Papers

AVS1997 Session EM+AS-MoM: Ultra Thin Silicon Oxides: Interface Structure and Kinetics

Monday, October 20, 1997 8:20 AM in Room C3/4

Monday Morning

Start	Invited?	Item
8:20 AM		EM+AS-MoM-1 Atomic Scale Processes During Oxidation at Si(001)-SiO₂ Interfaces A. Pasquarello (IRRMA and University of Geneva, Switzerland); M.S. Hybertsen (Bell Laboratories, Lucent Technologies); R. Car (IRRMA and University of Geneva, Switzerland) The atomic scale processes occuring during oxidation at the Si(001)-SiO₂ interface are not well understood. We have done extensive quantum molecular dynamics simulations to explore possible processes. First principles calculations have been performed starting from an ideal interface structure in a root 8 by root 8 unit cell containing 180 atoms. By imposing an appropriate temperature gradient on the cell, we have induced substantial atomic diffusion. Following a period of rapid diffusion, the system is quenched. The full schedule of molecular dynamics simulation spans about 20 picoseconds. During the period of rapid diffusion, we observe the oxidation of approximately 3 monolayers of crystalline silicon. The interface moves into the crystalline silicon in two steps: first the nearby oxygen induces substantial disorder in the adjacent silicon layer. Then the oxygen incorporates into that layer. We observe that oxygen diffusion frequently involves an intermediate configuration where the oxygen atom is three-fold coordinated by silicon. The final structure that is found following the quench has no dangling bond defects in the interface region. The transition region has roughly equal numbers of silicon in +1, +2 and +3 oxidation states. Further details of the dynamic atomic scale processes observed will be described in the talk.
8:40 AM		EM+AS-MoM-2 Experimental and Theoretical Studies of the Initial Water-Induced Oxidation of Si(100)-2x1 M.K. Weldon, B.B. Stefanov, K. Raghavachari, Y.J. Chabal (Bell Laboratories, Lucent Technologies); A.B. Gurevitch (Columbia University) Despite the considerable technological importance of silicon oxidation, remarkably little is known about the mechanism of the initial oxidation process due, primarily, to the dearth of suitable probes with the requisite sensitivity. For example, although infrared spectroscopy has historically been one of the most powerful (non-destructive) techniques for the study of interfacial reactivities, acquisition of spectra at low frequency (where the relevant Si-O modes occur) is hampered by silicon multiphonon absorption. However, we have recently developed novel optical geometries that afford high resolution spectra with submonolayer sensitivity and provide a first tantalizing insight into structures formed upon initial oxidation of the Si(100)-2x1 surface. In addition, we have also undertaken detailed theoretical studies in order to definitively characterize the novel oxide structures so-formed. Using this combined experimental and theoretical approach, we are able to demonstrate that the initial water-induced oxidation of the reconstructed Si(100) surface proceeds via insertion of oxygen into the Si dimer bond which, in turn, facilitates subsequent incorporation into the Si backbonds. Upon annealing to 500 °C, the oxygen agglomerates into five-oxygen containing dimers with the hydrogen being lost at elevated temperatures (600-800 °C). Interestingly, in a parallel set of studies we have found that similar structures are also formed during the chemical oxidation of silicon using hydrogen peroxide-containing solutions under ambient conditions, underlining the potential importance of these studies to understanding technologically relevant processes.
9:00 AM		EM+AS-MoM-3 Real-Time Core-Level Spectroscopy of Initial Thermal Oxide on Si(100) Y. Enta, Y. Miyanishi, H. Irimachi, M. Suemitsu, M. Niwano (Tohoku University, Japan); N. Miyamoto (Tohoku Gakuin University, Japan); E. Shigemasa, H. Kato (National Laboratory for High Energy Physics, Japan) Initial stage of thermal oxidation of Si(100) have been investigated by real-time high-resolution Si 2p core-level photoelectron spectroscopy. In particular, the dependence of the oxidation kinetics on the oxidation temperature was obtained from the time evolutions of the suboxide components. Experiments were carried out on the beam line BL-3B at Photon Factory, the National Laboratory for High Energy Physics (KEK). The nominal energy resolution, including both the monochromatized light (130eV) and the analyzer was about 0.1eV. The sample was resistively heated at 1000 C in ultrahigh vacuum chamber after conventional RCA treatment. The oxidation was conducted by exposing the Si(100) clean surface to molecular oxygen in the pressure range of 7.8x10^-8 to 2.8x10^-6 Torr and in the temperature range of room temperature to 720 C. The obtained Si 2p core-level spectra of the thin oxide consisted of several well-separated surface components, as opposed to ambiguous peak structures in many previous reports. The obtained spectra demonstrate that our measurements were done with a high resolution. The deconvoluted spectra showed that the intensities of Si¹⁺ and Si³⁺ components are enhanced at higher temperatures. With the knowledge that an ideal Si/SiO₂ interface should be free of Si¹⁺ and Si³⁺ components, this result implies that the interface roughness develops at high temperatures. The time evolutions of Si⁴⁺ component by real-time measurements showed a good agreement with those of O 2p intensity in our previous report at all the temperatures investigated ¹. This provides a strong support for our model for the oxidation kinetics that there exist two growth modes for the oxidation, whose domination switches at a critical temperature around 650 C: the Langmuir-type adsorption mode in the lower temperature region and 2D island growth mode in the higher temperature region. ¹Y. Enta et al., Appl. Surf. Sci. 100/101(1996)449.
9:20 AM		EM+AS-MoM-4 High Resolution XPS Study of Ultrathin Oxynitride Films on Si(100). E.P. Gusev (Rutgers University); K. Morino, S. Yokoyama, M. Hirose (Hiroshima University, Japan); M.L. Green (Bell Laboratories, Lucent Technologies) This work was to study local chemical configurations and depth distributions of nitrogen atoms in ultrathin (1.5 - 4 nm) oxynitride films thermally grown on Si(100) in NO and N₂O at 700 - 1000 C. High-resolution ESCA-300 (SCIENTA) spectrometer with Al photon source was used to monitor photoemission of N1s, O1s, and Si2p core-levels, with particular emphasize on the N1s level. For all processing conditions studied, we observed a single N1s peak with the width (FWHM~1.4 eV) similar to the O1s and Si2p (oxide) peaks, and much broader than the silicon substrate peak ( ~ 0.4 eV). The shape of the N1s peak is asymmetric and depends on the processing conditions, implying more than one bonding configuration of nitrogen. The N1s spectra can be deconvoluted into two Gaussian components. The component at 397.6 eV corresponds to nitrogen species near the SiOxNy/Si interface, as it follows from angular resolved measurements. Due to its similarity to a reference CVD Si₃N₄ film, we interpret this feature in terms of nitrogen atoms triple bonded with silicon near the interface. The other component at 398.2 - 398.9 eV (depending on film thickness) is located further into the film and it could be due to some N-O bonds, strained bonds near the interface or/and other structural factors. We will also discuss nitrogen depth profiles deduced from simulations of angular dependencies of relative N1s, O1s, Si2p(oxide) and Si2p(substrate) intensities.
9:40 AM		EM+AS-MoM-5 Direct Observation of Si Lattice Strain and its Distribution in the Si(001)-SiO₂ Interface Transition Layer with Medium Energy Ion Scattering Spectroscopy D.W. Moon, H.K. Kim (Korea Research Institute of Standards and Science); Y.P. Kim, S.K. Choi (Korea Advanced Institute of Science and Technology) The Si(001)-SiO₂ interface has been intensively investigated with various experimental techniques due to its importance in improving the electronic properties of 3-4 nm ultrathin SiO₂ dielectric layers and silicon-on-insulators. Previous experiments with transmission electron microscopy, photoemission spectroscopy, Rutherford Backscattering Spectroscopy and X-ray scattering have reported a .5 - 25 nm thick transition region with different aspects of the compositional and structural transition in the interface, depending on the used analysis methods. However, detailed informations on the transition region between a crystalline Si substrate and an amorphous SiO₂ overlayer are still not enough. We report here a direct observation of strained Si lattices and their strain distribution in the 1.3 nm transition regions of the Si(001)-SiO₂ interfaces for thermal oxides and similar observations for ion beam oxides formed by oxygen ion beam bombardment with medium energy ion scattering spectroscopy. From the energy spectra and the angular distributions of the Si peak, the Si crystallinity and its gradual change in the transition region could be analyzed with Si and O concentration changes . From the shift of the blocking dip position, the strain of the crystalline Si lattices and its distribution in the transition region could be measured. The strain was in the vertical direction and the maximum values at the SiO₂ side of the transition region were 0.96% and 2.8% for thermal oxides and ion beam oxides, respectively. We believe the informations in this report is useful to understand the still controversial Si(001)-SiO₂ interface structure more clearly.
