Astron. Astrophys. 321, L17-L20 (1997)

4. Abundance variations and nucleosynthesis

In broad terms, the composition of Sakurai's object shows evidence of severe contamination by material exposed to hydrogen and helium burning and associated nuclear reactions. Close examination provides some interesting constraints on the nucleosynthesis experienced by the star.

The present atmosphere is not a simple mix of initial unprocessed gas, gas run through the H-burning CNO-cycles, and H-exhausted gas exposed to He-burning, but must have been accompanied by further processing. This is demonstrated by the low observed ¹² C/¹³ C ratio, which encompasses the equilibrium value of 3.5 for CNO-cycling. As the equilibrium abundance of ¹³ C is very low following He-burning, the observed ratio suggests that ¹² C from He-burning has been exposed to hot protons. It would seem that C-rich material from He-burning has been mixed with ingested hydrogen such that the proton supply is effectively exhausted in converting inhibited (see Renzini 1990).

Not all protons are consumed in He-rich regions. Production of lithium is ascribable to the Cameron-Fowler (1971) mechanism. Here ³ He synthesised in a low mass main sequence star is converted to ⁷ Li in an envelope that convects ⁷ Li to low temperatures where it survives until re-exposed to high temperatures. Production of lithium implies H-burning in regions not previously exposed to H-burning temperatures; ³ He which is destroyed in regions that have undergone H-shell or H-core burning can hardly be resynthesised. The observed Li is not a fossil from an earlier stage as a Li-rich AGB star: the predicted Li/H ratio for AGB stars which have undergone hot-bottom burning is 10^-8 while the observed ratio is 10^-5 to 10^-6 and hydrogen consumption necessarily destroys fossil lithium. The overabundant Na and Al have likely been synthesised through ²² Ne(p, )²³ Na and ²⁵ Mg(p, )²⁶ Al. As in other H-deficient stars Ne is very high, which can not be explained by -captures on N from initial CNO, but must be due to products of He-burning and possibly additional CNO-cycling.

A remarkable feature of Sakurai's object is the large overabundance of light s -process elements and the high ratio of light to heavy s -process elements. Probably, ¹³ C( O is the neutron source. The s -processing may be characterized by the neutron exposure . We find a good fit to the abundances from Ni to La for October with mb^-1 using Malaney's (1987) predictions for a single exposure. The Rb abundance indicates a low neutron density of cm^-3 (Malaney 1987), while no useful limit could be set on the Tc abundance. An exponential distribution of exposures provides less good agreement with the observed abundances. For the R CrB star U Aqr, which also shows light s -element enhancements, Bond et al. (1979) obtained mb^-1. Such exposures imply that about 10 neutrons were captured by each Fe seed nucleus. Given that the observed ratio ¹³ C/Fe , the exposure, even in the presence of neutron poisons such as ¹⁴ N, seems an achievable goal. The fact that Ni, Cu and Zn are well fit by the predictions indicates that the envelope consists mostly of material exposed to neutrons. This fact also likely explains the anomalous high ratios of K/Fe and Sc/Fe.

The final He-shell flash may occur in a luminous post-AGB star or the white dwarf that evolves from the post-AGB star. In the latter case, hydrogen may be mixed with deep layers of He and C and consumed. In contrast, the H-burning layer in the post-AGB star prevents deep mixing. About 10% of all AGB stars may experience their final He-shell flash as a white dwarf and, if H consumption is severe, may convert the born-again AGB star to an R CrB star (Renzini 1990; Iben et al. 1996). Iben & MacDonald (1995) have presented a model in which mixing and nucleosynthesis were followed: their chosen model of a 0.6 star ended with an outer layer having the abundance ratios (by number of atoms) H/He , C/He , N/C , and O/C . This resembles the composition of Sakurai's object, apart from the predicted H deficiency not being as severe as observed. The model O/C ratio is lower than observed but might be raised by adjustment of the uncertain rate for the reaction ¹² C()¹⁶ O. In summary, final flash models offer a tantalising prospect of accounting for the observed composition of Sakurai's object.

Life as a born-again AGB star is brief: the model by Iben & MacDonald (1995) brightens by a factor of 10 and cools from of 40,000K to 6300K in just 17 yr. Evolution over the narrow temperature range covered by Sakurai's object between May and October is, of course, much faster. The evolutionary timescale seems similar to that of V605 Aql (Lundmark 1921). The timescale for compositional changes for Sakurai's object is likely even shorter; processed material rising from below will mix very quickly with the atmosphere. The other final flash candidate FG Sge has also showed rapid abundance alterations, e.g. some s -process elements increased by 0.8 dex in 7 years (Langer et al. 1974).

It is now important to extend the few available calculations of the final flash to a wider range of initial conditions, and to include Li-production and s -processing. The hints that the surface composition is evolving rapidly must be pursued by continued spectroscopic observations, which may shed further light on its evolutionary status and relation to the R CrB stars; we may have witnessed the birth of an R CrB star. Furthermore, a determination of the nebular composition would reveal the original composition of Sakurai's object prior to the final flash. Monitoring the visual variability of the star searching for R CrB-like declines is naturally of importance.

© European Southern Observatory (ESO) 1997

Online publication: June 30, 1998
helpdesk.link@springer.de