Da. Turnshek et al., THE HST QUASAR ABSORPTION-LINE KEY PROJECT .4. HST FAINT OBJECT SPECTROGRAPH AND GROUND-BASED OBSERVATIONS OF THE UNUSUAL LOW-REDSHIFT BROAD ABSORPTION-LINE QUASI-STELLAR OBJECT PG-0043+039, The Astrophysical journal, 428(1), 1994, pp. 93-112
HST FOS observations have shown that the spectrum of the low-redshift
(z(em) congruent-to 0.384) QSO PG 0043 + 039 exhibits weak broad absor
ption lines (BALs). The BALs were discovered during the course of UV s
pectrophotometry made for the HST Quasar Absorption Line Key Project.
The HST data are analyzed along with ground-based optical and IUE spec
trophotometry. The object is found to have a number of atypical proper
ties relative to normal non-BAL QSOs. For example, the object is one o
f the strongest Fe II emitters known, narrow-line [O III] and [O II] e
mission is not detected, and the C IV broad emission line is exceeding
ly weak. The observed continuum is atypical in the sense that it is mu
ch weaker than that of a normal optically selected QSO at rest wavelen
gths less than or similar to 2200 angstrom. Intrinsic reddening of E(B
- V) congruent-to 0.11 mag by dust similar to that found in the SMC a
t the redshift of PG 0043 + 039 conservatively accounts for the observ
ed continuum shape moderately well. These observed characteristics are
typical of low-ionization BAL QSOs, but convincing evidence for BALs
due to low-ionization transitions of Mg II, Al III, Al II, or C II doe
s not exist. Therefore, this object may be a misaligned BAL QSO having
many of the characteristics of low-ionization BAL QSOs with the sight
line passing through a putative dusty region, but evidently missing c
louds of high enough column density to produce observable low-ionizati
on BALs. If the intrinsic dust-extinction model is correct, the observ
ations suggest that the dust is not confined to the presumably higher
density, low-ionization BAL clouds, but that it has drifted to nearby
high-ionization BAL regions. We also consider other possible mechanism
s for producing the shape of the continuous energy distribution which
cannot be ruled out. We compare the Fe II emission in PG 0043 + 039 wi
th that in another Key Project QSO, NGC 2841-UB 3, which has optical F
e II emission comparable in strength to that in PG 0043 + 039, but has
anomalously weak UV Fe II emission. In addition, from an analysis of
UV and optical spectrophotometric data at 5 epochs over approximately
11 yr, there is tentative evidence that PG 0043 + 039 has varied in br
ightness by as much as 1.1 mag during this time interval. Two differen
t interpretations of PG 0043 + 039 and the low-ionization BAL QSOs are
considered: One is that the low-ionization BAL QSOs comprise a distin
ct population of QSOs that have large BAL region covering factor, no a
ppreciable narrow emission-line region, and other atypical properties.
An alternative possibility is that PG 0043 + 039 has a small BAL regi
on covering factor, which is consistent with previously proposed model
s for most BAL QSOs, but that nonspherical geometries coupled with mec
hanisms for anisotropic emission give rise to aspect-angle dependent Q
SO spectral properties, such as weak narrow-line [O III] emission. Whi
le the actual value of the BAL region covering factor in PG 0043 + 039
and low-ionization BAL QSOs is unclear, given the absence of a specif
ic model which would give rise to anisotropic continuum or line emissi
on of the type required, the evidence presented here would seem to ind
icate that the BAL region covering factor in a subclass of QSOs is lar
ger than resonance line scattering models generally indicate. Various
model scenarios for explaining the weak narrow-line [O III] emission a
re considered, but there is no definitive explanation.