non-linear response

Deals with issues related to computation of optical spectra, in RPA (-o c) or by solving the Bethe-Salpeter equation (-o b). Includes local field effects, excitons, etc.

Moderators: Davide Sangalli, andrea marini, Conor Hogan, myrta gruning

Re: non-linear response

Postby sitangshu » Wed May 29, 2019 9:14 am

Dear Sir,

While doing a SHG non linear calculation, I intentionally did not calculated the static screening and I encountered with this line:

[WARNING] HXC collisions not found. Potential is computed on the fly.

I did not understand the "fly" concept. Is this a wrong approach?

With regards,
Sitangshu
Sitangshu Bhattacharya
Indian Institute of Information Technology-Allahabad
India
Web-page: http://profile.iiita.ac.in/sitangshu/
Institute: http://www.iiita.ac.in/
sitangshu
 
Posts: 93
Joined: Thu Jan 05, 2017 8:08 am

Re: non-linear response

Postby claudio » Wed May 29, 2019 9:50 am

Dear Sitangshu

if you are doing RPA (TD-HARTREE) or TDDFT there are two ways to do it.

1) you calculate the COLLISIONS integrals before the calculation and then use them to reconstruct the Hartree
potential in the SHG calculation.

2) you calculate the Hartree potential at each step in the dynamics

the result is the same.
I did not put online the tutorial for the first option but I will do it soon.
in which calculation are you interested?

best
Claudio
Claudio Attaccalite
Institut Neel, CNRS/UJF ABINEEL group
25 rue des Martyrs BP 166, Bâtiment D
38042 Grenoble cedex 9 France
web site: http://www.attaccalite.com
Freely download scientific books from: http://www.freescience.info
User avatar
claudio
 
Posts: 215
Joined: Tue Mar 31, 2009 11:33 pm

Re: non-linear response

Postby sitangshu » Wed May 29, 2019 12:14 pm

Dear Professor Claudio,

Thank you for your quick response. I was doing a TDSHF calculation, but I started first with an IPA (with DG). The SHG-IPA result seems to be quite convincing over a large number of laser frequencies. The next approximation level I am using is TD-Hartree. At TD Hartree, I encountered this "on the fly" situation. I am aware that a static dielectric function should be calculated first to figure out the screening, but I did not find any one in yambo community to use on the fly situation, so I was in a bit of dilemma. It is not mentioned anywhere also... I am pasting a snippet of my TDDFT calculation. The simulation time used in our calculation is larger than the noise in polarization vs time plot. Can you please comment on this?

[05.01] Initialization
======================

Dephasing Time [fs]: 32.910594
Sampling Time [fs]: 8.3213347
Total simulation time [fs]: 55.000000

[05.02] Messaging
=================

---General---
Integrator :CRANKNIC
Simulation time [fs]: 55.000000
Step length [fs]: 0.01000
NE Steps : 5501
Correlation :TDDFT
Damping [eV]: 0.2000000
Dephasing [fs]: 6.5821189
Efield :SOFTSIN
Efield FWHM [fs]: 2.0000000
Efield Amplitude [V/m]:0.13725E+7
Frequencies range : 0.5000000 6.0000000
Frequencies steps : 10
Frequency step : 0.5500000
Use Dipoles :no
Induced Field :no
Force Snd Order Dipoles :no
Lenght of G vectors in each direction: 0.94663934 0.94663934 0.22166123
Lenght of G vectors in each direction: 0.94663934 0.94663934 0.22166123
[WF] Performing Wave-Functions I/O from ./SAVE

[WF loader] Normalization (few states) min/max :0.5012E-14 0.9963


[WARNING] Covariant Dipoles not in 3d. Assuming non periodic system.
Using forth order approximation for covariant dipoles
[WF] Performing Wave-Functions I/O from ./SAVE

[FFT-Oscillators/R space] Mesh size: 7 7 25

[WR./SAVE//ndb.dip_iR_and_P]--------------------------------
Brillouin Zone Q/K grids (IBZ/BZ): 144 144 144 144
RL vectors (WF): 601
Fragmentation :yes
Electronic Temperature [K]: 0.2000000
Bosonic Temperature [K]: 0.2000000
X band range : 1 33
X band range limits : 33 1
X e/h energy range [ev]:-1.0000000 -1.0000000
RL vectors in the sum : 601
[r,Vnl] included :no
Using shifted grids :no
Using covariant dipoles:yes
Using G-space approach :no
Using R-space approach :no
Direct v evaluation :no
Field momentum norm :0.10000E-4
Wavefunctions :Perdew, Burke & Ernzerhof(X)+Perdew, Burke & Ernzerhof(C)
- S/N 009235 -------------------------- v.04.03.02 r.00134 -
[WR./SAVE//ndb.Overlap]-------------------------------------
Brillouin Zone Q/K grids (IBZ/BZ): 144 144 144 144
RL vectors (WF): 601
Fragmentation :yes
Electronic Temperature [K]: 0.2000000
Bosonic Temperature [K]: 0.2000000
X band range : 1 33
X band range limits : 33 1
X e/h energy range [ev]:-1.0000000 -1.0000000
RL vectors in the sum : 601
[r,Vnl] included :no
Using shifted grids :no
Using covariant dipoles:yes
Using G-space approach :no
Using R-space approach :no
Direct v evaluation :no
Field momentum norm :0.10000E-4
Wavefunctions :Perdew, Burke & Ernzerhof(X)+Perdew, Burke & Ernzerhof(C)
- S/N 009235 -------------------------- v.04.03.02 r.00134 -
Lenght of G vectors in each direction: 0.94663934 0.94663934 0.22166123
[WF] Performing Wave-Functions I/O from ./SAVE

[FFT-NL] Mesh size: 7 7 25


[05.03] External/Internal QP corrections
========================================

[QP apply] Ind. Gap Correction [ev]: 1.8400000

[WARNING] HXC collisions not found. Potential is computed on the fly.

* * * Polarization * * *
==========================
Volume : 1441.94978746 bohr^3
Direction: 1
Numer of k-points per string : 12
Numer of different strings : 12
Electronic Phase : -0.04064865 mod(2)
Ionic Phase : -0.00000011 mod(1)
Total Phase : -0.04064876 mod(1)
Polarization : -0.31153870 mod( 7.66416224) (e/Volume).bohr
Polarization : -0.00021605 mod( 0.00531514) e/bohr^2
Polarization : -0.01236147 mod( 0.30410436) C/m^2

Direction: 2
Numer of k-points per string : 12
Numer of different strings : 12
Electronic Phase : 0.02316682 mod(2)
Ionic Phase : 0.00000011 mod(1)
Total Phase : 0.02316693 mod(1)
Polarization : 0.17755513 mod( 7.66416224) (e/Volume).bohr
Polarization : 0.00012314 mod( 0.00531514) e/bohr^2
Polarization : 0.00704517 mod( 0.30410436) C/m^2

The total polarization has module 1 !!

Initial Polarization :-.27762E-3 0.10664E-3 0.0000
[WR./SAVE//ndb.Nonlinear]-----------------------------------
Brillouin Zone Q/K grids (IBZ/BZ): 144 144 144 144
RL vectors (WF): 601
Global Gauge :length
[RT] Time steps : 5501
[RT] Time step [fs]: 0.01000
[NL] Damping [fs]: 0.2000000
[NL] Bands ange : 24 33
[NL] Energy range : 0.5000000 6.0000000
[NL] Number of freqs : 10
[NL] Force 2nd approx :no
[NL] Use DIPOLES :no
[NL] Induced Field :no
[RT] Integrator :CRANKNIC
[NL] Correlation :TDDFT
[FIELDs] Type :SOFTSIN
[FIELDs] Versor : 1.0000000 0.0000000 0.0000000
[FIELDs] Intensity [kWCMm2]: 500.00000
[FIELDs] Damping [fs]: 82.682747
[FIELDs] Frequency range [eV]: 0.100000 0.100000
[FIELDs] Frequency steps : 1
[FIELDs] Frequency step [eV]: 0.0000000
Sitangshu Bhattacharya
Indian Institute of Information Technology-Allahabad
India
Web-page: http://profile.iiita.ac.in/sitangshu/
Institute: http://www.iiita.ac.in/
sitangshu
 
Posts: 93
Joined: Thu Jan 05, 2017 8:08 am

Re: non-linear response

Postby claudio » Wed May 29, 2019 1:18 pm

Did you specify the simulation time in input or you leave -1?

regarding the TDSHF I will put only a tutorial within one week, so you can follow it

best
Claudio
Claudio Attaccalite
Institut Neel, CNRS/UJF ABINEEL group
25 rue des Martyrs BP 166, Bâtiment D
38042 Grenoble cedex 9 France
web site: http://www.attaccalite.com
Freely download scientific books from: http://www.freescience.info
User avatar
claudio
 
Posts: 215
Joined: Tue Mar 31, 2009 11:33 pm

Re: non-linear response

Postby sitangshu » Wed May 29, 2019 1:33 pm

No, I specified it after some initial testing with various simulation times. In any case if the time is short, the ypp eventually warns that the time is not long enough.

