Instanton calculation neads collapsing

[Raj Roy]

Hello,

I am trying to run  instanton calculation with 20 beads starting from TS geometry and its hessian. I have attached my input as well, however the beads are collapsing by the time the calculation reaches 14 steps. I am following the website tutorial. Also I am way below crossing over temperature. 

Regards,
RAJ

Ipi-input:

<simulation mode='static' verbosity='high' floatformat="%16.10f">
  <output prefix='simulation'>
    <properties stride='1' filename='out'>[ step, potential{atomic_unit}, forcemod{atomic_unit} ]</properties>
    <trajectory stride='1' format='xyz' filename='pos'>positions{atomic_unit}</trajectory>
    <trajectory stride='1' format='xyz' filename='f'>forces{atomic_unit}</trajectory>
    <checkpoint stride='5' overwrite='true' filename='chk'/>
  </output>
  <total_steps>500</total_steps>
  <prng>
    <seed>20223</seed>
  </prng>
  <ffsocket mode='inet' name='orca' pbc='False'>
    <address>localhost</address>
    <port>10207</port>
    <latency>0.01</latency>
    <timeout>7200</timeout>
  </ffsocket>
  <system>
    <initialize nbeads='20'>
      <file mode='xyz' units='angstrom'>init.xyz</file>
      <cell mode='abc' units='angstrom'>[200.0, 200.0, 200.0]</cell>
    </initialize>
    <forces>
      <force forcefield='orca'/>
    </forces>
    <ensemble>
      <temperature units='kelvin'>300</temperature>
    </ensemble>
    <motion mode='instanton'>
      <instanton mode='rate'>
        <alt_out>1</alt_out>
        <!-- ORCA VeryTightOpt: TolE=5e-7 Eh, TolMaxG=3e-5 Eh/bohr, TolMaxC=6e-6 bohr -->
        <tolerances>
          <energy units='atomic_unit'>5e-7</energy>
          <force units='atomic_unit'>4.8e-4</force>
          <position units='atomic_unit'>6e-6</position>
        </tolerances>
        <!-- Initial path stretch along imaginary mode; units are bohr (atomic_unit) -->
        <delta units='atomic_unit'>0.75</delta>
        <opt>nichols</opt>
        <hessian_update>powell</hessian_update>
        <hessian_asr>poly</hessian_asr>
        <hessian_init>false</hessian_init>
        <hessian_final>true</hessian_final>
        <biggest_step units='atomic_unit'>0.75</biggest_step>
        <!-- Single-bead TS Hessian from 03_TS; used only to build the initial path -->
        <hessian mode='file' shape='(45, 45)'>hessian.dat</hessian>
      </instanton>
    </motion>
  </system>
</simulation>

[yairl...@gmail.com]

Hi Raj,

Thanks for reaching out,
Let me ask, what is the value of the imaginary frequency?
if you provide all the inputs and outputs will be easier to diagnose the problem. 
Best,
Yair

[yairl...@gmail.com]

Hi Raj,


A freq 940.0 cm^-1 corresponds to Tc 215.404257765995 K so you might to start your first optimization at 200 K and not 300K.

Best,

Yair

PS: copying your message below so it becomes searchable

Hi Yair,

Thanks for your reply. I have attached a zip file that has 03_TS and 04_first_instanton_optimization directory. We first did TS optimization on Orca and refined it in ipi, (we built an interface between ORCA and ipi) and hessian generated was then used for instanton calculation as mentioned in the tutorial. The instanton directory has outputs for step 0 and 80.

The value of imaginary frequency = -943.2012177391579 cm^-1

Regards,
Raj