AP-CALCBC-6.9

U6.9 Integration by Substitution

Master u-substitution for AP Calculus BC: identify the inner function, substitute, integrate, convert back, and change limits for definite integrals.

What you'll do in this lesson

A voice-first session with the Crimsora tutor on U6.9 Integration by Substitution, then targeted practice and FRQs — with the tutor adapting to where you get stuck.

What this lesson covers

Integration by substitution reverses the chain rule. Whenever you see a composite function multiplied by (a constant times) the derivative of its inner function, u-substitution unlocks the integral. It is the single most-used integration technique on the AP exam, and mastering it now sets you up for integration by parts and partial fractions later.

In this lesson you will learn how to spot a good substitution, carry it out cleanly, and handle the special step definite integrals require: changing the limits of integration. We will also cover the most common mistakes—forgetting dudu, mismatched constants, and reverting incorrectly—so you can avoid them under time pressure.

Why Substitution Works

The chain rule says ddx[F(g(x))]=F(g(x))g(x)\frac{d}{dx}\big[F(g(x))\big] = F'(g(x))\cdot g'(x). Integration by substitution simply runs this backward. If an integrand looks like f(g(x))g(x)f(g(x))\cdot g'(x), then letting u=g(x)u = g(x) turns the whole expression into f(u)duf(u)\,du, which is usually a basic antiderivative.

The key move is the differential. If u=g(x)u = g(x), then dudx=g(x)\frac{du}{dx} = g'(x), so du=g(x)dxdu = g'(x)\,dx. You are literally trading the xx-world for the uu-world. Every xx (including the dxdx) must disappear and be replaced by uu and dudu before you integrate.

A common misconception is thinking you need the derivative of the inner function to appear exactly. In fact you only need it up to a constant multiple, because constants can be factored in and out of integrals. For example, if du=2xdxdu = 2x\,dx but the integrand only has xdxx\,dx, you write xdx=12dux\,dx = \tfrac{1}{2}\,du and continue. You cannot, however, fix a mismatch involving a variable factor—if the leftover cannot be written entirely in terms of uu, substitution alone will not finish the job.

The Step-by-Step Procedure

Follow the same reliable routine every time.
StepAction
1Choose uu = the inner function (often inside a power, root, exponent, or denominator)
2Compute du=g(x)dxdu = g'(x)\,dx
3Solve for dxdx or for the grouped factor, and substitute everything
4Integrate in terms of uu
5Replace uu with g(x)g(x) to return to xx (indefinite integrals)
6Add +C+C
How do you pick uu? Look for a function whose derivative also appears (up to a constant) in the integrand. Good candidates: the expression under a radical, the exponent of ee, the argument of a trig function, or a denominator.

Consider 2xcos(x2)dx\int 2x\cos(x^2)\,dx. Let u=x2u = x^2, so du=2xdxdu = 2x\,dx. The integral becomes cos(u)du=sin(u)+C=sin(x2)+C\int \cos(u)\,du = \sin(u) + C = \sin(x^2) + C. Notice how the 2xdx2x\,dx was perfectly consumed by dudu. The AP exam loves integrands engineered exactly this way, so training your eye to match derivatives is the whole skill.

Definite Integrals: Changing the Limits

For a definite integral you have two valid options, and choosing well saves time.

Option A—change the limits. When you substitute u=g(x)u = g(x), convert the xx-limits into uu-limits using u=g(x)u = g(x). Then you never convert back; you evaluate entirely in uu. This is the AP-preferred, error-resistant method.

Option B—keep the original limits, but then you MUST convert uu back to xx before plugging in the xx-bounds. A fatal error is leaving xx-limits on a uu-antiderivative.

Example with changed limits: 02xex2dx\int_{0}^{2} x\,e^{x^2}\,dx. Let u=x2u = x^2, du=2xdxdu = 2x\,dx, so xdx=12dux\,dx = \tfrac{1}{2}\,du. When x=0x=0, u=0u=0; when x=2x=2, u=4u=4. The integral becomes 1204eudu=12[eu]04=12(e41)\tfrac{1}{2}\int_{0}^{4} e^{u}\,du = \tfrac{1}{2}\big[e^{u}\big]_0^4 = \tfrac{1}{2}(e^4 - 1).
Change limitsKeep limits
Convert back to xx?NoYes
Bounds useduu-valuesoriginal xx-values
RiskForgetting to convert boundsForgetting to convert antiderivative
On free-response questions, always label your new limits clearly. Graders want to see that the bounds correspond to whichever variable your antiderivative is written in.

