Second-harmonic generation of 2D-hBN (5.3) - Revision history

Yambowikiadmin at 11:17, 19 May 2025

2025-05-19T11:17:51Z

← Older revision		Revision as of 11:17, 19 May 2025
Line 288:		Line 288:
	= References =		= References =
	<references/>		<references/>

			=== Links for 2025 school ===
			* Back to [[Modena 2025#Tutorials]]

Yambowikiadmin at 11:10, 19 May 2025

2025-05-19T11:10:18Z

Yambowikiadmin at 11:07, 19 May 2025

2025-05-19T11:07:25Z

← Older revision		Revision as of 11:07, 19 May 2025
Line 283:		Line 283:
	[[File:02 SHG SEX.png\|600px\|center\|SHG at the Hartree + Screened exchange from the current run with 12 frequencies (crosses) and a run with 112 frequencies (keeping the other parameters the same)]]		[[File:02 SHG SEX.png\|600px\|center\|SHG at the Hartree + Screened exchange from the current run with 12 frequencies (crosses) and a run with 112 frequencies (keeping the other parameters the same)]]

	~~Also in this case, we~~ note differences between the two simulations. ~~As discussed for the independent particle approximation, this~~ is due to the too short simulation time. Repeating the calculations with a longer time gives more accurate results.		We note differences between the two simulations. This is due to the too short simulation time. Repeating the calculations with a longer time gives more accurate results.
	In addition, note that in the interest of time, all parameters are '''under converged'''. In principle, to calculate the collisions one should use the same parameters of a converged BSE calculation. As well, to eliminate spurious interaction with the periodic images, one should use the Coulomb cut-off as it was used for this system in the [[Quasi-particles of a 2D system\| tutorial on quasiparticle calculations]]. As a reference, a converged calculation for this system is in <ref name="Gruning2014">M. Grüning and C. Attaccalite, [https://journals.aps.org/prb/abstract/10.1103/PhysRevB.89.081102 Phys. Rev. B '''89''', 081102]</ref>.		In addition, note that in the interest of time, all parameters are '''under converged'''. In principle, to calculate the collisions one should use the same parameters of a converged BSE calculation. As well, to eliminate spurious interaction with the periodic images, one should use the Coulomb cut-off as it was used for this system in the [[Quasi-particles of a 2D system\| tutorial on quasiparticle calculations]]. As a reference, a converged calculation for this system is in <ref name="Gruning2014">M. Grüning and C. Attaccalite, [https://journals.aps.org/prb/abstract/10.1103/PhysRevB.89.081102 Phys. Rev. B '''89''', 081102]</ref>.

	= References =		= References =
	<references/>		<references/>

Yambowikiadmin at 11:06, 19 May 2025

2025-05-19T11:06:10Z

← Older revision		Revision as of 11:06, 19 May 2025
Line 281:		Line 281:
	plot '04_hsex_files/o-04_SHG_hsex.YPP-X_probe_order_2' u 1:(sqrt($22+$32)/pmVm1toCGS) w p title "calc"		plot '04_hsex_files/o-04_SHG_hsex.YPP-X_probe_order_2' u 1:(sqrt($22+$32)/pmVm1toCGS) w p title "calc"

	[[File:SHG ~~hsex~~.png\|SHG at the Hartree + Screened exchange from the current run with 12 frequencies (crosses) and a run with 112 frequencies (keeping the other parameters the same)]]		[[File:02 SHG SEX.png\|600px\|center\|SHG at the Hartree + Screened exchange from the current run with 12 frequencies (crosses) and a run with 112 frequencies (keeping the other parameters the same)]]

	Also in this case, we note differences between the two simulations. As discussed for the independent particle approximation, this is due to the too short simulation time. Repeating the calculations with a longer time gives more accurate results.		Also in this case, we note differences between the two simulations. As discussed for the independent particle approximation, this is due to the too short simulation time. Repeating the calculations with a longer time gives more accurate results.

