To Prove:
$\displaystyle \lim_{m\rightarrow\infty} \int_0^z \dfrac{t^m}{1+t}dt = 0 : z\in\mathbb{C}, |z|=1.$
My Real Question: Is the following analysis valid?
My Research:
I know little about complex analysis. I (superficially) browsed some online texts and
(given $z=x+iy : x,y\in\mathbb{R}$), decided to construe
$\;\displaystyle \int_0^z \dfrac{t^m}{1+t}dt \;$ as $\;\displaystyle\left(\int_0^x \dfrac{t^m}{1+t}dt\right) \;+ i\times \left(\int_0^y \dfrac{t^m}{1+t}dt\right).$
My Work:
Since $|z|=1,\;|x|\leq 1\;$ and $\;|y|\leq 1.\;$ Further, when $\;m>0,$
$\displaystyle \left\lvert\int_0^x \dfrac{t^m}{1+t}dt\right\rvert < \left\lvert\int_0^x t^m dt\right\rvert \;= \left\lvert\;\frac{x^{m+1}}{m+1}\right\rvert$ and
$\displaystyle \left\lvert\int_0^y \dfrac{t^m}{1+t}dt\right\rvert < \left\lvert\int_0^y t^m dt\right\rvert \;= \left\lvert\;\frac{y^{m+1}}{m+1}\right\rvert.$
Since $\;|x|\leq 1\;$ and $\;|y|\leq 1,$ $\;\left\lvert\dfrac{x^{m+1}}{m+1}\right\rvert\;$ and $\;\left\lvert\dfrac{y^{m+1}}{m+1}\right\rvert\;$ both go to zero as $\;m\rightarrow\infty.$
Therefore $\;\displaystyle \left\lvert \int_0^z \frac{t^m}{1+t}dt\right\rvert\;$ goes to zero as $\;m\rightarrow\infty.$
Correction:
The above analysis assumes (for example) that as $t$ ranges from $0$ to $x,$ $\dfrac{1}{1+t} < 1.$ This is true when $x>0,$ but false when $x<0.$
However, if $-1 < x < 0,\;$ then as $t$ ranges from $0$ to $x,$ $\dfrac{1}{1+t} < \dfrac{1}{1+x}.$
This provides a remedy for the analysis except in the special cases of $x=-1$ or $y=-1.$ This forum query was originally intended as the final piece of analysis needed for a 2nd query: Request help verifying the Taylor series expansion of $\text{ln}(1 + e^{2ix})$.
In the 2nd query, $\ln(1+z)$ is disallowed when $z=-1,$ but allowed when $z=-i.$ Thus, even if my correction is valid, it still remains to prove (in the 2nd query) that $\displaystyle \ln(1-i) = \sum_{k=0}^{\infty} \frac{(-1)^k}{k+1} (-i)^{k+1}.$
This means that for marty cohen's answer in the 2nd query to apply, it must be shown that the $\displaystyle \lim_{m\rightarrow\infty} \int_0^{y} \frac{t^m}{1+t}dt = 0,$ specifically when $y=-1.$ I would love feedback here.
2nd Correction:
I have augmented the 2nd query: Request help verifying the Taylor series expansion of $\text{ln}(1 + e^{2ix})$, to manually resolve the case of $z = -i,\;$ which is the only pending case unresolved by my first Correction.
I don't know if $\int_0^z \dfrac{t^m}{1+t}dt\;$ exists when $z=-i,\;$ and would still like feedback on this. However, with respect to the 2nd query, the issue is moot.
For $|z|=1,z\ne -1,$ we have
$$\tag 1 \left |\int_0^z \dfrac{t^m}{1+t}dt\right|=\left | \int_0^1 \frac{(sz)^m z}{1+sz}\, ds\right | \le \int_0^1 \frac{s^m }{|1+sz|}\, ds.$$
Now for $s\in [0,1],$ $|1+sz| = |sz-(-1)|.$ The last expression is the distance from $sz$ to $-1,$ which is bounded below by the distance from the line segment $[0,z]$ to $-1.$ This last distance is a positive number I'll call $d_z.$ It follows that the right side of $(1)$ is bounded above by
$$\frac{\int_0^1s^m\,ds }{d_z}= \frac{1}{d_z(m+1)}\to 0.$$
This shows the left side of $(1)$ $\to 0,$ which is the desired result.