3
$\begingroup$

Hello guys I want to know if my proof is valid, and whether it is coherent and where I can improve it.

Prove that for all natural numbers, $a,b, n$, where $a>b$, and $n>1$ is a not a prime, $a^n-b^n$ is not a prime number.

My attempt:

Since $n$ is a non-prime greater than $1$, we have $n=xy$ where $1<x,y<n. $We know that $$a^n-b^n=a^{xy}-b^{xy}=(a^x-b^x)(a^{x(y-1)}+b^xa^{x(y-2)}+...+b^{x(y-2)}a^x+b^{x(y-1)})$$ Hence $(a^x-b^x)$ will always divide $a^n-b^n$. We merely need to prove that $a^n-b^n>a^x-b^x>1$. That $a^n-b^n>a^x-b^x$ is obvious since $a^{x(y-1)}+b^xa^{x(y-2)}+...+b^{x(y-2)}a^x+b^{x(y-1)}$ is a positive integer greater than $1$ ($y>1$). Furthermore, we can write $a^x-b^x$ as $(a-b)(a^{x-1}+ba^{x-2}+...+ab^{x-2}+b^{x-1})$. Since $x>1$, $a^{x-1}+ba^{x-2}+...+ab^{x-2}+b^{x-1}$ is a positive integer greater than $1$. Therefore, we have $a^x-b^x=(a-b)(a^{x-1}+ba^{x-2}+...+ab^{x-2}+b^{x-1})>a-b \geq1$. Hence $a^x-b^x>1$. This concludes the proof.

$\endgroup$
4
  • 1
    $\begingroup$ Perfectly valid. This is a generalization of the already generalized mersenne numbers. Like in the Mersenne case, a nonprime gives a nontrivial factor , as you have shown (note that the second factor is obviously larger than $1$, so equality is impossible). $\endgroup$
    – Peter
    Commented Jan 17, 2022 at 12:32
  • 2
    $\begingroup$ Please identify what part of the proof you have doubts about, and explain why. $\endgroup$ Commented Jan 17, 2022 at 12:34
  • $\begingroup$ @BillDubuque I added it now $\endgroup$ Commented Jan 17, 2022 at 12:47
  • 3
    $\begingroup$ Your edit does not help. What specific line of the proof do you have doubts about, and why? The site is not intended as a general open-ended proof checker. You need to pose a specific mathematical question. $\endgroup$ Commented Jan 17, 2022 at 12:56

1 Answer 1

1
$\begingroup$

I don't see anything wrong with your proof. Perhaps I could mention some shortcuts or alternatives that you might find useful.

Let $d$ be a non trivial divisor of $n$, then we may write $n = dk$ for some integer $k > 1$. From $a^d \equiv b^d \pmod{a^d - b^d}$, we obtain $a^n \equiv b^n \pmod{a^d - b^d}$ by raising each side to the $k$-th power. The latter congruence relation is of course equivalent to saying that $a^d - b^d$ is a divisor of $a^n - b^n$.

$x \mapsto x^n - x^d = x^d(x^{n-d} - 1): [1, \infty) \to \mathbb{R}$ is a strictly increasing function whenever $n > d > 0$, since it is the product of strictly increasing functions mapping only to non-negative values. Here $n$ and $d$ need not be integers. Hence both $a^n - b^n > a^d - b^d$ and $a^d - b^d > a - b$ follow from rewriting them as $a^n - a^d > b^n - b^d$ and $a^d - a > b^d - b$.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .