Neutron-induced capture cross sections of short-lived nuclei are of great importance in many areas: fundamental nuclear physics, nuclear astrophysics and nuclear applications, such as nuclear reactors. However, very often the high radioactivity of the samples makes the direct measurement of these cross sections extremely difficult. The surrogate reaction method [1] is an indirect way of determining neutron-induced cross sections for reactions that proceed through a compound nucleus. In this method, the decaying nucleus of the desired reaction is produced via a transfer or an inelastic scattering reaction (figure 1). This technique presents the advantage that the target material can be stable or less radioactive than the material required for a neutron-induced measurement. Nevertheless, a significant question in the use of the surrogate reaction method lies in the difference between the spin and parity population in the neutron-induced and surrogate reactions. The CENBG collaboration has successfully applied this technique to determine the neutron-induced fission cross sections of several short-lived nuclei such as 233Pa[2], 242;243Cm and 241Am[3]. The results are in very good agreement with neutron-induced data. We are currently investigating whether this powerful technique can also be used to determine radiative capture cross sections. For this purpose, we have studied the transfer reactions 174Yb(3He,pg)176Lu and 174Yb(3He,4Heg)173Yb as surrogates for the 175Lu(n,g) and 172Yb(n,g) reactions, respectively. The g probabilities obtained in our experiment show clear discrepancies with already existing neutroninduced data. In this contribution we will present the experimental set-up and the data analysis. Then, we will focus on the origin of these discrepencies. Perspectives for the surrogate method applied to capture cross sections will be discussed.< Réduire