To explore the mechanism, a series of control experiments were carried out. The radical scavenger 2, 2, 6, 6-tetramethylpiperidinooxy (TEMPO) and 1, 1-diphenylethylene completely inhibited the reaction, providing adduct
5 (see Sections 5.1 and 5.2 in Supporting information), indicating a difluoroalkyl radical is involved. Failure of difluoroalkylation with steric hindered
6 suggests that the probable coordination of ruthenium with anilides would trigger the reactions (
Scheme 4a) [
51]. According to previous investigation, three modes were proposed for ruthenium-substrate coordination (
Scheme 4b), such as (ⅰ) internal chelation-assisted cycloruthenation
7 [
27-
29,
48], (ⅱ) cross-over chelation-assisted cycloruthenation
8 [
61], and (ⅲ) ruthenium complex coordination with substrate
9 [
31,
50]. D/H exchange was not detected by NMR analysis in deuterium experiments (
Scheme 4c) and cross-over H/D exchange experiments (
Scheme 4d), indicating the cycloruthenation didn't occur. Thus, coordination mode
9, similar to the Frost's and our observations [
31,
50], was reasonable for this catalytic process, (
Scheme 4b). Indeed, XPS spectroscopy investigations further supported mode
9. N 1s peaks can be divided into two peaks at 399.5 eV (C—N bond) and 400.8 eV (Ru-N bond), demonstrating the coordination of ruthenium with nitrogen (Fig. S1 in Supporting information) [
62]. For XPS spectrum of O 1s, four peaks were identified for the Ru-O = C, P-O, C—O, and Ru-O at binding energies 531.0, 531.8, 532.1, and 533.3 eV, respectively, showing coordination of ruthenium with oxygen of acyl group (Fig. S1) [
63,
64]. Furthermore, the absence of either nitrogen or oxygen of amide groups changed the reaction site, confirming that the coordination mode
9 could afford an electron-deficient (hence more active) arene core and is essential to controlling regioselectivity (see Section 5.7 in Supporting information) [
31,
51]. Without
p-cymene was observed indicating ruthenium precatalyst without decoordination (see Sections 5.8 in Supporting information) [
65,
66].