The Astronomical Journal, 127:2936–2947, 2004 May

# 2004. The American Astronomical Society. All rights reserved. Printed in U.S.A.

SPECTROSCOPIC AND PHOTOMETRIC ANALYSIS OF HS 1136+6646: A HOT YOUNG DAO+K7 V

POST–COMMON-ENVELOPE, PRE–CATACLYSMIC VARIABLE BINARY1,2
D. K. Sing,3,4 J. B. Holberg,3,4 M. R. Burleigh,5 S. A. Good,5 M. A. Barstow,5 T. D. Oswalt,6 S. B. Howell,7
C. S. Brinkworth,8 M. Rudkin,6 K. Johnston,6 and S. Rafferty6
Received 2003 November 26; accepted 2004 February 17

ABSTRACT
Extensive photometric and spectroscopic observations have been obtained for HS 1136+6646. The observa-
tions reveal a newly formed post–common-envelope binary system containing a hot
DAO.5 primary and a
highly irradiated secondary. HS 1136+6646 is the most extreme example yet of a class of short-period hot H-rich
white dwarfs with K–M companion systems such as V471 Tau and Feige 24. HS 1136+6646 is a double-line
spectroscopic binary showing emission lines of H i, He ii, C ii, Ca ii, and Mg ii, due in part to irradiation of the
K7 V secondary by the hot white dwarf. Echelle spectra reveal the hydrogen emission lines to be double-peaked
with widths of
200 km s1, raising the possibility that emission from an optically thin disk may also contribute.
The emission lines are observed to disappear near the inferior conjunction. An orbital period of 0:83607 
0:00003 days has been determined through the phasing of radial velocities, emission-line equivalent widths, and
photometric measurements spanning a range of 24 months. Radial velocity measurements yield an amplitude of
KWD ¼ 69  2 km s1 for the white dwarf and KK7V ¼ 115  1 km s1 for the secondary star. In addition to
orbital variations, photometric measurements have also revealed a low-amplitude modulation with a period of
113.13 minutes and a semiamplitude of 0.0093 mag. These short-period modulations are possibly associated with
the rotation of the white dwarf. From fits of the Balmer line profiles, the white dwarf is estimated to have an
effective temperature and gravity of
70,000 K and log g
7:75, respectively. However, this optically derived
temperature is difficult to reconcile with the far-UV spectrum of the Lyman line region. Far Ultraviolet Spec-
troscopic Explorer spectra show the presence of O vi absorption lines and a spectral energy distribution whose
slope persists nearly to the Lyman limit. The extremely high temperature of the white dwarf, from both optical
and UV measurements, indicates that the binary system is one of the earliest post–common-envelope objects
known, having an age around 7:7  105 yr. Although the spectrum of the secondary star is best represented by a
K7 V star, indications are that the star may be overly luminous for its mass.
Key words: binaries: spectroscopic — white dwarfs

1. INTRODUCTION spiral closer together, decreasing their orbital period. As the

stars spiral inward and mass is transferred to the less massive
Post–common-envelope binary systems evolve from rela-
star, eventually producing an extended atmosphere, a common
tively widely separated main-sequence stars. As both stars
envelope (CE) develops around both companions. The two
mature, the more massive component evolves faster, becom-
stars then transfer orbital angular momentum to the envelope
ing a red giant or supergiant and overflowing its Roche lobe.
while they spiral inward, helping to eject its outer layers. The
The onset of unstable mass transfer then causes the stars to
resulting system is a detached binary with the primary star
cooling toward the white dwarf stage.
CE evolution is a poorly understood process affecting the
1
Based on observations made with the NASA-CNES-CSA Far Ultraviolet evolution of many close binary systems including cataclysmic
Spectroscopic Explorer. FUSE is operated for NASA by Johns Hopkins
University under NASA contract NAS 5-32985.
variables (CVs), Type Ia supernovae, millisecond pulsars, and
2
Based on observations obtained at the SARA Observatory at Kitt Peak, X-ray binaries with neutron star or black hole components. It
which is owned and operated by the Southeastern Association for Research in is not known exactly how efficiently the red giant’s CE is
Astronomy.
3
ejected, how much mass is lost in the process, or how much
Lunar and Planetary Laboratory, University of Arizona, 1629 East Uni- the orbital separation will decrease. Post-CE binaries provide
versity Boulevard, Tucson, AZ 85721; [email protected], holberg@
argus.lpl .arizona.edu. important observational constraints on theoretical models of
4
Visiting Astronomer, Kitt Peak National Observatory, National Optical CE evolution as well as improving our understanding of the
Astronomy Observatory, which is operated by the Association of Universities evolution, space density, and period distribution of CVs. Only
for Research in Astronomy, Inc., under cooperative agreement with the about 25 post-CE binaries have measured orbital periods and
National Science Foundation.
5
Department of Physics and Astronomy, University of Leicester, Leicester
far fewer have accurate component masses. Thus, the study
LE1 7RH, UK; [email protected], [email protected], [email protected]. of every newly discovered system provides the potential for
6
Department of Physics and Space Sciences, Florida Institute of Technol- important new information.
ogy, 150 West University Boulevard, Melbourne, FL 32901; oswalt@luyten The DAO star HS 1136+6646 (hereafter HS 1136) was
.astro.fit.edu. first discovered by Heber, Dreizler, & Hagen (1996) during
7
WIYN Observatory and NOAO, P.O. Box 26732, 950 North Cherry
Avenue, Tucson, AZ 85726; [email protected]. follow-up spectroscopy of stellar sources from the Hamburg-
8
Department of Physics and Astronomy, University of Southampton, Schmidt objective-prism survey. The authors classified HS
Highfield, Southampton SO17 1BJ, UK; [email protected]. 1136 as a binary star consisting of a hot DAO (He ii k4686
2936
 SPECTROSCOPY AND PHOTOMETRY OF HS 1136+6646 2937
TABLE 1
HS 1136+6646 System Parameters

Parameter Value Note

R.A. (J2000) ....................................................................... 11h39m05:s 98

Decl. (J2000)....................................................................... +66 300 17B98
V magnitude ........................................................................ 13.63  0.03
J magnitude......................................................................... 12.314  0.021
H magnitude........................................................................ 11.695  0.024
K magnitude........................................................................ 11.543  0.021
VdK....................................................................................... 15.46  0.04 Inferior conjunction estimate
VWD ..................................................................................... 13.85  0.03
V-band photometric semiamplitude (orbital) (mag) ........... 0.11
V-band photometric semiamplitude (rotational) (mag) ...... 0.0093
White dwarf DAOTeff (K)................................................... 70,000 log g
7.75
White dwarf rotational (?) period (minutes) ...................... 113.13
K dwarf spectral type ......................................................... K4–7 V
Radial velocity ephemeris T0.............................................. HJD 2452360.566  0.83607E
KWD velocity (km s1) ....................................................... 69  2
KdK velocity (km s1)......................................................... 130  7
 system (km s1) .............................................................. 20.9  1
 (km s1)........................................................................... 12.1  5
MdK (M) ............................................................................ 0.34
MWD (M) ........................................................................... 0.63
a (R) .................................................................................. 3.7  0.1 Orbital separation
Estimated distance (pc) ...................................................... 388þ100
50

