The Collapse of the Wien Tail in the Coldest Brown Dwarf? Hubble Space Telescope Near-Infrared Photometry of WISE J085510.83-071442.5
Adam C. Schneider, Michael C. Cushing, J. Davy Kirkpatrick, Christopher R. Gelino
DDraft version September 10, 2018
Preprint typeset using L A TEX style AASTeX6 v. 1.0
THE COLLAPSE OF THE WIEN TAIL IN THE COLDEST BROWN DWARF? HUBBLE SPACE TELESCOPENEAR-INFRARED PHOTOMETRY OF WISE J085510.83 − Adam C. Schneider , Michael C. Cushing , J. Davy Kirkpatrick , & Christopher R. Gelino Department of Physics and Astronomy, University of Toledo, 2801 W. Bancroft St., Toledo, OH 43606, USA; [email protected] Infrared Processing and Analysis Center, MS 100-22, California Institute of Technology, Pasadena, CA 91125, USA NASA Exoplanet Science Institute, Mail Code 100-22, California Institute of Technology, 770 South Wilson Ave, Pasadena, CA 91125,USA
ABSTRACTWe present
Hubble Space Telescope (HST) near-infrared photometry of the coldest known brown dwarf,WISE J085510.83 − − − HST using the F105W, F125W, and F160W filters, which approximate the Y , J , and H near-infrared bands. WISE 0855 − σ magnitude limit of ∼ − ∼ ± J − bandmagnitude reported by Faherty and coworkers. WISE J0855 − ± − − − F125Wand F125W − F160W colors relative to other known Y dwarfs. We find that when compared to themodels of Saumon et al. and Morley et al., the F105W − F125W and F125W − F160W colors of WISE0855 − Keywords: stars: brown dwarfs INTRODUCTIONWithout a stable internal energy source, brown dwarfscontinuously cool over time. At their lowest temper-atures, their atmospheres are composed primarily ofmolecules in both the gas phase and solid phase (in theform of condensate clouds) and thus their emergent spec-tra are sculpted by broad molecular absorption bands ofH O, CH , and NH . The exact chemical composition oftheir atmospheres evolves as they cool, which gives riseto the smooth variation in spectral morphology that isreflected in the MLTY spectral sequence.The vast majority of brown dwarfs in the field ( > T eff =2400–1400 K) and T ( T eff =1400–700 K) dwarfs because theyoperated at red-optical (0.7 − µ m) and near-infrared(1 − µ m) wavelengths. The evolution of a browndwarf through the M → T sequence is controlled by twomain processes: 1) the formation of dust and 2) a shiftin the carbon chemistry (see however, Tremblin et al.(2016) for an alternative explanation). The M/L tran- sition is marked by the formation of dust clouds thatremove refractory elements (TiO, VO, Fe, Si, Al) fromthe gas phase and generate their own continuum opac-ity. The L/T transition is signaled by a sudden lossof this opacity (by some as-yet unknown mechanism)and a shift in the carbon chemistry from being CO-dominated to CH -dominated. Our confidence that thebasic physics for L and T dwarfs is well understood isbuoyed by the fact that atmospheric models do a reason-ably good job of matching their observed spectra fromthe red-optical to the mid-infrared (e.g., Cushing et al.2008, Stephens et al. 2009).Identifying even cooler brown dwarfs required mov-ing to mid-infrared (2.5 − µ m) wavelengths and indeedthe Spitzer Space Telescope and the Wide-field InfraredSurvey Explorer (WISE) have discovered twenty-two ofthe twenty four brown dwarfs known with effective tem-peratures less than ∼
500 K (Cushing et al. 2011, Luh-man et al. 2011, Kirkpatrick et al. 2012, Tinney et al.2012, Kirkpatrick et al. 2013, Cushing et al. 2014, Pin-field et al. 2014, Luhman 2014b, Dupuy et al. 2015, andSchneider et al. 2015), the exceptions being CFBDSIRJ1458+1013 (Liu et al. 2011) and WISE J1217+1626B,a companion to a
WISE identified late-T dwarf (Liu a r X i v : . [ a s t r o - ph . E P ] M a y et al. 2012, Leggett et al. 2015). These brown dwarfs,which populate the Y spectral class (Cushing et al. 2011,Kirkpatrick et al. 2012), have proven much harder to un-derstand primarily because of their paucity and intrinsicfaintness ( M J (cid:38) − − WISE observations by Luhman (2014a) andKirkpatrick et al. (2014) as a high proper motionobject, but it was Luhman (2014b) who measureda distance of only 2 pc, securing it as the fourthclosest system to the Sun. The extremely red colorsof [3.6] − [4.5] = 3.55 mag and J − [4.5] > M [4 . ∼ T eff ∼
250 K) brown dwarf known. Todate, there have been four published efforts to imageWISE 0855 − Y > z AB > H > J MKO = 24.8 +0 . − . mag at 2.6 σ ; Faherty et al. 2014(see Table 1)). WISE 0855 − − Hubble Space Telescope (HST) , we obtained
HST near-infrared images of WISE 0855 − HST/WFC3 OBSERVATIONSWe observed WISE 0855 − λ p =1055.2 nm), F125W ( λ p = 1248.6 nm), and F160W ( λ p = 1536.9 nm) filters of Wide Field Camera 3 (WFC3;Kimble et al. 2008) aboard HST , where λ p is the “pivotwavelength” (see Tokunaga & Vacca 2005). These fil-ters coincide roughly with the Y , J , and H photometric where [3.6] and [4.5] refer to channel 1 and channel 2 of the In-fraRed Array Camera (IRAC; Fazio et al. 2004) aboard the SpitzerSpace Telescope bands. The F105W and F125W observations took placeon UT 2016 Mar 1 with total exposure times of 3835sand 3635s, respectively. The F160W observations tookplace on UT 2016 Mar 15 and UT 2016 Mar 27, with atotal exposure time of 25920s.
Table 1 . WISE J085510.83 − > b mag 2Y > c mag 3F105W > a mag 1J 25.0 +0 . − . (or > d ) mag 4F125W 26.41 ± > c mag 5F160W 23.86 ± > c mag 6W1 17.819 ± ± .
6] 17.44 ± .
5] 13.89 ± a σ upper limit (Section 2). b AB magnitude; S/N < c S/N < d S/N < References — (1) This work; (2) Kopytova et al.(2014); (3) Beam´ın et al. (2014); (4) Faherty et al.(2014); (5) Wright et al. (2014); (6) Luhman (2014b)
All images for each filter were aligned and then com-bined with the tweakreg and astrodrizzle routines avail-able in the AstroDrizzle software (Gonzaga et al. 2012).However, there were several individual exposures forwhich the World Coordinate System (WCS) was in errorduring the first visit for the F160W observations. Thesefew frames were omitted when constructing our F160Wimages. This has no affect on any resulting science con-clusions because WISE 0855 − − σ ). Thirty square arcsecond cut-outs of the combinedF105W, F125W, and F160W images around the positionof WISE 0855 − HST positions
F105W 2016 Mar 1NE5" F125W 2016 Mar 1 F160W 2016 Mar 15
Figure 1 . F105W, F125W, and F160W images centered on the expected position of WISE 0855 − − are due to either proper motion and parallax uncertain-ties, positional uncertainties associated with the IRACcoordinates from which the expected positions were cal-culated, or offsets of the HST
WCS. We do not detectWISE 0855 − − ∼
4. We determine F125W and F160W mag-nitudes following the method of Schneider et al. (2015),whereby we place random apertures around the imagein order to determine the background flux. This processis used because the noise in adjacent pixels becomes cor-related as part of the drizzling process, thereby makingthe common practice of using a sky annulus to determinethe noise of the background unsuitable. The F160W im-age in Figure 1 and the F160W photometry in Table 1come from the first visit.For the F105W image, we measure an upper limitfollowing the method in the
WISE explanatory supple-ment , where we take the flux measurement plus twotimes the measurement uncertainty as the 95% confi-dence upper limit. The aperture for this measurementis placed on the F105W image at the position of WISE0855 − HST pho-tometry from this program, as well as a summary of allpreviously published photometry for WISE 0855 − ANALYSIS3.1.
Colors Versus Spectral Type
With near-infrared photometry in hand, we can nowplace WISE 0855 − http://wise2.ipac.caltech.edu/docs/release/prelim/expsup/sec4 3c.html Y dwarfs. Figure 2 shows the F105W − F125W andF125W − F160W colors of a sample of T and Y dwarfsand WISE 0855 − HST system throughput tables andnear-infrared spectra found in the SpeX Prism Library (Burgasser 2014). Y dwarf colors are determined syn-thetically from the HST spectroscopic sample of Schnei-der et al. (2015). Because there are very few late-Tdwarfs ( > T7) in the SpeX Prism Library and in the
HST sample of Schneider et al. (2015), we supplementthese two datasets with late-T spectra from Kirkpatricket al. (2011). Also shown is the near-infrared photom-etry of two other cold brown dwarfs with estimated ef-fective temperatures less than 300 K; WD 0806 − . The photometry for WISE182831.08+265037.6 is determined synthetically fromits spectrum (Cushing et al., in preparation), whilethe photometry for WD0806 − HST images (Gelino et al., in preparation).Because a spectral type of WISE 0855 − > Y2. Similarly, WISE182831.08+265037.6 and WD0806 − > Y1.The severe turn towards the blue of theF105W − F125W color at the T/Y boundary is typicallyascribed to the loss of the K I and Na I opacity in thered optical as these atoms condense into KCl and Na S(Lodders 1999, Burningham et al. 2010). A similar http://stsci.edu/hst/wfc3/ins performance/throughputs http://pono.ucsd.edu/ ∼ adam/browndwarfs/spexprism/ Note, however, that an independent estimate of the effectivetemperature of WISE 182831.08+265037.6 based on its bolomet-ric luminosity of 520 +60 − K (Dupuy & Kraus 2013) implies thatWISE 182831.08+265037.6 is actually much warmer. It has beensuggested that WISE 182831.08+265037.6 is in fact an unresolvedbinary, which would help explain its unusual luminosity (Leggettet al. 2013).
T0 T1 T2 T3 T4 T5 T6 T7 T8 T9 Y0 Y1 >Y1>Y2
Spectral Type F W − F W WISE 0855WISE 1828WD 0806B T0 T1 T2 T3 T4 T5 T6 T7 T8 T9 Y0 Y1 >Y1>Y2
Spectral Type F W − F W WISE 0855WISE 1828WD 0806BW2354
Figure 2 . HST
F105W − F125W and F125W − F160W colors as a function of spectral type for T and Y dwarfs. trend is seen in the z − J (Lodieu et al. 2013) and Y − J (Burningham et al. 2010, Liu et al. 2012, Leggettet al. 2013, Leggett et al. 2015, Schneider et al. 2015)colors of Y dwarfs. Note, however, that the measured Y − J colors for Y dwarfs are typically much bluer thanmodels that incorporate this loss predict (e.g., Morleyet al. 2014), suggesting that there may be other factorsaffecting the flux around 1 micron (see Liu et al. 2012).At even lower temperatures, models predict this trendturns back towards the red with the precipitous loss ofthe Wien tail (Burrows et al. 2003). The position ofWD 0806 − − − F125W color limit is redder than those of ev-ery other brown dwarf with spectral type ≥ Y1, show-ing that the blueward trend must turn back toward thered somewhere between the Y1 dwarfs, WD0806 − − − F160W colors plateau for T7 to Y0 typeobjects, but there have been hints that this trend alsoturns back to the red (e.g., Cushing et al. 2011, Kirk-patrick et al. 2012, Leggett et al. 2013, Cushing et al.2014, Leggett et al. 2015, Schneider et al. 2015). Inter-estingly, atmospheric models predict a turn to the bluearound 300 K, since the water clouds that form at thesetemperatures absorb light in F160W but not in F125W(Morley et al. 2014). We do not see this blueward trend.Instead, the inclusion of WISE 182831.08+265037.6,WD 0806 − − − F160W colorturns toward the red for the coldest brown dwarfs.In summary, both the F105W − F125W color limit andF125W − F160W are extremely red compared to knownY dwarfs, and can likely be attributed to the col- lapse of the Wien tail at the low temperature of WISE0855 − Color-Magnitude Diagrams
In order to properly compare the
HST photometry ofWISE 0855 − HST photometry for those sam-ples. For the models, colors are found synthetically fromthe model spectra. To obtain absolute F125W mag-nitudes, we measure the J MKO − F125W color for eachmodel spectrum synthetically, then apply these offsetsto the absolute J MKO provided for each model . Colorsand absolute magnitudes for known brown dwarfs andlow-mass stars are found in a similar fashion: we firstdetermine colors synthetically from spectra available inthe SpeX Prism Library, Kirkpatrick et al. (2011), andSchneider et al. (2015). For the HST spectra, we mea-sure F125W and F160W magnitudes directly since thespectra are absolutely flux calibrated. For spectra fromthe SpeX Prism Library and Kirkpatrick et al. (2011),we first compute the J MKO − F125W color syntheticallyfrom the spectrum, and determine F125W magnitudesfrom measured J MKO values. The majority of J MKO andparallax measurements come from the Database of Ul-tracool Parallaxes maintained by Trent Dupuy (Dupuy& Liu 2012). Additional photometry and parallaxeswere taken from Kirkpatrick et al. (2011), Scholz et al.(2012), Mace et al. (2013), Marsh et al. (2013), Beich-man et al. (2014), and Tinney et al. (2014). The averagevalues of J MKO − F125W and H MKO − F160W for knownY dwarfs are − − ∼ cmorley/cmorley/Models.html ∼ tdupuy/plx/Database of Ultracool Parallaxes.html F125W − F160W M F W
225 K250 K275 K300 K325 K350 K375 K400 K450 K
Empirical DataM dwarfsL dwarfsT dwarfsY dwarfsModel TracksS12 (log g = 4)M14 (log g = 4.5)M14 (log g = 4)M14 (log g = 3.5)
W0855W0647 WD0806BW1828W2209W1639W1405 W1541W0359W0410W1738 W2056W2220
F125W − [4.5] M F W
225 K250 K275 K300 K325 K350 K375 K400 K450 K
Empirical DataM dwarfsL dwarfsT dwarfsY dwarfsModel TracksS12 (log g = 4)M14 (log g = 4.5)M14 (log g = 4)M14 (log g = 3.5)
W0359 W1828 W0855WD0806BW0410 W0647W1405 W1541W1639W1738 W2056 W2209W2220
F160W − [4.5] M F W
225 K250 K275 K300 K325 K350 K375 K400 K450 K
Empirical DataM dwarfsL dwarfsT dwarfsY dwarfsModel TracksS12 (log g = 4)M14 (log g = 4.5)M14 (log g = 4)M14 (log g = 3.5)
W0359W0410 W0647W1405 W1541W1639W1738 W1828W2056 W2209 W2220WD0806B W0855
F125W − [4.5] M [ . ]
225 K250 K275 K300 K325 K350 K375 K400 K450 K
Empirical DataM dwarfsL dwarfsT dwarfsY dwarfsModel TracksS12 (log g = 4)M14 (log g = 4.5)M14 (log g = 4)M14 (log g = 3.5)
W0359W0410 W0647W1405W1541W1639W1738 W1828W2056 W2209W2220 WD0806B W0855
Figure 3 . Color magnitude diagrams for WISE 0855 − f sed = 5 model tracks (dottedline). Previous studies have shown that J − H colors of thecoldest brown dwarfs are not well reproduced by lowtemperature models (e.g., Leggett et al. 2015, Schnei-der et al. 2015). Models generally predict much bluercolors for a given absolute magnitude than what ismeasured. The position of WISE 0855 − − F160W color-magnitude diagram in the top leftpanel of Figure 3 shows this issue extends down to thetemperature of WISE 0855 − − T eff estimates for WISE 0855 − f sed ), so we choose to only compare to f sed = 5models. We also chose only to employ models with sur-face gravities consistent with the evolutionary models ofSaumon & Marley (2008).As seen in the figure, there is no single model downto T eff ∼
225 K that reaches a F125W − F160W color asred as the color of WISE 0855 − − F160W color of the T eff = 225 K, log g =4.5, f sed = 5 model of Morley et al. (2014) comes theclosest to WISE 0855 − − F125W color of ∼ J − H colors of Y dwarfs than those of Saumon et al.(2012) (see their Figure 4). Applying our J MKO − F125Wand H MKO − F160W offsets to WISE 0855 − M J ≈ J − H ≈ M J value of 29 mag has a log g = 4.5 and iscomputed with equilibrium chemistry. While that modeltrack extends beyond the boundaries of the figure, weestimate its J − H color to be ∼ − J − H ≈ J − H ≈
