Partial Fractions as a Decomposition Tool

Break any rational expression into a sum of simple pieces you can integrate, invert, or sum. Learn to read denominators as a menu of building blocks, pick the right coefficient method fast, and verify your work by recombining with minimal algebra.

A rational function can look like one object, yet behave like a stack of simpler ones glued together. Partial fractions works because the denominator’s factors act like separate bottlenecks. Each factor creates its own kind of local behavior near where it vanishes, and the numerator can be re-expressed as a weighted mix of those behaviors. Once you split the function into those building blocks, a lot of downstream math stops feeling like magic because each block has a standard playbook.

To make that idea concrete, interact with a single decomposition and watch how the pieces line up with the denominator’s factors and their poles.

When you see it this way, partial fractions is less a trick and more a change of basis. You trade one complicated fraction for a sum of fractions that each have one job.

Make the fraction decomposable first

Partial fractions is picky about the input. Two cleanups decide the shape of the answer.

Cancellation versus division

Cancellation removes common factors that truly appear in both numerator and denominator. That changes the function itself by removing a removable discontinuity, so the decomposed form no longer needs terms for that canceled factor.

Long division is different. If the fraction is improper, meaning $\deg(\text{numerator})\ge\deg(\text{denominator})$, you rewrite it as a polynomial plus a proper fraction. The decomposition only applies to the proper remainder, but the polynomial part is not optional. It is part of the original function.

Use the pipeline below to see how canceling and long division change the final template you should expect.

Rule of thumb Cancel first if you truly have a common factor, then divide if the remaining fraction is still improper.

A quick self-check is to ask what would happen at large $x$. If your original rational function tends to a nonzero constant or grows like a line, your final answer must include a polynomial term from division.

Choose the right template from the denominator

If you can factor the denominator into real linear factors and irreducible quadratics, the partial fraction form is basically forced. The only freedom is the coefficients.

Here is the featured-snippet version you can rely on. Factor the denominator completely. For each distinct linear factor $(x-r)$, include a term $\frac{A}{x-r}$. For a repeated linear factor $(x-r)^k$, include a ladder $\frac{A_1}{x-r}+\frac{A_2}{(x-r)^2}+\cdots+\frac{A_k}{(x-r)^k}$. For each irreducible quadratic $x^2+px+q$, include $\frac{Ax+B}{x^2+px+q}$, and repeat that block in a ladder if the quadratic power repeats.

The comparison templates make the pattern easier to scan when you are setting up a new problem.

Two small intuitions keep you from memorizing blindly.

• Linear factors produce constants upstairs because there is no room for anything else.
• Irreducible quadratics require a full linear numerator $Ax+B$ because you need enough flexibility to match any remainder of degree less than 2.

Degree check In every block, the numerator degree must be strictly less than the denominator degree of that block.

If you accidentally put just $\frac{A}{x^2+px+q}$, you have already restricted the space of possible numerators, and many decompositions will become impossible.

Solve for coefficients without suffering

Once the template is set, you are solving for numbers. There are three common approaches, and each wins in a different situation.

Cover-up and root plugging

If you have distinct linear factors and no repeated powers, the cover-up idea is fast. Multiply both sides by the factor you want, then plug in the root that kills every other term. You get one coefficient per root with almost no algebra.

Repeated factors break this simplicity because plugging in the root wipes out multiple terms at once. You need extra equations, often from derivatives or from coefficient matching.

Coefficient matching

Multiply through by the full denominator, expand, then match coefficients of like powers of $x$. This is systematic and works in every case, including quadratics and repeats, but it can get algebra-heavy if you expand mindlessly.

A hybrid move is to plug in a few convenient $x$ values to get quick equations, then finish with coefficient matching for whatever remains.

Pick a method and watch how it generates the needed equations, including when it stops being enough by itself.

When the algebra starts to sprawl, step back and count unknowns. If you have five coefficients to find, you need five independent equations. Plugging roots might give two quickly, but it cannot conjure the other three.

Traps that waste the most time

Most errors are not arithmetic. They are missing terms that your denominator demands, or using a shortcut outside its conditions.

Common traps to catch early.

• Forgetting the full ladder for a repeated factor.
• Using $\frac{A}{\text{quadratic}}$ instead of $\frac{Ax+B}{\text{quadratic}}$.
• Canceling factors that are not truly common factors of the original numerator and denominator.
• Skipping long division when the fraction is improper.
• Solving for coefficients correctly, then copying them back into the wrong template.

A toggleable checklist helps you spot these quickly, and it includes a recombination check you can do in under a minute.

Recombining is the sanity check that pays rent. Multiply your proposed decomposition by the original denominator and confirm the numerators match. If you do not want to expand fully, plug in a few random $x$ values that avoid poles. Disagreement at one point means the whole decomposition is wrong.

Fast verify Evaluate both forms at two or three safe $x$ values. It catches template mistakes instantly and catches most arithmetic mistakes.

Why the pieces are so useful later

Partial fractions matters because each building block has a standard downstream behavior. $\frac{A}{x-r}$ turns into a log under integration and an exponential under inverse Laplace. A repeated factor like $\frac{A}{(x-r)^k}$ turns into a power term in integration and a polynomial times an exponential in Laplace. An irreducible quadratic block turns into arctan or sine and cosine patterns, depending on the context.

That is the mental model. You decompose once, then you apply a library of single-block rules.

Use the map below to connect each block type to the most common next steps.

Next-step practice strategy that builds intuition

Practice goes best when you separate pattern fluency from algebra.

Start with template drills. Take denominators and write the decomposition form without solving anything. If you can do that quickly, most of the hard errors disappear.

Then practice method choice.

• Use cover-up or root plugging when factors are distinct linear.
• Use a hybrid when there is a mix of linear and quadratic factors.
• Use coefficient matching when repeats appear or when plugging values runs out of equations.

Build a verification habit that is faster than redoing the whole problem. Plugging in a couple safe $x$ values is often enough to catch a missed ladder term or a wrong numerator shape, and recombining fully is the gold standard when you need certainty.