A&A 468, 163-170 (2007)
DOI: 10.1051/0004-6361:20066844

Southern infrared proper motion survey. II. A sample of low mass stars with $\rm {\mu \geq 0.1''/ {yr}}$[*]

N. R. Deacon1,2 - N. C. Hambly2


1 - Department of Astrophysics, Faculty of Science, Radboud University Nijmegen, PO Box 9010, 6500 GL Nijmegen, The Netherlands
2 - SUPA[*], Institute for Astronomy, University of Edinburgh, Royal Observatory Edinburgh, Blackford Hill, Edinburgh EH9 3HJ, UK

Received 30 November 2006 / Accepted 2 March 2007

Abstract
We present details of the second part of the Southern Infrared Proper Motion Survey (SIPS). Here accurate relative astrometry allows us to reduce the minimum proper motion to 0.1 arcsec per year. This yields 6904 objects with proper motions between our minum cut and half an arcsecond a year. A small overspill sample with proper motions greater than this is also included. We examine our sample to identify interesting individual objects such as common proper motion binaries, potential L dwarfs and candidate nearby stars. Finally we show our survey is incomplete due to many factors, factors which we will take into account when simulating these survey results in the next paper in this series.

Key words: astrometry - stars: low mass, brown dwarfs

1 Introduction

Proper motion surveys provide a wealth of information for the study of Low Mass Stars and Brown Dwarfs. Firstly studies of the Luminosity Function (and hence Mass Function, birthrate and space density) require large, clean samples of cool objects. Proper motion surveys provide this by allowing the easy exclusion of distant, intrinsically bright contaminants such giants. Secondly they can also identify common proper motion binaries. The make-up of these systems (and of their individual components) can provide interesting insights, not only into multiplicity, but into the star formation processes that created them (Burgasser et al. 2005).

In Deacon et al. (2005) (hereafter DHC) we combined J, H, and $K_{\rm S}$ data from the 2MASS survey with IN data from SuperCOSMOS (Hambly et al. 2001) scans of UKST plates to produce an infrared proper motion survey (SIPS) with a lower proper motion limit of half an arcsecond a year. Using such an infrared proper motion survey allowed us to study low mass stars and Brown Dwarfs in the passbands in which they are brightest. This yielded approximately 70 new high proper motion objects. While many of these objects were interesting in themselves, it was clear that a reduction in the lower proper motion limit (and hence an increase in the number of objects detected) was required to produce a significant sample. Henced we have produced a survey with a lower proper motion limit of 0.1''/yr. For the sake of comparison with SIPS I (DHC) we take the maximum proper motion of our sample to be 0.5/yr, although there will be a number of objects which spill over this limit.

The majority of current proper motion surveys have focussed on identifying objects with high proper motions. Many, such as Lepine & Shara (2002), have proper motion limits well above our maximum proper motion limit of 0.5''/yr. Others (Scholz et al. 2002; Subasavage et al. 2005; Lepine 2005) have lower limits of 0.45-0.4''/yr, just encroching on the region covered by our sample. Of those which go to lower limits two, Ruiz et al. (2001) and Wroblewski & Costa (2001), are limited in the areas of sky they cover. This leaves only three surveys which cover the majority of the southern hemisphere to lower proper motion limits comparable to ours: Luyten's New Luyten Two Tenths catalogue (NLTT, Luyten 1979b), Pokorny et al.'s (2003) Liverpool-Edinburgh High Proper Motion survey (LEHPM) and Finch et al. (2007). Both Luyten and Pokorny have lower proper motion limits of 0.2''/yr, however Luyten's survey is incomplete below $\delta=-30^\circ$. The survey by Finch et al. (2007) expands the work done by Subasavage et al. (2005) to a lower proper motion limit of 0.18''/yr. This survey covers the area south of $\delta=47^{\circ}$, i.e. the area where Luyten's surveys are most incomplete. Many of the objects identified here are also present in this study. All of these surveys primarily make use of optical data.


  \begin{figure} \par\includegraphics[width=8.5cm,clip]{twomassswirl.ps}\end{figure} Figure 1: The errors across one single UKST I plate (field 411). The axes represent the position on the plate in centimetres and the size and direction of the lines on each lollipop show the positional offsets. The scale for the length of the lines is 4 cm = 1''.
Open with DEXTER


  \begin{figure} \par {\includegraphics[width=5.7cm,clip]{IJhist.ps} } \par\vspace*{2mm} {\includegraphics[width=5.7cm,clip]{fig2b.ps} } \par\end{figure} Figure 2: Histograms showing the magnitude (J) and colour (I-J) of the objects found in this survey with proper motions between 0.1 and 0.5 arcsec per year.
Open with DEXTER

Hence the approximately 7000 objects presented here provide a large proper motion selected sample of low mass stars of comparable area and depth to those currently available but with better completeness for cool, red dwarfs. Here we present details of this sample along with a selection of interesting objects contained within it. In the third paper of this series we will outline the method used to simulate the survey results and the constraints on underlying distributions fundamental to star formation that can be set from these.

2 The SIPS selection method

In Deacon et al. (2005) we outlined the SIPS selection method. Here we give a brief recap.

The first stage of candidate selection is using 2MASS photometry. Here objects are plotted on a J-H vs. $H-K_{\rm s}$ colour-colour diagram. An object's position on this diagram leads it to be classified as a potential M/L dwarf, early T dwarf or mid-late T dwarf. There is also a fourth category in an overlap region between the mid M and mid T dwarf range. Any object that does not fall into one of the four categories is rejected. Next each candidates object had to be paired with an I plate counterpart. In the first run of the SIPS survey we simply used the positions from both the UKST I plates and the 2MASS survey to calculate the movement of an object and hence its proper motion. As the minimum proper motion in this case was half an arcsecond per year this was generally well above a $5\sigma$limit. However by reducing the lower proper motion limit to 0.1''/yr we run the risk of large errors in the measured proper motions and therefore spurious detections. Hence we employ a relative error mapping technique to reduce the rms error on the proper motions. The details of this are outlined in Sect. 2.1. Other than this the candidate selection method is identical to that outlined in DHC with I plate candidates being selected and then filtered on I-J colour, ellipticity, and being a good single image far from bright stars. Following the initial selection process all candidates were inspected by eye to reduce the number of spurious detections. The proper motions were then calculated using the same method as DHC.

Table 1: Objects with I-J>3.5 found in this survey.

Table 2: Objects with photometry suggesting they are L dwarfs. Our photometric spectral classifications are based on the colours in Kirkpatrick et al. (2000).


  \begin{figure} \par\includegraphics[width=6cm,clip]{CPMhist.ps}\end{figure} Figure 3: A histogram showing the seperations of objects with common proper motions in the SIPS survey. The line shows the expected distribution of randomly distributed objects, ${\rm d}n \propto r {\rm d}r$.
Open with DEXTER


  \begin{figure} \par\includegraphics[width=6cm,clip]{PMhist.ps}\end{figure} Figure 4: The cumulitive number of objects with proper motions greater than the minimum limit of each bin. The solid line shows the $N \propto \mu ^{-3}$ trend which would be expected for a totally complete sample. It is clear that the sample is incomplete below 0.2''/yr. The proper motion ($\mu $) is in arcseconds/year.
Open with DEXTER

   
2.1 Relative astrometry

In order to map out the systematic astrometric errors between 2MASS and the SSS, we employed the 2MASS catalogue positions as a standard. Robust median offsets in X and Y in 1cm boxes were computed over the field of view of each I plate, and the residuals were smoothed and filtered within $3\times3$ boxes to create a map of any systematic offsets between the photographic plate and 2MASS astrometry. A typical example of the resulting systematic positional error map is shown in Fig. 1, with errors at the field edges of up to 0.5 arcsec. These positional error maps were then applied to the photographic positions before using them, in conjunction with the 2MASS epoch positions, to measure proper motions.

3 Results

The objects found in this survey with $\rm0.1''/yr<\mu <0.5''/yr$ are shown in Table 7. Figure 2a shows an I-J histogram for all objects in this sample, while Fig. 2b shows a histgram for J magnitudes. It is clear that the objects make up a large sample of mid-late M dwarfs.

Table 3: SIPS objects which share common proper motions with other SIPS objects. PA = Position Angle. For NLTT objects see Luyten, for LEHPM objects see Pokorny et al. (2003), for WT objects see Wroblewski & Torres (1991) and for 1 see Giclas et al. (1971). Pairs marked with * were identified as binaries by Luyten (1988). The pair marked with ** was found simulatiously by Artigau et al. (2007).

Table 4: SIPS objects which share common proper motions with objects found in other studies which are not themselves SIPS objects. These pairings were found during the cross referencing process and should not be regarded as a complete list. PA = Position Angle. For LEHPM objects see Pokorny (2003), for NLTT objects see NLTT, 1 see Lasker et al. (1990), 2 see Gizis et al. (2000) and 3 see Schonfeld (1886). All pairs marked * were identified as common proper motion pairs by Luyten (1988).

3.1 Higher proper motion objects

To prevent objects which had a proper motion below 0.5''/yr before the realtive astronometry correction, but above 0.5''/yr after it from slipping through the net, objects with proper motions just above our upper limit were also examined. Hence we also identified several objects with proper motions greater than our upper limit of 0.5''/yr, these are shown in Table 8.

3.2 Interesting objects

Rather than examine every object in detail we have sought to select interesting individual objects, the details of which are listed in the following three subsections.

3.2.1 Red objects
In order to identify potential L and T dwarfs, we selected every object found with I-J>3.5. The photometry and proper motions of these objects are shown in Table 1. Comparing the $J-K_{\rm s}$colours with those for different spectral types in Kirkpatrick et al. (2000) it becomes clear that SIPS2255-5713, SIPS0128-5545 and SIPS0539-0059 are early-mid L dwarfs (i.e. earlier than L6). In fact Fan et al. (2000) spectroscopically identified that SIPS0539-0059 (SDSS 0539-0059) is an L5 dwarf. Additionally Reid et al. (in prep.) find that SIPS0719-5051 (2MASS J07193188-5051410) is an L0 dwarf. Kendall et al. (2006) identify SIPS2255-5713 and SIPS0128-5545 to be L5.5 and L1 respectively. It appears from the $J, H, K_{\rm s}$ colours that SIPS1625-2508 is a late M dwarf.

Table 5: Objects in this sample with proper motions in the range $\rm0.1''/yr<\mu <0.5''/yr$ which have estimated distances closer that 20 pc.

Table 6: Objects in this sample with proper motions above 0.5''/yr which have estimated distances closer that 20 pc.

We also compared object's J, H, and $K_{\rm S}$ photometry with the values which Kirkpatrick et al. (2000) quote as typical for different spectral types. All objects redder in J-H, $J-K_{\rm S}$, $H-K_{\rm S}$ than Kirkpatrick's values for an M9 dwarf are listed in Table 2. These objects are given spectral types based on their colours and those given in Kirkpatrick et al. (2000). SIPS2255-5713, SIPS0128-5545 and SIPS0539-0059 appear in thesamples in both Tables 1 and 2.

3.2.2 Common proper motion objects
During the visual inspection phase of the data reduction process it became apparent that there were many SIPS objects which shared a common proper motion. However it is often difficult to distinguish coincidence objects with the same proper motion from gravitiationally bound wide binaries. To separate these two classes of objects we plotted the separations of objects which had proper motions within $2\sigma$ of each other. This is shown in Fig. 3. The straight line marks the expected distribution of coincidence objects randomly placed around the other object. Clearly the vast majority follow this pattern. However the higher that expected number of pairs with separations less than three arcminute indicates that a population of binaries also contributes to this. Hence we choose a maximum separation of three arcminutes for our binary sample. A list of SIPS objects with separations less than this and proper motions within $2\sigma$ of each other is shown in Table 3.

Additionally during the cross-referencing process, several objects which, while clearly not the target SIPS object, had a similar proper motion to it were found. These were further investigated and any companion found to have a SuperCOSMOS proper motion (Hambly et al. 2001b) differing by less than $2\sigma$ from the SIPS object and to be closer than three arcminutes to it is listed in Table 4. Note that due to the manual nature of this selection mechanism this list should not be regarded as complete.

In some cases it appears that the redder object of a particular pair is actually brighter than the bluer object. However closer examination reveals that these differences are not inconsistent with the typical photometric errors of $\sim$0.3 mag (Hambly et al. 2001c).

Table 7: An example section of Table 7 which contains all objects in the survey with $\rm0.1''/yr>\mu >0.5''/yr$. The full version of Table 7 along with a citation key is available from CDS.

Table 8: An example section of Table 8 which contains all objects in our overspill sample which have $\rm\mu >0.5''$/yr. The full version of Table 8 along with a citation key is available from CDS.

3.2.3 Potential nearby stars
In order to identify nearby objects within the sample, we estimated the distances of objects using the RECONS colour-absolute magnitude relations (Henry et al. 2004). These relations allowed us to calculate absolute $K_{\rm S}$ magnitudes of a star from the I-J, I-H and $I-K_{\rm S}$ colours. Each of these estimates could then be combined with the apparent magnitude to yield a distance modulus (and hence distance) for each star. However we needed to gain a clear picture of the potential errors in such a calculation. Hence we used these relations to calculate the distance modulii for the sample of nearby stars compiled by Reid using photometry from Leggett (1992) and Bessell (1991) (with an additional simulated error on the I band data to mimic the less accurate plate photometry) and compared them with those distance modulii calculated from their trigonometric parallaxes. We found that these three distance relations when combined produced distance modulii that were 0.3 mag too close, with each having an error of one magnitude. This error is then combined with the random error from each measurement to produce an estimate for the total error on the distance (after the 0.3 mag offset was removed) of each star.

The calculated distance estimates are shown in Table 5 for the sample with 0.1''/yr < $\mu <0.5''$/yr and in Table 6 for those objects with proper motions above 0.5''/yr. In total there are 12 stars with distance estimates closer than 20 pc which have not been found before.

3.3 Completeness

The final aim of this survey is to study the local space density, mass function and birthrate of cool dwarfs. If we are to properly examine these we must first estimate the completeness of the survey. Firstly there is the problem of crowded regions. These were excluded by utilising the proximity flag in the 2MASS data files. There will also be problems relating to the UKST images. The SuperCOSMOS software flags objects which are blended with other objects. These have been excluded along with those falling near bright stars. The incompletenesses caused by these three effects have been examined and are quantified in Sect. 5 of DHC.

There will also be incompleteness caused by both the limiting magnitudes of the survey and the short epoch separation on some plates. To illustrate this a histogram of the cumulative number of objects with proper motions greater than the minimum proper motion in each bin is plotted in Fig. 4. If the survey was totally complete it would be expected that the number of objects in each bin would scale as $\mu^{-3}$ (see the solid line). However it is clear that this is not the case and that the incompleteness is significant below 0.2''/yr.

In order to gain information on the mass function and birthrate of cool dwarfs all the sources of incompleteness mentioned in this section must be taken into account. In the third paper of this series we will detail the simulations which use both the crowding incompleteness calculated in DHC and the other selection effect to produce simulated samples. These will then used to constrain underlying distributions such as the birthrate and mass function.

4 Conclusions

Here we have presented a large sample of cool low mass objects. Within this sample we have identified 45 common proper motion systems of which 11 had neither candidate previously identified. In addition 38 objects (12 of them new) which may lie within 20 pc have been found, along with 4 new potential L dwarfs. It has been shown that the surevy is incomplete and that the sources of incompleteness must be taken into account before examining the constraints that can be set on the mass function and stellar birthrate.

In the next paper in this series we will model the low mass stellar population. This will allow us to take into account all the selection criteria and their effects, along with the various sources of incompleteness. This will allow us to constrain the underlying mass function and birthrate.

Acknowledgements
The authors would like to thank Sue Tritton and Mike Read for their help in selecting plates, Harvey MacGillivray and Eve Thomson for their scanning of the material on SuperCOSMOS and Todd Henry and John Cooke for their helpful discussions. This publication makes use of data products from the Two Micron All Sky Survey, which is a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center/California Institute of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation. SuperCOSMOS wass funded by a grant from the UK Particle Physics and Astronomy Research Council. This publication makes use of the SLALIB positional astronomy library (Wallace 1998).

References

 
Copyright ESO 2007