0. The top and bottom right panels of Figure 3 show theabsolute F125W and [4.5] magnitudes as a function ofF125W − [4.5] color. Model F125W − [4.5] colors repro-duce the measured colors of cold brown dwarfs muchbetter than the F125W − F160W model colors. Whencompared to the log g = 4 Morley et al. (2014) mod-els that extend down to 200 K, the position of WISE0855 − ∼
225 K and ∼
250 K, similar to previous estimates(Beam´ın et al. 2014, Faherty et al. 2014, Kopytova et al.2014, Luhman 2014b, Luhman & Esplin 2014).We also show the F160W absolute magnitude as afunction of F160W − [4.5] color in the bottom left panelof Figure 3. This diagram is notable because the F160Wand [4.5] bands have the smallest photometric uncer-tainties of all the measurements in Table 1. In this dia-gram, WISE 0855 − − DISCUSSIONFaherty et al. (2014), using their 2.6 σ J − banddetection of WISE 0855 − − WISE channel 2 (W2) magnitudes and J − W2color to those from models. However, using an updatedparallax measurement and the J − band measurement ofFaherty et al. (2014), Luhman & Esplin (2014) find that the absolute [4 .
5] magnitude and the J − [4.5] color wereconsistent with Saumon et al. (2012) cloudless modelswith an atmosphere whose carbon and nitrogen chem-istry is out of equilibrium due to vertical mixing in theatmosphere. Our new position for WISE 0855 − [4 . vs. F125W − [4.5] color-magnitude diagram(bottom right panel of Figure 3) does little to clear upwhether it better matches cloud-free, water-cloud, ornon-equilibrium chemistry models because we do nothave non-equilibrium models. Nevertheless, the factthat its position agrees extremely well with the Morleyet al. (2014) water-cloud models in some diagrams inFigure 3, and not as well in others, suggests that draw-ing conclusions from a single color-magnitude diagramwhen considering competing models may be premature.Instead, a more comprehensive view of its full spectralenergy distribution is likely needed.Recent work has shown that non-equilibrium chem-istry driven by vertical mixing is likely important in coldbrown dwarf atmospheres, and may play a large role inshaping their final spectroscopic shapes (e.g., Schneideret al. 2015, Tremblin et al. 2015, Leggett et al. 2016).Non-equilibrium chemistry in the atmosphere of WISE0855 − − F160 color differences). A larger suite of mod-els including the effects of non-equilibrium chemistryand cloud formation that extends down to T eff valuesof ∼
200 K will be necessary to draw any further con-clusions of the atmospheric properties of this extremelycold object. ACKNOWLEDGMENTSWe thank the anonymous referee whose comments im-proved the clarity of this paper. We thank Mark Marley,Didier Saumon, and Caroline Morley for fruitful dis-cussions and for graciously making their models pub-licly available online. This work is Based on observa-tions made with the NASA/ESA Hubble Space Tele-scope, obtained at the Space Telescope Science Insti-tute, which is operated by the Association of Universi-ties for Research in Astronomy, Inc., under NASA con-tract NAS 5-26555. These observations are associatedwith program
Wide-fieldInfrared Survey Explorer , which is a joint project ofthe University of California, Los Angeles, and the JetPropulsion Laboratory/California Institute of Technol-ogy. This research has benefitted from the SpeX PrismSpectral Libraries, maintained by Adam Burgasser athttp://pono.ucsd.edu/ ∼ adam/browndwarfs/spexprism.REFERENCESadam/browndwarfs/spexprism.REFERENCES