I worked with Dr. Donald Garnett of Steward Observatory for this research. I began this project in June 2003, and finished in April 2005. The intent of this project was to shed light on the problem of chemical abundances in planetary nebulae (PNe). The ionic abundance determined from the spectrum of recombination lines of an ion is often much greater than the abundance determined from the forbidden (collisionally excited) spectral lines. The abundance discrepancy is not consistent; in some PNe the abundance determined from the recombination lines is equal to that from forbidden lines, within observational error, while in other PNe it can be over 20 times greater. Since the recombination lines are considerably fainter than the forbidden lines, until recently most spectra of PNe were not deep enough to accurately measure intensities for the recombination lines. Using 8 new deep spectra of 6 PNe, we analyzed the chemical abundances, with a focus on the ion O+2, which has a rich optical recombination spectrum. Although I was not involved in the observing run at the telescope that collected the data, I measured the spectral lines and performed the data analysis.

From the literature, we found 16 more PNe that were recently observed with high resolution spectra, and in combination with our PNe, we looked for correlations between the abundance discrepancy and various nebular properties, such as electron temperature, density, stellar luminosity, and about 15 other parameters. We found strong correlations with the surface brightness and nebular diameter, and moderate correlations with the electron density and expansion velocity. This suggests that the evolutionary state of the nebula is related to the abundance discrepancy. Also in our work we showed some progress in deriving abundances for transitions where dielectronic recombination dominates. Our results were published in the Astrophysical Journal, and I presented a poster on the work at an American Astronomical Society conference.

I did some work on a second phase of that project, where I investigated the abundance discrepancy as a function of radial position, to see if the discrepancy has any structure. We took spectra of three PNe with the Hubble Space Telescope and analyzed them for abundances across the diameter of the nebulae.

Download the full published paper, 463k PDF file

Abstract:
Recombination lines (RLs) of C II, N II, and O II in planetary nebulae (PNs) have been found to give abundances that are much larger in some cases than abundances from collisionally excited forbidden lines (CELs). The origins of this abundance discrepancy are highly debated. We present new spectroscopic observations of O II and C II recombination lines for six planetary nebulae. With these data we compare the abundances derived from the optical recombination lines with those determined from collisionally excited lines. Combining our new data with published results on RLs in other PNs, we examine the discrepancy in abundances derived from RLs and CELs. We find that there is a wide range in the measured abundance discrepancy Delta(O+2) = log{O+2(RL)} - log{O+2(CEL)}, ranging from approximately 0.1 dex (within the 1 sigma measurement errors) up to 1.4 dex. This tends to rule out errors in the recombination coefficients as a source of the discrepancy. Most RLs yield similar abundances, with the notable exception of O II multiplet V15, known to arise primarily from dielectronic recombination, which gives abundances averaging 0.6 dex higher than other O II RLs. We compare Delta(O+2) against a variety of physical properties of the PNs to look for clues as to the mechanism responsible for the abundance discrepancy. The strongest correlations are found with the nebula diameter and the Balmer surface brightness; high surface brightness, compact PNs show small values of Delta(O+2), while large low surface brightness PNs show the largest discrepancies. An inverse correlation of Delta(O+2) with nebular density is also seen. A marginal correlation of Delta(O+2) is found with expansion velocity. No correlations are seen with electron temperature, He+2/He+, central star effective temperature and luminosity, stellar mass-loss rate, or nebular morphology. Similar results are found for carbon in comparing C II RL abundances with ultraviolet measurements of C III].