3
$\begingroup$

Suppose I have a local field $\mathcal{O}_K$ and two different prime elements $\pi$ and $\overline{\pi},$ i.e they differ by a unit $\overline{\pi} = u \pi$ for some $u \in \mathcal{O}_K^{\times}$ not equal to $1.$ Let us suppose I have a Lubin Tate module $F_\pi$ for the prime element $\pi$ and a Lubin-Tate module $F_{\overline{\pi}}$ for $\overline{\pi}.$ I have heard it been claimed that in this case, $F_{\overline{\pi}}$ and $F_{\pi}$ are never isomorphic. Is there an easy way to see why this is true?

$\endgroup$

2 Answers 2

4
$\begingroup$

Another way of seeing this is by way of the totally ramified extension $\mathcal L_\pi$ generated by the torsion points of $F_\pi$ over $K$. In $K$, $\pi$ is the only prime element that’s the norm from every finite extension of $K$ in $\mathcal L_\pi$. Isomorphic formal modules give the same totally ramified field, and so the same universal norm $\pi$.

$\endgroup$
2
$\begingroup$

If the formal groups were isomorphic then the (conjugation by the) corresponding change of variables would send $[\pi]_{F_\pi}$ into $[\pi]_{F_{\pi'}}$. Thus modulo the maximal ideal $(\pi)$ we have $[\pi]_{F_{\pi'}}(X)\equiv X^q\equiv [\pi']_{F_{\pi'}}(X)$. Hence the reduction of $[\pi-\pi']_{F_{\pi'}}$ would be zero, which is certainly wrong (easy).

$\endgroup$
3
  • $\begingroup$ Why is it easy to see that $[\pi-\pi']$s reduction can't be zero? would you mind giving some details? I think I could see this, but I'm a bit unsure since over the completion of the maximal unramified extension, $ F_\pi$ and $F_{\pi'}$ are isomorphic? $\endgroup$
    – user44591
    Commented Nov 7, 2015 at 0:17
  • $\begingroup$ If you increase the residue field then the reduction of $h\circ [\pi]_F\circ h^{-1}$ ($h$ is a "change of variables") does not have to equal $X^q$. As about $[\pi'-\pi]$: since $\pi'-\pi$ has finite valuation, some iteration of $[\pi']$ factors through $[\pi'-\pi]$. $\endgroup$ Commented Nov 7, 2015 at 8:07
  • 1
    $\begingroup$ For a formal group of finite height, the map from the endomorphisms in charactecteristic zero to the endomorphisms in characteristic $p$ is always an injection. $\endgroup$
    – Lubin
    Commented Nov 8, 2015 at 3:17

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .