COLOR EVOLUTION IN HIGH REDSHIFT GALAXIES (INSTRUMENTATION, INFRARED, PHOTOMETRY).

Abstract

A Simultaneous Photometer for Infrared and Visual light (hereafter SPIV) has been constructed. SPIV uses three dichroic filters to divide light from a common aperture in the telescope focal plane into four colors with bandpasses (in microns) of: 0.42 to 0.7 (V(B)); 0.7 to 0.95 (I(B)); 1.45 to 1.8 (H); 1.97 to 2.27 (K). The H and K bands are detected by liquid helium cooled InSb diodes. I(B) is detected by a helium cooled Si diode, and V(B) by an uncooled EMI 9658R photomultiplier tube. The instrument response function (IRF) including the effect of atmospheric transmission is shown. The maximum usable aperture size of 2 mm corresponds to 7 arcseconds on the Multiple Mirror Telescope and to 20 arcseconds on the UAO 61" and 90" telescopes. Information about the sky brightness is combined with the IRF to calculate the expected background noise. These calculations show SPIV should be background noise limited in all bands. Observations show this is true except for I(B), which is detector noise limited. The optical alignment of the four channels is shown to be satisfactory. The SPIV instrumental magnitude system is defined. Observations of 40 radio galaxies and 39 non-radio first ranked cluster galaxies with known redshifts ranging from 0.019 to 1.6 are reduced on this system, correcting for atmospheric extinction, reddening, and aperture. The transformation to standard magnitude systems is derived. The reduced V(B)-H, I(B)-H and H-K colors as a function of redshift are compared for the radio and non-radio galaxies using statistical tests. No convincing differences are found, with the possible exception of H-K in the z = 0.2 to 0.4 range. These colors are also compared to a "no evolution" prediction generated by redshifting a composite spectrum of nearby elliptical galaxies, and to evolutionary models from Bruzual (1981). Passively evolving models with little residual star formation and a galaxy formation epoch z(f) > 3 are slightly favored by H-K observations. These models predict about one magnitude brightening at H by a redshift of one. The I(B)-H color becomes redder with redshift much faster than the reddest model, and shows large scatter for z 3. Residual star formation models require an older universe than is allowed by qₒ = 0.5, Hₒ = 80 km/sec/Mpc to fit the red V(B)-H envelope at high redshift. Some galaxies show strong blueward deviations for z > 0.4. This behavior is most easily explained by episodes of star formation involving small fractions of the total number of stars. Because of the flatness of H-K and I(B)-H at high redshift, and the scatter in V(B)-H, determining redshifts > 0.4 from broadband colors will be difficult.