• gamma
• CKM
• CDF
• Bs
• cp violation

The CP violation has played a central role in particle physics since 1964, year of its discovery in the kaon system. The large asymmetries predicted, and later observed, in the B meson system have confirmed the Cabibbo-Kobayashi-Maskawa (CKM) picture, but opened further questions and stimulated an increase of precision of the measurements in the search for new physics sources of CP violation.
While the CKM scheme can accommodate a certain amount of CP violation, it does not require it, nor it provides any fundamental motivation for its existence. In fact, the "cosmological question" of the dominance of matter over antimatter in our universe is a strong hint for the presence of CP-violating effects of a much larger strength than the standard CKM mechanism. Experimental measurements allow to determine the angles and the sides of the so-called "unitarity triangle", formed by the CKM matrix parameters. Today a good level of precision has been reached for two of the angles, but the resolution on the third one (angle gamma) is still rather poor.

One of the most important decay to probe such a field is the Cabibbo-suppressed mode Bs -> Ds K, which can have a large CP-violating interference via Bs mixing.
Isolating a clean sample of such decays is a crucial step toward a precise measurement of gamma and this thesis deals this challenge.
Selecting an enriched sample of Bs -> Ds K decays, keeping high the signal efficiency and rejecting the huge amount of background, which overwhelms the signal of interest by several order of magnitude, is a crucial ingredient. It also essential to disentangle Bs -> Ds K decays from kinematically similar decays, which lay down in the same mass region of the signal.

In particular the thesis describes the analysis of the Bs->Ds K decays collected by CDF II experiment at the Tevatron collider. The thesis uses the full data sample of the CDF II corresponding to about 10/fb of integrate luminosity.

To improve the signal to background ratio the selection of the Bs -> Ds K decays is optimized using an Artificial Neural Network, allowing an optimal use of the multidimensional information contained in the input variables.
As far as the presence of similar and larger physics backgrounds a careful determination of the backgrounds features and a precision calibration, using real data, of the particle identification observables is done. Thus all information, coming from kinematics and particle identification of the final state particles, is combined in an unbinned maximum likelihood fit to disentangle relative fractions of all contributions.

As a result, we measure the branching ratio B(Bs -> Ds K)/B(Bs -> Ds pi) decay.