Substrate processing method and substrate processing apparatus

Abstract

To suppress surface roughness after etching of a silicon film forming a recess, and control the shape of the recess.SOLUTION: A first process of supplying process gas including halogen-containing gas and basic gas to a substrate on which a recess, in which sidewalls are formed of a silicon film and a depth wall is formed of a germanium-containing film, is formed, and altering the surface of the silicon film to generate a reaction product; a second process of removing the reaction product to widen the depth of the recess; a process of performing a cycle consisting of the first process and the second process a plurality of times; and a process of performing the first process of the cycle performed first under a first processing condition, and performing the first process of the cycle performed later under a second processing condition different from the first processing condition are executed.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present disclosure relates to a substrate processing method and a substrate processing apparatus. [Background technology]

[0002] In manufacturing a semiconductor device, there are cases where the Si film of the Si film and the SiGe film formed on the surface of a semiconductor wafer (hereinafter referred to as a wafer) serving as a substrate is selectively etched. For example, Patent Document 1 describes that the selective etching is performed by using F2 gas and NH3 gas as etching gases and setting the ratio of NH3 gas to the etching gas to a predetermined value. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 6426489 Summary of the Invention [Problem to be solved by the invention]

[0004] The present disclosure provides a technique for suppressing surface roughness after etching of a silicon film in which recesses are formed, and for controlling the shape of the recesses. [Means for solving the problem]

[0005] The substrate processing method of the present disclosure includes a first step of supplying a processing gas containing a halogen-containing gas and a basic gas to a substrate having a recess whose side wall is formed of a silicon film and whose inner wall is formed of a germanium-containing film, thereby modifying a surface of the silicon film and generating a reaction product; a second step of removing the reaction product to widen the width of the recess; performing a cycle consisting of the first step and the second step a plurality of times; performing the first step of the first cycle under first processing conditions and the first step of the second cycle under second processing conditions different from the first processing conditions; Includes. Effect of the Invention

[0006] The present disclosure makes it possible to suppress the roughness of the surface of a silicon film in which recesses are formed after etching, and to control the shape of the recesses. [Brief description of the drawings]

[0007] [Figure 1] FIG. 2 is a cross-sectional side view of a wafer undergoing processing in accordance with one embodiment of the present disclosure. [Diagram 2] 5A and 5B are explanatory views showing changes in the recesses before and after etching. [Diagram 3] FIG. 2 is a cross-sectional side view of the wafer after etching. [Figure 4] 4 is an explanatory diagram showing a state of change in the Si film forming the recess; FIG. [Diagram 5] 4 is an explanatory diagram showing a state of change in the Si film forming the recess; FIG. [Figure 6] 4 is an explanatory diagram showing a state of change in the Si film forming the recess; FIG. [Figure 7] 5 is an explanatory diagram showing how the recess changes during etching. FIG. [Figure 8] FIG. 4 is an explanatory diagram showing how the recess has a rectangular shape. [Figure 9] FIG. 4 is an explanatory diagram showing how the recess has a rounded shape. [Figure 10] FIG. 2 is a flow chart showing the processing of the present embodiment. [Figure 11] FIG. 2 is a plan view showing an embodiment of a substrate processing apparatus for performing the above-mentioned processing. [Figure 12] 2 is a vertical sectional side view showing an example of a processing module provided in the substrate processing apparatus; FIG. [Figure 13] FIG. 11 is a graph showing the results of an evaluation test. [Figure 14]FIG. 11 is a graph showing the results of an evaluation test. [Figure 15] FIG. 11 is a graph showing the results of an evaluation test. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] An embodiment of the substrate processing method of the present disclosure will be described. FIG. 1 is a vertical cross-sectional side view of the surface of a wafer W, which is a substrate, showing a state before the etching process according to this embodiment is performed. An underlayer film 11 is formed on the wafer W. A SiGe (silicon germanium) film 12 and a Si (silicon) film 13 are alternately and repeatedly laminated on the underlayer film 11. With such a repeated structure, each of the SiGe film 12 and the Si film 13 is formed in multiple stages in the thickness direction of the wafer W. The thickness direction of the wafer W will be described as the vertical direction. The Si film 13 is formed, for example, by epitaxial growth.

[0009] A mask film 14 is laminated on the top Si film 13 of the above-mentioned repeated structure, and prevents etching of the Si film 13 from above. A recess extending vertically from the mask film 14 to the upper side of the lower film 11 is formed. This recess is a groove 15 extending in the front-to-back direction of the paper surface of FIG. 1. The groove 15 opens on the upper surface of the mask film 14. The side walls of the groove 15 are formed by the mask film 14, the Si film 13, and the surface layer parts of the lower film 11, and each SiGe film 12 is located away from the side walls of the groove 15. As a result, in a vertical cross-sectional view, multiple recesses facing the left side are formed at intervals on the left wall of the groove 15, and multiple recesses facing the right side are formed at intervals on the right wall of the groove 15. These recesses facing the left and right sides, respectively, are called recesses 16. Therefore, the recesses 16 are formed at multiple levels in the thickness direction of the wafer W, and are formed so as to open in directions intersecting the thickness direction.

[0010] The grooves 15 and the recesses 16 are formed to form a symmetrical structure, and the vertical widths of the recesses 16 are the same or approximately the same. Hereinafter, the upper recess 16, the lower recess 16, and the recess 16 between the upper and lower sides of the Si film 13 may be referred to as the top recess 16, the bottom recess 16, and the middle recess 16, respectively. As described above, the height of the stack formed by the SiGe film 12 and the Si film 13 (= the height from the upper surface of the lower film 11 to the lower end of the mask film 14) is, for example, 4 μm.

[0011] FIG. 2 shows an enlarged view of one of the recesses 16. The left side shows the state before etching corresponding to FIG. 1 above, and the right side shows the state after etching, which will be described later. Because of the film structure described above, the side walls of the recesses 16 are composed of the Si film 13, and the back wall (bottom wall) is composed of the SiGe film 12. Hereinafter, for convenience of explanation, the walls forming each recess 16 may be referred to as the upper wall 21, the lower wall 22, and the lateral wall 23. The upper wall 21 and the lower wall 22 are composed of the Si film 13, and the upper wall 21 of any one recess 16 is the lower wall 22 of the recess 16 above that one recess 16. The lateral wall 23 is composed of the SiGe film 12 before etching. In the recess 16 before etching, the upper wall 21 and the lower wall 22 are perpendicular or approximately perpendicular to the lateral wall 23, and in a vertical cross-sectional view, each of the upper wall 21 and the lower wall 22 forms a corner with the lateral wall 23, the angle being 90° or approximately 90°.

