Which dft are available for cubic response

Find answers or ask questions regarding Dalton calculations.
Please upload an output file showing the problem, if applicable.
(It is not necessary to upload input files, they can be found in the output file.)

Post Reply
kaushikhatua
Posts: 35
Joined: 10 Jan 2014, 07:05
First name(s): kaushik
Last name(s): hatua
Affiliation: IIEST India
Country: India

Which dft are available for cubic response

Post by kaushikhatua » 25 Dec 2015, 04:29

Which DFT,s are available for cubic response calculation. I have done some on BLYP, BHHLYP, B2PLYP, CAM, B3LYP etc. When I go for B97-D or PBEPBE , PBE0 it fails.

taylor
Posts: 589
Joined: 15 Oct 2013, 05:37
First name(s): Peter
Middle name(s): Robert
Last name(s): Taylor
Affiliation: Tianjin University
Country: China

Re: Which dft are available for cubic response

Post by taylor » 25 Dec 2015, 18:49

The statement "it fails", with no information about how/why such as error logs or outputs, is unlikely to get much of a response here: we are able, but not telepathic!

You do not say what you are calculating via cubic response, but I assume it relates to some electric, magnetic, or mixed perturbation? If so you are most likely wasting your time running DFT with any functional. Virtually all functionals are semiempirical, and to my knowledge none is parametrized for high-order electric or magnetic response. As a consequence, any results you obtain will be either completely fortuitous (if they look plausible), or garbage (when they don't). If your preliminary investigations indicate that you need, e.g., CC response, well then, however unfortunate from a computer time point of view, that's what is needed! I doubt you will find a DFT approach that can substitute...

Best regards
Pete

arnfinn
Posts: 231
Joined: 28 Aug 2013, 08:02
First name(s): Arnfinn
Middle name(s): Hykkerud
Last name(s): Steindal
Affiliation: UiT
Country: Norway
Location: UiT The Arctic University of Norway

Re: Which dft are available for cubic response

Post by arnfinn » 25 Dec 2015, 20:16


kaushikhatua
Posts: 35
Joined: 10 Jan 2014, 07:05
First name(s): kaushik
Last name(s): hatua
Affiliation: IIEST India
Country: India

Re: Which dft are available for cubic response

Post by kaushikhatua » 30 Dec 2015, 09:37

Sorry for late reply and also the for stupid behavior of myself. By the way I got the answer from previous post that second hyperpolarizability calculation is limited by the functionals as shown there. I am using Dalton 2013.4 . Actually I am trying to see which DFT produce close results to CCSD. Most of the functionals are considered or tested for linear response. Whether these functionals have been implemented in Dalton 2016?
Attachments
1Li-acetylene-PBE0.out
(41.08 KiB) Downloaded 271 times
1Li-acetylene-B97D.out
(45.96 KiB) Downloaded 239 times

taylor
Posts: 589
Joined: 15 Oct 2013, 05:37
First name(s): Peter
Middle name(s): Robert
Last name(s): Taylor
Affiliation: Tianjin University
Country: China

Re: Which dft are available for cubic response

Post by taylor » 31 Dec 2015, 14:04

I am posting again to make sure there is no confusion here. The list to which you were referred is, I think, a list of which functionals have been tested in response calculations of a certain order to make sure the code works. This says nothing about the value of using particular functionals for particular purposes, just that the code works.

Let me try to explain again what I posted above. DFT functionals are parametrized using some set of data to which the author(s) of the functional want to see good agreement. This might be experimental geometries and frequencies for some set of molecules, it might be barrier heights for chemical reactions, it might be chemical shifts in NMR. In this sense DFT methods are semiempirical and are indeed the best semiempirical methods we have. However, it is in the nature of fitting anything, like a functional, say, to a set of data, that the best results are invariably obtained by interpolation. By this, I mean that one is calculating a system, not part of the fitting set, but similar to the molecules in the fitting set and for the same properties for which the fitting was done. For example, I would expect reasonable geometries and frequencies for organic molecules using the B3LYP functional (and appropriate basis sets). I would not necessarily expect good dipole moments, because they were not part of the original fitting process and I would be extrapolating, not interpolating, and extrapolation is a much less reliable procedure in general.

If you are interested in obtaining good second hyperpolarizabilities, you will need a functional that was parametrized with this in mind. I know of no such functional, but I am not a DFT expert and there are so many functionals out there that I do not rule out there might be something suitable. That said, it needs to be parametrized against your property of interest. I do not believe that even a functional that is tuned to give reasonable polarizabilities or dispersion interaction will necessarily give reasonable hyperpolarizabilities, for example, and unfortunately unless you use a large set of comparisons the odd "<fill-in-the-blanks> functional gives good agreement with CCSD" observation is more than likely to be fortuitous.

I recognize that one may be interested in hyperpolarizabilities of systems that are too large to be treated with CCSD, say, although the divide-expand-consolidate (DEC) approach in LSDalton, together with an applied static electric field, should be applicable to large systems. However, it is pointless, both in terms of wasting computer time, and in terms of potentially misleading experimental collaborators, to try to use DFT methods to calculate properties for which they are entirely unsuited.

Best regards
Pete

hjaaj
Posts: 371
Joined: 27 Jun 2013, 18:44
First name(s): Hans Jørgen
Middle name(s): Aagaard
Last name(s): Jensen
Affiliation: Universith of Southern Denmark
Country: Denmark

Re: Which dft are available for cubic response

Post by hjaaj » 03 Feb 2016, 08:29

kaushikhatua wrote:Sorry for late reply and also the for stupid behavior of myself. By the way I got the answer from previous post that second hyperpolarizability calculation is limited by the functionals as shown there. I am using Dalton 2013.4 . Actually I am trying to see which DFT produce close results to CCSD. Most of the functionals are considered or tested for linear response. Whether these functionals have been implemented in Dalton 2016?
A late answer. As long as you are only interested in static second hyperpolarizabilites, then you can calculate them by using second order numerical differentiation of the polarizability for the much larger set of DFT functionals for which polarizability is implemented. The Dalton keyword for finite field is ".FINITE FIELD", and the property labels needed are XDIPLEN, YDIPLEN and ZDIPLEN. The field strengths you enter are in atomic units. I recommend to use 2nd order central difference approach such that the accuracy will be of second order in the finite field value.

hjaaj
Posts: 371
Joined: 27 Jun 2013, 18:44
First name(s): Hans Jørgen
Middle name(s): Aagaard
Last name(s): Jensen
Affiliation: Universith of Southern Denmark
Country: Denmark

Re: Which dft are available for cubic response

Post by hjaaj » 03 Feb 2016, 08:29

PS. And DO upgrade to Dalton2016!

Post Reply

Who is online

Users browsing this forum: No registered users and 12 guests