Does there exist a natural number $a$ so that: a) $\Big ((a^2-3)^3+1\Big) ^a-1$ is a perfect square? b) $\Big ((a^2-3)^3+1\Big) ^{a+1}-1$ is a perfect square?
Problem
Source: Latvia BW TST 2021 P13
Tags: number theory
Blastoor
29.09.2021 00:21
We can easily see that in the first case $a$ must be odd, in the second - even.
a) $\Big ((a^2-3)^3+1\Big) ^a-1=k^2$
Then $\Big ((a^2-2)((a^2-3)^2-(a^2-3)+1)\Big) ^a=k^2+1$.
Notice that $a^2-2 \equiv 3 \pmod 4$, but by Fermat's Two Squares theorem $k^2+1$ isn't divisible by a prime $3 \pmod 4$, thus no such $a$ exists.
b) We can easily notice that $\Big ((a^2-3)^3+1\Big) ^{a+1}-1 \equiv 3 \pmod 4$, however, $3$ isn't a quadratic residue $\pmod 4$, so no such $a$ exists.
starchan
29.09.2021 04:42
Well Mihailescu theorem basically kills this problem.
Ali_11
06.08.2023 06:10
Blastoor wrote:
We can easily see that in the first case $a$ must be odd, in the second - even.
a) $\Big ((a^2-3)^3+1\Big) ^a-1=k^2$
Then $\Big ((a^2-2)((a^2-3)^2-(a^2-3)+1)\Big) ^a=k^2+1$.
Notice that $a^2-2 \equiv 3 \pmod 4$, but by Fermat's Two Squares theorem $k^2+1$ isn't divisible by a prime $3 \pmod 4$, thus no such $a$ exists.
b) We can easily notice that $\Big ((a^2-3)^3+1\Big) ^{a+1}-1 \equiv 3 \pmod 4$, however, $3$ isn't a quadratic residue $\pmod 4$, so no such $a$ exists.
Why in first case a is odd in other a is even?