The Pre-cataclysmic Binary HS 1136+6646 May Have a Companion

Abstract

Because of the similarity of the primary star of HS 1136+6646 to the planetary nebula central star BE Ursae Majoris, we did wide field imaging of the former with an H-alpha filter. No nebulosity was detected. On the other hand, the point spread function of the star appeared extended. A partially-resolved red component is detected in the image with the five-band Sloan Digital Sky Survey. Most importantly, a companion is easily resolved in the HST acquisition image for the published STIS observation. A companion to the pre-cataclysmic binary is present at a separation of 1.349'' at position angle 54.4 deg. Evidence indicates that it is likely of K spectral type. We cannot demonstrate conclusively that this component has common proper motion with the close binary. However, the similar apparent z magnitudes and spectral types of HS 1136+6646B and the resolved component make it likely that we have in reality a hierarchial triple system. In any case, the presence of this component needs to be taken into account in future ground-based studies. Comment: 4 pages, 5 figures, requires emulateapj.sty. Accepted for publication in PASP. Figure 1 reduced in resolution

Full text

Page 1

arXiv:astro-ph/0609653v1 25 Sep 2006
Accepted for publication in the Publications of the Astronomical Society of the Pacific
Preprint typeset using LATEX style emulateapj v. 6/22/04
THE PRE-CATACLYSMIC BINARY HS 1136+6646 MAY HAVE A COMPANION
James Liebert1, Kurtis A. Williams1,2J. B. Holberg3, and D. K. Sing3
Accepted for publication in the Publications of the Astronomical Society of the Pacific
ABSTRACT
Because of the similarity of the primary star of HS 1136+6646to the planetary nebula central star BE
Ursae Majoris, we did wide field imaging of the former with an Hα filter. No nebulosity was detected.
On the other hand, the point spread function of the star appeared extended. A partially-resolved red
component is detected in the image with the five-band Sloan Digital Sky Survey. Most importantly,
a companion is easily resolved in the HST acquisition image for the published STIS observation. A
companion to the pre-cataclysmic binary is present at a separation of 1.349” at position angle 54.4o.
Evidence indicates that it is likely of K spectral type. We cannot demonstrate conclusively that
this component has common proper motion with the close binary. However, the similar apparent z
magnitudes and spectral types of HS 1136+6646B and the resolved component make it likely that we
have in reality a hierarchial triple system. In any case, the presence of this component needs to be
taken into account in future ground-based studies.
Subject headings: white dwarfs – binaries(including multiple): close – binaries: spectroscopic – stars:
individual(HS 1136+6646)
1. INTRODUCTION
HS 1136+6646(hereafter HS 1136) is a hot young DAO
white dwarf plus K7V secondary star, in a close detached
binary. The discoveryfrom the Hamburg Schmidt Survey
was reported in Heber, Dreizler, & Hagen (1996). Sing
et al. (2004) determined parameters for the individual
components and found an orbital period of 0.83607 days.
Radial velocity measurements of both components were
used to determine masses of 0.85⊙and 0.37M⊙for the
primary and secondary, respectively (Sing 2005). The
white dwarf parameters – Teff ∼70,000 K and log g
∼7.75 – could only be coarsely estimated, due to the
dependence of the parameters determined from fitting
the Balmer lines on the heavy element abundances.
Moreover, the optically-derived temperature was diffi-
cult to reconcile with the far-UV spectrum of the Lyman
line region. The Far Ultraviolet Spectroscopic Explorer
spectrum shows the presence of O VI absorption lines and
a spectral energy distribution whose slope persists nearly
to the Lyman limit, suggesting that the Balmer line Teff
estimate is too low. Fits to the Lyman lines by Good
et al. (2004) indicate a far higher value of 120,000 K –
among the highest determinations for a DAO white dwarf
– with log g of 6.5. These parameters suggest that the
system could have left the common envelope even more
recently than the Sing et al. (2004) estimate of 7.7×105
years. Indications are that the secondary star may be
overluminous compared to a main sequence star of its
mass, still bloated or out of thermal equilibrium from
the common envelope phase.
The secondary is irradiated by strong UV radiation
causing strong variations in the strengths of the emis-
sion lines vs. orbital phase. BE Ursae Majoris (Fer-
1Department of Astronomy and Steward Observatory, Univer-
sity of Arizona, Tucson, AZ 85721, liebert@as.arizona.edu
2Current address: Department of Astronomy, University of
Texas, Austin TX 78712, kurtis@astro.as.utexas.edu
3
LunarandPlanetary Laboratory,
zona, Tucson AZ85721,
singd@vega.lpl.arizona.edu
University ofAri-
holberg@vega.lpl.arizona.edu,
guson et al. 1999, and references therein) is a similar
close DAO+K binary also showing a strong orbitally-
modulated, “reflection effect.” Again, only coarse esti-
mates of Teffand log g were given by the above authors,
suggesting that the primary star could be similar to that
of HS 1136.
An additional property of the BE UMa system is the
presence of a large planetary nebula of low surface bright-
ness (Liebert et al.1995). As part of a program by the
second author of that paper to look for ancient planetary
nebulae (Tweedy & Kwitter 1994), a 50 arc minute field
around BE UMa was imaged in Hα 6563˚ A, [O III] 5007˚ A,
and [N II] 6584˚ A. The [N II] observation yielded a non-
detection, the [O III] a marginal one, but the Hα image
revealed the very faint nebula, centered on the binary
system, of about a 5 arc minute diameter.
This led us to try a similar observation of HS 1136, to
see if the systems were similar in yet another property. In
Section 2 we report that no nebula was detected. Instead,
we made a different discovery, discussed in Section 3.
2. THE SEARCH FOR A PLANETARY NEBULA
90Prime is a wide field imager, mounted at the prime
focus of the Steward Observatory 2.3-m Bok reflector
(Williams et al. 2004)4. The detector is a mosaic of four
thinned Lockheed 4096×4096 pixel CCDs. The f/2.98
system provides a plate scale of 0.45′′/pixel. We centered
the binary star on Chip-1, the cosmetically best chip hav-
ing no bad columns. The resulting image through an Hα
filter is shown in Figure 1. Scattered light in the upper
left of the figure is from a 6th magnitude star, 3 Draconis,
25’ northeast of HS 1136. This complicates the potential
detection of faint nebulosity. In any case, no evidence
of nebulosity is seen. Because of the complication, we
did not try to estimate a sensitivity limit above which a
nebula is not present.
One feature with this camera at prime focus is a con-
4We note that Ed Olszewski is the principal investigator of the
grant enabling construction

Page 2

2 Liebert et al.
Fig. 1.— The Hα image centered on HS 1136 obtained with
the 90Prime camera. Three 15-minute exposures were stacked to
make this combined image. North is up, with east to the left. The
image displayed is 15 arc minutes on a side.
Fig. 2.— The contour plots of counts, with each contour a
successive factor of two below the peak for (a, left) HS 1136 and (b,
right) one of the several reference stars in the field also measured.
North is up and east is to the left. That HS 1136 is extended
relative to a seeing-dominated point source is apparent.
tinual need to monitor the instrument focus.
checking for this after the observations with isophotal
contour plots, it became apparent to us that the tar-
get star appeared elongated at a position angle of 60 (or
240) degrees from due north (Figure 2a), while several
stars of similar magnitude in the field appeared almostly
perfectly circular (one of them is shown in Figure 2b).
This strongly suggested to us that HS 1136 was not a
point source, but either would turn out to be a barely-
resolved source or a partially-resolved blend of two or
more sources. In Figure 3 is shown the radial profiles
for HS 1136 (solid points) vs. those for four comparison
stars. That HS 1136 is more extended than the others is
indisputable. The result was not due to the instrument
being out of focus. However, the seeing at this time from
extracting the point spread functions was ∼1.6” – see
Figure 3.
Had we discovered a very compact nebula in Hα
around this unusual binary, instead of an old nebula?
A logical test was to try a very short exposure in broad-
band V , a wavelength range in which only weak nebular
emission could be anticipated. We did this observation
(not shown here), but the HS 1136 PSF again showed
extension, the comparison stars did not. Some other ex-
planation besides nebular emission had to be sought.
When
Fig. 3.— The azimuthally-summed radial distribution of counts
for HS 1136 (filled circles) and four reference stars (dark curves).
The full width at half maximum of the latter indicates a seeing of
about 1.8 arc second. The extension of the HS 1136 counts past
1” is consistent with the presence of the close companion.
Fig.4.— The multi-color image of HS 1136 from the Sloan
Digital Sky Survey. A faint red companion is marginally resolved
from the blue-colored close binary.
3. THE RESULT OF HIGHER RESOLUTION IMAGING: A
COMPANION
The field is covered by the Sloan Digital Sky Survey
(Gunn et al. 1998, York et al. 2000, Stoughton et al
2002). If one downloads the color image from the DR5
Finding Chart Tool (Figure 4), the presence of a fainter
companion to HS 1136 at the same orientation as indi-
cated by Fig. 2a is apparent. This image shows that the
companion is much redder than the optical light of the
close binary, the latter dominated by the very hot pri-
mary. As we will show later, visual inspection of other
SDSS stellar sources suggests strongly that the compan-
ion is of K or M spectral type.
The acquisition image taken for the previously-

