Information processing method and substrate processing apparatus

Abstract

To properly perform the film deposition process in the substrate processing equipment. [Solution] An information processing method comprising: using a plurality of pre-data, which are measurement results of the dope concentrations of a plurality of substrates processed in a pre-processing set, which is a set of a plurality of past film deposition processes in a film deposition apparatus, and a plurality of verification data, which are measurement results of the dope concentrations of a plurality of substrates processed in a verification processing set, which is a set of a plurality of film deposition processes of interest in the film deposition apparatus, to obtain an estimated distribution of the dope concentrations for the verification processing set; obtaining the difference between the estimated distribution and the measurement results of the dope concentrations of substrates processed in a verification run, which is a film deposition process of interest in the verification processing set; and determining that the measurement results of the verification run of interest are abnormal if the difference exceeds a predetermined threshold.

Description

[Technical Field]

[0001] This disclosure relates to an information processing method and a substrate processing apparatus. [Background technology]

[0002] Patent Document 1 discloses a film deposition apparatus for performing a SiC film deposition process on a silicon carbide (SiC) substrate. The film deposition apparatus includes a mounting table on which the SiC substrate is placed, a gas supply mechanism configured to form a flow of raw material gas from outside the mounting table in a direction perpendicular to the central axis of the mounting table, and an induction coil for heating the SiC substrate. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Patent No. 7001517 [Overview of the Initiative] [Problems that the invention aims to solve]

[0004] The technology described herein enables appropriate film deposition processing in a substrate processing apparatus. [Means for solving the problem]

[0005] One aspect of the present disclosure is an information processing method, which includes obtaining a plurality of pre-data that are measurement results of the doping concentrations of a plurality of substrates processed by a preprocessing set that is a set of a plurality of past film-forming processes in a film-forming apparatus, and a plurality of verification data that are the measurement results of the doping concentrations of a plurality of substrates processed by a verification processing set that is a set of the plurality of film-forming processes of interest in the film-forming apparatus, and using these to obtain an estimated distribution of the doping concentration for the verification processing set, obtaining a difference between the estimated distribution and the measurement result of the doping concentration of a substrate processed in a verification run that is the film-forming process of interest in the verification processing set, and determining that the measurement result of the verification run of interest is abnormal if the difference exceeds a predetermined threshold value.

Advantages of the Invention

[0006] According to the present disclosure, film-forming processing in a substrate processing apparatus can be appropriately performed.

Brief Description of the Drawings

[0007] [Figure 1] It is a diagram showing an outline of the configuration of a film-forming apparatus according to an embodiment. [Figure 2] It is a cross-sectional view showing an outline of the configuration inside a processing container of a film-forming apparatus according to an embodiment. [Figure 3] It is a diagram showing an outline of the configuration of an injector according to an embodiment. [Figure 4] It is a flowchart showing an outline of a configuration example of main steps of a film-forming method according to an embodiment. [Figure 5] It is a schematic diagram for explaining the memory effect. [Figure 6] It is a schematic diagram for explaining the memory effect. [Figure 7] It is a schematic diagram for explaining the memory effect. [Figure 8] It is a planar schematic diagram showing an outline of gas supply to a wafer and an outline of measurement points. [Figure 9]This flowchart shows an overview of the main steps of the flow rate determination method according to the embodiment. [Figure 10] This graph shows an overview of the Bayesian model and the measurement results during quality control (QC). [Figure 11] This graph shows an overview of the measurement results during quality control (QC) when some of the measurement results are replaced. [Modes for carrying out the invention]

[0008] The configuration of the substrate processing apparatus according to this embodiment will be described below with reference to the drawings. In this specification and the drawings, elements having substantially the same functional configuration are denoted by the same reference numerals, and redundant explanations will be omitted.

[0009] <Film forming equipment> Figure 1 is a schematic diagram of the configuration of a film deposition apparatus 1 as a substrate processing apparatus according to this embodiment. The film deposition apparatus 1 includes a processing container 11 that is roughly rectangular in shape. An exhaust line 12 is connected to the processing container 11, and the processing container 11 can be adjusted to a predetermined reduced pressure state by the exhaust line 12. The exhaust line 12 has an exhaust pipe 12a, one end of which is connected to the processing container 11. The exhaust pipe 12a consists of an exhaust manifold or the like, and a vacuum pump 12b, consisting of a mechanical booster pump or the like, is connected to the opposite side from the processing container. Between the processing container 11 and the vacuum pump 12b in the exhaust pipe 12a, a pressure adjustment unit 12c is provided, consisting of an APC (Automatic Pressure Control) valve or a proportional control valve, for adjusting the pressure inside the processing container 11. In addition, a pressure gauge 13 is provided in the processing container 11, and the pressure adjustment inside the processing container 11 by the pressure adjustment unit 12c is performed based on the measurement result from the pressure gauge 13.

[0010] The processing container 11 has a hollow rectangular prism-shaped processing container body 11a with openings at both ends, and side wall portions 11b connected to each end of the processing container body 11a so as to close the openings. The processing container body 11a and the side wall portions 11b are made of stainless steel or a dielectric material such as quartz.

[0011] An induction coil 14 connected to a high-frequency power supply 14a is provided outside the processing container body 11a. The induction coil 14 heats the substrate to be processed. For example, it inductively heats a susceptor case 23 and the like described later, and heats the substrate to be processed by the radiant heat from the inductively heated susceptor case 23.

[0012] Inside the processing container 11, a gas supply mechanism 15 is configured to supply a raw material gas or the like that is a raw material for film formation. The gas supply mechanism 15 is connected to an injector 16 provided in the processing container 11, which will be described later. The gas supply mechanism 15 has a first gas line L1 and a second gas line L2. The first gas line L1 and the second gas line L2 each include a first supply pipe 15a1 and a second supply pipe 15a2 connected to the injector 16, and a supply pipe 15b 16 ~15b 16 、15b 21 ~15b 22 and have.

[0013] Supply pipe 15b 11 ~15b 16 、15b 21 ~15b 22 Each has a mass flow controller (MFC) 15c as a flow controller 11 ~15c 16 、15c 21 ~15c 22 and a valve 15d 11 ~15d 16 、15d 21 ~15d 22 and are provided.

[0014] In the first gas line L1, a supply source 15e 11 is connected to the supply pipe 15b 11 and N2 gas is supplied from the supply source 15e 11 . Similarly, supply sources 15e 12 ~15b 16 are each connected to the supply pipe 15b 12 ~15e 16 and each supply source 15e12 ~15e 16 H2 gas, SiH4 gas, C3H8 gas, HCl gas, and Ar gas are supplied from here.

[0015] In the second gas line L2, supply pipe 15b 21 Source 15e 21 The supply source 15e is connected. 21 N2 gas is supplied from here. Similarly, supply pipe 15b 22 Source 15e 22 The system is connected and H2 gas is supplied.

[0016] In one embodiment, a supply source 15e that supplies N2 gas. 11 and source 15e 21 They may be supplied from the same source. That is, from a common source, supply pipe 15b 11 and supply pipe 15b 21 N2 gas may be supplied to it. Similarly, source 15e that supplies H2 gas. 12 and source 15e 22 They are the same supply source, and from the common supply source, supply pipe 15b 12 and supply pipe 15b 22 H2 gas may be supplied to it.

[0017] In one embodiment, another gas line (not shown) is provided to supply another single gas or gas mixture (not shown) to the injector 16. This other gas line has, for example, another supply source (not shown) that supplies C2H2 gas, C2H4 gas, or C2H6 gas.

[0018] When forming an n-type SiC film on a SiC substrate by epitaxial growth, SiH4 gas, C3H8 gas, H2 gas, N2 gas, and HCl gas are supplied to the processing container 11 from gas supply pipes 15b1 to 15b5 as raw material gases for film formation. A gas supply source and gas supply pipes for TMA (trimethylaluminum) gas may also be provided for the formation of a p-type SiC film.

[0019] Furthermore, when removing foreign matter adhering to the structure inside the processing container 11, for example, one of H2 gas or Ar gas, or a mixture of these gases, is supplied to the processing container 11 from the gas supply pipes 15b3 and 15b6.

[0020] The film deposition apparatus 1 described above is provided with at least one control unit 100, as shown in Figure 1. The control unit 100 processes computer-executable instructions that cause the film deposition apparatus 1 to perform various processes described herein, such as a flow rate determination method or an information processing method. The control unit 100 may be configured to control each element of the film deposition apparatus 1 to perform the various processes described herein. In one embodiment, part or all of the control unit 100 may be included in the film deposition apparatus 1. The control unit 100 may include a processing unit, a storage unit, and a communication interface. The control unit 100 is implemented, for example, by a computer. The processing unit may be configured to read a program from the storage unit that provides logic or routines that enable various control operations, and to perform various control operations by executing the read program. This program may be stored in the storage unit in advance, or it may be retrieved via a medium when needed. The retrieved program is stored in the storage unit and read from the storage unit and executed by the processing unit. The medium may be various storage media read by a computer, or it may be a communication line connected to a communication interface. The storage medium may be temporary or non-temporary. The processing unit may be a CPU (Central Processing Unit) and may consist of one or more circuits. The storage unit may include RAM (Random Access Memory), ROM (Read Only Memory), HDD (Hard Disk Drive), SSD (Solid State Drive), or a combination thereof. The communication interface may communicate with the film deposition apparatus 1 via a communication line such as a LAN (Local Area Network).

[0021] The information processing method disclosed herein may be executed by a device including a desired computer located outside the film deposition apparatus 1, which is different from the control unit 100 provided in the film deposition apparatus 1. Such a desired computer device may have a configuration similar to that of the control unit 100.

[0022] <Processing container> Next, the configuration inside the processing container 11 will be described. Figure 2 is a schematic cross-sectional view showing the general configuration inside the processing container 11 in the film deposition apparatus 1 of Figure 1. Inside the processing container 11, as shown in Figure 2, there is a susceptor 20 on which a SiC substrate (hereinafter referred to as wafer W) to be processed is placed via a holder H, a rotating shaft 21 that rotates the susceptor 20 and supports the susceptor 20, and a lifting unit 22 that raises and lowers the holder H on which the wafer W is placed. In addition, a susceptor case 23 is provided inside the processing container 11 as a housing unit, and the susceptor case 23 has a processing space S that houses the susceptor 20, and a processing gas is supplied to the processing space S from one end of the susceptor 20, passing over the center of the susceptor 20 and reaching the other end of the susceptor 20.

[0023] The susceptor 20 is formed in the shape of a disc with a recess 20a on its upper surface that is recessed vertically downward, and is installed horizontally inside the processing container 11. A holder H is fitted into the recess 20a. The holder H is rotated by the rotation shaft 21 around the central axis P of the susceptor 20 and the rotation shaft 21, thereby rotating the holder H as well.

[0024] Furthermore, the susceptor 20 is made of a conductive material that has high heat resistance and is easily heated by induction heating, and is composed of, for example, a graphite component whose upper surface is coated with SiC.

[0025] The holder H has a mounting area Ha formed on its upper surface on which the wafer W is placed. The holder H is made of a conductive material that has high heat resistance and is easily heated by induction heating. For example, the upper surface on which the wafer W is mounted is made of a graphite component coated with SiC. The holder H is formed in a disc shape with a smaller diameter than the susceptor 20, for example.

[0026] One end of the rotating shaft 21 is connected to the lower center of the susceptor 20, and the other end extends through the bottom of the processing container 11 to its lower side and is connected to a rotary drive mechanism (not shown). The rotation of the rotating shaft 21 by the rotary drive mechanism causes the susceptor 20 to rotate.

[0027] The lifting unit 22 is used to transfer the wafer W between the wafer W transport device outside the film deposition apparatus 1 and the susceptor 20, and in this example, it transfers the holder H on which the wafer W is placed. The holder H, i.e., the wafer W, is raised and lowered by the lifting drive mechanism (not shown) of this lifting unit 22.

[0028] The susceptor case 23 is formed in the shape of a rectangular parallelepiped with openings on two opposing sides. Processing gas is supplied from the opening on one side and discharged from the opening on the other side. In this structure, the processing gas supplied onto the wafer W is supplied and discharged along a direction parallel to the wafer W, which is perpendicular to the central axis P.

[0029] The susceptor case 23 is made of a conductive material that has high heat resistance and is easily heated by induction heating. For example, it is made of a graphite component with the wafer W-side surface coated with SiC.

[0030] Furthermore, an insulating material 24 is provided on the outer circumference of the susceptor case 23 to insulate the susceptor case 23 from the processing container 11. The insulating material 24 is formed, for example, using a fibrous carbon material with a high porosity. Although not shown in the figures, a holding structure is provided on the outside of the insulating material 24 to hold the insulating material 24 in a state where it is separated from the processing container 11.

[0031] Figure 3 is a schematic diagram of the configuration of the injector 16 according to this embodiment. The injector 16 is configured to extend along the susceptor 20 in a direction parallel to the wafer W to be placed on the susceptor 20. In addition, a plurality of supply ports, the first supply port 16h1 and the second supply port 16h2, are formed on the side portion 16s of the injector 16 facing the processing space S. The first supply port 16h1 and the second supply port 16h2 introduce gas supplied from the first gas line L1 and the second gas line L2 into the processing space S, respectively. In one embodiment, a diffusion space is provided inside the injector 16, configured to distribute the gas supplied from the first gas line L1 and the second gas line L2 to the respective supply ports 16h1 and 16h2.

