AVS2001 Session PS-MoP: Plasma Diagnostics and Plasma-Surface Interactions Poster Session
Monday, October 29, 2001 5:30 PM in Room 134/135
Monday Afternoon
Time Period MoP Sessions | Topic PS Sessions | Time Periods | Topics | AVS2001 Schedule
|
PS-MoP-1 Molecular Structure of Fluorocarbon Deposits Analyzed by EIMS and CIMS Combined with Thermal Desorption Technique
N. Takada, K. Sasaki, K. Shibagaki, K. Kadota (Nagoya University, Japan) In order to understand plasma-surface interaction in fluorocarbon plasmas, the analysis of fluorocarbon deposits is an important issue. To date, x-ray photoelectron spectroscopy (XPS) and Fourier transform infrared spectroscopy (FTIR) are exclusively adopted for the analysis of fluorocarbon films. However, detail of the molecular structure of fluorocarbon films cannot be understood by XPS and FTIR. In the present work, we adopted electron-impact mass spectrometry (EIMS) and chemical ionization mass spectrometry (CIMS) for the analysis of the molecular structure of fluorocarbon films produced by C4F8 and C4F8-H2 plasmas. A thermal desorption technique was used to evaporate fluorocarbon molecules from samples. In EIMS, thermally desorbed molecules were ionized by an electron beam with an energy of 70 eV. In CIMS, isobutene is ionized by electron impact to produce slow electrons. Negative fluorocarbon ions were produced via dissociative electron attachment of the slow electrons. The threshold temperature for thermal desorption of fluorocarbon molecules from a sample produced by a C4F8 plasma was 100-160 °C. Many peaks corresponding to mass numbers up to 800 were observed in the EIMS spectrum. The major peaks were classified into two groups having structures of CF3(CF2)n+ and CF(CF2)m+. This result indicates that molecules desorbed from the sample contain the polymerized structure of CF2. The overall shape of the CIMS spectrum of the fluorocarbon sample was similar to that of perfluorokerosene (PFK), which has a molecular structure of CF3(CF2)nCF3. This suggests that fluorocarbon molecules desorbed from the sample have a molecular structure similar to PFK. By comparing the CIMS spectra of the sample and PFK, the mass number of the desorbed molecule was estimated to be 1100. |
|
PS-MoP-2 Real Time Analysis of the Remote Oxygen and Hydrogen Plasma Cleaning using Mass Spectroscopy
H. Soh, H. Seo, Y. Kim, H. Jeon, Y.C. Kim (Hanyang University, Korea) The plasma cleaning technologies have been attracting a great attention due to the demands of the compatible process with the most cluster tool and of the environmentally safe process. The photo resist (PR) ashing and PR strip processes are generally followed the silicon etching process to remove the PR and polymerized residues, respectively, during integrated circuit fabrication. However, the PR strip is the wet chemical process and causes environmental problems. Especially, the polymerized residues formed at the contact and via holes during the photo resist (PR) ashing and PR strip processes must be removed prior to the metal contact. In this study, we continuously monitored and systematically analyzed the volatile gases from the oxidized PR molecules during the low temperature remote plasma cleaning process. Mass spectroscopy (QMS200) was used for the real time monitoring of the volatile gases containing carbon and fluorine. In-situ Auger electron microscopy, X-ray photoelectron spectroscopy, atomic force microscope analysis systems were used to evaluate the cleaning effects and to avoid recontamination such as carbon absorption in the air. The surface morphologies of the samples before and after plasma cleaning were also observed using scanning electron microscope. This paper will present the oxygen and hydrogen remote plasma cleaning efficiency and its chemical reaction mechanisms.
