计算化学公社

标题: 请教关于基组方面的问题（描述共价键） [打印本页]

作者
Author: qinjiu 时间: 2020-10-12 21:52
标题: 请教关于基组方面的问题（描述共价键）
J. Phys. B: At. Mol. Opt. Phys. 43 (2010) 065101
起因是这篇文章，他算了HeH+这个体系,在AV5Z基组上DIY了一下
因为这个体系比较小，我就用他给的活性空间随便测试了一下，发现2Sigma+丢失了，举个例子就是他的结果是-2.10，-2.09，-2.05au，我算出来的是-2.11，-2.04，-2.02au这种
基组就是正常的av5z,  CASSCF+MRCISD/FCI
hydrogen          (9s,5p,4d,3f,2g) -> [6s,5p,4d,3f,2g]
helium             (9s,5p,4d,3f,2g) -> [6s,5p,4d,3f,2g]
然后我就拿他附录里给的基组测试了一下，重复出了他的结果。因为我对基组这块了解不深，确实没搞懂为什么会出现这种结果。我猜测可能是他从解离极限的角度去做的基组，所以在渐进区能符合的很好，然后这个事情就搁置下来了，直到近几天在做一个包含共价键体系的计算，一个AB中性小双原子分子，最低的几个通道A+B里夹了一个A- + B+ 通道的，我一直算不对这个通道（>0.2eV）。其他极限能和NIST符合的很好（0.01eV以内，毕竟小体系），我就怀疑可能是基组的问题。（另一方面我还考虑过不是因为大小一致性的问题，需要做Davidson校正，尝试了一下结果也不理想）
半路出家搞量化，确实有很多常识不理解，这里请教各位老师可能是哪方面的问题?
顺便附一下这篇文章里的基组：

BASIS
spdfg,He,av5z;c
s,he,99.181545,10.536516,1.905592,0.424822,0.029920 !pour la 2s
c,1.5,0.00276,0.030509,0.196191,0.775469,-3.72247 !coeff contraction
p,he,19.984152,4.739990,1.021806,0.292204,0.101345,0.039590,0.016351  !pour la 2p
c,1.7,0.000152,0.002135,0.025319,0.160069,0.592704,1.066547,0.527114
s,he,99.838162,49.068205,2.653277,0.309173,0.038398,0.006064 !pour la 3s
c,1.6,-0.001952,0.012286,0.151694,0.516699,-3.156883,4.383564
p,he,4.673075,0.848515,0.227656,0.070757,0.007911    !pour la 3p
c,1.5,0.002929,0.037026,0.244477,0.806379,-2.050784
d,he,0.102716,0.026362,0.008950 !pour la 3d
c,1.3,0.062634,0.43099,0.551594
s,he,99.861298,49.188130,0.732304,0.035479,0.009221,0.001472
c,1.6,-0.004235,0.011613,0.316811,-2.495917,4.534324,4.267928
p,he,4.629236,0.396874,0.081760,0.017020,0.013667,0.002285
c,1.6,0.002547,0.076795,0.488599,3.304293,-4.848362,2.305369
d,he,0.082347,0.021503,0.003283
c,1.3,0.077839,0.454447,-0.894818
f,he,0.033780,0.010582,0.004106
c,1.3,0.026447,0.170021,0.204038
spdfg,H,av5z;c
s,h,19.907407,8.988620,0.645915,0.071221,0.024689
c,1.5,0.0001437,-0.005010,-0.059868,0.123387,0.891720
p,h,3.101143,0.56727,0.158596,0.052537,0.020001
c,1.5,0.005698,0.067894,0.407521,1.082414,0.800233
s,h,1.00084,0.272819,0.093336,0.023846,0.019287,0.00425
c,1.6,0.017083,0.120859,0.381641,0.782727,-2.036422,1.594513
p,h,0.986107,0.181062,0.051409,0.007783,0.003281
c,1.5,0.02459,0.260387,0.903505,-1.874014,-0.287389
d,h,0.199534,0.051732,0.018241,0.007264
c,1.4,0.050502,0.495617,1.609201,1.213582
s,h,11.93994,1.060859,0.188361,0.029081,0.007375,0.001197
c,1.6,0.013653,0.167164,0.662095,-3.032473,4.828012,-4.362099
p,h,1.015267,0.186800,0.053622,0.009383,0.002477
c,1.5,0.015689,0.170182,0.625077,-2.040089,2.535609
d,h,0.083401,0.021517,0.003285
c,1.3,0.077845,0.454482,-0.894944
f,h,0.033813,0.010591,0.004109
c,1.3,0.026401,0.169866,0.204085
END

(, 下载次数 Times of downloads: 39)

作者
Author: qinjiu 时间: 2020-10-12 22:00
本帖最后由 qinjiu 于 2020-10-27 20:23 编辑

(, 下载次数 Times of downloads: 32)

第二个问题大概解决了
MRCI+Q 有五种校正方法，一般默认是第二种，不妨试一下新版的第四种或者第五种

欢迎光临计算化学公社 (http://bbs.keinsci.com/)