# a question about potential of different group

hi, Steve,
There are some questions on my project,
I build a DLC model (model-1) by liquid quenching method, and the model consists of three group: group-DLC, group-sub, group-boundary. All the atoms is carbon, and when i depostion the DLC, the potential between the carbon is tersoff.

I use the read_data commond read the data of model-1, to build the model-2. And compute the interfacial strength of the group-DLC and group-sub weth the commond ‘compute group/group’, the potention of the DLC and DLC is tersoff , the potential of sub and sub is tersoff, the potential of boundary and boundary is tersoff, but the potential of DLC and sub is morse(cutoff is 4), then the DLC comes off the sub.

By now, I know the reason of the DLC depart from the sub is the defferent potential of the DLC and sub. In model-1, the potential of DLC and sub is tersoff. In model-2, the potenial of DLC and sub is morse.

my question is whether there is commond to compute the interfacial strength of defferent group,

terfacial strength,

confusing, then please respond to it and I am happy to clarify it.
I don't want to discourage you from asking for a second opinion from
others, but do you realize you have posted this question 10 times in
the last few hours to the entire mailing list?

There is a big problem with your question:
The tersoff potential is not pairwise-additive. It does not make
sense to compute the interaction energy between two groups of atoms if
the forces between them are not pairwise-additive. You have to think
more about what you want to know. What is the "interfacial strength"
between two groups? You have to define this. In my last reply to
your question I suggested you use this definition:
E(DLC+sub) - ( E(DLC) + E(sub) )
In my earlier email I included some instructions how to calculate this
in LAMMPS.
Let me know if you have any questions.

Andrew

Hi, Andrew,

Thanks for your reply, I am sorry to submit the same question for 8 times,
In the model-2, there are 4 group, in the order from the top to the bottom, the group-ball,
the group-DLC, the group-sub, the group-boundary. The group-ball is used to sliding over
the DLC under certain loads and velocity. The group-boundary is fixed.
The DLC is deposited on the group-sub, so there are bonds between atoms of the DLC and
group-sub, I want to know if the DLC would be detach from the group-sub under the sliding of the ball,
that is the ‘interfacial strength’ of the group-DLC and group-sub.
So how to calculate the interfacial force between group-DLC and group-sub, and between the group-ball and the group-DLC?

Thanks
King
Harbin Institute of Technology
China

the potention of the DLC and DLC is tersoff,
the potential of sub and sub is tersoff,
the potential of boundary and boundary is tersoff,
but the potential of DLC and sub is morse(cutoff is 4),
then the DLC comes off the sub.

So in model2, are you using
pair_style hybrid tersoff morse ?
A new question: Do all of the carbon atoms have the same type number
in your simulation? Are you sure that you successfully turned off the
tersoff interactions between "DLC" and "sub"?

It is difficult to be certain what you are doing. Can you post the
pair_style & pair_coeff commands from your input script? (Have you
tried using them in LAMMPS yet?)

Hi, Andrew,

In the model-2, there are 4 group, in the order from the top to the bottom,
the group-ball,
the group-DLC, the group-sub, the group-boundary. The group-ball is used to
sliding over
the DLC under certain loads and velocity. The group-boundary is fixed.
The DLC is deposited on the group-sub, so there are bonds between atoms of
the DLC and
group-sub,

What do you mean by "bonds"?
I thought you were using pair_style tersoff, or pair_style morse

Are you instead using bond_style morse?

I want to know if the DLC would be detach from the group-sub
under the sliding of the ball,
that is the 'interfacial strength' of the group-DLC and group-sub.
So how to calculate the interfacial force between group-DLC and group-sub,
and between the group-ball and the group-DLC?

Perhaps I should apologize too. Perhaps I completely misinterpreted
your question. When you say "interfacial strength" I just thought you
meant the interaction energy between group "DLC" and group "sub". But
if you are trying to calculate the shear (or tensile) interfacial
strength (at the boundary between different crystal phases of carbon),
then that is something I don't know how to calculate.

I suspect you should run a simulation using something like "fix
deform" (LAMMPS Molecular Dynamics Simulator) or "fix move"
(LAMMPS Molecular Dynamics Simulator) and measure the forces on
the groups of atoms as your structure deforms and breaks. This way
you could generate a stress-strain curve and calculate the area
underneath.

If so, then to calculate the force acting on a group of atoms, you can
save them in your dump file by including fx fy fz in the list of
arguments to the "dump" command and read them later. You could do
this, for example using:

dump 1 all custom 500 DUMP_FILE id type x y z ix iy iz fx fy fz

This works even if you are using pair_style tersoff.
Then after your simulation, you would have to write a script to read
this file and sum these forces over the two groups of atoms to
calculate the total force on each group. (Perhaps pizza.py can help
you do this.)

(You are correct that can not use compute group/group to measure the
force between groups if you are using the tersoff potential between
them. In one of your examples, model1, it seemed as though this is
what you were doing.)

Hope this helps
Andrew

So in model2, are you using pair_style hybrid tersoff morse ?

