JOVANA
Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
All guides

Trigonometric Substitution & Partial Fractions

Two reliable tactics for integrals that shrug off u-substitution and parts: trade an awkward square root for a clean trig identity, and split any rational function into pieces you already know how to integrate.

When the easy methods run out

You already own two great tools. u-substitution reverses the chain rule — it works when the integrand hides an inner function whose derivative is also sitting there. Integration by parts reverses the product rule — it shines when you can trade a hard integral for an easier one. But some integrands answer to neither. Two of the most common stumbling blocks are a stubborn square root like sqrt(a^2 - x^2), and a ratio of polynomials like (3x + 5) / (x^2 - x - 2). This guide hands you the standard tactic for each.

The common thread is the same one you have always relied on: change the problem into a shape you recognize. A plain u-substitution swaps one variable for another and hopes the new integral is friendlier. Here we are bolder — we substitute a whole trig function, or we tear a fraction apart — but the spirit is identical. Transform, integrate, then translate back. Keep that three-beat rhythm in mind and both methods feel less like magic and more like bookkeeping.

Trig substitution: three roots, three keys

Trigonometric substitution is built on three Pythagorean identities, each one a key cut to fit a particular root. Because 1 - sin^2(theta) = cos^2(theta), setting x = a sin(theta) turns sqrt(a^2 - x^2) into a cos(theta) — the root vanishes into a clean cosine. Because 1 + tan^2(theta) = sec^2(theta), setting x = a tan(theta) turns sqrt(a^2 + x^2) into a sec(theta). And because sec^2(theta) - 1 = tan^2(theta), setting x = a sec(theta) turns sqrt(x^2 - a^2) into a tan(theta). One look at the root tells you which substitution to grab.

sqrt(a^2 - x^2)   set x = a*sin(theta)   ->  root = a*cos(theta)
sqrt(a^2 + x^2)   set x = a*tan(theta)   ->  root = a*sec(theta)
sqrt(x^2 - a^2)   set x = a*sec(theta)   ->  root = a*tan(theta)
The three standard trig substitutions and the clean root each one produces.

There is one more piece of bookkeeping: when you replace x you must also replace dx by its differential. For x = a sin(theta) that means dx = a cos(theta) d(theta). Substitute both, and the entire integral becomes a problem in theta — usually a manageable mix of sines, cosines, or secants, which you then attack with the trig-power techniques in the next guide on powers of trig functions.

Walking through one — and the sign you must watch

Picture the integral of 1 / sqrt(4 - x^2) dx. The root sqrt(4 - x^2) matches the first shape with a = 2, so set x = 2 sin(theta), giving dx = 2 cos(theta) d(theta) and root = 2 cos(theta). The integral collapses to (2 cos(theta) d(theta)) / (2 cos(theta)), which is simply the integral of d(theta) — equal to theta. Now translate back. Since x = 2 sin(theta), we have sin(theta) = x/2, so theta = arcsin(x/2). The answer is arcsin(x/2) + C, an inverse function appearing right where the geometry promised it would.

To translate a messier answer back to x, draw the right triangle the substitution encodes. From sin(theta) = x/2, label the opposite side x and the hypotenuse 2; the Pythagorean theorem fills in the adjacent side as sqrt(4 - x^2). Then any trig function of theta you need — cos(theta), tan(theta), sec(theta) — you simply read off the triangle. This little picture is the bridge that carries every theta-answer back into the original x.

Partial fractions: splitting a rational function

Now the second wall: a ratio of polynomials. Partial fraction decomposition is the exact reverse of adding fractions over a common denominator. You take one tangled fraction and split it back into a sum of simpler ones, each with a basic factor of the original denominator on the bottom — and each of those small pieces is something you already know how to integrate. This is the engine behind integrating rational functions in general.

  1. Check the degrees. If the numerator's degree is not lower than the denominator's, do polynomial long division first and set the polynomial quotient aside (you integrate it trivially).
  2. Factor the denominator completely into linear factors (x - a) and irreducible quadratic factors (x^2 + bx + c).
  3. Write one term per factor: a linear (x - a) gives A/(x - a); a repeated (x - a)^2 also needs B/(x - a)^2; an irreducible quadratic gets a linear top, (Bx + C)/(x^2 + bx + c).
  4. Solve for the unknown constants — clear denominators and either match coefficients or plug in convenient x-values — then integrate each piece into a logarithm, an arctangent, or a power.

Take (3x + 5) / (x^2 - x - 2). The bottom factors as (x - 2)(x + 1), so we write it as A/(x - 2) + B/(x + 1). Clearing denominators gives 3x + 5 = A(x + 1) + B(x - 2). Plug in x = 2: 11 = 3A, so A = 11/3. Plug in x = -1: 2 = -3B, so B = -2/3. The integral is now just (11/3) ln|x - 2| - (2/3) ln|x + 1| + C — two ordinary logs, no cleverness required.

Choosing your weapon — and an honest limit

The decision is mostly visual. A square root of a quadratic shape (a^2 ± x^2 or x^2 - a^2) says trig substitution. A ratio of polynomials with no root says partial fractions. The two even cooperate: when a quadratic in a root is not centered, complete the square first — sqrt(x^2 + 4x + 13) becomes sqrt((x + 2)^2 + 9), and a shift u = x + 2 hands you the clean a^2 + u^2 shape ready for a tangent substitution.

There is a remarkable promise hiding here. Because every partial-fraction piece integrates to a log, an arctangent, or a power, every rational function has an elementary antiderivative — there are no exceptions and nothing exotic is ever needed. Partial fractions also resurface later: the inverse Laplace transform leans on exactly this split, breaking a transfer function into simple terms you read off one by one. And for some integrands of trig fractions, the Weierstrass substitution t = tan(theta/2) turns the whole thing into a rational function of t — handing it straight to partial fractions.