Re: [ESPResSo] Espresso seems to give unphysical results

[Limbach, Hans Joerg, LAUSANNE, Food Materials]

 Hi,

First idea would be to check whether the dihedral angle definition in ESPResSO 
and from the force field are the same. Most secure check for this is of course 
the source code.
As you mention that you customized the source code, one would need the code 
that you use to give you a hint.
Is it right that an NVE simulation gives the right conformation and things go 
wrong only if you turn on the thermostat? (please provide the full script)
I am sure this can be sorted out!

Cheers,
Hanjo

> -----Original Message-----
> From: address@hidden 
> [mailto:address@hidden On Behalf Of 
> Michael Winokur
> Sent: lundi, 24. août 2009 20:21
> To: address@hidden
> Subject: [ESPResSo] Espresso seems to give unphysical results
> 
> Hello all,
> 
> I've spent a bit of time learning the hoops and hurdles of 
> using tcl scripts to run Espresso. At present I have 
> customized a number of the source code interaction potentials 
> so that Espresso generates energies consistent with a 
> standard molecular mechanics force field (i.e. Allinger's MM3 
> which uses many anharmonic terms).  This should allow me to 
> use Espresso for both finer and coarser grain simulations and 
> also compare it with other molecular modelling packages.  The 
> molecule I am looking at is orthoterphenyl (a three ring 
> systems and a well-known fragile glass former).
> 
> At pressent I note that Espresso yields appropriate starting 
> energies for a single test molecule.  However, when I run the 
> NVT simulation e.g.
> "setmd time_step 0.00006
>  setmd skin 0.4
>  integrate set nvt
>  thermostat langevin 0.0001 1.0"
> the system energy (or temperature) first increases, next 
> decreses and then with subsequent iterations, on average, 
> increases considerably.  The simulation planarizes this OTP 
> molecule by increasing the interring bond angles. This 
> resulting molecular conformation is clearly not a low energy one.  
> I've tried a number of different time/thermostat trials but I 
> cannot figure out why this behavior should always be so.  
> 
> Would anyone have any ideas where I've gone wrong?
> 
> Here are typical condensed outputs from the start of a run:
>  0 E = 476.3665 kin =   7.747 tem =  0.6203 HAR 496.65 TOR 
> -48.32 BOND  16.18  
> subLJ -7275.60 Nonbond 7279.71             
>  1 E = 536.5164 kin =   7.940 tem =  0.6986 HAR 557.56 TOR 
> -49.64 BOND  17.60  
> subLJ -8316.61 Nonbond 8319.67             
>  2 E = 584.2990 kin =   7.498 tem =  0.7608 HAR 605.32 TOR 
> -50.76 BOND  19.63  
> subLJ -9229.96 Nonbond 9232.57             
>  3 E = 600.5471 kin =   7.850 tem =  0.7820 HAR 620.55 TOR 
> -52.07 BOND  21.71  
> subLJ -9398.53 Nonbond 9401.03             
>  4 E = 623.2451 kin =   7.605 tem =  0.8115 HAR 643.59 TOR 
> -53.27 BOND  23.01  
> subLJ -9764.05 Nonbond 9766.36             
>  5 E = 638.5015 kin =   7.423 tem =  0.8314 HAR 658.70 TOR 
> -54.25 BOND  24.45  
> subLJ -10191.71 Nonbond 10193.89           
>  6 E = 667.8782 kin =   7.968 tem =  0.8696 HAR 687.21 TOR 
> -55.15 BOND  25.76  
> subLJ -10746.96 Nonbond 10749.06           
>  7 E = 670.9218 kin =   7.748 tem =  0.8736 HAR 688.72 TOR 
> -55.83 BOND  28.25  
> subLJ -10719.87 Nonbond 10721.90           
>  8 E = 671.9178 kin =   7.832 tem =  0.8749 HAR 688.26 TOR 
> -56.48 BOND  30.33  
> subLJ -10827.99 Nonbond 10829.97           
>  9 E = 668.8554 kin =   7.219 tem =  0.8709 HAR 685.62 TOR 
> -57.30 BOND  31.38  
> subLJ -10683.19 Nonbond 10685.12           
>  10 E = 657.8612 kin =   7.559 tem =  0.8566 HAR 673.60 TOR 
> -58.08 BOND  32.75  
> subLJ -10439.64 Nonbond 10441.67          
>  11 E = 677.1924 kin =   7.862 tem =  0.8818 HAR 692.41 TOR 
> -58.74 BOND  33.70  
> subLJ -10730.58 Nonbond 10732.54          
>  12 E = 705.0812 kin =   7.499 tem =  0.9181 HAR 719.37 TOR 
> -59.42 BOND  35.72  
> subLJ -11236.45 Nonbond 11238.36          
>  13 E = 705.2942 kin =   7.727 tem =  0.9184 HAR 719.10 TOR 
> -60.20 BOND  36.68  
> subLJ -11551.19 Nonbond 11553.18          
>  14 E = 702.9975 kin =   7.822 tem =  0.9154 HAR 715.65 TOR 
> -60.60 BOND  38.11  
> subLJ -11735.88 Nonbond 11737.89          
>  15 E = 709.0762 kin =   8.061 tem =  0.9233 HAR 721.43 TOR 
> -61.43 BOND  39.08  
> subLJ -11785.12 Nonbond 11787.05          
>  16 E = 705.3679 kin =   7.737 tem =  0.9184 HAR 718.43 TOR 
> -62.23 BOND  39.52  
> subLJ -11737.30 Nonbond 11739.22          
>  17 E = 683.9043 kin =   7.667 tem =  0.8905 HAR 698.70 TOR 
> -62.84 BOND  38.43  
> subLJ -11354.50 Nonbond 11356.43          
>  18 E = 687.4399 kin =   7.171 tem =  0.8951 HAR 702.18 TOR 
> -63.24 BOND  39.39  
> subLJ -11471.37 Nonbond 11473.31          
>  19 E = 675.0700 kin =   7.678 tem =  0.8790 HAR 689.93 TOR 
> -63.90 BOND  39.39  
> subLJ -11329.82 Nonbond 11331.79          
>  20 E = 670.8316 kin =   7.703 tem =  0.8735 HAR 683.56 TOR 
> -64.50 BOND  42.10  
> subLJ -10756.45 Nonbond 10758.43          
>  21 E = 653.4761 kin =   7.491 tem =  0.8509 HAR 665.92 TOR 
> -65.02 BOND  43.03  
> subLJ -10436.87 Nonbond 10438.93          
>  22 E = 630.6521 kin =   7.743 tem =  0.8212 HAR 642.06 TOR 
> -65.83 BOND  44.55  
> subLJ -9983.90 Nonbond 9986.03            
>  23 E = 630.8224 kin =   7.731 tem =  0.8214 HAR 640.89 TOR 
> -66.29 BOND  46.40  
> subLJ -10065.90 Nonbond 10067.98          
>  24 E = 619.5612 kin =   7.413 tem =  0.8067 HAR 628.01 TOR 
> -66.81 BOND  48.81  
> subLJ -9748.74 Nonbond 9750.88            
> ...
>  311 E = 835.4695 kin =   7.666 tem =  1.0879 HAR 361.56 TOR 
> -115.79 BOND 
> 577.46  subLJ -6919.05 Nonbond 6923.62
> 
> The BOND energy is the one that is problematic.
> 
> Many thanks in advance for any suggestions.
> 
> Michael
> 
> 
> -- 
> 
> //////////////////////////////////////////////////////////////
> /////////////////////////////////////////
> / Michael Winokur, Phone:(608)263-7475          
> / Department of Physics    FAX:  (608)265-2334                  
> / University of Wisconsin  Email:address@hidden  
> / Madison, WI 53706         also: address@hidden  
> / Home Page: http://romano.physics.wisc.edu 
> //////////////////////////////////////////////////////////////
> //////////////////////////////////////////
> 
> 
> _______________________________________________
> ESPResSo mailing list
> address@hidden
> https://fias.uni-frankfurt.de/mailman/listinfo/espresso
>