Dear Brad,
Thanks for your valuable comments, this confirms what I was thinking. Thanks also for letting me know About the need for re-compiling. I will look into this. I take that when specifying the parameters list within this block, the C6 parameter always refers to the first atom defining the pair (and therefore remains constant for a specific sub-block), while the van-der-Waals radii naturally refer to the sum of both. Hence, %block MM.potentials 1 1 C6(1) R0(1)+R0(1) 1 2 C6(1) R0(1)+R0(2) 1 3 C6(1) R0(1)+R0(2) 2 2 C6(2) R0(2) +R0(2) 2 3 C6(2) R0(2) +R0(3) %endblock MM.Potentials Is this correct? I am still a bit unsure about the 1.66, but have found d=20 in the paper indeed. Thank for pointing this out to me. Best Tobias ________________________________ Dr. Tobias Kraemer Research Associate Institute of Chemical Sciences School of Engineering & Physical Sciences Heriot-Watt University Edinburgh EH14 4AS United Kingdom * [email protected] * +44 (0)131 451 3259 ________________________________ From: [email protected] [mailto:[email protected]] On Behalf Of Brad Wells Sent: 04 February 2014 14:31 To: [email protected] Subject: Re: [SIESTA-L] Grimmer Dispersion Correction Hi Tobias, (1) Unfortunately for 7 different atoms you need explicitly specify the 28 unique pairwise potentials. This is tedious, which is why I have a program that does it for me. One thing to note in the release versions of SIESTA there is only the memory space for 10 pair potentials. If you want to correctly use Grimme's dispersion correction you will have to change maxMMpot in the molecularmechanics.F90 and re-compile the program. If you don't do this SIESTA will only use the first 10 potentials that you specify. (2) Grimme uses J nm^(6) mol^(-1), SIESTA uses eV angstrom^(6), so the conversion factor is approx. 10.3641 (3) I am not sure about the 1.66. From Grimme's 2006 paper for PBE this should be 0.75. The D value is the same as in Grimme's paper. I hope this helps you. Brad Wells. Monash University. On 5/02/2014 1:07 AM, Kraemer, Tobias wrote: Dear all, I would like to try out adding Grimme's dispersion correction to my calculation. From the manual and the list archive I understand that this can be done, however, some points are unclear to me. The parameters need to be defined in the MM.Potentials block, and if I am not mistaken, this needs to be done for each pair wise interaction between atom types. For my system containing 7 different atoms, this seems to be quite tedious. Is there a more straight forward way of doing this? Certainly I could restrict the number to only include nearest neighbours, but this is difficult to decide a priory. So, there would be a large number of pairs that would need to be considered here. Nonetheless, if I should do it this way how would I translate between different units. It appears that the original Grimme paper reports these parameters in Jnm^(6)mol^(-1) which differs from SIESTA standard units (I stumbled upon this discussion in the archive, however, I couldn't find A conclusive answer. Unfortunately I also seem to be unable to recover the post from the archive, despite some time spent searching). My last question addresses the 2 parameters MM.Grimme.D and MM.Grimme.S6. The latter is by default set to 1.66 for a DZP basis, however, For which functional is this the default. Can this value be ised straight away with PBE? Where does this value come from? Likewise, can the default value for .D be used straight away? Thanks for your help.... Tobias ________________________________ Dr. Tobias Kraemer Research Associate Institute of Chemical Sciences School of Engineering & Physical Sciences Heriot-Watt University Edinburgh EH14 4AS United Kingdom * [email protected]<mailto:[email protected]> * +44 (0)131 451 3259 ________________________________ ________________________________ Sunday Times Scottish University of the Year 2011-2013 Top in the UK for student experience Fourth university in the UK and top in Scotland (National Student Survey 2012) We invite research leaders and ambitious early career researchers to join us in leading and driving research in key inter-disciplinary themes. Please see www.hw.ac.uk/researchleaders<http://www.hw.ac.uk/researchleaders> for further information and how to apply. Heriot-Watt University is a Scottish charity registered under charity number SC000278. ----- Sunday Times Scottish University of the Year 2011-2013 Top in the UK for student experience Fourth university in the UK and top in Scotland (National Student Survey 2012) We invite research leaders and ambitious early career researchers to join us in leading and driving research in key inter-disciplinary themes. Please see www.hw.ac.uk/researchleaders for further information and how to apply. Heriot-Watt University is a Scottish charity registered under charity number SC000278.
