[gmx-users] Water molecule starting at atom 18511 can not be settled

Justin Lemkul jalemkul at vt.edu
Sat Apr 16 21:52:25 CEST 2016



On 4/16/16 3:24 PM, Alexander Alexander wrote:
> Dear Justin,
>
> Thanks for the explanation.
> Defiantly it is something with the epsilon, as for epsilon=17.0 not only
> minimisation but also a 5 ns NVT calculation afterwards also is running
> fine and it is in the very end steps now. But for epsilon = 64.65 it
> crashes immediately
> Please find the input files via below links.
>
> https://drive.google.com/open?id=0B_CbyhnbKqQDZlpiMG8zU29BM2c
>

It is as I suspected.  Even in the epsilon = 17 system, the waters crush down 
into the B/Ni surface and pack very tightly.  It's vastly worse in the epsilon = 
64 system.  You are actually smashing waters into the surface, leaving a micro 
vacuum between layers of water.  It is very obvious from visual inspection from 
EM (output some coordinates and you will see this very obviously as a function 
of EM step), as well as interaction energies.  The values for LJ(SOL-SOL) and 
LJ(SOL-Other) drop precipitously with eps = 64, and LJ(SOL-SOL) drops similarly 
for eps = 17.  So the eps = 17 system is probably metastable, but eps = 64 is 
pretty hopeless.  Such values for well depths really make no sense.  These are 
100-600x deeper than almost any other interaction in typical force fields.

