The Family $X_{qp^s}$ for Distinct Primes $q,p$ in Connection with Carmichael's Totient Conjecture

Abstract

Shannon, Shiue, Saito and the author \cite{SSHS2025} studied the sets $X_k=\{n\in\mathbb{N}:\varphi(n)=k\}$ in connection with Carmichael's totient conjecture, and proved that for an odd prime $p\neq3$ and $s\in\{1,2,3,4\}$, $X_{2p^s}$ is empty whenever $2p^s+1$ is not prime (with $s=1$ fully characterized in both directions), conjecturing that this persists for every $s\ge1$. We resolve that conjecture in full, obtaining a complete characterization of $X_{2p^s}$ for every $s\ge1$ and every odd prime $p\neq3$, together with a parity dichotomy: $X_{2p^s}=\varnothing$ unconditionally whenever $s$ is even, while for odd $s$ both outcomes genuinely occur, determined exactly by the primality of $2p^s+1$. We then close the one remaining gap at $p=3$, where the proof technique degenerates, with a separate, unconditional characterization of $X_{2\cdot3^s}$. Building on these results, we consider the natural generalization $X_{qp^s}$, where $q\neq p$ are distinct primes and $s\ge1$. We show that this generalization splits into three cases of very different character. If $q$ and $p$ are both odd, the problem is trivial: $qp^s$ is odd, so $X_{qp^s}=\varnothing$ by elementary parity, with no case-analysis needed. If $q=2$, the problem is exactly the family $X_{2p^s}$ resolved above. If $p=2$, giving the ``mirror'' family $X_{2^sq}$ for an odd prime $q$, the problem is genuinely different. Beyond $s=1$, we give a complete resolution for $s=2$, recovering, with an explicit proof and solution sets, the qualitative dichotomy of \cite[Thm.~2.3.1]{SSHS2025}, but the problem remains open for $s\ge3$.

Abstract: arXiv descriptive metadata (CC0). Thank you to arXiv for use of its open access interoperability. This service was not reviewed or approved by, nor does it necessarily express or reflect the policies or opinions of, arXiv.