[0032] The first supply port 16h1 is configured to introduce gas supplied from the first gas line L1 into the processing space S. The first supply port 16h1 is formed at approximately equal intervals in the width direction of the side surface 16s, excluding the central part in the width direction of the side surface 16s where the second supply port 16h2 is formed. In one embodiment, if other gas lines are provided, a plurality of other supply ports (not shown) for introducing gas supplied from those gas lines may be formed alternately with the plurality of supply ports 16h1.

[0033] The second supply port 16h2 is configured to introduce gas supplied from the second gas line L2 into the processing space S. Only one supply port 16h2 is formed in the center of the width direction of the side portion 16s. In one embodiment, multiple second supply ports 16h2 are formed near the center of the width direction of the side portion 16s.

[0034] <Film formation method> Next, a film deposition method as a substrate processing method using the film deposition apparatus 1 will be described. Figure 4 is a schematic flowchart showing an example of the configuration of the film deposition method according to this embodiment.

[0035] First, the holder H on which the wafer W is placed is brought into the processing container 11 (step St1 in Figure 4). Specifically, the holder H is brought into the processing container 11 from outside the film deposition apparatus 1 via a gate valve (not shown) using an external transport means (not shown) and positioned above the susceptor 20. Next, the lifting unit 22 is raised to support the holder H. Then, the transport means is moved out of the processing container 11, and the lifting unit 22 is lowered, placing the holder H on the susceptor 20.

[0036] After the holder H is loaded, the raw material gas and carrier gas are supplied from the gas supply mechanism 15 to the injector 16 in a direction perpendicular to the central axis P within the processing container 11. Along with the gas supply, high-frequency power is applied from the high-frequency power supply 14a to the induction coil 14 to heat the wafer W, and an n-type SiC film is formed on the wafer W by epitaxial growth (step St2 in Figure 4).

[0037] In step St2, in the first gas line L1, valve 15d 11 ~15d 15 Open the MFC15c 11 ~15c 15 The flow rate is adjusted to supply N2 gas, H2 gas, SiH4 gas, C3H8 gas, and HCl gas into the processing container 11. In addition, valve 15d in the second gas line L2 21 ~15d 22 Open the MFC15c 21 ~15c 22 The flow rate is adjusted and N2 gas and H2 gas are supplied into the processing container 11. At this time, the flow rate of the gas supplied from each gas line L1 and L2 is set to a target flow rate determined in advance by the method described later.

[0038] In step St2, high-frequency power is applied to the induction coil 14 from the high-frequency power supply 14a, thereby heating the wafer W through radiation and heat conduction from the induction-heated holder H, susceptor 20, and susceptor case 23. During film deposition, the pressure inside the processing container 11 is, for example, 10 Torr to 600 Torr, and the temperature of the wafer W is, for example, 1500°C to 1700°C.

[0039] After the film deposition is complete, the holder H supporting the wafer W is removed from the processing container 11 (step S3). Specifically, valves 15d1 to 15d5 are closed to stop the supply of raw material gas and carrier gas, then the lifting unit 22 is raised to raise the holder H supporting the wafer W. Then, an external transport means from the film deposition apparatus 1 is inserted into the processing container 11 via a gate valve and positioned below the holder H. Subsequently, the lifting unit 22 is lowered to transfer the holder H from the lifting unit 22 to the transport means, and the transport means is retracted from the processing container 11, thereby removing the holder H holding the wafer W from the processing container 11. Note that the supply of high-frequency power to the induction coil 14 may be interrupted during the removal of the wafer W, but it is preferable to supply high-frequency power to the induction coil 14 while controlling the temperature so that the susceptor 20 and susceptor case 23 reach the optimal temperature in the next process.

[0040] After removing holder H, the process may be returned to step S1, another holder H with another wafer W placed on it may be brought into the processing container 11, and the process from steps St1 to St3 may be repeated.

[0041] <Flow rate determination method> The following describes a method for determining the flow rate of the gas supplied in step St2 described above. The flow rate determination method according to this embodiment includes an information processing method of the present disclosure. Figures 5 to 7 are diagrams for illustrating the memory effect. Figure 8 is a schematic plan view showing the gas supply to the wafer W from both the first gas line L1 and the second gas line L2. Figure 9 is a schematic flowchart of an example configuration of the flow rate determination method according to this embodiment.

[0042] First, in a conventional film deposition process using a conventional film deposition apparatus, the memory effect when a SiC film is deposited on a wafer W by epitaxial growth in step St2 above will be explained using Figures 5 to 7.

[0043] Figure 5(a) shows the initial state, for example, immediately after the susceptor 20 has been replaced or cleaned. In the initial state, no byproduct BP is attached to the surface of the susceptor 20. Figure 5(b) shows the state after the same film deposition process as St2 has been repeated multiple times from the state in Figure 5(a), and Figure 5(c) shows the state after the film deposition process has been repeated multiple times further from the state in Figure 5(b). As shown in Figures 5(b) and (c), when the film deposition of St2 is repeated in the substrate processing, byproduct BP containing N2 may accumulate on the surface of the susceptor 20. Furthermore, the amount of such byproduct BP increases with the number of repeated film depositions. When a wafer W is placed on a susceptor 20 with accumulated byproduct BP and substrate processing is performed, a portion of the byproduct BP is etched by the H2 gas supplied during film deposition. At this time, nitrogen atoms N are released from the byproduct BP and doped into the SiC film of the wafer W near the byproduct BP. As a result, the amount of nitrogen doping in a certain portion of the SiC film on wafer W (hereinafter simply referred to as "doping amount") may become higher than the amount of doping expected from the N2 gas supplied by St2. This effect of increased doping amount due to the accumulation of byproduct BP during repeated film deposition is called the memory effect.

[0044] As shown in Figure 6, byproduct BP accumulates on the surface of the susceptor 20 and the surface of the susceptor case 23. Furthermore, since H2 gas for etching the byproduct BP is supplied from the injector 16, and the wafer W is rotated by the rotation axis 21 during film deposition, the SiC film is more heavily doped at the peripheral edge E of the wafer W near the injector 16. In Figure 7, areas with relatively high doping are shown with dark shading, and areas with relatively low doping are shown with light shading. As shown in Figure 7, due to the memory effect, the peripheral edge E of the wafer W is more heavily doped than the central part C. Also, due to the memory effect, the increase in doping at the peripheral edge E becomes more pronounced with increasing film deposition cycles.

[0045] In the flow rate determination method according to this embodiment, the flow rate of the gas to be supplied from the first gas line L1 and the second gas line L2 (hereinafter referred to as the "target flow rate") is determined taking into consideration the memory effect described above. Specifically, during the film deposition process of a desired number of wafers W (hereinafter referred to simply as "QC time" as quality control time), the dope concentration of the wafer W processed during QC time is measured, and the target flow rate for subsequent wafers W is determined based on the said dope concentration. As an example, the measurement of the dope concentration according to this embodiment is performed at a desired number of measurement points (P1, P2, ...) in the radial direction from the center C to the peripheral E of the wafer W, as shown in Figure 8. However, the position and number of measurement points are not limited to this, and the position and number of measurement points can be set to any desired position and number based on the required accuracy of the target flow rate, etc.

[0046] According to the film deposition apparatus 1 of the above embodiment, the target flow rate can be determined independently for the first gas line and the second gas line. As shown in Figure 8, the first gas line can supply gas to the entire width of the wafer W, while the second gas line can supply gas mainly to the center C of the wafer W. By determining the target flow rate for each of the first and second gas lines, taking into account the radial bias in dope concentration due to the memory effect, the distribution of dope concentration in the plane of the wafer W can be brought closer to the desired one.

[0047] In determining the target flow rate, the inventors conducted thorough research and found that when the next wafer W was subjected to film deposition using the target flow rate determined based on the dope concentration, the distribution of dope concentration within the surface of the wafer W may not be as desired. Further investigation by the inventors into such cases revealed that, during the measurement of the dope concentration of the wafer W processed during QC, the dope concentration at any measurement point in the radial direction of the wafer W may have been obtained as an incorrect value due to measurement anomalies or other reasons.

[0048] In view of the above problems, the present inventors conceived of detecting an anomaly in the measurement result when the dope concentration is obtained as an incorrect value in the measurement result after a certain run, and came up with the flow rate determination method according to the present embodiment.

[0049] Specifically, first, multiple sets of preliminary data are acquired (step St101 in Figure 9). The multiple sets of preliminary data acquired in step St101 are, for example, a dataset of the coordinates x (radial position of the wafer W) of multiple measurement points and the dope concentration y at each of those measurement points (see Figure 10). The multiple sets of preliminary data are, for example, multiple measurement results of dope concentrations after multiple past runs.

[0050] The pre-data according to this embodiment is obtained from a set of runs (hereinafter referred to as "processing sets") executed in a processing container 11 from the initial state of a certain susceptor 20 until it is replaced with the next susceptor 20. Preferably, the pre-data is obtained from the processing set of the susceptor 20 that was used immediately before the processing set of the susceptor 20 being used at the time when the flow rate determination method according to this embodiment is about to determine the target flow rate. This makes it possible to reduce the influence of memory effects other than those related to the susceptor 20 in the processing container 11, such as memory effects related to the susceptor case 23.

[0051] In the following description, the set of processing steps in the susceptor 20 for which the target flow rate is to be determined by the flow rate determination method according to this embodiment will be referred to as the "verification processing set," and one film deposition process in the verification processing set will be referred to as the "verification run." Furthermore, the processing set preceding the verification processing set for which prior data has been acquired will be referred to as the "pre-processing set," and one film deposition process in the pre-processing set will be referred to as the "pre-run."

[0052] Each of the multiple pre-run data sets, for example, is measured and acquired after the execution of a pre-run, and is stored in the control unit 100.

[0053] The acquisition of prior data is not limited to the example above; it may also be based on prior data acquired for a different processing container 11 than the processing container 11 on which the verification run is performed. Furthermore, prior data acquired for multiple processing sets, rather than just one processing set in processing container 11, may be used as the prior data for process St101.

[0054] Next, a verification run is performed to acquire multiple verification data (step St102 in Figure 9). The verification run is performed multiple times. In each verification run, the target flow rate is first determined, and the wafer W is processed using the above-described film deposition method with this target flow rate. In the repeated processing described later, the corrected target flow rate can be used.

[0055] After each verification run, verification data is acquired by measuring the dope concentration in the plane of the wafer W, and this verification data is stored, for example, in the control unit 100. The multiple verification data acquired in step St102 are, like the prior data, a dataset of the coordinates x (radial position of the wafer W) of multiple measurement points and the dope concentration y at each of those measurement points (see Figure 10).

[0056] Next, the posterior distribution of the linear regression coefficients is calculated from the prior and validation data based on Bayes' theorem, and the Bayesian estimate is determined (step St103 in Figure 9). In step St103 according to one embodiment, assuming that the linear regression coefficients follow a multivariate normal distribution, the posterior distribution is calculated as follows.

[0057] First, the distribution of linear regression coefficients is determined for multiple pre-processed data obtained in the pre-processing set. The linear regression coefficient is the coefficient vector a, expressed by equation (2) below, when the multiple regression equation for the radial position of wafer W (x; explanatory variable) and dope concentration (y; dependent variable) is set as shown in equation (1) below.

number

[0058] While not limited to this, when considering multiple regression using equation (1) above for measurement results as shown in Figure 10, the degree of x can be modeled as, for example, 3rd order or higher. The degree of x may also be determined in advance by calculating the fitting accuracy using prior data.

[0059] In this embodiment, the average value (y-bar) of the dope concentration at each radial position of the wafer W is calculated from multiple prior data, and equation (3) is obtained from this average value and equation (1) above. Furthermore, equation (4) is obtained for the coefficient vector μ0 as the average of the prior distribution.

number

[0060] Furthermore, the covariance Σ0 of the prior distribution is set to a desired value. In one embodiment, the standard deviation corresponding to each variance of the prior distribution can be easily calculated as 1 / 10 of the mean of the prior distribution obtained above. In another embodiment, the covariance of the prior distribution can be precisely calculated by obtaining the coefficient vector for each of the multiple prior data using equation (1) above.

[0061] Assuming that the linear regression coefficients follow a multivariate normal distribution, we obtain equation (5) below.

number

[0062] Next, we have a dataset of n validation data D{(x1,y1),···,(x n ,y n For ), the design matrix G and the dope concentration vector y are set as shown in equations (6) and (7) below.

number

[0063] In this case, the likelihood for dataset D is expressed by equation (8) below, where σ 2 σ is the error variance of the measurement of the dope concentration of wafer W. 2 This can be determined empirically. In one embodiment, the measurement error variance σ 2 This can be determined using a machine learning method based on multiple measurement results accumulated in the past. In one embodiment, the measurement error variance σ 2 This can be determined by performing multiple measurements on a wafer W after the film deposition process.

number

[0064] From Bayes' theorem, we obtain the following equation (9) for the posterior distribution from equations (5) and (8) above.

number

[0065] Rearranging equation (9) above, we can see that the posterior distribution follows a multivariate normal distribution (μ,Σ). In this case, the mean μ of the posterior distribution is the Bayesian estimate of the coefficient vector a. However, the covariance Σ and mean μ of the posterior distribution are expressed by the following equations (10) and (11).

number

[0066] The above explanation is merely one example of a case where the posterior distribution can be solved exactly, and the technology of this disclosure is not limited to this. For example, if the linear regression coefficients do not follow a normal distribution, they can be solved approximately. In this case, estimates can be obtained using approximation methods such as the Markov chain Monte Carlo method (MCMC method) or variational approximation.

[0067] Next, the Bayesian estimate determined above is used to determine if the verification data is abnormal (step St104 in Figure 9). In one embodiment, the abnormality determination in step St104 is performed on the verification data obtained by measurement after the verification run of interest in the verification processing set. In this embodiment, the verification run of interest may be a verification run during QC (hereinafter referred to as the "QC run").

[0068] In one embodiment, in step St104, the dope concentration at each measurement point in the verification data acquired in the QC run is compared with the dope concentration at each measurement point in the multiple regression equation (1) above using Bayesian estimates (hereinafter referred to as the "Bayesian model"). This comparison can be performed by taking the difference between the dope concentration at each measurement point in the verification data acquired in the QC run and the dope concentration at each measurement point in the Bayesian model, and determining whether the difference is below a predetermined threshold. If the difference is below the threshold at all measurement points ("YES" in Figure 9), the measurement results of the verification data acquired in the QC run can be determined to be normal. On the other hand, if there are measurement points where the difference exceeds the threshold ("NO" in Figure 9), the measurement results at least at the measurement points where the difference exceeds the threshold in the verification data acquired in the QC run can be determined to be abnormal.

[0069] For example, in Figure 10, the Bayesian model is shown as a solid line graph, and the dope concentration at each measurement point of the Bayesian model is shown as a square plot. Validation data obtained from measurements after the QC run is shown as a circle plot and a dashed line graph. In this case, the difference in dope concentration between measurement points P1 and P2 among measurement points P1 to P6 is represented by d1 and d2, indicated by the double-ended arrows in Figure 10. If d1 and d2 exceed the threshold, the measurement results at measurement points P1 and P2 can be judged as incorrect and abnormal.

[0070] The threshold can be determined based on the required accuracy for the target flow rate. In one embodiment, a Monte Carlo simulation is performed on a combination of an arbitrary measurement result and the dope concentration of a wafer W processed using the target flow rate based on the measurement result. By evaluating the contribution of the measurement result to the dope concentration, the threshold can be determined to fall within the range of the required accuracy.

[0071] In step St104, if the measurement results of the verification data obtained in the QC run are determined to be normal, the target flow rate for the wafer W to be processed in subsequent steps is determined using this verification data. That is, the target flow rate is corrected (step St105 in Figure 9). After step St105, the process may be repeated by returning to step St102 and executing a verification run using the corrected target flow rate.

[0072] In step St104, if the measurement results at one or more measurement points in the verification data obtained during the QC run are determined to be abnormal, abnormality processing is performed (step St110 in Figure 9).

[0073] In one embodiment, the abnormality handling in step St110 includes issuing a warning that there is an abnormality in the measurement result. After the warning is issued, the flow rate determination method is terminated, and for example, the operator of the film deposition apparatus 1 may inspect or replace the measuring device.

[0074] In one embodiment, the abnormality handling in step St110 includes replacing the measured values. For example, if the measurement results at measurement points P1 and P2 are abnormal, as shown in Figure 10, the measured values ​​at measurement points P1 and P2 are replaced with the dope concentration values ​​at measurement points P1 and P2 in the Bayesian model. Figure 11 shows a graph (circle plot and dashed line graph) of the verification data after the QC run when the measured values ​​at measurement points P1 and P2 are replaced. After replacing the measured values, the target flow rate is corrected using the replaced measured values ​​(step St105 in Figure 9).

[0075] Furthermore, when substituting measurement results, it is possible to substitute not only the measurement results at measurement points where the measurement results were determined to be abnormal, but also the measurement results at all measurement points with the dope concentration in the Bayesian model. In other words, if there is an abnormality, the measurement results may be discarded, and the Bayesian model may be used as the measurement result for that QC run.

[0076] According to the flow rate determination method including the above information processing method, abnormalities in the measurement results of the dope concentration can be detected in the verification run of interest, which is executed at any time in the verification processing set. Furthermore, if the measurement result is determined to be abnormal, an appropriate target flow rate can be determined by replacing at least a part of the measurement result with a Bayesian model. As a result, the distribution of dope concentration in the plane of the wafer W can be made to the desired state.

