Re: [ESPResSo-users] FENE+LJ bond

[David Sean]

Hey Ricky,

Like Peter mentioned, using a FENE+LJ(r_cut=2.5) will be the same as
to FENE+LJ(r_cut=1.5). With typical parameters, FENE will not allow
bond stretching more that 1.5.

If this is is not the case for you, you might want to explore this
hacky solution: You use LJ with a cutoff of 2.5 with between all
beads. You then use an FENE-like exotic bond for the polymer: it has a
standard FENE potential but also a subtractive part that removes the
extra unwanted LJ between the distance of of 1.5 and 2.5.

The end result is FENE+LJ(r_cut=1.5) for all bonded beads and
LJ(r_cut=2.5) for all beads.

To implement this, you can use i) a tabulated potential (easiest, but
a bit slower solution), ii) modify the existing FENE potential to
include the LJ contribution you want (hack solution), or iii) create
your own potential and add it to ESPReSsO for other to use (best
solution, but harder to do).

Before you try any of these, you may want to look at the "subtracted
LJ" bond 
http://espressomd.org/html/doc/inter_bonded.html#subtracted-lennard-jones-bond
If you are lucky, you may be able to get it to only subtract the extra
LJ piece you don't want.

An before any of that, plot FENE+LJ(r_cut=1.5) and FENE+LJ(r_cut=2.5)
for you params to see if they are really different.

David

Le mar. 21 mai 2019 à 07:45, 赵睿祺 <address@hidden> a écrit :
>
> Dear all,
>
>
>
> The reason why I use espresso is that I want to model the behavior under the 
> magnetic field.
>
> But I have had the configuration of the system after relaxation without 
> magnetic field using another software. And the cutoff I use is 2^(1/6) with 
> bonded beads and 2.5 with nonbonded beads with that software. I don’t want to 
> waste the configuration I had.
>
> So now I'd like to add a new bond called "FENE+LJ" in espresso. If it's too 
> difficult then I would change to use the attractive cutoff for all the 
> nonbonded interaction.
>
> Thanks a lot for all your reply!
>
>
>
> Regards!
>
> Ricky
>
>
>
>
> -----原始邮件-----
> 发件人:"Peter Košovan" <address@hidden>
> 发送时间:2019-05-21 21:39:00 (星期二)
> 收件人: "address@hidden" <address@hidden>
> 抄送: "address@hidden" <address@hidden>
> 主题: Re: Re: [ESPResSo-users] FENE+LJ bond
>
> You could use exclusions to prevent bonded pairs from interacting with the 
> default LJ but still you would need to separately add the WCA for each pair 
> of bonded particles.
>
> As I said, setting a different particle type for each polymer particle is not 
> a good solution. Having 4000 particle types will probably slow down your 
> simulation.
>
> I engourage you again to ask the authors how precisely they did it.
>
> Finally, setting fene + LJ instead of FENE + WCA for bonded interactions 
> makes probably very little difference because LJ and WCA differ only at 
> values of r where FENE >> kT approx LJ > (WCA=0). Try plotting the potentials 
> to see what I mean. So you can safely use FENE+LJ isntead of FENE+WCA without 
> observing a difference in your results.
>
> peter
>
> On Tue, May 21, 2019 at 3:33 PM address@hidden <address@hidden> wrote:
>>
>> I tired to set different type of each particle, since the number of 
>> particles is about 40000 and the time is very long which leads me to find a 
>> new way.
>> I'm appreciate for your help.
>>
>> Best regards!
>> Ricky
>> ________________________________
>> address@hidden
>>
>>
>> From: Peter Košovan
>> Date: 2019-05-21 21:24
>> To: address@hidden
>> CC: address@hidden
>> Subject: Re: Re: [ESPResSo-users] FENE+LJ bond
>> I do not know how it works in LAMMPS that the authors used but in Espresso 
>> you can only set LJ or WCA between all pairs of particles of a given type. 
>> That makes good sense because short-range interactions of the same type of 
>> particles should be the same. You cannot set WCA only between pairs of 
>> polymer beads which are connected by a bond.
>>
>> If you insist on it, you can set a different type to each particle of your 
>> polymer, and then explicitly set WCA for each pair of bonded particles. But 
>> that will make your script complicated with no real benefit.
>>
>> Greetings,
>>
>> peter
>>
>> On Tue, May 21, 2019 at 3:11 PM address@hidden <address@hidden> wrote:
>>>
>>> As the paper said,
>>>     "For the classical Kremer and Grest (KG) model, only a repulsive force 
>>> is employed, i.e., Rc = 2^(1/6)σ. Such a model is close to an athermal 
>>> model of polymer, lacking the ability to predict the glass transition 
>>> driven by temperature. Furthermore, some simulation results reveal a 
>>> negative thermal expansion coefficient without attraction."
>>> Because I want to predict the Tg,  I think the cutoff should be 2.5sigma.  
>>> I think I should add the fene+lj bond on my own. Do you have some better 
>>> ideas?
>>>
>>> In addition, Do you mean that you would use WCA between all the polymer 
>>> beads?
>>>
>>> Thanks so much!
>>>
>>> Best regards!
>>> Ricky
>>> ________________________________
>>> address@hidden
>>>
>>>
>>> From: Peter Košovan
>>> Date: 2019-05-21 20:59
>>> To: Dusan Racko
>>> CC: address@hidden; address@hidden
>>> Subject: Re: [ESPResSo-users] FENE+LJ bond
>>> Actually, they did not specify in the paper for which pairs of particles 
>>> they used the usual LJ potential (Eq.1).  They only say that the fene + LJ 
>>> (WCA) with cut=2^1/6 sigma) is used for polymer-polymer interactions. The 
>>> text suggests that they used Eq.1 for all particle pairs including the 
>>> polymer but that would be a very wierd setting.
>>>
>>> If you need to model exactly the same system, then I think that you should 
>>> ask the authors for clarification. Otherwise, I would use WCA potential for 
>>> polymer-polymer and LJ with cut=2.5sigma fro all other pairs.
>>>
>>> Greetings,
>>>
>>> peter
>>>
>>> On Tue, May 21, 2019 at 2:54 PM Dusan Racko <address@hidden> wrote:
>>>>
>>>> Hi, isn't that a cutoff for the attractive potential to model magnetic 
>>>> particles only? Sorry, I don't know the papers, just suggesting
>>>>
>>>> On Tue, May 21, 2019 at 2:51 PM address@hidden <address@hidden> wrote:
>>>>>
>>>>> Dear Peter,
>>>>>
>>>>> Actually, in the paper, the cutoff distance of LJ potential between 
>>>>> nonbonded beads is 2.5[sigma] while the cutoff distance of LJ potential 
>>>>> between bonded beads is 2^(1/6)[sigma]. This can't be done by your way.
>>>>> Thanks anyway!
>>>>>
>>>>> Regards!
>>>>> Ricky
>>>>>
>>>>> ________________________________
>>>>> address@hidden
>>>>>
>>>>>
>>>>> From: Peter Košovan
>>>>> Date: 2019-05-21 20:44
>>>>> To: address@hidden
>>>>> CC: address@hidden
>>>>> Subject: Re: [ESPResSo-users] FENE+LJ bond
>>>>> Hi Ricky,
>>>>>
>>>>> Just set up a FENE potential between the given pair of particles, and an 
>>>>> LJ potential between the given particle types (probably the same type) 
>>>>> and you get exactly what you want. This is the standard setup which is 
>>>>> just described in a weird way in the paper that you cited.
>>>>>
>>>>> The fene potential acts just between the two particles with a given id. 
>>>>> The LJ potential acts between any pair of particles. If your pair of 
>>>>> particles interacts both with fene and LJ, then you get the fene+LJ 
>>>>> interaction.
>>>>>
>>>>> See also the script  minimal-polymer.py in the samples directory.
>>>>>
>>>>>
>>>>> peter
>>>>>
>>>>> On Tue, May 21, 2019 at 10:37 AM address@hidden <address@hidden> wrote:
>>>>>>
>>>>>> Dear Sebastian,
>>>>>>
>>>>>> Thanks for your reply at first!
>>>>>> I want to model the hard-soft block copolymers with magnetic particles 
>>>>>> and the copolymer model I use is from this paper
>>>>>>    https://link.aps.org/pdf/10.1103/PhysRevE.88.022602
>>>>>> In this paper, the bond potential is FENE+LJ.  Beyond that, I can also 
>>>>>> find paper using FENE+LJ potentials in 
>>>>>> https://aip.scitation.org/doi/10.1063/1.458541. What's more, in LAMMPS, 
>>>>>> it also uses the FENE+LJ type fene bond.
>>>>>>
>>>>>> I will use the FENE+LJ bond with bonded beads and the LJ potential with 
>>>>>> nonbonded beads.
>>>>>>
>>>>>> Regards!
>>>>>> Ricky
>>>>>>
>>>>>> ________________________________
>>>>>> address@hidden
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Dr. Peter Košovan
>>>>>
>>>>> Department of Physical and Macromolecular Chemistry
>>>>> Faculty of Science, Charles University in Prague, Czech Republic
>>>>>
>>>>> Katedra fyzikální a makromolekulární chemie
>>>>> Přírodovědecká fakulta Univerzity Karlovy v Praze
>>>>>
>>>>> www.natur.cuni.cz/chemistry/fyzchem/
>>>>> Tel. +420221951029
>>>>> Fax +420224919752
>>>>
>>>>
>>>>
>>>> --
>>>> _____________________
>>>> Polymer Institute of the Slovak Academy of Sciences
>>>> Dubravska cesta 3
>>>> 845 41 Bratislava, Slovak Republic [‡]
>>>> tel: +421 2 3229 4321
>>>> RG: https://www.researchgate.net/profile/Dusan_Racko
>>>> Scholar: https://goo.gl/oRVrGV
>>>
>>>
>>>
>>> --
>>> Dr. Peter Košovan
>>>
>>> Department of Physical and Macromolecular Chemistry
>>> Faculty of Science, Charles University in Prague, Czech Republic
>>>
>>> Katedra fyzikální a makromolekulární chemie
>>> Přírodovědecká fakulta Univerzity Karlovy v Praze
>>>
>>> www.natur.cuni.cz/chemistry/fyzchem/
>>> Tel. +420221951029
>>> Fax +420224919752
>>
>>
>>
>> --
>> Dr. Peter Košovan
>>
>> Department of Physical and Macromolecular Chemistry
>> Faculty of Science, Charles University in Prague, Czech Republic
>>
>> Katedra fyzikální a makromolekulární chemie
>> Přírodovědecká fakulta Univerzity Karlovy v Praze
>>
>> www.natur.cuni.cz/chemistry/fyzchem/
>> Tel. +420221951029
>> Fax +420224919752
>
>
>
> --
> Dr. Peter Košovan
>
> Department of Physical and Macromolecular Chemistry
> Faculty of Science, Charles University in Prague, Czech Republic
>
> Katedra fyzikální a makromolekulární chemie
> Přírodovědecká fakulta Univerzity Karlovy v Praze
>
> www.natur.cuni.cz/chemistry/fyzchem/
> Tel. +420221951029
> Fax +420224919752