Exciton analysis

[Dhanjit]

Dear all,
I have been trying to do the ypp calculation to analyse the exciton.
But I am having some issues with in the output file

cat o-3D_BSE.exc_qpt1_weights_at_*

Code: Select all

# Electron-Hole pairs that contribute to Excitonic State 68 for iq=1 more than  5.000000%
#
#                       K-point [iku]                         Weight
# :    0.00000000         0.00000000         0.00000000        0.170892164   
# :   0.624999963E-1     0.108253032         0.00000000        0.623574913   
# :   0.124999993        0.216506064         0.00000000        0.377418816   
# :   0.187499985        0.324759096         0.00000000        0.711636543   
# :  -0.249999985       -0.433012128         0.00000000        0.502409399   
# :   0.124999993         0.00000000         0.00000000        0.317584991   
# :   0.187499985        0.108253032         0.00000000        0.553700149   
# :   0.249999985        0.216506064         0.00000000        0.567887247   
# :  -0.187499985       -0.541265190         0.00000000        0.859183729   
# :  -0.124999993       -0.433012128         0.00000000         1.00000000   
# :  -0.624999963E-1    -0.324759096         0.00000000        0.582013607   
# :    0.00000000       -0.216506064         0.00000000        0.132365316   
# :   0.249999985         0.00000000         0.00000000        0.268746912   
# :   0.312500000        0.108253032         0.00000000        0.744910300   
# :  -0.124999993       -0.649518192         0.00000000        0.739944696   
# :  -0.624999963E-1    -0.541265190         0.00000000        0.896785617   
# :    0.00000000       -0.433012128         0.00000000        0.525171518   
# :   0.374999970         0.00000000         0.00000000        0.358814508   
# :  -0.624999963E-1    -0.757771254         0.00000000        0.808874309   
# :    0.00000000       -0.649518192         0.00000000        0.373740405   
# :  -0.499999970         0.00000000         0.00000000        0.198819488   
#    
#    Band_V             Band_C             Kv-q ibz           Symm_kv            Kc ibz             Symm_kc            Weight             Energy
#    
# 
# 07/17/2023 at 12:59 ypp @ master [start]
# 07/17/2023 at 12:59              [end]
#
# .-Input file  ypp_AMPL.in
# | excitons                         # [R] Excitonic properties
# | amplitude                        # [R] Amplitude
# | ElecTemp= 0.025869         eV    # Electronic Temperature
# | States= "68 - 74"                # Index of the BS state(s)
# | BSQindex= 1                      # Q-Index of the BS state(s)
# | #DipWeight                     #  Weight the contribution to the exciton WFs with the dipoles
# | Degen_Step= 0.010000       eV    # Maximum energy separation of two degenerate states

The transition details are missing. This should have included some numbers as shown in the tutorials.
***One more thing I wanted to know that after running the BSE Screening and BSE Kernel calculation can I directly go to "Bethe-Salpeter on top of quasiparticle energies" module skipping the "scissor" module. If not then how can I get the Scissor value corresponding to my system?
The above exciton analysis has been done skipping the Scissor module.
Any help will be appreciated.

Thanks

[Dhanjit]

Dear all,
I have been trying to do the ypp calculation to analyse the exciton.
But I am having some issues with the output file

