Proofs Chapter 4: Curl Formulas

Identities for the Curl of Vector Fields

4. Curl Formulas

This chapter proves the basic identities (29)-(35) for the curl $\nabla \times \mathbf{A}$ of a vector field. The proofs use the Levi-Civita symbol $\varepsilon_{ijk}$ together with the epsilon-delta identity $\varepsilon_{ijk}\varepsilon_{klm} = \delta_{il}\delta_{jm} - \delta_{im}\delta_{jl}$ (both introduced in Chapter 1). Like the divergence formulas of Chapter 3, the curl is a differential operator, so linearity and the Leibniz rule combine to yield each identity. Intuitively, the curl measures the circulation per unit area (the local strength of rotation), whose geometric meaning is illustrated in Figure 1.

Notation and conventions
Throughout, $\mathbf{A} = (A_1, A_2, A_3)$ and $\mathbf{B} = (B_1, B_2, B_3)$ are vector fields and $f, g$ are scalar fields. The component form of the curl is $[\nabla \times \mathbf{A}]_i = \varepsilon_{ijk}\,\partial_j A_k$, with summation over the repeated indices $j, k$ (Einstein convention).
The circulation around a small loop defines the curl, which stands perpendicular to the loop by the right-hand rule ∇ × A (⊥ to the plane) A · dr circulation per unit area Curl fingers along ∮, thumb points to ∇×A (right-hand rule) x z y
Figure 1. Geometric meaning of the curl. The circulation of the tangential component of a vector field around a small loop $C$, namely $\oint_C \mathbf{A}\cdot d\mathbf{r}$, divided by the enclosed area and taken in the limit, gives the normal component of the curl $\nabla\times\mathbf{A}$. By the right-hand rule (fingers along the circulation, thumb along the normal), the curl vector stands perpendicular to the loop.

4.1 Linearity

(29) Linearity of the Curl over Sums

Identity: $\nabla \times (\mathbf{A} + \mathbf{B}) = \nabla \times \mathbf{A} + \nabla \times \mathbf{B}$
Conditions: $\mathbf{A}, \mathbf{B}$ are $C^1$ vector fields.
Proof

Use the linearity of partial derivatives componentwise.

The $i$-th component of the left-hand side reads

\begin{equation}[\nabla \times (\mathbf{A} + \mathbf{B})]_i = \varepsilon_{ijk}\,\dfrac{\partial (A_k + B_k)}{\partial x_j} \label{eq:4-29-1}\end{equation}

By linearity of partial differentiation, $\partial(A_k + B_k)/\partial x_j = \partial A_k/\partial x_j + \partial B_k/\partial x_j$, hence

\begin{equation}[\nabla \times (\mathbf{A} + \mathbf{B})]_i = \varepsilon_{ijk}\,\dfrac{\partial A_k}{\partial x_j} + \varepsilon_{ijk}\,\dfrac{\partial B_k}{\partial x_j} \label{eq:4-29-2}\end{equation}

Each term in $\eqref{eq:4-29-2}$ is precisely the $i$-th component of the curl of $\mathbf{A}$ or $\mathbf{B}$.

\begin{equation}[\nabla \times (\mathbf{A} + \mathbf{B})]_i = [\nabla \times \mathbf{A}]_i + [\nabla \times \mathbf{B}]_i \label{eq:4-29-3}\end{equation}

Since the equality holds for $i = 1, 2, 3$, the vector identity follows.

Remark: The curl is a linear differential operator that maps a sum of fields to the sum of their curls. In electromagnetism this provides the mathematical basis for the superposition of magnetic fields generated by multiple currents.

(30) Curl of a Constant Multiple

Identity: $\nabla \times (c\,\mathbf{A}) = c\,(\nabla \times \mathbf{A})$ where $c$ is constant.
Conditions: $c \in \mathbb{R}$ is a constant; $\mathbf{A}$ is a $C^1$ vector field.
Proof

Pull the constant out of each partial derivative: $\partial(cA_k)/\partial x_j = c\,\partial A_k/\partial x_j$.

The $i$-th component of the left-hand side is

\begin{equation}[\nabla \times (c\,\mathbf{A})]_i = \varepsilon_{ijk}\,\dfrac{\partial (c A_k)}{\partial x_j} = \varepsilon_{ijk}\,c\,\dfrac{\partial A_k}{\partial x_j} \label{eq:4-30-1}\end{equation}

The constant $c$ factors out of the contracted sum.

\begin{equation}[\nabla \times (c\,\mathbf{A})]_i = c\,\varepsilon_{ijk}\,\dfrac{\partial A_k}{\partial x_j} = c\,[\nabla \times \mathbf{A}]_i \label{eq:4-30-2}\end{equation}

The result holds componentwise, so the vector identity follows.

Remark: Combined with (29), the curl is a linear map $\nabla \times : C^1 \to C^0$. When $c$ depends on the spatial coordinates one must use (31) instead.

4.2 Product Rules

(31) Curl of a Scalar Times a Vector Field

Identity: $\nabla \times (f\mathbf{A}) = f(\nabla \times \mathbf{A}) + (\nabla f) \times \mathbf{A}$
Conditions: $f$ is a $C^1$ scalar field; $\mathbf{A}$ is a $C^1$ vector field.
Proof

Apply the Leibniz rule for partial derivatives within the Levi-Civita representation.

The $i$-th component of the left-hand side is

\begin{equation}[\nabla \times (f\mathbf{A})]_i = \varepsilon_{ijk}\,\dfrac{\partial (f A_k)}{\partial x_j} \label{eq:4-31-1}\end{equation}

By the Leibniz rule, $\partial(f A_k)/\partial x_j = f\,\partial A_k/\partial x_j + A_k\,\partial f/\partial x_j$. Substituting into $\eqref{eq:4-31-1}$ gives

\begin{equation}[\nabla \times (f\mathbf{A})]_i = \varepsilon_{ijk}\left( f\,\dfrac{\partial A_k}{\partial x_j} + A_k\,\dfrac{\partial f}{\partial x_j} \right) \label{eq:4-31-2}\end{equation}

Split the expression into two terms.

\begin{equation}[\nabla \times (f\mathbf{A})]_i = f\,\varepsilon_{ijk}\,\dfrac{\partial A_k}{\partial x_j} + \varepsilon_{ijk}\,\dfrac{\partial f}{\partial x_j}\,A_k \label{eq:4-31-3}\end{equation}

The first term equals $f\,[\nabla \times \mathbf{A}]_i$. For the second term, write $\partial f/\partial x_j = (\nabla f)_j$:

\begin{equation}\varepsilon_{ijk}\,(\nabla f)_j\,A_k = [(\nabla f) \times \mathbf{A}]_i \label{eq:4-31-4}\end{equation}

This is the standard Levi-Civita representation $[\mathbf{u} \times \mathbf{v}]_i = \varepsilon_{ijk} u_j v_k$ of the cross product. Combining $\eqref{eq:4-31-3}$ with $\eqref{eq:4-31-4}$:

\begin{equation}[\nabla \times (f\mathbf{A})]_i = f\,[\nabla \times \mathbf{A}]_i + [(\nabla f) \times \mathbf{A}]_i \label{eq:4-31-5}\end{equation}

The vector identity follows since this holds for every component.

Remark: This is the curl analogue of the divergence product rule $\nabla \cdot (f\mathbf{A}) = f(\nabla \cdot \mathbf{A}) + (\nabla f) \cdot \mathbf{A}$. The appearance of a cross product in the second term is characteristic of the curl. Identity (34) is a corollary of this rule combined with $\nabla \times \nabla g = \mathbf{0}$.

(32) Curl of a Cross Product

Identity: $\nabla \times (\mathbf{A} \times \mathbf{B}) = \mathbf{A}(\nabla \cdot \mathbf{B}) - \mathbf{B}(\nabla \cdot \mathbf{A}) + (\mathbf{B} \cdot \nabla)\mathbf{A} - (\mathbf{A} \cdot \nabla)\mathbf{B}$
Conditions: $\mathbf{A}, \mathbf{B}$ are $C^1$ vector fields.
Proof

Combine the epsilon-delta identity $\varepsilon_{ijk}\varepsilon_{klm} = \delta_{il}\delta_{jm} - \delta_{im}\delta_{jl}$ with the Leibniz rule.

Write the $i$-th component of the left-hand side using Levi-Civita. The $k$-th component of $\mathbf{A} \times \mathbf{B}$ is $\varepsilon_{klm} A_l B_m$:

\begin{equation}[\nabla \times (\mathbf{A} \times \mathbf{B})]_i = \varepsilon_{ijk}\,\dfrac{\partial}{\partial x_j} (\varepsilon_{klm} A_l B_m) \label{eq:4-32-1}\end{equation}

Step 1: apply the epsilon-delta identity

The constant $\varepsilon_{klm}$ pulls outside the derivative. Apply the epsilon-delta identity to $\varepsilon_{ijk}\varepsilon_{klm}$:

\begin{equation}\varepsilon_{ijk}\varepsilon_{klm} = \delta_{il}\delta_{jm} - \delta_{im}\delta_{jl} \label{eq:4-32-2}\end{equation}

Substituting $\eqref{eq:4-32-2}$ into $\eqref{eq:4-32-1}$ gives

\begin{equation}[\nabla \times (\mathbf{A} \times \mathbf{B})]_i = (\delta_{il}\delta_{jm} - \delta_{im}\delta_{jl})\,\dfrac{\partial (A_l B_m)}{\partial x_j} \label{eq:4-32-3}\end{equation}

Step 2: contract with the Kronecker deltas

The product $\delta_{il}\delta_{jm}$ contracts to $l = i$, $m = j$, while $\delta_{im}\delta_{jl}$ contracts to $m = i$, $l = j$:

\begin{equation}[\nabla \times (\mathbf{A} \times \mathbf{B})]_i = \dfrac{\partial (A_i B_j)}{\partial x_j} - \dfrac{\partial (A_j B_i)}{\partial x_j} \label{eq:4-32-4}\end{equation}

Step 3: Leibniz rule and interpretation

Apply the Leibniz rule to each term.

\begin{equation}\dfrac{\partial (A_i B_j)}{\partial x_j} = A_i\,\dfrac{\partial B_j}{\partial x_j} + B_j\,\dfrac{\partial A_i}{\partial x_j} \label{eq:4-32-5}\end{equation}

\begin{equation}\dfrac{\partial (A_j B_i)}{\partial x_j} = A_j\,\dfrac{\partial B_i}{\partial x_j} + B_i\,\dfrac{\partial A_j}{\partial x_j} \label{eq:4-32-6}\end{equation}

Substituting $\eqref{eq:4-32-5}$ and $\eqref{eq:4-32-6}$ into $\eqref{eq:4-32-4}$:

\begin{equation}[\nabla \times (\mathbf{A} \times \mathbf{B})]_i = A_i\,\dfrac{\partial B_j}{\partial x_j} + B_j\,\dfrac{\partial A_i}{\partial x_j} - A_j\,\dfrac{\partial B_i}{\partial x_j} - B_i\,\dfrac{\partial A_j}{\partial x_j} \label{eq:4-32-7}\end{equation}

Identify each term: $\partial B_j/\partial x_j = \nabla \cdot \mathbf{B}$ and $\partial A_j/\partial x_j = \nabla \cdot \mathbf{A}$. The directional derivative operators $\mathbf{B} \cdot \nabla = B_j\,\partial/\partial x_j$ and $\mathbf{A} \cdot \nabla = A_j\,\partial/\partial x_j$ give

\begin{equation}B_j\,\dfrac{\partial A_i}{\partial x_j} = [(\mathbf{B} \cdot \nabla) \mathbf{A}]_i, \quad A_j\,\dfrac{\partial B_i}{\partial x_j} = [(\mathbf{A} \cdot \nabla) \mathbf{B}]_i \label{eq:4-32-8}\end{equation}

