(반도체공학) Fundamentals of Semiconductor Devices Solutions

SOLUTIONS CHAPTER 1
1.1. Show that Equation (1.6) follows from Equation (1.3).

q2
Solution: Equation (1.3) is dEP = dr . Integrating both sides we obtain
4πε 0 r 2
q2 q2 1 q2 ⎡ 1 ⎤
∫ P ∫ 4πε 0r2
dE = dr =
4πε0 ∫ r
2 dr = − + const = EP
4πε 0 ⎢⎣ r ⎥⎦
To find the constant, we employ the boundary condition that at r = ∞ , EP=Evac:
−q 2
EP (r = ∞) = Evac = + const = 0 + const
4πε 0 ⋅ ∞
∴const = Evac and
−q 2
EP = Evac − = Evac , Equation (1.6)
4πε 0 ⋅ ∞
1.2. Consider a lithium nucleus, of charge +3q. Calculate the first three
electron energies for an electron in a Li++ ion, using the Bohr model.
We repeat the analysis that we used for the hydrogen atom, except that
now the charge of the nucleus Q1 is equal to 3q = +3(1.6 × 10−19 )C . The
results of the key steps are
QQ −3q 2
F = 1 22 =
4πε 0 r 4πε 0 r 2
3q 2
EP (r ) = Evac −
4πε 0 r
mv 2 3q 2
− =0
r 4πε 0 r 2
mvn rn = nh
3q 2 ⎛1⎞
vn =
( 4πε 0 ) h ⎜⎝ n ⎟⎠
( 4πε 0 ) h 2
rn = (n 2 )
3mq 2
m32 q 4 ⎛ 1 ⎞
EK = ⎜ 2⎟
2 ( 4πε 0 ) h ⎝n ⎠
2 2
9mq 4 ⎛ 1 ⎞
E n = EPn + EKn = Evac − ⎜ 2⎟
2 ( 4πε 0 ) h ⎝n ⎠
2 2
Thus E1 = Evac − 9 (13.6eV ) = Evac − 122eV
Anderson & Anderson 1 2/15/04

Solutions Chapter 1
9 (13.6eV )
and E2 = Evac − = Evac − 30.6eV
22
9 (13.6eV )
and E3 = Evac − = Evac − 13.6eV
32
1.3. Show that Equations (1.12) and (1.13) follow from (1.8) and (1.11).
mv 2 q2
Equation (1.8) is: − =0
r 4πε 0 r 2
Multiply both sides by r2 and divide by v :
q2
mvr = which from Equation (1.11) is
4πε 0 v
q2
mvr = = nh
4πε 0 v
Solving the right hand equality for v :
q2
v= (Equation (1.13))
4πε 0 nh
Solving the left hand equality for r substituting in v :
nh nh ( 4πε 0 ) nh 4πε 0 n 2 h 2
r= = = (Equation (1.12)).
mv mq 2 mq 2
1.3. In each of the potential energy distributions in Figure 1P.1, sketch the
magnitude and direction of the force on the electron.

Solutions Chapter 1
The force is minus the gradient of the potential energy (Equation (1.2)).
1.5. Consider the electron in the energy diagram of Figure 1P.2. Taking the
energy the electron has at Point A as Etotal, at each of the indicated positions,
find the total energy, the kinetic energy, the potential energy, and the
electron’s velocity. Indicate the direction of force (if any). Recall that total
energy is conserved. At point “D” the electron collides inelastically with
something (perhaps an atom in the crystal). After the collision, the electron’s
energy is equal to its potential energy, and its kinetic energy is zero. Its total
energy is much less than before the collision; where did the extra energy go?
The electron at Point A is at rest. Its total energy is Etotal, its kinetic energy
and velocity are zero, and its potential energy is equal to its total energy.
The force on the electron is to the right, because the slope of the potential
energy is negative.
At Point B, the total energy of the electron is the same (by conservation of
energy), but the potential energy is EP(B ) = Etotal − 5eV . The kinetic energy
is the difference between the total and the potential energies, so
EK (B) = Etotal − EP (B ) = Etotal − (Etotal − 5eV ) = 5eV . The velocity of the electron
2 EK 2 ⋅ (5eV ) ⋅ (1.6 × 10−19 J / eV )

is v = = = 1.33 × 106 m / s (or 3 million miles
m 9.1 × 10−31 kg
an hour).
At Point C, the electron’s energy is Etotal, but its potential energy is

Solutions Chapter 1
EP(C) = Etotal − 10eV . Therefore, the kinetic energy of the electron is
EK (C ) = Etotal − E P(C) = Etotal − (Etotal − 10eV ) = 10eV , and
2 EK ( C ) 2 ⋅ (10eV ) (1.6 × 10−19 eV / J
v= = −31
= 1.86 × 106 m / s
m 9.1 × 10 kg
There is no force on the electron because the slope of the potential energy
is zero, but the electron continues moving to the right because it still has
kinetic energy.
At Point D, the electron has made a collision. Its total energy is now 10eV
less than it was- the extra 10 eV of energy was transferred to the atom it
collided with. The potential energy is EP (D) = Etotal − 10eV . The electron’s
total energy is now Etotal(new) = Etotal (original ) − 10eV , which is the same as its
potential energy. Therefore, the kinetic energy of the electron is zero, and
so is the velocity. There is no force on the electron because the slope of
the potential energy is zero.
1.6. Find the kinetic energies in each of the following. Express all your answers
in electron volts.
(a) an electron in the first allowed energy states of the hydrogen atom
(according to the Bohr model, Eq. (1.14))
mq 4 1
We use the relation EK =
2 ( 4πε 0 )
2 2 2
h n
( 9.1E − 31) kg (1.6 E − 19 ) C 4
4
1
EK =
2 ( 4π 8.8E − 12 ) ( F / m ) (1.06 E − 34 ) J s 1
2 2 2 2 2 2 2
= 2.15 E − 18 J
= 13.6eV
(b) a free electron, initially at rest at the back of a cathode ray tube in
your television, accelerated through a potential of 10 kV to strike the phosphor
layer
An electron accelerated through 10kV acquires an energy of 10KeV. Since
this electron was initially at rest, its kinetic energy is now 10 KeV, or
10,000 × 1.6 × 10−19 J = 1.6 × 10−15 J .
(c) a tiny, drifting dust particle, of mass 1 µg and velocity a leisurely 1
mm/s.
1 1
We use EK = mv 2 = ⋅10−9 kg ⋅ (10−3 m / s ) 2 = 5 × 10−16 J = 3.1keV
2 2

Solutions Chapter 1
1.7. For the following semiconductor materials, indicate to what degree you
expect covalent or ionic bonding, and why:
Ge
GaP
InGaAsP
HgCdTe
Ge: This element is in Column IV of the periodic table. Each atom in the
crystal shares 4 electrons with each neighboring atom, and in return
“receives” 4 electrons. Because every atom is “giving” and “receiving” the
same number of electrons, the charge is evenly distributed between
atoms, and the bonding is covalent.
GaP: Gallium is in Column III and phosphorous is in Column V. Thus in
the bonding, indium has less positive charge than the phosphorus, so the
electrons spends slightly more time near the P atoms than the Ga. This
bonding is largely covalaent but with a slightly ionic flavor.
InGaAsP: Indium and gallium are in Column III and arsenic and
phosphorous are in Column V, so the bonding has the same character as
that of InP, mostly covalent but somewhat ionic.
HgCdTe: Mercury is in Column II, while cadmium and tellurium are in
column VI. This II-VI semiconductor material has a largely ionic but slightly
covalent bonding. The electrons spend quite a bit of their time nearer the
column VI atoms because of those atoms’ greater positive charge.
1.8. For each of the semiconductors below, draw (to scale) the energy band
diagrams:
Si: Eg=1.12 eV. χ=4.05 eV
Ge: Eg=0.67 eV, χ=4.0 eV
GaAs: Eg=1.43 eV, χ=4.07 eV
InP: Eg=1.35 eV, χ=4.35 eV

Solutions Chapter 1
1.9. What minimum energy must an electron at the bottom of the conduction
band in aluminum antimonide gain to become free of the crystal? Repeat for
an electron at the top of the valence band?
The energy required to remove an electron from EC to Evac is the electron
affinity, which from Table 1.2 is 3.6 eV for aluminum antimonide. The
energy required to remove it from EV is the ionization potential, which is
equal to
γ = χ + E g = 3.6eV + 1.6eV = 5.2eV
1.10. A nondegenerate semiconductor cannot conduct current at absolute zero

(degeneracy will be discussed in Chapter 2). How much energy must at least
one electron obtain in silicon before conduction is possible?
In intrinsic silicon, at least one electron must be excited from the valence
band into the conduction band, meaning it must gain an energy equal to at
least the band gap. Thus ∆E=1.12 eV.
1.11. At room temperature in a cubic centimeter of intrinsic silicon, there will

be about 10 billion electrons in the conduction band.
a) How many holes are in the valence band?
The number of holes in an intrinsic semiconductor must be equal to the
number of electrons, or 10 billion.
b) If electrons are constantly seeking lower energies and recombining
with holes (empty states at lower energies), then how can the number 10
billion remain constant?
At a given temperature, the thermal energy of the crystal is a constant. If
some electrons are losing energy by recombining, others are gaining
energy, generating new electron-hole pairs.
1.12. Suppose the electron in Figure 1.12 is traveling to the right at constant
energy. What happens to it as it approaches the surface of the material?
Explain your answer, using the energy diagram.
The electron travels at constant energy and thus at constant speed
(constant kinetic energy) in the region where EP is constant. When the
electron approaches the barrier, the potential energy increases. The
kinetic energy thus decreases, and the electron slows down. When it
reaches the edge, its total energy is equal to its potential energy so it
stops. There is a force to the left (gradient in the potential energy) so the
electron is accelerated back to the left. Thus the electron is reflected at
this barrier.
1.13. Show that Equation (1.38) is a solution to Equation (1.37). What is the
significance of the positive and negative signs of K?
2m0 (E − E0 )
Equation (1.38): ψ (x) = Ae jKx + Be − jKx . Letting K = 2 (Equation
h
1.40), the solution can be written as

Solutions Chapter 1
⎛ 2m0 (E −E 0 ) ⎞ ⎛ 2m (E −E
0 0 )⎞
jx ⎜ ⎟ − jx ⎜ ⎟
⎜ 2 ⎟ ⎜ 2 ⎟
⎝ ⎠ ⎝ ⎠
ψ (x) = Ae
h h
+ Be
We wish to show this is a solution to Equation (1.37):
d 2ψ (x) 2m0
+ 2 (E − E0 )ψ (x) = 0 . We begin by finding the derivatives:
dx2 h
dψ (x) ⎡ 2m0 (E − E0 ) ⎤ jx 2 m 0 (E2 −E 0 ) ⎡ 2 m (E − E )
2m0 (E − E0 ) ⎤ − jx 0 h 2 0
= A⎢ j ⎥e h
+ B⎢ − j ⎥e
dx ⎢⎣ h2 ⎥⎦ ⎢⎣ h2 ⎥⎦
and
2 2
d 2ψ (x) ⎡ 2m (E − E ) ⎤ jx 2 m 0 (E2− E0 ) ⎡ 2 m (E − E )
2m 0 (E − E0 ) ⎤ − jx 0 h 2 0
2 =A j ⎢ 0
2
0
⎥ e h
⎢
+ B −j 2
⎥ e
dx ⎢⎣ h ⎥⎦ ⎢⎣ h ⎥⎦
2 m 0 (E − E 0 ) 2 m 0 (E − E0 )
⎛ 2m 0 (E − E0 )⎞ jx h2
⎛ 2m0 (E − E0 )⎞ − jx h2
= −A⎜ 2 ⎟e + B⎜ 2 ⎟e
⎝ h ⎠ ⎝ h ⎠
⎧ jx 2 m0 (E −E 0 ) 2 m 0 (E − E0 ) ⎫⎡ 2m E − E ⎤
⎪ ⎪ 0( 0)
− jx
= ⎨Ae h2
+ Be h2
⎬⎢ − 2 ⎥
⎪⎩ ⎪⎭⎣ h ⎦
Substituting into Equation (1.37) we obtain

⎛ jx 2m0 (E − E0 ) − jx
2m 0 (E − E0 ) ⎞
⎛ 2m0 (E − E0 )⎞
⎜ Ae h2
+ Be h2
⎟⎜ − ⎟+
⎜ ⎟⎝ h2 ⎠
⎝ ⎠
2m0 (E − E0 ) ⎛ jx ⎞
2m 0 (E − E0 ) 2m 0 (E − E0 )
− jx
h2 h2
⎜ Ae + Be ⎟ =0
h
2
⎜ ⎟
⎝ ⎠
0=0
From Equation (1.40), K can be taken as either the positive or negative
root. If we choose the opposite sign, then from Equation (1.38),
ψ (x) = Ae jKx + Be − jKx , it would be the same as changing the sign of x. Thus
the sign of K reverses the direction of propagation of the wave.
1.14. (a) Calculate the de Broglie wavelength of

i) a free electron with 1 eV of kinetic energy
p2
EK = 1eV = 1.6 × 10−19 J =
2m
kg ⋅ m
p = 2(9.11×10−31 kg )(1.6 × 10−19 J ) = 5.39 × 10−25
sec

Solutions Chapter 1
h 6.63 × 10 −34 J − s
λ= = = 1.23 nm
p 539 × 10 −27 kg ⋅ m/ s
b) an electron with 10 keV of kinetic energy
⎛ 1.6 × 10−19 J ⎞ −15

EK = 1 × 104 eV ⎜ ⎟ = 1.6 × 10 J
⎝ eV ⎠
p = 2mEK = 2 ( 9.1 × 10−31 kg )(1.6 × 10−15 J ) = 5.39 × 10−23 kg ⋅ m / s
λ = 0.0123nm
iii) a tiny, drifting dust particle of mass 1 µg and a leisurely velocity of 1

mm/s
m = 1µ g = 10−9 kg
v = 1mm / s = 10−3 m / s
p = mv = 10−12 kg − m / s
λ = h / p = 6.63 × 10−22 m
iv) yourself, walking at 4 mph on your way to class..
self, about 70 kg, v =4mph=1.8m/s
p=mv=70 kg×1.8m/s=126 kg-m/s
λ=h/p=5.26×10-32m
(b)What is a typical size of an atom? You begin to see why quantum

mechanics and the wave description are not useful for large objects.
An atom is on the order of 10-10 m. Even the 10keV electron's wavelength
is a 10th of that, and the dust particle's wavelength is ridiculously small.

Solutions Chapter 1
1.15. What is the wave length of an electron at the bottom of the E-K
relationship of Figure 1.13? What is its kinetic energy there?
2π
At the bottom of the parabola, K=0, and since λ = , λ is infinity. .
K
Note,however that for a finite crystal, the point at K=0 is not permitted
since the electron wave length cannot exceed the crystal dimension. Since
the kinetic energy is the total energy minus the potential energy (E0), the
kinetic energy is zero.
1.16. Consider the E-K diagram shown in Figure 1P.3:

E
Kx
5π 4π 3π 2π π π 2π 3π 4π 5π
− − − − −
a a a a a a a a a a
a) Verify that it meets the required criteria:
i) E(K) is periodic in K-space with period 2π/a
ii) Equivalent extrema exist at K=0, ±2π/a, ±4π/a…
iii) Equivalent extrema exist at K=±π/a, ±3π/a, ±5π/a…
iv) The slope of the E-K curve is zero at K=0, ±π/a, ±2π/a, …
b) Indicate the first Brillouin zone.

Solutions Chapter 1
E
first
Brillouin
zone Kx
5π 4π 3π 2π π π 2π 3π 4π 5π
− − − − −
a a a a a a a a a a
c) Sketch the corresponding vg-K diagram.
1 dE
The group velocity vg = is proportional to the slope of E:
h dK
vg
5π 4π 3π 2π π π 2π 3π 4π 5π Kx
− − − − −
a a a a a a a a a a
d) In what regions of the E-K diagram are electrons most likely to be

found for this material?
Electrons will be at the lowest energies in general, so for this material they
will be mostly found near the edges of the Brillouin zones, but not at K=0.
most likely
E to find electrons
first
Brillouin Kx
zone
5π 4π 3π 2π π π 2π 3π 4π 5π
− − − − −
a a a a a a a a a a
1.17. Explain the analogy between using a conducting layer thicker than the
skin depth to shield a region of space from electromagnetic waves, and the
ability of an electron to penetrate a potential barrier.
Although there cannot be an electric field in a bulk metal, electromagnetic

radiation will penetrate a short distance into the metal (skin depth). To
prevent the radiation from penetrating the metal and leaking through to the
other side, the metal should be thicker than the skin depth by some safety
factor.

Solutions Chapter 1
Similarly, an electron wave cannot exist in a bulk material whose potential
energy is higher than that of the electron, but the electron can penetrate a
short distance into this forbidden region. If the barrier is thin enough, the
electron can leak through to the other side.
1.18. The infinitely thick potential barrier of Figure 1.18 (a) can be considered a
crude approximation to the potential barrier at the surface of a semiconductor
(see Figure 1.12).
(a) How, then, might you construct a thin potential barrier like that in
Figure 1.18 (b)? Thin potential barriers are used in a wide variety of
semiconductor devices, including tunnel diodes, contacts, and field
effect transistors.
One could place two pieces of semiconductor close together with a narrow
air gap (another insulator will do) between them.
(b) How would you construct a potential well (thin region of lower
potential energy bounded by region of higher potential energy?
Potential wells are widely used in lasers, photodectors, and
heterojunction bipolar transistors.
In this case one would use a thin layer of semiconductor with air on either
side.
1.19. a) From the Bohr model, what emission wavelength would you expect
for a transition in hydrogen from E2 to E1? Transitions ending at E1 are
collectively called the Lyman series, and are generally found in the ultraviolet
region of the spectrum.
E2 − E1 = Evac − 3.4 − (E vac − 13.6) = 10.2eV

The wavelength of this transition is
hc ( 6.63 × 10 J − s )( 3.0 × 10 m / s )
−34 8
λ= = = 122nm
E 10.2eV (1.6 ×10−19 J / eV )
b) What emission wavelength would you expect from a transition from E4 to

E3? This is the first emission line in the Paschen series, and is in the infrared.
E4 − E3 = Evac − 0.85 − (Evac − 1. 51) = 0.66eV
,
hc
and λ = = 1.9µ m
E
1.20. What wavelength of light should you shine on hydrogen to cause

electrons to go from E1 to E2 by optical absorption?

Solutions Chapter 1
From Problem 1.19, we know that the energy corresponding to this
transition is 10.2 eV, or λ= 122nm.
b) What would happen if you passed a beam of λ=430 nm through a hydrogen

sample? Explain your answer.
From the Golden Rule, light of wavelength λ=430 nm has an energy of
1.24 1.24
E(eV ) = = = 2.88eV . It could not be absorbed, because
λ (µm) 0.43µm
there is no combination of allowed initial and final states corresponding to
this wavelength or energy. Thus, hydrogen is transparent to this radiation.
1.21. In discussing Figure 1.19 (a), we pointed out that in a material with a
band gap of 2.5 eV, an electron near the top of the valence band could not
absorb a photon of energy 2.06 eV, since it would have to end up at a
forbidden energy state.
a) What about an electron deep in the valence band, more than 2.06 eV
below the band edge EV? Why is it unlikely for this electron to absorb the
photon?
The electron must have an empty state at 2.06eV above its current state.
If this electron is deep in the valence band, the desired new state will
exist, but it is most likely occupied, since it is in the valence band.
b) Why is unlikely for a photon of 2.06 eV to be absorbed by an electron

in the conduction band?
For this absorption to take place, there must be an electron in the
conduction band. The states in the conduction band are mostly empty, so
the number of electrons available there for this absorption process is
small.
1.22. For a simple cubic crystalline structure of lattice constant a=0.50nm,

a) How many atoms are there per unit volume? (Hint: an easy way to
proceed is to calculate the volume of the unit cell, and the number of atoms
per unit cell. Although there are eight atoms involved in any given unit cell,
each atom in the simple cubic structure is part of eight different cells, one
corner of each. Thus, there are 8 atoms×1/8 atom per corner or 1 atom per unit
cell.)
The volume of the unit cell is the volume of the cube, or (0.5×10-9 m)3, or
1.25×10-28 cubic meters. The density of atoms is thus 1/1.25×10−23 m3, or
8×1027 atoms/cubic meter.
b) How many atoms per unit area are there in the (100) plane? The (110)
plane? The (111) plane?
In the (100) plane, we have a grid of atoms on squares a on a side.
Each square contains 1/4 of an atom at each corner, or 4 times 1/4 atoms
per area a2. Thus the area density of atoms is 1/a2.

Solutions Chapter 1
In the 110 plane, the atoms appear arranged in rectangles that are a by
a⋅a 2
The area density for the (110) plane is 1/ a 2 2 .
For the (111) plane, the atoms are arranged in regular pattern, the
fundamental element of which is an isosceles triangle 2a on a side.
1 1a 2 a2
The area of a triangle is bh = (a 2) = . But, each triangle
2 2 2 2
⎛1⎞
⎜ ⎟
contains 1/6 of at atom at each point, so the area density is ⎝ 2 ⎠ = 2 .
6 1
⎛ a ⎞ 3a
⎜ ⎟
⎝ 2 ⎠
c) What if the lattice is FCC instead (still with a=0.5 nm)? Now how
many atoms per unit volume are there?
The volume of the unit cell is still the volume of the cube, but now the
number of atoms per unit cell is different. There are still 8 corner atoms,
each contributing 1/8 of an atom to the unit cell, but there are also 6 face
atoms. Each of those face atoms is half in a given cell, and half in an
adjacent one. Thus the number of atoms/unit cell is 8×1/8 + 6×1/2=4
atoms per unit cell. The number of atoms per unit volume is thus 4
atoms/1.25×10-28m3=32×1027 m-3.

Solutions Chapter 1
SOLUTIONS CHAPTER 2
2.1. Of the two materials whose E-K diagrams are shown in Figure P2.1, which
will have the lowest effective mass for electrons? Which will have the lowest
effective mass for holes? Explain how you arrived at your conclusion.
E E
C C
E E
E E
V V
K K
The effective mass is proportional to the reciprocal of the curvature of the

band. The band with the greatest curvature therefore has the smallest
effective mass. The E-K diagram on the right has a higher curvature in
the conduction band, thus the effective mass for electrons in the
conduction band is smaller than for the material on the left.
For holes, the material on the left has higher curvature and thus the
effective mass for holes is smaller for the material on the left.
2.2. Assume a material has an E-K diagram given by

E(K )conduction = EC + E1 sin 2 (Ka)
E(K) valence = EV − E2 sin 2 (Ka)
Let a=0.5nm, E1=5eV, and E2=4eV.
a) Sketch the E-K diagram for the first Brillouin zone. Label the axes
completely.

Solutions Chapter 2
E
5 eV
E
C
EV
4eV
−π/a π/a K
b) What is the effective mass for an electron near the bottom of the
conduction band?
We use the expression for the effective mass :
−1
⎛⎜ d 2 E ⎞
m* = h 2
⎝ dK 2 ⎠
K =0
Taking the derivatives:
−1
⎡ d2 ⎤
⎢⎣ dK 2 ( C
m* = h 2
E + E1 sin 2 (Ka))
⎥⎦
K= 0
−1
⎡ d
= h 2 E1 (2asin(Ka)cos(Ka))⎤
⎣ dK ⎦
K =0
[
= h 2 E1 2a(−asin 2 (Ka) + acos 2 (Ka)) ]
−1
K= 0
2
h
=
2E1 a2
⎛⎜ 6.63 × 10 −34 Js ⎞
2
⎝ 2π ⎠
=
2(5eV ⋅1.6 × 10 J / eV )(5 × 10 m)
−19 −10 2
−32
= 2.8 × 10 kg = 0.03m0
c) What is the effective mass for holes near the top of the valence band?

Solutions Chapter 2
−1
⎡ d2 ⎤
m* = h
⎢⎣ dK 2
2
(EV − E2 sin 2 (Ka))
⎥⎦
K= 0
−1
⎡
(−2asin(Ka)cos(Ka))⎤
d
= h 2 E2
⎣ dK ⎦
K =0
[
= h 2 E2 2a(+asin 2 (Ka) − a cos2 (Ka)) ]
−1
K =0
−h
2
=
2E2 a 2
⎛ 6.63 × 10 −34 Js ⎞
2
−⎜
⎝ 2π ⎠
=
2(4eV ⋅1.6 × 10 J / eV )(5 × 10 m)
−19 −10 2
−32
= −3.5 × 10 kg = −0.04m0
This is the effective mass for electrons, which is negative as expected.

The effective mass for holes is positive, or mh = +0.04m 0
*
2.3. Explain in your own words why the effective mass of an electron is
different than its mass in vacuum.
The effective mass includes the effects of the periodic potential of the
crystal, so that Newtonian or classical mechanics can be used. If a given
external force is applied, the effect on the electron is the sum of the
external force plus all the internal forces due to the varying potential
energy in the crystal. The effective mass can be used to predict the
behavior of an electron in the crystal in reaction to the external force.
2.4. For the material of Figure P2.2, there are two different effective masses.
Which effective mass will be the one displayed by most of the electrons in the
conduction band? Why?

Solutions Chapter 2
E
C
E
E
V
Most of the electrons in the conduction band will be at the low energies
near the bottom of the band. Thus most of them will have the effective
mass corresponding to K=0 in this case.
2.5 Consider the energy band diagram in Figure 2P.3.

EC
1 eV
E
EV
2 µm
x
a) Find the electric field, and express the result in V/m. What direction
does it point?
Electric field is voltage/distance. The potential energy changes by 1 eV
over a distance of 2µm. Since the voltage ∆V=-q∆EP, the electric potential
change is
⎛ −1eV ⎞
(
∆V = − q∆EP = 1.6 × 10−19 C ⎜) −19 ⎟ = 1V .
⎝ 1.6 × 10 eV / J ⎠
The electric field is
dV −1V
E =− = −6
= −5 × 105V / m .
dx 2 × 10 m
The field points to the left (uphill).

Solutions Chapter 2
b) Find the force on the electron. In what direction is the electron
accelerated?
( )
From F = − qE = −(1.6 × 10−19 C ) −5 × 105V / m = 8.0 × 10−14 N . The force is to
the right.
c) Find the force on the hole. In what direction is the hole accelerated?
The force on the hole is
( )( )
F = + qE = +1.6 ×10−19 C 5 ×105V / m = −8.0 × 10−14 N , to the left.
2.6. For an optical emission to occur, four things are needed:

a) An electron must be at an elevated energy state
b) There must be an empty state at a lower energy for the electron to go
to
c) Energy must be conserved (energy of electron before emission =
energy of electron after emission + energy of photon).
d) K must be conserved.
It turns out that photons have negligible K compared with electrons of interest
in semiconductors. That implies an electron making an optical transition must
end up at the same value of K as when it started (the energy can change, since
the energy difference goes to the photon). On the basis of your expectation of
where electrons and holes are likely to be found, and the four conditions
above, which material will be a more efficient light emitter, Si or GaAs? Their
E-K diagrams are shown in Figure 2P.4. Explain your reasoning. Silicon is an
indirect gap material, meaning the conduction band minimum is not at the
same value of K as the maximum in the valence band. GaAs is a direct gap
material.
E E
EC EC
0 2π/a 0 2π/a
K <100>
(a) GaAs
(b) Silicon
If K is to be conserved, the total K of the electron plus the photon must be
equal to the K of the electron after the absorption (the photon is
annihilated). Since photons have nearly zero K, the electrons end up at
(almost) the same value of K as when they start, or that transitions on the
E-K diagram must be vertical. For optical emission, electrons must be in
the conduction band and holes must be available in the valence band.
Since most electrons are found at the bottom of the conduction band, the
most probably transition is from there to the top of the valence band,
where the empty states are expected to be. For silicon, the bottom of the
conduction band is at a different value of K than the top of the valence
band, so emission is unlikely.

Solutions Chapter 2
2.7. Consider the energy band diagram in Figure 2P.5.
0.2 eV
E EC
1 eV
EV
0.1 eV
x
i) What is the potential energy of the electron?
The potential energy for an electron near the bottom of the conduction
band is EC.
ii) What is the kinetic energy of the electron?

The kinetic energy for the electron is the total energy minus the potential
energy, in this case 0.2 eV.
iii) What is the potential energy of the hole?
The hole’s potential energy is EV.
iv) What is the kinetic energy of the hole?
The hole energy increases downward, and the total hole energy is 0.1 eV
below the valence band edge, so the kinetic energy of the hole is 0.1 eV.
v) The vertical axis is energy-- meaning the electron energy. What
direction represents increasing hole energy, up or down?
Hole energy increases downward. Holes trying to reach a lower energy
bubble upward on the (electron) energy band diagram.
2.8. A sample of GaAs is subjected to an electric field of 100V/cm.

a) What is the velocity of an electron in the conduction band if it is
initially at rest and is accelerated by this field for a period of 0.1 ps? (Use the
conductivity effective mass, since this problem is related to conduction.)
The force on the electron is
⎛ 100cm ⎞
F = − qE = ( −1.6 × 10−19 C ) (100V / cm ) ⎜ −15
⎟ = −1.6 × 10 N
⎝ 1m ⎠
The minus sign just indicates direction. We’ll neglect direction for the rest
of the problem.
From pseudo-classical mechanics, the acceleration is
F F 1.6 × 10−15 N
a= = * = = 2.6 × 1016 m / s 2
m * mce ( 0.067 ) ( 9.1 × 10 kg )
−31
where we have used the conductivity effective mass for electrons.

dv
The velocity after 0.1 ps is found from a = .
dt

Solutions Chapter 2
v sat final
10−13 s
∫
0
dv = ∫
0
adt
2.6 ×1016 dt = (1×10−13 s )( 2.6 × 1016 m / s 2 ) = 2.6 × 103 m / s

10−13 s
v final = ∫
0
b) What is the terminal velocity for a hole under the same conditions?
The force on the hole is the same, since the charge is the same (with a
sign change). The acceleration is different because the effective mass is
different:
F F 1.6 ×10−15 N
a= = = = 0.52 × 1016 m / s 2
m * mch *(m0 ) ( 0.34 ) ( 9.1× 10 kg )
−31
And the corresponding velocity is

0.52 ×1016 dt = (1×10−13 s )( 0.52 × 1016 m / s 2 ) = 0.52 × 103 m / s
10−13 s
v final = ∫
0
The more massive hole is traveling more slowly.
2.9 Near the bottom of the conduction band in a semiconductor , the electron
energy can be expressed as E = EC + AK where A is independent of K and is
2
positive.
a) Find an expression for the electron effective mass in terms of A.
From Equation (2.11), the effective mass is
−1
⎛ d 2E ⎞ 2
m *= h ⎜
⎝ dK 2 ⎟⎠
dE d 2E
The derivatives are = 2AK ; = 2A
dK dK 2
h2
Thus m *= .
2A
b) Find an expression for the electron velocity as a function of E.

The electron velocity is the group velocity given by Equation (2.8), or
1 dE 2 AK
vg = = .
h dK h
2.10. Consider a semiconductor material whose three-dimensional E-K

relationship has the same curvature in the x and z directions but a different
curvature in y. Let us consider the x and y directions, and let the relationship
be
E( x, y ) = AKx + BK y .
2 2

Solutions Chapter 2
(a) Using a computer program such as MATLAB, produce a mesh
plot of this function for B=2A. Rotate the perspective to get an
understanding of the shape of this function.
Here is some MATLAB code that will generate the plot
function problem210
Kx = 20;
Ky = 20;
A = 1;
x = -Kx:Kx;
y = -Ky:Ky;
for i= 1:2*Kx+1
for j = 1:2*Ky+1
E(i,j) = A *((x(i)*x(i) + 2*y(j)*y(j)));

end
end
mesh(x,y,E)
view(40,10);

Solutions Chapter 2
(b) In which direction is the effective mass smaller?
The smaller effective mass is associated with the large curvature.
d 2E d 2E
= 2 A = 2B = 4A . Thus the curvature is greater in the y-
dKx2 dK 2y
direction and the effective mass is smaller for an electron traveling in the y
direction.
(c) What determines which effective mass a given electron actually has in
this crystal?
The direction in which the electron is going.
2.11. For each of the following, identify whether the impurity will produce n-type or
p-type material, or be amphoteric:
a) Phosphorus replacing a gallium atom in gallium arsenide
Phosphorus is in Column V, and if it replaces gallium, which is in Column

III, then it will act as a donor.

Solutions Chapter 2
b) Zinc in germanium
Zinc is in Column II so it will act as an acceptor.
c) Carbon in indium phosphide
Carbon is in Column IV, so it will be amphoteric- acting as an acceptor

when it replaces phosphorus and acting as a donor when it replaces
indium.
d) Silicon in zinc selenide
Silicon is in Column IV and ZnSe is a II-VI semiconductor. Therefore

silicon will be amphoteric.
2.12. Show that the difference between the intrinsic level and midgap is small,
using GaAs as an example. Compare the result to that of silicon.
The difference between midgap and the actual intrinsic level, from
Equation (2.72), is
E + EV 3 ⎛ m* ⎞
Ei − C + kT ln ⎜ dsh * ⎟
2 4 ⎝ mdse ⎠
E + EV
The term C is the average of these two energies, or midgap. The
2
offset of the intrinsic level from this is
(0.026eV )ln ⎛⎝
3 0.48 ⎞ −2
= 3.8 × 10 eV = 38meV
4 0.067 ⎠
While still small compared to the band gap, this is considerably larger than
the case for silicon (from Example 2.4 the result is 1 meV). The larger
offset in GaAs results from the large difference in the effective masses for
holes and for electrons.
2.13. What is the probability that a state at the conduction band edge of intrinsic
GaAs is occupied at room temperature?
1
The probability function is f (EC ) = EC − Ef , but since the GaAs is
1 + e kT
intrinsic, the Fermi level is well below the conduction band edge, so we
can use the Boltzmann approximation:
EC − Ef
−
f (EC ) = e kt
The Fermi level is 38 meV above midgap, or at EC-(Eg/2 + 0.038):

Solutions Chapter 2
⎡ ⎛ Eg ⎞ ⎤
⎢ EC − ⎜⎜ EC − ⎟⎟ − 0.038⎥
−
⎢
⎣ ⎝ 2 ⎠ ⎦⎥ −
( Eg / 2−0.038) −
(1.43/ 2 − 0.038 eV )
f ( EC ) = e =e kT kT
=e (0.026 eV )
= 4.9 × 10−12 , or very few
states are occupied indeed.
2.14. The donor ground states for tellurium in GaAs are 5.9 meV below the
conduction band edge. (There are two of them because of spin.) At room
temperature, what is the probability that a given ground state is occupied if
the Fermi level is 0.1 eV below EC ?
This donor state is ( EC − E f ) − ( EC − ED ) = ( ED − E f ) = 0.1 − 0.0059 = 0.094eV

above the Fermi level.
The probability of occupancy is
1 1
f ( ED ) = ED − E f
= 0.094
= 0.026
1+ e kT 1+ e 0.026
2.15. What is the probability of occupancy of a state 0.1 eV above the Fermi
level? Below the Fermi level?
The energy difference E-Ef is 0.1eV, which is greater than 3kT, so the
Boltzmann approximation is valid:
E− Ef −0.1
−
f (E ) = e kT
=e 0.026
= 0.021 or a 2.1% chance of occupancy.
For a state 0.1 eV below the Fermi level, E-Ef=-0.1, which is not greater
than 2.3kT. Therefore we must use the full expression
1 1
f (E) = E −E f = −0.1 = 0.98
1+e kT 1+e 0.026
, or 98 percent of these states are
occupied.
2.16. Verify the definition of non-degeneracy: show that if the Fermi level is
2.3 kT below the conduction band edge, the probability of occupancy of the
lowest energy state in the conduction band is 10%.
We have EC-Ef=2.3kT=2.3(0.026eV)=0.0598eV
1 1
f (E) = E−E f
= 0.0598
= 0.091 which is close to 10%.
1 + e kT 1 + e 0.026
2.17. Calculate the effective densities of states for electrons and holes at room
temperature for

Solutions Chapter 2
a) Silicon.
We use the expressions
⎛ * ⎞ ⎛ T ⎞ 3/ 2
3/ 2
−3 m dse
NC = (2.54 × 10 )cm ⎜ ⎟ ⎜ ⎟ and
19
⎝ m0 ⎠ ⎝ 300 ⎠
⎛ m*dsh ⎞ ⎛ T ⎞ 3/ 2
3/ 2
⎟ ⎜ ⎟
−3
NV = (2.54 × 10 )cm ⎜
19
⎝ m0 ⎠ ⎝ 300 ⎠
For silicon, m*dse = 1.09m 0 , thus
NC = (2.54 × 10 )cm (1.09) (1) = 2.89 × 10 cm
19 −3 3/ 2 3/ 2 19 −3
Which is close to the accepted value of 2.86 ×1019 cm −3 . The difference is

due to round-off error.
For the holes, m*dsh = 1.15m0 , so NV = 3.13 × 1019 cm−3
b) GaAs
For GaAs, m*dse = 0.067m0 → N C = 4.41 × 1017 cm −3

And m*dsh = 0.48m0 → NV = 8.45 × 1018 cm −3
c) Ge
We have m*dse = 0.56m0 → NC = 1.06 × 1019 cm −3 and
−3
mdsh = 0.292m0 → NV = 4.01 × 10 cm
* 18
2.18. Compare the effective density of states for electrons, NC and for holes NV
for GaAs.
3/ 2
⎛ m * kT ⎞
From Equations (2.61) and (2.62), N C = 2⎜ dse 2 ⎟ and
⎝ 2π h ⎠
3/ 2
⎛ m * kT ⎞
N V = 2⎜ dsh 2 ⎟ . Thus
⎝ 2πh ⎠
3/ 2
⎛ m* kT ⎞
2⎜ dse 2 ⎟ 3 3
NC ⎝ 2π h ⎠ ⎛ mdse
*
⎞ 2 ⎛ 0.067m0 ⎞ 2
= = ⎜ * ⎟ =⎜ ⎟ = 0.052
⎝ mdsh ⎠ ⎝ 0.48m0 ⎠
3/ 2
NV ⎛ m *dshkT ⎞
2⎜
⎝ 2πh 2 ⎟⎠
or the effective density of states in the conduction band is only about 5%
of that in the valence band.

Equation (2.51):
1
f p (E ) = 1 − E− E f
kT
1+ e

Solutions Chapter 2
Ef − E
We are trying to get a term like e kT

into the denominator. Multiplying the
second term by 1 we write
⎛ E f −E ⎞ E f −E E f −E
⎜ ⎟ =1 −
kT kT kT
1 e e e
f p (E ) = 1 − E − E f ⎜ E f −E E ⎟ Ef − E E − E f E f −E
= 1 − E f −E
kT ⎜ ⎟
1+ e ⎝e kT
⎠ e kT
+e kT
e kT
e kT + 1
We find a common denominator:
⎛ E f −E ⎞ Ef − E E f −E E f −E
+ 1⎟ + 1− e k T
f p (E ) = 1⎜ E − E
kT kT kT
e e e
− E −E =
⎜ f ⎟ Ef − E
⎜ kT ⎟
f
⎝e + 1⎠ e kT + 1 e kT + 1
1
f p (E ) = E f −E
e kT
+1
− (EC − E f ) − (E f − EV )
2.20 Recall that the equations n0 = NC e and p0 = NV e
kT
are only kT
valid for non-degenerately doped semiconductors. One reason is that the

effective densities of states are only defined for non-degenerately doped
materials. What is the other reason?
The Boltzmann approximation is only valid for non-degenerate
semiconductors.
2.21. (a) Silicon is doped with ND=5×1015cm-3. Find n0 and p0 , and locate the
Fermi level. Draw the energy band diagram.
The material will be n-type. Since the number of donors is much greater
than the intrinsic concentration, and assuming all dopants to be ionized,
⎛ EC − E f ⎞
−⎜ ⎟
−3 ⎝ kT ⎠
then n0=ND. Thus n 0 = 5 × 10 cm = N C e 15
assuming non-degenerate
material (which we will check later).
⎛ND ⎞ ⎛ 5 × 1015 ⎞
So EC − E f = −kT ln ⎜ ⎟ = −0.026eV ln ⎜⎝ ⎟ = 0.22eV
⎝ NC⎠ 2.86 × 1019 ⎠
Since 0.22>2.3kT, the assumption of non-degeneracy is verified.

Solutions Chapter 2
(ni ) 2 (1.08 × 1010 ) 2
p0 = = = 2.3 × 104 cm -3
n0 5 × 1015
b) A new batch of silicon is doped with boron to NA=4×1015 cm-3. Find n0 and
p0 and locate the Fermi level. Draw the energy band diagram.
Again we assume non-degenerate material and complete ionization. thus
p0=4×1015, and
⎛N ⎞ ⎛ 4 × 1015 ⎞
E f − EV = −kT ln ⎜ A ⎟ = −0.026eV ln ⎜ = 0.23eV
⎝NV ⎠ ⎝ 3.1× 1019 ⎟⎠
E
C
E
i
E 1.12eV
f 0.23eV
E
V
(ni ) 2
n0 = = 2.9 × 104 cm -3
p0
2.22. Silicon is doped with NA=4×1017cm-3. Find n0 and p0. Draw the energy band
diagram indicating the position of the Fermi level.
We observe that NA>>ni, so we take p0=NA. Then, if the material is not

degenerately doped, we can use
ni2 (1.08 × 10 )
10 2
−3
n0 = = = 292cm
4 × 10
17
p0
Once the carrier concentrations are known, we can find the Fermi level
from the majority carrier concentration:
⎛ p ⎞ ⎛ 4 × 1017 ⎞
E f − EV = −kT ln ⎜ 0 ⎟ = − 0.026 ln ⎜ 19 ⎟
= 0.11eV
⎝ NV ⎠ ⎝ 3.1× 10 ⎠
This exceeds 2.3kT so our assumption of non-degeneracy is correct.

Solutions Chapter 2
2.23. GaAs is doped with ND=1015 cm-3 and NA=4×1014 cm-3. Draw the energy band
diagram and locate the Fermi level.
To locate the Fermi level, we need to find n0 and p0. The intrinsic
concentration ni for GaAs is 2.2×106cm-3, so NA and ND are both >>ni.
Thus we can say that n0=ND-NA=1015-4×1014=6×1014 cm-3.
If the material is not degenerately doped, then

ni2 ( 2.2 × 10 )
6 2
−3
p0 = = = 8.1 × 10 cm −3
n0 6 × 1014
⎛ n ⎞ ⎛ 6 ×1014 ⎞
EC − E f = −kT ln ⎜ 0 ⎟ = − 0.026eV ln ⎜ 17 ⎟
= 0.17eV
⎝ NC ⎠ ⎝ 4.4 × 10 ⎠
2.24. Ge is doped with 1015 Sb/cm-3. Assuming Sb to be a shallow donor, find n0 and
p0 at room temperature. Draw the energy band diagram.
N D = 1015 cm−3

Solutions Chapter 2
Is ND >> ni ? For Ge, nI=1.64×1013cm-3 (see next problem), so this is OK.
Since the donors are shallow, we take them to be all ionized ( n0 = N D ), so

if the material is not degenerately doped,
⎛N ⎞ ⎛ 1015 ⎞
EC − E f = − kT ln ⎜ D ⎟ = −0.026eV ln ⎜ 19 ⎟
= 0.24eV
⎝ NC ⎠ ⎝ 1.05 ×10 ⎠
This is still greater than 2.3kT=0.06eV so the assumption of non-
degeneracy is OK.
Then
n0 = N D = 1015 cm -3
ni2 (1.65 × 1013 ) 2
p0 = = 15
= 2.71 × 1011 cm -3
n0 10
2.25. a) Calculate the intrinsic carrier concentration for Ge at room
temperature.
⎛ − Eg ⎞
We use the relationship ni = NC NV ⎜ e 2kT ⎟
⎝ ⎠
For Ge, Eg=0.67eV, resulting in
−3 −3
⎛ − 2(0.026
0.67V
⎞ −3
ni = 1.06 ×10 cm ⋅ 4.01× 10 cm ⎜ e
19 18 eV )
⎟⎟ = 1.64 × 10 cm
13
⎜
⎝ ⎠
b) Make a semilogarithmic plot of nI vs. temperature over the range 200-
500K. The intrinsic carrier concentration should be on the vertical
(logarithmic) axis. Add curves for silicon and GaAs.
Note that the densities of states are functions of temperature.

Solutions Chapter 2
2.26. A diode is made by using an n-type Si substrate (ND=1015cm-3) and then
16 -3
adding acceptors to a concentration of NA=10 cm to one region. What are p0
and n0 in the p-type region?
We assume room temperature since T is not given. Therefore we can

assume all donors and acceptors are ionized, unless it turns out that the
Fermi level is too close to the dopant energy levels, which we will check
later.
For the p region:
p0 = N A − N D = 1016 − 1015 = 9 × 1015 cm−3
ni2 (1.08 × 10 )
10 2
−3
n0 = = = 1.3 × 10 cm
4
9 × 10
15
p0
To check the assumption of non-degeneracy, we find the Fermi level:
⎛ p ⎞ ⎛ 9 × 1015 ⎞
E f − EV = −kT ln ⎜ 0 ⎟ = (−0.026eV )ln ⎜ = 0.2eV , which is also
⎝ NV ⎠ ⎝ 3.1 × 1019 ⎟⎠
much greater than 2.3kT.
16 -3 15 -3
2.27. A sample of InP is doped with ND=10 cm and NA=10 cm .
a) Find n0, p0.
Assuming room temperature, all dopants are ionized. The material is net
n-type, so we solve for the electrons concentrations first. We have
n0= N D' =ND-NA=1016-1015=9.0×1015 cm-3.

Solutions Chapter 2
If the material is not degenerately doped, then we can find p0 from
n0p0=ni2. We will check degeneracy in part (b) of the problem. For InP,
ni2 (1.97 × 10 cm )
7 −3 2
7 -3 −2 −3
nI=1.97×10 cm . Therefore p0 = = −3 = 4.3 × 10 cm
9.0 × 10 cm
15
n0
n0=9.0×1015 cm-3 and p0 = 4.3×10-2 cm-3.
b) Locate Ef.
If the material is non-degenerate, we can use the Boltzmann

approximation (we have to check the degeneracy later) and thus use the
equation
−
(E C − E f )
n0 = NCe kT , or, looking up NC=2.02×1018 cm-3 for InP, we have
⎡n ⎤ ⎡ 9 × 1015 cm −3 ⎤
EC − E f = − kT ln ⎢ 0 ⎥ = −0.026eV ln ⎢ −3 ⎥
= −0.026eV ln ( 0.0045 )
⎣ 2.02 × 10 cm ⎦
18
⎣ NC ⎦
= −0.026eV ⋅ ( −5.4 ) = +0.14eV
We note in the process that EC-Ef is greater than 2.3kT, so we can get by
with the use of the Boltzmann approximation and we have also justified
our assumption of non-degeneracy in part (a).
EC-Ef=0.14eV
c) Sketch the energy band diagram. Label both axes and identify EC, EV Ef, Ei,
and Evac on your drawing. Indicate the electron affinity and the band gap.
We look up the band gap for InP and find it to be 1.35 eV, and the electron
affinity χ is 4.35 eV. Thus the energy band diagram looks like:

Solutions Chapter 2
2.28. A sample of silicon is doped with ND=4×1016 cm-3 and NA=8×1015 cm-3. Find
the equilibrium concentrations of electrons and holes, and locate the Fermi level.
Both ND and NA are much greater than ni, and these concentrations are not
great enough to degenerately dope the sample. Thus n0=ND-NA=4×1016-
8×1015=3.2×1015 cm—3. From the law of mass action,
n 2i (1.08 × 10 )
10 2
−3
po = = = 3.6 × 10 cm and
4
3.2 × 10
15
n0
⎛ n ⎞ ⎛ 3.2 × 1015 ⎞
EC − E f = −kT ln ⎜ 0 ⎟ = −0.026eV ⋅ ln ⎜ = 0.23eV
⎝NC⎠ ⎝ 2.86 × 1019 ⎟⎠
2.29. How heavily would you need to dope silicon with donors to violate the
assumption of non-degenerate doping? How many acceptors would be needed
to just cause a degenerately doped type material?
The threshold of degenerate doping occurs when EC-Ef=2.3kT. At this
threshold, we will take the Boltzmann approximation to apply (just barely).
Thus
−(EC − EF )
N D = n0 = N C e kT = N C e = (2.86 × 10 cm )(.10) = 2.86 × 10 cm

−2.3 19 −3 18 −3
For p-type we have

−(E f −EV )
N A = p0 = N V e kT = N V e −2.3 = (3.1 × 1019 cm−3 )(.10) = 3.1× 1018 cm−3
2.30. What donor concentration is required to elevate the Fermi level in Si to

2.3kT above Ei? What acceptor concentration will lower Ef 2.3kT below Ei?
For this problem it is convenient to use Equations (2.75):
E f − Ei
n0 = ni e kT
Ei − E f
p0 = ni e kT
For n-type, assuming complete ionization, we will have n0=ND and

2.3 kT
n0 = N D = (1.08 x10 )e 10 kT
= (1.08 x1010 ) x10 = 1.08 x1011 cm −3
For p-type, we have
p0 = N A = (1.08 x1010 )10 = 1.08 x1011 cm −3

Solutions Chapter 2
2.31 Given that for a particular sample of silicon ND =1015 cm-3 and n0=106 cm-3,
find NA and p0.
We know that (assuming non-degenerately doped material), that
n0 p0 = ni (Equation (2.66)).
2
Thus
n2
p0 = i = 1.2 × 1014 . Assuming all dopants are ionized, from Equation
n0
(2.90),
p0 + N D = n0 + N A
N A = p0 + N D − n0 = 1.2 × 1014 + 1015 − 106 = 1.1 × 1015
2.32. Manganese makes a donor trap state 0.53 eV below the conduction band
edge in silicon. If the silicon is doped with ND=4x1016 cm-3, what is the
probability of occupancy of the trap state? Assume the concentration of Mn is
small enough not to affect the overall doping and that the trap state is single-
valued.
We find the Fermi level:
⎛ n0 ⎞ ⎛ 4 × 1016 ⎞
EC − E f = −kT ln ⎜ ⎟ = −0.026 ln ⎜ 19 ⎟
= 0.17eV
⎝ NC ⎠ ⎝ 2.86 × 10 ⎠
The probability of occupancy of the Mn trap state is
1 1 1
f (E Mn ) = E Mn − E f = ( E C − E f ) −( E C − E Mn ) = 0.17− 0.53 ≈1
1+ e kT
1+e kT 1+e kT
So, the manganese states are essentially all occupied.
2.33. A sample of Si doped with phosphorus has its Fermi level 0.2 eV below
EC and the donor ground state is 0.045 eV below EC. Find n0.
Since the donor level is well above the Fermi level, we can take all the
dopants to be ionized. Thus
(E C −E f ) 0.2 eV
− 19 − −3
n 0 = N C e kT = 2.86 × 10 e 0.026 eV = 1.3 × 10 cm
16
2.34. The probability that an energy state at the conduction band edge ( EC ) of
Si is occupied is 10-4.
a) Is this Si n-type, p-type, or intrinsic?

Solutions Chapter 2
1
f (EC ) =
−4
EC − E f = 10
1+e kT
⎛ 1 ⎞
Thus EC − E f = kT ln ⎜ −4 − 1⎟ = 0.026eV ln ( 9,999 ) = 0.24eV . Since the
⎝ 10 ⎠
band gap of silicon is 1.12 eV, this place the Fermi level above midgap
and the material is n-type.
b) Find ND-NA.
Assuming ND-NA>>ni (which it will be since the Fermi level is well above
the intrinsic level),
( EC − EF ) 0.24
− −
−3
N D − N A = n0 = N C e kT
= 2.86 × 10 cm e
19 0.026
= 2.8 × 1015 cm −3
2.35. The effective density of states function NC for a given semiconductor is

7×1017 cm-3. What is the electron density-of-states effective mass?
3 3
⎛ m* ⎞ 2 ⎛ T ⎞ 2
From Equation (2.63), N C = 2.54 × 1017 ⎜ dse ⎟ ⋅ ⎜ ⎟ cm or at T=300K,
-3
⎝ m0 ⎠ ⎝ 300 ⎠
2 2
m *
⎛ NC ⎞ ⎛ 7 × 1017 ⎞ 33
=⎜dse
⎟ = ⎜ 19 ⎟
= 0.09
m0 ⎝ 2.54 × 1019 ⎠ ⎝ 2.54 × 10 ⎠
or mdse = 0.09 m0
*
2.36. Given two semiconductors A and B, let them have the same density-of-
states effective masses. Let Eg = 1 eV for A and 2 eV for B. Find the ratios of the
intrinsic carrier concentrations.
From Equation (2.67),
E gA
Eg − E gB − E gA 2−1
− n N CN V e 2kT
n i = N CN V e 2kT
. Thus iA = EgB =e 2kT
=e 2(0.026)
= 2.2 × 108
niB −
2kT
N CN V e
2.37. Semiconductor devices for many applications must be able to withstand

and operate over a wide range of temperatures, to operate from Antarctica or
deep space to a tropical climate in a hot truck. For example, military
specifications for semiconductor devices cover the range -55°C to +150°C.
Repeat Problem 2.26 at these two temperature extremes.
From Problem 2.26, ND=1015cm-3 and NA=1016 cm-3.
At –55°C, the temperature is 218K, and kT=0.019eV. From Figure 2.20 we
observe that carrier freezeout is not a serious issue. Thus we can still
assume all the dopants are ionized.
In the p-region

Solutions Chapter 2
p0 = N A − N D = 1016 − 1015 = 9 × 1015 cm −3
To find n0 in the p-region, we can use the law of mass action, but require
the value of ni at this temperature. From Equation (2.65)
⎛ Eg ⎞
− ⎜⎜ ⎟⎟
⎝ kT ⎠
n = N C NV e
2
i To find NC at –55 C, we use Equation (2.63)
3 3
⎛ T ⎞2 −3 ⎛ 218 ⎞
N C = [ N C (T = 300) ] ⋅ ⎜
2
−3
⎟ = 2.86 × 10 cm ⎜ ⎟ = 1.77 × 10 cm
19 19
⎝ 300 ⎠ ⎝ 300 ⎠
and similarly
3 3
⎛ 218 ⎞ 2 19 ⎛ 218 ⎞
NV = [ NV (T = 300 K ) ] ⋅ ⎜
2
−3
⎟ = 3.1× 10 ⎜ ⎟ = 1.92 × 10 cm
19
⎝ 300 ⎠ ⎝ 300 ⎠
Neglecting any changes in band gap with temperature, we find
⎛ Eg ⎞ 1.12 eV
− ⎜⎜ ⎟⎟ −
n = N C NV e
2
i
⎝ kT ⎠
(
= 1.77 × 10
19
)(1.92 × 10 )e 19 8.62 ×10 −5 eV / K (218K ) )
1.12 eV
( )(1.92 × 10 )e
−
= 1.77 × 10 19 19 1.88 ×10−2 e V
= 4.55 × 1012 cm−6
ni = 2.1 × 106 cm −3
Then
n 2 ( 2.1× 10 )
6 2
n0 = i = = 4.9 × 10−4 cm −3
p0 9 ×10 15
The Fermi level is found from Equation (2.60)

⎡p ⎤ ⎡ 9 × 1015 ⎤
E f − EV = −kT ln ⎢ 0 ⎥ = − ( 8.62 × 10−5 eV / K ) ( 218 K ) ln ⎢ 19 ⎥
⎣ NV ⎦ ⎣1.92 × 10 ⎦
= 0.14eV
We justify the use of the Boltzmann approximation by noting the
0.12eV>2.3·(kT)=2.3·(0.0188eV)=0.056eV.
At high temperatures, all dopants are ionized and n0=ND-NA=1015 cm-3 on

the n side. On the p-side, p0=NA-ND=9x1015 cm-3, but to find n0 we need to
find ni again. The new temperature is 273+150=423K.
3 3
⎛ 423 ⎞ 2 ⎛ 423 ⎞ 2
[
N C = N C ( T = 300) ⋅ ⎜ ⎟
⎝ 300 ⎠ ] = 2.86 × 1019 cm −3 ⎜ ⎟
⎝ 300 ⎠
= 4.79 × 1019 cm −3
3 3
⎛ 423 ⎞ 2 19 ⎛ 423 ⎞
NV = [ NV (T = 300 K ) ] ⋅ ⎜
2
−3
⎟ = 3.11 × 10 ⎜ ⎟ = 5.21 × 10 cm
19
⎝ 300 ⎠ ⎝ 300 ⎠
⎛ Eg ⎞ 1.12e V
− ⎜⎜ ⎟⎟ −
( )(5.21× 10 )e eV / K (423K ) )
−5
⎝ kT ⎠
n = N C NV e = 4.79 × 10
2 19 19 8.62×10
i
= 1.14 × 10 26 cm −6
ni = 1.07 × 1013 cm−3

Solutions Chapter 2
n 2 (1.07 × 10 )
13 2
n0 = i = = 1.3 ×1010
p0 9 × 10 15
We observe a great difference in the minority carrier concentration due to

increased ni.
The Fermi level is found from
⎡ p0 ⎤ ⎡ 9 × 1015 ⎤
⎥ = − ( 8.62 × 10 eV / K ) ( 423K ) ln ⎢
−5
E f − EV = − kT ln ⎢ 19 ⎥
⎣ NV ⎦ ⎣ 5.21× 10 ⎦
= 0.32eV
2.38. In a non-degenerate semiconductor, the electron distribution peaks at
EC+kT/2 as indicated in Figure 2.17, and the average electron kinetic energy is
at 3kT/2. For intrinsic material, find the probability at a state at these two
energies is occupied. Repeat for EC+10kT.
The probability of occupancy is (non-degenerate) when the Boltzmann

approximation applies
For the average kinetic energy of 3kT/2, the probability of the state being
occupied is
⎛ 3(0.026) ⎞
⎜ + 0.56 ⎟
⎡ 3 kT ⎤
⎢ + ( EC − Ei ) ⎥ −⎜ 2 ⎟
⎣ 2 ⎦ ⎜ 0.026 ⎟
− ⎜ ⎟
f (E) = e kT
=e ⎝ ⎠
= 1 × 10−10
⎛ 0.026 ⎞
⎜ + 0.56 ⎟
KT −⎜ 2 ⎟
+ ( EC − E f )
− 2 ⎜ 0.026 ⎟
KT ⎜ ⎟
For E = EC + , fE = e =e = 2.7 × 10−10
KT ⎝ ⎠
2
For a state at 10kT above the conduction band edge, the probability of
finding an electron at such a high energy is
⎡⎣10 kT + ( EC − Ei ) ⎤⎦ ⎛ 10×.026 + 0.56 ⎞
− −⎜ ⎟
f (E) = e kT
=e ⎝ 0.026 ⎠
= 2 × 10−14
2.39. Calculate the electron concentrations in Si and GaAs if the Fermi level is
0.2 eV below the conduction band edge.
Beginning with Si, we find from Equation (2.59) that

⎛ EC − Ef ⎞
−⎜ ⎟ −
0.2eV
⎝ kT ⎠ −3 −3
n0 = NCe = 2.86 × 10 cm e = 1.3 × 10 cm
19 0.026eV 16
For GaAs, the value of NC is different:

⎛ EC − Ef ⎞
−⎜ ⎟ −
0.2eV
⎝ kT ⎠ −3 −3
n0 = NCe = 4.4 × 10 cm e = 2.0 × 10 cm
17 0.026eV 14

Solutions Chapter 2
2.40. Complete the mathematical steps in Example 2.7 and verify the results.
a) The kinetic energy at which most electrons are found is determined by
setting the derivative of the distribution n(EK)=0.
dn(E K ) ⎡ ⎛ 1⎞ − K
E
− K ⎛
E
1 ⎞⎤
= 0 = C ⎢ EK ⎜ − ⎟ e kT + e kT ⎜ ⎟⎥
⎢ ⎝ kT ⎠ ⎝ ⎥
K ⎠⎦
dEK 2 E
⎣
− K ⎡ ⎛ 1 ⎞ ⎛ 1 ⎞⎤
E
0 = e kT ⎢ E K ⎜ − ⎟ + ⎜ ⎟⎥
⎢ ⎝ kT ⎠ ⎝ 2 E ⎠ ⎥
⎣ K ⎦
⎛ 1 ⎞ ⎛ 1 ⎞
EK ⎜ − ⎟ = − ⎜ ⎟
⎝ kT ⎠ ⎝ 2 EK ⎠
⎛ 1 ⎞ 1
EK ⎜ 2 2 ⎟ =
⎝ k T ⎠ 4EK
(kT )
2
E2
K =
4
1
EK = kT
2
b) The average kinetic energy is given by

EK n(EK )dEK
∞
EK =
∫
0
∫ n(E )dE
∞
K K
0
Substituting in Equation (2.84) as in the example, yields
− EK
∞
∫ EK3/ 2e kT
dEK
EK = 0
−E K
∞
∫
1/ 2 kT
E e K dE K
0
The numerator and denominator are both in the form
x me ax m m −1 ax
∫ e dx = a − a ∫ x e dx
m ax
x
For us m=3/2 (in the numerator), x=EK,. and a=-1/kT.
∞
− EK
3/ 2 k T − EK
E e 3/ 2 ∞
− ∫
K 1/ 2 kT
EK e dEK
(−1 / kT ) (−1/ kT ) 0
EK = 0
−E K
∞
∫0
EK1/2 e kT
dEK

Solutions Chapter 2
⎡ 3/ 2 −∞ −0
⎤ −E K
⎢∞ e kT
0 e ⎥
3/ 2 k T
3/ 2 ∞
⎢ −1 / kT − −1/ kT ⎥ − −1 / kT ∫0 EK e dEK
1/ 2 kT
⎢⎣ ( ) ( )⎥ (
⎦
)
EK = − EK
∞
∫ K E 1/ 2 kT
e
0
dE K
−E K
[ ]
3 ∞
0 − 0 + kT ∫ EK1/2 e kT
dEK
EK = 2 0
−E
∞ K
∫ 0
E 1/K 2e kT
dEK
3
= kT
2
2.41. At what doping concentration does apparent band-gap narrowing become

significant in n-type silicon? (We will define significant as a narrowing of
0.03eV.)
From Figure 2.25, we see that ∆Eg* = 0.03 at a doping level of about 3.3 x
1018 cm-3.
2.42. Suppose you were to dope some silicon with ND=NA=1016 cm-3.
a) Where do you expect the Fermi level to be?
Since the number of donor and acceptors is the same, we expect the
Fermi level to be at the intrinsic level.
b) What do you expect n0 and p0 to be?
If the Fermi level is at the intrinsic level, then it must the case that n0=p0
=ni.
c) Verify (or adjust) your expectations by performing the calculations.

What do you think is going on physically? What is happening to the electrons
in the donor and acceptor states?
Since NA is comparable to ND , we use charge neutrality (Equation (2.89)):
N A + n0 = N D + p0
10 + n0 = 10 + p0
16 16
∴n0 = p0
We also know that n0 p0 = ni2 , thus n0=ni and p0=ni, as we predicted. What
is happening physically is that the electrons from the donor states are

Solutions Chapter 2
primarily being used to occupy the acceptor states. Thus both donors and
acceptors are ionized but the concentration of free carriers is unchanged.
2.43. Plot and compare ni for Si as a function of temperature from 250 K to 500
K as calculated from Equation (2.88) with that calculated from Equations (2.85),
(2.86), (2.63), and (2.64). Neglect the small temperature dependences of the
density of states effective masses.
Two ways: combining
3 3
19 ⎛ m dse ⎞ ⎛ T ⎞ 2
Eg *
− 2
n i = N CN V e , N C = 2.54 × 10 ⎜
kT
⎟ ⎝ ,
⎝ m0 ⎠ 300 ⎠
3 3
⎛ m* ⎞ 2 ⎛ T ⎞ 2 4.73 × 10−4 T 2
N V = 2.54 × 1019 ⎜ dsh ⎟ , and Eg = 1.17 −
⎝ m0 ⎠ ⎝ 300 ⎠ T + 636eV
and using Equation (2.88) directly:

2.365
19 ⎛ T ⎞
6733
−
n i = 5.71 × 10 e T
⎝ 300 ⎠
The results are virtually indistinguishable in the plot.

Solutions Chapter 2
15 14 -3
2.44. A silicon sample doped with NA=10 and ND=10 cm is at a temperature
of 600 K. Find n0 and p0.
At this elevated temperature, it is not the case that ND-NA>>ni; in fact from
Figure 2.19, ni≈4×1015cm-3. Thus we must use space charge neutrality. At
such high temperatures, however, all dopants are ionized, and from
Equation (2.89),
p 0 − n0 + N D − N A = 0 . We solve this simultaneously with n 0 p0 = n 2i to
obtain

Solutions Chapter 2
ND − N A 1 ⎡
1
p0 = + ( N D − N A ) + ni2 ⎤ 2
2
2 2 ⎣ ⎦
1
1014 − 1015 1 ⎡ 14
( ) + ( 4 × 10 ) ⎤2
2 2
= + 10 − 1015 15
2 ⎢
2⎣ ⎥⎦
−3
= 1.6 × 10 cm
15
and
ni2 ( 4 × 10 )
2
15
n0 = = = 1.6 × 1015 cm −3
p0 1.6 × 1015
Thus at this temperature the material is nearly intrinsic.
2.45. In this chapter we differentiated between conductivity effective mass and

density-of-states effective mass. This problem illustrates the calculation of the
conductivity of a semiconductor by using the conductivity effective mass. (A
more thorough discussion appears in Chapter 3). The current density is the
current passing through a plane of unit area per second. Consider an n-type
semiconductor. The current density is
J = − qn0 v = −σE
where n0 is the electron concentration, v is the average electron velocity, and
σ is the conductivity. the negative sign indicates that the electron is charged
negatively. Since this is an n-type semiconductor, n0=ND and the conductivity
can be expressed as
qN v
σ =− D
E
17 -3
Find the conductivity of a semiconductor with ND=10 cm if the average time
between collisions of electrons is 10-13 seconds.
To find the average velocity, we recognize that the acceleration is related

to the force by
qE
F = qE = m*a and thus a = * . If each electron is accelerated at this
m
constant rate over an average time interval t = 10−13 s , then the average
terminal velocity is
10−13 qE
vterm = ∫ dt and the average velocity is half that:
0 m*
v =
1 qE
2 m*
( )
10−13
Plugging this result into our expression for the conductivity yields

Solutions Chapter 2
−13
⎛ qE ⎞ 10
qN D ⎜ * ⎟
qN D v ⎝ m ⎠ 2 q 2 N D 10−13
σ= = =
E E 2 m*
2⎛
−3 100 cm ⎞
( )⎜ ⎟ (10 s )
3 3
−19 −13
1.6 × 10 C 1017
cm ⋅ 3
= ⎝ m ⎠
*
2m
−28
1.3 × 10
=
m*
For GaAs the result is

1.3 × 10−28 1.3 × 10−28 −1 ⎛ 1m ⎞
( ) ( )
−1
σ= = = 2.13 × 10 3
Ω − m ⎜ ⎟ = 21.3 Ω − cm
m *
(
0.067 9.1 × 10−31 kg ) ⎝ 100cm ⎠

Solutions Chapter 2
Solutions Chapter 3
3.1. Calculate the speed of an electron in Si with kinetic energy 0.013 eV. Draw
an equilibrium energy band diagram for silicon and indicate where this
electron will be. Compare your calculated thermal speed to the typical drift
velocities cited in the text of 104 cm/s. How does it compare to typical
saturation velocities?
The electron speed is
2 EK 2 ( 0.013eV ⋅ 1.6 × 10−19 J / eV )

v= = =1.3 × 105 m / s
mce* 0.26 ( 9.1 × 10-31 kg )
100cm
= 1.3 × 105 m / s × = 1.3 × 107 cm / s
m
Even with this very small kinetic energy, the electron’s instantaneous
speed is still about 1000 times faster than the drift velocity. It is
comparable to the drift saturation velocity for silicon of 1×107cm/s.
EK=0.013eV
EC
Eg=1.12eV
EV
3.2. Calculate the resistivity for a uniformly doped silicon sample with 1017
donors per cubic centimeter.

Solutions Chapter 3
From Equation (3.14), σ = qµn n + qµ p p
The material is n-type, and not degenerately doped, so n0=1017 and

ni2 (1.08 × 10 cm )
10 −3 2
p0 = = = 1.2 × 103 cm −3
n0 1017 cm −3
The mobilities are µn =6.3×102 for majority carriers (electrons) at ND=1017

and µp =4.6×102 for minority holes.
σ = qµn n + qµ p p
= 1.6 × 10−19 ( 630 ) (1017 ) + 1.6 × 10−19 ( 460 ) (1.2 × 103 )
= 10 ( Ω − cm )
−1
1
The resistivity ρ is ρ = = 0.1 Ω-cm .
σ
Note that the contribution of holes to the conduction is negligible.
3.3. A lightly doped Si sample (ND = 1014 cm-3) is heated from 300 K to 400 K. Is
its resistivity expected to increase or decrease? Explain your answer. Repeat for
Si with ND = 1018 cm-3.
For the lightly doped sample, as the temperature is increased the intrinsic
concentration increases. It is still true that N D ni , though, and to
reasonable approximation n0 is constant. For this doping level, however,
lattice (phonon) scattering predominates and mobility decreases with
increasing temperature (Figure 3.8). Thus, the conductivity decreases or
the resistivity increases
For the heavily doped sample, impurity scattering predominates. As the

temperature is raised, the carriers’ thermal energy is increased so the
Solutions Chapter 3
effect of the impurity scattering is reduced, and the mobility should
increase. Therefore the resistivity will decrease.
3.4. Germanium is an interesting semiconductor because it has a small band
gap (Eg=0.67eV). (In fact, for a while it was not considered to be a
semiconductor but was classified as a metal. Now it is a semiconductor again.)
As a result, it has a higher intrinsic concentration ni than either silicon or
GaAs. Do you expect the conductivity of intrinsic germanium to be less than
or greater than that of intrinsic silicon? How about compared to GaAs? Why?
Because ni depends exponentially on bandgap, and Eg for Ge is

appreciably less than that of Si (1.12 eV) or GaAs (1.43 eV), one expects
Ge to be more conductive than either Si or GaAs, because it has more
carriers available for conduction, and its mobilities are comparable.
3.5. Find the electron and hole concentrations and the conductivity of silicon
doped with ND=1016 cm-3 and NA=1019 cm-3 . Be sure to take into account the
minority and majority carriers.
σ = qµn n + qµ p p .
This material is p-type and degenerately doped. The electrons are minority
carriers, and their mobility is 250 cm2/V-s, from Figure 3.4 at NA=ND=1019
cm-3. The holes are majority carriers, and µp ≈60 cm2V-1s-1.
To find the carrier concentrations, we note that the material is

degenerately doped. Thus,
p0=NA-ND=1019 cm-3, and from Equation (2.112),

Solutions Chapter 3
⎛ ∆Eg* ⎞
⎜ ⎟
N ⎜ kT
⎟
n0 p0 = A e ⎠ ni2 . The apparent band gap narrowing ∆Eg* is 85 meV from
⎝
NV
Figure 2.25. Thus
⎛ ∆EG* ⎞
⎛ 0.085 ⎞
N 1 ⎜⎜⎝ ⎟
1019
e⎝ 0.026 ⎠ (1.08 × 1010 ) = 99cm −3
kT ⎟⎠ 1 ⎜ ⎟ 2
n0 = A e n = 19
2
−3
10 3.1×10 cm
i 19
p0 NV
Thus the conductivity of this sample is
σ = qµn n + qµ p p
= 1.6 ×10−19 ⎡⎣( 250cm 2 / V − s )( 99cm −3 ) + ( 60cm 2 / V − s )(1019 cm −3 ) ⎤⎦
= 96 ( Ω − cm )
−1
3.6. A voltage of 2.5V is applied to a sample of silicon whose cross-sectional area is
0.1 µm ×1 µm. The length of the path is 0.1 µm. If the material is doped n-type with
ND=1018 cm-3, what is the current in the sample? What is the current density? These
dimensions could represent the channel of a field effect transistor.
The cross sectional area is A = 0.1µm × 1.0µm = 10−9 cm2 , and the length is
L=0.1µm=10-5 cm..
We know the current is given by I=JA, and J=σE. To find σ, we need to

find n0 and p0. We can write that n0=ND=1018, and p0 will be negligible so it
won’t contribute to conductivity. From Figure 3.4 for silicon doped to
1018cm-3, we find that the majority carrier mobility µn≈230 cm2/V·s, and
σ = qµn n = (1.6 × 10 −19 C)(230)(1018 ) = 37(Ω − cm)

−1
Thus
V 2.5V
= 37 ( Ω − cm ) ⋅
−1
J = σE = σ = 9 × 106 A / cm2
L 0.1×10−4 cm

Solutions Chapter 3
and.
I = JA = 9 × 106 Amps / cm 2 ⋅10−9 cm 2 = 9mA
3.7. Calculate the electron drift velocity for ND -NA = 1016 cm-3 in a bar of Si of
cross-sectional area 1.0 mm2 for a current of 50 mA.
I 50 × 10−3 A
From Equation (3.2), I=JA, or J = = = 5 A / cm 2 .
A ⎛ 1cm ⎞
2
1mm 2 ⎜ 2 ⎟
⎝ 100mm ⎠
The carrier density is n=ND=1016 cm-3.
Adapting Equation (3.8) to n-type material,
J n ( drift ) 5 A / cm 2
v= = = 3.1× 103 cm / s
qn 1.6 × 10−19 C (1016 cm −3 )
3.8. Explain how an electron can "collide" with a crystal defect. Hint: the
normal periodic potential of the lattice is disturbed by a crystalline defect. A
poor quality crystal will have lower carrier mobilities than a good one.
The defect can create a potential barrier well or trap. This will scatter
electrons and holes the same way a charged impurity does.
3.9. Explain why for a given non-compensated semiconductor with a given
doping level N, the electron mobility is larger in p-type Si than in n-type Si.
Refer to Figure 3.4.
In n-type silicon, the electrons are majority carriers, whereas in p-type

silicon they are minority carriers. From Figure 3.4, for the same doping

Solutions Chapter 3
level, the µn is larger when the electrons are minority carriers, which is in
p-type material.
3.10. With reduced N in problem 3.9 above, the electron mobilities approach
each other. Explain.
As the doping becomes less, the electrons are less affected by the
impurity miniband, and thus see the same scattering mechanisms as the
electrons in p-type material do.
3.11. Compare the mean free time between collisions for electrons and for
holes in intrinsic GaAs. How do these values compare to those for silicon?
Following the method of Example 3.2, we have
⎛ 1m 2 ⎞
( 0 )⎜ × ⋅ ×
10 4 cm2 ⎠ 0.067 (9.1× 10 kg )(0.8m / V ⋅ s)
⎟
3 2
0.067m 8 10 cm / V s −31
⎝
2
mceµn
*
tn = = =
q 1.6 × 10−19 C 1.6 × 10 −19 C
= 0.3ps
where we obtained the value of µn from Figure 3.7 and mce* from the
table inside the book’s cover. This is comparable to the case for silicon
(0.2ps).
For holes,
⎛ 1m 2 ⎞
*
µp
( 0.34 m0 )⎜ 4 × 10 2
cm 2
/ V ⋅ s × ⎟
mch ⎝ 104 cm 2 ⎠
tp = = −19
= 7.7 × 10−14 s = 0.077 ps
q 1.6 × 10 C
compared to 0.1 ps in silicon.
3.12. For the case of Problem 3.11, find an average gain in kinetic energy
between collisions for electrons and holes for an applied field of 100 V/cm.
The electrons attain an average velocity of (Equation (3.27))

Solutions Chapter 3
⎛ 100cm ⎞
qE
(1.6 ×10 C ) (100V / cm ) ⎜
−19
⎟
⎝ 1m ⎠ 0.3 × 10−12 s = 7.9 × 103 m / s
vdn = − tn = − ( )
mce* ( 0.067 ) ( 9.1×10−31 kg )
The corresponding kinetic energy is
(0.067)(9.1× 10−31 kg)

1 * 2
(7.9 × 103 m / s)
2
EK = m cev dn =
2 2
⎛ 1J ⎞
= 1.9 × 10−24 J
⎝ 1.6 × 10 −19 eV ⎠
= 1.2µeV
For holes the result is
⎛ V ⎞ ⎛ 100cm ⎞
qE
(1.6 ×10 C ) ⎜100
−19
⎝
⎟⎜
cm ⎠ ⎝ m ⎠
⎟
vdp = tp = 7.70 × 10−14 s
*
mch ( 0.34 ) ( 9.1×10 kg )
−31
= 4.0 × 102 m / s
and the kinetic energy is
(0.34)(9.1 × 10−31 kg )
1 * 2
( 4.0 × 106 m / s )
2
EK = mch vdp =
2 2
⎛ 1J ⎞
= 2.5 × 10−26 J ⎜ −19 ⎟
⎝ 1.6 × 10 eV ⎠
= 0.15µ eV
3.13. Estimate the saturation velocity of electrons in intrinsic GaAs. How does
your estimate compare with the experimental data?
Following Example 3.3, we make the assumptions again that an electron

loses all of its kinetic energy after each collision, that an electron gains an
energy equal to Epho between collisions, and the electrons remain in a
region of constant and valid effective mass.

Solutions Chapter 3
From the table inside the book cover, Epho for GaAs is 0.034 eV, and we
have from Equation (3.35),
v E pho 0.034ev (1.6 × 10−19 J / eV ) ⎛ 100cm ⎞

vsat = max = = ⋅⎜ ⎟ = 2.1 × 10 cm / s
7
2 2mce* 2 ( 0.067 ) ( 9.1 × 10 kg ) ⎝

−31
m ⎠
From Figure 3.9, we see that the saturation velocity for GaAs peaks at
about this value, 2×107 (a very good match), but then decreases and
levels off at about 6×106. The reduction is due to scattering into the other
conduction band minima, invalidating our assumption that the electron
remains in a region in which the concept of effective mass is valid- to go
from one such region to another involves crossing through a region in
which this assumption is not true.
3.14. The electron velocity in Si has its saturation value ( v sat≈ 1 x 107 cm/s)
over the range of 5 x 104 V/cm to 2 x 105 V/cm. Plot the mobility-field (µ - E) and
t −E relations over this range of fields.
Since the velocity is saturated, both µ and t must decrease. From

v
Equation (3.11), we have. µ = Thus the plot looks like, over the range
E
indicated:
Mobility µ (cm2/V·s )

Solutions Chapter 3
qE
Similarly, from Equation (3.27), v = tn which results in
mce*
3.15. Consider a p-type Si sample of NA = 1018cm-3 and ND = 0. Over a length of
1 µm the electron concentration drops linearly from 1016cm-3 to 1013cm-3.
Calculate the electron diffusion current density.
From Equation (3.40), we have
dn(x)
Jn(diff ) = qDn . Since the electron concentration is dropping linearly,
dx
the expression for n(x) is
103 cm−3
n(x) = − −4 x , and the derivative is
10 cm
dn −7 −4
= −10 cm .
dx
At NA =1018cm-3, the diffusion constant is, from Figure 3.11 (electrons are
minority carriers since the material is p-type), Dn=9cm2 /s. Thus

Solutions Chapter 3
⎛ cm 2 ⎞
= (1.6 × 10−19 C ) ⎜ 9 ⎟ ( −10 cm ) = −0.014nA / cm
dn( x) −4
J n ( diff ) = qDn 7 2
dx ⎝ s ⎠
The minus sign indicates that the current is flowing from a region of higher
concentration to a region of lower concentration.
3.16. A material is doped such that the electron concentration varies linearly
across the sample. The sample is 0.5 µm thick. The donor concentration varies
from ND=0 at x=0 to ND=1016cm-3 at x=0.5 µm.
a) Write equations for n(x) and p(x).
At x=0, n=ni. We will simplify by assuming it to be zero, and take the

carrier concentration to be linear from zero to 1016.
∆N D 1 × 1016 cm −3
n0 = 0 + x= −4
x = 2.0 × 1020 cm −4 ( x )
∆x 0.5 × 10 cm
ni2 ni2
p ( x) = =
n ( x ) 2.0 × 1020 ( x)
b) Find the electron diffusion current density.
= qDn ( 2.0 × 1020 cm −4 )

dn( x)
J n ( diff ) = qDn
dx
For this doping range, the mobility µn is fairly constant, and thus so is Dn.
From Figure 3.11 Dn = 30 cm2/s.
Thus the electron diffusion current is

Solutions Chapter 3
⎛ cm 2 ⎞
J n ( diff ) = qDn ( 2.0 × 1020 cm −4 ) = (1.6 × 10−19 C ) ⎜ 30 ⎟ ( 2.0 × 10 cm ) = 960 A / cm
20 −4 2
⎝ V ⋅ s ⎠
c) Find the hole diffusion current density at x-0 and x =0.5 µm. Can the
minority carriers contribute significant diffusion current?
The hole diffusion coefficient is, from Figure 3.11 for minority carriers,
about Dp =12 cm2/V-s .
From Equation (3.41) the hole diffusion current is given by

dp ( x ) ⎛ ni2 ⎞ ⎡ d ⎛ 1 ⎞⎤
J p ( diff ) = −qD p = − qD p ⎜ ⎟
20 ⎢ ⎜ ⎟⎥
dx ⎝ 2.0 × 10 ⎠ ⎣ dx ⎝ x ⎠ ⎦
⎛ ni2 ⎞
20 ⎟ (
= −qD p ⎜ − x −2 )
⎝ 2.0 × 10 ⎠
This equation is only valid once ND becomes noticeably greater than ni,
however, which becomes clear if we try to plug in x=0 and the equation
blows up. At x=0, in reality the hole concentration is also ni and not infinite.
It does, however, decline rapidly (as 1/x2) as the electron concentration
increases. Clearly for small x the minority carriers do indeed contribute a
large amount of diffusion current.
At x=0.5µm, the hole diffusion current is

⎛ ni2 ⎞
20 ⎟ (
J p ( diff ) = −qD p ⎜ − x −2 )
⎝ 2.0 × 10 ⎠
⎡ (1.08 × 1010 cm −3 )2 ⎤
= (1.6 × 10−19 C )(12cm 2 / Vs ) ⎢ ⎥ 1
⎢ 2.0 × 10 cm
20 =3
⎥ ( 0.5 × 10−4 cm )2
⎣ ⎦
=10
= 4.5 × 10 A / cm 2
which is negligible.
d) Find an expression for Ec(x)-Ef as a function of x.
⎛ n ( x) ⎞ ⎛ 2.0 × 1020 x ⎞
EC ( x) − E f = − kT ln ⎜ ⎟ = − 0.026 eV ln ⎜ 19 ⎟
⎝ NC ⎠ ⎝ 2.86 × 10 ⎠

Solutions Chapter 3
We recognize that at x=0, the carrier concentration is ni and EC-Ef =0.56,
and at x=0.5µm, EC-Ef=0.2 eV. The Fermi level varies logarithmically from
EC-Ef=0.56 to 0.2 eV.
e) At equilibrium, the total current must be zero. Show that there must
therefore be an internal electric field present in this sample. (The field is
generated by the variation in doping as we will see in the next chapter.)
We have
J = J n + J p = J n ( drift ) + J p ( drift ) + J n ( diff ) + J p ( diff )

= qE ( µn n + µ p p ) + J n ( diff ) + J p ( diff )
Also J=0 at equilibrium. The diffusion currents are not zero; therefore the
drift current is also not zero. The drift current is proportional to the electric
field E, and since q, n, p, and the µ’s cannot be zero, the electric field
cannot be zero.
3.17. Comment on the probability of absorption (zero, low, medium, high) by a
photon of λ=600 nm (red) by the following materials: Si, Ge, GaAs, InAs, SiC,
GaN, CdS.
Solution: a photon of 600 nm has a photon energy of 1.24/0.6µm=2.07eV,

so if the semiconductor has a band gap larger than 2.07 eV, this photon
cannot be absorbed at all. For those whose band gap is correct, we look
for indirect vs. direct gap materials. Indirect materials have a low
probability of absorption; direct materials have a high probability.
Si: low: Si has an indirect gap, and Eg=1.12 eV<2.07 eV.
Ge: High: although it is an indirect semiconductor, there is a direct
transition at K=0 for photon energies greater than 0.8 eV.
GaAs: high: direct and band gap is small enough
InAs: high, direct and band gap is small enough

Solutions Chapter 3
SiC: zero: band gap is too large
GaN: zero: direct but band gap is too high
CdS: zero: direct but band gap too high.
3.18. Explain physically why the carrier lifetimes are much less in direct gap
materials than in indirect.
It is more likely for carriers to recombine optically in direct gap materials

than in indirect materials, where they must recombine by phonon
processes, either by a combination of phonons or a combination of
phonons and photons. Since the indirect processes require more particles
and are therefore less probable, the lifetime of a carrier awaiting one of
these unusual events is longer.
3.19. Light of hν=1.5eV (this in the near infrared, at λ=826 nm) at a power level
of 10 mW shines on an intrinsic sample of GaAs of area 1 cm2. Let the electron
lifetime be 10 ps.
a) What is the number of photons arriving at the semiconductor surface
per sec? (recall energy=power×time).
One photon has energy of 1.5 eV, so to produce a power of 10 mW, we

require
N (photons) × 1.5eV / photon × (1.6 × 10 −19 J/ eV )

10mW =
1s
N = 4.2 × 10 photons / s
16
b) Verify that photons at this energy can be absorbed.

Solutions Chapter 3
For the photons to be absorbed, the energy of a photon must be greater
than the bandgap. The energy gap of GaAs is 1.43 eV, so these photons
can be absorbed.
c) Assuming every photon is absorbed and creates an electron-hole pair, and
assuming the GaAs sample is 1 mm thick, what is the optical generation rate?
There are N=4.2×1016 photons/sec arriving on an area 1 cm2, and

absorbed in a volume 1 cm2x0.1cm=0.1 cm3. Since each photon produces
and electron and a hole, the average optical generation rate is
N 4.2 × 1016 photons / s / cm 2
Gop = = = 4.2 × 1017 e-h pairs/sec/cm3
thickness 0.1cm
d) What are the equilibrium electron and holes densities (in the dark)?
Since the GaAs is intrinsic, n0 =p0=ni=2.2×106 cm-3.
e) What are the excess carrier concentrations when the light is on?
dn
Since this is steady state, =0. From Equation (3.63) we have
dt
∂n 1 ⎛ ∂J n ⎞ ⎛ ∆n ⎞
= ⎜ ⎟ + ⎜ Gop − ⎟=0
∂t q ⎝ ∂x ⎠ ⎝ τn ⎠
If the photons are uniformly absorbed, there is no diffusion current, and
∆n = Gopτ n = 4.2 × 1017 cm =3 s −1 (10 × 10−12 s ) = 4.2 × 106 cm −3
f) What are the recombination rates for electrons and holes when the light is
off? When the light is on?

n ∆n
R= 0 +
τn τn
When the light is off, n0=ni (since this is intrinsic GaAs), or n0=2.2×106 cm-3
and ∆n=0.
n 2.2 × 106 cm −3
Roff = 0 = = 2.2 × 1017 cm −3 s −1
τn −12
10 × 10 s
When the light is on,

Solutions Chapter 3
n0 ∆n 4.2 × 106 cm −3
Ron = + = 2.2 × 1017 + = 2.2 × 1017 + 4.2 × 1017 = 6.4 × 1017 cm −3 s −1
τn τn 10 × 10 −12
g. What are the steady-state carrier densities n and p?
For electrons we have

n=n0+∆n=2.2×106 cm-3 + 4.2×106 cm-3=6.4×106 cm-3
Since the material is intrinsic, n0=p0=ni, and since the excess electrons
and holes are created in pairs, ∆n=∆p, so
p=p0+∆p=n0+∆n=n=6.4×106 cm-3 .
h. How much does the conductivity of this sample change compared with its
dark value?
We know that
σ = q ( µn n + µ p p ) = q ( µn n0 + µ p p0 ) + q ( µn ∆n + µ p ∆p )
The mobilities for intrinsic GaAs are, from Figure 3.7, µn=8000 cm2/V-s
and µp=400 cm2/V-s. In the dark, the conductivity is
σ = q ( µn n0 + µ p p0 ) = 1.6 × 10−19 C ⎡⎣( 8000cm 2 / V ⋅ s )( 2.2 × 106 cm −3 ) + ( 400cm 2 / V ⋅ s )( 2.2 × 106 cm −3 )

= 3.0 × 10−9 ( Ω − cm )
−1
When the light is on, the result is
σ = q ( µn n + µ p p )
= 1.6 × 10−19 C ⎡⎣( 8000cm 2 / V ⋅ s )( 6.4 × 106 cm −3 ) + ( 400cm 2 / V ⋅ s )( 6.4 × 106 cm −3 ) ⎤⎦
= 8.6 × 10−9 ( Ω − cm )
−1
showing that the conductivity nearly triples when the light is on in this
case.

Solutions Chapter 3
i) Suppose the power level is kept the same, but the wavelength of the light is
shifted further into the infrared, at E=hν=1 eV (λ=1240 nm). What is the
generation rate now?
It is zero. These photons have energy smaller than the band gap and
cannot be absorbed.
3.20. Solve Equation (3.72) to prove that Equation (3.73) is a solution under the
appropriate initial conditions. Repeat for Equation (3.75).
Equation (3.72) is
d∆n ∆n
= Gop −
dt τn
Recognizing that
dx 1
∫ = ln (a + bx ),
a + bx b
Equation (3.72) can be put into the form
d∆n ⎛ ∆n ⎞
∫1− ∆n = Gop τ n ln ⎜
⎝
1−
Gopτ n ⎟⎠ ∫ op
= G dt
Gopτ n
Rise Time:
Using the initial condition at t=0, ∆n=0 and solving for ∆n, gives after a bit
of algebra,
⎛ − ⎞
t
∆n(t ) = Gopτ n ⎜ 1 − e ⎟
τn
⎝ ⎠
which is Equation (3.73).
Fall time:
Since Gop =0 and at t0, ∆n = ∆n(t0 ) , Equation (3.72) becomes

Solutions Chapter 3
−
(t −t 0 )
∆n = ∆n(t0 )e τn
in agreement with Equation (3.75).
3.21. A direct-gap semiconductor sample is illuminated at one end with light of
λ=500 nm (green), with an intensity of 1 mW/cm2. The area of the illuminated
surface is 1 cm2. Assume the carrier lifetimes are 10 ns.
a) Find the number of photons striking the sample per second.
The energy of a photon is E(eV)=1.24/λ(µm)=1.24/0.5=2.48eV. At 1

mW=1mJ/s = 10-3 J/s , the number of photons striking the surface is
J
10−3 ×1cm 2
N= s ⋅ cm 2
= 2.5 × 1015 photons / s
( 2.48eV / photon ) (1.6 ×10 J / eV )
−19
b) If every photon is absorbed uniformly (with x) within 1 µm of the
surface, what are the excess carrier concentrations ∆n and ∆p in this region?
Every absorbed photon produces an electron-hole pair, so in a volume of

1cm2x1mm= 10-1cm3, there are 2.5 x 1015 electron-hole pairs created per
second .
d∆n ∆n d∆n
From Eq. (3.72), = Gop − , and in the steady-state, =0,
dt τn dt
∆n = Gopτ n = 2.5 × 1015 cm −3 s−1 × 10 −7 s = 2.5 × 108 cm −3
and ∆p = ∆n .

Solutions Chapter 3
3.22. As we saw, it is possible to produce a non-uniform spatial distribution of
charge carriers by shining a light on one part of a semiconductor, producing
more electrons and holes in that part than elsewhere. What might be another
way to produce more electrons and holes in one location, without applying an
electric field? Would there be a net diffusion current?
One could heat one end of the bar, increasing ni locally. Thus electrons
would be diffusing from the warm end (higher concentration) to the cool
end, producing a net electron diffusion current. But, holes will be diffusing
in the same direction as the electrons, so the hole diffusion current is
going in the opposite direction. The net diffusion current is zero.
3.23. Consider a bar of semiconductor illuminated as shown in Figure P3.1.
a) Sketch the concentrations of electrons and holes as functions of
position.

Solutions Chapter 3
b) In which direction(s) will the electrons diffuse? Holes?
All carriers diffuse to regions of lower concentration, so they diffuse away

from the center of the bar.
c) In what directions do the electron and hole diffusion currents go?
The electron current flows toward the center of the bar, and the hole
diffusion current flows away from the center.
c) Explain why a plot of the voltage versus position along the length of the
sample is as shown in the figure.

Solutions Chapter 3
In the figure we observe less voltage drop across the illuminated region.
That is because there is less resistance in this region due to the increased
conductivity arising from the increased carrier concentrations produced by
the light.
3.24. a) Find the conductivity of a sample of GaAs doped with NA=1017 cm-3.
We need the carrier concentrations. Assuming room temperature, all

dopants are ionized, and p0=1017 cm-3. Since the material is not
degenerately doped,
ni2 (2.2 × 10 cm )
6 −3 2
−5 −3
n0 = = 17 −3 = 4.8 × 10 cm →n0=4.8×10-5 cm-3
p0 10 cm
The conductivity is given by σ 0 = q(µn n0 + µ p p0 ). We find the µ’s from

Figure 3.7 for GaAs with a total doping concentration of 1017cm-3. The
results are
µn≈5300 cm2/V·s and µp≈230 cm2/V·s. Plugging into the expression for σ0,
we have
[ ]
σ 0 = 1.6 × 10 −19 C (5300cm2 / V ⋅ s)(4.8 × 10−5 cm−3 )+ (230cm 2 / V ⋅ s)(1017 cm −3 )
= 3.68(Ω − cm )−1
Note that the electron contribution here is negligible.

Solutions Chapter 3
b) If light the sample is illuminated such that the excess electron concentration
is 1016 cm-3, what is the excess hole concentration?
When electrons are generated optically, for each electron that is created
there is a hole created too. Therefore the excess hole concentration is
∆p=1016 cm-3 also.
c) What is the conductivity of this sample when the light is on?
Now we have excess carriers, and the conductivity is
σ = q ( µn n + µ p p ) = q ( µn (n0 + ∆n) + µ p ( p0 + ∆p ) )
= 1.6 ×10−19 C ⎡⎣5300cm 2 / V ⋅ s ( ≈ 0 + 1016 cm −3 ) + 230cm 2 / V ⋅ s (1017 + 1016 cm −3 ) ⎤⎦
= 1.6 ×10−19 C ⎡⎣5.3 × 1019 + 230 ⋅ (1.1× 1017 ) ⎤⎦ = 13 ( Ω − cm )

−1
3.25. Consider an n-type Si sample with ND - NA = 1017cm-3. Under
illumination, ∆n=∆p = 5 x 1016cm-3. Find the quasi Fermi levels for electrons
and holes. Compare these locations to the Fermi level when the light is off.
⎛N ⎞
From Equation (3.86), E fn = EC − kT ln⎛⎝ C ⎞⎠ and E fp = EV + kT ln⎜ V ⎟ . We
N
n ⎝ p ⎠
know that n=n0+∆n. The material is n-type, so n0 =ND-NA =10 . Since this
17
is not doped enough to produce band gap narrowing,
ni2 (1.08 × 10 )
10 2
−3
p0 = = = 1.2 × 10 cm
3
17
n0 10
Thus n=1017+5×1016=1.5×1017cm-3 and p0=1.2×103+5×1016=5×1016cm-3.
The quasi-Fermi level for electrons is
⎛ 2.86 × 1019 cm−3 ⎞

EC − E fn = kT ln⎛⎝ C ⎞⎠ = 0.026 ln⎜
N
⎟ = 0.136eV
n ⎝ 1.5 × 1017 cm −3 ⎠

Solutions Chapter 3
With no light,
⎛N ⎞ ⎛ 2.86 × 1019 cm−3 ⎞

EC − E f = kT ln ⎜ C ⎟ = 0.026ln ⎜ ⎟ = 0.147eV
⎝ n0 ⎠ ⎝ 1 × 1017 cm −3 ⎠
The light moved the quasi-Fermi level for electrons (majority carriers) by
0.01eV.
For holes (minority carriers),
⎛N ⎞ ⎛ 3.1 × 1019 ⎞
E fp − EV = kT ln ⎜ V ⎟ = 0.026 eV ln ⎜ 16 ⎟
= 0.167eV
⎝ p ⎠ ⎝ 5x10 ⎠
With no light, the Fermi level was 0.147 eV away from the conduction
band edge, and with light the quasi Fermi level for holes is close to the
valence band edge- it has moved by 0.8eV.
3.26. A sample of InP is doped such that Ef –EV=0.2 eV. It is also illuminated
such that ∆n=∆p=103cm-3. Find the quasi-Fermi levels and sketch the energy
band diagram. Repeat for ∆n=∆p=1010.
We find the equilibrium carrier concentrations:

⎛ E f −E V ⎞
−⎜ ⎟ −
0.2
⎝ kt ⎠ −3 −3
p0 = N V e = 6.8 × 10 cm e = 3.1× 10 cm
18 0.026 15
ni2 (1.97 × 10 )
7 2
−3
n0 = = = 0.12cm
3.1 × 10
15
p0
The total carrier concentrations are:
n = n0 + ∆n = 0.12 + 103 ≈ 103 cm −3 ; p = p0 + ∆p = 3.1× 1015 + 103 ≈ 3.1× 1015 cm −3
From which we can find the quasi -Fermi levels:
⎛ n ⎞ ⎛ 103 ⎞
EC − E fn = − kT ln ⎜ ⎟ = −0.026eV ⋅ ln⎜ 18 ⎟ = 0.91eV , and Efp should
⎝ NC ⎠ ⎝ 2.07 × 10 ⎠
be the same as Ef since the majority carrier concentration is almost
unaffected.
When the excess carrier concentrations are 1010, n=n0+∆n≈1010cm-3, and

Solutions Chapter 3
⎛ n ⎞ ⎛ 1010 ⎞
EC − E fn = − kT ln ⎜ ⎟ = −0.026eV ⋅ ln⎜ ⎟ = 0.50eV . The excess
⎝ NC ⎠ ⎝ 2.07 × 1018 ⎠
hole concentration is still negligible compared to the equilibrium
concentration of holes and so Efp is still equal to Ef.
The energy band diagram for both cases is shown below:
0.91 eV 0.5 eV
EC
1.43 eV Efn (δn=1010 )
Ei Efn (δn=103 )
E f =Efp
EV
0.2 eV

Solutions Chapter 3
Chapter 4 Solutions
4.1. Consider a p-type Si sample of length 0.1 µm in which the net doping varies
exponentially from 5×1017 to 5×1015. Find the electric field.
x − x0 ⎛ N ' (x )⎞
Following the example of Section 4.2.2, = ln⎜ A' 0 ⎟
λ ⎝ N A (x ) ⎠
and λ =
( x − x0 ) = ( 0.1×10−4 cm ) = 2.2 ×10−6 cm
⎛ N' (x ) ⎞ ln (100 )
ln ⎜ A' 0 ⎟
⎝ NA ( x) ⎠
kT 1 ⎛ 1 ⎞
E =− = −0.026V ⎜ ⎟ = −12kV / cm
q λ −6
⎝ 2.2 × 10 cm ⎠
The field points in the negative x-direction.
4.2. Consider the graded-composition Si:Ge alloy discussed in Section 4.3. The
hole concentration is assumed constant ( N 'A is essentially constant and all
dopants are ionized). There is no electric field or diffusion for holes at
equilibrium but there is a field for electrons in the conduction band. Since at
equilibrium there is no net current, the drift electron current must be offset by
an opposing diffusion current. Identify and explain the source of the varying
electron concentration that produces the diffusion current.
Since the bandgap is changing, ni changes. Thus even though p0 is
ni2
constant, n 0 = is not. At the end where the band gap is smaller, ni is
p0
Solutions Chapter 4 1 2/17/07

Anderson & Anderson
larger, and thus n0 is larger. This produces an electron diffusion flux to the
left that offsets the drift flux gong to the right.
4.3. Find the time required for an electron to traverse the p-region of Example
4.2 due to drift alone.
The electric field is 20 kV/cm. at this field the electrons travel near their
saturation velocity. From Figure 3.9, v = 8.5x106 cm / s ≅ vsat .
∆x 0.14 ×10−4 cm
Then t = = = 1.6 ps .
v 8.5x106 cm / s
4.4. A graded alloy is manufactured in the AlGaAs system. At x=0, the material
is pure GaAs (χ=4.07, Eg=1.43), and over a distance of 2 µm the composition
changes to Ga0.6Al0.4As (χ=3.4 eV and Eg=1.92 eV). The material is intrinsic.
Find the effective electric fields for electrons and holes, and find the true
electric field.
We first draw the energy band diagram by drawing the Fermi level as a
constant, and then constructing the energy band diagram of the two
materials around it assuming the intrinsic level to be at midgap. At the
GaAs end, EC ≅ 1.43/2=0.715 eV above Ef, and Evac is 4.07 eV above EC.
At the AlGaAs end, the band edges are each ≈ 1.92/2=0.96 eV from the
Fermi level and Evac is 3.4 eV above Ec.

Anderson & Anderson
E vac
4.07 eV 3.4 eV
EC
0.715 eV 0.96 eV
Ef
EV
0 .0 0 .5 1 .0 1 .5 2 .0
distance (um)
The effective electric field for electrons is, from Equation (4.27),
1 dEC 1 ( 0.96 − 0.715) eV ⋅ (1.6 ×10 −19

J / eV )
E *
= = −19
⋅
q dx 1.6 ×10 C 2 ×10−4 cm
e
= 1.2kV / cm
For holes, we have
1 dEV 1 ( −0.96 − −0.715) eV ⋅ (1.6 ×10 −19

J / eV )
E *
= = −19
⋅
q dx 1.6 × 10 C 2 × 10−4 cm
p
= −1.2kV / cm
And the true field is
⎡( 3.4 + 0.96 ) − ( 4.07 + 0.715 ) ⎤⎦ eV ⋅ (1.6 × 10 −19

J / eV )
⋅⎣
1 dEvac 1
E = = −19
q dx 1.6 × 10 C 2 ×10−4 cm
= −2.1kV / cm

Anderson & Anderson
4.5. Consider the equilibrium energy band diagram of Figure P4.1:
EC
0.2 eV
Ef
Ei
EV
0.2 eV
1 µm
a) Find the effective electric field for electrons
The electric field is given by
1 dEC 1 ⎛ 0.2eV ⋅1.6 × 10−19 J / eV ⎞

E e* = = ⎜ ⎟ = 2kV / cm
q dx 1.6 ×10−19 C ⎝ 10−4 cm ⎠
b) Find the effective electric field for holes
1 dEV 1 ⎛ 0.2eV ⋅1.6 ×10−19 J / eV ⎞

E h* = = ⎜ ⎟ = 2kV / cm
q dx 1.6 × 10−19 C ⎝ 10−4 cm ⎠
c) Sketch the carrier concentrations.
(below)
d) Indicate the directions of each of the following:
i. The hole diffusion

ii. The hole diffusion current
iii. The hole drift
iv. The hole drift current
v. The electron diffusion
vi. The electron diffusion current
vii. The electron drift
viii. The electron drift current
ix. The electric field for electrons

Anderson & Anderson
Fn diffusion
Jndiffusion
Fn drift
J n drift
EC
Ef
Ei
ξ
EV
Fp diffusion
Jp diffusion
Fp drift
Jp drift
4.6. For the sample whose energy band diagram is shown in Figure P4.2:
EC
0.2 eV
Ef
Ei
EV
0.1 eV
1 µm
1 dEC 1 −0.2eV (1.6 ×10−19 J / eV )

E e* = = = −2kV / cm
q dx 1.6 ×10−19 C 10−4 cm
The field is negative so it points in the negative x direction.
b) Find the effective electric field for holes
1 dEV 1 +0.1eV (1.6 × 10−19 J / eV )

E *
= = = +1kV / cm
q dx 1.6 × 10−19 C 10−4 cm
p

Anderson & Anderson
The field for the holes is positive, pointing in the positive x direction.
c) Sketch the carrier concentrations
(below)
d) Indicate the directions of each of the following:
i. Hole diffusion
ii. Hole diffusion current
iii. Hole drift
iv. Hole drift current
v. Electron diffusion
vi. Electron diffusion current
vii. Electron drift
viii. Electron current
ix. Electric field for electrons
x. Electric field for holes
It may be perplexing at first to know how to draw the hole concentrations-
on the right hand end the material is more heavily n-type, so the hole
concentration should be small, but on the other hand the band gap is
small. The relative size of the “sail” can be estimated by the distance to

Anderson & Anderson
the Fermi level, which is closer to the valence band on the right hand side
than on the left hand side.
4.7. Draw the energy band diagram for GaAs that is doped such that EC-Ef=0.2+
(0.8 eV/µm)x for 0<x<0.5 µm.
The electric field is
1 dEC 1 0.4eV (1.6 × 10−19 J / eV )

E e* = = = 8kV / cm
q dx 1.6 ×10−19 C 0.5 × 10−4 cm
b) Indicate the directions of Jn(diff), Jn(drfit), Jp(diff), and Jp(drift).
(below)
c) Indicate the direction of the electric field.

