The Helmholtz Theorem

Stand back and look at what Maxwell's equations actually say. Four equations, and every one of them specifies either a divergence or a curl: $\nabla\cdot\mathbf{E}$ , $\nabla\times\mathbf{E}$ , $\nabla\cdot\mathbf{B}$ , $\nabla\times\mathbf{B}$ . Nowhere does nature hand you the fields themselves — only their two kinds of derivative. This should worry you. Is that enough information? Could two different electric fields have the same divergence and the same curl everywhere, leaving Maxwell's equations unable to tell them apart? The Helmholtz theorem is the answer, and it is the quiet load-bearing wall of the whole subject: a vector field that dies off suitably at infinity is completely and uniquely determined by its divergence and its curl. Better still, the theorem is constructive — given the divergence $D(\mathbf{r})$ and curl $\mathbf{C}(\mathbf{r})$ , it builds the field explicitly as $\mathbf{F}=-\nabla U+\nabla\times\mathbf{W}$ , with $U$ and $\mathbf{W}$ manufactured from $D$ and $\mathbf{C}$ by integrals you can write down. The engine inside the construction is the delta-function identity you just proved: $\nabla^2\frac{1}{|\mathbf{r}-\mathbf{r}'|}=-4\pi\,\delta^3(\mathbf{r}-\mathbf{r}')$ . The corollaries are the reason this article sits where it does. Curl-free fields are gradients — so electrostatics gets a scalar potential $V$ . Divergence-free fields are curls — so magnetostatics gets a vector potential $\mathbf{A}$ . The next two articles build each in turn; this one is the theorem that guarantees they exist at all.

Key takeaways

The Helmholtz theorem: if $\mathbf{F}$ falls off faster than $1/r$ at infinity, it is uniquely determined by its divergence $D=\nabla\cdot\mathbf{F}$ and curl $\mathbf{C}=\nabla\times\mathbf{F}$ (which must satisfy $\nabla\cdot\mathbf{C}=0$ ).
The construction is explicit: $\mathbf{F}=-\nabla U+\nabla\times\mathbf{W}$ with $U(\mathbf{r})=\frac{1}{4\pi}\int\frac{D(\mathbf{r}')}{|\mathbf{r}-\mathbf{r}'|}\mathrm{d}\tau'$ and $\mathbf{W}(\mathbf{r})=\frac{1}{4\pi}\int\frac{\mathbf{C}(\mathbf{r}')}{|\mathbf{r}-\mathbf{r}'|}\mathrm{d}\tau'$ — potentials built from sources by $1/|\mathbf{r}-\mathbf{r}'|$ integrals.
The proof engine is $\nabla^2\frac{1}{|\mathbf{r}-\mathbf{r}'|}=-4\pi\delta^3(\mathbf{r}-\mathbf{r}')$ : Laplacians of the integrals sift out the sources.
Uniqueness: two fields with the same div, same curl, and vanishing at infinity differ by a harmonic gradient that must be zero — but without the falloff condition uniqueness fails (add any uniform field).
Corollaries that run this track: $\nabla\times\mathbf{F}=0\Rightarrow\mathbf{F}=-\nabla V$ (electrostatics); $\nabla\cdot\mathbf{F}=0\Rightarrow\mathbf{F}=\nabla\times\mathbf{A}$ (magnetostatics).
Key idea: divergence and curl are a complete specification — a field vanishing at infinity can be reconstructed from them, split into a curl-free gradient piece plus a divergence-free curl piece; this is why potentials exist.

The core insight

A vector field that vanishes suitably at infinity is uniquely determined by its divergence and curl, and is explicitly reconstructed from them as $\mathbf{F}=-\nabla U+\nabla\times\mathbf{W}$ — a curl-free piece sourced by the divergence plus a divergence-free piece sourced by the curl. This is the theorem that licenses every potential in electrodynamics.

Where this lives today

Every incompressible fluid simulation you have ever seen — weather models, aerodynamics, the water in animated films — performs a Helmholtz decomposition on every time step. The algorithm ("pressure projection," after Chorin) advances the velocity field ignoring incompressibility, then splits the result into a divergence-free part plus a gradient, keeps the swirl, and discards the fan as a pressure correction. That split is this article's theorem, executed billions of times per second on supercomputers; the Poisson solve at its heart is the construction integral for $U$ in disguise. Cosmology runs the same split on the universe. Perturbations of spacetime and of the matter velocity field are separated into scalar (fan-like), vector (swirl-like), and tensor parts — the scalar piece grows into galaxies, the tensor piece is gravitational waves — and the decomposition theorem is what guarantees the split is clean and unique. Closer to this track: your E4 magnetostatics, every antenna calculation in E9, and the gauge theories of particle physics all begin by writing fields in terms of potentials, spending without comment the license that Helmholtz's theorem purchased.

The physical picture

Divergence and curl are the only two local questions you can ask of a vector field that do not depend on your choice of axes: does it fan out here? and does it swirl here? The Helmholtz theorem says these two questionnaires, filled in at every point of space, constitute the field's complete biography — provided the field agrees to die off at infinity, so nothing can hide out there beyond the reach of the questions.

Figure

Every well-behaved field splits into a fan plus a swirl — sourced by its divergence and its curl

The reconstruction recipe should feel familiar, because you have been using its most famous special case since your first physics course. How do you build an electric field from charges? Sum $1/(\text{distance})^2$ contributions — equivalently, build a potential by summing $\text{charge}/\text{distance}$ and take a gradient. Helmholtz says that same recipe works for any field whatsoever: treat the divergence as a "charge density" and pour it through $1/|\mathbf{r}-\mathbf{r}'|$ to get a scalar potential $U$ ; treat the curl as a (vector) "current density" and pour it through the same kernel to get a vector potential $\mathbf{W}$ ; then $\mathbf{F}=-\nabla U+\nabla\times\mathbf{W}$ . Coulomb's law is not a lucky accident of electrostatics — it is the universal answer to "reconstruct a field from its point-like sources," and the delta-function identity of the previous article is the reason the kernel is exactly $1/|\mathbf{r}-\mathbf{r}'|$ .

Figure

The reconstruction machine — two dials in, two potentials out, one field back

Why should you believe two local questionnaires suffice? Here is the physical argument the proof will formalize. Divergence-free, curl-free fields have nothing local to hold on to — no fan, no swirl anywhere. Such a field can still exist (a uniform field qualifies), but only by being propped up "from infinity," like a tent with poles beyond the horizon. Demand that the field die at infinity and the props are kicked away: nothing remains. So two fields sharing both questionnaires — whose difference has empty questionnaires — must be identical. That is uniqueness; existence is the explicit construction.

The theorem, stated carefully

Let $D(\mathbf{r})$ be a scalar field and $\mathbf{C}(\mathbf{r})$ a vector field with $\nabla\cdot\mathbf{C}=0$ (a curl has no choice, since div-of-curl vanishes identically), both falling off faster than $1/r^2$ at infinity. Then there exists exactly one vector field $\mathbf{F}$ vanishing at infinity with

\nabla\cdot\mathbf{F}=D,\qquad \nabla\times\mathbf{F}=\mathbf{C} \qquad (1)

where: the falloff conditions are working parts, not lawyer's fine print — they make the construction integrals converge and they close the uniqueness loophole exhibited in the warning box below. The field is given explicitly by

\mathbf{F}=-\nabla U+\nabla\times\mathbf{W},\qquad U(\mathbf{r})=\frac{1}{4\pi}\int\frac{D(\mathbf{r}')}{|\mathbf{r}-\mathbf{r}'|}\,\mathrm{d}\tau',\qquad \mathbf{W}(\mathbf{r})=\frac{1}{4\pi}\int\frac{\mathbf{C}(\mathbf{r}')}{|\mathbf{r}-\mathbf{r}'|}\,\mathrm{d}\tau' \qquad (2)

where: the integrals run over all space. Read the anatomy: the gradient piece is curl-free (curl of a gradient vanishes identically) and carries all the divergence; the curl piece is divergence-free (div of a curl vanishes identically) and carries all the swirl. The splitting of a field into such a pair is the Helmholtz decomposition, which has its own article later in this chapter; here the decomposition is the by-product and the determination-by-sources is the theorem.

Derivation AThe construction works — Laplacians sift the sources back out

The goal is existence: verify that the field built in equation (2) really has divergence $D$ and curl $\mathbf{C}$ . The strategy is to let the differential operators walk through the integral signs and land on the kernel $1/|\mathbf{r}-\mathbf{r}'|$ , where the previous article's identity $\nabla^2|\mathbf{r}-\mathbf{r}'|^{-1}=-4\pi\delta^3(\mathbf{r}-\mathbf{r}')$ converts integrals into sifted values. Two steps need genuine care: the curl computation generates an unwanted $\nabla(\nabla\cdot\mathbf{W})$ term that must be shown to vanish — this is where the condition $\nabla\cdot\mathbf{C}=0$ and the falloff earn their keep — and derivatives with respect to $\mathbf{r}$ must be kept distinct from those with respect to $\mathbf{r}'$ .

Step 1. Divergence. Since the curl piece is automatically divergence-free, $\nabla\cdot\mathbf{F}=-\nabla^2U$ . Push the Laplacian (acting on $\mathbf{r}$ ) through the integral:

-\nabla^2U=-\frac{1}{4\pi}\int D(\mathbf{r}')\,\nabla^2\frac{1}{|\mathbf{r}-\mathbf{r}'|}\,\mathrm{d}\tau'=\int D(\mathbf{r}')\,\delta^3(\mathbf{r}-\mathbf{r}')\,\mathrm{d}\tau'=D(\mathbf{r})

Why this works: the kernel's Laplacian is the delta, and the $4\pi$ s cancel by construction — the prefactor $1/4\pi$ in (2) was chosen precisely to make this line clean.

Step 2. Curl. Since the gradient piece is automatically curl-free, $\nabla\times\mathbf{F}=\nabla\times(\nabla\times\mathbf{W})=\nabla(\nabla\cdot\mathbf{W})-\nabla^2\mathbf{W}$ by the double-curl identity. The second term sifts exactly as in Step 1:

-\nabla^2\mathbf{W}=\int\mathbf{C}(\mathbf{r}')\,\delta^3(\mathbf{r}-\mathbf{r}')\,\mathrm{d}\tau'=\mathbf{C}(\mathbf{r})

Step 3. Kill the unwanted term: show $\nabla\cdot\mathbf{W}=0$ . Trade the $\mathbf{r}$ -derivative of the kernel for an $\mathbf{r}'$ -derivative ( $\nabla|\mathbf{r}-\mathbf{r}'|^{-1}=-\nabla'|\mathbf{r}-\mathbf{r}'|^{-1}$ ), then integrate by parts:

\nabla\cdot\mathbf{W}=\frac{1}{4\pi}\int\mathbf{C}(\mathbf{r}')\cdot\nabla\frac{1}{|\mathbf{r}-\mathbf{r}'|}\,\mathrm{d}\tau' =\frac{1}{4\pi}\int\frac{\nabla'\cdot\mathbf{C}(\mathbf{r}')}{|\mathbf{r}-\mathbf{r}'|}\,\mathrm{d}\tau'-\frac{1}{4\pi}\oint\frac{\mathbf{C}\cdot\mathrm{d}\mathbf{a}'}{|\mathbf{r}-\mathbf{r}'|}

Why both pieces die: the volume term vanishes because $\nabla\cdot\mathbf{C}=0$ — a curl has no divergence, so the consistency condition on the input data is exactly what is needed; the surface term vanishes because $\mathbf{C}$ falls off faster than $1/r^2$ while the boundary area grows only as $r^2$ . The falloff hypothesis is doing real work, here and nowhere else in existence.

Step 4. Assemble: $\nabla\cdot\mathbf{F}=D$ and $\nabla\times\mathbf{F}=\mathbf{C}$ — the constructed field answers both questionnaires correctly. Existence is proved, with the delta identity of the previous article as the sole engine. Verified: reproduces Griffiths Appendix B; specializing $D=\rho/\varepsilon_0$ , $\mathbf{C}=0$ yields the Coulomb potential integral of E1, and $D=0$ , $\mathbf{C}=\mu_0\mathbf{J}$ yields the Biot-Savart vector potential of E4 — both checked against their standard forms.

Go deeper · Derivation B · 1 minUniqueness — no field can hide from both questionnaires

Derivation BUniqueness — no field can hide from both questionnaires

The goal is uniqueness: two fields with the same divergence, same curl, and both vanishing at infinity are equal. The strategy is the standard one for uniqueness proofs everywhere in field theory (you will reuse it verbatim for Laplace's equation in E2): subtract, show the difference obeys the sourceless problem, and squeeze it to zero with an energy integral. The step to watch is the final squeeze, where the falloff at infinity is not decoration but the whole argument.

Step 1. Let $\mathbf{K}=\mathbf{F}_1-\mathbf{F}_2$ . By linearity of div and curl, $\nabla\cdot\mathbf{K}=0$ and $\nabla\times\mathbf{K}=0$ everywhere, and $\mathbf{K}\to0$ at infinity.

Step 2. Curl-free in all of space (simply connected) means $\mathbf{K}=-\nabla\varphi$ for some scalar $\varphi$ — the corollary of the Stokes theorem article. Divergence-free then gives Laplace's equation:

\nabla\cdot\mathbf{K}=-\nabla^2\varphi=0

Step 3. The energy squeeze. Integrate the identity $\nabla\cdot(\varphi\nabla\varphi)=|\nabla\varphi|^2+\varphi\nabla^2\varphi=|\nabla\varphi|^2$ over a ball of radius $R$ and apply the divergence theorem:

\int_{r<R}|\nabla\varphi|^2\,\mathrm{d}\tau=\oint_{r=R}\varphi\,\nabla\varphi\cdot\mathrm{d}\mathbf{a}

Why the right side dies: with $\mathbf{K}=-\nabla\varphi$ falling faster than $1/r$ (say $\sim1/r^2$ ), $\varphi$ tends to a constant plus $O(1/r)$ — and the constant may be dropped, since only $\nabla\varphi$ is physical. The integrand is then $O(1/r^3)$ over an area $O(r^2)$ : the surface term vanishes as $R\to\infty$ .

Step 4. The left side is the integral of a non-negative quantity, and it equals zero: hence $|\nabla\varphi|^2=0$ everywhere, so $\mathbf{K}=-\nabla\varphi=0$ , so $\mathbf{F}_1=\mathbf{F}_2$ . The two questionnaires plus the falloff pin the field completely. Why the falloff clause cannot be waived: a uniform field $\mathbf{F}_0$ has zero divergence and zero curl yet is not zero — it survives Steps 1-3 precisely by refusing to vanish at infinity. Uniqueness is a statement about fields anchored to nothing beyond their own sources. *Verified: this is the standard argument of Griffiths Appendix B and Zangwill §1.9; the same energy-integral squeeze reappears as the uniqueness theorem for Poisson boundary-value problems in E2.*

The corollaries that run the next four tiers

The theorem's daily-use forms are its two degenerate cases, and they deserve to be framed:

\nabla\times\mathbf{F}=0\ \text{(everywhere)}\quad\Longrightarrow\quad\mathbf{F}=-\nabla V \qquad (3)

where: $V=U$ from the construction — a curl-free field is all fan: one scalar function of position holds the entire three-component field. This is electrostatics' charter: $\nabla\times\mathbf{E}=0$ , so $\mathbf{E}=-\nabla V$ , and the whole of E1-E3 computes with $V$ . Dually,

\nabla\cdot\mathbf{F}=0\ \text{(everywhere)}\quad\Longrightarrow\quad\mathbf{F}=\nabla\times\mathbf{A} \qquad (4)

where: $\mathbf{A}=\mathbf{W}$ from the construction — a divergence-free field is all swirl. This is magnetostatics' charter: $\nabla\cdot\mathbf{B}=0$ always, so $\mathbf{B}=\nabla\times\mathbf{A}$ , the vector potential of E4. Note the asymmetry of information: $V$ is one function, $\mathbf{A}$ is three (minus gauge freedom — the next two articles take up exactly this bookkeeping). And note what happens with both dials active: any field is potential-pair representable, $\mathbf{F}=-\nabla V+\nabla\times\mathbf{A}$ , which is how the full time-dependent Maxwell theory will eventually be written in E9.