-Justin

> Cheers,
> Alex
>
> On Sat, Apr 16, 2016 at 7:49 PM, Justin Lemkul <jalemkul at vt.edu> wrote:
>
>>
>>
>> On 4/16/16 1:44 PM, Alexander Alexander wrote:
>>
>>> Dear Christopher,
>>>
>>> If it is not a default, I have not applied any "SETTLE" in my calculation
>>> as far as I know. Here is the .mdp file of my minimisation. Also the
>>>
>>
>> SETTLE is on by default unless using -DFLEXIBLE, as defined in all water
>> topologies.  Water models present in GROMACS are all intended to be rigid
>> unless the user specifies otherwise.
>>
>> "define = -DFLEXIBLE" did not help.
>>>
>>>
>> This suggests a problem with the force field.  The value you have set for
>> epsilon is massive, several orders of magnitude larger than normal values.
>> Likely the attraction between the Ni and OW is so large that it draws water
>> molecules very close together, to the point that they are compressed into
>> repulsive regions with each other.
>>
>> Providing a tarball of all your inputs (force field, coordinates,
>> topologies uploaded to a file-sharing service and URL sent to the mailing
>> list) would be helpful in diagnosing the problem and testing this idea.
>>
>> -Justin
>>
>>
>> -----------------------------------
>>> define = -DFLEXIBLE
>>> integrator = steep
>>> ;integrator = cg
>>> ;integrator = l-bfgs
>>> emtol = 10
>>> emstep = 0.0001
>>> nsteps = 100000000
>>> nstlog = 1
>>> nstenergy = 1
>>> nstlist = 10
>>> ns_type = grid
>>> coulombtype = PME
>>> rcoulomb = 1.26
>>> vdwtype = cutoff
>>> rvdw = 1.26
>>> pme_order = 4
>>> optimize_fft = yes
>>> ;constraints = all-bonds
>>> ;constraint_algorithm = LINCS
>>> cutoff-scheme = Verlet
>>> ;rlist = Cut-off distance for the short-range neighbor list.
>>> pbc = xyz
>>> ------------------------
>>> Cheers,
>>> Alex
>>>
>>>
>>> On Sat, Apr 16, 2016 at 7:02 PM, Christopher Neale <
>>> chris.neale at alum.utoronto.ca> wrote:
>>>
>>> Then don't use settle for your minimization?
>>>> (mdp option define = -DFLEXIBLE for most water models)
>>>>
>>>> You can always turn it on again and run another EM after whatever
>>>> problematic configurational feature has relaxed.
>>>>
>>>> ________________________________________
>>>> From: gromacs.org_gmx-users-bounces at maillist.sys.kth.se <
>>>> gromacs.org_gmx-users-bounces at maillist.sys.kth.se> on behalf of
>>>> Alexander
>>>> Alexander <alexanderwien2k at gmail.com>
>>>> Sent: 16 April 2016 11:57
>>>> To: gmx-users at gromacs.org
>>>> Subject: Re: [gmx-users] Water molecule starting at atom 18511 can not be
>>>>        settled
>>>>
>>>> Dear Mark,
>>>>
>>>> Thanks for your response.
>>>> If I understood your statement correctly, I would say these numbers used
>>>> here for the epsilon and sigma are not arbitrary but they come from some
>>>> well quantum calculations (DFT) of water on this surface, actually,
>>>> adsorption energy and optimized geometry of water on the surface were
>>>> used
>>>> in Lennard-Jones potentials to get sigma and epsilon, this means it is a
>>>> sound model physics.
>>>>
>>>> Cheers,
>>>>
>>>> Alex
>>>>
>>>>
>>>> On Sat, Apr 16, 2016 at 5:01 PM, Mark Abraham <mark.j.abraham at gmail.com>
>>>> wrote:
>>>>
>>>> Hi,
>>>>>
>>>>> You will always need a sound model physics and a starting configuration
>>>>> that is close enough to consistent with that. Which aspect do you think
>>>>>
>>>> is
>>>>
>>>>> wrong?
>>>>>
>>>>> Mark
>>>>>
>>>>> On Sat, Apr 16, 2016 at 3:58 PM Alexander Alexander <
>>>>> alexanderwien2k at gmail.com> wrote:
>>>>>
>>>>> Dear Gromacs user,
>>>>>>
>>>>>> The minimisation (steepest descent) of water on surfaces of a nickel
>>>>>> intermetallic compounds works fine if epsilon in the nonbond_params
>>>>>>
>>>>> below
>>>>
>>>>> is 17.0 (harvested from LJ(10-4)) but it will crash immediately if
>>>>>>
>>>>> epsilon
>>>>>
>>>>>> is elevated to 64.65 (harvested from LJ(12-6) or LJ(9-6)) with settling
>>>>>> problem of one water molecule.
>>>>>>
>>>>>> [ nonbond_params ]
>>>>>> ;     ai       aj    funct     sigma(nm)      epsilon (kj/mol)
>>>>>>         Ni     OW     1          0.2072             64.65
>>>>>>
>>>>>> The settling problem of the molecule has been already discussed many
>>>>>>
>>>>> times
>>>>>
>>>>>> in mailing list but no helpful solution has been proposed yet.
>>>>>> So any solution would be highly appreciated.
>>>>>>
>>>>>> Cheers,
>>>>>> Alex
>>>>>> --
>>>>>> Gromacs Users mailing list
>>>>>>
>>>>>> * Please search the archive at
>>>>>> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>>>>>> posting!
>>>>>>
>>>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>>>>
>>>>>> * For (un)subscribe requests visit
>>>>>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>>>>>> send a mail to gmx-users-request at gromacs.org.
>>>>>>
>>>>>> --
>>>>> Gromacs Users mailing list
>>>>>
>>>>> * Please search the archive at
>>>>> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>>>>> posting!
>>>>>
>>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>>>
>>>>> * For (un)subscribe requests visit
>>>>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>>>>> send a mail to gmx-users-request at gromacs.org.
>>>>>
>>>>> --
>>>> Gromacs Users mailing list
>>>>
>>>> * Please search the archive at
>>>> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>>>> posting!
>>>>
>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>>
>>>> * For (un)subscribe requests visit
>>>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>>>> send a mail to gmx-users-request at gromacs.org.
>>>> --
>>>> Gromacs Users mailing list
>>>>
>>>> * Please search the archive at
>>>> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>>>> posting!
>>>>
>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>>
>>>> * For (un)subscribe requests visit
>>>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>>>> send a mail to gmx-users-request at gromacs.org.
>>>>
>>>>
>> --
>> ==================================================
>>
>> Justin A. Lemkul, Ph.D.
>> Ruth L. Kirschstein NRSA Postdoctoral Fellow
>>
>> Department of Pharmaceutical Sciences
>> School of Pharmacy
>> Health Sciences Facility II, Room 629
>> University of Maryland, Baltimore
>> 20 Penn St.
>> Baltimore, MD 21201
>>
>> jalemkul at outerbanks.umaryland.edu | (410) 706-7441
>> http://mackerell.umaryland.edu/~jalemkul
>>
>> ==================================================
>>
>> --
>> Gromacs Users mailing list
>>
>> * Please search the archive at
>> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>> posting!
>>
>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
>> * For (un)subscribe requests visit
>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>> send a mail to gmx-users-request at gromacs.org.
>>

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================


More information about the gromacs.org_gmx-users mailing list