A SEXTET OF MAIN-BELT BINARY ASTEROID CANDIDATES

Abstract

Analysis of CCD photometry observations at the Center for Solar System Studies-Palmer Divide Station (CS3-PDS) made in 2014 June-October found six main-belt binary candidates. 1355 Magoeba (Hungaria) is a possible binary, showing indications of a secondary period. 2131 Mayall is a known binary. The rotation period of the primary was confirmed. (15778) 1993 NH is a possible member of the subclass of wide binaries that are characterized by a large amplitude, long period lightcurve superimposed by a low amplitude, short period component. (18890) 2000 EV26 is a confirmed new binary, showing mutual events in the satellite’s lightcurve. Analysis of data from 2011 shows, at best, only weak evidence of the satellite. (27568) 2000 PT6 is a probable binary with strong evidence of a second period but no confirming mutual events. (30535) 2001 OR5 is a possible binary that shows a secondary period but it may be a harmonic artifact of Fourier analysis.

CCD photometric observations of six main-belt asteroids made at the Center for Solar System Studies-Palmer Divide Station (CS3-PDS) in 2014 June through October showed them to be binary candidates. Table I lists the telescope/CCD camera combinations used for the observations. The cameras all use KAF blue-enhanced CCD chips and so have essentially the same response. The pixel scales for the combinations range from 1.24-1.60 arcsec/pixel.

Table I.

List of CS3-PDS telescope/CCD camera combinations.

Desig	Telescope	Camera
Squirt	0.30-m f/6.3 Schmidt-Cass	ML-1001E
Borealis	0.35-m f/9.1 Schmidt-Cass	FLI-1001E
Eclipticalis	0.35-m f/9.1 Schmidt-Cass	STL-1001E
Australius	0.35-m f/9.1 Schmidt-Cass	STL-1001E
Zephyr	0.50-m f/8.1 R-C	FLI-1001E

All observations were unfiltered since a clear filter can result in a 0.1-0.3 magnitude loss. The exposure duration depended on the asteroid’s brightness and sky motion. Guiding was on a field star, which sometimes caused the asteroid image to trail.

Measurements were done using MPO Canopus. If necessary, an elliptical aperture with the long axis parallel to the asteroid’s path was used. The Comp Star Selector utility in MPO Canopus found up to five comparison stars of near solar-color for differential photometry. Catalog magnitudes were usually taken from the MPOSC3 catalog, which is based on the 2MASS catalog (http://www.ipac.caltech.edu/2mass) but with magnitudes converted from J-K to BVRI using formulae developed by Warner (2007). When possible, magnitudes are taken from the APASS catalog (Henden et al., 2009) since these are derived directly from reductions based on Landolt standard fields. Using either catalog, the nightly zero points have been found to be consistent to about ± 0.05 mag or better, but on occasion are as large as 0.1 mag. This consistency is critical to analysis of long period and/or tumbling asteroids. Period analysis is also done using MPO Canopus, which implements the FALC algorithm developed by Harris (Harris et al., 1989).

In the plots below, the “Reduced Magnitude” is Johnson V as indicated in the Y-axis title. These are values that have been converted from sky magnitudes 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 given phase angle, e.g., alpha(6.5°), using G = 0.15, unless otherwise stated. The X-axis is the rotational phase, ranging from −0.05 to 1.05.

For the sake of brevity, only some of the previously reported results may be referenced in the discussions on specific asteroids. For a more complete listing, the reader is directed to the asteroid lightcurve database (LCDB; Warner et al., 2009). The on-line version at http://www.minorplanet.info/lightcurvedatabase.html allows direct queries that can be filtered a number of ways and the results saved to a text file. A set of text files of the main LCDB tables, including the references with bibcodes, is also available for download. Readers are strongly encouraged to obtain, when possible, the original references listed in the LCDB for their work.

Individual Results

1355 Magoeba.

The 2014 apparition was the fifth one observed by the author. Previous results have varied considerably, e.g., 31.65 h (Warner, 2010), 5.99 h (Warner, 2011), and 2.975 h (Warner, 2013) with a possible solution at 5.95 h. Analysis of the 2014 data, which indicated two periods in the data, may help explain some of the disparate solutions.

