The Two-terminal

MOS structure
BY DR. KUMARI NIBHA PRIYADARSHANI
 The Two-terminal MOS structure
Two-terminal MOS structure, often referred to as a MOS capacitor.

Fig 2.1: A two terminal MOS structure.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 2
Mcandrew, 3rd Edition (2011)
 The Flatband Voltage
In Fig 2.2 (a), We assume that somehow the same material is used to
connect the gate to the body. No charges are shown in the silicon, there
is no reason for the holes to pile up in any particular region.
In Fig 2.2 (b), From the gate material through the external connection
to the body. We encounter several contact potentials.
➢Sum of all contact potentials from gate, through external connection,
to bulk = 𝝋𝑺 − 𝝋𝑴
➢existence of a nonzero potential between the gate material and the
bulk causes net charges to appear on both sides of the oxide.
➢If, e.g., this potential is negative, the polarity of the charges will be
as shown.
External potential, 𝝋𝑴𝑺 = 𝝋𝑴 − 𝝋𝑺 must be applied for the net Fig 2.2: (a) MOS structure with gate, body, and short-circuiting
external connection made of same semiconductor material; (b) a
charges due to contact potential to disappear, as shown in Fig. Fig 2.2
MOS structure (with gate and body made of different materials,
(c). and a bottom metal contact) (c) the structure of “b” with a voltage
source such that the surface charge becomes zero.
Adapted from “Operation And Modeling of The
MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 3
Mcandrew, 3rd Edition (2011)
Another reason for net concentration of charges in the body in the absence of external bias are "parasitic" charge
that exists within the oxide as well as at the oxide-semiconductor interface. This charge consists of four parts:
➢An oxide fixed charge exists very close to the oxide-semiconductor interface due to the mechanisms of oxide
formation at the time such formation is completed. This charge is found to be rather independent of oxide
thickness, body doping type (nor p), and doping concentration.
➢A so-called oxide trapped charge can exist throughout the oxide, but usually close to either of its interfaces to
the body or the gate. This charge can be acquired through radiation, photoemission, or the injection of high-
energy carriers from the body.
➢A mobile ionic charge can exist within the oxide due to contamination by alkali ions (often sodium)
introduced by the environment during fabrication. This charge can move within the oxide under the presence of
an electric field.
➢An interface trap charge (also called fast surface-state charge) exists at the oxide-semiconductor interface. It
is caused by defects at that interface, which give rise to charge "traps"; these can exchange mobile carriers
with the semiconductor, acting as donors or acceptors.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 4
Mcandrew, 3rd Edition (2011)
➢ Q0 taken as the effective interface charge. This charge is almost
always positive for both p- and n-type substrates.
➢ In modern devices, Effective interface ion density is about 1010
to 1011 ions/cm2 , corresponding to an effective interface charge
density of 1.6 x 10-9 to 1.6 x 10-8 C/cm2.
➢ The charge Q0 will cause a total charge -Q0 to appear in the
system as demanded by charge neutrality. Part of that charge will
appear on the gate and part in the body.
➢ If all the required balancing charge -Q0 were provided on the
gate, no charge would be induced in the semiconductor.
➢ Since, at the gate and body ends of the oxide we must have
charge -Q0 and Q0, respectively, the potential drop across the
oxide, 𝜓𝑜𝑥 defined from the gate through the oxide to the body,
−𝑄𝑜 −𝑄′ 𝑜
must be equal to ൗ𝐶𝑜𝑥 = ൗ𝐶 ′ .
𝑜𝑥
➢ So,
Fig 2.2: (d) effect of effective interface chargeQ0; (e) the
structure in (d) with additional external bias so that the
surface charge becomes zero.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 5
Mcandrew, 3rd Edition (2011)
 Fig 2.3: (a) A
two-terminal
MOS
structure with
➢ The external voltage source keeps the Fermi levels
a p-type
of the gate (EFM) and of the body (EF) separated by
body; (b)
EF -EFM = q(𝝋𝑴 − 𝝋𝑺 ) and achieves the so-called
energy band
flatband condition.
diagram for
➢ A conduction band for the oxide is shown to be
the flatband
horizontal, as there is no field in the oxide under
condition,
our assumption of Q0 = 0.
assuming
➢ If now Q0 ≠ 0, then to keep the semiconductor
𝜑𝑀𝑆 > 0 and
bands flat, one has to adjust the external voltage to
𝑄′ 𝑜 = 0; (c)
the value VGB = VFB .
➢ The conduction band in the oxide will not be energy band
horizontal because there will be a nonzero oxide diagram for
field and a nonzero potential across it. the flatband
condition.
assuming
𝜑𝑀𝑆 > 0 and
𝑄′ 𝑜 ≠ 0.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 6
Mcandrew, 3rd Edition (2011)
 Potential Balance and Charge Balance
➢When VGB assumes values different from the flatband voltage
VFB will cause charges to appear in the semiconductor.
➢Practically, all of these charges will be contained within a region
adjacent to the top surface of the semiconductor, which is shaded.
Outside this region, the substrate is practically neutral.
➢The surface potential 𝜓𝑠 , is the total potential drop across the
region, defined from the surface to a point in the bulk outside that
region.
The potential drops are encountered in the loop:
o The voltage of the external source VGB·
o The potential drop across the oxide 𝜓𝑜𝑥 · Fig 2.4: (a) A p-body, two-terminal MOS structure under
o The surface potential 𝜓𝑠 general gate-body bias; (b) potential distribution, assuming
o Several contact potentials. Their sum, when going the gate, the body-metal contact, and the external wires are
clockwise, is 𝜑𝑀𝑆 , all made of the same material. The special case of 𝜓𝑠 > 0
Applying KVL, 𝑉𝐺𝐵 = 𝜓𝑜𝑥 + 𝜓𝑠 + 𝜙𝑀𝑆 → Potential balance equation has been assumed in drawing this plot.
Adapted from “Operation And Modeling of The
MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 7
Mcandrew, 3rd Edition (2011)
➢ 𝜑𝑀𝑆 is a known constant; therefore, any changes in VGB must be balanced by changes in 𝜓𝑜𝑥 and 𝜓𝑠 :

➢ The charges in the system encountered:

o The charge on the gate QG.
o The effective interface charge QO
o The charge in the semiconductor under the oxide QC.
These charges must balance one another for overall charge neutrality in the system:
Charge balance equation
In terms of charge per unit area:

The equivalent interface charge Q’O is for the present assumed fixed:

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 8
Mcandrew, 3rd Edition (2011)
 Effect of Gate-Body Voltage on Surface Condition
FLATBAND CONDITION ACCUMULATION

➢ The negative change of VGB will cause a negative

change in Q’G. Which, must be balanced by a
positive change in Q’C .
➢ Thus, holes will accumulate at the surface to
provide a net positive charge.
➢ The negative change in VGB causes negative
changes in 𝜓𝑜𝑥 and 𝜓𝑠 .

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 9
Mcandrew, 3rd Edition (2011)
 DEPLETION and INVERSION

➢ The positive change of VGB will cause a positive change in Q’G. Which, must be balanced by
a negative change in Q’C .
➢ The positive change in VGB causes positive changes in 𝜓𝑜𝑥 and 𝜓𝑠 .

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 10
Mcandrew, 3rd Edition (2011)
 DEPLETION INVERSION
➢For VGB not much higher than VFB , the positive ➢As VGB is increased further, more acceptor atoms are
potential at the surface with respect to the body uncovered, and 𝜓𝑠 , becomes sufficiently positive to
attract a significant number of free electrons to the
will simply drive holes away from the surface, surface; each of these electrons will also contribute a
leaving it depleted. The hole density will keep charge -q to QC.
decreasing well below the doping concentration
value NA. ➢These electrons come from the relatively slow
process of electron-hole generation in the depletion
➢For practical purposes, then, the charge QC is due region, caused by the thermal vibration of the lattice
to the uncovered acceptor atoms, each of which (assuming no radiation is present). Eventually, with a
sufficiently high VGB the density of electrons will
contributes a charge -q, and we can assume the exceed that of holes at the surface.
presence of a "depletion region" as in the p side of
the pn junction. ➢This is a situation opposite from that normally
expected in a p-type material; we now have
surface inversion.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 11
Mcandrew, 3rd Edition (2011)
 Fig 2.5: A MOS two-terminal structure in (a) accumulation; (b) depletion; and (c) inversion.
Adapted from “Operation And Modeling of The
MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 12
Mcandrew, 3rd Edition (2011)
 Fig 2.6: Energy band diagrams for a two-terminal MOS structure with a p-type body, assuming 𝜑𝑀𝑆 = 0 and 𝑄′ 𝑂 = 0, for various values of
VGB· (a) Flatband condition; (b) accumulation; (c) depletion; (d) inversion.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 13
Mcandrew, 3rd Edition (2011)
 Carrier Concentration
➢The oxide is assumed to be thick enough to block the flow of current, and thus equilibrium is maintained in
the substrate.
➢Deep in the neutral bulk outside the depletion region, the values of the hole and electron concentrations will
be p0 and n0 .

− s t − s t  s t ni 2  s t
psurface = p0 e  N Ae and nsurface = n0 e = e
p0
p0 2 ( s − 2F )
 N A e(
−F t t  s − 2F ) t
Also ni = pi = p0 e  nsurface = e
p0

and nsurface  N A e (
 s − 2F ) t
psurface  N A e − s t

Now as VGB   s  psurface  and nsurface 

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 14
Mcandrew, 3rd Edition (2011)
➢ At 𝜓𝑠 = 𝜑𝐹 , nsurface becomes equal to the intrinsic
concentration nsurface = psurface = ni·
➢ Ei bends just enough to touch EF .
➢ nsurface will be nonzero even in depletion but will be much
smaller than ni, even for 𝜓𝑠 , smaller than 𝜑𝐹 by only a few
𝜑𝑡 .
➢ With increasing 𝜓𝑠 , above 𝜑𝐹 , nsurface increases drastically,
and at 𝜓𝑠 = 2 𝜑𝐹 we have nsurface = p0 ≈ NA .

Fig 2.7: Electron concentration at the surface vs. surface potential.

Linear axes are used.
Adapted from “Operation And Modeling of The
MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 15
Mcandrew, 3rd Edition (2011)
 General Analysis
➢A general analysis is possible through which one can determine 𝜓𝑠 and Q’C for any value of VGB .
Fig. 2.4

➢Consider a point of ordinate y in the body (Fig. 2.4), and let 𝜓(y) be the potential there referenced to a point
deep in the bulk.
We will have, for the electron and hole concentration at y,

For the presence of both acceptor and donar atoms* *Ifnot all ions are ionized (e.g., at very
low temperatures), NA and ND should
be replaced by the densities of ionized
We can write poisons equation as: acceptor and donor atoms, respectively.

➢Multiply both sides of this equation by 2(d𝜓/dy); the resulting left-hand side can be recognized as (d/dy)
(d𝜓/dy)2. Replace y by a dummy variable ŷ and integrate from a point deep in the bulk (theoretically at
infinity, where 𝜓 = 0 and d𝜓/dy = 0) to a point y.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 16
Mcandrew, 3rd Edition (2011)
➢ Solve for d𝜓/dy at point y, and ℰ(y) = -d 𝜓/dy.

where, 𝜓 = 𝜓(y) and sgn(𝜓𝑠 ) is the "signum" function, defined as + 1 for 𝜓𝑠 > 0, zero for 𝜓𝑠 = 0, and -1 for
𝜓𝑠 < 0.
➢ Taking point 1 at the surface and point 2 deep in the bulk, where ℰ = 0.
This gives ℰsurface = Q’C/𝜖𝑠. Evaluating ℰsurface with 𝜓 = 𝜓𝑠 , we obtain

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 17
Mcandrew, 3rd Edition (2011)
➢ when 𝜓𝑠 = 0, the body is neutral everywhere; this means

Using this, NA-ND can be eliminated

We can write no and po as

➢ Thus, we obtain:

Fig 2.8: Total semiconductor charge vs. surface

potential (solid line) and constraint imposed by the rest
of the system (broken line).

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 18
Mcandrew, 3rd Edition (2011)
➢ The charge per unit area above the oxide, Q’G, can be simply related to the potential across the oxide, 𝜓𝑜𝑥 , and the
oxide capacitance per unit area, Cox,

Four equations that completely characterize the MOS structure under our assumptions.
o Potential balance
o Charge balance
o The semiconductor charge-potential relation
o The gate charge-potential relation

➢ The quantities 𝜑𝑀𝑆 , Q’O, C’ox, and NA are assumed known for a given fabrication process. For a given VGB, we can
solve these four equations to determine 𝜓𝑜𝑥 , 𝜓𝑠 , Q’G, and Q’C.
➢ Solve the charge balance equation for Q’C, and in the result we use the gate charge-potential relation.

Fig 2.8

➢ In principle Q’C and 𝜓𝑠 ., are the solution of the system for a given VGB as in in Fig 2.8 using a broken line for the
given VGB·

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 19
Mcandrew, 3rd Edition (2011)
 Using Q’C (𝜓𝑠 ):

Where,

Where, 2𝑞𝜖𝑠 = 5.8×10-16 F cm-1/2Vl/2

➢ 𝛾 is called the body effect coefficient,

➢ The above equation can be transformed as:

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 20
Mcandrew, 3rd Edition (2011)
 Fig 2.9: Surface Potential vs. VGB –VFB . Fig 2.10: Total semiconductor charge vs. VGB –VFB .

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 21
Mcandrew, 3rd Edition (2011)
 Accumulation and Depletion
General equation

➢In accumulation and depletion region, the contribution of electrons can be neglected. Also ND = 0.

➢Deep in Accumulation: Deep in accumulation (for 𝜓𝑠 below zero by several 𝜑𝑡 ), exponential term
dominates and the other terms can be neglected

➢Deep in Depletion: Deep in depletion (for 𝜓𝑠 above zero by several 𝜑𝑡 ) but less than 𝜑𝐹 . The exponential
term becomes negligible; if we also neglect 𝜑𝑡 .

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 22
Mcandrew, 3rd Edition (2011)
➢ If this equation is substituted in VGB

➢ Denoting the solution by 𝜓𝑠𝑎

𝜓𝑠𝑎 is the surface potential obtained deep in depletion.

predicts the surface potential accurately

➢ It predicts the surface potential accurately in deep depletion and even in weak inversion. However, in moderate
and strong inversion 𝜓𝑠 can be very different from 𝜓𝑠𝑎 because in those regions there exists a considerable
inversion layer charge that has not been taken into account in developing this equation.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 23
Mcandrew, 3rd Edition (2011)
 Inversion
➢ In Inversion, hole concentration is negligible. Also, ND = 0.

The total charge (per unit area) below the oxide is the sum of the charge due to the electrons in the inversion
layer Q’I and the charge due to the ionized acceptor atoms in the depletion region Q’B:

𝜓(y) decreases from 𝜓𝑠 toward zero, and n(y) decreases rapidly owing to its exponential dependence on
𝜓(y). Hence, one can choose a pointy = yc, below which the electron concentration will be negligible.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 24
Mcandrew, 3rd Edition (2011)
 Charge Sheet Approximation
➢A convenient approximation that leads to manageable models is that the inversion layer is a
sheet of negligible thickness.
➢Let the inversion layer thickness approach zero. The potential drop across the inversion layer,
which is the area of the corresponding part in the field plot, approaches zero, too,
➢We can assume that all of the surface potential 𝜓𝑠 is dropped across the depletion region. The
depth of the depletion region now becomes all of dB in the figure.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 25
Mcandrew, 3rd Edition (2011)
The field at the top of the
depletion region ℰB . Thus, this
field is the area under the
volume charge density, qNAdB,
divided by the permittivity 𝜖𝑠 :

The surface potential drop 𝜓𝑠 , is

then the area under the
corresponding triangle in the
field plot,

Fig 2.11: (a) Two-terminal MOS structure with p-type substrate; (b) charge density; (c) electric
field; (d) potential, all vs. depth.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 26
Mcandrew, 3rd Edition (2011)
Solving for dB :

Charge in pure depletion region:

Charge in inversion layer:

It is convenient to divide the inversion region into

three subregions: These are marked weak. moderate,
and strong inversion in Fig. 2.12.
Fig 2.12: Magnitude of inversion layer charge, depletion region charge, and their
sum (all per unit area) vs. surface potential.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 27
Mcandrew, 3rd Edition (2011)
➢ One can define the onset of weak inversion at 𝜓𝑠 = 𝜑𝐹 · The upper limit of weak inversion is defined in much
of the literature at 𝜓𝑠 =2𝜑𝐹 . It is seen that for surface potentials less than about this value, practically all the
surface charge is due to the charge in the depletion region.
➢ As 𝜓𝑠 , is raised above 2𝜑𝐹 , |Q’I| starts to become significant because of the exponential. For 𝜓𝑠 ., exceeding
2𝜑𝐹 by a few 𝜑𝑡 , Q’I becomes a very strong function of 𝜓𝑠 ·
➢ A point above 2𝜑𝐹 by a quantity 𝜑𝑍0 , which is several 2𝜑𝑡 , can be defined as the onset of strong inversion.
Relative contribution of change in Q’C can be:

The equation for Q’I :

The equation for VGB :

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 28
Mcandrew, 3rd Edition (2011)
 Fig 2.13: Surface potential vs. gate-substrate voltage (right) and charges vs. surface potential (left).

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 29
Mcandrew, 3rd Edition (2011)
 Strong Inversion
➢In strong inversion large changes in VGB result in very small changes in 𝜓𝑠 · A very common assumption is that in strong
inversion 𝜓𝑠 , is practically "pinned" to a constant value:

➢The value of 𝜑0 is close to 2𝜑𝐹 + 𝜑𝑍0 :

➢where ∆𝜑 will be considered constant for simplicity and, for the case of uniform substrates considered so far, is several 𝜑𝑡 .
➢The depletion region width dB is assumed to reach a maximum value dBm and not to increase with VGB anymore.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 30
Mcandrew, 3rd Edition (2011)
 Where,

Fig 2.14: Magnitude of inversion layer charge per unit area vs. gate-
substrate voltage.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 31
Mcandrew, 3rd Edition (2011)
 Weak Inversion
➢In weak inversion, for Q’I (VGB) consider
Q’I (VGB)

➢In the term 𝜓𝑠 + 𝜉 , in weak inversion 𝜓𝑠 is smaller than 2𝜑𝐹 . Thus 𝜉 ≪ 𝜓𝑠 . The term 𝜓𝑠 + 𝜉 can be
approximated by Taylors expansion around 𝜉 = 0.

➢To obtain a relation between Q’I; and the external bias VGB, we need to relate 𝜓𝑠 in the preceding equation
to VGB·
➢In weak inversion the inversion layer charge is extremely small compared with the depletion region charge,
it does not influence the surface potential. The total semiconductor charge is practically equal to the
depletion region charge, and the equations developed earlier for depletion apply here;

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 32
Mcandrew, 3rd Edition (2011)
 t 2q si N A ( sa − 2F ) t
Using  s =  sa , QI = − e
2  sa
We will be interested with the slope of  sa
( and hence  s in weak inv ) vs. VGB ,
−1
 d sa 
the inverse of which is denoted by n =   Fig 2.15: Surface potential and potential 𝜓𝑠𝑎 vs. gate-
 dVGB  substrate voltage.
−1 −1
   
 (
n =  2 −  2 +  2 4 + VGB − VFB
1
2  2 4 + VGB − VFB
) 

=
 sa
  2 4 +V −V


= 1+

2  sa
   GB FB 

 sa is not exactly a linear fn of VGB , because a part of VGB is also used in changing  ox.
→ VGB changes more than  sa , means n  1, typically, 1<n  1.5
−1
 d sa  
we have n = n0 =   = 1+
 VGB  2 2F
 sa = 2F

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 33
Mcandrew, 3rd Edition (2011)
  sa − 2F 1 V − VM 0
when  sa = 2F , VGB = V M 0 , then    sa − 2F = GB
VGB − V M 0 n0 n0
t 2q si N A ( sa − 2F ) t
Using this and x = 2F in QI = − e ,
2 2F

VGB −V M 0 ) ( n0t ) t 2q si N A

we get QI  QM 0e( , where QM 0 = −
2 2F
 − 2q si N A
note that n − 1 = =
2  sa  2  sa
C ox
− 2q si N A
 = ( n − 1) C ox
 and using  sa = 2F ,
2  sa
− 2q si N A
= ( n − 1) C ox
  we can also write
2 2F
 0 = ( n − 1) C ox
QM  t
Conclusions: QI variation with VGB
weak inv: exponentially
Strong inv: linearly Fig 2.16: Logarithm of inversion layer charge magnitude
per unit area vs. gate-substrate voltage. (a)(a), (b), (c)
Moderate inv: neither exponential nor linear
Adapted from “Operation And Modeling of The
MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 34
Mcandrew, 3rd Edition (2011)
 Small-Signal Capacitance
➢If VGB is increased by a small amount ∆𝑉𝐺𝐵, a positive charge ∆𝑄′ 𝐺 will flow into the gate terminal.
➢For overall charge neutrality, a charge of equal value must flow out of the body terminal or, equivalently, a
charge of value -∆𝑄′ 𝐺 must flow into the body terminal. An incremental (small-signal) capacitance per unit
area, C’gb, can thus be defined to relate charge changes to voltage changes.

➢The charge - ∆𝑄′ 𝐺 flowing into the body goes to change the semiconductor charge 𝑄′ 𝐶 by an amount ∆𝑄′ 𝐶 :

➢The gate-to-body voltage change will be distributed partly across the oxide (as ∆𝜓𝑂𝑋 ) and partly across the
semiconductor (as a change ∆𝜓𝑆 in the surface potential).

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 35
Mcandrew, 3rd Edition (2011)
Taking l/C’gb :

Where,

So,

Fig 2.16: Small-signal equivalent circuit for a two-

terminal MOS structure driven by a voltage source.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 36
Mcandrew, 3rd Edition (2011)
Determining C’C :

C’C Deep in Accumulation Deep in accumulation (for 𝜓𝑆 below zero by several 𝜑𝑡 ),

C’C Deep in Depletion and in Inversion (for 𝜓𝑆 above zero by several 𝜑𝑡 ),

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 37
Mcandrew, 3rd Edition (2011)
➢ To consider the individual contributions of the depletion region and inversion layer charges to C’c :

Fig 2.17: Small-signal capacitances per unit area vs.

surface potential.

Adapted from “Operation And Modeling of The

MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 38
Mcandrew, 3rd Edition (2011)
o If ∆𝑉𝐺𝐵 is a small-signal sinusoidal voltage, the steady-state
charge changes will also be sinusoidal. They will correspond
to equilibrium values if the frequency is low enough (e.g., 1
Hz). However, if the frequency is high (e.g., 100 kHz), things
will be different.

o The inversion layer charge now cannot keep up with the fast-
changing ∆𝑉𝐺𝐵 , and the required charge changes must be
provided by covering or uncovering acceptor atoms at the
bottom of the depletion region, just as in the case of depletion
operation.

o Thus, as in the depletion case, C’gb is effectively the series

combination of C’ox and C’b . In strong inversion, the width of
the depletion region reaches a maximum value, and C’b ,
attains a limit value. Fig 2.18: Small-signal equivalent circuit for the two-
terminal MOS structure, showing individual capacitances.
Adapted from “Operation And Modeling of The
MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 39
Mcandrew, 3rd Edition (2011)
The corresponding minimum value of C’gb, denoted by
C’gb,l

o The reason the inversion layer charge cannot follow fast

enough if the frequency is high, is that it is in a sense
isolated from the outside world by the oxide on top and
the depletion region below.
o Therefore, the electron concentration there can be
changed only by the mechanisms of thermal generation
and recombination, which in this case are very slow (no
external irradiation is assumed).
o If, instead, communication with the outside world were
possible, in the sense that inversion layer charge could
be provided or removed externally, then the behavior
exhibited by the solid curve in strong inversion would Fig 2.19: Total gate-body capacitance per unit area vs. gate-body
persist up to much higher frequencies. This bias. Solid line: "static" behavior; broken line: high-frequency
behavior.
communication with the outside world is provided by
the source and drain regions in a MOS transistor.
Adapted from “Operation And Modeling of The
MOS Transistor” by Yannis Tsividis & Colin DR. KUMARI NIBHA PRIYADARSHANI 40
Mcandrew, 3rd Edition (2011)