The central question of this paper is: given a polynomial with integer coefficients, how often does it produce "squarefree" values? A squarefree integer is one not divisible by any perfect square other than 1 (so 6 is squarefree, but 12 is not because it is divisible by 4). This is a classical and difficult problem in number theory, because showing that a polynomial avoids being divisible by large squares requires careful control over its arithmetic properties. The authors focus on polynomials of a specific degree structure, namely degree equal to twice a prime, and prove that under a natural algebraic condition on the polynomial (that the number field it defines contains a certain symmetric subfield), a positive fraction of integer inputs yield squarefree outputs. This is a strong and unconditional result, not just an infinitely-often statement.
The algebraic condition the authors require is that the splitting field of the polynomial contains a Galois extension of the appropriate degree. This structure allows them to use tools from algebraic number theory and the geometry of numbers to count how often troublesome square factors can appear, and to show rigorously that they cannot appear too often. The argument combines sieve methods, which are analytic tools for filtering out unwanted divisibility, with structural information about how primes behave in the relevant number fields.
As a striking application, the authors use their theorem to study the geometry of units in a special family of cubic number fields called exceptional cubic fields. Every number field has a group of units whose arrangement can be visualized as a lattice, called the unit shape. The hexagonal lattice is a particularly symmetric and beautiful configuration. The authors construct an infinite family of non-Galois totally real cubic fields whose unit shapes get arbitrarily close to the hexagonal lattice. This is the first known example of such a family in this setting, and it connects a purely arithmetic result about squarefree values to a geometric statement about the shape of algebraic number fields.