01The Big Idea — Boundary vs. Interior
Before any formulas, let's understand what Stokes' theorem is saying in plain English.
Imagine a river with a current that swirls — faster in some places, slower in others, maybe spinning in little eddies. You want to know: how much does the water circulate around the edge of some region? You have two ways to measure this:
Two Ways to Measure Circulation
Method 1 — Walk the boundary. Carry a paddle around the edge of the region. At each point, measure how much the current pushes the paddle in your direction of travel. Add it all up around the loop. This is a line integral — \(\oint_C \vec{F} \cdot d\vec{\ell}\).
Method 2 — Measure the spin inside. Look at every tiny patch of water inside the region. At each patch, measure how fast it's spinning (the curl). Add up all those spins over the whole surface. This is a surface integral — \(\iint_S (\nabla \times \vec{F}) \cdot d\vec{A}\).
Stokes' theorem says these two give the exact same answer. The total spin accumulated over a surface equals the net circulation around its edge. The interior determines the boundary — or equivalently, you can "trade" a surface integral for a line integral and vice versa.
The curl inside a surface (little rotation arrows) accumulates to produce the net circulation around the boundary loop \(C\). Interior eddies that don't touch the boundary cancel each other — only the boundary survives.
🌀
Why do interior spins cancel? Every interior edge is shared by two adjacent patches — one patch's clockwise contribution exactly cancels its neighbor's counterclockwise contribution. The only edges that don't cancel are on the boundary, where there's no neighbor. So the whole surface integral collapses to just the boundary. This is the deepest reason Stokes' theorem works.
02Curl — What It Actually Measures
Stokes' theorem involves the curl of a vector field, written \(\nabla \times \vec{F}\). Before we can state the theorem precisely, we need to understand what curl is — not just the formula, but the physical picture.
The Paddle-Wheel Test
Place a tiny paddle wheel in a flowing fluid. If the fluid rotates the wheel, the field has nonzero curl there. The curl vector points in the axis of rotation (using the right-hand rule), and its magnitude is twice the angular velocity of that infinitesimal paddle wheel.
Left: uniform flow — paddle wheel doesn't spin, curl = 0. Center: flow faster on top than bottom — shear causes spinning, curl ≠ 0. Right: circular flow — obvious spin, large curl pointing out of page.
The Formula from First Principles
Consider a tiny rectangle in the \(xy\)-plane with sides \(\Delta x\) and \(\Delta y\). We want the circulation of \(\vec{F} = (F_x, F_y, F_z)\) around this rectangle — going counterclockwise.
Bottom edge — going in +x direction, \(y = y_0\)
$$\text{contribution} = F_x(x_0, y_0)\,\Delta x$$
Right edge — going in +y direction, \(x = x_0 + \Delta x\)
$$\text{contribution} = F_y(x_0+\Delta x,\, y_0)\,\Delta y \approx \left(F_y + \frac{\partial F_y}{\partial x}\Delta x\right)\Delta y$$
Top edge — going in −x direction, \(y = y_0 + \Delta y\)
$$\text{contribution} = -F_x(x_0,\, y_0+\Delta y)\,\Delta x \approx -\left(F_x + \frac{\partial F_x}{\partial y}\Delta y\right)\Delta x$$
Left edge — going in −y direction, \(x = x_0\)
$$\text{contribution} = -F_y(x_0,\, y_0)\,\Delta y$$
Add all four edges — most terms cancel
$$\oint_{\text{rectangle}} \vec{F}\cdot d\vec{\ell} = \frac{\partial F_y}{\partial x}\Delta x\,\Delta y - \frac{\partial F_x}{\partial y}\Delta y\,\Delta x = \left(\frac{\partial F_y}{\partial x} - \frac{\partial F_x}{\partial y}\right)\Delta x\,\Delta y$$
Divide by area \(\Delta A = \Delta x\,\Delta y\) to get circulation per unit area
$$\frac{\oint \vec{F}\cdot d\vec{\ell}}{\Delta A} = \frac{\partial F_y}{\partial x} - \frac{\partial F_x}{\partial y} \equiv (\nabla \times \vec{F})_z$$
That quantity — \(\partial F_y/\partial x - \partial F_x/\partial y\) — is the \(z\)-component of the curl. It measures circulation per unit area in the \(xy\)-plane. The full 3D curl has three components:
The Curl — Full Formula
$$\nabla \times \vec{F} = \begin{vmatrix} \hat{x} & \hat{y} & \hat{z} \\ \partial_x & \partial_y & \partial_z \\ F_x & F_y & F_z \end{vmatrix}$$
$$= \left(\frac{\partial F_z}{\partial y} - \frac{\partial F_y}{\partial z}\right)\hat{x} + \left(\frac{\partial F_x}{\partial z} - \frac{\partial F_z}{\partial x}\right)\hat{y} + \left(\frac{\partial F_y}{\partial x} - \frac{\partial F_x}{\partial y}\right)\hat{z}$$
\((\nabla\times\vec{F})_x\)\(\partial_y F_z - \partial_z F_y\)\(yz\)-planeHow much the field curls around the \(x\)-axis. A tiny paddle wheel with axle along \(\hat{x}\) spins at rate proportional to this component.
\((\nabla\times\vec{F})_y\)\(\partial_z F_x - \partial_x F_z\)\(xz\)-planeHow much the field curls around the \(y\)-axis. Each component captures spin in one coordinate plane.
\((\nabla\times\vec{F})_z\)\(\partial_x F_y - \partial_y F_x\)\(xy\)-planeThe one we derived above. This is what you'd measure with a paddle wheel floating flat on a table — does the \(xy\)-flow spin it?
🧭
Geometric meaning of each term: Take \(\partial F_y/\partial x - \partial F_x/\partial y\). If \(\partial F_y/\partial x > 0\), the \(y\)-component of force grows as you go right — the right side of your paddle gets pushed up more than the left, spinning it counterclockwise (positive). If \(\partial F_x/\partial y > 0\), the \(x\)-component grows going up — the top gets pushed right more, spinning it clockwise (negative). The curl is the difference of these two competing tendencies.
⚡ Physics — Curl is Faraday's and Ampère's Law
In electromagnetism, curl is everywhere. Faraday's law says \(\nabla \times \vec{E} = -\partial\vec{B}/\partial t\) — a changing magnetic field creates a curling electric field. Ampère's law says \(\nabla \times \vec{B} = \mu_0\vec{J}\) — electric current creates a curling magnetic field (the field circles the wire). Both are statements about local curl, and Stokes' theorem is what connects them to the measurable loop integrals used in circuit theory.
03Line Integrals — Integrating Along a Curve
The left side of Stokes' theorem is a line integral. Let's make sure this is completely clear before we put everything together.
What a Line Integral Measures
Given a vector field \(\vec{F}\) and a curve \(C\), the line integral \(\int_C \vec{F}\cdot d\vec{\ell}\) measures the total "push" of the field along the curve. At each point, we take the component of \(\vec{F}\) in the direction of travel, and accumulate it over the whole path.
Parameterize the curve — let \(\vec{r}(t)\) trace the path for \(t \in [a,b]\)
$$d\vec{\ell} = \vec{r}'(t)\,dt = \left(\frac{dx}{dt},\frac{dy}{dt},\frac{dz}{dt}\right)dt$$
Form the dot product at each point
$$\vec{F}\cdot d\vec{\ell} = F_x\frac{dx}{dt}\,dt + F_y\frac{dy}{dt}\,dt + F_z\frac{dz}{dt}\,dt$$
Integrate over the parameter
$$\int_C \vec{F}\cdot d\vec{\ell} = \int_a^b \vec{F}(\vec{r}(t))\cdot\vec{r}'(t)\,dt$$
At each point on the curve, we project \(\vec{F}\) onto the tangent direction \(\hat{T}\). The line integral sums up \(|\vec{F}|\cos\theta\) (the projection) times \(|d\vec{\ell}|\) over the whole path.
