I am working on an integration by parts problem that, compared to the student solutions manual, my answer is pretty close. Could someone please point out where I went wrong?

Find $\int e^{2\theta} \cdot \sin{3\theta} \ d\theta$

$u_1 = \sin{3\theta}$

$du_1 = \frac{1}{3}\cos{3\theta} \ d\theta$

$v_1 = \frac{1}{2} e^{2\theta}$

$dv_1 = e^{2\theta} \ d\theta$

$\underbrace{\sin{3\theta}}_{u_1} \cdot \underbrace{\frac{1}{2} e^{2\theta}}_{v_1} - \int\underbrace{\frac{1}{2} e^{2\theta}}_{v_1} \cdot \underbrace{\frac{1}{3} \cos{3\theta} \ d\theta}_{du_1}$

$\frac{1}{2}\sin{3\theta} \cdot e^{2\theta} - \frac{1}{6} \int \cos{3\theta} \cdot e^{2\theta} \ d\theta$

Doing integration by parts again...

$u_2 = \cos{3\theta}$

$du_2 = -\frac{1}{3} \sin{3\theta} \ d\theta$

$v_2 = \frac{1}{2} e^{2\theta}$

$dv_2 = e^{2\theta} \ d\theta$

$\underbrace{\sin{3\theta}}_{u_1} \cdot \underbrace{\frac{1}{2} e^{2\theta}}_{v_1} - \frac{1}{6}\left(\underbrace{\cos{3\theta}}_{u_2} \cdot \underbrace{\frac{1}{2} e^{2\theta}}_{v_2} - \int \underbrace{\frac{1}{2} e^{2\theta}}_{v_2} \cdot \underbrace{-\frac{1}{3} \sin{3\theta} \ d\theta}_{du_2}\right)$

$\frac{1}{2}\sin{3\theta} \cdot e^{2\theta} - \frac{1}{6}\left(\frac{1}{2}\cos{3\theta} \cdot e^{2\theta} + \frac{1}{6} \int \sin{3\theta} \cdot e^{2\theta} \ d\theta\right)$

$\frac{1}{2}\sin{3\theta} \cdot e^{2\theta} - \frac{1}{12}\cos{3\theta} \cdot e^{2\theta} - \frac{1}{36} \int \sin{3\theta} \cdot e^{2\theta} \ d\theta$

$\frac{37}{36} \int \sin{3\theta} \cdot e^{2\theta} \ d\theta = \frac{1}{2}\sin{3\theta} \cdot e^{2\theta} - \frac{1}{12}\cos{3\theta} \cdot e^{2\theta}$

$\int \sin{3\theta} \cdot e^{2\theta} \ d\theta = \begin{equation} \boxed{\frac{18}{37}\sin{3\theta} \cdot e^{2\theta} - \frac{1}{37}\cos{3\theta} \cdot e^{2\theta}} \end{equation}$

However, the boxed answer is incorrect. The answer should read:

$\frac{1}{13} e^{2\theta} \left(2\sin{3\theta} - 3\cos{3\theta}\right)$

  • 2
    $\begingroup$ I know the intended purpose of this exercise is to practice integration by part, but there's a trick for this particular primitive. Write $\int e^{2 \theta} \sin(3 \theta) \, d \theta = \text{Im}\left(\int e^{(2+3i) \theta} \, d \theta \right)$. And compute $\int e^{(2+3i) \theta} \, d \theta = \frac{1}{2+3i} e^{(2+3i) \theta}$, whose imaginary part is $\frac{1}{13} e^{2 \theta} (2 \sin(3 \theta) - 3 \cos(3\theta))$. $\endgroup$ – Joel Cohen Mar 3 '12 at 18:22

Ah ha! I got it!

$du = 3 \cos{3\theta} \ d\theta$ not $\frac{1}{3} \cos{3\theta} \ d\theta$

For some reason I must have been thinking of integrating the dv rather than differentiating.

| cite | improve this answer | |
  • 2
    $\begingroup$ You made the same mistake in the second integration by parts. Nice write up of the question and answer by the way. $\endgroup$ – David Mitra Mar 3 '12 at 18:19

One could also use the right tool, which is the following fact:

For every nonzero complex number $z$, a primitive of $\theta\mapsto\mathrm e^{z\theta}$ is $\theta\mapsto z^{-1}\mathrm e^{z\theta}$.

Here, one is looking at the imaginary part of $\mathrm e^{z\theta}$ with $z=2+3\mathrm i$, hence the answer is the imaginary part of $$ \theta\mapsto z^{-1}\mathrm e^{z\theta}=|z|^{-2}\bar z\mathrm e^{z\theta}=|z|^{-2}(2-3\mathrm i)\mathrm e^{2\theta}(\cos(3\theta)+\mathrm i\sin(3\theta)). $$ Since $|z|^2=2^2+3^2=13$ and the imaginary part of $(2-3\mathrm i)(\cos(3\theta)+\mathrm i\sin(3\theta))$ is $2\sin(3\theta)-3\cos(3\theta)$, an answer is $\theta\mapsto \frac1{13}\mathrm e^{2\theta}(2\sin(3\theta)-3\cos(3\theta))$.

| cite | improve this answer | |
  • $\begingroup$ An easy way is to use the method of undetermined coefficients, guess that the answer has the form $y=e^{2\theta}(A\sin(3\theta)+B\cos(3\theta))$, take the derivative of $y$, and figure out what $A$ and $B$ are. $\endgroup$ – Stefan Smith Jun 8 '12 at 23:04
  • $\begingroup$ I wouldn't say that using Euler's formula is the right tool. What the OP did above is a valid method, and it is good to have more than one method. $\endgroup$ – Baby Dragon Jun 19 '13 at 14:09
  • $\begingroup$ @BabyDragon In principle, sure, except that the question itself shows that the other method is more complicated and more error-prone (as well as, if I may add, less conceptual). $\endgroup$ – Did Jun 19 '13 at 15:32

$$\int \mathrm{e}^{ax}\sin(bx)dx =\frac{\mathrm{e}^{ax}}{\mathrm{a^2+b^2}}[a\sin(bx)-b\cos(bx)]+c$$

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.