Annu. Rev. Astron. Astrophys. 2001. 39: 249-307
Copyright © 2001 by . All rights reserved

Next Contents Previous

3.5. Background Measurements from IRTS

The most recent instrument to provide diffuse infrared background measurements is the Near Infrared Spectrometer (NIRS) (Noda et al. 1994), which operated for 30 days in 1995 on the Japanese Infrared Telescope in Space (Murakami et al. 1994, 1996, Noda et al. 1996). The IRTS was flown on the Space Flyer Unit spacecraft, which was launched into a 486-km-altitude, 28.5° inclination orbit. To maintain maximum possible offset angles of the field of view from the Sun and Earth, sky coverage during the mission was limited to 7%. The IRTS telescope was a heavily baffled, on-axis, 15-cm Ritchey-Chrétien system with four instruments sharing the focal plane. The NIRS, a grating spectrometer, had an 8-arcmin square beam and covered the spectral range from 1.4 to 4.0 µm with 0.12 µm resolution. There were no intermediate optical stops in the system to trap light diffracted or scattered from structures in the optical beam. The NIRS contained a full-beam cold shutter, which allowed frequent measurement of the instrumental zero point and responsivity.

Matsumoto et al. (2000) recently provided a preliminary analysis of the NIRS data, reporting detection of the CIB based on analysis of the 5 days of data that were least disturbed by atmospheric, lunar, and nuclear radiation effects. The sky area analyzed included Galactic latitudes from 40° to 58°, and ecliptic latitudes from 12° to 71°. The NIRS could detect individual stars down to a fainter level than the DIRBE (~ 10.5 mag at 2.24 µm). Matsumoto et al. used a later version of the statistical Galaxy model used by Arendt et al. (1998) to calculate the contribution from stars below their detection limit (Cohen 1997). They adopted the IPD model of Kelsall et al. (1998) to calculate the zodiacal light, interpolating between the DIRBE wavelengths to the wavelengths of the NIRS measurements. After subtraction of the IPD contribution, the residuals showed some remaining variation with ecliptic latitude, which they interpreted as evidence for a fairly isotropic background. To obtain a quantitative value for the background at each wavelength, they correlated their star-subtracted brightness at each point with the IPD model brightness and used the extrapolation to zero IPD contribution as a measurement of the CIB.

The CIB intensities reported by Matsumoto et al. near 2.2 and 3.5 µm are similar to the values found by Gorjian et al. (2000) (Figure 3). At shorter wavelengths, the Matsumoto et al. (2000) results continue to rise steeply to ~ 65 nW m-2 sr-1 at 1.4 µm. This is somewhat above the 95% confidence level upper limit at 1.25 µm of Wright (2001b). The NIRS result implies an integrated background energy over the 1.4 to 4.0 µm range of ~ 30 nW m-2 sr-1 (see Section 4.1 for discussion of the integrated background energy).

The preliminary report by Matsumoto et al. (2000) does not provide detail regarding systematic uncertainties in their results. For example, information such as the zero-point uncertainty, limits on stray light, and uncertainty due to a shutter light leak (Noda et al. 1996) would be valuable. The largest uncertainty in the reported CIB values is attributed to the IPD model. Use of the Kelsall et al. (1998) model introduces uncertainty associated with the relative photometric calibrations of the DIRBE and NIRS measurements, in addition to the intrinsic uncertainties in the model. Although consistently calibrated reference stars were used for gain calibration of the two instruments, the relative surface brightness photometry can be discrepant because of systematic errors in determination of the beam shape and spectral response for each instrument, as well as random errors in the measurements of the response of each instrument to the standard reference stars. Matsumoto et al. (2000) noted that there is some evidence of calibration inconsistency in that the correlation between the star-subtracted brightness of the NIRS maps and the DIRBE IPD model brightness did not have unity slope at all wavelengths. It will be valuable to have a more detailed account of the NIRS analysis and its uncertainties.

Next Contents Previous