Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://star.arm.ac.uk/preprints/2011/578.pdf Äàòà èçìåíåíèÿ: Mon Jun 20 13:25:31 2011 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 02:58:48 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: http www.badastronomy.com bad tv foxapollo.html

Mon. Not. R. Astron. Soc. 000, 000­000 (0000)

Printed 16 June 2011

A (MN L TEX style file v2.2)

New short p eriod stellar pulsators at large Galactocentric distances
Gavin Ramsay1, Ralf Napiwotzki2, Thomas Barclay1,3, Pasi Hakala4, Stephen Potter5, Mark Cropper3 1
2 3 4 5

Armagh Observatory, Col lege Hil l, Armagh, BT61 9DG Centre for Astrophysics Research, STRI, University of Hertfordshire, Hatfield, AL10 9AB Mul lard Space Science Laboratory, University Col lege London, Holmbury St. Mary, Dorking, Surrey, RH5 6NT ¨ Finnish Centre for Astronomy with ESO, University of Turku, V¨is¨l ¨ntie 20, FI-21500 PIIKKIO, Finland aaa South African Astronomical Observatory, P.O. Box 9, Observatory 7935, Cape Town, South Africa

Accepted 2011 June 16. Received 2011 June 15; in original form 2011 April 25

ABSTRACT

We report the discovery of 31 blue, short period, pulsators made using data taken as part of the Rapid Temporal Survey (RATS). We find they have periods between 51­83 mins and full-amplitudes between 0.05­0.65 mag. Using the period-luminosity relationship for short period pulsating stars we determine their distance. Assuming they are pulsating in either the fundamental or first over-tone radial mode the majority are located at a distance greater than 3kpc, with several being more than 20 kpc distant. Most stars are at least 1 kpc from the Galactic plane, with three being more than 10 kpc. One is located in the direction of the Galactic anti-center and has Galactocentric distance of 30 kpc and is 20 kpc below the plane: they are therefore potential tracers of Galactic structure. We have obtained low-resolution spectra for a small number our targets and find they have temperatures between 7200­7900K and a metal content less than Solar. The colours of the pulsators and the spectral fits to those stars for which we have spectra indicate that they are either SX Phe or Scuti stars. We estimate the number of SX Phe stars in our Galaxy and find significantly fewer per unit mass than reported in massive globular clusters or dwarf spheroidal galaxies. Key words: stars: surveys ­ oscillations ­ stars: variables (SX Phe stars, Scuti stars) ­ stars: evolution

1

INTRODUCTION

Pulsations have b een detected from stars on a wide range of timescales, from several tens of seconds in the case of white dwarfs to several years in the case of red-giant stars. These pulsations manifest themselves through a p eriodic variation of the stellar brightness. Pulsating stars can also b e found over a wide range of parameter space (temp erature, luminosity) in the HR diagram (eg Jeffery 2008). A detailed study of the photometric variability of individual systems can give insight to the physical conditions deep inside the star (eg Kurtz 2004). In recent years many photometric surveys have b een undertaken, leading to a corresp onding increase in the numb er of known stellar pulsators. Factors such as cadence, depth, sky coverage and duration makes any individual survey more (or less) likely to discover sp ecific typ es of stellar pulsator.
c 0000 RAS

The Rapid Temp oral Survey (RATS) is a deep, high-cadence photometric survey covering nearly 40 square degrees which took place b etween 2003 and 2010 (Ramsay & Hakala 2005, Barclay et al 2011). This strategy allows us to detect sources which vary in their intensity on a timescale of a few minutes to several hours. In our first set of wide-field camera data taken in 2003, we identified a small numb er of blue stars which pulsate on a p eriod b etween 40­70 mins. An analysis of their optical sp ectra indicated they were SX Phe stars or Sct stars (Ramsay et al 2006). SX Phe stars are old, metal-p oor stars which are likely to b e halo ob jects (see Nemec & Mateo 1990 for a review). The Sct stars show similar characteristics to the SX Phe stars but have solar metallicities and more likely to b e located in the thin disk (see Breger 2000 for a review). Although Sct-like pulsations have b een de-


2
tion with other typ es of sources at low Galactic latitudes we allow a maximum extinction corresp onding to AV =0.40 (EB-V =0.13 for R=3.1). For blue stars EB-V =0.13 equates to Eg-r =0.13. If we add a conservative uncertainty of 0.13 mag in our observed colours (Barclay et al 2011) our search region therefore covers -0.12 < (g - r ) < 0.40, while the brightness of stars were in the range 15 < g < 23. Here we note that the shap e of the light curves and the colour of our sources is similar to some cataclysmic variables (CVs; eg Szkody et al 2002 who present the first sample of CVs discovered using SDSS data). The hydrogen accreting a CVs have a minimum orbital p eriod of 80 mins (G¨nsicke et al 2009), implying a p ossible over-lap with our target selection. (The helium dominated CVs with orbital p eriods in the range 40­70 mins do not show a photometric modulation implying we are not overlapping with these ob jects). However, CVs show optical sp ectra dominated by line emission ­ although we obtained a sp ectrum for only 7 out of the 31 new sources ­ none show evidence for emission lines (cf §5). Moreover, although pulsating blue stars are not X-ray sources, CVs are weak to moderately strong X-ray sources (Verbunt et al 1997). We therefore cross-correlated the p osition of our sources with that of catalogues derived from the Rosat All-Sky X-ray Survey (RASS). None of our sources has an X-ray counterpart within 20 arcsec of the optical p osition. In contrast 30 of the 48 CVs with orbital p eriods in the range 70­120 mins in the 2009 catalogue of Ritter & Kolb (2003) were detected in the RASS. Although at this stage we cannot preclude that none of our 31 sources are CVs we consider this rather unlikely. For sources which fell within our search range we manually insp ected each light curve to verify that the light curve was consistent with that of a stellar pulsator (cf Rodr´guez i et al 2007 for a recent example of a light curve for an SX Phe star) and to exclude light curves of low quality. We found 31 sources which showed a modulation in their lightcurve on a p eriod b etween 51­83 mins and had a mean brightness of g = 15.9 - 20.8. The full-amplitude of the modulation is in the range 0.05­0.65. We show their photometric prop erties in Table 1 and the light curves in Figure 3 and 4.

Figure 1. The colour-magnitude diagram of all stars in the RATS archive which are located in field with extinction less than AV 0.4. The blue pulsators presented in this paper are shown in Table 1 are shown as crosses.

tected in pre-main sequence stars with p eriods as short as 18 min (Amado et al 2004), approximately 90 p ercent of Sct stars have a pulsation p eriod in the range 40 min to 5.3 hrs (cf Table 1, Rodr´guez et al 2000). SX Phe and Scu i stars have a well defined p eriod-luminosity relationship and can therefore b e used as distance indicators and hence map Galactic structure (eg Nemec, Linnell Nemec & Lutz 1993). Since we took our first set of data in 2003, we have obtained a significant amount of further data (Barclay et al 2011). We therefore have made a systematic search for blue, short p eriod, pulsating stars. Our light curves are typically 2­2.5 hrs in duration, so the longest p eriod we can determine with confidence is less than 2 hrs. For stars with p eriods shorter than 40 mins, it b ecomes increasingly difficult to determine the nature of the source based on colour and p eriod information (cf Table 1 and 2, Barclay et al 2011). In this pap er, we therefore have decided to restrict our search for blue pulsating variables in the range of 40 min to 2 hrs.

3 2 OBSERVATIONS The RATS observing strategy is to take a series of 30 sec exp osures of a given field using the wide-field cameras on the Isaac Newton Telescop e in La Palma and the MPG/ESO 2.2m on La Silla for a duration of 2 hrs. To date our survey has discovered around 1.2â105 variable stars (see Barclay et al 2011 for a full description of our reduction process). Based on their photometric prop erties, a small sub-sample has b een selected for followup sp ectroscopic observations to determine their nature. To narrow our search for blue pulsators in our RATS data, we restricted our search to a range in b oth magnitude and colour. The intrinsic colour of SX Phe stars is typically (B - V )o 0.1 to 0.35 (eg Poretti et al 2008), which corresp onds to (g - r )o =­0.12 to 0.14. This is virtually identical to the colours for Sct stars (eg Rodr´guez i et al 2000). Many of our fields lie at low Galactic latitude and hence have high extinction. To reduce the contamina-

DISTANCES

There is a clear relationship b etween MV and pulsation p eriod which is applicable to SX Phe, Sct and RR Lyr stars (eg McNamara 1997). This Period-Luminosity (PL) relationship is consistent with a study of different typ es of short p eriod pulsators in the Fornax dwarf spheroidal galaxy (Poretti et al 2008). Since the PL relationship of McNamara (1997) is calibrated with resp ect to MV , we applied a small correction to transform our g band magnitudes to that of the V band (Jester et al 2005). Further, we used a NASA/IPAC tool1 which uses the maps of Schlegel, Finkb einer & Davis (1998) to determine extinction to the edge of the Galaxy. The PL relationship assumes that the p eriod is the fundamental radial pulsation mode rather than the first overtone which can also b e observed in these stars. Given the short duration of our light curves, it is difficult to assess

1

http://irsa.ipac.caltech.edu/applications/DUST c 0000 RAS, MNRAS 000, 000­000


Stel lar Pulsators at large distances
Short ID J000114 J000134 J000147 J030556 J050351 J065521 J065541 J120232 J120709 J120902 J135646 J135912 J155955 J160103 J175816 J175836 J175932 J180331 J180416 J181727 J181736 J181753 J181816 J182250 J182347 J195235 J200210 J220915 J230507 J233907 J234635 (J2000) 00: 00: 00: 03: 05: 06: 06: 12: 12: 12: 13: 13: 15: 16: 17: 17: 17: 18: 18: 18: 18: 18: 18: 18: 18: 19: 20: 22: 23: 23: 23: 01: 01: 01: 05: 03: 55: 55: 02: 07: 09: 56: 59: 59: 01: 58: 58: 59: 03: 04: 17: 17: 17: 18: 22: 23: 52: 02: 09: 05: 39: 46: 14. 34. 47. 56. 51. 21. 41. 32. 09. 02. 46. 12. 55. 03. 16. 36. 32. 31. 16. 27. 36. 53. 16. 50. 47. 35. 10. 15. 07. 07. 35. 7 8 5 8 4 2 8 2 2 4 3 9 4 5 2 5 1 0 2 9 2 2 0 6 7 3 0 4 9 3 6 (J2000) +53:43:05.4 +53:51:42.2 +53:23:18.7 ­00:36:16.2 +35:08:02.1 +10:41:58.9 +10:44:21.5 ­24:29:17.4 ­22:54:49.5 ­23:11:39.0 +22:54:40.4 +23:36:55.3 ­25:43:20.2 ­25:42:44.6 +28:17:52.8 +28:09:13.0 +01:19:40.4 +02:08:40.2 +02:08:32.4 +06:34:01.0 +06:24:26.0 +06:31:49.5 +07:30:43.0 +07:54:36.8 +07:53:45.5 +18:43:54.8 +18:43:07.3 +55:34:38.5 +34:17:23.4 +57:08:02.3 +56:23:23.7 l (J2000) 115.46 115.54 115.48 179.16 169.90 203.82 203.82 288.96 289.81 290.43 20.61 23.76 347.69 347.89 53.97 53.85 28.14 29.34 29.43 34.97 34.84 34.98 35.92 36.79 36.88 56.56 57.72 101.37 99.14 113.21 114.00 b (J2000) -8.42 -8.29 -8.76 -48.23 -3.83 5.72 5.81 37.05 38.82 38.65 74.62 74.30 20.35 20.17 23.35 23.23 12.14 11.63 11.47 10.53 10.42 10.41 10.77 9.93 9.71 -4.34 -6.29 -0.35 -23.62 -4.36 -5.35 g g-r Period (mins) 64.3 73.5 69.6 82.9 62.9 60.9 56.0 55.7 74.7 72.3 53.8 63.9 60.8 58.7 54.4 66.1 74.7 72.5 82.9 76.9 60.8 59.6 53.0 64.5 61.2 63.2 74.5 51.6 75.3 67.6 68.6 Am p (mag) 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 50 07 45 28 14 13 10 33 24 17 13 07 08 08 12 07 50 65 65 10 06 08 50 13 26 05 07 07 21 15 18 d, z (kpc) FM 6.0, -0. 4.2, -0. 3.4, -0. 30.0,-22. 3.4, -0. 2.7, 0. 11.6, 1. 21.2, 12. 8.2, 5. 5.8, 3. 9.0, 8. 8.7, 8. 5.2, 1. 5.7, 2. 12.7, 5. 9.2, 3. 6.5, 1. 7.0, 1. 8.2, 1. 8.6, 1. 8.2, 1. 7.4, 1. 7.1, 1. 10.8, 1. 15.7, 2. 4.9, -0. 5.1, -0. 0.6, 0. 32.1,-12. 4.1, -0. 1.8, -0. 9 6 5 4 2 3 2 8 2 6 7 3 8 0 0 6 4 4 6 6 5 3 3 9 6 4 6 0 8 3 2 d, z (kpc) FO 5.0, -0.7 3.5, -0.5 2.8, -0.4 24.8, -18.5 2.8, -0.2 2.2, 0.2 9.6, 1.0 17.5, 10.6 6.8, 4.3 4.8, 3.0 7.4, 7.2 7.2, 6.9 4.3, 1.5 4.7, 1.6 10.5, 4.1 7.6, 3.0 5.4, 1.1 5.8, 1.2 6.8, 1.4 7.1, 1.3 6.8, 1.2 6.1, 1.1 5.9, 1.1 8.9, 1.5 13.0, 2.2 4.1, -0.3 4.2, -0.5 0.5, 0.0 26.5, -10.6 3.4, -0.3 1.5, -0.1 Spec?

3
Low AV ?

17. 17. 16. 20. 18. 15. 19. 20. 17. 17. 18. 17. 17. 17. 19. 18. 18. 17. 18. 18. 18. 18. 18. 19. 19. 18. 18. 15. 20. 18. 15.

8 0 8 3 4 9 3 2 8 2 2 9 3 6 1 1 2 8 1 3 6 5 5 1 9 0 2 8 8 4 8

-0.12 0.19 0.35 -0.08 0.30 0.12 0.30 0.04 0.30 0.39 -0.09 0.03 0.37 0.36 0.04 0.05 0.36 -0.09 0.23 0.22 0.22 0.34 0.24 0.31 0.24 0.38 0.39 0.03 0.25 0.18 0.38

SDSS

Y

SAAO NOT NOT SAAO WHT WHT

Y Y Y Y Y Y Y Y Y

Y

Table 1. Candidate blue pulsating stars identified in RATS data. We show: the stars `short' ID; equatorial and galactic co-ordinates; g magnitude and g - r colour; period and full-amplitude of modulation. We also show the distance from the Sun, d, and the height above the Galactic plane, z , (where FM implies we assume the period we detect is the fundamental period and FO imples the period is the first overtone) and whether we have a spectrum and if so its origin (§5). The last column indicates whether it is located in a field with low extinction ­ they have been used in estimating the Galactic population of SX Phe stars (§7).

whether the p eriod we detect is either the fundamental or first over-tone p eriod (or whether the p eriod is even due to a radial mode). Some help is found from the fact that the p eriod of the first over-tone is less than the p eriod of the fundamental p eriod by a factor of 0.775 (Poretti et al 2005). We determined the distance to each source assuming the p eriod we detect was the fundamental radial mode and also by assuming the p eriod was the first over-tone (we show the distance and corresp onding height from the Galactic plane to each source in Table 1 under b oth assumptions). The error on the distance assuming we do not know the pulsation mode of the star is 17 p ercent. As a zeroth order test, we show in Figure 2 the relationship b etween the dereddened V mag (assuming the extinction to the edge of the Galaxy as determined ab ove) and the derived distance assuming the p eriod is the fundamental mode and also the first over-tone. Whilst one can argue that any individual ob ject may give a b etter overall linear relationship if one assumes the p eriod is one or the other mode, it gives us confidence that our distances are not grossly in error.
c 0000 RAS, MNRAS 000, 000­000

Taking into account the uncertainties in our photometry (g 0.1 for stars brighter than g =20, increasing to g 0.2 for fainter sources), the uncertainty on our p eriod determinations, coupled with the uncertainty on the pulsation mode, we estimate the errors on our distances mayb e up to 25 p ercent. If on the other hand the p eriods we determine are half the true p eriod then we significantly underestimate their distances. Similarly, if the extinction is less than that to the edge of the Galaxy then we also underestimate the distances. Of course, if the p eriod we detect is not due to radial pulsation then the distance is highly uncertain.

Our sources have a large spread of distances (Table 1). Assuming the p eriod is due to the fundamental radial pulsation mode then the closest star is 0.6 kp c, while the most distant lies at 32 kp c (the median is 7.0 kp c). Similarly, the sample shows a large spread in height from the Galactic plane, the least distant only 200 p c, with the most distant at 20 kp c. The median height is 1.4 kp c, which equates to twice the scale height of the thick disc of the Galaxy (eg de Jong et al 2010). On the other hand, if the p eriods are due


4
in Sutherland, South Africa and two using the 2.5m Nordic Optical Telescop e and ALFOSC on La Palma. A sp ectrum of another source was obtained from the SDSS data archive2 . Both arms of ISIS were used giving sp ectral coverage from 3800­5200 ° and 5500­9000 ° in the blue (R158B A A grating) and red (R158R) arms resp ectively. Grating #7 was used with the SAAO sp ectrograph giving a wavelength range ° of 3400­7500°. The sp ectral resolution was 5A for the A WHT and SAAO sp ectra. Grating #7 was used with the ALFOSC imaging sp ectrograph giving a wavelength range 3800­7000° and a sp ectral resolution of 8°. All the data A A were reduced using optimal extraction and standard techniques. Several sp ectra of J0305 were downloaded from the SDSS archive and co-added and re-binned into 4 ° bins. A We modelled the sp ectra using a grid of LTE models calculated with the atlas9 code (Kurucz 1992) with convective overshooting switched off. Sp ectra were calculated with the linfor line-formation code (Lemke 1991). Data for atomic and molecular transitions were compiled from the Kurucz line list. The sp ectra were fitted with the fitsb2 routine (Napiwotzki et al 2004). The error limits of all fit parameters were determined with a b ootstrapping method. The stellar temp eratures were estimated from the hydrogen Balmer lines of the stars (H to H). No gravity sensitive features are accessible in our low resolution sp ectra so gravity was fixed at log g = 4.0, which is a typical value for SX Phe and large amplitude Sct stars (eg McNamara 1997). Apart from the Ca H and K lines, the resolution of our sp ectra is clearly too low to allow a meaningful fitting of individual metal lines for abundance determinations. However, the sp ectra allow a determination of an overall abundance value. The general metallicity of the models was varied until an optimum fit was achieved. Since the extinction towards these high Galactic latitude targets is low we do not exp ect the Ca H and K lines to b e significantly contaminated by interstellar absorption. However, by including these lines we obtain an upp er limit to the metallicity. The sp ectral ranges used for determining the metallicity contain a mix of sp ectral features, but the dominant sp ecies is Fe I. We thus exp ect our metallicity [Met/H] to b e an approximate indicator of iron abundance. We show the results of our fits in Table 2 and the b est fits to the sp ectra in Figure 6. The b est-fit temp erature of our sources are in the range Teff =7200­7900K. Although the metal abundance is less well constrained, each source has a metallicity less than solar at the 1 level and for all but one source this is also true at the 3 level. Fixing log g at 3.5 and 4.5 rather than 4.0 changes the temp erature by less than 90K and log g by less than 0.1 dex.

10 Distance (kpc)

1

Fundamental period First Over-tone

12

14

16 v (o)

18

20

Figure 2. The relationship between the dereddended V mag and the distance determined assuming the measured period is the fundamental mode or the first over-tone. It suggests our distance determinations are not grossly in error.

to the first over-tone radial mode then these distances are less by 20 p ercent. The two sources with the greatest distances (J030556 and J230507 at 30 kp c) have a Galactic longitude which app ears to place them at a distance much further than the accepted limits of the spiral structure of the Milky Way (eg Churchwell et al 2009). Secondly, three sources (J030556, J120232, J230507) lie more than 10 kp c distant from the Galactic plane: this places them deep into the Galactic halo.

4

FOLLOWUP PHOTOMETRY

To confirm the p eriod of J0305, we obtained followup photometry of this source on 31 Dec 2010 using the Nordic Optical Telescop e and ALFOSC. We used white light and an exp osure time of 10 sec. The resulting light curve, binned into 120 sec bins, covers 3.8 hrs (Figure 5). A clear modulation is present in the light curve. A Lomb Scargle p ower sp ectrum of the light curve indicates a p eriod of 90 min. This compares with 83 min (Table 1) which was derived using the original INT light curve. Given the uncertainties in the p eriod derived using each data-set, the p eriods are consistent. For completeness, we note that using a p eriod of 90 min rather than 82.9 min (Table 1) places J0305 at a distance of 31.9 kp c rather than 30.0 kp c assuming we have identified this p eriod as the fundamental radial pulsation mode. We encourage additional photometry of all the sources shown in Table 1 to identify the mode of the pulsation seen in each star.

6 5 SPECTRAL OBSERVATIONS

THE NATURE OF THE VARIABLE SOURCES

We have sp ectroscopic observations of a small sample of our candidate blue pulsating stars (cf Table 1). We obtained sp ectra for two sources using the 4.2m William Herschel Telescop e (WHT) and the Intermediate disp ersion Sp ectrograph and Imaging System (ISIS) on La Palma, two using the SAAO 1.9m telescop e and the Cassegrain sp ectrograph

We have presented evidence that the ma jority of sources describ ed in this pap er are blue compact stellar pulsators. To place our sources in the general context of blue stellar pulsators, we used the most recent edition of the General

2

http://www.sdss3.org/dr8 c 0000 RAS, MNRAS 000, 000­000


Stel lar Pulsators at large distances
-0.1 Delta Mag

5

000134+535142

050351+350802

065521+104158

0.0

0.1 -0.1 Delta Mag

065541+104421

120902-231139

135912+233655

0.0

0.1 -0.1 Delta Mag

155955-254320

160103-254244

175816+281752

0.0

0.1 -0.1 Delta Mag

175836+280913

181727+063401

181736+062426

0.0

0.1 -0.1 Delta Mag

181753+063149

182250+075436

195235+184354

0.0

0.1 -0.1 Delta Mag

200210+184307

220915+553438

233907+570802

0.0

0.1 -0.1 Delta Mag

234635+562323

0

40 80 120 Time (mins)

160 0

40 80 120 Time (mins)

160

0.0

0.1 0 40 80 120 Time (min) 160
Figure 3. The light curves of our blue candidate stellar pulsators identified in our survey and which had an amplitude less than 0.2 mag.

Catalogue of Variable Stars (GCVS, Samus et al 2009) to determine the distribution of p eriods in different classes of short p eriod pulsating variables. SX Phe and Sct stars and the longer p eriod pulsating sdB stars all have a distribution of pulsation p eriods which include p eriods less than 60 mins. In contrast, RR Lyr and classical Cepheids have much longer pulsation p eriods. The Cepheid stars have a minimum

p eriod of 2 hrs. The GCVS notes 131 field Sct stars and 20 SX Phe stars. SX Phe stars and Sct stars have temp eratures typically 7200­7900K (implying late A/early F sp ectral typ es), while the long p eriod sdB stars have temp eratures typically 25000­30000K (Green et al 2003). Our analysis of those sources for which we have sp ectra result in parameters which confirm a SX Phe/ Scuti nature (§5). The observed colours

c 0000 RAS, MNRAS 000, 000­000


6

-0.3 Delta Mag 0.0 0.3 -0.3 Delta Mag 0.0 0.3 -0.3 Delta Mag 0.0 0.3 -0.5 Delta Mag

000147+532318

030556-003616

120232-242917

120709-225449

175932+011940

181816+073043

182347+075345

230507+341723

000114+534305

180331+020840

180416+020832

0.0

0.5 0 40 80 120 Time (min) 160 0 40 80 120 Time (min) 160 0 40 80 120 Time (min) 160

Figure 4. As for Fig 3 but for those stars which had an amplitude greater than 0.2 mag.

Source
-0.2

Temp (K) 7210+270 -230 7310+480 -350 7750+270 -240 7210+200 -290 7410+310 -360 7830+680 -400 7890+280 -230

Z (M ) 0. 2. 1. 2. 1. 1. 1. 5+0.4 -2.3 1+1.8 -3.3 1+0.6 -3.5 1+1.0 -2.1 1+1.1 -4.0 0+0.9 -4.7 3+1.1 -3.6

-0.1 Delta Mag

0.0

J030556 J120709 J135646 J135912 J160103 J175816 J175836

0.1

0.2

Table 2. The temperature and metallicity for seven of our sources derived from model fits to their optical spectra. The errors refer to the 3 confidence interval.
0 50 100 Time (mins) 150 200

Figure 5. The light curve of RAT J030556-003616 made using the NOT and ALFOSC in Dec 2010. The data has been binned into 120 sec bins.

are also consistent with this classification for all pulsators, although for fields with high extinction we cannot rule out some contamination from intrinsically redder sources. However, given that the amplitude of the long p eriod sdB stars is very low (eg Fontaine et al 2003) we consider it highly unlikely that blue pulsators are sdB stars.

The light curves of our sources (Figure 2 and 3) app ear to show regular pulsations p eriods, some displaying high (60 p ercent) full-amplitude modulations, while others are much lower (a few p ercent). They are similar to the light curves of SX Phe and Sct stars which app ear in the literature. We have low resolution sp ectra for a small numb er of our sources which were taken for identification purp oses: they are consistent with A/F typ e stars. Although our sp ectra are not high resolution, our model fits indicate that all sources have metallicities which are less than Solar (for most sources at the 3 confidence level). Higher resolution sp ectra
c 0000 RAS, MNRAS 000, 000­000


Stel lar Pulsators at large distances

7

Figure 6. We show the spectra and best model fit for those sources for which spectral data was available.

with good signal-to-noise are necessary to determine their metal content with higher confidence. In section 3 we noted that the vast ma jority of our sources are located at distances greater than 3 kp c, with some b eing over 20 kp c distant (if we have detected either the fundamental or first over-tone mode). Similarly, many are at a height of 1 kp c or more from the Galactic plane, with several b eing at least 10 kp c from the plane. Some of our sources are therefore at the remote edge of our Galaxy. SX Phe stars have b een found at b oth large distances and well into the Galactic plane (eg Berstein, Knezek & Offutt 1995; Jeon, Kim and Nemec 2010). They are therefore, in principal, p otential tracers of Galactic structure such as streams or the remnants of mergers. We note that J1356 and J1359 lie around 6 from the large globular cluster M3 (which is 10.4 kp c from the Sun) which places them at a distance of 2 kp c from M3. Although tidal tails have b een detected at distances of several kp c from globular clusters (eg Odenkirchen et al 2003), no tidal tails have b een detected from M3 (Jordi & Greb el 2010). The other source of particular note is J0305 which is located in the direction of the Galactic anti-center and at a distance of 30kp c (implying a Galactocentric distance of 38 kp c) and a height of 21 kp c b elow the Galactic plane. Recent work shows that the stellar density of the Galaxy decreases sharply at Galactocentric distances greater than 25 kp c (eg Watkins et al 2009, Sesar et al, 2010) which may indicate that J0305 is associated with a sub-structure of the halo. Alternatively it may b e in the process of b eing ejected from our Galaxy. A more detailed radial velocity study is required to answer this question.
c 0000 RAS, MNRAS 000, 000­000

7

THE GALACTIC POPULATION OF SX PHE STARS

SX Phe stars have b een identified in globular clusters and nearby dwarf galaxies (eg Olech et al 2005, Poretti et al 2008). Based on the numb er of blue stellar pulsators which we have identified in our survey, we now make an estimate of the total numb er of SX Phe stars which are present in our Galaxy. Since our survey is biased towards low Galactic fields, highly reddened (but intrinsically blue) pulsators could b e confused with apparently much redder sources making p otential contamination a significant concern. For this reason, we therefore base our simulation on those fields for which the total extinction is less than AV =0.45. A total of 35 fields had a column density less than our limit which corresp onds to an area of 10 square degrees. We find eleven blue pulsators in these fields (these are flagged in the last column of Table 1). Low resolution sp ectroscopic data exist for seven of these eleven pulsators and a sp ectroscopic analysis indicates they have a metal content consistent with that of SX Phe stars. Here we assume that all eleven pulsators at high Galactic latitudes are SX Phe stars. We use a simulation of the Galactic p opulations found in the fields to extrap olate the numb er of SX Phe stars found in the whole Galaxy. Our simulation uses a modified version of a model originally develop ed for p opulations of hot evolved stars (Napiwotzki 2008). Thin disc, thick disc and halo p opulations are included in our simulations, while the scale height adopted for the thick disk is 800 p c. The halo is modelled as an oblate ellipsoid with axis ratio = 0.76. Reddening is included in a simple approximation. Further details are given in Table 3


8
of Robin et al. (2003). No self-consistant modelling of the evolution of main sequence stars and p ossible binary channels, which might lead to the formation of SX Phe stars, is p erformed. We assume that the absolute brightness of SX Phe is randomly distributed within the observed interval MV = 2.1 - 3.2 (which is the implied absolute magnitude for SX Phe stars with p eriods of 2­1 hr resp ectively, McNamara 1997) and that the space density of SX Phe is simply prop ortional to the density of the parent p opulation. This should b e a good approximation for the field p opulation in which dynamical interaction plays a very small role. SX Phe stars are observed in globular clusters and are known to b e metal p oor. Thus it is clear that the thin disk can b e ruled out as the parent p opulation, but b oth the thick disk and halo p opulations, are feasible. The thick disk p opulation is almost as old as the Galactic halo and some stars of this p opulation have quite low metallicities. However, some thick disk stars have overall abundances comparable to thin disk stars (Bensby, Feltzing & Lundstr¨m 2003, Fuhrmann o 2004). Dep ending on the degree of this fraction and noting that the SX Phe phenomenon is linked to metallicity, different formation efficiencies can b e exp ected. We make two extreme assumptions to constrain the Galactic SX Phe p opulation: 1) thick disk and halo have the same formation efficiency or 2) SX Phe stars are only formed from halo stars. A total of 250 million SX Phe was simulated, making the statistical error of the Monte Carlo simulation negligible. We obtained a cumulative distribution of SX Phe stars as function of limiting magnitude. A catalogue of simulated stars in 1 fields around the central coordinates of the RATS fields and brighter than the detection limit of SX Phe variables (V = 22 [for blue stars this implies g =22]) was produced. Stars in this list were weighted with the effective field of view of the cameras (Barclay et al 2011). Simulated star numb ers were scaled according to the predicted numb er of stars and the observed numb er of stars (11 in the low extinction fields). The result is that we predict 6.6â104 SX Phe stars brighter than (V = 22) in our Galaxy if a mix of halo and thick disk is assumed and 4.0â104 if only the halo p opulation contributes. Recent determinations of the dynamical mass of the Milky Way include 1 â 1012 M (Watkins, Evans & An 2010) and 2.5 â 1012 M (Sakamoto et al 2003). However, given that the mass of the Milky Way is thought to b e dominated by dark matter, the stellar mass is exp ected to b e 1/20 of the dynamical mass (eg Moore et al 1999), giving a stellar mass in the range 5 - 12 â 1011 M . Our simulations therefore imply one SX Phe star p er 7.6-18 â 105 stars and that SX Phe stars are significantly less abundant p er unit mass in our Galaxy compared to than that found in globular clusters (eg one p er 4 â 104 M for Cen, Olech et al 2005) and dwarf spheroidal galaxies (eg one p er 2.1 â 104 M for the Fornax dSph, Poretti et al 2008). (We have assumed no dark matter is present in globular clusters and used the results of Lokas 2009 in determining the stellar mass of the Fornax dSph). Given we have assumed that all eleven stars at low Galactic latitude are SX Phe stars (and the sp ectra of seven are consistent with this) our estimate of the numb er of SX Phe stars in our Galaxy may b e an overestimate, indicating an even greater discrepancy b etween the relative numb er of SX Phe stars in our the Galaxy and other nearby stellar groups. Although there is some uncertainty in the numb er of b ona fide SX Phe stars in our survey, the discrepancy b etween the numb ers of SX Phe stars predicted in our Galaxy and nearby stellar systems is over an order of magnitude. The fact that SX Phe stars are less abundant in our Galaxy is presumably a consequence of the metallicity and star formation history of these systems. A comprehensive study of the numb er of SX Phe stars in different environments could lead to a b etter understanding of how these stars are formed.

8

CONCLUSIONS

We have identified 31 blue pulsating ob jects for which we have evidence that they are candidate SX Phe or Sct stars. These pulsators which have p eriods b etween 51­83 mins are well suited to b eing discovered using surveys like RATS which have high cadence but have a relatively short overall duration. Unlike the RR Lyrae stars which have a longer pulsation p eriod and corresp onding brighter absolute magnitude, they have b een little used to identify Galactic sub-structure. Our results suggest that existing survey telescop es would b e well suited to the discovery of SX Phe and Sct stars if their cadence was high enough. Further, if the mode of pulsation can b e identified then they would provide a useful cross-calibration set for luminosity-p eriod relationships and how this is affected by metallicity.

9

ACKNOWLEDGEMENTS

This pap er is based on observations obtained using the Isaac Newton Telescop e, the William Herschel Telescop e on La Palma (the ING); the MPG/ESO 2.2m and ESO 3.6m telescop es at the Europ ean Southern Observatory, La Silla, Chile under programmes 075.D-0111(A) and 079.D-0621(A). We thank the staff of the ING and ESO for their help in obtaining these observations. This pap er also uses data obtained using the 1.9m telescop e of the South African Astronomical Observatory, Sutherland, South Africa and the 2.5m Nordic Optical Telescop e. We thank the referee for p ointing out the issues surrounding the identification of the mode of pulsation and Michael Lemke for the grid of models sp ectra computed for a `A quick pro ject'.

REFERENCES
Amado, P. J., Moya, A., Suarez, J. C., Mart´n-Ruiz, S., Garrido, ´ i R., Rodr´guez, E., Catala, C., Goupil, M. J, 2004, MNRAS, i 352, L11 Barclay, T., Ramsay, G., Hakala, P., Napiwotzki, R., Nelemans, G., Potter, S., Todd, I., 2011,, MNRAS, 413, 2696 Bensby, T, Feltzing, S., Lundstrom, I., 2003, A&A, 410, 527 ¨ Breger, M., 2000, ASP Con Series, 210, 3 Churchwell, E., et al, 2009, PASP, 121, 213 de Jong, J. T. A., Yanny, B., Rix, H.-W., Dolphin, A. E., Martin, N. F., Beers, T. C., 2010, ApJ, 714, 663 Fontaine, G., Brassard, P., Charpinet, S., Green, E. M., Chayer, P., Billeres, M., Randall, S. K., 2003, ApJ, 597, 518 Fuhrmann, K., 2004, Astronomische Nachrichten 325, 3 Gansicke, B. T., et al, 2009, MNRAS, 397, 2170 ¨ Green, E. M., et al, 2003, ApJ, 583, L31 Kurtz, D. W., 2004, Solar Physics, 220, 123 c 0000 RAS, MNRAS 000, 000­000


Stel lar Pulsators at large distances
Kurucz, R. L., 1992, Rev Mex Astron Astro., 23, 181 Lemke, M. 1991, Internal Report, Department of Astronomy, University of Texas at Austin Lokas, E. L., 2009, MNRAS, 394, L102 Jeffery, C. S., 2008, CoAst, 157, 240 Jeon, Y.-B., Lee, S.-L., Nemec, J. M., 2010, PASP, 122, 17 Jester, S., et al, 2005, AJ, 130, 873 Jordi, K., Grebel, E. K., 2010, A&A, 522, 71 McNamara, D. H., 1997, PASP, 109, 1221 Moore, B., Ghigna, S., Governato, F., Lake, G., Quinn, T., Stadel, J., Tozzi, P., 1999, ApJ, 524, L19 Napiwotzki, R., et al, 2004, In Spectroscopically and Spatially Resolving the Components of the Close Binary Stars, ASP Conf Series, 318, 402 Napiwotzki, R., 2008, In Hot subdwarf stars and related ob jects, ASP Conf Series, 392, 139 Nemec, J., Mateo, M., 1990, In Confrontation between stellar pulsation and evolution, ASP Con Series, 11, 64 Nemec, J. M., Linnell Nemec, A. F., Lutz, T. E., 1993, ASPC, 53, 145 Odenkirchen, M., et al, 2003, AJ, 126, 2385 Olech, A., Dziembowski, W. A., Pamyatnykh, A. A., Kaluzny, J., Pych, W., Schwarzenberg-Czerny, A., Thompson, I. B., 2005, MNRAS, 363, 40 Poretti, E., et al, 2005, A&A, 440, 1097 Poretti, E., Clementini, G., Held, E. V., Greco, C., Mateo, M., Dell'Arciprete, L., Rizzi, L., Gullieuszik, M., Maio, M., 2008, ApJ, 685, 947 Ramsay, G., Hakala, P., 2005, MNRAS, 360, 314 Ramsay, G., Napiwotzki, R., Hakala, P., Lehto, H., 2006, MNRAS, 371, 957 Ritter, H., Kolb, U., 2003, A&A, 404, 301 Robin, A. C., Reyle, C., Derriere, S., Picaud, S., 2003, A&A, 409, 523 Rodr´guez, E., Lopez-Gonzalez, M. J., Lopez de Coca, P., 2000, i ´ ´ ´ A&AS, 144, 469 Rodr´guez, E. et al 2007, A&A, 471, 255 i Samus, N. N., et al, 2009, adsabs.harvard.edu/abs/2009yCat....102025S Sakamoto, T., Chiba, M., Beers, T. C., 2003, A&A, 397, 899 Schlegel, D. J., Finkbeiner, D. P., Davis, M., 1998, ApJ, 500, 525 Sesar, B., et al, 2010, ApJ, 708, 717 Szkody, P., et al, 2002, ApJ, 123, 430 Watkins, L. L., et al, 2009, MNRAS, 398, 1757 Watkins, L. L., Evans, N. W., An, J. H., 2010, MNRAS, 406, 264 Verbunt, F., Bunk, W. H., Ritter, H., Pfeffermann, E., 1997, A&A, 327, 602

9

c 0000 RAS, MNRAS 000, 000­000