cat o-3D_BSE.exc_qpt1_weights_at_*

Code: Select all

# Electron-Hole pairs that contribute to Excitonic State 68 for iq=1 more than  5.000000%
#
#                       K-point [iku]                         Weight
# :    0.00000000         0.00000000         0.00000000        0.170892164   
# :   0.624999963E-1     0.108253032         0.00000000        0.623574913   
# :   0.124999993        0.216506064         0.00000000        0.377418816   
# :   0.187499985        0.324759096         0.00000000        0.711636543   
# :  -0.249999985       -0.433012128         0.00000000        0.502409399   
# :   0.124999993         0.00000000         0.00000000        0.317584991   
# :   0.187499985        0.108253032         0.00000000        0.553700149   
# :   0.249999985        0.216506064         0.00000000        0.567887247   
# :  -0.187499985       -0.541265190         0.00000000        0.859183729   
# :  -0.124999993       -0.433012128         0.00000000         1.00000000   
# :  -0.624999963E-1    -0.324759096         0.00000000        0.582013607   
# :    0.00000000       -0.216506064         0.00000000        0.132365316   
# :   0.249999985         0.00000000         0.00000000        0.268746912   
# :   0.312500000        0.108253032         0.00000000        0.744910300   
# :  -0.124999993       -0.649518192         0.00000000        0.739944696   
# :  -0.624999963E-1    -0.541265190         0.00000000        0.896785617   
# :    0.00000000       -0.433012128         0.00000000        0.525171518   
# :   0.374999970         0.00000000         0.00000000        0.358814508   
# :  -0.624999963E-1    -0.757771254         0.00000000        0.808874309   
# :    0.00000000       -0.649518192         0.00000000        0.373740405   
# :  -0.499999970         0.00000000         0.00000000        0.198819488   
#    
#    Band_V             Band_C             Kv-q ibz           Symm_kv            Kc ibz             Symm_kc            Weight             Energy
#    
# 
# 07/17/2023 at 12:59 ypp @ master [start]
# 07/17/2023 at 12:59              [end]
#
# .-Input file  ypp_AMPL.in
# | excitons                         # [R] Excitonic properties
# | amplitude                        # [R] Amplitude
# | ElecTemp= 0.025869         eV    # Electronic Temperature
# | States= "68 - 74"                # Index of the BS state(s)
# | BSQindex= 1                      # Q-Index of the BS state(s)
# | #DipWeight                     #  Weight the contribution to the exciton WFs with the dipoles
# | Degen_Step= 0.010000       eV    # Maximum energy separation of two degenerate states

The transition details are missing. This should have included some numbers as shown in the tutorials.
***One more thing I wanted to know that after running the BSE Screening and BSE Kernel calculation can I directly go to "Bethe-Salpeter on top of quasiparticle energies" module skipping the "scissor" module. If not then how can I get the Scissor value corresponding to my system?
The above exciton analysis has been done skipping the Scissor module.
Any help will be appreciated.

Thanks

[Daniele Varsano]

Dear Dhanjit,

it is possible that there are no transitions with a weight larger than 5% and this is why they are not reported.
You can set in input the minimum weight to be printed (the default is 5%):
EXCITON_weight_treshold=0.01
will print all transition above 1%

if you set it to zero, you will have all the transitions.

Setting a scissor is alternative to use the QP database. You can set the parameters of the scissor and stretching by plotting E_qp vs E_ks.
see e.g. Step 3 of this tutorial:
https://www.yambo-code.eu/wiki/index.ph ... rial:_h-BN

the scissor will be the difference Gap_gw-Gap_dft and the stretching is obtained by a linear fit.

Best,
Daniele

[Reza_Reza]

Dear Daniele,

I am working on a low-bandgap system, and my BSE calculations indicate that the first exciton has a negative formation energy, suggesting the possibility of an excitonic insulator. However, I need to further investigate the broken-symmetry phase.

I would like to know if there is any way to study this within Yambo. One suggested approach is to compute the spectrum of neutral excitations at zero transferred momentum, as discussed in "Assignment of Excitonic Insulators in Ab initio Theories: The Case of NiBr₂" (Physical Review B 107, 115121 (2023)). However, it is not clear how to perform this calculation in practice.

I would appreciate any guidance you can provide.

Best regards,
Reza

[Davide Sangalli]

Dear Raza,

"neutral excitations at zero transferred momentum" means computing the eigenvalues and eigenvectors of the BSE at q=0.

Having a negative energy peak is an indication that there might be an instability, which could lead to a phase transition towards a lower symmetry phase.
However, before reaching this conclusion, one should carefully check the properties of the system.

As we discuss in the manuscript you mention, a first condition is that the excitonic peak corresponds to a state with lower symmetry than the ground state. It is not trivial to compute the excitonic symmetry (this is under development in yambo at present), but for sure the exciton has lower symmetry if it is degenerate.

In case this is confirmed, the next step would be to perform a self-consistent simulation (i) allowing for symmetry breaking and (ii) including some long range eh-exchange.
But this is beyond the yambo code.

Best,
D.

[Reza_Reza]

Hello,

I have calculated the excitonic weights for a low band gap system. As shown in the attached file, the peaks appear on the negative energy side, which suggests exciton instability. However, when I checked the transitions, I observed some transitions from the LUMO to the HOMO for exciton 5 (in my system, from band 67 to 66).

Do you think this could be due to the small band gap and the resulting exciton instability?

Exciton 2
# Electron-Hole pairs that contribute to Excitonic State 2 for iq=1 more than 5.000000%
# K-point [iku] Weight
# : 0.370000005 0.00000000 0.00000000 0.516459234E-1
# : 0.379999995 0.00000000 0.00000000 0.672221109E-1
# : 0.389999986 0.00000000 0.00000000 0.870630369E-1
# : 0.400000006 0.00000000 0.00000000 0.112242289
# : 0.409999996 0.00000000 0.00000000 0.144140750
# : 0.419999987 0.00000000 0.00000000 0.184161171
# : 0.430000007 0.00000000 0.00000000 0.234241918
# : 0.439999998 0.00000000 0.00000000 0.296461463
# : 0.449999988 0.00000000 0.00000000 0.373494595
# : 0.460000008 0.00000000 0.00000000 0.468677610
# : 0.469999999 0.00000000 0.00000000 0.588238716
# : 0.479999989 0.00000000 0.00000000 0.752155960
# : 0.490000010 0.00000000 0.00000000 1.00000000
# : -0.500000000 0.00000000 0.00000000 0.219500139
#
#
# Band_V Band_C Kv-q ibz Symm_kv Kc q ibz Symm_kc Weight Energy [eV]
#
66 67 50 1 50 1 0.104599 0.082385
66 67 50 5 50 5 0.104598 0.082385
66 67 49 1 49 1 0.078662 0.129223
66 67 49 5 49 5 0.078661 0.129223
66 67 48 1 48 1 0.061506 0.181934
66 67 48 5 48 5 0.061506 0.181934
#
# 07/24/2025 at 12:24 ypp @ mc-n048.mesu [start]
# 07/24/2025 at 12:24 [end]


Exciton 5
# Electron-Hole pairs that contribute to Excitonic State 5 for iq=1 more than 5.000000%
# K-point [iku] Weight
# : 0.349999994 0.00000000 0.00000000 0.570607856E-1
# : 0.360000014 0.00000000 0.00000000 0.672515109E-1
# : 0.370000005 0.00000000 0.00000000 0.786437765E-1
# : 0.379999995 0.00000000 0.00000000 0.910738334E-1
# : 0.389999986 0.00000000 0.00000000 0.104087196
# : 0.400000006 0.00000000 0.00000000 0.116803758
# : 0.409999996 0.00000000 0.00000000 0.127951056
# : 0.419999987 0.00000000 0.00000000 0.134973317
# : 0.430000007 0.00000000 0.00000000 0.134727836
# : 0.439999998 0.00000000 0.00000000 0.122590125
# : 0.449999988 0.00000000 0.00000000 0.942901149E-1
# : 0.479999989 0.00000000 0.00000000 0.705811903E-1
# : 0.490000010 0.00000000 0.00000000 1.00000000
# : -0.500000000 0.00000000 0.00000000 0.675059259
#
#
# Band_V Band_C Kv-q ibz Symm_kv Kc q ibz Symm_kc Weight Energy [eV]
#
67 66 51 1 51 1 0.210118 -0.058916
66 67 50 5 50 5 0.155624 0.082385
66 67 50 1 50 1 0.155620 0.082385
#
# 07/24/2025 at 12:26 ypp @ mc-n048.mesu [start]
# 07/24/2025 at 12:26 [end]


Best,
Reza

[Daniele Varsano]

Dear Reza,

are you including the "coupling" in the kernel of the BSE?
If this i the case, vc and cv pairs can couple and you can have a contribution from c->v transitions.

Best,

Daniele

[Reza_Reza]

Dear Daniele,

I have attached the input file and the report file for your review.

Best,
Reza

[Daniele Varsano]

Dear Reza,

you are getting c->v pairs contribution because after applying QP corrections, your system acquires a metallic character so you have partially filled bands.

Code: Select all

 [K+QP] Metallic Bands                             :   66   67

If this is not expected check your QP calculations.

Two settings to check are:
1) XTermKind= "BG" , this is discouraged, the algorithm is not super stable and does not provide great acceleration.
2) PPAPntXp= 2.500000 Ry , there is a particular reason to set an hugh value to the PP frequency?

Best,

Daniele

[Reza_Reza]

Hi Davide,

Thank you for your message.

Yes, the system indeed shows metallic behavior — the band gap is very small, less than 0.1 eV. What I find unusual is the trend in the gap evolution:

With PBE, the band gap is around 0.05 eV.

Applying G₀W₀ slightly increases the gap.

However, when I move toward self-consistency (scGW), the gap closes again, and the system becomes effectively metallic.

Best,

Reza