Fundamentals of Modern VLSI Devices - Solution Manual (1332223)

SOLUTION SET
to
Exercises in
FUNDAMENTALS OF MODERN VLSI DEVICES, 2nd ed.
published by
Cambridge University Press
Yuan Taur and Tak H. Ning
July 2009
SOLUTION 1
Solutions to Chapter 2 Exercises
2.1. From Eq. (2.4),
1 eΔE / kT
f ( E f − ΔE ) = = ,
1 + e− ΔE / kT eΔE / kT + 1
and
1
f ( E f + ΔE ) = .
1 + eΔE / kT
Adding the above two equations yields
eΔE / kT + 1
f ( E f − ΔE ) + f ( E f + ΔE ) = = 1.
eΔE / kT + 1
2.2. Neglecting the hole (last) term in Eq. (2.19), one obtains
− ( E c − E f )/ kT − ( E c + E d − 2 E f )/ kT
N ce + 2N ce = Nd .
Treating exp(Ef/kT) as an unknown, the above equation is a quadratic equation with the solution
E f / kT −1 + 1 + 8( N d / N c )e( E c − E d )/ kT
e = .
4e− E d / kT
Here only the positive root has been kept. For shallow donors with low to moderate
concentration at room temperature, (Nd/Nc)exp[(Ec − Ed)/kT] << 1, and the last equation can be
approximated by
E f / kT 4( N d / N c )e( E c − E d )/ kT N d E c / kT
e = = e ,
4e− E d / kT Nc
which is the same as Eq. (2.20). If we compare the above relation with Eq. (2.19), it is clear
that in this case, exp[−(Ed − Ef)/kT] << 1, and Nd+ ≈ Nd or complete ionization.
If the condition for low to moderate concentration of shallow donors is not met, then
exp[−(Ed − Ef)/kT] is no longer negligible compared with unity. That means Nd+ < Nd (Eq.
(2.19)) or incomplete ionization (freeze-out). [Note that incomplete ionization never occurs for
shallow impurities: arsenic, boron, phosphorus, and antimony at room temperature, even for
doping concentrations higher than Nc or Nv. This is because in heavily doped silicon, the
SOLUTION 2
impurity level broadens and the ionization energy decreases to zero, as discussed in Section
2.1.2.3.]
2.3. (a) Substituting Eqs. (2.5) and (2.3) into the expression for average kinetic energy, one
obtains
∞
∫
− ( E − E f )/ kT
( E − Ec )3/ 2 e dE
Ec
K.E. = ∞
.
∫
− ( E − E f )/ kT
( E − Ec )1/ 2 e dE
Ec
Applying integration by parts to the numerator yields

∞
(3 / 2)kT ∫ ( E − Ec )1/ 2 e
− ( E − E f )/ kT
dE 3
Ec
K.E. = ∞
= kT .
∫
1/ 2 − ( E − E f )/ kT
( E − Ec ) e dE 2
Ec
(b) For a degenerate semiconductor at 0 K, f(E) = 1 if E < Ef and f(E) = 0 if E > Ef.
Here Ef > Ec. Therefore,
Ef
K.E. =
∫ Ec
( E − Ec )3/ 2 dE 3
= ( E f − Ec ) .
Ef
∫ Ec
( E − Ec )1/ 2 dE 5
2.4. With the point charge Q at the center, construct a closed spherical surface S with radius
r. By symmetry, the electric field at every point on S has the same magnitude and points
outward perpendicular to the surface. Therefore,
∫∫ E ⋅ dS = 4πr E ,
2
S
where E is the magnitude of the electric field on S. 3-D Gauss’s law then gives
Q
E= ,
4πε si r 2
which is Coulomb’s law.
Since E =−dV/dr, the electric potential at a point on the sphere is
Q
V= ,
4πε si r
if one defines the potential to be zero at infinity.
SOLUTION 3
2.5. (a) Construct a cylindrical Gaussian surface perpendicular to the charge sheet as shown:
A
Qs
The cross-sectional area is A. At the two ends of the cylinder, the electric field E is
perpendicular to the surface and pointing outward. Along the side surface of the cylinder, the
field is parallel to the surface, so E⋅dS = 0. For an infinitely large sheet of charge, E is uniform
across A from symmetry. Therefore,
∫∫ E ⋅ dS = 2AE .
S
The charge enclosed within the surface is QsA. From Gauss’s law, one obtains E = Qs/2ε.
(b) The field due to the positively charged sheet is Qs/2ε pointing away from the sheet.
The field due to the negatively charged sheet is also Qs/2ε, but pointing toward the negatively
charged sheet. In the region between the two sheets, the two fields are in the same direction
and the total field adds up to Qs/ε, pointing from the positively charged sheet toward the
negatively charged sheet. In the regions outside the two parallel sheets, the fields are equal and
opposite to each other, resulting in zero net field.
2.6. (a) Using the integral expressions of Qd and Qi, one has
dQd qN a (1 − e− qψ s / kT )
Cd ≡ − = ,
dψ s Es
and
dQ q( ni / N a )( eqψ s / kT − 1)
2
Ci ≡ − i = .
dψ s Es
Here Es = E (ψ = ψs).
(b) Take the square of Eq. (2.182) and differentiate with respect to ψs, one obtains
SOLUTION 4
⎡ ⎤
( ) ( )
2
dQs n
2Qs = 2ε si qN a ⎢ 1 − e− qψ s / kT + i 2 eqψ s / kT − 1 ⎥ .
dψ s ⎣ Na ⎦
Using Qs =−εsiE s (Gauss’s law) and the above two equations in (a), it is straightforward to show
that
dQs qN a ⎡ ⎤
( ) ( )
2
− qψ s / kT ni qψ s / kT
Csi ≡ − = ⎢ 1 − e + 2 e − 1 ⎥ = Cd + Ci .
dψ s Es ⎣ Na ⎦
(c) When ψs = 2ψB, exp(qψs/kT) = (Na/ni)2 >> 1 and exp(−qψs/kT) << 1. From (a), one
has
qN a
Ci ≅ Cd ≅ .
Es
(d) It is clear from Fig. 2.33 that Qs and therefore Es takes off rapidly beyond strong
inversion. This means that Cd ∝ E s−1 decreases rapidly beyond strong inversion. We say that
the depletion layer (charge) is “screened” by the inversion layer. Note that Ci, on the other
hand, increases rapidly beyond strong inversion because of the exp(qψs/kT) factor.
2.7. In strong inversion and near the surface, Eq. (2.191) can be approximated by
dψ
2
2kTni qψ / 2 kT
=− e .
dx ε si N a
The solution to the above differential equation is
ni qx
e− qψ / 2 kT = e− qψ s / 2 kT + ,
2ε si kTN a
where ψs = ψ(x = 0) is the surface potential.
Using Eq. (2.178) and the above solution, one has

2
ni qψ / kT n( 0)
n( x ) = e =
( )
,
Na 1 + q n( 0) / 2ε si kT x
2
2
where n(0) equals (ni2/Na)exp(qψs/kT).
2.8. Substituting Eq. (2.189) into Eq. (2.195), one obtains
SOLUTION 5
2ε si qN aψ s
Vg − V fb = +ψ s ,
Cox
which is basically a quadratic equation that can be solved for ψs:
2
ε si qN a ε si qN a ⎛ ε qN ⎞
ψ s = Vg − V fb + 2
− 2 2
(Vg − V fb ) + ⎜⎜ si 2 a ⎟⎟ .
Cox Cox ⎝ Cox ⎠
For differential changes in ψs and Vg, one goes back to the previous equation and obtains
ε si qN a / 2ψ s
ΔVg / Δψ s = 1 + .
Cox
This is simply Δψs = ΔVg/(1 + Cd/Cox), where Cd is the depletion charge capacitance given by
Eq. (2.201).
2.9. By keeping only the dominant inversion charge term in Eq. (2.182) and substituting it in
Eq. (2.195), one obtains
2ε si kTN a ⎛ ni ⎞ qψ s / 2 kT
Vg = V fb + ψ s + ⎜ ⎟e .
Cox ⎝ Na ⎠
When Vg increases beyond the point where ψs = 2ψB, most of that increase appears as inversion
charge in the third term. There is very little change in the band bending beyond ψs = 2ψB. A
good approximate solution for ψs is then obtained by letting ψs = 2ψB for the second term and
solving ψs from the third term:
2kT ⎛ Cox (Vg − V fb − 2ψ B ) ⎞

ψ s ≈ 2ψ B + ln⎜
⎜ ⎟⎟ .
q ⎝ 2ε si kTN a ⎠
Here exp(qψB/kT) = Na/ni has been applied.
It is clear from the above equation that ψs − 2ψB is a weak function of Vg. As an
example, for Na = 1017 cm−3, tox = 100 Å, and Vg − Vfb = 4 V, one has 2ψB = 0.82 V, and ψs −
2ψB ≈ 7.2kT/q ≈ 0.19 V.
SOLUTION 6
2.10. In Fig. 2.37(b), the total capacitance seen by the gate equals Cox and (Ci + Cd) connected
in series, i.e.,
d ( −Qs ) Cox ( Ci + Cd )
C≡ = .
dVg Cox + Ci + Cd
Since Qs = Qi + Qd and ΔQi/ΔQd = Ci/Cd, it follows that
d ( −Qi ) Cox Ci
= .
dVg Cox + Ci + Cd
When Vg is below the threshold of strong inversion, Ci << Cox, therefore d(−Qi)/dVg ≈
Ci/(1 + Cd/Cox) << Cox. Above strong inversion, Ci >> Cox (> Cd) (see Exercise 2.6), and one
has d(−Qi)/dVg ≈ Cox. Such a behavior is sketched schematically below.
d(-Qi ) Cox
dVg
Vg
The sharp transition takes place at the threshold voltage. Qi(Vg) is simply the integrated area
under the above curve. Beyond threshold, Qi increases linearly with Vg with a slope equal to
Cox (see Fig. 3.17 in the text).
2.11. Equation (2.256) gives
1
Mp [ W
0 ]
W
0 0
x
[
= exp − ∫ (α p − α n )dx − ∫ α n exp − ∫ (α p − α n )dx ′ dx . ] (1)
As Mp → ∞, 1/Mp → 0, the above equation gives
0 = exp ⎡− ∫ (α p − α n )dx ⎤ − ∫ α n exp ⎡− ∫ (α p − α n )dx ′ ⎤dx

W W x
(2)
⎣⎢ 0 ⎦⎥ 0 ⎣⎢ 0 ⎦⎥
(
= exp −(α p − α n )W + ) αn
(α p − α n )
[exp(−(α p ) ]
− α n )W − 1 .
SOLUTION 7
Therefore, α n = α p exp( −(α p − α n )W ) , (4)
or ( ) (
W = ln α n / α p / α n − α p . )
Similarly, Eq. (2.257) gives the same expression for W as Mn → ∞.
x
2.12. Let u( x ) = − ∫ f ( x ' )dx ' . (1)
0
du( x )
Therefore = − f (x) (2)
dx
deu( x )
and = − f ( x )e u ( x ) . (3)
dx
Using (3), we have
f ( x )exp ⎡− ∫ f ( x ' )dx '⎤ dx = ∫ f ( x )eu( x ) dx

W x W
∫0 ⎣⎢ 0 ⎦⎥ 0
(4)
W deu( x )
= −∫ dx = 1 − eu(W ) .
0 dx
f ( x )exp ⎡− ∫ f ( x ' )dx '⎤ dx = 1 − exp ⎡− ∫ f ( x ' )dx '⎤ .

W x W
Therefore, ∫0 ⎣⎢ 0 ⎦⎥ ⎣⎢ 0 ⎦⎥
(5)
A similar procedure can be used to show that
f ( x )exp ⎡− ∫ f ( x ' )dx '⎤ dx = 1 − exp ⎡− ∫ f ( x ' )dx '⎤

W W W
∫0 ⎢⎣ x ⎥⎦ ⎢⎣ 0 ⎥⎦
(6)
W
by letting u( x ) = − ∫ f ( x ' )dx ' .
x
2.13. From Eqs. (2.80), (2.81), and (2.83), we have
2ε si ( Na + Nd )ψ m
Wd = ,
qNa Nd
ψ m = ψ bi − Vapp ,
SOLUTION 8
dQd (p - side) ε si
and Cd ≡ = .
dVapp Wd
2( N a + N d )
Therefore,
1
= (ψ bi − Vapp ) . (1)
Cd
2
ε si qN a N d
A plot of 1 / C d2 versus Vapp gives a straight line, which intercept the horizontal Vapp axis at Vapp
= ψbi and has a slope of −2(Na + Nd)/εsiqNaNd.
Now, from Eq. (2.73), we have
⎛n p ⎞ ⎛N N ⎞
qψ bi = kT ln ⎜ n 0 2 p 0 ⎟ = kT ln ⎜ d 2 a ⎟ . (2)
⎝ ni ⎠ ⎝ ni ⎠
Therefore, from the plot of 1 / C d2 versus Vapp plot, we have
− 2( N a + N d )
Slope = , (3)
ε si qN a N d
1 kT ⎛ N a N d ⎞ 2( N a + N d )
and = ln⎜ ⎟⎟ . (4)
C (Vapp = 0)
2
d q ⎜⎝ ni2 ⎠ ε si qN a N d
The two equations (3) and (4) can be solved for the two unknowns Na and Nd.
1/Cd2
Vapp
0 ψ bi
2.14.
SOLUTION 9
Na (x)
x
0 W(V)
n+ region p region
For the one-sided n+-p diode, the incremental increase in the magnitude of the depletion-
layer charge dQd on the p-side as a result of an incremental increase in the applied voltage dV is
dQd = qN a (W )dW , (1)
where the depletion-layer width W is a function of V. The depletion-layer capacitance is
dQd dW
C= = qN a (W ) , (2)
dV dV
which gives
C
N a (W ) = . (3)
q( dW / dV )
ε si
From C= ,
W
dW ε dC ε si C d (1 / C 2 )
we have = − si2 = . (4)
dV C dV 2 dV
Substituting (4) into (3), we have
2
N a (W ) = .
qε si [d (1 / C 2 ) / dV ]
2.15. Referring to Fig. 2.17 and Eq. (2.90), we have
Em (Wd + d )
ψ m = ψ bi + V = , (1)
2
where V is the applied voltage. From Gauss’s law we have
SOLUTION 10
Qd = ε siEm . (2)
Therefore,
1 dV d [Em (Wd + d ) / 2] 1 ⎡ dWd ⎤

= = = ⎢(Wd + d ) + Em
dEm ⎥⎦
. (3)
Cd dQd d [ε siEm ] 2ε si ⎣
Now, from Eq. (2.89), we have
dWd d ( xn + x p ) dx dx
Em = Em = Em n + Em p = ( x n − d ) + x p . (4)
dEm dEm dEm dEm
1 Wd
= .
Cd ε si
2.16.
( pn - pn0 )
x
−W n 0
n-region p-region
From Eqs. (2.115) and (2.116), the equation governing the steady-state hole distribution
in the n-region is
d 2 pn pn − pn 0
− = 0, (1)
dx 2 L2p
kTμ pτ p
where Lp ≡ τ p D p =
q
is the hole diffusion length. Let x = −Wn be the location of the ohmic contact on the n-side.
The boundary conditions for the hole distribution are
SOLUTION 11
pn = pn 0 exp( qVapp / kT ) at x = 0, (2)
and pn = pn 0 at x = −Wn. (3)
The general solution to (1) has the form of
− x / Lp x / Lp
pn − pn 0 = Ae + Be . (4)
Using (2), (4) gives
[
pn 0 exp(qVapp / kT ) − 1 = A + B . ] (5)
Using (3), (4) gives
Wn / L p −Wn / L p
0 = Ae + Be . (6)
Solving (5) and (6), we have
Wn / L p
[
B = pn 0 exp( qVapp / kT ) − 1 ]e e
Wn / L p
−e
− Wn / L p , (7)
− Wn / L p
and [
A = − pn 0 exp( qVapp / kT ) − 1 ]eWn / L p
e
−e
− Wn / L p
. (8)
Substituting (7) and (8) into (4) and rearranging, we have
[
pn − pn 0 = pn 0 exp( qVapp / kT ) − 1 ] [
sinh ( x + Wn ) / Lp ]. (9)
sinh(Wn / Lp )
Since there is no electric field in the n-region, the hole current density entering the n-region is,
from Eq. (2.55) in the text and (9),
⎛ dp ⎞
J p ( x = 0) = − qD p ⎜ n ⎟ =
− qD p pn 0 exp( qVapp / kT ) − 1
.
[ ]
⎝ dx ⎠ x = 0 Lp tanh(Wn / Lp )
2.17. (a) Ignoring heavy doping effect, from Eq. (2.129) the electron saturation current
density is
qDn ni2 qni2 Ln

Jn = = , (1)
N a Ln Na τ n
SOLUTION 12
where we have used the fact that (D/L) = (L/τ). From Table 2.1, we have ni = 1.0×1010 cm−3,
and from Fig. 2.24, for Na = 1017 cm−3, we have Ln = 75 μm = 7.5×10−3 cm, and τn = 3×10−6 s.
Substituting these values into (1) gives Jn = 4.0×10−13 A/cm2. Similarly, the hole saturation
current density is
qDp ni2 qni2 Lp

Jp = = . (2)
N d Lp Nd τ p
From Fig. 2.24, for Nd = 1020 cm−3, we have Lp = 0.38 μm = 3.8×10−5 cm, and τp = 5×10−10 s.
Substituting these values into (2) gives Jp = 1.2×10−14 A/cm2.
(b) If the n+-region is shallow, with a width of Wn = 0.1 μm = 1×10−5 cm, the hole
saturation current density is
qDp ni2 qni2 L2p

Jp = = . (3)
N dWn N d τ pWn
Using the same values as in (a), we have Jp = 4.6×10−14 A/cm2.
(c) If heavy-doping effect is included in (b), then the hole saturation current density is
given by
qDp nie2 qni2 L2p

Jp = = exp( ΔE g / kT ) . (4)
N dWn N d τ pWn
From Fig. 6.3, for an n+-region doped to 1020 cm−3, ΔEg ≈ 92 meV. At room temperature (300
K), kT = 26 meV, and exp(ΔEg/kT) = 34. That is, the effect of heavy doping increases Jp by a
factor of 34, giving Jp = 1.6×10−12 A/cm2.
2.18. Consider a wide-emitter but narrow-base n+-p diode, with emitter doping concentration
NE = 1020 cm−3, base doping concentration NB = 1017 cm−3, and base width WB = 100 nm.
(a) Ignoring heavy-doping effect, Eq. (2.143) gives the diffusion capacitance ratio
C Dn 2 N E WB
= . (1)
C Dp 3 N B L pE
From Fig. 2.24, we have LpE = 0.38 μm for holes in the n+ emitter. Substituting all these values
into (1) gives CDn/CDp = 175.
(b) When heavy-doping effect cannot be ignored, we start with Eq. (2.165) which gives
SOLUTION 13
2 q 2 qA
C Dn = I nt B = J n tB , (2)
3 kT 3 kT
where A is the cross-sectional area of the diode. From Eq. (2.131), we have
qDn nieB
2
Jn = − exp(qVapp / kT ) . (3)
N BWB
From Eq. (2.147), we have
WB2
tB = . (4)
2 Dn
Substituting (3) and (4) into (2), we obtain
2 qA qnieB
2
WB
C Dn = exp(qVapp / kT ) . (5)
3 kT 2 N B
Similarly, Eq. (2.166) and the hole-equivalent of Eq. (2.127) give
1 q 1 qA
C Dp = I pτ pE = J p τ pE (6)
2 kT 2 kT
1 qA qD p nieEτ pE
2
= exp(qVapp / kT )
2 kT N E L pE
1 qA qnieE L pE
2
= exp(qVapp / kT )
2 kT N E
where we have used the fact that L2 = Dτ. (5) and (6) give
C Dn 2 ⎛ nieB
2
⎞⎛ N E ⎞ WB
= ⎜⎜ 2 ⎟⎟⎜⎜ ⎟⎟ . (7)
C Dp 3 ⎝ nieE ⎠⎝ N B ⎠ L pE
Now, nieB
2
= ni2 exp( ΔE gB / kT )
and nieE
2
= ni2 exp( ΔE gE / kT ) .
nieB
2
⎛ ΔE gB − ΔE gE ⎞
Therefore, = exp⎜ ⎟. (8)
nieE
2
⎝ kT ⎠
SOLUTION 14
From Fig. 6.3, we have ΔEgB ≈ 2 meV for p-type base with doping concentration of 1017 cm−3,
and ΔEgE ≈ 92 meV for n-type emitter with doping concentration of 1020 cm−3. Also, at room
temperature (300 K), kT = 26 meV. Therefore
nieB
2
⎛ ΔE − ΔE gE ⎞
= exp⎜ gB ⎟ = 0.031. (9)
nieE
2
⎝ kT ⎠
That is, the effect of heavy-doping reduces the ratio CDn/CDp from 175 to 5.4.
2.19. We will do this by showing that
[ (
nT ( t ) = N T 1 − exp −σ N inj ( t ) , )] (1)
jG ( t ′ )
t
where N inj ( t ) ≡ ∫ dt ′ , (2)
0 q
is a solution to the differential equation. From (1), we have
dnT dN inj ( t )
dt
(
= σN T exp −σN inj ( t )
dt
) . (3)
From (2), we have
dN inj ( t ) jG ( t )
= . (4)
dt q
Substituting (4) into (3) and using (1), we have
dnT j
= G σ ( N T − nT ) .
dt q
1E+13
NT = 5E12 cm-3
σ = 1E-13 cm2
nT (cm-3)
1E+12
1E+11
2E+10
1E+11 1E+12 1E+13 5E+13
Ninj (cm-2)
SOLUTION 15
2.20. From Eq. (2.256), we have
= exp⎛⎜ − ∫ (α p − α n )dx ⎞⎟ − ∫ α n exp⎛⎜ − ∫ (α p − α n )dx′ ⎞⎟dx .

1 W W x
(1)
Mp ⎝ 0 ⎠ 0 ⎝ 0 ⎠
Since αp and αn are independent of x, (1) can be integrated to give
1 −W (α p −α n ) W − x (α − α )
=e − α n ∫ e p n dx (2)
Mp 0
=e
−W (α p −α n )
−
αn
α p − αn
[1 − e −W (α p −α n )
]
αp −W (α p −α n ) αn
= e − .
α p − αn α p − αn
Equation (2) can be rearranged to give
−W (α p −α n ) 1 α p − αn αn
e = + , (3)
Mp αp αp
⎛α 1 α p − α n ⎞⎟
which means − W (α p − α n ) = ln⎜ n + . (4)
⎜α M α ⎟
⎝ p p p ⎠
As avalanche breakdown is approached, Mp approaches infinity and W approaches a value given

by
ln(α p / α n )
W = . (5)
α p −αn
2.21. The band-to-band tunneling can be represented as tunneling through a triangular barrier
of height Eg, slope qE and tunneling distance Eg / qE. This is illustrated in the figure
below.
SOLUTION 16
Energy
Ec
slope = qE
Ev
Eg
0 W x
Ec
Ev
p-region n-region
The WKB tunneling coefficient is, from Eq. (2.245),
⎡ − 4π W ⎤
T ( E ) = exp ⎢
⎣ h
∫
0
2m* − qψ ( x) − E dx ⎥
⎦
(1)
with W = Eg / qE. For the coordinates indicated in the figure, we have
− qψ ( x) = E g − qxE . (2)
Also, for the tunneling electron being considered, E = 0. Substituting into (1), we have
⎡ − 4π W ⎤
T ( E = 0) = exp ⎢
⎣ h
∫0
2m* Eg − qxE dx ⎥
⎦
(3)
⎡ − 4 2m E ⎤
* 3 / 2
= exp ⎢ g
⎥.
⎢⎣ 3hqE ⎥⎦
SOLUTION 17
2.22. (a)
p-type
Ec
Ef
Ev
2V
Ec
Ef
Ev
n-type
kT ⎛ N a N d ⎞
(b) ψ bi = ln⎜ ⎟ = 0.82 V
q ⎜⎝ ni 2 ⎟
⎠
2ε si ( N a + N d )(ψ bi + VR )
Wd = = 0.27 μm
qN a N d
(c)
Emax = qN a (Wd / 2) = 2.1×105 V/cm
ε si
2.23. (a)
Wd
Ec
Ei ψ bi
Ef Ec E f
Ev
Ev
p-type n+
x=0
⎛N ⎞
Ei − E f = kT ln⎜⎜ a ⎟⎟ = 0.38 eV
⎝ ni ⎠
SOLUTION 18
(b)
Eg
ψ bi = + 0.38 eV = 0.94 eV
2q
(c)
2ε siψ bi
= 0.20 μm = 2×10 cm
-5
Wd =
qN a
⎛N ⎞
(d) E f − Ev = kT ln⎜⎜ v ⎟⎟ increases as temperature increases.
⎝ Na ⎠
So Ei − E f decreases as temperature increases.
And ψ bi decreases as temperature increases.
2.24.
⎛ Eg ⎞ ⎛ kT N d ⎞
V fb = −⎜⎜ −ψ B ⎟⎟ = −⎜⎜ 0.56 − ln ⎟ = −(0.56 − 0.35) = −0.21 V.
⎝ 2q ⎠ ⎝ q ni ⎟⎠
Vfb Vg
Band diagram at Vg = 0:
Ef Ef
n+ n-type
SOLUTION 19
At ψs = 2ψB = 0.7 V:
Qd 2ε si qN dψ s 4.8 × 10 −8
Vox = = = = 0.14 V.
Cox Cox Cox
2.25.
kT N a
(a) ψB = ln = 0.42 V
q ni
E
V fb = − g −ψ B = −0.98 V
2q
2ε si qN a (2ψ B )
Vt = V fb + 2ψ B + = −0.98 + 0.84 + 0.97 = 0.83 V.
Cox
(b)
C
Cmax
Cmin
Vfb Vt Vg
(c) Cmax = Cox = 1.73 ×10 −7 F/cm2
2ε si (2ψ B )
Wdm = = 0.104 μm
qN a
1 1 Wdm
= +
Cmin Cox ε si
Cmin = 0.63 ×10 −7 F/cm2
SOLUTION 20
2.26.
Let the surface potential at zero gate voltage be ψs.
2ε si qN aψ s
Then − V fb = ψ s +
Cox
Solve the above quadratic equation in ψs :
2ε si qN a 2ε si qN a
ψs = − + − V fb
2Cox 4Cox
2
2ε si qN a
2
= 0.278 V,
4Cox
Therefore, ψs = 0.353 V
2
ni qψ s / kT
Electron density at the surface: n = e = 8.3 ×108 cm-3
Na
hole density at the surface: p = N a e − qψ s / kT = 1.2 × 1011 cm-3
SOLUTION 21
3.1.
kT N a
(a) ψB = ln = 0.42 V
q ni
E
V fb = − g −ψ B = −0.98 V
2q
2ε si qN a (2ψ B )
Vt = V fb + 2ψ B + = −0.98 + 0.84 + 0.97 = 0.83 V.
Cox
Ec
E
i
Ef
Ev
Ef
p-type silicon
Oxide
SOLUTION 22
(b)
2ε si qN a (2ψ B + Vbs ) V
Vt = V fb + 2ψ B + = −0.98 + 0.84 + 1.44 = 1.30
Cox
Ec
E
i
Ef
Ev
Efn
Ef
p-type silicon
Oxide
2ε si (2ψ B )
(c) Wdm = = 0.104 μm
qN a
λ = Wdm + 3tox = 0.164 μm
Lmin ≈ 2λ = 0.33 μm
SOLUTION 23
3.2. Eq. (3.14) is an implicit equation relating ψs and V. Taking a differential yields
−1 / 2
2ε si kTN a ⎡ qψ s ni 2 q(ψ s −V )/ kT ⎤ q ⎛
2
ni q(ψ s −V )/ kT ⎞
0 = dψ s + ⎢ + 2e ⎥ ⎜ dψ s + 2 e ( dψ s − dV ) ⎟.
2Cox ⎣ kT N a ⎦ kT ⎝ Na ⎠
It is straightforward to re-group the terms and show that
dψ s ( ni / N a )eqψ s / kT
2 2
=
dV 1 + ( ni 2 / N a 2 )eqψ s / kT + ( Cox 2 / ε si qN a )( Qs / Cox )
in the limit of V → 0, where

1/ 2
⎡ qψ n
2
⎤
Qs = 2ε s kTN a ⎢ s + i 2 eqψ s / kT ⎥ .
⎣ kT N a ⎦
3.3. Applying 1-D Gauss’s law to a box bounded by a plane at depth x from the surface and
another plane deep in the substrate where E = 0, one can write
1 ⎛
⎜ Qd + q∫x n( x′)dx ′ ⎞⎟ ,
xi
E (x) =
ε si ⎝ ⎠
where Qd is the depletion charge density and xi is the inversion layer depth. Therefore,
xi Qd xi q xi xi
∫0
n( x )E ( x )dx =
ε si ∫
0
n( x )dx +
ε si ∫0
n( x )∫ n( x ′)dx′dx .
x
If we let n(x) = df(x)/dx, the second term on the RHS becomes
q df q ⎡ x i df 1 xi d ( f 2 ) ⎤
[ f ( xi ) − f ( x )]dx =
xi
ε si ∫0 ∫0 dx 2 ∫0 dx dx⎥⎦ ,
⎢ f ( x ) dx −
dx ε si ⎣
i
which can be further simplified to
q ⎧
[ ⎫ q
] q ⎡ xi 2
[ ]
1 2
[ ] ⎤.
2ε si ⎣⎢∫0
− − − = − =
2
⎨ f ( x ) f ( x ) f ( 0 ) f ( x ) f 2
( 0 ) ⎬ f ( x ) f ( 0 ) n( x ) dx
ε si ⎩ i i
2
i
⎭ 2ε si
i
⎦⎥
Substituting it back, one has

xi
Eeff ≡
∫0
n( x )E ( x )dx
=
Qd
+
Qi
,
xi
ε si 2ε si
∫
0
n( x )dx
SOLUTION 24
where
xi
Qi = q∫ n( x )dx .
0
3.4. This exercise is closely related to Exercise 2.10, in which Eq. (3.61) is derived:
d ( −Qi ) Cox Ci
= .
dVg Cox + Ci + Cd
When Vg is below the threshold of strong inversion, d(−Qi)/dVg ≈ Ci/(1 + Cd/Cox) << Cox.
Above strong inversion, Ci >> Cox (> Cd), and one has d(−Qi)/dVg ≈ Cox. This behavior is
sketched below.
d(-Qi ) Cox
dVg
Vg
Qi(Vg) is simply the integrated area under the curve. Beyond threshold, Qi increases linearly
with Vg with a slope equal to Cox (see Fig. 3.17 in the text).
The sharp transition above can be used to define a kind of inversion charge threshold
voltage, Vtinv, where Ci = Cox. Since Cd → 0 due to screening by inversion charge (Exercise
2.6), d|Qi|/dVg ≈ CoxCi/(Cox+Ci) = Cox/2 at Vg = Vtinv. Also, from Eq. (2.206), one has Qi ≈
(2kT/q)Cox at this point. Such a threshold voltage is slightly higher than the “2ψB” threshold
where Ci = Cd.
3.5. From the last term of Eq. (3.63), the polysilicon depletion effect causes an inversion
charge loss of
Cox (Vg − Vt )2
3
ΔQi = .
2ε si qN p
If we neglect the depletion charge in bulk silicon, then Qp ≈ Qi ≈ Cox(Vg − Vt) and the above
equation can be written as ΔQi/Qi ≈ CoxQp/2εsiqNp. From the discussions below Eq. (2.212),
the polysilicon depletion capacitance is Cp = εsiqNp/Qp. Therefore, one has ΔQi/Qi ≈ Cox/2Cp.
SOLUTION 25
There is a factor of two difference between the loss of charge and the loss of capacitance
because charge is the integration of capacitance and the polysilicon depletion capacitance is
voltage dependent (the next increment of charge always appears at the far edge of the poly
depletion region). This shows that treating the polysilicon depletion region as an equivalent
oxide layer will over estimate its effect on the drain current.
3.6.
(a) I ds =
[
μeff Cox (W / L ) (Vg − Vt )Vds − ( m / 2)Vds 2 ] (m = 1)
1 + ( μeff Vds / vsat L )
8
Vg -Vt =5.0 V
Drain current (mA)
6
4.0 V
4 3.0 V
2.0 V
2
1.0 V
0
0 1 2 3 4 5
Drain voltage (V)
(b) 1 + 2μeff (Vg − Vt ) / ( mvsat L ) − 1

I dsat = CoxWvsat (Vg − Vt ) (m = 1)
1 + 2μeff (Vg − Vt ) / ( mvsat L ) + 1
15
Vg -Vt =3.0 V
Saturation current (mA)
no velocity saturation
10
velocity saturation
0
0 1 2 3 4 5
Channel length (um)
SOLUTION 26
3.7. Differentiating Eq. (3.79) with respect to Vg, one obtains
dI dsat ⎡ 1 ⎤
gmsat ≡ = CoxWvsat ⎢1 − ⎥.
dVg ⎢⎣ 1 + 2μ eff (Vg − Vt ) / ( mvsat L ) ⎥⎦
When L → 0, gmsat → CoxWvsat (Eq. (3.96)).
When vsat → ∞, the square root factor can be expanded in a Taylor series to yield
⎡ ⎛ μ (V − Vt ) ⎞⎤ W Vg − Vt
gmsat = CoxWvsat ⎢1 − ⎜1 − eff g ⎟⎥ = μ eff Cox ,
⎢⎣ ⎝ mvsat L ⎠⎥⎦ L m
which can also be derived from the long-channel current expression, Eq. (3.80).
3.8. Since Ids = −WQiv and Qi = −Cox(Vg − Vt) at the source, one has
1 + 2μeff (Vg − Vt ) / ( mvsat L ) − 1

v = vsat
1 + 2μeff (Vg − Vt ) / ( mvsat L ) + 1
for the carrier velocity at the source end of the channel.
When L → 0, v → vsat. When vsat → ∞, the denominator approaches 2 and the square-
root expression in the numerator can be expaned in a Taylor series to yield
μ eff (Vg − Vt )
v= ,
2mL
which also follows from Ids = −WQiv = WCox(Vg − Vt)v and the long-channel current, Eq. (3.80).
3.9. Using an n = 2 velocity saturation model instead of n = 1, Eq. (3.73) becomes
μeff ( dV / dy )
I ds = −WQi (V ) .
1 + ( μeff / vsat )2 ( dV / dy )2
Substituting Eq. (3.76) for Qi(V), one can solve for dV/dy:
dV vsat I ds / μeff
= .
dy Cox W 2 vsat (Vg − Vt − mV )2 − I ds
2 2 2
SOLUTION 27
Since Ids is a constant based on current continuity, the above equation can be solved for V(y):
V ds L vsat I ds vsat I ds L
∫ Cox W 2 vsat (Vg − Vt − mV )2 − I ds dV = ∫ dy =
2 2 2
.
0 0 μeff μeff
Even though the integral can be carried out analytically by letting CoxWvsat(Vg − Vt − mV) =
Idscosh(u), the resulting transcendental equation cannot be solved for a closed form of Ids(Vds).
A numerical approach would be to evaluate Vds from the above integral equation for a given Ids.
3.10. By definition, the total integrated inversion charge under the gate is
L
Qi ( total ) ≡ ∫ Qi ( y )Wdy .
0
Change variable from y to V:
V ds Qi (V )WdV
Qi ( total ) = ∫ .
0 dV / dy
Substituting dV/dy from Eq. (3.74), one obtains
V ds ⎡ μeff W μeff W 2 2 ⎤
Qi ( total ) = − ∫ ⎢ Qi (V ) + Qi (V )⎥ dV .
0
⎢⎣ vsat I ds ⎥⎦
From Eq. (3.76), Qi(V) = −Cox(Vg − Vt − mV), the above integral can be evaluated to yield
CoxWμeff Cox W 2 μ eff

2
Qi ( total ) =
2mvsat
[
(Vg − Vt ) − (Vg − Vt − mVds ) −
2 2
3mI ds
] [
(Vg − Vt )3 − (Vg − Vt − mVds )3 . ]
In the saturation region, Vds and Ids are given by Eqs. (3.78) and (3.79). It can be shown that
the above equation reduces to
1 + 2μeff (Vg − Vt ) / ( mvsat L ) + (1 / 3)

Qi ( total ) = −WLCox (Vg − Vt ) .
1 + 2μeff (Vg − Vt ) / ( mvsat L ) + 1
SOLUTION 28
3.11.
-9
x 10
1
0.8
0.6
0.4
0.2
-0.2
-0.4
-0.6
-0.8
-1
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
-8
x 10
f (λn ) = ε si tan(πti / λn ) + ε i tan(πWd / λn )
From the figure above, λ1 ~ 15.9(nm) , λ2 ~ 7.3(nm) , λ1 ~ 5.0(nm) .

exp(− πL 2λ1 )
By substituting the numbers, ~ 40.14 , when L ~ 2λ1 .
exp(− πL 2λ2 )
SOLUTION 29
4.1. Under the scaling transformation, W → W/κ, L → L/κ, tox → tox/κ, Vds → Vds/κ, Vg →
Vg/κ, and Vt → Vt/κ, Eq. (3.23) becomes
W / κ ⎛ Vg Vt ⎞ Vds I ds
I ds → μeff (κCox ) − = ;
L / κ ⎜⎝ κ κ ⎟⎠ κ κ
and Eq. (3.28) becomes

2
W / κ 1 ⎛ Vg Vt ⎞ I
I ds → μ eff (κCox ) ⎜ − ⎟ = ds .
L / κ 2m ⎝ κ κ ⎠ κ
Note that both m = 1 + 3tox/Wdm and μeff which is a function of Eeff given by Eq. (3.53) are nearly
invariant under constant field scaling.
4.2. Under the same scaling rules as above, Eq. (3.40) becomes
2
W /κ ⎛ kT ⎞ q(V −V )/ κmkT
I ds → μeff (κCox ) ( m − 1)⎜ ⎟ e g t .
L/κ ⎝ q ⎠
The exp(−qVds/kT) term has been neglected since typically Vds >> kT/q. In subthreshold, Vg <
Vt and exp[q(Vg − Vt)/κmkT] > exp[q(Vg − Vt)/mkT] (note that κ > 1), therefore, the subthreshold
current increases with scaling faster than κIds.
If the temperature also scales down by the same factor, i.e., T → T/κ, then Ids → Ids/κ,
same as the drift current in Exercise 4.1.
4.3. Since the factor μeff(Vg − Vt)/(mvsatL) is invariant under the scaling transformation, W →
W/κ, L → L/κ, tox → tox/κ, Vds → Vds/κ, Vg → Vg/κ, and Vt → Vt/κ, the square-root expression
and therefore the fraction in Eq. (3.79) is unchanged after scaling. The saturation current Idsat
of Eq. (3.79) then scales the same way as the fully saturation-velocity limited current, Eq. (3.81),
i.e.,
W ⎛V V ⎞ I
I dsat → (κCox ) vsat ⎜ g − t ⎟ ∝ dsat .
κ ⎝κ κ⎠ κ
SOLUTION 30
4.4. Under the generalized scaling rules, W → W/κ, L → L/κ, tox → tox/κ, Vg → (α/κ)Vg, and
Vt → (α/κ)Vt, Eq. (3.79) transforms as
α 1 + 2αμ eff (Vg − Vt ) / ( mvsat L ) − 1

I dsat → CoxWvsat (Vg − Vt ) .
κ 1 + 2αμ eff (Vg − Vt ) / ( mvsat L ) + 1
In the limit of μeff(Vg − Vt)/(mvsatL) >> 1 (short channel or velocity saturation), the fractional
factor in the above equation equals 1 independent of α, so Idsat → (α/κ)Idsat. On the other hand,
if αμeff(Vg − Vt)/(mvsatL) << 1 (long channel), the fractional factor can be approximated by
αμeff(Vg − Vt)/(2mvsatL), and one has Idsat → (α2/κ)Idsat. In general, the scaling behavior is
between the two limits.
qN a
4.5. (a) Vt = V fb + 2ψ B + (Wdm − xs )
Cox
4ε siψ B
Wdm = + xs
2
qN a
Eliminate Na from the above eqs,
4ε siψ B
xs = − Wdm
Cox (Vt − V fb − 2ψ B )
Vfb = -Eg/2q – ψB = -1.06 V, so xs = 17 nm.
Substituting in the 2nd eq., Na = 1.3×1017 cm-3.
(b) m = 1 + 3tox/Wdm = 1.21

Inv. Subth. Slope = m×60mV/decade = 73 mV/decade
(c) λ = Wdm + 3tox = 0.12 μm

Lmin ≈ 2λ = 0.24 μm
4.6. From Eq. (3.25) with Vt replaced by Von, the linear region characteristics of the two parts
are:
W1 ⎛ m ⎞
I ds1 = μ eff Cox ⎜Vg − Von1 − Vds ⎟Vds ,
L⎝ 2 ⎠
and
W2 ⎛ m ⎞
I ds 2 = μ eff Cox ⎜Vg − Von 2 − Vds ⎟Vds .
L⎝ 2 ⎠
Since they are connected in parallel, Vg and Vds are the same and the total current is Ids = Ids1 +
Ids2. It is straightforward to show that
SOLUTION 31
W1 + W2 ⎛ m ⎞
I ds = μ eff Cox ⎜Vg − Von − Vds ⎟Vds ,
L ⎝ 2 ⎠
where Von = (W1Von1 + W2Von2)/(W1 + W2).
4.7. From Eq. (3. 23) with Vt replaced by Von, the linear region characteristics of the two parts
are:
I ds = μ eff Cox
W
L1
( )
Vg − Von1 Vds1,
and
I ds = μ eff Cox
W
L2
( )
Vg − Von 2 Vds2 .
Here the second order terms in Vds1 and Vds2 have been neglected (low-drain bias). Since they
are connected in series, Ids are the same and the total voltage is Vds = Vds1 + Vds2. Therefore,
I ds ⎛ L1 L2 ⎞
Vds = ⎜⎜ + ⎟.
μeff CoxW ⎝ Vg − Von1 Vg − Von 2 ⎟⎠
Since Von1 ≈ Von2, one considers only the first-order terms of (Von1 − Von2) by letting Von2 = Von1 +
δ and expanding the second term in a power series of δ:
I ds ⎛ L1 + L2 L2 ⎞ I ds L1 + L2 ⎛ L2 / ( L1 + L2 ) ⎞
Vds ≈ ⎜⎜ + δ ⎟⎟ = ⎜1 + δ ⎟⎟ .
μeff CoxW ⎝ Vg − Von1 (Vg − Von1 ) ⎠ μeff CoxW Vg − Von1 ⎜⎝
2
Vg − Von1 ⎠
This, in turn, can be approximated by
I ds ( L1 + L2 ) / (Vg − Von1 ) I ds L1 + L2
Vds ≈ = ,
μeff CoxW 1 − 2L / ( L + L ) μeff CoxW Vg − Von1 − (Von 2 − Von1 ) L2 / ( L1 + L2 )
1 2
δ
Vg − Von1
which can be written as
I ds L1 + L2
Vds = ,
μeff CoxW Vg − Von
where the equivalent linear threshold voltage is Von = (L1Von1 + L2Von2)/(L1 + L2).
SOLUTION 32
4.8. When Ids is considered as a function of the external voltages: Ids(Vg,Vds), one has
⎛∂ I ⎞ ⎛∂I ⎞
dI ds = ⎜ ds ⎟ dVg + ⎜ ds ⎟ dVds .
⎜ ∂ Vg ⎟
⎝ ⎠Vds ⎝ ∂ Vds ⎠V g
From Eqs. (4.63) and (4.64),
dVds = dVds′ + ( Rs + Rd )dI ds ,
and
dVg = dVg′ + Rs dI ds .
Substituting them into the top equation and solving for dIds, one obtains
(∂ I ds / ∂ Vg )V ds
dI ds = dVg′ + (a similar term in dVds′ ).
1 − Rs (∂ I ds / ∂ Vg )Vds − ( Rs + Rd )(∂ I ds / ∂ Vds )V g
Now Ids is expressed as a function of the internal voltages: Ids(Vg′,Vds′). One can identify the
first coefficient as the intrinsic transconductance, i.e.,
⎛∂ I ⎞ gm
gm′ ≡ ⎜ ds ⎟ = ,
⎜ ∂ Vg′ ⎟
⎝ ⎠Vds′ 1 − gm Rs − gds ( Rs + Rd )
where
⎛∂ I ⎞
gm ≡ ⎜ ds ⎟
⎜ ∂ Vg ⎟
⎝ ⎠Vds
is the extrinsic transconductance, and
⎛∂I ⎞
gds ≡ ⎜ ds ⎟
⎝ ∂ Vds ⎠V g
is the extrinsic output conductance.
SOLUTION 33
4.9. Let V = 0 in Eq. (3.12):
eqψ / kT
2
qn ψs
Qi = i
Na ∫ψ B E (ψ ) dψ .
In the subthreshold region, both the band bending ψ and the electric field E are mainly set by the
depletion charge. Because of the exponential factor in ψ, the above integration of inversion
charge is dominated by that over a thin surface layer within which E(ψ) is essentially constant
and can be taken outside the integral. Thus one obtains
2
kTni qψ s / kT
Qi = e ,
Es N a
where Es is the surface electric field. Note that exp(qψB/kT) << exp(qψs/kT) in weak inversion.
This equation is generally valid for nonuniform (vertically) dopings with Na being the p-
type concentration at the edge of the depletion layer. (Note that the factor Na merely reflects
the fact in Fig. 2.32 that the band bending ψs is defined with respect to the bands of the neutral
bulk region of doping Na.)
4.10. From Eq. (3.8),
dV
I ds ( y ) = μeff WQi ( y ) .
dy
Current continuity requires that Ids be a constant, independent of y. For a short-channel or a

nonuniformly-doped (laterally) MOSFET,
L dy μ W V ds Vds
∫0 Qi ( y )
= eff
I ds ∫0
dV = μeff W
I ds
.
Generalizing the result of the last exercise to laterally varying ψs and Es, one obtains
Vds 1 L dy Na L
=
I ds μeff W ∫ 0
=
Qi ( y ) μeff WkTni 2 ∫0 s
E ( y )e− qψ s ( y )/ kT dy .
Since Es(y) ≈ [Vg − Vfb − ψs(y)]/3tox is not a strong function of ψs, the exponential factor
dominates. This implies that the subthreshold current is controlled by the point of highest
barrier (lowest ψs) in the channel.
SOLUTION 34
4.11.
Electron energy level under triangular approximation, the electron energy levels are given by
2/3
⎡ 3hqEs ⎛ 3 ⎞⎤
Ej = ⎢ ⎜ j + ⎟⎥ where j = 0,1,2, L
⎢⎣ 4 2mx ⎝ 4 ⎠⎥⎦
kT ⎛ N a ⎞
ψB = ln⎜ ⎟ = 0.4679(V )
q ⎜⎝ ni ⎟⎠
2ε si
Wd = (2ψ B ) = 34.9(nm)
qN a
q Wd qN W
Es = ∫ N ( x)dx = a d = 0.537(MV cm)
ε si 0 ε si
Substituting Es and j gives,

E0 = 0.1143(eV ) , E1 = 0.2010(eV ) , E2 = 0.2717(eV ) , …
4.12.
kT ⎛ N a ⎞
ψB = ln⎜ ⎟ = 0.4083(V )
q ⎜⎝ ni ⎟⎠
Qs = −Cox (Vg − V fb − 2ψ B ) = −1.09 × 10−6 (C / cm 2 )
− Qs / q
x poly = = 0.680(nm)
N poly
qN poly x 2poly
Sinceψ p = = 35.5(mV ) , the band-bending in the gate can be neglected to the
2ε si
first order in the charge calculation.
SOLUTION 35
5.1. From Eqs. (5.4) and (5.5), the average CMOS inverter delay is
τ n +τ p CVdd ⎛ 1 1 ⎞
τ= = ⎜⎜ + ⎟⎟ .
2 4 ⎝ Wn I nsat Wp I psat ⎠
If the inverter is driving another stage with the same n- to p-width ratio and if both the n- and p-
devices have the same capacitance per unit width, the load capacitance C is proportional to (Wn
+ Wp), i.e.,
⎛ 1 1 ⎞ 1 1 W / Wn Wn / Wp
τ ∝ (Wn + Wp )⎜⎜ + ⎟⎟ = + + p + .
⎝ Wn I nsat Wp I psat ⎠ I nsat I psat I nsat I psat
The delay is a minimum when the n- to p-width ratio is such that the last two terms are equal,
i.e., when Wp/Wn = (Insat/Ipsat)1/2. Note that τn < τp for minimum delay.
5.2. For the following RC circuit,
Vdd
R V(t)
I(t) C
if the voltage source is switched to Vdd at t = 0, then
dV
Vdd = V ( t ) + RI ( t ) = V ( t ) + RC .
dt
With the initial condition V(t = 0) = 0, the solution for V(t) is
( )
V ( t ) = Vdd 1 − e− t / RC .
SOLUTION 36
The energy dissipated in R is
2
∞ Vdd ∞ 1
E = ∫ RI 2 dt = ∫ e−2 t / RC dt = CVdd ,
2
0 R 0 2
which is independent of R. The same amount of energy is stored in C.
If the voltage source is now switched to 0, one has
dV
V ( t ) + RC = 0.
dt
With the initial condition V(t = 0) = Vdd, the solution for V(t) is
V ( t ) = Vdd e− t / RC .
The energy stored in C is now all dissipated in R:

2
∞ Vdd ∞ 1
E = ∫ RI 2 dt = ∫ e−2 t / RC dt = CVdd .
2
0 R 0 2
5.3. From Eq. (3.25) and the inversion charge expression above Eq. (3.58), the transit time is
WLCox (Vg − Vt − mVds / 2) L2

τ tr = =
μ eff Cox (W / L )(Vg − Vt − mVds / 2)Vds μ eff Vds
for a MOSFET device biased in the linear region.
From Eq. (3.28) and the inversion charge expression above Eq. (3.60), the transit time is
( 2 / 3)WLCox (Vg − Vt ) 4mL2

τ tr = =
μ eff Cox (W / L )(Vg − Vt )2 / 2m 3μ eff (Vg − Vt )
for a long-channel MOSFET biased in saturation.
SOLUTION 37
5.4. From Eq. (3.79) and the inversion charge expression in Exercise 3.10, the transit time is
Qi L 1 + 2μeff (Vg − Vt ) / ( mvsat L ) + 1 / 3

τ tr ≡ =
I dsat vsat 1 + 2μeff (Vg − Vt ) / ( mvsat L ) − 1
for a short-channel MOSFET biased in saturation. The limiting value of τtr is L/vsat when the
device becomes fully velocity saturated as L → 0.
5.5. The transmission line model of contact resistance in a planar geometry is represented by
the distributed network below. The current flows from a thin resistive film (diffusion with a
sheet resistivity ρsd) into a ground plane (metal) with an interfacial contact resistivity ρc between
them (Fig. 5.16).
Metal contact
Gdx
Rdx
I
x=0 V(x) x=lc
I(x)
Following a similar approach as in Eqs. (5.23)-(5.25), one can write
dV
V ( x + dx ) − V ( x ) = dx = − I ( x ) Rdx ,
dx
and
dI
I ( x + dx ) − I ( x ) = dx = −V ( x )Gdx .
dx
Here R = ρsd/W and G = W/ρc. From the above two equations, one obtains
d2 f ρ
= RGf = sd f ,
dx 2
ρc
where f(x) = V(x) or I(x).
SOLUTION 38
5.6. The solution to the second-order differential equation in the above exercise is of the
sinh[(ρsd/ρc)1/2x] and cosh[(ρsd/ρc)1/2x] form. With the boundary condition I(x=lc) = 0 where x =
0 is the leading edge and x = lc is the far end of the contact window (see figure in Exercise 5.5),
the solution is
⎡ ρ ⎤
I ( x ) = I0 sinh ⎢ sd ( lc − x )⎥ ,
⎢⎣ ρ c ⎥⎦
where I0 is a constant multiplying factor. From the second equation in Exercise 5.5, the voltage
is
ρc dI ρc ρsd ⎡ ρ ⎤
V (x) = − = I0 cosh ⎢ sd ( lc − x )⎥ .
W dx W ⎢⎣ ρc ⎥⎦
The contact resistance is then
V ( x = 0) ρ c ρ sd ⎡ ρ sd ⎤
Rco = = coth ⎢lc ⎥.
I ( x = 0) W ⎢⎣ ρ c ⎥⎦
5.7. From Eqs. (5.43) and (5.45), τbmin < τ if
Cout + CL > 2Cout + 2 Cin CL .
This inequality is quadratic in CL1/2, which can be solved to yield
( )
2
CL > Cin + Cin + Cout .
Under this condition, the insertion of one or a few properly designed buffer stage(s) will help
reduce the overall delay.
SOLUTION 39
5.8. For an n-stage buffer with successive width ratios of k1, k2, k3, ⋅ ⋅ ⋅ ⋅ ⋅, kn, the switching
resistances are Rsw, Rsw/k1, Rsw/k1k2, Rsw/k1k2k3, ⋅ ⋅ ⋅ ⋅ ⋅, etc. The output capacitances are Cout,
k1Cout, k1k2Cout, k1k2k3Cout, ⋅ ⋅ ⋅ ⋅ ⋅, etc.; and the input capacitances are k1Cin, k1k2Cin, k1k2k3Cin,
k1k2k3k4Cin, ⋅ ⋅ ⋅ ⋅ ⋅, etc. The delay of the first stage is
τ b,1 = Rsw ( Cout + k1Cin ) ;
the delay of the second stage is
Rsw
τ b, 2 = ( k1Cout + k1k2Cin ) = Rsw ( Cout + k2Cin ) ;
k1
⋅ ⋅ ⋅ ⋅ ⋅, and so on, until the n-th stage:
τ b, n = Rsw ( Cout + kn Cin ) .
The delay of the last stage is
Rsw ⎛ CL ⎞
τ b, n +1 = ( k1k2 ⋅ ⋅ ⋅ kn Cout + CL ) = Rsw ⎜ Cout + ⎟.
k1k2 ⋅ ⋅ ⋅ kn ⎝ k1k2 ⋅ ⋅ ⋅ kn ⎠
The total buffer delay is the sum of all the stage delays, i.e.,
⎛ CL ⎞
τ b ( n ) = Rsw ⎜ ( n + 1)Cout + ( k1 + k2 + ⋅ ⋅ ⋅ ⋅ ⋅ + kn )Cin + ⎟.
⎝ k1k2 ⋅ ⋅ ⋅ ⋅ ⋅ kn ⎠
[Note that delay equations of the form Eq. (5.39) used above are defined in terms of propagation
delays through cascaded stages of identical width and loading (Fig. 5.29), and as such are not
strictly applicable to a chain of drivers with varying widths and rise times. When the input rise
time is shorter than that of the output as is the case with the above tapered drivers, the switching
resistance Rsw ≡ dτ/dCL is somewhat less than that defined in Section 5.3.1.3.]
SOLUTION 40
5.9. Take any pair of k’s (positive numbers), e.g., k1 and k2, for a given product k1k2, the sum
k1 + k2 is a minimum when k1 = k2. This can be generalized to the result that for a given
product k1k2k3 ⋅ ⋅ ⋅ kn, the sum k1 + k2 + k3 + ⋅ ⋅ ⋅ + kn is a minimum when k1 = k2 = ⋅ ⋅ ⋅ ⋅ ⋅ = kn.
Let k1 = k2 = ⋅ ⋅ ⋅ ⋅ ⋅ = kn = k, the delay in Exercise 5.8 becomes
τ b ( n ) = Rsw ⎛⎜ ( n + 1)Cout + nkCin + nL ⎟⎞ .