Yambowikiadmin at 10:53, 19 May 2025

2025-05-19T10:53:09Z

← Older revision		Revision as of 10:53, 19 May 2025
Line 186:		Line 186:
	plot '01_SHG_ip_files/o-01_SHG_ip.YPP-X_probe_order_2' u 1:(sqrt($22+$32)/pmVm1toCGS) w p title "calc"		plot '01_SHG_ip_files/o-01_SHG_ip.YPP-X_probe_order_2' u 1:(sqrt($22+$32)/pmVm1toCGS) w p title "calc"

	[[File: ~~SHG_intensity_of_2D_h-BN_obtained_from_the_previous_run_(and_compared_with_a_run_on_112_frequencies)~~.png \| SHG_intensity_of_2D_h-BN obtained from the current run (12 frequencies) compared with a run with 112 frequencies]]		[[File:02 SHG IP.png\|600px\|center\|SHG_intensity_of_2D_h-BN obtained from the current run (12 frequencies) compared with a run with 112 frequencies]]

	In the figure above, the results of the current run are compared with those for a run with 112 frequencies (one can obtain these results by repeating the calculations, changing the number of frequencies in <code>NLEnSteps</code> though this take about 10 times more computational time than the run with 12). One can notice some differences between the two runs.		In the figure above, the results of the current run are compared with those for a run with 112 frequencies (one can obtain these results by repeating the calculations, changing the number of frequencies in <code>NLEnSteps</code> though this take about 10 times more computational time than the run with 12). One can notice some differences between the two runs.
Line 192:		Line 192:
	As an exercise, you can repeat the simulation for e,g, a time of <code>35 fs</code> nand <code>60 fs</code>, that is setting <code>NLtime</code>. The comparison of the susceptibilities at different simulation times (as in the plot below) shows the importance of running long enough simulations. In too short simulations, the response signal coming from the switch-on of the electric field (delta-kick) is still of comparable intensities as the relevant second-order signal.		As an exercise, you can repeat the simulation for e,g, a time of <code>35 fs</code> nand <code>60 fs</code>, that is setting <code>NLtime</code>. The comparison of the susceptibilities at different simulation times (as in the plot below) shows the importance of running long enough simulations. In too short simulations, the response signal coming from the switch-on of the electric field (delta-kick) is still of comparable intensities as the relevant second-order signal.

	[[File:02 SHG IP l.png\|SHG of 2D-hBN with IPA. Effect of different simulation time]]		[[File:02 SHG IP l.png\|600px\|center\|SHG of 2D-hBN with IPA. Effect of different simulation time]]

	As a final note, the SHG depends on the size of the vacuum added in the supercell. Instead, one should consider the surface SHG which obtained by considering an effective thickness for the layer. This is usually chosen to be similar to the layer separation in the bulk counterpart.		As a final note, the SHG depends on the size of the vacuum added in the supercell. Instead, one should consider the surface SHG which obtained by considering an effective thickness for the layer. This is usually chosen to be similar to the layer separation in the bulk counterpart.

Yambowikiadmin at 10:49, 19 May 2025

2025-05-19T10:49:29Z

← Older revision		Revision as of 10:49, 19 May 2025
Line 188:		Line 188:
	[[File: SHG_intensity_of_2D_h-BN_obtained_from_the_previous_run_(and_compared_with_a_run_on_112_frequencies).png \| SHG_intensity_of_2D_h-BN obtained from the current run (12 frequencies) compared with a run with 112 frequencies]]		[[File: SHG_intensity_of_2D_h-BN_obtained_from_the_previous_run_(and_compared_with_a_run_on_112_frequencies).png \| SHG_intensity_of_2D_h-BN obtained from the current run (12 frequencies) compared with a run with 112 frequencies]]

	In the figure above, the results of the current run are compared with those for a run with 112 frequencies (one can obtain these results by repeating the calculations, changing the number of frequencies in <code>NLEnSteps</code> though this take about 10 times more computational time than the run with 12). One can notice some differences between the two runs. In fact, the two runs evaluate different frequnecies and the difference indicates spurious oscillations of the value of the polarizability due to the simulation time being too short and thus to the presence of 'noise' due to the eigenfrecies.		In the figure above, the results of the current run are compared with those for a run with 112 frequencies (one can obtain these results by repeating the calculations, changing the number of frequencies in <code>NLEnSteps</code> though this take about 10 times more computational time than the run with 12). One can notice some differences between the two runs.
	As an exercise, you can repeat the simulation for e,g, a time of <code>60 fs</code>, that is setting <code>NLtime=60</code>. The plot below compare the SHG extracted at 42 fs and 60 fs, where the latter gives clearly more accurate results.

