“…Though very slow, this process should be significant at very low temperature. At this point, one can stress that this process is different from quantum tunneling, 4,[60][61][62] which is also invoked in a magnetization switch at very low temperatures; indeed, in contrast to quantum tunneling, the present process involves (1) a collision with an electron and (2) a dissipation, i.e., an energy change between the initial and final magnetization states that is balanced by heating a substrate electron. In a classical view, there is still a barrier separating the two magnetization states, lowered by the E contribution; the system does not tunnel through the barrier at constant energy, but, due to the symmetry lowering introduced by E, collision with a substrate electron is able to induce a direct transition between the two magnetization states.…”