A T8.5 Brown Dwarf Member of the Xi Ursae Majoris System

0.68 MB·

by  Edward L. Wright

A T8.5 Brown Dwarf Member of the Xi Ursae Majoris System
Edward L. Wright1, M. F. Skrutskie2, J. Davy Kirkpatrick3, Christopher R. Gelino3, Roger
L. Griffith3, Kenneth A. Marsh4, Tom Jarrett3, M. J. Nelson2, H. J. Borish2, Gregory
3
Mace1, Amanda K. Mainzer5, Peter R. Eisenhardt5, Ian S. McLean1, John J. Tobin6,
1
0 Michael C. Cushing7
2
[email protected]

n
a
J

1
3 ABSTRACT

] The Wide-field Infrared Survey Explorer has revealed a T8.5 brown dwarf (WISE
R
J111838.70+312537.9) that exhibits common proper motion with a solar-neighborhood (8 pc)
S quadruple star system - Xi Ursae Majoris. The angular separation is 8.5′, and the projected
.
h physical separation is 4000 AU. The sub-solar metallicity and low chromospheric activity of ξ
≈
p UMa A argue that the system has an age of at least 2 Gyr. The infraredluminosity and color of
-
the brown dwarf suggests the mass of this companion ranges between 14 and 38 M for system
o J
r ages of 2 and 8 Gyr respectively.
t
s
Subject headings: brown dwarfs – infrared:stars – solar neighborhood – stars:late-type – stars:low-mass
a
[
1. Introduction Cardoso et al. 2009). In the absence of a dynami-
2
callymeasuredmass,spectralmodelingcansignif-
v Theeffectivetemperatureandthusspectrumof
icantlyconstrainabrowndwarf’smassifthereex-
4
a brown dwarf evolves with time as it cools as a
6 ists sufficientrestrictiononthe object’s age. Con-
degenerate object (Kumar 1962). For an isolated
7 straints on age and metallicity are available if the
5 brown dwarf, determination of mass and age are brown dwarf is a member of a multiple star sys-
. intertwined, such that a broad locus of mass and
3 tem. In this case, the properties of the primary,
0 age will be consistent with a single measured ef- particularly chromospheric activity and kinemat-
2 fective temperature(Burrows et al.2003). Inrare
ics, provide an indication of age.
1 instances browndwarfsmay reside inclose binary
: The Wide-field Infrared Survey Explorer mis-
v systems,resolvingthisambiguitywithadirectdy-
sion (WISE, Wright et al. 2010) has been a pro-
i namical mass estimation (Konopacky et al. 2010;
X ductive engine for the discovery of the coolest
ar 1UCLA Astronomy, PO Box 951547, Los Angeles CA b(4r.o6wnµmdw)afirflste.rsThaereWoIpStEimWall1y(3t.4unµemd)taondselWec2t
90095-1547
2Department of Astronomy, University of Virginia, the coolest candidates, specifically those with
Charlottesville,VA,22904 spectra significantly shaped by methane absorp-
3InfraredProcessingandAnalysisCenter,CaliforniaIn- tion at low effective temperature (Mainzer et al.
stituteofTechnology, PasadenaCA91125 2005). To date spectroscopic follow-up of WISE-
4School of Physics & Astronomy, Cardiff University, selected sources has revealed more than 100 ul-
CardiffCF243AA,UK
tracool brown dwarfs (Kirkpatrick et al. 2011)
5Jet Propulsion Laboratory, California Institute of
including several exceptionally cool Y-dwarfs
Technology, 4800 Oak Grove Dr., Pasadena, CA, 91109-
8001,USA (Cushing et al.2011;Kirkpatrick et al.2012;Tinney et al.
6National Radio Astronomy Observatory, Char- 2012)demonstratingthatWISEcolorsprovidefor
lottesville,VA22903 reliable photometric selection of ultra-cool brown
7Department of Physics and Astronomy, MS 111, Uni- dwarf candidates.
versityofToledo,2801W.BancroftSt.,ToledoOH43606-
WISE J111838.70+312537.9, hereafter WISE
3328
1
Table 1: Photometric Observations of WISE
1118+31.
Filter Vega Magnitude Instrument
Y(MKO) 19.18 0.12 FanCam
±
J(MKO) 17.792 0.053 WHIRC
±
J 18.22 0.16 Bigelow
±
J 18.37 0.08 FanCam
±
H(MKO) 18.146 0.060 WHIRC
±
H 18.13 0.23 Bigelow
±
Fig. 1.—WISEAll-skyImageAtlascutoutsshow- Ks(MKO) 18.746 0.150 WHIRC
ing a 10′ FoV centered halfway between WISE W1 16.160±0.071 WISE
±
1118+31 (lower left) and ξ UMa (upper right). ch1 15.603 0.026 IRAC
±
TheleftpanelshowsW1(3.4µm),whilethe right ch2 13.368 0.018 IRAC
±
panel shows W2 (4.6 µm). The lines point to W2 13.308 0.032 WISE
±
WISE 1118+31. The faint source near the lines W3 12.359 0.314 WISE
±
in the left panelis the nearby2MASS star,#1 on W4 >8.821 WISE
Figure 2. The odd shape above WISE 1118+31
in the W2 image is a ghost image of ξ UMa.
