Early and Extensive Ultraviolet Through Near Infrared Observations of the Intermediate-Luminosity Type Iax Supernovae 2024pxl

Introduction

Type Ia supernovae (SNe Ia) are a large class of transients resulting from the explosion of carbon-oxygen white dwarf (CO-WD) stars [47]. A non-negligible fraction (5-30% of the SN Ia rate; [131, 113, 23] (W. Li et al. 2011)) of these events are the peculiar Type Iax supernovae (SNe Iax; see [32, 52]), with SN 2002cx (W. Li et al. 2003) serving as the prototypical member of the class.¹ These explosions are generally less luminous and show lower ejecta velocities than normal SNe Ia.

A favored progenitor scenario for SNe Iax is the single-degenerate scenario with one WD and one non-degenerate companion ([130]). Unlike for normal SNe Ia, this scenario has robust observational support based on the progenitor detection of

Corresponding author: W. B. Hoogendam

Email: [email protected]

• NSF Graduate Research Fellow

† CCAPP Fellow

‡ FINESST FI SN 2012Z C. McCully et al. [84]. Furthermore, two SNe Iax, SNe 2004cs and 2007J, show strong He I emission at $\sim$+40 days after maximum light (something not seen in normal SNe Ia M. A. Tucker et al. [121]; R. J. Foley et al. [32], although see also C. J. White et al. [131] and R. J. Foley et al. [31]), which has been argued to support a non-degenerate He-star companion R. J. Foley et al. [32]. SNe Iax are predominantly found in late-type host galaxies J. D. Lyman et al. [75], whereas normal SNe Ia occur in both late- and early-type galaxies L. Galbany et al. [35]. Late-type galaxies may preferentially form proposed progenitor systems necessary for SNe Iax, such as binary systems composed of a CO WD accreting from a He star companion L. Piersanti et al. [93]. These systems are favored in part due to pre- and post-explosion Hubble Space Telescope (HST) imaging of SN 2012Z that reveals a blue progenitor C. McCully et al. [84] and a late-time flux excess which may be consistent with a bound remnant, late-time circumstellar medium interaction, or a peculiar late-time decay C. McCully et al. [85]; M. Schwab et al. [104].

Many studies favor the pure deflagration of a CO WD to explain SNe Iax (e.g., D. Branch et al. [9]; M. M. Phillips et al. [91]; D. K. Sahu et al. [101]; G. C. Jordan et al. [56]; M. Kromer et al. [62]; M. Fink et al. [29]; M. Bulla et al. [12]; S.-C. Leung & K. Nomoto [69]; Y. Zeng et al. [135]; M. R. Magee et al. [77]; F. Lach et al. [66]; C. Feldman et al. [28]; K. Maguire et al. [81]), where the flame front propagates as a deflagration wave. This deflagration front is subsonic and produces large-scale instabilities that mix elemental distributions within the ejecta (e.g., V. N. Gamezo et al. [36]). Alternatively, pulsational delayed detonation models (e.g., L. N. Ivanova et al. [51]; P. Hoeflich et al. [42]; E. Baron et al. [7]; M. D. Stritzinger et al. [116]) have been invoked to explain brighter SNe Iax, in particular SN 2012Z. In this explosion mechanism, the CO WD pulsates once before thermonuclear runaway. The flame front stars as a deflagration and then transitions into a supersonic detonation. The primary difference between these two models lies in the ejecta stratification, with deflagration explosions producing fully-mixed ejecta and pulsational delayed detonations having a layered outer-most ejecta.

Photometrically, SNe Iax are bluer in the ultraviolet than normal SNe Ia before maximum brightness and then redder than normal SNe Ia $\sim$10 days after maximum brightness P. A. Milne et al. [87]. They lack the secondary peaks in their $r$-, $i$-, and near-infrared (NIR) light curves commonly seen in normal SNe Ia M. M. Phillips et al. [91]; S. González-Gaitán et al. [38]; C. Ashall et al. [3]. Spectroscopically, the ejecta velocities are slower than normal SNe Ia and range from 2000 km s$^{-1}$ to 7000 km s$^{-1}$ (e.g., M. Stritzinger [114]; R. J. Foley et al. [32]; S. W. Jha [52], compared to $\sim$10000 km s$^{-1}$ at peak light for normal SNe Ia; N. Morrell et al. [88]), and signatures of C are common in SNe Iax, indicating some primordial CO WD material remains unburnt R. J. Foley et al. [32].

SNe Iax are generally featureless in their early NIR spectra (e.g., M. D. Stritzinger et al. [116]; K. Maguire et al. [81]), unlike normal SNe Ia (e.g., E. Y. Hsiao et al. [48, 49, 50]; W. B. Hoogendam et al. [46]) and as early as $\sim$18 days after maximum light the slow ejecta velocities enable the clear resolution of $H$-band Co II features (e.g., SNe 2010ae, 2012Z, and 2014ck M. D. Stritzinger et al. [115, 116]; L. Tomasella et al. [118]) that are heavily blended in normal SNe Ia. Comprehensive NIR observations of SNe Iax are rare; only SN 2012Z has NIR spectral coverage spanning a week before until a month after maximum light. Unlike normal SNe Ia, which become “nebular” with spectra showing solely forbidden lines, SNe Iax never reach this stage, and even a year past maximum show permitted Fe II, Na I D and Ca II NIR triplet lines, indicating a dense, slowly expanding emission region long after the SN explosion which can be naturally explained if some material stays gravitationally bound to the remnant WD S. W. Jha et al. [53]; D. K. Sahu et al. [101]; R. J. Foley et al. [30, 31]; Y. Camacho-Neves et al. [15].

