1: \documentclass[12pt,preprint]{aastex}
2:
3: %\usepackage{natbib}
4: %\citestyle{aa}
5: %\slugcomment{submitted to ApJ Letters, \today}
6:
7: \begin{document}
8:
9: \title{Direct measurement of the size of 2003 UB313 from the Hubble Space Telescope}
10: \author{M.E. Brown\altaffilmark{1}, E.L. Schaller\altaffilmark{1}, H.G. Roe
11: \altaffilmark{1}, D.L. Rabinowitz\altaffilmark{2}, C.A. Trujillo\altaffilmark{3}}
12: \altaffiltext{1}{Division of Geological and Planetary Sciences, California Institute
13: of Technology, Pasadena, CA 91125}
14: \altaffiltext{2}{Department of Physics, Yale University, New Haven, CT 06520}
15: \altaffiltext{3}{Gemini Observatory, 670 North A'ohoku Place, Hilo, HI 96720}
16: \email{mbrown@caltech.edu}
17:
18:
19: \begin{abstract}
20: We have used the Hubble Space Telescope to directly measure the angular size
21: of the large Kuiper belt object 2003 UB313. By carefully calibrating the
22: point spread function of a nearby field star, we measure the size
23: of 2003 UB313
24: to be 34.3$\pm$1.4 milliarcseconds, corresponding to a diameter of
25: 2400$\pm$100 km or a size $\sim5$\% larger than Pluto.
26: The V band geometric albedo of 2003 UB313 is $86\pm7$\%. The extremely high
27: albedo is consistent with the frosty methane spectrum, the lack of red coloring,
28: and the lack of observed photometric variation on the surface of 2003 UB313.
29: Methane photolysis should quickly darken the surface of 2003 UB313, but
30: continuous evaporation and redeposition of surface ices appears capable
31: of maintaining the extreme alebdo of this body.
32:
33: \end{abstract}
34:
35: \keywords{comets: general -- infrared: solar system -- minor planets}
36:
37: \section{Introduction}
38: The planetary-sized scattered Kuiper belt object 2003 UB313 was discovered
39: in an ongoing survey at the Samuel Oschin telescope at Palomar Observatory
40: \citep{2005ApJ...635L..97B}. The heliocentric distance of 97.5 AU and V magnitude of
41: 18.8 at the time of discovery implied a size larger than Pluto for any albedo
42: %2209 km
43: lower than 96\%, but the initial discovery was incapable of providing a
44: measurement of the albedo. The infrared reflectance
45: spectrum of 2003 UB313 is dominated
46: by absorption due to frozen methane \citep{2005ApJ...635L..97B}, like the spectra of Pluto and Triton,
47: suggesting the likelihood of a high albedo for 2003 UB313 like the
48: 50 - 80\% albedos of these two objects. With such a high albedo
49: 2003 UB313 would have a diameter in the range of 2500 to 3200 km.
50:
51: To date, most Kuiper belt objects with measured sizes have had their diameters determined through radiometry. In this technique, the measured thermal infrared flux (or fluxes)
52: from an object is converted to a diameter using models in which the thermal emission properties and surface temperature distributions of the object are assumed
53: \citep{2004A&A...415..771A,2005ApJ...624L..53C}. Owing to the uncertainties in the appropriate
54: model parameters, errors in these size determinations are large.
55: The object (50000)
56: Quaoar had its size directly measured by Hubble Space Telescope (HST)
57: observations \citep{2004AJ....127.2413B} which are unaffected by uncertainties in
58: thermal emission modeling. The major source of uncertainties in these
59: measurements is the unknown center-to-limb photometric function of
60: Quaoar, which changes the apparent angular diameter of the disk.
61:
62: Initial Spitzer Space Telescope observations of
63: 2003 UB313 resulted in a 70 $\mu$m
64: upper limit of 2 mJy, consistent with a size upper
65: limit of 3310 km and lower limit to the albedo of 44\% (Brown et al, in prep).
66: More recent observations from the IRAM telescope succeeded in detecting
67: a flux of $1.27 \pm 0.29$ mJy at 1.2mm, suggesting a size between
68: 2600 and 3400 km and an albedo between 42 and 72\% \citep{Bertoldi}.
69: An object of this size would have an angular diameter seen from the earth
70: of between 37 and 48 mas, similar to or larger than Quaoar,
71: thus
72: should be directly resolvable from HST observations. We present here
73: such observations and the size and albedo of 2003 UB313 directly
74: measured by HST.
75:
76: \section{Observations}
77: We obtained 16 119-second exposures and
78: 12 133-second exposures of 2003 UB313 using the High Resolution Camera
79: (HRC) of the Advanced Camera for Surveys (ACS) in the F435W
80: filter in two orbits of the Hubble Space Telescope (HST).
81: The first observations began at 8:29 on 9 December 2005 (UT) and one observation
82: was obtained every 170 seconds. The second observations
83: began at 6:52 on 10 December 2005
84: and were obtained every 184 seconds.
85: The timing of the observations was chosen
86: so that 2003 UB313 would be a few arcseconds from a moderately
87: bright field star located at 01h35m37.63 -05d38m57.8, and
88: the telescope was tracked at the sidereal rate so that
89: accurate point spread function (PSF)
90: measurements could be obtained from the untrailed image of the star.
91: The motion of 2003 UB313 caused
92: a smearing of about 30 mas of the image of 2003 UB313
93: during each observation, a
94: factor which needed to be taken carefully into account in the analysis of
95: the angular size of the object.
96:
97: The star had a brightness of $58.8\pm0.2$
98: counts s$^{-1}$ ($B=20.152\pm 0.004$)
99: in each of the images
100: and was single, well exposed, and not saturated. 2003 UB313 had a brightness
101: of $149.9\pm0.7$ counts s$^{-1}$ ($B=19.136\pm 0.005$) and traveled across a
102: field uncontaminated by
103: detectable background
104: determined from our full PSF fitting analysis,
105: described below.
106:
107: The method used to determine the size of 2003 UB313 from these observations
108: is almost identical to that described and validated
109: in detail in \citet{2004AJ....127.2413B}. Here, we very briefly
110: summarize the procedure, and particularly discuss a few small
111: improvements.
112: First, for each of the 28 individual stellar
113: images, the PSF is modeled using the HST PSF
114: simulator package TinyTim (Krist \& Hook, 2001) . While TinyTim's
115: default parameters provide a
116: close match to the average HRC PSF, the HST PSF is not absolutely stable,
117: so a fit is performed to the field star in each image.
118: To obtain a best-fit PSF, we use a downhill simplex $\chi^2$
119: minimization method
120: to fit the first 8 Zernike
121: terms in the aberration of the TinyTim model at the location of the star.
122: Figure 1 shows the best-fit focus for the individual images, while Figure
123: 2 gives an example of the best-fit model for one of the images.
124: In Brown \& Trujillo and in subsequent experiments on other images
125: we found that the focus varied smoothly over
126: the course of an orbit. Here we see the same trend but with a few significant
127: outliers. While some of this scatter is likely real and
128: caused by the increase in
129: jitter with HST now in two-gyro mode, we chose to discard as
130: unreliable the images in which the measured focus appears to vary greatly
131: between adjacent observations (though we note that the final results are
132: not sensitive to the removal of these points).
133:
134: With the aberrations measured for each stellar image, TinyTim is used
135: to construct a PSF at the location of 2003 UB313, accounting for the
136: (small) PSF
137: changes due to the field-dependent aberration of the HRC.
138: This PSF is convolved
139: with the measured motion vector of the object and then convolved with
140: model disks with diameters between 10 and 50 milliarcseconds.
141: This model image is then convolved with the three-by-three pixel gaussian
142: kernel representing the effects of CCD charge diffusion.
143: For each
144: image a best-fit model disk diameter is found which minimizes the
145: sum of the square of the residuals between the modeled image and the data.
146: An example of a best-fit model is shown in Figure 2.
147:
148: In this process, the two steps which are the largest source of potential error
149: are the choice of the charge diffusion kernel and the choice of the
150: center-to-limb brightness profile of the model disks. TinyTim provides an
151: estimate of the position-dependent charge diffusion kernel on the HRC CCD,
152: but this estimate is based on extrapolations from measurements at other
153: wavelengths \citep{2004SPIE.5499..328K}.
154: Because this parameter is critical, we chose to
155: independently measure the charge diffusion at the location of our star
156: and of 2003 UB313. We used archival
157: images of the globular cluster M92 obtained with HRC in the
158: F435W filter and found several
159: isolated stars near the positions of the central PSF star and the positions
160: of 2003 UB313 on the two separate orbits.
161: We used the procedure described above to fit the PSFs of the
162: globular cluster stars near the location of the central PSF star,
163: and then we transfered this PSF to the locations of 2003 UB313
164: using TinyTim.
165: We then performed a $\chi^2$ minimization of the charge diffusion kernel
166: of the globular cluster stars at the locations of 2003 UB313. This procedure
167: does not give us a reliable measure of the absolute width of the charge
168: diffusion kernel (as we have to assume an initial charge diffusion
169: kernel for the
170: central region), but only a measure of the difference between the center
171: location
172: of the PSF star and 2003 UB313. Fortunately, our measurements are only
173: sensitive to relative difference and are not dependent on the absolute
174: values. We find that the Gaussian
175: width of the charge diffusion kernel at the positions
176: of 2003 UB313 are identical for the two measured positions
177: and are
178: $1.04\pm0.01$ times
179: larger than the width at the center
180: PSF star.
181:
182:
183: The other important factor to consider
184: is the center-to-limb brightness profile of the model disks. An object
185: with a flat center-to-limb profile will appear larger than an object
186: whose brightness drops steeply near the limb.
187: In \citet{2004AJ....127.2413B} we considered an extremely wide range of possible profiles
188: as nothing was known about the surface of Quaoar at the time. For 2003 UB313,
189: however, we know that the surface is covered in methane frost \citet{2005ApJ...635L..97B} like
190: the surfaces of Pluto and Triton. As we have precise measurements of
191: the center-to-limb profile of Triton from well resolved Voyager measurements,
192: we take this profile, parameterized with a Hapke model \citep{1994Icar..109..296H}, as our
193: best analog for the profile of 2003 UB313.
194:
195: The size measured for 2003 UB313 in each of the
196: 17 images not discarded due to
197: focus deviations is shown in Figure 1.
198: Taking into account the random errors and charge diffusion
199: width uncertainty of 1.0 mas each
200: we derive an angular size for 2003
201: UB313 of $34.3\pm1.4$ mas. At a geocentric distance of 96.4 AU, this angular
202: diameter corresponds to a diameter of %(69.92 km/mas)
203: $2400 \pm 100$ km. For an absolute
204: $V$ magnitude of $H_v=-1.12\pm 0.01$
205: the albedo of of 2003 UB313 is $0.85\pm0.07$.
206:
207: \section{Discussion}
208: Though 2003 UB313 is 20\% brighter than Pluto in absolute terms, it is only
209: 6\% larger, as the albedo of 2003 UB313 is even higher than that of Pluto.
210: This extremely high derived albedo gives us confidence in our choice
211: of Triton as an analog for the center-to-limb function. Lower albedo
212: icy satellites have center-to-limb profiles which are much flatter than
213: that of Triton. Use of an icy-satellite like profile would give a smaller
214: size and thus even
215: higher albedo for 2003 UB313, inconsistent with the low albedos
216: of the icy satellites. We can, in fact, self-consistently solve for
217: the center-to-limb profile and the albedo by assuming a Hapke model to
218: give both
219: the profile and the geometric albedo. The profile gives a unique measurement
220: of the angular size, and thus the albedo, but the Hapke model independently
221: gives an albedo. The only Hapke models which give consistent albedos have
222: parameters very similar to Triton, with high single-scattering albedo strongly
223: backscattering particles and smooth surfaces.
