The Dirac Delta Function and Distributions

⚡ Key Takeaways

One-line statement: The Dirac delta $\delta(x)$ is not a function but a distribution — a continuous linear functional on test functions — that "samples" a function at a point: $\int f(x)\delta(x)\,dx = f(0)$ .

Key result in words: Generalized functions extend classical calculus so that pointwise impossible objects (infinite spikes, derivatives of jumps) become rigorous and differentiable infinitely many times.

Where it applies: Partial differential equations (Green's functions, fundamental solutions), Fourier analysis, signal processing, quantum mechanics, probability (point masses), electromagnetism.

Who/when: Heuristics by Oliver Heaviside (1890s) and Paul Dirac (1927); rigorous theory by Sergei Sobolev (1936) and Laurent Schwartz (1945–1950, Fields Medal 1950).

💡 The Core Insight

Stop asking what value a function has at a point. Instead ask what number it produces when integrated against every smooth probe. A distribution is fully determined by these "measurements." The delta is the measurement that reads off $\varphi(0)$ . This shift — from pointwise evaluation to duality pairing — is the single idea that makes the entire theory rigorous, and it automatically delivers differentiation by transferring derivatives onto the test function via integration by parts.

DefinitionTest Function Spaces

Let $\Omega\subseteq\mathbb{R}^n$ be open.

$\mathcal{D}(\Omega)=C_c^\infty(\Omega)$ : smooth functions with compact support in $\Omega$ , with its (LF-)topology of convergence.
$\mathcal{S}(\mathbb{R}^n)$ : the Schwartz space of smooth functions $\varphi$ such that every seminorm

1

\quad \|\varphi\|_{\alpha,\beta}=\sup_{x\in\mathbb{R}^n}\bigl|x^\alpha\,\partial^\beta\varphi(x)\bigr|<\infty

is finite, for all multi-indices $\alpha,\beta$ . Derivation hint: $x^\alpha=x_1^{\alpha_1}\cdots x_n^{\alpha_n}$ and $\partial^\beta=\partial_{x_1}^{\beta_1}\cdots\partial_{x_n}^{\beta_n}$ ; finiteness of all these seminorms is exactly "smooth + faster-than-polynomial decay of all derivatives." A sequence $\varphi_k\to\varphi$ in $\mathcal{D}$ iff all supports lie in one compact set and all derivatives converge uniformly.

DefinitionDistributions and the Delta

A distribution is a continuous linear functional $T:\mathcal{D}(\Omega)\to\mathbb{C}$ ; the space is $\mathcal{D}'(\Omega)$ . Continuity means: there exist $C>0$ and $N\in\mathbb{N}$ (depending on each compact $K$ ) with

2

\quad |\langle T,\varphi\rangle|\le C\sum_{|\alpha|\le N}\sup_{x\in K}|\partial^\alpha\varphi(x)|,\qquad \operatorname{supp}\varphi\subseteq K.

Derivation hint: this is the standard seminorm-bound characterization of continuity on the inductive-limit space $\mathcal{D}$ ; the least valid $N$ is the local order of $T$ . The Dirac delta $\delta_a\in\mathcal{D}'(\mathbb{R}^n)$ at the point $a$ is

3

\quad \langle\delta_a,\varphi\rangle=\varphi(a),\qquad \delta:=\delta_0.

Derivation hint: linearity and continuity (with $N=0$ , $C=1$ ) are immediate from $|\varphi(a)|\le\sup|\varphi|$ ; thus $\delta$ is a distribution of order $0$ — in fact a Radon measure. A distribution $T$ is a tempered distribution ( $T\in\mathcal{S}'$ ) if it is continuous on $\mathcal{S}$ ; these are exactly the distributions admitting a Fourier transform.

TheoremDistributional Derivative

For any $T\in\mathcal{D}'(\Omega)$ and multi-index $\alpha$ , the derivative $\partial^\alpha T$ defined by

4

\quad \langle\partial^\alpha T,\varphi\rangle=(-1)^{|\alpha|}\langle T,\partial^\alpha\varphi\rangle

is again a distribution. Every distribution is therefore $C^\infty$ in the distributional sense. Derivation hint: motivated by integration by parts $\int (\partial^\alpha f)\varphi=(-1)^{|\alpha|}\int f\,\partial^\alpha\varphi$ , with boundary terms vanishing because $\varphi$ has compact support; the right side is continuous in $\varphi$ since $\partial^\alpha:\mathcal{D}\to\mathcal{D}$ is continuous.

TheoremHeaviside, Delta, and Localization

Let $H$ be the Heaviside step ( $H(x)=1$ for $x>0$ , else $0$ ). Then

5

\quad H'=\delta,\qquad \delta'(\varphi)=-\varphi'(0),\qquad x\,\delta(x)=0,\qquad \delta(ax)=\tfrac{1}{|a|}\delta(x).

Derivation hint: each follows from the definitions: $\langle H',\varphi\rangle=-\int_0^\infty\varphi'=\varphi(0)$ ; $\langle x\delta,\varphi\rangle=\langle\delta,x\varphi\rangle=(x\varphi)|_{0}=0$ ; the scaling from change of variables $y=ax$ .

ProofDelta as a Limit of Approximate Identities

Claim. If $\eta\in L^1(\mathbb{R})$ with $\int\eta=1$ and $\eta_\varepsilon(x)=\varepsilon^{-1}\eta(x/\varepsilon)$ , then $\eta_\varepsilon\to\delta$ in $\mathcal{D}'$ as $\varepsilon\to0^+$ .

Step 1. For $\varphi\in\mathcal{D}$ , write the pairing as an integral of the locally integrable function $\eta_\varepsilon$ :

6

\quad \langle\eta_\varepsilon,\varphi\rangle=\int_{\mathbb{R}}\frac{1}{\varepsilon}\eta\!\left(\frac{x}{\varepsilon}\right)\varphi(x)\,dx =\int_{\mathbb{R}}\eta(y)\,\varphi(\varepsilon y)\,dy.

Derivation hint: substitute $y=x/\varepsilon$ , $dx=\varepsilon\,dy$ ; the prefactor $1/\varepsilon$ cancels.

7

\quad \lim_{\varepsilon\to0}\langle\eta_\varepsilon,\varphi\rangle=\varphi(0)\int_{\mathbb{R}}\eta(y)\,dy=\varphi(0)=\langle\delta,\varphi\rangle.

Derivation hint: pull the limit inside the integral using the integrable dominator $\|\varphi\|_\infty|\eta|$ , then use $\int\eta=1$ .

Step 3. Since this holds for all $\varphi\in\mathcal{D}$ , we have $\eta_\varepsilon\to\delta$ in the weak- $*$ topology of $\mathcal{D}'$ . $\qquad\blacksquare$ The same argument with $\eta(x)=\frac{1}{\pi}\frac{1}{1+x^2}$ (Cauchy/Lorentzian), $\eta(x)=\frac{1}{\sqrt{2\pi}}e^{-x^2/2}$ (Gaussian), or $\frac{1}{\pi}\frac{\sin x}{x}$ (Dirichlet, with care) all yield $\delta$ .

∎

ProofFourier Transform of the Delta

With convention $\widehat{f}(\xi)=\int f(x)e^{-i\xi x}\,dx$ , the Fourier transform on $\mathcal{S}'$ is defined by duality $\langle\widehat{T},\varphi\rangle=\langle T,\widehat{\varphi}\rangle$ . Then

8

\quad \widehat{\delta}=1,\qquad \widehat{1}=2\pi\,\delta,\qquad \widehat{\delta_a}(\xi)=e^{-ia\xi}.

Derivation hint: $\langle\widehat\delta,\varphi\rangle=\langle\delta,\widehat\varphi\rangle=\widehat\varphi(0)=\int\varphi$ , identifying $\widehat\delta$ with the constant $1$ ; the second follows by Fourier inversion / duality. This makes precise the physicists' identity

9

\quad \delta(x)=\frac{1}{2\pi}\int_{-\infty}^{\infty}e^{i\xi x}\,d\xi,

Derivation hint: interpret as the inverse transform of $\widehat\delta=1$ ; the divergent integral is meaningful only in $\mathcal{S}'$ , as a distributional limit of $\frac{1}{2\pi}\int_{-R}^{R}$ (a Dirichlet kernel).

∎

PA

Paul Adrien Maurice Dirac

1902–1984

Dirac introduced the delta in his 1927 paper on quantum mechanics and codified it in The Principles of Quantum Mechanics (1930), treating $\delta(x)$ as an "improper function" satisfying $\int\delta(x)\,dx=1$ and $\int f(x)\delta(x-a)\,dx=f(a)$ . He explicitly warned it was "not a quantity which can be generally used in mathematical analysis like an ordinary function, but its use must be confined to certain simple types of expression." Dirac needed it to express the eigenfunction expansion of operators with continuous spectra — the resolution of the identity $\sum_n|n\rangle\langle n|+\int|x\rangle\langle x|\,dx=\mathbb{1}$ — where the orthonormality $\langle x|x'\rangle=\delta(x-x')$ replaces the Kronecker delta of discrete bases. His pragmatic confidence that the manipulations were sound was vindicated two decades later by Schwartz.

🧠 Intuition

Think of a distribution as an idealized measuring device. A locally integrable function $f$ "measures" a test signal $\varphi$ by computing the weighted average $\int f\varphi$ . The delta is the perfect point sensor: it reports the value of the signal exactly at one point, ignoring everything else. You can never build such a sensor from an ordinary function (any real average smears over a region), but you can build a sequence of ever-sharper sensors — taller, narrower bumps of unit area — whose readings converge to the ideal one. The distribution is the limit of the readings, not of the graphs. For the derivative, integration by parts is the whole story: you can't differentiate a spike pointwise, but you can hand the derivative to the smooth probe instead. $\delta'$ is a "dipole sensor" that reports $-\varphi'(0)$ — the negative slope at the origin — which is why a derivative of a spike behaves like a balanced $+\!/\!-$ pair infinitely close together.

Figure

Approximate identities: Gaussian bumps

\eta_\varepsilon

sharpening to the delta spike as

\varepsilon\to0

, each of unit area

ExampleCharge density of a point charge and its potential

Given: A point charge $q=2\times10^{-9}\,\mathrm{C}$ at the origin in 3D. Model its density as $\rho(\mathbf{x})=q\,\delta^{(3)}(\mathbf{x})$ .

Step 1 — Total charge. Integrate the density:

10

\quad Q=\int_{\mathbb{R}^3}\rho\,d^3x=q\int\delta^{(3)}(\mathbf{x})\,d^3x=q.

Derivation hint: $\langle\delta^{(3)},1\rangle$ via the $n=3$ sampling property at $\mathbf{x}=0$ with $\varphi\equiv1$ on the support.

Step 2 — Poisson equation. The electrostatic potential satisfies $-\varepsilon_0\nabla^2\phi=\rho$ . The fundamental solution of the Laplacian in 3D is known:

11

\quad \nabla^2\!\left(\frac{-1}{4\pi|\mathbf{x}|}\right)=\delta^{(3)}(\mathbf{x}).

Derivation hint: $1/(4\pi r)$ is the Green's function of $-\nabla^2$ ; verify by integrating $\nabla^2(1/r)=0$ off the origin plus Gauss's flux $\oint\nabla(1/r)\cdot d\mathbf{S}=-4\pi$ .

Step 3 — Solve. Hence $\phi(\mathbf{x})=\dfrac{q}{4\pi\varepsilon_0|\mathbf{x}|}$ .

Numerical answer at $r=1\,\mathrm{m}$ ( $\varepsilon_0=8.854\times10^{-12}\,\mathrm{F/m}$ ):

\boxed{\phi=\frac{2\times10^{-9}}{4\pi(8.854\times10^{-12})(1)}\approx 17.98\ \mathrm{V}}

The delta cleanly produces the Coulomb potential as the Green's function evaluation.

ExampleDerivative of a jump: distributional bookkeeping

Given: $f(x)=\begin{cases}x^2 & x<1\\ 3x & x\ge1\end{cases}$ , which has a jump of size $3-1=2$ at $x=1$ (since $\lim_{x\to1^-}=1$ , $\lim_{x\to1^+}=3$ ).

Step 1 — Classical derivative away from the jump. $f'_{\text{cl}}(x)=2x$ for $x<1$ and $=3$ for $x>1$ .

Step 2 — Jump-derivative formula. For a piecewise $C^1$ function with jumps $[f]_{x_j}=f(x_j^+)-f(x_j^-)$ ,

12

\quad f'=\{f'\}_{\text{cl}}+\sum_j [f]_{x_j}\,\delta_{x_j}.

Derivation hint: integrate $\langle f',\varphi\rangle=-\int f\varphi'$ by parts on each smooth piece; the boundary terms collect into $\sum_j[f]_{x_j}\varphi(x_j)$ , which is the action of $[f]_{x_j}\delta_{x_j}$ .

Step 3 — Assemble. Here the single jump is $[f]_1=2$ :

\boxed{f'(x)=\bigl(2x\,\mathbb{1}_{x<1}+3\,\mathbb{1}_{x>1}\bigr)+2\,\delta(x-1)}

Check by re-integrating: pairing $f'$ against a $\varphi$ with $\varphi(1)=5$ , the delta contributes $2\cdot5=10$ to $\langle f',\varphi\rangle$ , exactly the missing boundary term a naive classical derivative would drop.

⚠️ Common mistake

$\delta$ is not a function. Writing " $\delta(0)=\infty$ " is meaningless; $\delta$ has no pointwise values. It only makes sense inside a pairing $\langle\delta,\varphi\rangle$ or under an integral sign understood as that pairing.

Products of distributions are generally undefined. There is no associative product making $\mathcal{D}'$ an algebra (Schwartz's impossibility result). In particular $\delta^2$ and $\delta(x)\,H(x)$ are ill-defined; attempting $\delta\cdot\delta$ leads to contradictions. Only special products like (smooth function) $\times$ (distribution) are legitimate.

$\delta(x)\,x = 0$ but you cannot "divide" to conclude $\delta(x)=0$ . Division is not generally available; the equation $xT=0$ has general solution $T=c\,\delta$ , not $T=0$ .

Convergence is weak- $*$ , not pointwise. $\eta_\varepsilon\to\delta$ in $\mathcal{D}'$ even though $\eta_\varepsilon(x)\to0$ for every fixed $x\neq0$ and $\eta_\varepsilon(0)\to\infty$ . The graphs do not converge to anything sensible; only the integrals against test functions converge.

CounterexampleNon-uniqueness in

xT=0

Suppose a student "solves" $x\,T(x)=0$ by dividing: $T=0/x=0$ . But take $T=5\delta$ . Then for any $\varphi\in\mathcal{D}$ ,

13

\quad \langle x\cdot5\delta,\varphi\rangle=\langle5\delta,x\varphi\rangle=5\,(x\varphi)\big|_{x=0}=0,

Derivation hint: multiplication by the smooth function $x$ is transferred onto $\varphi$ , and $x\varphi$ vanishes at $0$ regardless of $\varphi$ . so $x\cdot5\delta=0$ as a distribution, yet $5\delta\neq0$ . Thus the "solution" $T=0$ is not unique; the kernel of "multiply by $x$ " is the line $\mathbb{C}\delta$ . This is exactly why naive cancellation fails in distribution theory.