In this manuscript, we present extensive UV through NIR observations of the nearby SN Iax 2024pxl from shortly after discovery (the earliest optical spectrum is at $-10.1$ days and the earliest NIR spectrum at $-9.1$ days, the earliest ever for a SN Iax) to a month and a half after peak light when it entered a seasonal break. Recent preprints by M. Singh et al. [108] and L. A. Kwok et al. [65] present complementary data of SN 2024pxl to those in our work. § 2 presents our observations and data reductions, § 3 and § 4 present our photometric and spectroscopic analyses, respectively. § 5 provides a summary and concluding remarks. Throughout this work, we adopt $H_0 = 72$ km s$^{-1}$ Mpc$^{-1}$, $\Omega_m = 0.30$, and $\Omega_{\mathrm{vac}} = 0.70$ for this work.

Data

The Zwicky Transient Facility (ZTF; E. C. Bellm et al. [8]) discovered SN 2024pxl on UT 2024-07-23 at 06:34:48 (MJD=60514.3) with a $r$-band magnitude of 17.7 and a last non-detection of $r > 19.4$ mag two days prior on UT 2024-07-21 06:04:13 (MJD 60512.3; N. Rehemtulla et al. [97]). It was classified on UT 2024-07-25 20:10:13 ($\sim$4.5 days later) as a SN Iax based on a spectrum from the Wide Field Spectrograph (WiFeS; M. Dopita et al. [24, 25]) on the Australian National University (ANU) 2.3m telescope (S. Smartt et al. [110]).

MJD	Filter	$m$	$\sigma_m$	Source
60524.32	Y	15.36	0.01	UKIRT
60561.27	Y	15.56	0.01	UKIRT
$\vdots$	$\vdots$	$\vdots$	$\vdots$	$\vdots$
60608.21	Y	16.89	0.01	UKIRT

Photometric Data

We collected optical survey photometry of SN 2024pxl from the ZTF, the Asteroid Terrestrial-impact Last Alert System (ATLAS; J. L. Tonry et al. [120]), and All-Sky Automated Search for SuperNovae (ASAS-SN B. J. Shappee et al. [106]; C. S. Kochanek et al. [60]; K. Hart et al. [39]) surveys. We obtained the ZTF data using the ZTF forced photometry service (F. J. Masci et al. [83]).² We retrieved the ATLAS data from the ATLAS Transient Science Server (K. W. Smith et al. [111]). ATLAS data taken on the same night were stacked using a weighted average. ATLAS only observed SN 2024pxl using the $o$ band. Finally, we obtained ASAS-SN photometry from the SkyPatrol service (C. S. Kochanek et al. [60]; K. Hart et al. [39]).³

We also obtained $ZYJHK$ photometry with the UKIRT 3.8m telescope using the Wide Field Camera (WFCAM; M. Casali et al. [17]; S. T. Hodgkin et al. [41]). We processed these data with Photpipe (A. Rest et al. [98]) and calibrated them using The Two Micron All Sky Survey (2MASS; M. F. Skrutskie et al. [109]) following S. T. Hodgkin et al. [41] and E. R. Peterson et al. (2023). We used unforced, non-difference-imaged photometry from DOPHOT (P. L. Schechter et al. [102]). The background light is negligible compared to the flux from SN 2024pxl at the phases of our photometry, so the lack of a template does not significantly affect our photometry.

Finally, we obtained UV photometry from the Neil Gehrels Swift Observatory (Swift; N. Gehrels et al. [37]) using Swift’s Ultraviolet/Optical Telescope (UVOT; P. W. A. Roming et al. [100]). We used HEASOFT version 6.31.1 to combine the images with the UVOTIMSUM package and to obtain aperture photometry with the UVOTSOURCE package. An aperture of $5^{\prime\prime}$ is used for the source, and a $15^{\prime\prime}$ region centered on $(\alpha,\delta)=(17^\mathrm{h}32^\mathrm{m}28.47^\mathrm{s},+07^\circ05^\prime27.68^{\prime\prime})$ is used for the background. We report AB magnitude system using the A. A. Breeveld et al. [10] zero points. The UVW2 and

Spectroscopic Data

Fourteen optical spectra were taken on the University of Hawaii (UH) 2.2m telescope using the SuperNova Integral Field Spectrograph (SNIFS; B. I. Lantz et al. [67]) as part of the Spectroscopic Classification of Astrophysical Transients (SCAT; M. A. Tucker et al. [122]) Survey. The spectra were reduced with the SCAT pipeline (M. A. Tucker et al. [122]). We also include the classification spectrum, which we re-reduced with pWyFes (M. J. Childress et al. [19]). Table 2 compiles the optical spectroscopic observations from $-10.1$ to $+45.7$ days from $g$-band maximum.

UT Date	MJD	Phase [days]	Telescope	Spectrograph
2024-07-23	60515.4	$-10.1$	ANU 2.3m	WiFeS
2024-07-26	60517.3	$-8.2$	UH 2.2m	SNIFS
2024-07-28	60519.4	$-6.1$	UH 2.2m	SNIFS
2024-07-29	60520.4	$-5.1$	UH 2.2m	SNIFS
2024-07-30	60521.4	$-4.1$	UH 2.2m	SNIFS
2024-07-31	60522.4	$-3.1$	UH 2.2m	SNIFS
2024-08-01	60523.3	$-2.2$	UH 2.2m	SNIFS
2024-08-08	60530.3	$+4.8$	UH 2.2m	SNIFS
2024-08-09	60531.3	$+5.8$	UH 2.2m	SNIFS
2024-08-12	60534.3	$+8.8$	UH 2.2m	SNIFS
2024-08-13	60535.3	$+9.8$	UH 2.2m	SNIFS
2024-08-16	60538.3	$12.8$	UH 2.2m	SNIFS
2024-08-18	60540.3	$+14.8$	UH 2.2m	SNIFS
2024-08-21	60543.3	$+17.8$	UH 2.2m	SNIFS
2024-08-22	60544.3	$+18.8$	UH 2.2m	SNIFS
NIR
2024-07-25	60516.4	$-9.1$	Gemini-N	GNIRS
2024-08-01	60523.3	$-2.2$	IRTF	SpeX
2024-08-01	60523.4	$-2.1$	Gemini-N	GNIRS
2024-08-14	60536.4	$+10.9$	IRTF	SpeX
2024-09-05	60558.3	$+32.8$	IRTF	SpeX
2024-09-10	60563.2	$+37.7$	IRTF	SpeX
2024-09-11	60564.1	$+38.6$	Gemini-N	GNIRS
2024-09-14	60567.2	$+41.7$	IRTF	SpeX
2024-09-15	60568.2	$+42.7$	Gemini-N	GNIRS
2024-09-18	60571.2	$+45.7$	IRTF	SpeX

Of the total ten NIR spectra in our dataset, four were taken on the Gemini North 8.1m telescope using the Gemini Near-InfraRed Spectrograph (GNIRS; J. H. Elias et al. [26]) and six on the 3m NASA InfraRed Telescope Facility (IRTF) with SpeX (J. T. Rayner et al. [96]). The GNIRS and SpeX data were reduced with telluric corrections from an A0V star using the standard PypeIt (J. Prochaska et al. [95]) and Spextool (M. C. Cushing et al. [20]) procedures, respectively. Table 2 catalogs our NIR spectroscopic observations, which include the earliest ever NIR spectrum of a SN Ia at $\sim$9 days before $g$-band peak.

Host Galaxy

SN 2024pxl exploded in NGC 6384, a spiral galaxy at $z=0.005554$ (M. P. Haynes et al. [40]). NGC 6384 has a Milky Way line-of-sight extinction of $E(B-V)=0.106$ mag (E. F. Schlafly & D. P. Finkbeiner [103]). We use the Blast tool (D. O. Jones et al. [55]) ⁴ to estimate the properties of NGC 6384 using archival host-galaxy photometry from the Galaxy Evolution Explorer (D. C. Martin et al. [82]), 2MASS (M. F. Skrutskie et al. [109]), Wide-field Infrared Survey Explorer (E. L. Wright et al. [132]), Sloan Digital Sky Survey (M. Fukugita et al. [34]; D. G. York et al. [134]), and Panoramic Survey Telescope and Rapid Response System (K. C. Chambers et al. [18]) using PSF-matched elliptical apertures. Host-galaxy spectral energy distribution (SED) parameters are estimated using Prospector (J. Leja et al. [68]; B. D. Johnson et al. [54]; B. Wang et al. [128]). For NGC 6384, we find the following global parameters: a mass of $\log_{10}(M/M_{\odot}) = 10.57^{+0.06}_{-0.09}$, a star-formation rate (SFR) of $\log_{10}(\mathrm{SFR}/\mathrm{yr}/M_{\odot}) = -0.69^{+0.42}_{-0.48}$, a specific star-formation rate (sSFR) of $\log_{10}(\mathrm{sSFR}/\mathrm{yr}) = -11.25^{+0.44}_{-0.52}$, and a mass-weighted mean stellar age of $12.46^{+0.33}_{-3.57}$ Gyr. The galaxy age may be dominated by the bulge light, as this age is rather old for a star-forming galaxy like NGC 6384, and the local age in the area of SN 2024pxl is $\sim2$ Gyr younger than the global age, suggesting that the older bulge likely skews the global age.

SN 2024pxl's Siblings: SNe 1971L and 2017drh

In this section, we use two previous SNe Ia, 1971L and 2017drh, to estimate the distance to NGC 6384. A growing number of SN Ia share a host galaxy with legacy SNe Ia and are referred to as “sibling” SNe Ia (e.g., SNe 1997cn and 2015bo; M. Turatto et al. [124]; W. B. Hoogendam et al. [44]; see C. R. Burns et al. 2020; L. Kelsey [59] for compilations of other SNe Ia siblings).

SN 1971L was discovered by W. Logan [73] on 1971-06-24 with subsequent IAU Circulars from K. Locher [72] and F. Seiler [105], and photometric observations from R. Barbon et al. [5] and V. M. Lyutyj [76]. We omit reporting fit results for SN 1971L; the observations were taken from multiple sites using photometric plates, and there is a large, systematic scatter in the plate photometry that makes SN 1971L unreliable for a precise distance estimate.

