Re: [gmx-users] Ewald summation in 2 dimensions with correction

[Dan Gil]

Hi,

I tried using the regular PME method with and without corrections. The
results don't depend on the choice of ewald-geometry  at all!

I used (Gromacs 5.1):
wall-ewald-zfac = 3
ewald-geometry = 3d ; This prompts a warning because I am using 2 walls
- 2d system. I forced this with maxwarn.
E-z= 1 5 0

and I also used
wall-ewald-zfac = 3
ewald-geometry = 3dc   ; No warnings since this is recommended
E-z= 1 5 0

I checked that this input was being processing correctly on topol.tpr. For
the first case, I see ewald-geometry = 0 and the second case ewald-geometry
= 1.

The electric field from ionic liquid has a magnitude of ~7.5 V/nm in the
opposite direction to the applied electric field for both cases. This is
greater than the applied 5. I obtained the electric field using gmx
potential.

I would very much appreciate any recommendations.

Dan


Virus-free.
www.avg.com

<#m_-5936464299898835406_DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

On Mon, Nov 13, 2017 at 10:22 AM, Dan Gil  wrote:

> Hi,
>
> I am using gmx potential to measure the electric field. I've reported the
> values from field.xvg in my previous email. The field at the edges of the
> ionic liquid near the walls will oscillate because of the electric double
> layer but it is the bulk field I am concerned with.
>
> I read the supporting information - I thought that the Ewald summation
> correction for 2 dimensions was supposed to cancel out interactions between
> periodic images in the z direction?
>
> Were you using the regular 3-D Ewald summation in this work? I might try
> carrying out the calculations that way.
>
> Dan
>
> On Mon, Nov 13, 2017 at 2:23 AM, David van der Spoel  > wrote:
>
>> On 11/11/17 16:44, Dan Gil wrote:
>>
>>> Hi,
>>>
>>> Just wanted to see if anyone had any thoughts...
>>>
>>> For more information, if I use:
>>> wall-ewald-zfac = 3
>>> E-z= 1 5 0
>>> The electric field from ionic liquid has a magnitude of 7.5 V/nm in the
>>> opposite direction to the applied electric field. This makes no sense to
>>> me
>>> at all.
>>>
>>> With
>>> wall-ewald-zfac = 5
>>> E-z= 1 5 0
>>> The ionic liquid E is -6.2
>>>
>> How do you measure the electric field? Using gmx potential?
>> What numbers does the (optional?) file field.xvg contain?
>>
>> There can of course be a difference between applied field and effective
>> field, in particular with PBC. I sugges you read the supporting information
>> for our paper: Carl Caleman and David van der Spoel: Picosecond Melting of
>> Ice by an Infrared Laser Pulse - A Simulation Study Angew. Chem. Intl. Ed.
>> 47 pp. 1417-1420 (2008)
>>
>>>
>>> With
>>> wall-ewald-zfac = 10
>>> E-z= 1 5 0
>>> The electric field is still slightly above the expected 5 V/nm.
>>>
>>> The dimensions of the box is about 5x5x20 nm3.
>>>
>>> Reading various sources say that the wall factor should not matter as
>>> long
>>> as it is sufficiently large (x3) but in my case it seems to matter a lot.
>>> Should I report this as a bug or do you think I can investigate this
>>> further?
>>>
>>> Best Regards,
>>>
>>> Dan
>>>
>>> >> rce=link_campaign=sig-email_content=webmail>
>>> Virus-free.
>>> www.avg.com
>>> >> rce=link_campaign=sig-email_content=webmail>
>>> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
>>>
>>>
>>> On Wed, Nov 8, 2017 at 11:37 AM, Dan Gil  wrote:
>>>
>>> Hello,

 I am studying room-temperature ionic liquids in slab geometry and I am
 experiencing that I cannot get rid of the interactions between periodic
 images in the z-direction. Is there something wrong with what I am
 doing or
 is this inevitable?

 The reason I think this is because when I change "wall-ewald-zfac" the
 results change also.

 Gromacs 5.1
 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
 -
 ;Integration Method and Parameters
 integrator   = md
 nsteps   = 400
 dt= 0.002
 nstenergy= 100
 nstcalcenergy= 100
 nstlog   = 5000

 ;Output Control
 nstxout = 100
 nstvout = 1000

 ;Cutoff Schemes
 cutoff-scheme= verlet
 rlist= 1.0
 vdw-type = cut-off
 rvdw = 1.0
 DispCorr = EnerPres

 ;Coulomb interactions
 coulombtype  = pme
 rcoulomb = 1.0
 fourierspacing   = 0.4

 ;Constraints
 constraints  

Re: [gmx-users] Ewald summation in 2 dimensions with correction

[Dan Gil]

Hi,

I am using gmx potential to measure the electric field. I've reported the
values from field.xvg in my previous email. The field at the edges of the
ionic liquid near the walls will oscillate because of the electric double
layer but it is the bulk field I am concerned with.

I read the supporting information - I thought that the Ewald summation
correction for 2 dimensions was supposed to cancel out interactions between
periodic images in the z direction?

Were you using the regular 3-D Ewald summation in this work? I might try
carrying out the calculations that way.

Dan

On Mon, Nov 13, 2017 at 2:23 AM, David van der Spoel 
wrote:

> On 11/11/17 16:44, Dan Gil wrote:
>
>> Hi,
>>
>> Just wanted to see if anyone had any thoughts...
>>
>> For more information, if I use:
>> wall-ewald-zfac = 3
>> E-z= 1 5 0
>> The electric field from ionic liquid has a magnitude of 7.5 V/nm in the
>> opposite direction to the applied electric field. This makes no sense to
>> me
>> at all.
>>
>> With
>> wall-ewald-zfac = 5
>> E-z= 1 5 0
>> The ionic liquid E is -6.2
>>
> How do you measure the electric field? Using gmx potential?
> What numbers does the (optional?) file field.xvg contain?
>
> There can of course be a difference between applied field and effective
> field, in particular with PBC. I sugges you read the supporting information
> for our paper: Carl Caleman and David van der Spoel: Picosecond Melting of
> Ice by an Infrared Laser Pulse - A Simulation Study Angew. Chem. Intl. Ed.
> 47 pp. 1417-1420 (2008)
>
>>
>> With
>> wall-ewald-zfac = 10
>> E-z= 1 5 0
>> The electric field is still slightly above the expected 5 V/nm.
>>
>> The dimensions of the box is about 5x5x20 nm3.
>>
>> Reading various sources say that the wall factor should not matter as long
>> as it is sufficiently large (x3) but in my case it seems to matter a lot.
>> Should I report this as a bug or do you think I can investigate this
>> further?
>>
>> Best Regards,
>>
>> Dan
>>
>> > source=link_campaign=sig-email_content=webmail>
>> Virus-free.
>> www.avg.com
>> > source=link_campaign=sig-email_content=webmail>
>> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
>>
>>
>> On Wed, Nov 8, 2017 at 11:37 AM, Dan Gil  wrote:
>>
>> Hello,
>>>
>>> I am studying room-temperature ionic liquids in slab geometry and I am
>>> experiencing that I cannot get rid of the interactions between periodic
>>> images in the z-direction. Is there something wrong with what I am doing
>>> or
>>> is this inevitable?
>>>
>>> The reason I think this is because when I change "wall-ewald-zfac" the
>>> results change also.
>>>
>>> Gromacs 5.1
>>> - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
>>> ;Integration Method and Parameters
>>> integrator   = md
>>> nsteps   = 400
>>> dt= 0.002
>>> nstenergy= 100
>>> nstcalcenergy= 100
>>> nstlog   = 5000
>>>
>>> ;Output Control
>>> nstxout = 100
>>> nstvout = 1000
>>>
>>> ;Cutoff Schemes
>>> cutoff-scheme= verlet
>>> rlist= 1.0
>>> vdw-type = cut-off
>>> rvdw = 1.0
>>> DispCorr = EnerPres
>>>
>>> ;Coulomb interactions
>>> coulombtype  = pme
>>> rcoulomb = 1.0
>>> fourierspacing   = 0.4
>>>
>>> ;Constraints
>>> constraints  = all-bonds
>>>
>>> ;Temperature coupling
>>> gen-vel  = yes
>>> tcoupl   = v-rescale
>>> tc-grps  = System
>>> tau-t= 0
>>> ref-t= 300
>>>
>>> ;Walls
>>> pbc  = xy
>>> nwall= 2
>>> wall-type= 10-4
>>> ewald-geometry   = 3dc
>>> wall-atomtype= cp cm
>>> wall-density = 5 5
>>> wall-ewald-zfac  = 10
>>>
>>> ;Electric Field
>>> ;Cosine with freq=0, Field V/nm, Phase=n/a
>>> E-z  = 1 0 0
>>> - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
>>>
>>> Best Regards,
>>>
>>> Dan
>>>
>>>
>
> --
> David van der Spoel, Ph.D., Professor of Biology
> Head of Department, Cell & Molecular Biology, Uppsala University.
> Box 596, SE-75124 Uppsala, Sweden. Phone: +46184714205.
> http://www.icm.uu.se
> --
> Gromacs Users mailing list
>
> * Please search the archive at http://www.gromacs.org/Support
> /Mailing_Lists/GMX-Users_List before posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-requ...@gromacs.org.
>
-- 
Gromacs Users mailing list

* Please search the archive at 

Re: [gmx-users] Ewald summation in 2 dimensions with correction

[David van der Spoel]

On 11/11/17 16:44, Dan Gil wrote:

Hi,

Just wanted to see if anyone had any thoughts...

For more information, if I use:
wall-ewald-zfac = 3
E-z= 1 5 0
The electric field from ionic liquid has a magnitude of 7.5 V/nm in the
opposite direction to the applied electric field. This makes no sense to me
at all.

With
wall-ewald-zfac = 5
E-z= 1 5 0
The ionic liquid E is -6.2

How do you measure the electric field? Using gmx potential?
What numbers does the (optional?) file field.xvg contain?

