A lingering ultraviolet glow at the Milky Way’s poles

A faint glow, and a difficult accounting problem

Even a seemingly empty patch of sky can shine. Ultraviolet light may come from distant galaxies, hot gas in the Milky Way, or stars whose light has been scattered by interstellar dust. Astronomers call the combined glow the cosmic ultraviolet background. Working out what contributes to it matters: a signal left over after familiar sources are accounted for could point to an unrecognized source, but only if the known contributions and measurement uncertainties have been handled well.

In The Cosmic Ultraviolet Background at the Galactic Poles, Jayant Murthy and colleagues revisit this problem using archival observations from NASA’s Galaxy Evolution Explorer (GALEX). They focus on regions around the north and south Galactic poles, at Galactic latitudes above 80 degrees in absolute value. These high-latitude sightlines generally contain less dust than regions closer to the Milky Way’s plane, making it easier to distinguish dust-scattered starlight from a more uniform background.

Separating dust from the rest

The team used GALEX maps in two ultraviolet bands: far ultraviolet (FUV), near 1530 angstroms, and near ultraviolet (NUV), near 2360 angstroms. They compared the measured ultraviolet brightness with a map of dust-related reddening, a measure of how much dust lies along each line of sight. In broad terms, light that rises and falls with reddening can be associated with dust; the remaining baseline is treated as an offset.

They also modeled the scattered light directly. In a single-scattering model, ultraviolet light from stars encounters dust and is redirected toward the observer. Two parameters describe important aspects of that process: the albedo, or the fraction of light scattered rather than absorbed, and the asymmetry factor, which indicates how strongly scattering favors the forward direction. The model combines stellar information, maps of dust, and assumptions about how dust is distributed along the line of sight.

A statistical detail matters here: nearby map pixels are not independent measurements. Dust-scattered light is spatially correlated, so treating every pixel as a separate data point would make the results look more precise than they are. The authors instead used a block-resampling method, repeatedly resampling patches of sky and varying measurement values within their stated errors. They report that the results are not especially sensitive to block sizes over roughly 60–150 arcminutes.

What the analysis finds

For the north Galactic pole, the reported 1-sigma ranges suggest moderately reflective dust grains that scatter light strongly forward. In the FUV, the fitted albedo is 0.54–0.71 and the asymmetry factor is 0.74–0.83. In the NUV, the ranges are 0.66–0.73 and 0.71–0.77, respectively. The south-pole estimates are broadly consistent, but less tightly constrained in places; the paper attributes this in part to the smaller range of reddening sampled there.

The fitted offsets are about 277–284 photon units in the FUV and 513–520 in the NUV for the north-pole results. Here, a photon unit means a photon flux per unit area, time, solid angle and wavelength. The authors report that the offsets show no clear spatial structure in the selected regions, with variations of about 20–30 photon units—similar to the expected GALEX measurement uncertainty.

The offset is not automatically a new or exotic signal. The authors subtract estimates for known contributions, including light from unresolved external galaxies and, in the FUV, emission from ionized carbon; they also include an estimate of hydrogen two-photon emission. After this accounting, they report an unexplained excess of 165 ± 18 photon units in the FUV and 346 ± 34 in the NUV. The resulting NUV-to-FUV ratio is 2.1 ± 0.5.

A possible match is not a detection

The paper compares that brightness and colour with predictions for Axion Quark Nuggets (AQNs), a proposed dark-matter candidate. In this model, matter interacting with antimatter nuggets could produce broadband radiation. The authors say the observed signal appears consistent with predictions for some AQN masses and temperatures. That is a comparison between a measurement and a model—not evidence that AQNs have been detected, or that dark matter is the source. Other explanations remain possible, and the paper notes that some excluded sky regions show extra emission that may come from local sources such as molecular-hydrogen fluorescence or nearby stars.

There are further reasons for caution. The dust properties are inferred through a simplified single-scattering model, and the paper itself notes that dust-scattering results depend on uncertain source and dust geometry. The uncertainties on individual map fits also require care because the data points are correlated. Finally, estimates of the known background components carry their own uncertainty, which affects how much signal remains after subtraction.

This work is an arXiv preprint, so its claims should be treated as provisional rather than as settled consensus. The study sharpens the accounting of ultraviolet light at the Galactic poles and offers a way to test an AQN interpretation using the two bands’ intensity ratio. Establishing the origin of the excess will require independent scrutiny and further observations or analyses that can distinguish the proposed explanation from other sources of diffuse ultraviolet light.