Oxidation of Si Surface Nitrogenated by Plasma Immersion N + Ion Implantation

Similar documents
NITROGEN CONTAINING ULTRA THIN SiO 2 FILMS ON Si OBTAINED BY ION IMPLANTATION

Oxidation of hydrogenated crystalline silicon as an alternative approach for ultrathin SiO 2 growth

Defect Formation in 18 MeV Electron Irradiated MOS Structures

Nanocrystalline Si formation inside SiN x nanostructures usingionized N 2 gas bombardment

1.1 FEATURES OF SPECTROSCOPIC ELLIPSOMETRY

3.155J/6.152J Microelectronic Processing Technology Fall Term, 2004

Fabrication Technology, Part I

Supporting information. Infrared Characterization of Interfacial Si-O Bond Formation on Silanized. Flat SiO 2 /Si Surfaces

Structure of Thin Plasma Polymer Films on the Base of Hexamethyldisiloxane

NON-DESTRUCTIVE EVALUATION IN MANUFACTURING USING SPECTROSCOPIC. John A. Woollam and Paul G. Snyder

Ellipsometric spectroscopy studies of compaction and decompaction of Si-SiO 2 systems

Evaluation of plasma strip induced substrate damage Keping Han 1, S. Luo 1, O. Escorcia 1, Carlo Waldfried 1 and Ivan Berry 1, a

OPTICAL ANALYSIS OF ZnO THIN FILMS USING SPECTROSCOPIC ELLIPSOMETRY AND REFLECTOMETRY.

Defense Technical Information Center Compilation Part Notice

Plasma Deposition (Overview) Lecture 1

Laser-synthesized oxide-passivated bright Si quantum dots for bioimaging

In-situ Multilayer Film Growth Characterization by Brewster Angle Reflectance Differential Spectroscopy

Limiting Si/SiO 2 interface roughness resulting from thermal oxidation

Near-perfect modulator for polarization state of light

H loss mechanism during anneal of silicon nitride: Chemical dissociation

X-ray photoelectron spectroscopic characterization of molybdenum nitride thin films

DEPOSITION OF THIN TiO 2 FILMS BY DC MAGNETRON SPUTTERING METHOD

Ion Implant Part 1. Saroj Kumar Patra, TFE4180 Semiconductor Manufacturing Technology. Norwegian University of Science and Technology ( NTNU )

Large Scale Direct Synthesis of Graphene on Sapphire and Transfer-free Device Fabrication

Effects of plasma treatment on the precipitation of fluorine-doped silicon oxide

Self-study problems and questions Processing and Device Technology, FFF110/FYSD13

Feature-level Compensation & Control. Process Integration September 15, A UC Discovery Project

thin films alloys prepared by pulsed excimer laser deposition

Supporting Information

doi: /

ECE 695 Numerical Simulations Lecture 35: Solar Hybrid Energy Conversion Systems. Prof. Peter Bermel April 12, 2017

Thin Film Characterization Using Rotating Polarizer Analyzer Ellipsometer with a Speed Ratio 1:3

Bonding structure and hydrogen content in silicon nitride thin films deposited by the electron cyclotron resonance plasma method

T: +44 (0) W:

Ion Implantation. alternative to diffusion for the introduction of dopants essentially a physical process, rather than chemical advantages:

Ellipsometry Tutorial

Lecture 6. Rapid Thermal Processing. Reading: Chapter 6

Traps in MOCVD n-gan Studied by Deep Level Transient Spectroscopy and Minority Carrier Transient Spectroscopy

Starting solution. Hydrolysis reaction under thermostatic conditions. Check of viscosity and deposition test SOL. Deposition by spin coating

DUV ( nm ) Characterization of Materials: A new instrument, the Purged UV Spectroscopic Ellipsometer,

J. Price, 1,2 Y. Q. An, 1 M. C. Downer 1 1 The university of Texas at Austin, Department of Physics, Austin, TX

On the reversible photo-darkening in amorphous Ge 5 As 41 S 15 Se 39 film

Low temperature anodically grown silicon dioxide films for solar cell. Nicholas E. Grant

Refractive index profiling of CeO2 thin films using reverse engineering methods

3 Physical and Chemical Structure

CVD-3 LFSIN SiN x Process

SUPPLEMENTARY MATERIALS FOR PHONON TRANSMISSION COEFFICIENTS AT SOLID INTERFACES

Laser matter interaction

Wet-chemical passivation of Si(111)- and Si(100)-substrates

Supplementary Figure S1. AFM characterizations and topographical defects of h- BN films on silica substrates. (a) (c) show the AFM height

Surface modification of polyethylene terephthalate (PET) and oxide coated PET for adhesion improvement

ANALYTICAL METHODOLOGY DEVELOPMENT FOR SILICON-RICH-OXIDE CHEMICAL AND PHYSICAL CHARACTERIZATION

Local Anodic Oxidation with AFM: A Nanometer-Scale Spectroscopic Study with Photoemission Microscopy

Lecture 150 Basic IC Processes (10/10/01) Page ECE Analog Integrated Circuits and Systems P.E. Allen

Influence of silicon-nanocrystal distribution in SiO 2 matrix on charge injection and charge decay

U-Shaped Nano-Apertures for Enhanced Optical Transmission and Resolution

In-situ Monitoring of Thin-Film Formation Processes by Spectroscopic Ellipsometry

Graded SiO x N y layers as antireflection coatings for solar cells application

Asymmetrical heating behavior of doped Si channels in bulk silicon and in silicon-on-insulator under high current stress

Transparent Electrode Applications

THE INTERPRETATION OF ELLIPSOMETRIC MEASUREMENTS OF ION BOMBARDMENT OF NOBLE GASES ON SEMICONDUCTOR SURFACES *

Characteristics of a Silicon Wafer <111> and <004> after Planting Nitrogen

Crystalline to amorphous transition and band structure evolution in ion-damaged silicon studied by spectroscopic ellipsometry

Oxide growth model. Known as the Deal-Grove or linear-parabolic model

Supporting Information

Supporting information for: Semitransparent Polymer-Based Solar Cells with. Aluminum-Doped Zinc Oxide Electrodes

SCALE-UP OF THE BARIUM TITANATE ATOMIC LAYER DEPOSITION PROCESS ONTO 200 MM WAFER

Figure 1: Graphene release, transfer and stacking processes. The graphene stacking began with CVD

Semilab Technologies for 450mm Wafer Metrology

Supplementary Information. Experimental Evidence of Exciton Capture by Mid-Gap Defects in CVD. Grown Monolayer MoSe2

GISAXS STUDY OF CADMIUM SULFIDE QUANTUM DOTS

CVD: General considerations.

EE 527 MICROFABRICATION. Lecture 25 Tai-Chang Chen University of Washington

Effects of Antimony Near SiO 2 /SiC Interfaces

The characterization of MnO nanostructures synthesized using the chemical bath deposition method

PHYSICAL AND CHEMICAL PROPERTIES OF ATMOSPHERIC PRESSURE PLASMA POLYMER FILMS

PHOTOACOUSTIC TECHNIQUE FOR MEASURING BAND-GAP ENERGY OF POROUS SILICON LAYER ON n-si SUBSTRATE. Department of Physics,

Secondary ion mass spectrometry (SIMS)

Cu 2 O/g-C 3 N 4 nanocomposites: An insight into the band structure tuning and catalytic efficiencies

Lecture 6 Plasmas. Chapters 10 &16 Wolf and Tauber. ECE611 / CHE611 Electronic Materials Processing Fall John Labram 1/68

Cathodoluminescence spectroscopy of nitrided SiO 2 Si interfaces

OFFICE OF NAVAL RESEARCH GRANT N J-1178 R&T CODE 413Q TECHNICAL REPORT NO. #55

Manufacturable AlGaAs/GaAs HBT Implant Isolation Process Using Doubly Charged Helium

Sensors & Transducers 2015 by IFSA Publishing, S. L.

Manipulating Refractive Index in Organic Light. Emitting Diodes

Supporting Information. Plasmon Ruler for Measuring Dielectric Thin Films

Ellipsometry. Introduction. Measurement Principles

Bandgap engineering through nanocrystalline magnetic alloy grafting on. graphene

This document is downloaded from DR-NTU, Nanyang Technological University Library, Singapore.

Quantum Dots for Advanced Research and Devices

FTIR INVESTIGATION OF THE AGEING PROCESS OF CARBON NANOWALLS

Characterization of organic low-dielectric-constant materials using optical spectroscopy

Nanoscale IR spectroscopy of organic contaminants

Microstructural characterization of plasma-deposited SiO2-like thin films a detailed study by means of ellipsometric porosimetry

CHARACTERIZING PROCESS SEMICONDUCTOR THIN FILMS WITH A CONFOCAL MICRO X-RAY FLUORESCENCE MICROSCOPE

UNIVERSITY OF CALIFORNIA. College of Engineering. Department of Electrical Engineering and Computer Sciences. Professor Ali Javey.

Supplementary Information

A comparison of the defects introduced during plasma exposure in. high- and low-k dielectrics

SHG Spectroscopy. Clean surfaces Oxidation SOI wafer

TMT4320 Nanomaterials November 10 th, Thin films by physical/chemical methods (From chapter 24 and 25)

Transcription:

Bulg. J. Phys. 39 (2012) 178 185 Oxidation of Si Surface Nitrogenated by Plasma Immersion N + Ion Implantation A. Szekeres 1, S. Alexandrova 2, E. Vlaikova 1, E. Halova 2 1 Institute of Solid State Physics, 72 Tzarigradsko Chaussee, Sofia 1784, Bulgaria 2 Technical University Sofia, 8 Kl. Ohridski Blvd., Sofia 1797, Bulgaria Received 5 May 2012 Abstract. Oxidation of nanoscaled Si surface layer modified by shallow low energy nitrogen plasma-beam ion implantation is studied. By postimplantation oxidation in dry O 2 ambient SiO xn ylayers are synthesized. Results on the kinetics of SiO xn y formation based on Deal&Grove oxidation model are presented. Oxidation rate decrease with ion fluence is observed and explained with nitrogen induced process retardation. The optical and structural properties of the layers are examined by FTIR and ellipsometric spectroscopies determining the layers as a mixture of SiO 2 and Si 3N 4with ratio of volume fractions varying with ion fluence. In the FTIR spectra a broad band between 650 and 1000 cm 1 associated to asymmetric stretching mode of Si N is detected supporting the ellipsometric modelling. PACS codes: 68.35.-p, 81.16.-c 1 Introduction Advanced silicon materials research is a prerequisite to meet the requirements of future nanoelectronics and microphotonics. With adjustable structure and tunable refractive index silicon oxynitride provides a flexibility to fabricate active and passive devices on a single Si wafer. It has been shown [1] that the properties of SiO x N y layers strongly depend on the degree of nitridation (oxygen to nitride ratio x/y). Nitrogen plasma-beam ion implantation (PII) modification of Si surface and post-implantation anneal offer an alternative to traditional CVD deposition methods with potential to produce high quality Si-based nanostructures with easily controlled parameters. Excellent electrical characteristics were reported for layers obtained by implantation of 3 kev N + ion [2]. The purpose of this paper is to report the synthesis of nanoscaled SiO x N y layers on a Si substrate using low energy PII. Degree of nitridation using PII can be 178 1310 0157 c 2012 Heron Press Ltd.

Oxidation of Si Surface Nitrogenated by Plasma Immersion N + Ion Implantation controlled more accurately as compared to annealing the oxides in nitrogen containing ambient. At the first stage of synthesis the Si surface layer was modified by introducing nitrogen with different ion fluences. The second stage involved a high temperature oxidation of the Si substrates under oxidizing conditions, so that the incorporated nitrogen atoms together with the oxygen atoms form SiO x N y layers. 2 Experimental Details The experiments were carried out on Cz-grown (100)p silicon substrates, with resistivity 5 8 Ohm.cm. The Si wafers were bombarded with nitrogen ions through PII implantation. The implantation was accomplished at ion energy of 4 kev and fluencies ranging from 10 16 to 10 18 cm 2. In order to synthesize SiO x N y layers, the implanted Si substrates were oxidized in a conventional furnace at atmospheric pressure in dry oxygen at 1050 C for different durations. Simultaneously, SiO 2 reference layers were also grown into unimplanted Si wafers, as their thickness and refractive index served for comparison to those grown into implanted Si. Before oxidation, possible surface contaminations were removed in diluted HF. The optical and structural properties of the layers were investigated by spectroscopic ellipsometry (SE) and infrared (IR) spectroscopy. The ellipsometric measurements were carried out with a Rudolph Research manual variable-angle ellipsometer with polarizer-compensator-sampleanalyzer configuration in the spectral range of 300 640 nm and at incidence angles of 65, 70 and 75. The accuracy of the polarizer, analyzer and incidence angle was within ±0.01. The refractive index (n) of the grown layers were calculated by solving the inverse problem of ellipsometry [3]. The accuracy in determination of film thickness was ±0.2 nm, while the accuracy of the n values was ±0.005. Further, the obtained refractive index dispersion data were analyzed by Brugemann effective medium approximation (BEMA) giving information about the layers composition. The formed layers were considered as a physical mixture of known components of Si, SiO, SiO 2, Si 3 N 4 and voids. The dielectric functions of the components were taken from the literature [4,5]. The infrared (IR) spectra were recorded using a Perkin-Elmer 1430 IR spectrophotometer in transmission mode at normal incidence. The spectra were collected at room temperature in the spectral range from 400 to 4000 cm 1 with a resolution of 4 cm 1. 3 Results and Discussion The thickness of the layers grown on N + implanted Si substrates showed a tendency to decrease by the increase of the implanted ion fluence. This tendency 179

A. Szekeres, S. Alexandrova, E. Vlaikova, E. Halova Layer thickness (nm) 70 60 50 40 10 16 cm -2 30 20 10 10 18 cm -2 0 2 4 6 8 10 12 14 16 18 20 22 1017 cm -2 Time (min) Figure 1. Growth rate of the synthesized layers in oxidizing ambient at 1050 C for different N + ion fluences. is well seen in Figure 1, where the thickness dependences on oxidation time after their approximation by using the Deal&Grove oxidation model of Si [6] are presented. According to this model the oxidation rate is described through the equation x 2 0 + Ax 0 = B(t + τ), where B is the so called parabolic oxidation rate, B/A is the linear oxidation rate and τ takes into account any oxide thickness at the start of the oxidation process. It is assumed that the application of this model is relevant to our case since it assumes that the oxidation is limited by two steps, namely chemical oxidation rate at the substrate surface and diffusion of the oxidant through the already formed layer. These assumptions are general and we believe that they can be applicable to the oxidation of implanted Si. In Figure 1 the thickness dependence on N + fluences is evident, as with increasing the N + fluence the thickness of the formed layer decreases. The approximation parameters determine the growth kinetics constants, which are summarized in Table 1. For comparison purposes the data for unimplated oxidized wafer are also given in the Table 1. It is evident that for oxidation of implanted Si the parabolic rate is substantially higher and the linear rate smaller compared to that for oxidation of unimplated Table 1. Growth kinetics parameters for different N + ion fluences N + ion fluence Linear rate constant Parabolic rate constant (cm 2 ) (nm/min) (nm 2 /min) 0 20.3 80.9 10 16 4.4 769.2 10 17 7.0 1030.0 10 18 7.5 1119.0 180

Oxidation of Si Surface Nitrogenated by Plasma Immersion N + Ion Implantation Si. This observation can be understood in terms of two oxidation rate limiting processes. Obviously, the oxidation of the implanted Si is mainly determined by the oxidant diffusion through the already formed oxide layer. It can be suggested that the diffusion that proceeds through the damaged Si topsurface layer, is faster than for unimplanted Si. The obtained smaller linear rate constants confirm this suggestion indicating less influence of the chemical reaction rate at the Si surface. The oxidation rate constants characterize the formation of the layers as diffusion limited process through the Si region modified by the PII. This process is competitive with the retardation caused by the N atoms, especially at the start of the oxidation process. The oxidation is accompanied with flattening of the implanted profile [7] due to the out-diffusion of N atoms on one hand and, their diffusion deeper into Si on the other hand, broadening the N profile and causing an overall decrease of N concentration. After about 20 min oxidation the nitrogen in the Si is exhausted and the thickness of the formed layer approaches the 1,56 10 18 cm -2 10 min n 1,53 1,50 10 17 cm -2 10 16 cm -2 1,47 SiO 2 (a) 300 400 500 600 1,50 10 18 cm -2 10 17 cm -2 1,48 λ (nm) 20 min n 10 16 cm -2 1,46 SiO 2 (b) 300 400 500 600 λ (nm) Figure 2. Refractive index of the layers grown on p-si(100) as a function of wavelength for different N + ion fluences. 181

A. Szekeres, S. Alexandrova, E. Vlaikova, E. Halova value for the unimplanted Si. The role of the nitrogen profile and concentration is thus decisive for the exact stoichiometric content of the SiO x N y layer, which is further discussed based on the optical characterization of the layers. Increasing the fluence further up to 10 18 cm 2, the oxidation rate strongly increases, which could be ascribed to the plasma-beam generated defects on the Si surface that facilitates the oxygen diffusion and formation of an oxide layer. In Figure 2a the refractive index dispersion data is displayed for 10 min oxidation. The n spectrum for the oxide layer grown on unimplanted Si wafer is also presented, the n values are typical for stoichiometric SiO 2 thermally grown at 1050 C [5]. The closely grouped n spectra for the lower implant fluences point out the similarity of the layers microstructure. By increasing the N + fluence to 10 18 N + /cm 2 the n-spectrum showed an essential difference from that of SiO 2 layer giving evidence for silicon oxynitride layer formation. The EMA modeling of the component fractions in these layers found that they were free of voids and consisted of SiO 2 and Si 3 N 4 phases the ratio of which was N + dose dependent. For N + fluence of 10 18 cm 2, the volume fraction of Si 3 N 4 was the highest and was about 12 %. Prolonged oxidation (20 min) resulted in close thickness values (Figure 1) indicating for a significantly weakening of the N + dose influence. The refractive index spectra in Figure 2b did not show clear dependence on implantation dose and the resemblance of the obtained spectra suggested similarity of the layers microstructure, close to reference layer grown on unimplanted Si. The optical modelling revealed that in this case the nitride volume fraction is rather low, varying within 1-3 v.% with a tendency to increase by increasing the N + fluence. This pointed out that the structure of the grown layers after 20 min became silicon oxide enriched with nitrogen. The correlation between the refractive index value and the volume fraction of Si 3 N 4 in the layers formed into N + implanted Si wafers was linear, as the n value increased with the structure moving away from stoichiometric SiO 2 (Figure 3). Experimental support for the formation of SiO x N y layers by the applied technological processes was found from the IR transmission measurements. In Figure 4 the IR spectra of the layers grown during 10 min oxidation into implanted Si wafers are presented. The IR spectra of the layers grown for longer time (20 min) were very similar to that of a characteristic SiO 2 spectrum and, therefore they are not considered herein. No bands connected to Si H (2170 cm 1 ), N H (3350 cm 1 ) or O H (3640 cm 1 ) bonds were detected in all the measured IR spectra. Taking in view this and the observation that the characteristics vibration bands of Si-O and Si-N chemical bonds appear in the spectral range below 1400 cm 2 [8], data in Figure 4 is given up to 2000 cm 1. Emergence of the broad band in the wavenumber range of 650 1000 cm 1 is an evidence of SiO x N y formation (Figure 4). This band is observed in all spectra and is attributed to Si-N stretching modes. Band related to Si-N stretching bonds 182

Oxidation of Si Surface Nitrogenated by Plasma Immersion N + Ion Implantation 1,54 λ=632.8 nm 1,52 n 1,50 1,48 1,46 0 4 8 12 16 20 24 Si 3 N 4 fraction (%) Figure 3. Correlation between the refractive index value and the volume fraction of Si 3N 4 in the layers formed into N + implanted Si wafers. overlapping into a broad band around 940 cm 1 has been reported [9]. Also, Si-N related bands in complex spectra centered at 820 cm 1, 880 cm 1 and 915 cm 1 have been observed in buried Si 3 N 4 films [10]. Moreover, a peak observed at 960 cm 1 has been attributed to doubly bounded N atoms with two Si atoms without affecting by oxygen atoms [11]. The broad and intense band centered around 1080 cm 1 in the spectra (Figure 4) is related to the asymmetric stretching mode of Si-O-Si group. Deconvolution of this band revealed two Gaussian peaks (not shown here), one centered at 1060 cm 1 and connected with stretching modes of the Si O Si bonds in SiO x and the other centered at 1096 cm 1 and attributed to the asymmetrical stretching of Si O Si bonds in stoichiometric SiO 2 [12]. It should be noted that with Transmittance (arb. u.) 10 18 cm -2 10 17 cm -2 10 16 cm -2 400 800 1200 1600 2000 Wavenumber (cm -1 ) Figure 4. IR spectra of SiO xn y layers grown during 10 min into p-si(100) implanted with different nitrogen fluences. 183

A. Szekeres, S. Alexandrova, E. Vlaikova, E. Halova increasing the N + fluence the intensity of the 1060 cm 1 band increased at the expense of the 1096 cm 1 band, most probably due to the increased implantation damage and increased SiO x formation [12,13]. The band of Si-O-Si bending vibration mode at 800 cm 1 overlaps the broad Si-N band. The rocking mode of Si-O-Si bonds at 460 cm 1 is also present in the spectra and they are known for not being influenced by structural changes [12]. The band at 614 cm 1 is related to vibration mode of Si-Si bonds, an indication for the presence of Si-Si inclusions in the oxide network. After 20 min oxidation, when the whole N + implanted Si region (< 50 nm [7]) is fully oxidized, the IR spectra became unified, having strong characteristic peaks of Si-O vibrational modes and very weak vibrational modes of Si-N related bonds. From these results the layers grown into the N + implanted silicon can be characterized as nanoscaled silicon-rich SiO x N y films with relatively low N content. This conclusion is supported by the results from the extended in the IR region ellipsometric measurements [7], which have registered a shoulder at about 1150 cm 1 in the Ψ(ν) spectra even for the layers grown during 20 min oxidation, connected with Si-N bond vibration mode. 4 Conclusions The results gained by means of spectroscopic ellipsometry and FTIR spectra analysis reveal that under the conditions of our experiments the grown layers can be characterized as silicon rich SiO x N y with relatively low N content. The oxidation kinetics was considered using the Deal&Grove oxidation model of Si, revealing that for oxidation of implanted Si the parabolic rate was substantially higher and the linear rate smaller compared to that for oxidation of unimplated Si. The observed decrease of growing layer thickness with increasing the implanted N + fluence was explained with the retardation effect of nitrogen atoms. The refractive index values was found to increase with increasing the N + ion fluence due to increase of the volume fraction of Si 3 N 4. For the highest N + fluence of 10 18 cm 2, the amount of Si 3 N 4 was the maximum of 12 v.%. References [1] M. Abu El-Oyoun, T. Inokuma, Y. Kurata, and S. Hasegawa (2003) Solid-St. Electron. 47 1669. [2] E. Kapetanakis, D. Skarlatos, C. Tsamis, P. Normand, and D. Tsoukalas (2003) Appl. Phys. Lett. 82 4764. [3] R.M.A. Azzam, N.M. Bashara (1987) Ellipsometry and Polarized Light, North- Holland, Amsterdam. [4] D.E. Aspnes and A.A. Studna (1983) Phys. Rev. B 27 985. 184

Oxidation of Si Surface Nitrogenated by Plasma Immersion N + Ion Implantation [5] E.D. Palik, (ed.) (1994) Handbook of Optical Constants of Solids, New York: Academic Press. [6] B.E. Deal and A.S. Grove (1965) J. Appl. Phys. 36 3770. [7] S. Alexandrova, E. Halova, A. Szekeres, E. Vlaikova, M. Gartner, and M. Anastasescu (2012) J. Phys.: Conference Series 356 012031. [8] S.H. Mohamed (2011) Physica B 406 211. [9] R.Y. Honda, R.P. Mota, R.G.S. Batocki, D.C.R. Santos, T. Nicoleti, K.G. Kostov, M.E. Kayama, M.A. Algatti, N.C. Cruz, and L. Ruggiero (2009) J. Phys.: Conference Series 167 012055. [10] M.M. Larijani and G. Haidari (2010) World Appl. Sci. J. 8 972. [11] F. Rebib, E. Tomasella, E. Bêche, J. Cellier, and M. Jacquet (2008) J. Phys.: Conference Series 100 082034. [12] P. Lange, U. Schnakenberg, S. Ullerich, and H.-J. Schliwinski (1990) J. Appl. Phys. 68 3532. [13] Y. Rui, D. Chen, J. Xu, W. Li, Zh.Cen, X. Huang, and K. Chen, (2007) Appl. Surf. Sci. 253 8647. 185