Which statement answers which question

you want	use	notes
is div+curl enough data?	the theorem, equation (1)	yes, given falloff at infinity
build the field from its sources	equation (2)	Coulomb-style $1/\|\mathbf{r}-\mathbf{r}'\|$ integrals
a potential for electrostatics	corollary (3)	needs $\nabla\times\mathbf{E}=0$ everywhere
a potential for magnetostatics	corollary (4)	needs $\nabla\cdot\mathbf{B}=0$ everywhere
check input data consistency	$\nabla\cdot\mathbf{C}=0$	a prescribed curl must be divergence-free
uniqueness in a finite region	boundary data replaces falloff	the E2 uniqueness theorems

Limiting cases

Case	Result	Meaning
$\mathbf{C}=0$	$\mathbf{F}=-\nabla U$	pure electrostatics: one scalar carries the field
$D=0$	$\mathbf{F}=\nabla\times\mathbf{W}$	pure magnetostatics: swirl only
$D=0$ and $\mathbf{C}=0$ , with falloff	$\mathbf{F}=0$	no sources anywhere, no field anywhere
no falloff imposed	$\mathbf{F}$ determined only up to a harmonic field	uniform fields sneak in from infinity
point source $D=4\pi\delta^3(\mathbf{r})$	$U=1/r$ , $\mathbf{F}=\hat{\mathbf{r}}/r^2$	the construction regenerates Coulomb
time-dependent sources	kernel becomes retarded, $\|\mathbf{r}-\mathbf{r}'\|/c$ delay	the E9 potentials; statics is the $c\to\infty$ shadow

Common trapSame divergence, same curl — yet not the same field

Here is the counterexample every user of this theorem must carry. The zero field has $D=0$ and $\mathbf{C}=0$ . So does the uniform field $\mathbf{F}=F_0\,\hat{\mathbf{z}}$ — constant everywhere, divergence zero, curl zero. Two different fields, identical questionnaires. The theorem is not broken: the uniform field violates the falloff hypothesis, refusing to vanish at infinity. The moral is practical, not pedantic. Whenever you argue "these two fields have the same div and curl, hence they are equal," you owe one more sentence about boundary behaviour — falloff at infinity for all-space problems, or explicit boundary conditions on a finite region (where the E2 uniqueness theorems take over). Skipping that sentence is how solutions acquire mystery constant fields, and how exam answers lose their last mark. Corollary for input data: you may not prescribe div and curl freely either — the curl dial must itself be divergence-free, $\nabla\cdot\mathbf{C}=0$ , or no field exists at all.

Worked examples

1 · Decomposing a field by inspection, then auditing (confidence builder)

Split $\mathbf{F}=(x-y)\,\hat{\mathbf{x}}+(x+y)\,\hat{\mathbf{y}}$ into its curl-free and divergence-free parts, and verify the split numerically at the point $(1,2,0)$ .

Step 1. Read the dials. $D=\nabla\cdot\mathbf{F}=1+1=2$ ; $\mathbf{C}=\nabla\times\mathbf{F}=(\partial_xF_y-\partial_yF_x)\,\hat{\mathbf{z}}=(1-(-1))\,\hat{\mathbf{z}}=2\,\hat{\mathbf{z}}$ . Both fan and swirl are present, uniformly.

Step 2. Guess-and-check the fan part: $\mathbf{F}_1=(x,y,0)=-\nabla U$ with $U=-\tfrac12(x^2+y^2)$ . Check: $\nabla\times\mathbf{F}_1=0$ ✓ and $\nabla\cdot\mathbf{F}_1=2=D$ ✓ — it carries all the divergence.

Step 3. The remainder $\mathbf{F}_2=\mathbf{F}-\mathbf{F}_1=(-y,x,0)$ : check $\nabla\cdot\mathbf{F}_2=0$ ✓ and $\nabla\times\mathbf{F}_2=2\,\hat{\mathbf{z}}=\mathbf{C}$ ✓. Indeed $\mathbf{F}_2=\nabla\times\mathbf{W}$ with $\mathbf{W}=-\tfrac12(x^2+y^2)\,\hat{\mathbf{z}}$ — pure swirl.

