VOLUME 76, NUMBER 26 PHYSICAL REVIEW LETTERS 24 JUNE 1996 Theoretical and Experimental Optical Spectroscopy Study of Hydrogen Adsorption at Si(111)-(7 3 7) C. Noguez, 1 C. Beitia, 2 W. Preyss, 2 A. I. Shkrebtii, 3 M. Roy, 2 Y. Borensztein, 2, * and R. Del Sole 3, † 1 Instituto de Fisı ´ca, Universidad Nacional Autónoma de México, Apdo. Postal 20-364, 01000 México D.F., México and Department of Physics and Astronomy, and CMSS Program, Ohio University, Athens, Ohio 45701-2979 2 Laboratoire d’Optique des Solides, UA CNRS 781, Université Pierre et Marie Curie, Case 80, 4 place Jussieu, 75252 Paris Cedex 05, France 3 Instituto Nazionale di Fisica della Materia, Departimento di Fisica, II Universitá di Roma “Tor Vergata,” Via della Ricerca Scientifica 1, 00133 Rome, Italy (Received 2 November 1995; revised manuscript received 14 March 1996) The combination of in situ real-time surface differential reflectivity (SDR) spectroscopy and microscopic calculations has been used for the first time to investigate gas adsorption on a Si surface. The optical signatures of some microscopic structural units of Si(111)-(7 3 7) have been identified. The development of the corresponding features in the SDR spectra upon the amount of H exposure allowed us to demonstrate the occurrence of two different mechanisms in the hydrogenation and to determine their relative kinetics. [S0031-9007(96)00523-6] PACS numbers: 68.35.Bs, 68.45.Da, 73.20.At, 78.66.Db In recent years, optical spectroscopies such as re- flectance anisotropy spectroscopy (RAS) and surface dif- ferential reflectivity (SDR) have proved their ability to investigate the intrinsic electronic properties of solid sur- faces [1,2] and to monitor the interaction of surfaces with gas [1,3] or the growth of thin films [4]. However, most of the experimental studies are limited to a phenomeno- logical analysis, without a deep microscopic understand- ing of the optical response of the surfaces. Although the optical measurements are macroscopic in nature, they are directly linked to the microscopic structure of the surface and to its electronic structure [5,6]. A theoretical treat- ment of the optical response of a solid surface, based on its microscopic structure, is therefore necessary to fully understand the experimental data and to interpret them in terms of surface reconstruction, adsorption of atoms, growth of thin films, etc. The interaction of hydrogen with the Si111-7 3 7 surface, well described by the dimer-adatom-stacking fault (DAS) model [7], is a complex process of broad interest. Up to now, theoretical studies on the hydro- genation of Si111-7 3 7 had been hampered by its very large unit cell and only a few attempts have been made [8,9]. On the other hand, the hydrogenation of this surface has been investigated by numerous experimental techniques, which have shown the existence of two reac- tion paths [10–13]. We present the first combination of theoretical and real- time experimental optical studies of Si111-7 3 7 and of the H adsorption. Our experimental SDR spectra are described by means of a theoretical microscopic model, in terms of electronic transitions involving the bulk and surface states of Si111-7 3 7. This permits us to understand the intrinsic optical response of Si111-7 3 7, and to gain deep information on the hydrogenation process. We clearly distinguish the two main mechanisms and we follow their kinetics: (1) adsorption of H on the dangling bonds (DBs), and (2) H breaking of the Si-Si backbonds (BBs) of the adatoms (ADs). The SDR experiments were performed with a rapid in situ spectrometer [14], which delivers the change of the reflectivity upon H adsorption: DRR  R clean Si -R HSi R clean Si , where R clean Si and R HSi are the reflectivities of the clean and H-covered Si surfaces. This quantity is related to the optical suscep- tibility of the surface and hence provides information on the changes of its electronic structure [5]. The incidence angle u of the p-polarized light beam was 60 ± . The SDR signal was registered during the H adsorption, providing a real-time monitoring of the process. The Si samples were heated during 5 min at 900 ± C in the vacuum chamber (base pressure of 10 210 torr) for removing the oxide layer, then cooled down slowly (1 K/s) to get a sharp 7 3 7 low-energy-electron-diffraction (LEED) pattern. Atomic H was produced by decomposition of H 2 by a tungsten filament at 1800 ± C previously outgassed at higher temperature. Figure 1 shows typical experimental SDR spectra (point curves) taken from 500 spectra registered during the H exposure, given in H 2 exposure (note that the vertical scales differ from one spectrum to another). The sample was maintained at 220 ± C. Similar spectra were obtained at other temperatures between 2140 and 300 ± C. Spectrum (a) 20 L 1 L langmuir  10 26 torr s is dominated by a peak A at 1.8 eV, with a small negative minimum C at 3.2 eV. A second peak B is developing around 2.8 eV in spectrum (b) (90 L), together with a double structure D at 3.7–4.1 eV. While peak A saturates after about 150 L, structures B and D continue growing and become progressively dominant, as shown in spectrum (e), where peak A now reduces to a shoulder. The edge of another structure E appears above 5 eV in spectra (d) and (e). Our 0031-9007 96 76(26) 4923(4)$10.00 © 1996 The American Physical Society 4923