Context. The Gaia astrometric mission of the European Space Agency was launched on December 2013. It will provide a catalog of 500 000 quasars. Some of these targets will be chosen to build an optical reference system that will be linked to the International Celestial Reference Frame (ICRF). The astrometric coordinates of these sources will have roughly the same uncertainty at both optical and radio wavelengths, and it is then mandatory to observe a common set of targets to build the link. In the ICRF, some targets have been chosen because of their pointlikeness. They are quoted as defining sources, and they ensure very good uncertainty about their astrometric coordinates. At optical wavelengths, a comparable uncertainty could be achieved for targets that do not exhibit strong astrophysical phenomena, which is a potential source of photocenter flickering. A signature of these phenomena is a magnitude variation at optical wavelengths. Aims. The goal of this work is to present the time series of 14 targets suitable for the link between the ICRF and the future Gaia Celestial Reference Frame. The observations have been done systematically by robotic telescopes in France and Chile once every two nights since 2011 and in two filters. These time series are analyzed to search for periodic or quasi-periodic phenomena that must be taken into account when computing the uncertainty about the astrometric coordinates. Methods. Two independent methods were used in this work to analyze the time series. We used the CLEAN algorithm to compare the frequency obtained to those given by the Lomb-Scargle method. It avoids misinterpreting the frequency peaks given in the periodograms. Results. For the 14 targets we determine some periods with a confidence level above 90% in each case. Some of the periods found in this work were not previously known. For the others, we did a comparative study of the periods previously studied by others and always confirm their values. All the periods given by the two methods are in good agreement with the difference always below 7.9% in the worst case. Finally we also present the structure functions for our two sets of objects (BL Lac and Seyfert galaxies). Conclusions. For all but one target, we find variations that could be the signature of periodic or quasi-periodic phenomena. Our time series could allow to bring some constraints on astrophysical models that could explain such variations. Binary black hole system, instabilities in the jet or in the accretion disk, changes in the torus structure are some of these astrophysical phenomena. They must be kept in mind when evaluating the uncertainty about the astrometric position of the targets suitable for the link between reference systems in a conservative way.