|Annu. Rev. Astron. Astrophys. 1991. 29:
Copyright © 1991 by . All rights reserved
The portion of the background spectrum shortward of L presents special technical difficulties for its detection as described above. But the region is of particular interest. One interesting possible origin for the diffuse ultraviolet background could be red-shifted hydrogen L recombination radiation from an intergalactic medium or other redshifted sources, which of course will not be observed in the spectral range (blueward of L) that is under present consideration. In Section 3, a claim was made that there exists a diffuse background, from 1750 to 2900 Å (at least) averaging 400 units. If in the range < 1216 Å there is no such background, a prima facie case would exist that the longer-wavelength radiation (particularly if it can be shown to continue down from 1750 to 1216 Å) is redshifted L. With this in mind we examine the data.
While there are other important observations (e.g. 3, 88) the most comprehensive and useful are those made with the ultraviolet spectrometer on the Voyager spacecraft [Figure 12, from the data of Sandel et al (101) and Holberg (42)].
Figure 12. Voyager observations of cosmic background from Holberg (42), (filled symbols), and Sandel et al (10) (open symbols); upper limits are given by the lower edges of the semicircles. The Voyager limit on the background at latitudes above 30° is 100 units.
The Voyager spectrometer admits L, so scattered L is a problem. The processing of the data is discussed well by Holberg & Barber (43) and Holberg (41) and references therein, building confidence that the results are reliable. The fact that spectroscopy is involved, rather than simply broad-band photometry, permits some real understanding of the origin of the signal.
The results shown in Figure 12 are fundamental. Upper limits are given by the lower edges of the semicircles. Notice that there are no positive detections north or south of b = 20°. Looking only at the lowest upper limits we see evidence that above |b| = 30° there is no cosmic diffuse ultraviolet background brighter than 100 units. The notion that there might be a general background of 300 or 400 units at higher latitudes, as appears to exist at longer wavelengths, seems decisively excluded.
Thus, a prima facie case exists for the notion that the longer-wavelength ultraviolet background (described in detail below) is due to redshifted L radiation, which, if present, presumably would be from slowly recombining, highly ionized intergalactic clouds.
From 912 to 1216 Å, the Voyager data provide only an upper limit on any background. If the longer-wavelength ultraviolet background radiation is redshifted L, what source might we expect for these shorter wavelengths? One intriguing potential source comes from a recent suggestion by Sciama (103), who proposes neutrinos as the dark matter, decaying with emission of photons that are just capable of ionizing hydrogen. Speculative as it may be, this notion has many attractive properties, including an explanation of the remarkably ionized state of hydrogen in the universe (72, 97, 98). Sciama's photons, if they do exist, would be created short of 912 Å, and would be seen in the 912-1216 Å range for sufficient redshift. The work of Stecker (107), Kimble et al (56), Henry & Feldman (36), and Murthy & Henry (80) describes earlier searches for neutrino decay radiation. Results on neutrino decay from supernova 1987A are given by Chupp et al (15) and Kolb & Turner (58). Related discussion appears in Madsen & Epstein (65). If neutrinos have nonzero masses, they may oscillate (e.g. 5). There is the suggestion by Bahcall and Bethe, as reported by Nash (84), that Soviet study of solar neutrinos implies oscillation and hence a nonzero mass.
Sciama's photons also help with a problem that is faced by the L recombination origin, suggested above to be the source of the longer-wavelength background. To get adequate intensity requires considerable clumping [although probably not too much to violate an important observational constraint by Martin & Bowyer (66) on the uniformity of the ultraviolet background], and Sciama's photons would prevent cooling-time difficulties that otherwise arise. Of course if Sciama's photons are the ultimate energy source of a background of redshifted L radiation, that would destroy most or all of his photons!
Returning to the data in Figure 12, the positive detections below |b| = 17° are important: what is the source of this radiation? One particular observation, the 2000 unit observation in Ophiuchus at b = -16.°3, is especially valuable, as it involves a long, slow, spatial scan, over all of which the flux is seen (42). Thus we have decisive evidence for a truly diffuse origin. Holberg analyzes the spectral appearance of this source, concluding that this background ultimately arises from stellar sources of early spectral type. I assume, therefore, that what we are seeing is 2000 units of diffuse cosmic ultraviolet background radiation arising from the light of OB stars scattering from interstellar dust. The location on the sky is near one end of the ``bright half'' of the Gould belt shown in Figure 7. The other six positive detections in Figure 12 all occur near Orion, that is, at the other end of the bright part of the Gould belt; also, none of the upper limits is at a location near the bright half of the Gould belt. Additional observations at other locations along the bright half of the belt would be of great value.
So we have a compelling case for the detection of some diffuse ultraviolet background radiation that surely has its origin in the light of OB stars scattering from interstellar dust. That makes even more interesting the fact that such radiation is not seen by Voyager at moderate or high galactic latitudes (Figure 12) (and also, notice, it is not seen at two low-galactic latitude locations that are in the ``dim half'' of the Gould belt.) Why is no such radiation seen? There is certainly dust at many locations at high galactic latitudes, from IRAS observations, and from interstellar polarization studies. A natural answer is that the grains strongly forward-scatter the radiation out of the galaxy. This would require that the sources of the positive detections in Figure 12 be behind the scattering dust, which is of course possible.
In ending this section which has been positive in its discussion of the Voyager data it is important, however, to point out some of the difficulties encountered with them. In the analysis of a data sample (Section 2) we discussed a variety of contaminants - airglow, stars, and dark current - that may create a false diffuse ultraviolet background. The contrary can not occur: a spectrophotometric system, if it has demonstrated inflight the correctness of its calibration, can hardly fail to detect a true diffuse background if it is there. Therefore, when two diffuse background experiments that have been pointed at the same celestial location disagree, the burden of proof lies, of course, with the experiment that claims the higher background.
In this light, consider again the Voyager results. In his Ophiuchus scan, Holberg provides convincing proof that his signal is truly diffuse background. But such a proof is absent for the ``fixed-location'' positive detections in the Orion region by Sandel et al (101). In particular, there exists the possibility that what is being detected is not diffuse background radiation but direct radiation from a star, or stars, in the field of view. The Voyager (lowest) upper limits correspond to permitting one unreddened B0 star of magnitude 15.5 to be in the field of view (43). A false background of 3000 units would therefore require an unrecognized unreddened 11.8 magnitude B0 star (or several somewhat reddened such stars) to be in the field of view.
Why should that occur in Orion, and not elsewhere? For the very same reason that the claims of a diffuse background in that region of sky must be seriously entertained: that is where exceptionally large numbers of hot stars are located (Figure 7).
We turn next to the important spectral region, 1216 to 1800 Å, where many conflicting observations exist.