Online material

Appendix A: Observations and data reduction

We collected low-resolution near-infrared spectra of six faint Pleiades proper motion candidates from Zapatero Osorio et al. (2014a) using the Long-slit Intermediate Resolution Infrared Spectrograph, LIRIS (Manchado et al. 2004), at the Cassegrain focus on the 4.2-m William Herschel Telescope (WHT) and the echelle and grating Near-Infrared Spectrometer, NIRSPEC (McLean et al. 1998), at the Nasmyth focus on the 10-m Keck II telescope. LIRIS and NIRSPEC have 1024 × 1024 HAWAII and ALADDIN-3 InSb detectors for the 0.8−2.5 and 0.95−5.5 μm ranges, respectively. The pixel scales are 0.̋25 (LIRIS) and 0.̋19 (NIRSPEC). Table 1 provides the journal of the spectroscopic observations including abridged targets names, observing dates in Universal Time (UT), integration times, and air masses. We selected the LIRIS ZJ (0.9−1.35 μm) and HK (1.45−2.4 μm) grisms and the NIRSPEC K-band (1.95−2.31 μm) low-resolution mode, which combined with slit widths of 1′′ (LIRIS) and 0.̋76 (NIRSPEC), that is four pixels along the dispersion directions of the detectors, provide final spectral dispersions and resolutions of 6.1 Å pix^-1 and 25 Å (ZJ), 9.8 Å pix^-1 and 39 Å (HK), and of 4.3 Å pix^-1 and 17 Å (K). The length of the slits were 4′ (LIRIS) and 42′′ (NIRSPEC), which is long enough to simultaneously observe our targets and bright reference sources in the field of view. To align targets and their respective bright reference stars along the slits (all at less than 1′-distance from the Pleiades sources), we conveniently rotated the instruments. The weather during the WHT campaigns were photometric, and cirrus were present during the Keck observations. Raw seeing oscillated between 0.̋9 and 1.̋2.

Spectra were acquired at two different nodding positions on the detectors for a proper sky subtraction. Raw data were reduced using standard procedures for the near-infrared wavelengths, including sky subtraction and flat-fielding correction, and packages within the IRAF² environment. Individual flat-fielded, sky-subtracted frames were properly registered using the bright reference stars and stacked to produce deep data. Spectra of the Pleiades targets were optimally extracted using the IRAF APALL task and were calibrated in wavelength to precisions of ±0.3−0.5 Å using arc lines of Ne+Ar. The spectra of hot B and early-A stars observed immediately after our targets were used for division into the corresponding science spectra. We previously removed the hydrogen lines intrinsic to these hot stars. Finally, we multiplied the science spectra by the blackbodies of temperatures typical of B/A stars. Figure 1 illustrates the resulting LIRIS and NIRSPEC spectra of Pleiades candidates. Given the faint magnitudes of our spectroscopic targets (J = 17.5−20.3 and K = 16.1−17.8 mag), some spectra have a poor signal-to-noise ratio (S/N = 2.5 − 8). To improve the data quality, we depict binned spectra as indicated in the figure caption. For a proper scaling of the ZJ and HK separate spectra of Calar 16 and 20, we employed the objects’ J and H photometry published in Zapatero Osorio et al. (2014a).

We also acquired Sloan z-band images of Calar 21 and 25 using the Auxiliary-port Camera (ACAM) at the folded-Cassegrain focus on the WHT. Table 1 provides the log of the imaging observations. ACAM has a 2048 × 4096 EEV CCD detector with a pixel pitch of 0.̋25 on the sky. The ACAM observations of 2014 February were conducted under photometric conditions and a seeing of 1.̋2 (z), while there were thin cirrus and variable seeing during the 2014 January campaign. The z-band data were taken following a nine-point dither pattern for a proper sky subtraction. Groups of three to nine images were combined to produce the mean sky contribution every 3 to 18 min, which were later removed from the raw data. Images were corrected for flat-field, properly aligned, and those with the best seeing were stacked to deliver one final deep image per target. Aperture and point-spread-function photometry were obtained within the IRAF PHOT package. Photometric calibration was performed using the Z-band data of bright sources in the field of view of our targets provided by the UKIRT Infrared Deep Sky Survey (UKIDSS, Lawrence et al. 2007), which fully overlaps with the Pleiades. We list the measured Z-band magnitudes in Table 1; the Sloan magnitudes were converted into the UKIDSS system using the AB to Vega correction and the color terms appropriate for mid- to late-L types given by Hewett et al. (2006). Calar 25 remains undetected after 1.4-h on source integration time. In Table 1, we list a 4-σZ-band upper limit of 23.35 mag (~3 mag deeper than the UKIDSS survey), where σ stands for the sky noise at the position expected for the source.

Together with the Pleiades candidates, we also observed known spectroscopic and photometric L-type reference dwarfs by employing the same LIRIS and ACAM instrumental configurations as for the targets. Table 1 contains the journal of these observations. All raw data were reduced in the same manner as the Pleiades candidates. The sources 2MASS J22244381−0158521 and 2MASS J09083803+5032088 are field L4.5 and L5 dwarfs (Kirkpatrick et al. 2000; Cruz et al. 2003). The latter has Sloan z and UKIDSS Z photometry, which allowed us to confirm the color-term correction between the two photometric systems valid for L5 dwarfs. The objects 2MASS J00452143+1634446 (L2) and 2MASS J03552327+1133437 (L5) are two young field dwarfs with lithium detection at 670.8 nm and another spectroscopic features indicative of low-gravity atmospheres (Cruz et al. 2009; Faherty et al. 2013; Zapatero Osorio et al. 2014b), whose ages are estimated at 10−100 and 50−500 Myr, respectively (Zapatero Osorio et al. 2014b). The LIRIS spectra of the reference L dwarfs are plotted along with the Pleiades targets in Fig. 1. Additional low-resolution spectra of DENIS J122815.2−154733 (L6), SDSS J042348.57−041403.5 (T0), and the young G 196−3B (L3) and 2MASS J12073347−3932540b (L7) taken from the literature (Leggett et al. 2001; Patience et al. 2010; Zapatero Osorio et al. 2010) are also included in Fig. 1.

© ESO, 2014