Dear Prof. Matthias Troyer,
many thanks for your answer :)!
Do you want to measure the real-space correlation function or is the magnetic susceptibility as an integrated measure enough?
We would like to measure the real-space correlation function in order to separate between short- and long-range interactions and their influence on the internal energy. Is it possible to measure such a function using alps?
I have two more minor questions concerning the output: Is the susceptibility \chi given out by alps equivalent to \chi=\beta*(<m^2>-<m>^2)?
Is it possible to make a snapshot (spin position+spin alignment) of the spin system at a certain time-step?
Again many, many thanks for your support. Best regards, Fritz Koermann
--- Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf
Handelsregister B 2533 Amtsgericht Düsseldorf
Geschäftsführung Prof. Dr. Jörg Neugebauer Prof. Dr. Anke R. Pyzalla Prof. Dr. Dierk Raabe Prof. Dr. Martin Stratmann Dipl.-Kfm. Herbert Wilk
Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000 -------------------------------------------------
On Jul 21, 2008, at 5:05 PM, Fritz Koermann wrote:
Dear Prof. Matthias Troyer,
many thanks for your answer :)!
Do you want to measure the real-space correlation function or is the magnetic susceptibility as an integrated measure enough?
We would like to measure the real-space correlation function in order to separate between short- and long-range interactions and their influence on the internal energy. Is it possible to measure such a function using alps?
The internal energy is just the short-range correlation.
We can add correlation measurements but that is lots of data and will slow down the code quote a bit. Do you actually need all correlations?
I have two more minor questions concerning the output: Is the susceptibility \chi given out by alps equivalent to \chi=\beta*(<m^2>-<m>^2)?
On any finite lattice actually <m>^2 is 0, and one would need to measure <|m|>^2. We actually only do \chi=\beta <m^2>, so that the estimate is valid only above Tc.
Is it possible to make a snapshot (spin position+spin alignment) of the spin system at a certain time-step?
This is easily possible. In which format would you like to see the snapshot?
Best regards
Matthias Troyer
Hi,
The internal energy is just the short-range correlation.
Also below Tc?
We can add correlation measurements but that is lots of data and will slow down the code quote a bit. Do you actually need all correlations?
No, I thought about nearest-neighbor-correlation <SiSi+1> or something like that.
I have two more minor questions concerning the output: Is the susceptibility \chi given out by alps equivalent to \chi=\beta*(<m^2>-<m>^2)?
On any finite lattice actually <m>^2 is 0, and one would need to measure <|m|>^2. We actually only do \chi=\beta <m^2>, so that the estimate is valid only above Tc.
Would it be possible to measure also \chi below Tc (without large modifications within the code)?
Is it possible to make a snapshot (spin position+spin alignment) of the spin system at a certain time-step?
This is easily possible. In which format would you like to see the snapshot?
Wow :). I think any format would be ok, maybe a 6 column table x y z Sx Sy Sz?
Many,many thanks, best, Fritz Koermann
--- Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf
Handelsregister B 2533 Amtsgericht Düsseldorf
Geschäftsführung Prof. Dr. Jörg Neugebauer Prof. Dr. Anke R. Pyzalla Prof. Dr. Dierk Raabe Prof. Dr. Martin Stratmann Dipl.-Kfm. Herbert Wilk
Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000 -------------------------------------------------
On Jul 21, 2008, at 6:17 PM, Fritz Koermann wrote:
Hi,
The internal energy is just the short-range correlation.
Also below Tc?
Yes.
We can add correlation measurements but that is lots of data and will slow down the code quote a bit. Do you actually need all correlations?
No, I thought about nearest-neighbor-correlation <SiSi+1> or something like that.
That is just the energy.
I have two more minor questions concerning the output: Is the susceptibility \chi given out by alps equivalent to \chi=\beta*(<m^2>-<m>^2)?
On any finite lattice actually <m>^2 is 0, and one would need to measure <|m|>^2. We actually only do \chi=\beta <m^2>, so that the estimate is valid only above Tc.
