Shocks or photoionization: direct temperature measurements of the low-ionization gas in quiescent galaxies

The ionization mechanism of the low-ionization gas in quiescent red-sequence galaxies has been a long-standing puzzle. Direct temperature measurements would put strong constraints on this issue. We carefully selected a sample of quiescent red-sequence galaxies from Sloan Digital Sky Survey. We bin them into three bins with different [N II]/H α and [N II]/[O II] ratios, and we measure the temperature-sensitive [O III] λ4363, [N II] λ5755, [S II] λλ4068, 4076, and [O II] λλ7320, 7330 lines in the stacked spectra. The [S II] doublet ratios indicate the line-emitting gas is in the low-density regime (∼10-200 cm^-3). We found the temperatures in the S⁺ zones to be around 8000 K, the temperatures in the O⁺ zone to be around 1.1-1.5 × 10⁴ K, and the temperatures in the N⁺ zones to be around 1-1.4× 10⁴ K. The [O III] λ4363 line is not robustly detected. We found that the extinction corrections derived from Balmer decrements would yield unphysical relationships between the temperatures of the S⁺ zones and O⁺ zones, indicating that the extinction is significantly overestimated by the measured Balmer decrements. We compared these line ratios with model predictions for three ionization mechanisms: photoionization by hot evolved stars, shocks, and turbulent mixing layers. For the photoionization and shock models, the hot temperatures inferred from [S II] and [N II] coronal-to-strong line ratios require metallicities to be significantly subsolar. However, the [N II]/[O II] line ratios require them to be supersolar. None of the models could simultaneously explain all of the observed line ratios, neither could their combinations do.