Multiplicity of Nearby Free-floating Ultra-cool Dwarfs a HST

a r X i v :a s t r o -p h /0305484v 2 10 J u n 2003accepted for publication in AJ (September 2003),First sub-

mission:August 2002

Multiplicity of Nearby Free-?oating Ultra-cool Dwarfs:

a HST-WFPC2search for companions

Herv′e Bouy E.S.O,Karl Schwarzschildstra?e 2,D-85748Garching,Germany hbouy@6f4ce10c76c66137ee0619c7 Wolfgang Brandner Max-Planck Institut f¨u r Astronomie,K¨o nigstuhl 17,D-69117Heidelberg,Germany brandner@mpia.de Eduardo L.Mart′?n Institute for Astronomy,University of Hawai‘i,2680Woodlawn Drive,Honolulu,HI 96822,USA ege@6f4ce10c76c66137ee0619c7 Xavier Delfosse Laboratoire d’Astrophysique de l’Observatoire de Grenoble,414rue de la piscine,F-38400Saint Martin d’H`e re,France Xavier.Delfosse@obs.ujf-grenoble.fr

France Allard

Centre de Recherche Astronomique de Lyon (UML 5574),Ecole Normale Sup′e rieure,

69364Lyon Cedex 07,France

fallard@ens-lyon.fr

Gibor Basri

University of California at Berkeley,Astronomy Department,MC 3411Berkeley,CA

94720,USA

basri@6f4ce10c76c66137ee0619c7

–2–

ABSTRACT

We present HST/WFPC2observations of a sample of134ultra-cool objects (spectral types later than M7)coming from the DENIS,2MASS and SDSS sur-

veys,with distances estimated to range from7pc to105pc.Fifteen new ultra-

cool binary candidates are reported here.Eleven known binaries are con?rmed

and orbital motion is detected in some of them.We estimate that the closest

binary systems in this sample have periods between5and20years,and thus dy-

namical masses will be derived in the near future.For the calculation of binary

frequency we restrict ourselves to systems with distances less than20pc.After

correction of the binaries bias,we?nd a ratio of visual binaries(at the HST

limit of detection)of around10%,and that～15%of the26objects within20

parsecs are binary systems with separations between1and8A.U.The observed

frequency of ultra-cool binaries is similar than that of binaries with G-type pri-

maries in the separation range from2.1A.U.to140A.U.There is also a clear

de?cit of ultra-cool binaries with separations greater than15A.U.,and a possi-

ble tendency for the binaries to have mass ratios near unity.Most systems have

indeed visual and near-infrared brightness ratios between1and0.3.We discuss

our results in the framework of current scenarios for the formation and evolution

of free-?oating brown dwarfs.

Subject headings:stars:very low mass,ultra-cool dwarfs,brown dwarfs,binary

1.Introduction

E?ective strategies to detect brown dwarfs are proper motion surveys(e.g Ruiz et al. (1997)),wide?eld CCD surveys in star forming regions,open clusters like the Pleiades(e.g Stau?er et al.(1994);Zapatero Osorio et al.(1997);Bouvier et al.(1998)),or the new generation of optical and near-infrared all sky surveys like the Sloan Digital Sky Survey (York et al2000),the DEep Near Infrared Survey(Epchtein et al.1997),and the2Micron All Sky Survey(Skrutskie et al.1997).So far,DENIS has produced a list of～300nearby very low-mass objects(Delfosse et al.1997a,1999,2003;Mart′?n et al.1999a),and a similar number has been detected by2MASS(Kirkpatrick et al.1997,1999,2001;Burgasser et al.1999;Gizis et al.2000;Kirkpatrick et al.2003).To date,several hundreds of nearby free-?oating ultra-cool and brown dwarfs have been identi?ed,and many more candidates are waiting for con?rmation.This nearby sample is ideal for resolving ultra-cool and brown dwarfs binaries,and is large enough for statistical studies.

–3–

While systematic surveys of their physical properties are just starting,we already have indications that binary brown dwarfs are not rare.PPL15,the?rst brown dwarf in the Pleiades with con?rmed lithium(Basri et al.1996),turned out to be a spectroscopic binary (Basri&Mart′?n1999).Several direct imaging surveys in the?eld or in open clusters using the high spatial resolution provided by both HST and/or adaptive optics led to the discovery of a signi?cant number of binaries(Mart′?n et al.1997,1999a,2000a;Koerner et al.1999; Reid et al.2001;Neuh¨a user et al.2002;Close et al.2002a,b;Potter et al.2002;Burgasser et al.2003;Gizis et al.2003;Close et al.2003).

There are several key questions which need to be answered and for which brown dwarf binaries are an important piece of the puzzle:where do free-?oating brown dwarfs originate? Are they ejected stellar embryos(Reipurth&Clarke2001;Bate et al.2002;Delgado-Donate et al.2003),or do they form more isolated like ordinary stars due to fragmentation of collapsing molecular cloud clumps(Bodenheimer et al.1999)?

One can then ask,how the binary properties of brown dwarfs,such as the frequency, distribution of separations,distribution of mass-ratios(is there a lower mass limit for com-panions to brown dwarfs?),or the relation between orbital period and eccentricity,compare to the properties of stellar binaries.These properties hold important clues on the origin of (binary)brown dwarfs.An agreement of brown dwarf and stellar binary properties would suggest the same formation mechanism for both types of objects.

Resolved brown dwarf binaries provide a very valuable opportunity to measure sub-stellar dynamical masses.The?rst such measurement has been reported by Lane et al. (2001)in the brown dwarf binary Gliese569B,discovered by Mart′?n et al.(2000a).Dynam-ical masses for brown dwarfs of di?erent ages are very much needed,as the mass estimates currently depend on untested theoretical evolutionary tracks,model atmospheres,and as-sumptions on the internal structure of brown dwarfs.In particular,a brown dwarf with a given age and mass might have very similar colours and luminosity as a younger brown dwarf with a lower mass.This degeneracy in the mass-luminosity relation for sub-stellar objects makes it very hard to pin down the physical properties of brown dwarfs,and to achieve the interplay between observation and theory which is necessary in order to improve,adapt,and ?ne-tune models,and to guide the interpretation of the observations.A search for binary brown dwarfs,and a detailed study of their properties directly addresses these questions.

The Pleiades,the?rst cluster in which a signi?cant population of(coeval)brown dwarfs was identi?ed,is at a distance of125pc.This makes it hard to detect and resolve binary brown dwarfs which separations less than20A.U.Orbital periods of resolved brown dwarf binaries in the Pleiades will be excessively long(≥several100yrs),making it impractical to compute their orbits and to derive dynamical masses.A recent search for binaries among

–4–

34very low mass members of the Pleiades by Mart′?n et al.(2000a)with HST-WFPC2and Adaptive Optics system of CFHT did not lead to the discovery of any binary,indicating a lack of wide brown dwarfs binaries.

Nearby free-?oating brown dwarfs do not form a coeval sample,but they are close enough to measure their distance(angular parallax)precisely,and-by using HST/WFPC2 -it is possible to detect and resolve binary brown dwarfs with separations down to0.4A.U.

(0.′′06at7pc).

The objects we study here have distances between7and105pc.For a prototypical brown dwarf binary with a separation of0.′′060,component masses of0.045M⊙and0.030 M⊙,and a circular orbit,the orbital period would be respectively between～1and～100 years.Within the next5to10years dynamical masses for a wide range of brown dwarf binary companions could be determined.

In section2,we describe the observations and the data analysis.In section3we present and discuss the results we obtained on each binary,in section4we perform an analysis of the results,and in section5we discuss these results in the framework of current models of formation and evolution of free-?oating brown dwarfs.

2.Observations and data analysis

2.1.Sample selection

Our initial sample consists primarily of34objects detected by the near-infrared sky surveys DENIS,2MASS and SDSS.They have been selected by analysing their positions in a colour-magnitude diagram,looking for the reddest objects.

In order to increase the sample of objects and the quality of the statistical study,the total sample presented in this paper put together data from our own program(GO8720, P.I Brandner)with data from two other HST programs.We thus analyse data from the public HST archive coming from programs GO8146,P.I Reid,including21objects(Reid et al.2001),and program GO8581,P.I Reid(Gizis et al.2003),including84objects.The total sample thus had134objects.

The complete list of targets of the program is shown in Table1.Eleven of these objects were already previously identi?ed as binary brown dwarfs.

–5–

2.2.Observations

The observations occurred between February2000and August2001during HST Cycles 8and9.They have been carried out in snapshot mode.Each object was observed with the Planetary Camera of HST/WFPC2,in the F675W(600s)and F814W(300s)for our own program,in the F606W(100s)and F814W(300+350s)?lters for program GO8146(Cycle8, P.I Reid,see Reid et al.(2001))and in the F814W(100s,200s or400s)and F1042M(500s)?lters for program GO8581(Cycle9,P.I Reid,Gizis et al.(2003)).Some objects have been observed twice:once during our own HST program and the second time during HST program GO9157(P.I Mart′?n).Our targets are very red,thus the observations in F814W were sensitive to even lower mass companions than the observations in F675W,despite the shorter exposure time and the lower quantum e?ciency of WFPC2at longer wavelengths. For our program(GO8720),only one exposure was taken in each?lter for each object,thus not allowing to reject cosmic rays automatically.This choice was done to minimise overheads and maximise exposure times.

2.3.Data Analysis

2.3.1.Method

We processed the data in two steps.We?rst identi?ed the multiple systems either directly when resolved or by measuring the ellipticity of all the targets to look for elongated objects,using the IRAF1phot task.All the objects with an ellipticity greater than0.10 were suspected to be a binary,but all images have been inspected manually anyway to be sure we did not miss any candidate.

We used a custom-made PSF?tting program to compute the precise separation,position angle and?ux ratio of each binary system.This program makes a non-linear?t of the binary system,?tting both components at the same time.It uses di?erent PSF stars from throughout the sample and for each of them compute a model of the binary system(see Figure 1).The?ve free parameters for this model are the?ux ratio and the pixel coordinates of the two components of the binary system.A cross correlation between the model and the binary system gives us the best values for these?ve parameters.To minimise the e?ects due to the slight position sensitivity of the PSF shape in the detector and the slight change of HST focus from one orbit to another,we built a library of8di?erent PSF stars in each

–6–

?lter.As8PSF stars were not available in every?eld of view,we used the same library of PSF stars to analyse all the 6f4ce10c76c66137ee0619c7parison with analysis of images where8PSF stars were present in the?eld showed that this has no signi?cant e?ect on the accuracy of the results.

2.3.2.Calibration of the method

To evaluate the accuracy of our PSF?tting program,and look for systematic errors,we used the program on4950simulated binaries,covering a range of11various input?ux ratios (varying from0.05to1.0,expressed as F sec/F prim),135various separations(varying from 0.′′060to0.′′600,by steps of0.′′004,thus oversampling the pixel scale)and3position angles (0,22.5and45degrees,measured in the detector’s referential).It is important to note that these3angles,modulo45degrees,are su?cient to characterise16di?erent position angles by symmetry in the detector.The simulated binaries were built using unresolved objects from the sample(di?erent than the PSF stars used to make the PSF?tting).We made all these calibration in the F814W?lter and expect it to be the same in the other?lters.The S/N does not in?uence signi?cantly the conclusions of the calibrations since the library of PSF stars used to make the PSF?tting spans a large range of S/N in each?lters.

Figure2gives an overview of these calibrations for the separation and the position angle. Both plots show a periodic pattern which period is correlated to the pixel scale(P∝2×pixel scale,with a trigonometric scaling factor depending on the position angle).This e?ect is mostly due to the periodic errors introduced by the shift by a fraction of a pixel that we use to?t the binaries.

It appears that the program gives good astrometric results.For our study,we will consider that the systematic errors on the separation and on the position angle are equal to the average of the errors,and that the1σuncertainties are equal to the standard deviation of the errors.The systematic errors are negligible(less than0.′′0005on the separation and almost zero on the position angle).The1σuncertainty on the separation is0.′′0028.For the position angle it appears more appropriated to distinguish two cases:before and after 0.′′150.The spreading of the values is indeed much larger in the?rst case than in the other (see Figure2).This value of0.′′150is certainly related to the size of the FWHM at these wavelengths(～2.4pixels=0.′′110).The1σuncertainty on the position angle before0.′′150 is1.2degree,and becomes only0.3degree after0.′′150.As we could expect it(for symmetry reasons),there is no perceptible dependency on the position angle(except a slight change in the amplitude and period of the periodic pattern).The main variations are related to the di?erence of magnitude and of course to the separation.

–7–

The systematic errors and uncertainties on the di?erence of magnitudes require a more detailed analysis and description.Figure3gives an overview of these results.The errors are very dependent on the di?erence of magnitude itself and of course on the separation,but are almost independent on the position angle(only the period and amplitude of the periodic pattern depends slightly on the position angle,see above).

Again it is more appropriate to distinguish two di?erent parts in the range of separation: before and after0.′′150.The plots drawn in Figure3show also clearly that it is necessary to distinguish four di?erent cases in the range of di?erences of magnitudes:?Mag=0.00 (hereafter case1),?Mag=0.11(case2),0.11

Cases1and2are the easiest to describe.The results are excellent after0.′′150:the systematic errors and1σuncertainties are-0.01±0.01mag for the case1and0.09±0.02 mag for the case2.Before0.′′150the errors can be precisely described by a3rd order law with dispersions of only0.05mag in both cases.We will not linger over a more detailed description of these two cases since they do not describe any of our objects.

The third and fourth cases are more di?cult to analyse.

In the range of separation before0.′′150,the systematic errors in the third case can be relatively precisely described by a3rd order law with a dispersion of0.07mag.On the same range of separation,the fourth case is not reproduced as well as the other cases by a 3rd order law.Nevertheless,the dispersion is0.11mag which is still reasonable.For the rest of our study,we will assimilate the systematic errors on these ranges of separation and di?erences of magnitude to the values given by the corresponding third order laws,and the 1σuncertainties to the corresponding dispersions(see Table2).

In the range of separation after0.′′150,the third case shows also an obvious periodic pattern due mostly to the shift by a fraction of a pixel and to the normalization.Although it is obvious,this pattern cannot be easily?tted by a sinusoidal function without underestimating the systematic error,mainly because of the many points that are far away from the pattern. We thus decide to assimilate the systematic error on these values to the average(0.17mag) and the1σuncertainties to the standard deviation(±0.07mag).

The pattern still appears but less obviously in the fourth case where the spreading of the errors is much larger.Here also we assimilate the systematic error to the average of these values(0.20mag)and the1σuncertainties to their standard deviation(±0.09mag).

Table2and3gives an overview of the conclusions of these calibrations.All the values given in the text and in the tables have been corrected for the systematic errors and the given

–8–

uncertainties correspond to the1σuncertainties we calculated as explained in this section, unless speci?ed explicitly.

2.3.3.Accuracy of the results

An overview of the results is given in Table4and Figure4,and examples are given in Figure1.The uncertainties given in Table4correspond to the1σuncertainties calculated as explained in the previous section.As the observations in the F814W?lter o?er a much better S/N than in the F675W and F1042M?lters,and a much better sampling than the F675W(the di?raction limit is indeed0.′′0703in the F675W?lter,and0.′′0860in the F814W ?lter,and the pixel scale of the WFPC2/PC camera is0.′′0455),we consider them more accurate.

We therefore used only the values obtained with the F814W?lter to compute the?nal parameters given in Table4.For the data from the archive only the F814W and/or F1042M ?lters are available(the objects were too faint in the F606W?lter).

The PSF at these wavelengths is under-sampled by the0.′′0455pix?1plate scale of the Planetary Camera,thus not allowing to use deconvolution in the Fourier space.By using non-linear PSF?tting as described above,we were able to push the limit of detection down to～0.′′060arcsec for non-equal luminosity systems.Figure5shows that we should have been able to detect all the binary systems with di?erences of magnitudes between1.5 ?m F814W 3.0 mag easily even in the faintest cases,and binary systems with di?erences of magnitudes between3.0 ?m F814W 5.5mag in the brightest cases.

In the next section we will comment on the results for each object.

3.Results

3.1.Photometry,Spectral Classi?cation and Distances

An overview of values of photometry,spectral types,and distances for all the targets of the sample is given in Tables1,5,6and7.The I,J,K s values of the DENIS objects come from the DENIS survey.The J,H,K s values of the2MASS and some SDSS objects come from the 2MASS survey(Kirkpatrick et al.2000)and from Davy Kirkpatrick’s on-line archive of L and T dwarfs(Kirkpatrick2002).The I values of the2MASS and SDSS targets have been calculated as follow:using the F675W and F814W magnitudes obtained in the data(see section4.4)and the I magnitudes of the DENIS objects of our program(see Table1),we

–9–

derived the following relation:

I DENIS=?0.81+1.02×m F814W+0.28×(m F675W?m F814W)(1) with a dispersion of0.18mag(thus of the same order than the uncertainties on the DENIS magnitudes themselves)and used it to compute the I magnitudes2given in Table1.

The magnitudes have been measured with standard procedures using the aperture pho-tometry task phot in IRAF,using sky subtraction and a3σrejection.We used a recom-mended aperture of0.′′5in the case of unresolved objects,and an aperture of1.′′365(30 pixels)in the case of binaries,to measure the total?ux of the system.The counts were transformed to magnitudes using the relation:m[mag]=?2.5·log(counts/exp)+ZP?0.1 for the unresolved objects,where counts is the number of counts measured with IRAF, exp the exposure time,ZP the Vega Zero Point magnitude(ZP F675W=22.042mag, ZP F814W=21.639mag,and ZP F1042M=16.148mag,Biretta et al(2002))and?0.1 is to correct from the?nite to in?nite aperture,according to the HST data handbook,and the relation:m[mag]=?2.5·log(counts/exp)+ZP?0.028in the case of binaries,where ?0.028is the correction we evaluated as suggested in the HST data Handbook to correct from the?nite to in?nite aperture in that case.For multiple systems,we deduced the mag-

nitude of each component using our values of?ux ratios.In many cases we just had one image thus not allowing to remove cosmic ray events automatically.Nevertheless cosmic rays should not have in?uenced our photometry signi?cantly since we would have been able to detect any cosmic ray event too close to the object to be removed by the3σautomatic rejection.This situation happened in only one case(2MASSW2147+1431,F1042M?lter). All the cosmic ray events su?ciently far away from the object have been corrected by the 3σautomatic rejection of the IRAF phot task.The results are given in Tables4,5,6and7. Uncertainties for the unresolved objects are0.1mag.

A few DENIS objects and most2MASS and SDSS objects had accurate values of spectral types found in Davy Kirkpatrick’s on-line archive of L and T dwarfs(Kirkpatrick2002),and obtained through spectroscopic measurements.For the other objects,we used the Dahn et al.(2002)spectral type vs.(I-J)colour relation to deduce the spectral types given in Table 1.These latter spectral types will have to be con?rmed by spectroscopic measurements.The sample thus covers a large range of spectral types,going from～M8to～L8.

Two e?orts by the United States Naval Observatory and by an Australian group led by Chris Tinney are currently under way in order to derive precise angular parallaxes for these objects.Some of the targets already have published angular parallaxes and distances

–10–

(Dahn et al.2002).For the unresolved objects without published trigonometric parallax, we evaluated the photometric parallax using the Dahn et al.(2002)M J vs.Spectral Type relation when both J and spectral type where available,and the Dahn et al.(2002)M J vs. (I-J)relation in the other cases.For the multiple systems,we evaluated the distance using the Dahn et al.(2002)M I vs.Spectral Type relation and then multiplied the photometric

√

distance obtained by a correction factor of

–11–

pixels,increasing the uncertainties.As the F814W image is more sensitive than the F675W one,and as the secondary is brighter(and thus easier to detect)in the F814W?lter,we consider hereafter the corresponding value to be the best one.We measured di?erences of magnitudes of?m F675W=1.47±0.19mag and?m F814W=0.65±0.16mag,but these values could be altered by the fact that the PSF of the secondary is elongated.

The second set of data(GO9157)was better and more accurate values were obtained. Several exposures were available,thus allowing to reject cosmic rays.We measured a sep-aration of370.0±2.8mas and a position angle of255.8±0.3degrees.The di?erence of magnitude is?m F814W=0.63±0.09mag and is in good agreement with the previous one, suggesting that even if the PSF of the secondary was elongated,the value was not altered too much.Both Leggett et al.(2001);Koerner et al.(1999)report a di?erence of magnitude of about0mag in the IR.Figure1shows clearly that this is not the case in the F814W ?lter.Our results on the separation of the system are in good agreement with the one given by Leggett et al.(2001)within the uncertainties but do not agree with the one given by Koerner et al.(1999).We consider that our observations and the observations reported by Leggett et al.(2001),based on higher spatial resolutions than the one of Koerner et al. (1999),are more accurate.

To evaluate the orbital period of this system,we evaluate its semi-major axis by mul-tiplying the projected separation by a statistical correction factor of1.26as explained in Fischer&Marcy(1992)(see Table8).Then assuming a distance of19.8pc,with a total mass of～0.14M⊙(masses of the two components are unknown,but the presence of methane implies a mass less than0.07M⊙and the absence of lithium implies a mass greater than 0.055M⊙,assuming an age greater than0.5Gyr),and a semi-major axis of9.2A.U,the period of this system would be～74.6years.

3.3.DENISPJ035729.6-441730

DENISPJ035729.6-441730is one of the new binary system candidates we have discovered in this survey.It consists of a very close binary,with a separation of98±2.8mas and a position angle of174.7±1.2degrees.The di?erences of magnitude are?m F675W=1.23±0.11 mag and?m F814W=1.50±0.11mag,suggesting that the two components have slightly di?erent masses.Despite the very small separation,(we are here very close to WFPC2/PC pixel scale),the values obtained for the separation and the position angle in the two?lters are in very good agreement one with each other(cf.Figure4).The companion appears clearly after PSF?tting(see Figure6).

–12–

As described in section3.1,we evaluated its photometric distance corrected for multi-

plicity of22.2pc.To the projected separation of22.2×0.098=2.2A.U should correspond

a semi-major axis of2.2×1.26=2.8A.U.Assuming a face-on orbit,with a total mass of

～0.20M⊙(masses are here unknown and we assume masses of about～0.1M⊙for each component),and a semi-major axis of2.8A.U,the period of this system would be～10.5

years.

3.4.2MASSW0746425+2000032

2MASSW0746425+2000032(spectral type L0.5(Kirkpatrick et al.2000))has been

discovered by Kirkpatrick et al.(2000),and was suggested to be a binary by Reid et al.

(2000)based on its position in a colour-magnitude diagram.It has been con?rmed as a

multiple system by Reid et al.(2001),with a separation of0.22arcsec and a position

angle of15degrees.On the same data set we measured a separation of219±2.8mas and

a position angle of168.8±0.3degrees(cf.Figure4).There is a discrepancy with the

position angle given by Reid et al.(2001).To measure the position angle,we used the

positions of each component in the detector,and measured the corresponding position angle

of23.6degrees(in the detector).We then added145.212degrees due to the orientation

angle of the HST spacecraft(given by the ORIENTAT header keyword),to obtain the168.8

degrees mentioned above.Figure7shows the image of2MASSW0746425+2000032and its

orientation.Position angles are measured from the North to the East.

The distance of2MASSW0746425+2000032was estimated by trigonometric parallax

measurements to be12.3pc.The projected radius of219mas thus corresponds to a semi-

major axis about0.219×12.3×1.26=3.4A.U.

We measured?m F814W=1.0±0.09mag.On the same set of data Reid et al.(2001)

measured a di?erence of magnitude of?I=0.62mag,lower than the one we measure,but

the uncertainties are not given.

Assuming a semi-major axis of3.4A.U,with a total mass of～0.20M⊙(masses are

here unknown and we assume masses of about～0.1M⊙for each component),the period of

this system would be～14.0years.

3.5.2MASSW0850359+105716

2MASSW0850359+105716was a known binary brown dwarf(Reid et al.2001)of

spectral type L6in the Kirkpatrick classi?cation scheme(L5in the Mart′?n classi?cation

–13–

scheme)implying T eff<1800K.This object shows a strong lithium line(Kirkpatrick et al.1999)implying a mass M≤0.06M⊙.Its distance(25.6pc)and proper motion (μ=144.7±2.0mas/yr)were estimated by USNO parallax measurements(Dahn et al. 2002).Reid et al.(2001)measured a separation of0.16arcsec and a position angle of250 degrees on the1st February2001using HST/WFPC2.On the same data set we measured a separation of157±2.8mas and a position angle of114.7±0.3degrees(cf.Figure4).There is again a discrepancy between the values measured for the position angle.Figure7shows the image of2MASSW0850359+105716and its orientation.

We measured di?erence of magnitude of?m F814W=1.47±0.09mag.This implies again a slight di?erence in the masses of the two components.On the same set of data Reid et al.(2001)measured a di?erence of magnitude of?I=1.34mag,in good agreement with our value.Assuming a semi-major axis of0.157×41.0×1.26=8.1A.U,with an assumed total mass of～0.14M⊙(masses of the two components are unknown,but the presence of lithium in a L5dwarf implies a mass less than0.06M⊙for one of them at least),the period of this system would be～61.6years.

3.6.2MASSW0856479+223518

2MASSW0856479+223518is a good candidate binary we report in this paper.It has been observed on the24th of April2001in the F814W and F1042M?lters,but was not resolved in the latter one.We measured a separation of98±9mas and a position angle of 275±2.0degrees.The di?erence of magnitude is?m F814W=2.76±0.11mag,suggesting that the two components have probably di?erent masses.This large di?erence of magnitudes at such a small separation makes it very di?cult to measure precisely the di?erent parame-ters,explaining the higher uncertainties on the values.Nevertheless the two centroids appear clearly on the image(cf.Figure8)and the PSF?tting makes appear a faint but obvious companion(see Figure6).As we reach here the limit of detection of our method,the mul-tiplicity of this object should be considered with caution and further followup observations with higher spatial resolution are required in order to con?rmed that it is a real binary.

As described in section3.1,we evaluated a photometric distance corrected for the mul-tiplicity of d=34.7pc,which leads to a semi-major axis of4.4A.U,and with a total mass of～0.20M⊙,to a period of～20.8years(masses are here unknown and we assume masses of about～0.1M⊙for each component).

–14–

3.7.2MASSW0920122+351743

2MASSW0920122+351743(L6.5)has been identi?ed as a brown dwarf by Kirkpatrick et al.(2000)and as a binary by Reid et al.(2001).Nakajima et al.(2001)report the observation of methane in the H and K bands,which make more di?cult the de?nition of the transition between L and T dwarfs.

For that object,Reid et al.(2001)report a separation of75mas and a position angle of90degrees on the2nd of September2000.On the same set of data we measured a separation of75±2.8mas and a position angle of248.5±1.2degrees(cf.Figure4and 7).Once again there is a discrepancy in the position angle.Figure7shows the image of2MASSW0920122+351743and its orientation.We measured a di?erence of magnitude ?m F814W=0.88±0.11mag.On the same set of data Reid et al.(2001)measured a di?erence of magnitude of?I=0.44mag,lower than the one we measure,but the uncertainties are not given.Considering only our uncertainties,the two values are di?erent by4σ.

This makes of2MASSW0920122+351743one of the closest resolved binaries observed in the sample.Figure6shows the companion clearly appearing after PSF?tting.Kirkpatrick et al.(2000)estimated a distance of21pc for the unresolved 6f4ce10c76c66137ee0619c7ing our photometric measurement we derive a distance of16.7pc.Correcting for multiplicity it gives a distance of20.1pc for the multiple system.We can estimate the semi-major axis to be about1.9 A.U.

Assuming a total mass of～0.14M⊙(masses of the two components are unknown,but the presence of methane in this L6.5dwarf implies a mass of less than0.07M⊙),and a semi-major axis of1.9A.U,the period of this system would be～7.2years.

3.8.DENISPJ100428.3-114648

DENISPJ100428.3-114648is a new binary system candidate with a separation of146±2.8mas and position angle of304.5±1.2degrees.It has also been reported in the2MASS Survey as2MASSW1004282-114648.The di?erences of magnitudes are?m F675W=0.25±0.07mag and?m F814W=0.66±0.11mag.Once again the agreement between the values for separation and position angle obtained in the two?lters is very good(cf.Figure4).As explained in section3.1,we estimated its photometric distance corrected for binarity to be 46.8pc.

Assuming a semi-major axis of8.6A.U,with a total mass of～0.20M⊙(masses are

–15–

here unknown and we assume masses of about～0.1M⊙for each component),the period of this system would be～56.5years.

3.9.2MASSW1017075+130839

2MASSW1017075+130839is a new binary candidate we report in this paper.It has been observed on the16th of April2001in the F1042M and F814W?lters.We measure a separation of104±2.8mas and a position angle of92.6±1.2degrees.The di?erence of magnitude are?m F1042M=0.74±0.11mag,and?m F814W=0.77±0.11mag,suggesting that the two components have probably similar masses.

We evaluated the photometric distance corrected for binarity at～21.4pc.Assuming a semi-major axis of2.9A.U,with a total mass of～0.20M⊙(masses are here unknown and we assume masses of about～0.1M⊙for each component),the period of this system would be～11.0years.

3.10.2MASSW1112257+354813

2MASSW1112257+354813is also one of the15new binary candidates we report in this paper.It has been observed on the14th of February2001in the F814W and F1042M ?lters.For this object we measure a separation of70±2.8mas and a position angle of 79.6±1.2degrees.The di?erences of magnitude are?m F814W=1.07±0.11mag and ?m F1042M=1.04±0.11mag.Figure6shows the companion appearing clearly after PSF ?tting.

This object is particularly interesting since2MASSW1112257+354813is also known as Gl417B,con?rmed as a L-dwarf and as companion of the G0-dwarf Gl417A by Kirk-patrick et al.(2001),who measured a separation between the G-dwarf primary and the2MASSW1112257+354813unresolved system of90.′′0and a position angle of245de-grees.For such a system to be stable,the ratio of the semi-major axis has to be greater than～5,which is here clearly the case(ρAB/ρBC=90/0.070～1300).Kirkpatrick et al.(2001)estimated an age of0.08-0.3Gyr for the G-dwarf primary and,assuming that the primary and its companions are coeval,they computed a mass of0.035±0.015M⊙for 2MASSW1112257+354813,well below the hydrogen burning limit.The G-dwarf primary appears clearly on the HST/WFPC2image and saturates completely one of the four CCD of the HST/WFPC2camera.

This triple system is thus similar to Gl569Bab(Mart′?n et al.2000b;Kenworthy et al

–16–

2001)and HD130948(Potter et al.2002)(see Table9).

The distance of the G-dwarf primary has been evaluated by trigonometric parallax (π=46.04mas or d=21.7pc;Hipparcos,Perryman et al.(1997))and Kirkpatrick et al. (2000)assigned a spectral type of L4.5.Assuming a semi-major axis of1.26×21.7×0.070= 1.9A.U,with a total mass of～0.20M⊙(masses are here unknown and we assume masses of about～0.1M⊙for each component),the period of this system would be～5.8years only. As its distance is known precisely,and as its age can be constrained by studying the G-dwarf primary,this system is a very good and very promising candidate to constrain theory and the mass/luminosity function.

3.11.2MASSW1127534+741107

2MASSW1127534+741107is a M8.0dwarf(Gizis et al.2000)and also one of the15 new binary candidates we report in this paper.It has been observed on the5th of May 2001in the F1042M and F814W?lters.We measure a separation of252.5±2.8mas and a position angle of79.9±0.3degrees.The di?erences of magnitude between the primary and the secondary are?m F1042M=0.39±0.07mag and?m F814W=0.82±0.09mag,suggesting that the two components have probably similar masses.

We evaluated the photometric distance corrected for multiplicity at14.6pc.Assuming a semi-major axis of4.8A.U,with a total mass of～0.20M⊙(masses are here unknown and we assume again masses of about～0.1M⊙for each component),the period of this system would be～23.5years.

3.12.2MASSW1146344+223052

2MASSW1146344+223052has been discovered as a L3dwarf by Kirkpatrick et al. (1999)and as a binary by Koerner et al.(1999).It shows lithium lines and Basri et al.(2000)estimated its e?ective temperature to1950K.Its distance(27.2pc)has been measured by trigonometric parallax(Dahn et al.2002).Reid et al.(2001)measured a separation of0.′′29and a position angle of199.5degrees.On the same set of data we measured a separation of294±2.8mas and a position angle of199.5±0.3degrees(cf.Figure4).This time both position angle and separation are in agreement with the values given by Reid et al. (2001).We also measured a di?erence of magnitude?m F814W=0.75±0.09mag.On the same set of data Reid et al.(2001)measured a di?erence of magnitude of?I=0.31mag, again smaller than the one we measure,but the uncertainties are not given.Considering

–17–

only our uncertainties,the two measurements are di?erent by5σ.Such a di?erence is much larger than the uncertainties on our measurement,as explained in section2.3,especially considering that with a separation of～0.′′29and a di?erence of magnitude less than1.0, 2MASSW1146344+223052was an“easy”case for the PSF?tting program.

Assuming a semi-major axis of10.1A.U,with a total mass of～0.12M⊙(masses of the two components are unknown,but the presence of lithium in a L7dwarf implies a mass equal or less than0.06M⊙),the period of this system would be～92.6years.

3.13.DENISPJ122813.8-154711

DENISPJ122813.8-154711has been discovered by Delfosse et al.(1997a)and resolved for the?rst time by Mart′?n et al.(1999b).It is a L4.5brown dwarf in the Mart′?n classi?-cation scheme and a L5in the Kirkpatrick classi?cation scheme(Kirkpatrick et al.1999). This object has been observed and studied several times.We provide a summary of the astrometric and photometric measurements in Table10.

This object has also been observed in the2MASS survey and is reported as2MASSWJ-1228152-154734.The designations are di?erent mostly because the DENIS astrometric pipeline used at the time of discovery was not the?nal version and the uncertainties of the astrometry were high.

Lithium absorption has been reported by both Mart′?n et al.(1997)and Tinney et al.(1997),implying a mass M≤0.06M⊙.DENISPJ122813.8-154711was the second?eld brown dwarf to be con?rmed by the lithium test.

For this object again we had two sets of data.The results obtained with the?rst set are not very accurate because of coordinate mismatch.The target fell in a corner of one of the Wide Field Camera of WFPC2instead of in the centre of the Planetary Camera(PC),where the distortions are more important and the pixel scale coarser,implying a worse sampling of the PSF and thus higher uncertainties.Thus the results in the two?lters are not in very good agreement(cf.Table10and Figure4).The second set of data was much better since the target was in the centre of the PC and the exposure time was1700s instead of300s,spread over several exposures and allowing to reject cosmic rays easily.The accuracy of the result is then very good.The di?erences of magnitude in both?lters indicate that the two components are probably very similar:?m F814W=0.36±0.07mag and?m F675W=0.53±0.09mag.

DENISPJ122813.8-154711has been observed now over more than5years,and we have astrometric data related to the binary spread over more than3years.Within a few years we

–18–

should thus be able to compute orbital parameters and dynamical masses.The motion of DENISPJ122813.8-154711-B is already obvious.We see in Table10a separation change by ～8%and a position angle change by23degrees in three years.If we assume a semimajor axis of6.6A.U,and a total mass of0.12M⊙,we?nd an orbital period of～49years.

3.1

4.2MASSW1239272+551537

2MASSW1239272+551537is one of the new binary candidates we report in this paper. It has been observed on the18th of March2001in the F814W?lter.We measure a separation of211±2.8mas and a position angle of187.6±0.3degrees.The di?erence of?ux between the primary and the secondary is?m F814W=0.34±0.07mag and?m F1042M=0.54±0.09 mag,suggesting that the two components might have very similar masses.The whole system is classi?ed as L5by Kirkpatrick et al.(2000),who also evaluated the photometric distance of the whole system considering it was a single object(17pc).From our photometric measurement we calculate a distance(corrected for the multiplicity)of21.3pc.Assuming a semi-major axis of5.9A.U,with a total mass of～0.20M⊙(masses are here unknown and we assume masses of about～0.1M⊙for each component),the period of this system would be～32.0years.

3.15.2MASSW1311392+8032222

2MASSW1311392+8032222is a M8.0dwarf(Gizis et al.2000)and is also one of the new binaries we report in this paper.It has been observed on the30th of July2000in the F1042M and F814W?lters.We measure a separation of300±2.8mas and a position angle of167.3±0.3degrees.The di?erence of?ux between the primary and the secondary is ?m F1042M=0.45±0.09mag and?m F814W=0.0.39±0.07mag,suggesting that the two components have similar masses.

We evaluate a photometric distance corrected for binarity of13.7pc.Assuming a semi-major axis of5.5A.U,with a total mass of～0.20M⊙(masses are here unknown and we assume masses of about～0.1M⊙for each component),the period of this system would be ～28.8years.

–19–

3.16.2MASSW1426316+155701

2MASSW1426316+155701is a M9dwarf and has been reported as a binary for the?rst

time by Close et al.(2002a,b)with Hokupa’a/Gemini.On the22nd of September2001, they measured a separation of152±6mas and a position angle of344.1±0.7degrees.They estimated an age of0.8+6.7

?0.2

Gyr and a photometric distance of23.6±6.0pc,giving a separation

of3.6±0.9A.U and a period of17+10

?7

years.They also estimated the mass of the whole

system M tot=0.140+0.011

?0.026M⊙as well as the masses of each components:M A=0.074+0.011

?0.005

M⊙at the limit of the Hydrogen burning limit for the primary and M B=0.066+0.015

?0.006

M⊙for the secondary.

On the17th of July2001we measured a separation of157.1±2.8mas and a position angle of339.9±0.3degrees.We also estimate a photometric distance corrected for binarity of26.7pc.These values are close to the one of Close et al.(2002b)taken～2months later (within1σ).Gizis et al.(2000)measured a proper motion ofμ=0.′′121/yr,implying20 mas of motion between the observations,con?rming that the system is a common proper motion pair.We also measured di?erences of magnitudes of?m F814W=1.40±0.09mag, and?m F1042M=1.30±0.09mag.Assuming a total mass of0.14M⊙as given by Close et al.(2002a,b)and a semi-major axis of4.3A.U,we estimate the period of this system to be 33.3years.

3.17.2MASSW1430436+291541

2MASSW1430436+291541is one of the15new binary candidates we report in this paper.It has been observed on the19th of April2001in the F1042M and F814W?lters. We measure a separation of88±2.8mas and a position angle of327.5±1.2degrees.The di?erence of?ux between the primary and the secondary is?m F1042M=0.76±0.11mag and?m F814W=0.78±0.11mag,suggesting that the two components have slightly similar masses.The companion appears clearly after PSF?tting,as shown in Figure6.

We evaluated the distance(corrected for multiplicity)of29.4pc.Assuming a semi-major axis of3.4A.U,with a total mass of～0.20M⊙(masses are here unknown and we assume masses of about～0.1M⊙for each component),the period of this system would be ～13.9years.

–20–

3.18.DENISPJ144137.3-094559

DENISPJ144137.3-094559is a binary brown dwarf of spectral type L1(Mart′?n et al. 1999a),?rst resolved in February2000with Keck/NIRC(Mart′?n,E.,private communica-tion).It has been observed with Keck by Stephens et al.(2001)on the13th of April2000 and resolved with a separation of0.42arcsec.Kirkpatrick et al.(2000)deduced a dis-tance of25.5pc using photometric measurements,considering a single object.We estimate a photometric distance corrected for multiplicity of29.2pc.

For this object we had two epochs of data,one from our HST program(two?lters)and one from an on going program(GO9157).

In the?rst set of data obtained on the16th of January2001we measured a separation of375±2.8mas and a position angle of290.3±0.3degrees.The values obtained in both ?lters are again in very good agreement(cf.Figure4).We measured?m F675W=0.63±0.09 mag and?m F814W=0.37±0.07mag.In the second set of data,obtained four months later on the22nd of May2001,we measured a separation of370±2.8mas,a position angle of 290.8±0.3degrees,and a di?erence of magnitude?m F814=0.30±0.07.These data are in good agreement with the previous one.Assuming a semi-major axis of14.1A.U,with a total mass of～0.2M⊙(masses are here unknown and we assume masses of about～0.1 M⊙for each component),the period of this system would be～118years.

3.19.2MASSW1449378+235537

2MASSW1449378+235537is another of the new binary candidates we report in this paper.The spectral type of the whole system is L0(Kirkpatrick et al.2000).It has been observed on the21st of December2000in the F814W and F1042M?lters.We measure a separation of134±2.8mas and a position angle of64.4±1.2degrees.The di?erence of?ux between the primary and the secondary is?m F814W=1.51±0.11mag and?m F1042M= 1.08±0.11mag,suggesting that the secondary is redder than the primary.

Kirkpatrick et al.(2000)evaluated the photometric distance of the unresolved system to62pc.From our photometric measurements we estimate a photometric distance corrected for binarity of63.7pc.Assuming a semi-major axis of11.0A.U,with a total mass of～0.20 M⊙,the period of this system would be～81.1years.

–21–

3.20.2MASSW1600054+170832

2MASSW1600054+170832is one of the15new binary candidates we report in this paper.It is a L1.5dwarf according to Kirkpatrick et al.(2000).It has been observed on the14th of January2001in the F814W and F1042M?lters.We measure a separation of 57±2.8mas and a position angle of156.2±1.2degrees.The di?erence of?ux between the primary and the secondary is?m F814W=0.69±0.11.At such a close separation,it might be better to consider2-or3-σuncertainties on these values,since we reach here the very limit of resolution of the WFPC2/PC camera.Nevertheless,the companion appears once again clearly after PSF subtraction(see Figure6).The object was too faint for PSF?tting in the F1042M?lter.

Kirkpatrick et al.(2000)evaluated the photometric distance of the unresolved system to62pc.We estimate a photometric distance(corrected for multiplicity)of the primary of 60.6pc.Assuming a semi-major axis of4.3A.U,with a total mass of～0.20M⊙(masses are here unknown and we assume masses of about～0.1M⊙for each component),the period of this system would be～19.9years.

3.21.2MASSW1728114+394859

2MASSW1728114+394859is also one of the new binary candidates.The spectral type of the whole system is L7(Kirkpatrick et al.2000).It has been observed on the12th of August2000in the F814W and F1042M?lters.We measure a separation of131.3±2.8mas and a position angle of27.6±1.2degrees.The di?erence of?ux between the primary and the secondary in this range of wavelength is?m F814W=0.66±0.11mag,suggesting that the two components have similar masses.This object was also too faint for PSF?tting in the F1042M?lter.

Kirkpatrick et al.(2000)evaluated the photometric distance of the unresolved system (20pc).We estimate the photometric distance(corrected for multiplicity)at20.4pc.As-suming a semi-major axis of3.4A.U,with a total mass of～0.20M⊙,the period of this system would be～13.8years.

3.22.2MASSW2101154+175658

2MASSW2101154+175658is a L7.5dwarf according to Kirkpatrick et al.(2000)and is a new binary we report in this paper.It has been observed on the7th of May2001

本文来源：https://www.bwwdw.com/article/p9hl.html