4

We define the irrationality measure of a number $x\in\mathbb{R}$ as the supremum of the $r\in\mathbb{R}$ such that exists $c\in \mathbb{R}$ verifying

$$|x-\frac{p}{q}|\geq \frac{1}{q^r}$$ I was trying to prove that the irrationality measure of the Euler's constant $e$ is 2 using simple results of continued fractions (such as the expression of the denominators), and the simple continued fraction of $e$,

$$[2;1,2,1,1,4,1,1,6,1,1,8...]$$ Thank you very much for your help.

  • See https://mathworld.wolfram.com/IrrationalityMeasure.html and its references. $\mu(e)$=2 is proven in Borwein, J. M. and Borwein, P. B. "Irrationality Measures." ยง11.3 in Pi & the AGM: A Study in Analytic Number Theory and Computational Complexity. New York: Wiley, pp. 362-386, 1987. Not easy stuff. โ€“ Mittens Nov 30 '23 at 18:06

1 Answers1

5

Let $\alpha\in\Bbb R\setminus\Bbb Q$. It can be shown that (iirc this paper contains an error but the conclusions are all correct): $$\begin{align}\mu(\alpha)=1+\limsup_{n\to\infty}&\frac{\ln(q_{n+1})}{\ln(q_n)}=2+\limsup_{n\to\infty}\frac{\ln(a_n)}{\ln(q_{n-1})}\\0\le\limsup_{n\to\infty}\frac{\ln a_{n+1}}{\ln(F_n\prod_{j=1}^na_j)}&\le\mu(\alpha)-2\le\limsup_{n\to\infty}\frac{\ln a_{n+1}}{\ln(F_n+\prod_{j=1}^na_j)}\end{align}$$Where $\mu$ is the irrationality measure, $F_n$ is the $n$th Fibonnaci number and $\alpha=[a_0;a_1,a_2,\cdots]$ as an infinite simple continued fraction (SCF) (this representation is unique) with convergents $p_n/q_n$, indexing these so that the recurrence $q_{n+1}=a_{n+1}q_n+q_{n-1}$ holds.

As a corollary, any number $\alpha$ whose SCF satisfies $a_n=\exp(o(n))$, in particular any $\alpha$ whose SCF has coefficients which grow in polynomial time - as is the case for $e$ - must have irrationality measure $2$.

I haven't looked at this for a while, so I'm sorry to say I can't quite remember if - for the specific case of numbers whose SCF coefficients grow slowly - there is a more elementary way to deduce $\mu(\alpha)=2$. However this paper is not super technical, it only requires a few lemmas of the theory of SCFs (which I believe could be found in Hardy's book on number theory). I can review my notes if you wish.

FShrike
  • 48,385
  • Thank you very much for your answer, the article you mentioned seems of much interest itself. I was looking for a simple proof for $e$ that could be presented in a short course on number theory. I think the proof was easy when the $a_n$ satisfied $a_n\leq cn+d$ for certain $c,d$ positives, but I cannot remember the argument. โ€“ jantomico May 25 '23 at 22:45
  • 1
    @joseasd You're welcome. I might try to look for an easier proof in the morning, from my old notes. โ€“ FShrike May 25 '23 at 22:46