I have been doing surface potential modeling for Dual Metal Double Gate FinFET with 20nm channel length. We have designed mathematical model considering effects of source, drain as well as the junction of two metals. But our modeling and simulation result is varying by 0.1V. Also at low gate voltage (0.01V) the step at the junction is not prominent as it should be. So I need to know what kind of SCE are causing these problems and how to rectify them. Kindly find the surface potential graph due to modeling and simulation for different gate voltages.
Looking at the enclosed graphs I can see that they differ mostly in colors used to denote the two approaches, they are interchanged: what is green in one graph is blue in the second. So, maybe there is no problem here, no matter that applied voltages differ? Besides: what is the difference between simulation and modelling? None of those procedures seem to include any real measurements ...
Marek Wojciech Gutowski sir I am trying to design Dual Material Double Gate MOSFET and for that I have developed a mathematical model in order to analyze the surface potential characteristics of the said device. Furthermore I have designed and simulated the said device on SILVACO TCAD. If you look at any one of the two graphs enclosed you will see that there are two curves in the graph, one corresponding to the simulation result through silvaco and other corresponding to the mathematical model that I had develpoed.
My question here is that why the two curves are not matching exactly, and what kind of parasitic effect are causing this mismatch that I might have overlooked while developing the mathematical model?
For your convenience just have a look at the graph attached here, it has all the details written in it.
As for me, both curves share quite similar features. The difference seems to arise because two different materials are used, but they may be treated not on equal footing when simulating with first or second tool. Different materials may mean different geometry (width, thickness) but they are also most likely different because of their different carrier concentrations and/or different Fermi levels (chemical potential, if you prefer), even if both are metallic.
Besides: you know what is inside of your mathematical model (since you are its developer), but you probably don't know the inner details of SILVACO TCAD. Is it a reliable tool, working well on other, preferably more obvious, examples with known outcome? Does it work on exactly the same set of input data as are needed by your model? There might be some simplifications in SILVACO TCAD, maybe even a subtle bug manifesting itself only when working in nanoscale? Are both approaches classical or of quantum mechanical kind? At this device size both methods can generate varying results, that's natural. Similarly, the results will vary, if one method operates on ions and the other one on finite elements.
Sir actually SILVACO TCAD is used as the simulator in many research papers, and this is why I have chosen to use it as the simulator. And moreover as you said I don't know the inner details of SILVACO TCAD, but as for my mathematical model I have considered only classical mechanics approach, as of now I have not considered quantum mechanical effects but certainly that is something I will now try to implement in my mathematical model.
As for different materials I am using it is true that they will have different Fermi levels (and that is the very reason why I am using different materials) but the metal contacts formed on the gate using both materials have same dimensions. All that I want to understand is that what kind of behavior will junctions between two material will show in the p-type substrate.
Please find attached the potential contour in the device as shown by the simulator SILVACO TCAD.
I am using material 1 that has work-function=4.8 and material 2 that has work-function=4.5
Beautiful picture, indeed. Looks like FEM (Finite Element Method) is at work. If so, did you try to vary the mesh size?
Indeed sir I have varied mesh sizing once but yes I didn't have any idea that can be a factor in getting a proper curve in surface potential graph. I will try that.
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication