Asteroid Companions in the Visible: WFPC
2 Images
A. D. Storrs and K. Makhoul
8000 York Road
Towson, MD
21252
410-704-3003
FAX 410-704-3511
M. Gaffey and C. Wood
Univ. of N. Dakota
Space Studies
Dept., Box 9008
gaffey@space.edu
R. Landis
Jet Propulsion
Lab.
4800 Oak Grove Dr.
Pasadena, CA 91109
landis@mail2.jpl.nasa.gov
F. Vilas
NASA Headquarters
Code SE
300 E. St. NW
Washington, D.C.
20546
fvilas@hq.nasa.gov
E. Wells
3700 San Martin
Dr.
Baltimore, MD
21218
wells@stsci.edu
B. Zellner
Statesboro, Ga,
30460
13 manuscript pages, 8 figures, 4 tables
Proposed Running Head: Asteroid companions
Direct editorial correspondence to:
A. D. Storrs
8000 York Road
Towson, MD 21252
410-704-3003
FAX 410-704-3511
Abstract: We present visible and near-IR images of main-belt asteroids 45 Eugenia, 87 Sylvia, and 107 Camilla. These images show not only the primary object but also companion objects (satellites). We present reconstructions of these images and photometric information on the relative reflectance of the companions and the primaries. While the companions of 45 Eugenia and 107 Camilla are noticeably redder than the main body in the visible region, the companion of 87 Sylvia has substantially the same color. We hypothesize that a red color difference may be due to the companion having an older surface than the primary object. This implies that the companion to 87 Sylvia is much older than the other two, and/or that 45 Eugenia and 107 Camilla have been resurfaced more recently than their companions.
Keywords:
Asteroids, Composition; Satellites, General
Introduction:
Companions to asteroids moved from the theoretical to the practical with the discoveries of Dactyl, the satellite of 243 Ida (Belton et al. 1996), and Petit-Prince, the satellite of 45 Eugenia (Merline et al. 1999). While many companions had been proposed based on occultation and/or lightcurve information, these have not been confirmed by subsequent direct imaging observations (Storrs et al. 1999). Thus recent searches have concentrated on a broader sample of objects than that constrained by lightcurve or occultation observations.
Recent work on asteroidal companions has been summarized by Merline et al. (2001). By far the bulk of the detections of companions to main-belt asteroids has been through ground-based adaptive optics (AO) observations from large telescopes. This technique works best in the near infrared, however, and is difficult shortward of the J band (1.1 mm wavelength). Little or no information is available through direct imaging in the visible from ground-based observations. We report here the results of direct imaging observations of asteroidal companions, using the Wide-Field and Planetary Camera 2 (WFPC-2) of the Hubble Space Telescope (HST). This instrument has corrective optics that compensate for the error in polishing HST’s primary mirror, and offers diffraction limited images (HST’s aperture is 2.4m in diameter) over a 30 arcsec field of view from the vacuum ultraviolet out to 1 mm wavelength.
The WFPC-2 has an excellent dynamic range of over three orders of magnitude, or greater than seven and a half astronomical magnitudes, in the same image. We have made use of this capability by measuring the flux from the companions to 45 Eugenia and 87 Sylvia (discovered after the HST observations were planned but before they were executed) and to 107 Camilla (discovered by the HST observations, see below). We can measure the flux from the primary body in the same images and directly compare the two to investigate surface differences between the primaries and their companions. We also enhance the resolution of the images of the primary bodies and ratio the restored images from different wavelength filters, to see if there is any variegation in the surface of the primary bodies.
HST program 8583 is a “snapshot” survey of large main-belt asteroids. A sequence of images in the planetary camera of the WFPC-2 was designed to fit into short segments of the HST schedule that could not accommodate regular HST observations. The program used five filters, F439W, F673N, F791W, F953N, and F1042M. The letter “F” indicates that these optical elements are filters (as opposed to polarizers or grisms), the number is the central wavelength in nm, and the final letter indicates the with of the filter, W being wide-band, M medium, and N narrow. Particulars of the observation are given in Table 1. Note that the CCD sensitivity drops at long wavelengths—the most sensitive filter is the F791W so an extra exposure with this filter ended each sequence, in which the primary was deliberately saturated to enhance any faint satellite objects, increasing the effective dynamic range of detection of companions to more than 8 magnitudes. Thus the actual observing sequence was F439W, F673N, F953N, F1042M, F791W, and F791W (overexposed). Thus non-detections of companions in the longest wavelength images are always surrounded by definite detections in the shorter wavelength filters, ruling out confusion from background objects. Note that background stars are detected in simultaneous images in other chips of the WFPC-2. These stars are always trailed—the companions are definitely co-moving with the primary asteroids. Note also that the WFPC-2 filters are “wedged” to minimize the effects of multiple internal reflections. This has the result of shifting the image position in different filters, even when the spacecraft pointing remains the same. Thus the faint companion images appear at different places on the PC chip, although always at the same position in relation to the primary asteroid.
The photometric measurements are difficult and the details are discussed in the next section. The reconstruction process is discussed in Storrs et al. (2003) and only a brief discussion is included here. The results for each body are discussed next, and possible explanations for the data are given in the conclusion section.
Photometry:
Figure 1 shows the discovery sequence of images of the companion to 107 Camilla (Storrs et al., IAUC 7599). The images, stretched to show the primary body, are on the left; on the right are the same images to the same spatial scale, but brightness stretched to show the lowest levels. Note the companion to the right of the scattered light halo of the primary—it is only a few standard deviations above the noise level of the image. These images are the result of the STScI pipeline processing system, and have not been restored or enhanced other than the brightness scaling mentioned above.
To measure the relative brightnesses of the companion and primary object, we integrated the flux in concentric square apertures surrounding each object. Note that the point-spread function (PSF) of HST has spikes caused by the spider holding the secondary mirror, which run along the diagonals of the WFPC-2 images. Thus square apertures will retain more of the flux (direct and diffracted) than will circular apertures of the same area. The background flux can be determined trivially by dividing the flux difference between the outer and inner boxes by the difference in their areas. This process will remove both zero-order (constant) background as well as first-order background (e.g. scattered light from a nearby bright source). Note that the flux must be corrected for charge-transfer efficiency (CTE) (Holtzman et al. 1995)—some charge is lost in the CCD readout process. This correction is generally small for bright objects but can be large for faint objects.
The primary source of error in the photometric reduction process is the aperture correction and shot noise in the background. These opposing effects drive the choice of boxes used for photometric measurements—small (3x3 or 5x5 pixel) boxes for faint objects (to minimize shot noise), and large (11x11 or 15x15 pixels) for bright objects (to minimize the aperture correction). A finite sized aperture necessarily only measures part of the total flux, but the smaller the aperture the less sky noise there is. The fraction of the flux measured in an aperture of a given size can be determined by using the same photometric method on images of standard stars as is used on the (extended) asteroidal images. Figure 2 shows the result of reducing standard star images in various filters covering the same wavelength range as the asteroid observations. Note that the scatter for short wavelength observations is about 0.05 magnitudes (most of the PSF fits inside the small boxes) but gets larger (0.1 magnitudes) for longer wavelengths (where much of the light is outside of the aperture). To correct for the effect of a finite sized aperture, magnitudes were corrected by adding the values shown in Table 2.
The best signal-to-noise ratio (SNR) was obtained with the smallest aperture for these data. Thus the photometric information was determined by integrating the counts in a 3x3 pixel box centered on the companion, and the sky determined from a concentric 5x5 pixel box. The brightness of the primary asteroid was determined from the flux in an 11x11 pixel box, with the sky determined from a concentric 15x15 pixel box.
We restored the images (largely removing the effects of spherical aberration) using the Myopic Iterative STep Preserving ALgorithm (“MISTRAL”) routine developed by a team at the Office National d’Etudes et de Recherches Aerospatiales (ONERA). A good description is contained in Conan et al. (2000), and our application of the program is discussed in Storrs et al (2003). Parameters for the reconstruction of the various images are given in Table 3.
Results:
Figure 3 shows a summary plot for observations of 45 Eugenia and its companion Petit-Prince. The companion is clearly detected, at least at the three shortest wavelengths in the unreconstructed images, 0.83 arcsec to the left of the primary and well away from the diffraction spikes. Note that, as discussed above, these detections are at the beginning and end of the observation sequence, and that there is no apparent motion between the primary and the secondary. Also shown in this figure are the reconstructed images of the primary, showing it to be elongated, 240 km x 213 km. This is significantly larger than the SIMPS diameter of 215 km (Tedesco et al. 2002), and the axis ratio (a/b=1.13) is smaller than that given by the Planetary Data System Small Bodies Node (PDSSBN, http://pdssbn.astro.umd.edu/sbnhtml/). The albedo of 0.035 (at 440 nm wavelength) is significantly below the value of 0.0398 +/- 0.002 reported in the SIMPS.
Figure 4 shows the relative brightness of the companion to the primary at the wavelengths observed. Note that the detections at 0.9 and 1.0 mm are really upper limits—we only believe the slope in the reflectance of the secondary relative to the primary in the three shortest wavelength bands. For Petit-Prince, this slope is 0.6 magnitudes between 440 and 790 nm, much larger than the photometric error. Overall, Petit-Prince appears seven magnitudes fainter than 45 Eugenia, a magnitude fainter than reported by Merline et al. (1999, IAUC 7129).
Figure 5 shows a summary plot for observations of 87 Sylvia and its companion. The companion is marginally detected in the unreconstructed images, 0.33 arcsec to the left of the primary. These detections are at the beginning and end of the observation sequence and there is no apparent motion between the primary and the secondary. Also shown in this figure are the reconstructed images of the primary, showing it to have no significant elongation, and a diameter of 400 km. This is much larger than the SIMPS diameter of 261 km (Tedesco et al. 2002), and the albedo of 0.017 (at 440 nm wavelength) is less than half the value of 0.0435 +/- 0.005 reported in the SIMPS. The axis ratio given by the PDSSBN (a/b=1.42) is much greater than that for a symmetrical object. 87 Sylvia appears significantly larger, darker, and more symmetrical than previously reported.
Figure 6 shows the relative brightness of the companion to the primary at the wavelengths observed. Note that the detections at 0.9 and 1.0 mm are really upper limits. The companion is definitely detected in the three shortest wavelength filters and has the same color as the primary within the photometric error. Overall, the companion appears just over five magnitudes fainter than 87 Sylvia, considerably brighter than the brightness ratio of 470 (6.6 magnitudes) reported by Brown and Margot (IAUC 7588).
Figure 7 shows a summary plot for observations of 107 Camilla and its companion. The companion is marginally detected in the unreconstructed images, 0.6 arcsec to the right of the primary. Again the detections are at the beginning and end of the observation sequence and there is no apparent motion between the primary and the secondary. Also shown in this figure are the reconstructed images of the primary, showing it to be elongated, with a short axis of 256 km and a long axis of 365 km. This is considerably larger than the SIMPS diameter of 223 km (Tedesco et al. 2002), and the albedo of 0.041 (at 440 nm wavelength) is significantly less than the value of 0.0525 +/- 0.009 reported in the SIMPS. The axis ratio given by the PDSSBN (a/b=1.46) is only slightly greater than that of 1.43 observed for this object.
Figure 8 shows the relative brightness of the companion to the primary at the wavelengths observed. Note again that the detections at 0.9 and 1.0 mm are really upper limits. Nevertheless, the companion is definitely detected in the three shortest wavelength filters and has a significantly redder color (0.4 mag) over this wavelength range than the primary. Overall, the companion appears six and a half magnitudes fainter than 107 Camilla, slightly brighter than the brightness difference of seven magnitudes reported by Storrs et al. (IAUC 7599) from their preliminary analysis of these data. Note that observations of 375 Ursula were made one HST orbit preceding the observations of 107 Camilla and do not show any similar companion, indicating that this is not an instrumental artifact. Stars in the Camilla images are detectably trailed while the companion is not, indicating that the companion is not a background object. The companion is extended over several pixels in a way that cosmic ray hits in general are not. The probability of four cosmic ray hits of the same profile, brightness, and position with respect to the main asteroid (who’s position on the chip changed due to the different wedge of the different filters) is vanishingly small.
The data on the primary asteroids as well as the values in the literature are summarized in Table 4.
Conclusions:
Three asteroidal companions have definitely been detected in direct images from HST. In two of the cases ((45) 1 Petit-Prince and the companion to 107 Camilla) the companion appears significantly redder than the primary asteroid, while in the third case (the companion to 87 Sylvia) it appears to have the same color. This may be due to “space weathering” (e.g. Madey et al. 2002) of the surface of the companion: silicate surfaces exposed to solar wind and flares, as well as cosmic rays, get redder (in the visible region). If something (perhaps small impacts or downslope motion of surface dust) was “resurfacing” the primary body but not the companion, the companion would accumulate “weathering products” and appear redder than the primary. That the companion to 87 Sylvia has the same color as the primary indicates that either the surfaces are of substantially the same age, or that the “resurfacing” processes don’t have a big effect on the surface of 87 Sylvia (or have an equal effect on the companion and primary). We note that this asteroid is the most symmetric of those we observed.
The primary asteroids are all resolved in the HST images, and their shapes and sizes can be clearly seen after reconstruction with the MISTRAL algorithm. These asteroids all appear to be larger and darker than expected from data in the literature, especially that derived from IR radiometry (the SIMPS, Tedesco et al. 2002). Ratios of images taken in and out of the 0.7 mm hydration feature (Vilas et al. 1994) show little or no variation across the surfaces of these small bodies. If water of hydration is present, it is uniformly distributed on the spatial scales we can observe.
The use of HST imaging to detect companions and to constrain the ages relative to their primary objects is promising. We hope to pursue this in future HST programs.
Acknowledgments:
The Hubble Space Telescope is operated by NASA and ESA. We wish to acknowledge the support of NASA Contract NAS5-26555 and STScI grant GO-08583.01. The NASA Planetary Data System Small Bodies Node (PDSSBN) provided an invaluable summary of diverse data.
References:
Belton, M.J.S., and 20 colleagues 1996. The Discovery and Orbit of 1993 (243)1
Dactyl. Icarus 120, pp. 185-199
Conan, J.-M. T. Fusco, L.M. Mugnier, F. Marchis, C. Roddier, and F. Roddier,
2000. Deconvolution of Adaptive Optics images: from Theory to Practice. In Adaptive Optical System Technology (P. Wizinowich, Ed.), pp. 913-924, SPIE
Holtzman, J.A. 1995. The Photometric Performance and Calibration of WFPC2.
PASP 107, pp. 1065-1093
Madey, T.E., R.E. Johnson, and T.M. Orlando 2001. Far-out surface science:
radiation-induced surface processes in the solar system. Surface
Science
500, pp. 838-858
Merline, W.J., L.M. Close, C. Dumas, C.R. Chapman, F. Roddier, F. Menard,
D.C. Slater, G. Duvert, C. Shelton, and T. Morgan 1999. Discovery of a
moon orbiting the asteroid 45 Eugenia. Nature 401, pp. 565-568
Storrs, A., B. Weiss, B. Zellner, W. Burleson, R. Sichitiu, E. Wells, C. Kowal,
and D. Tholen, 1999. Imaging Observations of Asteroids with Hubble
Space Telescope. Icarus 137, 260-268
Storrs, A.D., C. Dunne, J.-M. Conan, L. Mugnier, B.P. Weiss, and B. Zellner,
2003. A closer look at main belt asteroids I: WF/PC images. Icarus
(submitted)
Tedesco, E.F., P.V. Noah, M. Noah, and S.D. Price, 2002. The Supplemental
IRAS Minor Planet Survey. Ap. J. 123, 1056-1085
Vilas, F., 1994.
A Cheaper, Faster, Better Way to Detect Water of Hydration on
Solar System Bodies. Icarus 111, pp.
456-467
Tables:
I: Observational Circumstances
|
Asteroid |
Observed (2001) |
R (AU) |
D (AU) |
Phase (deg) |
Dataset name |
|
45 Eugenia |
22 June |
2.58 |
1.67 |
12.3 |
u62h27 |
|
87 Sylvia |
23 Feb. |
3.76 |
2.80 |
4.47 |
u62h35 |
|
107 Camilla |
1 Mar. |
3.27 |
2.29 |
2.97 |
u62h39 |
II: Aperture Corrections (in magnitudes)
|
|
F439W |
F673N |
F791W |
F953N |
F1042M |
|
3x3 box |
0.3257 |
0.5796 |
0.6847 |
0.8137 |
0.8773 |
|
5x5 box |
-0.0265 |
0.1680 |
0.3688 |
0.5076 |
0.5738 |
III: MISTRAL Parameters
|
Asteroid image |
Filter |
Exp. Time (s) |
Peak Counts |
regobj |
threshold |
|
u62h2701 |
F439W |
23 |
2332 |
0.1 |
0.7 |
|
u62h2702 |
F673N |
40 |
1654 |
0.05 |
0.6 |
|
u62h2705 |
F791W |
1.6 |
855 |
0.05 |
0.5 |
|
u62h2703 |
F953N |
40 |
185 |
0.05 |
0.4 |
|
u62h2704 |
F1042M |
40 |
204 |
0.05 |
0.3 |
|
u62h3501 |
F439W |
30 |
1546 |
0.1 |
0.7 |
|
u62h3501 |
F673N |
40 |
711 |
0.05 |
0.2 |
|
u62h3505 |
F791W |
3 |
719 |
0.05 |
0.1 |
|
u62h3503 |
F953N |
40 |
83 |
0.01 |
0.4 |
|
u62h3504 |
F1042M |
40 |
94 |
0.01 |
0.3 |
|
u62h3901 |
F439W |
20 |
1794 |
0.1 |
0.7 |
|
u62h3902 |
F673N |
40 |
1278 |
0.05 |
0.6 |
|
u62h3905 |
F791W |
1.4 |
675 |
0.05 |
0.5 |
|
u62h3903 |
F953N |
40 |
126 |
0.05 |
0.4 |
|
u62h3904 |
F1042M |
40 |
1157 |
0.05 |
0.3 |
IV: Asteroid size and albedo
|
|||||||||||||||||||||||||||||||||||
1 From the NASA Planetary Data System Small Bodies Node (http://pdssbn.astro.umd.edu/sbnhtml/)
2From the Supplemental IRAS Minor Planet Survey
(Tedesco et al. 2002).