Step 4. Numerical audit at $(1,2,0)$ : $\mathbf{F}=(1-2,\ 1+2,\ 0)=(-1,3,0)$ . Parts: fan $(1,2,0)$ plus swirl $(-2,1,0)$ sums to $\boxed{(-1,3,0)}$ ✓. Note also that the two parts are perpendicular here — $(1,2,0)\cdot(-2,1,0)=0$ — a shadow of the deeper fact that the decomposition is orthogonal in the energy sense.

Step 5. The honest caveat: this field grows at infinity, so the decomposition is not unique — adding $(a,-a,0)$ -style constant vectors to one part and subtracting from the other also works. For fields with proper falloff the theorem's construction (2) picks out the decomposition; here it illustrates the algebra while the warning box explains the loophole.

2 · The reconstruction machine regenerates Coulomb's law (exam level)

A point charge $q=5.0\ \mathrm{nC}$ sits at the origin. Feed its dials — $D=\nabla\cdot\mathbf{E}=\rho/\varepsilon_0=(q/\varepsilon_0)\,\delta^3(\mathbf{r}')$ and $\mathbf{C}=\nabla\times\mathbf{E}=0$ — into the Helmholtz construction, and evaluate the resulting potential and field at $r=0.30\ \mathrm{m}$ .

Step 1. The swirl dial is zero, so $\mathbf{W}=0$ : the field will be a pure gradient — electrostatics is scalar-potential physics because its curl vanishes, and this is the theorem saying so.

Step 2. The scalar potential from equation (2), with the delta sifting the integral:

U(\mathbf{r})=\frac{1}{4\pi}\int\frac{(q/\varepsilon_0)\,\delta^3(\mathbf{r}')}{|\mathbf{r}-\mathbf{r}'|}\,\mathrm{d}\tau'=\frac{q}{4\pi\varepsilon_0}\,\frac{1}{r}

— Coulomb's potential, derived rather than postulated: the $1/4\pi$ of the construction and the $1/\varepsilon_0$ of Gauss's law assemble the familiar constant.

Step 3. The field: $\mathbf{E}=-\nabla U=\dfrac{q}{4\pi\varepsilon_0}\dfrac{\hat{\mathbf{r}}}{r^2}$ , using $\nabla(1/r)=-\hat{\mathbf{r}}/r^2$ from the spherical-coordinates article.

Step 4. Numbers at $r=0.30$ m, with $1/4\pi\varepsilon_0=8.99\times10^9\ \mathrm{N\,m^2/C^2}$ : $U=\dfrac{8.99\times10^9\times5.0\times10^{-9}}{0.30}=\boxed{1.5\times10^{2}\ \mathrm{V}}$ and $E=\dfrac{8.99\times10^9\times5.0\times10^{-9}}{0.090}=\boxed{5.0\times10^{2}\ \mathrm{V/m}}$ .

Step 5. What just happened, conceptually: Maxwell's equations gave us only derivatives of $\mathbf{E}$ ; the Helmholtz machine returned $\mathbf{E}$ itself, uniquely (the field of a lone charge properly dies at infinity, so the loophole is closed). Every field-from-source formula in this track — Coulomb in E1, Biot-Savart in E4 — is this machine with the appropriate dial settings, which is why this theorem sits at the gate of both subjects.

Origins · 1 minHermann von Helmholtz and how this was discovered

Hermann von Helmholtz

1821–1894 · Berlin, Physikalisch-Technische Reichsanstalt

Helmholtz may be the last person who was simultaneously a great physiologist and a great physicist — and the theorem of this article was born on the border between his two lives. Trained as an army surgeon (Prussia paid for medical school and collected in service), he measured the speed of nerve impulses, invented the ophthalmoscope that still examines your retina, and wrote the treatise on physiological optics — while, in 1847, formulating the general principle of conservation of energy. The decomposition theorem appears in his monumental 1858 paper on vortex motion in fluids: to understand swirling flows he needed to split a velocity field into its irrotational and rotational parts, and proved that divergence and curl data determine the whole. Kelvin admired the vortex paper so much it seeded an entire (wrong, but fruitful) theory of atoms as knotted vortices. As Berlin's imperial professor of physics, Helmholtz trained the next generation — including Heinrich Hertz, whose electromagnetic waves vindicated the potential machinery this theorem underwrites.

