Correct input file for poly_aniso

chucksty · 2020-03-09 15:19:17

Hi Team,

I am looking at computing the exchange interaction between two Er3+ ions in a dinuclear magnetic complex of Ci-symmetry.

First, I have computed for two single_aniso part for two fragments (Lu-Er and Er-Lu) using the input file below for each case;

&seward
Cholesky High
Basis Set
Er.ano-rcc
 Er1     -1.468767    3.805721    3.757264      Angstrom
End of Basis Set

Basis Set
Lu.ANO-RCC.
 Lu2      1.753301    1.766787    5.400763      Angstrom
End of Basis Set
...
basis for other atoms also included
...
...
AngM
 -1.468767    3.805721    3.757264  Angstrom
AMFI
End of Input

&RASSCF
Spin
4
nActel
11 0 0
Inactive
177
Ras1
0
Ras2
7
Ras3
0
CIROOT
35 35 1
OrbL
ALL
ORBA
FULL
End Of Input

&GRID_IT
NAME=ras
End of Input

&RASSI  &END
MEES
Properties
3
'AngMom ' 1
'AngMom ' 2
'AngMom ' 3
NR  OF  JOBIPHS
1 35
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
SpinOrbit
EJOB
End of Input

&SINGLE_ANISO
MLTP
2
4 2
ZJPR
-0.2
ENCU
100 100
XFIE
0.1
HINT
0 7.0 71
TINT
0 300 301
MAVE
1 12
TMAG
6 1.0 1.2 1.8 2.5 2.9 3.6
CRYS
Er
QUAX
1
UBAR
PLOT
End Of Input

Now, I would like to use poly_aniso to compute the exchange interaction. The manual at https://molcas.gitlab.io/OpenMolcas/sph … aniso.html explains how we could do this. But I am not sure I understand the illustration therein.

1. To begin with, I can't seem to find any aniso_i.input file after the single_aniso job. Is this the same as the input file for &Single_aniso or is this the ANISOINPUT text file that was generated by the &single_aniso job?
2. Under NNEQ, I sure understand what the T and F stand for but I am kinda confused on what magnetic centres are counted for the first number in the next line after the NNEQ keyword.
For my case, I would say, I have two magnetic centres, (2 X Er) as Lu3+ is not a magnetic ion. However, this complex has a centre of inversion and so possess an asymmetric unit with one Er3+ centre which means that the 2 Er3+ in each asymmetric unit that fused together to form the formed complex are equivalent. If this is true, then is it correct to specify my input as follows;

&seward
Cholesky High
Basis Set
Er.ano-rcc
 Er1     -1.468767    3.805721    3.757264      Angstrom
End of Basis Set

Basis Set
Lu.ANO-RCC.
 Lu2      1.753301    1.766787    5.400763      Angstrom
End of Basis Set
...
basis for other atoms also included
...
...
End of Input



&POLY_ANISO
NNEQ
1  T  F (1 because although there are 2 Er centres both of them are identical or should this be 2 because the two Er have different coordinate. Or Am I meant to consider the Lu centres I have used to replace the Er centres as well? What is the right number here? I have used T because I have computed the single_aniso at ab initio level)
2        (If  I go by the fact that there are 2 magnetic centres, 2 of which are of the same type, then I will have 2 on this next line)
4         (The next line is supposed to give the number of spin-orbit states to be computed. However, I am not sure I understand how to choose this number. I have used 4 because the ground state multiplicity of Er3+ is 4 and so I expect to have 4^2 = 16 Nex. Is this correct)
UBAR
PLOT
End Of Input

How do I link the ANISOINPUT files?
what else do I need to specify to run the job?
Could you help me with a sample input for a dinuclear lanthanide complex in which the two lanthanide centres are identical?

I would really appreciate if you could help me out with this. Thanks in advance for your kind help.

Chucksty

Last edited by chucksty (2020-03-10 11:22:40)

liviu.ungur · 2020-03-15 10:01:41

Dear chucksty,

Here is a short answer to your particular problem. A more detailed tutorial on how to run POLY_ANISO will be posted in a separate thread.
1) Each successful run of CASSCF/RASSI/SINGLE_ANISO produces an ASCII file named ANISOINPUT. It is located in the $WorkDir and has to be saved to your local folder for subsequent use. As a result of your two fragment calculations (Lu-Er and Er-Lu) you should have two ANISOINPUT files from each of the two calculations. Let's denote them as "ANISOINPUT_Lu_Er" and "ANISOINPUT_Er_Lu"

2) Linking the files can be done either by simply copying them to $WorkDir or as soft links in your shell script.