Would it be possible to measure also \chi below Tc (without large modifications within the code)?
Yes, we can do that.
Is it possible to make a snapshot (spin position+spin alignment) of the spin system at a certain time-step?
This is easily possible. In which format would you like to see the snapshot?
Wow :). I think any format would be ok, maybe a 6 column table x y z Sx Sy Sz?
Just for Heisenberg models or also for other models such as Ising or XY?
Matthias Troyer
The internal energy is just the short-range correlation.
Also below Tc?
Yes.
Does it also hold if one defines J_ij up to more than one shell? E.g. do other correlation functions <SiSi+j> with j=[1..number of shells] not entering the internal energy?
We can add correlation measurements but that is lots of data and will slow down the code quote a bit. Do you actually need all correlations?
No, I thought about nearest-neighbor-correlation <SiSi+1> or something like that.
That is just the energy.
I have two more minor questions concerning the output: Is the susceptibility \chi given out by alps equivalent to \chi=\beta*(<m^2>-<m>^2)?
On any finite lattice actually <m>^2 is 0, and one would need to measure <|m|>^2. We actually only do \chi=\beta <m^2>, so that the estimate is valid only above Tc.
Would it be possible to measure also \chi below Tc (without large modifications within the code)?
Yes, we can do that.
Is it possible to make a snapshot (spin position+spin alignment) of the spin system at a certain time-step?
This is easily possible. In which format would you like to see the snapshot?
Wow :). I think any format would be ok, maybe a 6 column table x y z Sx Sy Sz?
Just for Heisenberg models or also for other models such as Ising or XY?
At the moment we are mostly interested in the Heisenberg model.
Thanks really a lot, I kindly acknowledge your help.
Best regards, Fritz Koermann
--- Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf
Handelsregister B 2533 Amtsgericht Düsseldorf
Geschäftsführung Prof. Dr. Jörg Neugebauer Prof. Dr. Anke R. Pyzalla Prof. Dr. Dierk Raabe Prof. Dr. Martin Stratmann Dipl.-Kfm. Herbert Wilk
Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000 -------------------------------------------------
On Jul 22, 2008, at 7:44 AM, Fritz Koermann wrote:
The internal energy is just the short-range correlation.
Also below Tc?
Yes.
Does it also hold if one defines J_ij up to more than one shell? E.g. do other correlation functions <SiSi+j> with j=[1..number of shells] not entering the internal energy?
If they are in your model then they will contribute, but we don't have a model taking interactions to certain shells in ALPS. Is that the model you need?
Matthias Troyer
The internal energy is just the short-range correlation.
Also below Tc?
Yes.
Does it also hold if one defines J_ij up to more than one shell? E.g. do other correlation functions <SiSi+j> with j=[1..number of shells] not entering the internal energy?
If they are in your model then they will contribute, but we don't have a model taking interactions to certain shells in ALPS. Is that the model you need?
Yes, I need such a model. I defined one using the lattice.xml based on a given 3d bcc cell. I built up an model including the first 10 interactions with edges starting from the center (0,0,0) to the nearest-neighbors in the corresponding shells. Now I would like to figure out how strong the short-range correlations contribute to the internal energy and, also, to the free energy respectively. For such reasons a short-order correlation function would be very interesting.
I have more physics questions: Concerning the free energy, is it possible to set the the magnetic entropy in the classical Heisenberg model to constant for T->0 in order to obtain the free energy via integration (over the internal energy)? Or should one better integrate over the heat capacity to get the entropy?
Best, Fritz Koermann
--- Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf
Handelsregister B 2533 Amtsgericht Düsseldorf
Geschäftsführung Prof. Dr. Jörg Neugebauer Prof. Dr. Anke R. Pyzalla Prof. Dr. Dierk Raabe Prof. Dr. Martin Stratmann Dipl.-Kfm. Herbert Wilk
Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000 -------------------------------------------------
We have added two features for you that will be available soon in the next release:
1. if you define the parameter PRINT_SWEEPS=xxx then after xxx sweeps the configuration will be dumped, for 3D Heisenberg in 7 columns: the site index, then the spatial coordinates and then the spin components.
2. we also record the energy for each bond type. Bond type 0 will be the nearest neighbor correlation you wants, etc..
Matthias
On 21 Jul 2008, at 23:20, Fritz Koermann wrote:
The internal energy is just the short-range correlation.
Also below Tc?
Yes.
Does it also hold if one defines J_ij up to more than one shell? E.g. do other correlation functions <SiSi+j> with j=[1..number of shells] not entering the internal energy?
If they are in your model then they will contribute, but we don't have a model taking interactions to certain shells in ALPS. Is that the model you need?
Yes, I need such a model. I defined one using the lattice.xml based on a given 3d bcc cell. I built up an model including the first 10 interactions with edges starting from the center (0,0,0) to the nearest-neighbors in the corresponding shells. Now I would like to figure out how strong the short-range correlations contribute to the internal energy and, also, to the free energy respectively. For such reasons a short-order correlation function would be very interesting.
I have more physics questions: Concerning the free energy, is it possible to set the the magnetic entropy in the classical Heisenberg model to constant for T->0 in order to obtain the free energy via integration (over the internal energy)? Or should one better integrate over the heat capacity to get the entropy?
Best, Fritz Koermann
Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf
Handelsregister B 2533 Amtsgericht Düsseldorf
Geschäftsführung Prof. Dr. Jörg Neugebauer Prof. Dr. Anke R. Pyzalla Prof. Dr. Dierk Raabe Prof. Dr. Martin Stratmann Dipl.-Kfm. Herbert Wilk
Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000
Hi Matthias, I think there should be an easier way to get to the alps for dummies webpage, from the main page...
Matthias Troyer wrote:
We have added two features for you that will be available soon in the next release:
- if you define the parameter PRINT_SWEEPS=xxx then after xxx sweeps
the configuration will be dumped, for 3D Heisenberg in 7 columns: the site index, then the spatial coordinates and then the spin components.
- we also record the energy for each bond type. Bond type 0 will be
the nearest neighbor correlation you wants, etc..
Matthias
On 21 Jul 2008, at 23:20, Fritz Koermann wrote:
> The internal energy is just the short-range correlation.
Also below Tc?
Yes.
Does it also hold if one defines J_ij up to more than one shell? E.g. do other correlation functions <SiSi+j> with j=[1..number of shells] not entering the internal energy?
If they are in your model then they will contribute, but we don't have a model taking interactions to certain shells in ALPS. Is that the model you need?
Yes, I need such a model. I defined one using the lattice.xml based on a given 3d bcc cell. I built up an model including the first 10 interactions with edges starting from the center (0,0,0) to the nearest-neighbors in the corresponding shells. Now I would like to figure out how strong the short-range correlations contribute to the internal energy and, also, to the free energy respectively. For such reasons a short-order correlation function would be very interesting.
I have more physics questions: Concerning the free energy, is it possible to set the the magnetic entropy in the classical Heisenberg model to constant for T->0 in order to obtain the free energy via integration (over the internal energy)? Or should one better integrate over the heat capacity to get the entropy?
Best, Fritz Koermann
Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf
Handelsregister B 2533 Amtsgericht Düsseldorf
Geschäftsführung Prof. Dr. Jörg Neugebauer Prof. Dr. Anke R. Pyzalla Prof. Dr. Dierk Raabe Prof. Dr. Martin Stratmann Dipl.-Kfm. Herbert Wilk
Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000
Hi Dummie,
Maybe I should first write an "e-mail for Dummies" page explaining the difference between mailing lists and private mail.