Yes, in the model-2, I use the pair_style hybrid tersoff morse 4,
in the model-1, I use pair_style tersoff, and the pair_coeff * * SiC_1994.tersoff C C C

A new question: Do all of the carbon atoms have the same type number

And in the model-2, there are 4 types of atoms, type-1 is group-DLC, type-2 is group-sub,
type-3 is group-boundary, type-4 is group-ball
In the model-1, there are 3 types of atoms, type-1 is group-DLC, type-2 is group-sub,
type-3 is group-boundary.

Can you post the pair_style & pair_coeff commands from your input script? (Have you
tried using them in LAMMPS yet?)

Yes, in the model-2, the pair_style & pair_coeff commands is showed below,
and i have using them in LAMMPS,

pair_style hybrid tersoff morse 4
pair_coeff * * tersoff SiC_1994.tersoff C C C C
pair_coeff 1 2 morse 3.68 1.165 3.25
pair_coeff 2 4 morse 3.68 1.165 3.25
pair_coeff 1 4 morse 3.68 1.165 3.25

Are you sure that you successfully turned off the
tersoff interactions between “DLC” and “sub”?

And I have used the pair_style hybrid tersoff morse in the model-2 sucessfully.
But there has some problem in the model-2, in the second step during the simulation,
the group-DLC depart form the group-sub, and in the end of the simulation, the distance
of the group-DLC and group-sub is about 2 Å。 But in the first step of the simulation, there no
gap between the group-DLC and group-sub. And in the model-1, there no gap between the group-DLC
and group-sub,too.

What do you mean by “bonds”?
I thought you were using pair_style tersoff, or pair_style morse
Are you instead using bond_style morse?

I use the pair_style commond in my simulation only.

When you say “interfacial strength” I just thought you

meant the interaction energy between group “DLC” and group “sub”. But
if you are trying to calculate the shear (or tensile) interfacial
strength (at the boundary between different crystal phases of carbon),
then that is something I don’t know how to calculate.

Yes, I want to know the shear and tensile interfacial strength at the boundary
between the group-DLC and group-sub. In the manual, the command ‘compute group/group’
could compute the the total energy and force interaction between two groups of atoms, but this
command is not suitable for 3-body and other many-body potentials, such as Tersoff .
that is why i use the pair_style hybrid tersoff morse in the model-2, but there is gap between
the group-DLC and group-sub.
Did I misunderstanding the command ‘compute group/group’ ?

I suspect you should run a simulation using something like “fix
deform” (http://lammps.sandia.gov/doc/fix_deform.html) or “fix move”
(http://lammps.sandia.gov/doc/fix_move.html) and measure the forces on
the groups of atoms as your structure deforms and breaks. This way
you could generate a stress-strain curve and calculate the area
underneath.

Sorry, I can not understand the relationship the force and the command mention above
in the model-2, between the group-DLC and group-sub,I will read the manual careful to
understand the ‘fix deform’ and ‘fix move’ .

If so, then to calculate the force acting on a group of atoms, you can
save them in your dump file by including fx fy fz in the list of
arguments to the “dump” command and read them later. You could do
this, for example using:

dump 1 all custom 500 DUMP_FILE id type x y z ix iy iz fx fy fz

This works even if you are using pair_style tersoff.

Then after your simulation, you would have to write a script to read
this file and sum these forces over the two groups of atoms to
calculate the total force on each group. (Perhaps pizza.py can help
you do this.)

I would try to do what you have advised.

And thanks for you help so careful and kind

King
Harbin Institute of Technology
China

Hi, Andrew, this are somting I can not understand.

If so, then to calculate the force acting on a group of atoms, you can
save them in your dump file by including fx fy fz in the list of
arguments to the “dump” command and read them later. You could do
this, for example using:

dump 1 all custom 500 DUMP_FILE id type x y z ix iy iz fx fy fz

This works even if you are using pair_style tersoff.
Then after your simulation, you would have to write a script to read
this file and sum these forces over the two groups of atoms to
calculate the total force on each group. (Perhaps pizza.py can help
you do this.)

Using the command ‘dump’, then the ‘fx, fy, fz’ is the force on the atoms,
‘dump 1 all custom 500 DUMP_FILE id type x y z ix iy iz fx fy fz’ is to dump
the forces on all the atoms, then the sum of the forces on the atoms of group-DLC
is the force on the group-DLC, is it right ? If not, then what is the mean of ‘‘sum these
forces over the two groups of atoms to calculate the total force on each group’’ ?

Thanks for you help .

King
Harbin Institute of Technology

China

This question has split into multiple questions.

I will rephrase and repost part of your question later.

I am too busy to reply to any more questions on this topic, but I hope
I helped you.