Figures:
1. Discovery images of the companion to 107 Camilla. Pipeline processed HST images are in the left column. The filter sequence is F439W, F671N, F953N, F1042M, F791W, and F791W (overexposed) from top to bottom (first to last temporally). Images in the right column are stretched to have –10 DN be black, and +20 DN to be white, save for the last which is stretched between –10 and +100 DN. The scattered light from the main asteroid dominates the images in the right column (note the diffraction spikes along the diagonals) but the companion is clearly seen 0.6 arcsec to the right (NPA 261o) of the primary in the first and last pair of images, which are the deepest exposures. Note that 107 Camilla is saturated in the last exposure. Each image is 1.47 arcsec across, and north is 13o CCW from straight up in the images.
2. Magnitude difference between multiple observations of the same standard stars, measured with a 3x3 pixel square box. These standard star images are in various filters covering the same wavelength range as the asteroid observations. This plot shows that the scatter expected from this photometric technique is less than 0.05 mag. in the blue and visible, increasing to 0.2 mag. in the near IR as would be expected as the PSF gets much larger than the photometric box.
3. Summary plot of images of 45 Eugenia in the F439W filter (center wavelength 439 nm). The left image is a pipeline processed HST image stretched to show the companion (0.83 arcsec to the left of the primary, NPA 21.6o). Each image is 1.47 arcsec across, and north is 68o CCW from straight up in the images. The right image is the primary asteroid to the same scale, after restoration with the MISTRAL algorithm.
4. Brightness of the companion to 45 Eugenia ((45) 1 Petit-Prince)
compared to the primary, in magnitudes. Note that the detections in the
near IR (953 nm and 1042 nm) are very weak, essentially upper limits. The
companion is noticeably redder than the primary, in the visible region.
5. Summary plot of images of 87 Sylvia in the F439W filter (center wavelength 439 nm). The left image is a pipeline processed HST image stretched to show the companion (0.33 arcsec to the left of the primary, NPA 306o). Each image is 1.47 arcsec across, and north is 152o CCW from straight up in the images. The right image is the primary asteroid to the same scale, after restoration with the MISTRAL algorithm. Note that the bright spot to the lower right of 87 Sylvia does not appear on the other images in the sequence, and so is not another companion.
6. Brightness of the companion to 87 Sylvia compared to the primary, in
magnitudes. Note that the detections in the near IR (953 nm and 1042 nm)
are very weak, essentially upper limits. The companion is noticeably redder
than the primary, in the visible region.
7. Summary plot of images of 107 Camilla in the F439W filter (center wavelength 439 nm). The left image is a pipeline processed HST image stretched to show the companion (0.60 arcsec to the right of the primary, NPA 261o). Each image is 1.47 arcsec across, and north is 13o CCW from straight up in the images. The right image is the primary asteroid to the same scale, after restoration with the MISTRAL algorithm.
8. Brightness of the companion to 107 Camilla compared to the primary, in
magnitudes. Note that the detections in the near IR (953 nm and 1042 nm) are very weak, essentially upper limits. The companion is noticeably redder than the primary, in the visible region.