When the Loop is Closed
Stokes' theorem uses a closed loop — the curve comes back to where it started. We write this as \(\oint_C \vec{F}\cdot d\vec{\ell}\), and call it the circulation of \(\vec{F}\) around \(C\).
🔄
Positive or negative circulation? If the field generally flows in the direction you're walking, circulation is positive. If it flows against you, it's negative. For a conservative field (like gravity, or an electrostatic field with no time variation), the circulation around any closed loop is exactly zero — the field can't make you net-gain energy by going in circles. A nonzero circulation signals a non-conservative field with genuine rotation.
\(\oint_C\)Closed line integral—Integration around a closed curve \(C\) (the boundary of the surface). The circle on the integral sign signals that the path is closed — start = finish.
\(\vec{F}\)Vector fieldvariesCould be velocity (fluid flow), electric field, force field, magnetic field — any vector quantity that varies through space.
\(d\vec{\ell}\)Oriented path elementmAn infinitesimal vector tangent to the curve, pointing in the direction of traversal. Its magnitude is \(ds\) (arc length); its direction is \(\hat{T}\), the unit tangent.
\(\vec{F}\cdot d\vec{\ell}\)Tangential component × lengthvariesThe projection of \(\vec{F}\) onto the travel direction, times the tiny step length. Positive when field aids motion, negative when it opposes. Summing these gives total energy-like quantity along the path.
04Surface Integrals of Vector Fields — Flux Through a Surface
The right side of Stokes' theorem is a surface integral of the curl. To understand it, we first need to understand surface integrals of vector fields.
The Flux Idea
Imagine a surface \(S\) sitting in a vector field \(\vec{F}\) (think: a net held in a river). The flux through the surface measures how much of the field passes through it — not along it, but through it. Only the component perpendicular to the surface contributes.
The surface element \(d\vec{A}\) — an area patch with a normal direction
$$d\vec{A} = \hat{n}\,dA$$
\(\hat{n}\) is the unit normal to the surface at each point. \(dA\) is the area of the tiny patch. Together they encode both "how big" and "which way the surface faces."
Flux through the surface
$$\iint_S \vec{F}\cdot d\vec{A} = \iint_S \vec{F}\cdot\hat{n}\,dA$$
Flux = field dotted with the surface normal \(\hat{n}\). Field parallel to the surface contributes zero. Field perpendicular passes through fully. The integral sums all these perpendicular contributions.
For Stokes' Theorem: Flux of the Curl
In Stokes' theorem, we don't integrate \(\vec{F}\) itself through the surface — we integrate the curl of \(\vec{F}\). The curl \(\nabla \times \vec{F}\) is itself a vector field (a "rotation density" at each point), and we're computing how much of that rotation-vector pierces the surface:
Right side of Stokes' theorem
$$\iint_S (\nabla \times \vec{F})\cdot d\vec{A} = \iint_S (\nabla \times \vec{F})\cdot\hat{n}\,dA$$
This picks out the component of curl that is perpendicular to the surface — the spinning that's "aligned with" the surface's orientation.
📐
The orientation rule (right-hand rule): The normal \(\hat{n}\) and the direction of traversal around \(C\) must be consistent. Curl the fingers of your right hand in the direction you walk around \(C\) — your thumb points in the direction of \(\hat{n}\). This is how you make sure the two sides of Stokes' theorem have the same sign. Flip either the surface orientation or the loop direction and you negate both sides simultaneously.
05Building the Theorem — The Cancellation Argument
Now let's prove why Stokes' theorem is true. The argument is geometric and beautiful — no heavy computation required.
Step 1: Tile the Surface with Tiny Rectangles
Divide the surface \(S\) into a grid of tiny rectangular patches. Each patch is so small it's essentially flat. We showed in Section 2 that for a tiny flat rectangle with normal \(\hat{n}\) and area \(\Delta A\):
Circulation around one tiny patch = curl flux through it
$$\oint_{\partial(\text{patch})} \vec{F}\cdot d\vec{\ell} = (\nabla\times\vec{F})\cdot\hat{n}\;\Delta A$$
This is just our derivation from Section 2, restated: circulation per unit area = curl component normal to the patch.