C
⎝ k ⎠
Differentiating it with respect to k, one finds that the above expression has a minimum
⎡ ⎛ CL ⎞
1/ ( n +1)
⎤
τ b min ( n ) = Rsw ⎢( n + 1)Cout + ( n + 1)Cin ⎜ ⎟ ⎥,
⎢⎣ ⎝ Cin ⎠ ⎥⎦
when k = (CL/Cin)1/(n+1). [Note that the same result can be obtained alternatively by partially
differentiating the delay in Exercise 5.8 with respect to k1, k2, k3, ⋅ ⋅ ⋅ ⋅ ⋅, kn respectively and
solving the n coupled equations.] τbmin(n), as expected, is reduced to Eq. (5.45) if n = 1.
5.10. To find the optimum n which gives the shortest overall delay, we let k = (CL/Cin)1/(n+1)
and treat it as a continuous variable. Then n + 1 = ln(CL/Cin)/ln(k) and the delay in Exercise
5.9 becomes
ln( CL / Cin )
τ b min ( k ) = Rsw ( Cout + kCin ) .
ln( k )
This has a minimum when dτbmin/dk = 0, which yields an equation for k:
Cout
k[ln( k ) − 1] = .
Cin
n is then the closest integer to
ln( CL / Cin )
n= − 1.
ln( k )
For typical Cout/Cin ratios not too different from unity, k is in the range of 3-5. Note that
k also gives the optimum width ratio between the successive buffer stages, i.e., k1 = k2 = ⋅ ⋅ ⋅ ⋅ ⋅ =
kn = k. Since Cout + kCin = kln(k)Cin, the minimum buffered delay can be written as
τ b min ≈ kRsw Cin ln( CL / Cin ) ,
which only increases logarithmically with load capacitance.
SOLUTION 41
5.11.
Power = CVdd2f where f = 1/(2τ) and C = Cin+Cout+CL
So Power = Vdd2/2Rsw
If (Wn, Wp) Æ (kWn, kWp),
Rsw Æ Rsw/k
Cin+Cout Æ k(Cin+Cout)
Therefore,
Delay τ Æ Rsw(Cin+Cout+CL/k)
Power Æ k times original value.
SOLUTION 42
6.1. Consider hole current flow in an n-region. Eq. (2.64) gives
dφ p
J p = − qpn μ p . (1)
dx
Now, as shown in Fig. 2.18, dφn(n-region)/dx ≈ 0 when there is negligible IR drop in the n-
region. Therefore, (1) can be rewritten as
d
J p ( n − region ) ≈ − qpn μ p
dx
( )
φ p − φn . (2)
Eq. (6.15) gives
kT ⎛ pn nn ⎞
(φ p )
− φn = ln ⎜
q ⎝ nie2 ⎠
⎟ (3)
for the n-region. Substituting (3) into (2), we obtain
⎡ kT ⎛ 1 dnn 1 dnie2 ⎞ ⎤ dp
J p ( n − region ) = qpn μ p ⎢− ⎜ − 2 ⎟ ⎥ − qDp n , (4)
⎣ q ⎝ nn dx nie dx ⎠ ⎦ dx
where we have used the Einstein relation Dp = kTμp/q. If we compare (4) with Eq. (2.55), i.e.,
dpn
J p ( n − region ) = qpn μ pE ( n − region ) − qDp , (5)
dx
⎡ kT ⎛ 1 dnn 1 dnie2 ⎞ ⎤
we have E ( n − region ) = ⎢− ⎜ − 2 ⎟⎥ .
⎣ q ⎝ nn dx nie dx ⎠⎦
6.2. Eq. (2.64) gives, for an n-region,
dφ p
J p = − qpn μ p . (1)
dx
As shown in Fig. 2.18, when there is negligible IR drop in the n-region, we have
dφn (n − region)
≈ 0. (2)
dx
SOLUTION 43
Therefore, (1) and (2) give
d (φ p − φ n )
J p ( n − region ) ≈ − qpn μ p . (3)
dx
From Eq. (6.15), for the n-region,
kT ⎛ pn nn ⎞
(φ p − φ n ) = ln ⎜ ⎟. (4)
q ⎝ nie2 ⎠
Substituting (4) into (3), and using Dp = kTμp/q, we have
nie2 d ⎛ nn pn ⎞
J p ( n − region ) ≈ − qDp ⎜ ⎟.
nn dx ⎝ nie2 ⎠
6.3.
Δ p n (x)
WE 1
polysilicon silicon
x
0
− (W E + WE1 ) −W E
E -B junction
(a) Let Δpn(−WE) be the excess hole density at the polysilicon-silicon interface. Also
let
x ' = −( x + WE ) . That is, x’ is the distance between a point inside the polysilicon layer and the
polysilicon-silicon interface. The steady-state equation governing the excess hole density is
given by the hole-equivalent of Eq. (2.115), i.e.
d 2 Δpn Δpn
− 2 = 0. (1)
dx 2 Lp
Applying (1) to hole distribution inside the polysilicon layer, we have
SOLUTION 44
d 2 Δpn Δpn
− 2 = 0. (2)
dx'2 LpE1
The general equation for (2) has the form
x ' /L pE 1
Δpn ( x' ) = Ae + Be− x ' /LpE1 . (3)
At the ohmic contact at x’ = WE1, we have
WE 1 / L pE 1 −WE 1 / L pE 1
0 = Ae + Be . (4)
At the polysilicon-silicon interface (x’ = 0), we have
Δpn ( −WE ) = A + B . (5)
Δpn ( −WE )
(4) and (5) give A= 2W / L
(6)
1 − e E 1 pE 1
2W / L pE 1
− Δpn ( −WE )e E 1
and B= 2W / L
. (7)
1 − e E 1 pE 1
Substituting (6) and (7) into (3), we have
Δpn ( x' ) = Δpn ( −WE )

[
sinh (WE1 − x' ) / LpE1 ]. (8)
sinh(WE1 / LpE1 )
(b) The hole current density entering the polysilicon layer is given by Eq. (6.36), i.e.,
J p ( x = −WE ) = − qΔpn ( −WE )S p . (9)
Using (8), this current density is also
dΔpn ( x )
J p ( x = −WE ) = − qDpE1 (10)
dx x = −WE
dΔpn ( x ' )
= qDpE1
dx' x ' = 0
Δpn ( −WE )
= − qDpE1 .
LpE1 tanh(WE1 / LpE1 )
DpE1
Equations (9) and (10) give Sp = .
LpE1 tanh(WE1 / LpE1 )
SOLUTION 45
6.4. Neglecting parasitic resistances, Eqs. (6.144) and (6.145) give
I E = − I EBO [ exp( qVBE

′ / kT ) − 1] − α R IC (1)
and IC = − ICBO [ exp( qVBC

′ / kT ) − 1] − α F I E , (2)
which can be rearranged to give
kT ⎡ I E + α R IC ⎤
′ =
VBE ln 1 −
q ⎢⎣ I EBO ⎥⎦
(3)
kT ⎡ IC + α F I E ⎤
′ =
VBC ln 1 −
q ⎢⎣ ICBO ⎥⎦
and . (4)
Now, ′ = VC′ − VE′ = (VC′ − VB′ ) + (VB′ − VE′ ) = VBE

VCE ′ − VBC
′ . (5)
Substituting (3) and (4) into (5) and rearranging, we obtain
kT ⎡ ICBO ( I EBO − I E − α R IC ) ⎤
′ =
VCE
q ⎢⎣ I EBO ( ICBO − IC − α F I E ) ⎥⎦
ln . (6)
Now, Eq. (6.90) gives α R I R0 = α F IF 0 , (7)
and Eqs. (6.142) and (6.143) give
I EBO ≡ I F 0 (1 − α Rα F ) , (8)
and ICBO ≡ I R 0 (1 − α Rα F ) . (9)
ICBO I R 0 α F
Therefore, = = . (10)
I EBO I F 0 α R
kT ⎡ α F ( I EBO − I E − α R IC ) ⎤
Substituting (10) into (6) gives ′ =
VCE ln ⎢ ⎥.
q ⎣ α R ( ICBO − IC − α F I E ) ⎦
6.5. From Exercise (6.4), we have
SOLUTION 46
′ =
VCE ln ⎢ ⎥. (1)
q ⎣ α R ( ICBO − IC − α F I E ) ⎦
Now, when emitter and collector resistances are included, we have
VCE = VC − VE = VC′ + rc IC − (VE′ + re I E ) = VCE

′ + rc IC + re ( I B + IC ) , (2)
where we have used the fact that
I B + IC + I E = 0 . (3)
Substituting (1) into (2), we obtain
VCE = ln ⎢ ⎥ + rc IC + re ( I B + IC ) . (4)
q ⎣ α R ( ICBO − IC − α F I E ) ⎦
If we neglect the saturation currents IEBO and ICBO, then using (3) we can rearrange (4) to give
kT ⎡ α F ( I B + IC − α R IC ) ⎤
VCE ≈ ln ⎢ ⎥ + rC IC + re ( I B + IC ) (5)
q ⎣ α R (α F ( I B + IC ) − IC ) ⎦
kT ⎡ I B + I C (1 − α R ) ⎤
= ln ⎢ ⎥ + re ( I B + IC ) + rc IC .
q ⎣ α R [ I B − I C (1 − α F ) / α F ] ⎦
For open-circuit collector, IC = 0, and (5) reduces to

kT ⎛ 1 ⎞
VCE ( IC = 0) = ln ⎜ ⎟ + re I B .
q ⎝α R ⎠
6.6. From Eq. (2.119), we have
sinh[(WB − x ) / LnB ]
n p − n p 0 = n p 0 [ exp( qVBE / kT ) − 1] . (1)
sinh(WB / LnB )
If we assume diffusion current only, then
dn p cosh[(WB − x ) / LnB ] ⎛ −1 ⎞
J n ( x ) = qDnB = n p 0 [ exp( qVBE / kT ) − 1] ⎜ ⎟. (2)
dx sinh(WB / LnB ) ⎝ LnB ⎠
Therefore, the base transport factor is
SOLUTION 47
−1
J ( x = WB ) ⎛ W ⎞
αT ≡ n = ⎜ cosh B ⎟ . (3)
J n ( x = 0) ⎝ LnB ⎠
For a base region of WB = 100 nm and NB = 1018 cm−3, we have, from Fig. 2.24, LnB ≈ 15 μm,
and
−1
⎛ .⎞
01
α T = ⎜ cosh ⎟ = 0.9998 .
⎝ 15 ⎠
6.7. Eq. (6.150) give the condition for collector-emitter breakdown as
γα T M ( BVCEO ) = 1 . (1)
Now, from Eqs. (6.55) and (6.140), we have
β0
= γα T , (2)
1+ β0
where β0 is the common-emitter current gain at negligible base-collector junction avalanche.

(1) and (2) give
1
M ( BVCEO ) − 1 = .
β0
SOLUTION 48
7.1. For the polysilicon-emitter model described in Exercise 6.3,
⎛ n2 ⎞ ⎛ W
GE (WE ,WE1 ) = N E ⎜ 2i ⎟ ⎜ E +
(
LpE1 tanh WE1 / LpE1 ) ⎞⎟ . (1)
⎝ nieE ⎠ ⎜⎝ D pE D pE1 ⎟
⎠
Δ p n (x)
WE 1
polysilicon silicon
x
0
− (W E + WE1 ) −W E
E -B junction
Assume the lifetime in polysilicon layer is the same as in silicon, and assume the mobility in
polysilicon is smaller than that in silicon.
(a) For non-polysilicon-emitter: NE = 1020 cm−3, WE = 300 nm, WE1 = 0. From Eq.
(6.14), we have
ni2
= exp( − ΔE gE / kT ) . (2)
nieE
2
From Fig. 6.3, we have, for NE = 1020 cm−3, ΔEgE = 92 meV. Using kT = 26 meV, we have
ni2
= exp( − ΔE gE / kT ) = 0.029 . (3)
nieE
2
From, Fig. 2.24, we have μpE = 130 cm2/V-s. Therefore,
kTμ pE
DpE = = 338
. cm2/s. (4)
q
⎛ n2 ⎞ W
Therefore, GE ( 300 nm, 0) = N E ⎜ 2i ⎟ E = 2.6 × 1013 cm4-s.
⎝ nieE ⎠ DpE
SOLUTION 49
(b) For polysilicon-emitter: NE = 1020 cm−3 and WE = 30 nm. With μpE1/μpE = 1/3, we
have
kTμ pE1
DpE1 = = 113
. cm2/s. (5)
q
Now, from Fig. 2.24, we have LpE = 0.38 μm, we have
kT L
LpE1 = μ pE1τ pE = pE = 0.22 μm. (6)
q 3
Substituting all these values into (1) we have
WE1 (nm) GE(30 nm, WE1) (cm4-s)

50 1.52×1013
100 2.66×1013
200 4.36×1013
300 5.22×1013
(c)
2.2
GE(poly)/GE(non-poly)
2
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0 50 100 150 200 250 300 350
Polysilicon-layer thickness (nm)
7.2. For a box-like profile, Eq. (7.5) gives
−1
ρB
RSbi = ⎜⎛⎝ q ∫ p p ( x )μ p ( x )dx⎞⎟⎠
WB 1
= = , (1)
0 qN B μ pWB WB
where ρB is the resistivity of the p-type base layer. Setting RSbi = 104 Ω/ , and using Fig. 2.9
for ρB, we obtain the values in the following table.
SOLUTION 50
NB (cm−3) ρB (Ω-cm) WB (nm)
6×1016 0.3 300
1×1017 0.2 200
2.9×1017 0.1 100
7×1017 0.05 50
1E+18
Base doping (cm-3)
5E+17
3E+17
2E+17
1E+17
5E+16
30 50 100 200 300 500
Base width (nm)
7.3. From Exercise 7.2, and Fig. 2.24(a), we have the following table.
WB NB (cm−3) μnB (cm2/V-s) DnB (cm2/s) tB (ps)

(nm)
50 7×1017 37.5 9.75 1.3
100 2.9×1017 505 13.1 3.8
200 1×1017 760 19.8 10.1
where we have used DnB = kTμnB/q and tB = WB2/2DnB. Also, since Fig. 2.24(a) does not have
values for NB < 1×1017 cm−3, we have dropped the corresponding base width and base doping
concentration values in the table and the following plot.
12
Base transit time (ps)
10
0
0 50 100 150 200 250
Base width (nm)
SOLUTION 51
7.4. From Eq. (6.72), we have
QpB qN BWB
VA ≈ = . (1)
CdBC CdBC
To avoid significant base widening, collector current density is kept at
JC
J C = 0.3qvsat N C , or NC = , (2)
0.3qvsat
with vsat = 107 cm/s.
(a) For a one-sided B−C junction, with VCB = 2 V, Eqs. (2.83) and (2.85) give
ε si qN Cε si J Cε si
CdBC = = = . (3)
WdBC 2(ψ bi + VCB ) 0.6vsat (ψ bi + VCB )
We can obtain ψbi from Fig. 2.15 (or set it to about 0.95 V as an approximation) and use εsi =
1.04×10−12 F/cm to generate the following table.
JC (mA/μm2) NC (cm−3) ψbi (V) CdBC (fF/μm2)

0.1 2.08×1016 0.925 0.243
0.2 4.16 0.945 0.343
0.3 6.24 0.95 0.420
0.4 8.32 0.96 0.484
0.5 1.04×1017 0.97 0.540
0.8 1.66 0.98 0.659
1.0 2.08 0.99 0.761
2.0 4.16 1.0 1.07
3.0 6.24 1.02 1.31
4.0 8.32 1.02 1.52
5.0 1.04×1018 1.02 1.69
SOLUTION 52
3
2
CdBC (fF/um2)
1
0.5
0.3
0.2
0.1
0.1 0.2 0.3 0.5 1 2 3 5
JC (mA/um2)
(b) For a base design with qNBWB = 1.6×10−6 C/cm2, we have the following table
JC (mA/μm2) CdBC (fF/μm2) VA (V)

0.1 0.243 65.8
0.2 0.343 46.6
0.3 0.420 38.1
0.4 0.484 33.1
0.5 0.540 29.6
0.8 0.659 24.3
1.0 0.761 21.0
2.0 1.07 15.0
3.0 1.31 12.2
4.0 1.52 10.5
5.0 1.69 9.5
100
50
VA (V)
30
20
10
5
0.1 0.2 0.3 0.5 1 2 3 5
JC (mA/um2)
SOLUTION 53
7.5. An npn transistor with WB = 500 nm and negligible heavy doping effect in base layer,
and a graded base profile of
N B ( x ) = A − αx , (1)
A = N B ( 0) = 2 × 1017 cm−3, (2)
N B (WB ) = 2 × 1016 cm−3 (3)
A − N B (WB )
and α= = 3.6 × 1021 cm−4. (4)
WB
From Eq. (7.15), the electric field due to the dopant distribution is
kT 1 dN B kT α
E ( dopant ) = =− . (5)
q N B ( x ) dx q NB(x)
We have the following table and plot of results.
x (nm) NB (cm−3) E (V/cm)

0 2.0E17 -468
50 1.82E17 -514
100 1.64E17 -571
150 1.46E17 -641
200 1.28E17 -731
250 1.10E17 -851
300 9.2E16 -1017
350 7.4E16 -1265
400 5.6E16 -1671
450 3.8E16 -2463
500 2.0E16 -4680
0
Electric field (V/cm)
-1,000
-2,000
-3,000
-4,000
-5,000
0 100 200 300 400 500 600
x (nm)
SOLUTION 54
7.6. An npn transistor with WB = 100 nm and non-negligible heavy doping effect in base
layer, and a graded base profile of
N B ( x ) = A − αx , (1)
A = N B ( 0) = 5 × 1018 cm−3, (2)
N B (WB ) = 5 × 1017 cm−3 (3)
A − N B (WB )
and α= = 4.5 × 1023 cm−4. (4)
WB
From Eq. (7.15), the electric field due to the dopant distribution is
kT 1 dN B kT α
E ( dopant ) = =− , (5)
q N B ( x ) dx q NB(x)
and the field due to heavy-doping effect is
1 dΔE gB
E ( HD ) = − . (6)
q dx
From Eq. (6.17), the bandgap narrowing in the base layer is
(
ΔEg ( N a ) = 9 × F + F 2 + 0.5 ) meV (7)
where F = ln( N B / 1017 ) . (8)
⎡ F ⎤ α
Therefore, we have E ( HD ) = 9 × 10−3 ⎢1 + ⎥ V/cm. (9)
⎣ F 2 + 0.5 ⎦ N B
We have the following table and plot of the results.
SOLUTION 55
x (nm) NB (cm−3) E (dopant)(V/cm) E (HD)(V/cm) E (Tot)(V/cm)
0 5.0E18 -2340 1607 -733
10 4.55 -2571 1765 -806
20 4.1 -2854 1958 -895
30 3.65 -3205 2198 -1007
40 3.2 -3656 2506 -1151
50 2.75 -4255 2913 -1342
60 2.3 -5087 3479 -1608
70 1.85 -6324 4317 -2008
80 1.4 -8357 5687 -2670
90 9.5E17 -12316 8330 -3985
100 5.0 -23400 15516 -7884
20,000
Electric field (V/cm)
10,000
-10,000
E(dopant) E(HD) E(Total)
-20,000
0 20 40 60 80 100 120
x (nm)
7.7. Polysilicon emitter with NE = 1020 cm−3, AE = 1 μm2, single-crystal region thickness WE =
30 nm, polysilicon layer thickness WE1 = 200 nm, metal-poly contact resistivity ρcontact = 2×10−7
Ω-cm2. From Fig. 2.9, we have for the emitter doping concentration ρsingle-crystal = 8×10−4 Ω-
cm. Therefore, we have ρpolysilicon = 24×10−4 Ω-cm. From the information, we can calculate
the emitter resistance components.
W
rsin gle − crystal = ρ sin gle − crystal E = 0.24Ω
AE
WE1
rpolysilicon = ρ polysilicon = 4.8Ω
AE
rcontact = ρ contact / AE = 20Ω
SOLUTION 56
q
7.8. Eq. (6.33) gives JC 0 = . (1)
WB pp
∫ 0 DnB nieB
2
dx
For uniformly doped base with NB = 1018 cm−3, WB = 100 nm, and assuming low current levels
so that pp = NB, we have
qDnB nieB
2
JC0 = . (2)
N BWB
(a) Neglecting heavy doping effect, we have from Table 2.1 nieB = ni = 1.0×1010 cm−3.
Using the mobility values in Fig. 2.24a, we have
kTμ nB
DnB = = 8.58 cm2/s.
q
Putting these values into (1), we have JC0 (neglecting HD effect) = 1.37×10−11 A/cm2.
(b) When heavy doping effect is included, we have from Eq. (6.14) and Fig. 6.3,
nieB
2
= ni2 exp( ΔE gB / kT ) = ni2 exp( 43 / 26) = 5.23ni2 . (3)
That is, JC0 is 5.23 times larger, or JC0 = 7.19×10−11 A/cm2.
7.9.
xjE
Metal or metal silicide
n+
n+ emitter poly
p Ge
ΔEg0 = ΔEgmax
x
0 WB0
WB
Wcap
Fig. 7.11. Schematic illustrating the emitter and base regions of a polysilicon-emitter SiGe-
base bipolar transistor having a constant Ge distribution in the base. The emitter depth xjE is
assumed to be smaller than the thickness Wcap of the starting Ge-free layer, resulting in a
small region of thickness Wcap − xjE near the emitter end without Ge.
SOLUTION 57
−1
J C 0 (SiGe, x jE ) ⎧⎪⎛ W − x jE ⎞ ⎛ WB 0 − Wcap ⎞ ⎫
= γ η ⎨⎜ cap ⎟+⎜ ⎟ exp(− ΔEg 0 / kT )⎪⎬ (1)
J C 0 (Si, x jE )
x jE <Wcap
⎪⎩⎜⎝ WB 0 − x jE ⎟ ⎜W −x ⎟
⎠ ⎝ B0 jE ⎠ ⎪⎭
−1
⎧⎪⎛ Wcap − x jE ⎞ ⎛ (WB 0 − x jE ) − (Wcap − x jE ) ⎞ ⎫
= γ η ⎨⎜ ⎟+⎜ ⎟ exp(− ΔE g 0 / kT )⎪⎬
⎪⎩⎜⎝ WB 0 − x jE ⎟ ⎜
⎠ ⎝ WB 0 − x jE ⎟
⎠ ⎪⎭
−1
⎧⎪⎛ Wcap − x jE ⎞ ⎛ Wcap − x jE ⎞ ⎫
= γ η ⎨⎜ ⎟ + ⎜1 − ⎟ exp(− ΔE g 0 / kT )⎪⎬ .
⎪⎩⎜⎝ WB 0 − x jE ⎟ ⎜ W −x
⎠ ⎝ B0 jE
⎟
⎠ ⎪⎭
200
dE/kT=0.5 dE/kT=1 dE/kT=2.5 dE/kT=5
100
50
Current Ratio
20
10
5
2
1
0 0.2 0.4 0.6 0.8 1
(Wcap − x jE )/(WB0 − xjE)
V A (SiGe, x jE ) ⎛ Wcap − x jE ⎞ ⎛ ΔE ⎞ ⎛ W − Wcap ⎞

=⎜ ⎟ exp⎜ g 0 ⎟ + ⎜ B 0
⎜ kT ⎟ ⎜ W − x
⎟ (2)
V A (Si, x jE ) ⎜W −x ⎟ ⎟
x jE <Wcap ⎝ B0 jE ⎠ ⎝ ⎠ ⎝ B0 jE ⎠
⎛ Wcap − x jE ⎞ ⎛ ΔE ⎞ ⎛ (W − x jE ) − (Wcap − x jE ) ⎞
=⎜ ⎟ exp⎜ g 0 ⎟ + ⎜ B 0
⎜ kT ⎟ ⎜
⎟
⎜W −x ⎟ W − x ⎟
⎝ B0 jE ⎠ ⎝ ⎠ ⎝ B0 jE ⎠
⎛ Wcap − x jE ⎞ ⎛ ΔE ⎞ ⎛ W − x jE ⎞
=⎜ ⎟ exp⎜ g 0 ⎟ + ⎜1 − cap
⎜ kT ⎟ ⎜ W − x ⎟
⎟
⎜W −x ⎟
⎝ B0 jE ⎠ ⎝ ⎠ ⎝ B 0 jE ⎠
200
dE/kT=0.5 dE/kT=1 dE/kT=2.5 dE/kT=5
100
Early Voltage Ratio
50
20
10
5
2
1
0 0.2 0.4 0.6 0.8 1
(Wcap − x jE )/(WB0 − xjE)
SOLUTION 58
7.10. The base bandgap narrowing parameter corresponding to Fig. 7.30 is given by
n+ Ge
n+ emitter poly
ΔEgmax = ΔEg0
p
x
0 WB
WB1
x
ΔE gB ,SiGe ( x) = ΔE g 0 x < WB1 (1)
WB1
= ΔE g 0 WB > x > WB1.
J C 0 (SiGe) γ η WB
≈ WB
. (2)
J C 0 (Si)
∫ 0
exp[− ΔE gB ,SiGe ( x) / kT ]dx
Using (1), we have

WB ⎛ − ΔE gB ,SiGe ( x ) ⎞ WB 1 ⎛ − xΔE g 0 ⎞ WB ⎛ − ΔE g 0 ⎞
∫
0
exp⎜⎜
⎝ kT
⎟⎟dx = ∫ exp⎜⎜
⎠ 0
⎝ WB1kT ⎠
⎟⎟dx + ∫ exp⎜⎜
WB 1
⎝ kT ⎠
⎟⎟dx (3)
=
kTWB1
ΔE g 0
[ ]
1 − exp(− ΔE g 0 / kT ) + (WB − WB1 )exp(− ΔE g 0 / kT ) .
−1
⎧⎪ kT WB1 ⎛ W ⎞ ⎫⎪
J C 0 (SiGe)
= γη ⎨ [ ]
1 − exp(− ΔE g 0 / kT ) + ⎜⎜1 − B1 ⎟⎟ exp(− ΔE g 0 / kT )⎬ . (4)
J C 0 (Si) ⎪⎩ ΔE g 0 WB ⎝ WB ⎠ ⎪⎭
Similarly, the VA ratio can be obtained from Eq. (7. 35), namely
V A (SiGe) exp(ΔE g max kT ) WB

V A (Si)
=
WB ∫0 exp[−ΔE gB,SiGe ( x) / kT ]dx . (5)
SOLUTION 59
V A (SiGe) exp(ΔE g 0 / kT ) ⎧⎪ kTWB1 ⎫⎪
= ⎨ [ ]
1 − exp(− ΔE g 0 / kT ) + (WB − WB1 )exp(− ΔE g 0 / kT )⎬ (6)
V A (Si) WB ⎪⎩ ΔE g 0 ⎪⎭
⎛ W ⎞
=
kT WB1
ΔE g 0 WB
[ ]
exp(ΔE g 0 / kT ) − 1 + ⎜⎜1 − B1 ⎟⎟ .
⎝ WB ⎠
The tB ratio can be obtained from Eq. (7.40), namely
t B (SiGe) 2 WB WB
t B (Si)
=
η WB2 ∫0
exp[ΔE gB , SiGe ( x) / kT ]∫ exp[−ΔE gB , SiGe ( x' ) / kT ]dx' dx .
x
(7)
The integration wrt x can be written as the sum of two parts, with one part for x < WB1 and
another part for x > WB1. That is, we have
WB WB
∫0
exp[ΔE gB ,SiGe ( x) / kT ]∫ exp[−ΔE gB ,SiGe ( x' ) / kT ]dx' dx
x
(8)
WB 1 ⎛ ΔE gB ,SiGe ( x) ⎞ WB ⎛ − ΔE gB ,SiGe ( x' ) ⎞
=∫ exp⎜⎜ ⎟⎟ ∫ exp⎜⎜ ⎟⎟dx' dx
0
⎝ kT ⎠ x
⎝ kT ⎠
WB ⎛ ΔE gB ,SiGe ( x) ⎞ WB ⎛ − ΔE gB ,SiGe ( x' ) ⎞
+ ∫ exp⎜⎜ ⎟⎟ ∫ exp⎜⎜ ⎟⎟dx' dx .
WB 1
⎝ kT ⎠ x
⎝ kT ⎠
For the first integral on RHS of (8), we have
W B1 ⎛ ΔE ( x) ⎞ WB ⎛ − ΔEgB , SiGe ( x' ) ⎞

∫0
exp⎜⎜ gB , SiGe ⎟⎟ ∫ exp⎜⎜
⎝ kT ⎠x ⎝ kT
⎟⎟dx' dx
⎠
(9)
W B1 ⎛ ΔE ( x ) ⎞⎧ W B1 ⎛ − ΔEgB , SiGe ( x' ) ⎞ WB ⎛ − ΔEgB , SiGe ( x' ) ⎞⎫

= ∫ exp⎜⎜ gB , SiGe ⎟⎟⎨∫ exp⎜⎜ ⎟⎟ + ∫ exp⎜⎜ ⎟⎟⎬dx' dx
0
⎝ kT ⎠⎩ x ⎝ kT ⎠ WB1 ⎝ kT ⎠⎭
WB1 ⎛ ΔE ( x) ⎞⎧⎪W kT ⎛ − ΔEg 0 x ⎞ ⎡ W kT ⎤ ⎛ − ΔEg 0 ⎞⎫⎪
= ∫ exp⎜⎜ gB , SiGe ⎟⎟⎨ B1 exp⎜⎜ ⎟⎟ + ⎢(WB − WB1 ) − B1 ⎥ exp⎜⎜ ⎟⎟⎬dx
0
⎝ kT ⎠⎪⎩ ΔEg 0 ⎝ kTWB1 ⎠ ⎣⎢ ΔEg 0 ⎦⎥ ⎝ kT ⎠⎪⎭
WB21kT ⎡ W kT ⎤ ⎡W kT ⎤ ⎡ ⎛ − ΔE g 0 ⎞⎤
= + ⎢(WB − WB1 ) − B1 ⎥ ⎢ B1 ⎥ ⎢1 − exp⎜⎜ ⎟⎟⎥ .
ΔE g 0 ⎢⎣ ΔEg 0 ⎦⎥ ⎣⎢ ΔE g 0 ⎦⎥ ⎣ ⎝ kT ⎠⎦
For the second integral on RHS of (8), we have
WB ⎛ ΔE ( x) ⎞ WB ⎛ − ΔE gB , SiGe ( x' ) ⎞
∫W B1
exp⎜⎜ gB , SiGe ⎟⎟ ∫ exp⎜⎜
⎝ kT ⎠x ⎝ kT
⎟⎟dx' dx
⎠
(10)
⎛ ΔE ( x) ⎞ ⎛ − ΔEg 0 ⎞
⎟⎟(WB − x )dx
WB
= ∫ exp⎜⎜ gB , SiGe ⎟⎟ exp⎜⎜
W B1
⎝ kT ⎠ ⎝ kT ⎠
SOLUTION 60
=
1
(WB − WB1 )2 .
2
Therefore, (7) gives
t B (SiGe) 2 ⎧⎪WB21kT ⎡ WB1kT ⎤ ⎡WB1kT ⎤ ⎡ ⎛ − ΔEg 0 ⎞⎤ 1 ⎫

2⎪
= ⎨ + ⎢ (W − W ) − ⎥ ⎢ ⎥ ⎢1 − exp⎜
⎜ kT ⎟⎥ 2 B ⎟ + (W − W )
b1 ⎬
t B (Si) η WB2 ⎪⎩ ΔEg 0 ⎢⎣ B B1
ΔE g 0 ⎦⎥ ⎣⎢ ΔEg 0 ⎦⎥ ⎣ ⎝ ⎠⎦ ⎪⎭
1 ⎧⎪⎛ W ⎞ 2kT ⎛ WB1 ⎞ ⎫⎪

2 2
= ⎨⎜⎜1 − B1 ⎟⎟ + ⎜ ⎟ ⎬
η ⎪⎝ WB ⎠ ΔE g 0 ⎜⎝ WB ⎟⎠ ⎪
⎩ ⎭
1 ⎧⎪ 2kT WB1 ⎛⎜ WB1 kT WB1 ⎞⎟ ⎡ ⎛ − ΔE g 0 ⎞⎤ ⎫⎪
+ ⎨ 1− − ⎢1 − exp⎜⎜ ⎟⎟⎥ ⎬ .
η ⎪⎩ ΔE g 0 WB ⎜⎝ WB ΔE g 0 WB ⎟⎠ ⎣⎢ ⎝ kT ⎠⎦⎥ ⎪⎭
7.11. The base bandgap narrowing parameter corresponding to Fig. 7.17 is given by
ΔEgmax
xje Ge
n+ emitter poly
n+
ΔEg0
p
x
0 WB0
WB
Wcap
⎛ x − Wcap ⎞
ΔE gB , SiGe ( x) = ΔEg 0 + ⎜ ⎟(ΔEg max − ΔEg 0 ) x > Wcap (1)
⎜W −W ⎟
⎝ B 0 cap ⎠
= ΔE g 0 x < Wcap.
From Eq. (7.30) we have
J C 0 (SiGe, x jE ) γη (WB 0 − x jE )
= WB 0
. (2)
J C 0 (Si, x jE )
∫
x jE
exp[−ΔE gB ,SiGe ( x) / kT ]dx
The integral in (2) can be written as
WB 0 ⎛ − ΔE gB ,SiGe ( x) ⎞ Wcap ⎛ − ΔE gB ,SiGe ( x) ⎞ WB 0 ⎛ − ΔE gB ,SiGe ( x) ⎞

∫
x jE
exp⎜⎜
⎝ kT
⎟⎟dx = ∫ exp⎜⎜
⎠ x jE
⎝ kT
⎟⎟dx + ∫ exp⎜⎜
⎠ Wcap
⎝ kT
⎟⎟dx . (3)
⎠
SOLUTION 61
The first integral on RHS of (3) is
Wcap ⎛ − ΔE gB ,SiGe ( x) ⎞ ⎛ − ΔE g 0 ⎞
∫x jE
exp⎜⎜
⎝ kT
⎟⎟dx = (Wcap − x jE ) exp⎜⎜
⎠ ⎝ kT ⎠
⎟⎟ . (4)
The second integral in (2) is
WB 0 ⎛ − ΔE gB ,SiGe ( x) ⎞ WB 0 ⎪⎧ − ΔE g 0 (ΔE g max − ΔE g 0 )( x − Wcap ) ⎪⎫