SN 2017drh was discovered by S. Valenti et al. [125] on 2017-05-03. We perform photometry on public images of SN 2017drh from the Las Cumbres Observatory Global Telescope (LCOGT) with photpipe on the pixel-level calibrated LCOGT data. We create difference images for photpipe using the latest observed epoch without SN Ia flux as the template image. We use Super-Novae in object-oriented Python (SNooPy; C. R. Burns et al. [13, 14]) to fit SN 2017drh (with EBV_model2) to estimate its properties.

For SN 2017drh, we find a distance modulus of $\mu= 32.31 \pm0.32$ mag, light curve shape $s_{BV} = 0.89 \pm0.03$ mag, and host-galaxy extinction of $E(B-V)_{\mathrm{host}} = 1.19 \pm0.06$ mag. This distance is consistent with the Tully-Fisher distance of $\mu= 32.00 \pm0.45$ (R. B. Tully et al. [123]). We adopt the distance from SN 2017drh ($\mu= 32.31 \pm0.32$ mag; $d = 28.95 \pm5.14$ Mpc) for this work.

Host-Galaxy Extinction

In addition to the Milky Way extinction, we estimate the host-galaxy extinction from the Na I D feature using the relationship between $E(B-V)_{\mathrm{host}}$ from D. Poznanski et al. [94] with uncertainties calculated using the method from M. Phillips et al. [92]. We find $E(B-V)_{\mathrm{host}} = 0.19 \pm0.03$ mag, which we adopt for the rest of this work. Assuming $R_V = 3.1$ and using the J. A. Cardelli et al. [16] extinction law, this corresponds to $A_g = 0.36$ mag and $A_r = 0.25$ mag.

Photometric Analysis

Light Curves

Figure 1 shows our photometric observations. We highlight our extensive NIR photometric coverage: SN 2024pxl is the third SN Iax with a NIR light curve with high cadence over a 3-month timescale after SNe 2005hk (M. M. Phillips et al. [91]) and 2012Z (M. Yamanaka et al. [133]). We find a $g$-band peak of $m_g = 16.05 \pm0.01$ mag on MJD 60525.5 $\pm0.2$ and an $r$-band peak of $m_r = 15.57 \pm0.01$ mag on MJD 60529.2 $\pm0.2$ using a Gaussian Process fit as implemented in SNooPy. M. Singh et al. [107] propose that SN Iax form two groups based on their absolute magnitude and decline rates [see also 79], with the groups consisting of more luminous SNe Iax ($-17 \gtrsim M_r \gtrsim-19$ mag) and low-luminosity SNe Iax ($M_r \gtrsim-15.5$ mag). Figure 2 presents the $\Delta m_r(15)$ vs $M_r/R$ phase space with normal SNe Ia from the Carnegie Supernova Project [61] and literature SNe Iax. SN 2024pxl is on the lower edge of the luminous group with $\Delta m_r(15) = 0.69 \pm0.05$ mag and $M_r = -16.99 \pm0.32$ mag.

Quantity	Value	Units	Description
$t_0$	$60511.80^{+0.16}_{-0.18}$	MJD	Time of first light
$n_{o1}$	$1.32^{+0.06}_{-0.05}$		$o$-band rise slope
$n_{o2}$	$-0.04^{+0.01}_{-0.01}$		$o$-band decline slope
$A_o$	$1.62^{+0.02}_{-0.02}$	mJy	$o$-band constant
$n_{g1}$	$1.50^{+0.08}_{-0.07}$		$g$-band rise slope
$n_{g2}$	$-0.06^{+0.01}_{-0.01}$		$g$-band decline slope
$A_g$	$1.21^{+0.01}_{-0.02}$	mJy	$g$-band constant
$n_{r1}$	$1.30^{+0.06}_{-0.05}$		$r$-band rise slope
$n_{r2}$	$-0.04^{+0.01}_{-0.01}$		$r$-band decline slope
$A_r$	$1.67^{+0.02}_{-0.02}$	mJy	$r$-band constant

Rising Light Curve Fits

To determine the power-law slope in a given filter for SN 2024pxl (assuming the same time of first light in all bands), we use a power-law model defined as

to fit the ATLAS and ZTF forced photometry, where

The value of $\alpha$ in a one-component power law (i.e., $\alpha= a$) is sensitive to the proximity of peak light the fitting data extends (e.g., P. J. Vallely et al. [127, 126]), the $\alpha_1$ parameter from the two-component model is not as sensitive to what percentage of the peak flux that is included in the fit range [126].

We use a flexible power law to avoid this issue, fitting through 5 days after $g$-band maximum light. We also define $\Lambda$ as

This scale factor breaks the degeneracy between the multiplicative scale factor and power-law slope [86]. We use the emcee fitting package [33] as described in W. B. Hoogendam et al. [45] with the results shown in Figure 3. The fitted time of first light is MJD 60511.8 $\pm0.2$, 2.6 $\pm0.2$ days before discovery, 3.7(4.7) $\pm0.2$ days before our first optical (NIR) spectrum, and 13.8 $\pm0.2$ days before $g$-band maximum.