[0012] When gas is supplied to the wafer W, the gas is introduced into the groove 15 and further into the recesses 16 of each step. In this embodiment, the process gas is introduced into each recess 16 from the surface of the wafer W in this manner, so that the lower surface side of the upper wall 21 and the upper surface side of the lower wall 22 are etched. That is, the thickness of the Si film 13 forming the side wall of the recess 16 is reduced, and the recess 16 is widened in the vertical direction. Due to this widening, the side wall 23 of the recess 16 after etching is composed of the SiGe film 12 and the Si film 13. Note that FIG. 2 is a schematic diagram showing a desirable state in which the shape of the recess 16 after etching is a rectangular shape, which will be described later, and the roughness of the lower surface of the upper wall 21 and the upper surface of the lower wall 22 is suppressed.

[0013] Specifically, the etching process is as follows: F2 (fluorine) gas, which is a halogen-containing gas, and NH3 gas, which is a basic gas, are supplied to the wafer W as processing gases, and the surface of the Si film 13 constituting the upper wall 21 and the lower wall 22 is altered to generate a reaction product. The process of altering the Si film 13 with the processing gas is called a processing gas supply process. Then, a heating process is performed in which the wafer W is heated to sublimate the reaction product and widen the recess 16. The reaction product to be sublimated is various products such as AFS, which will be described later. The vertical width of the recess 16 is made to a desired size by repeating a cycle consisting of a processing gas supply process, which is a first process, and a heating process, which is a second process. The processing gas supply process and the heating process are each performed in a state in which the wafer W is stored in a processing container in which a vacuum atmosphere of a predetermined pressure is formed.

[0014] The above process gas acts mainly on the Si film 13 out of the Si film 13 and the SiGe film 12. Therefore, etching of the Si film 13 is selective to the SiGe film 12. The above reaction product is ammonia fluorosilicate [AFS: ((NH4)SiF6]. In the following description, unless otherwise specified, the supply time of the process gas refers to the supply time of the process gas in one cycle, and is the time during which both the halogen-containing gas F2 gas and the basic gas NH3 gas are supplied to the wafer W.

[0015] It is desired that the recess 16 after etching meets the first to third requirements described below. The first requirement is that the roughness (surface roughness) of the lower surface of the upper wall 21 and the upper surface of the lower wall 22, which are the Si film 13, is relatively small. The second requirement is described below. After etching, as before etching, the upper wall 21 and the lower wall 22 and the side wall 23 form an angle of 90° or approximately 90° at the back side of the recess 16, and the formation of such an angle results in high uniformity of the vertical width between the opening side and the back side of the recess 16. For convenience of explanation, the shape of the recess 16 in which such an angle is formed and the vertical width is high is described as a rectangular shape. On the other hand, the shape of the recess 16 in which such an angle is not formed, the recess 16 is a horizontally inclined U-shape in a vertical cross section, and the vertical width at the back side is smaller than that at the opening side is described as a round shape. The second requirement is that the shape of the recess 16 is closer to a rectangular shape (higher rectangularity).

[0016] The third requirement is that the variation in the vertical width of the recesses 16 at each height be suppressed. In other words, the variation in the amount of etching is suppressed among the top, middle, and bottom recesses 16. Fig. 3 is a schematic diagram showing a state in which the wafer W in Fig. 1 has been etched to meet the above first to third requirements, and the recess 16 after etching in Fig. 2 described above shows one of the recesses 16 of the wafer W shown in Fig. 3.

[0017] The first requirement, roughness of the Si film 13, will be explained in detail. The Si film 13 reacts with the process gases F2 gas and NH3 gas as shown in the following formula 1 to generate a product (SiF4). This SiF4 reacts as shown in formula 2 to generate AFS. The roughness of the Si film 13 after etching varies depending on the process conditions, but it is considered that the amount of AFS generated during the process gas supply process is involved. Si+F2+NH3→SiF4+NH3...Formula 1 SiF4+H2+F2+2NH3→(NH4)SiF6...Formula 2

[0018] Hereinafter, differences in roughness that are believed to be caused by differences in the amount of AFS generated will be described with reference to Fig. 4 to Fig. 6. Fig. 4 to Fig. 6 are schematic diagrams showing changes in the upper surface of the lower wall 22 of the recess between when the processing gas is supplied and after the etching is completed, and the lower surface of the upper wall 21 also changes in the same manner as the upper surface of the lower wall 22. The amount of AFS generated during the processing gas supply process increases in the order of Fig. 4, Fig. 5, and Fig. 6.

[0019] 4 shows a state in which the AFS layer 31 is formed in a dotted manner on the upper surface of the lower wall 22 during the process gas supply process due to a relatively small amount of AFS generation. The portion of the upper surface covered with the AFS layer 31 is prevented from coming into contact with the process gas, so that the portion is less likely to be altered by reaction with the process gas. That is, on the upper surface of the lower wall 22, the portion where the AFS layer 31 is formed and the portion where the AFS layer 31 is not formed have different reactivity to the process gas. Due to the difference in reactivity, unevenness is formed on the surface of the lower wall 22 after etching. Since the AFS layer 31 is formed in a dotted manner as described above, the surface roughness of the upper surface of the lower wall 22 is small. That is, the width of the concave portion and the width of the convex portion in the above-mentioned unevenness are relatively small.

[0020] 5, during the process gas supply step, the AFS layer 31 is formed as a thin layer that covers the entire upper surface of the lower wall 22, thereby suppressing reaction with the process gas to the same extent in each part of the upper surface. As a result, the entire upper surface of the lower wall 22 is etched with high uniformity, and the surface after etching becomes flat or approximately flat.

[0021] 6 shows a state in which a relatively large amount of AFS is generated during the process gas supply step, causing the AFS layer 31 to cover the entire upper surface of the lower wall 22, and the thickness of the AFS layer 31 to vary at various parts of the upper surface. That is, the surface roughness of the AFS layer 31 is large. The thicker the AFS layer 31 is, the more difficult it becomes for the process gas to come into contact with the lower wall 22, and therefore the surface roughness of the lower wall 22 after etching also increases. That is, with regard to the unevenness, the width of the concave portions and the width of the convex portions become relatively large.

