Temp vs vth / Temp vs id of MOS?

Hi all:

I simulate the Temp vs vth and Temp vs id of relationship of NMOS.

1. Temp = 27C, vth = 0.73V, id = 30uA when vgs = 1.6V, vds = 5V @TT
2. Based on item#1, I sweep temp to know Temp vs vth and id @ TT
a. when temp is increased, the id is decreased
b. when temp is increased, the vth is decreased too.<------------------------why
==>the result let me confuse.
Could you help me?
3. I change vgs = 1.6 -->1V, the result is difference:
a. when temp is increased, the id is increased, too<--------------------------why
b. when temp is increased, the vth is decreased.

My problems are:
1. why get difference relationship of Temp vs id when vgs is bigger than vth (vgs = 1V or 1.6V)?
2. why vth is decreased when temp is increased?

Thanks for your reply.

mpig

Hi,
You ask good questions

Answer for second question you can find here.
For first question i also would like to know answer...

- - - Updated - - -

Hi again,
Seems that i find answer on your first question.
There are we have another effect - temperature dependence of carrier mobility (~1/T^2 for silicon).

Hi sarge and all:

I study the the answer for the second question,
I have a little un-understanding:
the vsb = 0 in my simulation, so I think there is no body effect,
does VT0 (threshold voltage for zero substrate bias) has temperature coefficient?
or Material has temperature coefficient?
(Material: NA (doping parameter), Ni (intrinsic doping parameter for substrate) ...)
---------------------------------------------------------------------------------
Hi sarge:
Could you discuss more clearly about "temperature dependence of carrier mobility"
when id curve are difference when vgs is difference?


If you know any information, please teach me.
Thanks for your reply.

mpig

Ok, here my point of view:
1. Temperature dependence of treshold voltage:
Even if Vbs = 0, Vth has a negative temperature coefficient (as you can see in the wiki equation), and, except Vbs, you have **broken link removed** - body effect factor

2. Temperature dependence of carrier mobility:
As we know:
Id ≈ u*W/L*Cox*(Vgs-Vth)^2

where u - carrier mobility, and it has a temperature dependence.
I think that for low values of Vgs, most significant role plays a second part of equation ( (Vgs-Vth)^2 ) and we have increasing of current.
For high values Vgs, changes of Vgs-Vth lower than changes of u and we have decreasing of current.