From ehs73 at cornell.edu Mon Sep 23 19:09:57 2013 From: ehs73 at cornell.edu (E. H. Smith) Date: Mon, 23 Sep 2013 13:09:57 -0400 Subject: [Wannier] Trouble plotting MLWF in real space Message-ID: Hello, When I plot the MLWF of my compound in real space using xcrysden, the isosurfaces do not correctly "wrap" from one side of the unit cell to another. That is, if I'm visualizing the MLWF of my Pb atom and the Pb atom is very close to the "top" of my unit cell, the MLWF is only plotted up to the "top" of the unit cell and is truncated at the edge of the unit cell (instead of continuing on the "bottom" of the next unit cell). This problem persists when I change the number of unit cells plotted and whether or not I extend the isosurface over all the unit cells plotted or keep it on just one. Here is a screen shot of this problem in action: http://i.imgur.com/bqjxgYA.jpg Has anyone encountered this problem? Is this an xcrysden problem or a wannier90 problem? Thanks, E H Smith -------------- next part -------------- An HTML attachment was scrubbed... URL: From jonathan.yates at materials.ox.ac.uk Mon Sep 23 19:18:19 2013 From: jonathan.yates at materials.ox.ac.uk (Jonathan Yates) Date: Mon, 23 Sep 2013 17:18:19 +0000 Subject: [Wannier] Trouble plotting MLWF in real space In-Reply-To: References: Message-ID: <78877D222077A642A4CFAB09BD0077E001208C1B@MBX10.ad.oak.ox.ac.uk> On 23 Sep 2013, at 18:09, E. H. Smith wrote: > Hello, > > When I plot the MLWF of my compound in real space using xcrysden, the isosurfaces do not correctly "wrap" from one side of the unit cell to another. That is, if I'm visualizing the MLWF of my Pb atom and the Pb atom is very close to the "top" of my unit cell, the MLWF is only plotted up to the "top" of the unit cell and is truncated at the edge of the unit cell (instead of continuing on the "bottom" of the next unit cell). This problem persists when I change the number of unit cells plotted and whether or not I extend the isosurface over all the unit cells plotted or keep it on just one. Here is a screen shot of this problem in action: > > http://i.imgur.com/bqjxgYA.jpg > > Has anyone encountered this problem? Is this an xcrysden problem or a wannier90 problem? Try increasing wannier_plot_supercell. The default is 2, I'm guessing 3 might be ok. Remember that a WF does not have the periodicity of the unit cell. Jonathan -- Department of Materials, University of Oxford, Parks Road, Oxford, OX1 3PH, UK tel: +44 (0)1865 612797 http://users.ox.ac.uk/~oums0549/ From ehs73 at cornell.edu Mon Sep 23 23:30:59 2013 From: ehs73 at cornell.edu (E. H. Smith) Date: Mon, 23 Sep 2013 17:30:59 -0400 Subject: [Wannier] Trouble plotting MLWF in real space In-Reply-To: <78877D222077A642A4CFAB09BD0077E001208C1B@MBX10.ad.oak.ox.ac.uk> References: <78877D222077A642A4CFAB09BD0077E001208C1B@MBX10.ad.oak.ox.ac.uk> Message-ID: Thank you for the suggestion. Since this is a slow calculation to run, before I run it I would like to understand what's going on a bit better. I was using wannier_plot_supercell=2 (the default value). The spread in real space of the MLWF in question is about a quarter of my primitive cell size. If it got truncated in a 2x2x2 supercell, is it likely that a 3x3x3 will be better? On Mon, Sep 23, 2013 at 1:18 PM, Jonathan Yates < jonathan.yates at materials.ox.ac.uk> wrote: > > On 23 Sep 2013, at 18:09, E. H. Smith wrote: > > > Hello, > > > > When I plot the MLWF of my compound in real space using xcrysden, the > isosurfaces do not correctly "wrap" from one side of the unit cell to > another. That is, if I'm visualizing the MLWF of my Pb atom and the Pb atom > is very close to the "top" of my unit cell, the MLWF is only plotted up to > the "top" of the unit cell and is truncated at the edge of the unit cell > (instead of continuing on the "bottom" of the next unit cell). This problem > persists when I change the number of unit cells plotted and whether or not > I extend the isosurface over all the unit cells plotted or keep it on just > one. Here is a screen shot of this problem in action: > > > > http://i.imgur.com/bqjxgYA.jpg > > > > Has anyone encountered this problem? Is this an xcrysden problem or a > wannier90 problem? > > Try increasing wannier_plot_supercell. The default is 2, I'm guessing 3 > might be ok. Remember that a WF