∫Wcap
exp⎜⎜
⎝ kT
⎟⎟dx = ∫ exp⎨
⎠ Wcap
⎪⎩ kT
−
kT (WB 0 − Wcap )
⎬dx
⎪⎭
(5)
⎛ − kT ⎞ ⎡ ⎛ − (ΔE g max − ΔE g 0 ) ⎞ ⎤ ⎛ − ΔE g 0 ⎞
= (WB 0 − Wcap )⎜ ⎟ ⎢exp⎜
⎜ ⎟⎟ − 1⎥ exp⎜⎜ ⎟⎟ .
⎜ ΔE ⎟
⎝ g max − ΔE g 0 ⎠ ⎣ ⎝ kT ⎠ ⎦ ⎝ kT ⎠
Substituting (4) and (5) into (3), we have from (2)
J C 0 (SiGe, x jE ) γη exp(ΔE g 0 / kT )
= .
J C 0 (Si, x jE ) ⎛ Wcap − x jE ⎞ ⎛ WB 0 − Wcap ⎞⎛ kT ⎞⎡ ⎛ ΔE − ΔE g max ⎞⎤
⎜ ⎟+⎜ ⎟⎜ ⎟ ⎢1 − exp⎜ g 0 ⎟⎟⎥
⎜W −x ⎟ ⎜ W −x ⎟⎜ ΔE ⎟ ⎜
⎝ B0 jE ⎠ ⎝ B0 jE ⎠⎝ g max − ΔE g 0 ⎠ ⎣ ⎝ kT ⎠⎦
7.12. For evaluating the transit time ratio, it is convenient to write the base bandgap narrowing
parameter in the form
ΔEgmax
xje Ge
n+ emitter poly
n+
ΔEg0
p
x
0 WB0
WB
Wcap
′ ,SiGe ( x) ,
ΔE gB ,SiGe ( x) = ΔE g 0 + ΔE gB (1)
⎛ x − Wcap ⎞
where ′ , SiGe ( x) = ⎜
ΔEgB ⎟(ΔEg max − ΔEg 0 ) x > Wcap (2)
⎜W −W ⎟
⎝ B0 cap ⎠
=0 x < Wcap.
SOLUTION 62
The transit time ratio is, from Eq. (7.40)
t B ( SiGe, x jE )
(3)
t B ( Si, x jE )
x jE <Wcap
2 WB 0 WB 0
=
η (WB 0 − x jE ) 2 ∫ x jE
exp[ΔE gB ,SiGe ( x) / kT ]∫
x
exp[− ΔE gB , SiGe ( x' ) / kT ]dx' dx
2 WB 0 WB 0
=
η (WB 0 − x jE ) 2 ∫ x jE
exp[ΔE ' gB , SiGe ( x) / kT ]∫
x
exp[−ΔE ' gB , SiGe ( x' ) / kT ]dx' dx
2 Wcap WB 0
=
η (WB 0 − x jE ) 2 ∫ x jE
x
exp[− ΔE ' gB , SiGe ( x' ) / kT ]dx' dx
2 WB 0 WB 0
+
η (WB 0 − x jE ) 2 ∫
Wcap
exp[ΔE ' gB ,SiGe ( x) / kT ]∫
x
exp[− ΔE ' gB , SiGe ( x' ) / kT ]dx' dx .
The first double integral in (3) is for x < Wcap, and its value is
2 Wcap WB 0
η (WB 0 − x jE ) 2 ∫ x jE
x
exp[− ΔE ' gB , SiGe ( x' ) / kT ]dx' dx (4)
2 Wcap WB 0
=
η (WB 0 − x jE ) 2 ∫ ∫ x jE x
2 Wcap Wcap
=
η (WB 0 − x jE ) 2 ∫ ∫ x jE x
2 Wcap WB 0
+
η (WB 0 − x jE ) 2 ∫ ∫
x jE Wcap
2 Wcap
=
η (WB 0 − x jE ) 2 ∫
x jE
(Wcap − x)dx
2 Wcap WB 0 ⎡ ( x'−Wcap )(ΔEg max − ΔEg 0 ) ⎤

+
η (WB 0 − x jE ) 2 ∫x jE ∫Wcap ⎢⎢−
exp
(W − W ) kT
⎥ dx' dx
⎥⎦
⎣ B 0 cap
2(Wcap − x jE ) 2
=
η (WB 0 − x jE ) 2
2(WB 0 − Wcap )(Wcap − x jE ) ⎛ ⎞⎡ ΔE − ΔEg max ⎞⎤
+ ⎜ kT ⎟ ⎢1 − exp⎛⎜ g 0 ⎟⎟⎥ .
η (WB 0 − x jE ) 2 ⎜ ΔE − ΔE ⎟ ⎜ kT
⎝ g max g 0 ⎠⎣ ⎝ ⎠⎦
The second double integral in (3) is for x > Wcap, and its value is
2 WB 0 WB 0
η (WB 0 − x jE ) 2 ∫ Wcap
x
exp[− ΔE ' gB , SiGe ( x' ) / kT ]dx' dx (5)
SOLUTION 63
2 ⎡ ( x − Wcap )(ΔE g max − ΔE g 0 ) ⎤ WB 0
WB 0 ⎡ ( x'−Wcap )(ΔE g 0 − ΔE g max ) ⎤
= ∫Wcap ⎢ (WB 0 − Wcap )kT
η (WB 0 − x jE ) 2
exp ⎢ ⎥ ∫
⎥⎦ x
exp ⎢
⎢⎣ (WB 0 − Wcap )kT
⎥ dx' dx
⎥⎦
⎣
2(WB 0 − Wcap ) 2 ⎛ kT ⎞⎧⎪ ⎛ kT ⎞⎡ ⎛ ΔE − ΔE g max ⎞⎤ ⎫⎪
= ⎜ ⎟⎨1 − ⎜ ⎟ ⎢1 − exp⎜ g 0 ⎟⎟⎥ ⎬ .
η (WB 0 − x jE ) 2 ⎜⎝ ΔE g max − ΔE g 0 ⎟⎠⎪⎩ ⎜⎝ ΔE g max − ΔE g 0 ⎟⎠ ⎢⎣ ⎜ kT
⎝ ⎠⎥⎦ ⎪⎭
Substituting (4) and (5) into (3), we have
t B ( SiGe, x jE ) 2(Wcap − x jE ) 2
=
t B ( Si, x jE )
x jE <Wcap
η (WB 0 − x jE ) 2
2(WB 0 − Wcap )(Wcap − x jE ) ⎛ ⎞⎡ ΔE − ΔEg max ⎞⎤

+ ⎜ kT ⎟ ⎢1 − exp⎛⎜ g 0 ⎟⎟⎥
η (WB 0 − x jE ) 2 ⎜ ΔE − ΔE ⎟ ⎜ kT
⎝ g max g 0 ⎠⎣ ⎝ ⎠⎦
2(WB 0 − Wcap ) 2 ⎛ kT ⎞⎧⎪ ⎛ kT ⎞⎡ ⎛ ΔE − ΔE g max ⎞⎤ ⎫⎪

+ ⎜ ⎟⎨1 − ⎜ ⎟ ⎢1 − exp⎜ g 0 ⎟⎟⎥ ⎬ .
2 ⎜
η (WB 0 − x jE ) ⎝ ΔE g max − ΔE g 0 ⎟⎪ ⎜ ΔE ⎟⎢ ⎜
⎠⎩ ⎝ g max − ΔE g 0 ⎠⎣ ⎝ kT ⎠⎦⎥ ⎪⎭
7.13.
n+ Ge
n+ emitter poly
ΔEgmax = ΔEg0
p
x
0 WB
WB1
x
ΔE gB ,SiGe ( x) = ΔE g 0 x < WB1 (1)
WB1
= ΔE g 0 WB > x > WB1.
−1
⎧⎪ kT WB1 ⎛ W ⎞ ⎫⎪
J C 0 (SiGe)
= γη ⎨ [ ]
1 − exp(− ΔE g 0 / kT ) + ⎜⎜1 − B1 ⎟⎟ exp(− ΔE g 0 / kT )⎬ (2)
J C 0 (Si) ⎪⎩ ΔE g 0 WB ⎝ WB ⎠ ⎪⎭
SOLUTION 64
100000
dE/kT=2.5 dE/kT=5 dE/kT=10
10000
Current Ratio
1000
100
10
1
0 0.2 0.4 0.6 0.8 1
WB1/WB
⎛ W ⎞
V A (SiGe)
V A (Si)
=
kT WB1
ΔE g 0 WB
[
exp(ΔE g 0 / kT ) − 1 + ⎜⎜1 − B1 ⎟⎟ . ] (3)
⎝ WB ⎠
10000
3000
Early Voltage Ratio
1000
dE/kT=2.5
300
dE/kT=5
100
30 dE/kT=10
10
3
1
0 0.2 0.4 0.6 0.8 1
WB1/WB
t B (SiGe) 1 ⎧⎪⎛ WB1 ⎞ 2kT ⎛ WB1 ⎞ ⎫⎪

2 2
= ⎨⎜⎜1 − ⎟ + ⎜ ⎟ ⎬ (4)
t B (Si) η ⎪⎝ WB ⎟⎠ ΔE g 0 ⎜⎝ WB ⎟⎠ ⎪
⎩ ⎭
1 ⎧⎪ 2kT WB1 ⎛⎜ WB1 kT WB1 ⎞⎟ ⎡ ⎛ − ΔE g 0 ⎞⎤ ⎫⎪
+ ⎨
⎜
1− −
⎟ ⎢1 − exp⎜⎜ ⎟⎟⎥ ⎬ .
η ⎪⎩ ΔE g 0 WB ⎝ WB ΔE g 0 WB ⎠ ⎣⎢ ⎝ kT ⎠⎦⎥ ⎪⎭
1
0.9 dE/kT=2.5 dE/kT=5 dE/kT=10
Transit Time Ratio
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1
WB1/WB
SOLUTION 65
7.14. The Ge profile of interest, from Fig. 7.23, is
Electron flow
(normal) Electron flow
(reverse)
ΔEgmax
Ge
B
n+ p n+
x
0 WB
The current ratio, from Eq. (7.30) is
J C 0 (SiGe) γ η WB
= WB
. (1)
J C 0 (Si)
∫0
exp[− ΔE gB ,SiGe ( x) / kT ]dx
For the forward current ratio, we use in (1)
xΔE g max
ΔE gB , SiBe, FORWARD ( x) = . (2)
WB
For the reverse current ratio, we should use in (1)
(WB − x)ΔE g max

ΔE gB , SiBe, REVERSE ( x) = (3)
WB
J C 0 (SiGe, reverse) γ η WB
= (4)
J C 0 (Si, reverse) WB ⎛ ( x − WB )ΔE g max ⎞
∫0
exp⎜⎜
⎝ WB kT
⎟⎟dx
⎠
γ η WB
=
⎛ − ΔE g max ⎞ WB ⎛ xΔE g max ⎞
exp⎜⎜ ⎟⎟ ∫ exp⎜⎜ ⎟⎟dx
⎝ kT ⎠ ⎝ WB kT ⎠
0
SOLUTION 66
γη WB
=
⎛ − ΔE g max ⎞⎛ WB kT ⎞ ⎡ ⎛ ΔE g max ⎞ ⎤
exp⎜⎜ ⎟⎟⎜ ⎟ ⎢exp⎜ ⎟⎟ − 1⎥
⎜ ⎟ ⎢ ⎜ kT
⎝ kT ⎠⎝ ΔE g max ⎠⎣ ⎝ ⎠ ⎥⎦
γ η ΔE g max / kT
= .
1 − exp(− ΔE g max / kT )
7.15. From Fig. 7.25, the two transistors of interest are:
ΔEgmax ΔEgmax

Ge
Ge
B B
n+ emitter poly
n+ emitter poly
n+ n+
p p
x x
0 WB 0 WB
(a) (b)
From Eq. (7.35), we have for transistor b,
⎛ ΔE gB ,SiGe (WB ) ⎞
exp⎜⎜ ⎟⎟
V A (SiGe, b) DnB (SiGe, WB ) γ (WB ) ⎝ kT ⎠ WB exp[−ΔE
V A (Si, b)
=
DnB (SiGe) γ WB ∫0 gB , SiGe ( x ) / kT ]dx (1)
⎛ ( x − W B )ΔE g max ⎞
DnB (SiGe, WB ) γ (WB ) 1 WB
=
DnB (SiGe) 0
exp⎜⎜
γ WB
⎝ W B kT ⎠
∫
⎟⎟dx
D (SiGe, WB ) γ (WB ) exp(− ΔE g max / kT ) WB ⎛ xΔE g max ⎞

= nB
DnB (SiGe) γ WB ∫0 ⎜⎜⎝ WB kT ⎟⎟⎠dx
exp
D (SiGe, WB ) γ (WB ) exp(− ΔE g max / kT ) WB ⎛ xΔE g max ⎞

= nB
DnB (SiGe) γ WB ∫0
exp⎜⎜
⎝ WB kT ⎠
⎟⎟dx
DnB (SiGe, WB ) γ (WB ) exp(− ΔE g max / kT ) WB kT

=
DnB (SiGe) γ WB ΔE g max
exp(ΔE g max / kT ) − 1 [ ]
≈
kT
ΔE g max
[
exp(−ΔE g max / kT ) exp(ΔE g max / kT ) − 1 . ]
The VA ratio for transistor a is given by Eq. (7.36), namely
SOLUTION 67
V A (SiGe, a)
V A (Si, a )
=
kT
ΔE g max
[
exp(ΔE g max / kT ) − 1 . ] (2)
Since VA(Si, a) = VA(Si, b), we have
V A (SiGe, b) ⎛ − ΔE g max ⎞
= exp⎜⎜ ⎟⎟ . (3)
V A (SiGe, a) ⎝ kT ⎠
From Eq. (7.40), we have for transistor b
t B (SiGe, b) 2 WB WB
t B (Si, b)
=
η WB2 ∫0
exp[ΔE gB , SiGe ( x) / kT ]∫ exp[−ΔE gB , SiGe ( x' ) / kT ]dx' dx
x
(4)
2 WB WB ⎡ ( x'−WB )ΔE g max ⎤

=
η WB20 ∫
exp[ΔE gB , SiGe ( x) / kT ]∫ exp ⎢
x
⎣ WB kT
⎥ dx' dx
⎦
2 WB kTWB ⎡ ⎛ − ΔE g max ⎞ ⎛ x ΔE g max ⎞⎤
2 ∫0 ⎢1 − exp⎜⎜
= ⎟⎟ exp⎜⎜ ⎟⎟⎥ exp[ΔE gB , SiGe ( x) / kT ]dx
η WB ΔE g max ⎢⎣ ⎝ kT ⎠ ⎝ WB kT ⎠⎥⎦
2 WB kTWB ⎡ (WB − x)ΔE g max ⎤ 2 WB kTWB
2 ∫0
η WB2 ∫0 ΔE g max
= exp ⎢ ⎥ dx − dx
η WB ΔE g max ⎣ WB kT ⎦
2 kTWB ⎧⎪ kTWB ⎛ ΔE g max ⎞ ⎡ ⎛ − ΔE g max ⎞⎤ ⎫⎪
= ⎨ exp⎜⎜ ⎟⎟ ⎢1 − exp⎜⎜ ⎟⎟⎥ − W B ⎬
η WB2 ΔE g max ⎪⎩ ΔE g max ⎝ kT ⎠ ⎢⎣ ⎝ kT ⎠⎥⎦ ⎪⎭
2 kT ⎧⎪ kT ⎛ ΔE g max ⎞ ⎡ ⎛ − ΔE g max ⎞⎤ ⎫⎪
= ⎨ exp⎜⎜ ⎟⎟ ⎢1 − exp⎜⎜ ⎟⎟⎥ − 1⎬ .
η ΔE g max ⎪⎩ ΔE g max ⎝ kT ⎠ ⎣⎢ ⎝ kT ⎠⎦⎥ ⎪⎭
From Eq. (7.41), we have for transistor a
2kT ⎪⎧ ⎫⎪
t B (SiGe, a )
= ⎨1 −
kT
η ΔE g max ⎪⎩ ΔE g max
[
1 − exp(− ΔE g max / kT ) ⎬ . ] (5)
t B (Si, a ) ⎪⎭
Therefore, we have
kT ⎛ ΔE g max ⎞ ⎡ ⎛ − ΔE g max ⎞⎤
exp⎜⎜ ⎟⎟ ⎢1 − exp⎜⎜ ⎟⎟⎥ − 1
t B (SiGe, b) ΔE g max ⎝ kT ⎠ ⎣ ⎝ kT ⎠⎦
= . (6)
t B (SiGe, a ) kT ⎡ ⎛ − ΔE g max ⎞⎤
1− ⎢1 − exp⎜⎜ ⎟⎟⎥
ΔE g max ⎣ ⎝ kT ⎠⎦
SOLUTION 68
8.1. Ignoring the collector-substrate capacitance term, Fig. 6.20 gives the short-circuit-load
equivalent circuit
rb Cμ rc
B C
CdCS,tot
Cπ r'π (1+g'm re) r'0
g'm vbe /(1+g'm re)
re
E E
From Eq. (6.107), we have for the small-signal base−emitter voltage
′ = vbe − ic re − ib ( rb + re ) ≈ vbe − ic re ,
vbe (1)
where we have assumed the base current to be very small compared to the collector current.
Similarly, we have
′ = vcb − ic rc + ibrb ≈ vcb − ic rc = − vbc − ic rc .

vcb (2)
Now, for the short-circuit load, vbc = vbe, so (2) gives
′ = − vbe − ic rc .
vcb (3)
From the equivalent circuit, we have
g m′ ⎡ g m′ ⎤
ic = vbe + jωC μ vcb
′ =⎢ − jωC μ ⎥ vbe − jωC μ rc ic , (4)
1 + g m′ re ⎣1 + g m′ re ⎦
which can be rearranged to give
⎡ g m′ ⎤
[1 + jωC r ]i
μ c c =⎢ − jωC μ ⎥ vbe . (5)
⎣1 + g m′ re ⎦
For a transistor having collector doping concentration of 2.5×1016 cm−3, Fig. 2.16 shows that
CdBC is about 0.45 fF/μm2. If the base−collector junction area is 2 μm2 and the frequency is 1
GHz, then ωCμ = ωCdBC,tot = 5.6×10−3 mhos. For a bipolar transistor biased to have IC = 2 mA,
SOLUTION 69
g′m = qIC/kT = 7.7×10−2 mhos. That is, we can assume ωCμ to be small compared to g′m. We
can also show that ωCμrc is small compared to unity. Therefore, (5) gives
g m′
ic ≈ vbe . (6)
1 + g m′ re
Similarly, we have from the equivalent circuit
vbe
ib = + jωCπ vbe
′ + jωC μ vbc
′ . (7)
rb + re + rπ′ (1 + g ′m re )
Since 1/g′m is on the order of 13 Ω, r′π = β0/g′m is on the order of 1300 Ω, which is large
compared to rb and re. Therefore, using (1) and (3), we can rewrite (7) as
vbe
ib ≈ + jωCπ (vbe − ic re ) + jωC μ (vbe + ic rc )
rπ′ (1 + g m′ re )
jω (C μ rc − Cπ re )g m′ ⎤
(8)
⎡
+ jω (Cπ + C μ ) +
1
=⎢ ⎥ vbe
⎣ rπ′ (1 + g m′ re ) 1 + g m′ re ⎦
ic g m′
and β (ω ) = = . (9)
+ (1 + g m′ re ) jω (Cπ + C μ ) + jω (C μ rc − Cπ re )g m′
ib 1
rπ′
(9) can be rearranged to give
g m′ rπ′
β (ω ) =
[
1 + jωrπ′ (Cπ + C μ ) + g m′ (re + rc )C μ
.
] (10)
In the high-frequency limit, (10) becomes
g m′
β (ω ) ≈
[
jω (Cπ + C μ ) + g m (re + rc )C μ
,
] (11)
which drops to unity at a frequency fT given by
g m′
1=
[
2πf T (Cπ + C μ ) + g m′ (re + rc )C μ ] (12)
1 kT
or =τ F + (C dBE ,tot + C dBC ,tot ) + C dBC ,tot (re + rc ) , (13)
2π f T qI C
SOLUTION 70
where we have used Eqs. (6.99), (6.102), (6.103), and (6.116), i.e.
∂ IC qI
g m′ = = C,
′
∂ VBE kT
C μ = C dBC ,tot ,
Cπ = CdBE ,tot + C DE
qI C
and C DE = τ F = τ F g m′ .
kT
8.2. N−P−N transistor with AE = 2 μm2, ABC = 10 μm2, deep emitter with NE = 1020 cm−3, NB
= 1018 cm−3, WB = 100 nm, β0 = 100, and NC = 5×1016 cm−3. Assume one-sided junction
approximation, with VBE = 0.8 V and VCB = 2 V.
(a) From Eq. (8.16), we have
τ pE
τ E (deep− emitter) = . (1)
2β 0
From Fig. 2.24(b), we have τpE = 5×10−10 s. Therefore
τ pE
τ E (deep− emitter) = = 2.5 × 10 −12 s. (2)
2β 0
From Fig. 2.24(a), we have μnB = 330 cm2/V-s. Therefore DnB = kTμnB/q = 8.58 cm2/s.
Therefore, Eq. (8.18) gives
WB2
tB ≈ = 5.83 × 10−12 s. (3)
2 DnB
As discussed in Section 6.4.4, the base delay time is
2
τ B = t B = 3.89 ×10 −12 s. (4)
3
2ε si (ψ bi + VCB )
WdBC = . (5)
qN C
From Fig.2.15, we have ψbi ≈ 0.95 V. Therefore
SOLUTION 71
WdBC = = 2.77 × 10−5 cm. (6)
qN C
From Eq. (8.17) we have
WdBC
t BC = . × 10−12 s,
= 138 (7)
2vsat
where we have used vsat = 107 cm/s. Similarly, we have
2ε si (ψ bi − VBE )
WdBE = . × 10−6 cm,
= 171 (8)
qN B
WdBE
and t BE = = 0.86 × 10−12 s, (9)
2v sat
where we have used ψbi = 1.025 V from Fig. 2.15. From (6.114), we have
τ F = τ E + τ B + τ BE + τ BC = τ E + τ B + t BE + t BC = 8.63 ×10 −12 s. (10)
(b) From Eq. (2.83), we have
AE ε si
C dBE ,tot = = 1.22 ×10 −14 F, (11)
WdBE
ABC ε si
and C dBC ,tot = = 3.75 ×10 −15 F. (12)
WdBC
(c) For JC = 0.3qvsatNC = 2.4×104 A/cm2, we have
IC = AE J C = 4.8 × 10−4 A (13)
1 kT
and =τ F + (C dBE ,tot + C dBC ,tot ) = (8.63 ×10 −12 + 0.86 × 10 −12 ) s (14)
2π f T qI C
or fT = 16.9 GHz.
(d) For (re + rc) = 50 Ω, we have
(re + rc )C dBC ,tot = 0.19 × 10 −12 s, (15)
SOLUTION 72
which is very small compared to τF in this example. From (14) and (15), we have
1 kT
=τ F + (C dBE + C dBC ) + C dBC (re + rc ) = (9.59 × 10 −12 + 0.19 ×10 −12 ) s,
2π f T qI C
(16)
or fT = 16.3 GHz.
8.3. Transistor in Exercise 8.2 is replaced by a pedestal-collector design, with NC(int) =

5×10 cm−3, and NC(ext) = 5×1015 cm−3, i.e., NC(ext) is reduced by 10 times. The
16
corresponding areas are ABC(int) = AE = 2 μm2 and ABC(ext) = (ABC − AE) = 8 μm2. From (5) of
Ex. 8.2, we have
2ε si (ψ bi + VCB ) 2ε si ( 0.95 + 2)
WdBC (int) = = = 2.77 × 10−5 cm. (1)
qN C (int) qN C (int)
Similarly, we have
2ε si (ψ bi + VCB ) 2ε si (0.89 + 2)
WdBC (ext ) = = = 8.67 × 10 −5 cm. (2)
qN C (ext ) qN C (ext )
AE ε si
Therefore, C dBC ,tot (int) = = 0.751×10 −15 F, (3)
WdBC (int)
A (ext )ε si
C dBC ,tot (ext ) = BC = 0.960 × 10 −15 F (4)
WdBC (ext )
and C dBC ,tot = C dBC ,tot (int) + C dBC ,tot (ext ) = 1.71× 10 −15 F. (5)
To repeat Ex. 8.2, we simply use this value of CdBC,tot, instead of (11) of Ex. 8.2.
(a) Compared to Ex. 8.2, τE, τB, tBE and tBC all remain unchanged by the pedestal
collector design, since NC(int) is unchanged. Therefore, τF is also unchanged.
(b) CdBE,tot is also unchanged, and CdBC,tot is given by (5) above.
(c) JC is determined by NC(int) and hence is unchanged. Therefore, IC is also

unchanged, and
1 kT
=τ F + (C dBE ,tot + C dBC ,tot ) = (8.63 ×10 −12 + 0.75 × 10 −12 ) s (6)
2π f T qI C
or fT = 17.0 GHz.
SOLUTION 73
(d) For (re +rc) = 50 Ω,
(re + rc )C dBC ,tot = 0.09 × 10 −12 s, (7)
1 kT
and =τ F + (C dBE ,tot + C dBC ,tot ) + C dBC ,tot (re + rc ) = (9.38 ×10 −12 + 0.09 × 10 −12 ) s (8)
2π f T qI C
or fT = 16.8 GHz.
8.4. Same as Ex. 8.3, but with NC(int) = 1×1017 cm−3. In this case, WdBC(int) is reduced and
JC is increased.
2ε si (ψ bi + VCB ) 2ε si ( 0.97 + 2)
WdBC (int) = = . × 10−5 cm
= 196 (1)
qN C (int) qN C (int)
AE ε si
C dBC ,tot (int) = = 1.06 ×10 −15 F (2)
WdBC (int)
and C dBC ,tot = C dBC ,tot (int) + C dBC ,tot (ext ) = (1.06 + 0.960) × 10 −15 F = 2.02 × 10 −15 F. (3)
(a) Compared to Ex. 8.3, τE, τB, and tBE all remain unchanged by the pedestal collector
design. But tBC is reduced because of the larger NC(int). We have from (1)
WdBC (int)
t BC = = 0.98 × 10−12 s. (4)
2vsat
Therefore,
τ F = τ E + τ B + t BE + t BC = (2.5 + 3.89 + 0.86 + 0.98) × 10 −12 s = 8.23 × 10 −12 s. (5)
(b) CdBE,tot is unchanged, and CdBC,tot is given by (3) above.
(c) JC is increased to JC = 0.3qvsatNC(int) = 4.8×104 A/cm2, and IC is increased to
IC = AE J C = 9.6 × 10−4 A. (6)
1 kT
Therefore, =τ F + (C dBE ,tot + C dBC ,tot ) = (8.23 ×10 −12 + 0.38 × 10 −12 ) s (7)
2π f T qI C
or fT = 18.5 GHz.
SOLUTION 74
(d) For (re + rc) = 50 Ω,
(re + rc )C dBC ,tot = 0.10 × 10 −12 s, (8)
1 kT
and =τF + (C dBE ,tot + C dBC ,tot ) + C dBC ,tot (re + rc ) = (8.61× 10 −12 + 0.10 ×10 −12 ) s
2π f T qI C
(9)
or fT = 18.3 GHz.
8.5. Fig. 7.6 is shown here for use in this exercise.
1E+4
SiGe/Si Improvement Factor
Current gain Early voltage Base transit time

1E+3
1E+2
1E+1
1E+0
1E-1
0 2 4 6 8 10
ΔEgmax/kT
Fig. 7.6. Relative improvement factors for current gain, Early voltage, and base transit time of a SiGe-
base bipolar transistor over a Si-base bipolar transistor, as a function of the maximum base bandgap
narrowing. A linearly graded Ge profile is assumed. Also, γ and η are set to unity.
For ΔEgmax = 100 meV, we have ΔEgmax/kT ≈ 4. The incorporate of Ge does not change the
capacitances or the depletion layer thickness. It increases the current gain and reduces the base
and emitter transit times.
(a) Compared to Ex. 8.4, tBE and tBC are unchanged, but
τ E (SiGe) β 0 (Si) 1
= ≈ (1)
τ E (Si) β 0 (SiGe) 4
2.5 × 10 −12
or τ E (SiGe) = s = 0.625 × 10 −12 s, (2)
4
SOLUTION 75
where we have used τE(Si) = 2.5×10−12 s from (2) of Ex. 8.2. Also, from the plot above, we
have
τ B (SiGe) t B (SiGe)
= ≈ 0.4 , (3)
τ B (Si) t B (Si)
or τ B ( SiGe) = 0.4 × 3.89 × 10 −12 s = 1.56 ×10 −12 s (4)
where we have used τB(Si) = 3.89×10−12 s from (4) of Ex. 8.2. Therefore,
τ F = τ E + τ B + t BE + t BC = (0.625 + 1.56 + 0.98 + 0.86) × 10 −12 s = 4.03 × 10 −12 s. (5)
(b) Compared to Ex. 8.4, CdBE,tot and CdBC,tot are unchanged.
(c) Compared to Ex. 8.4, JC and IC are unchanged, and
1 kT
=τ F + (C dBE ,tot + C dBC ,tot ) = (4.03 ×10 −12 + 0.38 × 10 −12 ) s, (6)
2π f T qI C
or fT = 36.1 GHz.
(d) Compared to Ex. 8.4, (rc + re)CdBC,tot is unchanged, and
1 kT
=τF + (C dBE ,tot + C dBC ,tot ) + C dBC ,tot (re + rc ) = (4.41× 10 −12 + 0.10 × 10 −12 ) s (7)
2π f T qI C
or fT = 35.3 GHz.
8.6.
Carrier density
Carriers travel at constant velocity v
x x+dx
x
0 W
SOLUTION 76
The number of carriers between x and x + dx is ndx. At a constant velocity of v, the time
needed for these carriers to reach the depletion-layer edge after the current is switched off is (W
− x)/v. Therefore, the transit time averaged over all the carriers in the depletion layer is
W (W − x )
∫0 v ndx W
tavg = W = .
2v
ndx ∫0
8.7. Let us first consider the case of NC1(int) = 5×1016 cm−3. With a collector-base reverse
bias of VCB = 2V, the depletion-layer thickness is [see (5) in Ex. 8.2]
WdBC1 = = 2.77 × 10−5 cm (1)
qN C1 (int)
where we have used a ψbi value of 0.95 V from Fig. 2.15. With the one-sided junction
approximation, the maximum electric field is given by Eq. (2.78), i.e.
qN C1
Em1 = WdBC1 = 213
. × 105 V/cm. (2)
ε si
This is the field we want to maintain when we increase NC(int) and insert an i-layer at the same
time.
(a) Now consider NC2(int) = 1×1017 cm−3 with a i-layer of thickness d. Let WdBC2 be
the depletion-layer thickness. Since the electric field is given by (2), for our one-sided junction
approximation, Eq. (2.89) gives
qN C 2 (int)
ε si
[WdBC 2 − d ] = Em1 = 213
. × 105 V/cm, (3)
or . × 10−5 ) cm.
WdBC 2 = ( d + 138 (4)
Let WdBC20 denote the depletion-layer thickness when d is zero. We have
WdBC 20 = . × 10−5 cm,
= 196 (5)
qN C 2 (int)
where we have used ψbi = 0.97 V for NC2(int) = 1017 cm−3 [see (1) in Ex. 8.4]. Now, from Eq.
(2.92) we have for our p−i−n diode
WdBC 2 = WdBC
2
20 + d .
2
(6)
SOLUTION 77
Equating (4) and (6), we have
( d + 138
. × 10 ) = (196
. × 10 )
−5 2 −5 2
+ d2, (7)
which can be solved to give d = 0.70×10−5 cm = 70 nm. This is the i-layer thickness to
maintain the same electric field as in the NC1(int) case.
(b) For the transistor with NC1(int) = 5×1016 cm−3, the transit times have been
calculated in Ex. 8.2 [see (2), (4), (7), and (9) in Ex. 8.2], i.e.
τ pE
τ E1 (deep− emitter) = = 2.5 × 10 −12 s, (8)
2β 0
2
τ B1 = t B1 = 3.89 ×10 −12 s, (9)
3
W
. × 10−12 s,
t BC1 = dBC1 = 138 (10)
2v sat
W
t BE1 = dBE = 0.86 × 10−12 s. (11)
2v sat
and τ F 1 = τ E1 + τ B1 + t BE1 + t BC1 = 8.63 ×10 −12 s. (12)
For the transistor with NC2(int) = 1×1017 cm−3, τE, τB, and tBE are the same as transistor with
NC1(int), but tBC is different because of the different WdBC values. From (4), we have
. × 10−5 ) cm = 2.08 × 10−5 cm.

WdBC 2 = ( d + 138 (13)
WdBC 2
Therefore, t BC 2 = . × 10−12 s,
= 104 (14)
2v sat
and τ F 2 = τ E 2 + τ B 2 + t BE 2 + t BC 2 = (2.5 + 3.89 + 0.86 + 1.04) × 10 −12 s = 8.92 × 10 −12 s.

(15)
(c) The pedestal collectors have an extrinsic-collector doping concentration of NC(ext)

= 5×1015 cm−3. The depletion-layer thickness associated with the extrinsic collector-base
junction is [see (2) in Ex. 8.3]
2ε si (ψ bi + VCB ) 2ε si (0.89 + 2)
WdBC (ext ) = = = 8.67 × 10 −5 cm, (16)
qN C (ext ) qN C (ext )
where we have used ψbi = 0.89 V. The corresponding capacitance is [see (4) in Ex. 8.3]
SOLUTION 78
ABC (ext )ε si
C dBC ,tot (ext ) = = 0.960 × 10 −15 F. (17)
WdBC (ext )
The intrinsic collector-base junction capacitances are
AE ε si
C dBC1,tot (int) = = 0.751 × 10 −15 F, (18)
WdBC1
AE ε si
and C dBC 2,tot (int) = = 1.02 ×10 −15 F.
WdBC 2
(19)
Therefore, C dBC1,tot = C dBC ,tot (ext ) + C dBC1,tot (int) = 1.71×10 −15 F, (20)
and C dBC 2,tot = C dBC ,tot (ext ) + C dBC 2,tot (int) = 1.98 × 10 −15 F.
(21)
The emitter-base capacitances for the two transistors are the same, and has been calculated in
Ex. 8.2 [see (11) in Ex. 8.2], i.e.
AE ε si
C dBE ,tot = = 1.22 ×10 −14 F. (22)
WdBE
(d) For transistor with NC1(int) = 5×1016 cm−3, JC1 = 0.3qvsatNC1(int) = 2.4×104 A/cm2
and IC1 = AEJC1 = 4.8×10−4 A, and
1 kT
= τ F1 + (CdBE ,tot + CdBC1,tot ) = (8.63 × 10 −12 + 0.75 × 10 −12 ) s, (23)
2π f T 1 qI C1
or fT1 = 17.0 GHz.
For the transistor with NC2(int) = 1×1017 cm−3, JC2 = 0.3qvsatNC2(int) = 4.8×104 A/cm2 and IC2 =
AEJC2 = 9.6×10−4 A, and
1 kT
=τ F2 + (C dBE ,tot + C dBC 2,tot ) = (8.92 × 10 −12 + 0.38 × 10 −12 ) s, (24)
2π f T 2 qI C 2
or fT2 = 17.1 GHz.
8.8. We want to compare the Early voltages for the two base doping profiles.
SOLUTION 79
Concentration Concentration
NB NB
Base Base
NC NC
Collector Collector
x x
0 WB 0 WB
(WB − d)
(i) (ii)
QpB
(a) From VA ≈ , (1)
CdBC
QpB ( d = 0) qWB N B
we have VA ( d = 0 ) = = , (2)
CdBC ( d = 0) CdBC ( d = 0)
QpB ( d ) qN B (WB − d )
and VA ( d ) = = . (3)
CdBC ( d ) CdBC ( d )
VA ( d ) (W − d ) CdBC ( d = 0)
Therefore, = B . (4)
Vd ( d = 0) WB CdBC ( d )
Profile (ii) is a p-i-n diode, and Eq. (2.97) gives
CdBC ( d = 0) d2
= 1+ 2 . (5)
CdBC ( d ) WdBC ( d = 0)
Putting (5) into (4) gives
VA ( d ) ⎛ d ⎞ d2
= ⎜1 − ⎟ 1+ 2 . (6)
Vd ( d = 0) ⎝ WB ⎠ WdBC ( d = 0)
(b) For case of d/WB = 0.1 and the VA ratio equals to unity, we have
d2 d
1 = 0.9 1 + and = 0.48 .
WdBC
2
( d = 0) WdBC ( d = 0)
For the case of d/WB = 0.2 and the VA ratio equals to unity, we have
SOLUTION 80
d2 d
1 = 0.8 1 + 2 and = 0.75 .
WdBC ( d = 0) WdBC ( d = 0)
(c) Assume NB = 1×1018 cm−3, NC = 5×1016 cm−3, WB = 100 nm, and d = 10 nm. If we
assume one-sided B−C junction with VBC = 0, Fig. 2.16 gives WdBC(d = 0) ≈ 0.15 μm = 150 nm.
Putting these values into (6), we obtain
VA ( d )
≈ 0.90 for d = 10 nm. (7)
Vd ( d = 0)
For the case of d = 20 nm, we have
VA ( d )
≈ 0.81 for d = 20 nm. (8)
Vd ( d = 0)
8.9 Referring to Appendix 18 (on fT and fmax of bipolar transistors), we want to show that
re Re(Y11′ Y22′ / Y21′ ) << 1 for a typical bipolar transistor.
From Eqs. (A18.2) to (A18.5), we have
1 g′
Y11′ = + jω (Cπ + Cμ ) = m + jω (Cπ + Cμ ) , (1)
rπ′ β0
Y21′ = g ′m − jωCμ , (2)
1
and Y22′ = + jωCμ . (3)
r0′
Y11′ Y22′ Y11′ Y22′ Y21′*

