Dirac Delta and Distributions

Dirac delta distribution, point sources, impulses, and operational sampling properties in physical models.

What it is

The Dirac delta, denoted by $\delta(x)$, is a mathematical generalized function or distribution (introduced heuristically by Paul Dirac and rigorously formalized within distribution theory by Laurent Schwartz). Intuitively, it is described as a peaked object that vanishes everywhere except at the origin, where it becomes infinitely tall and narrow such that its total integral across the real line equals unity: $\delta(x) = 0 \quad \text{for } x \neq 0, \qquad \int_{-\infty}^\infty \delta(x) dx = 1$ Rigourously, the Dirac delta is not a classical function, but a continuous linear functional acting on a space of smooth test functions $\phi(x) \in \mathcal{D}(\mathbb{R})$ through the fundamental sampling (sifting) property: $\langle \delta, \phi \rangle = \int_{-\infty}^\infty \delta(x - x_0) \phi(x) dx = \phi(x_0)$

Why it appears in physics

It appears in physics because macroscopic physical modeling regularly involves concentrated sources of mass, charge, or momentum acting over regions or intervals so small that treating them as point-like idealizations dramatically simplifies calculation without forfeiting predictive accuracy:

• An isolated point electric charge $q$ at $\mathbf{r}_0$ has spatial charge density $\rho(\mathbf{r}) = q \delta(\mathbf{r} - \mathbf{r}_0)$.
• A point gravitational mass $m$: $\rho_m(\mathbf{r}) = m \delta(\mathbf{r} - \mathbf{r}_0)$.
• An impulsive mechanical collision (hammer blow, ball impact) transferring finite momentum $\Delta p$ over infinitesimal time $\Delta t \to 0$: $F(t) = \Delta p \cdot \delta(t - t_0)$.
• In field theory and electrodynamics, it defines Green's functions for differential operators like the Laplacian: $\nabla^2 \left(\frac{1}{|\mathbf{r} - \mathbf{r}'|}\right) = -4\pi \delta(\mathbf{r} - \mathbf{r}')$.

Visual or geometric idea

Envision a sequence of normalized ordinary functions (such as Gaussian bells or rectangular pulses) constrained to unit area that grow progressively narrower and taller: $f_\epsilon(x) = \frac{1}{\sqrt{2\pi \epsilon^2}} e^{-\frac{x^2}{2\epsilon^2}}$ As standard deviation $\epsilon \to 0$, the curve contracts into an infinitely sharp spike at $x=0$, perpetually enclosing an area of 1.

The Dirac delta represents the distributional limit of this family. When multiplied by any smooth test function $\phi(x)$ under an integral, the needle acts like an infinitesimal probe, "sifting out" the value of $\phi(x)$ precisely at the needle location while completely zeroing out the remainder of the domain.

Minimal symbolic language

• Fundamental sifting property: $\int_{-\infty}^\infty f(x) \delta(x - a) dx = f(a)$.
• Relation to Heaviside step: $\frac{d}{dx} \theta(x) = \delta(x)$, with $\theta(x) = 1$ for $x > 0$ and $0$ for $x < 0$.
• Scaling property: $\delta(a x) = \frac{1}{|a|} \delta(x)$ for non-zero scalar $a \neq 0$.
• Delta of a composite function: $\delta(g(x)) = \sum_k \frac{\delta(x - x_k)}{|g'(x_k)|}$, where $x_k$ are the simple roots of $g(x)$ ($g(x_k)=0, g'(x_k)\neq 0$).
• Derivative of the delta distribution: $\int_{-\infty}^\infty f(x) \delta'(x - a) dx = -f'(a)$.
• Fourier transform pair: $\mathcal{F}\{\delta(t - t_0)\} = e^{-i \omega t_0}$, with dual $\mathcal{F}\{1\} = 2\pi \delta(\omega)$.
• Fourier integral representation: $\delta(x) = \frac{1}{2\pi} \int_{-\infty}^\infty e^{i k x} dk$.
• 3D spherical delta: $\delta(\mathbf{r}) = \frac{\delta(r)}{4\pi r^2}$.

What the result means

Encountering a Dirac delta in a physical result signifies strict spatial or temporal localization of an extensive quantity. In electrostatics, a charge density containing $\delta(x - x_0)$ indicates an idealized surface charge sheet or line charge, producing a step jump discontinuity in the electric field via Gauss's law: $\Delta E = \sigma / \epsilon_0$.

