Artigo de Astronomia 1

MNRAS 513, 3090–3103 (2022) https://doi.org/10.
1093/mnras/stac1076
Advance Access publication 2022 April 19
Magnetic dynamos in white dwarfs – III. Explaining the occurrence of

strong magnetic fields in close double white dwarfs
Matthias R. Schreiber,1,2 ‹ Diogo Belloni ,1 Monica Zorotovic,3 Sarai Zapata,1 Boris T. Gänsicke 4
Downloaded from https://academic.oup.com/mnras/article/513/2/3090/6570706 by Lund University Libraries, Head Office user on 21 November 2023
and Steven G. Parsons 5
1 Departamento ˜ 1680, Valparaı́so, Chile
de Fı́sica, Universidad Técnica Federico Santa Marı́a, Avenida Espana
2 Millennium Nucleus for Planet Formation, NPF, Valparaı́so, Chile
3 Instituto de Fı́sica y Astronomı́a, Universidad de Valparaı́so, Avenida Gran Bretana
˜ 1111, Valparaı́so, Chile
4 Department of Physics, University of Warwick, Coventry CV4 7AL, UK
5 Department of Physics and Astronomy, University of Sheffield, Sheffield S3 7RH, UK
Accepted 2022 April 12. Received 2022 April 1; in original form 2022 February 3
ABSTRACT
The origin of strong (>∼ 1 MG) magnetic fields in white dwarfs has been a puzzle for decades. Recently, a dynamo mechanism
operating in rapidly rotating and crystallizing white dwarfs has been suggested to explain the occurrence rates of strong magnetic
fields in white dwarfs with close low-mass main-sequence star companions. Here, we investigate whether the same mechanism
may produce strong magnetic fields in close double white dwarfs. The only known strongly magnetic white dwarf that is part
of a close double white dwarf system, the magnetic component of NLTT 12758, is rapidly rotating and likely crystallizing and
therefore the proposed dynamo mechanism represents an excellent scenario for the origin of its magnetic field. Presenting a
revised formation scenario for NLTT 12758, we find a natural explanation for the rapid rotation of the magnetic component. We
furthermore show that it is not surprising that strong magnetic fields have not been detected in all other known double white
dwarfs. We therefore conclude that the incidence of magnetic fields in close double white dwarfs supports the idea that a rotation-
and crystallization-driven dynamo plays a major role in the generation of strong magnetic fields in white dwarfs.
Key words: stars: individual: NLTT 12758, NLTT 11748, SDSS J125733.63+542850.5 – stars: magnetic field – white dwarfs –
binaries: close.
most a few per cent, and all known systems are close to Roche lobe
1 I N T RO D U C T I O N
filling and contain cold white dwarfs with temperatures < ∼ 10 000 K
White dwarfs have been speculated to potentially have strong (Parsons et al. 2021). In contrast, the fraction of strongly magnetic
magnetic fields (exceeding 1 MG) since 1947 (Blackett 1947), but the white dwarfs among their semidetached descendants is very high,
first detection of a magnetic field in a white dwarf was only obtained exceeding most likely one-third (Pala et al. 2020).
more than 20 yr later (Kemp et al. 1970). Ever since, the question Present versions of the models suggested for generating strong
why some white dwarfs become strongly magnetic while others magnetic fields in white dwarfs, the fossil field (e.g. Braithwaite &
do not has been one of the fundamental unsolved issues of stellar Spruit 2004; Tout, Wickramasinghe & Ferrario 2004), the double-
evolution. Answering this question appears to be very complicated degenerate merger (Garcı́a-Berro et al. 2012), and the common-
largely because strongly magnetic white dwarfs are found in different envelope dynamo scenarios (e.g. Tout et al. 2008), cannot repro-
relative numbers among single white dwarfs, white dwarfs in close duce these observations. Either the predicted numbers of strongly
detached binaries, and white dwarfs in close semidetached binary magnetic white dwarfs are too small as in the fossil field scenario
stars. (Kawka & Vennes 2004; Tout et al. 2004; Wickramasinghe &
The incidence of strongly magnetic white dwarfs in magnitude- Ferrario 2005; Aurière et al. 2007), or the number of predicted
limited surveys of isolated white dwarfs is roughly 5 per cent (Kepler strongly magnetic white dwarfs is far too large as in the common-
et al. 2013), but the fraction is clearly higher among nearby white envelope scenario (Belloni & Schreiber 2020), or fails to reproduce
dwarfs that contain more old systems (Kawka et al. 2007; Bagnulo & the large number of magnetic white dwarfs in interacting binary stars
Landstreet 2020, 2021), which indicates that the fractions derived as in the double-degenerate merger scenario. In addition, in their
from magnitude-limited surveys potentially largely underestimate current form, none of the above scenarios can explain the absence of
the true fraction of strongly magnetic white dwarfs. strongly magnetic white dwarfs in young detached white dwarf/main-
Among close detached white dwarf binaries, the fraction of sequence binary stars (Liebert et al. 2005, 2015; Belloni & Schreiber
systems containing a strongly magnetic white dwarf is very small, at 2020; Schreiber et al. 2021a).
We have recently developed a new model for the generation
of magnetic fields in white dwarfs that can potentially explain
E-mail: [email protected] the different occurrence rates of strongly magnetic white dwarfs
© 2022 The Author(s)

Published by Oxford University Press on behalf of Royal Astronomical Society
Magnetic white dwarfs in double degenerates 3091
with main-sequence star companions (Schreiber et al. 2021a). This apply the suggested dynamo mechanism to close double white dwarf
scenario is based on the idea first put forward by Isern et al. (2017), binaries and investigate whether it can explain the occurrence rate and
who proposed that a crystallization- and rotation-driven dynamo characteristics of strongly magnetic white dwarfs in these systems.
similar to those operating in planets and low-mass stars could be
generating magnetic fields in white dwarfs.
2 K N OW N D O U B L E W H I T E DWA R F S A N D
According to the scenario developed by Schreiber et al. (2021a),
M AG N E T I C F I E L D S
white dwarfs in close binaries emerge from common-envelope
evolution and evolve towards shorter orbital periods without hosting In order to investigate whether the new scenario for magnetic field
a strong magnetic field. As soon as mass transfer starts, these post- generation may explain the observations of close (here defined as
common-envelope binaries become cataclysmic variables (CVs) and an orbital period below 35 d) double white dwarfs, we compiled
the white dwarf accretes mass and angular momentum. If the spun- a list of the currently known systems (see Table A1). At present,
up white dwarf has cooled enough to contain a crystallizing core, its only one system is known to host a strongly magnetized white
structure consists of an outer carbon-rich convective zone and a solid dwarf. NLTT 12758 consists of a magnetic white dwarf with a field
inner oxygen-rich nucleus (Isern et al. 2017; Schreiber et al. 2021a). strength of about 3.1 MG and an apparently non-magnetic white
This situation resembles that of the interior of planets and low-mass dwarf companion. The non-magnetic H-rich (DA) white dwarf is
stars and the convective motion in the liquid mantle can generate a more massive (M = 0.83 ± 0.03 M ) than its magnetic (DAP)
magnetic field. companion (M = 0.69 ± 0.05 M ), the orbital period of the binary
Isern et al. (2017) estimated the field strength generated by this is 1.154 d, the spin period of the magnetic white dwarf has been
dynamo based on scaling laws that are used for planets and low- measured to be 23 min, and the cooling ages of both white dwarfs
mass stars (Christensen & Aubert 2006; Christensen, Holzwarth & are comparable (Kawka et al. 2017).
Reiners 2009) and derived an upper limit of around 1 MG. However, The recently proposed magnetic dynamo requires relatively rapid
existing scaling laws are most likely not appropriate for white dwarfs rotation and a crystallizing core to generate strong magnetic fields in
as the much higher magnetic Prandtl number in white dwarfs could white dwarfs. The saturation of the dynamo is assumed to occur for
imply a much larger field strength (Bovino, Schleicher & Schober spin periods of just a few seconds (see Isern et al. 2017; Schreiber
2013; Brandenburg 2014). Very recently, Ginzburg, Fuller & Kawka et al. 2021a, for more details). The magnetic and lower mass white
(2022) argued that the convective turnover time is much larger than dwarf in NLTT 12758 is rapidly rotating with a spin period of 23 min
assumed by Isern et al. (2017), which would imply that white dwarfs that is much shorter than the typical spin period of non-magnetic
are almost always in the fast rotation regime. Combining this finding single white dwarfs (Hermes et al. 2017) but longer than the short
with a scaling law based on the balance between the Lorentz and periods of the order of seconds proposed for white dwarfs in CVs
Coriolis forces even permits the generation of strong magnetic fields (Schreiber et al. 2021a). One would therefore expect the magnetic
(reaching 100 MG) without postulating a magnetic field enhancement field strength of the magnetic white dwarf in NLTT 12758 to be
due to the white dwarf’s Prandtl number. smaller than the strongest magnetic fields observed in CVs. Indeed,
If indeed strong magnetic fields are generated in the white dwarfs in magnetic CVs the field strengths of the white dwarfs reach several
in CVs due to this dynamo, these fields may connect with the field of hundred MG with an average field strength of ∼30 MG (Ferrario, de
the secondary star and synchronizing torques transfer spin angular Martino & Gänsicke 2015; Belloni et al. 2020) and the average field
momentum from the white dwarf to the orbital motion, which can strength of pre-polars is ∼60–70 MG (Parsons et al. 2021), while for
cause the system to detach. At first, the detached system will appear the magnetic component in NLTT 12758 a field strength of just 3.1
for a short period of time as a radio pulsing detached white dwarf MG has been measured (Kawka et al. 2017).
binary similar to AR Sco (Marsh et al. 2016), and then evolve into a The second criterion for the dynamo to work is that the core of the
synchronized detached pre-polar (Schwope et al. 2009; Parsons et al. white dwarf must have started to crystallize. The onset of crystal-
2021) and subsequently into a semidetached magnetic CV (polar). lization depends on the temperature and the white dwarf mass. The
In addition to offering an explanation for the origin and evolution currently available cooling tracks that include crystallization differ
of magnetic white dwarfs in close binary stars with a main-sequence slightly in the predicted temperatures for the onset of crystallization.
star companion, we could recently show that the dynamo also In Fig. 1, we compare the onset of crystallization from the models
naturally explains the low occurrence rate of high-accretion rate by Salaris et al. (2010) and Bédard et al. (2020) with the position of
magnetic CVs in globular clusters (Belloni et al. 2021) and the all C/O white dwarfs in close double white dwarfs with known mass
increased incidence of magnetism among cold metal-polluted white and effective temperature.
dwarfs (Schreiber et al. 2021b). The measured parameters of the magnetic white dwarf component
Despite this success, we note that some individual systems contain- in NLTT 12758 (highlighted in blue in Fig. 1) are consistent with
ing magnetic white dwarfs require an alternative mechanism for the a crystallizing core in the white dwarf and therefore the rotation-
magnetic field generation. The recently discovered post-common- and crystallization-driven dynamo offers a plausible explanation for
envelope binary CC Cet (Wilson et al. 2021) and the polar EY Eri its magnetic nature. Most other white dwarfs that are members of
(Beuermann et al. 2020) both contain most likely magnetic helium- close double white dwarfs are either He-core white dwarfs for which
core white dwarfs and one of the main ingredients of the dynamo the dynamo scenario does not apply and/or are too hot for having
scenario is a crystallizing core consisting of carbon and oxygen. In crystallizing cores, which is also consistent with the dynamo scenario
addition, the potentially weakly magnetic white dwarf in the post- for magnetic field generation in white dwarfs. However, of the few
common-envelope binary V471 Tau (Sion et al. 2012) is too hot to double white dwarfs containing C/O white dwarfs, three additional
be crystallizing. ones have sufficiently cooled to be crystallizing.
Given this situation, it is of utmost importance to further investigate Apart from the DAP white dwarf in NLTT 12758, components
to which degree the dynamo scenario may solve the long-standing close to or beyond the onset of crystallization are its higher
puzzle of the origin of strong magnetic fields in white dwarfs and mass DA companion and the more massive white dwarfs in
in which cases an alternative mechanism is required. Here, we SDSS J125733.63+542850.5 (hereafter, SDSS J1257+5428) and
MNRAS 513, 3090–3103 (2022)

3092 M. R. Schreiber et al.
Figure 1. Comparison of the mass and effective temperature of the known white dwarfs that form part of a close (period less than 35 d) double white dwarf
(from Table A1) with the onset of crystallization according to the cooling sequences by Salaris et al. (2010, black dashed line) and Bédard et al. (2020, black
solid line). The vast majority of the white dwarfs are too hot to be crystallizing. Among the few exceptions is the magnetic component in NLTT 12758. For this
white dwarf, the measured parameters are consistent with crystallization in its core. As in addition the spin period has been measured to be very short (23 min),
it appears plausible that the rotation- and crystallization-driven dynamo generated the strong magnetic field. Three more white dwarfs are also consistent
with having a crystallizing core. This might indicate that those white dwarfs, in contrast to the magnetic component in NLTT 12758, did not accrete angular
momentum during their evolutionary history or that their magnetic field has not been detected yet.
NLTT 11748 (the white dwarfs highlighted with cyan and magenta periods found among some close white dwarf plus FGK secondary
in Fig. 1). This leads to an important follow-up question: why is star binary systems (Hernandez et al. 2021). Two consecutive
NLTT 12758 the only close double white dwarf where a strongly common envelopes have been suggested as the formation scenario
magnetic component has been detected? If our hypothesis that the for NLTT 12758 (Kawka et al. 2017) and we therefore start our
rotation- and crystallization-driven dynamo generated the magnetic discussion with a review of this classical scenario for the formation
field in NLTT 12758 is correct, the other white dwarfs with crys- of close double white dwarfs.
tallizing cores must either not have sufficiently spun up during The change in orbital separation during common-envelope evo-
their formation or we simply have not been able to detect their lution can be calculated using a simple energy conservation equa-
strong magnetic fields. To understand magnetic field generation tion that relates the loss of orbital energy of the binary and the
in close double white dwarfs, we therefore need to investigate binding energy of the envelope. This energy conservation equation is
the evolutionary history of NLTT 12758 and consider potential typically parametrized with the common-envelope efficiency α CE and
evolutionary differences to NLTT 11748 and SDSS J1257+5482 as the binding energy parameter λ (e.g. Han, Podsiadlowski & Eggleton
well as possible observational biases. We start by taking a closer look 1995; Dewi & Tauris 2000; Zorotovic et al. 2010).
at the formation of NLTT 12758. While it is clear that the common-envelope efficiency α CE should
be between zero and one – there are strong indications for a relatively
low value between 0.2 and 0.3 (Zorotovic et al. 2010) – there
3 WA S N LT T 1 2 7 5 8 F O R M E D T H RO U G H T W O is no agreement in the literature on the most realistic value of
COMMON-ENVELOPE PHASES? the binding energy parameter λ mostly because it remains unclear
Most close white dwarf binary stars form through common-envelope how efficient recombination energy can contribute to the process of
evolution (Webbink 1984; Zorotovic et al. 2010; Ivanova et al. 2013) expelling a common-envelope (Webbink 2008; Ivanova, Justham &
that occurs when the more massive star fills its Roche lobe on the Podsiadlowski 2015; Sabach et al. 2017; Grichener, Sabach & Soker
first giant branch (FGB) or the asymptotic giant branch (AGB). The 2018; Ivanova 2018; Soker, Grichener & Sabach 2018).
resulting mass transfer is typically dynamically unstable and leads to Assuming two common-envelope phases and using the BSE code,
the formation of a gaseous envelope around the secondary star and the Kawka et al. (2017) found that the history of NLTT 12758 can
core of the giant. This envelope is expelled at the expense of orbital be reproduced assuming an initial binary with stellar masses of
energy and orbital angular momentum. Therefore, the emerging post- 3.75 and 2.88 M and an orbital period of 2656 d. We could
common-envelope binary stars have typically short orbital periods reproduce their result only by assuming that ∼75 per cent of
between a few hours and a few days (Nebot Gómez-Morán et al. the available recombination energy contributed to expelling the
2011). envelope. This assumption can be considered unrealistic as even
Common-envelope evolution has been shown to successfully ex- works that emphasize the potential importance of recombination
plain the observed close white dwarf plus M dwarf binary population energy exclude such large values (Ivanova et al. 2015). We believe
(Zorotovic et al. 2010) and seems to also reproduce the short that this large fraction of recombination energy is related to a
MNRAS 513, 3090–3103 (2022)

misinterpretation of the input parameters of BSE (see Fig. B1 for In the following section, we therefore discuss an alternative forma-
more details). tion scenario for NLTT 12758 that includes stable non-conservative
In order to investigate whether more realistic parameter combina- mass transfer and show that it offers a natural explanation for the
tions allow to reproduce the formation of NLTT 12758 through two observed binary and stellar parameters of the system as well as the
common-envelope phases, we reconstructed both potential common- short spin period of the magnetic white dwarf.
envelope phases following Zorotovic et al. (2010) and Zorotovic,
Schreiber & Parsons (2014). We found that without contributions
4 A N A LT E R N AT I V E F O R M AT I O N M O D E L :
from recombination energy the system cannot have formed by
C O M B I N I N G S TA B L E M A S S T R A N S F E R A N D
two consecutive common-envelope phases. Assuming a relatively
C O M M O N - E N V E L O P E E VO L U T I O N
small fraction of recombination energy (≤10 per cent, but at least
2 per cent) and a value of the common-envelope efficiency exceeding As noted first by Nelemans et al. (2000) two common-envelope
0.5, reasonable solutions exist. phases using the classical energy budget equation fail to explain
These possibilities exist because the λ parameter is extremely sen- several observed double white dwarf systems. To solve this issue, they
sitive to the inclusion of recombination energy. Even for an efficiency suggested the γ algorithm based on angular momentum conservation
of recombination energy of just 2 per cent, the λ parameter for the that allows for common-envelope evolution without spiral in during
more massive white dwarf progenitor (first common-envelope phase) the first mass transfer phase. However, by implicitly assuming energy
is >
∼ 3, leading to a much smaller reduction of the orbital separation conservation the γ algorithm is rather hiding the energy problem
with respect to simulations that do not consider recombination energy instead of solving it (Ivanova et al. 2013). As an alternative, Webbink
at all (where λ is closer to 1). (2008) suggested that instead the first mass transfer could be stable
This situation is similar to that of the two white dwarf plus main- and non-conservative, which occurs when the mass ratio of the two
sequence post-common-envelope binaries known so far that require progenitor stars is close to one (Ge et al. 2020). Detailed models
extra energy sources in order to explain their current orbital periods for double He-core white dwarfs were subsequently performed and
(IK Peg and KOI 3278). However, both systems can be reproduced by showed that stable non-conservative mass transfer indeed offers an
assuming that < ∼ 2 per cent of the recombination energy contributed explanation (Woods et al. 2012). We here explore the evolutionary
to the ejection process (Zorotovic et al. 2014) and for NLTT 12758 scenario in which the first white dwarf forms through stable mass
we need a slightly larger fraction. We therefore conclude that two transfer for NLTT 12758 using detailed stellar evolution calculations.
consecutive common-envelope phases remain a possible scenario
for the evolutionary history of NLTT 12758 but that an (uncertain
4.1 Simulating the evolution with MESA
and unusual) source of additional energy (possibly recombination
energy) is required. We performed our simulations of the binaries with the one-
If two common-envelope phases produced NLTT 12758, the white dimensional stellar and binary evolution code Modules for Experi-
dwarf with the smaller mass must have formed through the second ments in Stellar Astrophysics (MESA; Paxton et al. 2011, 2013, 2015,
common-envelope phase and the more massive white dwarf must 2018, 2019, r15140).
have formed through the first one. This is because the onset of We accounted for mass-loss through winds using the standard
common-envelope evolution is defined by dynamically unstable MESA implementations. In particular, for red giants on the FGB, we
mass transfer, which quickly reaches mass transfer time-scales that adopted the Reimers (1975) prescription, assuming a wind efficiency
are orders of magnitude shorter than the thermal time-scale of the of η = 0.5, which is consistent with metallicity-independent esti-
secondary star. The secondary star is therefore supposed to accrete mates using star clusters (McDonald & Zijlstra 2015). For AGB stars,
only a small amount of mass (according to Chamandy et al. 2018, we adopted the Bloecker (1995) recipe, assuming a wind efficiency
the total accreted mass is roughly in the range of 10−2 –10−4 M ). of η = 0.02, which is consistent with the calibration performed by
As in addition the spiral-in process significantly reduces the orbital Ventura, D’Antona & Mazzitelli (2000) using the luminosity function
separation, the initially less massive star cannot evolve further up the of lithium-rich stars in the Magellanic Clouds.
giant branches than the first one. In other words, in two consecutive The Roche lobe radius of each star was computed using the fit of
common-envelope phases the more massive white dwarf is formed Eggleton (1983) and mass transfer rates during Roche lobe overflow
first. For NLTT 12758, this implies that the fast-spinning and strongly (RLOF) are determined following the prescription of Ritter (1988).
magnetic white dwarf formed in the second common-envelope phase. Regarding wind accretion, we assumed two different prescriptions
If that was indeed the evolutionary history of NLTT 12758, it remains depending on the donor stars. For FGB stars, we adopted the Bondi–
an open question why the magnetic white dwarf is rapidly rotating. Hoyle–Lyttleton prescription (Hoyle & Lyttleton 1939; Bondi &
However, the fact that the cooling ages of both white dwarfs in Hoyle 1944), since in these cases the wind velocity is much larger
NLTT 12758 are very similar, i.e. 2.2 ± 0.2 and 1.9 ± 0.4 Gyr (Kawka than the orbital velocity of the accretor. On the other hand, winds
et al. 2017), might indicate that the formation history of the system from AGB stars can have velocities smaller or comparable to the
was actually different. If the two progenitor stars evolved off the orbital velocity of the accretor. In such situations, hydrodynamical
main sequence at about the same time, their masses must have been simulations have shown that efficient wind accretion is possible
similar and therefore the first mass transfer phase could have been through the wind Roche lobe overflow (wRLOF) mechanism (Mo-
stable. hamed & Podsiadlowski 2007), which is a regime between the
In fact, according to table 5 in Kawka et al. (2017), even in their Bondi–Hoyle–Lyttleton regime and RLOF. In other words, in the
solution for two common envelopes the first mass transfer started wRLOF mechanism it is assumed that the slow winds fill the donors
when both stars had virtually identical masses, which, according Roche lobe, which implies that wind accretion is enhanced.
to Ge et al. (2020), should lead to stable mass transfer instead Given the importance of this mechanism on the binary evolution
of common-envelope evolution if the mass transfer is not fully we investigate here, we implemented the wRLOF model in MESA, as
conservative (i.e. a fraction of the transferred mass leaves the described in Abate et al. (2013), Iłkiewicz et al. (2019), and Belloni
system). et al. (2020). Regarding how much of the donor wind is allowed
MNRAS 513, 3090–3103 (2022)

to be accreted, since the accretor cannot accrete more mass than 4.2 Searching for a model that can explain the properties of
is lost by the red giant star, we enforced that the accretion rate NLTT 12758
efficiency in the Bondi–Hoyle–Lyttleton regime cannot be greater
In order to search for a reasonable model that could reproduce
than 80 per cent. In addition, we assume that the accretion rate
the observational properties of NLTT 12758, we carried out several
efficiency in the wRLOF regime has to be ≤50 per cent, which is
sparse grids covering different regions of the parameter space,
consistent with hydrodynamical simulations (see Abate et al. 2013,
which include the ZAMS masses, the initial orbital period, the
and references therein). Finally, during each step of the simulations,
accretion efficiency during stable non-conservative mass transfer, and
after calculating the accretion rate efficiency in both regimes, we
the orbital energy conversion efficiency during common-envelope
adopted the higher one to be used in the wind mass transfer scheme
evolution (α CE ).
in MESA.
We found out that, if the magnetic WD in NLTT 12758 was formed
The zero-age main-sequence (ZAMS) stars are assumed to have
through stable non-conservative mass transfer, then its progenitor
no rotation. However, as the stars evolve, their rotations are allowed
ZAMS mass cannot be less massive than 3.60 M . This is because
to change according to MESA standard prescriptions (Paxton et al.
the core mass growth during the first white dwarf progenitor evolution
2013). In addition to rotation, we also allow the stars to eventually
is virtually entirely interrupted when it becomes a Roche lobe filling
synchronize with the orbit, due to tidal interaction. As usual, we
AGB star. This implies that the first white dwarf progenitor core mass
distinguish the phases in which the star has a radiative envelope from
at the onset of RLOF must be already comparable to the observed
those in which it has a convective envelope since the synchronization
value of the DAP white dwarf in NLTT 12758. This also implies
time-scales in the former case are generally orders of magnitude
that the first white dwarf progenitor ZAMS mass cannot be much
larger than those in the latter. In particular, given the star masses we
larger than ∼4.00 M , since stars with masses larger than that would
investigate, we assume that main-sequence and core-helium-burning
develop a more massive core before the onset of RLOF.
stars have radiative envelopes, while subgiant, FGB, and AGB stars
In addition, stable mass transfer from an evolved red giant only
have convective envelopes. During each star evolution, the synchro-
occurs when the ZAMS star masses are comparable, which means
nization time-scales are computed using the MESA standard prescrip-
that the mass ratio (more massive over the less massive) has to be
tion, which are based on Hut (1981) and Hurley, Tout & Pols (2002).
࣠1.1. Otherwise, the mass transfer will be most likely dynamically
When the initially more massive star becomes a white dwarf, we
unstable. Moreover, to reproduce the magnetic white dwarf mass
ignore the contribution of the white dwarf to the tidal interaction, as
and to have stable mass transfer, the red giant has to be on the early
its synchronization time-scale should be much longer than that of the
AGB. In case the mass transfer starts when the magnetic white dwarf
initially less massive star. However, since the white dwarf can accrete
progenitor is on the thermally pulsing AGB phase, the mass transfer
mass and angular momentum from the winds of its companion,
will be most likely dynamically unstable, given the huge changes in
especially in the case of a symbiotic binary with the giant donor
the star size during this phase.
being on the AGB, we allowed the WD to spin up following King,
Regarding the initial orbital period, assuming that the onset of
Regev & Wynn (1991) and Schreiber et al. (2021a). Measuring short-
RLOF occurs when the star is on the early AGB phase implies that
time variability in symbiotic stars is very difficult, due to the fact that
the initial orbital period cannot be longer than ∼700 d. On the other
steady nuclear burning on the white dwarf surface usually reduces the
hand, in order to reproduce the non-magnetic white dwarf mass in
amplitude of the variation, reducing in turn its detection probability.
NLTT 12758, the core mass of its progenitor needs to substantially
Despite that, there is some evidence that the white dwarf can spin up
grow before the onset of common-envelope evolution. This implies
during the symbiotic star phase from the studies of the stable periodic
that when stable non-conservative mass transfer ends the orbital
oscillations in two systems. In BF Cyg, the white dwarf’s spin period
period has to be ∼3500–4000 d, which in turn implies that the initial
is ≈108 min (Formiggini & Leibowitz 2009), while Z And hosts a
orbital period of the ZAMS binary cannot be shorter than ∼500 d.
magnetic white dwarf with a spin period of ≈28 min (Sokoloski &
However, we shall emphasize that this lower limit for the initial
Bildsten 1999). In both cases, the white dwarf most likely spun up
orbital period strongly depends on the accretion efficiency during
due to wind mass and angular momentum accretion.
stable mass transfer. In particular, the larger the accretion efficiency,
Depending on the accretion rate on to the white dwarf, hydrogen
the longer the orbital period at the end of stable mass transfer. This
shell burning could be stable, which results in an increase of its mass
correlation comes from the fact that the increase in orbital period
(e.g. Shen & Bildsten 2007; Wolf et al. 2013). We implemented the
due to mass transfer is much stronger than the increase caused by
critical accretion rate derived by Nomoto et al. (2007, their equa-
mass-loss from the system.
tion 5), above which white dwarfs are thermally stable; i.e. hydrogen
To have dynamically stable mass transfer when the donor is an
burns steadily in a shell. For accretion rates lower than this critical
early AGB star, we found that the accretion efficiency must be
value, the white dwarf is considered thermally unstable, which
smaller than ∼0.5, which means that at least half of the mass
means that the hydrogen shell burning is unstable to flashes and, for
leaving the early AGB donor is not accreted by the donor, i.e. is
sufficiently strong flashes, nova eruptions are triggered, causing that
lost from the binary. Such huge mass-loss may seem unrealistic at
most of the accreted mass will be expelled. In addition, we assume
first glance, but there are several mechanisms able to drive mass-
that there is a maximum possible accretion rate (Nomoto et al. 2007,
loss from double red giant binaries, such as jets, stellar winds,
their equation 6) such that white dwarfs accreting at rates above it will
circumbinary disc, and accretion disc flashes/outbursts/outflows.
burn stably at this maximum rate, and the remaining non-accreted
Recently, while modelling the system 2M17091769+312758, which
matter will be piled up to form a red-giant-like envelope. Since we
is a semidetached binary composed of a red giant star transferring
expect such high rates when the TP-AGB companion of the first white
mass to a subgiant star, Miller et al. (2021) found that to reproduce
dwarf is close to filling its Roche lobe, i.e. the binary is close to the
the observational properties of this system, around half of the mass
onset of common-envelope evolution, we assume that this piled-up
leaving the donor star actually escapes the binary, corresponding to
matter will be eventually lost during common-envelope evolution.
an accretion efficiency of ∼0.5, which is consistent with our finding.
Finally, we assume solar metallicity and set all other parameters
However, we of course agree with one of the greats in the field
in MESA as the default values in version 15140.
MNRAS 513, 3090–3103 (2022)

Figure 2. The revised evolutionary scenario for NLTT 12758 consists of stable mass transfer followed by a symbiotic phase and common-envelope evolution
after which the orbital period is similar to the one we observe today. According to this scenario, the less massive white dwarf forms first and accretes mass
and angular momentum during the second mass transfer phase. When this fast-spinning white dwarf starts to crystallize, the dynamo mechanism suggested by
Schreiber et al. (2021a) generates the strong magnetic field that has been detected by Kawka et al. (2017).
who stated concerning mass-loss that ‘details of this process remain tidal forces, which synchronize the rotation of the giant star with
obscure’ (Webbink 2008). the orbital motion, the orbital period decreases during this symbiotic
phase. This finding is consistent with observational indications that
most red giants in symbiotic stars are synchronized with the orbit
4.3 Reproducing the evolutionary history of NLTT 12758 (Zamanov et al. 2007).
The best-fitting model in our approach, combining dynamically As soon as the giant fills its Roche lobe dynamically, unstable
stable non-conservative mass transfer followed by dynamically mass transfer leads to common-envelope evolution, which then leaves
unstable mass transfer, is illustrated in Fig. 2. The ZAMS stars have behind a close double white dwarf with a fast-spinning 0.69 M
masses of 3.85 and 3.50 M and the orbital period is 670 d. The first and a more massive 0.825 M component. For a common-envelope
mass transfer is stable but highly non-conservative. During this mass efficiency of 0.3, considering no contributions from recombination
transfer phase, the lower mass white dwarf of the system is formed, energy, and using λ as calculated by BSE, the orbital period of this
the secondary (the initially less massive) star reaches 4.49 M , and emerging double white dwarf binary is 1.16 d. Within the following
the period increases to 3814 d. ∼2 Gyr, its period decreases to 1.154 d due to angular momentum
As soon as this star evolves into a TP-AGB star, the binary becomes loss by the emission of gravitational radiation. The age difference
a symbiotic star and the white dwarf efficiently accretes mass and between the two white dwarfs is just 50 Myr. The two white dwarfs
angular momentum from its wind and spins up, reaching the observed cool and start to crystallize, at first the more massive DA white dwarf
values of the mass (0.69 M ) and spin period (∼23 min). Due to that does not generate a strong magnetic field as this white dwarf is
MNRAS 513, 3090–3103 (2022)

Figure 3. Evolution with the DA progenitor (star 2) mass of some properties of the best-fitting model for NLTT 12758, prior the onset of the common-envelope
phase, namely star core and first WD masses (top left-hand panel), orbital period (top right-hand panel), mass transfer rate during stable non-conservative mass
transfer and wind mass-loss/accretion after the first WD is formed (bottom left-hand panel), and spin period of the DAP (first WD) progenitor as well as of the
DAP WD (bottom right-hand panel). The ranges of DA and DAP WD masses, as inferred from observations, are indicated as grey areas in the top left-hand
panel, and the spin period of the DAP WD as a dotted horizontal line in the bottom right-hand panel. We reasonably well reproduce those properties in our
modelling with the MESA code.
not rotating rapidly and roughly ∼0.7 Gyr later the lower mass white The evolution of the mass transfer is shown in the bottom left panel
dwarf. When the latter happens, the rotation- and crystallization- of Fig. 3. The red line shows that the stable mass transfer reaches
driven dynamo generates the strong magnetic field of the DAP white large values of ∼10−3 M yr−1 , and that roughly 3 per cent of the
dwarf we observe today. wind mass-loss rate (blue) is accreted by the white dwarf (green)
The evolution of both stars prior to the final common-envelope during the symbiotic phase. This mass accretion leads to the spin-up
phase is further illustrated in Fig. 3. We plotted the core masses, of the white dwarf that formed first as displayed in the bottom left
orbital period, mass transfer rate, and the spin period as a function panel. The final spin period depends therefore on the mass transfer
of the mass of the initially less massive star. rate and on the assumed spin-up efficiency.
As shown in the upper left panel, the mass of the white dwarf that The full evolution of both stars prior to the white dwarf formation
later forms the magnetic component (the white dwarf that formed is also illustrated in an HR diagram (Fig. 4). The mass ratio of both
first) increases between the main two mass transfer events due to wind stars is so close to one that the lower mass star is already on the
accretion (red dashed–dotted line). The core mass of the initially less FGB when the first white dwarf is formed. During the stable mass
massive star (blue line) has already reached more than 0.4 M when transfer phase, as the FGB accretor star is growing in mass, it also
stable mass transfer starts but increases further after the mass transfer rejuvenates and becomes hotter.
phase ends, until the system enters common-envelope evolution. The very good agreement in the stellar masses, orbital and spin
The upper left plot illustrates how the orbital period changes during period, and of the ages of the white dwarfs is of course the result of
the evolution. During the stable mass transfer phase, the period adjusting the values of the common-envelope efficiency, the mass-
significantly increases (red line) but it decreases before the onset loss parameter, the initial masses, and orbital period as well as the
of common-envelope evolution due to tidal forces when the initially angular momentum accretion efficiency. However, we would like to
lower mass star evolves on the AGB. emphasize that all other parameters for stellar evolution and mass
MNRAS 513, 3090–3103 (2022)

the low-mass white dwarf to be hotter (Teff 8700 ± 140 K) and with
an H-rich (DA) atmosphere while the companion is cooler (Teff =
7600 ± 120 K) and could therefore be crystallizing. Unfortunately,
the more massive and cooler white dwarf is not visible in the spectra
that have been taken of this object and it is therefore impossible to
tell whether it hosts a strong magnetic field or not.
It also remains unclear which of the two white dwarfs formed
first. White dwarfs with masses below 0.2 M , such as the brighter
component in NLTT 11748, can experience stable H burning for
several Gyr (Driebe et al. 1999; Serenelli et al. 2002; Panei et al.
2007). We can therefore not fully exclude that the cooler and more
massive white dwarf formed after the low-mass white dwarf. This
scenario requires the first mass transfer to be stable, similar to the
scenario we developed in this paper which was inspired by earlier
work (e.g. Woods et al. 2012). In this case, the more massive
white dwarf might not have accreted sufficient mass and angular
momentum and the dynamo scenario would not apply. If, in contrast,
the more massive white dwarf formed first, the accretion of some
Figure 4. HR diagram of both stars until the white dwarf formation. The mass and angular momentum during the second mass transfer phase
initially lower mass star (blue) is already on the FGB when mass transfer starts could be expected and the conditions for the dynamo to work would
(dashed–dotted). At the onset of common-envelope evolution, the initially
be met.
lower mass star evolved higher up the AGB than the first star. This way, the
The third double white dwarf with one component most likely
higher mass white dwarf is formed second.
crystallizing is SDSS J1257+5482. In this system, the lower mass
white dwarf seems to be significantly older than the more massive
transfer are fixed and that the presented solution does not represent a component. It is currently not clear how this system formed (Bours
rare or unusual evolution. In fact, we found many solutions that are in et al. 2015) and we therefore restrain ourselves from a discussion
very good agreement with the observations and could easily produce of the potential accretion of mass and angular momentum of the
a broad population of double white dwarfs that formed through the massive component of the system.
channel we propose for NLTT 12758. In addition, any model for the Interestingly, both Kulkarni & van Kerkwijk (2010) and Marsh
formation of double white dwarfs that we are aware of contains at et al. (2011) found that the Balmer lines of the massive component
least the same number of free parameters. are broadened significantly relative to what is expected from just
The agreement between theoretical predictions and observa- pressure broadening. Mechanisms that have been suggested for the
tions of the DAP white dwarf in NLTT 12758 is as good as it additional broadening are fast rotation and/or a magnetic field. It
gets. Our new evolutionary scenario combined with the rotation- might thus be that the massive component in SDSS J1257+5428 is
and crystallization-driven dynamo scenario explains the magnetic the second fast-spinning and crystallizing magnetic white dwarf in
field generation in NLTT 12758. This dynamo mechanism is also a double-degenerate binary. Testing this hypothesis would provide
consistent with the observational result that most white dwarfs important constraints on the evolutionary history of the binary.
that are members of close double white dwarfs are not magnetic Given the above discussion, we conclude that the predictions of the
because the vast majority of them are not crystallizing. The only crystallization- and rotation-driven dynamo scenario are consistent
exceptions are the DA companion to the magnetic white dwarf in with the observed incidence of magnetism among close double
NLTT 12758 and the more massive components in NLTT 11748 and white dwarfs. This separates the dynamo from the common-envelope
SDSS J1257+5428. In the next section, we briefly discuss these dynamo scenario previously suggested. According to the common-
white dwarfs. envelope idea for the origin of the magnetic fields in white dwarfs
(e.g. Tout et al. 2008), a large number of hotter white dwarfs that
are members of double white dwarfs should be magnetic as shown
5 SHOULD THE MASSIVE COMPONENTS IN
by Belloni & Schreiber (2020). This does not seem to be the case:
N LT T 1 2 7 5 8 , N LT T 1 1 7 4 8 , A N D
among the relatively well-studied 57 double white dwarfs listed in
S D S S J 1 2 5 7 +5 4 8 2 B E M AG N E T I C ?
Table A1 only for one white dwarf, the fast-spinning component in
The more massive DA white dwarf in NLTT 12758 is clearly NLTT 12758, the detection of a magnetic field has been reported.
crystallizing but according to our evolutionary scenario has never
had the possibility to accrete significant amounts of angular momen-
6 C O U L D A L L ( S T RO N G ) M AG N E T I C F I E L D S
tum. According to the rotation- and crystallization-driven dynamo
I N W H I T E DWA R F S B E G E N E R AT E D B Y T H E
scenario, it is therefore not surprising that this white dwarf is not
DY N A M O ?
strongly magnetic.
In the case of NLTT 11748, the situation is more difficult to While representing a promising scenario for a relatively large
evaluate because neither the evolutionary history of the system nor number of magnetic fields in white dwarfs in different settings that
the magnetic or non-magnetic nature of the massive white dwarf is are otherwise inexplicable, the rotation- and crystallization-driven
known. Spectral observations of NLTT 11748 by Kawka & Vennes dynamo is certainly not the only mechanisms producing strongly
(2009) revealed the brighter component to be a low-mass (<0.2 M ) magnetic white dwarfs. As illustrated in fig. 16 of Ferrario et al.
helium-core white dwarf and Steinfadt et al. (2010) demonstrated (2015), white dwarfs with field strength exceeding 100 MG are found
that the companion is a relatively massive C/O white dwarf. Kaplan to be slow rotators, which is inconsistent with one of the conditions
et al. (2014) analysed high-precision eclipse light curves and found proposed for the dynamo. In addition, carefully inspecting the full
MNRAS 513, 3090–3103 (2022)

sample of known magnetic white dwarfs in close binaries reveals angular momentum, and common-envelope evolution as the second
that not even all magnetic fields of white dwarfs in these systems can main mass transfer phase led to the formation of NLTT 12758. We
be explained by the dynamo. performed MESA simulations that illustrate how well this revised
The magnetic CV HY Eridiani contains a white dwarf that is most evolutionary scenario can reproduce the observations of the system.
likely consisting of an He core as its mass is below 0.47 M at Our simulations show how the combination of stable mass transfer
a 90 per cent confidence level. The dynamo proposed by Schreiber and a subsequent common-envelope event can lead to a double
et al. (2021a), however, relies on the crystallization of a C/O core, white dwarf binary consisting of two relatively massive C/O white
which appears unlikely (although not impossible) for HY Eridiani. dwarfs. In particular, and most importantly in the context of this
One of the biggest strengths of the crystallization- and rotation- paper, they offer an explanation for the increased rotation rate of the
driven dynamo scenario is that it explains the large number of magnetic white dwarf, which represents a key ingredient for magnetic
magnetic CVs and the absence of strongly magnetic white dwarfs field generation according to the rotation- and crystallization-driven
among young detached post-common-envelope binaries. However, dynamo.
in two cases, weakly magnetic white dwarfs have been detected in Our finding that the dynamo scenario can consistently explain
young close detached white dwarf binaries with main-sequence star the magnetic nature of NLTT 12758 (and the absence of clear signs
companions. of magnetism among close double white dwarfs otherwise) adds
The first example is the V471 Tau, with an estimated field strength another piece of evidence to the already available support for the
of a few hundred kG (Sion et al. 2012), and a white dwarf dynamo scenario for the generation of strong magnetic fields in
temperature of 34 500 K, which clearly excludes crystallization and white dwarfs: the dynamo is the only scenario that can explain
thus the dynamo to be operating. An alternative scenario that can the absence of strongly magnetic white dwarfs among young post-
be excluded to be the dominant formation mechanism of magnetic common-envelope binaries, it offers an explanation for the existence
white dwarfs might be at work here: fossil fields. V471 Tau is of the radio pulsing white dwarf binary AR Sco, it is consistent with
most likely the descendant of a compact triple-star system as the the high occurrence rate of magnetic white dwarfs in CVs (Schreiber
white dwarf is the youngest and most massive white dwarf in et al. 2021a), it naturally explains the absence of high-accretion
the Hyades (O’Brien, Bond & Sion 2001). This potential history rate intermediate polars in globular clusters (Belloni et al. 2021),
offers an explanation for the magnetic nature of the white dwarf and seems to be the reason behind the observed relation between
as magnetic main-sequence Ap/Bp stars are likely the result of magnetism and metal pollution (Schreiber et al. 2021b). Considering
main-sequence star mergers (Ferrario et al. 2009; Schneider et al. that also single white dwarfs that are currently not metal polluted
2019). It might therefore be that the origin of the magnetic field of might have accreted planetary material in the past and therefore
the white dwarf in V471 Tau is best explained by the fossil field gained angular momentum (Schreiber et al. 2021b), the model might
scenario. also explain the magnetism of old white dwarfs that currently do not
The second magnetic white dwarf in a young close detached show signs of accretion (Bagnulo & Landstreet 2021).
binary, CC Ceti, was discovered only recently with a field strength All this recent evidence shows that the idea initially put forward
of 600–700 kG (Wilson et al. 2021). The white dwarf in this system by Isern et al. (2017) and further developed by Schreiber et al.
is most likely an He-core white dwarf and even if it was not, it is (2021a) has the potential to significantly help to finally solve the
clearly too hot (25 203 ± 42 K) to be crystallizing. long-standing mystery of the origin of strong magnetic fields in
While the crystallization- and rotation-driven dynamo offers a white dwarfs.
consistent explanation for the origin of the magnetic fields in many However, despite this success, the case cannot be closed yet.
white dwarfs, the examples listed earlier clearly demonstrate that it Magnetic white dwarfs exist that the dynamo model fails to explain.
cannot be the only mechanism producing magnetic white dwarfs. For This concerns slowly rotating magnetic single white dwarfs as well
massive and young magnetic single white dwarfs, stellar or common- as three magnetic white dwarfs in close binary stars (V471 Tau,
envelope merger events may offer an explanation. The magnetic field CC Ceti, and HY Eri) that either contain He-core white dwarfs or
in V471 Tau might be of the fossil field origin as the progenitor star of a non-crystallizing C/O white dwarf. At least for these objects,
the white dwarf was likely formed by a stellar merger. The origin of alternative scenarios for the generation of magnetic fields in white
the magnetic fields in the close binaries CC Ceti and HY Eri remains dwarfs need to be considered.
unclear. In addition, on the modelling side, the scenario remains rather
phenomenological as we currently do not have theories that allow us
to properly calculate how long it takes the magnetic field to emerge
7 CONCLUDING DISCUSSION
as soon as the dynamo kicks in. We also have no clue whether there
The magnetic white dwarf in NLTT 12758 is rapidly rotating and its are upper and lower limits for the crystallized mass fractions for the
temperature and mass are consistent with the core having started dynamo to properly work (and if so, what are their values?). Finally,
to crystallize. Therefore, the crystallization- and rotation-driven despite the rather detailed discussion presented in Schreiber et al.
dynamo can explain the occurrence of the detected magnetic field. (2021a), proper population synthesis that includes magnetic field
Furthermore, the model also explains that in all other known double generation through the dynamo has still to be performed.
white dwarfs no magnetic component has been found. From an observational point of view, we need to expand the work of
To understand the fast rotation of the magnetic white dwarf in Bagnulo & Landstreet (2021), who presented the first volume-limited
NLTT 12758, we investigated the evolutionary history of the system (but still rather small) sample of white dwarfs and investigated how
and found that a previously suggested sequence consisting of two the incidence of magnetism depends on key parameters. As in Sloan
common-envelope events appears to be unlikely as additional energy Digital Sky Survey (SDSS)-V white dwarfs are for the first time prime
sources are required and given that the cooling ages of both white targets of an SDSS survey, we may indeed be able to analyse a large
dwarfs are rather similar. volume-limited sample of white dwarfs in the near future. Combining
We propose that instead a phase of stable mass transfer, a symbiotic this sample with population models of single white dwarfs and those
phase during which the first formed white dwarf accretes mass and that are part of binary systems would provide key constraints on the
MNRAS 513, 3090–3103 (2022)

dynamo scenario and should bring us closer to finally understand the Bragaglia A., Greggio L., Renzini A., D’Odorico S., 1990, ApJ, 365,
origin of strongly magnetic white dwarfs. L13
Braithwaite J., Spruit H. C., 2004, Nature, 431, 819
Brandenburg A., 2014, ApJ, 791, 12
AC K N OW L E D G E M E N T S Brown W. R., Kilic M., Hermes J. J., Allende Prieto C., Kenyon S. J., Winget
D. E., 2011, ApJ, 737, L23
We thank the researchers that provided the microphysics necessary
Brown W. R., Kilic M., Allende Prieto C., Gianninas A., Kenyon S. J., 2013,
to develop MESA. The MESA equation of state is a blend of the OPAL ApJ, 769, 66
(Rogers & Nayfonov 2002), SCVH (Saumon, Chabrier & van Horn Brown W. R. et al., 2020, ApJ, 889, 49
1995), PTEH (Pols et al. 1995), HELM (Timmes & Swesty 2000), Buchler J. R., Yueh W. R., 1976, ApJ, 210, 440
and PC (Potekhin & Chabrier 2010) equation of states. Radiative Burdge K. B. et al., 2019, Nature, 571, 528
opacities are primarily from OPAL (Iglesias & Rogers 1993, 1996), Burdge K. B. et al., 2020, ApJ, 905, L7
with low-temperature data from Ferguson et al. (2005) and the high- Cassisi S., Potekhin A. Y., Pietrinferni A., Catelan M., Salaris M., 2007, ApJ,
temperature, Compton-scattering-dominated regime by Buchler & 661, 1094
Yueh (1976). Electron conduction opacities are from Cassisi et al. Chamandy L. et al., 2018, MNRAS, 480, 1898
Chandra V. et al., 2021, ApJ, 921, 160
(2007). Nuclear reaction rates are a combination of rates from
Christensen U. R., Aubert J., 2006, Geophys. J. Int., 166, 97
NACRE (Angulo et al. 1999) and JINA REACLIB (Cyburt et al.
Christensen U. R., Holzwarth V., Reiners A., 2009, Nature, 457, 167
2010), plus additional tabulated weak reaction rates (Fuller, Fowler & Chugunov A. I., Dewitt H. E., Yakovlev D. G., 2007, Phys. Rev. D, 76,
Newman 1985; Oda et al. 1994; Langanke & Martı́nez-Pinedo 2000). 025028
Screening is included via the prescription of Chugunov, Dewitt & Claeys J. S. W., Pols O. R., Izzard R. G., Vink J., Verbunt F. W. M., 2014,
Yakovlev (2007). Thermal neutrino loss rates are from Itoh et al. A&A, 563, A83
(1996). Coughlin M. W. et al., 2020, MNRAS, 494, L91
MRS and MZ acknowledge support from FONDECYT (grant Cyburt R. H. et al., 2010, ApJS, 189, 240
1221059). MRS is furthermore supported by ANID – Millennium Dewi J. D. M., Tauris T. M., 2000, A&A, 360, 1043
Science Initiative Program – NCN19 171. DB was supported by Driebe T., Blöcker T., Schönberner D., Herwig F., 1999, A&A, 350, 89
Eggleton P. P., 1983, ApJ, 268, 368
ESO/Gobierno de Chile. BTG was supported by the UK STFC
Ferguson J. W., Alexander D. R., Allard F., Barman T., Bodnarik J. G.,
grant ST/T000406/1. SGP acknowledges the support of an STFC
Hauschildt P. H., Heffner-Wong A., Tamanai A., 2005, ApJ, 623, 585
Ernest Rutherford Fellowship. This project has received funding Ferrario L., Pringle J. E., Tout C. A., Wickramasinghe D. T., 2009, MNRAS,
from the European Research Council (ERC) under the European 400, L71
Union’s Horizon 2020 research and innovation programme (grant Ferrario L., de Martino D., Gänsicke B. T., 2015, Space Sci. Rev., 191,
agreement number 101020057). 111
For the purpose of open access, the author has applied a Creative Formiggini L., Leibowitz E. M., 2009, MNRAS, 396, 1507
Commons Attribution (CC BY) licence to any Author Accepted Fuller G. M., Fowler W. A., Newman M. J., 1985, ApJ, 293, 1
Manuscript version arising. Garcı́a-Berro E. et al., 2012, ApJ, 749, 25
Ge H., Webbink R. F., Chen X., Han Z., 2020, ApJ, 899, 132
Ginzburg S., Fuller J., Kawka A., 2022, preprint (arXiv:2202.12902)
DATA AVA I L A B I L I T Y Grichener A., Sabach E., Soker N., 2018, MNRAS, 478, 1818
Hallakoun N. et al., 2016, MNRAS, 458, 845
The MESA files required to reproduce our simulations will be provided Han Z., Podsiadlowski P., Eggleton P. P., 1995, MNRAS, 272, 800
upon request. Hermes J. J. et al., 2017, ApJS, 232, 23
Hernandez M. S. et al., 2021, MNRAS, 501, 1677
Holberg J. B., Saffer R. A., Tweedy R. W., Barstow M. A., 1995, ApJ, 452,
REFERENCES
L133
Abate C., Pols O. R., Izzard R. G., Mohamed S. S., de Mink S. E., 2013, Hoyle F., Lyttleton R. A., 1939, Proc. Camb. Phil. Soc., 35, 405
A&A, 552, A26 Hurley J. R., Tout C. A., Pols O. R., 2002, MNRAS, 329, 897
Angulo C. et al., 1999, Nucl. Phys. A, 656, 3 Hut P., 1981, A&A, 99, 126
Aurière M. et al., 2007, A&A, 475, 1053 Iglesias C. A., Rogers F. J., 1993, ApJ, 412, 752
Bagnulo S., Landstreet J. D., 2020, A&A, 643, A134 Iglesias C. A., Rogers F. J., 1996, ApJ, 464, 943
Bagnulo S., Landstreet J. D., 2021, MNRAS, 507, 5902 Iłkiewicz K., Mikołajewska J., Belczyński K., Wiktorowicz G., Karczmarek
Bédard A., Bergeron P., Brassard P., Fontaine G., 2020, ApJ, 901, 93 P., 2019, MNRAS, 485, 5468
Belloni D., Schreiber M. R., 2020, MNRAS, 492, 1523 Isern J., Garcı́a-Berro E., Külebi B., Lorén-Aguilar P., 2017, ApJ, 836,
Belloni D., Mikołajewska J., Iłkiewicz K., Schreiber M. R., Giersz M., Rivera L28
Sandoval L. E., Rodrigues C. V., 2020, MNRAS, 496, 3436 Itoh N., Hayashi H., Nishikawa A., Kohyama Y., 1996, ApJS, 102, 411
Belloni D., Schreiber M. R., Salaris M., Maccarone T. J., Zorotovic M., 2021, Ivanova N., 2018, ApJ, 858, L24
MNRAS, 505, L74 Ivanova N. et al., 2013, A&AR, 21, 59
Bergeron P., Wesemael F., Liebert J., Fontaine G., 1989, ApJ, 345, L91 Ivanova N., Justham S., Podsiadlowski P., 2015, MNRAS, 447, 2181
Bergeron P., Saffer R. A., Liebert J., 1992, ApJ, 394, 228 Izzard R. G., 2004, PhD thesis, Univ. Cambridge
Beuermann K., Burwitz V., Reinsch K., Schwope A., Thomas H. C., 2020, Kaplan D. L. et al., 2014, ApJ, 780, 167
A&A, 634, A91 Kawka A., Vennes S., 2004, in Zverko J., Ziznovsky J., Adelman S. J., Weiss
Blackett P. M. S., 1947, Nature, 159, 658 W. W., eds, Proc. IAU Symp. 224, The A-Star Puzzle. Cambridge Univ.
Bloecker T., 1995, A&A, 297, 727 Press, Cambridge, p. 879
Bondi H., Hoyle F., 1944, MNRAS, 104, 273 Kawka A., Vennes S., 2009, A&A, 506, L25
Bours M. C. P. et al., 2014, MNRAS, 438, 3399 Kawka A., Vennes S., Schmidt G. D., Wickramasinghe D. T., Koch R., 2007,
Bours M. C. P. et al., 2015, MNRAS, 450, 3966 ApJ, 654, 499
Bovino S., Schleicher D. R. G., Schober J., 2013, New J. Phys., 15, Kawka A., Briggs G. P., Vennes S., Ferrario L., Paunzen E., Wickramasinghe
013055 D. T., 2017, MNRAS, 466, 1127
MNRAS 513, 3090–3103 (2022)

Kemp J. C., Swedlund J. B., Landstreet J. D., Angel J. R. P., 1970, ApJ, 161, Rogers F. J., Nayfonov A., 2002, ApJ, 576, 1064
L77 Sabach E., Hillel S., Schreier R., Soker N., 2017, MNRAS, 472, 4361
Kepler S. O. et al., 2013, MNRAS, 429, 2934 Saffer R. A., Liebert J., Olszewski E. W., 1988, ApJ, 334, 947
Kilic M. et al., 2014, MNRAS, 438, L26 Salaris M., Cassisi S., Pietrinferni A., Kowalski P. M., Isern J., 2010, ApJ,
Kilic M., Bédard A., Bergeron P., 2021a, MNRAS, 502, 4972 716, 1241
Kilic M., Brown W. R., Bédard A., Kosakowski A., 2021b, ApJ, 918, Saumon D., Chabrier G., van Horn H. M., 1995, ApJS, 99, 713
L14 Schneider F. R. N., Ohlmann S. T., Podsiadlowski P., Röpke F. K., Balbus S.
King A. R., Regev O., Wynn G. A., 1991, MNRAS, 251, 30P A., Pakmor R., Springel V., 2019, Nature, 574, 211
Koester D., Voss B., Napiwotzki R., Christlieb N., Homeier D., Lisker T., Schreiber M. R., Belloni D., Gänsicke B. T., Parsons S. G., Zorotovic M.,
Reimers D., Heber U., 2009, A&A, 505, 441 2021a, Nat. Astron., 5, 648
Kulkarni S. R., van Kerkwijk M. H., 2010, ApJ, 719, 1123 Schreiber M. R., Belloni D., Gänsicke B. T., Parsons S. G., 2021b, MNRAS,
Langanke K., Martı́nez-Pinedo G., 2000, Nucl. Phys. A, 673, 481 506, L29
Liebert J. et al., 2005, AJ, 129, 2376 Schwope A. D., Nebot Gomez-Moran A., Schreiber M. R., Gänsicke B. T.,
Liebert J., Ferrario L., Wickramasinghe D. T., Smith P. S., 2015, ApJ, 804, 2009, A&A, 500, 867
93 Serenelli A. M., Althaus L. G., Rohrmann R. D., Benvenuto O. G., 2002,
McDonald I., Zijlstra A. A., 2015, MNRAS, 448, 502 MNRAS, 337, 1091
Marsh T. R., 1995, MNRAS, 275, L1 Shen K. J., Bildsten L., 2007, ApJ, 660, 1444
Marsh T. R., Dhillon V. S., Duck S. R., 1995, MNRAS, 275, 828 Sion E. M., Bond H. E., Lindler D., Godon P., Wickramasinghe D., Ferrario
Marsh T. R., Gänsicke B. T., Steeghs D., Southworth J., Koester D., Harris L., Dupuis J., 2012, ApJ, 751, 66
V., Merry L., 2011, ApJ, 736, 95 Soker N., Grichener A., Sabach E., 2018, ApJ, 863, L14
Marsh T. R. et al., 2016, Nature, 537, 374 Sokoloski J. L., Bildsten L., 1999, ApJ, 517, 919
Maxted P. F. L., Marsh T. R., 1999, MNRAS, 307, 122 Steinfadt J. D. R., Kaplan D. L., Shporer A., Bildsten L., Howell S. B., 2010,
Maxted P. F. L., Marsh T. R., Moran C. K. J., Han Z., 2000, MNRAS, 314, ApJ, 716, L146
334 Timmes F. X., Swesty F. D., 2000, ApJS, 126, 501
Maxted P. F. L., Marsh T. R., Moran C. K. J., 2002a, MNRAS, 332, 745 Tout C. A., Wickramasinghe D. T., Ferrario L., 2004, MNRAS, 355, L13
Maxted P. F. L., Burleigh M. R., Marsh T. R., Bannister N. P., 2002b, MNRAS, Tout C. A., Wickramasinghe D. T., Liebert J., Ferrario L., Pringle J. E., 2008,
334, 833 MNRAS, 387, 897
Miller A., Kounkel M., Sun M., Dixon D., Boggio C., Covey K. R., Stassun Vennes S. et al., 2011, ApJ, 737, L16
K. G., Mathieu R., 2021, AJ, 162, 131 Ventura P., D’Antona F., Mazzitelli I., 2000, A&A, 363, 605
Mohamed S., Podsiadlowski P., 2007, in Napiwotzki R., Burleigh M. R., eds, Webbink R. F., 1984, ApJ, 277, 355
ASP Conf. Ser. Vol. 372, 15th European Workshop on White Dwarfs. Webbink R. F., 2008, in Milone E. F., Leahy D. A., Hobill. D. W., eds,
Astron. Soc. Pac., San Francisco, p. 397 Short-Period Binary Stars: Observations, Analyses, and Results. Springer,
Morales-Rueda L., Marsh T. R., Maxted P. F. L., Nelemans G., Karl C., Berlin, p. 233
Napiwotzki R., Moran C. K. J., 2005, MNRAS, 359, 648 Wesemael F. et al., 1994, ApJ, 429, 369
Moran C., Marsh T. R., Bragaglia A., 1997, MNRAS, 288, 538 Wickramasinghe D. T., Ferrario L., 2005, MNRAS, 356, 1576
Mullally F., Badenes C., Thompson S. E., Lupton R., 2009, ApJ, 707, L51 Wilson D. J., Toloza O., Landstreet J. D., Gänsicke B. T., Drake J. J., Hermes
Napiwotzki R. et al., 2020, A&A, 638, A131 J. J., Koester D., 2021, MNRAS, 508, 561
Nebot Gómez-Morán A. et al., 2011, A&A, 536, A43 Wolf W. M., Bildsten L., Brooks J., Paxton B., 2013, ApJ, 777, 136
Nelemans G., Verbunt F., Yungelson L. R., Portegies Zwart S. F., 2000, A&A, Woods T. E., Ivanova N., van der Sluys M. V., Chaichenets S., 2012, ApJ,
360, 1011 744, 12
Nelemans G. et al., 2005, A&A, 440, 1087 Zamanov R. K., Bode M. F., Melo C. H. F., Bachev R., Gomboc A., Stateva
Nomoto K., Saio H., Kato M., Hachisu I., 2007, ApJ, 663, 1269 I. K., Porter J. M., Pritchard J., 2007, MNRAS, 380, 1053
O’Brien M. S., Bond H. E., Sion E. M., 2001, ApJ, 563, 971 Zorotovic M., Schreiber M. R., Gänsicke B. T., Nebot Gómez-Morán A.,
Oda T., Hino M., Muto K., Takahara M., Sato K., 1994, At. Data Nucl. Data 2010, A&A, 520, A86
Tables, 56, 231 Zorotovic M., Schreiber M. R., Parsons S. G., 2014, A&A, 568, L9
Pala A. F. et al., 2020, MNRAS, 494, 3799
Panei J. A., Althaus L. G., Chen X., Han Z., 2007, MNRAS, 382, 779
Parsons S. G., Marsh T. R., Gänsicke B. T., Drake A. J., Koester D., 2011, A P P E N D I X A : C L O S E D O U B L E W H I T E DWA R F
ApJ, 735, L30 BINARIES
Parsons S. G. et al., 2020, Nat. Astron., 4, 690
We compiled a comprehensive list of close (periods below 35 d)
Parsons S. G., Gänsicke B. T., Schreiber M. R., Marsh T. R., Ashley R. P.,
Breedt E., Littlefair S. P., Meusinger H., 2021, MNRAS, 502, 4305 double white dwarfs. The resulting Table A1 should be complete
Paxton B., Bildsten L., Dotter A., Herwig F., Lesaffre P., Timmes F., 2011, with respect to double white dwarfs that contain a C/O white dwarf
ApJS, 192, 3 component with measured effective temperature. Such systems are
Paxton B. et al., 2013, ApJS, 208, 4 highlighted in bold and represent those that are plotted in Fig. 1 of
Paxton B. et al., 2015, ApJS, 220, 15 this paper.
Paxton B. et al., 2018, ApJS, 234, 34 We did not include all systems from the recently published large
Paxton B. et al., 2019, ApJS, 243, 10 sample of almost hundred extremely low-mass (ELM) white dwarfs
Pols O. R., Tout C. A., Eggleton P. P., Han Z., 1995, MNRAS, 274, 964 (Brown et al. 2020) as in all but one of these systems the stellar
Potekhin A. Y., Chabrier G., 2010, Contrib. Plasma Phys., 50, 82
components with a measured temperature are helium-core low-mass
Rebassa-Mansergas A., Parsons S. G., Garcı́a-Berro E., Gänsicke B. T.,
white dwarfs that are irrelevant for the rotation- and crystallization-
Schreiber M. R., Rybicka M., Koester D., 2017, MNRAS, 466, 1575
Reimers D., 1975, Mem. Soc. R. Sci. Liege, 8, 369
driven dynamo. The only exception, SDSS J1638+3500, is included
Ritter H., 1988, A&A, 202, 93 in the table.
MNRAS 513, 3090–3103 (2022)

Table A1. Parameters of the known close double white dwarfs with measured orbital periods. We ignored many ELM systems where the only component with
measured temperature is certainly a helium-core white dwarf. For a comprehensive list of such systems, see Brown et al. (2020). Components highlighted in
bold are (potential) C/O white dwarfs with measured Teff , which are the white dwarfs of relevance for this paper and which are plotted in Fig. 1.
Name M1 (M ) M2 (M ) Teff,1 (K) Teff,2 (K) Period Reference
NLTT11748 (WD 0342+176) 0.136–0162 0.707–0.740 8706 ± 136 7597 ± 119a 0.24 d 1
CSS 41177 0.38 ± 0.02 0.32 ± 0.01 24 500 11 500 2.78 h 2, 3
GALEXJ171708.5+675712 0.185 ± 0.01 ≥0.86 14 900 ± 200 – 5.90 h 4
SDSS J065133.33+284423.3 0.25 0.55 16 400 ± 300 – 12.75 min 5
SDSS J075141.18−014120.9 0.19 ± 0.02 0.97+ 0.06
−0.01 15 750 ± 240 – 1.92 h 6
SDSS J115219.99+024814.4 0.47 ± 0.11 0.44 ± 0.09 25 500 ± 1000 14 350 ± 500 2.40 h 7, 8
ZTF J153932.16+502738.8 0.61+ 0.017
−0.022 0.210+ 0.014
−0.015 48 900 ± 900 ≤10 000 6.91 min 9
ZTF J190125.42+530929.5 – – 28 000 ± 500 17 600 ± 400 40.60 min 10
ZTF J2243+5242 0.349+ 0.093
−0.074 0.384+ 0.114
−0.074 22 200+ 1800
−1600 16 200+ 1200
−1000 8.8 min 11
PG 1632+177 0.392+ 0.069
−0.059 0.526+ 0.095
−0.082 8800 ± 500 11 200 ± 500 2.05 d 12
WD 1534+503 0.392+ 0.069
−0.059 0.617+ 0.110
−0.096 8900 ± 500 8500 ± 500 0.71 d 12
SDSS J033816.16−813929.9 0.23 ± 0.015 0.38 + 0.05–0.03 18 100 ± 300 10 000 ± 1000 30.6 min 13
SDSS J063449.92+380352.2 0.452+ 0.070
−0.062 0.209+ 0.034
−0.021 27 300+ 4000
−2900 10 500+ 300
−200 26.5 min 13
WD 1434+503 (SDSS J143633.29+501026.8) 0.23(01) – 17 120(200) – 1.15 h 14
WD 1050+522 (SDSS J105353.89+520031.0) 0.22(01) – 16 150(200) – 0.96 h 14
SDSS J1337+3952 0.46 ± 0.02 0.26 ± 0.01 9450 ± 80 7520 ± 170 99 min 15
WD0028-474 0.60 ± 0.06 0.45 ± 0.04 18 500 ± 500 17 000 ± 500 9.35 h 16
SDSS J0318−0107 . 0.4 ± 0.05 0.49 ± 0.05 14 500 ± 500 13 500 ± 500 45.9 h 16
HE0410-1137 0.51 ± 0.04 0.39 ± 0.03 16 000 ± 500 19 000 ± 500 12.2 h 16
NLTT 12758 0.83 ± 0.03 0.69 ± 0.05 7950 ± 50 7220 ± 180 1.154 d 17
WD1202+608 (Feige 55) 0.3–0.487 ≥0.25 56 300 ± 1000 – 1.493 d 18
SDSS J125733.63+542850.5 0.24 1.06 ± 0.05 6400 ± 50 13 030 ± 150 4.6 h 19
WD 1704+481 0.39 ± 0.05 0.56 ± 0.07 9000 10 000 0.145 d 20, 21
WD0136+768 0.47 0.37 18 500 10 500 1.41 d 21, 22
WD1704+481 0.39 ± 0.05 0.54 9000 10 000 0.145 d 21
WD0957-666 0.37 ± 0.02 0.32 ± 0.03 30 000 11 000 1.46 h 21, 23, 24
WD1204+450 0.46 0.52 31 000 16 000 1.603 d 21
WD0135-052 (L870-2) 0.47 ± 0.05 0.52 ± 0.05 7470 ± 500 6920 ± 500 1.556 d 25, 26
WD1101+364 (= PG1101+364) 0.29 0.33 15 500 12 000 0.145 d 27
PG1115+166 0.70 0.70 – – 722.2 h 28
SDSS J174140.49+652638.7 0.17 ± 0.02 ≥1.11 10 540 ± 170 – 1.47 h 6
WD0225-192 (HE0225-1912) 0.55 0.23 20 488 – 0.22 d 29
WD0315-013 (HE0315-0118) 0.50 0.49 12 720 – 1.91 d 29
WD0320-192 (HE0320-1917) 0.31 0.45 13 248 – 0.86 d 29, 30, 31
WD0326-273 0.364 ≥0.96 9158 – 1.88 d 29, 30, 31
WD0453-295 0.40 0.44 16 360 1330 0.36 d 29, 32
WD1013-010 0.32 ≥ 0.62 8080 – 0.44 d 29, 30, 31
WD1022+050 0.37 ≥0.28 14 693 – 1.16 d 29, 31, 33, 34
HS1102+0934 0.38 ≥0.45 16 961 – 0.55 d 29, 30, 31, 35
WD1210+140 0.33 ≥0.44 32 127 – 0.64 d 29, 30, 31
HS1334+0701 0.35 – 16 891 – 0.23 d (uncertain) 29
WD1349+144 0.55 0.33 19 917 – 2.21 d 29, 30, 31
HE1414-0848 0.52 0.74 11 133 – 0.52 d 29, 31, 34
HE1511-0448 0.50 ≥0.67 50 899 – 3.22 d 29, 30, 31
WD1824+040 0.4 ≥0.73 14 787 – 6.27 d 25, 29, 31, 34
WD2020-425 0.81 0.54 34 004 – 0.3 d 29, 30
HE2209-1444 0.43 0.72 8471 – 0.28 d 29, 31, 34
WD1428+373 0.35 ≥0.23 – – 1.14 d 34
WD2032+188 0.41 ≥0.47 – – 5.08 d 34
SDSS J0755+4800 0.42 ≥0.90 19 890 ± 350 – 0.55 d 35
SDSS J1104+0918 0.46 ≥0.55 16 710 ± 250 – 0.55 d 35
SDSS J1557+2823 0.49 ≥0.43 12 550 ± 200 – 0.41 d 35
WD2331+290 0.39 ≥0.32 – – 0.17 d 36
WD1713+332 0.35 ≥0.18 – – 1.12 d 36
WD1241-010 0.31 ≥0.37 – – 3.35 d 36
MNRAS 513, 3090–3103 (2022)

Table A1 – continued
Name M1 (M ) M2 (M ) Teff,1 (K) Teff,2 (K) Period Reference
WD1317+453 0.33 ≥0.42 – – 4.87 d 36

SDSS J1638+3500 0.698 ± 0.030 – 37 250 ± 570 – 0.91 d 37
a Values correspond to assuming a thin envelope. References: (1) Kaplan et al. (2014), (2) Parsons et al. (2011), (3) Bours et al. (2014), (4) Vennes et al.
(2011), (5) Brown et al. (2011), (6) Kilic et al. (2014), (7) Hallakoun et al. (2016), (8) Parsons et al. (2020), (9) Burdge et al. (2019), (10) Coughlin et al.
(2020), (11) Burdge et al. (2020), (12) Kilic, Bédard & Bergeron (2021a), (13) Kilic et al. (2021b), (14) Mullally et al. (2009), (15) Chandra et al. (2021), (16)
Rebassa-Mansergas et al. (2017), (17) Kawka et al. (2017), (18) Holberg et al. (1995), (19) Bours et al. (2015), (20) Maxted et al. (2000) (21) Maxted, Marsh &
Moran (2002a), (22) Bergeron, Saffer & Liebert (1992), (23) Moran, Marsh & Bragaglia (1997), (24) Bragaglia et al. (1990), (25) Saffer, Liebert & Olszewski
(1988), (26) Bergeron et al. (1989), (27) Marsh (1995), (28) Maxted et al. (2002b), (29) Napiwotzki et al. (2020), (30) Koester et al. (2009), (31) Nelemans et al.
(2005), (32) Wesemael et al. (1994), (33) Maxted & Marsh (1999), (34) Morales-Rueda et al. (2005), (35) Brown et al. (2013), (36) Marsh, Dhillon & Duck
(1995), and (37) Brown et al. (2020)
A P P E N D I X B : U N D E R S TA N D I N G T H E BSE recombination energy contributes to expelling the envelope (Ivanova

I N P U T PA R A M E T E R S et al. 2015). Other works provide much stricter limits such as Soker
et al. (2018), who found that at maximum 10 per cent of the available
According to the Kawka et al. (2017), agreement with the observed
recombination energy can help expelling the envelope.
characteristics of NLTT 12758 was achieved with BSE assuming a
The fact that we could not reproduce the evolution suggested by
common-envelope efficiency of α CE = 0.15, a Reimer’s mass-loss
Kawka et al. (2017) unless an unrealistic fraction of the recombi-
parameter η = 1.0, and solar metallicity. However, the value that
nation energy of the envelope contributes to the ejection (in both
was adopted for the binding energy parameter λ is not given in their
common-envelope phases) might indicate that the authors perhaps
paper. This parameter is of crucial importance in the BSE code as its
misinterpreted the input parameter λ of the BSE code. As noted by
value determines to what degree recombination energy is included
Zorotovic et al. (2014), according to the comments written in the
in the energy budget of common-envelope evolution. code by its developers one can read in the main routine (bse.f) that
We therefore evolved the initial binary they proposed as a progen- the input parameter called lambda is ‘the binding energy factor for
itor for NLTT 12758 using BSE assuming a large range of values for common-envelope evolution (0.5)’. This leads many users to believe
λ. We used the same parameters as Kawka et al. (2017) but different that λ is fixed with the value they write in the input file. However,
values for the fraction of recombination energy by varying α rec from the code has been updated after publication. In the README BSE
0 to 1 in steps of 0.1. Fig. B1 shows the results we obtained from file, it says that the updates made after the publication of the BSE
this exercise. The two panels show the orbital separation of both paper are documented in the header of the evolv2.f routine. One of
stars after the first (left-hand panel) and second (right-hand panel) the updates documented there (since 2001 March) says ‘The value
common-envelope phases as a function of the white dwarf mass. of lambda used in calculations of the envelope binding energy for
The assumed contributions from recombination energy are colour giants in common envelope is now variable (see function in zfuncs)’.
coded and the dashed lines indicate the separation given for the Also, if we take a look at the file called zfuncs.f it is explained
corresponding stage in Kawka et al. (2017). that the input parameter lambda (that in this routine is called ‘fac’)
Apparently, a high-efficiency α rec between 0.7 and 0.8 was used is the fraction of the recombination/ionization energy used in the
by Kawka et al. (2017); i.e. they assumed that more than 70 per cent energy balance. Digging into the equations, which were fitted by
of the available recombination energy contributes to the envelope Onno Pols and published in Robert Izzard’s thesis (Izzard 2004) and
ejection process. Under these assumptions, during the first common Claeys et al. (2014, their appendix A), one can confirm that the input
envelope the binding energy is reduced by a factor of ∼64 and ∼73, parameter called lambda, if positive, corresponds to the fraction of
respectively; i.e. λ is 64 and 73 times larger for α rec = 0.7 and 0.8 with recombination energy that is assumed to contribute to the envelope
respect to λ for α rec = 0. This implies that while the orbital separation ejection [called λion in Claeys et al. (2014) or α rec in Zorotovic
is reduced to ∼1 per cent during common-envelope evolution for the et al. (2014)]. A fixed value for λ can still be assumed in BSE by
model without recombination energy (af /ai ∼ 0.01), the simulations selecting a negative value. For example, if the input parameter is set
with α rec = 0.7 and 0.8 predict the separation after common envelope to −0.5, then the common-envelope equations will use λ = 0.5. If,
to be ∼41 and ∼44 per cent of the separation at the onset of common- on the other hand, the input parameter is set to +0.75, 75 per cent of
envelope evolution (af /ai ∼ 0.41 and ∼0.44). the recombination energy is going to be used in the calculation of λ,
We consider this assumption highly unlikely given that even which leads to an unrealistically small reduction in orbital separation
simulations in favour of contributions from recombination energy during common-envelope evolution.
conclude that at most 60 per cent of the initially available hydrogen
MNRAS 513, 3090–3103 (2022)

Figure B1. White dwarf mass versus orbital separation after first (left) and second (right) common-envelope phases for different fractions of the
ionization/recombination energy. The grey area is for the measured white dwarf masses (with their error) and the black dashed lines correspond to the
orbital separations obtained by Kawka et al. (2017) after each of the two common-envelope episodes, while the red dotted line in the right-hand panel is the
current orbital separation (Kawka et al. 2017). Apparently (and potentially without being aware of it), between 70 and 80 per cent of the available recombination
energy was assumed to contribute to expelling the envelope, which can be considered as unrealistic.
This paper has been typeset from a TEX/LATEX file prepared by the author.
MNRAS 513, 3090–3103 (2022)

Artigo de Astronomia 1

Uploaded by

Copyright:

Available Formats

Artigo de Astronomia 1

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Artigo de Astronomia 1

Uploaded by

Copyright:

Available Formats

MNRAS 513, 3090–3103 (2022) https://doi.org/10.

Magnetic dynamos in white dwarfs – III. Explaining the occurrence of

© 2022 The Author(s)

MNRAS 513, 3090–3103 (2022)

MNRAS 513, 3090–3103 (2022)

MNRAS 513, 3090–3103 (2022)

MNRAS 513, 3090–3103 (2022)

MNRAS 513, 3090–3103 (2022)

MNRAS 513, 3090–3103 (2022)

MNRAS 513, 3090–3103 (2022)

MNRAS 513, 3090–3103 (2022)

MNRAS 513, 3090–3103 (2022)

MNRAS 513, 3090–3103 (2022)

Name M1 (M ) M2 (M ) Teff,1 (K) Teff,2 (K) Period Reference

MNRAS 513, 3090–3103 (2022)

Name M1 (M ) M2 (M ) Teff,1 (K) Teff,2 (K) Period Reference

WD1317+453 0.33 ≥0.42 – – 4.87 d 36

A P P E N D I X B : U N D E R S TA N D I N G T H E BSE recombination energy contributes to expelling the envelope (Ivanova

MNRAS 513, 3090–3103 (2022)

MNRAS 513, 3090–3103 (2022)

You might also like