3.3 The One-Sided Laplace Transform 189
L
[
d^2 f(t)
dt^2
u(t)
]
=s^2 F(s)−sf( 0 −)−
df(t)
dt
∣
∣t= 0 − (3.12)
In general, iff(N)(t)denotes anNth-order derivative of a functionf(t)that has a Laplace transformF(s), we
have
L[f(N)(t)u(t)]=sNF(s)−
N∑− 1
k= 0
f(k)( 0 −)sN−^1 −k (3.13)
wheref(m)(t)=dmf(t)/dtmis themth-order derivative,m> 0 , andf(^0 )(t),f(t).
The Laplace transform of the derivative of a causal signal is
L
[
df(t)
dt
u(t)
]
=
∫∞
0 −
df(t)
dt
e−stdt
This integral is evaluated by parts. Letw=e−st, thendw=−se−stdt, and letv=f(t)so thatdv=
[df(t)/dt]dt, and
∫
wdv=wv−
∫
vdw
We would then have
∫∞
0 −
df(t)
dt
e−stdt=e−stf(t)
∣∣∞
0 −−
∫∞
0 −
f(t)(−se−st)dt
=s
∫∞
0 −
f(t)e−stdt−f( 0 −)
=sF(s)−f( 0 −)
wheree−stf(t)|t= 0 −=f( 0 −)ande−stf(t)|t→∞=0 since the region of convergence guarantees that
lim
t→∞
f(t)e−σt= 0
For a second-order derivative we have that
L
[
d^2 f(t)
dt^2
u(t)
]
=L
[
df(^1 )(t)
dt
u(t)
]
=sL[f(^1 )(t)]−f(^1 )( 0 −)
=s^2 F(s)−sf( 0 −)−
df(t)
dt
|t= 0 −
where we used the notationf(^1 )(t)=df(t)/dt. This approach can be extended to any higher order to
obtain the general result shown above.