ICARUS 126, 212±217 (1997) ARTICLE NO. IS965646 The Extraordinary Colors of Trans-Neptunian Objects 1994 TB and 1993 SC S. C. TEGLER Department of Physics and Astronomy, Northern Arizona University, Flagstaff, Arizona 86011 E-mail: [email protected] AND W. ROMANISHIN Department of Physics and Astronomy, University of Oklahoma, Norman, Oklahoma 73019 Received June 20, 1996; revised October 31, 1996 In 1992, Jewitt and Luu (1993) carried out a survey of We present B±V and V±R colors for trans-neptunian objects 0.7 square degrees of the sky to a limiting magnitude of (TNOs) 1994 TB and 1993 SC. We have found that these TNOs R p 25 and discovered the ®rst TNO (1992 QB1). Thirty- have colors almost as red as 5145 Pholus which are therefore six TNOs have been discovered as of June, 1996. The consistent with surfaces rich in complex organic molecules. We TNOs have effective diameters ranging from 100 to 400 have measured B±V colors of 1.10 6 0.15 and 1.27 6 0.11 for km (Jewitt and Luu 1995), making them signi®cantly larger 1994 TB and 1993 SC, respectively, as well as V±R colors of 3 0.68 6 0.06 and 0.70 6 0.04 for 1994 TB and 1993 SC, respec- than typical comets such as 1P/Halley (15 8 km; Keller tively. Currently, there are only four TNOs with published et al. 1987). Jewitt and Luu have estimated that there are broadband colors (1992 QB1, 1993 FW, 1993 SC, and 1994 p35,000 TNOs larger than 100 km in the 30 to 50 AU TB). TNO colors range from solar (a single V±R measurement heliocentric distance range. of 1993 FW in the literature) to nearly the reddest objects in Almost nothing is known about the chemical nature the solar system (our measurements of 1993 SC and 1994 TB). of the TNOs that have been discovered so far. It is Although we have concluded 1993 SC is a very red TNO, a possible, perhaps likely, that they are mostly similar in measurement in the literature has suggested that there is noth- composition to the short-period comets that we believe ing special about the color of 1993 SC. Because there are so are derived from the Kuiper Belt, though cosmic-ray few TNOs with color measurements and because colors for 1993 SC are inconsistent, we cannot say yet with any certainty processing may have greatly altered their surfaces. How- whether the surfaces of TNOs exhibit a diversity of colors and ever, TNOs span wide ranges in size and heliocentric hence a diversity of surface compositions. 1997 Academic Press distance and some or all of them may have compositions very different from those of short-period comets. Compo- sition may also be a function of heliocentric distance, thereby providing a signature of formation mechanisms, 1. INTRODUCTION as in the asteroid belt. It is possible that TNOs have Edgeworth (1949) and Kuiper (1951) were the ®rst to surfaces that are (1) composed of pure H2O ice (similar hypothesize the existence of trans-neptunian objects to Pluto's moon Charon), (2) similar to carbonaceous (TNOs). They rationalized that planetesimal formation in chondritic material, or (3) similar to that of short period the solar nebula could not have ended abruptly near the comets (frozen H2O, CO, CH3OH, CO2,NH3,CH4, HCN, orbit of Neptune and that the low density of nebular mate- and silicates) but richer in complex organic molecules as rial beyond 40 to 50 AU could have supported the forma- a result of cosmic ray and solar wind bombardment of tion of small bodies (less than a few hundred kilometers their surfaces. in diameter). Edgeworth and Kuiper suggested that the In this paper we report B, V, and R band photometry solar system beyond Neptune consisted of progressively of the TNOs 1994 TB and 1993 SC. Our measurements smaller icy bodies (comets) con®ned near the plane of the allow us to constrain the surface composition of these ecliptic (the Kuiper Belt). two TNOs. 212 0019-1035/97 $25.00 Copyright 1997 by Academic Press All rights of reproduction in any form reserved. TRANS-NEPTUNIAN OBJECTS 213 2. OBSERVATIONS magnitudes brighter than the TNO, but with peak count rates well below any possible nonlinear part of the CCD Observations were obtained during four nights (24±27 response. Standard colors and magnitudes were obtained November 1995 UT), using the Steward Observatory 2.3- for these stars using the data from the ®rst night of the m telescope on Kitt Peak. The 1200 3 800 pixel Loral± Steward run, as well as a December, 1995 Lowell run. For Fairchild CCD was used in a 2 3 2 binning mode, yielding the Steward data, an aperture of 8 arcsec diameter (12 pixel a 600 3 400 pixel image, covering 3.0 3 2.0 arcmin on radius) was used for the standard stars