Anyway, as you told me in person you want an "ALPS svn for Dummies". So here we go with "Googling for Dummies"
1. open your web browser 2. go to www.google.com 3. enter "ALPS svn for Dummies" 4. click on search
Matthias
PS: this is a Wiki, just create the easier link you want
On 25 Jul 2008, at 16:12, Adrian E. Feiguin wrote:
Hi Matthias, I think there should be an easier way to get to the alps for dummies webpage, from the main page...
Matthias Troyer wrote:
We have added two features for you that will be available soon in the next release:
- if you define the parameter PRINT_SWEEPS=xxx then after xxx
sweeps the configuration will be dumped, for 3D Heisenberg in 7 columns: the site index, then the spatial coordinates and then the spin components.
- we also record the energy for each bond type. Bond type 0 will
be the nearest neighbor correlation you wants, etc..
Matthias
On 21 Jul 2008, at 23:20, Fritz Koermann wrote:
>> The internal energy is just the short-range correlation. > > Also below Tc?
Yes.
Does it also hold if one defines J_ij up to more than one shell? E.g. do other correlation functions <SiSi+j> with j=[1..number of shells] not entering the internal energy?
If they are in your model then they will contribute, but we don't have a model taking interactions to certain shells in ALPS. Is that the model you need?
Yes, I need such a model. I defined one using the lattice.xml based on a given 3d bcc cell. I built up an model including the first 10 interactions with edges starting from the center (0,0,0) to the nearest-neighbors in the corresponding shells. Now I would like to figure out how strong the short-range correlations contribute to the internal energy and, also, to the free energy respectively. For such reasons a short-order correlation function would be very interesting.
I have more physics questions: Concerning the free energy, is it possible to set the the magnetic entropy in the classical Heisenberg model to constant for T->0 in order to obtain the free energy via integration (over the internal energy)? Or should one better integrate over the heat capacity to get the entropy?
Best, Fritz Koermann
Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf
Handelsregister B 2533 Amtsgericht Düsseldorf
Geschäftsführung Prof. Dr. Jörg Neugebauer Prof. Dr. Anke R. Pyzalla Prof. Dr. Dierk Raabe Prof. Dr. Martin Stratmann Dipl.-Kfm. Herbert Wilk
Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000
Dear all,
1. Brilliant - exactly what I needed (been modifying the 1.3 spinmc code to get that) 2. How do we specify nvalues for Bond Type? I.e. the shell maximum radius in order to get the appropriate correlations in 1.3.3 version? What is the parameter?
Best and thanks, Kruno
--------------------------------------- Krunoslav Prsa, Ph. D. Student Laboratory for Neutron Scattering Paul Scherrer Institute and ETH-Zürich CH-5232 Villigen PSI, Switzerland tel: +41 56 310 20 91 mob: +41 76 386 17 99 ----------------------------------------
-----Original Message----- From: comp-phys-alps-users-bounces@phys.ethz.ch [mailto:comp-phys-alps-users-bounces@phys.ethz.ch] On Behalf Of Matthias Troyer Sent: Saturday, July 26, 2008 12:55 AM To: comp-phys-alps-users@phys.ethz.ch Subject: Re: [ALPS-users] correlation function+susceptibility
We have added two features for you that will be available soon in the next release:
1. if you define the parameter PRINT_SWEEPS=xxx then after xxx sweeps the configuration will be dumped, for 3D Heisenberg in 7 columns: the site index, then the spatial coordinates and then the spin components.
2. we also record the energy for each bond type. Bond type 0 will be the nearest neighbor correlation you wants, etc..
Matthias
On 21 Jul 2008, at 23:20, Fritz Koermann wrote:
The internal energy is just the short-range correlation.
Also below Tc?
Yes.
Does it also hold if one defines J_ij up to more than one shell? E.g. do other correlation functions <SiSi+j> with j=[1..number of shells] not entering the internal energy?
If they are in your model then they will contribute, but we don't have a model taking interactions to certain shells in ALPS. Is that the model you need?
Yes, I need such a model. I defined one using the lattice.xml based on a given 3d bcc cell. I built up an model including the first 10 interactions with edges starting from the center (0,0,0) to the nearest-neighbors in the corresponding shells. Now I would like to figure out how strong the short-range correlations contribute to the internal energy and, also, to the free energy respectively. For such reasons a short-order correlation function would be very interesting.
I have more physics questions: Concerning the free energy, is it possible to set the the magnetic entropy in the classical Heisenberg model to constant for T->0 in order to obtain the free energy via integration (over the internal energy)? Or should one better integrate over the heat capacity to get the entropy?
Best, Fritz Koermann
Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf
Handelsregister B 2533 Amtsgericht Düsseldorf
Geschäftsführung Prof. Dr. Jörg Neugebauer Prof. Dr. Anke R. Pyzalla Prof. Dr. Dierk Raabe Prof. Dr. Martin Stratmann Dipl.-Kfm. Herbert Wilk
Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000
On Jul 29, 2008, at 6:24 AM, Prsa Krunoslav wrote:
Dear all,
- Brilliant - exactly what I needed (been modifying the 1.3 spinmc
code to get that)
Do a diff on the version 1.3 and 1.3.3 and apply the patch to your sources. It only applies
- How do we specify nvalues for Bond Type? I.e. the shell maximum
radius in order to get the appropriate correlations in 1.3.3 version? What is the parameter?
You need to define a lattice with bonds up to the shall you want. I think Fritz Koermann has already done that for a bcc lattice, maybe he could post it?
Matthias
Best and thanks, Kruno
Krunoslav Prsa, Ph. D. Student Laboratory for Neutron Scattering Paul Scherrer Institute and ETH-Zürich CH-5232 Villigen PSI, Switzerland tel: +41 56 310 20 91 mob: +41 76 386 17 99
-----Original Message----- From: comp-phys-alps-users-bounces@phys.ethz.ch [mailto:comp-phys-alps-users-bounces@phys.ethz.ch ] On Behalf Of Matthias Troyer Sent: Saturday, July 26, 2008 12:55 AM To: comp-phys-alps-users@phys.ethz.ch Subject: Re: [ALPS-users] correlation function+susceptibility
We have added two features for you that will be available soon in the next release:
- if you define the parameter PRINT_SWEEPS=xxx then after xxx sweeps
the configuration will be dumped, for 3D Heisenberg in 7 columns: the site index, then the spatial coordinates and then the spin components.
- we also record the energy for each bond type. Bond type 0 will be
the nearest neighbor correlation you wants, etc..
Matthias
On 21 Jul 2008, at 23:20, Fritz Koermann wrote:
> The internal energy is just the short-range correlation. Also below Tc?
Yes.
Does it also hold if one defines J_ij up to more than one shell? E.g. do other correlation functions <SiSi+j> with j=[1..number of shells] not entering the internal energy?
If they are in your model then they will contribute, but we don't have a model taking interactions to certain shells in ALPS. Is that the model you need?
Yes, I need such a model. I defined one using the lattice.xml based on a given 3d bcc cell. I built up an model including the first 10 interactions with edges starting from the center (0,0,0) to the nearest-neighbors in the corresponding shells. Now I would like to figure out how strong the short-range correlations contribute to the internal energy and, also, to the free energy respectively. For such reasons a short-order correlation function would be very interesting.
I have more physics questions: Concerning the free energy, is it possible to set the the magnetic entropy in the classical Heisenberg model to constant for T->0 in order to obtain the free energy via integration (over the internal energy)? Or should one better integrate over the heat capacity to get the entropy?
Best, Fritz Koermann
Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf
Handelsregister B 2533 Amtsgericht Düsseldorf
Geschäftsführung Prof. Dr. Jörg Neugebauer Prof. Dr. Anke R. Pyzalla Prof. Dr. Dierk Raabe Prof. Dr. Martin Stratmann Dipl.-Kfm. Herbert Wilk
Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000
OK, clear. Best and thanks! Kruno
--------------------------------------- Krunoslav Prsa, Ph. D. Student Laboratory for Neutron Scattering Paul Scherrer Institute and ETH-Zürich CH-5232 Villigen PSI, Switzerland tel: +41 56 310 20 91 mob: +41 76 386 17 99 ----------------------------------------
-----Original Message----- From: comp-phys-alps-users-bounces@phys.ethz.ch [mailto:comp-phys-alps-users-bounces@phys.ethz.ch] On Behalf Of Matthias Troyer Sent: Tuesday, July 29, 2008 4:04 PM To: comp-phys-alps-users@phys.ethz.ch Subject: Re: [ALPS-users] correlation function+susceptibility
On Jul 29, 2008, at 6:24 AM, Prsa Krunoslav wrote:
Dear all,
- Brilliant - exactly what I needed (been modifying the 1.3 spinmc
code to get that)
Do a diff on the version 1.3 and 1.3.3 and apply the patch to your sources. It only applies
- How do we specify nvalues for Bond Type? I.e. the shell maximum
radius in order to get the appropriate correlations in 1.3.3 version? What is the parameter?
You need to define a lattice with bonds up to the shall you want. I think Fritz Koermann has already done that for a bcc lattice, maybe he could post it?
Matthias
Best and thanks, Kruno
Krunoslav Prsa, Ph. D. Student Laboratory for Neutron Scattering Paul Scherrer Institute and ETH-Zürich CH-5232 Villigen PSI, Switzerland tel: +41 56 310 20 91 mob: +41 76 386 17 99
-----Original Message----- From: comp-phys-alps-users-bounces@phys.ethz.ch [mailto:comp-phys-alps-users-bounces@phys.ethz.ch ] On Behalf Of Matthias Troyer Sent: Saturday, July 26, 2008 12:55 AM To: comp-phys-alps-users@phys.ethz.ch Subject: Re: [ALPS-users] correlation function+susceptibility
We have added two features for you that will be available soon in the next release:
- if you define the parameter PRINT_SWEEPS=xxx then after xxx sweeps
the configuration will be dumped, for 3D Heisenberg in 7 columns: the site index, then the spatial coordinates and then the spin components.
- we also record the energy for each bond type. Bond type 0 will be
the nearest neighbor correlation you wants, etc..
Matthias
On 21 Jul 2008, at 23:20, Fritz Koermann wrote:
> The internal energy is just the short-range correlation. Also below Tc?
Yes.
Does it also hold if one defines J_ij up to more than one shell? E.g. do other correlation functions <SiSi+j> with j=[1..number of shells] not entering the internal energy?
If they are in your model then they will contribute, but we don't have a model taking interactions to certain shells in ALPS. Is that the model you need?
Yes, I need such a model. I defined one using the lattice.xml based on a given 3d bcc cell. I built up an model including the first 10 interactions with edges starting from the center (0,0,0) to the nearest-neighbors in the corresponding shells. Now I would like to figure out how strong the short-range correlations contribute to the internal energy and, also, to the free energy respectively. For such reasons a short-order correlation function would be very interesting.
I have more physics questions: Concerning the free energy, is it possible to set the the magnetic entropy in the classical Heisenberg model to constant for T->0 in order to obtain the free energy via integration (over the internal energy)? Or should one better integrate over the heat capacity to get the entropy?
Best, Fritz Koermann
Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf
Handelsregister B 2533 Amtsgericht Düsseldorf
Geschäftsführung Prof. Dr. Jörg Neugebauer Prof. Dr. Anke R. Pyzalla Prof. Dr. Dierk Raabe Prof. Dr. Martin Stratmann Dipl.-Kfm. Herbert Wilk
Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000
Dear all,
first I would like to thank Matthias for the new release. Its really wonderful :)
You need to define a lattice with bonds up to the shall you want. I think Fritz Koermann has already done that for a bcc lattice, maybe he could post it?