	[[File:~~Comparison of the~~ SHG ~~intensity of 2D h-BN obtained with different simulation time~~.png\|~~Comparison of the~~ SHG ~~intensity~~ of 2D h-~~BN obtained~~ with different simulation time]]		As an exercise, you can repeat the simulation for e,g, a time of <code>35 fs</code> nand <code>60 fs</code>, that is setting <code>NLtime</code>. The comparison of the susceptibilities at different simulation times (as in the plot below) shows the importance of running long enough simulations. In too short simulations, the response signal coming from the switch-on of the electric field (delta-kick) is still of comparable intensities as the relevant second-order signal.

			[[File:02 SHG IP l.png\|SHG of 2D-hBN with IPA. Effect of different simulation time]]

	As a final note, the SHG depends on the size of the vacuum added in the supercell. Instead, one should consider the surface SHG which obtained by considering an effective thickness for the layer. This is usually chosen to be similar to the layer separation in the bulk counterpart.		As a final note, the SHG depends on the size of the vacuum added in the supercell. Instead, one should consider the surface SHG which obtained by considering an effective thickness for the layer. This is usually chosen to be similar to the layer separation in the bulk counterpart.

Yambowikiadmin at 10:11, 19 May 2025

2025-05-19T10:11:13Z

← Older revision		Revision as of 10:11, 19 May 2025
Line 119:		Line 119:
	\| DIP_ROLEs= "v.c.k" # [PARALLEL] CPUs roles (k,c,v)		\| DIP_ROLEs= "v.c.k" # [PARALLEL] CPUs roles (k,c,v)

	Also, look at the simulation time. This is approximately 42 fs.		Also, look at the simulation time. This is approximately 45-50 fs.

	Plot the polarization for a couple of frequencies, e.g. the first and last of the chosen range (the corresponding frequency can be read in the file, search for "Frequency value"):		Plot the polarization for a couple of frequencies, e.g. the first and last of the chosen range (the corresponding frequency can be read in the file, search for "Frequency value"):

Yambowikiadmin at 09:02, 19 May 2025

2025-05-19T09:02:54Z

← Older revision		Revision as of 09:02, 19 May 2025
Line 95:		Line 95:
	Run the simulation, possibly in parallel e. g. in a interactive session on 4 cores.		Run the simulation, possibly in parallel e. g. in a interactive session on 4 cores.

	'''[In order to get 4 cores while in an interacting node, exit your current ~~allocation~~ and rerun <code>~~salloc~~</code> asking for the proper number of tasks:'''		'''[In order to get 4 cores while in an interacting node, exit your current interactive run and rerun <code>srun</code> asking for the proper number of tasks:'''
	~~salloc~~ -A ~~tra23_Yambo~~ -p ~~m100_usr_prod~~ -~~q m100_qos_dbg~~ --~~nodes=~~1 -~~-ntasks-per-node=<span style="color:red">~~4~~</span>~~ -~~-cpus-per-task=~~4 -t 02:00:00		srun -A tra25_yambo -p dcgp_usr_prod --reservation=s_tra_yambo -N 1 -n 4 -c 4 -t 04:00:00 --gres=tmpfs:10g --pty /bin/bash
	'''Otherwise, you can submit the job with a slurm script as explained in the introductory page]'''		'''Otherwise, you can submit the job with a slurm script as explained in the introductory page]'''
			Note that when executing commands that don't need mpi (such as generating an input file) you may need to add -nompi to the command line

	Now execute the command:		Now execute the command:

Yambowikiadmin at 08:57, 19 May 2025

2025-05-19T08:57:47Z

← Older revision		Revision as of 08:57, 19 May 2025
Line 282:		Line 282:

