From kinetic measurements by optical second harmonic generation of 02 interaction with Si(111)-(7 X 7) at temperatures between 610 and 735 'C and pressures from 10 9 to 10 6 tort, we determine the boundaries for oxide nucleation and for quasiequilibrium between surface phases of oxide and silicon. The distinction between these boundaries rejects the influence of the kinetic parameters of the surface reactions on the critical conditions. At T~700 C, the oxide nucleation requires oxygen pressure well above that required for oxide growth.The study of silicon oxidation near critical conditions has been advanced by recent investigations [1 -4]. The reactions of oxygen with silicon at high temperatures and low 02 pressures is characterized by two regimes: gas phase etching with volatile SiO formation [5] and oxide growth on the surface. The boundary between the regimes is referred to as the critical condition and is described in terms of the critical oxygen pressure (P,), which is a function of the surface temperature (T,), by the equation P, (T,) = Po exp( -AE/kT, ) [6,7]. The activation energy AE and the prefactor I'o were practically independent on the crystallographic orientation of the surface. These values are /s. E = 3.93 eV and Po = 4.4 X 10' torr for the Si(111) surface [7]. However, the critical conditions were shown to depend on the nature of the oxidant; in the case of the Si-N20 interaction, the critical conditions were shifted to much higher pressures and were described by significantly different parameters of AE = 2.14 eV and Po --1.5 X 10 torr [8].Two distinct approaches have been suggested for understanding the existence of critical conditions. In the first, the critical pressure of 02 is derived from an equality between the 02 How adsorbed by the silicon surface and the SiO IIow leaving the surface [9]. This approach has given a semiquantitative agreement with experimental results [6,10]. An alternative approach, referred to as a kinetic model, has been used to describe the critical conditions deriving from an equality between the growth and decay rates of Si02 clusters on the Si surface [7]. Recently, the initial stage of Si oxidation was considered to be a first-order phase transition [3]. An unexpected result obtained with scanning tunneling microscopy (STM) [2,4] and reliection high-energy electron diffraction [1] was the finding that the critical conditions were shifted compared with that expected by extrapolating high temperature data [7]. This shift Feltz, Memmert, and Behm [4] connected with the procedure of data acquisition; with decreasing pressures the times required to build up detectable amounts of surface oxide become much longer than the experimental waiting times at temperatures below 727 C. As a result, they detect clean Si surfaces even at experimental conditions where oxidation of the Si surface should have happened according to extrapolation of high temperature data [6,7].In this Letter we use the optical second harmonic generation (SHG) to monitor in situ the kinetics of Si(ill)-...