Scattering of Particles by Barriers and Wells #%

L
R= M
E - E - V0 OP 2
L 15 - 15 - 10 OP
= M
2

(b)
MN E + E - V0 PQ MN 15 + 15 - 10 PQ
= M
L 15 - 5 OP 2

MN 15 + 5 PQ
F 3873
= G
. - 2.236 I
J
2

H 3873
. + 2.236 K
F 1637
= G
. IJ = 0.072
2

H 6109
. K
T = 1 – R = 1 – 0.072 = 0.928

(c) Here E = V0. Therefore,

R = 1 , T= 0

8.2 THE SQUARE POTENTIAL BARRIER

We now consider a one-dimensional potential barrier of finite width and height
given by
R| 0 x<0
V(x) = S V
|T 0 0 0< x <a (8.20)
x>a

V(x)

Region I Region II Region III

V0

Aeikx
Feikx
Be–ikx

x
0 a
Figure 8.4 The square potential barrier.

Such a barrier is called a square or a rectangular barrier and is shown in

Figure 8.4. Although the potential barriers in the real world do not have such
simple shapes, this idealized treatment forms the basis for the understanding of
more complicated systems and often provides a fairly good order-of-magnitude
estimate.
As in the previous section, we consider a particle of mass m incident on the
barrier from the left with energy E. As mentioned therein, according to classical
#& Quantum Mechanics: A Textbook for Undergraduates

mechanics, the particle would always be reflected back if E < V0 and would
always be transmitted if E > V0. We shall show that, quantum mechanically,
both reflection and transmission occur with finite probability for all values of E
except in some special cases.
We shall discuss the two cases, E > V0 and E < V0, separately.

Case 1: E > V0
Let us divide the whole space into three regions: Region I (x < 0), Region II
(0 < x < a) and Region III (x > a). In regions I and III the particle is free and
so the time-independent Schrödinger equation is
h 2 d 2y( x )
– = E y(x)
2m dx 2
d 2y( x ) 2 mE
or + k 2y(x) = 0, k2 = (8.21)
dx 2
h2
The general solution of this equation is

RS Ae ikx
+ Be -ikx x<0
T Fe
y(x) = ikx
+ Ge - ikx x>a
where A, B, F, G are arbitrary constants. For x < 0, the term Aexp(ikx)
corresponds to a plane wave of amplitude A incident on the barrier from the left
and the term B exp (–ikx) corresponds to a plane wave of amplitude B reflected
from the barrier. For x > a, the term F exp (ikx) corresponds to a transmitted
wave of amplitude F. Since no reflected wave is possible in this region we must
set G = 0.
In region II the Schrödinger equation is
h 2 d 2y( x )
– + V0 y = Ey
2m dx 2

d 2y 2m ( E - V0)
or + k¢2y(x) = 0, k¢2 = (8.22)
dx 2
h2
Since E > V0, the quantity k¢2 is positive. Therefore, the general solution of this
equation is
y(x) = Ceik ¢x + De–ik¢x 0<x<a
The complete eigenfunction is given by
R| Ae ikx
+ Be - ikx x<0
y(x) = S Ce ik ¢x
+ De - ik ¢x
|T Fe
0< x<a (8.23)
ikx
x>a
 Scattering of Particles by Barriers and Wells #'

The real part of the barrier eigenfunction for E > V0 is shown schematically
in Figure 8.5(a).

E > V0 (a)

E < V0 (b)

x
x=0 x=a

Figure 8.5 Schematic plots of the real parts of the barrier eigenfunctions for
(a) E > V0 and (b) E < V0.

Continuity of y(x) and dy(x)/dx at x = 0 and x = a gives

A+B = C+D (8.24)
ik (A – B) = ik¢ (C – D) (8.25)
Ceik¢a + De–ik¢a = Feika (8.26)
ik¢ (Ceik¢a – De–ik¢a) = ikFeika (8.27)
From (8.24) and (8.25) we obtain
1
A= [C(k + k¢) + D (k – k¢)] (8.28)
2k
1
B= [C(k – k¢) + D(k + k¢)] (8.29)
2k
From (8.26) and (8.27) we obtain
1
C= F(k¢ + k) ei(k – k¢)a (8.30)
2k¢
1
D= F(k¢ – k) ei(k + k¢)a (8.31)
2k¢
Dividing (8.31) by (8.30)
D k ¢ - k 2ik¢a
= e (8.32)
C k¢ + k
$ Quantum Mechanics: A Textbook for Undergraduates

Dividing (8.29) by (8.28)

FG D IJ (k + k ¢)
B
( k - k ¢) +
H CK
A
=
F DI
( k + k ¢) + G J ( k - k ¢)
H CK
On substitution for D/C from (8.32), this becomes
B ( k 2 - k ¢ 2 ) (1 - e 2ik ¢a )
= (8.33)
A ( k + k ¢)2 - ( k - k ¢)2 e 2ik ¢a
We need a similar expression for F/A. Equations (8.24) and (8.25) yield
1
C= [A(k + k¢) – B(k – k¢)]
2k ¢
Substituting in (8.30)
A(k + k¢) – B(k – k¢) = F(k + k¢)ei(k – k¢)a
F B
or (k + k¢)e i(k – k¢)a = (k + k¢) – (k – k¢)
A A
LM (k - k¢ )(1 - e ) OP (k – k¢)
2 2 2 ik ¢a

