An Archival Optical Counterpart Search for Extragalactic Fast X-Ray Transients Discovered by Einstein Probe

Abstract: Extragalactic fast X-ray transients (eFXTs) represent a rapidly growing class of high-energy phenomena, whose physical origins remain poorly understood. With its wide-field, sensitive all-sky monitoring, the Einstein Probe (EP) has greatly increased the discovery rate of eFXTs. The search and identification of the optical counterparts of eFXT are vital for understanding their classification and constraining their physical origin. Yet, a considerable fraction of eFXTs still lack secure classifications due to the absence of timely follow-up observations. We carry out a systematic search of publicly available optical survey data and transient databases (including the Zwicky Transient Facility, ZTF, and the Transient Name Server, TNS) for optical counterparts to eFXT candidates detected by EP. In this paper, we describe our ongoing program and report the first results. Specifically, we identified the eFXT EP240506a to be associated with a UV/optical counterpart, AT 2024ofs. Spectroscopy of its host galaxy with VLT yields a redshift of $z = 0.120 \pm 0.002$. By combining archival survey data with early-time multiwavelength observations, we find that the luminosity and light-curve evolution of AT~2024ofs are consistent with a core-collapse supernova origin. From detectability simulations, we estimate a local event rate density $$ρ<em>{0}=8.8^{+21.2}</em>{-3.9}\ \mathrm{yr^{-1}\, Gpc<sup>{-3}}$$ for EP240506a-like events, and completeness-corrected rate of about $36$--$78\ \mathrm{yr^{-1}\ Gpc<sup>{-3}}$ for EP-detected X-ray transients associated with supernovae. Our results demonstrate the potential of EP to uncover prompt high-energy emission from core-collapse supernovae and underscore the critical importance of timely follow-up of future eFXT events.

• The paper presents a systematic archival search and identification of optical counterparts for extragalactic fast X-ray transients (eFXTs) discovered by the Einstein Probe (EP), utilizing data from the Zwicky Transient Facility (ZTF) and the Transient Name Server (TNS).
• Out of 5,235 EP-detected eFXT candidates, the study identified 16 matches, including nine high-confidence eFXTs and seven unverified transient sources, with most occurring within known associations and a discovery of a previously unknown transient EP240506a/AT 2024ofs.
• Monte Carlo simulations confirm high-confidence matches present a low contamination rate, while unverified sources are largely false positives.

Motivation and scope

Extragalactic fast X-ray transients (eFXTs)—non-repeating, soft (0.5–4 keV) X-ray flashes lasting seconds to hours—have historically been difficult to classify because their coarse localizations and short durations rarely permit timely multiwavelength follow-up. The Einstein Probe (EP), with its ~3600 deg² Wide-field X-ray Telescope (WXT) field of view, has increased the known eFXT population to over 90 events within its first 1.5 years of operation, five of which are already associated with broad-lined Type Ic supernovae. This paper describes a systematic archival program by Liang et al. to recover optical counterparts of EP-detected eFXTs in ZTF and Transient Name Server (TNS) data, and reports its first major result: the identification of the eFXT EP240506a with the supernova candidate AT 2024ofs (2602.06321).

Sample construction and cross-matching

The X-ray sample comprises all high-confidence eFXT candidates (S/N > 6–7, no prior X-ray history, non-Galactic association, fluence enhancement confined to a single ~3.6 ks orbit) plus unverified sub-threshold sources (S/N ≈ 5–6) detected between January 9, 2024 and June 20, 2025—95 high-confidence candidates and 5,140 unverified sources in total. The optical sample was built from the ZTF alert stream via the Lasair broker using Sherlock contextual classifications (NT/SN/ORPHAN), excluding star-like Pan-STARRS matches, supplemented by TNS objects reported after EP's launch; this yields 12,346 ZTF candidates and 33,048 TNS objects.

Cross-matching adopts a conservative spatial radius of $r \le 3.5'$ (the WXT localization accuracy during commissioning) and a temporal window of $0 < \delta t \le 30$ days between the X-ray trigger and optical discovery, motivated by rest-frame core-collapse rise times of $\lesssim 20$ days. The procedure returns 16 matches: nine high-confidence eFXTs and seven unverified sources.

Classification of matched candidates

Most high-confidence matches correspond to previously published events: GRB associations at $z = 4.859$ (EP240315a), $z = 1.411$ (EP241030a), $z = 3.315$ (EP250226a), and $z = 1.52$ (EP250427a); and SN-associated transients including EP240414a/SN 2024gsa ($z = 0.401$, Ic-BL), EP250108a/SN 2025kg ($z = 0.176$, Ic-BL), and EP250304a/SN 2025fhm ($z = 0.2$, Ic-BL). None of the seven unverified-source matches survive scrutiny: one is consistent with a cataclysmic variable (WXT J160346+193555), and the remaining six are attributed to chance spatial/temporal alignments or instrumental artifacts.

Monte Carlo simulations quantify the contamination rate. Resampling mock optical transient catalogs that preserve the real sky and time distributions, the authors estimate via 200 realizations that at $0 < \delta t \le 30$0 days roughly five random matches are expected over the full candidate sample—consistent with the seven unrelated associations found—whereas only ~0.1 random associations are expected among the high-confidence eFXTs. This validates treating the high-confidence matches as predominantly genuine while flagging the unverified sample as dominated by false positives.

EP240506a / AT 2024ofs

The central new result is EP240506a, an "orphan" eFXT whose significance (S/N = 7.8) fell just below the on-board trigger threshold of 8.0 during commissioning, so no automated FXT follow-up occurred. The flare, truncated in default telemetry after passage through the South Atlantic Anomaly but recovered by manual reprocessing, has $0 < \delta t \le 30$1 s, a peak flux of $0 < \delta t \le 30$2, and a peak luminosity of $0 < \delta t \le 30$3 at the host redshift. The spectrum is fit by an absorbed power law with photon index $0 < \delta t \le 30$4. A Swift/XRT observation ~45 hr later yields only a limiting flux of $0 < \delta t \le 30$5. Non-detection by Fermi-GBM constrains $0 < \delta t \le 30$6 keV and $0 < \delta t \le 30$7 erg; these limits exclude much of the Type-I GRB parameter space on the Amati relation, though not all of it.

Despite extensive follow-up in the first three days, no optical/IR counterpart was found. AT 2024ofs emerged in ZTF and Pan-STARRS imaging ~19 days after trigger, coincident with a marginal Swift/UVOT UVM2 detection ($0 < \delta t \le 30$8 mag, S/N = 2.5) at ~2 days post-trigger—a detection near the sensitivity limit whose reality cannot be fully confirmed against deeper 2025 template imaging, as the authors acknowledge. Forced photometry of ZTF and ATLAS data reveals earlier emission at ~7.7 days. VLT/FORS2 spectroscopy of the host galaxy SDSS J141550.45−163937.0 gives $0 < \delta t \le 30$9, placing the transient 11.9 kpc (projected) from the host nucleus. The chance-coincidence probability of the X-ray–optical match is $\lesssim 20$0, and the host-galaxy association probability is $\lesssim 20$1, both strongly favoring a physical association.

The compiled light curve shows a longer rise than AT 2018cow or SNe Ibn but faster evolution than normal SNe Ibc, indicating an intermediate-timescale stripped-envelope event. Arnett-model fitting with the Redback package yields $\lesssim 20$2, $\lesssim 20$3, $\lesssim 20$4 erg, and $\lesssim 20$5, powered by $\lesssim 20$6Ni decay with an adopted mass of $\lesssim 20$7. The authors are explicit that the free Bayesian fits suffer from strong parameter degeneracies and drift toward unphysical regions due to sparse sampling around peak; the quoted parameters rely on fixing the nickel mass by analogy with EP240414a and EP250108a, and no spectrum of the transient itself exists. The X-ray luminosity exceeds that of XRO 080109 and is comparable to GRB 060218, EP240414a, and EP250108a, placing EP240506a in the relativistic shock-breakout regime and supporting the picture of a continuum of collapsar-related transients spanning ordinary core-collapse SNe to low-luminosity GRBs.

Event rate density

Using detectability simulations (detectable to $\lesssim 20$8 at S/N = 7) and the standard $\lesssim 20$9 formalism with the WXT field of view, duty cycle $z = 4.859$0, and 1.5 yr baseline, the authors derive a local rate density for EP240506a-like events of $z = 4.859$1. Combining the four SN-associated EP transients known by June 2025 gives $z = 4.859$2–$z = 4.859$3, rising to a completeness-corrected $z = 4.859$4–$z = 4.859$5 once the fact that only about one third of high-confidence eFXTs have reliable redshifts is accounted for. This rate is orders of magnitude below the local core-collapse SN rate of $z = 4.859$6 and broadly consistent with estimates for low-luminosity GRBs ($z = 4.859$7). The implication is that EP is probing a rare subset of core-collapse events—those with weak relativistic jets or cocoon-powered emission—rather than a generic SN channel.

Limitations and open questions

Several caveats bear directly on the results. The classification of AT 2024ofs rests on photometry alone; without a transient spectrum, the SN subclass and energy source remain inferred rather than measured. The Arnett-model parameters are not statistically robust, given missing peak coverage and fixed nickel mass. The marginal UVM2 detection at 2 days may be spurious. The rate estimate assumes EP240506a arises from core collapse and inherits the Poisson uncertainty of small-number statistics, with asymmetric errors spanning nearly a factor of three. Finally, the completeness correction depends on the still-unpublished redshift completeness fraction of the EP eFXT catalog (Wu et al., in prep), so the corrected rate should be regarded as provisional.

Conclusion

This paper establishes a reproducible archival framework for recovering optical counterparts of poorly localized EP transients and demonstrates its value through the recovery of EP240506a/AT 2024ofs, a previously unrecognized eFXT–SN association at $z = 4.859$8 with prompt X-ray luminosity in the relativistic shock-breakout regime. The case illustrates concretely how sub-threshold triggers and short-lived follow-up can cause even nearby, luminous SN counterparts to be missed in real time, and it quantifies both the genuine-match yield (~0.1 expected random associations among high-confidence eFXTs) and the contamination floor (~5 random matches across the full sample at $z = 4.859$9 days). The open question the work leaves most sharply posed is whether denser photometric sampling and prompt spectroscopy of future EP-discovered eFXTs will confirm the collapsar-continuum interpretation for the population as a whole, or reveal distinct subclasses currently blended together by sparse data.