10:00 AM		EM+AS-MoM-6 Optimization of Si-SiO₂ Interface Chemistry and Structure by Combined Plasma and Rapid Thermal Processing. G.L. Lucovsky (North Carolina State University) This paper address two aspects of interface bonding and structure that effect performance and reliability of field effect transistors with ultra-thin gate oxides: i) the incorporation of nitrogen atoms at the Si-SiO₂ interface; and ii) the minimization of excess sub-oxide bonding arrangements that contribute to interface roughness. The incorporation of nitrogen atoms at Si-SiO₂ interfaces at monolayer levels improves device reliability, and also reduces interfacial roughness as reflected in the variation of the effective mobility as a function of the electric field at the surface of the channel layer. Nitrogen incorporation has been accomplished by a several different processing methods including high temperature (>1000°C) furnace and rapid thermal oxidation, and low-temperature (300°C) plasma assisted processing. This paper discusses the plasma-assisted approach in combination with low thermal budget rapid thermal annealing. Nitrogen incorporation and interface roughness are have been studied by separating the interface formation and oxide growth into a sequence of discrete processing steps, each of whch can be independentally monitored and controlled: i) interface formation is accomplished by plasma assisted oxidation at 300°C; ii) nitrogen incorporation is achieved during oxidation by using N₂O as the source gas, or following oxidation in O₂ by exposure to excited species from an N₂/He remote discharge; and iii) chemical and structural relaxation are accomplished by a post plasma-processing 900°C rapid thermal anneal. Interface bonding and stuctural relaxation have studied on-line by Auger electron spectroscopy (AES) in a mutichamber system that has separate chambers for plasma processing, rapid thermal processing, and surface analysis. Ex-situ characterization by other techniques including X-ray photoelectron spectroscopy and medium energy ion scattering provide complementary determinations of interface roughness and nitrogen incorporation.
10:40 AM		EM+AS-MoM-8 Processing Dependence of the Structure and Electrical Characteristics of the Si/SiO₂ Transition Layer. J. Bevk, M.L. Green, K.S. Krisch, M.K. Weldon, Y.J. Chabal, P.K. Roy (Bell Laboratories, Lucent Technologies); S.D. Kosowsky, P.S. Pershan (Harvard University) We have used x-ray reflectivity to probe the Si/SiO₂ interface of the thermally grown oxides. Fits of model electron density profiles to the data reveal the existence of a transition layer at the Si/SiO₂ interface up to 15Å thick, with density higher than either the crystalline silicon or the main oxide layer. Quantitative values obtained for the excess scattering strength, a measure of the extent of this layer, are a function of the O₂ partial pressure of the growth environment in the as-grown oxides. Post-oxidation annealing in an inert gas atmosphere dramatically reduces the excess scattering strength of this interfacial layer, and additionally reduces the density of the main oxide layer. Infrared studies are underway to gain additional insight into the nature of the transition region. Finally, the structural studies are correlated with the electrical characteristics of the bare oxides and of the MOS capacitors.
11:00 AM		EM+AS-MoM-9 Si/SiO₂ Interface Roughness: Comparison between Surface Second Harmonic Generation and X-ray Scattering S.T. Cundiff, F.H. Baumann, W.H. Knox, K.W. Evans-Lutterodt, M.L. Green (Bell Laboratories, Lucent Tecnologies) The roughness of the Si(100)/SiO₂ interface is measured using both surface second harmonic generation (SSHG) and x-ray scattering. SSHG is an optical technique that, for materials with bulk inversion symmetry, is sensitive to regions where the inversion symmetry is broken, i.e. a surface or interface. SSHG has been shown to be sensitive to interface roughness ¹ using etch roughened Si. SSHG shows potential for nondestructive, in situ testing on a VLSI fabrication line. We compared the SSHG results with the interface roughness determined by x-ray scattering for a series of typical industrial oxides with thicknesses ranging from 50 to 600 Å. The x-ray scattering studies show that the interface roughness decreases with increasing oxide thickness ². The interface roughness measured by SSHG shows clear correlation with the x-ray scattering results ³, confirming the sensitivity of SSHG to the naturally occurring roughness at the Si/SiO₂ interface. The correlation between these two techniques is interesting as they are sensitive to differing regions of the spatial frequency roughness spectrum. The implications of this on the roughness spectrum and comparisons between various measurement techniques will be discussed. The SSHG measurements are made using ~10 fs pulses from a modelocked Ti:sapphire laser with a spectrum centered at 850 nm and a repetition rate of 80 MHz. These short pulses produce similar signal to noise ratio as earlier measurements, but using lower average power (30-40 mW) thereby avoiding possible artifact such as sample heating. ¹J.I. Dadap, B. Doris, Q. Deng, M.C. Downer, J.K. Lowell and A.C. Diebold, Appl. Phys. Lett., 64, 2139 (1994). ²M.T. Tang, K.W. Evans-Lutterodt, G.S. Higashi, and T. Boone, Appl. Phys. Lett., 62, 3144 (1993). ³S.T. Cundiff, W.H. Knox, F.H. Baumann, K.W. Evans-Lutterodt, M.-T. Tang, M.L. Green, H.M. van Driel, Appl. Phys. Lett. 70,1414 (1997). ⁴
11:20 AM		EM+AS-MoM-10 Measurements of Nitrided SiO₂/Si Interface Roughness By Crystal Truncation Rod Profiling J.L. Jordan-Sweet, R. Ludeke (IBM T.J. Watson Research Center); T.B. Hook (IBM Microelecronics Division) As feature sizes continue to shrink with each generation of microelectronic device, the roughness of the interfaces between ultrathin layers of materials plays an increasingly critical role in the device's performance. Thus, a probe that can measure accurately the roughness of a buried interface, with statistical meaning and over a large dynamical range of scale, is extremely valuable. Recently, a new technique has been developed for the non-destructive evaluation of the roughness of the interface between a crystalline substrate and amorphous film¹. X-ray diffraction from a truncated crystal produces "rods" of intensity in reciprocal space that are normal to the surface and which pass through the three-dimensional reciprocal lattice points produced by the bulk crystal. These truncation rods are a result of the two-dimensional nature of the terminated surface. If an overlayer is truly amorphous, the variation of intensity along the rods yields information about properties of the buried interface, such as roughness (rms deviation from perfect flatness). Although diffraction from the buried interface is orders of magnitude weaker than diffraction from the bulk wafer, it can be measured easily using synchrotron radiation. Silicon oxides grown under a variety of conditions have been assessed using this technique². We have applied this technique to the evaluation of a set of 70Å thin nitrided gate oxides on Si(100). We will describe the technique and analysis and present results relating interface roughness and nitridation/oxidation process. 1. M.-T. Tang, K.W. Evans-Lutterodt, G.S. Higashi, and T. Boone, Appl. Phys. Lett. 62, 3144 (1993). 2. M.-T. Tang, K.W. Evans-Lutterodt, M.L. Green, D. Brasen, K. Krisch, L. Manchanda, G.S. Higashi, and T. Boone, Appl. Phys. Lett. 64, 748 (1994).
11:40 AM		EM+AS-MoM-11 Recent Improvements in Measuring Ultrathin Oxynitride Layers Using Secondary Ion Mass Spectrometry (SIMS) S.W. Novak, E. Botnick, T.H. Buyuklimanli, M.S. Denker, C.W. Magee, J.T. Mayer, W. Ou (Evans East) Much recent interest has focused on manufacture of oxynitride films significantly thinner than 10 nm for use as gate dielectrics. Previously we have presented a technique for accurately measuring the distribution and areal density of nitrogen in 10 nm and sub-10 nm oxynitride films using SIMS. However this technique begins to reach the limits of resolution for films less than about 5 nm. By utilizing ion bombardment at 75 degrees we have obtained significant improvement in the depth resolution for thinner films. In addition, we have performed preliminary experiments using impact energies less than 1keV. We will demonstrate the effects of using low energy Cs bombardment and higher incidence angles on measurements of oxynitride films. Because the N is contained mainly at the oxide/Si interface, any standard used to calibrate N must also have the N at the interface, due to the difference in ion yield between SiO2 and Si. Our previous measurements calibrated the N contents of an oxynitride standard using an ion implanted standard of N in SiO2 and a series of re-oxidized oxynitride films. The N areal density measurements found for this standard compares well with N measurements performed using NRA. We will also compare the SIMS measurements with those carried out by other techniques and attempt to summarize the strengths and limitations of each type of measurement compared to SIMS.. One primary point of this talk will be that, for the SIMS measurements, the N concentration given for the interface peak is inaccurate due to the effects of ion mixing. However, the areal density of the N interface peak is accurately measured and this number should be used for quantitative comparisons of oxynitride samples.