search

qmcurrents / gimic

Gauge-including magnetically induced currents.
https://gimic.readthedocs.io
GNU General Public License v2.0
28 stars 19 forks source link

How to do Integration analysis? #238

Closed SKumar1208 closed 5 years ago

SKumar1208 commented 5 years ago

Hii... I have run integral calculation using integral gimic.inp (benzene) MOL and XDENS file. I got three output files i.e. gimic.out, grid.xyz and mol.xyz. Plz, tell me whether I a right or not? Plz, also tell me how can I plot integration analysis using these files? Is I have to do many run by changing the distace from the center of the molecule? If like that in which part of gimic.inp or other files I have to make changes? Plz, tell me stepwise instructions, so I can plot this current flow.

SKumar1208 commented 5 years ago

Hiii... Plz, reply to me. I need to calculate the current profile analysis. Plz, give me stepwise instructions.

heikef commented 5 years ago

So you have been able to run the integration module. Very good. It is correct that the output files are gimic.out, grid.xyz and mol.xyz. In gimic.out you will find the result of the integration across the respective bond you have defined. For benzen assuming you have started in the centre of the ring and integrated to the outside the result should be around 11.8 nA/T (total net current strength). Please compare. When you look at mol.xyz these are the molecular coordinates of your molecule in Angstrøm. In grid.xyz you will find the molecular coordinates + dummy atoms X which are visualizing your integration plane. You need to check always the grid.xyz file to make sure that the placement of your integration plane is such as you have intended while doing the definition in gimic.inp file. You can visualize both mol.xyz and grid.xyz using for example molden or any program that is able to read in coordinates in Angstrøm.

heikef commented 5 years ago

For a simple molecule it is sufficient to start in the centre of the ring and integrate across a bond. This holds typically for single ring molecules. Better is when you do a integration scan. This implies you stepwise exend you plane. Then you will obtain a curve and it will become clear where the integration should start and where it ends. Please have a look at Wiley Interdisciplinary Reviews (WIREs), 6, 639, (2016) there the slicing procedure is explained with some viewgraphs.

heikef commented 5 years ago

It is hard to help you when I do not know what sort of molecule you are interested in. When doing the integration analysis you look at selected bonds of your molecule. At the end you will have a number the integrated current strength susceptibility as a result. How you later on present this in your paper is up to you. It can be a table where you define the bond with atom numbers and present the result or you add it in a viewgraph next to your bond. I often do the latter option because it is easier for a reader to quickly catch the results, but this is a matter of taste.

heikef commented 5 years ago

I think a very convenient way to run several integrations across bonds for you is to make the bash scripts work on your computer. https://gimic.readthedocs.io/en/latest/scripts.html

heikef commented 5 years ago

In case you do not get the interactive bash scripts quickly to work:

  1. place the integration plane across the bond of interest in gimic.inp
  2. check if this is reasonable by running "gimic --dryrun" this will generate only the grid.xyz file without doing the integration
  3. visualize grid.xyz and check if the placement of the plane is like you intended it to be
  4. if this is the case change "calc=integral" in your gimic.inp file to calc=cdens and use a more loose grid because this will speed up your calculation - you are only after a visualization not a number - if you are after a number you have to increase the grid again
  5. run gimic > gimic.out
  6. you will see that a file called jmod.txt is generated
  7. visualize this file using the script here: https://github.com/qmcurrents/gimic/blob/master/tools/Visualization/Gnuplot/gnu_contour.cmd
  8. you will see a contour plot of the modulus of the current density - when your integration plane is too large or too small this should be obvious to see - remember you have always a inner part like a circle on the left side of your bond and an outer part like a circle on the right hand side of your bond. Imagine a lying number eight. Everything above or below even left and right which is also in this picture is something you do not want to include in your integration plane - look for example of the contour plots here: J. Phys. Chem. A, 113, 8668, (2009)
  9. adjust your plane and check again - repeat this procedure until you are happy with the placement
  10. increase the grid eg. spacing 0.02 and change "calc=cdens" back to "calc=integral"
  11. run the integration calculation

Note it is more convenient to do a quick run with the interactive bash scripts. This procedure is recommended. But you will also learn something doing it the other way.