We have = (4)
Y21′ Y21′
2
⎧⎡ g m′ ⎤⎡ 1 ⎤ ⎫
=
1
⎨⎢ [ ]
+ jω (Cπ + C μ )⎥ ⎢ + jωC μ ⎥ g m′ + jωC μ ⎬
Y21′ ⎩⎣ β 0 ⎦ ⎣ r0′
2
⎦ ⎭
⎧⎪⎡ g m′ ⎛ Cπ + C μ g m′ C μ ⎞⎤ ⎫⎪
− ω 2 C μ (Cπ + C μ ) + jω ⎜⎜ ⎟⎟⎥[g m′ + jωC μ ]⎬
1
= ⎨⎢ +
Y21′ ⎪⎩⎣ β 0 r0′ ⎝ r0′ β0 ⎪⎭
2
⎠⎦
1 ⎧ g m′ 2 ⎛ Cπ + C μ g ′m C μ ⎞⎫
= 2 ⎨
− ω 2 g m′ C μ (Cπ + C μ ) − ω 2 C μ ⎜⎜ + ⎟⎟⎬ + j (....) .
Y21′ ⎩ β 0 r0′ ⎝ r0′ β0 ⎠⎭
SOLUTION 81
Therefore,
⎛ Y ′Y ′ ⎞ ⎧ g m′ 2 ⎛ C + C μ g m′ C μ ⎞⎫
− ω 2 g m′ C μ (Cπ + C μ ) − ω 2C μ ⎜⎜ π
r
re Re⎜⎜ 11 22 ⎟⎟ = e 2 ⎨ + ⎟⎟⎬ (5)
⎝ Y21′ ⎠ Y21′ ⎩ β 0 r0′ ⎝ r0′ β0 ⎠⎭
r g ′2
< e2 m
Y21′ β 0 r0′
re g ′m2
=
( )
β 0 r0′ g m′ 2 + ω 2C μ2
.
For a typical transistor, assume re = 10 Ω, β0 = 100, IC = 1 mA, VA = 40 V, frequency f = 10

GHz, and Cμ = CdBC,tot = 2×10−15 F, we have
β 0 r0′ β 0 VA
= = 4× 10 5 , (7)
re re I C
2
β 0 r0′ β 0 r0′ ⎛ kT ⎞ β r′
and
re g ′ 2
ω Cμ =
2 2
⎜⎜
re ⎝ qI C ⎠ re
(
⎟⎟ ω 2 C μ2 = 0 3.3 × 10 −3 )2
= 4.4 . (8)
m
⎛ Y ′Y ′ ⎞ 1
Therefore, re Re⎜⎜ 11 22 ⎟⎟ < .
⎝ Y21′ ⎠ 4 × 10
5
SOLUTION 82
9.1. Consider a bipolar transistor biased to operate in the saturation region. Ignoring parasitic
resistances, the collector current is given by Eq. (9.5) and the base current is given by Eq. (9.6),
namely,
⎡ I ⎤
I C = I B 0 F e qVBE / kT ⎢ β 0 (1 − e − qVCE / kT ) − B 0 R e − qVCE / kT ⎥ ,
⎣ I B0F ⎦
and
⎛ I ⎞
I B = I B 0 F e qVBE / kT ⎜⎜1 + B 0 R e − qVCE / kT ⎟⎟ .
⎝ I B0F ⎠
The current gain is
I B 0 R − qVCE / kT
β 0 (1 − e − qV CE / kT
)− e
IC I BoF
β≡ = .
IB I B 0 R − qVCE / kT
1+ e
I B0F
Plot the current gain β as a function of qVCE/kT for qVCE/kT = 0 to qVCE/kT = 4, using β0 = 100
and IB0R/IB0F = 1. Repeat using β0 = 100 and IB0R/IB0F = 10. This exercise shows that for
practical transistors the current gain is less than 1 only for very small VCE values.
100
90
80
70
60
IB0R/IB0F = 1
IC / IB
50
40 IB0R/IB0F = 10
30
20
10
0
-10
0 0.5 1 1.5 2 2.5 3 3.5 4
qVCE / kT
SOLUTION 83
10
9
8
7
6
IB0R/IB0F = 1
IC / IB
5
4 IB0R/IB0F = 10
3
2
1
0
-1
0 0.05 0.1 0.15 0.2
qVCE / kT
SOLUTION 84
10.1.
(a)
The expression for the drain current is given by
2 βs
W 4ε si ⎛ 2kT ⎞ ⎡ β 2 ε siti 2 2 ⎤
I ds = μ ⎜ ⎟ ⎢ β tan β − + β tan β ⎥ . (1)
L t si ⎜⎝ q ⎟⎠ ⎣ 2 ε it si ⎦ βd
The boundary conditions at the source and drain are:
q(Vgs − Δφ ) ⎛ 2 2ε si kT ⎞
− ln⎜⎜ ⎟ = ln β s − ln(cos β s ) + 2ε siti β s tan β s (2)
2kT t q 2
n ⎟ ε t
⎝ si i ⎠ i si
q(Vgs − Δφ − Vds ) ⎛ 2 2ε si kT ⎞
− ln⎜⎜ ⎟ = ln β d − ln (cos β d ) + 2ε siti β d tan β d . (3)
2kT ⎝ t si q ni ⎟⎠
2
ε itsi
Here Δφ = 4.33 − (4.05+0.56) = −0.28 V.
If the values of βs and βd are solved for given biases Vgs and Vds, the value of the current is
obtained by Eq. (1). The problem is then to solve the implicit equations (2) and (3) with given
boundary conditions.
1.0
Vg=0.0(V)
Vg=0.5(V)
0.8
Vg=1.0(V)
Vg=1.5(V)
0.6 Vg=2.0(V)
Ids (mA)
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0
Vds (V)
Various numerical methods can be employed:
(i) Scan βs over 0 < β s < π 2 and use Eq. (2) to generate a look-up table mapping βs to Vgs.
Pick βs values that give Vgs = 0, 0.5, 1.0, 1.5, 2.0 V. Interpolation may be needed. For a given
Vgs and the corresponding βs, scan βd over (0, π/2) and use Eq. (3) to calculate corresponding Vds.
An Ids-Vds curve can then be generated for each Vgs (βs).
(ii) The implicit equations can be solved by numerical methods such as Newton Raphson
method. Use of internal functions of mathematical tools is also acceptable. (e.g., fsolve or solve
of Matlab.)
SOLUTION 85
(b)
∂I ds ∂I ds ∂β s ∂I ds ∂β d
i. Calculating gm: = + .
∂Vgs ∂β s ∂Vgs ∂β d ∂Vgs
From Eq. (1) in (a),
2
∂I ds ⎛ W ε si ⎞⎛ 2kT ⎞ ⎛ dβ tan β s ⎞
= ⎜⎜ 4 μeff ⎟⎜ ⎟ ⎜⎜ tan β s + β s tan 2 β s + 2r ( β s tan β s ) s ⎟⎟
∂β s ⎝ L tsi ⎟⎠⎜⎝ q ⎟⎠ ⎝ dβ s ⎠
2
∂I ds ⎛ W ε si ⎞⎛ 2kT ⎞ ⎛ dβ tan β d ⎞
= −⎜⎜ 4μeff ⎟⎜ ⎟ ⎜⎜ tan β d + β d tan 2 β d + 2r ( β d tan β d ) d ⎟⎟ .
∂β d ⎝ L tsi ⎟⎠⎜⎝ q ⎟⎠ ⎝ dβ d ⎠
From Eqs. (2) and (3) in (a),

dVgs ⎛ 2kT ⎞⎛ 1 dβ tan β s ⎞
= ⎜⎜ ⎟⎟⎜⎜ − tan β s + 2r s ⎟⎟
dβ s V = const . ⎝ q ⎠⎝ β s dβ s ⎠
ds
dVgs ⎛ 2kT ⎞⎛ 1 dβ tan β d ⎞

= ⎜⎜ ⎟⎟⎜⎜ − tan β d + 2r d ⎟⎟ .
dβ d V = const . ⎝ q ⎠⎝ β d dβ d ⎠
ds
Substitution:
2
∂I ⎛ W ε si ⎞⎛ 2kT ⎞
g m = ds = ⎜⎜ 4μeff ⎟⎜ ⎟ [β s tan β s − β d tan β d ] .
∂Vgs ⎝ L t si ⎟⎠⎜⎝ q ⎟⎠
Vds = const .
∂I ds ∂I ds ∂β s ∂I ds ∂β d
ii. Calculating gds: = + .
∂Vds ∂β s ∂Vds ∂β d ∂Vds
From Eqs. (2) and (3) in (a),

dVds
=0
dβ s V = const .
gs
dVds ⎛ 2kT ⎞⎛ 1 dβ tan β d ⎞

= −⎜⎜ ⎟⎟⎜⎜ − tan β d + 2r d ⎟⎟ .
dβ d V gs = const . ⎝ q ⎠⎝ β d dβ d ⎠
Substitution:
2
∂I ⎛ W ε si ⎞⎛ 2kT ⎞
g ds = ds = ⎜⎜ 4 μeff ⎟⎜ ⎟ β d tan β d .
∂Vds V gs = const . ⎝ L t si ⎟⎠⎜⎝ q ⎟⎠
SOLUTION 86
(c) tan(πti / λ ) tan(πtsi / 2λ ) = ε i / ε si
-9
1.0x10
f=εsitan(πtox/λ))-εoxcot(πtsi/2λ)
-10
5.0x10
λ=19.80(nm)
0.0
-10
-5.0x10
5 10 15 20 25
λ (nm)
By numerically plotting the function in the graph above, λ1 = 19.80(nm) .

Therefore, the device can be scaled down to Lmin ≈ 2λ1 = 39.6(nm) without severe SCEs.
SOLUTION 87
Solutions to Appendix 7 Exercises
xm
0 x
qψ field (x) qψ image(x)

qφ Bn qΔφ
Ec
Ef
Total
Ev
Metal Silicon
A7.1. Referring to the figure, the electrostatic potential associated with the image force is
q
ψ image ( x) = , (1)
16πε 0 x
and the electrostatic potential associated with the electric field in the silicon is given by
d 2ψ field ( x) qN d
− = . (2)
dx 2
ε si
Integrating (2) twice, subject to the boundary condition that the electric field is zero at x = Wd,
where Wd is the depletion layer width, and the boundary condition that ψfield(0) = 0, we have
qN d ⎛ x 2 ⎞
ψ field ( x) = − ⎜⎜ − Wd x ⎟⎟ . (3)
ε si ⎝ 2 ⎠
Therefore, the total electrostatic potential energy is
⎡ q qN d ⎛ x 2 ⎞⎤
PE ( x) = −q[ψ image ( x) + ψ field ( x)] = −q ⎢ + ⎜⎜Wd x − ⎟⎟⎥ . (4)
⎣16πε si x ε si ⎝ 2 ⎠⎦
The peak of PE(x) is given by dPE(x)/dx = 0, that is
−q qN dWd ⎛ x ⎞
+ ⎜⎜1 − m ⎟⎟ = 0 . (5)
16πε x 2
si m ε si ⎝ Wd ⎠
SOLUTION 88
We will first make the assumption that xm/Wd << 1, and then come back to verify that our
assumption is correct. With this assumption, (5) gives
q
xm ≈ , (6)
16πε siEm
− dψ field ( x) qN dWd
where Em = = (7)
dx x =0
ε si
is the absolute value of the electric field in the silicon at the interface. The energy barrier
lowering is
⎡ q qN d ⎛ x 2 ⎞⎤
qΔφ =| PE ( xm ) |= q ⎢ + ⎜⎜Wd xm − m ⎟⎟⎥ (8)
⎣16πε si xm ε si ⎝ 2 ⎠⎦
⎡ q qN dWd xm ⎤ q 3Em
≈ q⎢ + ⎥= .
⎣16πε si xm ε si ⎦ 4πε si
To show that xm/Wd << 1, we note that Wd is related to the total band bending. For
simplicity we assume there is no external applied voltage, so that
2ε siψ bi
Wd = , (9)
qN d
where ψbi is the built-in potential. For ψbi = 0.4 V and Nd = 1%1016 cm−3, we have Wd =
2.3%10−5 cm and xm = 3%10−7 cm. That is, xm/Wd << 1 as assumed.
SOLUTION 89

Fundamentals of Modern VLSI Devices - Solution Manual (1332223)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fundamentals of Modern VLSI Devices - Solution Manual (1332223)

Uploaded by

Copyright:

Available Formats

SOLUTION SET

FUNDAMENTALS OF MODERN VLSI DEVICES, 2nd ed.

Cambridge University Press

Yuan Taur and Tak H. Ning

2.1. From Eq. (2.4),

Adding the above two equations yields

Applying integration by parts to the numerator yields

which is Coulomb’s law.

Since E =−dV/dr, the electric potential at a point on the sphere is

if one defines the potential to be zero at infinity.

The solution to the above differential equation is

where ψs = ψ(x = 0) is the surface potential.

Using Eq. (2.178) and the above solution, one has

where n(0) equals (ni2/Na)exp(qψs/kT).

2.8. Substituting Eq. (2.189) into Eq. (2.195), one obtains

which is basically a quadratic equation that can be solved for ψs:

2kT ⎛ Cox (Vg − V fb − 2ψ B ) ⎞

Here exp(qψB/kT) = Na/ni has been applied.

Since Qs = Qi + Qd and ΔQi/ΔQd = Ci/Cd, it follows that

2.11. Equation (2.256) gives

As Mp → ∞, 1/Mp → 0, the above equation gives

0 = exp ⎡− ∫ (α p − α n )dx ⎤ − ∫ α n exp ⎡− ∫ (α p − α n )dx ′ ⎤dx

Using (3), we have

f ( x )exp ⎡− ∫ f ( x ' )dx '⎤ dx = ∫ f ( x )eu( x ) dx

f ( x )exp ⎡− ∫ f ( x ' )dx '⎤ dx = 1 − exp ⎡− ∫ f ( x ' )dx '⎤ .

A similar procedure can be used to show that

f ( x )exp ⎡− ∫ f ( x ' )dx '⎤ dx = 1 − exp ⎡− ∫ f ( x ' )dx '⎤

2.13. From Eqs. (2.80), (2.81), and (2.83), we have

Now, from Eq. (2.73), we have

Therefore, from the plot of 1 / C d2 versus Vapp plot, we have

dQd = qN a (W )dW , (1)

where the depletion-layer width W is a function of V. The depletion-layer capacitance is

Substituting (4) into (3), we have

2.15. Referring to Fig. 2.17 and Eq. (2.90), we have

where V is the applied voltage. From Gauss’s law we have

1 dV d [Em (Wd + d ) / 2] 1 ⎡ dWd ⎤

Now, from Eq. (2.89), we have

Substituting (4) into (3), we have

and pn = pn 0 at x = −Wn. (3)

The general solution to (1) has the form of

Using (2), (4) gives

Solving (5) and (6), we have

Substituting (7) and (8) into (4) and rearranging, we have

qDn ni2 qni2 Ln

qDp ni2 qni2 Lp

qDp ni2 qni2 L2p

qDp nie2 qni2 L2p

From Eq. (2.147), we have

Substituting (3) and (4) into (2), we obtain

Similarly, Eq. (2.166) and the hole-equivalent of Eq. (2.127) give

2.19. We will do this by showing that

From (2), we have

Substituting (4) into (3) and using (1), we have

= exp⎛⎜ − ∫ (α p − α n )dx ⎞⎟ − ∫ α n exp⎛⎜ − ∫ (α p − α n )dx′ ⎞⎟dx .

Since αp and αn are independent of x, (1) can be integrated to give

Equation (2) can be rearranged to give

As avalanche breakdown is approached, Mp approaches infinity and W approaches a value given

The WKB tunneling coefficient is, from Eq. (2.245),

with W = Eg / qE. For the coordinates indicated in the figure, we have

(c) Cmax = Cox = 1.73 ×10 −7 F/cm2

Cmin = 0.63 ×10 −7 F/cm2

Solve the above quadratic equation in ψs :

hole density at the surface: p = N a e − qψ s / kT = 1.2 × 1011 cm-3