Page 3

Optical Companion of HS 1136+66463
PM
1.3486"
N
Fig. 5.— The acquisition image with the STIS spectrograph on
the Hubble Space Telescope. The red companion is fully-resolved at
the separation marked in the figure and at the orientation indicated
in the previous figures. The STIS wavelength band is very broad
but centered near V . The companion is 1.9 magnitudes fainter than
the HS 1136 binary. The proper motion measurement is indicated
by the arrow, and discussed in the text
published spectrum with the Space Telescope Imaging
Spectrograph (STIS) instrument on the Hubble Space
Telescope is shown in Figure 5. The USNO-B1 (Monet
et al. 2003) position of HS 1136 is α2000= 11:39:05.570,
+66:30:17.75. The wide companion is separated by µ
= 1.349” at θ = 54.391o, or at position 11:39:05.945,
+66:30:18.45. Since the STIS imaging used the broad-
band clear imaging 50CCD mode, the band pass encom-
passes a wavelength range from 2000˚ A to past 1µ, though
the sensitivity peaks around the B,V and R bands. The
relatively brightness or count ratio of HS 1136AB to the
resolved companion is 5.5:1 (nearly 2 magnitudes).
The measured proper motion from the USNO-B Cat-
alog is 3.23” in 50 years at a position angle of 248.2 de-
grees, the wide companion to the close binary is not re-
solved. If we assume that the resolved companion is an
unrelated field star with negligible proper motion, the
two stars are presently moving apart. If we consider the
POSS I fields (epoch 1954.1), the red and blue images
are symmetrical and coincident with no evidence of an
offset, even though the proper motion difference between
HS 1136 and a stationary companion would have resulted
in a 1.9” separation on the opposite side of HS 1136.
However, the effective image size at the Palomar Ochsen
Schmidt scale of 67”/mm is several times this separa-
tion. Thus, the companion has not been demonstrated
to share a common proper motion.
Estimating the 3.84 yearsof proper motion between the
SDSS and STIS ACQ images for the close binary and the
resolved component, shows each has moved around 0.25”
in roughly similar directions (close binary: µ = 0.23” /
3.84 yr, θ = 239◦; resolved component: µ = 0.26” /
3.84 yr, θ = 288◦). Note that what we call the resolved
component is close, but not technically resolved by the
Rayleigh criteria in the z band image, complicating its
proper motion calculation.
We have retrieved the SDSS u g r i z magnitudes cor-
responding to the image of HS1136+6646 using the Ex-
plore tool in Sky Server. The resolved component is vir-
tually undetected in u, but is nearly resolved in the z
band image. Sky server has separated the two compo-
nents into the close binary SDSS J113905.78+663017.8
and the resolved component SDSS J113905.98+663018.3.
Point spread function (PSF) photometry from the SDSS
website lists u = 13.298± 0.014, r = 13.914± 0.03, i =
14.084± 0.08, and z = 14.159± 0.034 for the close binary
and u = 18 ±6, r = 16.12±0.31, i = 14.457±0.048, and z
= 14.052±0.028 for the resolved component. The g band
flux is saturated and therefore not used. The z band flux
from the close binary and resolved component can be es-
timated by convolving a synthetic spectrum of the white
dwarf (Sing 2005) with the z band filter and applying the
photometric corrections of Holberg & Bergeron (2006).
We find that the white dwarf “A” component has an es-
timated z magnitude of 15.2 while the binary companion
“B” is 14.7.
Unfortunately, the separate PSF magnitudes other
than z are unreliable. This is illustrated by the colors of
the resolved component ( r−i = 1.659 and i−z = 0.432)
which do not give main sequence colors. The z magni-
tudes of the two main sequence stars are observed to be
similar and should therefore both contribute to the over-
all flux of HS1136+6646 in the spectroscopic and photo-
metric observations reported by Sing et al. (2004). These
authors showed that the secondary star in HS1136+6646
(which we now know is a composite of two stars) was con-
sistent with a K7 V from the optical down to the infrared.
The two main sequence stars, with similar z magnitudes,
would therefore have to be of similar K spectral type
to give these consistent colors across the wide spectral
range.
Analyzing the photometry of the 50CCD image can
provide further hints as to the spectral type of the
HS 1136C component. We convolved the clear band pass
of the 50CCD image with the observed spectral energy
distributions of (1) the white dwarf primary, (2) a K7 V
close binary with z=14.7, and (3) a resolved G-M com-
ponent with z=14.025. A synthetic brightness ratio of
HS 1136AB to HS1136C was then calculated for different
resolved companion spectral types and compared with
the observed 5.5:1 ratio. We find a K4 V or K5 V is
most consistent for HS 1136C giving ratios of 5.3:1 and
5.8:1 respectively, while G or M types are ruled out with
a G5 V giving a ratio of 3.3:1 while a M3 V gives 12.9:1.
Even a K3 V with a ratio of 4.7 or a K7 V with 7.0 are
unlikely matches. With the main sequence stars having
similar magnitudes, type, and proper motion, it would
seem likely that HS1136+6646 is a triple star system,
although further studies will be needed for a definitive
proof.
Note that Tokovinin et al.
tive optics observations that 13 of 62 solar-type spec-
troscopic binaries have tertiary companions. Similar ob-
servations of HS 1136AB and its neighbor with an AO
system should be sufficient to establish more accurately
the color and likely spectral type of the latter, and es-
tablish whether its magnitude is consistent with it being
a companion at similar distance.
(2006) found via adap-
4. CONCLUSIONS
We failed to detect any diffuse nebula surrounding
the young post-CE system HS 1136. We did, however,
uncover evidence of an apparent third component con-
tributing to the optical light. It is prsently unclear if

Page 4

4 Liebert et al.
this component, easily resolved in the HST acquisition
image, is an unrelated field star or a common proper
motion companion making a hierarchial triple system.
Evidence indicates that it is likely of K spectral type.
Regardless, ground-based spectroscopic and photomet-
ric studies of this systems need to take into account the
presence of the additional star.
We thank Richard Cool and Ed Olszewski for providing
the Hα flat fields for the 90Prime images. JBH wishes
to acknowledge support from Space Telescope Science In-
stitute grant GO09762. JL and KAW acknowledge sup-
port from the National Science Foundation through grant
AST03-07321.
REFERENCES
Ferguson, D.H., Liebert, J., Haas S., Napiwotzki, R., & James,
T.A. 1999, ApJ, 518, 866
Good, S.A., Barstow, M.A., Holberg, J.B., Sing, D.K., Burleigh,
M.R., & Dobbie, P.D. 2004, MNRAS, 365, 1031
Gunn, J.E. et al. 1998, AJ,116, 3040
Heber, U., Dreizler, S., & Hagen, H.J. 1996, A&A, 311, L17
Holberg, J.B., & Bergeron, P. 2006, AJ, in press
Liebert, J., Tweedy, R.W., Napiwotzki, R., & Fulbright, M.S. 1995,
ApJ, 441, 424
Monet, D.G. et al. 2003, AJ, 125, 984
Sing, D.K. 2005 PhD dissertation, University of Arizona
Sing, D.K., Holberg, J.B., Burleigh, M.R., Good, S.A., Barstow,
M.A., Oswalt, T.D., Howell, S.B., Brinkworth, C.S., Rudkin,
M., Johnston, K., & Rafferty, S. 2004, AJ, 127, 2936
Stoughton, C. et al. 2002, AJ, 123, 485
Tokovinin, A., Thomas, S., Sterzik, M., & Udry, S. 2006, A&A,
450, 681
Tweedy, R.W., & Kwitter, K. 1994, AJ, 108, 188
Williams, G.G., Olszewski, E., Lesser, M.P, & Burge, J.H. 2004,
SPIE, 5492, 787
York, D.G. et al. 2000, AJ, 120, 1579