Show that the sequence $\{a_n\}_{n\in \mathbb{N}} = \{x: x=$ the nth decimal digit of Champernowne's constant$\}$ is not periodic.

For those who don't know what Champernowne's constant is, it's the following number:

$C_{10} = 0.123456789101112131415161718192021...$

Well, before someone says that $\{a_n\}_{n\in \mathbb{N}}$ is definitely aperiodic because Champernowne's constant is irrational(and even transcendental), I have to say that I'm well aware of the fact that a number is rational if and only if the sequence of its digits is periodic, but I'm looking for another proof that the sequence of the digits of $C_{10}$ is aperiodic that doesn't use anything but a smart way of showing that $a_{n+k} = a_n$ can't hold for any $k\in \mathbb{N}$.

I've observed that by grouping the terms of a sequence in a clever way we can show that no such $k$ can exist for some sequences. For example I created the following sequence:


The sequence is created this way: First you put one $1$, then two $0$'s, and in the nth step you add either $1$'s or $0$'s n times depending on whether n is odd or even respectively.

It's easy to show this sequence is not periodic, because no matter how large $k$ is you can always find $n$ such that $a_n = 1$ but $a_{n+k}=0$ or $a_n=0$ but $a_{n+k}=1$. That proves the sequence is not periodic and as a corollary it' proved that the real number $0.1001110000111110000001111111\cdots$ is irrational.

After showing that I've been trying to apply this technique of grouping to the sequence in problem without much success so far.

I think if the periodicity of a sequence $\{a_n\}$ is denoted by $\Omega(\{a_n\})$ then if we define a new sequence $\bar{a}_m(n) := \bar{a}_{m,n} = [a_n]_m$ then $\Omega(\{\bar{a}_{m,n}\}) \mid \Omega({a_n})$. This shouldn't be very hard to prove I think.

Now, to show that


is not periodic it's sufficient to prove that the sequence


is not periodic. But I haven't progressed much from here :/


$C_{10}$ contains arbitrarily long sequences of zeroes, but is not zero.

  • $\begingroup$ Doesn't that observation need a proof? $\endgroup$ – user66733 Aug 30 '13 at 17:09
  • 1
    $\begingroup$ @some1: No, it is a proof in itself. A simple and beautiful proof. $\endgroup$ – TonyK Aug 30 '13 at 17:12
  • 3
    $\begingroup$ @some1.new4u At some point in the decimal expansion of Champernowne's constant, the number $10^k$ is used for the digits. $\endgroup$ – Daniel Fischer Aug 30 '13 at 17:12
  • 5
    $\begingroup$ Here's an elaboration. As Daniel said, $C_{10}$ contains arbitrarily long sequences of zeroes, because for every $k$, it contains $10^k$ in its decimal expansion. Now suppose $a_n$ is periodic with period $p$. But we can find $p$ zeroes in a row in the sequence, which means the entire period is just zeroes. Therefore, the entire sequence must be 0. Contradiction! $\endgroup$ – Lopsy Aug 30 '13 at 17:14
  • 2
    $\begingroup$ Ah, now you're spoiling it Lopsy! $\endgroup$ – TonyK Aug 30 '13 at 17:14

Less elegant, but more direct to my taste.

Let's say there is some period. At some point after the period start you'll get a sequence beginning with that period and followed by the first digit of the period plus 1 modulo 10. It's true because any given sequence appears an infinite number of times in Champernowne's constant. Which contradicts the possibility of a period.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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