MNRAS 527, 10867–10877 (2024) https://doi.org/10.

1093/mnras/stad1795
Advance Access publication 2023 June 22

Supermassive black holes in merger-free galaxies have higher spins which

are preferentially aligned with their host galaxy
R. S. Beckmann ,1 ‹ † R. J. Smethurst ,2 ‹ † B. D. Simmons ,3 A. Coil,4 Y. Dubois,5 I. L. Garland ,3
C. J. Lintott ,2 G. Martin ,6,7 S. Peirani5,8 and C. Pichon5
1 Institute of Astronomy and Kavli Institute for Cosmology, University of Cambridge, Madingley Road, Cambridge CB3 0HA, UK
2 Oxford Astrophysics, Department of Physics, University of Oxford, Denys Wilkinson Building, Keble Road, Oxford OX1 3RH, UK
3 Physics Department, Lancaster University, Lancaster LA1 4YB, UK
4 Center for Astrophysics and Space Sciences, University of California, San Diego, 9500 Gilman Dr, MC 0424, La Jolla, CA 92093-0424, USA
5 Institut d’Astrophysique de Paris/CNRS, 98 bis blvd Arago, F-75014 Paris, France

Downloaded from https://academic.oup.com/mnras/article/527/4/10867/7205294 by guest on 10 March 2024

6 Steward Observatory, University of Arizona, 933 N. Cherry Ave, Tucson, AZ 85719, USA
7 Korea Astronomy and Space Science Institute, 776 Daedeokdae-ro, Yuseong-gu, Daejeon 34055, South Korea
8 Université Cote
ˆ d’Azur, Observatoire de la Coteˆ d’Azur, CNRS, Laboratoire Lagrange, Bd de l’Observatoire, CS 34229, F-06304 Nice Cedex 4, France

Accepted 2023 June 9. Received 2023 April 25; in original form 2022 November 24

ABSTRACT
Here, we use the Horizon–active galactic nucleus (AGN) simulation to test whether the spins of supermassive black hole (SMBH)
in merger-free galaxies are higher. We select samples using an observationally motivated bulge-to-total mass ratio of <0.1, along
with two simulation-motivated thresholds selecting galaxies which have not undergone a galaxy merger since z = 2, and those
SMBHs with < 10 per cent of their mass due to SMBH mergers. We find higher spins (>5σ ) in all three sample compared
to the rest of the population. In addition, we find that SMBHs with their growth dominated by BH mergers following galaxy
mergers are less likely to be aligned with their galaxy spin than those that have grown through accretion in the absence of galaxy
mergers (3.4σ ). We discuss the implications this has for the impact of active galactic nucleus (AGN) feedback, finding that
merger-free SMBHs spend on average 91 per cent of their lifetimes since z = 2 in a radio mode of feedback (88 per cent for
merger-dominated galaxies). Given that previous observational and theoretical works have concluded that merger-free processes
dominate SMBH-galaxy co-evolution, our results suggest that this co-evolution could be regulated by radio mode AGN feedback.
Key words: methods: data analysis – methods: statistical – galaxies: bulges – galaxies: evolution – quasars: supermassive black
holes.

that develop following major or minor galaxy mergers (Walker,

1 I N T RO D U C T I O N
Mihos & Hernquist 1996; Hopkins, Quataert & Murray 2012; Tonini
The tight correlations observed between the mass of a galaxy’s et al. 2016, including those rare gas-rich mergers after which a
supermassive black hole (SMBH) with velocity dispersion (Magor- stellar disc reforms, e.g. Sparre & Springel 2017) can be constructed.
rian et al. 1998; Merritt & Ferrarese 2001; Hu 2008; Kormendy, Galaxies with bulge-to-total ratios <0.1 are thought not to have had
Bender & Cornell 2011; McConnell et al. 2011; van den Bosch a merger since at least z ∼ 2 (Martig et al. 2012; Martin et al. 2018)
2016) and stellar bulge mass (Marconi & Hunt 2003; Häring & Rix and can be selected observationally to isolate merger-free systems.
2004) have long been interpreted as evidence for the merger-driven Simmons, Smethurst & Lintott (2017) showed that 101 unobscured
co-evolution of SMBHs and galaxies. However, recent simulations AGN hosted by such disc-dominated galaxies with assumed merger-
have shown that less than 35 per cent of SMBH growth since z = free evolutionary histories had substantial SMBH masses, with the
3 is due to accretion triggered by galaxy mergers (possibly less majority lying well above (up to 2 dex) the stellar bulge mass–SMBH
than 15 per cent; Martin et al. 2018; McAlpine et al. 2020), which mass correlation that is observed for galaxies with significant merger
challenges the observational paradigm of merger-driven galaxy- histories, and lying on the typical total stellar mass–SMBH mass
SMBH co-evolution regulated by active galactic nucleus (AGN) correlation.
feedback. To study this observationally, a sample of host galaxies Follow-up studies of a subset of this disc-dominated sample by
with merger-free evolutionary histories, selected to lack the bulges Smethurst et al. (2019, 2021) revealed outflows ionized by the AGN
from these merger-free systems. AGN feedback is thought to be a key
regulator of co-evolution and considered necessary in cosmological
 E-mail: [email protected] (RSB); [email protected]
volume simulations employing Lambda cold dark matter (CDM),
(RJS) yet the role of AGN feedback in the absence of mergers is currently
† First-authorship is shared between Beckmann and Smethurst. unknown. Merger-free AGN growth and feedback are severely
© 2023 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://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium,
provided the original work is properly cited.
10868 R. S. Beckmann, R. J. Smethurst et al.
understudied, in part because the bulk of previous observational of outflows in purely secularly fed systems, only those with mixed
studies either explicitly examine systems with merger-dominated accretion histories.
growth histories (e.g. ULIRGs; Tadhunter et al. 2018; Perna et al. We note that this would be possible with the high resolution
2021) or focus on samples where the merger-driven and merger-free provided by the Hubble Space Telescope (HST) ramp filters used
accretion histories cannot be disentangled (i.e. a mixed morphology to isolate outflowing [O III] emission ionized by the AGN. However,
sample across the entirety of SDSS such as Rakshit & Woo 2018). since such data does not yet exist, and given the debate over whether
For this reason, the processes powering the SMBHs accretion that it is observationally possible to determine the SMBH spin with
leads to the AGN outflows revealed by Smethurst et al. (2019, X-ray reflection spectroscopy, we turn to simulations to test our
2021) is still poorly understood. Given that a flurry of new results hypothesis of non-merger–driven SMBH growth. Following the ideas
are suggesting that SMBHs grow predominantly through galaxy of Nayakshin et al. (2012) and the results of Smethurst et al. (2019),
merger-free processes (Simmons et al. 2017; Martin et al. 2018; we test whether SMBHs in merger-free systems have higher spin
Smethurst et al. 2019, 2021; McAlpine et al. 2020; and see also a magnitudes and are subsequently aligned with their galactic discs
companion paper to this work Smethurst et al. 2023), it is imperative using the Horizon-AGN simulation.2 Horizon-AGN is a modern,
that we understand the galaxy-merger-free mechanisms responsi- large-scale galaxy evolution simulation that evolves the evolution of

Downloaded from https://academic.oup.com/mnras/article/527/4/10867/7205294 by guest on 10 March 2024

ble for the majority of SMBH growth and the subsequent AGN a large sample of galaxies from cosmic dawn to redshift z = 0. It has
feedback. been shown to reproduce a wide range of observable properties of
One important property of SMBHs influenced by their co- the galaxy and BH population, such as the galaxy mass functions and
evolution with their host galaxy is the magnitude of the BH spin, cosmic star formation history, the BH mass and luminosity functions,
along with its orientation with respect to its host galaxy. The spin and correlations between BHs and their host galaxies such as the
magnitude determines the efficiency with which the SMBH converts BH-stellar mass relation. While Horizon-AGN did not track BH spin
accreted mass to feedback energy, while the spin direction determines evolution on the fly, BH spins were post-processed for all BHs in
where feedback energy is deposited if the SMBH is driving a jet. We Horizon-AGN and presented in Dubois et al. (2014a).
do have clear theoretical predictions for the expected distribution We describe the Horizon-AGN simulation in Section 2.1, the
of the spins and geometries of SMBHs in systems evolving in the calculation of BH spins in Section 2.2 and our galaxy sample
absence of mergers which could power such outflows. Nayakshin, selection in Section 2.3. Our results are shown and discussed in
Power & King (2012) discussed the implications of merger versus Section 3, and we summarize our conclusions in Section 5.
non-merger-driven accretion on to a black hole (BH). Merger-driven
accretion occurs chaotically, with material infalling towards the
SMBH at random angular momentum vectors, on average spinning 2 S I M U L A T I O N DA TA
down the BH (Berti & Volonteri 2008; Dotti et al. 2013). Conversely,
material inflowing to the SMBH in the centre of a non-merger grown 2.1 Horizon-AGN simulation
system will come from within the galactic disc at a constant angular
Horizon-AGN is a cosmological-volume hydrodynamical simula-
momentum vector, spinning up the SMBH to maximum (Nayakshin
tion, which has been described in detail in Dubois et al. (2014b).
et al. 2012; Dubois, Volonteri & Silk 2014a; Bustamante & Springel
Here, we only reiterate its most important features.
2019).1 The Bardeen-Petterson effect is then thought to realign the
Horizon-AGN was run using the adaptive mesh refinement code
spin of the BH according to that of the accretion disc formed from
RAMSES (Teyssier 2002), using a standard CDM cosmology with
infalling gas (and therefore the galactic disc in a system fed by
total matter density m = 0.272, dark energy density  = 0.728,
planar accretion; Bardeen & Petterson 1975). The time-scales of this
amplitude of the matter power spectrum σ 8 = 0.81, baryon density
process have been the subject of much discussion (see e.g. Rees
b = 0.045, Hubble constant H0 = 70.4 kms−1 Mpc−1 , and spectral
1978; Papaloizou & Pringle 1983; Scheuer & Feiler 1996; Fragile
index ns = 0.967 using WMAP-7 cosmology (Komatsu et al. 2011).
et al. 2007; Sorathia, Krolik & Hawley 2013) but are thought to be
The simulation box has a size of Lbox 100 h−1 Mpc (comoving) and is
much shorter than the lifetimes of outflows or jets (Natarajan &
refined on a root grid of 10243 . From here, cells are further adaptively
Pringle 1998). Theoretical understanding therefore suggests that
refined up to a maximum resolution of x = 1 proper kpc (level 17).
outflows will be produced perpendicular to galactic discs specifically
Cells are (de)refined when the mass in a cell is more (less) than
in isolated, merger-free systems, and produced at random orientations
8 times the initial mass resolution. The simulation has a DM mass
in systems that have undergone a merger.
resolution of MDM =∼ 8.27 × 107 M , and includes prescription
There is very little agreement in the literature over the accuracy
for gas cooling including the contribution from metal released by
of methods that claim to be able to probe the spin of SMBHs
SN feedback, background UV heating, star formation, and stellar
observationally (e.g. using X-ray reflection spectroscopy to probe
feedback. Star formation is modelled according to a Schmidt law
the spin; see review by Reynolds 2014). In addition, although the
with a 1 per cent efficiency, using a Salpeter initial mass function.
orientation of outflows has been studied extensively in the literature
Stellar feedback is modelled to include stellar winds, type Ia and
with many studies finding no correlation with disc orientation (e.g.
type II supernovae (Dubois & Teyssier 2008; Kimm et al. 2015). The
see Kinney et al. 2000; Schmitt et al. 2003; Ruschel-Dutra et al.
minimum stellar resolution is ∼ 2 × 106 M .
2021), there has not yet been an observational study on the orientation
BHs are created at z > 1.5 in cells that exceed the density threshold
for star formation (n0 = 0.1 Hcm−1 ) with an initial seed mass of
105 M . To avoid multiple BHs forming in the same galaxies, a
1 Interestingly,Nayakshin et al. (2012) concluded from this theoretical con- 50 comoving kpc exclusion zone for new BH formation is enforced
sideration that secularly grown SMBHs should be overmassive in comparison around each existing BH. BH accretion and feedback is modelled
to merger grown BHs, which they pointed out was in contradiction to
measurements of SMBH masses in galaxies with pseudo-bulges available
at the time (e.g. Kormendy et al. 2011). However, the work of Simmons et al.
(2017) and Martin et al. (2018) has significantly eased this contradiction. 2 https://www.horizon-simulation.org/

MNRAS 527, 10867–10877 (2024)

 High-spin SMBHs in merger-free galaxies 10869
as in Dubois et al. (2012). BHs accrete gas via the Bondi–Hoyle– merger remnant is computed from the properties of the primary and
Lyttleton formalism ṀBH = 4π αG2 MBH 2
ρ̄/(c̄2s + ū2 )3/2 , where MBH secondary BH following Rezzolla et al. (2008).
is the BH mass, G is the gravitational constant, ρ̄, c̄s and ū are The galaxy spin axis is measured by computing the angular
the average gas density, sound speed and gas velocity, and α is a momentum vector agal of all star particles associated with the galaxy.
dimensionless boost factor. We set α = n/n0 if the gas number density The angle, φ gal,BH , between the vectors defining the galaxy spin, agal ,
n > n0 , and α = 1 otherwise (Booth & Schaye 2010). Accretion and SMBH spin, aBH , in {x, y, z} Cartesian coordinates within the
on to the BH is limited at the Eddington accretion rate ṀEdd . The simulation, was calculated as in the inverse cosine of the dot product
Eddington ratio is defined to fedd = ṀBH /ṀEdd ≤ 1, and measures of the two vectors divided by the product of their magnitudes:
the efficiency of BH accretion. BH spin is not followed throughout the  
agal .aBH
simulation and is instead computed in post-processing (see Section cos φgal,BH = . (1)
2.2). |agal ||aBH |
AGN feedback energy is released at a rate of ĖAGN = f r ṀBH c2
where r = 0.1 is the assumed radiative efficiency and c is the
speed of light. When fedd > 0.01, the AGN is quasar mode, and 2.3 Galaxy sample selection

Downloaded from https://academic.oup.com/mnras/article/527/4/10867/7205294 by guest on 10 March 2024

energy is injected isotropically as thermal energy with coupling Galaxies and dark matter halos in Horizon-AGN were identified
efficiency f = 0.15. If fedd ≤ 0.01, the AGN is in radio mode, using ADAPTAHOP (Aubert, Pichon & Colombi 2004; Tweed et al.
where energy is released in bi-conical outflows, f = 1, and drive 2009), using a local density threshold of ρ t = 178 times the average
powerful jets when 10−4 < fedd < 10−2 . BHs in Horizon-AGN are dark matter density, the 20 closest neighbours and a force softening
able to move freely within their host galaxy, rather than being pinned of 2 kpc. A minimum stellar mass cut of M∗ =∼ 109 M was applied,
to their centre. To account for un-resolved dynamical friction forces, where M∗ is the total mass of all star particles associated with a given
a subgrid force of magnitude FDM = fgas 4π αρ(GMBH /c̄s )2 is added galaxy as identified by ADAPTAHOP. Bulge masses are computed as
following Ostriker (1999). fgas = 0−2 (Chapon, Mayer & Teyssier in Volonteri et al. (2016), using a first Sersic profile with n = 1 for
2013) is a factor that depends on the Mach number. BHs merge when the disc component, and a second Sersic profile with the best fit of
located within 4 kpc of each other, and when their relative velocity n = 1, 2, 3, or 4 for the bulge component.
is smaller than the escape velocity of the binary. To identify BHs with galaxies for analysis, we combine two
spatial criteria: to be assigned to a galaxy, a BH must be located
2.2 Computing SMBH spin evolution within 10 per cent of the galaxy’s DM host halo virial radius, and
simultaneously within two effective radii of the galaxy. Galaxy
BH spin was not included in the Horizon-AGN run. Instead, it effective radii are computed by taking the geometric mean of the
is post-processed from quantities recorded at each coarse time- half-mass radius of the projected stellar densities along each of the
step throughout the simulation. For our analysis, we use the spins simulation’s Cartesian axes. If several BHs meet both criteria for a
calculated in Dubois et al. (2014a), using an analytic model that given galaxy, the most massive BH is retained (see Volonteri et al.
estimates the SMBH spin evolution using gas quantities at the 2016, for details) and all other (i.e. secondary or wandering) BHs
resolution scale of the simulation, as well as information on BH- are discarded from the data set analysed here. Galaxy mergers were
BH mergers. We briefly reiterate the key features of the model here identified using the DM halo merger trees constructed from the halo
but refer readers to Dubois et al. (2014a) for details. catalogues for z < 6, which are built from snapshots that are spaced
All BHs are assumed to form with zero spin (BH spin parameter on average every 130 Myr. Galaxy mergers are classified into major
|a| = 0). The evolution of BH spin direction and magnitude is (stellar mass ratios >1:4) and minor (stellar mass ratios 1:4 to 1:10)
integrated throughout its evolution using the BH accretion rate, galaxy mergers (see Martin et al. 2018, for details). During a BH
instantaneous BH mass and BH angular momentum JBH , and the merger, the less massive BH is considered to merge into the more
angular momentum vector of the accreted gas at each time-step massive one, whose identifier is retained.
recorded throughout the simulation. In brief, the model works as We identified a sample of 6851 galaxies at z = 0.0556, (the average
follows. At each time-step, the angular momentum of the gas Jgas redshift of the observed merger-free ‘bulgeless’ galaxy sample of
to be accreted is measured from the simulation within the accretion Simmons et al. 2017 for ease of comparison) from the Horizon-
region of the BH. This measured angular momentum is only used AGN simulation, which had central SMBH and for which bulge
to determine the direction of angular momentum accreted by the masses were computed. From this sample of galaxies, we selected
BH, not the magnitude. If the angular momentum of the gas and BH sub-samples using three different criteria:
are misaligned, the gas excerpts a torque on the BH spin through
the Bardeen–Petterson effect (Bardeen & Petterson 1975). In this (i) A galaxy merger-based criterion, which selects galaxies ac-
case, the inner accretion disc will warp and both BH spin and disc cording to the total number of galaxy mergers (both major, mass ratios
momentum will reorient until JBH and Jgas are aligned (or anti- >1:4, and minor, mass ratio 1:10 − 1:4) since redshift z = 2. The
aligned) with Jtot = JBH + Jgas , which determines the new direction threshold of z = 2 is motivated by the need for hierarchical structure
of the spin vector of the BH. At this point, BH and accretion disc can formation at earlier epochs, as per CDM, and is observationally
be either aligned or anti-aligned, which we determine using (King motivated by the studies of Martig et al. (2012); Martin et al. (2018)
et al. 2005). who showed that galaxies with low bulge-to-total ratios have had no
The change in BH spin magnitude due to angular momentum major mergers since at least z ∼ 2. Galaxy mergers were identified
accreted from the disc on to the BH is computed following Bardeen using the DM halo merger tree in Horizon-AGN. We combined the
(1970). Each accretion event is treated as the formation of a new number of major and minor mergers into a single parameter, Nmerger ,
accretion disc. We do not model the spin-down of BHs through to quantify the total number of mergers a galaxy has undergone
the Blandford–Znajek mechanism (Blandford & Znajek 1977), since z = 2. We use this to select a sample of galaxies which have
during which BHs can be spun down as spin energy is extracted been evolving in isolation with Nmerger = 0, which resulted in 1781
by BH-driven jets. During BH–BH mergers, the BH spin of the (26 per cent) galaxies. We compare these to galaxies that have had

MNRAS 527, 10867–10877 (2024)

10870 R. S. Beckmann, R. J. Smethurst et al.
their history dominated by galaxy mergers, with Nmerger ≥ 3 (2117 statistical significance in a Kolmogorov–Smirnov test in all cases).
galaxies, ∼ 31 per cent). This is also apparent for total and bulge stellar masses, with non-
(ii) A BH merger-based criterion, which applies cuts to fBH,merge , merger grown galaxies having lower stellar masses across Figs 1–3.
the cumulative mass fraction of a BH gained through BH mergers. Non-merger grown SMBHs selected on galaxy B/T and number
For an equal mass BH-BH merger, fBH,merge = 0.5 while for a 1:4 of galaxy mergers are confirmed to have low fBH,merge (see Figs 1 and
mass ratio merger, fBH,merge = 0.2. fBH,merge can exceed 0.5 if a BH 2). However, those galaxies selected with fBH,merge < 0.1 do not
undergoes repeated mergers. No lower mass ratio cut is applied uniformly have low bulge-to-total mass ratios. The distribution in
to mergers tracked in this way. We use this to select a sample of Fig. 3 is bimodal, with a peak at low and high B/T ratio, either side
SMBHs that have evolved with minimal BH mergers with fBH,merge of the peak for the rest of the population. This suggests that galaxies
< 0.1, which resulted in 2137 (∼ 31 per cent) galaxies. Note that that have grown their SMBHs without BH mergers ( 10 per cent by
1609 galaxies in the simulation had fBH,merge = 0, with no mass mass) can have significant bulges, perhaps grown by disc instabilities,
contribution from mergers, and 3508 galaxies (∼ 51 per cent of the although we caution that at ∼ 1 kpc resolution, disc instabilities
Horizon-AGN population at z = 0.0556) have fBH,merge <0.25 and so are likely to suppressed in HORIZON-AGN.3 Using a semi-analytic
have had their SMBH growth dominated by non-merger evolution. model, Parry, Eke & Frenk (2009) show in the Millennium simulation

Downloaded from https://academic.oup.com/mnras/article/527/4/10867/7205294 by guest on 10 March 2024