and the secondary the sky at 0.3 arcsec pixel21. This CCD has outstanding standard stars. Given the typical seeing (1.5 arcsec FWHM) quantum ef®ciency, particularly in the blue, where the QE this aperture encompassed essentially all the light of the exceeds 95%. This greatly aided our blue photometry of stars. Comparison of the secondary standard stars with these faint objects. The V and R ®lters provided a superb other stars in the same ®elds was used a check for any match to the standard Kron±Cousins±Landolt system variability of the secondary standard during the runÐnone V±R colors, with a transformation slope of unity and negli- was found. gible offset between instrumental and standard colors. The A completely independent check on the magnitudes of B and V ®lters yielded a transformation slope different our secondary standard stars was provided by observations form unity by only about 5%, and the overall ®t to the of the 1994 TB and 1993 SC ®elds on the night of 8 Decem- standard system was again excellent. ber 1995 UT, with the 0.76-m telescope of Lowell Observa- Observing conditions were variable. The ®rst night was tory and the Photometrics CCD of the National Under- photometric. Observations were obtained of the standard graduate Research Observatory. Observations were star ®eld in M 67 (Schild 1983) and several Landolt (1992) obtained of the Landolt standard star ®eld PG 0231 1 051. standard star ®elds (PG 0231 1 051 and PG 2213 2 006). The magnitudes derived for the secondary stars were in During the remainder of the run, there is evidence in the excellent agreement with those found on the ®rst night of data of variable thin clouds, with cloud extinction of 0.2 the November Steward run. mag at most. As discussed below, our photometry proce- The TNOs are so faint that the uncertainty in their mag- dure is not affected by these clouds, as we reference magni- nitudes is dominated by sky noise. Under such conditions, tudes to secondary standard stars whose magnitudes were it is common in stellar photometry to use a rather small determined on the ®rst night, and also during a separate aperture to perform the instrumental magnitude measure- run described below. ment, in order to minimize the sky noise, and to make a The initial CCD reductions were fairly standard, using correction to the ``total'' magnitude using observations in a combination of twilight and dark sky observations for small and large apertures of a brighter star in the same ¯at®elding. After bias subtraction and ¯at®elding, the re- ®eld. We essentially used this method to get instrumental gions around the TNO and secondary standard star in magnitudes for the TNOs. For each averaged TNO image, each image were searched visually for nearby cosmic rays. we measured the TNO with a small aperture (2.4 arcsec Cosmic rays on this chip were easy to spot, and usually diameter; 4 pixel radius) on the image with the TNO regis- affected only 3 or 4 pixels. Cosmic rays near the objects tered, and measured the secondary standard ®eld star with of interest were replaced by the average of neighboring the same aperture on the averaged frame with the stars good pixels. One image contained a cosmic ray coincident registered. This yields a magnitude difference between with the center of a TNO, and this image was discarded. TNO and secondary standard star which should be unaf- To increase the signal to noise, groups of individual 600 fected by seeing variations, tracking errors, variable clouds, sec (V and R ®lter) or 900 sec (most B ®lter images) frames or focus problems. Aperture photometry was done with taken together in time were averaged. Groups usually con- the PHOT task in the IRAF DIGIPHOT package. For all sisted of three images, but some had two or four. Because photometry, the sky value was found in an annulus from of the motion of the TNOs, two averaged images of each 20 to 33 pixels radius. The sky value was the peak of a group were produced. One averaged image had the individ- gaussian ®t to the histogram of pixel values in the sky ual frames shifted so that the stars were registered, the annulus. If necessary, the sky annulus was ®rst cleaned of other so that images of the moving TNO were registered. objects which might bias the sky measurement by replacing (Typical TNO motion was about four pixels per hour.) By them with a patch of nearby sky. averaging in this way, the motion of the TNO from frame To check on the accuracy of our quoted uncertainties to frame did not contribute to any degradation of the TNO in the V magnitudes of the TNOs, we measured the V image on the TNO registered image.