Cosmoglobe DR2. II. CIB monopole measurements from COBE-DIRBE through global Bayesian analysis

We derive new constraints on the CIB monopole spectrum from reprocessed COBE-DIRBE sky maps with lower instrumental and astrophysical contamination than the legacy DIRBE maps. These maps have been generated through a global Bayesian analysis framework that simultaneously fits cosmological, astrophysical, and instrumental parameters, described in a series of papers referred to as Cosmoglobe DR2. We have applied this method to the DIRBE Calibrated Individual Observations, complemented by selected HFI and FIRAS sky maps to break key astrophysical degeneracies, as well as WISE and Gaia compact object catalogs. In this paper, we focus on CIB constraints that result from this work. We report detections of an isotropic signal in five out of the ten DIRBE bands (1. 25, 2. 2, 3. 5, 140, and 240 m). For the 2. 2 m channel, we find an amplitude of 133\, nW\, m^-2\, sr^{-1}, 36% lower than reported from the official DIRBE release. For the 240 m channel, we find 103\, nW\, m^-2\, sr^{-1}, 26% lower than the official DIRBE release. We interpret these lower values as resulting from improved zodiacal light and Galactic foreground modeling. For the bands between 4. 9 and 100 m, the presence of excess radiation in solar-centric coordinates reported in a companion paper precludes the definition of lower limits. However, the analysis still provides well-defined upper limits. For the 12 m channel, we find an upper 95% confidence limit of 103 nW\, m^-2\, sr^{-1}, a factor of four lower than the corresponding legacy result of 468 nW\, m^-2\, sr^{-1}. The results presented in this paper redefines the state-of-the-art CIB monopole constraints from COBE-DIRBE, and it provides a real-world illustration of the power of global end-to-end analysis of multiple complementary data sets which is the foundational idea of the Cosmoglobe project.