Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
Thanks Colbhaidh for the post. I now doing a project to study about the NBTI degradation issue on PMOSFET when DC or AC stress is apply on it. Do you have any idea on how to write the scrip using Silvaco software for the test
1. Apply the DC stress on the PMOS
2. Apply the AC stress on the PMOS
You can set up a simulation deck to model the DC stress and AC stress but this will not model NBTI - in other words, nothing will change in the simulation after the stress as Silvaco does not simulate what actually happens to cause the instability. For example, one mechanism of NBTI is the cause of interface traps from dangling bonds that are created during the stress when Si-H loses the H due to an impact event.
I do not think even the latest version of Silvaco can model this (and this is only a simple event of NBTI, there are others).
What you can do is fit a simulation deck to the post stress results. So you could introduce an interface trap density that simulates the effect of an NBTI result but you are doing this retrospectively to a result you already know. But I do not think this is what you are trying to do.
Perhaps the most recent Silvaco software does have something but I am not aware of it and would be surprised as leading edge research in NBTI cannot fully explain the instability (or the experts cannot agree!)
Can u assist me in writing the code for PMOS device stress with DC voltage and also the PMOS device stress with AC voltage? ( the step to set up a simulation deck to model the DC stress and AC stress)
I still cannot see the point in this.
If you set up a TCAD deck for a pmos device, then bias it under the NBTI stress conditions for DC ( all you would do is set Drain, Source, Body to Vdd and Gate to zero) set the simulation temperature (say to 100C) then run the simulation for the stress time, then re-run the sumulation at room temp to find the VT (and so the deltaVt).
Then repeat this for different stress times.
In the real world a plot of deltaVt against stress times would be a straight line with increasing deltaVt with time.
However if you run the above simulations, there will not be any difference in Vt !!! Silvaco does not simulate the breaking of the Si-H bond at the gate oxide-Si interface because it does not simulate any Si-H bonds at all - it is not in the physics of the model - i.e. it is not in the source code for the program. Silvaco is derived from FLOODS/FLOOPS and I have some familiarity with that C code (unless they have implemented something since 2007).
The density of Si-H bonds and the Si/SiO2 interface is very process dependent and cannot be easily simulated. The loss of H during stress and subsequent recapture of the H at the dangling bond during recovery is also highly process dependent.
Are you sure your Silvaco does this modelling? What version are you using?
Silvaco ATLAS is capable of modelling the depassivation process when the reaction driving agent is any combination of hot channel current, tunnelling current, oxide field or inversion layer hole capture. Quoted from webpage **broken link removed**
This what I found from Silvaco webpage. Is that mean that the latest version of Silvaco can model the NBTI degradation? Please advice.
In the Figure 3b from this webpage, the NBTI degradation simulated Threshold Voltage shift is plot in logarithm form, is that the voltage shifting is more clearly to see if plot in logarithm format?
The server for Silvaco software in my university is being down this few days, I can not find out what version of Silvaco I am using.
Interesting...,it seems my version may have this after all. You have taugfht me something I did not know...
Now remember that Silvaco is modelling the effect of depassivation at a rate you determine - so it is not simulating actual loss of hydrongen to give dangling bonds but is empirically modelling what would happen if de-passivation occurs under the rate you give it.
In a real situation, the number of passivated dangling bonds cannot be known as it is process dependent. Nor can the rate of depassivation under stress nor the re-passivation under recovery. So in real life, if you change the gate oxide pre-clean and gate oxidation process, you may change the performance to NBTI stress. If you did the same in you simulation of the gate oxide process nothing would change in the simulation.
However, you could use Silvaco as a means top back track from a known NBTI effect to guess what the actual rates and density of traps was.
So the best thing to do is just try it out. Start with DC and stress your pmos with a gate voltage of -Vdd or more at 100C for X hours after entering the same parameters given in the webpage you mentioned. If this shows a voltage threshold change then you can proceed from there. I do not have access to this version of Silvaco but there model is quite straight forward. I will try do something here.
Since the variation in Vt with NBTI stress in in the order or 10 - 100's of mV, then it is better to use log scale when showing differences.
Finally the server in my lab is become normal and I can use the Silvaco software.
But unfortunately, the version of silvaco software I am using is quite old, and I can not find the Heiman model which mention in this web page to model the NBTI model. Do you have any others way to model the Heiman model or have the parameter of Heiman Model so that I can type in manually to model the Heiman model?
My version of Silvaco is 1999 if I'm not mistaken. Neither Heiman model nor Hansch model appear in my Silvaco software.
This all the model that contain in my silvaco software.
Appreciate if you can advise what should I do next to get my project done. I had asking my lecturer to purchase the 2010 version but the possibility to purchase is low.
these are bipolar models. The models for nbti degradation, or any other degradation due to hydrogen detrapping at the si-sio2 interface will be mos models. Bipolar models do not have this kind of degradation.
Also does you version of silvaco have the C interpreter?
Hi Colbhaidh
Sad to said there is no license for C-interpreter in my silvaco software. And the model category that exist in my version only have mobility, recombination, statistic, impact ionisation, tunneling models, energy transport and heat flow equation. I cant find any degradation category in my silvaco software. Is I'm looking in the wrong menu? I am search under commands---models----models. Or the Hansch model or Heiman model need to enable it by typing any code? If so, what code needed to enable it? Or do my Silvaco version is not capable to simulate the NBTI effect? Please advice.
Does your softwrae have the INTERFACE statement? Within the statement you can set CHARGE + or - which would simulate the dangling bonds and shift the Vt.
I'm using the Hansch model by typing code devdeg.h and devdeg.e to activate the model. Does it is the right way to activate it?
I try the DC stress and I get the voltage shifting for the result. But when i try to d AC stress, the threshold voltage I getting still be the same with the DC stress.
What is the right code for applying AC stress to the P-MOSFET? The code i write for AC stress is :
" solve vdrain=2 ac freq=1e3 direct
solve vsource=2 ac freq=1e3 direct
solve vsubstrate=2 ac freq=1e3 direct
solve vgate=0 ac freq=1e3 direct"
hye, can someone help me on how to make pin photodiode by using athena. actually, before this i had create the photodiode by using atlas, and for me it is quite easy than athena.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
