Author Topic: Can anyone help me with convergence error problems?  (Read 2239 times)

0 Members and 1 Guest are viewing this topic.

Offline Lim changmin

  • Heavy QuantumATK user
  • ***
  • Posts: 34
  • Country: kr
  • Reputation: 0
    • View Profile
Can anyone help me with convergence error problems?
« on: February 7, 2025, 13:52 »
Hello, I am trying to calculate transmission spectrum at 0 eV with Silicon/Silicide contact with different doping concentration(5e19,1e20,5e20,1e21)

I tried to converge the LCAO calculator but the device would not converge.

I changed density mesh cutoff, device algorithm(neutral atom, equivalent bulk etc..), contour parameter(ozaki point) for about a week with same structure, but nothing would work.

I tried to solve this problem on my own, but after spending a week with same structure, I realized it was beyond my beyond my capabilities.
Therefore, I have decided to seek assistance.

I am wondering if I am making a mistake or the structure is not properly optimized(0.02eV/A, optimize length : 10A). (I attached txt file of the calculation information)

Can anyone could give me some tips regarding with convergence problem?

When I check the log file, for some reason, the convergence energy suddenly changed to higher value. (40E~42E)

26 E = -940.453 dE =  3.211179e-01 dH =  8.870016e-01                       |
|  27 E = -942.197 dE =  1.744240e+00 dH =  8.995829e-01                       |
|  28 E =  -941.24 dE =  9.571292e-01 dH =  7.565638e-01                       |
|  29 E = -942.876 dE =  1.635776e+00 dH =  8.258672e-01                       |
|  30 E = -943.925 dE =  1.048822e+00 dH =  8.581108e-01                       |
|  31 E =  -944.33 dE =  4.050945e-01 dH =  9.593208e-01                       |
|  32 E = -942.058 dE =  2.272090e+00 dH =  8.716657e-01                       |
|  33 E = -942.011 dE =  4.704906e-02 dH =  8.909611e-01                       |
|  34 E = -941.766 dE =  2.451188e-01 dH =  7.616103e-01                       |
|  35 E = -942.428 dE =  6.625315e-01 dH =  7.147292e-01                       |
|  36 E = -942.211 dE =  2.172081e-01 dH =  6.725936e-01                       |
|  37 E = -942.583 dE =  3.722324e-01 dH =  6.988655e-01                       |
|  38 E = -942.315 dE =  2.688179e-01 dH =  7.281209e-01                       |
|  39 E = -946.068 dE =  3.753141e+00 dH =  7.930580e-01                       |
|  40 E = -944.435 dE =  1.633055e+00 dH =  6.637546e-01                       |
|  41 E =  -961.58 dE =  1.714573e+01 dH =  2.050847e+00                       |
|  42 E = -2420.58 dE =  1.458995e+03 dH =  8.792229e+01                       |
|  43 E = -2558.35 dE =  1.377787e+02 dH =  9.421465e+01                       |
|  44 E = -2414.11 dE =  1.442476e+02 dH =  8.872245e+01                       |
|  45 E = -1620.06 dE =  7.940496e+02 dH =  5.332355e+01                       |
|  46 E = -1685.53 dE =  6.547581e+01 dH =  5.754290e+01                       |
|  47 E = -2391.43 dE =  7.058958e+02 dH =  9.732136e+01                       |
|  48 E = -2297.45 dE =  9.397452e+01 dH =  9.073635e+01                       |
 

Thank you for reading the question

Offline Anders Blom

  • QuantumATK Staff
  • Supreme QuantumATK Wizard
  • *****
  • Posts: 5620
  • Country: dk
  • Reputation: 107
    • View Profile
    • QuantumATK at Synopsys
Re: Can anyone help me with convergence error problems?
« Reply #1 on: February 11, 2025, 01:54 »
Looks like a reasonable set up but I would definitely make sure it converges with GGA before trying HSE! Also, I have no experience with the Ozaki contour, again I would stick with the simplest option first, i.e. the default contour method. You could try to use fewer history steps, sometimes it works better with say 10 only, but I expect the convergence looks different with GGA, so that is step 1.