Detection of the Yarkovsky Effect on 1998 SD9 from Optical Observations

1998 SD9 is an unnumbered Aten-type asteroid, first observed at Lincoln Laboratory ETS, New Mexico, on 1998 September 18. A second apparition in 2008 resulted in a 10-year data arc and an orbit that was secure enough to predict an encounter with the Earth on 2018 August 29. Not much is known about the physical properties of 1998 SD9. It has an absolute magnitude of 24.2 mag, suggesting a diameter of 40 to 90 meters. With an orbital period of 0.59 years and an aphelion at 1.05 au, it stays close to the Sun for all of its orbit.

Its approach within 0.0108 au of the Earth on 2018 August 29 at 07:30 UT was predicted by NEODyS-2,³ who gave a 1 sigma uncertainty of about ±10 arcsec for the evening of 2018 August 29. The asteroid approached the Earth from the direction of the Sun, and it became visible on 2018 August 28, at high northern declination, rapidly moving southwards.

We observed 1998 SD9 for about 40 minutes, starting on 2018 August 29 at 23:21 UT, using the 10-inch Ritchey–Chrétien telescope at Northolt Branch Observatory, IAU code Z80. The asteroid could not be found within the 5 sigma uncertainty region (±50 arcsec) to a limiting magnitude of 18 in any of the three star fields that have been imaged. Extending the search radius to the size of the field of view (15 by 11 arcmin) yielded a clear detection in four separate stacks (varying in duration between 40 × 5 s and 86 × 5 s to account for changes in seeing conditions) at a distance of 4.5 arcmin, more than 30 standard deviations, from the predicted position, at apparent magnitude 17.2. Due to the offset, the target was missed in about 30 percent of the collected exposures.

Using the EXORB 8.1 orbit determination software (Vitagliano 1996),⁴ we confirm that the full data arc cannot be fit by a purely Keplerian orbit, but that the inclusion of the A2 parameter in the fit results in a good orbit (rms 091 using 72 observations spanning 7285 days). A value of A2 = (−2.759 ± 0.054) × 10⁻¹³ au day⁻² was found, suggesting a detection of the Yarkovsky effect.

By including data obtained by other stations up to 2018 September 1, using 137 observations spanning 7288 days, the JPL Small-Body Database gives a value of A2 = (−2.816 ± 0.059)×10⁻¹³ au day⁻², which is consistent with, but not significantly better than, our result.⁵

The small size and high irradiation make 1998 SD9 susceptible to acceleration by anisotropic emission of thermal photons, commonly referred to as the Yarkovsky effect. (Bottke et al. 2006) However, neither the 1998 nor the 2008 apparitions allowed radar observations, and the optical data were fit well by a purely Keplerian orbit. Radar observations were planned at Arecibo on 2018 August 29–30, with SNR's sufficient for high-resolution imaging,⁶ and at Goldstone on 2018 August 28–29. The Goldstone radar observations planning page for 1998 SD9 mentions the possibility to detect the Yarkovsky effect, but also that optical observations would have been difficult to obtain prior to the scheduled radar observations.⁷ The uncertainty in position may have prevented a successful radar detection. (Naidu et al. 2016) If successful, observations from Arecibo could provide a size estimate and shape model, as well as a rotation period and axis, and should improve the orbit uncertainty.

We considered outgassing as a possible alternative explanation, but disregard it as unlikely, given that the strength of the Yarkovsky effect was within the range expected for a small asteroid close to the Sun, and that neither the A1 nor the A3 non-gravitational parameters significantly differed from zero. Assuming a cometary nature, EXORB 8.1 finds A1 = (−1.95 ± 2.82) × 10⁻¹¹, A2 = (−2.47 ± 0.11) × 10⁻¹³, A3 = (+5.22 ± 3.05) × 10⁻¹¹. If cometary activity is found by future observations, then that would make 1998 SD9 the comet with the shortest orbital period, by some margin. The current record holder is 311P/PANSTARRS, with a period of 3.24 years.⁸ That would not be any less interesting than the preferred explanation (Yarkovsky effect).

Less than 100 asteroids have significant (signal-to-noise ratio larger than 3) detections of the Yarkovsky effect; of those, only about 30 have not been observed by radar. Adopting the JPL value of A2 = (−2.816 ± 0.059) × 10⁻¹³ au day⁻², this detection of the Yarkovsky effect has a signal-to-noise ratio of SNR = 48. This is the third best detection of the Yarkovsky effect to date, after (101955) Bennu with SNR = 182 and (480883) 2001 YE4 with SNR = 73, and the best ever that does not use radar observations (Table 1).