Measuring energy production in the Universe over all wavelengths and all time
TThe Realm of the Low-Surface-Brightness UniverseProceedings IAU Symposium No. 355, 2019D. Valls-Gabaud, I. Trujillo & S. Okamoto, eds. © Measuring energy production in the Universe over allwavelengths and all time
Simon P. Driver
International Centre for Radio Astronomy Research (ICRAR), University of Western Australia, 35 Stirling Highway, Crawley,Perth, WA6009, Australia
Abstract.
The study of the extragalactic background light (EBL) is undergoing a renaissance. New results from very highenergy experiments and deep space missions have broken the deadlock between the contradictory measurements in the opticaland near-IR arising from direct versus discrete source estimates. We are also seeing advances in our ability to model the EBLfrom γ -ray to radio wavelengths with improved dust models and AGN handling. With the advent of deep and wide spectroscopicand photometric redshift surveys, we can now subdivide the EBL into redshift intervals. This allows for the recovery of theCosmic Spectral Energy Distribution (CSED), or emissivity of a representative portion of the Universe, at any time. With newfacilities coming online, and more unified studies underway from γ -ray to radio wavelengths, it will soon be possible to measurethe EBL to within 1 per cent accuracy. At this level correct modelling of reionisation, awareness of missing populations or light,radiation from the intra-cluster and halo gas, and any signal from decaying dark-matter all become important. In due course,the goal is to measure and explain the origin of all photons incident on the Earth’s surface from the extragalactic domain, andwithin which is encoded the entire history of energy production in our Universe. Keywords. cosmology: diffuse radiation, cosmology: observations, galaxy: evolution
Figure 1.
A compendium of recent EBL measurements, mainly based on data assembled by Hill, Masui & Scott (2018) andalso including the Andrews et al. (2018) UV-far-IR model (purple line), the recent semi-analytic Shark model (Lagos et al. γ -ray to radio model (gold line). a r X i v : . [ a s t r o - ph . C O ] F e b Simon P. Driver
1. Introduction
The extra-galactic background light (EBL, see review by Mattila & V¨ais¨anen 2019), represents the radiation incidentfrom a steradian of extragalactic sky onto a square metre of the Earth’s atmosphere, i.e., the photon-flux that originatesfrom outside our Galaxy. Encoded in the EBL is the entirety of photon production pathways that have existed fromthe Big Bang to the present day. This is a remarkable concept, as to measure and model the EBL, is to explain allphoton-energy production over all time.Fig. 1 shows a recent compendium of EBL measurements, primarily taken from Hill, Masui & Scott (2018), butaugmented with some additional data from COBE/Planck (Odegard et al. et al. (2018) (purple line), a recent semi-analytic model from Shark (Lagos et al. since decoupling.At about 10 per cent of the CMB, are the Cosmic Optical Background (COB) and the Cosmic Infrared Background(CIB). There are three key processes at play here (see for example Gilmore et al. et al. et al. et al. γ -ray background (CGB; see for example Ajello et al. et al. et al. (2018).Because of the energy dominance of the COB and CIB these are, after the CMB, the most studied wavelengthregions and the main focus of this article. However, before moving on, it is worth advocating the need for further workin particular soft x-ray, extreme ultraviolet, radio, and very high ( > T eV ) scales. All of these windows, except forthe EUV (see Cooray et al. Messier mission — see these proceedings — will extend down to 0.15micron). One motivation to push further into theEUV might be the recent study of Mattila et al. (2017a) who used dark cloud observations to detect an anomalousUV photon-flux increasing into the UV (Mattila et al.
2. The optical controversy
Most obvious from Fig. 1 is a significant disparity in the COB and to some extent the CIB. Fig. 2 shows a zoom inof this region which now includes the errorbars associated with the measurements. These estimates predominantly fallinto two camps: Observations from direct background measurements (direct-EBL); and observations from integratedgalaxy counts (IGL-EBL). The former method should capture all the photon flux, whereas the latter only captures thephoton-flux from discrete detectable sources, i.e., galaxies and quasars. The fact that the direct-EBL measurementscan be a factor of 4—10 × above the IGL-EBL, can have a number of possible explanations. The least exciting is thatone method is simply in error. The most exciting is that there are significant unknown photon-production pathwaysoccurring outside of the detectable galaxy population. Possibilities for the latter might include missing populations ofdiffuse galaxies, excessive stripped gas, photon-production from a diffuse component of the IGM via some unknownprocess, or more exotic possibilities such as decaying dark-matter or dark-mater/matter interactions. Regardless, it isclearly important to understand the nature of this discrepancy.Direct-EBL estimates are typically taken from measurements of the absolute sky background from well calibrateddetectors, most notably data from the Hubble Space Telescope , which sits above the Earth’s atmosphere. However,this still requires further background subtraction of the Zodiacal Light (Zodi) and any Diffuse Galactic Light (DGL). nergy production over all time Figure 2.
A zoom in of the COB and CIB region from Fig. 1, but including data from the Very High Energy experiments anddeep space probes Pioneer 10/11 and New Horizons. These seem to corroborate the IGL-EBL measurements.
Fig. 3 highlights this issue by showing the flux of various components including that from the Earth’s night sky,the Zodi (dependent on Ecliptic coordinates), the Diffuse Galactic Light (dependent on Galactic coordinates), theEBL (isotropic), and the component of the EBL from reionisation (isotropic but with a highly uncertain shape andamplitude). Very roughly the Zodi is about 10 per cent of a dark site moon-less night sky, the EBL 1-10 per cent ofthe Zodi, the EBL and DGL comparable (with the latter highly dependent on direction), and the reionisation signal1 per cent of the EBL (see Zemcov et al. et al. et al. δ log N ( m ) /δm = 0 . .
4, the gradient of maximum contribution to the overall luminosity density. Assuming the counts continue to declinein a monotonic fashion the IGL-EBL is therefore bounded and measurable with a relatively small extrapolation error.Fig. 4 highlights how an IGL-EBL constraint is made for a particular waveband. The left-panel shows the raw galaxynumber-counts from a number of surveys, and the right-panel shows the contribution of each magnitude interval to theoverall IGL-EBL measurement. The IGL-EBL is recovered by integrating under the right-panel curve with a model-fit,or an extrapolated spline-fit (see for example Driver et al.
Figure 3. ( main panel ) The spectral energy distributions of various backgrounds including, the night sky, zodiacal light, theextragalactic background light, the diffuse Galactic light, and the expected signal of reionisation. The figure is shown in atypicalunits of Surface Brightness in AB mag/arcsec . ( top panel ) The attenuation of external radiation by the Earth’s atmosphere.
3. Very High Energy and deep space missions to the rescue
Recently, the disparity in Fig. 2 appears to at least be partially resolved from the inclusion of two new lines ofargument: Very High Energy (VHE) constraints, and measurements from the Pioneer 10 & 11 (Matsuoka et al. et al. et al. et al. ∼ e + e − -pairs. This inte-grated interaction along the pathlength of the TeV radiation results in a decrement in the received Blazar spectrumat energies where the interaction is expected to be maximum. This method requires the adoption of a COB model,and the VHE data constrains the amplitude of the model. Recently the H.E.S.S. and MAGIC teams reported the needfor a small upward normalisation over the IGL-EBL data of just 20 per cent, with about a 20 per cent uncertainty,i.e., essentially consistent. More recently the largest study to date from the Fermi-LAT collaboration of ∼
750 Blazarsreported full consistency with the IGL data (Fermi-LAT Collaboration 2018). The VHE results are shown as thethree solid bands on Fig. 2. Current efforts are also underway to constrain not only the normalisation, but the actualCOB spectral energy distribution shape. This is significantly harder but starting to place useful constraints on theCOB SED, see for example Biteau & Williams (2015) and VERITAS Collaboration (2019).The Pioneer 10/11 and New Horizons deep space probes, both contain relatively stable imaging cameras and thesecameras which can be used for direct-EBL measurements (see Matsuoka et al. et al. et al. et al. et al. (2021) have now decreased the uncertainty on the EBL-IGL measurements to < nergy production over all time Figure 4.
The IGL method relies on assembling galaxy number-counts ( left panel ) and then integrating the flux contribution ofeach magnitude interval ( right panel ). Note that in most optical/near-IR bands the contribution is well bounded with minimalextrapolation required, and that the peak contribution is at relatively intermediate magnitudes. Improvements will come notfrom deeper or wide data but improved methodologies in flux measurements and sample sizes at the mid-mag range.
4. Entering the era of precision EBL studies
The recent VHE and deep space data appear to provide compelling evidence that the IGL-EBL measurements prettymuch represents the full COB to within ∼
20 per cent (Driver et al. modest population of diffuse low surface brightness galaxies, or other photon-production pathways. This then motivates thereduction of the errors through direct-EBL, VHE and IGL-EBL methods to somewhere around the 1 per cent level.A goal which would represent a remarkable empirical feet but also entirely attainable with the next decade.
5. The SkySURF program
While the discrepancy in the HST direct-EBL estimates and the IGL-EBL data appears to be resolved, the expla-nation for the discrepancy is still not known but cannot be extragalactic. There are a number of obvious possibilities.The first is a limited understanding of the Zodiacal Light in the inner Solar System, which in turn implies a limitin our understanding of the Solar System dust distribution and properties. The second is an additional source ofcontamination in HST data, plausibly a component of Earth-shine given HST’s relatively low-orbit and the tendencyto pack orbits close to the Earth-limb. A more radical and exciting prospect might be additional optical radiationemanating from the Galactic Halo, or a brighter than expected contribution from the DGL.Led by Prof Rogier Windhorst and Dr Rolf Jansen at Arizona State University, a team of US and Australianscientists are looking to address this by reprocessing the entire HST Advanced Camera for Surveys and Wide-fieldCamera 3 archive, as part of an HST Cycle 27—29 SkySURF Archival program (AR-15810). The goals are to conductboth direct-EBL sky measurements, as well as obtain refined medium/deep galaxy number-counts using our new sourcefinding code ProFound (Robotham A. et al. u to K and from AB 10 th to 30 th magnitude. However, significant systematicsneed to be overcome related to issues such as star-galaxy separation, galaxy fragmentation, over-blending, and theaforementioned sensitivity to diffuse populations. Simon P. Driver Figure 5.
The latest cosmic star-formation history plot used as the starting point for our model and including the recentVHE constraints from the Fermi-LAT Collaboration (2018).
6. The state of EBL modelling: phenomenological and apriori
In addition to SkySURF, there are a number of upcoming wide and deep missions which will contribute significantnew data in the coming years. In particular high-resolution imaging data from new space-platforms: Euclid, JWST,and Roman, and extremely wide ground-based data from LSST. These should provide the statistical power to reachthat 1 per cent goal for the COB within a decade.At the 1 per cent level interesting astrophysics arises and in particular the direct contribution of reionisation becomesquantifiable, as does the contribution from Intra Cluster Light (ICL), Intra Group Light (IGL), Intra Halo Light (IHL),tidal streams, and a myriad of other physical processes.At the present time there are a number of approaches to modelling the EBL, which can be grouped under apriori models and phenomenological models. Examples of the apriori approach arise from numerical, semi-analytic andhydro-dynamical simulations such as Gilmore et al. (2012), Inoue et al. (2013), Cowley et al. (2019), Lagos et al. (2019), and Baes et al. (2019). Examples of the phenomenological model include our own work Andrews et al. (2018);Koushan et al. (2021), along with those of Dom´ınguez et al. (2011) and Khaire & Srianand (2019).Here in brief we summarise our own model, described in full in Andrews et al. (2018), which starts with two simpleaxioms: (1) the formation of today’s spheroid and bulge stars dominated the cosmic star-formation history at highredshift, and (2) AGN growth and activity is closely linked to spheroid and bulge star-formation. The basis for theformer is that the oldest known stars are found in the Galactic Centre (Zoccali et al. et al. et al. et al. (2019), and a metallicity evolution whichlinearly tracks the star-formation history. With these two axioms and the choices above, we can produce the purple lineshown in Fig. 1. This maps the currently measured COB/CIB portion of the EBL to within 30 per cent accuracy, i.e., nergy production over all time Figure 6.
A static frame from our EBL/CSED movie, which shows the build up of the EBL over time. ( upper panel ) the EBLdata as observed today with the EBL as it would be if observed when the Universe was 7.5Ga. ( lower panel ) the instantaneouscosmic spectral energy distribution (CSED) at an age of 7.5Ga, showing the contribution to the EBL from various componentsat this time-step (as indicated). The endpoint of the movie is the purple curve shown on Fig.1 comparable to the measurement error, across the entire optical to far-IR wavelength range, with only some tweakingof the AGN component required to match the UV data.Essentially this provides a self-consistency test by which the adopted Cosmic Star-formation History (see Fog. 5),under the most simplistic assumptions, fully predicts the present day EBL. Moreover the model not only predicts theEBL, but energy (photon) production at any time over the past 13 billion years. This is highlighted by the snapshotfrom our linked movie (Fig. 6), which shows the EBL (upper) and Cosmic Spectral Energy Distribution (CSED; lower)at a time when the Universe was 7.5Ga. In the upper panel we show the redshift zero EBL measurements, i.e., theendpoint to where the EBL will eventually grow to match, and the EBL as it would be observed at an age of 7.5Ga,i.e., in its fairly fledgling state. In the lower panel we show the instantaneous energy being produced by the fourcomponents (as indicated). The CSED, sub-divided into spheroid, disc obscured and unobscured AGN contributions,is potentially far more powerful than the EBL, as it dissects the EBL across time and providing a clear falsifiableprediction as our measurements improve. In due course comparisons between CSED measurements (Andrews et al.
Simon P. Driver2017) and models (Andrews et al. et al. et al. et al.
7. Prognosis and future directions
The prognosis for the EBL and its subdivision into time slices are remarkably good, with significant effort underwayon a number of fronts which will rapidly advance both the empirical measurements and our capacity to model the data.Critical will be the complement of upcoming deep and complete spectroscopic campaigns to allow for the deconstructionof the EBL into the CSED.With the analysis of the VST KiDS, VISTA VIKING and HST SkySURF data as they stand, we should be able toattain a ∼ et al. et al. γ -rays to radio waves. Thistoo is likely to be transformed over the next decade with various deep and wide radio surveys allowing us to constructcomparable IGL-EBL constraints over a broad wavelength range from 20cm to 10m. eROSITA will also improve uponprevious measurements of the CXB, particularly at the soft x-ray end where some discrepancies are seen (see Fig. 1).However, perhaps the most exciting prospect, is the potential to also construct CSED measurements into the x-rayand radio domains through the stacking of x-ray and radio data at the locations of known galaxies, or groups ofgalaxies, to directly measure the diffuse x-ray and radio continuum contributions as a function of time.We thank the organisers for a very enjoyable and productive meeting, and look forward to continuing these discus-sions over what looks to be a very busy and productive decade ahead. References
Ajello M., et al.
ApJ , 800, 27Andrews S.J., et al.
MNRAS , 470, 1342Andrews S.J. et al.
MNRAS , 474, 898Abeysekara A.U. et al.
ApJ , (accepted, arXiv:1910.00451)Baes et al.
MNRAS , 484, 4069Berstein R. 2007
ApJ , 666, 663Biteau J., Williams D.A. 2015,
ApJ , 812, 60Bock J et al.
SPIE , 4144, Astronomical Opportunities From The Outer Solar SystemBrdar V., Kopp J., Liu J. & Wang X-P. 2018,
PhyRvL , 120, 061301Cappelluti N. et al.
ApJ , 837, 19Cooray A. Gong Y., Smidt J., Santos M.G. 2012,
ApJ , 756, 92Cooray A. 2016,
RSOS , 3, 150555Cowley W.I., et al.
MNRAS , 487, 3082Dale D.A., et al.
ApJ , 784, 83Davies L.J.M., et al.
MNRAS , 480, 768Dom´ınguez A. et al.
MNRAS , 410, 2556Driver S.P. et al.
ApJ , 678, 10Driver S.P. et al.
ApJ , 827, 108Driver S.P., et al.
ESO Messenger , 175, 46Dunne L. et al.
Nature , 424, 285Fermi-LAT Collaboration 2019,
Science , 362, 1031Gebhardt K. et al.
ApJ , 539, 13Gilmore R.C., Somerville R.S., Primack J.R., Dom´ınguez A. 2012,
MNRAS , 422, 3189Hill R., Masui K. & Scott D. 2018,
Applied Spectroscopy , 72, 663Inoue Y., et al.
ApJ , 768, 197Lagos C., et al.
MNRAS , 489, 4196Lauer T., et al.
ApJ , 906, 77Khaire V. & Srianand R. 2019,
MNRAS , 484, 4174Koushan S., et al.
MNRAS , in pressMagorrian J., et al.
Contemporary Physics , 60, 23 (arXiv:1905.08825) nergy production over all time Mattila K. et al.
MNRAS , 470, 2133Mattila K. et al.
MNRAS , 470, 2152Matsuoka Y., Ienka N., Kawara K., & Oyabu S 2011
ApJ , 736, 119Odegard N. et al.
ApJ , 877, 40Robotham A. et al.
MNRAS , 476, 3137Zemcov M. et al.
Proc of Science , arXiv:1101.1560Zemcov M. et al.
Nature Communications , 8, 15003Zoccali M. et al.
A&A , 457, 177
Discussion
A. Dom´ınguez:
What is the contribution of AGN to the EBL ?
S. Driver:
In general the AGN are sub-dominant throughout, except in the UV and mid-infrared. However, we dohave some unexplained flux in the ultraviolet which we currently fixfix