WorksPhysics and Astronomy
ASAS-SN Rates V: Tidal Disruption Event Volumetric Rates and Luminosity Functions by Optical Spectroscopic Type, X-ray Luminosity, and Host-Galaxy Properties
Abstract
We present volumetric rates and luminosity functions (LFs) of tidal disruption events (TDEs) based on 21 TDEs from the All-Sky Automated Survey for Supernovae (ASAS-SN) between 2014 and 2024. We calculate completeness corrections from injection-recovery simulations of the ASAS-SN light curves using individual TDE templates. We find a total volumetric rate of for . We also measure the rates of two growing TDE subpopulations. The one off-nuclear TDE in our standard sample implies (% of the total rate). We measure the rate of recurrent TDEs (rTDEs) to be (% of the total rate) from the four rTDEs in our standard sample, increasing to (%) when the well-studied ASASSN-14ko is included. Galaxies with substantial (>0.1%) burst-mass fractions and steeply falling star-formation histories host the majority of TDEs. We recover the known overrepresentation of TDEs in post-starburst galaxies, by a factor of relative to matched control galaxies, and a weaker enhancement of in quiescent Balmer-strong galaxies. TDE-H+He events dominate the volumetric TDE rate, with TDE-H, TDE-He, and TDE-Featureless events comprising <15% of nearby TDEs. The -band LFs are statistically indistinguishable when split at the median X-ray to UV/optical blackbody luminosity ratio of 1%, corroborating unification models where the observed multi-wavelength emission is set primarily by viewing angle effects. Extrapolating the measured TDE LFs to higher luminosities and more massive black holes supports a TDE origin for luminous ambiguous nuclear transients and extreme nuclear transients.
Introduction
The vast majority of massive galaxies in the local universe host a supermassive black hole (SMBH) in their nucleus (e.g., [99] (Kormendy & Richstone 1995; Magorrian et al. 1998; Gültekin et al. 2009; Kormendy & Ho 2013)). A small fraction, (e.g., [99, 129]), are accreting matter at a substantial rate and can be detected as active galactic nuclei (AGNs; e.g., Rees 1984; Antonucci 1993; Reynolds 1997; Stern et al. 2005; Kewley et al. 2006). The remaining majority of inactive, or quiescent, SMBHs are nearly impossible to detect, save for a few dozen nearby and/or massive SMBHs for which stars (e.g., Kormendy et al. 1996; Ghez et al. 2005; Gebhardt et al. 2011) and/or gas (e.g., Ford et al. 1994; Atkinson et al. 2005) within the SMBH’s sphere of influence can be resolved, and two where emission near the horizon has been imaged ([51] (Event Horizon Telescope Collaboration et al. 2019)). Fortunately, the recent growth of time-domain astronomy has provided a promising new probe of quiescent SMBHs, including many at cosmological distances. These are the flares resulting from a tidal disruption event (TDE), when a star passes too close to an SMBH and is ripped apart as its self-gravity is overwhelmed by the intense tidal forces of the SMBH (e.g., Rees 1988; Phinney 1989; Evans & Kochanek 1989; Ulmer 1999). TDE candidates have been observed across the electromagnetic spectrum (e.g., Auchettl et al. 2017; Alexander et al. 2020; Gezari 2021), with many showing emission in multiple energy bands. Optical selection has proven particularly effective, with more than 100 optical TDEs discovered to date by surveys like the All-Sky Automated Survey for Supernovae (ASAS-SN; Shappee et al. 2014; Kochanek et al. 2017), the Asteroid Terrestrial Impact Last Alert System (ATLAS; Tonry et al. 2018), the Gravitational-wave Optical Transient Observer (GOTO; Dyer et al. 2024), the Panoramic Survey Telescope and Rapid Response System (Pan-STARRS; Chambers et al. 2016), and the Zwicky Transient Facility (ZTF; Bellm et al. 2019).
Although TDEs provide undeniable evidence of a dormant SMBH, several complications hinder a robust connection between TDE observables and SMBH parameters. These include various properties of the disrupted star (e.g., MacLeod et al. 2012; Guillochon & Ramirez-Ruiz 2013; Mockler et al. 2019; Law-Smith et al. 2019) and the SMBH (e.g., Ulmer 1999; Mockler et al. 2019; Gafton & Rosswog 2019) that affect the observed TDE emission. An accurate estimate of the rate at which TDEs occur is also fundamental to their use as probes of the SMBH population. Theoretically, TDE rates depend primarily on the diffusion of stars into the “loss cone” (e.g., Frank & Rees 1976; Lightman & Shapiro 1977; Cohn & Kulsrud 1978; Stone et al. 2019). Initially, there was tension between the theoretically predicted rate of TDEs (; Magorrian & Tremaine 1999; Wang & Merritt 2004; Stone & Metzger 2016) and the observed rate of TDEs in optical bands (; van Velzen & Farrar 2014; Holoien et al. 2016a; Yao et al. 2023). However, the increasing number of TDE candidates identified in other wavebands, such as the X-ray (e.g., Sazonov et al. 2021; Grotova et al. 2025; Eyles-Ferris et al. 2025), infrared (e.g., Masterson et al. 2024), and radio (e.g., Dykaar et al. 2024), together with improved treatment of selection effects, helps alleviate this tension.
Detailed estimates of the volumetric TDE rate, particularly its dependence on SMBH mass, can shed light on important, broader questions about SMBHs. In the lower-mass intermediate-mass black hole (IMBH) regime, the SMBH mass function provides insight into BH seeding mechanisms in the early universe (e.g., Ricarte & Natarajan 2018; Greene et al. 2020). At the high-mass end, since Solar-like stars are directly captured, TDE rate constraints can inform studies on the disruption of evolved (e.g., MacLeod et al. 2012) and massive stars (e.g., Hinkle et al. 2025; Wiseman et al. 2025; Graham et al. 2026), and the spin of high-mass quiescent SMBHs (e.g., Kesden 2012; Stone et al. 2019; Huang & Lu 2024). Large TDE samples may enable population-level constraints on the spins of high-mass quiescent SMBHs (e.g., Shvartzvald et al. 2024; Karmen et al. 2026), which will complement AGN spin measurements (e.g., Brenneman & Reynolds 2006; Reynolds 2021).
The TDE rate is relatively flat across SMBH masses in the range between and , falling sharply above (Kochanek 2016; van Velzen 2018; Yao et al. 2023). This decline is approximately consistent with the so-called “Hills mass” (Hills 1975) for Solar-like stars, where disruption occurs within the SMBH event horizon and is unobservable. This population-level trend suggests that the optical TDE population is dominated by the tidal disruption of low-mass () main-sequence stars rather than non-TDE nuclear transients.
Existing estimates of the optical TDE luminosity function (LF) show a decline in the volumetric rate towards higher -band luminosities. The decline is reasonably approximated by a power-law of , with between 2.3 and 3 (van Velzen 2018; Lin et al. 2022; Charalampopoulos et al. 2023; Yao et al. 2023). More detailed modeling by Yao et al. 2023 shows that the optical LF is better fit as a broken power-law with a break luminosity of . The LF based on black-body models for the bolometric luminosity also shows a decline at higher luminosities, with a slope of (Lin et al. 2022; Yao et al. 2023). The optical TDE LF is also broadly similar to the X-ray TDE LF (e.g., Auchettl et al. 2018; Sazonov et al. 2021; Grotova et al. 2025). TDE rates as a function of host-galaxy properties probe the processes producing TDEs. This includes understanding the overrepresentation of TDEs in quiescent Balmer-strong (QBS) and post-starburst (PSB) galaxies (e.g., Arcavi et al. 2014; French et al. 2016, 2018; Law-Smith et al. 2017; Grau et al. 2018), for which several models have been proposed (e.g., Stone et al. 2018; Bortolas 2022; Mockler et al. 2023; Melchor et al. 2024; Wang et al. 2024; Teboul & Perets 2025), but no certain explanation exists (Shepherd et al. 2026).
ASAS-SN (Shappee et al. 2014; Kochanek et al. 2017; Hart et al. 2023) provides an ideal dataset to measure the rate of transients. It has surveyed nearly the entire visible sky since 2014 in an untargeted fashion, reaching a roughly daily cadence following its expansion in late 2017. Additionally, the shallow depth ( mag) allows for high spectroscopic completeness (Holoien et al. 2017b,c,a, 2019a; Neumann et al. 2023; Desai et al. 2026a). This wide-area, long-baseline, and high-cadence coverage complements rate estimates from deeper surveys. ZTF (e.g., Perley et al. 2020; Yao et al. 2023; Dimitiadis et al. 2025) observes only the northern sky, while ATLAS (e.g., Srivastav et al. 2022, 2026) monitored only the northern sky until its southern units became online in 2022, and both have had longer average cadences than ASAS-SN since its expansion. Previous papers in this series have presented the rates of Type Ia supernovae and their subtypes (Desai et al. 2024, 2026a), core-collapse supernovae of various types (Pessi et al. 2025; Pessi et al., in prep.), and kilonovae (Desai et al. 2026b).
Here, we leverage the long history of successful ASAS-SN TDE discoveries to compute the local volumetric TDE rate. In § 2, we describe the TDE sample, measurements of TDE and host-galaxy properties, and our refitting of the ASAS-SN light curves to compute peak apparent magnitudes. § 3 outlines the completeness corrections and rate computations. We next present the TDE rates and associated results on the -band LF with respect to spectroscopic type and Xray luminosity (§ 4), their dependence on UV/optical blackbody parameters (§ 5), and how they vary with the properties of their host galaxies (§ 6). Finally, we summarize our conclusions in § 7. Throughout this manuscript, we adopt a flat CDM cosmology with km s Mpc and , and define km s Mpc.
The TDE Sample
In this section, we describe our sample of TDEs discovered or recovered by ASAS-SN (§ 2.1) and the properties against which we compare their rates: the photometric and spectroscopic properties of the TDEs (§ 2.2), the properties of their host galaxies (§ 2.3), and the masses of their central SMBHs (§ 2.4). We also detail the creation of our light curve templates, their application to ASAS-SN data (§ 2.5), and the cuts that define our standard sample for rate calculations (§ 2.6).
ASAS-SN TDE Catalog
ASAS-SN has operated in two distinct phases: a -band era (2012–late 2017) and a -band era (late 2017–present). We begin our analysis in 2014, when ASAS-SN expanded to both hemispheres, and assign 2014–2017 to the -band era and 2018–2024 to the -band era. During the -band era, ASAS-SN achieved a typical single-visit depth of mag with a typical cadence of 2–3 days over much of the survey footprint. The transition to the -band significantly improved survey performance, reaching a typical single-visit depth of mag at a cadence of 1–2 days. The full ASAS-SN transient archive and its photometric pipeline are described in Hart et al. (2023).
Our TDE sample spans the full analysis period from UTC 2014-01-01 to UTC 2024-12-31, covering both the -band and -band eras. We compiled TDEs that were either discovered by ASAS-SN or independently recovered in ASAS-SN data within this time range. To identify independently recovered events, we cross-matched all TDEs reported to the Transient Name Server5 (TNS) with the ASAS-SN transient archive6. A TDE is considered “recovered” if another group first reported it, but it would have been independently flagged in the ASAS-SN data. In total, 11 TDEs were discovered by ASAS-SN and 14 were independently recovered, yielding 25 total TDEs. Table 1 and Table 2 list the TDE and host-galaxy properties of the full sample of 25 TDEs, respectively, and Appendix § 11 gives the source of the data.
| IAU Name | Discovery Nameᵃ | RA [hh:mm:ss] | Dec [dd:mm:ss] | Spec. Type | [mag] | [mag] | [mag] | [mag] | |||||
| ASASSN-14aeᵇ | 11:08:40.11 | +34:05:52.2 | TDE-H | 0.0436 | 0.058 | −0.101 | 16.96 | −19.43 | 43.96 ± 0.02 | 4.36 ± 0.01 | 14.83 ± 0.02 | < 41.1 | |
| ASASSN-14liᵇ | 12:48:15.23 | +17:46:26.44 | TDE-H+He | 0.0206 | 0.085 | −0.058 | 16.43 | −18.36 | 44.08 ± 0.11 | 4.66 ± 0.05 | 14.29 ± 0.11 | 43.4 | |
| ASASSN-15oiᵇ | 20:39:09.183 | −30:45:20.10 | TDE-He | 0.0479 | 0.225 | −0.111 | 16.18 | −20.58 | 44.29 ± 0.02 | 4.36 ± 0.01 | 15.00 ± 0.02 | 42.7 | |
| 2018dyb | ASASSN-18pg | 16:10:58.774 | −60:55:23.16 | TDE-H+He | 0.0179 | 0.763 | −0.044 | 15.42 | −19.75 | 44.237 ± 0.004 | 4.41 ± 0.01 | 14.88 ± 0.01 | 41.0 |
| 2018fyk | ASASSN-18ul | 22:50:16.090 | −44:51:53.50 | TDE-H+He | 0.059 | 0.044 | −0.159 | 17.60 | −19.39 | 44.60 ± 0.02 | 4.59 ± 0.01 | 14.69 ± 0.02 | 43.6 |
| 2018hyz | ASASSN-18zj | 10:06:50.871 | +01:41:34.08 | TDE-H | 0.0457 | 0.114 | −0.101 | 16.58 | −19.97 | 44.16 ± 0.01 | 4.32 ± 0.01 | 15.02 ± 0.01 | 41.4 |
| 2018zr | PS18kh | 07:56:54.537 | +34:15:43.61 | TDE-H | 0.071 | 0.154 | −0.107 | 17.58 | −19.99 | 43.86 ± 0.01 | 4.13 ± 0.01 | 15.24 ± 0.02 | 41.6 |
| 2019ahk | ASASSN-19bt | 07:00:11.546 | −66:02:24.14 | TDE-H | 0.0262 | 0.404 | −0.052 | 15.42 | −20.22 | 44.020 ± 0.002 | 4.230 ± 0.001 | 15.124 ± 0.002 | 40.8 |
| 2019azh | ASASSN-19dj | 08:13:16.945 | +22:38:54.03 | TDE-H+He | 0.0223 | 0.146 | −0.061 | 15.50 | −19.53 | 44.380 ± 0.004 | 4.580 ± 0.002 | 14.604 ± 0.004 | 43.0 |
| 2019dsg | ZTF19aapreis | 20:57:02.974 | +14:12:15.86 | TDE-H+He | 0.0512 | 0.340 | −0.145 | 17.93 | −19.05 | 44.44 ± 0.02 | 4.72 ± 0.03 | 14.36 ± 0.06 | 43.5 |
| 2019qiz | ZTF19abzrhgq | 04:46:37.880 | −10:13:34.90 | TDE-H+He | 0.0151 | 0.372 | −0.036 | 15.89 | −18.52 | 43.53 ± 0.01 | 4.362 ± 0.001 | 14.615 ± 0.004 | 40.7 |
| 2019vcb | ZTF19acspeuw | 12:38:56.380 | +33:09:57.28 | TDE-H+He | 0.088 | 0.058 | −0.204 | 18.46 | −19.42 | 14.84 ± 0.08 | 43.0 | ||
| 2020neh | ZTF20abgwfek | 15:21:20.090 | +14:04:10.52 | TDE-H+He | 0.0620 | 0.146 | −0.144 | 17.92 | −19.30 | 43.78 ± 0.01 | 4.38 ± 0.01 | 14.71 ± 0.01 | < 41.7 |
| 2020nov | ZTF20abisysx | 16:58:12.970 | +02:07:03.04 | TDE-H | 0.0826 | 0.373 | −0.122 | 18.12 | −20.00 | 43.37 ± 0.10 | 4.13 ± 0.04 | 15.00 ± 0.05 | 43.3 |
| 2020vwl | Gaia20etp | 15:30:37.800 | +26:58:56.89 | TDE-H+He | 0.0325 | 0.158 | −0.091 | 17.59 | −18.25 | 43.90 ± 0.16 | 4.65 ± 0.06 | 14.23 ± 0.15 | < 41.3 |
| 2022bdw | ATLAS22dth | 08:25:10.360 | +18:34:57.50 | TDE-H+He | 0.0378 | 0.155 | −0.107 | 17.51 | −18.65 | 44.30 ± 0.10 | 4.69 ± 0.05 | 14.35 ± 0.05 | < 41.6 |
| 2022dbl | ASASSN-22ci | 12:20:45.010 | +49:33:04.68 | TDE-H+He | 0.0284 | 0.063 | −0.075 | 17.18 | −18.28 | 43.69 ± 0.01 | 4.51 ± 0.03 | 14.34 ± 0.05 | 41.5 |
| 2022dsb | ASASSN-22cs | 12:42:21.743 | −22:30:14.04 | TDE-H+He | 0.0234 | 0.167 | −0.108 | 17.48 | −18.27 | 44.54 ± 0.03 | 4.75 ± 0.03 | 14.34 ± 0.06 | 43.3 |
| 2022hvp | ZTF22aagyuaco | 09:54:45.240 | +55:26:25.37 | TDE-He | 0.112 | 0.025 | −0.284 | 16.63 | −21.69 | 45.35 ± 0.01 | 4.535 ± 0.002 | 15.181 ± 0.004 | < 42.7 |
| 2022lri | ATLAS22pzn | 02:20:08.009 | −22:43:15.42 | TDE-H+He | 0.0328 | 0.061 | −0.094 | 17.73 | −18.02 | 44.31 ± 0.04 | 4.72 ± 0.05 | 14.29 ± 0.11 | 44.2 |
| 2023clx | ASASSN-23bd | 11:40:09.397 | +15:19:38.54 | TDE-H+He | 0.0111 | 0.103 | −0.019 | 16.46 | −17.03 | 42.740 ± 0.004 | 4.15 ± 0.02 | 14.65 ± 0.03 | 39.6 |
| 2023mhs | ZTF23aaqdjh | 13:43:15.681 | +19:15:00.90 | TDE-H+He | 0.0482 | 0.075 | −0.089 | 15.89 | −20.74 | 44.32 ± 0.01 | 4.202 ± 0.001 | 15.330 ± 0.004 | 42.4 |
| 2024as | ATLAS24acrt | 10:05:06.320 | +17:29:21.05 | TDE-H | 0.1225 | 0.120 | −0.241 | 18.53 | −20.14 | 43.85 ± 0.01 | 4.28 ± 0.10 | 14.95 ± 0.19 | 42.5 |
| 2024puv | ZTF18abxnovz | 23:31:11.930 | +19:52:32.20 | TDE-H+He | 0.0488 | 0.198 | −0.096 | 17.18 | −19.60 | 43.95 ± 0.02 | 4.24 ± 0.02 | 15.07 ± 0.04 | 42.3 |
| 2024tvd | ZTF22aaigsqr | 17:10:42.585 | +28:50:15.11 | TDE-H+He | 0.0449 | 0.164 | −0.092 | 17.13 | −19.44 | 43.719 ± 0.003 | 4.26 ± 0.02 | 14.92 ± 0.05 | 43.5 |
Table 1. Properties of the TDEs discovered or recovered in ASAS-SN (UTC 2014-01-01 to UTC 2024-12-31).
Note. — This table will be made available in machine-readable form on Zenodo upon publication. ᵃ Name assigned by the discovering survey. TDEs with an “ASASSN-” name were discovered by ASAS-SN; the others were independently recovered in ASAS-SN data. ᵇ Discovered during the ASAS-SN V-band era (2014–2017). For these events, is converted from the measured -band peak magnitude using the color of the TDE’s blackbody at peak (§ 2.5). Columns: is the -band Milky Way line-of-sight extinction ((Schlafly & Finkbeiner 2011)); is the blackbody -correction (Equation (2)); is the peak apparent magnitude from fitting the ASAS-SN light curve with the per-event light curve templates (§ 2.5); is the peak absolute magnitude from Equation (1). , , and are the UV/optical blackbody luminosity, temperature, and radius at peak (§ 2.2). is the maximum observed X-ray luminosity or upper limit, as appropriate, for each TDE (§ 2.2).
| IAU Name | Discovery Name | [Gyr] | [mag] | H EW [Å] | Lick H [Å] | ||||||
| ASASSN-14ae | |||||||||||
| ASASSN-14li | |||||||||||
| ASASSN-15oi | |||||||||||
| 2018dyb | ASASSN-18pg | ||||||||||
| 2018fyk | ASASSN-18ul | ||||||||||
| 2018hyz | ASASSN-18zj | ||||||||||
| 2018zr | PS18kh | ||||||||||
| 2019ahk | ASASSN-19bt | ||||||||||
| 2019azh | ASASSN-19dj | ||||||||||
| 2019dsg | ZTF19aapreis | ||||||||||
| 2019qiz | ZTF19abzrhgq | ||||||||||
| 2019vcb | ZTF19acspeuw | ||||||||||
| 2020neh | ZTF20abgwfek | ||||||||||
| 2020nov | ZTF20abisysx | ||||||||||
| 2020vwl | Gaia20etp | ||||||||||
| 2022bdw | ATLAS22dth | ||||||||||
| 2022bdl | ASASSN-22ci | ||||||||||
| 2022dbs | ASASSN-22cs | ||||||||||
| 2022hvp | ZTF22aagyuauo | ||||||||||
| 2022lri | ATLAS22ppz | ||||||||||
| 2023clx | ASASSN-23bd | ||||||||||
| 2023msh | ZTF23aaqdjhi | ||||||||||
| 2024as | ATLAS24acr | ||||||||||
| 2024pvu | ZTF18asbxmvoz | ||||||||||
| 2024vtd | ZTF22aaiqxs |
Table 2. Host-galaxy properties for the ASAS-SN TDE sample (UTC 2014-01-01 to UTC 2024-12-31).
We individually inspected every TDE in our sample and confirmed both its TDE nature and its spectroscopic type. Based on this, we remove TDE 2020afhd (ZTF20abwtifz), which has also been reclassified as a Bowen Fluorescence Flare (Arcavi et al. 2024). Bowen Fluorescence Flares (e.g., Trakhtenbrot et al. 2019; Gromadzki et al. 2019; Makrygianni et al. 2023) are characterized by strong He II emission and the fluorescent Bowen lines (e.g., O III , O III , N III , and N III , with most lines being blends; Bowen 1934, 1935; Netzer et al. 1985). They have primarily been observed in pre-existing AGNs and have been proposed to occur from enhanced accretion onto the SMBH, possibly from a TDE or unrelated perturbation to the accretion flow (e.g., Trakhtenbrot et al. 2019; Makrygianni et al. 2023). Bowen Fluorescence Flares have an uncertain physical origin (e.g., Sni˙gowska et al. 2025; Veres 2026) and share similarities with ambiguous nuclear transients (ANTs; Neustadt et al. [155]; Hinkle et al. 2022; Hinkle 2024; Wiseman et al. 2025), which may also be TDEs occurring in AGN hosts. Although TDEs may power a subset of these exotic accretion flares, we restrict our estimate of the volumetric rates to the well-established class of UV/optical TDEs. This exclusion also applies to events like ASASSN-14ko, which is likely a repeating partial TDE (e.g., Payne et al. 2021, 2022, 2023), and ASASSN-15lh, which may be a superluminous supernova (e.g., Dong et al. 2016; Godoy-Rivera et al. 2017). Neither event is a standard UV/optical TDE based on their observational signatures.
TDE Properties
We adopted spectroscopic classifications from TNS and the literature, primarily van Velzen et al. (2020). Optical TDE spectra are commonly grouped into four spectroscopic types (e.g., Leloudas et al. 2019; van Velzen et al. 2021; Hammerstein et al. 2023) defined by the broad emission features present in their optical spectra. TDE-H events are dominated by broad hydrogen Balmer emission (e.g., H, H), while TDE-He events are dominated by broad He II emission with no clearly detected broad Balmer component. TDE-H+He events exhibit both broad hydrogen and helium features, and may also show Bowen fluorescence features, including blended N III and O III, while TDE-Featureless events exhibit a smooth, blue continuum with no discernible broad transient features (e.g., Leloudas et al. 2019; van Velzen et al. 2021; Hammerstein et al. 2023; Yao et al. 2026a). Some TDEs show significant temporal evolution in their spectra (e.g., Nicholl et al. 2019; Short et al. 2020; Hinkle et al. 2021a); in these cases, we adopt the spectroscopic classification from the spectrum obtained closest to the optical peak. Our sample contains 6 TDE-H, 2 TDE-He, 17 TDE-H+He events, and no TDE-Featureless events.
We fit the TDE UV/optical SEDs from the Swift Ultraviolet and Optical Telescope (UVOT; Roming et al. [179], Hinkle et al. 2021b, 2022; Holoien et al. 2022). In brief, we corrected the photometry for Galactic extinction, subtracted the host-galaxy flux in each band, and modeled the SED as a single-temperature blackbody (BB), sampling using a Markov Chain Monte Carlo (MCMC) method implemented in emcee (Foreman-Mackey et al. 2013) with broad, flat priors on temperature ( K K) and radius ( cm cm). We created a bolometric light curve from a well-sampled ground-based light curve by linearly interpolating a bolometric correction between Swift epochs and assuming a constant temperature outside of the Swift coverage. We propagated the uncertainty on the assumed constant temperature into the bolometric light curve.
We estimated the peak UV/optical blackbody luminosity by fitting a Gaussian Process (GP) to each bolometric light curve. We adopted the maximum of the GP posterior mean as the peak luminosity, with its uncertainty from the GP posterior at that time, and estimated the uncertainty on the time of peak by bootstrap resampling of the light curve. When at least one blackbody-fit epoch fell on each side of the time of peak luminosity, we similarly fit a GP to the temperature and radius time series and evaluated them at peak. For several TDEs, Swift observations are only available after peak, notably days post-peak for TDE 2022lri. In these cases, because most TDEs exhibit comparatively modest temperature evolution (e.g., (van Velzen et al. 2021; Hinkle et al. 2020; Stein et al. 2026)), we instead adopted the temperature of the earliest available Swift epoch as the temperature at peak. Assuming a constant temperature between peak and the first Swift epoch, we derived the corresponding blackbody radius at peak luminosity using the Stefan–Boltzmann law.
In addition to the UV/optical parameters, we computed rest-frame unabsorbed 0.3–10 keV X-ray luminosities or upper limits for each TDE. When possible, we collected the maximum X-ray luminosity from the literature. Because the X-ray coverage of TDEs is heterogeneous, the maximum X-ray luminosity may not represent the intrinsic peak X-ray luminosity for every event. For the few sources without published X-ray constraints, we used the online Swift XRT tools7 to measure an X-ray count rate or upper limit from the stacked coverage. Adopting the Galactic (HI4PI Collaboration et al. [85]), we converted the estimated count rate to a flux assuming an absorbed 50 eV blackbody model, a commonly adopted approximation for soft X-ray TDE emission (e.g., (Auchettl et al. 2017; Kara et al. 2018; Saxton et al. 2020)), using WebPIMMS (Mukai 1993).
Host-Galaxy Properties
We fit optical spectra of the TDE host galaxies with Bagpipes to estimate the properties of any recent burst of star formation. Bagpipes ((Carnall et al. 2018, 2019)) uses the MultiNest nested sampling algorithm (Feroz et al. 2019) and the Bruzual & Charlot (2003) stellar population models. We extrapolated metallicity priors from the galaxy’s stellar mass by using the stellar mass–metallicity trend in Table 2 of Gallazzi et al. (2005). Full details on the Bagpipes priors and the complete fit instructions are in Shepherd et al. (2026).
We adopted a two-component star formation history (SFH): an old stellar component modeled by a delayed exponential (; Simha et al. 2014) and a young stellar component modeled by a double power-law (; Wild et al. 2020) representing either a burst or an ongoing period of star formation. Here, is the age of the galaxy, is the e-folding time, is the falling slope, and is the rising slope. Shepherd et al. (2026) provides more details about the choice of the SFH functional form and other Bagpipes parameters.
The equivalent width (EW) of the H emission line and the Lick H index from the host-galaxy spectra also hold important information on recent star formaNote. — This table will be made available in machine-readable form on Zenodo upon publication.
Columns: , , , , and are the Bagpipes burst age, stellar mass formed in the recent burst, stellar mass formed in the old component, declining slope of the burst SFH, and dust attenuation (§ 2.3). is the SMBH mass (§ 2.4); is the bulge Sérsic index; H EW and Lick H are the H emission equivalent width and Lick H absorption index.
tion. When possible, we adopt literature values (e.g., French et al. 2016; Graur et al. 2018; French et al. 2020a; Shepherd et al. 2026) for the TDE host galaxies. Otherwise, we calculated the H emission line EW using Bagpipes (see Shepherd et al. 2026) and the Lick H absorption line index using pyLick (Borghi et al. 2022). To facilitate a robust comparison, we created a control sample of galaxies from SDSS, whose H EWs and H indices were included in SDSS galSpec (Kauffmann et al. 2003; Brinchmann et al. 2004; Tremonti et al. 2004). The control sample was constructed by splitting the redshift range of the TDE hosts into five equal bins and splitting the stellar mass range of the TDE hosts into seven equal bins. We randomly selected 950 non-TDE-host galaxies per TDE host galaxy falling in each cell, such that a cell containing TDE hosts receives control galaxies. This yields 19,000 control galaxies across the cells.
In addition to modeling recent star formation in the TDE hosts, we estimated morphological parameters that may correlate with the physical processes that produce TDEs. We measured the host-galaxy bulge Sérsic indices using -band imaging from the DESI Legacy Imaging Surveys (Dey et al. 2019). We were primarily interested in the bulge light profile, since the disrupted stars come from near the SMBH. We use the ser sic_exp profile model of PySérsic (Pasha & Miller 2023), which decomposes the galaxy into an inner Sérsic component and an outer exponential component with the Sérsic index fixed to , representing the bulge and disk, respectively. The remaining analyses use the bulge Sérsic index. One caveat of the ser sic_exp profile model is that it cannot capture higher-order components such as bars, arms, or asymmetrical structures.
For each PySérsic fit, we provided the -band galaxy image, its uncertainties, the PSF model, and a source mask. We used the -band inverse-variance maps from the Legacy Survey to construct the uncertainties. Bad pixels in the inverse-variance maps were assigned the median value of the remaining pixels. We used Photutils (Bradley et al. 2025) segmentation and deblending to identify the host galaxy and mask contaminants. The model was convolved with the local -band Legacy Survey PSF in PySérsic before comparison with the image. We estimated the posterior distributions of the model parameters using the Laplace approximation implemented in PySérsic and report the median of the posterior distribution as the adopted bulge Sérsic index.
Supermassive Black Hole Masses
When available, we adopt SMBH mass estimates inferred from late-time TDE emission (Mummery et al. 2024; Mummery & van Velzen 2025). These estimates use TDE observables directly, although they remain model-dependent and may have systematic uncertainties distinct from those of host-galaxy scaling relations. SMBH masses inferred from late-time plateau emission are consistent with expectations from host-galaxy stellar mass (e.g., Mummery et al. 2024; Mummery & van Velzen 2025; Guolo et al. 2025), although studies have shown that SMBH masses estimated from TDE emission may be underestimated (Angus et al. 2026). Most of our SMBH mass estimates come from Mummery et al. (2024) or Mummery & van Velzen (2025) using the late-time plateau emission. For the few sources without existing estimates, typically recent events without extensive late-time monitoring, we use host-galaxy estimates from the - relation (e.g., Gültekin et al. 2009; Kormendy & Ho 2013) or stellar-mass scalings (e.g., McConnell & Ma 2013; Reines & Volonteri 2015). For all methods, the scatter on the SMBH mass estimates is at least 0.3 dex, and typically higher for the host-galaxy scalings. Nevertheless, the large mass bins used for the analysis and the significant statistical uncertainties make this a subdominant source of uncertainty in our TDE SMBH mass function.
Light Curve Templates and Fitting
TDE light curves exhibit diversity that no universal template can cleanly reproduce (e.g., van Velzen et al. 2021; Hammerstein et al. 2023). Therefore, rather than adopting a single class-wide template as in the supernova analyses of Papers I–III (Desai et al. 2024; Pessi et al. 2025; Desai et al. 2026a), we construct an individual template for each of the 25 TDEs in our sample. This per-TDE approach serves two purposes: it can provide more robust peak magnitude estimates than local interpolation when the peak is not well sampled, and it allows the injection-recovery simulations described in § 3.1 to compute the completeness individually for each event, capturing the full diversity of light curve morphologies. The sample is also small enough that running separate completeness simulations for each TDE is computationally tractable.
Each template was constructed by fitting the available -band or -band photometry. These data come primarily from ASAS-SN or the Zwicky Transient Facility (ZTF; Bellm et al. 2019), but in some cases, we used published photometry from individual discovery papers (e.g., Holoien et al. 2019b, 2020). We fit the data with the Modular Open Source Fitter for Transients (MOSFiT; Guillochon et al. 2017; Mockler et al. 2019) TDE model. An advantage of this approach is that we can extrapolate for events with either no pre-peak data and/or poor temporal coverage to both early and late times using the physically motivated, yet model-dependent, light-curve shape of the MOSFiT TDE model. The resulting normalized templates capture the full rise and decline morphology of each event and are scaled in the injection-recovery simulations by varying only the peak flux and time of peak.
To measure the ASAS-SN peak magnitude for each TDE, we fit its ASAS-SN light curve with the template derived from that same event using minimization with two free parameters: the time of peak, , and the flux at peak, . The resulting peak flux is converted to a peak apparent magnitude, .
We compute absolute magnitudes for all TDEs using
where is the apparent magnitude from light-curve fits in the respective discovery band, is the distance modulus computed from the redshift using the astropy.cosmology package ([10, 12] (Astropy Collaboration et al. 2018)), and is the Milky
Way foreground extinction from Schlafly & Finkbeiner (2011), assuming a Fitzpatrick (1999) extinction law with . We do not correct for host-galaxy extinction, but we examine the dependence of the rate on host-galaxy dust in § 6.2. We compute the -correction for each TDE from its peak blackbody temperature (§ 2.2), evaluated at the -band effective frequency ():
where is the Planck function (e.g., Hogg et al. 2002). For the -band discovered TDEs, we likewise convert to -band magnitudes using the color of the redshift blackbody, with Å.
Standard Sample
First, we require mag (for -band-era events, converted to as in § 2.5), which corresponds to the regime where our photometric completeness is well-characterized by the injection-recovery simulations (§ 3.1). This removes TDE 2019vcb (ZTF19acspeuw), TDE 2020nov (ZTF20abisysx), and TDE 2024as (ATLAS24acr). Second, we require a Galactic latitude to avoid regions of high and spatially variable Milky Way extinction where the photometric selection function is poorly constrained, following the Galactic-latitude criterion adopted in previous papers in this series (Desai et al. 2024, 2026a,b; Pessi et al. 2025). This removes TDE 2018dyb (ASASSN-18pg). Third, we restrict the redshift range to . No TDEs in our sample lie at , so peculiar-velocity errors in the distance estimates are at most (e.g., Tully et al. 2016), much smaller than the statistical uncertainties on our rates. We chose the upper limit of which includes every TDE passing the other two cuts. It defines the outer boundary of our effective survey volume, to which the volume correction of each TDE is normalized (§ 3.1). After applying these cuts, our standard sample consists of 21 TDEs.
Figure 1 shows the peak apparent and absolute magnitude distributions as a function of redshift for the full, classification-vetted 25 TDE sample. The standard sample of 21 TDEs is shown with filled markers and excluded events are shown with open markers.
Figure 1. Peak apparent (top) and absolute (bottom) magnitude versus redshift distributions of the 25 TDEs discovered or recovered by ASAS-SN. The filled markers and filled histograms show our standard sample of 21 TDEs, while open markers and dashed histograms include the TDEs excluded by our standard cuts on Galactic latitude , peak apparent magnitude , and redshift , as marked by the gray dotted lines. Marker shapes and color give the spectral classifications: purple diamonds for TDE-H, cyan squares for TDE-He, and orange circles for TDE-H+He. Points with an outer dotted shape denote ASAS-SN-discovered TDEs. The gray dashed curve in the bottom panel marks the envelope corresponding to mag based on the distance modulus for the redshift.
Rate Computations
To compute volumetric TDE rates and LFs, we must correct the observed sample for the survey’s incomplete detection and recovery of events. We follow the injection-recovery framework established in Papers I–III (Desai et al. 2024; Pessi et al. 2025; Desai et al. 2026a) and apply three completeness corrections: a photometric correction , the probability that ASAS-SN recovers a TDE at its observed peak magnitude; a volume correction , the fraction of the volume out to within which it could have been selected; and a spectroscopic correction , the fraction of nuclear transients that receive a spectroscopic classification. Each lies between 0 and 1, and each TDE is weighted by . § 3.1 describes and , and § 3.2 describes . § 3.3 combines them to derive the volumetric rates and § 3.4 gives the rates of off-nuclear and recurrent TDEs.
Photometric Completeness
To estimate the photometric completeness for each TDE, we employ an injection-recovery simulation following the methodology of Papers I–III. We extract ASAS-SN light curves at random sky coordinates distributed uniformly across the sky, and compute using only those with , matching the Galactic-latitude cut of our standard sample (§ 2.6). These baseline light curves empirically capture the true observing conditions of the survey – including photometric noise, observing cadence, weather interruptions, and seasonal gaps – without tracing any specific astrophysical source.
astrophysical source. Using random sky positions rather than galaxy centers has a negligible effect on the completeness. Bright transients have signal-to-noise ratios high enough that host flux does not change whether an epoch is detected, while for faint transients the noise is dominated by the sky rather than the host. For each sky position, we inject the per-TDE light curve templates described in § 2.5 across trials. For each trial, the redshift is drawn uniformly from the comoving volume out to the redshift at which the template would reach the simulation’s limiting magnitude, so that the injected events span the range of peak apparent magnitudes of interest, and the peak time is drawn uniformly from the survey baseline. The injected light curves account for cosmological time dilation at each trial redshift. A trial is considered recovered if at least one epoch satisfies the discovery or recovery criterion described below.
The single-epoch recovery probability , the probability that an epoch with a given SNR yields a detection that passes the ASAS-SN pipeline and vetting, differs between the two survey eras. For the -band era (2014–2017), we adopt the piecewise function derived in Paper I (Desai et al. 2024), which saturates at for SNR . For the -band era (2018–2024), we use the per-year skew-normal efficiency models from Paper III (Desai et al. 2026a), which peak at averaged across all years and account for year-to-year changes in survey operations and transient-selection procedures, particularly after the onset of the COVID-19 pandemic. By modeling as a continuous function of SNR, this approach captures pipeline systematics and vetting efficiency that a simple flux threshold would miss.
Each TDE is assigned its own unique template rather than a class-wide template. This is motivated by the substantial morphological diversity of TDE light curves (§ 2.5), and it means that is evaluated individually per event. TDE light curves are also considerably longer-lived than SNe or kilonovae, with rise-and-fall timescales of weeks to months. This extended duration generally increases the number of epochs above threshold for a given peak flux, improving completeness relative to shorter-lived transients at comparable peak magnitudes.
For each injection, we record its peak apparent magnitude and whether it was recovered. is then the fraction of injections recovered in bins of peak apparent magnitude, i.e., the probability that ASAS-SN would have recovered the TDE at its observed brightness, and we evaluate it at the measured of each TDE. Figure 2 shows the resulting completeness curves for all TDEs in our sample, separated by survey era. The spread among curves at a given magnitude reflects the diversity of TDE light curve morphologies, with broader and more slowly evolving events achieving higher completeness at fainter apparent magnitudes.
Figure 2. Photometric completeness () as a function of peak apparent magnitude for each TDE in our sample, shown separately for the -band era (top) and the -band era (bottom). Each curve corresponds to an individual TDE and uses the per-TDE light curve template described in § 2.5.
corrects for recovery at a given apparent brightness, but not for the volume within which a TDE of that luminosity could have been selected. A TDE that falls below our magnitude limit beyond samples only part of the survey volume, so, following the approach of Papers I–III, we compute for each TDE
where is the redshift at which the TDE, with , would have , with mag and . This normalizes the effective detection volume of every TDE to the common maximum redshift of our sample.
Spectroscopic Completeness
Not all transients discovered or recovered by ASAS-SN receive spectroscopic classification. To account for this incompleteness, we compute a spectroscopic completeness correction using the population of nuclear transients in the ASAS-SN archive as a proxy for the spectroscopic completeness of TDEs specifically.
We construct a sample of nuclear transients from two sources. First, from the full sample of 3355 ASAS-SN-discovered or recovered -band transients in Paper III (Desai et al. 2026a)—which includes both classified and unclassified non-stellar objects with a host galaxy association, but excludes spectroscopically classified TDEs—we select the 771 events that lie within of a host galaxy nucleus, as determined by cross-matching with the REGALADE galaxy catalog (Tranin et al. 2026). Second, we add the 25 ASAS-SN-discovered or recovered TDEs from our sample, none of which are in the Paper III sample, yielding a combined sample of 796 nuclear transients. Of these, 662 (83%) are spectroscopically classified.
We define for a given year as the ratio of spectroscopically classified to total nuclear transients in that year. For the -band era (2014–2017), the ASAS-SN sample was nearly spectroscopically complete (Holoien et al. 2017a,b,c, 2019a), with , , , and for 2014 through 2017, respectively. For the -band era, , , , , , , and for 2018 through 2024, respectively. The decline in later years mirrors the pattern seen in Paper III, driven by the fainter -band limiting magnitude, disruption of follow-up resources during the COVID-19 pandemic, and the reduced availability of spectroscopic follow-up for southern targets. Since the statistical uncertainties from the small TDE sample size dominate the total error budget, precise characterization of is not critical, and this approach provides a sufficiently accurate estimate.
Volumetric Rates
The final statistical weight for the TDE is . depends on the discovery filter because the single-epoch detection efficiency differs between the -band and -band eras. and are filter-independent. The combined volumetric rate is then
where is the total comoving volume defined by our redshift limit (), yr is the full survey baseline from UTC 2014-01-01 to UTC 2024-12-31, and is the sky fraction correction for our Galactic latitude cut of .
We estimate statistical uncertainties on all rates via bootstrapping, following Papers I–III. We first draw the sample size as a Poisson deviate with expectation equal to the observed number of TDEs in the standard sample. We then draw that many TDEs from the observed sample with replacement and compute the rate using Equation (4). The procedure is repeated times, and the uncertainties are the and percentiles of the resulting rate distribution. For subsets or bins with TDEs, where the bootstrap is unreliable, we instead scale the Poisson confidence limits of Gehrels (1986) to the measured rate (– and – times the rate for and ). Rates based on so few events are correspondingly uncertain and sensitive to individual objects. Uncertainties on the relative fractions use a joint bootstrap, in which each object’s resampling weight enters both the subset and total rates, preserving their correlation. For , the subset members instead receive Gamma-distributed weights that reproduce the Gehrels (1986) limits, avoiding the spurious zero lower bound that arises when a lone object is dropped from a resample. Because these uncertainties are large and asymmetric, we quote rates as throughout. The resulting rates and fractions by spectral type are given in Table 3.
| Sample | Fraction [%] | |||
| TDE-H+He | 0.033 | 15 | ||
| TDE-H | 0.045 | 4 | ||
| TDE-He | 0.080 | 2 | ||
| Total | 0.044 | 21 | 100 |
Table 3. Volumetric TDE rates and relative fractions by spectral type. All fractions are relative to the total rate. Because each subtype rate is a strict subset of the total, and are correlated; we account for this with a joint bootstrap rather than linear error propagation.
For our full standard sample of 21 TDEs, we find a total volumetric rate of
i.e., . This total rate is effectively limited to ( mag), set by our faintest event, the TDE 2023clx (ASASSN-23bd; Hoogendam et al. 2024), which alone contributes of . From their sample of optical
TDEs in ZTF, Yao et al. (2023) find for events with . Applying the same cut, which removes only TDE 2023clx from our standard sample, we find (), consistent with Yao et al. (2023) at , although the samples and selection functions differ.
To estimate a per-galaxy rate, we restrict to the 19 TDE hosts with . This excludes TDE 2023clx, with , and the off-nuclear TDE 2024tvd, for which we cannot make a robust stellar-mass estimate (e.g., Patra et al. (2025)), giving . Dividing this by the number density of galaxies in the same mass range from the stellar mass function of Baldry et al. (2012) gives an average per-galaxy rate of
i.e., . This is broadly consistent with early estimates of the per-galaxy TDE rate from SDSS (van Velzen & Farrar 2014) and ASAS-SN (Holoien et al. 2016a). Our per-galaxy rate estimate is also within of the that Yao et al. (2023) find for galaxies with , although ours is an average over the given mass range rather than a value at fixed galaxy mass.
Off-Nuclear and Recurrent TDE Rates
Among the standard sample of TDEs is TDE 2024tvd (ZTF22aaigqsr; Yao et al. 2025a), one of the few optically-selected off-nuclear TDEs (Yao et al. 2025a; Stein et al. 2026; Li et al. [134]). From this event, we compute the rate of off-nuclear TDEs to be
This corresponds to of the total volumetric TDE rate, with uncertainties from a joint bootstrap. This is consistent with the conservative fractional range of 0.2–9.6% estimated by Stein et al. (2026) from the ZTF Bright Transient Survey (Fremling et al. 2020; Perley et al. 2020).
Our standard sample also includes four recurrent TDEs (rTDEs), each with two flares as of 2024 December 31: TDE 2019azh (ASASSN-19dj; Hinkle et al. 2021a; Yao et al. 2026b), TDE 2018fyk (ASASSN-18ul; Wevers et al. 2023; Pasham et al. 2024), TDE 2022dbl (ASASSN-22ci; Lin et al. 2024; Hinkle et al. 2024; Makrygianni et al. 2025; Hinkle et al. 2026), and TDE 2024pvu (ZTF18abxnvoz; Yao et al. 2026b). These events are plausibly powered by repeating partial TDEs (e.g., Yao et al. 2026b; Hinkle et al. 2026). Counting each rTDE once, using a single-flare template based on its first flare in ASAS-SN, we measure the volumetric rate of rTDEs to be
which corresponds to of all TDEs. If we include ASASSN-14ko (e.g., Payne et al. 2021, 2022, 2023), then the total TDE rate becomes (), and the rTDE rate is , increasing the rTDE fraction to . Because recurrence can only be established when multiple flares are observed, this represents a lower limit on the intrinsic rTDE rate. Our central estimates (–) are systematically lower than the – fractional rate derived by Yao et al. (2026b) from a combination of Catalina Real-Time Transient Survey (CRTS; Drake et al. 2009) and ZTF observations, but demonstrate that rTDEs are a non-negligible fraction of all TDEs.
Of the four rTDEs in our standard sample, only TDE 2018fyk (ASASSN-18ul) and TDE 2022dbl (ASASSN-22ci) had a second optical flare during the 2014–2024 ASAS-SN baseline; the additional flares of TDE 2019azh (ASASSN-19dj) and TDE 2024pvu (ZTF18abxnvoz) occurred before ASAS-SN began survey operations. The second flare of TDE 2018fyk peaked fainter than mag and therefore does not pass the cuts of our standard sample. This leaves a single additional flare, the 2024 re-brightening of TDE 2022dbl (Lin et al. 2024; Hinkle et al. 2024), which peaked at mag. Treating this flare exactly as any other TDE in our sample, with , , and evaluated at its own peak magnitude and year, and counting each flare as an independent event, we find
which corresponds to of all TDE flares in our sample, counting the TDE 2022dbl re-brightening as an additional event. This is a factor of higher than the rate obtained when each rTDE is counted once (Equation 8), because the fainter second flare of TDE 2022dbl ( mag) is only detectable out to and therefore carries a large weight, accounting for about three-quarters of .
Optical Luminosity Functions
We construct the -band LFs in terms of the peak rest-frame -band luminosity, which we define as at , and compute from the peak absolute magnitude as , where the constant follows from the AB zero point (3631 Jy) at 10 pc.
The sections that follow leverage the design of ASAS-SN and our well-calibrated knowledge of its systematics to examine various comparisons of the rate of optical TDEs as a function of key TDE parameters of interest. First, we determine the peak -band luminosity function (this Section) both for the full sample and as a function of optical spectroscopic type (§ 4.1) and X-ray luminosity (§ 4.2). Next, we compare the TDE rate to properties at peak light inferred through blackbody modeling of the TDE UV/optical SED (§ 5). Finally, we investigate how the TDE rate varies with host-galaxy properties (§ 6), such as the QBS or PSB nature of the host (§ 6.1), a recent burst component (§ 6.2), the bulge Sérsic index (§ 6.3), and the SMBH mass (§ 6.4).
Figure 3 presents the -band LF of TDEs in our sample; the binned rates for this and the later figures are given in Appendix § 12. Over the luminosity range probed by our sample, the -band LF declines toward higher peak optical luminosities, consistent with previous results ((van Velzen 2018; Lin et al. 2022; Charalampopoulos et al. 2023; Yao et al. 2023)). The shape is similar to the LF produced by ZTF ((Yao et al. 2023)), although our rates are lower by a factor of at erg s, consistent with the ratio of the integrated rates for erg s, (§ 3.3). Yao et al. (2023) measured a break in the rates at with a rate normalization of 287 yr Gpc dex. Employing the same model,
Figure 3. ASAS-SN -band TDE LF. The TDE rate increases towards lower luminosity, becoming shallower or possibly plateauing below erg s. The black dashed line shows a double power-law fit to our TDE LF. The double-power-law fit to the ZTF -band TDE LF of Yao et al. (2023) is shown as the dotted gray line for comparison. Left Panel: The -band LF, colored by the optical spectroscopic type of the TDE. Shown are TDE-H (purple), TDE-He (cyan), and TDE-H+He (orange). Right Panel: The -band LF, colored by the X-ray-to-blackbody luminosity ratio of the TDE. TDEs with are shown in blue and events with are shown in red. In both panels, the total rate is given by the black line. Subset points are offset slightly from the bin centers for clarity.
where is the break luminosity and () well below (above) the break, we fit our LF and estimate the parameters from bootstrap resamplings of the sample, adopting the median and 16th–84th percentiles of each parameter. Because our faint end contains only a single event, we retain only the bootstrap fits with , excluding unphysically steep solutions. We find a break luminosity of , a normalization of yr Gpc dex, a bright-end slope of , and a faint-end slope of , all consistent with Yao et al. (2023) within the respective uncertainties. A larger sample is needed to test whether the faint end of the -band LF turns over or continues to rise.
Exotic accretion-powered events with high peak luminosities such as luminous ANTs ((Wiseman et al. 2025)) and extreme nuclear transients (ENTs; (Hinkle et al. 2025)) are broadly consistent with extrapolations of the high-end of the TDE LF, although this does not necessarily imply that ANTs and ENTs are drawn from the same physical population as typical UV/optical TDEs. The rate of luminous ANTs was estimated from a ZTF-selected sample to be yr Gpc ((Wiseman et al. 2025)). This rate occurs at , near the median ANT peak luminosity (e.g., [155] (Wiseman et al. 2025)). The ENT rate was loosely constrained to be yr Gpc ((Hinkle et al. 2025)) from a sample of only 3 events. An extrapolation of our -band LF yields this rate at , in reasonable agreement with the confirmed ENTs. Future rate comparisons will need to account for the poorly constrained completeness for these events, with the recent discovery of several ENT candidates at lower redshifts ((Quintin et al. 2026)) suggesting that the true rate may be higher. The indications of pre-flare AGN activity, if it is related to the production of ANT and ENT-like flares, could decrease the rates by roughly a factor of the AGN fraction ( [99, 129]).
Optical Spectroscopic Types
Next, we examine the rates of the different optical spectroscopic types of TDEs: TDE-H, TDE-He, TDE-H+He, and TDE-Featureless (e.g., Leloudas et al. 2019; van Velzen et al. 2021; Hammerstein et al. 2023). Here we adopt the near-peak spectroscopic type, although some TDEs show evolution in their spectra (e.g., Nicholl et al. 2019; Short et al. 2020; Hinkle et al. 2021a). Observational studies have found differences in the physical properties of different TDE spectroscopic types, including the larger blackbody radii of TDE-H events (Hinkle et al. 2020; van Velzen et al. 2021; Nicholl et al. 2022), the high temperatures of TDE-He and TDE-Featureless events (van Velzen et al. 2021; Hammerstein et al. 2023), and a preference for lower (higher) mass SMBHs for TDE-H (TDE-Featureless) events (Nicholl et al. 2022; Hammerstein et al. 2023; Yao et al. 2026a).
Theoretical studies indicate that high-luminosity and compact photospheres are highly ionized and can suppress prominent optical line features, producing featureless spectra (e.g., Aspegren & Kasen 2026; Thomsen et al. 2026). Consistent with the observations, these studies also predict that the strongest emission lines vary with luminosity and photospheric radius, with H lines present for sufficiently extended photospheres and progressively stronger He II emission in more compact photospheres (Roth et al. 2016; Aspegren & Kasen 2026; Thomsen et al. 2026). For Eddington-limited TDE emission, the observed SMBH mass trend is also consistent with He-dominated or featureless spectra corresponding to higher luminosity events.
As shown in Table 3, TDE-H+He events heavily dominate the rate of nearby TDEs, comprising of the volumetric rate. The TDE-H and TDE-He spectroscopic types occur at and of the total optical TDE rate, respectively. The extremely high fraction of events showing both strong H and He emission is consistent with a comparison of the typical TDE peak blackbody luminosities and radii to parameters expected to produce such spectra (Aspegren & Kasen 2026, cf. their Figure 3) and broadly agrees with the predictions of Roth et al. (2016) (cf. their Figures 7 and 8) and Thomsen et al. (2026) (cf. their Figure 2d).
The TDE-Featureless events comprise two sub-populations (e.g., Yao et al. 2026a), one at high luminosities where over-ionization is likely (Aspegren & Kasen 2026) and another with lower luminosities but more compact photospheres (Yao et al. 2026a). ASASSN- SN has not discovered or recovered an unambiguous TDE-Featureless event. We place a 95% confidence upper limit on the rate of luminous ( to mag) featureless TDEs of , or of the total rate, assuming Poisson statistics for zero detected events. To compute the selection efficiency, we inject events uniformly in comoving volume out to with peak magnitudes between and mag, and recover them using the completeness curves built from the light curve of the luminous TDE-Featureless event TDE 2020qhs (Hammerstein et al. 2023) for each survey year.
The enigmatic transient ASASSN-15lh, which shows no clear emission features, has been interpreted both as a luminous TDE candidate (e.g., Leloudas et al. 2016) and as a superluminous supernova (e.g., Dong et al. 2016). Here, assuming that it is a featureless TDE, we conservatively use ASASSN-15lh as a constraint on the potential rate of luminous TDE-Featureless events. At , it lies beyond the volume of our standard sample, so we compute its rate using the method, summing over the two survey eras, with set by the redshift to which ASAS-SN could have detected it in each era ( in the -band era and in the -band era). We find , or of the total volumetric TDE rate, well below the upper limit placed on luminous featureless TDEs.
At the other end of the luminosity function, ASASSN-20hx (AT 2020ohl; Hinkle et al. 2022) has been claimed to be a low-luminosity featureless TDE (Yao et al. 2026a). It has been alternatively classified as an ANT (Hinkle et al. 2022; Roy et al. 2024; Hinkle 2024). The continued optical excess above host-galaxy levels and ongoing years-long re-brightening11 are highly atypical of TDEs, making it uncertain whether it is a bona fide TDE-Featureless. Nevertheless, assuming ASASSN-20hx is a low-luminosity TDE-Featureless event, we derive an approximate constraint on the rate of low-luminosity TDE-Featureless events within ASAS-SN of , or of the total volumetric TDE rate.
In the left panel of Figure 3, we see that TDE-H+He events are dominant for all -band luminosities below . By , the rates of TDE-H, TDE-He, and TDE events are similar, and only the TDE-He events in our sample have higher -band luminosities. The low rates and high luminosities of TDE-He events, and the absence of luminous TDE-Featureless events from our sample, are generally consistent with the preference of such TDEs to occur on more massive SMBHs.
X-ray Luminosity
While the TDEs in our sample are, by definition, optically-selected, they all have Swift XRT (Burrows et al. 2005) coverage. Therefore, we can also examine the -band LF with respect to the X-ray luminosity of the TDEs. This comparison is shown in the right panel of Figure 3, with two bins separating the TDEs at an X-ray-to-blackbody luminosity ratio of , approximately the median of our sample. The normalization and shape of our -band LF is broadly similar to the X-ray LFs (e.g., Auchettl et al. 2018; Sazonov et al. 2021; Guolo et al. 2024; Grotova et al. 2025). Given the heterogeneous coverage, should be viewed as a lower limit on the intrinsic maximum X-ray luminosity for many events. We therefore use as an approximate X-ray-brightness indicator, although the upper limits for all TDEs without detected X-ray emission lie firmly in the bin. We additionally caveat that the observed X-ray luminosity can vary strongly as a function of time, with some TDEs showing a significant X-ray brightening several months after peak optical light (e.g., Auchettl et al. 2017; Gezari et al. 2017; Holoien et al. 2018; Hinkle et al. 2021a). The rates of X-ray-faint and X-ray-bright optical TDEs are nearly identical. A qualitatively similar result is recovered when the sample is split at erg s (cf. Hammerstein et al. 2023, approximately the median X-ray luminosity of our sample). This suggests that the physics setting the multi-wavelength SED appearance, as traced here by the volumetric rate of X-ray-bright and X-ray-faint optical TDEs, is likely a viewing angle effect (e.g., Auchettl et al. 2017; Dai et al. 2018; Thomsen et al. 2022). In this picture, TDEs are powered by promptly formed accretion disks, and the broadband appearance depends on the optical depth of a reprocessing outflow along the line of sight to an observer. Observers whose lines of sight pass through an optically thick reprocessing layer preferentially observe UV/optical-dominated emission, whereas observers viewing the inner accretion flow without significant obscuration can detect stronger X-ray emission. This event-wise constraint is reminiscent of the global similarity of the volumetric rates of X-ray-selected and optically-selected TDEs (e.g., Auchettl et al. 2018; Sazonov et al. 2021; Yao et al. 2023; Guolo et al. 2024; Grotova et al. 2025).
Rates vs. Blackbody Properties
Despite the optical luminosity being a key direct observable, optically-selected TDEs emit the majority of their radiation in the UV, with SEDs well-described by single-temperature blackbodies (e.g., Gezari et al. 2012; Holoien et al. 2016a; Hinkle et al. 2020; van Velzen et al. 2021). We therefore consider the dependence of optical TDE rates on their blackbody properties at peak light (see § 2.2 for more details) in Figure 4. Similar to previous results (Lin et al. 2022; Yao et al. 2023), we find that the decline in TDE rate towards higher blackbody luminosities is shallower than the -band LF discussed in § 4. Unlike the blackbody LFs produced from ZTF data (Lin et al. 2022; Yao et al. 2023), a single power-law is not a good fit to the data. The integrated rates are consistent (§ 3.3), but our blackbody LF is flatter for erg s. While there are likely small differences in the methodology of estimating blackbody parameters, we suspect that this primarily reflects the limited number of low-luminosity
Figure 4. Optical TDE rates compared to the peak blackbody luminosity (left panel), blackbody radius at peak (middle panel), and blackbody temperature at peak (right panel). The rates are colored by the optical spectroscopic type of the TDE. Shown are TDE-H (purple), TDE-He (cyan), and TDE-H+He (orange). The total rate is shown in black. In the left panel, the gray dashed line shows the double power-law fit to the blackbody LF of Yao et al. (2023), and the top axis gives the SMBH mass for which the blackbody luminosity equals the Eddington luminosity (§ 5). Subset points are offset slightly from the bin centers for clarity.
events in our sample. Our two lowest-luminosity bins have only 3 sources, as compared to TDEs in the two lowest-luminosity bins for Yao et al. (2023). Many of these low-luminosity TDEs peak at apparent magnitudes similar to or fainter than the ASAS-SN limiting magnitude and often have faster light curve evolution than more luminous TDEs (e.g., Charalampopoulos et al. 2023; Hinkle et al. 2023; Hoogendam et al. 2024), making them difficult to detect. As compared to ASAS-SN, the larger volume probed by ZTF makes it more sensitive to these intrinsically faint, yet intrinsically common, events at modest redshift. Regardless, the blackbody LF at erg s remains consistent with Eddington-limited luminosities (see § 6.4). The top axis of the left panel of Figure 4 gives the SMBH mass for which equals the Eddington luminosity, erg s. We define the UV/optical Eddington ratio of each TDE as , using the peak blackbody luminosity and the SMBH mass from § 2.4.
Although the conversion between mass fallback and luminosity is complicated, the maximum accretion rate of a TDE is set primarily by the mass of the disrupted star (e.g., Kochanek 2016; Mockler et al. 2019; Bandopadhyay et al. 2024). The low rates of the highest-luminosity TDEs ( yr Gpc ) are suggestive of the disruption of massive stars, which should be intrinsically rare due to the stellar initial mass function and short stellar lifetimes.
We recover the known trend for TDE-H events to occur at lower blackbody temperatures (e.g., Hinkle et al. 2020; van Velzen et al. 2021; Hammerstein et al. 2023). There is no significant relationship between the TDE rate and blackbody temperature, with an approximately equal volumetric rate of TDEs for any observed UV/optical blackbody temperature.
Like previous studies (van Velzen 2018; Yao et al. 2023), the optical TDE rate declines monotonically with increasing blackbody radius. The low rate at blackbody radii of cm broadly agrees with the break radius found by Yao et al. (2023). Their interpretation that the presence of a maximum radius at this scale supports a limit imposed by the semi-major axis distance of the most bound debris is consistent with our observations and the rate suppression seen at the high-mass end of our TDE BH mass function (see § 6.4).
Rates vs. Host-Galaxy Properties
Transient host galaxies likely hold information relevant to understanding the physical origins of the transients themselves. Although the basic physical picture of a TDE is well understood, large open questions remain about the rate at which stars are placed on the low angular momentum orbits that ultimately lead to their destruction in a TDE. Here we discuss the rate of optical TDEs as a function of various properties of their host galaxies, including the per-galaxy rate enhancement in QBS and PSB galaxies, the presence of a recent burst of star formation, the host-galaxy light profile, and the mass of the host-galaxy SMBH.
We exclude TDE 2024tvd from the host-galaxy measurements that follow, since properties measured at the host nucleus do not describe the environment of its disruption. This leaves 20 of the 21 TDEs in our standard sample. TDE 2024tvd is retained in § 6.4, as its SMBH mass is estimated from the TDE itself rather than from host-galaxy scalings.
Rate Enhancements in QBS and PSB Galaxies
It has long been recognized that many TDEs occur in galaxies with little current star formation, but evidence for substantial recent (within Gyr) star formation (e.g., Arcavi et al. 2014; French et al. 2016, 2018; Law-Smith et al. 2017; Graur et al. 2018). It is now well-established that the per-galaxy TDE rate is enhanced in QBS and PSB galaxies, at levels of – the average rate (e.g., French et al. 2020b), with galaxies at more extreme Lick H index typically having higher rate enhancements.
Figure 5 compares H emission EW, tracing current star formation, to the Lick H absorption index, tracing star formation within the past Gyr, colored by the volumetric TDE rate. The KDE is a two-dimensional Gaussian kernel density estimate of the host positions in this plane, with each TDE entering with its completeness weight . We use the gaussian_kde implementation in SciPy (Virtanen et al. 2020) with Scott’s rule for the bandwidth (Scott 2015): the kernel covariance is the weighted sample covariance of the host positions scaled by , where is the effective number of hosts given the unequal weights. This corresponds to kernel widths of Å in Lick H and Å in H EW. The resulting surface is renormalized to integrate to the total volumetric rate of the sample, so that both panels show a rate per unit area of the plane and share a single color scale. Both for the discrete bins and the smoothed KDE, the volumetric TDE rate largely follows the underlying galaxy distribution. This independently confirms that QBS and PSB galaxies do not dominate the volumetric TDE rate, despite their significant per-galaxy rate enhancements. Volumetrically, the typical TDE host is a weakly star-forming galaxy, consistent with known TDE hosts (e.g., French et al. 2016; Graur et al. 2018; French et al. 2020b).
The over-representation of TDEs in the lower-right-hand corner, where QBS and PSB galaxies reside, is clear in Figure 5, so we quantify the rate enhancements. Following French et al. (2016), we define QBS galaxies by H EW Å and Å Lick H Å, and PSB galaxies as those with H EW Å and Lick H Å. Of the 20 hosts, three are classified as QBS and four as PSB.

Figure 5. Hα emission line EW, which traces current star formation, as compared to the Lick HδA absorption index, which traces star formation in the past Gyr. The TDE host galaxies in our sample are shown in white, and the gray points are the SDSS control sample. Volumetric TDE rates are shown in coarse bins (left panel) and as a Gaussian-smoothed KDE (right panel).
Because our TDE sample is completeness-corrected while the control sample is not, the two cannot be compared directly by number. Each TDE carries the statistical weight defined in § 3. We therefore assign a corresponding weight to each control galaxy. For control galaxy in cell of redshift and stellar mass parameter space,
with the number of control galaxies in cell . This distributes the total TDE weight in each cell uniformly among that cell’s control galaxies, and guarantees by construction. For cells containing a single TDE host, Equation (11) reduces to .
The TDE rate and control galaxy fractions in a region of the Hα–HδA plane are then
and
with their ratio being the rate enhancement
indicates that the estimate of the fraction of volumetric TDE rate in region is exactly proportional to the weighted fraction of galaxies in that region, at matched stellar mass and redshift; indicates overrepresentation. This rate enhancement is shown in Figure 6, and Table C3 gives its value in every region. For the KDE panel, the TDE and control distributions are

Figure 6. The same as Figure 5, except that the colors refer to the TDE rate enhancement (see Equation (14)). The rate enhancements are shown in coarse bins (left panel) and as a Gaussian-smoothed KDE (right panel).
Table C3. TDE rates and rate enhancement factors (Equation (14)) in every H EW–Lick H region used in § 6.1, relative to the stellar-mass- and redshift-matched SDSS control sample. Uncertainties are the 16th–84th percentiles of the joint Poisson bootstrap.
smoothed with the same fixed Gaussian kernel used for Figure 5, before taking their ratio. We estimate uncertainties with the same joint Poisson bootstrap used in § 3, recomputing Equations (11)–(14) in full for each realization so the numerator and denominator fluctuate coherently. We report the 16th and 84th percentiles as the uncertainties.
In Table 4, we quote the full-sample values as our primary result and present the TDE 2023clx-excluded values as a robustness check, since this event carries over a quarter of the total sample weight and is not hosted by a QBS or PSB galaxy. The qualitative conclusion is unchanged in either case: PSB galaxies, and the combined PSB+QBS population, are overrepresented in the volumetric TDE rate at fixed stellar mass and redshift, with the lower bounds of their uncertainties well above . The QBS-only enhancement, based on three hosts, is of lower significance. These results reinforce the known overrepresentation of QBS and PSB galaxies as TDE hosts (e.g., (e.g., Arcavi et al. 2014; French et al. 2016, 2018; Law-Smith et al. 2017; Graur et al. 2018). Using the calibrated TDE weights, we estimate more modest values of than many previous works, although they are generally consistent within uncertainties. The slightly lower values are likely due to our use of a matched control galaxy sample and weighted TDE frequency.
| Population | (Full sample) | (Excl. TDE 2023clx) | |
| PSB | 4 | ||
| QBS | 3 | ||
| PSB+QBS | 7 |
Table 4. Volumetric TDE rate enhancement factors in the Hα EW–Lick HδA plane, relative to a stellar-mass- and redshift-matched SDSS control sample. corresponds to no enhancement, i.e. a volumetric TDE rate exactly proportional to the abundance of such galaxies at matched and . Uncertainties are the 16th and 84th percentiles of the joint Poisson bootstrap realizations. The right-hand column excludes TDE 2023clx, which carries over a quarter of the total sample weight but whose host lies in neither region.
A Recent Burst of Star Formation
Recent work has investigated the empirical delay time distribution of TDEs to test various theoretical models for TDE rate enhancements in QBS and PSB galaxies (Shepherd et al. 2026). Following a similar procedure (detailed in § 2.3), we compare the volumetric TDE rates to various properties of any recent burst of star formation. These results are shown in Figure 7, with the time since burst () shown in the left panel, the burst mass fraction in the middle panel, and , the declining SFH slope of the recent burst component, in the right panel. An important caveat is that galaxies with low estimated burst mass fractions may not have experienced a burst; the recent SFH component likely traces the most recent epoch of star formation. We therefore restrict the following analyses of and to host galaxies with burst mass fractions above 0.1%. In our standard sample, only the host galaxies of TDE 2019dsg and TDE 2022hvp have burst mass fractions below 0.1%. Each has a large and low , consistent with a low level of ongoing star formation.

Figure 7. Optical TDE rates compared to the burst age (left panel), the burst mass fraction (middle panel), and , the falling slope of the burst SFH (right panel). The black squares and gray band show the total rate. The purple circles and green diamonds show the rates for TDEs in quiescent Balmer-strong or post-starburst (QBS/PSB) hosts and in all other (non-QBS/PSB) hosts, respectively. The left and right panels include only the 18 hosts with burst mass fractions above 0.1% (Section 6.2), while the middle panel includes all 20 hosts. In the right panel, the horizontal bars show the widths of the non-uniform bins. Subset points are offset slightly from the bin centers for clarity.
The TDE rate as a function of is relatively flat, with a slight increase in the volumetric rate at smaller burst ages. An increase in the TDE rate towards lower burst ages broadly aligns with theoretical models predicting enhanced TDE rates within several hundred Myr following a starburst (e.g., Stone et al. 2018; Bortolas 2022; Mockler et al. 2023; Melchor et al. 2024; Wang et al. 2022; Teboul & Perets 2025). For QBS and PSB galaxies, all with burst mass fractions 0.1%, the rate nominally increases towards larger , reminiscent of the rate enhancements presented in Shepherd et al. (2026). If we split the sample at a burst mass fraction of 1% (cf. Shepherd et al. 2026), there is no trend with . Taken together, these trends suggest that multiple factors increase the TDE rate in galaxies with recent bursts of star formation (Shepherd et al. 2026).
The volumetric rate of TDEs is lowest in galaxies that have not recently hosted a burst of star formation. The rate increases by more than two orders of magnitude between a burst mass fraction of and . However, above a burst mass fraction of , the TDE rate is consistent with a constant rate. This possibly indicates a threshold above which TDE production is more efficient, in broad agreement with the majority of TDE hosts having similarly high burst mass fractions (Shepherd et al. 2026). The total TDE rate is dominated by galaxies with very steeply falling SFHs, indicated by a high , although at most a weak trend exists for QBS/PSB galaxies. This suggests that optically-selected TDEs may prefer rapidly-declining bursts and that dynamical channels for increasing the TDE rate (e.g., Stone et al. 2018; Bortolas 2022; Mockler et al. 2023; Melchor et al. 2024; Teboul & Perets 2025) are most efficient shortly after a burst quenches.
Finally, the selection of TDE candidates in the IR, many of which lack optical counterparts (Masterson et al. 2024), demonstrates the unsurprising fact that optical selection alone is an incomplete picture of the TDE population. Heavily dust-obscured TDEs will be missed by optical surveys. We examined the rate of TDEs as a function of the fitted in our Bagpipes models and found no correlation between the rate and the amount of dust in the host galaxy (Figure A1, Appendix § 10). Although this total is not the line-of-sight dust relevant for obscuring a TDE flare, the similar rate in dusty galaxies suggests that the optical TDE rates are not substantially biased by a modest amount of host-galaxy dust ( mag). Multi-wavelength selection increases TDE completeness but also probably reduces purity since the cleanest separation between TDEs and other physically distinct accretion flares is based on optical spectroscopic classification.
Figure A1. Optical TDE rates as a function of the host-galaxy dust attenuation from our Bagpipes SED modeling (§ 2.3). The black squares show the total rate. Purple diamonds, cyan squares, and orange circles show TDE-H, TDE-He, and TDE-H+He events, respectively. The rate is consistent with being flat as a function of . Subset points are offset slightly from the bin centers for clarity.
Bulge Sérsic Index
Previous studies have shown that TDE host galaxies tend to be centrally concentrated (e.g., Law-Smith et al. 2017; Graur et al. 2018; Hammerstein et al. 2021; Newsome et al. 2025; Chang et al. [35]). In Figure 8, we examine the volumetric TDE rate as a function of the bulge Sérsic index . The TDE rate is dominated by galaxies with small and continues to decrease with increasing before possibly plateauing for . A Sérsic index of is often used to separate pseudobulges () and classical bulges (), which may form through secular processes and mergers, respectively (e.g., Fisher & Drory 2008; Gadotti 2009). This matches the known prevalence of TDEs in low-mass host galaxies where major mergers are less common (e.g., French et al. 2020b; van Velzen et al. 2021; Hammerstein et al. 2021), although some TDE host galaxies do exhibit morphological signatures consistent with recent mergers (e.g., Prieto et al. 2016). The high TDE rate in galaxies that likely host pseudobulges is also consistent with Chang et al. [35], who find that the majority of TDE hosts lack signs of morphological disturbances and are more likely than matched control galaxies to show bar- or ring-like structures, suggesting that their recent evolution has been dominated by bar-driven secular processes rather than mergers.
Figure 8. Optical TDE rates as a function of the Sérsic index of the host-galaxy bulge, which traces the central concentration of the stellar light profile. The black squares and gray band show the total rate. The purple circles and green diamonds show the rates for TDEs in quiescent Balmer-strong or post-starburst (QBS/PSB) hosts and in all other (non-QBS/PSB) hosts, respectively. Subset points are offset slightly from the bin centers for clarity.
The plateau in the TDE rate for galaxies with large is primarily associated with QBS and PSB hosts. In fact, all galaxies with are QBS or PSB after considering the host galaxy of TDE 2019ahk (ASASSN-19bt). TDE 2019ahk occurred in a galaxy with H EW = Å and Lick H Å, which places it near the population of shocked post-starburst galaxies (SPOGs; Alatalo et al. 2016; Holoien et al. 2019c). Although Holoien et al. (2019c) could not robustly identify the host galaxy of TDE 2019ahk as a member of the SPOG population, the high burst mass fraction at a Gyr-old age (see Table 2) supports its classification as a SPOG. The plateau in the total volumetric TDE rate at large is therefore plausibly associated with the overrepresentation of QBS and PSB galaxies.
Supermassive Black Hole Mass
The SMBH mass on which a TDE occurs is fundamental in setting key observational parameters, including the timescale (e.g., Wang & Merritt 2004; Stone & Metzger 2016; Kochanek 2016) and may regulate the observed bolometric luminosity through Eddington-limited accretion and reprocessing (e.g., Wevers et al. 2017, 2019). Thus, it is an important parameter for rate comparisons (van Velzen 2018; Yao et al. 2023). TDEs have long been predicted to occur at higher rates on lower-mass SMBHs (e.g., Magorrian & Tremaine 1999; Wang & Merritt 2004; Stone & Metzger 2016; Kochanek 2016) and the precipitous drop in the rates at the highest SMBH masses is well-known (e.g., Hills 1975; van Velzen 2018). TDEs on IMBHs are expected (e.g., MacLeod et al. 2016; Chang et al. [34]), although the rate predictions vary. Because the tidal disruption radius approaches the stellar radius below M, there may be qualitative changes in TDE phenomenology in the IMBH regime. Figure 9 shows the volumetric rate of optical TDEs compared to their SMBH mass. As discussed in § 2.4, we favor masses estimated through late-time TDE observations.
Figure 9. Optical TDE rates as a function of the SMBH mass. The rates are colored by the UV/optical Eddington ratio of the TDE, split at the median Eddington ratio of the sample. Subset points are offset slightly from the bin centers for clarity. We recover the scaling (orange dashed line) measured by Yao et al. (2023). We additionally show two expectations for the TDE SMBH mass function from Gallo & Sesana (2019) with (solid) and without (dotted) event horizon suppression. All three comparison curves are normalized to our total measured rate in the bin centered at .
Like previous measurements of the TDE SMBH mass function, we find that the TDE rate is relatively flat below M (van Velzen 2018; Yao et al. 2023). At low SMBH masses, our mass function is consistent with the rate dependence found by Yao et al. (2023), suggestive of a flat SMBH mass function. Similarly, we find that TDE rates drop for M, broadly consistent with the expected suppression near and above the Hills mass for low-mass stars. The observed Eddington ratios are highest around less massive SMBHs and drop for higher-mass SMBHs, consistent with the expected fallback rates over this mass range (e.g., Ulmer 1999; Strubbe & Quataert 2009; Kochanek 2016; Stone & Metzger 2016). The high Eddington-ratio TDEs dominate the total rate, confirming that local TDEs are excellent probes of super-Eddington accretion. ANTs and ENTs are promising transient probes of high-mass SMBHs (e.g., Subrayan et al. 2023; Wiseman et al. 2025; Hinkle et al. 2025). An extrapolation of the
high-mass end of the TDE SMBH mass function, following the Gallo & Sesana (2019) functional form with event horizon suppression, is again broadly consistent with existing rate estimates for luminous ANTs and ENTs. The estimated ANT rate of (Wiseman et al. 2025) agrees with our TDE SMBH mass function at . The general agreement in rates supports their interpretation as TDEs in AGNs and/or massive star TDEs (Subrayan et al. 2023; Wiseman et al. 2025; Hinkle et al. 2025; Graham et al. 2026).
Summary
We present the volumetric rate of optical TDEs as measured from the ASAS-SN survey. Our standard sample comprises 21 TDEs within a redshift of , spanning -band absolute magnitudes from to mag. The ASAS-SN completeness is calibrated and accounted for when measuring the TDE rate. We additionally present the rates of two interesting subpopulations: off-nuclear and recurrent TDEs. We also examine the dependence of the optical LFs on spectroscopic type, X-ray luminosity, UV/optical blackbody properties, and host-galaxy properties. Our conclusions are summarized as follows:
We measure a total local volumetric rate of optical TDEs of (). The ASAS-SN -band LF agrees in shape with previous estimates (e.g., van Velzen (2018); Lin et al. (2022); Charalampopoulos et al. (2023); Yao et al. (2023)), although its normalization is lower than that of Yao et al. (2023) by a factor of at .
Since the off-nuclear TDE 2024tvd (ZTF22aaigqsr) is in our standard sample, we estimate the local volumetric rate of off-nuclear TDEs to be . This rate is of the total rate, consistent with existing coarse estimates (Stein et al. 2026).
There are 4 recurrent TDEs in our standard sample: TDE 2019azh (ASASSN-19dj), TDE 2018fyk (ASASSN-18ul), TDE 2022dbl (ASASSN-22ci), and TDE 2024pvu (ZTF18abxnvoz), which are plausibly also repeating partial TDEs (e.g., Yao et al. (2026b); Hinkle et al. (2026)). The rTDE rate is ( of the total rate). Including the well-studied rTDE ASASSN-14ko, the rate is , with a fractional rate of . Our central value of – is lower than the previous estimates (–; Yao et al. (2026b)). If each flare of the rTDEs is counted as an independent TDE in the same host galaxy, the inferred rTDE rate rises to ( of all TDE flares).
The volumetric rate of optically-selected TDEs is dominated by galaxies with significant () burst-mass fractions and steeply falling SFHs.
The volumetric rate of TDEs predominantly follows the number density of galaxies in a redshift- and mass-matched control sample. However, known rate enhancements are recovered for PSB galaxies () and, at lower significance, for QBS galaxies ().
TDE-H+He events dominate the total TDE rate, with TDE-H and TDE-He events comprising 3.4% of the overall optical TDE rate. Luminous ( mag) TDE-Featureless events are <0.5% of local TDEs. Low-luminosity TDE-Featureless events are not well-constrained, but may represent of the total TDE rate.
The -band LFs are statistically indistinguishable for X-ray bright and X-ray faint events, supporting unification models where the viewing angle sets the multi-wavelength appearance of the TDE (e.g., Dai et al. (2018); Thomsen et al. (2022)).
We recover a known dependence of the volumetric rates, suggestive of a flat SMBH mass distribution (; Yao et al. (2023)).
Extrapolations of the optical volumetric TDE rate to higher luminosities and SMBH masses are broadly consistent with the estimated rates of nuclear flares like luminous ANTs (e.g., Wiseman et al. (2025)) and ENTs (e.g., Hinkle et al. (2025)). This tentatively supports claims that they are powered by TDEs in AGNs and/or TDEs of massive stars (e.g., Hinkle (2024); Wiseman et al. (2025); Hinkle et al. (2025); Graham et al. (2026); Quintin et al. (2026)).
A robustly estimated volumetric TDE rate provides important information on the physics of TDEs themselves (see § 4 and § 5) and insights into the relationship between host-galaxy properties and TDE production (see § 6). Existing surveys provide well-constrained local volumetric TDE rates, a key input for estimating the expected yield (e.g., Bricman & Gomboc (2020); Shvartsvald et al. (2024); Karmen et al. (2026)) from the new generation of surveys such as the Legacy Survey of Space and Time on the Vera C. Rubin Observatory (LSST; Ivezić et al. (2019)), the Roman Space Telescope’s High-Latitude Time-Domain Survey (Rose et al. (2021)), and the Ultraviolet Transient Astronomy Satellite (ULTRASAT; Shvartsvald et al. (2024)). These surveys will constrain the redshift evolution of TDE rates and eventually probe the SMBH mass function at higher redshifts (e.g., Karmen et al. (2026)). Sufficiently large TDE samples may also provide a path towards constraining the spin distribution of quiescent SMBHs (e.g., Kesden (2012); Huang & Lu (2024)), although there are degeneracies with massive star TDEs. In any case, local TDE samples will continue to complement these large higher-redshift samples by providing the most detailed view of the physics of TDEs and their host galaxies.
Acknowledgments
J.T.H. acknowledges support from NASA through the NASA Hubble Fellowship grant HST-HF2-51577.001-A, awarded by STScI. STScI is operated by the Association of Universities for Research in Astronomy, Incorporated, under NASA contract NAS5-26555. The Shappee group at the University of Hawai‘i is supported with funds from NSF grant AST-2407205. ASAS-SN is funded by Gordon and Betty Moore Foundation grants GBMF5490 and GBMF10501 and the Alfred P. Sloan Foundation grant G-202114192. C.S.K. and K.Z.S. are supported by NSF grants AST-2307385 and AST-2407206. M.S. and K.D.F. acknowledge support from NSF grant AST 22-06164. M.S. thanks the LSST-DA Data Science Fellowship Program, which is funded by LSST-DA, the Brinson Foundation, the WoodNext Foundation, and the Research Corporation for Science Advancement Foundation; her participation in the program has benefited this work. J.F.B. is supported by NSF grant PHY-2609870 and was supported in part by the World Research Hub (WRH) Program of Institute of Science Tokyo. S.D. is supported by the National Natural Science Foundation of China (Grant No. 12133005). W.B.H. acknowledges support from the NSF Graduate Research Fellowship Program under Grant No. 2236415. Parts of this research were supported by the Australian Research Council Discovery Early Career Researcher Award (DECRA) through project number DE230101069. D.O.J. acknowledges support from NSF grants AST-2407632, AST-2429450, AST-2510993, and AST-2607081, NASA grants 80NSSC24M0023 and 80NSSC24K0353, and JWST-GO-05324.031, awarded by the Space Telescope Science Institute (STScI), which is operated by the Association of Universities for Research in Astronomy, Inc., for NASA, under contract NAS5-26555. This work is also funded in part by the Gordon and Betty Moore Foundation through Grant GBMF13900 to D.O.J.
Funding for SDSS-III has been provided by the Alfred P. Sloan Foundation, the Participating Institutions, the National Science Foundation, and the U.S. Department of Energy Office of Science. The SDSS-III web site is http://www.sdss3.org/.
SDSS-III is managed by the Astrophysical Research Consortium for the Participating Institutions of the SDSS-III Collaboration including the University of Arizona, the Brazilian Participation Group, Brookhaven National Laboratory, Carnegie Mellon University, University of Florida, the French Participation Group, the German Participation Group, Harvard University, the Instituto de Astrofisica de Canarias, the Michigan State/Notre Dame/JINA Participation Group, Johns Hopkins University, Lawrence Berkeley National Laboratory, Max Planck Institute for Astrophysics, Max Planck Institute for Extraterrestrial Physics, New Mexico State University, New York University, Ohio State University, Pennsylvania State University, University of Portsmouth, Princeton University, the Spanish Participation Group, University of Tokyo, University of Utah, Vanderbilt University, University of Virginia, University of Washington, and Yale University.
Some of the data presented herein were obtained at Keck Observatory, which is a private 501(c)3 non-profit organization operated as a scientific partnership among the California Institute of Technology, the University of California, and the National Aeronautics and Space Administration. The Observatory was made possible by the generous financial support of the W. M. Keck Founda-
tion. This research has made use of the Keck Observatory Archive (KOA), which is operated by the W. M. Keck Observatory and the NASA Exoplanet Science Institute (NExScI), under contract with the National Aeronautics and Space Administration.
This publication makes use of data products from the Two Micron All Sky Survey, which is a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center/California Institute of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation.
We acknowledge the use of public data from the Swift data archive.
The Legacy Surveys consist of three individual and complementary projects: the Dark Energy Camera Legacy Survey (DECaLS; Proposal ID #2014B-0404; PIs: David Schlegel and Arjun Dey), the Beijing-Arizona Sky Survey (BASS; NOAO Prop. ID #2015A-0801; PIs: Zhou Xu and Xiaohui Fan), and the Mayall z-band Legacy Survey (MzLS; Prop. ID #2016A-0453; PI: Arjun Dey). DECaLS, BASS and MzLS together include data obtained, respectively, at the Blanco telescope, Cerro Tololo Inter-American Observatory, NSF’s NOIRLab; the Bok telescope, Steward Observatory, University of Arizona; and the Mayall telescope, Kitt Peak National Observatory, NOIRLab. Pipeline processing and analyses of the data were supported by NOIRLab and the Lawrence Berkeley National Laboratory (LBNL). The Legacy Surveys project is honored to be permitted to conduct astronomical research on I’oligam Du’ag (Kitt Peak), a mountain with particular significance to the Tohono O’odham Nation.
NOIRLab is operated by the Association of Universities in Astronomy (AURA) under a cooperative agreement with the National Science Foundation. LBNL is managed by the Regents of the University of California under contract to the U.S. Department of Energy.
This project used data obtained with the Dark Energy Camera (DECam), which was constructed by the Dark Energy Survey (DES) collaboration. Funding for the DES Projects has been provided by the U.S. Department of Energy, the U.S. National Science Foundation, the Ministry of Science and Education of Spain, the Science and Technology Facilities Council of the United Kingdom, the Higher Education Funding Council for England, the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign, the Kavli Institute of Cosmological Physics at the University of Chicago, Center for Cosmology and Astro-Particle Physics at the Ohio State University, the Mitchell Institute for Fundamental Physics and Astronomy at Texas A&M University, Financiadora de Estudos e Projetos, Fundacao Carlos Chagas Filho de Amparo, Financiadora de Estudos e Projetos, Fundacao Carlos Chagas Filho de Amparo a Pesquisa do Estado do Rio de Janeiro, Conselho Nacional de Desenvolvimento Cientifico e Tecnologico and the Ministerio da Ciencia, Tecnologia e Inovacao, the Deutsche Forschungsgemeinschaft and the Collaborating Institutions in the Dark Energy Survey. The Collaborating Institutions are Argonne National Laboratory, the University of California at Santa Cruz, the University of Cambridge, Centro de Investigaciones Energeticas, Medioambientales y Tecnologicas-Madrid, the University of Chicago, University College London, the DES-Brazil Consortium, the University of Edinburgh, the Eidgenössische Technische Hochschule (ETH) Zurich, Fermi National Accelerator Laboratory, the University of Illinois at Urbana-Champaign, the Institut de Ciencies de l’Espai (IEEC/CSIC), the Institut de Fisica d’Altes Energies, Lawrence Berkeley National Laboratory, the Ludwig Maximilians Universitat Munchen and the associated Excellence Cluster Universe, the University of Michigan, NSF’s NOIRLab, the University of Nottingham, the Ohio State University, the University of Pennsylvania, the University of Portsmouth, SLAC National Accelerator Laboratory, Stanford University, the University of Sussex, and Texas A&M University.
BASS is a key project of the Telescope Access Program (TAP), which has been funded by the National Astronomical Observatories of China, the Chinese Academy of Sciences (the Strategic Priority Research Program “The Emergence of Cosmological Structures” Grant # XDB09000000), and the Special Fund for Astronomy from the Ministry of Finance. The BASS is also supported by the External Cooperation Program of Chinese Academy of Sciences (Grant # 114A11KYSB20160057), and Chinese National Natural Science Foundation (Grant # 12120101003, # 11433005).
The Legacy Survey team makes use of data products from the Near-Earth Object Wide-field Infrared Survey Explorer (NEOWISE), which is a project of the Jet Propulsion Laboratory/California Institute of Technology. NEOWISE is funded by the National Aeronautics and Space Administration.
The Legacy Surveys imaging of the DESI footprint is supported by the Director, Office of Science, Office of High Energy Physics of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231, by the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility under the same contract; and by the U.S. National Science Foundation, Division of Astronomical Sciences under Contract No. AST-0950945 to NOAO.
The authors wish to recognize and acknowledge the very significant cultural role and reverence that the summits of Maunakea and Haleakalā have always had within the Native Hawaiian community. We are most fortunate to have the opportunity to conduct observations from this mountain.
Appendix
Rates vs
Figure A1 shows the volumetric TDE rate as a function of the host-galaxy dust attenuation from our Bagpipes modeling (§ 2.3). The rates are computed exactly as in § 3. Each TDE enters its bin with the statistical weight , and uncertainties come from the same joint Poisson bootstrap used throughout. Of the 21 TDEs in our standard sample, 20 have host-galaxy SED fits; TDE 2024tvd is excluded as an off-nuclear event. The fitted attenuations span –2 mag. The rate is consistent with being flat across the full range, both in total and for the individual spectroscopic types.
Source of Data
The sources of the data used in this work are not homogeneous. Table B1 lists, for every TDE in the sample, the first publication (or TNS classification report) identifying it as a TDE and the sources of its host-galaxy properties, X-ray luminosity, and SMBH mass. For hosts previously analyzed in Shepherd et al. (2026), the original and host spectrum sources are given in their Table D1, and the H EW and Lick H sources in their § 3.2, unless otherwise noted in Table B1.
| IAU Name | Discovery Name | Disc./Class. | H EW | Lick H | Host spectrum | |||
| ASASSN-14ae | 1 | 2 | 2 | 2 | 2 | 1 | 3 | |
| ASASSN-14li | 4 | 2 | 2 | 2 | 2 | 4 | 3 | |
| ASASSN-15oi | 5 | 2 | 2 | 2 | 2 | 6 | 3 | |
| 2018dyb | ASASSN-18pg | 7 | 2 | 2 | 2 | 2 | 8 | 3 |
| 2018fyk | ASASSN-18ul | 9 | 2 | 2 | 10 | 2 | 9 | 3 |
| 2018hyz | ASASSN-18zj | 11 | 2 | 2 | 2 | 2 | 12 | 3 |
| 2018zr | PS18kh | 13 | 14 | This work | This work | This work | 14 | 3 |
| 2019ahk | ASASSN-19bt | 15 | 2 | 2 | 15 | 2 | 15 | 3 |
| 2019azh | ASASSN-19dj | 16 | 2 | 2 | 2 | 2 | 17 | 3 |
| 2019dsg | ZTF19aapreis | 16 | 2 | 2 | This work | 2 | 18 | 3 |
| 2019qiz | ZTF19abzrhgq | 16 | 2 | 2 | 2 | 2 | 19 | 3 |
| 2019vcb | ZTF19acspeuw | 20 | 20 | This work | This work | This work | 21 | 22 |
| 2020neh | ZTF20abgwfek | 23 | 24 | 24 | 24 | 24 | 23 | 3 |
| 2020nov | ZTF20abisysx | 25 | 2 | 2 | 2 | 2 | 25 | 25 |
| 2020vwl | Gaia20etp | 22 | 2 | 2 | 2 | 2 | This work | 3 |
| 2022bdw | ATLAS22dth | 26 | 2 | 2 | 2 | 2 | This work | 27 |
| 2022dble | ASASSN-22ci | 28 | 2 | 2 | 2 | 2 | 10 | 27 |
| 2022dsb | ASASSN-22cs | 29 | 29 | This work | This work | This work | 29 | 27 |
| 2022hvp | ZTF22aagyuao | 30 | 30 | This work | This work | This work | This work | 27 |
| 2022lri | ATLAS22pnz | 31 | 31 | This work | This work | This work | 31 | 31 |
| 2023clx | ASASSN-23bd | 32 | 2 | 2 | 2 | 2 | 33 | 27 |
| 2023mhs | ZTF23aaqdjhi | 34 | 2 | 2 | 2 | 2 | This work | 2 |
| 2024as | ATLAS24acr | 35 | This work | This work | This work | This work | This work | This work |
| 2024pvu | ZTF18abxnvoz | 36 | 37 | This work | This work | This work | This work | 37 |
| 2024tvdj | ZTF22aaigqsr | 38 | 38 | 39 |
Table B1. Sources of the data used in this work.
Binned Rates
Table C1 lists the binned -band LF shown in Figure 3, and Table C2 lists the binned peak blackbody-luminosity LF (Figure 4, left panel), both for the total sample and by spectroscopic type. Table C3 gives the TDE rate and rate enhancement in every region of the H EW–Lick H plane (§ 6.1). All other binned rates shown in this paper, including those for each subset, will be made available in machine-readable form on Zenodo upon publication.
| Total | TDE-H | TDE-He | TDE-H+He | ||||
| 42.2–42.6 | 1 | ||||||
| 42.6–43.0 | 7 | ||||||
| 43.0–43.4 | 7 | ||||||
| 43.4–43.8 | 5 | ||||||
| 43.8–44.2 | 1 |
Table C1. Binned peak -band LF (Figure 3), given as with , for the total standard sample, each spectroscopic type, and the two X-ray classes. Uncertainties are the 16th–84th percentiles of the bootstrap.
| Total | TDE-H | TDE-He | TDE-H+He | ||
| 42.5–43.1 | 1 | ||||
| 43.1–43.7 | 2 | ||||
| 43.7–44.3 | 11 | ||||
| 44.3–44.9 | 6 | ||||
| 44.9–45.5 | 1 |
Table C2. Binned peak blackbody-luminosity LF (Figure 4, left panel), given as with , for the total standard sample and each spectroscopic type. Uncertainties are the 16th–84th percentiles of the bootstrap ((Gehrels 1986) limits for ). Bins in which a type has no events are marked with an ellipsis.
References: (1) Holoien et al. (2014); (2) Shepherd et al. (2026); (3) Mummery et al. (2024); (4) Holoien et al. (2016a); (5) Holoien et al. (2016b); (6) Holoien et al. (2018); (7) Leloudas et al. (2019); (8) Holoien et al. (2020); (9) Wevers et al. (2019); (10) Hinkle et al. (2024); (11) Short et al. (2020); (12) Hung et al. (2020); (13) van Velzen et al. (2019); (14) Holoien et al. (2019b); (15) Holoien et al. (2019c); (16) van Velzen et al. (2021); (17) Hinkle et al. (2021a); (18) Cannizzaro et al. (2021); (19) Nicholl et al. (2020); (20) Hammerstein et al. (2023); (21) Quintin et al. (2023); (22) Yao et al. (2023); (23) Angus et al. (2022); (24) SDSS galSpec (Kauffmann et al. (2003); Brinchmann et al. (2004); Tremonti et al. (2004)); (25) Earl et al. (2025); (26) Arcavi et al. (2022); (27) Mummery & van Velzen (2025); (28) Lin et al. (2024); (29) Malyali et al. (2024); (30) Yao et al. (2025b); (31) Yao et al. (2024); (32) Zhu et al. (2023); (33) Hoogendam et al. (2024); (34) Newsome et al. (2023); (35) Brivio et al. (2024); (36) Jordana-Mitjans et al. (2025); (37) Yao et al. (2026b); (38) Yao et al. (2025a); (39) Patra et al. (2025).
Note. “This work” entries for H EW and Lick H were measured with Bagpipes and pyLick, respectively (§ 2.3).
Binned Rates
Note. is the number of TDEs in the bin. Bins in which a subset has no events are marked with an ellipsis. Subsets with events in a bin use the (Gehrels 1986) limits scaled to the measured rate (§ 3.3).
Note. The PSB, QBS, and PSB+QBS rows repeat Table 4 with the TDE rate added. excludes TDE 2023clx, whose host carries over a quarter of the total sample weight; where this changes , the reduced number is given in parentheses. Grid cells include their lower edges and exclude their upper edges. For regions with , the uncertainties on , , and are the (Gehrels 1986) limits scaled to the measured value (§ 3.3); uses the bootstrap throughout. All 20 TDE hosts with host-galaxy fits lie within the grid, but 26% of the control weight lies outside the plotted ranges, so sums to 0.74 over the grid cells. and are for the full sample.
References
- [10]Astropy Collaboration et al., 2013, A&A, 558, A33
- [12]Astropy Collaboration et al., 2022, ApJ, 935, 167
- [34]Chang J. N. Y., Dai L., Pfister H., Kar Chowdhury R., Natarajan P., 2025, ApJ, 980, L22
- [35]Chang J. N. Y., et al., 2026, ApJ, 1004, L14
- [51]Event Horizon Telescope Collaboration et al., 2022, ApJ, 930, L12
- [85]HI4PI Collaboration et al., 2016, A&A, 594, A116
- [99]Ho L. C., 2008, ARA&A, 46, 475
- [129]Lacerda E. A. D., Sánchez S. F., Cid Fernandes R., López-Cobá C., Espinosa-Ponce C., Galbany L., 2020, MNRAS, 492, 3073
- [134]Li W., et al., 2026, Research Notes of the American Astronomical Society, 10, 153
- [155]Neustadt J. M. M., et al., 2020, MNRAS, 494, 2538
- [179]Roming P. W. A., et al., 2005, Space Sci. Rev., 120, 95






