The file xxx_OUT4GEN.rrm should be written in the same format in every case (TASK: "MAKE GUESS", "MICROITERATION", "ENERGY", "ENERGY and GRADIENT", or "ENERGY, GRADIENT, and HESSIAN").

The required format, for example, for 3-atom system is:

RESULTS
CURRENT COORDINATE
atom1  x₁  y₁  z₁
atom2  x₂  y₂  z₂
atom3  x₃  y₃  z₃
ENERGY =  E₁  E₂  E₃
       =  E₄  E₅  E₆
S**2   =  <S²>
GRADIENT
  g(x₁)
  g(y₁)
  g(z₁)
  g(x₂)
  g(y₂)
  g(z₂)
  g(x₃)
  g(y₃)
  g(z₃)
DIPOLE =  μ(x)  μ(y)  μ(z)
HESSIAN
  h(x₁,x₁)
  h(y₁,x₁)  h(y₁,y₁)
  h(z₁,x₁)  h(z₁,y₁)  h(z₁,z₁)
  h(x₂,x₁)  h(x₂,y₁)  h(x₂,z₁)  h(x₂,x₂)
  h(y₂,x₁)  h(y₂,y₁)  h(y₂,z₁)  h(y₂,x₂)  h(y₂,y₂)
  h(z₂,x₁)  h(z₂,y₁)  h(z₂,z₁)  h(z₂,x₂)  h(z₂,y₂)
  h(x₃,x₁)  h(x₃,y₁)  h(x₃,z₁)  h(x₃,x₂)  h(x₃,y₂)
  h(y₃,x₁)  h(y₃,y₁)  h(y₃,z₁)  h(y₃,x₂)  h(y₃,y₂)
  h(z₃,x₁)  h(z₃,y₁)  h(z₃,z₁)  h(z₃,x₂)  h(z₃,y₂)
  h(z₂,z₂)
  h(x₃,z₂)  h(x₃,x₃)
  h(y₃,z₂)  h(y₃,x₃)  h(y₃,y₃)
  h(z₃,z₂)  h(z₃,x₃)  h(z₃,y₃)  h(z₃,z₃)
DIPOLE DERIVATIVES
  dμ(x,x₁)  dμ(y,x₁)  dμ(z,x₁)
  dμ(x,y₁)  dμ(y,y₁)  dμ(z,y₁)
  dμ(x,z₁)  dμ(y,z₁)  dμ(z,z₁)
  dμ(x,x₂)  dμ(y,x₂)  dμ(z,x₂)
  dμ(x,y₂)  dμ(y,y₂)  dμ(z,y₂)
  dμ(x,z₂)  dμ(y,z₂)  dμ(z,z₂)
  dμ(x,x₃)  dμ(y,x₃)  dμ(z,x₃)
  dμ(x,y₃)  dμ(y,y₃)  dμ(z,y₃)
  dμ(x,z₃)  dμ(y,z₃)  dμ(z,z₃)
POLARIZABILITY
  α(x,x)
  α(x,y)  α(y,y)
  α(x,z)  α(y,z)  α(z,z)

If the file xxx_OUT4GEN.rrm is written in different formats, GRRM17 cannot read xxx_OUT4GEN.rrm. Therefore you need to put zero (0.0) into matrix elements of GRADIENT, DIPOLE, HESSIAN, DIPOLE DERIVATIVES, and/or POLARIZABILITY that cannot be obtained from your "sub link = aaa" program. Note: ENERGY requires 6 elements (E₁ – E₆). In usual cases, the energy from your "sub link = aaa" program have to be put into the first element (E₁), and put zero into the others (E₂ – E₆). The other E₂ – E₆ elements are reserved for the multistate calculations or some special usages.

In the case "TASK: MICROITERATION", GRRM17 requests you to perform microiterations by your "sub link = aaa" program and to give xxx_OUT4GEN.rrm which contains the coordinates after microiterations, energy, and gradients of the external atoms (although δV/δQ_n is almost zero for coordinates of external atoms Q_n after microiterations). GRRM17 does not perform microiterations, and this is a specification (this might seem to be strange, but this is designed for general use).

For further details, try a microiteration calculation with Gaussian program, and look at the xxx_LinkJOB.rrm file. You will understand how GRRM17 does a microiteration calculation. Even if you use the general interface, GRRM17 requires the same information.

The GRRM17 program uses Gaussian program to only compute single point calculation, and the file name_GauJOB.log is the Gaussian log file of the last step of your calculation. In this case, it would correspond to the freq calculation log file at the minimum point of the backward IRC path, so that you cannot see the IRC path information in the file name_GauJOB.log with GaussView. The IRC information is stored in the file name.log. To visualize the structure in IRC path computed by GRRM17 with GaussView, you need to plot the corresponding structure and energy separately from the file name.log.

In the "Bond Condition ... END" option, "12 75 < 2.06" means "the distance between atoms 12 and 75 to be less than or equal to 2.06 angstroms". This bond condition will be appled to the SC-AFIR search, but not to the IRC path calculation. Do not confuse this option with the IRC=Phase=(N1,N2 [,N3 [,N4]]) option in Gaussian. In GRRM17, an IRC calculation always computes forward and backward IRC paths.

The options "NoBondRearrange", "Bond Condition ... END", and "Add Interaction ... END" have no effect on the IRC calculation, because "IRC" keyword just performs the IRC path calculation. Such options are effective in the "SC-AFIR" calculation. In SC-AFIR, the "Bond Condition ... END" option cannot be used simultaniously with the "NoBondRearrange" option (the latter keyword in the input file is only effective).

You may need to add the basis to the valence part of Au:

GauInpB
C H N O F Si 0
6-31g**
****
Au 0
SDD
****

Au 0
SDD
END

This issue seems to be solved by the above comment. (However, you can still add comments to this thread.)

All services on our website are running now.

Thank you for your cooperation.

GRRM17 does not recognize the keyword "%link=molpro2015" or later.  Consequently, GRRM17 is officially compatible with Molpro 2006, 2008, 2010, and 2012. However, you can use Molpro 2015 or later at your own risk.

With Molpro 2015: 

Whereas we have not fully tested Molpro2015 jobs, the formats of output files of Molpro 2015 seems to be the same as those of Molpro2012, and GRRM17 will work with Molpro2015 by using the link keyword "%link=molpro2012". 

With Molpro 2018 or later: 

GRRM17 knows a Molpro job has been terminated normally if the string "Variable memory released" is appeared in the Molpro output file (XXX_MolJOB.out), because Molpro2015 or earlier prints the string in the last line of the output. But, Molpro2018 or later uses "Molpro calculation terminated" instead of "Variable memory released". Thus, GRRM17 will never understand the termination of the Molpro2018 (or later) calculation, and then GRRM17 program will stop with an error.  This issue will be evaded by using the TEXT card in Molpro input. For example: 

Molpro Input
BASIS=6-311+G(d,p)

{MCSCF,maxit=39,GRADIENT=1.d-8
{ITERATIONS;DO,UNCOUPLE,11,TO,39;}
START,*
ORBITAL,*
Occ,12;
Closed,4;
WF,16,1,0;
STATE,2;}

{RS2,Mix=2,Root=2,SHIFT=0.3,MaxIt=99,MaxIti=999;
ORBITAL,IGNORE_ERROR;
WF,16,1,0;
STATE,2;}

FORCE;
IF(STATUS.LT.0) STOP
$STR='Variable memory'
TEXT,$STR released
---
RSPT2 STATE 2.1
END

We are aware of another issue; the format of the punch file (XXX_moljob.pun) is different from that of Molpro2015 or earlier, then GRRM17 cannot read the energies from the punch file. This issue will be evaded by setting "submol" command to the shellscript in which the punch file format to be converted. For example: 

   setenv submol "./script.csh"   (csh/tcsh case)

and prepare "script.csh" like as 

#!/bin/csh -f
set PUN=`echo $2:r.pun | tr \[A-Z\] \[a-z\]`
set TMP='tmp'

molpro -W./ -d./ $1 $2

grep "Energy" ${PUN} > ${TMP}
set i = 1
set n = `wc -l ${TMP} | awk '{print $1}'`
while ( $i <= $n )
   set line = `head -n $i ${TMP} | tail -n 1`
   echo $line >> ${PUN}
   @ i = $i + 1
end
\rm -f ${TMP}
unset i, n, line
unset PUN, TMP

exit

Alternatively, you can use the general interface in GRRM17. Please see "General interface with external ab initio programs".

We are aware of another issue; the format of the punch file (XXX_moljob.pun) is bit different from that of Molpro2015 or earlier, then GRRM17 cannot read the energies from the punch file. For example:

RSPT2 STATE 1.1 Energy    -114.21130105   (Molpro2015 or earlier)

RSPT2 STATE  1.1 Energy    -114.21130105   (Molpro2018 or later)

So, make sure whether you provide an appropriate string to the Molpro Input section. For example,

Molpro Input
︙
IF(STATUS.LT.0) STOP
$STR='Variable memory'
TEXT,$STR released
---
RSPT2 STATE  2.1
END

We have made several tests using Molpro 2015.1 (patch level 33), Molpro 2018.2, and Molpro 2019.1.2, but we cannot give any warranty.

To combine ONIOM and MC-AFIR, please use the following format.

  atom  x  y  z  part-number  layer  [link-atom info]

where

  atom :  atom type
  x :  X-coordinate
  y :  Y-coordinate
  z :  Z-coordinate
  part-number :  part number for MC-AFIR (see Part designation for random structure generation)
  layer :  layer assignment (H/M/L) for ONIOM
  [link-atom info] :  (as needed) link-atom parameters for ONIOM

In your case, for example:

H  -5.09198232  1.46148646  0.64415956  1  H
H  -5.05103711  1.51846702 -1.80580446  1  H
H  -5.08619301  3.61195144 -0.53172902  1  H
H  -3.07552771  2.21038282 -0.53072587  1  H
Si -4.57617575  2.20057460 -0.55601978  1  H
Si  0.92262713 -0.42951851 -0.06586408  2  H
H   0.32586471  0.89377867  0.44846657  2  H
H   0.92141623 -0.42839950 -1.60585939  2  H
Si -0.36596427 -2.22060001  0.71381911  2  H
Si  0.54148635 -4.23086015 -0.06762047  2  H
...
Si  7.89452207  2.71361008 -0.06626650  2  L H 16
Si  6.60446476  0.92311796  3.05275285  2  L H 16
Si  3.11867252 -0.64788568  3.05254379  2  L H 12
Si -2.56235026 -2.00324774 -0.06393534  2  L H  9
Si  1.83058292 -2.43960890  3.83181251  2  L H 19
Si  1.45021923 -6.24149307  3.83086153  2  L H 19
Si -0.74565359 -6.02268238  0.71234080  2  L H 10
Si  2.35719613 -8.25146150  3.04887994  2  L H 23
Si  4.55185684 -8.47158271  3.82932265  2  L H 33
...

Although GRRM17 doesn't contain an interface with DMol3, GRRM17 can be used together with any energy computation code if a user prepare a simple interface script. Please see the following page:

https://afir.sci.hokudai.ac.jp/documents/manual/48

Dear Prof. Maeda,

I am sorry but I clarify my question.

I search TS and EQ structures for the initial reaction of A + B system by using SC-AFIR/UB3LYP/6-31G(p).
A is a tri-atomic molecule and B is an organic molecule including one nitrogen and some carbon and hydrogen atoms.
After one month starting a SC-AFIR calculation (B= one nitrogen + three carbons with some hydorgens) with options (i.e. NRUN=32,GauMem=1000,GauProc=4,Add Interaction,and Gamma=100 alog with SC=NegativeON,SC=LUP-ON,NoBondRearrange,DownDC=15,RandDC=15,Stable=OPT), 343EQ and 629TS (as well as 5900 TS in PT_list) were found. But this job does not done yet. EQ and TS searches are not stopping.
Could you please inform how to finish the SC-AFIR calculation properly? Should I perporm FirstOnly calculations in advance to found out proper initial conformation?

Thank you very much in advance.
Ayako Furuhama

Please see https://afir.sci.hokudai.ac.jp/documents/manual/50.

In general, I recommend to do RePATH in which the previous DS-AFIR results are read.

The ridge point will be used as the TS guess only when the DS = SingleSTEP option is used. Without this option, LUP is performed after the AFIR path is obtained, and energy maxima along the final LUP path are used as the TS guesses.