Prime values of a sparse polynomial sequence (Q2073279)

It was shown by the reviewer [Acta Math. 186, No. 1, 1--84 (2001; Zbl 1007.11055)] that there are infinitely many primes of the shape \(x^3+2y^3\). The goal of the present paper is to refine this result by showing that one can restrict to pairs \(x,y\) in which \(y\) is much smaller that \(x\). Specifically it is shown that there is a constant \(B>0\) such that, if \(0<\gamma<5/67\) is fixed, \(Y=X^{1-\gamma}\), and \(\eta=(\log X)^{-B}\), then the number of pairs \((x,y)\in(X,(1+\eta)X]\times(Y,(1+\eta)Y]\) for which \(x^3+2y^3\) is prime will be of exact order \(\eta^2 XY/\log X\). It is noted that the constant \(5/67\) is capable of improvement. This result gives the sparsest polynomial sequence currently known to contain infinitely many primes. The proof is based on the reviewer's argument, but with two main differences. Firstly, while the reviewer's approach gives an asymptotic formula, the present paper gives only a sieve lower bound, by estimating certain terms via a trivial lower bound. The second new feature is more technical. Because \(Y\) is small compared to \(X\), one needs a Hoheisel type result for some rather general multiplicative functions, summed over a small 3-dimensional region.