Appendix C

Tutorial on the Dirac delta

function and the Fourier
transformation

C.1 Dirac delta function

The delta function δ(x) studied in this section is a function that takes on zero values at all x 6= 0,
R
+∞
and is infinite at x = 0, so that its integral δ(x)dx = 1. This function allows one to write down
−∞
spatial density of a physical quantity that is concentrated in one point. For example, the density of a
one-dimensional particle of mass m located at x = a is written as mδ(x − a). In quantum mechanics,
we use δ(x) to write, for example, the wave function of a state with a well-defined position.
The delta function belongs to the class of so-called generalized functions. This means that it is
meaningful only as a part of an integral expression. While we may write a delta function outside
of an integral, we always keep in mind that it will eventually become a part of an integral, and
only then will it produce a valid result that can be used, for example, to predict an outcome of an
experiment. It is not possible to provide a rigorous mathematical theory of generalized functions in
the framework of this course. Below, we discuss only those properties of the delta function that are
useful for physicists.
Definition C.1 The Dirac delta function is a generalized function such that for any function f (x)
that is smooth1 and takes a value of 0 at ±∞
Z
+∞

δ(x)f (x)dx = f (0) (C.1)

−∞

Exercise C.1 Show that

a)
Z
+∞

δ(x)dx = 1; (C.2)
−∞

b) for any function f (x),

Z
+∞

δ(x − a)f (x)dx = f (a); (C.3)

−∞

1A smooth function is one that has derivatives of all finite orders.

128 A. I. Lvovsky. Quantum Mechanics I

c) for any real number a,

δ(ax) = δ(x)/|a|. (C.4)
n
0 ifx<0
Exercise C.2 For the Heaviside step function θ(x) = 1 ifx≥0 show that

d
θ(x) = δ(x). (C.5)
dx
Hint: use Eq. (C.1) with any smooth f (x) vanishing at ±∞.

Note C.1 The above result can be generalized to any function g(x) that has a discontinuity at
the point x = x0 :
¯
dg(x) dg(x) ¯¯
= + [g(x0 + 0) − g(x0 − 0)]δ(x − x0 ). (C.6)
dx dx ¯x6=x0

Definition C.2 Function

2
/b2
Gb (x) = e−x , (C.7)

is called the Gaussian function.

0 x

2
/b2
Figure C.1: Gaussian function e−x .

Exercise C.3 Show that the integral of the Gaussian function

Z
+∞
2 2 √
e−x /b dx = b π (C.8)
−∞

R
+∞ 2 √
(Hint: use e−x dx = π.)
−∞

Note C.2 The delta function can be visualized as a Gaussian function b√1 π Gb (x) of infinitely
√
narrow width b. The factor 1/(b π) is chosen to make the function’s integral equal to 1. We can
write
1 2 2
√ e−x /b → δ(x) for b → 0. (C.9)
b π

Exercise C.4 Show that for a smooth function f (x) which takes zero values at ±∞,

+∞·
Z ¸ ¯
d df ¯¯
δ(x) f (x)dx = − . (C.10)
dx dx ¯x=0
−∞
C.2. FOURIER TRANSFORMATION 129

Note C.3 Because the delta function is meaningful only as a part of an integral expression, Eq.
(C.10) can be rewritten as follows:
· ¸
d df
δ(x) f (x) = −δ(x) , (C.11)
dx dx
d
In other words, the expression dx δ(x) can be seen as an operator acting on functions:

d d
δ(x) = −δ(x) . (C.12)
dx dx

C.2 Fourier transformation

This is an important integral transformation used in all branches of physics. It is used, for example,
to determine the frequency spectrum of a time-dependent signal.
Definition C.3 The Fourier transform f˜ ≡ F [f ] of a function f (x) is a function of parameter k
defined as follows:
Z
+∞
˜ 1
f (k) = √ e−ikx f (x)dx. (C.13)
2π
−∞

R
+∞
Note C.4 f˜(0) = √1
2π
f (x)dx
−∞

Exercise C.5 For a real f (x), f˜(k) = f˜∗ (−k).

Exercise C.6 Show that the Fourier transform of a Gaussian function from Ex. C.3 is a Gaussian
function
2 2 b 2 2
F[e−x /b ] = √ e−k b /4 . (C.14)
2
Exercise C.7 a) Show that in the limit b → 0, Eq. (C.14) takes the form
1
F[δ(x)] = √ . (C.15)
2π

b) Show that in the opposite limit, b → ∞, one obtains

√
F[1] = 2π δ(k). (C.16)

[Hint: use Eq. (C.9).]

Exercise C.8 Show that
Z
+∞

eik0 x dx = 2πδ(k0 ) (C.17)

−∞

Note C.5 The above equation can be straightforwardly extended:

Z
+∞

eiak0 x dx = 2πδ(k0 )/|a|. (C.18)

−∞

Definition C.4 The inverse Fourier transform F −1 [g] of a function g(k) is a function of parameter
x such that
Z
+∞
−1 1
ğ(x) = F [g](x) = √ eikx g(k)dk. (C.19)
2π
−∞
130 A. I. Lvovsky. Quantum Mechanics I

Exercise C.9
F −1 [F[f ]](x) = f (x). (C.20)
Exercise C.10
F −1 [f (x)] = F[f (−x)]. (C.21)
Exercise C.11 Show that, if f˜(k) = F[f (x)] exists, then
a)
1 ˜
F[f (ax)] = f (k/a); (C.22)
|a|

b)
F[f (x − a)] = e−ika f˜(k); (C.23)

c)
F[eiξx f (x)] = f˜(k − ξ). (C.24)

d)
F[df (x)/dx] = ik f˜(k). (C.25)

Write similar rules for the inverse Fourier transformation.

Exercise C.12 Find the Fourier transform of δ(x + a) + δ(x − a).
Definition C.5 The convolution of two functions is the integral

Z
+∞

[f ∗ g](x) = f (x − y)g(y)dy (C.26)

−∞

Exercise C.13 Show that the definition of convolution is symmetric, i.e. f ∗ g = g ∗ f .

Exercise C.14 Show that any function is a convolution of itself with the delta function.

Exercise C.15 Show that, for any two functions f (x) and g(x),
a) √
F [f ∗ g] = 2πF [f ] × F [g]; (C.27)

b)
1
F [f × g] = √ F [f ] ∗ F [g]. (C.28)
2π
Exercise C.16 Verify the above result explicitly for two Gaussian functions Ga (x) and Gb (x).