Regards,
Sitangshu.
Sitangshu Bhattacharya
Indian Institute of Information Technology-Allahabad
India
Web-page: http://profile.iiita.ac.in/sitangshu/
Institute: http://www.iiita.ac.in/
sitangshu
 
Posts: 93
Joined: Thu Jan 05, 2017 8:08 am

Re: non-linear response

Postby sitangshu » Tue Jun 04, 2019 6:41 am

Dear Sir,

In continuation to my above mail, I did the HXC collision cohsex calculation, and generated the SHG using Screen-EXchange. The resultant SHG curve is convincing exhibiting spin-spliting in the lowest peak. The rise is parabolic with increasing laser frequency.
In order to understand the dielectric at w and w/2, I followed a pump-probe (-p option) analogy. The pump is chosen to be quasisine centred at w frequency. The integrator chosen is RK2 RWA. The probe field is generated using –u option. The field chosen is delta. As far as I understood, this delta should come after the pump and thus, I chose ExtF_Tstart > Field1_Tstart. The NL integrator and correlation are Crank-Nicholsan type and SEX. The script generates an additional calculation of negf. An RT simulation time of 0.3 ps is used. I hope this is correct. A quick run generates additional files ndb.RT_JPSM_Time* and spin_and_P2* . There occurs a start and restart calculation followed by a RT Dynamics computation. A snippet shows this:
[07.03] Dephasing matrix setup
==============================

[RT] Quasiparticle Dephasing applied

[07.04] Electronic Collisions loading and setup
===============================================

[RD./SAVE//ndb.COLLISIONS_HXC]------------------------------
Brillouin Zone Q/K grids (IBZ/BZ): 144 144 144 144
RL vectors (WF): 36937
COLLISIONS name :SEX
Total number of COLLISIONS : 7920
COLLISIONS kind : 327
Bands : 24 33
Bands mixing : 100.00000
Exchange Cutoff :0.50000E-3
Exchange RL vectors : 29189
X matrix size :511
X band range :1 - 200
X e/h energy range [ev]:-1.000000 - -1.000000
X Time ordering :T
X xc-Kernel :none
X Drude frequency : 0.000000 0.000000
X poles [o/o]: 100.0000
RL vectors in the sum :36937
[r,Vnl] included : no
Field direction :0.1000E-4 0.000000 0.000000
BZ energy Double Grid : yes
BZ energy DbGd points :15696
BZ Q point size factor : 1.000000
RIM RL components : 111
RIM random points : 1000000
- S/N 009235 -------------------------- v.04.03.02 r.00134 -

[WARNING][COLLISIONS] Found (and using) an Hartree-free collisions database

[07.05] Messaging
=================

---General---
Integrator :EULER + SIMPLE
Simulation time [fs]: 55.000000
Basic time step [fs]: 0.01000
steps : 5501

---Flags (if any)---
Hxc approximation :COHSEX

---I/O---
T between I/O of J,P,f [fs]: 0.050000
I/O of GFs [fs]: 0.050000
outputs [fs]: 0.100000
Extended collisions [I/O]:yes

---Fields---
Gauge :length
Field1 field :DELTA
Field1 polarization :linear
Field1 energy window/period [eV/fs]: 2.9669200 2.9669200 1.3939261 1.3939261
Field1 elemental oscillation [fs]: 0.0000000
Field1 width [fs]: 0.8493218
Field1 energy steps : 1
Field1 electric field [V/m]:0.13725E+7
Field1 max intensity [kW/cm^2]: 500.00000
Field1 max fluence [nJ/cm^2]: 0.0000000
Field1 final time [fs]: 2.5200000
Field1 half maximum full width [fs]: 2.0000000

[07.06] Start and Restart
=========================

[07.07] Real-Time dynamics
==========================

[07.07.01] Initialization
=========================

Unfortunately, there appears to be an error mentioning:
<30m-31s> P0100: [07.07.01] Initialization
P0100: [ERROR] STOP signal received while in :[07.07.01] Initialization
P0100: [ERROR]Unknow integrator

I am wondering what am I missing?
Can you please let me know about this error?


Regards,
Sitangshu
Sitangshu Bhattacharya
Indian Institute of Information Technology-Allahabad
India
Web-page: http://profile.iiita.ac.in/sitangshu/
Institute: http://www.iiita.ac.in/
sitangshu
 
Posts: 93
Joined: Thu Jan 05, 2017 8:08 am

Re: non-linear response

Postby claudio » Fri Jun 07, 2019 2:29 pm

Dear Sitangshu

try to follow this new tutorial

http://www.attaccalite.com/lumen/real_time_bse.html

and use the CRANKNIC integrator

let me know if it is fine

best
Claudio
Claudio Attaccalite
Institut Neel, CNRS/UJF ABINEEL group
25 rue des Martyrs BP 166, Bâtiment D
38042 Grenoble cedex 9 France
web site: http://www.attaccalite.com
Freely download scientific books from: http://www.freescience.info
User avatar
claudio
 
Posts: 215
Joined: Tue Mar 31, 2009 11:33 pm

Re: non-linear response

Postby sitangshu » Wed Jul 03, 2019 7:40 am

Dear Sir,

Thank you for scripting the tutorial. It took me a while to follow this and to understand the methods.
So, I did the calculation of a static BSE first then I compared it with the RT BSE. In both the cases the field is along 1 1 0. Therefore, first I carefully generated a double grid by breaking and mapping the symmetry to the field direction along 110.
I find that the BSE and RT-BSE is having a slight mismatch. However, the peak position are intact in both the cases.
I am not able to find out what I did wrong.

1. For the HXC static dielectric case, I choose (only the main parts I am showing):
% COLLBands
24 | 33 | # [COLL] Bands for the collisions
%
HXC_Potential= "HARTREE+SEX" # [SC] SC HXC Potential
EXXRLvcs= 500 RL # [XX] Exchange RL components
CORRLvcs= 500 RL # [GW] Correlation RL components

The same block size and BandsRnXs and bands range is used for calculating static dielectric and in BSE.

2. For the NL part, I choose (only the main parts I am showing):

NLstep= 0.0100 fs # [NL] Real Time step length
NLtime= 55.000000 fs # [NL] Simulation Time
NLintegrator= "CRANKNIC" # [NL] Integrator ("EULEREXP/RK2/RK4/RK2EXP/HEUN/INVINT/CRANKNIC")
NLCorrelation= "SEX" # [NL] Correlation ("IPA/HARTREE/TDDFT/LRC/LRW/JGM/SEX")
NLEnSteps= 1 # [NL] Energy steps
NLDamping= 0.100000 eV # [NL] Damping
ExtF_FrStep= 0.000000 eV # [NL ExtF] Frequency step
ExtF_Int= 1000. kWLm2 # [NL ExtF] Intensity
ExtF_Damp= 2.000000 fs # [NL ExtF] FWHM
ExtF_kind= "DELTA" # [NL ExtF] Kind(SIN|SOFTSIN|RES|ANTIRES|GAUSS|DELTA|QSSIN)
ExtF_Tstart= 0.0100 fs # [NL ExtF] Initial Time
GfnQPdb= "none" # [EXTQP G] Database
GfnQP_N= 1 # [EXTQP G] Interpolation neighbours
% GfnQP_E
1.84 | 1.000000 | 1.000000 | # [EXTQP G] E parameters (c/v) eV|adim|adim

3. For ypp, I chose:
Xorder= 1 # Max order of the response functions
% TimeRange
-1.000000 |-1.000000 | fs # Time-window where processing is done
%
ETStpsRt= 500 # Total Energy steps
% EnRngeRt
0.00000 | 10.00000 | eV # Energy range
%
DampMode= "NONE" # Damping type ( NONE | LORENTZIAN | GAUSSIAN )
DampFactor= 0.100000 eV # Damping parameter

The report file says:
[05.01] Initialization
======================

Dephasing Time [fs]: 65.821189
Sampling Time [fs]: 41.406673
Total simulation time [fs]: 55.000000

I am attaching the o.polarization and o.YPP_eps_along E.
Can you please comment on this?

With regards,
SItangshu
You do not have the required permissions to view the files attached to this post.
Sitangshu Bhattacharya
Indian Institute of Information Technology-Allahabad
India
Web-page: http://profile.iiita.ac.in/sitangshu/
Institute: http://www.iiita.ac.in/
sitangshu
 
Posts: 93
Joined: Thu Jan 05, 2017 8:08 am

Re: non-linear response

Postby sitangshu » Wed Jul 03, 2019 7:55 am

P.S: While doing BSE, I additionally did :

BSEmod= "coupling"

and uncommented:
WehCpl

Is it OK to match this condition with RT BSE?

Regards,
Sitangshu
Sitangshu Bhattacharya
Indian Institute of Information Technology-Allahabad
India
Web-page: http://profile.iiita.ac.in/sitangshu/
Institute: http://www.iiita.ac.in/
sitangshu
 
Posts: 93
Joined: Thu Jan 05, 2017 8:08 am

Re: non-linear response

Postby claudio » Wed Jul 03, 2019 9:36 am

Dear Sitangshu

everything seems fine to me,

may you send me your input files, in such a way that I can replicate your results
and make some additional tests.

best
Claudio
Claudio Attaccalite
Institut Neel, CNRS/UJF ABINEEL group
25 rue des Martyrs BP 166, Bâtiment D
38042 Grenoble cedex 9 France
web site: http://www.attaccalite.com
Freely download scientific books from: http://www.freescience.info
User avatar
claudio
 
Posts: 215
Joined: Tue Mar 31, 2009 11:33 pm

PreviousNext

Return to Linear Response

Who is online

Users browsing this forum: No registered users and 1 guest

cron