Why Closed Surfaces Have Zero Total Curl Flux
A vector field can have nonzero curl at many points on a surface, yet its total curl flux through a closed surface is zero, provided the field is smooth in a neighborhood of that surface. This is a direct consequence of Stokes' theorem: the total flux of curl through a surface is determined by circulation around its boundary, and a closed surface has no boundary curve.
The conclusion concerns the signed total flux. It does not say that the curl vanishes at every point, or that the flux through every portion of the surface is zero. Different portions can contribute nonzero amounts that cancel when the entire surface is included.
What “Closed Surface” Means
A surface is closed here if it has no boundary curves. A sphere is closed. So is the complete boundary of a solid cylinder: its curved side together with its two circular caps. By contrast, a disk, a hemisphere on its own, or just the curved side of a cylinder has a boundary and is not a closed surface.
An orientation is a consistent choice of unit normal on the surface. For a sphere or the boundary of a solid cylinder, the outward orientation points away from the enclosed solid. On a closed surface, the surface integral adds the flux using this chosen normal, so reversing the orientation reverses the sign of the integral. It does not change the fact that the total is zero.
Stokes' theorem applies to an oriented, piecewise smooth surface when the vector field is continuously differentiable on an open set containing the surface. The neighborhood condition matters: the components of \(\mathbf{F}\), and therefore its curl, must be defined and sufficiently smooth at the surface and nearby. Under these hypotheses, Stokes' theorem states
The integral on the left is the flux of \(\nabla\times\mathbf{F}\) through \(S\). The integral on the right is the circulation of \(\mathbf{F}\) around the oriented boundary \(\partial S\). For a closed surface, \(\partial S\) is empty, so there is no boundary curve to traverse and the boundary integral is zero.
The Deduction from Stokes' Theorem
Take an oriented, piecewise smooth closed surface \(S\), and suppose \(\mathbf{F}\) is continuously differentiable on an open neighborhood of \(S\). Stokes' theorem is therefore applicable.
The curl flux through \(S\) equals the circulation around \(\partial S\): \(\iint_S(\nabla\times\mathbf{F})\cdot\mathbf{n}\,dS=\oint_{\partial S}\mathbf{F}\cdot d\mathbf{r}\).
Since \(S\) is closed, \(\partial S=\varnothing\). There are no boundary curves, and the boundary integral is zero.
Substitution into Stokes' theorem gives \(\iint_S(\nabla\times\mathbf{F})\cdot\mathbf{n}\,dS=0\).
This is a complete deduction: it uses Stokes' theorem and the definition of a closed surface. If a closed surface is handled as several oriented smooth pieces, the curves where pieces meet occur with opposite induced orientations on adjacent pieces. Their boundary contributions cancel, leaving no boundary contribution for the whole surface.
Worked Example: Curl Flux Through a Sphere
Worked Example: Curl Flux Through a Sphere
Let \(S\) be the sphere of radius \(R>0\) centered at the origin, oriented outward, and let \(\mathbf{F}(x,y,z)=(-y,x,0)\). Compute the curl flux directly and check the closed-surface result.
First compute the curl. For \(\mathbf{F}=(F_1,F_2,F_3)=(-y,x,0)\), \[ \nabla\times\mathbf{F} = \left( \frac{\partial F_3}{\partial y}-\frac{\partial F_2}{\partial z}, \frac{\partial F_1}{\partial z}-\frac{\partial F_3}{\partial x}, \frac{\partial F_2}{\partial x}-\frac{\partial F_1}{\partial y} \right). \] The derivatives are \(\partial F_3/\partial y=0\), \(\partial F_2/\partial z=0\), \(\partial F_1/\partial z=0\), \(\partial F_3/\partial x=0\), \(\partial F_2/\partial x=1\), and \(\partial F_1/\partial y=-1\). Thus \[ \nabla\times\mathbf{F}=(0,0,1-(-1))=(0,0,2). \]
Use spherical coordinates on the sphere: \[ \mathbf{r}(\theta,\phi) =(R\sin\theta\cos\phi,\ R\sin\theta\sin\phi,\ R\cos\theta), \quad 0\leq\theta\leq\pi,\quad 0\leq\phi\leq2\pi. \] The outward unit normal is \(\mathbf{n}=\mathbf{r}/R=(\sin\theta\cos\phi,\sin\theta\sin\phi,\cos\theta)\), and the surface element is \(dS=R^2\sin\theta\,d\theta\,d\phi\). Therefore \[ (\nabla\times\mathbf{F})\cdot\mathbf{n}=2\cos\theta. \] The flux is \[ \begin{aligned} \iint_S(\nabla\times\mathbf{F})\cdot\mathbf{n}\,dS &=\int_0^{2\pi}\int_0^\pi 2\cos\theta\,R^2\sin\theta\,d\theta\,d\phi\\ &=2R^2\left(\int_0^{2\pi}d\phi\right) \left(\int_0^\pi\cos\theta\sin\theta\,d\theta\right)\\ &=2R^2(2\pi)\left[\frac12\sin^2\theta\right]_0^\pi\\ &=4\pi R^2(0-0)=0. \end{aligned} \] This verifies the result directly. It also agrees with Stokes' theorem because the sphere has no boundary and \(\mathbf{F}\) is smooth everywhere.
