The accretion of a solar mass per day by a

17-billion solar mass black hole

arXiv:2402.15101v1 [astro-ph.CO] 23 Feb 2024

Christian Wolf (christian.wolf@anu.edu.au)1,2 , Samuel Lai1 ,

Christopher A. Onken1 , Neelesh Amrutha1 , Fuyan Bian3 , Wei Jeat
Hon4 , Patrick Tisserand5 , and Rachel L. Webster4
1
Research School of Astronomy and Astrophysics, Australian National
University, Cotter Road Weston Creek, ACT 2611, Australia
2
Centre for Gravitational Astrophysics, Australian National University,
Building 38 Science Road, Acton, ACT 2601, Australia
3
European Southern Observatory, Alonso de Córdova 3107, Casilla
19001, Vitacura, Santiago 19, Chile
4
School of Physics, University of Melbourne, Parkville VIC 3010,
Australia
5
Sorbonne Universités, UPMC Univ Paris 6 et CNRS, Institut
d’Astrophysique de Paris, 98 bis bd Arago, F-75014 Paris, France

draft February 26, 2024

Abstract
Around a million quasars have been catalogued in the Universe by probing
deeper and using new methods for discovery. However, the hardest ones to find
seem to be the rarest and brightest specimen. In this work, we study the properties
of the most luminous of all quasars found so far. It has been overlooked until
recently, which demonstrates that modern all-sky surveys have much to reveal. The
black hole in this quasar accretes around one solar mass per day onto an existing
mass of ∼17 billion solar masses. In this process its accretion disc alone releases
a radiative energy of 2 × 1041 Watts. If the quasar is not strongly gravitationally
lensed, then its broad line region (BLR) is expected to have the largest physical
and angular diameter occurring in the Universe, and will allow the Very Large
Telescope Interferometer to image its rotation and measure its black hole mass
directly. This will be an important test for BLR size-luminosity relations, whose
extrapolation has underpinned common black-hole mass estimates at high redshift.

1
 In 1963, Maarten Schmidt identified the first quasar 1 , known as 3C 273. It appeared
as a remarkably bright star of 12th magnitude, while its redshift suggested that it was
among the most distant objects known in the Universe at the time. The two facts
together implied an implausibly huge output of light, and ever since then, newly found
quasars have impressed with their immense energy release from a small region of space.
This could only be explained as gravitational energy being converted into heat and light
within a highly viscous accretion disc around a supermassive black hole 2,3 (SMBH).
Quasars are signposts of fast growth in SMBHs on public display and allow the study
of these growth processes.
Finding large samples of quasars then provides population and growth statistics to
explain the origin of SMBHs in the Universe 4–7 . Generally, the most luminous quasars
contain the fastest-growing SMBHs, although the relation between mass accretion rate
and luminosity is affected by the mass and spin of the black hole as well as the structure
and viewing angle of the accretion disc and disc winds 8–14 .
Today, around a million quasars are known 15 , although some specimens stand out
from the crowd: in 2015, the ultra-luminous quasar J0100+2802 at redshift z = 6.3
was identified 16 with a SMBH of 10 billion solar masses 17 . In 2018, an even more
luminous object 18 was found, J2157–3602 at z = 4.7, with a SMBH of 24 billion
solar masses 14 . Although their luminosity implies rapid growth, their existence is hard
to explain: when black holes start from the remnant of a stellar collapse and grow
episodically within the Eddington limit, they are not expected to reach the evident
masses in the time from the Big Bang to the epoch of their observation, which has
triggered a search for alternative scenarios 19–22 .
While exceptionally rare, the most extremely luminous quasars are interesting for
several reasons beyond their intrinsic nature as discussed later. In this paper, we present
the properties of the recently discovered 23 quasar SMSS J052915.80–435152.0, here-
after J0529–4351, which is a 16th magnitude object at redshift z = 3.962 (see Fig. 1),
and reveal it to be the most luminous quasar currently known in the Universe (see
Fig. 2).

Results
When quasars appear extremely bright, it may be suspected that their observed bright-
ness is magnified by gravitational lensing from a massive galaxy on the line of sight.
Strong lensing causes multiple separate images of a quasar in the sky 24,25 . Two other
quasars with a redshift and apparent brightness similar to J0529-4351 are known to be
strongly magnified by lensing, the double-image APM 08279+5255 at z = 3.91, with
a separation of 0.5 arcsec 26 , and the quadruply imaged B 1422+231 at z = 3.62, with
separations up to 1.5 arcsec 27 . Estimated magnification factors for these two objects
range from 40 to 100, which implies that these quasars are not intrinsically extreme,
but are members of the bulk population 26,28 .
J0529-4351, in contrast, shows no sign of strong lensing; data from the European
Space Agency (ESA) Gaia satellite suggest it to be a point source, in terms of object
morphology and astrometric excess noise, which has been used to find dual quasars
or lensed quasars that appear unresolved to Gaia 29 (see Figure 2 and online methods

2
for more detail). We also find no strong foreground absorber system, which our high
signal-to-noise spectrum probes in Mg II at z > 1.15; the strongest system, at z =
2.118, has an equivalent width of EW(2796)=0.8Å; this suggests an impact parameter
of 20 kpc (or 2.4 arcsec) from the line of sight 30 to the quasar, while a plausible image
separation in a lensed scenario is ∼ 0.1 arcsec. We can also estimate the probability
of lensing a source at z ∼ 4 using common models for the galaxy distribution as
isothermal masses 31,32 , obtaining p ≈ 1.3 × 10−3 . An image separation of 0.2 arcsec
or less reduces it further to p ≈ 2 × 10−4 . These estimates change by only a factor
of a few when changing the input galaxy velocity dispersion function or the bright
end of the quasar luminosity function. Even if a very steep intrinsic quasar luminosity
function at the bright end enhanced the magnification bias, the probability that this
source is strongly lensed will be less than 1%. We thus take the strength of the quasar
emission at face value, although final confirmation from a high-resolution space-based
or adaptive-optics image would still be desirable.
We also investigate the recent history in the brightness of J0529–4351 to see whether
it may have been previously overlooked due to extreme variability. The 0.5-metre
NASA (National Aeronautics and Space Administration) Asteroid Terrestrial-impact
Last Alert System 33 (ATLAS) telescope provides a light curve since mid-2017 (see
Fig. 3). Brightness variations of ∼ 15% are found over the last six years, which are not
unexpected for luminous quasars. Earlier records, from photographic plates observed
in 1980 and 1998 (measured by the SuperCOSMOS Sky Surveys), found the R-band
brightness to be consistent with recent observations by the SkyMapper Southern Sur-
vey 34 (SMSS). The somewhat longer light curve from the Wide-field Infrared Survey
Explorer 35 (WISE) shows also only modest variability.
The quasar is undetected in the Rapid ASKAP Continuum Survey 36 (RACS), and
hence has a flux of less than 1µJy (4σ) in the broad 887 Mhz band. Using a common
definition of radio loudness 37 , we find the object to be safely in the radio-quiet regime
(R < 1). Thus, we have no reason to suspect that its luminosity is affected by jet
emission, let alone relativistically boosted.
For further analysis, we use a spectrum of optical and near-infrared light (see Fig. 4)
from the 8.2m Very Large Telescope (VLT) at the European Southern Observatory
(ESO). We split the spectrum with a publicly available spectrum-fitting code 38 into
an accretion disc continuum and emission-line contributions. From the disc contin-
uum we quantify the disc luminosity of the quasar and get a first proxy for the ac-
cretion rate. We find monochromatic luminosities of log(L135 /erg s−1 ) = 47.93 and
log(L300 /erg s−1 ) = 47.76 at restframe wavelengths of λ = 135 and 300 nm, re-
spectively. An approximate estimate of the full radiative output from the accretion
disc using standard bolometric corrections 39 and a 0.75 anisotropy factor 40 yields
log(Lbol /erg s−1 ) = 48.37. Using a standard radiative efficiency value 41 of 0.1, this
translates to an accretion rate of ∼ 413 solar masses per year. This result makes J0529–
4351 the most luminous quasar and by inference, the fastest-growing black hole in the
Universe known to date (in terms of mass growth per unit time).
A more refined analysis of the bolometric luminosity and accretion rate requires
modelling the possible spectra of the disc continuum over a grid of black-hole masses
and spins and simulating their observation for different accretion rates and viewing
angles 42–45 . Using slim-disc models as appropriate for SMBHs with high accretion

3
rates 46,47 and a publicly available code 14,48 (see online methods), we find a best-fitting
solution to our spectrum being an Eddington-accreting disc around an SMBH with a
moderate spin of a ≈ 0.4 viewed at an intermediate angle of i ≈ 45◦ (see Fig. 5).
The intrinsically emitted bolometric luminosity is log(Lbol /erg s−1 ) = 48.16, which
is 62% of the value derived above; the lower value mostly results from generic bolo-
metric corrections being overestimates for very massive black holes with their colder
discs and suppressed UV emission. The best-fit slim disc model yields an accretion
rate of 370 M⊙ per year and a radiative efficiency of ∼ 0.09. Given the lack of higher-
frequency data in the SED, the integrated model SED may still underpredict the bolo-
metric flux, while the standard bolometric corrections must be overcorrecting for black
holes of over a billion solar masses. The best estimate is expected to lie between the
two, and we thus choose the average of log(Lbol /erg s−1 ) = 48.27.
When modelling SEDs for a fixed mass accretion rate, different black-hole spins
lead to different bolometric corrections and radiative efficiencies, while the monochro-
matic UV-optical luminosities are only modestly affected. The main uncertainty arises
from the unknown viewing angle, which affects both the apparent monochromatic lu-
minosity and the derived accretion rate. Given the broad confidence intervals for spin
and viewing angle in the model fit, we consider the full plausible angle range from pole-
on to i = 60◦ as a 95% error margin; this corresponds to a 95% range in luminosity of
±0.12 dex and in accretion rate of 280 to 490 solar masses per year.

Black-hole mass
We estimate the mass of the black hole powering this quasar using two fundamentally
different methods:
(1) Assuming that the continuum emission is affected at its blue end by the inner
truncation of the accretion disc due to the innermost stable circular orbit around the
black hole 8 , we can infer a combined estimate for mass and spin of the black hole.
More massive black holes impose larger truncation radii and move the peak of the con-
tinuum emission to cooler temperatures and longer wavelengths 49 . With this method
the mass of the black hole in 3C 273 was found to be between 200 and 500 million
solar masses 50 , 35 years before a value of ∼300 million solar masses was measured
by interferometric observations 51 of the broad-emission line region in 3C 273. Since
then, the continuum-fitting method has not only proven useful for estimating masses
using thin disc models 43–45 but has also been applied 14 with slim discs models 46 that
are expected to be a more realistic description of the near-Eddington accretion discs
of fast-growing SMBH. From the continuum shape of J0529–4351, we find a mass of
log M/M⊙ = 10.28+0.17 −0.10 (or ∼ 19 billion solar masses).
(2) Assuming that the broad emission-line region represents virialised gas moving
at the velocity of Keplerian orbits around the SMBH, we can use the width of emission
lines and the continuum luminosity to infer the mass of the black hole. This method
is known as the virial single-epoch method 52,53 and has also been used to estimate the
mass of the SMBH in J0100+2802 17 at z = 6.3. Its application at the high-luminosity
end of the quasar population is mainly limited by a lack of independent calibrations for
the SMBH mass and relies on extrapolations from lower-luminosity quasars calibrated

4
with reverberation mapping 54 .
Our spectra offer two emission lines for this method, the triple-ionised carbon line
C IVλ1549 and the singly-ionised magnesium line Mg IIλ2799. The C IVλ1549 line
appears asymmetric and blueshifted relative to the Mg IIλ2799 line (∆v = −3120 ±
80 km/s) and a line full-width at half maximum (FWHM) of 7245 ± 175 km/s. The
Mg IIλ2799 line turns out to be hard to calibrate at this redshift because of atmospheric
absorption, but we do measure a FWHM of 4, 395±435 km/s. Using calibrations com-
monly applied to high-luminosity quasars, the line properties and luminosity translate
into SMBH mass estimates from log M/M⊙ = 10.03 ± 0.06 to 10.45 ± 0.02 for the
CIV line 53,55 and from log M/M⊙ = 10.03±0.09 to 10.36±0.09 for the MgII line 56,57
(errors are standard deviations due to propagated observational uncertainties, and the
scatter among the values is in line with large systematic calibration uncertainties). Ta-
ble 1 summarises all our estimates of masses and luminosities.
The line-based estimates are consistent with each other and with the continuum-
based estimate. Given that the systematic uncertainties in these methods are larger
than the statistical error propagation and may be as high as 0.4 dex, we combine the
mean mass estimates of the two lines and that from the SED without weighting into
a final result for the mass of log M/M⊙ = 10.24 ± 0.02 (simple mean and standard
deviation), or ∼ 17 billion solar masses. This also implies that the SMBH is accreting
near the Eddington limit (Eddington ratio of ∼ 0.9).

Discussion
In terms of luminosity and likely growth rate, J0529-4351 is the most extreme quasar
known. The accretion of J0529-4351 is near the Eddington limit, which is common
among quasars of the highest luminosity 16,58,59 . The growth rates are mostly uncertain
due to the unknown viewing angle. Assuming persistent accretion at its current Edding-
ton ratio, the mass doubling time is ∼ 30 Myr. However, with 19 billion solar masses,
the black hole in J0529-4351 at z = 3.962 is not the largest SMBH found in the most
luminous quasars. It has over 50% more mass than the black hole in J0100+2802 at
z = 6.3 but one third less mass than that in J2157-3602 at z = 4.7. The age of the
Universe at these three redshifts is 859 Myr, 1,244 Myr and 1,530 Myr, assuming a
flat Universe with a concordance cosmology (a cosmological constant of ΩΛ = 0.3).
Given that J0529-4351 is observed at a later epoch in the Universe than J2157-4351
and J0100+2802, it is less of a challenge to models of early SMBH growth.
It may be tempting to speculate on the spin of extreme SMBH: while it has been
argued that growing black holes should spin up over time, this would also increase
radiative feedback and slow down accretion, making it harder to grow the most massive
SMBH within the age of the Universe; instead, one way for growing black holes from
stellar seeds to the greatest masses we measure is “chaotic accretion” with randomly
changing orientation that keeps the black-hole spins and radiative feedback low 22 .
A long-standing question has been what mechanism fuels the high accretion rate,
which also must have persisted for some time already, though not necessarily in the
immediate past. Mass and kinematics of gas in the host galaxy of J0529-4351 could be
observed with the Atacama Large Millimetre Array (ALMA). ALMA has already re-

5
vealed the largest spiral galaxy in the early Universe 60 , in the quasar BRI 1335-0417 at
z = 4.4, which is ∼ 10× less luminous than J0529-4351. In contrast, ALMA observa-
tions have revealed a merger signature in the host galaxy of a quasar at z = 7.54 61 . Fur-
ther observations of extremely luminous quasars are progressing 62 . If extreme quasars
were caused by unusual host galaxy gas flows, ALMA should see this; if nothing un-
usual was found in the host gas, this would sharpen the well known puzzle of how to
sustain high accretion rates for long enough to form such extreme SMBHs.

