AMBER Archive (2009)

Subject: Re: [AMBER] dftb_disper problem

From: Fernando Martín García (fmgarcia_at_cbm.uam.es)
Date: Fri May 08 2009 - 07:43:19 CDT


Hi Gustavo:
I have changed the parameter that you indicated to me, but the problem
persist. I was investigating which is the atom that provokes the
problems and it's a oxygen atom that belongs to a water molecule, that
needs to be protonated (H3O+)at the end of the reaction I'm simulating.
These values depend on the type of oxygen atom I'm using?

Thanks
Fernando

El jue, 07-05-2009 a las 09:01 -0400, Gustavo Seabra escribió:
> Hi Fernando,
>
> This is actually a common problem with the use of dispersion in DFTB.
> What is happening here is that, at every step, the program tries to
> guess a hibridization state for an atom by counting the number of
> atoms within a certain distance threshold, in this case, 1.2A. Then,
> from the DISPERSION.INP_ONCHSP, it chooses the polarizability data for
> this hibridization. The DISPERSION.INP_ONCHSP file has the following
> format:
>
> $ cat DISPERSION.INP_ONCHSP
> O 0.560 0.560 0.000 0.000 3.8 3.8 3.8 3.8 3.15
> N 1.03 1.03 1.090 1.090 3.8 3.8 3.8 3.8 2.82
> C 1.382 1.382 1.382 1.064 3.8 3.8 3.8 3.8 2.5
> H 0.386 0.386 0.000 0.000 3.5 3.5 3.5 3.5 0.8
> P 1.6 1.6 1.6 1.6 4.7 4.7 4.7 4.7 4.5 !PO4
> S 3.0 3.0 3.0 3.0 4.7 4.7 4.7 4.7 4.8 !S not SO2
>
> The first 4 numbers after the atom name are the polarizabilities for
> each hibridization state, defined by the number of neighbors. (To
> understand the parameterization, look at the original reference: M.
> Elstner, P. Hobza, T. Frauenheim, S. Suhai, and E. Kaxiras, J. Chem.
> Phys. vol. 114 (2001) 5149 ) Notice that those parameters have been
> derived for DNA bases, so there is no guarantee they will be
> appropriate for anything else.
>
> Basically, these parameters were derived depending on the number of
> neighbors you may expect to find for an atom. However, during an MD
> calculation, it eventually happens that an extra atom gets close
> enough to confuse the program. In your case you have an Oxygen (atom
> 2) with 3 neighbors, and there is no polarizability data in the file
> for this situation. in this case, sander complains and crashes.
>
> ============
> WORKAROUND
> ============
> In the Oxygen line, just copy the parameter for 2 neighbors in to the
> next column. it will not be precise, but it will preventing sander
> from crashing. Line 1 in the DISPERSION.INP_ONCHSP file will will then
> look like:
> O 0.560 0.560 0.560 0.000 3.8 3.8 3.8 3.8 3.15
>
> Good luck,
> Gustavo Seabra
> Postdoctoral Associate
> Quantum Theory Project - University of Florida
> Gainesville - Florida - USA
>
>
>
> 2009/5/7 Fernando Martín García <fmgarcia_at_cbm.uam.es>:
> > Hi all:
> >
> > I have a problem with the flag "dftb_disper". When i launch my job with
> > a value of dftb_disper= 0, the job run well, but if i try to use
> > dftb_disper =1, the following message appears:
> >
> > --------------------------------------------------------------------------------
> > 4. RESULTS
> > --------------------------------------------------------------------------------
> >
> > *******************************************
> > WARNING: a parameter is 0.0 for atom 2
> > IZP = 2
> > nei = 4 ( 3 neighbors.)
> > hh1 = 0.00
> > hh2 = 3.80
> > Ni = 3.15
> > *******************************************
> > SANDER BOMB in subroutine <dispersion_params>
> > qm2_dftb_dispersion_params.f
> > Exiting.
> >
> > These are the parameters that i use in this job:
> >
> > &cntrl
> > ntx = 7, irest = 1, ntrx = 1,
> > ntxo = 1, nmropt = 1, ntr = 0,
> > ntpr = 100, ntwx = 100,
> > ntf = 1, ntb = 2, dielc = 1.0,
> > cut = 9, nsnb = 10, scnb = 2.0, scee = 1.2,
> > imin = 0, ibelly=0,
> > nstlim = 1000, dt = 0.001,
> > temp0 = 300.0, tempi = 300.0,
> > ntt = 1, vlimit = 20.0,
> > ntp = 1,
> > ntc = 1, tol = 0.00001, pres0=1, comp=44.6,
> > jfastw=0, nscm=1000,
> > ifqnt=1
> > /
> > &qmmm
> > iqmatoms= 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17,
> > 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28,
> > qmcharge= 0,
> > qm_theory='DFTB',
> > dftb_disper= 1,
> > dftb_maxiter= 70,
> > qmshake= 0,
> > qmmm_int = 2,
> > peptide_corr = 0,
> > qmcut= 8.0
> >
> > I've downloaded the parameter file "mio" from www.dftb.org and installed
> > in $AMBERHOME/dat/slko, and also the other two files are in the same
> > directory (DISPERSION.INP_ONCHSP and CM3). Can anybody tell me where is
> > the problem?
> >
> > Thanks
> >
> > --
> > Fernando Martín García.
> > Centro de Biología Molecular "Severo Ochoa".
> > C/ Nicolás Cabrera, 1.
> > Campus UAM. Cantoblanco, 28049 Madrid. Spain.
> >
> >
> > _______________________________________________
> > AMBER mailing list
> > AMBER_at_ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber
> >
>
> _______________________________________________
> AMBER mailing list
> AMBER_at_ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber

-- 
Fernando Martín García.
Centro de Biología Molecular "Severo Ochoa". 
C/ Nicolás Cabrera, 1. 
Campus UAM. Cantoblanco, 28049 Madrid. Spain.

_______________________________________________ AMBER mailing list AMBER_at_ambermd.org http://lists.ambermd.org/mailman/listinfo/amber