MNRAS 536, 3588–3600 (2025) https://doi.org/10.1093/mnras/stae2784 
Advance Access publication 2024 December 20 
SN 2023tsz: a helium-interaction-dri v en superno v a in a v ery lo w-mass 
galaxy 
B. Warwick , 1 ‹ J. Lyman , 1 M. Pursiainen, 1 D. L. Coppejans, 1 L. Galbany, 2 , 3 G. T. Jones , 1 
T. L. Killestein , 1 , 4 A. Kumar , 1 S. R. Oates , 5 K. Ackley, 1 J. P. Anderson, 6 , 7 A. Aryan, 8 
R. P. Breton , 9 T. W. Chen, 8 P. Clark, 10 V. S. Dhillon , 11 , 12 M. J. Dyer , 11 A. Gal-Yam, 13 
D. K. Galloway, 14 , 15 C. P. Guti ´errez , 2 , 3 M. Gromadzki , 16 C. Inserra , 17 F. Jim ´enez-Ibarra, 14 
L. Kelsey , 10 R. Kotak, 4 T. Kravtsov, 4 H. Kuncarayakti, 4 , 18 M. R. Magee , 1 K. Matilainen, 4 
S. Mattila, 4 , 19 T. E. M ¨uller -Bra v o, 2 , 3 M. Nicholl, 20 K. Noysena, 21 L. K. Nuttall, 10 P. O’Brien, 22 
D. O’Neill, 1 E. Pall ´e, 12 T. Pessi, 6 T. Petrushevska, 23 G. Pignata, 24 D. Pollacco, 1 F. Ragosta, 25 , 26 
G. Ramsay , 27 A. Sahu, 1 D. K. Sahu, 28 A. Singh , 29 , 30 J. Sollerman , 30 E. Stanway , 1 
R. Starling, 22 D. Steeghs , 1 R. S. Teja 28 , 31 and K. Ulaczyk 1 
Affiliations are listed at the end of the paper 
Accepted 2024 December 7. Received 2024 December 4; in original form 2024 September 23 
A B S T R A C T 
SN 2023tsz is a Type Ibn supernova (SN Ibn), an uncommon subtype of stripped-envelope core-collapse supernovae (SNe), 
disco v ered in an extremely low-mass host. SNe Ibn are characterized by narrow helium emission lines in their spectra and 
are believed to originate from the collapse of massive Wolf–Rayet (WR) stars, though their progenitor systems still remain 
poorly understood. In terms of energetics and spectrophotometric evolution, SN 2023tsz is largely a typical example of the 
class, although line profile asymmetries in the nebular phase are seen, which may indicate the presence of dust formation or 
unshocked circumstellar material. Intriguingly, SN 2023tsz is located in an extraordinarily low-mass host galaxy that is in the 
second percentile for stripped-envelope SN host masses and star formation rates (SFRs). The host has a radius of 1.0 kpc, a 
g-band absolute magnitude of −12 . 72 ± 0 . 05, and an estimated metallicity of log ( Z ∗/ Z ) ≈ −1 . 6. The SFR and metallicity of 
the host galaxy raise questions about the progenitor of SN 2023tsz. The low SFR suggests that a star with sufficient mass to 
evolve into a WR would be uncommon in this galaxy. Further, the very low metallicity is a challenge for single stellar evolution 
to enable H and He stripping of the progenitor and produce an SN Ibn explosion. The host galaxy of SN 2023tsz adds another 
piece to the ongoing puzzle of SNe Ibn progenitors, and demonstrates that they can occur in hosts too faint to be observed in 
contemporary sky surveys at a more typical SN Ibn redshift. 
Key words: circumstellar matter – stars: massive – supernovae: general – transients: supernovae. 
1
T  
t  
l  
(  
e  
c  
1  
c  
e  
n  
a  

t  
t  
e  
1  
3
 
S  
i  
S  
s  
S  
i  
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/article/536/4/3588/7929872 by Turun Yliopisto user on 14 M
arch 2025 I N T RO D U C T I O N  
ype Ibn supernovae (SNe Ibn) are a subclass of supernovae (SNe)
hat are characterized by the presence of narrow helium (He) emission
ines in their spectra but the absence of strong hydrogen (H) features
e.g. Gal-Yam 2017 ; Smith 2017 ). These spectral properties are
xplained by the interaction of SN ejecta with He-rich, but H-poor,
ircumstellar material (CSM). The first SN Ibn was observed in
999 (SN 1999cq; Matheson et al. 2000 ). Ho we ver, the label was not
oined until the analysis of SN 2006jc (F ole y et al. 2007 ; Pastorello
t al. 2007 ), which is considered the prototypical SN Ibn. Despite
early two decades since their identification, there are still questions
bout the nature of their progenitors. These questions persist due to E-mail: ben.w arwick@w arwick.ac.uk 
1
J
Published by Oxford University Press on behalf of Royal Astronomical Socie
Commons Attribution License ( https:// creativecommons.org/ licenses/ by/ 4.0/ ), whihe lack of a confirmed direct detection of an SN Ibn progenitor as
his type is relatively rare, with only 66 classified to date. 1 It was
stimated by Maeda & Moriya ( 2022 ) that they make up around
 per cent of core-collapse (CC) SNe, with an observed rate of
 per cent (Perley et al. 2020 ). 
The original progenitor suggested for the prototypical SN Ibn,
N 2006jc was an H-poor massive Wolf–Rayet (WR) star embedded
n an He-rich CSM (Pastorello et al. 2007 ; Tominaga et al. 2008 ).
uch progenitors are proposed for SNe Ibn, as the mass-loss that WR
tars undergo prior to explosion can explain the properties seen in
N Ibn light curves (Maeda & Moriya 2022 ). This progenitor model
s supported by the fact that the majority of SNe Ibn are found in Based on a Transient Name Server, https://www.wis-tns.org /, query on 2024 
une 18. 
© 2024 The Author(s). 
ty. This is an Open Access article distributed under the terms of the Creative 
ch permits unrestricted reuse, distribution, and reproduction in any medium, 
provided the original work is properly cited. 
SN 2023tsz: an SN Ibn in a very low-mass host 3589 
a  
2  
d  
a  
e  
  
b
C  
C  
s  
&  
b  
t
c  
e  
e  
p
t
d
c  
a  
s  
T
a  
t  
s  
s  
e
 
b
e  
r
(  
d
S  
1  
a  
c
L  
s  
(  
b
o
a
a
p  
0
i  
a
=  
2  
0
I  
U
f
2
I  
p
Figure 1. Image showing the host of SN 2023tsz (circled) and its immediate 
surroundings, including LEDA 152972. The image was created using g, r , and 
i observations from DR9 of the DESI Le gac y Surv e y (De y et al. 2019 ). The 
circle is the radius around SN 2023tsz’s host that contains 99 per cent of its 
light in the g band. The three image cut-outs in the bottom right of the image 
show the science (left), template (middle), and difference (right) images from 
GOTO that were used to disco v er SN 2023tsz. The GOTO science image was 
taken + 3 . 2 d with respect to our estimate of the peak (see text). 
L  
(  
(  
f  
a
2  
p
u
2  
p  
u  
U
f
A  
T
s
p  
S  
t  
S  
2
O
t
C
E
T  
e
p  
I
2 https:// asas-sn.osu.edu/ 
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/article/536/4/3588/7929872 by Turun Yliopisto user on 14 M
arch 2025ctiv e star-forming re gions (e.g. P astorello et al. 2015a ; Taddia et al.
015 ). Ho we ver, there is a notable exception, PS1-12sk, which was
isco v ered in the outskirts of an elliptical galaxy in a region with
 low star formation rate (SFR; Sanders et al. 2013 ; Hosseinzadeh
t al. 2019 ). Another proposed progenitor for SNe Ibn is a low-mass,
 5 M , He star likely arising from a binary system. In this scenario,
inary interactions are the cause of the mass-loss that creates the 
SM prior to the explosion. The SN Ibn then arises from either the
C of the He star, or an explosion triggered by the merger of the He
tar and its binary companion (Maund et al. 2016 ; Dessart, Hillier
 Kuncarayakti 2022 ; Dong et al. 2024 ). Such a progenitor could
e more likely in an older stellar population. It is also plausible
hat multiple progenitor channels could lead to SNe that would be 
lassified as an SN Ibn, such as the confirmed case of an SN IIb that
xploded inside a dense CSM and appeared to be an SN Ibn (Prentice
t al. 2020 ). Consequently, the nature and homogeneity of SN Ibn
rogenitors remain uncertain. 
Compared to most CC SNe, which are powered primarily through 
he decay of 56 Ni produced during the explosion, a combined nickel 
ecay and CSM interaction model is required to explain the light 
urves of SN Ibn (Clark et al. 2020 ). The CSM interaction is dominant
t early times of SNe Ibn evolution, and is thought to explain their
horter time-scales for both the rise and decline of their light curves.
hese shorter time-scales make SN Ibn observationally rare as they 
re harder to characterize upon disco v ery, and are particularly hard
o capture on their rise to the peak. The current generation of all-
k y surv e ys, which are observing at higher cadences than previous
urv e ys, will help to find and characterize SNe Ibn earlier in their
 volution allo wing impro v ed probing of their physics. 
The focus of this paper is the SN Ibn SN 2023tsz. It was disco v ered
y the Gra vitational-wa ve Optical Transient Observer (GOTO; Dyer 
t al. 2022 , 2024 ); Steeghs et al. 2022 ) on 2023 September 28 and
eported to the Transient Name Server under the name GOTO23anx 
Godson et al. 2023 ). There was a prior non-detection and two
etections from All-Sky Automated Survey for Supernovae (ASAS- 
N; Shappee et al. 2014 ; Kochanek et al. 2017 ) on 2023 September
3, 19, and 25, respectively. The SN is associated with a faint galaxy,
 possible satellite of LEDA 152972, located 33 arcsec from its
entre, visible in the Dark Energy Spectroscopic Instrument (DESI) 
e gac y Surv e y Data Release 9 (DR9) images (Dey et al. 2019 ) as
hown in Fig. 1 . It is also the first object to be disco v ered, reported
Godson et al. 2023 ), and classified (Pursiainen et al. 2023c ) entirely
y the GOTO Collaboration. 
This paper is structured as follows: Section 2 presents the 
bservations and data reduction methods; Section 3 presents the 
nalysis of the data; Section 4 discusses the implications of our 
nalysis; and Section 5 concludes our findings. Throughout this 
aper, we have corrected for a Milky Way extinction of E( B − V ) =
.036 mag (Schlafly & Finkbeiner 2011 ). When analysing the spectra 
n Section 3.2 there is no evidence for significant host extinction. We
ssume a flat  cold dark matter (  CDM) cosmology with m 
 0.31 and H 0 = 67.66 km s −1 Mpc −1 (Planck Collaboration VI
020 ). In the absence of host lines, we adopt the redshift value of z =
.028, derived from spectrum template matching using the Supernova 
dentification ( SNID ) code (Blondin & Tonry 2011 ), see Section 3.2 .
sing this cosmology and redshift, we calculate a luminosity distance 
or SN 2023tsz of 130 Mpc. 
 OBSERVATIONS  A N D  DATA  R E D U C T I O N  
n addition to GOTO L band (described in Steeghs et al. 2022 ),
hotometry of SN 2023tsz was collected in bands ugriz using the iverpool Telescope (LT; Steele et al. 2004 ), the Rapid Eye Mount
REM) telescope (Covino et al. 2004 ), and Las Cumbres Observatory
LCO; Brown et al. 2013 ) Global Telescope Network. The data
rom LT and LCO were provided already pre-reduced for bias, dark,
nd flat-field corrections using their own pipelines (McCully et al. 
018 ). The data from REM telescope were reduced using our own
ipeline. The light curves from these observations were calculated 
sing the photometry-sans-frustration ( PSF ) pipeline (Nicholl et al. 
023 ) making use of the inbuilt template subtraction of PSF . Further
hotometry was collected in the near -ultra violet (NUV) bands uvm 2,
v w2, and uv w1 with the Neil Gehrels Swift Observatory ( Swift )
ltraviolet and Optical Telescope (UV O T; Roming et al. 2005 ), g 
rom ASAS-SN, and o from the Asteroid Terrestrial-impact Last 
lert System (ATLAS; Tonry et al. 2018 ; Smith et al. 2020 ).
he ASAS-SN light curve was generated using their Sky Patrol 
ervice. 2 The ATLAS light curve was generated using their forced 
hotometry server (Shingles et al. 2021 ). The light curves from
wift UV O T were reduced using a 7 arcsec aperture to extract
he photometry. This aperture was used due to a slight issue with
wift that caused smearing of the sources on the images (Cenko
023 ). 
Spectra of SN 2023tsz were obtained using the Alhambra Faint 
bject Spectrograph and Camera (ALFOSC) on the Nordic Op- 
ical Telescope (NOT), the Himalayan Faint Object Spectrograph 
amera (HFOSC) on the Himalayan Chandra Telescope (HCT), the 
SO Faint Object Spectrograph Camera 2 (EFOSC2) on the New 
 echnology T elescope (NTT) at La Silla Observatory, taken by the
xtended Public ESO Spectroscopic Survey for Transient Objects 
lus (ePESSTO + ; Smartt et al. 2015 ), and the Optical System for
maging and low-Intermediate-Resolution Integrated Spectroscopy MNRAS 536, 3588–3600 (2025) 
3590 B. Warwick et al. 
M
Figure 2. The multiband light curve of SN 2023tsz. In the legend, the number denotes the offset applied to the values in that filter. The third-order polynomial 
used to fit the peak date in the g band is shown by the black dotted curve. Dashed lines are used to show the light curve decline in each band over 15 d intervals 
after the peak. The decline rate for each interval (in mag d −1 ) is indicated on the plot next to the dashed line. Each 15 d interval is shown by the vertical light 
grey lines. We only fit the decline if there are at least three observations in the 15 d increment. The thick grey lines at the top of the plot represent the epochs 
when spectroscopic observations were obtained. 
p  
f  
p  
a  
a  
f  
e  
f  
o  
T
 
K  
i  
(  
W  
(  
W  
t  
b  
a
3
4
3
3
T  
A  
a  
d  
a  
s  
s  
t  
d  
p  
p  
w  
0  
p  
w  
a  
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/article/536/4/3588/7929872 by Turun Yliopisto user on 14 M
arch 2025lus (OSIRIS + ) on the Gran Telescopio Canarias (GTC). The data
rom ALFOSC were reduced using the PYNOT-REDUX reduction
ipeline. 3 The spectroscopic data from HFOSC were reduced in
 standard manner using the packages and tasks in IRAF with the
id of the PYTHON scripts hosted at REDPIPE (Singh 2021 ). The data
rom EFOSC2 were reduced using the PESSTO pipeline 4 (Smartt
t al. 2015 ). GTC optical spectra from OSIRIS + were reduced
ollowing the routines in Piscarrera et al. (in preparation), based
n PYPEIT (Prochaska et al. 2020 ). The spectral log is provided in
able A1 . 
Photometry of the host of SN 2023tsz was obtained from the
ilo-De gree Surv e y (KiDS; de Jong et al. 2015 ) in the u , g, r , and
 bands, and from the VISTA Kilo-degree Infrared Galaxy Surv e y
VIKING; Edge et al. 2013 ) in the z band, detailed in Table A2 .
e also obtain upper limits from the Galaxy Evolution Explorer
 GALEX ; Martin et al. 2005 ) in the FUV and NUV bands, and the
ide-field Infrared Survey Explorer ( WISE ; Wright et al. 2010 ) in
he WISE 1, 2, 3, and 4 bands. Finally, one epoch of ALFOSC V -
and imaging polarimetry was obtained. The data were reduced and
nalysed following Pursiainen et al. ( 2023a ). NRAS 536, 3588–3600 (2025) 
 github .com/jkrogager/PyNO T/ 
 https:// github.com/ svalenti/ pessto 
e  
a  
m  
U  
g  
w   ANALYSI S  
.1 Photometry 
he multiband light curve of SN 2023tsz is shown in Fig. 2 . Only the
SAS-SN g band had pre-peak observations, one a non-detection
nd the other a detection. This is due to the fact that SN 2023tsz was
isco v ered as it exited solar conjunction, resulting in a late disco v ery
nd corresponding poor pre-peak observations. The decline was well
ampled in multiple bands up to ∼40 d post-peak, with later epochs
ampled in g, r , and i. We define our peak epoch as the peak in
he g band obtained from fitting a third-order polynomial to the
ata within ±20 d of the observed peak. This fitting gave us a
eak date of MJD 60212 ± 5. The large error on the time of the
eak is due to the lack of constraining observations; going forward
e adopt MJD 60212 as our peak date and use this for phase =
. To estimate its decline rate in 15 d increments post-peak, we
erformed a linear regression on the data. These 15 d increments
ere chosen to investigate the visible flattening of the light curve,
nd to allow for comparison to other SN Ibn from Hosseinzadeh
t al. ( 2017 ). The results of this are shown in Fig. 2 . In the sample
nalysis of Hosseinzadeh et al. ( 2017 ), decline rates were estimated
ostly in the R band, closely matched with our r-band observations.
nfortunately, determining the r-band peak brightness is difficult
iven our data coverage. The observed r-band peak was at + 4.76 d
hen the SN was −19 . 03 ± 0 . 03 mag. The r-band decline rate in the
SN 2023tsz: an SN Ibn in a very low-mass host 3591 
Figure 3. The initial 15 d post-peak decline rate and observed peak absolute 
magnitude of SN 2023tsz compared with other SNe Ibn. The tw o mark ers for 
SN 2023tsz encapsulate the uncertainty from not capturing the peak in the r 
band. The filled marker is an estimate of the peak by extrapolating the decline 
rate back to the epoch of the g -band peak, and the open square symbol is a 
certain lower limit on the absolute magnitude from the r -band observation at 
+ 4.76 d. Although we are comparing r-band observations to an R-band data 
set, differences between the filter throughputs are not at a level to change our 
inferences. The data for the SNe Ibn sample are from Hosseinzadeh et al. 
( 2017 ). 
fi  
b
o  
a  
s  
u  
T  
W  
a  
S  
o  
d
r
c
t
b  
h
u  
t  
c  
e
w
t
a
p
A
G
T  
w
t  
d
c  
Figure 4. The bolometric light curve (top), blackbody temperature (middle), 
and blackbody radius (bottom) of SN 2023tsz o v er all observational epochs. 
For the epochs that only have g-band observations, the markers are unfilled. 
t  
b
o  
i
B  
l  
p  
t  
o  
a  
b  
d  
t  
t  
e  
t  
e  
u
M
c  
i  
b
b
t  
i  
r  
2
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/article/536/4/3588/7929872 by Turun Yliopisto user on 14 M
arch 2025rst 15 d post-peak is 0 . 145 ± 0 . 002 mag d −1 , which if we extrapolate
ack to the peak makes the estimated peak r-band absolute magnitude 
f SN 2023tsz −19 . 7 ± 0 . 7 mag. This estimate is likely to be at
n epoch before the true r-band peak as the SN is cooling and
o would peak in the g band earlier than the r band. It is also
nlikely that the SN declines at this rate immediately from the peak.
herefore, we consider this value as an upper limit on the magnitude.
e show the range of peak magnitudes and the initial decline rate
longside a sample of other SN Ibn in Fig. 3 . The figure shows that
N 2023tsz is slightly brighter and faster than the majority of the
ther SN Ibn from Hosseinzadeh et al. ( 2017 ), but is well within the
istribution. 
The bolometric light curve was created using the SUPERBOL 
outine (Nicholl 2018 ). SUPERBOL works by fitting polynomial light- 
urve models to each band independently and then interpolating 
he magnitudes to the times of observation in a defined reference 
and before using them to fit blackbody models. As the g band
as the most observations, and the only pre-peak observations, we 
sed it as our reference band. In bands where we do not have late-
ime co v erage (all bands e xcept g, r , and i), we assume that the
olour relative to the g band remains constant at the last measured
poch for that band. The temperatures of the first two epochs, 
hich had only g -band data, were estimated by extrapolating the 
emperature curve with a second-degree polynomial. This was done 
s at these epochs only g-band photometry was obtained resulting in 
oor temperature estimates from fitting the bolometric light curve. 
ll bands in which observations were obtained, other than the 
OTO L band, were used to generate the bolometric light curve. 
he output of this routine was then smoothed by taking the error-
eighted average of the bolometric luminosity ( L bol ), blackbody 
emperature ( T BB ), and blackbody radius ( R BB ) for each individual
ay of observations. The full bolometric light curve, along with the 
orresponding T BB , and R BB are presented in Fig. 4 . The error inhe distance was not considered when estimating the errors in the
olometric luminosity. 
We fit to the bolometric light curve the semi-analytical models 
f Chatzopoulos, Wheeler & Vinko ( 2012 ) for a combined CSM
nteraction and 56 Ni decay (CSM + Ni) model, using an EMCEE 
ayesian analysis (F oreman-Macke y et al. 2013 ). We set the power-
a w inde x for the inner density profile of the ejecta as δ = 1, the
ower-la w inde x for the outer component of the ejecta as n = 10,
he power-law index of the CSM as s = 0, and the optical opacity
f the CSM as κ = 0 . 2 cm 2 g −1 . These values were chosen to
llow for a comparison to the work of Pellegrino et al. ( 2022a ,
 ). We use a constant density CSM as, in the model we use, the
ensity profile is degenerate with the radius of the star, and changing
he density will ef fecti vely only af fect the radius. The radius of
he progenitor was chosen to be R 0 = 1 . 25 × 10 12 cm and the
 xpansion v elocity of the ejecta v = 10 000 km s −1 , aligning with
he values used in Chatzopoulos et al. ( 2012 ). We constrained the
xplosion epoch to MJD 60203 . 63 ± 2 . 99. This explosion epoch
ses the non-detection, at MJD 60200.64, and first detection, at 
JD 60206.62 from ASAS-SN, as its bounds. The best-fitting model 
an be seen in Fig. 5 with the best-fitting model parameters presented
n Table 1 . Our first two data points only provide an estimate of the
olometric luminosity as they are from observations only in the g 
and and an extrapolated temperature. We have assumed that this 
emperature would decrease from the first observation, as is the case
n standard Type II SNe. Ho we ver, it has been observed that some
ise in temperature after explosion (Hiramatsu 2022 ; Hiramatsu et al.
023 ). MNRAS 536, 3588–3600 (2025) 
3592 B. Warwick et al. 
M
Figure 5. The best-fitting combined CSM + Ni model to the bolometric 
light curve of SN 2023tsz using the models of Chatzopoulos et al. ( 2012 ). 
For the epochs that only have g-band observations, the markers are unfilled. 
The individual CSM and Ni models are shown by the dash–dotted and dotted 
lines, respectively. The combined model is shown by the dashed red line. The 
best-fitting CSM (M CSM ), ejecta (M ej ), and Ni (M Ni ) masses are shown in the 
plot. The vertical dashed line shows the epoch of the ASAS-SN non-detection. 
Table 1. The best-fitting parameters for the CSM + Ni model. 
Parameter Description Value 
M CSM (M ) CSM mass 0 . 05 + 0 . 05 −0 . 02 
M ej (M ) Ejecta mass 0 . 59 + 0 . 69 −0 . 38 
E (10 51 erg) Total energy of the SN 0 . 35 + 0 . 13 −0 . 12 
˙M (10 −4 M ) Progenitor mass-loss rate 2 . 23 + 4 . 08 −1 . 66 
t expl (d) Time of explosion relative to peak −10 . 39 + 2 . 11 −1 . 49 
M Ni (M ) Total mass of nickel produced 0 . 04 + 0 . 01 −0 . 01 
3
T  
I  
&  
S  
a  
f  
w  
r  
p  
H  
v  
w  
w  
l  
0  
f  
a  
a  
a  
b  
c  
b  
h
 
