Sandage paper 2007
From Tauwiki
Contents |
[edit] Abstract
We have imaged the cosmic diffuse background radiation in both the far and near ultraviolet using NASA'S {\it GALEX} satellite. Observations in the direction of a dense dust cloud at high galactic latitude allow us to rule out the possibility that any substantial fraction of the radiation is extragalactic, and to show that the radiation is not starlight scattered from interstellar dust. We have no theory of the origin of this radiation field.
[edit] Introduction
The diffuse ultraviolet background radiation has been under investigation for many years (Bowyer 1991, Henry 1991). The ultraviolet background has also been discussed in the context of the diffuse backgrounds in {\it other} wavelength ranges by Henry (1999)--- in some wavelength ranges the cosmic backgrounds have turned out to be of fundamental importance. It remains to be seen how deeply significant the diffuse ultraviolet background (or, at least, components of it) will turn out to be. We here use {\it GALEX} observations to eliminate the possibility that any substantial fraction of the observed cosmic diffuse ultraviolet background can be of extragalactic origin.
There are many potential sources of celestial diffuse ultraviolet background radiation, the most obvious such source being starlight scattered from interstellar dust---and understanding {\it that} source (which must be present at some level) is critical to isolating any additional sources of diffuse ultraviolet background radiation.
Over recent years we have carried out a program of observation and modeling, with the aim of understanding in detail dust-scattered ultraviolet starlight. We have applied this model to three different regions: the Coalsack Nebula (Shalima et al. 2004, Sujatha et al. 2007), the Orion Nebula (Murthy et al. 2006, Shalima et al. 2006) and a region in Ophiuchus (Sujatha et al. 2005). In each case, we have been able to fit the observed diffuse radiation as starlight from the nearby hot young stars scattered by interstellar dust in the neighborhood.
We report here on observations of the nebulosity observed by Sandage (1976). This is a region at a moderate Galactic latitude (38.2°) with few nearby stars and thus with very little dust scattered radiation expected. In fact, Murthy et al. (1999) were able to place an upper limit of not more than 100 photon units on the diffuse far ultraviolet background.
We therefore selected this region as one where one might be able to observe extragalactic radiation in the ultraviolet. Such emission would be anticorrelated with the dust in the nebulosity and so should be readily identifiable (Matilla 1990). Our current understanding is that known extragalactic sources cannot contribute more than about 50 photon units (Paresce and Jakobsen 1984; Bowyer 1991) yet the observational evidence is that 300 - 400 photon units are present (Henry 1991). Such emission would have obvious cosmological implications (Henry ..., Paresce and Jakobsen 19xx).
In this work, we will present our observations of the Sandage nebulosity with the Galaxy Evolution Explorer (GALEX) satellite. We have obtained many observations of different regions using GALEX and we will fully describe our data analysis techniques in preparation for the analysis of those data.
[edit] Observations
Our purpose in this work is to observe a cloud at high Galactic latitude to probe the Galactic and extragalactic background. For this purpose we chose the nebulosity first noted by Sandage (1976) in the V band and identified by him as due to the light of Galactic plane stars scattered by the interstellar cloud. Details of this observation and the relevant instrumental parameters of GALEX are given in Table 1. Note that the exposure time is significantly less in the FUV channel because of the power supply problems at the time.
| FUV | NUV | |
|---|---|---|
| Wavelength range | 1350 - 1750 Â | 1750 - 3200 Â |
| FOV | 1.26° | 1.26 ° |
| Resolution | xx | xx |
| Exposure Time (s) | 14,821 | 35,210 |
| Number of Visits | 10 | 22 |
| RA | 09 28 07 | |
| Dec | 70 21 26 | |
| l | 142.3 | |
| b | 38.2 | |
A GALEX observation is broken up into a number of visits of approximately 30 minutes each in order to avoid day time observations. The standard GALEX pipeline will produce a single image for each of the FUV and NUV channels for each visit comprised of the individual photon events during that visit. These individual visits can then be combined to form a single image of the entire observation and the resultant FUV and NUV images are shown in Fig. 1. These images include both stars and diffuse sources of all types; however, it is important to note that the integrated stellar flux is no more than 7% of the total emission in the plates.
Our primary interest is in the diffuse radiation and so we have blanked out all the stars in the GALEX merged catalog for the field as well as the edge enhancement due to instrumental effects seen in this and all other GALEX data and averaged the data into bins of 2' in size. These images are shown in Fig. 2 and represent the total diffuse radiation from all sources in this direction. It is difficult in any imaging experiment to separate the different components of the diffuse radiation and we have to use ancillary information to constrain the sources, listed in Table 2.
[edit] Airglow
Photon counting instruments such as delay-line detectors used in GALEX are intrinsically low noise and the instrumental dark count, due largely to cosmic rays and other fast particles, is only the order of 20 counts per second in the FUV channel and 60 counts per second in the NUV corresponding to a uniform background of about 5 photon units (Morrissey et al. 2007). A significantly greater part of the emission is due to airglow, almost entirely from the OI lines at 1304, 1356, and 2471 Å (ref) with a night time surface brightness of 100 - 200 photon units.
Although we cannot explicitly separate the airglow from the other components of the diffuse sky, we would expect its contribution to change over the course of a single visit, unlike any of the other sources. Indeed, plotting the total count rate from the GALEX TEC counter (Fig. 4) shows that the airglow in a given visit is a function of the local time, with a minimum at local midnight. However, there is also a variation between visits which appears to be correlated with the level of solar activity (archived at http://www.dxlc.com). There is a sharp rise in the solar flux between Jan. 11 and Jan. 14, 2005 which is reflected in an increase in the level of the airglow. A similar result is obtained with the FUV data, although there are a fewer number of visits because of the HVPS problem. Note that, despite our initial expectations, there is no dependence of the airglow on the zenith angle. This line arises in the ionosphere and is due to resonant scattering of solar 2471 Å photons.
There are two components to the airglow contribution: a baseline component even at local midnight which cannot be independently distinguished from the other diffuse sources and a variable component which is correlated with the time from local midnight (Fig. 4). Integrating under the curves in Fig. 4 gives the time averaged airglow contribution from each visit over the baseline from the visit on Jan. 4, 2006. These values are tabulated in Table xx and imply an average contribution of about 80 photon units in the FUV and 115 photon units in the NUV from the airglow to the total diffuse signal. Because of the uncertainty in the baseline, we would estimate that the total airglow contribution is between 100 and 200 photon units in both channels.
[edit] Zodiacal light
Moving out from the Earth, zodiacal light, sunlight scattered by interplanetary grains, contributes to the NUV channel but not to FUV, as the solar spectrum drops off rapidly at wavelengths below 2000 Å. The zodiacal light has been mapped in the optical as a function of helioecliptic coordinates by Leinert (19xx) with the assumption that the color of the zodiacal light is close to 1; that is, that the ratio between the Solar spectrum and the zodiacal light is the same in the UV as in the optical. With these assumptions, the effective contribution of the zodiacal light is about 10 photon units in the FUV and 450 photon units in the NUV.
[edit] H2 Emission
Part of the emission in the FUV may be due to the H2 fluorescent emission observed by Martin, Hurwitz, and Bowyer (19xx) using the BEST instrument on STS-xx. The GALEX field is included in their Target 2, a scan over the Ursa Majoris target. The equivalent surface brightness of the level of molecular hydrogen emission they observed is 110 photon units in the FUV channel (the Werner bands) with no emission in NUV. Martin et al. did not find any variation in the H2 emission over the entire field despite the presence of a CO enhancement and speculated that the emission was from a halo of molecular hydrogen extending far beyond the cloud seen in CO. Following this result, we have assumed that the H2 contribution is uniform over the GALEX field.
[edit] Resultant
The different sources of emission are tabulated in Table xx. Of these, the airglow, zodiacal light and molecular emission are all uniform over the GALEX field of view and thus form a background which can be subtracted from the total. The average level of the remainder is about 700 photon units in the FUV and 750 photon units in NUV but with significant variation over the field. We will discuss these in the next section.
[edit] Results and Discussion
The diffuse UV flux is plotted as a function of the IRAS 100 micron flux in Fig. nuv_fuv after subtraction of the foreground emission (200 units in FUV and 600 units in NUV). There is a considerable scatter in both channels which we have estimated to be about 40 photon units in the NUV channel and 30 in the FUV channel, much greater than the approximately 10 photon units expected from photon statistics alone. This scatter may be either a true scatter in the diffuse background, perhaps due to inhomogeneities in the dust distribution or the radiation field or, more likely, is due to uncertainties on the order of 10% in the flat field. It may even be that the diffuse background will turn out to be the best measure of the instrument flat field.
As mentioned above, Hurwitz et al. (19xx) observed this region using the BEST instrument finding a spectrally flat diffuse background (after subtracting the molecular hydrogen fluorescence) of 500 photon units. Although they did scan over the region, the signal to noise in the BEST spectrograph was too low to record spatial structure of the diffuse radiation, if any. Our data are consistent with this interpretation except that we do find a correlation --- albeit quite poor --- between the NUV and the 100 micron flux (Fig. NUV_FUV) and the NUV and the FUV (but not between the NUV and the FUV).
Edited up to here. JM
[edit] Acknowledgments
This work was supported by NASA grants.
{\it Facility:} \facilityTemplate:\it GALEX.
\begin{thebibliography}{}
\bibitem[Armand, Milliard, \& Deharveng (1994)]{arm94} Armand, C., Milliard, B., \& Deharveng, J. M. 1999, \aap, 284, 12
\bibitem[Bowyer (1991)]{bow91} Bowyer, S. 1991, \araa, 29, 59
\bibitem[Henry (1991)]{hen91} Henry, R. C. 1991, \araa, 29, 89
\bibitem[Henry (1999)]{hen99} Henry, R. C. 1999, \apjl, 516, L49
\bibitem[Henry (2006)]{hen06} Henry, R. C. 2006, \cjaa, 6, Suppl. 1, 40
\bibitem[Henry \& Murthy (1995)]{hen95} Henry, R. C., \& Murthy, J. 1995, in Extragalactic Background Radiation, ed. Calzetti et al., Cambridge: Cambridge University Press, 51
\bibitem[Henyey \& Greenstein (1941)]{hen41} Henyey, L. C. \& Greenstein, J. L. 1941, \apj, 93, 70
\bibitem[mattila (1990)] Mattila, K. 1990, The Galactic and Extragalactic Background Radiation, IAU Symp., 139
\bibitem[Morrissey et al (2005)]{mor05} Morrissey, P., et al. 2005, \apjl, 619, L7
\bibitem[Murthy et al (1999)]{mur99} Murthy, J., Hall, D., Earl, M., Henry, R. C., \& Holberg, J. B. 1999, \apj, 522, 904
\bibitem[Murthy \& Sahnow (2004)]{mur04} Murthy, J., \& Sahnow, D. J. 2004, \apj, 615, 315
\bibitem[Murthy et al (2005)]{mur05} Murthy, J., Sahnow, D. J., \& Henry, R. C. 2005, \apjl, 618, L99
\bibitem[Schiminovich et al (2001)]{sch01} Schiminovich, D., Friedmann, P. G., Martin, C., \& Morrissey, P. F. 2001, ApJ, 563, L164
\bibitem[Schlegel, Finkbeiner, \& Davis (1998)]{sch98} Schlegel, D. J., Finkbeiner, D. P., \& Davis, M. 1998, ApJ, 500, 525
\bibitem[Shalima et al (2006)]{sha06} Shalima, P., Sujatha, N. V., Murthy, J., Henry, R. C., \& Sahnow, D. J. 2006, \mnras, 367, 1686
\bibitem[Sujatha et al (2004)]{suj04} Sujatha, N. V., Chakraborty, P., Murthy, J., \& Henry, R. C. 2004, BASI, 32, 151s
\bibitem[Sujatha et al (2005)]{suj05} Sujatha, N. V., Shalima, P., Murthy, J., \& Henry, R. C. 2005, \apj, 548, 296
\bibitem[Sujatha et al (2007)]{suj07} Sujatha, N. V., Murthy, J., Shalima, P., \& Henry, R. C. 2007, in press (astro-ph/0705.1752)
\bibitem[Weingartner \& Draine (2001)]{wei01} Weingartner, J. C., \& Draine, B. T. 2004, \apj, 548, 296
\bibitem[Welsh et al. (2007)]{wel07} Welsh, B. Y., et al. 2007, \aa, submitted (astro-ph 0706.0927)
\end{thebibliography}
\clearpage
\begin{figure} \epsscale{.80} \plotone{target.ps} \caption{The location of our {\it GALEX} target (celestial coordinates, with North at the top). A grid of Galactic coordinates is overlaid. The circle is the location of our {\it GALEX} observation, coincident with our {\it Voyager} observation (rectangle). The red shading is the IRAS-based dust map of Schlegel et al. (1998). The ultraviolet brightnesses of TD1 stars are included as white dots. Our {\it GALEX} target is not quite centered on our {\it Voyager} target, so as to avoid the TD1 stars.\label{fig1}} \end{figure} \clearpage
\begin{figure} \plottwo{FUV.no.mesh.eps}{NUV.no.mesh.eps} \caption{The FUV and NUV appearance of the sky at our target. Rather more spatial structure is seen in the FUV image than in the NUV image. The FUV brightness is about 725 photons s$^{-1}\,$ cm$^{-2}\,$ sr$^{-1}\,$ \AA$^{-1}$, while the NUV brightness is about 1160 photons s$^{-1}\,$ cm$^{-2}\,$ sr$^{-1}\,$ \AA$^{-1}$, of which perhaps 350 photon units are due to zodiacal light. The spatial distribution over the $GALEX$ field does not resemble any diffuse interstellar medium radiation that we know of. Is it conceivable that the radiation field is of heliospheric origin, rather than interstellar? The appearance of the target in both FUV and NUV is very similar to what was reported by Henry (2006) for two deep imaging survey targets observed with {\it GALEX}. We have eliminated from the image the outermost part of each $1.26^{\circ}$ {\it GALEX} image, as that part is invariably contaminated with an artifact of unknown origin. Note that the NUV image includes a significant amount of zodiacal light, which does not contribute to the FUV image. As the zodiacal light is not expected to have any structure on the sub-degree scale, it will diminish the contrast in the NUV image. \label{fig3}} \end{figure} \clearpage
\begin{figure} \plottwo{fuv.EBV.eps}{nuv.EBV.eps} \caption{Lack of correlation of {\it GALEX}-image ultraviolet intensity (FUV, and NUV) with E(B-V). Note the very substantial values of E(B-V) for this dust cloud; very significant attenuation of any extragalactic radiation would be expected. Our expectation had been that we would see very strong absorption of the presumed-extragalactic diffuse ultraviolet background by the intervening dust cloud. Instead, there is no trace of correlation. The clear implication is that the diffuse ultraviolet background radiation field originates {\it between~us} and the dust cloud, which is estimated (Sandage 1976) to be about 100 pc distant. \label{fig2}} \end{figure} \clearpage
\begin{figure}
\plottwo{correlation.eps}{sandage.spectrum.eps}
\caption{On the left, the figure shows that there is little correlation between the FUV and
NUV spatial structure that appears in the images. Our previous experience (Henry 2006) is
that there {\it is} strong correlation between FUV and NUV, in those cases where there is
strong correlation between ultraviolet intensity and E(B-V) of dust in the field,
but {\it no} correlation between FUV and NUV intensity if (as here) neither correlates with
dust E(B-V). On the right, our {\it Voyager} spectrum of the region we have observed with {\it GALEX}. The data
reduction technique we use (see Murthy et al. 1999) is to force the intensity below
912 \AA\ to zero. If the radiation should be heliospheric, that approach is
not justified. On the other hand, the spectrum shown rises slowly above
912 \AA\ to smoothly join the intensity of 725 units we observe with {\it GALEX}
longward of 1400 \AA,
which suggests that what is presented is indeed a fair representation of the radiation
spectrum. This would argue that the radiation is galactic rather than heliospheric in origin.
\label{fig2}}
\end{figure}
%\begin{figure} %\plottwo{fuv.bw.eps}{nuv.bw.eps} %\caption{The FUV and NUV appearance of the sky at our target. The spatial distribution %does not resemble any diffuse interstellar medium radiation that we know of. It is %possible that the radiation field is of heliospheric origin, not interstellar. The %appearance of the target in both FUV and NUV is very similar to what was %reported by Henry (2006) for two deep imaging survey targets. %\label{fig6}} %\end{figure} %\clearpage
