Module 1: Mathematical methods · Lesson 1.5 · about 35 min
Gradient, divergence and curl
What the gradient, divergence and curl of a field really measure, how to compute them in seconds, the identities JAM loves, and the traps in Cartesian and curved coordinates.
Module 1 · about 35 minutes. By the end you will be able to compute gradient, divergence and curl in your head for typical exam fields, explain what each one means physically, use the standard identities without memorising a table, and avoid the two classic traps in spherical and cylindrical coordinates.
1. One symbol, three jobs
Every formula in this lesson uses one operator, called del or nabla:
Think of it as a “vector that differentiates”. It can act on a field in exactly the three ways you can multiply with a vector:
| Operation | Acts on | Gives | Like |
|---|---|---|---|
| (gradient) | a scalar field | a vector | a number times a vector |
| (divergence) | a vector field | a scalar | the dot product |
| (curl) | a vector field | a vector | the cross product |
If you remember that table, you will never write the curl of a scalar or the gradient of a vector by mistake: those are the first things examiners use as wrong options.
2. Gradient: the direction of steepest climb
Picture a hill whose height is . At each point, the gradient is an arrow that points straight uphill, and its length is how steep the slope is there.
The rate of change of in any direction (a unit vector) is the directional derivative . It is largest along and zero along a level surface, so the gradient is always perpendicular to the surfaces = constant. That single fact gives you normals to surfaces for free: the normal to at any point is along , straight out from the centre.
- ∇φ
- an arrow pointing the way φ grows fastest
- |∇φ|
- how fast φ grows per metre in that direction
- ∇φ · n̂
- how fast φ grows if you walk along n̂ instead
- Physical meaning
- Forces push "downhill" in potential energy: F = −∇U. Heat flows down the temperature gradient. Electric field is E = −∇V.
Worked example. Find the rate of change of at along .
- at that point.
- The direction has length , so .
- .
The most common slip is step 2: forgetting to turn the direction into a unit vector, which gives 16.
3. Divergence: how much flows out
Now think of as the velocity of a fluid. Put a tiny imaginary box at a point and ask: does more fluid leave the box than enter it? The net outflow per unit volume is the divergence. Tap the picture below to move the probe square to different points and try each field.
Why is the formula just a sum of three derivatives? Step through it:
A tiny box. Take a box with sides dx, dy, dz. Fluid crosses its six faces.
The two x-faces. Out of the right face flows Fx(x + dx) · dy dz; into the left face flows Fx(x) · dy dz. The difference is (∂Fx/∂x) dx dy dz.
The other faces. In the same way, the y-faces add (∂Fy/∂y) dx dy dz and the z-faces add (∂Fz/∂z) dx dy dz.
Per unit volume. Divide the total by the volume dx dy dz: net outflow per volume = ∂Fx/∂x + ∂Fy/∂y + ∂Fz/∂z. That is ∇·F.
Add up all the boxes. Inner faces cancel between neighbours, so the total over a big region is the flow through its outer surface. That is the divergence theorem (lesson 1.6).
A result worth knowing by heart. For and :
Put : everywhere except the origin. That is the field of a point charge, and it is why Gauss’s law gives the same flux through any surface around the charge.
4. Curl: how much it swirls
Drop a tiny paddle wheel into the flow. If the water on one side moves faster than on the other, the wheel turns. The curl is a vector along the wheel’s axis whose size is (twice) how fast it spins.
Two fields every student should picture:
- Rigid rotation , for example with : curl , divergence 0.
- Shear flow : the lines are all straight, yet the curl is , because the top of the paddle wheel is pushed harder than the bottom.
“Straight streamlines mean zero curl.” Wrong. Curl is about whether neighbouring layers move at different speeds, not whether the flow goes round in circles. The shear flow has straight lines and non-zero curl; the field around a wire goes round in circles and has zero curl everywhere except on the wire.
A field with zero curl everywhere (in a region without holes) is conservative: it is the gradient of some potential, and line integrals depend only on the end points. JAM uses this every year: check the curl first, find the potential, and a long path integral becomes a subtraction.
5. The second derivatives and the identities
Applying twice gives five combinations. Two are always zero, and you can see why without algebra:
| Combination | Value | Why |
|---|---|---|
| always | a slope field has no swirl (mixed partials are equal) | |
| always | a pure swirl has no source | |
| the Laplacian | zero only for harmonic | |
| “BAC minus CAB” | ||
| not simpler |
Product rules work like ordinary ones, keeping the order of vectors in cross products. The ones that save the most time:
6. Curved coordinates: the two traps
In spherical coordinates the divergence is not just :
Trap 1. For the naive answer is wrong; the correct matches . The extra factors are there because the unit vectors themselves change direction from point to point.
Trap 2. The field has zero divergence for , but its flux through any closed surface around the origin is . The whole source is concentrated at the origin: . If the origin is outside your surface, the flux is zero.
- acts like a vector: gradient (on scalars), divergence (dot), curl (cross). Never mix their inputs and outputs.
- The gradient points uphill and is perpendicular to level surfaces; the directional derivative is with a unit vector.
- Divergence is net outflow per unit volume; curl is local spin (twice the angular velocity for rigid rotation).
- Curl of a gradient and divergence of a curl are always zero. .
- Zero curl in a region without holes means conservative: find a potential and subtract.
- In spherical and cylindrical coordinates use the full formulas, and watch for sources hidden at a singular point.
Practise: exam questions on this lesson
These start easy and climb to top-rank level. Check each answer, then read the full solution. More on the vector calculus practice page.
Next: lesson 1.6 turns these local ideas into the big integral theorems of Gauss, Stokes and Green, which you will use all through electromagnetism. Going on to GATE? The GATE Physics course builds on this with tensors and curvilinear coordinates in general.
Check yourself
Sources
- D. J. Griffiths, Introduction to Electrodynamics, 4th ed. (2017), chapter 1: vector analysis
- M. L. Boas, Mathematical Methods in the Physical Sciences, 3rd ed. (2006), chapter 6
- H. M. Schey, Div, Grad, Curl, and All That, 4th ed. (2005)
- IIT JAM Physics syllabus (Mathematical Methods section), published by the organising IIT each year
Your ticks stay in this browser. Save them as a file on the course home.
Updated 2026-11-22. Not affiliated with IIT or the JAM organising institute. Exam pattern as published for recent years: always check the official JAM website before the exam. All questions are original.