Temperature Measurement and Control in Rapid Thermal Processing

ABSTRACT : This paper presents a brief overview on the various aspects related to Temperature Measurements and control in Rapid thermal Processing(RTP). A description of the main features of this system is also given for better understanding. Relating to the temperature measurements, techniques like use of dual pyrometers and use of thin-film thermocouples for calibration are discussed. The biggest problem in RTP industry is maintaining uniform temperature on the wafer surface, this subject has been addressed by a note on the effect of patterns and emissivity effects on temperature uniformity is stated. To give a feel of the latest developments in the field of RTP, the fourth and fifth sections discuss about Spike Annealing and RTCVD (Rapid Thermal Chemical Vapor Deposition). The paper ends with some detail on the Control methods adopted to maintain uniform temperature .

Keywords : Dual Pyrometer, Spike RTP, Thermocouples, Pattern Induced RTP, Cold Wall, Carrier Absorption, Temperature Uniformity.

[4]Rapid thermal processing (RTP) is a single wafer alternative approach to batch furnaces. The use of lamps for wafer heating allows rapid temperature cycling and short time wafer processing over a wide range of temperatures in controlled ambient. The conventional RTP technologies have gained limited acceptance due to process control and temperature non-uniformity problems. Temperature measurement and control problems due to the wafer emissivity effects and lack of dynamic control degrade the overall within-wafer process uniformity and wafer to wafer reliability. Exciting developments over the last few years, however, have improved the prospects for acceptance of leading edge RTP technologies in IC manufacturing with emissivity-compensated, sensor based, model-based, dynamic, multi variable wafer temperature and process uniformity control. The transition toward larger wafers and the proven performance advantages of RTP are raising the prospects for accelerated increase in RTP market size over the next several years, particularly starting with 0.25µ technology node.

A Structure for a RTP system may be defined in a variety of ways depending upon the number of lamps or number of Chambers or the sensing method adopted and so. The above figure shows the generic structure of the RTP system. The Wafer is placed in the middle of the chamber with a series of lamps above it and the temperature sensing equipment below. The Chamber contains gases and is tightly enclosed. More detail can be obtained from the Overall structure of the RTP. This structure has been developed at Stanford University.

As can be seen from the figure, a series of lamps are used to heat the wafer and there is arrangement is made in such a way that there is uniform heating of the wafer. Even taking care of distribution of lamps it is observed that the edges of the wafers are at a lower temperature compared to the central part of the wafer. This problem can be overcome by use of extra heat source at the edges. Heat transfer through the gas chamber is done through quartz windows.

The main parts in case of RTP equipment are the Heat Source, Reflector, Chamber and Temperature sensing equipment. The mainstay in RTP heaters is the tungsten (W)- halogen lamp. The choice for a heat source depends on various factors like initial heating rate, thermal mass, and lifetime and most importantly their symmetry. The figure shown has pseudo-ring symmetry. The designer has an option to choice of using high energy Arc

lamps instead of a bank of Halogen lamps. The difference between the two being the wavelength of radiation. The emission spectrum for an arc lamp ranges around 0.2µm to 1.4µm, whereas tungsten lamp radiation is between 0.3µm to 4.0µm. At shorter wavelengths , the excess energy is dissipated in the material in the form of heat. Therefore, almost all of the emitted radiation from the arc lamp is absorbed in the silicon wafer. Above the cutoff wavelength for band-to-band transitions, silicon can still absorb photons, but, through a different mechanism called free carrier absorption. In this regime, by absorbing the energy of photon, an already free carrier can make a transition to a higher level in the same valley. Clearly, the transition requires a momentum change which can be provided by phonons or ionized impurity scattering. Naturally, amount of free carrier absorption in the wafer is proportional to the number of free carriers which is determined by the doping level of the substrate and the process temperature. In silicon, free carrier absorption occurs at wavelengths above 2µm. Therefore considerable amount of free carrier absorption takes place when tungsten halogen lamps are used for heating the wafer.

Commercial RTP systems employ three generic chamber designs, namely cold wall, warm wall, and hot wall. The Cold Wall reactor is made from water-cooled metals such as stainless steel, aluminum, or other alloys in cylindrical or rectangular shapes. Reflectivity is accomplished by electro polishing the metal, or by coating it with reflective materials such as gold or aluminum. Often the metal surface is passivated by anodizing it, with a thin quartz layer.

The chamber top face is an air or water cooled quartz window plate that transmits the optical flux into the chamber. The cold walls have the advantage of minimum thermal memory effects that could further complicate temperature measurement. Water cooling of the quartz plate is more effective than air cooling. Especially with high-throughput processing the top plate can be heated up by the many wafers. Thus, parasitic deposition on hot spots may occur, which causes particles and decreases the uniformity in temperature.

But water cooling also has disadvantages. These are possible formation of bubbles, contamination in the water, and, most importantly, the less efficient coupling of the heat source to the wafer, due to the optical absorption above 1.4µm by the water column. This problem can be alleviated by using coolants that do not absorb in the near infrared and have enough heat capacity, for example, Fomblin oil.

Other window plate materials are also used, like sapphire. It has about 20 times more thermal conductivity and will create less thermal memory than quartz. On the other hand, depending on the process, too cold a wall temperature may also give problems in terms of unwanted deposition or condensation of reactants. A tunable wall temperature, say from room temperature to 100C, may be an effective compromise here, and so are the use of quartz liners along the metal walls and the use of slower ramp rates up to deposition temperature. The details regarding temperature measurement, calibration, and uniformity issues are given in the next session.

In order to meet the NTRS requirements of temperature uniformity and uncertainty, the current measurement methods should be highly reliable and accurate. Moreover they should be capable of quick multiple readings with a resolution of + 0.25 C at around 1000 C. The solution for the needs is Optical Pyrometry. Pyrometry is the favored method because it is a non- contacting and also the errors generated in the readings are very less. [3]These can be easily integrated into chamber designs. Commercial pyrometers are capable of reading upto a resolution of + 0.1 C with a frequency greater than 20Hz and lifetime of 1yr.

However the repeatability of temperature measurement depends upon a variety of issues, the chief being,

Minimizing the interference from the background radiation.
Minimizing the impact wafer emissivity has on temperature measurement.
Finding an improved way of performing in situ calibration of the pyrometers.

Background radiation from the lamps or any other source in the RTP equipment can generate optical signals thereby leading to erroneous readings in the optical pyrometer. This is of most importance especially at low temperatures. As a remedy to this problem, the chamber is designed in such way that the only single sided heating occurs. Temperature is measured by a pyrometer that detects wafer radiance and subtracts background interference from the furnace radiation. This is done by integrating the lamps on one side and the measuring equipment on the other side of he wafer being manufactured. Moreover care should be taken that no other radiation reaches the pyrometer, the incident radiation should not reach the pyrometer and hence the area under the wafer is made light tight. This can be done by applying a coat of dense material inside the chamber surface. Wafer rotation is used for azimuthal uniformity. An alternative to this approach is called Dual Pyrometer method, wherein we use two pyrometers instead of one. Experimental set up for this mode of operation is shown in Fig. 3.

Fig. 3Dual Pyrometer temperature measurement with in situ, real-time emissivity measurement and compensation by simultaneous transmission and reflection measurement. [2],[6]

[6]The two radiation probes shown in the illustration are used to determine the wafer temperature called the ripple pyrometry. The probe with a view of the wafer collects a mixture of radiation emitted by the wafer and the lamps. The other probe samples radiation from the lamps for reference. The power to the lamps is modulated at a frequency in the range 20 to 200 Hz to create flicker in the lamp output that in turn produces an ac ripple signature in the wafer and lamp detectors. The ratio of the ac ripple signals is related to wafer reflectivity. The ripple pyrometer determines wafer temperature by subtracting the interfering lamp reflections in the wafer sensor and computing the effective wafer emissivity. [7]Although we have seen that ripple pyrometry and other methods have reduced the uncertainties in temperature measurement due to wafer emissivities and stray radiation, but as of present, the reliability of these methods do not provide traceability to the International Temperature scale of 1990 (ITS-90). Another alternative temperature measurement is Light-Pipe Radiation Thermometers.

Although temperature measurement using Pyrometers gives precise and fast readings, this process suffers from the drawback that pyrometer reading is dependant upon the wafer emissivity. This wafer emissivity is an unknown parameter. [2] The emissivity of a material is a function of the optical properties of

the starting wafer material (intrinsic emissivity),
the layers on top of or buried in the wafer ( extrinsic emissivity), and
the specific optical properties of the reflective chamber with all components inside (effective emissivity)

Both intrinsic and extrinsic emissivity values should be measured in a cold, black environment, excluding the extra effects of the reflective chamber and stray light and other hot chamber parts, which together comprise the effective emissivity.

[9]For most purposes pyrometers are chosen to operate at one particular wavelength; and if the emitted radiation from a sample is of insufficient magnitude, the pyrometer may not respond to it. The spatial resolution of the pyrometer is limited by its surface area. Pyrometers measure the amount of radiation emitted from a wafer within a narrow wavelength window. The ratio of the wafer emitted radiation to that of a blackbody under the same conditions of temperature, wavelength, angle of incidence and direction of both temperature and wavelength is called emissivity. It is also a function of the surface roughness. [10]For many applications such as RTA, the surface emissivity remains a constant throughout the process. However, for the applications where the characteristics of the surface being changed, such as RTCVD (will be discussed in the next section), the emissivity is being varying as a function of the type of film being deposited, the thickness of the film, and the decomposition of the surface it is being deposited on.

In traditional RTP tools, the energy source is a bank of tungsten halogen lamps that reach a temperature of 2000°C or higher while the process chamber walls remain relatively cool at <400°C. Process temperatures on the wafer are typically between 500° and 1100°C, and heat transfer is dominated by thermal radiation from the lamps to the wafer and from the wafer to the chamber walls.

The large temperature differences enable the fast wafer heating and cooling and the remarkable flexibility of process control that are characteristic of RTP. However, because the optical properties of the wafer affect both its absorption of energy from the lamps and emission of energy to the chamber, coatings or patterns on the wafer surfaces lead to across-wafer temperature variations. For example, even a very simple coating, consisting of a layer of polysilicon on top of a thin silicon dioxide film, can cause the lamp power coupling to vary from 0.4 to 0.8, depending on the thickness of the two constituent films. There is a similar impact on the integrated thermal emissivity, which determines how much the wafer radiates to the chamber walls.

If a coating is present only on parts of the wafer, then the uncoated regions will experience different temperature-time cycles during processing than the coated regions, which can have a significant impact on process uniformity. Thus, it is essential for RTP systems to include closed-loop wafer-temperature control capability. This can be achieved with an integral pyrometer that measures the wafer temperature and provides feedback to the control system. The challenge is to perform highly accurate pyrometric measurements on a wafer of unknown spectral emissivity in surroundings where a large amount of stray radiation may be emitted by the heat source. Suppliers of RTP tools have addressed that challenge using various innovative techniques, which, when combined with sophisticated power control, have led to excellent temperature results on monitor wafers..

In order to cancel out the effect emissivity has on pyrometer reading, a closed loop control system is adopted as shown in Fig. 4. T p is the temperature inferred from the pyrometer, this will be equal to the actual wafer temperature if the assumed emissivity is equal to the wafer emissivity. For more detail the reader is referred to [10].

Finding an improved way of performing in situ calibration of the pyrometers : [2]

The optical pyrometers described above give only a relative signal and need calibration to measure absolute temperature. The calibration of a pyrometer in a RTP system is done at two distinct levels. First is a system level calibrating, obtained by matching the pyrometer reading to that of a thermocouple in contact with sacrificial reference wafer. For frequent system recalibration one prefers indirect methods such as temperature sensitive reactions. If the emissivity is likely to show small variations for the individual wafer within a batch, a wafer level calibration is needed. Here, one measures the extrinsic emissivity directly by an ex situ reflectivity measurement prior to processing, or one measures the effective emissivity in situ before or during processing.

System level Calibration : The classical means of system-level calibration uses a cantilevered thermocouple, usually chromel-alumel. The method has a few disadvantages such as irreproducible contact, degradation upon repeated high temperature cycling, heat loss along the leads, and noninert contact causing chemical reaction within the wafer. Some of these problems can be solved more or less successfully by physically embedding the thermocouple in the wafer. The cantilevered or embedded thermocouple is normally used only for one-time calibration.

Wafer level Calibration : Calibration at the wafer level implies some method of emissivity correction. This can occur at the wafer batch level, or at the individual wafer level. In this section let us discuss in situ calibration method of LPRT’s against wire/thin-film thermocouple combinations with traceability to the ITS-90. Use of wire thermocouples alone has resulted in large uncertainties as these conduct and thereby reduce the temperature in the vicinity regions of the wafer. This uncertainty is observed to be ranging from 1 to 20C. Thin-film thermocouples have an upper hand in this regard. They give the temperature at the location of interest with reasonable accuracy while making very low thermal perturbation because they have very low thermal conductivity. They measure the temperature difference between the location of interest and another location on the periphery of the wafer, where high accuracy platinum-palladium wire thermocouples are used. Here the problem of wafer temperature reduction in the vicinity doesn’t cause any change to the wafer profile.

Experimental Method : [8] This type of calibration relies on obtaining a thermoelectric output for a Thin-film Thermocouple(TFTC) and compares it with that of a pure platinum wire. Fig. 4 below, illustrates the calibration apparatus for a particular set up.

The reference junction is nothing but the platinum-palladium thermocouple. This reference junction has been previously calibrated at NIST. For more detail on this experimental technique the user is referred to [8].

In ex situ measurement compensation of emmissivity is achieved by measuring the emissivity. Previously discussed Dual Pyrometry method is one of these techniques.

Uniformity is one of the major concerns in RTP. In the chamber, due to the radiant heat loss from the edge of the wafer, the temperature is usually a maximum at the wafer center and rapidly decreases with radius. A common mistake among the equipment manufacturers is to assume that good process uniformity will be obtained if the temperature uniformity is a few percent. It is observed that lower activation energies result in more uniform film deposition. Typical RTP reactor geometries utilize lamps that usually extend one or more wafer diameters beyond the wafer edge. Often reflectors are used to produce lamp images which extend further out from the wafer edge. These measures are taken largely to assure uniform heating of the wafers. One method of determining the heat flux distribution on the wafer is to use geometric optics to calculate the radiation from each lamp as a function of position on the wafer surface. Using this method has shown a 5-10% decrease from wafer centre to edge [13]. Therefore , RTP lamp configurations that often appear to provide uniform do not accomplish this.

Different techniques have been employed to compensate for the heat losses at the wafer edge. Some of the common are: [13]

Alter the center of the furnace windows so that the light entering here is scattered to some extent. Therefore, less light will reach the centre of the wafer than does the edge.
Supply increased power to the outer lamps to increase the radiation at the edges of the wafer.
Use a susceptor of some kind, which effectively makes the wafer appear larger, from a heat transfer perspective, than it actually is.
Employ a mirror system to focus or reflect more light on the wafer edges than reaches the centre.

All of these methods have been used with some success by vendors, and all have experienced limitations. The first and last methods have been found to be successful must be tailored to a specific process. Consequently, temperature uniformity can be achieved for only one set of process conditions. Moreover, temperature uniformity is not maintained during heat-up or cool-down.

RTCVD is relatively a new development in rapid thermal processing. One of the principal advantages of RTCVD is that sharp transitions are obtained between layers of differing composition or doping, while exposing the substrate to a much lower thermal budget than a furnace low-pressure chemical vapor deposition (LPCVD) process. Moreover RTCVD is a cold wall system which generates fewer particles. It also reduces problems of auto doping and interdiffusion through a lower thermal budget than LPCVD.A typical RTCVD system is as shown below.

An important area for the application of RTCVD is the growth of Ge xSi 1-x heterostructures on silicon. RTCVD demonstrates the ability to form alternating layers of semiconductor materials with sharp transitions in the stoichemistry. This is of para-mount importance for strained-layer Si/ Ge xSi 1-x/Si heterostructures providing the ability to tailor the band gap due to a reduction in the indirect band gap relative to the unstrained layers. Such materials offer the promise for new devices such as FETs., avalanche photodiodes, Infrared diodes. If the thickness of the strained layer is too large, relaxation takes place, and for that reason it is necessary to precisely define the layer thickness and stoichemistry.

Rapid Thermal Annealing (RTA) is of critical importance in electrical activation and damage annealing. During the activation of implanted dopants, Transient Enhanced Diffusion (TED) occurs, which is highly undesirable when shallow junctions are expected. TED is observed to occur due to excess Si interstitial distributions. Equilibrium dopant diffusion is recovered as the excess Si interstitial population is dissipated The goal in junction formation is to achieve a low sheet resistance with a high concentration of electrically active dopants, with as little dopant diffusion as necessary.

[12]Spike-annealing techniques were introduced to control dopant diffusion in shallow junction formation and it is integrated with contact junction and gate electrode activation Current spike annealing methods use infrared heating that is characterized by near thermal equilibrium across the thickness of the wafer. The variables are heating rates, the switching time from heating to cooling, and the cooling rate.

In spike annealing, we try to limit the undesired TED while at the same time reaching the peak temperature required for dopant activation. This is achieved by heating to the desired temperature in minimum possible time and cools down immediately as in without any dwell time at the peak temperature. An Idealized Spike process is illustrated here below.

For a spike application the wafer is heated rapidly to the peak temperature and then cooled rapidly. The transition from heating to cooling is limited by the thermal time constant of the wafer, typically 15ms, and the time constant for the heat source. The wafer thermal response time thus determines the turn around time. It has been assumed that the heat source can be turned completely off when the peak temperature is reached without causing transient thermal gradients across the wafer. This requires an inherently uniform process chamber that maintains a uniform loss rate without the need to use feedback control on lamps. If uniformity is maintained by varying the irradiance distribution across the wafer then that irradiance will increase the turn around time. The fast time response of the arc lamp is also essential to maintaining accurate closed loop control of the peak temperature. A lamp and temperature diagnostic response time of less than 1 ms can easily be maintained which is much faster than the wafer response time of 15 ms.

Fig. 8.Illustration of temperature vs. time profile for an assisted flash anneals. Wafer is initially heated to an assist temperature. Application of flash lamp radiation produces sharp surface-temperature peak. Wafer cools as both sources of heat are removed. [6]

The above figure clearly explains what we mean by Flash Annealing. These methods use a high optical flux for faster heating to either briefly melt the wafer surface or selectively raise the surface temperature. Upon termination of the light pulse, the surface temperature rapidly cools down by thermal diffusion into the bulk of Si. However this rapid cooling may cause problems like lattice defects and dopant metastability. Consequently, these alternative techniques are still in the research and development stage.

A Model Based Control method is presented for accurate control of RTP systems. The model uses 4 states: lamp filament temperature, quartz temperature and TC temperature. A set of four first order, non-linear differential equations describe the model. Feedback is achieved by updating the model, based on a comparison between actual (measured) system and modeled system response.

The Model Based Control uses two steps to reach a tight match between the setpoint temperature and wafer temperature Fig. 9. In the first step, the 4 state model of the system is invertedto predict that level of lamp voltage that will drive the wafer temperature to reach the setpoint temperature. The inverted model transforms Si temperature into predicted lamp voltage. The calculated lamp voltage is limited to a minimum value (0 V) and a maximum value (210 V) to reach the actual lamp voltage sent to the lamps. Due to this limitation on lamp voltage, a perfect match is not always possible. This causes a boundary on the first and second order derivatives of temperature. In the second step, the actual lamp voltage (after clipping) is used as input of the forwardmodel to calculate a prediction of the indicated TC temperature. Any offset between the predicted TC temperature and the measured TC temperature is used to adjust the model. The forward model is used to calculate the TC temperature (not the Si temperature). Sampling rate was 40 ms.

The advantage of the first step (inverted model) is that even without feedback, a reasonable match between setpoint and actual wafer temperature is achieved, so the feedback control only needs to correct small residual errors. The advantage of the second step (forward model and model adjustment) is that the model is kept up to date with the actual system at any time. Even when the actuators are saturated (lamp voltage maximum or zero), the match between model and system is maintained. As soon as the actuators are out of saturation, the inverted model (predictor) is ready to provide the exact signals for tracking the wafer to the setpoint temperature. In general, any parameter of the model (in the model data base) could be tuned in real time while the control is active. An optimized method may be to tune the chamber efficiency h (mainly dependent on wafer absorptivity) during temperature ramps and to tune the wafer emissivities e s , e q during steady state. The simplified solution used here (Fig. 10) was to not adjust the internal parameters of the model, but add a temperature offset at the entrance of the inverted model and the opposite offset at the exit of the forward model. A very interesting aspect of this feedback loop is that the “plant gain” as seen by the feedback controller always equals 1, over the full frequency spectrum. When the limiter is not active (no actuator saturation), the loop goes through the inverted model, the limiter and the RTP system. When the model is accurate, the combination of the inverted model and the actual system is close to a unity gain system. The branch through the forward model has zero net effect due to the subtraction of the offset value. When the limiter is active (actuator saturation), the path through the inverted model and RTP system is not active, but there is a straight path through the subtraction after the forward model, that also yields a unity “plant gain”. In all conditions, the system remains well behaved and allows an easy general design of the feedback controller with a high gain and still without oscillations. In these experiments, the feedback controller was a PI controller with a gain of 22 and an integration time constant of 1 second. The high feedback gain and short integrator time constant causes a fast and complete annihilation of any residual errors.

RTP has the advantages over conventional furnace technology of operating in a controlled microenvironment with low thermal mass and rapid heating rates. As high temperatures are attainable in a short time, single wafer thermal processing can occur in multiple-use chambers or in clusters of single-use chambers, thermal budgets can be achieved for defect free junctions with high dopant activation, and sharp interfaces can be created in layered structures. The main obstacles for unanimity of RTP in manufacturing are still temperature reproducibility and uniformity during all processing, i.e. in the dynamic and stationary parts to the thermal cycle, in which layers are being formed. One of the challenges of this decade is to overcome these crucial obstacles such that the semiconductor industry can successfully start the anticipated processing of DRAMs of Gigabits range on 250nm diameter wafers with 0.18µm design rules. Being a single wafer process it provides the best processing route. Additionally, cycle time and yield considerations make RTP a technology of choice.

[1] “ Rapid Thermal Processing of Semiconductors”, Victor E.Borisenko and Peter J. Hesketh, Plenum Publishing Corporation.

[2] “ Rapid Thermal Processing- Science and Technology”, Richard B. Fair, Academic Press.

[3] “ Advances in RTP Temperature Measurement and Control”, Bruce Peuse, Gary Miner, Mark Yam, and Curtis Elia.

[5] “A Learning Approach of Wafer Temperature Control in a Rapid Thermal Processing System”, Jin Young Choi and Hyun Min Do, IEEE Transactions on Semiconductor Manufacturing, Vol. 14, No. 1, Feb. 2001.

[6] “ Recent developments in Rapid Thermal Processing”, A.T. Fiory, Department of Physics, NJIT.

[7] “ITS-90 Traceable Calibration of Radiometers using Wire/Thin – Film Thermocouples in the NIST RTP tool : Experimental Procedures and Results”, C.W. Meyer, D.W. Allen, D.P. DeWitt, K.G. Kreider, F.L. Lovas, and B.K. Tsai.

[8] “Calibration of thin film Thermocouples on Silicon Wafers”, K.G. Kreider, D.C. Ripple, and D.P. DeWitt, NIST.

[9] “ Issues in Emissivity of Silicon”, S. Abedrabbo, J.C. Hensel, F.M. Sopori, A.T. Fiory, and N.M. Ravindra.

[10] “ Model based Emissivity Correction in Pyrometer Temperature Control of Rapid Thermal Processing Systems”, F. Yates Sorrell and Ronald S. Gyuresik, IEEE Transactions on Semiconductor Manufacturing, Vol. 6, No. 3, Aug. 1993.

[12] “Spike Thermal Processing Using Arc Lamps”, D.M. Camm and M. lefrancois, Vortek Industries Ltd.

[13] “ Manufacturability Issues in Rapid Thermal Chemical Vapor Deposition”, Mehmet C. Ozturk, F.Yates Sorrell, Jimmie J. Wortamn, F.Scott Johnson, and Douglas T. Grider, IEEE Transactions on Semiconductor Manufacturing, Vol. 4, No. 2, May. 1991.

[14] “ A Methodology for Analysis of RTCVD Systems”, F. Yates Sorrell, Mark J. Fordham, and Seungil Yu, IEEE Transactions on Semiconductor Manufacturing, Vol. 8 No. 3, Aug. 1995.

[15] “ Model-Based Control of Fast-Ramp RTP Systems’, D. De Roover, A. Enami-Naeini, J.L. Ebert, S. Ghosal, and G.W. Van Der Linden, SC Solutions Inc.

[16] “ Manufacturing Control of Ultra-Shallow Junction Formation By Rapid Thermal Annealing”, Sing-Pin Tay, Yaozi Hu, Vidula Kirtikar, Peter Borden, Gale Lane, and Laurie Bechtler, 7 th International Conferencce on Advanced Thermal Processing of Semiconductors (RTP’99).

[17] “ Model Based Temperature Control In RTP Yielding +0.1 C AccuracyOn A 1000 C, 2 Second, !00 C/sec. Spike Anneal”, Peter Vandenabeele, Wayne Renken, SensArray Corp.