[0022] As shown in FIG. 4, a state where the surface roughness is relatively small is called micro-roughness, and as shown in FIG. 6, a state where the surface roughness is relatively large is called large roughness. As described above, when the amount of AFS generated shifts from the appropriate range to the smaller side, the surface roughness of the Si film 13 due to micro-roughness increases, and when it shifts to the larger side, the surface roughness of the Si film 13 due to large roughness increases. According to the above formulas 1 and 2, the amount of AFS generated is affected by the flow rate ratio of F2 gas and NH3 gas, which are processing gases supplied to the wafer W, and when the flow rate of NH3 gas is relatively large relative to the flow rate of F2 gas, the amount of AFS generated increases. Also, the amount of AFS generated increases as the supply time of the processing gas increases. Incidentally, the flow rate of NH3 gas relative to the flow rate of F2 gas in the processing gas (NH3 gas flow rate / F2 gas flow rate) may be referred to as the flow rate ratio of NH3 gas below. Therefore, a large flow rate ratio of NH3 gas means that the flow rate of NH3 gas is large relative to the flow rate of F2 gas (= the ratio of NH3 gas to the ratio of F2 gas in the processing gas is large).

[0023] Next, the shape of the recess 16 after the second etching will be described. Depending on the processing conditions, the shape of the recess 16 after etching changes between a rectangular shape and a round shape, and the process that is estimated to occur before the rectangular shape and the round shape are formed will be described below. First, with reference to FIG. 7 and FIG. 8, the process of forming the rectangular shape will be described. F2 gas 41 and NH3 gas 42, which are processing gases, enter the recess 16 (upper part of FIG. 7) and react with the lower surface of the upper wall 21 and the upper surface of the lower wall 22, which are the Si film 13, as shown in the above-mentioned formula 1, to generate SiF4. A part of this SiF4 is released into the recess 16 as a gas, while the other part reacts as shown in the above-mentioned formula 2 to form the AFS layer 31.

[0024] The F2 gas 41 and the NH3 gas 42 act on the side wall 23 in addition to the upper wall 21 and the lower wall 22, and the surface layer of the SiGe film 12 constituting the side wall 23 is altered and slightly etched, releasing GeF4 (germanium tetrafluoride) gas 43. By etching the side wall 23 in this way, a part of the Si film 13 is newly exposed in the recess 16 at the back of the recess 16, thereby forming the upper wall 21 and the lower wall 22 of the recess 16 (lower part of FIG. 7). This part is called a newly exposed part 44.

[0025] The newly exposed portion 44 is exposed to the recess 16 during the supply of the processing gas, and is exposed to the F2 gas 41 and the NH3 gas 42 for a shorter period of time than the portions of the Si film 13 that formed the upper wall 21 and the lower wall 22 before etching. Therefore, the amount of reaction with the F2 gas 41 and the NH3 gas 42 is relatively small. Also, the F2 gas 41 and the NH3 gas 42 are less likely to enter the recess 16 toward the back side. From the above, the upper wall 21 and the lower wall 22 including the newly exposed portion 44 react less with the F2 gas 41 and the NH3 gas 42 toward the back side of the recess 16, but the GeF4 gas 43 has reactivity with the Si film 13 and reacts relatively strongly with the Si film 13 at the back side of the recess 16 by being generated from the SiGe film 12. Therefore, the Si film 13 reacts with high uniformity from the opening of the recess 16 to the back side (upper part of FIG. 8). Therefore, after the reaction products generated in the process gas supply step are removed by the heating step, the recess 16 becomes rectangular (lower part of FIG. 8). The AFG layer, which is a reaction product generated in the SiGe film 12, is shown as 45, and the reaction product between the GeF4 gas 43 and the Si film 13 is shown as 46 in the figure. The AFG layer 45 and the reaction product 46 are also removed together with the AFS layer 45. The AFG layer 45 will be described later.

[0026] Next, the estimation of the process of forming the round shape will be described. When the reaction between the surface layer of the SiGe film 12 and the processing gas proceeds from the state shown in the lower part of FIG. 7, the surface layer gradually changes to the AFG [(NH4)GeF6] layer 45. As the change to the AFG layer 45 proceeds, the amount of GeF4 gas 44 released decreases. Therefore, for the upper wall 21 and the lower wall 22, the reaction amount with GeF4 gas 44 and / or F2 gas 41 and NH3 gas 42 decreases toward the back side of the recess 16 (upper part of FIG. 9). Therefore, after the reaction product is removed by the heating process, the recess 16 becomes round (lower part of FIG. 9). From the above, the longer the supply time of the processing gas or the larger the flow rate ratio of the NH3 gas, the easier it is to form the AFG layer 45 (=the easier it is to reduce the release of GeF4 gas 44), and therefore the easier it is to form the round shape.

[0027] Next, the variation in the amount of etching among the top, middle, and bottom regarding the third requirement will be explained. Since the processing gas is less likely to penetrate toward the back of the groove 15, when the supply time of the processing gas is relatively short, the upper wall 21 and the lower wall 22 of the recess 16 are more likely to come into contact with the processing gas and react with the processing gas in the order of top>middle>bottom. Therefore, the amount of etching increases in this order.

[0028] However, as shown in the evaluation test described later, it has been confirmed that the longer the supply time of the processing gas, the greater the amount of etching becomes toward the bottom side (i.e., the order of the amount of etching becomes top < middle < bottom). This is considered to be due in part to the fact that the longer the supply time of the processing gas, the thicker the AFS layer 31 formed becomes toward the top side (the opening side of the groove 15), preventing the upper wall 21 and the lower wall 22 from reacting with the processing gas. In addition, the greater the flow rate of the NH3 gas, the easier it is for the thickness of the AFS layer 31 on the top side to increase, and the greater the amount of etching tends to be toward the bottom side.

[0029] As described above, by setting the processing conditions in which the flow rate ratio of NH3 gas is relatively small and / or the supply time of the processing gas is relatively short, the recess 16 after etching tends to have micro-roughness, a rectangular shape, and a large amount of etching on the top side. Conversely, by setting the processing conditions in which the flow rate ratio of NH3 gas is relatively large and / or the supply time of the processing gas is relatively long, the recess 16 after etching tends to have large roughness, a round shape, and a large amount of etching on the bottom side. Regarding the flow rate ratio of NH3 gas and the supply time of the processing gas, the appropriate range for keeping the roughness within the allowable range, the appropriate range for forming a rectangular shape, and the appropriate range for making the etching amount between the top and bottom uniform do not necessarily match. Therefore, when the cycle consisting of the processing gas supply step and the heating step is repeatedly performed, if the processing gas supply step of each cycle is set to the same processing conditions, it is difficult to fully meet all of the above first to third requirements.

