CCL: NPT TIP5P, 1atm MD

From: Manish Agarwal <zmanish(0)gmail.com>
Subject: CCL: NPT TIP5P, 1atm MD
Date: Thu, 11 Feb 2010 13:38:39 +0530
 Sent to CCL by: Manish Agarwal [zmanish]~[gmail.com]
 My sincere apologies for the delayed update.
 Many thanks to all the responses I have got.
 This is a long mail, and is purely an update - so I am still running
 simulations.
 A small recounting of events:
 NPT Molecular dynamics Simulation of TIP5P water model,  Cubic
 Boundary Conditions,  Ewald Summation for Electrostatics, Quaternion
 for rigid body motion, Timestep 0.001ps  (1fs)
 Target
 Temperature 300K, pressure 1atm (0.001 katm)
 256 water molecules: 1 million steps
 Berendesen Themostat (20ps) and barostat(1ps)
 Pressure 9.1328E-04 katm
 Temperature 2.7879E+02
 Volume: 7.7622E+03
 Density:0.985
 256 water molecules: 2 million steps
 Berendesen Themostat (0.1ps) and barostat(0.5ps)
 Pressure 4.2339E-03 katm
 Temperature 2.9982E+02
 Volume :  7.7838E+03
 Density: 0.983
 >>>>>Here is what I did, following excellent discussion here,
 500 Molecules of water (tip5p)
 1million steps (1ns)
 Berendesen Themostat (2ps) and barostat(0.5ps)
 Pressure 1.1679E-03 katm
 Temperature 2.9812E+02
 Volume :  1.5194E+04
 Density: 0.983
 1000 Molecules of water (tip5p)
 1million steps (1ns)
 Berendesen Themostat (2ps) and barostat(0.5ps)
 Pressure 1.1679E-03 katm
 Temperature 2.9812E+02
 Volume : 3.0412E+04
 Density: 0.982
 I am starting up longer runs 4-6ns-
 and currently I have 40h/ns for 1000 molecules
 and 24h/ns for 500 molecules
 Regards
 Manish Agarwal
 <zmanish|-|gmail.com>
 Senior Research Scholar
 Department of Chemistry
 Indian Institute of Technology, Delhi
 - - - - - - - - - - - - - - - - - - - - - - - - - - -
 On 7 February 2010 17:26, Manish Agarwal zmanish#,#gmail.com
 <owner-chemistry|-|ccl.net> wrote:
 >
 > Sent to CCL by: Manish Agarwal [zmanish|*|gmail.com]
 > Dear CCLers
 >
 > Excellent suggestions, that too on the weekend ... My thanks to all.
 > I should have mentioned  that I am using cubic Periodic Boundary.
 >
 > Here is what I intend to do:
 >
 > The time step of 1 fs is, in my opinion, not too low or too high for
 > bulk water simulations at temperatures where there are no slow
 > dynamics.
 >
 > I _am_ going to increase the number of molecules, and see if it
 > changes the outcome - 500, and 1000.
 > I am also going to increase the number of simulation timesteps - 4-6
 > million should suffice for the test. I will also try Nose-Hoover
 > thermostat-barostat combination.
 >
 > In the past I have run Berendsen NVT simulations with tau=200ps with
 > no problems - although, reviewers have complained it to be 'too high'.
 > The large relaxation time allows for internal dynamics to be as near
 > NVE as possible, specially near temperatures below the model's melting
 > point. But this is my first brush with NPT.
 >
 > My main concern has been that I am trying to compare MD-NPT results
 > with those obtained from MC-NPT.
 >
 > I will try and post my results in the next 24-48 hours...
 >
 > Looking forward to more discussion,
 > Regards, and thanks again.
 > Manish Agarwal
 > - - - - - - - - - - - - - - - - - - - - - - - - - - -
 >
 >
 > On 6 February 2010 23:02, Venable, Richard (NIH/NHLBI) E
 > venabler!A!nhlbi.nih.gov <owner-chemistry . ccl.net> wrote:
 >>
 >> Sent to CCL by: "Venable, Richard (NIH/NHLBI) [E]"
 [venabler-*-nhlbi.nih.gov]
 >> I have to disagree.
 >>
 >> If there are suspected stability problems, the last thing one should do
 is to increase the timestep.
 >>
 >> The weak coupling scheme typified by the Berendsen thermostat and
 barostat have been justly criticized by many, including Allen and Tildesley;
 Nose-Hoover extended system approaches are regarded as considerably less flawed.
