LAPLACE TRANSFORMS
Laplace Transform
Let f (t) be a function whose domain includes (0, ) then the Laplace trans-
form of f (t) is:
Z
L(f (t))(s) = f (t)est dt
0
127
128 CHAPTER 5. LAPLACE TRANSFORMS
Solution:
c
1 st
Z
st
L(1)(s) = 1e dt = lim e
0 c s 0
In order for this limit to exist, we must insist that s 6= 0 and that s > 0 so
that esc has a limit (of zero). When s > 0, we obtain
1 1
lim (esc 1) =
s c s
So
1
L(1)(s) = ; s > 0.
s
Solution: Z
L(t)(s) = test dt
0
We integrate by Parts (letting u = t and dv = est dt) to obtain:
1 1
Z
test dt = test 2 est ,
s s
so Z c
st 1 st 1 st
te dt = lim te 2 e
0 c s s
0
In order for this limit to exist, we again must insist that s 6= 0 and that s > 0
so that esc has a limit (of zero). We obtain
1 1
lim (cesc 0) 2 lim (esc 1)
s c s c
which exists for s > 0 and after LHopitals rule yields
1
L(t)(s) = ; s > 0.
s2
Example 5.3 Show that if f (t) = tn , for any positive integer n then
n
L(tn )(s) = L(tn1 )(s)
s
Solution: Z
n
L(t )(s) = tn est dt
0
n
We integrate by Parts (letting u = t and dv = est dt) to obtain:
1 n st 1
Z Z
n st
t e dt = t e + ntn1 est dt,
s s
so c
1 n st n
Z
n st n1
t e dt = lim t e + s L(t )(s).
0 c s 0
In order for this limit to exist, we again must insist that s 6= 0 and that s > 0
so that esc has a limit (of zero). Using LHopitals rule n times, we see that
1 n sc
lim c e = 0.
c s
Proof: Z
L(af (t) + bg(t))(s) = [af (t) + bg(t)]est dt
0
Z Z
= af (t)est dt + bg(t)est dt = aF (s) + bG(s)
0 0
From this result we can take the Laplace transform of any arbitrary poly-
nomial in the next example,
Solution:
6 2 1
=4 +8 7
s4 s3 s
24 16 7
= + 3 ,
s4 s s
where s > 0.
Proof: Z
at
L(e )(s) = eat est dt
0
Z
= e(as)t dt,
0
1
L(eat )(s) = , s>a
sa
t
Solution: Rewriting 2t = eln2 = etln2 and taking a = ln 2,
1
L(e(ln2)t )(s) = , s > ln 2
s ln 2
Lastly,
We derive the second formula and leave the derivation of the first formula
as an exercise. By definition:
Z
L(sin(t))(s) = est sin(t) dt.
0
So moving the term with the integral on the right side to the left side
gives Z 2Z
st
e sin(t) dt + 2
est sin(t) dt = ( 2 )
0 s 0 s
Factoring,
2
Z
1+ 2 est sin(t) dt = ( 2 )
s 0 s
So
( s2 )
Z
st
e sin(t) dt = = 2
0 1+ 2
s + 2
s2
5.1. INTRODUCTION AND DEFINITION 133
|f (t)| M et
for all t > 0. Then the Laplace transform of f (t) exists with a domain of at
least s > .
Proof: Suppose that there are numbers and M so that
|f (t)| M et
Z
M
est M et dt = M L[et ](s) = ,
0 s
for s > .
Therefore, since Z b
est |f (t)| dt
0
is increasing in b and bounded for all b,
Z b
lim est |f (t)| dt
b 0
Note that most all exponential functions, polynomials, and the trig func-
2
tions sine and cosine satisfy this condition but ln x, tan x and et do not.
These functions do not have Laplace transforms.
Exercises
1. Compute L[f (t)](s) for f (t) = 0.
5. Compute L[sinh(t)](s),
In 7-12, take the Laplace transform of the function f (t) using the results
in this section (do not derive using the definition)
7. f (t) = t3 7t2 + 8
ez +ez
11. f (t) = cosh(t), where is a constant. (Recall cosh(z) = 2
).
ez ez
12. f (t) = sinh(t), where is a constant. (Recall sinh(z) = 2
).
