Measurements of Neglected Double Stars

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 411

Brandon Bonifacio, Calla Marchetti, Ryan Caputo, and Kalée Tock

Stanford Online High School Stanford, California, United States

Abstract: Double stars with a dim, high delta-magnitude companion are difficult to resolve and measure, and are therefore often neglected despite their high abundance in the galaxy. We measured fourteen of these dim, high delta-magnitude doubles, some from the WDS and some discovered by Gaia but never studied before. Although all of our systems’ components have similar parallaxes and proper motions, many of the systems have only a few observations other than what is presented here, making them historically neglected. To resolve the systems, we used PixInsight and AstroImageJ to perform image stacking. Using the measurements from Gaia Data Release 2, we present an escape velocity estimate to assess the likelihood of a system being gravitationally bound. A Monte Carlo method was employed to characterize the error associated with this calculation.

Introduction velocity and relative velocity are approximate because Red dwarf stars have masses less than 50 percent of the many estimates involved (Caputo, 2020). For that of our Sun and surface temperatures in the range example, escape velocity depends on mass, which can 2500-4000K. Despite their small masses, red dwarfs be computed from luminosity if the star is on the main burn through their hydrogen supplies relatively slowly, sequence. However, the formula for the mass luminos- giving them long lifespans of about 100 billion years. ity relationship itself is an approximation, depending These extended lifespans imply that the population of on the type of star to which it is applied. Also, if a red dwarfs is relatively large; it is estimated that they star’s luminosity is not directly available in Gaia DR 2, make up 70% of stars in the Universe (Kaufman, then it must be estimated from its observed magnitude 2017). Despite their abundance, however, red dwarfs together with its distance, both of which have their own are less commonly observed in double star studies be- associated errors. Therefore, in order to evaluate the cause imaging them can be problematic. For example, error on both escape velocity and relative velocity, it is red dwarfs are often located in high delta-magnitude useful to sample the input variable distributions using a Monte Carlo technique. systems, making them difficult to separate from the overwhelming brightness of their corresponding prima- Target Selection ries. Also, it is not uncommon for red dwarfs to be Target selection was based on a number of criteria. fainter than 14th magnitude, which necessitates long First, the coordinates had to be visible to Ryan Caputo exposure times. Because red dwarfs are more difficult on his telescope in Texas in February, at which time to measure, many of these systems are either not im- the best RA was between 3 and 15 hours. Declination aged or have few observations (Wasson, et al., 2020). needed to be above -10 degrees. The instruments used, We searched the Washington Double Star Catalog which are described in the next section, required the (WDS) and Gaia Data Release 2 (DR 2) specifically separation to be greater than 4 arcseconds. If the sys- for these red dwarf stars or stars having other notewor- tem delta magnitude was larger than 3, then its separa- thy characteristics, such as being relatively unobserved tion needed to be correspondingly higher, as a large or neglected systems. difference in magnitudes makes stars more difficult to For each of the systems studied here, we assess the split. Individual magnitudes were limited to 16 or less, likelihood that the system is gravitationally bound by for the sake of imaging time. comparing the system escape velocity to the relative To find double star systems satisfying all of these velocity of the component stars. However, both escape criteria, the GDS1.0 tool was used.

Received *****DATE HERE******

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 412

Measurements of Neglected Double Stars

GDS1.0 is software (Dave Rowe, 2018) which star required 20 minute exposures for the particular tel- compiles lists of Gaia DR2 double stars with separation escope used.

less than 10” from the 1.3 billion stars in Gaia DR2, Targets matching these to systems in the WDS where possible The targets systems of this study are shown in Ta- (Wasson et. al., 2020). Using this program, users can ble 1 together with their parallax and proper motion extract stars that satisfy specific conditions of RA, Dec, (PM) data from Gaia DR2. Systems that exist in the primary star magnitude (Comp 1 Mag), delta- WDS are listed by discoverer code, while systems that magnitude, separation, and Gravitationally Bound In- are not in the WDS are identified by their GDS1.0 ID. dex (GBI). The GBI is a parameter that estimates the For each system, a proper motion ratio (rPM) was cal- likelihood of a physical relationship based on similarity culated. The rPM of the stars is a representation of how of parallax and proper motion. Further sorting on Gaia similarly they move across the celestial sphere, and it is parallax and proper motion produced candidates likely calculated using the equations below. First, the magni- to be true binaries. tude of the stars’ relative motion is calculated by taking Even after narrowing down double star systems the magnitude of the difference vector between the pri- according to these criteria, an abundance of possible mary and secondary proper motion vectors, as shown in targets remained. Among those, we were particularly Equation 1. interested in stars with temperatures below 4000K, in- In Equation 1, pmRA1 is the angular velocity of the dicating that they are red dwarfs. We also targeted one primary star in right ascension measured in milliarcsec- star with a temperature above 7000K and a relatively onds per year, pmDec1 is the angular velocity of the dim absolute magnitude, indicating that it is most likely primary star in declination measured in milliarcseconds a white dwarf. Some of the stars we selected have parallax greater than 30 milliarcseconds (mas); these are of particular interest since they are in our solar neighborhood.

Instruments Used Equation 1: Proper motion of stars A ZWO ASI 1600mm camera attached to a 6-inch classical Cassegrain was used to image all the double per year, and so on for the variables with a 2 subscript star systems for this study. The Cassegrain has a focal for the secondary star. ratio of f/12, giving it a relatively high focal length of After this, the rPM of the stars is calculated by 1800 mm. Combined with the small 3.4 μm pixels of dividing the magnitude of the proper motion by the the ASI 1600mm, the sampling is 0.4“ per pixel. This magnitude of the larger proper motion vector. For ex- yields excellent resolution on close double stars, allow- ample, if the primary star has a larger magnitude of ing seeing-limited measurements to be made. proper motion than the secondary star, then the rPM is To overcome the difficulty of imaging high delta given by the below equation. mag systems, an Astronomik IR pass 742nm filter was used in observations specified in Table 2. This filter often removes more light from the primary than the secondary, lowering the delta mag of the system and allowing measurements to be made on otherwise im- Equation 2: Relative proper motion of stars possibly-difficult pairs. Additionally, a guiding system was employed to improve tracking accuracy. This par- If the secondary star’s proper motion is larger, then tially corrects periodic error and polar alignment error the subscript becomes a 2 instead of a 1 in the denomi- for very long exposures. With an aperture of only six nator of Equation 2. A small rPM value indicates that inches and a slow focal ratio, photons are at a premium, the difference in component motions is small relative to and reaching 15th magnitude stars requires 10 minute the component motions themselves. An rPM below 0.2 exposures with an open filter. Imaging in the infrared is the criterion for classifying the stars as having Com- has its benefits: namely, the ability to record higher del- mon Proper Motion, or CPM (Harshaw, 2016). All of ta mag stars at closer separations. But this comes at a the stars in this study are CPM stars. Note that STF cost: longer exposure times. Not only does less light 326AB and LDS 883AC are part of a single triple sys- pass to the camera, but the ASI 1600mm camera is less tem, although an orbit has not been established for ei- sensitive at infrared wavelengths. These effects com- ther system.

bine to make it difficult to image the already-faint red dwarfs. Through the 742nm IR filter, a magnitude 14.5

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 413

Measurements of Neglected Double Stars

Parallax of Parallax of Proper Motion Proper Motion System Primary Secondary of Primary of Secondary rPM (mas) (mas) (mas/year) (mas/year)

(264.5±.10, (282.0±.14, STF 326AB 44.4±.05 44.4±.07 .098 -193.4±.07) -166.5±.1)

(264.5±.10, (276.9±.15, LDS 883AC 44.4±.05 44.3±.06 .056 -193.4±.07) -179.9±.11)

(86.2±.09, (85.1±.1, 6763049 9.7±.04 9.7±.06 .010 -116.6±.05) -117.4±.06)

(-34.2±.03, ( - 30.9±.03, 6814658 12.2±.05 12.3±.04 .095 -9.3±.07) -9.9±.08)

(10.5±.09, (11.1±.06, KPP 837 6.0±.05 5.94±.03 .029 -37.2±.05) -38.1±.03

(18.6±.08, (17.3±.05, 6709693 6.5±.04 6.5±.03 .040 -27.5±.07) -27.9±.04)

(-105.5±.10, (-106.6±.06, 6765766 11.4±.07 11.3±.03 .011 3.8±.07) 3.5±.03)

(-6.8±.09, ( - 7.8±.10, HDS 597 9.5±.05 9.7±.06 .059 -16.9±.04) -16.5±.06)

(-93.5±.09, ( - 96.2±.74, 6770204 16.2±.04 16.2±.08 .028 -27.8±.08) -28.7±.67)

(-228.6±.13, ( - 205.3±.33, LDS 905 60.2±.08 62.0±.21 .095 -120.0±.09) -127.6±.19)

(22.4±.04, (22.2±.04, 6782185 9.2±.04 9.3±.03 .013 -4.1±.05) -3.6±.04)

(97.7±.06, (98.5±.13, SKF 365 34.1±.03 34.0±.05 .073 -57.8±.06) -49.6±.11)

(-19.0±.10, (-19.1±.07, STT 181 6.5±.05 6.5±.06 .085 4.3±.06) 6.0±.05)

(1303.3±2.05, (1335.0±.08, STI 2051AB 180.4±.59 181.3±.05 .041 -2043.8±1.23) -1947.6±.09)

Table 1: Gaia parallax, proper motion, and rPM (ratio of the PM difference vector magnitude to the longer proper motion vector magnitude) of our target systems.

Noteworthy Stars separation of 4.58″ and a low delta mag of 1.93, while A few of our targets have noteworthy characteris- LDS 883AC has a high separation of 43.75″ and a high tics. We have selected a few to discuss in detail below. delta mag of 5.17. Because of this, the three stars can The first interesting system is STF 326AB and LDS not be seen at the same time in the same image. As 883AC. These systems have the same primary star, so shown in Figure 1, if the contrast is increased so that together they form a triple rather than a double. Howev- the primary and secondary are distinguishable, the ter- er, it is not certain whether the tertiary is gravitationally tiary is so faint that it is barely visible. However, when bound. These stars also proved to be complicated to the contrast is decreased so the tertiary is observable, image process due to the large difference in magnitudes the primary and secondary seem to have melded togeth- and separation between the stars. STF 326AB has a low er.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 414

Measurements of Neglected Double Stars

Figure 1: Measuring large delta magnitude systems, with triple in lower right. Figure 3: The position of the tertiary with respect to the primary The historical measurements of STF 326AB and LDS 883AC were plotted along with the Gaia measurement and our measurement. Figure 2 and Figure 3 are the graphs of the stars’ relative positions, which are plotted by fixing the primary’s location at the origin. Our measurement for STF 326AB seems to align with the historical trend, but for LDS 883AC there are not enough historical measurements for a trend to be discernible. Hopmann proposed an orbital solution for this pair in 1967, but it is listed as having incomplete elements (Hopmann, 1967). Hartkopf provided a linear solution in 2017, shown below in Figure 4 (Hartkopf, 2017). The system having both an orbital and linear solution implies that if it is binary, the period is very long.

Figure 4: Linear solution for STF 326AB. Note that the orientation is flipped relative to the historical data plot Another interesting system is STI 2051 AB; this pair is the amalgam of all the characteristics we focused on. Its high parallax of 180 mas puts it the closest to earth of all of our targets, at 5.5 parsec away. The primary and secondary have surface temperatures of 3300K and 7800K, which are the highest and lowest temperatures of any of our targets. The low luminosity of the secondary confirms its status as a white dwarf; type O stars have similar temperatures but much higher luminosities. This makes STI 2051AB a pair with both Figure 2: The position of the secondary with respect to the a red dwarf and a white dwarf. A linear solution was primary provided by Hartkopf in 2017 (Hartkopf, 2017), and is shown in Figure 5. An updated plot with the Gaia measurement and our measurement is also shown in Figure 6

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 415

Measurements of Neglected Double Stars

stars themselves appear brighter, thereby increasing the signal to noise ratio (SNR). Note that the primary star is not saturated despite it appearing overexposed. The black and white contrast settings are adjusted from their defaults to show the secondary star, and this has the effect of making the brighter primary appear extra bright. However, it is important to note that the pixel values are not saturated in the unstretched image.

Figure 5: Linear solution for STI 2051AB. Note that the orientation is flipped relative to the historical data plot

Figure 7: Image of 6814658, of magnitudes 7.3 and 12.3, before stacking (top) and after stacking (bottom) in AIJ. Figure 6: Plot of STI2051 position, with the GAIA measurement and our measurement Rather than stacking all of the files into a single image, multiple stacks were made for each system to Image Stacking create a pseudo-exposure. These pseudo-exposures Like LDS883AC, several of these systems have a were then measured and averaged to produce a more bright primary and a high delta magnitude, so an expo- accurate overall measurement. Ideally, just enough im- sure time that is optimal for the secondary star would ages are stacked so that the SNR is sufficient for accu- saturate the primary. To increase the signal to noise of rate astrometry. If too many images are stacked, then the secondary star without increasing exposure time, the extra frames are wasted because there is no need for the images must be stacked. To do this, the images the extra SNR. Images are platesolved in AIJ, and once were first aligned relative to a reference frame using the images are platesolved AIJ automatically calculates PA software PixInsight. Then these aligned images were and separation between two points. Table 2 shows the imported into AstroImageJ and stacked into a single systems with their delta magnitudes listed, and Figure 7 image, in which each pixel has the average brightness shows a high delta mag system before and after stack- of the pixels in the component images. Because bright- ing in AIJ. Table 3 shows how many images were used ness fluctuations due to noise are random, this has the per stack for each of the systems, as well as the meas- effect of averaging out the noise without making the urements obtained from these stacks.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 416

Measurements of Neglected Double Stars

Spectral Spectral Magnitude Magnitude of Delta Filters Used System Type of Type of of Primary Secondary Magnitude Primary Secondary STF 326AB K K 7.33 9.26 1.93 infrared 742

LDS 883AC K G 7.33 12.50 5.17 infrared 742

6763049 K K 11.91 14.70 2.79 open

6814658 F K 7.32 12.34 5.02 open

KPP 837 K K 11.84 13.35 1.52 open

6709693 G K 11.23 13.53 2.30 infrared 742

6765766 G K 10.77 13.39 2.62 infrared 742

HDS 597 F K 8.95 11.96 3.00 infrared 742

6770204 K M 11.56 13.22 1.65 open

LDS 905 M M 10.55 12.97 2.42 infrared 742

6782185 F K 8.36 13.83 5.47 infrared 742

SKF 365 M M 11.89 14.57 2.68 infrared 742

STT 181 A G 7.97 11.77 3.80 open

STI 2051AB M F 9.70 12.35 2.65 open

Table 2: Spectral types, estimated from Gaia BP-RP and an HR diagram, and magnitudes (in Gaia G) of stars.

Standard Images per Position Standard Number of Error on Separation System Date Stack Angle Error on Images Position (“) (°) Separation Angle

STF 326AB 2020.11 100 20 221.28 .100 4.58 .040

LDS 883AC 2020.11 100 20 266.35 .016 43.75 .011

6763049 2020.11 5 1 130.33 .108 8.95 .018

6814658 2020.11 100 20 257.78 .049 9.89 .017

KPP 837 2020.11 30 7 88.53 .052 7.1 .006

6709693 2020.15 2 1 185.3 .250 8.02 .014

6765766 2020.16 15 1 134.34 .057 7.22 .010

HDS 597 2020.16 21 6 79.82 .057 6.61 .007

6770204 2020.16 10 1 317.56 .048 7.59 .009

LDS 905 2020.16 14 1 348.29 .014 9.98 .002

6782185 2020.16 29 9 356.33 .045 9.23 .006

SKF 365 2020.16 3 1 286.07 .079 8.98 .016

STT 181 2020.16 250 20 262.68 .030 6.75 .002

STI 2051AB 2020.01 10 1 58.60 .024 10.99 .010

Table 3: Measurements of double stars.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 417

Measurements of Neglected Double Stars

Overview of Important Variables Before the release of Gaia DR2, many optical doubles were incorrectly assumed to be binaries because their component stars appeared relatively close together Equation 3: Transverse separation of stars where p1 is the in the sky and had similar proper motions. However, parallax of the primary star. Gaia DR2 parallax measurements indicated that for some of these systems, the component stars were actu- parallaxes of both stars are very similar. Moreover, in ally many parsecs apart from each other in the radial the overall calculation of three-dimensional spatial sep- direction, making it impossible for them to be gravita- aration for these targets, the transverse separation is tionally bound to each other (Dugan et. al, 2019). usually greatly overshadowed by the radial separation. Now that Gaia DR2 parallax, PM, and radial ve- However, when examining stars with larger differences locity measurements are available for many WDS and in parallax, it might be useful to modify this equation to other systems, we can evaluate the probability that two choose the parallax of the star further away from Earth, stars are gravitationally bound together more accurately so that it represents the maximum transverse separation. in terms of the rPM discussed earlier in the paper and Once the transverse separation has been found, the three other variables: the three-dimensional distance three-dimensional distance in parsecs between the stars between the two stars derived below, the system escape can be computed using equation 4. velocity derived in Appendix A, and the actual relative velocity of the stars derived later in the paper.

Three-Dimensional Distance The three-dimensional distance between the stars is the spatial separation between them. This distance alone provides a useful filter because both binary stars Equation 4: 3D separation between the stars. would have to be fairly massive and also moving very slowly relative to each other in order to remain gravita- Escape Velocity tionally bound beyond roughly one-tenth of a parsec, so The escape velocity of the stars is calculated from in general, it is unlikely that two stars with a separation the property that the total mechanical energy of the bi- of more than one-tenth of a parsec will be gravitational- nary star system must be zero for the stars to orbit each ly bound. To be specific, if the two stars each had the other. Otherwise, the stars would not be bound. mass of our sun, which is unlikely considering the most Through the use of this principle, we can calculate the common star is a red dwarf or a another star that has maximum velocity that a star can have in order to be in less mass that our sun, the relative velocity of the stars a physical orbit with another star (Thornton, 2004; would have to be less than 400 m/s in order to be gravi- Richmond, 2020). The escape velocity equation is giv- tationally bound to each other. As can be seen in Table en below, and the full derivation of it is left to Appen- 5, not many star systems have minimum actual veloci- dix A. ties less than 400 m/s, so most stars with separations greater than one-tenth of a parsec are not gravitationally bound. Furthermore, most of the stars in this study had a mass less than that of our sun. However, the three- dimensional distance alone can not act as a definite rule for deciding whether stars are bound together because some systems are actually composed of slow, massive Equation 5: Escape velocity equation of the binary star system where r is the stars that can remain coupled at above this distance. three-dimensional distance between the The three-dimensional distance is also used in the stars and M is the total mass of the sys- expression for the escape velocity of the system. To tem, which is equal to the mass of the calculate it, the transverse distance between the stars in primary plus the mass of the secondary. parsecs must first be computed using equation 3. To calculate the total mass used in the escape ve- In equation 3, sep is the separation of the stars in locity equation, we use a mass-to-luminosity approxi- arcseconds. This is calculated by AIJ as described mation for main-sequence stars that can be summarized above in Image Stacking. The variable p1 is the parallax by eqation 6 below (Salaris & Cassisi, 2005; Duric of the primary star measured in arcseconds. The choice 2004; “The Eddington Limit,” 2003). of the primary star’s parallax here is arbitrary, since we are computing the transverse separation only, and the

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 418

Measurements of Neglected Double Stars

Equation 9: Transverse velocity of the stars, where pmMag is as found using Equation 1 and p1 is the parallax of the primary star. The choice of the primary’s parallax is arbitrary for the same reasoning described earlier. relative three-dimensional velocity of the stars by tak- Equation 6: Mass-Luminosity equations where M and L are in ing the square root of the sum of the squares of the two- units of solar masses and solar luminosities, respectively. As a dimensional relative velocity and the relative RV, as note, only the first two equations were used in this study due to the nature of the stars we were observing. To use the last equa- shown in equation 10. tion, truly massive stars would be required. Once this final 3D velocity is computed, it is com-

The luminosity for the stars is determined from Gaia’s catalog, but if Gaia does not have the luminosity of the star in its catalog, the following approximation is used from the visual apparent magnitude of the stars Equation 10: 3D velocity of the stars. using Gaia’s filter. First, the absolute magnitude of the star is calculated from equation 7. pared to the escape velocity of the system to determine whether the stars are bound. For example, based on the equations above, KPP 837 has a total mass of roughly 2.8 solar masses (from the mass-luminosity relation- Equation 7: Absolute magnitude of a star ship). The distance between its component stars is estimated from the parallax of the star, p, about 1.25 parsecs. Therefore, the system escape ve- and the apparent visual magnitude of the locity of KPP 837 is roughly 107 m/s. Since the stars star, m. are moving at roughly 822 m/s relative to each other in the transverse direction (no radial velocities were avail- Once the absolute magnitude is calculated, the lu- able for this star system, so transverse velocity must be minosity of the star is estimated from Equation 8 be- used), it would seem unlikely that KPP 837 is a binary. low. Then the mass is found from the relation shown in However, based on a complete analysis of this system Equation 6, and used in the escape velocity calculation (which we will conduct below), KPP 837 actually does of Equation 5. have a decent probability of being bound. The reason the analysis above seems to contradict this is that it does not fully account for the error on each of the con- stituent measurements. The errors of the final velocities Equation 8: Luminosity of the star in solar luminosities. depend on the errors of each of the inputs in complex ways, and the inputs themselves vary depending on Relative Velocity which measurements are available for each star. Be- Once the escape velocity for the system has been cause of the complexity inherent in the way the error estimated from Equation 5, it is compared to the rela- propagates through these equations, we employ a Mon- tive velocity of the stars to each other. To calculate the te Carlo series to understand the true likelihood of a binary relationship between the stars. actual relative velocity of the system, we first calculate the actual relative transverse velocity, equation 9 be- Monte Carlo Analysis low, which is taken by converting the magnitude of the For the stars studied here, the largest sources of difference in the angular velocity vectors (from Equa- error are their parallax values and proper motion vec- tion 1) from units of mas/yr to kilometers per second. tors. The variation is significant enough to drastically This is done by multiplying by unit conversion factors change the resulting analysis. The way that error affects and the radial distance from the stars to the Earth as the computation is nontrivial to assess because it affects shown below. Just like Equation 3, p1 is taken to be the each star system in a different way. For example, it parallax of the primary star. might skew one system’s distribution of escape veloci- This produces the two-dimensional relative velocities to the left, but a different system’s velocities to the ty, and for some stars, a two-dimensional velocity is the right. If that were not enough, understanding the ways best we can do. However, for the stars for which Gaia has measured radial velocity (RV), we can calculate the

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 419

Measurements of Neglected Double Stars

in which the input distributions themselves are skewed related to the input variables in complex ways and re- is nontrivial, especially in the case of Gaia parallaxes quires this method of analysis. (Luri et. al, 2018). Therefore, we developed a “catch- The 95% confidence interval produced is where all” Monte Carlo method of statistical analysis to understand how error affects the inferences that can be drawn regarding these systems. This Monte Carlo method was roughly based on the curriculum taught by Adam Rengstorf during the Summer Science Program (Rengstorf, 2019). A Monte Carlo series randomly selects values for the input parameters of separation, parallax, proper motion, radial velocity, and luminosity from their corresponding distributions. Escape velocity and relative velocity are then computed from these values and stored in a large list. After having done this a large number of times, statistical analysis of the list yields a distribution of the resulting velocities that can be used to understand each of their corresponding errors. After Figure 8: Example Monte Carlo histogram. This histogram was constructed to show how complex relation- a series of tests, one million Monte Carlo iterations ships exist between the input variables and output variables seemed to output results that were consistent, so one because, although the input variables were all normally distrib- million is the number of iterations used here. uted, the output variable is skewed right. Thus, advanced meth- To provide an example of this so that others may ods are required in order to accurately take the error of the reproduce this method for other star systems, we show input variables into account. For clarity, the vertical scale is the number of times each range of three-dimensional separations the procedure and results of this analysis for KPP 837 were calculated out of the one million iterations. So, the range of below. This is an ideal example due to its relative sim- distances between 0 parsecs and 0.275 parsecs occurred roughly plicity and also because it demonstrates the usefulness 120,000 times, or 12% of the time. of this method.

Example Monte Carlo Series conclusions can be drawn. For this example, the 95% The Monte Carlo series for this research requires confidence interval for the three-dimensional distance twenty-two input parameters for each binary star sys- between the stars of KPP 837 is (0.058, 3.979), and this tem. In the case that the radial velocity for both stars is range is retrieved from the summary of all confidence not known, the actual relative velocity calculated is a intervals for each star system in Table 5. Since the low- two-dimensional transverse velocity instead of a three- er bound of this range is below one-tenth of a parsec, dimensional velocity. The input parameters for KPP we can be confident that it is entirely possible that these 837 are listed below in Table 4. stars are close enough to be gravitationally bound. The standard error of a value is taken to be equal to However, in order to be more sure, the velocities need the standard deviation of that value’s Gaussian distribu- to be analyzed. The escape velocity and actual velocity tion. With this interpretation, together with the assump- ranges for KPP 837 (Table 5) are, respectively, (60, tion that the data are normally distributed (in reality, the 497) and (720, 930), and the actual velocity is two- data are not always normally distributed), one million dimensional because the radial velocity of the second- numbers are selected randomly from the distributions ary star was not available. Since these two intervals do created by each value’s range in each system’s respec- not overlap at all, we would conclude that it is improba- tive table like the one in Table 4. For each set of values, ble that these stars are gravitationally boundary. How- the calculation is performed for the three-dimensional ever, since the actual velocity is two-dimensional, it is separation of the stars (Equation 4), the actual relative entirely possible that the stars could be moving towards velocity (Equation 9 or 10), and the escape velocity each other at 223 m/s or more in the radial direction, (Equation 5). The resulting list of these values is or- which would mean that their velocity intervals would dered, and a 95% confidence interval is created from overlap. Since 223 m/s is a possible speed with which the center 95% of the data. The example histogram pro- the stars could be moving towards each other (speeds duced for the three-dimensional separation of the com- greater than 500 m/s, while also technically possible, ponents of KPP 837 is shown in Figure 8. As shown, are more unlikely), we can conclude that it is possible the resulting range of possible three-dimensional dis- that the stars are gravitationally bound. If the velocity tances is skewed right, so the final variable calculated is

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 420

Measurements of Neglected Double Stars

did overlap, we would conclude that it is likely that the become more concrete once radial velocities for more stars are gravitationally bound. This method will stars become known, perhaps in Gaia DR3.

Data Value Name Data Value Units

Angular Separation 7.1 ± 0.006 arcseconds

Parallax of Primary 5.9755 ± 0.0463 milliarcseconds

Parallax of Secondary 5.9439 ± 0.0315 milliarcseconds Primary Right Ascension 10.460 ± 0.092 milliarcseconds/year Proper Motion Primary Declination Proper -37.112 ± 0.049 milliarcseconds/year Motion Secondary Right Ascension 11.058 ± 0.061 milliarcseconds/year Proper Motion Secondary Declination Proper -38.110 ± 0.032 milliarcseconds/year Motion Primary Radial Velocity 4.89 ± 0.31 kilometers/second

Secondary Radial Velocity Unknown ± Unknown kilometers/second

Primary Luminosity Range (0.421, 0.430) solar luminosities

Secondary Luminosity Range (0.125, 0.128) solar luminosities Table 4: Values for KPP 837 Monte Carlo. Since the second radial velocity is not known, the Monte Carlo will only calculate the actual transverse velocity and use that for the two-dimensional velocity.

Results Table 5 shows the confidence interval results of the Furthermore, the confidence intervals for the escape Monte Carlo code for each star system. and relative velocities are not close to overlapping, so Based on the results of the Monte Carlo, we can de- based on this data analysis we should be skeptical of an termine quite a bit about the characteristics of the star orbital solution for HDS 597 if this star system has not system we are studying. Take star system LDS 883 AC, been observed for a long time or if a linear model for for example. Having a minimum three-dimensional sep- the motion of the stars works just as well as an elliptical aration of less than one-tenth of a parsec is a good sign model for the motion. that they could be gravitationally bound to each other as Based on the data generated by the Monte Carlo shown in the table above. None of the stars with a mini- analysis, the binary star systems can be classified into mum three-dimensional separation greater than 0.1 par- three groups: probable, possible, and improbable to be secs have overlapping relative and escape velocities, physically bounded. The probable group consists of and the same can be said for many of the systems with stars whose separation confidence interval includes val- minimum separations slightly smaller than 0.1 parsecs. ues near (perhaps 0.15 parsecs, for example, if the sys- In LDS 883 AC’s 3D separation confidence interval, tem consists of massive stars) or below one-tenth of a the minimum separation between the stars is about parsec and whose escape and relative velocity intervals 0.011 parsecs. Therefore, LDS 883 AC’s component overlap. Possible systems are systems where the confi- stars could be close enough to form a physical binary dence intervals are almost fulfilling the requirements star system. Furthermore, the confidence intervals for for a likely system in a way that the radial velocity not LDS 88C’s escape velocity and actual velocity signifi- accounted for could explain the difference between the cantly overlap with a common range of 542 m/s. Tak- intervals, and an improbable system’s confidence inter- ing this into account, we can consider it likely that LDS vals are not close to fulfilling these requirements. These 883 AC is a binary star system. groupings are shown in Table 6. On the other hand, consider HDS 597. The mini- When drawing conclusions from this data, it is best mum 3D separation of the component stars is 1.448 to keep in mind that even a slight error in the data col- parsecs. This is an extremely large distance between the lection of the parallax angle could skew the results stars, so the results of this analysis would not support a drawn for a star, so researchers should not place too conclusion that HDS 597 is gravitationally bound. much weight on the values of this table until

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 421

Measurements of Neglected Double Stars

95% Confidence 95% Confidence 95% Confidence Interval for 3D Interval for Interval for Type of Actual System Separation Escape Velocity Actual Relative Relative Velocity (Parsecs) (m/s) Velocity (m/s)

STF 326 AB (0.00138, 0.0901) (402, 3266) (3035, 3087) 2D

LDS 883 AC (0.0113, 0.785) (110, 923) (381, 1122) 2D

6763049 (0.0288, 1.917) (71, 577) (532, 727) 2D

6814658 (0.084, 1.611) (116, 504) (1137, 1198) 2D

KPP 837 (0.058, 3.979) (60, 497) (720, 930) 2D

6709693 (0.068, 3.672) (64, 470) (744, 978) 2D

6765766 (0.0703, 2.012) (84, 452) (317, 487) 2D

HDS 597 (1.45, 4.779) (63, 114) (486, 4338) 3D

6770204 (0.0113, 0.787) (110, 922) (381, 1124) 2D

LDS 905 (0.362, 0.596) (90, 115) (1673, 1767) 2D

6782185 (0.0283, 1.734) (105, 829) (188, 302) 2D

SKF 365 (0.0168, 0.208) (148, 524) (991, 1051) 2D

STT 181 (0.0569, 4.066) (82, 695) (1002, 1188) 2D

STI 2051AB (0.00166, 0.0620) (203,1241) (2313, 2435) 2D

Table 5: Monte Carlo results for the 3D separation, escape velocity, actual relative velocity, and actual relative velocity type for the systems of stars studied in this research. The values in each range are the minimum and maximum values for each confidence interval.

Probably Gravitationally Bound Possibly Gravitationally Bound Improbably Gravitationally Bound

STF 326 AB STI 2051AB HDS 597

LDS 883 AC 6814658 LDS 905

6763049 KPP 837

6765766 6709693

6770204 SKF 365

6782185 STT 181

Table 6: Binary Star Systems categorized as probably, possibly, or improbably gravitational bounded.

scientists have developed excellent methods for ble star systems that previously have been difficult to measuring the parallaxes of stars with less error and measure, such as red dwarf binaries. By using filters more consistency. The mass-luminosity relationships and stacking images in AstroImageJ, angular separation might also miscalculate the mass of the stars since the and position angle can be measured for these stars. equations are estimations. Furthermore, most of the Furthermore, a Monte Carlo method of error analysis actual velocities for these stars were 2D velocities be- can be used to estimate the likelihood that the compo- cause many of the systems lacked radial velocity meas- nent stars of a double star system are gravitationally urements in Gaia. Hopefully there will be more radial bound to each other. This technique involves calculat- velocity measurements in future Gaia catalogs. ing confidence intervals for the separation between the

Conclusion stars, the escape velocity for the system, and the actual These methods provide a means of observing dou- relative velocity of the stars. However, because slight

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 422

Measurements of Neglected Double Stars

deviations in parallax and proper motion can have pro- Kaufman, M., 2017 “Red Dwarf Stars and the Planets found effects on the results of these methods, the tech- Around Them” nique outlined here can be expected to improve corre- Luri, X., A. G. A. Brown, L. M. Sarro, F. Arenou, C. A. sponding to the precision of the data available in future L. Bailer-Jones, A. Castro-Ginard, J. de Bruijne, T. Gaia data releases. Prusti, C. Babusiaux,, and H. E. Delgado, 2018.

Acknowledgements “Gaia Data Release 2: Using Gaia parallaxes” As Rick Wasson graciously gifted the Astronomik IR tronomy and Astrophysics, 616, A9, 1 - 19. pass 742nm filter, along with several other filters and a Richmond, M. (n.d.). The "Reduced Mass" approach. filter wheel. These were integral to the success of this Retrieved March 8, 2020, from http://spiff.rit.edu/ project. classes/phys440/lectures/reduced/reduced.html This research was made possible by the Washing- Rowe, D., 2018. Gaia (DR2)Search Program. Private ton Double Star catalog maintained by the U.S. Naval Communication. Observatory, the Stelledoppie catalog maintained by Salaris, Maurizio; Santi Cassisi (2005). Evolution of Gianluca Sordiglioni, Astrometry.net, and AstroImageJ stars and stellar populations. John Wiley & Sons. software which was written by Karen Collins and John pp. 138–140. ISBN 978-0-470-09220-0. Kielkopf. This work has also made use of data from "The Eddington Limit (Lecture 18)" (PDF). the European Space Agency (ESA) mission Gaia jila.colorado.edu. 2003. Retrieved 2020-02-21 (https:// www.cosmos.esa.int/gaia ), processed by the Thornton, S., Marion, B., 2004 Classical Dynamics of Gaia Data Processing and Analysis Consortium Particles and Systems. Cengage Learning, Fifth edi (DPAC, https:// www.cosmos.esa.int/web/gaia/dpac/ tion, Chapter 8. consortium ). Funding for the DPAC has been provid- Wasson, R., Rowe, D., Russell, G., 2020 “Observation ed by national institutions, in particular the institutions of Gaia (DR2) “New” Red or White Dwarf Bina participating in the Gaia Multilateral Agreement. ry Stars in the Solar Neighborhood”, Journal of

References Double Star Observations, 16 (3), 208-228. Hopmann, J, 1967. Untersuchungen an zwolf visuellen Rengstorf, A. Lecture Notes on Orbital Determination. Dopplesternen, Mitteilungen der Universitaets- New Mexico Institute of Mining and Technolo Sternwarte Wien, Vol. 13, pp 49-88. gy, July 2019.

Caputo, R., Bonifacio, B., Datar, S., Dehnadi, S., Lian

E, Green, T., Koodli, K., Koubaa, E., Marchetti, C., APPENDIX A Nair, S., Olson, G., Perian, Q., Singh, G., Wang, C., Derivation of Escape Velocity Equation Yeung, P., Robertson, P., Chalcraft, E., Tock, K.,

2020. “Observation and Investigation of 14 Wide First we represent the position of both stars as vectors Common Proper Motion Doubles in the Washington in Cartesian coordinates with our Sun at the origin. Double Star Catalogue”, Journal of Double Star Obs This is shown in Figure 1 below, where CM is the cen- ervations, 16 (2), 173-182. ter of mass of the system, R is the position vector be- Dugan, O., Robinson, T., Carmeci, F., Tock, K., 2019. tween the Earth and the center of mass, and r, as the “CCD Measurements and Reclassification of WDS difference between r and r : r = r -r . 07106 +1543 to an Optical Double”, Journal of 1 2 1 2

Double Star Observations, 15 (1), 119-129 Duric, Nebojsa (2004). Advanced astrophysics. Cam bridge University Press. p. 19. ISBN 978-0-521- 52571-8. Harshaw, Richard, 2016, “CCD Measurements of 141 Proper Motion Stars: The Autumn 2015 Observing Program at the Brilliant Sky Observatory, Part 3”, Journal of Double Star Observations, 12 (4), 394- 399. Hopmann, J., 1967. “Untersuchungen an zwolf visuel len Dopplesternen” (“Studies of Twelve Visual Double Stars”). Mitt. Univ. Sternw. Wien, Vol. Figure 1: The primary star has a position r1 and a 13, 49 - 88. mass m1, and the secondary star has a position r2 and Hartkopf, W.I. and Mason, B.D. “Catalog of Rectiline a mass m2 relative to the origin of the coordinate ar Elements”, v 2017.09.20. system.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 423

Measurements of Neglected Double Stars

If we put the center of mass of the system as the Since the total energy of the system is zero, the po- origin of the coordinate system, the positions of the tential energy U is equal to the inverse of the kinetic stars can be redefined for simpler calculations as in Fig- energy in Equation 5. To find an expression for U, we ure 2. take the primary star as a single particle in free space. This particle is not bound by any system, so its gravitational potential energy is zero. When a new particle (the secondary star) is brought from a faraway distance to a position near the primary, the gravitational potential energy decreases because of the attraction of the two masses to each other. The magnitude of the decrease is equal to the gravitational force integrated over the distance separating the two stars as they are brought into proximity, as shown in Equation 6.

Figure 2: Primary and secondary stars relative to Equation 6: The potential energy of a star system whose second- the center of mass of the ary star, of mass m2, is brought from a faraway distance into a double star system. position close to the primary star, of mass m1. From the inertial reference frame of the center of Finding the escape velocity of this system involves mass, the following relations must hold: calculating the relative velocity of the stars subject to the constraint that the system has a total energy of zero. The derivation is shown below. Equation 1: Definition of center of mass.

Equation 2 Equation 7: The kinetic energy of the system is equal to the positive value of the potential energy

Equation 3

The total kinetic energy of this system is the sum of the kinetic energies of the individual stars as measured Equation 8: The escape velocity of the star system, where from the inertial center of mass reference frame, ex- M = m1 + m2 is the total mass of the system. pressed in Equation 4.

Equation 4 : Total kinetic energy for this system where the magnitudes of vectors are shown without arrows and time derivatives are designated by dots.

Plugging the relations of Equations 2 and 3 into Equation 4, we obtain the result shown in Equation 5.

Equation 5: Total kinetic energy for this system in terms of the relative velocities of the stars.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 424

Astrometric Measurement of WDS 01335-0331

Kyle Adams, Britton Baltich, Vanessa Wright, and Hajdi Isufi

Brigham Young University – Idaho Rexburg, Idaho, United States

Abstract: The double star WDS 01335-0331 was measured using a 0.4 meter telescope. A set of 10 10 second exposures was taken and we calculated an average separation of 5.28 arcseconds and an average positional angle of 111.8 degrees.

Introduction known. These additional parameters were chosen due Research was conducted on the double star WDS to our hope of estimating the spectral type and deter- 01335- 0331 in the constellation of Cetus. The star was mine if it is an optical double or a physical binary. chosen from the Washington Double Star Catalogue of To find the star, the Gaia Double Star Catalog Se- neglected double stars. The system was observed to lection Tool (GDS Catalog Selection Tool) created by determine if an orbital path exists as well as the spec- Dave Rowe, was downloaded and used [2]. The entire- tral types. This binary was initially observed in 1827 ty of the Washington Double Star and Gaia catalog is by John Herschel (discovery number HJ 640) and last included in the program. observed in 2017. In 193 years, there have been 30 rec- Once downloaded, the first four parameters were orded observations of this system. put into the GDS Star Catalog, resulting in 186 double Assuming that the stars are main sequence, the stars. After excluding stars without the required paral- masses of the stars can be calculated using the mass- lax and those with catalogued spectral types, four luminosity relationship. If the luminosity and effec- choices remained. Of the four, WDS 01335-0331 was tive temperature are known or calculated, then the selected based on the quantity and timing of the previ- spectral type can be estimated by looking at a Hertz- ous observations. Using the WDS name, the website sprung–Russell diagram. Stelle Doppie [3] was used to find that the formal classification of the system was unknown. Methods and Materials Telescopes of the Las Cumbres Observatory The reason WDS 01335-0331 was picked was be- (LCO) were used for our measurements [4]. Ten imag- cause its properties matched our seven parameters suit- es were taken for exposure times of 2, 3, and 4 sec- ed to making astrometric measurements with a small onds. This was done through a SBIG STL-6303 cam- telescope. The parameters included magnitudes be- era with a clear filter on the Haleakala tween 7 and 11, a separation between 5 and 8 arcsec- 0.4 meter telescope in Hawaii [5]. Unfortunately, it onds, and delta magnitude of less than 1 to ensure that was realized that the images that were received needed the double star could be resolved into two stars. Fur- longer exposure times. Exposure times of 10 seconds thermore, a right ascension (RA) between 1 and 13 was for 10 pictures (see Figure 1) were used on the second chosen to ensure the star could be seen during the ob- attempt. serving time frame. A parallax of 5 milliarcseconds Once the images were received and converted into (mas) or higher was also chosen because parallax a format the program AstroImageJ (AIJ) [6] could use, measurements below this number were not reliable [1]. measurements were taken and put into a spreadsheet. Only doubles that had 50 observations or less, and dou- The measurements important to our research are seen ble star systems with observations that were spread in Table I. Historical data was used in conjunction with throughout time were considered. The search was fur- Harshaw’s Plotting Tool to produce Figure 2. ther refined by looking for a system with an unknown spectral type and the nature of the double was un-

Received September 14, 2020

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 425

Astrometric Measurement of WDS 01335-0331

Data Points SEP(arcsec) PA(Deg)

Average Standard Deviation ±0.02 ±0.4 Mean Error ±0.007 ±0.1 Figure 1: The double star WDS 01335-0331 HJ 640 Table 1: Astrometric measurements. taken with an exposure time of 10 seconds. This image was taken on the Haleakala Observatory in Hawaii on the is- land of Maui. Discussion Results Our data supports previous measurements of the The data we collected from our observations (see separation being around 5.32”. Our average PA of Table I) gave us an average separation (SEP) of 5.27” 111.8° with an error range of 111.4° to 112.2° also and an average position angle (PA) of 111.8°. The aver- supports previous observations of the position angle age of our observational data (see Table I) was used to being around 112.01° [3]. calculate an additional data point represented by the When comparing the two parallaxes from Table 2, triangle on Figure 2. The other data-points were provid- the primary has 10.37 milliarcseconds (mas) with an ed courtesy of Dr. Brian Mason at the United States error of 10.33 mas to 10.41 mas while the secondary Naval Observatory (USNO). has 10.44 mas with an error of 10.34 mas to 10.54 mas. When taken into account, the two parallaxes overlap. Furthermore, using Harshaw’s Plotting Tool [7] in conjunction with WDS Gaia Data Release 2 Version 3B [8], it is found that the two stars have a possible weighted separation of 511 AU. This separation could mean that the system is gravitationally bound, but the exact alignment of the two stars to the Earth’s line of sight is unknown. The actual separation could be different and therefore the system could be two separate stars. The orbital velocities would also need to be known but the necessary data for calculating this is yet to be documented by a future mission [9]. Referring to Table II, the RA Proper Motion (PM) of the primary was 39.7 mas/year(yr) with an error of 39.6 mas/yr to 39.8 mas/yr. The secondary star has a RA PM of 39.6 mas/yr with an error of 39.4 mas/yr to 39.7 mas/yr. The ranges of the PM overlap, meaning that the stars could be moving together. According to Harshaw, “Two stars that are in orbit around one another should have identical, or very nearly identical, proper motions. Large differences in proper motion would sug- Figure 2: Cartesian plot of the double stars positions gest the stars are not gravitationally related” [10]. with added observation represented by the triangular point. Therefore, because the PM overlap, it is once again The horizontal vector represents how much the stars should likely that system is gravitationally bound. Additional- have moved over the time that they have been observed. The ly, the two vectors in Figure 2 are almost the exact angled vector represents how much the stars have moved. The magnitudes have been multiplied by 2 for ease of same magnitude but have a difference in direction of 37 viewing. degrees. According to Harshaw’s Instructions in his

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 426

Astrometric Measurement of WDS 01335-0331

plotting tool [7], there is likely no problem with Gaia’s Pulling the above information together, the star sys- PM data for this pair, which means the data can be tem WDS 01335-0331 is likely a physical binary. The trusted for PM. parallax’s overlap, the separation between the stars is The Declination (DEC) PM of the primary is -79.9 likely small enough for them to be gravitationally mas/yr with an error of -79.6 mas/yr to -79.7 mas/yr for bound (orbital velocities still need to be calculated) and the primary star. The secondary star has a DEC PM of - their PM’s are nearly identical. Additionally, in the 78.9 mas/yr with an error of -79.0 mas/yr to -78.8 mas/ WDS Gaia Data Release 2 Version 2 is a probability yr. The DEC PM do not overlap but they are close. function done by Richard Harshaw to give the probabil- Based on both stars luminosity’s, they appear to be ity that a double star is a physical binary. It relies on main sequence stars. Using the masses and separation, four weight factors of parallax, PM, R2, and radial ve- the approximate orbital period can be calculated using locity with parallax being assigned the highest relative the Distance PA Calculator [8] (approximate because it weight [10]. The double star WDS 01335-0331 proba- is unknown if the star is at its periastron or apastron bility of being a physical binary is 89 percent. position within its assumed orbit). The orbital period is Primary Star Secondary Star greater than or equal to 8,900 years for the reasons dis- Parallax cussed in the previous sentence (see Table II). (arcsec) The stars were plotted on the Hertzsprung-Russel Parallax Error Diagram (HR Diagram) using the solar luminosity and (arcsec) Proper Motion RA effective temperature. Upon viewing the plot (see Fig- 0.0397 0.0396 (arcsec/yr) ure 4), one can see that both stars fall within the spec- PM Error tral classification of K0V, which means both are orange (arcsec/yr) dwarf main sequence stars. Proper Motion -0.0797 -0.0789 The data presented by Figure 2 further indicate that Dec (arcsec/yr) the orbital period is extensive as there is no drastic PM Error change in the position of the stars, excluding outliers, (arcsec/yr) Radial Velocity with subsequent observations. It is assumed that a trend 25.6 NA would be visible if the data was gathered over a signifi- (km/s) cantly longer time period (a couple of centuries), for RV Error (km/s) 0.52 NA Temperature (K) 5280 5240 more accurate data analysis as well as providing time 0.53 0.44 for the stars to change orientation to a more significant Calculated Values from Gaia Data degree. Based on the approximate orbital period of Solar Mass 0.86 0.82 8,900 years, more data is needed to clearly establish WDS 01335-0331 as a physical binary. SEP (arcsec) 5.32 PA (Deg) 112.01 Orbital Period (yrs) 8900 Table 2: Gaia data. Conclusion The double star WDS 01335-0331 was chosen from a set of parameters determined by the team in hopes to classify the type of double. Images were taken and re- taken with the LCO 0.4 meter telescope in Hawaii. Af- ter analyzing the data from these images and calculating error, our team was able to match this data with the Gaia data. Certain tools were used to plot the cartesian positions of the double star, reliability of PM data, separation of stars, solar masses, approximate orbital period, SEP, PA, and probability of double star being a Figure 3:. Using the HR Diagram [11] to ap- physical binary. The PM, the parallax, and the separa- proximate the spectral type of both stars. The addition point to WDS 01335-0331 possibly being a physi- tion symbol (+) is the primary star while the small cal binary and with a probability of 89 percent. Ulti- multiplication symbol (x), is the secondary star. The mately, it can not be concluded that the star system is a stars are similar in temperature and luminosity

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 427

Astrometric Measurement of WDS 01335-0331

physical binary but it is probably one. Acknowledgements We thank Rachel Freed for her help in providing the links and information that was needed to complete this research. Thanks goes to Dr. Stephen McNeil for his guidance and instruction that led this team to the Double Star that was observed and studied. A special thanks to Dr. Brian Mason for providing historical data. We thank the Las Cumbres Observatory for allowing our team time on one of their 0.4 meter telescopes and for providing the images used. References [1] X. Luri and et al., Gaia data release 2-using gaia parallaxes, Astronomy and Astrophysics 616, 19 (2018). [2] D. Rowe, Dave Rowe’s Gaia Double Star Selection Tool, Drop- box. [3] 01335-0331 hj 640, Stelle Doppie. [4] Observation portal, Las Cumbres Observatory (2020). [5] Las Cumbres Observatory: 0.4-meter (2020). [6] Astroimagej for astronomy (2019). [7] In publication. [8] Personal correspondence. [9] F. Rica and A. Brown, “Determining the Nature of a Double Star: The Law of Conservation of Energy and the Orbital Velocity”, Journal of Double Star Observations 7, 254 (2011). [10] R. Harshaw, Gaia dr2 and the Washington Double Star Catalog: A tale of two databases, Journal of Double Star Observations 14, 734 (2018). [11] R. C. J. Erickson and R. R., Hertzsprung-russell diagram., Salem Press Encyclopedia of Science (2018).

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 428

Analysis of an Unmeasured Component of WDS 12290+3735 KZA 36

Richard L. Nugent, Ernest W. Iverson

International Occultation Timing Association [email protected] [email protected]

Abstract: Our recent observations of the system WDS 12290+3735 KZA 36 identified the presence of a previously unmeasured component. Historical images from the Palomar Obser- vatory Sky Survey and the Two Micron all Sky Survey images show this component merged with the primary star. A new image acquired shows this component separated from the primary star. Analysis of Gaia proper motions also show similar directional movement of this component and the primary star, however the Gaia parallaxes indicate the distance between the two stars exceeds 31 parsecs.

Introduction The multiple star system WDS 12290+3735 KZA 36 is listed in the Washington Double Star Catalog with 3 components, A, B and C. Author Nugent measured this system in April 2019 (Nugent and Iverson 2020) and he noticed a previously unreported component faintly visible several arc-seconds from the primary star A. We propose the identifier “D” for this previously unmeasured component. We decided not to pub- Figure 1. Left: POSS-II IR image, 15 April 1998. Right: lish the position angle and separation for the AD pair in 2MASS image, 2001. our 2020 paper pending further analysis. Methodology

Examining old astronomical plates can often be a

very useful source of historical double star information.

Archive images from the Palomar Observatory Sky

Survey (POSS-II) and the Two Micron All Sky Survey

(2MASS) are shown in Figure 1. Compared to the

round images of the B and C components, the A com-

ponent has an oval, non-spherical shape. This oval

shape is most noticeable in the 2MASS image. Images

of merged stars are frequently seen in both the POSS-II Figure 2. KZA 36, CCD image courtesy Dr. Allen Gilchrist. Image and 2MASS survey plates. In the case of KZA 36, satu- acquired 26 June 2020 ration from the brighter primary star in Figure 1 makes it appear joined with the D component. The European Results and Discussion Space Agency’s (ESA) Global Astrometric Interferom- New observations of the KZA 36 system were eter for Astrophysics (Gaia) satellite lists the magni- made using a 14-inch telescope with a plate scale of tude of the proposed D component at +13.54 and the 0.2"/pixel. Our measurement of the AD pair for the primary star A at +9.22. The KZA 36 system is shown epoch 2020.49 is PA = 237.7° and separation = as Figure 2 at the epoch 2020.49. 5.9" (see Table 1).

Received July 4, 2020

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 429

Analysis of an Unmeasured Component of WDS 12290+3735 KZA 36

PA PA Sep Sep Date Mag A Mag D (°) SEM (″) SEM WDS 12290+3735 AD 237.7 0.94 5.9 0.08 2020.49 9.22 13.54

Table 1: Position angle and separation for AD component.