Worked Example: Constant Curl on a Closed Cylinder
Worked Example: Constant Curl on a Closed Cylinder
Let \(S\) be the complete boundary of a cylinder of radius \(a>0\), extending from \(z=0\) to \(z=h\), where \(h>0\). Orient \(S\) outward, and again take \(\mathbf{F}(x,y,z)=(-y,x,0)\). Compute the flux separately on the curved side, top, and bottom.
As in the sphere calculation, \(\nabla\times\mathbf{F}=(0,0,2)\). On the curved side, use \[ \mathbf{r}(\theta,z)=(a\cos\theta,a\sin\theta,z), \quad 0\leq\theta\leq2\pi,\quad 0\leq z\leq h. \] The outward unit normal is \(\mathbf{n}=(\cos\theta,\sin\theta,0)\), and \(dS=a\,d\theta\,dz\). Since \[ (\nabla\times\mathbf{F})\cdot\mathbf{n} =(0,0,2)\cdot(\cos\theta,\sin\theta,0)=0, \] the side contribution is \[ \int_0^h\int_0^{2\pi}0\cdot a\,d\theta\,dz=0. \]
On the top disk, the outward unit normal is \((0,0,1)\), so the integrand is \(2\). The disk has area \(\pi a^2\), giving a top contribution \(2\pi a^2\). On the bottom disk, the outward unit normal is \((0,0,-1)\), so the integrand is \(-2\). Its contribution is \(-2\pi a^2\). Adding all three parts gives \[ 0+2\pi a^2-2\pi a^2=0. \] The calculation checks the theorem: the side contributes zero, while the oppositely oriented caps contribute equal and opposite amounts.
Worked Example: A Nonconstant Curl on a Closed Cylinder
Worked Example: A Nonconstant Curl on a Closed Cylinder
Use the same outward-oriented cylinder, but let \(\mathbf{F}(x,y,z)=(0,0,xy)\). This field is smooth everywhere. Compute the curl flux on each part of the surface.
The curl is \[ \nabla\times\mathbf{F} = \left( \frac{\partial(xy)}{\partial y}, -\frac{\partial(xy)}{\partial x}, 0 \right) =(x,-y,0). \] On the top and bottom disks, the normals are vertical, while the curl has zero vertical component. Thus both cap contributions are zero.
On the side, \(x=a\cos\theta\), \(y=a\sin\theta\), the outward unit normal is \((\cos\theta,\sin\theta,0)\), and \(dS=a\,d\theta\,dz\). Hence \[ (\nabla\times\mathbf{F})\cdot\mathbf{n} =a\cos^2\theta-a\sin^2\theta. \] The side flux is therefore \[ \begin{aligned} \int_0^h\int_0^{2\pi} a(\cos^2\theta-\sin^2\theta)a\,d\theta\,dz &=a^2h\int_0^{2\pi}(\cos^2\theta-\sin^2\theta)\,d\theta\\ &=a^2h\int_0^{2\pi}\cos(2\theta)\,d\theta\\ &=a^2h\left[\frac12\sin(2\theta)\right]_0^{2\pi}\\ &=a^2h(0-0)=0. \end{aligned} \] The cap fluxes are also zero, so the total flux is \(0+0+0=0\). Here the curl is not constant, and its normal component varies around the side, but the positive and negative contributions cancel in the full integral. This direct computation agrees with Stokes' theorem.
A Common Mistake: Treating Both Caps as Outward
On the complete cylinder, the outward normal on the top is upward, but the outward normal on the bottom is downward. A frequent error is to use the upward normal on both caps. For the field \(\mathbf{F}=(-y,x,0)\), the curl is \((0,0,2)\), so that incorrect choice would give \(2\pi a^2\) for each cap and the incorrect total \[ 0+2\pi a^2+2\pi a^2=4\pi a^2. \] The correction is to use the actual outward normal on each piece: the bottom contribution is \((0,0,2)\cdot(0,0,-1)\) times the disk area, or \(-2\pi a^2\). The correct total is \(0+2\pi a^2-2\pi a^2=0\).
Another related error is to call only the curved side of a cylinder a closed surface. That side has two boundary circles. Stokes' theorem on the side alone relates its curl flux to circulation around those circles; the zero conclusion applies to the complete closed surface, including both caps.
Why the Result Matters
The closed-surface result is a useful check on a flux calculation. If a smooth field is defined near an entire closed oriented surface, a computed nonzero total curl flux signals that something needs attention: perhaps a piece was omitted, a normal was given the wrong direction, or a surface integral was evaluated incorrectly.
The smoothness condition cannot be dropped casually. If a field is undefined or nonsmooth at a point on the surface, the hypotheses used to apply Stokes' theorem may fail. The conclusion is guaranteed by the theorem only when its assumptions hold. For the examples above, the fields are polynomial expressions, so they are smooth on all of \(\mathbb{R}^3\), and the required condition is satisfied.
Check Your Understanding
Use the definitions, orientation choices, and computations in this tutorial to answer the following.
- What property of a surface makes it closed in the sense used here?
- Which smoothness condition on \(\mathbf{F}\) is needed to apply Stokes' theorem?
- On the boundary of a solid cylinder, what are the outward unit normals on the top and bottom disks?
- For \(\mathbf{F}=(-y,x,0)\), why do the top and bottom curl-flux contributions cancel?
- Does zero total curl flux imply that \(\nabla\times\mathbf{F}\) is zero at every point of the closed surface? Explain.