Chapter 6

Laplace Transforms

6.1 Introduction

The operaion of differentiation transforms a given function to another function:

D[f (x)] = f ′ (x).

Differentiation is an example of a transform. Another example is that of integration:

Z x
I[f (x)] = f (s) ds.
0

A transform is a mapping from one set of functions to another:

T : F1 → F2 .

For the differerential operator D, we have F1 = {all differentiable functions}, whereas for the integral oper-
ator I we might have F1 = {all continuous functions}.

A useful tool for solving differerential equations is that of an integral transform, particularly on of the form:
Z β
T [f ](s) = K(x, s)f (x) dx.
α

The function K(x, s) is called the kernel, and the transform is usually denoted as F = T [f ]. Several
transforms have been found useful: Laplace, Fourier, Mellin, Hankel, . . . , others. We will only consider
Laplace transforms.

Definition 6.1. The Laplace transform of a function f is defiend as

Z ∞
L[f ](s) := e−sx f (x) dx,
0

provided the integral exists. ◭

82
Math 334 6.1. INTRODUCTION 83

Remark. This is an improper integral, so it doesn’t converge for just any old function f . ◭

We will use Laplace transforms to solve ODEs in the following manner. First take the Laplace transform of
the ODE to get an algebraic equation in Y = L[y]. This algebraic equation is linear in Y , and so is eaasily
solved. Then take the inverse Laplace transform to get the solution to the original equation. The procedure
is represented by the schematic in Figure 6.1.

Solution of ODEs by Laplace transform

Laplace
linear ODE transform algebraic equation
F(x,y,y’,y’’)=0 G(s,Y)=0

solve

inverse Laplace
solution transform solution
y(x) Y(s)

Figure 6.1: A schematic of the solution of ODEs by Laplace transforms

Consider the behaviour of a function f near a point x0 , in particular consider the following limits:

f (x+
0 ) := lim+ f (x), f (x−
0 ) := lim− f (x).
x→x0 x→x0

Definition 6.2. If
(i) f (x+ −
0 ) = f (x0 ) = f (x0 ), then f is continuous at x0 ;
+
(ii) f (x0 ) = f (x− 0 ) 6= f (x0 ), then f has a removable discontinuity at x0 ;
(iii) f (x+ −
0 ) 6= f (x0 ), then f has a jump discontinuity at x0 ;
+
(iv) either f (x0 ) or f (x− 0 ) does not exist, then f has a “bad” discontinuity at x0 . ◭

Definition 6.3. A function f is piecewise continuous (or sectionally continuous) on (a, b) if

(i) f is continuous for x ∈ (a, b) except possibly at a finite number of points;
(ii) f (x+0 ) exists for all x ∈ [a, b);
(iii) f (x−0 ) exists for all x ∈ (a, b]. ◭

Remark. A function f is piecewise continuous on an interval (a, b) if f has at most a finite number of
Math 334 6.1. INTRODUCTION 84

discontinuities, none of which is worse than a jump discontinuity. ◭

Definition 6.4. A function f is piecewise continuous on an infinite interval (a, ∞) if f is piecewise

continuous on every finite subinterval (a, b), for b > a. ◭

Example 6.5. The function 

x
 06x61
f (x) = −1 1 < x 6 2

1 2<x<3


is piecewise continuous on the interval (0, 3).

Example 6.6. The functions f (x) = 1/x and g(x) = sin(1/x) are not piecewise continuous on the interval
(0, 1), since
1 1
f (0+ ) = lim+ f (x) = lim+ and g(0+ ) = lim+ g(x) = lim+ sin do not exist.
x→0 x→0 x x→0 x→0 x

Definition 6.7. P C(a, b) := {set of all piecewise continuous functions on (a, b)} ◭

The set of piecewise continuous functions P C(a, b) satisfy the following properties:
(1) f, g ∈ P C(a, b) =⇒ αf + βg ∈ P C(a, b) ∀α, β ∈ R;
(2) f, g ∈ P C(a, b) =⇒ f g ∈ P C(a, b);
Rb
(3) f ∈ P C(a, b) =⇒ a
|f (x)| dx exists.

Remark. It follows from property (1) that P C(a, b) is a vector space. ◭

Definition 6.8. A function f is said to be of exponential order if there exist constants M, x̄ and a, with
M, x̄ > 0 such that |f (x)| 6 M eax for x > x̄. ◭

Example 6.9. The functions x, e2x , and cos 3x are of exponential order since

|x| 6 ex , |e2x | 6 e2x , | cos 3x| 6 1 6 ex ,

2 2
whereas ex is not of exponential order since ex > M eax for sufficiently large x no matter what the valus of
M or a.

Theorem 6.10.
If (i) f ∈ P C(0, ∞); and (ii) f is of exponential order (i.e. |f (x)| 6 M eax for x > x̄),
then the Laplace transform L[f (x)] =: F (s) exists for s > a.

Proof
Z ∞ Z x̄ Z ∞
−sx −sx
e f (x) dx = e f (x) dx + e−sx f (x) dx =: I1 + I2 .
0 0 x̄
Math 334 6.2. PROPERTIES OF LAPLACE TRANSFORMS 85

Since f ∈ P C(0, x̄) and e−sx ∈ P C(0, x̄) for every s, it follows from properties 2 and 3 listed earlier that
e−sx f (x) ∈ P C(0, x̄) and hence I1 exists.

As for the second integral, we have

Z ∞
−sx −sx ax −(s−a)x
x > x̄ =⇒ |e f (x)| 6 e Me = Me =⇒ |I2 | 6 M e−(s−a)x dx < ∞
x̄

for s > a, and so I2 exists. The result follows.

Example 6.11. For f (x) = 1 we have

∞
1
Z
L[1] = e−sx 1 dx = , for s > 0.
0 s

For f (x) = eax we have

∞ ∞
1
Z Z
L[eax ] = e−sx eax dx = e−(s−a)x dx = , for s > a.
0 0 s−a

For f (x) = cos ax we have

Z ∞ ∞
−sx e−sx  s
L[cos ax] = e cos ax dx = 2 2
(a sin ax − s cos ax) = 2 , for s > 0.
0 a +s 0 a + s2

For f (x) = x we have

∞   ∞
x 1 1
Z

L[x] = x e−sx dx = − + 2 e−sx  = 2 , for s > 0.
0 s s 0 s

6.2 Properties of Laplace Transforms

In this section we examine some of the properties of the Laplace Transform.

Theorem 6.12.
If (i) f ∈ P C(0, ∞); (ii) f is of exponential order; and (iii) F (s) = L[f (x)], then

lim F (s) = 0.
s→∞

Proof
Exercise.

Example 6.13.
1 1
For f (x) = 1 we have F (s) = and lim F (s) = lim = 0.
s s→∞ s→∞ s
1 1
For f (x) = eax we have F (s) = and lim F (s) = lim = 0.
s−a s→∞ s→∞ s − a
s s
For f (x) = cos ax we have F (s) = 2 2
and lim F (s) = lim 2 = 0.
a +s s→∞ s→∞ a + s2
Math 334 6.2. PROPERTIES OF LAPLACE TRANSFORMS 86

If we define F1 := {f ∈ P C(0, ∞); f is of exponential order}, then we have the following:

Theorem 6.14.
The Laplace transform is a linear operator on F1 , i.e.
L[af + bg] = aL[f ] + bL[g] ∀a, b ∈ R, ∀f, g ∈ F1 .
Proof
Z ∞ Z ∞ Z ∞
−sx −sx
L[af + bg] = e (af (x) + bg(x)) dx = a e f (x) dx + b e−sx g(x) dx = aL[f ] + bL[g].
0 0 0

Example 6.15. Find the Laplace transform of f (x) = 7 − 3e2x .

Solution
7 3
L[f (x)] = L[7 − 3e2x ] = 7L[1] − 3L[e2x ] = − .
s s−2

Example 6.16. Find the Laplace transform of cosh ax and sinh ax.
Solution
eax + e−ax
 
1 ax −ax

1 1 1 s
L[cosh ax] = L[ ]= L[e ] + L[e ] = + = 2 ,
2 2 2 s−a s+a s − a2
eax − e−ax
 
