Gas hole design method and gas hole design device

The method optimizes gas hole layouts in semiconductor manufacturing equipment through statistical modeling and optimization, addressing inefficiencies in manual adjustments and enhancing film quality.

JP7582748B2Active Publication Date: 2024-11-13TOKYO ELECTRON LTD
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
JP2021066447
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2021-04-09
Publication Date
2024-11-13
Estimated Expiration
2041-04-09

AI Technical Summary

Technical Problem

Existing methods for designing gas holes in semiconductor manufacturing equipment are inefficient and require significant manual adjustments, leading to high man-hours and difficulty in achieving optimal film quality.

Method used

A method involving the design of gas holes that includes modeling and optimization using a statistical approach, utilizing black box optimization to determine kernel shapes and tree structure search to optimize hole layouts, reducing manual adjustments and improving film quality.

Benefits of technology

This method allows for precise determination of gas hole layouts, reducing manual effort and improving film thickness uniformity and quality by leveraging statistical models and optimization algorithms.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007582748000001
    Figure 0007582748000001
  • Figure 0007582748000002
    Figure 0007582748000002
  • Figure 0007582748000003
    Figure 0007582748000003
Patent Text Reader

Abstract

To provide a method for designing gas holes which efficiently designs an arrangement of gas holes.SOLUTION: A method for designing gas holes which supply a gas into a processing container of a semiconductor manufacturing apparatus includes a step of creating a model for evaluating film quality, on the basis of, an arrangement of the gas holes, and a plurality of data sets for comparing film formation results obtained by supplying the gas from the gas holes in the arrangement of the gas holes and actually forming films, and a step of calculating an evaluation value of film quality in the arrangement of the gas holes using the model, and optimizing the arrangement of the gas holes.SELECTED DRAWING: Figure 5
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present disclosure relates to a method for designing a gas hole and an apparatus for designing a gas hole. [Background technology]

[0002] As one of the methods by which a semiconductor manufacturing apparatus forms a thin film such as a silicon oxide film on a substrate such as a wafer, a CVD (Chemical Vapor Deposition) method or an ALD (Atomic Layer Deposition) method, in which a plurality of types of process gases are sequentially supplied to the surface of the wafer and a reaction product is deposited, is known. Such a semiconductor manufacturing apparatus may have a shower head for gas supply (see, for example, Patent Document 1). Patent Document 1 discloses a semiconductor manufacturing apparatus having a source gas supply means and an auxiliary gas supply means. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2018-78233 A Summary of the Invention [Problem to be solved by the invention]

[0004] The present disclosure provides a method for designing gas holes that efficiently designs the arrangement of gas holes. [Means for solving the problem]

[0005] In view of the above problems, a method for designing gas holes that supply gas into a processing vessel of a semiconductor manufacturing apparatus includes the steps of: generating a model for evaluating film quality based on a plurality of data sets that correspond the arrangement of the gas holes to the results of film formation that is actually formed by supplying gas from the gas holes in the arrangement of the gas holes; and calculating an evaluation value of film quality for the arrangement of the gas holes using the model, and optimizing the arrangement of the gas holes. , has. Effect of the Invention

[0006] It is possible to provide a method for efficiently designing the arrangement of gas holes. [Brief description of the drawings]

[0007] [Figure 1] FIG. 1 is a cross-sectional view showing an example of the configuration of a semiconductor manufacturing apparatus according to an embodiment; [Diagram 2] Plan view of the semiconductor manufacturing equipment of FIG. [Diagram 3] A cross-sectional view of the rotary table of the semiconductor manufacturing equipment of FIG. [Figure 4] Hardware configuration diagram of an example of a showerhead design device [Diagram 5] A flowchart illustrating an example of the overall flow of a showerhead design method. [Figure 6] A plan view of an example of a shower head. [Figure 7] A functional block diagram of an example that explains the functions of a showerhead design device by dividing them into blocks. [Figure 8] An example diagram explaining the kernel and its shape [Figure 9] FIG. 1 is an explanatory diagram of an example of a location on a wafer where film thickness is evaluated; [Figure 10] A schematic diagram of an example explaining the estimation of film thickness using a kernel [Figure 11] Schematic diagram of an example of black-box optimization [Figure 12] An example of a flowchart showing the procedure for optimizing the kernel shape using the black-box optimization method by the kernel shape determination unit. [Figure 13] Diagram showing the effect of internal parameters on the kernel objective function [Figure 14] The shape of the kernel obtained after adjustment is shown in relation to the position on the Y line of the showerhead. [Figure 15] An example diagram explaining the accuracy of the model [Figure 16]FIG. 1 is an explanatory diagram of an example of determining a hole layout by searching a tree structure. [Figure 17] FIG. 1 is a flowchart showing an example of a procedure for optimizing a hole layout by a gas hole optimization unit. [Figure 18] A schematic diagram of an example explaining the hole layout optimization method [Figure 19] A diagram showing an example of an optimized hole layout. [Figure 20] A diagram showing the deposition results calculated by the design tool using the optimized hole layout. [Figure 21] A diagram showing an example of the actual film formation results when a semiconductor manufacturing device forms a film using a shower head with an optimized hole layout. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Hereinafter, non-limiting exemplary embodiments of the present disclosure will be described with reference to the accompanying drawings. In all of the accompanying drawings, the same or corresponding members or parts are designated by the same or corresponding reference numerals, and duplicated descriptions will be omitted.

[0009] [Example of the configuration of a semiconductor manufacturing device with a shower head] First, an example of a semiconductor manufacturing apparatus equipped with a showerhead will be described with reference to Figures 1 to 3. The semiconductor manufacturing apparatus 1 of Figure 1 is an example of a processing apparatus that executes the method for controlling a turntable of the present disclosure.

[0010] The semiconductor manufacturing apparatus 1 of the present disclosure includes a chamber 1 having a substantially circular planar shape, and a turntable 2 that is provided within the chamber 1 and has a rotation center at the center of the chamber 1. The chamber 1 includes a container body 12 having a cylindrical shape with a bottom, and a top plate 11 that is detachably and airtightly disposed on the upper surface of the container body 12 via a seal member 13 such as an O-ring.

[0011] The rotary table 2 is fixed at the center to a cylindrical core 21, and the core 21 is fixed to the upper end of a rotary shaft 22 extending in the vertical direction so as to be rotatable. The rotary shaft 22 penetrates the bottom 14 of the chamber 1, and its lower end is attached to a drive unit 23 that rotates the rotary shaft 22 around the vertical axis. The rotary shaft 22 and the drive unit 23 are housed in a cylindrical case body 20 with an open upper surface. A bellows 16 is provided between the bottom 14 of the container main body 12 and the case body 20. As a result, the case body 20 is airtightly attached to the lower surface of the bottom 14 of the chamber 1, and the internal atmosphere of the case body 20 is isolated from the external atmosphere. The drive unit 23 may be a motor.

