Improper integrals sit at the junction of three habits the AP Calculus BC exam rewards: careful reading of the limits, disciplined handling of infinity, and the willingness to decide whether an answer exists at all. For a YÖS or TR-YÖS candidate building a preparation plan that borrows from the College Board syllabus, this single topic is unusually efficient, because the same handful of techniques cover roughly half a unit of marks. A candidate who can identify the type of improper integral in ten seconds, set up the correct limit, evaluate it cleanly, and then judge convergence, will collect the points other students leave on the table. This article walks through that workflow with worked examples, common traps, and a diagnostic view of the question families you are most likely to face.
What the AP Calculus BC syllabus means by "improper"
An integral is called improper whenever one of two situations appears: either one of the limits of integration is infinite, or the integrand itself is undefined somewhere inside the interval of integration. Both situations break the Fundamental Theorem of Calculus, which assumes a continuous function on a closed bounded interval. The fix is the same in both cases: rewrite the improper integral as a limit, evaluate the limit, and only then declare a value.
The first case, an infinite limit, is often written as the integral from a to infinity, from minus infinity to b, or from minus infinity to plus infinity. The second case, an interior discontinuity, usually involves a vertical asymptote: a denominator that vanishes, the logarithm of a quantity that touches zero, or an integrand that diverges at the upper or lower endpoint.
For a YÖS candidate, the practical consequence is that every improper integral is, before it is anything else, a limit problem. The integrand may be a clean rational function or it may involve e raised to a negative power; the underlying machinery is identical. Students who internalise this point early stop panicking at the sight of an infinity symbol and start reading the problem for what it actually asks.
It also helps to notice how the topic interlocks with earlier units. The substitution techniques from Unit 6 and the integration by parts pattern from Unit 7 are both fair game inside an improper integral. The new content is not the antiderivative; the new content is the limit. A solid preparation sequence is therefore: revisit antiderivatives, then layer the improper-integral framework on top, then drill convergence decisions.
Two types, three situations
- Type 1, infinite upper limit: ∫a∞ f(x) dx = limb→∞ ∫ab f(x) dx.
- Type 1, infinite lower limit: ∫-∞b f(x) dx = lima→-∞ ∫ab f(x) dx.
- Type 2, interior discontinuity: ∫ab f(x) dx when f is unbounded on (a, b); split at the bad point c and take two limits.
If you can read an item and decide which of these three situations applies within five seconds, you are already past the first hurdle. Most AP-style items set up so that the classification is the only real thinking step; the rest is mechanical.
Setting up the limit: notation that will not lose you marks
Notation errors cost more on improper integrals than on almost any other topic, because the marker is looking for the limit step itself. If you write the answer without ever showing the limit, you are betting on full credit. In my experience, that bet fails often enough to matter. A clean setup reads as: define the integral as a limit of definite integrals, evaluate the inner definite integral symbolically, then compute the limit.
Take the canonical example ∫1∞ 1/x² dx. The correct setup is limb→∞ ∫1b x-2 dx = limb→∞ [-1/x]1b = limb→∞ (-1/b + 1). The limit is 1. Most students will write "1" and lose a mark or two; the candidates who write the limit explicitly will not.
For a Type 2 example, consider ∫01 1/√x dx. The integrand blows up at x = 0, so the setup is lima→0+ ∫a1 x-1/2 dx = lima→0+ [2x1/2]a1 = lima→0+ (2 - 2√a) = 2. Note the one-sided limit. The integrand is only defined for x > 0 in this form, so a two-sided limit would be meaningless.
A third template, useful when both infinities appear, is the split. The integral from minus infinity to plus infinity is not one limit; it is two limits glued together, evaluated separately. You choose a finite point, usually zero, write the integral as a sum of two improper integrals at that point, and check that both halves converge. If only one converges, the whole integral diverges. AP markers are fond of items where this rule decides the answer.
Convergence versus divergence: the decisions that decide your score
Once the limit is set up, the question becomes whether the limit exists as a finite real number. The four outcomes a marker will accept are: a finite positive number, a finite negative number, plus infinity, minus infinity, or "does not exist". Of these, the first two count as convergent. The remaining three count as divergent, and divergent answers are valid in their own right, but the question usually requires you to name the type of divergence.
For most YÖS-aligned practice, the items test p-integrals, exponential decay versus growth, and ratios that compare against known series. A p-integral ∫1∞ 1/xp dx converges when p > 1 and diverges when p ≤ 1. Memorise this; it is the single most useful sentence in the topic. A similar rule applies at the lower end: ∫01 1/xp dx converges when p < 1 and diverges when p ≥ 1. Together these two rules cover a large fraction of the convergence questions you will face.
For items that are not pure p-integrals, the comparison test is the workhorse. If 0 ≤ f(x) ≤ g(x) on the interval and ∫g converges, then ∫f converges. If 0 ≤ g(x) ≤ f(x) and ∫g diverges, then ∫f diverges. Many AP items are designed so that a quick comparison with 1/x or 1/x² is enough. Candidates who reach for L'Hôpital inside the comparison are usually over-complicating things; the comparison is the answer.
One more decision rule: the integrand must not oscillate to a finite limit while failing to settle. The integral of sin(x)/x from zero to infinity converges, even though 1/x alone diverges, because the oscillations cancel out. The AP rarely asks for that level of subtlety, but the limit comparison test handles the cases it does ask, and knowing the two outcomes (convergent or divergent, plus the value when convergent) is enough to mark the item.
Worked example 1: a Type 1 infinite limit with substitution
Consider ∫0∞ x·e-x dx. This is a staple of multiple-choice sections, because the gamma-function flavour rewards students who recognise the form. The setup is limb→∞ ∫0b x·e-x dx. Integrate by parts with u = x, dv = e-x dx, giving du = dx and v = -e-x. The antiderivative is -x·e-x - ∫ -e-x dx = -x·e-x - e-x + C.
Evaluate from 0 to b: [-x·e-x - e-x]0b = (-b·e-b - e-b) - (0 - 1) = 1 - b·e-b - e-b. Now take the limit as b → ∞. The term b·e-b is the classic case for L'Hôpital: rewrite as b/ex, which goes to 0. The term e-b goes to 0. The result is 1.
The trap students fall into is to treat this as an ordinary integral and write "1" without ever writing "limb→∞". The marker will accept the answer, but the work is incomplete, and in a free-response section that omission can cost a point. Treat the limit as a mandatory line of work, not a stylistic flourish.