It is well-known that the Euler $\phi$-function is multiplicative: that is, for co-prime positive integers $m,n$ we have $\phi(mn) = \phi(m)\phi(n)$. Thus it is defined by its values on prime powers. We know that $\phi(p^k) = p^{k-1} (p-1)$ for all primes $p$.

What about the multiplicative function $\psi$ defined on the primes by $\psi(p^k) = p^{k-1} (p+1)$? Does it have a name? If so, what's known about it?

For example, can one evaluate $\sum_{n \leq X} \psi(n)$?


$\psi$ is the multiplicative convolution of $\mu^2$ and the identity function, hence its Dirichlet series is $$\sum_{n=1}^\infty\frac{\psi(n)}{n^s}=\frac{\zeta(s)\zeta(s-1)}{\zeta(2s)},\qquad\Re(s)>2.$$ This implies by Perron's formula and standard bounds that $$\sum_{n \leq X} \psi(n)\sim\frac{\zeta(2)}{2\zeta(4)}X^2=\frac{15}{2\pi^2}X^2.$$

  • 2
    $\begingroup$ I will just add this function appears in Exercise 11 in Chapter 3 in Apostol's book. The last part of this exercise is the asymptotic formula $\sum_{n\le x} \varphi_1(n)=\frac{\zeta(2)}{2\zeta(4)}x^2+O(x\log x)$. (Apostol denotes the function by $\varphi_1$.) $\endgroup$ – Martin Sleziak Oct 18 '18 at 13:58
  • 2
    $\begingroup$ In the field of automorphic forms, this function is often denotes $\nu(n)$ (for example, in Iwaniec-Luo-Sarnak's seminal paper on low-lying zeroes of $L$-functions of holomorphic cuspidal eigenforms). $\endgroup$ – Peter Humphries Oct 18 '18 at 14:07
  • $\begingroup$ @MartinSleziak: Thanks for your comment. Clearly, one can also prove this statement by elementary means, by making use of the convolution I mentioned. $\endgroup$ – GH from MO Oct 18 '18 at 14:35
  • $\begingroup$ @PeterHumphries: Indeed the function looked familiar! It is also the index of $\Gamma_0(n)$ in $\Gamma_0(1)$ if I am not mistaken. $\endgroup$ – GH from MO Oct 18 '18 at 14:36
  • 1
    $\begingroup$ @GHfromMO, yep, which is why it comes up in the field automorphic forms so often! $\endgroup$ – Peter Humphries Oct 18 '18 at 15:03

This is also called the Dedekind $\psi$ function: $$ \psi(n):=n\prod_{p|n}(1+p^{-1}) $$

See also A001615 and A158523.

  • $\begingroup$ This should be a comment in my opinion, but thanks for it anyways. (Usually answers should be true answers, which really answer the original question.) $\endgroup$ – GH from MO Oct 18 '18 at 14:38
  • 11
    $\begingroup$ @GHfromMO You obviously know more about the etiquette of this site than me, but I do believe I answered the main two questions in the OP: "does this have a name", and "what is known about it". I didn't say anything about the asymptotics, because that was already covered by you, and because that seemed to me as an example question, and not the main one. But I appreciate the feedback anyway. $\endgroup$ – AccidentalFourierTransform Oct 18 '18 at 14:42
  • 2
    $\begingroup$ I agree with you. I forgot about those parts of the question! (BTW it is usually discouraged to ask several questions in one post, exactly to avoid confusions like this.) $\endgroup$ – GH from MO Oct 18 '18 at 14:44
  • 4
    $\begingroup$ I find it amazing that I used the same letter as the generally accepted one! $\endgroup$ – Stanley Yao Xiao Oct 18 '18 at 15:09

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

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