🏖️ Relaxation

📖 The Relaxation step propagates or optimizes cone point geometries on the lower electronic state to identify possible ground-state photoproducts.

---

Concept Overview

• Each cone point generated during cone sampling is used as a starting geometry.
• NANR provides three relaxation workflows:
• Optimization only (geometry minimization)
• No initial velocity MD + optimization
• Random initial velocity MD + optimization
• These methods emulate how molecules relax after nonadiabatic transitions, revealing accessible minima on the ground-state PES.

---

🛌 Step-By-Step: Optimization Only

1. From each cone_sampling/ directory, create an opt.in template:

   coordinates start.xyz
   basis 6-31gs
   spinmult 1
   charge 0
   method wb97x
   run minimize
   dftd no
   purify no
   precision mixed
   scf diis+a
   

2. Modify or reuse the submission script:

   #!/bin/bash
   #SBATCH -p l40-gpu
   #SBATCH -N 1
   #SBATCH -n 1
   #SBATCH -J opt
   #SBATCH --mem=50G
   #SBATCH -t 10:00:00
   #SBATCH --qos gpu_access
   #SBATCH --gres=gpu:1
   #SBATCH -o slurm-%j.out
   #SBATCH -e slurm-%j.err
   
   module load tc/25.03
   terachem opt.in > opt.out
   

3. Prepare and run your jobs:

   chmod +x submit.sh
   sbatch submit.sh
   

4. The final geometry (bottom N+2 lines of optim.xyz) corresponds to your optimized photoproduct.

---

🚶‍♀️ + 🛌 Step-By-Step: No Initial Velocity MD + Optimization

1. Create an md.in file:

   coordinates XXX
   basis 6-31gs
   charge 0
   method wb97x
   spinmult 1
   run md
   precision mixed
   nstep 100
   

2. Launch MD and follow with optimization:

   sbatch submit.sh
   

3. After the MD completes, extract the final coordinates from scr.coords/coors.xyz, then run optimization as above.

---

🏃‍♀️ + 🛌 Step-By-Step: Random Initial Velocity MD + Optimization

1. Generate random velocity files:

   filenames:
     work_path: <molecule>/seam_sampling/clustering/ciopt#/cone_sampling/
   velocity_sampling:
     samples_per_point: 5
     added_energy: 1.8289  # in eV
   

   python velocity_sampling.py
   

2. For each velocity index (V0, V1, ...), perform MD as in the previous workflow.

3. After MD completion, extract the last coordinates to initialize optimization directories (O0/, O1/, etc.).

4. Submit optimization jobs for each set:

   sbatch submit.sh
   

---

Outputs

• opt.out → TeraChem optimization log
• optim.xyz → converged geometry
• scr.start/optim.xyz → final relaxed structure

Each relaxed structure corresponds to a photoproduct or intermediate accessible from a particular cone point direction.

---

Continue back to 🏠 Home →