The right-most column shows 1′ postage stamps
of WISE images at 3.4, 4.6 & 12 µm.
Note.—Allmagnitudes areVegamagnitudes anduse
the2MASSJHKs filtersexcept asnoted. TheY-band
1118+31, easily meets the WISE brown dwarf calibrationusestheHamuyetal.(2006)transformation
color selection criterion (Kirkpatrick et al. 2011) ofY −Ks vs.J−Ks.
with a W1-W2 color of 2.85 compared to a selec-
tion threshold of 2.0. Few confusing objects meet nal at the position of the W1/2 source in W3 (12
theseselectionrestrictions. Inadditionthissource µm) and only an upper limit in W4 (22 µm). Of
lies 8.5′ from one of the nearer stars to the Sun, the twenty apparitions, eighteen were sufficiently
ξ UMa, prompting an investigation into a possi- separated from a detector edge to permit source
ble system membership. This paper reports the extraction. In all eighteen cases the source was
spectral characterization of WISE 1118+31 and detected in W2, producing a combined SNR=34
the analysis of a series of astrometric observa- detection with W2=13.31. Because exceptionally
tions spanning 26 months aimed at determining coolbrowndwarfs areconsiderablyfainter in W1,
whether this source exhibits common proper mo- whichwasoptimizedtoproduceasubstantialflux
tionwithξ UMa. Theseobservationsdemonstrate difference between W1 and W2, WISE 1118+31
that WISE 1118+31 is a newly found member of is detected in W1 in only 12 of the 18 opportu-
thisalreadyremarkablemultiplestarsystem. The nities with a combined SNR=15 and W1=16.16,
characteristics of the primary system provide an yielding a color of W1-W2=2.85. Figure 1 shows
indication of the metallicity and age of the newly portions of the WISE image atlas covering both
discoveredultra-coolbrowndwarfconstrainingthe WISE 1118+31and ξ UMa.
mass of this object.
2.2. Follow-up Imaging
2. Observations
Multiple epochs of near-infrared imaging pro-
2.1. WISE vide the astrometric data for WISE 1118+31
needed to confirm common proper motion with
WISE imaged the region of the sky containing
ξ UMa. Photometric information from these im-
WISE 1118+31 on twenty occasions between 21
ages in summarized in Table 1 while the astro-
May 2010 04:20 UT and 25 May 2010 09:39 UT.
metric data are listed in Table 2. No correc-
The WISE All-Sky Catalog reports this source as
tionsfornon-linearitywereappliedtotheground-
well detected in W1 and W2, with marginal sig-
based photometry because the sky is brighter
2
several 2MASS stars for photometric and astro-
metric reference. A median sky frame was sub-
tractedfromeachindividualexposurepriortoflat
fielding with the median background level sub-
sequently restored to the image. Y&J aperture
photometrywascomputedusinganaperturewith
a radius of 3 pixels. The zero points for the J-
band were computed using stars in the FOV with
measured 2MASS magnitudes since the J-band
filterinFanCamarebasedonthe 2MASSsystem.
In order to derive the Y-band zero point, we first
computed the Y-band magnitudes of stars in the
FOVusingtheir2MASSJ andK magnitudesand
s
the transformation given by Hamuy et al. (2006).
The final uncertainty in the magnitudes include
thephotonnoisefromtheskyandsource,theread
noise,andtheuncertaintyinthezeropointdueto
the computed Y-band magnitudes of the calibra-
Fig. 2.— Y-band image of the WISE 1118+31
tors. The resulting magnitudes and uncertainties
field obtained at Fan Mountain Observatory 15
are given in Table 1.
March 2012 UT. North is up and East is to the
left. Perpendicular lines mark the position of
2.2.2. Mount Bigelow/2MASS
WISE 1118+31. Star 1 nearby is the J=15.94
star 2MASS J11183876+3125441, while star 2 is The former 2MASS camera on the 1.54m
2MASS J11185086+3126520 with J=10.32. The Kuiper Telescope on Mt. Bigelow, Arizona, has
faintest detected objects have Y 20. The faint three 256 256-pixel NICMOS3 arrays simultane-
≈ ×
diffraction spike entering from the upper right ously observing in 2MASS J, H, and K filters
s
(northwest) is from ξ UMa 8.5′ away. (Milligan et al. 1996). The plate scale for all
three arrays is 1.65′′/pixel, resulting in a 7′ field
of view. Exposures of 10s duration, 216 in all,
(Sa´nchez et al. 2008) than both WISE 1118+31
of WISE 1118+31 were obtained on 20 May 2011
and the faint 2MASS objects used as calibrators.
using 6repeatsofa3 3boxdither pattern, with
The absolute calibration uncertainties of 3% for ×
four consecutive images taken at each of the nine
Spitzer (Reach et al. 2005) and 2.4, 2.8, 4.5 &
dither positions. The data were reduced using
5.7% for W1..4 (WISE Explanatory Supplement
customIDLroutinesimplementingstandardnear-
IV.4.h.v) relative to Spitzer have not been in-
infraredflatfielding,backgroundremoval,andco-
§
cluded in the errors quoted in Table 1.
additiontechniques. Flat fields in eachband were
constructed using on-source frames. 2MASS stars
2.2.1. Fan Mountain Observatory/FanCam
providedphotometricreferenceinallthreebands,
Y&J photometry and astrometry of WISE leading to the magnitudes reported in Table 1.
1118+31wereobtainedatvariousepochsbetween
28 Nov 2010 and 15 Mar 2012 with FanCam, 2.2.3. WIYN/WHIRC
a HAWAII-1 based near-infrared imager operat-
JHK broad-band imaging of WISE 1118+31
s
ing at the University of Virginia’s Fan Mountain
was obtained on UT 31 Dec 2011 with the WIYN
31-inch telescope (Kanneganti et al. 2009). The
High-ResolutionInfraredCamera(WHIRC,Meixner et al.
source position was dithered by approximately
2010)andtheWIYN3.5-mObservatory. Thedata
10′′ between 30 s exposures comprising total ex- quality is excellent: both seeing ( 0.5′′ with 0.1′′
posure times ranging from 60 to 80 minutes. The ∼
pixel scale) and photometric stability conditions
FanCam field of view (FOV) is 8.7′ (0.51′′/pixel).
were optimal. For each band, individual frames
The central 7′ 7′ (Figure 2) of the combined,
hadexposuretimesof120,120and40seconds,for
×
dithered exposures was fully covered, providing
3
J,H andK ,respectively. WHIRCusesMKOfil- 2.2.5. Keck Adaptive Optics Imaging
s
ters for JHK . Using an efficient on-array dither
s High resolution imaging observations of WISE
pattern,atotalof7,9and43framesforJ,H and
1118+31wereobtainedusingtheKeckIILGS-AO
K , respectively, were obtained, thus providing a
s system (Wizinowich et al. 2006; van Dam et al.
totalexposuretime of840,1080and1720seconds
2006) with NIRC2 on 14 April 2012 (UT). WISE
for the J, H and K mosaics respectively.
s 1118+3125 is not bright enough to serve as the
Individual frames were corrected for pupil-
tip-tilt reference star for the LGS-AO system, so
ghosts, dark and median sky flat subtracted, nor-
we used the nearby 2MASS star (# 1 on Fig-
malizedbydome flat,anddistortioncorrectedus-
ure 2), which has R=16.7 and is located about
inginformationfromtheWHIRCuserinformation 7′′ from the target. The seeing was generally
guide. Astrometric and photometric solutions us- very good throughout the night (0′.′3–0′.′5, how-
ing 2MASS standards were then found. The fully
ever high clouds were present during the observa-
reducedframeswerecombinedintoadeepmosaic,
tions of WISE 1118+31. We used the MKO H
with outlier (bad pixel) rejection applied using filter and narrow plate scale (0.009942′′/pixel for
temporal statistics. A final astrometric and flux a single-frame field-of-view of 10′′ 10′′) for the
calibration was then applied to the deep mosaic. ×
observations. The data were obtained by using
Thephotometricuncertainty(comparingwiththe
a 3-point dither pattern that avoided the noisy,
2MASS PSC) was typically better than 5% for
lower left quadrant of the array. Each image had
each mosaic produced. The achieved spatial res-
anintegrationtimeof120sandtheditherpattern
olution was 0.5 0.7′′ for the final mosaics,
andtheastro∼metric−uncertaintywas 0.05′′. The was repeated five times to give a total exposure
∼ time of 1800 s.
target source, WISE 1118+31, was detected in
The images were reducedin a standardfashion
allthree bands. Using a 1.1′′ radius circularaper-
using custom IDL scripts. These steps included
ture,thebackground-subtractedintegrated(Vega)
darkframesubtraction,flatfielding(usingadome
magnitudes found are reported in Table 1. Note
flat),andskysubtractionfromaskyframecreated
that WHIRC uses MKO filters, and the expected
from the dithered science frames. The individual
difference J2MASS JMKO for a T8.5browndwarf
− frames were then shifted to move WISE 1118+31
is 0.36 mag (Stephens & Leggett 2004). If we
≈ to the center of the array and the stack was me-
correct the J2MASS values by this amount we get
dian averaged to create the final mosaic seen in
three values for JMKO: 17.79, 17.86 & 18.01. The
Figure 3. The FWHM in the final mosaic is 54
mid-range of these values in JMKO = 17.9 which
mas and shows no irregularities or evidence for a
we adopt, giving JMKO W2=4.6.
− close companion.
2.2.4. Spitzer
2.3. Spectroscopy
TheInfraredArrayCamera(IRAC)(Fazio et al.
