Context. The [Ne II] line 12.81 μm was proposed to be a good tracer of gas in the environments of proto-planetary disks; its origin is explained by different mechanisms: jets in outflows, photo-evaporative disk winds driven by stellar X-rays/EUV or by the X-ray irradiated proto-planetary disk atmosphere. Previous Spitzer studies gave hints toward the neon emitting mechanism by exploring correlations between the line luminosity and properties of the star-disk system. These studies concluded that the origin of the emission is likely related to accretion and outflows, with some influence from X-rays.

Aims. We provide direct constraints on the origin of the [Ne II] emission using high-spatial and spectral resolution observations that allow us to study the kinematics of the emitting gas. In addition we compare the [Ne II] line with optical forbidden lines.

Methods. We obtained high-resolution ground-based observations with VISIR-VLT for 15 stars and UVES-VLT for three of them. The stars were chosen for having bright neon emission lines detected with Spitzer/IRS. The velocity shifts and profiles are used to disentangle the different emitting mechanisms producing the [Ne II] line. A comparison between results from this study and previous high-resolution studies is also presented.

Results. The [Ne II] line was detected in seven stars, among them the first confirmed detection of [Ne II] in a Herbig Be star, V892 Tau. In four cases, the large blueshifted lines indicate an origin in a jet. In two stars, the small shifts and asymmetric profiles indicate an origin in a photo-evaporative wind. CoKu Tau 1, seen close to edge-on, shows a spatially unresolved line centered at the stellar rest velocity, although cross-dispersion centroids move within 10 AU from one side of the star to the other as a function of wavelength. The line profile is symmetric with wings extending up to ~±80 km s^-1. The origin of the [Ne II] line is unclear and could either be due to the bipolar jet or to the disk. For the stars with VLT-UVES observations, in several cases, the optical forbidden line profiles and shifts are very similar to the profile of the [Ne II] line, suggesting that the lines are emitted in the same region. A general trend observed with VISIR is a lower line flux when compared with the fluxes obtained with Spitzer. We found no correlation between the line full-width at half maximum and the line peak velocity. The [Ne II] line remains undetected in a large part of the sample, an indication that the emission detected with Spitzer in those stars is likely extended.