We stacked 12CO and 13CO spectra of NGC 3627 after redefining the velocity axis of each spectrum of the mapping data so that the zero corresponds to the local mean velocity of 12CO spectra. The signal-to-noise ratios of the resulting spectra are improved by a factor of up to 3.2 compared to those obtained with normal stacking analysis. We successfully detect a weak 13CO emission from the interarm region where the emission was not detected in the individual pointings. We compare the integrated

more »

... tensity ratios I12 CO/I13 CO among six characteristic regions (center, bar, bar-end, offset, arm, and interarm). We find that I12CO/I13CO in the bar and interarm are higher than those in the other regions by a factor of ~2 and I12CO/I13CO in the center is moderately high. These high I12CO/I13CO ratios in the bar and center are attributed to a high intensity ratio (T12CO/T13CO) and one in the interarm is attributed to a high ratio of the full width at half maximum of spectra (FWHM12CO/FWHM13CO). The difference between FWHM12CO and FWHM13CO of the interarm indicates the existence of two components, one with a narrow line width (~FWHM13CO) and the other with a broad line width (~FWHM12CO). Additionally, the T12CO/T13CO ratio in the broad-line-width component of the interarm is higher than the other regions. The high T12CO/T13CO in the center and bar and of the broad-line-width component in the interarm suggest the existence of non-optically thick 12CO components. We find that more than half of the 12CO emissions of the interarm are likely to be radiated from the diffuse component. Our result suggests that the use of a universal CO-to-H2 conversion factor might lead to an overestimation of molecular gas mass and underestimation of star-formation efficiency in the interarm by a factor of a few.