Guides · 7 min read
u-Substitution: The First Integration Technique to Master
If you learn only one integration technique well, make it u-substitution. It is the most frequently used method by a wide margin, it is the foundation the harder techniques build on, and it is nothing more than the chain rule run in reverse. Getting comfortable with it early makes the rest of integral calculus far less intimidating.
This guide explains what substitution is really doing, how to pick the right u, and the details that trip people up, especially the handling of limits in definite integrals.
Substitution Is the Chain Rule Backward
Recall that the chain rule differentiates a composite function by multiplying the outer derivative by the inner derivative. u-substitution reverses this. When an integrand contains a composite function together with the derivative of its inner part, substitution unwinds the composition and leaves a simple integral you already know how to solve.
Seeing the connection removes the mystery. Every time you successfully substitute, you are recognizing the footprint of a chain-rule derivative and stepping back through it. That is why the technique feels natural once the chain rule is solid, and awkward when it is not.
How to Choose u
The whole method hinges on picking the right inner function to call u. The reliable choice is the inner function of a composition whose derivative also appears in the integrand, up to a constant factor. Look for something raised to a power, tucked under a root, or sitting inside a trig, exponential, or logarithmic function, and check whether its derivative is present elsewhere as a multiplier.
Once you choose u, compute du, the derivative of u times dx, and use it to replace the matching part of the integrand. If the derivative appears only up to a constant, you can adjust with a constant multiple, which is fine. If it appears with a variable mismatch you cannot fix, u was the wrong choice, and you should try a different inner function or a different technique entirely.
Set u to the inner function whose derivative is present in the integrand.
A Worked Pattern
Consider integrating 2x times the cosine of x squared. The inner function is x squared, and its derivative is 2x, which is exactly the other factor. Set u equal to x squared, so du equals 2x dx. The integral becomes simply the integral of cosine of u du, which is sine of u, and converting back gives sine of x squared plus a constant.
Notice how the substitution turned an integral that has no obvious antiderivative into one of the most basic forms. That transformation, from unfamiliar to standard, is the entire value of the technique. When the constant does not match perfectly, say the integrand had only x rather than 2x, you compensate by pulling out a factor of one-half, which is a routine adjustment.
Definite Integrals: Change the Limits or Convert Back
Substitution in a definite integral introduces a fork that causes many avoidable errors. You have two valid options. The first is to change the limits of integration into u-values as you substitute: plug the original x-limits into your expression for u and integrate directly in u with the new bounds, never converting back. This is usually the cleaner path.
The second option is to keep the original limits, find the antiderivative, then convert back to x before evaluating at the original bounds. Both give the same answer. The fatal mistake is mixing them, evaluating a u-antiderivative at the original x-limits without converting either one. Pick one method per problem and commit to it.
When Substitution Does Not Work
Substitution is powerful but not universal, and part of mastering it is recognizing its limits. If no inner function has its derivative present in the integrand, substitution will not simplify anything, and forcing it wastes time. That is your cue to consider integration by parts for products of unlike functions, partial fractions for rational functions, or trig substitution for certain square roots.
The practical workflow is to always attempt substitution first among the real techniques, because it is fast and common, but to abandon it quickly when the derivative match is absent. Building that judgment, knowing when to persist and when to switch, comes from working many integrals and paying attention to which structural clues led to which method.
- No inner derivative present: substitution will not help; try another technique.
- Product of two unlike functions: consider integration by parts.
- Ratio of polynomials: consider partial fractions.
- Square root of a sum or difference of squares: consider trigonometric substitution.
Frequently asked questions
How do I choose what to set as u?
Choose the inner function of a composition whose derivative also appears in the integrand, up to a constant. It is usually the expression inside a power, root, trig, exponential, or logarithmic function.
What do I do about the limits in a definite integral?
Either convert the limits to u-values and integrate entirely in u, or keep the original limits and convert your antiderivative back to x before evaluating. Never evaluate a u-antiderivative at the original x-limits.
Why is u-substitution the same as the chain rule?
The chain rule differentiates a composite function by multiplying by the inner derivative. Substitution reverses that: it recognizes a composite function paired with its inner derivative and unwinds the composition.
What if the derivative is off by a constant?
That is fine. You can factor out or multiply in a constant to make du match. Only a mismatch involving the variable itself, not just a constant, means u was the wrong choice.