The first quantities that can be derived from these spectra are k-corrections, i.e., how magnitudes and colours change with the redshift because of the shift of the observed wavelength ranges. These corrections are crucial to link the properties of the local universe with the distant one (see, for example, Peebles, 1993), and their accurate knowledge is important to study the evolutionary effects. Our spectra allow us to compute these corrections in the near-IR filters starting from z = 0 with a greater accuracy than by using simple broad-band photometry as, for example, Cowie et al. (1994). In Figure 7 we show the k-corrections computed for the J, H and K bands. These filters have an increasing importance in the surveys for distant galaxies and in the study of the known high-redshift objects, especially when more 8-meters class space- and ground-based telescopes optimized for the infrared will be available.
Figure 7. k-corrections in the J, H and K bands between redshifts 0 and 2 for each morphological type: E (solid line), S0 (dotted), Sa (short dashed), Sb (long dashed) and Sc (dot-dashed). Our spectra define these k-corrections up to z = 0.3 in J, z = 0.7 in H and 1.2 in K, while at higher redshift they reflect the K96 spectra corrected to match the observed colours.
The k-corrections were computed using Bessell & Brett (1988) filters, more standard that the CIT system. We note that in the K band the k-corrections are almost independent from the galaxy type up to z = 2, reflecting the dominance of giant stars in these wavelengths in all the observed galaxies, as previously noted (Glazebrook et al. 1995). Note that no contributions are included from galaxy evolution or increasing absorption from the intergalactic medium; the computation is limited to z = 2 because above this limit these effects are certainly dominant.
We have compared our k-correction in the K band with others previously published. At low redshift (z < 0.6) we found consistency with the k-correction by Glazebrook et al. (1995) derived from Bruzual & Charlot (1993) model for a SSP with age of 5 Gyrs, and used, for example, by Loveday (2000) and Szokoly et al. (1998) to define the local luminosity function. On the contrary, big differences (up to 0.2 mag) are found with k-correction from 'UV-hot' elliptical model of Rocca-Volmerange & Guiderdoni (1988). At higher redshift (z > 0.6) our k-corrections are 'redder' up to 0.3 magnitudes than previous ones, however at these redshifts they should be taken with caution, and considered as lower limits, due to the possible evolutionary effect on galaxy spectrum.
Important differences are also found with the k-corrections by Poggianti (1997). Regarding the SED of the elliptical, the difference in J is generally about 0.1 and in H is always below 0.2, while differences greater than 0.3 are found in the K band in the redshift range z = 0.3 to 1.3, up to a value of about 0.45 magnitudes at z = 0.4-0.7. Although the differences in the filter response functions adopted here and in Poggianti (1997) contribute to the differences between the K-corrections, most of the K-band discrepancy arises from intrinsic differences in the SED, probably due to the sparse sampling of the Poggianti (1997) models at > 17000 Å. The comparison for Sa and Sc galaxies is less meaningful because of large spectral variations within each morphological class, possible aperture effects (see sec. 4), especially for Sc galaxies, and the presence of emission lines, not included in the Poggianti (1997) model.