MOS Capacitors
The ICS Instructions contain the basic procedures for making measurements. This
screen is intended to point out things unique to using the MOSC model file.
Device Selection
There are five types of capacitors in the mask set:
- metal over field oxide with guard rings (A),
- metal over gate oxide (B),
- metal over field oxide (C),
- metal over diffusion 1 (D) and
- metal over diffusion 2 (E).
Each type has been implemented with three different geometries:
- The square capacitor is 300 x 300 μm
- The round capacitor radius equals 150 μm
- The finger capacitor has a center region of 100 x 100 μm and twelve 20 x 100
μm fingers.
- Make the measurements on the round "B" capacitors.
You will determine the doping
level of the silicon from the capacitance measurements in the report. You will
also determine the oxide thickness (Gate Oxide in the case of the "B" capacitors.)
Probe Assignments
- SMU1 and CM(H) → probe1 → Round "B Capacitor" pad
- SMU2 and CM(L) → probe2 → Substrate (small square under the "B")
Theory
These are metal oxide semiconductor capacitors (MOSC), and the ece440 textbook has
a description of what the capacitance vs. voltage curves
should look like, as well
as some of the information that can be determined from them.
Here are some brief highlights:
The capacitance of the oxide is given by:
Cox = εA/d
where:
- ε = dielectric constant of insulator = ε0εr
- A = area of the device
- d = distance between the plates (dielectric thickness)
Under a particular bias polarity, a depletion layer will form in the silicon below
the oxide, adding another capacitor in series with the oxide capacitor. The differential
capacitance, Cd, of the semiconductor-space charge region is
Cd = (dielectric constant of silicon)*(Area)/(Thickness of depletion
region)
The total capacitance is
Ctot = (Cox * Cd)/(Cox+Cd)
Here is a typical C-V curve for a MOS capacitor on a grounded n-type substrate with
±10V applied to the aluminum contact on top of the oxide. The flat regions
corresponding to the oxide capacitance and the total capacitance at maximum depletion
width are evident.
It is possible to model the capacitors by separating the capacitance into center
and edge effects. The equations are similar to those used for modeling the p-n junction
capacitance.
The total capacitance per area is
Ctotal = P * Cedge + A * Carea,
where P equals the perimeter and A equals the area. Note that the units of Cedge
and Carea are pF/μm and pF/μm2, respectively. By using two
of the capacitors, it is possible to solve for P and A.
|