<div dir="ltr"><DIV>Hello,</DIV>
<DIV>I'd like to update my progress on this issue.</DIV>
<DIV>It turns out the rotation of the molecule is responsible for the peak splitting, as Ari said. To control the rotation motion of the molecule, we have to combine "ANGVEL_ZERO" with "none-periodic" calculation. In that way, I got the correct IR vibrational frequencies for H2O molecule.</DIV>
<DIV> </DIV>
<DIV>But, I have another question again about the intensity of the IR. It looks the intensity ratio of the 3 modes is not correct. The IR was obtained from Travis program. My question is that how the Travis program work on IR calculation? Has it considered the prefactor of the FT of the dipole correlation function to get the correct intensity? Thanks a lot!</DIV>
<DIV> </DIV>
<DIV>Regards,</DIV>
<DIV>Jianjun</DIV>
<DIV> <BR>On Tuesday, 27 August 2013 14:26:33 UTC-6, JianJun Yang wrote:</DIV>
<BLOCKQUOTE style="BORDER-LEFT: #ccc 1px solid; MARGIN: 0px 0px 0px 0.8ex; PADDING-LEFT: 1ex" class=gmail_quote>
<DIV dir=ltr>Dear Ari & Florian,
<DIV><BR></DIV>
<DIV>Thank you very much for the comments.</DIV>
<DIV>Yes, the molecule is rotating as well. When I made EPS_SCF tighter to 1E-7 I did make the EPS_DEFAULT tighter to 1E-11.</DIV>
<DIV>I will try your suggestions and will let you know the effect later.</DIV>
<DIV><BR></DIV>
<DIV>Regards,</DIV>
<DIV>Jianjun</DIV><BR>On Tuesday, 27 August 2013 05:08:24 UTC-6, Florian Schiffmann wrote:
<BLOCKQUOTE style="BORDER-LEFT: #ccc 1px solid; MARGIN: 0px 0px 0px 0.8ex; PADDING-LEFT: 1ex" class=gmail_quote>
<DIV dir=ltr>Dear Jianjun,<BR><BR>what Ari says is correct. About the cutoff, a reasonable value is mainly determined by the PP for the elements. If I have seen correctly you are using BLYP. These PP might require even higher values to get PW convergence. Simply run several singlepoint calculations with various cutoffs and see when the energy doesn't change anymore.<BR><BR>Flo<BR></DIV></BLOCKQUOTE></DIV></BLOCKQUOTE></div>