I am trying to extract electron mobility values from FET transfer curves using the equation µ = [L / W*Ci*Vd]*[∂ISD / ∂VGg] where L and W are the respective length and width of the channel and Ci is the gate channel capacitance per unit area. I am using Si wafers with a 300 angstrom SiO2 layer as the dielectric. Should I be using a known value for this or should I be measuring it on each device? If I measure capacitance, should I be measuring a single value that I then divide by the area of my contact pads to the surface?
(Note, this equation comes from Podzorov, V. et al. Critical Assessment of Charge Mobility Extraction in FETs, Nature Materials 17, 2-7 (2017))
Dear Danny,
when measuring the dielectric layer capacitance and also when applying the formula you mentioned to extract the mobility, you should reasonably sure that the hypotheses for the validity of the models you are using are met. For instance: what W and L are you considering? Are fringing fields negligible? What kind of de-embedding procedure are you using to measure the dielectric capacitance?
Having said that, I believe that the best way to measure the dielectric layer capacitance is to measure it under positive gate bias and with drain and source shorted together, so that to operate the resulting Gate-oxide-bulk capacitor either in inversion or in enhancement mode. For higher accuracy you can interpolate the measured capacitance under different biases and to derive Ci as the asymptotic value.
Hi Danny,
First of all, are you sure that your SiO2 is 300 angstrom thick? I suspect you meant that it is 300 nm thick! Also, to confirm that everything is OK, first calculate the capacitance per area (for 300 nm SiO2 you should get 11.5 nF/cm2) and then measure it for minimum 5-10 contact pads. You can measure just one but measuring 5-10 and taking the average will give you more precise value of the measured capacitance. Hint: to measure capacitance you can use commercial capacitance meters that work at 800 Hz or more sophisticated RLC meters which can measure capacitance at any chosen frequency (then choose 1 kHz or so).
I hope it helps!
Regards,
Leszek
- Badges
- Science topic
- Similar topics
- Phytochemistry
- Extracts