Recovering Lost Light: Discovery of Supernova Remnants with Integral Field Spectroscopy

We present results from a systematic search for broad (≥ 400 km s−1) Hα emission in integral field spectroscopy data cubes of ∼1200 nearby galaxies obtained with PMAS and MUSE. We found 19 unique regions that pass our quality cuts, four of which match the locations of previously discovered supernovae (SNe): one Type IIP and three Type IIn, including the well-known SN 2005ip. We suggest that these objects are young Supernova remnants (SNRs), with bright and broad Hα emission powered by the interaction between the SN ejecta and dense circumstellar material. The stellar ages measured at the locations of these SNR candidates are systematically lower by about 0.5 dex than those measured at the locations of core-collapse (CC) SNe, implying that their progenitors might be shorter lived and therefore more massive than a typical CCSN progenitor. The methods laid out in this work open a new window into the study of nearby SNe with integral field spectroscopy.

1. INTRODUCTION Supernovae (SNe) are energetic stellar explosions that mark the endpoints in the life of certain types of stars.Although they are rare events, occurring once or twice per century in a typical galaxy, SNe are essential to understanding the chemical evolution of the Universe (Kobayashi et al. 2006;Andrews et al. 2016;Prantzos et al. 2018) and the injection of energy into the interstellar medium (ISM, Thornton et al. 1998).This local deposition of energy plays a crucial role in galaxy evolution, triggering star formation (Stinson et al. 2006;Dalla Vecchia & Schaye 2012;Hopkins et al. 2014) and seeding and sustaining turbulence (Mac Low & Klessen 2004).
Our understanding of the role that SNe play in the ecology of the ISM is limited by the many questions that remain open about their stellar progenitors.The baseline physical scenarios for the two major classes of SNe posit that core collapse (CC) SNe arise from gravitational collapse in the cores of stars more massive than ∼ 8 M ⊙ , and Type Ia SNe are the aftermath of a thermonuclear runaway ignited in the central regions of massive CO white dwarfs (WDs) that are somehow destabilized by accretion in close binary systems.However, many important details of these scenarios remain obscure, including the role played by binary interactions and pre-explosion progenitor mass loss in CCSNe (Smartt 2009;Smith 2014), and the specific identity of the progenitors in SN Ia (Maoz et al. 2014).As a result, we have not been able to fully characterize fundamental properties that regulate the feedback mechanism into the ISM, like the distribution of progenitor masses and delay times (Badenes et al. 2009;Jennings et al. 2014;Zapartas et al. 2017;Auchettl et al. 2019;Strolger et al. 2020;Castrillo et al. 2021) or the detailed nucleosynthetic yields (Romano et al. 2010;Andrews et al. 2017).
Recent advances in astronomical instrumentation have opened new opportunities to explore these issues.In particular, the wide availability of data from Integral Field Spectroscopy (IFS) makes it possible to study the properties of the host galaxies of SNe with an unprecedented level of detail, revealing a wealth of information on the environments in which SN explode (e.g.Stanishev et al. 2012;Kuncarayakti et al. 2013a,b;Rigault et al. 2013;Galbany et al. 2014Galbany et al. , 2016aGalbany et al. ,b, 2017;;Krühler et al. 2017;Galbany et al. 2018a;Kuncarayakti et al. 2018;Lyman et al. 2018;Rigault et al. 2018;Lyman et al. 2020).Previous IFS studies of SNe have relied on dedicated SN surveys and literature searches to identify SN host galaxies and locate their explosion sites.Here, we describe the first results of an effort to use the IFS observations themselves to discover recent SNe.This is possible because some SNe undergo strong interaction with a dense circumstellar medium (CSM) shortly after the explosion, leading to large luminosities that can be sustained for a long time (Milisavljevic & Fesen 2017;Dessart et al. 2023).These objects are often described as transitional, between SNe and Supernova Remnants (SNRs), and range from young SNe with clear evidence of CSM interaction such as SN 1993J (Fransson et al. 1996;Matheson et al. 2000), SN1996cr (Bauer et al. 2008;Quirola-Vásquez et al. 2019), and SN 1978K (Kuncarayakti et al. 2016) to SNRs like Cas A, which remain optically bright centuries after the explosion (Milisavljevic et al. 2012).For clarity, we will refer to this class of objects as young SNRs.Our work is motivated by the fact that some of these young SNRs are luminous enough to stand out in narrow-band images of their host galaxies (e.g., SNR NGC4449, Milisavljevic & Fesen 2008), and in principle could be picked up serendipitously in IFS observations of nearby galaxies.To explore this possibility, we have conducted a systematic search for regions in IFS data cubes that are characterized by bright Hα line emission with a significant broad component (≥ 400 km s −1 ) that could be associated with CSM interaction.This paper is organized as follows.In Section 2, we describe the IFS data that constitute our initial sample.In Section 3, we outline the methods we have employed to identify candidate young SNRs.In Section 4, we discuss the main properties of our sample of young SNR candidates.In Section 5 we comment on our results and discuss possible avenues for future research.

OBSERVATIONS
Our initial sample is constituted by IFS observations of nearby galaxies obtained using two different instruments: (i) the Potsdam Multi Aperture Spectograph (PMAS; Roth et al. 2005) in PPak mode (Verheijen et al. 2004;Kelz et al. 2006), mounted on the 3.5 m telescope of the Centro Astronomico Hispano-Aleman (CAHA) at the Calar Alto Observatory.PPak consists of a fiber bundle of 382 fibers with 2.7" diameter.Among these, 331 are ordered in a single hexagonal bundle, with the remaining fibers used for sky measurements and calibration purposes.Most objects are observed with two overlapping setups and then combined.The V500 grating has a spectral resolution of ∼6 Å in the wavelength range 3750−7300 Å.The V1200 grating has a higher spectral resolution of ∼2.7 Å, with a bluer range spanning 3400-4750 Å.The final products are 3D datacubes with a 100% covering factor within a hexagonal FoV of ∼1.3 arcmin 2 with 1"×1" pixels, which correspond to ∼4000 spectra per object.
(ii) the Multi-Unit Spectroscopic Explorer (MUSE; Bacon et al. 2014), located at the Nasmyth B focus of Yepun, the VLT UT4 telescope at Cerro Paranal Observatory.It has a modular structure composed of 24 identical IFU modules that together sample, in Wide Field Mode (WFM), a near-contiguous 1 arcmin 2 FoV with spaxels of 0.2 × 0.2 arcsec, and a wavelength coverage of 4650-9300 Å with a mean resolution of R ∼3000.MUSE produces ∼100,000 spectra per pointing.
Observations obtained with PMAS come in their majority from the 3rd data release of the Calar Alto Legacy Integral Field Area survey (CALIFA; Sánchez et al. 2012Sánchez et al. , 2016)).The PMAS sample consists of 667 galaxies selected from DR7 of the Sloan Digital Sky Survey (SDSS, Abazajian et al. 2009) with redshifts between 0.005 and 0.03, declinations >-7 • , and diameters between 45" and 79.2".These selection criteria are optimized for the PMAS/PPak instrument.We added to this sample data from the PMAS/PPak Integral field Supernova hosts COmpilation (PISCO; Galbany et al. 2018a), an extended project of CALIFA that aimed at increasing the sample of supernova host galaxies used for environmental studies (Galbany et al. 2014(Galbany et al. , 2016b)).As of January 2020, the PISCO sample contained 220 galaxies, bringing the total PMAS sample to 887 galaxies.
Observations obtained with MUSE come in its totality from the All-weather MUse Supernova Integral-field of Nearby Galaxies (AMUSING; Galbany et al. 2016a;Galbany et al. in prep.)survey.This survey has been running for 11 semesters (P95-P106) with the focus of obtaining IFS of SN host galaxies, each semester with a different science focus including among others: hosts of SNe that showed strong sodium absorption lines in their spectra indicating the presence of large amounts of dust; hosts of SNe discovered by the ASASSN-SN survey to study SN rates as a function of local environment (Pessi et al. 2023); SN hosts with low surface brightness (Holoien et al. 2022); hosts of SNe included in the Carnegie Supernova Project (CSP; Phillips et al. 2019) sample.The compilation used in this work consisted of 342 nearby SN host galaxies.

Pre-processing
All IFS datacubes are pre-processed in preparation for the analysis.First, regions with foreground stars are masked out.Then, single stellar population (SSP) synthesis models are fit to all individual spectra (around 4,000 and 100,000 spectra per PMAS and MUSE cube, respectively) to separate the underlying stellar continuum from the ionized gas-phase emission.This is done with STARLIGHT (Cid Fernandes et al. 2005Fernandes et al. , 2009)), which determines the fractional contribution of different SSP models to the spectrum, x i , and to the galaxy mass, µ i .STARLIGHT accounts for dust extinction (A * V ) as a foreground screen, using a reddening law from Cardelli et al. (1989) that assumes R V =3.1.To reduce computing time, we used a reduced grid of 248 SSP models from Cid Fernandes et al. (2013), with 62 ages spanning 1 Myr to 14 Gyr and four metallicities (0.2, 0.4, 1.0, and 1.5 solar).The best fit model is subtracted from each observed spectrum and the resulting residual is stored in a 3D cube containing only the ionized gas-phase emission.
The resulting residual spectra are mostly flat, dominated by the narrow line emission from the host galaxy.We fit the most prominent lines (Hβ, [O iii] λ5007, [O iii] λ4969, [N ii] λλ6548,84, Hα, [S ii] λλ6719,31) with the Python version of MPFIT (Markwardt 2009(Markwardt , 2012;;Newville et al. 2016), which performs non-linear leastsquares fitting using the Levenberg-Marquardt minimization algorithm (Levenberg 1944;Marquardt 1963).For each residual spectrum, we focus on the wavelength range corresponding to each line and fit it using a Gaussian profile to retrieve line fluxes, central wavelengths, velocity shifts and widths, and their corresponding errors.When required, more than one Gaussian is fitted simultaneously for transitions that are close to one another (Hα with the two [N ii], the two [S ii], and Hβ with [O iii]).
We correct all the fitted line fluxes for reddening intrinsic to the host galaxy by measuring the Hα to Hβ ratio and applying an extinction law from Cardelli et al. (1989), assuming R V = 3.1, Case B recombination, and densities of ∼10 3 cm −3 around a heating source with T∼10 4 K and a large optical depth (Osterbrock & Ferland 2006).We note that the extinction correction and the selection of a particular extinction law do not affect the method we use to detect young SNR candidates.

Line emission from young SNRs in IFS data: an illustrative example
Young SNRs have distinctive optical spectra.In Figure 1 we show the integrated line emission in two spectral windows around the [O iii] λ5007 and Hα lines from the Galactic SNR Cas A (age ∼340 yr, Thorstensen et al. 2001, Milisavljevic et al. 2012, the SNR in NGC 4449 (age ∼80 yr, Milisavljevic & Fesen 2008, Bietenholz et al. 2010), and SN 2005ip (age ∼17 yr), which we use as an illustrative example of a young SNR imaged by our IFS data.SN 2005ip is a bright, well-observed Type IIn SN that exploded in NGC 2906 (Fox et al. 2009(Fox et al. , 2010) ) and has been showing signs of strong CSM interaction and enhanced dust formation for well over a decade (Smith et al. 2009;Stritzinger et al. 2012;Katsuda et al. 2014;Smith et al. 2017;Bevan et al. 2019;Fox et al. 2020).All three objects show bright emission in both lines, with Cas A and SNR NGC4449 being brighter in [O iii], and SN 2005ip being brighter in Hα.The Hα emission is noticeably broadened by several hundred km s −1 due to shock interactions in all three objects, although in all cases this broad emission appears superimposed on several components of narrow emission from Hα and the neighboring [N ii] λλ6548,84 doublet.The [O iii] λ5007 line is broadened by several thousand km s −1 in both Cas A and SNR NGC4449, but the spectrum from SN 2005ip appears much narrower, even if the neighboring Hβ line has a clear blueshifted broad component.
A detailed physical interpretation of the line emission from young SNRs is outside the scope of the present work.In particular, the mapping between shock physics and line emission is likely to be complex -we note that our IFS data only provide an unresolved view of objects that probably have a great deal of spatial structure, which can only be studied in detail in nearby cases like Cas A (Milisavljevic et al. 2012;Milisavljevic & Fesen 2015).We refer the reader to Fransson et al. (2002) and Milisavljevic et al. (2012) for further details.For our purposes, the qualitative comparisons shown in Figure To evaluate our IFS observations of SN 2005ip in context, we have compiled the 30 single-slit spectra of the SN analyzed by Smith et al. (2009) and Smith et al. (2017), which span ages between 17 days and more than 10 years after discovery.As noted in these references,  The broad component to the Hα emission associated with SN 2005ip does not appear at any other location in the IFS data cubes for the host galaxy NGC 2906.To illustrate this, we added a broad emission component to the fits to the residual spectra in the Hα region for all spaxels, in addition to the narrow Hα and the two narrow [N ii] lines.This broad component has a minimum width of 400 km s −1 , but no minimum amplitude, so that MPFIT returns a zero value for the amplitude when the spectral fit does not require it.For the reminder of the paper, we will not distinguish be-tween 'broad' and 'intermediate' components to the Hα line emission, as defined in the case of SN 2005ip, but will use the notation 'broad' to describe any contribution to the line flux that is clearly broader than usual (i.e., that is not kinematically narrow).Figure 4 shows the flux maps for NGC 2906 generated by this procedure for both the narrow (left column) and broad (right column) Hα components, in the PMAS (top row) and MUSE (bottom row) data.The narrow component is distributed throughout the entire disk of the galaxy, showcasing emission from individual HII regions, but the broad component is restricted exclusively to the location of SN 2005ip.Note how the higher spatial resolution of MUSE (0.2" compared to 1" for PMAS) allows to accurately pinpoint the site of the SN.
Without spatial coincidence with a previously recorded SN, as seen here for SN 2005ip, it is impossible to determine with absolute confidence that a region showing broad line emission in an IFS data cube is in fact a young SNR.In general, broad line emission in IFS data cubes is often associated with Active Galactic Nuclei in the central regions (see e.g.Papaderos et al. 2013;Singh et al. 2013;Lacerda et al. 2020).Off-center broad line emission has been associated with sources like stellar outflows from supergiants or Wolf-Rayet clusters (e.g.Diaz et al. 1987, Terlevich et al. 1991, Kehrig et al. 2020), but the vast majority of these examples have low luminosities, which would be hard to disentangle from the background in IFS data cubes, and modest widths, with FWHM below 200 km s −1 .In rare cases, bright Hα emission with FWHM in excess of 1000 km s −1 has been reported in single-slit spectra, like the giant HII region NGC 5471 in M101 (Castaneda et al. 1990).For this specific example, follow-up X-ray observations indicate the presence of at least one SNR at this location (Williams & Chu 1995).Of course, larger scale ionized flows outflows have been found in starburst and post-starburst galaxies like the 'green peas' (Rodríguez Del Pino et al. 2019;Amorín et al. 2012;Hogarth et al. 2020), but these outflows will not appear as high contrast point sources like the IFS detection of SN 2005ip shown in Figure 4.

Search method
The illustrative example of SN 2005ip demonstrates that IFS data cubes have the potential to both recover emission from young SNRs and extract relevant physical information about them.Motivated by this realization, we have conducted a systematic search for broad Hα emission in all the IFS data of all the nearby galaxies in our sample.To do this, we defined a set of criteria designed to single out young SNR candidates without making strong assumptions about the specific properties of their line emission.First, we require a minimum FWHM of 400 km s −1 , which should be enough to remove most non-SN local outflows.Second, we use the flux (F ), flux uncertainty (∆F ), width (σ), and width uncertainty (∆σ) of the Gaussian fits to the broad Hα component (denoted as Hα,SNR) to define two diagnostic ratios designed to single out bright line emission from young SNR candidates: • log (σ Hα,SNR /∆σ Hα,SNR ) • log F 2 Hα,SNR /∆F Hα,SNR We show the values for these two diagnostic ratios in all the spaxels of the MUSE data cube for NGC 2906 in Figure 5.This plot illustrates the discriminating power of our chosen diagnostic ratios, with the spaxels that cover the site of SN 2005ip clearly deviating from the distribution of values measured in the rest of the host galaxy along both axes.
We have examined the values of these diagnostic ratios in all IFS data cubes in our sample of 887 PMAS and 342 MUSE observations of nearby galaxies, consisting of more than 35 million individual spectra.To minimize false detections, we restrict our search to spaxels where the signal-to-noise ratio (S/N) for Hα, SNR is higher than 5 in the residual spectra.We have set a conservative threshold of 3σ above the median of the distribution in both diagnostic ratios to flag regions of interest as a young SNR candidates.
As a side note, we attempted a similar search centered in the [O iii] λ5007 line, but we failed to find any clear candidates.This indicates that objects with strong, broad [O iii] emission like Cas A and SNR NGC 4449 (Figure 1) must be rare, or short-lived, or both.

Young SNR Candidates
Our search yielded 20 contiguous regions that pass our quality cuts, 7 in MUSE and 13 in PMAS.We list these regions in Table 2, and classify them as young SNR candidates, by analogy with the properties of SN 2005ip described in Section 3.2.The only object that appears in both MUSE and PMAS data is SN 2005ip itself, which brings the total number of unique objects identified in our search to 19.The individual fits to the residual spectra around Hα in the spaxels identified as young SNR candidates are shown in Figures 6, 7, and 8, along with maps of the entire host galaxy, both in the broad Hα component and in white light.The regions outside the SNR candidates that appear in some of these broad Hα maps did not pass our quality cuts.In Figure 9 we show the distribution of host galaxy redshifts, along with the FWHM, luminosity, and systemic velocity of the broad Hα component corresponding to these regions.Four of these 19 objects coincide spatially with previously known SNe: the already discussed SN 2005ip, a Type IIn SN in the spiral NGC 2906 imaged by PMAS and MUSE 2593 and 3104 days after discovery, respectively; ASAS-SN 14fd (Holoien et al. 2019), a Type IIn SN in the dwarf irregular galaxy LEDA 43070 (PGC 43070) imaged by MUSE 514 days after discovery; SN2011fh (Pessi et al. 2022), a Type IIn SN in the spiral NGC 4806 imaged by MUSE 1362 days after discovery; and ASAS-SN 14jb (Holoien et al. 2019), a Type IIP SN in the spiral ESO 467-51, imaged by MUSE 391 days after discovery.A fifth object, the SNR candidate in NGC 5908, coincides spatially with PSNJ15164204+5525011, a 'SN impostor' reported in 2012, likely a luminous blue variable (Benetti et al. 2012).
It is worth noting that three out of four SNe that coincide with our SNR candidates are Type IIn, despite the fact that this subtype only accounts for ∼9% of CC SNe (Smith et al. 2011;Kiewe et al. 2012).The fourth SN, ASAS-SN 14jb, is a rare extraplanar Type IIP SN in an edge-on spiral, whose MUSE observations were analyzed and discussed in Meza et al. (2019).The remaining 15 candidates in our sample are probably young SNRs whose SN either exploded before the era of modern transient surveys, or were missed, perhaps because of weather, or poor sampling, or coincidence with the Sun.With distances up to ∼100 Mpc (for SNR Arp 142, at a redshift of 0.0233, see Figure 9), these are among the furthest SNRs identified as such.Interestingly, one of our candidate SNRs is located in NGC 6946, the 'Fireworks Galaxy', a nearby spiral with a high star formation rate that has hosted 10 known SNe (Eldridge & Xiao 2019; Right column: image of the IFS data footprint in white light.Eibensteiner et al. 2022).Our results bring the total number of SNe in this galaxy up to eleven.With all the caveats attached to small number statistics, the four objects associated with known SNe in our sample suggest that our method is turning up SNRs whose Hα luminosity is driven predominantly by CSM interaction, likely core collapse SNe whose progenitors have lost a great deal of mass in their pre-SN evolution, either due to winds or to binary interactions (Smith et al. 2009;Langer 2012;Dessart et al. 2023).Some (possibly most) of these SNe might have shown signs of interaction during their optically thick phase, appearing as Type IIn or Type Ibn SNe (Kiewe et al. 2012;Taddia et al. 2013;Smith 2017).Others, like ASAS-SN 14jb, might not have developed those signs until later on, particularly if the progenitor drove some sort of fast outflow clearing a low-density cavity surrounded by denser and slower material (Dwarkadas 2005(Dwarkadas , 2007;;Patnaude et al. 2015Patnaude et al. , 2017)).There is of course no way to tell how old these SNRs might be, but given the properties of Cas A, SNR NGC 4449, and SN2005 ip discussed in Section 3.2, ages ranging between a few years and a few centuries seem reasonable.
Eight of our young SNR candidates have FWHM values that are within 25% of our lower threshold of 400 km s −1 : 2MASX J23331223-6034201, LEDA 1015413, Arp 142, NGC 2276B, NGC 5735, NGC 6946, and UGC 09182.While these candidates might be considered somewhat more marginal than the others, it is important to emphasize that in each case the fit to the Hα spectral window does require the presence of a broad component with a high level of significance that shows spatial clustering in the 2D maps.The FWHM values in the other eleven candidates range between 953 ± 358 km s −1 for NGC 4806/SN2011fh and 3423 ± 713 km s −1 for NGC 5908, comparable to the values measured in X-ray bright SNRs decades to centuries after the SN explosion (Vink 2012).We note that the errors produced by MPFIT for the FWHM and systemic velocities of the broad Hα component are likely underestimated in the SNR candidates with the largest FWHM values.A Bayesian analysis might reveal substantial correlations in the posterior distributions for these parameters, but that is outside the scope of the present work.With one exception, the luminosities we measure for the broad Hα component in our SNR candidates range between 3 × 10 35 and 3 × 10 37 ergs/s.The outlier, with a luminosity of 9 × 10 32 ergs/s, is the SNR candidate in NGC 6946, which is also by far the closest galaxy in our sample (7.9 ± 4.0 Mpc, Eldridge & Xiao 2019).For comparison, the Hα luminosities of the 143 SNRs in five nearby galaxies compiled by de Grijs et al. (2000) range between 10 36 and 10 38 ergs/s (see Figure 9).Although there is considerable overlap in these luminosity ranges, it is important to keep in mind that all our SNR candidates show considerably broad emission, while most local SNRs (including those in the sample from de Grijs et al. 2000) do not.
In Figure 10 we show the distribution of the metallicities, star formation rates and average stellar ages of the candidate SNRs in our sample, derived from the IFS spectra at their locations, compared to larger samples of CC and Type Ia SNe from PISCO (190 CC SNe and 234 SN Ia, respectively, Galbany et al. 2018b).All the parameters for our SNR candidates have been measured following the procedures described in Galbany et al. (2018b).The metallicities at the location of the candidate SNRs are somewhat lower than those found in the environments of the PISCO CC SNe, and the star formation rates are intermediate between the CC and Ia SNe in the PISCO samples, but these differences are small and hard to interpret for a sample as small as ours.The most striking systematic difference between our candidate SNRs and the bulk population of PISCO SNe is in the average stellar ages, which are clearly lower by about half a dex than those found in the environments of CC SNe, and about a dex lower than those of SN Ia.This indicates that the progenitors of our candidate SNRs might be shorter lived, and hence more massive, than those of a typical CC SNe.A similar trend has been found for the environments of Type IIn SNe by Moriya et al. (2023).

DISCUSSION AND CONCLUSIONS
We have conducted a systematic search for regions with broad (≥ 400 km s −1 ) Hα emission in IFS data cubes of 1229 nearby galaxies imaged by the PMAS and MUSE instruments.We have identified 19 such regions, which we classify as SNR candidates by analogy with the properties of known objects like Cas A and SNR NGC 4449-1.Indeed, four of the regions we have found coincide with the sites of previously known CC SNe, one Type IIP and three Type IIn, including the well known interacting object SN 2005ip.These coincidences, and the physical properties of the SNR candidates we have identified, suggest that the broad Hα emission in these regions is produced by a strong interaction between SN ejecta and some sort of dense surrounding medium.This medium could be material lost by the SN progenitor before the explosion due to stellar winds or binary interactions, which seems to be a common feature in Type IIn SNe, or a dense component of the interstellar medium associated with the formation site of the SN progenitors.The stellar ages measured from the IFS data at

Figure 1 .
Figure 1.Optical spectra of Cas A and SNR NGC4449 (both from Milisavljevic et al. 2012), and SN 2005ip (from our PMAS IFS data) around the [O III] λ5007 line (left), and the Hα line (right).

Figure 2 .
Figure 2. Left: Line fits to the residual spectrum of NGC 2906 at the location of SN 2005ip in PMAS (top) and MUSE (bottom).Blue, red and orange lines show the narrow [N ii] doublet and Hα, while turquoise shows the broad Hα component corresponding to SN 2005ip.This broad component is blueshifted with respect to the narrow Hα emission from the host galaxy.The full width at half maximum (FWHM) of the Hα emission from SN 2005ip is shown as the horizontal dashed line.The overall fit and emission spectrum are shown with purple and a black, dashed lines.Right: The same residual spectra shown in the left panels, fitted without including the broad Hα component.Notice the inability of the fitted narrow line emission (purple) to reproduce the observed residual spectrum (dashed black).

Figure 3 .
Figure 3. Temporal evolution of the FWHM for the two components of the Hα emission from SN 2005ip, calculated from the spectra analyzed in Smith et al. (2009) and Smith et al. (2017) (red circles for the broad component, blue squares for the intermediate component), together with the IFS observations from PMAS and MUSE shown in Figure2(yellow pentagon and green cross).Three lines are included for illustrative purposes: a constant velocity set to the FWHM of the broad component in the first spectrum (t = 17 days), a Sedov solution (v ∝ t −3/5 ) anchored to the FWHM of the MUSE spectrum, and a v ∝ t −1/7 powerlaw, also anchored to the FWHM of the MUSE spectrum.

Figure 4 .
Figure 4. Hα flux maps for NGC 2906.The Hα contribution from SN 2005ip (right) can be separated from that of its host galaxy (left).Top: PMAS.Bottom: MUSE.

Figure 5 .
Figure 5. Diagnostic plot showing the MUSE data for the entire galaxy NGC2906.F , ∆F , σ, and ∆σ represent the fluxes and widths of the broad components in the Hα line with their uncertainties.Spaxels where the broad Hα component was detected are shown in turquoise, while the rest of the spaxels in the galaxy are shown in orange.The black contours indicate the 1-, 2and 3-σ confidence contours of the distribution of each quantity.

Figure 6 .
Figure 6.Left column: spectra of young SNRs identified in our IFS sample in the Hα region, with the broad Hα component highlighted in magenta.Middle column: spatial distribution of the broad Hα component in the entire footprint of the IFS data.Right column: image of the IFS data footprint in white light.

Table 1 .
Full Width Hal Maximum of the two components, broad and intermediate, of SN 2005ip measured in spectra obtained from the literature plus our PMAS and MUSE measurement.
gether with our IFS observations.The broadest of the two components is only seen clearly in spectra taken before April 2006 (MJD 53852, or 172 days after discovery, see Table1), but the intermediate component is always

Table 2 .
Coordinates, broad Hα fluxes and luminosities of the 20 SNR candidate regions with broad Hα emission.