Series sum of sin log and cos log

274 Views Asked by At

I was solving a series based question and I am now totally confused on how to proceed further to find the analytical values of $a$ and $b$ (where $a,b\in \mathbb{R}$) such that:

$$ \sum_{k=2}^{\infty}\frac{(-1)^k\cos(b \ln(k))}{k^a} = 1 $$

and

$$ \sum_{k=2}^{\infty}\frac{(-1)^k\sin(b \ln(k))}{k^a} = 0 $$

where $\ln x$ is the natural logarithm.

any help would be greatly appreciated. Thanks in advance :)

1

There are 1 best solutions below

0
On

We begin with the equations in the posted question

$$\sum_{n=2}^\infty \frac{(-1)^n\sin(b\log(n))}{n^a}=0\tag1$$

and

$$\sum_{n=2}^\infty \frac{(-1)^n\cos(b\log(n))}{n^a}=1\tag2$$

Multiplying $(1)$ by $i$ and subtracting from $(2)$, we find that the pair of equations $(1)$ and $(2)$ are equivalent to the equation

$$\sum_{n=1}^\infty \frac{(-1)^{n-1}}{n^z}=0\tag3$$

where $z=a+ib$.

The series on the left-hand side of $(3)$ is a representation of the Dirichlet Eta Function, $\eta(z)$, for $\text{Re}(z)>0$. Therefore, the posted question implicitly asks to find the zeros of the Eta function in the right-half plane.


To proceed, we note the relationship

$$\underbrace{\sum_{n=1}^\infty \frac{(-1)^{n-1}}{n^z}}_{\eta(z), \text{Re}(z)>0}=(1-2^{1-z})\underbrace{\sum_{n=1}^\infty \frac1{n^z}}_{\zeta(z),\text{Re}(z)>1}\tag4$$

where the series on the right-hand side of $(4)$ is a representation of the Riemann Zeta function, $\zeta(z)$, for $\text{Re}(z)>1$.



ASIDE:

  1. I showed in THIS ANSWER and THIS ANSWER that the series representation for $\zeta(z)$ in $(4)$ diverges for $\text{Re}(z)\le 1$.
  2. We can extend the definition of $\zeta(z)$ to the right-half plane by rewriting $(4)$ as

$$\zeta(z)=(1-2^{1-z})^{-1}\sum_{n=1}^\infty \frac{(-1)^{n-1}}{n^z}\tag5$$

for $\text{Re}(z)>0$. I showed in THIS ANSWER that from $(5)$ we can write

$$\zeta(z)=\frac1{z-1}+\gamma +O(z-1)$$

which explicitly shows that $\zeta(z)$ has a simple pole at $z=1$ with residue $1$.



It is easy to see from $(5)$ that zeros of $1-2^{1-z}$ are also zeros of $\eta(z)$ unless they are also poles of $\zeta(z)$. Therefore, we find that the zeros of $\eta(z)$, as represented by the series in $(3)$, include the points $z_k$ where

$$z_k=1+i\frac{2\pi k}{\log(2)}$$

for $k\in \mathbb{Z}\setminus \{0\}$.

Inasmuch as the zeros of the Zeta function lie in the region $0<\text{Re}(z)<1$, the Dirichlet Eta function has no zeros on the half plane $\text{Re}(z)>1$. It had been conjectured (The Riemann Hypothesis, 1859) that all of the zeros of $\zeta(z)$ are located at even negative integers (the trivial zeros) and at complex numbers with real part equal to $1/2$. This conjecture remains unproven at the time of this post. If this conjecture is true, then all other zeros of $\eta(z)$ as represented by the series in $(3)$ are located at points $z=\frac12+ib$.



ZEROS OF THE ETA FUNCTION

The Eta function can be analytically continued to the left-half plane through the relationship in $(5)$ (i.e., $\eta(z)=(1-2^{1-z})\zeta(z)$) with the Zeta function and the analytic continuation of the Riemann Zeta function through the functional equation

$$\zeta(z)=2^z\pi^{z-1}\sin\left(\frac{\pi z}{2}\right)\Gamma(1-z)\zeta(1-z)\tag6$$

Note from $(6)$ that $\zeta(z)$ has zeros at the negative even integers. These are the so-called trivial zeros of $\zeta(z)$. So, we see that the zeros of $\eta(z)$ include the points

$$z_k=\begin{cases}1+i\frac{2k\pi}{\log(2)}&,k\in \mathbb{Z}\setminus \{0\}\\\\ 2k&,k\in \mathbb{Z_{<0}} \end{cases}$$

Again, if the Riemann hypothesis is correct, all other zeros of $\eta(z)$ are located at points $z=\frac12+ib$.