On prime factors of terms of binary recurrence sequences

C.L.Stewart

DOI: https://doi.org/10.48550/arXiv.2206.01275

2022-06-03

Abstract:We give estimates from below for the greatest prime factor of the n-th term of a binary recurrence sequence.

Number Theory

What problem does this paper attempt to address?

The problem that this paper attempts to solve is about the growth of the largest prime factor of the terms in binary recurrence sequences. Specifically, the author aims to prove that for some specific types of binary recurrence sequences, the largest prime factor \(P(u_n)\) of the \(n\)-th term increases rapidly as \(n\) grows.

### Specific problem description

1. **Background and existing results**:
- In 1934, Mahler [14] proved that if \(u_n\) is the \(n\)-th term of a non - degenerate binary recurrence sequence, then the largest prime factor of \(u_n\) tends to infinity as \(n\) increases.
- In 1967, Schinzel [18] gave an effective lower bound estimate, showing that there exists a computable positive number \(C_0\) such that:
\[
P(u_n)>C_0n^{c_1}(\log n)^{c_2}
\]
where \((c_1, c_2)=\begin{cases}
(1 / 84,7 / 12)&\text{if }\alpha\text{ and }\beta\text{ are integers}\\
(1 / 133,7 / 19)&\text{otherwise}
\end{cases}\)

2. **Improved results**:
- Stewart [21], Yu and Hung [25] and others further improved these results. In particular, Stewart in his 2013 study [23] proved that there exists a computable positive number \(C\) such that when \(n\) exceeds \(C\):
\[
P(u_n)>n^{1/2}\exp\left(\frac{\log n}{104\log\log n}\right)
\]

3. **Special case of Lucas sequences**:
- For Lucas sequences \(t_n\), Stewart [22] proved that when \(n\) exceeds a computable positive number:
\[
P(t_n)>n\exp\left(\frac{\log n}{104\log\log n}\right)
\]

4. **Main contributions of this paper**:
- This paper proves the following theorems:
- **Theorem 1**: Let \((u_n)_{n = 0}^{\infty}\) be a non - degenerate binary recurrence sequence, and \(a/b\) and \(\alpha/\beta\) be multiplicatively related. There exists a computable positive number \(C\) such that when \(n\) exceeds \(C\):
\[
P(u_n)>n\exp\left(\frac{\log n}{104\log\log n}\right)
\]
- **Theorem 2**: For all non - degenerate binary recurrence sequences \((u_n)_{n=0}^{\infty}\), except possibly for a set of positive integers with asymptotic density zero, we have:
\[
P(u_n)>n\exp\left(\frac{\log n}{104\log\log n}\right)
\]

### Summary

The core problem of this paper is to explore the growth rate of the largest prime factor of the terms in binary recurrence sequences. Through a series of complex mathematical tools and methods (such as the p - adic form of the linear form in logarithms estimate, primitive divisors in algebraic number theory, etc.), it is proved that these largest prime factors grow very fast as \(n\) grows. This not only deepens...