On Chow motives of surfaces (Q2450159)

The paper under review should be seen as a drastic strengthening of the author's earlier paper [Invent. Math. 181, No. 1, 1--19 (2010; Zbl 1193.14050)] and these two papers should be regarded as a pair. It proves a phenomenon baptized ``Rost nilpotence'' for all surfaces of finite type over a field of characteristic zero. The previous paper accomplished the same, but only for the much smaller class of geometrically rational surfaces. The phenomenon of ``Rost nilpotence'' was first isolated by Rost for quadrics. It says that an endomorphism of a Chow motive (over some field \(K\)) which trivializes under base change to a larger base field \(L\)/\(K\) must be nilpotent (under composition of endomorphisms). This was first shown for quadrics by Rost and then later for projective homogeneous varieties by Chernousov, Gille and Merkurjev in [\textit{V. Chernousov} et al., Duke Math. J. 126, No. 1, 137--159 (2005; Zbl 1086.14041)]. As far as the reviewer can judge this at all, Gille's proof of Rost nilpotence for surfaces came as quite a surprise to many and is a breakthrough result. Needless to say, the methods are completely different from the previous works on the matter. It is based on first expressing the endomorphism groups of motives as Chow groups (this is trivially true by definition), and studying the latter via the explicit cycle complexes of Rost. A tricky point is to lift the action by correspondences to the cycle complexes themselves instead of just their cohomology; the necessary technical construction to resolve this had already been established by Rost, but it is used in a very original and beautiful way by the author. Next, this gets combined with studying the action of the Galois group of the extension \(L\)/\(K\) on these explicit cycle complexes -- after observing that it suffices to study finite Galois extensions. This part relies on more classical results of Bloch and Colliot-Thelene which belong to the field of studying algebraic cycles under base change of the base field. They are in themselves very deep, relying on the action of the Galois group on \(K\)-theory/motivic cohomology. This structure of proof is largely parallel to the author's previous paper, but assembled in a much more powerful way. One can very comfortably read them simultaneously and find certain aspects that are explained only briefly in the present paper with more detail in the previous paper of the author.