1
1 1 1 a
L[sinh ax] = L[ ]= L[eax ] − L[e−ax ] = − = 2 .
2 2 2 s−a s+a s − a2

Theorem 6.17.
If L[f (x)] = F (s), then the following hold:
1. L[eax f (x)] = F (s − a);
1 s
2. L[f (ax)] = F ( );
a ′a
3. L[xf (x)] = −F (s);
4. L[fZ ′ (x)] = sF (s) − f (0);
x
1
5. L[ f (t) dt] = F (s).
0 s
Proof Z ∞ Z ∞
ax −sx ax
1. L[e f (x)] = e e f (x) dx = e−(s−a)x f (x) dx = F (s − a).
0 0
∞ ∞
1 1 s
Z Z
sξ
2. L[f (ax)] = e−sx f (ax) dx = e− a f (ξ) dξ = F ( ).
0 a 0 a a
∞ ∞
d
Z Z
−sx
3. L[xf (x))] = e xf (x)) dx = − e−sx f (x) dx = −F ′ (s).
0 ds 0
Z ∞ ∞ Z ∞
−sx ′ −sx

′
4. L[f (x))] = e f (x)) dx = e f (x) + s
 e−sx f (x) dx = −f (0) + sF (s).
0 0 0
Z ∞
5. Let g(x) := f (t) dt and G(s) := L[g(x)]. Then g′ (x) = f (x) and g(0) = 0. We now have
0
F (s) = L[f (x)] = L[g′ (x)] = sG(s) − g(0) = sG(s) (using property 4). Therefore
Z x
1
L[ f (t) dt] = L[g(x)] = G(s) = F (s).
0 s
Math 334 6.2. PROPERTIES OF LAPLACE TRANSFORMS 87

Remark.
1. Formula 4 can be extended as follows:

Let g(x) = f ′ (x). Then G(s) = L[g(x)] = L[f ′ (x)] = sF (s) − f (0).

This leads to

L[f ′′ (x)] = L[g′ (x)] = sG(s) − g(0) = s(sF (s) − f (0)) − f ′ (0) = s2 F (s) − sf (0) − f ′ (0).

Proceeding by induction, we get

n−1
X
L[f (n) (x)] = sn F (s) − sn−1 f (0) − sn−2 f ′ (0) − · · · − f (n−1) (0) = sn F (s) − sn−j f (j) (0).
j=0

2. In a similar way, formula 3 can be extended to

L[xn f (x)] = (−1)nF (n) (s). ◭

These properties of the Laplace transform allow us to build up a table of Laplace transforms without having
to actually integrate.

Example 6.18.
 
d d 1 1
1. L[x] = L[x · 1] = − L[1] = − = 2 . (using property 3)
ds ds s s
Z x
1 1 1 1
2. L[x] = L[ 1 dt] = L[1] = · = 2 . (using property 5)
0 s s s s
 
2 d d 1 2
3. L[x ] = L[x · x] = − L[x] = − 2
= 3 . (using property 3)
ds ds s s
Z x
2 2 1 2
4. L[x2 ] = L[ 2t dt] = L[x] = · 2 = 3 . (using property 5)
0 s s s s
1
5. L[eax ] = L[eax · 1] = L[1](s − a) =
. (using property 1)
s−a
 
ax d ax d 1 1
6. L[x e ] = − L[e ] = − = . (using property 3)
ds ds s − a (s − a)2
1
7. L[x eax ] = L[x](s − a) = . (using property 1)
(s − a)2
s−a
8. L[eax cos bx] = L[cos bx](s − a) = . (using property 1)
b2 + (s − a)2

One situation for which Laplace transforms are particularly useful is in solving differerential equations with
discontinuous coefficients. Let (
x2 0 6 x < 2
f (x) = . (6.1)
0 x>2
Math 334 6.2. PROPERTIES OF LAPLACE TRANSFORMS 88

Calculating the Laplace transform requires integration by parts twice leading to:
Z ∞ Z 2
F (s) = L[f (x)] = e−sx f (x) dx = x2 e−sx dx
0 0
 2
 2  
x 2x 2  2 4 4 2
=− + 2 + 3 e−sx  = 3 − + 2 + 3 e−2s .
s s s 0 s s s s

It turns out that the above calculation can be simplified by using the unit step function. Define
(
0 x<a
ua (x) := .
1 x>a

Now consider the functions f (x) and f (x)ua(x) in the Figure 6.2 below.

y y

y=f(x) y=f(x)ua(x)

a x a x

Figure 6.2: A plot of y = f (x) and y = f (x)ua(x).

We can think of the unit step function ua (x) as a switch, so that f (x)ua (x) “turns on” f (x) at x = a.
Similarly, f (x)ub(x) “turns on” f (x) at x = b. Therefore f (x)[ua(x) − ub (x)] “turns on” f (x) at x = a and
“turns off” f (x) at x = b. See Figure 6.3.
( 
0 x<a
f (x)ua(x) =

 
0 x<a

f (x) x > a

 

( =⇒ f (x)[u a (x) − u b (x)] = f (x) a6x<b.
0 x<b
 
0 x>b

 
f (x)ub(x) = 

f (x) x > b


y=f(x)[ua(x)-ub(x)]

a b x

Figure 6.3: A plot of y = f (x)[ua(x) − ub (x)].

Math 334 6.2. PROPERTIES OF LAPLACE TRANSFORMS 89

Theorem 6.19.
If L[f (x)] exists, then L[f (x)ua(x)] = e−ax L[f (x + a)].

Proof

Z ∞ Z ∞
−st
L[f (x)ua (x)] = e e−stf (t) dt
f (t)ua(t) dt =
0 a
Z ∞ Z ∞
= e−s(x+a) f (x + a) dx = e−as e−sx f (x + a) dx = e−as L[f (x + a)].
0 0

Let us return to the function defined in Eq. (6.1). Using unit step functions f (x) = x2 [u0 (x) − u2 (x)].
Calculating the Laplace transform we get

L[f (x)] = L[x2 u0 (x)] − L[x2 u2 (x)] = e−0s L[(x + 0)2 ] − e−2s L[(x + 2)2 ]
 
2 4 4 2
= L[x2 ] − e−2s L[x2 + 4x + 4] = 3 − e−2s + 2+ 3 .
s s s s

This calculation was less tedious than integration by parts twice.

Another class of functions important in applications is the class of periodic funtions.

Definition 6.20. A funtion f is periodic with period p if f (x + np) = f (x) for all n ∈ Z. ◭

Theorem 6.21. p
1
Z
If (i) f ∈ F1 , and (ii) f is periodic with period p, then L[f (x)] = f (x)e−sx dx.
1 − e−ps 0

Proof
First, we have
Z (n+1)p Z p Z p
e−sx f (x) dx = e−s(t+np) f (t + np) dt = e−nps e−st f (t) dt.
np 0 0

Now, taking the Laplace transform we get

Z ∞ Z p Z 2p Z 3p
L[f (x)] = e−sx f (x) dx = e−sx f (x) dx + e−sx f (x) dx +
e−sx f (x) dx + · · ·
0 0 p 2p
∞ Z (n+1)p Z p !Z
X ∞ 
X  X ∞ p
−sx −nps −st −nps
= e f (x) dx = e e f (t) dt = e e−st f (t) dt
n=0 np n=0 0 n=0 0
p ∞
1 1
Z X
= e−sx f (x) dx. (using = tn for |t| < 1 with t = e−ps )
1 − e−ps 0 1 − t n=0

2π
Example 6.22. The function cos ax is periodic with period p = . Using the result of the previous
a
Math 334 6.2. PROPERTIES OF LAPLACE TRANSFORMS 90

theorem we get
2π    2π
1 1 e−sx
Z a
−sx
a
L[cos ax] = e cos ax dx = 2 2
(a sin ax − s cos ax) 
1−e
−2πs
a 0 1−e
−2πs
a s +a 0
( −2πs )
1 e a (−s) − (−s) s
= = 2 .
1−e
−2πs
a s2 + a2 s + a2

Example 6.23. Finf L[f (x)] for the function

(
1 0<x61
f (x) = , f (x + 2n) = f (x) ∀n ∈ Z.
−1 1 < x 6 2

Solution
The function f is periodic with period 2, so we have
Z 2 Z 1 Z 2 
1 −sx 1 −sx −sx
L[f (x)] = e f (x) dx = e dx − e dx
1 − e−2s 0 1 − e−2s 0 1
(1 − e−s )2 es/2 − e−s/2
 −2s 
1 e − 2e−s + 1 1 − e−s 1 1
= = = = = tanh( ).
1 − e−2s s s(1 − e−2s ) s(1 + e−s s(es/2 + e−s/2 ) s s