[en] The understanding of the ICRF-edge plasma interaction has increased greatly in recent years. One important result has been the virtual elimination of ICRF-specific impurities during ICRF heating on JET. It has been shown that impurity release from ICRF antennas is an effect of rf sheaths. An rf-sheath model gives good agreement with JET impurity influx data and suggests a practical prescription for reducing impurities to negligible levels. Another important development, made possible by the reduction in impurities, was the achievement of H-modes induced by ICRF heating alone under conditions in which rf sheaths were expected to be robust (monopole phasing) and to be negligible (dipole phasing). It was found that the quality, duration and termination mechanism of the H-mode was different in the two cases. In this paper, the authors describe simulation results and experimental data which provide a possible explanation for the unique properties of the monopole ICRF H-mode. The simulation shows that the rf sheaths near the antenna drive large convective cells, which penetrate the SOL and cause enhanced edge transport. The edge convection is a good candidate to explain the reduction in quality of the highly-sheared, poloidally uniform H-mode transport barrier, and the resulting flattening of the SOL density and temperature profiles may explain the longer duration and the absence of a radiative collapse for the monopole H-mode. An analysis of JET data shows the existence of phasing-dependent ICRF-enhanced edge transport in agreement with the theoretical model. The 2D simulation results are discussed and compared with the available experimental data. This work suggest that ICRF antennas in monopole or FWCD phasing provide another practical means of biasing the edge plasma to control the global confinement, central density and impurity accumulation in H-mode operation