strange convergence problem about cp2k qmmm

[Geng Sun]

Hello, everyone,

      I am using the cp2k qmmm to optimize the structure of Pt6 supported 
on Al2O3 surface. But I encounter a problem about the 
      convergence of the qmmm.

     1)  Firstly, the scf of the QM part converges strangely like this:  
All of the QM SCF converges within two steps, and the second number in the 
Covergence column is always zero.

     Step     Update method      Time    Convergence         Total 
energy    Change
  
------------------------------------------------------------------------------
     1 Broy./Dav.  0.40E+00    4.1     2.02291583       -81.2822944519 
-8.13E+01
     2 Broy./Dav.  0.40E+00    5.8     0.0000E+00       -70.8618426104  
1.04E+01

  *** SCF run converged in     2 steps ***

    2) Then, the convergence for the structure is also strange. 

     I want the max force is below 0.05 eV/A, so I use something like this

  &GEO_OPT
    MAX_ITER 20
    OPTIMIZER LBFGS
    MAX_FORCE [eV*angstrom^-1] 0.05
  &END GEO_OPT

     But fininally, the program ends "normally" with this massage (shown 
below). Obviously, In the previous Convergence check, the maxforce  is 
larger
 Than what I want. 

  Convergence check :
  Max. step size             =         0.0000000000
  Conv. limit for step size  =         0.0030000000
  Convergence in step size   =                  YES
  RMS step size              =         0.0000000000
  Conv. limit for RMS step   =         0.0015000000
  Convergence in RMS step    =                  YES
  Max. gradient              =         0.5334857508
  Conv. limit for gradients  =         0.0009723453
  Conv. for gradients        =                   NO
  RMS gradient               =         0.0328663882
  Conv. limit for RMS grad.  =         0.0003000000
  Conv. for gradients        =                   NO
 ---------------------------------------------------

 --------------------------
 OPTIMIZATION STEP:      2
 --------------------------

 ***********************************************
 * Specific L-BFGS convergence criteria
 * WANTED_PROJ_GRADIENT and WANTED_REL_F_ERROR
 * satisfied .... run CONVERGED!
 ***********************************************

    So, what is the source of this error in my calculation. 

    Thank you for any advice !

    The input files and the input structure is in the attachment. 


Geng Sun
              
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20150901/74f68075/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: test.tgz
Type: application/x-gtar
Size: 851595 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20150901/74f68075/attachment.gtar>