Have we overlooked still more extreme quasars?

Finding rare and exceptionally bright quasars such as J0529-4351 does not require large
telescopes, but is instead a needle-in-the-haystack problem that needs precise data with
discriminative power across large areas of sky. Such objects are often hiding in plain
sight and are mistaken for stars. E.g., the quasar J1144-4308 at redshift z = 0.83 was
recognised as one of the brightest quasars in the night sky (at 14 mag) only in 2022,
despite having been imaged on photographic plates since the 19th century 63 .
In the 1960s finding the first quasars was driven by radio detections of objects that
otherwise appeared like regular nearby stars in our Milky Way Galaxy. Later quasar
surveys employed colour selection, template-fitting and Bayesian methods 64–72 . But
the main barrier for obtaining complete samples has always been contamination of
quasar candidate selections with stars from our Milky Way, which appear similar if
no discriminating information is available and vastly outnumber true quasars in the
extreme regime of the brightest quasars. With a million quasars known by now and
the wealth of data from modern all-sky surveys, machine-learning approaches are now
most popular 73–75 . These tend to get the classification right for the majority of objects,
while training samples imply that they perform less well for rare extreme cases. In the
case of extremely luminous quasars, the obvious bias of a training sample is simply
that they do not seem to exist until they are found.
A low-resolution spectrum of J0529-4351 revealing its quasar nature and redshift
has been part of the public all-sky Gaia DR3 data set published on 13 June 2022. The
machine-learning classification of this data set by the Gaia Discrete Source Classifier
(DSC) has assigned J0529-4351 a 99.98% probability to be a Milky Way star 76 , al-
though a human astronomer eyeballing the Gaia spectrum would recognise the quasar
and redshift at first sight.
These days, combining data from Gaia and WISE makes an all-sky search for bright
quasars straight-forward. An absence of parallax and proper motions removes most
of the bright Galactic stars from the search, and the mid-infrared photometry from
WISE is then sufficient to discriminate the disk continuum and dust emission of quasars
from the Rayleigh-Jeans tail of stellar photospheres 69 . The All-sky Bright Complete
Quasar Survey 23 (AllBRICQS), which reported J0529-4351, selects its candidates with
a simple heuristic WISE colour cut of W1−W2> 0.3 and an absence of > 4σ evidence
for stellar motion. While it aims for completeness and its recall of known quasars is
nearly perfect, its efficiency is still high, given that 97% of its candidates turn out to
be quasars 23 . Given that the AllBRICQS follow-up is complete at Rp < 16 in both
hemispheres, we doubt that a quasar of higher UV-optical luminosity will be found in

6
the future, unless it is hiding behind the Galactic Plane: AllBRICQS did not search at
Galactic latitude of |b| < 10◦ , where source crowding and higher dust extinction makes
the discovery of quasars still very challenging 77 .

Outlook: future observations with bright quasars

Extremely luminous quasars enable further observations for specific challenging quests:
after decades of instrumental development to increase the spatial resolution of tele-
scopic observations, the VLTI/GRAVITY Collaboration 51 made headlines with their
spatio-kinematic mapping of the broad-line region (BLR) in the iconic nearby quasar
3C 273 (redshift z = 0.157). This observation revealed the orbital rotation of the
disk-like BLR in 3C 273 in a spatially resolved pattern and thus provided a direct mea-
surement of the black hole mass from the BLR orbits. The measurement was possible
as 3C 273 was bright enough and displayed its Paschen-α emission line in GRAVITY’s
K-band window.
While the redshift of J0529-4351 imposed challenges for the work presented here,
as nearly all strong UV emission lines are in places heavily affected by atmospheric
absorption, it does place its Hβ line at 2.4µ in the K-band window of the GRAVITY
instrument. We predict the size of the Hβ-emitting broad-line region in J0529–4351 by
extrapolating the radius-luminosity relation 78 and find a radius of 2.2 pc. This implies
an angular diameter of 0.64 milli-arcseconds, which is an order of magnitude larger
than the BLR in 3C 273 and thus the largest-appearing quasar BLR in the Universe. We
expect the soon-to-be-upgraded VLTI/GRAVITY+ to obtain a superbly well-resolved
picture of the BLR rotation in J0529-4351 and thus a much more reliable measurement
of its black hole mass. Crucially, black-hole masses at the high end are estimated from
relations that have been extrapolated by orders of magnitude such that the whole scale
for objects like this is at risk. Getting a direct mass measurement for a black hole
with likely 50× the mass of the black hole in 3C 273 would be extremely valuable
for constraining the relations commonly used to estimate the masses of black holes in
the early Universe. The true mass scale of the earliest SMBH would also impact the
question of how hard exactly it is to form them.
Future plans include watching the Universe expand with repeat observations of
quasars lasting for a decade: the expansion shifts the redshift of individual gas clouds
observed as Lyman-α forest absorption lines in quasar spectra 79–81 (Sandage test). The
signal of such observing campaigns depends on the availability of a sufficient number
of very bright quasars and will still require advanced facilities such as the forthcom-
ing ESO Extremely Large Telescope (ELT). J0529–4351 will quite obviously be an
important part of this long-term endeavour.

7
Methods
Evidence for point-source geometry
While Gaia provides no images as such, we use the source characterisation from
its Data Release 3 to provide evidence for a point-source geometry in J0529-4351.
First, there is no noticeable flux outside an aperture with 0.175 arcsec radius around
the object centroid in the high-resolution imaging 82 , as determined by comparing the
fluxes between BP /RP and the G-band aperture 23 . The closest neighbouring object
is 3.2 arcsec away; it is not detected by Gaia, but by the DECam component of the
DESI Legacy Imaging Surveys 83 , with g = 23.581 ± 0.039, r = 22.319 ± 0.015,
i = 20.686 ± 0.007, and z = 20.099 ± 0.007 in DR10, which is a good match to the
colours of an M4 star in the Milky Way.
Second, astrometric excess noise is absent in J0529-4351, although 319 good ob-
servations were available to determine astrometric solutions. All Gaia parameters that
may hint at unresolved multiple sources, including frac_multipeak (0),
astrometric_excess_noise (0), astrometric_sigma5d_max (0.069),
ruwe (1.013), and ipd_gof_harmonic_amplitude (0.0167) are in line with the
average values (0.132, 0.042, 0.074, 1.015, 0.045) for bright quasars with Rp < 16 at
redshift z > 2 and far below any selection cuts used to search for Gaia-unresolved
binary stars 84 and dual or lensed quasars 29 . Figure 1 shows the astrometric excess
noise (AEN) of quasars at Rp < 18 including all dual and lensed quasars 25,29 with im-
age separations below 0.5 arcsec from the list at https://research.ast.cam.
ac.uk/lensedquasars/, that are unresolved by Gaia; even at the smallest known
image separation of 0.18 they show an (AEN)> 1.0 mas.
For the lensing calculation we followed a procedure used in the analysis in a re-
cently discovered high-redshift lensed quasar 32 , while adjusting the source redshift to
the z ∼ 4 of J0529-4351.

Spectroscopic observations and data processing

The optical/near-IR spectrum of J0529–4351 was obtained with the X-Shooter instru-
ment 85 on ESO’s Very Large Telescope (UT3) on 2023, Jan 14. The airmass at the time
of observation was 1.06 and the estimated seeing was about 1.5 arcsec. In the UVB,
VIS, and NIR arms, the slit widths and total exposure times were: 1.0 arcsec and 500 s;
0.9 arcsec and 400 s; and 0.9 arcsec and 800 s, respectively.
The data were reduced using PypeIt, a Python-based spectroscopic data reduction
pipeline 86 . Because it was not possible to a priori know the underlying true shape
of the C IVλ1549 emission line beneath the atmospheric A-band absorption, we could
not reliably fit the telluric model to the J0529–4351 VIS spectrum alone. To obtain
an improved correction, we took the telluric model generated from a standard star
observed on the same night and modified it by adopting a power-law intrinsic stellar
spectrum across a broad span of wavelengths unaffected by strong stellar atmospheric
lines. Applying this modified model to J0529–4351 yielded a final spectrum with a
smooth C IVλ1549 profile containing only narrow absorption lines and no spurious
emission from overcorrected telluric absorption.

8
 The observed spectrum was corrected for slit losses by calibration to external pho-
tometric data, in particular, the quasar’s VHS DR6 J-band magnitude of 14.812 ±
0.003 mag (Vega). With the VHS calibration anchoring the NIR arm, the VIS and
UVB data were sequentially matched to the spectroscopic flux levels in the wavelength
regions of overlap between the arms. This J-band calibration is consistent (within the
photometric errors) with that of the iSDSS X-Shooter acquisition image, as calibrated
by synthetic photometry for a neighbouring, non-variable star with Gaia low-resolution
spectroscopy, and the flux scale measured in the Gaia spectrum of J0529–4351 itself.

Spectral decomposition and emission-line fitting

The reduced and telluric-corrected X-Shooter spectrum is then transformed into the
rest-frame using a redshift of z = 3.962. We model the broad emission-line profile
of both the C IVλ1549 line and Mg IIλ2799 line using the code PyQSpecFit 38 , a
python-based spectral modelling package that is designed for rest-frame UV and op-
tical quasar spectra. We model each line individually, selecting independent windows
in wavelength to constrain the quasar continuum and the emission-line flux. We mask
the narrow absorption features present in our spectrum by applying a boxcar sigma-
clipping routine with a width of 50 pixels and a 3σ threshold.
For C IVλ1549, we constrain the continuum on either side of the feature with a
power-law model over line-free wavelength windows of 1445Å–1455Å and 1973Å–
1983Å. The emission-line model is composed of a maximum of three broad Gaussian
components with a minimum full-width at half-maximum (FWHM) of 1000 km s−1 .
We fit the C IVλ1549 profile between 1480Å–1528Å and 1537Å–1565Å, avoiding the
narrow absorption feature between 1528Å–1537Å. We also measure the monochro-
matic luminosity at 1450Å using the power-law model of the continuum. We estimate
the errors in the spectral fitting by creating 50 realisations of the spectrum after ran-
domly redistributing the flux data according to each point’s Gaussian uncertainties. The
standard deviation in each of the derived properties of the C IVλ1549 emission-line is
adopted as the statistical error.
For Mg IIλ2799, the continuum model is composed of a power-law and a template
of the UV flux contribution from broad Fe II emission, which is constrained across
the wavelength ranges 1973Å–1983Å, 2060Å–2340Å, 2600Å–2740Å, 2840Å–3100Å.
The emission-line model for Mg IIλ2799 is composed of up to one narrow and three
broad Gaussian components, where the narrow and broad distinction is set at a FWHM
of 1000 km s−1 . We fit the Mg IIλ2799 line between 2700Å–2870Å and the monochro-
matic luminosity as 3000Å is also measured from the continuum power-law model. We
derive the uncertainty in each of the line properties by adopting four different templates
of the UV Fe II emission 87–90 . As with C IVλ1549, we also resample the spectrum 50
times and measure the standard deviation in the Mg IIλ2799 line properties, summing
in quadrature with the uncertainty derived from Fe II templates. More information
about the PyQSpecFit modelling of C IVλ1549 and Mg IIλ2799 lines are detailed
in a study of the now second-most luminous quasar, SMSS J2157–3602 14 , and also
in a study of line properties in the luminous quasar sample of the European Southern
Observatory Large Programme XQ-100 59 , which is at a comparable redshift range to
J0529–4351.

9
Luminosity calculations
For calculating monochromatic continuum luminosities, we corrected the spectra for
extinction by dust in the Milky Way. We used the estimate of E(B − V ) = 0.041 ±
0.0027 from the Schlegel maps 91 with the correction factor 92 of 0.86 and the extinction
law by Fitzpatrick 93 , which yields restframe absorption values of A1450 = 0.074 mag
and A3000 = 0.025 mag. We then apply bolometric corrections for mean quasar
SEDs 39 of k1350 = 3.81 and k3000 = 5.15 and apply an anisotropy correction factor
of 0.75 assuming mean orientation 40 ). The accretion disc continuum fitting described
below determines the bolometric luminosity directly from the integrated disc model
SEDs and implies a spin- and inclination-dependent anisotropy correction. We then
take the average of the luminosity measurements.

Spectral decomposition and continuum fitting

We use a publicly available code 48 to model the shape of the accretion disc continuum
to spectral energy distribution models predicted for slim discs with slimbh 47 , which
is a grid of synthetic spectra from ray-traced numerical solutions of slim accretion disc
equations. The free parameters are mass and spin of the black hole as well as accre-
tion rate and disc inclination. A Markov chain Monte Carlo (MCMC) method is used
to map out likelihood contours across the 4D parameter space. We collect the Milky
Way extinction-corrected spectrum and extend the IR coverage with WISE photometry,
specifically the CatWISE2020 94 W1 and W2 photometric passbands. We then create
synthetic data points to represent the accretion disc flux within selected line-free win-
dows in the observed spectrum. For the WISE wavelengths, we adopt the spectrum
of the Selsing high-luminosity quasar template 95 , scaled independently to the flux of
W1 and W2 bandpasses, and create synthetic line-free data points in the same fashion.
Furthermore, we use the X-Shooter UVB arm to estimate the continuum flux through
the Lyα forest and create one more synthetic data point to constrain the continuum. We
then use MCMC to infer the Bayesian posterior probability distributions of the intrinsic
black hole properties using the set of synthetic data points to represent the flux of the
accretion disc continuum.
Further detailed information on the continuum fitting method is presented in an in-
depth study of the quasar SMSS J2157–3602 14 . However, unlike this previous study,
the static and smooth thermal accretion disc models are unable to fully reproduce the
hardening of the continuum in the VIS arm or the flux passing through the Lyα forest.
It has been shown previously that quasars with high C IVλ1549 blueshift, which are
indicative of strong outflows and winds, tend towards a bluer UV continuum to the
blue side of λrest ≈ 200 nm and a slightly redder continuum on the red side 96 , thus
producing a spectral break that is missing in the thermal disc models. Therefore, our
slimbh synthetic spectra utilises the disc atmosphere model, BHSPEC 97,98 , to help
reproduce the emerging Compton-hardened radiation.
We also attempted to fit the data with alternative models. E.g., we adopt kerrbb
thin disc models 99 despite the fact that the slim disc model reproduces the thin disc
SED at low Eddington ratios, and find indeed similar results for the mass of the black
hole and the luminosity of the disc, although the colour of the UV continuum is not

10
properly reproduced in the absence of Comptonisation considerations. We also exam-
ine the option of J0529–4351 being lensed by demagnifying the spectrum and find that
it does not improve the quality of fit.

Host galaxy dust extinction

We assume no dust extinction in the quasar host galaxy given that typical host redden-
ing levels are found to be consistent with E(B −V ) ≈ 0.0 among luminous quasars 100 ,
while less than 1% of quasars seem to have E(B −V ) > 0.1. Also, an object appearing
as the brightest object in the Universe has a low probability of being extinguished by
notable levels of dust. If dust were present, it would make the continuum redder and
the measured luminosity lower. This will lead to the black hole mass being overesti-
mated by the continuum-fitting method and underestimated by the virial single-epoch
method. Such a discrepancy is not observed here.

Light-curve construction
The light curve with photometry from the NASA (National Aeronautics and Space
Administration) Asteroid Terrestrial-impact Last Alert System 33 (ATLAS) 0.5m tele-
scope was obtained from the ATLAS website. For any object in the ATLAS observing
area, up to four observations per night are available, depending on weather. The orange
passband is observed in all clear nights, while the cyan passband is only used during
the half period around New Moon. For slowly varying objects, noise can be suppressed
by combining observations from longer periods; we determine median measurements
to reduce the influence of outlier measurements, using per-week intervals in the orange
passband and per-moon period intervals in the cyan passband. Error bars on the medi-
ans express the level of variability within the interval by showing inter-quartile ranges
of the values.

Data availability
Data from Data Release 3 (DR3) of the European Space Agency’s Gaia mission are
publicly available (https://gea.esac.esa.int/archive/). NASA ATLAS data are available
from https://fallingstar-data.com/forcedphot/. The SkyMapper Southern Survey data
are available from https://skymapper.anu.edu.au/ (doi:10.25914/5f14eded2d116). The
raw spectrum and calibration files from the ESO/VLT are available in the ESO archive
at http://archive.eso.org/. A reduced spectrum is available from the authors on reason-
able request.

Code availability
The spectral fitting code and the quasar continuum fitting code were written by SL in
python and are publicly available on github at https://github.com/samlaihei/PyQSpecFit 38
and https://github.com/samlaihei/BADFit 48 .

11
Acknowledgements
This work was supported by the Australian Research Council (ARC) through Discov-
ery Project DP190100252 (CW, FB, CAO, SL). SL is grateful to the Research School
of Astronomy & Astrophysics at Australian National University for funding his Ph.D.
studentship. We thank Giovanni Ferrami from the University of Melbourne for dis-
cussing solutions for strong gravitational lensing.
Data for this project were obtained at the European Southern Observatory through
DDT proposal 2110.B-5032.
This work has made use of data from the European Space Agency mission Gaia
(https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analy-
sis Consortium (DPAC, https://www.cosmos.esa.int/web/gaia/dpac/consortium). Fund-
ing for the DPAC has been provided by national institutions, in particular the institu-
tions participating in the Gaia Multilateral Agreement.
This publication makes use of data products from the Wide-field Infrared Survey
Explorer, which is a joint project of the University of California, Los Angeles, and the
Jet Propulsion Laboratory/California Institute of Technology, and NEOWISE, which
is a project of the Jet Propulsion Laboratory/California Institute of Technology. WISE
and NEOWISE are funded by the National Aeronautics and Space Administration.
SuperCOSMOS Sky Survey material is based on photographic data originating
from the UK, Palomar and ESO Schmidt telescopes and is provided by the Wide-Field
Astronomy Unit, Institute for Astronomy, University of Edinburgh.
This work made use of Astropy (http://www.astropy.org), a community-developed
core Python package and an ecosystem of tools and resources for astronomy 101 .
This research has made use of the SVO Filter Profile Service (http://svo2.cab.inta-
csic.es/theory/fps/) supported from the Spanish MINECO through grant AYA2017-
84089.
The national facility capability for SkyMapper has been funded through ARC LIEF
grant LE130100104 from the Australian Research Council, awarded to the University
of Sydney, the Australian National University, Swinburne University of Technology,
the University of Queensland, the University of Western Australia, the University of
Melbourne, Curtin University of Technology, Monash University and the Australian
Astronomical Observatory. SkyMapper is owned and operated by The Australian Na-
tional University’s Research School of Astronomy and Astrophysics. The survey data
were processed and provided by the SkyMapper Team at ANU. The SkyMapper node
of the All-Sky Virtual Observatory (ASVO) is hosted at the National Computational In-
frastructure (NCI). Development and support of the SkyMapper node of the ASVO has
been funded in part by Astronomy Australia Limited (AAL) and the Australian Gov-
ernment through the Commonwealth’s Education Investment Fund (EIF) and National
Collaborative Research Infrastructure Strategy (NCRIS), particularly the National eRe-
search Collaboration Tools and Resources (NeCTAR) and the Australian National Data
Service Projects (ANDS).
This work uses data from the University of Hawaii’s ATLAS project, funded through
NASA grants NN12AR55G, 80NSSC18K0284 and 80NSSC18K1575, with contribu-
tions from the Queen’s University Belfast, STScI, the South African Astronomical Ob-
servatory and the Millennium Institute of Astrophysics, Chile.

12
Author Contributions Statement
All authors contributed to data collection. SL led the data analysis with contributions
from CAO and CW. CW selected the quasar candidates and led the drafting and editing
of the article.

Competing Interests Statement

The authors declare no competing interests.

13
Table 1: Luminosity, black hole mass and accretion rate for the quasar J052915.80–
435152.0: given are best-fit estimates and 68% confidence intervals. Systematic uncer-
tainties for the mass estimates are proposed to be as high as 0.4 dex.

Luminosity log L/(erg s−1 )

λL135 47.93
λL300 47.76
Lbol (from 135 nm) 48.39
Lbol (from 300 nm) 48.35
Lbol (slim disc model) 48.16
Lbol (best estimate) 48.27 ± 0.06
Black hole mass log M/M⊙
C IVλ1549 line (average) 10.24 ± 0.15
Mg IIλ2799 line (average) 10.20 ± 0.08
Continuum fit (slim disc 47 ) 10.28+0.17
−0.10
All methods (best estimate) 10.24 ± 0.02
Mass accretion rate log Ṁ /(M⊙ yr−1 )
for plausible inclination range 2.57 ± 0.06

14
 Figure 1: The exceptional quasar J0529–4351: image from the Dark Energy Camera
Legacy Survey DR10 with size 20 × 20 arcsec2 ; North is up and East is to the left; a
neighbouring M star appears in red.

12
Lbol = 0.75 × k × L (erg s 1)

J0529-4351
13 Known z>1 quasars
1048 Lensed/dual quasars
14
RP (mag)

15

1047 AllBRICQS 3C 273 16

J0529-4351 J0100+2802
J1144-4308 XQ-100 17
J2157-3602 SDSS DR14Q
1046 18
0 1 2 3 4 5 6 10 2 10 1 100 101
Redshift Astrometric excess noise (mas)

Figure 2: J0529–4351 in relation to other quasars. Left: J0529–4351 vs the population

from SDSS and from XQ-100, which has the highest-quality spectra 59 for luminous
quasars at z ≈ 4. The luminosity axis shows monochromatic luminosity with standard
bolometric corrections applied. Right: Astrometric excess noise (AEN) of J0529–4351
compared to known bright quasars at z > 1; the known Gaia-unresolved lensed or dual
quasars are all at AEN> 1 mas. Zero AEN was replaced with a value of 0.01.