Figure 1 shows the unsubtracted data from 2014 after a period search covering a range of 2-17 hours. The small deviations at the minimum near 0.2 rotational phase prompted a search for a second period.

Figure 1.

The data from the 2014 apparition for 1355 Magoeba are plotted without subtracting any effects due a possible second period. The period is close to the final result for P1

The dual period search feature in MPO Canopus was used to determine if there was a second period present in the data. The process started by finding a best fit period of all data without subtraction that was restricted to a range of 2-7 hours, the range of the more likely previous results. The resulting Fourier model curve was subtracted from the data when conducting a search for a second period, this time with a range of 2 to 25 hours. This covers typical rotation periods of a fully-asynchronous satellite or the orbital period and rotation period of a tidally-locked satellite. The resulting Fourier model was subtracted from the data for another search of the short period. The process of going back-and-forth was repeated until both periods stabilized.

The result was a short period of 2.9712 h (Figure 2), which shows somewhat less scatter overall compared to Figure 1 as well as near the minimum at 0.2 rotation phase. Figure 2 shows the lightcurve for the second period, 15.05 hours. There are no obvious signs of mutual events, i.e., occultations and/or eclipses, but the lightcurve period is in the proper range for a satellite (see Pravec et al., 2010) and can be reasonably interpreted to be due to an elongated satellite that is tidally-locked to its orbital period. Given the lack of mutual events and the asymmetrical shape of the second period lightcurve, this is listed as a possible instead of probable binary.

Figure 2.

The short period for 1355 after subtracting a second period of 15.05 hours.

2131 Mayall.

This Hungaria is likely a member of the orbital group but not the collisional family. Observations by Masiero et al. (2011) show an albedo of about 0.2, which is more typical of a type S asteroid. The collisional family members are type E. The asteroid was discovered to be a binary in 2009 (Warner et al., 2010) fulfilling the adage of “the third time is a charm” since observations in 2005 (Warner, 2005) and 2006 (Warner et al., 2007a) did not show any signs of a satellite, even with additional analysis after the binary discovery.

Figure 4 shows the data for the asteroid after a preliminary search for a short period in the range of 2-17 hours. Either the data were very bad or there is a second period involved. Figure 5 shows the lightcurve for the primary after the dual period search. To find this, the period for the satellite was assumed to be within a small range around the one found in 2010, i.e., 23.48 hours. Since this is nearly commensurate with an Earth day, no one station could hope to cover the entire secondary lightcurve and so it was expected that there would be large gaps in the long period lightcurve.

Figure 4.

The data for 2131 after the initial search for a short period. The result is chaotic at best.

Figure 5.

The lightcurve for the primary body of 2131 Mayall.

Since data from another location were not available, the best that could be done was to subtract the result of the search, despite the Fourier model having very large gyrations, when searching for the short period. Once this was found, the search for the second period was “refined”. The lightcurve appears to show a deep event near rotation phase 0, but that was not confirmed. Were it not known that this is a binary system, the best that could be said based on the 2014 data is that Mayall is a probable binary.

(15778) 1993 NH.

In the past few years, there has been growing evidence for a subset of binary asteroids known as wide binaries. These are discussed in detail by Jacabson and Scheeres (2011). The photometric observations of such objects show a lightcurve with a long period (P > 100 hours) and relatively large amplitude (Figure 8) superimposed by a short period, low amplitude lightcurve (Figure 7).

Figure 8.

The long period component for 1993 NH. At this scale, the low amplitude, short period lightcurve cannot be seen.

Figure 7.

The short period lightcurve for 1993 NH.

Since the long period lightcurve dominates the data, the search for its period is done first. Once an approximate period is found, often using only a 2^nd order Fourier fit, the resulting model lightcurve is subtracted in the search for the short period. Quite often, that lightcurve has a low amplitude and so it is difficult to determine if the Fourier analysis is finding a true period or locking onto noise. In this case, it’s believed that the short period is real, especially since it is appropriate for a fully-asynchronous satellite, i.e., one not tidally-locked to its orbit. Still, this is listed only as a possible binary.

The long period component is not due to the satellite but to the primary, which has a slow rotation period due to conservation of rotational energy. The chances for seeing mutual events and so confirming the orbital period of the satellite, which will be very long, are exceedingly remote.

(18890) 2000 EV26.

This Hungaria was first reported to have a period of 10.53 h (Warner, 2012). Analysis of the 2014 data indicated that this is a confirmed binary system with the secondary period showing mutual events (occultations and/or eclipses) that allow determining the orbital period.

Figure 9 shows the 2014 data set when finding a period with no more than a bimodal lightcurve. Searching out to 15 hours found solutions of up to 6 extrema pairs. The deviations from the main curve prompted a second period search.

Figure 9.

The unsubtracted lightcurve of 2000 EV2 after an initial search for a period shows obvious signs of a second period.

Figures 10 and 11 give the final results of the dual period search. The former shows the rotation due to the primary. The low amplitude and symmetrical shape indicate a nearly spheroidal body, which is typical among small binary asteroids. Figure 11 shows the lightcurve due to the satellite with the mutual events at 0.4 and 0.9 rotation phase. The latter event allows giving an effective diameter ratio of the satellite to the primary as Ds/Dp ≥ 0.27 ± 0.02. Since the secondary event is not total (it is not flat at the minimum point), this makes the estimate a minimum and so the satellite could be larger.

Figure 10.

The lightcurve for 2000 EV26 after subtracting the effects of the satellite. The scatter is considerably less than in Figure 9.

Figure 11.

The lightcurve for P2 clearly shows the mutual events and defines the orbital period of the proposed satellite.

Given the findings from 2014, the data from 2011 were re-examined to see if they also showed signs of the satellite. Unfortunately, that data set had considerably fewer data points per night and, because a satellite was not suspected at that time, it did not have as many nights. This reinforces the advice that when preliminary results for an asteroid show a period of 2-5 hours and – in most, but not all cases – a low amplitude, some extra observation time should be allowed to make sure that evidence of a satellite is not overlooked.

Figure 12 shows the 2011 lightcurve for the shorter period after a dual period search where the solutions were forced near those found for 2014. The fit in Figure 12 is within reason and so a new period of 3.882 h has been adopted for the 2011 data. On the other hand, given the sparse data, the solution for the satellite is far from conclusive. If the search were being done without knowledge of the 2014 results, it’s highly unlikely that this would be accepted as evidence for even a possible satellite. At best, subtracting the P2 model lightcurve did improve the fit of the P1 lightcurve, but that could also be due to a fit to random noise.

Figure 12.

The 2011 data for 2000 EV2 were searched for evidence of the satellite found in 2014. This plot shows the primary lightcurve after forcing the P1 and P2 values near those found in 2014.

(27568) 2000 PT6.

The author first observed this asteroid in 2011 (Warner, 2012) when a period of 3.624 h was reported. Observations in 2013 (Warner and Stephens, 2013) with five sessions from March 11-19 found a period of 3.4885 h and indicated the possibility of a satellite with an orbital period of 16.356 h. The 2011 data were re-examined but no indications of a satellite were found. The asteroid was observed again in 2014 August and September with the aim of confirming, if possible, the satellite and the orbital period.

The initial search without subtracting the effects due a satellite found a period of 3.499 h. Figure 14 shows that result. The dual period search feature of MPO Canopus found a primary period, also of 3.499 hours (Figure 15). This is good agreement with 2013 results. That is where the similarities ended.

Figure 14.

The unsubtracted lightcurve of 2000 PT6.

Figure 15.

The lightcurve of 2000 PT6 after subtracting the effects of a second period. This is much cleaner than in Figure 14.

The 2014 data set that consisted of 10 individual nights from August 19 through September 6 led to a secondary period of 11.73 h, or almost a 2:3 ratio of the longer period found in 2013. This might indicate that a rotational alias (i.e., uncertainty about the number of rotations over a given period) had been found with one set or the other since both answers cannot be right. The dual period search was applied to the 2013 and 2014 data sets but forcing the search for P2 for one data set to be near the value from the other data set, e.g., the search for P2 using the 2014 data was limited to the range of 15-17 hours. For the 2013 data, the closest fit was to 12.2 h with a gap of about 0.3 rotation phase. The 2014 data could be forced to a period of 15.73 hours and a bimodal lightcurve. Both of these follow-up solutions are significantly off the original periods.

An explanation for the different results may be due to what’s called a fit by exclusion. This is where the Fourier analysis finds a false best-fit period by minimizing the number of overlapping data points. This result is often seen as a gap in the lightcurve. Such a gap was in the 2013 data when fit to its original period of 16.356 hours. Even though there were parts of the lightcurve that were covered more than once, a fit by exclusion may explain why the latest data do not support that result.

This brings up another argument for the 2014 results: the 2014 data set covered more nights over a longer period of time. Since the two suspect periods are nearly commensurate with an Earth day, it would be difficult to avoid finding a rotational alias. If the period is just enough off from an integral ratio with a 24-hour cycle and the range of observations eventually allows seeing more of the lightcurve, the number of rotational aliases can be reduced. The 2014 data may have succeeded in that effort while the 2013 data did not.

It is very likely that 2000 PT6 is a binary asteroid, i.e., it is listed as probable, since subtracting P2 significantly improves the P1 lightcurve. However, the question of the orbital period of the satellite remains very much in question and so follow-up observations are encouraged at future apparitions.

(30535) 2001 OR5.

There were no previous results in the LCDB against which to compare the results from the 2014 observations. The difference between a lightcurve with no subtraction and one with the eventual second period removed is very small.

Figure 17 shows the subtracted lightcurve for 2001 OR5. Even without subtraction, the solution is unique and so not in question. The dual period process described previously eventually found a second period of 13.27 hours (Figure 18) with an irregular lightcurve that wouldn’t seem to have a reasonable physical counterpart. A telling clue that this may be a false secondary period is that P2:P1 is almost an integral ratio, i.e., 9:2. This is the result of what is sometimes called a harmonic alias, which is where the Fourier analysis locks onto higher orders of noise in a lightcurve when a suspect period is subtracted from the data. The true difference is 8.937:2.0, but this is close enough so that when combined with the unconvincing shape of the P2 lightcurve, the best that can be said at this point is this is a possible binary and so additional observations of higher-precision are encouraged.

Figure 17.

The lightcurve for 2001 OR5 after subtracting the effects of a secondary period is almost the same as with no subtraction.

Figure 18.

The lightcurve for P2 is marginally convincing, if that.

Figure 3.

The lightcurve for the second period of 15.05 h does not show any obvious mutual events but may indicate a satellite with a rotation period locked to its orbital period.

Figure 6.

The lightcurve for 2131 Mayall after subtracting the effects of the primary’s rotation. This is hardly a conclusive result but subtracting the Fourier curve produces a good solution for the primary.

Figure 13.

The P2 lightcurve from the 2011 data shows what may be a hint of events at 0.1 and 0.6 rotation phase. With no follow-up to support that claim, this could easily be attributed to “bad data.”

Figure 16.

The adopted period for P2. This suggests a tidally-locked satellite with an orbital period of 11.73 h.

Table II.

Observing circumstances.

Number	Name	2014 mm/dd	Pts	Phase	LPAB	BPAB	Period	P.E.	Amp	A.E.
1355	Magoeba	09/02-09/17	569	30.4,27.8	42	9	2.9712	0.0003	0.09	0.01
2131	Mayall	07/23-07/31	248	37.2,37.0	8	30	2.5907	0.0005	0.07	0.01
15778	1993 NH	08/15-08/28	235	34.5,36.9	282	26	3.32	0.001	0.04	0.01
18890	2000 EV26	07/23-08/02	297	22.0,18.9	330	23	3.8216	0.0005	0.11	0.01
18890	2000 EV26	12 06/28-07/04	73	20.1,20.9	264	32	3.882	0.003	0.09	0.02
27568	2000 PT6	08/19-09/06	403	24.9,17.8	10	−3	3.499	0.001	0.11	0.02
30535	2001 OR5	06/18-07/03	323	23.6,25.2	255	34	2.9697	0.0002	0.12	0.01

¹²Dates are in 2012. The period and amplitude are for the primary of the binary system. The phase angle (α) is given at the start and end of each date range, unless it reached a minimum, which is then the second of three values. If a single value is given, the phase angle did not change significantly and the average value is given. L_PAB and B_PAB are each the average phase angle bisector longitude and latitude, unless two values are given (first/last date in range).