SKumar1208 commented 5 years ago

Thank you so much for your reply. Yes, I am getting this result for benzene (Induced current (au) : 0.420288 Positive contribution: 0.596631 ( 16.812711 ) Negative contribution: -0.176343 ( -4.969249 )

Induced current (nA/T)  :    11.843462
   (conversion factor)  :    28.179409)

I did this calculation for my molecules and got the results for one bond ( positive value (~5 for aromatic and negative value (~(-9) for antiaromatic). Plz tell me can I write these value. Plz, also tell me how can I get a graph out of this.

On Fri, Apr 19, 2019 at 2:30 PM Heike Fliegl notifications@github.com wrote:

It is hard to help you when I do not know what sort of molecule you are interested in. When doing the integration analysis you look at selected bonds of your molecule. At the end you will have a number the integrated current strength susceptibility as a result. How you later on present this in your paper is up to you. It can be a table where you define the bond with atom numbers and present the result or you add it in a viewgraph next to your bond. I often do the latter option because it is easier for a reader to quickly catch the results, but this is a matter of taste.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-484817755, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFBOH32P5TTIXTHMSH3PRGC3DANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

SKumar1208 commented 5 years ago

Ok, thank you.. I will do and let you know.

On Fri, Apr 19, 2019 at 2:56 PM Heike Fliegl notifications@github.com wrote:

In case you do not get the interactive bash scripts quickly to work:

  1. place the integration plane across the bond of interest in gimic.inp
  2. check if this is reasonable by running "gimic --dryrun" this will generate only the grid.xyz file without doing the integration
  3. visualize grid.xyz and check if the placement of the plane is like you intended it to be
  4. if this is the case change "calc=integral" in your gimic.inp file to calc=cdens and use a more loose grid because this will speed up your calculation - you are only after a visualization not a number - if you are after a number you have to increase the grid again
  5. run gimic > gimic.out
  6. you will see that a file called jmod.txt is generated

  7. visualize this file using the script here:

    https://github.com/qmcurrents/gimic/blob/master/tools/Visualization/Gnuplot/gnu_contour.cmd

  8. you will see a contour plot of the modulus of the current density - when your integration plane is too large or too small this should be obvious to see - remember you have always a inner part like a circle on the left side of your bond and an outer part like a circle on the right hand side of your bond. Imagine a lying number eight. Everything above or below even left and right which is also in this picture is something you do not want to include in your integration plane - look for example of the contour plots here: J. Phys. Chem. A, 113, 8668, (2009)
  9. adjust your plane and check again - repeat this procedure until you are happy with the placement
  10. increase the grid eg. spacing 0.02 and change "calc=cdens" back to "calc=integral"
  11. run the integration calculation

Note it is more convenient to do a quick run with the interactive bash scripts. This procedure is recommended. But you will also learn something doing it the other way.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-484825649, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFG6CEZR6Y4GWNPDFRTPRGF47ANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

SKumar1208 commented 5 years ago

Hiii... I have this type of molecules plz, see the attached .pdb file). I am also attaching the MOL and XDENS file of this molecule XDENS https://drive.google.com/file/d/1kPjxN0LUTAS966MnOCm4YEpvtIbr56DL/view?usp=drive_web . I don't have interactive base script. Plz, can you tell me the gimic.inp for this molecule.

On Fri, Apr 19, 2019 at 2:56 PM Heike Fliegl notifications@github.com wrote:

In case you do not get the interactive bash scripts quickly to work:

  1. place the integration plane across the bond of interest in gimic.inp
  2. check if this is reasonable by running "gimic --dryrun" this will generate only the grid.xyz file without doing the integration
  3. visualize grid.xyz and check if the placement of the plane is like you intended it to be
  4. if this is the case change "calc=integral" in your gimic.inp file to calc=cdens and use a more loose grid because this will speed up your calculation - you are only after a visualization not a number - if you are after a number you have to increase the grid again
  5. run gimic > gimic.out
  6. you will see that a file called jmod.txt is generated

