Indivisibility of class numbers and Iwasawa \(\lambda\)-invariants of real quadratic fields (Q2731688)

One of the famous unsolved problems on class numbers of real quadratic fields is Gauss' problem, which asks if there are infinitely many real quadratic fields \(k\) whose class number \(h_k\) is 1. The triviality of 2-parts of \(h_k\) is known by genus theory. According to the ``Cohen-Lenstra heuristic'' [\textit{H. Cohen} and \textit{H. W. Lenstra}, Lect. Notes Math. 1068, 33-62 (1984; Zbl 0558.12002)], we predict that there are infinitely many real quadratic fields \(k\) with \(p\nmid h_k\) for a given odd prime number \(p\). It was proved that the set of such fields has a positive density for \(p=3\) by \textit{H. Davenport} and \textit{H. Heilbronn} [Proc. R. Soc. Lond., Ser. A 322, 405-420 (1971; Zbl 0212.08101)], also for \(p=3\) with \(p\) nonsplitting in \(k\) by \textit{J. Nakagawa} and \textit{K. Horie} [Proc. Am. Math. Soc. 104, 20-24 (1988; Zbl 0663.14023)], for \(5\leq p\leq 4999\) with \(p\) ramified in \(k\) by \textit{K. Ono} [Compos. Math. 119, 1-11 (1999; Zbl 1002.11080)]. NEWLINENEWLINENEWLINEOn the other hand, we have another problem, so-called Greenberg's conjecture, which states that the Iwasawa \(\lambda\)-invariants \(\lambda_p\) of totally real number fields vanish for any prime number \(p\). This is true for \(\mathbb{Q}\), but we do not know even the case of quadratic fields. So, it is natural to ask if there exist infinitely many real quadratic fields with \(\lambda_p=0\) (and some additional condition) for a given odd prime number \(p\). It was proved that the set of such fields has a positive density for \(p=3\) which does not split in \(k\) by J. Nakagawa and K. Horie (cited above), for \(p=3\) which splits in \(k\) by the reviewer [\textit{H. Taya}, Proc. Am. Math. Soc. 128, 1285-1292 (2000; Zbl 0958.11069)], for \(5\leq p\leq 4999\) which is ramified in \(k\) by K. Ono (cited above). NEWLINENEWLINENEWLINEFor a real quadratic field \(k\), let \(d_k\) be its discriminant and \(R_p(k)\) the \(p\)-adic regulator of \(k\). In this excellent paper under review, the author shows by refining Ono's idea (loc. cit.), which is done by constructing a half-integral weight Eisenstein series whose Fourier coefficients are given by generalized Bernoulli numbers attached to quadratic characters, that the density of real quadratic fields \(k\) with \(0< d_k< X\), \(p\nmid h_k\), \(v_p(R_p(k))= 1\) and \(p\) splitting (resp. \(p\) inert) is at least a constant depending on \(p\) times \(\sqrt{X}/\log X\) for any prime number \(p\geq 5\) (resp. \(p\equiv 3\pmod 4\), \(p\geq 5\)). NEWLINENEWLINENEWLINEIn the inertia case, Iwasawa's classical theorem [\textit{K. Iwasawa}, Abh. Math. Semin. Univ. Hamb. 20, 257-258 (1956; Zbl 0074.03002)] says that \(p\nmid h_k\) implies \(\lambda_p=0\). Further, in the splitting case, a result of \textit{T. Fukuda} and \textit{K. Komatsu} [J. Number Theory 23, 238-242 (1986; Zbl 0593.12003)] says that \(p\nmid h_k\) and \(v_p(R_p(k))= 1\) imply \(\lambda_p=0\). Therefore, the author also mentions that the density of real quadratic fields \(k\) with \(0< d_k< X\), \(\lambda_p=0\), \(v_p(R_p(k))= 1\) and \(p\) splitting (resp. \(p\) inert) has the same lower bounds as above for any prime number \(p\geq 5\) (resp. \(p\equiv 3\pmod 4\), \(p\geq 5\)).