[0012] Further, a lifting mechanism 17 capable of raising and lowering the turntable 2 and changing the height of the turntable 2 is provided outside the bellows 16. The lifting mechanism 17 raises and lowers the turntable 2, and is configured to be able to change the distance between the ceiling surface 45 and the wafer W in response to the raising and lowering of the turntable 2. The lifting mechanism 17 may be realized by various configurations as long as it can raise and lower the turntable 2, and may be configured, for example, to extend and contract the length of the rotation shaft 22 by a gear or the like.

[0013] A first exhaust port 610 is provided on the outer periphery of the chamber 1 and communicates with an exhaust pipe 630. The exhaust pipe 630 is connected to a vacuum pump 640 via a pressure regulator 650, and is configured so that the inside of the chamber 1 can be exhausted from the first exhaust port 610.

[0014] FIG. 2 is a diagram for explaining the structure inside the chamber 1, and shows the structure inside the chamber 1 when the semiconductor manufacturing apparatus 1 according to the embodiment is viewed from above, with the top plate 11 omitted. On the surface of the turntable 2, a circular recess 24 is provided along the rotation direction (circumferential direction) as shown in FIG. 2 for placing a plurality of semiconductor wafers (hereinafter referred to as "substrates" or "wafers") W (five in the illustrated example). Note that FIG. 2 shows a wafer W in one recess 24. This recess 24 has an inner diameter slightly (e.g., 2 mm) larger than the diameter of the wafer W (e.g., 300 mm) and a depth approximately equal to the thickness of the wafer W, and is a placement portion on which the wafer W can be placed. Therefore, when the wafer W is placed on the recess 24, the surface of the wafer W and the surface of the turntable 2 (the area on which the wafer W is not placed) are at the same height. On the bottom surface of the recess 24, a through hole (none of which is shown) is formed through which, for example, three lift pins pass to support the back surface of the wafer W and lift the wafer W. The transfer port 15 is used for transferring the wafer W between the external transfer arm 10 and the rotary table 2 .

[0015] Above the turntable 2, a reaction gas nozzle 31, a reaction gas nozzle 32, and a separation gas nozzle 41, 42, each made of, for example, quartz, are arranged. In the illustrated example, the separation gas nozzle 41, the reaction gas nozzle 31, the separation gas nozzle 42, and the reaction gas nozzle 32 are arranged in this order at intervals in the circumferential direction of the chamber 1, clockwise from the transfer port 15 (the direction of rotation of the turntable 2 indicated by the arrow 9). These nozzles 41, 31, 42, 32 have gas introduction ports 41a, 31a, 42a, 32a, which are the base ends, fixed to the outer circumferential wall of the vessel body 12. As a result, the nozzles are installed so that gas is introduced into the chamber 1 from the outer circumferential wall of the chamber 1 and extends parallel to the turntable 2 along the radial direction of the vessel body 12.

[0016] A first reactive gas supply source storing a first reactive gas is connected to the reactive gas nozzle 31 via an on-off valve and a flow rate regulator (both not shown). A second reactive gas supply source storing a second reactive gas that reacts with the first reactive gas is connected to the reactive gas nozzle 32 via an on-off valve and a flow rate regulator (both not shown).

[0017] Here, the first reaction gas is preferably a gas containing a semiconductor element or a metal element, and when it becomes an oxide or nitride, it can be used as an oxide film or a nitride film. The second reaction gas is selected as an oxidizing gas or a nitriding gas that can react with a semiconductor element or a metal element to generate a semiconductor oxide or a semiconductor nitride, or a metal oxide or a metal nitride. Specifically, the first reaction gas is preferably an organic semiconductor gas or an organic metal gas that contains a semiconductor element or a metal element. In addition, the first reaction gas is preferably a gas that has an adsorbing property to the surface of the wafer W. The second reaction gas is preferably an oxidizing gas or a nitriding gas that can undergo an oxidation reaction or a nitriding reaction with the first reaction gas that is adsorbed to the surface of the wafer W, and can deposit a reactive compound on the surface of the wafer W.

[0018] Specifically, for example, the first reaction gas is a reaction gas containing silicon, and is an organic aminosilane such as diisopropylaminosilane or bis(tertialbutylaminosilane) (BTBAS) that forms an oxide film of SiO2 or a nitride film of SiN. Alternatively, the first reaction gas is a reaction gas containing hafnium, and is tetrakisdimethylaminohafnium (hereinafter referred to as "TDMAH") that forms an oxide film of HfO. Alternatively, the first reaction gas is a reaction gas containing titanium, and is TiCl4 or the like that forms a nitride film of TiN. In addition, the second reaction gas is an oxidation gas such as ozone gas (O3) or oxygen gas (O2). Alternatively, the second reaction gas is a nitriding gas such as ammonia gas (NH3).

[0019] In addition, a supply source of a rare gas such as Ar or He or an inert gas such as nitrogen (N2) gas is connected to the separation gas nozzles 41 and 42 via an opening / closing valve or a flow rate regulator (both not shown). The inert gas supplied from the separation gas nozzles 41 and 42 is also referred to as a separation gas. In this embodiment, for example, N2 gas is used as the inert gas.

[0020] The reaction gas nozzle 31 is connected to a first reaction gas supply source, a second reaction gas supply source, and a supply source of an inert gas used as a separation gas such as a rare gas such as Ar or He, or nitrogen (N2) gas. A switching unit (not shown) is operated to switch which gas is supplied. The reaction gas nozzle 32 is connected to a second reaction gas supply source, a first reaction gas supply source, and a supply source of an inert gas used as a separation gas, and a switching unit (not shown) is operated to switch which gas is supplied.

[0021] The area defined below the reaction gas nozzle 31 becomes a first processing area P1 for adsorbing the first reaction gas onto the wafer W. The area defined below the reaction gas nozzle 32 becomes a second processing area P2 for oxidizing or nitriding the first reaction gas adsorbed onto the wafer W in the first processing area P1.

[0022] In the chamber 1, two convex portions 4 are attached to the back surface of the top plate 11 so as to protrude toward the turntable 2. The convex portions 4, together with the separation gas nozzles 41 and 42, constitute a separation region D. That is, the region below the separation gas nozzles 41 and 42 separates the first processing region P1 from the second processing region P2, and serves as the separation region D that prevents the first reaction gas from mixing with the second reaction gas. The convex portions 4 have a substantially fan-shaped planar shape with an arc-shaped top, and in the present disclosure, the inner arc is connected to a protruding portion 5 (see FIG. 1) that protrudes beyond the ceiling surface 45, and the outer arc is disposed along the inner circumferential surface of the vessel body 12 of the chamber 1. As shown in FIG. 1, the protruding portion 5 is provided so as to surround the outer periphery of the core portion 21 that fixes the turntable 2.

[0023] 2 is attached to the rear surface of the top plate 11. Therefore, the lower surface of the convex portion 4 is lower than the ceiling surfaces 45 located on both sides of the lower surface in the circumferential direction.

[0024] A groove (not shown) is formed in the circumferential center of each of the convex portions 4, and accommodates separation gas nozzles 41, 42. The separation gas nozzles 41, 42 have gas ejection holes formed therein.

[0025] The reaction gas nozzles 31 and 32 are provided near the wafer W, spaced apart from the ceiling surface 45. The N2 gas supplied from the separation gas nozzle 42 acts as a counter flow to the first reaction gas from the first processing region P1 and the second reaction gas (oxidizing gas or nitriding gas) from the second processing region P2. Therefore, the first reaction gas from the first processing region P1 and the second reaction gas from the second processing region P2 are separated by the convex portion 4 and the N2 gas. Therefore, the first reaction gas and the second reaction gas are prevented from mixing and reacting with each other in the chamber 1.

[0026] A first exhaust port 610 and a second exhaust port 620 are formed between the turntable 2 and the inner peripheral surface of the container body 12. A pressure regulator (APC, Auto Pressure Controller) 650 is provided in an exhaust pipe 630 between the first exhaust port 610 and a vacuum pump 640 shown in FIG. 1. Similarly, the second exhaust port 620 is connected to a vacuum pump (not shown) via an exhaust pipe provided with a pressure regulator. The exhaust pressures of the first exhaust port 610 and the second exhaust port 620 are configured to be independently controllable.

[0027] 1, a heater unit 7, which is a heating means, is provided in the space between the turntable 2 and the bottom 14 of the chamber 1. The heater unit 7 provided below the turntable 2 heats the wafer W on the turntable 2 to a temperature (e.g., 450° C.) determined by a process recipe via the turntable 2.

[0028] The case body 20 is provided with a purge gas supply pipe 72 for supplying N2 gas as a purge gas for purging. Furthermore, a plurality of purge gas supply pipes 73 for purging the arrangement space of the heater unit 7 are provided at a predetermined angular interval in the circumferential direction below the heater unit 7 on the bottom 14 of the chamber 1. A ring-shaped cover member 71 is provided below the periphery of the turntable 2, and a lid member 7a for covering the heater unit 7 is provided between the heater unit 7 and the turntable 2. The lid member 7a can be made of, for example, quartz. This prevents gas from entering the area where the heater unit 7 is provided.

[0029] By supplying N2 gas from the purge gas supply pipe 72 and the purge gas supply pipe 73, the flow of N2 gas can suppress mixing of the gas in the space 481 and the space 482 shown in FIG. 2 through the space below the center of the chamber 1 and the space below the turntable 2.

[0030] A separation gas supply pipe 47 is connected to the center of the top plate 11 of the chamber 1, and is configured to supply N2 gas, which is a separation gas, to a space 481 between the top plate 11 and the core part 21. The separation gas supplied to the space 481 is discharged toward the periphery along the surface of the turntable 2 on the wafer placement area side through a narrow space 46 between the protrusion 5 and the turntable 2. The space 46 can be maintained at a higher pressure than the spaces 481 and 482 by the separation gas. Therefore, the space 46 prevents the first reaction gas supplied to the first processing region P1 and the second reaction gas supplied to the second processing region P2 from mixing through the space 48.

[0031] The semiconductor manufacturing equipment 1 is provided with a control unit 90 that controls the operation of the semiconductor manufacturing equipment 1. Various sensors are provided in the semiconductor manufacturing equipment 1. As an example of the various sensors, a temperature sensor 60 that measures the temperature of the turntable 2 is provided. The temperature sensor 60 may be, for example, a radiation thermometer that is provided above the turntable 2 and uses the difference in material between the turntable 2 and the wafer W to determine detachment of the wafer W placed on the placement section or to measure the temperature of the turntable 2. The temperature sensor 60 is provided in contact with or without contact with the turntable 2 and measures the temperature of the turntable 2.

[0032] When the temperature sensor 60 is a radiation thermometer, for example, the radiation thermometer is provided on an external window of the chamber 1, and the temperature of the object is measured by measuring the intensity of infrared rays or visible light emitted from the object. By using a radiation thermometer as the temperature sensor 60, the temperature of the turntable 2 can be measured quickly and without contact. The temperature sensor 60 transmits the measured temperature to the control unit 90. The control unit 90 acquires the temperature measured by the temperature sensor 60 and uses it to control the elevation of the turntable 2.

[0033] FIG. 3 is a cross-sectional view taken along a concentric circle of the turntable 2 of the semiconductor manufacturing apparatus 1 of FIG. 1, from the separation region D through the first processing region P1 to the separation region D. FIG.

[0034] A generally sector-shaped convex portion 4 is provided on the top plate 11 in the separation region D. The convex portion 4 is attached to the rear surface of the top plate 11. In the space below the top plate 11, a flat, low ceiling surface (hereinafter referred to as a "first ceiling surface 44") which is the lower surface of the convex portion 4, and ceiling surfaces (hereinafter referred to as "second ceiling surfaces 45") which are higher than the first ceiling surface 44 and located on both sides of the first ceiling surface 44 in the circumferential direction are formed.

[0035] 3, the convex portion 4 forming the first ceiling surface 44 has a sector-shaped planar shape with its top cut into an arc shape. A groove portion 43 is formed in the convex portion 4 at the circumferential center so as to extend in the radial direction. Separation gas nozzles 41 and 42 are housed in the groove portion 43. The periphery of the convex portion 4 is bent in an L-shape so as to face the outer end surface of the turntable 2 and to be slightly spaced from the container body 12 in order to prevent mixing of the respective process gases.

[0036] A nozzle cover 230 is provided above the reactive gas nozzle 31 so that the first process gas flows along the wafer W and the separation gas flows along the top plate 11 side while avoiding the vicinity of the wafer W. As shown in FIG. 3, the nozzle cover 230 includes a cover body 231 and a flow straightening plate 232. The cover body 231 has a roughly box shape with an opening on the lower side to accommodate the reactive gas nozzle 31. The flow straightening plate 232 is a plate-like body connected to the upstream side and downstream side of the rotation direction of the turntable 2 at the opening end on the lower side of the cover body 231. The side wall surface of the cover body 231 on the rotation center side of the turntable 2 extends toward the turntable 2 so as to face the tip of the reactive gas nozzle 31. In addition, the side wall surface of the cover body 231 on the outer edge side of the turntable 2 is cut out so as not to interfere with the reactive gas nozzle 31. The nozzle cover 230 is not essential and may be provided as necessary.

[0037] [Example of shower head design device configuration] 4 is a diagram showing an example of a hardware configuration of a showerhead design apparatus 50. The showerhead design apparatus 50 has a typical computer configuration. The showerhead design apparatus 50 has a CPU (Central Processing Unit) 501, a ROM (Read Only Memory) 502, and a RAM (Random Access Memory) 503.

[0038] Showerhead design apparatus 50 also has auxiliary storage device 504, operation device 505, display device 506, I / F (Interface) device 507, and drive device 508. The hardware components of showerhead design apparatus 50 are connected to each other via bus 509.

[0039] The CPU 501 executes various programs installed in the auxiliary storage device 504 .

[0040] The ROM 502 is a non-volatile memory and functions as a main storage device. The ROM 502 stores various programs and data necessary for the CPU 501 to execute various programs installed in the auxiliary storage device 504.

[0041] The RAM 503 is a volatile memory such as a dynamic random access memory (DRAM) or a static random access memory (SRAM), and functions as a main storage device. The RAM 503 provides a working area in which various programs installed in the auxiliary storage device 504 are expanded when the CPU 501 executes them.

[0042] The auxiliary storage device 504 is a non-volatile large-capacity storage device that stores various programs. The auxiliary storage device 504 may be a non-volatile large-capacity storage medium such as a hard disk drive (HDD) or a solid state drive (SSD). The various programs include a program that optimizes the arrangement of gas holes in the showerhead 36 (hereinafter simply referred to as hole layout).

[0043] The operation device 505 is an input device used when an administrator inputs various instructions to the showerhead design device 50. The display device 506 is a display device that displays internal information of the showerhead design device 50 and information obtained from the outside.

[0044] The I / F device 507 is a connection device for connecting to a communication line and communicating with the group management controller 40 .

[0045] The drive device 508 is a device for setting a recording medium. The recording medium includes media that record information optically, electrically, or magnetically, such as a CD-ROM, a flexible disk, a magneto-optical disk, etc. The recording medium may also include semiconductor memories that record information electrically, such as ROMs, flash memories, etc.

[0046] The various programs to be installed in the auxiliary storage device 504 are installed, for example, by setting a distributed recording medium in the drive device 508 and reading out the various programs recorded on the recording medium by the drive device 508. Alternatively, the various programs to be installed in the auxiliary storage device 504 may be installed by downloading them from a predetermined server.

[0047] The showerhead design device 50 may also have a GPU (Graphics Processing Unit). A GPU is a processor used for performing graphic processing, and is capable of performing a large amount of parallel calculations, enabling high-speed calculations required for machine learning.

[0048] [Overall work flow] Incidentally, a showerhead has hundreds to thousands of gas outlet holes, and the hole layout of the showerhead 36 can affect the gas concentration distribution and gas flow rate distribution on the wafer, which can in turn affect the film quality.

[0049] Adjustments to the hole layout of the showerhead 36 were often made based on the experience of experienced process engineers. However, the number of holes in a showerhead ranges from hundreds to thousands, and it was difficult even for experienced engineers to accurately determine the number and positions of holes to be adjusted while understanding the effect of each hole on the film formation results. In addition, adjusting the hole layout required preparing multiple showerheads in advance or opening a jig to block the holes, both of which required a lot of man-hours. Therefore, in layout adjustments, it was necessary not only to aim for good film quality, but also to reduce the number of adjustments.

[0050] Therefore, the overall flow of the design method of the showerhead 36 of the present disclosure, which reduces the number of adjustments (man-hours) and enables adjustment of the optimal hole layout, will be described with reference to Figure 5. Figure 5 is an example of a flowchart illustrating the overall flow of the design method of the showerhead 36.

[0051] The developer manufactures an adjustment showerhead (S101). An adjustment showerhead is a showerhead with test gas holes (hereafter simply called holes) to realize various hole layouts (see Figure 6).

[0052] Next, the developer mounts shower heads for adjusting a plurality of different hole layouts on the semiconductor manufacturing equipment 1, and actually performs film formation (S102). The film formation results are used for a data set, which will be described later. In this disclosure, for example, seven hole layouts are actually used for film formation, but the number of types of hole layouts may be as small as one. However, it is considered that a certain number of types of hole layouts for actually forming a film make it easier to obtain a hole layout with good film formation quality (film quality). Film quality is an attribute of the film formation used in the evaluation of the film formation, such as the uniformity of the film thickness distribution, the reflectance of the film, or the density of the film formation.

[0053] In this disclosure, the design device 50 determines the hole layout in two steps. Step 1 is a step of generating a model that uses the film formation results to output an evaluation value of the film formation (film thickness distribution may also be output) in an arbitrary hole layout. The evaluation value is a numerical representation of the uniformity of the film thickness distribution. Step 2 is a step of outputting the evaluation value while using this model in an optimization method such as tree structure search, and optimizing the hole layout.

[0054] Therefore, first, in step 1, the design device 50 generates a model using the film formation results (S103).

[0055] Next, in step 2, the design device 50 uses the model to output evaluation values ​​and optimize the hole layout (S104).

[0056] The developer reproduces the optimized hole layout in an adjustment showerhead or an actual showerhead, and performs film formation in semiconductor manufacturing equipment 1 (S105).

[0057] The developer compares the predicted (expected) film thickness distribution with the actual film thickness distribution and determines whether the difference between the predicted and actual film thickness distributions is within an allowable range, i.e., whether the difference between the predicted film thickness distribution and the actual film thickness distribution is within a preset allowable range (S106).

[0058] If the difference between the predicted film thickness distribution and the actual film thickness distribution is not within the allowable range, a shower head for adjusting the multiple hole layout based on the optimized hole layout is attached to semiconductor manufacturing equipment 1 and film formation is actually performed in order to generate a model again (S107). Thereafter, steps are executed again from step S103. The shower head for adjusting the multiple hole layout based on the optimized hole layout is one in which the open / closed state of one or more holes is changed based on the hole layout optimized in S105.

[0059] If the difference between the predicted film thickness distribution and the actual film thickness distribution is within an allowable range, the developer decides to adopt the optimized hole layout for the actual device (S108).

[0060] In the following, steps 1 and 2 (steps S103 and S104) described with reference to FIG. 5 will be mainly described in detail.

[0061] [Example of a shower head] FIG. 6 shows a plan view of the showerhead 36. FIG. 6 shows the holes 61 of the showerhead 36 as viewed from a plane parallel to the wafer W. A large number of holes 61 are arranged in the showerhead 36. Although the number of holes 61 is small in FIG. 6(a) for convenience of drawing, in reality, hundreds to thousands of holes are arranged. The entire length of the wafer W is covered by this showerhead 36. Therefore, when the diameter of the wafer W is 300 mm, the length of the showerhead 36 in the longitudinal direction is also about 300 mm. The length in the lateral direction may be determined as appropriate.

[0062] In FIG. 6(a), the size and shape of the holes 61 in the shower head 36 are one type, but as shown in FIG. 6(b), the size and shape of the holes 61 may vary. In FIG. 6(b), the holes from the top to the fourth row are round holes, and the size of the holes 61 is designed such that the first row < the second row = the third row < the fourth row. The hole in the fifth row is a square hole 62. The square hole 62 may be present in any row, or the round holes 61 and the square holes 62 may be mixed in one row. The size of the square holes 62 may differ from row to row. The shape of the holes 62 may be a triangle, a diamond, or the like. Therefore, the shape of the holes 61 can also be optimized in this disclosure.

