Ever tried summing up an infinite list of numbers that just keeps going? Now imagine each of those numbers is actually a function of x. That's a power series — and once you start doing calculus on it, things get interesting fast.
Most calculus students meet power series as a weird side quest. But the integration and differentiation of power series is where the real payoff shows up. You can bend these infinite polynomials into shapes that solve differential equations, model physics, and even sneak up on values of functions you thought were out of reach.
Here's the thing — once you see how tame these operations actually are, a lot of the fear drops away.
What Is a Power Series
A power series is basically a polynomial that never ends. Instead of stopping at x³ or x⁴, it keeps going with bigger and bigger powers of x, each multiplied by some coefficient.
The usual form looks like this:
∑ cₙ(x − a)ⁿ — starting from n = 0 and going forever.
That a is the center. That's why when a is zero, you've got what people call a Maclaurin series*. When it's somewhere else, it's a Taylor series* centered at a.
Why It Behaves Like a Polynomial
The short version is: inside its interval of convergence, a power series acts a lot like a really long polynomial. You can add them, multiply by constants, and — crucially — take derivatives and integrals term by term.
That's not automatically true for every infinite sum. But power series are special. The convergence is "nice enough" in the open interval where they converge that calculus slides right in.
The Interval of Convergence
Every power series has a radius of convergence, R. Inside (−R, R) around the center, it converges absolutely. Which means at the endpoints, who knows — could converge, could blow up. That interval matters because term-by-term calculus is only guaranteed to work inside* it.
Why It Matters
So why should anyone care about integrating or differentiating these things?
Because closed-form functions are lazy. They don't always want to show up. A lot of differential equations have solutions that aren't neat combinations of sin, cos, eˣ, and the like. But they do have power series solutions.
And here's what most people miss: if you can differentiate or integrate a power series, you can build new series from old ones. Need the integral of e^(x²)? Think about it: good luck with ordinary methods. The power series route just works.
In practice, this is how calculators approximate messy functions. Plus, they're not magic. They're summing truncated series — and those series come from differentiation and integration of known ones.
Turns out, understanding this also explains why some approximations are shockingly good and others fall apart the moment you step outside the radius.
How It Works
Alright, the meaty part. Let's say you've got a power series:
f(x) = ∑ cₙ(x − a)ⁿ
defined and convergent for |x − a| < R.
Differentiating Term by Term
You take the derivative of each term like it's a regular polynomial term. The derivative of cₙ(x − a)ⁿ is n·cₙ(x − a)ⁿ⁻¹.
So:
f′(x) = ∑ n·cₙ(x − a)ⁿ⁻¹
And the radius of convergence stays the same. Same R. The endpoints might change behavior, but inside the interval, you're golden.
Why does this matter? Because now you can differentiate sin(x)'s series, or ln(1+x)'s series, and get the series for cos(x) or 1/(1+x) without memorizing anything new.
Integrating Term by Term
Integration is just as friendly. You integrate each term:
∫ f(x) dx = C + ∑ cₙ (x − a)ⁿ⁺¹ / (n + 1)
Again, same radius R. You pick up a constant of integration, like always. And if you're doing a definite integral from a to x, the constant sorts itself out.
Real talk — this is how you get the series for arctan(x). Start with 1/(1+x²) = ∑ (−1)ⁿx²ⁿ, integrate term by term, and boom: arctan(x) = ∑ (−1)ⁿ x²ⁿ⁺¹/(2n+1).
Shifting the Index
One annoying-but-essential skill: reindexing. In practice, after you differentiate, your sum starts at n=1 because the n=0 term died. After you integrate, you get an n+1 in the denominator and a power bumped up.
If you found this helpful, you might also enjoy how does phosphorus get into animals or what are the differences between primary succession and secondary succession.
To compare series or solve equations, you'll shift indices so everything lines up. That said, it's bookkeeping, not deep magic. But honestly, this is the part most guides get wrong by skipping it.
Example: From Geometric Series to Logarithm
Start with the geometric series:
1/(1−x) = ∑ xⁿ for |x| < 1
Integrate both sides from 0 to x:
−ln(1−x) = ∑ xⁿ⁺¹/(n+1)
So ln(1−x) = −∑ xⁿ⁺¹/(n+1). Which means differentiate it back and you're at square one. That back-and-forth is the whole game.
Common Mistakes
Most people get a few things wrong here, and it's rarely the calculus itself.
First: forgetting the radius. No. It works where the series converges. Here's the thing — they differentiate or integrate and assume the new series works everywhere the original function exists. The function might be smooth on all of ℝ; the series might only cover (−1, 1).
Second: ignoring endpoints. The radius is the same, but convergence at x = a ± R can change. A series might converge at an endpoint before differentiating and diverge after, because the terms don't shrink as fast.
Third: messy indices. They leave sums starting at n=1 when they need n=0, then wonder why the differential equation doesn't balance. Slow down and reindex.
And a quiet one — assuming term-by-term works outside power series. It doesn't generalize to arbitrary infinite sums. Now, uniform convergence is required in the general case. Power series get a free pass inside their radius. That's the gift.
Practical Tips
Here's what actually works when you're doing this stuff by hand or teaching it.
Start with a series you know. Practically speaking, geometric, eˣ, sin, cos. Don't build from scratch if you can manipulate a known one.
Always write the radius next to the series. Every. And single. Consider this: time. It keeps you honest.
When solving differential equations with series, align all your sums to the same power of x before equating coefficients. Use substitution like k = n−1 to shift. It feels tedious. It prevents errors.
For definite integrals, integrate from the center a to x. Because of that, the constant vanishes and the lower limit often kills most terms. Cleaner that way.
And if you're approximating, truncate and then check: am I inside the radius? If not, the whole approximation is suspect.
I know it sounds simple — but it's easy to miss when you're deep in algebra.
FAQ
Can you differentiate and integrate power series anywhere? Inside the interval of convergence, yes — term by term, with the same radius. At the endpoints, check separately each time.
Does differentiation change the radius of convergence? No. The radius stays the same. Endpoint convergence can change, though.
Why is term-by-term calculus allowed for power series? Because they converge uniformly on any closed subinterval inside the radius, which lets you swap limits and integrals/derivatives. That's the technical reason.
How do I get the series for a function that isn't standard? Start from a related series you know — geometric, exponential, trig — then differentiate, integrate, substitute, or shift. Build it.
What's the easiest series to practice on? The geometric series 1/(1−x). Differentiate it, integrate it, plug in −x², whatever. It teaches every move.
The cool part is that once these operations feel like muscle memory, infinite series stop being a haunted house and start looking like a toolbox. You're not praying the math works — you know why it does, and where
it doesn't. That shift from anxiety to control is the real payoff.
One last thing worth saying out loud: none of this is magic, and none of it is a trap either. Power series are a contract. But you respect the radius, you handle the endpoints with care, and you keep your indices straight — and in return, calculus on infinity behaves exactly like calculus on polynomials. The rules are stricter than they look, but they're also more forgiving than they feel at 2 a.m. with a half-finished problem set.
So the next time you see a function that won't yield to ordinary methods, don't flinch. Differentiate, integrate, shift, solve. Think about it: write the series, mark the radius, and reach for the toolbox. The haunted house is just a workshop — and you've got the keys.
