1.

11 Dirac Delta Function 75

Z∞
dx
1.10.4 , for a > 0.
eax + 1
0

Z∞
sin x
1.10.5 d x.
x2
π

Z∞
e−x sin x
1.10.6 d x.
x
0

Zx
1.10.7 erf(t) dt.
0

The result can be expressed in terms of special functions in Table 1.2.

Zx
1.10.8 E 1 (t) dt.
1

Obtain a result in which the only special function is E 1 .

Z∞
e−x
1.10.9 d x.
x +1
0

Z∞ 2
tan−1 x
1.10.10 Show that d x = π ln 2.
x
0

Hint. Integrate by parts, to linearize in tan−1 . Then replace tan−1 x by tan−1 ax and
evaluate for a = 1.
1.10.11 By direct integration in Cartesian coordinates, find the area of the ellipse defined by
x2 y2
+ = 1.
a 2 b2
1.10.12 A unit circle is divided into two pieces by a straight line whose distance of closest
approach to the center is 1/2 unit. By evaluating a suitable integral, find the area of
the smaller piece thereby produced. Then use simple geometric considerations to verify
your answer.

1.11 DIRAC DELTA FUNCTION

Frequently we are faced with the problem of describing a quantity that is zero everywhere
except at a single point, while at that point it is infinite in such a way that its integral over
76 Chapter 1 Mathematical Preliminaries

any interval containing that point has a finite value. For this purpose it is useful to introduce
the Dirac delta function, which is defined to have the properties

δ(x) = 0, x 6= 0, (1.149)
Zb
f (0) = f (x) δ(x) d x, (1.150)
a

where f (x) is any well-behaved function and the integration includes the origin. As a
special case of Eq. (1.150),
Z∞
δ(x) d x = 1. (1.151)
−∞

From Eq. (1.150), δ(x) must be an infinitely high, thin spike at x = 0, as in the description
of an impulsive force or the charge density for a point charge. The problem is that no such
function exists, in the usual sense of function. However, the crucial property in Eq. (1.150)
can be developed rigorously as the limit of a sequence of functions, a distribution. For
example, the delta function may be approximated by any of the sequences of functions,
Eqs. (1.152) to (1.155) and Figs. 1.21 and 1.22:

 0, x < − 2n
 1

δn (x) = n, − 2n 1
< x < 2n
1
(1.152)

0, x > 2n ,
1


n
δn (x) = √ exp(−n 2 x 2 ), (1.153)
π

y = δn(x)

n e−n2x2
π

x x

FIGURE 1.21 δ-Sequence function: left, Eq. (1.152); right, Eq. (1.153).
 1.11 Dirac Delta Function 77

n⋅ 1 sin nx
π 1+ n2x2 πx

x x

FIGURE 1.22 δ-Sequence function: left, Eq. (1.154); right, Eq. (1.155).

n 1
δn (x) = , (1.154)
π 1 + n2 x 2
Zn
sin nx 1
δn (x) = = ei xt dt. (1.155)
πx 2π
−n

While all these sequences (and others) cause δ(x) to have the same properties, they dif-
fer somewhat in ease of use for various purposes. Equation (1.152) is useful in providing
a simple derivation of the integral property, Eq. (1.150). Equation (1.153) is convenient to
differentiate. Its derivatives lead to the Hermite polynomials. Equation (1.155) is particu-
larly useful in Fourier analysis and in applications to quantum mechanics. In the theory of
Fourier series, Eq. (1.155) often appears (modified) as the Dirichlet kernel:

1 sin[(n + 21 )x]
δn (x) =
. (1.156)
2π sin 12 x

In using these approximations in Eq. (1.150) and elsewhere, we assume that f (x) is well
behaved—that it offers no problems at large x.
The forms for δn (x) given in Eqs. (1.152) to (1.155) all obviously peak strongly for
large n at x = 0. They must also be scaled in agreement with Eq. (1.151). For the forms
in Eqs. (1.152) and (1.154), verification of the scale is the topic of Exercises 1.11.1 and
1.11.2. To check the scales of Eqs. (1.153) and (1.155), we need values of the integrals

Z∞ Z∞
π
r
−n 2 x 2 sin nx
e dx = and d x = π.
n x
−∞ −∞

These results are respectively trivial extensions of Eqs. (1.148) and (11.107) (the latter of
which we derive later).
78 Chapter 1 Mathematical Preliminaries

For most physical purposes the forms describing delta functions are quite adequate.
However, from a mathematical point of view the situation is still unsatisfactory. The limits

lim δn (x)
n→∞

do not exist.
A way out of this difficulty is provided by the theory of distributions. Recognizing that
Eq. (1.150) is the fundamental property, we focus our attention on it rather than on δ(x)
itself. Equations (1.152) to (1.155) with n = 1, 2, 3 . . . may be interpreted as sequences of
normalized functions, and we may consistently write

Z∞ Z
δ(x) f (x) d x ≡ lim δn (x) f (x) d x. (1.157)
n→∞
−∞

Thus, δ(x) is labeled a distribution (not a function) and is regarded as defined by

Eq. (1.157). We might emphasize that the integral on the left-hand side of Eq. (1.157)
is not a Riemann integral.8

Properties of δ(x)
• From any of Eqs. (1.152) through (1.155) we see that Dirac’s delta function must be
even in x, δ(−x) = δ(x).
• If a > 0,
1
δ(ax) = δ(x), a > 0. (1.158)
a
Equation (1.158) can be proved by making the substitution x = y/a:

Z∞ Z∞
1 1
f (x)δ(ax) d x = f (y/a)δ(y) dy = f (0).
a a
−∞ −∞

If a < 0, Eq. (1.158) becomes δ(ax) = δ(x)/|a|.

