MathLabs

Open problem, Arithmetic and number theory, posed 1958

Schinzel's hypothesis H and Bunyakovsky's conjecture

Open

For every finite collection {f1(x),…,fk(x)}\{f_1(x), \dots, f_k(x)\} of nonconstant irreducible polynomials in Z[x]\mathbb{Z}[x] with positive leading coefficients, if there is no prime pp that divides the product f1(n)⋯fk(n)f_1(n)\cdots f_k(n) for every integer nn, then there are infinitely many positive integers nn such that f1(n),…,fk(n)f_1(n), \dots, f_k(n) are simultaneously prime. The case k=1k=1 is Bunyakovsky's conjecture.

Research frontier as of 2026

As of 2026, Bunyakovsky's conjecture is unproved for every single polynomial of degree ≥2\ge 2, and Schinzel's hypothesis H is unproved for every pair of polynomials (k≥2k \ge 2). The closest unconditional results either relax primality to almost-primality — such as Iwaniec's 1978 theorem that n2+1n^2+1 has at most two prime factors infinitely often — or replace the single-variable polynomial with a two-variable polynomial of large density, such as the Friedlander–Iwaniec theorem (a2+b4a^2+b^4, 1998) and the Heath-Brown theorem (a3+2b3a^3+2b^3, 2001). For linear systems {x+h1,…,x+hk}\{x+h_1, \dots, x+h_k\}, the Maynard–Tao sieve (2013) proves that for any mm, if kk is large enough, at least mm of the linear forms are simultaneously prime infinitely often, establishing a weak existential fragment of Dickson's conjecture.

Best known results

  • For k=1k=1 and deg⁡f1=1\deg f_1 = 1, Bunyakovsky's conjecture holds by Dirichlet's theorem on primes in arithmetic progressions (1837).
  • For any irreducible quadratic polynomial f(x)f(x) without fixed prime divisors, f(n)f(n) has at most two prime factors for infinitely many nn (Iwaniec, 1978).
  • For any admissible tuple of kk linear polynomials, at least mm of them are simultaneously prime infinitely often once k≥Ce4mk \ge C e^{4m} (Maynard, Tao, Polymath8b, 2013–2014).

Tools and where they stop

ToolAchievedWhere it stops
Half-dimensional and weighted linear sievesProduce almost-prime values (PrP_r) for single polynomials and systems of polynomials of bounded degree.Cannot cross the parity barrier to guarantee actual primes (P1P_1), and the level of distribution for polynomial sequences of degree d≥2d \ge 2 is too small.
Multidimensional Maynard–Tao sieveFinds clusters of mm simultaneous primes within a larger admissible set of kk polynomials when the polynomials satisfy a Bombieri–Vinogradov-type theorem.Requires kk to be much larger than mm (so it never proves all kk forms are simultaneously prime) and relies on level-of-distribution estimates known mainly for linear polynomials.

Open questions

  • Does there exist any single polynomial f(x)∈Z[x]f(x) \in \mathbb{Z}[x] of degree ≥2\ge 2 that takes infinitely many prime values?
  • Can the Maynard–Tao bounded-cluster theorem be extended unconditionally to admissible families of irreducible quadratic polynomials?

References

  1. Andrzej Schinzel, Wacław Sierpiński (1958). Sur certaines hypothèses concernant les nombres premiers · DOI:10.4064/aa-4-3-185-208
  2. Paul T. Bateman, Roger A. Horn (1962). A heuristic asymptotic formula concerning the distribution of prime numbers · DOI:10.1090/S0025-5718-1962-0148632-7
  3. Henryk Iwaniec (1978). Almost-primes represented by quadratic polynomials · DOI:10.1007/BF01578070