15
 Year
2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024

16.1

16.2 ATLAS-orange
ATLAS-cyan 0.51
Magnitude

12.44

12.48 W1 0.36
W2
57000 57500 58000 58500 59000 59500 60000
Modified Julian Date

Figure 3: Light curves of J0529–4351: data from NASA ATLAS are combined into
weekly (orange passband) or moon-period (cyan passband) median averages of typi-
cally eight data points. Data from WISE are median averages of typically 25 visits
completed within less than two weeks. Error bars show interquartile ranges within the
averaging period.

16
 1.0
Ly
Si IV +
O IV]
Luminosity Density (erg s 1 Å 1)
C IV
1045 C III] +
0.8 Si III]

0.6
1044
1000 1200 1400 1600 1800 2000
1045
0.4 Mg II

0.2

1044
0.0
2000
0.0 2500
0.2 3000
0.4 3500
0.6 4000
0.8 45001.0
Restframe Wavelength (Å)

140 60 Data
Residual ( ) f , obs (10 17erg s 1 cm 2 Å 1)

120 Power-law
50 Pseudo-Continuum
100 Line

80 40
60
40 30
20
10 5
0 0
10 5
1400 1600 1800 2200 2400 2600 2800 3000
Restframe Wavelength (Å) Restframe Wavelength (Å)
J0529-4351
AllBRICQS
Gaia BP/RP
Flux (arbitrary)

4000 5000 6000 7000 8000 9000 10000

Observed Wavelength (Å)

Figure 4: Spectrum of the J0529-4351 from the ESO 8.2m Very Large Telescope. Top:
observed-frame wavelength range of 0.42 − 2.27µm; a power law with αν = 0.5 is
shown as a dashed line for comparison. Centre: decomposition, emission-line fits, and
noise level (grey line, very low in the C IVλ1549 part): derived black hole masses are
log M/M⊙ = 10.24 for C IVλ1549 (left) and 10.20 for Mg IIλ2799 (right). Bottom:
earlier optical-only spectra from AllBRICQS and from Gaia DR3.

17
Figure 5: Results of the continuum fitting for J0529–4351 using MCMC and slim
disc models. Parameter estimates are given as peak of the probability distribution,
and confidence intervals are central 68-percentiles. The highest likelihood model is a
disc viewed at ∼45◦ angle around an intermediate-spin SMBH with ∼19 billion solar
masses and an accretion rate of 370 M⊙ per year; the constraints around viewing angle
and spin are, however, weak. Spin is denoted by a and Ldisc is in units of Eddington
luminosity.

18
References
[1] Schmidt, M. 3C 273 : A Star-Like Object with Large Red-Shift. Nature 197,
1040 (1963).

[2] Hoyle, F. & Fowler, W. A. Nature of strong radio sources. Nature 197, 533–535
(1963).
[3] Shakura, N. I. & Sunyaev, R. A. Black holes in binary systems. Observational
appearance. Astron. Astrophys. 24, 337–355 (1973).
[4] Soltan, A. Masses of quasars. Mon. Not. R. Astron. Soc. 200, 115–122 (1982).

[5] Boyle, B. J. et al. The 2dF QSO Redshift Survey - I. The optical luminosity
function of quasi-stellar objects. Mon. Not. R. Astron. Soc. 317, 1014–1022
(2000).
[6] Richards, G. T. et al. The sloan digital sky survey quasar survey: Quasar lumi-
nosity function from data release 3. Astron. J. 131, 2766–2787 (2006).
[7] Tucci, M. & Volonteri, M. Constraining supermassive black hole evolution
through the continuity equation. Astron. Astrophys. 600, A64 (2017).
[8] Bardeen, J. M. Kerr Metric Black Holes. Nature 226, 64–65 (1970).

[9] Davis, S. W. & Laor, A. The Radiative Efficiency of Accretion Flows in Indi-
vidual Active Galactic Nuclei. Astrophys. J. 728, 98 (2011).
[10] Dexter, J. & Agol, E. Quasar Accretion Disks are Strongly Inhomogeneous.
Astrophys. J. Lett. 727, L24 (2011).
[11] Slone, O. & Netzer, H. The effects of disc winds on the spectrum and black hole
growth rate of active galactic nuclei. Mon. Not. R. Astron. Soc. 426, 656–664
(2012).
[12] Campitiello, S., Ghisellini, G., Sbarrato, T. & Calderone, G. How to constrain
mass and spin of supermassive black holes through their disk emission. Astron.
Astrophys. 612, A59 (2018).

[13] Starkey, D. A., Huang, J., Horne, K. & Lin, D. N. C. Rimmed and rippled
accretion disc models to explain AGN continuum lags. Mon. Not. R. Astron.
Soc. 519, 2754–2768 (2023).
[14] Lai, S., Wolf, C., Onken, C. A. & Bian, F. Characterising SMSS J2157-3602, the
most luminous known quasar, with accretion disc models. Mon. Not. R. Astron.
Soc. 521, 3682–3698 (2023).
[15] Flesch, E. W. Identification confusion and blending concealment in the sdss-
dr16 quasar catalogues - 40 new quasars and 82 false quasars identified. Mon.
Not. R. Astron. Soc. 504, 621–635 (2021).

19
[16] Wu, X.-B. et al. An ultraluminous quasar with a twelve-billion-solar-mass black
hole at redshift 6.30. Nature 518, 512–515 (2015).
[17] Eilers, A.-C. et al. EIGER. III. JWST/NIRCam Observations of the Ultralumi-
nous High-redshift Quasar J0100+2802. Astrophys. J. 950, 68 (2023).

[18] Wolf, C. et al. Discovery of the most ultra-luminous qso using gaia, skymapper,
and wise. Publ. Astron. Soc. Aust. 35, e024 (2018).
[19] Volonteri, M. Formation of supermassive black holes. Astron. Astrophys. Rev.
18, 279–315 (2010).
[20] Volonteri, M., Silk, J. & Dubus, G. The Case for Supercritical Accretion onto
Massive Black Holes at High Redshift. Astrophys. J. 804, 148 (2015).
[21] Amarantidis, S. et al. The first supermassive black holes: indications from mod-
els for future observations. Mon. Not. R. Astron. Soc. 485, 2694–2709 (2019).
[22] Zubovas, K. & King, A. High-redshift SMBHs can grow from stellar-mass seeds
via chaotic accretion. Mon. Not. R. Astron. Soc. 501, 4289–4297 (2021).
[23] Onken, C. A. et al. AllBRICQS: The All-sky BRIght, Complete Quasar Survey.
Publ. Astron. Soc. Aust. 40, e010 (2023).
[24] Walsh, D., Carswell, R. F. & Weymann, R. J. 0957+561 A, B: twin quasistellar
objects or gravitational lens? Nature 279, 381–384 (1979).

[25] Lemon, C. et al. Gravitationally lensed quasars in Gaia - IV. 150 new lenses,
quasar pairs, and projected quasars. Mon. Not. R. Astron. Soc. 520, 3305–3328
(2023).
[26] Irwin, M. J., Ibata, R. A., Lewis, G. F. & Totten, E. J. APM 08279+5255: an
Ultraluminous Broad Absorption Line Quasar at a Redshift z = 3.87. Astrophys.
J. 505, 529–535 (1998).
[27] Patnaik, A. R., Browne, I. W. A., Walsh, D., Chaffee, F. H. & Foltz, C. B. B
1422+231 : a new gravitationally lensed system at z = 3.62. Mon. Not. R. Astron.
Soc. 259, 1P–4 (1992).

[28] Egami, E. et al. APM 08279+5255: Keck Near- and Mid-Infrared High-
Resolution Imaging. Astrophys. J. 535, 561–574 (2000).
[29] Chen, Y.-C. et al. Varstrometry for Off-nucleus and Dual Subkiloparsec AGN
(VODKA): Hubble Space Telescope Discovers Double Quasars. Astrophys. J.
925, 162 (2022).

[30] Chen, H.-W. et al. An Empirical Characterization of Extended Cool Gas Around
Galaxies Using Mg II Absorption Features. Astrophys. J. 714, 1521–1541
(2010).

20
[31] Mason, C. A. et al. Correcting the z ∼ 8 Galaxy Luminosity Function for Grav-
itational Lensing Magnification Bias. Astrophys. J. 805, 79 (2015).
[32] Yue, M., Fan, X., Yang, J. & Wang, F. Revisiting the Lensed Fraction of High-
redshift Quasars. Astrophys. J. 925, 169 (2022).

[33] Tonry, J. L. et al. Atlas: A high-cadence all-sky survey system. Publ. Astron.
Soc. Pac. 130, 064505 (2018).
[34] Onken, C. A. et al. SkyMapper Southern Survey: Second data release (DR2).
Publ. Astron. Soc. Aust. 36, e033 (2019).
[35] Wright, E. L. et al. The Wide-field Infrared Survey Explorer (WISE): Mission
Description and Initial On-orbit Performance. Astron. J. 140, 1868–1881 (2010).
[36] McConnell, D. et al. The Rapid ASKAP Continuum Survey I: Design and first
results. Publ. Astron. Soc. Aust. 37, e048 (2020).
[37] Kellermann, K. I., Sramek, R., Schmidt, M., Shaffer, D. B. & Green, R. VLA
Observations of Objects in the Palomar Bright Quasar Survey. Astron. J. 98,
1195 (1989).
[38] Lai, S. samlaihei/pyqspecfit: Pyqspecfit v1.0.0 (2023). URL https://doi.
org/10.5281/zenodo.7772752.
[39] Richards, G. T. et al. Spectral energy distributions and multiwavelength selec-
tion of type 1 quasars. Astrophys. J. Suppl. Ser. 166, 470–497 (2006).
[40] Runnoe, J. C., Brotherton, M. S. & Shang, Z. Updating quasar bolometric lumi-
nosity corrections. Mon. Not. R. Astron. Soc. 422, 478–493 (2012).
[41] Yu, Q. & Tremaine, S. Observational constraints on growth of massive black
holes. Mon. Not. R. Astron. Soc. 335, 965–976 (2002).
[42] Calderone, G., Ghisellini, G., Colpi, M. & Dotti, M. Black hole mass estimate
for a sample of radio-loud narrow-line Seyfert 1 galaxies. Mon. Not. R. Astron.
Soc. 431, 210–239 (2013).
[43] Capellupo, D. M., Netzer, H., Lira, P., Trakhtenbrot, B. & Mejı́a-Restrepo, J.
Active galactic nuclei at z ∼ 1.5 - I. Spectral energy distribution and accretion
discs. Mon. Not. R. Astron. Soc. 446, 3427–3446 (2015).
[44] Mejı́a-Restrepo, J. E., Lira, P., Netzer, H., Trakhtenbrot, B. & Capellupo, D. M.
The effect of nuclear gas distribution on the mass determination of supermassive
black holes. Nature Astronomy 2, 63–68 (2018).

[45] Campitiello, S., Celotti, A., Ghisellini, G. & Sbarrato, T. Estimating black hole
masses: Accretion disk fitting versus reverberation mapping and single epoch.
Astron. Astrophys. 640, A39 (2020).

21
[46] Abramowicz, M. A., Czerny, B., Lasota, J. P. & Szuszkiewicz, E. Slim accretion
disks. Astrophys. J. 332, 646–658 (1988).
[47] Sadowski, A. et al. Relativistic slim disks with vertical structure. Astron. Astro-
phys. 527, A17 (2011).

[48] Lai, S. samlaihei/badfit: Badfit v1.0.0 (2023). URL https://doi.org/

10.5281/zenodo.7772748.
[49] Laor, A. & Davis, S. W. Cold accretion discs and lineless quasars. Mon. Not. R.
Astron. Soc. 417, 681–688 (2011).
[50] Malkan, M. A. The ultraviolet excess of luminous quasars. II. Evidence for
massive accretion disks. Astrophys. J. 268, 582–590 (1983).
[51] Gravity Collaboration et al. Spatially resolved rotation of the broad-line region
of a quasar at sub-parsec scale. Nature 563, 657–660 (2018).
[52] McLure, R. J. & Jarvis, M. J. Measuring the black hole masses of high-redshift
quasars. Mon. Not. R. Astron. Soc. 337, 109–116 (2002).
[53] Vestergaard, M. & Peterson, B. M. Determining Central Black Hole Masses
in Distant Active Galaxies and Quasars. II. Improved Optical and UV Scaling
Relationships. Astrophys. J. 641, 689–709 (2006).
[54] Peterson, B. M. et al. Central Masses and Broad-Line Region Sizes of Ac-
tive Galactic Nuclei. II. A Homogeneous Analysis of a Large Reverberation-
Mapping Database. Astrophys. J. 613, 682–699 (2004).
[55] Coatman, L. et al. Correcting C IV-based virial black hole masses. Mon. Not.
R. Astron. Soc. 465, 2120–2142 (2017).

[56] Vestergaard, M. & Osmer, P. S. Mass Functions of the Active Black Holes in
Distant Quasars from the Large Bright Quasar Survey, the Bright Quasar Survey,
and the Color-selected Sample of the SDSS Fall Equatorial Stripe. Astrophys. J.
699, 800–816 (2009).
[57] Shen, Y. et al. A Catalog of Quasar Properties from Sloan Digital Sky Survey
Data Release 7. Astrophys. J. Suppl. Ser. 194, 45 (2011).
[58] Shen, Y., Greene, J. E., Strauss, M. A., Richards, G. T. & Schneider, D. P. Biases
in virial black hole masses: An sdss perspective. Astrophys. J. 680, 169–190
(2008).
[59] Lai, S. et al. Virial black hole mass estimates of quasars in the XQ-100 legacy
survey. Mon. Not. R. Astron. Soc. 526, 3230–3247 (2023).
[60] Tsukui, T. & Iguchi, S. Spiral morphology in an intensely star-forming disk
galaxy more than 12 billion years ago. Science 372, 1201–1205 (2021).

22
[61] Bañados, E. et al. The z = 7.54 Quasar ULAS J1342+0928 Is Hosted by a
Galaxy Merger. Astrophys. J. Lett. 881, L23 (2019).
[62] Bischetti, M. et al. The WISSH quasars project. IX. Cold gas content and envi-
ronment of luminous QSOs at z ∼ 2.4-4.7. Astron. Astrophys. 645, A33 (2021).

[63] Onken, C. A. et al. Discovery of the most luminous quasar of the last 9 Gyr.
Publ. Astron. Soc. Aust. 39, e037 (2022).
[64] Webster, R. L., Francis, P. J., Petersont, B. A., Drinkwater, M. J. & Masci, F. J.
Evidence for a large undetected population of dust-reddened quasars. Nature
375, 469–471 (1995).

