How I can solve the equation $y^3=8x^6+2x^3y-y^2$ in integers?
I made the substition $x^3=z$ and got the equation $8z^2+2yz-y^3-y^2=0$. So I decided to apply the general formula for quadratic equations and thus I got $$z=\frac{-2y\pm \sqrt{32y^3+36y^2}}{16}=\frac{-y\pm y\sqrt{8y+9}}{8}.$$
So in order to have $z\in \mathbb{Z}$, $8y+9$ must be a perfect square, let's say $8y+9=k^2$, with $k\in \mathbb{Z}$. From this we get $8y=(k+3)(k-3)$, but what could I do next? Also, is there another approach, like an appropriate factorization, to solve this equation? Thanks in advance.
Well, after a couple of days thinking I think the diophantine equation can be solved in a different way. First of all, asumme that $y=0$, so the equation becomes $8x^6=0$, then $x=0$. Now, let's consider $y\neq 0$, we note that the LHS is divisible by $y$ and thus the RHS also has to be divisible by $y$, so we get that $y\mid 8x^6$. The idea is to prove that $\gcd(x,y)=1$. We can write the equation in the form $$(2x^3+y^2+y)(4x^3-y)=4x^3y^2.$$
Let's suppose $\gcd(x,y)=d>1$, so there exists a prime number $p$ such that $p\mid x$ and $p\mid y$. Set $v_{p}(x)=\alpha>0$ and $v_{p}(y)=\beta>0$. First, let's assume that $p$ is odd, then $v_{p}(4x^3y^2)=3\alpha+2\beta$.
On the other hand, $v_{p}((2x^3+y^2+y)(4x^3-y))=v_{p}(2x^3+y^2+y)+v_{p}(4x^3-y)$, but $v_{p}(2x^3+y^2+y)=\min\{v_{p}(2x^3), v_{p}(y^2), v_{p}(y)\}=3\alpha$ or $\beta$ if $3\alpha\le \beta$ or if $3\alpha>\beta$, respectively. Also $v_{p}(4x^3-y)=\min\{v_{p}(4x^3), v_{p}(y)\}=3\alpha$ or $\beta$ if $3\alpha\le \beta$ or if $3\alpha>\beta$, respectively. In summary, we deduce that $$v_{p}((2x^3+y^2+y)(4x^3-y))= \begin{cases} 6\alpha &\text{if } 3\alpha\le \beta \\ 2\beta &\text{if } 3\alpha>\beta \end{cases} $$
So we conclude that $6\alpha=3\alpha+2\beta$ or $2\beta=3\alpha+2\beta$. In the first case we get $\beta\le 0$, and in the second case we get $\alpha=0$, contradiction in both cases. Now, Let's suppose $p=2$. In this case we get $v_{p}(4x^3y^2)=3\alpha+2\beta+2$, and after an analogous reasoning we deduce that $$v_{p}((2x^3+y^2+y)(4x^3-y))= \begin{cases} 6\alpha+3 &\text{if } 3\alpha+1< \beta \\ 3\alpha+\beta+1 &\text{if }\, 3\alpha+1=\beta \\ 2\beta &\text{if } 3\alpha+1>\beta \end{cases} $$
So we have either $6\alpha+3=3\alpha+2\beta+2$ or $3\alpha+\beta+1=3\alpha+2\beta+2$ or $2\beta=3\alpha+2\beta+2$. In any case we arrive to a contradiction. In conclusion such $p$ doesn't exists and therefore $\gcd(x,y)=1$. Using this and the fact that $y\mid 8x^6$ we get that $y\mid 8$. Therefore, the possible values for $y$ are $\{\pm1, \pm2, \pm4, \pm8\}$.
A routine check using the possible values of $y$ gives us solutions only for $y=-1$ and $y=2$. In the first case we get $x=0$, and in the second case we get $x=1$. Hence, all the integer solutions for the equation are: $$(x,y)=(0,0), (0,-1), (1,2).$$