Clumpy galaxies at z ~ 0.6: kinematics, stability and comparison with analogues at other redshifts

International audienceDistant clumpy galaxies are thought to be Jeans-unstable discs and an important channel for the formation of local galaxies, as suggested by recent spatially resolved kinematic observations of z ~ 2 galaxies. I study the kinematics of clumpy galaxies at z ~ 0.6 and compare their properties with those of counterparts at higher and lower redshifts. I selected a sample of 11 clumpy galaxies at z ~ 0.6 from the representative sample of emission-line, intermediate-mass galaxies IMAGES. Selection was based on rest-frame UV morphology from Hubble Space Telescope/Advanced Camera for Surveys images, mimicking the selection criteria commonly used at higher redshifts. Their spatially resolved kinematics were derived in the frame of the IMAGES survey, using the Very Large Telescope/FLAMES-GIRAFFE multi-integral field spectrograph. For those showing large-scale rotation, I derived the Toomre Q parameter, which characterizes the stability of their gaseous and stellar phases. I find that the fraction of UV-selected clumpy galaxies at z ~ 0.6 is 20 +/- 12 per cent. Roughly half of them (45 +/- 30 per cent) have complex kinematics inconsistent with Jeans-unstable discs, while those in the remaining half (55 +/- 30 per cent) show large-scale rotations. The latter reveal a stable gaseous phase, but the contribution of their stellar phase makes them globally unstable to clump formation. Clumpy galaxies appear to be less unstable at z ~ 0.6 than at z ~ 2, which could explain why the UV clumps tend to vanish in rest-frame optical images of z ~ 0.6 clumpy galaxies, conversely to z ~ 2 clumpy galaxies, in which the stellar phase can substantially fragment. This suggests that the former correspond to patchy star formation regions superimposed on a smoother mass distribution. A possible and widespread scenario for driving clump formation relies on instabilities by cold streams penetrating the dark matter haloes where clumpy galaxies inhabit. While such a gas accretion process is predicted to be significant in massive, z ~ 2 haloes, it is also predicted to be strongly suppressed in similar, z ~ 0.6 haloes, which could explain why lowest z clumpy galaxies appear to be driven by a different mechanism. Instead, I found that interactions are probably the dominant driver leading to the formation of clumpy galaxies at z 1 clumpy galaxies remains more uncertain. While cold flows could be an important driver at z ~ 2, I also argue that the observed and cumulative merger fraction between z = 2 and z = 3 is large enough so that every z ~ 2 galaxy might be the result of a merger that occurred within their past 1Gyr. I conclude that it is premature to rule out mergers as a universal driver for galaxy evolution from z ~ 2 down to z = 0