Tests edge cases, boundary conditions, and common misconceptions for integration.
UT-1: Integration by Parts — Cyclic Integral Trap
Question:
(a) Evaluate ∫e2xsinxdx.
(b) A student sets up I=∫e2xsinxdx, applies integration by parts twice, and gets I=(something)−I, then concludes I=0. Explain the error.
[Difficulty: hard. Tests cyclic integration by parts and the algebraic step of collecting the I terms.]
Solution:
(a) Let I=∫e2xsinxdx.
First application: u=e2x, dv=sinxdx. Then du=2e2xdx, v=−cosx.
I=−e2xcosx+∫2e2xcosxdx
Second application on ∫e2xcosxdx: u=e2x, dv=cosxdx. Then du=2e2xdx, v=sinx.
∫e2xcosxdx=e2xsinx−∫2e2xsinxdx=e2xsinx−I
Substitute back:
I=−e2xcosx+2e2xsinx−2I
3I=e2x(2sinx−cosx)
I=3e2x(2sinx−cosx)+C
(b) The student's error is that when they got I=(something)−I, they incorrectly concluded I=0. The correct step is to add I to both sides to get 2I=(something), then divide by 2 (or in this case 3). The cyclic nature of the integral means I appears on both sides, but this does not mean I=0 — it means I can be solved for algebraically.
UT-2: Partial Fractions with Irreducible Quadratic
Question:
(a) Express (x+1)(x2+1)3x−1 in partial fractions.
(b) Hence evaluate ∫(x+1)(x2+1)3x−1dx.
(c) A student attempts to split the denominator as x+1A+x2+1B and gets A=3, B=−1. Show that this is incorrect.
[Difficulty: hard. Tests partial fraction decomposition with an irreducible quadratic factor and the correct form of the decomposition.]
Solution:
(a) Since x2+1 is irreducible (discriminant =−4<0), the correct form is:
(x+1)(x2+1)3x−1=x+1A+x2+1Bx+C
3x−1=A(x2+1)+(Bx+C)(x+1)
3x−1=Ax2+A+Bx2+Bx+Cx+C
3x−1=(A+B)x2+(B+C)x+(A+C)
Comparing coefficients:
x2: A+B=0 \quad (i)
x: B+C=3 \quad (ii)
constant: A+C=−1 \quad (iii)
From (i): B=−A. From (iii): C=−1−A.
Substitute into (ii): −A+(−1−A)=3⟹−2A−1=3⟹−2A=4⟹A=−2.
Then B=2 and C=−1−(−2)=1.
(x+1)(x2+1)3x−1=x+1−2+x2+12x+1
(b)
∫(x+1)(x2+1)3x−1dx=∫x+1−2dx+∫x2+12x+1dx
=−2ln∣x+1∣+∫x2+12xdx+∫x2+11dx
=−2ln∣x+1∣+ln(x2+1)+arctanx+C
(c) If the student uses x+1A+x2+1B, then:
3x−1=A(x2+1)+B(x+1)=Ax2+A+Bx+B
Comparing coefficients: A=0 (from x2 term), B=3 (from x term), and A+B=3=−1 (from constant term), which is a contradiction (3=−1). This shows the form x2+1B is incorrect — the numerator must be linear: x2+1Bx+C.
UT-3: Improper Integral — Convergence with Parameter
Question:
(a) Determine whether ∫1∞xp1dx converges or diverges, stating the condition on p.
(b) Evaluate ∫01x1dx and explain why it converges despite the integrand being unbounded.
(c) A student claims that since x1→∞ as x→0+, the integral must diverge. Explain the error.
[Difficulty: hard. Tests improper integrals at both endpoints and the p-test for convergence.]
Solution:
(a)
∫1∞xp1dx=limb→∞∫1bx−pdx
Case p=1:
=limb→∞[1−px1−p]1b=limb→∞(1−pb1−p−1−p1)
If p>1: 1−p<0, so b1−p→0 and the integral converges to p−11.
If p<1: 1−p>0, so b1−p→∞ and the integral diverges.
The integral converges to 2. Although the integrand is unbounded at x=0, the area under the curve is finite because the singularity is integrable (the exponent −21>−1).
(c) The student confuses the behaviour of the integrand with the behaviour of the integral. An unbounded integrand does not necessarily produce a divergent integral. The key question is whether the area accumulates to a finite value. For x1 near x=0, the function grows, but slowly enough that the total area remains bounded. The p-test shows that ∫01x−pdx converges when p<1.