We note that while there is a link between galaxy mergers and SMBH that bulge growth is dominated by disc instabilities4 for galaxies with
mergers, one of the reasons that the galaxy-merger based sample and total stellar masses < 1011 M , with mergers only dominating for the
the SMBH-merger based sample are not identical is that numerical most massive galaxies. This leads to secular growth of both bulges
SMBH mergers can be significantly delayed in comparison to the and SMBHs, resulting in the distribution of B/T for non-merger
merger of their host galaxies, even without considering further delays grown systems seen in Fig. 3.
due to processes on scales that are unresolved in HORIZON-AGN The distribution of the SMBH spin magnitudes are >5σ different
(Volonteri et al. 2020). for each of the three selections of non-merger grown systems. It is
(iii) An observationally motivated bulge to total (B/T) cut to most apparent for those galaxies selected using fBH,merge < 0.1; this
produce a disc-dominated sample, based on the ratio between the is unsurprising since this traces the true amount of mass in a SMBH
galaxy bulge mass Mbulge and the total galaxy stellar mass M∗ . grown by BH mergers (since B/T and Nmerge are both proxies for
We use this to select a sample of galaxies with assumed galaxy merger SMBH growth). The spins of non-merger grown SMBHs are
merger-free histories with B/T < 0.1 (as is done observationally, e.g. maximal (confirming the hypothesis of Smethurst et al. 2019, which
Simmons et al. 2013; Simmons et al. 2017) which resulted in 179 built on the work of Nayakshin et al. 2012), in agreement with the
(∼ 3 per cent) galaxies. This small sample size is a combination results of Dubois et al. (2014a) and Bustamante & Springel (2019).
of the low numbers of truly disc-dominated galaxies produced in We note that these results hold when incomplete, total stellar mass-
simulations (due to disc instability triggered bulge formation), and matched samples of merger-free and merger-dominated systems are
the poor resolution of the simulation outputs on which the mass also compared.
decomposition is performed (leading to an overestimate of the bulge However, the distributions of alignment between the SMBH spin
mass in the simulation). and galaxy spin, φ gal,BH , are not statistically significantly differ-
ent between the non-merger-selected galaxies and the rest of the
81 galaxies are found in each of the three samples (i.e. with
population in either of Figs 1–3. This appears to contradict the
Nmerger = 0, fBH,merge < 0.1 and B/T < 0.1). 1346 galaxies are found
secular feeding hypothesis of Smethurst et al. (2019) and the results
in both the galaxy merger-free and BH merger-free samples (i.e.
from Bustamante & Springel (2019). We investigated this further
Nmerger = 0 and fBH,merge < 0.1). Of the 179 galaxies selected to be
in Fig. 4. Under the secular feeding hypothesis, SMBHs grown by
disc dominated, 112 (63 per cent) are classed as both disc dominated
non-merger processes should be spun up due to alignment between
and have not undergone a major or minor galaxy merger since z = 2
galaxy and SMBH spin, so should have maximal spin (1 − |a| ∼
(i.e. Nmerger = 0 and B/T < 0.1). This suggests that the disc-dominated
0) and be aligned (cos φ gal,BH = 1). While the majority of such
galaxies observed by Simmons et al. (2013); Simmons et al. (2017);
systems display this behaviour there is still a large tail to misaligned
Smethurst et al. (2019, 2021) are not a unique subset of galaxies, but
and anti-aligned systems (the bottom left corner of Fig. 4) due to
are instead representative of the merger-free galaxy population.
the fact that individual accretion episodes can realign BH spin when
alignment time-scales are short, which increases the scatter in φ gal,BH
(see appendix C of Dubois et al. 2014a). For a model with longer
3 R E S U LT S
accretion time-scales, we would expect a larger discrepancy between
Figs 1–3 show the properties of galaxies with merger-free and BHs in non-merger systems, who would have time to align with
merger-dominated evolutionary histories identified using our three their host galaxy, and those in merger-dominated system, where BH
different criteria: bulge-to-total ratio (B/T; Fig. 1), number of minor spin alignment would lag after any changes in galactic spin due to
and major galaxy mergers since z = 2 (Nmerger ; Fig. 2), and the fraction galaxy merger. In such a scenario, merger-grown systems would be
of the SMBH mass gained through BH mergers (fBH,merge ; Fig. 3).
We investigate the differences in the distributions of SMBH mass
3 The lack of resolution within the vertical structure of the disc acts like
(MBH ), total stellar mass (M∗ ), bulge stellar mass (Mbulge ), bulge-
to-total ratio (B/T), fraction of the SMBH mass gained through BH an extra source of temperature in the disc that most likely prevents it from
secularly barring.
mergers (fBH,merge ), SMBH spin magnitude (|aBH |), and the angle 4 Bulge growth due to instabilities is computed in the Millenium simulation
between the SMBH spin and galaxy spin (φ gal,BH ; see equation 1).
using an inequality quantifying the dynamics of the disc, which when satisfied
The masses of non-merger grown SMBHs are lower than for the leads to either partial or full (depending on the model used) collapse of the
rest of the population. This is apparent when non-merger grown mass in the disc into a spheroid. Horizon-AGN, being a hydrodynamical
SMBHs are selected using an observational cut with galaxy B/T simulation, natively follows the formation of bulges through distributed star
ratio and with the cuts that simulations make possible of number formation within the galaxy, but is limited by its spatial resolution of x =
of mergers and fraction of SMBH mass grown by mergers (>5σ 1kpc.

MNRAS 527, 10867–10877 (2024)

 High-spin SMBHs in merger-free galaxies 10871

Downloaded from https://academic.oup.com/mnras/article/527/4/10867/7205294 by guest on 10 March 2024

Figure 1. The distribution of properties across the Horizon-AGN sample for galaxies selected to have low bulge-to-total ratios, B/T < 0.1 (blue; assumed to have
accretion grown SMBHs), compared to the rest of the sample with B/T > 0.1 (black; i.e. merger and accretion grown SMBHs). Shown are the SMBH masses,
MBH ; total stellar masses, M∗ ; bulge stellar mass, Mbulge ; the fraction of the SMBH mass resulting from mergers, fBH,merge (the dashed line shows fBH,merge =
0.5 and denotes the threshold between non-merger on the left and merger dominated growth on the right); the spin magnitude of the SMBH, |a|; and the angle
between the spin of the SMBH and the galaxy, cos φ gal,BH . A value of cos φ gal,BH = 1 means the SMBH and galaxy spins are aligned; cos φ gal,BH = 0 means
they are misaligned by 90◦ ; and cos φ gal,BH = −1 means they are misaligned by 180◦ . A value of |a| = 1 represents a maximally spinning BH. ‘Bulgeless’
systems with B/T < 0.1 have statistically significantly lower SMBH masses, lower stellar masses, lower bulge masses, and lower SMBH merger mass fractions.
However, they have statistically significantly higher spin magnitudes, yet no statistically significant difference in spin alignment. We note that these results hold
when incomplete, total stellar mass-matched samples of merger-free and merger-dominated systems are compared.

expected to be spun down and misaligned with their galaxies due we once again look at the distribution of cos φ gal,BH but for the true
to the variation in angular momentum of accreted gas introduced extremes of the population; those non-merger–driven systems with
by galaxy mergers. In our sample, merger-dominated systems show fBH,merge < 0.1 (2137 systems) and fBH,merge >0.6 (318 systems). Their
a similar scatter in φ gal,BH values as non-merger grown systems, distributions of cos φ gal,BH are statistically significantly different (σ =
suggesting that secular processes dominate the evolution of SMBHs 3.4 in a KS test), with BH merger grown systems more likely
in the epochs between galaxy mergers. While it is evident that BH to be misaligned (cos φ gal,BH ∼ 0) or anti-aligned (cos φ gal, BH ∼
merger-grown BHs are spun down by the mergers in Fig. 4, there is no −1), due to the spin flips caused by the misaligned orbital angular
correlation between φ gal,BH and BH spin magnitude |a| (solid line). momentum of SMBH during mergers. It is only when probing the
The caveat is that the spatial resolution of x = 1 kpc under-resolves extremes of the population that this is apparent. The rest of the
the angular momentum structure in the centre of the galaxy, let alone SMBH population evolves with a mix of merger and non-merger
on scales closer to the event horizon of the BH. While (Dubois et al. histories, with non-merger processes dominating (51 per cent of
2014b) have shown that the spin evolution of a BH is reasonably SMBH have fBH,merger < 0.25, i.e. have not undergone a major merger
well converged on resolutions of 10–80 pc, the significantly lower with a mass ratio of at least 1:3 since z ∼ 2), where high spins
resolution of Horizon-AGN means that we probably overpredict the and alignment are expected. We note that while BH–BH mergers
alignment between galactic and BH spin. We also note that the could not happen without the galaxy mergers that deliver multiple
BH spin evolution model used here treats each accretion event as BHs to the same galaxy, the effect discussed here is entirely driven
independent. If the time-scales for a newly formed accretion disc by the re-alignment of BH spins during a BH–BH merger, not
to be consumed are significantly longer than the time-step of the by the rearranging of the galactic spin during galaxy mergers. As
simulation this would artificially increase the scatter in Fig. 4. can be seen in the bottom right-hand panel of Fig. 2, there is no
The right-hand panel of Fig. 4 shows that there are differences in noticeable difference in the distribution of alignment angles between
spin and spin alignment with fBH,merge , but they are gradual enough BH and galaxy for merger-rich and merger-poor galaxies. In addition,
across the parameter space that the single-threshold distributions of BH spins for BHs in both samples of galaxies are highly aligned
Figs 1–3 are not able to capture them. However, it is clear in Fig. 4 with their host galaxy spin. This result will be at least partially
the SMBHs with the highest average mass fraction grown by BH influenced by the limited resolution of Horizon-AGN, and should
mergers, fBH,merge , are found in the lowest spin, most misaligned bins. be treated as an upper limit: at 1 kpc of resolution, the angular
This is only apparent when controlling for fBH,merge and so in Fig. 5 momentum distribution in the centres of galaxies is insufficiently well

MNRAS 527, 10867–10877 (2024)

10872 R. S. Beckmann, R. J. Smethurst et al.

Downloaded from https://academic.oup.com/mnras/article/527/4/10867/7205294 by guest on 10 March 2024

Figure 2. The distribution of properties across the Horizon-AGN sample for galaxies selected to have had no major or minor mergers since z = 2, Nmerger = 0
(blue; assumed to have accretion grown SMBHs), compared to those which have had more than three major or minor mergers since z = 2, Nmerger > 3 (black;
i.e. merger dominated SMBH growth). Shown are the SMBH masses, MBH ; total stellar masses, M∗ ; bulge stellar mass, Mbulge ; the fraction of the SMBH mass
resulting from mergers, fBH,merge (the dashed line shows fBH,merge = 0.5 and denotes the threshold between non-merger on the left and merger dominated growth
on the right); the spin magnitude of the SMBH, |a|; and the angle between the spin of the SMBH and the galaxy, cos φ gal,BH . A value of cos φ gal,BH = 1 means
the SMBH and galaxy spins are aligned; cos φ gal,BH = 0 means they are misaligned by 90◦ ; and cos φ gal,BH = −1 means they are misaligned by 180◦ . A value
of |a| = 1 represents a maximally spinning BH. Accretion dominated systems with Nmerger = 0 have statistically significantly lower SMBH masses, lower stellar
masses, lower bulge masses, and lower SMBH merger mass fractions. However, they do have statistically significantly higher SMBH spin magnitudes, yet once
again no statistically significant difference in spin alignment. We note that these results hold when incomplete; total stellar mass matched samples of merger-free
and merger-dominated systems are compared.

resolved to significantly decouple from the larger galactic angular a BH to re-align its spin vector with that of the host galaxy is much
momentum. shorter than the characteristic time-scale for it to increase its spin
magnitude, as can be seen by that fact that on average all BHs in our
sample, no matter their merger history, preferentially align with their
4 DISCUSSION host galaxy. A BH that has undergone a not-too-recent major merger,
or series of more minor mergers, would only need a comparatively
4.1 Time-scales of spin alignment small amount of coherent accretion, possibly following a period of
chaotic accretion during a galaxy merger that reduces its spin, to
Our results have shown that SMBHs in merger-free galaxies have then realign itself with its host galaxy while retaining a low spin.
higher spins (>5σ ) than the rest of the galaxy population and are The rate of this accretion would be too low to cause significant
more likely to be aligned with their galaxy spin (3.4σ ). These results mass gain or spin-up the BH spin magnitude. Finally, systems
support the secular feeding hypothesis of Nayakshin et al. (2012) and might end up spinning highly if they are dominated by their orbital
Smethurst et al. (2019), where material inflowing to the SMBH in angular momentum: Gammie, Shapiro & McKinney (2004) argue
the centre of a non-merger grown system will come from within the that following a merger between two BHs, it is reasonable to assume
galactic disc at a constant angular momentum vector, spinning up the that the final BH has an angular momentum that is equal to that of
SMBH to maximum, and subsequently align the spin of the SMBH the binary. Specifically, ‘the merger of two BHs of comparable mass
according to that of the accretion disc formed from infalling gas will immediately drive the spin parameter of the merged hole to 0.8,’
through the Bardeen–Petterson effect (Bardeen & Petterson 1975). where the spin parameter is J/MBH .
However, two populations of BHs remain difficult to explain in
this hypothesis: (i) merger-grown SMBHs that are aligned with their
galaxy but have low-spin magnitudes, and (ii) merger-grown SMBHs
4.2 Spin alignment and the impact of AGN feedback
with high-spin magnitudes. Some of these effects will be statistical:
for some BH–BH mergers, the spins of the two BHs and orbital Understanding the relation between the SMBH spin magnitude and
angular momentum will roughly align, which means the spin of the alignment of the galaxy and SMBH spin vectors in merger versus
post-merger BH will remain high. For others, the post-merger spin non-merger grown systems is crucial for our understanding of the
will happen to align closely with that of the galaxy. For aligned, low- impact of AGN feedback and therefore of galaxy evolution in
spin BHs it is most likely a question of time-scales: the time-scale for its entirety. AGNs in Horizon-AGN are thought to be in one of

MNRAS 527, 10867–10877 (2024)

 High-spin SMBHs in merger-free galaxies 10873

Downloaded from https://academic.oup.com/mnras/article/527/4/10867/7205294 by guest on 10 March 2024

Figure 3. The distribution of properties across the Horizon-AGN sample for galaxies selected to have low SMBH mass fractions grown by mergers, fBH,merge
< 0.1 (blue; and therefore accretion dominated SMBH growth), compared to the rest of the sample with fBH,merge > 0.1 (black; i.e. merger and accretion grown
SMBHs). Shown are the SMBH masses, MBH ; total stellar masses, M∗ ; bulge stellar mass, Mbulge ; the bulge-to-total stellar mass ratio B/T; the spin magnitude
of the SMBH, |a|; and the angle between the spin of the SMBH and the galaxy, cos φ gal, BH . A value of cos φ gal,BH = 1 means the SMBH and galaxy spins are
aligned, cos φ gal,BH = 0 means they are misaligned by 90◦ , and cos φ gal,BH = −1 means they are misaligned by 180◦ . A value of |a| = 1 represents a maximally
spinning BH. Systems with SMBH mass fractions grown by mergers, fBH,merge <0.1, have statistically significantly lower SMBH masses, lower stellar masses,
and lower bulge masses. They have a bimodal distribution in bulge-to-total ratio, B/T, which is statistically significantly different to the rest of the population.
They also have statistically significantly higher spin magnitudes, yet once again no statistically significant difference in spin alignment. We note that these results
hold when incomplete, total stellar mass-matched samples of merger-free and merger-dominated systems are compared.

Figure 4. The SMBH spin magnitude, 1 − |a|, plotted against the angle between the SMBH spin vector and galaxy spin vector, cos φ gal,BH , for our entire
Horizon-AGN sample. In the left-hand panel, the data are plotted as a scatter plot with the points coloured by the fraction of the SMBH built by mergers,
fBH,merge , with the black line showing the change in the average alignment with SMBH spin magnitude. In the right-hand panel, the data are binned and the
median fBH,merge value is shown for each bin. While there appears to be no obvious correlation between SMBH spin and alignment in the left-hand panel, the
right-hand panel reveals that the highest fBH,merge values are found at the lowest spin in the most misaligned systems (either perpendicular to the galaxy spin, or
anti-aligned). Note that a low value of 1 − |a| corresponds to a maximally spinning SMBH. We show 1 − |a| here, as opposed to |a|, on a logarithmic scale to
better appreciate the range of spins in maximally spinning systems. A value of cos φ gal,BH = 1 means the SMBH and galaxy spins are aligned, cos φ gal,BH =
0 means they are misaligned by 90◦ , and cos φ gal,BH = −1 means they are anti-aligned by 180◦ . We note that these results hold when incomplete, total stellar
mass matched samples of merger-free and merger-dominated systems are compared.

MNRAS 527, 10867–10877 (2024)

10874 R. S. Beckmann, R. J. Smethurst et al.

Downloaded from https://academic.oup.com/mnras/article/527/4/10867/7205294 by guest on 10 March 2024

Figure 5. The distribution of the angle between the galaxy and BH spin vector, cos φ gal,BH , for galaxies with SMBHs with less than 10 per cent of their mass
due to mergers (fBH,merge < 0.1) and with more than 60 per cent of their mass due to mergers (fBH,merge ≥ 0.6). These are the two extremes of the population, with
distributions that are statistically significantly different (3.4σ ). SMBHs dominated by non-merger growth are more likely to be aligned to their galaxy spins than
SMBHs dominated by merger growth. A value of cos φ gal,BH = 1 means the SMBH and galaxy spins are aligned, cos φ gal,BH = 0 means they are misaligned by
90◦ (marked by the dashed line), and cos φ gal,BH = −1 means they are misaligned by 180◦ .

two feedback modes. Those SMBHs with Eddington ratios, fEdd <
1 per cent are thought to cause a ‘radio/kinetic/jet/maintenance
mode’ of AGN feedback, mostly affecting the galaxy halo. Whereas,
those SMBHs with Eddington ratios, fEdd > 1 per cent are thought
to cause a ‘quasar/thermal/radiative mode’ of AGN feedback, heating
the central regions of the galaxy during the peak luminosity of the
AGN. In both modes, the radiative efficiency changes as a function of
BH spin magnitude, with higher spinning BHs typically converting a
larger fraction of their accreted mass into feedback energy. Given
that we have shown that galaxy-merger-free grown SMBHs are
more likely to have a higher spin, and therefore higher rotational
energy, this could have important implications for the AGN feedback
efficiency in such systems.
Our results also suggest that such outflows from merger-free
systems will be preferentially aligned with the galaxy spin and will
therefore be ejected out of the plane of the galaxy if the AGN is in Figure 6. The distribution of the time spent in a radio mode of AGN feedback
the directional ‘radio/kinetic/jet/maintenance’ mode, rather than the (i.e. with fEdd < 1 per cent) since z = 2 for galaxies with merger-free
isotropic ‘quasar/thermal/radiative’ mode; such increased feedback evolutionary histories (Nmerger = 0; blue) and those with merger-dominated
efficiency may then only impact both the very centre of the galaxy evolutionary histories (Nmerger > 3; black). The median value for each
and the galaxy halo, rather than progressing through the galactic distribution is shown by the dashed lines in the corresponding colours. The
distributions are statistically significantly different (>5σ ) with merger-free
disc. Davé et al. (2019) using the SIMBA simulation found that the
systems more likely to spend a larger proportion of their lifetimes in a radio
‘radio/kinetic/jet/maintenance mode’ of AGN feedback can drive
mode of AGN feedback.
population-wide galactic quenching from the ‘outside-in’ as it heats
the halo and cuts off the gas supply from the galactic halo to the disc. (Nmerger > 3) evolutionary histories. We calculated the fraction of time
This quenches galaxies even in the absence of the more energetic since z = 2 that SMBHs spend in a radio mode of AGN feedback (i.e.
‘quasar/thermal/radiative mode’ of AGN feedback which heats the with fEdd < 10−2 ). The distributions for each sample are shown in
surrounding galactic regions around the SMBH. Similar results on Fig. 6 along with the median values shown by the dashed lines. The
the importance of quenching in the absence of galaxy mergers were two distributions are statistically significantly different [with σ > 5 in
reported for IllustrisTNG (Weinberger et al. 2018; Xu et al. 2022) a Kolmogorov–Smirnov (KS) test], with merger-free galaxies more
and Horizon-AGN (Dubois et al. 2016; Beckmann et al. 2017; Martin likely to spend a higher proportion of their evolutionary histories
et al. 2018). in a radio mode of AGN feedback. Despite this, the averages for
This difference in spin direction will only impact where the AGN both samples remain high: SMBHs in merger-free galaxies spend,
feedback energy is delivered if the AGN spend a significant fraction on average, 91 per cent of their lifetime in radio mode AGN feedback,
of their time in ‘radio/kinetic/jet’ mode. To understand whether while the sample in merger-dominated galaxies spends on average,
this is the case, we investigated the SMBH Eddington ratios in 88 per cent of their lifetime in radio mode AGN feedback. While
those galaxies with merger-free (Nmerger = 0) and merger-dominated in radio mode, AGN drive powerful jets (here defined as 10−4 <

MNRAS 527, 10867–10877 (2024)

 High-spin SMBHs in merger-free galaxies 10875
using the angular momentum of the accreted gas for any given
accretion and subsequent feedback event (gas-based model, which
was used as the simulation was run). The gas-based model leads
to less alignment between jets and the galactic angular momentum
than the spin-based model. For disc galaxies, this would mean more
jet energy is deposited directly into the circumgalactic medium in a
spin-based than in a gas-based jet model. The difference in alignment
between gas-based and spin-based models arises because in a spin-
based model, the inherent angular momentum of the SMBH stabilizes
the jets against sudden realignment, while the direction of the central
accreted angular momentum can be quite stochastic even for a well-
ordered disc galaxy. Based on this insight, we would expect an on-
the-fly spin-based model to drive jets into the circumgalactic more
often than the current gas-based model, which would lead to a more

Downloaded from https://academic.oup.com/mnras/article/527/4/10867/7205294 by guest on 10 March 2024

indirect impact on galactic star formation, by cutting off large-scale
Figure 7. Distribution of gal,jet , the angle between the jet and the galactic gas flows, rather than direct quenching of star formation in galaxies.
angular momentum, for two jet models: a gas-based model where jets are (ii) Secondly, if SMBH spin were taken into account in BH’s
injected along the angular momentum axis of gas accreted during the last radiative luminosity r , a given SMBH might produce anywhere from
accretion event (red solid) and a spin-based model where jets are injected 0.3 to ∼4 times the feedback energy at the same accretion rate in
along the spin axis of the SMBH (grey hatched). Median values for each
comparison to a fixed r . This would both change how much feedback
distribution are shown by the solid lines of corresponding colours. As it
energy the galaxy is receiving, and how fast the SMBH is growing
disregards the inherent angular momentum of a spinning SMBH, the gas-
based model (red solid) shows a higher degree of scatter and less long-term (as the SMBH gains mass at a rate of 1 − r ). For the model used
alignment between SMBH jets and galactic angular momentum than the here, feedback efficiency peaks for maximally spinning SMBH at a
spin-based model (grey hatched). value of about 4 times the fixed r = 0.1 used to run the simulation, so
we would expect secularly grown SMBHs to grow more slowly than
fedd < 10−2 ) for a significant fraction of their time: 64 per cent those in merger-driven galaxies. This could reduce some of the effects
for SMBHs in merger-free galaxies and 59 per cent for SMBHa in of SMBHs in merger-free galaxies being overmassive in comparison
merger-dominated galaxies. This similarity does not strongly depend to their host galaxy reported in the companion paper (Smethurst et
on the exact threshold value of fedd chosen when transitioning from al. 2023), or delay (but not prevent) the efficient spin-up of SMBHs
jet to quasar mode, with merger-free galaxies continuing to spend in merger-free galaxies shown here in Fig. 3. The exact impact is
∼5 per cent more time in jet mode than merger-dominated galaxies hard to predict as SMBHs undergo self-regulation: strong feedback
for over an order of magnitude in difference in transition threshold. episodes reduce accretion on to the SMBHs in the short term, but
This means that for a non-negligible fraction of their evolution, AGN in long term the balance between gas inflows and SMBH feedback
feedback energy for merger-free galaxies is injected with a higher might simply settle at a new equilibrium that produces a population
efficiency at a given accretion rate than for merger-grown SMBH of SMBHs with ultimately similar properties. How exactly a full
due to the higher spin of SMBH in merger-free galaxies. This could on-the-fly model will different from post-processing will have to be
potentially enhance the importance of radio-mode quenching (as tested using a follow-up simulation.
advocated by Weinberger et al. 2018; Davé et al. 2019) in merger-
free galaxies, and compensate for the short-lived luminosity bursts 4.4 Follow-up observational work
of AGN post-galaxy merger (McAlpine et al. 2011; Volonteri et al.
2016), which can lead to the effective quenching of the merger Given the predictions of our simulations presented here which
remnant (Dubois et al. 2016). In addition, the similarity between suggest that the spin magnitude and orientation of galaxy-merger-
the average time spent in a radio-mode feedback for the merger- free SMBHs are indeed different from the wider galaxy population
free and merger-dominated samples again suggests that in the long and that this may lead to increased AGN feedback efficiency, high-
epochs between galaxy mergers, merger-free accretion dominates the resolution observational studies are therefore essential to test these
growth of SMBHs even in galaxy merger-rich systems and possibly predictions. First, a study to test the alignment of AGN outflows with
makes an important contribution to their long-term co-evolution. respect to their galactic discs in merger-free galaxies. For example,
using narrow-band filters centred on [O III] ionization with the high-
spatial resolution provided by the Hubble Space Telescope to probe
4.3 Post-processed versus on-the-fly SMBH spin evolution the alignment of AGN outflows. While previous studies have found
no alignment between outflows and the galaxy minor axis (i.e. an
One limitation of the work shown here is that the spin-evolution is
alignment of SMBH spin and galaxy spin; see Kinney et al. 2000;
post-processed from the BH mass evolution history, rather than self-
Schmitt et al. 2003; Ruschel-Dutra et al. 2021), such studies were of
consistently run on the fly. If the BH spin model had been run on
samples with mixed morphologies and therefore mixed evolutionary
the fly, AGN feedback in the simulation would have differed in the
histories. By isolating merger-free systems, e.g. by observationally
following two ways:
selecting bulgeless galaxies, and studying the orientation of their
(i) First, AGN jets would realign more slowly and coherently, outflows we can observationally test the hypotheses discussed here.
as the BH spin vector is computed as an integrated quantity while Secondly, an observational study on the subsequent impact of the
the local gas angular momentum is an instantaneous quantity. This outflow on the galaxy star formation rate is required. A high spatial
can be seen in Fig. 7, which shows the angle between the galactic and spectral resolution integral field unit, such as MUSE or VIRUS,
angular momentum and the jet for two cases: when using the will be able to spatially resolve the areas of disc-dominated merger-
instantaneous SMBH spin (spin-based model, post-processed), and free galaxies impacted by AGN outflows and spectrally separate

MNRAS 527, 10867–10877 (2024)

10876 R. S. Beckmann, R. J. Smethurst et al.
the emission ionized by star formation and the outflow. Such a of AGN feedback occurs in this mode, in agreement with the findings
study would allow us to observationally test the assertions of Davé from the SIMBA simulation (Davé et al. 2019). Given that we find
et al. (2019) and the hypothesis discussed here: whether merger-free that galaxy merger-dominated systems spend a similar average time
powered radio mode AGN feedback could indeed be the cause of in a radio mode, this once again suggests that in the epochs between
galaxy population wide quenching. galaxy mergers, merger-free accretion dominates the growth of their
Smethurst et al. (2021) found that their observational sample of SMBHs. If future observational studies reveal that this feedback is
four ‘bulgeless’ (assumed merger-free) AGN with outflows, had capable of causing galaxy quenching, this then suggests that the
outflow rates, energy injection rates and momentum fluxes that were correlations between, e.g. total stellar mass and SMBH mass, seen
comparable to a sample of low-z Type 1 AGN from Rakshit & for samples of merger-free galaxies (Simmons et al. 2013; Simmons
Woo (2018, within the scatter). Smethurst et al. (2021) suggested et al. 2017; Smethurst et al. 2023) is caused in part by secularly
that this result implied that it is possible that the majority of low- driven co-evolution regulated by radio mode AGN feedback.
redshift AGNs (both SMBH growth and outflows) are powered by
non-merger processes. Given our results above, this suggests that the High-resolution observational studies on the impact of AGN
majority of this low-z AGN feedback will occur via a radio mode. outflows and alignment with respect to their galactic discs in a

Downloaded from https://academic.oup.com/mnras/article/527/4/10867/7205294 by guest on 10 March 2024

If future observational studies reveal that this feedback is capable pure sample of merger-free galaxies (e.g. disc-dominated galaxies)
of causing galaxy quenching, this could explain why a correlation is therefore essential to both test these hypotheses of secularly
is still observed between e.g. total stellar mass and SMBH mass powered AGN feedback which seems to dominate galaxy-SMBH
for merger-free systems (see Simmons et al. 2013; Simmons et al. co-evolution.
2017 and companion paper Smethurst et al. 2023); co-evolution
regulated by radio mode AGN feedback is occurring due to secular
processes. AC K N OW L E D G E M E N T S
First authorship is shared between RSB and RJS. RSB assem-
bled simulation data catalogues, interpreted results and wrote the
5 CONCLUSIONS
manuscript. RJS conceived of the project, analysed data, interpreted
We have investigated the SMBH spin magnitude and spin alignment results and wrote the manuscript.
in the context of the spin of their host galaxy for merger-free and RJ Smethurst gratefully acknowledges funding from Christ
merger-dominated galaxies using the Horizon–AGN simulation. Our Church, Oxford and the Royal Astronomical Society. RS Beck-
conclusions are summarized as follows: mann gratefully acknowledges funding from Newnham College,
Cambridge. BDS acknowledges support from a UK Research and In-
(i) Galaxies that have evolved in the absence of mergers host novation Future Leaders Fellowship (grant number MR/T044136/1)
SMBHs with preferentially higher spin magnitudes than those This work is partially supported by grant Segal ANR-19-CE31-
with merger-dominated evolutionary histories. This is true for both 0017 of the French Agence Nationale de la Recherche. ILG ac-
SMBH mergers and galaxy mergers and supports the hypothesis of knowledges support from an STFC PhD studentship (grant number
Smethurst et al. (2019), building on the work of Nayakshin et al. ST/T506205/1) and from the Faculty of Science and Technology at
(2012). Lancaster University.
(ii) SMBHs with low-mass fractions built by BH mergers This work used the HPC resources of CINES (Jade supercomputer)
(fBH,merge <0.1), as well as galaxies with few galaxy mergers, have under the allocation 2013047012 made by GENCI, and the horizon
a bi-modal distribution of bulge-to-total ratios (see Figs 3 and and Dirac clusters for post processing. This work is partially
2) suggesting that galaxies which have grown their SMBHs in supported by the Spin(e) grants ANR-13-BS05-0002 of the French
the absence of mergers can still have significant stellar bulges, in Agence Nationale de la Recherche and by the National Science
agreement with Parry et al. (2009). This leads to secular co-evolution Foundation under grant no. NSF PHY11-25915, and it is part of the
of both bulges and SMBHs. Horizon-UK project, which used the DiRAC Complexity system,
(iii) At first glance, SMBHs in galaxies with and without mergers operated by the University of Leicester IT Services, which forms
have a very similar distribution of the angle between the spin vectors part of the STFC DiRAC HPC Facility (www.dirac.ac.uk). This
of the SMBH and their galaxy, φ BH,gal and show similar scatter in equipment is funded by BIS National E-Infrastructure capital grant
φ BH,gal (see Fig. 4). It is only when probing the extremes of the ST/K000373/1 and STFC DiRAC Operations grant ST/K0003259/1.
population that it becomes apparent that the distribution of φ BH,gal DiRAC is part of the National E-Infrastructure. This work has made
are statistically significantly different (3.4σ ) for SMBH dominated use of the Horizon Cluster hosted by Institut d’Astrophysique de
(fBH,merge > 0.6) or not (fBH,merge < 0.1) by BH mergers. This once Paris. We thank Stéphane Rouberol for smoothly running this cluster
again suggests that secular processes drive the evolution of SMBHs for us. This research made use of ASTROPY,5 a community-developed
in the epochs between galaxy mergers, which for most objects core PYTHON package for Astronomy (Astropy Collaboration 2013,
dominates their overall evolution. 2018).
(iv) Given that previous observational and theoretical works have
also concluded that galaxy merger-free processes dominate SMBH-
galaxy co-evolution, this suggests secular processes power the
majority of the subsequent AGN feedback. We investigated the DATA AVA I L A B I L I T Y
Eddington ratios of galaxies with merger-free and merger-dominated All data used in this paper is available upon request to the first
evolutionary histories, determining the time spent in a radio mode authors.
of AGN feedback (i.e. fEdd < 1 per cent; see Fig. 6). We found that
merger-free systems spend 91 per cent of their evolutionary history
in a radio mode, similar to merger-dominated systems which spend
88 per cent of their time in radio mode, suggesting that the majority 5 http://www.astropy.org

MNRAS 527, 10867–10877 (2024)

 High-spin SMBHs in merger-free galaxies 10877
REFERENCES Martin G. et al., 2018, MNRAS, 476, 2801
Merritt D., Ferrarese L., 2001, MNRAS, 320, L30
Astropy Collaboration, 2013, A&A, 558, A33 Natarajan P., Pringle J. E., 1998, ApJ, 506, L97
Astropy Collaboration, 2018, AJ, 156, 123 Nayakshin S., Power C., King A. R., 2012, ApJ, 753, 15
Aubert D., Pichon C., Colombi S., 2004, MNRAS, 352, 376 Ostriker E. C., 1999, ApJ, 513, 252
Bardeen J. M., 1970, Nature, 226, 64 Papaloizou J. C. B., Pringle J. E., 1983, MNRAS, 202, 1181
Bardeen J. M., Petterson J. A., 1975, ApJ, 195, L65 Parry O. H., Eke V. R., Frenk C. S., 2009, MNRAS, 396, 1972
Beckmann R. S. et al., 2017, MNRAS, 472, 949 Perna M. et al., 2021, A&A, 646, A101
Berti E., Volonteri M., 2008, ApJ, 684, 822 Rakshit S., Woo J.-H., 2018, ApJ, 865, 5
Blandford R. D., Znajek R. L., 1977, MNRAS, 179, 433 Rees M. J., 1978, Nature, 275, 516
Booth C. M., Schaye J., 2010, MNRAS, 405, L1 Reynolds C. S., 2014, Space Sci. Rev., 183, 277
Bustamante S., Springel V., 2019, MNRAS, 490, 4133 Rezzolla L., Barausse E., Dorband E. N., Pollney D., Reisswig C., Seiler J.,
Chapon D., Mayer L., Teyssier R., 2013, MNRAS, 429, 3114 Husa S., 2008, Phys. Rev. D, 78, 044002
Davé R., Anglés-Alcázar D., Narayanan D., Li Q., Rafieferantsoa M. H., Ruschel-Dutra D. et al., 2021, MNRAS, 507, 74
Appleby S., 2019, MNRAS, 486, 2827 Scheuer P., Feiler R., 1996, MNRAS, 282, 291

Downloaded from https://academic.oup.com/mnras/article/527/4/10867/7205294 by guest on 10 March 2024

Dotti M., Colpi M., Pallini S., Perego A., Volonteri M., 2013, ApJ, 762, 68 Schmitt H. R., Donley J. L., Antonucci R. R. J., Hutchings J. B., Kinney A.
Dubois Y., Teyssier R., 2008, A&A, 477, 79 L., Pringle J. E., 2003, ApJ, 597, 768
Dubois Y., Devriendt J., Slyz A., Teyssier R., 2012, MNRAS, 420, 2662 Simmons B. D. et al., 2013, MNRAS, 429, 2199
Dubois Y., Volonteri M., Silk J., 2014a, MNRAS, 440, 1590 Simmons B. D., Smethurst R. J., Lintott C., 2017, MNRAS, 470, 1559
Dubois Y. et al., 2014b, MNRAS, 444, 1453 Smethurst R. J., Simmons B. D., Lintott C. J., Shanahan J., 2019, MNRAS,
Dubois Y., Peirani S., Pichon C., Devriendt J., Gavazzi R., Welker C., 489, 4016
Volonteri M., 2016, MNRAS, 463, 3948 Smethurst R. J. et al., 2021, MNRAS, 507, 3985
Fragile P. C., Blaes O. M., Anninos P., Salmonson J. D., 2007, ApJ, 668, 417 Smethurst R. J. et al., 2023, MNRAS, 527, 10855
Gammie C. F., Shapiro S. L., McKinney J. C., 2004, ApJ, 602, 312 Sorathia K. A., Krolik J. H., Hawley J. F., 2013, ApJ, 777, 21
Häring N., Rix H.-W., 2004, ApJ, 604, L89 Sparre M., Springel V., 2017, MNRAS, 470, 3946
Hopkins P. F., Quataert E., Murray N., 2012, MNRAS, p. 2655
Tadhunter C. et al., 2018, MNRAS, 478, 1558
Hu J., 2008, MNRAS, 386, 2242
Teyssier R., 2002, A&A, 385, 337
Kimm T., Cen R., Devriendt J., Dubois Y., Slyz A., 2015, MNRAS, 451, 2900
Tonini C., Mutch S. J., Croton D. J., Wyithe J. S. B., 2016, MNRAS, 459,
King A. R., Lubow S. H., Ogilvie G. I., Pringle J. E., 2005, MNRAS, 363, 49
4109
Kinney A. L., Schmitt H. R., Clarke C. J., Pringle J. E., Ulvestad J. S.,
Tweed D., Devriendt J., Blaizot J., Colombi S., Slyz A., 2009, A&A, 506,
Antonucci R. R. J., 2000, ApJ, 537, 152
647
Komatsu E. et al., 2011, ApJS, 192, 18
van den Bosch R. C. E., 2016, ApJ, 831, 134
Kormendy J., Bender R., Cornell M. E., 2011, Nature, 469, 374
Volonteri M., Dubois Y., Pichon C., Devriendt J., 2016, MNRAS, 460, 2979
McAlpine W., Satyapal S., Gliozzi M., Cheung C. C., Sambruna R. M.,
Volonteri M. et al., 2020, MNRAS, 498, 2219
Eracleous M., 2011, ApJ, 728, 25
Walker I. R., Mihos J. C., Hernquist L., 1996, ApJ, 460, 121
McAlpine S., Harrison C. M., Rosario D. J., Alexander D. M., Ellison S. L.,
Weinberger R. et al., 2018, MNRAS, 479, 4056
Johansson P. H., Patton D. R., 2020, MNRAS, 494, 5713
Xu Y., Luo Y., Kang X., Li Z., Li Z., Wang P., Libeskind N., 2022, ApJ, 928,
McConnell N. J., Ma C.-P., Gebhardt K., Wright S. A., Murphy J. D., Lauer
100
T. R., Graham J. R., Richstone D. O., 2011, Nature, 480, 215
Magorrian J. et al., 1998, AJ, 115, 2285
Marconi A., Hunt L. K., 2003, ApJ, 589, L21
Martig M., Bournaud F., Croton D. J., Dekel A., Teyssier R., 2012, ApJ, 756,
26 This paper has been typeset from a TEX/LATEX file prepared by the author.

© 2023 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://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.

MNRAS 527, 10867–10877 (2024)