Lunar Occultations of Eighteen Stellar Sources from the 2.4-m Thai National Telescope
A. Richichi, O. Tasuya, P. Irawati, B. Soonthornthum, V. S. Dhillon, T. R. Marsh
aa r X i v : . [ a s t r o - ph . S R ] D ec Lunar Occultations of Eighteen Stellar Sourcesfrom the 2.4-m Thai National Telescope
A. Richichi , O. Tasuya , P. Irawati , B. Soonthornthum , V.S. Dhillon , , T.R. Marsh [email protected] ABSTRACT
We report further results from the program of lunar occultation (LO) observations started at the 2.4-mThai National Telescope (TNT) in 2014. We have recorded LO events of 18 stellar sources, leading tothe detection of four angular diameters and two binary stars. With two exceptions, these are first-timedeterminations. We could resolve angular diameters as small as 2 milliarcseconds (mas) and projectedseparations as small as 4 mas. We discuss the individual results, in the context of previous observationswhen available. The first-time angular diameters for o Psc, HR 6196 and 75 Leo are in good agreementwith expected values, while that of π Leo agrees with the average of previous determinations but has ahigher accuracy. We find a new secondary in o Psc, as previously suspected from Hipparcos data. Wealso obtain an accurate measurement of the companion in 31 Ari, revealing inconsistencies in the currentlyavailable orbital parameters. The TNT, equipped with the fast ULTRASPEC imager, is the leading facilityin Southeast Asia for high time resolution observations. The LO technique at this telescope achieves asensitivity of i ′ ≈
10 mag, with a potential to detect several hundreds of LO events per year.
Subject headings: techniques: high angular resolution – occultations – binaries: general – stars: fundamental param-eters
1. Introduction
In a previous paper (Richichi et al. 2014b, R14hereafter) we have described the novel combinationof the ULTRASPEC instrument installed at the 2.4-mThai National Telescope (TNT), and operated in driftmode to achieve fast imaging at the level of few mil-liseconds. This capability, entirely new for the givenlongitude range and at a telescope which is the largestin the southeast-asian region, has enabled a programof routine lunar occultation (LO) observations. Theaim is to measure angular diameters, to detect and National Astronomical Research Institute of Thailand, 191 Si-riphanich Bldg., Huay Kaew Rd., Suthep, Muang, Chiang Mai50200, Thailand Department of Physics and Materials Science, Faculty of Sci-ence, Chiang Mai University, Chiang Mai 50200, Thailand Department of Physics and Astronomy, University of Sheffield,Sheffield S3 7RH, UK Instituto de Astrofisica de Canarias, 38205 La Laguna, SantaCruz de Tenerife, Spain Department of Physics, University of Warwick, Gibbet HillRoad, Coventry CV4 7AL, UK characterize circumstellar components, and to inves-tigate binary or multiple stars. The achieved angularresolution is at the milliarcsecond level (mas) and thesensitivity i ′ ≈
10 mag.This is currently the only routine program of itskind around the world. Another large program wasactive at the ESO Very Large Telescope in the pre-vious decade and produced a large number of results(Richichi et al. 2014a, and references therein), but wasterminated with the decommissioning of the ISAACinstrument. Only few observatories around the worldremain suitably equipped to observe LO, and none ofthem has a comparable program in place. Clearly, thehigh time resolution of this facility can be employedto observe other classes of objects as well, from cata-clysmic variables to transits, and many kinds of tran-sient phenomena.We report here on LO results recorded in the sec-ond TNT observing cycle, between December 2014and May 2015.1 . Observations and Data Analysis
The observations and the data analysis followclosely what was already described in R14, and weprovide here only a brief summary. An in-depth gen-eral description of ULTRASPEC at TNT has beenpresented by Dhillon et al. (2014). For LO, we usethe so-called drift mode of the instrument, in whichmost of the detector is masked and light is recordedin two small square sub-windows. Only one of thetwo is effectively used. This allows us to record unin-terrupted continuous sequences, with typical samplingtimes around 6 ms and minimal overheads.For the observations reported here, we have usedstandard SDSS r ′ , i ′ , z ′ broad-band filters, as well astwo narrow-band filters, R cont and N86. They havecentral wavelengths of 6010 and 8611 Å and full-widthhalf-maxima (FWHM) of 118 and 122 Å, respectively.Narrow-band filters have the advantage of increasingthe contrast of the diffraction fringes, as well of courseof reducing the risk of saturation in case of strong lunarbackground.We typically record about 30 s data around the pre-dicted time of the event, to account for possible un-certainties, e.g. due to proper motions and limb devi-ations. The sub-windows can be further rebinned toimprove the time resolution – see the parameters Suband Reb in Table 1.The result is a sequence of several thousand frames,which are converted to a FITS cube and analyzed withspecifically developed software (see Richichi et al.2014a, and references therein). A crucial step is toadopt an extraction mask tailored to measure onlythe pixels effectively exposed to the stellar light,thus greatly reducing the noise contribution from thegeneral background. Each frame is accurately time-stamped with sub-millisecond precision thanks to adedicated GPS signal. Time-stamping is also relevantfor other applications such as lunar limb profiling,which however we do not discuss here.Note that we concern ourselves only with about 1 sof data around the occultation event: the LO diffrac-tion pattern extends in fact over 0.5 s or less. This cor-responds in practice to about 2 ′′ on the sky. Thus weare not sensitive to, e.g., wide companions.We use two approaches to analyze the data: amodel-independent maximum-likelihood method toderive the brightness profile of the source (Richichi1989), and a model-dependent least-squares methodto derive parameters such as angular diameters and bi- nary separations (Richichi et al. 1996). The softwareincludes a number of features, e.g. to derive upper lim-its for unresolved sources or to model low-frequencyscintillation. Details are provided in the referencesabove.
3. Results
The LO events are listed in chronological order inTable 1, which follows the same format as R14. Insummary: D and R refer to disappearances and reap-pearances, respectively; the magnitudes and spectraare quoted from Simbad; the filters were describedin Sect. 2; Sub and Bin list, respectively, the size ofthe detector sub-array and the on-chip rebinning - i.e.,Sub 16x16 and Bin 2x2 lead to frames of size 8x8; τ and ∆ T are the integration and sampling times, re-spectively; S/N is the signal-to-noise ratio, measuredas the unocculted stellar signal divided by the rms ofthe fit residuals; and finally UR, Diam, and Bin denoteunresolved, resolved diameter, and binary star, respec-tively.The stars with a positive determination (i.e., not un-resolved) are listed in Table 2, which also follows aformat used in our previous papers. Columns 2 and3 list the measured rate of the event V and its devia-tion from the predicted rate V t . This deviation is duemainly to slopes in the local lunar limb ψ , which canthus be retrieved and are listed in Column 4. All thedeviations and limb slopes are within the norm, basedon the experience of several thousands LO events.Columns 5 and 6 list the Position Angle and the Con-tact Angle of the event, with the limb slope already in-cluded. For the sources found to be resolved, column7 lists the best-fitting angular diameter in the uniformdisk approximation. This is strictly valid for the ob-served wavelength only; conversions to limb-darkenedvalues can be derived depending on models. For thebinary stars, columns 8 and 9 list the projected separa-tion (along the PA direction) and the brightness ratio,respectively.In the following we discuss in some detail our re-sults, also in the context of available previous studies.We show figures only for one case of a resolved angu-lar diameter (Fig. 1) and for one case of a binary star(Fig. 2). o Piscium
This star (HR 510, HIP 8198) is a relatively nearbylate-type giant star (G8, 85 pc) which has been the2
ABLE IST OF OBSERVED EVENTS
Date Time Type Source V Sp Filter Sub Bin τ ∆ T S/N Notes(UT) (mag) (pixels) (ms)09-Dec-14 22:10 R SAO 97302 8.0 G5V N86 16x16 2x2 6.6 6.9 5.0 UR09-Dec-14 22:13 R SAO 97303 8.7 K0V N86 16x16 2x2 6.6 6.9 2.6 UR02-Jan-15 14:53 D IRC +20090 9.2 K0 N86 16x16 2x2 6.6 6.9 11.9 UR26-Jan-15 14:14 D o Psc 4.3 G8III N86 16x16 2x2 6.1 6.3 60.9 Bin, Diam27-Jan-15 11:49 D 31 Ari 5.7 F7V R cont i ′ z ′ z ′ z ′ π Leo 4.7 M2IIIab N86 8x8 no 6.1 6.3 146.6 Diam24-Apr-15 14:33 D SAO 96634 9.0 A2V r ′ z ′ - z ′ T ABLE UMMARY OF RESULTS : ANGULAR SIZES ( TOP ) AND BINARIES ( BOTTOM ) (1) (2) (3) (4) (5) (6) (7) (8) (9)Source V (m/ms) (V/V t )–1 ψ ( ◦ ) PA( ◦ ) CA( ◦ ) φ UD (mas) Proj.Sep.(mas) Br. Ratioo Psc 0.2138 - ± ± π Leo 0.6112 - ± - ± - ± ± - ± ± >
20 mas. Hartkopf & McAlister (1984)obtained speckle observations at a 4-m telescope, find-ing it unresolved with an upper limit of 30 mas. In-terestingly, no LO event of this star was ever reporteduntil now.We obtained a good quality (S/N=61), first everLO light curve of o Psc. Thanks to a relatively largecontact angle, the efffective limb velocity was about3 times slower than normal, leading to an increasedsampling of the fringes which was crucial in assessingdepartures from a point source. The data are consis-tent with a resolved angular diameter and in additionwith a secondary component 4.0 mag dimmer than theprimary, with a separation of 12 mas projected alongPA=144 ◦ . Our first-time determination of the angulardiameter, when combined with the Hipparcos distance,results in 19.8 R ⊙ .The companion, given the magnitude difference, islikely to be a main sequence star, although more detailscould be ascertained only with further observations atdifferent wavelengths. Some constraint on the actualPA can be inferred from the upper limit on the separa-tion by Hartkopf & McAlister (1984). Assuming thattheir speckle measurements were sensitive to a com-panion with such a brightness ratio, the PA would haveto be in the –admittedly broad– range 78 ◦ to 210 ◦ . As-suming a combined 1 M ⊙ , a real separation between1 and 2.6 AU (converted from our measurement andfrom the speckle upper limit), a circular orbit and noinclination effects, the minimum period would be be-tween 1 and 4 years. Barring significant inclinationsof the orbital plane on the sky, orbital motion mightbe detectable within relatively short periods of time,using i.e. adaptive optics at a very large telescope atvisual wavelengths. This star is a late-type giant (G8II/III) and a brightnear-infrared source (IRC - K = 2 . χ almost 2.5 times lower than for a point source. Wealso revisited the old data set mentioned in CHARM2,which was obtained with an InSb photometer in the K band at the Calar Alto 1.2-m telescope on July 16,1997. Unfortunately, those data are affected by signif-icant scintillation and only lead to an upper limit be-tween 2.7 to 3.4 mas, consistent with our result but notvery constraining. π Leonis
This cool giant (HR 3950, IRC +10224) hashad a number of spectral classifications, with ageneral agreement around M2. It has featured inmany publications especially in the near-IR, whereit has often been used as a calibrator source due toits brightness ( K = 0 . π Leo has been repeat-edly determined in the past, with results listed in theuniform φ UD , limb-darkened φ LD or fully-darkened φ FD disk hypothesis. Using the NPOI interferome-ter at various wavelengths with an average of 740 nm,Nordgren et al. (1999) reported φ UD = 4 . ± .
05 mas,in very good agreement with our own determina-tion. A large number of LO observations also ex-ist, with varying levels of data quality and accuracy.Vilas & Lasker (1977) reported φ LD = 5 . R -band (no error), while Africano et al. (1978) re-ported φ FD = 3 . ± . φ UD =4 . ± .
33 mas, while Baug & Chandrasekhar (2013)obtained φ UD = 4 . ± . K -4and filters. Schmidtke & Africano (2011) reported φ UD = 3 . ± .
15 mas, using a 1991 LO light curveobtained in H α with a 1.3-m telescope. The weightedaverage of these LO results, neglecting the differencein wavelengths and limb-darkening, gives at first ap-proximation φ = 4 . ± . π Leo with a quality(S/N=147) better than any previous others. As shownin Fig. 1, the source is clearly resolved, and our resultlisted in Table 2 is consistent with all good-quality pre-vious determinations, but with significantly improvedaccuracy.
This is a bright M0 giant (HR 4371), especially lu-minous in the near-IR (IRC +00203, K = 1 . φ UD = 2 . ± .
03 mas.Our light curve for 75 Leo has a formally veryhigh S/N of 266. In fact, the data were taken with arather slow sampling (and correspondingly long inte-gration time, hence the high S/N value), due to passingclouds that made the star temporarily disappear dur-ing acquisition and forced us to observe in blind modewith an unusually large sub-window. As a result, thediffraction fringes are smoothed and poorly sampled.Nevertheless, these factors can be properly includedin the data analysis and the χ clearly shows that thestar is indeed resolved. Our resulting angular diame-ter value is φ UD = 2 . ± .
03 mas, in excellent agree-ment with the estimate of Cohen et al. (1999). Thisrepresents the only direct determination for 75 Leo.Barring a very unfavorable combination of projectiondirections, the bright companion possibly claimed by(Guhl & Stecklum 1991, who did not mention a PA for their event) should have been seen in our data.
This bright star (HR 763, SAO 93022) has a spec-tral type F7V. Being main sequence, it is as expectedrelatively nearby: its Hipparcos parallax places it at35 pc. 31 Ari was first discovered to be binary froma LO by Africano et al. (1978), with projected sepa-ration 21 mas along PA=265 .
7. The magnitude differ-ence was found to be small both in the blue and in thered, prompting the authors to speculate similar spectraltypes. For comparison with our measurement (shownin Fig. 2), they quoted ∆ mag = 0 . ± . IRC+20090 was found to have φ UD = 3 . ± .
100 125 150 175 I n t en s i t y [ c oun t s / ] -5 0 5 2700 2800 2900 3000 x Relative Time [ms]
Fig. 1.— Top panel: occultation data (dots) for π Leo,and best fit by a model with an angular diameter of4.35 mas (solid line). The lower panel shows the fitresiduals, enlarged by a factor of two for clarity, as asolid line. For comparison, also the residuals of thebest fit by a point-like source are shown (dashed line).A color version of this figure is available online.
A BRelative Time [ms]
Fig. 2.— The occultation data (dots) for 31 Ari, re-peated twice and offset for clarity. The top set showsthe best fit by a point-source model, and the bottomset the best fit by a binary model with 3.76 mas sepa-ration. The two segments mark the times of the geo-metrical occultation of each component. The verticalaxis is in arbitrary units. A color version of this figureis available online. PA= 71 ◦ . Following that however, the source was notdetected again by speckle in several attempts includingwith the russian 6-m telescope (McAlister et al. 1993;Balega et al. 1994; Fu et al. 1997). The star is at justunder 40 pc distance and one could speculate a highlyeccentric orbit which over the course of a few yearsbrought the component below the diffraction limits ofthe telescopes employed. We note that our LO lightcurve had sufficient S/N and sampling rate to detectthe companion if the projected separation had been & . ◦
8, unchanged over nineyears. The system was rediscovered by Africano et al.(1978, apparently unaware of Couteau’s work) whorecorded two LO events, each in two colors. Theprojected separations (0 . ′′ . ′′
2) however did notagree with Couteau’s results, and also not betweenthemselves. Finally, Mason (1996) could not resolvethe system by speckle observations at a 3.6-m tele-scope. Our data are consistent with a companion with60 ms projected separation along PA= 85 ◦ , in approx-imate agreement with Couteau’s values. However, wedo not list this among our binary detections becausethere is no significant χ improvement over a point-source model. The companion should be detectable byadaptive optics imaging.
4. Conclusions
We reported the latest results from the program ofroutine LO observations at the 2.4-m Thai NationalTelescope. They include 4 angular diameters, threeof which are first-time determinations, and two binarystars, of which one was not previously known.The angular diameters of o Psc, HR 6196, π Leoand 75 leo are found to be in the range 2 to 4 mas,with an average accuracy of 0.5%. Thus, if combinedwith accurate bolometric fluxes, our results have thepotential to constrain the effective temperatures to the1% level, or less than 50 K. Concerning the binarystars, we detected companions with projected separa-tions of about 4 and 12 mas in 31 Ari and o Psc, re-spectively. The magnitude differences (in one case ashigh as 4 mag) could be measured to the 1% level.We also discussed those stars for which an angu-lar diameter or a companion had been previously pub-lished, but which we found unresolved. The case ofSAO 98400 merits attention, as this binary was de-6ected repeatedly in the 1980’s but not in the followingdecades.As stated in R14, the TNT equipped with the UL-TRASPEC EMCCD fast imager can record LO lightcurves with few ms sampling times to about i ′ ≈
10 mag at S/N=1. For illustration, we computed fu-ture events taking the SAO catalog as an approximateproxy for the sample of stars potentially reachable byLO at the TNT. After filtering for good circumstances(Sun and Moon elevations, lunar phase), we find thatover 500 LO events would be well observable duringeach dry season period.VSD and TRM acknowledge the support of theRoyal Society and the Leverhulme Trust for the op-eration of ULTRASPEC at the TNT. This researchmade use of the Simbad database, operated at the CDS,Strasbourg, France. The data of 28-31 March 2015were collected during the NIATW (NARIT Interna-tional Astronomical Training Workshop), with the in-volvement of the participants.
REFERENCES
Africano, J. L., Evans, D. S., Fekel, F. C., Smith,B. W., Morgan, C. A. 1978, AJ, 83, 1100Balega, I. I., & Balega, Y. Y. 1988, Pisma v Astro-nomicheskii Zhurnal, 14, 927Balega, I. I., Balega, Y. Y., Belkin, I. N., et al. 1994,A&AS, 105, 503Baug, T., & Chandrasekhar, T. 2013, Research in As-tronomy and Astrophysics, 13, 1363Cohen, M., Walker, R. G., Carter, B., et al. 1999, AJ,117, 1864Couteau, P. 1966, Journal des Observateurs, 49, 341Couteau, P. 1975, A&AS, 20, 391Dhillon, V. S., Marsh, T. R., Atkinson, D. C., et al.2014, MNRAS, 444, 4009Eggleton, P. P., & Tokovinin, A. A. 2008, MNRAS,389, 869Eitter, J. J., & Beavers, W. I. 1979, ApJS, 40, 475Evans, D. S., & Edwards, D. A. 1981, AJ, 86, 1277Frankowski, A., Jancart, S., Jorissen, A. 2007, A&A,464, 377 Fu, H.-H., Hartkopf, W. I., Mason, B. D., et al. 1997,AJ, 114, 1623Guhl, K., & Stecklum, B. 1991, AstronomischeNachrichten, 312, 315Hartkopf, W. I., & McAlister, H. A. 1984, PASP, 96,105Hartkopf, W. I., Mason, B. D., McAlister, H. A., et al.2000, AJ, 119, 3084Hartkopf, W. I., Tokovinin, A., Mason, B. D. 2012, AJ,143, 42Horch, E. P., Bahi, L. A. P., Gaulin, J. R., et al. 2012,AJ, 143, 10Makarov, V. V., & Kaplan, G. H. 2005, AJ, 129, 2420Mason, B. D. 1996, AJ, 112, 2260Mason, B. D. 1997, AJ, 114, 808McAlister, H. A., & Hendry, E. M. 1982, ApJS, 49,267McAlister, H. A., Hartkopf, W. I., Sowell, J. R., Dom-browski, E. G., & Franz, O. G. 1989, AJ, 97, 510McAlister, H. A., Mason, B. D., Hartkopf, W. I., &Shara, M. M. 1993, AJ, 106, 1639Meyer, C., Rabbia, Y., Froeschle, M., Helmer, G., &Amieux, G. 1995, A&AS, 110, 107Monnier, J. D., Millan-Gabet, R., Tuthill, P. G., et al.2004, ApJ, 605, 436Montargès, M., Kervella, P., Perrin, G., et al. 2014,A&A, 572, A17Muller, P. 1958, Journal des Observateurs, 41, 109Nordgren, T. E., Germain, M. E., Benson, J. A., et al.1999, AJ, 118, 3032Pasinetti Fracassini, L. E., Pastori, L., Covino, S., &Pozzi, A. 2001, A&A, 367, 521Richichi, A. 1989, A&A, 226, 366Richichi, A., Baffa, C., Calamai, G., Lisi, F. 1996, AJ,112, 2786Richichi, A., Percheron, I., & Khristoforova, M. 2005,A&A, 431, 7737ichichi, A., Fors, O., Cusano, F., Ivanov, V. D. 2014a,AJ, 147, 57Richichi, A., Irawati, P., Soonthornthum, B., Dhillon,V. S., & Marsh, T. R. 2014b, AJ, 148, 100, R14.Ridgway, S. T., Wells, D. C., Joyce, R. R., & Allen,R. G. 1979, AJ, 84, 247Schmidtke, P. C., & Africano, J. L. 2011, AJ, 141, 10Tej, A., & Chandrasekhar, T. 2000, MNRAS, 317, 687Vilas, F., & Lasker, B. M. 1977, PASP, 89, 95Woodruff, H. C., Tuthill, P. G., Monnier, J. D., et al.2008, ApJ, 673, 418