CHAOS I. DIRECT CHEMICAL ABUNDANCES FOR H II REGIONS IN NGC 628

The CHemical Abundances of Spirals (CHAOS) project leverages the combined power of the Large Binocular Telescope (LBT) with the broad spectral range and sensitivity of the Multi Object Double Spectrograph (MODS) to measure direct abundances (based on observations of the temperature-sensitive auroral lines) in large samples of H II regions in spiral galaxies. We present LBT MODS observations of 62 H II regions in the nearby spiral galaxy NGC 628, with an unprecedentedly large number of auroral lines measurements (18 [O III]lambda 4363, 29 [N II]lambda 5755, 40 [S III]lambda 6312, and 40 [O II]lambda lambda 7320, 7330 detections) in 45 H II regions. Comparing derived temperatures from multiple auroral line measurements, we find: (1) a strong correlation between temperatures based on [S III]lambda 6312 and [N II]lambda 5755; and (2) large discrepancies for some temperatures based on [O II]lambda lambda 7320, 7330 and [O III]lambda 4363. Both of these trends are consistent with other observations in the literature, yet, given the widespread use and acceptance of [O III]lambda 4363 as a temperature determinant, the magnitude of the T[O III] discrepancies still came as a surprise. Based on these results, we conduct a uniform abundance analysis prioritizing the temperatures derived from [S III]lambda 6312 and [N II]lambda 5755, and report the gas-phase abundance gradients for NGC 628. Relative abundances of S/O, Ne/O, and Ar/O are constant across the galaxy, consistent with no systematic change in the upper IMF over the sampled range in metallicity. These alpha-element ratios, along with N/O, all show small dispersions (sigma similar to 0.1 dex) over 70% of the azimuthally averaged radius. We interpret these results as an indication that, at a given radius, the interstellar medium in NGC 628 is chemically well-mixed. Unlike the gradients in the nearly temperature-independent relative abundances, O/H abundances have a larger intrinsic dispersion of similar to 0.165 dex. We posit that this dispersion represents an upper limit to the true dispersion in O/H at a given radius and that some of that dispersion is due to systematic uncertainties arising from temperature measurements.