In 13-15, use the definition to take the Laplace transform of the function
f (t) using integration by parts and formulas given in this section
16. Recall that L[f (t) + g(t)] = L[f (t)] + L[g(t)]. Is it true in general that
L[f (t) g(t)] = L[f (t)] L[g(t)]?
(a) Try f (t) = t and g(t) = 1 to see if L[f (t) g(t)] = L[f (t)] L[g(t)].
(b) Is it ever true that L[f (t) g(t)] = L[f (t)] L[g(t)]?
136 CHAPTER 5. LAPLACE TRANSFORMS
Proof: By definition,
Z
0
L[y (t)](s) = est y 0 (t) dt
0
obtain Z
0 st
L[y (t)](s) = e y(t) 0
(s)est y(t) dt
0
For s big enough, limc esc y(c) = 0, since the Laplace transform of
y(t) exists (this follows since the continuous integrand of a convergent
improper integral must tend to zero we will not prove this fact here.)
So we obtain
L[y 0 (t)](s) = y(0) + sL[y(t)](s)
1
L[y 00 (t)](s) =
s2
so
1
s2 L[y(t)](s) sy(0) y 0 (0) =
s2
Solving for L[y(t)](s) we obtain
1 y(0) y 0 (0)
L(y(t))(s) = + + 2
s4 s s
Now if we could reverse engineer which function y(t) has Laplace trans-
form equal to the right hand side, then we would be done.
By playing a bit, we can see that y(t) = 16 t3 + y(0)
2
+ y 0 (0)t has this
Laplace transform. So voila! We have solved the differential equation.
(Note that the general solution obtained from undetermined coefficients
would be y(t) = 61 t3 + C1 + C2 t which agrees with our solution since
y(0) and y 0 (0) are unspecified constants.)
The above example illustrates a common checklist for solving DEs using
the Laplace transform:
we see that
L1 [aF (s) + bG(s)] = af (t) + bg(t)
or
L1 [aF (s) + bG(s)] = aL1 [F (s)] + bL1 [G(s)],
so the Laplace transform of a linear combination is the linear combi-
nation of the Laplace transforms.
The following table lists several inverses which come from the previous
section:
F (s) L1 [F (s)]
1
s
1
n! n
sn+1
t
1
sa
eat
s
s2 + 2
cos t
s2 + 2
sin t
aF (s) + bG(s) af (t) + bg(t)
7
Example 5.9 Find L1 s3
5.2. LAPLACE TRANSFORMS, THE INVERSE LAPLACE TRANSFORM, AND ODES139
Solution:
1 7 1 1
L 3
= 7L
s s3
2
We multiply the inner fraction by 2
to obtain the form in the previous
table.
1 2 7 1 2
= 7L = L
2s3 2 s3
7
= t2
2
2s1
Example 5.10 Find L1 s2 +4
Solution:
1 2s 1 1 s 1 1
L = 2L L
s2 + 4 s2 + 4 s2 + 4
2
Again, the inner fraction of the second must be multiplied by 2
to put
it in the form in the table ( = 2 since 2 = 4).
1 2
= 2 cos 2t L
2(s2 + 4)
1 1 2
= 2 cos 2t L
2 s2 + 4
1
= 2 cos 2t sin 2t
2
Example 5.11 Solve y 00 (t) + 9y(t) = e4t with y(0) = 1 and y 0 (0) = 0
140 CHAPTER 5. LAPLACE TRANSFORMS
1
s2 L[y(t)] sy(0) y 0 (0) + 9L[y(t)] =
s4
1
(s2 + 9)L[y(t)] = +s
s4
1 s
L[y(t)] = 2
+ 2
(s 4)(s + 9) s + 9
We use partial fractions to decompose the first term into a sum of terms
that are recognizable inverses.
Noting that by partial fractions:
1 A Bs + C
2
= + 2 ,
(s 4)(s + 9) s4 s +9
1 1 4
where A = 25
B = 25 C = 25 . Substituting in, we obtain
1 1 1 s 4 1 s
L[y(t)] = 2
2
+
25 s4 25 s +9 25 s +9 s2 +9
1 1 24 s 4 1
L[y(t)] = + 2
2
25 s4 25 s +9 25 s +9
So
1 4t 24 4
y(t) = e + cos 3t sin 3t
25 25 75
(note that the
3 appeared
in the demoninator since = 3 and in order
1
to find L1 2 we multiply the inner fraction by 33 .)
s +9
L[y (n) (t)](s) = sn L[y(t)](s) sn1 y(0) sn2 y 0 (0) ... sy (n2) (0) y (n1) (0)
Exercises
In 1-7, find the Inverse Laplace transform of the function F (s).