And in the model-2, there are 4 types of atoms, type-1 is group-DLC,
type-2 is group-sub,
type-3 is group-boundary, type-4 is group-ball
In the model-1, there are 3 types of atoms, type-1 is group-DLC, type-2 is
group-sub,
type-3 is group-boundary.

in the model-2, the pair_style & pair_coeff commands is showed below,
and i have using them in LAMMPS,
pair_style hybrid tersoff morse 4
pair_coeff * * tersoff SiC_1994.tersoff C C C C
pair_coeff 1 2 morse 3.68 1.165 3.25
pair_coeff 2 4 morse 3.68 1.165 3.25
pair_coeff 1 4 morse 3.68 1.165 3.25

But there has some problem in the model-2, in the second step during the
simulation,
the group-DLC depart form the group-sub, and in the end of the
simulation, the distance
of the group-DLC and group-sub is about 2 Å。 But in the first step of the
simulation, there no
gap between the group-DLC and group-sub. And in the model-1, there no gap
between the group-DLC
and group-sub,too.

I do not know why this happened.
When there are physical problems with the model, I always have the
same suggestions:

Are you sure your morse energy parameters are correct?
If these are in units of kJ/mole, the numbers seem small to me.

Your "D0" energy in your morse potential is 3.68. Are these in
kCal/mole? ("units real")?
I ask because typical carbon-bond dissociation energies are over 356
kJ/mole (85kCal/mole).

Also: Have you tried visualizing your dump file?
(Run a short simulation and save frequently.)
If you can create a LAMMPS DATA file for your system (you can use
restart2data), then you can use VMD to visualize your trajectory. I
have attached a README file containing some general instructions for
doing this.

that is why i use the pair_style hybrid tersoff morse in the model-2, but
there is gap between
the group-DLC and group-sub.
Did I misunderstanding the command 'compute group/group' ?

I don't know. Good question. I'll post this question separately.

Sorry, I can not understand the relationship the force and the command
mention above
in the model-2, between the group-DLC and group-sub,I will read the manual
careful to
understand the 'fix deform' and 'fix move' .

Yes, I want to know the shear and tensile interfacial strength at the boundary
between the group-DLC and group-sub.

I looked up the definition of "interfacial strength" using google.
should let someone more knowledgeable answer it.

According to the web pages I found using google, the "interfacial
strength" is proportional to the area under the "stress-strain" curve
during fracture. Is this what you are trying to do? Are you trying
to break your system and measure the force? If this is not what you
are doing, then please ignore my previous email.

If you are trying to break your system, then you can use "fix move" or
"fix deform", to pull the atoms apart. These fixes can hold on to the
atoms at the top and bottom of your simulation box and force them to
move (to break the system). As this is happening, you can estimate
the stress by calculating the total force on atoms at the top and
bottom as a function. You can do this using the dump command. You do
not have to use "compute group/group".

>If so, then to calculate the force acting on a group of atoms, you can
>save them in your dump file by including fx fy fz in the list of
>arguments to the "dump" command and read them later. You could do
>this, for example using:

>dump 1 all custom 500 DUMP_FILE id type x y z ix iy iz fx fy fz

>This works even if you are using pair_style tersoff.

>Then after your simulation, you would have to write a script to read
>this file and sum these forces over the two groups of atoms to
>calculate the total force on each group. (Perhaps pizza.py can help
>you do this.)

I would try to do what you have advised.

And thanks for you help so careful and kind

This question has split into multiple questions.
I will rephrase and repost part of your question later.
I am too busy to reply to any more questions on this topic, but I hope
I helped you.

Thanks for you help very much. I have know how to deal with my problems.

Are you sure your morse energy parameters are correct?
If these are in units of kJ/mole, the numbers seem small to me.

Your “D0” energy in your morse potential is 3.68. Are these in
kCal/mole? (“units real”)?
I ask because typical carbon-bond dissociation energies are over 356
kJ/mole (85kCal/mole).

The units in my script is metal, so the unit of the energy is eV, but the
energy parattmeter of morse is readed from other papers.

Also: Have you tried visualizing your dump file?
(Run a short simulation and save frequently.)
If you can create a LAMMPS DATA file for your system (you can use
restart2data), then you can use VMD to visualize your trajectory. I
have attached a README file containing some general instructions for
doing this.

Yes, I dump the ‘dump.lammpstrj’ file use the command ‘dump’ , and visualize
this file using VMD.

I don’t know. Good question. I’ll post this question separately.

Thank you very much.

I looked up the definition of “interfacial strength” using google.
should let someone more knowledgeable answer it.

According to the web pages I found using google, the “interfacial
strength” is proportional to the area under the “stress-strain” curve
during fracture. Is this what you are trying to do? Are you trying
to break your system and measure the force? If this is not what you
are doing, then please ignore my previous email.

If you are trying to break your system, then you can use “fix move” or
“fix deform”, to pull the atoms apart. These fixes can hold on to the
atoms at the top and bottom of your simulation box and force them to
move (to break the system). As this is happening, you can estimate
the stress by calculating the total force on atoms at the top and
bottom as a function. You can do this using the dump command. You do
not have to use “compute group/group”.

Thanks, I have done what you adivsed, to calulate the force on atoms.