GALEX
From Tauwiki
[edit] A Study of SANDAGE Region (MBM 30)
THIS PAGE WILL NOT BE FURTHER UPDATED. IT IS OF HISTORICAL INTEREST ONLY AND FOR REFERENCE PURPOSES. HOWEVER THE INFORMATION HERE MAY HAVE BEEN SUPERCEDED BY LATER WORK.
[edit] Observation
The GALEX imaged the SANDAGE region in the vicinity of M82 (MBM 30 cloud) in January 2005-2006 as part of our own GALEX guest investigator program. The center of this observation was around an RA, DEC of 142.0, 70.4 (9 28 6.7, 70 21 26.18) (gl, gb => 142.3,38.2). There are a total of 41 observations in the GALEX database but most are under 1000 seconds and only ours is over 5000 seconds. The total exposure time is 35,209 s.
[edit] Details of the Observed Region
As per Sandage (1976) the cloud in this region is supposed to be at a distance of 100 pc. We have also tried to locate the cloud distance by plotting the Hipparcos distance versus the E(B-V) of stars present within +/- 3 deg from the center of our target and found good agreement with Sandage (1976) (see Fig.1). The spreading of the E(B-V) beyond ~100 pc in the figure is due to the filamentary nature of the cloud in the region.
[edit] Data Analysis
The GALEX image consists of point sources and diffuse radiation which is the sum of dust scattered star light, airglow, zodiacal light (which is negligible in the FUV), integrated light of faint galaxies, dark counts, etc. In order to separate the diffuse background from the observed images need for our analysis, we have to remove the point sources from the images. The positions of stars from each of the GALEX field are extracted using the open source SExtractor program (Bertin and Arnouts 1995) and are tied to the catalogs - TD-1 (Thompson et al. 1978), Tycho2 (Hog et al. 2000), and SK2 (Myers et al. 2002) and are tabulated. We have blanked out these positions (see Fig.2) and then rebinned the background image (see Fig.3a & 3b) to increase the signal-to-noise ratio for our studies. Fig.3c is the GALEX pipeline derived background image.
[edit] Components of Diffuse UV Radiation
The BULK of the radiation in ALL of our GALEX images is the diffuse radiation. The point sources are only a fraction of the total signal. The most extreme case is SANDAGE itself, where less than 7% of the signal is due to point sources! The components of the diffuse radiation present in the field and their strengths are discussed below.
[edit] Dark Count
The instrumental dark count - which is one of the contributors to the diffuse background image - is low, 20/60 cps which corresponds to 5(4) 
(hereafter referred to as 'photon units') in FUV/NUV bands, when compared to the lowest total nightsky backgrounds of 1000/10,000 cps which corresponds to 246 & 601 photon units (Martin et al. 2005, ApJL).
[edit] Airglow
[edit] Airglow Observations from Space
Airglow emission from the Earth's atmosphere is one of the major sources of radiation in the UV and varies considerably with time, on short and long timescales, mainly due to changes in the atmosphere and in solar activity. About a factor of 2 variation occurs with the solar cycle, with airglow faintest at sunspot minimum. It is also a strong function of the zenith angle.
Different lines arise in different parts of the atmosphere with some in the ionospheric E layer at ~ 90 km, some in the F region above 150 km and others at much higher altitudes. Figure 4a shows ultraviolet nightglow spectrum observed at a height of 330 km in Jan. 1986 at minimum solar activity. The viewing line of spacecraft on the night side of the atmosphere may cross the terminator and continue through the sunlit parts of the atmosphere. Under these twilight conditions, dayglow features become important. E.g. the NO bands then are excited by resonance fluorescence and then are much stronger, the N2 Lyman-Birge-Hopfield bands are clearly visible, and the forbidden [OII] emission at 2470 Å is strong (see Fig.4b) (Meier 1991). Since the GALEX altitude is at ~700 km, the other contributions like N2 LBH can be considered as negligible (the scale height of N2 is ~30km (Brown et al 2000).
The primary contributors in the UV at the altitude of GALEX (700 km) are Ly α (1216 Å ) and the O I lines at 1304 and 1356 Å. The strength of these lines as observed by HST at 600 km is summarized in Table 2. The 2471 Å line of OI is the only airglow line of consequence in the NUV portion of the spectrum. Reference is Boffi, F.R., et al. 2007, "ACS Instrument Handbook", Version 8.0, (Baltimore: STScI). Note that the values given in the table are typical values observed by HST and it will vary depending on the position of the telescope with respect to the day-night terminator, solar activity and the position of the target relative to the Earth limb.
[edit] Predictions
[edit] Instrument Details
Three of the GALEX back focal assembly optics are coated with multilayer filters designed to enhance the in-band throughput and off-band rejection of the GALEX instrument. The dichroic coating applied to the entrance face of the fused-silica aspheric corrector plate separates the FUV (reflection) and NUV (transmission) optical paths. The FUV channel has a CsI photocathode deposited directly on the MCP and a MgF2 window for UV transmission down to the instrument cutoff at approximately 1350 Å. This all-dielectric dichroic coating provides a significant improvement over conventional 40%-40% UV beam splitter coatings, with a mean reflectance of 61% over the 1400 - 1700 Å band and a mean transmittance of 83% over the 1800-2750 Å band. A transmissive blue-edge filter coated on MgF2 provides 10% rejection of the OI 1304 Å airglow line for the FUV channel. A reflective broad-band red-blocking filter on the M3 folding mirror has an edge at 2800 Å. This edge yields an additional factor of 10-20% rejection for the NUV zodiacal light background above and beyond the natural Cs2Te detector photocathode cut-off.
[edit] GALEX Effective Area
The GALEX effective area is given taken from the GALEX GI pages with specific text files for NUV and FUV and are plotted in Fig.5.
[edit] AG Calculations
If we use this effective area and a value of 1 R for the O I (1356 Å), we find that the equivalent contribution of this line to the FUV band pass would correspond to a surface brightness of 113 photon units. There would be no contribution at all from the Ly &alpha or O I 1304 Å lines as per the published effective area. In fact, because of the CaF2 filter, it is unlikely that there is any contribution from Lyα. However, as noted above, there is an estimated 10% leak from O I (1304 Å) implying a roughly equal contribution to the total background. The total airglow contribution to the FUV is therefore equivalent to about 200 photon units but will depend at least on the solar cycle and zenith angle. Assuming 1 R from OI 2471 Å would contribute about 100 photon units to the NUV.
[edit] Airglow Variation
We would expect that the airglow would be correlated with the zenith angle. In order to calculate the zenith angle, we use the fact that the zenith direction is the same as the Local Sidereal Time (see Zenith Angle) so the angle is given by the dot product between the LST direction and the look direction. However, and surprisingly, it turns out that there is no correlation between the zenith angle and the total count rate. Instead the count rate (in a single visit) is only dependent on the local time with a minimum at the local midnight.
There is an additional difference between visits which depends on the level of solar activity. There is a sharp rise in this level between Jan. 11 and 14, 2005 which is when the first set of visits occurred (Fig. solar_2005) with a much lower level on Jan. 3, 2006 when the second visit occurred (Fig. solar_2006). The solar activity plots are from [Jan Alvestad's site].
[edit] Optical Depth
The effective cross-section for the dust grains is almost the same in FUV and NUV at 1.4 x 10-21 cm2.
[edit] GALEX Observations
We have used the event counters on the instrument to track the total signal during each visit. Of course, the zodiacal and cosmic backgrounds will not change but the airglow may be expected to vary with zenith angle. This is shown in Fig. 6 where we have plotted the count rate in photon units for one of the visits. The shape of the curve is a characteristic inverted bell shape and we have assumed that the minimum values correspond to the times when the airglow is essentially zero. Based on this curve, we can calculate an average contribution of the airglow to the total signal. But actually this minimum value itself is varying during the visits of our target. So we have separately calculated the constant airglow value for both FUV & NUV visits and are tabulated below.
There are 10 visits in FUV, and 22 visits in NUV for this observation and the average airglow intensity in FUV and NUV bands for each visit is plotted in Fig. 7. The average variable component in the total AG is about 46 photon units in the FUV and 58 photon units in the NUV with a variation between 0 and 250 photon units depending on zenith angle. The average total AG variation in FUV and NUV respectively 84 pu and 116 pu. Most of the airglow in the FUV channel will be from the OI lines at 1304 and 1356 Å and from OI 2471 Å in the NUV. The exposure details of each visit is shown in Fig. 8.
[edit] Zodiacal Light
In the NUV image most of the observed radiation is due to the zodiacal light, the Sunlight scattered by interplanetary dust grains. Since the spectrum of the zodiacal light follows that of the Sun and, as the Sun is a cool star (G type), its contribution in the FUV i.e., below 2000 Å is negligible. The level of zodiacal light in the NUV images is dependent on the angle from the Sun and the distance from the ecliptic plane. Considering the fact that GALEX is constrained to perform its operation at more than 90° from the Sun, the contribution can vary from 300 to 10,000 units in the NUV. Leinert et al. (1998) has tabulated the Zodiacal Light as a function of helioecliptic coordinates and we have used these values to estimate its contribution to any given observation. The effective photon counts in the GALEX NUV band is given in Fig. 7.