You may need partial fractions
1
1. F (s) = s+2
1
2. F (s) = s4
s+5
3. F (s) = s2 +16
4
4. F (s) = s+6
1
5. F (s) = (s+1)(s2 +1)
s
6. F (s) = (s+1)(s2 +1)
3
7. F (s) = s3 +s
8. Use the Laplace transform to find the general solution of
y 00 + 6y 0 + 5y = t
11. Use the Laplace transform to find the solution of the IVP
y 00 + y = 2, y(0) = 1, y 0 (0) = 1
12. Use the Laplace transform to find the solution of the IVP
y 00 + 2y 0 + y = 2, y(2) = 1, y 0 (2) = 1
(Hint: find the general solution first).
142 CHAPTER 5. LAPLACE TRANSFORMS
Effectively, this says L[eat f (t)](s) is equal to L[f (t)](w), (using w for the
traditional s in the definition) and replacing w = s a.
Proof: By definition:
Z
at
L[e f (t)] = eat est f (t) dt
0
Z
= e(sa)t f (t) dt.
0
Also by definition,
Z
L[f (t)](s a) = e(sa)t f (t) dt.
0
s+2
= .
s2 + 4s + 13
L1 [F (s 8)] = e8t t4 .
s
Example 5.15 Find L1 s2 +2s+5
s s s
= 2 =
s2 + 2s + 5 s + 2s + 1 + 4 (s + 1)2 + 4
We also need an s + 1 in the numerator to use the backwards shift with
cosine, so we write the numerator as:
s+11 s+1 1
= =
(s + 1)2 + 4 (s + 1)2 + 4 (s + 1)2 + 4
Now we can take the inverse of each term using the backward shift
theorem to get
1 s 1 s+1 1 1
L =L L
s2 + 2s + 5 (s + 1)2 + 4 (s + 1)2 + 4
5.3. ADVANCED PROPERTIES OF THE LAPLACE TRANSFORM 145
1
= et cos 2t et sin 2t
2
Differentiation Theorem
Let F (s) be the Laplace transform of f (t). Then
d
[F (s)] = L[tf (t)](s)
ds
or
d
[F (s)] = L[tf (t)](s)
ds
Proof: By definition:
F (s + s) F (s)
F 0 (s) = lim
s0 s
Note that
Z Z Z
F (s+s)F (s) = e(s+s)t f (t) dt est f (t) dt = [e(s+s)t est ]f (t) dt
0 0 0
So
F (s + s) F (s) e(s+s)t est
Z
= f (t) dt
s 0 s
Therefore
F (s + s) F (s) e(s+s)t est
Z
lim = lim f (t) dt
s0 s 0 s0 s
146 CHAPTER 5. LAPLACE TRANSFORMS
e(s+s)t est d
Noting that lims0 s
is the definition of ds
(est ), we know
this limit is test .
Substituting, we obtain
Z
0
F (s) = (test )f (t) dt.
0
By the Theorem,
L[tf (t)](s) = F 0 (s).
1
Here f (t) = e2t , so F (s) = s2
.
Now
F 0 (s) = (s 2)2
so
1
L[te2t ](s) = ((s 2)2 ) =
(s 2)2
Note that we could have also obtained this answer by using the Forward
Exponential Shift Theorem.
By the Theorem,
L[tf (t)](s) = F 0 (s).
s
Here f (t) = cos(3t), so F (s) = s2 +9
.
Now
(s2 + 9) s(2s) 9 s2
F 0 (s) = =
(s2 + 9)2 (s2 + 9)2
5.3. ADVANCED PROPERTIES OF THE LAPLACE TRANSFORM 147
so
9 s2 s2 9
L[t cos(3t)](s) = =
(s2 + 9)2 (s2 + 9)2
so
2
s2 L sy(0) y 0 (0) + 6sL 6y(0) + 10L =
s2 +4
(here L is shorthand for L[y(t)](s)), so
2
(s2 + 6s + 10)L = sy(0) + y 0 (0) + 6y(0) +
s2 + 4
Solving for L,
2
sy(0) + y 0 (0) + 6y(0) + s2 +4
L=
s2 + 6s + 10
sy(0)
(s + 3)2 + 1
which we can rewrite as
sy(0) (s + 3 3)y(0) (s + 3)y(0) 3y(0)
2
= 2
= .
(s + 3) + 1 (s + 3) + 1 (s + 3) + 1 (s + 3)2 + 1
2
148 CHAPTER 5. LAPLACE TRANSFORMS
We can invert this using the shift theorem, so the inverse of the first
term is:
y 0 (0) + 6y(0)
(s + 3)2 + 1
which has inverse transform
2 As + B Cs + D
= 2 + 2
(s2 + 4)(s2+ 6s + 10) s +4 s + 6s + 10
1 1 1
where A = 15 ,B = 15 , C = 15 , D = 13 . (The reader should be familiar
with the method of Partial Fractions, normally covered in a second
semester calculus class).
We again can easily invert the first term
As + B s B 2
2
=A 2 +
s +4 s +4 2 s2 + 4
which inverts to
B
A cos 2t + sin 2t.
2
The second term, after again setting up for the shifting theorem by
completing the square, becomes
Cs + D C(s + 3 3) D
= +
s2 + 6s + 10 2
(s + 3) + 1 (s + 3)2 + 1
(s + 3) 1
=C 2
+ (D 3C)
(s + 3) + 1 (s + 3)2 + 1
Taking the inverse, we obtain:
5.3. ADVANCED PROPERTIES OF THE LAPLACE TRANSFORM 149
Putting it altogether,
1 1
cos 2t + sin 2t.
15 30
1 1
y(t) = C1 e3t cos t + C2 e3t sin t cos 2t + sin 2t.
15 30
The previous example could (and probably should) have been solved
using other methods. The next example is one where these other meth-
ods would not apply.
y 00 + ty 0 2y = 2, y(0) = 0, y 0 (0) = 0
150 CHAPTER 5. LAPLACE TRANSFORMS
2
L[y 00 ] + L[ty 0 ] 2L[y] = ,
s
d 2
s2 L[y] (L[y 0 ]) 2L[y] = ,
ds s
d 2
s2 L[y] (sL[y]) 2L[y] = ,
ds s
d
Using the product rule, note ds
(sL[y]) = sL0 [y] + L[y] (recall: L[y] is
also a function of s). So,
2
s2 L[y] sL0 [y] L[y] 2L[y] = ,
s
or
2
(s2 3)L[y] sL0 [y] = ,
s
This is a first order linear differential equation in L with variable s!
0 3 2
L [y] + s + L[y] = ,
s s2
s2 s2
This has integrating factor e 2 +3 ln s which simplifies to s3 e 2 , so
s2
s3 e 2 ( s22 ) ds + C
R
L[y] = s2
s3 e 2
This resolves to
s2
2e 2 + C
s2
s3 e 2
2
We take C = 0 and obtain L[y] = s3
(all other choices do not give valid
outputs of Laplace transforms).
5.3. ADVANCED PROPERTIES OF THE LAPLACE TRANSFORM 151
Which implies that y(t) = t2 solves the DE. (One may easily check
that, indeed y(t) = t2 does solve the DE/IVP.
Exercises
In 1-8, solve the ODE/IVP using Laplace Transform
1. y 00 + 4y 0 + 3y = 0, y(0) = 1, y 0 (0) = 0
2. y 00 + 4y 0 + 3y = t2 , y(0) = 1, y 0 (0) = 0
3. y 00 3y 0 + 2y = sin t, y(0) = 0, y 0 (0) = 0
4. y 00 3y 0 + 2y = et , y(0) = 1, y 0 (0) = 0
5. y 00 2y = t2 , y(0) = 1, y 0 (0) = 0
6. y 00 4y = e2t , y(0) = 0, y 0 (0) = 1
7. y 00 + 3ty 0 y = 6t, y(0) = 0, y 0 (0) = 0
8. y 00 +ty 0 3y = 2t, y(0) = 0, y 0 (0) = 1 (You will need integration
by parts or use technology)