Eclipsing Post Common Envelope Binaries from the Catalina Surveys

Abstract

We analyse the Catalina Real-time Transient Survey light curves of 835
spectroscopically confirmed white dwarf plus main-sequence binaries from the
Sloan Digital Sky Survey (SDSS) with g<19, in search of new eclipsing systems.
We identify 29 eclipsing systems, 12 of which were previously unknown. This
brings the total number of eclipsing white dwarf plus main-sequence binaries to
49. Our set of new eclipsing systems contains two with periods of 1.9 and 2.3
days making them the longest period eclipsing white dwarf binaries known. We
also identify one system which shows very large ellipsoidal modulation (almost
0.3 magnitudes), implying that the system is both very close to Roche-lobe
overflow and at high inclination. However, our follow up photometry failed to
firmly detect an eclipse meaning that this system either contains a cool white
dwarf and hence the eclipse is very shallow and undetectable in our
red-sensitive photometry or that it is non-eclipsing. Radial velocity
measurements for the main-sequence stars in three of our newly identified
eclipsing systems imply that their white dwarf masses are lower than those
inferred from modelling their SDSS spectra. 13 non-eclipsing post common
envelope binaries were also identified, from either reflection or ellipsoidal
modulation effects. The white dwarfs in our newly discovered eclipsing systems
span a wide range of parameters, including; low mass (~0.3Msol), very hot
(80,000K) and a DC white dwarf. The spectral types of the main-sequence stars
range from M2 to M6. This makes our sample ideal for testing white dwarf and
low-mass star mass-radius relationships as well as close binary evolution.

Full text

arXiv:1211.0316v1 [astro-ph.SR] 1 Nov 2012
Mon. Not. R. Astron. Soc. 000, 1–14 (2012) Printed 5 November 2012 (MN L
A
T
E
X style file v2.2)
Eclipsing Post Common Envelope Binaries from the
Catalina Surveys
S. G. Parsons1,2⋆, B. T. G¨ansicke1, T. R. Marsh1, A. J. Drake3, V. S. Dhillon4,
S. P. Littlefair4, S. Pyrzas1,5, A. Rebassa-Mansergas2and M. R. Schreiber2,6
1Department of Physics, University of Warwick, Coventry, CV4 7AL
2Departmento de F´ısica y Astronom´ıa, Universidad de Valpara´ıso, Avenida Gran Bretana 1111, Valpara´ıso, Chile
3California Institute of Technology, 1200 E. California Blvd, CA 91225, USA
4Department of Physics and Astronomy, University of Sheffield, Sheffield S3 7RH, UK
5Instituto de Astronom´ıa, Universidad Cat´olica del Norte, Avenida Angamos 0610, Casilla 1280, Antofagasta, Chile
6Millennium Nucleus “Protoplanetary Disks in ALMA Early Science”, Universidad de Valparaiso, Av. Gran Bretana 111, Chile
Accepted 2012 October 31. Received 2012 October 24; in original form 2012 September 11
ABSTRACT
We analyse the Catalina Real-time Transient Survey light curves of 835 spectro-
scopically confirmed white dwarf plus main-sequence binaries from the Sloan Digital
Sky Survey (SDSS) with g < 19, in search of new eclipsing systems. We identify 29
eclipsing systems, 12 of which were previously unknown. This brings the total number
of eclipsing white dwarf plus main-sequence binaries to 49. Our set of new eclipsing
systems contains two with periods of 1.9 and 2.3 days making them the longest pe-
riod eclipsing white dwarf binaries known. We also identify one system which shows
very large ellipsoidal modulation (almost 0.3 magnitudes), implying that the system is
both very close to Roche-lobe overflow and at high inclination. However, our follow up
photometry failed to firmly detect an eclipse meaning that this system either contains
a cool white dwarf and hence the eclipse is very shallow and undetectable in our red-
sensitive photometry or that it is non-eclipsing. Radial velocity measurements for the
main-sequence stars in three of our newly identified eclipsing systems imply that their
white dwarf masses are lower than those inferred from modelling their SDSS spectra.
13 non-eclipsing post common envelope binaries were also identified, from either re-
flection or ellipsoidal modulation effects. The white dwarfs in our newly discovered
eclipsing systems span a wide range of parameters, including; low mass (∼0.3M⊙),
very hot (80,000 K) and a DC white dwarf. The spectral types of the main-sequence
stars range from M2 to M6. This makes our sample ideal for testing white dwarf and
low-mass star mass-radius relationships as well as close binary evolution.
Key words: binaries: close – binaries: eclipsing – stars: white dwarfs – stars: low
mass
1 INTRODUCTION
Around 25% of main-sequence binary systems have stars
that are close enough to each other that they will inter-
act at some point in their evolution (Willems & Kolb 2004).
This interaction is caused by one or both of the stars filling
its Roche lobe and causing material to flow from one star to
the other. This process can often lead to a common-envelope
(CE) phase. The CE phase gives birth to very close binaries
and is thought to lead to the creation of some of the Galaxy’s
most exotic objects, such as cataclysmic variables (CVs),
⋆steven.parsons@uv.cl
low-mass X-ray binaries, B-type subdwarfs (sdB stars), dou-
ble degenerates, short gamma ray burst (GRB) progenitors
and millisecond pulsars.
One of the most common outcomes of the CE phase are
the close detached white dwarf plus main-sequence binaries,
known as Post Common Envelope Binaries (PCEBs). These
systems offer a unique opportunity to study close binaries
without the added complications of accretion, hence these
systems can provide us with superb tests of both the com-
mon envelope phase itself and the longer-term angular mo-
mentum loss mechanisms that drive the evolution of many
interacting binary stars (Schreiber & G¨ansicke 2003).
There are now over 2000 known white dwarf plus
c
2012 RAS

2S. G. Parsons et al.
main-sequence binaries (Silvestri et al. 2006; Heller et al.
2011; Morgan et al. 2012; Liu et al. 2012), with the
largest and most homogeneous catalogue presented by
Rebassa-Mansergas et al. (2007, 2010, 2012a) using data
from the Sloan Digital Sky Survey (SDSS; York et al. 2000;
Adelman-McCarthy et al. 2008; Abazajian et al. 2009).
Among these ∼1/3 are thought to be close PCEBs
(Schreiber et al. 2010; Rebassa-Mansergas et al. 2011). This
rise in the discovery rate of PCEBs is reflected in a corre-
sponding rise in the number of eclipsing systems: 30 of the 37
currently known eclipsing PCEBs were identified in the last
3 years. Many of these were identified by observing large
radial velocity variations in the SDSS sub-spectra (each
SDSS spectrum is the average of typically three 15 minute
exposures or sub-spectra) (Nebot G´omez-Mor´an et al. 2009;
Pyrzas et al. 2009, 2012) or from searches for pulsations
from the white dwarf (Steinfadt et al. 2008). However,
an increasing number of eclipsing systems are now be-
ing discovered in large scale time-domain surveys such
as the Palomar Transit Factory (PTF) (Law et al. 2011,
2012), the multi-epoch SDSS photometric survey (Stripe 82)
(Becker et al. 2011)1and a survey at the Isaac Newton Tele-
scope (Almenara et al. 2012).
By far the most successful search for eclipsing PCEBs
was made by Drake et al. (2009, 2010) using data from the
CSS (Catalina Sky Survey) and the Catalina Real Time
Transient Survey (CRTS). They discovered 26 eclipsing sys-
tems; 13 were previously unknown eclipsing PCEBs, 6 were
previously known eclipsing PCEBs, 3 were eclipsing cata-
clysmic variables, 3 were sdB+dM eclipsing binaries and
1 turned out to be a double white dwarf eclipsing binary
(Parsons et al. 2011a). The primary aim of that study was
to detect transiting planets around white dwarfs, since even
Earth-sized planets would produce deep eclipses due to the
small size of white dwarfs. Therefore, Drake et al. (2010)
selected their targets from the white dwarf catalogue of
Eisenstein et al. (2006) supplemented with additional pho-
tometric objects from the SDSS that were selected using
the (u−g,g−r) colour plane with a cut that included the
majority of the Eisenstein et al. (2006) white dwarfs. This
selection rejects PCEBs where the companion star notice-
ably contributes in the gor rbands, and hence implies that
the optical colours of Drake et al’s targets are dominated by
their white dwarf components. Therefore, the Drake et al.
(2010) sample is heavily biased towards hot white dwarfs
with late-type companions.
Here we present a search for eclipsing PCEBs combin-
ing the large catalogue of WDMS binaries spectroscopically
identified in SDSS (Rebassa-Mansergas et al. 2012a) with
the detailed CSS light curves of these objects. We recover
all 17 previously known eclipsing PCEBs contained in our
target list, and identify 12 additional ones, plus one can-
didate eclipsing PCEB. The 12 newly discovered eclipsing
PCEBs have been followed-up with high-speed photometry.
1Note that only 6 of the 42 candidate white dwarf plus M dwarf
binaries in Becker et al. (2011) show evidence of a white dwarf
in their SDSS spectrum. Of these only one (SDSS J013851.54-
001621.6) has been confirmed as eclipsing (Parsons et al. 2012b).
2 DATA REDUCTION
We selected all targets from Rebassa-Mansergas et al.
(2012a) with g < 19, a total of 966 systems. Not all of these
systems have been observed as part of the CSS and several
targets were highly blended with nearby stars and hence the
resultant light curves were very poor. We discarded these
systems, resulting in 835 light curves in total.
The Catalina Sky Survey has been running since mid
2005 and is designed to discover Near-Earth Objects. It uses
the 0.7m f/1.9 Catalina Schmidt Telescope with an eight
square degree field of view. Full details of the CSS can be
found in Drake et al. (2009). The observing strategy is to
observe each field in a sequence of four 30-second expo-
sures, spaced evenly over approximately 30 minutes, typi-
cally reaching V magnitudes of 19 to 20. The CSS dataset
consists of fields covered from a few times to more than 400
times. We used data obtained up to November 2011.
In order to improve the calibration of the CSS pho-
tometry and hence the associated uncertainties, we decided
to perform differential photometry on the reduced (bias-
subtracted and flat-fielded) CSS images. This also allowed
us to identify images in which the target was not detected
(e.g. deeply eclipsing systems).
For each target we produced a series of 10′×10′image
cutouts from the CSS data, centred on the target. All of
these images contained a substantial number of additional
nearby stars to the main target. Since, by definition, all the
fields had been observed as part of the SDSS, the additional
sources had SDSS magnitudes and could be used to calcu-
late the zeropoint, and hence flux calibrate each frame. We
chose to use the SDSS rband magnitudes since this filter
most closely approximated the filterless response of the CSS
detectors.
The zero point of each frame was determined by extract-
ing all sources within it using Sextractor (Bertin & Arnouts
1996) and cross-matching them with the SDSS catalogue, we
selected all non-blended stars with magnitudes of 15 < r <
19.5 and δr < 0.05, as determined by the appropriate flags
in casjobs (Li & Thakar 2008). We then took the difference
between the extracted magnitudes and SDSS magnitudes,
removed any values more than 2.5σfrom the mean, and
took the median value as the zeropoint for that frame. This
corrected for any variations in the observing conditions and
also reduced the impact on the zeropoint of any genuinely
variable sources in the frame. We also flagged up frames in
which the target was not detected.
3 ECLIPSING SYSTEMS
We visually inspected each light curve in order to identify
eclipses. This was achieved by identifying any points signifi-
cantly fainter than the average magnitude of the star (∼2.5σ
from the mean), or any frames in which the target was not
detected. We then inspected the reduced images of each faint
point identified to ensure that the target was not on a bad
pixel or the edge of a CCD.
Once we had detected that a system was eclipsing we
attempted to determine its period. Initially we calculated a
periodogram from the light curve using the Press & Rybicki
(1989) method with inverse variance weights whereby data
c
2012 RAS, MNRAS 000, 1–14

Eclipsing PCEBs from the CRTS 3
Table 1. Identified eclipsing systems. The newly discovered systems are shown in bold and their parameters are taken from
Rebassa-Mansergas et al. (2012a) except for SDSS J1021+1744, SDSS J1028+0931 and SDSS J1411+1028 where we have radial velocity
information and hence are able to constrain the white dwarf masses using the mass function. The white dwarf is not visible spectro-
scopically in SDSS J0745+2631, and SDSS J1307+2156 contains a featureless DC white dwarf, hence these systems have no parameters
listed for their white dwarfs. For the previously discovered systems we list the current best constraints from the literature. References:
(1) Pyrzas et al. (2009), (2) Parsons et al. (2010b), (3) this paper, (4) Drake et al. (2010), (5) Parsons et al. (2012c), (6) Pyrzas et al.
(2012), (7) Nebot G´omez-Mor´an et al. (2009), (8) Parsons et al. (2012a).
SDSS Name WD mass WD Teff MS star rmag Period T0 Ref
(M⊙) (K) sp type (days) MJD(BTDB)
SDSS J011009.09+132616.1 0.47 ±0.20 25900 ±427 M4.0 16.86 0.332686752(1) 53993.949090(2) 1,2
SDSS J030308.35+005444.1 0.91 ±0.03 <8000 M4.5 18.06 0.13443767232(25) 53991.1172793(19) 1,2
SDSS J074548.63+263123.4†M2.0 17.46 0.2192638284(1) 53387.2495(10) 3
SDSS J082145.27+455923.4 0.66 ±0.05 80938 ±4024 M2.0 17.52 0.5090912(69) 55989.038796(23) 3
SDSS J083845.86+191416.5 0.39 ±0.04 13904 ±424 M5.0 18.36 0.13011225(40) 53495.4541(33) 4
SDSS J085746.18+034255.3 0.514 ±0.049 35300 ±400 M8.0 18.26 0.065096538(3) 55552.7127652(8) 4,5
SDSS J090812.04+060421.2 0.37 ±0.02 17505 ±242 M4.0 17.28 0.1494381329(27) 53466.333170(36) 4
SDSS J092741.73+332959.1 0.59 ±0.05 27111 ±494 M3.0 18.22 2.3082217(65) 56074.906137(21) 3
SDSS J093947.95+325807.3 0.52 ±0.03 28389 ±278 M4.0 18.03 0.330989655(21) 55587.308823(10) 4
SDSS J094634.49+203003.4 0.62 ±0.10 10307 ±141 M5.0 18.89 0.2528612195(1) 56032.945590(25) 3
SDSS J095719.24+234240.7 0.43 ±0.03 25891 ±547 M2.0 18.06 0.150870740(6) 55604.830124(6) 4
SDSS J095737.59+300136.5 0.42 ±0.05 28064 ±848 M3.0 18.78 1.9261278(10) 56014.975114(32) 3
SDSS J102102.25+174439.9 0.50 ±0.05 32595 ±928 M4.0 19.01 0.140359073(1) 56093.90558(12) 3
SDSS J102857.78+093129.8 0.42 ±0.04 18756 ±959 M3.0 15.58 0.235025762(1) 56001.093511(94) 3
SDSS J105756.93+130703.5 0.34 ±0.07 12536 ±978 M5.0 18.66 0.125162115(23) 56010.062214(14) 3
SDSS J121010.13+334722.9 0.415 ±0.010 6000 ±200 M5.0 16.16 0.124489764(1) 54923.033686(6) 6
SDSS J121258.25-012310.2 0.439 ±0.002 17707 ±35 M4.0 16.94 0.33587093(13) 54104.20917(48) 7,8
SDSS J122339.61-005631.1 0.45 ±0.06 11565 ±59 M6.0 18.04 0.0900780(13) 55707.0169865(72) 3
SDSS J124432.25+101710.8 0.40 ±0.03 21168 ±435 M5.0 18.34 0.2278562(2) 53466.3618(11) 4
SDSS J130733.49+215636.7 <8000 M4.0 17.42 0.2163221322(1) 56007.221371(16) 3
SDSS J132925.21+123025.4 0.35 ±0.08 12250 ±1032 M8.0 17.51 0.0809662550(14) 55271.05481841(97) 4
SDSS J134841.61+183410.5 0.59 ±0.02 15071 ±167 M4.0 17.19 0.24843148(1) 53833.3425(1) 4
SDSS J140847.14+295044.9 0.49 ±0.04 29050 ±484 M5.0 18.96 0.191790270(24) 56112.91291(18) 3
SDSS J141057.73-020236.6 0.47 ±0.06 29727 ±508 M3.0 18.85 0.363497(25) 53464.4880(36) 4
SDSS J141134.70+102839.7 0.36 ±0.04 30419 ±701 M3.0 19.13 0.16750990(10) 56031.172782(48) 3
SDSS J141536.40+011718.2 0.564 ±0.014 55,995 ±673 M4.5 17.30 0.344330838759(92) 42543.3377143(30) 8
SDSS J142355.06+240924.3 0.41 ±0.02 32972 ±318 M5.0 17.87 0.38200426(32) 53470.39985(18) 4
SDSS J143547.87+373338.5 0.40 ±0.04 12392 ±328 M5.0 17.25 0.12563114665(67) 54148.2035726(35) 1
SDSS J145634.30+161137.7 0.37 ±0.02 19149 ±262 M6.0 18.04 0.2291202(2) 51665.6720(34) 4
SDSS J223530.61+142855.0 0.45 ±0.06 21045 ±711 M4.0 18.83 0.1444564852(34) 55469.065554(86) 3
†Not confirmed as an eclipsing system
with smaller errors are given larger weightings. We then
folded the light curve on the peak frequency and visually
inspected the resultant light curve. In most cases the light
curves showed out-of-eclipse variations due to reflection or
ellipsoidal modulation effects. In these cases the out-of-
eclipse effects allowed us to find the correct period. However,
this approach is not ideal for systems that show no out-of-
eclipse variations (e.g. longer period systems). The period of
these systems could be measured using a box fit similar to
those used in exoplanet transit searches (e.g. Kov´acs et al.
2002). However, since we had knowledge of when the sys-
tem is both in and out of eclipse, we used a simpler (and
quicker) approach. We folded the data points over a large
range of periods and measured the phase dispersion of the
in-eclipse points. We rejected any period in which the in-
eclipse points are dispersed by more than 20% of the orbital
period. Furthermore, we insisted that there were no out-of-
eclipse points between the in-eclipse points. In all the cases
where we used this approach it led to a unique period, this
is due to the large number of observations for most targets.
Finally, in order to search for shallower eclipses, we
folded all of our light curves on the peak frequency of their
periodogram and inspected for eclipses. In cases where we
would expect ellipsoidal modulation (e.g. systems with dom-
inant main-sequence star contributions) we also folded the
light curves on half the value of the peak frequency, since
ellipsoidal modulation causes a double peaked shape in the
light curve.
In total we found 29 eclipsing systems, 12 of which were
previously unknown, and one candidate eclipsing PCEB
which needs better quality photometry for confirmation.
This increases the number of confirmed eclipsing PCEBs
by more than 20% from 37 to 49. All our identified eclips-
ing systems are detailed in Table 1. CSS light curves of the
newly identified eclipsing systems are shown in Figure 1, and
their SDSS spectra are shown in Figure 2.
The distribution of all known eclipsing PCEBs with
SDSS spectroscopy in the (u−g , g −r) colour plane is
shown in Figure 3. This illustrates that our newly discov-
ered eclipsing PCEBs generally contain M-stars with slightly
earlier spectral types than the previously known eclipsing
systems. This is unsurprising given that the majority of
c
2012 RAS, MNRAS 000, 1–14

4S. G. Parsons et al.
Figure 1. Phase-folded CSS light curves of the newly identified eclipsing PCEBs.
c
2012 RAS, MNRAS 000, 1–14

Eclipsing PCEBs from the CRTS 5
Figure 2. SDSS spectra of the newly identified eclipsing PCEBs. The hydrogen Balmer lines are indicated by red lines (absorption
features are from the white dwarf, emission features indicate an active or irradiated main-sequence star) and the Na ilines are indicated
by the green lines (absorption from the main-sequence star).
the previously known eclipsing PCEBs in this sample were
found by Drake et al. (2010) in a search for transiting plan-
ets around white dwarfs. The colour selection used in the
Drake et al. (2010) study was biased towards systems domi-
nated by the white dwarf, meaning that systems with earlier
main-sequence star spectral types were missed.
The orbital period distribution for all the SDSS eclips-
ing PCEBs is shown in Figure 4, as well as the period distri-
bution of all SDSS PCEBs from Nebot G´omez-Mor´an et al.
(2011). Unsurprisingly our systems generally have shorter
periods, this is primarily due to the fact that shorter period
systems can be eclipsing over a wider range of inclinations.
However, we have detected two eclipsing PCEBs with pe-
riods in excess of 1.9 days. This is much longer than the
previous longest period eclipsing PCEB, V471 Tau, which
has a period of only 0.52 days, although, as Figure 4 shows,
several non-eclipsing PCEBs have been found with peri-
ods this long, or longer (Nebot G´omez-Mor´an et al. 2011;
Rebassa-Mansergas et al. 2012b). Our ability to detect these
systems is due to the long baseline provided by the Catalina
Sky Survey.
3.1 System parameters
The masses, white dwarf temperatures and main-sequence
star spectral types were all taken from the catalogue
of Rebassa-Mansergas et al. (2012a). These were deter-
mined by decomposing and fitting the SDSS spectra, see
c
2012 RAS, MNRAS 000, 1–14

6S. G. Parsons et al.
Figure 3. Distribution of quasars (light gray dots), stars (dark
gray dots), and WDMS binaries (open circles, coloured sym-
bols) in the (u−g, g −r) colour plane. All WDMS bina-
ries shown here have SDSS DR7 spectropscopy. Our input tar-
get sample (open circles) included 835 WDMS binaries from
Rebassa-Mansergas et al. (2012a) with g < 19 and good qual-
ity CSS light curves. Analysing the CSS light curves of these 835
systems, we identify 29 eclipsing PCEBs of which 17 were pre-
viously known (magenta dots) and 12 are new discoveries (red
dots). One additional eclipsing PCEB candidate identified here is
marked by the red circle. Additional known eclipsing PCEBs that
have DR7 SDSS spectra but that were too faint for our magnitude
cut are shown as cyan dots. The eclipsing PCEBs announced by
Drake et al. (2009, 2010) are shown by blue crosses.
Rebassa-Mansergas et al. (2007) for a detailed explanation.
In brief, the technique first determines the spectral type
of the main-sequence star by fitting the SDSS spectrum
with a two-component model. The main-sequence compo-
nent is then subtracted and the residual white dwarf spec-
trum is fitted with a model grid of white dwarfs from
Koester (2010) to determine its temperature and surface
gravity, the mass is then determined using a mass-radius re-
lation for white dwarfs (Bergeron et al. 1995; Fontaine et al.
2001). This method gives a good first approximation of the
stellar parameters, and in most cases gives results consis-
tent with those obtained from high-precision studies (e.g.
Parsons et al. 2012a). However, as we will show in Section 5,
there are cases in which the deconvolution technique can give
erroneous results.
In principle the light curves can also be used to con-
strain the radii of the two stars and hence the masses via
a mass-radius relation. However, the CSS photometry does
not sample the white dwarf eclipse sufficently. Our follow-up
photometry can only be used to place a lower limit on the
size of the main-sequence star (Rsec /a, where ais the orbital
separation) and an upper limit on the size of the white dwarf
(RWD/a). Therefore, we adopt the decomposition values for
all further discussions.
4 FOLLOW UP PHOTOMETRY
We obtained follow up high-speed photometry of all our
newly identifed eclipsing systems. The majority of these
systems were observed with the high-speed camera RISE
Figure 4. Period distribution of all SDSS spectroscopically con-
firmed eclipsing PCEBs (black) and the orbital period distribu-
tion of all SDSS PCEBs from Nebot G´omez-Mor´an et al. (2011)
(grey). Since our detection efficiency is high (see Section 7.2) the
difference between the two distributions mainly reflects the (geo-
metric) probability of a system being eclipsing.
(Steele et al. 2008) on the Liverpool Telescope (LT). The
robotic nature of the LT makes it ideal to observe these sys-
tems, particularly the longer period ones. RISE is a frame
transfer CCD camera with a single wideband V+R filter and
negligible deadtime between frames. We observed one eclipse
of each of the newly identified systems using exposure times
of between 5 and 25 seconds, depending upon the brightness
of the target. The raw data are automatically run through
a pipeline that debiases, removes a scaled dark frame and
flat-fields the data.
We also observed two systems, S DSS J1223-0056 and
SDSS J2235+1428, with th e high-speed camera ULTRA-
CAM (Dhillon et al. 2007), mounted as a visitor instrument
on the New Technology Telescope (NTT) at La Silla.
In all cases the source flux was determined with aper-
ture photometry, using a variable aperture, whereby the ra-
dius of the aperture is scaled according to the FWHM, using
the ULTRACAM pipeline (Dhillon et al. 2007). Variations
in observing conditions were accounted for by determining
the flux relative to nearby comparison stars.
The follow up light curves of all the confirmed eclipsing
systems are shown in Figure 5. We list the mid-eclipse times
(T0) from our follow up observations in Table 1.
5 NOTES ON INDIVIDUAL SYSTEMS
SDSS J074548.63+263123.4
SDSS J0745+2631 was classified as a WDMS binary due to
a slight blue excess, there is no spectroscopic evidence of
a white dwarf in this system. The CSS light curve of this
system shows very large ellipsoidal modulation, but only
marginal evidence of an eclipse. The top-left panel of Fig-
ure 6 shows the CSS light curve of this system folded over
its 5.2 hour period. The amplitude of this ellipsoidal modu-
lation is related to the Roche lobe filling factor of the main-
sequence star. For SDSS J0745+2631 the amplitude is al-
c
2012 RAS, MNRAS 000, 1–14

Eclipsing PCEBs from the CRTS 7
Figure 5. Follow up light curves of the newly identified eclipsing systems. All the data were obtained using RISE on the Liverpool
Telescope except for SDSS J1223-0056 and SDSS J2235+1428 which were obtained using ULTRACAM on the NTT (the rband eclipses
are shown here). The dip seen in the light curve of SDSSJ1021+1744 is likely caused by material ejected from the main-sequence star
moving in front of the white dwarf. There also appears to be a flare from the main-sequence star during the egress of the white dwarf.
most 0.3 magnitudes, which is the maximum possible value,
implying that the main-sequence star almost fills its Roche
lobe. The amplitude is also related to the orbital inclination,
the large amplitude in this case implying that the inclina-
tion is high. Given that the flux of this system is dominated
by the main-sequence star at visible wavelengths, with no
white dwarf features seen in the SDSS spectrum (top-right
panel of Figure 6). Therefore, we would expect any eclipse
to be shallow, and may be beyond the precision of the CSS
data.
We determined which of the minima in the CSS light
curve of SDSS J0745+2631 corresponded to phase zero (the
c
2012 RAS, MNRAS 000, 1–14

8S. G. Parsons et al.
Figure 6. Top left: CSS light curve of SDSS J0745+2631 showing large ellipsoidal modul ation. The amplitude of this modulation is
almost at the maximum possible value (∼0.3 mags), implying that the main-sequence star is close to filling its Roche lobe. It also implies
that the inclination of the system is high. Top right: SDSS spectrum of SDSSJ0745+2631 showing that the M star dominates the overall
flux at optical wavelengths. Bottom left: Radial velocity measurements of the Na i8200˚
A doublet folded on the orbital period. Although
these measurements cannot be used to measure the radial velocity amplitude of the M star they allowed us to determine which of the
minima in the CSS light curve corresponded to phase zero (the eclipse of the white dwarf). Bottom right: follow up LT/RISE light curve
of SDSS J0745+2631 around the orbital phase of the putative white dwarf eclipse. The light curve appears to show a shallow eclipse-like
feature superimposed on top of the ellipsoidal modulation. However, small flares from the M star could cause a similar feature and we
are unable to say with certainty that the system is eclipsing. Photometry at shorter wavelengths, where the contribution from the white
dwarf is larger, and hence the eclipse deeper, will prove if this system is eclipsing or not.
putative eclipse of the white dwarf) using the radial ve-
locity measurements from the SDSS sub-spectra. In this
case there were only three measurements of the Na i8200˚
A
doublet all occuring at a similar orbital phase, with values
∼250 km s−1(bottom-left panel of Figure 6). These obser-
vations were obtained near quadrature (either phase 0.25 or
0.75) and hence must have been taken at phase 0.25 (since at
phase 0.75 we would expect the lines to be blueshifted rather
than redshifted), which allowed us to determine which min-
ima corresponded to phase zero.
Our follow-up RISE light curve is shown in the lower-
right panel of Figure 6. There is some evidence of a shal-
low dip around phase zero. However, the rapidly rotating
M2 star is likely to be active and hence we would expect it
to flare occasionally. A couple of unfortunately timed flares
would give the same shape as a shallow eclipse meaning
that, from this one observation alone, we cannot confirm
the eclipsing nature of this binary. Observations at bluer
wavelengths, where the contribution from the white dwarf
is larger, may reveal the eclipse.
SDSS J082145.27+455923.4
With a temperature of 80,000 K, the 0.66 M⊙white dwarf
in SDSS J0821+4559 is the hottest white dwarf known in an
eclipsing PCEB and a reflection component is easily visible
in the CSS light curve. The binary has a fairly long period of
12.2 hours. It also contains a main-sequence star with one
of the earliest spectral types, M2, in an eclipsing PCEB.
The SDSS spectrum is dominated by the hot white dwarf.
The Na i8200˚
A absorption doublet from the M star is also
tentatively detected.
SDSS J092741.73+332959.1
SDSS J0927+3329 is the longest p eriod eclipsing white dwarf
binary currently known, with a period of 2.3 days (55.4
hours). The 0.59 M⊙white dwarf is relatively hot but no
out-of-eclipse variations are seen in the CSS data. Both com-
ponents are easily detected in the SDSS spectrum.
c
2012 RAS, MNRAS 000, 1–14

Eclipsing PCEBs from the CRTS 9
SDSS J094634.49+203003.4
SDSS J0946+2030 contains a relatively cool 0.62 M⊙white
dwarf with an M5 main-sequence companion in a 6.1 hour
binary. Little variation is seen in the light curve outside of
the eclipse. Both components are well detected in the SDSS
spectrum.
SDSS J095737.59+300136.5
Its period of 1.9 days (46.2 hours) means that
SDSS J0957+3001 has the second longest period of all
known eclipsing PCEBs, behind SDSS J0927+3339. Unsur-
prisingly, there is no evidence of out-of-eclipse variations
in the CSS data. Both components are visible in the SDSS
spectrum. The hot white dwarf has a fairly low mass of
0.42 M⊙, and is hence likely to have a helium core.
SDSS J102102.25+174439.9
The CSS light curve of SDS S J1021+1744 is d ominated by
the main-sequence star and there is a clear ellipsoidal mod-
ulation component. Our follow-up LT/RISE photometry re-
vealed a large dip in the brightness of the system ∼15 min-
utes after the end of the eclipse (see Figure 5). There is also
some evidence for a flare from the main-sequence star occur-
ing during the egress of the white dwarf. It is possible that
this dip is caused by material ejected during this flare (or
a previous flare) passing in front of the white dwarf, simi-
lar features have been seen in the eclipsing PCEB QS Vir
(O’Donoghue et al. 2003; Parsons et al. 2011b). If this is the
case then a large amount of material must have been ejected,
since almost half of the white dwarf’s flux is blocked.
The SDSS spectrum is in fact a composition of 10 sub-
spectra. The radial velocity measurements of the Na i8200˚
A
absorption doublet reported by Rebassa-Mansergas et al.
(2012a) combined with our ephemeris allowed us to mea-
sure the radial velocity amplitude of the main-sequence
star as Ksec = 235 ±9 km s−1, with a systemic velocity of
γ=−20 ±6 km s−1(see Figure 7). With this information,
and the orbital period, we can constrain the mass of the
white dwarf, using the mass function,
f(MWD) = (MWD sin i)3
(MWD +Msec)2=Porb K3
sec
2πG .(1)
For a given inclination, i, Equation 1 defines the relation-
ship between MWD and Msec. The lower-left hand panel of
Figure 7 shows t his relationship for SDSS J1021+1744 for an
inclination of 90◦and 75◦, the probable range over which the
system is eclipsing. Also shown are the limits on the masses
of the two stars from the deconvolution of the SDSS spec-
trum. It is clear that the mass of the white dwarf from the
deconvolution (1.06 ±0.09M⊙) is a substantial overestimate
since the main-sequence star would have to have a mass in
excess of 1M⊙, certainly not an M dwarf. Assuming that the
constraint on the mass of the secondary star is correct, the
mass of the white dwarf is closer to 0.5M⊙. This discrepancy
could be caused by the fact that this system is faint and the
signal-to-noise of the SDSS spectrum is low. Furthermore,
the main-sequence star dominates the spectrum hence the
fit to the few white dwarf features visible is relatively poor.
SDSS J102857.78+093129.8
SDSS J1028+0931 is the brightest of our new eclipsing sys-
tems and has a period of 5.6 hours. The flux is dominated by
the main-sequence star at visible wavelengths, although the
white dwarf’s features are still visible in the SDSS spec-
trum. The CSS light curve shows evidence of ellipsoidal
modulation. There are 13 SDSS sub-spectra for this object.
Rebassa-Mansergas et al. (2012a) measured the radial ve-
locity of the main-sequence star from the Na i8200˚
A ab-
sorption doublet for each of these sub-spectra. Using these
measurements and our ephemeris we were able to deter-
mine that the radial velocity amplitude of main-sequence
star is Ksec = 164 ±5 km s−1, with a systemic velocity of
γ= 12 ±4 km s−1(upper-centre panel of Figure 7).
As with SDSS J1021+1744, we can use the Ksec
measurement to constrain the mass of the white dwarf.
The lower-centre panel of Figure 7 shows the relation-
ship between MWD and Msec for high inclinations. Like
SDSS J1021+1744 the mass of the white dwarf from the
deconvolution (0.80 ±0.04M⊙) is a substantial overesti-
mate. Assuming that the constraint on the mass of the sec-
ondary star is correct, the mass of the white dwarf roughly
0.42M⊙. However, like SDSS J1021+1744, the SDSS spec-
trum is dominated by the main-sequence star to such an
extent that it may have affected the fit to the white dwarf
features.
SDSS J105756.93+130703.5
The mass of the white dwarf in SDSS J1057+1307 deter-
mined from the spectral decomposition of the SDSS spec-
trum is 0.34 M⊙(Rebassa-Mansergas et al. 2010), making
it the lowest mass white dwarf in all our new eclipsing sys-
tems and likely to have a helium core. The SDSS spectrum
is dominated by the white dwarf but there are some features
from the main-sequence star at long (>7000˚
A) wavelengths.
The orbital period is almost exactly 3 hours.
SDSS J122339.61-005631.1
SDSS J1223-0056 was observed with ULTRACAM mounted
on the NTT as part of a project to detect pulsating white
dwarfs in white dwarf plus main-sequence binaries. No pul-
sations were seen, but an eclipse was recorded. Therefore, we
knew in advance that we might see eclipses in the CSS light
curve of this system. SDSS J1223-0056 is a partially eclips-
ing system and hence the eclipse only lasts around 5 minutes
(see Figure 5). Nevertheless, 3 CSS observations were taken
in eclipse. The CSS light curve also shows some evidence
of ellipsoidal modulation. This system has the shortest or-
bital period of all our newly discovered eclipsing systems, 2.1
hours. The white dwarf has a low mass of 0.45M⊙, making
it likely to be a helium core white dwarf. The main-sequence
star has a spectral type of M6, making it the latest spectral
type from all our new eclipsing systems.
SDSS J130733.49+215636.7
The white dwarf in SDS S J1307+2156 is a DC white dwarf
and hence has a featureless spectrum. Therefore, we have no
information on its mass. The temperature of the white dwarf
c
2012 RAS, MNRAS 000, 1–14

10 S. G. Parsons et al.
Figure 7. Radial velocity curves (top) and mass function plots (bottom) for SDSS J1021+1744 (left), SDSS J1028+0931 (centre) and
SDSS J1411+1028 (right). The dashed lines on the mass function plots indicate the limits on the mass of the main-sequence star
(horizontal) and white dwarf (vertical) based on the spectral deconvolution. In all cases the implied mass of the white dwarf is lower
than that determined from the spectral deconvolution.
is limited to <8000 K based on the lack of Balmer absorption
lines and a Galaxy Evolution Explorer (GALEX) near-UV
magnitude of 21.05 ±0.22. Our follow up LT/RISE photom-
etry of the eclipse revealed a very sharp ingress and egress,
lasting ∼25 seconds each (see Figure 5). The short dura-
tion of these features implies that the white dwarf is quite
small and is therefore likely to be quite massive. However,
radial velocity information is needed in order to constrain
the white dwarf’s mass. The CSS light curve also shows ev-
idence of ellipsoidal modulation over the 5.2 hour orbital
period.
SDSS J1307+2156 is only the second known eclips-
ing non-DA white dwarf after SDSS J0303+0054
(Pyrzas et al. 2009; Debes et al. 2012). There are 5
other non-DA white dwarfs in (non-eclipsing) PCEBs
(Nebot G´omez-Mor´an et al. 2011). Interestingly, all
these white dwarfs are featureless DC white dwarfs,
there is currently no known DB white dwarf in a
PCEB2. This deficit is significant because 27 WDMS
systems with DB white dwarfs have been spectro-
scopically followed up in order to determine whether
they are close PCEBs (Rebassa-Mansergas et al. 2012a;
Nebot G´omez-Mor´an et al. 2011). None of these showed
2Raymond et al. (2003) claim to observe a 150 km s−1
radial velocity variation in the DB+MS binary
SDSS J144258.47+001031.5, based on 2 measurements. However,
no variations are seen in the 18 radial velocity measurements
listed in Rebassa-Mansergas et al. (2012a), hence we conclude
that the measurements of Raymond et al. (2003) are erroneous
and this system is in fact a wide binary.
any radial velocity variations, despite the fact that we
would expect ∼1/3 to be PCEBs (Schreiber et al. 2010;
Rebassa-Mansergas et al. 2011), hence this deficit appears
to be genuine.
The lack of DB white dwarfs in PCEBs is likely due to
the white dwarf accreting some of the wind of its main-
sequence companion. Wind accretion rates on to white
dwarfs in PCEBs are of the order of 10−15 M⊙yr−1(Debes
2006; Tappert et al. 2011; Pyrzas et al. 2012; Parsons et al.
2012a) meaning that the white dwarf will accrete 10−7M⊙
of hydrogen in 100 Myr. The hydrogen in this wind will form
a layer on the surface of the white dwarf, turning a DB white
dwarf in to a DA white dwarf. However, DC white dwarfs are
much cooler and have much deeper outer convection zones
(Dufour et al. 2007) which mixes the accreted hydrogen to
such a low level that it is invisible. Hence we would still
expect to see DC white dwarfs in PCEBs, as we do.
SDSS J140847.14+295044.9
The 0.49 M⊙white dwarf in SDSS J1408+2950 is relatively
hot causing a small reflection effect, evident in the CSS light
curve. It has an orbital period of 4.6 hours. The SDSS spec-
trum is dominated by the white dwarf, however, a large emis-
sion component is visible in the Hαline. This emission is
much stronger than in any of the other new eclipsing sys-
tems. Some of the main-sequence stars in the other systems
are more highly irradiated than the one in SDSS J1408+2950
meaning that this emission could also be due to activity on
the main-sequence star.
c
2012 RAS, MNRAS 000, 1–14

Eclipsing PCEBs from the CRTS 11
Table 2. Non-eclipsing PCEBs identified from the CSS photometry. Systems were identified using either a reflection effect (R) or
ellipsoidal modulation (E). Stellar parameters taken from Rebassa-Mansergas et al. (2012a)
SDSS Name WD mass WD Teff Sp type of rmag Period Type Amplitude
(M⊙) (K) MS star (days) (mags)
SDSS J074807.22+205814.2 0.52 ±0.06 86726 ±7788 M2.0 18.55 0.07205455(1) R 0.035
SDSS J080304.61+121810.3 1.04 ±0.27 15071 ±2261 M2.0 17.14 0.5723126(29) E 0.023
SDSS J083618.61+432651.5 0.46 ±0.03 24726 ±521 M3.0 17.98 0.19689803(95) R 0.049
SDSS J091211.01+442057.8 M0.0 17.30 0.7311376(50) E 0.052
SDSS J091216.37+234442.5 0.69 ±0.03 30071 ±245 M3.0 17.66 0.2635582(5) R 0.065
SDSS J113316.27+270747.6 0.56 ±0.05 72971 ±4638 M1.0 18.00 0.1781628(12) R 0.023
SDSS J114509.77+381329.2 M4.0 15.95 0.19003799(27) E 0.027
SDSS J115857.33+152921.4 0.80 ±0.10 36996 ±1203 M3.0 18.86 0.06666328(14) R 0.053
SDSS J122630.86+303852.5 0.40 ±0.01 30071 ±66 M3.0 16.41 0.2586905(9) R 0.061
SDSS J122930.65+263050.4 1.04 ±0.08 21045 ±820 M3.0 17.30 0.6711480(66) E 0.045
SDSS J155904.62+035623.4†0.68 ±0.09 48212 ±2446 18.58 0.0943473(1) R 0.120
SDSS J162558.25+351035.7 K7.0 17.40 0.3856815(25) R 0.062
SDSS J173002.48+333401.8 0.44 ±0.03 47114 ±1176 18.39 0.1569473(3) R 0.140
†Found by Nebot G´omez-Mor´an et al. (2011)
SDSS J141134.70+102839.7
Tappert et al. (2011) first presented evidence that
SDSS J1411+1028 was a PCEB. Nebot G´omez-Mor´an et al.
(2011) determined its period as 4.0 hours and measured
the radial velocity amplitude of main-sequence star as
Ksec = 168 ±4 km s−1. Due to its eclipsing nature, our CSS
photometry gives tighter constraints on its period. We also
detect a reflection component in the out-of-eclipse light
curve. The spectrum is dominated by the relatively hot
white dwarf.
Since we have a measurement of the radial velocity am-
plitude of the main-sequence star we can use it to constrain
the mass of the white dwarf. The lower-right hand panel of
Figure 7 shows the relationship between MWD and Msec for
high inclinations. As with the two other systems where we
have radial velocity information, the mass of the white dwarf
determined from the deconvolution of the SDSS spectrum
(0.54 ±0.08M⊙) is an overestimate, although the discrep-
ancy is smaller in this case. Assuming that the constraint
on the mass of the secondary star is correct, the mass of the
white dwarf roughly 0.36M⊙, making it a firm helium core
candidate. In this case the low signal-to-noise of the SDSS
spectrum may have contributed to this overestimation.
SDSS J223530.61+142855.0
Like SDSS J1223-0056 we had prior knowledge that
SDSS J2235+1428 was an eclipsing system. The radial ve-
locity of the main-sequence star was observed to change by
almost 500 km s−1between two nights, implying that the
system was not only a PCEB, but that it was also a high in-
clination system. Subsequent photometric follow up revealed
that the system was eclipsing and that its period was 3.4
hours. The deep white dwarf eclipse is clearly visible in the
CSS light curve. However, there is little out-of-eclipse varia-
tion. The SDSS spectrum is dominated by the white dwarf.
The Na i8200˚
A absorption doublet from the main-sequence
star is just visible. With a mass of 0.45 M⊙the white dwarf
in SDSS J2235+1428 is another helium core candidate.
6 NON-ECLIPSING PCEBS
As noted in Section 3, we performed a period search on all
of our light curves to search for shallower eclipses. How-
ever, this approach also revealed the periods of several non-
eclipsing systems via reflection or ellipsoidal modulation ef-
fects.
Since these effects can be quite small we used 4 differ-
ent period analyses: Scargle (Scargle 1982), a straight power
spectrum, analysis of variance (AoV, Schwarzenberg-Czerny
1989) and ANoVA (Schwarzenberg-Czerny 1996). We con-
sider the period robust if all 4 methods give, within one
harmonic, the same period.
Using photometry to identify and determine the pe-
riods of PCEBs is not particularly efficient. A much
better approach is to try to detect radial velocity
variations, since these are easier to detect and much
less biased towards hot white dwarfs, short periods
and large Roche-lobe filling factors. This approach has
already led to the discovery of dozens of PCEBs
(Rebassa-Mansergas et al. 2007, 2010, 2011; Schreiber et al.
2008, 2010; Nebot G´omez-Mor´an et al. 2011). Nevertheless,
using the CSS photometry we were able to determine the
periods of 13 PCEBs, including the previously know PCEB
SDSS J1559+0356, for which we measure the same period as
Nebot G´omez-Mor´an et al. (2011). These non-eclipsing sys-
tems are detailed in Table 2.
7 DISCUSSION
7.1 The percentage of eclipsing systems
Among our 835 objects, roughly one third (∼280) will
be close PCEBs (Schreiber et al. 2010). Of these 29 are
eclipsing (∼10%). To test if this relatively high percent-
age is consistent with our current knowledge of PCEBs
we simulated the light curves of a population of PCEBs
with random inclinations and measured the percentage
that were eclipsing. We used the period distribution of
Nebot G´omez-Mor´an et al. (2011) and the white dwarf
mass distribution of Zorotovic et al. (2011a). For the
c
2012 RAS, MNRAS 000, 1–14

12 S. G. Parsons et al.
main-sequence star we adopted the mass distribution of
Zorotovic et al. (2011b) and included their correlation be-
tween the orbital period and main-sequence star mass. The
combination of these distributions provided the two stel-
lar masses and the orbital period. We then used the mass-
radius relationship for white dwarfs of Eggleton, quoted
in Verbunt & Rappaport (1988) and the mass-radius rela-
tionship for a 3 Gyr main-sequence star from Baraffe et al.
(1998). We accounted for Roche distortion and reject any
systems in which the secondary star exceeded its Roche lobe,
then tested if the system was eclipsing.
We found that 12% of the simulated PCEBs were eclips-
ing, consistent with the number found in the CSS photome-
try. We also simulated populations with different mass dis-
tributions and found that the number of eclipsing systems
is very insensitive to the mass of the white dwarf, due to
its small size. However, the number of eclipsing systems is
quite dependent upon the mass of the main-sequence star,
with a larger number of systems eclipsing with more mas-
sive (and hence larger) main-sequence stars. For example,
scaling the mass distribution of Zorotovic et al. (2011b) to
peak at 0.4M⊙increases the percentage of eclipsing systems
to 15%.
7.2 Completeness
We used our simulated PCEB light curves to test our eclipse
detection completeness. For each CSS light curve in our sam-
ple, we took the temporal sampling and created 100,000
synthetic PCEB light curves and tested how many of the
eclipsing systems were detected. We classified a system as
a confirmed eclipsing system if 2 or more separate eclipses
were detected. The results of this are shown in Figure 8.
We found that for light curves with 100 data points 90% of
eclipsing systems were detected, and with 200 data points
99% of eclipsing systems were detected. Only for light curves
that have less than 100 points is there a reasonable chance
of missing some eclipsing systems (these missed systems are
usually systems with periods in excess of 1 day). Since the
majority of our light curves comprise of more than 200 ob-
servations (76%) our overall detection percentage is ∼97%.
Therefore it is very unlikely that the well-sampled light
curves that show no eclipses are in fact eclipsers (and that
we have just missed the eclipses).
This simulation did not take into account our ability
to visually detect the eclipse in the CSS data. There could
potentially be some very shallow eclipsing systems which we
would not detect due to the signal-to-noise of the CSS data
(e.g. systems with very cool white dwarfs). Therefore, our
completeness calculations should be viewed as upper limits.
7.3 Long period systems
Our sample of new eclipsing systems contain two sys-
tems with periods in excess of 1.9 days. These systems
are particularly well suited for studying any long term
orbital period variations, which in some cases have been
attributed to the presence of planets in orbit around
the binary (Beuermann et al. 2010; Parsons et al. 2010b;
Kami´nski et al. 2007; Potter et al. 2011; Beuermann et al.
2011). However, the effects of quadrupole moment fluctua-
Figure 8. Probability of detecting eclipsing PCEBs based on the
number of CSS data points. Each point represents the temporal
sampling of a CSS light curve.
tions driven by stellar activity cycles, known as the Apple-
gate mechanism (Applegate 1992), can complicate the anal-
ysis of eclipse time variations by adding noise on potentially
the same scale as that caused by any third body.
The amount of energy required to drive period changes
via Applegate’s mechanism scales as (a/R)2, where ais
the orbital separation and Ris the radius of the main-
sequence star (Applegate 1992). Therefore, any period vari-
ations caused by Applegate’s mechanism in these longer pe-
riod systems will be negligible. This makes these systems
ideal to search for planets since Applegate’s mechanism can
be ruled out as the cause of any observed period variations,
leaving few alternative explanations other than the reflex
motion of the binary caused by an unseen body in orbit
around them.
7.4 Future Evolution
Detached PCEBs are the direct progenitors of cataclysmic
variables (CVs). As such they may provide crucial infor-
mation for our understanding of CV evolution because, in
contrast to CVs, it is possible to determine the evolution-
ary state of each PCEB using the temperature of the white
dwarf which provide us with a robust age estimate. Further-
more, for a given angular momentum loss prescription, one
can easily reconstruct the history of the systems and predict
their future.
Following Schreiber & G¨ansicke (2003) and
Zorotovic et al. (2011a) we search for PCEBs repre-
sentative for the progenitors of the current CV population
among the 11 new eclipsing PCEBs in our sample with
reasonable estimates of the stellar masses and WD temper-
ature (the only system we excluded is the DC white dwarf
system SDSS J1307+2156). We require the CV formation
time to be shorter than the Hubble time and the mass ratio
q=Msec/MWD to be <0.67 as otherwise the (second)
mass transfer will be dynamically unstable. With this
restriction only three PCEBs within our sample can be
considered representative for the progenitors of the current
CV population: SDSS J0821+4559, SDSS J0946+2030 and
c
2012 RAS, MNRAS 000, 1–14

Eclipsing PCEBs from the CRTS 13
SDSS J1223-0056. SDSS J0821+4559 is one of t he few
known progenitors of CVs that will clearly start mass
transfer above the orbital period gap (Porb = 3.94 hrs),
while SDSS J0946+2030 will b ecome a CV in the gap
(Porb = 2.32 hrs). The interpretation of SDSS J1223-0056
as being a CV progenitor is ambiguous, as this particular
system could also be a detached CV evolving through the
gap (Davis et al. 2008). This second possibility appears to
be very reasonable as the system is close to Roche-lobe
filling and will start mass transfer in only ∼2.7 Myrs, close
to the lower boundary of the period gap at Porb = 2.24 hrs.
The remaining systems in our sample have either very
long CV formation time scales (true for the remaining
systems containing carbon-oxygen core white dwarfs), or
the second mass transfer will be dynamically unstable
(true for the remaining systems with helium core white
dwarfs), or both which happens to be the case for the
long orbital period PCEB containing a helium core primary
(SDSS J0957+3001).
Finally, we note that the above calculations should be
taken as first order estimates mostly because of two sys-
tematic uncertainties. Firstly, the stellar masses were de-
rived from fitting the SDSS spectrum which we have shown
to be potentially quite rough (see Figure 7). Secondly, we
used an empirical spectral type-radius relation to estimate
the radii of the secondary stars, the intrinsic scatter around
any such relation is know to be significant (for details see
Rebassa-Mansergas et al. 2007).
8 CONCLUSIONS
We have analysed the Catalina Sky Survey light curves of all
white dwarf plus main-sequence binaries in the catalogue of
Rebassa-Mansergas et al. (2012a) with g < 19, in a search
for new eclipsing systems. We identify a total of 29 eclipsing
systems, 12 of which were previously unknown, and one can-
didate eclipsing system which needs better quality photom-
etry for confirmation. This increases the number of known
eclipsing post common envelope binaries to 49. We present
high-speed follow-up light curves of all our newly identified
systems, confirming both their eclipsing nature and their
ephemerides.
We find two new eclipsing systems with periods in ex-
cess of 1.9 days. These systems are ideal targets for detecting
planets in orbit around the binary via orbital period varia-
tions. This is because a common source of noise in the eclipse
time variations, known as Applegate’s effect, will have a re-
duced impact on the timing variations in these long period
systems.
Our newly discovered systems cover a large variety of
parameters. We find one system with a very hot white dwarf
(Teff = 80938 K), a system with a featureless DC type white
dwarf and a system with a very low mass white dwarf
(MWD ∼0.3M⊙). The main-sequence stars span spectral
types from M2-M6.
For three systems we were able to place constraints on
the mass of the white dwarf using measurements of the radial
velocity amplitude of the main-sequence star. In all cases
the mass is lower than implied from the deconvolution of
the SDSS spectrum. Therefore, the system parameters of all
our newly identified systems are subject to some uncertainty
until more detailed studies (e.g. Parsons et al. 2010a, 2012a;
Pyrzas et al. 2012) are carried out on them.
ACKNOWLEDGMENTS
SGP acknowledges support from the Joint Committee ESO-
Government of Chile. ULTRACAM, BTG, TRM, VSD and
SPL are supported by the Science and Technology Facili-
ties Council (STFC). ARM acknowledges financial support
from FONDECYT in the form of grant number 3110049.
MRS thanks FONDECYT (project 1100782) and the Mil-
lennium Science Initiative, Chilean Ministry of Economy,
Nucleus P10-022-F. The results presented in this paper are
based on observations collected at the European Southern
Observatory under programme IDs 086.D-0161 and 087.D-
0557. The CSS survey is funded by the National Aeronautics
and Space Administration under Grant No. NNG05GF22G
issued through the Science Mission Directorate Near-Earth
Objects Observations Program. The CRTS survey is sup-
ported by the U.S. National Science Foundation under
grants AST-0909182. The Liverpool Telescope is operated
on the island of La Palma by Liverpool John Moores Uni-
versity in the Spanish Observatorio del Roque de los Mucha-
chos of the Instituto de Astrofisica de Canarias with finan-
cial support from the UK Science and Technology Facilities
Council.
REFERENCES
Abazajian, K. N., et al., 2009, ApJS, 182, 543
Adelman-McCarthy, J. K., et al., 2008, ApJS, 175, 297
Almenara, J. M., et al., 2012, MNRAS, 420, 3017
Applegate, J. H., 1992, ApJ, 385, 621
Baraffe, I., Chabrier, G., Allard, F., Hauschildt, P. H., 1998,
A&A, 337, 403
Becker, A. C., Bochanski, J. J., Hawley, S. L., Ivezi´c, ˇ
Z.,
Kowalski, A. F., Sesar, B., West, A. A., 2011, ApJ, 731,
17
Bergeron, P., Wesemael, F., Beauchamp, A., 1995, PASP,
107, 1047
Bertin, E., Arnouts, S., 1996, A&AS, 117, 393
Beuermann, K., et al., 2010, A&A, 521, L60
Beuermann, K., et al., 2011, A&A, 526, A53
Davis, P. J., Kolb, U., Willems, B., G¨ansicke, B. T., 2008,
MNRAS, 389, 1563
Debes, J. H., 2006, ApJ, 652, 636
Debes, J. H., Hoard, D. W., Farihi, J., Wachter, S., Lei-
sawitz, D. T., Cohen, M., 2012, ApJ, 759, 37
Dhillon, V. S., et al., 2007, MNRAS, 378, 825
Drake, A. J., et al., 2009, ApJ, 696, 870
Drake, A. J., et al., 2010, ArXiv e-prints
Dufour, P., et al., 2007, ApJ, 663, 1291
Eisenstein, D. J., et al., 2006, ApJS, 167, 40
Fontaine, G., Brassard, P., Bergeron, P., 2001, PASP, 113,
409
Heller, R., Schwope, A. D., Østensen, R. H., 2011, in
Schmidtobreick, L., Schreiber, M. R., Tappert, C., eds.,
Evolution of Compact Binaries, vol. 447 of Astronomical
Society of the Pacific Conference Series, p. 177
Kami´nski, K. Z., et al., 2007, AJ, 134, 1206
c
2012 RAS, MNRAS 000, 1–14