|
|
PS-MoP-3 Investigation and Modeling of Plasma-Wall Interactions in Inductively Coupled Fluorocarbon Plasmas and the Effects of Chamber Dimension
E.A. Joseph, S.P. Sant, L.J. Overzet, M. Goeckner (University of Texas, Dallas); M.J. Kushner (University of Illinois, Urbana Champaign) Plasma-wall interactions in fluorocarbon based feedgas chemistries, namely CF4, are examined in both a standard and a modified inductively coupled Gaseous Electronics Conference (GEC) reference cell using In-Situ Fourier Transform Infrared Spectroscopy (FTIR). Initial measurements in the standard GEC reference cell show the dissociation of the CF4 feedgas into radical CFx species as has been observed elsewhere.1 Experimental results from both the standard GEC reference cell as well as the modified GEC reference cell, which differs from the standard cell in that it has easily configurable plasma exposed surfaces,2 are compared to results from the Hybrid Plasma Equipment Model3 to better elucidate the influence of the plasma exposed surfaces on plasma parameters including, but not limited to, plasma density, wall temperature, and CFx polymer thickness. This material is based upon work supported by the National Science Foundation under Grant No. 0078669 |
|
PS-MoP-4 Controllable Ion Source for Process Enhancement in a Downstream Plasma Ash Chamber
A.K. Srivastava, P. Sakthivel, I. Berry (Axcelis Technologies, Inc.); H.H. Sawin (Massachusetts Institute of Technology) Plasma ash tools have in the past been designed to generate a gentle downstream plasma impinging on the wafer. However, additional demands to remove residue, both in front-end and back-end processes have recently been imposed on these ash tools. This has led to a gradual transition of the old plasma asher into a ash-etch hybrid tool that can tolerate harsh etchants in moderate amounts, and also have the flexibility of controlling the ion flux to the wafer without charge damage. Different approaches have been attempted to meet these needs with limited success, mostly by taking the chamber towards the reactive ion etch (RIE) regime where the risk of device damage is high. A new approach to the dynamic control of ion generation and transport to the wafer, while maintaining operating regimes that are conducive to an ash tool and hence non-destructive to the wafer, is reported in this paper. In a typical downstream plasma ash tool, an upstream microwave discharge is fed into a plenum that disperses the reactive species through a hard-coated aluminum baffle plate above the wafer. When ion bombardment is required, a secondary ion source coupling radio frequency (RF) is switched on, creating a glow discharge in the plenum above the baffle plate. Under optimal conditions, electrons from the glow discharge are trapped inside the sheath of the baffle plate holes. These electrons complete the current path on the bottom surface of the baffle plate, and have sufficient energy to ionize some of the neutral gas. This secondary source of ions provides a gentle, low energy, uniform bombardment on the wafer. This paper describes the ion source in detail as it is used during specific ash/residue removal processes. The unique ion generation mechanism is discussed. Preliminary results of ion bombardment on the removal of the carbonized layer of implanted resist at low temperature are presented. Optical emissions data characterizing the ash chemistries are also presented. |
|
PS-MoP-5 A Multi-Technique Investigation of the Pulsed Plasma and Plasma Polymers of Acrylic Acid, Propanoic Acid and Hexamethylenedisiloxane
S. Fraser, D. Barton (University of Sheffield, England); A.J. Roberts (Kratos Analytical, England); R.D. Short (University of Sheffield, England) The synthesis of thin plasma polymer films from radio frequency (rf) sustained glow discharges of small organic compounds is well documented. Films containing a high degree of retention of the starting monomers original functionality and structure can be deposited using a low power plasma. These plasmas can be sustained by continuous wave (CW) or by pulsing a higher input power to achieve a lower average power. Plasma polymers of acrylic acid, propanoic acid and hexamethylenedisiloxane have been fabricated from pulsed 13.56 MHz RF plasmas in a "capicatively" coupled plasma deposition chamber. Plasma "on" and "off" (ton/off) times in the millisecond (ms) pulse time regime have been investigated using a peak plasma power of 50 W. Employing a fixed ton (5ms) the effect of toff (0-2000 ms) on the solid-phase plasma polymer has been investigated using X-ray photoelectron spectroscopy (XPS) and time-of-flight secondary ion mass spectrometry (ToF SIMS). Mass spectrometry has been employed to monitor the neutral and charged species in the plasma-gas as a function of toff and deposition rates have been monitored by means of a quartz crystal mass balance (QCMB). For comparison continuous (CW) wave plasmas of equivalent "average power" have been studied. The XPS (C1s core line) show that with increased toff the extent of functional group retention increased. Valence band XPS and SIMS data indicate at low "average" powers (0.2 W) a linear polymer, closely resembling the conventional polymer, is deposited. The mass spectral data indicate the principal affect of toff is on the amount of intact monomer in the system. For example, with acrylic acid this was shown by monitoring the signal from the molecular ion of acrylic acid (m/z 72) in the neutrals. |
|
PS-MoP-6 Synthesis and Characterization of New Material (BON) for Semiconductor Applications
G.C. Chen, S.-B. Lee, J.-H. Boo (Sungkyunkwan University, Korea) Recently, oxynitride compounds sparked among functional materials, such as Ti(ON), Si(ON) and SiBON. They presented a promising potential application of diffusion barrier, solar material and interlevel dielectric material. Since BON was declared as a by-product in synthesis of hard material, we have firstly developed the research on the molecular structure of new material (i. e. BON). In this study, we would like to report on the electrical property of BON thin films grown on Si(100) substrates by low frequency RF derived plasma assisted CVD. The effect of growth condition, such as flux of fed gas, deposition temperature as well as growth time, on the I-V characteristics are mainly discussed. The experimental results showed that the electrical resistance decreased with increasing the nitrogen flux and growth time. Amorphous BON thin film grown at relatively low temperature has more higher resistance than microcrystal containing thin film that obtained at high temperature, and the most smooth morphology was benefit of getting low resistance. By controlling nitrogen content in the film layer it can be possible to making the BON thin films that have either semiconducting or insulating properties. The as-grown new materials were characterized by XPS, RBS, FTIR, AFM, and TEM/TED. During CVD, moreover, the optical emission spectra (OES) were also measured in-situ for plasma diagnostics and analysing gas phase reaction. Based on the OES result, we confirmed that the formation of BON thin film was suit to be performed under nitrogen plasma. |
|
PS-MoP-7 The Effect of Controlling the Ion Energies of a Plasma Polymerised Deposition Surface Upon the Film Structure
D. Barton, R.D. Short (University of Sheffield, UK); J.W. Bradley (UMIST, UK) The formation of plasma polymerised films has been the subject of intensive research over many years. Despite this work and the technological relevance of such films, little attention has been paid to the actual mechanisms involved in both the polymer formation and deposition. In particular, the question of whether the polymer is formed either in the bulk plasma, the sheath region, or at the surface remains unresolved. We present an experimental technique which allows the independent control of positive ion energies at a deposition surface. The technique relies upon feeding an RF signal to the surface which is matched in both phase and amplitude to the RF potentials in the plasma. In this way, it is possible to control the energy of the positive ions at the surface, independently of other plasma parameters. Ions may be a significant constituent of polymer films as previous measurements of the ion flux to a deposition surface have suggested that these particles may be responsible for about 25% of the total polymer mass. Embedded into the deposition surface are plasma diagnostics including a plasma mass spectrometer, quartz mass balance and an ion flux probe. The deposited films are analysed using XPS. Using an acrylic acid monomer, we will relate energy changes of ions striking the surface to the film properties such as extent of cross linking and functionality retention. |
|
PS-MoP-9 Uniformity Control of Electron Temperature and Density within a Commercial-scale Helicon Plasma Processing Reactor
M.J. Neumann, J.E. Norman, D.N. Ruzic (University of Illinois at Urbana) A study of electron temperature and density profiles was performed with PlasmaQuest 256 helicon plasma processing research reactor to better understand power absorption within a plasma reactor operating in helicon mode. The plasma source is a PMT Mori 200 antenna, coupled with an inner and outer opposing magnetic field coil placed around the it and a third coil positioned below it. The use of external magnetic field coils gives the ability to quickly change from one operating mode and condition to another. External electrical and magnetic variation effects are observed with a radially and z-varying Langmuir probes and spectroscopy. For nitrogen at 800 W RF and 80 A on the inner and lower magnets, the electron density profile was seen to decrease inward from 1.1x1011 cm-3 at the edge of the chamber to 6.0x1010 cm-3 in the center, while the electron temperature increased inward from 1.0 eV to 4.8 eV at the center of the chamber. In contrast, at 800 W RF and 40 A on the inner magnet alone, the electron density profile was seen to increase inward from 2.0x1010 cm-3 at the edge of the chamber to a peak of 7.0x1010 cm-3, while the electron temperature decreased inward from 4.0 eV to 2.6 eV at the center of the chamber. Through manipulation of electron density and temperature, the surface properties of various biopolymers can be changed, while leaving the underlying material unchanged. One observable change is a decrease in the water contact angle with common biopolymers, such as low-density polyethylene and high-density polyethylene. These changing plasma conditions have also been observed with argon and oxygen. Modeling has been performed to show the physical mechanisms involved in the electromagnetic energy transfer to the plasma. Thus, within the same reactor chamber and via only external manipulation, very distinct plasma conditions yielding different interaction properties can be produced. |
|
PS-MoP-10 Atomic-Order Plasma Nitridation of Ultrathin Silicon Dioxide Films
T. Seino, T. Matsuura, J. Murota (Tohoku University, Japan) Atomic-order nitridation of SiO2 by a nitrogen plasma without substrate heating has been investigated using an ECR plasma apparatus. A 3nm-thick SiO2 was thermally grown by wet oxidation of Si(100) at 700°C. The SiO2 film was nitrided by the nitrogen plasma for 1-324 min at the N2 pressure of 1.3-2.6Pa with the microwave power of 200W. After plasma nitridation, some samples were annealed at 400-800°C for 1hour in nitrogen. The depth profile was obtained by the repetition of etching by a 1%-diluted HF solution and XPS measurements. When the incident ions were exposed directly on the surface, the number of the incident ions (the product of the ion density and the plasma exposure time) was 3x1015-9x1018cm-2, and the N1s peaks were observed at 397-398eV and at 402eV. With increasing nitrogen plasma exposure time, the N1s peak at 398eV shifted to 397eV corresponding to Si3N4. By annealing at 400-800°C after nitridation, the N1s peak at 402eV disappeared. The depth profiles for the N1s peak at 397-398eV were almost the same before and after annealing. By a shutter placed above the wafer, the number of the incident ions was suppressed to 5x1014-1x1016cm-2 and the ion energy (below 20eV) became lower than that (below 30eV) without the suppression of the incident ions, and the N1s peak was observed only at 399eV. By annealing at 600-800°C after nitridation, the N atom concentration at 399eV was decreased due to the diffusion of the N atoms. This result is similar to the thermal nitridation of SiO2 by NH3. It is also found that the N atom concentration was normalized by the number of the incident ions in spite of the ion energy and the radical density. From these results, it is considered that the nitridation is caused mainly by the incident ions. It is suggested that the ion energy changes the binding conditions of N atoms in SiO2. |
|
PS-MoP-11 Plasma Processing Tests from a Large Area High Density Plasma Source Based on Electron Beam Ionization
D. Leonhardt (Naval Research Laboratory); S.G. Walton, D.D. Blackwell (National Research Council); D.P. Murphy, R.F. Fernsler, R.A. Meger (Naval Research Laboratory) Electron beam ionization is both efficient at producing plasma and scalable to large area (square meters) when the electron beam is magnetically collimated. The beam ionization process is also fairly independent of gas composition, capable of producing low temperature plasma electrons (Te ~ 0.5eV in molecular gases, 1-2eV in atomic gases) in high densities (ne ~ 109-1012 cm-3).1 A 'Large Area Plasma Processing System' has been developed which consists of a planar plasma distribution generated by a sheet of 2-5kV electrons injected into a neutral gas background. The electron beam is magnetically collimated by a 150 Gauss field and operates in a gas pressures of 20-100 mtorr. A photoresist (PR) ashing process was studied under various system conditions In order to test plasma processing capabilities and control over processing parameters. The process system consisted of pure O2 or O2/Ar gas mixtures interacting with PR coated silicon substrates which were exposed to the plasma on an rf biasable (capacitively coupled) stage. Process conditions such as gas mixture composition, operating pressure, beam-to-substrate distance, duty cycle and incident ion energy were varied to determine the various effects of this plasma source on the PR ashing process. In situ Langmuir probe and mass spectrometry data2 are correlated to the material removal as determined by ex situ surface analysis (profilometry, SEM). Results show that the material removal process is ion energy dependent and highly anisotropic for ion energies ≥50eV. Uniformity tests of the large area source will also be presented if time permits. Work Supported by the Office of Naval Research. |
|
PS-MoP-12 Energy Distributions of Ions and Neutrals from a Sputter Source
G.J. Peter, N. Mueller, H. Zogg (Inficon Limited, Principality of Liechtenstein) The kinetic energy of the deposited particles impinging on the surface of the substrate is an essential process parameter in sputter deposition processes. The energy as well as the mass of these particles can be measured with an energy selective mass spectrometer. Ions from the plasma can directly and easily be guided by an Ion Transfer Optics into the energy and mass filter resulting in a high sensitivity. In contrast to this, neutral particles have to be ionised first, before any energy or mass filtering can be done. It is common to use electron impact ion sources with well defined electrical potentials for ionisation, although this is a low efficiency process. For optimisation of the sensitivity for neutral particles, a wide acceptance angle of the ion is mandatory. This can be obtained by minimising the distance between the entrance aperture and the ion source. But this design is in no way optimum for the detection of ions, as at least the ion formation chamber of the ion source is a poor ion optics. So any design of a combined instrument will result in a poor sensitivity for the ions and a high sensitivity for the neutrals and vice versa. In order to overcome these difficulties, an otherwise standard plasma process monitor PPM 422 was equipped with two exchangeable, electrically insulated entrance apertures. One of it was used for the measurement of ions and the other one for detection of neutrals. The electrical insulation allows to set the aperture to any electrical potential. So a realistic simulation of the potentials on the substrate ca be performed. The energy distributions of particles from a gun type planar magnetron were measured with this modified device. |
|
PS-MoP-13 Species Characterization in Inductively Driven Fluorocarbon Etch Plasmas
G.A. Hebner, I.C. Abraham (Sandia National Laboratories) A number of techniques have been used to characterize etching plasmas containing fluorocarbon gases. Laser induced fluorescence was used to measure the spatially resolved SiF and SiF2 densities in inductively driven discharges containing C2F6 and C4F8. Measurements of the spatially resolved SiF and SiF2 densities were performed as functions of the induction coil power, pressure, and bias power above a silicon surface. The SiF density had a maximum at a radial distance of 2 - 3 cm from the center of the plasma, and then monotonically decreased towards the edge of the plasma region. The SiF2 density had a maximum at a radial distance of approximately 7 cm from the center of the plasma. Electron and negative ion densities was measured in C4F8 and the identity of the negative ions in C2F6, CHF3 and C4F8 containing discharges was investigated using a novel photodetachment experiment. To investigate the influence of surface material, the rf biased electrode was covered with a silicon wafer or a fused silica (SiO2) wafer. In most cases, the trends in the electron and negative ion density were independent of the wafer material. A novel microwave resonant cavity structure was developed to identify the negative ions using laser photodetachment spectroscopy. Unlike traditional microwave cavity techniques, this method offers the possibility of spatial resolution. This work was supported by SEMATECH and the United States Department of Energy (DE-AC04-94AL85000). |
|
PS-MoP-14 Investigation on the Plasma Uniformity in Reactive Gas ECR Plasmas
M. Shindo, Y. Kawai (Kyushu University, Japan) Reactive gas plasmas are widely used in etching process for fabricating ULSI. In order to reduce the production costs, a uniform and high-density plasma with a large diameter is required. An ECR plasma can be a candidate for such a plasma, since it easily reaches more than 1012cm-3. In addition, one of the mechanism of the uniformity in an Ar ECR plasma was clarified by Ueda and Kawai.1 On the other hand, the transportation and diffusion mechanism in reactive gas plasmas become complicated since a large quantity of negative ions are produced. Thus, it is necessary to investigate the plasma uniformity under the existence of much negative ions. In this study, an attempt to measure the spatial distribution of charged species in O2/Ar and C4F8/Ar ECR plasmas was made. Here, the density of negative ions was estimated from the phase velocity of the ion acoustic waves (fast mode) which were launched from a wire antenna to which the positive pulse voltages (30µ sec in duration time and Vpp=40V) were applied, and were detected with a plane Langmuir probe biased at -90V. The detected signals were observed with an oscilloscope, and a trough and a crest were found, corresponding to the waves excited at the leading and falling edge, respectively. As a result, the negative ions existed near the wall rather than the center region. Moreover, it was found that the positive ion density was uniform in the center region, and the uniform area widened as the reactive gas mixture rate was increased. |
|
PS-MoP-16 A Novel Approach to Time Resolved Langmuir Probe Measurements
A.K. Jindal, S.K. Kanakasabapathy, M. Goeckner, L.J. Overzet (University of Texas at Dallas) Langmuir Probes, when used carefully, have provided spatially resolved plasma density, potential, and electron temperature measurements. However, time resolved Langmuir Probe measurements are time consuming in triggered box car averaged systems. We present a novel approach to time resolved Langmuir Probe I-V curve acquisition. Probe Current and voltage are measured as functions of time, for a fixed applied DC bias, with a digitizing oscilloscope. These traces are collected and a new applied bias is set. By doing this repeatedly, we build up I-V-t profiles. Then, in an offline fashion, time is eliminated between these various traces and I-V curves are obtained for all instances of time in the pulse period. In contrast to state of the art triggered box car averaged systems, I-V curves are obtained with better time resolution and much faster. For example: I-V curves for 512 instances of time within a 1 msec pulse period with 300 points in voltage, is obtained in as little as 15 minutes. Once these time resolved I-V curves are obtained, appropriate probe theory is used for extracting time resolved plasma parameters using a Levenberg-Marquardt1 fitting algorithm. The high speed of data acquisition minimizes the effects of long-term plasma drift associated with slower techniques. We shall be presenting data from pulsed Chlorine and Argon discharges using this technique. . |
|
PS-MoP-17 Evaluation of Langmuir Probe Theories via Comparison with Microwave Interferometry and Plasma Oscillation Probe Methods in ICPs
J.D. Evans, F.F. Chen (University of California, Los Angeles); W. Zawalski (Hiden Analytical, Ltd., England) Measurements of plasma density (Np) are performed under identical conditions using microwave interferometry (MWI), plasma oscillation probe (POP) and Langmuir probes. Np via MWI is treated as the "known" result to which the other techniques are compared. Data is obtained in a chamber that lies downstream from a PlasmaTherm ICP source, over a wide range of parameters: Prf<1kW, Te=2-5eV, Np=10^10 - 10^12 cm^-3. Ne obtained via POP agrees well with MWI in relatively low-collisional regimes, as expected. I-V curves are acquired using Hiden Analytical's ESPion Langmuir probe, which employs broadband RF compensation via inline chokes and a compensation electrode. A variety of probe tip radii are used, which allows for a broad base of comparison. Conventional probe theories (OML, ABR and BRL) are used, as well as an approximation based on the assumption that the ion current (Ii) is determined mainly by the unipolar sheath expansion, whose thickness follows a modified Child-Langmuir (C-L-mod) law in which Ii is nearly proportional to Vo^(4/3), where Vo = |Vbias-Vplasma|. Different probe theories yield vastly differing values of ion density (Ni), whereas Te is reliably determined in most cases. In most cases the dependence of Ii with Vo agrees very well the C-L-mod law. Excellent agreement between Ni thus obtained and the MWI results is observed throughout the parameter range. However, it is also found that Hiden's ESPsoft algorithm, which employs OML theory that predicts Ii^2 proportional to Vo, yields values for Ne in good agreement with those of MWI and POP. Reasons for the success of the C-L-mod law in light of this apparent paradox are presented. |
|
PS-MoP-18 Electron Energy Control in an Inductively Coupled Plasma by Means of Induction Field Reversal
H. Shindo (Tokai University, Japan); T. Urayama (ADTEC Plasma Technology CO., LTD, Japan); Y. Horiike (The University of Tokyo, Japan); S. Fujii (ADTEC Plasma Technology CO., LTD, Japan) In the deep submicron etching for ultra large-scale-integrated circuit (ULSI) processes, the development of low-pressure and high density plasma sources has been highly required. In these plasmas, however, several crucial problems have risen up in conjunction with electron energy. These are related to the fact that the electrons are prone to be excessively energetic in high density plasmas generated at low pressures. In this work, we studied electron energy control in an inductively coupled plasma by employing different azimuthal mode antenna. It is expected for electron energy to reduce at higher azimuthal mode antenna, because induction field is reversed with a shorter length. The experiments were carried out in an inductively coupled plasma (ICP) which was produced in a stainless-steel chamber of 350 mm in diameter by supplying the RF power of 13.56 MHz through the quartz window at one end. The three types of antenna were prepared.The electron temperature could be reduced by increasing the azimuthal mode number of one-turn antenna with no notable change in electron density. These effects were remarkable in low pressures below 10 mTorr, hence in a condition of longer electron mean free path. The mechanism for these behaviors is that the electron temperature at lower pressures depends on both the electron mean free path and the field reverse distance, as anticipated. The electron gains more energy with larger field reverse distance so long as the mean free path is same, and the electron mean free path is inversely proportional to the pressure.The RF magnetic field was also measured by a pick up coil, and its behavior was quite consistent with the electron energy, meaning that the electron energy change is not due to the capacitive effect. It was concluded that the induction field reverse distance in conjunction with electron free path was essential in electron energy control. |
|
PS-MoP-21 Novel In Situ Diagnostics for Plasma Processing of Advanced Materials
