1. E. Solano¹, N. Lodieu², M. Aberasturi¹, M.R. Zapatero-Osorio¹, E. Martín¹ 1.- Centro de Astrobiología (INTA-CSIC), Madrid, Spain 2.- Instituto de Astrofísica de Canarias, Tenerife, Spain Discovery of ultracool subdwarfs using Virtual Observatory tools

2. Ultracool subdwarfs: properties Metal-poor dwarfs with spectral type later than M7. Metal-poor dwarfs with spectral type later than M7. Less luminous (actually hotter) than their solar-metallicity counterparts due to the dearth of metals in their atmospheres. Less luminous (actually hotter) than their solar-metallicity counterparts due to the dearth of metals in their atmospheres. Metal-poor dwarfs with spectral type later than M7. Metal-poor dwarfs with spectral type later than M7. Less luminous (actually hotter) than their solar-metallicity counterparts due to the dearth of metals in their atmospheres. Less luminous (actually hotter) than their solar-metallicity counterparts due to the dearth of metals in their atmospheres.

3. Ultracool subdwarfs: Identification Halo regions Halo regions No contamination by disk stars. No contamination by disk stars. Low luminosity. Low luminosity. Trigonometrical parallaxes and spectra unobtainable. Trigonometrical parallaxes and spectra unobtainable. Solar neighbourhood Solar neighbourhood Low luminosity. Low luminosity. Hard to find because of their local scarcity (4 out of 348 at d<10 pc, Henry et al. 2006, AJ, 132, 2360). Hard to find because of their local scarcity (4 out of 348 at d<10 pc, Henry et al. 2006, AJ, 132, 2360). Halo regions Halo regions No contamination by disk stars. No contamination by disk stars. Low luminosity. Low luminosity. Trigonometrical parallaxes and spectra unobtainable. Trigonometrical parallaxes and spectra unobtainable. Solar neighbourhood Solar neighbourhood Low luminosity. Low luminosity. Hard to find because of their local scarcity (4 out of 348 at d<10 pc, Henry et al. 2006, AJ, 132, 2360). Hard to find because of their local scarcity (4 out of 348 at d<10 pc, Henry et al. 2006, AJ, 132, 2360). Henry et al.

4. Ultracool subdwarfs: Identification Key parameter: Proper motion Key parameter: Proper motion The high space velocities of subdwarfs allow them to distinguish from the multitude of similarly faint but slowly moving disk stars. The high space velocities of subdwarfs allow them to distinguish from the multitude of similarly faint but slowly moving disk stars. Key parameter: Proper motion Key parameter: Proper motion The high space velocities of subdwarfs allow them to distinguish from the multitude of similarly faint but slowly moving disk stars. The high space velocities of subdwarfs allow them to distinguish from the multitude of similarly faint but slowly moving disk stars.

5. Ultracool subdwarfs: Identification (II) Drawback Drawback Identification typically requires the combination of attributes (colors, proper motions,...) available from different archives of large-area surveys (e.g. UKIDSS, SDSS, 2MASS, DENIS,...). Identification typically requires the combination of attributes (colors, proper motions,...) available from different archives of large-area surveys (e.g. UKIDSS, SDSS, 2MASS, DENIS,...). Very time-consuming task if performed by hand but, on the contrary, an approach that perfectly fits in the framework of the Virtual Observatory. Very time-consuming task if performed by hand but, on the contrary, an approach that perfectly fits in the framework of the Virtual Observatory. Drawback Drawback Identification typically requires the combination of attributes (colors, proper motions,...) available from different archives of large-area surveys (e.g. UKIDSS, SDSS, 2MASS, DENIS,...). Identification typically requires the combination of attributes (colors, proper motions,...) available from different archives of large-area surveys (e.g. UKIDSS, SDSS, 2MASS, DENIS,...). Very time-consuming task if performed by hand but, on the contrary, an approach that perfectly fits in the framework of the Virtual Observatory. Very time-consuming task if performed by hand but, on the contrary, an approach that perfectly fits in the framework of the Virtual Observatory.

6. The Virtual Observatory (VO) VO is a consolidated initiative that represents the framework where to develop archive-based research. VO is a consolidated initiative that represents the framework where to develop archive-based research. By ensuring easy and transparent access to astronomical resources the Virtual Observatory is fostering new and/or more efficient science. By ensuring easy and transparent access to astronomical resources the Virtual Observatory is fostering new and/or more efficient science. VO is a consolidated initiative that represents the framework where to develop archive-based research. VO is a consolidated initiative that represents the framework where to develop archive-based research. By ensuring easy and transparent access to astronomical resources the Virtual Observatory is fostering new and/or more efficient science. By ensuring easy and transparent access to astronomical resources the Virtual Observatory is fostering new and/or more efficient science. Std s Tools Science

7. This work: Objectives Identification of a statistically significant number of bona-fide ultracool subdwarfs to: Bridge the gap between M and L spectral types. Help the discovery of cooler Population II brown dwarfs. Establish optical and near-infrared spectral classification at cooler temperatures and lower metallicities. Investigate the role of metallicity in the atmospheres of ultracool dwarfs. Shed light on the star formation history of the Milky Way. Good candidates for exoplanet searches (low chromospheric activity). Identification of a statistically significant number of bona-fide ultracool subdwarfs to: Bridge the gap between M and L spectral types. Help the discovery of cooler Population II brown dwarfs. Establish optical and near-infrared spectral classification at cooler temperatures and lower metallicities. Investigate the role of metallicity in the atmospheres of ultracool dwarfs. Shed light on the star formation history of the Milky Way. Good candidates for exoplanet searches (low chromospheric activity).

8. This work: Methodology Proper motion and photometric searches using VO tools (ALADIN, TOPCAT). 8500 (SDSS-2MASS) + 1500 (SDSS-UKIDSS) square degrees explored. Proper motion and photometric searches using VO tools (ALADIN, TOPCAT). 8500 (SDSS-2MASS) + 1500 (SDSS-UKIDSS) square degrees explored.

9. This work: Methodology (II) Workflow: All point sources in SDSS without 2MASS/UKIDSS counterparts within 5”. Colours and reduced proper motion based on the M7-M9.5 subdwarfs identified by Lépine & Scholz (2008, AJ, 681, L33). (J-K) ≤ 0.7 mag Workflow: All point sources in SDSS without 2MASS/UKIDSS counterparts within 5”. Colours and reduced proper motion based on the M7-M9.5 subdwarfs identified by Lépine & Scholz (2008, AJ, 681, L33). (J-K) ≤ 0.7 mag

10. This work: Results SDSS (Data Release 7) - 2MASS (Point Source Catalogue)‏ 356 candidates after the initial search Further refinement - Whenever possible, UKIDSS and SuperCosmos were used for a third epoch. - Visual inspection to discard false detections. Final number of candidates: 30. Six were already known. Optical spectroscopic follow-up of new candidates: - 3 + 1 + 2 = 6 candidates observed with NOT/ALFOSC in January, July and August 2009. - 6 out of 6 exhibit spectral features typical of subdwarfs. SDSS (Data Release 7) - 2MASS (Point Source Catalogue)‏ 356 candidates after the initial search Further refinement - Whenever possible, UKIDSS and SuperCosmos were used for a third epoch. - Visual inspection to discard false detections. Final number of candidates: 30. Six were already known. Optical spectroscopic follow-up of new candidates: - 3 + 1 + 2 = 6 candidates observed with NOT/ALFOSC in January, July and August 2009. - 6 out of 6 exhibit spectral features typical of subdwarfs.

11. This work: Results (II) SDSS (Data Release 7) – UKIDSS /LAS (Data Release 5)‏ Aim: Detection of cooler (down to 1000K) and less massive (up to 60 Jupiter masses) subdwarfs in a radius of 25 pc assuming [Fe/H] = -1. 31 new candidates (after removal of false detections and known subdwarfs) Optical spectroscopic follow-up:  GTC / OSIRIS  11 hours awarded in semester 09A. - 8 objects observed but S/N lower than expected. Very hard to confirm/discard candidates.  16 hours awarded in 09B  19 objects sent to the queue.  VLT / FORS2  20 hours  22 objects sent to the queue for the next semester. SDSS (Data Release 7) – UKIDSS /LAS (Data Release 5)‏ Aim: Detection of cooler (down to 1000K) and less massive (up to 60 Jupiter masses) subdwarfs in a radius of 25 pc assuming [Fe/H] = -1. 31 new candidates (after removal of false detections and known subdwarfs) Optical spectroscopic follow-up:  GTC / OSIRIS  11 hours awarded in semester 09A. - 8 objects observed but S/N lower than expected. Very hard to confirm/discard candidates.  16 hours awarded in 09B  19 objects sent to the queue.  VLT / FORS2  20 hours  22 objects sent to the queue for the next semester.

12. Summary The census of ultracool subdwarfs is largely incomplete.  39 out of which only three show L-types. ‏ We have developed a methodology to identify ultracool subdwarfs using Virtual Observatory tools. Six confirmed subdwarfs out of six observed spectroscopically. Spectral types: M5-M6 (only the brightest targets – SDSS / 2MASS - have been analysed so far). If the success rate remains so for fainter objects our procedure will double the number of subdwarfs with spectral types later than M7. This methodology opens new prospects to further increase these numbers using upcoming large-scale surveys (VISTA, PanSTARRS,…). The census of ultracool subdwarfs is largely incomplete.  39 out of which only three show L-types. ‏ We have developed a methodology to identify ultracool subdwarfs using Virtual Observatory tools. Six confirmed subdwarfs out of six observed spectroscopically. Spectral types: M5-M6 (only the brightest targets – SDSS / 2MASS - have been analysed so far). If the success rate remains so for fainter objects our procedure will double the number of subdwarfs with spectral types later than M7. This methodology opens new prospects to further increase these numbers using upcoming large-scale surveys (VISTA, PanSTARRS,…).