Observational Properties of a Bright Type lax SN 2018cni and a Faint Type Iax SN 2020kyg

Type Ia supernovae (SNe) are the final outcome of the thermonuclear explosion of a carbon–oxygen (CO) white dwarf in a binary system (Hoyle & Fowler 1960; Wang & Han 2012; Maoz et al. 2014; Maeda & Terada 2016; Jha et al. 2019). As a group, Type Ia SNe display great uniformity and are well known as members of one parameter family (Phillips 1993; Phillips et al. 1999). Despite the uniformity, diversity has also been noted from time to time, based on the explosion scenario (Taubenberger 2017). Among different subtypes of thermonuclear SNe, Type Iax SNe emerge as low luminosity (M_V,peak = −13 to −19 mag; Jha 2017) and less energetic members (Foley et al. 2013). They are also known as 2002cx-like SNe after their prototype SN 2002cx (Li et al. 2003; Jha et al. 2006). McClelland et al. (2010) suggested a correlation between peak luminosity, the photospheric velocity at maximum, and the shape of the light curve of Type Iax SNe; however, there are a few outliers, such as SN 2009ku (Narayan et al. 2011), SN 2014ck (Tomasella et al. 2016), etc. Narayan et al. (2011) suggested a scaling relation between the decline rate and ejecta velocity. A possible correlation between the decline rate and absolute magnitude at maximum was also shown by Foley et al. (2013) and Magee et al. (2016). The diversity in the observed properties, such as peak brightness, light curve decline rates, and ejecta velocities, indicates that Type Iax SNe form a heterogeneous class.

During the early phase, spectra of Type Iax SNe exhibit similar spectral signatures as SN 1991T-like Type Ia SNe (Phillips et al. 2007; Foley et al. 2010). They show Fe ii and Fe iii lines during early spectral evolution and do not display any high-velocity features (Silverman et al. 2015). The brighter Type Iax members, such as SNe 2005hk (Sahu et al. 2008), 2012Z (Stritzinger et al. 2015; Yamanaka et al. 2015), and 2020rea (Singh et al. 2022), have weaker Si ii and less distinguished C ii features as compared to the fainter members of this class, such as SNe 2008ha (Foley et al. 2009), 2019gsc (Srivastav et al. 2020; Tomasella et al. 2020), and 2020kyg (Srivastav et al. 2022). Furthermore, the W-shaped feature, attributed to S ii (Hachinger et al. 2006), looks stronger in the spectra of the fainter Type Iax than the brighter ones.

The wide range of luminosity among members of this class presents a challenge in understanding them collectively in terms of a single physical scenario. For SN 2012Z, McCully et al. (2014a) reported the detection of a blue source coinciding with the SN location in the preexplosion images. An accreting white dwarf with a helium star in a binary system has been suggested as the plausible progenitor of SN 2012Z (McCully et al. 2014a, 2022). Jordan et al. (2012), Kromer et al. (2013), and Fink et al. (2014) proposed that the deflagration of a Chandrasekhar-mass white dwarf, leaving behind a remnant, could explain the low luminous nature of the Type Iax class. Foley et al. (2014) investigated the emission observed at the explosion site of SN 2008ha 4 yr after the explosion and discussed the possibilities of its association with the remnant of the white dwarf. Further, the presence of a bound remnant is also supported by late-time flattening found in the light curves of Type Iax SNe (McCully et al. 2014b, 2022; Singh et al. 2018; Kawabata et al. 2021). A similar explosion of a hybrid carbon–oxygen–neon (CONe) white dwarf (Meng & Podsiadlowski 2014; Kromer et al. 2015) yields an outcome consistent with faint Type Iax SNe. Given the variety of possible progenitor channels (Pumo et al. 2009; Valenti et al. 2009; Moriya et al. 2010; Denissenkov et al. 2015; Bravo et al. 2016; Kashyap et al. 2018; Bobrick et al. 2022; Lach et al. 2022), detailed analysis of a statistically complete sample of Type Iax SNe is required to thoroughly understand the nature of this peculiar class of SNe.

Bright and faint members of the Type Iax class show diverse nature in terms of their luminosity, the synthesized mass of ⁵⁶Ni, spectral features, etc. This paper presents an optical follow-up of two Type Iax SNe, 2018cni and 2020kyg. SN 2018cni belongs to the bright Type Iax SNe, and SN 2020kyg is situated at the fainter end of the Type Iax class. We intend to examine the heterogeneity or homogeneity in the two extremes of luminosity distribution of Type Iax SNe. Section 2 provides information about the discovery of the SNe, observations, and reduction of the data. Distance and extinction estimations are given in Section 3. Section 4 presents a detailed description of the photometric properties of the SNe and analytical modeling of the bolometric light curve. Section 5 discusses spectroscopic properties and spectral modeling. A comparative analysis followed by a brief summary is presented in Section 6.

SN 2018cni (also known as PTSS-18fdb, ATLAS18qql, and Gaia18bpu) was detected by Purple Mountain Observatory (PMO) Tsinghua Supernova Survey ²⁴ (Tan et al. 2018) on 2018 June 13. The location of the SN is at 1'' east and 5'' south of the host galaxy GALEXASC J150122.78-101044.8. It was classified as a Type Iax SN by Zhang et al. (2018a, 2018b) using a spectrum obtained on 2018 June 17. This spectrum was found to match well with SN 2005hk, 3 days before maximum light. The location of SN 2018cni in its host galaxy is presented in Figure 1.

Zoom In Zoom Out Reset image size

SN 2020kyg was spotted by Asteroid Terrestrial-impact Last Alert System (ATLAS; Tonry et al. 2018) at 00 north and 356 east to the center of the host galaxy NGC 5012 on 2020 May 24 (Smith et al. 2020; Tonry et al. 2020). ATLAS20nuc and AT2020kyg are other aliases of this source. A nondetection was reported on 2020 May 20 (MJD = 58989.36; Smith et al. 2020). Ochner et al. (2020) and Hiramatsu et al. (2020) classified it as a Type Iax SN before maximum. Figure 1(b) shows the position of SN 2020kyg in the host galaxy NGC 5012.

Optical photometric observations of SNe 2018cni and 2020kyg were initiated ∼6 and 3 days from discovery with the 1 m telescopes of the Las Cumbres Observatory (LCO; Brown et al. 2013) under the Global Supernova Project and continued up to ∼80 and 60 days, respectively. The photometry was performed using the lcogtsnpipe pipeline (Valenti et al. 2016). The BV magnitudes are calibrated to the Vega system, while the gri magnitudes are calibrated to the AB system. SN 2018cni exploded in the proximity of the host galaxy, hence template subtraction was performed in order to eliminate the host galaxy contamination. The template was observed in BVgri bands using the 1 m LCO telescope on 2019 August 6, which is more than a year after the discovery. The image subtraction was performed within the lcogtsnpipe using PyZOGY (Guevel & Hosseinzadeh 2017). The optical photometry of SNe 2018cni and 2020kyg are presented in Tables 1 and 2, respectively.

Spectroscopic observations of SNe 2018cni and 2020kyg were initiated ∼5 days and ∼1 day after the discovery, respectively. Four spectra of SN 2018cni and three spectra of SN 2020kyg were taken with FLOYDS spectrograph on the 2 m Faulkes Telescope North (FTN; Brown et al. 2013). One spectrum of SN 2018cni was acquired with the Blue Channel (BCH) spectrograph on 6.5 m MMT, and two spectra of SN 2018cni were obtained with the Robert Stobie Spectrograph(RSS) mounted on 9.2 m Southern African Large Telescope (SALT). Two spectra of SN 2018cni were obtained using the dual-beam Kast spectrograph (Miller & Stone 1993) on the Lick Shane 3 m telescope. The 300/7500 grating on the red side and the 452/3306 grism on the blue side were used, providing continuous coverage between 3500 and 10400 Å. The slit was aligned along the parallactic angle (Filippenko 1982). Two spectra of SN 2020kyg were taken with Gemini Multi-Object Spectrograph (GMOS-N) using B600 and R400 gratings on Gemini North. Spectral reduction of all spectra taken with the FLOYDS spectrograph was done using the floydsspec ²⁵ pipeline. The spectrum of SN 2018cni observed with MMT was reduced using standard tasks described in Andrews et al. (2019). The two spectra from the Kast spectrograph were reduced in a standard manner using tools in a custom spectroscopy pipeline ²⁶ (Siebert et al. 2019). The two spectra of SN 2018cni obtained from RSS were reduced using a custom pipeline that contains the PYSALT package and standard PYRAF spectral reduction tasks (Crawford et al. 2010). The Gemini spectra of SN 2020kyg were reduced using the standard Gemini IRAF package. The log of spectroscopic observations of both the SNe are provided in Tables 3 and 4, respectively.

The host galaxy of SN 2018cni, GALEXASC J150122.78-101044.8, is located at a redshift of 0.032 (Zhang et al. 2018a). The extinction due to the Milky Way in the direction of SN 2018cni is E(B − V) = 0.09 mag (Schlafly & Finkbeiner 2011), which yields A_V = 0.278 mag using the Cardelli extinction law (Cardelli et al. 1989). The spectra of SN 2018cni do not exhibit significant Na ID absorption at the redshift of the host galaxy, implying negligible host extinction. SN 2020kyg is located on the outskirts of the host galaxy NGC 5012, hence noticeable extinction from the host is not expected. The extinction due to the Milky Way is E(B − V) = 0.012 mag (Schlafly & Finkbeiner 2011), which results in A_V = 0.038 mag assuming R_v = 3.1. We used the redshift-dependent luminosity distance for GALEXASC J150122.78-101044.8 and NGC 5012 as 130.56 ± 12.15 Mpc and 35.83 ± 2.66 Mpc, respectively (assuming H₀ = 73 km s⁻¹ Mpc⁻¹, Ω_m = 0.27, and Ω_v = 0.73).

4.1. Light Curve and Color Curve

Figures 2 and 3 present the light-curve evolution of SNe 2018cni and 2020kyg, respectively. Observations began before both the SNe attained peak brightness. The time and magnitude at maximum along with decline rates for both the SNe are estimated using a low-order spline fit to the light curves. The values are listed in Table 5. Light curve decline rates, Δm₁₅(B), measured for SNe 2018cni and 2020kyg in B-band are 1.99 ± 0.14 mag and 2.54 ± 0.13 mag, respectively. The I-band light curves of SNe 2018cni and 2020kyg do not show a secondary maximum, usually seen in Type Ia SNe. Figure 4 presents the light-curve comparison of SNe 2018cni and 2020kyg in BVRI bands with a few well-studied Type Iax SNe 2002cx (Li et al. 2003), 2005hk (Sahu et al. 2008), 2008ha (Foley et al. 2009), 2010ae (Stritzinger et al. 2014), 2012Z (Yamanaka et al. 2015), and 2019muj (Barna et al. 2021), covering the luminosity range for this class. The r- and i-band magnitudes of SNe 2010ae, 2018cni, and 2020kyg are converted to R- and I-band using transformation equations given by Jordi et al. (2006). The magnitudes of each SN in the respective bands are normalized by their peak magnitudes. In B- and V-bands, SN 2018cni declines faster than SNe 2002cx, 2005hk, and 2012Z, while SN 2020kyg is the fastest declining SN among all the SNe used for comparison, as shown in Figure 4. In R- and I-bands, the light-curve evolution of SN 2018cni is similar to SNe 2005hk and 2012Z, while SN 2020kyg resembles SN 2008ha.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 5. Parameters of SNe 2018cni and 2020kyg

SN 2018cni	B-band	V-band	g-band	r-band	i-band
JD of maximum light (2458000+)	289.99 ± 0.5	289.99 ± 0.5	289.99 ± 0.5	290.00 ± 0.5	290.01 ± 0.5
Magnitude at maximum (mag)	18.36 ± 0.07	18.03 ± 0.04	18.13 ± 0.02	17.98 ± 0.03	18.09 ± 0.08
Absolute magnitude at maximum (mag)	−17.59 ± 0.21	−17.81 ± 0.21	−17.78 ± 0.20	−17.82 ± 0.20	−17.65 ± 0.22
Δm15(mag)	1.99 ± 0.14	0.79 ± 0.07	1.61 ± 0.09	0.48 ± 0.06	0.18 ± 0.08
SN 2020kyg
JD of maximum light (2458000+)	998.94 ± 2.0	996.88 ± 0.5	998.95 ± 2	1000.9 ± 0.5	1000.9 ± 0.5
Magnitude at maximum (mag)	18.36 ± 0.06	18.28 ± 0.14	18.37 ± 0.02	18.31 ± 0.03	18.38 ± 0.04
Absolute magnitude at maximum (mag)	−14.45 ± 0.17	−14.52 ± 0.21	−14.44 ± 0.16	−14.49 ± 0.16	−14.41 ± 0.17
Δm15(mag)	2.54 ± 0.13	0.98 ± 0.19	1.77 ± 0.08	0.84 ± 0.06	0.61 ± 0.06

Download table as:  ASCIITypeset image

Figure 5 presents the reddening-corrected color evolution of SNe 2018cni and 2020kyg and their comparison with other well-studied Type Iax SNe. The (B − V), (V − I), (V − R), and (R − I) color evolution of SNe 2018cni and 2020kyg are found to be similar to the SNe used for comparison. This indicates that bright and faint members behave in a similar fashion in terms of color evolution.

Zoom In Zoom Out Reset image size

4.2. Bolometric Light Curve and Analytical Modeling

The pseudobolometric light curves of SNe 2018cni and 2020kyg are constructed with SuperBol code (Nicholl 2018) in BVri bands using the luminosity distance discussed in Section 3. The pseudobolometric light curves of SNe 2002cx, 2005hk, 2008ha, and 2010ae are constructed in a similar manner and are presented in Figure 6. We find that the bolometric luminosity of SN 2018cni is similar to bright Type Iax SN 2005hk, whereas SN 2020kyg lies between SNe 2008ha and 2010ae and belongs to the fainter class of Type Iax SNe. We have also used the SuperBol code to calculate the BgVri luminosities of SNe 2018cni and 2020kyg, which are shown in Figure 7.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Explosion parameters of both the SNe are estimated using Arnett's model (Arnett 1982) along with the formulation given in Valenti et al. (2008). Homologous expansion, spherically symmetric ejecta, no mixing of ⁵⁶Ni, and optically thick ejecta are basic assumptions used in the model. The mass of ⁵⁶Ni and the timescale of the light curve (τ) are free parameters for the fitting. Optical opacity is fixed to 0.07 cm² g⁻¹ and photospheric velocities used are 7000 km s⁻¹ and 4500 km s⁻¹ (Section 5.1) for SNe 2018cni and 2020kyg, respectively.

Analytical modeling of the pseudobolometric light curve of SN 2018cni gives a mass of ⁵⁶Ni ∼ 0.07 ± 0.01M_⊙ and the ejecta a mass of 0.48 M_⊙. For SN 2020kyg, these estimates are a mass of ⁵⁶Ni ∼ 0.002 ± 0.001 M_⊙ and an ejecta mass of 0.14 M_⊙. For SN 2020kyg, our estimated values are lower than that presented in Srivastav et al. (2022), where ugrizyJHK fluxes were used for estimating the explosion parameters. The evolution of the blackbody temperature and radius for SNe 2018cni and 2020kyg are shown in Figure 8. Despite the large difference in luminosity for the two SNe, their blackbody-temperature evolution is similar, whereas the blackbody radius is larger for the bright SN 2018cni. This suggests that diversity in luminosity is directly related to the radius of the photosphere, which, in turn, is related to expansion velocity.

Zoom In Zoom Out Reset image size

Because of the intrinsic heterogeneity of SNe Iax, the contribution of UV and near-infrared (NIR) flux to the bolometric flux is not well constrained. However, depending on the availability of the data in UV and NIR bands, the fractional contribution of UV and NIR flux to the bolometric flux has been reported for a number of SNe Iax. For SN 2005hk, Phillips et al. (2007) estimated a 20% contribution of UV flux to the UVOIR light curve at the early time of evolution. Yamanaka et al. (2015) showed that the NIR band contributes ∼20% to Opt+NIR around the maximum for SN 2012Z. A total UV+IR contribution of ∼35% has been reported for SN 2014ck (Tomasella et al. 2016). Srivastav et al. (2020) and Dutta et al. (2022) quoted the ratio of peak quasi-bolometric luminosity to peak blackbody bolometric luminosity as 0.69 (SN 2019gsc) and 0.62 (SN 2020sck), respectively. For SN 2020kyg, Srivastav et al. (2022) estimated a 60% contribution from the optical luminosity to the total bolometric luminosity around the maximum. If ∼35% flux is added to the pseudobolometric flux of SNe 2018cni and 2020kyg, the mass of ⁵⁶Ni increases to 0.09 M_⊙ and 0.003 M_⊙, respectively. In this study, we have used pseudobolometric fluxes and reported the lower limit for the explosion parameters.

We compare the BgVri bolometric light curves of SNe 2018cni and 2020kyg with model light curves of deflagration of a CO white dwarf (Fink et al. 2014) and a hybrid CONe white dwarf (Kromer et al. 2015), respectively, in Figure 9. These model light curves are taken from the HESMA database. In the model light curves of CO white dwarfs, N1-def, N3-def, N5-def, and N10-def correspond to 1, 3, 5, and 10 ignition spots, respectively. The increasing number of ignition spots indicates a stronger explosion. We find that SN 2018cni is similar to the N3-def model and the low-luminosity SN 2020kyg is better explained by the deflagration of the CONe white dwarf. Since the models used for comparison include flux from UV to NIR and we have only BgVri fluxes, we intend to only investigate the probable explosion scenario by finding the best match between the theoretical models and the observed luminosity of both the SNe.

