NEXUS: A Search for Nuclear Variability with the First Two JWST NIRCam Epochs

Zachary Stone, Yue Shen, Ming-Yang Zhuang, Lei Hu, Justin Pierel, Junyao Li, Adam J. Burgasser, Jenny E. Greene, Zhiwei Pan, Alice E. Shapley, Fengwu Sun, Padmavathi Venkatraman, Feige Wang

First listed 2025-09-23 | Last updated 2025-09-23

Abstract

The multi-cycle JWST Treasury program NEXUS will obtain cadenced imaging and spectroscopic observations around the North Ecliptic Pole during 2024-2028. Here we report a systematic search for nuclear variability among $\sim 25\,$k sources covered by NIRCam (F200W+F444W) imaging using the first two NEXUS epochs separated by 9 months in the observed frame. Difference imaging techniques reach $1σ$ variability sensitivity of 0.18~mag (F200W) and 0.15~mag (F444W) at 28th magnitude (within 0".2 diameter aperture), improved to $0.01$~mag and $0.02$~mag at $<25$th magnitude, demonstrating the superb performance of NIRCam photometry. The difference imaging results represent significant improvement over aperture photometry on individual epochs (by $>30\%$). We identify 465 high-confidence variable sources among the parent sample, with 2-epoch flux difference at $>3σ$ from the fiducial variability sensitivity. Essentially all these variable sources are of extragalactic origin based on preliminary photometric classifications, and follow a similar photometric redshift distribution as the parent sample up to $z_{\rm phot}>10$. While the majority of these variability candidates are likely normal unobscured AGNs, some of them may be rare nuclear stellar transients and tidal disruption events that await confirmation with spectroscopy and continued photometric monitoring. We also constrain the photometric variability of ten spectroscopically confirmed broad-line Little Red Dots (LRDs) at $3\lesssim z \lesssim 7$, and find none of them show detectable variability in either band. We derive stringent $3σ$ upper limits on the F444W variability of $\sim 3-10\%$ for these LRDs, with a median value of $\sim 5\%$. These constraints imply weak variability in the rest-frame optical continuum of LRDs.

Short digest

NEXUS uses two NIRCam epochs (Wide-1.1 in Sep 2024 and Deep-1 in Jun 2025; 9-month baseline) to run a nuclear-variability search across ~25k sources in F200W and F444W with PSF-matched difference imaging. The pipeline reaches ~0.18 mag (F200W) and ~0.15 mag (F444W) sensitivity at 28 mag (0.2″ aperture), improving to ~0.01–0.02 mag at <25 mag, and outperforms single-epoch aperture photometry by >30%. They flag 465 high-confidence extragalactic variables whose photo-z distribution mirrors the parent sample to zphot>10. Crucially, ten spectroscopically confirmed broad-line LRDs at 3≲z≲7 show no detectable variability, yielding 3σ F444W limits of ~3–10% (median ~5%), implying weak rest-optical continuum variability; rare nuclear transients remain to be confirmed with continued monitoring and spectroscopy.

Key figures to inspect

• Figure 1: Map the Wide-1.1 and Deep-1 overlap to see the exact area used for variability, noting which pointings contribute to F200W/F444W depth and where cadence gaps could bias detections.
• Figure 2: Inspect the cross-convolution/SFFT workflow and the decorrelated SFFT difference to verify PSF matching; use the SNR map and background Gaussian fit to judge residual systematics and masking efficacy around artifacts (e.g., wisps, claws).
• Figure 3: Compare source cutouts between epochs and subtraction methods to see real-variable residuals versus artifacts; check AUTO magnitudes and Δ-flux S/N, and note cases where SFFT performs better for bright, compact nuclei.
• Figure 4: Examine method-to-method Δmag consistency across apertures (including the 0.2″ nuclear aperture) and identify saturated cases masked in SFFT; look for any aperture-dependent biases that could affect the variable catalog.