Future semiconductor manufacture includes challenging requirements to maintain film interface quality and minimize contamination. In addition, there will also be a drive to reduce costs by reducing the number of operational steps or by increasing throughput. Processes driven by yield considerations are metal stacks, poly-metal dielectric (cluster tools exist today) and, in the future perhaps, gate stacks, poly emitters, and salicide contacts. Processes driven to reduce operational expenses are those that are used many times, such as lithocells (track and exposure tools) and, perhaps, metrology cells. This paper reviews the status and process intent of typical cluster tools and their architectures. It addresses many of the issues that exist and provides a theme that much learning is required to achieve a substantial cluster tool environment in a future factory environment.

Today's advanced devices require increasing levels of process complexity. Semiconductor (SC) manufacturers are investigating higher levels of process integration using cluster tools as an approach to decrease process cycle times and increase yields.

Ti and TiN layers have several applications in VLSI devices as anti reflective coatings, step coverage enhancement, adhesion layers for blanket tungsten CVD, and with increasing importance, as diffusion barriers. This work describes the development of a simple and reproducible reactive sputtering process for TiN, using a BALZERS ARQ 150 DC planar magnetron with a multichamber process system. Resistivity and uniformity, deposition rate, stoichiometry and stress, and their dependence upon sputter source power, sputtering pressure, argon to nitrogen ratio, and substrate temperature were investigated. A process window is specified with resistivity values below 70|lohm-cm, film uniformity better than ±5% over a 200mm wafer, deposition rates up to 100 nm/minute, and residual stress below 5 x 109 dynes/cm2.

A review is presented of cluster tool concepts, their potential advantages for future IC manufacturing, approaches to cluster tools and cluster tool technologies. As wafer size increases and device feature size decreases, cluster tools should play a more central role in future IC manufacturing, although there are several problems to be overcome before cluster tools are available for a broad spectrum of IC technologies.

The next generation sub-micron devices requires control of wafer environment and the interfaces between film layers. A wafer must not be exposed to atmosphere between some processing steps. To meet these requirements, multiple processing integrated system or "Cluster" tool for IC fabrication has been a major tooling development in the last few years. This leads to the concept of open architecture system and the development of interface standards. This also creates new demands on the system control and user interfaces. An open architecture integrated system must operate in a seamless fashion with real-time control and communication among the process modules and the wafer transport platform. Furthermore, it must present the user with a single interface mechanism with all the flexibility and functions required from the process modules as well as the wafer transport platform. The user interface will need to be configurable to die unique requirements of each process modules. In essence, the user should have all the flexibility in selecting their process module of choice without any penalty in performance. This paper describes the control and communication architecture and methods used in an open architecture cluster host platform. The architecture provides the basis of an environment that is easily adaptable to the various requirements in this type of tool. Functions such as flexible wafer routing, process recipe selection and editing, interface with multiple modules are shown. Other useful functions such as data logging, system performance record and "intelligent process parameter control" are also described. Finally, the authors will discuss developments directions that will provide complete and quick integration between the process modules and cluster host platform while still retaining the unique identity of each supplier in the future.

Cluster tool demands on oxidation process for device fabrication leads to requirement of single wafer high pressure hot process tool. A joint development project between GaSonics/IPC and Texas Instruments has produced a working tool capable of delivering high quality oxides and excellent reflow performance to meet the future needs of the industry.

Cluster tool processing offers the semiconductor manufacturer solutions to some of the most significant problems of complex device manufacturing, primarily that of achieving high device yields at a competitive cost. Integrated processing, lower particulates, better wafer to wafer uniformity, precise control over each wafer, excellent small lot economics, and capability of simple expansion are benefits inherent with clustered architecture. Cluster tool architecture has proven to enable evolution of process and equipment to meet increasingly difficult device requirements. Process modules can be developed to address new applications, and subsequently installed on an existing platform, at a much lower cost than developing a totally new tool. Evolutionary enhancements can be simply made to the platform which benefit multiple process applications, and platform improvements required for new process applications can be easily transferred to existing process applications, often without any application specific re-design. Clustering promotes a synergy among available applications; for example, several features designed for a new Metal Etch/Photoresist Strip application were easily transferred to other processes that are available on the cluster, allowing multiple processes to benefit from development work on a single process. This inherent flexibility allows the equipment manufacturer to provide a tool capable of evolving to meet the user's changing needs; a tool with a longer life cycle, which can reduce the semiconductor manufacturers capital equipment expenditures.

As an extension to approaches in wafer processing, a new technique is being developed: Dynamic Design Processing. This technique is based on the recalculation of design specifications, whenever the results of any process step diverge specifications. These random divergences are inherent in any processing step. Recalculation of the design specification values restores the final performance of the device to the desired one. Simulation results show that applying this technique is practically equivalent to eliminating the process randomness.

In this paper an attempt is made to study the Moss-Burstein shift in quantum wells, quantum well wires, quantum dots, inversion layers, magnetic quantization, magneto-size quantization and magneto-inversion layers in infrared compounds,by formulating the appropriate electron statistics, it is found, taking Hg, Cd Te as an example, that the Moss—Burstein shift exhibits oscillation of various manners for the said quantum confinements of the band states with respect to electron statisties, film thickness and magnetic field respectively. The oscillations are totally band-structure dependent and the theoretical analysis is in agreement with the experimental results as given elsewhere. in addition, the well-known results for relatively wide bandyap materials have also been obtained as special cases of our generalized analysis.

In this paper, an attempt is made to investigate the photoemission from III-V, II-VI, pbTe/SnTe, strained layer and HgTe/CdTe superlattices with graded structures under magnetic field and compare the same with the bulk specimens of the constituent materials by formulating the respective magnetic-dispersion relations. It is found, that the photoemission exhibits oscillatory dependence with inverse quantizing magnetic field and the electron concentration respectively in all the cases. The oscillations in HgTe/CdTe superlattices show up much more significantly as compared to other systems even in the presence of broadening. In addition, the well-known expressions for the bulk specimens of non-degenerate parabolic materials have also been derived as special cases from our generalized expressions.

In a proposed process diagnostic system, a template correlation routine is used to identify the pattern produced on a wafer by bad die. The same pattern recognition routine can be applied to the pattern produced on the wafer from the result or different types of failure of the die. The overall wafer pattern, the individual test patterns, and electrical measurements made during final test form the input data to an expert system. In turn, the expert system uses heuristic algorithms that identify probable process problems, The template correlation routines have been successfully developed and implemented. Current efforts are dedicated to the knowledge engineering phase of the project. The expert system will, at first, identify defects to the process level. Future refinements will permit the diagnostic tool to identify the defect by name as well as process. Successful development and implementation of this system will save the labor of manual investigation of anomalous events in the fabrication of VLSI devices. Used in conjunction with statistical process control, this system should improve VLSI device yield.

The excitation processes in a DC glow discharge ignited using a saddle-field electrode configuration have been examined using optical emission spectroscopy (OES) for discharges established in SiH4, and in Ar/SiH4 and N2/SiH4 gas mixtures. The emission intensities of excited gas-phase species are correlated with plasma probe mass and energy spectral analysis of the resulting reactive radicals impinging onto the substrate holder. Discharges ignited in SiH4 exhibit strong Si optical emission lines relative to the SiH lines, reflecting extensive gas-phase decomposition of the starting SiH4. The corresponding mass spectra of positively charged radicals exhibit a dominant peak at 28 amu that is associated with Si+. The resulting deposition rate of a-Si:H scales linearly with the flow rate of SiH4. The addition of argon to the glow discharge in SiH4 assists the gas-phase dissociation of the SiH4 as indicated by higher partial pressures at 28 and 29 amu, corresponding to the enhanced formation of Si and SiH. Moreover, gas-phase interactions with excited argon result in greater excitation of the background H2, leading to a higher concentration of atomic hydrogen in the discharge. Ionized atomic hydrogen dominates the discharge current at higher Ar to SiH4 gas flow ratios. OES spectra of DC saddle-field discharges in N2/SiH4 gas mixtures indicate strong activation of N2+ and good dissociation of the SiH4 over a wide range in flow ratios, facilitating the preparation of Si:N:H films with stoichiometries ranging from N/Si = 0 to 1.8.

This paper will explore the methodologies of real-time measurement of photoresist film thickness on silicon wafers using multi-wavelength reflection interferometry. Reflected light from the wafer’s surface, containing the interference profile, is collected in-situ via a fiber optic cable and film thickness is determined in real-time via a pattern recognition algorithm. The instrumentation used to make this measurement and its application towards optimizing track performance during spin-coating and bake will be discussed. Data demonstrating basic thickness versus spin-time and thickness versus bake-time profiles acquired on-line without process disruption will be presented along with its utilization towards minimizing process set-up and machine qualification. Moreover, the advantages of characterizing film thickness on-line and in real-time will be reviewed.

Mass spectrometric and optical emission studies have been performed on argon discharges in a GEC rf reference reactor. Kinetic-energy distributions for ions produced in the sheath region are broad and exhibit structure, while ions produced in the bulk plasma exhibit narrow, featureless energy distributions. The addition of small amounts of O2 to an argon discharge significantly alters the observed positive-ion kinetic- energy distributions. Optical emission studies indicate increasing spatial non-uniformity in the plasma at higher pressures. Time-resolved optical emission studies indicate a varying relationship between the applied rf voltage and the time-varying optical emission with changing pressure and position between the electrodes.

Radio-frequency (rf) electrical sources are commonly used to generate plasmas for processing of industrial materials and for related experimental work. Published descriptions of such plasmas usually include generator-power measurements, and occasionally include plasma dc-bias measurements. One or both of these quantities are also used in industrial feedback control systems for setpoint regulation. Recent work at Sandia and elsewhere with an experimental rf discharge device (the "GEC RF Reference Cell") has shown that power and dc-bias levels are often insufficient information for specifying the state of the plasma. The plasma can have nonlinear electrical characteristics that cause harmonic generation, and the harmonic levels can depend sensitively on the impedance of the external circuitry at harmonic frequencies. Even though the harmonics may be low in amplitude, they can be directly related to large changes in plasma power and to changes in optical emission from the plasma. Consequently, in order for a worker to truly master the plasma-generation process, it is necessary to understand, measure, and control electrical characteristics of the plasma. In this paper we describe techniques that have been developed from work with the Reference Cell for making electrical measurements on rf plasmas, and we describe surprising observations of harmonic behavior.

Real-time plasma etch process monitoring, based on sensors which measure plasma properties that relate directly to desire etch features, is critical to the manufacturability of future-generation integrated microelectronics. This study reports on the development of spatially resolved optical emission spectroscopy tools and spectral signature analysis techniques which, taken together, show promise for meeting the need for such a sensor-based tool. This approach to plasma monitoring requires the development of optical emission calibration data sets, which in this case were obtained using the GEC Reference Cell plasma etch reactor (a RIE RF discharge system developed for interlaboratory comparisons). Systematic electrical probe measurements of the current and voltage waveform characteristics of Ar and CF4/CHF3 discharges in the Cell are reported and shown to be a sensitive indicator of both variations in Cell to Cell characteristics and process-induced variability within a Cell. Variations in the optical emission intensity show a near unity correlation with probe measurements if a broad spectrum of the emission is examined using multivariate statistical algorithms (chemometrics) and the spatial dependence of the emission is considered. Preliminary results involving further application of this spectral signature analysis technique to silicon dioxide etching are discussed.

This paper describes the application of thermal imaging by an infrared CCD TV camera with PtSi Schottky-barrier detectors for in-situ monitoring of plasma etching parameters. In-situ radiometric measurements were successfully employed for wafer temperature measurements and end-pint detection during plasma etching of polycrystalline silicon film on oxidized Si substrate for normal as well as oblique viewing. It was shown that thermal imaging can also be used for remote sensing of etch rate, heat of reaction and for measuring thermal time constants. The heat transfer coefficient of the thermal contact between the Si wafer and the water-cooled electrode can be determined from these measurements. The values of the heat transfer coefficients for the wafers just placed on the electrode without any provision of a good thermal contact were found to fall in the range of 8 to 12 J/m2-s-K and increased to about 200 J/m2-s-K when vacuum oil was used to improve the thermal contact.

Ultra Large Scale Integrated (ULSI) device requirements constantly challenge the technological frontiers of equipment capabilities. In order to have high uptime it is now required for machines to contain a certain level of intelligence. By utilizing on-board diagnostics, the equipment is able to inform operators of the change in the status of the equipment before a hard failure occurs. This is particularly important in high volume manufacturing where it is not acceptable to incur hard failures which can cause production losses. Drytek has developed software which incorporates in situ diagnostics and monitoring in order to meet the needs of today's leading edge IC manufacturers. These products allow for user definable parameters to be monitored in real-time, both on the tool and through a SECS I and II protocol port for host computer communications. This paper describes Drytek's approach to onboard diagnostics and process monitoring and presents the benefits and uses of such on-board capabilities.

An in-situ ellipsometer was used to provide real-time thickness measurements and control of process end point during plasma etching of silicon dioxide. In-situ ellipsometer thickness values were obtained at intervals of 0.15 seconds. The measured thickness was used to determine end point to obtain a desired film thickness. The oxide films etched using in-process ellipsometer control had a thickness accuracy and reproducibility of 3 A. In comparison, timed etching of silicon dioxide films in the same reactor, a thickness reproducibility of only 69 A was obtained.

A new methodology for process optimization is introduced which has the goal of dividing process variables into three categories. Robustness factors are those used to render the process robust against disturbances. Tuning factors are those used to tune the uniformity of a batch. Adjustment factors are those used to adjust the mean of the batch. An objective function called a 3-step SN ratio is derived to enable the 3-step optimization. In the first step, a designed experiment is performed and the results are used to find the best tuning and adjustment factors and to set the levels of the robustness factors. In the second step, the tuning factor (or factors) is used to tune those aspects of the batch uniformity which can be tuned. In the third step, the adjustment factor is used to adjust the batch mean to the target The methodology is illustrated and demonstrated in application to single wafer plasma etching.

Polysilicon etching in a single-wafer, parallel-plate, magnetically- enhanced RIE tool has been examined using two different approaches to the non-physical modeling of the system characteristics. The behavior of both process responses (polysilicon and oxide etch rates) and plasma parameters (voltage and current metrics) have been examined as a function of five variables (rf power, pressure, magnetic field, gas flow rate, and He backside cooling). The variable-response mapping was examined using both neural network and response surface approaches. The greater fitting power of the former method is demonstrated in a side-by-side, internally consistent comparison of the same data set using these two approaches.

Automated control schemes would greatly improve the reproducibility of plasma-assisted etching processes. In this paper we report on the application of an in situ metal film thickness sensor to control a plasma tungsten etch process. The process consists of an anisotropic step to control line profile and remove as much tungsten as possible, followed by an isotropic step which etches through to the underlying layer. In typical operations, a pilot wafer is measured off-line to determine the initial tungsten thickness. An etch time for the first step is then calculated before processing the entire lot. Single wafer lots require the elimination of a pilot wafer. Recently, we integrated a metal film thickness sensor (based on the technology of eddy currents) into a single-wafer plasma tungsten etch module. Our control strategy uses the sensor in a feedforward manner. A measurement of the tungsten film thickness is made in situ prior to processing. Process control software adjusts the etch time for the wafer based on the measured thickness and the predicted etch rate for the equipment settings. The etch rate is calculated from an empirical model obtained using response- surface methodology. A three-fold decrease in wafer-to-wafer variability in final thickness after the etch step was realized compared to that for the deposited thickness.

Ideally, the desired output of any process should be monitored directly during the process for closed loop feedback control. Unfortunately, for most processes in the semiconductor industry the output is difficult to monitor. Plasma etching is one such process with the additional complexity of being a nonlinear, multiple input, multiple output system. A phenomenological model can predict process behavior over a wide range of operating conditions and hence can be used for the design and control of plasma reactors. By contrast, empirical models are highly unreliable outside the range of operating conditions over which they are parameterized. A phenomenological process model for the etch rate of SiO2 in CF4 and CF4/O2 plasmas is presented. Direct process measurements including gas flow, chamber pressure, wafer temperature, RF power, optical spectroscopy and charge density are incorporated into the model to predict the etch rate. The model is used in conjunction with a previously developed, generalized, PC-based, controller to provide control of the process.

Plasma processes typically involve a number of coupled control parameters which exhibit complex interactions. These parameters jointly effect the plasma process parameters which may or may not be measurable. Ultimately, the process parameters affect the wafer parameters which we would ultimately like to control. The inherent multivariate nature of the problem makes conventional control methodologies difficult to apply. Artificial neural networks (ANNs) offer a promising alternative because of their ability to learn the desired control behavior by direct observation of the process. Once enabled as the controller, the ANN continues to improve its model of the process behavior and thus compensates for slow drifts in the process.

The accurate control of the photolithographic process is critical to the production of VLSI circuits. A control scheme is needed so that deviations from specifications may be compensated by adjustments to the process. Such a control scheme would take advantage of equipment models for the various steps involved in photolithography. Unfortunately, the equipment used for photolithography often changes with time, and is always subject to various disturbances which in turn introduce significant fluctuation in the process performance. In this report, we present an adaptive regression model which will evaluate itself and decide whether it should be refitted to the equipment to better reflect equipment behavior. The model is adaptively modified through recursive estimation based on in-line wafer measurements. Decisions for model changes are based on formal statistical tests which use the principles of the regression control chart [1]. This strategy is being tested on the spin-coat and bake equipment in the Berkeley Microfabrication Laboratory and will soon be extended to the entire lithographic sequence.

A closed-loop adaptive control technique for photolithography is proposed and evaluated. In this strategy, development time is manipulated in order to keep the output critical dimension at a desired value, despite the disruptive effects of unmeasured process disturbances. The adaptive control strategy incorporates a three parameter reduced-order model of the photolithography process, a parameter estimator, and a nonlinear model-inversion controller. Simulation studies show that the adaptive controller is able to reject the detrimental effects of process disturbances and bring the critical dimension back to its desired value. In the simulations, PROLITH was used to represent the real lithography process.

Raman microprobe scattering is shown to be a useful non-invasive real-time probe during thin film processing. Applications during laser-assisted processing that are relevant to microelectronics are emphasized.

Light scattering methods are being used in semiconductor device fabrication for such diverse purposes as particulate monitoring and linewidth control. In this overview, we focus on applications of light scattering for monitoring reactively ion etched trench profiles and plasma passivation of III-V materials.

Measurements of ultra-thin films (<100A) and small geometries (< 1/mm) of IC product wafers require more than simply a smaller measurement spot size. An optical artifact has been discovered when using spectrophotometers to measure ultra-thin films near feature edges. A model of this effect will be presented. This artifact is a subtle effect that produces measurable reflectivity errors tens of microns from a feature edge. While this error is small, it is not negligible for film thickness measurements below 400A. Experiments have been performed on typical spectrophotometers and data from these experiments will be presented. These data will be compared to a newly developed laser-based dielectric film thickness measurement system that significantly reduces this edge effect.

The semiconductor industry is moving towards thinner dielectric films (less than 100 angstroms), more complex film structures (oxide on polysilicon on oxide on silicon) and smaller lateral geometries (less than 1 micron). The attendant measurement requirements demand more than incremental improvement of existing methods. In this paper, a novel laser-based dielectric film measurement system is described that meets these requirements by bridging the gap between spectrophotometer speed and ellipsometer precision while measuring with a 0.9 micron focused laser spot. The operating principles of a new technique which we call Beam Profile Reflectometry are discussed and data are presented for a number of different single- and multi-layer film structures relevant to microelectronics processing.

The dynamics of pulsed excimer laser ablation of thin films of organic material from a metal surface was investigated using optically excited surface plasmons. The metal substrate is formed by a 500 A thick silver film adjacent to the base of a glass prism, which serves for optical excitation and detection of the surface plasmons propagating at the silver-vacuum interface. The organic films were prepared by cooling the sample to 77 K and depositing the materials (isopropanol, tetrafluoromethane, acetone) from the vapor phase; ablation of the films was accomplished by KrF excimer laser ( = 248nm). Detection of the surface plasmon resonance allows to monitor the ablation process on a nanosecond time scale and with a resolution far better than a monolayer. At the same time, the surface plasmons provide an in- situ probe for time-resolved measurements of the substrate temperature. This allows to determinate the temperature at which the ablation sets in. For tetrafluoromethane, the ablation temperature was found to be independent of the laser fluence, suggesting a thermal desorption process. On the other hand, for isopropanol and acetone, a strong dependence of the ablation temperature on the fluence was observed. From the large delay between the leading edge of the laser pulse and the onset of ablation, we conclude that ablative photodecomposition is not present in our experiment. There is, however, evidence for laser- induced chemical transformations in the organic films. Solid films of transformed material, which were stable at room-temperature and under atmospheric conditions, were formed during the isopropanol experiments. We suggest these transformation process to be connected to the observed fluence dependence of the ablation temperature.

We describe the use of optically excited surface plasmons to measure the thickness of ultrathin films deposited on gold and silver surfaces with submonolayer resolution. Additional structural information on the film is obtained by looking at the scattering of the surface plasmons. Applications of this method to physisorbed and quench-condensed molecular hydrogen films and to spreading of liquids are presented.

In-situ infrared absorption techniques have been used to examine fundamental chemical and physical phenomena in semiconductor process plasmas. Fourier Transform infrared absorption spectroscopy (FTIR) is used to characterize the chemical environment in halocarbon containing plasmas which produce particles. A correlation is observed between the distribution of chemical species in the plasma and the extent of particle formation as demonstrated by laser light scattering. The addition of oxygen affects both the chemical species distribution and the amount of light scattering in the plasma. Also, high resolution infrared laser absorption spectroscopy is used to characterize rotational and vibrational temperatures in a parallel-plate N20 discharge. The relevance of these studies to semiconductor process plasmas is also discussed.

For applications in ultra-large scale integration, low pressure, high density plasmas are being developed for etching and deposition of thin films. To control critical parameters such as the flux and energy distribution of ions impacting surfaces, it is necessary to understand how these parameters are influenced by physical and electromagnetic design. In this work, we report measurements of ion velocity distributions in Ar/He and Cl2/He electron cyclotron resonance plasmas. Using Doppler-shifted laser-induced fluorescence spectroscopy, we measure metastable Ar and Cl ion velocity distributions parallel and perpendicular to the magnetic field as a function of magnetic field amplitude, pressure, and microwave power. We also examine the effects of the wafer platen on the distribution functions by repeating the measurements after removing the platen. We find nearly isotropic ion velocity distributions when the source is operated as a magnetic mirror and the He partial pressure is low; higher He pressures tend to cool the parallel velocity distribution. Downstream, we consistently observe bimodal ion velocity distributions: the fast component, created in the source, follows magnetic flux lines into the reactor; the slow component, created mostly where the plasma expands from the source into the reaction chamber, is more isotropic. The relative amplitudes of these two components, the average ion energy, and the ion energy distribution are easily controlled by changing pressure and magnetic field.

Plasma-generated particulates are receiving increased attention as a source of contamination in plasma etching and deposition systems. These particles are suspended electrically in the plasma, and they are subject to electrical, thermophoretic, gravitational, and frictional drag forces from both plasma ions and neutrals. Consequently, they form complex spatial arrangements in the glow. We have investigated these particle arrangements by spatially mapping the particle distribution with a simple HeNe scattering system. We have also employed direct sampling as a way to determine particle size. The systems we have studied include copper and aluminum particles formed in an argon sputtering system, and carbon containing particles grown in a methane discharge. In addition we have modeled a particulate containing discharge in an effort to understand how the particulates act under the influence of electrical forces, ion drag, gravitational forces, and thermophoretic forces. The model is used as an aid in understanding the light scattering results.

This paper will discuss and show applications of optical emission spectroscopy techniques and methods to monitor plasma emissions during semiconductor processing. A brief discussion of the instrumentation that was used and the software to control the instrumentation will also be presented. Optical emission spectroscopy techniques that will be discussed include chemical species identification in plasma etching, process fingerprinting, contamination detection, endpoint analysis/control and sputter/deposition plasma monitoring.

The deposition of TiN on a silicon substrate in a Triode Ion Plating (TIP) deposition system was investigated. The effects of the deposition parameters such as the pressure, electron beam current and the gas flow rates were studied by looking at the optical emission lines collected through an Optical Emission Spectrometer (OES). Two kinds of spectra were recorded: Scan spectra between 200 and 600 nm for the identification of the species in the plasma and time based spectra for monitoring the titanium and nitrogen emission lines. The deposition rate was also monitored in situ by using a crystal thickness monitor. Secondary Ion Mass Spectrometry (SIMS) profiling was used to analyse the distribution of Ti and N in the deposited films. It was found that OES can be used for the in situ monitoring of the deposition process. Moreover, the titanium emission line at 364.2 nm can be used to correct the fluctuation in the titanium evaporation rate.

Pulsed laser deposition (PLD) has recently been shown to be an effective means of depositing thin films from refractory targets. In this study, Nd/YAG and excimer lasers have been used to deposit thin films from refractory, high Tc superconducting, dielectric, ferroelectric and magnetic targets. Optical emission spectroscopy of the laser induced plume has been used to determine the identity and energy (temperature) of the excited state species present in the laser induced plumes. Temporally and spatially resolved optical emission spectra were obtained using a gated intensified photodiode array detector coupled to a grating spectrometer. The individual emission spectra were analyzed to identify the atomic and ionic species present. The temporal and spatial evolution of individual emission lines were used to determine the velocity of the species in the plume. These results were combined with the results from in situ molecular beam mass spectrometric analysis of the plumes. In addition, studies of the stoichiometry and morphology, as well as the electrical properties, of these PLD thin films were carried out for correlation with the spectroscopic observations.

Two separate laser spectroscopic approaches have been used to investigate plasmas in the downstream region of a diverging field ECR etcher driven by a standard 1 kW ASTeX 2450 MHz source. Infrared absorption spectroscopy with a diode laser source and a multipass optical system has been used to monitor concentrations of CF2, CF3, and CF4 in pure CF4 or CHF3 or in mixtures of these with each other or with O2. These species concentrations, which in the cases of CF2 and CF4 are absolute, have been obtained as functions of both pressure and power. In the same tool details of the ion dynamics were studied using laser induced fluorescence (LIF) with a pulsed, etalon narrowed dye laser source, pumped by a nitrogen laser. Doppler limited line profiles of several rotational transitions of the (0-0) band of the B 2£+ — X 2E+ system of Nj were used to extract axial and transverse ion velocity distributions. Both pure nitrogen and nitrogen-helium mixtures were studied, and the transverse distributions were also obtained as a function of radial position. In the pressure region studied (0.5-4 mTorr) ion energies are generally much smaller than had been anticipated, a result we attribute to the major role played by collisions. We have also measured the translational temperature of neutral helium (by LIF) and the rotational temperature of Nj (by LIF and optical emission) for comparison purposes.

The tessellated probe is a large area single sided array of electrostatic probe elements arranged to simulate a wafer of 200mm diameter during processing within a plasma system. Various modes of operation are available including the standard negative DC bias for ion saturation measurements. Total ion flux and spatial distribution can be measured, giving two very useful parameters for process development. Measurements can also be made in the presence of simultaneously applied RF bias.

This paper will briefly describe the theory of interferometry, and then detail the method, apparatus and preliminary results of experiments with the Tempest. Infrared laser interferometry is a well known technique for measuring layer thickness based upon assessing the optical thickness, or path length, L.