Yes, here's the example for the bcc lattice in the lattice.xml-file. You may first define your lattice cell, for instance like:
<LATTICE name="body-centered cubic lattice" dimension="3"> <PARAMETER name="a" default="1"/> <BASIS> <VECTOR> a/2 a/2 -a/2</VECTOR> <VECTOR> a/2 -a/2 a/2</VECTOR> <VECTOR>-a/2 a/2 a/2</VECTOR> </BASIS> <RECIPROCALBASIS> <VECTOR> 2*pi/a 2*pi/a 0</VECTOR> <VECTOR> 2*pi/a 0 2*pi/a</VECTOR> <VECTOR> 0 2*pi/a 2*pi/a</VECTOR> </RECIPROCALBASIS> </LATTICE>
The interactions can than be defined as <VERTEX> in an approperiate <UNITCELL>. Please note that the positions of the nearest neighbors defining your interactions are defined in terms of the lattice vectors defined above. Including the first 3 shells the interactions may be defined as below.
<UNITCELL name="body-centered cubic lattice" dimension="3"> <VERTEX/> <EDGE type="0"><SOURCE vertex="1" offset="0 0 0"/><TARGET vertex="1" offset="1 0 0"/></EDGE> <EDGE type="0"><SOURCE vertex="1" offset="0 0 0"/><TARGET vertex="1" offset="0 1 0"/></EDGE> <EDGE type="0"><SOURCE vertex="1" offset="0 0 0"/><TARGET vertex="1" offset="0 0 1"/></EDGE> <EDGE type="0"><SOURCE vertex="1" offset="0 0 0"/><TARGET vertex="1" offset="1 1 1"/></EDGE> <EDGE type="1"><SOURCE vertex="1" offset="0 0 0"/><TARGET vertex="1" offset="1 1 0"/></EDGE> <EDGE type="1"><SOURCE vertex="1" offset="0 0 0"/><TARGET vertex="1" offset="1 0 1"/></EDGE> <EDGE type="1"><SOURCE vertex="1" offset="0 0 0"/><TARGET vertex="1" offset="0 1 1"/></EDGE> <EDGE type="2"><SOURCE vertex="1" offset="0 0 0"/><TARGET vertex="1" offset="2 1 1"/></EDGE> <EDGE type="2"><SOURCE vertex="1" offset="0 0 0"/><TARGET vertex="1" offset="1 2 1"/></EDGE> <EDGE type="2"><SOURCE vertex="1" offset="0 0 0"/><TARGET vertex="1" offset="1 1 2"/></EDGE> <EDGE type="2"><SOURCE vertex="1" offset="0 0 0"/><TARGET vertex="1" offset="1 0 -1"/></EDGE> <EDGE type="2"><SOURCE vertex="1" offset="0 0 0"/><TARGET vertex="1" offset="1 -1 0"/></EDGE> <EDGE type="2"><SOURCE vertex="1" offset="0 0 0"/><TARGET vertex="1" offset="0 1 -1"/></EDGE> </UNITCELL>
Here the interactions are always between the centered atom at position "0 0 0" and his neighbors at "1 0 0", e.g. the first 8 neighbors in the first shell (Be aware of the avoided double counting, i.e. here the interactions to "-1 0 0", "0 -1 0" etc. are neglected. It might be that therefor the interactions have to be scaled by a factor of 2 depending on the used Hamiltonian).
The interactions may be finally defined in your parameter-file like
J0=5 J1=4 J2=-2 J3=7 J4=...
where Jk corresponds to the 'type="k"' defined in the Unitcell above.
Best, Fritz Koermann
--- Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf
Handelsregister B 2533 Amtsgericht Düsseldorf
Geschäftsführung Prof. Dr. Jörg Neugebauer Prof. Dr. Anke R. Pyzalla Prof. Dr. Dierk Raabe Prof. Dr. Martin Stratmann Dipl.-Kfm. Herbert Wilk
Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000 -------------------------------------------------
comp-phys-alps-users@lists.phys.ethz.ch