This element was discovered independently by J. Priestley in Leeds, England in 1774 and by C. Schecle in Uppsala, Sweden in 1771. The name comes from the Greek oxy genes (acid forming).
OI 13.6 eV, OII 35.1 eV, OIII 54.9 eV, OIV 77.4 eV, OV 113.9 eV, OVI 138.1 eV, OVII 739 eV, OVIII 871 eV.
Absorption lines of 0I
Source: Data are from Jaschek et al. (1992). The data are averages and refer to the whole blend 7772-7774-7775. The * data are from Arellano Ferro et al. (1991). The data refer to the 7772 line alone.
OI (7772) appears in B-type stars, has a maximum around A 5 and declines toward later stars. It has a pronounced positive luminosity effect.
The luminosity effect was first noticed by Keenan and Hynek (1950) and has since then been used frequently for calibration purposes. For instance Arellano Ferro et al. (1991) have measured equivalent widths of OI 7774 in F and G stars to obtain a quadratic relationship between W(7774) and absolute magnitude. Faraggiana et al. (1988) have measured equivalent widths for early type stars.
The infrared triplet at 7772 is among the strongest lines of the chromospheric spectrum of the sun (Pierce 1968).
Another important line is (OI) 6300. The line is seen in absorption in late type giants and has been used very often for abundance studies (see for instance Barbuy (1988)).
The resonance line of OI (UV2) is at 1302.
Emission lines of 0I
In Oe stars the feature at 7774 is seen in emission (Andrillat et al. 1982).
In Be stars, the 7772(1) and 8446(4) lines are often seen in emission (Polidan and Peters 1976, Andrillat et al. 1990) with W(8446) < 3 Å. Sometimes 8446 is much stronger than 7772, which fact is attributable to fluorescence of Lyman beta. Jaschek et al. (1993) found that O1 7772 is always in emission in Be stars (W(7772) < 1.4 Å). The line is also seen in very strong emission in Herbig Be-Ae and in B[e] stars. The lines 7772 and 8446 do not always show the same behavior (Praderie et al. 1987). In B[e] stars one observes [OI] in addition.
In compact infrared sources 8446 is usually in emission (McGregor et al. 1988) and the same happens in symbiotic stars (Kenyon and Fernandez-Castro 1987).
The ultraviolet 1152(6) line is visible in the solar chromospheric spectrum (Feldman and Doschek 1991). This line is also observed in other cool stars, where it indicates the existence of a chromosphere. The line is characteristicof `non-coronal' stars. See Part Two, section 3.1. [OI] lines appear in emission around minimum light in long-period variables (Querci 1986), in planetary nebulue, T Tau stars, novue, supernovue, gaseous nebulue and symbiotic stars.
Attention is called to the fact that [OI] lines also occur in the atmospheric air-glow and in auroras of the earth.
Absorption lines of 0II
OII (for instance 4076) appears in late O-type stars, has a maximum at B 1 and disappears around B 3. A strong positive luminosity effect exists (see also 4649 and 4415).
Forbidden lines of 0II
Forbidden lines of OII appear in emission in B[e] stars. For discussion of a typical object, see Swings (1973).
T Tau stars often exhibit 3726 and 3729 in emission (Sun et al. 1985).
Absorption lines of 0III
The only supergiant is plotted as a cross in the figure.
OIII lines (5592) are strongest for types O 6-O 7 and show a positive luminosity effect. The lines disappear at about B 0. The OIII line 3759 has W = 0.30 for O 7.
OIII lines (M.3 and 15) can be excited by fluorescence from HeII Lyman alpha. These lines are prominent in emission in symbiotic stars (Merrill 1950, Wallerstein et al. 1991) and in one recurrent nova (Joy and Swings 1945).
Forbidden lines of 0111
T Tau stars sometimes exhibit the lines 5007 and 4959 in emission (Sun et al.1985). These two [OIII] lines are usually the strongest nebular lines. They occur in objects like planetary nebulue, novae and gaseous nebulue, and in post-maximum spectra of novae and symbiotic stars.
Absorption lines of OIV and OV
Source: Data are from Prinja (1990) and are average values derived from MK standards.
OIV is present in O-type stars (Dean and Bruhweiler 1985). The ultraviolet line 1371(7) of OV decreases from O 3 to O 7, where it disappears (Heck et al.1984).
Emission lines of OV
The presence of OV 1371(7) indicates the existence of a transition region, characteristic of coronal stars (see Part Two, section 3.1).
One OV emission line at 1.5541 µm is observed weakly in WC stars (Eenens et al. 1991).
Absorption lines of OVI
The blend at 1035 (1032-1037) (W M.1) is present in early O-type stars, whereas in Be stars it remains visible up to O 9.5 (Marlborough 1982).
Emission lines of O VI
OVI lines in emission (centered on broad absorption lines) are characteristic of pre-degenerates (Motch et al. 1993). The resonance lines 1032 and 1037 are seen in very strong emission in the solar ultraviolet spectrum (Feldman and Doschek 1991). Their presence in stars (for instance in symbiotic stars) can be inferred from the presence of strong fluorescence ultraviolet Fe II emission lines (Feibelman et al. 1991).
Emission lines of OVIII
Possibly the 6068 line is present in emission in pre-degenerates, according to Motch et al. (1993).
The Wolf-Rayet oxygen sequence
Barlow and Hummer (1982) have introduced a small subgroup of the WR sequence characterized by a very strong emission of OVI 3811 and OIV at 3400 (a blend of 3381 and 3426-M.2 and 3). Eenens and Williams (1991) have recently studied one of these objects and also found OV in emission. Nugis and Niedzielski (1990) suggest that the lines of OVI may be present in many WN stars, but that these lines are masked by blends of other lines.
Behavior in non-normal stars
OII lines are about normal in the N variety of the CNO stars (also called OBN stars by some authors) (Schoenberner et al. 1988). Walborn (1980) reported that OIII is reinforced in CNO stars of N type.
Both OV and OVI lines are present in the spectra of pre-degenerates (Werner et al. 1991, Werner 1991).
O is weak in some sdO and sdB stars (Baschek and Norris 1970, Baschek et al. 1982).
Gerbaldi et al. (1989) and Roby and Lambert (1990) have investigated the behavior of OI lines in Bp and Ap stars. They find that OI is somewhat weak in stars of the Hg-Mn and Si subgroups, whereas in stars of the Cr-Eu-Sr subgroup it is very weak. OI lines are somewhat weak (Roby and Lambert 1990) in Am stars. In lambda Boo stars, O is about normal (Baschek et al. 1984).
O seems to be slightly weak in galactic supergiants by a factor of two, but it is not certain that this underabundance is real (Luck and Lambert 1992).
O seems to be underabundant in blue stragglers of an old open cluster (Mathys 1991).
O is overabundant with respect to iron in metal-weak (late) giants by factors of the order of two, both in halo stars and in globular cluster giants (Brown et al. 1991, Gonzalez and Wallerstein 1992) (see also Part Two, section 2.1). It has to be remarked, however, that these determinations are delicate because they are based either upon the analysis of forbidden O lines or upon molecules (TiO, CO). In the latter case one has to assume something about the behavior of the other elements that enter into the molecule. This has produced discrepancies between the results of different authors, which have not been completely cleared up (Barbuy 1992). For the use of CO see for instance Tsuji (1991) and for that of forbidden oxygen lines, Spite and Spite (1991).
Oxygen-rich giants are usually surrounded by circumstellar shells, which are rich in oxygen, characterized by silicate emission features peaking at 10 and 291 µm.
O is strengthened in R CrB stars (Cottrell and Lambert 1982) and seems to have a normal abundance in C stars (Lambert et al. 1986), in Ba stars (Lambert 1985) and in S and MS stars.
O is probably deficient within Magellanic Cloud supergiants (Lennon et al. 1991) but some authors find that it behaves in a manner parallel to that of iron (Barbuy et al. 1981, Spite and Spite 1990). Luck and Lambert (1992) find that this is true only for the Small Magellanic Cloud, whereas in the Large Magellanic Cloud the O/Fe ratio may differ from star to star.
OI absorption lines are strong in the principal spectrum phase of novae, whereas OII lines are strong in the Orion phase. Lines of [OI] (6300 and 6363) and [OIII] appear in emission in the same phase, accompanied in the ultraviolet region by emission lines of O III, O IV and OV. Sometimes an `OI flash' appears before the Orion phase, during which the OI lines rival in intensity the Balmer lines. Toward the end of the Orion phase there appears the [OIII] flash. Both [OI] and [OIII] lines are present in emission in the nebular stage (Payne-Gaposchkin 1957, Warner 1989). O is overabundant by one order of magnitude in the spectra of the C-O nova subgroup (Andreae 1993).
OI appears in absorption in the spectra of supernovae of type Ib and in the nebular stages of types Ib and II (Branch 1990). The intensification of O is sometimes so great that a special type of supernova, called Ic, has been created (Hill 1993).
Oxygen has three stable isotopes: O 16, 17 and 18. In the solar system their respective frequencies of occurrence are 99.7%, 0.04% and 0.2%. There exist also five unstable isotopes.
Harris et al. (1988) studied the isotope ratio in five K-type giants, using the CO molecules formed with the different oxygen isotopes in the 5 µm region. The result is that the O16 / O17 ratio is about 400 and the O16 / O18 ratio about 500. The solar system values are about 2600 and 500 respectively.
Dominy (1984) has searched for isotope anomalies in early C-type stars and found no deviation from solar system values.
Harris et al. (1987) re-analyzed the early C stars, using the CO bands at 4289 cm-1. They found O16 / O17 ratios between 550 and 4100 and O16 / O18 ratios between 700 and 2400.
stars have been studied by the same authors, who found O16 / O17 ratiosbetween 350 and 850.
Harris et al. (1985a) have studied these ratios in Ba stars and found O16 / O17 ratios btween 100 and 500 and O16 / O18 ratios between 60 and 550.
In yet another study, Harris et al. (1985b) investigated some MS and S stars. They found O16 / O17 ratios between 500 and 3000 and O16 / O18ratios between 850 and 4600.
It seems clear that it is difficult to draw general conclusions from these values, which are scattered rather widely.
For completeness it should be added that isotopic ratios can also be determined radioastronomically from a variety of molecules - for instance CO and CS (Kahane et al. 1992).
O16 iS produced by He burning, O17 by hydrogen burning or by explosive hydrogen burning and O18 by He or N burning.
Published in "The Behavior of Chemical Elements in Stars", Carlos Jaschek and Mercedes Jaschek, 1995, Cambridge University Press.