Lecture26

Review of HLI (week 7) and Eber Moll Model (week 7).

C-V plot of the structure

In the small signal current, we have not DC current, but the current is influenced only by the charge
and the discharge of the capacitors, in the difference bias condition of operation. We have analyzed
3 situations:
- LF (low frequency) in which the C is equal to COX because we make the AC measurement
during a time sufficiently long to reach the equilibrium (xd = xdmax);

- HF (high frequency) in which the C remain constant at the minimum value because the AC
signal is faster than the time required to form the inversion layer (xd = xdmax);
- if in addition to having a fast AC signal to measurement the capacitance C we also vary the bias
voltage in a fast way, the system goes into a DD (deep depletion condition) where in a transient
way xd > xdmax; at the same point increasing further the value of the bias voltage, the depletion
region becomes large enough that the generation process increases to the point that the HF and DD
curve merge (si uniscono).

Analytical expression for charge concentration on the MOS structure

We take in account two different situation; first of all, we analyze the capacitive behavior in the
thermal equilibrium, then we’ll analyze it out from thermal equilibrium, applying an external bias
voltage.
We have in thermal equilibrium between bulk and surface in the silicon, so the Fermi level is
horizontal; we define the potential in the silicon as Formula p.141
Take some annotations:
- Ef is constant due to the thermal equilibrium; - Ei(x) depends by ‘x’ because the potential changes
point to point; - Φp is the potential in the bulk (Φp < 0, being a p-type substrate); - Φ(0) = Φs is
the surface potential; - Φn is the free-electron charge density in the inversion layer; - Φd = Φn –
Φs depletion-charge density; - Ns and Np surface free-carrier density; - p*n = ni^2 for the thermal
equilibrium. We can see that in the bulk Formula p.142
We can also express the concentration at the surface Formula p.142
Plot p.140

Table.

Plot p.143

We make the approximation that after the bands in the semiconductor bends (piegare) by |2 Φp| the
bands do not bend anymore. After |2 Φp| banding the concentration of electron on the surface is the
same as that of holes in the bulk. Remember that 60 mV change in Φs gives a factor of 10 change in
the surface concetration. We call Xdmax and Qmax the width of the depletion region and the
charge in the semiconductor after |2 Φp| → band-bending (when the bands stop bending); we have
Formulas p.143
Qmax is charge per unity area in the depletion region at maximum bending.
 VG – VFB
Surface charge Surface charge
Voltage applied at Qs
condition density majority
MOS structure
Negative
Negative Accumulation Ps > Na
|Qs| > |Qp|
In this case we have the density of carrier in the surface is higher than the concertation in the
bulk. In this case we have a p-type material in the surface much high doping then in the bulk.
Negative Flat-Band
0 Ps = Na
|Qs| = |Qp| Neutral
The potential in the surface is still negative. In this case we have no electric field, no drop
voltage.
Positive Negative
Depletion Ni < Ps < Na
+ |Qs| < |Qp|
We apply a small positive voltage and the potential in the surface still remains negative. The
intrinsic Fermi level goes above to the Fermi level, with a distance than the distance that there is
in the bulk. It is the beginning with the depletion.
Positive
0 Intrinsic Ns = Ps = Ni
++
At this voltage, the intrinsic Fermi level touches the Fermi level and we have no charge density in
the surface. The surface is an undoped material.
Positive Positive
Weak inversion Ni < Ns < Na
+++ |Qs| < |Qp|
Now the Qs become positive; now in the surface we have a higher concentration of N carrier than
P carrier. The majority population is changed.
Positive Positive Onset of strong
Ns = Na
++++ |Qs| = |Qp| inversion
Here we have reversed the concentration and the concentration of electrons in the surface is equal
to the concentration of holes in the bulk; a sort of p-n junction.
Positive Positive
Strong inversion Ns > Na
+++++ |Qs| > |Qp|
We have a Strong inversion; the concentration of electrons is too higher than the concertation of
holes in the bulk.
Three terminals MOS structure
Plot p.145
Put a n+ contact in the proximity of the inversion layer. This allow the injection of electrons to
build the inversion layer very quickly without waiting for generation in the depletion region.
The n+ contact allows to bias the surface induced p-n junction with a reverse bias (higher voltage
to n contact and lower voltage to bulk). This takes us out of the thermal equilibrium condition
between bulk and surface → the Fermi level splits into two quasi Fermi level since we have a
leakage current from the bulk to the surface; this current is small but sufficient to put the system
out from the thermal equilibrium. Considering the plot shows Plot p.145
Reducing Vb and keeping constant Vg and Vc, the depletion region between surface and bulk
becomes wider and also that under the gate. We make Vb more negative keeping all the other
voltages constant and, since the voltage at the silicon surface is given by Vc, also this stays
constant. The voltage on the oxide also stay constant, this makes constant both the charge on
the metal and the total charge on the semiconductor. Vb more negative extends the depletion
region below the ‘channel’ (i.e. inversion layer) increasing the charge in the depletion region.
More charge in the depletion region and constant charge in the semiconductor has the effect of
reducing the charge on the channel (mobile charge); we call this Body effect.
Resume, modulating the bulk bias voltage we induced an indirect modulation in the channel, in
particular, in the stored charges in the channel.
Since we have no more thermal equilibrium between surface and bulk, the Fermi level splits giving
a quasi-Fermi level EFP in the bulk and a quasi-Fermi EFN in the surface separated by an energy =
q*(Vc – Vb), this corresponds to the follow band diagram Plot p.146 . To get into strong
inversion the surface potential must be Φs = |Φp| + Vc – Vb instead of only |Φp|, therefore the total
band-bending from bulk to surface is |2Φp| +Vc – Vb instead of |2Φp| it follows that Formula p.146

Since for a voltage between gate and bulk = VFB there is no charge in the structure. The effective
voltage between gate and bulk that creates a charge in the structure is Vg – Vb – VFB. This is also
the effective voltage that creates a voltage drop on the oxide and on the silicon → Vg – Vb – VFB =
Vox + Φs – Φp (Vox is drop on oxide and Φs – Φp drop on silicon).
Formula p.146 (field constant on the oxide); Formula p.146 (Vg – Vb – VFB - Φs + Φp); for the
continuity of D at the silicon surface Formula p.146