Step 2: Sum Over All Patches
Sum this equation over all patches in the grid:
Sum of all patch circulations
$$\sum_{\text{patches}} \oint_{\partial(\text{patch})} \vec{F}\cdot d\vec{\ell} = \sum_{\text{patches}} (\nabla\times\vec{F})\cdot\hat{n}\;\Delta A$$
Every interior edge is shared by two adjacent patches, traversed in opposite directions — contributions cancel exactly. Only the boundary edges (shown in teal) have no neighbor to cancel them. The entire sum collapses to just the boundary loop.
Step 3: Interior Edges Cancel
Look at any interior edge shared by patches \(P_1\) and \(P_2\). When we traverse patch \(P_1\)'s boundary, we cross this edge in one direction. When we traverse patch \(P_2\)'s boundary, we cross it in the opposite direction. Same edge, same field values, opposite signs — perfect cancellation.
Interior edges cancel in pairs
$$\sum_{\text{patches}} \oint_{\partial(\text{patch})} \vec{F}\cdot d\vec{\ell} = \oint_{\partial S} \vec{F}\cdot d\vec{\ell} = \oint_C \vec{F}\cdot d\vec{\ell}$$
Only the outermost boundary edges survive. The left side collapses to the single loop integral around \(C = \partial S\).
Step 4: Right Side Becomes a Surface Integral
As patches shrink to zero, the sum becomes an integral
$$\sum_{\text{patches}} (\nabla\times\vec{F})\cdot\hat{n}\;\Delta A \;\longrightarrow\; \iint_S (\nabla\times\vec{F})\cdot\hat{n}\,dA$$
Combining Steps 3 and 4:
Stokes' Theorem — Derived
$$\oint_C \vec{F}\cdot d\vec{\ell} = \iint_S (\nabla\times\vec{F})\cdot d\vec{A}$$
The entire proof reduces to: tile → each tile satisfies a local Stokes → sum → interior cancels → only boundary survives. No magic — just a rigorous version of the cancellation picture.
06The Theorem — Precisely Stated
Stokes' Theorem
$$\boxed{\oint_C \vec{F}\cdot d\vec{\ell} = \iint_S (\nabla\times\vec{F})\cdot d\vec{A}}$$
where \(C = \partial S\) is the oriented boundary of surface \(S\), and \(\hat{n}\) (encoded in \(d\vec{A} = \hat{n}\,dA\)) is determined by the right-hand rule with respect to the traversal direction of \(C\).
\(\oint_C\)Closed line integral around \(C\)—The boundary loop. Must be closed, oriented (direction matters). The circle on \(\oint\) is the reminder that it's a loop, not an open path.
\(\vec{F}\)Vector fieldanyMust be continuously differentiable on an open set containing \(S\). Stokes' theorem fails if \(\vec{F}\) has singularities on or inside the surface.
\(d\vec{\ell}\)Oriented path elementmTangent vector to \(C\), in the traversal direction, times arc-length element \(ds\). The orientation of \(C\) and the choice of \(\hat{n}\) for \(S\) must be consistent via the right-hand rule.
\(\nabla\times\vec{F}\)Curl of \(\vec{F}\)same as F/mThe local rotation density of the field. A vector at each point whose direction is the axis of rotation and whose magnitude is twice the angular speed of an infinitesimal paddle wheel.
\(d\vec{A} = \hat{n}\,dA\)Oriented surface elementm²A tiny area patch on \(S\), with the outward normal \(\hat{n}\) attached. The dot product \((\nabla\times\vec{F})\cdot\hat{n}\,dA\) picks out the curl component perpendicular to the surface at that patch.
\(C = \partial S\)Boundary of \(S\)—Stokes' theorem is a statement about boundaries. The loop \(C\) must be the complete boundary of \(S\) — the curve you get by tracing the edge of the surface. \(\partial\) is the mathematical "boundary of" operator.
Conditions for Stokes' Theorem to Hold
Requirements
1. \(\vec{F}\) must have continuous first-order partial derivatives on an open region containing \(S\).
2. \(S\) must be an orientable surface (it has two distinct sides — no Möbius strips!).
3. \(C = \partial S\) must be a piecewise smooth closed curve.
4. The orientation of \(C\) and the normal \(\hat{n}\) to \(S\) must follow the right-hand rule consistently.
07Worked Examples
Example 1 — Verify Stokes' on a Flat Square
Let \(\vec{F} = (-y,\,x,\,0)\) and \(S\) be the square \(0 \le x \le 1\), \(0 \le y \le 1\), \(z = 0\). The boundary \(C\) is the unit square traversed counterclockwise.
Compute the curl
$$\nabla\times\vec{F} = \begin{vmatrix}\hat{x}&\hat{y}&\hat{z}\\\partial_x&\partial_y&\partial_z\\-y&x&0\end{vmatrix} = (0-0)\hat{x} - (0-0)\hat{y} + (1-(-1))\hat{z} = 2\hat{z}$$
Right side: surface integral of curl over the unit square (\(\hat{n} = \hat{z}\), \(dA = dx\,dy\))
$$\iint_S (\nabla\times\vec{F})\cdot\hat{z}\,dA = \iint_S 2\,dx\,dy = 2 \times \text{Area} = 2\times 1 = 2$$
Left side: compute the line integral around the four edges
$$\text{Bottom } (y=0,\, x: 0\to1): \int_0^1 (-0)\,dx + x\cdot 0 = 0$$
$$\text{Right } (x=1,\, y: 0\to1): \int_0^1 (-y)\cdot 0 + 1\,dy = 1$$
$$\text{Top } (y=1,\, x: 1\to0): \int_1^0 (-1)\,dx = -\int_1^0 dx = 1$$
$$\text{Left } (x=0,\, y: 1\to0): \int_1^0 (-y)\cdot 0 + 0\,dy = 0$$
$$\oint_C \vec{F}\cdot d\vec{\ell} = 0 + 1 + 1 + 0 = 2 \checkmark$$
✅
Both sides give 2. The field \(\vec{F} = (-y, x, 0)\) is a counterclockwise rotation — like water spinning in a drain viewed from above. It has constant curl \(2\hat{z}\) everywhere, so the surface integral is just \(2 \times \text{area}\). The line integral picks up the same thing by measuring how much the field pushes you along as you walk the boundary.
Example 2 — Stokes on a Curved Surface (Hemisphere)
Let \(\vec{F} = (xz,\,-y,\,x^2)\) and \(C\) be the unit circle in the \(xy\)-plane (\(x^2+y^2=1\), \(z=0\)). Use Stokes to compute \(\oint_C \vec{F}\cdot d\vec{\ell}\) by integrating over the upper hemisphere \(S\).
Step 1 — Compute the curl of \(\vec{F} = (xz, -y, x^2)\)
$$(\nabla\times\vec{F})_x = \partial_y(x^2) - \partial_z(-y) = 0 - 0 = 0$$
$$(\nabla\times\vec{F})_y = \partial_z(xz) - \partial_x(x^2) = x - 2x = -x$$
$$(\nabla\times\vec{F})_z = \partial_x(-y) - \partial_y(xz) = 0 - 0 = 0$$
$$\nabla\times\vec{F} = (0,\,-x,\,0)$$
Step 2 — Parametrize the upper hemisphere. Use \(\hat{n} = (x,y,z)\) (outward normal on unit sphere) and \(dA\) = surface area element
$$(\nabla\times\vec{F})\cdot\hat{n} = (0,\,-x,\,0)\cdot(x,y,z) = -xy$$
Step 3 — Switch to the flat disk (Section 8 explains why we can) — use \(S' =\) unit disk in \(z=0\), \(\hat{n}=\hat{z}\)
$$(\nabla\times\vec{F})\cdot\hat{z} = 0$$
$$\iint_{S'} 0\,dA = 0$$
$$\therefore\;\oint_C \vec{F}\cdot d\vec{\ell} = 0$$
🎯
The power of Stokes: choose the easiest surface. The unit circle is the boundary of both the flat disk and the hemisphere. The flat disk gave \((\nabla\times\vec{F})\cdot\hat{z} = 0\) immediately. Computing the same line integral directly around the circle would require a messy trigonometric parameterization. Stokes lets us pick whichever surface makes the integral simplest.
