发现程序会报一些warning:
- Warning (ORCA_SYM): Tolerance set to a very high value (1.0E+00)! Beware of odd results!
- The point group will now be determined using a tolerance of 1.0000e+00.
- Splitting atom subsets according to nuclear charge, mass and basis set.
- Splitting atom subsets according to distance from the molecule's center.
- Identifying relative distance patterns of the atoms.
- Splitting atom subsets according to atoms' relative distance patterns.
- Bring atoms of each subset into input order.
- The molecule is planar.
- The molecule has a center of inversion.
- Analyzing the first atom subset for its symmetry.
- WARNING (ORCA_SYM): Unexpected number of nearest neighbors (17) in function
- TPolytope::CenteredPolygon(TSymmetryFinder &).
- WARNING (ORCA_SYM): None of the atom subsets could be assigned a point group.
- Please take all warnings, especially the ORCA_SYM warnings, very seriously!!!
- WARNING (ORCA_SYM): An inconsistency has occured during symmetry recognition.
- Using C2h as a fallback. If this point group appears to be too low, please try
- increasing the cleaning tolerance using option SymThresh in the %symmetry input
- block. If the problem persists, you may post a message in the ORCA forum.
- Testing point group C2h.
- Success!
- This point group has been found: C2h
- Largest non-degenerate subgroup: C2h
复制代码
可以看到,首先程序明确说这个symthresh设得太大了,结果可能会比较奇怪,然后报了好几个警告,因为判断点群的时候出现了内部不自洽的问题,所以放弃判断真实点群,但是程序至少知道C2h是分子的子群,所以后续用C2h群计算,虽然未必是用户想要的结果,但是起码能让计算正常进行。
另外我这里测试的结果是,symthresh设成1e-2是可以正确识别点群为D9h的。所以程序应该没问题。做计算一定要看输出文件里的警告,不能只看最终结果 |
|