If the sample is Raman active then it depends on the setup you are using, if it is nano-, micro- or macro- configurations. In nano and micro you would need the least material, if they are powders even a small quantity equivalent to a single grain of sand is more than enough. In micro you focus the laser and collect the Raman signal using optical objectives (an optical microscope). In nano, specially in tip-enhanced Raman spectroscopy, you can increase the sensitivity up to several orders of magnitude so requiring very small quantities. If you don't have access to nano or macro configurations then even in macro you can use very small quantities if you choose a proper substrate in order to have surface-enhanced Raman spectroscopy.

Thank you so much for your answer.

However metals are very weak scatterers. Probably you would like this paper

J Raman Spec, 2011, online: "Study of Raman spectrum of uranium using a surface-enhanced Raman scattering technique"