$$2^x-3^y=2021$$ where $x,y$ are non-negative integers. I only found $2^{11}-3^3=2021$.
-
1Look modulo $43,47$ and $3$ and then proceed like here. – Dietrich Burde Mar 26 '21 at 14:49
-
3Is this a contest problem? – Kenta S Mar 26 '21 at 15:01
-
2No, it isn't. I was just thinking about differences between powers of 2 and powers of 3. – spock Mar 26 '21 at 15:07
-
2The equation, also involving $2021$, doesn't look like just thinking about differences between powers of $2$ and powers of $3$. Rather it looks like a recent homework or context question. – Dietrich Burde Mar 26 '21 at 16:25
-
If we allow all perfect powers , not just the powers of $2$ and $3$ , the problem is related to Pillai's conjecture , which is however only solved for difference $1$. – Peter Mar 26 '21 at 16:40
2 Answers
Hmm, three answers with -1 vote. This is interesting.
The fact that we have a solution with $y=3$ suggests that an exclusionary proof must focus on $y\ge 4$ so that the power of $3$ is a multiple of $81$.
So, assume $y\ge 4$ and render
$2^y\equiv 2021\equiv-4\bmod 81.$ (1)
The Euler totient if $81$ is $54$. We have $2^2=4$ and $2^{27}\equiv-1\bmod 81$, so the congruence (1) implies
$x\equiv29\bmod54.$ (2)
This implies that $x$ is two greater than a multiple of $3$ so $2^x\equiv4\bmod 7$. In addition $2021\equiv5\bmod7$, so
$3^y\equiv-1\bmod7,y\equiv3\bmod6.$ (3)
From (3) we find that $y$ has to be a multiple of $3$, so render $y=3z$ from which
$2^x-27^z=2021; x,z\in\mathbb{N}.$ (4)
Now we try modulo ... $109$. Why this number? Experimenting with various other candidates led to the conclusion that the sought contradiction can be realized only if we choose a modulus whose Euler totient is a multiple of $27$, and $109$ is the smallest such prime number.
With $x\equiv29\bmod54$ from (2) and the Euler totient of $109$ being $108$, we have $2^x=\pm2^{29}\equiv\pm23\bmod109$. We also have $2021\equiv59\bmod109$ and thus two possibilities for $27^z$:
$27^z\equiv(23-59)\equiv73\bmod109.$
$27^z\equiv(-23-59)\equiv27\bmod109.$
The first of these results fails, but the second holds when
$z\equiv1\bmod9.$ (5)
Our next prime modulus candidate is $\bmod163$. This has an Euler totient of $162=3×54$, so we have three possible residues for $2^x$ allowed by (2):
$2^x\in\{73,94,159\}\bmod163.$
With $2021\equiv65\bmod163$ we then have:
$27^z\in\{8,29,94\}\bmod163.$
In this case the residues $29$ and $94$ fail whereas $8$ allows $z\equiv23\bmod54$ ($3$ is a primitive root $\bmod163$ so $27^z$ will assume $54$ different residues). But this implies
$z\equiv5\bmod9,$ (6)
whereas above, we found $z\equiv1\bmod9$ (5). The contradiction between (5) and (6) implies that if $y\ge4$, then Eq. (4) cannot hold both $\bmod109$ and $\bmod163$, and without any solution to that equation we can't solve our given equation $2^x-3^y=2021\bmod7$!
We conclude that $2^x-3^y=2021$ has no solutions with $y\ge4$, assuring after trials for smaller $y$ that $(x,y)=(11,3)$ is unique.

- 39,403
-
And what is wrong that causes the downvote? Please offer a way this can be improved. – Oscar Lanzi Mar 29 '21 at 01:44
-
1Good question. I have $3$ downvotes, although the solution by Scott is really nice and elementary. Perhaps the problem is indeed, that this is an ongoing contest and we shouldn't answer. – Dietrich Burde Mar 29 '21 at 08:12
According to Pillai, the equation $2^x-3^y=d$ has at most one solution except for $d=-1,5,13$. Since we have a solution for $d=2021$, it is the only solution. For a proof by elementary methods see the reference below (the corollary to Theorem $4$).
Reference: R. Scott, On the Equations $p^x-b^y=c$ and $a^x+b^y=c^z$, J. Number Theory 44 (2) (1993), 153–165.

- 130,978
-
1
-
Did you have a look at the proof there? It seems to be elementary. – Dietrich Burde Mar 26 '21 at 16:02