Common Traps and Exam Tips

The most frequent error is dropping the differential. Writing u=3x+1u = 3x+1 then integrating (3x+1)5dx\int (3x+1)^5\,dx as if it were u5\int u^5 ignores that du=3dxdu = 3\,dx, so dx=13dudx = \tfrac{1}{3}\,du. The correct answer carries a factor of 13\tfrac{1}{3}.

A second trap is a leftover variable. In xx+1dx\int x\sqrt{x+1}\,dx, letting u=x+1u = x+1 gives du=dxdu = dx, but the stray xx must be rewritten as x=u1x = u - 1. Then (u1)udu\int (u-1)\sqrt{u}\,du expands into integrable powers. Do not abandon substitution just because the derivative is not sitting there—sometimes solving for xx rescues the problem.

Third, watch constant handling. You can pull constants out (5fdu=5fdu\int 5f\,du = 5\int f\,du) but never pull out variables.

For multiple choice, working backward by differentiating each answer choice with the chain rule is a fast check. For calculator-active questions, you can substitute and then numerically integrate. On no-calculator sections, expect substitutions that produce clean logs, exponentials, or standard trig antiderivatives. Recognizing g(x)g(x)dx=lng(x)+C\int \frac{g'(x)}{g(x)}\,dx = \ln|g(x)| + C as a substitution pattern is especially high-yield.

Recognizing Substitution Patterns Quickly

Speed on the AP exam comes from pattern recognition. Train yourself to associate integrand shapes with the substitution that clears them.
Integrand shapeChoose uuResult type
[g(x)]ng(x)[g(x)]^n\,g'(x)u=g(x)u = g(x)power rule un+1n+1\frac{u^{n+1}}{n+1}
eg(x)g(x)e^{g(x)}\,g'(x)u=g(x)u = g(x)eue^{u}
g(x)g(x)\frac{g'(x)}{g(x)}u=g(x)u = g(x)lnu\ln|u|
cos(g(x))g(x)\cos(g(x))\,g'(x)u=g(x)u = g(x)sin(u)\sin(u)
g(x)1+[g(x)]2\frac{g'(x)}{1+[g(x)]^2}u=g(x)u = g(x)arctan(u)\arctan(u)
The unifying idea: find a chunk whose derivative is also present. Radicals suggest letting uu equal what is inside. Fractions often reward setting uu equal to the denominator and checking whether the numerator is (a multiple of) its derivative.

If a first choice of uu does not clear all the xx's, do not panic—either solve for xx in terms of uu to eliminate a leftover, or reconsider your choice. When substitution genuinely fails (no derivative match and no algebraic fix), that signals a different technique from later lessons, such as integration by parts or partial fractions. Building this decision instinct is exactly what U6.14 will formalize.

Key terms

u-substitution.
An integration technique that reverses the chain rule by replacing an inner function g(x)g(x) with a new variable uu, turning a composite integrand into a simpler one.
Inner function.
The function g(x)g(x) nested inside another function; it is the natural choice for uu because its derivative typically appears in the integrand.
Differential.
The expression du=g(x)dxdu = g'(x)\,dx, which converts the dxdx portion of the integral into dudu during substitution.
Composite function.
A function of the form f(g(x))f(g(x)), where one function is applied to the output of another; substitution targets these.
Changing the limits.
For definite integrals, converting the original xx-bounds into uu-bounds via u=g(x)u = g(x) so the integral can be evaluated entirely in uu.
Integration constant.
The +C+C added to every indefinite integral to account for the family of antiderivatives; omitted only after applying definite bounds.

Worked example

Evaluate 1elnxxdx\int_{1}^{e} \frac{\ln x}{x}\,dx.
Identify the inner function. The integrand contains lnx\ln x and its derivative 1x\frac{1}{x}, which is a classic substitution signal. Let u=lnxu = \ln x.

Compute the differential: du=1xdxdu = \frac{1}{x}\,dx. Notice the 1xdx\frac{1}{x}\,dx already sitting in the integral becomes exactly dudu, and lnx\ln x becomes uu.

Change the limits (preferred method). When x=1x = 1, u=ln1=0u = \ln 1 = 0. When x=ex = e, u=lne=1u = \ln e = 1. So the new bounds are from u=0u=0 to u=1u=1.

Rewrite the integral entirely in uu: 01udu\int_{0}^{1} u\,du.

Integrate: udu=u22\int u\,du = \frac{u^2}{2}. Because we changed the limits, we evaluate in uu directly and do not convert back.

