A TRIO OF BINARY ASTEROIDS

Abstract

CCD observations of three asteroids were made at the Center for Solar System Studies in mid-2014. The Hungaria member 1727 Mette is a known binary system. We saw no evidence of the satellite in 2014, which – assuming the satellite’s orbit is nearly in the primary’s equatorial rotational plane – leads to an estimate for the primary’s spin axis orientation. The near-Earth asteroid (NEA) (68063) 2000 YJ66 has a primary spin period just at the spin barrier of 2.2 hours. Another NEA, 2014 PL51, looks to be one of a small number of wide binary candidates, which are characterized by a large amplitude lightcurve with a period of hundreds of hours superimposed by a short period, low amplitude lightcurve.

The near-Earth asteroid (NEA) program at the Center for Solar System Studies, located in Landers, CA, is primarily dedicated to finding the rotation period of as many NEAs as possible. This includes checking for and confirming satellites and non-principal axis rotation (NPAR, or tumbling). The results are added to the growing number of rotational statistics for NEAs, which are important to understanding NEA evolution, overall characteristics, and – to some degree – estimation of impact hazards. The program also monitors members of the Hungarias. These inner main-belt asteroids, which are not subject to planetary tidal encounters, serve as a control group to compare and contrast their rotation, binary, and tumbling statistics against those for the NEAs.

During the period of 2013 May 25 through 2014 October 3, the CS3 NEA program determined 190 rotation periods for 181 NEAs. Of these, 15 were binaries (rated as possible, probable, or confirmed) and 13 suspected or confirmed tumblers.

As part of the NEA program, we observed three asteroids: one Hungaria, 1727 Mette, and two NEAs, (68063) 2000 YJ66 and 2014 PL51. Mette was a known binary. To our knowledge, the two NEAs had not been observed before, and so their binary status was not known. Table I lists the equipment that was used by the authors for this study. Table II gives the dates of observations.

Table I.

List of observers and equipment.

Observer	Telescope	Camera
Warner	0.50-m f/8.1 Ritchey-Chretien 0.35-m f/9.6 Schmidt-Cass	FLI-1001E STL-1001E
Stephens	0.35-m f/10 Schmidt-Cass 0.40-m f/10 Schmidt-Cass	STL-1001E FLI-1001E

Table II.

Dates of observations for each asteroid. Observations were not necessarily made on every date within a range.

Asteroid	Obs	Date (2014 mmm/dd)
1727 Mette	RDS	Jul 19–22
(68063) 2000 YJ66	BDW RDS	Aug 18–21 Aug 28–31; Sep 01–06
2014 PL51	BDW RDS	Aug 31; Sep 02–06, 24 Sep 15–18

Image processing and measurement were done using MPO Canopus (Bdw Publishing). Master flats and darks were applied to the science frames prior to measurements. The MPO Canopus export data sets were collected by Warner for period analysis, also in MPO Canopus, which incorporates the FALC Fourier analysis algorithm developed by Harris (Harris et al., 1989).

Conversion to an internal standard system with approximately ±0.05 mag zero point precision was accomplished using the Comp Star Selector in MPO Canopus and the MPOSC3 catalog provided with that software. The magnitudes in the MPOSC3 are based on the 2MASS catalog converted to the BVRcIc system using formulae developed by Warner (2007). This internal calibration works well within a given telescope-camera system. However, in some cases, there is a nearly constant offset from one system to the next, as in the case of merging the Stephens data with Warner’s. For (68063) 2000 YJ66, an offset of approximately 0.2 magnitudes was applied to the Stephens sessions.

In the lightcurve plots presented below, the “Reduced Magnitude” is Johnson V corrected to unity distance by applying –5*log (rΔ) to the measured sky magnitudes with r and Δ being, respectively, the Sun-asteroid and Earth-asteroid distances in AU. The magnitudes were normalized to the phase angle given in parentheses, e.g., alpha(68.5°), using G = 0.15. The horizontal axis is the rotation phase, ranging from –0.05 to 1.05.

1727 Mette.

This asteroid is a member of the Hungaria orbital group, not of the collisional family. Tholen (1984) listed it as a type S asteroid; members of the Hungaria collisional family are type E. Its rotation period was well established (see references in the asteroid lightcurve database; LCDB; Warner et al., 2009) before being discovered as a binary (Warner and Stephens, 2013). The 2014 observations were made as follow-up to confirm, if possible, the satellite’s lightcurve and orbital period.

Figure 1.

The lightcurve for 1727 Mette showed no obvious signs of the satellite discovered by the authors in 2013.

Analysis of the observations by Stephens from 2014 Jul 19–22 yielded a period of 2.9808 ± 0.0002 h and lightcurve amplitude of 0.31 ± 0.02 mag. The period is in excellent agreement with earlier results. The total time of observations was about 16 hours (4 hours per night). This should have been enough to capture at least one event in the 21-hour orbital period. However, there were no obvious signs of deviations on the order of 0.05 mag, the level seen in 2013. Events of much less deviation would have been too small to observe reliably.

The satellite discovery observations were made in 2013 January, when the phase angle was about 10° and the phase angle bisector longitude (L_PAB) was 114° (see the appendix in Harris et al., 1984, for the derivation of the phase angle bisector). The amplitude of the primary was 0.33 mag. Over the years, the amplitude has stayed within the range of 0.22–0.35 mag (Warner et al., 2009). In 2014, the phase angle was 17° and L_PAB was 322° (the opposing angle being 142°). This is about 30° from the longitude in 2013. Usually, this is not enough of a difference to go from seeing to not seeing events. An exception would be if the 2013 observations were just within the boundaries for event seasons.

Assuming that the satellite’s orbital plane is near the primary’s equatorial rotational plane, the lack of events in 2014 would imply that the primary’s spin axis ecliptic longitude is near 320°, or 140°. This is supported by the fact that lower amplitudes are expected when viewing the asteroid more pole-on and that the lowest amplitudes in the LCDB were made near a L_PAB of 140°. Observations at future apparitions are encouraged to help model the binary system and primary spin axis.

(68063) 2000 YJ66.

By general definition, an NEA is an asteroid with a perihelion distance of q < 1.3 AU. As such, 2000 YJ66 is just within the group with a q ~ 1.27 AU. Warner did the initial observations in the middle of 2014 August. Because of observatory structural limitations, the asteroid moved too far north to continue with the available telescopes. Stephens continued the observations, starting in late August and going into early September.

A plot (Fig. 2) using all data phased to an initial best fit in a Fourier analysis shows a large amount of scatter. However, the scatter within the individual data sets was much less, indicating the possibility of a second period. In MPO Canopus, the search for a second period can be done two ways. The one we chose was to find a period using all data. The resulting Fourier model curve was subtracted from the data in a second search. Usually, we start on a likely short period, e.g., between 1 and 6 hours, then search for a longer period. In this case, we tried finding the long period first as well as the short to see if it would change the results. It did not.

Figure 2.

A plot using the full data set for 2000 YJ66 showed a fair amount of scatter that might be periodic instead of random. This prompted a dual period search.

When and if a second period is found, that Fourier curve is subtracted from the data and a new search for the short period is run. This iterative process continues until both periods stabilize. There is a bit of self-fulfilling prophecy in this method since each solution depends on the other period being correct. It has happened where, after a more extensive data set was obtained and previous periods put aside, an entirely different set of periods was found. In almost all cases, it was the long period that changed the most dramatically.

Figure 3.

The period spectrum for the shorter period slightly favors a solution of 2.11 hours. The sequence of minima represents successive differences of half a rotation over a 24-hour span.

Figure 4.

The lightcurve for the primary in the 2000 YJ66 binary system, obtained after subtracting the effects of the satellite.

The primary’s period of 2.1102 h is just at or above the so-called spin barrier of about 2.2 hours, which separates rubble pile and strength-bound bodies. The fact that the there is a satellite indicates that this is almost certainly a rubble pile asteroid that was spun-up by the YORP effect (Yarkovsky–O’Keefe–Radzievskii–Paddack; Rubincam, 2000) to the point where it shed mass that formed the satellite.

We did try forcing the short period to the other periods on either side of 2.11 hours. Even after subtracting the effects of the satellite, the scatter in the primary lightcurve made it difficult to say that any one of the solutions was unique and so a “more ordinary” value of about 2.2 hours is almost as likely. It is important to note that regardless of which short period we used, the long period solution did not change by more than 0.02 hours.

Figure 5.

The lightcurve attributed to the tidally-locked satellite of 2000 YJ66.

The secondary lightcurve for 2000 YJ66 does not show any mutual events, i.e., occultations or eclipses, which are considered direct evidence of a satellite. It does show a lightcurve of 0.12 mag amplitude with a period of 15.69 ± 0.02 hours. The period is within the expected boundaries given the primary period (Pravec et al., 2010). The shape of the curve fits one for an elongated satellite that is tidally-locked to the orbital period.

2014 PL51.

Jacobson and Scheeres (2011) have examined the possibilities for different types of binary asteroid systems. Among them is what are sometimes called wide binaries. In this case, the primary has a long period (due to conservation of energy) while the satellite has a short period and is well-removed from the primary. Unless the rarest of circumstances prevail, the period of the orbit cannot be determined from the observations due to the absence of mutual events, but it is probably long based on the asynchronous rotation of the two components.

In such a system, the primary has a lightcurve with a period of hundreds of hours and an amplitude A ≥ 0.3 mag. The satellite’s lightcurve amplitude is on the order of A ≤ 0.10 mag with a period in the range 2–8 hours. In addition to 2014 PL51, there are seven other such binary candidates (Table III).

Table III.

List of suspected wide binary asteroids. P1 is the primary’s period (hours). P2 is the satellite’s, which is not tidally-locked to its orbital period. All references are Warner (et al.) except for 1220 Crocus, which is Binzel and was only recently added to the list.

Number	Name	P1	P2	Ref
1220	Crocus	785	7.97	Icarus 72, 135–208
8026	Johnmckay	372	2.2981	MPB 38, 33–36
15778	1993 NH	113	3.320	MPB, this issue
67175	2000 BA19	275	2.7157	MPB 40, 36–42
119744	2001 YN42	624	7.24	MPB 41, 102–112
190208	2006 AQ	182	2.621	MPB, this issue
218144	2002 RL66	588	2.49	MPB 37, 109–111
	2014 PL51	205	5.384	This work

In most of these, the amplitude of the secondary lightcurve has been weak in comparison to the noise in the data, and so the results were inconclusive. The results for 2014 PL51 make one of the stronger cases to date.

The raw plot of the data for 2014 PL51 shows a very strong, long period component. Because of the possibility for a wide binary, a search is always made for a secondary period by first eliminating the long period, but with caution to avoid forcing a solution. The period search process described above led to finding a period of 205 ± 5 hours for the primary. The precision is somewhat arbitrary given the gaps in the coverage of the lightcurve. Just as much so, when the data set extends for a month or so, there is the chance that the lightcurve evolves due to changing phase angle or phase angle bisector longitude and latitude.

Figure 6.

The raw plot of the data obtained for 2014 PL51 clearly shows a long period component.

Figure 7.

The long period lightcurve for 2014 PL51. The short period component is not obvious at this point.

Figure 8.

The short period lightcurve is attributed to a satellite that is not tidally-locked to a very long orbital period.

The search for a secondary period found 5.384 h with an amplitude of about 0.09 mag. The shape is somewhat asymmetrical, which is not unexpected for a low amplitude lightcurve. More important is that the minimums cannot be attributed to mutual events since this would indicate that the orbital period is also 5.4 hours. This is far too short if typical densities are presumed for the two bodies.