The level of the zodiacal light in each of the visits is tabulated in Table 2 (Zodiacal Light in NUV) and can vary by up to 50 photon units between observations.
| Date | L | β | Brightness |
|---|---|---|---|
| 11/01/05 | 176.2 | 51.4 | 454 +- 50 |
| 19/04/05 | 85.61 | 51.4 | 474 +- 50 |
| 21/11/05 | 124.13 | 51.42 | 403 +- 50 |
| 03/01/06 | 167.8 | 51.4 | 454 +- 50 |
The NUV detector is preceded by a red blocking filter on the M3 folding mirror, which produces a sharper long wavelength cutoff and thereby reduces the zodiacal light background and optical contamination. Assuming a red leak of 10-4 at 4000 Å and 10-5 at 5000 Å the contribution from the visible part of the zodiacal light is no more than about 15 photon units.
[edit] Extragalactic Light
Extragalactic background light (EBL) present in the UV is mainly from integrated light of galaxies and QSO's, redshifted starlight from unresolved galaxies, stars and dust from IGM, decaying elementary particles such as neutrinos etc and because of its weakness, the separation of different components is very difficult. It is assumed to be an isotropic component over the entire sky but its contribution in any sight lines is depending on the Galactic extinction.
The possible components of diffuse FUV EBL and their observed intensities (upper limits in photon units) are listed in Table 3 (Leinert et al 1998). Assuming these values, we have derived the amount of EBL in our region after correction for Galactic foreground extinction and are listed in Table 4. If we assume a maximum of 300 units, the extragalactic component in our field is reduced to about 120 units for the low N(H) regions and to about 10 units for the highest N(H) regions in FUV & NUV, leading to an average contribution of 65 photon units which are very less compared to the total diffuse signal observed in the region.
[edit] H2 Fluorescence
Our target is well inside the Ursa Major Molecular cloud. The CO map of the region is shown in Fig.8. The GALEX field includes a region where an HI cloud is contiguous to the CO cloud (see Fig.8). According to De Vries (1988), the filamentary CO core of the cloud appears to be surrounded by H2 halo with insufficient column density to form detectable amounts of CO, which indicate the cloud itself is currently being destroyed by the radiation field.
The Berkeley Spectrometer scanned the entire cloud during the period of January 12-20 in 1989 as part of the UVX Shuttle experiment (Martin et al 1990). They found strong CO emission as well as signature of molecular hydrogen fluorescence in the form of strong emission in the 1570 - 1610 Å band. The scan also reveals that there wasn't any significant enhancement in H2 intensity toward the core, which they interpret might be due to the absorption of background by dust in the cloud. The spectra obtained from the scan for the target is shown in Fig.9. We calculated the effective contribution of H2 fluorescence in the FUV field from this spectra as 135
20 photon units and considered it as a uniform component in the GALEX field.
[edit] Dust Scattered Radiation
We have modelled the dust scattered radiation in the observed region using our single scattering model (Sujatha et al 2005) by distributing the total hydrogen column density at 100 pc in 3 pc. Our analysis shows that stars which are lying within 30 deg latitude can produce not less than 91% of the total ISRF towards the observed line of sight, in which ~60% is comming from the stars within 15 deg. Also it is clear from the modelling that the stars which are forward scattering is contributing ~ 63% to the total dust scattered radiation. If we consider only the distance of the stars, we can say that the stars with distance less than 110pc is contributing only 23% and stars with distance between 110 and 700 pc is contributing ~73% to the total ISRF towards the observed direction.
Our model output shows that the dust scattered radiation is not varying linearly with the amount of dust in the region and is almost under saturation in UV due to the large optical depth of the region. The FUV & NUV optical depth in the observed region is varying from 0.812 to 3.248
If we assume the optical properties of dust grains in the region are similar to our best-fit parameters (Sujatha et al 2007) i.e., (a,g) = (0.3,0.6), we could see that the dust scattered radiation in the field is almost constant around 550 photon units in the FUV and is around 410 photon units in the NUV.
Based on the above informations, we have tabulated the various contributions in our field in Table 5.
[edit] Correlation Studies
[edit] FUV and NUV correlations
There is a correlation between the NUV and FUV fluxes (Fig. FUV_NUV). However, the scatter is quite large for the NUV/FUV data. If we force the NUV data to have the same dependence as the FUV, we find a difference of 130 photon units between the two. However, this is clearly not a great fit. The actual ratio from the ISRF (NUV/FUV) is 0.8 which actually gives us a worse fit because the observed ratio is 1.2. Note that the slope is independent of the constants.
If we look at the ratio of the FUV and NUV, we find that it increases with increasing FUV. This would be consistent with the excess being due to H2 fluorescence. This only contributes to the FUV but not NUV - thus the ratio increases with increasing FUV.
The radiation field varies very little across the field and so we can assume that the ratio between the FUV and NUV should be the same. Therefore, we can subtract the minimum ratio of about 5 and assume that the rest is due to H2 fluorescence.
This has no correlation to the IR map which, according to de Vries, maps the CO cloud.
[edit] NUV Radiation
There is no strong correlation between the FUV and NUV fluxes (Fig.14). If we were to force a linear fit, we would find NUV = 0.4*FUV + 894 but the correlation coefficient is only 0.45. If the τ value is the same in both FUV and NUV, the scale factor should be unity with only a difference in baseline.
If we look at the data in more detail, we find that the mean NUV flux is 1196; with a standard deviation of only 38 photon units. The statistical deviation is about 7 photon units plus flat fielding and other systematic errors might yield another error of 5% or 50 photon units. Thus our background may indeed be flat to within the uncertainty. There are a number of outliers present; if we remove these, the mean is almost the same at 1189 photon units but the standard deviation drops to 30 photon units. These outliers are at the edge of the field and so may be affected by edge issues. (Figs: NUV_outliers and NUV_IR_Correlation).
Going still further, it is clear from the NUV-IR plot that there is no correlation with the IR at lower values of both but that, at the upper end, there is an increase in NUV with IR. These points are localized in a cloud which can, with some imagination, be seen in both the NUV and at 100 microns (images below). If we further exclude these points, then the mean NUV is 1185 with a standard deviation of 28 photon units.
[edit] FUV Radiation
The FUV data show no correlation at all with the 100 micron emission (Fig. 11) and the cloud seen in the NUV data is not visible in FUV. The FUV data do show a correlation with the NUV data as in Fig. FUV_NUV but with significant scatter. The blue pixels are, by and large, from the center of the image (Fig. FUV_blue_points). The ratio is shown in Fig. Ratio_map. We have subtracted a uniform background from each of the two images before taking the ratio. In the case of the FUV, we subtracted 200 photon units to account for the airglow and H2 contributions and 600 units from NUV to account for airglow and zodiacal light.
The ratio peaks in the center of the image. If we exclude those points, the ratio is reasonable fit by a straight line. The parameters are (NUV - 600) = 0.53 (FUV -200) + 287 with a correlation coefficient of 0.7725. The data are certainly not consistent with a ratio of 1, which is what we would expect.
[edit] Stellar Contribution to Dust scattered light
If we fix a specific a and g, we can estimate the number of stars which contribute to the total scattered flux.
[edit] a = 0.1; g = 0 (isotropic scattering)
For a single point [(gl, gb) = (142.48273, 37.771142)] 10% of the scattered light is from a single star at (gl, gb) = (136.96, -18.559) but many stars contribute to the total:
| Cumulative Percentage | Number of Stars |
|---|---|
| 11 | 1 |
| 20 | 4 |
| 30 | 12 |
| 40 | 26 |
| 50 | 61 |
| 80 | 659 |
| 90 | 1824 |
| 99 | 11842 |
[edit] a = 0.1; g = 0.6 (forward scattering)
Now for the single point [(gl, gb) = (142.48273, 37.771142)] about 6% of the flux is from the brightest star (the same as in the previous case).
| Cumulative Percentage | Number of Stars |
|---|---|
| 6.7 | 1 |
| 20 | 7 |
| 30 | 19 |
| 40 | 39 |
| 50 | 74 |
| 80 | 697 |
| 90 | 1817 |
| 99 | 10093 |
[edit] Where does the starlight come from
In both cases, the distribution of the contribution stars is essentially the same - peaking at about 100 - 150 pc and coming from the bright stars in Crucis.
[edit] Speculation
What could cause this effect? Remember that the τ values are almost the same in both channels. We understand the uniform effects reasonably well - it amounts to about 200 photon units in FUV and 600 in NUV. This leaves between 400 and 700 photon units in FUV and NUV. We can break this emission into two parts. There is a cloud visible in both IR and NUV but not in FUV, where there is no correlation at all. There is a correlation between the NUV and FUV in the rest of the image.
The difficulty is that the FUV and NUV should track each other because τ is the same. The only possibility is that the cloud is illuminated by cooler stars which do not have any FUV emission. The remaining emission would be a mix of stars which would match the ratio. In fact, when I look at the effective emission from stars of different spectral types, I get the following table. The ratio between FUV and NUV fluxes matches that seen in the data.
| spectral type | FUV | NUV |
|---|---|---|
| O | 122 | 51.6 |
| B | 55 | 30.4 |
| A | 0.98 | 1.89 |
| F | 0.01 | 0.6 |
| G | 0 | 0.16 |
[edit] Correlation between NUV and IR Intensities
In Fig.12 we have plotted the NUV brightness against the corresponding IR 100 \micorn intensity of the region. Here we can see that approximately 60% of the data are correlated with each other with a correlation coefficient, r=0.59 which is far better than that of the FUV case where the correlation coefficient is only 0.06. From these it is clear that there is some major difference between the sources of radiation in these bands in the region. This could be due to the additional contribution of late type stars such as F type in the NUV in addition to the contrubution of early type stars. The list of stars in the region shows that some of the early F type stars are lying in front of the clouds.
The correlation between the two UV channels and the IR is plotted in Fig. FUV_NUV_IR. There is considerable scatter in the data but it is clear that there is a correlation between the NUV and the IR but not the FUV and IR. The lines show the best fits. In the FUV, it is essentially a straight line with an intercept of about 530 photon units. In the NUV, there is more of a correlation with an intercept of 524 photon units and a slope of 9.7.
The stars (O-F) present in the region within +/- 3 deg are listed in Table 7. It is clear from the table that most of the early type stars (B & A type) are lying behind the cloud and we found that its contribution to the scattered output is negligible. This is also clear from Fig. 13 in which we have plotted the ISRF as a function of distance from the Earth. This confirms that the illuminating source of the cloud is the integrated light from the galactic plane.
[edit] Previous Observations
Hurwitz et al. finds 500 photon units from BEST. Also spectrum is in Martin et al. and is plotted in
[edit] Discussion and Conclusion
(1) Distance vs. E(B-V) of stars within +-3 deg of the target shows that the cloud is at ~ 100 pc away.
(2) Single scattering model prediction of scattered light after distributing the dust from 100 to 104 pc in 0.1 tau step is ~ 550 photon units in FUV and 450 photon units in the NUV for our best-fit parameters (a,g=>0.3,0.6) for the lowest tau locations in the region.
(3) The EBL in the field is varying between 7 and 120 photon units in FUV & NUV if we assume a total of 300 photon units of EBL.
(4) The contribution of H2 fluoresence in the GALEX FUV band is ~ 13520 photon units. No contribution in the NUV.
[edit] Separation into different visits
If we break the FUV into two sets (visits 1-5 and 6-10), we find an excellent correlation with an error bar of about 15 units. There is a difference of about 40 units (y intercept) between the two data sets. There is a similar error bar for the NUV data but a difference in the y intercept of only 17 units. I believe that the zero has already been taken into account so the difference should actually be zero. This level of about 40 units may be considered as an uncertainty in the actual level of the airglow.
A further comparison with the data suggests that the best fit (table below) comes with an additional (flat) background subtraction of 150 photon units, implying an underestimate of the airglow. Note that we only looked at the NUV data because the FUV had the additional component of H2 fluorescence.
[edit] Other papers
[edit] de Vries et al. ApJ 319, 723
- IR to CO and IR to N(H2), respectively.
- Optical depth of 0.3 mag.
- I100 = aN(HI) + bW(CO)
- a = 10-20
- b = 1
- distance of cloud is 100 pc
[edit] Sandage 1976 AJ 81, 954
- AV = 0.3 mag.
- N(HI) = 5e21
[edit] Wieland et al. ApJ 306, L101
- Cloud 30
- 100 micron much more extended than CO because HI extends outside.
- No temperature dependence in cloud.
- Linear correlation between WCO and 100 micron.
[edit] Haikala et al. ApJ 443 L33
- FAUST images of diffuse clouds.
- FOV of 7.6°
- Spatial resolution of 1°2 for extinction.
- FAUST pixels of 4.5' square
- Correlation of I(FAUST) = 128 I(100) -265
- AV = 4.3E(b-y) = 0.39 mag
- HI column density of 6.5 x 1020 cm-2
[edit] Martin, Hurwitz, & Bowyer ApJ 354, 220
- Molecular H2 fluorescence in UMa.
[edit] References
(1) Bertin & Arnouts 1996, .
(2) Leinert et al. 1998, A&AS, 127, 1-99.
(3) Martin et al. 2005, ApJL.
(4) Murthy, J. & Henry, R. C. 1995, ApJ, 448, 848.
(4) Sandage, 1976, AJ.
(5) Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525.
(6) Thompson, G. I. et al. 1978, Catalogue of Stellar Ultraviolet Fluxes (Science Research Council).