Zoom In Zoom Out Reset image size

5.1. Spectral Features and Comparison with Other SNe

Figures 10 and 11 show the spectral sequence of SNe 2018cni and 2020kyg from −2 days to +55 days and −4.0 days to +35 days with respect to the maximum in the B-band, respectively. The first spectrum of SN 2018cni at −2 days shows a blue continuum with Ca ii H&K, Fe iii, and Si iii lines, while beyond 5200 Å, the spectrum is featureless. In the postmaximum phase, the continuum in the blue region weakens. The postmaximum spectra of SN 2018cni exhibit Fe ii features near 5000 Å, Co ii features near 6500 Å, and Ca ii NIR features. Spectra of SN 2018cni are compared with spectra of SNe 2002cx (Li et al. 2003) and 2005hk (Sahu et al. 2008), at comparable epochs in Figure 12. In the premaximum phase (top panel), the strength of the Fe iii λ4420 line appears to be similar in all three SNe, while the Si iii λ4553, 4568 lines in SN 2018cni appear stronger than the other two SNe. During the postmaximum epoch (middle and bottom panel), spectra of SN 2018cni are dominated by iron group elements (IGE) and display a good one-to-one correspondence in the spectral features with SNe 2002cx and 2005hk.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In the spectral evolution of SN 2020kyg, narrow P-Cygni profiles of intermediate-mass elements (IMEs) and IGEs are present. In Figure 12(b), we compare the spectra of SN 2020kyg with other faint Type Iax SNe 2008ha (Foley et al. 2009), 2010ae (Stritzinger et al. 2014), and 2019gsc (Srivastav et al. 2020). In the premaximum phase, strong features due to C ii and Si ii can be seen in all the spectra. The presence of C ii is indicative of unburnt carbon in the progenitor system. In the early postmaximum phase, C ii and Si ii lines become weak, and Fe ii around 5000 Å can be seen with almost equal strength in spectra of SN 2020kyg and other SNe used for comparison. In the postmaximum phase, the Ca ii NIR feature starts developing in all of the SNe. An overall similarity can be seen in the spectral signatures of SN 2020kyg with other SNe used for comparison. Spectra of SN 2020kyg at +16 and +35 days are very similar, mainly dominated by permitted and forbidden lines due to Fe ii, Co ii, and Ca ii. The presence of forbidden lines in the +16 days spectrum indicates a considerably fast evolution of SN 2020kyg and its early transition into the nebular phase.

We measure the expansion velocity of SN 2020kyg by fitting a Gaussian function to the absorption minima of the Si ii profile. The estimated velocities at three epochs (−4.0, −3.1, and +5.9 days) are 4330, 4290, and 3580 km s⁻¹, respectively. The measurement errors associated with the estimation of expansion velocities are ∼600 km s⁻¹. Our estimations are consistent with the values given in Srivastav et al. (2022). As the identification of the Si ii line in SN 2018cni is questionable, we did not attempt to measure its expansion velocity.

5.2. Spectral Modeling

We perform spectral modeling for SN 2018cni at three epochs and SN 2020kyg at four epochs (Figures 13 and 14) using the one-dimensional radiative transfer code TARDIS (Kerzendorf & Sim 2014; Kerzendorf et al. 2018). TARDIS assumes a blackbody-emitting photosphere with a custom-designed, spherically symmetric ejecta. The one-dimensional density and abundance structure was the subject of fitting, following the abundance tomography method described in Barna et al. (2018): as the homologously expanding SN ejecta becomes optically thinner, the velocity of the photosphere decreases and observations can probe deeper into the ejecta.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The assumed density function is designed to resemble those of the various pure deflagration models (Fink et al. 2014), similar to previous attempts of fitting SN Iax spectra (Barna et al. 2021). This density function includes a simple exponential inner and a more strongly decreasing outer part:

$$\begin{eqnarray}\rho (v,{t}_{\exp })=\left\{\begin{array}{ll}{\rho }_{0}\cdot {\left(\displaystyle \frac{{t}_{\exp }}{{t}_{0}}\right)}^{-3}\cdot \exp \left(-\displaystyle \frac{v}{{v}_{0}}\right) & \mathrm{for}\,v\leqslant {v}_{\mathrm{cut}}\\ {\rho }_{0}\cdot {\left(\displaystyle \frac{{t}_{\exp }}{{t}_{0}}\right)}^{-3}\cdot \exp \left(-\displaystyle \frac{v}{{v}_{0}}\right)\cdot {8}^{-\tfrac{8{\left(v-{v}_{\mathrm{cut}}\right)}^{2}}{{v}_{\mathrm{cut}}^{2}}} & \mathrm{for}\,v\gt {v}_{\mathrm{cut}}\end{array}\right.,\end{eqnarray} \tag{ 1 }$$

where the innermost density ρ₀ (at reference time t₀ = 100 s), the slope of the function v₀, and the start of the shallow cutoff past the velocity v_cut are free parameters in the fitting process.

Abundances are constrained by fitting the prominent P-Cygni features at all spectral epochs of the actual SN, simultaneously. The model ejecta is split into 1000 km s⁻¹ zones, where the mass fractions of the prominent chemical elements are fitting parameters.

Due to the high number of mass fraction parameters, the whole parameter space cannot be fully explored manually. Thus, a different combination of physical and chemical properties, which would provide a similarly good match with the observed spectra, cannot be fully ruled out. However, considering that the main spectral features and the flux continuum of each epoch are well reproduced, we conclude that our results provide at least feasible solutions.

Other input parameters are the luminosity and the time since the explosion. The latter parameter controls the dilution of the density structure, and thus, affects the estimated temperature profile of the model ejecta. In the absence of good photometric coverage during the early phase, the explosion date is constrained from the spectral synthesis, as t_exp has a critical impact on the premaximum epochs. The best-fit date is considered as the date of the explosion. In the case of SN 2020kyg, the inferred date of the explosion is in agreement with that of Srivastav et al. (2022; JD = 2458988.3), within the estimated ±1.0 day uncertainty. The explosion epoch for SN 2018cni estimated in a similar manner is JD = 2458275.3. Luminosities are measured from the bolometric light curve, allowing variation within the uncertainty range of the bolometric estimations.

The main parameters of the best-fit model structures for SNe 2018cni and 2020kyg can be found in Table 6. The two objects fit in the general trend of other SNe Iax studied by abundance tomography, which displays increasing ejecta mass and rise time with luminosity. The constrained density functions (Figures 15 and 16) exhibit a relatively good agreement with the predictions of the pure deflagration models with corresponding luminosities.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 6. Parameters of the Best-fit Models of SNe 2018cni and 2020kyg from Our Spectral Synthesis Method

	SN 2018cni	SN 2020kyg
Δt [days]	16	10.8
vmax [km s−1]	7500	3100
Core density [g cm−3]	1.40	0.15
Density slope [km s−1]	2700	2800
Cutoff velocity [km s−1]	7000	⋯

Notes. Δt and v_max represent t_exp and v_phot at maximum light. Assuming an epoch of maximum light B_max = 2458289.99 and 2458998.94 JD for SNe 2018cni and 2020kyg, respectively.

Download table as:  ASCIITypeset image

Srivastav et al. (2022) fitted a different subset of the spectral time series of SN 2020kyg, following the same principles of the abundance tomography technique and the same radiative transfer code. The density functions show nearly the same profile with a slope of v₀ = 2800 km s⁻¹ (3000 km s⁻¹ in Srivastav et al. 2022), which is the key parameter in reproducing the observed P-Cygni features. The model of Srivastav et al. (2022) also has a slightly lower core density of ρ₀ = 0.03 g cm⁻³. As a significant difference, our model fitting constrained lower v_phot velocities over the epochs (assuming a linear decrease with time): for the only epoch included in both samples (${t}_{\exp }\,=\,6.5$ days), we derived v_phot = 3500 km s⁻¹ instead of 4200 km s⁻¹, which results in a hotter photosphere. The source of discrepancy is in the different characterization of the temperature profile, which also greatly affects the inferred chemical abundances. In the nearly uniform chemical structure of Srivastav et al. (2022), carbon is the dominant element (X(C) = 0.60) with significant contributions of O and minor abundances of Si, Fe, and ⁵⁶Ni. Note that such an extreme abundance of C is in conflict with the plausible explosion scenarios.

In this study, we aimed to constrain the abundance structure by adopting the general characteristics of the pure deflagration models (Fink et al. 2014; Lach et al. 2022) as a starting point for our fitting process. The final results are plotted in Figures 15 and 16. The outer layers of the models are dominated by C and O, suggesting that not the entire ejecta was reached by the fusion flame. At lower velocities, O is still the most abundant element in the entire regime of our models, while carbon is excluded from the inner regions (the lower limit is 10,000 km s⁻¹ for SN 2018cni and 4000 km s⁻¹ for SN 2020kyg), otherwise, the C ii lines would be extreme around the maximum light. These results strongly contradict the predictions of the pure deflagration scenario, but one should treat them with caution: O is also used as a puffer element to fill the unconstrained mass fractions; thus, X(O) is highly uncertain.

In general, the chemical distribution of the inner structure (i.e., below the carbon-populated layers) resembles the predictions of the pure deflagration models (Fink et al. 2014), as the elements forming the dominant spectral features (O, Si, S, Fe, and Ni) show nearly constant abundances. The total IGE mass fraction is lower than expected (X(IGE) ∼ 0.55 for N1-def and N3-def models), which is the consequence of the weakly constrained X(⁵⁶Ni). Despite the significantly lower densities of the outer regions adopted in our approach, Si, S, Fe, and Ni are either absent or present only with a significantly lower mass fraction. Magee et al. (2022) argued for strong mixing and uniform abundances, but their work was limited to the analysis of a few lines instead of fitting the whole optical wavelength range and studying a stratification model fixed to the radial coordinates. However, this study, just like previous fitting approaches, does not contradict the results of Magee et al. (2022), that the inner regions may inhibit a nearly uniform abundance structure.

Moreover, the complete lack of or very low abundances of IMEs and IGEs in the outer regions are relatively well constrained in our models by the blue wings of absorption features, and further supported by the low-density values at these higher velocities. This result indeed indicates some level of stratification; however, not necessarily a continuously changing multilayer structure. Instead, a C/O-dominated top layer with a deflagration-like well-mixed ejecta is also supported by the spectral synthesis.

This paper presents a combined analysis of a bright Type Iax SN 2018cni and a faint Type Iax SN 2020kyg. We attempted to comprehend the nature of the members of the Type Iax class, which are located at the two extremes in the luminosity space of this class. We see that in terms of photometric properties, spectral features, and ejecta configuration these two SNe behave differently. This also reflects that there might be a difference in the progenitor channels for both events. In the following, we summarize our findings for both events and try to make a collective picture of these two events as members of the same class with a gap in luminosity distribution.

SN 2018cni is a bright member of the Type Iax class with peak absolute magnitude M_V = −17.81 ± 0.21 mag. The estimated bolometric luminosity shows that SN 2018cni shares a similarity with SN 2005hk and is brighter than SN 2002cx. The mass of ⁵⁶Ni estimated from the analytical modeling of the BgVri bolometric light curve is ∼0.07 ± 0.01 M_⊙. The spectral features of SN 2018cni match well with SNe 2002cx and 2005hk.

With peak absolute magnitude, M_g = − 14.44 ± 0.16 mag, SN 2020kyg lies at the fainter end of Type Iax luminosity distribution. It is brighter than SNe 2008ha (Stritzinger et al. 2014), 2019gsc (Srivastav et al. 2020), and 2021fcg (Karambelkar et al. 2021) and fainter than SN 2010ae (Stritzinger et al. 2014). Analysis of the bolometric light curve (BgVri, obtained through Superbol) of SN 2020kyg indicates that ∼0.002 ± 0.001 M_⊙ of ⁵⁶Ni was produced in the explosion.

Magee et al. (2016) suggested a correlation between the absolute magnitude, decline rate parameter, and rise time of Type Iax SNe in the R/r-band. Since then the sample has grown significantly, especially in the fainter end. An attempt is made to check for any possible correlation in the extended sample. Table 7 includes all the Type Iax SNe we have used in our study and their references. In Figure 17, we have plotted the absolute magnitude and decline rate parameter in the r/R-band for the extended sample. We have also included the N1-def, N3-def, and N5-def models (Fink et al. 2014) for comparison. We recover the weak correlation between the absolute magnitude and light-curve decline rate parameter as reported in Magee et al. (2016) and shown in Figure 17. However, dividing the current sample based on the brightness of the objects reveals an interesting trend: bright and faint objects behave differently. For the bright Type Iax cluster (M_r ≤ −17.1 mag), there is a negative correlation with a correlation coefficient of 0.49 (p = 0.044), viz. the objects with a faster decline rate are found to be fainter. On the other hand, for the faint cluster (M_r ≥ −14.64 mag), a trend of positive correlation emerges (Pearson correlation coefficient of −0.85, p = 0.070), viz. objects with a faster decline rate appear brighter. The sample of Type Iax studied so far is limited and a study with a bigger sample covering both the bright and faint end of luminosity distribution will be required to confirm this trend. In Figure 18, we have presented a plot between the rise time and absolute magnitude for a sample of well-studied Type Iax SNe. The rise times of SNe 2018cni and 2020kyg plotted in this figure are estimated from the spectral fitting (Section 5.2). From the plot, it is clear that the two quantities are correlated (Pearson correlation coefficient = −0.53, p = 0.02), the brighter objects tend to have a longer rise time. Type Iax SNe with fair premaximum coverage are included in this analysis, while some unfiltered measurements are not considered.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The spectral features of SNe 2018cni and 2020kyg appear similar, except for the Si ii feature, which is prominent in SN 2020kyg. Modeling of the spectral sequence using TARDIS indicates that some level of stratification at the outer layers and mixed ejecta in the inner layers are required.

Comparison of the bolometric light curves obtained through Superbol shows a resemblance of SN 2018cni with the N3-def model (Fink et al. 2014), whereas the luminosity of SN 2020kyg falls in the range of pure deflagration of a CONe white dwarf. For brighter members of the Type Iax class, several explosion scenarios have been proposed, such as pulsational-delayed detonation (Ivanova et al. 1974; Khokhlov 1991a, 1991b; Khokhlov et al. 1993; Hoeflich et al. 1995; Hoeflich & Khokhlov 1996; Bravo & García-Senz 2006; Baron et al. 2012; Dessart et al. 2014), deflagration-to-detonation transition (Khokhlov 1991a, 1991c; Khokhlov et al. 1993; Hoeflich et al. 1995; Hoeflich & Khokhlov 1996; Höflich et al. 2002; Seitenzahl et al. 2013; Sim et al. 2013), pure deflagration of a white dwarf (Fink et al. 2014; Kromer et al. 2015), core-collapse origin (Foley et al. 2009, 2016; Valenti et al. 2009; Moriya et al. 2010), etc. Most of the observed properties of relatively bright Type Iax SNe match well with the predictions of pure deflagration of a white dwarf leaving behind a bound remnant. Observational properties of faint Type Iax can be explained by weak deflagration of a CONe white dwarf (Kromer et al. 2015), but the spectrophotometric evolution of these models is faster as compared to the observed ones. Another possible explanation is a double degenerate channel, which involves the merger of CO and ONe white dwarfs. This merger results in a failed detonation and is associated with a fast-declining transient (Kashyap et al. 2018) fainter than faint Type Iax SNe.

In any case, incomplete burning of the white dwarf explains most of the properties of Type Iax SNe, but the possibility for another progenitor scenario for faint members of this class cannot be ruled out completely. The different nature of the correlation between peak luminosity and light-curve decline rate for bright and faint members indicates heterogeneity within SNe Iax. Further study and in-depth analysis of this class will be required to understand their observed properties, progenitor system, and explosion mechanism.

We thank the anonymous referee for providing useful comments toward the improvement of the manuscript. We acknowledge Wiezmann Interactive Supernova data REPository http://wiserep.weizmann.ac.il (WISeREP; Yaron & Gal-Yam 2012). This research has made use of the CfA Supernova Archive, which is funded in part by the National Science Foundation through grant AST 0907903. This research has made use of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. This work makes use of the Las Cumbres Observatory global telescope network. The LCO group is supported by NSF grant AST-1911225. R.D. acknowledges funds by ANID grant FONDECYT Postdoctorado No. 3220449. D.A.H., G.H., and C.M. were supported by NSF Grants AST-1313484 and AST-1911225.

This project has been supported by the János Bolyai Research Scholarship of the Hungarian Academy of Sciences (T.S.), as well as by the FK134432 grant of the National Research, Development and Innovation Office of Hungary and the ÚNKP 22-5 New National Excellence Programs of the Ministry for Culture and Innovation from the source of the National Research, Development and Innovation Fund (B.B. and T.S.). This research made use of TARDIS, a community-developed software package for spectral synthesis in supernovae (Kerzendorf et al. 2018, 2019). 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. This work made use of the Heidelberg Supernova Model Archive (HESMA). ²⁷ We are grateful to the Southern African Large Telescope (SALT) astronomers and support staff who obtained the SALT data presented here. The SALT observations of SN 2018cni were obtained with Rutgers University program 2018-1-MLT-006 (PI: Jha). The Gemini Observatory data for SN 2020kyg were obtained through program GN-2020A-Q-224 (PI: Jha). Observations reported here were obtained at the MMT Observatory, a joint facility of the University of Arizona and the Smithsonian Institution. We would like to express our gratitude to the Lick Observatory staff for their support. We gratefully acknowledge the usage of native lands for our science. Shane 3 m observations were conducted on the stolen land of the Ohlone (Costanoans), Tamyen, and Muwekma Ohlone tribes. A major upgrade of the Kast spectrograph on the Shane 3 m telescope at Lick Observatory was made possible through generous gifts from William and Marina Kast as well as the Heising-Simons Foundation. Research at Lick Observatory is partially supported by a generous gift from Google.

Software: TARDIS (Kerzendorf et al. 2018, 2019), PYSALT (Crawford et al. 2010), SuperBol (Nicholl 2018), lcogtsnpipe (Valenti et al. 2016), PyZOGY (Guevel & Hosseinzadeh 2017), floydsspec pipeline (https://www.authorea.com/users/598/articles/6566), spectral pipeline (https://github.com/msiebert1/UCSC).