In the frequency domain, a Fourier spectrum featuring a delta peak $\delta(\omega - \omega_0)$ denotes a pure, infinitely sustained monochromatic harmonic component: a sinusoidal wave $\cos(\omega_0 t)$ whose spectral power density is concentrated with zero bandwidth at that exact frequency.

When not to use it

Do not attempt to multiply Dirac deltas together: nonlinear expressions like $\delta(x)^2$ or $\delta(x)\delta(y)$ with overlapping arguments are mathematically undefined in Schwartz distribution theory and produce uncontrollable infinities (a foundational hurdle in quantum electrodynamics requiring renormalization).

Moreover, never evaluate the integral of a delta centered at a jump discontinuity of the integrand without explicit regularization: expressions such as $\int \theta(x) \delta(x) dx$ are ambiguous unless an operational limit or symmetric prescription is declared.

Operational summary

Apply the sifting identity by substituting the integration variable with the point where the delta argument vanishes: $\int f(x) \delta(x - x_0) dx = f(x_0)$, verifying that the integration limits bracket $x_0$. If the argument has a scaling factor $\delta(a x + b)$, divide by $|a|$. For derivatives of the delta distribution, shift the derivative onto the test function with a minus sign: $\int f(x) \delta'(x) dx = -f'(0)$.

Procedure

Before calculating

Inspect the integration limits of the expression containing the Dirac delta: if the root $x_0$ lies outside the interval $[a, b]$, the integral is identically zero.

If the singular point coincides exactly with a boundary limit, check problem specifications or adopt the standard symmetric half-weight convention ($1/2$).

Step 1: identify the mathematical object

Analyze the argument inside the Dirac delta:

• For simple linear arguments $\delta(x - x_0)$, the sampling point is $x_0$.
• For scaled arguments $\delta(a x - b)$, factor out the coefficient: $\delta(a(x - b/a)) = \frac{1}{|a|} \delta(x - b/a)$.
• For nonlinear arguments $\delta(g(x))$, find all real roots $x_k$ where $g(x_k) = 0$ and evaluate derivatives $|g'(x_k)|$.

Step 2: choose the operation

If the delta is integrated against a continuous function: Apply the sifting property: $\int_a^b f(x) \delta(x - x_0) dx = f(x_0)$ (provided $a < x_0 < b$).

If the derivative of the delta appears: Apply distributional integration by parts: $\int f(x) \delta'(x - x_0) dx = -f'(x_0)$.

Step 3: preserve units and signs

Remember that the Dirac delta possesses physical units inverse to its argument: if $x$ is in meters, $\delta(x)$ has units of $\text{m}^{-1}$, ensuring $\int \delta(x) dx$ is dimensionless.

Maintain absolute value signs when scaling: $\delta(-2x) = \frac{1}{|-2|} \delta(x) = +\frac{1}{2} \delta(x)$; the delta is intrinsically an even distribution ($\delta(-x) = \delta(x)$).

Step 4: calculate

Substitute evaluated values $f(x_0)$ into the expression.

For composite arguments like $\delta(x^2 - a^2)$, expand into simple roots: $\delta(x^2 - a^2) = \frac{\delta(x - a) + \delta(x + a)}{2|a|}$.

Step 5: interpret the result

Confirm dimensional correctness of the final evaluated expression.

Ensure the result captures the physical meaning of total concentrated quantity (total charge, net momentum impulse, or point force).

Self-check checklist

• Did the root of the delta argument fall inside the integration bounds?
• Did you use the absolute value $|a|$ when applying the scaling property?
• Did you insert a minus sign when integrating against $\delta'$: $\int f \delta' = -f'$?
• Are the dimensional units of $\delta(x)$ equal to $[x]^{-1}$?

Guided example

Physical situation

A baseball bat strikes a ball of mass $m = 0.145\text{ kg}$ at time $t_0 = 0.5\text{ s}$. The contact force is modeled as an ideal impulsive blow: $F(t) = J_0 \delta(t - t_0)$, with transferred momentum impulse $J_0 = 6.0\text{ N}\cdot\text{s}$. The ball starts from rest ($v(0) = 0$). We seek the ball velocity $v(t)$ for all $t > 0$ and its kinetic energy following the collision.

Mathematical translation

Newton's second law dictates $m \frac{dv}{dt} = F(t) = J_0 \delta(t - t_0)$. Integrating both sides with respect to time from $t=0$ up to arbitrary time $t$: $m [v(t) - v(0)] = \int_0^t J_0 \delta(t' - t_0) dt'$.

Step-by-step calculation

1. Evaluate the delta integral across the two temporal regions:

• For $t < t_0 = 0.5\text{ s}$, the singular point $t'=0.5$ lies outside $[0, t]$, so the integral vanishes: $m v(t) = 0 \implies v(t) = 0\text{ m/s}$.
• For $t > t_0 = 0.5\text{ s}$, the singular point $t'=0.5$ falls inside $[0, t]$, so the sifting integral yields 1: $m v(t) = J_0 (1) = J_0$.

2. Calculate post-impact velocity for $t > 0.5\text{ s}$: $v(t) = \frac{J_0}{m} = \frac{6.0\text{ N}\cdot\text{s}}{0.145\text{ kg}} \approx 41.38\text{ m/s}$ (about $149\text{ km/h}$).
3. Express velocity compactly via the Heaviside step distribution: $v(t) = 41.38 \ \theta(t - 0.5)\text{ m/s}$.
4. Compute post-impact kinetic energy: $E_k = \frac{1}{2} m v^2 = \frac{1}{2}(0.145)(41.38)^2 \approx 124.1\text{ J}$.

Reading the result

The impulsive Dirac delta force causes an abrupt jump discontinuity in the velocity profile from $0$ to $41.38\text{ m/s}$ at $t = 0.5\text{ s}$, proving that the time derivative of a step velocity jump is an impulsive delta acceleration.

Common error avoided

We avoided attempting to evaluate the acceleration at $t=0.5\text{ s}$ as an ordinary finite real number, properly treating acceleration as an impulsive distribution $a(t) = \frac{J_0}{m} \delta(t - t_0)$ whose physically meaningful observable is its integrated impulse $\Delta v$.

Inverse example

Given result

The electrostatic potential throughout space due to a point-like distribution is $V(r) = \frac{Q}{4\pi \epsilon_0 r}$. Evaluating the Laplacian for $r > 0$ yields $\nabla^2(1/r) = 0$, yet the total charge enclosed in any surrounding sphere is $Q \neq 0$.

What physical question it can answer

It resolves this apparent paradox: because $\nabla^2(1/r)$ vanishes everywhere except at $r=0$ where it diverges, the distributional identity $\nabla^2\left(\frac{1}{r}\right) = -4\pi \delta(\mathbf{r})$ proves that all charge $Q$ is concentrated at a singular mathematical point at the origin: $\rho(\mathbf{r}) = Q \delta(\mathbf{r})$, establishing the rigorous basis of Coulomb's law for point charges.

Common errors

Error 1: ignoring integration limits when applying the sifting property

It seems reasonable to replace $x$ with $x_0$ without inspecting integration boundaries. It fails because the Dirac delta vanishes everywhere outside its singular point; if the interval $[a, b]$ excludes $x_0$, the integral is strictly zero. It is detected by checking whether $a < x_0 < b$. It is corrected by verifying that the support point lies inside the integration bounds before evaluating. Mini-example: $\int_1^3 x^2 \delta(x) dx = 0$ because $x_0 = 0$ is outside $[1, 3]$.

Error 2: omitting the absolute value in the scaling property delta(ax)

It seems reasonable to write $\delta(a x) = \frac{1}{a} \delta(x)$, allowing negative scalars when $a < 0$. It fails because the Dirac delta is an even distribution with normalized positive unit area: substitution $u = ax$ reverses the integration limits when $a < 0$, cancelling the minus sign and yielding $|a|$ in the denominator. It is detected if a mass or probability density becomes negative due to a negative coordinate scale factor. It is corrected by strictly applying $\delta(a x) = \frac{1}{|a|} \delta(x)$. Mini-example: $\delta(-5x) = +\frac{1}{5}\delta(x)$, never $-\frac{1}{5}\delta(x)$.

Error 3: forgetting the minus sign when integrating against delta derivatives

It seems reasonable to assume $\int f(x) \delta'(x - x_0) dx = f'(x_0)$. It fails because distributional integration by parts transfers the derivative onto the test function with an obligatory negative sign: $\int f(x) \delta'(x - x_0) dx = -f'(x_0)$. It is detected if dipole moments or force gradients acquire the opposite physical sign. It is corrected by applying the rule that each derivative of a delta contributes a factor of $(-1)^n$. Mini-example: $\int x \delta'(x) dx = -(1) = -1$, not $+1$.

Error 4: omitting Jacobian derivatives in nonlinear arguments delta(g(x))

It seems reasonable to sum function values across roots without weighting: $\int f(x) \delta(g(x)) dx = \sum f(x_k)$. It fails because near a simple zero $x_k$, the function expands linearly as $g(x) \approx g'(x_k)(x - x_k)$, so argument scaling introduces a Jacobian factor of $\frac{1}{|g'(x_k)|}$. It is detected if physical dimensions of the output are inconsistent. It is corrected by using the general identity $\delta(g(x)) = \sum_k \frac{\delta(x - x_k)}{|g'(x_k)|}$. Mini-example: $\delta(2x - 4) = \frac{1}{2}\delta(x - 2)$; omitting the $1/2$ factor erroneously doubles the result.

Error 5: attempting to square or multiply Dirac deltas directly

It seems reasonable to compute quadratic energy densities by writing $\rho^2 = q^2 \delta(\mathbf{r})^2$. It fails because products of distributions with overlapping singular supports are ill-defined in Schwartz distribution theory and produce unresolvable infinities. It is detected whenever $\delta(0)$ or $\delta(x)^2$ appears in analytical calculations. It is corrected by regularizing the point source as a smooth distribution of finite radius $\epsilon$ and taking the physical limit only on macroscopic observables. Mini-example: the electrostatic self-energy of a point charge $\int E^2 dV$ diverges to infinity due to this singular product.

Quick detection rule

If you evaluate a delta whose root lies outside the integration interval, if you produce a negative factor when scaling $\delta(-k x)$, or if you omit the minus sign in $\int f \delta' = -f'$, your distribution calculus contains an error.

Physical applications

1. Classical electrostatics and point source charge distributions

It appears when solving Poisson's equation $\nabla^2 V = -\frac{\rho}{\epsilon_0}$ for localized point charges, line charges, and surface boundary sheets. It is calculated by modeling a point charge $q$ at the origin as $\rho(\mathbf{r}) = q \delta(\mathbf{r})$, allowing Maxwell's equations to apply uniformly across all space within a single differential framework. The result reproduces Coulomb's law and proves that the electric field normal component undergoes a step discontinuity across a charged sheet: $\Delta E_n = \frac{\sigma}{\epsilon_0}$. The dangerous error is omitting the factor of $4\pi$ in the field divergence $\nabla \cdot (\hat{r}/r^2) = 4\pi \delta(\mathbf{r})$, which would erroneously predict zero net electric flux from a point charge. As a final check, the student must verify Gauss's law by integrating $\rho(\mathbf{r})$ over an enclosed sphere.

2. Classical mechanics and instantaneous collision impulses

It appears when analyzing sudden mechanical impacts: hammer strikes, tennis ball impacts, or elastic particle rebounds against rigid surfaces. It is calculated by modeling the contact force as a Dirac delta impulse $F(t) = \Delta p \cdot \delta(t - t_0)$. The result determines post-collision velocities immediately following impact without requiring microscopic modeling of material contact deformations. The dangerous error is computing mechanical work by integrating force along distance across a discontinuous velocity profile without regularization, causing ambiguous mathematical infinities. As a final check, the student must verify the impulse-momentum theorem $\int_{t_0^-}^{t_0^+} F(t) dt = \Delta p$.

3. Green's functions and impulse response in linear physical systems

It appears in acoustics, seismology, and linear circuit theory when characterizing system responses to instantaneous point impulses: $L\{G(t, t')\} = \delta(t - t')$, where $G$ is the Green's function. It is calculated by exciting the system with a delta input and computing the impulse response $h(t)$. The result yields the system output under any arbitrary dynamic driving force $f(t)$ via a convolution integral: $y(t) = \int_{-\infty}^\infty G(t, t') f(t') dt'$. The dangerous error is omitting causality constraints ($G(t, t') = 0$ for $t < t'$), which would violate the physical principle that effects cannot precede causes. As a final check, the student must determine the impulse response of a damped harmonic oscillator subject to a delta force.

4. Quantum mechanics and continuum state orthogonality

It appears in Dirac's bra-ket formulation of quantum mechanics for eigenstates belonging to the continuous spectrum, such as position $|x\rangle$ and momentum $|p\rangle$ states. It is calculated by evaluating the continuous inner product: $\langle x | x' \rangle = \delta(x - x')$ and $\langle p | p' \rangle = 2\pi\hbar \delta(p - p')$. The result provides the completeness resolution of the identity $\int |x\rangle \langle x| dx = \hat{I}$, allowing arbitrary state wavefunctions to be expanded in the continuous coordinate basis. The dangerous error is treating $|x\rangle$ as square-integrable vectors in standard Hilbert space, forgetting that continuous states are non-normalizable distributions normalized exclusively to delta functions. As a final check, the student must derive the canonical commutation relation $[\hat{x}, \hat{p}] = i\hbar \hat{I}$ in the position representation.

5. Fourier optics and point spread function (PSF) in imaging systems

It appears when modeling optical image formation in telescopes, microscopes, and cameras. It is calculated by representing an idealized point light source (such as a distant star) as a two-dimensional spatial Dirac delta $\delta(x, y)$. The resulting image profile is the optical system's Point Spread Function (PSF), which due to aperture diffraction broadens into an Airy disk pattern. The dangerous error is assuming that two closely spaced point emitters can always be separated without accounting for Rayleigh's resolution criterion, which sets the diffraction limit where overlapping PSFs merge into an unresolved blur. As a final check, the student must explain how convolving the camera's PSF with an original scene describes image blur.

Practice

Level 1: recognition

Evaluate the following integrals containing Dirac delta distributions using the sifting property: a) $\int_{-\infty}^\infty (x^3 - 2x + 5) \delta(x - 2) dx$ b) $\int_0^5 \cos(\pi x) \delta(x - 1) dx$ c) $\int_2^{10} e^{-x^2} \delta(x) dx$ (inspect integration limits carefully)

Exercise: For case c), explain why the integral is identically zero without evaluating the exponential.

Level 2: direct application

Use the scaling and derivative identities of the delta distribution to solve: a) $\int_{-\infty}^\infty \sin(x) \delta(3x - \pi) dx$ b) $\int_{-\infty}^\infty (x^2 + 4x) \delta'(x - 1) dx$

Exercise: For case b), confirm the minus sign introduced when transferring the derivative onto the polynomial.

Level 3: choosing the tool

A linear mass density along the $x$-axis is given by $\lambda(x) = M_0 \delta(x) + \lambda_0 [\theta(x) - \theta(x - L)]$, describing a concentrated point mass $M_0$ at the origin attached to a uniform rod of length $L$ and density $\lambda_0$.

Exercise: Find the total mass $M_{\text{tot}} = \int_{-\infty}^\infty \lambda(x) dx$ and compute the center-of-mass position $x_{\text{cm}} = \frac{1}{M_{\text{tot}}} \int_{-\infty}^\infty x \lambda(x) dx$.

Level 4: physical interpretation

In classical electrodynamics, a point charge $q$ moving along trajectory $\mathbf{r}_p(t)$ with velocity $\mathbf{v}(t)$ produces charge and current densities: $\rho(\mathbf{r}, t) = q \delta(\mathbf{r} - \mathbf{r}_p(t))$, $\mathbf{J}(\mathbf{r}, t) = q \mathbf{v}(t) \delta(\mathbf{r} - \mathbf{r}_p(t))$.

Exercise: Substitute these into the continuity equation $\nabla \cdot \mathbf{J} + \frac{\partial \rho}{\partial t} = 0$ and prove that local conservation of electric charge is rigorously satisfied for a point particle.

Level 5: intentional error

A student evaluates $\int_{-\infty}^\infty \delta(x^2 - 4) e^x dx$ by declaring: "Since $x^2 - 4 = 0$ at $x = 2$, the integral simply evaluates to $e^2$".

Exercise: Identify the two major mistakes: omitting the negative root $x = -2$ and neglecting the derivative Jacobian factor $|g'(x_k)| = |2x_k|$ which introduces a factor of $1/4$ at each root. Carry out the correct calculation using $\delta(g(x)) = \sum \frac{\delta(x - x_k)}{|g'(x_k)|}$ to obtain the true numerical value.