Laplace Transforms

Imagine you are handed a differential equation — an object built from derivatives, rates of change, and unknown functions — and told to solve it. That usually means integration, guessing forms, and juggling constants. The Laplace transform offers a radically different route: it converts the whole differential equation into an ordinary algebra problem. You solve the algebra with the same skills you learned in high school, then translate the answer back. It is, in the most literal sense, a machine for turning calculus into algebra.

This one idea powers control systems, electrical engineering, signal processing, and mechanical vibration analysis. Once you internalize why it works — that differentiation in the original world becomes multiplication in the transformed world — the Laplace transform stops feeling like a table to memorize and starts feeling like a genuinely elegant tool.

Learning Objectives

• Define the Laplace transform as an integral operator and compute it for basic functions.
• Explain why the transform turns differentiation into multiplication, and therefore turns ODEs into algebraic equations.
• Use the transform to solve initial-value problems, including the role of initial conditions.
• Apply the inverse transform, especially via partial fractions, to recover solutions.
• Understand the historical motivation from Laplace's probability work and Heaviside's operational calculus.

Quick Answer

The Laplace transform of a function $f(t)$ is defined by $\mathcal{L}\{f(t)\} = F(s) = \int_0^\infty e^{-st} f(t)\, dt.$ Its defining superpower is the derivative rule: $\mathcal{L}\{f'(t)\} = sF(s) - f(0)$. Because differentiation becomes multiplication by $s$ (plus initial-condition terms), a linear differential equation in $t$ becomes an ordinary algebraic equation in $s$. You solve that algebra for $F(s)$, then apply the inverse transform — usually with partial fractions and a table — to recover $f(t)$. Initial conditions are baked in from the start, so no arbitrary constants need to be chased down at the end.

Where It Came From

The story begins not with differential equations but with probability. In the late 1700s and early 1800s, Pierre-Simon Laplace was studying generating functions — sums and integrals that packaged an entire sequence or distribution into a single function. His integral $\int_0^\infty e^{-st} f(t)\,dt$ appeared as a tool for manipulating these probability generating functions and for approximating integrals. Laplace was not primarily trying to solve differential equations; he was trying to tame infinite processes and asymptotics.

The transform's rise to fame as a differential-equation solver came a century later through Oliver Heaviside (1850–1925), a self-taught English engineer working on telegraph and telephone signal transmission. Heaviside faced a practical crisis: the equations governing current and voltage on long cables were differential equations that were painful to solve by classical methods, and engineers needed fast, repeatable answers. He invented an "operational calculus" in which the derivative operator $\tfrac{d}{dt}$ was treated as an algebraic symbol $p$ that you could multiply, divide, and even take reciprocals of. He would write $\tfrac{1}{p}$ to mean integration and manipulate expressions with breathtaking boldness.

The catch: Heaviside could not rigorously justify his methods, and mathematicians of his era were scandalized. His famous retort — "Shall I refuse my dinner because I do not fully understand the process of digestion?" — captured his engineer's pragmatism. It fell to mathematicians in the early 20th century (notably Bromwich, Carson, and others) to prove that Heaviside's operational shortcuts were exactly the Laplace transform in disguise, giving the operator $p$ the rigorous meaning of the variable $s$. So the tool we use today is a marriage: Laplace's integral supplied the rigor, and Heaviside's engineering need supplied the purpose.

The Transform: A Machine That Eats Functions

The Laplace transform takes a function of time $t$ and produces a function of a new variable $s$:

$F(s) = \int_0^\infty e^{-st} f(t)\, dt.$

The factor $e^{-st}$ decays rapidly, which is what makes the integral converge for a wide class of functions (those that don't grow faster than an exponential). The variable $s$ can be thought of as a "complex frequency."

Worked example — the transform of $1$. Let $f(t) = 1$ for $t \ge 0$. Then

$F(s) = \int_0^\infty e^{-st} \cdot 1 \, dt = \left[ -\frac{1}{s} e^{-st} \right]_0^\infty = 0 - \left(-\frac{1}{s}\right) = \frac{1}{s}, \quad s > 0.$

Worked example — the transform of $e^{at}$. Let $f(t) = e^{at}$. Then

$F(s) = \int_0^\infty e^{-st} e^{at}\, dt = \int_0^\infty e^{-(s-a)t}\, dt = \frac{1}{s-a}, \quad s > a.$

Notice a beautiful pattern already: exponentials in time become simple poles (fractions) in $s$. Here is a starter table you will lean on constantly.

$f(t)$	$F(s) = \mathcal{L}\{f(t)\}$
$1$	$\dfrac{1}{s}$
$t^n$	$\dfrac{n!}{s^{n+1}}$
$e^{at}$	$\dfrac{1}{s-a}$
$\sin(\omega t)$	$\dfrac{\omega}{s^2 + \omega^2}$
$\cos(\omega t)$	$\dfrac{s}{s^2 + \omega^2}$

The transform is linear: $\mathcal{L}\{af(t) + bg(t)\} = aF(s) + bG(s)$. This follows immediately from the linearity of the integral, and it means you can transform complicated combinations piece by piece.

Why Differentiation Becomes Multiplication

This is the heart of everything. Apply the definition to $f'(t)$ and integrate by parts:

$\mathcal{L}\{f'(t)\} = \int_0^\infty e^{-st} f'(t)\, dt = \left[ e^{-st} f(t) \right]_0^\infty + s\int_0^\infty e^{-st} f(t)\, dt.$

The boundary term at $\infty$ vanishes (thanks to $e^{-st}$), and at $0$ it gives $-f(0)$. So:

$\mathcal{L}\{f'(t)\} = sF(s) - f(0).$

Applying this again gives the second derivative rule:

$\mathcal{L}\{f''(t)\} = s^2 F(s) - s f(0) - f'(0).$

Look at what happened: the operation of differentiating in time-world turned into multiplying by $s$ in $s$-world, with the initial conditions attached as small correction terms. A derivative — a calculus operation — has become multiplication, an algebra operation. That is the whole trick that lets us convert differential equations into algebraic ones, and it is precisely the rigorous version of Heaviside's operator $p$.

Solving Initial-Value Problems: The Three-Step Recipe

The method always follows the same shape:

1. Transform every term of the differential equation, using the derivative rules to fold in the initial conditions.
2. Solve algebraically for $F(s)$.
3. Invert $F(s)$ back to $f(t)$, usually via partial fractions and the table.

Worked example. Solve $y' + 3y = 0$ with $y(0) = 2$.

Step 1 — transform. Using $\mathcal{L}\{y'\} = sY(s) - y(0)$:

$sY(s) - 2 + 3Y(s) = 0.$

Step 2 — solve for $Y(s)$.

$(s+3)Y(s) = 2 \quad\Rightarrow\quad Y(s) = \frac{2}{s+3}.$

Step 3 — invert. From the table, $\dfrac{1}{s-a} \leftrightarrow e^{at}$ with $a = -3$:

$y(t) = 2e^{-3t}.$

You can check: $y(0) = 2$ ✓ and $y' + 3y = -6e^{-3t} + 6e^{-3t} = 0$ ✓. Notice there was no arbitrary constant to determine at the end — the initial condition entered in Step 1.

A second-order worked example. Solve $y'' + 4y = 0$ with $y(0) = 1$, $y'(0) = 0$ (undamped oscillation).

Step 1. Using the second-derivative rule:

$\big(s^2 Y(s) - s\cdot 1 - 0\big) + 4Y(s) = 0.$

Step 2.

$(s^2 + 4) Y(s) = s \quad\Rightarrow\quad Y(s) = \frac{s}{s^2 + 4}.$

Step 3. From the table, $\dfrac{s}{s^2 + \omega^2} \leftrightarrow \cos(\omega t)$ with $\omega = 2$:

$y(t) = \cos(2t).$

Exactly the oscillation we expect from a mass on a spring released from rest.

The Inverse Transform and Partial Fractions

Most real problems produce an $F(s)$ that is a ratio of polynomials but not yet in table form. The workhorse technique is partial fraction decomposition: break the ratio into simple pieces each matching a table entry.

Worked example. Solve $y' - y = e^{2t}$ with $y(0) = 0$.

Step 1. $\mathcal{L}\{e^{2t}\} = \dfrac{1}{s-2}$, so:

$sY(s) - 0 - Y(s) = \frac{1}{s-2}.$

Step 2.

$(s-1)Y(s) = \frac{1}{s-2} \quad\Rightarrow\quad Y(s) = \frac{1}{(s-1)(s-2)}.$

Step 3 — partial fractions. Write

$\frac{1}{(s-1)(s-2)} = \frac{A}{s-1} + \frac{B}{s-2}.$

Multiplying through: $1 = A(s-2) + B(s-1)$. Set $s = 2$: $1 = B(1)$, so $B = 1$. Set $s = 1$: $1 = A(-1)$, so $A = -1$. Therefore

$Y(s) = \frac{-1}{s-1} + \frac{1}{s-2}.$

Inverting term by term:

$y(t) = -e^{t} + e^{2t}.$

Check: $y(0) = -1 + 1 = 0$ ✓, and $y' - y = (-e^t + 2e^{2t}) - (-e^t + e^{2t}) = e^{2t}$ ✓.

Real-World Applications

• Electrical circuits. Kirchhoff's laws for circuits with resistors, inductors, and capacitors produce differential equations. Transforming them turns the whole circuit into an algebraic network of "impedances" in $s$, which is exactly how engineers analyze filters and transient response.
• Control systems. The "transfer function" of a system — the backbone of control engineering — is literally the Laplace transform of its output over its input. Stability is read off directly from the location of poles in the $s$-plane.
• Mechanical vibrations. Suspension systems, buildings under earthquake loading, and machinery with damping are all governed by second-order ODEs solved cleanly by transforms, especially when forces switch on suddenly (step and impulse inputs).
• Signal processing. The Laplace transform generalizes the Fourier transform and is used to design and analyze how systems respond across frequencies.

Common Mistakes

1. Forgetting the initial-condition terms. Students write $\mathcal{L}\{y'\} = sY(s)$ and drop the $-y(0)$. Why it's wrong: those boundary terms come directly from integration by parts and carry the initial data. Correction: always write the full rules $sY - y(0)$ and $s^2Y - sy(0) - y'(0)$, even when a condition is zero.

2. Confusing $\mathcal{L}\{f(t)g(t)\}$ with $F(s)G(s)$. The transform of a product is not the product of transforms. Why it's wrong: the integral of a product doesn't factor that way. Correction: the product $F(s)G(s)$ corresponds to a convolution $\int_0^t f(\tau)g(t-\tau)\,d\tau$, not to $f(t)g(t)$.

3. Botching partial fractions with repeated or complex roots. For a repeated factor $(s-a)^2$ you need both $\dfrac{A}{s-a}$ and $\dfrac{B}{(s-a)^2}$; for irreducible quadratics you need a numerator $Bs + C$. Why it's wrong: a single term can't represent the full behavior. Correction: match the decomposition form to the denominator's factor structure before solving for coefficients.

Comparison and Connections

The Laplace transform sits in a family of tools that all trade one domain for another.

Method	Best for	Handles initial conditions?
Laplace transform	Linear ODEs, discontinuous/impulse forcing, transients	Yes, built in
Fourier transform	Steady-state, frequency content of signals	No (assumes no transients)
Undetermined coefficients	Simple constant-coefficient ODEs by hand	Applied at the end
Variation of parameters	General forcing, variable inputs	Applied at the end

The Fourier transform is essentially the Laplace transform evaluated along the imaginary axis ($s = i\omega$), which is why Laplace is the more general tool for systems that start from rest and settle over time. Compared with classical solution methods, Laplace's advantage grows sharply when the forcing function is discontinuous — a switch flipped at $t = 5$, or a sudden hammer-blow impulse — where the Heaviside step and Dirac delta functions make the transform approach far cleaner.

Practice Questions

Recall

State the definition of the Laplace transform and the rule for $\mathcal{L}\{f''(t)\}$. Answer: $F(s) = \int_0^\infty e^{-st} f(t)\,dt$; and $\mathcal{L}\{f''(t)\} = s^2 F(s) - s f(0) - f'(0)$.

Understanding

Explain in one or two sentences why the Laplace transform converts a differential equation into an algebraic one. Answer: Because integration by parts turns differentiation into multiplication by $s$ (plus initial-condition terms), so every derivative in the equation becomes a polynomial factor in $s$, leaving an algebraic equation for $F(s)$.

Application

Solve $y' + 2y = 0$ with $y(0) = 5$. Answer: Transform: $sY - 5 + 2Y = 0 \Rightarrow Y = \dfrac{5}{s+2}$. Invert: $y(t) = 5e^{-2t}$.

Analysis

Solve $y'' + y = 0$ with $y(0) = 0$, $y'(0) = 3$, and explain physically what the solution represents. Answer: Transform: $s^2 Y - 0 - 3 + Y = 0 \Rightarrow Y = \dfrac{3}{s^2+1}$. Since $\dfrac{\omega}{s^2+\omega^2} \leftrightarrow \sin(\omega t)$ with $\omega = 1$, we get $y(t) = 3\sin t$. It represents an undamped oscillator started at equilibrium with an initial push, oscillating with amplitude 3.

FAQ

Why do we integrate from $0$ to $\infty$ and not over all time? Because most engineering problems have a definite "start" (a switch flips, a system is released) and we care about behavior for $t \ge 0$. This one-sided transform is exactly what lets initial conditions at $t = 0$ enter the algebra naturally.

What is the variable $s$, really? It is a complex "frequency" variable. Its real part measures growth or decay and its imaginary part measures oscillation. The locations in the $s$-plane where $F(s)$ blows up (its poles) tell you how the solution behaves in time.

Do I have to memorize the whole table? No. Memorize a handful — $1$, $e^{at}$, $\sin$, $\cos$, $t^n$ — and the derivative rules. Everything else you either derive or look up. Understanding the derivative rule matters far more than memorizing entries.

When does the transform fail to exist? When $f(t)$ grows faster than any exponential (for example $e^{t^2}$), the integral diverges for all $s$. For "exponentially bounded" functions — nearly everything in engineering — it converges for $s$ large enough.

How is this connected to Heaviside's operator $p$? Heaviside treated $\tfrac{d}{dt}$ as a multiplicable symbol $p$; the Laplace transform proves that this symbol behaves exactly like the variable $s$ (for zero initial conditions). Heaviside had the right intuition decades before it was made rigorous.

Quick Revision

• Definition: $F(s) = \int_0^\infty e^{-st} f(t)\,dt$.
• Linearity: $\mathcal{L}\{af + bg\} = aF + bG$.
• Key transforms: $1 \to \tfrac{1}{s}$, $e^{at} \to \tfrac{1}{s-a}$, $\sin\omega t \to \tfrac{\omega}{s^2+\omega^2}$, $\cos\omega t \to \tfrac{s}{s^2+\omega^2}$, $t^n \to \tfrac{n!}{s^{n+1}}$.
• Derivative rules: $\mathcal{L}\{f'\} = sF - f(0)$; $\mathcal{L}\{f''\} = s^2F - sf(0) - f'(0)$.
• Recipe: transform → solve algebra for $F(s)$ → invert (partial fractions + table).
• Initial conditions are built in from the start — no chasing arbitrary constants.
• Product of transforms = convolution, NOT product of functions.

Related Topics

Prerequisites

• Differential Equations overview
• Integration by parts and improper integrals (Calculus)
• Partial fraction decomposition (Algebra)

Related Topics

• First-order and second-order linear differential equations
• The Heaviside step function and Dirac delta (impulse) inputs
• Fourier analysis and transfer functions

Next Topics

• Convolution and the convolution theorem
• Systems of differential equations via matrix methods (Linear Algebra)
• Applications to control theory and circuit analysis