Is
$$\sum\limits_{n=1}^\infty \frac{i^n}{n} $$
convergent?
Im confused as to how to solve this question, I've been trying to use ratio test but that doesn't seem to be helping.
Is
$$\sum\limits_{n=1}^\infty \frac{i^n}{n} $$
convergent?
Im confused as to how to solve this question, I've been trying to use ratio test but that doesn't seem to be helping.
Thew ratio and root tests are for series of positive terms. The series is not absolutely convergent. You can use Dirichlet's test to prove that it is (conditionally) convergent.
You can do this somewhat directly:
Look at $\frac{1}{4n} + \frac{i}{4n+1} + \frac{i^2}{4n+2} + \frac{i^3}{4n+3}.$ Since $i^4 = 1$, we know we can chunk up the series in this way.
So, put all of these together as one fraction: you get $\frac{(16+16i)n^2 + (16+8i)n + 3}{4n(2n+1)(4n+1)(4n+3)}$.
But this is a sum in which the terms are order $\frac{1}{n^2}$, which is certainly convergent, by eg, the comparison test applied to the absolute value of this.
This shows us that the partial sums $S_{4n}$ converge, but then the entire series must converge: Consider $|S_m - S_n|$ as two partial sums. Then, we can find the two closest $4k$ and $4j$ to $m$ and $n$, $|S_m - S_n| \leq |S_m - S_{4k}| + |S_{4k} - S_{4j}| + |S_{4j} - S_n|$. We showed that the middle term tends to zero, but then the first and last are of order $\frac{1}{n}$ and thus converge.
Moreover, this method applies for any rational multiple of $\pi$ argument.
The series $$A=\sum_{n\geqslant1}\frac1{4n-2}-\frac1{4n}=\sum_{n\geqslant1}\frac1{2n(4n-2)}$$ and $$B=\sum_{n\geqslant1}\frac1{4n-3}-\frac1{4n-1}=\sum_{n\geqslant1}\frac2{(4n-3)(4n-1)}$$ both converge, by comparison with the converging Riemann series $\sum\limits_{n\geqslant1}\frac1{n^2}$, and the series in the question is $-A+\mathrm i B$, hence it converges as well.
The above is a hands-on approach. One should be aware though, that for every complex number $z\ne1$ such that $|z|=1$, the series $$\sum_{n\geqslant1}\frac{z^n}n,$$ converges, and master at least one argument to show that it does, see this previous question for example.
The partial sum of the series to $2k$ terms is $$\frac i1-\frac12-\frac i3+\frac14+\frac i5-\frac16-\frac i7+\frac18+\cdots +(-1)^k\frac1{2k}$$ which can be rewritten as $$\Bigl(-\frac12+\frac14-\frac16+\frac18+\cdots+(-1)^k\frac1{2k}\Bigr) +i\Bigl(\frac11-\frac13+\frac15-\frac17+\cdots+(-1)^{k-1}\frac1{2k-1}\Bigr)\ .$$ Both the series are closely related to the alternating harmonic series whose limits are familiar from real calculus. The expression just given tends to $$-\frac{\ln2}{2}+i\frac\pi4$$ as $k\to\infty$. You can easily confirm that the partial sum to an odd number of terms gives the same answer, which is therefore the sum of the series.