E.A. Evans, G. Zhang (University of Akron); A. Salifu (Cree Research) Kinetic and transport parameters for nitride growth and processing are required for process design and optimization. A new approach for accurately determining these parameters at processing conditions will be presented. We have combined a large inductively coupled plasma reactor with a digital microbalance to make measurements of substrate weight during plasma processing of aluminum and aluminum nitride samples. The current set-up is capable of sensing growth and/or etch rates on the order of 1 microgram per second regardless of substrate area. We will present results that demonstrate the usefulness of these measurements for understanding the kinetic and transport mechanisms involved during plasma processing. These relationships are important not only for controlling oxidation during and after deposition but also for identifying growth regimes for high quality thin film nitride materials. |
|
PS-MoP-22 Analysis of Chlorine-Containing Plasmas
G.F. Franz (Infineon Corp., Germany) Capacitively coupled discharges of strongly reactive atmospheres containing mixtures of boron trichloride and chlorine are investigated employing impedance measurements, Langmuir probe measurements, optical emission spectroscopy (OES), and self-excited electron resonance spectroscopy (SEERS). The analysis covers the whole area spanned by these gases (including some mixtures), discharge pressure and RF power over more than one order of magnitude, and their impact on important plasma parameters like plasma density, plasma potential, electron temperature, temperature of the plasma bulk, electron collision rate with neutrals, and actual RF power coupled into the discharge. From these, other properties (electrical conductivity, capacitance, plasma bulk resistance, sheath resistance and its thickness) can be derived. Since the methods are partially complementary, a mutual control of the obtained data is made possible, and the limits of the methods can be evaluated. Compared to discharges of inert gases, which are used as calibration standard, electron plasma density and electron temperature are both definitely lower, which is mainly caused by electron attachment of the electronegative molecules. Absolutely no chlorine ions could be found in the plasma which is referred rather to the effective cooling of the Cl-containing species than to the nonexistence of these species. Furthermore, we compared values for the electron temperature and the plasma density obtained with OES and SEERS, respectively, and with the Langmuir-probe system. The concordance in both properties is surprisingly good, despite the fact that the electron energy distribution should be described with two temperatures and only Langmuir is spatially resolved. The variation of the calculated dc conductivity either from impedance measurements or combined Langmuir/SEERS lies within a factor of 2 and is mainly referred to uncertainties of the current path rather than to principal faults of the various methods. |
|
PS-MoP-23 Fluorocarbon Decomposition Products and Effluent Analysis from Atmospheric Pressure Dielectric Barrier Discharges
S.F. Miralai, S. Mukhopadhyay (Wright State University); V. Shanov (University of Cincinnati); S. Datta (Procter & Gamble) Dielectric barrier discharges (DBDs) are gaining increased attention as an economical and reliable method for generating non-equilibrium plasma conditions in atmospheric pressure gases. This has led to a number of important applications including industrial ozone generation, surface modification of polymers, excitation of CO2 lasers and large area plasma flat panel displays. One of the challenges in the development of applications of DBD is an understanding of reactive species generated in the plasma and analysis of effluent gases. Some of the reaction products in the effluent gases are in very low concentration, increasing the complexity of the detection process. This paper deals with the detection of reactive species and products in the plasma and in the effluent gases from a fluorocarbon dielectric barrier discharge plasma . Dissociation efficiencies of fluorocarbons at various frequencies and using different carrier gases , such as helium, nitrogen, argon and oxygen will be presented. Analysis of the reactive species and products will be described, using Optical Emissions Spectroscopy (OES) and Quadrupole Mass Spectrometry. Correlation of experimental data with preliminary modeling results will be presented. |
|
PS-MoP-24 Endpoint Detector for Controlling Clean and Passivation in HDP-CVD Processes
R. Rulkens (Novellus Systems, Inc.) Optical Emission Spectroscopy was used to develop a sophisticated End Point Detector for the clean and passivation steps of HDP CVD processes. During the clean process steps, NF3 plasma removes build-up of SiO2 on the reactor walls. The end point detector in real time determines the optimal time at which the silica film is removed from the reactor walls and prevents over or under etching. After the clean steps, a Hydrogen plasma is used in the passivation step to remove residual fluorine. In a similar manner, the End Point Detector determines the optimal time needed for passivation. The diagnostics are incorporated into the equipment hard- and software and automatically control the correct timing of the clean and passivation steps. |