Calculus Integration Techniques: Trigonometric Substitution

Try to integrate $\int \sqrt{9 - x^2}\, dx$.

Go on, try your usual moves. U-substitution needs a derivative of the inside sitting outside, and there isn't one. Power rule doesn't apply to a square root of a difference. Even the reverse chain rule stalls out immediately.

Nothing works. And that's strange, because $\sqrt{9-x^2}$ isn't some exotic function. It's just the top half of a circle.

That's the clue. Circles, right triangles, and the Pythagorean theorem all speak the same language. If we can translate our square root into that language, the problem should get easier. That's exactly what trigonometric substitution does.

Why does a circle equation break every rule you know?

Here's the thing about $\sqrt{9-x^2}$: it comes from the circle equation $x^2 + y^2 = 9$, solved for $y$. And circles are built out of angles, not out of $x$ and $y$ directly. Every point on a circle of radius 3 can be written as $(3\cos\theta, 3\sin\theta)$ for some angle $\theta$.

So if $x$ is really a stand-in for an angle in disguise, our integral is fighting the wrong variable. Swap it out. Let $x = 3\sin\theta$ instead of treating $x$ as free.

That one substitution is the entire idea behind trig substitution. When you see a square root shaped like a circle, an ellipse-ish expansion, or a hyperbola, replace $x$ with the trig function that matches that shape, and the square root collapses.

The fix: let the Pythagorean identity do the work

Watch what happens when $x = 3\sin\theta$. Substitute directly into the radical:

And here's the payoff: $1 - \sin^2\theta$ is just $\cos^2\theta$. That's the Pythagorean identity, the same one you memorized in trig class, now doing real work for us.

The square root is gone. No radical, just a plain cosine. That's not a coincidence and it's not a party trick. It's true for every value of $\theta$, and you can check it yourself.

Drag $\theta$ anywhere on that semicircle. The point's height, computed two completely different ways, one from the raw square root and one from $3\cos\theta$, never disagrees by more than a rounding error. The identity isn't an approximation. It's exact, always.

Three shapes, three substitutions

$\sqrt{9-x^2}$ is one shape. But radicals come in three flavors, and each one matches a different Pythagorean identity:

• $\sqrt{a^2 - x^2}$ calls for $x = a\sin\theta$, because $1 - \sin^2\theta = \cos^2\theta$
• $\sqrt{a^2 + x^2}$ calls for $x = a\tan\theta$, because $1 + \tan^2\theta = \sec^2\theta$
• $\sqrt{x^2 - a^2}$ calls for $x = a\sec\theta$, because $\sec^2\theta - 1 = \tan^2\theta$

You don't need to memorize these as separate rules. Each one comes from a right triangle with sides $a$, $x$, and the radical. Once you draw that triangle, the substitution falls out on its own.

Notice that all three triangles have the exact same skeleton: an adjacent side, an opposite side, a hypotenuse, and $\theta$ sitting at the same corner. The only thing that changes is which side is $a$, which is $x$, and which is the radical. Match the radical in front of you to one of these three shapes, and you already know which substitution to use.

Does swapping variables actually preserve the area?

This is the part that should bother you a little. We just replaced $x$ with $\theta$, a completely different variable with a completely different range. Is the integral we're computing after the swap still measuring the same area we started with?

Yes. And you can watch it happen instead of taking it on faith.

The left panel is $\int_0^3 \sqrt{9-x^2}\, dx$, computed the ordinary way, rectangle by rectangle, in $x$. The right panel is what that same integral turns into after the substitution: $\int_0^{\pi/2} 9\cos^2\theta\, d\theta$, rectangles in $\theta$. Hit play and watch both sums climb toward $9\pi/4$, the exact area of a quarter circle of radius 3. Different variable, different-looking curve, same number. That's what makes the substitution legal, not just convenient.

Finishing the problem without leaving an angle in your answer

Now let's actually solve $\int \sqrt{9-x^2}\, dx$. With $x = 3\sin\theta$ and $dx = 3\cos\theta\, d\theta$:

That's a standard integral. Using the identity $\cos^2\theta = \frac{1+\cos 2\theta}{2}$:

We've got an antiderivative. But it's written in $\theta$, and the original problem was written in $x$. We need to translate back, and this is exactly where that right triangle earns its keep. From $x = 3\sin\theta$, we know $\sin\theta = x/3$. The other two sides of the triangle give us $\cos\theta$ directly, no extra algebra required.

Drag $x/a$ and watch every ratio update at once: $\sin\theta$, $\cos\theta$, $\tan\theta$, and $\theta$ itself. That's the whole back-substitution step. You're not solving anything new here, just reading numbers off a triangle you already built.

Plugging $\sin\theta = x/3$ and $\cos\theta = \sqrt{9-x^2}/3$ back in:

No angle left anywhere. Just $x$, exactly as the original problem asked.

The other two triangles, briefly

The $\tan$ and $\sec$ cases follow the identical playbook, just with a different triangle. For $\int \frac{dx}{\sqrt{x^2+4}}$, you'd spot the $\sqrt{a^2+x^2}$ shape, set $x = 2\tan\theta$, and the identity $1+\tan^2\theta=\sec^2\theta$ turns the radical into $2\sec\theta$. For $\int \frac{dx}{x\sqrt{x^2-16}}$, you'd spot $\sqrt{x^2-a^2}$, set $x=4\sec\theta$, and $\sec^2\theta - 1 = \tan^2\theta$ does the collapsing.

Same three moves every time: match the shape, substitute, simplify with the identity, integrate, then read the answer back off the triangle.

The short version

A square root of $a^2 - x^2$, $a^2+x^2$, or $x^2-a^2$ resists ordinary integration because it's secretly a circle, a hyperbola, or an ellipse-shaped identity wearing algebra as a disguise. Trig substitution unmasks it: replace $x$ with $a\sin\theta$, $a\tan\theta$, or $a\sec\theta$ to match the shape, and the corresponding Pythagorean identity collapses the radical into a single clean trig term. The substitution doesn't just simplify the symbols, it preserves the actual area, which you can verify numerically by comparing Riemann sums before and after. And when you're done integrating in $\theta$, the same right triangle that motivated the substitution hands you every ratio you need to write the final answer back in terms of $x$.

The whole technique is really just one triangle, built once, used twice: once going in, once coming back out.

All visualizations are interactive React components running entirely in your browser, computed live with no lookup tables. The area comparison uses midpoint Riemann sums with up to 90 rectangles per side. No libraries beyond React.