>>COPY $FileDir/ANISOINPUT_Er_Lu  $WorkDir/aniso_1.input
>>COPY $FileDir/ANISOINPUT_Lu_Er  $WorkDir/aniso_2.input

The following input for POLY_ANISO is explained as follows. The NNEQ keyword specifies the number of independent magnetic sites. In your case, there are two independent magnetic sites. Therefore, the input is:

&POLY_ANISO 
NNEQ
2  T  F
1  1
4  4

The first line tells that there are two independent sites, the first (T) means that both sites have been computed ab initio and we have an ANISOINPUT file for it, the last (F) means that we are not using the HDF5 file format (still in testing).
The second line tells the program that we have only one site of each kind.
The third line tells that we wish to consider four spin-orbit eigenstates as the exchange basis for each of the sites. In this case, the size of the exchange matrix is 16x16.

The next keyword is the one defining the magnetic interaction. It is one or more from the list: LIN1, LIN3, LIN9, PAIR, etc., as listed in the manual. The interaction values have to be provided. They could be adjusted such as to simulate the experimental XT or M, if available, or estimated from other calculations, e.g. broken-symmetry DFT.

PAIR
1
1 2   -1.232

The above input tells the code that there is only one interaction, between sites 1 and 2, of the strength -1.232 cm-1. This interaction will be included for the ground Kramers Doublet as well as between the first excited Kramers Doublets of each Er ions included in the calculation. In the case, the experimental values for the macroscopic properties (susceptibility, magnetization, etc.) are available, the interaction parameter could be adjusted such as to minimize the standard deviation between calculation and experiment. Use the keywords TEXP or HEXP accordingly. The exchange coupling parameter could be varied either inside you submitting shell script or inside the input as follows:

>> FOREACH J in ( -2.0, -1.0,  0.0,  1.0, 2.0 ) 

&POLY\_ANISO
 NNEQ
   2  T  F
   1  1
   4  4
 PAIR
 1
 1  2  $J
 TEXP
 NT
 values for T(i)   values for XT (i)  (NT lines, NT = integer) 
 ...
 End Of Input

>> ENDDO

Keyword COOR can be used to tell the program that the dipole-dipole interaction to be accounted for using the ab initio input data.
UBAR, PLOT, etc. can be used as wished. The generated plots are located in the $WorkDir and are overwritten each time the POLY_ANISO is executed. I use them in the final calculation when the exchange parameters are known.

3) In case the binuclear compound has exact symmetry, then the ab initio calculation can be done only on one site. The other center is exactly the same and need not be computed. However, it is related to the original site by some symmetry operation, e.g. inversion. This information is given to the POLY_ANISO by the keyword SYMM. The input, in this case, would be:

>>COPY $FileDir/ANISOINPUT_Er_Lu  $WorkDir/aniso_1.input

>> FOREACH J in ( -2.0, -1.0,  0.0,  1.0, 2.0 ) 

&POLY_ANISO
 NNEQ
   1  T  F
   2
   4
 PAIR
 1
 1  2  $J
 SYMM
 1
 1.0 0.0 0.0
 0.0 1.0 0.0
 0.0 0.0 1.0
-1.0 0.0 0.0
 0.0 -1.0 0.0
 0.0 0.0 -1.0
 TEXP
 NT
 values for T(i)   values for XT (i)  (NT lines, NT = integer) 
 ...
 End Of Input

>> ENDDO

The keyword NNEQ tells the code that there are two sites of type 1. The keyword SYMM provides the transformation matrices (3x3) for each of the sites. The second site is transformed into the first one by the inversion matrix (i.e. minus identity matrix).

Last edited by liviu.ungur (2020-03-16 03:13:32)

chucksty · 2020-03-16 11:58:30

Thank you so much, Liviu.

I will try tis out now.

