Indefinite integrals

At a Glance

Why This Chapter Matters

Integration trick questions appear in compulsory Q1 (Section A) and occasionally Q2/Q3. They are universally solvable — no new theorem is needed, just pattern recognition. Five archetypes account for every UPSC question on this atom. The hardest (King’s reflection, Beta/Gamma reduction) are worth 10 marks and take 5–8 minutes. Recognising the archetype at a glance is the single most valuable skill here: the wrong technique can consume 15 minutes without progress, while the correct one reaches the answer in 3.

Minimum Theory

The Fundamental Theorem (FTC) and its subtlety. If F′(x)=f(x)F'(x)=f(x) on (a,b](a,b] and FF extends continuously to aa, then ∫abf(x) dx=F(b)−F(a)\int_a^b f(x)\,dx = F(b)-F(a). The key subtlety: F′(a)F'(a) need not exist (the antiderivative can be non-differentiable at a single point), yet FTC still holds. The improper-FTC version via lim⁡ε→0∫a+εb\lim_{\varepsilon\to0}\int_{a+\varepsilon}^b is how such cases are handled rigorously.

King’s (reflection) property. For any integrable ff and a≤ba\le b: ∫abf(x) dx=∫abf(a+b−x) dx.\int_a^b f(x)\,dx = \int_a^b f(a+b-x)\,dx. Adding the two expressions and dividing by 2 gives a self-referential integral that collapses cleanly when f(x)+f(a+b−x)f(x)+f(a+b-x) simplifies to a constant. This is the most versatile trick in UPSC integration questions.

Beta and Gamma functions. Γ(s)=∫0∞ts−1e−t dt\Gamma(s)=\int_0^\infty t^{s-1}e^{-t}\,dt (valid for s>0s>0), with Γ(n)=(n−1)!\Gamma(n)=(n-1)! for positive integers and Γ(1/2)=π\Gamma(1/2)=\sqrt\pi. The master reduction: ∫01xm(−log⁡x)n dx=Γ(n+1)(m+1)n+1\int_0^1 x^m(-\log x)^n\,dx = \dfrac{\Gamma(n+1)}{(m+1)^{n+1}}, obtained by the substitution x=e−tx=e^{-t}. The Beta function B(p,q)=∫01tp−1(1−t)q−1 dt=Γ(p)Γ(q)/Γ(p+q)B(p,q)=\int_0^1 t^{p-1}(1-t)^{q-1}\,dt=\Gamma(p)\Gamma(q)/\Gamma(p+q), with the scaling ∫ab(x−a)m(b−x)n dx=(b−a)m+n+1B(m+1,n+1)\int_a^b (x-a)^m(b-x)^n\,dx=(b-a)^{m+n+1}B(m+1,n+1).

Question Archetypes

ArchetypeRecognition cue
recognize-antiderivativeIntegrand “looks like” a product-rule expansion d/dx[F]d/dx[F]
symmetry-propertySymmetric or complementary limits; trig integrand where sin⁡↔cos⁡\sin\leftrightarrow\cos swap helps
beta-gamma-reduction∫01xm(−log⁡x)n dx\int_0^1 x^m(-\log x)^n\,dx or ∫ab(x−a)m(b−x)n dx\int_a^b(x-a)^m(b-x)^n\,dx form
integration-by-partsInverse-trig, log, or polynomial times transcendental; boundary term contributes
integral-functional-equationAn equation of the form ∫0xf=(…)\int_0^x f = (\ldots); differentiate to find ff

recognize-antiderivative (1 question(s); 2013)

Recognise the integrand as an exact derivative and apply FTC

Recognition Cues

The integrand consists of two terms that look “mismatched” but together form the derivative of a product. Look for structures like u′v+uv′u'v + uv' (product rule) or ddx[uv]\frac{d}{dx}[uv] after grouping. A telltale sign: direct integration of each piece is intractable, but their sum is immediately an exact derivative.

Solution Template

  1. Stare at the two additive terms and ask: could this be (fg)′=f′g+fg′(fg)'=f'g+fg' for some f,gf,g?
  2. Identify ff and gg; verify by differentiating fgfg.
  3. Extend fgfg continuously to the endpoint if needed.
  4. Apply FTC (possibly as a limit ε→0\varepsilon\to0).

