Welcome To Support Community

Materials Studio

Advanced Search
Ask Search:
karankaran 

Extract Forcefield constants of PCFF

Hello all,

Is there any way to extract PCFF forcefield constants from Materials studio?
I saved PCFF forcefield in the Materials studio working window from " Modules - Forcite - Forcefield manager", then I tried to find out different parameters of PCFF,  but I couldn't find any value for different structure. For PCFF forcefield constant I mean the constant of energy parameters such as bend Torsion, stretch bend stretch and etc.

for example, I know the R Theta R formulation for Stretch Bend Stretch, but how can I find the constants of the formulation (Kij0 and Kjk0)?? 

Thank you very much,
Bests,
Karan
Best Answer chosen by karan
Jason DeJoannisJason DeJoannis
Hi karan,
 
Yes you can get the PCFF forcefield parameters in Materials Studio. They are viewable and editable in the Forcefield Editor and also accessible in the scripting api.
 
For instance, the stretch-bend-stretch potential with the functional form:
User-added image
has two distinct parameters, Kij0 and Kjk0. (The other two parameters q0 and Rij0/Rjk0 come from the angle and stretch potentials.)
 
They are only visible in the Forcefield Editor when you change the Functional Form filter to “R Theta R”:
User-added image 
Best,
Jason
 

All Answers

Jason DeJoannisJason DeJoannis
Hi karan,
 
Yes you can get the PCFF forcefield parameters in Materials Studio. They are viewable and editable in the Forcefield Editor and also accessible in the scripting api.
 
For instance, the stretch-bend-stretch potential with the functional form:
User-added image
has two distinct parameters, Kij0 and Kjk0. (The other two parameters q0 and Rij0/Rjk0 come from the angle and stretch potentials.)
 
They are only visible in the Forcefield Editor when you change the Functional Form filter to “R Theta R”:
User-added image 
Best,
Jason
 
This was selected as the best answer
karankaran
Dear Jason,
Thank you a lot for your accurate response. Your answers always are helpful.

Now I have faced another problem concerning PCFF forcefield in Materials Studio. For some structures, the existed parameters have not defined in the PCFF database. For example, when you want to simulate epoxy with PCFF in Materials Studio, the (cp c cp) or (cp oc c2) for bond-bond and bond-angle, (cp c cp cp) or (cp oc c2 hc) and .... for Torsion and dihedral (and ...) exist in the structur, but PCFF parameters have not been defined in the PCFF database! Surprisingly, Materials Studio shows no error for any data missing and the simulation will be finished without any problem! 

Now, here is the question. How does Materials studio calculate these constants? Does it ignore them?! If no, how can we find out the considered values for these parameters in PCFF?

​Thank you again for your time and concern.

Bests,
Karan
karankaran
I think some equivalences have been considered in the Forcefiled database! (as shown in equivalence tab in forcefield manager). Am I right?? 

User-added image
Is there any other assumption?!

Thank you
Bests,
Karan 
Jason DeJoannisJason DeJoannis
Hi Karan,

Yes, the equivalence table is what determines the mapping. The cp-c-cp-cp torsion you mentioned actually does have parameters directly.

However the cp-oc-c2-hc torsion gets mapped onto the parameters for cp-o-c-h, since that is its Level 2 torsion equivalent.

Best,
Jason
karankaran
Thank you again for your answer.

And the last question. what is the difference between a Double bond and Partial Double bond in Materials studio? For example in the epoxy scripting code, you sketch the initial structure of DGEBA and DEDTA with benzene ring of Partial Double Bond, while we know that the benzene rings in these structures have been consisted  of three single and three double bonds! The difference in the initial structures leads to different partial charges of atoms! 
How can we explain this situation?

Thank You,
Bests,
Karan
Jason DeJoannisJason DeJoannis
Hi Karan,

Partial double bonds represent resonance and so are used in aromatic rings.

I don't think you are right about the charges. Can you show me? The image below demonstrates that PCFF assigns the same types in both cases.

User-added image
karankaran
Hi Jason,

The forcefield types for atoms are the same, but the partial charges of atoms are different. Please see the following images for DEDTA structure with partial double bond and combination of single and double bond in Benzene ring. 
The charges in both structure are assigned by "forcefield assign" (PCFF). 
Am I right?

By the way, for assigning atom charges in polymers (especially epoxy system), is it better to use "forcefield assign" or " Charge using Qeq"? It seems that Qeq calculate charge more precisely. 

Thank you again for your  valuable information and help.

User-added image
User-added image
Bests,
Karan
Jason DeJoannisJason DeJoannis
That is not what PCFF assigns in MS2017R2. 

Also it doesn't make sense. If the types are the same, the charges will be the same, since it is using the charge increments method.

Also take a look at your charges. They should be identical for atoms that are structurally the same. Instead you have small differences like -0.644 vs -0.645 and 0.137 vs 0.144. That suggests they were assigned with QEq which takes the 3D structure into account.
karankaran
Oh, you're right. I made a terrible mistake! I had used different charge methods for each!

For "using Qeq method" we have:  
User-added image

User-added image

Of course the difference in atom charges are negligible. 

By the way, for assigning atom charges in polymers (especially epoxy system), is it better to use "forcefield assign" or " Charge using Qeq"? It seems that Qeq calculate charge more precisely. 
As you said,  QEq  takes the 3D structure into account. Is that better or not?

