Abstract

In this paper we investigate the origin of the mid-infrared (IR) hydrogen recombination lines for a sample of 114 disks in different evolutionary stages (full, transitional, and debris disks) collected from the Spitzer archive. We focus on the two brighter H I lines observed in the Spitzer spectra, the H I (7-6) at 12.37 mu m and the H I (9-7) at 11.32 mu m. We detect the H I (7-6) line in 46 objects, and the H I (9-7) in 11. We compare these lines with the other most common gas line detected in Spitzer spectra, the [Ne II] at 12.81 mu m. We argue that it is unlikely that the H I emission originates from the photoevaporating upper surface layers of the disk, as has been found for the [Ne II] lines toward low-accreting stars. Using the H I (9-7)/H I (7-6) line ratios we find these gas lines are likely probing gas with hydrogen column densities of 10(10)-10(11) cm(-3). The subsample of objects surrounded by full and transitional disks show a positive correlation between the accretion luminosity and the H I line luminosity. These two results suggest that the observed mid-IR H I lines trace gas accreting onto the star in the same way as other hydrogen recombination lines at shorter wavelengths. A pure chromospheric origin of these lines can be excluded for the vast majority of full and transitional disks. We report for the first time the detection of the H I (7-6) line in eight young (20 Myr) debris disks. A pure chromospheric origin cannot be ruled out in these objects. If the H I (7-6) line traces accretion in these older systems, as in the case of full and transitional disks, the strength of the emission implies accretion rates lower than 10(-10) M-circle dot yr(-1). We discuss some advantages of extending accretion indicators to longer wavelengths, and the next steps required pinning down the origin of mid-IR hydrogen lines.