Far-Ultraviolet to Near-Infrared Observations of SN 2023ixf: A high energy explosion engulfed in complex circumstellar material

Massive stars ($\rm\gtrsim 8~{}M_{\odot}$) that meet their fate with the explosive phenomena are termed as Core-collapse Supernovae (CCSNe). CCSNe are either hydrogen-rich (Type II) or hydrogen-poor (Ib, Ic) (Filippenko, 1997). Recent advancements in all-sky surveys (e.g., ZTF, ATLAS) have made it possible to discover young supernovae when rapid changes occur in their light curves, spectral energy distribution, and spectral evolution apart from increasing brightness (Khazov et al., 2016; Bruch et al., 2022). The early evolution of a good fraction ($>36\%$) of Type II SNe is dominated by narrow emission features associated with confined dense circumstellar material (CSM, Bruch et al., 2021, 2022). The characteristics of the nearby dense CSM are visible in the spectral sequence as “flash” features consisting of narrow high-ionization lines that last a few to several days depending on the radius and density of the CSM (Gal-Yam et al., 2014; Yaron et al., 2017; Jacobson-Galán et al., 2022). The flash features are caused by the ionizing photons that result when the shock breaks out from the stellar surface and flashes/ionizes the nearby circumstellar material. Some authors (Kochanek, 2019; Jacobson-Galán et al., 2022) have noted that the ionization from shock breakout lasts for a few hours only, and to get prolonged flash features, another photon source is required, such as ejecta-CSM interaction. The earliest detailed time series observations of “flash spectroscopy” were observed for SN 2013fs (Yaron et al., 2017). The confined CSM ($<10^{15}$  cm) of SN 2013fs was indicated by the disappearance of flash features, and it was consistent with the radio non-detection (Yaron et al., 2017). It was argued that this could only result if the progenitor had undergone a short-lived episode of enhanced mass loss just a few years before the explosion. Even though the rise of all-sky surveys has led to an order-of-magnitude increase in the early detection (and follow-up) of such events (Blagorodnova et al., 2018; Nicholl, 2021), the physics behind the specifics of such interaction and origins of CSM are still not definitively understood, and the associated observables such as light curves are not very well constrained (Fuller, 2017; Wu & Fuller, 2021; Dessart & Hillier, 2022; Ko et al., 2022; Moriya et al., 2023).

The CCSNe have been studied extensively in optical and near-infrared (NIR) regimes, but extensive studies in the ultraviolet (UV) regime are still limited (Pritchard et al., 2014; Brown et al., 2007; Vasylyev et al., 2023). The crucial aspect of the observational investigation in UV is the requirement of observation from space-based missions (Vasylyev et al., 2022; Bostroem et al., 2023a) for which scheduling time disruptive Target-of-Opportunity (ToO) observations is not rapid for a majority of missions. The flux in UV declines very rapidly, requiring prompt observations and follow-ups. The UV emission from young CCSNe allows the investigation of hot and dense ejecta and/or the presence of CSM when the photosphere is located in the outer layers of the progenitor star (Bufano et al., 2009). Many Type II SNe show a nearly featureless early optical spectral sequence, unlike the far-UV and near-UV, which showcases a plethora of metal features. The detection of these features can be used to determine the composition of the outer envelope of the pre-SN star, the temperature of the outer layers of the ejecta, or the CSM and its characteristics (Dessart et al., 2022; Bostroem et al., 2023a).

SN 2023ixf was discovered on 2023 May 19 17:27:15.00 UT (JD 2460084.23) in the galaxy M 101 at $\sim 14.9$ mag in ‘clear’ filter (Itagaki, 2023) and classified as a Type II SN (Perley & Gal-Yam, 2023; Teja et al., 2023a). The pre-discovery photometry from Zwicky Transient Facility (ZTF) and other Transient Name Server (TNS) alerts provide tight constraints on the explosion epoch. Using the last non-detection (JD 2460083.31) and first detection (JD 2460083.32) (Chufarin et al., 2023), we find the explosion epoch, $\rm t_{exp}=JD~{}2460083.315\pm 0.005$ which has been used throughout this work. We note that the last-non detection used is not very deep ($>$ 18 mag) and if we consider the deeper non-detection ($>20.5$ mag, Mao et al., 2023) on JD 2460083.16, the explosion epoch has a marginal change (of $\sim$ 0.08 d) to JD 24600083.235.

Several professional and amateur astronomers have followed up on SN 2023ixf, being one of the nearest CCSNe in the last 25 years. Various time-domain groups across the globe have been monitoring it soon after its discovery, and the results based on the early observations have already been presented. The early phase optical and NIR photometry and optical spectroscopy have been presented by Yamanaka et al. (2023); Hosseinzadeh et al. (2023); Jacobson-Galan et al. (2023). The presence of flash features in the spectra and increased luminosity is interpreted as due to the presence of nitrogen/helium-rich dense CSM and the interaction of supernova ejecta with it (Yamanaka et al., 2023; Jacobson-Galan et al., 2023). By comparing the early light curve with the shock cooling emission (Hosseinzadeh et al., 2023) suggested that the progenitor of SN 2023ixf could be a red supergiant with radius 410$\pm$ 10 R_⊙. The high-resolution spectroscopy revealed that the confined CSM is asymmetric (Smith et al., 2023). Pre-imaging data was analyzed at the SN 2023ixf site in recent works, constraining the mass of the progenitor between $12-17$ M_⊙ (Jencson et al., 2023; Pledger & Shara, 2023; Soraisam et al., 2023). These estimates are well within the earlier detected CCSNe progenitors (Smartt, 2009; Van Dyk, 2017).

This letter presents the panchromatic evolution of SN 2023ixf spanning the far-ultraviolet (FUV) to near-infrared (NIR) wavelengths during the first three weeks since its discovery. The flow of the paper is as follows: In Section 2, we estimate the distance to the host galaxy and its extinction and briefly describe the source of data acquisition and the reduction procedure. Further, we present our spectroscopic observation in Section 3 along with its analysis and modeling in different regimes. Later on in Section 4, we describe the light curve evolution and its early phase analysis. We summarize and discuss this early phase work in Section 5.

SN 2023ixf exploded in the outer spiral arm of the host galaxy, M 101, a face-on giant spiral galaxy that lies comparatively close to the Local Group. Tikhonov et al. (2015) estimated a mean distance of 6.79 $\pm$ 0.14 Mpc ($\rm\mu$ = 29.15 $\pm$ 0.05 mag) to M 101 using the tip of the RGB method (Lee et al., 1993) with low-uncertainty. Riess et al. (2022) used Cepheids to estimate a distance of 6.85 $\pm$ 0.15 Mpc ($\rm\mu$ = 29.18 $\pm$ 0.04 mag). We adopt a mean distance of 6.82 $\pm$ 0.14 Mpc ($\rm\mu$ = 29.17 $\pm$ 0.04 mag). The gas phase metallicity was computed by Garner et al. (2022) using various H II regions in the galaxy and estimated an oxygen abundance of 12 + log[O/H] $\sim\,$ 8.7 in the outer spiral arms of the galaxy which is similar to solar metallicity (Asplund et al., 2009).

Galactic reddening in the line-of-sight of SN 2023ixf inferred from the dust-extinction map of Schlafly & Finkbeiner (2011) is $E(B-V)$ = 0.0077$\pm$0.0002 mag. Using high-resolution data, Lundquist et al. (2023) computed equivalent widths of $\rm Na\,$I D1 and D2 lines to be 0.118 Å and 0.169 Å, respectively. Using the relation from Poznanski et al. (2012), we infer a mean host reddening of $E(B-V)$ = 0.031 $\pm$ 0.011 mag using $\rm Na\,$I D1 and D2. A total reddening of $E(B-V)$ = 0.039 $\pm$ 0.011 mag is adopted for SN 2023ixf which is consistent with Smith et al. (2023).

The multiband light curves based on observations from the various facilities described in Section 2 are shown in Figure 3. We converted all the pre and post-discovery public data to AB magnitude scale and included it with our dataset using the transformations described in Blanton & Roweis (2007). The public dataset reported is very helpful in putting tight constraints on the explosion epoch (Hosseinzadeh et al., 2023).

We do not see the fast declining phase after maximum, the $s1$ phase, usually attributed to the initial cooling phase post-breakout (Anderson et al., 2014), in the $V$-band light curve of SN 2023ixf. Instead, it declines very slowly for the initial few days, at $\rm 1.18^{+0.49}_{-0.51}~{}\rm{mag}~{}100~{}d^{-1}$, right after it reached a peak $V-$band magnitude of $-18.06\pm 0.07$ mag around $\sim 5$ d after explosion. The peak magnitude falls at the brighter end of Type II SNe. The peak V-band brightness is comparable with SN 2013by (Valenti et al., 2015) and SN 2014G (Terreran et al., 2016), which were classified as Type IIL, although with many similarities to the Type IIP sub-class. SN 2014G also showed flash ionization features in early spectral evolution. While the initial decline of SN 2023ixf is inconsistent with that of Type IIL, its evolution at later phases is yet to be probed. Although the early spectra indicate interaction with a nearby dense CSM, SN 2023ixf is not extremely bright in the UV bands like Type IIn SNe.

The observed rise time of $\sim 4-5$ d is shorter than other normal Type II SNe, which, on average, take $\sim$ 10 days to reach the peak (Valenti et al., 2016). We compare the $g-r$ color with similar events that showed flash features such as SN 2013by (Valenti et al., 2015; Black et al., 2017), SN 2014G (Terreran et al., 2016), and the bluest Type II SN 2020pni (Terreran et al., 2022). The color evolution is similar to these events for the initial $\sim$ 20 d but slightly redder than SN 2020pni. The NIR light curves are also presented in Yamanaka et al. (2023) up to a week post-explosion. We show the evolution beyond that and observe that the flux increases in the NIR, possibly due to pre-existing dust around the ejecta. The presence of pre-SN dust is also described in Neustadt et al. (2023).

The early prolonged flash features indicated the presence of a dense CSM around the progenitor. Recently, Moriya et al. (2023) provided a comprehensive set of grids for model light curves that could shed light on the structure of CSM and its effects on the early light curve of interacting Type II SNe. In their work, a confined CSM is attached over radius, $R_{0}$, for five progenitors with mass ranging from 10 to 18 M_⊙. The CSM density structure follows from Moriya et al. (2018), whereas the wind velocity, $v_{wind}$ at a distance $r$ was taken to be in the form as given below:

where $v_{0}$ and $v_{\infty}$ are the initial wind velocity at the surface of the progenitor and terminal velocity, respectively, and $\beta$ is a wind structure parameter that determines the efficiency of wind acceleration.

These model light curves can be used to constrain the very early light curve behavior of Type II SNe. Our work utilizes the well-sampled $g$-band light curve of SN 2023ixf to compare with the model grid of interacting Type II SNe generated by Moriya et al. (2023). We used the models with $\rm{}^{56}Ni$ mass in the typical range of 0.01 to 0.04 M_⊙ (Anderson et al., 2014). Furthermore, we found that the initial light curves are insensitive to the $\rm{}^{56}Ni$ mass. We iterated over each parameter ($\rm E_{exp}$, $\beta$, $R_{CSM}$, and $\dot{M}$) in succession, keeping others fixed with their full range for a single run. This procedure is repeated for 12, 14, 16, and 18 $\rm M_{\odot}$ progenitor models. We categorically reject models which show significant deviations from the observed light curves based on their peak luminosities and rise times. Subsequently, we do this for other parameters constraining the values for previous parameters. The best-fitting models for each progenitor are shown in Figure 3. We note that the slow early rise till day 2 is not captured by any of the models, and the later evolution is such that either the rise or plateau could be matched but not the entire light curve. Since we are concerned about the initial rise, we do not probe it further; detailed hydrodynamical modeling specific to this particular event will be required to understand the entire light curve evolution. Further, the degeneracy in the progenitor masses could not be lifted by these models, but these models give a very tight constraint on the radius of the outer CSM utilizing the rise times of the model light curves. The dense CSM is confined to $\rm 4.0-10.0\times 10^{14}~{}cm$. Further, $\beta$ varies from 0.5 to 1.5 depending on the progenitor mass, which is close to the typical values for RSGs ($\beta>1$). The $\beta<1$ value obtained for $\rm M_{ZAMS}$ would accelerate winds slightly faster and cause less dense CSM in the vicinity, which is not the case for SN 2023ixf. The mass loss rate is also slightly on the higher end ($\rm 10^{-3.0\pm 0.5}~{}M_{\odot}~{}yr^{-1}$). The average density of the CSM comes out to be $\rm\sim 10^{-14}~{}g~{}cm^{-3}$ which is in line with the values obtained in Bostroem et al. (2023b) but below the values inferred in Jacobson-Galan et al. (2023) obtained from the detailed spectral modeling. The mass-loss rates align with the density limits of CSM derived from the non-detection of radio emission (230 GHz) at early times (Berger et al., 2023). For a typical RSG ($\sim 500~{}M_{\odot}$), the above would translate to a mass loss $\sim 14-18$ years before the explosion. But as seen in Smith et al. (2023), wind speeds measured using high-resolution early spectra are one order higher than what is assumed in the model parameters, which would give an eruptive mass loss timeline to be around 2 years before the explosion. However, wind acceleration cannot be ruled out. Another parameter that is tightly constrained by the models is the explosion energy. Only the models with explosion energies more than 2.0 foe could match the observed $g$-band flux. The explosion energy increases as the progenitor mass is increased. The explosion energy obtained is higher than for the usual Type II SNe.

In a recent work, Khatami & Kasen (2023) presented various light curves of transients arising from interacting SN. These include the interaction of SN ejecta with no CSM to a very heavy CSM. Considering the latent space of luminosity and rise-time presented in that work, we find that the light curve evolution of SN 2023ixf (for the period presented in this work) appears to be similar to the model light curves for shock-breakout in a light-CSM scenario. Comparing the rise-times and peak luminosity of SN 2023ixf with the shock-breakout happening inside the CSM, we find that it falls within 0.01 $<$ M_CSM [M_⊙] $<$ 0.1. Using the parameters obtained from light curve analysis, we get a CSM mass ranging from 0.001-0.03 M_⊙ (assuming $v_{\rm{wind}}=10$ km $s^{-1}$), where the upper limit is well within the range obtained from Khatami & Kasen (2023). It indicates the mass loss rate could have been even higher than $\rm 10^{-2.5}~{}M_{\odot}~{}yr^{-1}$, as also being reported in Jacobson-Galan et al. (2023); Hiramatsu et al. (2023).

This work presents an extensive set of early-phase observations for the closest CCSN in the last 25 years, SN 2023ixf, that exploded in M 101. The panchromatic observations covered wavelengths from the FUV to NIR regime using both ground and space-based observatories. The multi-band photometry spans FUV to NIR, spanning up to $\sim$ 23 d since the explosion. Light curves were compared with a large model light curves grid to infer nearby dense CSM properties.

Detailed spectral coverage in FUV, NUV, and optical during the first $\sim 25$ days since the explosion is presented, beginning within 2 days from the explosion. The lines due to Mg II, Fe II in the NUV, and C III, C II, Si IV, He II in the FUV were identified. The early ($<$ 7 d) spectral sequence of SN 2023ixf indicates the presence of a dense CSM. There are no significant signatures subsequently, except for an intermediate-width emission feature of H$\rm\,\alpha$ after $+7$ d. The high-resolution spectra presented by Smith et al. (2023) show the presence of an intermediate-width P-Cygni profile during this phase, lasting for about a week, arising in the post-shock, swept-up CSM shell. The line profile during the photospheric phase beginning $\sim$ 16 d shows a multi-peaked/boxy profile of H$\rm\,alpha$, indicating an ongoing CSM interaction with a shell-shaped CSM with an inner radius of $\sim$ 75 AU and an outer radius of $\sim$ 140 AU. Considering a standard RSG wind velocity, the progenitor likely experienced enhanced mass-loss $\sim$ 35 - 65 years before the explosion. All the above inferences from our multi-wavelength observations indicate a multi-faceted circumstellar matter around the progenitor of SN 2023ixf.

The early phase light curve of SN 2023ixf is influenced by the presence of dense nearby CSM, which was likely accumulated due to enhanced mass loss(es) during the later stages of the progenitor’s evolution. SN 2023ixf was found to have a very bright peak luminosity ($M_{V}\approx-18.1$ mag), much higher than the average luminosity for Type II SNe ($M_{V}\approx-16.7$ mag). Light curves were compared with a large model grid of interacting SNe with varied progenitor masses and CSM properties to infer the properties of the dense CSM in SN 2023ixf. Based on our comparison with light curve models, the high luminosity is likely a mix of interaction with a confined CSM and an inherently energetic explosion. We cannot conclusively decipher the weightage of the above components to the overall luminosity of SN 2023ixf; hence, further monitoring is required. We will continue to carry out the multi-wavelength follow-up of SN 2023ixf.

.

We thank the anonymous referee for an in-depth review that helped improve the manuscript. RST thanks Sergiy S. Vasylyev for providing HST spectra of SN 2021yja. The GROWTH India Telescope (GIT) is a 70-cm telescope with a 0.7-degree field of view, set up by the Indian Institute of Astrophysics (IIA) and the Indian Institute of Technology Bombay (IITB) with funding from Indo-US Science and Technology Forum and the Science and Engineering Research Board, Department of Science and Technology, Government of India. It is located at the Indian Astronomical Observatory (IAO, Hanle). We acknowledge funding by the IITB alumni batch of 1994, which partially supports the operation of the telescope. Telescope technical details are available at https://sites.google.com/view/growthindia/. The HCT observations were made under our accepted ToO proposal HCT-2023-C2-P25. We thank the staff of IAO, Hanle, and CREST, Hosakote, that made these observations possible. The facilities at IAO and CREST are operated by the Indian Institute of Astrophysics, Bangalore. DKS acknowledges the support provided by DST-JSPS under grant number DST/INT/JSPS/P 363/2022. This research has made use of the High-Performance Computing (HPC) resources⁵⁵5https://www.iiap.res.in/?q=facilities/computing/nova made available by the Computer Center of the Indian Institute of Astrophysics, Bangalore. This research made use of RedPipe⁶⁶6https://github.com/sPaMFouR/RedPipe (Singh, 2021), an assemblage of data reduction and analysis scripts written by AS. This work uses the SXT and UVIT data from the AstroSat mission of the Indian Space Research Organisation (ISRO). The SN was observed through multiple ToO proposals, and the data were made public through the ISSDC data archive. We thank the SXT and UVIT payload operation centers for verifying and releasing the data via the ISSDC data archive and providing the necessary software tools. This work has also used software and/or web tools obtained from NASA’s High Energy Astrophysics Science Archive Research Center (HEASARC), a service of the Goddard Space Flight Center and the Smithsonian Astrophysical Observatory. This work was also partially supported by a Leverhulme Trust Research Project Grant. This research also made use of the NASA/IPAC Extragalactic Database (NED⁷⁷7https://ned.ipac.caltech.edu), which is funded by the National Aeronautics and Space Administration and operated by the California Institute of Technology.