What is the single line emitter found with JWST/NIRSpec multiobject coronagraphy?

A real line with no visible source

The object is a single emission line detected at 1.60 μm, measured with a line flux of [2.5 ± 0.3] × 10⁻¹⁸ erg s⁻¹ cm⁻² and a full width at half maximum of 13.2 ± 2.0 Å in the observed frame. The combined measurement has a signal-to-noise ratio of 8.4. The line appears in four separate dither stacks and in two position angles roughly 89° apart, which is the paper's main argument that it cannot be an instrumental artifact.

What makes the detection unusual is what is missing. The line sits at a sky position where the JADES NIRCam F150W image — reaching a 5σ depth of 4.6 nJy in 0.3-inch apertures — shows nothing. The F150W filter fully covers 1.60 μm, and the line flux alone should have produced a detectable signal at about magnitude 29.6, roughly 0.7 mag brighter than the JADES limiting magnitude. The source is therefore either far fainter in continuum than in the line, or it was not present when the imaging was taken.

Opening nearly every shutter to remove preselection

Multi-Object Coronagraphy for Slitless Spectroscopy (MOCSS) is the observing strategy that made the search possible. NIRSpec normally observes a preselected list of targets through its microshutter array, which biases any sample toward objects already known from imaging or color selection. MOCSS instead leaves almost all shutters open, turning the instrument into a wide-field slitless spectrograph, and closes only the shutters covering known low-redshift sources that would otherwise contaminate the spectra.

The masking targeted galaxies detected at wavelengths below 9600 Å in HST F814W images, which removes bright foreground objects. Accounting for failed shutters and the intentional closures, about 90% of the microshutter array area remained open. Observations used the G140M/F100LP pairing, covering roughly 0.97 to 1.89 μm at a resolving power near 1000, and the field was observed at two orientations with eight exposures each.

Filtering 42 candidates down to seven secure lines

Line candidates were found directly in the dispersed images rather than extracted at the positions of known sources, using a source-extraction routine that kept regions with more than five contiguous pixels above 3σ. Each detector stack contained roughly 200 to 300 candidates. Within one orientation, a real line should appear in both stacks, offset by the large dither of about 200 pixels along the dispersion direction and 2 pixels across it; candidates whose centroids matched within 4 pixels were kept, yielding 21 candidates per orientation and 42 in total.

The harder step was assigning wavelengths. In slitless spectroscopy the wavelength of a given pixel depends on where the source lies in the field, so a single dispersed position can correspond to different wavelengths for different objects. The team mapped each candidate to every shutter that could host it in each orientation and kept only lines that landed on the same shutter in both orientations with wavelengths agreeing within 6 Å. That left seven secure lines. Six have hosts already identified in JADES imaging; the seventh is the object discussed here.

Why the equivalent width limit is so restrictive

Equivalent width compares line flux to the underlying continuum, so a line with almost no detectable continuum has a very large equivalent width. Using the 5σ imaging depth of 4.6 nJy, the authors derive a 3σ continuum limit of 3.23 × 10⁻²² erg s⁻¹ cm⁻² Å⁻¹. Combined with the measured line flux, that gives a lower limit on the observed-frame equivalent width of 7620 Å.

This number is the paper's central discriminator. A normal star-forming galaxy produces strong lines with equivalent widths of tens to a few hundred Å in the observed frame at these redshifts; 7620 Å is described by the authors as near the ceiling of expectation. The limit depends on the assumption that the source was present and persistent when the imaging was taken. If the line came from a transient that was absent during the earlier imaging, the continuum constraint weakens dramatically, and the 3σ spectroscopic noise alone gives a much smaller observed-frame limit of about 41 Å.

Four standard identifications that do not fit

If the line belongs to a galaxy, its redshift follows from which transition it is. The authors test the four strong lines commonly seen in galaxy spectra and find each one strained. Reading the line as [OII] would place the galaxy at z = 3.29 with a rest-frame equivalent width above about 1770 Å; the [OII] doublet separation of about 11.6 Å in the observed frame should also be at least partly resolvable at the survey's roughly 10 Å resolution. Reading it as [OIII] gives z = 2.20, Hα gives z = 1.44, and Lyα gives z = 12.16.

Two further constraints apply across these cases. The spectrum shows no additional emission lines anywhere in the covered wavelength range, with 3σ line-flux limits between about 8.2 and 10.6 × 10⁻¹⁹ erg s⁻¹ cm⁻² depending on wavelength. And the extreme equivalent width required by the imaging non-detection is hard to reconcile with ordinary stellar populations. The authors note that if the source is real and persistent, it would have to be a very young, very metal-poor galaxy.

AGN and transients shift the burden of explanation

The paper also considers active galactic nuclei and transient phenomena, which would naturally explain a line without an imaging counterpart. A flaring star in the Milky Way is effectively ruled out: the Brackett (13–4) transition at 1.61 μm is close in wavelength, but that scenario predicts Brackett-12 at 1.64 μm, which is not observed.

A Type IIn supernova at z ≈ 0.25 could produce narrow lines through shock interaction with circumstellar material, with the line read as Paβ; however, Case B recombination predicts Paγ at 1.37 μm with a Paβ/Paγ flux ratio near 1.6, which should be detectable above 3σ, though dust could suppress it. A tidal disruption event offers another route: in dust-obscured cases, optically thick gas can suppress forbidden lines such as [OIII] and dust can weaken Hβ, allowing Hα to appear essentially alone. The two orientations were taken about 88 days apart and the line fluxes agree between them, which does not exclude a slow-evolving transient but does argue against a rapidly moving solar system object.

The redshift is the measurement still missing

The paper's own priority is clear: the decisive next step is spectroscopy that can establish the redshift and reveal or rule out companion lines. The authors single out JWST/NIRSpec integral field unit observations because they would simultaneously cover this line-emitter and the nearby galaxies, three of which lie within about one arcsecond of the inferred position and could in principle host an ionizing source illuminating a gas cloud.

The most useful way to read the result is as a fork defined by the imaging non-detection. If the source was persistent, the 7620 Å equivalent-width limit stands and the object must be an extreme, metal-poor system — possibly at z ≈ 12.2 if the line is Lyα. If the source was a transient absent during the earlier imaging, the equivalent-width constraint relaxes to about 41 Å and the difficulty moves to explaining the line profile and the missing companion lines. Either branch leaves the same object as an unusually extreme case, and the paper does not choose between them.

Sources used in this answer