Tutorials › Calculus III › Irrotational Fields and Zero Curl

Home  /  Tutorials  /  Curl and Divergence

Curl and Divergence · Tutorial 920 of 999

Irrotational Fields and Zero Curl

Compute and interpret curl, distinguish local irrotationality from global conservativeness, and see how the domain affects whether a potential exists.

Advanced 10 min read

What You'll Learn

  • Define an irrotational vector field using its curl.
  • Compute curl component by component and test whether it vanishes throughout a domain.
  • Interpret curl as twice the local angular-velocity vector for a differentiable velocity field.
  • Verify that constant and gradient-generated fields have zero curl.
  • Identify why zero curl does not guarantee a global potential on every domain.
  • Test a punctured-domain example by calculating its circulation around a closed curve.

What Does It Mean for a Field to Be Irrotational?

A vector field can move points through space without making them spin locally. Curl describes the rotational part of a field’s first-order variation: if the curl vanishes, that local rotational component is absent. The field itself need not vanish, and it may still stretch, compress, or shear nearby points.

For a differentiable vector field \(\mathbf{v}=(P,Q,R)\) on an open subset \(D\) of \(\mathbb{R}^3\), its curl is

$$ \nabla\times\mathbf{v} = \left( \frac{\partial R}{\partial y}-\frac{\partial Q}{\partial z}, \frac{\partial P}{\partial z}-\frac{\partial R}{\partial x}, \frac{\partial Q}{\partial x}-\frac{\partial P}{\partial y} \right). $$

The curl is itself a vector field. Its direction indicates the axis of local rotation, while its magnitude measures the strength of the rotational component. In two dimensions, for a field \(\mathbf{v}=(P,Q)\), the scalar curl is \(\frac{\partial Q}{\partial x}-\frac{\partial P}{\partial y}\); it is the \(z\)-component of the corresponding three-dimensional curl when the field is extended with zero \(z\)-component and no \(z\)-dependence.

Definition: A differentiable vector field \(\mathbf{v}\) is irrotational on \(D\) if \(\nabla\times\mathbf{v}=\mathbf{0}\) at every point of \(D\). Saying the field has zero curl means this equality holds throughout the specified domain.

The domain matters. A formula may have zero curl wherever it is defined but fail to be defined at certain points. The test is applied only on the field’s domain; it does not make a singular point part of that domain.

Why Curl Measures Local Rotation

For a differentiable velocity field, examine how the velocity changes near a point \(\mathbf{x}\). Let \(J\) be its Jacobian matrix at that point. For a small displacement \(\mathbf{h}\), differentiability gives the first-order approximation \(\mathbf{v}(\mathbf{x}+\mathbf{h})\approx\mathbf{v}(\mathbf{x})+J\mathbf{h}\). The matrix \(J\) splits into a symmetric part and an antisymmetric part:

$$ J=\frac{J+J^{\top}}{2}+\frac{J-J^{\top}}{2}. $$

The symmetric part describes local stretching and shearing. The antisymmetric part describes the local rigid-rotation component. If \(\boldsymbol{\omega}=\frac12(\nabla\times\mathbf{v})\), direct comparison of matrix entries shows that the antisymmetric part acts on a displacement as \(\mathbf{h}\mapsto\boldsymbol{\omega}\times\mathbf{h}\). Thus the local angular-velocity vector is half the curl.

This explains the word irrotational: when curl is zero, the antisymmetric part of the first-order change vanishes. It does not follow that \(J\) is zero, since its symmetric part can remain nonzero. A flow can therefore deform nearby shapes even when it has no local rigid-rotation component.

Interpretation: Zero curl rules out local rotational motion in the first-order behavior of a differentiable field. It does not mean that the field is zero, constant, or free of all motion.

Worked Example: A Constant Field

Worked Example: A Constant Field

Let \(\mathbf{v}=(3,-2,5)\) on all of \(\mathbb{R}^3\). Each component is constant, so all its partial derivatives are zero:

$$ \frac{\partial R}{\partial y}=0,\quad \frac{\partial Q}{\partial z}=0,\quad \frac{\partial P}{\partial z}=0,\quad \frac{\partial R}{\partial x}=0,\quad \frac{\partial Q}{\partial x}=0,\quad \frac{\partial P}{\partial y}=0. $$

Substituting these values into the curl formula gives

$$ \nabla\times\mathbf{v} = (0-0,\;0-0,\;0-0) = (0,0,0). $$

Therefore the field is irrotational. It is also generated by the scalar potential \(\phi(x,y,z)=3x-2y+5z\), since \(\nabla\phi=(3,-2,5)=\mathbf{v}\). Differentiating this potential recovers every component of the field, which checks the claim.

Gradient-Generated Fields

A field of the form \(\mathbf{v}=\nabla\phi\), where \(\phi\) is a scalar function, is called a gradient field or a field generated by a potential. If \(\phi\) has continuous second partial derivatives, then its mixed partial derivatives agree. The identity \(\nabla\times(\nabla\phi)=\mathbf{0}\), established in the earlier material on curl of a gradient, then shows that every such field has zero curl.

Worked Example: Checking a Gradient-Generated Field

Let \(\phi(x,y,z)=x^2y+yz^2\), and define \(\mathbf{v}=\nabla\phi\). Compute the gradient component by component:

$$ P=\frac{\partial\phi}{\partial x}=2xy,\qquad Q=\frac{\partial\phi}{\partial y}=x^2+z^2,\qquad R=\frac{\partial\phi}{\partial z}=2yz. $$

Now compute each component of curl:

$$ \begin{aligned} (\nabla\times\mathbf{v})_1 &=\frac{\partial R}{\partial y}-\frac{\partial Q}{\partial z} =2z-2z=0,\\ (\nabla\times\mathbf{v})_2 &=\frac{\partial P}{\partial z}-\frac{\partial R}{\partial x} =0-0=0,\\ (\nabla\times\mathbf{v})_3 &=\frac{\partial Q}{\partial x}-\frac{\partial P}{\partial y} =2x-2x=0. \end{aligned} $$

Hence \(\nabla\times\mathbf{v}=(0,0,0)\). As a check, the mixed derivatives responsible for the cancellations are \(\phi_{yx}=\phi_{xy}=2x\), \(\phi_{zx}=\phi_{xz}=0\), and \(\phi_{zy}=\phi_{yz}=2z\). The field is irrotational everywhere because the polynomial potential is defined and has continuous second derivatives everywhere.

Zero Curl and the Existence of a Potential

The implication from a potential to zero curl holds under the differentiability condition just stated. The reverse implication needs care: zero curl by itself does not guarantee one potential defined throughout the entire domain. A useful converse theorem requires a condition on the domain.

Theorem: If \(D\) is a simply connected open subset of \(\mathbb{R}^3\) and \(\mathbf{v}\) has continuous first partial derivatives on \(D\), then \(\nabla\times\mathbf{v}=\mathbf{0}\) throughout \(D\) implies that \(\mathbf{v}\) is a gradient field on \(D\). Simply connected means, informally, that loops in the domain can be continuously contracted to a point without leaving the domain.

The domain condition cannot simply be dropped. If a hole or excluded axis is present, a field may have zero curl everywhere it is defined yet still lack a single-valued potential on the whole domain. The following example demonstrates this distinction directly, without using the theorem as a substitute for checking the field.

Worked Example: Zero Curl Without a Global Potential

Consider the field

$$ \mathbf{v}(x,y,z) = \left( -\frac{y}{x^2+y^2}, \frac{x}{x^2+y^2}, 0 \right), \qquad x^2+y^2>0. $$

Its domain is \(\mathbb{R}^3\) with the \(z\)-axis removed. Write \(P=-y/(x^2+y^2)\), \(Q=x/(x^2+y^2)\), and \(R=0\). The first two curl components vanish because \(P\) and \(Q\) do not depend on \(z\), and \(R\) is identically zero. For the third component, differentiate explicitly:

$$ \begin{aligned} \frac{\partial Q}{\partial x} &=\frac{(x^2+y^2)-2x^2}{(x^2+y^2)^2} =\frac{y^2-x^2}{(x^2+y^2)^2},\\ \frac{\partial P}{\partial y} &=\frac{-(x^2+y^2)+2y^2}{(x^2+y^2)^2} =\frac{y^2-x^2}{(x^2+y^2)^2}. \end{aligned} $$

Thus \(\frac{\partial Q}{\partial x}-\frac{\partial P}{\partial y}=0\), and all three components of curl are zero at every point in the domain.

Nevertheless, the field has no global potential on this domain. To see why, consider the unit circle in the plane \(z=0\), parametrized by \(\mathbf{r}(t)=(\cos t,\sin t,0)\) for \(0\leq t\leq2\pi\). Along it, \(x^2+y^2=1\), so \(\mathbf{v}(\mathbf{r}(t))=(-\sin t,\cos t,0)\), while \(\mathbf{r}'(t)=(-\sin t,\cos t,0)\). Their dot product is \(\sin^2t+\cos^2t=1\). Therefore the circulation around this closed curve is

$$ \int_0^{2\pi}\mathbf{v}(\mathbf{r}(t))\cdot\mathbf{r}'(t)\,dt = \int_0^{2\pi}1\,dt = 2\pi. $$

If \(\mathbf{v}=\nabla\phi\) for a single-valued potential on the domain, the chain rule would make this integral \(\phi(\mathbf{r}(2\pi))-\phi(\mathbf{r}(0))=0\), since the curve ends where it starts. The computed value \(2\pi\) contradicts that consequence. So this field is irrotational on its domain but not conservative there.

How to Test for Irrotationality

To classify a field, first identify the domain where the component formulas are defined. Then compute the curl and determine whether every component vanishes at every point in that domain. A curl that vanishes only at some points does not make the field irrotational on the whole domain. Nor does a formula that is undefined somewhere pass a test at that excluded point.

1
Identify the domain.
Record where denominators, coordinate expressions, or other components are defined.
2
Compute all curl components.
Use the standard component formula, keeping the order of each difference.
3
Check the result throughout the domain.
If every component is identically zero there, the field is irrotational on that domain.
4
Separate local and global conclusions.
Zero curl gives local irrotationality. A global potential requires suitable domain conditions as well.

A Common Mistake: Confusing Zero Curl with a Zero Field

Zero curl does not mean that all components of the field equal zero. For the constant field \((3,-2,5)\), for example, the field is nonzero at every point, even though its curl vanishes. The wrong reasoning would be to say that zero curl forces \(P=Q=R=0\); curl involves differences of partial derivatives, not the component values themselves.

A related error is to conclude that every zero-curl field has a global potential. The punctured-domain example disproves this: its curl is zero everywhere in its domain, but its circulation around the unit circle is \(2\pi\), which is incompatible with a global potential. Check the domain before applying a converse result.

Why the Distinction Matters

Irrotationality is a local differential property: it concerns the curl at each point. Conservativeness is a global property: it asks whether one potential function describes the field throughout its domain. On an appropriate simply connected domain, continuous first derivatives and zero curl guarantee such a potential. On a domain with a hole, the local curl test alone may not settle the global question.

Keeping these statements separate makes curl tests more informative. You can identify the absence of local rotational tendency without claiming that the field is motionless, and you can recognize when domain geometry prevents a global potential even though the curl vanishes everywhere the field is defined.

Check Your Understanding

Use the definitions and distinctions in this tutorial to answer the questions.

  1. What condition on the curl defines an irrotational field on a domain?
  2. For a differentiable velocity field, how is the local angular-velocity vector related to curl?
  3. Why does a field generated by a twice continuously differentiable potential have zero curl?
  4. What additional domain condition appears in the stated theorem that gives a potential from zero curl?
  5. Why does the nonzero circulation around the unit circle rule out a global potential for the punctured-domain example?

Previous

Next

Back to Tutorials