Recognizing that the uncertainties in isolating the bright contribution from the IPD are a major obstacle to CIB detection in the far infrared, Finkbeiner, Davis, & Schlegel (2000) analyzed the DIRBE data using two approaches which avoid the need for a detailed IPD model. Both methods were based on analysis of the annual variation of dimensionless ratios of the data after removal of the ISM contribution. The approach of Schlegel et al. (1998) was used to determine the ISM contribution. Subtraction of the ISM contribution from each DIRBE weekly map yielded maps nominally containing only the IPD signal, which varies over the course of the year, and any time-independent signals due to other Galactic contributions and the CIB. They constructed dimensionless ratios of the brightnesses in opposing directions on the sky, and compared the annual variation or ecliptic latitude variation of these ratios to models containing contributions from interplanetary dust plus a temporally constant background. Though these methods do not require detailed models of the IPD cloud, they do require assumptions about the spatial symmetries and temporal invariance of the IPD cloud similar to those of the Kelsall et al. (1998) model. In one approach, they analyzed the annual variation of [N-S] / [N+S], where N (S) is the brightness toward the North (South) ecliptic pole. This analysis yielded statistically significant temporally constant backgrounds at 60 and 100 µm, but does not demonstrate isotropy of that signal. To explore the ecliptic latitude dependence of the background, they analyzed a second dimensionless ratio based upon brightnesses at 90° solar elongation and symmetric ecliptic latitudes forward and backwards relative to the Earth's direction of motion. Interpretation of the results of this method does depend modestly on assumptions about the ecliptic latitude dependence of the IPD contribution, i.e., on a model for the IPD cloud at high ecliptic latitudes. These two methods gave comparable results for the constant backgrounds at 60 and 100 µm. Finkbeiner et al. tentatively identified these backgrounds with the CIB (Table 1), but noted that the values are rather high and not consistent with limits implied by the opacity of the intergalactic medium to TeV -rays (Aharonian et al. 1999; Samuelson et al. 1998). They concluded that there is not yet a satisfactory explanation for these constant backgrounds.
A distinctly different approach to studying the CIB is to look for fluctuations attributable to the non-uniform spatial distribution of discrete sources contributing to the background. If one can estimate the expected amplitude of the fluctuations relative to the brightness of the background itself, one can use observations of the fluctuations to determine or set limits upon the CIB. Of course, measuring fluctuations in the extragalactic background brightness presents similar challenges of discriminating contributions from foreground sources and instrumental effects as those that exist for direct detection of the background. However, at mid-infrared wavelengths, the dominant IPD foreground is less of an obstacle to fluctuation studies than to total CIB brightness measurements since the IPD emission varies relatively smoothly over the sky.
Kashlinsky et al. (1996a) provided a framework for such analyses. In order to relate fluctuations to the intensity of the CIB, they evaluated the expected fluctuations from galaxies, finding that the fluctuations on the 0.7° angular scale of the DIRBE beam are expected to be of the order of 5-10% of the CIB. Kashlinsky, Mather, & Odenwald (1996b) analyzed the fluctuations in the DIRBE maps from 12-100 µm after removing the IPD light using the model of Kelsall et al. (1998) and blanking bright discrete sources. They calculated the zero lag autocorrelation function, C(0), finding C(0)1/2 (1 - 1.5) nW m-2 sr-1. Kashlinsky & Odenwald (2000) further analyzed the fluctuations in this spectral range, finding a slightly lower limit C(0)1/2 (0.5 - 1) nW m-2 sr-1. These analyses imply upper limits to the 12-100 µm CIB contributed by sources clustered like galaxies of the order of 10-15 nW m-2 sr-1, the most restrictive limits based upon infrared measurements in this spectral range (Table 1).