THE HIGH ANGULAR RESOLUTION MULTIPLICITY OF MASSIVE STARS

Massive stars appear to love company. There is growing evidence that the incidence of binary and multiple stars among the massive O- and B-type stars is much larger than that for solar-type stars (see Zinnecker & Yorke 2007 and references therein). This difference in multiplicity properties may ultimately reflect differences in the star-formation process between massive and low-mass stars. For example, while low-mass stars may lose angular momentum by magnetic- and disk-related processes, it may be that these are ineffective in massive star formation because of the very short timescale of formation. Instead, the initial angular momentum of the natal cloud may end up (through a variety of processes) in the orbital angular momentum of binaries among the more massive stars (Bate et al. 2002; Zinnecker & Yorke 2007; Gies 2007).

The observational evidence for the high incidence of binaries among the massive stars comes from spectroscopic investigations of short-period systems and high angular resolution measurements of longer period (and wide) binaries. We made one of the most comprehensive surveys of the bright, Galactic O-type stars in a speckle interferometric study made in 1994 with the NOAO 4 m telescopes in both the northern and southern hemispheres (Mason et al. 1998). This investigation considered both speckle measurements and published data on radial velocity measurements to determine the overall binary properties among stars in clusters and associations, field O-stars, and runaway O-stars. The results indicated a much higher incidence of binaries among O-stars in clusters and associations, and we suggested that the true binary frequency may reach 100% among cluster stars once account is made for the observational bias against detection of binaries with periods larger than those found spectroscopically but smaller than those found through high angular resolution measurements. This work was complemented by similar speckle interferometric surveys of Wolf-Rayet stars (Hartkopf et al. 1999) and Be stars (Mason et al. 1997b).

Ten years later (and armed with an improved detector) we decided it was an opportune time for follow up and expanded speckle observations. A second epoch survey is desirable for a number of reasons. Some systems observed in 1994 may have been situated in orbital phases of close separation, and hence were unresolved. Since the systems detectable by speckle correspond to periods of decades for massive stars, it is important to repeat the survey after a similar time span. Furthermore, there are a significant number of specific systems where new observations are particularly important. For example, there are several cases where a triple is indicated by spectroscopy, but we have yet to resolve the wide system (e.g., δ Cir; Penny et al. 2001). The placement of many of the very hot, O2 and O3 stars in the Hertzsprung–Russell diagram suggests that they are very massive because they are so bright, but sometimes this extreme luminosity is instead due to the presence of a companion (Nelan et al. 2004; Niemela & Gamen 2005; Maíz-Apellániz et al. 2007). The massive binaries in the Orion Trapezium detected in the near-IR by Schertl et al. (2003) have separations that are within the resolution limit of a 4 m telescope, and detection or not of these companions at another wavelength can help set limits on the magnitude difference Δm, the color, and hence object type. For systems with two speckle measurements, a third one may allow the motion to be recognized as either linear or nonlinear (i.e., Keplerian), indicating whether the pair is optical or physical. This is extremely important in the case of ι Ori, where dynamical analysis (Gualandris et al. 2004) of this complex runaway system virtually requires that the speckle companion (first reported in Mason et al. 1998 at only 011 separation) be optical rather than physical. Finally, such high angular resolution measurements can provide direct astrometric orbits (for the nearby systems) and hence mass measurements for binaries that are clearly noninteracting (Vanbeveren et al. 1998). These provide fundamental data on the masses and other properties of the most massive stars.

For all these reasons, we embarked on a new survey of speckle interferometry measurements of the massive stars that were mainly selected from the Galactic O-star Catalog (Maíz-Apellániz et al. 2004). We describe the observational program in Section 2 and outline the main tabular results in Section 3. We use these results to reassess the binary properties of the O-stars in Section 4, and then we discuss the results for specific targets in Section 5. The observational program included a significant number of other, less-massive stars, and these measurements and several updated astrometric orbits are given in Appendices A and B, respectively.

The instrument used for most of these observations was the USNO speckle interferometer, described most recently by Hartkopf et al. (2008). Three different filters were selected, all having approximately the same central wavelength but with different full width at half-maximum (FWHM) band passes. Of these, two are standard filters (Strömgren y, 550 ± 24 nm, and Johnson V, 545 ± 85 nm). An intermediate filter, designated USNO green (560 ± 45 nm), was also used. While the Johnson V allows the camera to observe much fainter targets, the resolution limit is degraded to about 005. Both of the other filters reached the goal resolution limit of 003. We selected a filter for each target with a bandwidth suitable to the magnitude of the star and which allowed us to detect an adequate number of speckles. These resolution limit values are most significant when no companion was detected. Instances when the wider Johnson filter was used are indicated with a note to these tables.

Observations of northern hemisphere objects were obtained with the Kitt Peak National Observatory (KPNO) 4 m Mayall reflector during the period 2005 November 8–13; southern hemisphere pairs were observed at the Cerro Tololo Inter-American Observatory (CTIO) 4 m Blanco reflector during the period 2006 March 9–13. Atmospheric conditions during both runs were exceptional, with excellent transparency and significant periods of subarcsecond seeing with both telescopes, especially at Cerro Tololo. On these two runs, 1876 observations were obtained, resulting in 652 measures of double stars and 1050 high-quality observations where a pair was definitively not seen. The remaining observations were of insufficient quality for a definitive measure. Additional observations of massive stars were obtained during other 4 m observing runs as listed below.

Calibration of the KPNO data was determined through the use of a double-slit mask placed over the "stove pipe" of the 4 m telescope during observations of a bright known-single star (as described in Hartkopf et al. 2000). This application of the well-known experiment of Young allowed the determination of scale and position angle zero point without relying on binaries themselves to determine calibration parameters. Multiple observations through the slit mask (during five separate KPNO runs from 2001 to 2008) yielded mean errors of 011 in the position angle zero point and 0.165% in the scale error. These "internal errors" are undoubtedly underestimates of the true errors of these observations. Plate scales for the five Kitt Peak runs, 2001 January, 2001 July, 2005 November, 2007 August, and 2008 June, were found to be 0.01257, 0.01282, 0.01095, 0.01090, and 0.01096 arcseconds pixel⁻¹, respectively. While the camera remained the same for all five runs, the latter three were obtained with a newer computer and frame grabber and a different set of microscope objectives. The effective field of view for the detection of binaries is 15 for nominal conditions and 30 when the targets are fainter and a lower microscope objective is used with the Johnson V filter. Wider, easily detected pairs can be accommodated with a larger 60 field of view with a low-power microscope objective and 2 × 2 pixel averaging.

Since the slit-mask option was not available on the CTIO 4 m telescope, we calibrated the southern hemisphere data using observations of numerous well-observed, wide, and equatorially located binaries that we observed with both the KPNO and CTIO telescopes. Published orbital elements for these pairs were updated as needed using the recent KPNO measures, then predicted ρ and θ values from those orbits deemed of sufficiently high quality were used to determine the CTIO scale and position angle zero point. The calibration errors for these southern observations were (not surprisingly) considerably higher than those achieved using the slit mask. Mean errors for three CTIO runs from 2001 to 2006 were 067 in position angle and 1.44% in scale. Plate scales for the three Cerro Tololo runs, 2001 January, 2001 July, and 2006 March, were 0.01262, 0.01253, and 0.01084 arcseconds pixel⁻¹, respectively. The differences are attributable to changes in equipment as described above. The field of view was comparable for the southern and northern observations.

Speckle interferometry is a technique which is very sensitive to changes in observing conditions, particularly coherence length (ρ₀) and time (τ₀). These are typically manifested as a degradation of detection capability close to the resolution limit or at larger magnitude differences. To ensure we are reaching our desired detection thresholds, a variety of systems with well-determined morphologies and magnitude differences were observed throughout every observing night. In all cases, the observations of these test objects indicated that our measurements met or exceeded these thresholds, as indicated in Figure 1.

Zoom In Zoom Out Reset image size

The target list consists of the original sample of O-stars from Mason et al. (1998), additional O-stars from the catalog of Maíz-Apellániz et al. (2004), WR stars, and B-stars. The B-star sample includes candidates for orbit and mass determination, Pleiades cluster members observed previously (Mason et al. 1993a), and Be stars (Mason et al. 1997b). A number of low-mass targets were also observed that are discussed in Appendix A.

Table 1 presents coordinates and magnitude information from CDS⁵ for all those binaries which are resolved or measured for the first time. Column 1 gives the coordinates of the primary of the pair. Column 2 lists the discoverer designation number (with WSI = Washington Speckle Interferometry), and Column 3 gives an alternative designation. Column 4 provides the spectral classification, and Column 5 the combined visual magnitude. Finally, Column 6 refers to notes below the table.

Table 1. Newly Resolved Pairs

Coordinates α, δ (2000)	Discoverer Designation	Other Designation	Spectral Classification	VAB (mag)	Notes
031959.27+653908.3	WSI 51 Aa,Ab	HD 20336	B2.5 Vne	4.73
034716.57+240742.3	WSI 52 Da,Db	HD 23608	F3 V	8.72	1
042837.00+191049.6	WSI 53 Aa,Ab	Tau	G9.5 III	3.54	2
075220.28 − 262546.7	WSI 54	HD 64315	O6 Vn	9.23
080929.33 − 472043.0	WSI 55 Ba,Ab	HD 68243	B1 IV	4.20	Section 5.4
104512.87 − 594419.2	WSI 56	CPD−59 2636	O8 V	9.29
131345.52 − 633511.8	WSI 57	HD 114737	O9 III	8.00	Section 5.4
131444.39 − 633451.8	WSI 58 Aa,Ab	HD 114886	O9 II–III	6.86	Section 5.4
141501.61 − 614224.4	WSI 59 Ba,Ab	HD 124314B	...	8.66	3, Section 5.4
171905.50 − 384851.2	WSI 60	CD−38 11748	O4 If+	11.17
171946.16 − 360552.3	WSI 61 Ba,Ab	HD 319703B	O6.5 V	11.34	Section 5.4
172444.34 − 341156.6	WSI 62 CD	HD 319718C	...	...	Section 5.4
172444.34 − 341156.6	WSI 62 CE	HD 319718C	...	...	Section 5.4
175136.72 − 163236.3	WSI 63 AB	TYC 6249-233	...	11.75	4, 5
175136.72 − 163236.3	WSI 63 AC	TYC 6249-233	...	11.75	5
175331.95 − 162247.0	WSI 64	GSC S81N021274	...	13.28	5
180015.80+042207.0	WSI 65	66 Oph	B2 Ve	4.78
203308.78+411318.1	WSI 66	Cyg OB2-22	O3 If* + O6 V((f))	11.68	Section 5.4
203323.46+410912.9	WSI 67	Cyg OB2-841	O5.5 V	11.89

Notes. (1) Spectroscopic triple noted by G. Torres (2006, private communication). Not examined in the earlier speckle survey of the Pleiades (Mason et al. 1993a). (2) Companion not detected in the earlier speckle survey of the Hyades (Mason et al. 1993b). (3) A new close pair associated with the B component of this multiple system. The precise coordinates above are for the A component. (4) New companion was "preconfirmed" with 2MASS data.⁶ (5) This was a possible occultation target for the New Horizons mission.

Download table as:  ASCIITypeset image

Table 2 lists the astrometric measures of the observed massive binaries. They are subdivided into four groups consisting of the original 1998 sample of O-stars, the newer set of O-stars, WR stars, and B-stars. The first three columns identify the system by providing the epoch-2000 coordinates, discovery designation, and an alternate designation. Columns 4–6 give the epoch of observation (expressed as a fractional Besselian year), the position angle θ (in degrees), and the separation ρ (in seconds of arc). Note that the position angle has not been corrected for precession, and is thus based on the equinox for the epoch of observation. Objects whose measures are of lower quality are indicated by colons following the position angle and separation. These lower-quality measurements may be due to one or more of the following factors: close separation, large Δm, one or both components very faint, a large zenith distance at the time of observation, and poor seeing or transparency. They are included primarily because they confirm an earlier observation or because a long time has elapsed since the last measurement. Column 7 provides the V-band magnitude difference. This is usually a catalog value from the Washington Double Star Catalog (WDS; Mason et al. 2001), although for new pairs and some other infrequently measured interferometric pairs it is a crude value based upon the strengths of the secondary peak and "anti-peak" in Fourier Transform space, as seen in the generated directed vector autocorrelations (Bagnuolo et al. 1992). Differential magnitudes were "calibrated" by direct comparison with other pairs of known magnitude difference and are probably accurate to ±0.5 mag. Column 8 indicates the number of observations used to derive the mean position (usually 1). For systems with orbits, the observed minus calculated residuals O − C for both θ and ρ are given in Columns 9 and 10 according to the orbit whose reference is given in Column 11. Finally, Column 12 refers to specific notes for these systems. Some measures from other KPNO/CTIO 4 m runs are noted and listed here and in Table 3.

Table 3 provides a complete list of single star observations for the massive star sample. The precise coordinate (α, δ) is given in Column 1, while Columns 2–4 list various designations. A code for the massive star subsample is given in Column 5, and the Besselian date of observation appears in Column 6. Column 7 indicates with a K or C if the 4 m telescope used for the observation is the Mayall reflector at KPNO (K) or the Blanco reflector at CTIO (C). Finally, Column 8 provides notes for the stars.

It is important to consider the environment of massive stars in the determination of binary frequency. While most massive stars are found close to their birthplaces in stellar clusters and OB associations, there are significant numbers of "field" O-stars (which have no apparent nearby cluster; de Wit et al. 2005) and "runaway" O-stars (high velocity or remote from the Galactic plane; Gies & Bolton 1986) that were probably ejected from clusters. The ejection process may have involved close gravitational encounters of binaries and/or supernovae explosions in binaries (Hoogerwerf et al. 2000; Zinnecker & Yorke 2007), and such ejected stars will generally be single objects. In our original speckle survey (Mason et al. 1998), we found that indeed the binary fraction decreased among field and runaway O-stars compared to those in clusters and associations.

Here, we revisit the question of the binary frequency of massive stars based upon the results from our speckle interferometric survey. We will restrict our sample to the O-stars appearing in the Galactic O Star Catalog of Maíz-Apellániz et al. (2004), since we now have speckle data for 360 of the 370 stars in the catalog. These stars and their binary properties are listed in Table 4, using the same names and order (based upon increasing Galactic longitude) as given in the Galactic O Star Catalog.

The second column of Table 4 gives a code for the short-period, spectroscopic binary status based upon a literature search through 2008 August. These codes are similar to those adopted by Mason et al. (1998), and we use an "SB" prefix for known or probable spectroscopic binaries, a "C" for constant velocity stars, and a "U" for stars of unknown status (usually with fewer than four radial velocity measurements). The SB stars with a published orbit have an "O" suffix attached to the code and a middle numeral that represents the number of spectral components identified. Usually a code of "SB2O" represents a double-lined spectroscopic binary, but we also apply it to cases such as QZ Car = HD 93206 that consists of two single-lined binaries in a quadruple system. The "SB3O" code is applied to triple systems where a third, stationary, spectral component is visible at the greatest velocity separation of the double-lined system. An "E" suffix denotes the presence of orbital flux variations (eclipses or ellipsoidal variations), and the "SBE" code indicates that we know that the star is a binary from the light curve but no spectroscopic investigation exists yet. The suspected spectroscopic binaries are coded by "SB2?" (where observers report line doubling) and "SB1?" (where the range in measured radial velocity exceeds 35 km s⁻¹). Note that "SB2?" systems are not uncertain in their spectroscopic multiplicity but simply lack complete orbital determination. The most recent published reference is indicated by the SAO/NASA Astrophysics Data System bibliographic code in Column 6 of Table 4.

The number of angularly resolved components is given in Column 3 of Table 4. This represents the sum of the number of close components found by speckle interferometry, wider and fainter components found by Turner et al. (2008b) in an adaptive optics survey, and other (usually wider) components listed in the WDS (Mason et al. 2001). These sources were supplemented by detailed studies of specific stars or clusters, such as ζ Ori (long baseline optical interferometry; Hummel et al. 2000), Trumpler 14 (HST FGS; Nelan et al. 2004), the Orion Trapezium (infrared single aperture interferometry; Petr et al. 1998; Simon et al. 1999; Weigelt et al. 1999; Kraus et al. 2007), and NGC 6611 (Duchêne et al. 2001). A quotation mark in this column indicates that the star is a member of a visual system whose primary component also appears in the table (usually just above or below such an entry), and a colon marks those stars that lack speckle observations. Note that a large number of visual components may indicate that the star resides at the center of a dense cluster.

Column 4 of Table 4 associates the star with the field or the name of the home cluster, while Column 5 lists whether or not the star is considered to be a runaway object. These determinations come directly from the Galactic O Star Catalog (Maíz-Apellániz et al. 2004) with new runaway identifications noted by Mdzinarishvili (2004) and de Wit et al. (2005). Note that some runaway stars can be traced to a cluster of origin, so that they will be assigned to that cluster in Column 4.

The binary statistics derived from Table 4 are summarized in Table 5 (an updated version of Table 3 from Mason et al. 1998). We caution that the sample is magnitude limited (and therefore biased to more luminous stars) and incompletely surveyed (for example, the Turner et al. 2008b adaptive optics work is limited to stars with declination > − 42°). The stars are grouped into cluster/association, field, and runaway categories to compare the binary properties. For the immediate purpose of this work, we simply assigned any star that was not a field or runaway object to the cluster/association category. This includes stars described as more distant than some foreground cluster, since such stars generally reside along a spiral arm of the Galaxy where cluster membership is common. The top section of Table 5 summarizes the visual multiplicity properties of each category for the 347 unique, visual systems in the Galactic O Star Catalog. The results are presented in rows that correspond to the sum based upon the number of visual components n found. We divide the sample into single and multiple groups in determining the percentages without and with companions (making the tacit assumption that most of the visual companions are gravitationally bound and not line-of-sight optical companions).

The middle section of Table 5 presents the corresponding sums for the spectroscopic binary properties for all 370 entries in the Galactic O Star Catalog. The percentages for each subgroup represent fractions with the unknown "U" status objects excluded from the totals. Finally, the lower section in Table 5 shows the percentages for the presence of any companion (spectroscopic or visual) again excluding the stars with unknown spectroscopic status.

The results from this larger sample tend to confirm the trend found by Mason et al. (1998) that the binary frequency is lower among field and runaway stars than that found in the cluster/association group. The binaries found among the runaway stars tend to be close systems with nearly equal mass components (HD 1337, ι Ori, Y Cyg) and binaries with neutron star companions (HD 14633, HD 15137, X Per, HD 153919). The former groups are predicted to be infrequently ejected in close gravitational encounters (Leonard & Duncan 1990) while the latter are the result of a supernova explosion in a binary, so both processes must contribute to the ejection of massive stars from clusters. A number of runaways have visual companions that must be optical, chance alignments, since the ejection processes are too energetic for soft, wide binaries to survive.

The binary statistics for the cluster and association group offer us the best estimate of the binary properties at birth (before dynamical and stellar evolution processes alter the statistics). Our results indicate that most O-stars (and by extension most massive stars) are born in binary or multiple star systems. This result is especially striking since those binaries with orbital periods too long for easy spectroscopic detection and too short for direct angular resolution are absent from the totals, so the fractions reported here are clearly lower limits for the binary frequency. Thus, the processes that lead to the formation of massive stars strongly favor the production of binary and multiple star systems.

5.1. ι Ori = CHR 250

The complex dynamical relationship of AE Aur, μ Col, and ι Ori is one of the best examples of a binary–binary collision (Gies & Bolton 1986; Leonard & Duncan 1990; Leonard 1995; Clarke & Pringle 1992). As ι Ori is a known close pair (P = 29.13376 d; Marchenko et al. 2000), the much wider speckle component would be hierarchical if physical, with an estimated period of at least 40 y (Gualandris et al. 2004). As the high energy needed to eject AE Aur and μ Col with their runaway velocities seemed inconsistent with the less energetic dynamical interaction required for the CHR 250 pair to remain bound, Gualandris et al. (2004) postulated that this pair was nonphysical, despite their close proximity. Figure 2 shows a least-squares, linear fit (see Hartkopf et al. 2006) to the published data (Mason et al. 1998 and Table 2). The data are also consistent with a long-period orbit, but much longer than ≈ 40 yr.

Zoom In Zoom Out Reset image size

5.2. δ Ori = HEI 42

We present a first orbit for the wide component of this triple system that is based on all available published data and the new measures listed in Table 2. The previous measurements were extracted from the WDS (Mason et al. 2001) and were weighted following the precepts of Hartkopf et al. (2001a). The orbital elements were determined with an iterative three-dimensional grid-search algorithm (Seymour et al. 2002). The seven orbital elements are presented in Table 6: P (period, in years), a (semimajor axis, in arcseconds), i (inclination, in degrees), Ω (longitude of the node, equinox 2000, in degrees), T (epoch of periastron passage, in fractional Besselian year), e (eccentricity), and ω (longitude of periastron, in degrees). An ephemeris for the period 2010–2018, in two-year increments, is provided in Table B2. The orbit is illustrated in Figure 3.

Zoom In Zoom Out Reset image size

Table 6. Orbital Elements for δ Ori = HEI 42

Element	Value
P (y)	201
a ('')	0.26
i (deg)	108
Ω (deg)	139
T (BY)	1957
e	0.56
ω (deg)	236

Download table as:  ASCIITypeset image

Due to the preliminary nature and incomplete phase coverage of the orbital fit, the errors are large and difficult to quantify. It is entirely possible that the companion may continue moving to the southeast for longer than the orbit plot and ephemeris would indicate. The orbit here then may prove wildly erroneous, however, it does serve to highlight the need for periodic monitoring of the pair to verify the orbit predictions. The preliminary orbit indicates a total mass of 32 M_☉ for a distance of 414 pc (Menten et al. 2007).

The A component is itself a close binary with an orbital period of about 5.7 days (see Harvin et al. 2002 for a thorough analysis of the close pair). Curiously, the preliminary orbital period, 201 yr, is close to the derived apsidal period of the close binary (227 ± 37 yr, Monet 1980; 225 ± 27 yr, Harvey et al. 1987).

5.3. δ Sco = LAB 3

Bedding (1993) published the first set of orbital elements for δ Sco, followed a few years later by an updated solution from Hartkopf et al. (1996). Both solutions were based solely on interferometric data (speckle interferometry plus two measures made using aperture masking). Miroshnichenko et al. (2001) obtained complementary radial velocity data which tied down T quite precisely and also gave a more accurate estimate of the eccentricity, while adopting the values for period and semimajor axis obtained by Hartkopf et al. (1996).

Since the 1996 solution, observations have covered over one additional revolution. Published data include a speckle measure by Horch et al. (1999) and one measure by Hipparcos (ESA 1997). This paper includes new speckle measures from the Kitt Peak and Cerro Tololo 4 m telescopes, the Mount Wilson 100 inch, and the USNO (Flagstaff Station) 61 inch, as well as unpublished KPNO and CTIO 4 m observations made with the CHARA speckle camera. A new orbital solution was determined, utilizing all available interferometric data and adopting the T and eccentricity values of Miroshnichenko et al. (2001). Elements from this new orbit as well as the previously published solutions are given in Table 7; future ephemerides for the new orbit are given in Table B2. The new solution and all data used in its determination are shown in Figure 4. Here speckle data from this paper (Table 2) are shown as filled stars, while other interferometry measures are indicated by filled circles; the Hipparcos measure is shown as a letter "H." Measures are connected to their predicted locations along the orbit by "O − C" lines; the dotted lines indicate measures given zero weight in the final orbital solution. The dot-dashed line indicates the line of nodes and the shaded circle surrounding the origin indicates the Rayleigh separation limit for a 4 m telescope. At two epochs in early 1990, observations obtained with the KPNO and CTIO 4 m telescopes did not resolve the pair; these are indicated by dotted O − C lines from the origin to their predicted locations along the orbit. According to the orbital solution, these observations should have been marginally resolved. However, given a magnitude difference Δm > 2 mag, the lack of resolution so close to the Rayleigh limit is not at all surprising. The total mass of the system is approximately 27 M_☉ for a distance of 140 pc (Shatsky & Tokovinin 2002).

Zoom In Zoom Out Reset image size

Table 7. Orbital Elements for δ Sco = LAB 3

Element	Bedding (1993)	Hartkopf et al. (1996)	Miroshnichenko et al. (2001)	This Work
P (y)	10.5	10.583 ± 0.075	10.58a	10.68 ± 0.05
a ('')	0.11	0.1067 ± 0.0067	0.107a	0.104 ± 0.006
i (deg)	70	48.5 ± 6.6	38 ± 5	39 ± 8
Ω (deg)	0	159.3 ± 7.6	175	153 ± 9
T (BY)	1979.3	1979.41 ± 0.14	2000.693 ± 0.008	2000.693b
e	0.82	0.92 ± 0.02	0.94 ± 0.01	0.94b
ω (deg)	170	24 ± 13	−1 ± 5	29 ± 12

Notes. ^aParameter adopted from Hartkopf et al. (1996) solution. ^bParameter adopted from Miroshnichenko et al. (2001) solution.

Download table as:  ASCIITypeset image

5.4. Notes on Stars Listed in Table 1

5.5. Notes on Stars Listed in Table 2

5.6. Notes on Stars Listed in Table 3

The USNO speckle interferometry program has been supported by NASA and the Space Interferometry Mission through Key Project MASSIF and is based upon work supported by the National Aeronautics and Space Administration under Grant No. NNH06AD70I issued through the Terestrial Planet Finder Foundation Science program. This material is based upon work supported by the National Science Foundation under Grant No. AST-0506573 and AST-0606861. This research has made use of the SIMBAD database, operated at CDS, Strasbourg, France. Thanks are also extended to the U. S. Naval Observatory for its continued support of the Double Star Program. The telescope operators and observing support personnel of KPNO and CTIO continue to provide exceptional support for visiting astronomers. Thanks to Skip Andree, Bill Binkert, Gale Brehmer, Ed Eastburn, Angel Guerra, Hal Halbedal, David Rojas, Patricio Ugarte, Ricard Venegas, George Will, and the rest of the KPNO and CTIO staffs.

Tables A1 and A2 are identical in form to Tables 2 and 3, respectively, but the targets listed here are selected from other sample sets as indicated in the more extensive collection of notes. As many of these systems were used for either primary (CTIO) or secondary (KPNO) scale and angle calibration, many have calculated orbits, with residuals derived from orbital solutions given in Table B1.

We found that some of those pairs identified as calibration systems in Table A1 (used to investigate differential magnitude detection rates at various separations; Figure 1) had poorly defined orbits. The KPNO measures, independently calibrated by use of the slit-mask, allowed us to optimize these orbits and to generate ephemerides, which helped us calibrate the CTIO measures. Going one step further, these CTIO measures could then be incorporated in a new, improved orbit solution using the same methodology as that described in Section 5.2. These orbital elements are presented in Table B1, together with their grades (see Hartkopf et al. 2001a for a description of the grading scale). Also provided in Table B1 is the reference to the previous "best" published orbit. Formal errors are listed below each element. Future ephemerides are presented in Table B2 and relative orbit plots are illustrated in Figures 5 and 6, with the dashed curve indicating the prior orbit listed in Table B1.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size