JP2002523908A - Low-temperature process for forming an epitaxial layer on a semiconductor substrate - Google Patents

Abstract

(57) [Summary] A low-temperature processing method in which an epitaxial layer is formed on the surface of a base material without requiring an ultra-vacuum or ultra-cleanliness in a processing chamber when forming the epitaxial layer. In this method, an epitaxial layer can be simultaneously formed on a plurality of base materials (28). The base material is placed in a processing chamber (12) having a plurality of controlled heaters (212) and controllers (308, 306). The processing gas (250) is supplied through the pipe and the panel (248) to control the flow rate.

Description

DETAILED DESCRIPTION OF THE INVENTION

BACKGROUND OF THE INVENTION An epitaxial layer is composed of a single crystal semiconductor layer grown on a semiconductor substrate. Such layers are well known and have been used for many years, for example in the manufacture of bipolar junction devices, and more recently in the manufacture of CMOS devices.
The epitaxial layer used for the silicon substrate is a silicon epitaxial layer (aka
, Homoepitaxial layers). For advanced applications, homoepitaxial layers of SiGe or SiGeC are used.

[0002] Semiconductor substrates with an epitaxial layer, when compared to their polished wafer counterparts,
There is a feature that the doping of impurities can be performed uniformly and the crystal quality can be accurately controlled. These features are particularly advantageous when a large number of semiconductor wafers are used for manufacturing a desired integrated circuit device.

[0003] In order to understand these features, it has been found that the use of highly sophisticated equipment and techniques is required for growing large amounts of crystals and preparing large numbers of wafers with low defect densities. Must be. Even with such sophisticated equipment and techniques, the yield of crystals grown to be suitable for good wafer preparation is very low. For this reason, the dimensions of the wafer must be significantly increased.
However, advanced CMOS circuits require only about 3 microns of crystalline material when used in the manufacture of active devices, and therefore, rather than polishing an epitaxially affected wafer, the epitaxial layer Is desirably formed to that depth. Epitaxial layers can also be used to form nearly defect-free single crystal devices on large numbers of wafer surfaces that can have unacceptable defect densities.

[0004] Despite the advantages described above, epitaxial growth is the most complex and costly process in advanced processing in the semiconductor field. The most commonly used epitaxial growth method for silicon-based materials is chemical vapor deposition (CVD). In general, the efficiency of a CVD process depends on both the thermodynamic mass transport of the reactants and the surface kinetics of the chemical reaction. The higher the temperature, the faster the reaction rate is, and the effect of the deposition of the epitaxial layer depends mainly on the material transport at that temperature. As a result, the deposition process is, to a lesser extent, not dependent on temperature fluctuations, but is very dependent on the uniformity of the mass transport action (gas flow) at high temperatures. On the other hand, when the temperature is low, the reaction rate decreases exponentially as the temperature decreases. Therefore, the deposition effect of the epitaxial layer depends on the reaction at a low temperature. As a result, the deposition process is no longer dependent on the mass transport effect, but rather on the temperature uniformity. Because controlling the mass transfer of reactants in a processing furnace is often easier than controlling the processing temperature, the semiconductor industry has adopted a high-temperature epitaxial growth method that uses temperatures higher than about 1,100 ° C. It is usually adopted.

In order to carry out the high temperature epitaxial growth method as described above using a wafer having a large diameter, a single-wafer radiant heating wafer reactor (RTCVD) is generally used. Such a single-wafer reactor is more often used than a batch reactor because the batch reactor cannot make the gas transfer action uniform under such a high temperature environment. While single wafer reactors have been the solution for forming epitaxial layers on large diameter wafers, the throughput of such reactors is relatively low (the number of processed wafers is 25 Or less). Similarly, processing costs are high and the price difference between a wafer with an epitaxial layer and a polished wafer is very large.
In addition, the high processing temperature may cause, for example, the formation of slip lines, the outflow of dopants from the substrate, and the concomitant incorporation of dopants into the growing epitaxial layer (i.e., autodoping), and It can have a number of undesirable effects.

The present invention eliminates the need for utilizing high temperatures in the epitaxial growth method. Therefore, a low-temperature epitaxial growth method particularly suitable for forming a homogeneous or heterogeneous epitaxial layer on a large semiconductor wafer is provided.

(Summary of the Invention) A method for forming an epitaxial layer on the surface of a base material is disclosed. Briefly described, the method comprises the steps of loading a processing chamber with one or more substrates, evacuating gas from the processing chamber, and removing at least a portion of the evacuated gas from a dry reactive gas or Replacing with a purge gas, adjusting the pressure in the processing chamber to sub-atmospheric pressure, and adjusting the temperature to a value not exceeding about 1,000 ° C., and removing the oxides originally contained in the base material, The pressure inside the processing chamber is about 0.133 ~
Adjusting the temperature within a range of 133 hPa (about 0.1 to 100 Torr) and a value not exceeding 850 ° C., and introducing a precursor gas into a processing chamber to form an epitaxial layer. .

(Description of Embodiment) Outline of Typical Epitaxial Growth Reaction Vessel FIG. 1 shows an embodiment of a vertical heating reaction vessel system 10 for forming an epitaxial layer on a semiconductor wafer or substrate by a low-temperature epitaxial growth method of the present invention. Show. This heated reactor system was patented on April 19, 1988,
U.S. Patent No. entitled "Vertical Thermal Processor (Vertical Heat Processor)"
No. 4,738,618. The disclosure of this US patent is incorporated herein by reference. The heating reaction vessel system 10 includes a bell-shaped quartz processing tube 12 that can be vertically positioned by pneumatic pressure and a heating furnace 14 that can be vertically positioned by pneumatic pressure. The quartz processing tube 12 and the heating furnace 14 are heated. In order to improve the loading operation, the processing operation, the cooling operation, the cleaning operation, and the like in the reaction vessel system 10, various positions are taken concentrically with each other in the vertical direction.
Further, since the quartz processing tube 12 and the heating furnace 14 are positioned within the frame structure 16 and supported by an overhead supporting type double rail slide mechanism 18 similar to a draw drawer, they are mutually connected as described later. In addition to being independently lifted and lowered, the overhead supporting slide mechanism 18 conveys the sheet to the outside of the cabinet enclosure for periodic inspection and cleaning.

A plurality of working cylinders 22 a-2 are provided near the outer periphery of the container shield 24 of the heating furnace 14.
2n are arranged, and when the reaction vessel 10 is operated, the heating cylinder 14, the vessel shield 24, the built-in ceramic heating element 26 (FIG. 2), and the bell-shaped heating furnace liner tube 28 are respectively moved to the desired positions by these operation cylinders. Can be set to. Heating furnace 1
In the vicinity of the outer periphery of the container shield 24 of FIG. 4, other pneumatic working cylinders 30a to 30n are also arranged, and these pneumatic working cylinders include an inner processing pipe elevating ring 33 including the quartz processing pipe 12, The elevating ring 32 is positioned vertically. The water-cooled bottom plate 34 and the firing table 36 are positioned at the center inside the processing cavity surrounding body 38. The electronic control panel 40 is attached to the upper front panel 44. A vacuum sealing loading door 42 is provided on a lower front plate 45 of the cabinet.
Pressurized gas is supplied via the supply pipe 46 when the pneumatic cylinder is operated and the processing apparatus is purged. On the other hand, the processing gas flowing in the processing pipe 12 in the axial direction is supplied through supply pipes 48 and 50. Cylinders 22a to 22n, 30a to 30
A pneumatic control panel 52 for controlling the pneumatic pressure for operating n is provided near the container shield 24 in the upper rear part of the cabinet. A computer card rack 54 and an auxiliary control panel 56 are attached to the lower rear access door 58 of the cabinet. Numerals 60a to 60n denote a plurality of adjustable support columns, which are interposed between the V-shaped frame 62, the gas panel shield 64, and the bottom plate 66 of the cabinet. 64 and associated components.

FIG. 2 shows a side sectional view of the heating furnace in a state where the heating furnace is covered with a quartz processing tube, and the reference numerals used here correspond to the reference numerals used in the above description. The bell-shaped quartz processing chamber 12 is mounted on a circular quartz ring 68. The quartz ring 68 is mounted on the inner processing tube elevating ring 33, and the inner processing tube elevating ring 33 is formed of the outer processing tube elevating ring 32 concentric with the quartz ring 68 and the quartz processing tube 12. It is positioned inside. With the quartz ring holding ring 37 facing the inner processing tube elevating ring 33, the processing tube quartz ring 68 and the processing tube 12 are attached to the inner processing tube elevating ring 33. Since the outer processing tube elevating ring 32 is attached to the inner processing tube elevating ring 33 with a plurality of detachable pins 70a to 70n, the quartz processing tube 12 is taken out of the vertical heating processor housing as described later. It can be cleaned. These outer and inner processing tube lifting rings 3
2 and 33 are O-shaped ring seals 72 provided concentrically on the water-cooled bottom plate 34.
, 74. Inner quartz processing tube elevating ring 33 and processing Kanji English ring 68
Another seal 76 is provided between them to ensure airtightness in the room. The processing tube elevating cylinders 30a to 30n are supported by the container shield top plate 24a and the lower container flange ring 35 and are in contact with the outer elevating ring 32, as well as the outer elevating ring 32, the inner elevating ring 33, and the processing tube quartz. The ring 68 and the quartz processing chamber 12 can be positioned vertically with respect to the water-cooled bottom plate 34.

The heating furnace 14 that can be positioned in the vertical direction is not necessarily limited to this, but a ceramic heater mounted and mounted on a bell-shaped element tube or a liner tube 28 for the heating furnace and a quartz ring 78 of the element tube. 26, a heater element raising / lowering ring 80, a container flange ring 35, and a heater element positioning ring 82, all of which are concentric with the quartz processing tube 12. The element tube quartz ring 78 is positioned within the element elevating ring 80 with an O-shaped ring seal 84 interposed therebetween. The container flange ring 35 is on the heater element elevating ring 80 and
4 is easily attached to the heater element elevating ring 80. The O-shaped ring seal 84 not only seals the flange ring 35 but also seals the element tube quartz ring 78 and the element elevating ring 80. An O-shaped ring seal 80 is also interposed between the heater element elevating ring 80 and the quartz ring holding ring 37. The heater element positioning ring 82 is on the flange ring 35, and aligns and incorporates the ceramic heater element 26 along the vertical axis of the heating furnace 14. The stationary ends of the vessel-heating furnace working cylinders 22a-22n are attached to an upper vessel shield 24a. The other end is attached to the flange ring 35 while the operating rod is connected to the overhead sliding track mechanism 18 as shown in FIG.
The container 24 and the heating furnace 14 are positioned vertically with respect to the quartz processing tube 12 and the water-cooled bottom plate 34.

The cylindrical loading cavity surround 3 having a front loading notch 88 and a rear loading notch 89
1 is positioned so as to cover the cylinders 22a to 22n and 30a to 30n, the container shield 24, the vertically positionable heating furnace 14, and the bell-shaped quartz processing tube 12, as shown in FIG. An inflatable seal 90 is disposed on top of the loading cavity enclosure 38 and expands radially during loading and unloading of a silicon or gallium arsenide wafer or substrate to expand and contract the container enclosure. 38, the central portion of the vertical heating reaction vessel system 10 can be kept clean and free of contaminants.

The ceramic heater element 26 includes a plurality of resistance wire heaters 92a to 92n and a plurality of thermocouple elements 94a to 94n, which detect the temperature of the ceramic heater and control it. The inside is covered with a heat-resistant elastic protective layer 97 made of zirconia oxide to prevent the transfer of metal contaminants. The water-cooled bottom plate 34 includes a bottom portion 34a, a top portion 34b, and a water chamber 98 facing between the bottom portion 34a and the top portion 34b. A quartz heat insulating plate ring 98 is provided above the upper bottom plate 34b and below the processing tube quartz ring 68. Further, another circular quartz heat insulating plate 100 is disposed above the center of the upper water-cooled bottom plate 34b to define a central atmosphere chamber 102. A polygonal cross section quartz firing table 36 having a plurality of side walls 104a to 104n, a quartz bottom 106, and a quartz top 108 is mounted on a quartz circular bottom shield 100 and extends to the middle of the quartz processing tube 12. It extends upward. A quartz heat shield 110 at the bottom of the base is mounted on the base bottom 106 and held stationary by a ring 112.

1 and 2, a group of wafers, schematically indicated at 114, is placed in a processing tube 12 in preparation for forming an epitaxial layer. The processing gas to the processing pipe 12 is supplied to the common port 126 via the connecting tools 118 and 120 via the supply pipes 122 and 124.
Then, after merging at the port 126, the gas is jetted to an upper portion inside the processing tube 12 so as to form an axial flow along the length direction of the processing tube 12. As shown in FIG.
An exhaust pipe 128 extending in the axial direction is provided in a bottom region of the exhaust pipe 128. The exhaust pipe 128 extends outwardly through the connecting member 130 and is connected to a vacuum source, so that the processing is performed. The processing gas inside the pipe 12 flows axially. The bottom plate 34 has a gas purge inlet 1
The quartz processing tube 12 can be purged by the presence of the connector 32 and the connector 134. Another purge inlet 136 and a connector 138 are provided on the same bottom plate, but the barge gas sent through the purge inlet 136 purges a portion between the processing pipe 12 and the element pipe 28 and then purges. Outlet 134 and connector 142
Is discharged through. The plurality of thermocouples 154a to 154n are made of a quartz thermocouple tube 150.
Is positioned vertically like a periscope on the quartz thermocouple tube, and an actuator 156 acts on a bottom rod 158 of the surrounding body tube 160 to prepare for temperature control by a computer. The temperature data in the temperate zone 162 is detected. Thermocouple leads 164a-164n are provided in thermocouple bottom bar 158 for computer temperature analysis.

Typical Batch Type Low Temperature Epitaxial Growth Method The low temperature epitaxial growth method can be performed using the above-described typical heated reactor. FIG. 3 shows a table of one embodiment of such a low-temperature epitaxial growth method.
The specific parameters shown in FIG. 3 are for the case of forming a 2-micron thick epitaxial layer on each wafer by a silicon epitaxial growth method using 200 wafers as one batch.

As shown in FIG. 3, prior to loading the wafer into the heated reactor 10, only the original silicon dioxide layer preferably remains at a thickness of about 2 nanometers, for example, using a wet cleaning process. Prepare the wafer to the point where it will do. Thereafter, the batch of wafers indicated by 114 is loaded into the heating reaction vessel 10, and the furnace 14 and the processing tube 12 at that time are loaded.
Is set to the ascending position. When loading a wafer, the temperature control system is 500
° C and set to an initial temperature while the wafer is being introduced into the processing area (wafer 1
This state is maintained for 0.1 minute per sheet). During that time, the wafer is flushed with 2% nitrogen gas.
Put under atmospheric pressure flowing in slm.

After all of the wafers have been loaded into the processing area, the processing tube 12 is lowered to surround the batch 114 of wafers, and the processing chamber defined therewithin the wafer processing inside the processing tube 12. Perform a pump and purge process to remove contaminants from the environment. During the pump-purge process, the aforementioned nitrogen gas is allowed to flow. Then, the furnace element 14 is lowered to cover the processing tube 12, and thereafter, a vacuum leak test is performed on the processing tube 12. Preferably, the pump purge process and the lowering of the furnace element 14 are performed twice. A vacuum leak test can ensure that a leak in the processing tube 12 does not cause the internal pressure to rise above a predetermined rate.

At the end of the pre-treatment described above, the actual processing of the batch 114 of wafers begins. First, the native oxide is removed from the wafer by exposing the wafer to a dry reducing gas atmosphere at sub-atmospheric pressure at a temperature not exceeding about 1,000 ° C. For this purpose, the temperature of the heating reaction vessel is rapidly raised from the original 550 ° C. to 950 ° C. FIG. 3 shows a typical temperature increase rate according to the detected temperature or the set temperature used in the present embodiment. During this temperature raising and constant temperature state, for example, only the hydrogen gas having a flow rate of 2 slm is introduced into the processing tube 12, while the introduction of the nitrogen gas is cut off. The pressure in the processing tube 12 is increased to, for example, about 1.333 hPa (about 1 Torr). The duration of this stage of the wafer processing cycle is preferably about 25 minutes, which is the case when the thickness of the wafers and the native oxide is as described above.

Removal of the original oxide in this manner allows the heating schedule for the wafer to be as scheduled. This is because oxide removal is performed at a temperature of about 1000 ° C. or less. Generally, removal of such contaminants requires a special etching gas and / or a high temperature of over 1000 degrees Celsius.

After removing the native oxide, the wafer is ready to produce an epitaxial layer. The conditions for forming the epitaxial layer in the heating reactor system are selected so that the surface reaction controls the process of depositing and forming the epitaxial layer. Therefore, the deposition temperature is selected so as not to exceed about 850 ° C., and so that the pressure in the processing tube 12 is within a range of about 0.1133 to 133 hPa (about 0.1 to 100 Torr).

According to the typical method shown in FIG. 3, the temperature is rapidly lowered to, for example, about 800 ° C. using the temperature rise / fall rate shown in FIG. Using this ramp rates, it takes about 15 minutes to lower the temperature from 950 ° C., the oxide removal temperature, to 950 ° C., the desired low temperature setting for epitaxial layer formation. After the temperature is lowered in this way, the heating reaction vessel 10 is left for a predetermined time, for example, 5 minutes in the case of the present embodiment, and is stabilized at a desired set deposition forming temperature. In this temperature raising and stabilizing step in a typical epitaxial growth method, the pressure in the processing tube 12 is preferably maintained at about 1.33 hPa (about 1 Torr), and the flow rate of nitrogen gas is maintained at about 2 slm. desirable.

After a predetermined stabilization time has elapsed, a precursor gas is introduced into the processing tube 12 to generate a desired epitaxial layer on the wafer. Where a silicon epitaxial layer is desired, the precursor is preferably silane or disilane. The use of such precursors can increase the deposition rate and increase the rate of HC
A mixture of L and SiCl ₄ , that is, SiHCL ₃ can be effectively replaced. For example, if a non-homogeneous epitaxial layer of silicon and germanium is desired, a germanium precursor such as GeH ₄ may be introduced into the processing tube 12 simultaneously with the silicon gas precursor.

In a typical embodiment, the introduction of the silane gas into the processing tube 12 is performed for about 2 minutes while maintaining the set temperature at about 800 ° C. However, the introduction time varies depending on the thickness of the epitaxial layer to be formed.

When the deposition cycle is completed, the temperature of the heating reactor system 10 is controlled so as to rapidly decrease the temperature to a rest temperature such as about 550 ° C. At the same time as the temperature
For example, the processing tube 12 is purged using nitrogen gas. This rapid cooling operation is shown in FIG.
It is desirable to use the temperature rise / fall rate shown in (1). If this temperature rise / fall rate is used, the above-mentioned rapid cooling and purging operation is completed in about 10 minutes. This rapid cool down / purge operation is followed by one or more pump / purge cycles. During this pump-purge cycle, the heating furnace 14 is raised from the state covering the processing tube 12 to the raised position.

The low temperature method utilized in the method of the present invention allows for the formation of very thin epitaxial layers on heavily doped substrates with a wide range of doping profiles. This is because the adverse effect of autodoping is significantly reduced as compared with the high-temperature epitaxial growth method.

The method according to the invention can be used for a conventional CMOS method. The method of the present invention utilizes low processing temperatures and allows for in-situ cleaning operations so that it does not interfere with device structures already introduced prior to epitaxial layer formation. Absent. In the case of high temperatures, on the other hand, the device structure is affected by the diffusion process, which causes crystal defects under such high temperature conditions.
In addition, cleaning operations used in high temperature epitaxial growth (mainly dry etching or annealing at higher temperatures) require that the top of the device structure be etched back.
k) Such reverse etching does not occur in the method of the present invention. Therefore, the method of the present invention can be used not only for integration into the conventional CMOS method but also for pre-processing the substrate.

Typical Heating Control System In order to carry out the epitaxial growth method described above, the temperature rise / fall rate of 3 ° C. per minute (3 ° C./min) used in the conventional heating reactor system is substantially higher. A heating furnace system 10 capable of controlling a large temperature (at least twice or more) rapid temperature rise / fall rate is required. The heating reactor used in the aforementioned epitaxial growth method is 10 ° C / min.
It is desirable that the rate of temperature rise / fall can be controlled within the range of -100 ° C / min. As described above, it is also desirable that the temperature of the wafer in the processing tube 12 can be accurately controlled in order to form an appropriate epitaxial film on the wafer at a low temperature. Preferably, the control system is capable of performing accurate temperature control over a wide temperature range used in the epitaxial growth method. It is also desirable to minimize temperature overshoot. Further, the heated reaction vessel 10 is more capable of maintaining a uniform wafer temperature over the entire wafer and from wafer to wafer better than about 0.5 ° C. in a temperature range between 500 ° C. and 850 ° C. Is desired. Such criteria may be, for example, to name each invention as "Tempera
ture Control System Using Multiple Temperature Controllers And Observer
Gain Control (Temperature Control System with Temperature Overshoot Contr) and Temperature Control System with Multiple Temperature Controllers
ol (temperature control system equipped with a temperature overshoot control device) "
U.S. Provisional Patent Application Ser. Nos. 60 / 084,907 and 60
No. 084,909 (both applications corresponding to Japanese Patent Application No. 2000-548914) can be satisfied by using a heating reaction vessel and a control system. These U.S. Provisional Patent Applications are hereby incorporated by reference.

By way of example, FIG. 4 illustrates an embodiment of a heated reaction vessel system 10 that is controlled using a programmable temperature control system suitable to achieve the functions and functions described above.
Shown in The temperature control system described below corresponds to that disclosed in the above-mentioned provisional application.

The heating reaction vessel system 10 includes a heating reaction vessel 212. The heating reaction vessel 212 includes the processing tube 12 that defines a reaction chamber as described above. Also,
The heating reaction vessel system 10 also includes a boat loader or paddle 218 used for putting the wafer group 220 into and out of the processing tube 12.

The wafer group 114 includes a plurality of boats 226 made of, for example, quartz or silicon carbide, and a plurality of silicon wafers 26 mounted on each boat 26. In the illustrated embodiment, the wafers on each boat 226 are isolated from each other, and these boats with wafers form a wafer processing row or semiconductor matrix processing row.

The heated reaction vessel 212 is provided with one or more heating elements 230, which surround the processing tube 12. In the embodiment shown, these heating elements 230 comprise resistive heating coils and extend the length of the processing chamber parallel to the longitudinal axis of the processing tube 12. Preferably, the heating element 230 is partitioned to define a plurality of independently controllable heating zones 232, which are configured to divide the heating coil into independently controllable zones. This can be achieved by providing a connection portion along the length direction. These zones can be controlled independently of each other by supplying power to both ends of the coil per zone or each part of the larger coil. In particular, the heating reaction vessel system 1
0 includes high current voltage transformers and silicon controlled rectifiers (SCRs) to provide controlled power to each heating zone 232.

The heating element 230 is incorporated in the ceramic heat insulating material 235. This insulation is designed to reflect or otherwise direct heat toward the row of wafers and form a more uniform layer that minimizes heat flux variations away from the processing row. are doing.

One or both of the temperature feedback and the temperature input used in controlling the process temperature are input from at least two output sources, ie, a plurality of spike thermocouples 236 and a plurality of profile thermocouples 242. It has become so. The term thermocouple, as used herein, should be understood to include various temperature sensors that include the more specific meaning of a thermocouple. Another temperature sensor configuration is also described using the term thermocouple.

A spike thermocouple 236 is located at a suitable location, such as between the heating element 230 and the processing tube 12, to measure the temperature of the heating element in each zone. Accordingly, these spike thermocouples 236 are spaced along the longitudinal direction of the heating element 30, but each heating zone 232 is arranged so that at least one spike thermocouple 236 faces. These spike thermocouples 236
Outputs specific and accurate information representing the temperature of each heating zone or the temperature of the heating element in each heating zone.

The profile thermocouple 242 is disposed along the profile rod 240 and is housed in a sleeve 238 extending inside the processing tube 12. This sleeve 23
8 is preferably made of quartz or silicon carbide. Profile rod 24
0 extends in parallel along the longitudinal direction of the processing tube 12. At least one thermocouple 242 is disposed in each heating zone 232. However, these profile thermocouples 242 need not necessarily be aligned with spike thermocouples 236. These profile thermocouples 242 measure the internal temperature of the process tube 12 and
Information indicating the temperature of the wafer group 114 in each heating zone is output.

If desired, a plurality of thermocoupled wafers 24 may be included during modeling of the heated reaction vessel.
4 may be used. FIG. 5 shows a state in which a thermocouple-attached wafer 244 is accommodated in the heating reaction vessel 12 of FIG. 4 at the time of modeling. These thermocoupled wafers 244 are equally spaced apart in the wafer group 116 so that the actual temperature of the wafer 228 can be accurately measured. Each thermocouple-attached wafer 244 includes a silicon wafer and two thermocouples 246, and the thermocouples 246 are attached to the edge and the center of the silicon wafer, respectively. Wafer 24 with each thermocouple
The thermocouple 246 is attached to the thermocouple 4 using, for example, a ceramic adhesive so that the actual temperature can be accurately obtained.

In the heating reaction vessel system 10, a controlled amount of processing gas is supplied to the processing tube 12 in order to grow or clean the surface of the silicon wafer 228 during the above-described epitaxial growth method. A gas supply device 250 is provided. As shown in FIG. 7, the gas supply control panel 248 includes a valve 252 and a mass flow controller 25.
4 is provided. The mass flow controller 252 measures the flow rate of the processing gas to the processing pipe 12 and is used to control the flow rate.

The pressure of the processing tube 12 is set so that a low pressure chemical vapor deposition type epitaxial growth method can be performed. As shown in FIG. 7, the heating reaction vessel system 10 includes a pressure controller 256 and a baratron or another baratron that measures the pressure in the processing tube and outputs the measured pressure to the pressure controller 256. Preferably, a suitable pressure sensing device 258 is provided. In addition, pumps and valves 60 are provided in the heating reaction vessel system 10 in a state of being connected to the pressure controller 56 so that a desired pressure is set in the processing tube 12 for the epitaxial growth method disclosed herein. It is desirable to make it exude.

Control of the operation of the hardware described above includes the user inputting a programmed prescription,
It is preferably done via an interface between the programmable control system suitable for execution and the hardware. 6, the heating reaction vessel system 10 includes a control system 264 for controlling the temperature in the heating reaction vessel 12 and other processing processes. The control system 264 preferably has two subsystems, a processing sequence system 266 that accepts and executes a processing sequence, and a temperature subsystem 26 that performs temperature control according to the processing sequence.
It is desirable to be divided into eight and eight. Both the processing sequence subsystem 266 and the temperature control subsystem 268 are adapted to follow a user-defined processing recipe entered via, for example, the user interface 282.

In the illustrated embodiment, the processing sequence subsystem includes a random access memory 272, a programmable EPROM 274 for storing control logic, a plurality of digital I / O channels 276, a plurality of analog I / O channels 278, An interface 282 is included. When external (remote) communication is required, a plurality of serial input / output channels 80 may be included in preparation for the communication.
Although other user interfaces are possible, the user interface 82 in the illustrated embodiment comprises a touch screen terminal interface that allows a user to enter a user-defined process prescription. In this processing prescription, the user defines a processing time, a gas flow rate, a chamber pressure, a temperature set value, a temperature rise / fall rate, and the like for each processing step. Parameters necessary for temperature control are transmitted to the temperature control subsystem 68 as input parameters.
Executes the temperature control in the prescription using this input parameter. Such a parameter input / output path is indicated by 87. In the illustrated embodiment, such parameters are:
It comprises at least the temperature set value that the user inputs according to the prescription.

As described above, the heating reaction vessel system 10 is provided with the gas supply system and the pressure control system. Such a system may include a processing sequence subsystem 266 and a control system interface shown in FIG. 7 that interfaces with one or both of the gas supply system and the pressure control system. As shown in FIG. 7, the gas supply control panel interface 3
04 is interposed between the gas supply control panel 148 and the processing sequence subsystem 168. The gas supply control board interface 304 provides an interface between the control system 264 and the mass flow controller 254, the gas valves 252, the pressure controller 256, the boat loader 218, and the like. In addition, the gas supply control board interface 304 provides a hardware safety interlock for the heated reaction vessel (e.g., supply of oxygen-mixed hydrogen at an appropriate mixing ratio,
Detection of the flame of the torch 62).

The temperature control system 268 controls the temperature in the heating reaction vessel 12 according to a user-defined recipe. This temperature control is preferably based on dynamic modeling of the heated reaction vessel 12, where the desired temperature condition is modeled by measurable system parameters. In operation, the recipe comprises temperature settings used by one or more dynamic models to drive the heated reaction vessel 12 to a desired temperature state.

Referring to FIG. 8, the temperature control subsystem 268 has two input signals for each heating zone, ie, a profile thermocouple input from which the temperature in the profile thermocouple of each heating zone can be determined, and Two input signals are provided: a spike thermocouple input at which the temperature at the spike thermocouple is anti-electron. These profiles and spike thermocouple temperatures are provided along line 415 via the illustrated thermocouple interface 330, along with a temperature set point input on line 335 and one or more heating elements output on line 340. Used to generate output control values. The output control value on line 340 is provided to the input of heating element control interface 345 and is used to control the power supplied to the elements in each heating zone via line 350.

FIG. 8 only illustrates one embodiment of the temperature control subsystem 268. As shown, a plurality of multivariable controllers 296, 298, 300 are used. Each controller 2
96, 298, 300 are preferably based on one or more dynamic models obtained from experimental data to optimize control accuracy over a predetermined temperature range that does not overlap with the temperature ranges of other controllers. Is desirable. To that end, controller 296 can optimize temperature control over the low level temperature range, and controller 298 can optimize temperature control over the medium level temperature range.
00 are each configured to optimize temperature control over a high level temperature range. The temperature range that the controller carries is generally only one range,
It may be desirable for each temperature range to overlap somewhat. According to one embodiment of the system, the low level temperature range may be centered at about 500 ° C, the medium level temperature range may be centered at about 800 ° C, and the high level temperature range may be centered at about 1050 ° C. The temperature ranges in which each controller operates will have a broad range of approximately ± 150 ° relative to the center temperature of these ranges.

Each controller 296, 298, 300 is constructed by using a robust optimization control theory on a model derived from the experience of the furnace and the base material to be heated. Specifically, in the illustrated embodiment, the multivariable controllers 296, 298, 300
Preferably, it is constructed using H-infinite control theory. These controllers 296
For the method of deriving the model used in, 298 and 300, the title of the invention is “Model Based Temperature Controller For Semiconductor Thermal Reactors”
No. 08 / 791,024, filed Jan. 27, 1997. As disclosed in this U.S. patent application, two dynamic models are used for each controller configuration: a model for the power input to the spike output and a model for the spike input to the profile output. Is desirable. However, it is also possible to use other multivariable control logic.

The control logic flow shown in FIG. 8 is largely independent of the particular control theory used to construct each controller 296, 298, 300. The only requirement that is generally sought is that each controller 296, 298, 300 be capable of generating an accurate control output based on one or more measured variable inputs.

In the illustrated system, the variable input to each controller is a set value input and a thermocouple 2
36, 242 and one or more temperature data inputs representing the temperature detected. The setpoint input illustrated as logic block 155 holds the setpoint temperature value to be achieved. This value is determined by the particular recipe to be implemented by system 10 and is provided to the inputs of each controller 296, 298, 300 simultaneously. Similarly, a plurality of data values representing the temperature values detected by the thermocouples 236, 242 are simultaneously input to the controllers 296, 298, 300. The set temperature values and thermocouple data values are supplied to the respective dynamic models corresponding to each of the controllers 296, 298, 300 to generate output control values output via lines 360, 365, 370, respectively. Has become.

As described above, the dynamic model used by each controller is optimized so that it can be used over a temperature range substantially excluding the temperature range formulated in the dynamic model of another controller. Thus, the power control values on lines 360, 365, and 370 provide a practical possibility that can be used to drive reaction vessel 12 to a desired set point. However, only one of these possibilities is that the model used by the controller is different for each model and is optimized only for a given temperature range, so it is optimized for given reactor conditions. You may have said that.
Therefore, the temperature subsystem 268 must be able to select which set of control values should be supplied to the heating element and thus the heating element ignition interface 245 to control the reaction vessel temperature.

In the illustrated system, the selection of which set of control values to supply to the interface 345 is made by the control output select switch logic indicated at 375. As shown, the control values on lines 360, 365, and 370 are provided to the input of a control output select switch 375, and the switch 375 selects a set of control values based on predetermined selection criteria on a control value output line 340. Output to Selection criteria are: Controller 2
Preferably, any dynamic model used to derive 96, 298, 300 is based on measurable input values indicating that it is optimal under certain conditions.

One of the criteria used by the control output selection switch 375 to select any control value to be supplied to the heating element ignition interface 345 is the set value input value at 335. As indicated by line 335, the set value input value may have been output as an input to control output select switch 375. Lines 360, 365
The set of control values appearing at the output of control output select switch 375 within the control values shown at 370 depend on the particular temperature set point input value on line 335.

In one mode of operation, the switching logic that depends on the value of the temperature setpoint input value is to assign an upper and / or lower set temperature threshold to each controller 296, 298, 300. Become. For example, the output value 360 of the low-level temperature controller 296 is selected and output to the output line 340 if the set temperature value is equal to or less than a predetermined threshold T _Low . If the set temperature value is equal to or higher than T _Low but equal to or lower than another threshold value T _High ,
The control value 365 of the middle level temperature controller 298 is selected and output to the output line 340. Further, if the set temperature value is equal to or higher than T _High , the control value 170 of the high-level temperature controller 300 is selected and output to the output line 340. In this example, it is desirable that the value of T _Low is selected so that the dynamic model used by the low-level temperature controller 296 is accurate or approximates to an upper limit temperature value that is at least one of optimal. . Alternatively (or in addition), the value of T _Low
It is desirable that the dynamic model used by the medium-level temperature controller 298 is selected to be accurate or approximate to a minimum temperature value that is at least one of the optimum.
T The value of _High, or dynamic model mid-level temperature controller 298 is utilized accurately, what is chosen to approximate to the maximum temperature value which is at least one of the best is desirable. Alternatively (or in addition), the value of T _High may be such that the dynamic model utilized by high-level temperature controller 300 is accurate,
It is desirable that the temperature is selected so as to approximate the lower limit temperature value which is at least one of the optimum values.

Another criterion used by the control output select switch 375 to select any control value to be supplied to the heating element ignition interface 345 is a thermocouple detected by one or more thermocouples 236, 242. There is a temperature input value. As indicated by line 380, the thermocouple temperature input value may be provided as an input to control output select switch 375. What is output as the output of the control output selection switch 375 is
Which set of the control values in the lines 360, 365, and 370 is the line 3
80 depends on the particular thermocouple temperature input value that appears. Preferably, the thermocouple temperature input value is determined using a thermocouple provided near the middle of the reaction chamber, such as the spike thermocouple 236a in FIG. From a thermocouple arranged near the middle of the reaction chamber, a temperature input reproducing the average temperature of the entire reaction chamber can be obtained.

In operation, the switching logic, which depends on the value of the thermocouple temperature input value, will assign a temperature range that selects each controller 296, 298, 300 for output. If the thermocouple input value represents a temperature in the range below the predetermined threshold T _Low , the set of control values on line 360 of the low level temperature controller 296 is selected and the output line 34 is selected.
Come out at 0. If the thermocouple input value represents a temperature in the range above T _Low but below another threshold T _High , the line 365 of the medium level temperature controller 298
Are selected and appear on the output line 340. Similarly, if the thermocouple input value represents a temperature in the range above T _High , the set of control values on line 370 of high-level temperature controller 300 is selected and appears on output line 340. In this example, it is desirable that the value of T _Low is selected so that the dynamic model used by the low-level temperature controller 296 is accurate or approximates to an upper limit temperature value that is at least one of optimal. But, alternatively (or in addition), the dynamic model used by the mid-level temperature controller 298 is accurate,
The temperature may be selected so as to approximate the lower limit temperature value that is at least one of the optimum values. Similarly, the value of T _High is desirably selected so that the dynamic model used by the medium-level temperature controller 298 is accurate or approximates the optimal upper limit temperature value that is at least one of the optimum. The method (or in addition) may be selected such that the dynamic model utilized by the high-level temperature controller 300 is accurate or approximates a lower temperature limit that is at least one of optimal.

The present inventors have determined that all controllers 296, 298, 300 have calculated control values based on their respective dynamic models, even if the control values of a particular controller are not utilized on output line 340. Has been found to be desirable. However, if the switching criterion for the previously unused controller is satisfied in the control output selection switching logic 375, such as when switching to the control value output of the unused controller, the control system 68 responds. It has also been found by the inventors that the container 12 will be driven to the desired set point in a sub-optimal manner. This is because the dynamic model used in the unused controller is different from the dynamic model of the controller used before the control value switching. With such different dynamic models, substantially different control values can be obtained for the output on the output line 340.

When the set of control values is being switched, the reaction vessel 12 is driven suboptimally (sub-optimal d).
Each controller 296, 298, 300 is logically connected to the monitor gain feedback logic 385, 390, 395 and the temperature control logic 4 to reduce riving.
00, 405, and 410. Each temperature control logic 400, 405,
410 is a set value input value 355, a plurality of thermocouple data values appearing on line 415,
A dynamic model based on the monitor gain feedback output values from the monitor gain feedback logic provided via lines 420, 425, and 430, respectively, is implemented. Each monitor gain feedback logic 385,
390, 395 are provided with one or more input values that are used to generate supervisory gain feedback output values that are provided to corresponding temperature control logic 385, 390, 395. In the illustrated embodiment, each monitor gain feedback logic 385, 390, 395 has a corresponding temperature control logic 400, 405.
, 410, the one or more controller output values are represented by lines 435, 440,
The feedback value provided by control output select switch logic 375 is provided via lines 450, 455, and 460, as well as via line 445. The controller output values on lines 435, 440, 445 are preferably the corresponding temperature control logic used in the supervisory feedback path within the temperature control logic to ultimately calculate the control value at the controller output. It is desirable that the data value be calculated within the range. The monitor feedback value output from control output select switch logic 375 includes an indication of the control output of the particular controller selected as output on output line 340.

In operation, the monitor gain feedback values on lines 420, 425, and 430 for a particular controller indicate that the control value output of the particular controller is on output line 340 in preparation for output to the heating element ignition interface 345. Is different for each given set of controller output values on lines 435, 440 and 445. For example, the value input on line 450 is applied to controller 296 on line 360.
Is output on output line 340, a set of control values X is output via line 360 at a predetermined set of low temperature controller output values on line 435. Also, line 4
If the value input at 50 indicates that the output of controller 96 on line 360 is not present on output line 340, then another set of control values Y at the predetermined set of low temperature control output values on line 435. Is output via line 360. Therefore, line 435,
The controller output model at 440, 445 is a first control matrix with corresponding supervisory gain feedback logic 385, 390, 395 if the control value output of a particular controller is provided via output line 340. If the control is to be performed but the control value output of a particular controller is not to be provided via output line 340, then the corresponding monitor gain feedback logic unit is to be subjected to a second matrix calculation. It is desirable. These values are then output to the output lines 420, 425, 430 of the corresponding supervisor gain feedback logic 385, 390, 395.
To the corresponding temperature control logic. The gain control matrix constant for the first gain control matrix calculation is preferably selected so that the particular controller can operate in the best mode within the temperature range for which the dynamic model is established. Another gain matrix constant is used for the second gain matrix calculation. Preferably, the second gain matrix is selected so that the control value output of a particular controller can typically track the control value output of the controller selected by control output selection switch 375. Accordingly, each controller outputs the optimum control value to its output terminal when the control output selection switch outputs the control value to the output line 340, but any of the control values of the other controllers is output to the output line 340. Preferably, when issued at 340, a control value tracking the control value output of one or more other controllers is output. First and second for specific temperature controllers 296, 298, 300
The constants in the gain matrix calculation may be different from the corresponding constants used in other temperature controllers (generally will be different in general).

As described above, the embodiment shown in FIG. 8 generally comprises temperature controllers 296, 29
8, 300 is independent of the particular multivariable controller utilized. However, FIG. 9 shows one embodiment of realizing a temperature controller using two dynamic models, a model relating to the power input to the spike output and a model relating to the spike input to the profile output. The dynamic model is specifically designed to use the profile and the temperature value output by the spike thermocouple. FIG.
Although the architecture of is shown at 296 of the low-level temperature controller,
It is suitable for use in the configuration of each temperature controller 296, 298, 300.

As shown, the controller, indicated at 296, controls line 49 based on the comparison of the temperature set point on line 335 with the profile thermocouple value input via line 300.
5 is provided with a profile error signal generator 490 for generating a profile error value to be output to the CPU 5. Similarly, spike error signal generator 505 outputs a spike error value via line 510. The profile error value on line 495 is provided to input 515 of profile controller 520 and the spike error value on line 510 is provided to input 525 of spike controller 530. Such a configuration is called "Model Based Temperature Controller For Semiconduc
tor Thermal Processors (Model-Based Temperature Controllers for Semiconductor or Heat Treatment Equipment) "
As disclosed in U.S. patent application Ser. No. 08 / 791,024 filed on Jan. 27, 1997, a detailed description thereof will not be given here for convenience.

Controller 296 implements supervisory gain feedback logic in accordance with the system architecture shown in FIG. 9 and the corresponding disclosures previously described. For this purpose, an input terminal 51 is provided using an anti-windup gain calculation matrix 535, a gain schedule calculation matrix 540, and a supervisor mode switch 545.
5 is calculated. Non-final gain calculation matrix 53
5 outputs via a line 555 the output value calculated using the first set of gain parameters. The gain schedule calculation matrix 540 outputs via line 560 the output value calculated using the second set of gain parameters. The respective output values on the output lines 555 and 560 are supplied to the input terminal of the supervisor mode switch 545.
The monitor mode switch 545 selects one of the output values supplied through the output lines 555 and 560, and outputs the selected output value to the profile controller 52.
0 input 515. That is, if the value of the signal indicated at 450 indicates that the control value at the output 560 of the controller 296 is being provided to the heating element ignition interface 345, then the supervisor mode switch 545 sets the non-final gain calculation matrix on line 555. The output value of 535 is selected. Similarly, if the value of the signal indicated at 550 indicates that the control value at the output 560 of the controller 296 has not been supplied to the heating element ignition interface 545, the supervisor mode switch 54
5 selects the output value of the monitor gain calculation matrix 535 on line 560.

A similar logic architecture is provided for spike controller 530. As shown, a non-final gain calculation matrix 570, a gain schedule calculation matrix 575, and a supervisor mode switch 580 are used to calculate the value of the parameter provided to input 525. Non-final gain calculation matrix 57
0 outputs on line 585 the output value calculated using the first set of gain parameters. The gain schedule calculation matrix 575 outputs an output value calculated using the second set of gain parameters to a line 590. Each output value on lines 585, 590 is provided to an input of a supervisor mode switch 580. The monitor mode switch 580 selects one of the output values supplied via the output lines 585 and 590, and outputs the selected output value to the input terminal 525 of the spike controller 530.
To supply. That is, if the value of the signal indicated at 450 indicates that the control value at output 360 of controller 296 is being provided to heating element ignition interface 345, supervisor mode switch 580 causes the non-final gain calculation matrix at line 585 to be activated. The output value of 570 is selected. Similarly, the value of the signal indicated by 450 is
The monitor mode switch 580 selects the output value of the monitor gain calculation matrix 575 on line 590 if the control value at output 360 at 6 indicates that it is not being supplied to the heating element ignition interface 345.

Using the logic architecture shown in FIG. 9, a de-coupled supervisory feedback system can also be implemented. In such a system, the gain corresponding to the circled input value is set to zero.

It should be noted that a wide range of system logic architectures utilize both profile and spike thermocouple data values to calculate control values to be selectively provided to the heating element ignition interface 345, individually. It can be used to implement a temperature control system that includes the disclosed supervisor gain feedback logic. For example, rather than using a spike and profile controller that is designed for the same temperature range, a set of output values from one of the multiple profile controllers for different temperature ranges is simply a value. It is also possible to design the temperature control system to be selectively supplied to a single common spike controller input. In such a system,
The disclosed supervisor mode feedback logic will apply only to each profile controller.

As described above, the rapid growth rate of the temperature is utilized in the epitaxial growth method of the present invention. In a temperature control system that has been used for controlling the temperature of a heated reaction vessel, the temperature of the reaction vessel is raised or lowered to a desired set value in accordance with a linear raising / lowering action. FIG. 10 is a graph showing the relationship between the reaction vessel temperature and time when the temperature of the reaction vessel is controlled in this manner. As shown, the temperature of the reaction vessel overshoots before the temperature controller can finally adjust the reaction vessel temperature to a set point. The occurrence of such overshoot undesirably significantly changes the heat treatment of the semiconductor wafer or other base material. In general, overshoot increases as the elevating rate increases, and is therefore highly undesirable when using a rapid elevating rate.

FIG. 11 shows a method for realizing the temperature control system 268 in which the above-mentioned overshoot is minimized so that the rapid temperature rise / fall rate can be used. As shown,
The controller 640 is supplied with a plurality of data values corresponding to the temperature values measured by the thermocouples 236, 242 via a line 635. Further, the temperature output value T _output is also supplied from the temperature increasing / decreasing temperature value generator 645. As described below, the temperature value generator 645 provides the controller 640 with a temperature output value T _output at least in response to a temperature set value input value logically indicated by 650, for example, as input by a user as part of the process prescription. Supply. The generation of the temperature output value T _output which helps prevent the temperature overshoot in the heating reactor 12 is based on the logic operation performed by the temperature raising / lowering temperature value generator.

The controller 640 is built by using a robust optimization control theory to a model derived from the experience of the furnace and the base material to be heated. Specifically, in the illustrated embodiment, controller 640 is preferably implemented as described above using a multiple temperature range model and viewer gain feedback. However, it is also possible to use other multivariable control logic configurations. Thus, the control logic flow shown in FIG. 11 is independent of the particular control theory used to construct the controller. The only requirement that is generally sought is that the controller 640 be able to execute a dynamic model based on one or more measured variable inputs.

Controller 640 applies the temperature output values from line 655 and the thermocouple data values from line 635 to the dynamic model. The controller 640 uses these input values to generate an output line 660 connected to the input of the heating element ignition interface 665.
To generate an output composed of a plurality of control values to be output to the. Ignition interface 665
Outputs controlled power via output line 670 based on the value of the control value, thereby adjusting the amount of heat in each zone to adjust the reactor temperature.

Here, the controller 640 is not supplied with a set value input as a direct input to the dynamic model. Rather, the temperature setpoint input value is supplied to a temperature value generator 645, from which the controller 640 sends an incremental temperature output value T _{output which the} controller 640 uses to control the temperature of the heating reactor 12. Is supplied. The temperature output value T _output forms a modified temperature ramp function over time. At a temperature output values forming the modified heating function, at a temperature up to a temperature set point input value T _sp during part of the heating stage Yutakaritsu (maximum ramp rate), or it controls in close rate 640 is driven. However, the rate of temperature increase of the corrected temperature gradient function is reduced as the temperature approaches the set value, thereby preventing the temperature overshoot in the heating reaction vessel 12.

There are several novel aspects to implementing the modified temperature gradient function of the temperature value generator 645. FIG. 12 shows one of them.

In FIG. 12, the target of the temperature value generator 645 is to drive the controller 640, ie, the temperature output value T _output that drives the reactor temperature from the initial temperature T _initial to the temperature setting input value T _sp . The idea is to put out a sequence. In the process, the temperature value generator 645
Outputs a temperature output value T _output forming a corrected temperature gradient function over time. Such a modified temperature gradient function is shown by line 678.

As indicated by reference numeral 678, the temperature set value input value is supplied to the temperature value generator 645 at time t ₁ , and the incremental temperature output value T _output corresponding to the maximum heating rate R _max is sent to the controller 640. Supplied. Maximum value of the temperature Yutakaritsu R _max is to be based on the value input by the user, or may be a predetermined system constant.

[0071] As the value of the temperature output value T _output approaches the temperature setting input values T _sp, the temperature value generator 645 supplies the temperature output value T _{outpu t} corresponding to the minimum temperature Yutakaritsu R _min to the control unit 640 Begin to. This minimum heating rate is used until the temperature output value T _output supplied from the temperature value generator 645 becomes equal to the temperature setting input value T _sp . This minimum heating rate Rmax may be based on a value input by the user, or may be a predetermined system constant. It is preferable that the heating rate is set to a minimum rate at which overshooting of the maximum allowable temperature occurs in the heating reactor 12 so that the heat treatment of the base material within the processing allowable range is not impaired.

The point at which the temperature value generator 645 switches from the maximum heating rate R _max to the minimum heating rate R _min can be set in several ways. For example, the temperature value generator 645 may switch to a minimum heating rate when the temperature value becomes equal to or exceeds the threshold value. This threshold may be based on the percentage X of the temperature setting input value _Tsp . The value of X may be part of the prescription entered by the user, or may be a predetermined system constant. Similarly, the temperature value generator 645 may switch to the minimum heating rate at a predetermined time during the heating phase. The predetermined time value in this case may be based on a predetermined percentage Z of the total heating time (t ₂ −t ₁ ).

FIG. 13 shows another modified temperature gradient function implemented by the temperature value generator 645. As indicated by the function line 682, the temperature value generator 645 generates a temperature output value T _output that drives the controller 640 at the minimum rate of temperature increase R _min at the point in time when the temperature increase phase has just started. At a predetermined point in time, such as the temperature setting input value T _sp or the percentage Y of the _initial temperature value T _initial , the temperature output value T _output that drives the controller 640 at the maximum heating rate R _max.
Is generated from the temperature value generator 645. Then, at another predetermined time point of the temperature set value T _sp or the initial temperature value T _initial , for example, at the percentage X, the controller sets the controller at the minimum temperature increase rate R _min until the temperature output value reaches the temperature set input value T _sp . A temperature output value T _output for driving 640 is generated from a temperature value generator 645. As described above, the switching of the heating rate may be performed based on the total heating time, a user-defined prescription value, a predetermined system constant, or the like.

Another modified temperature gradient function implemented by the temperature value generator 645 is shown in FIG. In this example, the heating rate R _current used by the controller 640 varies over the heating phase, but does not fall below the minimum heating rate R _min . Therefore, the temperature increase rate R _current at the predetermined temperature output value T _output is expressed as follows.

R _current = (T _target −T _output ) / τ where τ is a time constant (user-defined or system constant),

[0076] Here, _Tsp is a temperature set value, and ΔT represents the entire time of the temperature raising period.

As a result of the above-mentioned calculation of the heating rate, if a value smaller than the minimum heating rate R _min is obtained, then R _current = R _min and, therefore, the heating rate used by the temperature value generator. The rate does not fall below the minimum heating rate.

As shown in FIG. 14, with the temperature output value T _output generated as described above,
The transition between the minimum and maximum heating rates over the heating phase can be smooth. By making the transition smooth, the controller 640 can more accurately control the temperature of the heating reactor 12.

FIG. 15 shows a predetermined minimum heating rate value R._minIs the temperature output value T_outpushadow to bring to t
It shows the sound. In this figure, line 710 represents a large minimum heating rate R_minTo
Temperature output value T when used_outputFIG. Line 715 is line 71
A minimum heating rate R smaller than the minimum heating rate used to obtain the temperature output value indicated by 0 _min Temperature output value T when using_output3 shows a graph. Similarly, line 718
, 720 represent the temperature output values when a smaller minimum heating rate is used, respectively.
are doing. As can be seen from FIG. 15, at a small minimum heating rate, the temperature set value T_spTo
It takes time to reach the final value, but the transition to that value must be smooth.
Thus, the temperature overshoot in the heating reactor 12 is prevented or eliminated.
You can leave. Even if the time constant τ is changed, the same effect can be obtained. Time
If the constant is large, the transition between the minimum heating rate and the maximum heating rate will be smooth and
, The temperature output value is the temperature set value T_spYou can get a smooth transition as you reach
.

Although the modified temperature gradient function in the above-described embodiment has been described as being used in the temperature increasing step, it is apparent that the modified temperature gradient function can be similarly used in the temperature decreasing step. When used in such a temperature lowering step, one of the purposes of the modified temperature gradient operation is to keep the reactor temperature from falling below the newly applied set temperature.

Various modifications can be made to the system described so far without departing from the basic configuration suggested by the present invention. Although the present invention has been described in one or more specific embodiments, various modifications may occur to those skilled in the art without departing from the scope of the invention.

[Brief description of the drawings]

FIG. 1 is a perspective view showing one embodiment of a heating reaction vessel system that can be carried out by a batch type low-temperature epitaxial growth method of the present invention.

FIG. 2 is a cross-sectional side view of the exemplary heated reactor system shown at time 1.

FIG. 3 is a table illustrating one embodiment of the method of the present invention implemented in the exemplary heated reactor system shown in FIG.

FIG. 4 is a partial cross-sectional side view of a heated reaction vessel system that can be controlled using the temperature control system of the present invention.

FIG. 5 is a partial cross-sectional side view of the heated reactor system of FIG. 4 controlled during modeling and specification using a thermocoupled wafer.

FIG. 6 is a block diagram showing a preferred architecture of the entire control system including the temperature control system of the present invention.

FIG. 7 is a schematic block diagram illustrating one embodiment of a processing sequence system and a gas supply interface that can be used with the control system of FIG.

FIG. 8 is a logic flowchart showing the operation of the temperature control system configured according to one embodiment of the present invention.

FIG. 9 is a logic flow diagram illustrating the operation of a controller utilizing H-infinity control logic that may be used to implement the system of FIG.

FIG. 10 is a graph showing the relationship between reaction vessel temperature and time showing a temperature overshoot generally occurring when a conventional temperature control system is used.

FIG. 11 is a logic flowchart showing the operation of the temperature control system according to one embodiment of the present invention.

FIG. 12 is a graph illustrating one embodiment of a modified temperature gradient function applied to an input of a controller of a temperature control system to limit temperature overshoot during a heating phase.

FIG. 13 is a graph illustrating another embodiment of a modified temperature gradient function applied to an input of a controller of a temperature control system to limit temperature overshoot during a heating phase.

FIG. 14 is a graph illustrating yet another embodiment of a modified temperature gradient function applied to an input of a controller of a temperature control system to limit temperature overshoot during a heating phase.

FIG. 15 is a graph showing the influence of a minimum gradient value on the shape of the curve shown in FIG. 14;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 ... Heating reaction container system 12 ... Heating reaction furnace 26 ... Wafer 30 ... Heating element 32 ... Heating zone 36 ... Spike thermocouple 42 ... Profile thermocouple

[Procedural Amendment] Submission of translation of Article 34 Amendment of the Patent Cooperation Treaty

[Submission date] September 20, 2000 (2000.9.20)

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] Claims

[Correction method] Change

[Correction contents]

[Claims]

──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Thomas Morgenstern, Federal Republic of Germany-Day 15234 Frankfurt, Warsauer Strasse 40 (72) Inventor Dirk Volanski, Federal Republic of Germany-Day 15234 Frankfurt, Renne・ Strasse No. 4 (72) Inventor Paul Earl McHugh 1961 Darlington Drive, Kalispell, Montana, United States 1912 (72) Inventor Kevin Stoddard, United States 85255 Scottsdale, Arizona, East Happy Valley Road・ No. 12, 10040 (72) Inventor Constantinos Zacaris West Eye, Chandler, Arizona, USA 85226 Nho-stream door 3516 No. F-term (reference) 5F045 AA03 AB02 AB05 AC01 AD09 AD10 AD11 AD12 AE17 AE19 AE21 AE23 AE25 AF03 BB07 DP19 DP20 EB15 GB05

Claims (20)

[Claims] 1. A method for forming an epitaxial layer on a surface of a base material, comprising: providing one or more base materials in a processing chamber; evacuating a gas from the processing chamber; Replacing a portion with a dry reaction gas or a purge gas, adjusting the pressure in the processing chamber to a sub-atmospheric pressure, and adjusting the temperature to a value not exceeding about 1,000 ° C .; Removing the pressure of the processing chamber within a range of about 0.1133 to 133 hPa (about 0.1 to 100 Torr);
A method further comprising adjusting the temperature to a value not exceeding 850 ° C. and introducing a precursor gas into the processing chamber to form an epitaxial layer. 2. The method of claim 1, wherein the one or more preforms are cleaned and the oxide layer is removed prior to providing one or more preforms in the processing chamber. The method further comprising the step of bringing only one of them into a deposited state. 3. The method of claim 2, wherein said oxide layer has a thickness of about 20.
A method consisting of being Angstrom. 4. The method of claim 1, wherein, when further adjusting the temperature of the processing chamber, the temperature of the processing chamber is adjusted at a rate of about 10 ° C. to about 100 ° C. per minute. How to be. 5. The method of claim 1, wherein the temperature of the processing chamber after further adjusting the temperature of the processing chamber is in the range of about 500 ° C. to about 800 ° C. 6. The method of claim 1, wherein the temperature of the processing chamber after further adjusting the temperature of the processing chamber is about 800 ° C. 7. The method of claim 1, wherein the temperature difference between each of the preforms and the temperature difference between the one or more preforms is in the range of 0.5 ° C. . 8. The method of claim 1, wherein removing the oxide comprises inserting hydrogen gas into a processing chamber. 9. The method of claim 1, wherein the pressure in the processing chamber is adjusted to sub-atmospheric pressure, wherein the pressure is adjusted to about 1.33 hPa. 10. The method of claim 1, wherein said precursor gas comprises at least silane or disilane to form a silicon-containing epitaxial layer. 11. The method of claim 1, wherein the precursor gas comprises at least germane to form a germanium-containing epitaxial layer. 12. The method of claim 1, wherein said dry reaction gas or purge gas comprises nitrogen. 13. The method of claim 1, wherein the step of forming an epitaxial layer on the host material is performed prior to performing other method steps for forming an electronic device on the host material. The way that consists. 14. The method of claim 1, wherein the step of forming an epitaxial layer on the host material is performed after performing other method steps for forming an electronic device on the host material. How to be. 15. The method of claim 1, wherein each matrix is arranged horizontally and stacked vertically with another matrix. 16. The method of claim 1, wherein said processing chamber is part of a vertical reaction vessel. 17. The method according to claim 16, wherein the vertical reaction vessel comprises:
A method comprising a double elevating vertical reaction vessel. 18. The method according to claim 16, wherein the vertical reaction vessel comprises:
A method comprising a microprocessor controlled temperature control system for controlling a temperature of a processing chamber. 19. The method according to claim 18, wherein the vertical reaction vessel is
A method comprising providing a plurality of thermocouples for feedback to a temperature control system. 20. The method according to claim 16, wherein the vertical reaction vessel comprises an axisymmetric gas injector.