in absorption) with a late K companion. No emission lines parameters derived for HS 1136 and compare it with similar
were seen in the 1993 March 11, 1:14 UT (U. Heber 2004, post-CE systems.
private communication) spectrum, despite the fact that our
ephemeris (see Table 1) places the phase at 0:64  0:1, when 2. SPECTROSCOPIC OBSERVATIONS
the emission lines should be visible. A combination of sig-
2.1. Bok Spectroscopic Observations
nificantly lower resolution along with all of the strongest
emission lines observed between 4000 and 7000 8 falling in The radial velocity measurements were obtained on the
the cores of the white dwarf’s broad photospheric absorption Steward Observatory 2.4 m Bok telescope, located on Kitt
lines is a possible reason for the lack of visual emission Peak, during three observing runs between 2001 May and
lines. Moreover, the Heber spectra do not extend shortward 2003 March (see Table 2). The spectra were obtained using the
of 4000 8, where some of the strongest emission lines are Boller & Chivens Spectrograph at the Ritchey-Chrétien f /9
observed. focus. A 1200  800 15
m pixel CCD was used in combi-
The coordinates of HS 1136, derived from the STScI nation with a first-order 1200 line mm1 grating blazed at
Digital Sky Survey plates, are included in Table 1 along with 5436 8 to record the blue channel spectra covering a wave-
a finding chart in Figure 1. Estimates of proper motion de- length range of 3880–5040 8. With a 2B5 slit width, a typical
rived from the USNO-B1.0 catalog (Monet et al. 2003) in- spectral resolution of 3.5 8 was achieved at a reciprocal dis-
dicate a proper motion of 0B065 yr1 at a position angle persion of 0.99 8 pixel1 on the CCD. Typical exposure times
of 248 . of 1000–1200 s yielded characteristic signal-to-noise ratios
We first observed HS 1136 on 2001 March 15 with the (S/ N) of 50–60. Before each observation, the instrument was
Steward Observatory 2.4 m telescope as part of a spectro- rotated to align the slit perpendicular to the horizon, mini-
scopic survey of DAO and H-rich planetary nebula central mizing the effects of atmospheric dispersion.
stars. At that time its spectrum was dramatically different from Standard Image Reduction and Analysis Facility (IRAF)
that published by Heber et al. (1996). Strong emission lines routines were used to reduce the data. The wavelength
due to H i, He ii, Ca ii, and Mg ii were observed. Follow-up calibration was established with HeAr arc-lamp spectra, in-
observations on 2001 May 5–6, 2002 March 26–29, and terpolated between exposures taken before and after each
2003 March 21–23 showed a large-amplitude orbital mo- observation, to account for any small wavelength shifts that
tion between the emission lines and the He ii k4686 photo- may occur while the telescope tracks an object. The spectra
spheric absorption line, due to the white dwarf, as expected were flux calibrated using the Massey et al. (1988) spectro-
in a double-line spectroscopic binary for discussion of these photometric standard Feige 34 observed over a range of zenith
initial observations see Holberg et al. (2001) and Sing et al. angles covering the program stars’ air-mass range. HR 4550
(2002). and HD 103095 were used as radial velocity standards.
In this paper we present a complete description of the Radial velocities of the K7 V companion were extracted
existing ground-based data including the 2003 observations, from the unresolved emission lines in the optical data by the
refine the system parameters, and present two newly dis- following procedure using standard IRAF packages. A tem-
covered properties of the white dwarf and the secondary star. plate image was formed by first flattening all the HS 1136
In xx 2 and 3 we discuss the spectroscopic and photometric observations with a thirtieth-order spline to remove the con-
observations that are analyzed in x 4. In x 5 we interpret the tinuum, applying a heliocentric velocity correction for each
 Fig. 1.—Digital Sky Survey finding chart (POSS II, blue) for HS 1136+6646

TABLE 2
Log of Spectroscopic Observations

MJD Start MJD End

Date (JD  2,450,000) (JD  2,450,000) Telescopea Spectral Rangeb No. Observations

2001 Jan 29............ 1921.85486 1921.96217 FUSE FUV 2

2001 Mar 15 .......... 1984.91318 1984.92952 Bok Blue 2c
2001 May 05.......... 2035.65081 2035.94868 Bok Blue 18d
2001 May 06.......... 2036.63678 2036.84621 Bok Blue 18d
2001 Dec. 20 ......... 2264.85638 2264.93835 Bok Blue 5
2001 Dec 21 .......... 2265.99446 2266.00834 Bok Blue 1
2002 Jan 03............ 2277.82778 2277.84861 Mayall Red 1
2002 Jan 04............ 2278.83333 2279.01389 Mayall Red 4
2002 Jan 29............ 2303.64792 2303.73803 FUSE FUV 2
2002 Mar 26 .......... 2360.63557 2360.99945 Bok Blue 14c
2002 Mar 27 .......... 2361.65752 2361.98831 Bok Blue 14c
2002 Mar 28 .......... 2362.69017 2362.98201 Bok Blue 11c
2002 Mar 29 .......... 2363.62567 2363.96778 Bok Blue 18c
2002 Dec 18 .......... 2627.80817 2628.05427 Bok Red 13c
2003 Mar 21 .......... 2720.62975 2721.01248 Bok Blue 16c
2003 Mar 22 .......... 2721.61615 2722.00100 Bok Blue 18c
2003 Mar 23 .......... 2722.63475 2723.00905 Bok Blue 19c
2003 May 14.......... 2773.31983 2774.93389 FUSE FUV 18
a
(Bok) Steward Observatory 2.3 m telescope at Kitt Peak; (Mayall) NOAO 4 m telescope at Kitt Peak. (FUSE ) Far
Ultraviolet Spectroscopic Explorer.
b
Blue: 3880–5040 8; red: 5733–6896 8.
c
1200 s exposure times.
d
1000 s exposure times.
 SPECTROSCOPY AND PHOTOMETRY OF HS 1136+6646 2939