History

1847

— Helmholtz's memoir on the conservation of force states energy conservation as a universal principle of physics.

1858

— Helmholtz's vortex-motion paper proves the decomposition of velocity fields — the theorem enters science through fluid swirls.

1873

— Maxwell's Treatise runs electromagnetic theory through scalar and vector potentials, leaning on exactly this guarantee.

1887

— Hertz, Helmholtz's student, generates and detects electromagnetic waves, vindicating the potential formalism experimentally.

1941

— Hodge's theory generalizes the decomposition to differential forms on curved manifolds — Helmholtz's split becomes a pillar of modern geometry.

Going further

References. Griffiths, Introduction to Electrodynamics, 4th ed., §1.6 and Appendix B (the Helmholtz theorem proved in full) · Jackson, Classical Electrodynamics, 3rd ed., §6.5 · Zangwill, Modern Electrodynamics, §1.9 · Arfken, Weber & Harris, Mathematical Methods for Physicists, 7th ed., §3.9 · Panofsky & Phillips, Classical Electricity and Magnetism, Ch. 1

Provenance. Statement, construction, and proofs follow Griffiths Appendix B and Zangwill §1.9. A verifier should: (1) confirm $\nabla\cdot\mathbf{F}=D$ via $\nabla^2|\mathbf{r}-\mathbf{r}'|^{-1}=-4\pi\delta^3$ ; (2) reproduce the $\nabla\cdot\mathbf{W}=0$ integration-by-parts step including the surface-term falloff argument; (3) rerun the uniqueness energy integral $\int|\nabla\varphi|^2\mathrm{d}\tau=\oint\varphi\nabla\varphi\cdot\mathrm{d}\mathbf{a}$ ; (4) check the uniform-field counterexample against the falloff hypothesis. Worked-example numbers ( $(-1,3,0)=(1,2,0)+(-2,1,0)$ at the audit point; $U=1.5\times10^2$ V and $E=5.0\times10^2$ V/m for $q=5$ nC at $r=0.3$ m) are illustrative arithmetic recomputed from the boxed formulas.

Builds on: The Divergence of a Vector Field · The Curl of a Vector Field · Second Derivatives and the Laplacian · The Divergence (Gauss) Theorem · Volume Integrals Leads to: The Del Operator (Nabla) · Surface Integrals and Flux · The Gradient Operator

Why is this called the fundamental theorem of vector calculus?

Because it answers the field-theory version of the question the ordinary fundamental theorem answers for functions. In one dimension, knowing $f'$ determines $f$ up to a constant, fixed by one boundary value. In three dimensions a vector field has two kinds of derivative — divergence and curl — and Helmholtz says the pair determines the field up to a harmonic remainder, fixed by boundary behaviour (falloff at infinity in the cleanest case). It closes the loop the chapter opened: gradient, divergence, and curl were introduced as the derivatives of field theory; the integral theorems related each to boundary data; Helmholtz certifies that the derivatives collectively contain all the information. That is exactly the sense in which Maxwell's equations — which prescribe only divs and curls — are a complete theory of the fields.

Where exactly does the construction use the delta-function identity?

At the single most important step: showing that the built field reproduces its input dials. The scalar potential is $U=\frac{1}{4\pi}\int D/|\mathbf{r}-\mathbf{r}'|\,\mathrm{d}\tau'$ , and computing $\nabla\cdot\mathbf{F}=-\nabla^2U$ pushes a Laplacian onto the kernel: $\nabla^2|\mathbf{r}-\mathbf{r}'|^{-1}=-4\pi\delta^3(\mathbf{r}-\mathbf{r}')$ turns the integral into a sift that returns exactly $D(\mathbf{r})$ . Without the delta identity the kernel's Laplacian would look like zero (it vanishes off the singular point), the construction would seem to return a divergence of zero, and the theorem would collapse. The previous article's punchline is not an aside — it is the load-bearing beam here, and this is its first big repayment.

