1999ApJ...525..691S

We study the characteristics of molecular gas in the central regions of spiral galaxies on the basis of our CO(J=1-0) imaging survey of 20 nearby spiral galaxies using the NRO and OVRO millimeter arrays. Condensations of molecular gas at galactic centers with size scales ≲1 kpc and CO-derived masses M_gas(R<500 pc)∼10⁸-10⁹ M_☉ are found to be prevalent in the gas-rich∼L^* galaxies. Moreover, the degree of gas concentration to the central kiloparsec is found to be higher in barred systems than in unbarred galaxies. This is the first statistical evidence for the higher central concentration of molecular gas in barred galaxies, and it strongly supports the theory of bar-driven gas transport. It is most likely that more than half of molecular gas within the central kiloparsec of a barred galaxy was transported there from outside by the bar. The supply of gas has exceeded the consumption of gas by star formation in the central kiloparsec, resulting in the excess gas in the centers of barred systems. The mean rate of gas inflow is statistically estimated to be larger than 0.1-1 M_☉.yr^–1. There is no clear correlation between gas mass in the central kiloparsec and the type of nuclear spectrum (H II, LINER, or Seyfert), suggesting that the amount of gas at this scale does not determine the nature of the nuclear activity. There is, however, a clear correlation for galaxies with larger gas-to-dynamical mass ratios to have H II nuclear spectra, while galaxies with smaller ratios show spectra indicating active galactic nuclei (AGNs). This trend may well be related to the gravitational stability of the nuclear gas disk, which is generally lower for larger gas mass fractions. It is therefore possible that all galaxies have active nuclei, but that dwarf AGNs are overwhelmed by the surrounding star formation when the nuclear molecular gas disk is massive and unstable. The theoretical prediction of bar dissolution by condensation of gas to galactic centers is observationally tested by comparing gas concentration in barred and unbarred galaxies. If a bar is to be destroyed so abruptly that the gas condensation at the nucleus does not have enough time to be consumed, then there would be currently unbarred but previously barred galaxies with high gas concentrations. The lack of such galaxies in our sample, together with the current rates of gas consumption at the galactic centers, suggests that the timescale for bar dissolution is larger than 10⁸-10¹⁰ yr or a bar in a L^* galaxy is not destroyed by a condensation of 10⁸-10⁹ M_☉ gas in the central kiloparsec.