  7. visualize this file using the script here:

    https://github.com/qmcurrents/gimic/blob/master/tools/Visualization/Gnuplot/gnu_contour.cmd

  8. you will see a contour plot of the modulus of the current density - when your integration plane is too large or too small this should be obvious to see - remember you have always a inner part like a circle on the left side of your bond and an outer part like a circle on the right hand side of your bond. Imagine a lying number eight. Everything above or below even left and right which is also in this picture is something you do not want to include in your integration plane - look for example of the contour plots here: J. Phys. Chem. A, 113, 8668, (2009)
  9. adjust your plane and check again - repeat this procedure until you are happy with the placement
  10. increase the grid eg. spacing 0.02 and change "calc=cdens" back to "calc=integral"
  11. run the integration calculation

Note it is more convenient to do a quick run with the interactive bash scripts. This procedure is recommended. But you will also learn something doing it the other way.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-484825649, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFG6CEZR6Y4GWNPDFRTPRGF47ANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

heikef commented 5 years ago

I am sorry but only with a XDENS file I can not see much.

heikef commented 5 years ago

The gimic.inp for the benzene example should be fine for you but you have to adapt it according to the molecule you are interested in.

SKumar1208 commented 5 years ago

Which files you need? plz tell me. I will send all those

On Fri, Apr 19, 2019 at 4:32 PM Heike Fliegl notifications@github.com wrote:

I am sorry but only with a XDENS file I can not see much.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-484852770, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFAIE4B5YKJKDAGQOM3PRGRDDANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

heikef commented 5 years ago

NEW INPUT

calc=integral # cdens, integral, edens, divj title="" basis="../MOL" xdens="../XDENS" debug=1 # debug print level openshell=false magnet_axis=X #[-] i,j,k || x,y,z -> align magnet along axis

grid type can be: base/std, bond or file (see examples below)

Grid(bond) { # define grid orthogonal to a bond type=gauss # gauss distribution of grid points bond=[1,2] <-- here you give the atom numbers of the bond you are interested in fixpoint=4 <-- the fixpoint should lie to the left of your bond

coord1=[0.0, 0.0, 2.145166] # "atom" coordinates

coord2=[0.0, 0.0, -2.145166]

fixcoord=[0.0, 0.0, 0.0]

distance=1.32               # place grid 'distance' between atoms <--- middle of your bond 
gauss_order=9               # order for gauss quadrature
spacing=[0.02, 0.02, 0.02]     # spacing of points on grid (i,j,k)
height=[-5.0, 5.0]      <-- you can extend it to eg. +-8 
width=[-2.2, 5.0].      <--- assuming your molecule is planar this is the important part where you set the dimensions of you plane

}

Advanced { lip_order=5 # order of Lagrange Interpolating Polynominals to use spherical=off # don't touch, unless you REALLY know what you are doing diamag=on # turn on/off diamagnetic contributions paramag=on # turn on/off paramagnetic contributions GIAO=on # turn on/off GIAOs. Don't change unless you know why. screening=on # use screening to speed up screening_thrs=1.d-8 # Screening threshold }

Essential { # this section is new and only works for master acid=off <-- switch this off since you do not seem to need it }

heikef commented 5 years ago

note my comments <-- should be removed when you run the calculation otherwise the program will crash

heikef commented 5 years ago

I need to see the gimic.inp, MOL, XDENS files to be able to comment on your question.

SKumar1208 commented 5 years ago

I am not a theoretical specialist. plz, send me how I can make a make a integration plane in gimic.inp any example. Like if I have to make a plane across the naphthalene from one side to other including the central bond (means bond between C2-C3, C6-C7 and central C9-C10) what should I have to do?

On Fri, Apr 19, 2019 at 4:33 PM Heike Fliegl notifications@github.com wrote:

The gimic.inp for the benzene example should be fine for you but you have to adapt it according to the molecule you are interested in.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-484853036, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFA4OXFZ5UPVPOSF7ODPRGRG3ANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

heikef commented 5 years ago

You do not need to share the molecule you are working on since github is public and everybody can see it. Instead you either share an example or try to investigate another bond of the benzene molecule adapting the gimic.inp file yourself.

