AP Calculus Indefinite Integrals

An indefinite integral on the AP Calculus exam asks for an antiderivative plus a constant of integration, and the work that earns credit happens in the recognition stage, not the arithmetic. Most candidates lose marks not because they cannot antidifferentiate once they see the trick, but because they reach for a method before they have read the integrand carefully. The exam rewards a calm three-second scan: power, exponential, trigonometric, logarithmic, rational, inverse trigonometric, then algebraic manipulation. This article walks through that scan, the six integrand families you will actually meet, and the tactical decisions between u-substitution and integration by parts that separate a strong response from a hesitant one. Worked examples sit alongside the usual error patterns, so you leave with a method, not just a list of rules.

The shape of an AP Calculus indefinite integral question

On the AP Calculus AB and BC papers, indefinite integration rarely appears as a stand-alone item. It is the second or third step in a free-response chain that begins with a graph, a table, a verbal scenario, or a function defined piecewise or by accumulation. The typical stem reads roughly: let f be the function defined by … find an antiderivative F of f such that F(1) = 4, or write an expression for the family of antiderivatives of g. The expected output is an algebraic expression in x with + C, or a specific antiderivative satisfying an initial condition. The constant is part of the answer; leaving it off is a deduction on the AP rubric, even when the initial condition is given, because graders check the form before they check the value.

Two structural facts shape how you should approach the question. First, the scoring rubric awards one or two method points for the right setup, one or two for correct execution, and a final point for the antiderivative in the requested form. Reading the verb matters: find, write, evaluate, and determine are not interchangeable. Find the particular antiderivative demands the constant resolved. Find an antiderivative accepts + C. Second, the calculator is not your friend here. The TI-84's fnInt( computes a definite integral numerically; it will not give you a clean antiderivative expression, and the grader will not accept a decimal. Symbolic work is the only path to the mark.

Most candidates reading this will sit in one of two camps: those who learned the power rule once and have been waiting for everything else to feel as easy, and those who panic at the first trig identity. Both groups benefit from a structured reading of the integrand. The next sections build that reading, family by family, with the exact decision point where u-substitution stops working and integration by parts has to take over.

What the rubric is actually rewarding

Method points recognise the correct setup: choosing u and du correctly, or applying the product-to-sum identity, or pulling out the constant 1/2 from a coefficient. Execution points credit the differentiation or simplification that follows. The final point recognises the antiderivative expression in closed form. The constant + C sits inside the final-point judgement. When a problem gives an initial condition, the constant resolves to a number, and the rubric expects that number, not a symbol. A candidate who writes the correct family of antiderivatives but skips the constant loses the same point as a candidate who writes the wrong antiderivative — the rubric treats form and finality as a single unit.

Family one: power, polynomial, and root integrands

The most common indefinite integral on AP Calculus is the one that looks like nothing special. ∫(3x² + 5x − 7) dx, ∫(√x + 1/x) dx, ∫(x⁴ − 2x) dx. The technique is the reverse power rule, applied term by term, and the + C at the end. Candidates who lose marks here usually do so for one of two reasons: they divide by zero when n = −1, or they forget that ∫x⁻¹ dx = ln|x| + C rather than x⁰/0.

The reverse power rule states that ∫xⁿ dx = xⁿ⁺¹/(n+1) + C whenever n ≠ −1. For a polynomial, you integrate each term independently and add the constants together. For a sum of two terms, the constant from each term collapses into a single + C, because two arbitrary constants add to one arbitrary constant. A common student error is to write + C₁ and + C₂, which the rubric will not penalise explicitly but which signals a misunderstanding. Write one C.

Roots and fractional powers are powers in disguise. ∫√x dx = ∫x^(1/2) dx = (2/3)x^(3/2) + C. ∫1/√x dx = ∫x^(−1/2) dx = 2x^(1/2) + C. ∫1/x² dx = −1/x + C. The pattern holds whenever the exponent is a constant not equal to −1. When you see x⁻¹, the rule changes: the antiderivative is ln|x| + C, and the absolute value is not optional on the AP exam. The function 1/x is defined for negative x as well, and the antiderivative must be defined there too; ln(−x) handles the negative side, and the absolute value is the compact way of writing both branches.

Rational functions with linear denominators, such as ∫1/(2x − 3) dx, sit on the boundary of the power family and the substitution family. Strictly, the technique is u-substitution, but a fast reader recognises that 1/(2x − 3) behaves like (1/2) · 1/u with u = 2x − 3. The answer is (1/2)ln|2x − 3| + C. The factor of 1/2 comes from du = 2 dx, so dx = du/2, and that 1/2 has to surface somewhere. Candidates who write ln|2x − 3| + C are reading the integrand as if du = dx, which is rarely true. Always check the chain rule that produced the integrand in the first place.

Worked example: a polynomial with a twist

Find an antiderivative of f(x) = 6x² + 4x + 3x⁻². Term by term: 2x³ + 2x² + 3 · x⁻¹/(−1) + C = 2x³ + 2x² − 3x⁻¹ + C, which can also be written 2x³ + 2x² − 3/x + C. The two forms are equivalent; graders accept either. If the stem had specified F(1) = 5, you would plug in x = 1 to get 2 + 2 − 3 = 1, so C = 4, giving F(x) = 2x³ + 2x² − 3/x + 4. Note the constant resolved to a number, not a symbol.

Family two: exponential and logarithmic integrands

The exponential family covers e raised to a linear function and the natural log. The reverse rule is that ∫e^(ax) dx = (1/a)e^(ax) + C, and ∫e^u du = e^u + C when u is the substitution. The coefficient a has to be pulled out as 1/a, exactly as the chain rule would have multiplied it back in during differentiation. A common error is to treat e^(3x) as if its antiderivative were e^(3x), ignoring the factor of 3. It is not. The derivative of e^(3x) is 3e^(3x), so the antiderivative of e^(3x) is (1/3)e^(3x) + C.

For natural log, the rule is ∫(1/x) dx = ln|x| + C, which we have already met. Composite forms such as ∫(1/(3x + 2)) dx extend the same idea through u-substitution: u = 3x + 2, du = 3 dx, dx = du/3, the integral becomes (1/3)∫(1/u) du = (1/3)ln|u| + C = (1/3)ln|3x + 2| + C. The 1/3 is the price of the chain rule. Products of the form ∫x · e^(x²) dx are the cleanest u-substitution case: u = x², du = 2x dx, the x dx becomes (1/2) du, and the integral reduces to (1/2)∫e^u du = (1/2)e^(x²) + C.

BC candidates should also recognise hyperbolic-style integrands, though these are rarer. ∫sinh(x) dx = cosh(x) + C and ∫cosh(x) dx = sinh(x) + C. The exam does not require deep familiarity, but a passing recognition of these forms prevents a candidate from wasting time hunting for a u-substitution that does not exist.

Worked example: e raised to a composite

Find an antiderivative of g(x) = x² · e^(x³). Let u = x³, du = 3x² dx, so x² dx = du/3. The integral is ∫e^u · (1/3) du = (1/3)e^u + C = (1/3)e^(x³) + C. The x² is the signal that u-substitution will work; the e^(x³) is the function whose derivative structure is being exploited. If the x² were missing, you would need integration by parts.

Family three: trigonometric integrands

Trig indefinite integrals fall into three sub-families. The first is the basic six: ∫sin(x) dx = −cos(x) + C, ∫cos(x) dx = sin(x) + C, ∫sec²(x) dx = tan(x) + C, ∫sec(x)tan(x) dx = sec(x) + C, ∫csc²(x) dx = −cot(x) + C, ∫csc(x)cot(x) dx = −csc(x) + C. These six are the trigonometric derivatives in reverse, and the AP exam expects them as automatic recall. A candidate who pauses to derive each one from scratch is using up minutes that the free-response section cannot spare.

The second sub-family is composites, where the trig function has a chain inside. ∫sin(3x) dx = −(1/3)cos(3x) + C, ∫cos(5x) dx = (1/5)sin(5x) + C, ∫sec²(4x) dx = (1/4)tan(4x) + C. The pattern is the same as the exponential case: the inner function contributes a coefficient that becomes 1 divided by that coefficient on the outside. The third sub-family is products of trig functions, which require identity work. ∫sin(x)cos(x) dx can be tackled two ways: substitution with u = sin(x), or the double-angle identity sin(x)cos(x) = (1/2)sin(2x). The substitution route is faster: du = cos(x) dx, and the integral becomes ∫u du = u²/2 + C = sin²(x)/2 + C. The identity route gives −(1/4)cos(2x) + C, which is equivalent up to a constant. The grader accepts both.

