A universal formula for the \(j\)-invariant of the canonical lifting (Q2253873): Difference between revisions

Let \(E\) be an ordinary elliptic curve defined over an algebraically closed field \(k\) of characteristic \(p\). Let \(W(k)\) be the ring of Witt vectors of \(k\) and, for any \(n\), denote by \(W_n(k)\) the subring of vectors of length \(n\). By the general Serre-Tate theorem, for any \(n\), it is possible to define a (unique up to isomorphism) canonical lifting \(\mathbb{E}_n\) of \(E\) over \(W_n(k)\) (it corresponds to the identity element of a certain group of homomorphisms and it is isomorphic to \(E\) when reduced modulo \(p\)): these liftings are compatible with respect to the natural maps \(W_m(k)\rightarrow W_n(k)\), hence they define a \textit{canonical lifting} \(\mathbb{E}\) of \(E\) over \(W(k)\). The paper deals with the \(j\)-invariant \(j(\mathbb{E})\) and, in particular, with the map \[ \Theta:\{j_0\in k\;\text{s.t.}\;\text{any}\;E\;\text{with}\;j(E)=j_0\;\text{is \;ordinary}\} \longrightarrow W(k), \] \[ \Theta(j_0)=\Theta(j_0(E)):=j(\mathbb{E})=(j_0,j_1,j_2,\dots) \;. \] The main theorem states that there exist polynomials \(f_i\in \mathbb{F}_p\left[ X,\frac{1}{\phi_p(X)}\right]\) (where \(\phi_p(X)\) is the Hasse polynomial of \(\mathbb{F}_p[X]\), whose roots are the \(j\)-invariants of supersingular curves in characteristic \(p\)), such that \(\Theta(j_0)=(j_0,f_1(j_0),f_2(j_0),\dots)\). For the proof, the author considers a field of characteristic \(p\) (not necessarily algebraically closed) and extends the definition of canonical lifting to families of elliptic curves defined over \(\mathbb{F}_p\)-schemes by working on the scheme of \(p^\infty\)-torsion points (using the Serre-Tate theorem) and their connected-étale sequence. This provides a formula for \(j(\mathbb{E})\) similar to the required one but defined on the perfect closure of the base field. Then a theorem of \textit{L. R. A. Finotti} [J. Number Theory 130, No. 3, 620--638 (2010; Zbl 1217.11059)] on canonical liftings over imperfect fields (which is proved in a different way in the paper under review) provides the final crucial tool for the proof of the main theorem.