Hello sir! Thank you for the informative lecture. If possible, I would love to see a video on BTBT models or junction leakage simulations in the future. Thank you!
hello Professor, thanks for the very clear tutorial. I have followed every step of the way but to create 3d Gan/AlGaN FinFET but when I copy all my commands from the editor, and create a new project and import the cods to sed tool, it does not run/Stays running(blue) for a day and then it just doesn't finish. what do you think is the problem and if you would be able to demonstrate 3d structure + meshing and device as well that would be great. I couldn't find anything neither on line or anywhere that explains it
I guess you might have defined some very dense meshes. Your GaN probably is in the order of um. If you set 10nm in each direction, it will need to create 1 million mesh points. You probably want to have dense mesh in the 2DEG region only and in the direction perpendicular to the heterointerface. To check this, you can double click node explorer and you will see it is running. And then use a terminal and type "top" to check the task and probably you can find the mesh engine is still running and consumes quite some memory. Try to stop the job, comment out all mesh definitions, and run it one time first. It should finish fast. Then add meshing boxes one by one to pinpoint and solve the problem. For 2D structure, 10k mesh points are usually reasonable. For 3D, do not exceed 200k unless you really know why it is necessary.
@@quantum_computing I appreciate your response thank you vey much. But it works fine when I build the mesh using the structure editor. it runs and finishes and show me the mesh in visual with no issues. but when I copy the same code to gedit, then create a new project in swb, and add a new tool "sde" and import the file that I saved, that's when it doesn't finish even though it ran and finished within seconds when I clicked on build mesh in the structure editor. I am trying to make it work in swb so then I can create device and continue the rest in swb. (GaN and the whole device is in nm scale, Lch=100nm)
@@salimalmenshad6854 I see. Looks like you are not using batch mode. Since sde is opened, swb is waiting for sde to be closed. If you close it, it will finish. Ideally, you want to put the meshing command in the script also (the command after you click mesh in sde GUI). Then make sde in batch mode in swb (right click and choose properties) so the whole process will be automatic. Thanks!
Thank you very much for this excellent tutorial! I'm currently doing project on Boron CVD and I wonder if it is possible to model pure-B layer in Sentaurus Structure Editor? I can't find pure Boron in the material menu.
@user-wl6sr4mf6k, you just do not put any doping. When it is undoped, it will generate an equal amount of electrons and holes. E.g. at 300K, it is about 1E10 per cm^3 for Si. SDevice will simulate this correctly. If you probe the undoped region, you will see it has about 1E10 per cm^3 e and h for Si at 300k.
Hi, how can I create regions with custom materials - where I can tweak the bandgap, electron affinity, recombination times, etc? The material I need is not a pre-defined material in Sentaurus TCAD. I'd apprecicate an answer, thanks!
You need to create a new material in datexcodes.txt. And then define those parameters in the parameter files. There are some discussions in the comments of this video ruclips.net/video/OjYjAaUU3EA/видео.html
@user-wl6sr4mf6k, you just do not put any doping. When it is undoped, it will generate an equal amount of electrons and holes. E.g. at 300K, it is about 1E10 per cm^3 for Si. SDevice will simulate this correctly. If you probe the undoped region, you will see it has about 1E10 per cm^3 e and h for Si at 300k.
@@quantum_computing Thank you so much for your help! for example, my intrinsic InP wafer has a carrier concentration with 1e+16, how to define this value? how to write the code? Thank you!
@@Hengliu-y1b You need to determine what is the source of the 1e16 carriers. If they are from the impurities, then you need to add doping. It doesn't look like it is intrinsic as InP has a bandgap of 1.35eV.
@@Hengliu-y1b This involves a lot of physics. Just to get the carrier, you can model the TD as a doped region. But this won't capture how the traps affect the electrostatics. Unfortunately, I do not have the bandwidth to discuss too deep into this. We had some papers before. But I do not believe you need to go that deep. In summary, there is no simple solution and you need to model it based on your need. I suggest you contact Synopsys support to see if they have new models for this. Here are some papers: 1. Victor Moroz, Hiu Yung Wong, Munkang Choi, Nelson Braga, R. V. Mickevicius, Yuhao Zhang, Thomas Palacios, “The Impact of Defects on GaN Device Behavior: Modeling Dislocations, Traps, and Pits”, ECS J. Solid State Sci. Technol. 2016, volume 5, issue 4, P3142-P3148 2. C55. 2. Victor Moroz, Hiu Yung Wong, and Munkang Choi, “Modeling Extended Defects in Semiconductor Devices, “ECS PRiME 2016. ECS Transactions 75 (4), 143-152 (INVITED PAPER) 3. Y. Zhang*, H.-Y. Wong*, M. Sun, S. Joglekar, N. A., R. V. Mickevicius, and T. Palacios, "Design space and origin of off-state leakage in GaN vertical power diodes," 2015 IEEE International Electron Devices Meeting (IEDM), Washington, DC, USA, 2015, pp. 35.1.1-35.1.4.
I actually do not have the script. But if you follow the video, you should be able to reproduce it because that script was created by copying the tutorial example.
Hello sir! Thank you for the informative lecture.
If possible, I would love to see a video on BTBT models or junction leakage simulations in the future. Thank you!
When I find time, I will try to post some. Thanks!
thx from Korea!
@user-wc1qo7mq5u annyeonghaseyo!
hello Professor, thanks for the very clear tutorial. I have followed every step of the way but to create 3d Gan/AlGaN FinFET but when I copy all my commands from the editor, and create a new project and import the cods to sed tool, it does not run/Stays running(blue) for a day and then it just doesn't finish. what do you think is the problem and if you would be able to demonstrate 3d structure + meshing and device as well that would be great. I couldn't find anything neither on line or anywhere that explains it
I guess you might have defined some very dense meshes. Your GaN probably is in the order of um. If you set 10nm in each direction, it will need to create 1 million mesh points. You probably want to have dense mesh in the 2DEG region only and in the direction perpendicular to the heterointerface. To check this, you can double click node explorer and you will see it is running. And then use a terminal and type "top" to check the task and probably you can find the mesh engine is still running and consumes quite some memory. Try to stop the job, comment out all mesh definitions, and run it one time first. It should finish fast. Then add meshing boxes one by one to pinpoint and solve the problem. For 2D structure, 10k mesh points are usually reasonable. For 3D, do not exceed 200k unless you really know why it is necessary.
@@quantum_computing I appreciate your response thank you vey much. But it works fine when I build the mesh using the structure editor. it runs and finishes and show me the mesh in visual with no issues. but when I copy the same code to gedit, then create a new project in swb, and add a new tool "sde" and import the file that I saved, that's when it doesn't finish even though it ran and finished within seconds when I clicked on build mesh in the structure editor. I am trying to make it work in swb so then I can create device and continue the rest in swb. (GaN and the whole device is in nm scale, Lch=100nm)
@@salimalmenshad6854 I see. Looks like you are not using batch mode. Since sde is opened, swb is waiting for sde to be closed. If you close it, it will finish. Ideally, you want to put the meshing command in the script also (the command after you click mesh in sde GUI). Then make sde in batch mode in swb (right click and choose properties) so the whole process will be automatic. Thanks!
See here for how to change to batch mode ruclips.net/video/e2Lt3y13oMw/видео.htmlfeature=shared&t=7048
Thank you very much for this excellent tutorial! I'm currently doing project on Boron CVD and I wonder if it is possible to model pure-B layer in Sentaurus Structure Editor? I can't find pure Boron in the material menu.
You need to define the material in datexcodes.txt and also the corresponding parameter file for device simulation.
hello sir, how to define the carrier concentration of the intrinsic semiconductor? it's not a N or P doping
@user-wl6sr4mf6k, you just do not put any doping. When it is undoped, it will generate an equal amount of electrons and holes. E.g. at 300K, it is about 1E10 per cm^3 for Si. SDevice will simulate this correctly. If you probe the undoped region, you will see it has about 1E10 per cm^3 e and h for Si at 300k.
Hi, how can I create regions with custom materials - where I can tweak the bandgap, electron affinity, recombination times, etc?
The material I need is not a pre-defined material in Sentaurus TCAD.
I'd apprecicate an answer, thanks!
You need to create a new material in datexcodes.txt. And then define those parameters in the parameter files. There are some discussions in the comments of this video ruclips.net/video/OjYjAaUU3EA/видео.html
Hello sir, i wanna ask you about re-open the 'Scheme command' window. I close the window in mistake, but i cannot find re-open it.
@whope110 try to right click on any of the tool icons, and you can check "Scheme command" (first one) to get it back. Let me know if it works.
I made a video to make it clearer: ruclips.net/video/0ZNwo9bDGdg/видео.htmlfeature=shared
Thanks sir. Really appreciate it. Yesterday i try to find scheme command for 2 hours... also thanks you for making the video😊
Glad that it works, @@whope110
hello sir, how to define the carrier concentration of a intrinsic semiconductor? it's not N or P doping
@user-wl6sr4mf6k, you just do not put any doping. When it is undoped, it will generate an equal amount of electrons and holes. E.g. at 300K, it is about 1E10 per cm^3 for Si. SDevice will simulate this correctly. If you probe the undoped region, you will see it has about 1E10 per cm^3 e and h for Si at 300k.
@@quantum_computing Thank you so much for your help! for example, my intrinsic InP wafer has a carrier concentration with 1e+16, how to define this value? how to write the code? Thank you!
@@Hengliu-y1b You need to determine what is the source of the 1e16 carriers. If they are from the impurities, then you need to add doping. It doesn't look like it is intrinsic as InP has a bandgap of 1.35eV.
@@quantum_computing it is come from dislocation density in the InP, how to define this carrier concentration? Thank you so much.
@@Hengliu-y1b This involves a lot of physics. Just to get the carrier, you can model the TD as a doped region. But this won't capture how the traps affect the electrostatics. Unfortunately, I do not have the bandwidth to discuss too deep into this. We had some papers before. But I do not believe you need to go that deep. In summary, there is no simple solution and you need to model it based on your need. I suggest you contact Synopsys support to see if they have new models for this. Here are some papers:
1. Victor Moroz, Hiu Yung Wong, Munkang Choi, Nelson Braga, R. V. Mickevicius, Yuhao Zhang, Thomas Palacios, “The Impact of Defects on GaN Device Behavior: Modeling Dislocations, Traps, and Pits”, ECS J. Solid State Sci. Technol. 2016, volume 5, issue 4, P3142-P3148 2. C55.
2. Victor Moroz, Hiu Yung Wong, and Munkang Choi, “Modeling Extended Defects in Semiconductor Devices, “ECS PRiME 2016. ECS Transactions 75 (4), 143-152 (INVITED PAPER)
3. Y. Zhang*, H.-Y. Wong*, M. Sun, S. Joglekar, N. A., R. V. Mickevicius, and T. Palacios, "Design space and origin of off-state leakage in GaN vertical power diodes," 2015 IEEE International Electron Devices Meeting (IEDM), Washington, DC, USA, 2015, pp. 35.1.1-35.1.4.
would you mind sharing here the script?
I actually do not have the script. But if you follow the video, you should be able to reproduce it because that script was created by copying the tutorial example.