BC candidates also meet ∫tan(x) dx, ∫cot(x) dx, ∫sec(x) dx, and ∫csc(x) dx. These are not reversible from the basic six. The technique is to multiply by a clever form of 1. ∫tan(x) dx = ∫sin(x)/cos(x) dx, let u = cos(x), du = −sin(x) dx, and the integral becomes −∫1/u du = −ln|cos(x)| + C = ln|sec(x)| + C. ∫sec(x) dx requires multiplying numerator and denominator by sec(x) + tan(x); the result is ln|sec(x) + tan(x)| + C. These are worth memorising because the work is not obvious and the rubric gives a method point for the right setup.

Worked example: a product that hides a derivative

Find an antiderivative of h(x) = sin(x) · cos⁴(x). The exponent on the cosine and the sin(x) sitting next to it are the signal. Let u = cos(x), du = −sin(x) dx, so sin(x) dx = −du. The integral is ∫u⁴ · (−du) = −u⁵/5 + C = −cos⁵(x)/5 + C. The sin(x) is not extra; it is the derivative of the cosine, which is exactly what u-substitution needs. If the integrand had been sin(x) · cos³(x) + cos(x), the cos(x) would not pair with anything and you would split the integral.

Family four: rational functions and partial fractions

Rational functions appear more on BC than on AB, but both exams include the simple linear-denominator case. ∫1/(x² − 1) dx requires either partial fractions or a trigonometric substitution. Partial fractions splits 1/(x² − 1) into 1/2 · [1/(x − 1) − 1/(x + 1)], and each piece integrates to a logarithm. The result is (1/2)ln|(x − 1)/(x + 1)| + C. Trig substitution handles 1/(x² + 1) by letting x = tan(θ), dx = sec²(θ) dθ, and the denominator becomes sec²(θ); the integrand simplifies to 1, and the integral is θ + C = arctan(x) + C.

The decision rule between partial fractions and trig substitution depends on the form of the denominator. Sum of squares, x² + a², suggests trig substitution with x = a tan(θ). Difference of squares, x² − a², suggests hyperbolic substitution or partial fractions. Sum of a square and a linear term, x² + 2x + 5, requires completing the square first; once you have (x + 1)² + 4, the substitution u = x + 1 collapses it to ∫1/(u² + 4) du = (1/2)arctan(u/2) + C. Completing the square is a free method point on the rubric; skipping it and trying a non-existent trig substitution is a common way to lose two minutes and one mark.

Worked example: completing the square

Find an antiderivative of k(x) = 1/(x² + 4x + 13). Complete the square: x² + 4x + 13 = (x + 2)² + 9. Let u = x + 2, du = dx. The integral becomes ∫1/(u² + 9) du, which matches the form a² = 9, a = 3, so the answer is (1/3)arctan(u/3) + C = (1/3)arctan((x + 2)/3) + C. The grader will look for the completed square; writing arctan(x² + 4x + 13) as if it factored is a sign that the work has gone wrong, and it costs the execution point.

Family five: inverse trigonometric integrands

The reverse of the derivative of arctan, arcsin, and arccsc gives a small but recurring family. ∫1/(1 + x²) dx = arctan(x) + C. ∫1/√(1 − x²) dx = arcsin(x) + C. ∫1/(x√(x² − 1)) dx = arcsec(x) + C. The general form is ∫1/√(a² − u²) du = arcsin(u/a) + C and ∫1/(a² + u²) du = (1/a)arctan(u/a) + C. The rubric will accept any of the equivalent forms, including arctan and arccot on one side, and arcsin and arccos on the other, provided the coefficient on the outside is correct.

These are usually reached by u-substitution: an integrand such as 1/(9 + x²) becomes 1/9 · 1/(1 + (x/3)²), let u = x/3, du = dx/3, the integral becomes (1/3) · (1/3) · ∫1/(1 + u²) du = (1/3) · (1/3) arctan(u) + C = (1/9) arctan(x/3) + C. The factors of 1/3 are the price of the chain rule, exactly as in the exponential and trig cases. Candidates who write arctan(x/3) without the 1/9 are skipping a step, and the rubric will not give full credit.

