The Laplace Transform, Its Properties & Region of Convergence
Complex frequency, damping, and the ROC that makes it all converge
What if a signal grows without bound — can we still take its transform? The Laplace transform says yes, by trading pure oscillation for complex frequency.
From Fourier to Laplace: Adding Damping
In Module 10, the Fourier transform decomposed signals into pure complex exponentials ejωt along the imaginary axis. But many real signals — capacitor charging curves, unstable system responses, signals with initial conditions — grow exponentially or fail to be absolutely integrable, leaving the Fourier integral undefined.
The Laplace transform solves this by introducing a real damping factor σ into the exponent: instead of ejωt, we use e−st where s = σ+jω. The factor e−σt can tamp down exponential growth, making the integral converge for a much broader class of signals.
The Complex Frequency Variable s = σ + jω
The substitution s=σ+jω means e−st=e−σte−jωt: a decaying (or growing) envelope e−σt multiplied by a rotating phasor e−jωt. When σ=0, we recover the Fourier kernel e−jωt exactly.
This is why the Laplace transform is a generalization: every signal that has a Fourier transform also has a Laplace transform, but the converse is not true. Signals like e2tu(t) — which blows up — still have a Laplace transform (with ROC Re(s)>2), even though their Fourier transform does not exist.
Region of Convergence: Why It Matters
The Laplace integral ∫₀^∞x(t)e−stdt converges only for certain values of s. This set is called the Region of Convergence (ROC). Without specifying the ROC, the transform is ambiguous — two different signals can share the same X(s) expression but have different ROCs.
For the standard pair above, convergence requires Re(s)>−a: the real part of s must be large enough to overcome the exponential rate a. The special case a=0 gives the unit step: u(t)↔1/s with ROC Re(s)>0.
Stating X(s)=1/(s+3) is incomplete without its ROC. The same rational function could correspond to a right-sided signal e−3tu(t) (ROC: Re(s)>−3) or a left-sided signal −e−3tu(−t) (ROC: Re(s)<−3). Always specify the ROC alongside X(s).
Key Properties: Linearity and Differentiation
The Laplace transform inherits linearity from the integral: L{a·x(t)+b·y(t)}=a·X(s)+b·Y(s), with an ROC at least the intersection of the individual ROCs.
The differentiation property is where Laplace truly earns its keep. It converts a time derivative into multiplication by s, minus an initial-condition term. This is the mechanism that turns differential equations — the language of circuits, Maxwell's equations, and mechanical systems — into simple algebra.
Because differentiation maps to multiplication by s (plus an initial-condition term), an Nth-order linear ODE in time becomes an Nth-degree polynomial equation in s. You solve the algebra, then invert — no guessing of homogeneous and particular solutions required. We will do exactly this in the next lesson.
Final Value Theorem and the Fourier Connection
Two more results connect Laplace back to steady-state behavior and to the Fourier transform you already know.
The Final Value Theorem (FVT) gives the steady-state value of x(t) directly from X(s): lim(t→∞)x(t)=lim(s→0)s·X(s), valid when the ROC of sX(s) includes the origin (i.e., the system is stable). We will use this as a sanity check in the next lesson.
The Fourier connection is immediate: if the ROC of X(s) includes the jω-axis (Re(s)=0), then the Fourier transform is simply X(s) evaluated at s=jω. If the ROC does not include the jω-axis — for instance, a growing exponential — the Fourier transform does not exist, but the Laplace transform still does.
- Write out the Laplace integral: X(s) = ∫₀^∞ e^(-3t)e^(-st) dt.
- Combine exponents: X(s) = ∫₀^∞ e^(-(s+3)t) dt.
- Evaluate the integral: X(s) = [−1/(s+3)] · e^(-(s+3)t) evaluated from 0 to ∞.
- At t→∞, e^(-(s+3)t) → 0 when Re(s+3) > 0, i.e., Re(s) > −3.
- At t=0, e^0 = 1, so X(s) = 0 − (−1/(s+3)) = 1/(s+3).
- Start with the transform pair: X(s) = 1/(s+5), ROC: Re(s) > -5.
- Evaluate at s = 0: X(0) = 1/(0+5) = 1/5.
- Since Re(0) = 0 > -5, s = 0 is inside the ROC, so the evaluation is valid.
- X(0) = 1/5 = 0.2.
Check your understanding
- The Laplace transform X(s) = ∫₀^∞ x(t)e^(-st)dt generalizes the Fourier transform by using s = σ + jω, adding real damping σ to imaginary frequency jω.
- The ROC is the set of s-values where the integral converges; for e^(-at)u(t) ↔ 1/(s+a), the ROC is Re(s) > -a.
- The differentiation property L{dx/dt} = sX(s) − x(0⁻) converts ODEs into algebraic equations in s.
- The Fourier transform is the Laplace transform evaluated on the jω-axis: X(jω) = X(s)|_{s=jω}, valid when the ROC includes the imaginary axis.
- The Final Value Theorem gives lim(t→∞) x(t) = lim(s→0) sX(s), provided the ROC of sX(s) includes the origin.