Regards
Chucksty

chucksty · 2020-06-25 01:13:39

Dear Liviu,
Sorry to bother you again on this. Please does HEXP not work with poly_aniso?
I have specified every other thing, but the HEXP section seems to be causing problems. Below is the input section of it;

  HEXP
  21
  0.000000        -0.00060409
  2.500000        0.84604
  5.000000        1.41341
  7.500000        1.74782
  10.00000        1.95572
  12.50000        2.09398
  15.00000        2.19366
  17.50000        2.27017
  20.00000        2.3374
  22.50000        2.39318
  25.00000        2.4424
  27.50000        2.48663
  30.00000        2.53083
  32.50000        2.57022
  35.00000        2.60952
  37.50000        2.64431
  40.00000        2.67981
  42.50000        2.7113
  45.00000        2.73597
  47.50000        2.76551
  50.00000        2.79518
  TMAG
  1.8

When I run the job, it crashes with the error message below;

-- ----------------------------------

>>> COPY /filestore/home/molcas_calculation/anisofiles/ANISOINPUT_A_Lu.input /scratch/home/job-160854/aniso_1.input

>>> COPY /filestore/home/molcas_calculation/anisofiles/ANISOINPUT_Lu_B.input /scratch/home/job-160854/aniso_2.input

>>> EXPORT J = -2.0

--- Start Module: poly_aniso at Thu Jun 25 00:57:16 2020 ---

()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

                                            &POLY_ANISO

                                   only a single process is used
                      available to each process: 1.0 TB of memory, 12 threads
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

POLY_ANISO (OPEN):
by:   Liviu Unugur       (chmlu@nus.edu.sg)
and   Liviu F. Chibotaru (Liviu.Chibotaru@kuleuven.be)
Last updated - 2 July 2018
The code allocated at least:         3809648 bytes of memory for this run.
The data provided in TMAG will be ignored.
  READIN: Error reading "poly_aniso.input"
  near line nr.                   179
--- Stop Module: poly_aniso at Thu Jun 25 00:57:16 2020 /rc=_RC_INPUT_ERROR_ ---
*** files: xmldump
    saved to directory /filestore/dci24/molcas_calculation/poly_aniso

>>> END FOREACH

    Timing: Wall=0.12 User=0.02 System=0.01

If I remove the HEXP section, the other parts of the calculation run smoothly without any problem.

Also, I am wondering why it is trying to ignore the TMAG value I provided ("The data provided in TMAG will be ignored."). If that is ignored, where does it get the temperature for the calculation? Or is this also happening because of the HEXP problem? By the way, I am using the latest version of molcas.

Thanks

Regards

chucksty · 2020-06-26 11:28:40

I have now solved this problem. Apparently, the HEXP keyword wasn't working because I didn't specify the temperature at which the magnetisation is to be computed within the HEXP block. I thought TMAG should suffix to specify the temperature but it isn't. The temperature(s) at which the magnetisation is to be computed still need to be specified under the HEXP keyword. This is not documented in the manual but looking at the source code reveal that the right input is

HEXP
NT T1 T2 ... TN (Where NT is the number of temperature points at which the experimental magnetisation was measured and needs to be computed; T1,T2...TN are the respective temperatures)
NH            (Where NH is the number of field points for which the magnetization was measured)
H1   MagT1  MagT2 ... MagTN
H2   MagT1  MagT2 ... MagTN
 .       .
 .       .
 .       .        (Where H1 is the field in tesla and MagT1 is the measured magnetization in Borh magneton at T1 etc)

For example

HEXP
4 1.8 2.0 2.4 3.0
5
0.1  0.004  0.003  0.0025  0.002
0.8  0.02    0.018  0.016   0.010
1.0  0.500  0.450  0.440   0.400
1.4  0.604  0.600  0.580   0.560
2.0  0.802  0.800  0.790   0.760

This time you don't need to specify TMAG anywhere. I just thought I should post it to help others and to also confirm that it is the correct thing.

Regards
Chucksty

Last edited by chucksty (2020-06-26 19:10:56)

Molcas Forum

Announcement

#1 2020-03-09 15:19:17

Correct input file for poly_aniso

#2 2020-03-15 10:01:41

Re: Correct input file for poly_aniso

#3 2020-03-16 11:58:30

Re: Correct input file for poly_aniso

#4 2020-06-25 01:13:39

Re: Correct input file for poly_aniso

#5 2020-06-26 11:28:40

Re: Correct input file for poly_aniso

Board footer