Worked Example

2013 Paper 1, 2013-P1-Q1c (10 marks)

Evaluate ∫01 ⁣(2xsin⁡1x−cos⁡1x)dx\displaystyle\int_0^1\!\left(2x\sin\frac{1}{x}-\cos\frac{1}{x}\right)dx.

Step 1 — Spot the antiderivative. For x>0x>0: ddx ⁣[x2sin⁡1x]=2xsin⁡1x+x2cos⁡1x⋅ ⁣(−1x2)=2xsin⁡1x−cos⁡1x.\frac{d}{dx}\!\left[x^2\sin\frac{1}{x}\right] = 2x\sin\frac{1}{x} + x^2\cos\frac{1}{x}\cdot\!\left(-\frac{1}{x^2}\right) = 2x\sin\frac{1}{x} - \cos\frac{1}{x}. The integrand is exactly F′(x)F'(x) with F(x)=x2sin⁡(1/x)F(x)=x^2\sin(1/x).

Step 2 — Handle x=0x=0. Set F(0)=0F(0)=0; then ∣F(x)∣≤x2→0|F(x)|\le x^2\to 0, so FF is continuous at 00. The integrand is bounded (with an essential discontinuity at 00, hence Riemann integrable).

Step 3 — Apply FTC via limit: ∫01=lim⁡ε→0∫ε1=F(1)−F(ε)=sin⁡1−ε2sin⁡(1/ε)→sin⁡1.\int_0^1 = \lim_{\varepsilon\to0}\int_\varepsilon^1 = F(1)-F(\varepsilon) = \sin 1 - \varepsilon^2\sin(1/\varepsilon)\to\sin 1. ∫01 ⁣(2xsin⁡1x−cos⁡1x)dx=sin⁡1.\boxed{\int_0^1\!\left(2x\sin\frac{1}{x}-\cos\frac{1}{x}\right)dx = \sin 1.}

Common Traps

symmetry-property (2 question(s); 2014, 2015)

Use a reflection/King property of definite integrals

Recognition Cues

The integral has limits [a,b][a,b] symmetric about some midpoint, and f(x)+f(a+b−x)f(x)+f(a+b-x) simplifies to a constant. Canonical form: ∫π/6π/3\int_{\pi/6}^{\pi/3} (midpoint π/4\pi/4, where sin⁡↔cos⁡\sin\leftrightarrow\cos) or ∫01\int_0^1 (midpoint 1/21/2). Trig functions with complementary arguments, logarithms, and cube roots of trig ratios all signal this archetype.

Solution Template

  1. Call the integral II.
  2. Substitute x→a+b−xx \to a+b-x (limits stay [a,b][a,b]; dx→−dxdx \to -dx with a sign flip from reversing limits). Call the result II again (same value).
  3. Add the two expressions for II; the integrand simplifies (often to a constant or constant times a simpler function).
  4. Divide by 2.

Worked Example 1

2015 Paper 1, 2015-P1-Q1d (10 marks)

Evaluate I=∫π/6π/3sin⁡x3sin⁡x3+cos⁡x3 dxI = \displaystyle\int_{\pi/6}^{\pi/3}\frac{\sqrt[3]{\sin x}}{\sqrt[3]{\sin x}+\sqrt[3]{\cos x}}\,dx.

The limits are symmetric about π/4\pi/4 (since π/6+π/3=π/2\pi/6+\pi/3=\pi/2). Substitute x→π/2−xx\to\pi/2-x; then sin⁡x↔cos⁡x\sin x \leftrightarrow \cos x: I=∫π/6π/3cos⁡x3cos⁡x3+sin⁡x3 dx.I = \int_{\pi/6}^{\pi/3}\frac{\sqrt[3]{\cos x}}{\sqrt[3]{\cos x}+\sqrt[3]{\sin x}}\,dx. Add: 2I=∫π/6π/3sin⁡x3+cos⁡x3sin⁡x3+cos⁡x3 dx=∫π/6π/31 dx=π3−π6=π6.2I = \int_{\pi/6}^{\pi/3}\frac{\sqrt[3]{\sin x}+\sqrt[3]{\cos x}}{\sqrt[3]{\sin x}+\sqrt[3]{\cos x}}\,dx = \int_{\pi/6}^{\pi/3}1\,dx = \frac{\pi}{3}-\frac{\pi}{6}=\frac{\pi}{6}. I=π12.\boxed{I = \frac{\pi}{12}.}

Note: the cube-root is a red herring — any g(sin⁡x)/[g(sin⁡x)+g(cos⁡x)]g(\sin x)/[g(\sin x)+g(\cos x)] yields I=(interval length)/2I=(\text{interval length})/2.