14 S. G. Parsons et al.
Koester, D., 2010, Mem. Soc. Astron. Italiana, 81, 921
Kov´acs, G., Zucker, S., Mazeh, T., 2002, A&A, 391, 369
Law, N. M., Kraus, A. L., Street, R. R., Lister, T., Shporer,
A., Hillenbrand, L. A., Palomar Transient Factory Collab-
oration, 2011, in Johns-Krull, C., Browning, M. K., West,
A. A., eds., 16th Cambridge Workshop on Cool Stars, Stel-
lar Systems, and the Sun, vol. 448 of Astronomical Society
of the Pacific Conference Series, p. 1367
Law, N. M., et al., 2012, ApJ, 757, 133
Li, N., Thakar, A. R., 2008, Computing in Science and
Engineering, 10, 18
Liu, C., Li, L., Zhang, F., Zhang, Y., Jiang, D., Liu, J.,
2012, MNRAS, 424, 1841
Morgan, D. P., West, A. A., Garc´es, A., Catal´an, S., Dhital,
S., Fuchs, M., Silvestri, N. M., 2012, AJ, 144, 93
Nebot G´omez-Mor´an, A., et al., 2009, A&A, 495, 561
Nebot G´omez-Mor´an, A., et al., 2011, A&A, 536, A43
O’Donoghue, D., Koen, C., Kilkenny, D., Stobie, R. S.,
Koester, D., Bessell, M. S., Hambly, N., MacGillivray, H.,
2003, MNRAS, 345, 506
Parsons, S. G., Marsh, T. R., Copperwheat, C. M., Dhillon,
V. S., Littlefair, S. P., G¨ansicke, B. T., Hickman, R.,
2010a, MNRAS, 402, 2591
Parsons, S. G., et al., 2010b, MNRAS, 407, 2362
Parsons, S. G., Marsh, T. R., G¨ansicke, B. T., Drake, A. J.,
Koester, D., 2011a, ApJ, 735, L30
Parsons, S. G., Marsh, T. R., G¨ansicke, B. T., Tappert, C.,
2011b, MNRAS, 412, 2563
Parsons, S. G., et al., 2012a, MNRAS, 420, 3281
Parsons, S. G., et al., 2012b, MNRAS, 426, 1950
Parsons, S. G., et al., 2012c, MNRAS, 419, 304
Potter, S. B., et al., 2011, MNRAS, 416, 2202
Press, W. H., Rybicki, G. B., 1989, ApJ, 338, 277P
Pyrzas, S., et al., 2009, MNRAS, 394, 978
Pyrzas, S., et al., 2012, MNRAS, 419, 817
Raymond, S. N., et al., 2003, AJ, 125, 2621
Rebassa-Mansergas, A., G¨ansicke, B. T., Rodr´ıguez-Gil, P.,
Schreiber, M. R., Koester, D., 2007, MNRAS, 382, 1377
Rebassa-Mansergas, A., G¨ansicke, B. T., Schreiber, M. R.,
Koester, D., Rodr´ıguez-Gil, P., 2010, MNRAS, 402, 620
Rebassa-Mansergas, A., Nebot G´omez-Mor´an, A.,
Schreiber, M. R., Girven, J., G¨ansicke, B. T., 2011,
MNRAS, 413, 1121
Rebassa-Mansergas, A., Nebot G´omez-Mor´an, A.,
Schreiber, M. R., G¨ansicke, B. T., Schwope, A., Gallardo,
J., Koester, D., 2012a, MNRAS, 419, 806
Rebassa-Mansergas, A., et al., 2012b, MNRAS, 423, 320
Scargle, J. D., 1982, ApJ, 263, 835
Schreiber, M. R., G¨ansicke, B. T., 2003, A&A, 406, 305
Schreiber, M. R., G¨ansicke, B. T., Southworth, J., Schwope,
A. D., Koester, D., 2008, A&A, 484, 441
Schreiber, M. R., et al., 2010, A&A, 513, L7
Schwarzenberg-Czerny, A., 1989, MNRAS, 241, 153
Schwarzenberg-Czerny, A., 1996, ApJ, 460, L107
Silvestri, N. M., et al., 2006, AJ, 131, 1674
Steele, I. A., Bates, S. D., Gibson, N., Keenan, F.,
Meaburn, J., Mottram, C. J., Pollacco, D., Todd, I., 2008,
in Proc. SPIE, vol. 7014, p. 70146J
Steinfadt, J. D. R., Bildsten, L., Howell, S. B., 2008, ApJ,
677, L113
Tappert, C., G¨ansicke, B. T., Rebassa-Mansergas, A.,
Schmidtobreick, L., Schreiber, M. R., 2011, A&A, 531,
A113
Verbunt, F., Rappaport, S., 1988, ApJ, 332, 193
Willems, B., Kolb, U., 2004, A&A, 419, 1057
York, D. G., et al., 2000, AJ, 120, 1579
Zorotovic, M., Schreiber, M. R., G¨ansicke, B. T., 2011a,
A&A, 536, A42
Zorotovic, M., et al., 2011b, A&A, 536, L3
c
2012 RAS, MNRAS 000, 1–14