A High Incidence of Mid-infrared Variability in Local Ultraluminous Infrared Galaxies

The gist The MIR variability in local ultraluminous infrared galaxies reveals compact, intensely active nuclear regions, suggesting that these variations are driven by either tidal disruption events or intrinsic changes in their active galactic nuclei.

Sample and Data

The study utilizes approximately 11 years of multi-epoch photometry from the NEOWISE catalog to investigate mid-infrared (MIR) variability in local ultraluminous infrared galaxies (ULIRGs). The sample is based on the Infrared Astronomical Satellite (IRAS) Point Source Catalog Redshift (PSCz) survey and the 1 Jy ULIRG sample at 60 µm. The WISE magnitudes are converted into fluxes from Vega magnitude system with zero-point fluxes of fν,0 =309.540 and 171.787 Jy for the W1 and W2 bands, respectively. The mission survey continued post-cryogenic NEOWISE Reactivaion mission (Mainzer et al. 2014) which monitored the entire sky in the W1 and W2 bands for approximately 11 years (2013–2024).

Methodology for Variability Detection

The variability analysis involves several steps to ensure robust detection of MIR fluctuations. Then, quality cuts are imposed against single epoch measurements for single-epoch phtometry with qual flag ≥ 5, and moon masked = 0. The NEOWISE single-epoch photometry consists of ‘epochs’ taken roughly every six months, and each target has about 22 epochs in total. Because the focus is on half-year or longer time scale variability, photometry is binned for each epoch by deriving the weighted average and standard errors to obtain the binned flux and flux errors for each epoch, while conducting 3 sigma clipping (we adopt 3 times median absolute deviation) at the same time to remove outliers. Unbiased Pearson r correlation coefficients are calculated using equations involving unbiased covariance CW1,W2 and unbiased variances σW1 and σW2. The criteria for detecting MIR variability are set as r > 0.75, σW1 > 0.010, and σW2 > 0.034.

Results of Variability Analysis

The analysis identified 30 variable ULIRGs with statistically significant MIR variability. The difference between maximum and minimum W2 (4.6 µm) band luminosity (∆LW2) of the 30 variable ULIRGs ranges from ∆LW2 = 7 × 1042 to 5 × 1044 erg s−1. Of these, the ∆LW2 of 25 variable ULIRGs out of 30 are greater than ∆LW2 = 1 × 1043 erg s−1. This threshold surpasses the MIR luminosity range observed in known supernovae (SNe; L3.6 µm and L4.5 µm < 1042.3 erg s−1). Light curves of the variable ULIRGs show diverse behaviors, with variability observed on timescales of a few years, suggesting that the MIR-emitting regions are compact, with sizes of order < ∼ 1 pc.

Discussion and Interpretation

The distribution of r differs significantly by optical type: H II galaxies and LINERs are mostly found at lower values (r < ∼ 0.65), whereas Seyfert galaxies tend to have higher r. This suggests that the variability is associated with AGN activity rather than stellar processes, although the influence of sample bias is unknown. The 25 ULIRGs exhibiting ∆LW2 > 1043 erg s−1 cannot be explained by previously known supernovae observed in the MIR. This indicates that variable ULIRGs with ∆LW2 > 1043 erg s−1 are not powered by stellar activity, but by intrinsic AGN luminosity variations or TDEs. All variable ULIRGs with ∆LW2 > 1043 erg s−1 show at least one independent AGN signature, supporting this interpretation. The variability is observed on timescales of a few years, suggesting that the MIR-emitting regions are compact, with sizes of order < ∼ 1 pc.

AGN Signatures

Optical classifications from Nardini et al. 2010 identify 14 out of 25 variable ULIRGs as Seyfert 1 or Seyfert 2 galaxies. Among the remaining non-Seyfert galaxies, IRAS 20551-4250, IRAS 12514+1027, and IRAS 09320+6134 (UGC 05101) are identified as AGNs by X-ray observations. All variable ULIRGs with ∆LW2 > 1043 erg s−1 fulfill at least one AGN criterion. In combination with the fact that these ULIRGs have ∆LW2 values that exceed the luminosity of the brightest known supernovae by a factor of ∼5, they most likely host AGNs.

TDE Candidates

Variable ULIRGs include previously reported TDE candidates in IRAS F01004-2237 and a TDE candidate in IRAS 05189-2524. Both of these candidates are successfully detected with the method, indicating that our approach is also effective for identifying TDE candidates. For IRAS 17028+5817, two apparent transient events exhibit light curves superficially resembling those of repeating partial TDEs.

Summary

The study analyzes 11 years of NEOWISE photometry for 164 ULIRGs from Nardini et al. 2010 to investigate MIR variability as a probe of buried AGNs. Significant variability was detected in 30 ULIRGs, among which 25 show amplitudes exceeding ∆LW2 = 1043 erg s−1. These include previously reported TDE candidates and hard X-ray detected AGNs. Because the variability is observed on timescales of a few years, the MIR-emitting regions is suggested to be compact, with sizes of order < ∼ 1 pc. The large amplitudes cannot be explained by SNe, and are most likely due to intrinsic AGN luminosity changes or TDEs occurring in AGN-hosting systems. All variable ULIRGs with ∆LW2 > 1043 erg s−1 show at least one independent AGN signature, supporting this interpretation. This work demonstrates that MIR variability can define a new field in uncovering buried nuclear activity in ULIRGs.

Table 1 Summary

Table 1 summarizes the variable sources, showing that for the 25 variable ULIRGs with ∆LW2 > 1043 erg s−1, at least one AGN criterion is fulfilled. The most luminous source in this group has a ∆LW2 of 5 × 1044 erg s−1. The presence of AGN signatures is a key finding, supporting the conclusion that the variability originates from active black holes.

Acknowledgments and Data Availability

This study is based on group work carried out at the GalaxyIGM Workshop 2023 in Hamamatsu. The authors thank Takumi S. Tanaka, Kohei Inayoshi, Takashi J. Moriya, Hiroya Umeda, Masami Ouchi, Yusei Koyama, Hirofumi Noda, Yuichi Harikane, Tomokazu Kiyota, Shogo Yoshioka, Shohei Aoyama, Nanase Harada and Shigeo Kimura for helpful discussions and useful suggestions. This publication makes use of data products from the Wide-field Infrared Survey Explorer which is a joint project of the University of California, Los Angeles and the Jet Propulsion Laboratory/California Institute of Technology and NEOWISE which is a project of the Jet Propulsion Laboratory/California Institute of Technology.

1. AI systems can be improved to detect buried AGN activity by employing MIR variability as an effective diagnostic tool, specifically by calculating unbiased Pearson r correlation coefficients (Equation 1) using binned photometry from multi-epoch WISE data.

2. The improved system will be capable of distinguishing between AGN activity and star formation by applying the criterion Variability provides a clear distinction between AGN activity and star formation, which is supported by the observation that AGNs are known to vary across the entire electromagnetic spectrum, whereas star formation typically extends over kiloparsec scales.

3. The system can perform luminosity thresholding to filter out contamination from bright supernovae, using the criterion that sources brighter than this threshold "are unlikely to originate from supernova activity since even the most luminous supernova observed in the MIR, SN 2010jl, have peak luminosity of L[3.6 µm] and L[4.5 µm] < 1042.3 erg s−1."

4. The AI can classify ULIRGs based on their variability correlation coefficients (r) to suggest the origin of activity, noting that "H II galaxies and LINERs are mostly found at lower values (r < ∼ 0.65), whereas Seyfert galaxies tend to have higher r."

5. The system can identify TDE candidates by successfully applying the methodology used in this study, as shown by the finding that previously reported TDE candidates in IRAS F01004-2237 (Dou et al. 2017; Sun et al. 2024) and a TDE candidate in IRAS 05189-2524 (Reynolds et al. 2022) are successfully detected with our method.

6. The system can constrain the physical size of MIR-emitting regions, as variability studies can provide much tighter constraints, potentially limiting the MIR-emitting region to a few parsecs or less, based on observed timescales of a few years variability.

We explore mid-infrared (MIR) variability in local ultraluminous infrared galaxies (ULIRGs; infrared luminsoity L IR>10 12 L) utilizing the about 11 years of photometry from the NEOWISE multi-epoch catalog of Wide-field Infrared Survey Explorer (WISE). We identify 30 variable ULIRGs with statistically significant MIR variability. The variability is observed on timescales of a few years, implying that the MIR-emitting regions are compact (1 pc). The difference between maximum and minimum W2 (4.6 μ m) band luminosity (ΔL W2) of the 30 variable ULIRGs range from ΔL W2 = 7 times10 42 to 5 times 10 44 erg s-1. The ΔL W2 of 25 variable ULIRGs out of 30 are greater than ΔL W2 = 1 times10 43 erg s-1, surpassing the MIR luminosity range observed in known supernovae (SNe; L 3.6 μm and L 4.5 μm < 10 42.3 erg s-1). Therefore, the MIR variabilities in these 25 ULIRGs are most likely driven by tidal disruption events (TDEs) or intrinsic changes in their active galactic nuclei (AGN) torus emission. Our sample includes hard X-ray detected AGNs (e.g., UGC 05101) and previously reported TDE candidates (IRAS F01004-2237, IRAS 05189-2524). All 25 also exhibit at least one AGN signature(s) besides the MIR variability, suggesting that even if the MIR variability originates from TDEs, the black holes responsible are likely AGNs. Our results suggest that MIR variability is an effective tool for detecting buried AGNs and highlights the intense nuclear activity in ULIRGs.