I have the Diophantine equation $$6b^2c^2 + 3c^2 - 36bc - 4b^4 - 4b^2 + 53=0.$$ Numerical calculations suggest this has only one positive integer solution, namely $(b,c)=(2,3)$. Is there a quick way to prove or disprove that?

- 393,674

- 7,889
- 2
- 28
- 70
-
1Did you try replacing $b=x+2, c =y+3$? – markvs Jun 26 '20 at 02:54
-
Interesting tip! I just did that, but don’t exactly see how to take that to the goal line… – Kieren MacMillan Jun 26 '20 at 02:58
-
Your polynomial in $x,y$ can be better looking. It may be reducible or a sum of squares. Anyway, what is it? – markvs Jun 26 '20 at 03:02
-
I’m playing around with primes of the form $u^2+2v^2$, and this popped out at one point in my investigations. – Kieren MacMillan Jun 26 '20 at 03:04
-
Primes representable by a binary quadratic form should be well known. I do not see how it is related to your question. But if you say what polynomial did you get after the substitution it may help. – markvs Jun 26 '20 at 03:09
-
It is not surprising that many people get similar things. The equation is quadratic in $c$. If $b, c$ are positive integers, we have $$c = \frac{54b + \sqrt{(6b^2+3)^3 - 1458}}{18b^2 + 9}.$$ Let $x = 6b^2 + 3$. Let $x^3 - 1458 = y^2$. This is an elliptic curve. – River Li Jun 26 '20 at 04:25
4 Answers
Solving for $c$, we get $$ c=-{\frac {-18\,b \pm \sqrt {24\,{b}^{6}+36\,{b}^{4}+18\,{b}^{2}-159} }{3 (2\,{b}^{2}+1)}}$$ If there is to be an integer solution, we need $24\,{b}^{6}+36\,{b}^{4}+18\,{b}^{2}-159$ to be a square. Taking $s = b^2$, let's look for integer solutions of $t^2 = 24 s^3 + 36 s^2 + 18 s - 159$, which is an elliptic curve. Taking $s = x/24 - 1/2$ and $t = y/24$, this becomes $y^2 = x^3 - 93312$: note that if $s$ and $t$ are integers, then $x = 24 s + 12$ and $y = 24 t$ are also integers. Now $y^2 = x^3 - 93312$ is an elliptic curve in Weierstrass form. According to Sage, its only integer solutions are $x = 108, y = \pm 1080$. This corresponds to $s=4$ and $t = \pm 45$. And indeed $s=4$ gives us $b = \pm 2$; $b=2$ gives us an integer solution with $c=3$, and $b=-2$ gives us an integer solution with $c=-3$. Those are the only integer solutions.

- 448,999
-
Is there no easier way than asking Sage? :)
If nobody comes up with a simpler solution in a day or two, I’ll pop back and award this answer. Thanks!!
– Kieren MacMillan Jun 26 '20 at 04:11 -
3
$6b^2c^2 + 3c^2 - 36bc - 4b^4 - 4b^2 + 53=0$
In order to quadratic equation for $c$ has rational solution, the discriminant must be square number.
$v^2=96b^6+144b^4+72b^2-636$
So, above quartic equation must have rational solutions.
This quartic equation can be transformed to below elliptic curve.
$Y^2=X^3+36X^2+432X-91584$ with $b^2=X/24.$
Using online Magma calculator as follows.
$E$:=EllipticCurve([$0, 36, 0, 432, -91584$])
IntegralPoints($E$)
It says that all integral point is $ [ (96 : -1080 : 1) ]$.
Hence all integral point is $(b,c)=(2,3).$

- 2,346
The equation is quadratic in $c$. If $b, c$ are positive integers, we have $$c = \frac{18b + \sqrt{3[(2b^2+1)^3 - 54]}}{6b^2 + 3}.$$ So, $(2b^2+1)^3 - 54 = 3m^2$ for some positive integer $m$.
Since $3 | 54$ and $3 | 3m^2$, we know that $3 | (2b^2 + 1)$. Let $x = \frac{2b^2+1}{3}$ and $y = \frac{m}{3}$. We have $x^3 - 2 = y^2$. Since $x$ is a positive integer, we know that $y$ is also a positive integer.
The equation $x^3 - 2 = y^2$ is well-known. The only positive integer solutions to $x^3 - 2 = y^2$ are $(x, y) = (3, 5)$. So, $(b, c) = (2, 3)$ is the only positive integer solution.
See:
[1] Theorem 3.4 (page 7), https://kconrad.math.uconn.edu/blurbs/gradnumthy/mordelleqn1.pdf

- 37,323
Hint
$$6b^2c^2+3c^2-36bc-4b^4-4b^2+53=(6b^2c^2-36bc+54)-(4b^4+4b^2+1)+3c^2=6(bc-3)^2-(2b^2+1)^2+3c^2=0\\\implies 6(bc-3)^2+3c^2=(2b^2+1)^2$$ Then $3\mid 2b^2+1$. Now proceed like this by checking modulo primes like $2,3$ and so on.

- 6,054
-
4Well, we know there is an integer solution $b=2, c=3$, so this must work modulo every prime. You can check modulo as many primes as you want, but I don't see how you will rule out other solutions. – Robert Israel Jun 26 '20 at 04:33