6.1 Definition and Existence

Definition 6.1.1 (Laplace transform). For \(f\) defined on \([0,\infty )\), \[\mathcal {L}\{f\}(s) = F(s) = \int ^{\infty }_{0}f(t)\,e^{-st}\,dt,\] for those \(s\) for which the integral converges.

Definition 6.1.2 (Exponential order). \(f\) is of exponential order \(\alpha \) if there are constants \(M\) and \(T\) with \(\left |f(t)\right |\leq M e^{\alpha t}\) for all \(t\geq T\).

Theorem 6.1.3 (Existence). If \(f\) is piecewise continuous on \([0,\infty )\) and of exponential order \(\alpha \), then \(\mathcal {L}\{f\}(s)\) converges absolutely for \(\operatorname {Re}s>\alpha \), and \(F(s)\to 0\) as \(s\to \infty \).

Proof. For \(t\geq T\) and \(\operatorname {Re}s = \sigma >\alpha \), \[\left |f(t)e^{-st}\right | \leq M e^{\alpha t}e^{-\sigma t} = M e^{-(\sigma -\alpha )t},\] and \(\int ^{\infty }_{T}Me^{-(\sigma -\alpha )t}dt = Me^{-(\sigma -\alpha )T}/(\sigma -\alpha )\) is finite. The integral over \([0,T]\) is finite because \(f\) is piecewise continuous there. Letting \(\sigma \to \infty \) in the same bound sends \(F(s)\) to zero. □

Note 6.1.4. The last clause of Theorem 6.1.3 is a useful check. A candidate transform that does not tend to zero at infinity — \(F(s)=1\), or \(F(s)=s/(s+1)\) — is not the transform of any function satisfying the hypotheses, and finding one in the middle of a calculation means an error earlier.

Example 6.1.5 (Powers, and where the gamma function enters). For \(f(t) = t^{a}\) with \(a>-1\), substitute \(u = st\) in the defining integral: \[\mathcal {L}\{t^{a}\}(s) = \int ^{\infty }_{0}t^{a}e^{-st}dt = \frac {1}{s^{a+1}}\int ^{\infty }_{0}u^{a}e^{-u}\,du = \frac {\Gamma (a+1)}{s^{a+1}},\] valid for \(s>0\). For integer \(a=n\) this is \(n!/s^{n+1}\).

Note 6.1.6. Example 6.1.5 is the gamma function of Chapter 2 doing the work, and it is worth noticing how little was required: a change of variable turned the Laplace integral into the gamma integral exactly. The restriction \(a>-1\) is the convergence condition established in Problem 1.6.9, arriving here for the same reason — the behaviour at the origin. Without the gamma function the transform of \(t^{1/2}\) could not be written down at all.

\(f(t)\) \(F(s)\) valid for
\(1\) \(1/s\) \(s>0\)
\(t^{a},\ a>-1\) \(\Gamma (a+1)/s^{a+1}\) \(s>0\)
\(e^{at}\) \(1/(s-a)\) \(s>a\)
\(\sin bt\) \(b/\left (s^{2}+b^{2}\right )\) \(s>0\)
\(\cos bt\) \(s/\left (s^{2}+b^{2}\right )\) \(s>0\)
\(\sinh bt\) \(b/\left (s^{2}-b^{2}\right )\) \(s>|b|\)
\(\cosh bt\) \(s/\left (s^{2}-b^{2}\right )\) \(s>|b|\)
\(H(t-a)\) \(e^{-as}/s\) \(s>0\)
\(\delta (t-a)\) \(e^{-as}\) all \(s\)
Table 6.1: Elementary Laplace transforms. \(H\) is the Heaviside step and \(\delta \) the Dirac delta.

Questions on this section

Stuck on something here? Ask below and it stays attached to this topic.