There can of course be a difference between applied field and effective 
field, in particular with PBC. I sugges you read the supporting 
information for our paper: Carl Caleman and David van der Spoel: 
Picosecond Melting of Ice by an Infrared Laser Pulse - A Simulation 
Study Angew. Chem. Intl. Ed. 47 pp. 1417-1420 (2008)


With
wall-ewald-zfac = 10
E-z= 1 5 0
The electric field is still slightly above the expected 5 V/nm.

The dimensions of the box is about 5x5x20 nm3.

Reading various sources say that the wall factor should not matter as long
as it is sufficiently large (x3) but in my case it seems to matter a lot.
Should I report this as a bug or do you think I can investigate this
further?

Best Regards,

Dan


Virus-free.
www.avg.com

<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

On Wed, Nov 8, 2017 at 11:37 AM, Dan Gil  wrote:


Hello,

I am studying room-temperature ionic liquids in slab geometry and I am
experiencing that I cannot get rid of the interactions between periodic
images in the z-direction. Is there something wrong with what I am doing or
is this inevitable?

The reason I think this is because when I change "wall-ewald-zfac" the
results change also.

Gromacs 5.1
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
;Integration Method and Parameters
integrator   = md
nsteps   = 400
dt= 0.002
nstenergy= 100
nstcalcenergy= 100
nstlog   = 5000

;Output Control
nstxout = 100
nstvout = 1000

;Cutoff Schemes
cutoff-scheme= verlet
rlist= 1.0
vdw-type = cut-off
rvdw = 1.0
DispCorr = EnerPres

;Coulomb interactions
coulombtype  = pme
rcoulomb = 1.0
fourierspacing   = 0.4

;Constraints
constraints  = all-bonds

;Temperature coupling
gen-vel  = yes
tcoupl   = v-rescale
tc-grps  = System
tau-t= 0
ref-t= 300

;Walls
pbc  = xy
nwall= 2
wall-type= 10-4
ewald-geometry   = 3dc
wall-atomtype= cp cm
wall-density = 5 5
wall-ewald-zfac  = 10

;Electric Field
;Cosine with freq=0, Field V/nm, Phase=n/a
E-z  = 1 0 0
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Best Regards,

Dan




--
David van der Spoel, Ph.D., Professor of Biology
Head of Department, Cell & Molecular Biology, Uppsala University.
Box 596, SE-75124 Uppsala, Sweden. Phone: +46184714205.
http://www.icm.uu.se
--
Gromacs Users mailing list

* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a 
mail to gmx-users-requ...@gromacs.org.

Re: [gmx-users] Ewald summation in 2 dimensions with correction

[Dan Gil]

Hi,

Just wanted to see if anyone had any thoughts...

For more information, if I use:
wall-ewald-zfac = 3
E-z= 1 5 0
The electric field from ionic liquid has a magnitude of 7.5 V/nm in the
opposite direction to the applied electric field. This makes no sense to me
at all.

With
wall-ewald-zfac = 5
E-z= 1 5 0
The ionic liquid E is -6.2

With
wall-ewald-zfac = 10
E-z= 1 5 0
The electric field is still slightly above the expected 5 V/nm.

The dimensions of the box is about 5x5x20 nm3.

Reading various sources say that the wall factor should not matter as long
as it is sufficiently large (x3) but in my case it seems to matter a lot.
Should I report this as a bug or do you think I can investigate this
further?

Best Regards,

Dan


Virus-free.
www.avg.com

<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

On Wed, Nov 8, 2017 at 11:37 AM, Dan Gil  wrote:

> Hello,
>
> I am studying room-temperature ionic liquids in slab geometry and I am
> experiencing that I cannot get rid of the interactions between periodic
> images in the z-direction. Is there something wrong with what I am doing or
> is this inevitable?
>
> The reason I think this is because when I change "wall-ewald-zfac" the
> results change also.
>
> Gromacs 5.1
> - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
> ;Integration Method and Parameters
> integrator   = md
> nsteps   = 400
> dt= 0.002
> nstenergy= 100
> nstcalcenergy= 100
> nstlog   = 5000
>
> ;Output Control
> nstxout = 100
> nstvout = 1000
>
> ;Cutoff Schemes
> cutoff-scheme= verlet
> rlist= 1.0
> vdw-type = cut-off
> rvdw = 1.0
> DispCorr = EnerPres
>
> ;Coulomb interactions
> coulombtype  = pme
> rcoulomb = 1.0
> fourierspacing   = 0.4
>
> ;Constraints
> constraints  = all-bonds
>
> ;Temperature coupling
> gen-vel  = yes
> tcoupl   = v-rescale
> tc-grps  = System
> tau-t= 0
> ref-t= 300
>
> ;Walls
> pbc  = xy
> nwall= 2
> wall-type= 10-4
> ewald-geometry   = 3dc
> wall-atomtype= cp cm
> wall-density = 5 5
> wall-ewald-zfac  = 10
>
> ;Electric Field
> ;Cosine with freq=0, Field V/nm, Phase=n/a
> E-z  = 1 0 0
> - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
>
> Best Regards,
>
> Dan
>
-- 
Gromacs Users mailing list

* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a 
mail to gmx-users-requ...@gromacs.org.