When analysing H ii regions, a possible source of systematic error on empirically derived quantities, such as the gas temperature and the chemical composition, is the limited size of the slit used for the observations. In order to evaluate this type of systematic error, we use the photoionization code aangaba to create a virtual photoionized region and mimic the effect of a slit observation. A grid of models was built varying the ionizing radiation spectrum emitted by a central stellar cluster, as well as the gas abundance. The calculated line surface brightness was then used to simulate slit observations and to derive empirical parameters using the usual methods described in the literature. Depending on the fraction of the object covered by the slit, the empirically derived physical parameters and chemical composition can be different from those obtained from observations of the whole object. This effect is mainly dependent on the age of the ionizing stellar cluster. The low-ionization lines, which originate in the outer layers of the ionized gas, are more sensitive to the size of the area covered by the slit than the high-ionization forbidden lines or recombination lines, because these lines are mainly produced closer to the inner radius of the nebula. For a slit covering 50 per cent or less of the total area, the measured [O iii], [O ii] and [O i] line intensities are less than 78, 62 and 58 per cent of the total intensity for a young H ii region (t < 3 Myr); for older objects the effect due to the slit is less significant. Regarding the temperature indicator T[O iii], the slit effects are small (usually less than 5 per cent) because this temperature is derived from [O iii] high-ionization lines. On the other hand, for the abundance (and temperature) indicator R23, which depends also on the [O ii] line, the slit effect is slightly higher. Therefore, the systematic error due to slit observations on the O abundance is low, being usually less than 10 per cent, except for H ii regions powered by stellar clusters with a relative low number of ionizing photons between 13.6 and 54.4 eV, which create a smaller O++ emitting volume. In this case, the systematic error on the empirical O abundance deduced from slit observations is more than 10 per cent when the covered area is less than 50 per cent.
Read full abstract