Dirac delta function by John S.

Colton, Feb 2008

Definitions
1. Definition as limit. The Dirac delta function can be thought of as a rectangular pulse that
grows narrower and narrower while simultaneously growing larger and larger.
rect(x, b) =
(x) = lim(b0) rect(x, b)
ote that the integral of the delta function is the area under the cur!e, and has been held constant
at 1 throughout the limit process.
1 ) (

x
Shifting the origin. "ust as a parabola can be shifted away from the origin by writing y = (x # x
0
)
$
instead of %ust y = x
$
, any function can be shifted by plugging in x # x
0
in place of its usual
argument x.
(x - x
0
) =
&hifting the position of the pea' doesn(t affect the total area if the integral is ta'en from # to .
1 ) (
0

x x
Disclaimer) *athematicians will ob%ect that the Dirac delta function defined this way (or any
other way, for that matter) is not a real function. That is true, but physicists recogni+e that for all
practical purposes you really can %ust thin' of the delta function as a !ery large pea'.
1
x
y
width = b
height = 1,b
(so total area = 1)
x
y
x
0
$. Definition as deri!ati!e of step function. The step function, also called the -.ea!iside step
function/ is usually defined li'e this)
h(x) = 0 for x 0 0,
h(x) = 1 for x 1 0
h(x) = 0.2 for x = 0
3t(s a function whose only feature is a step up from 0 to 1, at the origin)
4hat(s the deri!ati!e of this function5 4ell, the slope is +ero for x 0 0 and the slope is +ero for
x 1 0. 4hat about right at the origin5 The slope is infinite6 &o the deri!ati!e of this function is a
function which is +ero e!erywhere e7cept at the origin, where it(s infinite8and the integral of
the deri!ati!e function from # to must be 1 because h(x) is the anti9deri!ati!e and has a !alue
of 1 at x = and a !alue of 0 at x = # (thin' :undamental Theorem of ;alculus).
<n infinite pea' at the origin whose integral is 15 &ound familiar5 Therefore (x) can also be
defined as the deri!ati!e of the step function.
=. Definition as :ourier transform. 4e(!e seen that ta'ing the :ourier transform of a function
gi!es you the fre>uency components of the function. 4hat about ta'ing the :ourier transform of
a pure sine or cosine wa!e oscillating at
0
5 There is only one fre>uency component, so the
:ourier transform must be a single, !ery large pea' at
0
(or possibly two pea's, one at
0
and
one at #
0
). < delta function6
?. Definition as density. 4hat(s the density of a 1 'g point mass located at the origin5 4ell, it(s
a function that must be +ero e!erywhere e7cept at the origin8and it must be infinitely large at
the origin because for a mass that truly occupies only a single point, the mass must ha!e been
infinitely compressed. .ow about the integral5 The integral of the density must gi!e you the
mass, which is 1 'g. < function that is +ero e!erywhere e7cept at the origin, and has an integral
e>ual to 15 &ounds li'e the delta function again6
*ore precisely, this would be a three9dimensional analog to the regular delta function (x),
because the density must be integrated o!er three dimensions in order to gi!e the mass. This is
sometimes written (r) or as
=
(r))

=
(r) = (x) (y) (+)
$
x
y
1
Properties
1. 3ntegral. @ne of the most important properties of the delta function has already been
mentioned) it integrates to 1.
$. &ifting property. 4hen a delta function (x # x
0
) multiplies another function f(x), the product
must be +ero e!erywhere e7cept at the location of the infinite pea', x
0
. <t that location, the
product is infinite li'e the delta function, but it might be a larger or smaller infinity (now you see
why mathematicians don(t li'e physicists), depending on whether the !alue of f(x) at that point is
larger or smaller than 1. 3n other words, the area of the product function is not necessarily 1 any
more, it is modified by the !alue of f(x) at the infinite pea'. This is called the -sifting property/
of the delta function)

) ( ) ( ) (
0 0
x f dx x f x x
*athematicians may call the delta function a -functional/, because it is really only well9defined
inside integrals li'e this, in terms of what it does to other functions. (< -functional/ is something
that operates on functions, rather than a -function/ which is something that operates on
!ariables.)
=. &ymmetry. < few other properties can be readily seen from the definition of the delta function)
a. (#x) = (x) (ote that (x) beha!es as if it were an e!en function)
b. (x # x
0
) = (#x A x
0
)
?. Binear systems. 3f a physical system has linear responses and if its response to delta functions
(-impulses/) is 'nown, then in theory the output of this system can be determined for almost any
input, no matter how comple7. This rather ama+ing property of linear systems is a result of the
following) almost any arbitrary function can be decomposed into (or -sampled by/) a linear
combination of delta functions, each weighted appropriately, and each of which produces its own
impulse response. Thus, by application of the superposition principle, the o!erall response to the
arbitrary input can be found by adding up all of the indi!idual impulse responses.
Fourier Transforms
1. < single delta function. The :ourier transform of a single delta function in time can be
obtained using the sifting property and the definition of the :.T.)
{ }

$
1
$
1
) (
$
1
) (
) 0 (

i t i
e dt e t t FT
=
&imilarly, the :ourier transform of (t # t
0
) is)
{ }
) (
0 0
0
$
1
) (
$
1
) (
t i t i
e dt e t t t t FT

{ }
0
$
1
) (
0
t i
e t t FT

The :ourier transform has a constant amplitude but a phase that !aries linearly with .
<nother interesting way to define the delta function can be obtained by doing the re!erse :ourier
Transform and setting it e>ual to the original delta function)

,
_

'

d e e e FT
t i t i t i
0 0
$
1
$
1
$
1
1

d e t t
t t i ) (
0
0
$
1
) (
ote) the negati!e sign in the e7ponent is optional, as can be seen by substituting in = #) the
d becomes #d, but the limits of integration are re!ersed.
$. :inite comb function. < series of e>ually spaced delta functions, sometimes called the comb
function, is written as)

N
N n
n
t t t comb ) ( ) (
where t
n
= n t
0
(note that comb is also a function of N)
@ne potential use for the comb function might be to perform a sampling operation on another
function (see -Binear systems,/ abo!e).
The :ourier transform of the comb function is
{ }

N
N n
t i t i
N
N n
n
n
e dt e t t t comb FT

$
1
) (
$
1
) (
:or e7ample, if N = 1, the :ourier transform is)
{ } ( ) ( ) ) cos( $ 1
$
1
1
$
1
) 1 , (
0
0 0
t e e N t comb FT
t i t i

+ + +

?
x
y
C C
t
0
$ A 1 total delta funtions
:or arbitrary N, the :ourier transform is gi!en by
{ }

N
N n
t i
n
e N t comb FT

$
1
), (
which, after summing the finite series and doing lots of algebra, reduces to)
{ }

,
_

,
_

$
sin
$
) 1 $ ( sin
$
1
), (
0
0
t
t
N
N t comb FT

This is a periodic function made up of large primary pea's surrounded by secondary pea's that
decrease in amplitude as you mo!e away from the primary pea'.
D ? $ $ ? D
2
2
1 0
1 2
$ 0

otice that the function is periodic and repeats e!ery $.
<s can easily be found from B(.ospital(s Eule, the amplitude of the primary pea' depends on )
amplitude = ($A1),s>rt($)
The first +ero (a measure of the width of the primary pea') is gi!en by
0
) 1 $ (
$
t N
zero first
+

=. 3nfinite comb function) Fsing the pre!ious two e>uations, one can see that in the limit as
, the large pea's (li'e the ones at #$, 0, and $) get infinitely tall and infinitely narrow.
They are delta functions6 Thus, the :ourier transform of an infinite comb function in the time
domain, is a similar infinite comb function in the fre>uency domain.
2
:T(comb) plotted here as a
function of t
0
, for =10,
without the 1,s>rt($)