Substituting these into $\eqref{eq:4-32-7}$:

\begin{equation}[\nabla \times (\mathbf{A} \times \mathbf{B})]_i = A_i\,(\nabla \cdot \mathbf{B}) - B_i\,(\nabla \cdot \mathbf{A}) + [(\mathbf{B} \cdot \nabla) \mathbf{A}]_i - [(\mathbf{A} \cdot \nabla) \mathbf{B}]_i \label{eq:4-32-9}\end{equation}

The vector identity follows since this holds for every component:

\begin{equation}\nabla \times (\mathbf{A} \times \mathbf{B}) = \mathbf{A}(\nabla \cdot \mathbf{B}) - \mathbf{B}(\nabla \cdot \mathbf{A}) + (\mathbf{B} \cdot \nabla)\mathbf{A} - (\mathbf{A} \cdot \nabla)\mathbf{B} \label{eq:4-32-10}\end{equation}

Remark: This identity appears directly in fluid dynamics through Helmholtz's vorticity equation. The first two terms encode compression and expansion (divergence), while the last two represent advection (directional derivative). Identity (33) is a special case.

4.3 Combinations with Gradient

(33) Curl of the Cross Product of Two Gradients

Identity: $\nabla \times (\nabla f \times \nabla g) = \nabla f\,\nabla^2 g - \nabla g\,\nabla^2 f + (\nabla g \cdot \nabla)\nabla f - (\nabla f \cdot \nabla)\nabla g$
Conditions: $f, g$ are $C^2$ scalar fields.
Proof

Apply (32) with $\mathbf{A} = \nabla f$ and $\mathbf{B} = \nabla g$.

The right-hand side of (32) reads

\begin{equation}\nabla \times (\mathbf{A} \times \mathbf{B}) = \mathbf{A}(\nabla \cdot \mathbf{B}) - \mathbf{B}(\nabla \cdot \mathbf{A}) + (\mathbf{B} \cdot \nabla)\mathbf{A} - (\mathbf{A} \cdot \nabla)\mathbf{B} \label{eq:4-33-1}\end{equation}

Substituting $\mathbf{A} = \nabla f$ and $\mathbf{B} = \nabla g$ yields

\begin{equation}\nabla \times (\nabla f \times \nabla g) = \nabla f\,(\nabla \cdot \nabla g) - \nabla g\,(\nabla \cdot \nabla f) + (\nabla g \cdot \nabla)\nabla f - (\nabla f \cdot \nabla)\nabla g \label{eq:4-33-2}\end{equation}

By the definition of the Laplacian, $\nabla^2 = \nabla \cdot \nabla$, so $\nabla \cdot \nabla f = \nabla^2 f$ and $\nabla \cdot \nabla g = \nabla^2 g$. Substituting into $\eqref{eq:4-33-2}$:

\begin{equation}\nabla \times (\nabla f \times \nabla g) = \nabla f\,\nabla^2 g - \nabla g\,\nabla^2 f + (\nabla g \cdot \nabla)\nabla f - (\nabla f \cdot \nabla)\nabla g \label{eq:4-33-3}\end{equation}

Remark: This form arises in magnetohydrodynamics (MHD), where it appears in evolution equations for the magnetic field. The vector $\nabla f \times \nabla g$ is tangent to the line of intersection of the level surfaces $f = \mathrm{const}$ and $g = \mathrm{const}$.

(34) Curl of a Scalar Times a Gradient

Identity: $\nabla \times (f\,\nabla g) = \nabla f \times \nabla g$
Conditions: $f, g$ are $C^2$ scalar fields.
Proof

Apply (31) with $\mathbf{A} = \nabla g$ and use the vanishing curl of a gradient (Identity (38): $\nabla \times \nabla g = \mathbf{0}$, proved in Chapter 5).

From (31):

\begin{equation}\nabla \times (f\,\nabla g) = f\,(\nabla \times \nabla g) + (\nabla f) \times \nabla g \label{eq:4-34-1}\end{equation}

The curl of a gradient is identically zero:

\begin{equation}\nabla \times \nabla g = \mathbf{0} \label{eq:4-34-2}\end{equation}

Substituting $\eqref{eq:4-34-2}$ into $\eqref{eq:4-34-1}$ eliminates the first term:

\begin{equation}\nabla \times (f\,\nabla g) = \nabla f \times \nabla g \label{eq:4-34-3}\end{equation}

Remark: This identity is useful for analysing level-surface structure. The field $f\nabla g$ has nonzero curl precisely when $\nabla f$ and $\nabla g$ are non-parallel, i.e. when the level surfaces of $f$ and $g$ intersect transversally.

(35) Antisymmetry of the Curl in $f$ and $g$

Identity: $\nabla \times (f\,\nabla g) = -\nabla \times (g\,\nabla f)$
Conditions: $f, g$ are $C^2$ scalar fields.
Proof

Combine (34) with the antisymmetry of the cross product.

By (34), the left-hand side is

\begin{equation}\nabla \times (f\,\nabla g) = \nabla f \times \nabla g \label{eq:4-35-1}\end{equation}

Apply the same identity (34) with $f$ and $g$ swapped:

\begin{equation}\nabla \times (g\,\nabla f) = \nabla g \times \nabla f \label{eq:4-35-2}\end{equation}

By the antisymmetry of the cross product (Identity (2): $\mathbf{u} \times \mathbf{v} = -\mathbf{v} \times \mathbf{u}$),

\begin{equation}\nabla g \times \nabla f = -(\nabla f \times \nabla g) \label{eq:4-35-3}\end{equation}

Substituting $\eqref{eq:4-35-3}$ into $\eqref{eq:4-35-2}$:

\begin{equation}\nabla \times (g\,\nabla f) = -(\nabla f \times \nabla g) \label{eq:4-35-4}\end{equation}

Comparing $\eqref{eq:4-35-1}$ with $\eqref{eq:4-35-4}$ gives

\begin{equation}\nabla \times (f\,\nabla g) = -\nabla \times (g\,\nabla f) \label{eq:4-35-5}\end{equation}

Remark: The identity exhibits the antisymmetric exchange of the roles of $f$ and $g$. This reflects the symplectic structure that swaps the roles of position and momentum in Lagrangian descriptions.

References

  • Arfken, G. B., Weber, H. J., & Harris, F. E. (2013). Mathematical Methods for Physicists (7th ed.). Academic Press.
  • Griffiths, D. J. (2017). Introduction to Electrodynamics (4th ed.). Cambridge University Press.
  • Vector calculus identities - Wikipedia

Frequently Asked Questions

What are the product rules for the curl?

Key formulas: $\nabla\times(f\mathbf{F})=f(\nabla\times\mathbf{F})+(\nabla f)\times\mathbf{F}$ for scalar-vector products, and $\nabla\times(\mathbf{F}\times\mathbf{G})=\mathbf{F}(\nabla\cdot\mathbf{G})-\mathbf{G}(\nabla\cdot\mathbf{F})+(\mathbf{G}\cdot\nabla)\mathbf{F}-(\mathbf{F}\cdot\nabla)\mathbf{G}$ for cross products.

What is a curl-free (irrotational) vector field?

A vector field with $\nabla\times\mathbf{F}=\mathbf{0}$ is called irrotational or curl-free. In a simply connected region, $\nabla\times\mathbf{F}=0$ is equivalent to $\mathbf{F}=\nabla\phi$ (conservative field with scalar potential $\phi$). Examples include gravitational and electrostatic fields, where the line integral is path-independent.

What does Stokes' theorem state?

Stokes' theorem states $\oint_C \mathbf{F}\cdot d\mathbf{r}=\iint_S (\nabla\times\mathbf{F})\cdot d\mathbf{S}$, where $C$ is the boundary of surface $S$. The circulation (line integral) of $\mathbf{F}$ around a closed curve equals the flux of the curl through any surface bounded by that curve. It underpins Ampere's law in electromagnetism.