MODERN PHYSICS Fam Le Kien
MODERN PHYSICS Fam Le Kien
MODERN PHYSICS Fam Le Kien
MODERN PHYSICS
Fam Le Kien
Department of Applied Physics and Chemistry,
University of Electro-Communications, Chofu, Tokyo 182-8585, Japan
Textbook: CONCEPTS OF MODERN PHYSICS (sixth edition, 2003)
By Arthur Beiser
McGraw-Hill
i
Contents
I. The special theory of relativity 1
A. The Michelson-Morley experiment 1
B. The special theory of relativity 4
C. The Galilean transformation 5
D. The Lorentz transformation 7
E. Length contraction 9
F. Time dilation 11
G. Doppler eect 13
1. Doppler eect in sound 13
2. Doppler eect in light 14
H. The relativity of mass 15
I. Mass and energy 18
J. Velocity addition 21
II. Wave properties of particles 23
A. De Broglie waves 23
B. Wave function: de Broglie waves are waves of probability amplitude 24
C. Describing a wave 25
D. Phase and group velocities of de Broglie waves 27
III. Particle diraction 32
IV. Uncertainty principle 37
Average value and standard deviation 40
Compatible observables 40
Proof of the uncertainty principle 40
Uncertainty principle from the particle approach 41
Uncertainty principle for energy and time 42
V. Atomic spectra 46
A. Spectral series 46
B. The Bohr atom 47
ii
C. Energy levels and spectra 50
D. Origin of line spectra 51
VI. Correspondence principle 55
VII. The lasers 57
VIII. Quantum mechanics 61
IX. Schrodinger equation 64
X. Particle in a box 69
XI. Finite potential well 75
XII. Tunnel eect 79
iii
I. THE SPECIAL THEORY OF RELATIVITY
A. The Michelson-Morley experiment
In the past, it was assumed that light is a wave propagating in an all-pervading elastic
medium called the ether. In other words, it was assumed that there exists a universal frame
of reference. Let us see what this idea means by considering a simple analogy.
Consider a river of width D which ows with the speed v, see Fig. 1. Two boats start
out from one bank of the river with the same speed V (with respect to the water). Boat
A crosses the river to a point on the other bank directly opposite the starting point and
then returns to the starting point. Boat B heads downstream for the distance D and then
returns to the starting point. Lets calculate the time required for each round trip.
We rst consider boat A. In order to compensate the water current, boat A must head
somewhat upstream, see Fig. 2. The upstream component of the velocity of boat A should
be exactly v in order to cancel out the river current v. The perpendicular component V
V
2
v
2
= V
_
1 v
2
/V
2
. (2)
The time for the initial crossing is D/V
, that
is,
t
A
=
2D/V
_
1 v
2
/V
2
. (3)
The case of boat B is somewhat dierent. As boat B heads downstream, its speed relative
to the shore is V + v, and it travels the distance D in the time
D
V + v
. (4)
On the return trip, the speed of boat B relative to the shore is V v, and boat B travels
the distance D in the time
D
V v
. (5)
1
FIG. 1: Boat A goes directly across the river and returns to its starting point, while boat B heads
downstream for an identical distance and then returns.
FIG. 2: Boat A must head upstream in order to compensate for the river current.
The total round-trip time t
B
is the sum of these times, namely,
t
B
=
D
V + v
+
D
V v
=
2D/V
1 v
2
/V
2
. (6)
The ratio between the times t
A
and t
B
is
t
A
t
B
=
_
1 v
2
/V
2
. (7)
If we know the common speed V of the two boats and measure the ratio t
A
/t
B
, we can
determine the speed v of the river current.
2
ether current
FIG. 3: The Michelson-Morley experiment.
The reasoning used in this problem may be transferred to the analogous problem of the
passage of light waves through the ether. If there is an ether pervading space, we move
through it with at least the speed of the earths orbital motion about the sun. From the
point of view of an observer on the earth, the ether is moving past the earth. To detect this
motion, we can use a pair of light beams formed by a beamsplitter instead of a pair of boats,
see Fig. 3. One of these light beams is directed to a mirror along a path perpendicular to the
ether current. The other beam goes to a mirror along a path parallel to the ether current.
The optical arrangement is such that both beams return to the same viewing screen. The
path lengths of the two beams are chosen to be exactly the same.
If there is no ether current, the two beams will arrive at the screen in phase and will
interfere constructively to yield a bright eld of view. The presence of an ether current,
however, would cause the beams to have dierent transit times, so that they would no
longer arrive at the screen in phase but would interfere destructively. This is the essence of
the famous experiment performed by American physicists Michelson and Morley in 1887.
In the Michelson-Morley experiment, although the sensitivity was enough to detect the
expected ether current, no ether current was detected.
The negative result of the Michelson-Morley experiment had two consequences. First, it
3
said that the hypothesis of the ether is wrong. Second, it suggested that the speed of light
in free space is the same everywhere, regardless of any motion of the source or the observer.
B. The special theory of relativity
When we speak of motion, we mean motion relative to a frame of reference. Without a
frame of reference the concept of motion has no meaning. The frame of reference may be
a road, the earths surface, the sun, the center of our galaxy; but in every case we must
specify it. The absence of an ether means that there is no universal frame of reference.
Therefore, all motion exists solely relative to the person or instrument observing it. Should
we be isolated in the universe, there would be no way in which we could determine whether
we are in motion or not.
Inertial frame of reference: An inertial frame of reference is a frame in which Newtons
rst law of motion holds true. In such a frame, an object at rest remains at rest and an
object in motion continues to move at constant velocity if no force acts on it. Any frame
of reference that moves at constant velocity with respect to an inertial frame is itself an
inertial frame.
The special theory of relativity, developed by Einstein in 1905, treats problems involving
the motion of inertial frames of references at constant velocity with respect to one another.
It has a profound inuence on all of physics.
The special theory of relativity is based upon two postulates:
1) The rst postulate: The laws of physics may be expressed in equations having the
same form in all inertial frames of reference moving at constant velocity with respect to one
another. This postulate expresses the absence of a universal frame of reference.
2) The second postulate: The speed of light in free space has the same value in all inertial
frames of reference. This postulate follows directly from the result of the Michelson-Morley
experiment.
The above postulates subvert almost all of the intuitive concepts of time and space we
form on the basis of our daily experience. A simple example will illustrate this statement.
Example: We have two boats on a lake, with boat A stationary in the water and boat B
drifts at the constant velocity v. At the instant that B is abreast of A, a are is red by
somebody on one of the boats. The light from the are travels uniformly in all directions,
4
according to the second postulate of special relativity. An observer in either boat must see a
sphere of light expanding with himself at its center, even though one of the boat is changing
its position with respect to the point where the are went o.
The above situation is unusual. Why? Let us consider a more familiar analog. Instead
of ring a are, one of the observers drops a stone into water when the boats are abreast of
each other. A circular pattern of ripples spreads out. The center of the circular pattern is
the point where the stone was dropped. Therefore, the pattern appears dierent to observers
on each boat.
It is important to recognize that motion and waves in water are entirely dierent from
motion and waves of light in space; water is in itself a frame of reference while space is not,
and the wave speed in water varies with the observers motion while the wave speed of light
does not.
C. The Galilean transformation
Suppose that we are in a frame of reference S and nd that an event occurs at the time
t and has the coordinates x, y, z. Consider a dierent frame of reference S
moving with
respect to S at the constant velocity v, see Fig. 4. An observer located in S
will nd
that the same event occurs at the time t
, y
, z
, y
, z
, t
.
For simplicity, we assume that v is in the +x direction and that time in both systems is
measured from the instant when the origins of S and S
, z
, t
:
x
= x vt,
y
= y,
z
= z.
(8)
We also intuitively expect that times are the same in both frames of reference:
t
= t. (9)
The set of equations (8) and (9) is known as the Galilean transformation.
5
FIG. 4: Frame S
, we simply dierentiate Eqs. (8) with respect to time. The results are
V
x
= V
x
v,
V
y
= V
y
,
V
z
= V
z
.
(10)
The Galilean transformation and the corresponding velocity transformation are in accord
with our intuitive expectations. However, they violate both of the postulates of special
relativity. The rst postulate calls for identical equations of physics in both the S and S
frames of reference, but the fundamental equations of electricity and magnetism assume
very dierent forms when Eqs. (8) and (9) are used. The second postulate calls for the
same value of the speed of light whether determined in S or S
we have c
= c v, not correct.
Clearly, a dierent transformation is required if the postulates of special relativity are to be
satised.
6
D. The Lorentz transformation
We now develop a set of transformation equations directly from the postulates of special
relativity. A reasonable guess is
x
, so
that a single event in the frame S corresponds to a single event in the frame S
; it is simple;
and it has the possibility of reducing to the equation x
, we
need only change the sign of v to write the corresponding equation for x in terms of x
and
t
:
x = k(x
+ vt
). (12)
As in the case of the Galilean transformation, there is nothing to indicate that there
might be dierences between y and y
= y
z
= z.
(13)
To get the time transformation, we substitute Eq. (11) into Eq. (94). The result is
x = k
2
(x vt) + kvt
, (14)
from which we nd that
t
= kt +
1 k
2
kv
x. (15)
To nd k, we use the second postulate. At the instant t = t
= ct
, (17)
7
respectively. Substituting Eqs. (11) and (15) into Eq. (17) yields
k(x vt) = ckt +
1 k
2
kv
cx. (18)
Solving for x, we nd
x =
ckt + vkt
k
1k
2
kv
c
= ct
1 + v/c
1 (1/k
2
1)(c/v)
. (19)
The use of Eq. (17) gives
1 + v/c
1 (1/k
2
1)(c/v)
= 1. (20)
Hence, we obtain
k =
1
_
1 v
2
/c
2
. (21)
Thus, the transformation equations are
x
=
x vt
_
1 v
2
/c
2
y
= y
z
= z
t
=
t (vx/c
2
)
_
1 v
2
/c
2
.
(22)
The above transformation is called the Lorentz transformation.
The inverse Lorentz transformation is
x =
x
+ vt
_
1 v
2
/c
2
y = y
z = z
t =
t
+ (vx
/c
2
)
_
1 v
2
/c
2
.
(23)
The Lorentz equation reduce to the ordinary Galilean equation when the relative velocity
v of S and S
1
=
t
0
vx
1
/c
2
_
1 v
2
/c
2
, (24)
while the are at x
2
and t
0
appears at the time
t
2
=
t
0
vx
2
/c
2
_
1 v
2
/c
2
. (25)
Since x
1
,= x
2
, we have t
1
,= t
2
. Hence two events that occur simultaneously to one observer
are separated by a time interval
t
2
t
1
=
v(x
1
x
2
)/c
2
_
1 v
2
/c
2
(26)
to another observer moving at the speed v relative to the rst observer. Thus, simultaneity
is a relative concept.
E. Length contraction
A rod is lying at rest along the x
1
and x
2
. The length L
0
of the rod is
L
0
= x
2
x
1
. (27)
Suppose that we measure the length of the rod from a frame of reference S, parallel to which
the rod is moving with the velocity v. Will the length L measured in S be the same as the
length L
0
measured in S
?
The length L of the rod in the frame S is determined as
L = x
2
x
1
, (28)
9
where x
1
and x
2
are the coordinates of the rod ends measured at the same time t. According
to the inverse Lorentz transformation, we have
x
1
=
x
1
vt
_
1 v
2
/c
2
x
2
=
x
2
vt
_
1 v
2
/c
2
.
(29)
Hence, we obtain
x
2
x
1
=
x
2
x
1
_
1 v
2
/c
2
. (30)
The use of the denitions (27) and (28) yields
L
0
=
L
_
1 v
2
/c
2
(31)
or, equivalently,
L = L
0
_
1 v
2
/c
2
. (32)
According to the above equation, the length of an object in motion with respect to an
observer appears to be shorter than when it is at rest with respect to him. This phenomenon
is called the Lorentz-FitzGerald contraction or the length contraction.
The length of an object is a maximum when measured in a reference frame in which the
object is at rest.
The relativistic length contraction is negligible for ordinary speeds, but it is an important
eect at speeds close to the speed of light. A speed of 3000 km/s seems enormous to us, but
it results in a shortening in the direction of motion by a factor of only
L
L
0
=
_
1 v
2
/c
2
=
1
_
3000
3 10
5
_
2
= 0.99995 = 99.995%. (33)
On the other hand, a body traveling at 0.8 the speed of light is shortened by a factor of
L
L
0
=
_
1 v
2
/c
2
=
1
_
0.8c
c
_
2
= 0.6 = 60%. (34)
The Lorentz-FitzGerald contraction occurs only in the direction of the relative motion:
if v is parallel to x, the y and z dimensions of a moving object are the same in both S and
S
.
The Lorentz-FitzGerald contraction is a real physical phenomenon and is dierent from
the visual eects.
10
F. Time dilation
Time intervals, too, are aected by relative motion. Clocks moving with respect to an
observer appear to tick less rapidly than they do when at rest with respect to him. This
eect is called time dilation.
Lets consider an event happening at a point x
in the frame S
. An observer in S
measures the time. He nds that the event happens from the time t
1
to the time t
2
. The
duration of the event is
T
0
= t
2
t
1
. (35)
Meanwhile, an observer in S also measures the time. He nds that the above event happens
from t
1
to t
2
, where
t
1
=
t
1
+ (v/c
2
)x
_
1 v
2
/c
2
(36)
and
t
2
=
t
2
+ (v/c
2
)x
_
1 v
2
/c
2
. (37)
To the observer in S, the duration of the event is
T = t
2
t
1
=
t
2
t
1
_
1 v
2
/c
2
. (38)
Hence, we obtain
T =
T
0
_
1 v
2
/c
2
. (39)
Clearly, T > T
0
. Thus, a clock moving with respect to an observer appears to tick less
rapidly than it does when at rest with respect to him. In other words, a moving clock runs
more slowly than a stationary clock.
Example 1: mesons
We show an interesting manifestation of both the time dilation and the length contraction
in the decay of unstable particles called mesons. A meson decays into an electron an
average of T
0
= 2 10
6
s after it comes into being. mesons are created high in the
atmosphere by fast cosmic-ray particles arriving at the earth from space. mesons reach
sea level in profusion. The typical speed of mesons is v = 2.994 10
8
m/s, which is 0.998
of the velocity of light c. In the mean lifetime T
0
, mesons can travel a distance of only
L = vT
0
= (2.994 10
8
m/s) (2 10
6
s) = 600 m. (40)
11
However, mesons are actually created at attitudes more than 10 times greater than the
distance L. How to explain this paradox?
a) Lets examine the problem from the frame of reference of an observer on the ground.
The lifetime of the meson in our reference frame has been extended, due to the relative
motion, to the value
T =
T
0
_
1 v
2
/c
2
=
2 10
6
1 0.998
2
s =
2 10
6
0.063
m = 32 10
6
s. (41)
This value is almost 16 times greater than when it is at rest with respect to us. In 32 10
6
s, a meson can travel a distance
L
0
= vT = (2.994 10
8
m/s) (32 10
6
s) = 9600 m. (42)
This distance is larger than the attitude at which the meson is created.
b) Lets examine the problem from the frame of reference of the meson. In this frame,
the meson is at rest, its lifetime is T
0
= 2 10
6
s, the earth ground is moving toward the
meson. Compared to the distance L
0
= 9600 m in the frame of the ground, the distance L
appears to be shortened by the factor
_
1 v
2
/c
2
= 0.063, that is,
L
L
0
=
_
1 v
2
/c
2
. (43)
Hence, we have
L = L
0
_
1 v
2
/c
2
= 9600
1 0.998
2
m = 9600 0.063 m = 600 m. (44)
The travel time in the frame of reference of the meson is L/v = (600 m)/(3 10
8
m/s)
= 2 10
6
s. This time is the same as the lifetime of the meson. Thus, the two points of
view give identical results.
Example 2: Twin paradox
Consider the famous relativistic eect known as the twin paradox. This paradox involves
two identical clocks, one remains on earth and the other one is taken on a trip into space at
the speed v and eventually is brought back. It is customary to replace the clocks with the
pair of twins Dick and Jane. Dick is 20 years old when he takes o a space trip at a speed
of 0.8c to a star 20 light-years away. His trip takes 50 years. To Jane, who stays behind,
the pace of Dicks life is slower than her pace by a factor of
_
1 v
2
/c
2
=
1 0.8
2
= 0.6 = 60%. (45)
12
To Jane, Dicks heart beats only 3 times for every 5 beats of her heart; Dick thinks only
3 thoughts for every 5 thoughts of hers. Finally, Dick returns after 50 years according to
Janes calendar, but to Dick the trip has taken only 30 years. Dick is therefore 50 years old
whereas Jane is 70 years old.
Where is the paradox? If we consider the situation from the point of view of Dick in the
spacecraft, Jane on the earth is in motion relative to him at a speed of 0.8c. Should not
Jane then be 50 years old when the spacecraft returns, while Dick is then be 70the precise
opposite of what was conducted above?
But the two situations are not equivalent. Dick changed from one inertial frame to a
dierent one when he started out, when he reversed direction to head home, and when he
landed on the earth. Jane, however, remained in the same inertial frame during Dicks trip.
Therefore, the time dilation formula applies to Janes observations of Dick, but not to Dicks
observations of Jane.
To look at Dicks trip from his perspective, we must take into account that the distance
L he covers is shortened to
L = L
0
_
1 v
2
/c
2
= 20 light-years 0.6 = 12 light-years. (46)
Hence, to Dick, his trip took
2L/v = 2 12 light-years/0.8c = 30 light-years. (47)
Thus, the aging of the twins is nonsymmetric. This eect has been veried experimentally.
G. Doppler eect
1. Doppler eect in sound
We are familiar with the increase in frequency of a sound when its source approaches us
(or we approach the source) and the decrease in frequency when the source recedes from us
(or we recede from the source). These changes in frequency constitute the Doppler eect.
The origin of this eect is straightforward. Indeed, successive waves emitted by a source
moving toward an observer are closer together than normal because of the advance of the
source. Consequently, the separation between the waves, i.e. the wavelength, is shorter, and
hence the corresponding frequency is higher.
13
The relation between the source frequency
0
and the observed frequency is
=
0
1 + v/c
1 V/c
, (48)
where c is the speed of sound, v is the speed of the observer, and V is the speed of the source.
The signs of v and V are plus for approaching and minus receding. Transverse motion does
not cause a Doppler shift in sound.
2. Doppler eect in light
We consider a light source as a clock that ticks
0
times per second and emits a wave of
light with each tick.
a) Transverse Doppler eect in light
Assume that the observer is moving perpendicular to a line between him and the light
source. In the reference frame of the source, the proper time between two adjacent ticks is
T
0
= 1/
0
. In the reference frame of the observer, the time between two adjacent ticks is
T = T
0
/
_
1 v
2
/c
2
. The frequence measured in the reference frame of the observer is
=
1
T
=
1
T
0
_
1 v
2
/c
2
. (49)
Hence, we have
=
0
_
1 v
2
/c
2
. (50)
The observed frequency is always lower than the source frequency
0
. The same formula
is true when the source is moving perpendicular to the line between the source and the
observer. Thus, transverse motion does cause a Doppler shift in light, unlike in the case of
sound.
b) Longitudinal Doppler eect in light
Assume that the observer is receding from the source. We call S the reference frame
of the source, and call S
= 0. In the frame S
=
x vt
_
1 v
2
/c
2
=
vT
0
_
1 v
2
/c
2
t
=
t (vx/c
2
)
_
1 v
2
/c
2
=
T
0
_
1 v
2
/c
2
.
(51)
14
Since x
+[x
1 + v/c
1 v/c
. (52)
Hence, the observed frequency = 1/T is
=
0
1 v/c
1 + v/c
. (53)
The observed frequency is lower than the source frequency. The same formula is true for
the motion of the source away from the observer.
In the case where the observer is approaching the source, we have
=
0
1 + v/c
1 v/c
. (54)
In this case, the observed frequency is higher than the source frequency. The same formula
is true for the motion of the source toward the observer.
The expanding universe
The Doppler eect in light is an important tool in astronomy. Stars emit light of certain
characteristic frequencies called spectral lines. Motion of a star toward or away from the
earth results in a Doppler shift in these frequencies. The spectral lines of distant galaxies
of stars are all shifted toward the lower-frequency end and hence are called red shifts. Such
shifts indicate that galaxies are receding from us and from each other, that is, the universe
is expanding.
H. The relativity of mass
We have seen that fundamental physical quantities such as length and time have meaning
only when the reference frame in which they are measured is specied. We now show the
mass of a body also depends on the reference frame.
For this purpose, we consider an elastic collision between two particles A and B, see Fig.
5. In this collision, kinetic energy is conserved. The properties of A and B are identical
when determined in the reference frames in which they are at rest. We observe the collision
in two dierent reference frames S and S
, at the point (x
= 0, y
B
= L/V , A was thrown in the y direction at the speed V
A
while B is thrown in the
16
y
B
, where
V
A
= V
B
= V. (55)
Hence the behavior of A as seen from S is exactly the same as the behavior of B as seen
from S
. At the time t = t
A
+L/V = 0 and t
= t
B
+L/V = 0, the particles A and B reach
the origins O and O
, respectively. Since O = O
at the time t = t
B
as seen from S
is therefore
T
0
=
2L
V
A
=
2L
V
B
. (56)
Due to the time dilation eect, the round-trip time T for B as measured in S is longer than
T
0
by the factor 1/
_
1 v
2
/c
2
, that is,
T =
T
0
_
1 v
2
/c
2
. (57)
Hence, the speed V
B
of B as measured in S is
V
B
=
2L
T
=
2L
_
1 v
2
/c
2
T
0
= V
A
_
1 v
2
/c
2
. (58)
Since the collision is elastic, the total momentum in conserved, that is,
m
A
V
A
m
B
V
B
= m
A
V
A
+ m
B
V
B
, (59)
where m
A
and m
B
are the masses of A and B as measured in S. This leads to
m
A
V
A
= m
B
V
B
. (60)
Inserting Eq. (58) into Eq. (60), we nd
m
A
= m
B
_
1 v
2
/c
2
. (61)
In the above example, both A and B are moving in S. In order to obtain a formula for the
mass m of a moving body in terms of its mass m
0
when measured at rest, we consider the
limit where V
A
and V
B
are very small. Then, we have m
A
= m
0
and m
B
= m and so
m =
m
0
_
1 v
2
/c
2
. (62)
17
We call m
0
the rest mass or the proper mass. Thus, the mass m of a moving body is larger
than its rest mass m
0
by the factor 1/
_
1 v
2
/c
2
. The increase of the mass due to motion
is a relativistic eect. We call m the relativistic mass of the body.
The relativistic momentum is dened as
p = mv =
m
0
v
_
1 v
2
/c
2
. (63)
In the theory of special relativity, the Newtons second law of motion is expressed by the
formula
F =
dp
dt
=
d
dt
(mv) = m
0
d
dt
_
v
_
1 v
2
/c
2
_
. (64)
According to the above formula, we have
F =
d
dt
(mv) = m
d
dt
v + v
d
dt
m = ma +v
d
dt
m. (65)
Here a = dv/dt is the acceleration. If v varies with time, m also varies with time. In this
case, we have F ,= ma.
I. Mass and energy
The most famous formula Einstein obtained from the postulates of special relativity is
the relationship between mass and energy.
We recall from elementary physics that the work done on an object by a constant force F
that acts through a distance s is Fs. Here we have assumed that F is in the same direction
as s. If no other forces act on the object and the object starts from rest, all work done on it
becomes kinetic energy K, so K = Fs. In the general case where F need not be constant,
the formula for kinetic energy is the integral
K =
_
s
0
F ds. (66)
Using the relativistic form of the second law of motion
F =
d(mv)
dt
, (67)
18
Eq. (66) becomes
K =
_
s
0
d(mv)
dt
ds =
_
v
0
v d(mv) =
_
v
0
v d
_
m
0
v
_
1 v
2
/c
2
_
=
m
0
v
2
_
1 v
2
/c
2
m
0
_
v
0
v
_
1 v
2
/c
2
dv
=
m
0
v
2
_
1 v
2
/c
2
+ m
0
c
2
_
1 v
2
/c
2
v
0
=
m
0
c
2
_
1 v
2
/c
2
m
0
c
2
.
(68)
The above equation may be rewritten as
mc
2
= K + m
0
c
2
. (69)
We interpret mc
2
as the total energy of the body. It follows from Eq. (69) that, when the
body is at rest, that is, when K = 0, the body nevertheless possesses the energy m
0
c
2
. The
energy
E
0
= m
0
c
2
(70)
is called the rest energy. The total energy can be written as
E = mc
2
=
m
0
c
2
_
1 v
2
/c
2
. (71)
In terms of E and E
0
, Eq. (69) becomes
E = E
0
+ K. (72)
Equation (71) says that mass and energy are not independent. Mass can be created or
destroyed, but when this happens, an equivalent amount of energy simultaneously vanishes
or comes into being, and vice versa. Mass and energy are dierent aspects of the same thing.
The conversion factor between the unit of mass (kg) and the unit of energy (J) is c
2
, so
1 kg of matter has an energy of mc
2
= 1 kg (3 10
8
m/s)
2
= 9 10
16
J. This energy is
enough to send a payload of a million tons to the moon.
Kinetic energy at low speeds
The relativistic formula for the kinetic energy is
K = mc
2
m
0
c
2
=
m
0
c
2
_
1 v
2
/c
2
m
0
c
2
. (73)
19
Consider the case where v c. We use the approximation (1 +x)
n
= 1 +nx for small x to
expand the rst term in the above formula. Then we obtain the expression
K = mc
2
m
0
c
2
=
1
2
m
0
v
2
, (74)
in agreement with classical mechanics.
Energy and momentum
Total energy and momentum are conserved in an isolated system, and the rest energy
of a particle is invariant. Hence these quantities are in some sense more fundamental than
kinetic energy and velocity. Lets look into how the total energy, rest energy, and momentum
of a particle are related.
Total energy is
E =
m
0
c
2
_
1 v
2
/c
2
. (75)
Square of E is
E
2
=
m
2
0
c
4
1 v
2
/c
2
. (76)
Momentum is
p =
m
0
v
_
1 v
2
/c
2
. (77)
We have
p
2
c
2
=
m
2
0
v
2
c
2
1 v
2
/c
2
. (78)
When we subtract p
2
c
2
from E
2
, we obtain
E
2
p
2
c
2
=
m
2
0
c
4
m
2
0
v
2
c
2
1 v
2
/c
2
= m
2
0
c
4
. (79)
Hence, we have the relation
E
2
= m
2
0
c
4
+ p
2
c
2
. (80)
Massless particles
In classical mechanics, a particle must have rest mass in order to have energy and mo-
mentum. However, this requirement does not hold true in relativistic mechanics.
When m
0
= 0 and v < c, we nd from Eqs. (75) and (77) that E = p = 0. Such a
particle is meaningless. However, when m
0
= 0 and v = c, we have E = 0/0 and p = 0/0,
indicating that E and p can have any values. Thus, massless particles may exist if they
20
travel with the speed of light. The relation between the total energy E and the momentum
p of a massless particle is
E = pc. (81)
An example of massless particles is the photon.
J. Velocity addition
Special relativity postulates that the speed of light c in free space has the same value for
all observers, regardless of their relative motion. Common sense tells us that if we throw a
ball forward at 10 m/s from a car moving at 30 m/s, the balls speed relative to the road
will be 40 m/s. What if we switch on the cars headlights when its speed is v? The same
reasoning suggests that their light ought to have a speed of c + v relative to the road. But
this violates the above postulate. Common sense is not reliable in dealing with light or with
a body moving with a speed comparable to the speed of light. To get the correct results for
velocity addition, we must use the Lorentz transformation.
Consider something moving relative to both S and S
they are
V
x
=
dx
dt
, V
y
=
dy
dt
, V
z
=
dz
dt
. (83)
By dierentiating the Lorentz transformation equations for x
, y
, z
, and t
, we obtain
dx
=
dx vdt
_
1 v
2
/c
2
dy
= dy
dz
= dz
dt
=
dt (v/c
2
)dx
_
1 v
2
/c
2
.
(84)
21
So we have
V
x
=
dx
dt
=
dx vdt
dt (v/c
2
)dx
=
dx
dt
v
1 (v/c
2
)
dx
dt
=
V
x
v
1 (vV
x
/c
2
)
,
V
y
=
dy
dt
=
dy
_
1 v
2
/c
2
dt (v/c
2
)dx
=
dy
dt
_
1 v
2
/c
2
1 (v/c
2
)
dx
dt
=
V
y
_
1 v
2
/c
2
1 (vV
x
/c
2
)
,
V
z
=
dz
dt
=
dz
_
1 v
2
/c
2
dt (v/c
2
)dx
=
dz
dt
_
1 v
2
/c
2
1 (v/c
2
)
dx
dt
=
V
z
_
1 v
2
/c
2
1 (vV
x
/c
2
)
.
(85)
Thus, the formulae for velocity addition are
V
x
=
V
x
v
1 (vV
x
/c
2
)
,
V
y
=
V
y
_
1 v
2
/c
2
1 (vV
x
/c
2
)
,
V
z
=
V
z
_
1 v
2
/c
2
1 (vV
x
/c
2
)
.
(86)
The inverse formulae are
V
x
=
V
x
+v
1 + (vV
x
/c
2
)
,
V
y
=
V
y
_
1 v
2
/c
2
1 + (vV
x
/c
2
)
,
V
z
=
V
z
_
1 v
2
/c
2
1 + (vV
x
/c
2
)
.
(87)
Consider a ray of light emitted in the moving frame S
x
= c. An observer in S will measure the speed
V
x
=
V
x
+ v
1 + (vV
x
/c
2
)
=
c +v
1 + (vc/c
2
)
= c. (88)
Thus both observers in S and S
x
+v
1 + (vV
x
/c
2
)
=
0.5c + 0.9c
1 + (0.5c)(0.9c)/c
2
)
= 0.97c. (89)
This speed is less than c.
22
II. WAVE PROPERTIES OF PARTICLES
A. De Broglie waves
A photon of light of frequency has the energy E = h. According to the special relativity
theory, the energy E is related to the momentum p as E = pc. Hence, the momentum of a
photon is
p =
h
c
. (90)
Here, h = 6.626 10
34
J s is the Plancks constant and c = 2.998 10
8
m/s is the speed
of light in free space.
On the other hand, the wavelength of a photon is = c/. When we use Eq. (90), we
nd the following relation between the wavelength and momentum of a photon:
=
h
p
. (91)
The momentum p describes the particle property of the photon. The wavelength de-
scribes the wave property of the photon. Each photon has both particle and wave properties.
De Broglie suggested that Eq. (91) is completely general: It applies not only to photons
but also to material particles. A material particle, i.e., a moving body, can behave as it
has a wave nature. The matter waves are called the de Broglie waves. The wavelength of
a particle is given by Eq. (91) and is called the de Broglie wavelength. The greater the
particles momentum, the shorter its de Broglie wavelength.
The momentum of a particle of mass m and velocity v is p = mv. The de Broglie
wavelength is therefore given by
=
h
mv
. (92)
In the above equation, m is the relativistic mass, which is related to the rest mass m
0
as
m =
m
0
_
1 v
2
/c
2
. (93)
The wave and particle properties of moving bodies can never be observed at the same time.
In some situations a moving body resembles a wave and in others it resembles a particle.
Which set of properties is most conspicuous depends on how its de Broglie wavelength
compares with its dimensions and the dimensions of whatever it interacts with.
To illustrate this statement, we show two examples.
23
Example 1
Find the de Broglie wavelength of a dust particle with a mass of 10
15
kg and a velocity
of 1 mm/s.
Solution
Since v c, we can let m = m
0
. Hence
=
h
mv
=
6.6 10
34
J s
(10
15
kg) (10
3
m/s)
= 6.6 10
16
m. (94)
Assume that the diameter of the dust particle is 1 m. Then, the de Broglie wavelength
of the dust particle is very small compared with its dimensions. Therefore, we do not expect
to nd any wave aspects in its behavior.
Example 2
Find the de Broglie wavelength of an electron with a velocity of 10
7
m/s. The rest mass
of an electron is m
0
= 9.1 10
31
kg.
Solution
Since v c, we can let m = m
0
. Hence
=
h
mv
=
6.6 10
34
J s
(9.1 10
31
kg) (10
7
m/s)
= 7.3 10
11
m. (95)
The radius of the hydrogen atom is 5.3 10
11
m. The de Broglie wavelength of the
electron (with a velocity of 10
7
m/s) is comparable with the dimensions of atoms. Therefore,
the wave character of moving electrons is the key to understanding atomic structure and
behavior.
B. Wave function: de Broglie waves are waves of probability amplitude
In water waves, the quantity that varies periodically is the height of the water surface. In
sound waves, it is pressure. In light waves, electric and magnetic elds vary. What quantity
varies in the case of de Broglie matter waves?
The quantity whose variations make up matter waves is called the wave function and is
denoted by the symbol . This quantity is a function of space and time, i.e., = (r, t).
The value of the wave function at the particular point r = (x, y, z) in space at the time
t is related to the likelihood of nding the body there at the time. More precisely, [(r, t)[
2
,
24
the square of the absolute value of the wave function, is the probability of nding the body at
the point r at the time t. The value of the wave function (r, t) is the probability amplitude.
It can be negative, and is not an observable quantity.
C. Describing a wave
Consider a string stretched along the x direction. We shake the string at x = 0 up and
down along the y direction. Assume that the vibrations are harmonic in character. The
displacement of the string at x = 0 can be written as
y = Acos 2t. (96)
Here is the frequency of the vibrations and A is their amplitude.
A wave of vibrations propagates along the x direction with a wave velocity v
p
. Wave
formula:
y = Acos 2
_
t
x
v
p
_
. (97)
The wave velocity v
p
is called the phase velocity. Since the wave velocity v
p
is given by
v
p
= , we have
y = Acos 2
_
t
x
_
. (98)
The angular frequency is dened by the formula
= 2. (99)
The wave number k is dened by the formula
k =
2
=
v
p
. (100)
In terms of and k, the wave formula can be written as
y = Acos(t kx). (101)
In the three-dimensional space, k becomes a vector k normal to the wave fronts and x is
replaced by the radius vector r. Then, kx is replaced by the scalar product k r = kr.
In the case of de Broglie waves, the momentum of the particle is
p = h/ = hk/2 = hk. (102)
25
Here, h = h/2 = 1.054 10
34
J s is the reduced Plancks constant.
Exercise 1:
A photon and a particle have the same wavelength. (a) Compare their linear momenta.
(b) Compare the photons energy and the particles total energy. (c) Compare the photons
energy and the particles kinetic energy.
Answer:
(a) They have the same linear momenta: p = h/.
(b) The photons energy is E
ph
= h = hc/ = pc. The particles total energy is
E
p
=
_
p
2
c
2
+E
2
0
. Thus the photons energy is smaller than the particles total energy.
(c) The particles kinetic energy is K = E
p
E
0
=
_
p
2
c
2
+ E
2
0
E
0
< pc. Thus the
particles kinetic energy is smaller than the photons energy.
Exercise 2:
Show that the de Broglie wavelength of a particle of rest mass m
0
and kinetic energy K
is = hc/
_
K(K + 2m
0
c
2
).
Answer: Be denition, we have = h/p = hc/pc. To calculate pc, we use the kinetic
energy K = E E
0
=
_
p
2
c
2
+E
2
0
E
0
=
_
p
2
c
2
+ m
2
0
c
4
m
0
c
2
. The latter yields pc =
_
K(K + 2m
0
c
2
). Hence, we obtain
= hc/
_
K(K + 2m
0
c
2
). (103)
Note that, when v c, we have K = m
0
c
2
(1/
_
1 v
2
/c
2
1) m
0
c
2
. Hence, we obtain
= h/
2m
0
K. This formula can be derived by another way using the nonrelativistic
formulae = h/p and K = p
2
/2m
0
.
Exercise 3:
Show that if the total energy of a moving particle greatly exceeds its rest energy, its de
Broglie wavelength is nearly the same as the wavelength of a photon with the same total
energy.
Answer: For a particle, we have E =
_
p
2
c
2
+E
2
0
. Hence, pc =
_
E
2
E
2
0
. When
E E
0
, we obtain pc
= E. In this case, the de Broglie wavelength of the particle is
p
= h/p
= hc/E. Meanwhile, for a photon, we always have E = pc and hence,
ph
=
h/p = hc/E. Thus
p
=
ph
.
26
D. Phase and group velocities of de Broglie waves
How fast do de Broglie waves travel? Since a de Broglie wave is associated with a moving
body, one may expect that this wave has the same velocity as that of the body. Let us see
if this is true.
We call the de Broglie wave velocity v
p
. To nd v
p
, we can apply the usual formula
v
p
= . (104)
The wavelength is simply the de Broglie wavelength, i.e.,
=
h
mv
. (105)
To nd the frequency , we equate the quantum expression E = h with the relativistic
expression E = mc
2
. Then we obtain
=
mc
2
h
. (106)
The de Broglie wave velocity is therefore
v
p
= =
_
mc
2
h
__
h
mv
_
=
c
2
v
. (107)
The wave velocity v
p
is the phase velocity. The particle velocity v is the group velocity.
Because the particle velocity v must be less than the velocity of light c, the de Broglie wave
velocity is always larger than c. To understand this result, we must look into the distinction
between phase velocity and group velocity.
We consider a harmonic wave
y = Acos(t kx). (108)
The de Broglie waves associated with a moving body cannot be represented simply by a for-
mula resembling Eq. (108). Instead, the wave representation of a moving body corresponds
to a wave packet, or wave group. A wave group is a superposition of individual waves of
dierent wavelengths. The interference of the individual waves with one another results in
the variation in amplitude that denes the group shape. If the velocities of the individual
waves are the same, the velocity of the wave group is the common phase velocity. However,
27
FIG. 6: A wave group.
if the phase velocity varies with wavelength, an eect called dispersion, the dierent indi-
vidual waves do not proceed together. As a result, the wave group has a velocity dierent
from the phase velocities of the individual waves. This is the case with de Broglie waves.
As an example, we consider the case where the wave group consists of two waves that
have the same amplitude A but dier by a small amount in angular frequency and a
small amount k in wave number:
y
1
= Acos(t kx),
y
2
= Acos[( + )t (k + k)x]. (109)
The wave group is then given by
y = y
1
+ y
2
= 2Acos
1
2
( t k x) cos
1
2
[(2 + )t (2k + k)x]. (110)
Since and k k, we nd
y = 2Acos
1
2
( t k x) cos(t kx). (111)
The above equation represents a wave of angular frequency and wave number k whose
amplitude is modulated by an angular frequency /2 and a wave number k/2.
The eect of the modulation is to produce successive wave groups. The phase velocity v
p
is
v
p
=
k
, (112)
and the velocity of the successive wave groups is
v
g
=
k
. (113)
28
When and k have continuous spreads instead of the two values in the above discussion,
the group velocity is
v
g
=
d
dk
. (114)
We now use Eqs. (112) and (114) to calculate the phase and group velocities of de Broglie
waves. The angular frequency of the de Broglie waves associated with a body of rest mass
m
0
moving with the velocity v is
= 2 =
2mc
2
h
=
2m
0
c
2
h
_
1 v
2
/c
2
. (115)
The wave number of the de Broglie waves is
k =
2
=
2mv
h
=
2m
0
v
h
_
1 v
2
/c
2
. (116)
The group velocity of the de Broglie waves is
v
g
=
d
dk
=
d/dv
dk/dv
. (117)
It follows from Eqs. (115) and (116) that
d
dv
=
2m
0
v
h(1 v
2
/c
2
)
3/2
,
dk
dv
=
2m
0
h(1 v
2
/c
2
)
3/2
. (118)
Hence, we nd
v
g
= v. (119)
Thus, the group velocity of the de Broglie waves is the velocity of the moving body.
The phase velocity of the de Broglie waves is, as found earlier,
v
p
=
k
=
c
2
v
. (120)
The fact that v
p
> c does not violate the special relativity theory because v
p
is the motion
of the phase of the wave group, not the motion of the individual waves that make up the
group, and consequently, not the motion of the body.
Exercise 1:
An electron has a de Broglie wavelength of 2 10
12
m. Find its kinetic energy and the
phase and group velocities of its de Broglie waves.
29
Solution: First, we calculate pc:
pc =
hc
=
(6.6 10
34
)(3 10
8
)
2 10
12
= 1 10
13
J. (121)
The rest energy of the electron is E
0
= m
0
c
2
= (9.1 10
31
)(3 10
8
)
2
= 0.8 10
13
J. The
kinetic energy of the electron is
K = E E
0
=
_
p
2
c
2
+ E
2
0
E
0
=
_
(1 10
13
)
2
+ (0.8 10
13
)
2
0.8 10
13
= 0.48 10
13
J. (122)
In the units of eV (1 eV= 1.6 10
19
J), we have K = 3 10
5
eV = 300 keV.
To nd the electron velocity v, we use the formula
v
c
=
mcv
mc
2
=
pc
E
=
pc
_
p
2
c
2
+ E
2
0
=
1 10
13
_
(1 10
13
)
2
+ (0.8 10
13
)
2
= 0.78. (123)
Hence, the group velocity is v
g
= v = 0.78 c and the phase velocity is v
p
= c
2
/v = 1.28 c.
Exercise 2:
A proton and an electron have the same velocity. Compare the wavelengths and phase
and group velocities of their de Broglie waves.
Answer: (a) = h/p = h/mv. The electron has a smaller mass and consequently a longer
wavelength compared to those of the proton.
(b) Since v
g
= v, the de Broglie waves of the proton and electron have the same group
velocity.
(c) Since v
p
= c
2
/v, the de Broglie waves of the proton and electron have the same phase
velocity.
Exercise 3:
(a) A proton and an electron have the same kinetic energy. (b) Compare the wavelengths
and phase and group velocities of their de Broglie waves.
Answer: (a) = h/p and pc =
_
K(K + 2m
0
c
2
) (or K
= p
2
/2m
0
in nonrelativistic
considerations). The electron has a smaller mass than the proton. Therefore, we have
elec
>
proton
.
(b) From K = E E
0
= m
0
c
2
(1/
_
1 v
2
/c
2
1), we nd
v
2
c
2
= 1
_
1
K
K + m
0
c
2
_
2
.
30
Since m
elec
< m
proton
, we have v
elec
> v
proton
. From the nonrelativistic formula K
= m
0
v
2
/2,
we get the same conclusion. Since v
g
= v and v
p
= c
2
/v, the electron has a larger group
velocity and a smaller phase velocity than the proton does.
Exercise 4:
Verify the statement that, if the phase velocity is the same for all wavelengths of a certain
wave phenomenon, the group and phase velocities are the same.
Answer: v
p
= /k and v
g
= d/dk. If v
p
is a constant then v
g
= v
p
.
Exercise 5:
(a) Show that the phase group velocity of a particle of rest mass m
0
and de Broglie
wavelength is v
p
= c
_
1 + (m
0
c/h)
2
. (b) Compare the group and phase velocities of an
electron with the de Broglie wavelength of 1 10
13
m.
Answer: (a) We have = h/p and p = m
0
v/
_
1 v
2
/c
2
. Hence, we nd
v = c/
_
1 + (m
0
c/h)
2
. (124)
Since v
g
= v and v
p
= c
2
/v, we obtain
v
p
= c
_
1 + (m
0
c/h)
2
. (125)
(b) For m
0
= 9.110
31
kg, h = 6.610
34
Js, c = 310
8
m/s, = 110
13
m, we nd
1 + (m
0
c/h)
2
= 1.0016. Consequently, v
p
/v
g
= 1.0016, v
p
= 1.0008 c and v
g
= 0.9992 c.
31
FIG. 7: Scheme of the Davission-Germer experiment.
III. PARTICLE DIFFRACTION
A wave eect with no analog in the behavior of Newtonian particles is diraction. In
1927, Davisson and Germer demonstrated that electron beams are diracted when they are
scattered by the regular atomic arrays of crystal.
Davisson and Germer studied the scattering of electrons from a solid using an apparatus
sketched in Fig. 7. The energy of the electrons in the primary beam, the angle at which
they reach the target, and the position of the detector could all be varied.
Classical physics predicts that the scattered electrons will emerge in all directions with
only a moderate dependence of their intensity on scattering angle and even less on the energy
of the primary electrons. Using a block of nickel as the target, Davisson and Germer veried
these predictions. To prevent the crystal from being oxidized, the apparatus was kept in the
vacuum.
In the midst of their work an accident occurred that allowed air to enter their apparatus
and oxidize the metal surface. To reduce the oxide, the target was baked in a hot oven. After
this treatment, the target was returned to the apparatus and the measurements resumed.
Then the results were very dierent. Instead of a continuous variation of scattered electron
intensity with angle, distinct maxima and minima were observed whose positions depend on
the electron energy, see Fig. 8.
Two questions arise: What is the reason for this new eect? Why did it not appear until
the nickel target was baked?
De Broglies hypothesis suggested that electron waves were being diracted by the target.
32
FIG. 8: Results of the Davission-Germer experiment.
This was realized when heating a block of nickel at high temperature causes the many
individual crystals in the target to form a single large crystal. This crystal acts like a set
of many planes of atoms, called Bragg planes. The spacing of the planes in this crystal is
denoted by d = 0.091 nm. The angle of incidence and scattering relative to the Bragg planes
is denoted by . The Bragg equation for maxima in the diraction pattern is
n = 2d sin . (126)
When the kinetic energy of electrons is 54 eV, the angle is = 65
2mK =
_
(2)(9.1 10
31
kg)(54 eV)(1.6 10
19
J/eV) = 410
24
kg m/s. (129)
The electron wavelength is therefore
=
h
mv
=
6.6 10
34
Js
4 10
24
kg m/s
= 0.165 nm. (130)
33
FIG. 9: The diraction of the de Broglie waves by the target is responsible for the results of the
experiment.
Note that the realistic situation is more complicated than the above analysis. For ex-
ample, a complication arises from the fact that the energy of an electron increases when it
enters a crystal. Another complication is that there are several families of Bragg planes. The
interference between the waves diracted from dierent families may prevent the observation
of maxima even when the Bragg condition is satised.
Exercise:
What eect on the scattering angle in the Davisson-Germer experiment does increasing
the electron energy have?
Answer: leads to a decrease of and an increase of the scattering angle .
Exercise:
A beam of 50-keV electrons is directed at a crystal and diracted electrons are found at
an angle of 50
relative to the original beam. What is the spacing of the atomic planes of
the crystal? A relativistic calculation is needed for . Here m
0
c
2
= 0.5 MeV.
Answer: We have
=
hc
_
K(K + 2m
0
c
2
)
=
(6.6 10
34
Js)(3 10
8
m/s)
1.6 10
19
J/eV
_
(50 10
3
eV)(50 10
3
eV + 2(0.5 10
6
eV))
=
(6.6 10
34
)(3 10
8
)
1.6 10
19
[10
4
_
5 (5 + 100)]
m =
19.8
1.6 23
10
11
m
= 0.5 10
11
m = 0.005 nm. (131)
34
The spacing between the atomic planes is
d =
2 sin
=
0.005 nm
2 sin 65
=
0.005 nm
2 0.906
= 2.76 10
3
nm. (132)
Exercise:
A beam of 5.4-keV electrons is directed at a crystal and diracted electrons are found at
an angle of 50
relative to the original beam. What is the spacing of the atomic planes of
the crystal? We can ignore relativistic considerations.
Answer: We have
=
h
2m
0
K
=
6.6 10
34
Js
_
(2)(9.1 10
31
kg)(5.4 10
3
eV)(1.6 10
19
J/eV)
=
6.6 10
34
m
15.7 10
46
=
6.6 10
34
m
4 10
23
= 1.65 10
11
m = 0.0165 nm. (133)
The spacing between the atomic planes is
d =
2 sin
=
0.0165 nm
2 sin 65
=
0.0165 nm
2 0.906
= 9.1 10
3
nm. (134)
Particle in a box
The wave nature of a moving particle leads to some remarkable consequences when the
particle is restricted to a certain region of space instead of being able to move freely.
The simplest case is that of a particle in a box. We assume that the particle can move
only along one direction of the box, bouncing back and forth between the walls. We assume
that the walls are innitely hard, so the particle does not lose energy each time it strikes a
wall. We also assume that the velocity of the particle is suciently small that we can ignore
relativistic considerations.
From a wave point of view, a particle trapped in a box is like a standing wave. The
possible de Broglie wavelengths of the particle are therefore determined by the width L of
the box. The longest wavelength is = 2L, the next is = L, then = 2L/3, and so forth.
The general formula is
=
2L
n
(n = 1, 2, 3, . . . ). (135)
Because mv = h/, the restrictions on imposed by the box width L are equivalent to
limits on the momentum of the particle and, in turn, to limits on its kinetic energy. The
35
kinetic energy of the particle is
K =
mv
2
2
=
h
2
2m
2
. (136)
Since the permitted wavelengths are
n
= 2L/n and the particle has no potential energy in
this model, the permitted energies of the particle are
E
n
=
n
2
h
2
8mL
2
(n = 1, 2, 3, . . . ). (137)
Each permitted energy is called an energy level. The integer number n that specied an
energy level E
n
is called its quantum number.
We can draw three general conclusions:
1. A trapped particle cannot have an arbitrary energy, as a free particle can. The energies
are quantized (discrete) and can be characterized by a quantum number.
2. A trapped particle cannot have zero energy. The zero energy means v = 0 and
therefore = . There is no way to trap a wave with an innite wavelength in a box.
3. Because the Planck constant h is very smallonly 6.63 10
34
J s quantization of
energy is conspicuous only when m and L are also small. This is why we are not aware
of energy quantization in our own experience. The smaller the connement, the larger the
energy required for connement.
If a particle is conned into a rectangular volume (a three-dimentional box), the permitted
energies are
E
n
1
n
2
n
3
=
(n
2
1
+ n
2
2
+ n
2
3
)h
2
8mL
2
(n = 1, 2, 3, . . . ). (138)
Exercise: The lowest possible energy of a particle in a box is 1 eV. What are the next
two higher energies the particle can have?
Answer: 4 eV and 9 eV.
36
FIG. 10: (a) A narrow de Broglie wave group. (b) A wide wave group.
IV. UNCERTAINTY PRINCIPLE
Look at the wave group of Fig. 6. The particle that corresponds to this wave group can
be found anywhere within the group at a given time. The probability of nding the particle
is given by [[
2
.
When the wave group is narrower, the particles position can be specied more precisely,
see Fig. 10(a). However, the wavelength of the waves in a narrow packet is not well dened.
The reason is that the range of wavelengths of individual waves is large. This means that,
since = h/mv, the particles momentum is not a precise quantity. If we make a series of
momentum measurements, we will nd a broad range of values.
When the wave group is wider, the particles wavelength can be specied more precisely,
see Fig. 10(b). Therefore, the momentum can be measured more precisely. However, since
the wave group is wide, the position of the particle is not well dened. If we make a series
of position measurements, we will nd a broad range of values.
Thus we have the uncertainty principle (discovered by Heisenberg in 1927):
It is impossible to know both the exact position and exact momentum of an object at the
same time.
We present a mathematical expression for this principle. A moving body corresponds to
a single wave group, not a series of them. An isolated wave group is a superposition of an
innite number of wave trains with dierent frequencies, wave numbers, and amplitudes, see
37
FIG. 11: An isolated wave group is the result of superposing an innite number of waves with
dierent wave lengths.
Fig. 11.
At a certain time t, when (x) is a real function, the wave group (x) can be represented
by the Fourier integral
(x) =
_
0
g(k) cos kxdk. (139)
More precisely and more generally, we have the formula
(x) =
_
g(k)e
ikx
dk. (140)
The function g(k) describes the amplitudes and wavelengths of the harmonic waves that
contribute to the wave group. The narrower the wave group, the broader the range of
wavelengths involved, see Fig. 12.
We assume that the wave function (x) is spread in the interval x, and that the
Fourier transform g(k) is spread in the interval k. The spreads x and k are dened as
the standard deviations of x and k, respectively. These spreads are related to each other.
Due to the properties the Fourier transformation, we always have
x k
1
2
. (141)
To understand Eq. (141) qualitatively, we consider a simple example. Assume that
g(k) consists of only three individual components. The wave vectors of these individual
components are k
0
, k
0
k/2, and k
0
+k/2. Their amplitudes are proportional to 1, 1/2,
and 1/2, respectively. We then have
(x) = g
0
_
cos(k
0
x) +
1
2
cos
__
k
0
k
2
_
x
_
+
1
2
cos
__
k
0
+
k
2
_
x
__
= g
0
cos(k
0
x)
_
1 + cos
_
k
2
x
__
. (142)
38
FIG. 12: The wave functions and Fourier transforms for (a) a pulse, (b) a wave group, (c) a wave
train, and (d) a Gaussian distribution.
The above function is maximum at x = 0 and goes to zero at x = 2/k. The width of
this function is x = 4/k. Thus we have
x k = 4
1
2
. (143)
If the wave function (x) is a Gaussian function, then the Fourier transform g(k) is also
a Gaussian function, and we have x k = 1/2. Indeed, we take
(x) = N exp
_
x
2
4a
2
_
. (144)
The width of (x) is x = a. We nd
g(k) = N
exp
_
a
2
k
2
_
, (145)
which is a Gaussian function. We can prove this with the help of the formula
_
e
a
2
k
2
cos kxdk =
a
e
x
2
4a
2
. (146)
The width of g(k) is k = 1/(2a). Thus we have x k = 1/2.
The Fourier transform g(k) is related to the probability of nding the particle with the
momentum p = hk. More precisely, [g(k)[
2
is proportional to the probability for the particle
to have the momentum p = hk. When we use the relation p = hk and Eq. (141), we
nd
x p
h
2
. (147)
The above equation is a mathematical expression of the uncertainty principle. The spreads
x and k are the measures of the uncertainties in the position and momentum, respectively,
39
of the particle. Equation (147) says that the product of the uncertainty x in the position
of an object at some instant and the uncertainty p in its momentum component in the x
direction at the same time is equal or greater than h/2.
If we arrange the particle in such a way that x is small, then p will be large. If we
reduce p in some way, then x will be large.