Best Regards,
Karan
karankaran
Hi Jason,
Sorry I have another question, too! 

As I read in the help of Materials studio, to find un-existed bond, dihedral, torsion and ..., Materials studio go through step down algorithm. For example, for PCFF forcefield, it has defined levels of equivalences for each forcefiled type. Using step down algorithm for PCFF, the appropriate parameters will be found!
For instances, for cp forcefield type, level 2 is cp_ and level 3 is X:
User-added image 
Now, here are the questions:
1- what is the meaning of "cp_"? I can't find cp_ type in PCFF! (There is same problem for c_ and etc).

2- I know that "X" is for dummy forcefield type. But there is not any combination for dummy types in most of the forcefield parameters!

3- For example for stretch-stretch types, how can we map cp-c-cp? (this structure is existed in epoxy formulation)

Thank you, and sorry for the inconvenient 

Bests,
Karan 
Jason DeJoannisJason DeJoannis
Hi Karan,
 
1) There is indeed a cp_ and c_ in pcff. You can find them by entering “c*_” in the filter box of the Types tab.
 
2) There are plenty of interactions defined using the X type. As a dummy atom, X has no vdW interactions. But it does have bond, angle and torsion interactions. It has no cross-terms.
 
3) Yes, cp-c-cp does not have a Stretch-Stretch interaction. That can be confirmed by sketching the structure and running a Forcite | Energy job (use existing types) and you will see that the Stretch-Stretch energy is zero.
 
You will notice that there is no explicit section defining Stretch-Stretch equivalences. Instead they are defined by the equivalence of their constituent parts. For instance, a c3m-c-c= Stretch-Stretch interaction is mapped onto c-c-c=, since c3m has a Stretch equivalence to c.
 
Best,
Jason
 
karankaran
Hi again,

So if we want to convert data from Materials Studio to other MD software such as LAMMPS, then should we put the un-mapped parameters to zero?
for example, you said that "cp-c-cp does not have a Stretch-Stretch interaction" does this mean cp-c-cp really does not have stretch stretch interaction or Materials studio does not consider (or can not map it). because we have cp-c-cp in our structure!

Please answer this question too : "By the way, for assigning atom charges in polymers (especially epoxy system), is it better to use "forcefield assign" or " Charge using Qeq"? It seems that Qeq calculate charge more precisely. 
As you said,  QEq  takes the 3D structure into account. Is that better or not?"


Thank you again for your time and help.

Yours Sincerely,
Karan
karankaran
I look forward to answer my question

I really appreciate your time and concern
Jason DeJoannisJason DeJoannis
Hi Karan,

The BIOVIA implementation of PCFF is the original, so it should be taken as the reference for other codes like LAMMPS. By the way, I have implemented a LAMMPS conversion for PCFF and Dreiding and posted it on the ScienceCloud Protocol Exchange. 

To answer your specific questions:
  • Yes interactions of zero should be used
  • Yes the PCFF-assigned charges should be used
Best,
Jason
karankaran
Dear Jason,

I don't know how to thank you for your helps. I really appreciate it.

I went through ScienceCloud Exchange site and searched for the LAMMPS conversion for PCFF by you. I found the following link:

https://exchange.sciencecloud.com/exchange/browse#details:224/Solubility Parameter Lammps

Is that the file you mean? It seems that there is a problem with the attached file! The following error is seen after I try to download the file:
User-added image

Is there any other way to download the implemented conversion file?

Thank you again,

Yours Sincerely,
Karan
karankaran
Hello again,

How does dummy forcefield type (X) work in equivalences part of forcefield?
In other words, what does dummy forcefield type mean?! Unfortunately, I don't get the idea of dummy forcefiled type! 
for example, for bond stretch equivalence of all forcefield types, X has been defined as the Level 3. Then for a bond with no match of Level 1 and level 2, how level 3 is defined?! We don't have X~X interaction for sure!

Thank you again Jason.

Best regard,
Karan 
Jason DeJoannisJason DeJoannis
Hi Karan,

I have updated the zip file on ScienceCloud so it should work now.

I don't understand you question about dummy atoms. They are handled the same as any other atom type. What type of bond are you looking at?

Jason
karankaran
Hi Jason,
Thank you for your response.

For example, in epoxy/TETA system I can't find following parameters:

Bend-Bend
hc cp cp cp
cp cp cp oc
c2 na c2 c2
c3 c cp cp
....
or

Stretch Torsion Stretch
Torsion Stretch

cp c cp cp
cp cp oc c2
cp oc c2 hc
cp oc c2 c1
cp oc c2 c3h
na c2 c1 oh

These are some examples in the system. what should I do for these parameters (with PCFF forcefield)?? 

Thank you very much,

Karan
 
Jason DeJoannisJason DeJoannis
Hi Karan,
 
Let’s take a look at the first two Bend-Bend interactions:
  • hc cp cp cp
  • cp cp cp oc
 
These can be decomposed into the Angle Bend interactions:
  • hc cp cp
  • cp cp cp
  • cp cp oc
 
So let’s figure out the Bend-Bend equivalency. At Level 1, for Angle Bend equivalents, we have:
  • hc = h
  • oc = o
  • cp = cp
 
Now we look for the following Bend-Bend interactions:
  • h cp cp cp
  • cp cp cp o
 
It turns out that both are defined and the functional form “Ignore” is specified. So these interactions are ignored.
User-added image
 
Best,
Jason