Greenberg schemes and the motivis Serre invariant.

In this thesis, we use logarithmic methods to study motivic objects.
Let R be a complete discrete valuation ring with perfect residue field k, and denote by K its fraction field.
We give in chapter 2 a new construction of the motivic Serre invariant of a smooth K-variety and extend it additively to arbitrary K-varieties.
The main advantage of this construction is to rely only on resolution of singularities and not on a characteristic zero assumption, as did previous results.
As an application, we give a conditional positive answer to Serre's question on the existence of rational fixed points of a G-action on the affine space, for G a finite l-group.
We end the chapter by showing how the logarithmic point of view that we use in our construction leads to a new understanding of the motivic nearby cycles with support of Guibert, Loeser and Merle as a motivic volume.

In chapter 4 we use the theory of logarithmic geometry to derive a new formula for the motivic zeta function via the volume Poincaré series.
More precisely, we show how to compute the volume Poincaré series associated to a generically smooth log smooth R-scheme in terms of its log geometry, more specifically in terms of its associated fan in the sense of Kato.
This formula yields a much smaller set of candidate poles for the motivic zeta function and seems especially well suited to tackle the monodromy conjecture of Halle and Nicaise for Calabi-Yau K-varieties, for which log smooth models appear naturally through the Gross-Siebert programme on mirror symmetry.
We end the chapter by showing how this formula sheds new light on previous results regarding the motivic zeta function of a polynomial nondegenerate with respect to its Newton polyhedron, and of a polynomial in two variables.