Region of convergence

The region of convergence (ROC) of a Laplace transform is the set of complex $s$ for which the defining integral

$X(s)={\int }_{-\infty }^{\infty }x(t)e^{-st}dt$

converges. Typically, the ROC is the range of real parts $\sigma =\mathrm{Re}(s)$ for which the integral makes sense.

The ROC matters because the Laplace transform of a signal is not determined by $X(s)$ alone — it is determined by $X(s)$ together with its ROC. Two different signals can have the same algebraic $X(s)$ but different ROCs, and the ROC distinguishes them.

Two examples that share an algebraic form

Right-sided exponential: $x_1(t)=e^{\alpha t}u(t)$.

$X_1(s)={\int }_0^{\infty }{e}^{\alpha t}{e}^{-st}dt={\int }_0^{\infty }{e}^{(\alpha -s)t}dt.$

For convergence at $\infty$, need $\mathrm{Re}(\alpha -s)<0$, i.e. $\sigma >\alpha$. The result:

$X_1(s)=\frac{1}{s-\alpha },\text{ROC}:\sigma >\alpha .$

Left-sided exponential: $x_2(t)=e^{\alpha t}u(-t)$.

$X_2(s)={\int }_{-\infty }^0{e}^{(\alpha -s)t}dt.$

For convergence at $-\infty$, need $\mathrm{Re}(\alpha -s)>0$, i.e. $\sigma <\alpha$. The result:

$X_2(s)=-\frac{1}{s-\alpha },\text{ROC}:\sigma <\alpha .$

The two transforms differ only by sign; the ROC is what distinguishes the right-sided from the left-sided signal.

General rules

• Right-sided signal (zero for $t<$ some point): ROC is to the right of all poles. $\sigma >\max \mathrm{Re}(\text{poles})$.
• Left-sided signal (zero for $t>$ some point): ROC is to the left of all poles. $\sigma <\min \mathrm{Re}(\text{poles})$.
• Two-sided signal (nonzero on both sides): ROC is a strip $\alpha <\sigma <\beta$ between two real parts. If $\alpha >\beta$, the transform doesn’t exist.

The ROC never contains a pole, because the transform diverges at poles.

Why this matters for systems

For an LTI system’s transfer function $H(s)$, the ROC tells you whether the system is causal, stable, both, or neither.

• Causal: ROC is a right half-plane to the right of all poles.
• Stable: ROC includes the imaginary axis $\mathrm{Re}(s)=0$.
• Causal and stable: ROC is a right half-plane that includes the imaginary axis, which means all poles are in the open left half-plane ($\mathrm{Re}(\text{pole})<0$). This is the standard stability criterion.

So pole locations on the s-plane determine both stability and causality. A pole at $s=-3$: fine, contributes a decaying exponential. A pole at $s=+3$: not stable in the causal sense (would need anti-causal interpretation for an ROC strip to the left). A pole at $s=0$ on the imaginary axis: marginally stable (integrator).

Where ROCs come up in practice

In the unilateral transform, every signal is causal and the ROC is implicitly a right half-plane — we don’t track it. For the bilateral transform, or for problems mixing right- and left-sided signals, the ROC has to be reported alongside the algebraic transform.

When inverting a Laplace transform with multiple poles, the ROC tells you which poles inverse-transform to right-sided exponentials and which to left-sided ones. For the standard causal case (ROC to the right of all poles), every pole inverse-transforms to a right-sided (decaying or growing) exponential, and the answer has a $u(t)$ factor.