http://www.astro.ku.dk/~erik/gaia/72.2G.phot.txt http://www.astro.ku.dk/~erik/gaia/72.2G.fig.ps.gz A new photometric system (2G) ============================= SAG_CUO_72 E. Hoeg, J. Knude, V. Straizys 13 April 2000 ABSTRACT: Photometric filter systems with relevance for the discussions of systems proposed for GAIA have been described in tabular form in SAG_CUO_71. Here follows a description of a new proposed 2G system with the 5 bands of the BBP now including the UV. The 10 band MBP, based on the Stromvil system is very similar to the 1G system. Only one of the Paschen bands is shifted by 10 nm, and the total integration time agrees with the current GAIA baseline. A graphical overview of all photometric systems relevant in discussions of GAIA is given as CUO_72...ps. ===== Introduction There are several reasons that the proposed 1G system should be replaced by another, the 2G system. The discussion at a workshop on GAIA photometry held at Moletai, Lithuania, in October 1999 and further studies have lead to the following proposal. *** Discussion of BBP The 1G system contained 5 bands of 100 nm width, covering only the wavelengths >400 nm because the CCD considered at the time of proposing the 1G had no UV sensitivity. After introduction of the CCD#1B with UV sensitivity for astrometry, the same CCD has to be used in the BBP, and the bands must cover the whole spectrum. It has been shown that the 4-band BBP of the F-system is sufficient for chromaticity corrections of the astrometric measurements. The measurements with the BBP in the Astro-1 telescope are suited for this purpose because they are obtained at high angular resolution, 37 mas/sample along scan. The high angular resolution is required for correction of partly merged double or multiple stars. The 4 bands in the 1F system are sufficient to cover the needs of astrometric chromaticity corrections. However, we argue that for reasons of better astrophysical significance 5 broad bands should be taken, including a UV band. A possible choice is then the 5 Sloan bands (see CUO_71). But the UV band should be broader in order to exploite the absence of atmospheric absorption and for the sake of chromaticity determination. The Sloan system was designed specifically for photometry of galaxies and quasars, not for stars as GAIA needs. The classification possibilities using the Sloan system are described by Straizys, Lazauskaite, Liubertas and Azusienis in Baltic Astronomy 7-4, 605-624, 1998. The Sloan passbands were designed to exclude the strongest night-sky lines of OI 557.7 nm and HgI 546.0 nm (see Fukugita et al. 1996). Therefore a division was placed at 550 nm, but this argument does not count for observations from space where night-sky lines are absent. It was argued at the Moletai meeting that one of the bands should be placed at 550 nm corresponding to the familiar V. No such band is present in the Sloan system. Another band should be placed at Ic, defined by Cousins (1978, MNASSA, 37, 8). The ESO implementation of Ic has the centre=798.2 and FWHM= 142.6 nm. The inclusion of V and Ic would give a direct connection to the experience gained from the large amount of existing stellar photometry. V-I is an important parameter of temperature in a wide range of spectral classes, luminosities and metallicities. *** The 2G-BBP system The 2G-BBP system is listed in Table 1. It is very similar to the Johnson-Cousins UBVRcIc 5 colour system. The GAIA bands are called UgBgVgRgIg. An analysis of the classification possibilities using the 2G-BBP system will be finished soon, similar to the one for the Sloan system by Straizys et al., cited above. *** The 2G-MBP system The 1G system contained 10 bands, the 7 Stromvil bands and 3 red bands around the Paschen jump. The 2G system (see Table 2) has the same bands, only the centre of the p1 band is shifted 10 nm towards red. The total transit time available for MBP is now 60 s, corresponding to 2 deg field in direct imaging, instead of 90 s = 3 deg assumed for the 1G using a dichroic filter. *** CCD CCD#1B must be assumed for the BBP, as in the 1G and 1F systems. The red-sensitive CCD#2 is now assumed for the red parts of MBP, as in the 1F system, but CCD#1B was assumed in the 1G system. ------------------------------------------------------------------- TABLES 1 and 2 ===== The 2G-BBP system Table 1. The 2G-BBP system. The system differs from the 1G-BBP system described in CUO_71. The system is very similar to UBVRcIc. We list the two half-maximum wavelength (HMW) of each filter in a separate column. The QE of the CCD is given for the wavelength in the Centre column, and also for the red tail of the CCD which defines the red side of the Ig filter. The system is designed for the CCD#1B in BBP. Names Centre HMW FWHM QE t Notes nm nm nm % s BBP, CCD#1B 295 2G33B Ug 330 80 58 0.86 380 395 2G44B Bg 445 100 84 0.86 495 510 2G55B Vg 550 80 80 0.86 590 570 2G64B Rg 645 150 74 0.86 720 730 64 2G82B Ig 820 180 43 0.86 Red tail of CCD#1B: 850 35 900 22 910 950 13 ---- BBP total 4.30 s ===== The 2G-MBP system Table 2. The 2G-MBP system. The system is very similar to the 1G-MBP system described first in CUO_58 of 11 Feb 1999, and in CUO_71. The system is designed for the CCD#1B in the part with <550 nm. The more red-sensitive CCD#2 is assumed for >550 nm. Names Centre FWHM QE t Notes nm nm % s MBP, CCD#1B SSM0 Gw 550 500 - 0.004 G band, bright stars SSM1 Gw 550 500 - 3 G band = no filter 2G35 u 350 40 65 3*3 (1) 2G38 P 380 30 77 2*3 2G41 v 405 20 84 3 2G46 b 460 20 84 3 2G52 Z 515 20 82 3 2G55 y 545 20 80 3 MBP, CCD#2 2G66 S 655 20 84 3 2G81 p1 810 40 90 3 (2) 2G88 p2 875 30 82 3 2G94 p3 938 20 65 2*3 ---- MBP integr. of colours 14*3 s = 42 s SSM 3 s Spaces 11*1 = 11 s N+1=11 filters * 1 s Spare time 4 ---- MBP total transit time 60 s = 2 deg ===== Notes to Table 2: Note 1. The medium u passband must be different from the broad Ug passband. The main arguments are two: (A) The effective wavelength of such a broad passband strongly depends on the intrinsic energy distribution of a star (or its temperature, luminosity and metallicity). As a result, all color excesses and Q-parameters which include the broad u magnitude, depend strongly on spectral class, luminosity and metallicity. Consequently, it is impossible to calculate interstellar reddening-free Q parameters without knowledge of the spectral types of stars. This means that for calculating Qs we need the spectral types but the spectral types themselves must be determined by Qs. (B) Since differences of ultraviolet interstellar extinction laws increase when passing to shorter wavelengths, the passband at 330 nm with its long extension to UV is much more sensitive to the law variations than the narrower passband at 350 nm. Variations of interstellar reddening laws cause additional difficulties in classification of stars by their reddening-free Q-parameters. Note 2. Recent experiments with infrared passbands in the vicinity of the Paschen jump (V. Straizys, Baltic Astronomy, 8-4, 491, 1999) show that the mean wavelength of the p1 passband may be shifted from 800 nm to 810 nm. This makes no difficulties in measuring the Paschen jump height, but increases usefulness of this passband in recognizing carbon stars. As a result, the group of passbands at 810 nm (p1), 875 nm (p2) and 938 nm (p3) are of the same sensitivity to TiO and CN absorption bands as the passbands F82 and F89 proposed by Grenon et al. (1999). No further red and infrared filters are necessary. ---------------------------------------------------------