3
$\begingroup$

I have to prove $$\sum_{d|n}\frac{\mu(d)}{\varphi(d)^{2}}>C$$ for all positive odd $n$ where $C>0$ is a constant. (You can choose the constant $C$.)
I got $$\sum_{d|n}\frac{\mu(d)}{\varphi(d)^{2}}=\prod_{p|n}\left(1-\frac{p}{\varphi(p)^{2}}\right)\geq\prod_{p}\left(1-\frac{p}{(p-1)^{2}}\right),$$ but I cannot prove the infinite product converges to a positive real number.
Can anyone help me?

*Edited
The first question was just a conjecture I made to prove another question, but it seems to be false.
The original question was to prove that $$\sum_{q=1}^{\infty}\frac{\mu(q)c_{q}(n)}{\varphi(q)^{3}}>C$$ for all positive odd $n$ where $C>0$ is a constant you can choose. ($c_q(n)$ is Ramanujan's Sum.)
I got $$\sum_{q=1}^{\infty}\frac{\mu(q)c_{q}(n)}{\varphi(q)^{3}}=\sum_{d|n}\frac{\mu(d)}{\varphi(d)^{2}}\sum_{(n,e)=1}\frac{\mu(e)^2}{\varphi(d)^{3}},$$ but cannot solve the problem at all.
You can use another idea if you want.

$\endgroup$
  • $\begingroup$ "I cannot prove the infinite product converges to a positive real number" because it doesn't. $\lim\limits_{k\to\infty}\prod\limits_{p \leqslant k}\left(1-\frac{p}{(p-1)^2}\right) = 0$, since $\sum\frac1p = \infty$. You need to exploit the finiteness of the product, I don't yet see how, though. $\endgroup$ – Daniel Fischer Jul 7 '13 at 14:45
  • $\begingroup$ @DanielFischer My question was just a conjecture I made to prove another problem. I'm editing my question and posting the original problem. $\endgroup$ – Flan Jul 7 '13 at 14:55
2
$\begingroup$

I need new spectacles ;)

You wrote:

$$\sum_{d|n}\frac{\mu(d)}{\varphi(d)^{2}}=\prod_{p|n}\left(1-\frac{p}{\varphi(p)^{2}}\right)\geq\prod_{p}\left(1-\frac{p}{(p-1)^{2}}\right),$$

and "but I cannot prove the infinite product converges to a positive real number", and I focused on the infinite product you wrote, that converges to $0$.

But the equality in that is wrong. If $d$ is squarefree, then $\dfrac{\mu(d)}{\varphi(d)^2} = \prod\limits_{p\mid d}\frac{(-1)}{\varphi(p)^2}$ and hence

$$\sum_{d|n}\frac{\mu(d)}{\varphi(d)^{2}}=\prod_{p|n}\left(1-\frac{1}{\varphi(p)^{2}}\right)\geq\prod_{p > 2}\left(1-\frac{1}{(p-1)^{2}}\right),$$

and the latter product is easily seen to be positive since $\sum \frac{1}{n^2}$ converges.


Another way to obtain the existence of a positive $C$ such that

$$\sum_{q=1}^{\infty}\frac{\mu(q)c_{q}(n)}{\varphi(q)^{3}}>C$$

for all positive odd $n$, and a different, maybe easier to handle(?) expression for the sum is to use Kluyvert's expression for the Ramanujan sum, $c_q(n) = \sum\limits_{d\mid (q,n)} \mu\left(\frac{q}{d}\right)d$, and set

$$P = \prod_p \left(1 + \frac{1}{\varphi(p)^3}\right)$$

for brevity. Then one can rearrange (all sums are absolutely convergent because a rather crude estimate gives $\varphi(n) \geqslant C\frac{n}{\log n}$)

$$\begin{align} \sum_{q=1}^{\infty}\frac{\mu(q)c_{q}(n)}{\varphi(q)^{3}} &= \sum_{q=1}^\infty \frac{\mu(q)}{\varphi(q)^3}\left(\sum_{d\mid (q,n)} \mu(q/d)d\right)\\ &= \sum_{d\mid n}\sum_{k=1}^\infty \frac{\mu(kd)\mu(k)d}{\varphi(kd)^3}\\ &= \sum_{d\mid n} \sum_{\substack{k=1\\(k,d) = (1)}}^{\infty} \frac{\mu(d)d\mu(k)^2}{\varphi(d)^3\varphi(k)^3}\\ &= \sum_{d\mid n} \frac{\mu(d)d}{\varphi(d)^3} \sum_{\substack{k=1\\(k,d) = (1)}}^{\infty}\frac{\mu(k)^2}{\varphi(k)^3}\\ &= \sum_{d\mid n} \frac{\mu(d)d}{\varphi(d)^3} \prod_{p \not\mid d}\left(1 + \frac{1}{\varphi(p)^3}\right)\\ &= P \cdot \sum_{d \mid n} \frac{\mu(d)d}{\varphi(d)^3 \prod\limits_{p\mid d}\left(1+\frac{1}{\varphi(p)^3}\right)}\\ &= P \cdot \sum_{d \mid n} \frac{\mu(d)d}{\prod\limits_{p\mid d}\bigl(\varphi(p)^3+1\bigr)}\\ &= P\cdot \prod_{p\mid n} \left(1 - \frac{p}{\varphi(p)^3+1}\right). \end{align}$$

The last product again is bounded below (for odd $n$) by a positive number because for $p \geqslant 3$ we have $\frac{p}{\varphi(p)^3+1} \leqslant \frac{3}{p^2}$ and hence

$$\prod_{p \geqslant 3} \left(1 - \frac{p}{\varphi(p)^3+1}\right) > 0.$$

| cite | improve this answer | |
$\endgroup$

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.