Fermat's equation over the tower of cyclotomic fields (Q4328671)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 1734690
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | Fermat's equation over the tower of cyclotomic fields |
scientific article; zbMATH DE number 1734690 |
Statements
Fermat's equation over the tower of cyclotomic fields (English)
0 references
17 August 2003
0 references
Fermat's equation
0 references
Bernoulli numbers
0 references
local units
0 references
logarithms
0 references
0.8435828685760498
0 references
0.8325406908988953
0 references
0.7764379382133484
0 references
The author proves that the Fermat equation \(x^l+y^l+z^l=0\) (\(l>3\) prime) is unsolvable over some specific subfields of \({\mathbb Q}(\zeta_{l^\infty})\). The latter field is the inductive limit of the fields \(L_n={} {\mathbb Q}(\zeta_{l^n})\); let \(R_n=L_n^+\) and \(H_n\) the maximal \(l\)-subextension of \(L_n\) (or \(R_n\)). Then the author specifies upper bounds \(\phi_1(l) \leq \phi_2(l)\leq \phi_3(l)\) such that the first case of FLT holds for the fields \(L_{\phi_1(l)}\), \(R_{\phi_2(l)}\), and \(H_{\phi_3(l)}\) respectively. (Some results on the second case are also proved.) The author gives explicit lower bounds on the numbers \(\phi_j(l)\); these lower bounds are of the shape \(c(l)/(I(l)+1)\) where \(I(l)\) is the irregularity index of \(l\) and \(c(l)\) grows a bit slower than \(\sqrt{\log(l)}\). By the very definition, \(\phi_j(l)\) can never be larger than \((l-3)/2\). The starting point is Kummer's classical idea of factoring \(x^l+y^l\) over the \(l\)-th cyclotomic field. The author already considered the case of the field \(L_0\) in a previous paper [Proc. Steklov Inst. Math. 208, 146-165 (1995); translation from Tr. Mat. Inst. Steklova 208, 163-185 (1995; Zbl 0881.11037)]. In the present paper, the methods still are ``classical'' but very clever. The definition of the bounds \(\phi_j(l)\) is not easy to grasp for a reader, but everything is explicit and could easily be put on a computer. An essential new ingredient (with respect to Kummer's approach) seems to be a substantial chunk of local theory, starting at Proposition 3. The general idea is to start with a statement of the shape ``a certain Galois power of a rational expression \(\omega_k(x,y)\) is an \(l\)-th power in the \(l\)-completion of \(L_n\)'' and to deduce that certain explicit polynomials \(P_d\) have roots modulo \(l\). (The size of the bounds \(\phi_j(l)\) depends on the existence of zeros of these polynomials mod \(l\).) In the argument, it is important to control the image of the \(l\)-adic logarithm map on \(l\)-adic units in \(L_n\), which is somewhat tricky for \(n>0\). Notwithstanding its technical nature this is an important paper which provides significant results. The reviewer has a few minor remarks: The statement concerning the proportion of regular primes (p. 503) is backed not only by computations but also by some fairly reasonable heuristical arguments. Two comments on the translation from the Russian, which is rather good in general, cf. the review Zbl 0881.11037 quoted above: The last sentence in the paragraph just after Theorem 4 is not quite clear in the English version. The original states it clearly: if \(l\) runs through any strictly increasing sequence of primes with bounded irregularity index (assuming such sequences exist!), then \(\phi_j(l)\) tends to infinity as well. Also, instead of ``differentiation'' (differentsirovanie), ``derivation'' would have been more standard. On p.513 (below middle) read \(\log(V_0)=\ell_0^2S_0\) instead of \(\log(V_0)=1+\ell_0^2S_0\).
0 references