[0077] The embodiments disclosed herein should be considered in all respects as illustrative and not restrictive. The embodiments described above may be omitted, replaced, or modified in various ways without departing from the scope and spirit of the appended claims. For example, the constituent elements of the embodiments described above can be combined in any way. Such any combination will naturally yield the functions and effects of each constituent element in the combination, as well as other functions and effects that will be apparent to those skilled in the art from the description herein.

[0078] Furthermore, the effects described herein are merely descriptive or illustrative and not limiting. In other words, the technology relating to this disclosure may produce other effects that will be apparent to those skilled in the art from the description herein, in addition to or in lieu of the effects described herein. [Explanation of Symbols]

[0079] 1 Film deposition equipment 16 Injectors 16h1 1st supply port 16h2 2nd supply port 100 Control Unit L1 1st Gas Line L2 Second Gas Line W wafer

Claims

1. Information processing method, Multiple pre-processing data sets, which are measurement results of dope concentrations of multiple substrates processed in a pre-processing set that is a set of multiple past film deposition processes in a film deposition apparatus, Using the above-mentioned verification data, which is the measurement result of the dope concentration of multiple substrates processed by a verification processing set, which is a set of multiple film deposition processes of interest in the aforementioned film deposition apparatus, To determine the estimated distribution of the dope concentration for the verification processing set, The difference between the estimated distribution and the measured result of the dope concentration of the substrate processed in the verification run, which is the film deposition process of interest in the verification processing set, is obtained. An information processing method that includes determining that the measurement result of the verification run of interest is abnormal if the difference exceeds a predetermined threshold.

2. The aforementioned pre-data and verification data are datasets that include the coordinates of multiple measurement points on the substrate and the dope concentration at those measurement points. When the coordinates of the measurement point are used as the explanatory variable and the dope concentration is used as the dependent variable, The information processing method according to claim 1, comprising calculating the estimated distribution of the dope concentration as a posterior distribution in Bayes' theorem by assuming that the linear regression coefficients of the prior data follow a multivariate normal distribution.

3. The information processing method according to claim 2, further comprising obtaining the difference in the measurement results by comparing the measurement results of the verification run of interest with a Bayesian model using the mean of the linear regression coefficients of the posterior distribution for each of the multiple measurement points.

4. By acquiring the difference for each of the measurement points, an abnormality in the measurement result can be determined for each of the measurement points. The information processing method according to claim 3, comprising replacing, with respect to the measurement point where the measurement result is determined to be abnormal, the value of the dope concentration at at least that measurement point with the value of the dope concentration at that measurement point in the Bayesian model.

5. The information processing method according to claim 4, further comprising determining a target flow rate of gas to be supplied to substrates processed in the film deposition apparatus after the verification processing set using the replaced measurement results.

6. In the aforementioned film deposition apparatus, when the verification processing set is performed using a certain susceptor, The information processing method according to any one of claims 1 to 5, wherein the pre-processing set is a set of multiple film deposition processes performed using another susceptor that was used immediately before the susceptor used in the film deposition apparatus in the verification processing set.

7. A substrate processing apparatus that processes a substrate with gas, An injector having a supply port for introducing the gas into a processing space for processing the substrate based on a target flow rate, It comprises a control unit and, The control unit, Multiple pre-processing data sets, which are measurement results of dope concentrations of multiple substrates processed in a pre-processing set that is a set of multiple past film deposition processes in a film deposition apparatus, Using the above-mentioned verification data, which is the measurement result of the dope concentration of multiple substrates processed by a verification processing set, which is a set of multiple film deposition processes of interest in the aforementioned film deposition apparatus, To determine the estimated distribution of the dope concentration for the verification processing set, The difference between the estimated distribution and the measured result of the dope concentration of the substrate processed in the verification run, which is the film deposition process of interest in the verification processing set, is obtained. A substrate processing apparatus that performs control including determining that the measurement result of the verification run of interest is abnormal if the difference exceeds a predetermined threshold.

8. The substrate processing apparatus according to claim 7, comprising a susceptor on which the substrate is placed, and a susceptor case in which the susceptor is housed and the processing space is formed inside.

9. The system includes a first gas line and a second gas line that supply the gas to the injector, The supply port of the injector has a first supply port formed in a portion other than the central part of the injector, and a second supply port formed in the central part of the injector. The gas supplied from the first gas line to the injector is configured to be supplied to the processing space from the first supply port. The substrate processing apparatus according to claim 7, wherein the gas supplied from the second gas line to the injector is configured to be supplied from the second supply port to the processing space.

10. The substrate processing apparatus according to claim 9, further comprising a flow controller configured to individually control the flow rates of the gases supplied from the first gas line and the second gas line, respectively.

Images