Deux photons - deux atomes (2P2A) excitation s’avère être un banc de test importantde propriétés optiques quantiques. Depuis 1980, on sait que la probabilité de transition2P2A peut être grandement améliorée en utilisant la source monochromatique cohérenteavec la présence d’interactions atomiques qui induisent un pic de résonance de 2P2A. En2004 , Muthukrishnan et al. sont trouvés une grande amélioration de la transition ainsique d’une résonance de 2P2A, quand une source de deux photons intriqués, provenant dela radiation d’un atome cascade à trois niveaux, interagit avec les deux atomes différentsde l’ interaction. Leur conclusion est que : dans certaines situations, l’intrication peutremplacer un vrai hamiltonien d’interaction, qui est une déclaration de grande portée etune propriété importante liée à l’intrication. Ce phénomène mérite plus d’attention et defavoriser la compréhension. / Two-photonÐtwo-atom (2P2A) excitation turns out to be an important test bench of quantum optical properties. In 2004, Muthukrishnan et al. announced a great enhancement of the transition at 2P2A resonance, when an entangled two-photon state, coming from a three-level cascade, interacts with two non-interacting atoms. Their conclusion: in some situations, entanglement can replace a real interaction Hamiltonian, reveals an important property related to entanglement. In this thesis, We study in detail the efficiency of various multimode two-photon states of light to induce the simultaneous excitation of two atoms of different kinds under 2P2A resonance condition. We find that several separable, correlated states, produced either by an atomic cascade or parametric down-conversion, or even appropriate combinations of coherent states, have comparable efficiencies as the photon state produced by an atomic cascade. We conclude that the true physical origin of the enhancement is the frequency anti-correlations, neither the temporal correlations, time ordering, nor entanglement. To present, we are not sure whether or not the enhancement is a quantum effect. We also studied: 1. the optimized 2P2A transition probability for a given photon spectrum; 2. the influence coming from finite lifetimes of the excited detecting atoms; 3. the 2P2A transition probability in the presence of atomic interaction.