Lateral pressure simulation test: NPT balance for negative external pressure

Hello everyone!
I try to perform a “lateral pressure simulation test” to measure v21 of carbon nanotubes (CNTs). I found in the the mail list that " if that pressure is positive, it will try to compress the system, if it is negative, it will expand the system. since the pressure at inifinite dilution is zero, a negative external pressure will always expand your system. " “thus you can have a balance between internal and external pressure only for positive pressure values.”

  1. so I can not balance the system in Negative pressure and make CNTs enlarge (with command fix 1 all npt temp 4.3 4.3 1 y 0 0 1 x -5 -5 5 z -5 -5 5 couple xz) Am I right?

  2. then if i set a positive pressure value, box will shrink and finally CNTs turn into the shape of a square due to periodic boundary conditions.

Now i have no idea how to simulate that. Any comments will be highly appreciated. Thanks a lot!

units 		   metal
dimension  	   3
boundary 	   s  p	 s
atom_style	 atomic
neighbor     2.0 nsq      
neigh_modify every 1 delay 0 check yes
read_data CNt_sortY25-0_103.xyz                  

pair_style airebo 3.0 1 1		 
pair_coeff * * CH.airebo C     

reset_timestep	0
timestep        0.0005 
min_style       cg
minimize 1.0e-6 1.0e-6 10000 1000

change_box all boundary p p p
change_box all z delta -10 10 units box
change_box all x delta -10 10 units box

velocity all create 4.3 4928459 dist gaussian mom yes rot yes                           

reset_timestep	0
fix     1 all npt temp 4.3 4.3 0.1 y 0 0 1  x 0 0 1  z 0 0 1 couple xz                                 
dump    1 all custom 20 relaxation-deform.dat id type x y z
 
run		  200000   
unfix   1
undump  1
# ------------------------ DEFORMATION-------------------------------------- 
fix     2 all npt temp 4.3 4.3 1 y 0 0 1  x 0  -5  5  z 0  -5  5 couple xz
run		  200000                                                               
unfix   2                                                                   
fix     2 all npt temp 4.3 4.3 1 y 0 0 1  x -5 -10 5  z  -5 -10 5 couple xz
run		  200000                                                              
unfix   2                                                                   
fix     2 all npt temp 4.3 4.3 1 y 0 0 1  x -10 -15 5  z  -10 -15 5 couple xz
run		  200000                                                               
unfix   2                                                                   
fix     2 all npt temp 4.3 4.3 1 y 0 0 1  x -15 -20 5  z  -15 -20 5 couple xz
run		  400000                                                               
unfix   2                                                                   
fix     2 all npt temp 4.3 4.3 1 y 0 0 1  x -20 -25 5  z  -20 -25 5 couple xz
run		  600000                                                               
unfix   2                                                                   
fix     2 all npt temp 4.3 4.3 1 y 0 0 1  x -25 -30 5  z  -25 -30 5 couple xz
run		  800000                                                              
unfix   2                                                                                                                                    

print "All done"

It is more common to use fix deform on the dimension you want to strain by compressing or expanding and use fix npt only to adjust the other dimensions.

thank you Axel
I understand what you mean. In axial direction of CNTs, fix deform is common and easy to implement. BUT in the radial direction how to “strain” ?
thanks again Best wishes

The way fix deform changes the box and scales the coordinates is exactly the same as how fix npt does it. So i don’t see a difference there.