Multivariable Calculus · Double integral in polar coordinates

Double integral in polar coordinates

A double integral in polar coordinates uses the area element r dr dθ: cells of the same Δr and Δθ grow as they move away from the origin. Drag P to compare its cell with the same cell in the (θ, r) plot, and read the radial limits off the ray.

Double integrals in polar coordinates in this visualization

A double integral in polar coordinates is the same ∬Rf dA\iint_R f\,dA, with the region and the function described by the distance r from the origin and the angle θ instead of x and y. Putting x=rcos⁡θx = r\cos\theta and y=rsin⁡θy = r\sin\theta gives

The area element in polar coordinates is r dr dθ
∬Rf dA=∫αβ ⁣∫r1(θ)r2(θ)f(rcos⁡θ,rsin⁡θ) r dr dθ\iint_R f\,dA = \int_\alpha^\beta\!\int_{r_1(\theta)}^{r_2(\theta)} f(r\cos\theta, r\sin\theta)\,r\,dr\,d\theta

It pays off when R is bounded by circles about the origin or by rays from it, or when f depends on x2+y2x^2 + y^2.

The extra r comes from the cells. The polar grid cuts the plane along circles and rays, and a cell between radii r1r_1 and r2r_2 with angle Δθ has area (r22−r12) Δθ/2=rˉ Δr Δθ(r_2^2 - r_1^2)\,\Delta\theta/2 = \bar r\,\Delta r\,\Delta\theta, where rˉ\bar r is its middle radius. This is exact, not an approximation. So equal steps Δr and Δθ make cells that grow with their distance from the origin: in the (θ, r) plot under the view every cell is the same Δθ × Δr rectangle, and its partner in the xy-plane is rˉ\bar r times as large. Adding f rˉ Δr Δθf\,\bar r\,\Delta r\,\Delta\theta over the cells and refining the grid gives dA=r dr dθdA = r\,dr\,d\theta.

The inner limits are read along a ray from the origin: the r where it enters R and the r where it leaves. On the ring 1≤r≤21 \le r \le 2 they are 1 and 2 at every angle, so the ring is a rectangle in the (θ, r) plot; the same ring in x and y needs a vertical slice through the hole split in two. On the cardioid the upper limit 1+cos⁡θ1 + \cos\theta changes with θ. The outer limits are the angles of the rays that cross R. The disc (x−1)2+y2≤1(x - 1)^2 + y^2 \le 1 passes through the origin, and only rays with −π/2<θ<π/2-\pi/2 < \theta < \pi/2 cross it, each leaving at r=2cos⁡θr = 2\cos\theta; the rays at ±π/2\pm\pi/2 only touch the origin.

The factor r can make an integral possible. e−r2e^{-r^2} has no elementary antiderivative, but e−r2 re^{-r^2}\,r does, so over the disc of radius a

With the factor r the inner integral can be done
∬e−x2−y2 dA=∫02π ⁣∫0ae−r2 r dr dθ=π (1−e−a2)\iint e^{-x^2-y^2}\,dA = \int_0^{2\pi}\!\int_0^a e^{-r^2}\,r\,dr\,d\theta = \pi\,(1 - e^{-a^2})

As a grows the integral over the whole plane is π. That integral is I2I^2 with I=∫−∞∞e−x2 dx>0I = \int_{-\infty}^{\infty} e^{-x^2}\,dx > 0, so I=πI = \sqrt{\pi}, the Gaussian integral.

The r is the area factor of the map from (θ, r) to (x, y): the absolute value of its Jacobian determinant (in the plot’s order, θ then r, the determinant itself is −r). A general change of variables multiplies by the absolute value of the Jacobian determinant in the same way.

Things to notice

  • The innermost cells are sectors, with r1=0r_1 = 0. The area rˉ Δr Δθ\bar r\,\Delta r\,\Delta\theta still holds, with rˉ=Δr/2\bar r = \Delta r/2.
  • Polar coordinates are not always simpler. On a square centered at the origin the upper radial limit depends on which side the ray meets, so the integral splits into pieces.
  • A point has many pairs (r, θ). The angles cover one turn, or only the rays that cross R; going further would count part of R twice.

Controls

  • Drag P in the xy-plane or in the (θ, r) plot under it. With P selected, the up and down arrow keys change r, left and right change θ.

Related