6
$\begingroup$

Is there a non zero polynomial $R \in \mathbb{Z}[X,Y]$ such that there exists an infinite number of pair $(p,q)$ with $p$ and $q$ primes, $p \neq q$ and $R(p,q)=0$ ?

I know the curve must be of genus $0$ (Faltings-Mordell).

My question is related to Polynomial equations in $n$ and $\phi(n)$ that has been solved.

  • 7
    Conjecturally, $R(x, y) = x - y + 2$ or $R(x, y) = x - 2y - 1$...2011-12-14
  • 2
    For the record, the conjectures mentioned in Qiaochu Yuan's comment are about [twin primes](http://en.wikipedia.org/wiki/Twin_prime) and [safe primes](http://en.wikipedia.org/wiki/Safe_prime) or [Sophie Germain primes](http://en.wikipedia.org/wiki/Sophie_Germain_prime).2011-12-14
  • 2
    Conjecturally, there are infinitely many primes $p$ such that $p^2-2$ is prime, so $R(x,y)=x^2-y-2$ will do. More generally, it is widely believed that if $f$ and $g$ are irreducible polynomials with integer coefficients, and if there is no $d\gt1$ such that $d$ divides $f(n)g(n)$ for all $n$, then there are infinitely many $n$ such that $f(n)$ and $g(n)$ are both prime. But the proofs are way out of reach. See Schinzel's Hypothesis H.2011-12-14
  • 0
    @GerryMyerson: Thanks.2011-12-15
  • 1
    Less trivially, (p-q-2)(p-q-4)(p-q-6)(p-q-8)(p-q-10)(p-q-12)(p-q-14)(p-q-16)(p-q-18)(p-q-20) is zero infinitely often under Elliott-Halberstam.2012-02-09

1 Answers 1