Average value and standard deviation
Consider an observable A. Assume that the probability for A to take the value is
described by the probability density (). The mean value of A is
A) =
_
() d. (148)
We introduce the notation
F(A)) =
_
F()() d. (149)
In particular, we have
A
n
) =
_
n
() d. (150)
The standard deviation A of A is dened by the formula
(A)
2
= (A A))
2
) = A
2
) A)
2
. (151)
Compatible observables
If two observables A and B are compatible, they can be described by a joint probability
density (, ). We can dene
F
1
(A)F
2
(B)) =
_
F
1
()F
2
()(, ) dd. (152)
Proof of the uncertainty principle
Consider the position x and the momentum p of a particle. In quantum mechanics, x
and p are not compatible, that is, the averages xp) and px) cannot be dened in the
conventional way. Moreover, we have
xp) ,= px). (153)
40
More precisely, we have
xp) px) = i h. (154)
Dene x = x x) and p = p p). It follows from the above equation that
xp) px) = i h. (155)
As known, CC
. (162)
On the other hand, light is a wave phenomenon as well as a particle phenomenon. We
cannot determine the position of the electron with an accuracy better than the wavelength.
Consequently, we have
x . (163)
Combining Eqs. (162) and (163) gives
xp h. (164)
This result is consistent with the formula xp h/2.
Uncertainty principle for energy and time
Another form of the uncertainty principle concerns energy and time.
Consider the measurement of the energy E emitted during the time interval t in an
atomic process. Assume that the energy is in the form of electro-magnetic waves. The
energy is E = h. Therefore, the uncertainty in energy is
E = h. (165)
To measure the frequency , we account the number of waves N for the interval t and
divide this number by the time interval, that is, = N/t. Assume that the uncertainty in
number of waves in the wave group is one. Then, the uncertainty in frequency is
1
t
. (166)
It follows from Eqs. (165) and (166) that
E t h. (167)
A more rigorous treatment gives
E t
h
2
. (168)
42
Thus the product of the uncertainty in an energy measurement and the uncertainty in the
time at which the measurement is made is greater than or equal to h/2.
Consider a conservative system. For this system, the greater the energy uncertainty, the
more rapid the time evolution. More precisely, if t is a time interval at the end of which
the system has evolved to an appreciable extent and if E denotes the energy uncertainty,
t and E satisfy the relation
E t
h
2
. (169)
The above equation is a mathematical expression of the uncertainty principle for energy and
time.
The proof is given below. Consider a wave packet. The energy uncertainty E is asso-
ciated with the momentum uncertainty p via the formula
E =
dE
dp
p. (170)
Since E = h and p = hk, we have
dE
dp
=
d
dk
= v
g
. (171)
Hence
E = v
g
p. (172)
Now the characteristic evolution time t is the time taken by this wave packet to pass a
point in space. If x is the spatial extension of the wave packet, we have
t =
x
v
g
. (173)
From this we obtain
Et = px
h
2
. (174)
Example 1
A measurement establishes the position of a proton with an accuracy of 1 10
11
m.
Find the uncertainty in the position of the proton 1 second later. The rest mass of a proton
is m
0
= 1.672 10
27
kg. Assume v c.
Solution
The uncertainty in the protons position at t = 0 is x
0
= 1 10
11
m. According to
Eq. (147), the uncertainty in its momentum at this time is
p
h
2x
0
. (175)
43
Since v c, the momentum is p = mv = m
0
v. Therefore, we have p = m
0
v. Hence, the
uncertainty in the protons velocity is
v
h
2m
0
x
0
. (176)
After the time t, the position of the proton cannot be known more accurately than
x = tv
ht
2m
0
x
0
. (177)
Hence x is inversely proportional to x
0
. This means that the more we know about the
protons position at a given time, the less we know about its later position.
The value of x at t = 1 s is
x
(1.054 10
34
J s) (1 s)
2 (1.672 10
27
kg) (1 10
11
m)
= 3.15 10
3
m. (178)
Exercise: (a) Discuss the prohibition of E = 0 for a trapped particle in a box in terms of
the uncertainty principle. (b) How does the minimum momentum of such a particle compare
with the momentum uncertainty required by the uncertainty principle if we take x = L?
Answer: (a) Since the particle is trapped in the box, x is not innite. Therfore, p
cannot be zero and consequently p cannot be zero. This is why the particle cannot have
E = 0. (b) If we take x = L then p h/2x = h/2L. On the other hand, the rst
permitted value of is 2L. Therefore, the minimum momentum is p
min
= h/ = h/2L =
h/L > h/2L. Thus the minimum momentum of the trapped particle is larger than the
momentum uncertainty required by the uncertainty principle.
Exercise: Compare the uncertainties in the velocities of an electron and a proton in a
small box.
Answer: Take x = L. Then (p)
min
= h/2x = h/2L. Since p = mv, we have
(v)
min
= h/2mL. Hence, the uncertainty in the velocity of the electron is larger than that
of the proton.
Exercise: Verify that the uncertainty principle can be written as L h/2, where L
is the angular momentum and is the angular position.
Answer: Consider the rotational motion of a particle along a circle of radius a. We
have L = mvr and = x/r. Hence L = mrv = rp and = x/r. Therefore,
L = px h/2.
Exercise: A hydrogen atom is 5.3 10
11
m in radius. Use the uncertainty principle to
estimate the minimum kinetic energy an electron can have in this atom.
44
Answer: Here we have x = 5.3 10
11
m. The uncertainty in momentum is
p
h
2x
= 1 10
24
kg m/s. (179)
The momentum of the electron must be at least comparable to its uncertainty. Consequently,
the kinetic energy of the electron is
K =
p
2
2m
(1 10
24
)
2
(2)(9.1 10
31
)
J 5 10
19
J
= 3 eV. (180)
45
V. ATOMIC SPECTRA
When an atomic gas is excited by passing an electric current through it, the emitted
radiation has a spectrum which contains specic wavelengths only.
Each element has a characteristic line spectrum.
The number, strength, and exact wavelengths of the lines in the spectrum of an element
depend on temperature, pressure, the presence of electric and magnetic elds, and the motion
of the source.
Spectroscopy is therefore a useful tool for analyzing the composition and the state of a
source.
A. Spectral series
It has been experimentally found that the spectral lines of an element fall into sets called
spectral series.
For hydrogen, the Lyman series contains the wavelengths given by the formula
1
= R
_
1
1
2
1
n
2
_
with n = 2, 3, 4, . . . , (181)
and is in the ultraviolet region (40010 nm). Here, R = 1.097 10
7
m
1
is the Rydberg
constant.
The Balmer series contains the lines
1
= R
_
1
2
2
1
n
2
_
with n = 3, 4, 5, . . . , (182)
and is in the visible region (800400 nm).
In the infrared region (from 800 nm to 1 mm), three series have been found. They are
Paschen:
1
= R
_
1
3
2
1
n
2
_
with n = 4, 5, 6, . . . , (183)
Brackett:
1
= R
_
1
4
2
1
n
2
_
with n = 5, 6, 7, . . . , (184)
and
Pfund:
1
= R
_
1
5
2
1
n
2
_
with n = 6, 7, 8, . . . (185)
The existence of spectral lines and series poses a test for any theory of atomic structure.
46
FIG. 13: Spectral series of hydrogen.
Exercise: What is the shortest wavelength present in the Brackett series of spectral lines?
Answer: The Brackett series is 1/ = R(1/4
2
1/n
2
) with n = 5, 6, 7, . . . . The shortest
wavelength in this series is = 16/R.
Exercise: What is the shortest wavelength present in the Paschen series of spectral lines?
Answer: The Paschen series is 1/ = R(1/3
2
1/n
2
) with n = 4, 5, 6, . . . . The shortest
wavelength in this series is = 9/R.
B. The Bohr atom
The rst theory of the atom to meet with any success was put forward in 1913 by Niels
Bohr. This theory can be formulated in terms of the de Broglie waves as shown below.
Consider an electron in orbit around a hydrogen nucleus. The de Broglie wavelength of
this electron is
=
h
mv
. (186)
47
To determine v, we recall that the centripetal force is
F
c
=
mv
2
r
. (187)
This force is provided by the electric force
F
e
=
1
4
0
e
2
r
2
. (188)
The condition for a stable orbit is F
c
= F
e
, i.e.,
mv
2
r
=
1
4
0
e
2
r
2
. (189)
The electron velocity is therefore found to be
v =
e
4
0
mr
. (190)
Hence, the orbital electron wavelength is
=
h
e
_
4
0
r
m
. (191)
We assume that the motion of the electron in the hydrogen atom is analogous to the
vibrations of a wire loop. We know that, in a wire loop, the loops circumference is an
integer number of the wavelength of the resonant mode. Therefore, we assume that an
electron can circle a nucleus only if its orbit contains an integer number of the de Broglie
wavelength. Thus, the condition for orbital stability is
n = 2r, with n = 1, 2, 3, . . . (192)
The integer n is called the quantum number of the orbit. When we substitute Eq. (192)
into Eq. (191), we nd that the radii of the orbits are
r = r
n
= n
2
h
2
0
me
2
, with n = 1, 2, 3, . . . (193)
The radius of the innermost orbit is called the Bohr radius of the hydrogen atom and is
denoted by a
0
:
a
0
= r
1
=
h
2
0
me
2
= 5.292 10
11
m. (194)
Here we have used the parameters e = 1.610
19
C,
0
= 8.8510
12
F/m, m
0
= 9.110
31
kg, and h = 6.6 10
34
Js.
48
FIG. 14: Vibrations of a wire loop.
The other radii are given in terms of a
0
by the formula
r
n
= n
2
a
0
, with n = 1, 2, 3, . . . (195)
Exercise 1: In the Bohr model, the electron is in constant motion. How can such an
electron have a negative energy?
Answer: The potential energy of the electron interacting with the proton is U =
e
2
/4
0
r, a negative value. The kinetic energy of the electron is K = mv
2
/2 = e
2
/8
0
r.
Since the absolute of the potential energy is larger than the kinetic energy, the total energy
E = K +U of the electron is negative.
Exercise 2: Derive a formula for the speed of an electron in the nth orbit of a hydrogen
atom using the Bohr model.
Answer: In the Bohr model, r
n
= n
2
a
0
, where a
0
= h
2
0
/me
2
is the Bohr radius. From
the condition for orbital stability n = 2r, we have = 2na
0
. From the de Broglie
relation = h/mv, we obtain v = h/nma
0
.
Exercise 3: An electron at rest is released far away from a proton and moves toward the
proton. (a) Show that the de Broglie wavelength is proportional to
r, where r is the
distance of the electron from the proton. (b) Find when r = a
0
. How does this compare
with the wavelength of an electron in the ground-state Bohr orbit? (c) In order for the
electron to be captured by the proton to form a ground-state hydrogen atom, energy must
be lost by the system. How much energy?
Answer: (a) It follows from the energy conservation law E = K + U = 0 that K =
49
p
2
/2m = U = e
2
/4
0
r. Hence p = e
_
m/2
0
r. Therefore, = h/p =
_
2
0
h
2
r/me
2
=
2a
0
r.
(b) When r = a
0
, we have =
2 a
0
. The wavelength of an electron in the ground-state
Bohr orbit is
0
= 2a
0
. As we can see, <
0
.
(c) The energy to be released is E
1
= U/2 = (1/8
0
)(e
2
/r
1
) = e
2
/8
0
a
0
=
me
4
/8
2
0
h
2
= 2.18 10
18
J = 13.6 eV.
Exercise 4: Compare the uncertainty in the momentum of an electron conned to a region
of linear dimension a
0
with the momentum of an electron in the ground-state Bohr orbit.
Answer: p h/2a
0
. According to exercise 2, the momentum of an electron in the
ground-state Bohr orbit is p
1
= mv
1
= h/a
0
. Thus the uncertainty in the momentum of the
electron is less than the momentum of the electron in the ground-state Bohr orbit.
C. Energy levels and spectra
The total electron energy E is the sum of its kinetic energy
mv
2
2
(196)
and potential energy
e
2
4
0
r
. (197)
This means that
E =
mv
2
2
e
2
4
0
r
. (198)
Due to Eq. (189), we nd
E =
e
2
8
0
r
. (199)
Dierent permitted orbits have dierent energies: Due to Eq. (189), we nd
E
n
=
e
2
8
0
r
n
. (200)
When we use Eq. (193) for r
n
, we see that
E
n
=
me
4
8
2
0
h
2
1
n
2
=
E
1
n
2
, with n = 1, 2, 3, . . . (201)
Here, E
1
= 2.18 10
18
J = 13.6 eV.
The energies specied by Eq. (201) are called the energy levels of the hydrogen atom.
These levels are all negative. This means that the electron does not have enough energy to
50
escape from the nucleus. An electron in the hydrogen atom can have only these energies
and no others.
The lowest level E
1
is called the ground state, and the higher levels E
2
, E
3
, E
4
, . . . are
called excited states.
As the quantum number n increases, the level energy E
n
tends to 0. In the limit of
n = , we have E
1
n
2
f
_
, (203)
where n
i
is the quantum number of the initial state and n
f
is the quantum number of the
nal state. Therefore, we have
=
E
1
h
_
1
n
2
f
1
n
2
i
_
. (204)
Since = c/, we nd the formula for the spectrum:
1
=
E
1
hc
_
1
n
2
f
1
n
2
i
_
. (205)
The Lyman series corresponds to the case where n
f
= 1:
Lyman (n
f
= 1):
1
=
E
1
hc
_
1
1
2
1
n
2
_
, with n = 2, 3, 4, . . . (206)
Balmer (n
f
= 2):
1
=
E
1
hc
_
1
2
2
1
n
2
_
, with n = 3, 4, 5, . . . (207)
51
FIG. 15: Transitions between energy levels.
Paschen (n
f
= 3):
1
=
E
1
hc
_
1
3
2
1
n
2
_
, with n = 4, 5, 6, . . . (208)
Brackett (n
f
= 4):
1
=
E
1
hc
_
1
4
2
1
n
2
_
, with n = 5, 6, 7, . . . (209)
Pfund (n
f
= 5):
1
=
E
1
hc
_
1
5
2
1
n
2
_
, with n = 6, 7, 8, . . . (210)
The Rydberg constant is calculated to be
R =
E
1
hc
=
me
4
8
2
0
h
3
c
= 1.097 10
7
m
1
. (211)
Thus, the theoretical results of the Bohr model are in agreement with the experimental
results.
Example 1
Find the longest wavelength present in the Balmer series of hydrogen.
Solution
In the Balmer series, the quantum number of the nal state is n
f
= 2. The wavelengths
of the lines in this series is given by the formula
1
= R
_
1
2
2
1
n
2
_
, with n = 3, 4, 5, . . . (212)
52
The longest wavelength corresponds to n = 3:
1
= R
_
1
2
2
1
3
2
_
= 0.139 R. (213)
Hence, we obtain
=
1
0.139 R
=
1
0.139 (1.097 10
7
m
1
)
= 6.56 10
7
m = 656 nm. (214)
Exercise 1: When radiation with a continuous spectrum is passed through a volume of
hydrogen gas whose atoms are all in the ground state, which spectral series will be present
in the resulting absorption spectrum?
Answer: Lyman series: 1/ = R(1 1/n
2
).
Exercise 2: A proton and an electron, both at rest initially, combine to form a ground-
state hydrogen atom. A single photon is emitted in this process. What is its wavelength?
Answer: The energy of the ground state is E
1
. The wavelength of the emitted photon is
1/ = E
1
/ch = R = 1.097 10
7
m
1
. Hence = 0.912 10
7
m = 91.2 nm.
Exercise 3: How many dierent wavelengths would appear in the emission spectrum of
hydrogen atoms initially in the n = 5 state?
Answer: 4 lines (n = 5 n
1 0.889 = 1/
0.111 = 1/0.333 = 3.
Exercise 5: Doppler eect An excited atom of mass m and initial speed v emits a photon
in its direction of motion. Assume that v is small compared to c, so that relativistic con-
siderations are not required. If h/2mc
2
1, use the requirement that linear momentum
and energy must both be conserved to show that the frequency of the photon is higher by
/
0
= v/c than it would have been if the atom had been at rest.
Answer: We have two equations
mv = hk + mv
,
E
i
+
mv
2
2
= E
f
+ h +
mv
2
2
. (215)
53
Introduce the notation E
i
E
f
= h
0
. Elimination of v
=
0
(1 + v/c). (217)
Then, we have /
= v/c.
54
VI. CORRESPONDENCE PRINCIPLE
Quantum physics is very dierent from classical physics in the microworld. However,
quantum physics must give the same results as classical physics in the macroworld where
classical physics is valid. We show that this basic requirement is true for the Bohr model of
the hydrogen atom.
We calculate the frequency of radiation from a hydrogen atom.
a) Classical picture. According to the electromagnetic theory, the radiation frequency of
a hydrogen atom is equal to the frequency of the evolution of the electron rotating around
the nucleus, or to an integer multiple of this frequency. According to Eq. (190), the speed
of the electron is
v =
e
4
0
mr
. (218)
Hence, the evolution frequency of the electron is
f =
v
2r
=
e
2
4
0
mr
3
. (219)
The radius of an stable orbit is given in terms of the quantum number n by Eq. (193) as
r = n
2
h
2
0
me
2
. (220)
Therefore, the evolution frequency of the electron is
f =
me
4
4
2
0
h
3
n
3
=
E
1
h
2
n
3
. (221)
b) Quantum picture. According to the Bohr model, when the electron drops from an
orbit n
i
to an orbit n
f
, a photon is emitted. The frequency of the emitted photon is given
by Eq. (204) as
=
E
1
h
_
1
n
2
f
1
n
2
i
_
. (222)
Under what circumstances should the Bohr atom behave classically. Assume that the
electron orbit is so large that we might be able to measure it directly. In this case, quantum
eects ought to be negligible. An orbit 10 m across, for instance, meets this specication.
The quantum number of this orbit is n = 435, a large number.
Lets write n = n
i
and n p = n
f
. With this notation,
=
E
1
h
_
1
(n p)
2
1
n
2
_
=
E
1
h
2np p
2
n
2
(n p)
2
. (223)
55
When n is much larger than p, we can use the approximations 2npp
2
2np and (np)
2
n
2
. Hence, we nd the photon frequency
=
E
1
h
2p
n
3
. (224)
When p = 1, the photon frequency is exactly the same as the frequency of rotation f of
the orbital electron given by Eq. (221). Multiples of the rotation frequency f are radiated
when p = 2, 3, 4, . . . . Hence, both quantum and classical pictures of the hydrogen atom give
the same results in the limit of very large quantum numbers.
The requirement that quantum physics give the same results as classical physics in the
limit of large quantum numbers was called by Bohr the correspondence principle.
Since the de Broglie electron wavelength is = h/mv, we can rewrite the condition for
orbital stability n = 2r, see Eq. (192), in the form
mvr =
nh
2
. (225)
Exercise 1: Of the following quantities, which increase and which decrease in the Bohr
model as n increases: Frequency of evolution, electron speed, electron wavelength, angular
momentum, potential energy, kinetic energy, total energy.
Answer: Frequency of evolution f v/r 1/n
3
decreases, electron speed v 1/
r
1/n decreases, electron wavelength r/n n increases, kinetic energy K v
2
1/n
2
decreases, potential energy U 1/r 1/n
2
increases, total energy E 1/n
2
increases.
Exercise 2: Show that the frequency of the photon emitted by a hydrogen atom in going
from the level n+1 to the level n is always intermediate between the frequencies of evolution
of the electron in the respective orbits.
Answer: Photon frequency is
=
E
1
h
_
1
(n + 1)
2
1
n
2
_
=
[E
1
[
h
_
1
n
2
1
(n + 1)
2
_
. (226)
The frequencies of revolution is
f
n
=
E
1
h
2
n
3
=
[E
1
[
h
2
n
3
. (227)
Since
2
(n + 1)
3
<
1
n
2
1
(n + 1)
2
<
2
n
3
, (228)
we have f
n+1
< < f
n
.
56
FIG. 16: Transitions between two energy levels in an atom.
VII. THE LASERS
The laser is a device that produces a light beam with some remarkable properties:
1. The light is nearly monochromatic.
2. The light is coherent, with the waves all exactly in phase with each another. An
interference pattern can be obtained not only by placing two slits in a laser beam but also
by using beams from two separate lasers.
3. The beam is well collimated. It spreads out very little, even over long distances.
4. The beam is extremely intense, much more intense than any other source.
The key to the laser is the presence in many atoms of one or more excited energy levels
whose lifetimes may be 10
3
s or more instead of the usual 10
8
s. Such long-lived states
are called metastable (temporally stable).
Consider two energy levels E
0
and E
1
. There are three kinds of transition involving these
levels. If the atom is initially in the lower state E
0
, it can be raised to E
1
by absorbing a
photon of energy E
1
E
0
= h. This process is called induced absorption. If the atom is
initially in the upper state E
1
, the atom can drop to E
0
by emitting a photon of energy h.
This process is called spontaneous emission.
Einstein in 1917 pointed out that there is a third possibility, called induced emission,
in which an incident photon of energy h causes a transition from E
1
to E
0
. In induced
emission, the radiated light waves are exactly in phase with the incident ones, so the result is
an enhanced beam of coherent light. Induced emission has the same probability as induced
absorption. In other words, when a photon is incident to an atomic system, the probability
of emission of an initially excited atom is equal to the probability of absorption of an initially
unexcited atom.
A three-level laser
57
photons
FIG. 17: The principle of the laser.
The simplest kind of lasers is a three-level laser. A three-level laser uses an ensemble of
atoms (or molecules) that have a metastable state E
1
with the energy h above the ground
state E
0
and a still higher excited state E
2
that decays to the metastable state E
1
, see Fig.
17.
We need to create a situation where more atoms are in the metastable state than in the
ground state. Then, we shine light of frequency on the ensemble. There will be more
induced emissions from atoms in the metastable state than induced absorption by atoms
in the ground state. The result will be an amplication of the original light. This is the
principle of the operation of the laser.
In an atomic system, the ground state is normally occupied to the greatest extent. The
term population inversion describes an ensemble of atoms in which the majority are in an
excited state.
There are many ways to produce a population inversion. One of them is optical pumping.
In this method, an external light source is used. Some photons of this source have the right
frequency to raise ground-state atoms to an excited state that decays spontaneously to the
desired metastable state.
58
FIG. 18: The ruby laser.
Why are three levels needed? Suppose that there are only two levels. The pumping will
produce not only upward but also downward transitions. When half of the atoms are in
each state, the rate of induced emission is equal to the rate of induced absorption, so the
atomic ensemble cannot ever have more than half its atoms in the metastable state. Due to
the lack of population inversion, laser amplication cannot occur.
A practical three-level laser
The rst successful laser is the ruby laser. It is based on the three energy levels in the
chromium ion Cr
3+
. A ruby is a crystal of aluminum oxide, Al
2
O
3
, in which some of the Al
3+
ions are replaced by Cr
3+
ions. A Cr
3+
ion has a metastable level whose lifetime is about
0.003 s. In the ruby laser, a xenon ash lamp excites the Cr
3+
ions to a higher level from
which they fall to the metastable level by losing energy to other ions in the crystal. Photons
from the spontaneous emission of some Cr
3+
ions are reected back and forth between two
mirrors at the ends of the ruby rod, stimulating other excited Cr
3+
ions to radiate. After a
few microseconds, the result is a large pulse of monochromatic, coherent red light from the
partly transparent end of the rod.
The length of the rod is made to be an integer multiple of half-wavelength, so the radiation
trapped in its forms an optical standing wave. Since the induced emissions are stimulated
by the standing wave, the generated waves are all in phase with it.
Exercise 1: For laser action to occur, the medium used must have at least three energy
levels. What must be the nature of each of these levels? Why is three the minimum number?
Answer: One level is the ground state, another one is an excited state, the last one is an
metastable (long-lived) state.
If there is only two levels, one cannot achieve positive population inversion. Optical
pumping produces not only upward but also downward transitions. Therefore, in a two-
59
level system, the population of the upper level cannot be larger than the population of the
lower level.
Exercise 2: A ruby laser emits 1-J pulses of light whose wavelength is 694 nm. What is
the minimum number of Cr
3+
ions in the ruby?
Answer: The energy of a photon is h = hc/ = (6.63 10
34
)(3 10
8
)/(694 10
9
) =
2.87 10
19
J. The number of photons in each pulse is N = 1/(2.87 10
19
) = 3.5 10
18
.
This is the minimum number of Cr
3+
ions required for such laser pulses.
60
VIII. QUANTUM MECHANICS
The fundamental dierence between classical (or Newtonian) and quantum mechanics
lies in what they describe.
In classical mechanics, the future history of a particle is completely determined by its
initial position and momentum together with the forces that act on it. In the everyday
world, i.e. the world of macroscopic bodies, all these quantities can be determined well
enough for the predictions of classical mechanics to agree with observations.
The uncertainty principle suggests that the nature of an observable quantity is dierent
in the world of microscopic particles. It is impossible to determine exactly the future history
of a particle because its initial state cannot be established with sucient accuracy. This
world requires quantum mechanics.
The quantities whose relationships quantum mechanics explores are probabilities. For
example, instead of asserting that the radius of the electrons orbit in the ground-state
hydrogen atom is 5.3 10
11
m, as the Bohr theory does, quantum mechanics states that
this is the most probable radius. In a suitable experiment most trials will yield dierent
values, either larger or smaller, the value most likely to be found will be 5.3 10
11
m.
Classical mechanics is an approximate version of quantum mechanics. The certainties
in classical mechanics are illusory, and their apparent agreement with experiment occurs
because ordinary objects consist of so many individual atoms that departures from average
behavior are unnoticeable.
Wave function
The quantity with which quantum mechanics is concerned is the wave function of a
body. The wave function itself has no physical interpretation. However, the square of its
absolute value, [(x, t)[
2
, evaluated at a particular place x at a given time t, is proportional
to the probability of nding the body there at that time. The probabilities of all other
characteristics of the body, such as linear momentum, angular momentum, and energy, can
also be established from .
The problem of quantum mechanics is to determine for a body when its freedom of
motion is limited by the action of external forces.
Wave functions are usually complex, with both real and imaginary parts. Every complex
61
wave function can be written as
= A + iB, (229)
where A and B are real functions.
The complex conjugate is
= A iB. (230)
The probability density is
[[
2
=
= A
2
+ B
2
. (231)
Thus the probability density P [[
2
is always a positive real number, as required.
We now establish certain requirements for .
Since [[
2
is proportional to the probability of nding the body at a point, the integral
of [[
2
over all space,
_
[[
2
dV (232)
is proportional to the probability of nding the body somewhere in space. Therefore, this
integral must be nite. It cannot be 0 or . If this integral is zero, the particle does not
exist. If this integral is , it cannot not have the meaning of a probability.
It is convenient to have [[
2
be equal to the probability density P of nding the particle.
If so, then it must be true that
_
[[
2
dV = 1 (233)
because
_
PdV = 1 (234)
A wave function that satises Eq. (234) is said to be normalized. Every acceptable wave
function can be normalized by multiplying it by an appropriate constant.
Besides being normalizable, must be single-valued since P can have only one value at
a particular place and time. must also be continuous. On the other hand, the momentum
components are proportional to the partial derivatives /x, /y, /z. Therefore,
these derivatives must be nite, continuous, and single-valued. Only wave functions with
all these properties can yield physically meaningful results when used in calculations, so
only such well-behaved wave functions are admissible as mathematical representations of
real bodies.
To summarize:
62
1. must be single-valued and continuous everywhere.
2. /x, /y, /z must be single-valued and continuous everywhere.
3. must be normalizable, which means that must go to 0 as x, y, or z tends to
in order that
_
[[
2
dV over all space be a nite constant.
The above rules are not always obeyed by the wave functions of particles in model situ-
ations that only approximate actual ones.
Exercise: Which of the wave functions in Fig. 5.14 of the textbook cannot have physical
signicance in the interval shown? Why not?
Answer: (b), not single-valued; (c), derivative not continuous; (d), derivative not nite
and not continuous; (f), not continuous.
63
IX. SCHR
ODINGER EQUATION
Wave equation
Consider a wave whose variable quantity is y. Assume that this wave propagates in the
x direction with the speed v. Then, the propagation of the wave is described by the wave
equation
2
y
x
2
=
1
v
2
2
y
t
2
. (235)
All the solutions of the above equation must be of the form
y = F
_
t
x
v
_
, (236)
where F is any function that can be dierentiated. The solutions F(tx/v) represent waves
traveling in the +x direction, and the solutions F(t +x/v) represent waves traveling in the
x direction.
Consider the wave equivalent of a free particle moving in the +x direction. This wave is
described as an undamped, monochromatic harmonic wave
y = Ae
i(tx/v)
. (237)
Time-dependent Schrodinger equation
In quantum mechanics, a particle is described by a wave function . We assume that
the wave function of a free particle moving along the +x direction is specied by
= Ae
i(tx/v)
. (238)
When we use the relations = 2 and v = , we have
= Ae
2i(tx/)
. (239)
In terms of the total energy E = h = 2 h and the momentum p = h/ = 2 h/, the
wave function of the free particle is
= Ae
(i/h)(Etpx)
. (240)
We dierentiate Eq. (240) twice with respect to x. Then we obtain
x
2
=
p
2
h
2
. (241)
64
Dierentiating Eq. (240) once with respect to t gives
t
=
iE
h
. (242)
Assume that Eqs. (241) and (242) are valid for an arbitrary particle. Assume that the
velocity v of the particle is small compared to the light velocity c. Then, the total energy
E of the particle is the sum of its kinetic energy p
2
/2m and its potential energy U(x, t):
E =
p
2
2m
+ U(x, t). (243)
Substituting Eq. (243) into Eq. (242) and using Eq. (241), we obtain the equation
i h
t
=
h
2
2m
x
2
+ U. (244)
This is the time-dependent Schr odinger equation in one dimension.
In three dimensions, the time-dependent Schrodinger equation is
i h
t
=
h
2
2m
_
x
2
+
2
y
2
+
2
z
2
_
+ U. (245)
Here, the potential energy U is a function of x, y, z, and t.
The Schr odinger equation is a basic principle. It must be postulated. It cannot be derived
from other principles of physics.
Linearity and superposition
An important property of the Schrodinger equation is that it is linear in the wave function
. This means that the equation has terms that contain and its derivatives but no terms
independent of and no terms that involve higher powers of and its derivatives.
As a result, a linear combination of solutions of the Schr odinger equation is also itself a
solution. If
1
and
2
are two solutions, then
= a
1
1
+ a
2
2
(246)
is also a solution. Here, a
1
and a
2
are constants. Thus the wave functions
1
and
2
obey
the superposition principle that other waves do. Consequently, interference efects can occur
for wave functions just as they can for light, sound, water, and electromagnetic waves.
Exercise: Prove that the Schrodinger equation is linear by showing that
= a
1
1
+ a
2
2
(247)
65
is also a solution of Eq. (244) if
1
and
2
are themselves solutions.
Expectation values
The wave function of a particle contains all the information about the particle. How to
extract information from the wave function? As an example, let us calculate the expectation
value x) of the position of a particle that is described by the wave function (x, t) in one
dimension.
To make the procedure clear, we rst consider a number of identical particles. Assume
that there are N
1
particles at x
1
, N
2
particles at x
2
, and so on. The average position is then
given by
x =
N
1
x
1
+ N
2
x
2
+ N
3
x
3
+
N
1
+ N
2
+ N
3
+
=
N
i
x
i
N
i
. (248)
When we are dealing with a single particle, we must replace the number N
i
of particles at
x
i
by the probability P
i
that the particle be found in an interval dx at x
i
. This probability
is
P
i
= [
i
[
2
dx, (249)
where
i
is the wave function evaluated at x = x
i
. Making the substitution N
i
= P
i
=
[
i
[
2
dx and changing the summations to integrals, we nd that the expectation value of the
position of the single particle is
x) =
_
x[[
2
dx
_
[[
2
dx
. (250)
We say that is a normalized wave function if
_
[[
2
dx = 1. When is normalized, the
expectation value for position is
x) =
_
x[[
2
dx. (251)
In more general, consider an arbitrary quantity G(x) that depends on the position x but
not on the momentum p. The expectation value of this quantity is calculated as
G(x)) =
_
G(x)[[
2
dx. (252)
The expectation value p) for momentum cannot be calculated this way because, accord-
ing to the uncertainty principle, no such function as p(x) can exist. Indeed, if we specify x,
so that x = 0, we cannot specify a corresponding p since xp h/2. The same problem
occurs for the expectation value E) for energy.
66
Example
A particle limited to the x axis has the wave function = ax between x = 0 and x = 1;
= 0 elsewhere.
(a) Find a for which is a normalized wave function.
(b) Find the probability that the particle can be found between x = 0.45 and x = 0.55.
(c) Find the expectation value x) of the particles position.
Solution
(a) The constant a satises the normalization condition
1 =
_
1
0
[[
2
dx = a
2
_
1
0
x
2
dx = a
2
x
3
3
1
0
=
a
2
3
. (253)
Hence we nd a =
3.
(b) The probability is
_
x
2
x
1
[[
2
dx = a
2
_
0.55
0.45
x
2
dx = a
2
x
3
3
0.55
0.45
= 0.0251a
2
= 0.0753. (254)
(c) The expectation value for position is
_
1
0
x[[
2
dx = a
2
_
1
0
x
3
dx = a
2
x
4
4
1
0
=
a
2
4
=
3
4
. (255)
Steady-state Schrodinger equation
Note that the wave function of a free particle (U = 0) can be written as
= Ae
(i/h)(Etpx)
= Ae
iEt/h
e
ipx/h
= e
iEt/h
, (256)
where depends on position but not on time.
Consider the case where the particle is not free. Assume that the potential U is a function
of x but is independent of time. Consider the situation where the wave function can be
written in the form
= e
iEt/h
, (257)
where is independent of time. This situation is called the steady state. Substituting Eq.
(257) into Eq. (244), we nd
Ee
iEt/h
=
h
2
2m
e
iEt/h
x
2
+ Ue
iEt/h
. (258)
67
Dividing the above equation by the common exponential factor gives
x
2
+
2m
h
2
(E U) = 0. (259)
The above equation is called the steady-state Schrodinger equation in one dimension. In
three dimensions, the steady-state Schr odinger equation is
x
2
+
2
y
2
+
2
z
2
+
2m
h
2
(E U) = 0. (260)
An important property of the steady-state Schr odinger equation is that, if it has one or
more solutions for a given system, each of these solutions corresponds to a specic value of
the energy E. That is how energy quantization appears in quantum mechanics.
Eigenvalues and eigenfunctions
The values of energy E
n
for which the steady-state Schr odinger equation can be solved
are called eigenvalues or eigenenergies. The corresponding wave functions
n
are called
eigenfunctions or eigenstates. An example of a set of eigenenergies is the discrete energy
levels of the hydrogen atom:
E
n
=
me
4
32
2
2
0
h
2
1
n
2
, with n = 1, 2, 3, . . . (261)
68
FIG. 19: An innite potential.
X. PARTICLE IN A BOX
The simplest quantum-mechanical problem is that of a particle trapped in a box with
innitely hard walls. Consider a particle that is restricted to traveling along the x axis
between x = 0 and x = L by innitely hard walls. This situation can be described by the
potential energy U that is innite on both sides of the box and is a nite constant on the
inside. Because the particle cannot have an innite amount of energy, it cannot exist outside
the box. Hence, the wave function is zero for x 0 and x L. We will calculate the
steady-state wave function for the inside of the box.
Inside the box, U is a constant. For convenience, we take U = 0 for the inside of the box.
Then, the steady-state Schrodinger equation (259) becomes
d
2
dx
2
+
2m
h
2
E = 0. (262)
The total energy E = K+U = K is positive because so is the kinetic energy K. The general
solution of the above equation is
= Asin
2mE
h
x + Bcos
2mE
h
x. (263)
Here, A and B are coecients to be evaluated.
The boundary conditions are (0) = 0 and (L) = 0. From (0) = 0, sin 0 = 0 and
cos 0 = 1, we nd B = 0. The condition (L) = 0 will be satised if sin
2mEL/ h = 0,
that is, if
2mE
h
L = n, with n = 1, 2, 3, . . . . (264)
Hence, the energy of the particle can have only certain values
E
n
=
n
2
2
h
2
2mL
2
, with n = 1, 2, 3, . . . . (265)
69
The above values are the eigenenergies of the particle. They constitute the energy levels of
the particle.
Wave function of a particle in a box
Since B = 0, the wave function of the particle inside a box is
n
= Asin
2mE
n
h
x. (266)
Using Eq. (265) for E
n
, we nd
n
= Asin
nx
L
. (267)
It is easy to verify that the above wave function satises the steady-state Schrodinger equa-
tion and the boundary conditions (0) = (L) = 0. Note that Eq. (267) is valid only for
0 x L. Outside the box, we have
n
= 0. We also note that the spatial derivative
n
of
the wave function (267) is not continuous at x = 0 as well as at x = L. The reason is that
the box with innitely hard walls is just an approximate model, not a realistic situation (see
the next section).
To nd the constant A, we calculate the integral of [
n
[
2
from x = to x = +. We
nd
_
+
[
n
[
2
dx =
_
L
0
[
n
[
2
dx = A
2
_
L
0
sin
2
nx
L
dx =
A
2
2
_
L
0
_
1 cos
2nx
L
_
dx
=
A
2
L
2
. (268)
To satisfy the normalization condition
_
+
[
n
[
2
dx = 1, (269)
we choose
A =
_
2
L
. (270)
Thus the normalized wave functions of the particle are
n
=
_
2
L
sin
nx
L
, with n = 1, 2, 3, . . . . (271)
The normalized wave functions
1
,
2
, and
3
together with the probability densities
[
1
[
2
, [
2
[
2
, and [
3
[
2
are plotted in Fig. 20.
As seen,
n
= 0 at x = 0 and x = L. At a particular place in the box, the probability
of nding the particle may be very dierent for dierent quantum numbers. For example,
70
FIG. 20: Wave functions and probability densities of a particle in a box with rigid walls.
[
1
[
2
has its maximum value in the middle of the box, while [
2
[
2
= 0 there. In other words,
a particle in the lowest energy level of n = 1 is most likely to be in the middle of the box,
while a particle in the next higher state of n = 2 is never there. Classical physics, of course,
suggests the same probability for the particle being anywhere in the box.
Example 1
Consider a particle trapped in a box of with L. Find the probability that the particle
can be found between 0.45 L and 0.55 L for the ground and rst excited states.
Solution
71
FIG. 21: Wave functions and probability densities of a particle in a box with rigid walls.
This part is one tenth of the box width and is centered on the middle of the box. Clas-
sically we would expect the particle to be in this region 10 percent of the time.
However, quantum mechanics gives dierent predictions depending on the quantum num-
ber. In fact, the probability of nding the particle between x
1
and x
2
is
P
x
1
x
2
=
_
x
2
x
1
[
n
[
2
dx =
2
L
_
x
2
x
1
sin
2
nx
L
dx =
_
x
L
1
2n
sin
2nx
L
_
x
2
x
1
=
x
2
x
1
L
1
n
cos
n(x
2
+ x
1
)
L
sin
n(x
2
x
1
)
L
. (272)
For x
1
= 0.45 L and x
2
= 0.55 L, we obtain
P
x
1
x
2
=
1
10
1
n
cos n sin
n
10
=
1
10
+ (1)
n+1
sin
n
10
n
.
For the ground state, which corresponds to n = 1, we have
P
x
1
x
2
= 0.198 = 19.8%. (273)
This is about twice the classical probability.
72
For the rst excited state, which corresponds to n = 2, we have
P
x
1
x
2
= 0.0065 = 0.65%. (274)
This low probability is consistent with the fact that [
2
[
2
= 0 at x = 0.5 L.
Note that, when n is very large, we have P
x
1
x
2
(x
2
x
1
)/L. This result is in agreement
with classical physics, as required by the correspondence principle.
Example 2
Consider a particle trapped in a box of with L. Find the expectation value x) of the
position of the particle.
Solution
x) =
_
L
0
x[
n
[
2
dx =
2
L
_
L
0
x sin
2
nx
L
dx
=
2
L
_
x
2
4
x sin(2nx/L)
4n/L
cos(2nx/L)
8(n/L)
2
_
L
0
=
L
2
. (275)
This result means that the average position of the particle is the middle of the box. It
reects the symmetry of the system. There is no conict with the fact that [
n
[
2
= 0 at L/2
for the states of n = 2, 4, 6, . . . because x) is an average, not a probability.
Exercise 1: Consider a particle in a box. Show that as n , the probability of
nding the particle between x and x +x is x/L and so is independent of x, which is the
classical expectation.
Answer: According to Eq. (272), the probability of nding the particle between x and
x + x is
P
x,x+x
=
_
x+x
x
[
n
[
2
dx =
2
L
_
x+x
x
sin
2
nx
L
dx =
_
x
L
1
2n
sin
2nx
L
_
x+x
x
=
x
L
1
2n
sin
2n(x + x)
L
+
1
2n
sin
2nx
L
. (276)
As n , the last two terms vanish. Hence, we obtain in this limit the classical result
P
x,x+x
= x/L.
Exercise 2: Consider a particle in a 3D cubic box of length L. The wave function is
given by
n
1
n
2
n
3
=
_
2
L
_
3/2
sin
n
1
x
L
sin
n
2
y
L
sin
n
3
z
L
, with n
1
, n
2
, n
3
= 1, 2, 3, . . . . (277)
73
Find E
n
1
n
2
n
3
!
Answer: Using the steady-state Schrodinger equation, we nd
E
n
1
n
2
n
3
=
(n
2
1
+ n
2
2
+ n
2
3
)
2
h
2
2mL
2
. (278)
Exercise 3: What is the relation between the ground-state energy of a 1D box and that
of a 3D box with the same length L?
Answer: E
(3D)
111
= 3E
(1D)
1
.
Exercise 4: Consider a particle in its ground state in a 3D cubic box of length L. What
is the probability of nding the particle in the volume dened by 0 x, y, z L/4?
Answer: Look at Eq. (272). In the 1D case, the probability of nding the particle in the
volume dened by 0 x L/4 is
P =
1
4
1
2
. (279)
In the 3D case, the probability of nding the particle in the volume dened by 0 x, y, z
L/4 is
P =
_
1
4
1
2
_
3
. (280)
74
FIG. 22: A nite potential well.
XI. FINITE POTENTIAL WELL
Potential energies are never innite in the real world. Therefore, the box with innitely
hard walls has no physical counterpart. However, potential wells with barriers of nite
heights certainly do exist.
Consider a potential well with square corners that is U high and L wide and contains
a particle whose energy E is less than U. According to classical mechanics, when the
particle strikes the sides of the well, it bounces o without entering regions I and III. In
quantum mechanics, the particle also bounces back and forth, but now it has a probability
of penetrating into regions I and III even though E < U.
In regions I and III, the steady-state Schrodinger equation is
d
2
dx
2
+
2m
h
2
(E U
0
) = 0. (281)
We rewrite the above equation as
d
2
dx
2
a
2
= 0, (282)
where
a =
_
2m(U
0
E)
h
> 0. (283)
The solutions to Eq. (282) are
I
= Ae
ax
+ Be
ax
,
III
= Ce
ax
+ De
ax
. (284)
Both
I
and
III
must be nite everywhere. Since e
ax
as x and e
ax
as
75
FIG. 23: Wave functions and probability densities of a particle in a nite potential well
x , the coecients B and C must be zero. Hence we have
I
= Ae
ax
,
III
= De
ax
. (285)
These wave functions decrease exponentially inside the barrier at the sides of the well.
Inside the well, the steady-state Schr odinger equation is the same as Eq. (262), that is,
d
2
dx
2
+
2m
h
2
E = 0. (286)
The solution is again
II
= Gsin
2mE
h
x + F cos
2mE
h
x. (287)
In the case of a well with innitely high barriers, we found that F = 0 in order that = 0
at x = 0 and x = L. Here, because
II
= A at x = 0 and
II
= De
L
at x = L, both the sine
and cosine components in the solution (287) are possible.
For either solution, both and d/dx must be continuous at x = 0 and x = L. The
wave functions inside and outside each side of the well must have the same values and the
same slopes where they join. When these boundary conditions are taken into account, we
nd that exact matching occurs only for certain specic values E
n
of the particle energy.
76
We shows this energy quantization in the below. We rewrite the solution as
I
= Ae
ax
,
II
= Gsin kx + F cos kx,
I
= De
ax
, (288)
where
a =
_
2m(U
0
E)
h
, k =
2mE
h
. (289)
The boundary conditions
I
(0) =
II
(0) and
I
(0) =
II
(0) give
F = A,
Gk = Aa, (290)
The boundary conditions
III
(L) =
II
(L) and
III
(L) =
II
(L) give
Gsin kL +F cos kL = De
aL
,
Gk cos kL Fk sin kL = Dae
aL
. (291)
Substituting Eqs. (290) into Eqs. (291), we obtain
A(a sin kL + k cos kL) Dke
aL
= 0,
A(a cos kL k sin kL) + Dae
aL
= 0. (292)
The above have nonzero solutions for A and D only if
ae
aL
(a sin kL + k cos kL) + ke
aL
(a cos kL k sin kL) = 0, (293)
that is, if
tan kL =
2ak
k
2
a
2
=
2(a/k)
1 (a/k)
2
. (294)
Since tan kL = 2 tan(kL/2)/[1 tan
2
(kL/2)], two possible cases are:
tan
kL
2
=
a
k
or tan
kL
2
=
k
a
. (295)
We set
k
0
=
a
2
+ k
2
=
2mU
0
h
. (296)
77
L L L L L
FIG. 24: Graphic solutions for the energies of the bound states of a particle in a square potential
well.
(1) The case
tan
kL
2
=
a
k
(297)
leads to
cos
kL
2
=
k
k
0
,
tan
kL
2
> 0. (298)
In deriving the above equations we used the formula cos
2
x = 1/(1 + tan
2
x)
(2) The case
tan
kL
2
=
k
a
(299)
leads to
sin
kL
2
=
k
k
0
,
tan
kL
2
< 0. (300)
In deriving the above equations we used the formula sin
2
x = tan
2
x/(1 + tan
2
x)
In Fig. 24, we illustrate the graphic solutions for the energies of the bound states of a
particle in a square potential well. When U
0
is nite, k
0
is nite. In this case, the number
of solutions is nite. The number of solutions increases with increasing potential height U
0
.
Note that the wavelengths that t into the well with a nite height are longer than those
for the well with an innite height. Due to this fact, the corresponding momenta are lower,
and therefore the energy levels E
n
are lower for each n than they are for a particle in an
innite well.
78
XII. TUNNEL EFFECT
In the previous section, the walls of the potential well were of nite height but they were
assumed to be innitely thick. As a result the particle was trapped forever though it could
penetrate the walls.
We now look at the situation where the potential barrier has a nite height as well as a
nite width. What we will nd is that the particle has a certain probabilitynot necessarily
great, but not zero eitherof passing through the barrier and emerging on the other side.
The particle lacks the energy to go over the top of the barrier, but it can nevertheless tunnel
through it. The higher the barrier and the wider it is, the less the chance that the particle
can get through.
Consider a beam of identical particles all of which have the kinetic energy E. The beam
is incident from the left on a potential barrier of height U
0
> E and width L. On both sides
of the barrier the potential is zero. In these regions, the steady-state Schr odinger equation
is
d
2
I
dx
2
+
2m
h
2
E
I
= 0,
d
2
III
dx
2
+
2m
h
2
E
III
= 0. (301)
The solutions to these equations are
I
= Ae
ik
1
x
+ Be
ik
1
x
, (302)
III
= Fe
ik
1
x
+ Ge
ik
1
x
. (303)
Here
k
1
=
2mE
h
=
p
h
=
2
(304)
is the wave number of the de Broglie waves that represent the particles outside the barrier.
The incoming wave is described by the term
I+
= Ae
ik
1
x
. (305)
This wave corresponds to the incident beam of particles in the sense that [
I+
[
2
is their
probability density. The ux of the particles that arrive at the barrier is proportional to
[
I+
[
2
. (306)
79
The reected wave is described by the term
I
= Be
ik
1
x
. (307)
The total wave on the left side of the barrier is
I
=
I+
+
I
. (308)
On the right side of the barrier there can only be a wave
III+
= Fe
ik
1
x
(309)
that propagates in the +x direction at the group velocity v
III+
since region III contains
nothing that could reect the wave. Thus, we have G = 0 and
III
=
III+
= Fe
ik
1
x
. (310)
The ux of the particles that transmit through the barrier is proportional to
[
III
[
2
. (311)
Hence, the transmission probability T for a particle to pass through the barrier is the ratio
T =
[
III
[
2
[
I+
[
2
=
[F[
2
[A[
2
. (312)
Classically T = 0 because a particle with E < U
0
cannot exist inside the barrier. Lets
consider the quantum mechanical case.
In region II, the steady-state Schrodinger equation is
d
2
II
dx
2
+
2m
h
2
(E U
0
)
II
=
d
2
II
dx
2
2m
h
2
(U
0
E)
II
= 0. (313)
Since U
0
> E, the solution is
II
= Ce
k
2
x
+ De
k
2
x
. (314)
Here,
k
2
=
_
2m(U
0
E)
h
. (315)
Note that
II
does not oscillate and therefore does not represent a moving particle. However,
the probability [
II
[
2
is nor zero, so there is a nite probability of nding a particle within
the barrier. Such a particle may emerge into region III or may return to region I.
80
FIG. 25: Upper part: Reection of a particle from a potential barrier in classical mechanics. Lower
part: Reection and transmission of a particle from a potential barrier in quantum mechanics.
Applying the boundary conditions
In order to calculate the transmission probability T we have to apply the appropriate
conditions to
I
,
II
, and
III
. As discussed earlier, both and its derivative /x must
be continuous everywhere.
The boundary conditions at x = 0 are
I
(0) =
II
(0),
d
I
(0)
dx
=
d
II
(0)
dx
. (316)
The boundary conditions at x = L are
II
(L) =
III
(L),
d
II
(L)
dx
=
d
III
(L)
dx
. (317)
We substitute
I
,
II
, and
III
from Eqs. (302), (303), and (314) into the above equations.
This yields
A + B = C + D,
ik
1
(A B) = k
2
(C D), (318)
and
Ce
k
2
L
+ De
k
2
L
= Fe
ik
1
L
,
k
2
(Ce
k
2
L
De
k
2
L
) = ik
1
Fe
ik
1
L
. (319)
81
FIG. 26: At each wall of the barrier, the wave functions inside and outside it must match up
perfectly, which means that they must have the same values and slopes there.
Solving Eqs. (318) for A and B, we nd
A =
1
2
_
1 +
ik
2
k
1
_
C +
1
2
_
1
ik
2
k
1
_
D, (320)
B =
1
2
_
1
ik
2
k
1
_
C +
1
2
_
1 +
ik
2
k
1
_
D. (321)
Solving Eqs. (319) for C and D, we nd
C =
1
2
_
1
ik
1
k
2
_
Fe
(ik
1
+k
2
)L
,
D =
1
2
_
1 +
ik
1
k
2
_
Fe
(ik
1
k
2
)L
. (322)
Substituting Eqs. (322) into Eq. (320) yields
A
F
=
_
1
2
+
i
4
_
k
2
k
1
k
1
k
2
__
e
(ik
1
+k
2
)L
+
_
1
2
i
4
_
k
2
k
1
k
1
k
2
__
e
(ik
1
k
2
)L
. (323)
We assume that U
0
E. Then, we have k
2
/k
1
k
1
/k
2
and consequently
k
2
k
1
k
1
k
2
=
k
2
k
1
. (324)
We also assume that the barrier is wide enough for
II
to be severely weakened between
x = 0 and x = L. This means that k
2
L 1 and therefore
e
k
2
L
e
k
2
L
. (325)
Hence Eq. (323) can be approximated by
A
F
=
_
1
2
+
ik
2
4k
1
_
e
(ik
1
+k
2
)L
. (326)
82
It follows from the above equation that
[A[
2
[F[
2
=
_
1
4
+
k
2
2
16k
2
1
_
e
2k
2
L
. (327)
We recall that the transmission probability is
T =
[F[
2
[A[
2
. (328)
Therefore, we have
T =
[F[
2
[A[
2
=
16
4 + (k
2
/k
1
)
2
e
2k
2
L
. (329)
From the denition of k
1
, Eq. (304), and k
2
, Eq. (315) we see that
_
k
2
k
1
_
2
=
U
0
E
E
=
U
0
E
1. (330)
This quantity varies much less with E and U
0
than does the exponential. Therefore, a
reasonable approximation of the transmission probability is
T = e
2k
2
L
. (331)
Exercise 1: Electrons with energies of 1 eV and 2 eV are incident on a barrier 10 eV
high and 0.5 nm wide. (a) Find their respective transmission probabilities. (b) How are
these aected if the barrier is doubled in width?
Answer:
(a) For the 1-eV electrons
k
2
=
_
2m(U
0
E)
h
=
_
(2)(9.1 10
31
kg) [(10 1) eV] (1.6 10
19
J/eV)
1.054 10
34
J s
= 1.54 10
10
m
1
. (332)
Since L = 0.5 nm = 5 10
10
m, we have 2k
2
L = (2)(1.5410
10
)(510
10
) = 15.4. Hence,
the approximate transmission probability is
T
1
= e
2k
2
L
= e
15.4
= 2.05 10
7
. (333)
Thus one 1-eV electron out of 4.9 million can tunnel through the 10-eV barrier on average.
For the 2-eV electrons a similar calculation gives T
2
= 5.15 10
7
. The probability T
2
is
over twice the probability T
1
.
83
(b) If the barrier is doubled in width to 1 nm, the transmission probabilities become
T
1
= 4.2 10
14
and T
2
= 26.5 10
14
. (334)
Clearly, T is more sensitive to the width of the barrier than to the particle energy here.
Exercise 2: An electron and a proton with the same energy E approach a potential
barrier whose height U
0
is greater than E. Do they have the same probability of getting
through? If not, which has the greater probability?
Answer: the electron has the smaller mass, the smaller wave number k
2
=
_
2m(U
0
E)/ h inside the barrier, and consequently the greater transmission probability
T = e
2k
2
L
.
84