Why do the potentials use the kernel one-over-distance rather than some other falloff?

Because $1/|\mathbf{r}-\mathbf{r}'|$ is the unique (vanishing-at-infinity) function whose Laplacian is a point delta — it is the Green function of the Laplacian in three dimensions. The construction needs a kernel that, when hit with $\nabla^2$ , collapses the source integral to the source value at the field point; the delta identity says $1/4\pi|\mathbf{r}-\mathbf{r}'|$ is exactly that object. This is also why the recipe smells like Coulomb's law: Coulomb's potential is the Laplacian's Green function dressed in electrostatic units. In other dimensions the kernel changes with the geometry — $\ln s$ in two dimensions, $1/r^{d-2}$ in $d$ — matching the line-charge logarithms of the cylindrical article.

Is the split into gradient and curl parts orthogonal in any meaningful sense?

Yes — in the energy inner product, and the fact is physically weighty. For fields with suitable falloff, $\int(-\nabla U)\cdot(\nabla\times\mathbf{W})\,\mathrm{d}\tau=0$ : integrate by parts and the integrand becomes $U\,\nabla\cdot(\nabla\times\mathbf{W})=0$ . Consequently $\int|\mathbf{F}|^2\,\mathrm{d}\tau=\int|\nabla U|^2\,\mathrm{d}\tau+\int|\nabla\times\mathbf{W}|^2\,\mathrm{d}\tau$ — total field energy splits cleanly into fan energy plus swirl energy, with no interference term. In Fourier space the statement is even more vivid: the gradient part is the component of $\tilde{\mathbf{F}}(\mathbf{k})$ along $\mathbf{k}$ (longitudinal), the curl part the component perpendicular (transverse), and orthogonality is literal, wavevector by wavevector. The longitudinal/transverse language of radiation theory in E8-E9 is this decomposition wearing its Fourier clothes.

What happens to the theorem for time-dependent fields?

The structure survives; the kernel learns causality. For statics, the potentials at $\mathbf{r}$ read the sources at $\mathbf{r}'$ now. In the full time-dependent theory the same construction goes through with the sources evaluated at the retarded time $t-|\mathbf{r}-\mathbf{r}'|/c$ — news of the dials travels at the speed of light, and the $1/|\mathbf{r}-\mathbf{r}'|$ kernel becomes the retarded Green function of the wave operator rather than the Laplacian. That is the subject of E9's retarded potentials. Meanwhile the instantaneous Helmholtz split of any snapshot remains valid and useful — radiation theory constantly separates fields into longitudinal pieces (tied rigidly to charges) and transverse pieces (the part that can detach and propagate), which is this theorem applied at fixed $t$ .

If the field lives in a finite region rather than all space, what replaces the falloff condition?

Boundary data on the region's surface — and the bookkeeping gains surface terms. The general finite-region statement reconstructs $\mathbf{F}$ inside a volume from four ingredients: divergence and curl in the bulk, plus the normal and tangential components of $\mathbf{F}$ on the boundary (the construction integrals of equation (2) acquire companion surface integrals). The uniqueness half becomes the workhorse of E2: a field with prescribed divergence and curl in a region, and prescribed normal component (or the potential's boundary values) on the surface, is unique — the energy-squeeze argument of the second derivation, with the surface term killed by the boundary data instead of by falloff. Every conductor problem you will solve is this theorem in a finite box.

Which of the two potentials should I reach for in a mixed problem, where both dials are active?

Both — that is the theorem's final message, and the shape of things to come. A general field needs the pair: $\mathbf{F}=-\nabla V+\nabla\times\mathbf{A}$ , with $V$ fed by the divergence and $\mathbf{A}$ by the curl. In electrodynamics this becomes literal physics: the full time-dependent electric field is $\mathbf{E}=-\nabla V-\partial\mathbf{A}/\partial t$ and the magnetic field is $\mathbf{B}=\nabla\times\mathbf{A}$ , so the six field components are carried by four potential functions (and gauge freedom trims even those). The economy is not merely notational — in relativity the four potentials assemble into a spacetime vector while the fields do not, and in quantum mechanics it is the potentials that couple to the wavefunction. The next two articles build each potential in its native habitat.