08Why the Choice of Surface Doesn't Matter
This is one of the most striking consequences of Stokes' theorem. The loop \(C\) is fixed, but you can span it with any surface — a flat disk, a bowl, a wild crumpled shape — and you always get the same answer for the surface integral of the curl.
The Proof (in One Sentence)
Take two surfaces \(S_1\) and \(S_2\) that share the same boundary \(C\). Together they form a closed surface \(S_1 \cup (-S_2)\) (we flip \(S_2\)'s orientation so the normals are consistently outward). Apply the Divergence Theorem:
Divergence theorem on the closed surface
$$\iint_{S_1} (\nabla\times\vec{F})\cdot d\vec{A} - \iint_{S_2} (\nabla\times\vec{F})\cdot d\vec{A} = \iiint_V \nabla\cdot(\nabla\times\vec{F})\,dV$$
Key identity: the divergence of any curl is always zero
$$\nabla\cdot(\nabla\times\vec{F}) = 0 \quad\text{(always, for any smooth }\vec{F}\text{)}$$
Proof: expand and every mixed partial appears twice with opposite signs — all cancel. This is a fundamental identity of vector calculus.
Therefore
$$\iint_{S_1} (\nabla\times\vec{F})\cdot d\vec{A} = \iint_{S_2} (\nabla\times\vec{F})\cdot d\vec{A} \quad \checkmark$$
Three different surfaces all bounded by the same loop \(C\) — flat disk, shallow bowl, deep bowl. Stokes' theorem gives the same value on all of them, because \(\nabla\cdot(\nabla\times\vec{F}) = 0\) means curl flux through a closed surface is always zero.
💡
Physical interpretation: Curl flux has no "sources" or "sinks" — unlike ordinary vector fields, curl fields can't be created or destroyed at a point. Any curl that flows into a closed surface must flow out the other side. So two surfaces sharing the same boundary capture exactly the same total curl. The field's curl is "conserved" in this topological sense.
09Physics Connections
Faraday's Law of Induction
Faraday's law in differential form says \(\nabla \times \vec{E} = -\partial\vec{B}/\partial t\). Apply Stokes' theorem to any surface \(S\) bounded by a conducting loop \(C\):
Apply Stokes' to Faraday's law
$$\oint_C \vec{E}\cdot d\vec{\ell} = \iint_S (\nabla\times\vec{E})\cdot d\vec{A} = \iint_S \left(-\frac{\partial\vec{B}}{\partial t}\right)\cdot d\vec{A} = -\frac{d}{dt}\iint_S \vec{B}\cdot d\vec{A} = -\frac{d\Phi_B}{dt}$$
The line integral of \(\vec{E}\) around the loop is the EMF (voltage). The surface integral of \(\vec{B}\) is the magnetic flux \(\Phi_B\). So Stokes turns the local differential law into the global integral law: \(\mathcal{E} = -d\Phi_B/dt\). This is exactly the equation governing generators, motors, and transformers.
Ampère's Law
Similarly, \(\nabla\times\vec{B} = \mu_0\vec{J}\) (Ampère's law without displacement current). Stokes' theorem gives:
Apply Stokes' to Ampère's law
$$\oint_C \vec{B}\cdot d\vec{\ell} = \iint_S (\nabla\times\vec{B})\cdot d\vec{A} = \mu_0\iint_S \vec{J}\cdot d\vec{A} = \mu_0 I_{\text{enc}}$$
The circulation of \(\vec{B}\) around a loop equals \(\mu_0\) times the current threading through any surface bounded by that loop. You can compute the field around a long wire using any surface you want — a flat disk, a bulging balloon — same answer every time.
⚡ Real World — Generators, MRI, and Circuit Theory
Generators: Rotating a coil in a magnetic field changes \(\Phi_B\), which by Faraday's law (which is Stokes') induces an EMF that drives current. Every power plant on Earth relies on this. Stokes' theorem is the mathematical backbone of electrical power generation.
MRI: Gradient coils create spatially varying magnetic fields. The RF pulse creates a time-varying \(\vec{B}\), inducing a curling \(\vec{E}\) (Faraday/Stokes) that resonantly excites hydrogen nuclei. The signal they emit is detected as the induced EMF in a receiver loop — again Stokes' theorem.
Kirchhoff's Voltage Law: KVL (\(\sum V = 0\) around a loop) is Stokes' theorem applied to a static electric field. For a conservative field, \(\nabla\times\vec{E} = 0\), so \(\oint \vec{E}\cdot d\vec{\ell} = 0\). KVL is Stokes' special case.
Conservative Fields and Path Independence
A field \(\vec{F}\) is conservative if and only if \(\nabla\times\vec{F} = 0\) everywhere. By Stokes' theorem, this is equivalent to \(\oint_C \vec{F}\cdot d\vec{\ell} = 0\) for every closed loop — which means the line integral between any two points is path-independent. Stokes' theorem is the reason these three statements (conservative, curl-free, path-independent) are all equivalent.
10Summary
| Concept | Formula / Statement | Geometric meaning |
|---|
| Stokes' Theorem | \(\oint_C \vec{F}\cdot d\vec{\ell} = \iint_S (\nabla\times\vec{F})\cdot d\vec{A}\) | Boundary circulation = interior curl flux |
| Curl | \(\nabla\times\vec{F} = (\partial_y F_z - \partial_z F_y,\;\partial_z F_x - \partial_x F_z,\;\partial_x F_y - \partial_y F_x)\) | Local rotation density; paddle-wheel spin per unit area |
| Curl z-component | \((\nabla\times\vec{F})_z = \partial_x F_y - \partial_y F_x\) | Circulation per unit area in \(xy\)-plane |
| Surface element | \(d\vec{A} = \hat{n}\,dA\) | Tiny area patch with outward normal |
| Orientation rule | Right-hand rule: curl fingers along \(C\), thumb = \(\hat{n}\) | Links loop direction to surface normal sign |
| Key identity | \(\nabla\cdot(\nabla\times\vec{F}) = 0\) | Curl has no sources; surface choice doesn't matter |
| Conservative field | \(\nabla\times\vec{F} = 0 \Leftrightarrow \oint_C \vec{F}\cdot d\vec{\ell} = 0\) | Zero curl ↔ zero circulation ↔ path independence |
| Faraday's Law | \(\oint \vec{E}\cdot d\vec{\ell} = -d\Phi_B/dt\) | Stokes applied to \(\nabla\times\vec{E} = -\partial_t\vec{B}\) |
| Ampère's Law | \(\oint \vec{B}\cdot d\vec{\ell} = \mu_0 I_{\text{enc}}\) | Stokes applied to \(\nabla\times\vec{B} = \mu_0\vec{J}\) |
The One Thread
Stokes' theorem is about boundaries. What happens on the edge of a region is determined entirely by what happens inside it — and vice versa. Tile the interior into infinitesimal patches, each satisfying the local Stokes identity (curl = circulation per area). Sum up. Interior edges cancel. Only the boundary survives. That's the whole theorem — the rest is just making it rigorous.
In physics: curl is the local law (Faraday, Ampère). The loop integral is the global measurement (EMF, enclosed current). Stokes is the bridge between them — the theorem that lets you go from differential equations written at a point to integral laws written around a loop. All of classical electromagnetism lives in this bridge.