Apply the bounds: [u22]01=122022=12\left[\frac{u^2}{2}\right]_0^1 = \frac{1^2}{2} - \frac{0^2}{2} = \frac{1}{2}.

The value of the definite integral is 12\frac{1}{2}. As a check, note the antiderivative in xx is (lnx)22\frac{(\ln x)^2}{2}, and evaluating from 1 to ee gives 120=12\frac{1}{2} - 0 = \frac{1}{2}—consistent.

Practice questions

Evaluate 6x2x3+4dx\int 6x^2\sqrt{x^3+4}\,dx.
  1. 43(x3+4)3/2+C\frac{4}{3}(x^3+4)^{3/2} + C
  2. 23(x3+4)3/2+C\frac{2}{3}(x^3+4)^{3/2} + C
  3. (x3+4)3/2+C(x^3+4)^{3/2} + C
  4. 92(x3+4)3/2+C\frac{9}{2}(x^3+4)^{3/2} + C

Answer: 43(x3+4)3/2+C\frac{4}{3}(x^3+4)^{3/2} + C

Let u=x3+4u = x^3 + 4, so du=3x2dxdu = 3x^2\,dx, giving 6x2dx=2du6x^2\,dx = 2\,du. The integral becomes 2udu=223u3/2=43u3/2\int 2\sqrt{u}\,du = 2\cdot \frac{2}{3}u^{3/2} = \frac{4}{3}u^{3/2}. Substituting back yields 43(x3+4)3/2+C\frac{4}{3}(x^3+4)^{3/2} + C. The trap answers come from mishandling the constant factor from dudu.
Evaluate xx3dx\int \frac{x}{\sqrt{x-3}}\,dx using substitution, showing how to handle the leftover variable.

Answer: 23(x3)3/2+6(x3)1/2+C\frac{2}{3}(x-3)^{3/2} + 6(x-3)^{1/2} + C

Let u=x3u = x - 3, so du=dxdu = dx and x=u+3x = u + 3. The integral becomes u+3udu=(u1/2+3u1/2)du=23u3/2+6u1/2+C\int \frac{u+3}{\sqrt{u}}\,du = \int \left(u^{1/2} + 3u^{-1/2}\right)du = \frac{2}{3}u^{3/2} + 6u^{1/2} + C. Converting back gives 23(x3)3/2+6(x3)1/2+C\frac{2}{3}(x-3)^{3/2} + 6(x-3)^{1/2} + C. This problem tests whether you can rescue a substitution by solving for xx when a stray variable remains.
Evaluate 0π/2cosxesinxdx\int_{0}^{\pi/2} \cos x\, e^{\sin x}\,dx.

Answer: e1e - 1

Let u=sinxu = \sin x, so du=cosxdxdu = \cos x\,dx. Change limits: when x=0x=0, u=0u=0; when x=π/2x=\pi/2, u=1u=1. The integral becomes 01eudu=[eu]01=e1e0=e1\int_0^1 e^u\,du = [e^u]_0^1 = e^1 - e^0 = e - 1. Changing the limits avoids converting the antiderivative back to xx.

FAQ

How do I know what to pick for u?
Choose the inner function whose derivative (up to a constant multiple) also appears in the integrand. Common picks are the expression inside a power or root, the exponent of ee, the argument of a trig function, or a denominator. If your choice makes every remaining xx disappear after substitution, you picked well.
Do I have to change the limits for definite integrals?
No, but you must be consistent. If you change the limits to uu-values, evaluate the antiderivative in uu and never convert back. If you keep the original xx-limits, you must first convert the antiderivative back to xx before plugging in. Changing the limits is usually faster and less error-prone.
What if the derivative of my inner function isn't in the integrand?
First check whether it is off only by a constant—constants can be adjusted freely. If a variable factor is left over, try solving your substitution for xx and rewriting that factor in terms of uu, as in xx+1dx\int x\sqrt{x+1}\,dx. If nothing works, the integral likely needs a different technique such as integration by parts.
Why did I lose points even though my answer looked right?
The most common causes are forgetting the differential factor (e.g., missing a 13\tfrac{1}{3} from du=3dxdu = 3\,dx), omitting +C+C on an indefinite integral, or leaving xx-limits on an antiderivative written in uu. Always double-check that every xx and dxdx was replaced and that your bounds match your variable.

Learn this with a teacher, not a page

The Crimsora tutor teaches U6.9 Integration by Substitution live — explaining on a whiteboard, asking you questions, and adapting to where you get stuck.