Diffuse supernova neutrino background is detectable in Super-Kamiokande

The diffuse supernova neutrino background (DSNB) provides an immediate opportunity to study the emission of MeV thermal neutrinos from core-collapse supernovae. The DSNB is a powerful probe of stellar and neutrino physics, provided that the core-collapse rate is large enough and that its uncertainty is small enough. To assess the important physics enabled by the DSNB, we start with the cosmic star formation history of Hopkins and Beacom (2006) and confirm its normalization and evolution by cross-checks with the supernova rate, extragalactic background light, and stellar mass density. We find a sufficient core-collapse rate with small uncertainties that translate into a variation of $\pm 40\%$ in the DSNB event spectrum. Considering thermal neutrino spectra with effective temperatures between 4–6 MeV, the predicted DSNB is within a factor 4–2 below the upper limit obtained by Super-Kamiokande in 2003. Furthermore, detection prospects would be dramatically improved with a gadolinium-enhanced Super-Kamiokande: the backgrounds would be significantly reduced, the fluxes and uncertainties converge at the lower threshold energy, and the predicted event rate is $1.2–5.6\mathrm{events}{\mathrm{yr}}^{-1}$ in the energy range 10–26 MeV. These results demonstrate the imminent detection of the DSNB by Super-Kamiokande and its exciting prospects for studying stellar and neutrino physics.

Figure 1

Redshift evolution of the comoving SFR density. Data derived from various indicators are included as labeled, all scaled to a Salpeter IMF [50]. The majority are from the compilation of HB06 [37], with additional data from LBG [63] and $\mathrm{H}\alpha$ measurements [64]. We plot the fiducial CSFH (thick solid curve) and our generous adopted uncertainty range (thin solid curves). The curves take into account SFR data at higher redshifts [37, 65] that are not shown here, where we focus on the lower redshifts most relevant for the DSNB.

Figure 2

Evolution of the core-collapse supernova rate (top panel) and type Ia to CCSN ratio (bottom panel), shown against distance in Mpc (for $z<0.1$) and redshift (for $z>0.1$). Data as labeled; filled symbols indicate data or limits we adopt, and empty points indicate published rates we treat as lower limits. Error bars show statistical errors, except for horizontal bars on Dahlen et al. points that show bin size. In the top panel, predictions from the CSFH are shown by the solid curve, where the uncertainty band propagates from Fig. 1. The SFR measurements, scaled to a CCSN rate, are shown in the background in light brown (small squares). The BG IMF has been adopted; using other IMFs causes differences of only a few percent. In the bottom panel, the dashed lines denote conservative upper limits from the nonobservation of any type Ia supernovae within 10 Mpc [17]. The 10 Mpc point has been very conservatively selected from Ref. 17, and the rate has been further corrected downwards (see text).

Figure 3

The observed EBL spectrum. Filled symbols are based on integrated galaxy counts, while empty symbols and the black lines at high wavelength represent absolute measurements. In the UV to NIR, we also show the gamma-ray upper limits placed by HESS (red solid curve) [99], MAGIC (blue dashed curve) [101], and Edelstein et al. 2000 (filled downward triangle at $0.1\mu \mathrm{m}$) [155]. We show three shaded regions representing the minimum (dark shading), nominal (light shading) and maximum (lightest shading) EBL. The integrated total EBL from these are listed. The integrated galaxy counts shown include (in order of increasing $\lambda$): Gardner et al. 2000 (red) [90], Madau and Pozzetti 2000 (green) [91], Levenson and Wright 2008 (blue) [92], Fazio et al. 2004 (yellow) [93], Metcalfe et al. 2003 (red) [94], Elbaz et al. 2002 (blue) [95], Papovich et al. 2004 (green) [96], Dole et al. 2006 (yellow) [97], and Frayer et al. 2006 (purple) [98]. Absolute measurements shown include (in order of increasing $\lambda$): Cambresy et al. 2001 (dark blue) [85], Wright 2001 (purple) [86], Dwek and Arendt 1998 (dark green) [87], Lagache et al. 2000 (green) [88], and Hauser et al. 1998 (red) [89]. For the results of Bernstein et al. [81, 82, 83, 84], we show $2\sigma$ upper limits (filled downward-pointing triangles), nominal values (empty circles), and lower limits (filled upward-pointing triangles), all connected by dashed lines. We show the diffuse sky measurements of Matsumoto et al. 2005 [102] as small brown symbols and the FIR measurements by Fixsen et al. 1998 [156] connected by black lines.

Figure 4

DSNB flux spectrum for emitted neutrino spectra as labeled. For each spectrum, two curves are plotted representing the full range of uncertainties due to astrophysical inputs (the fiducial prediction lies in between). The shadings indicate backgrounds, with origins as labeled. Decays of invisible muons and spallation products would be reduced in a gadolinium-enhanced SK, opening the energy region 10 MeV and above to a rate-limited DSNB search; see Fig. 5.

Figure 5

DSNB event rates at SK (flux spectra weighted with the detection cross section) against positron energy. Note the linear axis. We hatch in the 2003 upper limit by SK, $<2$ events $(22.5\mathrm{kton}\mathrm{yr}{)}^{-1}$ in the energy range 18–26 MeV. The limit applies to all spectra (see text). In a gadolinium-enhanced SK, decays of invisible muon and spallation products would be reduced, opening up the energy range $\gtrsim 10\mathrm{MeV}$ for DSNB search (unshaded region).