Family six: u-substitution versus integration by parts

This is the decision that costs the most marks, because both techniques are on the syllabus and the integrand does not announce which one applies. The rule of thumb, learned the hard way by most candidates, is that u-substitution works whenever the integrand contains a function and a reasonable multiple of its derivative. The polynomial times e^(x²) example worked because x² was inside the exponential and 2x was outside. The sin(x) times cos⁴(x) example worked because cos⁴(x) was inside and sin(x) was the derivative of cos(x). When the integrand is a product where neither factor is the derivative of the other — x · e^x, x · sin(x), ln(x), x² · ln(x) — u-substitution fails, and integration by parts is the only tool.

Integration by parts uses the formula ∫u dv = uv − ∫v du. The choice of u and dv is the art. The LIATE acronym helps: Logarithmic, Inverse trig, Algebraic, Trig, Exponential — pick u in that order, and the leftover becomes dv. For ∫x · e^x dx, u = x (algebraic) and dv = e^x dx (exponential). Then du = dx and v = e^x. The integral is x · e^x − ∫e^x dx = x · e^x − e^x + C = e^x(x − 1) + C. The technique is correct, and the constant is required at the end.

For ∫x · sin(x) dx, u = x, dv = sin(x) dx, du = dx, v = −cos(x). The integral is −x · cos(x) − ∫(−cos(x)) dx = −x · cos(x) + sin(x) + C. For ∫ln(x) dx, there is no obvious dv, so let u = ln(x) and dv = dx. Then du = (1/x) dx and v = x. The integral is x · ln(x) − ∫x · (1/x) dx = x · ln(x) − x + C. This last example is the most common trap: candidates try to use u-substitution with u = ln(x), which gives du = (1/x) dx but leaves an x in the integrand that does not match. The method has to switch to integration by parts, and the LIATE acronym is what tells you to switch.

Tabular integration is a useful shortcut for repeated integration by parts, especially when u and dv lead to a cycle. If integrating by parts gives you back the original integral with a sign change, you can solve algebraically. ∫e^x · sin(x) dx: let u = e^x, dv = sin(x) dx; then du = e^x dx, v = −cos(x). The integral is −e^x · cos(x) + ∫e^x · cos(x) dx. The new integral still has the same shape, so apply parts again: u = e^x, dv = cos(x) dx; du = e^x dx, v = sin(x). The integral becomes −e^x · cos(x) + e^x · sin(x) − ∫e^x · sin(x) dx. Now there is an ∫e^x · sin(x) dx on both sides. Add it to both sides: 2∫e^x · sin(x) dx = e^x(sin(x) − cos(x)) + C, so the answer is (1/2)e^x(sin(x) − cos(x)) + C. The constant resurfaces at the end, never in the middle.

How u-substitution and integration by parts fit into the AP scoring rubric

The AP scoring rubric for a free-response integration question typically has three or four points. One point goes to setting up the technique: identifying u and du correctly, choosing the LIATE ordering, or stating the integration-by-parts formula. One point goes to the differentiation or simplification that follows: actually computing du from u, or v from dv. One point goes to the final antiderivative expression in closed form. A possible fourth point goes to applying an initial condition and reporting the resolved constant.

Most candidates reading this will lose the method point rather than the execution point. The execution is mechanical: you differentiate u, you find an antiderivative of dv, you multiply and subtract. The method is interpretive: you have to see the structure of the integrand and choose the right technique. A candidate who attempts integration by parts on a substitution-friendly integrand is reading the integrand wrongly, and the rubric reflects that by awarding zero on the first line and zero on the second because the setup is incorrect.

There is a tactical reason to prefer u-substitution when both are plausible. u-substitution is faster on the exam, leaves less room for sign errors, and matches the structure of the rubric's first line: let u = …, then du = … dx. If you can spot a u-substitution, take it. Integration by parts should be a second choice, used only when no clean u-substitution exists. For most candidates, this means: try u-substitution for ten seconds, fail, switch to integration by parts.