 >>
 >> Starting from 0 K is usually not necessary, and can be ill advised in
 some programs.  Anything below 200 K should be just fine.
 >>
 >> I don't believe one can run NPT without periodic boundaries.
 >>
 >> --
 >> Rick Venable
 >> compuational chemist
 >>
 >> ________________________________________
 >>> From: owner-chemistry+rick_venable==nih.gov||ccl.net
 [owner-chemistry+rick_venable==nih.gov||ccl.net] On Behalf Of Mahmoud A. A.
 Ibrahim m.ibrahim[a]compchem.net [owner-chemistry||ccl.net]
 >> Sent: Saturday, February 06, 2010 11:17 AM
 >> To: Venable, Richard (NIH/NHLBI) [E]
 >> Subject: CCL: NPT TIP5P, 1atm MD
 >>
 >> Dear Manish
 >> I am not familiar with DL Ploy.
 >> But, if you don't mind, you can have a go with the following issues to
 reproduce the water density:
 >> * Use NPT ensemble with relaxation time of 2.0ps
 >> * Use Berendsen Thermostat with 0.5ps time constant
 >> * Use Periodic Boundary Condition (Important)
 >> * If it is available, start your run from 0K and increase the
 temperature gradually until reaches 300K. (During heating step, use constant
 volume boundary condition).
 >>
 >> Have a go with these issues, Good look
 >> Sincerely;
 >> M. Ibrahim
 >>
 >>
 >> On Fri, Feb 5, 2010 at 10:27 PM, Manish Agarwal
 zmanish**gmail.com<http://gmail.com> <owner-chemistry{=cl.net<mailto:owner-chemistry{=cl.net>> wrote:
 >>
 >> Sent to CCL by: Manish Agarwal [zmanish^^gmail.com<http://gmail.com>;]
 >> Dear CCLers,
 >>
 >> I am having trouble reproducing the density at 1 atm, 300K for TIP5P
 >> water, or for that matter any water model in the NPT ensemble using
 >> Molecular Dynamics:
 >>
 >> have tried
 >> Berendsen Thermostat and Barostat
 >> Nose-Hoover Thermostat and Barostat
 >>
 >> am using Ewald summation for electrostatics,
 >> 256 water molecules. 2000000  steps, with or without velocity
 scaling
 >> in the beginning 1fs time step
 >> DLPOLY ver 2.19
 >>
 >> The problem is as follows : either the pressure does not stabilize at
 >> 1atm (too low or too high) or the temperature does not maintain at
 >> 300K
 >> When both are maintained, for example Berendsen Thermostat relaxation
 >> time 0.1ps, Barostat relaxation time 0.5ps, the density does not reach
 >> 0.99g/cc; also, the energy profile with time looks too perturbed (does
 >> NOT look like a nice fluctuation about the mean).
 >>
 >> In the NVT ensemble, however, I have been able to reproduce other
 >> static(orientational order) and dynamic properties(diffusivities) - so
 >> the model seems fine.
 >>
 >> Any thoughts, suggestions are appreciated.
 >>
 >> Regards,
 >> Manish Agarwal
 >> <zmanish||gmail.com<http://gmail.com>>;
 >>
 >>
 >>
 >> -= This is automatically added to each message by the mailing script =-
 >> E-mail to subscribers: CHEMISTRY{=cl.net<mailto:CHEMISTRY{=cl.net> or use:E-mail to administrators:
 CHEMISTRY-REQUEST{=cl.net<mailto:CHEMISTRY-REQUEST{=cl.net> or usehttp://www.ccl.net/chemistry/sub_unsub.shtmlhttp://www.ccl.net/spammers.txt
 >>
 >>
 >> - This is automatically added to each message by the mailing script
 ->      http://www.ccl.net/cgi-bin/ccl/send_ccl_message>;  
    http://www.ccl.net/cgi-bin/ccl/send_ccl_message>      http://www.ccl.net/chemistry/sub_unsub.shtml>;  
    http://www.ccl.net/spammers.txt>;
 >>
 >>
 >
 >
 >
 > - This is automatically added to each message by the mailing script ->
      http://www.ccl.net/cgi-bin/ccl/send_ccl_message>;  
    http://www.ccl.net/cgi-bin/ccl/send_ccl_message>;  
    http://www.ccl.net/chemistry/sub_unsub.shtml>;  
    http://www.ccl.net/spammers.txt>;
 >
 >