«1 Standard Photometric Systems Michael S. Bessell Research School of Astronomy and Astrophysics, The Australian National University, Weston, ACT ...»
The Cousins system UBVRI standards are de ned by the Menzies et al. (1989) Eregion list of 570 stars with magnitudes in the range 2V 11 supplemented by the Kilkenny at al. (1998) list of stars with the most extreme blue and red colors. Landolt (1983, 1992) has very usefully provided many faint photoelectric standards in small areas suitable for CCD imaging however, it must be remembered by CCD observers that the catalog measurements refer to apertures of 14 arcsec centered on the individual stars and the same aperture (including background stars) needs to be measured with the CCD. Stetson (2000) provides homogeneous multiple CCD magnitudes in BVRI for all these Landolt elds including values for many additional stars. These magnitudes are based on small synthetic apertures for bright isolated stars and pro le tting photometry (corrected by growth curves) for fainter stars or those with companions within a few arcsecs. The large deviations between some of Stetson's and Landolt's values must result from the di erent apertures used.
As the Johnson RI system is little used now, UBVRI normally refers to the JohnsonCousins UBV system and the Cousins RI system. Sometimes people refer to the JohnsonLandolt UBV system to distinguish between the Landolt (1973, 1983, 1992) version of the UBV system and the subtly di erent Cousins E-region and equatorial version. Taylor (1986) provides transformations between many sets of VRI photometry.
The VRI passbands of Bessell (1990) are a good representation of the Cousins system. Model atmosphere temperature calibrations and bolometric corrections for the UBVRI system have been provided by Bessell et al. (1998) and Houdashelt et al. (2000).
12 Bessell VandenBerg & Clem (2003) provide empirical BVRI temperature relations. Melendez & Ramirez (2003) have recently provided an infra-red ux method (IRFM) temperature calibration for the RI system. Ramirez & Melendez (2004) also provide an IRFM temperature calibration for V-K for solar-type stars. Bergeat, Knapik & Rutily (2001) provide a temperature calibration for carbon stars.
The color V-K or R-K is probably the best color for use in determining the e ective temperature of all but the hottest stars. This is particularly pertinent now that 2MASS and DENIS K magnitudes (see sections 5.5, 5.6) are available for a large number of stars. Not only is the color quite insensitive to metallicity, but the large baseline of the color means that reasonably large uncertainties in the K magnitude for instance will not greatly a ect the derived temperature. For late-M and L dwarfs I-K is the best color to use. If near-IR magnitudes are not available, V-I is the best color to use. It has half the baseline of V-K. It should also be noted that for stars hotter than the sun, V-I is between two and three times more sensitive to temperature than is b-y.
Isochrones and evolutionary tracks in BVRI are provided by VandenBerg et al. (2000), Bergbusch & VandenBerg (2001), Lejeune & Schaerer (2001) and Girardi et al. (2002).
Bessell, Castelli & Plez (1998) (Appendix D) discuss bolometric corrections and provide the ux calibration of the UBVRIJHKL system (Table A1). Note that the magnitude zeropoints in Table A1 for f and f are accidentally reversed.
2.1.1 S20-S25 based VRI systems Sandage (1997, 2001) discusses the UBV(RI) Mount Wilson photometric system derived using an extended-red S20 phototube. This VRI system is closely related to the Cousins VRI system as seen in the derived transformation coe cients, so precise astrophysically valid transformations to the Cousins system should be possible for most halo stars.
Menzies (1993) discusses in detail the systematic di erences that exist between a natural extended-red S20 RI system and the natural GaAs RI system with the same lters. For stars earlier than M, the transformations are single valued and robust, but di erences exist for M supergiants, M giants, M dwarfs and carbon stars due to the di erent red cuto s of the tubes and the di erent strength TiO and CN bands in the stars.
2.1.2 CCD based VRI systems The V and R passbands are generally well reproduced by most CCD users, but the CCD I bandpass, more often than not, extends further to the red than the standard I band.
This is because in the standard GaAs system, the cuto of the GaAs tube de nes the cuto of the I band and unless this cuto near 8800 A is mimicked by an interference edge coating on the I lter, the CCD I band continues out beyond 10000 A. Fortunately, this does not seem to cause a transformation problem for most stars which are smooth in the far-red region, but when there are signi cant non-grey variations in the ux beyond the 8800 A cuto of the standard system, systematic di erences will occur between the standard system and the natural CCD I system. The least that needs to be done is to ensure that the change in slope of the transformation equation is well de ned for the latest spectral-types. Small systematic di erences will also be evident for the hotter ABF-type stars due to the Paschen line contribution. Photometry of emission line objects such as SN will also be a problem with even small di erences in bandpasses, as seen by I band observations of SN1987a with di erent detectors.
The inclusion in the SDSS system of a Z band situated approximately between the cuto of the standard I band and the cuto of the CCD near 11000 A also makes it important to cut o the CCD I band with an edge coating.
Standard Photometric Systems 13
2.2 The Washington CMT1 T2 system The broadband Washington system of photometric standards was established by Canterna (1976) and Geisler (1990). The Washington system was devised to use the wideband sensitivity of GaAs phototubes and CCDs and makes use of the sensitivity of blue-violet colors to metallicity and gathers more violet light in cool stars. Most of its colors can be well transformed into related colors in the Cousins BVRI system some transformations were given in Bessell (1992). An empirical abundance calibration was originally provided by Geisler (1986), and a revised calibration of the system was given by Geisler, Claria, & Minniti (1991). Bessell (2001) revised the passbands of the Washingtom system and computed theoretical colors using the Castelli (1999) model atmospheres of Kurucz. The Washington system, supplemented by the DDO 51 lter is being used in the Spaghetti Survey (Morrison et al. 2000, 2001 Dohm-Palmer et al. 2000) to identify and measure the distances and abundances of K giants. Isochrones are provided by Girardi et al.
2.3 The Sloan Digital Sky Survey ugriz system This revolutionary project, by providing an unprecedented database of photometric observations of stars and galaxies, has essentially made its bandpasses the de facto standard for all future photometric surveys and most future photometric imaging. Much e ort is going into calibrating the SDSS system and providing transformation to other systems for particular kinds of stars. Two data releases have been made (Abazajian et al. 2003,
2004) and although the nal SDSS system passbands are still to determined, preliminary passbands - labelled u0 etc (http://home.fnal.gov/ dtucker/ugriz/index.html) - enable synthetic photometry to be done and isochrones in SDSS magnitudes and colors to be computed (Girardi et al. 2004).
The original g,r,i,z passbands were produced by using Schott GG, OG and RG glasses to de ne the blue edges and a multilayer interference edge coating to provide the red edge. Sets of nominally identical lters were also made to use on other telescopes to provide calibrating standards. Unfortunately, the Survey telescope lters were used in a vacuum and the layers in the interference lter coatings shrank slightly, shifting the red edge blueward by an amount approximately 1% of the wavelength (50-100A). The set of lters provided for the USNO 1m telescope and used for the standards were not stored in a vacuum and did not su er the same shifts, however it means that the two systems are not identical and the consequences of this are being reviewed. The 158 standard stars that de ne the u g r i z photometric system are given by Smith et al.
(2002) together with transformations between the SDSS system and UBVRI. The revised absolute ux distribution of the fundamental SDSS standard BD+17 4708 is given by Bohlin & Gilliland (2004).
In some ways it is unfortunate that the SDSS bandpasses were chosen basically to provide photometric redshifts for galaxies, rather than to isolate relevant stellar absorption features in particular bands or to match existing broad-band systems. It is also unfortunate that the SDSS bandpasses have been altered by the red edges shifting blueward from their designed placement, as this will cause uncertainty as to whether others should copy the originally speci ed SDSS bandpasses or the actual shifted bandpasses.
While considering that question it is also pertinent to consider modifying the ultraviolet and blue-green bands for stellar photometric purposes by placing the ultraviolet band below the Balmer Jump (like the Stromgren u) and by introducing a violet band sensitive to metallicity in cool stars and hydrogen lines in hot stars (like the DDO 38 band or the Walraven L band). For higher velocity resolution of nearby galaxies it would also be useful to break the g band into two.
2.4 The Hipparcos-Tycho Hp BT VT system ESA's Hipparcos mission provided parallaxes of unprecedented precision and accuracy for the nearby stars and also produced exceedingly precise magnitudes for hundreds of thousands of stars (Hipparcos and Tycho Catalogue: Perryman et al. 1997 Tycho2 Catalogue: H g et al. 2000).
The main Hipparcos detector was an un ltered S20 image dissector scanner which provided the Hp magnitudes. Most stars brighter than 8.5 were measured with a precision of a few tenths of a millimagnitude. In addition, light from the star mapper area was divided by a dichroic beam splitter onto two photomultiplier tubes, providing simultaneously measured BT and VT magnitudes. The Tycho catalog provides magnitudes for a larger number of stars, but for the fainter stars the precision is lower than the Hipparcos catalog. For the brighter stars it is comparable. This wealth of accurately calibrated and precisely measured photometric data covers the whole sky, north and south of the equator. This enables intercomparison of many of the ground-based standard photometric systems and a search to be made for systematic di erences with right ascension and declination.
The Hipparcos catalog provided tables and equations for relating Johnson-Cousins B and V magnitudes with HP, BT, VT based on pre- ight calibrations. However, Bessell (2000) discovered small systematic di erences between these tables and values measured using precise E-region stars. The catalog passbands were adjusted slightly until synthetic photometry indicated a good match to the observations. The biggest e ect was for the HP passband, which seemed to require a signi cant redward shift to the blue edge of the bandpass. The Hipparcos spacecraft su ered on-going damage from regular passage through the Van Allen radiation belts which must have a ected the detector response.
The redward shift of the blue edge of the HP passband to t the observations may be an oversimpli cation of the actual damage but until more information is available is an acceptable x. Grenon (private communication 2002) considered the complete loss of sensitivity as adopted was unlikely and favored a more complicated depression of the blue.
Platais et al. (2003) discuss V-I colors and issues concerning the Hipparcos and Tycho data for very red stars, in particular the carbon stars.
2.5 The HST WFPC2 160w, 336, 439, 450, 555, 675, 814 system WFPC2 has 38 lters comprising, 18 broad-band, 5 medium-band, 13 narrow-band (mostly for emission-line work) and two long-pass red lters. Five of the broad-bands, F336W, F439W, F555W, F675W, F814W were designed to be closely related to the UBVRI bands. The system was calibrated from the ground before the ight and regular in- ight observations of standards are made. Holtzman et al. (1995) present an in depth analysis of the WPFC2 photometric bands, their zeropoints and transformations to the UBVRI system. Details of the synthetic photometric comparisons are shown. The authors caution users of the danger of transforming WFPC2 magnitudes onto the UBVRI system for stars which are not representative of those for which the standardization transformation equations were derived. They also suggest caution with respect to reddening corrections.
The synthetic photometry comparisons indicate good agreement except for F336W and U where there was signi cant deviations for hot stars. They suggested that the U bandpass was probably at fault rather than F336W. Synthetic photometry has been computed using the Vilnius (Straizys & Sviderskiene 1972) and Pickles (1998) spectrophotometric catalogs with the F336W passband from the WFPC2 website and the UX90 passband from Bessell (1990). The agreement was much closer (within 0.1 mag) with this U bandpass than that used by Holtzman et al. (1995) (0.4 mag) but the character of the AB star deviations could only be duplicated by shifting the F336W bandpass about 100 A Standard Photometric Systems 15 to the red. Adjusting the U bandpass could not e ect the required change. However, for ground-based work the atmosphere greatly a ects the e ective F336W passband compared to that in space, consequently the ground-based tests are probably not indicative of the true bandpass.
In Fig. 2 are shown the synphot passband and the modi ed F336 band that has been bodily shifted to the red (lighter line) together with the UX90 passband, a modi ed UX90 band (blue edge shifted 50 A to the red) and the U3 band of Buser (1978) (very light line) that is recommended in synphot for the U band. Bessell, Castelli & Plez (1998) discuss synthetic U-B colors based on their model atmospheres in an appendix and show that U-B using the UX90 passband ts the Landolt U-B colors reasonably well, while the synthetic Cousins U-B colors require a scaling of 0.96, equivalent to shifting the blue edge of UX90 by about 50 A as shown in Fig. 2.
The far UV bandpass lter F160BW is discussed by Watson et al. (1994) and Keller et al. (2000), the latter who also provided theoretical F160BW - F555W colors and showed the excellent temperature sensitivity of F160BW - F555W for very hot stars. Isochrones in HST bandpasses have been computed by Girardi et al. (2002).