The Exterior Derivative

The exterior derivative $d$ is a single operator that subsumes gradient, curl, and divergence under one definition. Where vector calculus offers three operators with three different output shapes, the language of forms offers one operator that increases form-degree by 1 and satisfies one beautiful identity: $d^2 = 0$. This is the algebraic engine behind the generalised Stokes' Theorem. Once you can compute $d$ on forms of low degree, you have what you need to read modern geometric physics.

Definition on 0-forms (the gradient)

A 0-form is a smooth function $f$. Its exterior derivative is the 1-form

Up to the dictionary $dx \leftrightarrow \mathbf{e}_1$ etc., this is the gradient.

Definition on 1-forms (the curl)

For a 1-form $\omega = P,dx + Q,dy + R,dz$, the exterior derivative is the 2-form

which after expanding and simplifying becomes

The three coefficients are the components of $\nabla\times(P, Q, R)$. So $d$ on 1-forms IS the curl.

Definition on 2-forms (the divergence)

For a 2-form $\eta = A,dy\wedge dz + B,dz\wedge dx + C,dx\wedge dy$,

The coefficient is $\nabla\cdot(A, B, C)$. So $d$ on 2-forms IS the divergence.

The headline identity: $d^2 = 0$

For any smooth form $\omega$, applying $d$ twice gives zero. The proof is two lines: $d^2 f = \sum_{i,j} f_{x_i x_j},dx_i\wedge dx_j$; mixed partials are symmetric in $i, j$, the wedge product is antisymmetric, so the whole sum cancels in pairs and vanishes. Geometrically $d^2=0$ corresponds to “the boundary of a boundary is empty”. It implies $\nabla\times(\nabla f) = \mathbf{0}$ and $\nabla\cdot(\nabla\times\mathbf{F}) = 0$ at once.

Closed and exact

A form is closed if $d\omega = 0$. A form is exact if $\omega = d\alpha$ for some $\alpha$. Since $d^2 = 0$, every exact form is closed. The converse holds on contractible (e.g., star-shaped) domains by the Poincaré lemma, but fails in general, the gap measures the topology of the domain (de Rham cohomology). On $\mathbb{R}^2\setminus{0}$, the form $\omega = (-y,dx + x,dy)/(x^2+y^2)$ is closed but not exact; its integral around the origin is $2\pi$, witnessing the hole.

(Computing $d$ symbolically is a pencil-and-paper exercise; widget visualisation of forms is rare and not yet in lang-core. The references include online tools.)

Pause and think: If $\omega = d\alpha$ is exact, what is $\int_M \omega$ for a closed manifold $M$ (one with $\partial M = \emptyset$)? Use Stokes' Theorem to argue this without computing.

Try it

• Compute $df$ for $f = xy + z^2$. Then compute $d(df)$ and confirm it equals zero.
• For $\omega = y,dx + x,dy$, compute $d\omega$. Is $\omega$ closed? Find an $f$ such that $\omega = df$.
• Predict first: is $\omega = x,dy$ closed? Compute $d\omega$ to verify. If not closed, can it possibly be exact?
• Compute $d\eta$ for the 2-form $\eta = x,dy\wedge dz + y,dz\wedge dx + z,dx\wedge dy$. The result should be a multiple of the volume form.

A trap to watch for

When computing $d(P,dx + Q,dy + R,dz)$, beginners often forget the wedge-product sign: $dy\wedge dx = -,dx\wedge dy$. Re-order systematically to the canonical basis ($dx\wedge dy,,dy\wedge dz,,dz\wedge dx$) and track the sign each time. Skipping this is the most common error in form calculations.

What you now know

You can compute $d$ on low-degree forms in $\mathbb{R}^3$, verify $d^2 = 0$, and distinguish closed from exact. Section 6.3 makes the form-vector-field dictionary explicit, so you can move freely between the two languages.

References

• Garrity, T. (2002). All the Mathematics You Missed. Cambridge University Press, ch. 6.
• Bachman, D. (2012). A Geometric Approach to Differential Forms (2nd ed.). Birkhäuser.
• Spivak, M. (1965). Calculus on Manifolds. W. A. Benjamin, ch. 4.
• Lee, J. M. (2012). Introduction to Smooth Manifolds (2nd ed.). Springer, ch. 14.
• Hubbard, J. H., Hubbard, B. B. (2015). Vector Calculus, Linear Algebra, and Differential Forms (5th ed.). Matrix Editions.