My First TFET

[esp]

I made a TFET type device, by first creating doped p and n regions for source and drain (selective doping), adding top and bottom gates, and doing some calcs .... i get the IV curve and transmission spectrum (at Vg=0) attached ... .there is obviously something wrong here, so i have a couple questions ... (note that this was my first stab at it so i am sure i have done something wrong in geometry, etc, but regardless, i got some results so i have questions now) ...

1) regarding the geometry, should the gate typically overhang the channel significantly perpendicular to the transport direction?  i tried this before and it took 3x longer to calculate ... but if it is more realistic i should do it

2) when making source and drain, because of how atk works, the electrode region is copied into the bulk region .... now my gate(s) should not be over that copied part, it should be only over the channel, but then there is a region which is not electrode, and not channel ... i do not want to put the gate over that because that wouldnt make sense ... so what is typical?  does my question make sense?

3)  I took a look at the transmission spectra used to make the IV plots, and i see in fact there is a strange curve there that corresponds to the IV curve .. my guess is that although my bandstructure is how i wanted it, that the doping affected the transmission spectrum .. i need to be able to dope to my liking without this effect on the IV curve ... does anyone have good references on graphene device doping? most papers just say they doped with X dopants per atom

[esp]

and note this is obviously not a good TFET, but i am just learning

[Anirban Basak]

Hi esp, i think i will try to answer some of your questions.

1) In this case you have to consider the fringing fields from your gate. If you use high-k dielectric i guess the electric field will not fringe too much and you can use smaller gate.

2) You should remember that the voltage drop you apply actually appears across the scatterer (i.e. the left source, the channel and the right drain region) all encapsulated in the middle box as a single structure. The electrode you use in ATK does not get voltage drop and is only there to provide self-energy for the left side of source and the right side of drain. In simple word if in your device structure if you ignore the left and right electrode box and consider the remaining structure, that is where the voltage drops. 

for 3 I dont have much ideas to tell hehe 

[Nordland]

1) regarding the geometry, should the gate typically overhang the channel significantly perpendicular to the transport direction?  i tried this before and it took 3x longer to calculate ... but if it is more realistic i should do it

It depends a little on what you want exactly. If you want to model the FET effect in your system, you could go with a more narrow system, but if you want to model a real experiment, then it might be better with wider dielectric regions. Again it comes to the model, because it is a model after all. The dielectric regions are in the real world made of atoms, with which the carbon atoms would interact, so it is only an approximation and I will have a hard time trying to say something general

2) when making source and drain, because of how atk works, the electrode region is copied into the bulk region .... now my gate(s) should not be over that copied part, it should be only over the channel, but then there is a region which is not electrode, and not channel ... i do not want to put the gate over that because that wouldnt make sense ... so what is typical?  does my question make sense?

I think I understand. I would the system as you do.

3)  I took a look at the transmission spectra used to make the IV plots, and i see in fact there is a strange curve there that corresponds to the IV curve .. my guess is that although my bandstructure is how i wanted it, that the doping affected the transmission spectrum .. i need to be able to dope to my liking without this effect on the IV curve ... does anyone have good references on graphene device doping? most papers just say they doped with X dopants per atom

Here I need a little more detailed information to help you. Are you sure you are running self-consistent?

[esp]

yes i am always running self consistently

[Nordland]

I am a little uncertain of what is the problem with question 3 ? Care to elaborate/clarify ?

[Anders Blom]

I like your bands plot! Is it just a straight-up band structure for the central region or did you project it on anything?

On 3) do you mean that you would like to dope the electrodes but not the central region? And/or you want some effective doping rather than inserting explicit dopant atoms (since the doping concentration in some cases becomes artificially large)? I also don't understand the "unwanted effect" you see on T(E), but it may be related to these two questions...?

[esp]

Sorry ... well i suppose question 3 was too general, and so that is why I simply asked if anyone had ant good reference [papers] on graphene doping ... it seems graphene, especially with very small dimensions, is extremely sensitive to where the dopants are ... for example putting a Boron atom in one place is totally bandstructure then putting it in the next position ... in one place there is a gap of 0.5eV, in another almost no gap .. very strange .. so i suppose i don't know what i am doing regarding doping of graphene .. thats why i need references ... it seems you cannot just dope X% ... about regions, no i made a script that dopes source,drain, or channel, with different methods i have tried (edge doping, random, selective, etc)