HSPICE: Vth variation with time

[kmrkushal]

Hi all,

I am trying to model NBTI effects in PMOS(Just starting with Basic model of change of vth = k*((t)^(1/6))

For this I want to change Vth at each Time step of .tran evaluation.
I tried the following:
.param k = '0.1/((3*365*24*60*60)^(1/6))'
.param vth = '-0.5808 - (k*((TIME)^(1/6)))'
m2 vd 0 vd vd pmos L=40nm W=500nm
v2 vd 0 1.0V
.tran 1s 1e8
.probe I(m2) lv9(m2)

I have change in model file
+VTH0 = vth

But nothing is happening and Vth value is remaining constant

Would be of much help to me if anybody can point to a solution


Regards
Kushal

 

[micpro]

Have you looked into MOSRA (MOS reliability analysis for Hspice)?

 

[kmrkushal]

Have you looked into MOSRA (MOS reliability analysis for Hspice)?

I tried it once but it was not doing anything specific.
Actually I dont know whether for MOSRA do we require a special license and if my University has it or not?
Also the documentation on MOSRA was not helping me a lot.

I was able to solve the problem by using DELVTO in MOS definitions.

 

[timof]

A few things to keep in mind, when simulating NBTI:

1. the power factor (1/6 in your equation) is something that is not very well defined or known, it can vary anywhere between ~0.1 and ~0.5 depending on NBTI characterization methodology, process, etc. A "classical" value is 0.25, but other values are observed equally frequently in measurements.

2. The degradation depends not on the total time, but time under stress - which may be different for different PMOS transistor instances.

3. Even identical devices, especially small-size devices, under identical stress conditions may exhibit different amount of degradation, due to the random/statistical nature of degradation (defects and interface states generation). This may be important in circuits sensitive to mismatch.

Max
----------