does not have the periodicity of the unit > cell. > > Jonathan > > > > > > -- > Department of Materials, University of Oxford, Parks Road, Oxford, OX1 > 3PH, UK > tel: +44 (0)1865 612797 http://users.ox.ac.uk/~oums0549/ > > _______________________________________________ > Wannier mailing list > Wannier at quantum-espresso.org > http://www.democritos.it/mailman/listinfo/wannier > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jonathan.yates at materials.ox.ac.uk Tue Sep 24 00:23:42 2013 From: jonathan.yates at materials.ox.ac.uk (Jonathan Yates) Date: Mon, 23 Sep 2013 22:23:42 +0000 Subject: [Wannier] Trouble plotting MLWF in real space In-Reply-To: References: <78877D222077A642A4CFAB09BD0077E001208C1B@MBX10.ad.oak.ox.ac.uk> Message-ID: <78877D222077A642A4CFAB09BD0077E00120BD6A@MBX10.ad.oak.ox.ac.uk> On 23 Sep 2013, at 22:30, "E. H. Smith" wrote: > Thank you for the suggestion. Since this is a slow calculation to run, before I run it I would like to understand what's going on a bit better. I was using wannier_plot_supercell=2 (the default value). The spread in real space of the MLWF in question is about a quarter of my primitive cell size. If it got truncated in a 2x2x2 supercell, is it likely that a 3x3x3 will be better? The spread isn't particularly relevant here - assuming the WF are fairly localised. The issue is the location of the supercell relative to the position of your wannier function. By default W90 is plotting the WF in the supercell that contains the "home" unit cell, plus seven additional cells. But your WF is on the edge of this supercell. By plotting in a 3x3x3 box you should be plotting in a box that includes all of your WF (that's a guess, but if you look at the files you should be able to confirm it). Plotting in a 3x3x3 cell is expensive - and it's clearly not the smart way to do this. Ideally one would plot in a small box centred on the WF you are interested in. It wouldn't be hard to modify the code to cope with that (it is on the to-do list) Jonathan -- Department of Materials, University of Oxford, Parks Road, Oxford, OX1 3PH, UK tel: +44 (0)1865 612797 http://users.ox.ac.uk/~oums0549/ From ehs73 at cornell.edu Tue Sep 24 16:44:49 2013 From: ehs73 at cornell.edu (E. H. Smith) Date: Tue, 24 Sep 2013 10:44:49 -0400 Subject: [Wannier] Trouble plotting MLWF in real space In-Reply-To: <78877D222077A642A4CFAB09BD0077E00120BD6A@MBX10.ad.oak.ox.ac.uk> References: <78877D222077A642A4CFAB09BD0077E001208C1B@MBX10.ad.oak.ox.ac.uk> <78877D222077A642A4CFAB09BD0077E00120BD6A@MBX10.ad.oak.ox.ac.uk> Message-ID: All right. Thanks for the information. On Mon, Sep 23, 2013 at 6:23 PM, Jonathan Yates < jonathan.yates at materials.ox.ac.uk> wrote: > > On 23 Sep 2013, at 22:30, "E. H. Smith" > wrote: > > > Thank you for the suggestion. Since this is a slow calculation to run, > before I run it I would like to understand what's going on a bit better. I > was using wannier_plot_supercell=2 (the default value). The spread in real > space of the MLWF in question is about a quarter of my primitive cell size. > If it got truncated in a 2x2x2 supercell, is it likely that a 3x3x3 will be > better? > > The spread isn't particularly relevant here - assuming the WF are fairly > localised. The issue is the location of the supercell relative to the > position of your wannier function. By default W90 is plotting the WF in the > supercell that contains the "home" unit cell, plus seven additional cells. > But your WF is on the edge of this supercell. By plotting in a 3x3x3 box > you should be plotting in a box that includes all of your WF (that's a > guess, but if you look at the files you should be able to confirm it). > > Plotting in a 3x3x3 cell is expensive - and it's clearly not the smart > way to do this. Ideally one would plot in a small box centred on the WF you > are interested in. It wouldn't be hard to modify the code to cope with that > (it is on the to-do list) > > > Jonathan > > > > > -- > Department of Materials, University of Oxford, Parks Road, Oxford, OX1 > 3PH, UK > tel: +44 (0)1865 612797 http://users.ox.ac.uk/~oums0549/ > > _______________________________________________ > Wannier mailing list > Wannier at quantum-espresso.org > http://www.democritos.it/mailman/listinfo/wannier > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: