Annu. Rev. Astron. Astrophys. 1991. 29: 59-88
Copyright © 1991 by . All rights reserved

Next Contents Previous


In the 1980s a number of experiments were carried out to investigate the correlation between the intensity of the far ultraviolet background and parameters associated with the Milky Way Galaxy. The most common association investigated was that with Galactic neutral hydrogen column as derived from 21 cm radio studies. It is important to note, however, that many variables correlate with Galactic neutral hydrogen column and that a correlation of the far ultraviolet intensity with this parameter may be only a consequence of these general relationships. Additional evidence would have to be provided to demonstrate that a correlation between the far ultraviolet intensity and Galactic neutral hydrogen column is, in fact, a direct correlation. This point is addressed later in this section. Despite this caveat, an observed correlation with any galactic variable would be of profound significance since it would demonstrate that the flux is not extragalactic but is Galactic in origin.

The Berkeley Extreme Ultraviolet Telescope (Paresce, McKee, and Bowyer 1980) obtained data on this topic with a 37 cm diameter grazing incidence telescope. The telescope had a 2.5° circular field of view and a filter mechanism that was used to select various bandpasses for observation. In addition to the primary extreme ultraviolet bands, the instrument also included a far ultraviolet channel. The bandpass of this channel (1350 to 1550 Å) was defined by a calcium fluoride filter on the short wavelength side and by the rapid decline in the detector efficiency on the long wavelength side. Because of the rather unusual observing conditions characterizing this joint-nation manned mission, the data obtained were of uneven quality (Paresce et al. 1979). The data set for | b| > 30° (chosen to avoid extensive star contamination) was subjected to an array of screening routines designed to select the subset of the data that had no definable evidence of airglow, zodiacal light, spacecraft glow, or particle contamination. Stellar signals were then removed from these data. In 86% of the data, the stellar correction was less than 20% of the observed count rate. This restricted data set provided clear evidence of a correlation with Galactic hydrogen, though this correlation was not exact. An offset was also found, in that the intensity extrapolated to zero hydrogen column was not zero but ranged (depending on view direction) from about 200 to 600 photons cm -2 s-1 str-1 Å-1 (hereafter continuum units, CU).

Subsequently, Maucherat-Joubert, Deharveng, and Cruvellier (1980) used data obtained with a far ultraviolet multichannel spectrophotometer flown on the French Astronomical Satellite D2B-Aura to search for a correlation with Galactic hydrogen. The bandpass of the spectrophotometric channel employed was 1525-1855 Å. A calcium fluoride filter eliminated the potential contamination of the data with instrumentally scattered hydrogen Lyman alpha radiation. The combination of slit width and satellite rotation gave a 1° x 2.75° field of view per integration period. Data with | b | > 40° were examined in detail. Resolved stars were removed, and the signal was corrected for dark current. The resultant data were binned in areas covering 24 to 40 square degrees, depending on the uniformity of the galactic hydrogen column in the binned area. A correlation of intensity with Galactic hydrogen was found for the overall data set and also for a subset of the data that had a high gradient of hydrogen column. The zero hydrogen column extrapolated to zero intensity was found to be 600 (+0, -250) CU.

Zvereva et al. (1982) utilized data from a spectrophotometer flown on the Soviet Prognoz 6 satellite to look for a correlation of this type. The field of view of the instrument was 6°. The bandpass selected for study covered lambda.giflambda.gif approx.gif 1500-1700 Å. Data for | b| > 15° were corrected for a small hydrogen 1216 Å signal and for stellar contributions. A correlation with Galactic hydrogen column was found with an intensity extrapolated to zero hydrogen column of 470 CU.

Weller (1983) obtained data on the far ultraviolet background with a photometer on the Solrad II satellite. This photometer had ~ 0.003 cm2 effective area and a field of view of 8°. The bandpass of the instrument was ~ 1230-1500 Å. The observations were made during a relatively brief period between other modes of the experiment. No attempt was made to search for a correlation with Galactic neutral hydrogen column although a large variation with Galactic latitude was noted. A minimum intensity of about 200 CU was found at the north and south Galactic poles.

Joubert et al. (1983) reanalyzed the far ultraviolet data obtained with the D2B satellite using more of the data (all data with | b| > 30°) as well as improved and extended 21-cm radio data. A strong correlation with neutral Galactic hydrogen was confirmed. It was noted that the distribution of data was not symmetric with respect to the regression line, primarily because of regions with ultraviolet excesses at the higher intensity levels. When these excess flux regions were excluded from the data, the intensity extrapolated to zero hydrogen column was 677 ± 20 CU.

Three photometers sensitive in the far ultraviolet were employed at the focal plane of a one-meter telescope flown on an Aries sounding rocket to study this question (Jakobsen et al. 1984). These photometers covered wavelengths from 1450 to 2400 Å in three bands. The large collecting area of this telescope provided high sensitivity, which permitted the use of a very small field of view (0.5°), virtually eliminating the problem of stellar contamination. The use of three photometers permitted an accurate appraisal of airglow and zodiacal light effects. Data from all three photometers showed a correlation with Galactic hydrogen column. The shortest wavelength photometer, which spanned the 1450-1700 Å range, required less than a 10% correction for airglow, and the potential uncorrected stellar signal was less than 10%. Data from this photometer showed a flux extrapolated to zero Galactic hydrogen column of 610 ± 60 CU.

Fix, Craven, and Frank (1989) used data obtained with an ultraviolet imaging photometer flown on the Dynamics Explorer I to study this question. This imager had a bandpass from 1360 to 1800 Å and a small field of view (0.32°). A correlation with Galactic hydrogen column was found with an intensity extrapolated to zero hydrogen column of 530 ± 80 CU.

Onaka (1990) reported on data obtained with a far ultraviolet imaging detector at the focal plane of a 17 cm Ritchy-Chretien telescope. The instrument was flown on a sounding rocket, and a scan of the Virgo cluster was carried out. A correlation of the intensity of the background flux with Galactic hydrogen column was observed, and an extrapolation of this intensity to zero hydrogen column yielded ~ 30 CU. This result was rather uncertain, however, because of the relatively small range of hydrogen column densities scanned.

Lequeux (1990) and Perault et al. (1991) carried out a very detailed reanalysis of the data obtained with the far ultraviolet imager on the French D2B-Aura satellite. Great care was taken to account for stellar contributions, and carefully normalized fine-grid 21-cm radio contours were used to generate hydrogen column maps. The resultant data display a close correlation between intensity and Galactic hydrogen with less dispersion than is exhibited in other results.

Murthy et al. (1989) reported measurements of the far ultraviolet background obtained with the Johns Hopkins far ultraviolet spectrometer that was flown on the Space Shuttle as part of NASA's UVX payload. The Johns Hopkins instrument consisted of two Ebert-Fastie scanning spectrometers fed by off-axis parabolic mirrors. The field of view of these spectrometers was 4° x 0.3°. The short wavelength spectrometer covered the spectral range from 1200 to 1700 Å with a resolution of 17 Å. The mirror for this spectrometer was 5.2 x 7.8 cm. Murthy et al. found some, but not conclusive, evidence for a spatially varying background with this instrument.

Hurwitz, Bowyer, and Martin (1991) obtained data with the Berkeley imaging nebular spectrograph, the other instrument in the NASA UVX payload. The Johns Hopkins and Berkeley instruments were coaligned and operated simultaneously. The Berkeley instrument had a field of view of 0.1° x 4°, a resolution of 15 ± 2 Å, and a bandpass from 1400 to 1850 Å. This instrument had a number of advantages for studies of the diffuse ultraviolet background. It employed a photon-counting detector with very low background, and it had a large area-solid angle product, which provided high sensitivity and allowed investigation of very low reddening view directions (the instrument was, in fact, some five orders of magnitude more sensitive to extended diffuse radiation than the instrumentation on the IUE satellite). The narrow field of view perpendicular to the dispersion direction limited the number of stars that could contribute to the spectra. Those stellar signals that were present could be easily identified and removed from the data because of the imaging capabilities of the instrument. Special care was taken in the design and fabrication of this instrument to ensure that stray light from stellar sources and out-of-band scattering would be negligible. Consequently, view directions with relatively high column densities of neutral hydrogen (and with accompanying higher stellar signals) could be investigated. Finally, because the instrument had spectroscopic capabilities, different processes contributing to the background could be identified and treated independently. The intensity of the background obtained with this instrument was found to be strongly correlated with hydrogen column at low hydrogen column densities; at higher column densities, the flux saturated and remained constant thereafter with increasing hydrogen column. The flux extrapolated to zero hydrogen column was 272 ± 13 CU.

One of the key motives for utilizing two coaligned instruments in NASA's UVX payload was to provide independent measurements of the cosmic far ultraviolet background under identical observing conditions. In Figure 1, I superimpose the data from these two instruments. The solid squares are the Berkeley data (Hurwitz, Bowyer, and Martin 1991) and the hollow triangles are the Johns Hopkins results (Murthy et al. 1989). The vertical error bars indicate statistical uncertainties (Berkeley data) or uncertainties in subtracting various backgrounds (Johns Hopkins data). Horizontal errors bars on the Johns Hopkins data, where present, indicate a range of Galactic latitude included in scanned observations. Scanned observations were treated as multiple targets by Berkeley. The Johns Hopkins data span ~ 1200-1700 Å; the Berkeley data span 1415-1835 Å, excluding high-ionization emission lines and H2 fluorescence concentrated between 1535 and 1655 Å.

Figure 1. Measurements of the diffuse far ultraviolet background obtained with the co-aligned University of California at Berkeley (UCB) and Johns Hopkins University (JHU) spectrometers flown together as the UVX experiment on the Space Shuttle. The solid squares represent UCB data; the hollow triangles JHU. The vertical error bars indicate statistical uncertainties (UCB) or uncertainties in subtracting various backgrounds (JHU). The horizontal error bars on the JHU data, where present, indicate a range of Galactic latitude included in scanned observations. Scanned observations were treated as multiple targets by UCB.

In Figure 2, the Berkeley data presented in Figure 1 are plotted against Galactic neutral hydrogen column. The intensity of this flux is seen to be strongly correlated with neutral hydrogen until saturation occurs; thereafter, the flux remains constant with increasing hydrogen column.

Figure 2. The continuum far ultraviolet intensity as obtained with the UCB spectrograph on the UVX experiment versus neutral Galactic hydrogen column. The vertical error bars are statistical uncertainties only; the horizontal uncertainties indicate the range of hydrogen column at low hydrogen column densities; at higher column densities the optical depth becomes greater than one and the background intensity remains constant with increasing hydrogen column.

A summary of all the results discussed above is shown in Table 1. The majority of these data show a correlation of intensity with Galactic latitude and/or neutral hydrogen column, indicating that the far ultraviolet background is primarily Galactic in origin. Disagreement exists regarding the slope of the correlation, the degree of scatter about the mean, and the intensity of the flux extrapolated to zero hydrogen column. These differences could reflect true astrophysical effects since a variety of mechanisms will produce a diffuse far ultraviolet flux at some level. However, it is unclear whether these observations have reached a level of accuracy that reflects true astrophysical phenomena or if these are simply experimental artifacts.

Table 1. Correlation of intensity of the far ultraviolet background with Galactic neutral hydrogen column a

Investigators Slope of correlationb Intensity at NH = 0 intercept c Coments

Paresce, McKee
& Bowyer (1980)
90 ± 10 to 140 ± 20 106 ± 60 to 570 ± 80 Apollo-Soyuz Mission.
Variation among four
view directions
Deharveng &
Cruvellier (1980)
180 ± 80 600+0-350 D2B satellite
| b | > 40°
Zvereva et al (1982) 150 470 Prognoz 6 satellite
Weller (1983) Not evaluated
(Strong correlation with Galactic latitude)
Not evaluated
180 ± 75
280 ± 88
at North and South Galactic poles
Solrad 11 satellite
Joubert et al (1983) 96 ± 6 677 ± 20 Reanalysis of
D2B satellite data
| b | > 30°
Jakobsen et al (1984) 100 ± 10 450 ± 30 Aries rocket
Fix, Craven &
Frank (1989)
~ 60 530 ± 80 Dynamics Explorer 1
Murthy et al
None found Not evaluated
100 to 300 at high Galactic latitudes
Johns Hopkins UVX
Shuttle experiment
Lequeux (1990) 96 Not evaluated Detailed reanalysis of
D2B satellite data
Onaka (1990) (30) (see text) 400 Rocket flight scan of
Virgo cluster
Hurwitz, Bowyer
& Martin (1991)
102 ± 6 272 ± 13 Berkeley UVX
Shuttle experiment

a Data reported since 1980. For earlier results, see reviews by Davidsen, Bowyer & Lampton (1974) and Paresce & Jakobsen (1980).
b Units of photons (cm2 ster Å)-1 / 1020 HI cm-2.
c Units of photons (cm2 ster Å)-1.

Finally, I return to the question of the underlying cause of the correlation of far ultraviolet intensity with some Galactic variable. It is clear that the majority of the data summarized in Table 1 are dominated by a cosec b variation with Galactic latitude, which is expected for any quantity with a plane parallel Galactic distribution. The D2B data set as analyzed by Perault et al. (1991) is of sufficient quality to establish that the far ultraviolet intensities do not correlate with cosec b as well as they correlate with Galactic neutral hydrogen column density. Hurwitz, Bowyer, and Martin (1991) found the reduced chi2 of their UVX data was three times better in a dependence with Galactic neutral hydrogen column than with Galactic latitude. As is shown in Figure 2, they also found that their data saturated at a value of NHI 1021 atoms cm-2, which provides an additional clue as to the source of the diffuse far ultraviolet flux.

Next Contents Previous