t  
w  
t  
c  
T  
e  
c  
(  
7  
s  
F  
f  
(  
p  
s  
b  
o
 
H  
s  
d  
p  
t  
a
 
b  
o  
s  
i  
s  
s  
l  
c  
o  
o  
o
3
T  
+  
i  
a  
t  
m  
f  
>  
f  
i  
w  
a  
f  
s  
p  
b
 
m  
t  
M  
p  
t  
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/article/536/4/3588/7929872 by Turun Yliopisto user on 14 M
arch 2025.2 Spectra 
he spectral time series of SN 2023tsz is compared to example SNe
bn with spectra taken at similar epochs in Fig. 6 . SNID (Blondin
 Tonry 2011 ) was used to match the initial NOT spectrum of
N 2023tsz to SNe template spectra and found a best match with
n SN Ibn at a redshift of z = 0.028, which we choose to adopt
or the redshift in our analysis. This template matching, along
ith manual inspection of the lines present in the spectrum, gave
ise to the reported classification based on similarity to several
rominent known SN Ibn, also shown in Fig. 6 . Investigating the
e lines in the spectrum obtained at + 36 . 3 d found a redshift
alue of 0 . 0260 –0 . 0265. This is a difference of 450 –600 km s −1 ,
hich could be partially due to an inherent offset of the line. This
ould result in a luminosity distance of 120 Mpc and bolometric
uminosity 17 per cent lower than our values obtained with z =
.028. Our adopted redshift value differs from the redshift value
or LEDA 152972 of z = 0.3515 (Liske et al. 2015 ). This implies
 difference of 2200 km s −1 in recession velocity, significantly
bo v e the escape v elocity e xpected of LEDA 152972, suggesting
 chance alignment with the host of SN 2023tsz, rather than them
eing gravitationally bound. Ho we ver, without narro w host lines, we
annot rule this out. The spectra show no presence of suppressed
lue emission so we conclude there is no evidence for significant
ost extinction. NRAS 536, 3588–3600 (2025) The most prominent spectral features across all the spectra are
he He I 5876, 6678, and 7066 Å lines. To investigate these lines,
e fit a Gaussian emission and absorption component to reproduce
he P Cygni profile, along with an additional Lorentzian emission
omponent to account for the broad wings of the lines, to each line.
his fitting was performed o v er a range extending 6000 km s −1 on
ach side of the line centre. The best-fitting values for each of these
omponents were then determined using an EMCEE Bayesian analysis
F oreman-Macke y et al. 2013 ). Additionally, for the He I 5876 and
066 Å lines in the + 35 . 3 and + 70 . 2 d spectra, we performed the
ame analysis with two Gaussian emission components (shown in
ig. 7 ) to investigate the noticeable blue asymmetry of the spectral
eatures, likely due to a suppression of the red wing of the feature
in the + 70 . 2 d spectrum). In the classification spectrum, the most
rominent line is the He I 5876 Å. In this line, we observe a
ignificant P Cygni profile, with the profile minimum blueshifted
y −1140 + 100 −180 km s −1 . We do not find a fit to a P-Cyqni profile in the
ther spectra. 
The He I 5876 Å line becomes more prominent along with the
e I 6678 and 7066 Å lines up to at least + 36 . 3 d post-peak. These
trong He I lines are characteristic of typical SNe Ibn. The lines
o weaken again at + 70 . 2 d post-peak, but still remain the most
rominent features. At all times after the initial classification epoch,
he emission component of these lines dominates o v er the blueshifted
bsorption component. 
In the + 70 . 2 d spectrum, the He I lines appear to have a noticeable
lue asymmetry. This asymmetry was confirmed by comparing fits
f the lines in this spectrum to fits of the same lines in the + 36 . 3 d
pectrum. The fits for the He I 5876 and 7066 Å lines are shown
n Fig. 7 . The features in the + 36 . 3 d spectrum are well fit by a
ingle Gaussian with no evident second peak and so we determine
uch a component is not required. In the + 70 . 2 d spectrum, the same
ines are best fit by two Gaussian components, a dominant Gaussian
omponent and an additional, smaller, narrower Gaussian component
n the blue side of the dominant Gaussian. In both cases the centre
f the dominant Gaussian was fixed for each spectrum. The results
f these fits are shown in Table 2 . 
.3 Polarimetry 
he NOT/ALFOSC V -band polarimetry of SN 2023tsz taken at
 15 . 6 d appears to be consistent with zero polarization, as shown
n Fig. 8 . For the SN, we find Stokes parameters Q = −0 . 36 ± 0 . 28
nd U = 0 . 34 ± 0 . 29, but the ALFOSC field of view also co v ers
wo nearby bright stars, which are perfectly consistent with the
easured polarization of the SN. Based on the Gaia parallaxes
rom the Data Release 3 (Gaia Collaboration 2023 ), the stars are
 150 pc abo v e the Milky Way plane, and as such, they probe the
ull Galactic dust column and can be used to measure the Galactic
nterstellar polarization (ISP) component (Tran 1995 ). We adopt the
eighted average of the two stars and find Q ISP = −0 . 23 ± 0 . 13
nd U ISP = 0 . 27 ± 0 . 13 for the Galactic ISP. The lo w v alue is
urther supported by the Heiles catalogue (Heiles 2000 ), which
hows 10 stars within 5 ◦ from the SN that are consistent with a
olarization fraction of P < 0 . 2 per cent. After ISP and polarization
ias corrections, we find P = 0 . 08 ± 0 . 31 per cent for the SN. 
Whilst we cannot directly estimate the host galaxy ISP, its
aximum value should be related to the host galaxy extinction via
he following empirical relation: P ISP < 9 × E( B − V ) (Serkowski,
athewson & Ford 1975 ). The photometric and spectroscopic
roperties of the SN and the host galaxy properties all support
he assumption of low host extinction, implying that the host ISP
SN 2023tsz: an SN Ibn in a very low-mass host 3593 
Figure 6. The spectroscopic time series of SN 2023tsz compared to the SN Ibn 2015U (Shivvers et al. 2016 ), 2018bcc (Karamehmetoglu et al. 2021 ), and 
2018jmt (Chaso vniko v et al. 2018 ). Spectra for these three SNe were obtained from the Weizmann Interactive Supernova Data Repository (Yaron & Gal-Yam 
2012 ). The phase relative to the peak is shown next to each spectrum. Additionally, the instrument used to obtain each spectrum is listed. The He I , He II , H α, 
and H β features are marked by dashed lines, and tellurics are shown by the shaded regions. The spectra are shown in the rest frame of the SN. 
s  
o  
c  
5
i
p  
c  
s  
t
c
3
W  
C  
c
u  
a  
w  
i  
l
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/article/536/4/3588/7929872 by Turun Yliopisto user on 14 M
arch 2025hould also be low. As such, we can conclude that the polarization
f SN 2023tsz is likely intrinsically low. The V band co v ers mostly
ontinuum and the only clear line feature in the V bandpass is He I
876 Å and that manifests mostly as emission. Line emission is 
nherently unpolarized, so it can only decrease the observed V -band 
olarization, but as the line is narrow (2 . 25 + 0 . 73 −0 . 21 × 10 3 km s −1 ) and
o v ers only a small part of the whole band, the depolarizing effect
hould be negligible. As such, we assume the low polarization is
hat of the continuum and conclude that the SN photosphere was 
onsistent with spherical symmetry at + 15 . 64 d post-peak. u  .4 Host data 
e fit the photometric data for the host of SN 2023tsz using the
IGALE software (Boquien et al. 2019 ). CIGALE works by computing
omposite spectral energy distribution (SED) models of the galaxy 
sing specified modules that each deal with a different component of
 galaxy’s emission. It then fits these models to the input observations,
hich for the host of SN 2023tsz are the surv e y observations detailed
n Table A2 , and determines the best-fitting model, the one with the
owest reduced χ2 . Each module has multiple parameters, and the 
ser provides an array of values for each of them. CIGALE thenMNRAS 536, 3588–3600 (2025) 
3594 B. Warwick et al. 
M
Figure 7. Line fitting of the He I 5876 Å line (left) and He I 7066 Å line (right) for the spectra obtained from the NTT (top) and the GTC (bottom). Each line is 
fit with two Gaussian profiles, for the NTT spectra the best fit was found to be with a single Gaussian. For each individual plot, the top panel shows the combined 
profile o v er the spectral feature, the middle panel sho ws the residuals, and the bottom panel sho ws the two indi vidual Gaussian components. 
Table 2. The best-fitting parameters to the He I 5876 Å and He I 7066 Å lines of the + 36 . 3 and + 70 . 2 d spectra. 
Parameter + 36 . 3 d spectrum + 70 . 2 d spectrum 
5876 Å 7066 Å 5876 Å 7066 Å
Primary peak centre a (10 2 km s −1 ) −5.40 + 0 . 31 −0 . 32 −3.90 + 1 . 77 −1 . 12 
Primary amplitude (erg s −1 cm −2 Å) 6.58 + 0 . 22 −0 . 21 × 10 −15 5.30 + 0 . 12 −0 . 12 × 10 −15 4.55 + 0 . 45 −0 . 64 × 10 −16 3.81 + 0 . 32 −0 . 49 × 10 −16 
Primary FWHM (10 3 km s −1 ) 3.40 + 0 . 11 −0 . 10 2.75 + 0 . 07 −0 . 07 2.87 + 0 . 21 −0 . 27 2.17 + 0 . 27 −0 . 22 
Secondary peak centre (km s −1 ) – −1.49 + 0 . 06 −0 . 05 × 10 3 −1.38 + 0 . 15 −0 . 09 × 10 3 
Secondary amplitude (erg s −1 cm −2 Å) – 8.32 + 4 . 63 −2 . 37 × 10 −17 1.39 + 0 . 24 −0 . 22 × 10 −16 
Secondary FWHM (km s −1 ) – 8.55 + 2 . 75 −2 . 02 × 10 2 1.10 + 0 . 36 −0 . 19 × 10 3 
a The same peak centre was fit for both spectral features within each spectrum. 
c  
t  
(  
&  
e  
t
 
p  
t  
r  
o  
5  
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/article/536/4/3588/7929872 by Turun Yliopisto user on 14 M
arch 2025omputes a model for each combination of parameter values. Galaxy
emplates were generated using a delayed star formation history
SFH), using the simple stellar population models of BC03 (Bruzual
 Charlot 2003 ). Dust attenuation was included using the Calzetti
t al. ( 2000 ) law and dust emission following the Draine & Li ( 2007 )
emplates. NRAS 536, 3588–3600 (2025) To allow us to compare how the age of the burst affects the other
arameters in the SED fitting, three different burst age parametriza-
ions were tested: the first had the age of the burst allowed o v er a
ange of ages from 5 to 200 Myr (scenario 1), the second o v er a range
f ages from 100 to 500 Myr (scenario 2), and the third from 1 to
 Myr (scenario 3). The 5–200 Myr range provided the best fit, with a
SN 2023tsz: an SN Ibn in a very low-mass host 3595 
Figure 8. The Stokes Q –U diagram of the V -band polarimetry, before 
ISP correction, taken at + 15 . 6 d. The measured polarization fraction of 
SN 2023tsz is consistent with the two other stars present in the image. 
After correcting for the Galactic ISP and polarization bias, we find P = 
0 . 08 ± 0 . 31 per cent. Q = 0 per cent and P = 1 per cent have been marked 
with dotted lines. 
Figure 9. The best-fitting SED of the host of SN 2023tsz from CIGALE 
is shown by the solid line. The observations from KiDs and VIKING are 
shown by the data points. The dashed line indicates the region (bandpass 
and throughput) of the spectrum that would have been observed by the HST 
F 300 X filter had the host been at the redshift of PS1-12sk ( z = 0 . 054). 
b  
s  
s  
f
b  
b  
s  
v
(
 
t  
9  
i  
3  
f  
h  
c  
(  
o  
d
 
f  
b  
i  
C  
e  
3  
m  
h
i
 
i  
1  
(  
h  
w
o
v  
u  
fi  
W  
o  
T
m  
t  
o
4
4
T  
t  
f  
S  
a  
fi  
v
f  
f  
S
a  
o  
u  
p  
fi  
m  
s  
l  
w  
S  
c  
5
 
p  
b
p
a
u
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/article/536/4/3588/7929872 by Turun Yliopisto user on 14 M
arch 2025urst age of 50 Myr. The best-fitting model SED, from scenario 1, is
hown in Fig. 9 . The key parameters of the best fit in each scenario are
hown in Table 3 . As can be seen, all the fits have comparable values
or their SFR, stellar mass, and specific SFR (sSFR) meaning the 
est-fitting values are robust. For the rest of the analysis, we use the
est fit from scenario 1 and its values for SFR, 2 . 64 × 10 −3 M  yr −1 ,
tellar mass, 1 . 15 × 10 7 M , and sSFR, 2 . 29 × 10 −10 yr −1 . These
alues are shown in comparison to a sample of stripped-envelope 
SE) SN host galaxies, from Schulze et al. ( 2021 ), in Fig. 10 . 
We measure the galaxy to have a radius of 1.6 arcsec, equi v alent
o a physical radius of 1.0 kpc, defined by the radius encompassing
9 per cent of the galaxy’s light in the Le gac y Surv e y DR9 g-band
mages. This equates to an area of 8.2 arcsec 2 and a physical size of
.1 kpc 2 . This radius is shown in Fig. 1 . This, along with the SFR
rom our best-fitting SED, puts the SFR density of SN 2023tsz’s
ost at 9 . 3 × 10 −4 M  yr −1 kpc −2 . When we compare this to theombined data sets of Hosseinzadeh et al. ( 2019 ) and Galbany et al.
 2018 ) there are only two SN Ib/c (4.4 per cent of the sample) and
ne SN Ibn (5.9 per cent of the sample) host galaxies with lower SFR
ensities. 
Using the fraction of stars per unit mass that produce CC SNe
rom Botticella et al. ( 2012 ), K CC ∼ 0 . 01 M −1  , the SFR from our
est-fitting SED would mean that we expect a massive star to explode
n this galaxy every ≈35 kyr . As SNe Ibn account for ≈1 per cent of
C SNe (Maeda & Moriya 2022 ), we would expect to observe an
vent such as SN 2023tsz in its host approximately once in every
.5 Myr. This expectation could still be too frequent given that lower
ass star-forming regions, as expected in such a low star formation
ost, may underproduce the most massive stars owing to stochastic 
nitial mass function sampling (Stanway & Eldridge 2023 ). 
The only other SN Ibn that was disco v ered in a low SFR region
s PS1-12sk (Sanders et al. 2013 ). No host was found at PS1-
2sk’s location in the Hubble Space Telescope ( HST ) observations
Hosseinzadeh et al. 2019 ). We simulated an observation of the
ost of SN 2023tsz at the redshift of PS1-12sk to investigate if it
ould have been visible in the HST observations. This simulated 
bservation was done using the PYSYNPHOT package (STScI De- 
elopment Team 2013 ) into which we input the SED generated
sing CIGALE . The part of the SED is visible in the HST F 300 X
lter at z = 0 . 054, the redshift of PS1-12sk, is shown in Fig. 9 .
e obtained an apparent magnitude of 24.61 mag for the simulated
bservation, equating to a surface brightness of 26.90 mag arcsec −2 .
his surface brightness is higher than the 5 σ limit, 27.5 mag arcsec −1 
easured by Hosseinzadeh et al. ( 2019 ) meaning a host comparable
o the host of SN 2023tsz would have just been visible in their
bservations. 
 DI SCUSSI ON  
.1 Comparison of explosion parameters to other SN Ibn/Icn 
he decline rate and peak magnitude of SN 2023tsz are comparable to
he population of SNe Ibn (Fig. 3 ). When comparing the parameters
rom fitting the CSM + Ni model to the bolometric light curve of
N 2023tsz with those of other SNe Ibn (Pellegrino et al. 2022a )
nd SNe Icn (Pellegrino et al. 2022b ) using the same models, we
nd that our ejecta mass M ej = 0 . 59 + 0 . 69 −0 . 38 M  is comparable to their
alues when considering their associated uncertainties. Our value 
or the CSM mass M CSM = 0 . 05 + 0 . 05 −0 . 02 M  is lower than that found
or other SN Ibn fit using the same models with the exception of
N 2019wep, which had a comparable value when considering the 
ssociated errors. SN 2023tsz is on the upper end for decline rates
f SN Ibn, so this lower value of M CSM is expected. The large
ncertainties on our value of M ej are most likely due to our limited
hotometric co v erage of the rise and peak phases, meaning our model
tting is not well constrained at those epochs. The inferred 56 Ni
ass is higher than those found for the other SNe Ibn fit using the
ame model. We take our value for the 56 Ni produced as an upper
imit, as our final photometric observation lies below our model fit,
hich means it is consistent with the lower 56 Ni derived for other
Ne Ibn. The final photometric data point lying below the model
ould indicate incomplete trapping of the γ -rays from the decay of
6 Ni (Clocchiatti & Wheeler 1997 ). We were unable to obtain later
hotometry to help further constrain the M Ni due to SN 2023tsz
ecoming too faint for our instruments. In summary, the key 
arameters (Ni, CSM, and ejecta masses) of SN 2023tsz show general 
greement with those of other SNe Ibn when considering associated 
ncertainties. MNRAS 536, 3588–3600 (2025) 
3596 B. Warwick et al. 
M
Table 3. Key parameters of different SED fits to the host of SN 2023tsz, with different SFH burst age parametrizations. The chosen burst age for the best fit 
(i.e. the lowest reduced chi-squared) is given with the model. 
Model log(SFR) (M  yr −1 ) log(stellar mass) (M ) log(sSFR) ( yr −1 ) χ2 
Scenario 1 (burst age = 50 Myr) −2.58 7.06 −9.64 0.22 
Scenario 2 (burst age = 100 Myr) −2.83 7.08 −9.91 0.44 
Scenario 3 (burst age = 5 Myr) −2.95 6.64 −9.59 1.74 
Figure 10. Cumulativ e frequenc y plots showing the stellar mass (left), SFR (centre), and sSFR (right) of the host of SN 2023tsz (dashed line), SN Ibn hosts (9; 
dash–dot line), and SE SN hosts (9 SN Ibn and 118 SN Ib/c; solid line). The host properties are from Schulze et al. ( 2021 ). The percentage of supernovae hosts 
with a mass (left), SFR (centre), or sSFR (right) less than the host of SN 2023tsz is indicated in each plot. 
4
T  
s  
T  
i  
+  
s  
o  
T  
p  
I  
S  
e  
I  
s  
o  
e  
I  
a  
p  
S
4
T  
s  
O  
a  
p  
T  
E  
o  
t  
a  
d  
p  
p  
a  
i  
c  
i  
e  
fi  
fl  
a  
o  
f  
0  
G  
o  
c  
l  
c  
a
 
b  
t  
m  
d  
T  
p  
w  
d  
t  
p  
i  
l  
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/article/536/4/3588/7929872 by Turun Yliopisto user on 14 M
arch 2025.2 Spherical symmetry in SNe Ib/cn 
he polarimetry of SN 2023tsz is an important addition to the
ample of three other SNe Ibn/Icn with polarimetric observations.
he polarization factor of SN 2023tsz, P = 0 . 08 ± 0 . 31 per cent ,
s consistent with it having spherically symmetric photosphere at
 15 . 64 d post-peak. This observed spherically symmetric photo-
phere does not rule out a progenitor binary system, as at the time of
bservation the photosphere will have engulfed the binary system.
o date, there are only three other SNe Ibn/Icn with conclusive
olarimetry. SN Ibn 2023emq (Pursiainen et al. 2023b ) and SN
cn 2021csp (Perley et al. 2022 ) exhibited low polarization, while
N Ibn 2015G showed polarization up to ≈2 . 7 per cent , but the
xact value is unclear due to an uncertain, but possibly substantial,
SP contribution (Shivvers et al. 2017 ). While the sample is still
mall, most SNe Ibn/Icn appear to be consistent with a high degree
f spherical symmetry, but at most one appears to have possibly
xhibited an aspherical photosphere. Further polarimetry of SNe
bn/Icn will be needed to either confirm whether this one possible
spherical case is an outlier among a generally spherically symmetric
opulation or if there is a population of aspherical interacting
Ne. 
.3 Causes of asymmetric line profiles 
he blue asymmetry seen in SN 2023tsz’s + 70 . 2 d post-peak
pectrum could potentially be explained by a number of scenarios.
ne scenario is dust formation within the SN ejecta. Dust can partly
bsorb light from the far side of the ejecta, suppressing the redshifted
art of the emission lines. This effect has been observed in some
ype II SNe, such as SN 1987A and SN 1999em (Lucy et al. 1989 ;
lmhamdi et al. 2003 ) more than a year after explosion. It typically
nly appears at later times as the ejecta need to cool below the
hreshold for dust formation. At the epoch of the spectrum, we inferNRAS 536, 3588–3600 (2025)  blackbody temperature of T BB > 9000 K which is inconsistent with
ust formation in the ejecta. Ho we ver, gi ven the lo w ejecta mass, it is
ossible that SN 2023tsz quickly transitioned from the photospheric
hase to the nebular phase. If this is the case then the blackbody
pproximation no longer holds. The excess flux below 5500 Å seen
n the + 36.3 d spectrum, which is seen in other interacting SN at
omparable times (Perley et al. 2022 ), and is similar to the forest of
ron lines found in modelling the nebular phase of SE SNe (Dessart
t al. 2023 ), suggests this could be the case. As our temperature
tting is heavily dependent on our g-band observations, the excess
ux in this region could lead to a higher fitted temperature that is
ctually the case. If indeed the temperature is lower than that of
ur blackbody fit, then dust formation could have occurred. Dust
ormation has been observed in SN 2006jc and OGLE-2012-SN-
06 at a similar epoch to our observation (Mattila et al. 2008 ;
an, Wang & Liang 2021 ). These objects had near-infrared (NIR)
bservations that showed a second blackbody component that was
onsistent with thermal radiation from newly formed dust. As we
ack NIR observations, we cannot investigate whether this is the
ase in SN 2023tsz to confirm dust formation as a cause of the line
symmetry. 
Another possibility is that the unshocked CSM is causing the
lue asymmetry. If there is CSM in the line of sight, it will scatter
he photons from the receding ejecta (relative to our line of sight)
ore than the approaching ejecta, leading to the observed blue-
ominant asymmetry. A visual representation of this is given in
addia et al. ( 2020 ). Based on our polarimetric observation, the
hotosphere of SN 2023tsz is consistent with spherical symmetry,
hich implies a large fraction of the CSM likely follows a spherical
istribution, as expected of a WR star wind. This directly means that
here should be CSM in the line of sight, making this scenario a
ossibility. Ho we ver, unlike in Taddia et al. ( 2020 ), the asymmetry
n SN 2023tsz occurs after we see the broadening of the narrow-
ine features, caused by the receding of the photosphere revealing
SN 2023tsz: an SN Ibn in a very low-mass host 3597 
t  
t
o
 
b  
t
a
W  
p
w
4
T  
t  
w
t
w  
r  
c  
s
 
i  
1  
b
t
t  
p  
o  
t  
s
h  
s  
o  
g
a  
s  
f
t  
C  
2
u
o
l  
t  
s  
d
a  
t  
t
d
N
n  
o  
m
2  
i  
O
L
e  
t  
h  
2
w  
P  
S  
m  
w
H  
s
o  
h  
o
t  
m  
g
 
a  
S  
m
s
s  
i
a
A
p  
p
5
W
p  
t  
h  
p  
s  
r
fi
0  
c  
m
 
m  
0  
t  
t
a
i
n  
f  
t
a  
o  
l  
a  
o
w
o  
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/article/536/4/3588/7929872 by Turun Yliopisto user on 14 M
arch 2025he shocked CSM. If the asymmetry was due to the absorption of
he emission from the receding shocked CSM, then it should have 
ccurred with the broadening of the lines. 
It is also possible that the ejecta were asymmetric, but appeared to
e symmetric at + 15.6 d because the asymmetric region was within
he photosphere. At + 70.2 d the photosphere could have receded 
nd revealed this asymmetric ejecta, producing asymmetric lines. 
ithout polarimetry at the same epoch, we cannot rule out this
ossibility solely on our earlier observation of symmetry. However, 
ithout such observations, we also cannot confirm it. 
.4 Implications for SN Ibn progenitors and hosts 
he characterization study of Pastorello et al. ( 2015b ) found that, at
he time, all but one SNe Ibn were found in spiral galaxies. As they
ere typically found in star-forming regions this supported the idea 
hat their progenitors are massive stars. However, the one exception 
as PS1-12sk (Sanders et al. 2013 ). As PS1-12sk was found in a
egion with a low SFR it brought into question whether all SNe Ibn
ome from massive stars, and if not, then what is the mechanism that
trips their progenitors? 
SN 2023tsz is now the second SN Ibn that has been disco v ered
n a low SFR region comparable to PS1-12sk, although, unlike PS1-
2sk, SN 2023tsz has a visible host. If SN 2023tsz and PS1-12sk
oth come from non-massive star progenitors, then it could suggest 
hat alternate channels for SNe Ibn are more common than initially 
hought. Ho we ver, a lo w SFR does not rule out a massi ve star
rogenitor. The work by Hosseinzadeh et al. ( 2019 ) focused solely
n examining the SFR, excluding the sSFR due to the limitations of
heir HST observations. The sSFR is a better indicator of the current
tar formation of a galaxy than SFR alone, because sSFR reveals 
o w ef ficiently the galaxy is forming stars relati ve to its mass. The
SFR of the host of SN 2023tsz is in the 65th percentile for the hosts
f SE SN, as shown in Fig. 10 . This is not unexpected as lower mass
alaxies are observed to have higher sSFR than higher mass galaxies 
t low redshifts (Bauer et al. 2013 ; Geha et al. 2024 ). Based on the
SFR alone we cannot rule out a massive star progenitor. This is
urther supported by lower mass galaxies forming the majority of 
heir stars later than in more massive galaxies (Cowie et al. 1996 ;
imatti, Daddi & Renzini 2006 ; Thomas et al. 2019 ; Bellstedt et al.
020 ). 
Estimating the stellar metallicity of the host of SN 2023tsz 
sing a galaxy mass–metallicity relation (MZR) derived from the 
bservationally supported simulations of Ma et al. ( 2016 ) gives 
og ( Z ∗/ Z ) ≈ −1 . 6. The galaxies of comparable mass used to derive
he MZR come from Kirby et al. ( 2013 ), and have a range of
tellar metallicities from −1 . 3 to −1 . 7. For single stars, wind-
riven stripping is weaker at lower metallicity, making it inefficient 
t producing WRs (Mokiem et al. 2007 ), although rotation of
he star can help in this re gard (Me ynet & Maeder 2005 ). At
hese low metallicities, binary interactions are thought to be the 
ominant envelope stripping mechanism (e.g. Bartzak os, Moff at & 
iemela 2001 ), although more recent work suggests that this may 
ot necessarily be the case (Shenar et al. 2020 ). The binary fraction
f WR stars appears to be independent of metallicity down to the
etallicities of the Magellanic Clouds (Foellmi, Moffat & Guerrero 
003 ). Ho we ver, the inferred metallicity of the host of SN 2023tsz
s a factor ∼10 lower than that of the Small Magellanic Cloud.
bservational constraints on populations at these metallicities from 
ocal Group dwarfs are extremely difficult. The work of Foellmi 
t al. ( 2003 ) is based on only 12 sources, but it does suggest
hat WR stars are not ruled out as progenitors in these low-massosts, and consequently cannot be ruled out as a progenitor for SN
023tsz. 
Our simulated observation shows that the host of SN 2023tsz 
ould have just been visible in the observations of the region of
S1-12sk (Hosseinzadeh et al. 2019 ). Ho we ver, we note that SE
Ne have been discovered in hosts an order of magnitude less
assive (Schulze et al. 2021 ) than that of SN 2023tsz. Such a galaxy
ould plausibly not have been observable in the observations of 
osseinzadeh et al. ( 2019 ). We further note the similarity in larger
cale environment between SN 2023tsz and PS1-12sk: in the outskirts 
f a more massive local galaxy. Although the satellite nature of the
ost of SN 2023tsz is uncertain, such a location is to be expected
f faint, low-mass satellite galaxies. Given the low luminosity of 
hese galaxies it is possible that SNe in such hosts at higher redshifts
ay be misattributed to being highly offset from a more massive
alaxy. 
The exceptional nature of the host of SN 2023tsz then highlights
 need for a thorough investigation of the host environments of
Ne Ibn. It is not yet clear if they are over-represented in low-
ass, low-metallicity galaxies compared to other CC SN types. Any 
uch preference would require explanation by purported progenitor 
ystems. If a population of SNe Ibn exists in low-mass host galaxies,
t becomes important to determine whether SNe Ibn from more 
ctively star-forming regions share the same progenitor scenario. 
lternatively, the SNe Ibn population may consist of two separate 
rogenitor groups that give rise to SNe with the same observed
roperties. 
 C O N C L U S I O N  
e have presented the analysis of the photometric, spectroscopic, 
olarimetric, and host properties of SN 2023tsz, a typical SN Ibn
hat stands out due to its association with an exceptionally low-mass
ost galaxy. This study adds to the limited sample of SN Ibn with
olarimetric data, making it only the fourth such SN Ib/cn to be
tudied in this way. It is also the first SN Ibn to be disco v ered,
eported, and classified entirely by the GOTO collaboration. The key 
ndings of our analysis are summarized below. 
(i) The peak absolute magnitude of SN 2023tsz, M r = −19 . 72 ±
 . 04 mag, and its decline rate of 0 . 145 ± 0 . 002 mag d −1 are
onsistent with other known SNe Ibn, indicating it is a typical
ember of the class. 
(ii) Our modelling of the bolometric light curve using a CSM + Ni
odel resulted in inferred values of M CSM = 0 . 05 + 0 . 05 −0 . 02 M , M ej =
 . 59 + 0 . 69 −0 . 38 M , and M Ni = 0 . 04 + 0 . 01 −0 . 01 M , all of which are comparable
o other SN Ibn. The high initial decline rate of SN 2023tsz, compared
o other SNe Ibn, is consistent with it having a low CSM mass. 
(iii) At later times, the spectrum of SN 2023tsz showed blue 
symmetry in its prominent emission lines. While the exact cause 
s uncertain we suggest three plausible explanations. The first expla- 
ation is dust formation in a cool dense shell could absorb emission
rom the receding ejecta. Whilst our temperature fitting does not rule
his scenario out we cannot confirm it without contemporaneous NIR 
nd mid-infrared data. A second e xplanation involv es the scattering
f redshifted photons by CSM, ho we ver, the asymmetry appears
ater than we would expect for this explanation. The third is that we
re seeing asymmetric ejecta that were, at the time, below the earlier
bserved symmetric photosphere. Ho we ver, this cannot be confirmed 
ithout coe v al polarimetry. 
(iv) Polarimetric observations showed a polarization fraction 
f P = 0 . 08 ± 0 . 32 per cent at 15.6 d post-peak, suggesting aMNRAS 536, 3588–3600 (2025) 
3598 B. Warwick et al. 
M
s  
p
 
o  
o  
p  
e
 
o  
f
 
l  
p  
a  
s  
g  
b  
u
 
i  
r  
a
A
B  
S  
f
 
w  
r
 
G  
n  
w
 
(
 
S
 
g  
S
 
C  
M
 
3
 
s
 
s  
U  
o  
a  
t  
o  
h  
d  
U
 
p  
U  
U  
N  
I  
U  
S
 
S
 
D
 
s  
1
 
S
 
n
 
f  
t
 
s
 
A  
o  
T  
u
 
o  
0
 
A  
t  
N  
A
 
A  
p  
0  
0
 
u  
(  
S
 
o  
0
 
(
 
3
 
c
D
T  
d  
d  
a
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/article/536/4/3588/7929872 by Turun Yliopisto user on 14 M
arch 2025ymmetric photosphere. This is consistent with three of the four
olarimetric observations of SNe Ib/Icn to date. 
(v) From our SED fitting, the host galaxy of SN 2023tsz has SFR
f 2 . 64 × 10 −3 M  yr −1 , a stellar mass of 1 . 15 × 10 7 M , and a sSFR
f 2 . 29 × 10 −10 yr −1 . While its SFR and stellar mass are in the 2nd
ercentile for SE SN hosts, its sSFR is in the 65th percentile (Schulze
t al. 2021 ). 
(vi) We estimate a host metallicity of log( Z ∗/ Z ) ≈ −1 . 6 based
n the MZR, indicating that the progenitor system of SN 2023tsz is
rom an extremely low-metallicity environment. 
These findings show that SNe Ibn can occur in extraordinarily
o w-mass and lo w-metallicity galaxies. This environment poses
roblems for the stellar evolution of a single-stripped star progenitor,
lthough does not rule them out. An important question that remains
urrounding SNe Ibn: Are the y o v er-represented in these low-mass
alaxies, and if so, why? These host galaxies are often too faint to
e detected by contemporary sk y surv e ys, potentially skewing our
nderstanding of the host environments of SNe Ibn. 
The host of SN 2023tsz shows the need for further investigation
nto the occurrence rates of SNe Ibn in such hosts. An y o v er-
epresentation in lo w-mass, lo w-metallicity hosts will need to be
ddressed in future proposed progenitor scenarios. 
C K N OW L E D G E M E N T S  
W acknowledges the UK Research and Innovation’s (UKRI)
cience and Technology Facilities Council (STFC) studentship grant
unding, project reference ST/X508871/1. 
The authors w ould lik e to acknowledge the University of War-
ick Research Technology Platform (SCRTP) for assistance in the
esearch described in this paper. 
This work used observations from the Las Cumbres Observatory
lobal Telescope Network. Time on the Las Cumbres Observatory
etw ork w as provided via the Optical Infrared Co-ordination Net-
ork for Astronomy (OPTICON, proposal 23B030). 
JL, MP, and DO acknowledge support from a UKRI Fellowship
MR/T020784/1). 
DLC acknowledges support from the STFC, grant number
T/X001121/1. 
TLK acknowledges support via a Research Council of Finland
rant (340613; PI: R. Kotak), and from the STFC, grant number
T/T506503/1). 
For the purpose of open access, the author has applied a Creative
ommons Attribution (CC BY) licence to the Author Accepted
anuscript version arising from this submission. 
SM was funded by the Research Council of Finland project
50458. 
LK and LKN thank the UKRI Future Leaders Fellowship for
upport through the grant MR/T01881X/1. 
Based in part on observations made with the Nordic Optical Tele-
cope, owned in collaboration by the University of Turku and Aarhus
niversity, and operated jointly by Aarhus University, the University
f Turku, and the University of Oslo, representing Denmark, Finland,
nd Norway, the University of Iceland and Stockholm University at
he Observatorio del Roque de los Muchachos, La Palma, Spain,
f the Instituto de Astrofisica de Canarias. The NOT data presented
ere were obtained with ALFOSC, which is provided by the Instituto
e Astrofisica de Andalucia (IAA) under a joint agreement with the
niversity of Copenhagen and NOT. 
The Gra vitational-wa ve Optical Transient Observer (GOTO)
roject acknowledges the support of the Monash-Warwick Alliance,NRAS 536, 3588–3600 (2025) niversity of Warwick, Monash Univ ersity, Univ ersity of Sheffield,
niversity of Leicester, Armagh Observatory & Planetarium, the
ational Astronomical Research Institute of Thailand (NARIT),
nstituto de Astrof ´ısica de Canarias (IAC), University of Portsmouth,
niversity of Turku, and the STFC (grant numbers ST/T007184/1,
T/T003103/1, and ST/Z000165/1). 
DS acknowledges support from the STFC via grant numbers
T/T003103/1, ST/Z000165/1, and ST/X001121/1. 
JPA’s work was funded by the National Agency for Research and
evelopment (ANID), Millennium Science Initiative, ICN12 009. 
TP acknowledges the financial support from the Slo v enian Re-
earch Agency (grants I0-0033, P1-0031, J1-8136, J1-2460, and Z1-
853). 
PC was supported by the STFC (grants ST/S000550/1 and
T/W001225/1). 
RS was funded by a Leverhulme Research Project Grant. 
MJD was funded by the STFC as part of the GOTO project (grant
umber ST/V000853/1). 
MRM acknowledges a Warwick Astrophysics prize post-doctoral
ellowship made possible thanks to a generous philanthropic dona-
ion 
AS acknowledges the Warwick Astrophysics PhD prize scholar-
hip made possible thanks to a generous philanthropic donation. 
Based on observations collected at the European Organisation for
stronomical Research in the Southern Hemisphere, Chile, as part
f ePESSTO + (the advanced Public ESO Spectroscopic Surv e y for
ransient Objects Surv e y). ePESSTO + observations were obtained
nder ESO program ID 112.25JQ. 
AA acknowledges the Yushan Fellow Program by the Ministry
f Education, Taiwan for the financial support (MOE-111-YSFMS-
008-001-P1). 
This research has made use of data obtained from the High Energy
strophysics Science Archive Research Center (HEASARC) and
he Leicester Database and Archive Service (LEDAS), provided by
ASA’s Goddard Space Flight Center and the School of Physics and
stronomy, University of Leicester, UK, respectively. 
LG, TEM-B, and CPG acknowledge financial support from
GAUR, CSIC, MCIN, and AEI10.13039/501100011033 under
rojects PID2023-151307NB-I00, PIE 20215AT016, CEX2020-
01058-M, FJC2021-047124-I, 2021-BP-00168, and 2021-SGR-
1270. 
MN was supported by the European Research Council (ERC)
nder the European Union’s Horizon 2020 Framework Programme
grant agreement no. 948381) and by UK Space Agency grant no.
T/Y000692/1. 
TWC acknowledges the Yushan Fellow Program by the Ministry
f Education, Taiwan for the financial support (MOE-111-YSFMS-
008-001-P1). 
PO was supported by the STFC (grant number ST/W000857/1). 
RK acknowledges support from the Research Council of Finland
grant no. 340613). 
HK was funded by the Research Council of Finland projects
24504, 328898, and 353019. 
We would like to thank the anonymous re vie wer for their helpful
omments and feedback. 
ATA  AVAI LABI LI TY  
he underlying raw photometric, spectroscopic, and polarimetric
ata are available from the relevant data archives. The analysed
ata will be shared on reasonable request to the corresponding
uthor. 
SN 2023tsz: an SN Ibn in a very low-mass host 3599 
R
B
B
B
B
B  
B  
B
B
C  
C
C  
C
C
C
C
C  
C
d
D
D  
D
D
D
D  
D
E  
E
F
F  
F  
G
G
G  
G
G
G  
H
H
H
H
H  
K
K  
K
L
L  
 
M  
M  
M
M
M  
M
M  
M
M
N
N
P
P
P
P
P
P
P
P
P
P
P
P
P
R
S
S
S
S  
S
S  
S
S
S
S
S
S  
S
S
S
S  
S
T
T
T  
T
T
T
W
Y
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/article/536/4/3588/7929872 by Turun Yliopisto user on 14 M
arch 2025EFER ENCES  
artzakos P. , Moffat A. F. J., Niemela V. S., 2001, MNRAS , 324, 33 
auer A. E. et al., 2013, MNRAS , 434, 209 
ellstedt S. et al., 2020, MNRAS , 498, 5581 
londin S. , Tonry J. L., 2011, Astrophysics Source Code Library, 
record ascl:1107.001 
oquien M. , Burgarella D., Roehlly Y., Buat V., Ciesla L., Corre D., Inoue
A. K., Salas H., 2019, A&A , 622, A103 
otticella M. T. , Smartt S. J., Kennicutt R. C. J., Cappellaro E., Sereno M.,
Lee J. C., 2012, Mem. Soc. Astron. Ital. Suppl., 19, 158 
rown T. M. et al., 2013, PASP , 125, 1031 
ruzual G. , Charlot S., 2003, MNRAS , 344, 1000 
alzetti D. , Armus L., Bohlin R. C., Kinney A. L., Koornneef J., Storchi-
Bergmann T., 2000, ApJ , 533, 682 
enko B. , 2023, GCN Circ., 34633, 1 
haso vniko v A. et al., 2018, Transient Name Serv er Disco v ery Report, 2018-
1888, 1 
hatzopoulos E. , Wheeler J. C., Vinko J., 2012, ApJ , 746, 121 
imatti A. , Daddi E., Renzini A., 2006, A&A , 453, L29 
lark P. et al., 2020, MNRAS , 492, 2208 
locchiatti A. , Wheeler J. C., 1997, ApJ , 491, 375 
ovino S. et al., 2004, in Moorwood A. F. M., Iye M., eds, Proc. SPIE Conf.
Ser. Vol. 5492, Ground-based Instrumentation for Astronomy . SPIE, 
Bellingham, p. 1613 
owie L. L. , Songaila A., Hu E. M., Cohen J. G., 1996, AJ , 112, 839 
e Jong J. T. A. et al., 2015, A&A , 582, A62 
essart L. , Hillier D. J., Kuncarayakti H., 2022, A&A , 658, A130 
essart L. , Hillier D. J., Woosley S. E., Kuncarayakti H., 2023, A&A , 677,
A7 
ey A. et al., 2019, AJ , 157, 168 
ong Y. et al., 2024, ApJ , 977, 254 
raine B. T. , Li A., 2007, ApJ , 657, 810 
yer M. J. et al., 2022, in Marshall H. K., Spyromilio J., Usuda T., eds, Proc.
SPIE Conf. Ser. Vol. 12182, Ground-based and Airborne Telescopes IX. 
SPIE, Bellingham, p. 121821Y 
yer M. J. et al., 2024, SPIE, 13094, 130941X 
dge A. , Sutherland W., K uijken K., Driv er S., McMahon R., Eales S.,
Emerson J. P., 2013, The Messenger, 154, 32 
lmhamdi A. et al., 2003, MNRAS , 338, 939 
oellmi C. , Moffat A. F. J., Guerrero M. A., 2003, MNRAS , 338, 360 
 ole y R. J. , Smith N., Ganeshalingam M., Li W., Chornock R., Filippenko A.
V., 2007, ApJ , 657, L105 
 oreman-Macke y D. , Hogg D. W., Lang D., Goodman J., 2013, PASP , 125,
306 
aia Collaboration , 2023, A&A , 674, A1 
albany L. et al., 2018, ApJ , 855, 107 
al-Yam A. , 2017, in Alsabti A. W., Murdin P., eds, Handbook of Supernovae .
Springer, Cham, Switzerland, p. 195 
an W.-P. , Wang S.-Q., Liang E.-W., 2021, ApJ , 914, 125 
eha M. et al., 2024, ApJ, 976, 118 
odson B. et al., 2023, Transient Name Serv er Disco v ery Report, 2023-2419,
1 
eiles C. , 2000, AJ , 119, 923 
iramatsu D. , 2022, Am. Astron. Soc. Meeting, #240, 131.06 
iramatsu D. et al., 2023, ApJ , 955, L8 
osseinzadeh G. et al., 2017, ApJ , 836, 158 
osseinzadeh G. , McCully C., Zabludoff A. I., Arcavi I., French K. D.,
Howell D. A., Berger E., Hiramatsu D., 2019, ApJ , 871, L9 
aramehmetoglu E. et al., 2021, A&A , 649, A163 
irby E. N. , Cohen J. G., Guhathakurta P., Cheng L., Bullock J. S., Gallazzi
A., 2013, ApJ , 779, 102 
ochanek C. S. et al., 2017, PASP , 129, 104502 
iske J. et al., 2015, MNRAS , 452, 2087 ucy L. B. , Danziger I. J., Gouiffes C., Bouchet P., 1989, in Tenorio-Tagle
G., Moles M., Melnick J., eds, Structure and Dynamics of the Interstellar
medium. Springer-Verlag, Berlin, p. 164 
a X. , Hopkins P. F., Faucher-Gigu `ere C.-A., Zolman N., Muratov A. L.,
Kere ˇs D., Quataert E., 2016, MNRAS , 456, 2140 
cCully C. et al., 2018, LCOGT/banzai: Initial Release (0.9.4). Zenodo ( ht
tps:// doi.org/ 10.5281/ zenodo.1257560 ) 
aeda K. , Moriya T. J., 2022, ApJ , 927, 25 
artin D. C. et al., 2005, ApJ , 619, L1 
atheson T. , Filippenko A. V., Chornock R., Leonard D. C., Li W., 2000, AJ ,
119, 2303 
attila S. et al., 2008, MNRAS , 389, 141 
aund J. R. , Pastorello A., Mattila S., Itagaki K., Boles T., 2016, ApJ , 833,
128 
eynet G. , Maeder A., 2005, A&A , 429, 581 
okiem M. R. et al., 2007, A&A , 473, 603 
icholl M. , 2018, Res. Notes Am. Astron. Soc. , 2, 230 
icholl M. et al., 2023, ApJ , 954, L28 
astorello A. et al., 2007, Nature , 447, 829 
astorello A. et al., 2015a, MNRAS , 449, 1941 
astorello A. et al., 2015b, MNRAS , 449, 1954 
ellegrino C. et al., 2022a, ApJ , 926, 125 
ellegrino C. et al., 2022b, ApJ , 938, 73 
erley D. A. et al., 2020, ApJ , 904, 35 
erley D. A. et al., 2022, ApJ , 927, 180 
lanck Collaboration VI , 2020, A&A , 641, A6 
rentice S. J. et al., 2020, MNRAS , 499, 1450 
rochaska J. et al., 2020, J. Open Source Softw. , 5, 2308 
ursiainen M. et al., 2023a, A&A , 674, A81 
ursiainen M. et al., 2023b, ApJ , 959, L10 
ursiainen M. et al., 2023c, Transient Name Server Classification Report, 
2023-2442, 1 
oming P. W. A. et al., 2005, Space Sci. Rev. , 120, 95 
anders N. E. et al., 2013, ApJ , 769, 39 
chlafly E. F. , Finkbeiner D. P., 2011, ApJ , 737, 103 
chulze S. et al., 2021, ApJS , 255, 29 
erko wski K. , Mathe wson D. S., Ford V. L., 1975, ApJ , 196,
261 
happee B. J. et al., 2014, ApJ , 788, 48 
henar T. , Gilkis A., Vink J. S., Sana H., Sander A. A. C., 2020, A&A , 634,
A79 
hingles L. et al., 2021, Transient Name Server AstroNote, 7, 1 
hivvers I. et al., 2016, MNRAS , 461, 3057 
hivvers I. et al., 2017, MNRAS , 471, 4381 
ingh A. , 2021, Astrophysics Source Code Library, record ascl:2106.024 
martt S. J. et al., 2015, A&A , 579, A40 
mith N. , 2017, in Alsabti A. W., Murdin P., eds, Handbook of Supernovae .
Springer, Cham, Switzerland, p. 403 
mith K. W. et al., 2020, PASP , 132, 085002 
tanway E. R. , Eldridge J. J., 2023, MNRAS , 522, 4430 
teeghs D. et al., 2022, MNRAS , 511, 2405 
teele I. A. et al., 2004, in Oschmann J. M. J., ed., Proc. SPIE Conf. Ser. Vol.
5489, Ground-based Telescopes. SPIE, Bellingham, p. 679 
TScI Development Team , 2013, Astrophysics Source Code Library, 
record ascl:1303.023 
 addia F . et al., 2015, A&A , 580, A131 
 addia F . et al., 2020, A&A , 638, A92 
homas A. D. , K e wley L. J., Dopita M. A., Gro v es B. A., Hopkins A. M.,
Sutherland R. S., 2019, ApJ , 874, 100 
ominaga N. et al., 2008, ApJ , 687, 1208 
onry J. L. et al., 2018, PASP , 130, 064505 
ran H. D. , 1995, ApJ , 440, 565 
right E. L. et al., 2010, AJ , 140, 1868 
aron O. , Gal-Yam A., 2012, PASP , 124, 668 MNRAS 536, 3588–3600 (2025) 
3600 B. Warwick et al. 
M
A
T
D Instrument Grism R ( λ/
λ) Range ( Å) 
2 ALFOSC Gr4 360 3200–9600 
2 HFOSC Gr7/Gr8 500 3800–9250 
2 EFOSC2 Gr13 355 3685–9315 
2 OSIRIS + R1000B + R1000R 1018, 1122 3630–10000 
Error Surv e y 
0 .34 VIKING 
0 .05 KiDS 
0 .07 KiDS 
0 .04 KiDS 
0 .20 KiDS 
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
3
1
1
2
2
2
2
2
2
2
2
2
2
3
3
T
D
ow
nloaded from
 https://academPPENDIX  A :  TA BLES  
able A1. Spectroscopic time series of SN 2023tsz. 
ate MJD Phase (d) Telescope 
023-09-28 60216.23 + 4.2 NOT 
023-10-20 60237.95 + 25.9 HCT 
023-10-30 60248.28 + 36.3 NTT 
023-12-03 60282.19 + 70.2 GTC 
Table A2. Surv e y data of the host of SN 2023tsz. 
Filter Mag 
z 22 .20 
g 22 .79 
i 22 .31 
r 22 .53 
u 23 .37 NRAS 536, 3588–3600 (2025) 
 Department of Physics, University of Warwick, Gibbet Hill Road, Coventry CV4 7AL, UK 
 Institute of Space Sciences (ICE-CSIC), Campus UAB, Carrer de Can Magrans, s/n, E-08193 Barcelona, Spain 
 Institut d’Estudis Espacials de Catalunya (IEEC), E-08860 Castelldefels (Barcelona), Spain 
 Department of Physics & Astronomy, University of Turku, Vesilinnantie 5, FI-20014 Turku, Finland 
 Department of Physics, Lancaster University, Lancaster LA1 4YB, UK 
 European Southern Observatory, Alonso de C ´ordova 3107, Casilla 19, Santiago, Chile 
 Millennium Institute of Astrophysics MAS, Nuncio Monsenor Sotero Sanz 100, Off. 104, Providencia, Santiago, Chile 
 Graduate Institute of Astronomy, National Central University, 300 Jhongda Road, 32001 Jhongli, Taiwan 
 Jodrell Bank Centre for Astrophysics, Department of Physics and Astronomy, The University of Manchester, Manchester M13 9PL, UK 
0 Institute of Cosmology and Gravitation, University of Portsmouth, Portsmouth PO1 3FX, UK 
1 Astrophysics Research Cluster, School of Mathematical and Physical Sciences, University of Sheffield, Sheffield S3 7RH, UK 
2 Instituto de Astrof ´ısica de Canarias, E-38205 La Laguna, Tenerife, Spain 
3 Department of Particle Physics and Astrophysics, Weizmann Institute of Science, Rehovot 7610001, Israel 
4 School of Physics & Astronomy, Monash University, Clayton, VIC 3800, Australia 
5 Institute for Globally Distributed Open Research and Education (IGDORE), Box 1074, Kristineh ¨ojdsgatan 9A, Gothenburg, SE-412 82, Sweden 
6 Astronomical Observatory, University of Warsaw, Al. Ujazdowskie 4, PL-00-478 Warszawa, Poland 
7 Cardiff Hub for Astrophysics Research and Technology, School of Physics & Astronomy, Cardiff University, Queens Buildings, The Parade, Cardiff CF24 
AA, UK 
8 Finnish Centre for Astronomy with ESO (FINCA), University of Turku, FI-20014 Turku, Finland 
9 School of Sciences, European University Cyprus, Diogenes Street, Engomi, 1516 Nicosia, Cyprus 
0 Astrophysics Research Centre, School of Mathematics and Physics, Queens University Belfast, Belfast BT7 1NN, UK 
1 National Astronomical Research Institute of Thailand, 260 Moo 4, T. Donkaew, A. Maerim, Chiangmai 50180, Thailand 
2 School of Physics & Astronomy, University of Leicester, University Road, Leicester LE1 7RH, UK 
3 Center for Astrophysics and Cosmology, University of Nova Gorica, Vipavska 11c, 5270 Ajdov ˇs ˇcina, Slovenia 
4 Instituto de Alta Investigaci ´on, Universidad de Tarapac ´a, Casilla 7D, Arica, Chile 
5 Dipartimento di Fisica ‘Ettore Pancini’, Universit `a di Napoli Federico II, Via Cinthia 9, I-80126 Naples, Italy 
6 INAF – Osservatorio Astronomico di Capodimonte, Via Moiariello 16, I-80131 Naples, Italy 
7 Armagh Observatory & Planetarium, College Hill, Armagh BT61 9DG, UK 
8 Indian Institute of Astrophysics, II Block, Koramangala, Bengaluru 560034, Karnataka, India 
9 Hir oshima Astr ophysical Science Centre, Hir oshima Univer sity, 1-3-1 Ka gamiyama, Higashi-Hir oshima, Hir oshima 739-8526, Japan 
0 The Oskar Klein Centre, Department of Astronomy, Stockholm University, AlbaNova, SE-10691 Stockholm, Sweden 
1 Pondicherry University, R.V. Nagar, Kalapet, Pondicherry 605014, UT of Puducherry, India 
his paper has been typeset from a T E X/L A T E X file prepared by the author. 
© 2024 The Author(s). 
Published by Oxford University Press on behalf of Royal Astronomical Society. This is an Open Access article distributed under the terms of the Creative Commons Attribution License 
( https://cr eativecommons.or g/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 
ic.oup.com
/m
nras/article/536/4/3588/7929872 by Turun Yliopisto user on 14 M
arch 2025