If the A and D components of KZA 36 comprise a why the WDS measurements from 1928 – 2015 (10 true binary system, the proper motions and distances of published records) did not include the D component? the stars should be similar. With the recent release of Was the D component behind or in front of the A com- the Gaia satellite data archive, the most accurate proper ponent until recently? Did the large magnitude differ- motions and distances are now available. For KZA 36 ence (Δm = 4.4) discourage observers from measuring A, B, C and our proposed D component we used the the D component or was it simply too faint to be visi- Aladin Sky Atlas program to overlay the Gaia proper ble? motions on the POSS-II 1998 image. Figure 3 shows We combined the Gaia proper motions of both the A the A and D components have similar proper motions in both direction and magnitude. Gaia’s proper motion error estimates for the A and D components are 4% and 9% respectively. Based on proper motion data alone it appears that the AD pair characterize a true binary system.

Figure 4. Law of cosines calculation for the distance between the AD components.

and D components to determine relative movement between the two stars. The net result shows that the D component is moving westward away from the A component by 0.011"/yr. This corresponds to 1.1 AU’s per year at its distance. In 1928 when this system was first measured, the calculated AD separation was 4.9" and in 1956 when the first POSS images were exposed, the calculated AD separation was 5.2". The Gaia radial ve- Figure 3. Gaia proper motions overlaid on the KZA locities of A and D are -3.8 and -3.5 km/sec respective- 36 components. POSS-II image from 15 April 1998. ly. This corresponds to their actual motion toward Earth of 0.8 AU's /yr. Both the A and D component’s net mo- Using Gaia parallax data, we derived the distances tion away from each other and toward Earth is insignifi- between the components A and D. From Figure 4 we cant. know their individual distances from Earth (sides a and Thus, we conclude that the A and D components of b of the triangle). Using our measurement for the sepa- KZA 36 form an optical alignment and not a physical ration angle θ between the A and D components, we binary pair. With the large distances between the AB applied the law of cosines to solve for the distance be- (518 ± 22 pc) and AC (156 ± 4pc) components, both tween A and D (side c): of these pairs also represent unrelated stars. Figure 5 shows a distance diagram for the KZA 36 system. 2 2 c = a + b − 2abcos The AD distance is 43 ±12 parsecs. This large error is due to the D component’s 12% parallax error. The Gaia parallaxes and resulting distances of the components are shown in Table 2. The obvious question is

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 430

Analysis of an Unmeasured Component of WDS 12290+3735 KZA 36

Parallax Distance Parallax Distance RA DEC Component Mag Error Error (mas) (parsecs) (mas) (parsecs) 12h 28m 57.4s +37° 35’ 41.0” A 9.22 6.898 0.042 145 1

12h 28m 01.8s +37° 35’ 16.4” B 12.68 1.507 0.048 663 21

12h 28m 53.2s +37° 36’ 47.9” C 12.87 3.323 0.039 301 3

12h 28m 57.0s +37° 35’ 38.0” D 13.54 9.778 1.150 102 12

Table 2: Gaia Position, magnitudes, parallaxes and distances of KZA 36 A,B,C components plus new D component.

the California Institute of Technology with funds from the National Science Foundation, the National Geo- graphic Society, the Sloan Foundation, the Samuel Os- chin foundation, and the Eastman Kodak Corporation.

References Nugent, R. and Iverson, E. 2020, Journal of Double Star Observations, 16, No. 2, 141-147

Figure 5. Projection of Gaia distances for the KZA 36 components. The proposed component D is just a chance optical alignment with A component. The unlabeled star below component C (Gaia mag = +15.5) is UCAC4 638-047996. Its Gaia distance is 1392 ± 83 pc. Acknowledgements Dr. Allen Gilchrist kindly provided the recent CCD image of the system KZA 36 used as Figure 2. This work has made use of data from the European Space Agency (ESA) mission Gaia (https:// www.cosmos.esa.int/Gaia), processed by the Gaia Data Processing and Analysis Consortium, DPAC, (https:// www.cosmos.esa.int/web/Gaia/dpac/consortium). We acknowledged data from the Washington Double Star Catalog maintained at the US Naval Observatory. Also acknowledged is the Aladin Sky Atlas Interactive software program and the VisieR catalog database from the Center de Données Astronomiques in Strasbourg, France. We have used images from the Two Micron All Sky Survey, (which is a joint project of the University of Massachusetts and the Infrared Processing and Anal- ysis Center/California Institute of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation). The second Pal- omar Observatory Sky Survey (POSS-II) was made by

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 431

Astrometric Measurements of Double Stars LEP 57 and LDS 1317

Douglas Edwards¹, Vagmi Padmanabham¹, Cooper Knightly¹, Madison McAfee¹, Pat Boyce²

1. High Tech High, San Diego, CA 2. Boyce-Astro Research Initiatives and Education Foundation

Abstract: We performed astrometric measurements of the AB (LEP 57) and AC (LDS 1317) pairs of star system WDS 12323+1335 using images taken with the Las Cumbres Obser- vatory (LCO). We used a 0.4-meter Meade telescope to investigate both pairs. For LEP 57, we aimed to validate the existence of the neglected and rarely observed B star, which had not been measured since epoch 2000. The images taken indicated that the B star may not exist, and that the two previous measurements of the AB pair are unaccountable. For LDS 1317, a mean The- ta and Rho of 351.2° and 12.5” respectively were calculated. The AC pair is likely physical.

Introduction Locating the B star of LEP 57 through Aladin 10 Double Star WDS 12323+1335, Figure 1, is a dou- proved difficult, and the focus of the research was thus ble star system with recorded observations of an AB, shifted from simply updating older data to further ex- LEP 57, and an AC pair, LDS 1317. LDS 1317 was ploring the existence of LEP 57 AB and consequential- first discovered in 1962 as a two-star pair with a sepa- ly updating the data if the AB pair indeed existed. Ad- ration of 13”. In 1998, another observation found a ditionally, LDS 1317 would be analyzed to determine new star at a separation of 5.23”. This new star re- if it was a binary pair or simply an optical pair. placed the original B star, and the original B was rela- beled as the C star (Mason, 2012). The system was initially chosen as a research target due to LEP 57 having only two observations with the last occurring in 2000, by Sebastien Lepine. His observation reports that the system is a triple star and includes position angle and separation measurements for LEP 57. LDS 1317 has seven unique observations in the WDS and was last observed in 2015 as part of a GAIA sky survey. A review of the pair’s records in GAIA additionally showed consistently similar proper Figure 2: Both the A and C star have very high and very similar motion (PM) measurements, Figure 2. proper motion, as seen in Aladin 10 with a GAIA proper motion overlay.

Methods and Materials Equipment The images were taken using the Las Cumbres Ob- servatory (LCO) system. The primary telescope used is located at the McDonald Observatory in Texas, and additional images were taken from the Haleakala Ob- servatory in Hawaii, USA. Both presented a minimal air mass value at the scheduled observation angle. Each telescope has eight filters, a 29.2 x 19.5 arcminute Figure 1: Aladin 10 image of system 12323+1335 as well as the FOV, and an SBIG STL-6303 CCD camera. predicted location of the B star.

Received September 20, 2020

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 432

Analysis of an Unmeasured Component of WDS 12290+3735 KZA 36

Figure 3: AIJ Image showing LDS 1317 AC.

Observations A total of 16 images were taken using the Pan- Discussion STARRS-w filter on April 18th, 2020 with four differ- AB Pair ent exposure times: six, eight, ten, and twelve seconds. In the images acquired, the B star was not visually Both the Hawaii and Texas observatories were used. discernible within or around the previously reported The images were processed through the Our Solar separation of 5.23” from the A star while using the Pan- Siblings (OSS) Pipeline. The OSS Pipeline (Fitzgerald STARRS-w filter. Four images were also taken using 2018) processes images in multiple phases by cleaning, the Sloan r filter, which includes some of the infrared labelling, plate solving, and calibrating all images to spectrum, and yet the B star was not visible in those optimize them for later analysis. After OSS Pipeline images either. Additionally, the area around the A star processing, images were measured with AstroImageJ. was examined for signs of a B star within a 30” radius. The Distance & Angle tool was used to accurately lo- As there were no prior measurements of the B star’s cate the stellar center of each star, and measure the po- proper motion, it was impossible to perform a more sition angle and separation. Once all measurements accurate search, and no candidate was found with simi- were made, data was organized in an excel spreadsheet lar magnitude to the B star in the immediate area. to determine mean values as well as standard deviation In addition to it not being present in our images, and standard error of the mean for the complete image GAIA was unable to perform a parallax measurement set. of the B star nor identify any viable star within proxim-

Results ity of the A star. DSS images show that the A stars Table 1 displays our calculated mean position magnitude of 9.92 V and angular radius of ~6.5” are angle (Θ) and angular separation (ρ) of LDS 1317 great enough to feasibly obstruct view of the B star entirely. When using a data layer provided by the NASA/ AC, as well as the standard deviation and standard IPAC Extragalactic Database (NED) while observing error of the mean for the pair. The delta magnitude the system in Aladin 10, it was found that an infrared was also measured, and the mean of these meas- source is recorded to be in the exact location of the B urements was calculated at 6.1935. It was not pos- star’s last known position, Figure 4. In WDS, the B star sible to measure LEP 57 AB, as the B star was un- has a “K” note on it, indicating that it’s magnitude was observable in all images and using all filters. measured in the infrared band. Additionally, 2MASS, a

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 433

Analysis of an Unmeasured Component of WDS 12290+3735 KZA 36

WDS 12323 + 1335 LDS 137 Telescope(number of images taken per Epoch:2020.297 Θ(Degrees) ρ(arcseconds) exposure count using the W filter

Mean: 351.200 12.334 kb 88:(4,6 sec),(4,8 sec) kb 82:(4,10 sec),(4,12 sec) Standard Deviation: 1.007 0.101

Std. Error of Mean: 0.252 0.025 2015.15 (Last measurement prior to 351.201 12.571 this investigation)

Table 1: Table displaying filters used and the calculated average values of our measurements.

sky survey using a filter within the infrared spec- LDS 1713 AC trum, has a record of the B star, and is responsible for its only magnitude measurement of 14.3 using the H Epoch: Θ(°) ρ(arcseconds) filter. Due to a lack resources it was impossible to make 1962.22 352.0 13.000 a recording in the infrared band, but it was noted that as the angular radius of the A star is greater than the sepa- 1998.04 351.3 12.380 ration of the B star, it’s possible the B star would have 2000.00 351.7 12.300 been impossible to observe, even when using an infrared filter. Considering the lack of consistent infor- 2001.10 351.0 12.465 mation on the B star as well as the uncertainty of its 2012.40 352.1 12.470 physical status according to multiple unique sky surveys, there is not enough evidence to confirm the exist- 2015.00 351.2 12.516 ence of the B star. 2015.50 351.2 12.517

Table 2: Historical WDS measurement data of the AB pair.

Parallax Error

Star A 12.6031 0.0671

Star B Not Reported Not Reported

Star C 12.5329 0.0508

Figure 4. Aladin 10 image of system 12323+1335 taken through Table 3: Measured parallax values of system the NED database. 12323+1335, note the lack of data for the B star due to its invisibility. AC Pair The measurements made of LDS 1317 are consistent The Harshaw Statistic Calculator (Harshaw, 2014) with those made previously, seen in Table 2. The most was used to analyze the proper motion measurements of recent observation, a GAIA measurement for epoch the A and C star, and quantify the vector difference be- 2015.5, recorded LDS 1317 with a Theta and Rho of tween the two. In Harshaw’s paper, he states that there 351.2° and 12.5” respectively, both being well within is a clear association between low vector differences the margin of error relative to the measurements made and gravitational pairs, so a pair with a low vector dif- by our group. With values this similar to the records, ference, for example a value of 0.1, is likely to be a our Theta and Rho measurements appeared to be accu- physical pair. When the two proper motion measure- rate and reliable. A review of the parallax for both ments of the AC pair were entered, the result was a vec- showed a similar measurement of 12.6031 for the A tor difference of ~0.005, an extremely low value. With star and 12.5329 for the C star, Table 3. such a slight vector difference, it’s likely that the A and C stars are a physical pair.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 434

Analysis of an Unmeasured Component of WDS 12290+3735 KZA 36

The parallaxes of the A and C stars were used to fore we started our research was performed by GAIA in calculate the minimum separation distance between 2015. Both of these stars have a very similar parallax them. By applying the small angle approximation, their with parallax measurements of 12.6031 for the A star minimum separation at the time of our observations and 12.5329 for the C star. Using the Harshaw Statistic was 0.155 light years perpendicular to our line of sight. Calculator to analyze our proper motion measurements, The difference in the mean distance to the A and C the resulting and extremely low vector difference value stars in the radial direction from the Gaia DR2 data is of ~0.005 is a factor in determining the gravitational 1.45 light years. The errors in these measurements are relation of the A and C stars. This shows us that they relatively small due to the nearness of these two stars, are moving together which supports the assumption that roughly 80 parsecs. There are competing views on the the AC pair may be gravitationally related. Finally, we maximum spatial separation allowed for two stars to be calculated that there is a 32% probability that this pair gravitationally bound: one light year and three light is within one light year of each other in three dimen- years depending largely on their masses. We employed sional space, and an 81% chance that they fall within a Buchheim’s method of determining the probability den- three light year distance, indicating a significant chance sity function of their separation in the radial direction that they are gravitationally bound. This myriad of evi- based on the Gaia data. We found that there is a 32% dence leads us to conclude that LDS 1317 AC is most likelihood that A and B are within one light year of likely a gravitationally bound pair. each other and 81% likelihood of being within three light years apart radially. Acknowledgements Previous records of LDS 1317 AC support the pos- The authors thank the Boyce Research Initiatives sibility that the system is a physical pair, and our meas- and Education Foundation (BRIEF) for providing us urements supported their physical status further. With with access to their remote server, educational videos, highly similar proper motion and parallax in addition to and their constant support and guidance throughout the the close separation of less than 13” that has persisted research and writing process. We would also like to since 1962, it is highly likely that LDS 1317 is a physi- thank the United States Naval Observatory, particularly cal pair. Additionally, our calculations of the vector Brian Mason, for providing historical double star data, differences and probability of gravitational relation are and the Las Cumbres Observatory for providing us ac- highly reliable, tested methods that continue to support cess to their sites and equipment. Additionally, we give the proposition that LDS 1317 AC is a physical pair. thanks and recognition to Bob Buchheim for providing his methodology on determining gravitational relation- Conclusion ships of double stars, and for, as he puts it, “personal From the observation, two conclusions were communication”. We thank Pat Boyce for his invalua- formed about this star system. First, for LEP 57, it was ble guidance and mentorship throughout the writing of determined that the B star couldn’t be confirmed. Over- this paper. We would also like to thank Brian Delgado all, the recorded data of the B star was generally lim- at High Tech High for presenting this research oppor- ited, likely due to the luminosity of the A star, which tunity to all of us. This paper was supported by the restricted data gathering on the B star. There was also Boyce Research Initiatives and Education Foundation some conflicting data on the arc-second separation for (BRIEF). the closest star to the primary star. For example, SIM- BAD stated that the separation of LEP 57 AB was References 5.23”, but the GAIA catalog shows that the closest star Fitzgerald, M.T. (2018, accepted), "The Our Solar is 12.5” away, the separation value of the C star. In ad- Siblings Pipeline: Tackling the data issues of the dition, the 2MASS sky survey was able to detect the B scaling problem for robotic telescope based astrono- star using an infrared bandwidth filter, yet it is the only my education projects". Robotic Tele scopes, Student source to do so. The search for the B star did not reveal Research and Education any other star resembling it from the previously record- Proceedings. ed 5.23” or 30” radius around the primary star. Because Harshaw, Richard. 2014, “Another Statistical Tool for of all these past issues and our inability to get a clear Evaluating Double Stars”. Brilliant Sky image of the B star, it wasn’t possible to confirm the Observatory Cave Creek, Arizona. Journal of star’s existence or its lack thereof. Double Star Observations, 10 (1), 32. Regarding LDS 1317, we concluded that there is a Las Cumbres Observatory (LCO), 2020, 0.4 Meter, high probability that the AC pair is a physical pair. The JPEG, Las Cumbres Observatory, Global Head- last observation and documentation on the AC pair be- quarters California, https://lco.global/observatory/telescopes.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 435

Analysis of an Unmeasured Component of WDS 12290+3735 KZA 36

Mason, Brian. et al, 2012, Washington Double Star Catalog, Astronomy Department, United States Na val Observatory, http://ad.usno.navy.mil/proj/ WDS/ Salgado et al. (2017): Gaia Data Release 2. Gaia achive data access facilities; European Space Agency (ESA) mission Gaia, https://www.cosmos.esa.int/ web/gaia. SIMBAD Astronomical Database. Unistra/CNRS.2018. Web https://simbad.u-strasbg.fr/simbad/sim-basic? Ident=12+32+20.77+% 2B13+35+01.8&submit=SIMBAD%20search

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 436

Astrometric measurements for double stars not found in the WDS

Roberta Bonnell, Tyce Olaveson, Jakob Bergstedt, Dallin Fisher, Stephen R. McNeil

Brigham Young University-Idaho, Rexburg, Idaho

Abstract: While in pursuit of a suitable target to study, two previously unrecorded visual binary star systems were discovered and observed. One of these binary stars is most likely an optical double, whereas the other binary has strong supporting evidence to be a true binary system. A minimal orbital period is calculated for this system.

Introduction meter telescope located at the Siding Spring Observa- In seeking out neglected double-stars of an uncer- tory in Australia, using an SBIG 6303 camera with a tain nature, images were taken of the binary system clear filter. The pairs are located close together in the WDS 1108-6140 using the Las Cumbres Observatory sky, and so we were able to take ten, 30 second expo- (LCO) network. Upon analyzing the system, we found sures (centered on Double 2) to capture both binaries. that the stars were too close together to take accurate From these images we measured the position angle and measurements with AstroImageJ. Almost by accident, separation for each pair using AstroImageJ. Examples two nearby pairs of stars piqued our interest as they of these images are shown in Figures 1 and 2. were visually close to their companion star, and had similar magnitudes within each pair. We checked the data in Stelle Doppie and Aladin, and found that these stars are in close proximity to each other in terms of parallax. Determined to study these pairs, we scoured Stelle Doppie to find the star coordinates and historical data only to find that no star with these coordinates had been reported. Searching the Washington Double Star (WDS) catalogue resulted in a similar conclusion: these stars have not yet been observed as potential double star systems. This research observes the position angle (θ) and Figure 1. Image of Double 1 taken from the Siding Spring separation (ρ) between these pairs of stars and will en- Observatory on May 28 2020, boxed in red . deavor to calculate minimal orbital periods should data indicate they may be physical binaries and not optical doubles. Following are the primary stars’ coordinates, GAIA DR2 IDs, and the way we will reference them in this paper: Double 1 RA 11:06:30.6 DEC -46:10:28.97 GAIA DR2 ID: 5386565914691315968

Double 2 RA 11:06:34.7 DEC -46:11:55.3 GAIA DR2 ID: 5386565120119495296 Figure 2. Image of Double 2 taken from the Siding Spring Equipment and Methods Observatory on May 28, 2020, boxed in red We gathered new CCD images from the LCO 0.4-

Received September 23, 2020

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 437

Analysis of an Unmeasured Component of WDS 12290+3735 KZA 36

add more evidence in regard to this hypothesis, though Results such measurements will likely be taken with a new mis- The following data was collected using As- sion. troimageJ. The GAIA data for Double 1 does not rule out the The position angle and separation of each double possibility of it being a true binary system since the was measured multiple times on each image to assure stars have similar parallaxes and proper motions. consistency and the individual measurements are shown Hence, the following calculations will be for Double 1 in Table 1. only. Double 1 Double 2 To calculate a minimal orbital period for the po- Image tential system we make use of the following properties: θ (deg) ρ (arcsec) θ (deg) ρ (arcsec) number the distance between the two stars in AU’s, and the 1 154.78 7.04 72.57 11.1 masses of each star. The distance is calculated through simple geome- 2 155.44 6.98 72.84 11.04 try. First, we use the parallax angle in arcseconds to 3 155.17 6.97 72.45 11.09 calculate the distance (D) to our primary star: 4 155.21 6.96 72.6 11.09 5 154.45 7.00 72.78 11.05 6 154.92 6.97 72.69 11.13 7 154.66 7.02 72.68 11.02 This gives us 487.1 parsecs to our primary star. 8 154.95 6.98 72.72 11.09 Now using the separation angle in radians and the dis- 9 155.25 7.05 72.65 11.06 tance to the primary we can calculate the distance be- 10 154.93 6.98 72.72 11.11 tween the two. Mean 154.98 7.00 72.67 11.08 퐷푆 푝푐 = 퐷 푡푎푛 휌 Std. Dev. 0.30 0.031 0.11 0.034 −5 퐷푠 = (487.1푝푐) 푡푎푛(3.39 × 10 rads) Std. Err. 0.09 0.01 0.03 0.01 of Mean This would put the separation between the two stars at a distance of ± .00017 parsecs or 3400 ± 36 AU. Table 1. Set of data collected by measuring the Position Angle and Separation through AstroImageJ. The next step will be to calculate the masses. We will be using the mass luminosity relationships from In addition to collecting these data we also re- Eker, et. al.: viewed the available GAIA data to draw additional conclusions concerning the nature of the stars in question. The data is contained in Tables 3 and 4. The spectral 푙표푔 퐿 ∕ 퐿ʘ = 4.841 푙표푔 푀 ∕ 푀ʘ − 0.026 type indicated in the tables was estimated from plotting the luminosity and surface temperatures on an HR Dia- 푙표푔 퐿/퐿ʘ +0.026 gram. We also computed the position angle and the sep- 푀 = 푀ʘ10 4.841 aration based on the GAIA data, shown in Table 2, to assure our measurements are accurate. Using this formula and luminosity data from GAIA Double 1 Double 2 the masses of each star can be calculated. θ (deg) ρ (arcsec) θ (deg) ρ (arcsec) 154.87 7.04 72.61 11.19 Primary Star Secondary Star Table 2. θ and ρ based on data from GAIA.

Analysis 푀푃 푀푆 Based on the GAIA data it seems unlikely that 푙표푔 0.353 +0.026 푙표푔 0.195 +0.026 = 푀 10 4.841 = 푀 10 4.841 Double 2 is a physical binary system, and the method ʘ ʘ we will be using to calculate star mass will not work on 푀 = 0.816푀 푀 = 0.722푀 this system due to them being off of the main sequence. 푃 ʘ 푠 ʘ Further study with more precise measurements could

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 438

Analysis of an Unmeasured Component of WDS 12290+3735 KZA 36

GAIA Source ID 5386565914691315968 5386565914691315712

Reference epoch (yr) 2015.5 2015.5 Right ascension (deg) 166.63 ± 0.02 166.63 ± 0.02 Declination (deg) -46.17 ± 0.02 -46.18 ± 0.02 Parallax (mas) 2.053 ± 0.022 2.015 ± 0.030

Proper motion: right ascen- -20.745 ± 0.035 -20.87 ± 0.47 sion (mas/yr)

Proper motion: declination 5.050 ± 0.034 5.011 ± 0.045 (mas/yr)

Effective temperature (K) 4920 ± 80 4790 ± 120

Lum (solLum) 0.353±0.010 0.195±0.008 Spectral Type (estimated) K1 V K4 V

Table 3. GAIA data for the primary and secondary stars in Double 1.

GAIA Source ID 538656512119495296 538656515877768800

Reference epoch (yr) 2015.5 2015.5 Right ascension (deg) 166.64 ± 0.02 166.65 ± 0.017 Declination (deg) -46.20 ± 0.02 -46.20 ± 0.02 Parallax (mas) 0.284 ± 0.030 0.236 ± 0.026

Proper motion: right ascen- -4.841 ± 0.047 -7.074 ± 0.038 sion (mas/yr)

Proper motion: declination 3.200 ± 0.043 1.348 ± 0.036 (mas/yr)

Effective temperature (K) 4450 ±210 4680 ± 200

Lum (solLum) 88.57±15.90 26.29±7.25 Spectral Type (estimated) B8 III A0 III Table 4. GAIA data for the primary and secondary stars in Double 2.

Now that we know the masses of the two stars in Should this be a physical binary the above would solar masses and the distance between them in astro- be the minimum orbital period which assumes the stars nomical units, we can calculate the minimal orbital pe- are at their maximum separation. riod using Kepler’s Third Law: Conclusion 푎3 퐴푈 In this study we have located two possible binary 푃2 푦푟푠 = 푀 + 푀 systems and measured each system’s position angle and 푃 푆 separation. Upon dissection of the available GAIA data, we have determined that Double 2 is most likely an op- 3406퐴푈 3 푃 푦푟푠 = tical double, though one that still warrants research. We 0.816푀ʘ + 0.722푀ʘ have also shed light on what may be a physical binary based on current GAIA data and our measurements. We 푃 ≈ 160,000 푦푟 have also calculated a minimum orbital period in order

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 439

Analysis of an Unmeasured Component of WDS 12290+3735 KZA 36

to give some idea of time frames for future significant measurements. Future studies of these stars should go towards collecting data for orbital parameters in the case of Double 1 and confirming the nature of Double 2. These are two sources of data that can be added to the vast collective knowledge of the WDS. Acknowledgements This work has made use of data from the European Space Agency (ESA) mission Gaia (https:// www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, https:// www.cosmos.esa.int/web/gaia/dpac/consortium). Fund- ing for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement. This research has made use of the “SIMBAD Astronomical Database,” and the "Aladin Sky Atlas," operated at CDS, Strasbourg Ob- servatory, France. This research has made use of the Washington Double Star Catalog maintained at the U.S. Naval Ob- servatory. We’d like to thank Rachel Freed, M.S. President of INSTAR for guiding us through this research process. References Eker, Z., Soydugan, F., Soydugan, E., Bilir, S., Yaz Gökçe, E., Steer, I., Tüysüz, M., Şenyüz, T., and Demircan, O., “Main-Sequence Effective Tempera- tures From A Revised Mass–Luminosity Relation Based On Accurate Properties,” 2015 April, The Astronomical Journal, 149:131 (16pp) Gaia Collaboration et al. (2016): Description of the Gaia mission (spacecraft, instruments, survey and measurement principles, and operations) Gaia Collaboration et al. (2018b): Summary of the contents and survey properties. Las Cumbres Observatory | Many Eyes - One Vision (n.d.). Retrieved from https://lco.global/ Morgan, S., University of Northern Iowa, Spectral type characteristics. Retrieved from https://sites.uni.edu/ morgans/astro/course/Notes/section2/ spectralmasses.html

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 440

A New Double Star from an Asteroid Occultation TYC 6356-00058-1

Dave Gault1, Tony Barry1, Peter Nosworthy1, William Hanna2, Steve Messner2, David Oesper2, Robert Dunford2, Randy Trank2, Steve Preston2, and Dave Herald2

1) Trans-Tasman Occultation Alliance 2) International Occultation Timing Association (IOTA)

Abstract: On May 2nd 2020, (259) Aletheia was predicted to occult TYC 6356-00058-1, and four observers at sites in Eastern Australia attempted to observe the event. Observers at three sites each observed an occultation while one observer observed the star during the occultation period but did not observe an occultation. Close examination of the disappearance at each of the three sites revealed that the star has a faint companion of magnitude 13.5, at a separation of 2.6 mas.

Prediction The occultation was predicted by Preston using Oc- observing sites. It should be noted that sites N (Peter cult software. Figure 1 shows the path of the asteroid’s Nosworthy) and B (Tony Barry) perfectly overlapped shadow across the Earth and the location of the with respect to the path of the asteroid’s shadow.

Figure 1: Occultation period and location of observing sites.

Received July 22, 2020

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 441

A New Double Star from an Asteroid Occultation TYC 6356-00058-1

Nosworthy’s reappearance showed a 2-frame hesi- Observations tation marginally above the noise level, before com- Nosworthy and Gault used GPS time inserted PAL plete reappearance (analogue) video. Gault’s reappearance showed an increasing reap- Barry used a GPS disciplined digital video system. pearance over 3-frame duration Hanna used a QHY174 camera system with built in Barry’s reappearance was instantaneous. GPS receiver. Positive detections were observed at three sites as Examination of the Disappearance Steps shown in Figure 2. Figure 3 shows the disappearance light curves in detail with the measures of the step shown in red which depicts the detected light from the new stellar companion. Nosworthy detected the companion for 0.20 seconds after the disappearance of the prime star. Gault detected the companion for 0.24 seconds after the disappearance of the prime star. Barry detected the companion for 0.26 seconds after the disappearance of the prime star. Analysis of the Occultation Fit The basis for analysis of asteroid occultations in- volving double stars is presented by Herald, et al 20103, in addition the present observation has the com- plication of two chords superimposed and an additional close chord, means that a circular ellipse is typically used to determine the companion’s separation and position angle, however this produced a very poor fit of circular ellipse. It was noted that a previous 4-chord occultation observation by (259) Aletheia was observed in 2018, production a Best-Fit-Ellipse of 210.0 x167.0km (Figure 4).

Figure 2: The complete light curves of the three observers

Nosworthy and Barry both observed occultations of 11.96 seconds duration, while Gault observed an occul- Figure 4: 2018 January 28, 4 chord occultation by (259) Aletheia tation of 12.28 seconds duration. All three observations revealed a step event at Dis- Figure 5 shows applying the 2018 observed Best- appearance shown at the bottom of the curve near ex- Fit-Ellipse dimensions to the present observation pro- tinction but well clear of the noise level. Further discus- duced an excellent fit of all chords from both stellar sion of the disappearance steps will occur later in this components, giving a double star separation of 2.6 +/- paper. 0.3 mas, at PA 174° +/- 3.5°

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 442

A New Double Star from an Asteroid Occultation TYC 6356-00058-1

Figure 3: Detail light curves of (a) Peter Nosworthy, (b) Dave Gault, and (c) Tony Barry.

Figure 5: Best fit ellipse of chords from primary and new stellar companion.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 443

A New Double Star from an Asteroid Occultation TYC 6356-00058-1

Estimation of Component Magnitudes Fig. 6 Shows the estimation of the apparent visual magnitudes of both components, based on Gault’s light curve, being; A: Mag. 10.25 +/- 0.04 Mag. B: Mag. 13.53 +/- 0.30 Mag.

Figure 6: Estimation of component magnitudes. Double Star Characteristics Star TYC 6356-00058-1, SAO 189796, BD 22 5555 Coordinates (J2000) 20h 52m 44.1s, -22° 04m 51.9s Spectral Type G5 Mag. A 10.25 +/- 0.04 (V) Mag. B 13.53 +/- 0.30 (V) Separation 2.6 +/- 0.3 mas Position Angle 174° +/- 3.5° Epoch 2020.3382 References Herald D., 2010, Journal of Double Star Observations, 6, 88-96.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 444

The Fairborn Institute Robotic Observatory: First Observations

Calla Marchetti1, Ryan Caputo2, and Russell Genet3

1) University of California, Los Angeles 2) University of Colorado, Boulder 3) California Polytechnic State University, San Luis Obispo

Abstract: Speckle data were remotely collected across five nights as the first observations of the Fairborn Institute Robotic Observatory. Measurements of 24 close double star systems with separations ranging from 1.49 to 4.52 arcseconds are reported

Introduction Using this information, some targets were specifically The Fairborn Institute Robotic Observatory (FIRO) chosen because the secondary star, while dimmer, had a is focusing on measuring close double stars using hotter surface temperature than the primary. This trait is speckle interferometry — a technique developed by fairly uncommon, and it was speculated that the sec- Labeyrie (1970). Speckle interferometry enables double ondaries of these systems—listed below in Table 1-- stars below the seeing limit to be resolved, shifting the were white dwarfs. But when the stars were plotted on limiting factor from atmospheric seeing to the aperture HR diagrams, using the luminosity and surface temper- of the telescope, thus allowing the diffraction limit of ature given by Gaia, it was clear that all of the second- the telescope to be reached. Short exposures, typically aries were much too luminous to be white dwarfs. For less than 40 milliseconds, are taken of the target, which some of the systems, both stars were on the main se- freezes the seeing and prevents the stars from blurring quence and the difference between the stars’ luminosity together. A Fourier transform is then applied to the im- and surface temperature was minor. For the rest of the ages to generate an autocorrelogram. Bispectrum analy- systems, shown in Figure 2, the primary star did not fall sis is a further refinement beyond the autocorrelogram: on the main sequence, but with the red giants: stars in it removes the extraneous sideband, and transforms the the last stages of stellar evolution characterized by their autocorrelogram back into a real image. inflated sizes and cooler temperatures. With a few of these, such as STF 1783, the secondary may also be in Target Selection the process of becoming a red giant. All other observed The FIRO observatory’s horizon is restricted by stars are on the main sequence. trees, so target selection was limited to stars within a region corresponding roughly to declinations between Instrumentation +13 and 60 degrees and a right ascension angle of six FIRO has an 11-inch aperture C-11 telescope hours. mounted on a custom mount designed by Genet and Targets were further limited to having separations built by students at California Polytechnic State Uni- of 1.4″ and wider, delta magnitudes of 3 or less, and versity. The C-11 is f/10, having a focal length of secondary magnitudes of 12 or brighter. These con- 2800mm. The ZWO ASI 1600mm camera has 3.8 μm straints avoided pushing the limitations of the FIRO pixels, giving a scale of approximately 0.26 ″/pixel. telescope, based on the telescope’s aperture pixel scale, This sampling allows 1.0″ separation double stars to be and the seeing conditions. Finally, only stars that were resolved. The theoretical diffraction limit of an 11-inch already in the WDS were selected. All targets were telescope is 0.4″, however the sampling does not allow chosen using the Gaia Double Star (GDS) search tool, this to be reached. The sampling is relatively coarse for which draws from the Gaia DR2 database (Rowe, speckle, and at least four to five pixels between the cen- 2018). troids are needed to make confident measurements. The Gaia DR2 contains information about the sur- However, the course sampling has the advantage of face temperature and luminosity (Andrae et al, 2018). packing more light into each pixel,

Received October 17, 2020

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 445

The Fairborn Institute Robotic Observatory: First Observations

Primary Secondary Primary Surface Secondary Surface System Luminosity Luminosity Temperature (K) Temperature (K) (L⊙) (L⊙)

STF 1783 52.6 6.8 4939 5153

STF 1808 1.3 0.7 5828 5831

STF 2035 8.3 3.0 6449 6876

STT 314 63.5 3.8 4562 4854

STF 2112 15.1 6.6 5350 5861

STF 2210 44.0 7.8 4962 7381

HU 1286 1.5 1.4 5440 5726

Table 1: Gaia DR2 luminosity and surface temperature for targets where the secondary had a hotter surface temperature.

Figure 2: Target systems where one or more of the stars are potential red giants, plotted on an HR diagram.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 446

The Fairborn Institute Robotic Observatory: First Observations

meaning fainter doubles can be measured. In addi- is reported. For every night that a star was measured, tion, there was no filter used to restrict the bandpass. four captures of 1000 frames were taken. By definition, this transmits the most light, helping to The pixel scale was calibrated each night. It was measure faint double stars. approximately 0.245″ per pixel for the first two nights and .278″ per pixel for the remaining nights. This difference is most likely due to Genet adjusting the telescope during the day and moving the primary mirror. This has the effect of changing the back focus and therefore the magnification of the telescope. The data were processed with bispectrum analysis using STB 1.14. Results of the bispectrum analysis are summarized in Table 2, with systems listed in order of right ascension.

Discussion As stated above, some stars were measured across multiple nights. For these stars, we examined the standard errors associated with each night compared to the overall standard error to see how measuring across mul- Figure 3: The FIRO telescope in its enclosure. tiple nights affected the accuracy of the measurements. As shown in Table 3, some of the overall standard er- Procedure rors are larger than the individual errors, which initial- For these observations, Genet opened the observa- ly, we thought was problematic. But the larger overall tory’s roof and powered up the equipment. Other than error is probably caused by nightly variation in temper- that, the telescope was controlled remotely by Mar- ature or other atmospheric qualities that affect the re- chetti and Caputo. Observations took place between sulting measurement. The larger error is likely a more July 2-10, 2020. accurate representation of the error involved in measur- All mount control software was interfaced through ing double stars; taking back-to-back measurements the SiTechZWOCam software (Gray, 2020) and Sidere- within a single night produces a low error, but perhaps al Technology (SiTech). Cartes du Ciel (Chevalley, artificially so. 2020) was used to slew the telescope to the target and reference stars, and Nighttime Imaging ‘n’ Astronomy Conclusion (NINA) (Berg, 2020) was used to correct for pointing We presented measurements on 24 double stars error by doing a platesolve and resync routine. Sharp- with most separations below the seeing limit. Speckle Cap (Glover, 2020) was used to capture the speckle interferometry, specifically bispectrum processing, al- images. The process was similar to Caputo et al. lowed these measurements to be made. Considering the (2020). FWHM of the stars was around 3″ during image acqui- The speckle interferometry data reduction used Da- sition, our measurement of a 1.5″ split shows the power vid Rowe’s SpeckleToolBox (STB) (Rowe, 2019). This of speckle interferometry. Bispectrum processing is a software generates FITS cubes, performs the Fourier further application of speckle interferometry and is an analysis, and then presents the resulting autocorrelo- alternative way of measuring position angle and separa- gram. The data can further be processed using bispec- tion. It is useful as it generates a real image instead of trum, which transforms the power spectrum density an autocorrelogram, and resolves the position angle (PSD) that the autocorrelogram generates back into an ambiguity. image and removes the sidebands and resolves the 180o These 24 double stars represent the first science position angle ambiguity. We used bispectrum to pro- that was done with the Fairborn Institute Robotic Tele- cess these observations, as it results in a more accurate scope. Our measurements of 24 double stars, in addi- measurement than just performing autocorrelation. tion to being the first science, contributed to the development and usability of FIRO by fine-tuning the data Measurements acquisition process. We report measurements of 24 double stars. Some stars were measured across multiple nights to check for consistency, and an overall average and standard error

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 447

The Fairborn Institute Robotic Observatory: First Observations

Position Standard error Standard Error of Number of Separation System Date Angle of Position Separation Nights (″) (o) Angle (o) (″)

STF 1730 2020.52 1 338.37 0.12 1.996 0.010

STF 1783 2020.51 1 48.10 0.17 2.559 0.007

STF 1808AB 2020.52 1 82.58 0.25 2.723 0.009

STF 1826AB 2020.52 1 310.32 0.05 4.526 0.005

STF 1858AB 2020.51 2 38.15 0.04 3.083 0.029

STF 1884 2020.50 1 55.34 0.17 2.252 0.011

A 1629 2020.52 1 281.12 0.25 2.135 0.011

ES 774 2020.52 1 230.44 0.17 3.459 0.002

ES 1252 2020.52 1 15.51 0.28 1.645 0.013

HU 148 2020.51 1 207.36 0.13 1.699 0.007

STT 296AB 2020.51 1 272.73 0.11 2.302 0.005

STF 2000 2020.51 1 226.41 0.11 2.619 0.012

BU 811AB 2020.52 1 262.16 0.11 1.726 0.008

STF 2035 2020.52 1 35.51 0.07 2.698 0.003

STT 314 2020.51 5 233.94 0.11 3.841 0.023

J 1124 2020.51 1 276.11 0.04 3.859 0.002

STF 2112 2020.51 2 262.54 0.17 2.255 0.029

STT 322 2020.51 2 200.85 1.01 1.497 0.032

A 1875 2020.51 5 189.64 0.13 2.592 0.012

STF 2153 2020.50 1 243.93 0.36 1.568 0.041

HEI 247 2020.52 1 103.16 0.20 2.109 0.008

TDS 867 2020.52 1 247.91 0.33 1.956 0.010

STF 2210 2020.51 4 122.26 0.07 3.321 0.015

HU 1286 2020.51 4 268.75 0.14 3.234 0.014

Table 2: Position angle and separation measurements of the target systems, as well as date observed, number of nights observed, and standard errors.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 448

The Fairborn Institute Robotic Observatory: First Observations

July 2-3 July 3-4 July 4-5 July 7-8 July 9-10 Overall System Standard Standard Standard Standard Standard Standard Error Error Error Error Error Error

Position 0.04 ------0.03 0.04 Angle STF 1858AB Separation 0.008 ------0.005 0.029

Position 0.21 0.30 0.04 0.03 0.07 Angle 0.11 STT 314 Separation 0.025 0.015 0.007 0.004 0.004 0.023

Position 0.32 -- -- 0.08 -- 0.17 Angle STF 2112 Separation 0.053 -- -- 0.008 -- 0.029

Position 0.42 -- -- 0.41 -- 1.01 Angle STT 322 Separation 0.033 -- -- 0.018 -- 0.032

Position 0.39 0.19 0.11 0.16 0.08 0.13 Angle A 1875 Separation 0.035 0.022 0.011 0.004 0.008 0.012

Position 0.03 0.22 0.09 -- 0.09 0.07 Angle STF 2210 Separation 0.180 0.025 0.004 -- 0.005 0.015

Position 0.40 0.25 0.07 -- 0.09 0.13 Angle HU 1286 Separation 0.004 0.009 0.003 -- 0.003 0.014

Table 3: Standard error of individual nights along with the overall standard error for stars measured across multiple nights.

Acknowledgements Caputo et al. (2020). First Remote Student Speckle This research was made possible by the Washing- Interferometry Double Star Observations on the ton Double Star catalog maintained by the U.S. Naval InStAR Student Robotic Telescope Network. Observatory, along with the Stelledoppie catalog main- Journal of Double Star Observations. 16(5). tained by Gianluca Sordiglioni. This work has also Glover, R., SharpCap, https://www.sharpcap.co.uk/. made use of data from the European Space Agency Accessed June 28, 2020. (ESA) mission Gaia. Funding for the DPAC has been Harshaw, R., Rowe, D., and Genet, R. (2017). The provided by national institutions, in particular the insti- SpeckleToolBox: A Powerful Data Reduction Tool tutions participating in the Gaia Multilateral Agree- for CCD Astrometry. Journal of Double Star ment. Thanks to Dave Rowe, for the use of the Speckle- Observations, 13(1), 16. ToolBox and the Gaia Double Star (GDS) database and Labeyrie, A. (1970). Attainment of Diffraction Limited selection tool. Resolution in Large Telescopes by Fourier References Analyzing Speckle Patterns in Star Images. Andrae, R., et al. (2018). Gaia Data Release 2. First Astronomy and Astrophysics, 6, 85. Stellar Parameters from Apsis. Astronomy and Rowe, D., 2018. Gaia (DR2) Search Program. Private Astrophysics. Vol. 616. A8. Communication Berg, S. Nighttime Imaging ‘n’ Astronomy, https:// Rowe, D., 2019. SpeckleToolBox (STB) Program. nighttime-imaging.eu/. Accessed June 28, 2020. Private Communication Chevalley, P., Cartes Du Ciel. https://www.ap-i.net/ Gray, D., Genet, R., and Harshaw, R. (2020). skychart/en/start. Accessed June 28, 2020. SiTechZWOCam Operations Manual. In press.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 449

New Stellar Companion To Exoplanet Host Binary Star WASP 3 AB

F. M. Rica

Grupo Astronómico Orión, C/José Ruíz Azorín, 14, 4º D, E06800 Mérida (Spain) [email protected]

Abstract: This work reports the discovery of a new wide (ρ = 18.33”) stellar companion, with common proper motion and common parallax to WASP-3 AB exoplanet host binary star (WSP 3 AB = WDS 18345+3540). This system is located at about 235-240 pc. The new companion is a weak star (mag. V = 14.1) of spectral type K5V separated about 4330 AU of AB. VOSA, a Virtual Observatory tool for spectral energy distribution fitting was used to characterize the astrophysics of the new component. From Gaia DR2 astrometric data, the relative and projected velocity of the new companion with respect to WSP 3 AB was calculated. This relative velocity is greater than the escape velocity but the hypothesis of background and no- moving object can be rejected with extremely high significance. Gaia DR2 don’t list B component and likely the astrometry for the main stellar component be affected by binarity.

1 Introduction and estimate for the first time the spectral type of the B At the moment of writing this work more than 4296 component and the minimum orbital period. In Section exoplanets have been discovered in 3175 planetary sys- §3, I detail the astrophysical characterization of WASP tems. Some of them are composed of more than one 3 C using VOSA and other procedures. New astromet- star, that is, are binary or multiple systems. One of the ric measures are detailed in Section §4. In Section §5, I exoplanet host binary system is WASP 3 (see Figure 1) describe the dynamical study of C with respect to AB. listed in WDS catalog as WSP 3 AB (= WDS 18345+3540 AB) and composed of stars with K magnitude of 9.36 and 15.93 separated by only 1.2”. In this work, I report the discovery of a third wide (ρ = 18.33”) stellar companion of K5V spectral type with common proper motion and common parallax. To characterize this new component, I used Virtual Obser- vatory (VO) techniques. The VO is “an international initiative designed to provide the astronomical community with the data access and the research tools necessary to enable the exploration of the digital, multi- wavelength universe that is resident in the astronomical data archives.” I took advantage of the accurate proper motions of the European Space Agency’s Gaia mission Second Data Release (DR2) archive (Lindegren et al. 2018; Arenou et al. 2018). This new companion was briefly reported in the Double Star Information Circular num. 198 (June 2019) of Commission G1 of International Astronomical Un- ion (IAU). But here I publish for the first time many astrophysical and dynamical data. Figure 1: WASP-3 and the new weak companion (PanSTARRS In Section §2, I present the close pair WSP 3 AB DR1 image). Image obtained using Aladin Sky Atlas.

Received August 22, 2020

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 450

New Stellar Companion To Exoplanet Host Binary Star WASP 3 AB

(2010)), Gaia DR2 (bands G, GBp and GRp) and Pan- 2 The Close Binary WASP 3 AB Starrs PS1 (bands g, r, z, i, z, y; Flewelling et al. (2016), Ngo et al. (2015) using the 9.8 m Keck II telescope, Chambers et al. (2016)). In total 15 photometric points discovered in 2012.42 a low mass (0.108 ± 0.006 Mʘ) which 14 were used in the fits (W4 band is an upper and weak companion to WASP 3 A at 1.19 arcsec bound value and was rejected). (285.4 ± 0.7 AU) of separation. They performed high In VOSA, we used the BT-Settl theoretical models quality measurements (errors of about 1.5 mas) in (Allard 2012) in the fitting procedure (using minimiza- 2012.57 and 2013.41. Nog et al. determined that WASP 2 tion χ ) between 4200 and 4700 K in Teff and between 3 AB is a common proper motion system and estimated 4.0 and 5.0 in log g for solar metallicity. The uncertain- the apparent magnitude in JHKs bands of B component. ty in the best fit was the size of the grid, which was of Gaia satellite does not list the B component in Data 50 K in Teff and 0.25 in log g. Anyway, the VOSA log g Release 2 (DR2). This is not unusual. Arenou et al. values have to be taken with caution and refined using (2018) determined that only a small fraction of sub- other indicators. The results of the fits was T = 4400 arcsecond pairs were resolved. And the angular separa- eff ± 50 K, log g = 4.5 ± 0.25 and R = 0.82 ± 0.02 Rʘ. Teff tion of AB, 1.19”, is near this limit. Maybe in the final and stellar radius is in very good agreement with Gaia release of the Gaia catalog, B may be listed. DR2 (Teff = 4441+380 and R = Rʘ). From Gaia DR2 parallax of the A component, the - 173 JHKs absolute magnitudes for B was calculated match- Figure 2 plots the spectral energy distribution ing with a M5.5V star. This is the first time the spectral (SED) for WASP 3 C obtained with VOSA. Figure 3 type of B component is published in the literature. As- plots the result of the BT-Settl fit over plotting a theo- suming a face-on (i = 0 deg) and circular (e = 0) orbit, retical spectrum. an orbital period of 4,169 years was calculated which correspond to a change in positional angle of 0.09 deg yr-1 approximately. For an edge-on orbit, the separation will change up to 1.8 mas yr-1. No significant relative motion detected. If the relative astrometry is corrected by atmospheric refraction, only angular distance will be affected and it must be increased by 1 mas. 3 Characterizing the New Stellar Companion WASP 3 C The exoplanet host binary star WASP-3 AB (F7V + M5.5V, mag. V = 10.64) has a new weak companion

(V = 14.1) located at 18.33 arcsec (4330 AU at the distance of the system). Figure 1 shows the position of this Figure 2: SED for WASP 3 C obtained with VOSA. new companion. The system is located at about 237 pc. The reddening in the line of sight was estimated using the maps of Schlafly and Finkbeiner (2011). The resulting values were scaled to the initial distance using the formula published by Anthony-Twarog, and Twarog (1994). In this work, a reddening of Av = 0.08 was calculated The VO Spectral energy distribution Analyzer (VOSA2; Bayo et al. 2008) tool was used to derive astrophysical properties of WASP 3 C from fits of the observed spectral energy distributions to theoretical models. Apart from the CMC14 (band r) and 2MASS (bands J, H and Ks; Skrutskie et al. (2006)) photometric data, I also used those of the Wide-field Infrared Sur- Figure 3: BT-Settl fit of WASP 3 C over plotting a theoretical vey Explorer (bands W1–4; WISE, Wright et al. spectrum.

1Combined V magnitude obtained from Tycho-2 catalog and converted to standard system. 2http://svo2.cab.inta-csic.es/theory/vosa/

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 451

New Stellar Companion To Exoplanet Host Binary Star WASP 3 AB

VOSA allows performing a template fit to obtain prove the quality of this measure I used the REDUC the spectral type. In this work, I use the empirical tem- software, programmed by the French astronomer Flor- plate library of stellar spectra of Kesseli et al. (2017). In ent Losse. This software is widely used by the non- this case the fit used only fits five observational points professional double star community and in the normal toward the bluest part of the SED. The spectral type mode REDUC obtains very good centroid calculation resulting is K5V. even for bad shaped and overexposed stars, like those Table 1 lists all the astrophysical data for AB and C relatively bright stars on POSS photographic plates. My components obtained in the astronomical literature or in measurement (246.0 deg and 17.90”) is significantly this work. For the Gaia parallax and distance I took into closer to modern measures than the measure reported account the systematic offset of -0.03 mas (Lindegren by Damm (2019) in a private communication to USNO et al. 2018). This value is the mean offset but the exact using the old POSS-O plate. The last astrometric point offset for a combination of magnitude, colour, and posi- was obtained by the author using the coordinates of tion may be different by several tens of µas. For this Gaia DR2 catalog. reason I add quadratically the mean offset to the parallax listed by Gaia. 5 Dynamical Study of the New Component C From the astrometric data of Gaia DR2 a relative I also estimated the spectral type using other proce- -1 -1 dures. The Pan-STARRS colors (transformed to B, V, I motion of 2.07 ± 0.07 mas yr (2.32 ± 0.08 km s ) of standard photometry using the transformation of Kos- the new companion with respect to WASP-3 AB was tov & Bonev (2017)) and 2MASS photometry (see Ta- determined. This is 9% of the total proper motion sug- ble 1) were also used to obtain a spectral type K5V. gesting a common proper motion nature. From spectral types and stellar masses, an escape velocity (assuming The procedure used consisted in compare the V magni- -1 tude and the B - V, V - I, J - H and H - K colors (taken face-on and circular orbit) of 0.80 km s was calculat- into account their errors) for WASP 3 C component ed. As the relative motion is clearly greater than escape with those listed in the Mamajeck table (Version velocity I cannot conclude a gravitational relation be- 2019.3.22)3. The entry of the table with a minimum χ 2 tween AB and C. But we have to take into account that was chosen. An excel tool designed by the author was Gaia did not detect component B and therefore the sat- used. ellite measured the astrometric solution of the AB pho- Other nearly independent method use V magnitude to-center. This photo-center will orbit around the AB for WASP 3 C and Gaia DR2 parallax to obtain a V center of mass of several thousand years. The quality absolute magnitude of +7.12 ± 0.05 which matches astrometric parameters of Gaia with a K4V spectral type. In addition to this, using the (ASTROMETRIC_EXCESS_NOISE and RUWE) effective temperature listed in Gaia DR2, a spectral show no problem in the astrometric fit for AB while for type of K5V is also estimated. In both methods, the new component, the RUWE = 1.74 (higher than the Mamajeck table is used. In this work a spectral type limit of 1.4 propose by Lindegren) indicating a possible K5V was adopted for WASP 3 C. problem in the Gaia astrometric reduction. Although Again using the Mamajeck table, a stellar mass of the ASTROMETRIC_EXCESS_NOISE = 0.17 mas 0.76 solar mass was estimated. This is the first time in indicates no problem in Gaia astrometry solution. From historical astrometric points the separation (ρ) the literature that astrophysical parameters for WSP 3 C -1 are published. seems to increase in 2.4 ± 0.6 mas yr while no significant change was observed for position angle. The rela- 4 New Astrometric Measurements tive motion is of 2.8 mas yr-1 (in agreement with that Since the new C component was reported for the obtained using Gaia data) with a significance less than first time (Rica 2019) USNO catalogued this new com- 2σ. panion and add a few astrometric points. Table 2 list From the stellar masses and the differential magni- WDS astrometric points and detail the observational tude of A and B stellar components, I calculated that epoch, the position angle and distance, the magnitudes, the center of mass is located at 0.10” from A while the the observed number of nights (column “n”), the refer- photo-center is at 0.001-0.003” (depending if J, H, Ks ence of the publication, the aperture of the telescope (in band is used) from A. Therefore the distance between meter) and finally the source. The first astrometric point center of mass and photo-center is about 0.097”. In listed in WDS was obtained using the POSSI-O photo- summary, the astrometric solution of AB listed in Gaia graphic plate taken in 1950. With the objective to im- DR2 could be significantly altered.

3 http://www.pas.rochester.edu/~emamajek/EEM_dwarf_UBVIJHK_colors_Teff.dat. Most of the content of this table was incorporated into Table 5 of Pecaut & Mamajek (2013).

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 452

New Stellar Companion To Exoplanet Host Binary Star WASP 3 AB

AB C Reference

WASP 3 Designations BD+35 3293 UCAC4 629-058739

18h 34m 30.246s 18h 34m 31.618s Gaia DR2  2000 +35º 39’ 41.15” +35º 39’ 33.64” Gaia DR2  2000

V +10.64 ± 0.06 a) +14.07 ± 0.04 d)

B - V +0.41 ± 0.07 a) +1.12 ± 0.06 d)

V - I … +1.40 ± 0.06 d)

G +10.453 +13.604 Gaia DR2

J +9.60 ± 0.02 +11.76 ± 0.02 2MASS

H +6.41 ± 0.01 +11.20 ± 0.02 2MASS

K +6.36 ± 0.02 +11.06 ± 0.02 2MASS

() [mas yr -1] -6.07 ± 0.04 -7.39 ± 0.04 Gaia DR2

() [mas yr -1] -21.75 ± 0.05 -23.34 ± 0.06 Gaia DR2

Spectral type F7V + M5.5Vc) K5Vc)

Parallax [mas] 4.30 ± 0.04 4.17 ± 0.04 Gaia DR2

Distance [pc] 232.7 ± 2.3 239.8 ± 2.4

… Radial Velocity [km s-1] -4.40 ± 0.36 Gaia DR2

Torres et al. [Fe/H] -0.06 ± 0.08 … (2012)

1.20 ± 0.01e) 0.76c) Mass [Mʘ] 1.26 ± 0.10b)

e) c) Radius [Rʘ] 1.30 ± 0.07 0.82 ± 0.02

Teff [K] 6486 ± 50f) e) 4400 ± 50c)

Log g 4.27 ± 0.02b) 4.5 ± 0.25c)

Notes. a) From Tycho-2 transformed to standard system; b) Santos et al. (2013); c) This work; d) obtained transforming Pan Starrs photometry to standard system using the transformation of Kostov & Bonev (2017); e) Knutson et al. (2014); e) Gaia DR2; f) Stevens et al. (2017).

Table 1: Astrophysical data for the components of WASP 3 system WSP 3 ( = WDS 18345+3540).

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 453

New Stellar Companion To Exoplanet Host Binary Star WASP 3 AB

Epoch θ [ º ] σ [ " ] mg. A mg. B n Author Aperture Source 1950.452 247.9 17.31 1 Dam2019m 1.2 POSSI-O 1950.450 246.0 17.90 1 RICA 1.2 POSSI-O 1989.512 245.9 18.20 1 RICA 1.2 POSSII-J 1991.532 245.4 17.92 1 RICA 1.2 POSSII-J 1998.3 245.9 18.36 9.603 11.755 1 TMA2003 1.3 2MASS 2001.68 245.8 18.31 1 Dam2019m 0.2 2002.53 245.8 18.30 10.51 13.76 4 UC_2013a 0.2 UCAC4 2010.5 246.0 18.27 1 Dam2019m 0.4 WISE 2015.5 245.8 18.33 1 Fmr2019c 1.0 Gaia R2 Table 2: Astrometric data for WASP AB-C . If the new component be a background and no- Planetary Science Division of the NASA Science Mis- moving object, the relative motion with respect to AB sion Directorate, the National Science Foundation would be of 22.57 ± 0.07 mas yr-1 therefore it can be Grant No. AST-1238877, the University of Maryland, rejected the background object hypothesis with a signif- Eotvos Lorand University (ELTE), the Los Alamos Na- icance of 215 σ. tional Laboratory, and the Gordon and Betty Moore Foundation. Acknowledgements This publication makes use of VOSA, developed References under the Spanish Virtual Observatory project support- Allard et al,. 2012, RSPTA 370. 2765A ed by the Spanish MINECO through grant AyA2017- Anthony-Twarog B. J., Twarog B. A. 1994, AJ, 107, 84089. 1577. This publication makes use of data products from Arenou F. et al., 2018, A&A, 616, A17 the Wide-field Infrared Survey Explorer, which is a Bayo A., Rodrigo C., Barrado y Navascués D., Solano joint project of the University of California, Los Ange- E., Gutiérrez R., Morales-Calderón M., Allard, F., les, and the Jet Propulsion Laboratory/California Insti- 2008, A&A 492,277B. tute of Technology, funded by the National Aeronautics Chambers et al., 2016, arXiv: 1612.05560 and Space Administration. Damm, F., private communication, email #219, 2019 This publication makes use of data products from Flewelling et al., 2016, arXiv: 1612.05243 the Two Micron All Sky Survey, which is a joint pro- Kesseli et al. 2017, ApJS, 230, 16K ject of the University of Massachusetts and the Infrared Knutson H.A., Fulton B.J., Montet B.T. et al., 2014, Processing and Analysis Center/California Institute of ApJ, 785, 126K Technology, funded by the National Aeronautics and Kostov, A. & Bonev, T,. 2017, Transformation of Pan- Space Administration and the National Science Foun- STARRS1 gri to Stetson BVRI magnitudes. dation. Photometry of small bodies observations. Bulgarian The Pan-STARRS1 Surveys (PS1) and the PS1 Astronomical Journal, 28. public science archive have been made possible through Lindegren L. et al., 2018, A&A, 616, A2 contributions by the Institute for Astronomy, the Uni- Ngo et al., 2015, ApJ, 800, 138N versity of Hawaii, the Pan-STARRS Project Office, the Pecaut & Mamajek, 2013, ApJS, 208, 9. Max-Planck Society and its participating institutes, the Rica, F. M., IAU Double Star Information Max Planck Institute for Astronomy, Heidelberg and Circular num. 198 (June 2019) of Commission G1. the Max Planck Institute for Extraterrestrial Physics, Santos N. C., Sousa S. G., Mortier A. et al., 2013, Garching, The Johns Hopkins University, Durham Uni- A&A, 556A, 150S versity, the University of Edinburgh, the Queen's Uni- Schlafly & Finkbeiner, 2011, ApJ, 737, 103 versity Belfast, the Harvard-Smithsonian Center for Skrutskie et al., 2006, AJ, 131, 1163S Astrophysics, the Las Cumbres Observatory Global Stevens D.J., Stassun K.G., Gaudi B.S., 2017, AJ, 154, Telescope Network Incorporated, the National Central 259S University of Taiwan, the Space Telescope Science In- Torres G., Fischer D.A., Sozzetti A. et al., 2012, ApJ, stitute, the National Aeronautics and Space Administra- 757,161T tion under Grant No. NNX08AR22G issued through the Wright et al., 2010, AJ 140, 1868W

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 454

Speckle Observations of Albireo A

Rainer Anton and Johannes M. Ohlert

Abstract: A 120 cm Cassegrain reflector was used in 2019 and 2020 for recordings of the close double star Albireo A (= MCA 55 Aa,Ac) with a fast CMOS camera. In total, more than 100.000 images were analyzed mainly with speckle interferometry, in parts also with lucky imaging. The image scale was calibrated with data from the Gaia catalog DR2 on the basis of wide field images including component Albireo B. Positions of Aa,Ac were found to markedly deviate from earlier orbit calculations by Scardia et al. in 2008, and Roberts and Mason in 2018, but were reasonably close to new ephemeris data by Mason, and by Scardia et al. in 2019, respectively.

Introduction assumed, which was evaluated by Scardia et al. on the Albireo ( Cygni) is one of the most famous dou- basis of mostly speckle observations from 1976 to ble stars in the sky, especially for its brightness and 2008 [3]. The period would be 214 years. As the angu- striking color contrast of components A and B. Up to lar coverage was only small, the quality grade was esti- now, the pair has generally been deemed as true bina- mated to 4 – 5, which means preliminary indetermi- ry, being physically bound. In fact, parallax data from nate. In 2018, Roberts and Mason derived another so- the Hipparcos mission indicated about similar distanc- lution upon including a few measurements taken in es of A and B from earth, with overlapping error mar- 2004. The eccentricity would be rather large, and the gins. Also, proper motion data were not too different, period reduced to 69 years [4]. Things changed in both in strength and direction. But as no definite orbital 2019, when Scardia et al. came up with observations motion has yet been detected, the orbital period would from 2017, which led to an orbit with period 120 years be very long. Beside these astrometric considerations, [5]. Also, in 2019, Mason made a new computation there are also arguments from astrophysics, e. g. mass- with taking into account a measurement of ours from es, spectral classes, and others, which seem to support that year (see below) [6]. This resulted in an orbit with the physical nature. period 213 years, close to that of Scardia 2008, but However, this picture became partly inconsistent with differences in other parameters. Obviously, more with recent data from Gaia DR2, exhibiting significant observations are needed in the future for further refin- differences of both parallax and proper motion values ing the orbit parameters. Our present measurements are of A and B [1]. As a result, the separation would be as just aimed to contribute to this issue. large as 62 ly, which is not reasonable for a binary. Instrumentation Further, the proper motion of A would have changed One of us (Ohlert) used a telescope of Cassegrain direction by roughly 90 degrees within 14 years, or type with aperture 120 cm, focal length 951 cm, locat- even less, which seems to be rather unlikely. One prob- ed in Trebur, Germany, which is operated by the As- lem may be that Albireo A is too bright for Gaia, lead- tronomie Stiftung Trebur (Astronomy Foundation ing to saturation of the detectors, hence less accurate Trebur). On three nights in July 2019 and one night in measurements. Another problem is that A is double June 2020, several series with up to 13000 images each itself, but not resolved, neither by Hipparcos nor by were recorded with a CMOS camera (QHY 5L-IIm) Gaia. Thus, the apparent proper motion of A may be with exposure times of 5 or 10 ms, at rates of up to 200 influenced by orbital motion of Aa-Ac. In order to take frames per sec. Imaging of the pair Aa,Ac is difficult, this into account, precise knowledge of the orbital pa- namely because of the narrow separation in the range rameters is needed. However, observational data are of 0.3 – 0.4 arcsec, combined with the large difference scarce. These aspects have recently been addressed by in brightness of Aa and Ac (2.43/5.08 mag), and of Bastian and Anton in 2018 [2]. color (K3II/B9V). Therefore, a photometric Blue- In recent years, up to 2018, an orbit of Aa-Ac was Johnson filter was used in all series. This also helped

Received July 30, 2020

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 455

Speckle Observations of Albireo A

to partly compensate for the difference of sensitivity of the camera in the blue and yellow. In fact, by reducing the glare of A, the visibility of Ac was significantly improved. As disadvantage, seeing effects are more pronounced at short wavelengths compared with longer ones, but this is hoped to overcome by taking large numbers of frames. Data Processing Speckle images were mostly analyzed with the program “Reduc” by Florent Losse, without pre-selection, because this appeared impracticable with such large numbers of frames [7]. At each night, some series were also recorded with a larger field of view in order to in- Figure 1: Wide field view of Albireo Aa,Ac – B, record- clude component B at position angle 54.04 degrees, and ed on 2019-07-09. Stack of 211 lucky images. Contrast distance 34.59 arcsec according to Gaia DR2 [1]. These was enhanced by unsharp masking. An enlargement of were searched for “lucky images”, which were stacked the close pair is shown in the inset as indicated. The and used for calibration of the image scale as well as other inset shows the superposition of the speckle corre- the orientation of the camera. As a side effect, the lation of Aa,Ac using all 9400 frames of the series. North is up, east is left. speckle clouds of B were analyzed too, in order to check for possible artefacts, but none were found. While the data from Gaia refer to the epoch 2015.5, no significant change is expected in the course of five years, and the scatter of other literature data is too large as to derive an accurate movement even in the long term. Results Figure 1 shows a wide field image of Albireo Aa,Ac–B, obtained in July 2019 by superposition of 211 lucky images, selected by visual inspection from a recording of about 9400 frames. Exposure time was 5 Figure 2: Superposition of 43 lucky msec. The seeing was only mediocre, which resulted in images selected out of 15000 a rather diffuse background. Nevertheless, the pair frames in two recordings from Aa,Ac is clearly resolved, which can better be seen up- nights 2019-07-09 and 2019-07-24. on enlarging. The inset at bottom left shows the result Original frames were re-sampled of speckle autocorrelation of all 9400 frames. The sepa- before stacking with factor 4x4. ration and orientation of the spots corresponds to that in the lucky image. This and several similar views served Another observation run was performed in the night for calibration of the scale and orientation by referring of 2020-06-22/23. Nine series of consecutive record- to the position angle and separation for A and B as de- ings delivered consistent results. An example is shown rived from Gaia, as mentioned above. This resulted in in figure 3. Similar to the process described above, an original resolution on the camera chip of 0.081 speckle autocorrelation was done with all 4492 frames arcsec/pixel. Searching for lucky images in the speckle of the recording “Alb 1 A+B” (see table 2 below), with- clouds by visual inspection in all the thousands of out pre-selection. The pair was also clearly resolved in frames is tedious, and only few were found with more all other series recorded that night. or less isolated speckle pairs corresponding to the result In the following two tables, all evaluations of re- of the autocorrelation program. An example is shown in cordings made in 2019 and 2020 are listed with the re- figure 2. Generally, all frames were re-sampled before spective parameters. Wide field images with compo- stacking, which results in smoothened intensity pro- nents A and B are analyzed with lucky imaging for cali- files, and in higher precision in determining the peak bration, while the results for the close pair Aa,Ac are centers, both in lucky imaging and in speckle analysis. from speckle autocorrelations.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 456

Speckle Observations of Albireo A

date/ Besselian lucky imaging # of year speckle correlations of Aa,Ac of Aa-B for calibration recording 2019+ frames PA/ rho/ frames PA/ rho/

used degrees arcsec used degrees arcsec 9.7./S2 .523 370 54.04 ! 34.59 ! 9400 51.4 0.312 9.7./S1 .523 10000 52.5 0.317 24.7./S5 .564 10100 55.8 0.305 24.7./S6 " 1100 53.5 0.320 24.7./S7 " 10000 49.1 0.313 24.7./S8 " 12500 54.1 0.304 25.7./S10 .567 4300 56.8 0.325

53.3 ± 2.6 0.314 ± 0.008 mean values ± s.d. . Table 1: List of recordings from 2019-07-09 to 2019-07-25. The mean Besselian date was 2019.540. Numbers of frames used for analyses are also listed. Series S2, recorded on 2019-07-09, was also evaluated with lucky imaging and served for calibration with Gaia data (with exclamation marks, see text). lucky imaging of Aa-B file name U.T. speckle correlation of Aa,Ac for calibration frames PA/degrees rho/arcsec frames PA/degrees rho/arcsec Alb 1 A+B 00:25 304 54.04 ! 34.59 ! 4492 50.2 0.327 Alb 2 A+B 00:27 4484 54.8 0.317 Alb 3 A+B 00:29 4414 50.6 0.346 Alb 4 A+B 00:56 4489 51.8 0.320 Alb 1 AaAc 00:32 12050 53.9 0.321

Alb 2 AaAc 00:35 12466 51.4 0.325

Alb 3 AaAc 00:45 12500 52.1 0.339

Alb 4 AaAc 00:48 12500 52.4 0.333

Alb 5 AaAc 00:52 12500 48.9 0.322

mean values ± s.d. 51.8 ± 1.8 0.328 ± 0.010

Table 2: List of recordings at 2020-06-23 at times as indicated. Besselian date was 2020.478. Numbers of frames used for analyses are also listed. Series Alb 1 A+B was evaluated with lucky imaging and served for calibration with Gaia data (exclamation marks, see text). Measures for Aa,Ac are obtained from speckle analyses.

In figures 4 and 5, our position data are compared with ephemeris calculations from other authors, which have already been mentioned in the introduction. Our results seem to fit the new ephemeris data from 2019 by Mason, as well as by Scardia et al. reasonably well, when one would accept a certain scatter of all data, including those on which the orbit calculations are based. Apparently, those from 2008 (Scardia), and from 2018 (Roberts & Mason) are obsolete. Still, a better refinement of the orbital elements is expected with more measurements to be made in the future. It may well be Figure 3: Speckle autocorre- that the true orbit will lie in between the two solutions lation of MCA 55 Aa,Ac from all from 2019. 4492 frames of series Alb 1 A+B, recorded on 2020-06-23 (see table 2). North is up, east is left

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 457

Speckle Observations of Albireo A

[2] Bastian, U., and Anton, R., 2018, A&A, 620, L2. [3] Scardia, M. et al., 2008, Astronomische Nachrichten, 329, 54. [4] Roberts, L.C. & Mason, B.D., 2018, MNRAS, 473, 4497. [5] Scardia, M. et al., June 2019, IAU Commission G1, Double Stars Information Circular No. 198. [6] Mason, B.D., United States Naval Observatory, private communication, 2019. [7] Losse, Fl., http://www.astrosurf.com/hfosaf [8] Hartkopf, W.I. et al., Fourth Catalog of Interferometric Measurements of Binary Stars, U.S. Naval Observatory, online access Oct. 2019. [9] Hartkopf, W.I. et al., Sixth Catalog of Orbits of Visual Binary Stars, U.S. Naval Observatory, online access Oct. 2019. Figure 4: Plot of the position angle of MCA 55 Aa,Ac vs. time (Besselian year). Open circles are speckle data from the “speckle catalog” [8], crossed circles are our own measures. Authors Curves in color are ephemeris data from Scardia et al. (red, Rainer Anton is a retired professor of physics from from 2008 [3,9], and magenta, from 2019 [5]), Roberts & Ma- Hamburg University, Germany. He has son (blue, from 2018 [4,9]), and Mason (green, from 2019 [6]), respectively. been measuring double stars in the northern and south- ern hemispheres since 1995.

Johannes M. Ohlert is a retired professor of physics from University of Applied Sciences, Friedberg, Ger- many. He is co-founder of the Astronomy Foundation Trebur. In cooperation with the Nicolaus Copernicus University, Torun, Poland most of his observations are aimed to get precise measurements of transit time varia- tions of transits of exoplanets.

Figure 5: Plot of the separation rho of MCA 55 Aa,Ac vs. time. Meaning of the symbols as in figure 4.

Acknowledgements This work has made use of data from the Gaia satellite mission, which were published by the ESA consortium, as well as from the double star catalogs provided by the United States Naval Observatory. Special thanks are due to Dr. B. Mason for calculating a new orbit by including a measurement of ours. References [1] Gaia Archive, Data Release 2, 2018: http:// esac.esa.int/archive/

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 458

First Remote Student Speckle Interferometry Double Star Observations on the InStAR Student Robotic Telescope Network

Ryan Caputo1, Calla Marchetti2, Jed Teagarden3, J.D. Armstrong3, Caroline Wiese4, Kalée Tock4, Russell Genet5, Richard Harshaw6, Rick Wasson7, and Rachel Freed6

1) University of Colorado, Boulder 2) University of California, Los Angeles 3) University of Hawaii Institute for Astronomy 4) Stanford Online High School, Palo Alto 5) California Polytechnic State University, San Luis Obispo 6) Institute for Student Astronomical Research 7) Orange County Astronomers, Murrieta

Abstract: In this first operation of the Institute for Student Astronomical Research (InStAR) Student Robotic Telescope Network, students in Hawaii, California, and New York remotely operated the telescope at the Purple Sky Observatory in Midland, Texas. Speckle interferometry astrometric measurements were obtained for 31 close double stars with separations ranging from 1.89 to 6.15 arcseconds. The new network is briefly described, along with the remote observation instrumentation, software, procedures, and results.

Introduction where the diffraction limit of a telescope can be Having a dedicated robotic observatory network reached, operating below the seeing limit of the atmos- provides an environment where students can learn to phere (Harshaw, 2017). It is therefore relatively inde- develop and manage a modern research organization. pendent of atmospheric conditions and shifts the limit- The InStAR Student Robotic Telescope Network will ing factor from atmospheric seeing to the telescope’s complement existing observatories to provide research aperture. Speckle interferometry allows stars with data for all levels of astronomers — from amateur to smaller separations to be split when compared to CCD professional. Today’s undergraduate students will be- imaging. Measuring these stars are important because come tomorrow’s graduate students, and eventually doubles with smaller separations are more likely to be professional astronomers. Providing them with a orbiting and to be orbiting with short periods. means of collecting research-grade data is an important Instrumentation step in this process. The Purple Sky Observatory, led by Ryan Caputo, Currently there are five observatories in the In- consists of a six-inch classical Cassegrain telescope StAR Student Robotic Network. The Purple Sky Ob- with a focal length of 1800mm (f/12) and an ASI servatory is the first to host this network’s remote stu- 1600mm CMOS camera. The small 3.8 µm pixels and dent observations, forming the basis for this paper. The long focal length give a sampling of approximately Fairborn Institute Robotic Observatory, FIRO, (Genet) 0.4″ per pixel. This instrumentation is particularly suit- is just now starting to host remote student observa- able for speckle interferometry because of the fine tions. The Shepherd’s Lair (Gray) and Estrada Obser- sampling relative to the 6” aperture of the telescope. vatory (Estrada) are ready to host remote student ob- Furthermore, the camera is a CMOS sensor with a high servations, but have not done so yet, while the Plum readout speed and low read noise (~1 e). The low read Tree Observatory (Freed) is still coming online. Other noise allows faint signals to be recorded in short expo- observatories are welcome to join the network. sures, giving CMOS chips an advantage over CCD Speckle interferometry is an image processing chips even when the quantum efficiencies are similar technique, pioneered in the 1970’s (Labeyrie, 1970) (Genet 2016). Figure 1 shows the telescope.

Received September 24, 2020

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 459

First Remote Student Speckle Interferometry Double Star Observations on the InStAR Student Robotic Telescope Network

A Baader G-filter was employed to restrict the band- urements, we used drift calibration to determine a cam- pass. era angle and pixel scale for this run of, respectively, - 110.74° and 0.42866″/pixel. The software Cartes du Ciel (Chevalley) allowed targets to be selected by HD number and commanded the mount to the selected coordinates. NINA, “Nighttime Imaging ‘n’ Astronomy,” (Berg) allowed, among many other functions, an on-the-spot plate solve routine to determine pointing error. NINA commanded a quick exposure which it then passed off automatically to Platesolve2, written by Dave Rowe. The coordinates of the image were uploaded to the mount to update its pointing. A re-slew command was then sent manually through Cartes du Ciel to center the target. The capture software was then opened. We initially used FireCapture (Edelmann), but found the download times to be slower than the capture rate of the camera. SharpCap (Glover) performed the same basic functions as FireCapture, and due to the more streamlined nature of SharpCap, we switched to this software in the middle of the first observing run, and we used it for the entirety of the second observing run. Figure 2 shows SharpCap being controlled by Marchetti. Figure 1: Caputo and the Purple Sky Observatory.

Target Selection Because speckle interferometry allows diffraction limited measurements to be made, the main limitation was the six-inch aperture of Caputo’s telescope. How- ever, the full resolving power of the telescope could only have been achieved if the sampling was adequate. The airy disk at 540 nm (G-filter) is 1″, meaning there are about five pixels across the airy disk. Six to seven pixels is optimal for maximum resolution (Rowe, per- Figure 2: SharpCap imaging the double star STF 1785, 2.64″ sonal communication); therefore, the telescope was fur- separation (raw frame). ther limited by pixel scale, so higher separations were A problem arose when it came time to measure the targeted. Separations 1.9″ and wider were chosen. Since stars; all five of the images that were taken using the exposure time could be no greater than 40 millisec- SharpCap during the first observing run had position onds to avoid image smearing, the stars had to be rela- angles that varied drastically from what the previous tively bright, so stars above magnitude 10.5 were not measurements suggested. To try and find what was considered. High delta magnitude pairs, set as anything causing this problem, image stacks from the short greater than two magnitudes of difference, were also speckle exposures were made in Registax 6. Registax not considered. Finally, only stars visible from the Pur- can perform lucky imaging which allows better images, ple Sky Observatory in June were chosen, correspond- in this case stars with lower FWHM values, to be rec- ing to right ascensions of 11 to 22 hours and declina- orded (Malsbury, 2013). When these were opened in tions north of -20°. All of these stars were chosen using AstroImageJ, the orientation of the stars was mirrored a spreadsheet compiled by Richard Harshaw (2020). across the x axis, as shown in Figure 3. The problem Software was that SharpCap automatically produced .ser files, Dave Rowe’s SpeckleToolBox was used to gener- which we had been converting to .fits files. We found ate FITS cubes, process the results using Fourier trans- that this conversion was breaking image orientation, forms, and then present them as autocorrelograms for resulting in the discrepancy in position angle. When measurements (Harshaw, 2017). To calibrate the meas- SharpCap’s settings were adjusted to save directly

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 460

First Remote Student Speckle Interferometry Double Star Observations on the InStAR Student Robotic Telescope Network

to .fits files, the image orientation was preserved, and the problem was resolved. Aside from that, both Fire- Capture and SharpCap appear to be identical in basic functionality, and there is no distinction between data gathered by them.

Figure 4: A raw image frame of STF 2909AB (left), and the processed speckle interferometry image from SpeckleToolBox (right). Measurements Measurements are given in Table 1. The systems Figure 3: A five second exposure of D 22 (left) compared to the are listed in order of right ascension. The Number of image stack of the speckle files (right), which had been convert- Images is the number of fits cubes measured, with each ed from .ser to .fits, flipping the image about the horizontal axis. fits cube consisting of 1000 individual frames. Remote Observation Sample Orbit Above all, the goal of the project was to conduct STF 1785 has an estimated period of 155 years. remote speckle observations. Students from Hawaii, Gaia measured this pair in 2015.5, five years before our California, and New York logged into the Purple Sky measurement. Therefore, the pair is expected to have Observatory computer with AnyDesk, which allows traversed about 3% of its orbit between Gaia’s meas- multiple people to be connected to the host computer at urement and our measurement. once; they can see the screen and have mouse control. The Orbit Predictions spreadsheet devised by The significant difference that arises from remote oper- Drummond can quantify this expectation ation is that the team cannot see or hear the telescope. (Drummond 2020). This spreadsheet provides the ex- Not being able to interact with a telescope physical- pected position angle and separation for a known binary ly — only though predetermined software and rather star at a given time. The spreadsheet was updated with limited hardware — poses many potential problems. the most recently published orbital elements for STF The main issue encountered was the telescope’s poor 1785 (Izmailov, 2019), and it predicted a position angle pointing accuracy. Due to it not being a permanent set- of 190.8o and separation of 2.749″ at the time of our up, the telescope is never polar aligned well, nor does it measurement. Our measurement (191.3o, 2.6″) is in have a mount model. The procedure incorporated the good agreement with this prediction, so much so that plate solve routine to correct for the pointing error. the two points are almost on top of each other as shown The speckle routine allowed a much cleaner split of in Figure 5 below. the primary and secondary than traditional CCD imaging techniques, as is evident in Figure 4. On the left, a single eight millisecond exposure is displayed in As- troImageJ. The autocorrelogram generated by Speckle- ToolBox is displayed on the right. The bright spot in the middle is the primary, and the secondary star is one of the bright spots surrounding. One minor inconven- ience of the autocorrelogram is the secondary star is duplicated as two sidebands. Resolving this ambiguity requires at least rough knowledge of the expected position angle. All of the double stars measured here move slowly, so we can be confident the stars will have nearly the same position angle as before, resolving the ambiguity. Figure 5: At left, the historical measurements of STF 1785 are plotted with our measurement and the Gaia DR2 measurement labeled. At right, the WDS orbital solution is compared.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 461

First Remote Student Speckle Interferometry Double Star Observations on the InStAR Student Robotic Telescope Network

Standard error Standard Error Number of Position Separation System Date on Position of Separation Images Angle (o) (″) Angle (o) (″) STF 1510 2020.43 6 328.0 0.08 5.64 0.01

STT 261 2020.43 4 337.4 0.15 2.53 0.00*

STT 266 2020.43 6 355.3 0.06 1.92 0.00*

STF1785 2020.43 4 191.3 0.06 2.64 0.00*

STF1858 AB 2020.43 5 36.8 0.11 3.10 0.01

BU 346 2020.43 5 276.7 0.19 2.80 0.01

STF1890 2020.43 6 45.8 0.61 2.56 0.02

STF1905 2020.43 6 162.6 0.33 3.00 0.02

STF1896 AB 2020.43 5 276.4 0.14 4.14 0.02

STF1950 2020.43 4 91.0 0.18 3.30 0.00*

STF1954 AB 2020.43 4 171.4 0.10 3.97 0.01

STF1985 2020.43 4 354.4 0.07 6.15 0.01

STF2213 2020.43 5 327.0 0.03 4.77 0.00*

STF2404 2020.43 4 180.2 0.04 3.61 0.00*

STF2466 AB 2020.43 5 102.4 0.07 2.50 0.01

STF2522 2020.43 4 338.2 0.24 4.55 0.01

STF2525 AB 2020.43 6 289.2 0.16 2.26 0.01

STF2545 AB 2020.43 5 326.4 0.10 3.66 0.01

STF2613 AB 2020.43 5 354.3 0.10 3.52 0.00*

HJ 1485 2020.43 4 275.6 0.11 4.72 0.01

H N 138 2020.43 4 326.7 0.05 3.27 0.00*

WEI 35 AB 2020.43 5 213.7 0.02 4.19 0.01

HJ 1537 2020.43 5 22.4 0.08 3.57 0.01

STF2742 2020.43 6 213.4 0.09 2.95 0.02

STF2749 AC 2020.43 5 176.3 0.10 3.46 0.03

STF2760 AB 2020.43 4 33.1 0.07 5.62 0.01

STT 437 AB 2020.43 5 19.6 0.04 2.54 0.00*

STF2799 AB 2020.43 6 258.1 0.07 1.89 0.00*

BU 274 2020.43 5 180.3 0.07 3.66 0.01

STF2826 AC 2020.43 5 81.4 0.14 4.0 0.02

STF2909 AB 2020.43 4 156.4 0.50 2.38 0.01

Table 1: Position angle and separation measurements of the target systems *Stars with standard error listed as .00 have standard errors between .001 and .004

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 462

First Remote Student Speckle Interferometry Double Star Observations on the InStAR Student Robotic Telescope Network

Rick Wasson, along with several other filters and a fil- Discussion: Benefits of Speckle Interferometry ter wheel. We would like to thank Dave Rowe for writ- Speckle Interferometry allows close pairs, such as ing SpeckleToolBox. STB is an intuitive and powerful those presented here, to be measured. During the ob- program. serving run, one of us (Armstrong) requested standard We also thank Las Cumbres Observatory for tele- CCD images of several of the stars in multiple filters scope time. using the Las Cumbres Observatory 16” telescopes. This research was made possible by the Washing- Even though the images were coming from observatory ton Double Star catalog maintained by the U.S. Naval locations with excellent seeing, only a few of them Observatory, the Stelledoppie catalog maintained by were able to be measured because of the stars’ small Gianluca Sordiglioni, Astrometry.net, and the As- separations. Figure 6 shows STF 2404, a pair with a troImageJ software. 3.6″ separation, imaged by LCO on the left and Ca- puto’s 6” telescope on the right. Caputo’s 6” telescope References outperforms the 16” telescope because speckle interfer- AnyDesk. https://anydesk.com/en/. Accessed June 28, ometry removes the atmospheric limitation. 2020. Armstrong, J. D., & Tong, W. (2016). Accuracy and Precision of Multicolor Observations of Four Dou ble Stars. Journal of Double Star Observations. 12 (1), 4. Berg, S. Nighttime Imaging ‘n’ Astronomy, https:// nighttime-imaging.eu/. Accessed June 28, 2020. Chevalley, P. Cartes Du Ciel. https://www.ap-i.net/ skychart/en/start. Accessed June 28, 2020. Edelmann, T., FireCapture, http://www.firecapture.de/. Accessed June 28, 2020. Genet, R., Rowe, D., Ashcraft, C., Wen, S., Jones, G., Figure 6: STF 2404: A 3.6-arcsecond separation, imaged with the LCO 16-inch telescope (left) and with Caputo’s 6- inch tele- Schillings, B., Harshaw, R., Ray, J., & Hass, J. scope processed in SpeckleToolbox (right). Note they have dif- (2016). Speckle Interferometry of Close Visual ferent orientations due to different camera angles. Binaries with the ZW Optical ASI 224MC CMOS Camera. Journal of Double Star Observations, 12(3), Speckle interferometry is a single-target type of 10. observation that can only be done target by target. As Drummond , 2020. Personal Communication. such, efforts are being made to develop a network with Glover, R. SharpCap, https://www.sharpcap.co.uk/. the main science goal being speckle interferometry. Accessed June 28, 2020. Furthermore, Gaia Data Release 2 (DR2) has re- Harshaw, Richard, 2020. Personal communication. vealed thousands of previously unknown double stars Harshaw, R., Rowe, D., and Genet, R. (2017). The with similar proper motion and parallax. These targets SpeckleToolBox: A Powerful Data Reduction Tool are well situated for speckle interferometry of moderate for CCD Astrometry. Journal of Double Star -aperture telescopes. Observations, 13(1), 16. Izmailov, I. S. (2019). The Orbits of 451 Wide Visual Conclusion Double Stars. Astronomy Letters, 45(1), 30–38. InStAR is working to build a large-volume speckle https://doi.org/10.1134/S106377371901002X interferometry program for double stars too close to Labeyrie, A. (1970). Attainment of Diffraction Limited measure via traditional long exposure imaging. The Resolution in Large Telescopes by Fourier Purple Sky Observatory has performed the first remote Analysing Speckle Patterns in Star Images. Astrono speckle observations in the network with 31 measure- my and Astrophysics, 6, 85. ments of close double stars within the range of 2-6″. Malsbury, A. S. (2013) “Measurement of Double Stars Students from Hawaii, California, and New York con- using Webcams: 2011 and 2012”, Journal of Double nected to the observatory computer and controlled the Star Observations, 9(3), 176-182. telescope. The telescope operator, Ryan Caputo, was Registax. https://www.astronomie.be/registax/. Ac only present to supervise the telescope and the students. cessed June 28, 2020. Acknowledgements Rowe, D. Platesolve2, https://planewave.com/ The Baader photometric G filter was a gift from downloads/all/. Accessed June 28, 2020.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 463

Speckle Interferometry of Close Doubles on the Mount Wilson 60” Telescope A Live Virtual Star Party

Calla Marchetti1, Ivan Altunin1, Kingsley Panigrahi1, Emersen Panigrahi1, Sophia Panigrahi1, Sophia Risin1, Elliott Chalcraft1, Ryan Caputo1, Kalee Tock1,2, Rachel Freed2, Reed Estrada3, Chris Estrada4, Rick Wasson5, Dave Rowe6, Richard Harshaw2, Blake Estes7, and Tom Meneghini7

1) Stanford Online High School 2) Institute for Student Astronomical Research (Sonoma), CA 3) Northrop Aviation 4) Institute for Student Astronomical Research (Lancaster), CA 5) Orange County Astronomers 6) PlaneWave Instruments 7) Mount Wilson Observatory

Abstract: Speckle data were collected on four different nights in May and June of 2020 using three different telescopes: Mt. Wilson Observatory 60-inch telescope, the Orange County Astronomers 22-inch Kuhn telescope, and the Arizona Sonoran Desert Observatory of Glen- dale 11-inch telescope. Three of those observation events involved students from multiple locations throughout the world participating over Zoom. Later, students met with astronomer mentors online to reduce the data. Eight double star targets with separations ranging from 0.144 to 2.880 arcseconds were measured and reported here, including the triple system A 1609.

Introduction dale (ASDOG) 11-inch telescope. These instruments Close pairs are especially interesting to double-star are shown in Figure 1. astronomers because their orbits tend to be faster. For The MWO 60-inch telescope has a “bent” Casse- some, a full orbit is observable in its entirety over the grain f/16 configuration with a 24-meter focal length course of a human lifetime. However, many known and was used here without a Barlow. Completed in close binaries are difficult to resolve, and some have 1908, the mirror alone weighs 1,900 pounds. It was dim companions. Therefore, measuring them necessi- famously used by Harlow Shapley to create a map of tates the use of large telescopes or advanced tech- the Milky Way Galaxy, which established our Sun’s niques. In spring of 2020, a student observing run was position on its periphery. It was also the first tele- organized by Stanford Online High School (SOHS) scope to image star-like condensations in the “spiral and the Institute for Student Astronomy Research nebulae” (Simmons, 2020). This opened the field for (InStAR) for the purpose of using speckle interferome- later work by Edwin Hubble, who used the 100-inch try to measure several close pairs. Although the in- telescope on Mount Wilson to collect images of Ce- person session was cancelled due to COVID-19 re- pheid variable stars, confirming the “spiral nebulae” as strictions, the project was conducted online, with stu- separate galaxies and challenging Shapley’s initial po- dents participating via Zoom in the data collection and sition that the Milky Way was the extent of the Uni- subsequent reduction. verse (Trimble, 1995). For this project, the historic telescope was mated to Instrumentation a ZWO ASI 1600MM back-illuminated, cooled CMOS The three telescopes used for this project were the camera with a Baader R-filter, attached to the telescope Mt. Wilson Observatory (MWO) 60-inch, the Orange via a flip mirror and several T2 threaded adapter tubes. County Astronomers (OCA) 22-inch Kuhn telescope, The configuration is shown in Figure 2. and the Arizona Sonoran Desert Observatory of Glen-

Received September 24, 2020

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 464

Speckle Interferometry of Close Doubles on the Mount Wilson 60” Telescope A Live Virtual Star Party

Figure 1: Left to right: MWO 60-inch telescope, OCA 22-inch Kuhn telescope, ASDOG 11-inch telescope.

Figure 2: (Left) Red filter on the ZWO camera; (Right) ZWO ASI 1600 Camera attached to a flip mirror with an eyepiece at the top. The 22-inch aperture OCA Kuhn telescope is also Target Selection historic in that it has belonged to the Orange County The targets were selected in part to test the capabil- Astronomers astronomy club since being built by club ities of the telescopes used. Even with excellent see- members, led by William Kuhn, in the 1980s. Initially, ing, starlight is generally smeared out such that the the very heavy telescope, which has an Equatorial Fork “seeing limit” is at least 3″. This means that two stars mount, was operated manually, but, over the years, the would need to be at least 3″ apart in order to resolve telescope has been modified so that it is now computer- them in typical seeing conditions. However, speckle controlled. A 2x Barlow was employed so that the f/8 interferometry enables observers to operate below the Cassegrain optics effectively became f/16. The speckle seeing limit and obtain diffraction-limited information camera used was a ZWO ASI 290MM, so that the re- about the positions of the stars. The closest possible sulting image pixel scale was 0.0736"/pixel. For this separation for double star astrometry using a red filter imaging session, the filters were Clear (as in CCD im- and the speckle interferometry reduction technique is aging "Luminance", used only for finding) and Sloan therefore given by the Rayleigh limit shown in Equa- (SDSS) g' r' i' z' (Generation 2) from Astrodon. tion 1: The third telescope used for this project is located in the suburbs of Phoenix at the Arizona Sonoran De- sert Observatory of Glendale (ASDOG). This Celes- tron 11-inch aperture telescope belongs to Jimmy Ray Equation 1: Rayleigh limit for 650nm wavelength using a tele- and has a Celestron German Equatorial Mount. It is scope of aperture d, where d is given in inches. mated to a ZWO ASI290MM camera. Despite having significant light pollution, ASDOG is well-suited to A comparison between the Rayleigh limit and the speckle interferometry, which is less impacted by light closest-separation pair imaged in this study for each of pollution than most other forms of astronomy. the telescopes used is shown in Table 1.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 465

Speckle Interferometry of Close Doubles on the Mount Wilson 60” Telescope A Live Virtual Star Party

Rayleigh Closest separation Aperture The malfunction was corrected, providing a large Observatory limit imaged in this (inches) amount of back-focus (more than 11 inches) for the (″) study (″) June Star Party. In fact, during the MWO observing MWO 60 0.107 0.144 run on June 14th, triple star A1609AB, C was analyzed OCA 22 0.293 0.367 live using Dave Rowe’s Speckle ToolBox (STB) and shared with the Zoom audience. As the evening pro- ASDOG 11 0.585 2.880 gressed, the team developed a routine of calling out Table 1: Rayleigh limit along with the closest separation double SAO numbers to Telescope Operator Blake Estes at the star imaged by each of the telescopes in this study. control console, who entered these into The Sky software and then pointed the telescope to the correct coor- Observing Sessions dinates. From there, the engineering team performed Rick Wasson used the OCA Kuhn telescope to im- fine adjustments with the fine guidance hand controller age STF 1527, WRH12, and HU572. Jimmy Ray and situated at the telescope to locate and center the star in Richard Harshaw used ASDOG to image STF 1670AB, the camera field of view. with multiple students participating via Zoom. A team including Rick Wasson, Dave Rowe, Reed and Chris Results Estrada, MWO Director Tom Meneghini, Telescope Of the original targets, all were successfully re- Operator Blake Estes, Rachel Freed, and Kalée Tock solved except for 14267+1625 A2069, whose separa- gathered at MWO for an engineering run and then for tion was predicted to be very close to the diffraction an imaging session three weeks later, with students par- limit of the MWO 60-inch telescope. The pixel scale ticipating via Zoom on both occasions. was approximately 0.03″ per pixel, so the 3-4 pixels Fire Capture software was used to control the cam- that separate the two centroids would not have provided era and acquire the speckle images during all of the ob- adequate sampling for confident measurements. A2069 serving sessions. For each of the selected double star was chosen initially because it had a predicted separa- targets, short speckle exposure lengths were estimated tion 0.108″ for 2020.0, but by the time the system was based on the magnitude of the double stars, seeing and measured the prediction was approaching 0.101″, which wavelength. Drift calibrations were performed to deter- is below the MWO 60-inch telescope Rayleigh limit mine the pixel scale and camera angle. In most cases, shown in Table 1. Therefore, the binary was not re- reference stars were observed in order to conduct de- solved but “elongated”, as shown in Figure 3. This convolution, which removes the effects of optical aber- elongation indicates that the separation of A2069 was rations and some atmospheric effects. The deconvolu- less than 0.107″, the Rayleigh limit for the MWO tele- tion requires the reference stars to be single stars, ideal- scope. ly within 4 degrees of the target. It is important to col- The systems that were resolved were analyzed with lect the reference star data used for deconvolution with- autocorrelation and bispectrum analysis using STB in about 10 minutes of acquiring the target star to re- 1.14. One of the systems, STF 1609AB, was a triple duce the chances of a change in atmospheric conditions system. The autocorrelation and bispectrum results of between the two measurements. this system are shown in Figure 4. Speckle interferometry had not been done using the MWO 60-inch telescope before the engineering run of May 24. At that observing session, the bright star Arc- turus initially did not come into focus on the camera detector. Focus was apparently inside the focuser, and immoveable. Tom Meneghini and Blake Estes later discovered that the focus problem was caused by a malfunction jamming the mechanism which supports and moves the secondary mirror, which is the way the telescope is focused. As a work-around, a set of lenses was employed. This brought focus out to the camera, but also reduced the effective focal length and caused con- siderable optical distortion. Therefore, the data collected during the engineering run were ill-suited for Figure 3: Elongated bispectrum image analysis, though the speckle process was still demon- of A2069 strated for the Zoom audience.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 466

Speckle Interferometry of Close Doubles on the Mount Wilson 60” Telescope A Live Virtual Star Party

Figure 4: A 1609ABC Single image (left), autocorrelation (center) and bispectrum reconstructed image (right). In the Autocorrelation, the two nearly equal components (AB) create correlations with the C component as well (upper right and lower left).

Number of Position Standard Standard Tele- Separa- System Date Fits Cu- Angle error on Error on scope* tion (″) bes (o) PA (o) Sep (″)

STF 1527 2020.34 OCA 22 4 302.24 0.13 0.493 0.012 WDS 11190+1416

WRH 12 2020.34 OCA 22 4 8.02 0.28 0.352 0.015 WDS 12349+2238

HU 572 2020.34 OCA 22 4 329.40 0.59 0.554 0.006 WDS 13091+2127

STF 1670AB1 2020.36 ASD 11 1 356.17 NA 2.880 NA WDS 12417-0127

TOK 406 2020.45 MWO 60 1 14.36 NA 0.144 NA WDS 14382+1402

A 1609AB 2020.45 MWO 60 1 87.39 NA 0.256 NA WDS 13258+4430

A 1609 AC 2020.45 MWO 60 1 220.69 NA 2.618 NA

A 1609AB, C 2020.45 MWO 60 1 222.27 NA 2.647 NA WDS 13258+4430

A 2069 2020.45 MWO 60 1 <0.107 WDS 14267+1625

*OCA 22: Orange County Astronomers 22-inch Kuhn telescope. MWO 60: Mt. Wilson Observatory 60-inch telescope. ASD 11: Arizona Sonoran Desert Observatory of Glendale, Celestron 11-inch telescope. 1 STF 1670AB’s was measured with STB1.05 autocorrelation only, because STB1.14 does not work without a reference star. Table 2: Measurements of position angle and separation for the eight star systems studied here.

The astrometric measurements obtained are sum- each measurement to its corresponding residual, based marized in Table 2. Note that it was not possible to on the orbital ephemeris predicted by Bill Drummond’s report standard errors for the measurements of all of the spreadsheet, is shown in Table 3 (Drummond, 2020). pairs, as only a single fits cube was obtained in several Note that most of these stars were not resolved by Gaia cases. Ideally, it would be best to average the results DR2 (except for A1670AC and A1609AB,C), so it was from about five 1000-frame fits cubes. For the OCA 22 not possible to compare the measurements to each -inch telescope, the astrometry was averaged from fits pair’s corresponding measurement in Gaia DR2. cubes taken in four Sloan filters. A comparison of

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 467

Speckle Interferometry of Close Doubles on the Mount Wilson 60” Telescope A Live Virtual Star Party

System Position Angle (o) Residuals (o) Separation (″) Residuals (″) Predicted 302 -- 0.46 -- STF 1527 Autocorrelation 302.52 0.52 0.490 0.030 Bispectrum 302.24 0.24 0.493 0.033 Predicted 9.1 -- 0.325 -- WRH 12 Autocorrelation 8.06 -1.04 0.406 0.081 Bispectrum 8.02 -1.08 0.352 0.027 Predicted 6.2 -- 0.129 -- TOK 406 Autocorrelation 12.94 6.24 0.141 0.012 Bispectrum 14.36 8.16 0.144 0.015 Predicted 357 -- 2.975 -- STF 1670AB Autocorrelation 356.17 -0.83 2.880 -0.095 Bispectrum ------Predicted 328 -- 0.554 -- HU 572 Autocorrelation 327.72 -0.28 0.533 -0.021 Bispectrum 329.4 1.4 0.554 0 Predicted 89.3 -- 0.244 -- A 1609AB Autocorrelation 87.7 -1.56 0.256 0.011 Bispectrum 87.39 -1.91 0.256 0.012

Table 3: Predicted position angles and separations based on Bill Drummond’s spreadsheet and the WDS Sixth Orbit Catalog Orbital Elements.

For the pairs observed at the OCA 22-inch tele- Analysis scope (STF 1527, WRH 12, and HU 572), the use of Delta magnitude normally decreases as the filter multiple filters made possible the bispectrum analysis wavelength increases, but as seen in Table 4, the oppo- of approximate delta magnitude (Δm = mB-mA) values site was the case for STF 1527. This would suggest ei- at different wavelengths, as shown in Table 4. The un- ther that the primary is a red giant or that the secondary certainties of delta magnitude in bispectrum may be is a red dwarf. The system being 1400 lightyears away large, particularly when only a single observation is makes it too far for a white dwarf to be visible, espe- made and the colors are not transformed to a standard cially with such a bright primary, and the WDS says the photometric system, as is the case here. primary is spectral type A0IV. This confirms that it is a Filter Sloan g’2 Sloan r’2 Sloan i’2 Sloan z’2 sub-giant running out of hydrogen and beginning to Center WL 475nm 630nm 770nm ~900nm leave the main sequence. Most likely, both components FWHM 149 nm 133 nm 149 nm ~160 nm were originally type B stars, but the primary was a STF 1527 0.68 0.55 0.70 0.73 more massive, hotter, earlier B type. Now the primary has run out of fuel first, expanded, and cooled down to WRH 12 1.51 1.84 2.04 2.21 type A0, it’s redder than the B companion but enlarged HU 572 0.76 0.62 0.65 0.87 to still be about 2 mags brighter. As seen in Figures 5 - 10 below, the measurements Table 4: Delta magnitude (secondary - primary) of STF 1527, of all pairs reported here fall within 0.1 arcsecond of WRH 12, and HU 572 in multiple Sloan filters. For each filter, their predicted locations. In most cases, this corrobo- the center wavelength and full-width-half-max (FWHM) band- rates both the accuracy of the orbits and the accuracy of pass are given in the headings. For the IR long-pass z’2 filter, the long WL end is determined by the detector sensitivity limit, the measurements. Specifically, our points for which is assumed to be 980 nm. STF1527, STF1670AB, HU572, and A1609AB are all close to their predicted orbit points. However, there is a scattered trend with respect to time in the plotted

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 468

Speckle Interferometry of Close Doubles on the Mount Wilson 60” Telescope A Live Virtual Star Party measurements of WRH 12 (Figure 6 at right), as demonstrated by the interspersed locations of the darker and lighter points on the historical data plot. Therefore, it remains somewhat unclear whether the corresponding locations between the measurement point and predicted point are coincidental. Similarly, for TOK 406, the four earliest-date points of the historical data file had to be quadrant-flipped in order to match their positions on the WDS orbital plot, as shown in Figure 7. A new paper in prep by Tokovin may soon update the orbital elements of this system (Tokovin, 2020). Figure 8: Left: WDS orbital plot of STF 1670AB. Right: the historical data along with the measurement (green square), mostly hidden behind the prediction based on the WDS orbital ephemeris (orange square).

