Foundations — Improper integrals — Type I (infinite limits), Type II (discontinuous integrand)

Before you can read the parent note, you must own every piece of notation it fires at you. Below, each symbol is built from nothing: plain words → the picture → why the topic can't live without it. Read top to bottom; every item leans on the one above it.

1. The function $f(x)$ and its graph

Plain words. $x$ is the input (where you stand on the horizontal line). $f(x)$ is the height the rule assigns there.

The picture. A wiggly line floating above (or below) the horizontal axis. Standing at $x=2$ and looking straight up until you hit the curve tells you $f(2)$.

Why the topic needs it. Every question in this chapter is "how much area sits between this curve and the horizontal axis?" You can't ask about area until you can point at the curve whose area you mean.

2. The axes, $a$, $b$, and the interval $[a,b]$

Plain words. $a$ is your left wall, $b$ is your right wall. $[a,b]$ is everything between and on those two walls.

The picture. Two vertical fences planted at $x=a$ and $x=b$. The area you care about is the paint trapped between the fences, under the curve, above the ground line.

Why the topic needs it. A normal integral demands both walls sit at real, finite positions. Improper integrals are exactly the cases where one wall is pushed to infinity ($b=\infty$) — or where the curve punches through the ceiling inside the fences. The bracket-vs-parenthesis distinction ($[a,b)$ vs $[a,b]$) is how the parent note quietly signals "the endpoint is a bad point, don't stand on it."

3. The integral sign ${\int }_a^{b}f(x)dx$ = area

Plain words. Chop the region under the curve into millions of skinny rectangles. Each rectangle has height $f(x)$ and a whisker-thin width $dx$. Multiply to get its area, then total them all.

The picture. A picket-fence of thin bars filling the region; their combined area is the integral.

"Signed" area. Where the curve dips below the axis, $f(x)<0$, so the strip area counts as negative. This single fact is the red flag in the parent's Type II mistake: a curve that is always positive can never produce a negative total, so the bogus answer $-2$ was impossible from the start.

Why the topic needs it. This is the entire quantity we are trying to compute. Everything else — limits, splitting, $p$-rules — is machinery for evaluating this symbol when a wall runs to infinity or the height blows up.

4. Continuous vs. the vertical asymptote (blow-up)

Plain words. "Continuous" = smooth, unbroken. "Bounded" = never taller than some fixed height. A blow-up = a spot where the curve becomes infinitely tall.

The picture. Compare $y=1/x^2$ near $x=0$: as you walk toward $0$, the curve climbs faster and faster, never touching the vertical axis but racing upward forever. That vertical axis is the asymptote.

Why the topic needs it. The Fundamental Theorem of Calculus (see Fundamental Theorem of Calculus) only works on a closed, bounded, continuous piece. A blow-up violates bounded and continuous at once — that is the exact defect that makes an integral Type II. Spotting the asymptote before you integrate is what stops the parent note's $-2$ blunder.

5. Infinity $\infty$ — a direction, not a number

Plain words. $\infty$ is shorthand for "keeps going forever, with no last value."

The picture. The right fence from §2, but ripped out and dragged rightward without ever stopping. The region under the curve now has no right wall at all.

Why the topic needs it. An interval like $[1,\infty )$ has an infinite right side — this is precisely Type I. Because $\infty$ is not a number, you literally cannot write $F(\infty )-F(1)$. We need a legal detour, which is the next symbol.

6. The moving wall $t$ and the limit $\underset{t\to \infty }{\lim }$

Plain words. $t$ is a temporary wall that we are allowed to place at any real position. We compute the ordinary area up to $t$, get a formula in $t$, then slide $t$ toward the forbidden spot and watch the number.

The picture. A vertical fence at position $t$, painted, then dragged rightward. Watch the paint total on a meter: if the needle settles on, say, $1$, the integral converges to $1$; if the needle climbs forever, it diverges.

Why the topic needs it. This is the master move of the whole topic. Both definitions in the parent note are the same trick: ${\int }_a^{\infty }fdx={\lim }_{t\to \infty }{\int }_a^{t}fdx,{\int }_a^{b}fdx={\lim }_{t\to b^{-}}{\int }_a^{t}fdx.$ Infinity and blow-ups become ordinary limit problems (see Limits at Infinity). $t$ is the finite stand-in that lets FTC do its job legally.

7. The antiderivative $F$ and the power rule

Plain words. An antiderivative runs the derivative machine in reverse. Once you have $F$, the FTC hands you the area: ${\int }_a^{t}f=F(t)-F(a)$.

The power rule (the one tool the $p$-integral leans on entirely): $\int x^{-p}dx=\frac{x^{1-p}}{1-p}(p\ne 1),\int x^{-1}dx=\ln x(p=1).$

Why this tool and not another? The parent note's headline integral is $\int 1/x^{p}dx$, a pure power of $x$. The power rule is the one antidifferentiation rule built for exactly powers of $x$ — nothing fancier is needed. The lone exception is $p=1$: the power rule would divide by $1-p=0$, which is illegal, so that single case defers to the logarithm $\ln x$. That is precisely why $p=1$ is the famous knife-edge in the parent note.

8. The exponent $p$ and the phrase "$p$-integral"

The picture. A fan of curves $x^{-1/2},x^{-1},x^{-2}$ on one axis. To the far right they peel apart — steeper $p$ hugs the axis sooner. Near $0$ they cross the other way — steeper $p$ climbs the ceiling sooner.

Why the topic needs it. Sliding one dial $p$ reproduces every convergence verdict in the chapter. This is why the parent calls it the "80/20 powerhouse" and why the Comparison Test for Integrals and p-series and Integral Test are its direct children.

How it all feeds the topic

Read it upward: to understand an improper integral you need area, infinity, blow-ups, and the moving-wall limit; the limit itself runs on the antiderivative; dialing the exponent $p$ turns the whole machine into the master $p$-integral, which then powers every later convergence test.

Where these tools reappear

• The finite area ${\int }_0^{\infty }{e}^{-x}dx=1$ is what makes an exponential a legal probability density.
• The moving-wall-to-infinity limit is the beating heart of the Laplace Transform and the Gamma Function.
• The $p$-verdict at infinity is the integral cousin of the p-series.

Equipment checklist

Test yourself — cover the right side and answer before revealing.