224:
225:
226: The measured size is well below the Spitzer upper limit of 3310 km, but
227: apparently
228: inconsistent with the reported IRAM measured size of 3000$\pm$400 km
229: \citep{Bertoldi}.
230: A reevaluation of the IRAM measurement, however, suggests that the
231: measurements are indeed compatible. The IRAM size measurement
232: assumed an absolute magnitude for 2003 UB313 of $H_v=-1.16\pm0.1$, which is
233: about 4\% brighter than the best current measurement of $H_v=-1.12\pm0.01$
234: (Rabinowitz et al., in prep), suggesting a smaller size.
235: Reevaluating a thermal model with the fainter absolute magnitude
236: and assuming that 2003 UB313
237: is pole-on and is the largest size allowed by our uncertainties predicts
238: a 1.2mm emission of 0.94 mJy, consistent at the 1.1$\sigma$ level
239: with the $1.27\pm0.29$ mJy reported
240: by IRAM. Even an equator-on orientation would give a flux of 0.70 mJy,
241: consistent at the 2$\sigma$ level.
242:
243: The $86\pm7$\% albedo of 2003 UB313 is significantly
244: higher than the $\sim$60\% average of Pluto, but
245: consistent with the albedo of
246: the individual brightest areas on Pluto \citep{2001AJ....121..552Y}.
247: 2003 UB313 is also less red than Pluto \citep{2005ApJ...635L..97B}, and Pluto's darkest
248: regions appear to be a major source of
249: the red color \citep{1999AJ....117.1063Y}. These two
250: characteristics suggest that the surface of 2003 UB313 might resemble
251: the brightest regions on Pluto while containing few if
252: any of the darkest regions,
253: a suggestion consistent with the lack of detectable photometric
254: variation on 2003 UB313 (Roe et al., in prep.).
255:
256: The geometric albedo of 2003 UB313 is higher than that of all known bodies
257: in the solar system with the exception of Saturn's satellite Enceladus,
258: which has a geometric albedo of 103\%
259: and has active water plumes capable of providing continuous frost
260: resurfacing (Hansen et al. 2006).
261: Even
262: Triton, which has active geysers and appears
263: freshly resurfaced, has a slightly lower albedo of 0.77 \citep{1994Icar..109..296H}. Typical
264: inactive
265: water-ice covered satellites in the outer solar system have albedos
266: ranging from 0.2 to 0.4, with exceptions only related to activity
267: or exogenic processes.
268: On 2003 UB313 even highly reflective methane frosts will irreversibly darken
269: due to long-term photolysis;
270: some uncommon process which continues to maintain a
271: high albedo on 2003 UB313 is required.
272:
273: One process unique on large bodies in the Kuiper belt with eccentric orbits
274: is atmospheric
275: freeze-out. At 2003 UB313's 38 AU perihelion equilibrium temperature of
276: $\sim$43 K,
277: the vapor pressure over a solid nitrogen surface would be $\sim$30 Pa,
278: while
279: the pressure over a solid methane surface would be $
280: \sim$0.2 Pa (Lodders and Fegley, 1998).
281: At the current distance of
282: 97 AU and equilibrium temperature of 27 K these pressures drop by 6 and 8
283: orders of magnitude,
284: respectively. Any nitrogen or methane atmosphere that exists at perihelion
285: is essentially completely frozen to the surface by aphelion.
286:
287: This 560-year cycle of
288: evaporation and then freeze-out of the atmosphere of 2003 UB313
289: will leave fresh undarkened ices on the top layer at aphelion.
290: A 30 Pa nitrogen atmosphere
291: at perihelion corresponds to a layer of solid nitrogen several centimeters
292: thick at aphelion, while a 0.2 Pa methane atmosphere corresponds to a
293: layer tens of microns thick. These layers would sit on top of any dark
294: photolyzed methane and would maintain a high albedo for grain sizes smaller
295: than the surface thickness, which appears likely. At perihleion the dark areas
296: would be revealed, leading to a Pluto-like appearance. 2005 FY9, which
297: also has a surface covered in methane frost
298: \citep{2005DPS....37.5211B,
299: 2006A&A...445L..35L}
300: and is currently at a distance of 52 AU, should have a surface temperature
301: approximately midway between Pluto and 2003 UB313. The presence of
302: small but discernible
303: photometric variations on 2005 FY9 (Roe et al., in prep.) and the evidence
304: for a small surface coverage of warm dark material (Brown et al., in prep.)
305: suggests that 2005 FY9 could have undergone partial atmospheric freeze-out,
306: but that at the temperature of 2005 FY9 the process is not complete as on
307: 2003 UB313. The
308: process of atmospheric freeze-out, which has long been debated for
309: Pluto (i.e Stern \& Trafton 1984, Stansberry \& Yelle 1999, Eliot et al. 2004)
310: can now be studied in a growing population of methane rich objects at
311: a wide range of distances and temperatures in the outer solar system.
312:
313: {\it Acknowledgments:} We would like to thank the director and staff
314: at STScI for providing the opportunity to make these measurements
315: and the assistance in making them
316: happen. This research has been supported by a
317: grant from STScI and from NASA Planetary Astronomy.
318:
319: \begin{thebibliography}{17}
320: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
321:
322: \bibitem[{{Altenhoff} {et~al.}(2004){Altenhoff}, {Bertoldi}, \&
323: {Menten}}]{2004A&A...415..771A}
324: {Altenhoff}, W.~J., {Bertoldi}, F., \& {Menten}, K.~M. 2004, \aap, 415, 771
325:
326: \bibitem[{{Barkume} {et~al.}(2005){Barkume}, {Brown}, \&
327: {Schaller}}]{2005DPS....37.5211B}
328: {Barkume}, K.~M., {Brown}, M.~E., \& {Schaller}, E.~L. 2005, AAS/Division for
329: Planetary Sciences Meeting Abstracts, 37,
330:
331: \bibitem[{{Bertoldi} {et~al.}(2006){Bertoldi}, {Altenhoff}, {Weiss}, {Menten},
332: \& {Thum}}]{Bertoldi}
333: {Bertoldi}, F., {Altenhoff}, W., {Weiss}, A., {Menten}, K.~M., \& {Thum}, C.
334: 2006, {Nature}, 439, 563
335:
336: \bibitem[{{Brown} \& {Trujillo}(2004)}]{2004AJ....127.2413B}
337: {Brown}, M.~E. \& {Trujillo}, C.~A. 2004, \aj, 127, 2413
338:
339: \bibitem[{{Brown} {et~al.}(2005){Brown}, {Trujillo}, \&
340: {Rabinowitz}}]{2005ApJ...635L..97B}
341: {Brown}, M.~E., {Trujillo}, C.~A., \& {Rabinowitz}, D.~L. 2005, \apjl, 635, L97
342:
343: \bibitem[{{Cruikshank} {et~al.}(2005){Cruikshank}, {Stansberry}, {Emery},
344: {Fern{\'a}ndez}, {Werner}, {Trilling}, \& {Rieke}}]{2005ApJ...624L..53C}
345: {Cruikshank}, D.~P., {Stansberry}, J.~A., {Emery}, J.~P., {Fern{\'a}ndez},
346: Y.~R., {Werner}, M.~W., {Trilling}, D.~E., \& {Rieke}, G.~H. 2005, \apjl,