Common pitfalls and how to avoid them

The most common pitfall on indefinite integrals is the missing constant of integration. The AP rubric explicitly deducts a point for omitting + C on an indefinite integral, even if the candidate never applied an initial condition. A close second is the wrong coefficient from the chain rule: writing e^(3x) + C instead of (1/3)e^(3x) + C, or sin(5x) + C instead of −(1/5)cos(5x) + C. The third is a sign error: the derivative of cos(x) is −sin(x), so the antiderivative of sin(x) is −cos(x) + C, not cos(x) + C. Sign errors are particularly common on integration by parts, because the formula subtracts ∫v du, and a sign flip in v propagates.

Another pitfall is applying the wrong identity. Candidates who try to integrate sin²(x) by recalling that the derivative of sin(x) is cos(x) get stuck; the right move is the half-angle identity sin²(x) = (1 − cos(2x))/2, which makes the integral a linear combination of 1 and cos(2x), both of which are in the basic family. Trying to use power reduction on cos³(x) by writing it as cos(x) · cos²(x) and then replacing cos²(x) with (1 + cos(2x))/2 works, but most candidates save time by using the substitution u = sin(x) instead. The decision between identities and substitution is a time judgement: identities are usually one step faster when the exponent is even, substitution when the exponent is odd.

A fifth pitfall is over-engineering. If the integrand is a polynomial, do not reach for integration by parts. If the integrand is x · e^(x²), the polynomial is the derivative of the exponent, and u-substitution finishes in one line. If the integrand is x · e^x, there is no u-substitution, and integration by parts is the only path. Reading the integrand for the derivative of the inner function is the work that takes ten seconds and saves two minutes.

Comparing the six families at a glance

The table below summarises the six integrand families, the typical signal in the expression, and the technique the exam expects. Use it as a decision aid during practice, not as a memorisation list; the rubric wants the method, but the method is chosen by the structure, and the structure is what the table is describing.

Family	Signal in the integrand	Primary technique	Sample antiderivative
Power, polynomial, root	xⁿ, √x, 1/x with chain	Reverse power rule, term by term	∫(3x² + 1) dx = x³ + x + C
Exponential and logarithmic	e^(ax), 1/(ax + b)	Reverse rule, then chain adjustment	∫e^(3x) dx = (1/3)e^(3x) + C
Trigonometric	sin, cos, sec, csc, tan, cot with chain	Reverse rule, identity work, or substitution	∫sin(5x) dx = −(1/5)cos(5x) + C
Rational	1/(quadratic), partial fractions needed	Partial fractions or trig substitution	∫1/(x² + 1) dx = arctan(x) + C
Inverse trigonometric	1/√(a² − u²), 1/(a² + u²)	Direct reverse rule with a coefficient	∫1/√(1 − x²) dx = arcsin(x) + C
Products without a derivative pair	x · e^x, x · sin(x), ln(x)	Integration by parts, LIATE ordering	∫x · e^x dx = e^x(x − 1) + C

Practice sequencing that matches the AP Calculus exam

For most candidates, the right preparation order is: power and polynomial, exponential, basic trig, then rational and inverse trig, then u-substitution as a unifying lens, then integration by parts last. The reason for the order is that u-substitution is a technique that applies across all the families, and it is easier to learn it on simple integrands before generalising. Integration by parts is harder because it requires choosing u and dv, and the LIATE acronym makes more sense once the basic families are automatic.

In a typical ten-week preparation plan, weeks one and two cover the power and exponential families plus the basic six trig derivatives in reverse. Weeks three and four introduce u-substitution as a method, working through composite exponentials, composite trig, and product integrands where one factor is a derivative of the inner function. Weeks five and six cover rational functions, completing the square, and the inverse trig family. Week seven introduces integration by parts with simple products such as x · e^x and x · sin(x). Week eight covers repeated integration by parts and the cycle technique for e^x · sin(x) and e^x · cos(x). Week nine is mixed practice under timed conditions, ideally with past AP free-response items. Week ten is a slower review of the items where marks were dropped, organised by family.

The most useful diagnostic question after each practice set is: which of the six families did this integrand belong to, and what was the signal that pointed there? If the answer is unclear, the candidate is reading the integrand at the level of the arithmetic, not the structure, and that is the habit to fix before the next timed practice. In my experience, the candidates who improve fastest are the ones who write the family name next to each practice problem in the margin. The annotation slows them down for a week, then speeds them up for the rest of the preparation.

What this means for AP Calculus exam preparation

The syllabus treats indefinite integrals as one piece of a larger differentiation and integration framework that also covers the Fundamental Theorem of Calculus, definite integrals, the accumulation function, and differential equations. The free-response section tests indefinite integration as a means to an end: most of the time, the antiderivative is the answer to a sub-part of a larger problem about accumulation, average value, or the solution of a differential equation. The scoring rewards fluency in the technique, not novelty. A clean, two-line u-substitution earns full marks; a clever trick that takes three minutes and produces a wrong constant loses marks.

Candidates who want to score a 5 on the AP Calculus exam should be able to recognise the integrand family within ten seconds, choose a technique, execute it, and report the antiderivative with the correct constant in under three minutes per free-response sub-part. That target is achievable for any candidate who has practised the six families to automaticity and can switch to integration by parts when u-substitution fails. The work is mechanical once the recognition is in place, and the recognition is what the preparation plan builds.

For candidates who sit AP Calculus BC rather than AB, the syllabus adds partial fractions, improper integrals, and the logistic differential equation to the integration toolkit, but the structure of the questions is the same. The additional families are recognisable extensions of the rational and inverse trig families, and the same u-substitution-versus-integration-by-parts decision applies. Building automaticity in the six families is the foundation; the BC-specific items sit on top of it without changing the underlying technique.

Conclusion and next steps

Indefinite integrals on the AP Calculus exam are a recognition problem disguised as a computation problem. The six integrand families — power, exponential, trigonometric, rational, inverse trigonometric, and the integration-by-parts products — cover every antiderivative the syllabus expects, and the decision between u-substitution and integration by parts is the work the rubric rewards. Practice should be sequenced by family, with u-substitution as the default technique and integration by parts reserved for integrands that contain no derivative pair. The constant of integration is part of the answer, the chain-rule coefficient must surface on the outside, and the initial condition, when given, resolves the constant to a number.

TestPrep Europe's diagnostic assessment on AP Calculus indefinite integrals is a natural starting point for candidates who want a precise reading of which family recognition is sharp and which is still soft, with a practice plan tailored to the result.

Frequently asked questions

How is an indefinite integral different from a definite integral on the AP Calculus exam?

An indefinite integral asks for an antiderivative expressed as a function of x plus a constant of integration, and the answer is a family of curves. A definite integral asks for a numerical value over a specific interval, computed by evaluating an antiderivative at the upper and lower limits and subtracting. The AP rubric treats them very differently: indefinite integrals require the constant, definite integrals do not.

When should I use integration by parts instead of u-substitution?

Use u-substitution first whenever the integrand contains a function and a multiple of its derivative sitting next to it, such as x · e^(x²) or sin(x) · cos⁴(x). Switch to integration by parts only when the integrand is a product of two function families where neither factor is the derivative of the other, such as x · e^x, x · ln(x), or ln(x) on its own. The LIATE acronym gives the ordering: pick u from Logarithmic, Inverse trig, Algebraic, Trig, or Exponential, in that order.

Do I have to write the constant of integration on every AP Calculus indefinite integral?

Yes. The constant of integration, written as + C, is part of the answer whenever the problem asks for an antiderivative without specifying an initial condition. The AP rubric deducts a point for omitting it. When the problem does give an initial condition, the constant resolves to a specific number, and the rubric expects that number, not a symbol.

Why does the chain rule affect the antiderivative?

Differentiation by the chain rule multiplies by the derivative of the inner function, so the reverse operation must divide by that derivative. The result is a coefficient on the outside of the antiderivative, such as 1/3 when integrating e^(3x) or 1/5 when integrating cos(5x). Skipping this coefficient is one of the most common ways to lose the execution point on the AP rubric.

Is there a difference between the AB and BC syllabus on indefinite integrals?

Both AB and BC cover the same six integrand families and the same two techniques. BC extends the rational family with partial fractions and adds the logistic differential equation, which requires separation of variables and a partial-fractions step. The recognition work and the u-substitution-versus-integration-by-parts decision are identical across the two courses.

AP Calculus Indefinite Integrals: which technique to reach for first