	Also in this case, we note differences between the two simulations. As discussed for the independent particle approximation, this is due to the too short simulation time. Repeating the calculations with a longer time gives more accurate results.		Also in this case, we note differences between the two simulations. As discussed for the independent particle approximation, this is due to the too short simulation time. Repeating the calculations with a longer time gives more accurate results.
	In addition, note that in the interest of time, all parameters are ~~<bold>~~under converged~~</bold>~~. In principle, to calculate the collisions one should use the same parameters of a converged BSE calculation. As well, to eliminate spurious interaction with the periodic images, one should use the Coulomb cut-off as it was used for this system in the [[Quasi-particles of a 2D system\| tutorial on quasiparticle calculations]]. As a reference, a converged calculation for this system is in <ref name="Gruning2014">M. Grüning and C. Attaccalite, [https://journals.aps.org/prb/abstract/10.1103/PhysRevB.89.081102 Phys. Rev. B '''89''', 081102]</ref>.		In addition, note that in the interest of time, all parameters are '''under converged'''. In principle, to calculate the collisions one should use the same parameters of a converged BSE calculation. As well, to eliminate spurious interaction with the periodic images, one should use the Coulomb cut-off as it was used for this system in the [[Quasi-particles of a 2D system\| tutorial on quasiparticle calculations]]. As a reference, a converged calculation for this system is in <ref name="Gruning2014">M. Grüning and C. Attaccalite, [https://journals.aps.org/prb/abstract/10.1103/PhysRevB.89.081102 Phys. Rev. B '''89''', 081102]</ref>.

	= References =		= References =
	<references/>		<references/>

Yambowikiadmin at 08:57, 19 May 2025

2025-05-19T08:57:11Z

← Older revision		Revision as of 08:57, 19 May 2025
Line 282:		Line 282:

	Also in this case, we note differences between the two simulations. As discussed for the independent particle approximation, this is due to the too short simulation time. Repeating the calculations with a longer time gives more accurate results.		Also in this case, we note differences between the two simulations. As discussed for the independent particle approximation, this is due to the too short simulation time. Repeating the calculations with a longer time gives more accurate results.
	In addition, note that in the interest of time, all parameters are under converged. In principle, to calculate the collisions one should use the same parameters of a converged BSE calculation. As well, to eliminate spurious interaction with the periodic images, one should use the Coulomb cut-off as it was used for this system in the [[Quasi-particles of a 2D system\| tutorial on quasiparticle calculations]]. As a reference, a converged calculation for this system is in <ref name="Gruning2014">M. Grüning and C. Attaccalite, [https://journals.aps.org/prb/abstract/10.1103/PhysRevB.89.081102 Phys. Rev. B '''89''', 081102]</ref>.		In addition, note that in the interest of time, all parameters are <bold>under converged</bold>. In principle, to calculate the collisions one should use the same parameters of a converged BSE calculation. As well, to eliminate spurious interaction with the periodic images, one should use the Coulomb cut-off as it was used for this system in the [[Quasi-particles of a 2D system\| tutorial on quasiparticle calculations]]. As a reference, a converged calculation for this system is in <ref name="Gruning2014">M. Grüning and C. Attaccalite, [https://journals.aps.org/prb/abstract/10.1103/PhysRevB.89.081102 Phys. Rev. B '''89''', 081102]</ref>.

	= References =		= References =
	<references/>		<references/>

← Older revision		Revision as of 11:10, 19 May 2025
Line 127:		Line 127:
	replot '01_SHG_ip_files/o-01_SHG_ip.polarization_F12' u 1:2 w l lw 2 title "omega = 7.375 eV"		replot '01_SHG_ip_files/o-01_SHG_ip.polarization_F12' u 1:2 w l lw 2 title "omega = 7.375 eV"

	[[File:Berry_phase_polarization_of_2D_h-BN_obtained_from_the_previous_run_for_two_laser_frequencies.png\|Berry_phase_polarization_of_2D_h-BN_obtained from the run for two laser frequencies]]		[[File:Berry_phase_polarization_of_2D_h-BN_obtained_from_the_previous_run_for_two_laser_frequencies.png\|600px\|center\|Berry_phase_polarization_of_2D_h-BN_obtained from the run for two laser frequencies]]

	After the applied field is 'switched on', the polarization is oscillating at the frequency of the applied field and at the eigenfrequencies of the system as described above in the cartoon. This is visible particularly for the higher frequency. Due to the dephasing, after about 20 fs the polarization is oscillating (at least to the eye) at the frequency of the applied field. Nonlinear components are not visible since they are several orders of magnitude smaller than the first order.		After the applied field is 'switched on', the polarization is oscillating at the frequency of the applied field and at the eigenfrequencies of the system as described above in the cartoon. This is visible particularly for the higher frequency. Due to the dephasing, after about 20 fs the polarization is oscillating (at least to the eye) at the frequency of the applied field. Nonlinear components are not visible since they are several orders of magnitude smaller than the first order.
Line 188:		Line 188:
	[[File:02 SHG IP.png\|600px\|center\|SHG_intensity_of_2D_h-BN obtained from the current run (12 frequencies) compared with a run with 112 frequencies]]		[[File:02 SHG IP.png\|600px\|center\|SHG_intensity_of_2D_h-BN obtained from the current run (12 frequencies) compared with a run with 112 frequencies]]

	In the figure above, the results of the current run are compared with those for a run with 112 frequencies (one can obtain these results by repeating the calculations, changing the number of frequencies in <code>NLEnSteps</code> though this take about 10 times more computational time than the run with 12)~~. One can notice some differences between the two runs~~.		In the figure above, the results of the current run are compared with those for a run with 112 frequencies (one can obtain these results by repeating the calculations, changing the number of frequencies in <code>NLEnSteps</code> though this take about 10 times more computational time than the run with 12).

	As an exercise, you can repeat the ~~simulation~~ for e,g, a time of <code>35 fs</code> nand <code>60 fs</code>, that is setting <code>NLtime</code>. The comparison of the susceptibilities at different simulation times (as in the plot below) shows the importance of running long enough simulations. In too short simulations, the response signal coming from the switch-on of the electric field (delta-kick) is still of comparable intensities as the relevant second-order signal.		As an exercise, you can repeat the simulations for e,g, a time of <code>35 fs</code> nand <code>60 fs</code>, that is setting <code>NLtime</code> to the desired time. The comparison of the susceptibilities at different simulation times (as in the plot below) shows the importance of running long enough simulations. In too short simulations, the response signal coming from the switch-on of the electric field (delta-kick) is still of comparable intensities as the relevant second-order signal.

	[[File:02 SHG IP l.png\|600px\|center\|SHG of 2D-hBN with IPA. Effect of different simulation time]]		[[File:02 SHG IP l.png\|600px\|center\|SHG of 2D-hBN with IPA. Effect of different simulation time]]
Line 283:		Line 283:
	[[File:02 SHG SEX.png\|600px\|center\|SHG at the Hartree + Screened exchange from the current run with 12 frequencies (crosses) and a run with 112 frequencies (keeping the other parameters the same)]]		[[File:02 SHG SEX.png\|600px\|center\|SHG at the Hartree + Screened exchange from the current run with 12 frequencies (crosses) and a run with 112 frequencies (keeping the other parameters the same)]]

	We note differences between the two simulations~~. This~~ is due to the too short simulation time. Repeating the calculations with a longer time gives more accurate results.		We note some differences between the two simulations, this is due to the too short simulation time. Repeating the calculations with a longer time gives more accurate results.
	In addition, note that in the interest of time, all parameters are '''under converged'''. In principle, to calculate the collisions one should use the same parameters of a converged BSE calculation. As well, to eliminate spurious interaction with the periodic images, one should use the Coulomb cut-off as it was used for this system in the [[Quasi-particles of a 2D system\| tutorial on quasiparticle calculations]]. As a reference, a converged calculation for this system is in <ref name="Gruning2014">M. Grüning and C. Attaccalite, [https://journals.aps.org/prb/abstract/10.1103/PhysRevB.89.081102 Phys. Rev. B '''89''', 081102]</ref>.		In addition, note that in the interest of time, all parameters are '''under converged'''. In principle, to calculate the collisions one should use the same parameters of a converged BSE calculation. As well, to eliminate spurious interaction with the periodic images, one should use the Coulomb cut-off as it was used for this system in the [[Quasi-particles of a 2D system\| tutorial on quasiparticle calculations]]. As a reference, a converged calculation for this system is in <ref name="Gruning2014">M. Grüning and C. Attaccalite, [https://journals.aps.org/prb/abstract/10.1103/PhysRevB.89.081102 Phys. Rev. B '''89''', 081102]</ref>.

	= References =		= References =
	<references/>		<references/>