SN 2024pxl is best described as a smoothly rising SN Ia with no excess flux and is a “single” rising SN Ia in the W. B. Hoogendam et al. [43] morphological scheme. The smooth rise also rules out significant excess flux from interaction with a companion or the circumstellar medium. The $g$- and $r$-band rise slopes for SN 2024pxl are slightly steeper (1.72 $\pm0.07$ and 1.34 $\pm0.05$, respectively) than SN 2020udy (1.38 $\pm0.11$, 1.29 $\pm0.07$, respectively; K. Maguire et al. [81]). The primary power-law parameters $n_{g1}$ and $n_{r1}$ have been shown to be generally consistent with single-power-law fits [126]. To better compare SNe 2020udy and 2024pxl, we fit using a similar single-power-law method as K. Maguire et al. [81]. We find $g$- and $r$-band exponents of $\sim1.12$ and $\sim1.24$, respectively. The $g$-band changes significantly, whereas the $r$-band does not; this is most likely an effect of inflexible power laws depending on the amount of rise that is fit and the sparser sampling in the $g$ band. In general, SN 2024pxl is similar to SN 2020udy and consistent with the recent sample study by M. R. Magee et al. [80], which finds that SNe Iax have shallower rising light curve slopes than normal SNe Ia.

Color Curves

Figure 4 shows the $g-r$ color curve of SN 2024pxl compared to three SNe Iax (SNe 2005hk; M. M. Phillips et al. [91, 90]; M. D. Stritzinger et al. [116], and SN 2020udy; K. Maguire et al. [81]) and normal SNe Ia from the Carnegie Supernova Project [61]. We highlight these three since they are well-observed, SNe 2005hk and 2012Z constitute our NIR colors comparison sample (see below), and SN 2020udy also has early-time $g-r$ colors. SN 2024pxl and the other SNe Iax become redder even at early times. The general redward trend with no blueward evolution could be caused by SNe Iax having a well-mixed ejecta with ${}^{56}\mathrm{Ni}$ out to the surface layers, rather than a centrally concentrated ${}^{56}\mathrm{Ni}$ distribution. Such an elemental distribution alters the energy distribution in the ejecta. This reduces the amount of stored energy in the central regions relative to normal SNe Ia, and thus makes adiabatic cooling more dominant at early times. At the earliest epochs, SN 2024pxl and SN 2020udy have a “knee” in the color curve at $\sim$10 days before maximum light, becoming $\sim$0.20 mag redder in a day. SN 2024pxl matches the other two SNe Iax near maximum light.

To the best of our knowledge, we present the first analysis of the $J-H$ color curves of SNe Iax. Only SNe 2005hk ([91]) and 2012Z ([133]) have sufficient NIR data for $J-H$ color curves to compare to SN 2024pxl. SN 2024pxl exhibits a similar evolution for the first two weeks after maximum light; however, during the $H$-band break region, it is redder than SN 2005hk and 2012Z. In the $g-J$ color curve, SNe Iax start with $g-J \approx0$ mag, but they never become bluer, unlike normal SNe Ia, which become significantly bluer as part of their evolution. Instead, SN 2024pxl and the other SNe Iax become redder after maximum light in $g-J$, reaching their red maximum $\sim20$ days later. They slowly become bluer over time but remain several magnitudes redder than normal SNe Ia.

time but remain several magnitudes redder than normal SNe Ia. SN 2024pxl is $\sim$1 mag redder than both SNe 2005hk and 2012Z in $g-J$ before an extinction correction, but after, it has a similar color, supporting our extinction estimate.

Compared to normal SNe Ia, SNe Iax are on the red edge of the normal SNe Ia $g-r$ colors throughout their evolution. SNe Iax remain highly ionized for an extended period due to a combination of low ejecta velocities and ${}^{56}\mathrm{Ni}$ mixed throughout the ejecta. This is also the reason SNe Iax lack the $r$-band secondary peak that is common in normal SNe Ia and thought to be from recombination from Fe iii to Fe ii ([e.g. 58]). The $J-H$ colors of normal SNe Ia have a blue peak at $\sim$20 days after maximum light, corresponding to the phase of the $H$-band break in the NIR spectra ([129, 1 b]), but SN 2024pxl and the other SNe Iax have a linear, blue $J-H$ evolution starting at $\sim$+10 days. The stronger $H$-band break in normal SNe Ia may be because they generally produce more ${}^{56}\mathrm{Ni}$ than SNe Iax, or that the velocity distribution of ${}^{56}\mathrm{Ni}$ differs between objects. By ~+30 days, normal SNe Ia and SN 2024pxl and the other SNe Iax have similar $J-H$ colors again, but by ~+75 days, normal SNe Ia are redder than SN 2024pxl and the other SNe Iax by nearly a magnitude.

Finally, we also present a novel analysis of the $g-J$ colors for SNe Iax compared to a representative sample of four SN Ia from Figure 1 of M. M. Phillips [90], arranged by $\Delta m_{15}(B)$: SNe 2008hv, 2006ax, 2007on, and 2006mr). We limit our sample of normal SNe Ia to these four to ensure the general trend within normal objects is visible and not obscured by too many objects. Normal SNe Ia become bluer as they approach maximum light, with the fastest decliners reaching their bluest point near maximum light. There is no blueward $g-J$ evolution until ~+20 days after maximum light. SNe Iax may be a more homogeneous class with less intra-object variance than normal SNe Ia. Our sample is small and may be biased, but SNe 2005hk, 2012Z, and 2024pxl span a range of luminosities and decline rates, suggesting that the sample is potentially representative. A larger sample of SNe Ia and SNe Iax with optical and NIR photometry (e.g., combining data from ground-based surveys and the upcoming Nancy Grace Roman Space Telescope D. Spergel et al. [112]) would help improve this first-order explanation of the $g-J$ colors of SNe Ia.

Spectroscopic Analysis

We obtained fifteen optical spectra between −10.1 and +18.8 days and ten NIR spectra between −9.1 and +45.7 days relative to the $g$-band maximum.

Optical Spectra

Figure 5 shows the optical spectroscopic time series for SN 2024pxl, which has typical SN Iax features at early times, with the strongest being Fe III $\lambda\lambda$4404, 5129. S II $\lambda\lambda$5454, 5640 and Si II $\lambda$6355 emerge near maximum light along with a strong feature in the area of C II $\lambda$6580, which is not uncommon for SNe Iax (R. J. Foley et al. [32]) and often interpreted as C (K. Maguire et al. [81]). There is a feature near C II $\lambda$7234, similar to SN 2020udy with a tentative C II $\lambda$7234 feature identified (K. Maguire et al. [81]). However, this feature has a higher velocity than C II $\lambda$6580. Alternatively, T. Szalai et al. [117] argue that this ~7200 Å feature is a blend of Fe II in SNe Iax, which is more plausible given the velocity discrepancy with C II $\lambda$6580.

Figure 6 compares the maximum light optical spectrum of SN 2024pxl with other SNe Iax: SN 2005hk (M. M. Phillips et al. [91]), SN 2012Z (M. D. Stritzinger et al. [116]), SN 2019gsc (L. Tomasella et al. [119]), SN 2019muj (B. Barna et al. [6]), and SN 2020udy (K. Maguire et al. [81]). The objects are arranged by absolute magnitude, with SN 2012Z being the most luminous and SN 2019gsc being the least luminous. The less luminous objects have weaker Fe III $\lambda$5129 and stronger C II $\lambda$6580, and the more luminous SNe 2005hk and 2020udy have stronger Fe III $\lambda$5129 and weaker

C II $\lambda$6580. SN 2012Z does not follow this prescription exactly – it has strong Fe III and strong C II $\lambda$6580. This trend may suggest that even within the SNe Iax class, there is a burning efficiency difference between more luminous objects having more burning (weaker C II $\lambda$6580) and lower luminosity objects having less burning (more C II $\lambda$6580). To confirm this, a systematic abundance tomography study of SNe Iax is needed, which is beyond the scope of this work.

Figure 7 shows the velocity evolution of Fe III $\lambda$4404, Si II $\lambda$6355, and C II $\lambda$6580 for SN 2024pxl compared to the Si II $\lambda$6355 velocities of other SNe Iax. These three features trace different burning regions: C is a pure tracer of primordial unburnt material, Si traces

C+O burning, and Fe traces Si burning into ${}^{56}\mathrm{Ni}$ (which decays into Fe). The similar velocity evolutions for the three features in SN 2024pxl suggest that the ejecta is efficiently mixed, containing unburnt, partially burnt, and fully burnt material in the same layers. This is a typical sign of deflagration burning. The first velocity point is highly uncertain due to the shallow line profiles, which may underestimate the velocities. The feature velocities for SN 2024pxl are lower than both normal SNe Ia and bright SNe Iax, such as SNe 2005hk and 2012Z [91, 116], but similar to another SN Iax: SN 2010ae [115].

NIR Spectra

Figure 8 shows our NIR spectral time series data. The early-time NIR spectra show a featureless continuum, similar to the other two SNe Iax with spectra at a similar, but slightly later, phase [116, 81]. Near maximum light, features start to emerge, including intermediate mass elements (Mg II 0.9227, 1.0092, 1.0927 $\mu$m) and iron group elements (Fe II 0.9998, 1.0503, 1.0863, 1.1126 $\mu$m, and Co II 1.5759, 1.6064, and 1.6361 $\mu$m). These features increase in strength at $\sim30$ days after maximum light. Due to the low ejecta velocities, each feature is resolved rather than blended into a single feature as in normal SNe Ia.

The NIR regime covers strong features for two important elements: He and C. He is highly sensitive to the progenitor system, with He arising from an outer He shell or nondegenerate He-star companion. C is a pure tracer of unburnt material. Our NIR spectra do not show conclusive He absorption features for both He I 1.083 $\mu$m and He I 2.058 $\mu$m. This is consistent with SNe 2005hk, 2012Z, 2015H, and 2020udy, which do not have He signatures [78, 81] but stands in contrast to reported He detections in SNe 2004cs and 2007J [32], although it is unclear if these objects are bonafide SNe Iax, see C. J. White et al. [131]; R. J. Foley et al. [31]).

The NIR C I 1.0693 $\mu$m feature, the most common NIR C feature in SNe Ia and one of the strongest, is also absent in our NIR spectra. Even well after maximum light, no C I features are visible, similar to SN 2020udy [81]. This suggests the ejecta remains hot enough to ionize C, producing C II rather than C I. The persistence of C II $\lambda$6580 through maximum light in our optical spectra indicates a well-mixed ejecta with unburnt, singly ionized C throughout.

Finally, we may see an emerging [Ni II] 1.939 $\mu$m feature, attributed to stable ${}^{58}\mathrm{Ni}$ since all the synthesized ${}^{58}\mathrm{Ni}$ has radioactively decayed by this phase (e.g., S. Kumar et al. [63]). Previous work by R. J. Foley et al. [31] identified a [Ni II] feature at $\sim200$ days after the maximum light in the optical spectra of SNe 2008A and 2008ge. Generally, the presence of stable Ni is taken as an indicator of high-density burning (e.g., [64, 21, 22, 4]). Further nebular-phase observations will confirm if SN 2024pxl has [Ni II] 1.939 $\mu$m, similar to other SNe Iax.

NIR Spectral Comparison of SNe Iax

NIR spectra of SNe Iax, and in particular spectra over a week before maximum light, are rarer. We compare SN 2024pxl to other SNe Iax with NIR spectra in Figure 9. Our comparison objects are SN 2005hk [62], SN 2012Z [116], SN 2019muj ([6]), and SN 2020udy ([81]).

At early times, roughly a week before maximum light, SN 2024pxl and the other SNe Iax with early-time NIR spectra are similarly featureless. At maximum light, weak spectral features emerge, but the explosion remains hot, as the slow expansion velocities limit the efficacy of adiabatic expansion to cool the ejecta. It is not until nearly a month later that Co II features dominate the $H$-(see § 4.4) and $K$-bands.

The H-band Break in SNe Iax

The 20-30 day NIR Fe and Co emission in the $H$-band in SNe Ia is denoted as the “H-band break” ([129, 1 b]). These features come from $^{56}$Ni decay products (Co and Fe) above the photosphere and are usually strongly blended in normal SNe Ia ([129]). However, in SNe Iax, the low velocities minimize blending, allowing individual features to be resolved. These features trace the outer distribution of $^{56}$Ni, providing insights into the high-density burning and energetics of the explosion.

Figure 10 shows the $\sim$30 day spectra for SNe Iax and two normal SNe Ia for comparison. The objects are arranged by peak magnitude. There is a correlation between peak magnitude and Doppler broadening of the NIR Co II features in SNe Iax, with more luminous SNe Iax having higher velocities and generally broader line profiles ([115]). In addition to this known effect, we tentatively find a new one: the more luminous SNe Iax have more blueshifted Co II line profiles. It is not possible to detect this in the sample used by M. D. Stritzinger et al. [115] because their spectral sample is not large enough in the 20 to 30 day phase range. Comparing the blueshifting requires similar-epoch spectra such as those we compile and present in this work (ranging from $\sim$+22 to $\sim$+33 days).

To obtain velocity estimates, we remove the pseudo-continuum using a linear fit of the edges of the emission feature and plot the pseudo-continuum subtracted emission features in the bottom panels of Figure 10. We do this for three Co II lines: Co II 1.5759, 1.6064, and 1.6361 $\mu$m. SN 2012Z, a luminous SN Iax, has a blueshift in the emission of $\sim$3000 km s$^{-1}$, where as the slightly less luminous SN 2005hk has a shift only of $\sim$1000 km s$^{-1}$. The least luminous SNe Iax, 2014ck, 2019muj and 2024pxl, are near 0 km s$^{-1}$.

To demonstrate this effect better, we measure the peak Doppler velocity of the three Co II peaks using a Gaussian fit to the pseudo-continuum-subtracted feature profile and calculate the median velocity. We adopt the standard deviation of the three peak velocity measurements as the uncertainty. Our results are shown in Figure 11. The more luminous SNe Iax (SNe 2005hk and 2012Z) suggest there could be a correlation between velocity and absolute magnitude that saturates at lower luminosities (i.e., all SNe Ia with $M_r \ge-17.5$ mag have velocities of $\sim0$ km s$^{-1}$). The small sample size renders this correlation tentative; additional data are needed to improve the sample statistics and further analysis to clarify what physical origins may drive this trend.

0.3-2 micron SEDs

SN 2024pxl has three epochs with a complete optical to NIR SED at approximately one week before maximum light, at maximum light, and approximately one and a half weeks after maximum light. These are shown in Figure 12. The pre-peak SED shows a hot explosion dominated by the optical light. The NIR is featureless due to the high temperature of the explosion. At peak light, lines begin to form in the NIR as the explosion cools, but it is still hot and dominated by the optical. A week and a half later, the spectrum has cooled signifi-

cantly, and the NIR contribution is more pronounced, with absorption features forming in the NIR as the temperature drops (although the explosion remains hot enough to ionize C).

Conclusion

We present multi-wavelength observations of the nearby SN 2024pxl: an intermediate luminosity SN Iax with a typical light curve shape ($M_r = -16.99 \pm0.32$ mag and $\Delta m_{15}(r) = 0.69 \pm0.05$ mag). Our extensive NIR photometry and spectroscopy make SN 2024pxl one of the most comprehensively observed SN Iax in the NIR yet, alongside only SNe 2005hk and 2012Z. The rising light curve of SN 2024pxl does not show evidence for excess flux at early times, which is inconsistent with significant interaction with a companion star or circumstellar material.

