Welcome to MathOverflow

Question

0

Hello there, I have a problem and would like to know do anyone have an elementary proof of it and it goes like this:

Show that $ a^n + 1\neq m^2$ when $a=4,7,10$ for every $n$ and $m$ ($n$ and $m$ are natural numbers).

EDIT: I succeded in finding an elementary proof that $ a^n + 1\neq m^2$ when $a$ is of the form $a=3k+4$ , $k\geq0$ and I did it using properties of the "digital root" function

http://en.wikipedia.org/wiki/Digital_root

First, suppose that $a^n=m^2-1=(m-1)(m+1)$

We have the following cases:

1) $dr(m-1)=1$ imply $dr(m+1)=3$ which taken together imply $dr((m-1)(m+1))=3$

2) $dr(m-1)=2$ imply $dr(m+1)=4$ which taken together imply $dr((m-1)(m+1))=8$

3) $dr(m-1)=3$ imply $dr(m+1)=5$ which taken together imply $dr((m-1)(m+1))=6$

4) $dr(m-1)=4$ imply $dr(m+1)=6$ which taken together imply $dr((m-1)(m+1))=6$

5) $dr(m-1)=5$ imply $dr(m+1)=7$ which taken together imply $dr((m-1)(m+1))=8$

6) $dr(m-1)=6$ imply $dr(m+1)=8$ which taken together imply $dr((m-1)(m+1))=3$

7) $dr(m-1)=7$ imply $dr(m+1)=9$ which taken together imply $dr((m-1)(m+1))=9$

8) $dr(m-1)=8$ imply $dr(m+1)=1$ which taken together imply $dr((m-1)(m+1))=8$

9) $dr(m-1)=9$ imply $dr(m+1)=2$ which taken together imply $dr((m-1)(m+1))=9$

So we have $dr((m-1)(m+1))\in\lbrace 3,6,8,9 \rbrace$ but since $dr(3k+4)\in\lbrace 1,4,7 \rbrace$ implies that $dr((3k+4)^n)\in\lbrace 1,4,7 \rbrace$ the result follows because $\lbrace 3,6,8,9 \rbrace \cap \lbrace 1,4,7 \rbrace = \emptyset$

edited Jan 28 at 20:19

asked Jan 28 at 15:07

Antisha
17●4

1

For $n>1$, this is a special case of Catalan's conjecture (proved by Mihailescu). These special cases may well have elementary proofs, search the literature. The case $n=1$ you can do by inspection. – Felipe Voloch Jan 28 at 15:31

At least 4 and 7 seem simple, factor $m^2-1=(m+1)(m-1)$ and observe that not both can be a power of 2 (except for $m=3$) or a power of 7, resp. And, what is the motivation for this? Voting to close. – quid Jan 28 at 15:40

Thank you, I forgot that Catalan´s conjecture have status of a theorem, and sure there must be some elementary method for proving these three cases. – Antisha Jan 28 at 15:47

Motivation is to find elementary proof when $a$ is of the form $a=3k+4$ when k is natural number, or zero. – Antisha Jan 28 at 15:51

1

Your proof can be simplified greatly. If $a=3k+4$, then $a\equiv 1\pmod{3}$, hence $a^n+1\equiv 2\pmod{3}$. On the other hand, it is easy to see that $m^2\equiv 2\pmod{3}$ is impossible. – GH Jan 28 at 20:39

show 9 more comments

GH · Accepted Answer · 2013-01-28 18:12:32Z UTC

3

I recommend this survey about the solution of Catalan's conjecture. We learn from here that the special case you are considering, namely $x^p-y^q=1$ for $p=2$, was solved by Chao Ko in 1964. In 1976 Chein published a simpler proof, see here.

answered Jan 28 at 18:12

GH
14.9k●1●31●53

I commented above (and then mistakenly deleted) that the specific result desired by the poster was realized through mod 3 considerations. I suspect that the poster is even more interested in when elementary considerations provide quick answers. For example, knowing when 2 is not a qth power mod a covers the cases mentioned and many more. Gerhard "Looks For Really Simple Answers" Paseman, 2013.01.28 – Gerhard Paseman Jan 28 at 18:57

1

As a general aside: there is also a technically non-elementary (but in certain ways simpler) proof by Mignotte "A new proof of Ko Chao's theorem" (2004) – quid Jan 28 at 19:11

1

Also, while I presently do not see how to improve on Chein's argument, I still think that something simpler and possibly less complete is wanted. Gerhard "AskMe About System Design" Paseman, 2013.01.28 – Gerhard Paseman Jan 28 at 19:13

I found an elementary proof that $a^n + 1\neq m^2$ when $a$ is of the form $a=3k+4$. In it I used the properties of this function: mathworld.wolfram.com/DigitalRoot.html Thank you all for your valuable information. – Antisha Jan 28 at 19:23

Perhaps it would be useful to include your elementary proof in your original post. – GH Jan 28 at 19:27

show 4 more comments

Numbers of a certain form not expressible as squares

Remember to vote up questions/answers you find interesting or helpful (requires 15 reputation points)

1 Answer

You can accept an answer to one of your own questions by clicking the check mark next to it. This awards 15 reputation points to the person who answered and 2 reputation points to you.