[0063] In the design method of the present disclosure, a developer realizes multiple hole layouts by selectively blocking the holes 61, 62 of the adjustment shower head 36 using tape, a jig, an opening and closing mechanism, etc. The hole layout is the arrangement of the holes.

[0064] [About the function] 7 is an example of a functional block diagram explaining the functions of the design device 50 of the shower head 36 by dividing them into blocks. The design device 50 of the shower head 36 is an example of a gas hole design device. The design device 50 of the shower head 36 has a kernel shape determination unit 51, a model generation unit 52, a film thickness estimation unit 53, and a gas hole optimization unit 54. These functions of the design device 50 of the shower head 36 are functions or means realized by the CPU 501 executing a program loaded from the auxiliary storage device 504 to the RAM 503.

[0065] The kernel shape determination unit 51 and the model generation unit 52 are mainly used in the above step 1, and the film thickness estimation unit 53 and the gas hole optimization unit 54 are mainly used in step 2.

[0066] Although details will be described later, the kernel shape determination unit 51 determines the kernel (the effect of the holes in the shower head 36 on the film thickness) by a method such as black box optimization. The model generation unit 52 uses this kernel to generate a model that outputs an evaluation value when the film thickness is generated. The evaluation value is, for example, a numerical representation of the uniformity of the film thickness. The model has a function of outputting the film thickness distribution.

[0067] The film thickness estimation unit 53 uses the generated model to output an evaluation value of the film thickness when a film is formed with an arbitrary hole layout. The gas hole optimization unit 54 optimizes the hole layout of the shower head 36 using an optimization method (search algorithm) such as a tree structure search based on the evaluation value of the film thickness output by the film thickness estimation unit 53. Details will be described later.

[0068] [About the kernel] First, the kernel will be described with reference to Fig. 8. Fig. 8 is a diagram for explaining a kernel 63 and the shape of the kernel 63. The kernel 63 is a one-dimensional shape that represents the effect that one (or multiple) hole in the showerhead 36 has on the film thickness. In other words, it can be said to be a film thickness model of one hole.

[0069] The X-axis in Fig. 8 corresponds to the position in the longitudinal direction of the showerhead 36, with X=0 corresponding to the center of the hole. The Y-axis in Fig. 8 corresponds to the film thickness. Assuming that the showerhead 36 is attached to the turntable 2, the longitudinal direction of the showerhead 36 is the radial direction of the turntable 2 (see Fig. 9(a)). The kernel 63 has a larger (thicker) film thickness toward the center of the hole and a smaller (thinner) film thickness away from the center of the hole.

[0070] However, it is difficult to estimate by experience alone how each hole 61 of the shower head 36 affects the film thickness. Therefore, in the present disclosure, the kernel shape determination unit 51 determines the shape of the kernel 63 for each hole 61 of the shower head 36 by a method called black box optimization.

[0071] The standard form of the kernel 63 may be, for example, a normal distribution. Three kernels 63 with different shapes are shown in FIG. 8. κ (kappa) in FIG. 8, which will be described later, is one of the internal parameters that determine the shape of the kernel 63, and correlates with the size in the Y-axis direction (film thickness). σ is one of the internal parameters that determine the shape of the kernel 63, and correlates with the spread in the X-axis direction. The internal parameters also include skewness λ (a statistic that indicates how much the shape is distorted from a normal distribution). The shape of the kernel 63 is not limited to the three shown in the figure, and may be any shape.

[0072] [About the film thickness to be evaluated] FIG. 9 is a diagram for explaining the location of the film thickness to be evaluated on the wafer W. FIG. 9(a) is a simplified diagram of FIG. 2, and FIG. 9(b) is an enlarged view of one wafer W placed on the turntable 2. The film thickness at the location indicated by the arrow 65 is the film thickness actually measured in the film formed by the actual device. Also, in the phase where the model outputs the evaluation value in step 2, it is the film thickness estimated for evaluating the film formation. Hereinafter, the location on the wafer W indicated by the arrow 65 is referred to as the Y line. When the wafer W passes under the shower head 36, the Y line is approximately parallel to the longitudinal direction of the shower head 36.

[0073] The wafer W revolves around the turntable 2, and the side of the Y line closer to the axis of rotation is called the IN side, and the side farther from the axis of rotation is called the OUT side. The end of the IN side is set to 0 mm, and positions above the Y line are indicated. If the diameter of the wafer W is 300 mm, the end of the OUT side is 300 mm.

[0074] Fig. 10 is a diagram showing a schematic diagram of film thickness estimation using a kernel 63. Fig. 10(a) shows an arbitrary hole layout for which the film thickness is to be known. Note that in Fig. 10, the arrangement of the holes 61 is shown as squares, but this does not represent the shape of the holes 61, but simply indicates the holes 61 lined up in the longitudinal and lateral directions of the shower head 36. Also, the holes 61 marked with diagonal lines indicate that they are open, and the holes 61 marked without diagonal lines indicate that they are closed.

[0075] 10(b) is a diagram showing the shapes of the kernels 63 for each row, the number of which is the same as the number of holes opened in the showerhead 36. The kernels 63 are optimized by the kernel shape determination unit 51.

[0076] Therefore, by adding up the kernels 63 for each column, the film thickness distribution in the Y line direction can be estimated. Figure 10(c) shows the film thickness distribution of the Y line estimated by adding up the kernels by the film thickness estimation unit 53. As shown in Figure 10(c), the film thickness is obtained at any position on the Y line.

[0077] In this way, by making a physical assumption of the kernel 63, the model can output the film thickness with a small amount of data compared to the number of holes 61 that the showerhead 36 has.

[0078] In the present disclosure, the function of adding up the kernels 63 and outputting the film thickness distribution and film thickness evaluation value of the Y line is called a model. The model generation unit 52 generates this model. The model generation unit 52 uses the shape of the kernel 63 determined by the kernel shape determination unit 51 to generate a model that estimates the evaluation value of the film thickness when a film is formed using the shower head 36 having an arbitrary hole layout.

[0079] The model adds a film thickness model (the shape of the kernel 63 itself) generated by the kernel 63 determined for each hole to each row of holes in the shower head 36. That is, the model estimates a film thickness distribution for the input of the hole layout, digitizes an evaluation of the estimated film thickness distribution, and outputs an evaluation value of the film thickness.

[0080] [Determining the kernel shape through black-box optimization (Step 1)] Based on the above, the optimization of the kernel will be described in detail.

[0081] Figure 11 is a schematic diagram of black-box optimization. The inputs of black-box optimization are, for example, the following two: (1) Kernel 63 objective function (an example of the first objective function) (2) Conditions for internal parameters (1) The objective function of the kernel 63 is a function for evaluating the validity of the kernel 63. The objective function of the kernel 63 outputs the validity of the kernel 63 as a numerical value. The validity is, for example, a numerical value of the difference between the film thickness distribution obtained by the kernel and the film thickness of the data set.

[0082] In addition, the objective function of Kernel 63 uses the "dataset" and "model framework information." A data set is a set of a "hole layout" and a corresponding "film thickness distribution." As explained in Figure 5, the developer actually deposits a film using an adjustable shower head, preferably with multiple hole layouts. The developer measures the film thickness of the Y line and creates a film thickness distribution. This is the data set.

[0083] The data set may include information about the process conditions executed by the semiconductor manufacturing apparatus 1. The information about the process conditions may include at least one of the type of gas, the pressure in the processing chamber, the temperature of the stage on which the wafer W is placed, the plasma output, the film formation time, and the rotation speed of the rotary table. The model framework information is information on how to calculate the "film thickness distribution" using the "hole layout" and "kernel". The method for calculating the film thickness distribution was explained in Figure 10.

[0084] (2) The internal parameter conditions are the domain and initial values ​​of the internal parameters of the kernel 63. The internal parameters are the coefficients related to the magnitude, spread, skewness, etc. In this disclosure, the following four (a, b, σ, λ) are the internal parameters.

[0085] κ: The degree of influence of the hole on the film thickness around the hole (κ= a * r + b). κ affects the thickness (size) of the film.

[0086] a: Effect of position on film thickness (slope).

[0087] b: Position-independent effect on film thickness (intercept).

[0088] σ: width of kernel 63.

[0089] λ: Kernel 63 distortion.

[0090] Note that r is the distance from 0 mm on the Y line (it is not an internal parameter). In other words, the internal parameter a of κ is defined so that the size of the kernel 63 changes with a constant gradient with respect to the Y line. This is to reproduce the effect on the film thickness caused by the speed at which the wafer W passes under the shower head 36 at the outer side of the turntable 2 is faster.

[0091] Kernel shape determination unit 51 performs trial and error (learning) using (1) and (2) as inputs for black-box optimization to determine the internal parameters of kernel 63. In other words, the output of black-box optimization is the optimized kernel 63 (the values ​​of the internal parameters of kernel 63). The values ​​of kernel 63 are output for every hole. Alternatively, the values ​​may be output for each column in the Y line direction, ignoring differences between rows. In addition, kernel shape determination unit 51 may also output the value (validity) of the objective function of optimized kernel 63.

[0092] 12 is a flowchart showing an example of a procedure in which the kernel shape determination unit 51 uses the black-box optimization method to optimize the shape of the kernel 63. The kernel shape determination unit 51 optimizes the shape of the kernel 63 for each of one or more holes in the longitudinal direction of the showerhead 36.

[0093] First, the kernel shape determination unit 51 calculates the film thickness by adding up the influence of each kernel 63 on the film thickness in the hole layout of the data set based on the framework information of the model (S11). The kernel shape determination unit 51 performs the same process on all the obtained data sets.

[0094] Next, the kernel shape determination unit 51 inputs the film thickness of the hole layout obtained in the data set and the calculated film thickness into the objective function of the kernel 63 to determine the validity of the kernel 63 (S12).

[0095] The kernel shape determination unit 51 judges whether the validity of all the data sets is good, or whether the improvement in validity when the internal parameters are changed is equal to or smaller than a certain value (S13), or ends the process when steps S11 and S12 have been repeated a certain number of times.

[0096] If the determination in step S13 is No, kernel shape determination unit 51 gradually changes the parameters of kernel 63 within the range of the domain included in condition (2) of the internal parameters (S14). A method known as a parameter change method is Bayesian optimization.

[0097] Bayesian optimization selects the next search point (in this disclosure, the internal parameter) by optimizing a proxy function called an acquisition function instead of an objective function that requires a large computational cost to evaluate. A Gaussian process is often used to construct this acquisition function. A description of Bayesian optimization is omitted here.

[0098] If the determination in step S3 is Yes, kernel shape determination section 51 sets the finally obtained internal parameter values ​​as the shape of kernel 63 optimized.

[0099] FIG. 13 is an example of the effect of the internal parameters on the objective function of the kernel 63. Of the four internal parameters, FIG. 13 shows the effect of b and σ on the objective function of the kernel. The contour plot in FIG. 13 represents the magnitude of the difference (sum of squares) between the film thickness distribution obtained by the kernel and the film thickness of the data set, with areas of the same color representing approximately the same magnitude of difference. From FIG. 13, it can be said that b and σ corresponding to the lowest contour area 66 are the optimal internal parameters. For convenience of drawing, FIG. 13 is in black and white, but the degree of validity is indicated by color.

[0100] 14 is a diagram showing the shape of the kernel 63 obtained after adjustment in correspondence with the position on the Y line of the shower head 36. In FIG. 14, the shapes of the kernel 63 for three holes 61 at positions of 0 mm, 150 mm, and 300 mm are shown, but in reality, the shapes of the kernel 63 are obtained for the number of holes 61.

[0101] As shown in FIG. 14, the internal parameters are adjusted so that the influence of kernel 63 is smaller closer to the 300 mm side of the Y line (towards the OUT side of the turntable 2). This means that the gas concentration becomes thinner toward the outside of the turntable 2, which is physically appropriate. Also, depending on how the objective function of kernel 63 is selected, when the internal parameters of kernel 63 are optimized, the values ​​of some kernels may become negative. In order to reproduce the film thickness of the data set well, it may be difficult to explain such physical appropriateness.

[0102] Fig. 15 is a diagram for explaining the accuracy of the model. In Fig. 15, the horizontal axis indicates the position of the Y line, and the vertical axis indicates the film thickness. Each graph shown in Fig. 15 is as follows. Graph 1 shows the actual film thicknesses that were used in the data set. Figure 15 shows one of the seven film thicknesses. Graph 2 shows the actual film thickness measured based on a certain hole layout. Graph 3 shows the estimated film thickness output by the model when the same hole layout as in Graph 2 was input into the model.

[0103] Comparing graphs 2 and 3, we can see that there is no significant difference between the estimated values ​​and the actual measured values, and that the kernel shape generated by black-box optimization is effective.

[0104] [Determining hole layout by tree structure search (Step 2)] As described above, a model that outputs a film thickness distribution using the optimized kernel 63 has been generated, so that the gas hole optimization unit 54 can determine an optimal hole layout.

[0105] 16 is a diagram for explaining the determination of a hole layout by a tree structure search. The inputs for the tree structure search are, for example, the following two: (1) Hole layout objective function (an example of a second objective function) (2) Hole layout conditions (1) The objective function of the hole layout is a function for evaluating the film thickness distribution. The objective function of the hole layout uses a model. This model is the model generated by the model generation unit 52 in step 1. The evaluation value may be, for example, the standard deviation of the film thickness at positions spaced at regular intervals on the Y line.

[0106] (2) The hole layout conditions are the domain and initial value of the hole layout. The domain is, for example, how many holes may be drilled in the shower head 36. The initial value is the hole layout in the initial state when the tree structure search is performed.

[0107] The gas hole optimization unit 54 uses (1) and (2) as inputs for tree structure search, and performs trial and error to determine the hole layout. In other words, the output of the tree structure search is an optimized hole layout. The gas hole optimization unit 54 may also output the value of the objective function of the optimized hole layout (e.g., an evaluation value output by a model, etc.).

[0108] Fig. 17 is an example of a flow chart showing a procedure for optimizing the hole layout by the gas hole optimization unit 54. Fig. 18 is a diagram showing a schematic diagram of a hole layout optimization method, and Fig. 18 shows the corresponding step numbers in Fig. 17. The tree structure search method shown in Figs. 17 and 18 is called a Monte Carlo tree search.

[0109] First, the gas hole optimization unit 54 sets a reference pattern of the hole layout that is the starting point of optimization (S21). The reference pattern may be determined based on some knowledge, may be a pattern in which open holes are uniformly arranged, or may be a pattern in which the open and closed states are determined randomly. However, the same restrictions as those of the domain of definition of the hole layout conditions (2) are taken into consideration.

[0110] Next, the gas hole optimization unit 54 assumes a pattern in which the open / close state of only one hole is changed from the reference pattern for all holes, and the film thickness estimation unit 53 calculates an evaluation value using the model (S22). That is, the film thickness estimation unit 53 estimates the film thickness using the model generated by the model generation unit 52, and further calculates an evaluation value. As shown in FIG. 18, an evaluation value is obtained for each of the open / closed patterns 110, 111, and 112, which are different from the current pattern (here, the reference pattern) by only one open / close state. The pattern 110 is, for example, a pattern in which a certain hole is opened from the reference pattern (this hole was closed in the reference pattern), the pattern 111 is, for example, a pattern in which another certain hole is closed from the reference pattern (this hole was open in the reference pattern), and the pattern 112 is, for example, a pattern in which yet another certain hole is opened from the reference pattern (this hole was closed in the reference pattern).

[0111] Next, the gas hole optimization unit 54 selects N patterns to continue the search based on the calculated evaluation value (S23). The gas hole optimization unit 54 selects N patterns, for example, in descending order of evaluation value. In FIG. 18, the shaded circles represent the selected patterns.

[0112] The gas hole optimization unit 54 repeats steps S22 and S23 based on the patterns left in step S23. This deepens the pattern search depth by one for each cycle. Also, N patterns are selected for each layer. The gas hole optimization unit 54 repeats steps S22 and S23 (S24) until a predetermined upper limit of the depth is reached. For example, although a depth of 3 is shown in FIG. 18, this is merely for convenience of drawing, and the upper limit of the depth may be changed depending on the calculation speed of the design device 50, the allowable time, etc. In FIG. 18, the patterns selected in each layer in step S23 are shown by diagonal lines.

[0113] When the predetermined upper limit of the depth is reached (Yes in S24), the gas hole optimization unit 54 propagates the highest evaluation value for each layer from the pattern at the deepest position to the reference pattern (S25). A supplementary explanation will be given with reference to Fig. 18. First, the left branch 67 in Fig. 18 will be described.

[0114] "Evaluation value propagated from depth 4 1.5 < Evaluation value propagated from depth 4 1.7" → Propagate 1.7 to depth 2 "Evaluation value propagated from depth 3 1.1 < Evaluation value propagated from depth 3 1.7" → Propagate 1.7 to depth 1 Next, the right branch 68 in FIG. 18 will be described.

[0115] "The evaluation value propagated from depth 4 is 1.2 < the evaluation value propagated from depth 4 is 1.4" → Propagate 1.4 to depth 2 "Evaluation value propagated from depth 3 1.4 < Evaluation value propagated from depth 3 1.5" → Propagate 1.5 to depth 1 As a result, the two evaluation values ​​1.7 and 1.5 at depth 1 are determined, and 1.7 is propagated to the reference pattern.

[0116] The gas hole optimization unit 54 selects one of the changes that leads to the pattern with the highest evaluation value, and updates the reference pattern (S26). In the example of Fig. 18, the pattern 110 that propagated 1.7 is the next reference pattern.

[0117] The gas hole optimization unit 54 repeats steps S22 to S26, and ends the search when the reference pattern has been updated a certain number of times (S27). Therefore, the gas hole optimization unit 54 can repeatedly select the pattern with the highest evaluation value in the tree structure, and determine the optimal hole layout.

[0118] FIG. 19 shows an example of an optimized hole layout. FIG. 19 shows a portion of a showerhead 36. FIG. 19(a) shows the hole layout before optimization, and FIG. 19(b) shows the hole layout after optimization. The hole layout before optimization has a relatively biased pattern between open and closed states, but FIG. 19(b) shows a more complex pattern. In this way, according to the showerhead design method of the present disclosure, a hole layout with a pattern that is difficult for humans to adjust can be obtained.

[0119] Furthermore, the optimal hole layout may vary depending on factors such as gas flow rate, susceptor temperature, plasma output (e.g., the applied voltage of a plasma source not shown), gas type, and film formation recipe, and it is preferable to optimize the hole layout for each of these factors.

[0120] FIG. 20 is an example of a diagram showing a deposition result calculated by the design device 50 using an optimized hole layout. The deposition result calculated by the computer is a film thickness distribution output by a model based on the optimized hole layout. The horizontal axis of FIG. 20 is the position on the Y line, and the vertical axis is the film thickness. The contents of each graph are as follows. Graph 101 is an estimate of deposition results for the optimized hole layout. The dotted line graph 102 is the film thickness used in the data set.

[0121] Among the seven deposition results (dotted lines) before optimization, the deposition result with the least variation was compared with the one with the least variation in graph 101 (estimated deposition result). It was confirmed that the standard deviation was less than half. In this way, it is possible to obtain an optimal hole layout on a computer that improves the film thickness distribution.

[0122] Fig. 21 shows an actual result of film formation when semiconductor manufacturing equipment 1 forms a film using shower head 36 with an optimized hole layout. Note that in Fig. 21, the hole layout is optimized only for about 30% of all holes in shower head 36 on the side closer to 0 mm of the Y line.

[0123] The horizontal axis of Figure 21 is the position on the Y line, and the vertical axis is the film thickness. The contents of each graph are as follows. Graph 103 shows the actual deposition results based on the hole layout before optimization. Graph 104 shows the actual deposition results with the optimized hole layout.

[0124] Comparing graphs 103 and 104, the film thickness variation is improved when the position on the Y line is about 70 mm or less. Therefore, by optimizing the hole layout, the film thickness distribution in the actual device could be improved.

[0125] [Major Effects] As described above, the design method of the shower head of the semiconductor manufacturing apparatus 1 disclosed herein utilizes a statistical model using actual film formation data, and therefore can take into account physical phenomena that cannot be reproduced by general simulations, enabling the hole layout to be determined with high precision.

[0126] In addition, by using a model designed based on physical phenomena such as gas diffusion and adsorption, the number of experimental samples required for kernel optimization can be reduced.

[0127] In addition, by using an optimization algorithm to determine the kernel shape and hole layout, a lot of trial and error can be done on a computer, making it easier to find a good hole layout compared to when optimizing by hand.

[0128] Furthermore, by combining the present disclosure with a conventional method for optimizing controllable elements such as gas flow rate, susceptor temperature, or plasma output (e.g., applied voltage of a plasma source), optimization can be performed in a wider search space than ever before. For example, these elements can be included in the search space when optimizing the kernel shape. This is expected to make it easier to find a combination of element values ​​and hole layouts that will produce better film formation results.

[0129] [others] The embodiments may be modified and improved in various ways without departing from the spirit and scope of the appended claims. The matters described in the above embodiments may be configured in other ways as long as they are not inconsistent, and may be combined as long as they are not inconsistent.

[0130] For example, in this disclosure, the hole layout is optimized primarily to achieve uniform film thickness, but the hole layout may also be optimized to obtain a film with preferred film quality, such as film reflectance or film density.

[0131] In addition, in the present disclosure, the hole layout of the showerhead 36 through which the reaction gas nozzle 31 supplies gas is optimized, but optimization is possible if there is a showerhead through which the separation gas nozzle 42 and the reaction gas nozzle 32 supply gas.

[0132] In addition, in this disclosure, a black-box optimization method is described in step S1, and a tree structure search method is described in step S2. However, in optimizing the kernel shape and hole layout, an optimization method derived from these or a different optimization method may be used.

[0133] The semiconductor manufacturing apparatus 1 of the present disclosure can be applied to any type of apparatus, such as an atomic layer deposition (ALD) apparatus, a capacitively coupled plasma (CCP), an inductively coupled plasma (ICP), a radial line slot antenna (RLSA), an electron cyclotron resonance plasma (ECR), or a helicon wave plasma (HWP). In addition, the present disclosure can be applied to any semiconductor manufacturing apparatus that forms a film using a shower head. [Explanation of symbols]

[0134] 1. Semiconductor manufacturing equipment 2 Rotating table 36 Shower Head 51 Kernel shape determination unit 52 Model Generation Unit 53 Film Thickness Estimation Section 54 Gas hole optimization section

Claims

1. A method for designing a gas hole for supplying gas into a processing vessel of a semiconductor manufacturing apparatus, comprising: generating a model for evaluating film quality based on a plurality of data sets in which the arrangement of the gas holes corresponds to the results of film formation that is actually performed by supplying gas from the gas holes in the arrangement of the gas holes; calculating an evaluation value of a film quality for the arrangement of the gas holes using the model, and optimizing the arrangement of the gas holes; A method for designing a gas hole having the following structure:

2. In the step of generating the model, a film thickness model formed by the gas supplied from the gas holes is optimized for at least one or more holes in a longitudinal direction of the gas holes; The method for designing a gas hole according to claim 1 , wherein the model outputs an evaluation value of the film quality using the optimized film thickness model.

3. The method for designing gas holes according to claim 2 , wherein the model outputs a film thickness distribution in the longitudinal direction as information about the film quality by adding up the optimized film thickness models for each of the gas holes.

4. 4. The method for designing a gas hole according to claim 2, wherein in the step of generating the model, the film thickness model is optimized by black-box optimization.

5. In the black-box optimization, a first objective function for evaluating the film thickness model is input, 5. The method for designing gas holes according to claim 4, wherein the first objective function takes as inputs the arrangement of the gas holes, a film thickness when a film is formed through the gas holes in the arrangement, and a calculation method for a film thickness distribution using the film thickness model, and outputs an optimized film thickness model.

6. 6. The gas hole design method according to claim 5, wherein the first objective function further inputs process conditions including at least one of the type of gas, the pressure in the processing vessel, the temperature of the stage on which the substrate is placed, the plasma output, the film formation time, and the rotation speed of the rotary table.

7. the film thickness model has a shape in which the film thickness is greatest at the center of the gas hole and decreases with increasing distance from the center of the gas hole, The method for designing a gas hole according to claim 6 , wherein the first objective function outputs at least one of a magnitude, a spread, and a skewness of the film thickness model.

8. The semiconductor manufacturing apparatus has a rotary table that revolves a wafer on which a film is formed, The black box optimization further includes inputting a domain and an initial value of the magnitude, spread, or skewness of the film thickness model; The method for designing a gas hole according to claim 7 , wherein it is defined that in the domain, the size of the film thickness model changes at a constant gradient with respect to the radial direction of the rotary table.

9. 7. The gas hole design method according to claim 6, wherein in the step of optimizing the arrangement of the gas holes, the arrangement of the gas holes is changed using a tree structure search technique, and the calculation of the evaluation value of the film quality output by the model is repeated to optimize the arrangement of the gas holes.

10. In the tree structure search method, a second objective function for evaluating the arrangement of the gas holes is input, the second objective function uses the optimized film thickness model to output an evaluation value of the film quality output by the model; The gas hole design method according to claim 9, wherein in the step of optimizing the gas hole arrangement, the optimized gas hole arrangement is determined by searching for a gas hole pattern having a higher evaluation value.

11. The method for designing gas holes according to any one of claims 1 to 10, wherein the design method determines a shape of the gas holes in a shower head used in the semiconductor manufacturing apparatus.

12. In the step of optimizing the arrangement of the gas holes, (a) setting a reference pattern for the arrangement of the gas holes as a starting point for optimization; (b) a step of assuming a pattern in which the open / closed state of only one hole is changed from a reference pattern for all holes, and calculating an evaluation value of the film quality using the model; (c) selecting N patterns for continuing the search based on the film quality evaluation value; The steps (b) and (c) are repeated until the depth of the tree structure search reaches a predetermined upper limit, deepening the hierarchy of the tree structure. (d) A step of propagating the highest evaluation value for each layer from the deepest layer to the reference pattern. (e) selecting one of the modifications that results in the highest scoring pattern and updating the reference pattern; The method for designing gas holes according to claim 10 or 11, wherein the steps (a) to (e) are repeated to determine an optimal arrangement of the gas holes.

13. A design device for a gas hole for supplying gas into a processing vessel of a semiconductor manufacturing device, a model generating unit that generates a model for evaluating film quality based on a plurality of data sets that correspond to the arrangement of the gas holes and the results of film formation that is actually performed by supplying gas from the gas holes in the arrangement of the gas holes; a gas hole optimization unit that uses the model to calculate an evaluation value of a film quality in the arrangement of the gas holes and optimizes the arrangement of the gas holes; A gas hole design device having the above structure.

Citation Information

Patent Citations

  • Electrostatic chuck with improved wafer temperature uniformity

    JP1996055905A

  • Processing apparatus and processing method of object to be processed

    JP2005101626A

  • Gas feeder and substrate treatment device

    JP2008038200A

  • Epitaxial growth device and method of manufacturing epitaxial wafer

    JP2010034474A

  • Film deposition device, film deposition method, and storage medium

    JP2012084598A