connection

The delta sits at the crossroads of several theories. As a measure, it is the unit point mass $\delta_a$ in Measure Theory — a Radon measure, hence a distribution of order zero; the Lebesgue Integral against a continuous function reproduces $\varphi(a)$ . Through the Fourier Transform, $\widehat\delta=1$ links localization in space to spreading in frequency, the cleanest instance of the uncertainty principle. The delta is the identity for convolution, $f*\delta=f$ , which is the engine behind Green's Functions and Fundamental Solutions: solving $Lu=f$ reduces to finding $E$ with $LE=\delta$ , then $u=E*f$ . In Sobolev Spaces $H^s$ , $\delta\in H^{s}(\mathbb{R}^n)$ exactly when $s<-n/2$ , quantifying how singular it is. Finally, in Functional Analysis, $\mathcal{D}'$ is the strong dual of a nuclear Fréchet/LF space, and the delta is the evaluation functional whose representability (or lack thereof) characterizes Reproducing Kernel Hilbert Spaces.

🔬 Where this appears

⏳ History

- 1822 — Fourier: In Théorie analytique de la chaleur, Fourier writes the inversion formula

f(x)=\frac{1}{2\pi}\iint f(\alpha)e^{i\xi(x-\alpha)}\,d\alpha\,d\xi

, implicitly containing the delta as the inner integral.

- 1828 — Green / 1830s — Poisson: Green's functions for potential theory introduce source terms that are, in retrospect, delta-supported.

- 1880s–1890s — Heaviside: Operational calculus uses the step function

H

and its "derivative" (the impulse) freely; ridiculed by contemporaries for lack of rigor but spectacularly effective.

- 1894 — Kirchhoff / 1900s — physicists use impulsive idealizations in optics and mechanics.

- 1927 — Dirac: Introduces

\delta

explicitly in The physical interpretation of the quantum dynamics (Proc. Roy. Soc.); systematized in his 1930 textbook.

- 1936 — Sobolev: Defines generalized derivatives and "fonctionnelles" to solve hyperbolic PDEs — the first rigorous distribution-like theory, with what we now call Sobolev spaces.

- 1944–1950 — Schwartz: Laurent Schwartz develops the full théorie des distributions, unifying duality, Fourier transform, convolution, and the kernel theorem; publishes Théorie des distributions (1950–1951).

- 1950 — Fields Medal awarded to Schwartz, largely for distributions.

- 1955–1961 — Gelfand–Shilov, Hörmander: Generalized Functions (multivolume) and Hörmander's wavefront-set / microlocal refinement extend the theory; Mikio Sato's hyperfunctions (1959) give an alternative sheaf-theoretic foundation.

open-problem

Multiplication of distributions. Schwartz's impossibility theorem shows no associative, commutative product on $\mathcal{D}'$ extends pointwise multiplication of continuous functions while keeping $1$ as unit and $\delta$ behaving consistently. Yet physics (renormalization in QFT, nonlinear PDEs with rough data) demands products like $\delta^2$ or $H\delta$ . Partial resolutions include Colombeau algebras (sequences modulo negligibles, giving an associative differential algebra containing $\mathcal{D}'$ , but with a delicate "association" relation), Hörmander's product valid when wavefront sets do not collide (microlocal condition $\mathrm{WF}(u)+\mathrm{WF}(v)\not\ni 0$ ), and paraproducts / rough path & regularity structures (Martin Hairer's 2014 Fields Medal work) for stochastic PDEs. Determining the "right" canonical product remains domain-dependent: there is no universal answer, only context-specific algebras and renormalization schemes that fix the ambiguous finite parts. The frontier question — a fully canonical, physically faithful nonlinear theory of singular distributions — stays open.

The Dirac Delta Function and Distributions

Formal Statement

Why It Holds / Derivation

Building Intuition

Worked Examples

Common Misconceptions

Connections

Real-World Applications

Historical Development

Subtleties & Limits

See Also