Gradient, Divergence, Curl, and Laplacian

Definitions, Physical Meaning, and Identities

Contents

  1. The Nabla Operator
  2. Gradient (grad)
  3. Divergence (div)
  4. Curl (rot)
  5. The Laplacian
  6. Directional Derivative
  7. Summary
  8. Related Pages

1. The Nabla Operator

Definition: The nabla operator $\nabla$

In a three-dimensional Cartesian coordinate system, the nabla operator is the vector differential operator defined by

$$\nabla = \left( \dfrac{\partial}{\partial x},\; \dfrac{\partial}{\partial y},\; \dfrac{\partial}{\partial z} \right)$$

Although $\nabla$ is not a vector in itself, treating it formally as a vector lets us describe the gradient, divergence, and curl in a unified way. Acting on a scalar field it produces a vector field (the gradient); its dot product with a vector field gives the divergence; and its cross product with a vector field gives the curl.

$\nabla$ is an operator and acts on whatever stands to its right. The expressions $\nabla f$ and $f \nabla$ have different meanings, so the order matters.

2. Gradient (grad)

Definition: gradient

The gradient of a scalar field $f(x, y, z)$ is the vector field

$$\operatorname{grad} f = \nabla f = \left( \dfrac{\partial f}{\partial x},\; \dfrac{\partial f}{\partial y},\; \dfrac{\partial f}{\partial z} \right)$$

Physical meaning

  • Direction: $\nabla f$ points in the direction in which $f$ increases most rapidly.
  • Magnitude: $|\nabla f|$ equals the maximum rate of change in that direction.
  • On a level surface $f = \text{const}$, $\nabla f$ is everywhere normal to the surface.

Relation to the total differential

The total differential of $f$ is the dot product of the gradient with the infinitesimal displacement $d\mathbf{r} = (dx, dy, dz)$:

$$df = \nabla f \cdot d\mathbf{r} = \dfrac{\partial f}{\partial x} dx + \dfrac{\partial f}{\partial y} dy + \dfrac{\partial f}{\partial z} dz$$

Example: gravitational potential

For a point mass $M$ located at the origin, the gravitational potential is

$$\varphi = -\dfrac{GM}{r}, \quad r = \sqrt{x^2 + y^2 + z^2}$$

The gravitational force is $\mathbf{F} = -\nabla \varphi$, which points in the direction of decreasing potential (i.e., toward the mass).

$$\mathbf{F} = -\nabla \varphi = -\dfrac{GM}{r^3} \mathbf{r}$$

Property: the curl of a gradient is zero

For any $C^2$ scalar field $f$,

$$\nabla \times (\nabla f) = \mathbf{0}$$

That is, every gradient field is irrotational.

Proof

Compute each component of the curl of $\nabla f = \left(\dfrac{\partial f}{\partial x},\, \dfrac{\partial f}{\partial y},\, \dfrac{\partial f}{\partial z}\right)$.

The $x$-component:

$$\dfrac{\partial}{\partial y}\!\left(\dfrac{\partial f}{\partial z}\right) - \dfrac{\partial}{\partial z}\!\left(\dfrac{\partial f}{\partial y}\right) = \dfrac{\partial^2 f}{\partial y\,\partial z} - \dfrac{\partial^2 f}{\partial z\,\partial y}$$

Since $f$ is of class $C^2$, Schwarz's theorem (equality of mixed partial derivatives) gives

$$\dfrac{\partial^2 f}{\partial y\,\partial z} = \dfrac{\partial^2 f}{\partial z\,\partial y}$$

so the $x$-component vanishes. The $y$- and $z$-components vanish in the same way:

$$y\text{-component}: \quad \dfrac{\partial^2 f}{\partial z\,\partial x} - \dfrac{\partial^2 f}{\partial x\,\partial z} = 0$$ $$z\text{-component}: \quad \dfrac{\partial^2 f}{\partial x\,\partial y} - \dfrac{\partial^2 f}{\partial y\,\partial x} = 0$$

Hence $\nabla \times (\nabla f) = \mathbf{0}$. $\blacksquare$

3. Divergence (div)

Definition: divergence

The divergence of a vector field $\mathbf{A} = (A_x, A_y, A_z)$ is the scalar field

$$\operatorname{div} \mathbf{A} = \nabla \cdot \mathbf{A} = \dfrac{\partial A_x}{\partial x} + \dfrac{\partial A_y}{\partial y} + \dfrac{\partial A_z}{\partial z}$$

Physical meaning

  • $\nabla \cdot \mathbf{A} > 0$: the point is a source (net outflow).
  • $\nabla \cdot \mathbf{A} < 0$: the point is a sink (net inflow).
  • $\nabla \cdot \mathbf{A} = 0$: the field is solenoidal (incompressible); inflow and outflow balance.

Example: divergence of the position vector

The divergence of the position vector $\mathbf{r} = (x, y, z)$ is

$$\nabla \cdot \mathbf{r} = \dfrac{\partial x}{\partial x} + \dfrac{\partial y}{\partial y} + \dfrac{\partial z}{\partial z} = 3$$

and, away from the origin ($r \neq 0$),

$$\nabla \cdot \left( \dfrac{\mathbf{r}}{r^3} \right) = 0$$

This expresses the fact that an inverse-square field (such as the gravitational or electrostatic field of a point source) has no divergence except at the source itself.

Example: the continuity equation

The conservation law for a density $\rho$ and flux density $\mathbf{j}$ is

$$\dfrac{\partial \rho}{\partial t} + \nabla \cdot \mathbf{j} = 0$$

This continuity equation appears throughout physics, including conservation of charge and conservation of mass.

Property: the divergence of a curl is zero

For any $C^2$ vector field $\mathbf{A}$,

$$\nabla \cdot (\nabla \times \mathbf{A}) = 0$$

That is, every curl field is solenoidal (source-free).

Proof

The components of $\nabla \times \mathbf{A}$ are

$$\nabla \times \mathbf{A} = \left( \dfrac{\partial A_z}{\partial y} - \dfrac{\partial A_y}{\partial z},\;\; \dfrac{\partial A_x}{\partial z} - \dfrac{\partial A_z}{\partial x},\;\; \dfrac{\partial A_y}{\partial x} - \dfrac{\partial A_x}{\partial y} \right)$$

Take the divergence:

$$\nabla \cdot (\nabla \times \mathbf{A}) = \dfrac{\partial}{\partial x}\!\left(\dfrac{\partial A_z}{\partial y} - \dfrac{\partial A_y}{\partial z}\right) + \dfrac{\partial}{\partial y}\!\left(\dfrac{\partial A_x}{\partial z} - \dfrac{\partial A_z}{\partial x}\right) + \dfrac{\partial}{\partial z}\!\left(\dfrac{\partial A_y}{\partial x} - \dfrac{\partial A_x}{\partial y}\right)$$

Expanding gives six terms:

$$= \dfrac{\partial^2 A_z}{\partial x\,\partial y} - \dfrac{\partial^2 A_y}{\partial x\,\partial z} + \dfrac{\partial^2 A_x}{\partial y\,\partial z} - \dfrac{\partial^2 A_z}{\partial y\,\partial x} + \dfrac{\partial^2 A_y}{\partial z\,\partial x} - \dfrac{\partial^2 A_x}{\partial z\,\partial y}$$

Since each component of $\mathbf{A}$ is $C^2$, Schwarz's theorem lets us interchange the order of partial differentiation. Pairing terms with the same component yields

$$= \underbrace{\left(\dfrac{\partial^2 A_z}{\partial x\,\partial y} - \dfrac{\partial^2 A_z}{\partial y\,\partial x}\right)}_{=\,0} + \underbrace{\left(\dfrac{\partial^2 A_x}{\partial y\,\partial z} - \dfrac{\partial^2 A_x}{\partial z\,\partial y}\right)}_{=\,0} + \underbrace{\left(\dfrac{\partial^2 A_y}{\partial z\,\partial x} - \dfrac{\partial^2 A_y}{\partial x\,\partial z}\right)}_{=\,0}$$

Hence $\nabla \cdot (\nabla \times \mathbf{A}) = 0$. $\blacksquare$

4. Curl (rot)

Definition: curl (also written rot)

The curl of a vector field $\mathbf{A} = (A_x, A_y, A_z)$ is the vector field

$$\operatorname{curl} \mathbf{A} = \nabla \times \mathbf{A} = \begin{vmatrix} \mathbf{e}_x & \mathbf{e}_y & \mathbf{e}_z \\ \dfrac{\partial}{\partial x} & \dfrac{\partial}{\partial y} & \dfrac{\partial}{\partial z} \\ A_x & A_y & A_z \end{vmatrix}$$

Component form

Expanding the determinant gives the components

$$\nabla \times \mathbf{A} = \left( \dfrac{\partial A_z}{\partial y} - \dfrac{\partial A_y}{\partial z},\; \dfrac{\partial A_x}{\partial z} - \dfrac{\partial A_z}{\partial x},\; \dfrac{\partial A_y}{\partial x} - \dfrac{\partial A_x}{\partial y} \right)$$

Physical meaning

  • $\nabla \times \mathbf{A}$ measures the local circulation density (vorticity) of the vector field.
  • Direction: the axis of rotation of the local swirl (right-hand rule).
  • Magnitude: the strength of the swirl.

Example: rigid-body rotation

The velocity field of a rigid body rotating with angular velocity $\boldsymbol{\Omega}$ is

$$\mathbf{A} = \boldsymbol{\Omega} \times \mathbf{r}$$

Its curl is

$$\nabla \times \mathbf{A} = 2\boldsymbol{\Omega}$$

so the vorticity is twice the angular velocity.

Property: the curl of a gradient is zero (recap)

For any $C^2$ scalar field $f$,

$$\nabla \times (\nabla f) = \mathbf{0}$$

Conversely, on a simply connected domain, if $\nabla \times \mathbf{A} = \mathbf{0}$ then there exists a scalar potential $f$ with $\mathbf{A} = \nabla f$.

5. The Laplacian

Definition: scalar Laplacian

The Laplacian of a scalar field $f$ is defined as the divergence of the gradient:

$$\nabla^2 f = \operatorname{div}(\operatorname{grad} f) = \nabla \cdot (\nabla f) = \dfrac{\partial^2 f}{\partial x^2} + \dfrac{\partial^2 f}{\partial y^2} + \dfrac{\partial^2 f}{\partial z^2}$$

Definition: vector Laplacian

The Laplacian of a vector field $\mathbf{A} = (A_x, A_y, A_z)$ is obtained by applying the scalar Laplacian to each component:

$$\nabla^2 \mathbf{A} = (\nabla^2 A_x,\; \nabla^2 A_y,\; \nabla^2 A_z)$$

Note: $\nabla^2 = \nabla \cdot \nabla = \operatorname{div} \circ \operatorname{grad}$. The Laplacian is the divergence of the gradient, not the "gradient of the divergence."

The Laplacian in physics

Laplace's equation

A function whose Laplacian vanishes is called a harmonic function.

$$\nabla^2 f = 0$$

The electrostatic potential in a charge-free region and a steady-state temperature distribution both satisfy this equation.

Heat equation

The time evolution of a temperature $u$, with thermal diffusivity $\alpha$, obeys

$$\dfrac{\partial u}{\partial t} = \alpha \nabla^2 u$$

The Laplacian represents the deviation from the local average: at a point that is colder than its surroundings, $\nabla^2 u > 0$ and the temperature rises.

Wave equation

For a wave with propagation speed $c$,

$$\dfrac{\partial^2 u}{\partial t^2} = c^2 \nabla^2 u$$

This describes a wide range of wave phenomena, including sound waves, electromagnetic waves, and the vibrations of a string.

6. Directional Derivative

Definition: directional derivative

The directional derivative of a scalar field $f$ along a unit vector $\hat{\mathbf{n}}$ is

$$D_{\hat{\mathbf{n}}} f = \nabla f \cdot \hat{\mathbf{n}} = |\nabla f| \cos\theta$$

where $\theta$ is the angle between $\nabla f$ and $\hat{\mathbf{n}}$.

The directional derivative attains its maximum value $|\nabla f|$ when $\hat{\mathbf{n}}$ is aligned with $\nabla f$, and its minimum value $-|\nabla f|$ in the opposite direction. It is $0$ in any direction orthogonal to $\nabla f$ (i.e., along the level surface).

The operator $(\mathbf{A} \cdot \nabla)$ on a vector field

Acting on a vector field $\mathbf{B}$, the operator $(\mathbf{A} \cdot \nabla)$ differentiates each component of $\mathbf{B}$ along the direction of $\mathbf{A}$:

$$(\mathbf{A} \cdot \nabla) \mathbf{B} = \left( \mathbf{A} \cdot \nabla B_x,\; \mathbf{A} \cdot \nabla B_y,\; \mathbf{A} \cdot \nabla B_z \right)$$

This advection operator plays a central role in the convective term $(\mathbf{v} \cdot \nabla)\mathbf{v}$ of fluid mechanics and the Navier-Stokes equations.

Summary

The four basic operations of vector calculus are summarized below.

OperationNotationInput → OutputPhysical meaning
Gradient (grad)$\nabla f$scalar → vectordirection of steepest ascent
Divergence (div)$\nabla \cdot \mathbf{A}$vector → scalarstrength of source/sink
Curl$\nabla \times \mathbf{A}$vector → vectorstrength and axis of rotation
Laplacian$\nabla^2 f$scalar → scalardeviation from local average

Fundamental identities

  • $\nabla \times (\nabla f) = \mathbf{0}$ (every gradient field is irrotational)
  • $\nabla \cdot (\nabla \times \mathbf{A}) = 0$ (every curl field is solenoidal)
  • $\nabla^2 f = \nabla \cdot (\nabla f)$ (Laplacian = divergence of the gradient)