Abstract:We establish a weak form of Ennola's conjecture. We achieve this by showing that two main assumptions Louboutin made in his previous work hold true. These assumptions are about Laurent polynomials over the rationals, and we prove them by using Newton identities.
What problem does this paper attempt to address?
### The problems the paper attempts to solve
The paper aims to prove a weak form of the Ennola conjecture. Specifically, the author achieves this goal by verifying two main hypotheses proposed by Louboutin in his previous work. These two hypotheses are related to Laurent polynomials over the rational numbers, and the author uses Newton's identities to prove these hypotheses.
#### Background and problem statement
In algebraic number theory, an algebraic number $\epsilon$ is called an exceptional unit if both $\epsilon$ and $\epsilon - 1$ are units. For example, for an integer $l\geq3$, consider the non - Galois totally real cubic number field $\mathbb{Q}(\epsilon_l)$, where the minimal polynomial of $\epsilon_l$ is:
\[
X^3+(l - 1)X^2 - lX - 1\in\mathbb{Z}[X]
\]
It can be observed that both $\epsilon_l$ and $\epsilon_l - 1$ are units, that is, $\epsilon_l$ is an exceptional unit.
According to Dirichlet's unit theorem, the rank of the unit group $U_l$ of the ring of integers of $\mathbb{Q}(\epsilon_l)$ is 2. Therefore, it is natural to ask: Is $\{\epsilon_l,\epsilon_l - 1\}$ a pair of fundamental units? That is, $U_l=\langle - 1,\epsilon_l,\epsilon_l - 1\rangle$?
This is an unsolved problem. Ennola conjectured in [Enn91] that for the maximal order $O$ of $\mathbb{Q}(\epsilon_l)$, $\{\epsilon_l,\epsilon_l - 1\}$ is a pair of fundamental units. We call this conjecture the Ennola conjecture. He proved that this conjecture is correct for $3\leq l\leq500$, and the unit index $j_l:=[U_l:\langle - 1,\epsilon_l,\epsilon_l - 1\rangle]$ is always coprime to 2, 3 and 5. He also proved that if $[O:\mathbb{Z}[\epsilon_l]]\leq l/3$, then $\{\epsilon_l,\epsilon_l - 1\}$ is a pair of fundamental units.
To prove that the Ennola conjecture holds is equivalent to proving that for all $l\geq3$, $j_l = 1$. Louboutin proved in [Lou17] that for $l\geq3$, the unit index $j_l$ is coprime to 19!, and for $3\leq l\leq5\times10^7$, $j_l = 1$. He also proved that if the ABC conjecture is assumed to hold, then except for a finite number of $l$, the Ennola conjecture holds, that is, for large enough $l$, $j_l = 1$. Another important result is that, in [Lou21], it was conditionally proved that for any given integer $N\geq2$, there exists a large enough $l$ such that $\gcd(j_l,N!) = 1$.
Louboutin derived a weak form of the Ennola conjecture from Conjecture 12 and Conjecture 20 in [Lou21]. Although Louboutin verified the validity of these conjectures for a finite number of cases by calculation in [Lou21], Conjecture 12 and Conjecture 20 are still unsolved.
In this paper, the author provides proofs of these two conjectures, thereby establishing the following theorem: