Abstract

Context. The mechanisms of planet formation are still under debate. We know little about how planets form, even if more than 4000 exoplanets have been detected to date. Recent investigations target the cot of newly born planets: the protoplanetary disk. At the first stages of their life, exoplanets still accrete material from the gas-rich disk in which they are embedded. Transitional disks are indeed disks that show peculiarities, such as gaps, spiral arms, and rings, which can be connected to the presence of substellar companions. Aims. To investigate what is responsible for these features, we selected all the known transitional disks in the solar neighborhood (<200 pc) that are visible from the southern hemisphere. We conducted a survey of 11 transitional disks with the SPHERE instrument at the Very Large Telescope. This is the largest H alpha survey that has been conducted so far to look for protoplanets. The observations were performed with the H alpha filter of ZIMPOL in order to target protoplanets that are still in the accretion stage. All the selected targets are very young stars, less than 20 Myr, and show low extinction in the visible. Methods. We reduced the ZIMPOL pupil stabilized data by applying the method of the angular spectral differential imaging (ASDI), which combines both techniques. The datacubes are composed of the Cnt_H alpha and the narrow band filter H alpha, which are taken simultaneously to permit the suppression of the speckle pattern. The principal component analysis method was employed for the reduction of the data. For each dataset, we derived the 5 sigma contrast limit and converted it in upper limits on the accretion luminosity. Results. We do not detect any new accreting substellar companions around the targeted transition disks down to an average contrast of 12 magnitudes at 0.'' 2 from the central star. We have recovered the signal of the accreting M star companion around the star HD 142527. We have detected and resolved, for the first time in visible light, the quadruple system HD 98800. For every other system, we can exclude the presence of massive actively accreting companions, assuming that the accretion is not episodic and that the extinction is negligible. The mean accretion luminosity limit is 10(-6) L-circle dot at a separation of 0.'' 2 from the host.