After running several gimic calculations you will see that it is not difficult. Just remain patient.

SKumar1208 commented 5 years ago

OK. Thanks.

On Fri, Apr 19, 2019 at 4:41 PM Heike Fliegl notifications@github.com wrote:

You do not need to share the molecule you are working on since github is public and everybody can see it. Instead you either share an example or try to investigate another bond of the benzene molecule adapting the gimic.inp file yourself.

After running several gimic calculations you will see that it is not difficult. Just remain patient.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-484855039, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFAOTRGIDD6QNZKKNUDPRGSF5ANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

heikef commented 5 years ago
  1. you need three points to define a plane. This is done defining the atoms of your bond of interest and a fixpoint. The program sets then the plane automatically perpendicular to your bond.
  2. with the definition of distance which is typically 1/2 of your bond length you tell the program to place the plane in the middle of your bond.
  3. now we have defined a big plane, but the molecule is small - therefore we need to adapt the plane according to the geometry of the molecule we are looking at
  4. a plane has a height and a width - when the molecule is planar we keep the height fixed and play with the width - the centre of the plane is assumed in the middle of the bond width=[-3.0, 4.0] means go 3 bohr inside the ring and 4 bohr outside the ring for example, height=[-6.0, +6.0] would mean go -6.0 bohr below the midpoint of the bond and 6.0 bohr above. You see, nothing fancy. You do not need to deal a lot with theory to be able to define a plane. Do not forget to check the result by looking at grid.xyz

Look at the viewgraphs here: https://gimic.readthedocs.io/en/latest/interpretation.html

SKumar1208 commented 5 years ago

Plz, tell me. If I run the integration calculation for each bond (which I want) and write the values (obtained from the gimic.out) in table or over the structure of the molecule as you said earlier will be ok or not?

On Fri, Apr 19, 2019 at 4:41 PM Heike Fliegl notifications@github.com wrote:

You do not need to share the molecule you are working on since github is public and everybody can see it. Instead you either share an example or try to investigate another bond of the benzene molecule adapting the gimic.inp file yourself.

After running several gimic calculations you will see that it is not difficult. Just remain patient.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-484855039, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFAOTRGIDD6QNZKKNUDPRGSF5ANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

heikef commented 5 years ago

I think this should be ok. If you have another idea on how to illustrate the results you can also do this as long as it is clear at which bond you have been looking. You need to report the total current strength, when the molecule is very complicated it can happen that you also need to report dia- and paratropic current strength. I do not know what you are investigating so I can only comment regarding previous work done in our group.

SKumar1208 commented 5 years ago

Ok, thank you so much.

On Fri, Apr 19, 2019 at 4:52 PM Heike Fliegl notifications@github.com wrote:

I think this should be ok. If you have another idea on how to illustrate the results you can also do this as long as it is clear at which bond you have been looking. You need to report the total current strength, when the molecule is very complicated it can happen that you also need to report dia- and paratropic current strength. I do not know what you are investigating so I can only comment regarding previous work done in our group.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-484861562, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFHLZEFHAU4Q3GRI2OLPRGTQBANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

heikef commented 5 years ago

If you want to set up a plot for a current strength function you need to get the bash scripts working. Normally this is useful when looking at complicated molecules. For the future and for your general understanding you should learn on how to master this as well.

SKumar1208 commented 5 years ago

Ok, thank you.

On Fri, Apr 19, 2019 at 4:56 PM Heike Fliegl notifications@github.com wrote:

If you want to set up a plot for a current strength function you need to get the bash scripts working. Normally this is useful when looking at complicated molecules. For the future and for your general understanding you should learn on how to master this as well.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-484863903, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFETSV3LDCRGG3LSQEDPRGT5DANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

heikef commented 5 years ago

Following the github conversation you should be able to create a plot placing the inspection plane 1 bohr above your molecule plus looking at selected bonds and running a integration calculation. Often this is sufficient when you are under time pressure as you have indicated.

SKumar1208 commented 5 years ago

Hiii...I am able to run the integration calculation with some bonds like [1, 2], [2,3]. However, I am not able to run calculation with [3,4] bond in case of benzene as well as in my molecules. I am getting this error (<< ERROR: Basis vectors are linearly dependent, field direction undefined!