• Shift of origin:

Z∞
δ(x − x0 ) f (x) d x = f (x0 ), (1.159)
−∞

which can be proved by making the substitution y = x − x0 and noting that when y = 0,
x = x0 .

8 It can be treated as a Stieltjes integral if desired; δ(x) d x is replaced by du(x), where u(x) is the Heaviside step function
(compare Exercise 1.11.9).
 1.11 Dirac Delta Function 79

• If the argument of δ(x) is a function g(x) with simple zeros at points ai on the real axis
(and therefore g 0 (ai ) 6 = 0),
  X δ(x − a )
i
δ g(x) = . (1.160)
|g 0 (ai )|
i
To prove Eq. (1.160), we write
Z∞ ai +ε
X Z  
f (x)δ(x) d x = f (x)δ (x − ai )g 0 (ai ) d x,
−∞ i ai −ε

where we have decomposed the original integral into a sum of integrals over small in-
tervals containing the zeros of g(x). In these intervals, we replaced g(x) by the leading
term in its Taylor series. Applying Eqs. (1.158) and (1.159) to each term of the sum,
we confirm Eq. (1.160).
• Derivative of delta function:
Z∞ Z∞
f (x)δ (x − x0 ) d x = −
0
f 0 (x)δ(x − x0 ) d x = − f 0 (x0 ). (1.161)
−∞ −∞

Equation (1.161) can be taken as defining the derivative δ 0 (x); it is evaluated by per-
forming an integration by parts on any of the sequences defining the delta function.
• In three dimensions, the delta function δ(r) is interpreted as δ(x)δ(y)δ(z), so it de-
scribes a function localized at the origin and with unit integrated weight, irrespective
of the coordinate system in use. Thus, in spherical polar coordinates,
y
f (r2 )δ(r2 − r1 )r22 dr2 sin θ2 dθ2 dφ2 = f (r1 ). (1.162)
• Equation (1.155) corresponds in the limit to
Z∞
1  
δ(t − x) = exp iω(t − x) dω, (1.163)
2π
−∞
with the understanding that this has meaning only when under an integral sign. In that
context it is extremely useful for the simplification of Fourier integrals (Chapter 20).
• Expansions of δ(x) are addressed in Chapter 5. See Example 5.1.7.

Kronecker Delta
It is sometimes useful to have a symbol that is the discrete analog of the Dirac delta func-
tion, with the property that it is unity when the discrete variable has a certain value, and
zero otherwise. A quantity with these properties is known as the Kronecker delta, defined
for indices i and j as
(
1, i = j,
δi j = (1.164)
0, i 6= j.
80 Chapter 1 Mathematical Preliminaries

Frequent uses of this symbol are to select a special term from a summation, or to have one
functional form for all nonzero values of an index, but a different form when the index is
zero. Examples:
X X 1 2π
f i j δi j = f ii , Cn = .
1 + δn0 L
ij i

Exercises
1.11.1 Let
1


 0, x <− ,

 2n
1 1

δn (x) = n, − <x< ,
 2n 2n
1


0, < x.


2n
Show that
Z∞
lim f (x)δn (x) d x = f (0),
n→∞
−∞
assuming that f (x) is continuous at x = 0.
1.11.2 For
n 1
δn (x) = ,
π 1 + n2 x 2
show that
Z∞
δn (x) d x = 1.
−∞
1.11.3 Fejer’s method of summing series is associated with the function
sin(nt/2) 2
 
1
δn (t) = .
2π n sin(t/2)
Show that δn (t) is a delta distribution, in the sense that
Z∞
sin(nt/2) 2
 
1
lim f (t) dt = f (0).
n→∞ 2π n sin(t/2)
−∞
1.11.4 Prove that
1
δ[a(x − x1 )] = δ(x − x1 ).
a
Note. If δ[a(x − x1 )] is considered even, relative to x1 , the relation holds for negative
a and 1/a may be replaced by 1/ |a|.
1.11.5 Show that
δ[(x − x1 )(x − x2 )] = [δ(x − x1 ) + δ(x − x2 )]/ |x1 − x2 | .
Hint. Try using Exercise 1.11.4.
 1.11 Dirac Delta Function 81

un(x) n large
n→∞
1

n small

FIGURE 1.23 Heaviside unit step function.

n 2 2
1.11.6 Using the Gauss error curve delta sequence δn = √ e−n x , show that
π
d
δ(x) = −δ(x),
x
dx
treating δ(x) and its derivative as in Eq. (1.157).
1.11.7 Show that
Z∞
δ 0 (x) f (x) d x = − f 0 (0).
−∞

Here we assume that f 0 (x) is continuous at x = 0.

1.11.8 Prove that
−1
d f (x)
δ( f (x)) = δ(x − x0 ),
dx x=x0
where x0 is chosen so that f (x0 ) = 0.
Hint. Note that δ( f ) d f = δ(x) d x.
1.11.9 (a) If we define a sequence δn (x) = n/(2 cosh2 nx), show that
Z∞
δn (x) d x = 1, independent of n.
−∞

(b) Continuing this analysis, show that9

Zx
1
δn (x) d x = [1 + tanh nx] ≡ u n (x)
2
−∞

and
x < 0,

0,
lim u n (x) =
n→∞ 1, x > 0.
This is the Heaviside unit step function (Fig. 1.23).

9 Many other symbols are used for this function. This is the AMS-55 notation (in Additional Readings, see Abramowitz and
Stegun): u for unit.