image, and then applying a preliminary velocity correction to

shift each image onto the spectra with the strongest emission
lines. All of the flattened spectra were then combined with the
routine SCOMBINE using a logarithmic wavelength scale.
The heliocentric velocity for the template was removed for use
in FXCOR (which automatically applies a heliocentric cor-
rection). The original flattened spectra were then run through
the FXCOR package, which outputs the observed velocities
for each spectra relative to the template. These observed ve-
locities were then used to refine the preliminary velocity
corrections and produce an improved template. This refining
procedure was repeated, typically twice, until output velocities
did not change significantly with each improved template.
Radial mean velocities from this iteration procedure yielded
uncertainties typically of 3 or 4 km s1 (see Fig. 2), for those
spectra in which the emission lines are strong. Fig. 3.—FUSE spectrum of HS 1136+6646 along with a TeA ¼ 110; 000 K,
log g ¼ 7:0 model. Note the strong continuum that continues up to the Lyman
2.2. Mayall Echelle Spectroscopic Observations limit without any roll-off.
High-dispersion measurements of the emission-line profiles
were obtained on the NOAO 4 m Mayall telescope located at
tive 3939 s exposures. Both of the observations were obtained
KPNO in 2002 February (see Table 2). The CCD echelle
through the large (LWRS, 3000  3000 ) aperture in time tagged
spectrograph instrument was used with a 58.5 groves mm1
(TTG) mode.
grating along with the 226-1 226 groves mm1 cross-dispersal
Data from the 2001 and 2002 observations have been re-
grating, having a first-order blaze at 6700 8. The T2KB
duced using the most recent available version of the FUSE
2048  2048 24
m CCD recorded the spectra from 4270–
data reduction software (CALFUSE ver. 2.4). Extracted spectra
7467 8 with a typical resolution of 0.16 8 at a reciprocal
from each of the four instrumental channels (separate optical
dispersion of 0.079 8 pixel1. Exposure times of 1800 s
paths consisting of different combinations of mirrors and
produced typical S/ N of 20. Standard IRAF routines were
gratings) and the two instrumental sides (independent micro-
used to reduce the data. ThAr lamps were used for wavelength
channel plate detectors) were extracted. The resulting eight
calibration and G191-B2B was used as a spectrophotometric
spectra were combined into a single continuous spectrum as
standard (Massey et al. 1988).
follows. First, the spectra were compared and scaled in flux to
2.3. FUSE match the observed flux in the LiF1 channel over the 1050–
1070 8 range, since this wavelength region is common to all
Far-UV (900–1180 8) spectra of HS 1136+6646 were four channels. Second, the spectra were resampled onto a
obtained on two separate occasions with the Far Ultraviolet common uniform wavelength scale with increments of 0.05 8,
Spectroscopic Explorer (FUSE ) satellite. The FUSE mission which approximate the instrumental spectra resolution. Mu-
and its spectrographs, which achieve spectra resolutions of tual alignment of the component spectra was achieved with
k=
k
20; 000, are described in Moos et al. (1997). The respect to interstellar absorptions features. Third, for each
observations of HS 1136+6646 (see Table 2) include a 2001 exposure, a single composite spectrum was produced by a
January 12 Cycle 2 FUSE observation (B053080100) under co-addition process that weighted the fluxes with respect to the
a program to investigate DAO and H-rich planetary nebulae S/ N in each spectral region. Photospheric features in the spectra
central stars (PI: M. A. Barstow) and an observation obtained corresponding to each exposure were measured and apparent
on 2002 January 29, as an observatory supplementary target stellar velocities relative to the interstellar medium (ISM) ve-
(S60106061000). The 2001 observations consisted of two locity were obtained. A final aggregate spectrum was also
3225 s exposures, separated in time by 100 minutes. The 2002 produced by Doppler shifting each of the exposures into the
observations were continuous but divided into two consecu- laboratory velocity frame using its measured relative stellar
velocity. This aggregate spectrum, which has an S= N
50
near 1000 8, is shown in Figure 3.
The most striking feature of the HS 1136 FUSE spectra is
the clear presence of the O vi kk1032, 1036 resonance doublet
(see Fig. 4). Barstow et al. (2001) determined that temper-
atures in excess of 56,000 K are required to produce observ-
able O vi lines in DA stars having nonstratified atmospheres.
An extensive set of recent (2003 May) FUSE observations
covering an entire orbit of HS 1136 also exist. A full analysis
of the 2003 May observations will be reported in a future
paper, although preliminary results are reported in this work.

3. PHOTOMETRIC OBSERVATIONS
3.1. SARA Photometry
Fig. 2.—Folded radial velocities for both components, phased with a period UBVRI observations were obtained with the SARA 0.9 m
of 0.83607 days. automated telescope at Kitt Peak in Arizona. Both differential
2940 SING ET AL. Vol. 127

pixels and an image scale of 0B33 pixel1, giving an approx-

imately 10 0  10 0 field of view. Observations were made for
up to
9 hr each night, through either a Harris B, V, or R filter.
The exposure time for each individual frame was 40 s. Com-
bined with the fast readout time of just over 1 minute, this
gave a duty cycle of
2 minutes. Full details are given in
Table 4. Standard calibrations such as sky flats and bias frames
were also obtained each night.
Data reduction was carried out using standard Starlink
software routines (e.g., FIGARO, Shortridge et al. 2001).
Aperture photometry was performed using the PHOTOM
package (Eaton, Draper, & Allen 2000). Differential magni-
tudes were then calculated with respect to a brighter com-
parison star in the CCD field of view.
Additional short photometric observations of HS 1136 were
also made from the JKT by three of us (M. R. B., S. A. G.,
Fig. 4.—Portion of the 2002 FUSE spectrum of HS 1136+6646 showing
the O vi stellar lines along with a C ii interstellar line.
and C. S. B.) during 2002 May 12–19 and May 26–June 2, to
help to unambiguously determine the binary orbital period.
Typically, a set of five to ten 40 s exposures was made through
time-series and absolute photometry was obtained on several a V filter each night. Data reduction was performed in a
nights between 2002 March and June and between 2003 manner identical to that described above.
March and May (see Table 3) by both on-site and remote-
access observers. Images were collected using an Apogee 3.3. UH 2.2 m Photometry
AP7p camera with a back-illuminated SITe SIA 502AB CCD photometry of HS 1136 was obtained on 2003 April
512  512 pixel CCD. The pixels are 24
m square, corre- UT at the University of Hawaii 88 inch (2.25 m) telescope
sponding to 0B73 at the telescope focal plane. Read noise and located on Mauna Kea. The observations were made with the
gain for the camera are about 12.2 e (rms) and 6.1 e OPTIC CCD camera using the newly implemented ‘‘video
ADU1, respectively. mode’’ for high-speed CCD photometry (Howell et al. 2003).
Duty cycle times were typically 25–35 s for the time-series Integrations of 30 s in a standard V filter were obtained from
work, including a readout time of
5 s. Sky flats and dark and 8:30 to 10:45 UT and three local stars of similar brightness
bias exposures were taken every night. All data were cali- were used as comparisons for the differential photometry.
brated and reduced using standard IRAF routines. Differential HS 1136 was placed on the standard system using observa-
magnitudes were obtained using a brighter comparison star in tions of the photometric standard fields near PG 1047+694
the CCD field of view. and PG 1528+062 (Landolt 1992). The CCD photometry
yielded a V-band magnitude of 13:63  0:03 at an orbital
3.2. JKT Photometry phase of about 0.25.
Intensive photometric monitoring of HS 1136 was also
3.4. 2MASS Photometry
carried out by two of us (M. R. B. and S. A. G.) over five
nights in 2002 February with the 1.0 m Jakobus Kapteyn Infrared photometry of HS 1136 was obtained from the Two
Telescope (JKT) on La Palma. The JAG-CCD system was Micron (2MASS) All-Sky Point Source Catalog. The resulting
used with a SITe2 detector. This device has 2048  2048 24
m magnitudes are J ¼ 12:314  0:021, H ¼ 11:695  0:024,

TABLE 3
Log of Photometric Observations

MJD Start MJD End

Date (JD  2,450,000) (JD  2,450,000) Telescopea Bands No. Frames

2002 Mar 27 ................................... 2360.698148 2361.037407 SARA V 723

2002 Apr 8...................................... 2372.643831 2373.025324 SARA V 571
2002 Apr 17.................................... 2381.877269 2382.016227 SARA V 299
2002 May 16–18 ............................ 2410.935729 2412.904575 JKT BVR 12 V, 2 B and R
2002 May 23................................... 2417.647118 2417.925058 SARA UBVRI 59 V
2002 May 28–31 ............................ 2422.870693 2426.912329 JKT V 42
2002 Jun 07 .................................... 2432.655880 2432.831655 SARA UBVRI 48 V
2003 Mar 21 ................................... 2720.624375 2721.019988 SARA V 799
2003 Mar 22 ................................... 2721.615058 2722.037674 SARA V 756
2003 Mar 23 ................................... 2722.871736 2722.799479 SARA V 752
2003 Apr 08.................................... 2737.84757 2737.93868 OPTIC/UH88 V 110
2003 May 25................................... 2784.670671 2784.846204 SARA V 492
2003 May 26................................... 2785.746435 2785.896933 SARA V 374
2003 May 27................................... 2786.665949 2786.870926 SARA V 490
a
(SARA) Southeastern Association for Research in Astronomy telescope at Kitt Peak; (JKT) Jacobus Kapteyn Telescope at La
Palma; (OPTIC/ UH88): University of Hawaii 88 inch (2.25 m) telescope at Mauna Kea.
No. 5, 2004 SPECTROSCOPY AND PHOTOMETRY OF HS 1136+6646 2941
TABLE 4
Summary of 2002 February JKT Observing Run

Exposure Time Duration

UT Date Filter (s) (hr) No. Frames Conditions

2002 Feb 34............ V 40 7.0 214 Photometric, seeing 1B3

2002 Feb 4/5........... R 40 4.2 129 Observed through cirrus, seeing 100
2002 Feb 5/6........... B 40 5.8 179 Cirrus, seeing variable, 1B2 –400
2002 Feb 6/7........... B 40 8.8 270 Not photometric, seeing 200
2002 Feb 7/8........... R 40 6.9 181 Not photometric, seeing 300 –500

and K ¼ 11:543  0:021, from which the corresponding DA+dM4 binary system with similar emission-line behavior,
colors can be derived, JH = +0.619  0.032 and HK = PG 1224+309, was studied by Orosz et al. (1999), who de-
+0.152  0.032. These 2MASS magnitudes, colors, and un- termined an inclination of 77  7 . The 0.25 day period PG
certainties are used to estimate the apparent spectral type of 1224+309 system is only a few degrees away from giving rise
the main-sequence component of the binary system. to an eclipse, but the emission lines are seen throughout the
entire orbital phase. Emission-line disappearance in the HS
4. ANALYSIS 1136 system therefore suggests a relatively high inclination.
4.1. Radial Velocities Measurements of the equivalent widths of the spectral lines
show a modulation at the same orbital period, but with a one-
The radial velocity curves for both components of the HS
quarter phase difference (see Fig. 5). This difference is con-
1136 system are shown in Figure 2. The apparent velocities of
sistent with the notion of the reflection effect giving rise to the
the K star are derived directly from the unresolved Bok CCD
narrow peaks of the emission lines. The preliminary mass ratio
spectra of the emission lines. It should be noted here that
q ¼ MK7V =MWD is estimated from the ratio of the K velocities
the Mayall echelle spectra reveal these lines to be surprisingly
giving a value of 0.60.
broad and to consist of asymmetric centrally reversed line
profiles (see x 4.6). Thus, the velocities in Figure 2 pertain 4.2. Photometry
to centroids of the emission lines. Since these emission lines
UBVRI photometric observations revealed a time-dependant
may contain nonphotocentric components (see x 4.6), they
sinusoidal variation of the light emission from HS 1136 due
may also possess possible nonorbital biases.
to the orbital motion in the binary system and the reflection
The estimate of the K velocity for the white dwarf, based
effect. From the reflection effect, the hemisphere of the sec-
solely on the weak He ii k4686 line, is KWD ¼ 61  10 km s1.
A significantly better estimate, however, is available from a ondary star facing the hot white dwarf is optically brighter
than the outward-facing hemisphere. As the two stars orbit one
preliminary analysis of a recent 2003 May FUSE observation
another, the observed UBVRI fluxes vary with the percentage
of HS 1136, which covered nearly a full orbital cycle. A
of the illuminated hemisphere seen.
complete analysis of the 2003 May FUSE observations will
In addition to the large amplitude photometric variations
be reported elsewhere; however, we will use the preliminary
associated with the orbit, low-amplitude photometric variations
FUSE value of KWD ¼ 69  2 km s1 in lieu of our ground-
of a much shorter period are also observed. These short-
based value.
period modulations have a period of 113.13 minutes, a half-
All of the emission lines from the narrow irradiation com-
amplitude of
0.0093 mag, and can be clearly seen in the
ponent are observed to disappear near the inferior conjunc-
orbital photometric light curve of Figure 6. They appear to
tion and remain undetectable for
15% of the total orbit. At
this orbital phase, the observed spectrum is effectively a be present in all observations and to be independent of orbital
phase. They are coherent from night to night and, as seen from
composite of the white dwarf and the dark side of the K star. A
preliminary modeling (Fig. 7), can almost certainly be phased

Fig. 5.—Mean spectroscopic emission line equivalent widths as a function Fig. 6.—Resampled V-band SARA photometry phased after subtracting a
of orbital phase (see discussion in text). Data for 2001 March, May, December sine curve fit to the small modulations. Phase dispersion minimization fits
and 2002 March are phased with a period of 0.83607 days. result in an orbital period of 0.83604 days.
2942 SING ET AL. Vol. 127

Fig. 7.—Low-amplitude 113 minute modulation seen in the residuals of Fig. 8.—Models of the HS 1136+6646 components and composite
the SARA photometric V-band light curve. The modulation is believed to be spectra overplotted with the Bok spectroscopic data (gray) near the superior
related to the rotation of a magnetic spot on the white dwarf. conjunction.

over much longer periods with sufficient data. In Figure 7 we of degenerate star pulsations and is too short for any stable
show a residual light curve, after removal of the orbital var- rotation of the K star.
iations, compared with a best-fitting sine wave having a period
of 113.13 minutes and a semiamplitude of 0.0093 mag. There
4.3. Determination of Orbital Ephemeris
are indications that the wave form is doubly sinusoidal, or that
the amplitude may vary, but night-to-night matching of the The orbital period was determined using the 2-based phase
photometry makes departures from a pure sine wave difficult dispersion minimization (PDM) technique as described in
to characterize from the present data. Handler (1998) searched Stellingwerf (1978). The V-band photometric data was sup-
for short-period pulsations in HS 1136 using high-speed rel- plemented with the emission-line equivalent width data (see
ative photometry. No evidence of short-period variations were Table 5) to achieve better accuracy. The equivalent widths of
observed but the short duration of the observation (76 minutes) the emission lines were measured from the five brightest H i
together with other variations seen and the low-frequency and He i lines (see Fig. 8). An average strength for these
filtering of the data would have made it difficult to detect five lines computed for each spectrum was used in the PDM
any variations of the type we report. The low-amplitude, 113 technique, yielding an orbital period of 0:83607  0:00003
minute modulations are presumed to originate with the hot days using all of the spectroscopic ground-based data (see
white dwarf and to be associated with either a rotational Table 2; Fig. 5). Since the 113 minute modulations are not
photometric modulation due to a magnetic spot or pole on necessarily related to the orbital motion, phasing the V-band
the white dwarf’s surface. Alternative explanations such as data alone led to systematic errors in determining the period,
pulsational instability in the white dwarf or rotational modu- since the PDM couples to the small variations. In order to
lations of the K star are less likely. The observed 113 minute avoid this systematic error, a preliminary fit of the small
period of the modulations is well outside the known range modulations with a cosine function (see Fig. 7) was made and

TABLE 5
Bok Equivalent Width Measurements of Emission Lines at Orbital Phase = 0.5

Wavelength Continuum Flux Equivalent Width GFWHMa

Species (8) (ergs cm2 s1 81) (ergs cm2 s1 81) (8) (8)

H i ....................... 3970.072c 2.362  1014 9.958  1014 4.22 3.526

H i ....................... 4101.740 1.968  1014 7.439  1014 3.78 3.287
H i ....................... 4340.470 1.665  1014 7.385  1014 4.44 3.124
H i ....................... 4861.330 1.119  1014 6.008  1014 5.37 3.600
He i and H i ........ 3888.650 2.615  1014 9.188  1014 3.51 3.623
He i...................... 4026.191 2.241  1014 1.738  1014 0.77 3.019
He i...................... 4120.820 2.190  1014 6.686  1015 0.31 3.264
He i...................... 4143.760 2.200  1014 5.022  1015 0.23 2.378
He i...................... 4387.929 1.562  1014 7.074  1015 0.45 2.658
He i...................... 4471.479 1.463  1014 1.358  1014 0.93 2.626
He i...................... 4713.146 1.274  1014 6.121  1015 0.48 2.690
He i...................... 4921.931 1.180  1014 1.265  1014 1.07 3.173
Ca ii ..................... 3933.660 2.478  1014 2.685  1014 1.08 3.016
C ii....................... 4267.260 1.702  1014 6.312  1015 0.37 2.583
Mg ii.................... 4481.245b 1.464  1014 8.717  1015 0.60 2.258
a
FWHM based on a Gaussian fit.
b
Includes the Ca ii k 3968.470 line, explaining a relatively larger equivalent width and GFWHM as compared to the other
hydrogen lines.
c
Blend of 4481.160 and 4481.330.
No. 5, 2004 SPECTROSCOPY AND PHOTOMETRY OF HS 1136+6646 2943

subtracted from the V-band photometric data. This procedure

was performed by first fitting the orbital variations with a
cosine function of the form
 
t  T0
Vmag (t) ¼ V0 cos 2 þ hV i; ð1Þ
P

where Vmag is the observed V magnitude, V0 is the amplitude

of the light curve, and hV i is the mean magnitude. The fit was
then subtracted from the photometric data before determining
the period. Although some small modulations are still present,
a period of 0:83604  0:00004 days is obtained from the PDM
fit. This closely matches the period obtained with spectro-
scopic equivalent width data (see Fig. 6). The radial velocity
data agrees well with this period, as shown in Fig. 2, which
overplots the radial velocity data as sinusoids of the form
  Fig. 9.—Bok 2.4 m emission-line spectra at the superior conjunction (top)
t  T0
VWD (t) ¼  0 þ KWD sin 2 ; ð2Þ compared with the spectrum at the inferior conjunction (middle) and the
P extracted white dwarf spectrum (bottom). The best-fitting model white dwarf
spectrum of TeA ¼ 70; 000 K and log g ¼ 7:75 is overplotted on top of the
extracted white dwarf spectrum.
for the white dwarf, and
 
t  T0 spectrum. The new white dwarf spectrum was then refitted and
VK7V (t) ¼  þ KK7V sin 2 ; ð3Þ
P a new set of white dwarf parameters determined. A final set
of parameters corresponding to TeA ¼ 70; 000  5000 K and
for the secondary star, with the orbital parameters as listed in log g ¼ 7:75  0:25 was achieved after several iterations. This
Table 1, where KWD and KK7V are the K velocities, P is the final model is shown in Fig. 9. Note that the fitting procedure
orbital period, and T0 and t are, respectively, the epoch and employed normalized individual Balmer profiles, thereby
time;  and  0 are, respectively, the system gamma velocity minimizing any bias associated with the sloping spectral en-
and the algebraic sum of the system velocity plus gravitational ergy distribution. Best results were obtained from the H
redshift of the white dwarf. and H lines, suggesting that not all of the K7 V flux was
properly subtracted from the composite spectra.
4.4. Properties of the White Dwarf
In principal, it is also possible to use the H i Lyman series
In principal, it is possible to independently estimate the lines to estimate the Teff and log g of HS 1136. Barstow et al.
effective temperature and gravity of a hot white dwarf such (2003) have extensively investigated the relationship between
as HS 1136 from a detailed analysis of both the Balmer and Balmer line estimates and those derived from the FUSE Lyman
Lyman line profiles. However, in the case of HS 1136, the lines spectra for a set of 16 DA white dwarfs. In general, they
Balmer lines are contaminated both by the continuum from find very good agreement for stars with temperatures below
the secondary star and the emission lines. We have attempted
50,000 K and increasingly wide divergence for stars at higher
to minimize this contamination by the following procedure. temperatures. The sense of the difference in temperatures
Optical spectra obtained near the orbital inferior conjunc- found by these two methods is that Lyman line temperatures
tion (see Fig. 9) with the Bok telescope were used to extract a began to exceed the Balmer line temperatures at
50,000 K,
characteristic white dwarf spectrum. In order to remove the increasing to a discrepancy of 6% to 10% higher by 60,000 K.
K star from the composite flux as accurately as possible, The corresponding correlation for gravities shows a good
a template spectrum of the K7 V star BD +46 1635 was agreement between Lyman and Balmer results. There is a weak
obtained with the Bok telescope on 2003 March 23. This indication of higher Lyman line gravities for the very hottest
ensured that the template K star spectrum had the same res- stars, but this is based only on two stars with large error bars.
olution, dispersion, and wavelength range as the HS 1136 At present, we can suggest no explanation for the differences
spectra at inferior conjunction. Shortward of 4000 8, the flux between the two methods nor is there a clear indication of
of HS 1136 is dominated by the hot white dwarf, making it which method, Lyman or Balmer, provides the best estimate
possible to accurately flux calibrate an assumed preliminary of effective temperature for the hottest H-rich white dwarfs.
model of the white dwarf (TeA ¼ 50; 000 K, log g ¼ 8:0). The Keeping the results of Barstow et al. (2003) in mind, we
template K7 V star was then flux-calibrated so that the com- fitted the Lyman lines for HS 1136 in a 2 minimization
posite model flux corresponded to the observed levels of the technique similar to that of Barstow et al. (2003). We find
HS 1136 spectra. A preliminary model white dwarf spectrum TeA ¼ 110; 000 K and log g = 7.0–7.5 using the 2001 and 2002
was then obtained by subtracting the spectrum of the correctly FUSE observations. However, this fit is very unsatisfactory
flux-calibrated K7 V star from the HS 1136 spectra at inferior (see Fig. 3) and is also incompatible with the strong contin-
conjunction. This white dwarf spectrum was then fitted with a uum that continues up to the Lyman limit with no discernible
grid LTE, pure-hydrogen model stellar atmospheres produced roll-off at the convergence of the Lyman series. Models indi-
by TLUSTY (Hubeny & Lanz 1995). cate that a much higher effective temperature is necessary
A final set of white dwarf parameters was obtained by to match the spectral energy distribution shortward of 950 Å
the following iterative procedure. A synthetic white dwarf and the Lyman limit. Clearly, the Lyman line temperature and
spectrum was generated from fitted Teff and log g values, the far-UV (FUV) energy distribution are incompatible with
which was then used to rescale the subtracted K7 V template our results for the Balmer line fits. Nevertheless, both Balmer
2944 SING ET AL. Vol. 127

and Lyman analysis place HS 1136 among the hottest stars

considered by Barstow et al. (2003) and among the hottest
H-rich white dwarfs known. Although we are presently unable
to reconcile the Balmer and Lyman line results, we will adopt
the Balmer line parameters TeA ¼ 70; 000 K and log g ¼ 7:75
as a good match to the white dwarf at optical wavelengths and
use these results in the discussion that follows.
An estimate of the mass and thermal age of the white dwarf
can be obtained from its Balmer line temperature and gravity.
Using the thermally evolved models of Wood (1995), we
find a mass of MWD ¼ 0:63  0:05 M and a thermal age of
(8:2  0:7)  105 yr for the white dwarf. From synthetic pho-
tometry of Bergeron et al. (1995) we find a corresponding Mv
of 8.5 for the white dwarf.
In addition to the temperature and gravity, the optical
spectrum of the white dwarf can be used to estimate the
photospheric He abundance of the white dwarf. The measured
equivalent width of the He ii k4686 line is 32 m8, which for
a 70,000 K, log g ¼ 7:75 white dwarf corresponds to a He
abundance of He=H
2  103 .

4.5. Properties of the K Star

The approximate spectral type of the main-sequence com-
panion can be best estimated from the observed 2MASS JHK
colors in x 3.4. These infrared colors can be compared with the
intrinsic main-sequence colors of K dwarfs given in Bessell &
Brett (1988). After transforming the Bessell & Brett colors to
the equivalent 2MASS colors using the color transformations
of Carpenter (2001), the observed 2MASS colors correspond
the range of spectral types of K4–M0 V with an apparent
visual magnitude of 14.34–15.27. We will adopt K7þ4 2 V. This
estimate is quite consistent with the spectroscopic determi-
nation of Heber et al. (1996), based on the atomic and mo-
lecular features seen in their red spectrum of the star as well as
Fig. 10.—Mayall 4 m H emission-line profiles from the inferior con-
our modeling of the optical continuum near superior and in- junction (bottom) to the superior conjunction (top). The corresponding phases
ferior conjunction, which assumed a K7 V secondary (Pickles from bottom to top are 0.05, 0.25, 0.28, 0.39, and 0.45.
1998) and a 70,000 K, log g ¼ 7:75 white dwarf (see Fig. 8).
It should be noted that the 2MASS data were obtained at
between the two peaks, which tends to diminish near the
a phase of 0:22  0:05, which corresponds to the mean of
superior conjunction.
the V-band photometric orbital variations. Second, this also
One possible explanation for the double-peaked emission-
assumes that the main-sequence star can be characterized by
line structures observed in HS 1136 is that they are due to an
its JHK colors and is not over-luminous as is the case with
accretion disk. The aspects of the lines that lend themselves
V 471 Tau (O’Brien, Bond, & Sion 2001).
to this idea are the asymmetric nature of the peaks and the

300 km s1 widths. Under this assumption, we would assign
4.6. Emission-Line Profiles the higher, narrow peak to irradiation from the K7 V star with
Observations of HS 1136 were obtained on 2003 January the underlying (now symmetric) double-peaked structure due
03–04 with the 4 m Mayall telescope yielding five spectra to a disk of captured (wind) material orbiting the white dwarf.
that cover an orbital phase range from 0.05 to 0.45. These Taking the orbital parameters we have derived in this paper,
echelle spectra, which fully resolve the emission lines, re- the line widths observed place the disk far from the white
vealed lines that are unexpectedly broad and asymmetrically dwarf near a radius of 0.7RL , where RL is the Roche lobe
double-peaked (see Fig. 10). Neither of these aspects are radius, and its structure would have to be very optically thin or
obviously compatible with emission arising solely from the completely evaporated (a Strömgren sphere) in its interior; a
irradiated stellar photosphere of the K7 V star. The three very ringlike disk. The argument against an accretion disk is
hydrogen emission lines, H, H, and H, have FWHM of that we note the large decrease in the line strengths and their
5.34, 3.15, and 2.84 8, respectively. Line broadening due to symmetric nature at phase 0.05, requiring a nearly complete
differential orbital motion during the 1800 s exposure and eclipse of the disk by the K7 V star.
thermal Doppler broadening at 7000 K explain the width of The double-peaked emission line structure may also be at-
the He i, He ii, Ca ii, and Mg ii lines, but not those of the tributed to Stark broadening analogous to BE UMa, in which
broader H i lines. Unlike the other emission lines, the hy- case the lines are formed deep in the atmosphere of the K7 V
drogen lines clearly show the double-peaked feature. Figure 10 star (Ferguson & James 1994). Emission lines in BE UMa from
shows the development of the H emission line as a function elements heavier than hydrogen, such as C ii, C iii, and N iii, are
of orbital phase over almost half a cycle. As the strength formed higher up in the atmosphere and thus are not as broad as
of the emission grows, there is a pronounced asymmetry the Balmer lines. Such would be the case with HS 1136 were
No. 5, 2004 SPECTROSCOPY AND PHOTOMETRY OF HS 1136+6646 2945
TABLE 6
Similar Post-CE Systems

Period a Teff WD Mass Post-CE Age

System (days) (R) Components (K) (M) (yr)

BE UMa.................................. 2.291 7.5 sdO/DAO+K5 V 105,000 0.7 104

HS 1136+6646 ........................ 0.836 3.7 DAO+K7 V 70,000 0.64 105
V 471 Tau ............................... 0.521 3.3 DA+K2 V 34,000 0.84 107

the He i, He ii, Ca ii, and Mg ii lines are not as broad nor 127 km s1, resulting in a q value of 0.54, which corresponds
observed to be double-peaked. Unlike HS 1136, however, the to a white dwarf mass of 1.12 M. Equation (4) then gives an
double-peaked emission-line profiles of BE UMa are report- orbital separation of 4.5 R, while equation (5) gives an in-
edly symmetric throughout the orbital cycle. The asymmetry of clination of 46  3 . An inclination of 46 is too low for the
the Balmer lines in HS 1136 raises doubts about the validity of emission lines to be associated with the reflection effect, and
the Stark broadening mechanism and leaves a satisfactory below a lower limit of 54 set with the known orbital period,
explanation of the nature of the emission lines elusive. KK7V > 115 km s1, KWD ¼ 69 km s1, and a white dwarf
mass 0.63 M, which is more reasonable. At such an angle,
4.7. Binary Components the lines would remain visible throughout the orbital cycle. In
With a known preliminary mass ratio of q ¼ 0:6 (x 4.1), we addition, the high white dwarf mass, 1.12 M, would imply a
employ two scenarios in order to estimate a consistent set of gravity of log g > 8:5, which is not consistent with either our
binary system parameters. In this manner we can investigate Balmer or Lyman spectral analysis.
which assumptions made in estimating the parameters are
consistent with other observed features, such as emission-line 5. DISCUSSION
disappearance. In the first scenario, we assume a white dwarf 5.1. Parameters of the Binary System
mass based on Balmer line fits for Teff , log g, and a K star that
has a radius twice that of a K7 V zero-age main-sequence The parameters found in the first scenario in x 4.7 require
(ZAMS) star. This scenario is based on the possibility that HS the K7 V star to be out of thermal equilibrium and the mass of
1136 resembles other very young CE systems, such as BE the white dwarf to be around 0.6 M. A secondary star mass of
UMa, in which the main-sequence companion has been heated 0.33 M is not characteristic of a K7 V star, rather it indicates
by the accretion of matter during the CE phase (see Schreiber the secondary star may actually be an early-M star. This lower
& Gänsicke 2003). A white dwarf mass of MWD ¼ 0:63 M mass is clearly at odds with the observed luminosity as well as
and age of 7:7  105 yr is adopted from the synthetic pho- the spectral typing of the secondary as a K7 V star and is
tometry of Bergeron et al. (1995), based on a TeA ¼ 70; 000 K further evidence that the secondary may be overly luminous.
and log g ¼ 7:75 DA white dwarf. Since the radial velocity There are several reasons for believing this first scenario is
profile of the K7 V star is derived from observations of the more likely than the second. First, there are four binary sys-
emission lines, the observed value of q ¼ 0:6 represents an tems closely related in temperature to HS 1136 that have
upper limit. In order to shift the apparent radial velocity from measured orbital parameters: UU Sge, V477 Lyr, KV Vel, and
the center of light to the center of mass, a correction of +16 km BE UMa (Schreiber & Gänsicke 2003). All of these binary
s1 to the radial velocity is needed (see Orosz et al. 1999 for systems have a hot (>60,000 K) primary component indicating
velocity correction discussion and technique). This estimate the systems are very young. All have a secondary star with
assumes a radius of 0.79 R and mass 0.38 M (Pickles 1998) radii inflated by a factor of
2. The thermal timescale of the
for the secondary star, which is uniformly illuminated over envelope of the K7 V star is much higher than the cooling time
one hemisphere, along with a q value of 0.6. With an observed of the white dwarf (
7:7  105 yr) following the accretion of
value of KK7V ¼ 115 km s1, the ‘‘K-corrected’’ value is then matter during the CE phase for the HS 1136 system. This
131 km s1. The new q value is then 0.53 corresponding to supports the notion that the K star is out of thermal equilib-
a secondary star mass of MdK ¼ 0:33 M. Kepler’s third law rium. The second scenario, which results in a higher white
then gives the orbital separation a, dwarf mass, would only strengthen the above argument since
the cooling time would be shorter. The narrow-peak emission-
a3 ¼ G(MWD þ MdK )(P=2)2 ; ð4Þ line disappearance near orbital phase zero is another reason
to favor the first scenario. The inclination derived in the sec-
and the inclination follows from ond scenario is nearly 20 lower than in the first. If the bi-
nary system were in such a low-inclination orbit, a nontrivial
PKdK (KWD þ KdK )2 fraction of the secondary star’s irradiated hemisphere would
sin3 i ¼ ; ð5Þ always be visible, and the emission lines would always be
2GMWD
present.
where G is the gravitational constant and the other parameters
5.2. Age of the Binary System
are listed in Table 1. We find a ¼ 3:7 R and i ¼ 64  6 .
Eclipses have not been observed, placing an upper limit on The cooling time of the hot white dwarf is on the order of
the inclination of i < 80 , based on the parameters listed in 7:7  105 yr, which indicates that the HS 1136 system has just
Table 1. This is consistent with the value given by equation (5). recently emerged from the post–CE phase (see Table 6). This
In the second scenario it is assumed that the secondary system can test the origins of CE evolution and help determine
component is a K7 V ZAMS star with a mass of 0.609 M and such parameters as the efficiency of the CE process and
a radius of 0.681 R. The K corrected value for KK7V becomes whether post-CE secondaries satisfy a normal main-sequence
2946 SING ET AL. Vol. 127

mass-radius relation (Orosz et al. 1999). HS 1136 is similar to 5.4. The Future Evolution of HS 1136+6646
the well-known systems Feige 24 and V 471 Tau, which are We have modeled the orbital evolution of HS 1136 using
hot DA stars with dM2 and dK2 companions, respectively.
the parameters listed in Table 1. Specifically, we assume a
Feige 24 shows a well-studied reflection effect while the DA
white dwarf mass of 0.7 M and a respective K7 V star mass
in V471 Tau is too cool relative to its main-sequence com-
of 0.61 M and radius of 0.56 R. The initial period of the
panion to produce an observable reflection effect.
system was estimated to be 0.83 days. This period is some-
5.3. Similar Post–CE, Pre-CV Stars what longer then the present orbital period but less than a day,
which would be typical of the progenitor binary. We have
It is of interest to compare HS 1136 with two other well-
followed the orbital evolution of the system using the secular
observed systems that, in many ways, appear to bracket it.
evolution model code described in Howell et al. (2001). We
These two systems are BE UMa (Ferguson et al. 1999 and
find that the present system will continue to evolve toward
references therein) and V 471 Tau (O’Brien et al. 2001 and shorter orbital periods for approximately 1  109 yr. When the
references therein). All of these systems appear to have a post-
system reaches an orbital period of approximately 4.9 hr, the
CE origin and to contain a hot degenerate star with a main-
Roche lobe of the K7 V companion will contact the white
sequence K star. In Table 6 we list the system parameters and dwarf surface and a high rate of mass transfer will commence.
the characteristics of the degenerate stars. Since both BE UMa
HS 1136 should then become a standard CV system. It is
and V 471 Tau are eclipsing systems, the stellar parameters of
interesting to speculate that if the 113 minute modulation is in
the two components are relatively well determined.
fact due to a magnetic spot on the surface of the white dwarf
Like HS 1136, BE UMa exhibits a very strong reflection
and if the magnetic field is strong enough (1–10 MG), then
effect, both in the continuum and emission lines. The latter
HS 1136 may be an incipient intermediate polar or DQ Her
exhibits a host of first- and second-ionization stages of CNO
star. In DQ Her stars, the magnetic field of the white dwarf is
species as well as Si and Al. The key aspect of BE UMa is the
too weak to create the well-defined accretion columns char-
high luminosity of the degenerate star, which has led Ferguson
acteristic of the AM Her type systems. Such systems often
et al. (1999) to classify it as a borderline sdO/ DAO star. In
form a partial accretion disk. By the time mass transfer begins,
contrast, although the effective temperatures of the degenerate the white dwarf will have cooled to between 6000 and 7000 K
stars in HS 1136 and BE UMa are similar, the DAO star in
and will no longer dominate the optical flux from the system.
HS 1136 is not nearly as luminous or dominant compared to
However, for a typical 5 hr orbital period cataclysmic variable,
its main-sequence companion. Ferguson et al. (1999) estimate
the white dwarf temperature is 30,000 K or more (Sion 1999).
the gravity of the degenerate star in BE UMa to be 6.5, while
Apparently HS 1136 will reheat quickly once mass accretion
the composite spectrum of HS 1136 can be used to place a
begins.
lower limit of 7.0 on the gravity of the DAO.
V 471 Tau, on the other hand, is almost completely domi-
6. CONCLUSIONS
nated by the K star and shows no detectable reflection effect.
However, low-amplitude variations, similar to the ones de- HS 1136 offers a unique opportunity to study a recently
tected in HS 1136 (x 4.2), have been reported in the soft X-ray evolved common-envelope (CE) system that is relatively
in V471 Tau (Jensen et al. 1986) and in the EUV from the hot bright and nearby. We have presented new spectroscopic and
white dwarf GD 394 (Dupuis et al. 2000). In both cases these photometric observations that characterize the binary param-
modulations are believed to be associated with the rotation eters of HS 1136. These observations also show HS 1136 to be
of the white dwarf. In particular, the magnetic ‘‘spots’’ are the second youngest post-CE system known having emerged
thought to accrete high-z matter producing a nonuniform dis- from its CE phase only
105 yr ago. From analysis of the
tribution of heavy elements on the stellar surface. Ellipsoidal binary parameters, we find that HS 1136 has a K7V star that is
variations cannot contribute to the main orbital photometric out of thermal equilibrium. The photometric measurements
modulations seen in HS 1136, as the secondary does not fill its reveal low-amplitude 113 minute modulations that are pre-
Roche lobe. sumably due to the rotation of the white dwarf. Synthetic
The K5 V secondary star of BE UMa is out of thermal white dwarf model fits to the Balmer line profiles of the white
equilibrium, resulting in a radius nearly twice that expected. dwarf in HS 1136 give a TeA ¼ 70; 000 K and log g ¼ 7:75.
This large radius is a result of the thermal timescale of the This temperature and gravity remain difficult to reconcile with
star’s envelope being longer than the time since emerging from the UV spectrum Lyman line results, TeA ¼ 110; 000 K and
the CE phase (Ferguson et al. 1999). The thermal timescale log g ¼ 7:00 7:5.
for the K7 V star in HS 1136 is
106 yr, which is more than In 2003 May, HS 1136 was observed by the FUSE space-
the estimated age of 7:7  105 yr, leaving open the possi- craft. These UV observations provided nearly complete orbital
bility that it, too, is not in thermal equilibrium and has an coverage of the binary system and should allow a better de-
extended radius. termination of the white dwarf velocity curve. A preliminary
In Table 6 we also list the estimated post-CE ages for all value is cited in this work, as well as a very high S/ N FUV
three systems based on white dwarf cooling ages. In the case spectrum of the white dwarf. Cycle 12 observations with the
of BE UMa, the high luminosity of the degenerate star com- Space Telescope Imaging Spectrograph (STIS) are also planed
plicates the interpretation of its evolutionary status. It is with the Hubble Space Telescope. The echelle UV spectra
clearly a very young system since Liebert et al. (1995) have expected from these observations should allow placement of
detected a 30 nebula surrounding the star. For reasonable the FUSE velocity for the white dwarf on an absolute scale
estimates of the expansion rate of the nebula, they find good and also provide a gravitational redshift for the white dwarf.
agreement with the thermal age of the degenerate. V 471 Tau Such observations will allow detailed modeling of the white
is clearly the oldest system of the three, but the thermal age dwarf photosphere and a determination of its heavy-element
of the DA is in apparent conflict with its being a member of abundance. Studies of the low amplitude 113 minute modu-
the Hyades cluster (O’Brien et al. 2001). lations will also continue from the ground, using multiple
No. 5, 2004 SPECTROSCOPY AND PHOTOMETRY OF HS 1136+6646 2947

sights to obtain longer unbroken photometric series to better under NASA contract NAS 5-26555. Support for MAST for
determine the waveform and hopefully to identify its source. non-HST data is provided by the NASA Office of Space
A better understanding of the orbital-related changes in the Science via grant NAG 5-7584 and by other grants and con-
components of the emission lines will result from echelle tracts. This research has made use of the USNOFS Image
spectra that cover a full orbit. and Catalog Archive operated by the United States Naval
Observatory, Flagstaff Station (http://www.nofs.navy.mil/data/
fchpix). This paper has also made use of data products from
D. K. S. and J. B. H. wish to acknowledge support from the Two Micron All Sky Survey, which is a joint project of
NASA grant NAG 5-10700 and NAG 5-13213. M. A. B. and the University of Massachusetts and the Infrared Process-
M. R. B. were supported by the Particle and Astronomy Re- ing and Analysis Center=California Institute of Technology,
search Council, UK. The FUSE data presented in this paper funded by the National Aeronautics and Space Administration
were obtained from the Multimission Archive at the Space and the National Science Foundation. T. D. O. would like
Telescope Science Institute (MAST). STScI is operated by the to acknowledge partial support for this project from NASA
Association of Universities for Research in Astronomy, Inc., (NAGW 5-9408) and NSF (AST 00-97616, AST 02-06115).

REFERENCES
Barstow, M. A., Burleigh, M. R., Bannister, N. P., Holberg, J. B., Hubeny, I., Hubeny, I., & Lanz, T. 1995, ApJ, 439, 875
Bruhweiler, F. C., & Napiwotzki, R. 2001, in ASP Conf. Ser., Vol. 226, Jensen, K. A., Swank, J. H., Petre, R., Guinan, E. F., Sion, E. M., & Shipman,
12th Euopean Conference on White Dwarfs, ed. J. L. Provencal et al.(San H. L. 1986, ApJ, 309, L27
Francisco: ASP), 128 Landolt, A. U. 1992, AJ, 104, 340
Barstow, M. A., Good, S. A., Burleigh, M. R., Hubeny, I., Holberg, J. B., & Liebert, J., Tweedy, R. W., Napiwotzki, R., & Fulbright, M. S. 1995, ApJ,
Levan, A. J. 2003, MNRAS, 344, 562B 441, 424
Bergeron, P., Wesemael, F., & Beauchamp, A. 1995, PASP, 107, 1047 Massey, P., Strobel, K., Barnes, J. V., & Anderson, E. 1988, ApJ, 328, 315
Bessell, M. S., & Brett, J. M. 1988, PASP, 100, 1134 Monet, D. G. et al. 2003, AJ, 125, 984
Carpenter, J. M. 2001, AJ, 121, 2851 Moos, H. W. et al. 2000, ApJ, 538, L1
Dupuis, J., Chayer, P., Vennes, S., Christian, D. J., & Kruk, J. W. 2000, ApJ, O’Brien, M. S., Bond, H. E., & Sion, E. M. 2001, ApJ, 563, 971
537, 977 Orosz, J. A., Wade, R. A., Harlow, J. B., Thorstensen, J. R., Taylor, C. J., &
Eaton, N., Draper, P. W., & Allan, A. 2000, PHOTOM: A Photometry Package Eracleous, M. 1999, AJ, 117, 1598
(Starlink User Note 45.12; Chilton: Rutherford Appleton Lab.) Pickles, A. J. 1998, PASP, 110, 863
Ferguson, D. H., & James, T. A. 1994, ApJS, 94, 723 Schreiber, M. R. & Gänsicke, B. T. 2003, A&A, 406, 305
Ferguson, D. H., Liebert, J., Haas, S., Napiwotzki, R., & James, G. A. 1999, Shortridge K., et al. 2001, Starlink User Note 86.19
ApJ, 528, 866 Sing, D. K., Holberg, J. B., Barstow, M. A., Burleigh, M., Good, S., Oswalt, T.,
Handler, G. 1998, A&A, 339, 170 Howell, S. & Brinkworth, C. 2002, in Proc. 13th European White Dwarf
Heber, U., Dreizler, S., & Hagen, H. J. 1996, A&A, 311, L17 Workshop, ed. R. Silvotti & D. de Martino (Dordrecht: Kluwer), 349
Holberg, J. B., Sing, D. K., Barstow, M. A., & Good, S. 2001, Astron. Ges. Sion, E. S. 1999, PASP, 111, 532
Abstr. Ser., 18, No. P106 Stellingwerf, R. F. 1978, ApJ, 224, 953
Howell, S. B., Everett, M. E., Tonry, J. L., Pickles, A., & Dain, C. 2003, PASP, Wood, M. A. 1995 in 9th European Conference on White Dwarfs, ed.
115, 1340 Koester, D. & Werner, K. (Berlin: Springer), 41
Howell, S. B., Nelson, L. A., & Rappaport, S. 2001, ApJ, 550, 897H

Note added in proof.—C. S. Brinkworth, M. R. Burleigh, G. A. Wynn, & T. R. Marsh (MNRAS, 348, L33 [2004]) have reported a
low-amplitude photometric variability for the magnetic white dwarf GD 356, which closely resembles HS 1136+6646 in both
period and amplitude. They attribute the photometric variability to a dark spot on the stellar surface.