). iN CASE OF SOME OTHER BONDS ALSO (IN MY MOLECULE) i AM GETTING THE SAME ERROR. Plz, tell me what should I have to change in the gimic.inp to get the calculation done. Plz, also tell me is the bond number is same as molecule numbering shows? Actually I want to calculate the integration value for each aromatic/antiaromatic bonds in my molecule.

On Fri, Apr 19, 2019 at 4:58 PM Heike Fliegl notifications@github.com wrote:

Following the github conversation you should be able to create a plot placing the inspection plane 1 bohr above your molecule plus looking at selected bonds and running a integration calculation. Often this is sufficient when you are under time pressure as you have indicated.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-484864420, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFGOADH7RZYSILEB2BLPRGUE5ANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

SKumar1208 commented 5 years ago

Hiii...I am able to run the integration calculation with some bonds like [1, 2], [2,3]. However, I am not able to run calculation with [3,4] bond in case of benzene as well as in my molecules. I am getting this error (<< ERROR: Basis vectors are linearly dependent, field direction undefined!

). iN CASE OF SOME OTHER BONDS ALSO (IN MY MOLECULE) i AM GETTING THE SAME ERROR. Plz, tell me what should I have to change in the gimic.inp to get the calculation done. Plz, also tell me is the bond number is same as molecule numbering shows? Actually, I want to calculate the integration value for each bonds associated with aromatic/ antiaromatic ring in my molecule.

heikef commented 5 years ago

This error typically occurs when for example your fixpoint is identical to one point defining your bond. Please check.

If you look at a single ring system investigating one bond is sufficient. You will always obtain the same result. If not you have a reason to be concerned.

SKumar1208 commented 5 years ago

Actually, my system is a two ring system (naphthalene based) and I want to calculate the integration analysis in both rings and at the bridged C-C bond. Plz, tell me what should I have to change in gimic. Just considering my molecule as naphthalene, suggest me how can I get these integration values ( Phys. Chem. Chem. Phys., 2016, 18, 15934). Plz, see this paper they have done.

On Wed, Apr 24, 2019 at 3:13 PM Heike Fliegl notifications@github.com wrote:

This error typically occurs when for example your fixpoint is identical to one point defining your bond. Please check.

If you look at a single ring system investigating one bond is sufficient. You will always obtain the same result. If not you have a reason to be concerned.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-486148451, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFGZXGKHGP46EDK3Z5DPSATS5ANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

heikef commented 5 years ago

I do not know which input you were using for your molecule. What I can say is that the coordinates for atom 1 atom 2 and the fixpoint have to be different. You should always check the grid file if the integration plane was set properly. Assuming your molecule is planar you can also use the "X" option for the external magnetic field. My impression is when you want to reproduce the values of the paper you refer to you should make the bash scripts work on your computer.

If you want to run more complicated analysis you have to make sure that you understand what you are doing. Train first on calculating net currents and later on shift to local and global ones.

SKumar1208 commented 5 years ago

Plz, see the attached figures and files. I want to calculate the current strength at the central carbon bond C3-C4 (figure-naph.jpg). However I am not getting the value as reported in the figure (c6cp01968d-f4). I have found the value corresponding to the bond C1 and C2 as reported (see the 1 6-gimic.out). Plz, tell me in grid.xyz. and gimic.inp. what is worng. what should I have to do? how I can know by seeing the grid.xyz that my plane is right or worng?

On Wed, Apr 24, 2019 at 5:20 PM Heike Fliegl notifications@github.com wrote:

I do not know which input you were using for your molecule. What I can say is that the coordinates for atom 1 atom 2 and the fixpoint have to be different. You should always check the grid file if the integration plane was set properly. Assuming your molecule is planar you can also use the "X" option for the external magnetic field. My impression is when you want to reproduce the values of the paper you refer to you should make the bash scripts work on your computer.

If you want to run more complicated analysis you have to make sure that you understand what you are doing. Train first on calculating net currents and later on shift to local and global ones.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-486190317, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFFJTA5DV4IAMS4L5SLPSBCRVANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

heikef commented 5 years ago

I do not see an attachment. When you open the grid.xyz file with eg. molden or xmakemol you will see the integration plane. Always choose the fixpoint to the left of the bond you are interested in. The best would be to do a stepwise integration across the bond. When you install the bash scripts the fixpoint is chosen automatically. @mariavd can help you to get the scripts working

heikef commented 5 years ago

Were you able to reproduce the result for benzene? Did you investigate different bonds of benzene? There you should always get the same result. You could change the fixpoint location and see what sort of effect it has. This will help you to understand how the program works and how to define the input.

SKumar1208 commented 5 years ago

Yes, I have reproduce the benzene results. It was same. However, I could not reproduce for 3,4 bond in that also. I have taken the fix point right in that case. I will do it by placing it to left and will see. Thank you. I will let you know.

On 24 Apr 2019 11:56 p.m., "Heike Fliegl" notifications@github.com wrote:

Were you able to reproduce the result for benzene? Did you investigate different bonds of benzene? There you should always get the same result. You could change the fixpoint location and see what sort of effect it has. This will help you to understand how the program works and how to define the input.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-486370929, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFH7O7II2WEVVAZTZGLPSCQ43ANCNFSM4HGWVYUA .

heikef commented 5 years ago

What happened then for bond 3-4 ? I am still at the benzene example

SKumar1208 commented 5 years ago

There was error which I have sent you earlier. However, as you told me to change the fix point value, I tried this for naphthalene not for benzene. Now I will do as you told then let you know. My molecule is naphthalene based and I have to do integration analysis of both rings and for the central bond.

On 25 Apr 2019 12:07 a.m., "Heike Fliegl" notifications@github.com wrote:

What happened then for bond 3-4 ? I am still at the benzene example

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-486375077, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFHK4GEAUGV5JJM6253PSCSGRANCNFSM4HGWVYUA .

SKumar1208 commented 5 years ago

I have sent the figure which I have taken from the paper I mentioned. I don't know why you r not getting those files. They have reported the current strength value for naphthalene central bond by placing a plan ( may be parallel) the value for that is higher than the side bonds.

On 25 Apr 2019 12:16 a.m., "Sharvan Kumar" shravan.aug12@gmail.com wrote:

There was error which I have sent you earlier. However, as you told me to change the fix point value, I tried this for naphthalene not for benzene. Now I will do as you told then let you know. My molecule is naphthalene based and I have to do integration analysis of both rings and for the central bond.

On 25 Apr 2019 12:07 a.m., "Heike Fliegl" notifications@github.com wrote:

What happened then for bond 3-4 ? I am still at the benzene example

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-486375077, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFHK4GEAUGV5JJM6253PSCSGRANCNFSM4HGWVYUA .

heikef commented 5 years ago

When you do not understand why you got the error with benzene it does not make sense to continue with napthalene.

SKumar1208 commented 5 years ago

Yeah, you r right. I thought may be due to presence of other ring, I might get result and for this moment I can go through that. But I was wrong.

On 25 Apr 2019 12:25 a.m., "Heike Fliegl" notifications@github.com wrote:

When you do not understand why you got the error with benzene it does not make sense to continue with napthalene.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-486381570, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFGISL27W56IBPYJLHTPSCUKJANCNFSM4HGWVYUA .

heikef commented 5 years ago

Can you please share the gimic.inp file for benzene for the case it crashed? I will try to reproduce your problem.

SKumar1208 commented 5 years ago

I will send you the file tomorrow. Because at this moment I am not in my lab. Sorry. I will send you tomorrow.

On 25 Apr 2019 12:32 a.m., "Heike Fliegl" notifications@github.com wrote:

Can you please share the gimic.inp file for benzene for the case it crashed? I will try to reproduce your problem.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-486384225, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFCDQA53QB3KFHMZBG3PSCVF7ANCNFSM4HGWVYUA .

heikef commented 5 years ago

No problem.

SKumar1208 commented 5 years ago

Hiii... Thank you so much. Now I am able to run integration analysis of all the bonds of benzene and naphthalene. I have one more query. Plz tell me how can I place a plane 10 bohr above the molecular plane? (simply speaking, what point should I have to change in the gimic.inp to place a plane according to my need of distance from the molecular plane).

heikef commented 5 years ago

Are you sure that you really mean 10 bohr? Usually the plane is place 1 bohr above or below the molecule. There are two ways to do this:

  1. you modify the placement of the plane by hand in the gimic.inp file and use the cdens module
  2. you do a 3D calculation and visualize the result later with ParaView. This has the advantage that you can shift the plane as you like eg. keep it inside the molecular plan and stepwise increas the distance. It would also allow you to do 3D streamline plots.

I am glad to hear that you managed to run the integration analysis. Did you manage to get the bash scripts working?

heikef commented 5 years ago

https://github.com/qmcurrents/gimic/blob/master/examples/benzene/3D/gimic.inp Here is a 3D example, but you need to do the dryrun option first to make sure your molecule fits into the box. Use also a spacing of 0.3.

SKumar1208 commented 5 years ago

Actually, I mean, how can I change the plane distance if I want to calculate at more than 1 bohr? I think this gimic inp is fix for 1 bohr. I didn't understand in the gimic.inp which point is related to the plane distance. It would be helpful for me if I will know this point. So that I can tune the plane distance and understand the changes in the ring current. I need just the information about that point in the gimic.inp.

On Sun, Apr 28, 2019 at 1:53 PM Heike Fliegl notifications@github.com wrote:

Are you sure that you really mean 10 bohr? Usually the plane is place 1 bohr above or below the molecule. There are two ways to do this:

  1. you modify the placement of the plane by hand in the gimic.inp file and use the cdens module
  2. you do a 3D calculation and visualize the result later with ParaView. This has the advantage that you can shift the plane as you like eg. keep it inside the molecular plan and stepwise increas the distance. It would also allow you to do 3D streamline plots.

I am glad to hear that you managed to run the integration analysis. Did you manage to get the bash scripts working?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-487357096, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFG4HGRDXPVO5XULP4DPSVNGJANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

heikef commented 5 years ago

In Paraview you can set a plane after a 3D gimic calculation is finished. There you can easily shift the plane as you like eg. tilt it of do whatever you are interested in. When I understand you correctly you want to calculate only 1 2D plane which is exactly defined in the input. Assuming that you molecule is oriented in the xy plane the modification would look like this:

origin=[-8.0, -8.0, -1.0] --> you change the origin placing the plane 1 bohr below in this case the molecular plane lengths=[16.0, 16.0, 0.0] --> here you tell the program to span only one plane in xy space, therefore the last (z) entry is zero. By doing so you restrict yourself regarding the visualization options, but if this is ok for you this issue should be solved.

Always check first with gimic --dryrun if the plane has been placed as you intended it.

SKumar1208 commented 5 years ago

Ok, Thank you so much.

On Sun, Apr 28, 2019 at 6:15 PM Heike Fliegl notifications@github.com wrote:

In Paraview you can set a plane after a 3D gimic calculation is finished. There you can easily shift the plane as you like eg. tilt it of do whatever you are interested in. When I understand you correctly you want to calculate only 1 2D plane which is exactly defined in the input. Assuming that you molecule is oriented in the xy plane the modification would look like this:

origin=[-8.0, -8.0, -1.0] --> you change the origin placing the plane 1 bohr below in this case the molecular plane lengths=[16.0, 16.0, 0.0] --> here you tell the program to span only one plane in xy space, therefore the last (z) entry is zero. By doing so you restrict yourself regarding the visualization options, but if this is ok for you this issue should be solved.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/qmcurrents/gimic/issues/238#issuecomment-487375933, or mute the thread https://github.com/notifications/unsubscribe-auth/ALVXKFDZPXQ6PR54PM3N2R3PSWL73ANCNFSM4HGWVYUA .

-- Sincerely, ------------------------------------------------- Dr Sharvan Kumar

Research AssociateChemical Engineering and Chemistry Indian Institute of Technology, DelhiNew Delhi 110016

heikef commented 5 years ago

Ok, I will close this issue then.