SN 2017gci: a nearby Type I Superluminous Supernova with a bumpy tail

MNRAS 000,1–17 (2015) Preprint 24 December 2020 Compiled using MNRAS L A TEX style file v3.0 SN 2017gci: a nearby Type I Superluminous Supernova with a bumpy tail A. Fiore,1,2★T.-W. Chen,3,4A. Jerkstrand,3S. Benetti,1R. Ciolfi,1,5C. Inserra,6 E. Cappellaro,1A. Pastorello,1G. Leloudas,7S. Schulze,8M. Berton,9,10 J. Burke,12,13 C. McCully,13 W. Fong,14 L. Galbany,15 M. Gromadzki,16 C. P. Gutiérrez,11 D. Hiramatsu,12,13 G. Hosseinzadeh,17 D. A. Howell,12,13 E. Kankare,18 R. Lunnan,3T. E. Müller-Bravo,11 D. O’ Neill,19 M. Nicholl,20,21 A. Rau,4J. Sollerman,3G. Terreran,14 S. Valenti,22 D. R. Young19 1INAF - Osservatorio Astronomico di Padova, Vicolo dell’Osservatorio 5, I-35122 Padova, Italy 2Dipartimento di Fisica e Astronomia ‘G. Galilei’, Università di Padova, Vicolo dell’Osservatorio 3, I-35122 Padova, Italy 3Department of Astronomy, Oskar Klein Centre, Stockholm University, Albanova, 10691 Stockholm, Sweden 4Max-Planck-Institut für Extraterrestrische Physik, Giessenbachstraße 1, 85748, Garching, Germany 5INFN, Sezione di Padova, Via Francesco Marzolo 8, I-35131 Padova, Italy 6School of Physics & Astronomy, Cardiff University, Cardiff, UK 7DTU Space, National Space Institute, Technical University of Denmark, Elektrovej 327, 2800 Kgs. Lyngby, Denmark 8Department of Particle Physics and Astrophysics, Weizmann Institute of Science, Rehovot 7610001, Israel 9Finnish Centre for Astronomy with ESO (FINCA), University of Turku, Vesilinnantie 5, FI-20014 University of Turku, Finland 10Aalto University Metsähovi Radio Observatory, Metsähovintie 114, FI-02540 Kylmälä, Finland 11Department of Physics and Astronomy, University of Southampton, Southampton, Hampshire, SO17 1BJ, UK 12Department of Physics, University of California, Santa Barbara, CA 93106-9530, USA 13Las Cumbres Observatory, 6740 Cortona Dr, Suite 102, Goleta, CA 93117-5575, USA 14Center for Interdisciplinary Exploration and Research in Astrophysics (CIERA) and Department of Physics and Astronomy, Northwestern University, Evanston, IL 60208 15Departamento de Física Teórica y del Cosmos, Universidad de Granada, E-18071 Granada, Spain 16Astronomical Observatory, University of Warsaw, Al. Ujazdowskie 4, 00-478 Warszawa, Poland 17Center for Astrophysics | Harvard & Smithsonian, 60 Garden Street, Cambridge, MA 02138-1516, USA 18Department of Physics and Astronomy, University of Turku, Vesilinnantie 5, FI-20014 Turku, Finland 19Astrophysics Research Centre, School of Mathematics and Physics, Queens University Belfast, Belfast BT7 1NN, UK 20Birmingham Institute for Gravitational Wave Astronomy and School of Physics and Astronomy, University of Birmingham, Birmingham B15 2TT, UK 21Institute for Astronomy, University of Edinburgh, Royal Observatory, Blackford Hill, EH9 3HJ, UK 22Department of Physics and Astronomy, University of California, 1 Shields Avenue, Davis, CA 95616-5270, USA Accepted XXX. Received YYY; in original form ZZZ ABSTRACT We present and discuss the optical spectro-photometric observations of the nearby (𝑧=0.087) Type I superluminous supernova (SLSN I) SN 2017gci, whose peak K-corrected absolute magnitude reaches 𝑀𝑔=−21.5mag. Its photometric and spectroscopic evolution includes features of both slowand of fast-evolving SLSN I, thus favoring a continuum distribution between the two SLSN-I subclasses. In particular, similarly to other SLSNe I, the multi-band light curves of SN 2017gci show two re-brightenings at about 103 and 142 days after the maximum light. Interestingly, this broadly agrees with a broad emission feature emerging around 6520 Å after ∼51 days from the maximum light, which is followed by a sharp knee in the light curve. If we interpret this feature as H𝛼, this could support the fact that the bumps are the signature of late interactions of the ejecta with a (hydrogen-rich) circumstellar material. Then we fitted magnetarand CSM-interactionpowered synthetic light curves onto the bolometric one of SN 2017gci. In the magnetar case, the fit suggests a polar magnetic field 𝐵p≃6×1014 G, an initial period of the magnetar 𝑃initial ≃2.8ms, an ejecta mass 𝑀ejecta ≃9Mand an ejecta opacity 𝜅≃0.08cm2g−1. A CSM-interaction scenario would imply a CSM mass ≃5Mand an ejecta mass ≃12M. Finally, the nebular spectrum of phase +187 days was modeled, deriving a mass of ∼10 𝑀for the ejecta. Our models suggest that either a magnetar or CSM interaction might be the power sources for SN 2017gci and that its progenitor was a massive (40 𝑀) star. Key words: supernovae: general – supernovae individual: SN 2017gci ★E-mail: [email protected] ©2015 The Authors arXiv:2012.12755v1 [astro-ph.HE] 23 Dec 2020 2A. Fiore et al. 1 INTRODUCTION Superluminous supernovae (SLSNe) were initially defined as those supernovae (SNe) whose peak-absolute magnitude is brighter than -21 mag (Gal-Yam 2012). They are intrinsically rare objects often discovered in metal-poor dwarf host galaxies (Chen et al. 2013;Lunnan et al. 2014;Leloudas et al. 2015;Perley et al. 2016;Chen et al. 2017a;Schulze et al. 2018). The origin of such peculiar transients represents a major challenge for contemporary astrophysics since it raises some fundamental questions about the ultimate stages of the evolution of massive stars. From an observational point of view, SLSNe can be broadly classified according to their hydrogen abundance. SLSNe I are H poor, although some of them display a late (&100 days) occurrence of H𝛼(a fraction estimated to be ∼15%, Yan et al. 2017), while Type II SLSNe display Balmer lines in their optical spectra. Recently it has been proposed that 𝑀𝑔=−19.8mag can be used as a luminosity threshold for the SLSNe I subclass only (Gal-Yam 2018b). However, this does not seem to correspond to a sharp edge in the luminosity function of H-poor SNe (De Cia et al. 2018;Gal-Yam 2018b;Quimby et al. 2018) and the SLSN I classification is generally inferred with a spectrum taken at about the maximum luminosity. This is characterized by a hot blue continuum (with a blackbody temperature 𝑇BB ≃10000 −15000 K) with O II absorptions between 3000 −5000 Å. Determining which physical mechanisms drive the explosion of a SLSN is not obvious. Therefore the discovery of nearby SLSNe (with 𝑧.0.1) is of particular interest since it may allow for higher resolution spectra, possibly in a wider wavelength range. A handful of viable scenarios have been invoked to explain the luminosity of SLSNe, as e. g. the onset of the pair-instability mechanism (e g. Yoshida et al. 2016) in very massive stars (heavier than ∼130 M, Rakavy & Shaviv 1967;Gal-Yam et al. 2009). In such a scenario, the central pressure drop caused by the 𝑒+, 𝑒−pair creation promptly triggers the collapse of the star and the thermonuclear explosion of its core with an overwhelming production of nickel. Nonetheless, the amount of 56Ni mass required for an absolute peak-magnitude brighter than ∼ −21 mag could make the rise time of the light curves (LCs) too slow (Nicholl et al. 2013) compared to the observations. Moreover, the spectra of the slow SLSNe I cannot be fitted by pairinstability models (Dessart et al. 2013;Jerkstrand, Smartt & Heger 2016). Another possibility lies in the interaction of the SN ejecta with a circumstellar material (CSM, e. g. Chevalier & Fransson 2003; Chevalier & Irwin 2011;Ginzubrg & Balberg 2012;Chatzopoulos, Wheeler & Vinko 2012;Chatzopoulos et al. 2013;Nicholl et al. 2014;Chen et al. 2015) which was lost by the progenitor star, e g., via stellar winds or during a pulsational pair-instability phase. If so, the SN ejecta crashes into surrounding shells or clumps of dense matter and drives a shock at the collision edge. This can convert the kinetic energy of the SN ejecta to radiation. However, there are generally no ‘standard’ spectroscopic signatures (i. e. narrow emission lines, as in the case of Type IIn (SL)SNe, e g. SN 2006gy, Smith et al. 2007) of CSM interaction in the spectra of SLSNe I (Lunnan et al. 2019). On the other hand, the presence of the intermediate-width Mg II resonance doublet around ∼2800 Å (Lunnan et al. 2018), the late broad H𝛼emission (Yan et al. 2015) and the LC oscillations (bumps) of some SLSNe I (Nicholl et al. 2015;Yan et al. 2017) strongly support that the interaction with CSM must be taken into account. Finally, a model which has growing consensus within the astrophysical community considers that the luminosity of SLSNe I is sustained by the spin-down radiation of a nascent magnetar (e. g. Kasen & Bildsten 2010;Woosley 2010;Suzuki & Maeda 2017,2019). According to this scenario, a highly-magnetized, newly-born neutron star is the compact remnant left by the SLSN explosion. Similarly to the case of a pulsar-wind nebula (e g. Metzger et al. 2014), the energy radiated by the neutron star via magnetic-dipole braking inflates a low-density, radiation-dominated photon-pair plasma nebula that afterwards thermalizes into the expanding ejecta, thus acting as a (possibly dominant) power source to explain the luminosity of SLSNe I. The magnetar scenario is favoured also by the association of the superluminous SN 2011kl (Greiner et al. 2015) with an ultra long gamma ray burst. Initially, it was proposed that SLSNe I might share the environment with fast radio bursts (FRBs) (Nicholl et al. 2017;Margalit et al. 2018) but the recent discovery of two FRBs with a massive host galaxy (Ravi et al. 2019;Marcote et al. 2020) disfavours this association. SLSNe I are actually a heterogeneous class of transients. In fact, it is possible to distinguish between at least two subclasses, depending on whether their LCs evolve in a slow or a fast fashion. Slow-evolving SLSNe I have a rise time towards the maximum luminosity which exceeds 50 days, whereas the fast-evolving SLSNe I reach the maximum light in less than 30 days. Although a continuum distribution likely fills the gap between the two subclasses (Nicholl et al. 2015; De Cia et al. 2018), the distinction between fastand slow-evolving SLSNe I is still used (e. g. Kumar et al. 2020) and helpful to distinguish different rise or decline timescales within the SLSN I class. In addition, slow-evolving SLSNe I more often show bumps in their LC both before and after the maximum-luminosity epoch (Inserra et al. 2017;Inserra 2019). SN 2017gci is located at RA =06h46m45.02sand Dec = −27°14055.800 (J2000). It was discovered by Gaia on the August 16th, 2017 (Delgado et al. 2017) as an apparently hostless, blue transient and named Gaia17cbp. Initially, it was classified as a Cataclysmic Variable-candidate. Later it was reclassified as SLSN I (Lyman et al. 2017) by the extended Public ESO Spectroscopic Survey for Transient Objects (ePESSTO, Smartt et al. 2015). The last 𝑔0, 𝑟0, 𝑖0, 𝑧0, 𝐽, 𝐻, 𝐾s-band imaging frames (taken on September 29th, 2019) show that the host-galaxy flux contribution of SN 2017gci is not completely negligible at optical/NIR wavelengths (𝑔host ≃22.8mag, 𝑟host ≃22.2mag, 𝑖host ≃22 mag, 𝐽host ≃21.6 mag, 𝐻host ≃21.5mag, see Section 2). We hereby present the LCs and the spectra of the SLSN I SN 2017gci. The observations will be made public via WiseRep1. In addition, we provide an interpretation of the data both with a semianalytic magnetar-powered modelling and by means of the singlezone SUMO models (Jerkstrand et al. 2017) for the nebular spectra of SLSNe I. Hereafter, in Section 2 we describe and discuss the photometric observations; Section 3 deals with the spectra of SN 2017gci; in Section 4 we compare the spectra and the LCs of SN 2017gci with those ones of other SLSNe I and we provide our interpretation of this event within the magnetar scenario; finally we summarize our conclusion in Section 5. Throughout the paper we assume a flat Universe with Ωm=0.31 and 𝐻0=71 ±3 km s−1Mpc−1. Given such cosmological parameters and a redshift 𝑧=0.0873 ±0.0003 (see Section 3.2), we found a luminosity distance for SN 2017gci of 𝑑L=392.5+23.5 −15.9Mpc, corresponding to a distance modulus 𝜇=37.96 ±0.1 mag . Moreover we assume no extinction from the host galaxy since no narrow absorption interstellar line of the Na ID doublet (Poznanski et al. 2012) is seen in the optical spectra. 1https://wiserep.weizmann.ac.il/search/ . MNRAS 000,1–17 (2015) The bumpy, superluminous SN 2017gci 3 0 100 200 300 400 Rest-frame phase from maximum φ[days] 10 15 20 25 30 Apparent Magnitude GROND LCO Gaia EFOSC2 UVOT SOFI UVW2-10.4 UVM2-9.3 UVW1-7.5 U-6 B-4.4 g-3 V-1.2 r i+1.9 z+3.8 J+6.8 H+10.6 K+13.2 Figure 1. S-corrected LCs of SN 2017gci in 𝑈𝑉 𝑊 2, 𝑈𝑉 𝑀2, 𝑈𝑉 𝑊 1, 𝑈 , 𝐵, 𝑔, 𝑉 , 𝑟 , 𝑖, 𝑧, 𝐽 , 𝐻 , 𝐾sbands, respectively plotted in black, brown, cyan, dark green, dark blue, green, purple, red, blue, magenta, orange, silver and yellow. Magnitudes obtained with different instruments were plotted with different symbols, as labelled in grey in the upper-right corner. The green dotted line represents a 4th-order polynomial fit of the early LC to estimate the maximum epoch in 𝑔 band. Dotted lines represent the linear fit to the data with rest-frame phases later then 74 days, while solid lines represent the linear fits once the bumps have been excluded as explained in the text. The latter points are plotted as empty dots. Arrows correspond to 2.5𝜎detection limits. Magnitudes are in AB system. 0 50 100 150 200 250 300 350 400 Rest-frame phase from maximum φ[days] 41.5 42.0 42.5 43.0 43.5 44.0 log10(Lbol/1 erg s−1) 56Co-decay slope Figure 2. Pseudo-bolometric LC of SN 2017gci (computed after having applied the S-corrections and the K-corrections to multi-band photometry, see text). Red dots: pseudo-bolometric LC obtained integrating the SED with the trapezoidal rule. Luminosities are in logarithmic scale and arrows correspond to 2.5𝜎 limits. The light-blue shaded areas refer to the epochs during which the bumps occur. 2 PHOTOMETRY 2.1 Observations and preliminary reduction We performed most of the photometric follow-up with the MPG 2.2m telescope+GROND (Gamma-Ray Burst Optical and Near-Infrared Detector, Greiner et al. 2008) as a part of GREAT survey (Chen et al. 2018) and with NTT+EFOSC2 (Buzzoni et al. 1984). Pre maximumand maximum-epoch data are scarce, but some epochs near the peak were obtained thanks to the photometry of the Las Cumbres Observatory2(LCO) Global Telescope network. These observations were 2https://lco.global/ . obtainedwiththecameraSinistro(Brownetal.2011)builtfor the 1mclass LCO telescopes. The set of photometric data we have collected consists of 𝑔0, 𝑟0, 𝑖0, 𝑧0, 𝐽, 𝐻, 𝐾s-band images taken at ESO La Silla Observatory with 2.2m+GROND, 𝐵, 𝑉, 𝑔, 𝑟, 𝑖, 𝑧-filter images taken at LCO, 𝑈𝑉𝑊2, 𝑈𝑉 𝑀2, 𝑈𝑉𝑊1, 𝑈, 𝐵, 𝑉-filter images obtained with the Swift Ultraviolet/Optical Telescope (UVOT) and 𝐽, 𝐻, 𝐾s-filter frames obtained with NTT+SOFI (Son OF Isaac, Moorwood et al. 1998). To pre-reduce the EFOSC2 frames, we applied standard overscan, bias and flatfielding procedures within IRAF. The SOFI frames were pre-reduced with the PESSTO pipeline (Smartt et al. 2015). The GROND images were pre-reduced by the GROND pipeline (Krühler et al. 2008), which applies de-bias and flat-field corrections, stacks images and provides astrometry calibration. MNRAS 000,1–17 (2015) 4A. Fiore et al. 2.2 Data reduction We corrected the 𝑖and 𝑧EFOSC2 frames for the fringing pattern by means of fringing masks. These were created by downloading and reducing ∼100 archival 𝑖and 𝑧images from the ESO Archive Science Facility3for each filter at random coordinates, and selecting those with an exposure time &100 s. We took the median of all of them in order to get rid of the field stars present in the frames. After subtracting the median value from each averaged image we obtained the master fringing mask to be subtracted to the frames. 𝐵, 𝑔, 𝑉, 𝑟, 𝑖, 𝑧, 𝐽, 𝐻, 𝐾s-filter magnitudes were measured using the SNOoPY package (Cappellaro 2014) with the Point Spread Function (PSF)-fitting technique, via the DAOPHOT tool (Stetson 1987). Within this method, a reference PSF is obtained by averaging those ones of isolated field stars and then fitted onto the SN to obtain the instrumental magnitude. Meanwhile, the background underneath the SN can be estimated interpolating a low order polynomial to the surrounding regions. In alternative, we removed the host galaxy contribution with the template-subtraction technique, which was also performed within SNOoPY, and via the hotpants package (Becker 2015). The template-subtraction method envisages the subtraction of the scientific frames with a template image of the same field taken with the same filter when the SN is absent. After the template subtraction, the magnitudes are always derived with the PSF method in the residual frame. We found that the template subtraction method gives indeed more reliable photometric measurements, especially when the SN flux becomes fainter. For SN 2017gci this happens at 𝜙∼100 days after maximum. In Tables A1,A2,A4,A5 if not differently stated, the reported magnitudes have been derived after template subtraction. The 𝑔, 𝑟, 𝑖, 𝑧-template frames were downloaded from the Image Cutout Server4of the second Data Release of Pan-STARRS as stack images. Deep 𝐵, 𝑉-template frames were requested to LCO which observed the field of SN 2017gci on 2019 October, 4th (corresponding to 708 rest-frame days after maximum). For the 𝐽, 𝐻, 𝐾stemplate frames we used the combination of the last GROND 𝐽, 𝐻, 𝐾s-band frames taken on 2019 September, 25th (700 rest-frame days after the maximum) and 29th (703 rest-frame days after the maximum), assuming that at these very late epochs SN 2017gci faded well below the detection limit. Since the host galaxy is not visible in the deep frame taken about 2 year after explosion, we estimated an upper limit for the 𝐾smagnitude (𝐾host,uplim ≃18.6mag) of the host galaxy using the PSF technique. Hence we decided to use 𝐵, 𝑔, 𝑉, 𝑟, 𝑖, 𝑧, 𝐽, 𝐻 template-subtracted magnitudes and 𝐾PSF magnitudes. 𝐵, 𝑔, 𝑉, 𝑟, 𝑖, 𝑧 magnitudes were calibrated on the field stars identified with the Pan-STARRS (Panoramic Survey Telescope and Rapid Response System, Chambers et al. 2016) catalogue. The calibration was performed after having applied the color correction (see equation 6 in Tonry et al. 2012) between Pan-STARRS and SDSS filters. For the 𝐵, 𝑉 images the calibration was done after having converted the Pan-STARRS magnitudes to Sloan as before, and then the Sloan magnitudes to Johnsons system following Chonis & Gaskell (2008). The NIR magnitudes were instead calibrated with a local sequence of stars from the Two Micron All Sky Survey (2MASS, Skrutskie et al. 2006). To measure 𝑈𝑉𝑊2, 𝑈𝑉 𝑀2, 𝑈𝑉𝑊1, 𝑈, 𝐵, 𝑉 Swift/UVOT magnitudes we stacked the layers of the individual observing segments with the task uvotimsum and measured the brightness using 500-radius aperture with the task uvotsource in HEASoft version 3http://archive.eso.org/ . 4https://ps1images.stsci.edu/cgi-bin/ps1cutouts/ . 6.25 (Nasa High Energy Astrophysics Science Archive Research Center (Heasarc) 2014). Since we have used several instruments to collect the photometry of SN 2017gci, each one defining its own photometric system, it is necessary to convert all of them into a standard one. The procedure involved is sometimes called S-correction (Stritzinger et al. 2002) and we applied it following the method described in EliasRosa et al. (2006) and Pignata et al. (2004). Therefore we computed synthetic photometry using the observed-frame spectra by means of the library pysynphot5both for the standard photometric systems (𝑚𝑠,standard) and for the instrumental filters (𝑚𝑠,instr)6. For each instrument and each bandpass filter, the S-correction 𝑆corr was then computed as 𝑆corr =𝑚𝑠,standard −𝑚𝑠,instr. We linearly interpolated over the spectroscopic epochs the 𝑆corr grid to match the photometric epochs and then we applied the corresponding correction. We estimated a mean statistical uncertainty for this correction by looking at the dispersion around the interpolation and we assumed it to be 0.02 mag. This uncertainty was eventually summed in quadrature with the photometric one to have the final error. However, the above procedure can be performed only when the passband filters are entirely covered by the wavelength range of the spectra. If this is not the case (𝑈, 𝑧, 𝐽, 𝐻, 𝐾s), we computed the S-correction as before but using blackbody spectral energy distribution reported to the observer frame. We considered two temperature ranges: 𝑇=12000 −8000 K up to 40 days and 𝑇=8000 −4000 K at later phases, broadly corresponding to the blackbody temperatures derived from the SED blackbody fit (see Section 4.3.3). The maximum of the S-correction computed in the adopted range is taken as a proxy of the Scorrection error introduced by the non standard system, which we called Δ𝑆corr. The Δ𝑆corr values were propagated in our analysis. The reduced𝑈𝑉𝑊2, 𝑈𝑉 𝑀2, 𝑈𝑉𝑊1, 𝑈, 𝐵, 𝑔, 𝑉, 𝑟, 𝑖, 𝑧, 𝐽, 𝐻, 𝐾smagnitudes are reported in Tab. A1,A2,A4,A5. The S-corrections 𝑆corr and the Δ𝑆corr values are listed in Tab. A6,A7,A8,A9. The latter were dividedforsimplicityintwo temperaturebins(4000 K < 𝑇 < 8000 K and 8000K < 𝑇 < 12000K). The S-corrected LCs of SN 2017gci are shown in Fig. 1. Magnitudes are in AB system and the phases are corrected for time dilation (in the following, we will refer to the rest-frame phase with respect to maximum luminosity as 𝜙). From our photometric data, it is not possible to provide a robust estimate for the maximum luminosity and the corresponding epoch due to a lack of early time coverage. To obtain an upper limit on the rise time, we added a non-detection from the Gaia-archival data (Gaia collaboration 2016a,b;Salgado et al. 2017), whose epoch is June, 27th 2017 (MJD=57931), which was converted to 𝑔magnitude7. Then a 4th-order polynomial was fit over the early 𝑔-filter magnitudes allowing us to estimate the epoch and magnitude of maximum luminosity: MJDmax =57990.3+8 −15 for 𝑔max =17.1±0.3mag. 5https://pysynphot.readthedocs.io/ . 6Instrumental transmission functions for the different instruments were retrieved from http://svo2.cab.inta-csic.es/theory/fps3/. 7Useful relationships to convert Gaia magnitudes to those of the standard photometric systems are available in Section 5.3.5 of the Documentation Release (v. 1.2) of the Gaia Data Release 1. This is accessible from the following URL: https://gea.esac.esa.int/archive/documentation/GDR1/ . MNRAS 000,1–17 (2015) The bumpy, superluminous SN 2017gci 5 58000 58050 58100 58150 58200 −0.05 0.00 0.05 S-correction for Sinistro g r i B V 58000 58050 58100 58150 58200 −0.1 0.0 S-correction for GROND 58000 58050 58100 58150 58200 MJD observation 0.00 0.05 S-correction for Swift/UVOT Figure 3. S-correction for LCO+Sinistro (top panel), GROND+2.2m (middle panel) and Swift/UVOT (lower panel). Filled dots are coded as in the label (top right corner). 2.2.1 K-correction For the optical and 𝐽, 𝐻 magnitudes, we obtained the K-corrections from the spectra at our disposal (see Section 3). For each of them and for each band-pass filter, we derived a synthetic magnitude via pysynphot. This was done both for the rest-frame spectrum (for which we computed a synthetic magnitude 𝑚𝑠,rest) and for the observed one (for which we computed a synthetic magnitude 𝑚𝑠,obs) via pysynphot distributed via AstroConda8. For each epoch, the K-correction 𝐾was computed as 𝐾=𝑚𝑠,obs −𝑚𝑠,rest. The resulting K-corrections are listed in Tab. A10. Finally, to adjust the sparser time sampling of the spectral epochs to the denser one of the magnitudes we linearly interpolated this table. Similarly, the K-corrections for the 𝑈𝑉𝑊2, 𝑈𝑉 𝑀2, 𝑈𝑉𝑊1, 𝐾s-filter magnitudes were estimated by using the SED (retrieved by photometry) in place of the observed spectra. 2.3 Main characteristics of the LCs and of the bolometric curve The 𝑔, 𝑉, 𝑟, 𝑖-filters LCs remain nearly constant for the first ∼20 days, while the 𝐵LC starts to decline earlier (after about ∼12 days from the maximum light). The 𝑈−and 𝑈𝑉𝑊2, 𝑈𝑉 𝑀2, 𝑈𝑉𝑊1-filters LCs possibly peak a few days before, but the early-time data coverage is inadequate and so we cannot securely constrain the maximumluminosity epoch for those filters. The overall evolution is slower in the 𝑧, 𝐽, 𝐻, 𝐾smagnitudes, and the early flat phase around the maximum luminosity lasts about 30 −40 days. Then for 𝜙&17 days, the evolution steepens and at ≃54 −57 days the observed LCs present an abrupt change of their slopes. Such a ‘knee’ seems much sharper than the transition region which usually preludes to the so-called ‘magnetar tail’ (see e g. Inserra et al. 2013;De Cia et 8https://astroconda.readthedocs.io/en/latest/ . al. 2018). Thereafter, for 𝜙&71 days, the LCs settle on a steady, almost linear decline. During this phase, the LCs display two sharp re-brightenings at 𝜙∼103 and 142 days. Finally, after 𝜙 > 213 days, the SN is no longer detectable. After this epoch, we took 4 frames in the 𝑔, 𝑟, 𝑖, 𝑧 bands, 3 in 𝐽-, 2 in 𝐻and 1 in 𝐾sGROND bands until September 29th, 2019 (see Introduction). However, the template subtracted images provided only 2.5𝜎detection limits. Apparent magnitudes were converted to absolute magnitudes once the redshift and the Galactic absorption are known. Given a Galactic extinction 𝐴𝑉 G=0.360mag (Schlafly & Finkbeiner 2011), the distance modulus 𝜇and the 𝑔apparent magnitude for the maximum given in Section 1, the K-corrected absolute peak-magnitude is 𝑀𝑔=−21.5±0.3 mag in 𝑔band. We also built the pseudobolometric LC of SN 2017gci. This was computed by integrating its K-corrected 𝑈𝑉𝑊2, 𝑈𝑊 𝑀2, 𝑈𝑉𝑊1, 𝑈, 𝐵, 𝑔, 𝑉, 𝑟, 𝑖, 𝑧, 𝐽, 𝐻, 𝐾sphotometry. We adopted as reference the epochs of the 𝑟band photometry, and missing measurements at given epochs for the other filters were obtained through interpolation or, if necessary, by extrapolation assuming a constant colour from the closest available epoch. The fluxes at the filter effective wavelengths, corrected for the Galactic extinction, provide the Spectral Energy Distribution (SED) at each epoch. Then, we integrate the SED with the trapezoidal rule, assuming zero flux at the integration boundaries. We measured the early and late luminosity-decay slopes onto the bolometric LC . The steeper early decline (between 𝜙=30−51 days from the maximum light) is estimated to be 0.040 mag/day, whereas the late one (between 𝜙=60 −210 days from the maximum light) is 0.018 mag/day. As mentioned above, a handful of SLSNe I (e g. SN 2015bn, Nicholl et al. 2016a) show a bumpy LC. To measure the post-maximum decay slope for 𝜙&73 days, we excluded the bumps from the linear fit. To do that, we proceeded in the following way. For a given wavelength band, we fitted a first-order polynomial using all the magnitudes for 𝜙 > 73 days, and for each of those epochs we subtracted the interpolated magnitude to the corresponding observed one. Then, we computed the standard deviation 𝜎73−103 of the fit residuals for 73 .𝜙.103 days (corresponding to the gray shaded area in Fig. 4), since at these epochs the magnitudes do not seem to be affected much by the bumps. At this point we repeated the linear fit, this time excluding all the magnitudes whose difference with the previous fit is brighter than 1×𝜎73−103 mag. Bluer LCs tend to decay faster than the redder ones (except for the 𝑖and 𝑧bands). Nonetheless, we specify that the slope estimates for the 𝐵, 𝑉, 𝐽, 𝐻, 𝐾sLCs are less accurate because of the evident data paucity. The late-decline slopes were not measured for the 𝑈𝑉𝑊2, 𝑈𝑉 𝑀2, 𝑈𝑊𝑉1, 𝑈-filter LCs since no coeval measure is available. The luminosity excesses in the different bands 𝛿𝑚 over the late post-peak decline rate h𝑚iare such that |𝛿𝑚|=|h𝑚i − 𝑚|.1mag (see Fig. 4). 3 SPECTROSCOPY 3.1 Observations and data reduction Optical spectra were acquired with the ESO New Technology Telescope (NTT)+EFOSC2 at La Silla Observatory, Chile, the ESO Very Large Telescope (VLT)+X-Shooter(XS) (see Vernet et al 2011, for a description) at Paranal Observatory, Chile, the Keck I telescope+LRIS (Oke et al 1995) at W. M. Keck Observatory, Maunakea, Hawaii and the Multiple Mirror Telescope (MMT)+Binospec (Fabricant et al. 2019) at Maunakea Observatory. The EFOSC2, Binospec and LRIS spectra were reduced with the standard IRAF tools. Also, five LCO+FLOYDS spectra were secured (see Tab. A12). MNRAS 000,1–17 (2015) 6A. Fiore et al. 75 100 125 150 175 200 225 250 Rest-frame phase from maximum φ[days] −1.0 −0.8 −0.6 −0.4 −0.2 0.0 0.2 0.4 0.6 0.8 δm [mag] g-filter r-filter i-filter z-filter B-filter V-filter J-filter H-filter Figure 4. Luminosity undulations 𝛿𝑚 in each bandpass filter. Datapoints are coloured as in Fig. 1. The dashed black line marks the 𝛿𝑚 =0mag level. The gray-shaded area represents the phase range within which we consider the LC decline not to be affected by the bump features. Therein we computed the standard deviation of the linear-fit residuals to disentangle the bumps (see the text) from the LC decline. −50 −25 0 25 50 75 100 125 Rest-frame phase from maximum φ[days] 4000 6000 8000 10000 12000 14000 Blackbody temperature TBB [K] SN 2017gci SN 2015bn SN 2011ke (Inserra et al. 2013) Figure 5. Temporal evolution of the black-body temperature of SN 2017gci (black squares), SN 2015bn (red squares) and SN 2011ke (cyan squares). The blackbody temperatures for both SN 2017gci and SN 2015bn were retrieved by a blackbody fit of the SED, while the temperatures of SN 2011ke are taken from Inserra et al. (2013). The two-dimensional raw spectroscopic frames were then corrected for overscan, divided by a normalized flat-field, corrected for cosmic rays (by means of the L. A. Cosmic algorithm, Van Dokkum 2001), extracted across the spatial direction after having interpolated the background below the SN with a low-order polynomial fit on the surrounding regions, calibrated in wavelength against HeAr arcs. Then the extracted one-dimensional spectra were calibrated in flux and corrected for telluric absorption thanks to a set of spectrophotometric standard stars. Finally, the flux calibration of these spectra was also checked against the magnitudes retrieved by coeval photometry. The first XS spectrum (𝜙=187 days) was reduced following the procedure described in Krühler et al. (2015), whereas the second one (𝜙=367 days) was reduced via the esoreflex ESO-pipeline (v 2.9.1, Freudling et al. 2013). 3.2 The spectra The spectral evolution of SN 2017gci is shown in Fig. 6. In order to increase the signal-to-noise ratio we included in Fig. 6the average between the spectra observed at 𝜙=20 and 𝜙=23 days which is marked with a phase of 𝜙=22 days in the figure. The identification of the spectral features was done following Howell (2017), Quimby MNRAS 000,1–17 (2015) The bumpy, superluminous SN 2017gci 7 et al. (2018). Until the maximum light, the spectra of SN 2017gci show a hot blue continuum whose black-body temperature reaches 𝑇BB ≃12000−14000 K in the spectra about the maximum luminosity. On the redder side of the spectra, the broad Na ID𝜆𝜆 5890,5896 doublet, the C II 𝜆𝜆 6580,7121 lines and the O I𝜆7774 are evident. Tentatively we also identified the Si II 𝜆6355 feature. At shorter wavelengths the doublet H&K of the Ca II and the W-shaped O II features at 𝜆𝜆 4115,4357,4650 are also present. To test their identification we compared the pre-maximum spectrum of SN 2017gci at 𝜙=−7d with two synthetic spectra computed with TARDIS (Temperature And Radiative Diffusion In Supernovae, Kerzendorf & Sim 2014), an open-source, Monte Carlo-based, radiative-transfer spectral synthesis code for SN spectra. The first TARDIS spectrum (see Fig. 8, orange dots) was calculated assuming a pure Carbon chemical abundance while in the second one (magenta dots) a pure Oxygen abundance was input. Both of them assume Local Thermodynamic Equilibrium conditions. As shown in Fig. 8, the features at 𝜆𝜆 4115,4651 are well matched by the pure-O spectrum, but we cannot exclude a line blending with C II spectral features. Moreover, we identified two absorptions as the contribution of Fe II + Fe III. The latter is a common feature among the slow SLSNe I (Inserra 2019). After ∼20 days, the continuum becomes significantly redder with much less prominent O II absorptions. In their place, the Fe II and Mg II features start to be visible. From 𝜙&33 days, the spectra of SN 2017gci resemble those of a Type Ic BL SN at maximum luminosity, as expected by a SLSN I (Pastorello et al. 2010). Then, up to 𝜙∼156 there are no significant changes in the the spectra, except for the continuum becoming even fainter and redder. Surprisingly, at 𝜙∼51 days a spectral feature consistent with H𝛼emerges in the spectra and remains visible until 𝜙∼133 days. The occurrence of such a feature precedes (∼3−6days) a LC knee ( see Section 2.3). For 𝜙≥155 days the spectra becomes ‘pseudonebular’ (Nicholl et al. 2019) where emission features start to prevail on the absorption but with a residual fainter continuum. At these epochs the SN ejecta were cool enough to favour the recombination of the electrons. This reduces the free-electrons density, hence the optical depth and allow us to investigate the deepest emitting regions of the SN explosion. Moreover, in such a low-density environment the semi-forbidden and forbidden atomic transitions start to dominate the spectra. In fact, the emissions of the semi-forbidden 𝜆4571 Mg I], the [O I] doublet at 𝜆𝜆 6300,6364 and 𝜆𝜆 7291,7323 [Ca II] are present, as well as the strong NIR Ca II 𝜆𝜆 8498,8542,8662 triplet. Finally, in the late spectra (at 𝜙 > 130 days) of SN 2017gci the narrow H𝛼and [O III]- emission lines from the host galaxy become gradually visible. Using these features we calculated the redshift of the host galaxy, which turns out to be 𝑧=0.0873 ± 0.0003 (where the uncertainty is derived from the dispersion of the measurements). Moreover, the spectrum at 𝜙=187 days presents two features between 9000-11000 Å (see the insert in Fig. 6) where the contribution of Mg II 𝜆0.92 𝜇mand 𝜆1.09 𝜇mHe Imight be involved. He Iis not frequently seen among SLSNe I, except in the case of PTF10hgi (Quimby et al. 2018) and possibly in the case of SN 2012il(Inserraetal.2013;Quimbyetal. 2018). Furtherinterpretation of the spectrum at 𝜙=187 days will be provided in Section 4.4. In the spectrum taken at 𝜙=367 days almost all the broad features present in the previous spectrum are no longer present except a residual contribution from [O I]. The NIR part of this spectrum was too faint to be extracted. 3.2.1 Photospheric velocity To estimate the photospheric velocity we measured the wavelengths corresponding to the minima of the P-Cygni profiles which occur in the spectra of SN 2017gci. They were determined with a gaussian fit of the absorption features (see Fig. 7) after having been normalized and continuum-subtracted. We performed these measurements from 𝜙=−7to 𝜙=−4days, when the O II absorption minima are present in the spectra. Errorbars are estimated by changing the continuum level multiple times before performing the fit. The Doppler shift measured with respect to the rest-frame wavelength of the emissions corresponds to a photospheric velocity 𝑣(O II).8000km s−1(see Fig. 7). 4 DISCUSSION In the following, we will discuss the interpretation of the data presented in the previous Sections. 4.1 Metallicity of the host galaxy We estimated the metallicity of the SN 2017gci site by means of the narrow emission lines of the spectra at 𝜙=187,367 days, attributed to the host-galaxy contribution. To test simultaneously several metallicity diagnostics, we used the python-based tool PYMCZ (Bianco et al. 2016). PYMCZ takes as input a list of flux measurements with an associated uncertainty for [O II]𝜆3727, H𝛽, [O III]𝜆4959, [O III] 𝜆5007, H𝛼, [N II]𝜆6584, [S II]𝜆6717. For each flux measurement, PYMCZ generates a set of synthetic data via a Monte Carlo simulation.Hencea Gaussianprobabilitydistribution isdrawn(whosemean is the input flux and whose standard deviation is the uncertainty of the flux) and randomly sampled. These flux measurements are used to compute the 12 +log10(O/H)index via the D02 (De Nicoló et al. 2002), PP04 N2Ha, PP04 O3N2 (Pettini & Pagel 2004), M08 N2Ha, M08 O3O2 (Maiolino et al. 2008) and M13 N2 (Marino et al. 2013) calibrators. The resulting 12+log10(O/H)estimates (see the boxplot in Fig. 9) cluster around ∼8.1(∼0.3𝑍), thus pointing towards a low-metallicity environment as it is expected by the host galaxies of SLSNe I (see Introduction). A comparison of the PP04 O3O2 metallicity measurements of SN 2017gci with other SLSNe I and GRBs at redshift 𝑧.0.1is reported in Tab. A14. The environment of SN 2017gci is among those with the lower metallicity. 4.2 [O I] emission profile In the spectrum of SN 2017gci at 𝜙=187 days, a close look to the profile of the [O I]𝜆𝜆 6300,6364 emission doublet points out the presence of a double peak on its topside (see Fig. 10, top panel). The bluest hump of the [O I] doublet peaks at 𝜆∼6260 Å (i. e. ∼40 Å blueshifted with respect to its rest-frame wavelength) while the reddest hump peaks between 𝜆∼6300 −6310 Å. The 40 −50 Å separation between the two peaks (which is lower than the natural 64 Å separation of the doublet) is similar to what was found by Milisavljevic et al. (2010) for the velocity shifts measured on the asymmetric [O I] profiles of a sample of stripped-envelope SNe. In fact, double or multi-peaked [O I] profiles were also observed in the late spectra of SNe Ib/c, as in the case of SN 2005bf (Anupama et al. 2005) or SN 2009jf (Valenti et al. 2011;Sahu et al. 2011) (see also Taubenberger et al. 2009, for further studies on the asymmetric [O I] profiles). The ∼40 Å wavelength shift of the blue peak corresponds to a MNRAS 000,1–17 (2015) 8A. Fiore et al. 4000 5000 6000 7000 8000 9000 10000 11000 12000 Rest wavelength λ[˚ A] −1 0 1 2 3 4 Fλ[erg cm−2s−1˚ A−1] + constant ×10−15 CII HK Ca II OII Fe II+ Fe III Mg I] Hβ Mg II Fe II [O III] Na ID Si II [O I] Hα CII CII [Ca II] O I Ca II −7d −5d −4d −3d 22d 32d 51d 73d 103d 131d 133d 155d 187d 367d 9000 10000 11000 12000 −7.0 −6.5 −6.0 −5.5 −5.0×10−16 Mg II? Mg II? He I? Figure 6. Spectral evolution of SN 2017gci. On the right side we plotted the rest-frame phase from the discovery for each spectrum. To provide a clear representation, we smoothed the spectra with a Savitzky-Golay filter. The smoothed spectra (black lines) have been overimposed to the original ones (light-blue lines), and all of them were scaled and offset. The black dashed lines mark the wavelengths at which the spectral features occur in the spectra. For each of them, the ion responsible of the transition is labelled nearby. In the last two spectra (at 𝜙=187,367 days), the narrow emission lines [O III], H𝛼, H𝛽were cut up to a certain flux threshold. The epochs of the observations, the scaling factors and the offsets are summarized in Tab. A12. The insert in the upper-right corner zooms the NIR part of the XS spectrum taken at 187 days. velocity blueshift of ∼2000 kms−1. To test whether the observed double-peaked [O I] profile could be reproduced by two velocity components, we fitted a composite model made of five gaussians (see Fig. 10, lower panel): two gaussians for the 2000 km s−1-blueshifted [O I]-doublet component (with a FWHM of ∼4000km s−1), two gaussians for the rest-frame [O I]-doublet component (with a FWHM of ∼4000km s−1) and a broader (FWHM ∼11000 km s−1) restframe component. The FWHM of the two doublets was kept constant in the fitting procedure. The peaks of two couples of gaussians have a fixed separation of 64 Å and a flux ratio 3:1 (see Fig. 10, lower panel). The broad component was added to better fit the broad wings of the emission feature. The best-fit curve (see Fig. 10, lower panel) underestimates the flux emitted in the blue hump of the doublet, but broadly accounts for the 40 −50 Å separation of the two peaks. The physical interpretation of the such profiles is not unique. It was suggested (Taubenberger et al. 2009;Valenti et al. 2011) that they may be the signature of a certain degree of ejecta asphericity. In fact, an asymmetric jet-like explosion (where the major part of the material is launched in the direction opposite to the observer) or ejecta blobs could be responsible of the two velocity components. In addition, the wings could be explained by a more sphericallysymmetric ejecta component. Finally, further clues on the ejecta geometry of SN 2017gci will be given by polarimetric observations. In fact, continuum polarization measurements of SN 2017gci display an evolution in the polarization degree, which grows for 𝜙 > 27 days. This may be an evidence of the SN photosphere departure from spherical symmetry at late phases (Cikota et al., in preparation). 4.3 Comparisons with other SLSNe I 4.3.1 Comparing the bolometric light curves We compared the bolometric LC of SN 2017gci with those of a sample of SLSNe I. Among these, the slow-SLSNe I subsample consists of SN 2015bn (Nicholl et al. 2016a,b), PTF12dam (Nicholl et al. 2013;Chen et al. 2015;Vreeswijk et al. 2017), SN 2007bi (Gal-Yam et al. 2009), PTF09cnd (Quimby et al. 2018), SN 2018bsz MNRAS 000,1–17 (2015) The bumpy, superluminous SN 2017gci 9 4000 5000 6000 7000 8000 Rest-frame wavlelength λ[˚ A] −12 −10 −8 −6 −4 −2 0 2 Normalized Fλ[erg cm−2s−1˚ A−1]−continuum φ=-7 days φ=-5 days φ=-4 days −8−7−6−5−4−3 Rest-frame phase from the maximum [days] 5000 6000 7000 8000 9000 OII λ4357 OII λ4650 Photospheric velocity [km s−1] Figure 7. Left panel: normalized and continuum-subtracted spectra between 𝜙=−7and 𝜙=−4days. The O II absorption minima were fitted with a gaussian (black solid lines). Right panel: Photospheric velocities retrieved by the absorption minima of the O II (𝜆4357,4650) features. 3500 3750 4000 4250 4500 4750 5000 5250 5500 Rest wavelength λ[˚ A] 0.2 0.4 0.6 0.8 1.0 1.2 Fλ[erg cm−2s−1˚ A−1] ×10−15 OII λλ4115 OII λλ4357 OII λλ4650 FeIII λλ4430 CII λλ4745 CII λλ3920 pure C pure O Figure 8. The spectral range between 3500 −5000 Å; the black solid line is the spectrum of SN 2017gci with rest-frame phase 𝜙=−7days. The dotted black lines mark the spectral lines which they are labelled with, respectively of Ca II, O II, C II and Fe III. The orange and the magenta dotted lines trace the synthetic profiles, respectively for a pure Carbon and Oxygen composition, computed with TARDIS. For the sake of completeness, we reported also H&K Ca II doublet, not present in the TARDIS spectrum. The O II features are pretty well matched by the magenta profile. (Anderson et al. 2018) and LSQ14an (Inserra et al. 2017), whereas the fast subsample includes SN 2011ke and PTF11rks (Inserra et al. 2013) (see Fig. 11). The apparent magnitudes of the last two were taken from The Open Supernova Catalog (https://sne.space/, Guillochon et al. 2017). Soon after the maximum luminosity, the LC decline of SN 2017gci is much faster than the slow SLSNe, except for SN 2015bn which shows an initial change of slope after the maximum luminosity. This might suggest that SN 2017gci is a fast SLSN I, as confirmed by the comparison with SN 2011ke and PTF11rks which fairly well reproduce the decline of SN 2017gci. 4.3.2 Spectroscopic comparison Moreover, three spectra of SN 2017gci (at 𝜙=−7,51,133 days) have been compared with the spectra of other SLSNe I (see Fig. 12). To the previous comparison sample, we added also two spectra of the intermediate-evolving type I SLSN Gaia16apd (Kangas et al. 2017). At pre-maximum/maximum epochs the spectral features of SN2017gcishowsimilaritieswiththoseoftwoslowSLSNeI,namely SN 2015bn and SN 2018bsz (Anderson et al. 2018). In particular, the presence of the broad C II features on the red side of the spectrum makes SN 2017gci look like a slow SLSN I. At later phases, the spectra become more similar each other. After ∼40 days from maximum light, the spectra show several broad features and nearly reproduce the overall spectral behaviour of SNe Ic BL at maximum luminosity. This actually holds true both for the slow and fast-evolving SLSNe I (whose prototype is SN 2011ke). Similarly, for 𝜙&100 days, the spectrum of SN 2017gci has characteristics similar to the other SLSNe of the sample, with the presence of Mg I], [Ca II] and the O Iemissions. As already mentioned, the resemblance of the late (𝜙50 days) post-maximum spectra of SN 2017gci with those of a SN Ic BL was verified via the GEneric cLAssification TOol (GELATO, Harutyunyan et al. 2008) which, for the spectra at 𝜙=51,133 days respectively outputs as best-match template SN 2005az (a Type Ic SN, 𝜙=1day) and SN 1997ef (a type Ic BL SN, at 𝜙=41 days in Fig. 13). In particular, the remarkMNRAS 000,1–17 (2015) 16 A. Fiore et al. the Legacy Survey of Space and Time (LSST) at the Vera Rubin Observatory (VRO) will discover a huge number of SLSNe (Villar et al. 2018), which would be crucial especially for very early detections. In addition, three-dimensional hydrodynamical modelling including an improved treatment of radiative transport will allow us to better investigate the properties of SLSNe at both early and late phases, thus boosting our understanding of the underlying explosion mechanism (Soker & Gilkis 2017) as well as the nature of the progenitor stars. DATA AVAILABILITY STATEMENT The data presented in this article and listed in the Appendix A are available in the online supplementary material. ACKNOWLEDGEMENTS This article has been accepted for publication in MNRAS published by Oxford University Press on behalf of the Royal Astronomical Society. We thank the anonymous referee for the very useful comments, which contributed to improve the manuscript. AF is partially supported by the PRIN-INAF 2017 with the project Towards the SKA and CTA era: discovery, localisation, and physics of transients sources (P.I. M. Giroletti). These observations made use of the LCO network. DAH, CP, DH, and JB are supported by NSF Grant AST-1911225 and NASA Grant 80NSSC19k1639. TMB was funded by the CONICYT PFCHA / DOCTORADOBECAS CHILE/201772180113. MG is supported by the Polish NCN MAESTRO grant 2014/14/A/ST9/00121. TWC acknowledges the funding provided by the Alexander von Humboldt Foundation and the EU Funding under Marie Skłodowska-Curie grant agreement No 842471, and Thomas Krühler for reducing X-Shooter spectrum. LG was funded by the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 839090. This work has been partially supported by the Spanish grant PGC2018-095317-B-C21 within the European Funds for Regional Development (FEDER). CPG acknowledges support from EU/FP7ERC grant no. [615929]. GL was supported by a research grant (19054) from VILLUM FONDEN. MN is supported by a Royal Astronomical Society Research Fellowship. R.L. is supported by a Marie Skłodowska-Curie Individual Fellowship within the Horizon 2020 European Union (EU) Framework Programme for Research and Innovation (H2020-MSCA-IF-2017-794467). GT acknowledges partial support by the National Science Foundation under Award No. AST-1909796. Research by S.V. is supported by NSF grants AST-1813176 and AST-2008108. Some of the observations reported here were obtained at the MMT Observatory, a joint facility of the University of Arizona and the Smithsonian Institution under program 2018A-UAO-G16 (PI Terreran). Some of the data presented herein were obtained at the W. M. Keck Observatory, which is operated as a scientific partnership among the California Institute of Technology, the University of California, and the National Aeronautics and Space Administration under program NW440 (PI Fong). The Observatory was made possible by the generous financial support of the W. M. Keck Foundation. The authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of Maunakea has always had within the indigenous Hawaiian community. We are most fortunate to have the opportunity to conduct observations from this mountain. W. M. Keck Observatory and MMT Observatory access was supported by Northwestern University and the Center for Interdisciplinary Exploration and Research in Astrophysics (CIERA). Based on observations collected at the European Organisation for Astronomical Research in the Southern Hemisphere under ESO programmes 199.D-0143, 0100.D0751(B), 0101.D-0199(B), 099.A-9025(A), 0100.A-9099(A)099.A9099 and 0100.A-9099. This work makes use of observations from the LCO network. Part of the funding for GROND (both hardware as well as personnel) was generously granted from the LeibnizPrize to Prof. G. Hasinger (DFG grant HA 1850/28-1). The PanSTARRS1 Surveys (PS1) have been made possible through contributions of the Institute for Astronomy, the University of Hawaii, the Pan-STARRS Project Office, the Max-Planck Society and its participating institutes, the Max Planck Institute for Astronomy, Heidelberg, and the Max Planck Institute for Extraterrestrial Physics, Garching, The Johns Hopkins University, Durham University, the University of Edinburgh, Queen’s University Belfast, the HarvardSmithsonian Center for Astrophysics, the Las Cumbres Observatory Global Telescope Network Incorporated, the National Central University of Taiwan, the Space Telescope Science Institute, the National Aeronautics and Space Administration Grants No.s NNX08AR22G, NNX12AR65G, and NNX14AM74G, the National Science Foundation under Grant No. AST-1238877, the University of Maryland, Eotvos Lorand University (ELTE), the Los Alamos National Laboratory and the Gordon and Betty Moore foundation. The ATLAS surveys are funded through NASA grants NNX12AR55G. This work has made use of data from the European Space Agency (ESA) mission Gaia (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, https: //www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement. This research made use of TARDIS, a community-developed software package for spectral synthesis in supernovae (Kerzendorf & Sim 2014). The development of TARDIS received support from the Google Summer of Code initiative and from ESA’s Summer of Code in Space program. TARDIS makes extensive use of Astropy and PyNE. REFERENCES Anderson J. P. et al. 2018, A&A, 620, 67 Andrews J. E., Smith N. 2018, MNRAS, 477, 74 Anupama G. C., Sahu D. K., Deng J., Nomoto K., Tominaga N., Tanaka M., Mazzali P. A., Prabhu, T. P., 2005, 631, L125 Arnett, W. D., 1982, ApJ, 253, 785 Brown T. M. et al.2011, BAAS, 21813202B Becker A. 2015, HOTPANTS: High Order Transform of PSF ANd Template Subtraction, Astrophysics Source Code Library, ascl:1504.004 Bianco F. B., Modjaz M., Oh S. M., Fierroz D., Liu Y. Q., Kewley L., Graur O., 2016, 16, 54 Buzzoni B. et al. 1984, The Messenger, 38, 9 Calzetti D., Kinney A. L., Storchi-Bergmann T., 1994, ApJ, 429, 582 Cappellaro E. 2014, SNOoPy: a package for PSF fitting and template subtraction, http://sngroup.oapd.inaf.it/snoopy.html/ Chambers K. C. et al. 2016, preprint (arXiv:1612.05560) Chatzopoulos E., Wheeler J. C., Vinko J. 2012, ApJ, 746, 121 Chatzopoulos E., Wheeler J. C., Vinko J., Horvath Z. L., Nagy A. 2013, ApJ, 773, 76 Chen T. -W. et al. 2013, ApJ, 763, L28 Chen T. -W. et al. 2015, MNRAS, 452, 1567 Chen K.-J., Woosley S. E., Sukhbold T., 2016, ApJ, 832, 73 Chen T. W., Smartt S.J., Yates R. M., Nicholl M., Krühler T., Schady P., Dennefeld M., Inserra C.,2017, MNRAS, 470, 3566 Chen T. W., Schady P., Xiao L., Eldridge J. J., Schweyer T., Lee C.-H., Yu Po-Chieh, Smartt S. J., Inserra C., 2017, ApJ, 470, 3566 Chen et al. 2018, ApJ, 867, 31 Chevalier R. A., Irwin C. M. 2011, ApJL, 729, L6 Chonis T. S., Gaskell C. M., 2008, AJ, 135, 264 Chevalier R. A., Fransson C., 2003, 598, 171 Cikota A. et al. 2017, MNRAS, 469, 4705 De Cia A. et al. 2018, ApJ, 860, 100 Delgado A., Harrison D., Hodgkin S., Leeuwen M. V., Rixon G., Yoldas A., 2017, TNSTR, 873 MNRAS 000,1–17 (2015) The bumpy, superluminous SN 2017gci 17 Denicoló G., Terlevich R., Terlevich E. 2002, MNRAS, 330, 69 Dessart L., Waldman R., Livne E., Hillier D. J., Blondin S. 2013, MNRAS, 428, 3227 Elias-Rosa N. et al., 2006, MNRAS, 369, 1880 Fabricant, D., Fata, R., Epps, H., et al. 2019, PASP, 131, 075004 Freudling W., Romaniello M., Bramich D.M., Ballester P., Forchi V., GarcíaDabló C.E., Moehle S., Neese M. J. A&A, 2013, 559, A96 Gal-Yam A. et al. 2009, Nature, 462, 624 Gal-Yam A., 2012, Science, 337, 927 Gal-Yam A., 2018, preprint(arXiv181201428G) Gaia Collaboration et al. 2016, Gaia Data Release 1. Summary of the astrometric, photometric, and survey properties, A&A, 595, A2 Gaia Collaboration et al. 2016 The Gaia mission, A&A, 595, A1 Ginzburg S., Balberg S. 2012, ApJ, 757, 178 Greiner J. et al. 2008, PASP, 120, 405 Greiner J. et al. 2015, Nature, 523, 189 Guillochon J., Parrent J., Kelley L. Z., Margutti R., 2017, ApJ, 835, 64 Howell D. A. 2017 Handbook of Supernovae, ISBN 978-3-319-21845-8. Springer International Publishing AG, 2017, p. 431, 431 Harutyunyan A., Pfahler P., Pastorello A. et al. 2008, A&A, 488, 383 Inserra C. et al. 2013, ApJ, 770, 128 Inserra C. et al. 2017, MNRAS, 468, 4642-4662 Inserra C. 2019, Nature Astronomy, 3, 697 Izzo L. et al. 2018, A&A, 610, A11 Jerkstrand A., Fransson C., Kozma C. 2011, A&A530, A45 Jerkstrand A., Fransson C., Maguire K., Smartt S., Ergon M., Spyromilio J. 2012, A&A, 546, A28 Jerkstrand A., Smartt S. J., Heger, A. 2016, MNRAS, 455, 3207 Jerkstrand A. et al. 2017, ApJ, 835, 13 Kangas T. et al. 2017, MNRAS, 469, 1246 Kasen D., Bildsten L. 2010, ApJ, 717, 245 Kewley L. J., Dopita M. A. 2002, ApJS, 142, 35 Kewley L. J., Ellison S. L. 2008, ApJ, 681, 1183 Kerzendorf W. E., Sim S. A. 2014, MNRAS, 440, 387 Krühler T. et al. 2008, ApJ, 685, 376 Kumar A. et al. 2020, ApJ, 892, 28 Leloudas et al. 2015, MNRAS, 449, 917 Liu, Yu-Q., Modjaz, M., Bianco, F. B. 2017, ApJ, 845, 85 Lyman J. et al. 2017, The Astronomer’s Telegram, No. 10650 Lunnan R. et al. 2014, ApJ, 787, 138 Lunnan R. et al. 2018, 2018, Nature Astronomy, 2, 887 Lunnan R. et al. 2019, preprint(arXiv:1910.02968) Maiolino R., Nagao T., Grazian A. et al. 2008, A&A, 488, 463 Marino R. A., Rosales-Ortega F. F., S ‘anche, S. F, et al. 2013, A&A, 559, A114 Marcote et al. 2020, Nature, 577, 190 Margalit B., Metzger B. D., Thompson T. A., Nicholl M., Sukhbold T. 2018, MNRAS, 475,2659 Metzger B. D., Vurm, I., Hascoet R., Beloborodov A. M. 2014, MNRAS, 437, 703 Milisavljevic D., Fesen R. A., Gerardy C. L., Kirshner R. P., Challis, P. 2010, ApJ, 709, 1343 Moorwood A., Cuby J.-G., Lidman C. 1998, The Messenger, 91, 9 Nakamura T. et al. 2000, in IAU Symp. 195, Highly Energetic Physical Processes and Mechanisms for Emission from Astrophysical Plasmas, ed. P. C. H. Martens, S. Tsuruta, M. Weber (Michigan : ASP), 347 Nasa High Energy Astrophysics Science Archive Research Center (Heasarc) 2014, Astrophysics Source Code Library, ascl:1408.004 Nicholl M. et al. 2013,Nature, 502, 346–349 Nicholl M. et al. 2014, MNRAS, 444, 2096 Nicholl M. et al. 2015, MNRAS, 452, 3869 Nicholl M. et al. 2016a, ApJ, 826, 39 Nicholl M. et al. 2016b, ApJ, 828, L18 Nicholl M., Williams P. K. G., Berger E., Villar V. A., Alexander K. D., Eftekhari T., Metzger B. D. 2017, ApJ, 843, 84 Nicholl M., Guillochon J., Berger E. 2017, ApJ, 850, 55 Nicholl M., Berger E., Blanchard P. K., Gomez S., Chornock R. 2019, ApJ, 871, 102 Oke, J. B., Cohen, J. G., Carr, M., et al. 1995, PASP, 107, 375 Pettini M., Pagel B. E. J. 2004, MNRAS, 348, L59 Pastorello A. et al. 2010, ApJ, 724, L16 Perley D. A. et al. 2016, ApJ, 830, 13 Pignata G. et al., 2004, MNRAS, 355, 178 Poznanski D., Prochaska J. X., Bloom J. S. 2012, MNRAS, 426, 1465 Quimby R. M. et al. 2018 ApJ, 855, 2 Rakavy G., Shaviv G. 1967, ApJ, 148, 803 Ravi, V., Catha, M., D’Addario, L., Djorgovski, S. G., Hallinan, G., Hobbs, R., Kocz, J., Kulkarni, S. R. 2019 Nature 572, 352 Sahu D. K., Gurugubelli U. K., Anupama G. C., Nomoto K., 2011, MNRAS,413, 2583 Salgado J., González-Núñez J., Gutiérrez-Sánchez R.,Segovia J. C., Durán J., Hernández J. L. and Arviset C. 2017 The ESA Gaia Archive: Data Release 1. Astronomy and Computing 21, pp. 22-26. Schlafly E. F., Finkbeiner D. P. 2011, ApJ, 737, 103 Schulze S. et al. 2018, MNRAS, 473, 1258 Smartt S. J. et al. 2015, A&A, 579, A40 Smith N. et al. 2007, 666, 1116 Skrutskie M. F. et al. 2006, AJ, 131, 1163 Soker N., Gilkis A. 2017, ApJ, 851, 95 Stetson P. B. 1987, PASP, 99, 191 Stritzinger M. et al., 2002, AJ, 124, 2100 Suzuki A., Maeda K. 2017, MNRAS, 466, 2633 Suzuki A., Maeda K. 2019, ApJ, 880, 150 Tanga, M., Krühler T., Schady P., Klose S., Graham J. F., Greiner J., Kann D. A., Nardini M. 2018, A&A, 615, A136 Taubenberger S. et al. 2009, MNRAS, 397, 677 Thöne C. C., Christensen L., Prochaska J. X. et al. 2014, MNRAS, 441, 2034 Thöne C. C., de Ugarte Postigo A., García-Benito R., Leloudas G., Schulze S., Amorin R. 2015, MNRAS, 451, L65 Tonry J. L., et al., 2012, ApJ, 750, 99 Valenti S. et al. , 2011, MNRAS, 416, 3138 Van Dokkum P. G., 2001 PASP, 113, 1420 Vernet, J., Dekker, H., D’Odorico, S., et al. 2011, A&A, 536, A10 Villar V. A., Nicholl M., Berger E. 2018, ApJ, 869, 166 Vreeswijk P. M. et al. 2017, ApJ, 835, 58 Yan L. et al. 2015, ApJ, 814, 108 Yan L. et al. 2017, ApJ, 848, 6 Yoshida T., Umeda H., Maeda K., Ishii T. 2016, MNRAS, 457, 351-361, Wheeler J. C., Piran T. and Weinberg S. 1990, Supernovae-Proceedings of the 6th Jerusalem Winter School for Theoretical Physics World Scientific Woosley S. E. 2010, ApJL, 719, L204, doi: 10.1088/2041-8205/719/2/L204 APPENDIX A: TABLES This paper has been typeset from a TEX/L A TEX file prepared by the author. MNRAS 000,1–17 (2015) 18 A. Fiore et al. Table A1. 𝑈𝑉 𝑊 1, 𝑈𝑉 𝑀2, 𝑈𝑉 𝑊 2-filters observed (non K-corrected) aperture magnitudes (in AB system). Errors are in parentheses. MJD r. f. phase from maximum 𝑈𝑉 𝑊 1𝑈𝑉 𝑀2𝑈𝑉 𝑊 2instrument 57986.00 -3.96 - - 19.58(0.08) Swift/UVOT 57986.77 -3.25 - 19.46(0.05) - Swift/UVOT 57989.82 -0.44 18.84(0.07) - - Swift/UVOT 57989.82 -0.44 - - 20.04(0.10) Swift/UVOT 57990.82 -0.44 - 19.84(0.08) - Swift/UVOT 57992.79 2.29 19.32(0.52) - - Swift/UVOT 57992.79 2.29 - 20.11(0.32) - Swift/UVOT 57998.29 7.35 19.22(0.13) - - Swift/UVOT 57998.29 7.35 - 20.29(0.20) - Swift/UVOT 57998.29 7.35 - - 20.43(0.19) Swift/UVOT 58001.47 10.28 19.56(0.10) - - Swift/UVOT 58001.47 10.28 - - 20.54(0.13) Swift/UVOT 58001.48 10.28 - 20.77(0.15) - Swift/UVOT Table A2. 𝑔, 𝑟, 𝑖, 𝑧-filter observed (non K-corrected, non S-corrected) magnitudes (in AB system). Errors are in parentheses. MJD r. f. phase from maximum 𝑔 𝑟 𝑖 𝑧 instrument 57931.00 -54.55 21.10(0.20) - - - Gaia 57977.44 -11.83 17.30(0.20) - - - Gaia 57983.44 -6.31 17.07(0.02) 17.24(0.01) 17.33(0.02) 17.63(0.02) GROND 57984.14 -5.67 - 17.23(0.06) 17.20(0.05) - LCO+Sinistro 57986.80 -3.22 17.31(0.01) 17.23(0.01) 17.27(0.01) - LCO+Sinistro 57991.14 0.78 17.27(0.01) 17.27(0.02) 17.26(0.01) - LCO+Sinistro 57991.15 0.78 - - - 17.80(0.05) LCO+Sinistro 57994.78 4.13 17.37(0.01) 17.21(0.01) 17.32(0.01) - LCO+Sinistro 57999.39 8.36 - - - 17.48(0.03) LCO+Sinistro 58003.38 12.03 - - - 17.48(0.03) LCO+Sinistro 58007.78 16.08 17.43(0.06) 17.44(0.03) 17.39(0.07) - LCO+Sinistro 58008.38 16.63 17.43(0.01) 17.39(0.01) 17.39(0.01) 17.41(0.01) GROND 58017.37 24.90 17.99(0.01) 17.65(0.01) 17.77(0.01) 17.75(0.01) GROND 58019.68 27.03 18.27(0.04) 17.91(0.04) - - LCO+Sinistro 58020.36 27.65 18.18(0.01) 17.66(0.01) 17.78(0.01) 17.74(0.03) GROND 58024.08 31.08 18.20(0.05) 17.98(0.02) 17.82(0.03) - LCO+Sinistro 58025.30 32.20 18.19(0.13) 18.06(0.04) 18.01(0.03) 17.77(0.06) GROND 58027.12 33.88 18.37(0.05) 18.04(0.02) 17.86(0.03) - LCO+Sinistro 58029.32 35.89 18.39(0.02) 17.97(0.01) 17.75(0.03) 17.80(0.04) GROND 58031.00 37.45 18.53(0.08) 18.09(0.06) 17.97(0.04) - LCO+Sinistro 58032.08 38.43 18.70(0.02) 18.27(0.03) - - LCO+Sinistro 58033.30 39.56 18.88(0.02) 18.07(0.01) 18.04(0.02) 17.95(0.01) GROND 58036.28 42.30 18.86(0.03) - 18.13(0.02) - LCO+Sinistro 58040.02 45.74 18.89(0.02) 18.39(0.02) 18.28(0.02) - LCO+Sinistro 58040.24 45.94 18.94(0.01) 18.41(0.01) 18.37(0.01) 18.18(0.02) GROND 58044.01 49.41 19.36(0.04) 18.72(0.03) 18.69(0.04) - LCO+Sinistro 58044.34 49.71 19.36(0.01) 18.53(0.04) 18.69(0.02) 18.59(0.04) GROND 58047.75 52.85 19.62(0.06) 19.03(0.12) - - LCO+Sinistro 58047.75 52.85 - - 18.94(0.20) - LCO+Sinistro 58049.35 54.32 20.14(0.04) 19.32(0.04) 19.29(0.04) 18.77(0.05) GROND 58054.29 58.87 20.50(0.04) 19.74(0.02) 19.54(0.03) 19.27(0.03) GROND 58057.70 62.01 20.48(0.08) 19.90(0.06) - - LCO+Sinistro 58063.21 67.07 20.60(0.11) 19.76(0.07) 19.52(0.04) 19.35(0.06) GROND 58063.22 67.09 - 19.87(0.09) 19.59(0.08) - LCO+Sinistro 58068.31 71.77 20.59(0.05) 19.73(0.02) 19.50(0.02) 19.32(0.03) GROND 58071.24 74.46 20.70(0.06) 20.01(0.05) 19.70(0.05) - LCO+Sinistro 58072.30 75.44 20.70(0.05) 19.77(0.02) 19.72(0.05) 19.33(0.04) GROND 58078.20 80.86 20.76(0.07) 19.95(0.02) 19.77(0.02) 19.31(0.03) GROND 58079.16 81.75 20.78(0.06) 20.13(0.05) 19.82(0.07) - LCO+Sinistro MNRAS 000,1–17 (2015) The bumpy, superluminous SN 2017gci 19 Table A3. (continued) 58083.23 85.49 20.81(0.04) 20.04(0.02) 19.75(0.02) 19.07(0.02) GROND 58087.26 89.20 20.89(0.06) 20.00(0.02) 19.88(0.04) 19.46(0.03) GROND 58092.19 93.74 20.95(0.10) 20.18(0.06) 20.01(0.06) 19.77(0.07) GROND 58095.02 96.34 20.92(0.10) 20.28(0.11) 20.12(0.09) - LCO+Sinistro 58098.12 99.19 21.20(0.06) 20.34(0.02) 20.14(0.03) 19.92(0.05) GROND 58101.16 101.99 21.17(0.06) 20.50(0.05) 20.05(0.06) - LCO+Sinistro 58102.20 102.95 21.07(0.06) 20.48(0.03) 19.89(0.02) 19.76(0.04) GROND 58109.64 109.79 21.00(0.08) 20.34(0.06) 20.05(0.07) - LCO+Sinistro 58117.16 116.70 20.91(0.06) 20.16(0.05) 19.89(0.03) - LCO+Sinistro 58123.25 122.31 21.41(0.10) 20.60(0.03) 20.55(0.03) 20.04(0.06) GROND 58124.90 123.83 21.43(0.11) 20.89(0.11) 20.95(0.13) - LCO+Sinistro 58129.25 127.83 21.86(0.12) 21.00(0.05) 20.86(0.05) 20.44(0.04) GROND 58134.15 132.34 22.13(0.12) 21.20(0.05) 21.10(0.04) 20.52(0.05) GROND 58139.20 136.98 22.28(0.12) 21.13(0.05) 21.23(0.05) 20.53(0.06) GROND 58143.18 140.64 22.23(0.12) 21.24(0.06) 21.10(0.04) 20.74(0.07) GROND 58151.10 147.93 21.72(0.12) 20.93(0.04) 20.74(0.03) 20.37(0.04) GROND 58163.17 159.03 22.03(0.06) 21.21(0.05) 21.01(0.04) 20.70(0.04) GROND 58170.12 165.43 22.22(0.19) 21.30(0.04) 21.16(0.05) 20.57(0.08) GROND 58183.07 177.34 22.63(0.10) 21.72(0.06) - - GROND 58188.09 181.96 - 22.14(0.06) 22.62(0.07) 21.76(0.07) GROND 58191.18 184.80 - 22.34(0.05) 22.39(0.09) 21.67(0.09) GROND 58192.04 185.59 23.25(0.10) 22.43(0.05) 23.17(0.09) 22.27(0.13) GROND 58200.01 192.93 - 22.56(0.10) 23.17(0.12) - GROND 58206.01 198.45 - 22.56(0.09) - 22.73(0.18) GROND 58212.01 203.97 - 22.81(0.12) 22.81(0.18) 22.13(0.18) GROND 58221.02 212.25 - 22.76(0.07) - - GROND 58257.99 246.27 - &22.86 &22.90 - EFOSC2 58370.37 349.65 - &23.19 - - EFOSC2 58467.31 438.83 - - - - EFOSC2 58469.26 440.63 - - &23.25 - EFOSC2 MNRAS 000,1–17 (2015) 20 A. Fiore et al. Table A4. 𝑈, 𝐵, 𝑉 -observed (non K-corrected, non S-corrected) magnitudes (in AB system). Swift/UVOT photometry was measured with a 500-radius aperture (see text). Errors are in parentheses. MJD r. f. phase from maximum 𝑈 𝐵 𝑉 instrument 57984.12 -5.68 - 17.36(0.05) 17.30(0.05) LCO+Sinistro 57986.79 -3.23 - 17.23(0.01) 17.45(0.01) LCO+Sinistro 57987.78 -2.31 - 17.33(0.01) 17.42(0.00) LCO+Sinistro 57989.82 -0.44 17.48(0.05) - - Swift/UVOT 57989.82 -0.44 - 17.21(0.06) - Swift/UVOT 57989.82 -0.44 - - 17.21(0.12) Swift/UVOT 57991.14 0.77 - 17.07(0.03) 17.43(0.01) LCO+Sinistro 57994.78 4.12 - 17.36(0.01) 17.33(0.00) LCO+Sinistro 57998.29 7.35 17.76(0.09) - - Swift/UVOT 57998.29 7.35 - 17.34(0.10) - Swift/UVOT 57998.29 7.35 - - 17.17(0.16) Swift/UVOT 57999.34 8.32 - 17.43(0.02) 17.38(0.01) LCO+Sinistro 58001.47 10.28 17.92(0.06) - - Swift/UVOT 58001.47 10.28 - 17.39(0.06) - Swift/UVOT 58001.48 10.28 - - 17.28(0.11) Swift/UVOT 58003.12 11.79 - 17.51(0.01) 17.25(0.01) LCO+Sinistro 58007.10 15.46 - 17.74(0.00) 17.32(0.01) LCO+Sinistro 58007.76 16.06 - 17.80(0.01) 17.37(0.01) LCO+Sinistro 58011.73 19.71 - 17.85(0.01) 17.56(0.01) LCO+Sinistro 58015.68 23.35 - 18.27(0.01) 17.68(0.01) LCO+Sinistro 58027.12 33.87 - 18.74(0.01) 18.18(0.01) LCO+Sinistro 58031.00 37.44 - 18.76(0.04) 18.19(0.04) LCO+Sinistro 58032.06 38.42 - 18.70(0.01) 18.19(0.01) LCO+Sinistro 58036.26 42.29 - 18.91(0.02) 18.39(0.01) LCO+Sinistro 58040.00 45.72 - 19.31(0.01) 18.56(0.01) LCO+Sinistro 58044.00 49.40 - 19.68(0.04) 18.82(0.05) LCO+Sinistro 58047.74 52.84 - 20.25(0.02) 19.50(0.02) LCO+Sinistro 58057.70 62.00 - 20.92(0.08) 20.20(0.06) LCO+Sinistro 58071.22 74.44 - 21.14(0.06) 20.02(0.03) LCO+Sinistro 58095.01 96.33 - 21.56(0.10) 20.26(0.03) LCO+Sinistro 58101.14 101.97 - 21.72(0.06) 20.68(0.04) LCO+Sinistro 58109.62 109.77 - 21.16(0.08) 20.59(0.06) LCO+Sinistro 58117.14 116.68 - 21.20(0.06) 20.41(0.03) LCO+Sinistro 58124.88 123.80 - 21.82(0.09) 20.86(0.05) LCO+Sinistro MNRAS 000,1–17 (2015) The bumpy, superluminous SN 2017gci 21 Table A5. NIR-observed (non K-corrected) template-subtracted (𝐽 , 𝐻 ) magnitudes and PSF (𝐾s) magnitudes (in AB system). Errors are in parentheses. MJD r. f. phase from maximum 𝐽 𝐻 𝐾sinstrument 57983.44 -6.31 17.75(0.02) 18.19(0.03) 18.68(0.08) GROND 57996.40 5.61 17.72(0.05) &14.957 18.64(0.12) SOFI 58008.38 16.63 17.82(0.02) 18.13(0.04) 18.39(0.05) GROND 58017.37 24.90 17.99(0.04) 18.40(0.05) 18.73(0.09) GROND 58017.38 24.91 17.92(0.02) 18.23(0.03) 18.63(0.05) SOFI 58020.36 27.65 18.16(0.02) 18.30(0.04) 18.66(0.05) GROND 58025.30 32.20 17.95(0.07) 18.42(0.09) - GROND 58029.31 35.89 17.99(0.03) 18.50(0.03) 18.80(0.07) GROND 58033.30 39.56 18.16(0.02) 18.57(0.04) 18.94(0.08) GROND 58056.29 60.71 &17.64 19.46(0.04) 19.70(0.06) SOFI 58078.20 80.86 19.70(0.03) 19.52(0.04) - GROND 58083.23 85.49 19.29(0.04) 19.51(0.05) &19.76 GROND 58087.26 89.20 19.44(0.04) 19.45(0.04) 19.38(0.10) GROND 58092.19 93.74 19.47(0.06) 19.48(0.09) &19.52 GROND 58098.12 99.19 19.80(0.08) 19.63(0.05) 19.38(0.10) GROND 58102.21 102.95 - 19.51(0.06) - GROND 58108.11 108.38 19.64(0.04) 19.65(0.06) &19.50 GROND 58118.23 117.69 19.59(0.07) 19.27(0.04) 19.93(0.19) GROND 58123.25 122.31 20.04(0.09) 19.74(0.08) - GROND 58129.25 127.83 20.67(0.09) 20.26(0.09) &18.98 GROND 58134.15 132.34 20.55(0.08) 19.98(0.10) 19.37(0.17) GROND 58139.20 136.98 20.54(0.13) 20.32(0.08) &19.91 GROND 58143.18 140.64 20.52(0.06) 20.66(0.08) - GROND 58163.17 159.03 20.44(0.09) 19.86(0.07) &19.29 GROND 58170.12 165.43 20.65(0.12) - - GROND 58178.23 172.89 &20.40 20.04(0.11) &19.09 GROND 58183.07 177.34 &20.63 20.35(0.16) &19.06 GROND 58188.09 181.96 &21.03 &20.62 - GROND 58191.18 184.80 &21.05 &20.50 &19.44 GROND 58192.04 185.59 &19.49 &20.56 &18.14 GROND 58200.01 192.93 &21.69 -&19.59 GROND 58206.01 198.45 &21.58 &20.91 &19.43 GROND 58212.01 203.97 21.43(0.15) &21.05 &19.23 GROND 58221.02 212.25 - 20.74(0.09) - GROND 58411.32 387.32 21.78(0.13) &21.25 &20.22 GROND 58427.25 401.98 &21.73 &21.41 &20.22 GROND Table A6. S-corrections for GROND bands. MJD 𝑔 𝑟 𝑖 57892.39 -0.001 -0.013 -0.007 57984.39 0.053 -0.002 -0.030 57986.38 0.045 -0.007 -0.043 57987.38 0.047 -0.009 -0.040 58025.31 -0.017 0.028 0.020 58045.28 -0.057 0.008 0.022 58069.21 -0.081 0.003 0.032 58102.20 -0.066 -0.010 0.037 58132.45 -0.095 -0.031 0.054 58135.22 -0.075 -0.030 0.036 58159.23 -0.102 0.0560 -0.074 58192.10 -0.045 0.0570 -0.052 MNRAS 000,1–17 (2015) 22 A. Fiore et al. Table A7. S-corrections for Sinistro filters. MJD 𝐵 𝑔 𝑉 𝑟 𝑖 𝑧 57892.39 0.000 0.001 0.004 -0.016 -0.009 -0.005 57984.39 -0.002 -0.008 0.036 -0.006 -0.009 -0.059 57986.38 -0.002 -0.006 0.024 -0.005 -0.013 -0.057 57987.38 -0.003 -0.007 0.026 -0.007 -0.012 -0.057 58025.31 -0.003 0.013 0.010 -0.007 0.000 -0.054 58045.28 -0.008 0.021 -0.004 -0.017 -0.003 -0.052 58069.21 -0.006 0.018 -0.013 -0.014 -0.003 -0.049 58102.20 -0.009 0.015 -0.012 -0.026 -0.005 -0.051 58132.45 -0.006 -0.001 -0.014 -0.027 -0.007 -0.009 58135.22 -0.009 0.010 -0.014 -0.037 -0.007 -0.049 58159.23 -0.013 0.024 -0.014 0.056 -0.007 -0.008 58192.10 0.001 0.053 -0.008 0.020 -0.010 -0.005 Table A8. S-corrections for Swift/UVOT filters. MJD 𝐵 𝑉 57892.39 -0.000 0.013 57984.39 0.001 0.019 57986.38 -0.001 0.010 57987.38 -0.001 0.012 58025.31 0.006 0.049 58045.28 0.001 0.037 58069.21 -0.001 0.019 58102.20 -0.001 0.003 58132.45 -0.012 -0.017 58135.22 -0.011 -0.010 58159.23 -0.024 -0.004 58192.10 -0.000 0.020 Table A9. Estimated uncertainties Δ𝑆corr for the filters 𝑈 , 𝑧, 𝐽 , 𝐻 , 𝐾s(for each instrument) divided in two temperature ranges (see text). 4000K < 𝑇 < 8000K 8000K < 𝑇 < 12000K GROND Δ𝑆corr,𝑧 =0.010 mag Δ𝑆corr,𝐽 =0.004 mag Δ𝑆corr,𝐻 =0.001mag Δ𝑆corr,𝐾s=0.010 mag Δ𝑆corr,𝑧 =0.002 mag Δ𝑆corr,𝐽 =0.001 mag Δ𝑆corr,𝐻 =0.000mag Δ𝑆corr,𝐾s=0.001 mag SOFI Δ𝑆corr,𝐽 =0.020 mag Δ𝑆corr,𝐻 =0.080mag Δ𝑆corr,𝐾s=0.070 mag Δ𝑆corr,𝐽 =0.002 mag Δ𝑆corr,𝐻 =0.001mag Δ𝑆corr,𝐾s=0.120 mag LCO+Sinistro Δ𝑆corr,𝑧 =0.001 mag Δ𝑆corr,𝑧 =0.005mag Swift/UVOT Δ𝑆corr,𝑈 =0.2 mag Δ𝑆corr,𝑈 =0.05mag MNRAS 000,1–17 (2015) The bumpy, superluminous SN 2017gci 23 Table A10. K-corrections expressed in magnitudes. r. f. phase from maximum 𝑈𝑉 𝑊 2𝑈𝑉 𝑀2𝑈𝑉 𝑊 1𝑈 𝐵 𝑔 𝑉 𝑟 𝑖 𝑧 𝐽 𝐻 𝐾s -7 0.39 0.51 0.26 0.04 -0.18 -0.18 -0.21 -0.23 -0.24 -1.55 -0.24 -0.25 -0.25 -5 0.37 0.5 0.26 0.04 -0.17 -0.17 -0.21 -0.24 -0.24 -0.24 -0.22 -0.22 -0.25 -4 0.35 0.49 0.26 0.04 -0.16 -0.15 -0.17 -0.16 -0.27 -1.45 -0.22 -0.23 -0.24 32 0.06 0.28 0.27 0.02 0.08 0.02 -0.07 -0.21 -0.14 -1.54 -0.16 -0.20 -0.22 51 -0.07 0.17 0.27 0.02 0.22 0.13 -0.02 -0.13 -0.16 -1.52 -0.15 -0.18 -0.21 73 -0.24 0.02 0.28 0.01 0.30 0.19 0.01 -0.08 -0.06 -1.60 0.58 -0.10 -0.20 103 -0.4 -0.13 0.28 0.01 0.26 0.15 -0.01 -0.05 -0.09 -1.45 0.18 -0.16 -0.19 135 0.244 0.179 0.006 0.023 -0.094 -1.455 +175 0.164 0.097 -0.067 0.097 -0.131 0.096 1.64097 -0.16296 -0.19509 +358 0.106 0.066 -0.085 0.066 -0.17 -0.378 -0.11968 -0.1644 -0.19602 Table A11. Slopes of the observed LCs [10−2mag/day] 𝑈𝑉 𝑊 2𝑈𝑉 𝑀2𝑈𝑉 𝑊 1𝑈 𝐵 𝑔 𝑉 𝑟 𝑖 𝑧 𝐽 𝐻 𝐾s - - - - - 2.22 - 2.23 2.28 2.25 2.23 1.80 - Table A12. Spectra in Fig. 6. MJD r. f. phase from maximum instrument resolution 57982.39 -7 EFOSC2 18.2 57984.39 -5 EFOSC2 17.9 57986.38 -4 EFOSC2 18.9 57987.38 -3 EFOSC2 27.2 58011.73 20 LCO+FLOYDS 21.1 58015.73 23 LCO+FLOYDS 21.8 58021.73 29 LCO+FLOYDS 22.1 58025.31 32 EFOSC2 17.9 58030.71 37 LCO+FLOYDS 20.0 58035.73 42 LCO+FLOYDS 18.4 58045.28 51 EFOSC2 17.8 58069.21 73 EFOSC2 17.8 58102.20 103 EFOSC2 17.8 58132.45 131 LRIS - 58135.22 133 EFOSC2 18.0 58159.23 155 Binospec - 58192.10 187 X-Shooter - 58389.35 367 X-Shooter - MNRAS 000,1–17 (2015) 24 A. Fiore et al. Table A13. Logarithm of the bolometric luminosities integrated over the𝑈𝑉 𝑊 2, 𝑈𝑉 𝑀2, 𝑈𝑉 𝑊 1, 𝑈 , 𝐵, 𝑔, 𝑉 , 𝑟, 𝑖, 𝑧, 𝐽 , 𝐻 , 𝐾s, the blackbody temperatures (expressed in Kelvin). Errors are reported in parenthesis. We fixed a maximum error for the blackbody temperatures to 2000 K (see text). Epochs later than ∼160 days require even larger errorbars. r. f. phase from maximum log10 𝐿bol 𝑇BB -5.57 43.65(0.06) 11693.01 -4.94 43.63(0.06) 11487.76 -2.48 43.67(0.07) 11633.86 1.51 43.75(0.07) 11429.11 4.86 43.77(0.07) 10699.96 16.81 43.7(0.06) 10775.08 17.37 43.73(0.06) 9313.39 25.64 43.61(0.04) 8909.54 27.76 43.53(0.04) 8597.25 28.39 43.58(0.04) 8393.45 31.81 43.5(0.04) 8351.5 32.93 43.46(0.04) 8555.91 34.61 43.47(0.04) 8681.41 36.62 43.48(0.05) 8248.12 38.18 43.44(0.05) 8210.32 39.16 43.44(0.04) 7489.47 40.29 43.45(0.04) 7478.19 46.47 43.33(0.03) 6997.47 46.68 43.33(0.03) 6997.47 50.15 43.21(0.03) 6887.7 50.45 43.2(0.03) 6205.62 53.59 43.03(0.04) 6061.82 55.06 42.98(0.04) 5678.73 59.6 42.87(0.03) 5640.1 62.74 42.85(0.05) 5586.08 67.81 42.84(0.04) 5562.2 67.82 42.84(0.04) 5421.8 72.5 42.84(0.03) 5471.99 75.19 42.8(0.03) 5356.28 76.17 42.82(0.03) 5357.45 81.6 42.78(0.03) 5407.49 82.48 42.78(0.03) 5324.42 86.23 42.79(0.03) 5218.12 89.94 42.77(0.03) 5192.4 94.47 42.74(0.03) 4961.57 97.07 42.73(0.04) 4920.14 99.93 42.66(0.03) 4974.52 102.72 42.63(0.03) 5085.91 103.68 42.65(0.02) 5171.26 110.52 42.68(0.03) 4629.46 117.43 42.72(0.03) 4610.89 123.05 42.54(0.04) 4338.4 124.56 42.52(0.05) 4165.87 128.56 42.43(0.06) 4180.01 133.07 42.39(0.05) 4121.38 137.72 42.39(0.06) 4384.67 141.38 42.27(0.12) 4236.11 148.67 42.44(0.04) 4162.07 159.77 42.39(0.04) 3710.72 166.16 42.53(0.07) - 178.08 42.31(0.08) - 185.54 42.28(0.07) - 186.33 42.13(0.09) - 193.66 42.17(0.13) - 199.18 42.11(0.14) - 204.7 42.16(0.1) - 212.99 42.16(0.1) - MNRAS 000,1–17 (2015) The bumpy, superluminous SN 2017gci 25 Table A14. Comparison of the metallicity estimated for SN 2017gci with the metallicities of a sample of nearby SLSNe I and GRBs (data from Chen et al. 2017b). Object SN 2017gci SN 2017egm PTF11hrq PTF12dam GRB 100316D GRB 060505 GRB 111005A Reference Sec. 4.1 (Chen et al. 2017b) (Cikota et al. 2017) (Thöne et al. 2015) (Izzo et al. 2018) (Thöne et al. 2014) (Tanga et al. 2017) Redshift 0.087 0.031 0.057 0.107 0.059 0.089 0.013 PP04 O3N2 8.135 ±0.07 8.77 ±0.01 8.19 ±0.01 8.01 ±0.14 8.21 ±0.02 8.24 ±0.00 8.63 ±0.03 Table A15. Best-fit estimates of the physical parameters of SN 2017gci (with reference to Fig. 2). The TigerFit best-fit model is listed in the first column with the assumed phase from the explosions in square brackets. ejecta mass polar mag. initial phase from the opacity CSM progenitor diffusion spin-down mass 𝑀ejecta accretion rate field 𝐵pperiod 𝑃initial explosion 𝜙0𝜅mass radius timescale timescale [M] [Myear−1] [1014 G] [ms] [days] [cm2g−1][M] [1014 cm] [days] [days] MF1 9.0 - 5.5 2.8 15.7 0.08 - - 34.5 1.1 csm0 [30] 12.4 0.1 - - - 0.07 4.9 0.004 - - MNRAS 000,1–17 (2015)