347: 624, L53
348:
349: \bibitem[{{Eliot} {et~al.}(2004)}]{a}{Eliot}, J.L., et al. 2004, Nature, 424, 165
350:
351: \bibitem[{{Hansen} {et~al.}(2006)}]{b} {Hansen}, C.J., et al. 2006, Science, 311, 1422
352:
353: \bibitem[{{Hillier} {et~al.}(1994){Hillier}, {Veverka}, {Helfenstein}, \&
354: {Lee}}]{1994Icar..109..296H}
355: {Hillier}, J., {Veverka}, J., {Helfenstein}, P., \& {Lee}, P. 1994, Icarus,
356: 109, 296
357:
358: \bibitem[{{Krist}(2004)}]{2004SPIE.5499..328K}
359: {Krist}, J.~E. 2004, in Optical and Infrared Detectors for Astronomy. Edited by
360: James D. Garnett and James W. Beletic. Proceedings of the SPIE, Volume 5499,
361: pp. 328-335 (2004), 328--335
362:
363: \bibitem[{{Krist \& Hook}(2001)}] {bbb}
364: Krist, J. \& Hook, R. 2001, The Tiny Tim User's Guide (ver 6.0; Baltimore: STScI)
365:
366: \bibitem[{{Licandro} {et~al.}(2006){Licandro}, {Pinilla-Alonso}, {Pedani},
367: {Oliva}, {Tozzi}, \& {Grundy}}]{2006A&A...445L..35L}
368: {Licandro}, J., {Pinilla-Alonso}, N., {Pedani}, M., {Oliva}, E., {Tozzi},
369: G.~P., \& {Grundy}, W.~M. 2006, \aap, 445, L35
370:
371: \bibitem[{aaaa (2006)}] {aaaa}
372: Lodders, K. \& Fegley, B. 1998, The Planetary Scientist's Companion (New York: Oxford Univ. Press)
373:
374: \bibitem[{{Stern \& Trafton}(1984)}]{c}
375: {Stern}, S.A \& {Trafton}, L.M. 1984, Icarus, 58, 312
376:
377: \bibitem[{{Stansberry \& Yelle}(1999)}]{d}
378: {Stansberry}, J.A. \& {Yelle}, R.V. 1999, Icarus, 141, 299
379:
380: \bibitem[{{Young} {et~al.}(2001){Young}, {Binzel}, \&
381: {Crane}}]{2001AJ....121..552Y}
382: {Young}, E.~F., {Binzel}, R.~P., \& {Crane}, K. 2001, \aj, 121, 552
383:
384: \bibitem[{{Young} {et~al.}(1999){Young}, {Galdamez}, {Buie}, {Binzel}, \&
385: {Tholen}}]{1999AJ....117.1063Y}
386: {Young}, E.~F., {Galdamez}, K., {Buie}, M.~W., {Binzel}, R.~P., \& {Tholen},
387: D.~J. 1999, \aj, 117, 1063
388:
389: \end{thebibliography}
390:
391: \clearpage
392:
393: \begin{figure}
394: \plotone{f1.eps}
395: \caption{The top panel shows the best-fit focus (in units of RMS waves
396: of aberration at 547 nm) for the 28 images. Previous experience has
397: shown that the focus varies smoothly over the course of an orbit. We
398: expect that the large jumps in derived focus are a product of the
399: lower signal-to-noise of the PSF star and of the increased jitter in
400: two gyro mode. We discard as potentially unreliable all of the measurements
401: marked with a ''X'' in which the focus appears to shift significantly
402: between images. The bottom panel shows the best fit disk size for each of the
403: individual images. The mean is shown as the solid line and the standard
404: deviation is shown as the dashed line.}
405: \end{figure}
406:
407: \clearpage
408:
409: \begin{figure}
410: \plotone{f2.eps}
411: \caption{Comparison of the star and 2003 UB313 in image 21 with
412: the best-fit models. The data and models share identical
413: logarithmic scales. The residuals are linearly scaled from
414: $\pm5$\% of the maximum of the model image.}
415: \end{figure}
416:
417: \end{document}
418: