A counterexample to the first Zassenhaus conjecture (Q1989713)

Let \(G\) be a finite group, and let \(\mathcal{U}(\mathbb{Z}G)\) be the unit group of the integral group ring of \(G\). The first Zassenhaus conjecture says that for any unit \(u\) of finite order in \(\mathcal{U}(\mathbb{Z}G)\) there exists a unit \(a\) in \(\mathcal{U}(\mathbb{Q}G)\) such that \(a^{-1}\cdot u \cdot a=\pm g\) for some \(g \in G\). In this paper, the authors disprove this conjecture by constructing an infinite series of metabelian groups \(G\) such that \(\mathbb{Z}G\) contains a unit of finite order not conjugate in \(\mathbb{Q}G\) to any element of the form \(\pm g\) for \(g \in G\). These groups \(G\) are certain groups of the form \(G(p, q; d; \alpha, \beta )=N\rtimes A\), where \(p\) and \(q\) be odd primes, \(d\) an odd divisor of \(p -1\) and \(q-1\), \(N\) the additive group \(\mathbb{F}_{p^2} \oplus \mathbb{F}_{q^2}\), \(\alpha\) and \(\beta\) are primitive elements in the multiplicative groups \(\mathbb{F}^\times_{p^2}\) and \(\mathbb{F}^\times_{q^2}\), respectively, \(A = \langle a, b, c \mid a^{\frac{p^2-1}{d}} = b^{\frac{q^2-1}{d}} = 1, \ c^d= a b\rangle\), and the action of \(A\) on \(N\) is given by \((x, y)^a = (\alpha^d x, y)\), \((x, y^b = (x, \beta^d y)\), \((x,y)^c = (\alpha x,\beta y)\). The smallest example is a group \(G\) of order \(27 \cdot 32 \cdot 5 \cdot 72 \cdot 192\) such that \(\mathcal{U}(\mathbb{Z}G)\) contains an element of order \(7 \cdot 19\) which is not conjugate in \(\mathcal{U}(\mathbb{Q}G)\) to any element of the form \(\pm g.\)