Anderson & Anderson
4.8. A sample of AlyGa1-yAs has a composition that is graded linearly from pure
GaAs to Al0.4Ga0.6As over a length of 10 µm. The sample is not doped. Find the
value of the built-in electric field for electrons and holes. The band gap of
AlyGa1-yAs is given by Eg = 1.43 + 1.425 y (in eV) for y<0.43. Note the “y” here is
composition fraction, not distance. Assume that the electron affinity is
constant and Ei is at midgap.
The band gap expands by 1.43(0.4) eV=0.57 eV in a distance of 10 µm.
The conduction band edge moves upward by half this amount (since the
material is intrinsic) and the valence band moves down by half this
amount. The electric field for electrons is thus 0.285 V/10 µm=285 V/cm
(the conduction band edge has a positive slope) and the field for holes is –
285 V/cm (the valence band edge has a negative slope).
0.285eV E
C
1.43 eV
1.43+0.57=2.0eV
E
f
EV
0.285eV
10 µm

Anderson & Anderson
4.9. An optoelectronic device will be made of layers of GaxIn1-xAsyP1-y grown on
an InP substrate. To make sure the layers will have the same lattice constant as
the substrate, the gallium fraction x is kept equal to 0.47y. The arsenic fraction
is varied among the layers from y=0 to y=0.5. The band gap varies as
Eg = 1.35 − 0.72y + 0.12y 2 eV. Assuming for convenience that the effective
masses are constant at mdse and mdsh , find the ratio of ni(y=0) to ni(y=0.5) from
* *
the pure InP layers to the y=0.5 layer.
The layers vary from pure InP to Ga0.235In0.765As0.5P0.5. The band gap of
pure InP is 1.35 eV, and the band gap of the quaternary layer is
Eg = 1.35 − 0.72(0.5) + 0.12(0.52 ) = 1.02eV .
Since the effective masses are assumed to be equal, then NC and NV are
equal. The ratio of the nI’s is

E g (InP )
2kT E g (InP ) −E g (GaAsInP) (1.35 −1.02)
ni (InP) NC NV e
= Eg (GaAsInP) =e 2kT
=e 2 ×0.026
= 570
ni (GaAsInP) 2kT
NC NV e
So there are 570 times as many intrinsic carriers in the narrower band gap
alloy.
4.10. Consider the Si:Ge alloy discussed in Example 4.2. If the acceptor
concentration varies exponentially from 1018 cm-3 to 2×1016 cm-3, show that the
total effective electric field for electrons is E e* =E∗doping+ E ∗composition.
The electric field due to the grading of the composition was shown to be
20kV/cm in the example. Therefore it remains to find the electric field due

Anderson & Anderson
to the variation in doping. We will assume for simplicity the NC and NV are
the same for pure silicon and the Si:Ge alloy. The effective density of
states for electrons NC actually does not vary much from Si to Ge, but NV
varies by a factor of 10 from one to the other. Since the alloy in question,
however, is only 13% Ge, we can roughly take NV to be constant. We then
need to find Ef –EV for the sample. At the pure silicon end,
⎛ N (x = 0)⎞ ⎛ 1 × 1018 ⎞
E f − EV (x = 0) = − kT ln ⎜ A ⎟ = −0.026ln ⎜ ⎟ = 0.089eV . This is
⎝ NV ⎠ ⎝ 3.1 × 1019 ⎠
greater than 2.3kT so we can use this result. For the alloy end,
⎛ N (x = 0.05)⎞ ⎛ 2 × 1016 ⎞
E f − EV (x = 0.05 µm) = −kT ln⎜ A ⎟ = −0.026eV ln⎜⎝ ⎟ = 0.19eV
⎝ NV ⎠ 3.1 × 1019 ⎠
The slope in the valence band edge is therefore
∆EV 0.089 − 0.19
= = −2.02eV / µ m or 0.1eV in 0.05 µ m . This results in an
∆x 0.05µ m
effective electric field for holes of -20kV/cm. The resulting energy band
diagram, then, reflects the negative slope in the conduction band edge
due to graded composition and the negative valence band slope due to
the doping. Since, however, the band gap must be maintained at
∆Eg = 0.1eV / 5µm , as the valence band slopes the conduction band must
also slope to compensate.
We draw the energy band diagram by drawing Ef as constant first. Then
we add the valence band edge, with its slope of –2.02 eV/µm. Finally, we
add the conduction band edge such that the band gap decreases by 0.1
eV/µm. The total slope of the conduction band edge is thus equal to the
slope of EV and the change in Eg:
Si Si Ge0.13
0.87
∆EC =-0.2 eV E
E =1.12 eV C
g E =-1.02eV
g
Ef
E
∆EV =-0.1 eV V

Anderson & Anderson
4.11. We will discuss pn junctions in Chapter 5. However, you already have
enough knowledge to deduce some things from the energy band diagram,
shown in Figure P4.3.
a) In what region is there an electric field?
electron drift
electron diffusion
region of electric field
EC
qV bi
Ei
Ef
0.2 eV 0.3 eV
EV
Eg =1eV
hole diffusion
hole drift
There is a field wherever there is a non-zero slope in EC or EV.
b) What is the value of the built-in voltage?
The built-in voltage is found from Eg-0.2-0.3=0.5 eV, so Vbi=0.5 V
c) Sketch the electron and hole concentrations. Are the directions of the drift

Anderson & Anderson
and diffusion components of the electron and hole currents correct for
equilibrium?
The electron hole and electron drift and diffusion compensate for each
other as expected at equilibrium.

Anderson & Anderson
SOLUTIONS CHAPTER 5
5.1. A silicon pn junction is formed between n-type silicon doped with
ND=1017cm-3 and p-type silicon doped with NA=1016 cm-3.
a) Sketch the energy band diagram. Label both axes and all
important energy levels.
Neither side is degenerately doped. We find
⎛ 1017 ⎞ ⎛ 1016 ⎞
δn = −kT ln ⎜ ⎟ = 0.147eV and δ = −kT ln ⎜ ⎟ = 0.209eV .
⎝ 2.86 × 1019 ⎠ p
⎝ 3.1 × 1019 ⎠
The resulting energy band diagram is

qVbi=0.764 eV
EC
electron energy
Ei
EC
Ef Ef
δn=0.147 eV EV
Ei
δp=0.209 eV
EV
distance (x)
b) Find nn0, pn0, np0, and pp0. Sketch the carrier concentrations.
nn0 = N D = 1017 cm −3
ni2 (1.08 × 10 cm )
10 −3 2
−3
= = = 1.17 × 10 cm
3
pn0 17 −3
nn0 10 cm
−3
p p0 = NA = 10 cm
16
(1.08 × 1010 cm −3 )
2
ni2
n p0 = = = 1.17 × 10 4 cm−3
p p0 1016 cm −3
See above for carrier concentration sketches.

c) What is the built-in voltage?
qVbi = E g − δ n − δ p = 1.12 − 0.147 − 0.209 = 0.764eV
or Vbi =0.764V
Anderson& Anderson 1 February 17, 2007

Solutions Chapter 5
5.2 Recall that the circuit symbol for a diode is as shown in Figure II.3. Which
is the anode (the end labeled “+”) of a pn junction diode, the p side or the n
side? Explain your reasoning.
In a diode, current can flow from the anode to the cathode but not from the
cathode to the anode. In a pn junction, current can flow from p to n but
only a small leakage current can flow from n to p. Thus, the p-side is the
anode and the n-side is the cathode.
5.3. A p+n junction is formed in silicon. On the p side, the Fermi level is at the
(intrinsic) valence band edge Ef=EV0. The n side is doped with ND=5×1016 cm-3.
a) Sketch the energy band diagram.
First, we have to find δn
⎛ ND ⎞ ⎛ 5 × 1016 cm−3 ⎞
δn = −kT ln ⎜ ⎟ = −0.026eV ln ⎜⎝ ⎟ = 0.165eV
⎝ NC ⎠ 2.86 × 1019 cm−3 ⎠
The resulting energy band diagram is
qVbi=0.955 eV
EC
electron energy
Ei
EC
Ef
Ei δn=0.165 eV E V= E f
δp=0 eV
EV
distance (x)
b) Sketch the carrier concentrations.

see above
c) What is the built-in voltage?
Vbi = Eg − δ n − δ p = 1.12 − 0.165 − 0 = 0.955eV
5.4. Fill in the missing steps to derive Equation 5.35 for the junction width on
the n side of the junction.

Solutions Chapter 5
2εV j 2εV jn
We want to show that wn = using wn = and
⎛ N' ⎞ qN D'
qN D ' ⎜1 + D' ⎟
⎝ NA ⎠
V jn N A'
= . From this last equation, and Vj=Vjn +Vjp we can write
V jp ND'
⎛ Vp⎞ ⎛ N' ⎞
V j = V jn + V jp = V jn ⎜1 + jn ⎟ = V jn ⎜1 + D' ⎟
⎜ V ⎟
⎝ j ⎠ ⎝ NA ⎠
Vj 2εV jn
Rearranging, V = n
and substituting this into wn = to get
⎛ N D' ⎞
j
qN D'
⎜ 1 + ⎟
⎝ N A' ⎠
2ε V j
wn =
⎛ N' ⎞
qN D' ⎜ 1 + D' ⎟
⎝ NA ⎠
5.5. A step pn junction diode is made in silicon with the n-side having
N D' =2×1016 cm-3 and on the p-side the net doping is N A' =5×1015 cm-3.
a) Draw, to scale, the energy band diagram of the junction at
equilibrium.
We begin by finding the locations of the Fermi levels on each side of the
junction. On the n side,
⎛n ⎞ ⎛ N' ⎞ ⎛ 2 × 1016 ⎞
EC − E f = − kT ln⎜ 0 ⎟ = − kT ln⎜ D ⎟ = −0.026eV ln⎜ = 0.19eV
⎝ NC ⎠ ⎝ NC ⎠ ⎝ 2.89 × 1019 ⎠
This assumes complete ionization and nondegeneracy. On the p side, we

have
⎛ p0 ⎞ ⎛ N A' ⎞ ⎛ 5 × 1015 ⎞
E f − EV = − kT ln ⎜ ⎟ = − kT ln ⎜ ⎟ = −0.026eV ln ⎜ 19 ⎟
= 0.25eV
⎝ NV ⎠ ⎝ NV ⎠ ⎝ 3.1 × 10 ⎠
From these we can construct the entire energy band diagram. The built-in
voltage is then 1.12 − 0.19 − 0.25eV =0.68eV

Solutions Chapter 5
0.70eV
0.19eV
0.25eV
0.31eV
1.12eV 0.37eV
b) Find the built-in voltage, and compare to the value measured off your
drawing in part (a).
⎛ ⎞
kT ⎛ N D' N A' ⎞ ⎜ (2 ×1016 cm −3 ) ⋅ (5 ×1015 cm −3 ⎟
Vbi = ln ⎜ ⎟ = 0.026V ln ⎜ = 0.71V
q ⎝ ni2 ⎠
⎝ (1.08 × 10 10
cm )
−3 2 ⎟
⎠
which agrees reasonably well with the result above. The difference results
from round-off errors.
c) Find the junction width.

We use
2ε Vbi ( N A' + N D' ) 2(11.8)(8.85 ×10−14 F / cm)(0.71V )(2.5 ×1016 cm −3 )

w= =
qN A' N D' (1.6 ×10−19 C )(2 × 1016 cm −3 )(5 ×1015 cm −3 )
= 48 × 10−6 cm = 0.48µ m
d) Find the width of the n-side of the depletion region and the p-side of
the depletion region, and the voltage dropped across each side of the transition
region.
2εV j 2(11.8)(8.85 × 10−14 F / cm)(0.71V )
wn = =
⎛ N' ⎞ ⎛ 2 × 1016 cm −3 ⎞
qN D' ⎜1 + D' ⎟ 1.6 × 10−19 C (2 × 1016 cm −3 ) ⎜ 1 + −3 ⎟
⎝ 5 × 10 cm ⎠
15
⎝ NA ⎠
= 9.6 × 10−6 cm = 0.096 µ m

Solutions Chapter 5
2εV j 2(11.8)(8.85 × 10−14 F / cm)(0.71V )
wp = =
⎛ N' ⎞ ⎛ 5 × 1015 cm −3 ⎞
qN A' ⎜1 + A' ⎟ 1.6 × 10−19 C (5 × 1015 cm −3 ) ⎜1 + −3 ⎟
⎝ 2 × 10 cm ⎠
16
⎝ NC ⎠
= 38 × 10−6 cm = 0.38µ m
For the voltages, we have
qN D' 1.6 × 10−19 C ⋅ 2 × 1016 cm −3
V jn = ( wn ) 2 = (9.6 × 10−6 cm) 2
2ε 2(11.8)(8.85 × 10−14 F / cm)
= 0.14V
qN A' 1.6 × 10−19 C ⋅ 5 × 1015 cm −3
V jp = ( wp )2 = (38 × 10−6 cm) 2
2ε −14
2(11.8)(8.85 × 10 F / cm)
= 0.55V
e) Plot the electric field. What is its maximum value?
N'
We know that E ( x) = q D ( x − xn ) for xn < x < x0 , and that
ε
qN A'
E ( x) =
ε
(x p − x ) for x0 ≤ x ≤ x p . This results in a field distribution as
shown below:
3
30x10
ξmax=29000V/cm
Electric Field (V/cm)
25
20
15
10
0
0.0 0.1 0.2 0.3 0.4
Distance fromn x(µm)
Wn =0.098µm W =0.38µm
p

Solutions Chapter 5
The maximum field occurs at x0, which is 0.098 µm from xn. The value of
the field is 2.9×104 V/cm.
f) Plot the voltage distribution.

The voltage is found by integrating the field, resulting in
qN'D 1.6 × 10 −19 C ⋅ 2 × 1016 cm −3
(x − x n) = −
2
V(x) − V (x n ) = − (x − x n ) for
2
−14
2ε 2(11.8)(8.85 × 10 F / cm)
xn ≤ x ≤ x0 . We arbitrarily set xn to x=0. As we integrate in x, we start at
V(xn)=0. Thus
qND'
V(x) = − (x ) + [V (x n = 0) = 0 ]
2
2ε
When x=0.098µm, we switch to
qN 'A
( ) qN 'A
(0.48µm − x )2 − 0.71V . The plot
2
V(x) = + x − x + V(x ) = +
2ε p p
2ε
becomes
Voltage (volts)
0.2
0.0
-0.14 -0.2
-0.4
-0.6
-0.71
-0.8
-1.0
-0.2 0 0.2 0.4 0.6 0.8
xn xo xp
Distance (µm)
g) Plot the potential energy for electrons (EC).

The potential energy EP=-qV, so we invert the plot from part (f). The units
on the vertical axis become eV instead of V:

Solutions Chapter 5
1.0
0.8
0.6
0.4
0.2
0.0
-0.2
-0.2 0 0.2 0.4 0.6 0.8
xn xo xp
Distance (µm)
Note that this is identical to the shape of the conduction band edge since
for electrons, EC=EP.
h) Draw the energy band diagram for Va=0.5 V.
The new junction voltage is Vj = Vbi − Va = 0.71− 0.5 = 0.21V
2ε (Vbi − Va )(N 'A + N 'D ) 2(11.8)(8.85× 10 −14 F / cm)(0.71 − 0.5V)(2.5 ×1015 cm −3 )

w= =
qN 'A N'D (1.6 × 10 −19 C)(2 ×1016 cm −3 )(5 ×1015 cm −3 )
= 26 × 10−6 cm = 0.26µm
qVj =0.71-0.5=0.21eV W=26µm
i) Draw the energy band diagram for Va=-5V.

The new junction voltage is Vj = Vbi − Va = 0.71− (−5) = 5.71V

Solutions Chapter 5
The new junction width is
2ε (Vbi − Va )( N A' + N D' ) 2(11.8)(8.85 × 10−14 F / cm)(0.71 − (−5)V )(2.5 × 1015 cm −3 )

w= =
qN A' N D' (1.6 × 10−19 C )(2 × 1016 cm −3 )(5 × 1015 cm −3 )
= 136 × 10−6 cm = 1.36µ m
W=136µm
qV =5.71eV
j
In the diagrams below, the vertical axis is to scale but the junction widths
are not.

Solutions Chapter 5
V =V =0.71V
j bi
Equilibrium
V j =V -V =0.71-0.5V
bi a
Forward bias of 0.5 volts
Vj =Vbi -V =0.71-(-5)
a
Reverse Bias
5.6. Consider the equilibrium energy band diagram for a pn junction diode
shown in Figure P5.1 (a).
a) Indicate the region(s) where there exists an electric field.

Solutions Chapter 5
b) What is the value of the built-in voltage?
It is the band gap less the displacement of the Fermi level from the band
edges on each side: Vbi=Eg /q –0.2-0.3=1-0.5=0.5 V
c) Sketch the electron and hole concentrations. Indicate the directions of the
drift and diffusion components of the electron and hole fluxes and currents .
see above
d) The same device is shown in Figure P5.1(b), but now a voltage is applied
across it. If the total junction voltage Vj=Vbi-Va, what is the value of the applied
voltage?
The junction voltage is 5 V, and the built-in voltage is 0.5 volts, so the
applied voltage is –4.5 V.
5.7. A pn junction is formed in silicon between n-type ( N D' =1018 cm-3) and p-
type ( N A' =1017 cm-3). Find, for equilibrium,
a) wn , wp and w
The junction voltage at equilibrium is
Eg kT ⎛ N C NV ⎞ ⎛ ( 2.86 × 1019 )( 3.1 × 1019 ) ⎞

Vbi = − ln ⎜ ' ' ⎟ = 1.12V − 0.026V ln ⎜ ⎟
q q ⎝ ND N A ⎠ ⎜ 10181017 ⎟
⎝ ⎠
= 1.12 − 0.236V = 0.88V
Solutions Chapter 5
From this we find
2εV j ( N A' + N D' ) 2 (11.8 ) ( 8.85 × 10−14 F / cm ) ( 0.88 ) (1.1 × 1018 cm −3 )

w= =
qN D ' N A ' 1.6 × 10−19 C (1018 )(1017 )
= 0.112 × 10−4 cm = 0.112 µ m
2εVj 2(11.8)(8.85 × 10−14 F / cm)(0.88V )

wn = =
⎛ N D' ⎞ 18 ⎛ 1018 ⎞
And qND ' ⎜ 1 + ⎟ 1.6 × 10 C(10 )⎜ 1 + 17 ⎟
−19
⎝ NA ' ⎠ ⎝ 10 ⎠
= 1.02 × 10 −6 cm = 0.0102 µm
wp = w − wn = 0.112µ m − 0.0102µ m = 0.102 µ m
b) The built-in voltage

Found above, Vbi=0.88V
c) The maximum electric field

2 ⋅ 0.88V
E max = 2V j / w = = 16kV / cm
0.112 ×10−4 cm
d) How much of Vbi is dropped on the n side? On the p-side?

qN D' 2 (1.6 × 10−19 C )(1018 )
(1.02 × 10−6 cm ) = 0.080V
2
Vj =
n
wn =
2ε 2 (11.8 ) ( 8.85 × 10 F / cm )
−14
and on the p side

Vjp = V j − V jn = 0.88 − 0.08 = 0.80V
e) Sketch the energy band diagram carefully reflecting your calculations

above.

Solutions Chapter 5
5.8. An n+p junction has N A' =1017 cm-3. The Fermi level on the n+ side is taken
to be at the (intrinsic) conduction band edge. Find the junction width at
equilibrium.
To find w we need the built-in voltage:
Eg δp −kT ⎛ 1017 ⎞
Vbi = − = 1.12V − ln ⎜ ⎟
q q q ⎝ 3.1× 1019 ⎠
= 1.12V − 0.15V = 0.97V
2εV j
Then w ≈ wp = . We don’t know N D' , but it is much larger
⎛ N' ⎞
qN A' ⎜1 + A' ⎟
⎝ ND ⎠
than N A' , so this reduces to
2εV j 2 (11.8 ) ( 8.85 × 10−14 F / cm ) 0.97V

w = = 0.11µ m
qN A' 1.6 × 10−19 C (1017 )
5.9. A silicon diode has N D' =1017 on the n side and N A' -=1016 on the p side. It is
forward biased at Va=0.5 V.

Solutions Chapter 5
a) What is the diffusion current density due to minority carriers at the
plane x=xp?
Dn ⎛ qVkTa ⎞
We use Jn (x p ) = q np0 ⎜ e − 1⎟ . We therefore need to find Dn, Ln, and
Ln ⎝ ⎠
np0, At xp, electrons are minority carriers. Thus from Figures 3.11 and
3.23, Dn≈30cm2/s and Ln≈400 µm.
From this we can find
(1.08 × 1010 cm −3 )
2
ni2
We find npo using n p0 = = 16 −3 = 1.17 × 10 4 cm−3 .
p p0 10 cm
Thus
cm 2
⎛ ⎞ 30
s (1.17 × 10 4 cm−3 )⎡ e 0.026V − 1⎤
qVa 0.5V
Dn
Jn (x p ) = q np0 ⎜ e kT − 1⎟ = 1.6 × 10 −19 C ⋅ ⎢ ⎥
Ln ⎝ ⎠ 0.04cm ⎣ ⎦
= 1.4 × 10 −12 [2.2 × 108 − 1]= 3.1 × 10 −4 A / cm 2
b) What is the minority carrier diffusion density at the plane x=xn?
We find µp from Figure 3.22 for a doping of 1017, and find µp=220 cm2/Vs.
For holes in n-type silicon doped at 1017 cm-3, Dp≈12 cm2/s and Lp≈70
µm=0.007cm, from Figures 3.11 and 3.23 respectively. The minority
carrier density is
ni2 (1.08 × 10 cm )
10 −3 2
−3
= = = 1.17 × 10 cm
3
pn0 17
nn0 10
and
cm2
⎛ qVa ⎞ 12
s (1.17 × 10 3 cm −3 )⎡ e 0.026V − 1⎤
0.5V
Dp
J p (xn ) = q pn0 ⎜ e kT − 1⎟ = 1.6 × 10 −19 C ⋅ ⎢ ⎥
Lp ⎝ ⎠ 0.007cm ⎣ ⎦
= 7.2 × 10 −5 A / cm 2
c) What is the total current density in the junction neglecting

recombination and generation?
J diff = J n (x p ) + J p (x n ) = 3.1 × 10−4 + 7.2 × 10 −5 ≈ 3.8 × 10 −4 A / cm2
Solutions Chapter 5
d) What is the maximum recombination current density in the forward-
biased junction? Compare this result to the injection (diffusion current).
The recombination rate is:

⎡ 2(0.026V
0.5V
⎤
qV a
(1.08 × 10 cm )⎢e
10 −3 )
⎥
2kT
ni e ⎢⎣ ⎥⎦ 17 cm
−3
Rmax = = −4 = 8.1 × 10
2τ 0 2(10 s) s
where τ0 is taken as the average lifetime in the more lightly doped side
since the depletion region is predominantly on that side.
An equation for the associated current is not explicitly given in the text, but
we approximate it by adapting Equation 5.78 for the generation current.
We will have
J R = + qwRmax ,
where the plus sign reflects that this current flows from p to n. For
simplicity we have assumed that recombination has its maximum rate in
the entire depletion region. We will need to calculate the junction width w
and to do that we first need Vbi:
⎡ ⎤
kT ⎛ N D' N A' ⎞ ⎢ 10161017 ⎥ = 0.774V
Vbi = ln ⎜ ⎟ = 0.026V ln
q ⎝ ni2 ⎠ ⎢ (1.08 × 1010 cm −3 )2 ⎥
⎣ ⎦
2ε ( N D' + N A' ) (Vbi − Va )

w=
qN D' N A'
( 2 )(11.8) ⎛⎜ 8.85 × 10−14

F ⎞ 17
⎟ (10 + 10 cm ) ( 0.774 − 0.5 )V
16 −3
= ⎝ cm ⎠ = 0.20µ m
(1.6 × 10 C )(1016 cm−3 )(1017 cm−3 )
−19
The resulting current is

⎛ 10−4 cm ⎞
J R = +qWRmax = 1.6 × 10 −19 C⎜ 0.20µm ⋅
⎝ µm ⎠
⎟ (
8.1 × 1017 cm −3 / s) = 2.6 × 10 −6 A / cm 2
To compare this to the diffusion current, we take the ratio and find

Solutions Chapter 5
J diff 3.8 × 10 −4 A / cm 2
= = 146
JR 2.6 × 10 −6 A / cm 2
Thus under forward bias we expect the diffusion current to dominate over
the recombination current.
e) Which is larger, the injection current density into the lightly doped
side or the injection current into the heavily doped side?
From parts (a) and (b), we see that the injection from the heavily doped
side into the lightly doped side is greater.
f) Repeat part (a) for a reverse bias of Va=-5 V.

cm 2
⎛ ⎞ 30
s (1.17 × 10 5 cm −3 )⎡e 0.026V − 1⎤
qVa −5V
D
Jn (x p ) = q n np0 ⎜ e kT − 1⎟ = 1.6 × 10 −19 C ⋅ ⎢ ⎥
Ln ⎝ ⎠ 0.04cm ⎣ ⎦
= 1.4 × 10 −12 [3.03 × 10 −84 − 1]= −1.4 × 10−12 A / cm2
g) Repeat part (b) for Va=-5V.

cm2
⎛ ⎞ 12
−3 ⎡ 0.026V ⎤
qVa −5V
Dp
J p (xn ) = q pn0 ⎜ e
⎝
kT
− 1⎟
⎠
−19
= 1.6 × 10 C ⋅ s (
1.17 × 10 cm )⎢ e
3
− 1⎥
Lp 0.007cm ⎣ ⎦
= −3.2 × 10−13 A / cm 2
h) Estimate the generation current density under reverse bias.
We find the junction width at this bias:
2ε (N D' + N A' )(Vbi − Va )
w=
qNd ' NA '
(2)(11.8)⎛⎝8.85 × 10−14
F ⎞ 17
cm ⎠ (10 + 1016 cm −3 )(0.774 − (−5)V
= = 0.91µm
(1.6 × 10 −19
C)(1016 cm −3 )(1017 cm −3 )
JG =
− qni w
=−
(1.6 × 10 −19 C )(1.08 × 10−10 cm −3 )(0.91 × 10−4 cm) −10
= −7.4 × 10 A / cm
2
2(10 s)
−4
2τ 0
i) Compare the generation current density to the diffusion current

density in the reverse-biased junction.
JG JG 7.4 × 10 −10 A / cm2
= = 2 ≈ 430
J diff J n + J p (1.4 × 10 + 3.2 × 10 )A / cm
−12 −13

Solutions Chapter 5
5.10. a) Calculate the minority excess carrier concentrations at each edge of the
transition region for a silicon diode with N D' =5×1017 cm-3 and N 'A =1017 cm-3. The
diode is forward-biased with Va=0.5V.
−3
nn0 = N D = 5 × 10 cm
' 17
n 2 (1.08 × 10 cm )
10 −3 2
−3
= i = = 2.33× 10 cm
2
pn0 −3
5 × 10 cm
17
nn0
p p0 = 1017 cm −3
(1.08 × 1010 )
2
ni2
n p0 = = 17 = 1.17 × 103 cm −3
p p0 10
⎛ qV ⎞ ⎛ 0.5 ⎞
∆np (x p ) = n p0 ⎜ e kT − 1⎟ = 1.17 × 103 cm−3 ⎜ e 0.026 − 1⎟ = (1.17 × 10 3 )(2.2 × 108 )= 2.6 × 1011 cm −3
a
⎝ ⎠ ⎝ ⎠
⎛ qV a
⎞ ⎛ 0.5 ⎞
∆pn (x n ) = pno ⎜ e kT − 1⎟ = (2.33 × 102 )⎜ e 0.026 − 1⎟ = (2.33 × 102 )(2.2 × 10 8 ) = 5.1× 1010 cm −3
⎝ ⎠ ⎝ ⎠
b) Sketch the diffusion current as a function of distance on the p-side.

We expect the diffusion current to decay exponentially.
Jndiff
xp
x
c) Since the total current is constant, and since the difference between
the total current and the minority carrier diffusion current is due to drift of
majority carriers, sketch the majority carrier drift current as a function of
distance on the p-side.

Solutions Chapter 5
Jtotal
Jp(drift)
Jndiff
xp
x
5.11. a) Show that Equation (5.93) follows from Equation (5.46).
Equation 5.46:
⎛ EC − E f ⎞
−⎜⎜ ⎟⎟
⎝ kT ⎠
n0 = N C e
−
[EC (x )− E f ]
nn0 (x) = NC e kT
−
[EC (x )− E f −EC n + EC n ]
Thus = NC e kT
−
[(E cn −E f )] [− ECn +E C (x ) ]
−
= NC e kT
e kT
−
[(E cn −E f )]
We recognize nn0 = NC e kT
and thus
−
[EC (x )− EC n ]
nn0 (x) = nn0e kT
which is Equation 5.93.
b) Derive Equation 5.94.
We multiply together Equations 5.93

Solutions Chapter 5
−[EC ( x)− E Cn ] [
− E Vp − EV (x ) ]
n(x)p(x) = nn0 e kT p p0 e kT
[
− EC (x )− EV ( x)− E Cn + EVp ]
=N N e
' ' kT
D A
−[EC (x )− EV ( x) ] − − E Cn + E Vp [ ]
=N N e
' ' kT kT
D A e
[ ] [E Cn − E Vp ]
− Eg
= N N e kT e kT
' '
D A
[(ECn − Eg = EVn )− EVp ] [−(E Vp − EVn )]

= N D N Ae
' ' kT
= N D N Ae
' ' kT
[
− EVp − E Vn ]
=N N e'
D
'
A
kT
[ ]
− qV j
=N N e'
D
'
A
kT
From Equation 5.13, however, we have

kT ⎛ N 'D N'A ⎞
Vbi = ln ⎜ ⎟
q ⎝ ni2 ⎠
qVbi
ND' N'A = ni2 e kT
Substituting into our previous result,
−
[qVj ]
n(x)p(x) = N N e '
D
'
A
kT
qV bi
−
[qVj ]
=ne i
2 kT
e kT
qV bi
−
[q (Vbi −Va )]
=ne i
2 kT
e kT
qV a
=ne i
2 kT
5.12. Consider an n+p junction under reverse bias of 5 V. Let N A' =5×1017 cm-3
and the junction area be 75 µm2.
a) Find the reverse current due to diffusion.
From Equation 5.77, we have
⎛ qVa ⎞ ⎛ 0.026
−5
⎞
J = J 0 ⎜ e − 1⎟ = J 0 ⎜ e
kT
− 1⎟ ≈ − J 0
⎝ ⎠ ⎝ ⎠

Solutions Chapter 5
⎛ Dn n p0 Dp pn0 ⎞
J 0 = q⎜ +
⎝ Ln Lp ⎟⎠
For an n+p junction, pn0 will be negligible, and this reduces to

⎛ Dn n p0 ⎞
J 0 = q⎜ ⎟
⎝ Ln ⎠
For a doping on the p- side of 5×1017,

np0=ni2/pp0=(1.08×1010)2/5×1017=2.33×102cm-2.
From Figure 3.11, Dn≈10 cm2·s, and from Figure 3.23, Ln≈45 µm=0.0045
cm, remembering to use the values for minority carriers.
Thus
⎛ Dn n p0 ⎞
J 0 = q⎜ −19 (
⎢ )(
⎡ 10cm2 ⋅ s 2.33 × 10 2 cm−3 )⎤⎥ = 8.3 × 10
−14
⎟ = 1.6 × 10 C ⎢ ⎥
A / cm2
⎝ Ln ⎠ ⎣
0.0045cm
⎦
The current is Idiff=J0A=(-8.3×10-14A/cm2)(75×10-8 cm2)=-6.20×10-20A.
b) Find the reverse current due to generation.
1
JG = −
( '
)
ni ⎡⎢ qε N D + N A (Vbi − Va )⎤⎥
' 2
,or for a n+p junction, in which N D' N A' ,

τ0 ⎢ ' '
2N D N A ⎥
⎣ ⎦
1
ni ⎡ qε (Vbi − Va )⎤ 2
JG = − ⎢ ⎥
τ 0 ⎢⎣ 2N A' ⎥⎦
From Figure 5.10, Vbi=0.98V for the lightly doped side at 1017. Since most
of the junction is on the lightly doped (p) side, then for τ0 we should use
the electron (minority carrier ) lifetime for τ0, which from Figure 3.21 is
about 23 µs.
Thus

Solutions Chapter 5
1
n ⎡ qε (Vbi − Va )⎤ 2
JG = − i ⎢ ⎥
τ 0 ⎢⎣ 2N A' ⎥⎦
=−
(1.08 × 1010 cm −3 ) (1.6 × 10 C)(11.8)(8.85 × 10
−19 −14
)
F / cm (0.98 − (−5))
23µs 2(5 × 10 cm )
17 −3
= 4.7 × 10 −10 A / cm 2
and IG=J0A=(-4.7×10-10A/cm2)(75×10-8 cm2)=3.5×10-16A.
5.13. Note that diffusion coefficient, diffusion length and lifetime depend on
doping. (Equations 3.18, 3.42, 3.82 and 3.76.) Consider a one-sided n+p junction
in silicon.
a) Plot the reverse diffusion current density as a function of doping
for Va =-5 V.
Equation 3.18: µn (N A )= 232 +
1180
0.9
⎛ NA ⎞
1+ ⎜ ⎟
⎝ 8 × 1016 ⎠
Dn kT
Equation 3.42: =
µn q
Equation 3.82: Ln = Dnτ n
[ ]
−1
−12 −32
Equation 3.76: τ n = 3.45 × 10 N A + 9.5 × 10 N A
2
Reverse diffusion current: J=-J0, and from (Equation 5.78):

⎛ D n n p 0 D p pn 0 ⎞
J 0 = q⎜ + ⎟ , which for a one-sided n+p junction reduces to
⎝ Ln Lp ⎠
⎛ Dn n p 0 ⎞
J 0 = q⎜ ⎟ since pn0 is negligibly small.
⎝ Ln ⎠
b) Plot the reverse generation current density as a function of

doping for Va = -5 V.
1
JG = −
( '
)
ni ⎡⎢ qε N D + N A (Vbi − Va )⎤⎥
' 2
,or for a n+p junction, in which N D' N A' ,

τ0 ⎢ 2N D' N A' ⎥
⎣ ⎦

Solutions Chapter 5
1
n ⎡ qε (Vbi − Va )⎤ 2
JG = − i ⎢ ⎥ . We also require Vbi as a function of doping, which
τ 0 ⎢⎣ 2N A' ⎥⎦
is given by Equation 5.13:

kT ⎛ N C N ’A ⎞
Vbi = ln ⎜ ⎟
q ⎝ ni2 ⎠
Using τn for τ0 since most of the junction is on the lightly doped side, and it
is minority carriers that are of interest. The magnitudes of both currents
are plotted below
-7
10
Current density (A/cm2)
-8
10
-9 generation
10
-10
10
-11
10
diffusion
-12
10
-13
10
14 15 16 17 18
10 10 10 10 10
Doping concentration NA (cm-3)
5.14. A Si junction has N D' =1015 cm-3 and N A =1017 cm-3. The junction area is 100
'
µm2.
a) Find Vbi.
⎛ 15 −3 ⎞
kT ⎛ N 'A N D' ⎞ ⎜ 10 cm ⋅1017 cm−3 ⎟
Vbi = = 0.026V ⋅ ln = 0.71V
q ⎜⎝ n2i ⎟⎠
ln
⎜ 1.08 × 1010 cm −3 2 ⎟
⎝ ⎠ ( )
b) Find I0.
⎛ qVa ⎞
Neglecting generation and recombination, we have I = I0 ⎜ e kT − 1⎟
⎝ ⎠

Solutions Chapter 5
⎛D n D p ⎞
where I0 = qA ⎜ n p0 + p n0 ⎟ .
⎝ Ln Lp ⎠
(1.08 × 1010 cm −3 )
2
ni2
n p0 = = 17 −3 = 1.17 × 103 cm −3
p p0 10 cm
ni2 (1.08 × 10 cm )
10 −3 2
−3
p n0 = = = 1.17 × 10 cm
5
15 −3
nn 0 10 cm
From Figure 3.11 for minority carriers,
Dn=20 cm2/s
Dp=13 cm2/s
and from Figure 3.23,

Ln=0.012 cm
Lp=0.1 cm
Thus
⎛ 20⋅ 1.17 × 103 13 ⋅1.17 × 105 ⎞
I0 = (1.6 × 10−19 C )(100 × 10 −8 cm 2 )⎜ + ⎟
⎝ 0.012 0.1 ⎠
= 3.1 × 10−19 A
c) Find the current at Va=-5 V, 0V, and +0.5V. Remember to consider

both diffusion and recombination-generation current. Also bear in mind that
most of the junction appears on the lightly doped side (thus lifetime should be
chosen for that material).
At Va =-5V, the diffusion current is I=-I0=-3.1×10-19A, since

⎛ qV a
⎞
I = I0 ⎜ e kT − 1⎟ and the term in the brackets is essential equal to –1.
⎝ ⎠
The generation current density is

qni w
JG =
2τ 0
Since the junction is mostly on the lightly doped side, to choose a value
for τ0 we pick an average of the electron and hole lifetimes, at a doping of

Solutions Chapter 5
1015 (the lightly doped side). We will choose, from Figure 3.21, a value of
about 10-3 s.
The junction width is
w=
(
2ε (Vbi − Va ) N 'A + N D' )
' '
qN N
A D
2(11.8)(8.85 × 10−14 F / cm)(0.71 + 5)(1.01 ×1017 )

= = 2.7µm
(1.6 × 10 C)(10
−19 17
)(1015 )
and JG =
(1.6 × 10 −19
C)(1.08 × 1010 cm −3 )(2.7 × 10 −4 cm )
= 2.3 × 10
−10
A / cm
2
−3
2(10 s)
The generation current is IG=JG·A=2.3×10-16 A.
The total reverse current at –5 volts is
I0+IG=3.1×10−19+2.3×10-16= 2.3×10-16A.
At zero volts, the device is in equilibrium and the current is zero.
At Va=+0.5 Volts, RG current can be neglected and diffusion current is the

main contributor:
⎛ 0.5 ⎞
[ ]
I = I 0 ⎜ e0.026 − 1⎟ = (3.1× 10−19 A ) 2.2 × 10 8 = 7.0 × 10 −11 A = 70pA
⎝ ⎠
5.15. Plot the small signal resistance Rp as a function of applied voltage for Va
positive. Let the reverse leakage current be I0=10-15 A. Use a logarithmic scale
for the resistance axis. Comment on your graph in view of your expectations of
the resistance of a diode.
The differential resistance
−1 −1
⎡ dI ⎤ 1 ⎧⎪ qI ⎛ qVa ⎞ ⎫⎪
Rp = ⎢ ⎥ = = ⎨ 0 ⎜ e kT − 1⎟ ⎬
⎣ dVa ⎦ ⎧⎪ ⎡ qV ⎤ ⎫⎪ ⎪ kT ⎝ ⎠ ⎪⎭
− 1⎥ ⎬ ⎩
d a
⎨ o⎢
I e kT
dVa ⎪⎩ ⎣ ⎦ ⎪⎭
Plotting this with I0=10-15 A and kT/q=0.026 V, we obtain

Solutions Chapter 5
We see from the graph that the resistance is essentially infinite until the
diode starts to turn on, after which the resistance decreases exponentially.
5.16 Consider a symmetrical junction in GaAs in which N D' = N 'A =1016 cm-3.
a) Find Vbi.
Vbi =
kT ⎛ N 'A N D' ⎞
ln ⎜
(
= (0.026V )⋅ ln ⎢
⎢ )
⎡ 1016 cm −3 2 ⎤
⎥ = 1.16V
⎟ 2⎥
q ⎝ ni ⎠ ( )
2
⎢⎣ 2.2 × 10 ⎥⎦
6
b) Calculate the series resistance Rs of the bulk regions if the cross

sectional area of the junction is 50 µm2 and the lengths of the bulk regions are
each 2 µm. Note that the series resistance Rs is due to the finite resistivity of
the semiconductor materials; it is not the same as the differential resistance Rp.
ρL
From Chapter 3, we have R = , so we need to find the resistivity ρ. The
A
current is carried by the majority carriers in the bulk, and their numbers
are the equilibrium concentrations.
On the n-side, we have µn=7000 cm2/v·s from Figure 3.7 at a doping of
1016.

Solutions Chapter 5
−1
⎧⎪ −19 ⎡⎛ ⎤ ⎫⎪
{
= q(µn nn + µ p pn ) } cm2 ⎞ 16 −3
⎟ (10 cm )+ ( negligible)⎥ ⎬
1 −1
ρn = = ⎨1.6 × 10 C ⎢⎜ 7000
σn ⎪⎩ ⎣⎝ V ⋅ s⎠ ⎦ ⎪⎭
= 0.089(Ω − cm )
From this the series resistance is
ρn L (0.089 Ω ⋅cm)(2 × 10−4 cm)
Rs(n = = = 35.6 Ω
side)
A 50 × 10 −8 cm 2
On the p-side, we find µp=310 cm2/V·s.
−1
⎪⎧ ⎡ cm 2 ⎞ 16 −3 ⎤ ⎪⎫
{
= q(µn nn + µp pn ) } ⎛
⎟ (10 cm )⎥ ⎬
1 −1
ρp = = ⎨1.6 × 10 −19 C ⎢( negligible) + ⎜ 310
σp ⎩⎪ ⎣ ⎝ V ⋅ s⎠ ⎦ ⎭⎪
= 2.0(Ω − cm)
ρ pL (2Ω ⋅ cm)(2 × 10 −4 cm )
Rs(p = = = 800Ω
side )
A 50 × 10−8 cm2
5.17. Consider a forward biased pn junction:

⎛ q (Va − IR S ) ⎞
I = I0 ⎜⎜ e kT − 1⎟⎟ or
⎝ ⎠
kT ⎛ I ⎞
Va = ln ⎜ + 1⎟ + IRS
q ⎝ I0 ⎠
Let I0=10 A and Rs=20Ω. Plot I vs. Va for 0.3V ≤ Va ≤ 1V . (Hint: choose values
-16
for I and solve corresponding values for Va.)

Solutions Chapter 5
100
10-1
Current (Amperes)
10-2
10-3
10-4
10-5
10-6
10-7
10-8
10-9
0.4 0.5 0.6 0.7 0.8 0.9 1.0
Voltage (V)
5.18. A silicon pn junction has N D' =1017 cm-13 and N 'A =1018 cm-3. The junction
area is 100 µm2.What is the junction capacitance at Va=-5V?
qε N D' N 'A
The junction capacitance is given by C j = A
( )
2 N 'D + N A' (Vbi − Va )
The built-in voltage is

⎡ ⎤
kT ⎡ N D' N 'A ⎤ ⎢ 1017 ⋅1018 ⎥ = 0.89V
Vbi = ln ⎢ 2 ⎥ = 0.026V ln ⎢ 2⎥
q ⎣ ni ⎦
⎢⎣ 1.08 × 10
10
( ) ⎥⎦
Thus
C )(11.8)⎛ 8.85 × 10 −14

F ⎞ 17 −3
⎛ 10 cm ⎞ −8 2 (1.6 × 10
−19
⎝ cm ⎠
10 cm ⋅1018 cm −3
C j = 100µm2 ⎜ ⎟
⎝ µm 2 ⎠ 2(10 cm + 10 )(0.89V − (−5V ))
17 −3 18
= 0.036pF
5.19. A junction has N D' =1018 cm-3 and N 'A =1016 cm-3. Both sides are long, and
the fraction of reclaimable charge is 0.5. Compare the magnitudes of the
junction capacitance and the stored charge capacitance at Va=-5 V, 0 V, and +0.5

Solutions Chapter 5
V. The junction area is 100 µm2. Note that generation -recombination current
does not contribute to stored charge, only diffusion current contributes.
The junction capacitance is given by
qε N D' N 'A
Cj = A
( )
2 N 'D + N A' (Vbi − Va )
= 100 × 10 cm −8 2 (1.6 × 10 C )(11.8)(8.85 × 10 F / cm)(10 cm )(10

−19 −14 18 −3 16
cm−3 )
2(1.01× 10 cm )(0.92 − V )
18 −3
a
1
= 2.9 × 10 −14 F
(0.92 − V )
a
Cj(-5V)=1.2×10-14F
Cj(0V)=3.0×10-14F
Cj(+0.5V)=4.5×10−14F
The stored charge capacitance is given by

Csc = δIτ n
where I is the diffusion current. Since most of the depletion region is on
the more lightly dope side (the p-side), the lifetime associated with this
doping level is, from Figure 3.21,
τn≈1.2×10-4 s.
The diffusion current is
⎛ qVa ⎞
I = I 0 ⎜ e kT − 1⎟
⎝ ⎠
and
⎛D Dp ⎞ 30
I0 = Aq ⎜ n n p0 + pn 0 ⎟ = 100 × 10 −8 ⋅1.6 × 10 −19 ⋅ ⋅1.17 × 104 = 1.4 × 10 −18 A
⎝ Ln Lp ⎠ 0.04
where Dn and Ln were obtained from Figure 3.11 and 3.23 respectively,
and we have neglected the second term in the parentheses since
pn0<<np0.

Solutions Chapter 5
The stored charge capacitances:
−18 −4 −23
Csc (−5V ) = δIτ n = (0.5)(1.4 × 10 A )(1.2 × 10 s) = 8.4 × 10 F
Csc (0V ) = δIτ n = (0.6)(0A)(10 −6 s) = 0Farads (I=0)
0.5
−18
Csc (+0.5V ) = δIτ n = (0.5)(1.4 × 10 e 0.026
A )(1.2 × 10 −4 s) = 1.9 × 10 −14 F
To compare, we have
V=-5 V Csc≈0 Cj(-5V)=1.2×10-14 F
V=0 Csc=0 Cj(0V)=3×10-14 F
V=+0.5 Csc=1.9×10-14 F Cj(+0.5V)=4.5×10-14 F
As expected, the junction capacitance dominates at negative and low

voltages, but at significant forward bias the stored charged capacitance
becomes more important.
5.20. A bipolar transistor consists of two pn junctions back to back. In the

forward active mode, the emitter-base junction is forward biased and the base-
collector junction is reverse-biased. The base is generally very thin- much
shorter than a diffusion length. Figure P5.2 shows the energy band diagram.
EC
EV
a) Identify the emitter, base, and collector on the diagram.
Since the EB junction is forward biased, it must be the one on the left, and
since the CB junction is reverse-biased, it must be the one on the right.
Since they share the base as a common region, that must be the center
region.
b) Is this an npn or a pnp transistor?
npn
c) At the EB junction, are electrons injected or extracted?
Solutions Chapter 5
The EB junction is forward biased, and electrons are injected into the
base.
d) At the CB junction, are electrons injected or extracted?
Extracted by the reverse bias.
e) Sketch the electron concentration distribution you expect across the
base.
carrier
carriers concentration
injected distribution
carriers
extracted
EC
EV
f) This thin base region is the source of the term “short base diode.” The
goal in designing a transistor is to ensure that every electron injected into the
base from the emitter eventually ends up in the collector (so that IE≈IC).
Explain how this short-base structure helps. Hint: What else could happen to
the electrons?
The base is so short that electrons are extracted by the collector before
they have a chance to recombine in the base.
5.21. In Figure 5.39, explain how the diode current can be negative when the
applied voltage is zero.
Although the applied voltage is zero, there is still some residual charge
across the internal capacitance of the diode. The negative current is a
result of this capacitance discharging.
⎛ I ⎞
5.22. From Equation (5.120), the storage time is given by t s = τ n ln ⎜ 1+ F ⎟ . From
⎝ IR ⎠
Figure 5.39 (obtained from a SPICE plot), IF=4.2 mA and IR= -0.6 mA. Estimate
the electron lifetime in the p-type region.

Solutions Chapter 5
From Figure 5.39, the storage time is approximately 0.2 µs.
5.0mA
IF
4
3
ID
2
0.0 0.2 0.4 0.6 0.8 1.0 µs
ts≈0.2 µs
Thus, from Equation (5.120),
0.2 × 10 −6 s
τn = = 96ns
⎛ 4.2 × 10−3 ⎞
ln ⎜ 1+
⎝ −0.6 × 10−3 ⎟⎠

Solutions Chapter 5
Solutions Chapter 6
6.1. Consider a base-collector junction of a silicon BJT (bipolar junction

transistor) like that if Figure 6.1. Assuming a linearly graded junction with
a=1.2x1018 cm-3/um, find Vbi
This can be done iteratively from Equation (6.11)
⎡ 1
⎤
2kT ⎢ a ⎛ 12εV bi ⎞ 3 ⎥
Vbi = ln ⎢ ⎜ ⎟
q 2n ⎝ qa ⎠ ⎥
⎢⎣ i ⎥⎦
We choose a starting guess of 1V, and find:
Vbi cube root term logarithm new Vbi

1 0.000402474 10.0150146 5.21E-01
5.21E-01 0.00032381 9.79753926 5.09E-01
5.09E-01 0.000321449 9.7902212 5.09E-01
The built-in voltage is 0.51 V.
6.2. A silicon pn homojunction has a doping profile as indicated in Figure P6.1.

a) Find the value of the electric field in the bulk on the p side.
kT 1 −0.026V
From Equation (4.22), E p = − = = −0.72V / µ m = −7.2kV / cm
q λ 0.036 µ m
b) Find the electric field in the bulk on the n side.
Similarly on the n side,

Solutions Chapter 6
λn =
x − x0
=
( 0.6 − 0.3) µ m = −0.27 µ m
⎛ N (x ) ⎞ ⎡ 6 ×1016 ⎤
ln ⎜ D 0 ⎟ ln ⎢ 17 ⎥
⎝ N D ( x) ⎠ ⎣1.8 × 10 ⎦
kT 1 −0.026V
En =− = = +0.096V / µ m = 960V / cm
q λ −0.27 µ m
c) Plot NA-ND as a function of position [Hint: see Equation (4.16)], and

find the slope a.
We first find the grading constant a. One way to obtain it is express the
quantity NA-ND as a function of x, which we do using Equation (4.16):
− (x − x 0 )
(x − 0.2)
λp 18 − 0.036
N A (x) = N A (x 0 )e = 10 e
− (x − x 0 )
(x −.3)
λp −
N D (x ) = N D (x 0 )e = 6 × 1016 e −0.27
and the difference plotted below:
The slope is approximately –1018 cm-3/ µm, or a=1018cm-3/ µm (1022cm-4).
d) Find the built-in voltage.
This can be done iteratively from Equation (6.11)

Solutions Chapter 6
⎡ 1
⎤
⎛ ⎞
2kT ⎢ a 12εV bi 3 ⎥
Vbi = ln ⎢ ⎜ ⎟
q 2ni ⎝ qa ⎠ ⎥
⎢⎣ ⎥⎦
Picking a starting voltage of 1V, this converges rapidly to 0.831V:
Vbi cube root term logarithm new Vbi

1 1.98517E-05 16.0336938 8.34E-01
8.34E-01 1.86843E-05 15.9730874 8.31E-01
8.31E-01 1.86608E-05 15.9718251 8.31E-01
e) Find the junction width at equilibrium
1
⎡12 (11.8 ⋅ 8.85 × 10−14 F / cm ) 0.83V ⎤ 3
1
⎡12εVbi ⎤ 3
w=⎢ ⎥ =⎢ ⎥ = 1.9 × 10−5 cm = 0.19µ m
⎣ qa ⎦ ⎢⎣ (1.6 ×10 C )(10 cm ) ⎥⎦
−19 22 −4
f) For the junction width you found in (e), comment on the validity of
the linear approximation used over this distance.
From the figure above in part (c) we find that over a distance of 0.2 µm on
either side of the junction, the approximation is poor.
6.3. In section 6.2.2, it was claimed that hyper-abrupt junctions exhibit a large
fractional change in junction capacitance with applied voltage. Explain
physically why we should expect this to be the case.
In a hyper-abrupt junction, the doping concentration decreases as one

gets further from the junction. Let us consider the reverse bias regime. As
the bias voltage gets larger, the junction width must get large to uncover
the appropriate number of ionized dopants. Since the dopants get more
scarce away from the junction, as the bias increase the junction width
increases more and more rapidly. This changes the distance between the
capacitive plates more and more rapidly, thus changing the capacitance
more rapidly.
6.4. Figure P6.2 shows the equilibrium energy band diagram for a
heterojunction between n-type Semiconductor A (band gap 1.5 eV) and p-type
Semiconductor B (1.0 eV) at equilibrium.

Solutions Chapter 6
Evac
χ=4.0 eV
quantum well
χ=3.5 eV
EC
Eg=1.0eV
Ef
EV
Eg=1.5 eV
a) Indicate the directions of:

Electron diffusion →
Electron diffusion current Jn(diff) ←
Electron drift ←
Electron drift current Jn(drift) →
Hole diffusion ←
Hole diffusion current Jp(diff) ←
Hole drift →
Hole drift current Jp(drift) →
Effective electric field for holes →
True electric field →
b) In the conduction band edge, there is a "glitch", which looks like

a quantum well. The quantum well may have one or two discrete states.
i) Indicate on the drawing where the potential well is for the electrons
Indicated above.
ii) If an electron is trapped inside the well, list the mechanisms by

which it can get out.
The electron can tunnel out to the left, it can be thermionically emitted
over the barrier to the left or to the right, or it can recombine with a hole.

Solutions Chapter 6
6.5. Using the electron affinity model, draw (to scale) the energy band diagram
for a heterojunction between p-GaAs (Eg=1.43 eV, χ=4.07 eV) whose δp=0.1 eV,
and N-Al0.3Ga0.7As (Eg=1.8eV, χ=3.74 eV) whose δn=0.15 eV. Neglect interface
states.
Under neutrality, we have:
At equilibrium, it becomes:
Vbi= Vbi = ( χ 2 + Eg 2 − δ p ) − ( χ1 + δ n ) = ( 4.07 + 1.43 − 0.1) − ( 3.74 + 0.15 ) = 1.51V . .
The discontinuity in the conduction band edge is

Solutions Chapter 6
∆EC = ( χ 2 − χ 3 ) = 4.07 − 3.74 = 0.33eV
And the barrier for electrons is thus
EB( electrons ) = Vbi − ∆EC = 1.51 − 0.33 = 1.18eV .
The discontinuity in EV is
∆EV = ( χ1 + Eg1 ) − ( χ 2 + Eg 2 ) = ( 3.74 + 1.8 ) − (4.07 + 1.43) = 0.04eV
and thus the barrier to holes is
EB ( holes ) = Vbi + ∆EV = 1.51 + 0.04 = 1.55eV
6.6. Repeat the previous problem only now let the GaAs be n-type with δn=0.1
eV and let the AlGaAs be p-type with δp=0.15 eV.
Setting the Fermi levels equal gives

Solutions Chapter 6
Vbi= Vbi = ( χ1 + Eg1 − δ p ) − ( χ 2 + δ n ) = ( 3.74 + 1.8 − 0.15 ) − ( 4.07 + 0.1) = 1.22V
The discontinuity in the conduction band edge is
∆EC = ( χ 2 − χ1 ) = 4.07 − 3.74 = 0.33eV
And the barrier for electrons is thus
EB( electrons ) = Vbi − ∆EC = 1.22 + 0.33 = 1.55eV .
The discontinuity in EV is
∆EV = ( χ1 + Eg1 ) − ( χ 2 + Eg 2 ) = ( 3.74 + 1.8 ) − (4.07 + 1.43) = 0.04eV
and thus the barrier to holes is
EB ( holes ) = Vbi + ∆EV = 1.22 − 0.04 = 1.18eV
Now the glitch is in the valence band, and the barrier for holes is smaller
than for electrons.
6.7. Consider the Type I Np heterojunction of Figures 6.8 and 6.9 in which the
net doping N D' and N A' are uniform on the N and p sides respectively. Let εn be
the permittivity on the N side and εp that on the p-side. Solve Poisson's
equation to find the depletion widths wn and wp on the n and p sides and the

Solutions Chapter 6
total depletion width. This is similar to the procedure used in Chapter 5 for
homojunctions. (Hint: At the interface, displacement (εE) is continuous.)
Poisson’s Equation:
QV
E ( x) = ∫ dx
ε
x
qN D' N D'
On the N side, E n = ∫ dx ' = q ( x − xn ) for xn<x<x0
xn
εn εn
xp
−qN A' N A'
On the p side, E p = ∫ εp
dx ' = q
εp
(x p − x ) for x0<x<xp
x
From continuity of displacement, we have
ε nE n ( x0 ) = ε pE p ( x0 )
Therefore
(
qN 'D (x 0 − x n ) = qN A' x p − x 0 or )
N D' w n = N A' w p
We integrate the field to find the voltage: V ( x) = − ∫E ( x)dx ( )

qN D'
(x − x 0)
2
N side: Vn (x ) − Vn (x n ) = −
2ε n
qN 'A
( )
2
p side: Vp (x ) − Vp (x p ) = − x −x
2ε p p
On the N side the total built-in voltage is
qN D' qN D' 2
(x − x n ) =
2
V =n
w
j
2ε n 0 2ε n n
qN 'A
( ) qN A' 2
2
and on the p side V j = x − x0 =
p
w
2ε p p 2ε p p
The total junction voltage is

Solutions Chapter 6
qN D' 2 qN 'A 2
Vbi = V jn + V jp = w + w
2ε n n 2ε p p
Substituting in N D w n = N A w p , we find
' '
2
qN D' 2 qN A' ⎛ N D' ⎞ 2
Vbi = w + ⎜ ' ⎟ w n or
2ε n n 2ε p ⎝ NA ⎠
Vbi ⎛ 2ε nε p N A' ⎞
wn = ⋅⎜
qN D' qN D'2 ⎜ 2ε ε N ' ⎟⎟
+ ⎝ n p A⎠
2ε n 2ε p N A'
2ε nε p N A' Vbi
=
qN D' ( ε p N A' + ε n N D' )
Similarly
2ε n ε p N 'DV bi
wp =
(
qN A' ε p N A' + ε n N D' )
And the total junction width is
2ε nε p N A' Vbi 2ε nε p N D' Vbi

w = wn + wp = +
qN D' ( ε p N A' + ε n N D' ) qN A' ( ε p N A' + ε n N D' )
2ε nε p ( N D' + N A' ) Vbi

2
=
qN D' N A' ( ε n N D' + ε p N A' )
6.8. Show that the junction capacitance per unit area for Problem 6.7 can be
written as
dQV ⎡ qε nε p N A' N D' ⎤
Cj= = A ⎢ ⎥.
dVa
⎣ (
⎢ 2 ε n N D' + ε p N A' (Vbi − Va ) ⎥
⎦)

Solutions Chapter 6
dQV
Cj =
dVa
In our case the charge is qN D' wn = qN A' wp and we have
⎡ d ⎤ ⎡ d ⎤
dQV
= A⎢ ( qN D' wn ) ⎥ = A ⎢ ( qN A' wp )⎥
dVa ⎣ dVa ⎦ ⎣ dVa ⎦
d ⎡ ' 2ε nε p N A (Vbi − Va ) ⎤
'
= A ⎢ qN D ⎥
dVa ⎢ qN D' ( ε p N A' + ε n N D' ) ⎥
⎣ ⎦
⎡ 2ε nε p N A' ⎤ d 1
⎢
= A qN D
'
+⎥ (Vbi − Va ) 2
⎢ qN D' ( ε p N A' + ε n N D' ) ⎥ dVa
⎣ ⎦
⎡ q 2ε nε p N A' N D' ⎤ ⎛ 1 ⎞ 1
( )
−
= A⎢ ⎥ − −
(ε p N A' + ε n N D' ) ⎥⎦ ⎜⎝ 2 ⎟⎠ bi a
V V 2
⎢
⎣
⎡ qε nε p N A' N D' ⎤
= A⎢ ⎥
⎢ 2 ( ε p N A' + ε n N D' ) (Vbi − Va ) ⎥
⎣ ⎦
6.9. For the Si:Ge Nn junction of Figure 6.17, sketch the energy band diagram
that you would expect using the simple electron affinity model (i.e., ignoring
tunneling-induced dipoles and interface states). Discuss the difference in
current flow that would result from the two energy band diagrams (simple
model, and considering the effects of the presence of dipoles). Let EC-Ef=0.1
eV for silicon and 0.15 eV for Ge.
Using the EAM, we would obtain the diagram shown below:

Solutions Chapter 6
Evac
∆EC
EC
∆EV
EV
silicon germanium
Current would flow easily in both directions in the ideal case since there is
no barrier to overcome as there is in Figure 6.17. Although that spike may
be thin enough to support tunneling, the current across the junction would
not be as great as in the ideal case.
If we add tunneling-induced dipoles to the case above, we would allow for

electrons from the valence band of the Ge to tunnel a short distance into
the silicon. This would place a negative charge near the surface of the Si
and a resulting positive charge near the surface of the germanium. Thus
we would expect band bending and a reduction of the barriers above.
Evac
∆EC
EC
EV
silicon germanium

Solutions Chapter 6
6.10. We saw than in a heterojunction in which the spike and notch occurred in
the conduction band (as in Figure 6.12), it was possible for electrons in the
valence band of the narrow-gap material to tunnel a short distance into the
forbidden band of the wide-band-gap material. This tunneling induced a
dipole that produced a discontinuity in the bands at the junction. Consider a
Pn heterojunction such as that in Figure P6.3. Comment on the possibility of
tunneling-induced dipoles in this case.
The possibility for electrons in the valence band to tunnel into the
forbidden band exists, and we can expect similar band alterations as a
result.
6.11. Consider an n-GaAs Schottky barrier diode of area 10 µm2 and potential
barrier EB (0) = 1eV .
a. Find the value of I0 from Equation (6.15).
qm* (kT ) − E BkT(0)

2
J0 = e
2π 2 h 3
(1.6 × 10 C)(0.067 ⋅ 9.1× 10−31 kg )(1.38 × 10 −23 J/ K ⋅ 300K )
−19 2
exp ⎛ −
1eV ⎞
=
2π 2 (1.06 × 10 −34 J⋅ s ) ⎝ 0.026eV ⎠
3
= (7.1 × 109 )(1.98 × 10 −17 )

= 1.4 × 10 −7 A / cm 2 = 1.4 × 10−11 A / cm2
where we used the conductivity effective mass for electrons in GaAs,
because a Schottky barrier is a majority carrier device.

Solutions Chapter 6
b. Plot the forward I-V characteristic for a diode quality factor of
n=1.3.
The equation for the current is Equation (6.14) (but note footnote in text):
⎡ qV a ⎤
I = JA = AJ0 ⎢e nkT − 1⎥ . With J0=1.4×10-11A/cm2 and n=1.3, we obtain the
⎣ ⎦
plot below.
6.12. Consider the junction of Figure P6.4.
qVbi
Evac
E vac
EC
Ef Ef
E
V
Metal Semiconductor
E
C

Solutions Chapter 6
a) Draw its energy band diagram under forward and reverse bias.
Evac
E vac
EC
Ef Ef
E
V
Metal Semiconductor
E
C
Forward bias
Evac
E vac
EC
Ef
Ef
Metal E
V
Semiconductor
E
C
Reverse bias.
b) Will this junction be ohmic or rectifying? Why?

Solutions Chapter 6
This is a metal-p-type semiconductor junction. Electrons from the metal
can flow easily to and from the valence band under either bias polarity.
6.13. Consider an n-Si:Al Schottky barrier diode for which N D' =1017cm-3 and
the measured built-in voltage is 0.70 V. Let the junction area be 10µm2.
a) Find the depletion width w. Note that this junction can be treated as a
one-sided junction.
We use Equation (5.40) for a p+n junction because the depletion region is
entirely in n-type material:
1
⎡ 2(11.8)(8.85 × 10 −14 F / cm)(0.7)⎤ 2
1
⎡ 2εV j ⎤ 2
−6
w=⎢ ' ⎥ =⎢ ⎥ = 9.6 × 10 cm = 0.096 µm
⎣ D⎦
qN ⎢⎣ (1.6 × 10 −19
C)10 17
cm−3
⎥⎦
b) Find the junction capacitance at Va=-5 V.
First we find the junction width at Va=-5 V:
( )
1
⎡ 2ε (Vbi − Va ) ⎤ 2 ⎡ 2(11.8) 8.85 × 10 F / cm ( 0.7 − (−5) ) ⎤ 2
−14
w=⎢ ⎥ =⎢ ⎥
⎣ qN D' ⎦ ⎢⎣ ( )
1.6 × 10−19 C 1017 cm −3 ⎥⎦
= 2.7 × 10−5 cm = 0.27 µ m
εA (11.8) (8.85 ×10−14 F / C )(10 ×10−8 cm2 )

Cj = = = 3.9 × 10−15 F
w 2.7 × 10−5 cm
6.14. A heterojunction is formed between n-type GaAs of N D' =1016 cm-3 and p-
type germanium of N A' =1017 . The measured discontinuities in the band edges
are ∆EC=0.27 eV and ∆EV=0.49 eV.
a) Find the barrier to electrons in eV.
⎛ N D' ⎞ ⎛ 1016 ⎞
In the GaAs, EC − E f = δ n = − kT ln ⎜ ⎟ = −0.026eV ln ⎜ 17 ⎟
= 0.098eV
⎝ NC ⎠ ⎝ 4.4 × 10 ⎠

Solutions Chapter 6
⎛ N' ⎞ ⎛ 1017 ⎞
In the Ge, E f − EV = δ p = − kT ln ⎜ A ⎟ = −0, 026eV ln ⎜ 18 ⎟
= 0.096eV
⎝ NV ⎠ ⎝ 4 × 10 ⎠
The energy difference between the conduction band edge in the Ge and
the Fermi level is Eg-δp=0.67-0.096=0.57eV. The barrier is this height,
minus the difference between the conduction band edge in the GaAs and
the Fermi level, or the barrier to electrons=0.57-0.098≈0.47 eV.
qV bi
E
vac
χ=4.07
χ=4.0
barrier for
electrons
0.47eV E
C
0.57eV 0.096eV
Ef
EV
0.098eV ∆E =0.27eV
C
∆E =0.49eV barrier
V for holes =1.24 eV
b) Find the barrier to holes in eV.
The barrier to holes is equal to Eg(GaAs)-δn -δp=1.43-0.096-0.098=1.24eV.
c) Find the built-in voltage (measured at Evac) in eV.
From the figure, Vbi = χ GaAs + 0.47 − χ Ge = 4.07 + 0.47 − 4.0 = 0.54eV
d). To what is the physical source of the discontinuity in the vacuum

level attributed?
The discontinuity is caused by the tunneling-induced dipoles.

Solutions Chapter 6
Solutions Chapter 7
7.1. In any circuit, the transistor operation can be understood by examining the
superposition of the transistor characteristics and the conditions imposed by the
circuit (the load line). What differentiates, then, a digital transistor circuit from an
analog one?
The difference is in how the circuit is operated. In a digital circuit, the gate
voltage is varied between two values, one that produces cutoff and one in
the sublinear region. In an analog circuit, the gate voltage is normally
varied within the current saturation range, producing proportional changes
in the output (drain) voltage.
7.2. In modern FETs, the gate is usually degenerately doped silicon, whose Fermi
level is essentially at the bottom of the conduction band edge (for an NFET) or at the
top of the valence band edge (for a PFET).
a) Draw an energy band diagram similar to Figure 7.5b, except making the
transistor a PFET.
b) Suppose the device is an NFET, but the gate is made of metal (as was done
in the early days). Draw the energy band diagram. Take ΦM<ΦS.
The energy band diagram is:

Solutions Chapter 7
7.3. For the transistor of Figure P7.1, by how much should the gate voltage be
changed to produce inversion? Threshold? Assume half the applied voltage appears
across the oxide and half across the semiconductor. If the device in the figure is at
equilibrium, is this an enhancement or a depletion FET?
We need to bend the bands in the Si by an additional 0.1 eV to just invert

the surface, meaning the gate voltage should be increased by 2×0.1=0.2V.
To produce threshold, the surface should be inverted by 0.3 eV, requiring
another 0.3eV of band bending or 0.6V. Thus the change in gate voltage
required to reach threshold if 0.6+0.2=0.8V. As no channel exists at
equilibrium, this is an enhancement device.
7.4. For each of the transistors of Figure 7.10,

a) What is the polarity of the threshold voltage VT ?
b) What is the polarity of VDS that should be used?
c) When VGS=0 (equilibrium), is the transistor on or off?
d) Is the current ID carried by electrons or holes?
e) For VGS>VT, is the current ID carried primarily by drift or diffusion?
Enhancement Enhancement Depletion Depletion
NFET PFET NFET PFET
VT Positive Negative Negative Positive
VDS Positive Negative Positive Negative
VGS=0 Off Off On On
Carriers Electrons Holes Electrons Holes
Current Drift Drift Drift Drift

Solutions Chapter 7
7.5 An NFET is fabricated with a degenerately doped n-type gate. What doping
concentration in the p-type substrate is needed for there to be a channel even with
no voltage applied? Assume half the built-in voltage is dropped across the oxide and
half across the silicon. Let the definition of “a channel exists” be that the surface of
the silicon is inverted such that the electron concentration at the Si/SiO2 interface is
equal to the hole concentration in the bulk p-type silicon.. Is the doping you found a
minimum or a maximum required to create a depletion-type device?
We require that the surface be inverted by qφf. That means the band
bending in the semiconductor must be 2φf. Since that much is also
dropped across the oxide, the built-in voltage must be 4φf.
Evac
qφs=2qφf
Vbi
qφ ox=2qφf
E vac
ΦS χS
Φ G=χ G
EC
qφf
Ei
EC=Ef Ef
qφ f
EV
qφs=2qφf
EV
Φ G − Φ S = 2 ( 2φ f ) = 4φ f . From the figure,

1
The built-in voltage is Vbi =
q
Eg
ΦS = χ + + qφ f , and Φ G = χ . Therefore
2
Eg
Φ S − ΦG = χ + + qφ f − χ = 4qφ f
2
Eg
= 3qφ f
2
Eg 1.12eV
qφ f = = = 0.19eV
6 6
This is not going to be degenerately doped, so we can use the Boltzmann
approximation to obtain
Ei − E f (0.19 eV )
N A' = p0 = ni e kT = 1.08 × 1010 cm −3e (0.026 eV = 1.6 × 1013 cm −3

Heavier doping will reduce the amount of inversion on the surface, so
lighter doping than N A' is required to ensure a channel at the surface.
7.6. Consider two silicon MOSFETs, one n channel and the other p channel, with
substrate dopings of 1016cm-3. The NMOS has an n+ gate and the PMOS has a p+

Solutions Chapter 7
gate, both doped to 1019cm-3. Find the built-in voltage Vbi for each, and draw the
energy band diagram. Neglect the band-gap narrowing effects discussed in Chapter
2.
For the NMOS, we construct the energy band diagram, keeping in mind
that the substrate is p-type and that since both the gate and substrate are
silicon, they have the same electron affinity χ. We find the location of the
Fermi level with respect to the valence band edge from
⎛N ⎞ ⎛ 1016 ⎞
E f − EV = −kT ln ⎜ A ⎟ = −0.026eV ln⎜ = 0.21eV .
⎝N ⎠
V ⎝ 3.1× 1019 ⎟⎠
For the poly side, we take the Fermi level to be at the conduction band
edge since the material is degenerately doped.
The energy band diagram is
and the built-in voltage is 4.05 +1.12 − 0.21 − 4.05 = 0.91

⎛N ⎞ ⎛ 1016 ⎞
For the PMOS, EC − E f = − kT ln ⎜ D ⎟ = −0.026eV ln ⎜ ⎟ = 0.21eV
⎝ 2.86 × 10
19
⎝ NC ⎠ ⎠
The energy band diagram is

Solutions Chapter 7
and the built-in voltage= (4.05+1.12)- 4.05-0.21=0.91eV.
7.7. In the MOS process, structures like the gate of a transistor are used to make
capacitors as well. If the oxide thickness is 4 nm, what area is needed to achieve a
capacitance of 1 pF? The permittivity of silicon dioxide is 3.9ε0.
The capacitance per unit area is
Cox =
'
=
(
ε ox (3.9 ) 8.85 × 10 F / cm
−14
) = 8.63 × 10−7 F / cm2
t ox −7
4 × 10 cm
So the area required to obtain 1 pF is
C 1 ×10 −12 F −6
A= ' = = 1.16 × 10 cm .
2
−7
C 8.63 × 10 F / cm 2
This corresponds to a square about 11 µm on a side.
7.8. In MOS processing, the W/L ratio is often intentionally made different from
transistor to transistor. Plot the I-V characteristics for the device of Figure 7.19 with
the W/L ratio changed to 10 instead of 5. Compare this to the results for W/L=5.
Using Equation (7.19), we have, for VDS<VDSsat ,and VGS>VT,:
⎛W ⎞ ⎡ V2 ⎤ ⎛ cm 2 ⎞ ⎡ VD2 ⎤
⎝L⎠
( )
I D = ⎜ ⎟ Cox’ µ ⎢(VGS − VT ) VDS − DS ⎥ = 10 8.63 × 10−7 F / cm 2 ⎜ 500 ⎟ ⎢( G
V ⋅s ⎠⎣
V − VT ) D
V − ⎥
⎣ 2 ⎦ ⎝ 2 ⎦
⎡ V2 ⎤
= 4.31 × 10−3 ⎢(VGS − VT ) VDS − DS ⎥
⎣ 2 ⎦
We will choose VGS-VT to be 1V, 2V, 3V, and 4 V. This only applies up to
VDS=VDSsat for each curve, or up to VDSsat=VGS-VT= 1, 2, 3, and 4
respectively. After that, the current is constant for each curve. The
resulting plot is

Solutions Chapter 7
7.9. Explain why the electron affinity model can be used to good approximation to
determine the band line-up normal to the gate in a MOSFET. That is, why is the
tunneling- induced dipole effect negligible?
Recall from Figure SI.A.9 that the tunneling depth is smallest near the
center of the band gap and larger near the band edges. Because the band
gap of the oxide is so large, the electron barely penetrates into the oxide.
Since the dipole effect is proportional to the penetration depth, its effect is
negligible in this case.
7.10. Plot the ID-VDS characteristics for an NFET, using the long channel model, for
which W=10µm , L=1µm, tox=4 nm, VT =0.25V, and the channel length modulation
parameter is λ=0.04V-1. Use VGS= 1V, 2V, 3V, and 4V. Find the output conductance
in saturation for VGS=3V.
For tox=4 nm, Cox' =

ε ox
=
ε r( ox )ε 0
=
(3.9)(8.85 × 10 −14
F / cm ) = 8.6 × 10 −7
F / cm2
t ox t ox −7
4 × 10 cm
WCox’
µ⎡ VDS2
⎤
ID = (
⎢ GS
V − VT ) VDS − ⎥
L ⎣ 2 ⎦
⎛ cm 2 ⎞
(
(10µ m ) 8.6 × 10−7 F / cm ⎜ 500 ) ⎟
V ⋅s⎠ ⎡ VD2 ⎤
⎝
= ⎢(VG − VT ) VD − ⎥
1µ m ⎣ 2 ⎦
⎡ V2 ⎤
= 4.3 × 10−3 ⎢(VG − VT ) VD − D ⎥
⎣ 2 ⎦

Solutions Chapter 7
and VDSsat=VGS-VT.
Above saturation, from Equation (7.43), I D = I Dsat [1 + λ (VDS − VDSsat ) ] with

λ=0.04V-1, where from the long-channel model
WCox’ µ ⎛ VDSsat
2
⎞
I Dsat = ⎜ ⎟
L ⎝ 2 ⎠
For VGS=1V, VDSsat=0.75 V and IDsat=1.21mA
For VGs=2V, VDSat=1.75V and IDsat=6.58mA
For VGS=3V, VDSsat=2.75 V, and IDsat=16.25mA
For VG S=4V, VDSsat=3.75V and IDsat=30.2mA
The output conductance is defined as ∂I D / ∂VD , which for VGS=3V is

∆I D 17.7mA − 16.2mA
= = 0.68mS (From the data used to make the plot).
∆VDS 4.96V − 2.77V
7.11. An enhancement NFET with the characteristics in Table 7.1 has a threshold
voltage of VT=1V, a channel length of 1 µm and a width of 5 µm. Considering
velocity saturation, with vsat=5 x 106 cm/s, find the current ID for
a) VGS=0 V, VDS= 1V
Here VG<VT, so the device is below threshold, so the current is essentially
zero.
b) VGS=2V, VDS=1 V

Solutions Chapter 7
⎡ 1
⎤
vsat ⎢ ⎛ 2 µ (V − V ) ⎞ 2
⎥
= L ⎢⎜⎜ 1 + ⎟ − 1⎥
lf GS T
VDsat ⎟
µlf vsat L
⎢⎣⎝ ⎠ ⎥⎦
⎡ 1
⎤
( 5 × 106 cm / s ) ⎢ ⎛ 2 500cm 2
(
/ V ⋅ s ( 2 − 1) V)⎞ 2
⎥
= (
10−4 cm ⎢⎜ 1 + ) ⎟ − 1⎥
500cm / V ⋅ s
2
⎢⎜⎝ 5 ×10 cm / s 10 cm ⎟⎠
6 2 −4
( )⎥
⎣ ⎦
= 0.73V
Thus we are above saturation (VDS>VDSsat), and from Equation (7.67),
⎛ V ⎞
⎜ VGS − VT − DSsat ⎟ VDSsat
WCox µlf ⎝
’
2 ⎠
I D = I Dsat =
L µ V
1 + lf DSsat
Lvsat
⎛ cm 2 ⎞
( 5 ×10 −4
cm )( 8.6 × 10−7 F / cm ) ⎜ 500 ⎟
V ⋅s ⎠
⎛
⎜ 2 −1−
0.73 ⎞
⎟ 0.73V
= ⎝ ⋅ ⎝ 2 ⎠
10−4 cm cm 2
500 ( 0.73V )
1 + −4 V ⋅ s
10 cm ( 5 ×106 cm / s )
= 0.58mA
Alternately, using Equation (7.69)

I Dsat = WCox' vsat (VGS − VT − VDsat )
= 5x10-4 cm ( 8.6x10-7 F / cm )( 5 × 106 cm / s ) ( 2 − 1 − 0.73) V = 0.58mA
c) VGS=3V, VDS=1 V
⎡ 1
⎤
vsat ⎢⎛ 2 µlf (VGS − VT ) ⎞ 2 ⎥
VDSsat = L ⎜1 + ⎟⎟ − 1⎥
µlf ⎢⎜⎝ vsat L ⎠
⎢⎣ ⎥⎦
⎡ 1
⎤
⎢⎛ ⎛ cm 2 ⎞ ⎞2 ⎥
⎟ ( 3 − 1)
(5 ×106 cm / s )⎢⎜
2 ⎜ 500
V ⋅s ⎠
⎟ ⎥
= −4 ⎜(
10 cm ⎢ 1 + ⎝
) ⎟ − 1⎥
500
cm 2 ⎢⎜
⎜
(
5 ×106 cm / s 10−4 cm )( ) ⎟
⎟⎟
⎥
V ⋅s ⎢⎜ ⎥
⎢⎣⎝ ⎠ ⎥⎦
= 1.2V
This is greater than our VDS, so now the device is not saturated. The
current is

Solutions Chapter 7
⎛ V ⎞
V − V − DS ⎟ VDS
WCox’ µlf ⎜⎝ GS T 2 ⎠
I D = I Dsat =
L µ V
1 + lf DS
Lvsat
⎛ F ⎞⎛ cm 2 ⎞
( 5 ×10 −4
cm ) ⎜ 8.6 ×10−7
⎝
⎟⎜
cm ⎠ ⎝
500 ⎟
V ⋅s ⎠
⎛ 1 ⎞
⎜ 3V − 1V − V ⎟1V
= ⋅ ⎝ 2 ⎠
10−4 cm ⎛ cm 2 ⎞
⎜ 500 ⎟1V
⎝ V ⋅s ⎠
1 + −4
10 cm ( 5 ×106 cm / s )
= 1.61mA
7.12. An NFET is made with tox=4 nm, L=1µm, W=2 µm, VT=1V, and
µlf = 500cm2 / V ⋅ s .
If the simple model is used, what should the width of the PFET be to get the
same characteristics (apart from polarity). Let the low field mobility for holes be
200 cm2/V·s.
WCox' µ
(VGS − VT ) .
2
From the long channel model (Equation (7.25)), I Dsat =
L
Thus
Wn Cox' µn
(VGS − VT )
2
I Dsatn L Wµ
= =1= n n
I Dsatp W p Cox µ p
'
Wp µ p
(VGS − VT )
2
L
W µ ⎛ 500⎞
Thus W p = n n = 2µm = 5µ m
µp ⎝ 200⎠
7.13. An NFET and a PFET are made on the same chip, using the same process. The
NFET has Cox' =8.6×10-7F/cm2, tox=4 nm, L=0.2 µm, W=15µm, VT=1.5 V, and µlf=500
cm2/V·s. If the PFET is identical except for its mobility (200cm2/Vs) and width, W.
a) What should W be for the PFET to make the characteristics the same as
for the NFET, as predicted by the simple model?
W ⎡ V ⎤
I DNFET = NFET Cox’ µlfNFET ⎢(VGS − VT ) − DS ⎥ VDS
LNFET ⎣ 2 ⎦
= WNFET µlfNFET [ X ] = 15 × 500 [ X ]
⎡ V ⎤
where X = ⎢(VGS − VT ) − DS ⎥ VDS is a constant that is the same for both the
⎣ 2 ⎦
NFET and the PFET.
We want this to equal that for the PFET:

Solutions Chapter 7
I DPFET = WNPET µlfPFET (X ) = WNPET (200)(X )
Setting these equal, we obtain:
15 ⋅500 = WPFET ⋅ 200
WPFET = 37.5µm
b) Find VDSsat and IDsat for VGS-VT=1 V

Using the simple model, VDSsat = VGS − VT = 1V
WCox’
µlf ⎡ 2
VDSsat ⎤
I Dsat = ⎢(VGS − VT ) VDSsat − ⎥
L ⎣ 2 ⎦
⎛ cm 2 ⎞
(15 × 10 cm )(8.6 × 10 F / cm ) ⎜ 500 V ⋅ s ⎟ ⎡
−4 −7
⎝ ⎠ 1 1 −1 ⎤
2
= ⎢ ( )( ) ⎥
0.2 × 10−4 cm ⎣ 2⎦
= 16mA
c) If velocity saturation is considered, How different are VDSsat and IDsat for
the NFET and the PFET compared to the simple model results? Express your result
as a ratio (e.g., IDsatNFET/IDsatPFET and VDSsatNFET/VDSsatPFET).
For the NFET, we have W=15 µm and L=0.2 µm, with µlf=500 cm2/V·s
⎡ 1
⎤
vsat ⎢ ⎛ 2 µ (V − V ) ⎞ 2
⎥
VDSsatNFET = L ⎢⎜ 1 + ⎟⎟ − 1⎥
lfNFET GS T
µlfNFET ⎝ ⎜ vsat L
⎢⎣ ⎠ ⎥⎦
⎡ 1
⎤
⎢ ⎛ ⎛ cm 2
⎞ ⎞ 2
⎥
⎢ ⎜ 2 ⎜ 500 ⎟ (1V ) ⎟ ⎥
4 × 10 cm / s
6
⎜1 + ⎝ V ⋅s⎠ ⎟ − 1⎥
= 0.2 × 10 −4
cm ⎢
cm 2 ⎢⎜ 4 × 106 cm ⋅ 0.2 × 10−4 cm ⎟ ⎥
500 ⎢ ⎜⎜ ⎟⎟ ⎥
V ⋅s s
⎢⎣⎝ ⎠ ⎥⎦
⎡ 1
⎤
= 0.16 ⎢(1 + 12.5 ) 2 − 1⎥
⎣ ⎦
= 0.43V
For the PFET,

Solutions Chapter 7
⎡ 1
⎤
vsat ⎢ ⎛ 2 µ (V − V ) ⎞ 2
⎥
VDSsatPFET = L ⎢⎜ 1 + ⎟⎟ − 1⎥
lfPFET GS T
µlfPFET ⎝ ⎜ vsat L
⎢⎣ ⎠ ⎥⎦
⎡ 1
⎤
⎢ ⎛ ⎛ cm ⎞ 2
⎞ 2
⎥
⎢ ⎜ 2 ⎜ 200 ⎟ (1V ) ⎟ ⎥
4 × 10 cm / s
6
⎜1 + ⎝ V ⋅s⎠ ⎟ − 1⎥
= 0.2 × 10 −4
cm ⎢
cm 2 ⎢⎜ 4 × 106 cm ⋅ 0.2 × 10−4 cm ⎟ ⎥
200 ⎢ ⎜⎜ ⎟⎟ ⎥
V ⋅s s
⎢⎣⎝ ⎠ ⎥⎦
⎡ 1
⎤
= 0.16 ⎢(1 + 5 ) 2 − 1⎥
⎣ ⎦
= 0.58V
V 0.43V
The ratio of these is DSsatNFET = = 0.74
VDSsatPFET 0.58V
The saturation currents are

I ( simple model ) 16mA
I DsatNFET = D = = 4.34mA
µlfNFET VDSsatNFET ⎛ cm 2 ⎞
1+ ⎜ 500 ⎟ ( 0.43V )
Lvsat ⎝ V ⋅s⎠
1+
( 0.2 × 10 −4
cm )( 4 × 106 cm / s )
and
I D ( simple model ) 16mA
I DsatPFET = = = 6.53mA
µlfPFETVDsatPFET ⎛ cm 2 ⎞
1+ ⎜ 200 ⎟ ( 0.58V )
Lv sat ⎝ V ⋅s⎠
1+
( 0.2 × 10−4 cm )( 4 × 106 cm / s )
IDsatNFET 4.34mA
The ratio of these is = = 0.66
IDsatPFET 6.53mA
7.14. A good way to check the validity of a derivation is to verify that it reduces to
the expected result for a particular known case. For example, in the simple model
we neglected velocity saturation, and in the later model we took it into account.
Since the high field that causes velocity saturation occurs in shorter channel devices,
we would expect that the later model would reduce to the simple model for long-
channel devices.
a) Show that in the limit of large L, VDSsat as given by Equation (7.68)
reduces to VDSsat = (VGS - VT) as given by Equation (7.26) for the simple (long
channel) model .
You may find the following information useful:
For x < 1

Solutions Chapter 7
n(n − 1) x 2
(1 ± x) = 1 ± nx + ± ...
n
2!
n(n + 1)x 2
(1 ± x)−n = 1 m nx + m ...
2!
Expanding Equation (7.68) using the relation given yields
⎡ 1
⎤
vsat L ⎢ ⎛ 2 µ lf ( GS
V − VT ) ⎞ ⎥
2
VDsat = ⎢ ⎜⎜1 + ⎟⎟ − 1⎥
µlf ⎝ vsat L ⎠
⎢⎣ ⎥⎦
vsat L ⎢⎡⎛ 1 2µlf (VGS − VT ) 1 ⎛ 1 ⎞ ⎪⎧ 2µlf (VGS − VT ) ⎪⎫ ⎞ ⎤

2
= ⎜1 + + ⎜ − 1⎟ ⎨ ⎬ + H .O.T ⎟ − 1⎥
µlf ⎢⎜ 2 vsat L 2 ⎝ 2 ⎠ ⎩⎪ vsat L ⎭⎪ ⎟ ⎥
⎣⎝ ⎠ ⎦
If the length L is large, then the term
2µlf (VGS − VT )
becomes small when raised to higher powers and the
vsat L
squared and higher order terms can be neglected.
Thus
v L ⎡⎛ 1 2µlf (VG − VT ) ⎞ ⎤
VDsat = sat ⎢⎜ 1 + ⎟⎟ − 1⎥
µlf ⎢⎣⎜⎝ 2 vsat L ⎠ ⎥⎦
v L ⎛ µlf (VG − VT ) ⎞
= sat ⎜ ⎟⎟
µlf ⎜⎝ vsat L ⎠
= (VG − VT )
b) Show that in the limit of large L, IDsat as given by Equation (7.61)
reduces to that of Equation (7.25) for the simple model.
WCox’ µlf ⎡ 2
VDSsat ⎤
I Dsat =
µlf VDSsat ⎣ ⎢ (V GS − VT ) DSsat
V − ⎥
2 ⎦
L+
vsat
µlf VDSsat
If L becomes large compared to , this term can be neglected in the
vsat
denominator and this expression reduces immediately to Equation (7.25):
WCox' µlf ⎡ 2
VDSsat ⎤
I Dsat = (
⎢ GS
V − V T ) V DSsat − ⎥
L ⎣ 2 ⎦
To see whether L is in fact large compared to the neglected term, we
choose some typical numbers: µlf=500cm2/V·s, VDSsat≈2V, and vsat=4×106
cm/s. Then
µlf VDSsat ( 500 )( 2 )
= = 2.5 × 10−4 cm = 2.5µ m .
vsat 4 × 106
For long channel devices this length is easily exceeded.

Solutions Chapter 7
7.16. Find an approximate closed-form expression for the drain current, accounting
for series resistance and velocity saturation. That is, solve for ID in Equation (7.72).
Neglect the term in ID2.
Equation (7.72):
⎛ V ⎞
WCox' µlf ⎜ VGS − VT − DS ⎟ (VDS − 2 I D RS )
ID = ⎝ 2 ⎠
, which applies only for
⎡ µlf (VDS − 2 I D RS ) ⎤
L ⎢1 + ⎥
⎣ Lvsat ⎦
VDS<VDSsat.
Multiplying both sides by the denominator:
⎧⎪ ⎡ µlf (VDS − 2 I D RS ) ⎤ ⎫⎪ ⎛ VDS ⎞
I D ⎨ L ⎢1 + ⎥⎬ = WCox µlf ⎜ VGS − VT − ⎟ (VDS − 2 I D RS )
'
⎪⎩ ⎣ Lvsat ⎦⎭⎪ ⎝ 2 ⎠
and multiplying through
µlf (VDS − 2 I D RS ) ⎛ V ⎞ ⎛ V ⎞
ID L + ID = WCox' µlf ⎜ VGS − VT − DS ⎟ VDS − WCox' µlf 2 RS I D ⎜ VGS − VT − DS ⎟
vsat ⎝ 2 ⎠ ⎝ 2 ⎠
µlf VDS µlf RS ⎛ V ⎞ ⎛ V ⎞

ID L + ID − 2 I D2 = WCox' µlf ⎜ VGS − VT − DS ⎟ VDS − WCox' µlf 2 I D RS ⎜ VGS − VT − DS ⎟
vsat vsat ⎝ 2 ⎠ ⎝ 2 ⎠
Neglecting the ID2 term and collecting all terms in ID on the left,
µlf VDS ⎛ V ⎞ ⎛ V ⎞
ID L + ID − WCox' µlf 2 I D RS ⎜ VGS − VT − DS ⎟ = WCox' µlf ⎜ VGS − VT − DS ⎟ VDS
vsat ⎝ 2 ⎠ ⎝ 2 ⎠
Dividing by the coefficient of ID:
⎛ V ⎞
WCox' µlf ⎜ VGS − VT − DS ⎟ VDS
ID = ⎝ 2 ⎠
µlf VDS ⎛ V ⎞
L+ − WCox' µlf 2 RS ⎜ VGS − VT − DS ⎟
vsat ⎝ 2 ⎠
Take out the L in the denominator, and multiplying the third and fourth
terms in the denominator by vsat/vsat:
⎛ V ⎞
ID = ⎝ 2 ⎠
⎡ µlf VDS WCox' µlf 2 RS vsat (VGS − VT ) V ⎛ WCox' µlf 2 RS vsat ⎞ ⎤
L ⎢1 + − − DS ⎜ ⎟⎟ ⎥
⎣⎢ Lvsat Lvsat 2 ⎜⎝ Lvsat ⎠ ⎦⎥
Collecting terms in VD

Solutions Chapter 7
⎛ V ⎞
ID = ⎝ 2 ⎠
⎡ µlf VDS 2µlf RS Wvsat Cox' (VGS − VT ) ⎤
L ⎢1 + (1 − RS Wvsat Cox ) −
'
⎥
⎢⎣ vsat L vsat L ⎥⎦

Solutions Chapter 7
SOLUTIONS TO PROBLEMS FROM CHAPTER 8
8.1. Figure P8.1. shows the ID-VGS characteristic for an NMOS with VDS=50 mV.
It is known for this device that W=10µm, L=0.5µm, and tox=5 nm.
a) Find the threshold voltage
Since VDS=50 mV and L=0.5µm, the average field in the channel is 0.1
V/µm=1kV/cm and the device is not in the velocity saturation region. Thus
the expression (Equation (8.4))
WCox' µ0VDS ⎛ VDS ⎞
ID = ⎜ VGS − VT − ⎟
L (1 + θ (VGS − VT ) ) ⎝ 2 ⎠
can be used.
The threshold is extrapolated from the linear region, and using Equation
(8.8),
V 0.05
VT + DS = 0.5V = VT +
2 2
Therefore VT=0.475V.
b) Find µ0 , the electron channel mobility at threshold.

The slope of the linear region is about 160µA/(0.5V)=0.32mA/V. From
Equation (8.7), we have
dI D WCox' µ0VDS
=
dVGS L
Anderson & Anderson 1 February 17, 2007

Solutions Chapter 8
ε ox (3.9)(8.85 × 10−14 F / cm)
C =
'
= = 6.9 × 10 −7 F / cm2
−7
5 × 10 cm
ox
t ox
Solving for µ0 we have
dI
L D
dVGS 0.5 ⎛ 0.32 × 10−3 A / V ⎞ cm 2
µ0 = = ⎜ ⎟ = 460
WCox' VDS 10 ⎜⎝ 6.9 × 10−7 F / cm 2 ( 0.05V ) ⎟⎠ V ⋅s
8.2. A particular MOSFET process produces C'B =10-7F/cm2 and I0 = 4 × 10−20 A , and
a threshold voltage of VT=0.5V. For gate oxide thicknesses of 6.5 nm and 4 nm,
find n and S. Which device is better, and why?
For tox=6.5 nm,

C'ox =
ε ox
=
(
3.9 8.85× 10 −14 F / cm ) = 5.3 × 10−7 F / cm
−7
tox 6.5 × 10 cm
C 'B 10 −7
n = 1+ =1 + = 1.19
'
Cox 5.3 × 10 −7
kTn
S = 2.3 = 71mV / decade
q
For the 4 nm oxide, Cox' = 8.63 × 10−7 F / cm =, and
C'B 10−7
n =1+ ' =1+ = 1.12
Cox 8.63 × 10 −7
2.3kTn
S= = (2.3)(0.026)(1.12) = 67mV / decade
q
The 4 nm device is better. The swing S is smaller, resulting in a sharper
turn-on and thus permitting a reduced threshold voltage and a reduced
power supply voltage and reduced power consumption.
8.3. a) Find Wp/Wn needed to match IDsat for CMOS transistors

if µlfn = 500cm2 / V ⋅ s , µlfp=200 cm2/V·s, L=0.5 µm, and VGS-VT=2.6 V. Assume that
vsat = 4 × 106 cm / s s.
W
From Figure 8.6, we want p = 1.3 .
Wn
b) Find VDSsat for the NMOS and the PMOS.

Solutions Chapter 8
⎡ 1
⎤
v sat ⎢ ⎛ 2 µ (V − V ) ⎞ 2
⎥
VDSsat ( n ) = L ⎜1 +
lfn GS T
⎟ − 1⎥
µlfn nFET ⎢⎝ v sat LnFET ⎠
⎣⎢ ⎥⎦
⎡ 1
⎤
⎢ ⎛ ⎛ cm 2
⎞ ⎞ 2
⎥
4 ×10 6 cm
⎢ ⎜ 2 ⎜ 500 ⎟ ( 2.6V ) ⎟ ⎥
⋅
2 ( )
= s 0.5 × 10−4 cm ⎢⎜1 + ⎝ V s ⎠ ⎟ − 1⎥
cm ⎢⎜ 4 ×106 cm ( 0.5 ×10−4 cm ) ⎟ ⎥
500
⎢ ⎜⎜ ⎟⎟ ⎥
V ⋅s s
⎢⎣⎝ ⎠ ⎥⎦
= 1.09V
(or one could read it off Figure 7.31)
Similarly,
⎡ 1
⎤
vsat ⎢⎛ 2µlfp (VGS − VT ) ⎞ 2 ⎥
VDSsat ( p ) = L ⎜1 + ⎟⎟ − 1⎥
µlfp pFET ⎢⎜⎝ vsat LpFET ⎠
⎢⎣ ⎥⎦
⎡ 1
⎤
⎢ ⎛ ⎛ cm 2
⎞ ⎞ 2
⎥
⎟ ( 2.6V ) ⎟
cm
4 × 106 ⎢⎜ 2 ⎜ 200
V ⋅s⎠ ⎥
⎝
= ( )
s 0.5 × 10−4 cm ⎢⎜1 +
⎢⎜ 4 × 106 cm 0.5 × 10−4 cm ⎟
⎟ − 1⎥
200
cm 2
V ⋅s ⎢ ⎜
⎜ s
( ⎟
⎟
)
⎥
⎥
⎢⎣⎝ ⎠ ⎥⎦
= 1.49V
c) Adjust the length of the NFET to equalize the VDSsat’s. What should
Wp
the new be to keep the IDsat’s equal?
Wn
We want to set VDSsat(n)to 1.49V. From Figure 7.31, the new channel
length for the NFET should be about 1.2 µm. Then to set the IDsat’s equal,
I Dsat( n )
we set =1
IDsat ( p )
or
⎛ V ⎞
WnCox' µlfn ⎜ VGS − VT − DSsatn ⎟ VDSsatn
⎝ 2 ⎠
⎛ µ V ⎞
LnFET ⎜1 + lfn DSsatn ⎟
I Dsat ( n )
= ⎝ LnFET vsat ⎠
=1
I Dsat ( p ) ⎛ VDSsatp ⎞
W p Cox µlfp ⎜ VGS − VT −
'
⎟ VDsatp
⎝ 2 ⎠
⎛ µ V ⎞
LpFET ⎜1 + lfp DSsatp ⎟
⎜ LpFET vsat ⎟⎠
⎝
Since we have already adjusted the channel lengths to equalize the
VDSsat’s, we know that VDSsatn=VDSsatp and thus

Solutions Chapter 8
⎛ µlfpVDSsatp ⎞ ⎛ 200 ⋅ 1.49V ⎞
L pFET µlfn ⎜1 + ⎟⎟ ( 0.5 µ m )( 500 ) ⎜1 + ⎟
Wp ⎜
⎝ LpFET vsat ⎠= ⎝ (
⎜ 0.5 × 10 cm 4 × 106 cm / s
−4
) ⎟
⎠
=
Wn ⎛ µlfnVDSsatn ⎞ ⎛ 500 ⋅ 1.49V ⎞
LnFET µlfp ⎜1 + ⎟ (1.2 µ m )( 200 ) ⎜ 1 + ⎟
⎝ LnFET vsat ⎠ (
⎜ 1.2 × 10−4 cm 4 × 106 cm / s
⎝ ) ⎟
⎠
6.2 × 10−2
= = 1.01
6.12 × 10−2
Thus once this compensation is made the PFET will be the same width as
the NFET.
8.4. A CMOS inverter drives a load that consists of the gate of another FET.
The gate area is 0.2µm ×5µm. Find the dynamic power dissipation if the clock
frequency is 350MHz, the supply voltage is VDD=2.5V, and the oxide thickness
is 5 nm. If an MSI circuit (medium scale integrated circuit) has 1000 transistors,
what is the power consumption just due to switching? (Neglect feedthrough
current).
The load capacitance is
εA (3.9)(8.85 × 10 F / cm )(0.2 × 10 )(5 × 10 )
−14 −4 −4
CL = Cox A =
'
= −7 = 6.9 × 10 −15 F
tox 5 × 10 cm
Pdynamic = CL VDD f = (6.9 × 10 F )(2.5V ) (350 × 10 6 s−1 ) = 15 µW per transistor.
−15
2 2
For 1000 transistors the chip consumes 1000·15µW=15mW in dynamic

switching power.
Wp
8.5. A CMOS inverter has = 1.5 . The channel lengths are L=1µm, and
Wn
Wn=10µm. Find the propagation delay time if the load capacitance is 1pF. Let
tox=4 nm, VDD=2.5V, and VGS-VT=2V.
At a channel length of 1µm, this value of Wn/Wp equalized the saturation
currents, and IDsat(n)=IDsat(p).= Cox' Wvsat(VGS-VT-VDSsat)
For tox=4 nm, Cox' =8.63×10-7F/cm.
We can calculate either VDSsat to find IDsat since the IDsat’s have been
equalized:
⎡ 1
⎤
vsat ⎢ ⎛ 2 µ (V − V ) ⎞ 2
⎥
VDSsat ( n ) = ⎜1 + ⎟⎟ − 1⎥
lfn GS T
L
µlfn nFET ⎢⎜⎝ v sat LnFET ⎠
⎢⎣ ⎥⎦
⎡ 1
⎤
=
4 × 10 cm / s −4 ⎢
6
10 cm ⎢⎜1 +
⎛ 2 ( 500 cm 2
/ V ⋅ s ) ( 2V ) ⎞ 2
⎥
⎟ − 1⎥ = 1.16V
⎢⎝ ( 4 × 10 cm / s )(10 cm ) ⎠
500cm 2 / V ⋅ s ⎜ 6 −4
⎟
⎥
⎣ ⎦

Solutions Chapter 8
WCox' µlfn ⎡ 2
VDSsatn ⎤
I Dsat ( n ) = I Dsat ( p ) = ⎢(VGS − VT ) VDSsatn − ⎥
⎡ µlfnVDSsatn ⎤ ⎣ 2 ⎦
⎢ LnFET + ⎥
⎣ vsat ⎦
cm 2
10 × 10 ( 8.63 × 10 ) 500
−4 −7
V ⋅ s ⎡( 2 )(1.16 ) − 1.16 ⎤ = 2.9mA

2
= ⎢
⎡ −4 500 ⋅ 1.16 ⎤ ⎣ 2 ⎥⎦
⎢⎣10 +
4 × 106 ⎥⎦
Thus the propagation delay time is
1 ⎡C V C V ⎤ 1 ⎡ (10 F ) ( 2.5V ) ⎤
−12
td = ⎢ L DD + L DD ⎥ = ⎢ 2 × ⎥ = 0.5ns
2 ⎢⎣ 2 I Dsatn 2 I Dsatp ⎥⎦ 2 ⎢
⎣ 2 ( 2.5 × 10 −3
A ) ⎥⎦
8.6. For the transistor whose ID-VDS characteristics are shown in Figure P8.2,
find the small signal parameters:
a) gd for VGS=1.0V , and VDS<VDSsat
∂I D
We have g d = . Finding the slope we have 360µA/0.3V=0.0012 S
∂VDS
b) gd for VDS>VDSsat, VGS=1.1V
In this region the slope is zero, and so is the output conductance.
c) gm for VDS=0.2V, VGS=1.1V

Solutions Chapter 8
∂I D
gm = . We pick VGS=1.1 V, and see that if VGS varies above and
∂VGS
below this for a change of ∆VGS=0.2 V, that ID changes by about 120µA. At
this operating point gm=(120×10-6/0.2)=6×10-4 S
d) gmsat for VDS>VDSsat, VG=1.1V
∆I D (430 − 220) µ A
g msat = = = 1.05 × 10−3 S
∆VGS 0.2V
8.7. A CMOS circuit must operate at 500 MHz. What is the maximum channel
length that can be tolerated, neglecting short-channel effects?
From Figure 8.16, the maximum channel length is about 3.3 µm (we use
the PFET line because the PFETs are slower, assuming equal channel
lengths, so they represent the worst case).
8.8 Practice using circuit models. Suppose a FET was found to have a finite
input resistance, Rin. How would you include this in the model of Figure 8.13
(a)? How would you include the channel resistance?
The input is between the gate and source, so we need to add a resistor
from gate to source. The channel resistance extends between the source
and the drain.
8.9. A designer wishes to design a for a 0.3µm process (meaning that the FETs
have L=0.3 µm). Furthermore, the current fabrication line is limited to tox=4
nm. What should the junction depth xj, be such that the short-channel effects
can be neglected? Let VDD=3.3V, and let N'A in the channel be 1017 cm-3. If the
channel doping is reduced, what is the effect on the allowable junction depth?
If the supply voltage is reduced, what is the effect?

Solutions Chapter 8
Using the rule of thumb presented in Equation (8.40), we have
1/ 3
Lmin = 0.9 ⎡⎣tox x j ( wS + wD ) 2 ⎤⎦ = 0.9γ 1/ 3
where tox is in nm and xj, wS , and wD are in µm.
3
⎛ 0.3 ⎞
γ <⎜ ⎟ = 0.037 = tox x j ( wS + wD )
2
⎝ 0.9 ⎠
The depletion region width at the source end is that of an n+p junction
under zero bias:
⎡ (3.1 × 1019 )(1017 )⎤
1
⎡ 2εVbi ⎤ 2 kT ⎛ NV N A ' ⎞ ⎢ ⎥ = 0.98V
wS = ⎢ ⎥ and Vbi = q ln ⎜⎝ n2 ⎟⎠ = 0.026 ln ⎢
⎣ ( ) ⎥⎦
10 2
⎣ A ⎦
qN ' i 1.08 × 10
so
2 (11.8 ) ( 8.85 × 10−14 ) ( 0.98 )
1
⎡ 2εVbi ⎤ 2
wS = ⎢ ⎥ = = 0.11µ m
⎣ qN A ' ⎦ 1.6 ×10−19 (1017 )
At the drain end, the junction is the same but it has a junction voltage
Vj=Vbi-Va=0.98 + 3.3=4.28V (the junction is reverse biased). Thus
2 (11.8 ) ( 8.85 × 10−14 ) ( 4.28 )
1
⎡ 2εV j ⎤ 2
wD = ⎢ ⎥ = = 0.24µ m
⎣ qN A ' ⎦ 1.6 × 10−19 (1017 )
γ 0.037
then x j = = = 0.075µ m
tox ( wS + wD ) 4(0.11 + 0.24) 2
2
If the doping in the channel is reduced, the depletion regions get wider
and the junction depth needs to be reduced.
If the supply voltage is reduced, then wD will be smaller and the junction
can be a little deeper.
8.10. SOI devices are more rad-hard (insensitive to radiation environments

such as outer space or nuclear reactors) than conventional CMOS as discussed
in the text. Examples of ionizing radiation are bombardment with high kinetic
energy neutrons, electrons, or protons, or gamma rays (which are photons with
energies of a few MeV). The high-energy particles can disrupt the silicon
crystal in conventional CMOS, while in SOI the amount of crystalline Si is
much smaller. Furthermore, charges that accumulate in the buried SiO2 layer
from the bombardment will generally not affect device operation. Explain why
SOI devices will be also be less affected by the absorption of gamma rays.
Gamma radiation, when absorbed, will produce electron hole pairs. In the
semiconductor substrate of conventional CMOS, these pairs can be
separated and migrate, producing soft errors. In SOI, there is much less

Solutions Chapter 8
silicon and a great deal more insulator. Electron-holes pairs cannot
migrate in the insulator and thus the charges do not separate.
8.11. (a) Draw the energy band diagram for the HFET, perpendicular to the
gate, for the case when the transistor is strongly on . Indicate the polarity of
the voltage applied to the gate.
channel enhanced
EC
ED
∆EC
Ef
EA
Efm
EV
∆EV
semi-
p-GaAs insulating
GaAs
metal n+ AlGaAs
undoped AlGaAs
The gate voltage is made more positive.
b) Repeat for the depletion case (no channel). Now what is the polarity
of the gate voltage?

Solutions Chapter 8
channel depleted
EC
∆EC ED
Efm
Ef
EA
EV
∆EV
metal n+ AlGaAs
semi-
undoped AlGaAs p-GaAs insulating
GaAs
In the depletion case the gate is made more negative.
Note that since the undoped layer is only slightly conductive, some of the
applied voltage is dropped across this layer. The size of ∆EC is determined
by the materials and not the voltages, so these steps remain the same.
c) Draw the energy band diagram along the channel for the equilibrium case
and when the channel is conducting.
We observe that the source, drain, and channel are all GaAs- no
heterojunction in this direction. The energy band diagram in this plane
looks the same as for a conventional MOSFET:

Solutions Chapter 8
8.12. There are two primary reasons that GaAs-based HFETs are faster than
silicon MOSFETs. One reason, given in the text, was that the electrons are
traveling in a lightly doped channel. Explain how this increases their velocity.
What do you think is the other reason?
Since there are few impurities in the lightly doped channel, there is little
ionized impurity scattering. The other reason for the high speed is the
higher electron (low field) mobility in GaAs compared to Si.
8.13. The channel in the HFET is often referred to as supporting a 2-

dimensional electron gas. The electron gas part of this name refers to the sea
of electrons that exists in the channel when the transistor is conducting. Why
is it called 2-dimensional?
The channel is physically a thin sheet (i.e. 2-D) along the interface
between the AlGaAs and the GaAs. The energy band diagrams only show
one dimension of this.
8.14. Draw the energy band diagram at the source and the channel end of the
JFET under saturation (e.g., the same type of figure as Figure 8.29 for the
MESFET. Point out the similarities and differences in operation of the two
different types of devices.

Solutions Chapter 8
The two are qualitatively similar. In both, a voltage on the gate increases
or decreases the width of a depletion region. The difference is that in the
JFET, the depletion region is caused by a reverse-biased pn junction. In
the MESFET, it is the depletion region cause by a reverse-biased Schottky
diode depletion region.
8.15. An n-channel JFET has a channel doping of 1016 cm-3. Let the thickness of
the n layer (a) be 1 µm, the length of the channel be 4 µm, and the width be 40
µm.
a) What is the threshold voltage?
Since the gate is degenerately doped, we can assume that the depletion
region is entirely on the n side (channel). We find that the depletion width
is equal to the channel width when w = 1µm , or Vj=Vbi-Va=6.5V from Figure
5.12. Since from Figure 5.10 the built-in voltage is 0.92 V, the applied
voltage at threshold is VT=Va=Vbi-Vj=0.92-6.3=-5.4 volts.

Solutions Chapter 8
b) If the gate voltage is -2V, at what drain voltage does the current
saturate?
From Equation (8.50), VDSsat=VGS-VT= -2-(-5.3)=3.3V.
c) What is the saturation current under these conditions, using the

simple long-channel model?
From Equation 8.51,
⎧ ⎡ 3
⎤⎫
qW µlf N D' a ⎪ 2 ⎢ ⎛ V − V ⎞ 2
⎥⎪
= ⎨VGS − VT − (Vbi − VT ) ⎢1 − ⎜ ⎟ ⎥⎬
bi GS
I Dsat
L ⎪ 3 ⎝ Vbi − VT ⎠ ⎥ ⎪
⎩ ⎣⎢ ⎦⎭
From Figure 3.4 the low-field mobility at this doping (the electrons are
majority carriers in this case) is about 1100 cm2/V·s.
⎛ cm 2 ⎞
(1.6 × 10−19 C)(40µm )⎜⎝1100 V ⋅ s ⎟⎠ (1016 cm−3 )(10 −4 cm)
I Dsat =
4 µm
⎧ ⎡ 3 ⎤⎫
⎪ ⎢ ⎛ 0.92 − (−2 ) ⎞ 2 ⎥⎪
× ⎨−2 − (−5.3) − [0.92 − (−5.3)]⎢1− ⎜
2
⎟ ⎥⎬
⎪ 3 ⎢ ⎝ 0.92 − (−5.3)⎠ ⎥⎪
⎩ ⎣ ⎦⎭
−3
= 1.76 ×10 (0.42)
= 0.85mA

Solutions Chapter 8
SOLUTIONS CHAPTER 9
9.1. For each of the transistors in Figure P9.1, indicate the mode of operation
(forward active, cutoff, saturation, etc.)
9V 9V
4.9 V 5.0V
4.8V 0.2 V
4.1V
(a) (b)
9V
0.7 V 0.7 V
-5 V 0.5V
(c) (d)
-9 V
a) This is an npn transistor. The collector voltage is higher than the base
voltage so the CB junction is reverse biased. The base voltage is higher
than the emitter voltage so that junction is forward biased. Thus this
transistor operates in forward active mode.
b) Also npn. The EB junction is forward biased by 0.2 V, which is not

enough to turn it on. The CB junction is reverse biased, so this transistor is
in cutoff.
c) This one is pnp. The EB junction is forward biased and the CB junction
is reverse biased, so this one is in forward active mode.
d) Also pnp; both junctions forward biased- this one is in saturation.

Solutions Chapter 9
9.2. For each area in Figure P9.2, identify how the region should be doped to
make a good pnp transistor (n+, n, intrinsic, p, p+). and indicate the reason(s).
Indicate where the actual pnp transistor occurs.
E B C
b a e
f
c
d
transistor
a) p+. The emitter should be heavily doped to maximize β and also to

make a good contact with the metal.
b) n-type base should be more lightly doped than emitter
c) p-type collector should be lightly doped.
d) this layer should be p+ to make it highly conductive. It transfers the
collected holes to the collector contact.
e) n+ to ensure a good contact with the base contact.
f) p+ to conduct collected holes easily and to provide good contact with the
collector contact.
9.3. Draw the energy band diagram for a pnp transistor at equilibrium and
under forward active bias.
The drawing should reflect the relative dopings as well:

Solutions Chapter 9
p+ n p p+
emitter base collector collector
-W E WBM x
EC E
C
E ,E E ,E V
f V f
equilibrium
E
C
Fn In
E ,E V
f
Fp
Ip active mode bias
9.4. From Equation (9.13), α=γαTM, where M is the carrier multiplication factor
in the base collector junction. For small base-collector voltage, M=1 and α=γαT
α
and β = . Show that for avalanche breakdown,
1−α
1
M = 1+ .
β
Solution:
α αM
β= = 0
1 − α 1 − α0M
For breakdown, β = ∞ or α M = 1
Solutions Chapter 9
β βM
But α = and α M = =1
1+ β 1+ β M
Solving for M,
1
M = 1+
β
9.5. Using the prototype model, (i.e. ignore the apparent band-gap narrowing
and its effect on injection efficiency) find α and β for an npn BJT with
N DE = 1019 cm −3 , N AB = 2 × 1017 cm −3 , and N DC = 1017 cm−3 . Indicate clearly all your
steps. Let VBE=0.8V VCB=2.0 V. The metallurgical widths are WEM = 0.2 µm and
WBM = 0.2 µm.
We can find the minority carrier diffusion constants from Figure 3.11:
DpE=3.8 cm2/s and DnB=15 cm2/s
The built-in voltages for the two junctions are from Equation (5.13):
⎡ ( 3.1 × 1019 )( 2 × 1017 ) ⎤
kT ⎡ NV N AB ⎤ ⎢ ⎥ = 1.00V (one-sided
Vbi ( EB ) = ln ⎢ ⎥ = 0.026 ⋅ ln ⎢
q ⎣ ni2 ⎦
⎣ (1.08 × 10 ) ⎥⎦
10 2
junction) and
⎡ (1017 )( 2 × 1017 ) ⎤
kT ⎡ N DC N AB ⎤
Vbi ( CB ) = ln ⎢ ⎥ = 0.026 ln ⎢⎢ ⎥ = 0.85V
⎣ (1.08 × 10 ) ⎦
q ⎣ ni2 ⎦ 10 2 ⎥
If the forward bias on the EB junction is 0.8V, then the junction voltage is
VJ(EB)=1.00-0.8=0.2V, and the junction on the base side of the emitter
junction extends, from Equation (5.39)
1
⎡ 2ε V j ⎤ 2 ⎡ 2 × 11.8 × 8.85 × 10−14 F / cm × 0.2V ⎤
1/ 2
wp ( EB ) =⎢ ⎥ =⎢ ⎥ 3.6 × 10−6 cm = 0.036µ m

⎣ qN AB ⎦ ⎣ 1.6 × 10−19 × 2 × 1017 cm −3 ⎦
For a CB junction with 2V reverse bias, the transition region in the base is,
from Equation (5.36)

Solutions Chapter 9
1
⎡ ⎤ 2
⎢ ⎥
2ε V j ( BC )
wp ( BC ) =⎢ ⎥
⎢ ⎛ N ⎞⎥
⎢ qN AB ⎜1 + AB ⎟⎥
⎢⎣ ⎝ N DC ⎠ ⎥⎦
1
⎡ ⎤2
=⎢
( )
⎢ 2 (11.8 ) 8.85 × 10−12 F / m ( 0.85 − (−2) ) V ⎥
⎥
⎢ −3 ⎛ 2 × 1017 cm −3 ⎞⎥
⎢ (1.6 × 10 C )( 2 × 10 m ) ⎜1 +
−19
⎟⎥
23
⎣⎢ ⎝ 1017 cm −3 ⎠ ⎦⎥
= 0.079 µ m
The remaining base width is then
WB = WBM − wp ( EB ) − wp ( BC ) = 0.2 − 0.036 − 0.079 = 0.085µ m
We take the emitter width to be equal to the emitter metallurgical width, or
WE=WEM.
From Figure (3.23), LnB=85µm for a doping of NAB=2x1017cm.
To find α, we must find γ and αT:
1 1
γ ≈ = = 0.998
N AB D pE WB 2 × 10 3.8 0.085
17
1+ 1+
N DE DnE WE 1019 15 0.2
WB2 (0.085µ m) 2
and αT = 1 − = 1 − = 0.9999995 ≈ 1
2 ( 85µ m )
2
2 L2nB
In this case α≈γ=0.998
α 0.998
We find β = = = 500 .
1 − α 1 − 0.998
9.6. For the device of Example 9.1, estimate the emitter-base junction width, the
width appearing on the emitter side and that appearing on the base side.
Assume that the junction is forward biased with V j = Vbi − Va = 0.2V . Draw the
emitter and base to scale and indicate the locations of the edges of the
transition region.
From Example 9.1. we have N DE = 2 × 10 and N AB

' 18
'
= 1016 cm-3. The simple
model neglects band gap narrowing. From Figure 5.10, we look up the
built-in voltage for this one-sided junction, and find
Vbi=0.91 V.
The total junction width is given by

Solutions Chapter 9
1
⎡ 2ε ( N A' + N D' ) (Vbi − Va ) ⎤ 2
w = wn + wp = ⎢ ⎥
⎢⎣ qN A' N D' ⎥⎦
1
⎡ 2 (11.8 ) ( 8.85 × 10−14 F / cm )( 2 × 1018 cm −3 + 1016 cm −3 ) ( 0.91 − 0.2V ) ⎤ 2
=⎢ ⎥
⎢⎣ (1.6 × 10−19 C )( 2 × 1018 cm−3 )(1016 cm−3 ) ⎥⎦
= 0.31µ m
wn N A' 1016
We also have that = ' = = 0.005
wp N D 2 × 1018
Thus wn = 0.005wp , which we combine with
wn + wp = 0.31µ m
to obtain
0.005wp + wp = 0.31µ m
0.31
wp = = 0.308µ m
1 + 0.005
and
wn = 0.005wp = ( 0.005 ) × ( 0.308 ) = 0.0015µ m
x=0 Wn Wp
EB junction CB junction
9.7. Derive the simple model expression for β for a pnp transistor with a non-
degenerate emitter (the pnp equivalent of Equation (9.42).
For a pnp transistor, the injection efficiency would be given by
1
γ= I . Following the derivation in the text for an npn, we would have
1+ nE
I pE
for the forward injection of holes:
dpB qAE DpB[p B (0 + ) − 0 ] qAE DpB pB (0+ )
I pE = − qAE DpB = =
dx x = 0 + WB WB
and for the back injection
dnE qAE DnE [nE (0 − ) − 0] qAE DnE nE (0− )
InE = −qAE DnE = =
dx x =0 − WE WE
and
Solutions Chapter 9
1 1
γ= − = −
WB qAE DnEnE (0 ) n (0 ) D W
1+ +
1 + E + nE B
qAE DpB pB (0 ) WE pB (0 ) DpB WE
For the excess carriers we have
qV
+ ni2 kTa
pB (0 ) = ' e
N DB
qVa
n2
nE (0 − ) = i' e kT
N AE
nE (0− ) N 'DB
Thus =
pB (0 + ) N 'AE
and
1
γ= '
N DnE WB
1+ DB
N 'AE DpB WE
For αT, we write

I
α T = 1 − rec
IpE
The recombination current is proportional to the number of carriers, so
taking an average value of excess holes in the base we obtain
qA p W
Irec = E B B . Since the excess hole concentration varies linearly from
τn
pB (0)
pB(0+) at x=0 to pB(x=WB)=0, the average is pB = . Thus
2
qAE pB (0)WB
2τ p
αT = 1 −
I pE
qAE DpB pB (0)
But, I pE = from before, so
WB
qAE pB (0)WB
2τ WB2
α T = 1 − qA D pp (0) = 1 −
E pB B 2τ Dp pB
WB
Since τpDp=Lp2, we have
WB2
αT = 1 − 2
2LpB
We saw that in the npn this was so close to unity as to be negligible. The
diffusion length for p-type material is smaller than for n-type, but only by a
factor of 2 or so. Thus we can still say that αT ≈1 and
α γ I pC pB (0 + ) DpB WE N 'AE DpB WE
β= ≈ = = = '
InB nE (0 ) DnE WB N DB DnE WB
−
1 − α 1− γ
Solutions Chapter 9
9.8. A prototype npn transistor has its emitter doped to ND=5×1019cm-3. The
base is doped with NA=2×1017 cm-3. Find ∆Eg* accounting for narrowing of the
emitter conduction band. Find β if WE=0.12 µm, and WB=0.07 µm.
From Figure (2.25), the apparent band gap narrowing in the emitter for n-
type material doped to 5x1019 cm-3 is 0.08 eV.

∆E * −0.08
N D W − g 2.86 × 1019 11 0.12 0.026
β = C nB E e kT = e = ( 327 )( 0.046 ) = 15
N AB D pE WB 5 × 1017 3.3 0.07
where the minority D’s come from Figure 3.11.
Note that if the base band-gap narrowing is also taken into account, from
Figure 2.25, the base apparent band-gap narrowing is 0.03 V, and
∆EgBE
*
N D W − 2.86 × 1019 11 0.12 − (0.08 − 0.03)

β = C nB E e kT
= e 0.026
= ( 327 )( 0.146 ) = 48
N AB D pE WB 5 × 1017 3.3 0.07
9.9. In Example 9.2, we neglected the band gap narrowing in the base. Find the
current gain β, this time taking ∆EgB
*
into account. Draw the equilibrium
energy band diagram, and indicate the true band edges and apparent band
edges in both the emitter and the base. What is the effective barrier height for
electrons? Holes?
1
From Equation (9.40), we have γ = ∆E *
'
N AE gBE
D pE WB
1+ e kT
NC DnB WE
γ
where ∆EgBE
*
= EgB
*
− EbE
*
. Since β ≅ , following Example 9.3 we find
1− γ
β
γ= .
1+ β
∆EgBE *
N' D W 1
Letting Z = AB pE B e kT , we can write γ = , and
N C DnB WE 1+ Z
1 1
β ≅ 1+ Z = 1+ Z =
1
1 Z Z
1−
1+ Z 1+ Z
∆EgBE
*
N DnB WE −
and β ≅ 'C e kT
.
N AB D pE WB
Solutions Chapter 9
barrier to qV
BE
electrons E
g
EC
∆E*
gE Ef
∆EgB
* E
g
E*
gE Eg qVBE barrier
to holes
EV
E B C
9.10. Suppose that a degenerate Si layer of ND=1020 cm-3 and NA=0 is epitaxially
deposited onto the emitter of the transistor of Section 9.4. Explain how such a layer
could increase β slightly. Both emitter layers together are still much shorter than
the diffusion length of either electrons or holes in the emitter. [Hint: There is a small
amount of bandgap narrowing due to the acceptors in the base (and the
compensated acceptors in the emitter)].
We draw the energy band diagram to show the effect of the surface layer:
∆E*g
(due to nB
donors)
E
C
p
E
∆E*g
(due to EV
acceptors)
We see that because there are substantial acceptors in the original emitter
and base layers, there is a small amount of band gap narrowing- affecting
the position of the valence band edge. This effect will not be present in the
surface layer, creating a small energy barrier for holes in the emitter. This
will tend to reduce the back injection of holes from the base by a small
amount, with an attendant increase in β.
9.11. Working engineers must be able to teach themselves new things

throughout their careers to keep up to date. Research and learn about
secondary ion mass spectroscopy (SIMS), and write up a short explanation of
what it does and how it works.
(left for student)
Solutions Chapter 9
9.12. If an pnp transistor is made with the same dimensions and doping
concentrations as an npn, explain why the pnp will always have a lower β.
N 'DE DnB WB
For an npn transistor β =
N'AB DpE WE
N 'AE DpB WB
For a pnp transistor β =
N 'EB DnE WE
If the transistors are made identically, the first ratio and the third ratio
would be the same The diffusion constants, however, will be different.
Since hole diffuse more poorly than electrons, the second ratio will always
be smaller in the pnp and thus its β is lower.
9.13. If a pnp transistor is considered to have all step junctions, with the
properties in the table below, find:
'
N AE =5×1019 cm-3 '
N DB =4×1017 cm-3 '
N AC =1017 cm-3
WEM=0.13 µm WBM=0.20 µm
VEB=0.8V (forward bias) VCB=2 V (reverse bias)
a) the minority carrier lifetimes

From Figure 3.21, we look these up to find
τnE=20 ns
τpB=0.7 µs
τnC=13 µs
b) the corresponding diffusion coefficients and diffusion lengths

From Figures 3.11 and 3.23, we obtain
DnE=6.1 cm2/s; LnE=0.8µm
DpB=10 cm2/s; LpB=30 µm
DnC=23 cm2/s, LnC=110 µm
c) the built-in voltage of the collector-base junction

⎛ 17 ⎞
kT ⎛ N D' N A' ⎞ ⎜ 10 ⋅ 4 × 1017 ⎟
Vbi (CB ) = ln ⎜ ⎟ = ( 0.026V ) ln = 0.87V
q ⎝ ni2 ⎠ ⎜ (1.08 × 1010 )2 ⎟
⎝ ⎠
d) The built-in voltage of the emitter-base junction (neglecting band

gap narrowing)
Since this is a one-sided junction, we find this from Figure 5.10 for a lightly
doped side at 4x1017, and it is about 1.02V
e) WB
Solutions Chapter 9
The emitter-base junction has a junction voltage Vj=Vbi-Va=1.02-
0.8=0.22V. We can find
2εV j 2 (11.8 ) ( 8.85 × 10−14 ) ( 0.22 )
wn ( EB ) = = = 0.027 µ m µm
qN D' 1.6 × 10−19 ( 4 × 1017 )
For the collector- base junction,
1 1
⎡ ⎤ 2
⎡ ⎤2
⎢ ⎥
wn ( CB ) = ⎢ 2ε (VbiCB − VaCB ) ⎥
= ⎢ ( )
⎢ 2 (11.8 ) 8.85 × 10−14 F / cm ( 0.87 − (−2 ) V ⎥
⎥
⎢ ⎛ '
⎞⎥ ⎢ ⎛ 17
⎞ ⎥
⎢ (1.6 × 10 C )( 4 × 10 cm ) ⎜1 +
N AC −19 −3 10
⎢ qN DB ' ⎜1 + ' ⎟ ⎥ 17
17 ⎟ ⎥
⎢⎣ ⎝ N DB ⎠ ⎥⎦ ⎣⎢ ⎝ 4 × 10 ⎠ ⎦⎥
= 0.087 µ m
The final base width is
WB = WBM − wnEB − wnCB = 0.20 − 0.027 − 0.087 = 0.086µ m
f) Neglecting band gap narrowing, find γ
1 1
γ= =
N '
D W 4 × 10 6.1 0.08
17
1 + DB ⋅ nE ⋅ B 1 + ⋅ ⋅
'
N AE D pB WE 5 × 1019 10 0.13
= 0.997
g) Find αT
2
W2 1 ⎛ 0.086 µ m ⎞
αT = 1 − B2 = 1 − ⎜ ⎟ = 0.999996 ≈ 1
2 L pB 2 ⎝ 30 µ m ⎠
h) Assuming M=1, find α

Here αT is also unity, so α=MαTγ=γ=0.997
i) Find β.
α 0.997
β= = = 332
1 − α 1 − 0.997
9.14. Repeat problem 13, but this time account for band gap narrowing in the
emitter.
To account for band gap narrowing, we use the equation
∆E *
NV D pB WE − kTg
β= ' e
N DB DnE WB

Solutions Chapter 9
The band gap narrowing is ∆Eg*=0.113eV from Figure 2.25. Thus a more
realistic β for this transistor is:
3.1 × 1019 10 0.13 − 0.026

0.113
β= ⋅ ⋅ e = (206)(0.013) ≈ 2.7
4 × 1017 6.1 0.08
9.15. Find β for an npn transistor with a degenerately doped emitter at

'
N DE = 1020 cm-3, and N AB
'
= 5 × 1017 cm-3. Let WE=0.15µm and WB=0.1 µm.
∆E gBE
*
N D W −
Equation (9.43) is β = 'C nB E e kT
. From Figure 3.11, DnB=13 and
N AB D pE WB
DpE=3.6 cm2/s.
The band gap narrowing in the emitter is 93meV and in the base, 32 mV
from Figure 2.25.
− 0.032)
2.86 × 1019 13 0.15 − (0.0930.026
Thus β = e = 30
5 × 1017 3.6 0.1
9.16. Consider the graded base transistor whose SIMS profile is shown in
Figure P9.3.

Solutions Chapter 9
a) Find η.
NAB(0+)=2x1017 and NAB(WB)=1016. From Equation (9.53),
⎛ 2 × 1017 ⎞
η = ln ⎜ 16 ⎟ = 3.0 .
⎝ 10 ⎠
b) Find the built-in electric field in the base.
kT η 3.0
From Equation (9.54): E = = (0.026V ) = 650V / cm
q WB 1.2 × 10−4 cm
c) Estimate β.
β (η ) η
From Equation (9.60), = . We therefore first find β(0). The
β (0) 1 − e −η
emitter is degenerately doped, with average N DE'
≈ 1019 cm-3. From Figure
3.11, we find for the emitter that DpE=13 cm2/s, and using an average
doping in the base of 5×1016 we obtain DnB≈30cm2/s. From Figure P9.3,
we find WE=0.20µm, and WB=0.4-0.23=0.17µm. Thus

Solutions Chapter 9
∆E *
N C DnB WE − kTg 2.86 × 1019 30 0.20 − 0.026
0.05
β (0) = ' e = e = (1.55 × 103 ) ( 0.14 ) = 217
N AB D pE WB 5 × 10 16
13 0.17
and
⎡ η ⎤ ⎛ 3 ⎞
β (η = 3.0 ) = β ( 0 ) ⎢ −η ⎥
= 217 ⎜ −3 ⎟ = 685
⎣1 − e ⎦ ⎝1− e ⎠
9.17. A graded based npn transistor has its emitter degenerately doped to 1019 cm-3,
its base has a doping concentration of 5×1017 cm-3 at the emitter edge and a grading
parameter η=3. The emitter width is WE=0.2µm and base width is WB=0.15 µm.
Find β.
The β for a prototype transistor is given by Equation (9.43):
∆E*
N D W − g
β = 'C nB E e kT . We find DnB=9 cm/s and DpE=4 cm2/s from Figure
N AB D pE WB
3.11. From Figure 2.25, the band gap narrowing ∆Eg*=0.05eV, yielding for
the prototype transistor
∆E *
N C DnB WE − kTg 2.86 × 1019 9 (0.2) ⎛ −0.05 ⎞
β= ' e = exp ⎜ ⎟ = 25
N AB D pE WB 5 × 10 17
4 (0.15) ⎝ 0.026 ⎠
Using Equation (9.60), we take the grading in the base into account and
obtain
η ⎡ 3 ⎤
β (η ) = β ( 0 ) = 25 ⎢ = 25 [3.2] = 79
⎣1 − e ⎥⎦
−η −3
1− e
9.18. An npn BJT is operating in the forward active region. Assuming the
default SPICE values IS=10-16A, βF=100, find the terminal currents IC, IB, and IE
for VBE=0.7V.
⎛ 1 ⎞
qVBE
⎛ 1 ⎞ −16 0.026
0.7
From Example 9.5, I E = ⎜ 1 + ⎟ ISe = ⎜1 + ⎟ (10 ) e = 50 µ A
kT
⎝ βF ⎠ ⎝ 100 ⎠
qVBE 0.7
IC = I S e kT
= 10−16 e 0.026 = 49 µ A
IC 49 × 10−6
IB = = = 490nA
βF 100
9.19. Typical forward β’s are on the order of 100, while typical reverse β’s are
on the order of 0.1. Find the corresponding αF and αR.
From Equations (9.66),

Solutions Chapter 9
βF 100
αF = = = 0.99
β F + 1 100 + 1
β 0.1
αR = R = = 0.09
β R + 1 0.1 + 1
9.20. Fill in the missing steps to verify Equation (9.75).

⎛ y⎞
We take I ( y ) = I B ⎜1 − ⎟ , and substitute into Equation (9.71) to find VB(y):
⎝ L⎠
dVB ( y ) R R ⎛ y⎞
= − I ( y ) = − I B ⎜1 − ⎟
dy h h ⎝ L⎠
R ⎛ y⎞ R ⎡ y2 ⎤
VB ( y ) = − I B ∫ ⎜1 − ⎟ dy = − I B ⎢ y − ⎥ in agreement with Equation
h ⎝ L⎠ h ⎣ 2L ⎦
(9.74).
Substituting into Equation (9.70):
L⎡ R ⎛ y 2 ⎞⎤ ⎡ R ⎛ y ⎞⎤
∫
L dVB ( y ) − − − I B ⎜ 1 − ⎟ ⎥ dy
∫0 VB ( y )
dy
dy 0
⎢
⎣ h ⎝
I B ⎜ y ⎟ ⎥ ⎢
2L ⎠⎦ ⎣ h ⎝ L ⎠⎦
VB = =
L dV ( y ) L⎡R ⎛ y ⎞⎤
∫0 dy B
dy
∫0 ⎢⎣ h I B ⎜⎝1 − L ⎟⎠⎥⎦ dy
R L⎛ y2 ⎞ ⎛ y⎞ R L⎛ y 2 y 2 y3 ⎞
IB ∫ ⎜ y −
h ∫0 ⎝
⎟⎜ 1 − ⎟ dy I B ⎜ y − − + ⎟ dy
h 0
⎝ 2L ⎠ ⎝ L ⎠ 2 L L 2 L2 ⎠
= =
L⎛ y⎞ L⎛ y⎞
∫0 ⎝ L ⎠
⎜ 1 − ⎟ dy ∫0 ⎜⎝1 − L ⎟⎠ dy
L
R ⎛ y 2 y3 y3 y 4 ⎞ R ⎛ L2 L3 L3 L4 ⎞
IB ⎜ − − + ⎟ IB ⎜ − − + ⎟
h ⎝ 2 6 L 3L 8L2 ⎠ 0 h ⎝ 2 6 L 3L 8L2 ⎠
= =
⎛ y2 ⎞
L
⎛ L2 ⎞
⎜ − ⎟
⎜y−
L
⎟ ⎝ 2L ⎠
⎝ 2L ⎠ 0
R ⎛ 24 L2 8L2 16 L2 6 L2 ⎞ R ⎛ 6 L2 ⎞
IB ⎜ − − + ⎟ I B⎜ ⎟
h ⎝ 48 48 48 48 ⎠ h ⎝ 48 ⎠
= =
⎛ L⎞ ⎛L⎞
⎜L− ⎟ ⎜ ⎟
⎝ 2⎠ ⎝2⎠
R 12 L R ⎛L⎞
= IB = IB ⎜ ⎟
h 48 h ⎝4⎠
V RL
Now we put this result into RB = B = which is Equation (9.75).
IB 4h
9.21. We previously chose an equivalent resistance in the base using the

average voltage drop across the base. Here we will used the power dissipated
Solutions Chapter 9
in a resistor, P=I2R. We define a new equivalent resistance based on the power
PB that is dissipated by the base current flowing through a lumped resistance
P
RB = B2 where PB is the power dissipated in the base. Find an expression for
IB
RB. Assume IB to be small enough that Equation (9.73) is valid.
We are considering the power dissipated in an incremental length of the

⎛ y⎞
base region dy. We begin with letting I ( y ) = I B ⎜1 − ⎟ . We can write the
⎝ L⎠
power as
R
dPB ( y ) = dy ⋅ I 2 ( y ) . Integrating to get PB(y), we have
h
L
R 2 ⎛ 2 y y2 ⎞ R 2⎡ y 2 y3 ⎤
2
R 2 ⎛ y⎞
L L
PB ( y ) = I B ∫ ⎜ 1 − ⎟ dy = I B ∫ ⎜1 − + 2 ⎟ dy = IB ⎢ y − + 2 ⎥
h 0⎝ L⎠ h 0⎝ L L ⎠ h ⎣ L 3L ⎦ 0
R 2⎡ L⎤
= IB ⎢L − L + ⎥
h ⎣ 3⎦
R 2L
= IB
h 3
Then
R 2L
IB
PB h 3 =RL
RB = 2 = 2
IB IB 3h
9.22. For the transistor whose IC-VCE curves are shown in Figure P9.4, find the
Early voltage.

Solutions Chapter 9
We find the equation for a straight line approximation to the I-V
characteristics. For I B = 40 µ A, the slope is:
5.4 − 3.7
m= = 0.34mA / V
5−0
The y-intercept by inspection is 3.7x10-3 A. Solving the equation for a line
to find the x-coordinate at y=0, we find
0 = 0.34 × 10−3 ( x) + 3.7 × 10−3
x = −11V
−3
9.23. An npn transistor has N DE = 10 , N AB = 5 × 10 , and N DC = 10 cm . For
' 20 ' 17 ' 16
VBE=0.75 V, how small can the base width WBM be to keep punch-through
above 12 V?
Since band-gap narrowing has little effect on junction width, we ignore it.
qVbi EB = Eg − δ pB
⎛ NV ⎞ ⎛ 3.1 × 1019 ⎞
and δ p = E f − EV = kT ln ⎜ ⎟ = 0.026ln ⎜⎝ ⎟ = 0.11eV
⎝ NA ⎠ 5 × 1017 ⎠
Thus VbiEB = 1.12 − 0.11 = 0.91V
The junction voltage is therefore VJEB=VbiEB-VEB=0.91-0.75=0.16V. From
Figure 5.12, the junction width (which appears almost entirely in the base)
is about 0.025.µm.
The built-in voltage at the CB junction is

Solutions Chapter 9
⎡ (1016 ) ⋅ ( 5 × 1017 ) ⎤
kT ⎛ N D' N A' ⎞
Vbi (CB ) = ln ⎜ ⎟ = ( 0.026V ) ln ⎢ ⎥ = 0.82V
q ⎝ ni2 ⎠ ⎢ (1.08 × 1010 )2 ⎥
⎣ ⎦
At VCE=-12V, the junction voltage across the CB junction is VjCB=VbiCB-

VaCB= VbiCB +VCE-VBE=0.82+12-0.75=12.07V. The width in the base due to
the CB junction is
1 1
⎡ ⎤2 ⎡ ⎤2
⎢ ⎥
W pCB = ⎢ 2ε (V jCB ) ⎥ = ⎢ ( )
⎢ 2 (11.8 ) 8.85 × 10−14 (12.07V ) ⎥
⎥ = 0.024 µ m
⎢ ⎛ '
⎞ ⎥ ⎢ ⎛ × 17 ⎥
⎞
⎢ (1.6 × 10 )( 5 × 10 ) ⎜ 1 +
N −19 5 10
⎢ qN AB
'
⎜1 + ' ⎟ ⎥
AB 17
⎟⎥
⎢⎣ ⎥ ⎢ ⎠ ⎥⎦
16
⎝ N DC ⎠ ⎦ ⎣ ⎝ 10
Therefore this base must be made at least 0.016 µm+0.024 µm≈0.04 µm
thick.
9.24. As indicated in Section 9.11, to avoid excessive base push-out, JC is less than
about 0.25 mA/µm2. Consider a BJT with an emitter width (L) of 0.5µm and a base
sheet resistance of 10kΩ / . For β = 100 , find IB and the lateral voltage drop in the
intrinsic base region and discuss the current crowding effect for this device.
JC 0.25 ×10−3 mA / µ m 2
JB = = = 2.5µ A / µ m 2
β 100
I B = J B Lh
For negligible current crowding, from Equation (9.75),
RL
RB =
4h
R L J B LhR L J B R L2
VB = I B RB = I B = =
4h 4h 4
104 × (0.5) 2
−6
VB = 2.5 × 10 = 1.56mV
4

Solutions Chapter 9
kT
which is much less than = 26mV
q
and so for this device, current crowding is negligible.
9.25. For the transistor whose IC-VCE curves are shown in Figure P9.5, explain
why the lines are closer together as IB increases.
There are two reasons, both relating to high injection. At high injection, the
back-injection of holes into the emitter (for an npn transistor) becomes
significant (cannot be neglected with respect to the majority carrier
1
concentration in the emitter) reducing the injection efficiency γ = .
I pE
1+
I nB
The second reason is the base push-out or Kirk effect. The concentration
of minority carriers in the base becomes so high they cannot all be
extracted at the collector edge of the base region, so the base width is
effectively increased, which lowers β from Equation 9.32.
9.26. Find β for the device whose Gummel plot is shown in Figure P9.6.

Solutions Chapter 9
1 mA
Ι C (µA)
100 µA
Current
Ι Β (µA)
10 µA
1 µA
100 nA
10 nA
0 0.2 0.4 0.6 0.8 1.0
VBE (Volts)
We choose a convenient spot in the range where IC and IB are parallel. At the
point chosen, IC=100µA and IB=0.9µA. Since β=IC/IB, β=100/0.9=111.
9.27. Figure P9.7 shows some experimental data for the current gain
(normalized to that at 300 K) of a bipolar junction transistor as a function of
temperature. Also included is a straight-line approximation to the linear
region. From this data, estimate the value of ∆Eg* .

Solutions Chapter 9
Solution: From Equation (9.43), we have (neglecting the temperature
dependence of NC and of the diffusion coefficients):
−∆E* −∆E *
N C DnBWE kT g g
β= ' e = β 0e kT
N AB DPEWB
where β0 is a constant (the β that would be predicted if bandgap
narrowing were ignored). We can rearrange the above equation to the
form
β −∆Eg ⎛ 1 ⎞
*
ln ≅ ⎜ ⎟
β0 k ⎝T ⎠
or
β
d ln
β0
∆Eg* = − k
1
d
T
Extrapolating the room temperature values to lower temperatures using

the room temperature slope as indicated by the dotted line in Figure P9.7,
we can find ∆Eg* by finding the slope
Solutions Chapter 9
⎛ β ⎞ ⎛ β (T1 ) ⎞ ⎛ β (T2 ) ⎞ ⎛ β (T1 ) ⎞
∆ ln ⎜ ⎟ ln ⎜ ⎟ − ln ⎜ ⎟ ln ⎜ ⎟
⎝ β0 ⎠ = ⎝ β0 ⎠ ⎝ β 0 ⎠ = ⎝ β (T2 ) ⎠
⎛1⎞ 1 1
−
1 1
−
∆⎜ ⎟
⎝T ⎠ T1 T2 T1 T2
Choosing two points at 1/ T1 = 3.33 × 10−3 K −1 and 1/ T2 = 10 × 10−3 K −1 , we find

that β changes by a factor of
β (T1 ) / β 0 1
= = 71
β (T2 ) / β 0 0.014
Then
⎛ β ⎞
∆ ln ⎜ ⎟
⎝ β 0 ⎠ = −k ln ( 71)
∆Eg* (eV ) = − k −3 −3
= − ( 8.62 ×10−5 eV / K ) ( −640 K )
⎛ ⎞
1 3.3 × 10 − 10 × 10
∆⎜ ⎟
⎝T ⎠
= 0.055eV

Solutions Chapter 9
SOLUTIONS CHAPTER 10
10.1 For an npn transistor with N DE

'
= 1019 cm −3 , N AB
'
= 2 × 1017 cm −3 ,
'
N DC = 5 × 1016 cm −3 , WE = 0.13µ m , and WB = 0.15µ m under the bias conditions of
I B = 20 µ A and VBC = −2.5V , find
a) β
b) IC
c) rπ
d) gm
e) CBE (Cπ )
f) Cµ
g) fco
h) fT
Note that band gap narrowing should be accounted for, and that both sides of
the E-B junction are short. Let the area of the emitter junction be AE=2.5×10-
7cm-2 and the area of the collector junction be A =8×10-7 cm-2.
C
a) To find β, we use Equation (9.43) and from Figure 2.25, ∆Eg* = 50meV .
Then
−∆Eg*
N D W
β = 'C nB E e kT
N AB D pE WB
From Figure 3.11, we find DpE≈3.8, and DnB≈13 cm-2/s.
The widths are WE=0.13µm and WB=0.15 µm.
−∆Eg*
N D W
β = 'C nB E e kT
N AB D pE WB
2.86 × 1019 13 0.15 −0.026
0.05
= e
2 × 1017 3.8 0.13
= 83

Chapter 10: Solutions
b) Ι C
Since I C = β I B ,
I C = 83x20µ A = 1.7 mA.
c) rπ
kT 0.026
rπ = = = 1300 Ω
qI B 20 × 10−6
d) gm
From Equation 10.15
β 83
gm = = = 64mS
rπ 1300
e) Cπ
For a BJT operating in the active mode, CscEB C jEB and we can
approximate Cπ ≅ CscBE . From Equation 10.30 for constant base doping,
the stored charge capacitance is
( 0.13 × 10−4 cm ) ( 83)

2
WB2 β
Csc = = = 2.4 × 10−13 F = 0.24 pF
3DnB rπ 3 (15cm 2 / s ) (1300Ω )
f) Cµ
The capacitance Cµ is just the BC junction capacitance. This is not a one-

sided junction, so the result is:

1
⎡ qε N A N D ⎤ 2
Cµ = C jBC = A ⎢ ⎥
⎣ 2 ( N A + N D )(Vbi − Va ) ⎦
1
⎡ (1.6 × 10−19 ) (11.8 ) ( 8.85 × 10−14 F / cm )( 2 × 1017 cm −3 )( 5 × 1016 cm −3 ) ⎤ 2
= ( 8 × 10−7 cm 2 ) ⎢ ⎥
⎢⎣ 2 ( 2 × 1017 + 5 × 1016 ) ( 0.83 − ( −2.5 ) ) ⎥⎦
= 2.5x10-14 F = 0.025 pF
where we found the value of VbiBC from
⎛ ⎞
kT ⎛ N A' N D' ⎞ ⎜ 2x1017 ⋅ 5 × 1016 ⎟ = 0.83V
Vbi = ln ⎜ ⎟ = 0.026 ln
⎝ ( )
q ⎝ ni2 ⎠ ⎜ 1.08 × 1010 2 ⎟
⎠
g) fco
The cutoff frequency for this transistor is, from Equation (10.37),
1 1
f co = = = 462MHz
2π rπ ( Cbe + Cµ ) 2π (1300Ω ) ( 2.4 × 10−13 + 2.5 × 10−14 F )
h) fT
fT = β DC f co = ( 83) ( 462 × 106 ) = 38 × 109 Hz = 38GHz
ic
10.2. For the transistor of Problem 10.1, plot as a function of frequency.
ib
What is the unity-gain frequency?
ic β 83
We use = = and plot against frequency.
ib ⎛ f ⎞
2
⎛ f ⎞
2
1+ ⎜ ⎟ 1+ ⎜ 6 ⎟
⎝ f co ⎠ ⎝ 462 × 10 ⎠

Note that from the plot, fT = 38 GHz in agreement with the result from
Problem 10.1h.
10.3. An npn BJT with uniformly doped emitter, base, and collector has β=95,
base width WB of 0.15 µm, an electron diffusion coefficient of 10 cm2/s, an rπ of
1000 Ω, and a CB junction capacitance of 0.05 pF. What is its cut-off frequency?
The stored charge capacitance in the uniformly doped BJT is
( 0.15 ×10−4 ) ( 95 )
2
WB2 β
Csc ( BE ) = = = 0.71× 10−12 pF
3DnB rπ 3 (10 )(1000 )
1 1
f co = = = 210MHz
2π rπ ( CscBE + Cµ ) 2π (1000 ) ( 0.71`×10−12 + 0.05 ×10−12 )
10.4. If the base in the above problem is actually graded in doping, with
a grading parameter η = 2,
a) What is the field in the base?
kT η 0.026 × 2
From Equation (9.54), E = = = 0.35V / µ m = 3.5 kV / cm
q WB 0.15µ m

Note that this field is small enough that the low field mobility and diffusion
coefficients are reasonably applicable. (Figure 3.9.)
b) For the same value of I C , what is the value of β?
For a graded base with η = 2, from Equation (9.60)
β (η ) η 2
= = = 2.3
β (0) 1 − e −η
1 − e−2
β (η ) = 2.3β (0) = 2.3 × 95=219
c) What is the resultant cutoff frequency?
For η = 2, from Figure 10.8, δ = 0.45 and from Equation (10.29)
δ WB2 β 0.45 ⎡ WB2 β ⎤ 0.45

Csc ( BE ) = = ⎢ ⎥= ⎡3 × ( Csc (uniform ) ) ⎤
2 DnB rπ 2 ⎣ DnB rπ ⎦ 2 ⎣ ⎦
= 0.67 ( 0.71 × 10−12 ) = 0.48 × 10−12 F
where we have used Equation (10.30), and the resulting cut-off frequency
is
1 1
f co = = = 300 MHz
2π rπ ( CscBE + Cµ ) 2π (1000 ) ( 0.48 × 10−12 + 0.05 × 10−12 )
300
which is an improvement of a factor of = 1.43
210
10.5. Two transistors that are otherwise identical differ in that one has a
uniform base and the other has a graded base with η=6. Assume the field in
the base is small enough that the diffusion constant has its low-field value.
a) What is the improvement in β for the graded device?
From Figure 9.12, β improves by a factor of 7.
b) What is the change in the stored charge capacitance if both transistors are
operated at the same value of IC?
δ WB2 β DC δ WB2 qI C
CscBE = =
2 DnB rπ 2 DnB kT
If all other parameters are the same, using Figure 10.8 to find the values
of δ, we have
CscBE (graded base) δ graded base 0.20

= = = 0.30
CscBE (prototype) δ prototype 0.67
So the graded base transistor has 70% lower capacitance.
10.6. Equation (10.39) indicates that the electron velocity in the base increases
as ∆n( x) decreases (InB constant). Explain the physics of this.
The current is constant across the base. Since the current is a function of
drift plus diffusion and since E is constant, so is the force on the electrons
or the electron velocity due to the field is also constant. However, near the
collector where ∆n is small, diffusion predominates and the velocity due to
diffusion must increase.
10.7. Find the base transit time tT for an npn prototype BJT with WB = 0.05 µm
and base doping of 5 × 1017 cm-3.
W B2
From Equation (10.48), t T = . From Figure 3.11, the majority carrier
2Dn
diffusion length (electrons are majority carriers in the base) for a doping
concentration of 5×1017 Dn = 11cm 2 / s . Thus
W B2 (0.05 × 10 cm)
−4 2
tT = = = 1.1ps
2Dn ⎛ cm2 ⎞
2⎜ 11 ⎟
⎝ s ⎠
10.8. Repeat the above problem for a pnp prototype BJT with WB = 0.05 µm
and base doping of 5 × 1017 cm--3.
Now the carriers of interest are the holes, and Dp=5cm2/s, and

( 0.05 × 10−4 cm )
2
WB2
t tb = = = 2.5 ps
2Dp ⎛ cm 2 ⎞
2⎜5 ⎟
⎝ s ⎠
10.9 Find the transit time tBC across the base-collector transition region for an
npn prototype BJT with base doping of 5 × 1017cm-3 , collector doping of 5 ×
1016 cm -3 and collector-base voltage of 2.5 V.
In the high field of the base-collector transition region, the electrons travel
w
at their saturation velocity. From Equation 10.50, we have t BC = BC . The
vsat
junction width at 2.5 V is given by Equation (5.91),
1
⎡ 2ε (Vbi − Va ) ( N A' + N D' ) ⎤ 2
wBC =⎢ ⎥ .
⎢⎣ qN A' N D' ⎥⎦
We find the built-in voltage from Equation (5.13)
⎛ ( 5 × 1017 )( 5 × 1016 ) ⎞
kT ⎛ N A' N D' ⎞
Vbi = ln ⎜ ⎟ = 0.026 ln ⎜⎜ ⎟ = 0.86V
⎝ (1.08 × 10 )
q ⎝ ni2 ⎠ 10` 2 ⎟
⎠
1
⎡ 2 (11.8 ) ( 8.85 × 10 ) ( 0.86 − ( −2.5 ) ) ( 5 × 10 + 5 × 10
−14 17 16
) ⎤⎥ 2
wBC = ⎢
⎢⎣ (1.6 × 10−19 )( 5 × 1017 )( 5 × 1016 ) ⎥⎦
= 3.1 × 10−5 cm = 0.31µ m
The time it takes an electron to cross the transition region is
wBC 3.1 × 10−5 cm

t BC = = = 3.1 ps
vsat 107 cm / s

10.10. In a particular prototype transistor fabrication process, the base width is
cut in half. What is the effect on the base transit time?
W B2
From Equation (10.48), t T = , thus if the base width is cut in half, the
2Dn
transit time decreases by a factor of 4.
10.11. An npn transistor's base region is doped to 1018 cm-3. How thick would
the base region need to be for the base transit time to be equal to 1/100 of the
electron lifetime in the base? If such a transistor were manufactured, describe
the effect on β and the operating frequency?
From Figure 3.21, the electron lifetime at the doping concentration is τn=1
µs.
W B2
From Equation (10.48), t tb = = 0.01(τ n ) = 10ns . From Figure 3.11, Dn
2Dn
(the electrons are minority carriers in the base) =9 cm2/V·s.
⎛ cm 2 ⎞
W B = 2Dn [0.01(τ n )] = 2⎜ 9 ⎟ (10 × 10 s ) = 4.2 × 10 cm = 4.2 µm .
−9 −4
⎝ s ⎠
This is a very thick base, and the unity current gain frequency from
1 1
Equation 10.49 is fT = = = 16 MHz
2π tT 2π (10 × 10−9 )
10.12. For high speed BJTs, the double poly self-aligned transistor is preferred
to the conventional transistor discussed in Chapter 9. For similar emitter, base
and collector dopings compare the following parameters and explain your
reasoning.
(a) Collector-base junction capacitance.
This will be smaller for the double poly self-aligned transistor because the
collector and base areas can be smaller.
(b) Forward current gain βF

− ∆E *
N C DnB W E kT g
It will be the same, because from Equation (9.43), β = ' e
N AB D pE W B
and none of these factors are influenced by this structure. The self-aligned
structure reduces the lateral dimensions of the structure but not the
thickness of the regions.
(c) Reverse current gain βR.
Will be increased because the collector-base area is reduced, thus

increasing the reverse transport efficiency
(d) Base resistance
The base resistance is reduced because the base can be shorter.
(e) Early voltage
There is no effect because the Early effect is a function of base thickness,

not lateral dimension.
10.13. For a particular Schottky-clamped transistor, the leakage current density

in the Schottky diode is J0SD=10-5 A/cm2 , and the leakage current density for
the base-collector junction is J0BC =10-11A/cm2. For a given bias voltage across
both in parallel, what is the ratio of the current flowing through the Schottky
contact to the current flowing through the BC junction? What is the impact on
turn-off time? Assume that the area of the base-collector junction is ten times
that of the Schottky diode.
The current density in a diode is
⎛ qVa ⎞
J = J 0 ⎜⎜ e kT − 1⎟⎟ . Since the voltage is the same, the ratio of the currents is
⎝ ⎠
the same as the ratio of their reverse leakage currents
I pn 10 AJ pn 10 J 0 pn 10−10
= = = −5
= 10−5 . This speeds up the output
I Schottky AJ Schottky J 0 Schottky 10
transition low-to-high time since the stored minority charge because the
forward biased C-B junction is reduced by a factor of 105.
10.14. From Equations (10.43) and (10.23), find an expression for the base
transit time in a graded base transistor as a function of η.

QB
Equation (10.43) is t T = and Equation (10.23) says that
IC
I CW B2 ⎛ η − 1− e −η ⎞
QB = ⎜ ⎟ . Combining these we have
DnB ⎝ η2 ⎠
QB W B2 ⎛ η − 1 − e−η ⎞
tT = ⎜ ⎟
IC DnB ⎝ η2 ⎠

CHAPTER 11 PROBLEM Solutions
11.1. Consider a photodetector operating in the neighborhood of 60 GHz. How

many cycles of green light are there in a single cycle of 60 GHz? Can a
photodetector be used to follow the oscillations of the electromagnetic field
associated with this light?
For the RF signal,
1 1
TRF = = = 17 ps
vRF 60 × 109
For the optical signal, the wavelength is about 500 nm, from Figure 11.1.
The period is
1 λ 500 × 10−9
Top = = = = 0.0017 ps
vop c 3 × 108
17
The RF period is = 10000 times longer, so 10,000 cycles of optical
0.0017
light fit into one cycle of the RF. Therefore even the fastest detectors can
only measure a time average intensity compared to the oscillation of light.
11.2. Light with wavelength λ=700 nm is incident on a sample of GaAs.
a) Where in the spectrum does this radiation lie?
From Figure 11.1, this is at the red edge of visible light.
b) At what depth is the incident flux (neglecting Fresnel loss) reduced to 10%
of its value at the surface? 1%?
To find the absorption coefficient from Figure 11.4, we must convert this
wavelength into energy. using the Golden Rule:
1.24 1.24
E(eV) = = = 1.77eV
λ ( µm) 0.700
From Figure 11.4, α=2.1×104 cm-1. The decay in intensity is given by

Deleted: 2/26/2004
Anderson & Anderson

−α x
FL (x) = FL (0)e or
FL ( x) − ( 2.1×104 cm −1 ) x
= 0.1 = e
FL (0)
ln(0.1) −4
x10% = − = 1.1× 10 cm = 1.1µm
2.1 × 10 4 cm −1
For 1% remaining flux,
ln ( 0.01)
x1% = − = 2.2 × 10−4 cm = 2.2 µ m
2.1 × 104 cm −1
c) The color is changed to orange. Now how deep does the light penetrate (to
the 10% level)?
From Figure 11.1, orange light has a wavelength of about 600 nm, or a
photon energy of 1.24/0.6=2.07eV.
From Figure 11.4, at this energy the absorption coefficient of GaAs
increases to 5.3×104cm-1, and
ln (0.1)
x10% = − = 0.43µm
5.3 × 10 4 cm −1
11.3. a) Calculate the Fresnel reflection at normal incidence for light going from
air to glass (n=1.5).
(1− 1.5)2 0.52

R= 2 = = 0.04 or 4% is reflected.
(1+ 1.5) 2.52
b) Explain why you can see into a store window and see your reflection at the
same time, but at night looking out your window from a lighted room you can only
see your reflection?”
When you look into a store window, you see 4% of the sunlight reflected
from you and off the surface of the glass, but you also see the light from
the merchandise being transmitted through the window. To your eye both
images are superimposed.

Deleted: 2/26/2004
Anderson & Anderson

At night, there is very little light from the far side of the window, so your
4% reflection of the room light is the only thing you can see.
Equation (11.6) is:
1 dJn (x) ∆n
+ GL (x) − p = 0
q dn τn
But we know
GL ( x) = α (1 − R ) FL ( x) = α (1 − R ) FLi e −α x
and we also know
dn
Jn = qDn
dx
Using n=np,
dJ n d 2 np
= qDn
dx dx 2
Therefore
1⎛ d 2 np ⎞ ∆n p
⎟⎟ + α (1 − R ) FLi e −
−α x
⎜⎜ qDn =0
q⎝ dx 2 ⎠ τn
1⎡ ⎛ d 2 n p 0 d 2 ∆n p ⎞⎤ ∆n p
⎟⎟ ⎥ + α (1 − R ) FLi e −
−α x
⎢ qDn ⎜ + =0
q ⎢⎣ ⎜ dx 2 dx 2 τn
⎝ ⎠ ⎥⎦
But since np0 is a constant, this reduces to
1⎡ d 2 ∆n p ⎤ ∆n p
⎢ qDn 2 ⎥
+ α (1 − R ) FLi e −α x − =0
q ⎢⎣ dx ⎥⎦ τn
which is Equation (11.9.)
11.5. For the circuit shown in Figure P11.1

Deleted: 2/26/2004
Anderson & Anderson

a) Plot the I-Va characteristic for the diode with photocurrent IL=100µA, 200µA,
and 300µA. Let IDark=I0=10-14 A and the ideality factor n=1.
RL=10 kΩ
Va
b) On your graph, also plot the load line for Va =+5V and Va=-5 V
c) Find the current flowing through the circuit for each load line and plot it
against IL. Recalling that the photocurrent IL is proportional to the intensity of the
light, under which bias regime should one operate photodiodes if one wants the
output current to be proportional to intensity?
Deleted: 2/26/2004
Anderson & Anderson

To produce a linear variation in output current with incident light, reverse
(or zero) bias should be used.
11.6. A photodiode is made of Al0.1Ga0.9As. The refractive index at λ=800 nm is 3.65.

If the junction depth is 0.3 µm, and the junction width is 1.5 µm, find the quantum
efficiency ηQ and the responsivity Rph. Assume the light is incident from air.
(1 − n ) = (1 − 3.65)
2 2
(−2.65) 2
The Fresnel reflection is R = = = 0.32
(1 + n ) (1 + 3.65)2
2
4.652
To find the absorption coefficient, we use
1.24 1.24
E (eV ) = = = 1.55eV . From Figure 11.4, we find the
λ ( µ m) 0.8µ m
absorption coefficient is α=9000 cm-1. After the surface layer, the flux
remaining is
( )(
− 9×103 cm−1 0.3×10−4 cm )
FL ( xn ) = FL (0.3µ m) = (1 − R ) FLi e −α xn = (0.68)( FLi )e
= 0.52 FLi
The amount of light remaining past the depletion region is

Deleted: 2/26/2004
Anderson & Anderson

FL ( w) = FL ( xn )e −α w
so the amount absorbed in the depletion region is
FL (0.3µ m) ⎡1 − e
( ) (1.5×10 ) ⎤ = 0.52 F 1 − 0.26 = 0.38F
−4
Li ( )
− 9×103
⎢⎣ ⎦ ⎥ Li
If every photon absorbed produces current, then
ηQ = (1 − 0.32 ) ⎡⎣e −α x (1 − e−α w ) ⎤⎦ = (0.68) [ 0.38] = 0.26

n
The responsivity is
q qλ (1.6 × 10−19 )( 0.8 × 10−6 ) 0.26 = 0.17

R ph = ηQ = ηQ = ( )
hν hc ( 6.62 × 10−34 )( 3 × 108 )
11.7. For the following solar cell whose I-V characteristics are shown in Figure
P11.2, find Isc, Voc, and η. The incident power is 15 mW.
From the graph we find ISC = -4mA, VOC = 0.95V, Im=3.5 mA, and Vm=0.83
V. The fill factor is
Im Vm 3.5 × 0.83
FF = = = 76%
IscVoc 4.0 × 0.95
and the power conversion efficiency is Deleted: 2/26/2004
Anderson & Anderson

η = FF
I scVoc
= ( 0.76 )
( 4 ×10 ) ( 0.95) = 0.19 = 19%
−3
PLi 15 ×10−3
11.8. Which absorbs more of the total solar spectrum, GaAs or Ge?
From Figure 11.8, we see that Ge has a small bandgap, and thus
can absorb all wavelengths below about 1.8µm, whereas GaAs has
bandgap such that is can only absorb wavelengths smaller than
about 0.9 µm, so GaAs is transparent to a large fraction of the solar
spectrum.
11.9. If a photon of wavelength at the solar spectrum peak of λ=0.5µm (green) is

absorbed by Si, the electron and hole have excess energy as shown on the energy
band diagram of Figure P11.3. If both carriers scatter down (or up) to the band
edges, what percentage of the absorbed energy is lost as heat?
EC
EV
From the Golden Rule, the total energy absorbed from the photon is
1.24 1.24
E (eV ) = = = 2.48eV
λ ( µ m) 0.5
If both the electron and hole relax to the band edges, the energy different
between them now is just 1.12 eV. Therefore the total energy given up as
heat is
Eph-Eg=2.48-1.12=1.36eV, or more than half.
11.10. If the doping in the p region of an n+p photodiode is decreased, one would
expect the diffusion length in a solar cell to increase. Verify (or contradict) this by Deleted: 2/26/2004
Anderson & Anderson

repeating Example 11.4 with N A' =1016 cm-3. For a factor of 10 change in doping,
what was the change in ηQ?
The parameter that is affected is the diffusion length Ln, which is increased
to 400µm (Figure 3.23). We now have
(10−2 µm −1 )(400 µm)(1 − 0.2)e −10

−2
α L (1 − R )e− αx n
⋅0.4
ηQ (λ = 1µm) = n = = 64%
(αLn + 1) (10 −2 µm−1 )(400µm) + 1
and
αLn (1− R)e −α x (1µm )(400 µm)(1− 0.2)e

n
−1 −1⋅0.4
ηQ (λ = 0.5µm) = = = 53%
(αLn + 1) (1µm−1 )(400µm) + 1
We see that the lighter doping made a significant change in the efficiency
at infrared but made little difference at green.
11.11. If a GaAs photodiode has a junction depth of 0.2 µm, and if light absorbed in
the surface layer is considered lost to surface recombination, what is the total
fractional loss in photons in the surface layer? Let the photon energy be 1.4 eV, and
repeat for Eph=1.8 eV. Where on the spectrum are these two energies?
At Eph=1.4 eV, α=4000 cm-1 (From Figure 11.4), and at Eph=1.8 eV,
α=30,000 cm-1
The light remaining after crossing the surface layer at 1.4 eV is
FL (xn ) −α x − ( 4000) (0.2×10 −4 )

=e n = e = 0.92 , so 8% is lost.
FL (0)
At 1.8 eV,
FL ( xn ) −( 3×104 )( 0.2×10−4 )
= e−α xn = e = 0.55 is kept or 45% is lost.
FL (0)
Figure 11.1 shows that 1.4 eV is in the near infrared, and 1.8 eV is in the
red.
Deleted: 2/26/2004
Anderson & Anderson

11.12. a) Explain why the PIN diode would break down first at the corners if there
were no guard ring.
The field lines are concentrated at the corners. Field lines must start and
end on charges, but if the number of positive and negative charges have
to be equal, the field lines have to be closer (higher field strength) in the
corners.
b) Draw the energy band diagram for a PIN diode under high reverse bias,
and indicate the breakdown mechanism.
(Figure appears in part (c).) The breakdown mechanism is tunneling.
c) Draw the energy band diagram for a p+-p—-i-n junction under the same
reverse bias. Explain why this structure will break down at higher voltages.
The p- region drops some of the voltage, so that the field in the intrinsic
region is reduced.
11.13. a) What should the concentration x of arsenic be in a GaAsxP1-x LED

designed to emit band to band at λ=670 nm (Figure 11.13)?
Using the Golden Rule, we find the energy corresponding to this
wavelength:
Deleted: 2/26/2004
Anderson & Anderson

1.24 1.24
E( eV) = = = 1.85eV
λ ( µm) 0.67
From Figure 11.13, we find that this bandgap results from an As
concentration of 0.65.
b) If you also take into account the fact that the electrons are concentrated a
little above EC and the holes are concentrated a little below EV, how much does that
change the bandgap you would choose?
The peaks of the electron and holes distributions are about 1/2 kT away
from the band edges. If this adds a total of kT=0.026 eV to the emission
energy, that is still a small change and we can neglect this effect.
11.14. Recall that the peak of the electron distribution with energy is about 1/2 kT
above EC, and the peak of the holes distribution is about 1/2 kT below EV If each
distribution is approximated as having an overall width of kT, estimate the spectral
width of the emission. Assume the material emits at 1.3 µm. Hint: to find ∆λ, use
E=hc/λ, and take the derivative dE/dλ to obtain an expression for ∆λ in terms of ∆E.
From the figure, we could estimate the range of possible energy emissions
is about 2kT=2x0.026 eV=∆E. Taking the derivative as suggested, we
have
−1
∆E = hc ∆λ = −
1
(6.6 × 10 −34
J ⋅ s)(3 × 108 m / s)∆ λ = 0.052eV
λ (1.3 × 10 m)
2 −6 2
∆λ =
(0.052eV ⋅ 1.6 × 10 −19
J / eV )
= 71nm
−13
1.17 × 10 J/m
Deleted: 2/26/2004
Anderson & Anderson

11.15 a) Explain why nitrogen is not a donor in GaP
Nitrogen is in Column V of the periodic table, and so is phosphorus.
Therefore if an N atom replaces a P atom in the lattice, there are no extra
electrons in the binding and thus no donor.
b) Explain why the N has no long range forces in GaP
Since the nitrogen is not a donor, it will not be ionized in the lattice. Since
it is electrically neutral it will not attract electrons, (unless the electron gets
so close as to fall into the N well).
11.16. Optical fiber manufacturers battled the OH- ion (resulting from water) for
years. These ions, when incorporated into the glass, produce a strong absorption at
1.4 µm (see Figure 11.16). They have finally managed to nearly eliminate it. Is there
a similar absorption in the earth’s atmosphere?
A look at the solar spectrum in Figure 11.8 shows almost no light at λ=1.4
µm reaches the earth, although there is radiation at that wavelength
outside the earth’s atmosphere. Thus water vapor is a significant source of
absorption in the atmosphere, as well.
11.17. What semiconductor materials can be used to produce emission at 1 eV? Of

these, are any compatible with readily available substrates (e.g. GaAs or InP)?
From Figure 11.17, considering only direct gap materials, the possibilities
are in the following systems: InGaAs, InAsP, InGaAsP, InSbP, InAlSb. Of
these the InGaAsP quaternary compounds can be lattice-matched in InP
substrates.
Deleted: 2/26/2004
Anderson & Anderson

11.18. Find the frequency difference between the Fabry-Perot resonances of an
edge-emitting laser diode chip in which the effective index the mode sees is 3.5, the
wavelength is 900 nm, and the chip length is 100 µm. If the gain curve is 50 nm wide,
how many Fabry Perot resonances are there in this range for thisdiode?
We can follow the procedure in Example 11.7, but it is quicker to take the
derivative:
−1
dq = 2 nd∆λ and recognize that the change in q from one mode to the
λ2
next is ∆q=1. Then we have
−1 −6
dq = 1 = 2(3.5)(100 × 10 m)∆ λ
(900 × 10 m)
−9 2
or
∆λ =
(900 × 10 ) = 1.16nm where we have neglected the minus sign
−9 2
2(3.5)(100 × 10 )
−6
because we are interested in the separation, not the direction. There are
50/1.16≈43 peaks under the gain curve.
11.19. A diode begins to lase when the gain in the cavity exceeds the losses. One
source of loss is the partially reflective mirrors at either end of the cavity. Some
percentage of the light power is lost each time the light strikes one of the mirrors.
How would the power-current curve of a laser be changed if coatings are added to
the facets to increase the reflectivity?
We would expect it takes less current for the gain to exceed the loss since
the loss is less. Thus, the threshold would be reduced.
Deleted: 2/26/2004
Anderson & Anderson

11.20. Lasers are often characterized with an L-I-V curve, or one that plots light,
current, and voltage, like the one in Figure P-11.4 for a VCSEL. The L-I (power, or
light, vs. current) curve uses the left axis, and the V-I curve uses the right axis. For
the laser shown in the figure, what is the ratio of the optical power emitted to the
electrical power dissipated in the device, at an operating current of 10 mA? What
happens to the rest of the power?
At 10 mA, the diode emits about 5.4 mW of optical power. It consumes
P=IV=(10mA)×(3V)=30mW. Thus the “wall-plug efficiency” is 5.4/30=18%.
The rest of the power is converted to heat.
Deleted: 2/26/2004
Anderson & Anderson

11.21. Comment on the probability of absorption (zero, low, medium, high) by a
photon of λ=600 nm (red) by the following materials:
Solution: a photon of 600 nm has a photon energy of 1.24/0.6 µm = 2.07
eV, so if the semiconductor has a band gap larger than 2.07 eV, this
photon cannot be absorbed at all. For those whose band gap is correct,
we look for indirect vs. direct gap materials. Indirect materials have a low
probability of absorption; direct materials have a high probability.
Si: low: Si has an indirect gap, and Eg=1.12 eV<2.07 eV.
Ge: high: indirect, the band gap is small enough. However there is
a direct transition at hν = 0.8eV
GaAs: high: direct and band gap is small enough
InAs: high, direct and band gap is small enough
SiC: zero: band gap is too large
GaN: zero: direct but band gap is too high
CdS: zero: direct but band gap too high.
Deleted: 2/26/2004
Anderson & Anderson


(반도체공학) Fundamentals of Semiconductor Devices Solutions

Uploaded by

Copyright:

Available Formats

You might also like

(반도체공학) Fundamentals of Semiconductor Devices Solutions

Uploaded by

Document Information

Original Description:

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

(반도체공학) Fundamentals of Semiconductor Devices Solutions

Uploaded by

Copyright:

Available Formats

SOLUTIONS CHAPTER 1

1.1. Show that Equation (1.6) follows from Equation (1.3).

Thus E1 = Evac − 9 (13.6eV ) = Evac − 122eV

Anderson & Anderson 1 2/15/04

Multiply both sides by r2 and divide by v :

Solving the right hand equality for v :

Solving the left hand equality for r substituting in v :

Anderson & Anderson 2 2/15/04

2 EK 2 ⋅ (5eV ) ⋅ (1.6 × 10−19 J / eV )

Anderson & Anderson 3 2/15/04

Anderson & Anderson 4 2/15/04

Anderson & Anderson 5 2/15/04

1.10. A nondegenerate semiconductor cannot conduct current at absolute zero

1.11. At room temperature in a cubic centimeter of intrinsic silicon, there will

Anderson & Anderson 6 2/15/04

Substituting into Equation (1.37) we obtain

1.14. (a) Calculate the de Broglie wavelength of

Anderson & Anderson 7 2/15/04

b) an electron with 10 keV of kinetic energy

⎛ 1.6 × 10−19 J ⎞ −15

p = 2mEK = 2 ( 9.1 × 10−31 kg )(1.6 × 10−15 J ) = 5.39 × 10−23 kg ⋅ m / s

iii) a tiny, drifting dust particle of mass 1 µg and a leisurely velocity of 1

iv) yourself, walking at 4 mph on your way to class..

self, about 70 kg, v =4mph=1.8m/s

p=mv=70 kg×1.8m/s=126 kg-m/s

(b)What is a typical size of an atom? You begin to see why quantum

Anderson & Anderson 8 2/15/04

1.16. Consider the E-K diagram shown in Figure 1P.3:

b) Indicate the first Brillouin zone.

Anderson & Anderson 9 2/15/04

d) In what regions of the E-K diagram are electrons most likely to be

Although there cannot be an electric field in a bulk metal, electromagnetic

Anderson & Anderson 10 2/15/04

E2 − E1 = Evac − 3.4 − (E vac − 13.6) = 10.2eV

b) What emission wavelength would you expect from a transition from E4 to

1.20. What wavelength of light should you shine on hydrogen to cause

Anderson & Anderson 11 2/15/04

b) What would happen if you passed a beam of λ=430 nm through a hydrogen

b) Why is unlikely for a photon of 2.06 eV to be absorbed by an electron

1.22. For a simple cubic crystalline structure of lattice constant a=0.50nm,

Anderson & Anderson 12 2/15/04

Anderson & Anderson 13 2/15/04

The effective mass is proportional to the reciprocal of the curvature of the

2.2. Assume a material has an E-K diagram given by

Anderson & Anderson 1 2/15/04

We use the expression for the effective mass :

Taking the derivatives:

Anderson & Anderson 2 2/15/04

This is the effective mass for electrons, which is negative as expected.

Anderson & Anderson 3 2/15/04

2.5 Consider the energy band diagram in Figure 2P.3.

Anderson & Anderson 4 2/15/04

2.6. For an optical emission to occur, four things are needed:

Anderson & Anderson 5 2/15/04

ii) What is the kinetic energy of the electron?

2.8. A sample of GaAs is subjected to an electric field of 100V/cm.

where we have used the conductivity effective mass for electrons.

E(i,j) = A ((x(i)x(i) + 2y(j)y(j)));