1 minute read

Here, I am trying to determine all integers \(n > 1\) such that \(\frac{2^n+1}{n^2}\) is an integer.

First Part

Assume \(a \neq 2\) is the smallest prime that divides \(n\).

Then, since \(a \vert 2^n+1\), \(a \vert 2^{2n}-1\).

Also, we know that \(a \vert 2^{a-1}-1\) (Fermat’s Little Theorem)

Therefore, \(a \vert 2^{\text{gcd}(2n, \space a-1)}-1\).

It can be seen that \(\text{gcd}(2n, \space a-1)=2*\text{gcd}(n, \space a-1)=2*1=2\), hence \(a=3\). (\(a-1\) and \(n\) are coprime and \(a\) is the smallest prime that divides \(n\))

Therefore, we can express \(n=(3^x)(y)\).

Second Part

To proceed, we shall state the Lifting Exponent Lemma:

For a prime \(p\), if \(p \mid x+y\), neither \(x\) nor \(y\) is divisible by \(p\), and \(n\) is an odd positive integer, then

\[v_p(a^n+b^n)=v_p(n)+v_p(a+b).\]

In our case,

\[v_3(2^n+1)=v_3(n)+v_3(3) = x+1\]

Since \(2^n+1>n^2\), \(v_3(2^n+1)>v_3(n^2)\). Hence \(x+1>2x\). Hence \(x=1\).

Third Part

Now let \(q\) be the smallest prime factor of \(y\).

Now,

\[\begin{align} q|2^{\text{gcd}(2n,q-1)}-1 &= 2^{2*\text{gcd}(n,q-1)}-1 \\ &=2^6-1 \\ &=63 \\ &=7*3^2 \end{align}\]

This gives \(q=7\), but it is a contradiction, as a power of two can only be congruent to \(1,2,4\), or \(4\) modulo \(7\).

[\(\text{gcd}(n,q-1)\) can be \(1\) or \(3\), but the case for \(1\) is rejected also.]

Therefore \(n=3\) is the only possible solution.

Credits

This question is from International Mathematical Olympiad, 1990, Q3. The solution is inspired by this website and this PDF.

Exercises for the reader

  • Why is \(a\) not divisible by \(2\)?
  • Why if \(a \vert 2^n+1\), then \(a \vert 2^{2n}-1\).
  • Why is the conclusion \(a \vert 2^{\text{gcd}(2n, \space a-1)}-1\) true?
  • State the Lifting Exponent Lemma and list all the cases.
  • Show that \(2^n+1>n^2\).
  • Show that \(\text{gcd}(n,q-1)\) can be \(1\) or \(3\).
  • Why can’t we simply state \(n=3^k\)?
Tag Styling Example

Updated: