Vector calculus

In short

Vector calculus extends differentiation and integration to vector fields — functions that assign a vector (velocity, force, electric field) to every point of space. Its three derivative operators (gradient, divergence, curl) and three integral theorems (Green, Stokes, Gauss) let one convert between local behaviour and total flow across a region, which is why Maxwell's equations, fluid flow and heat conduction are all written in this language. It is a compulsory unit in B.Sc Mathematics (UGC CBCS "Multivariate Calculus"/"Analysis II") and in every engineering mathematics paper, and carries 8–12 marks in GATE MA/ME/EE/CE and in CSIR-UGC NET Unit 1. Practically it underlies CFD and FEA work at Indian employers (ISRO, DRDO, BARC, L&T, Ansys/Altair India, TCS and Infosys engineering-services units) and the physics behind every electromagnetics or fluid-mechanics interview.

Explained

Scalar and vector fields, the del operator

• Scalar field φ(x, y, z): one number per point — temperature, pressure, electrostatic potential. Vector field F(x, y, z) = F₁i + F₂j + F₃k: one vector per point — velocity, force, E and B.
• Del (nabla) ∇ = i ∂/∂x + j ∂/∂y + k ∂/∂z is a vector differential operator; it acts three ways: ∇φ (gradient, vector), ∇·F (divergence, scalar), ∇×F (curl, vector).
• Level surface φ = c; ∇φ is normal to it. Unit normal n̂ = ∇φ/|∇φ|.
• Directional derivative of φ along unit vector â is ∇φ·â; it is maximum along ∇φ with value |∇φ|, zero along the level surface.
• Vector point function of one variable r(t) = x(t)i + y(t)j + z(t)k describes a curve; dr/dt is tangent, |dr/dt| = speed, and for arc length s, dr/ds is the unit tangent T̂.
• Serret–Frenet formulae (examined in many B.Sc syllabi): dT̂/ds = κN̂, dB̂/ds = −τN̂, dN̂/ds = τB̂ − κT̂, with curvature κ and torsion τ; B̂ = T̂ × N̂.

Gradient, divergence, curl and the Laplacian

• Gradient measures rate and direction of steepest increase of a scalar. It is the only one of the three that turns a scalar into a vector.
• Divergence ∇·F measures net outflow per unit volume at a point — the "source strength". ∇·F > 0: source; < 0: sink; = 0: solenoidal (incompressible, divergence-free).
• Curl ∇×F measures local rotation (twice the angular velocity of a small fluid element). ∇×F = 0: irrotational. A rigid rotation v = ω × r has ∇×v = 2ω.
• Laplacian ∇²φ = ∇·(∇φ); ∇²φ = 0 is Laplace's equation (harmonic φ), ∇²φ = f is Poisson's equation.
• Standard identities (all provable by expanding components):
  • ∇×(∇φ) = 0 — gradient fields are irrotational.
  • ∇·(∇×F) = 0 — curl fields are solenoidal.
  • ∇·(φF) = φ(∇·F) + ∇φ·F; ∇×(φF) = φ(∇×F) + ∇φ × F.
  • ∇·(F×G) = G·(∇×F) − F·(∇×G).
  • ∇×(∇×F) = ∇(∇·F) − ∇²F.
  • ∇(F·G) = (F·∇)G + (G·∇)F + F×(∇×G) + G×(∇×F).
• Classification of fields: irrotational and solenoidal (∇²φ = 0, potential flow); irrotational only (F = ∇φ, φ a scalar potential); solenoidal only (F = ∇×A, A a vector potential); neither (general field, Helmholtz decomposition into both parts).

Line integrals, work and conservative fields

• Line integral ∫_C F·dr = ∫ (F₁dx + F₂dy + F₃dz) = work done by F along C. Parametrise by t, substitute, integrate over the t-range.
• Circulation = ∫_C F·dr around a closed curve. Flux across a plane curve = ∫_C F·n̂ ds.
• Conservative field (three equivalent statements in a simply connected region):
  1. ∇×F = 0 everywhere;
  2. F = ∇φ for some scalar potential φ;
  3. ∫_C F·dr depends only on the end points, so ∮ F·dr = 0.
• Finding the potential: integrate F₁ with respect to x, then fix the unknown function of (y, z) by matching F₂ and F₃. Then ∫_A^B F·dr = φ(B) − φ(A).
• Caveat: simple connectedness matters. F = (−y i + x j)/(x² + y²) has zero curl but ∮ F·dr = 2π around the origin, because the domain has a hole — a favourite GATE/NET trap.

Surface and volume integrals

• Surface integral of a scalar: ∬_S φ dS; flux of a vector: ∬_S F·n̂ dS, with n̂ the outward (or chosen) unit normal.
• Evaluation routes: (a) project onto a coordinate plane — dS = dx dy/|n̂·k|; (b) parametrise r(u, v) — dS = |r_u × r_v| du dv; (c) use symmetry (on a sphere of radius a, n̂ = r/a).
• Volume integral ∭_V φ dV or ∭_V F dV, evaluated as a triple integral; change variables with the Jacobian J = ∂(x, y, z)/∂(u, v, w).
• Standard volume elements: Cartesian dx dy dz; cylindrical ρ dρ dφ dz; spherical r² sin θ dr dθ dφ.
• Orientation must be consistent: for Stokes' theorem the curve is traversed anticlockwise when viewed from the side the normal points to (right-hand rule); for Gauss' theorem the normal is outward.

The three integral theorems

• Green's theorem (plane): ∮_C (M dx + N dy) = ∬_R (∂N/∂x − ∂M/∂y) dA for a positively oriented simple closed curve bounding region R, with M, N having continuous first partials on R.
  • Corollary: area of R = ½∮_C (x dy − y dx).
  • Two vector forms: circulation form (curl, gives Stokes in the plane) and flux form (divergence, gives Gauss in two dimensions).
• Stokes' theorem: ∮_C F·dr = ∬_S (∇×F)·n̂ dS — circulation round the rim equals the flux of curl through any capping surface. Needs S piecewise smooth, two-sided, with C as its boundary.
• Gauss divergence theorem: ∬_S F·n̂ dS = ∭_V (∇·F) dV — outward flux through a closed surface equals the integral of divergence inside. Needs V bounded by a closed, piecewise-smooth surface, F continuously differentiable on V.
• Relationships: all three are forms of the same statement (the generalised Stokes theorem); Green's theorem is the planar case of both Stokes' and the divergence theorem.
• Green's identities (from the divergence theorem with F = φ∇ψ): ∭(φ∇²ψ + ∇φ·∇ψ)dV = ∬φ(∂ψ/∂n)dS, and the symmetric second identity — the basis of uniqueness proofs for Laplace/Poisson problems.
• Use: switch to whichever side is easier. A messy surface integral over a closed surface becomes a short volume integral if ∇·F is simple (often a constant); a hard line integral becomes easy if ∇×F is simple.

Curvilinear coordinates and applications

• Orthogonal curvilinear coordinates (u₁, u₂, u₃) with scale factors h_i = |∂r/∂u_i|: cylindrical (h = 1, ρ, 1), spherical (h = 1, r, r sin θ).
• > ∇φ = Σ (1/h_i)(∂φ/∂u_i) ê_i; ∇·F = (1/h₁h₂h₃) Σ ∂(h_j h_k F_i)/∂u_i
• In spherical coordinates ∇²φ = (1/r²)∂/∂r(r² ∂φ/∂r) + (1/r²sin θ)∂/∂θ(sin θ ∂φ/∂θ) + (1/r²sin²θ)(∂²φ/∂φ²).
• Physics links: Maxwell's equations (∇·D = ρ, ∇·B = 0, ∇×E = −∂B/∂t, ∇×H = J + ∂D/∂t); continuity equation ∂ρ/∂t + ∇·(ρv) = 0; Fourier's law q = −k∇T; Euler/Navier–Stokes momentum equations; Gauss' law of electrostatics as the divergence theorem applied to E.
• Limitations: the theorems fail at singularities of the field (point charges inside V must be excluded or handled by the delta function), on non-orientable surfaces (Möbius strip), and where partial derivatives are not continuous; numerical work (FEM/FVM) is the practical route when the geometry is irregular.

Key points

• Scalar field – one number per point; vector field – one vector per point.
• ∇φ – vector of steepest ascent, normal to level surfaces, magnitude = max rate of change.
• Directional derivative – ∇φ·â; maximum |∇φ|, zero along a level surface.
• Divergence – scalar; net outflow per unit volume. ∇·F = 0 → solenoidal.
• Curl – vector; twice the local angular velocity. ∇×F = 0 → irrotational.
• Laplacian – ∇²φ = ∇·∇φ; ∇²φ = 0 defines a harmonic (potential) function.
• Two zero identities – curl grad = 0 and div curl = 0; they explain why potentials exist.
• Conservative field – irrotational in a simply connected domain ⇔ F = ∇φ ⇔ path-independent work ⇔ zero circulation.
• Scalar potential – found by partial integration; work = φ(end) − φ(start).
• Green's theorem – line integral round a plane curve = double integral of (N_x − M_y) over the enclosed area.
• Stokes' theorem – circulation on the boundary = flux of curl through any capping surface.
• Gauss divergence theorem – outward flux through a closed surface = volume integral of divergence.
• Area by line integral – A = ½∮(x dy − y dx).
• Scale factors – cylindrical (1, ρ, 1), spherical (1, r, r sin θ); needed for ∇ in those systems.
• Jacobian – converts dV between coordinate systems; ρ for cylindrical, r² sin θ for spherical.
• Orientation – outward normal for Gauss; right-hand rule linking C and n̂ for Stokes.
• Simple connectedness – zero curl alone does not guarantee a potential if the domain has a hole.
• Helmholtz decomposition – any well-behaved field = irrotational part + solenoidal part.

Formulas and examples

Example 1 — Directional derivative and unit normal. φ = x²y + y²z at P(1, 2, −1), direction a = 2i − j + 2k.

• ∂φ/∂x = 2xy = 4; ∂φ/∂y = x² + 2yz = 1 + (−4) = −3; ∂φ/∂z = y² = 4. So ∇φ = 4i − 3j + 4k.
• |a| = √(4 + 1 + 4) = 3, â = (2i − j + 2k)/3.
• Directional derivative = (4·2 + (−3)(−1) + 4·2)/3 = (8 + 3 + 8)/3 = 19/3 ≈ 6.33.
• Maximum rate of change = |∇φ| = √(16 + 9 + 16) = √41 ≈ 6.40, along 4i − 3j + 4k; unit normal to the level surface φ = −2 at P is (4i − 3j + 4k)/√41.

Example 2 — Conservative field, potential and work. F = (2xy + z³)i + x²j + 3xz²k.

• ∇×F: i-component ∂(3xz²)/∂y − ∂(x²)/∂z = 0; j-component ∂(2xy + z³)/∂z − ∂(3xz²)/∂x = 3z² − 3z² = 0; k-component ∂(x²)/∂x − ∂(2xy + z³)/∂y = 2x − 2x = 0. Hence irrotational, so conservative.
• φ from ∂φ/∂x = 2xy + z³ → φ = x²y + xz³ + g(y, z). Then ∂φ/∂y = x² ✓ (so g is independent of y) and ∂φ/∂z = 3xz² ✓ → φ = x²y + xz³ + C.
• Work from A(1, −2, 1) to B(3, 1, 4): φ(B) = 9·1 + 3·64 = 9 + 192 = 201; φ(A) = 1·(−2) + 1·1 = −1. W = 201 − (−1) = 202 J (F in N, coordinates in m).

Example 3 — Green's theorem. ∮_C (3x² − 8y²)dx + (4y − 6xy)dy, C the boundary of the region bounded by y = √x and y = x² (anticlockwise).

• M = 3x² − 8y², N = 4y − 6xy → ∂N/∂x = −6y, ∂M/∂y = −16y, difference = −6y + 16y = 10y.
• The curves meet at (0, 0) and (1, 1), with √x ≥ x² on [0, 1].
• ∬ 10y dA = ∫₀¹ ∫_{x²}^{√x} 10y dy dx = ∫₀¹ 5(x − x⁴) dx = 5(1/2 − 1/5) = 5 × 3/10 = 3/2.

Example 4 — Gauss divergence theorem. F = 4xz i − y² j + yz k over the unit cube 0 ≤ x, y, z ≤ 1.

• ∇·F = 4z − 2y + y = 4z − y.
• ∭ (4z − y) dV = ∫₀¹∫₀¹∫₀¹ (4z − y) dx dy dz = ∫₀¹ 4z dz − ∫₀¹ y dy = 2 − 0.5 = 1.5.
• (Direct face-by-face surface integration gives the same 3/2 after six separate integrals — the theorem replaces six with one.)

Example 5 — Stokes' theorem. F = (x² − y²)i + 2xy j, C the boundary of the rectangle 0 ≤ x ≤ a, 0 ≤ y ≤ b in the z = 0 plane, n̂ = k.

• ∇×F = k[∂(2xy)/∂x − ∂(x² − y²)/∂y] = k(2y + 2y) = 4y k.
• ∬ 4y dA = ∫₀^a∫₀^b 4y dy dx = a × 2b² = 2ab².
• Check for a = 3, b = 2: 2 × 3 × 4 = 24; the direct line integral over the four sides also gives 24.

Where it is used

• Aerospace and defence: ISRO/VSSC and DRDO use the divergence and momentum equations in CFD solvers for nozzle, airframe and re-entry heating analysis; roles such as CFD analyst and design engineer are interviewed on these basics.
• Electrical and electronics: Maxwell's equations in antenna and transmission-line design at ISRO, BEL, Tejas Networks and semiconductor design centres; GATE EE/EC electromagnetics questions are largely vector-calculus questions.
• Civil, water resources and energy: seepage and groundwater flow (Laplace equation), heat conduction in structures, hydro and thermal plant flow analysis at L&T, NTPC and the Central Water Commission.
• Simulation software and engineering services: Ansys, Altair, Dassault, Siemens India plus TCS, Infosys, Tech Mahindra and Wipro engineering-services divisions build and apply FEM/FVM solvers that discretise these operators.
• Weather, oceanography and remote sensing: IMD, INCOIS and NRSC use vorticity (curl of wind velocity) and divergence fields in numerical weather prediction.
• Graphics, ML and robotics: gradients drive optimisation (gradient descent) and lighting/normal computation; divergence-free constraints in fluid animation; gradient fields in path planning at Indian product and game studios.

Interview questions

• Q: What is the physical meaning of divergence and curl? A: Divergence is the net outflow per unit volume at a point, so it measures sources and sinks; curl measures the local rotation of the field — numerically twice the angular velocity of a small element placed in it.
• Q: Why is ∇×(∇φ) always zero? A: Each component is a difference of mixed second partials, such as ∂²φ/∂y∂z − ∂²φ/∂z∂y, which cancel when φ has continuous second derivatives.
• Q: How do you test whether a force field is conservative, and what follows if it is? A: Compute the curl: if it is zero throughout a simply connected region the field is conservative, a scalar potential φ exists with F = ∇φ, work is path-independent and equals φ(B) − φ(A), and every closed-loop integral is zero.
• Q: Can a field have zero curl and still a non-zero closed line integral? A: Yes, if the domain is not simply connected — F = (−y i + x j)/(x² + y²) has zero curl but circulation 2π on any loop enclosing the origin, where the field is undefined.
• Q: State the divergence theorem and name one condition it needs. A: The outward flux of F through a closed surface equals the volume integral of ∇·F inside, provided the surface is closed and piecewise smooth and F is continuously differentiable throughout the volume — so no singularities inside.
• Q: How is Green's theorem related to Stokes' and the divergence theorem? A: Green's theorem is the two-dimensional case of both: its circulation form is Stokes' theorem with n̂ = k, and its flux form is the divergence theorem in the plane.
• Q: What does a solenoidal field mean physically, and give an example. A: Zero divergence, so no net creation or destruction of the quantity — the magnetic field B, and the velocity field of an incompressible fluid.
• Q: Why does ∇·B = 0 while ∇·D = ρ? A: There are no isolated magnetic monopoles, so magnetic flux lines close on themselves, whereas electric charge is a genuine source of the electric displacement field.
• Q: When would you convert a surface integral into a volume integral rather than evaluating it directly? A: When the surface is closed and made of several faces or is awkward to parametrise, while ∇·F is simple — typically constant — so the triple integral reduces to a constant times the volume.
• Q: What is the directional derivative and in which direction is it largest? A: It is ∇φ·â, the rate of change of φ along the unit vector â; it is largest along ∇φ with value |∇φ| and zero along any direction tangent to the level surface.
• Q: What is the Laplacian and why does it matter in engineering? A: ∇²φ = ∇·(∇φ); ∇²φ = 0 (Laplace) describes steady-state potential, temperature, seepage and electrostatic problems, and ∇²φ = f (Poisson) the same problems with a source term.
• Q: How do you choose the capping surface in Stokes' theorem? A: Any piecewise-smooth two-sided surface with the given curve as its boundary works; pick the simplest one — usually a flat disc or the plane region — and orient the normal by the right-hand rule relative to the direction of traversal.

Common mistakes

• Treating ∇ as an ordinary vector: ∇·F and F·∇ are different (one is a scalar, the other an operator), and ∇ does not commute with the functions it acts on.
• Writing divergence as a vector or curl as a scalar — losing marks before any computation. In two dimensions the "scalar curl" is only the k-component.
• Using the wrong normal direction: an inward normal in Gauss' theorem flips the sign; a mismatch between curve orientation and n̂ in Stokes does the same.
• Concluding "zero curl ⇒ conservative" without checking that the region is simply connected and the field is differentiable everywhere in it.
• Applying the divergence theorem to an open surface, or to a volume containing a singularity (a point charge) without excluding it.
• Forgetting the Jacobian — using dr dθ dφ instead of r² sin θ dr dθ dφ, or dρ dφ dz instead of ρ dρ dφ dz.
• Mixing up the limits when converting a line integral to a double integral: sketch the region and identify the intersections before integrating.

Related topics

Multiple integrals and Jacobians · Partial differential equations (Laplace, heat, wave) · Electromagnetic field theory · Fluid mechanics and continuity · Differential geometry of curves and surfaces · Complex analysis basics