Calculus › Section O · Multiple integrals › Topic O.3

Double integrals in polar coordinates

Describe a region by distance and angle and circular boundaries become constants. One extra factor of r is the whole price of the switch, and it unlocks even the famous Gaussian integral.

Circular regions fight rectangular coordinates

Try describing a disk of radius 3 with vertical strips: each strip enters and exits at \(y = \pm\sqrt{9 - x^2}\), so the integral involves square roots that only get worse after the first integration. The region's boundary is screaming its natural coordinates. In polar coordinates, the same disk is \(0 \le r \le 3\), \(0 \le \theta \le 2\pi\): constant limits, a rectangular region in \((r, \theta)\)-space. The rule of thumb for the switch: use polar coordinates when the region is circular (disks, annuli, cardioids) or when the integrand features \(x^2 + y^2\), which becomes the single symbol \(r^2\).

Converting the pieces

Two conversions happen, and forgetting either one breaks the problem. First, the variables:

x = r\cos\theta, \qquad y = r\sin\theta

Substitute these into the integrand, so \(x^2 + y^2\) becomes \(r^2\). Second, the area element: \(dA = r\,dr\,d\theta\). The full template is

\iint_R f(x, y)\,dA = \int_\alpha^\beta \int_{r_1(\theta)}^{r_2(\theta)} f(r\cos\theta, r\sin\theta)\,r\,dr\,d\theta

The limits describe the region: \(\theta\) sweeps from \(\alpha\) to \(\beta\), and along each ray a thin strip runs from \(r_1(\theta)\) to \(r_2(\theta)\). Rays replace vertical strips; entry and exit are now distances from the origin.

Why dA picks up a factor of r

The area element is not \(dr\,d\theta\), and the reason is geometry. A small polar patch between \(r\) and \(r + \Delta r\), between \(\theta\) and \(\theta + \Delta\theta\), is a thin curved wedge. Its long arc has length \(r\Delta\theta\), because arc length is radius times angle, and its short side is \(\Delta r\). The patch's area is about \(r\,\Delta r\,\Delta\theta\): the wedge grows wider in proportion to \(r\), and the factor \(r\) records that. Near the origin the wedges are skinny and contribute almost nothing; far out they are wide and contribute proportionally more. Passing to the limit gives \(dA = r\,dr\,d\theta\).

Worked example: area of a disk

Compute the area of the disk of radius 3. With \(f = 1\):

A = \int_0^{2\pi} \int_0^3 r\,dr\,d\theta = \int_0^{2\pi} \frac{9}{2}\,d\theta = 9\pi

which agrees with \(\pi r^2\) at \(r = 3\). Notice how the two constant limits made the integral a product of two one-variable integrals, one in \(r\) and one in \(\theta\): this is the separable shortcut in action. An annulus such as \(2 \le r \le 5\) changes nothing but the inner limits: \(\int_0^{2\pi} \int_2^5 r\,dr\,d\theta = \pi(25 - 4) = 21\pi\), the difference of the two circles' areas, as it must be.

Worked example: volume under a paraboloid cap

Find the volume under \(z = 4 - x^2 - y^2\) above the disk \(x^2 + y^2 \le 4\). The surface and the region both speak \(r\): the roof is \(z = 4 - r^2\) and the floor is \(0 \le r \le 2\), so

V = \int_0^{2\pi} \int_0^2 (4 - r^2)\,r\,dr\,d\theta = 2\pi \int_0^2 (4r - r^3)\,dr = 2\pi \Big[2r^2 - \frac{r^4}{4}\Big]_0^2 = 8\pi

Check it a second way: the volume under a paraboloid \(z = h - r^2\) style cap is half the circumscribing cylinder, \(\frac{1}{2}\pi R^2 h\). Here \(R = 2\), \(h = 4\), giving \(\frac{1}{2}\pi \cdot 4 \cdot 4 = 8\pi\). The \(r\) from \(dA\) is exactly what made the antiderivative \(2r^2 - r^4/4\) elementary.

Regions traced by rays: the cardioid

For a region bounded by a polar curve \(r = R(\theta)\), the ray at angle \(\theta\) enters at the origin and exits at the curve, so the limits are \(0 \le r \le R(\theta)\) with \(\theta\) running over the shortest interval that traces the boundary once. The area inside \(r = 1 + \cos\theta\), for example, is

A = \int_0^{2\pi} \int_0^{1 + \cos\theta} r\,dr\,d\theta = \frac{1}{2}\int_0^{2\pi} (1 + \cos\theta)^2\,d\theta

Expanding, \(\int_0^{2\pi} (1 + 2\cos\theta + \cos^2\theta)\,d\theta = 2\pi + 0 + \pi = 3\pi\), so the area is \(\frac{3\pi}{2}\). The same one-variable polar area formula \(\frac{1}{2}\int R(\theta)^2\,d\theta\) met earlier in polar graphs falls out of the double integral by doing the \(r\)-integration first.

The Gaussian integral

The most famous integral in statistics has no elementary antiderivative: \(I = \int_{-\infty}^{\infty} e^{-x^2}\,dx\). Squaring it seems to make things worse, until you read the square as a double integral over the whole plane, using the fact that \(e^{-(x^2 + y^2)} = e^{-x^2}e^{-y^2}\):

I^2 = \int_{-\infty}^{\infty} \int_{-\infty}^{\infty} e^{-(x^2 + y^2)}\,dA = \int_0^{2\pi} \int_0^{\infty} e^{-r^2} r\,dr\,d\theta

The polar form is easy: substitute \(u = r^2\), \(du = 2r\,dr\), so the inner integral is \(\frac{1}{2}\int_0^\infty e^{-u}\,du = \frac{1}{2}\). Therefore \(I^2 = 2\pi \cdot \frac{1}{2} = \pi\), and since \(I > 0\), \(I = \sqrt{\pi}\). The bell curve's total area is exactly \(\sqrt{\pi}\), and this polar computation is the classic elementary proof.

Choosing limits without tears

Setting up a polar integral is another drawing exercise, this time with rays. Sketch the region, draw one ray from the origin through it, and read the two distances where the ray enters and leaves: those are \(r_1(\theta)\) and \(r_2(\theta)\). Then find the smallest \(\theta\)-interval that sweeps every ray of the region: a full disk needs \(0 \le \theta \le 2\pi\); a half-disk above the x-axis needs \(0 \le \theta \le \pi\); a quarter-disk needs \(0 \le \theta \le \pi/2\). Keep \(r \ge 0\) in the limits and let \(\theta\) do the sweeping: a point "behind" the origin is reached by adjusting the angle, not by a negative radius.

Common mistakes to avoid

The single most common error is dropping the factor \(r\): the area element is \(r\,dr\,d\theta\), never \(dr\,d\theta\), and its omission silently reports the area of the disk of radius 3 as \(6\pi\) instead of \(9\pi\). Second, substitute into the integrand as well as the limits: \(x^2 + y^2\) left as letters mixes coordinate systems. Third, when converting \(\theta\)-limits, sweep the region exactly once: integrating \(\theta\) from 0 to \(4\pi\) counts a full disk twice. Fourth, order matters: with \(dr\,d\theta\) innermost, the \(r\)-limits come first from the ray reading.

Key takeaways

Switch to polar coordinates when the region or the integrand is circular: substitute \(x = r\cos\theta\), \(y = r\sin\theta\), and replace \(dA\) by \(r\,dr\,d\theta\), where the \(r\) measures how the wedge widens with distance. Limits come from a ray sketch: entry and exit radii give the inner limits, and the shortest \(\theta\)-interval tracing the region gives the outer pair. Constant limits make disks and annulus integrals factor into products, the cardioid reduces to the familiar \(\frac{1}{2}\int R(\theta)^2\,d\theta\), and the Gaussian integral \(\int_{-\infty}^{\infty} e^{-x^2}\,dx = \sqrt{\pi}\) is the method's most celebrated catch.

Flashcards

Flashcard

Polar conversion of a point

x = r cos θ, y = r sin θ

Substitute into the integrand too.

Test your understanding

Question 1 of 5

What is the area of the disk of radius 3, computed as ∫₀^{2π} ∫₀³ r dr dθ?

Question 2 of 5

What is the value of the double integral of (x² + y²) over the disk of radius 2?

Question 3 of 5

Using A = (1/2)∫₀^{2π} r² dθ, what is the area inside the cardioid r = 1 + cos θ?

Question 4 of 5

A student computes the area of the disk of radius 3 as ∫₀^{2π} ∫₀³ 1 dr dθ = 6π. What went wrong?

Question 5 of 5

Why does squaring I = ∫ e^(-x²) dx from -∞ to ∞ turn it into a computable double integral?

Standards: Beyond AP

← Previous topicNext topic →