Figure 5: Left: WDS orbital plot of STF 1527. Right: the historical data along with the measurement (green square), and the prediction based on the WDS orbital ephemeris (orange square). For this system, the “measured” and “predicted” points are nearly identical, so the two position markers are overlapping. Figure 9: Left: WDS orbital plot of HU 572. Right: the historical data along with the measurement (green square), mostly hidden behind the prediction based on the WDS orbital ephemeris (orange square).

Figure 6: Left: WDS orbital plot of WRH 12. Right: the historical data along with the measurement (green square), and the prediction based on the WDS orbital ephemeris (orange square).

Figure 10: Left: WDS orbital plot of STF 1609AB. Right: the historical data along with the measurement (green square), mostly hidden behind the prediction based on the WDS orbital ephemeris (orange square). Figure 11 shows the orbital plot for the system whose position angle and separation measurements are not reported here: A 2069. The secondary star’s close proximity to periapsis on the night of our observation (predicted at 0.101″) was within the 0.107″ Rayleigh Figure 7: Left: WDS orbital plot of TOK 406 with the prediction in limit of the 60-inch MWO telescope. yellow, autocorrelation measurement as the open green triangle, and two separate bispectrum measurements in green and red on top of each other. Right: Data from the historical data file with the prediction in orange and the bispectrum measurement in green. Note that the earliest four points in this data set were quadrant flipped in the historical data file; they are here shown corrected.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 469

Speckle Interferometry of Close Doubles on the Mount Wilson 60” Telescope A Live Virtual Star Party

capture. Wonderplanets.de Genet, R., Rowe, D., Meneghini, T., Buchheim, R., Estrada, R., Estrada, C., Boyce, P., Boyce, G.,Ridgely, J., Smidth, N., Harshaw, R., and Ken ney, J., 2016, “Mount Wilson 100-inch Speckle In terferometry Engineering Checkout.” JDSO, 12, 263 -269, March 2016. Harshaw, Richard and Rowe, David and Genet, Russell, 2017, “The Speckle Toolbox: A Powerful Data Re duction Tool for CCD Astrometry.” JDSO, 13, 52- Figure 11: Left: WDS orbital plot of A 2069. Right: the histori- 67, January 1, 2017. cal data along with the predicted position of the secondary Merrill, P. W., 1921, “Interferometer Observations of based on the WDS orbital ephemeris (orange square). Double Stars”, PASP, 33, 209. Conclusion Michelson, A.A., & Pease, F.G., 1921, “Measurement Using speckle interferometry enables reaching a of the Diameter of α Orionis with the Interferome telescope’s diffraction limited resolution, making very ter”, ApJ, 53, 249. close double stars accessible with a high speed camera. Rowe, David A. and Genet, Russell M., 2015, “User’s As a result of our study, 6 double star systems were Guide to PS3 Speckle Interferometry Reduction successfully resolved using both bispectrum and auto- Program”, JDSO, 11, 266-276. correlation methods. Of the successfully separated sys- Rowe, D., 2017, “WDS1.2 Search Program.” Private tems, observed separation values were very close to the communication. predicted values and most were consistent with orbital Rowe, D., Genet, R., 2020. “STB1.13 Bispectrum Pro solutions obtained from the WDS. cessing,” in preparation for submission to JDSO. Simmons, Mike, 2020. “Building the 60-inch tele Acknowledgements scope.” Mount Wilson Observatory website, re This research was made possible by the Washing- trieved July 2020. ton Double Star catalog maintained by the U.S. Naval Tokovin, Andrei, Brian D. Mason, Rene A. Mendez, Observatory, along with the Stelledoppie catalog main- Edgardo Costa, Elliott P. Horch. “Speckle Inter tained by Gianluca Sordiglioni. ferometry at SOAR in 2019”, in preparation. The team would like to acknowledge the generous Trimble, Virginia, 1995. “The 1920 Shapley-Curtis support from Mount Wilson Observatory, for allowing Discussion: Background, Issues, and Aftermath”, us to do an engineering run in preparation for our Live Publications of the Astronomical Society of the Pa Virtual Star Party Research Programs. cific, V. 107, pp. 1133. Thanks also to Jimmy Ray of the ASDOG Observa- tory for generously donating his time and his telescope for use in this project, and to the Orange County As- tronomers for use of their telescope and facilities. References Dainty, J. C. 1981. Speckle Interferometry in Astrono my. Symposium on Recent Advances in Observational Astronomy, Ensenada, Mexico. 95- 111. Drummond, Bill, 2020. Personal communication. Labeyrie, Antoine. 1970. Attainment of Diffraction Limited Resolution in Large Telescopes by Fourier Analysing Speckle Patterns in Star Images. Astrono my & Astrophysics. 6, 85-87. Anderson, J.A., 1920, “Application of Michelson’s In terferometer Method to the Measurement of Close Double Stars.” Contributions from Mount Wilson Observatory, No. 185, ApJ, 51, 263-275. Edelmann, Torsten, 2017. FireCapture2.6. http:// fire

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 470

Observations of Potential Gaia DR2 Red Dwarf Binary Stars in the Solar Neighborhood – II

1 2 3 Ivan Altunin , Rick Wasson , and Russell Genet

1. University of California, Berkeley, [email protected] 2. Orange County Astronomers, Murrieta, California, [email protected] 3. California Polytechnic State University, San Luis Obispo, [email protected]

Abstract: Astrometric and photometric observations of double stars within 100 parsecs of the Earth, which included a late K or M dwarf component, were made in Sloan g’ r’ i’ z’ filters with the Orange County Astronomers 22-inch telescope in 2019 using quasi-speckle interferometry techniques where exposures up to 1 second long were processed with standard speckle bispectrum methods. This paper reports continuation of those observations in 2020 for 12 more binaries at the Fairborn Institute Robotic Observatory’s remotely accessed 11-inch telescope. Longer quasi-speckle integrations—up to 5 seconds—were used with a telescope of half the aperture and 1/4 the photon-capturing area, reaching about the same faint magnitudes (G=14.8) without introducing too much atmospheric smearing. Gaia summary data are re- viewed for the observed stars, bispectrum reconstructed images are presented, astrometric and photometric results are reported, and details of several systems are discussed.

Introduction same target list, filters, and data reduction methods, but Although the Red Dwarf (RD) stars—spectral a telescope with half the aperture, the 11-inch tele- types late K and M on the Main Sequence—are the scope at the Fairborn Institute Robotic Observatory most common stars in the Milky Way, they are intrinsi- (FIRO). cally faint and more difficult to observe than earlier Targets included systems containing one or two types, so their population, even within the Solar Neigh- RD components already in the Washington Double borhood, has not been fully explored (Cooper 2019). Star (WDS) catalog, as well as systems not in the WDS Recent advances in low cost, red-sensitive, high speed but in the Gaia Double Star (GDS) data base derived but low-read-noise, back-side illuminated CMOS cam- from Gaia Data Release 2 (DR2). Most of these sys- eras has enabled the study of brighter RD stars with tems have few previous observations—often by large- small telescopes. Astrometric observation of binaries scale surveys such as SDSS and 2MASS, which have with one or both components being a RD can lead to limited resolution (because of over-exposure in very orbital solutions and the determination of the mass ra- deep images) to obtain separation and position angles tio of the two stars, while photometry of the binary between the components or multi-band photometry of components can provide approximate temperatures and the individual components. The current observations spectral types. Taken together, these observations can offer confirmation of the previous astrometric observa- provide estimates of the mass of each star (Davidson et tions, the possibility of detecting orbital motion, and al. 2009). The RECONS survey has completed the cen- new Δ magnitudes in the Sloan bands. sus of all stars within 10 parsecs of the Sun, and the The Gaia astrometric/photometric satellite has pro- campaign to discover all the stars within 25 parsecs is vided observations for a comprehensive stellar catalog, continuing (Winters et al. 2019). Gaia DR2, which was made public in 2018. A sub- A campaign to observe the brightest nearby binary catalog, GDS1, has been developed (Rowe 2018) that systems containing a RD was initiated with the Orange contains 6.8 million Gaia double stars suitable for ob- County Astronomers 22-inch telescope in 2019 servation with smaller telescopes with separations less (Wasson et al. 2020, hereinafter RDSN-I). The obser- than 10". The accompanying GDS1 Tool was used to vations presented here continue that program, using the locate potential binary stars in the Solar neighborhood

Received October 26, 2020

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 471

Observations of Potential Gaia DR2 Red Dwarf Binary Stars in the Solar Neighborhood – II

that had accurate parallaxes, proper motions, and color extended to observations with a telescope of half the indices, and might contain a RD star. aperture (11 inches) and 1/4th the photon-capturing ar- Using GDS1, all stars with parallax greater than 10 ea. Hopefully, the same faint magnitudes could be milli-arcsec (100 parsecs) as well as those in the south- reached, albeit with longer integration times and with- ern sky (to avoid high airmass) were discarded, still out introducing too much atmospheric smearing. leaving several thousand double star candidates. Poten- The third objective was to further refine remote tial binaries, as opposed to chance optical doubles, were real-time speckle interferometry observational tech- identified as pairs having very nearly the same Gaia niques. These observations were initiated by Marchetti parallax and similar proper motion. et al. (2020) on the 11-inch telescope at the Fairborn The Gaia H-R diagram, published as part of DR2, Institute Robotic Observatory (FIRO). shows that the division between K and M dwarfs occurs at Gaia color index (Bp-Rp) = 2.3. Therefore, systems Image Processing Technique that possibly contain a RD (late K through M) were In the previous RDSN-I observations with the OCA identified by (Bp-Rp) > 2. Luminosity class was as- 22-inch telescope, it was originally hoped that speckle sumed to be normal main sequence dwarf (V) because interferometry could be employed for diffraction- none of these systems contained a component brighter limited observations (>0.35” in the i’ band) which than 10th magnitude. Thirty-five of the brightest candi- might reveal additional RD stars unresolved by Gaia. dates were observed in RDSN-I; about half of these However, because of the short exposures required to were not found in the Washington Double Star Catalog “freeze” the seeing and preserve the diffraction-limited (WDS) and may be considered “new” Gaia double speckle information in each frame, very few of the RD stars. An additional dozen candidates were observed in binaries were bright enough to be observable with such RDSN-II (this paper) with, again, half not being in the short exposures. Thus, most candidates required longer WDS, although all but one of these were found in the exposures, up to 1 second. These long exposures were newly revised Washington Double Star Supplement blurred by atmospheric motion, and thus were seeing- (WDSS), which was not surprising since many of the limited rather than diffraction-limited. WDSS entries consist of just Gaia observations of both Nevertheless, speckle processing techniques were components. still found to be useful – capable of resolving stars Quasi-speckle interferometry employs a single ref- more clearly with separations at or somewhat below the erence star for deconvolution, and bispectrum analysis seeing limit. Apparently, there is still information about of double star observations for recovery of phase infor- the large-scale structure (and perhaps some small-scale mation distorted by the atmosphere. Quasi-speckle in- structure as well) in these long-exposure images, which terferometry appears to provide better results than ei- can be correlated by speckle processing better than the ther simple stacking or lucky imaging. This somewhat usual methods of stacking or lucky imaging. The obser- surprising result may be due to: (1) the retention of vation of a single star used for mathematical deconvo- some moderate-scale information in these somewhat lution in speckle processing—standard practice in longer-exposure images which can be correlated by speckle interferometry—can, to some extent, cancel speckle processing better than by stacking or lucky im- optical aberrations and systematic atmospheric phe- aging; (2) single-star deconvolution which can, to some nomena such as dispersion. Finally, using all the frames extent, cancel optical aberrations and systematic atmos- probably increases signal/noise if most of them contain pheric phenomena such as dispersion, and (3) pro- some valid image information. Further exploration of cessing of the observations in the Fourier frequency the quasi-speckle interferometry technique is planned. domain which allows suppression of noise at spatial Instrumentation and Procedures frequencies that cannot be real, e.g., smaller than the The Fairborn Institute Robotic Observatory (FIRO) Airy disk or larger than the frame size. is equipped with the 11-inch telescope shown in Figure Objective 1. The optics are Celestron C-11 Schmidt Cassegrain The primary objective of this project was to contin- with an f/10 focal ratio and an effective focal length of ue the astrometric and photometric observations in mul- 2800 mm. The custom equatorial “L” mount was de- tiple Sloan filters begun in RDSN-I on the OCA 22- signed and built in conjunction with California Poly- inch telescope. The OCA telescope became temporarily technic State University students and staff. A Sidereal inoperative due to equipment difficulties. Technology Servo Controller II controls the telescope. The second objective was to determine whether the quasi-speckle interferometry technique could be

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 472

Observations of Potential Gaia DR2 Red Dwarf Binary Stars in the Solar Neighborhood – II

the double star separation limit for the system being ~ 1.0" instead of the diffraction limit of ~ 0.6". The Windows 10 computer, located in a cabinet just below the telescope, is a Dell OptiPlex 7070 with an Intel Core i7-9700 (8 cores operating at 4.7 GHz), 16GB DDR4 memory, an internal 256GB solid state drive, and two external (USB 3.2) 256GB solid state drives. Remote access to the system is via a static Inter- net IP address. Speckle interferometry observations are made remotely in real time using a Sidereal Technology software program, SiTech ZWOCam, that controls the camera, filter wheel, focuser, and telescope. The program’s ability to capture short exposures at high speed is critical for speckle interferometry. PlateSolve 3 allowed targets to be precisely centered within a se- Figure 1: The “L” of the L mount is on the lectable, small region of interest. An automatic focusing left sitting on top of an equatorial wedge, subroutine provides sharp images. Finally, a scripting while the Celestron C-11 optics is on the feature allows a sequence of observations with different right. filters and integration times to be repeated any desired number of times. The instrument payload, shown in Figure 2, consists of a focuser, filter wheel, and camera. The Clem- Remote Observation Methods ent “scissors” focuser is controlled by a gearhead servo- A typical observing sequence began with manually motor and a second Sidereal Technology Servo Con- opening the observatory and powering the equipment. troller II. The five-position ZWO motorized filter wheel The remotely located observers were connected to contains a clear “filter,” and four Sloan filters: g’, r’, i’, FIRO’s computer over the Internet via AnyDesk. and z’. The CMOS camera is a thermoelectrically- SiTech ZWOCam was launched, and the camera, filter cooled (40-50°C below ambient) ZWO ASI 1600 MM, wheel, and focuser activated. selected for its low read noise (only 1.2e), small pixel The telescope was then unparked and tracking was size (3.8 microns), and relatively large format (16 meg- initiated. Since the telescope is parked within the al- apixels, 4636x3520). lowed observing window, an automatic focus routine and plate solution readies the telescope for observations before slewing to the first target. After the target coordinates were entered, the telescope moved to the target and, after another plate solution and a short move within the same full field, the target was always near the center of a small region of interest (RoI), typically 256x256 pixels (about 70”x70”). An imaging sequence was set up that included the exposure times, camera gain, and number of exposures for each filter in the sequence (typically a few hundred to a thousand). This filter sequence was set to repeat the desired number of times (typically five) to provide error Figure 2: The instrument cluster in- statistics and improved precision. A path for storing cludes a Clement focuser, a ZWO filter images on an external drive was defined, and the obser- wheel, and CMOS camera. vations were initiated. During the observations, small Although the pixel scale of approximately 0.28"/ corrections in RA were sometimes required to keep the pixel is somewhat under-sampled, the full-frame field- target near the center of the region of interest. After the of-view of 21.5x16.3 arcminutes allows plate solutions sequence was completed for the double star, the coordi- for precise positioning and astrometric calibration to be nates of the reference star were entered, and the entire made without the need for a second, wider-field cam- process was repeated, although only one sequence in era. This one-camera-does-it-all compromise results in each filter was required for deconvolution.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 473

Observations of Potential Gaia DR2 Red Dwarf Binary Stars in the Solar Neighborhood – II

thus providing potential additional astrometric and pho- Astrometric Calibration tometric measures. WDSS-1 indicates that only the sin- Pixel scale and camera orientation calibration were gle Gaia observation had been reported before the new accomplished with the Plate Solving feature of the measures in this paper. SciTech ZWOCam program. As noted above, a full frame image of the target field, roughly 21’x16’, was routinely taken to locate the center of the field after a slew. The solution relied on the UCAC4 Catalog stars to magnitude ~15. Typically, in 5-second exposures, 50 to several hundred stars were available for solution, yielding high accuracy. The instrumentation was not touched during the several weeks of observation, so the calibration was very repeatable. The calibration was checked each night, and a total of ten Plate Solve frames were saved and processed again to provide overall statistics. The average pixel scale was 0.2779”/pixel ± 0.0004. The average camera orientation on the sky, which had been carefully aligned prior to the observations, was 0.092 Figure 3: The final step in the bispectrum analysis reduction is degree ± 0.039. Because the camera angle correction astrometric and photometric measurement of the system. The was so small (<0.1 degree) compared with the uncer- green circle is the astrometric/photometric aperture for the Pri- tainty of PA measurements, it was ignored in the results mary star (“Reference”), which is always centered in the image. shown in Table 2. The purple aperture is for the secondary star (“Target”), and yellow for the Background. The average Background pixel ADU Data Reduction level is subtracted from the ADU signal of each pixel within the star apertures; the two apertures must be the same size for prop- This research made use of Speckle Toolbox 1.14 er photometric measurement of Δ magnitude. (Rowe 2020) using the bispectrum analysis (BSA) method so that the target components could be resolved Observations as individual stars, enabling our measurements without Observations made of the target stars in Table 1 are the 180o ambiguity of autocorrelation. The first step in shown in Table 2. Only a single observation, consisting the data reduction was to take all the images of each of several hundred to a thousand frames, were made in star in each filter and process them into a single FITS each filter. Because most of these systems were faint Cube. Bispectrum analysis was initiated by taking the (RD component 11 < G < 15), the quality of the obser- Fourier transform and calculating the triple correlation vations varied widely, generally based on brightness of coefficients for all images in each FITS cube. These the secondary star at each wavelength, and on the see- coefficients were then averaged, and the average power ing. Therefore, a Quality Code was included in Table 3, spectrum was also calculated. The bispectrum phase offered as an aid to interpreting these observations: was then reconstructed by iteration to form the image. 1. Good quality, comparable with moderate resolu- Photon bias was removed, and the image sharpened tion speckle. with adjustment of the Gaussian lowpass and high-pass 2. Lower quality, faint, or seeing-smeared images. filters. Finally, the astrometry and photometry measure- 3. Marginal quality, because of low S/N of very ments were obtained as shown in Figure 3. faint or smeared images. Doubles Observed 4. Poor quality, secondary star barely detectable The Gaia DR2 characteristics of the doubles stud- above the noise. ied are summarized in Table 1. These pairs were chosen 5. Faint component not detected. from the same target list developed for RDSN-I. The far-right column gives the WDS identification for those It was assumed that the measurements in separa- stars previously identified. “WDSS-n” indicates that the tions and position angle did not change systematically star is in the WDS Supplemental (WDSS) Catalog, and in different wavelengths. Thus, the measurements in “n” is the number of previous observations given there. each filter were treated as “repeat” observations, allow- Most of the WDSS stars were first identified as double ing us to find the mean and standard deviation between by the Gaia satellite in DR2. Stars that have r > 3” may them. This assumption has not been explored or proven have been imaged, but not identified, by earlier surveys, and warrants further investigation.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 474

Observations of Potential Gaia DR2 Red Dwarf Binary Stars in the Solar Neighborhood – II

Coords (2000) ρ GA (Bp-Rp)A S A PMRAA PMDA πA ΤA WDS UCAC4 θ GB (Bp-Rp)B S B PMRAB PMDB πB ΤB Disc Period

170648.9+321159 3.279 10.779 1.822 K 53.2 -74.7 31.93 4153 WDSS-2 1206 612-054165 27.3 12.624 2.310 K/M 46.0 -82.8 31.93 3471

173307.3+091437 4.689 9.199 1.401 K 31.0 -21.2 31.32 4522 WDSS-2 1905 497-071347 72.3 13.022 2.751 M 26.2 3.3 31.41 3757

182740.7+501613 4.917 10.111 1.574 K 194.97 84.35 24.75 4231 18277+5016 2926 702-060220 236.7 12.14 2.568 M 200.44 86.87 24.3 3394 LEP 158

No 183513.4+241839 2.468 11.644 2.415 M -109.21 -273.73 39.70 3987 18352+2419 Estmt 572-068539 4.6 12.419 2.608 M -121.31 -277.05 39.63 3670 KPP3331

184930.9+414635 4.179 9.624 1.195 G -5.69 -21.11 20.37 4933 WDSS-1 2762 659-067595 89.5 12.303 2.352 K/M -6.34 -24.81 20.57 3600

185453.7+105842 3.808 8.793 1.750 K 29.56 132.01 53.77 4106 18550+1058 648 505-095824 44.9 11.437 2.586 M 29.48 84.21 53.88 3931 VYS8

No 191659.0+022216 4.388 12.041 1.534 K 56.3 -23.8 10.24 4364 WDSS-3 Estmt 462-091772 311.6 14.637 2.400 M 57.8 -24.4 10.36 3729

192558.0+355453 6.038 9.083 1.164 G -90.7 -161.0 25.03 4894 19260+3555 3461 630-070424 106.3 11.597 2.079 K/M -100.0 -156.9 25.04 4046 BU 1286 AC

195255.2+431624 5.091 10.081 1.203 G 121.7 106.5 16.30 4828 19529+4316 5266 667-080219 257.1 13.216 2.047 K/M 122.0 102.1 16.28 4185 KPP3349

202201.6+214720 5.444 11.273 2.275 K/M 2.48 -131.67 37.17 3744 WDSS-3 2421 559-117009 67.46 14.841 3.716 Late M 4.67 -133.83 37.15 3590

203335.4+385341 2.654 12.214 2.430 M 110.11 73.50 26.98 3398 20336+3854 1174 645-095063 79.39 12.375 2.458 M 103.95 56.08 27.16 3376 KPP4205

204528.2+370546 4.892 12.228 1.945 K 60.30 6.66 11.56 4030 None 8660 636-098623 318.26 12.748 1.935 K 65.74 8.30 12.09 3933

Table 1: Gaia DR2 characteristics of the candidate binaries observed. Left column: Primary RA and Dec coordinates for 2000.0 (hhmmss.s+ddmmss format)[above] and UCAC4 Catalog ID of Primary [below]. Column 2: Gaia DR2 separation (r) in arc-sec [above] and Gaia position angle ( ) in degrees [below]. The remaining seven columns give Gaia data for the Primary (A) [above] and Secondary (B) [below]: G magnitude, (Bp-Rp) Color Index, approximate Spectral Type based on the Gaia H-R diagram, Proper Motion in RA (mas/year), Proper Motion in Declination (mas/year), Parallax (mas), effective surface Temperature (K), WDS and Discovery designations if the pair is already recognized as a double star in the WDS Catalog (WDSS-n indicates an entry in the WDS Supplementary Catalog, where n is the total number of observations), and the estimated Period in years. Late M Dwarfs are noted in bold.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 475

Observations of Potential Gaia DR2 Red Dwarf Binary Stars in the Solar Neighborhood – II

aperture photometry. The net flux “signal” for each star Differential Photometry from Bispectrum in the reconstructed image is the sum of ADU counts Analysis from all pixels inside a circular aperture centered on the For cool red stars, observations at longer wave- centroid of the star, with the background level for the lengths are useful not only because they produce most same aperture area subtracted. However, the back- of their radiation in near-IR, but also because the at- ground is measured in a separate “blank” area of the mospheric seeing is better, causing less distortion dur- image, not in an annulus surrounding the star because ing exposures. The RDSN-I observations were made of the likelihood that the companion star, the Airy using Sloan filters (Astrodon Gen_2 interference fil- rings, or other diffraction artifacts may fall in or near ters), and the FIRO speckle instrumentation includes the annulus. The Δ magnitude (secondary-primary) is the same Sloan filter set from Astrodon. For the stars calculated automatically by STB from the ratio of net observed in this paper, the best combination of RD tar- counts of each star. get/detector sensitivity/seeing was generally in the i’ During reduction, the same photometry aperture band. was used for both stars, to capture the same proportion Bispectrum analysis recovers the approximately of light for each star; use of different apertures would correct flux and Δ magnitude in each filter. Bispectrum contain different percentages of total light, causing an photometry is like the method typically used in CCD error in Δ magnitude. A large aperture that included all

Date Coord (2000) Filter θ r Δ Mag Q 2020.589 173307.3+091437 Clear 70.82 4.487 4.61 4 UCAC4 497-071347 r’ 69.15 4.839 4.45 4 i’ 70.83 4.558 3.33 2 z’ 69.49 4.723 2.87 1 Average 70.07 4.652 Std Dev 0.88 0.159 Filter Clear r’ i’ z’

Images

Date Coord (2000) Filter θ r Δ Mag Q 2020.611 182740.7+501613 Clear No Observations UCAC4 702-060220 r’ 238.56 4.910 3.50 4 WDS: 18277+5016 i’ 235.50 5.054 2.79 2 Discov: LEP 158 z’ 239.53 5.196 2.49 3 Average 237.86 5.053 Std Dev 2.10 0.143 Filter Clear r’ i’ z’

Images

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 476

Observations of Potential Gaia DR2 Red Dwarf Binary Stars in the Solar Neighborhood – II

Date Coord (2000) Filter θ r Δ Mag Q 2020.591 183513.4+241839 Clear 4.24 2.530 1.48 2 UCAC4 572-068539 r’ 5.19 2.465 1.45 2 WDS: 18352+2419 i’ 3.87 2.479 0.97 1 Discov: KPP 3331 z’ 6.24 2.401 1.30 1 Average 4.89 2.469 Std Dev 1.06 0.053 Filter Clear r’ i’ z’

Images

Date Coord (2000) Filter θ r Δ Mag Q 2020.600 184930.9+414635 Clear No Observations UCAC4 659-067595 r’ 89.85 4.180 3.03 2 i’ 89.94 4.217 2.19 1 z’ 89.76 4.203 1.77 1 Average 89.85 4.200 Std Dev 0.09 0.019 Filter Clear r’ i’ z’

Images

Date Coord (2000) Filter θ r Δ Mag Q 2020.600 185453.7+105842 Clear No Observations UCAC4 505-095824 r’ 48.10 3.620 3.50 1 WDS: 18550+1058 i’ 47.26 3.735 2.76 1 Discov: VYS8 z’ 48.19 3.453 2.15 3 Average 47.85 3.603 Std Dev 0.51 0.142 Filter Clear r’ i’ z’

Images

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 477

Observations of Potential Gaia DR2 Red Dwarf Binary Stars in the Solar Neighborhood – II

Date Coord (2000) Filter θ r Δ Mag Q 2020.611 191659.0+022216 Clear No Observations UCAC4: 462-091772 r’ 311.61 4.198 3.47 2 i’ 312.18 4.667 3.39 2 z’ Secondary Not Detected 5 Average 311.90 4.433 Std Dev 0.40 0.332 Clear r’ i’ z’

Date Coord (2000) Filter θ r Δ Mag Q 2020.630 192558.0+355453 Clear No Observations UCAC4: 630-070424 r’ 106.52 5.961 3.48 2 WDS: 19260+3555 i’ 106.15 5.972 2.66 1 Discov:BU1286AC z’ 106.32 5.914 2.43 1 Average 106.33 5.949 Std Dev 0.19 0.031 Clear r’ i’ z’

Date Coord (2000) Filter θ r Δ Mag Q 2020.622 195255.2+431624 Clear No Observations

UCAC4: 667-080219 r’ 255.92 5.109 3.58 2 WDS: 19529+4316 i’ 257.57 5.108 3.00 1 Discov: KPP3349 z’ 257.82 5.081 2.61 1 Average 257.10 5.099 Std Dev 1.03 0.016 Clear r’ i’ z’

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 478

Observations of Potential Gaia DR2 Red Dwarf Binary Stars in the Solar Neighborhood – II

Date Coord (2000) Filter θ r Δ Mag Q 2020.622 202201.6+214720 Clear No Observations UCAC4: 559-117009 r’ 67.94 5.436 3.99 2 i’ 67.41 5.490 3.43 2 z’ 67.37 5.440 3.15 2 Average 67.57 5.455 Std Dev 0.32 0.030 Clear r’ i’ z’

Date Coord (2000) Filter θ r Δ Mag Q 2020.611 203335.4+385341 Clear No Observations UCAC4: 645-095063 r’ 80.65 2.650 0.69 1 WDS: 20336+3854 i’ 79.59 2.643 0.34 1 Discov: KPP 4205 z’ 79.99 2.628 0.24 1 Average 80.08 2.640 Std Dev 0.54 0.011 Clear r’ i’ z’

Date Coord (2000) Filter θ r Δ Mag Q 2020.622 204528.2+370546 Clear No Observations UCAC4: 636-098623 r’ 318.7 4.813 0.84 1 i’ 317.97 4.831 0.71 1 1.4 z’ 319.15 4.935 3

Average 318.61 4.860 Std Dev 0.60 0.066 Clear r’ i’ z’

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 479

Observations of Potential Gaia DR2 Red Dwarf Binary Stars in the Solar Neighborhood – II

Date Coord (2000) Filter θ  Δ Mag Q 2020.591 170648.9+321159 Clear 26.15 3.222 2.09 3 UCAC4 612-054165 r’ 25.23 3.326 2.59 1 i’ 28.54 3.301 2.13 1 z’ 26.45 3.225 1.92 1 Average 26.59 3.269 Std Dev 1.40 0.053 Filter Clear r’ i’ z’

Images

Table 2: Observed characteristics of the stars using the BiSpectrum Analysis (BSA) method. The first column is the Besselian observation date. Column 2 includes RA and Dec Coordinates (2000.0) in hhmmss.s+ddmmss format [above], and UCAC4 Catalog number of the primary star [below]; WDS and discovery designations are also given for previously known doubles. The remaining columns are Filter, observed Position Angle (q) in degrees, observed Separation (r) in arc-sec, BSA Δ magnitude (secondary-primary), and Observation Quality Code (defined above). The Average and Standard Deviation results include all reported data, weighted equally regardless of quality. Reconstructed BSA images are shown in the last row for each system, oriented North up, East left. the starlight was not used, because it could have ies (G~14.8). included more background noise than real starlight at The ZWO ASI 1600MM camera used here has a the outer edges. Typically, slightly larger aperture di- low-noise CMOS detector like the ZWO ASI 290MM ameters were needed for the i’ or z’ filters because the used in RDSN-I, but the 1600 has a larger format, size of the Airy disk grows in proportion to wavelength. cooled sensor, and a different QE spectral response. This can also restrict astrometry and photometry of Combining the filter transmission with the detector QE close pairs which approach the system’s resolution. Ap- to estimate center wavelength (WL) and effective sensi- ertures must not overlap because the same light in the tivity for each filter are shown in Table 3. overlapping area would contribute to each star, biasing The secondary stars observed were all quite red (Bp both the astrometry and photometry results. -Rp > 2.0 in Table 1); therefore, they should all be Observations using the ZWO CMOS camera and brighter at longer wavelengths. However, the z’_2 filter Astrodon Sloan Gen_2 filter set should yield Δ magni- generally gave poorer images than in RDSN-I. This was tudes roughly comparable with the Sloan Digital Sky probably caused by one or both of the following condi- Survey (SDSS) catalog. These filters match the pass- tions: (1) the lower effective QE of the ZWO ASI bands of the SDSS catalog reasonably well, and the 1600MM camera in the z’ filter (Table 3) sometimes CMOS silicon back-side-illuminated (BSI) sensor has resulting in inadequate exposure, or (2) the wide sepa- similar sensitivity to the BSI CCD array used in SDSS, ration of some pairs might have been outside the isopla- although spectral quantum efficiency (QE) response natic patch where the atmosphere could cause non- details are somewhat different. Instrumental photome- correlated movement of the two stars’ seeing disks, try was not transformed to the Sloan or any other stand- which would diffuse and weaken the secondary star ard photometric system here, but these observations image when registered on the primary star. The effec- should supplement the Gaia G, Bp and Rp data. tive sensitivity of the 1600 CMOS camera which, like For the red secondary stars observed here, the g’_2 most CCD cameras is optimized for visible wave- filter generally gave low S/N and uncertain results in lengths, is only about 10% at the center WL of the z’ RDSN-I; the g’ band was initially tried again in this filter, whereas the 290 camera, which is optimized for study but was soon abandoned. However, the r’_2, i’_2, near-IR sensitivity, has an effective sensitivity of about and z’_2 filters usually gave satisfactory results for the 32% in the z’ filter, better suited for these faint RD long exposures—up to 3 seconds in r’ and i’, and 5 sec- stars. onds in z’—that were required for the faintest secondar-

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 480

Observations of Potential Gaia DR2 Red Dwarf Binary Stars in the Solar Neighborhood – II

Filter Center WL FWHM Max Transmission Effective QE (nm) (nm) (%) (%)

Clear ~605 ~360 95 48 g’ 485 130 98 54 r’ 624 130 98 49 i’ 758 165 99 28 z’ 880 232 99 10

Table 3: Results of Convolution of the ZWO ASI 1600MM camera quantum efficiency (QE) with transmission of the Clear and Astrodon Sloan Gen_2 filters. Columns are: filter, effective center wavelength, full-width-half-max transmission, maximum transmission, and effective QE at the center wavelength. QE characteristics of the ZWO ASI 1600MM camera were estimated from the Relative QE given in the Sony manufacturer literature, with Peak QE assumed to be 60%. QE was extrapolated beyond the 800nm limit given, to QE = 0 at 1050 nm.

Speckle differential photometry uncertainties of ferent seeing cells. In principle, speckle processing adds ~0.1 magnitude, primarily driven by Poisson statistics, the pixel brightness of the fainter star relative to the have been reported for large telescopes (Horch et al. position of the brighter star; therefore in successive 2004), while uncertainties < 0.2 magnitude were found frames, the secondary star may appear in a slightly dif- for small telescopes (Davidson et al. 2009). A Δ magni- ferent position relative to the primary if moved differ- tude goal of ± 0.1 magnitude corresponds approximate- ently by the atmosphere. This effect may produce a ly to ±10% uncertainty. If both stars contribute equally blurred or blotchy fainter representation of the second- to the uncertainty, then the flux of each star must be ary star. It may average out by taking many frames, but within ± 5% of the correct value. That is, the flux (or we generally acquired fewer frames (only a few hun- number of electrons counted) for each star must be dred) as exposure times grew longer (greater than 1 within 1/20th of the correct signal. If the error is driven second). Thus, even if averaged, differential atmospher- primarily by photon statistics, then the net flux count ic motion would spread the light of the secondary over for each star must be at least 400 electrons for un- a wider area, making it appear fainter. Although these certainty < 1 400 < 1/20. A high camera gain was adverse effects do not seem to have been too severe for generally used, giving ~0.2e-/ADU. Therefore, to the stars observed here, it could become a problem un- achieve a signal of 400 electrons, the net ADU count der other conditions. within the photometry aperture must be greater than Two of the systems previously observed in RDSN-I ADU ~ 400/0.2 ~ 2000. This net ADU count signal lev- were observed again here, approximately one year later: el was achieved for most of the primary stars, and some 170648.9+321159 and 191659.0+022216. Neither of of the brighter secondaries in the i’ and z’ filters, but these systems show movements beyond their standard not for all secondary stars in the r’ filter. Generally, of deviation () uncertainties, relative to Gaia or the 2019 course, the fainter component will contribute the larger RDSN-I measures. In Figure 4 two other systems do error for Δ magnitude. show movement greater than their 1 uncertainties for Discussion , although not a strong statistical likelihood. These two As listed in Table 1, six systems had previous en- systems are both close to us, making their movements tries in the WDS Catalog, and five of the remaining six easier to detect. Several other systems in Figure 5 have were found in the WDSS. These are new Gaia discover- small  or  uncertainties in the current measures. ies, but were previously “data-mined” and were found However, they have movements so small that no con- in the WDS Supplement Catalog, recently placed on the clusion can be made about possible orbital motions un- WDS (and Georgia State University mirror) website. til confirmed by further observations. One system, 204528.2+370546, discussed below, was 173307.3+091437 is apparently a new Gaia double found in neither catalog, so it is a new Gaia discovery star which has significant movement -2.2 ±0.9 deg in , first reported here. seen at bottom of Figure 4. The separation uncertainty Some of these systems have a separation greater is much larger than the change since Gaia. This pair than 3”, where both stars may not be in the same isopla- may have a large enough separation (~3.3”), to have natic patch. Depending on seeing, the two stars of a been resolved in earlier large survey archives, yielding wide pair may be affected somewhat differently by dif- additional measures. The large Gaia parallax is 31.93

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 481

Observations of Potential Gaia DR2 Red Dwarf Binary Stars in the Solar Neighborhood – II

only about 100 light years distant. The respective com- If this motion were part of a circular orbit, it would in- ponents are roughly K3V and M2V, based on the Gaia dicate an orbital period of roughly 780 years. The orbits H-R diagram and (Bp-Rp) colors (1.401/2.751) in Table of all the other systems observed are probably even 1. longer. Table 1 shows that the Gaia parallaxes and proper motions in RA are nearly identical. However, the proper motions in declination are significantly different; the B component is moving northward at only about 64% the speed of the primary. This is consistent with the direction of motion indicated in Figure 6, but the motion is diagonal, rather than purely north south. This may be the result of orbital motion. This system deserves continued occasional observation, to see whether the roughly linear apparent motion gradually reveals an orbit.

Figure 4. Comparison of the difference between Ob- served (2020) and Gaia (2015.5) separation and position angles. Uncertainty bars are standard deviations of the observations in Table 2. Solid blue dots indicate possible orbital change in position.

Figure 6. Historical observations of 185453.7+105842 = WDS 18550+1058 = VYS8. The current measure is shown as a red triangle. Since first discovered in 1946, it has shown roughly linear motion, in the direction Figure 5. (O-G) data for the stars within the colored box of the arrow. of Figure 4, plotted on an expanded scale. Solid blue dots indicate possible orbital change in position, i.e., greater than 1s uncertainty of r or q, or both. For the 191659.0+022216 system, 1-second exposures in r’ and i’ barely detected the secondary star, as seen in 185453.7+105842 = WDS 18550+1058 = VYS8 is Table 2. It was not detected in the Sloan z’ band, even noted as spectral type M0 in the WDS, but the compo- though the exposure was 2 seconds, because of declin- nent types—based on the Gaia H-R diagram and (Bp- ing detector sensitivity and because it was the most dis- Rp) colors (1.750/2.586) in Table 1—are roughly K5V tant observed, at 97 parsecs. Although the secondary and M1V. star is not a particularly late type M dwarf (Bp-Rp = This system has the largest movement of all those 2.4), it is very faint at G=14.6. The slightly more faint observed, shown at the upper left of Figure 4: approxi- and much later type RD discussed next was detected in mately +3.0 ±0.5 degrees in q since the Gaia measure- z’ with 5 second exposures. ment about five years earlier, or ~0.6o/year. It also has 202201.6+214720 is composed of two Red Dwarfs, but the largest movement since discovery in 1946, about of extremely different spectral types. The Gaia color 34o over 74 years, giving a more accurate 0.46o/year. indices from Table 1 are (Bp-Rp) = 2.275/3.716, corre- The separation seems to be slowly closing at about - sponding roughly to M0V and M7V. Although nearby 0.02”/year. All the historic data are plotted in Figure 6. in the Solar neighborhood at 27pcs, the late component,

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 482

Observations of Potential Gaia DR2 Red Dwarf Binary Stars in the Solar Neighborhood – II

G= 14.84, was the faintest star observed; the magnitude riods. difference (DG = 3.57) was also one of the largest observed. Long exposures, 3 seconds in r’ and i’, and 5 Acknowledgements seconds in z’, were used to successfully capture this We acknowledge David Rowe’s development of pair with the 11-inch FIRO telescope. the Gaia Double Star (GDS) data base and tool and the 204528.2+370546 is the only system not found in Speckle ToolBox (STB), both critical to the success of either the WDS or WDSS, so it is truly a new Gaia dis- this project. RG thanks Rowe for contributing the ZWO covery, not previously data mined or purposely ob- ASI 1600 camera to the Fairborn Institute Robotic Ob- served. However, it is also a rather peculiar case. The servatory, Kevin Iott for contributing the PlaneWave components are similar in Gaia magnitude and color Instruments worm-drive assemblies, and Dan Gray for index (Table 1), but the secondary appears much fainter contributing the Sidereal Technology control system. in the images of Table 2, almost disappearing in z’ We also thank Dan Gray for providing the Sidereal where it should be relatively brighter. Perhaps the sec- Technology SiTech ZWOCam software for the project ondary star is variable? The proper motions are similar, and his responsiveness to requests for making software but the parallaxes are more than 4% different and near improvements. the 100 parsecs limit of distance selection in this study. We thank those whose reviews improved this This system deserves further observation. paper. They included Richard Harshaw, Robert Buch- heim, Thomas Smith, and Vera Wallen. Conclusion Finally, we thank the European Space Agency The objectives of this study were achieved. The and the Gaia team for the use of their observations, and observations of Gaia DR2 red dwarf binary stars in the the U.S. Naval Observatory for the use of the Washing- solar neighborhood begun in RDSN-1 was successfully ton Double Star Catalog and its supplement. continued in RDSN-2 for 12 more systems that conta- tined a late K or M dwarf star. It was determined that References the quasi-speckle interferometry technique could be Cooper, K., 2019. “Meet the Neighbors,” Sky and Tele extended to observations with a telescope of half the scope, January 2019, 35. aperture and 1/4 the photon-capturing area, reaching Davidson, J., Baptista, B., Horch, E., Franz, O., & van about the same faint magnitudes with longer integration Altena, W., 2009. “A Photometric Analysis of Sev times without introducing too much atmospheric smear- enteen Binary Stars Using Speckle Imaging,” The ing. Finally, remote real-time speckle interferometry Astronomical Journal, 138, 1354. observational techniques were refined. Horch, E., Meyer, R., & van Altena, W., 2004. We conclude that quasi-speckle interferometry, “Speckle observations of Binary Stars with the which employs a single reference star for deconvolution WIYN telescope—IV. Differential Photometry,” and bispectrum analysis of all observations for phase The Astronomical Journal, 127:1727–1735. reconstruction, appears to provide better results than Marchetti, C., Caputo, R., & Genet, R., 2020. “The either simple stacking or lucky imaging. This may be Fairborn Institute Robotic Observatory: First Obser due to: (1) the retention of information on the relative vations,” submitted to the Journal of Double Star position of the two stars in these long-exposure images Observations. which can be correlated by speckle processing better Rowe, D., 2018, “GDS1.0 Gaia (DR2) Double Stars than stacking or lucky imaging; (2) single-star deconvo- Search Program.” Private communication. lution which can, to some extent, cancel optical aberra- Rowe, D., & Genet, R., 2015. “User’s Guide to PS3 tions and systematic atmospheric phenomena such as Speckle Interferometry Reduction Program, Journal dispersion, and (3) processing of the observations in the of Double Star observations, 11, 266. Fourier frequency domain which allows suppression of Wasson, R., Rowe, D., & Genet, R., 2020. noise at spatial frequencies that cannot be real, e.g., “Observation of Gaia (DR2) Red and White Dwarf smaller than the Airy disk or larger than the frame size. Binary Stars in the Solar Neighborhood,” Journal of Red dwarf secondary stars as late as M7V were Double Star Observations, 16-3, 208. observed, but only a few systems showed motion great- Winters, J., Henry, T., Jao, W., Subasavage, J., Chate er than the uncertainties since the observations of Gaia lain, J., Slatten, K., Riedel, A., Silverstein, M., & DR2. The fastest moving system, which also happened Payne, M., 2019. “The Solar Neighborhood. XLV: to have a relatively long observational history (74 The Stellar Multiplicity Rate of M Dwarfs Within years), may have an orbital period on the order of 800 25 pc,” The Astronomical Journal, 157, 216. years. Therefore, all these systems likely have long pe-

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 483

A New Double Star Detected During an Occultation by the Asteroid (409) Aspasia

Michel Boutet, and Jacques Sanchez

Observatoire Les Pléiades, 31130 Latrape, France [email protected]

Abstract: On November 9, 2018 an occultation of the star UCAC4 435-115475 by the asteroid (405) Aspasia produced two successive magnitude drops, showing this star is a new double.

Occultation Circumstance bility of 96% (figure 1). On 2018 November 9, an occultation of the star According to the astorb database of October 1, UCAC4 435-115475 by the Asteroid (405) Aspasia 2018, the shadow center of the ≈190 km diameter as- was predicted at our Latrape observatory with a proba- teroid was passing only 22 km north from our site. We

Figure 1: Occult 4 prediction for the occultation path

Received October 28, 2020

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 484

A New Double Star Detected During an Occultation by the Asteroid (409) Aspasia

were two observers at the same site, Michel Boutet and Jacques Sanchez with 12" and 14" SCT. We both used analog CCD video cameras (Watec 910HX/RC) at prime focus. Unfortunately, the 14” setup failed at the last minute, but the recording with the 12” setup, by Michel Boutet, was successful. UCAC4 435-115475 parameters The star coordinates (J2000) are (figure 2): Alpha: 21h 15mn 05.8738s (astrometric) Delta: -03° 06' 18.927" (astrometric) The star of Vmag 12.4 is located in the constellation Aquarius (Aqr) and was at 31° elevation that day.

4UC435-115475 Figure 3: TANGRA(4) reduction of the video recording of the two Overview of the UCAC-4 catalog occultations (blue curve). The reference stars (green and pink) RA: 21h15m05.8846s +/- 54 milliarcseconds are stable. declination: - 3 06' 18.857" +/- 34 milliarcseconds (Above RA/dec is the J2000 value straight from the catalog, with no propermotion or other corrections applied)

Proper motion in RA: 13.0 +/- 4.6 milliarcsec/year Proper motion in dec: 6.9 +/- 3.5 milliarcsec/year UCAC instrumental "pseudo-R" fit-model magnitude: 12.430 UCAC instrumental "pseudo-R" aperture magnitude: 12.414 Error on above magnitudes is 7 millimags Good star Single star 6 UCAC images were taken; 4 were actually used (3) 3 catalogues were used for proper motions Figure 4: AOTA analysis of occultation 1 with Occult 4 . Not matched to Hipparcos or Tycho AC2000 match flag: 1 NPM Lick match flag: 1 Running number within UCAC4 catalog: 98440871 UCAC2: 174-152230

Figure 2: UCAC4 star parameters Double Star Occultation The star is considered as single in the UCAC4 catalog and is not listed in the Washington Double Star catalog. During the Aspasia occultation, two successive Figure 5: AOTA analysis of occultation 2 with Occult 4(3) . magnitude drops were observed while only one was expected: the first one lasted 7.60s at the expected time, Occ1 timings: at UT 20h 38mn 10s. A calculation with the Occult D1 - 20:38:09.82 ± 0.24 disappearance software and the astorb catalog of October 1 predicted R1 - 20:38:17.42 ± 0.16 reappearance an occultation at 20h 38mn 15s. A second unexpected occultation occurred 28.56s later, at UT 20h 38mn 46s Occ2 timings: and lasted 10.56s with a smaller mag drop. D2 - 20:38:45.98 ± 0.72 disappearance The expected maximum occultation duration was R2 - 20:38:56.54 ± 1.04 reappearance 11.0s confirming we were close to the central line, es- Double Star Magnitudes Evaluation pecially for the second star. Analysis of the fluxes by Eric Frappa(1) gave a The recording and the results analysis show clearly max flux (asteroid + stars) of 2961.6. The flux for Occ1 a double star (figures 3, 4, 5). (the asteroid plus the secondary star minus the main star) was 1871.2 and the flux for Occ2 (the asteroid

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 485

A New Double Star Detected During an Occultation by the Asteroid (409) Aspasia

main star minus the secondary star) was 2457.9. Using Further analysis with the real shape the standard magnitude equation: Aspasia shape has already been modeled after pho- Δm = -2.5 log(Flux1/Flux2), tometric observations, three DAMIT and ISAM models the Occ1 magnitude drop was estimated to 0.5 and are available (see an example on figure 7). All models the Occ2 magnitude drop to 0.2. show some small deviations from a circular shape. If we combine these flux drops into one, simulating a single star, we find a combined mag drop of 0.8 which is what was expected (figure 1). Using the known asteroid magnitude of 12.5 and the star magnitude of 12.4, with the combined magnitudes equation, m = m1 - 2.5 log (1 + antilog (-0.4 (m2 - m1))) we can derive the magnitude of each component main star mag = 12.8 secondary star mag = 13.7 Double star separation (Sep) and position angle (P.A.) evaluation From the measured chords and the asteroid parameters, the double star characteristics can be derived. As there was only one observation, a unique solution cannot be found: the measured path can occur in one side or in the other side of the asteroid, giving the same measured timings. Four solutions exist. They have been Figure 7: Example of DAMIT #215 shape model estimated 'manually' by geometry but confirmed with solution #1, showing Aspasia at the occultation time more accuracy with the Occult4 software of Dave Her- superimposed on the measured chords. ald(3) (Table 1 & figure 6), using the asteroid observations editor and its very good documentation. We as- Evaluation of the best model fit: sume a circular shape for the asteroid(2) and a diameter For each model and each solution, we superim- of 179.5 km. posed the asteroid modeled shape on the measured chords and measured graphically the deviations between them (Table 1).

Solution #1 #2 #3 #4 Total /Model Model DAMIT 215 6 23 16 5 50 DAMIT 715 9 19 41 16 95 ISAM 2 8 14 7 7 36 Total /Solution 23 56 64 28 Table 1: Cumulated distances between chords and model Solution #1 Solution #2 shape (mm). Graphical / manual estimation on screen: the lower figure the better fit.

The best fits are for DAMIT #215 solutions 1 and 4, DAMIT #715 solutions 1 and 4 and ISAM #2 solutions 1, 3 and 4. These rough estimations give a better probability for solutions 1 and 4 on which DAMIT 215 and ISAM 2 models give the best fit. Solution #3 Solution #4 Summary Figure 6: The four solutions for Sep and P.A. of the star components with a circular model of the asteroid. UCAC4 435-115475 is a double star found through an occultation by an asteroid. Four solutions for the All solutions are close to each other. characteristics are shown in Table 2, with more probability for solutions 1 and 4.

Vol. 16 No. 5 November 1, 2020 Journal of Double Star Observations Page 486

A New Double Star Detected During an Occultation by the Asteroid (409) Aspasia

NAME RA+DEC MAGS PA SEP DATE N NOTES

95.6 ± 0.3687 ± UCAC4 435-115475 211506-0306 12.8,13.7 2018.858 1 Solution 1 5.6 0.0061

98.0 ± 0.3703 ± UCAC4 435-115475 211506-0306 12.8,13.7 2018.858 1 Solution 2 5.4 0.0068 86.6 ± 0.3687 ± UCAC4 435-115475 211506-0306 12.8,13.7 2018.858 1 Solution 3 5.5 0.0060

84.2 ± 0.3703 ± UCAC4 435-115475 211506-0306 12.8,13.7 2018.858 1 Solution 4 5.5 0.0069

Table 2: UCAC4 435-115475 solutions with circular model

Acknowledgments / References The author would like to acknowledge and thank Eric Frappa for his assistance in the analysis of these events.

(1)- Analysis and calculations of the stars magnitudes performed by Eric Frappa. http://www.euraster.net/

(2)- An occultation observation of (409) Aspasia show ing its roughly circular shape: http://www.euraster.net/results/2008/ in dex.html#0212-409

(3)- Dave Herald Occult software for lunar and aster oids occultations prediction and analysis including a module for double stars analysis through occulta tions. http://www.lunar-occultations.com/iota/ occult4.htm

(4)- Tangra3 is a software designed by Hristo Pavlov for reducing astronomical video observations such as asteroid occultations. http:// www.hristopavlov.net/Tangra3/