N (k + k¢) - (k - k¢) e Q
= (k + k¢) – 2 2 2 ik ¢a

Simplifying, we obtain
F 4kk ¢ e i ( k ¢- k ) a
= (8.34)
A ( k + k ¢)2 - ( k - k ¢)2 e 2ik ¢a
The reflection and transmission coefficients are, respectively,

B2 LM
4k 2 k¢ 2 OP -1
L 4 E( E - V ) OP
= M1 + 0
-1

N Q N V sin k ¢a Q
R= = 1+ 2 (8.35)
A (k - k ¢ 2 ) 2 sin 2 k ¢a 2 2

L (k k ¢a O L V sin k¢a OP
0
and -1 -1

= M1 + PQ = M1 +
2 2 2 2
F - k ¢ 2 ) 2 sin 2 0

N N 4E (E - V ) Q
T= (8.36)
A 4k 2 k¢ 2 0

It can be easily shown that, as expected,

R+T=1
Note that T is in general less than unity. This is in contradiction to the
classical result that the particle always crosses the barrier when E > V0. Here
T = 1 only when k¢a = p, 2p, 3p, .... Now, if l¢ is the de Broglie wavelength
of the particle when it is passing through the barrier, then
2p
k¢ =
l¢
 Scattering of Particles by Barriers and Wells $

Therefore T = 1 when
FG l¢ IJ ,
a=n
H 2K n = 1, 2, 3, ...

Thus, there is perfect transmission only when the thickness of the barrier is
equal to an integral multiple of half the de Broglie wavelength in the internal
region. This is analogous to the interference phenomena in the transmission of
light through thin refracting layers.
Equation (8.36) shows that

LM mV0 a 2 OP -1

N Q
T Æ 1+ as E Æ V0 (from above) (8.37)
2h 2
As E increases, T oscillates between a steadily increasing lower envelope and
unity, as shown in Figure 8.6. The dimensionless parameter mV0a2/ h 2 is
considered as a measure of the ‘opacity’ of the barrier.

1.0
0.8
0.6
0.4
0.2
0 E/V0
0 1 2 3 4 5
Figure 8.6 Variation of transmission coefficient for a square potential barrier as
a function of E/V0 for mV0 a2/ h2 = 10.

Case 2: E < V0
In region I (x < 0) and III (x > a), the Schrödinger equation and its solution remain
the same as in case 1. In region II (0 < x < a) the Schrödinger equation is

d 2y 2m (V0 - E )
– K 2y(x) = 0, K2 = (8.38)
dx 2 h2
Therefore, the eigenfunction in region II is
y(x) = Ce–Kx + DeKx 0<x<a (8.39)
The real part of the complete eigenfunction for E < V0 is shown schematically
in Figure 8.5(b).
The reflection and transmission coefficients can be immediately obtained
if we replace k¢ by iK in (8.35) and (8.36). Remembering that sinix = i sinhx,
we obtain
$ Quantum Mechanics: A Textbook for Undergraduates

L
R = M1 +
4k 2 K 2 OP -1
L 4 E (V - E) OP
= M1 + 0
-1
(8.40)
N (k 2
Q
+ K 2 ) 2 sinh 2 ( Ka ) N V sinh ( Ka) Q
2 2

L (k + K ) sinh ( Ka) O L V sinh (Ka) OP

0
and -1 -1

T = M1 + PQ = M1 +
2 2 2 2 2 2
0

N N 4 E (V - E ) Q
2 2 (8.41)
4k K 0

It is again readily verified that R + T = 1. We note that T Æ 0 in the limit

E Æ 0. Further, T is a monotonically increasing function of E and approaches

LM1 + mV a OP
0
2
-1

N 2h Q 2
as E Æ V0 (from below) (8.42)

Thus T joins smoothly to the value given in (8.37) for the case E Æ V0 from
above (see Figure 8.6).
For a broad high barrier, Ka >> 1. This is true for most cases of practical
interest. We may take sinh Ka ª exp(Ka)/2. In that case,
F 4kK I 2

Tª GH k + K JK
2 2
e–2Ka =
16 E (V0 - E )
V02
e–2Ka (8.43)

and is very small.

Further, the factor 16E(V0 – E)/V 02 varies slowly with V0 and E and is of
order unity in most cases. The exponential factor is the dominant one and varies
rapidly with V0 and E. Therefore, for order of magnitude calculation, we can
take
T ª e–2Ka (8.44)

Variable Potential Barrier

In actual physical systems the potential barrier is not simple square type—the
potential energy is variable (see for example, Figure 8.7). It is possible to have
an estimate of the transmission coefficient in the case of such potentials by
noticing that the quantity Ka is geometrically an area. If K varies with x, then

this expression can be generalized to z x2

x1
K(x) dx, where

2m (V( x ) - E )
K(x) =
h
We therefore have

T ª exp -
2 x2
h x1
LM
N z
2m(V ( x ) - E ) dx (8.45)
OP
Q
x1 and x2 are the classical turning points, i.e., the points at which V(x) = E.
The most important and striking thing to note is that the particle has a
nonzero probability of passing through a potential barrier which is completely