2.3.1. LBT-LUCIFER
2004) onboard the Spitzer Space Telescope em-
ploys 256 256-pixel detector arrays to image a We obtained an H- and K-band spectrum of
field of vie×w of 5.′2 5.′2 (1′.′2 pixel−1). IRAC was WISE 1118+31 on 2012 Dec 12 (UT) using the
×
used during the warm Spitzer mission to obtain Large Binocular Telescope (LBT) Near-Infrared
deeper photometry in its 3.6 and 4.5 µm channels Spectroscopic Utility with Camera and Integral
(hereafter, ch1 and ch2, respectively) than WISE FieldUnitforExtragalacticResearch(LUCIFER,
was able to take in its W1 and W2 bands. These Mandel et al.2008). Aseriesoftwelve300sexpo-
observations were made as part of Cycle 7 and sureswasobtainedatdifferentpositionsalongthe
Cycle 8 programs 70062 and 80109 (Kirkpatrick, 4′ slittofacilitate skysubtraction. TheA0Vstar
PI). Our standard data acquisition and reduction HD97034wasalsoobservedfortelluriccorrection
methodology for IRAC observationsis outlined in and flux calibration purposes. A series of halogen
Kirkpatrick et al. (2011). lampexposureswerealsoobtainedforflatfielding
purposes.
The data were reduced using custom Interac-
4
tive Data Language (IDL) software based on the
Spextool data reduction package (Cushing et al.
2004). Pairs of images taken at two different po-
sitions along the slit were first subtracted and
flat-fielded. The spectra were then extracted and
wavelength calibrated using sky emission lines of
OH and CH4. The twelve spectra are combined
and then corrected for telluric absorption and
flux calibrated using the technique described by
Vacca et al. (2003). The final spectrum is shown
in Figure 4.
2.3.2. Hale - TripleSpec
A 1 2.5 µm spectrum of WISE 1118+31 was
−
obtained with the Triple Spectrograph (Triple-
spec, Herter et al. 2008) at the 5.08m Hale Tele-
scope at Palomar Observatory. The 1 2.5 µm
−
range is covered over four cross-dispersed orders
which are imaged simultaneously on the 1024
2048 HAWAII-2 array. The 1′′-wide slit provide×s Fig. 3.— H-band LGS-AO image of WISE
1118+3125takenwiththeNIRC2cameraonKeck
a resolving power of R 2700. A series of eight,
≈ II.Theimageis 0′.′6onasidewithNorthupand
300sexposureswereobtainedattwodifferentpo- ≈
sitionsalongthe30′′-longslittofacilitateskysub- Eastto the left. There is no evidence for an equal
brightness companion beyond a separation of 50
traction. The A0 V star HD 99966 was observed
mas.
fortelluriccorrectionandfluxcalibrationpurposes
and dome flats were obtained at the start of the 3.0
night. CH Palomar/TripleSpec
4
The data were reduced using a modified ver- 2.5 LBT/Lucifer
HO
sion of the Spextool (Cushing et al. 2004) pack- 2
age; a detailed description of the reduction steps >
0) 2.0 CH
can be found in Kirkpatrick et al. (2011). Briefly, 6 4
a two-dimensional wavelength solution is derived −1. CH4
usingskyemissionfeaturesofOHandCH4. Spec- 55 1.5
tra are then extracted from pair-subtracted, flat- 1.
(
fielded images. The resulting spectra are com- <f λ 1.0 H2O
bined and corrected for telluric absorption and / λ HO CH
fluxcalibratedonanorder-by-orderbasis. Finally, f 2 4
0.5
the spectra from each order are stitched together
to form a complete 1 2.5 µm spectrum. The
spectrum was then flux−calibrated as described in 0.0
Rayner et al. (2009) using the photometry in Ta-
1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4
ble 1. The final spectrum is shown in Figure 4.
Wavelength (µm)
3. Discussion
Fig. 4.— Spectrum of WISE 1118+31 obtained
3.1. PropertiesoftheCentral StarSystem
with TripleSpec (black) andLUCIFER (red). Re-
ξ Ursae Majoris1 is a complex stellar sys- gions of strong telluric absorption are shown in
tem with at least 4, and possibly 5 components grey. Prominent absorption bands of CH4 and
H2O are indicated. The agreement between the
1alsoAlulaAustralis,HR4374/5, Gl423,HD98230/1 two spectra is excellent.
5
(Mason et al. 1995) known prior to the discov- [Fe/H] = 0.32 0.05 dex.
− ±
ery of WISE 1118+31. Visible to the unaided
eye a short distance from the Big Dipper, this 3.1.2. Chromospheric Activity and Age of the
telescopic double star was among the first to be Central Stars
recognized as a gravitationally bound binary sys-
Cayrel de Strobel et al. (1994) see chromo-
tem (Herschel 1804). Given the 60 year period of
spheric calcium emission lines for ξ UMa B but
the visual pair it was not until 23 years later that
not for ξ UMa A. They propose that the short 4
Struve (1827) calculated a formal orbit. Subse-
day period orbit of the Bb system is driving the
quently both components of the visual pair with
chromosphericactivityinB.Fromthelackofemis-
a = 2.536′′ (Mason et al. 1995) were found to be
sioninA,Cayrel de Strobel et al.(1994)conclude
spectroscopic binaries. Heintz (1967) finds peri-
thatξ UMaisolderthanstarsintheclusterNGC
ods of 669.1 days for the Aa system and 3.9805
752, which has an age of 2 Gyr. Ball et al. (2005)
days for the Bb system. Griffin (1998) gives ve-
report X-ray observations of ξ UMa which show
locity amplitudes, K =4.85 0.14 km/sec and
Aa ± that all of the observed X-ray emission is coming
K = 4.33 0.09 km/sec in the 60 year orbit
Bb ± fromthe 4 day period binary Bb. X-rayflux from
which lead to a dynamical parallax estimate of
component A is more than 2 orders of magnitude
0.126′′ 0.0023′′.
± fainter, implying logLX <27.5. Since component
However So¨derhjelm (1999) re-analyzed the
A is close to solar luminosity, the ratio of X-ray
Hipparcos data combined with other data for vi-
to bolometric flux is R = log(L /L ) < 6
sual binaries and gives 0.1197′′ 0.0008′′ for the X X bol −
which gives an age greater than 4 Gyr using
±
parallax of ξ UMa, and total mass for the AaBb
Equation(A3) in Mamajek & Hillenbrand (2008).
system of 2.62 M⊙. The low space velocity (U,V,W) = ( 2.36
− ±
The weighted mean of the dynamical parallax 0.33, 28.45 1.22, 20.26 0.25) km/sec cal-
and the Hipparcos parallax is 0.1206′′ 0.00074′′ culate−d by ±Karata¸s−et al. ±(2004) implies that
butwithχ2 =8.6for1degreeoffreedo±minthefit ξ UMa is a member of the thin disk, so ages
forthemean,soweinflatetheerrorsbyafactorof much greater than 8 Gyr are unlikely. Finally
√8.6andadopt0.1206′′ 0.0022′′fortheparallax. Cayrel de Strobel et al. (1994) found that the lu-
±
Since the possible 5th component, seen by minosities and effective temperatures were con-
Mason et al.(1995)separatedby56milli-arcseconds sistent within the errors with the masses derived
from the B component of the visual binary, was by Heintz (1967) for a 5 Gyr isochrone calculated
only detected at one epoch we will not consider it with sub-solar abundance.
to be a member of the system.
3.2. Astrometry of the Companion
Bakos et al. (2002) give a proper motion of
0.75′′/yr in position angle 217.49◦, based on Astrometric information was extracted from
a 29 year interval. These values resolve into the observed images at the various epochs us-
0.456′′/yr in RA and 0.595′′/yr in Dec which ing the standard maximum likelihood technique
− −
we adopt. in which a point spread function (PSF) was fit
to each source profile. The positional uncertain-
3.1.1. Spectral Types and Metallicities of the ties were estimated using an error model which
Central Stars
includes the effects of instrumental and sky back-
ground noise and PSF uncertainty. The PSF and
Cayrel de Strobel et al. (1994) used high reso-
its associateduncertainty map were estimated for
lution spectroscopy to find effective temperatures
eachimage individually using a set of brightstars
and gravities of 5950 30 K, logg = 4.3 0.2
± ± in the field. In order to minimize systematic ef-
for ξ UMa A; and 5650 50 K, logg = 4.5 0.2
± ± fects, our astrometry was based on relative posi-
for B. These temperatures and gravities are con-
tions, using asa reference the nearbystar 2MASS
sistent with spectral types of F8.5V and G2V as-
J11183876+3125441 at a separation of approxi-
signed to ξ UMa A and B by Keenan & McNeil
mately6.5′′(star1inFigure2). Thisissufficiently
(1989). Cayrel de Strobel et al.(1994)found that
close that systematic errors due to such effects as
both starshadslightly sub-solarironabundances,
6
2 W1118+31 while the ∆χ2 = 1.28 for 3 extra degrees of
freedom between the best fit and a fit forced to
match the proper motion and parallax of ξ UMa
is perfectly consistent with WISE 1118+31 being
1
n ["] ∆δ a bound member of the ξ UMa system. Figure 5
o shows two fits: one forced to match ξ UMa and
oti
M one with the proper motion and parallax as free
0
∆α cosδ parameters.
Earth & LEO
Spitzer
3.3. The Nature of the Companion
-1
0.0 0.5 1.0 1.5 2.0 2.5 3.0
3.3.1. Spectral Classification
Years since 1/1/2010
As shown in Figure 4, the spectrum of WISE
Fig. 5.— Astrometric data and fits for WISE
1118+31 exhibits deep absorption bands of CH4
1118+31. The blue curves and point are for
and H2O indicative of T dwarfs. We derive a
ground-based or Low Earth Orbit observatories,
more precise spectraltype using the spectralclas-
while the red curves and points are for Spitzer.
sification schemes of Burgasser et al. (2006) and
The bold dashed lines show the fit with proper
the extension to this system by Cushing et al.
motion and parallaxas free parameters,while the
(2011)wherebyUGPSJ072227.51-054031.2(here-
lighter solid curves show the fit forced to match ξ
afterUGPS0722 05,Lucas et al.2010)isdefined
UMa. These fits are very similar. Note that ∆δ −
astheT9spectralstandardandWISE1738+2732
has been displaced by a constant for clarity.
is defined as the Y0 spectral standard. As shown
in Figure 6, WISE 1118+31 has a spectral type
focal-planedistortion,platescale,androtationer- of T8.5 based on the width of the J-band peak at
rors cancel out in the relative position to an ac- 1.27 µm.
curacy much greater than the random estimation
errors. We do, however, assume that the proper 3.3.2. Absolutemagnitude,LuminosityandMass
motionandparallaxofthereferencestararenegli-
For our adopted parallax the absolute magni-
giblecomparedtothoseofWISE1118+31. Insup-
tude is MW2 = W2 + 5log(10̟) = 13.715
portofthis assumption,we findno significantdif- ±
0.051. This agrees very well with the MW2 vs.
ferenceinrelativemotionofWISE1118+31when
type relation in Kirkpatrick et al. (2011) which
the differential astrometry is repeated using more
gives MW2 = 13.71 0.21 for a spectral type
distant 2MASS reference stars, at separations of ±
69′′ and 105′′. The astrometric estimation proce- of T8.5 ± 0.5. The MW2 vs. type relation in
Kirkpatrick et al. (2012) with W1828+2650 ex-
dure is discussed in more detail by Marsh et al.
cludedgivesMW2 =13.64 0.31whichalsoagrees
(2012). ±
with the observed W2 flux. This MW2 = 13.715
Table 2 gives the measured position offsets implies νLν =1.14 10−6 L⊙ at 4.6 µm.
×
whichwereinput to aparallaxandpropermotion
To convert νL to L we need the factor
ν
code that handles both Earth-based and Spitzer
νFν(W2)/Fbol, which is equivalent to the bolo-
observations. Table 3 gives the output from the metriccorrection,sinceνF =(λ F◦/f )10−0.4m
ν iso λ c
code for three different sets of free parameters. (Wright et al.2010)whichis1.118 10−7erg/cm2/sec
In one case the proper motion and parallax were for W2=0 with a ν−1 spectrum×, and Fbol =
forcedtozero. Inthesecondcasetheywereforced 2.48 10−510−0.4mbol erg/cm2/sec, so
to match ξ UMa. In the final case the proper mo- ×
tion and parallax were left as free parameters. νF (W2)
ν
2.5log = 5.87+mbol m(W2).
The astrometric fits for WISE 1118+31 are a (cid:18) Fbol (cid:19) − −
very good match to the motion of ξ UMa. The (1)
∆χ2 = 241.0 between the best fit and a fit with Thus
a fixed position shows that the motion of WISE
1118+31 has been detected with a SNR of 15.5, BC(W2) = mbol m(W2)
−
7
Table 2: Astrometric Observations of WISE 1118+31.
Date ∆αcosδ [′′] ∆δ [′′] Observatory Band
2010.39 0.306 0.162 0.444 0.182 WISE W
− ± − ±
2010.91 0.165 0.211 0.603 0.191 FanMt Y
− ± − ±
2010.91 0.208 0.162 0.477 0.184 WISE W
− ± − ±
2011.05 0.232 0.067 0.794 0.067 Spitzer ch2
− ± − ±
2011.12 0.414 0.234 0.600 0.194 FanMt J
− ± − ±
2011.24 0.393 0.113 0.765 0.100 FanMt J
− ± − ±
2011.38 0.651 0.086 0.829 0.084 MtBglw J
− ± − ±
2011.38 0.554 0.194 0.723 0.190 MtBglw H
− ± − ±
2011.85 0.639 0.190 1.280 0.170 FanMt J
− ± − ±
2012.00 0.650 0.005 1.253 0.004 WIYN J
− ± − ±
2012.00 0.676 0.010 1.235 0.012 WIYN H
− ± − ±
2012.00 0.766 0.037 1.239 0.038 WIYN K
− ± − ±
2012.20 0.850 0.165 1.339 0.153 FanMt Y
− ± − ±
2012.15 0.695 0.117 1.631 0.126 Spitzer ch2
− ± − ±
2012.52 1.081 0.136 1.611 0.139 Spitzer ch2
− ± − ±
Note.—Offsets relativeto11h 18m 38.69s +31◦ 25′ 37.7′′ (J2000).
Referencestar: 11h 18m 38.77s +31◦ 25′ 44.2′′ (2MASS).
Table 3: Astrometric Fits for WISE 1118+31.
Type χ2 #df cosδdα/dt [′′/yr] dδ/dt [′′/yr] ̟ [′′]
Fixed 270.836 28 0 0 0
Forced to ξ UMa 29.856 28 0.456 0.595 0.1206
− −
Free 28.579 25 0.419 0.048 0.563 0.045 0.124 0.033
− ± − ± ±
νFν(W2) bolometric correction. Thus, even if the effective
= 2.5log +5.87.(2)
(cid:18) Fbol (cid:19) temperatureTeff isaparameterofthemodel,that
does not imply that the model parameter is the
Note that νFν/Fbol is non-negative and also actual Teff of the star.
normalized since (νFν/Fbol)dlnν = 1 by defi- GivenamodelspectrumFν,theratioνFν(W2)/Fbol
nition. Since theRnormalization integral is domi- is given by
nated by the brightest peaks, errors in the faint
parts of the spectral energy distribution (SED) νF (νF /hν)R(ν)dlnν
ν ν
=
have only small effects on the derived value of Fbol ([1ergR/cm2]/hν)R(ν)dlnν F dν
ν
νFν(W2)/Fbol orequivalently the bolometriccor- R R (3)
rection. The brightest peaks in νFν/Fbol for late whereR(ν)istheresponseperphoton(Wright et al.
T dwarfs are in the J band and the W2 band, so
2010). The model we use is a linear combination
if we get the J W2 color rightthen our value of
of sulfide cloud models from Morley et al. (2012)
−
νFν(W2)/Fbol will be fairly accurate. with Teff =600K,log(g)=5.0. We formthe sum
We will use a model spectrum to evaluate of0.2times the flux fromthe cloudlessmodeland
νFν(W2)/Fbol in order to fill in the gaps between 0.8 times the flux from the fsed =4 model. Phys-
the photometric passbands, but we are only us- ically this is a brown dwarf that is 20% covered
ing the shape of the spectrum to calculate the by clear zones and 80% covered by cloudy bands.
8
This linear combination gives νFν/Fbol = 1.46,
JMKO W2 = 4.60, (Y J)MKO = 1.23,
− −
(J H)MKO = 0.18, (J Ks)MKO = 0.06
− − − −
and W1 W2 = 3.17. The model matches the
−
observedJ W2colorbydesignbut is too bright
−
in K and too faint in W1. If we make an ad hoc
s
correction to the SED by making the wavelength
WISE 1118+3125
range from 3.0 to 3.8 µm covering the W1 band
4
0.4 mag brighter that would match the W1 W2
−
color and change the normalization integral by
nt 2%. If we make the wavelength range from 2.0
a T7
st 3 to 2.3 µm covering the Ks band 0.8 mag fainter,
n then the J K color matches the data and the
o − s
C normalizationintegralchangesby-1.3%. Thuswe
>
7) adopt νFν(W2)/Fbol = 1.46 with a 10% uncer-
2 T8 taintytoallowforsucherrorsinthemodel,giving
−1. 2 L = (1.14/1.46) 10−6 = 10−6.107±0.043 L⊙ and
6 ×
2 BC(W2) = 6.28 so Mbol = 20.0 0.11 for WISE
1. 1118+31. Interpolating the L vs.±M and t figures
(
<f λ T9 in Saumon & Marley (2008) gives this luminosity
/ λ 1 for masses of 14 to 31 MJ with ages of 2 and 8
f Gyrusingthe cloudymodels,and17to38M for
J
the cloudless models.
Y0 Eventhoughthe modelparametersdo notnec-
0 essarily apply to the actual star, we find that
they do give a mass and luminosity consistent
1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
Wavelength (µm) with those derived above. The Teff = 600 K,
log(g)=5.0 models of Hubeny & Burrows (2007)
have radii of 0.91R , and a mass M = gR2/G=
J
Fig. 6.— Sequence of spectral standards from
32 M . The luminosity calculated directly from
J
T7 to Y0 (red) along with the spectrum WISE Teff and R is L = 4πR2Te4ff = 10−6.01 L⊙ so
1118+31 (black). The spectral standards are
this model is reasonably self-consistent. Calcu-
2MASS 0727+1710 (T7, Burgasser et al. 2006,
lating the temperature from R and L gives Teff =
2002)and2MASS0415-0935(T8,Burgasser et al. (0.91R /R)1/2(567 14) K.
J
2006, 2002), UGPS 0722-05 (T9, Cushing et al. ±
2011; Lucas et al. 2010), and WISE 1738+2732 3.4. Another Binary?
(Y0,Cushing et al.2011). ThespectrumofWISE
1118+3125 has been smoothed to a resolving If something like the Morley et al. (2012) sul-
power of λ/dλ=1000 for display purposes. Spec- fide clouds is not causing the red J W2 color of
−
tra are normalized to unity over the 1.26-1.27 WISE 1118+31, then there is a tension between
µm wavelength range and offset for clarity (dot- the absolute magnitude MW2, which requires a
ted lines). Regions of high telluric absorption are fairly high Teff, and the colors and spectral type
shown in grey. WISE 1118+31 is classified as which suggest a lower temperature. This could
T8.5 0.5. The mismatch between the spectrum be ameliorated if WISE 1118+31 were an equal
ofW±ISE 1118+31andthe templates at 1.63µm mass binary which would increase the total ra-
is due to poor subtraction of the OH sky∼lines. diating area. In this case the absolute magni-
tude of one component of the binary would be
MW2 = 14.47 and the colors would remain the
same. The best fitting Hubeny & Burrows (2007)
model is then a non-chemical equilibrium model
with Teff = 500 K and log(g) = 5.0. This model
9
has MW2 = 14.27, W1-W2 = 4.43, J(MKO)-W2 action that tightened the ξ UMa Bb binary. This
= 4.57, and R = 0.888RJ. νFν/Fbol = 1.546 in would make WISE 1118+31 similar in history to
theW2bandforthismodel,sotheluminosityofa an Oort Cloud comet, formed at a radius of tens
singlecomponentisL=0.5 1.14 10−6/1.546= of AU from the central star system, lifted into an
3.7 10−7 L⊙. The mass r×ange f×or ages from 2 orbit with a high apocenter by multi-body per-
×
to 8 Gyr in the Saumon & Marley (2008) cloud- turbations, followed by galactic tides raising the
less models is 12to 29M , while the cloudy mod- pericenter (Duncan et al. 1987). Given an orbital
J
els give 10 to 25 M . Thus the range of total period of order 105 years, the orbital eccentricity
J
massesis21to59M undertheequalmassbinary of WISE 1118+31 may ultimately be detectable.
J
hypothesis, which is not much different than the Withanapparentorbitalradiusof8.5′ theorbital
range under the single object hypothesis. How- motion will amount to tens of milliarcseconds per
ever, the absolute magnitude of the components year relative to the primary system, and a radial
would no longer agree with the MW2 vs. type re- velocitydifferenceofabout1km/sec. Characteriz-
lation (Kirkpatrick et al. 2011, 2012). ingthismotion,specificallythedifferentialproper
As stated in 2.2.5, we saw no evidence for bi- motion between WISE 1118+31 and its primary,
narityinthe 14§Apr 2010NIRC2 imagesofWISE is tractable given modern infrared and visual as-
1118+31. To determine the upper limit on the trometric capability, and certainly will be accom-
separation of an equal brightness companion, we plished with the passage of time.
added a shifted copy of the mosaic to the original
mosaic. At separations 4 pixels, we are unable 4. Conclusions
≤
to convincingly establish the presence of the com-
Alula Australis (ξ UMa), a solar neighbor-
panion because of the difficulty in differentiating
hood visual binary where each component is it-
betweenanirregularPSFcausedbyanoff-axistip-
selfaspectroscopicbinary,possessesanultra-cool
tiltreferencestarandanelongatedPSFcausedby
browndwarf(T8.5)companionataprojectedsep-
a companion when no other stars are present on
arationof 4100AU. This systemis similar to,but
the array. At separationsof 5 pixels ( 50 mas),
≥ ≥ somewhat more complex than, the binaries stud-
thepresenceofacompanionisobvious. Therefore,
ied by Allen et al. (2012), who found that 1 in
wesettheupperlimitontheseparationto50mas,
5 spectroscopic binary systems had distant com-
slightly less than the FWHM of the image. If this
mon proper motion companions. Faherty et al.
was not due to a temporary conjunction, so the
(2010) found that wide companions were much
actual separation were less than 0.4 AU, then the
morelikelywithbinaryortertiarycentralobjects,
orbital period could be less than 2 years.
and ξ UMa is a continuation of that trend. Thus
the ξ UMa system is uncommon primarily in the
3.5. The View from the Neighborhood of
apparent magnitude of the central star system,
WISE 1118+31
due to its proximity to the Sun. The ξ UMa sys-
With a projected separation of 4100 AU from tem provides a very accurate absolute magnitude
the four main components of the ξ UMa system, for a T8.5 brown dwarf, but both the age of the
WISE 1118+31 has a distant but interesting per- system and the mass of the brown dwarf are un-
spective. Thetwobinariesthatformthe 2′′ visual certainduetothedifficultyofdeterminingtheage
pair as seen from the Earth are separated by 20 of a star in the middle of its main sequence life-
AU, orby anangularscale of15′ fromthe distant time. WISE 1118+31 appears to be slightly less
perspective of WISE 1118+31. At that distance luminous and redder than the T8 dwarf WISEP
the A and B components would each shine with J1423+01whichisacommonpropermotioncom-
an apparent visual magnitude of -9, one hundred panion to BD+01 2920 (Pinfield et al. 2012).
times brighter than Venus in the skies of Earth.
The ξ UMa system is amenable to detailed
Giventhecomplexityoftheξ UMasystem,itis study since it is quite close to the Sun. An
reasonabletospeculatethatWISE1118+31could accurate (sub km/sec) radial velocity for WISE
beacomponentthatwasoncemorecloselybound 1118+31would be valuable informationfor deter-
tothesystem,whichwasthrownintoanorbitwith mining its orbit around ξ UMa. Improved proper
a very large apocenter during a three body inter-
10