# Dual concept for the p-primary component

Is there a dual concept for the p-primary component of an abelian group? Please name some books/papers where it is studied.

-
Posted at the same time to M.SE: math.stackexchange.com/questions/311702/… This practice is frowned upon! –  Steve D Feb 23 '13 at 5:12

Well, first of all you should start finding a dual notion for the torsion part. Indeed, a given Abelian discrete Abelian group $G$ is the central element in a short exact sequence: $$0\to t(G)\to G\to G/t(G)\to 0 ,$$ where $t(G)$ is the torsion part and $G/t(G)$ is torsion free. As the duality functor (in the Pontryagin-Van Kampen duality) is exact you get a short exact sequence of compact Abelian groups: $$0\to \widehat{G/t(G)}\to \widehat G\to \widehat{t(G)}\to 0 .$$ Recall also that the torsion part $t(G)$ can be definite to be the direct union of all the finite subgroups of $G$. Using that finite Abelian groups are self-dual and that the duality functor send direct to inverse limits, you obtain that $\widehat{t(G)}$ is an inverse limit of finite Abelian groups, that is, it is pro-finite. In fact, it is the maximal pro-finite quotient of $\widehat G$.
Now, coming to your question, you have to recall that $t(G)$ is the direct sum of its $p$-primary components $t_p(G)$ (each one defined as the direct union of all the finite $p$-subgroups of $G$) and that the duality functor sends direct sums (i.e., coproducts) to products. Thus $\widehat {t(G)}$ is the (topological) product of the $\widehat{t_p(G)}$ and each $\widehat{t_p(G)}$ is the inverse limit of $p$-torsion finite groups, that is, pro $p$-finite.
To conclude, the answer to your question is: the dual of the $p$-torsion subgroup is the maximal pro $p$-finite quotient of the dual group.