Worked Example 2

2014 Paper 1, 2014-P1-Q1d (10 marks)

Evaluate ∫01log⁡(1+x)1+x2 dx\displaystyle\int_0^1\frac{\log(1+x)}{1+x^2}\,dx.

Substitute x=tan⁡θx=\tan\theta: the integral becomes I=∫0π/4log⁡(1+tan⁡θ) dθI = \int_0^{\pi/4}\log(1+\tan\theta)\,d\theta.

Apply King with a+b−θ=π/4−θa+b-\theta = \pi/4-\theta; use tan⁡(π/4−θ)=(1−tan⁡θ)/(1+tan⁡θ)\tan(\pi/4-\theta)=(1-\tan\theta)/(1+\tan\theta): 1+tan⁡(π/4−θ)=21+tan⁡θ,log⁡(⋅)=log⁡2−log⁡(1+tan⁡θ).1+\tan(\pi/4-\theta) = \frac{2}{1+\tan\theta}, \quad \log(\cdot) = \log 2-\log(1+\tan\theta). Then I=∫0π/4[log⁡2−log⁡(1+tan⁡θ)] dθ=π4log⁡2−II = \int_0^{\pi/4}[\log 2-\log(1+\tan\theta)]\,d\theta = \frac{\pi}{4}\log 2 - I, giving: I=π8log⁡2.\boxed{I = \frac{\pi}{8}\log 2.}

The key: the upper limit is π/4\pi/4 precisely because tan⁡(π/4)=1\tan(\pi/4)=1 makes the symmetry work.

Common Traps

beta-gamma-reduction (2 question(s); 2016, 2021)

Reduce an integral to a Beta/Gamma function via substitution

Recognition Cues

The integrand contains xm(−log⁡x)nx^m(-\log x)^n on [0,1][0,1], or (x−a)m(b−x)n(x-a)^m(b-x)^n on [a,b][a,b], or a product of two power functions on an interval. The substitution x=e−tx=e^{-t} converts the first type to a Gamma integral; linear rescaling converts the second to a Beta integral.

Solution Template

For ∫01xm(−log⁡x)n dx\int_0^1 x^m(-\log x)^n\,dx: substitute x=e−tx=e^{-t}, −log⁡x=t-\log x = t, dx=−e−tdtdx = -e^{-t}dt, limits [0,1]→[∞,0][0,1]\to[\infty,0] (flip sign): ∫0∞(e−t)mtne−t dt=∫0∞tne−(m+1)t dt=Γ(n+1)(m+1)n+1.\int_0^\infty (e^{-t})^m t^n e^{-t}\,dt = \int_0^\infty t^n e^{-(m+1)t}\,dt = \frac{\Gamma(n+1)}{(m+1)^{n+1}}.

For ∫ab(x−a)m(b−x)n dx\int_a^b (x-a)^m(b-x)^n\,dx: substitute x=a+(b−a)tx=a+(b-a)t, t∈[0,1]t\in[0,1]: =(b−a)m+n+1∫01tm(1−t)n dt=(b−a)m+n+1B(m+1,n+1).= (b-a)^{m+n+1}\int_0^1 t^m(1-t)^n\,dt = (b-a)^{m+n+1}B(m+1,n+1).

Worked Example 1

2016 Paper 1, 2016-P1-Q1c (10 marks)

Evaluate I=∫01 ⁣xlog⁡ ⁣(1x)3 dxI=\displaystyle\int_0^1\!\sqrt[3]{x\log\!\left(\frac{1}{x}\right)}\,dx.

Rewrite: I=∫01x1/3(−log⁡x)1/3 dxI = \int_0^1 x^{1/3}(-\log x)^{1/3}\,dx (since log⁡(1/x)=−log⁡x≥0\log(1/x)=-\log x\ge0 on (0,1)(0,1)).

Substitute x=e−tx=e^{-t}: I=∫0∞e−t/3⋅t1/3⋅e−t dt=∫0∞t1/3e−4t/3 dt.I = \int_0^\infty e^{-t/3}\cdot t^{1/3}\cdot e^{-t}\,dt = \int_0^\infty t^{1/3}e^{-4t/3}\,dt. Using ∫0∞ts−1e−pt dt=Γ(s)/ps\int_0^\infty t^{s-1}e^{-pt}\,dt=\Gamma(s)/p^s with s=4/3s=4/3 and p=4/3p=4/3: I=Γ(4/3)(4/3)4/3=(34)4/3Γ ⁣(43)≈0.608.I = \frac{\Gamma(4/3)}{(4/3)^{4/3}} = \left(\frac{3}{4}\right)^{4/3}\Gamma\!\left(\frac{4}{3}\right) \approx 0.608. I=(34)4/3Γ ⁣ ⁣(43).\boxed{I = \left(\frac{3}{4}\right)^{4/3}\Gamma\!\!\left(\frac{4}{3}\right).}

The most common error: forgetting the (4/3)4/3(4/3)^{4/3} denominator (writing Γ(4/3)\Gamma(4/3) alone is wrong because the formula is Γ(s)/ps\Gamma(s)/p^s, not Γ(s)\Gamma(s)).

Worked Example 2

2021 Paper 1, 2021-P1-Q3a-iii (8 marks)

Express ∫ab(x−a)m(b−x)n dx\displaystyle\int_a^b(x-a)^m(b-x)^n\,dx in terms of the Beta function.

Substitute x=a+(b−a)tx=a+(b-a)t, dx=(b−a) dtdx=(b-a)\,dt. Then x−a=(b−a)tx-a=(b-a)t and b−x=(b−a)(1−t)b-x=(b-a)(1-t): ∫ab=(b−a)mtm⋅(b−a)n(1−t)n⋅(b−a) dt∣t∈[0,1]=(b−a)m+n+1∫01tm(1−t)n dt.\int_a^b = (b-a)^m t^m\cdot(b-a)^n(1-t)^n\cdot(b-a)\,dt\bigg|_{t\in[0,1]} = (b-a)^{m+n+1}\int_0^1 t^m(1-t)^n\,dt. ∫ab(x−a)m(b−x)n dx=(b−a)m+n+1B(m+1,n+1).\boxed{\int_a^b(x-a)^m(b-x)^n\,dx = (b-a)^{m+n+1}B(m+1,n+1).}

Special case a=0,b=1a=0,b=1: recovers the standard Beta integral B(m+1,n+1)B(m+1,n+1) directly.

Common Traps

integration-by-parts (1 question(s); 2020)

Evaluate a definite integral using integration by parts

Recognition Cues

The integrand contains an inverse-trig function (e.g., arctan⁡\arctan) or logarithm that resists direct integration. Integration by parts with u=u= the transcendental and dv=dxdv=dx (so v=xv=x) is the standard entry. The boundary term at the singular endpoint may be a 0⋅(finite)0\cdot(\text{finite}) limit.

Solution Template

  1. Set u=u= (transcendental), dv=dxdv=dx, compute v=xv=x and dudu.
  2. I=[xu]ab−∫abx duI = [xu]_a^b - \int_a^b x\,du; evaluate boundary terms with limits.
  3. Reduce ∫x du\int x\,du to a standard form (partial fractions, completing the square, arctan⁡\arctan integral).

Worked Example

2020 Paper 1, 2020-P1-Q2a (15 marks)

Evaluate ∫01tan⁡−1 ⁣ ⁣(1−1x)dx\displaystyle\int_0^1\tan^{-1}\!\!\left(1-\frac{1}{x}\right)dx.

IBP: u=arctan⁡ ⁣(1−1x)u=\arctan\!\bigl(1-\tfrac1x\bigr), dv=dxdv=dx, v=xv=x: du=11+(1−1/x)2⋅1x2 dx=dx2x2−2x+1.du = \frac{1}{1+(1-1/x)^2}\cdot\frac{1}{x^2}\,dx = \frac{dx}{2x^2-2x+1}. Boundary terms: at x=1x=1: arctan⁡(0)=0\arctan(0)=0; at x→0+x\to0^+: arctan⁡(−∞)=−π/2\arctan(-\infty)=-\pi/2, so product →0\to 0.

I=0−∫01x dx2x2−2x+1.I = 0 - \int_0^1\frac{x\,dx}{2x^2-2x+1}.

Split x=14(4x−2)+12x=\frac{1}{4}(4x-2)+\frac{1}{2} to expose the logarithm-derivative and arctan parts:

∫01x dx2x2−2x+1=14[ln⁡(2x2−2x+1)]01⏟=0+12∫01dx2x2−2x+1.\int_0^1\frac{x\,dx}{2x^2-2x+1} = \frac{1}{4}\underbrace{\Big[\ln(2x^2-2x+1)\Big]_0^1}_{=0} + \frac{1}{2}\int_0^1\frac{dx}{2x^2-2x+1}.

Complete the square: 2x2−2x+1=2(x−12)2+122x^2-2x+1 = 2(x-\tfrac12)^2+\tfrac12, so 12x2−2x+1=2(2x−1)2+1\frac{1}{2x^2-2x+1}=\frac{2}{(2x-1)^2+1}.

12∫012 dx(2x−1)2+1=u=2x−112∫−11duu2+1=12⋅π2=π4.\frac{1}{2}\int_0^1\frac{2\,dx}{(2x-1)^2+1}\stackrel{u=2x-1}{=}\frac{1}{2}\int_{-1}^1\frac{du}{u^2+1}=\frac{1}{2}\cdot\frac{\pi}{2}=\frac{\pi}{4}. I=−π4.\boxed{I = -\frac{\pi}{4}.}

The integral is negative because 1−1/x<01-1/x < 0 for x∈(0,1)x\in(0,1), so the arctan is negative throughout.

Common Traps

integral-functional-equation (1 question(s); 2021)

Differentiate an integral identity to recover f(x)f(x)

Recognition Cues

The question gives an equation of the form ∫0xf(t) dt=(expression in x)\int_0^x f(t)\,dt = (\text{expression in }x) and asks for ff at a point (usually f(1)f(1)). The trick is immediate: differentiate both sides w.r.t. xx using FTC.

Solution Template

  1. Differentiate both sides w.r.t. xx: LHS becomes f(x)f(x) (FTC); RHS becomes the derivative of the expression.
  2. Solve for f(x)f(x) pointwise (often from f(x)(1+h(x))=1f(x)(1+h(x))=1 or similar).
  3. Evaluate at the requested xx.

Worked Example

2021 Paper 1, 2021-P1-Q3a-ii (5 marks)

If ∫0xf(t) dt=x+∫x1tf(t) dt\displaystyle\int_0^x f(t)\,dt = x + \int_x^1 tf(t)\,dt, find f(1)f(1).

Differentiate both sides w.r.t. xx: f(x)=1−xf(x).f(x) = 1 - xf(x). (LHS: FTC gives f(x)f(x). RHS: d/dx[∫x1tf(t) dt]=−xf(x)d/dx[\int_x^1 tf(t)\,dt] = -xf(x) — FTC with xx as the lower limit gives a minus sign.)

Solve: f(x)(1+x)=1  ⇒  f(x)=11+x.f(x)(1+x)=1 \;\Rightarrow\; f(x)=\frac{1}{1+x}. f(1)=12.\boxed{f(1) = \frac{1}{2}.}

Common Traps

Marks-Aware Writing

For a 10-mark Q1 integral: Identify the archetype in the first 30 seconds. State the technique (e.g., “King’s property with x→π/2−xx\to\pi/2-x”), show the key algebraic step that makes both expressions add to something simple, then write the answer. All three parts earn marks; a bare answer earns at most 3.

For a 15-mark integral (Q2/Q3): Expect two or three distinct stages (e.g., IBP followed by partial fractions). Lay out each stage with a label and sub-result. Boundary terms at singular endpoints need a limit argument, not just substitution.

Naming the technique (“substituting x=e−tx=e^{-t},” “using the reflection x→a+b−xx\to a+b-x,” “differentiating both sides”) signals to the examiner that you understand what you are doing — this is worth 1–2 marks beyond the computation.

Practice Set

Ready to drill what you just read?

Daily Practice turns these patterns into one adaptive set a day — practised daily until they're automatic, free for everyone.

See Daily Practice →

This chapter is part of the Maths Coverage Map — 14 years, mapped. Get the take-away PDF free.