J. Phys. Chem. 1993,97, 12051-12060
12051
Precursor and Direct Activated Chemisorption of Chlorine Molecules onto Si( 111) (7x7) and Si( 100) (2X 1) Surfaces Daniel J. D. Sullivan,’ Harris C. Flaum, and Andrew C. Kummel Department of Chemistry, University of California, San Diego, La Jolla. California 92093 Received: July 15, 1993; In Final Form: September 9, 1993’
The zero-coverage sticking/chemisorption probabilities (SO)of a monoenergetic Clz beam are measured on two faces of silicon: S i ( l l 1 ) (7x7) and Si(100) (2x1). The initial sticking probabilities (SO) are measured as a function of the incident translational energy (Ei),the surface temperature (Ts),and the angle between the incident beam and the surface normal (0i). For Clz chemisorption on Si(ll1) (7x7) at 300 K, there is a moderate increase in SOfrom 54% at 0.038 CV to 75% at 0.66 eV. SOis nearly insensitive to the surface temperature for Ei > 0.1 1 eV. At Ei S 0.11 eV there is a drop in SOwhen the Tsis elevated. These data indicate that the primary mechanism for adsorption of Clz onto Si(ll1) (7x7) is direct activated chemisorption with an average barrier -0.04 eV and that there is a precursor mediated chemisorption channel a t low Ei. Conversely, the initial sticking probability is a strong function of the incident molecular beam energy for Clz onto Si( 100) (2X 1). For the Si( 100) (2X 1) surface at 300 K, there is a decrease in SOfrom 58% at 0.038 eV to 42% at 0.045 eV and then a sharp increase to 72% a t 0.16 eV. At the very lowest incident translational energy (0.038 eV), SOis a strong function of the surface temperature while for higher incident translational energy (10.38 eV) SOis independent of the surface temperature. These data indicate that Cl2 adsorbs on Si(100) (2x1) via precursor-mediated chemisorption at low translational energies and via direct activated chemisorption with an average barrier of -0.055 f 0,010 eV at high translational energies. The SOis independent of the incident angle of the molecular beam for all incident energies on both surfaces.
I. Introduction The goal of this study is to determine the mechanism(s) of chemisorptionfor chlorine molecules striking a clean, well-ordered silicon surface. Our experiments probe the initial stages of molecular chlorine chemisorption onto a silicon surface at zero chlorine coverage. Chlorine chemisorption is studied upon two dissimilar faces of silicon. The Si( 100) (2X 1) surface has the top layer of atoms arranged in dimer pairs with all the atoms in equivalent sites within the unit cell.’ The Si( 111) (7x7) surface structure is more complicated: there are 13distinct surface atom sites within the unit cell including the adatoms, rest atoms, the corner holes, and the stacking fault.2 There are two basic mechanisms for chemisorption on a surface: precursor-mediated and direct activated. “Precursor chemisorption”denotes a process in which the incident molecules first become trapped in the molecular physisorption well and then dissociatively chemisorb by migrating to defects, utilizing the thermal energy of the surface, or repartitioning molecular energy. For “direct chemisorption”,molecules instantaneously chemisorb upon hitting the surface. The dangling bond character of semiconductor surfaces may allow direct chemisorption of Cl2 without overcoming a barrier. “Direct activated chemisorption” refers to a chemisorption process in which incident molecules require a minimum incident energy in order to chemisorb upon striking the surface. The silicon surfacesundergo reconstructions which create sites with varying dangling bond order. Therefore it is reasonable to expect a variety of barrier heights to chemisorption to be present on these surfaces. The chemisorption mechanism@) can be determined by measuring the chemisorption probability versus the incident kinetic energy of the gas molecules.3-5 For precursor-mediated chemisorption, the sticking probability decreases with increasing molecular beam energy because high-energy molecules cannot dissipate enough of their kinetic energy into either the surface lattice or rotational/vibrational excitation to fall into the Abstract published in Aduance ACS Abstracts. October 15, 1993.
0022-365419312097-1205 1%04.00/0
physisorption well. In addition, for precursor-mediated chemisorption, the zero-coverage (“initial”) sticking probability normally decreases with increasing surface temperature since the barrier to desorption from the physisorption well is usually greater than the barrier to chemisorption from the physisorption well.5.6 For direct chemisorption, molecules chemisorb instantaneously upon hitting the surface. If this process is barrierless, the initial sticking coefficient will be independent of the translational energy of the molecular beam (0.02-4.0 eV). In general, for direct activated chemisorption,the initial sticking coefficient increases with kinetic energy until it approaches a constant value. For most direct chemisorption processes, the initial sticking probability is nearly independent of the surface temperature.%’ For direct activated chemisorption into a long-lived molecular state, there can be a slight decrease in the initial chemisorption probability with increasing surface temperature if the activation barrier to desorption from the chemisorption state is slightly greater than the barrier to dissociation from the molecular chemisorptionstatea6 We have measured the initial chemisorption probability of molecular chlorineonthe 300KSi(lll) (7x7) and Si( 100) (2X 1) surfaces at normal incidence as a function of molecular beam energy. In addition, we have determinedthe initial chemisorption probabilityas a function of the surface temperature and the angle of incidence. Even though at equilibriumfor Ts= 300 K, chemisorption and physisorption both result in dissociation, there may be an intermediate metastable molecular chemisorption state. An exampleofthisisthechemisorptionofOzonPt(111). J. Grimbolt, A. C. Luntz, and D. E. Fowler638 and C. T. Rettner and C. B. Mullinsgshow that both precursor-mediatedand direct activated chemisorption occur through an intermediate 02-chemisorption state; this chemisorption mechanism has been denoted as “quasidirect”. Therefore, in direct activated chemisorption,the barrier being probed may lie either between the gas-phase and the molecular chemisorption state or between the gas-phase and the dissociative chemisorption state even if all chemisorption results in dissociation. The presence of molecularly adsorbed states is suggested for HC1 and HBr on Si(ll1) by M. Miyamura, Y. 0 1993 American Chemical Society
Sullivan et al.
12052 The Journal of Physical Chemistry, Vol. 97, No. 46, 1993
Sakisaka, M. Nishijima, and M. 0nchi.lo The presence of molecularly adsorbed states for these species on the Si(111) surface would indicate a strong possibility for the existenceof molecularly adsorbed states for Cl2. C. C. Cheng, Q. Gao, W. J. Choyke, and J. T. Yates, Jr.” show that at least two states are accessed for C12 adsorbed on Si(lOO), and Miyamura et al. show that HCl also accesses at least two states on the Si( 11 1) surface. The two states observed by Cheng et al. are both dissociative with C1 in a bridging site, bonding with two silicon surface atoms, or in a tilted atop site, bonding with a single silicon atom. Miyamura et al. suggest that one of the two states accessed by HCl is molecular and the other is dissociative. ChemisorptionofCl~onSi(ll1) (7X7)isdiscussedbyP. Gupta, P. A. Coon, B. G. Koehler, and S.George.12 Gupta et al., and the references sited therein, show that the chemisorption process results in dissociation into two sites, 01and 82. The 81 and 8 2 sites are distinguished by the desorption energies observed for SiC12. The 61 and 8 2 sites fill sequentially with only the 81 site occupied at low coverages, 8 < 0.68,. The B1 state is attributed to recombinative desorption of Sic1 C1, with Sic1 being the only surface species present at surface coverages below 0.68,. At surfacecoveragesabove 0.68,, theSiCl,, x = 2,3, and4, desorption products are correlated with the presence of mono-, di-, and trichlorides on the surface. The groups of F. X. Campos, G. C. Weaver, C. J. Waltman, and S.R. Leone” and A. Szabo, P. D. Farrall, and T. Engel14 have recently reported the effects of increased incident energy on the etching of Si( 100) with atomic and molecular chlorine. In both of these studies the etch rates are determined by monitoring the etch products with mass spectroscopy. The molecular beams in both of these studies produce atomic as well as molecular chlorine and thus may influence the etch rates and etch products. Campos et al. use a molecular beam produced by laser desorption of cryogenic Cl2 films. This source provides a beam with a wide distribution of kinetic energies. They report increases in the etch rate of 317 K Si( 100) surfaces by molecular chlorine of a factor of 10 when the mean incident energy is increased from that of a thermal beam (-0.025 eV) to 0.4 eV. Campos et al. observe that the etch rate increases with increased incident energy with etch rates of 30 times that observed for thermal beams when incident beams contain considerableportions with kinetic energies above 3 eV. Campos et al. suggest that the increased etch rate is due to the kinetic energy of the incident molecular chlorine being used to break the silicon-silicon bonds on the surface. In the study reported by Szabo et al. hyperthermal beams of Cl/C12 are shown to have much higher etch rates on Si(100) surfacesat low temperatures (180 I K) than thermal beams. Szabo et al. suggest that two pathways for desorption of the etch products are active. One pathway being the thermal desorption of volatile products and the second being the direct process of induced desorption caused by impact of the hyperthermal particles from the beam. The interaction of chlorine with silicon is important in the etching procedures used in device fabrication.l5-18 The literature contains several studies on the interaction of molecular chlorine with silicon s~rfaces.IS-1~J9-21 These studies have focused on the products of the nature of the bonding on the ~urface,~~.25 the reaction sites,lZ,22-31 and reactions at various coverages (8cJ.I2v2OIn these references, all of the surfaces have already undergone reaction and have chlorine present on the surface (0~1 > 0) when the experiments are performed, Conversely, the experimentsdescribed in this paper probe the reactions on unreacted surfaces with well-characterized structures and with zero chlorine coverage (8cI = 0).
A)
+
11. Experimental Section
The experiments are performed in the vacuum chamber shown schematically in Figure 1. The instrument consists of two major parts, a molecular beam source and an ultrahigh vacuum (UHV) experimental chamber.
Figure 1. (A) A side view schematic of the experimental apparatus. The apparatus consists of two major parts, the molecular beam source and the ultrahigh vacuum experimental chamber. The molecular beam, the quartz flag, the sample, and the quadrupole mass spectrometer are collinear. The quartz flag and the sample may be moved from the path of the molecular beam to allow the quadrupole mass spectrometer to sample the direct beam. (B) An internal view of the molecular beam source. The molecular beam originates from the pulsed valve and travels through the skimmer, a slit on the chopper wheel, the gate valve, and finally through the collimator to enter the experimental chamber.
A. Molecular Beam Source. The molecular beam source (Thermionics) is divided into three independently pumped chambers (see Figure 1B): Chamber 1 houses the pulsed valve and skimmer; chamber 2 contains the chopper wheel, chopper motor, and gate valve; and chamber 3 holds the collimator. Chamber 1 is pumped by a water-baffled 16-in. diffusion pump (Varian VHS-IO, pumping speed 2120 L/s). Chambers 2 and 3 are each pumped by a water-baffled 6-in. diffusion pump (Edwards 100F,pumping speed 322 L/s). Typical pressures for chambers 1,2, and 3 are lo-’, 10-8, and l C 9 Torr, respectively. The molecular beam source is inserted into the experimental chamber through a 16.75-in. rectangular conflat flange so that the collimator is only 6-8 cm from the surface, and the nozzle is 12-1 5 cm from the surface. The molecular beam source remains attached to theUHV systemduring the 180 OC bakeout; however, no heat is directly applied to the source. The beam is composed of Clz (99.999% Matheson) seeded in He, Kr, or Xe expanded through a pulsed-valve nozzle (General Valve) with a 2.0-mm orifice. The translational energy of the chlorine molecules in the beam is varied between 0.038 and 0.66 eV by expanding mixtures of Clz and He,Kr, or Xe. The monoenergeticbeam passes through a 0.29-mm conical skimmer (Beam Dynamics), an open slit of the chopper wheel, and a 2.0mm collimator. The 12 cm diameter chopper wheel has two 0.95 cm wide slits 180° apart as well as two 0.05-cm slits 180° apart. The chopped beam pulses can be measured by a quadrupole mass spectrometer and averaged/displayed on a digital oscilloscope (LeCroy, 9400A). The pulsed valve and the chopper wheel are synchronized via a digital delay box with a photodiode and an electric eye mounted on opposite sides of the chopper wheel. This allows the gas pulses to pass through either the large slits, 120-rs
The Journal of Physical Chemistry, Vol. 97, No. 46, 1993 12053
Chemisorption of Chlorine onto Silicon Surfaces pulse length at 200 Hz, or the small slits, 7 ps at 200 Hz. The velocity of the molecules is determined by measuring the flight time from the chopper wheel to the quadrupole. The distance from the chopper wheel to the quadrupole (29.5 f 0.1 cm) is determined by measuring the fight times of gases of known speeds, He, Kr, Ar, Xe, and N2. From the rotational spectra and fast time-of-flight measurements (time response > 1. This approximation is valid when Vd/Va >> 1 and T, is - ( E d - &)/kB or Ts is > ( E d - &)/kB. The first condition, vd/Va >> 1, is almost always valid, as the number of accessible states is always larger for the gas-phase molecular ensemble than for the chemisorbed molecular ensemble. The second condition is shown to be true for the range of T,studied from the results of the fit. Equation A1 then becomes
-
