AVS1997 Session EM+SS-TuA: Semiconductor Surface Chemistry: Making and Breaking Covalent Bonds
Tuesday, October 21, 1997 2:00 PM in Room C3/4
Tuesday Afternoon
Time Period TuA Sessions | Abstract Timeline | Topic EM Sessions | Time Periods | Topics | AVS1997 Schedule
| Start | Invited? | Item |
|---|---|---|
| 2:00 PM |
EM+SS-TuA-1 Formation of Periodic Step and Terrace Structure on Si(100) Surface during Annealing in Hydrogen Diluted with Inert Gas
Y. Kumagai, K. Namba, T. Komeda, Y. Nishioka (Texas Instruments Tsukuba Research & Development Center Ltd., Japan) There has been considerable interest in the atomically flat Si(100) surface due to increasing demand for the SiO2/Si interfacial quality in the future metal oxide semiconductor (MOS) devices. To this end, annealing the Si wafer in the H2 ambient has attracted much attention due to its high potential for realizing atomically flat hydrogen terminated surface. However, the technique has been limited by the requirement for the high gas purity and for the leakage free system. In addition, the use of 100% H2 needs special care for safety. Here, we report the first result of Si surface planarization by annealing Si wafer in an inflammable ambient where 3% H2 is diluted with the He base gas. Si(100) wafers pre-cleaned in organic solvents are annealed at various temperatures (800C to 1100C) for 10 min at 0.1 MPa in a quartz furnace sealed by a gasket. The undesired impurities such as H2O and O2 in the gas are eliminated by a inline purifier. On the Si surface annealed at 1000C or above, a well defined step and terrace structure consisting of alternating straight A-type and zigzagged B-type step is observed by AFM. The step height is only mono-atomic (0.13 nm) and the terrace width reaches the wafer's off-angle limited value. Therefore, the surface is considered to be atomically flat over a wide range (>10µmx10µm). In the FT-IR ATR spectrum, an intense Si monohydride peak at 2099 cm-1 indicative of the hydrogen termination of the dangling bonds on the top Si dimers is observed besides a weak Si dihydride peak at 2108 cm-1 probably arising from the step edges or the defect sites. Both AFM and FT-IR results show that the surface is ideally terminated by the highly uniform monohydride layer. The surface morphology does not vary beyond 1000C annealing while no morphological improvement is observed below 950C. The result suggests that the atomically flat Si(100) surface may be formed through surface etching by hydrogen dissociating at the surface with the critical temperature higher than near 1000C. |
|
| 2:20 PM |
EM+SS-TuA-2 Imaging and Dissociation of D-S Bond on Si(111)-7x7 with Low Temperature STM
M.A. Rezaei, B.C. Stipe, W. Ho (Cornell University) We have studied the low coverage adsorption of D2S on Si(111)-7x7 with a low temperature scanning tunneling microscope (30 K-300 K). We observe a molecular species adsorbed only on center adatom sites, with no observable difference between the faulted and unfaulted halves of the unit cell. Previous studies suggest that this molecular species is DS1. This species can be dissociated into two fragments by placing the STM tip above the molecule and applying a suitable voltage pulse. The first fragment appears in place of the pre-dissociated species and is attributed to the sulfur containing complex. This fragment is stable even under high voltage and current conditions. The second fragment is observed away from the predissociated species. This fragment is consistent with deuterium adsorbed on a surface dangling bond and it can be desorbed with the STM. Occasionally after the STM dissociation of the molecular species, only the more stable fragment is visible, implying either a desorption or long range motion of deuterium.
|
|
| 2:40 PM |
EM+SS-TuA-3 A Scanning Tunneling Microscopy Investigation of H-Atom-Etched Br-Terminated Silicon Surfaces
M.T. McEllistrem, B.S. Itchkawitz, J.J. Boland (University of North Carolina, Chapel Hill) Bromine-passivated Si(100)-2x1 and Si(111)-7x7 surfaces were investigated with scanning tunneling microscopy following exposure to hydrogen atoms. The hydrogen atoms were found to remove bromine atoms, and in some cases etch the silicon surface. Previous studies by others, which measured changes in surface adsorbate concentration, indicate that halogen abstraction by hydrogen atoms follows Eley-Rideal kinetics.1,2 Such studies did not address the formation of surface dangling bonds, however. The scanning tunneling microscopy studies described here follow the creation of dangling bonds, the formation of surface hydride, and the etching of silicon. Differentiation of these various reactions is aided by the use of scanning tunneling spectroscopy. Reaction sites appear to be randomly distributed across the Si(100)-2x1 and Si(111)-1x1 surfaces, but appear correlated in the case of the Si(111)-7x7 surface. Dangling bond sites are found to be mobile at room temperature upon repeated imaging of the same surface region. This unexpected bromine atom mobility is likely facilitated by the tunneling tip and its associated electric field.
|
|
| 3:00 PM |
EM+SS-TuA-4 "Molecular" Chemistry of Hydrogen-Terminated Silicon Surfaces
C.E.D. Chidsey (Stanford University) Hydrogen-terminated silicon surfaces are metastable under ambient conditions. This metastability is due to the significant activation barriers to reaction with ambient species, such as O2 and H2O. However, like closed-shell molecular compounds that contain H-Si bonds, the hydrogen-terminated silicon surfaces do react with the some anions and radical species. On the Si(111) surface, reaction with the fluoride anion occurs only at step edges, leading to step-flow etching and the ideal H-Si(111) surface first described by Higashi et al. 1. Reaction with radicals explains other reactivity patterns of hydrogen-terminated silicon surfaces and offers several novel ways to chemically modify silicon surfaces. Three cases will be discussed. Case (1): Dissolved oxygen in water initiates the oxidation of the terraces of the otherwise ideal H-Si(111) surface. We propose that oxygen molecules are reduced in water to superoxide anion radicals which then abstract hydrogen atoms from the H-Si(111) terraces to form silicon radicals (dangling bonds) and further oxidation products. H-atom abstraction by superoxide anion radical explains the known enhancement by water of oxide growth on hydrogen-terminated silicon surfaces and may be responsible for many contamination processes on hydrogen-terminated silicon surfaces. Case (2): H-Si(111) can be readily chlorinated by a radical chain reaction initiated by UV-cleavage of Cl2 to form ideal Cl-Si(111). Case (3): Free-radical initiated addition of 1-alkenes to H-Si(111) yields densely packed alkyl monolayers attached to the surface by robust, covalent C-Si bonds.
|
|
| 3:40 PM |
EM+SS-TuA-6 Characterization of Silicon Oxide Etching in HF Vapor Process
Y.P. Han, A.S. Lawing, H.H. Sawin (Massachusetts Institute of Technology) We have studied oxide etching mechanisms of HF vapor etching process in two regimes : monolayer coverage(gas phase) and multilayer(condensed layer). The etching rate of oxide is greatly affected by the flow rate of the reactant stream and the total pressure of reactor, which can change the mass transfer rates of both reactants and products. The rate limiting steps of the etch rate have been studied at various conditions by changing the temperature of the reactor, the partial pressure of the reactants and the flow rate. We have observed a loading effect of the etching process at a low flow rate of reactants due to the depletion of reactants. The presence of condensed layer has also been investigated by in situ ellipsometry. Ellipsometric measurements has shown two adsorption regimes with respect to the partial pressures of both HF and water on a silicon wafer. Water adsorptions prior to the etch reaction play very important role in both etching rate and inhibition time which is called a preconditioning. The etching rate measurements for a small samples have shown long clearing times measured by ellipsometry. This long clearing occurs where the sample edge etch more slowly than the center for a small sample. The nonuniform etching of the sample can be explained by the low concentration of product at the edge of the sample, which could be explained by build up of an etching product. |
|
| 4:00 PM |
EM+SS-TuA-7 Low-Pressure UV-Chlorine Chemical Vapor Cleaning: Ni Removal from Si(100)
C.H. Courtney, H.H. Lamb (North Carolina State University) UV-chlorine (UV/Cl2) chemical vapor cleaning (CVC) is a low-temperature, vacuum-compatible process for removing transition metal and hydrocarbon contaminants from Si surfaces. Our typical UV/Cl2 CVC processing conditions are 20 sccm Cl2, 300 mTorr pressure, and a substrate temperature of 150-200°C. In this work, Ni removal from Si(100) by UV/Cl2 CVC was investigated using Auger electron spectroscopy (AES), x-ray photoelectron spectroscopy (XPS), and atomic force microscopy (AFM). Ni was applied to H-terminated and oxide-covered Si(100) surfaces by ultrahigh-vacuum physical vapor deposition. In situ XPS verified that Ni deposited on H-terminated Si at submonolayer coverages reacts to form nickel silicide species; however, Ni deposited on oxide-covered Si forms metallic clusters. UV/Cl2 CVC for 2 min at 200°C was found to reduce the Ni surface concentration on Si to below the AES detection limit; concomitantly, a surface chlorosilyl layer was formed. AES depth profiles revealed that Ni was not contained within or buried beneath the chlorosilyl layer. Ex situ AFM of surfaces after UV/Cl2 CVC indicated that photochemical etching was limited to ~20 Å and that the surface was smooth (RMS roughness = 1.6 Å). In contrast, UV/Cl2 exposure of Ni deposited on oxide-covered Si produced involatile NiClx species. These results indicate that Ni removal from Si occurs via a photochemical etching mechanism and not direct volatilization of Ni chlorides. Experiments are underway to determine a possible role of trichlorosilylnickel complexes. |
|
| 4:20 PM |
EM+SS-TuA-8 The Mechanism of Copper Removal in Ultraviolet Excited Chlorine
A.S. Lawing, H.H. Sawin (Massachusetts Institute of Technology); R.T. Fayfield (FSI International) We have investigated the relative effects of surface and gas phase photolysis on the efficiency of copper removal in the ultraviolet excited chlorine (UV/Cl2) process. Based on our results, we have developed a mechanism for the removal of copper from wafer surfaces which is significantly different from what has been proposed in the literature to date. The UV/Cl2 process can be operated in a parameter space where copper removal is driven by photon stimulated reduction and desorption of copper chlorides and not by gas phase production of chlorine radicals. The mechanism implies that substrate etching or the formation of SiXCuYClZ complexes is not required for copper removal with UV/Cl2. Efficient copper removal can be realized by balancing the flux of photons and chlorine to the wafer surface such that a high concentration of the volatile product, CuCl, is maintained on the wafer surface. These experiments were performed in an apparatus in which we can perform a cleaning process and transfer the sample into an analytical chamber for surface analysis while maintaining a base pressure in the 10-7 torr range. Sample contamination was achieved by sputter deposition of ~0.05-0.1 monolayer of copper in the analytical chamber. XPS analysis was performed before and after the cleaning process. Using this analysis we monitor the amount of metal removed, and the chemical state of the copper after processing. The ability to distinguish the chemical states and relative surface concentration of copper chlorides was crucial to the development of our removal mechanism. Two copper removal processes have been developed based on our mechanism; 1) copper is removed at 50 mTorr Cl2 and 75°C and 2) copper is removed with UV illumination under vacuum after pre-chlorination in the dark. Substrate surface integrity is maintained with these two processes. |
|
| 4:40 PM |
EM+SS-TuA-9 Study on Surface Symmetry and Electronic States Change in Laser-Induced Surface Reaction of the Si(111)/Cl2 System Using Second-Harmonic Generation
S. Haraichi, F. Sasaki (Electrotechnical Laboratory, Japan) The understanding of atomic-scale mechanisms in the dry etching process is required to develop the next-generation microfabrication technology which demands high accuracy and very low damage. However an accurate time-resolved analysis of below the nanosecond order which can elucidate the dynamic mechanism of the etching reaction has not been undertaken as yet. We have investigated Cl chemisorption and laser-induced surface reaction for the Si(111)/Cl2 system using femtosecond second-harmonic generation (SHG). The probe beam at 600-1300 nm was generated by the output of the femtosecond optical parametric generator (OPG) and amplifier (OPA) pumped by a Ti-sapphire regenerative amplifier. The SHG spectra from the clean Si(111)7x7 reconstructed surface and the Cl-chemisorbed Si surface have been measured as a function of the probe beam photon energy ranging from 1.55 to 2.07 eV. The SHG spectrum from the clean Si(111)7x7 surface shows a resonant peak at 1.7 eV, and that from the Cl-chemisorbed Si surface shows two resonant peaks at around 1.6 and 2 eV involving the Cl-induced electronic states. The surface symmetry at each resonant peak before and after the Cl chemisorption and the laser-induced surface reaction has been observed by rotating the normally incident linear-polarized probe beam. The SHG intensities increase with Cl exposure, however, surface symmetry shows no significant change after the Cl chemisorption and the laser-induced reaction even each resonant peak corresponds to the different Cl-induced electronic state. Rather low damage chemisorption and laser-induced reaction without the remarkable change of surface structure seem to be realized. The reaction yield and the threshold laser fluency dependence on the laser beam photon energy ranging from 1.55 to 2.07 eV have also been studied. No resonant reaction has been observed within the range of these photon energies. |
|
| 5:00 PM |
EM+SS-TuA-10 The Use of Reactive Ion Sputtering to Produce Clean Germanium Surfaces in a Carbon Rich Environment - An Ion Scattering Study
V.S. Smentkowski, A.R. Krauss, D.M. Gruen (Argonne National Laboratory); J.C. Holecek, J.A. Schultz (Ionwerks) We have used the ion spectroscopic techniques of direct recoil spectroscopy (DRS) and mass spectroscopy of recoiled ions (MSRI) to demonstrate that low energy reactive ion sputtering of Ge is capable of removing surface impurities such as carbon. The experiments were performed in a vacuum chamber intentionally maintained at 3.5 x 10-7 Torr. At these pressures physical sputtering using noble gas is not effective for cleaning Ge surfaces as carbon re-deposits onto the surface. We demonstrate that reactive sputtering of Ge using 4.0 keV nitrogen at a Ge surface temperature of ~ 740 K and above removes surface carbon and deposits nitrogen on the Ge surface. Heating the nitridated Ge surface to above ~ 880 K results in the desorption of nitrogen and generates a Ge surfaces free of H, C, N, and O under poor vacuum conditions. 1
|