Regarding the factorization $a^2+3b^2 = cd$.

77 Views Asked by At

Let $a,b,c,d$ be positive integers, with $\gcd(c,d)=1$, such that $$a^2+3b^2=cd.$$ By well-known classical results, we have that $c$ and $d$ are both of the form $u^2+3v^2$.

QUESTION: Is it valid to claim that that there exists at least one representation \begin{align} (c,d) = (p^2+3q^2,\ r^2+3s^2) \end{align} in [not necessarily positive] integers $p,q,r,s$ such that \begin{align} (a,b) &= (pr+3qs,\ ps-qr) &\text{or}&& (a,b) &= (pr-3qs,\ ps+qr), \end{align} or does this commit some sort of unique factorization error?

1

There are 1 best solutions below

0
On BEST ANSWER

yes, this is Theorem 69 on page 91 of Dickson, Introduction to the Theory of Numbers (1929). You have made it easier by demanding $c,d$ coprime. The reason this matters is that Dickson's forms need not be primitive, so $2x^2 + 2 xy + 2 y^2$ might come up

enter image description here