On variations in the fine-structure constant and stellar pollution of quasar absorption systems

At redshifts z_abs < 2, quasar absorption-line constraints on space-time variations in the fine-structure constant, alpha, rely on the comparison of MgII and FeII transition wavelengths. One potentially important uncertainty is the relative abundance of Mg isotopes in the absorbers which, if different from solar, can cause spurious shifts in the measured wavelengths and, therefore, alpha. Here we explore chemical evolution models with enhanced populations of intermediate-mass (IM) stars which, in their asymptotic giant branch (AGB) phase, are thought to be the dominant factories for heavy Mg isotopes at the low metallicities typical of quasar absorption systems. By design, these models partially explain recent Keck/HIRES evidence for a smaller alpha in z_abs < 2 absorption clouds than on Earth. However, such models also over-produce N, violating observed abundance trends in high-z_abs damped Lyman-alpha systems (DLAs). Our results do not support the recent claim of Ashenfelter, Mathews & Olive (2004b) that similar models of IM-enhanced initial mass functions (IMFs) may simultaneously explain the HIRES varying-alpha data and DLA N abundances. We explore the effect of the IM-enhanced model on Si, Al and P abundances, finding it to be much-less pronounced than for N. We also show that the 13C/12C ratio, as measured in absorption systems, could constitute a future diagnostic of non-standard models of the high-redshift IMF.