high pressure in NPT ensemble

LAMMPS LAMMPS Mailing List Mirror
1

Dear lammps users:

I constructed a water box that SiO2 membrane divided it two equal parts. the COM of SiO2 membrane is in origin (0 , 0 , 0). I do it with VMD. After minimization I set the fix NPT to equilibrate and to control pressure of system. When I use keyword iso (iso 1 1 1000) I faced with vacuum in my box ( waters pressed tow sides of box in -z and +z direction) and sometimes high pressure fluctuation, and when I use keyword z (z 1 1 1000) I faced with high pressure in my system and when I use
x 0 0 1000 y 0 0 1000 z 1 1 1000
I faced with change in x and y size of my membrane. I have 9500 particles in my box and dimensions of my box is 703535.

What is problem of my system ? How can I fix this problem?

I have also used nve, langevin and press/berendsen fix according to previous lammps mail list, but these command did not fix my pressure problems.

Best Regards
Xueqing Lu

2

Dear lammps users:
I constructed a water box that SiO2 membrane divided it two equal parts. the
COM of SiO2 membrane is in origin (0 , 0 , 0). I do it with VMD. After
minimization I set the fix NPT to equilibrate and to control pressure of
system. When I use keyword iso (iso 1 1 1000) I faced with vacuum in my box
( waters pressed tow sides of box in -z and +z direction) and sometimes high
pressure fluctuation, and when I use keyword z (z 1 1 1000) I faced with
high pressure in my system and when I use
x 0 0 1000 y 0 0 1000 z 1 1 1000
I faced with change in x and y size of my membrane. I have 9500 particles
in my box and dimensions of my box is 70*35*35.
What is problem of my system ? How can I fix this problem?
I have also used nve, langevin and press/berendsen fix according to previous
lammps mail list, but these command did not fix my pressure problems.

have you considered, that you have a very different compressibility
along different axes in your system. particularly, i doubt that there
is much of a need to relax the dimensions of your cell in the
directions of the SiO2 layer(s?). it is probably best to build the
system to retain the dimensions of a corresponding bulk SiO2 system
(unless the layers are extremely thin). a relaxation would then really
only be needed in the direction of the water layer (z direction?).
using an isotropic deformation is a very bad idea, since you don't
have an isotropic system. also, global pressure is rather meaningless,
you should look at the individual components of the pressure tensor
trace.

axel.

3

Dear Axel, thanks your reply.
My SiO2 layer is very thin, thus I use this command:
units real
fix 6 all npt temp 300.0 300.0 5.0 x 0.0 0.0 1000.0 y 0.0 0.0 1000.0 z 1.0 1.0 1000.0

and when I print the pressure tensor components every 500 steps for x, y & z component for 13 steps after minimization:

x:
-1299.3313908481218277
-579.64618525286527984
-357.21802859195821611
-179.82305364743194787
-435.48835406059009756
-235.05040419492291903
-41.206900548356486524
-34.929540453061058258
261.20964965639501543
-55.1594144253545835
72.91944458094900483
409.00005321021285454
29.068078882160964582

y:
855958.94613892945927
245513.85588570294203
105477.99443537011393
51452.71353299907787
26748.593631782241573
14677.597262392970151
8748.4943744813208468
5144.3168233466640231
3544.8587525681832631
2692.8575021731844572
1948.6108427437750379
1715.9906943607902576
1588.9759868298026504

z:
-1299.3313908481218277
-579.64618525286527984
-357.21802859195821611
-179.82305364743194787
-435.48835406059009756
-235.05040419492291903
-41.206900548356486524
-34.929540453061058258
261.20964965639501543
-55.1594144253545835
72.91944458094900483
409.00005321021285454
29.068078882160964582

and when I use “couple xy” in fix npt command, x+y is about 1 atm, and average of fluctuated z component is 1 atm but global pressure is very big.

and when I do not allocate any quantity to x and y component:
fix 6 all npt temp 300.0 300.0 5.0 z 1.0 1.0 1000.0

x:
168738138.29058641195
170296136.4341506362
172612182.13990220428
174538602.49550268054
176195729.579998523
177365742.27048924565
178967968.64061206579
179574172.39271244407
180971466.529053092
182004969.48433855176
181821918.2787591517
181585882.98385000229
182330983.37955382466

y:
339744927.7863008976
342883149.11074000597
347546684.02624231577
351424164.71335709095
354762479.55797356367
357116992.16631495953
360342948.72059506178
361563637.16639024019
364376083.68448036909
366457612.38392549753
366088341.51451385021
365614661.06039249897
367113610.14647561312

z:
222.64545079143118755
-395.87458437079163787
-799.38236319738746261
-431.91501496602268162
-274.70295517215646441
-548.29316845764572008
-351.74462625299514684
-659.81336176382660597
56.986159896164373606
510.42031054248093369
-282.63365223281579119
-358.87756637379408176
124.19432173492009497

what is your suggestion to fix problem in this system?

4

Dear Axel, thanks your reply.
My SiO2 layer is very thin, thus I use this command:
units real
fix 6 all npt temp 300.0 300.0 5.0 x 0.0 0.0 1000.0 y 0.0 0.0 1000.0 z 1.0 1.0 1000.0

and when I print the pressure tensor components every 500 steps for x, y & z component for 13 steps after minimization:

x:
-1299.3313908481218277
-579.64618525286527984
-357.21802859195821611
-179.82305364743194787
-435.48835406059009756
-235.05040419492291903
-41.206900548356486524
-34.929540453061058258
261.20964965639501543
-55.1594144253545835
72.91944458094900483
409.00005321021285454
29.068078882160964582

y:
855958.94613892945927
245513.85588570294203
105477.99443537011393
51452.71353299907787
26748.593631782241573
14677.597262392970151
8748.4943744813208468
5144.3168233466640231
3544.8587525681832631
2692.8575021731844572
1948.6108427437750379
1715.9906943607902576
1588.9759868298026504

z:
-1299.3313908481218277
-579.64618525286527984
-357.21802859195821611
-179.82305364743194787
-435.48835406059009756
-235.05040419492291903
-41.206900548356486524
-34.929540453061058258
261.20964965639501543
-55.1594144253545835
72.91944458094900483
409.00005321021285454
29.068078882160964582

and when I use “couple xy” in fix npt command, x+y is about 1 atm, and average of fluctuated z component is 1 atm but global pressure is very big.

and when I do not allocate any quantity to x and y component:
fix 6 all npt temp 300.0 300.0 5.0 z 1.0 1.0 1000.0

x:
168738138.29058641195
170296136.4341506362
172612182.13990220428
174538602.49550268054
176195729.579998523
177365742.27048924565
178967968.64061206579
179574172.39271244407
180971466.529053092
182004969.48433855176
181821918.2787591517
181585882.98385000229
182330983.37955382466

y:
339744927.7863008976
342883149.11074000597
347546684.02624231577
351424164.71335709095
354762479.55797356367
357116992.16631495953
360342948.72059506178
361563637.16639024019
364376083.68448036909
366457612.38392549753
366088341.51451385021
365614661.06039249897
367113610.14647561312

z:
222.64545079143118755
-395.87458437079163787
-799.38236319738746261
-431.91501496602268162
-274.70295517215646441
-548.29316845764572008
-351.74462625299514684
-659.81336176382660597
56.986159896164373606
510.42031054248093369
-282.63365223281579119
-358.87756637379408176
124.19432173492009497

what is your suggestion to fix problem in this system?

5

there is no need to repost your reply every six hours. remember there
are people on the list that are in different time zones and sometimes
they have to sleep.

Dear Axel, thanks your reply.
My SiO2 layer is very thin, thus I use this command:
units real
fix 6 all npt temp 300.0 300.0 5.0 x 0.0 0.0 1000.0 y 0.0 0.0
1000.0 z 1.0 1.0 1000.0
and when I print the pressure tensor components every 500 steps for x, y &
z component for 13 steps after minimization:
x:

[...]

29.068078882160964582
and when I use "couple xy" in fix npt command, x+y is about 1 atm, and
average of fluctuated z component is 1 atm but global pressure is very big.

and when I do not allocate any quantity to x and y component:
fix 6 all npt temp 300.0 300.0 5.0 z 1.0 1.0 1000.0
x:
168738138.29058641195

[...]

-358.87756637379408176
124.19432173492009497
what is your suggestion to fix problem in this system?

well, obviously your initial structure is not well equilibrated as far
as the sio2 layer is concerned, and the data your present here
indicate, that you can still use quite some more equilibration time.
the high pressure may be due to how you construct your system, or due
to your system parameters or a combination of those. there is not
enough information here to tell one from the other or identify yet
another possible reason.
overall, it is *your* simulation and you will have to defend and
explain the results and thus *you* have to make the call what is the
most reasonable choice.

axel.