Totally real bi-quadratic fields with large Pólya groups (Q2138612)

For the ring of integers \(O_K\) of a number field \(K\), the free \(O_K\)-module Int\((O_K)\) of polynomials on \(K\) which preserve \(O_K\) is a well-studied object. If it admits a basis \(\{f_i\}\) of polynomials with deg \(f_i = i\) for each \(i \geq 0\), one calls \(K\) a Pólya field. The existence of such a basis (called a regular basis) is intimately related to the class group in the following manner. For each prime number \(q\), the ideal \(\pi_q\) of \(O_K\) is defined to be the product of all prime ideals whose norm is \(q\) (the empty product is simply \(O_K\)). The ideal classes of such ideals generate the so-called Pólya group \(Po(K)\) of \(K\) which is a subgroup of the class group \(Cl(K)\). The number field \(K\) is Pólya if \(Po(K)\) is trivial; hence \(Po(K)\) measures the deviation from Int\((O_K)\) having a regular basis. Already, from classical work of Hilbert, it is known that for a quadratic field \(K\), the Pólya group \(Po(K)\) is a \(2\)-group. In the case of biquadratic fields, the \(2\)-class group has been extensively studied over a long period (see for example, works of Ezra Brown and Charles Parry from 1977--78). If one can construct fields for which \(Po(K)\) has arbitrarily large \(2\)-rank, then this a fortiori shows the \(2\)-rank of the class group can be arbitrarily large. In this paper, the authors explicitly (modulo Dirichlet's theorem on primes in arithmetic progressions) produce infinitely many totally real biquadratic fields for which the Pólya group is isomorphic to \(n\) copies of \(\mathbb{Z}/2 \mathbb{Z}\) for any given \(n \geq 1\). The study of \(Po(K)\) is carried with the help of Galois cohomology of the unit group \(O_K^{\ast}\).