[65] Warren, S. J., Hewett, P. C. & Foltz, C. B. The KX method for producing K-
band flux-limited samples of quasars. Mon. Not. R. Astron. Soc. 312, 827–832
(2000).
[66] Wisotzki, L. et al. The Hamburg/ESO survey for bright QSOs. III. A large flux-
limited sample of QSOs. Astron. Astrophys. 358, 77–87 (2000).

[67] Fan, X. et al. High-Redshift Quasars Found in Sloan Digital Sky Survey Com-
missioning Data. III. A Color-selected Sample at I∗ < 20 in the Fall Equatorial
Stripe. Astron. J. 121, 31–53 (2001).
[68] Wolf, C. et al. The evolution of faint AGN between z =˜1 and z =˜5 from the
COMBO-17 survey. Astron. Astrophys. 408, 499–514 (2003).
[69] Stern, D. et al. Mid-Infrared Selection of Active Galaxies. Astrophys. J. 631,
163–168 (2005).
[70] Richards, G. T. et al. Efficient Photometric Selection of Quasars from the Sloan
Digital Sky Survey. II. ˜1,000,000 Quasars from Data Release 6. Astrophys. J.
Suppl. Ser. 180, 67–83 (2009).
[71] Ivezić, Ž. et al. Optical selection of quasars: SDSS and LSST. In Mickaelian,
A. M. & Sanders, D. B. (eds.) Multiwavelength AGN Surveys and Studies, vol.
304, 11–17 (2014).

[72] Reed, S. L. et al. Eight new luminous z ≥ 6 quasars discovered via SED model
fitting of VISTA, WISE and Dark Energy Survey Year 1 observations. Mon.
Not. R. Astron. Soc. 468, 4702–4718 (2017).
[73] Calderone, G. et al. Finding the Brightest Cosmic Beacons in the Southern
Hemisphere. Astrophys. J. 887, 268 (2019).

[74] Yang, Q. & Shen, Y. A Southern Photometric Quasar Catalog from the Dark
Energy Survey Data Release 2. Astrophys. J. Suppl. Ser. 264, 9 (2023).
[75] Storey-Fisher, K. et al. Quaia, the Gaia-unWISE Quasar Catalog: An All-Sky
Spectroscopic Quasar Sample. arXiv e-prints arXiv:2306.17749 (2023).

23
[76] Delchambre, L. et al. Gaia DR3: Apsis III – Non-stellar content and source
classification. Astron. Astrophys. 674, A31 (2023).
[77] Fu, Y. et al. Finding Quasars behind the Galactic Plane. II. Spectroscopic Iden-
tifications of 204 Quasars at |b| < 20°. Astrophys. J. Suppl. Ser. 261, 32 (2022).

[78] Bentz, M. C. et al. The Low-luminosity End of the Radius-Luminosity Rela-

tionship for Active Galactic Nuclei. Astrophys. J. 767, 149 (2013).
[79] Sandage, A. The Change of Redshift and Apparent Luminosity of Galaxies due
to the Deceleration of Selected Expanding Universes. Astrophys. J. 136, 319
(1962).

[80] Liske, J. et al. Cosmic dynamics in the era of Extremely Large Telescopes. Mon.
Not. R. Astron. Soc. 386, 1192–1218 (2008).
[81] Cristiani, S. et al. Spectroscopy of QUBRICS quasar candidates: 1672 new
redshifts and a golden sample for the Sandage test of the redshift drift. Mon.
Not. R. Astron. Soc. 522, 2019–2028 (2023).

[82] Gaia Collaboration et al. Gaia Early Data Release 3. Summary of the contents
and survey properties. Astron. Astrophys. 649, A1 (2021).
[83] Dey, A. et al. Overview of the DESI Legacy Imaging Surveys. Astron. J. 157,
168 (2019).

[84] Fabricius, C. et al. Gaia Early Data Release 3. Catalogue validation. Astron.
Astrophys. 649, A5 (2021).
[85] Vernet, J. et al. X-shooter, the new wide band intermediate resolution spectro-
graph at the ESO Very Large Telescope. Astron. Astrophys. 536, A105 (2011).

[86] Prochaska, J. X. et al. Pypeit: The python spectroscopic data reduction pipeline.
Journal of Open Source Software 5, 2308 (2020).
[87] Vestergaard, M. & Wilkes, B. J. An Empirical Ultraviolet Template for Iron
Emission in Quasars as Derived from I Zwicky 1. Astrophys. J. Suppl. Ser. 134,
1–33 (2001).

[88] Tsuzuki, Y. et al. Fe II Emission in 14 Low-Redshift Quasars. I. Observations.

Astrophys. J. 650, 57–79 (2006).
[89] Mejı́a-Restrepo, J. E., Trakhtenbrot, B., Lira, P., Netzer, H. & Capellupo, D. M.
Active galactic nuclei at z ∼ 1.5 - II. Black hole mass estimation by means of
broad emission lines. Mon. Not. R. Astron. Soc. 460, 187–211 (2016).

[90] Bruhweiler, F. & Verner, E. Modeling Fe II Emission and Revised Fe II (UV)

Empirical Templates for the Seyfert 1 Galaxy I Zw 1. Astrophys. J. 675, 83–95
(2008).

24
 [91] Schlegel, D. J., Finkbeiner, D. P. & Davis, M. Maps of dust infrared emission
for use in estimation of reddening and cosmic microwave background radiation
foregrounds. Astrophys. J. 500, 525–553 (1998).
[92] Schlafly, E. F. & Finkbeiner, D. P. Measuring reddening with sloan digital sky
survey stellar spectra and recalibrating sfd. Astrophys. J. 737, 103 (2011).

[93] Fitzpatrick, E. L. Correcting for the effects of interstellar extinction. Publ.

Astron. Soc. Pac. 111, 63–75 (1999).
[94] Marocco, F. et al. The CatWISE2020 Catalog. Astrophys. J. Suppl. Ser. 253, 8
(2021).

[95] Selsing, J., Fynbo, J. P. U., Christensen, L. & Krogager, J. K. An X-Shooter

composite of bright 1 < z < 2 quasars from UV to infrared. Astron. Astrophys.
585, A87 (2016).
[96] Temple, M. J., Banerji, M., Hewett, P. C., Rankine, A. L. & Richards, G. T. Ex-
ploring the link between C IV outflow kinematics and sublimation-temperature
dust in quasars. Mon. Not. R. Astron. Soc. 501, 3061–3073 (2021).
[97] Davis, S. W., Blaes, O. M., Hubeny, I. & Turner, N. J. Relativistic Accretion
Disk Models of High-State Black Hole X-Ray Binary Spectra. Astrophys. J.
621, 372–387 (2005).

[98] Davis, S. W. & Hubeny, I. A Grid of Relativistic, Non-LTE Accretion Disk

Models for Spectral Fitting of Black Hole Binaries. Astrophys. J. Suppl. Ser.
164, 530–535 (2006).
[99] Li, L.-X., Zimmerman, E. R., Narayan, R. & McClintock, J. E. Multitempera-
ture Blackbody Spectrum of a Thin Accretion Disk around a Kerr Black Hole:
Model Computations and Comparison with Observations. Astrophys. J. Suppl.
Ser. 157, 335–370 (2005).
[100] Krawczyk, C. M. et al. Mining for dust in type 1 quasars. Astron. J. 149, 203
(2015).
[101] Astropy Collaboration et al. The Astropy Project: Sustaining and Growing a
Community-oriented Open-source Project and the Latest Major Release (v5.0)
of the Core Package. Astrophys. J. 935, 167 (2022).

25