SN 2024pxl has the earliest-ever NIR spectrum of a SN Iax to date, at $-9.1$ days. The early-time NIR spectra are featureless, indicating a hot explosion at early times. At later times, $\sim$30 days after maximum light, the NIR spectra show resolved Co II features in the $H$ and $K$ bands. Our optical spectra reveal a well-mixed, low-velocity, and highly ionized ejecta at early times characterized by Fe III absorption features. SN 2024pxl exhibits a strong C II $\lambda6580$ feature throughout its photospheric phase. The smoothly rising light curves, wellmixed ejecta, and persistent C features are consistent with a deflagration explosion.

Early-time and comprehensive NIR observations spanning the entire lifetimes of SNe Iax offer additional tests of the leading deflagration models. The $J-H$ and $g-J$ color curves of SNe Iax significantly differ from those of normal SNe Ia and suggest that SNe Iax begin adiabatic cooling immediately without centrally-concentrated $^{56}\mathrm{Ni}$ heating the ejecta. Around a month after maximum light, there may be a trend between the velocity of the Co II features and the absolute magnitude of SNe Iax, with more luminous objects showing a larger blueshift.

As more SNe Iax are discovered shortly after the explosion with multi-wavelength follow-up, larger population-level studies can be conducted at every phase, especially the earliest phases after the explosion.

Acknowledgments

We thank the anonymous referee for comments that improved the manuscript. We additionally thank Peter Hoeflich and Chris Burns for helpful discussions.

This material is based upon work supported by the National Science Foundation Graduate Research Fellowship Program under Grant Nos. 1842402 and 2236415.

Any opinions, findings, conclusions, or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

C.A. is supported by STScI grants (JWST-GO-02114, JWST-GO-02122, JWST-GO-04522, JWST-GO-03726, JWST-GO-6582, HST-AR-17555, JWST-GO-04217, JWST-GO-6023, JWST-GO-5290, JWST-GO-5057, JWST-GO-6677) and JPL-1717705.

D.O.J. acknowledges support from NSF grants AST-2407632 and AST-2429450, NASA grant 80NSSC24M0023, and HST/JWST grants HST-GO-17128.028, HST-GO-16269.012, and JWST-GO-05324.031, awarded by the Space Telescope Science Institute (STScI), which is operated by the Association of Universities for Research in Astronomy, Inc., for NASA, under contract NAS5-26555.

The Shappee group at the University of Hawai’i is supported with funds from NSF (grants AST-1908952, AST-1911074, & AST-1920392) and NASA (grants HST-GO-17087, 80NSSC24K0521, 80NSSC24K0490, 80NSSC24K0508, 80NSSC23K0058, & 80NSSC23K1431).

A.D. is supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Grant Agreement No. 101002652).

J.T.H. acknowledges support from NASA grant 80NSSC23K1431.

C.D.K. gratefully acknowledges support from the NSF through AST-2432037, the HST Guest Observer Program through HST-SNAP-17070 and HST-GO-17706, and from JWST Archival Research through JWST-AR-6241 and JWST-AR-5441.

Based on observations obtained at the international Gemini Observatory, a program of NSF NOIRLab, which is managed by the Association of Universities for Research in Astronomy (AURA) under a cooperative agreement with the U.S. National Science Foundation on behalf of the Gemini Observatory partnership: the U.S. National Science Foundation (United States), National Research Council (Canada), Agencia Nacional de Investigación y Desarrollo (Chile), Ministerio de Ciencia, Tecnología e Innovación (Argentina), Ministério da Ciência, Tecnologia, Inovações e Comunicações (Brazil), and Korea Astronomy and Space Science Institute (Republic of Korea). Data were taken by programs GN-2024A-Q-136, GN-2024A-Q-226, and GN-2024B-Q-109. This work was enabled by observations made from the Gemini North telescope, located within the Maunakea Science Reserve and adjacent to the summit of Maunakea. We are grateful for the privilege of observing the Universe from a place that is unique in both its astronomical quality and its cultural significance.

The Infrared Telescope Facility is operated by the University of Hawaii under contract 80HQTR24DA010 with the National Aeronautics and Space Administration.

UKIRT is owned by the University of Hawaii (UH) and operated by the UH Institute for Astronomy.

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.

We thank Las Cumbres Observatory and its staff for their continued support of ASAS-SN. ASAS-SN is funded in part by the Gordon and Betty Moore Foundation through grants GBMF5490 and GBMF10501 to the Ohio State University and also funded in part by the Alfred P. Sloan Foundation grant G-2021-14192.

This work has used data from the Asteroid Terrestrial-impact Last Alert System (ATLAS) project. The Asteroid Terrestrial-impact Last Alert System (ATLAS) project is primarily funded to search for near-earth asteroids through NASA grants NN12AR55G, 80NSSC18K0284, and 80NSSC18K1575; byproducts of the NEO search include images and catalogs from the survey area. This work was partially funded by Kepler/K2 grant J1944/80NSSC19K0112 and HST GO-15889, and STFC grants ST/T000198/1 and ST/S006109/1. The ATLAS science products have been made possible through the contributions of the University of Hawaii Institute for Astronomy, Queen’s University Belfast, the Space Telescope Science Institute, the South African Astronomical Observatory, and the Millennium Institute of Astrophysics (MAS), Chile.

Facilities: Swift(UVOT), IRTF(SpeX), Gemini:Gillett(GNIRS), UKIRT(WFCAM), UH:2.2m(SNIFS)