Duals of formal group Hopf orders in cyclic groups (Q1766860): Difference between revisions

This paper is concerned with finding Hopf orders in group rings \(K[G]\), where \(G\) is cyclic of order \(p^n\) and \(K\) is a \(p\)-adic field. A very good overview of what is known is given at the beginning of the paper, and we will not repeat it. The main point is the following: If \(G\) is elementary abelian of order \(p^n\), one knows how to construct Hopf orders in \(K[G]\) which depend on \(n(n+1)/2\) parameters. The Hopf orders constructed to this purpose are called triangular; they arise as iterated extensions of Hopf orders of rank \(p\), which are well understood. The first \(n\) parameters come from valuation criteria, and the rest (which one might call ``continuous'' parameters) come from certain units of \(O_K\) which go into the construction of the Hopf order. For cyclic \(G\) of order \(p^n\) and general \(n\), such families apparently had not yet been given. This is done in the present paper. The desired families are obtained as Cartier duals of so-called formal group Hopf orders, which are available from previous work of the same authors. (They are obtained by conjugating the multiplicative formal group and then taking a Hopf kernel.) It appears that these formal group Hopf orders already seem to involve the required number of parameters, but they are constructed in a way which is essentially different from triangular Hopf algebras. The main result states: Under suitable valuative conditions, the dual of a formal group Hopf order is a triangular Hopf order. The proof is fairly involved; it rests on explicit calculations. The link established in this paper is new (apart from the \(n=3\) case treated in the predecessor paper [Trans. Am. Math. Soc. 358, No. 3, 1117--1163 (2006; Zbl 1084.16013)]) and quite interesting. A minor remark: It might be argued that the condition \(i_{s-1}\geq 2npi_s\) in the main result (Thm.~2.2) is rather strict. If we assume \(n\geq 4\) (since the cases \(n\leq 3\) have been treated before) and also assume that \(i_1>0\) (the case \(i_1=0\) is in some sense degenerate, and less interesting), then even for \(p=2\), Theorem 2.2 only applies to Hopf orders \(H\) in \(K[G]\) for fields \(K\) that have very high ramification degree \(e_K\) over the rationals (\(e_K\geq 16^3=4092\) for \(n=4\), and the lower bound explodes with \(n\) faster than \(n!\)). It would be interesting to investigate whether a somewhat less strict hypothesis would suffice.