Annu. Rev. Astron. Astrophys. 1991. 29: 89-127 Copyright © 1991 by Annual Reviews. All rights reserved

5. THE BACKGROUND 1216 TO 1800 Å

A very attractive set of sounding rocket diffuse ultraviolet background radiation data at 1560 Å is provided by Onaka (86) and is shown in Figure 13. These data hold considerable potential for being an important measurement, and they are of a character that permits useful discussion in this section.

Figure 13. Rocket measurement of the diffuse ultraviolet background by Onaka (86). The general level agrees well with the background at longer wavelengths in Figure 11, but some correlation with neutral hydrogen column density appears. This figure is from (86), with permission.

The potential of these data for being particularly important, stems from the use of an imaging detector (87), which means that concerns about point-source contamination are minimized. Figure 13 shows a background intensity of ~ 400 units, consistent with the Johns Hopkins Aries rocket result at longer wavelengths, reported above. With the Voyager results just reviewed the case for an origin in redshifted hydrogen L radiation is strengthened somewhat.

However, we also see in Figure 13 some dependence of the diffuse ultraviolet background on neutral hydrogen column density. Such dependences have been reported before; the usual interpretation is that the correlated portion of the signal is due to the light of OB stars in the Galactic plane scattering off of dust located above the Galactic plane. The correlation with neutral hydrogen column density occurs because of the well-known correlation that exists between neutral hydrogen column density and dust (11, 111).

The theory of such dust-scattered radiation is given by Jura (55). A major problem with the Jura model when it is used at moderate Galactic latitudes is its assumption of a uniform longitude dependence in the original Galactic plane source. We have seen, in Figures 7, 8, and 9, that that assumption is wrong. Nevertheless, Jura's model is useful for discussion as a first approximation. Jura's theory has been applied by Onaka to the data of Figure 13, giving a (1 - g) = 0.065 ± 0.015, where a is the albedo of the dust grains, and g is the heuristic asymmetry factor of Henyey & Greenstein (40). Negative values of g correspond to predominant backscattering. Even quite small positive values of g indicate rather strong forward scattering.

There has long been widespread agreement, which may be wrong, that the albedo of the grains in the far ultraviolet is high; that a = 0.5 may even be an underestimate. This view arose from theory and observation (70 and reference therein to 62). It is supported by an ultraviolet photograph of Orion (14) that seems to show a bright general diffuse glow, and also by the comparison of the Apollo 17 ultraviolet radiation field (34) with that of TD-1 (27) which shows an excess that Henry (31) attributed to bright low Galactic latitude diffuse Galactic light. If we do accept for the moment that a = 0.5, Onaka's data provide g = 0.87 Å 0.03, which is very strong forward scattering.

What would Voyager have seen, if this interpretation of Onaka's result is correct? Extinction is stronger at Voyager wavelengths of ~ 1100 Å than it is at Onaka's 1560 Å (105, 118). Of course Voyager does not see the extragalactic component of ~ 380 units that Onaka's observation implies, but we have assumed that this is because the extragalactic source does not continue below ~ 1216 Å. The question we are addressing is this: should Voyager have seen the dust-scattered component? A simple calculation, assuming no change in a or in g between 1560 and 1100 Å, predicts a signal for Voyager of only ~ 90 units, which is just inside the Voyager upper limit.

The predicted intensity will be less if the ultimate source, the far-ultraviolet radiation field, should decline shortward of L. The observed spectrum of Henry et al (34) suggests that no large decline occurs.

Next, Onaka's relation can be used (for discussion purposes) to predict what should be seen from dust scattering at lower galactic latitudes. This must be done with caution, as Jura's theory does not include multiple scattering, so consider what should be seen at galactic latitude 40°, where hydrogen column densities lie in the range 4-10 x 10²⁰ cm^-2 (29). The prediction is 500-800 units, depending on longitude.

Finally, if indeed there are 500-800 units present at b ~ 40° (of which 120-420 are due to dust), what should Voyager have seen at those latitudes? The answer is, 200-700 units, depending on longitude! No such radiation is observed (Figure 12), and so either this interpretation of Onaka's result (and some others described below) is incorrect, or the grains change in their albedo and/or g value substantially between 1560 and 1100 Å. The grains may of course change, but the change required for, say, 300 units due to dust at 1560 Å is rather large: at 1100 Å a (1 - g) ~ 0.015, which gives a = 0.1 for g = 0.87, and even lower a for smaller values of g, at this wavelength.

Yet we have a very well established Voyager observation of diffuse Galactic light (that is, starlight scattered from dust) in Ophiuchus, described above. Voyager was fully capable of detecting light scattered from dust, if it is there. The data from Voyager are therefore facts of life which those who wish to ascribe much of the ultraviolet background to dust scattering must explain.