[0030] Therefore, in this embodiment, the process conditions are different between the process gas supply step performed in the first cycle and the process gas supply step performed in the second cycle. Fig. 10 shows a process flow in this embodiment. Assuming that the cycle is set to be performed Z times in total, the process gas supply step is performed under the first process condition, and then the heating step is performed. This is repeated X times in the first cycle (step S1). Then, the process gas supply step is performed under the second process condition, and then the heating step is performed. This is repeated Y times (Z times - X times) in the second cycle (step S2).

[0031] By changing the processing conditions as described above, it is possible to prevent a particular tendency from appearing strongly in the roughness, the shape of the recesses, and the amount of etching between the top, middle, and bottom, as compared with the case of processing under a single processing condition, and to meet the first to third requirements. However, as will be shown later as an evaluation test, it has been confirmed that the shape and roughness of the recesses 16 after etching are relatively more affected by the processing conditions performed in the previous cycle than by the latter cycle. In other words, when the processing gas supply process in the previous cycle is set under conditions that form a rectangular shape and micro-roughness, the recesses 16 after etching tend to have a rectangular shape and micro-roughness. Conversely, when the processing gas supply process in the previous cycle is set under conditions that form a round shape and large roughness, the recesses 16 after etching tend to have a round shape and large roughness.

[0032] Regarding roughness, even if it is formed slightly, it is acceptable as long as it falls within the tolerance. Therefore, for the purpose of forming a rectangular shape, the first processing condition in the front cycle may be a processing condition that tends to form a rectangular shape and micro-roughness, and the second processing condition in the rear cycle may be a processing condition that tends to form a round shape and large roughness compared to the first processing condition. Therefore, specifically, the flow rate ratio of NH3 gas in the first processing condition is made smaller than that in the second processing condition, or the supply time of the processing gas in the first processing condition is made shorter than that in the second processing condition. Also, in order to increase the tendency to form the rectangular shape, it is preferable to set the first processing condition in the front cycle and the second processing condition in the rear cycle in such a way, and then set, for example, the number of times the front cycle is performed (X times) ≧ the number of times the rear cycle is performed (Y times).

[0033] In addition, there are cases where the first requirement (requirement to suppress roughness) is more important than the second requirement (requirement to increase rectangularity). In that case, as described above, the effect of the first process condition in the previous cycle remains relatively large on the recess 16 after etching, so that the first process condition is selected as a condition that relatively suppresses roughness. Then, as the second process condition in the subsequent cycle, a condition that corrects the etching amount between the top, middle, and bottom and modifies the shape of the recess 16 so that the rectangularity is increased may be selected. Specifically, the flow rate ratio of NH3 gas is made larger in the first process condition than in the second process condition, or the supply time of the process gas is made larger in the first process condition than in the second process condition. In this way, even when a high suppression effect on roughness is to be obtained, it is preferable to set, for example, the number of times the previous cycle is performed (X times) ≧ the number of times the subsequent cycle is performed (Y times) in order to increase the effect.

[0034] In this embodiment, in order to achieve high rectangularity, the flow rate ratio of NH3 gas is set to be greater under the first processing conditions than under the second processing conditions, and the supply time of the processing gas is set to be shorter under the first processing conditions than under the second processing conditions, and the number of times the first cycle is performed (X times) is set to be greater than the number of times the second cycle is performed.

[0035] To describe a more specific example of the processing conditions, the flow rate ratio of NH3 gas (NH3 gas flow rate / F2 gas flow rate) is set within the range of 0.01 to 0.014 in each of the first and second processing conditions, based on the results of the evaluation test described later. Therefore, the flow rate ratio of NH3 gas in at least one of the first and second processing conditions is smaller than 0.014. In addition, the supply time of the processing gas is set within the range of, for example, 10 to 15 seconds in each of the first and second processing conditions. In this specification, the supply time of the processing gas refers to the time during which the halogen-containing gas (F2 gas in this embodiment) and the basic gas (NH3 gas in this embodiment) constituting the processing gas are simultaneously supplied to the wafer W (i.e., the time during which they are supplied into the processing vessel housing the wafer W).

[0036] A substrate processing apparatus 5, which is one embodiment of an apparatus for performing the processing described in the flow chart of Fig. 10, will be described with reference to the plan view of Fig. 11. The substrate processing apparatus 5 includes a load / unload section 51 for loading and unloading a wafer W, two load lock chambers 61 provided adjacent to the load lock chambers 51, two heat treatment modules 60 provided adjacent to the two load lock chambers 61, and two processing modules 7 provided adjacent to the two heat treatment modules 60, respectively. The processing module 7 corresponds to a first processing section, and the heat treatment module 60 corresponds to a second processing section.

[0037] The load / unload section 51 includes an atmospheric pressure transfer chamber 53 in which a first substrate transfer mechanism 52 is provided and which is kept under atmospheric pressure, and a carrier mounting table 55 on the side of the atmospheric pressure transfer chamber 53 on which a carrier 54 for storing a wafer W is mounted. In the drawing, reference numeral 56 denotes an aligner adjacent to the atmospheric pressure transfer chamber 53, which is provided to rotate the wafer W to optically obtain the amount of eccentricity and align the wafer W with respect to the first substrate transfer mechanism 52. The first substrate transfer mechanism 52 transfers the wafer W between the carrier 54 on the carrier mounting table 55, the aligner 56, and a load lock chamber 61.

[0038] A second substrate transfer mechanism 62 having, for example, an articulated arm structure is provided in each load lock chamber 61, and the second substrate transfer mechanism 62 transfers the wafer W between the load lock chamber 61, the heat treatment module 60, and the treatment module 7. The inside of the treatment vessel constituting the heat treatment module 60 and the inside of the treatment vessel constituting the treatment module 7 are in a vacuum atmosphere, and the inside of the load lock chamber 61 can be switched between a normal pressure atmosphere and a vacuum atmosphere so that the wafer W can be transferred between the inside of these vacuum atmosphere treatment vessels and the normal pressure transfer chamber 53.

[0039] In the drawing, reference numeral 63 denotes a gate valve that can be opened and closed, and is provided between the normal pressure transfer chamber 53 and the load lock chamber 61, between the load lock chamber 61 and the heat treatment module 60, and between the heat treatment module 60 and the treatment module 7. The heat treatment module 60 includes the above-mentioned treatment vessel, an exhaust mechanism for exhausting the inside of the treatment vessel to form a vacuum atmosphere, and a stage that is provided in the treatment vessel and is capable of heating the wafer W placed thereon, and is configured to be able to perform the heating process described above.

[0040] The processing module 7 will be described with reference to the vertical cross-sectional side view of FIG. 12. This processing module 7 executes the processing gas supply process described above. In the figure, 71 denotes a processing vessel constituting the processing module 7. In the figure, 72 denotes a transfer port for the wafer W that opens in the side wall of the processing vessel 71 and is opened and closed by the gate valve 63 described above. A stage 81 for placing the wafer W thereon is provided within the processing vessel 71, and the stage 81 is provided with lift pins (not shown). The wafer W is transferred between the second substrate transfer mechanism 62 and the stage 81 via the lift pins.

[0041] A temperature adjustment unit 82 is embedded in the stage 81, and the wafer W placed on the stage 81 is adjusted to the above-mentioned temperature. This temperature adjustment unit 82 is configured as a flow path that is part of a circulation path through which a temperature adjustment fluid such as water flows, and the temperature of the wafer W is adjusted by heat exchange with the fluid. However, the temperature adjustment unit 82 is not limited to being a flow path for such a fluid, and may be configured, for example, by a heater for performing resistance heating.

[0042] One end of an exhaust pipe 83 opens into the processing vessel 71, and the other end of the exhaust pipe 83 is connected to an exhaust mechanism 85 constituted by, for example, a vacuum pump via a valve 84, which is a pressure changing mechanism. By adjusting the opening of the valve 84, the pressure inside the processing vessel 71 is set to a pressure range described later, and processing is performed.

[0043] A gas shower head 86, which is a processing gas supply mechanism, is provided at the upper side of the processing vessel 71 so as to face the stage 81. The downstream sides of gas supply paths 91-94 are connected to the gas shower head 86, and the upstream sides of the gas supply paths 91-94 are connected to gas supply sources 96-99 via flow rate adjusters 95, respectively. Each flow rate adjuster 95 includes a valve and a mass flow controller. The gas supplied from the gas supply sources 96-99 is supplied to the downstream side and cut off by opening and closing a valve included in the flow rate adjuster 95.

[0044] F2 gas, NH3 gas, Ar (argon) gas, and N2 (nitrogen) gas are respectively supplied from gas supply sources 96, 97, 98, and 99. Therefore, these HF gas, NH3 gas, Ar gas, and N2 gas can be respectively supplied from the gas shower head 86 into the processing vessel 71. The Ar gas and N2 gas are supplied as carrier gases into the processing vessel 71 together with the F2 gas and NH3 gas.

[0045] As shown in FIG. 11, the substrate processing apparatus 5 includes a control unit 50, which is a computer, and the control unit 50 includes a program, a memory, and a CPU. The program includes commands (each step) for performing the above-mentioned processing of the wafer W and transport of the wafer W. The program is stored in a storage medium, such as a compact disc, a hard disk, a magneto-optical disc, or a DVD, and is installed in the control unit 50. The control unit 50 outputs control signals to each part of the substrate processing apparatus 5 according to the program, and controls the operation of each part. Specifically, the operation of the processing module 7, the operation of the heat treatment module 60, the operation of the first substrate transport mechanism 52, the operation of the second substrate transport mechanism 62, and the operation of the aligner 56 are controlled by the control signals. The operation of the processing module 7 includes, for example, the temperature of the fluid supplied to the stage 81, the supply and cut-off of each gas from the gas shower head 86, and the adjustment of the exhaust flow rate by the valve 84.

[0046] The transfer path of the wafer W in the substrate processing apparatus 5 will be described. As described in FIG. 1, the carrier 54 storing the wafer W on which each film is formed is placed on the carrier mounting table 55. The wafer W is transferred in the normal pressure transfer chamber 53 → aligner 56 → normal pressure transfer chamber 53 → load lock chamber 61 in this order, and is transferred to the processing module 7 via the heat treatment module 60. Then, as described above, the processing gas supply step is performed, and the surface layers of the upper wall 21 and the lower wall 22 forming the recess 16 are altered to become the AFS layer 31. Next, the wafer W is transferred to the heat treatment module 60, and the AFS layer 31 and other reaction products are sublimated. Thereafter, the wafer W is transferred back and forth between the processing module 7 and the heat treatment module 60, and a cycle consisting of the processing gas supply step and the heating step is repeated.

[0047] Among the above cycles, in the first to Xth cycles, the process gas supply step is performed under the first process conditions as described in Fig. 10. Then, in the X+1th and subsequent cycles, the process gas supply step is performed under the second process conditions in which the flow rate ratio of NH3 gas is increased and the process gas supply time is lengthened. The cycle under the second process conditions is repeated Y times, and when a total of Z (X+Y) cycles are completed and each recess 16 has the desired vertical width as shown in Fig. 3, the wafer W is transferred from the heat treatment module 60 to the load lock chamber 61, then to the normal pressure transfer chamber 53, and then returned to the carrier 54.

[0048] The process gas supply step in the above cycle will be explained in more detail. In this process gas step, the inside of the process vessel 71 of the process module 7 is evacuated to, for example, 100 mTorr (13.3 Pa) to 10 Torr (1333 Pa), and the temperature of the wafer W is adjusted to, for example, -20°C to 60°C. In this state, F2 gas and NH3 gas are supplied into the process vessel 71 in parallel (step T1). The time for supplying F2 gas and NH3 gas and the flow rate ratio of F2 gas and NH3 gas in this step T1 are the process gas supply time and the flow rate ratio of NH3 gas that have been described so far. After the supply of F2 gas and NH3 gas is stopped, N2 gas is supplied into the process vessel 71 together with the exhaustion of the inside of the process vessel 71, thereby purging the F2 gas and NH3 gas remaining in the process vessel 71 (step T2). After this purging, the wafer W is carried out from the process vessel 71. Further, regarding the heating step in the above cycle, in this heating step, the wafer W is heated to a temperature higher than the temperature of the wafer W in steps T1 and T2, for example, to 80°C to 300°C, thereby sublimating the reaction products.

[0049] According to the present embodiment described above, the first to third requirements can be met by setting the process conditions in the process gas supply step differently between the front and rear cycles. That is, for the recess 16 after etching, the roughness of the Si film 13 forming the upper wall 21 and the lower wall 22 can be suppressed, the recess 16 can be made rectangular, and the variation in the vertical width of the recess 16 between the top, middle, and bottom can be suppressed. Therefore, the yield of semiconductor products manufactured from the wafer W after etching can be increased.

[0050] Although it has been described that the number of times the front-stage cycle is performed is X times and the number of times the rear-stage cycle is performed is Y times, these X times and / or Y times may be one time. Therefore, the front-stage cycle and the rear-stage cycle do not have to be performed repeatedly. In addition, in the examples described so far, the flow rate ratio of NH3 gas or the supply time of the processing gas in the processing gas supply step is changed only once, resulting in a two-stage process consisting of the front-stage cycle and the subsequent rear-stage cycle. The process is not limited to being two-stage, and the processing conditions may be changed two or more times to result in a three or more stage process.

[0051] In addition, when performing two-stage processing, it is preferable to set the flow rate ratio of NH3 gas to be larger in the first cycle than in the second cycle, or to set the supply time of the processing gas to be shorter in the first cycle than in the second cycle, in order to form a rectangular shape. It was also stated that it is preferable to set the number of cycles to be equal to or greater than the number of cycles in the second cycle. When performing processing in three or more stages, for example, the processing conditions of the first stage and the processing conditions of any stage from the second stage onwards should have such a relationship, and the number of cycles should be, for example, equal to or greater than the number of cycles in the first stage.

[0052] In this embodiment, the atmosphere around the wafer W is kept in a vacuum atmosphere from the start of the first cycle to the end of the last cycle, but for example, each cycle may be performed in a different apparatus, and the wafer W may be moved between the apparatuses by passing through an air atmosphere. However, from the viewpoint of preventing a decrease in throughput, it is preferable to perform processing in a vacuum atmosphere around the wafer W from the first cycle to the last cycle, as in the case of using the above-mentioned substrate processing apparatus 5.

[0053] Incidentally, after the process gas supply step in any cycle, a second fluorine-containing gas may be supplied before the heating step to further suppress the roughness after etching of the Si film 13. The second fluorine-containing gas (i.e., halogen-containing gas) is, for example, HF (hydrogen fluoride) gas, and is therefore a different type of gas from the first fluorine-containing gas, F2 gas, constituting the process gas. The HF gas is supplied to the wafer W in the process module 7.

[0054] The process procedure in the process module 7 when supplying this HF gas will be described. After the above-mentioned supply of the process gas (step T1) and purging of the inside of the process vessel 71 (step T2) are performed in order, HF gas is supplied into the process vessel 71 (step T3). After the supply of HF gas is stopped, in step T4, the process vessel 71 is brought into a state where N2 gas is supplied and exhausted in the same manner as in step T2, and the remaining HF gas is purged. Thereafter, the wafer W is unloaded from the process module 7. Note that, in the case of supplying HF gas, which is a roughness suppression gas, in this manner, steps T1 to T4 are the first step, step T1 in which both a halogen-containing gas (F2 gas) and a basic gas (NH3 gas) are supplied corresponds to the first supply step, and step T3 in which only one of these gases is supplied corresponds to the second supply step.

[0055] It is believed that the supply of the above-mentioned HF gas causes the NH3 contained in the AFS layer 31 formed in step T1 to react with the HF gas to produce a reaction product, and that the reaction product slightly etches the surface layer of the Si film 13 covering the AFS layer 31, thereby suppressing roughness.

[0056] The processing module 7 performing the above steps T1 to T4 is configured such that, in addition to the gas supply sources 96 to 99 described above, a HF gas supply source is provided, and the HF gas supply source is connected to a gas shower head 86 via a flow path. A flow rate regulator 95 is provided in the HF gas flow path, as in the flow paths of other gases, and HF gas is supplied into the processing vessel 71 via the gas shower head 86 at a desired flow rate. When the HF gas is supplied into the processing vessel 71, for example, Ar gas and N2 gas are also supplied into the processing vessel 71 as carrier gases.

[0057] Moreover, instead of supplying HF gas to the wafer W as a gas for suppressing roughness, NH3 gas may be supplied to the wafer W. When NH3 gas is supplied in this manner, this NH3 gas reacts with the fluorine component contained in the AFS layer 31 formed in step T1 to generate a reaction product, and it is considered that the reaction product slightly etches the surface layer of the Si film 13 coated on the AFS layer 31, thereby suppressing roughness.

[0058] In addition, when NH3 gas is used as the roughness suppression gas, for example, F2 gas and NH3 gas may be supplied together as the process gas, and NH3 gas may be continued to be supplied to the wafer W even after the supply of F2 gas is stopped, so that the purge with N2 gas in step T2 may be omitted. In addition, the supply of each of the above roughness suppression gases may be performed in any cycle. For example, it may be performed in the last cycle of the cycles in each stage, or it may be performed in every cycle.

[0059] Although it has been described that the temperature of the wafer W is increased to a temperature higher than that during the process gas supply step in order to remove the reaction products, the present invention is not limited to such removal. For example, after the above steps T1 and T2 are performed in the process module 7, the opening of the valve 74 disposed in the exhaust pipe 73 may be increased to reduce the pressure in the process vessel 71, thereby lowering the sublimation temperature of the reaction products. Therefore, the present invention is not limited to performing one cycle in different process vessels.

[0060] As described above, the timings of the end and start of supply of the processing gases F2 gas and NH3 gas may be shifted from each other. The fluorine-containing gas contained in the processing gas is not limited to F2 gas, and may be, for example, IF7 gas, IF5 gas, ClF3 gas, or SF6 gas. Incidentally, "containing fluorine" does not mean that fluorine is contained as an impurity, but that fluorine is contained as a constituent. Similarly, "containing Ge" described later means that fluorine is contained as a constituent, not as an impurity.

[0061] Moreover, the recess 16 is not limited to being laid horizontally, but is open vertically, and the present technology can also be applied to the case of etching the Si film forming the sidewall of such a recess 16. In that case, the recess 16 can be etched so as to meet the first and second requirements (requirements of roughness and rectangular shape) among the first to third requirements described above.

[0062] Even if the side wall 23 of the recess 16 is formed by a Ge film instead of the SiGe film 12, it is considered that it is possible to produce GeF4 using a fluorine-containing gas and etch the inner side of the recess 16. Therefore, the side wall 23 of the recess 16 is not limited to being formed by the SiGe film 12, and may be formed by any film containing a Ge film.

[0063] The embodiments disclosed herein should be considered to be illustrative and not restrictive in all respects. The above-described embodiments may be omitted, substituted, modified, or combined in various forms without departing from the scope and spirit of the appended claims.

[0064] [Evaluation test] Evaluation tests 1 to 5 performed on this technology will be described below. In each test, a cycle consisting of a process gas supply step and a heating step was performed on a substrate having the film structure described in FIG. 1 using the device described in the embodiment, and the dimensions of each part were measured from SEM images acquired before and after the process. In evaluation tests 1 to 4, the process conditions were not changed between the front and rear cycles described in FIG. 10, and the process was performed under a single process condition. In evaluation tests 1 to 5, the period during which F2 gas was supplied and the period during which NH3 gas was supplied in the process gas supply step were the same. In other words, the timing at which the supply of these gases to the process vessel 71 started and the timing at which the supply of these gases to the process vessel 71 ended were also the same.

[0065] The following three dimensions were measured in each evaluation test. One was the width of the groove 15, and hereinafter, this width is referred to as the groove width dimension. The other was the vertical width of the recess 16 at a position slightly away from the opening of the recess 16 toward the back of the recess 16 by a certain distance, and hereinafter, this vertical width is referred to as the opening side dimension. Still another was the vertical width of the recess 16 at a position slightly away from the lateral wall 23 of the recess 16 toward the opening of the recess 16 by a certain distance, and hereinafter, this vertical width is referred to as the back side dimension.

[0066] Among the recesses 16 formed in multiple stages, a plurality of recesses 16 in a predetermined stage are regarded as the top recesses 16 in the evaluation test. The groove width dimension, the opening side dimension, and the rear side dimension of each of the recesses 16 in the top are calculated by (dimension after etching - dimension before etching) / 2. This calculation is performed for each of the plurality of recesses 16 in the top. Then, the average value among the recesses 16 is calculated for each of the calculated values ​​obtained from the groove width dimension, the calculated value obtained from the opening side dimension, and the calculated value obtained from the rear side dimension, and the obtained average value is regarded as the groove width etching amount, the opening side etching amount, and the rear side etching amount in the recesses 16 in the top. The groove width dimension of each recess 16 in the top is obtained from between the upper walls 21 of the recesses 16 facing each other in the horizontal direction.

[0067] Similarly, the recesses 16 in a given step are designated as middle recesses 16 and bottom recesses 16, respectively, and the groove width etching amount, opening side etching amount, and back side etching amount are calculated for each of the middle recesses 16 and bottom recesses 16. Each etching amount calculated in this way is divided by a predetermined reference amount to obtain a standardized value, which is shown in Figures 13 to 15 shown later. The larger the standardized value, the larger the actual etching amount. Note that each etching amount (i.e., actual measured value) before test standardization obtained in each evaluation test is a value of 20 nm or less. Incidentally, for evaluation tests 3 to 5, a table showing the shape of the recess 16 after etching and the state of roughness of the recess 16 is shown as the test results. In each table, those that have a clear rectangular shape and a clear round shape are shown as "rectangle" and "round", respectively, and those with other shapes are shown as "intermediate". Therefore, those that are not clearly rectangular but have a relatively high rectangularity and are at a level that is not problematic in practical use are also shown as "intermediate" in the table, so the processing conditions that result in a "rectangle" are not the only processing conditions that can be used in practice. In the table, those that have neither microroughness nor large roughness and have a high level of smoothness are shown as "smooth".

[0068] Evaluation test 1 In the evaluation tests 1-1 to 1-3, the substrate was etched by repeating a cycle of a process gas supply step and a heating step. The combination of the number of cycles and the pressure in the process vessel 71 during the process gas supply step was changed between the evaluation tests 1-1 to 1-3. In the evaluation test 1-1, the pressure in the process vessel 71 was 200 Pa (1.5 Torr), and the number of cycles was 9. In the evaluation test 1-2, the pressure in the process vessel 71 was 666.6 Pa (5 Torr), and the number of cycles was 9. In the evaluation test 1-3, the pressure in the process vessel 71 was 666.6 Pa (5 Torr), and the number of cycles was 3. Each process condition, such as the temperature of the wafer W during the process gas supply, the supply time of the process gas, and the flow rate ratio of the NH3 gas, was set to a value within the range described in the embodiment.

[0069] The results of this evaluation test 1 show the average values ​​of the values ​​obtained from the top, middle, and bottom recesses 16. In evaluation test 1-1, the groove width etching amount was 1.26, the opening side etching amount was 1.62, and the back side etching amount was 1.66. In evaluation test 1-2, the etching progressed so much that the upper wall 21 and the lower wall 22 were lost, and each etching amount could not be measured. In evaluation test 1-3, the loss of the upper wall 21 and the lower wall 22 was prevented, and the groove width etching amount was 1.78, the opening side etching amount was 1.9, and the back side etching amount was 1.9.

[0070] From the results of the above evaluation test 1, it is found that the etching rate can be increased by relatively increasing the pressure in the processing vessel 71 when the processing gas is supplied, and that excessive etching can be prevented by appropriately setting the number of cycles. It is also found that the pressure in the processing vessel 71 is preferably greater than 1.5 Torr (200 Pa), and more preferably greater than 5 Torr (666.6 Pa).

[0071] Evaluation test 2 In the evaluation tests 2-1 to 2-5, each substrate was processed by performing the process gas supply step and the heating step once each, without repeating the cycle of these steps. The process gas supply time was set to a different value among the evaluation tests 2-1 to 2-5, and was set to 5 seconds, 10 seconds, 15 seconds, 25 seconds, and 35 seconds in the evaluation tests 2-1, 2-2, 2-3, 2-4, and 2-5, respectively. In addition, each process condition, such as the flow rate ratio of NH3 gas, the temperature of the wafer W during the process gas supply, and the pressure inside the process container 71, was set to a value within the range described in the embodiment.

[0072] The graphs in Figures 13 to 15 below summarize the etching amounts in Evaluation Test 2. Each graph shows the relationship between the etching amount and the supply time of the processing gas, with the vertical axis showing the etching amount and the horizontal axis showing the supply time. Figure 13 shows the trench width etching, Figure 14 shows the opening side etching amount, and Figure 15 shows the back side etching amount.

[0073] As is clear from the graph of FIG. 13, the etching amount in the lateral direction of the groove 15 is less related to the supply time of the processing gas for each of the top, middle, and bottom. As is clear from the graphs of FIG. 14 and FIG. 15, the etching amount on the opening side of the recess 16 and the etching amount on the back side of the recess 16 are relatively small when the etching time is relatively long. This is considered to be because a lot of AFS is generated on the top side as described in the embodiment. Note that, as described in the embodiment, the graphs of FIG. 14 and FIG. 15 show that when the etching time is short, the etching amount is larger on the top, and when the etching time is long, the etching amount is larger on the bottom. From the graphs of FIG. 14 and FIG. 15, when the supply time of the processing gas is in the range of 10 seconds to 15 seconds for each of the top, middle, and bottom, the slopes of the graphs are approximately the same, and the difference in the etching amount between the top, middle, and bottom is suppressed. Therefore, this evaluation test 2 shows that it is preferable to set the supply time of the processing gas in one cycle to 10 seconds to 15 seconds.

[0074] Evaluation Test 3 In the evaluation tests 3-1 to 3-6, the combination of the flow rate ratio of NH3 gas and the supply time of the processing gas was different for each substrate, and the processing was performed on each substrate. The number of cycles was 3. The supply time of the processing gas was set to 10 seconds in the evaluation tests 3-1 to 3-3, and to 15 seconds in the evaluation tests 3-1 to 3-3. The flow rate ratio of NH3 gas was set to 0.01 in the evaluation tests 3-1 and 3-4, 0.014 in the evaluation tests 3-2 and 3-5, and 0.02 in the evaluation tests 3-3 and 3-6. The other processing conditions were the same as those in the evaluation test 2. Table 1 below summarizes the results obtained in the evaluation test 3. In the evaluation test 3, the etching amount of each part was obtained as in the evaluation test 2.

[0075] [Table 1]

[0076] Regarding the roughness of the upper wall 21 and the lower wall 22 of the recess 16 after etching, microroughness was confirmed in the middle and bottom recess 16 of Evaluation Test 3-1. Furthermore, large roughness was confirmed in the top recess 16 of Evaluation Test 3-3 and the top, middle, and bottom recess 16 of Evaluation Test 3-6. Regarding the shape of the recess 16 after etching, the top, middle, and bottom recess 16 of Evaluation Tests 3-1 and 3-4 and the bottom recess 16 of Evaluation Test 3-2 were rectangular.

[0077] Therefore, it was confirmed from the evaluation test 3 that micro-roughness was formed when the flow rate ratio of NH3 gas was relatively small and the etching time was relatively short. It was also confirmed that large roughness was formed when the flow rate ratio of NH3 gas was relatively large. As described in the embodiment, it is considered that the roughness state was different due to the different amount of AFS generated. It was also confirmed from the evaluation test 3 that there was a tendency for a rectangular shape to be formed when the flow rate ratio of NH3 gas was relatively small. And, it is found that in order to form a rectangular shape, it is preferable to make the flow rate ratio smaller than 0.014, and more preferably 0.01 or less.

[0078] However, in the evaluation tests 3-1, 3-4, and 3-5 in which the flow rate ratio of NH3 gas was set to 0.01 and a rectangular shape was obtained, the variation in the amount of etching between the top and bottom was relatively large. Therefore, the results of the evaluation tests 3-1 to 3-6 did not sufficiently meet one or more of the first to third requirements.

[0079] Evaluation Test 4 In the evaluation tests 4-1 to 4-5, the combination of the flow rate ratio of NH3 gas, the supply time of the processing gas, and the number of cycles was different for each substrate, and the substrates were processed. The supply time of the processing gas was set to 10 seconds, 15 seconds, 15 seconds, 12.5 seconds, and 10 seconds in the evaluation tests 4-1, 4-2, 4-3, 4-4, and 4-5, respectively. The flow rate ratio of NH3 gas in the processing gas was set to 0.01 in the evaluation tests 4-1, 4-2, and 4-4, and to 0.014 in the evaluation tests 4-3 and 4-5, respectively. The number of cycles was set to 3 in the evaluation tests 4-1 to 4-4, and to 4 in the evaluation test 4-5, respectively. The other processing conditions were the same as those in the evaluation tests 2 and 3. The following Table 2 summarizes the results of the evaluation test 4. In the evaluation test 4 and the evaluation test 5 described below, the etching amount of each part was obtained in the same manner as in the evaluation test 3.

[0080] [Table 2]

[0081] Regarding the roughness of the upper wall 21 and the lower wall 22 of the recess 16 after etching, microroughness was confirmed in the middle and bottom recess 16 of the evaluation tests 4-1 and 4-4 and in the bottom recess 16 of the evaluation test 4-5. The roughness of the other recesses 16 was suppressed within a desirable range. A comparison among the evaluation tests 4-1, 4-2, and 4-4, which have the same number of cycles and the same NH3 gas flow rate ratio, shows that microroughness is likely to be formed when the supply time of the processing gas is short, as described in the embodiment.

[0082] Regarding the shape of the recesses 16 after etching, the top, middle, and bottom recesses 16 in the evaluation tests 4-1 and 4-2 and the middle and bottom recesses 16 in the evaluation tests 4-4 and 4-5 were rectangular, while the other recesses 16 were round. Comparison among the evaluation tests 4-1, 4-2, and 4-4 shows that there is a tendency for the recesses 16 to have a rectangular shape when the supply time of the processing gas is relatively short as described in the embodiment. In addition, in the evaluation tests 4-1, 4-2, and 4-4 in which the flow rate ratio of the NH3 gas was set to 0.01, two or three of the top, middle, and bottom recesses were rectangular, so it is found that the flow rate ratio is preferably 0.01 or less in order to obtain a rectangular shape.

[0083] Looking at the opening side etching amount and the back side etching amount, in evaluation tests 4-2, 4-4, and 4-5, the top etching amount was larger than the middle and bottom etching amounts. In other words, there was a relatively large variation in the etching amount between the top, middle, and bottom. Also, in evaluation test 4-3, the top etching amount was smaller than the middle and bottom etching amounts. As a result of the above, the recesses 16 of each substrate in evaluation test 4 did not sufficiently meet one or more of the first to third requirements.

[0084] By the way, further verification of the results of evaluation test 4 shows that when the flow rate ratio of NH3 gas is 0.01, the recess 16 tends to be rectangular and the amount of etching of the top tends to be large. Also, when the flow rate ratio of NH3 gas is 0.014, it is shown that the recess 16 tends to be round and the tendency of the amount of etching of the top to be large is alleviated compared to when the flow rate ratio is 0.01. Furthermore, when the flow rate ratio of NH3 gas is 0.01 and 0.014, there are cases where the roughness is suppressed within a desirable range. However, when the flow rate ratio of NH3 gas in the processing gas is 0.01 and 0.014, the desired result was not obtained with respect to the roughness when the processing gas supply time was 10 seconds.

[0085] Considering the results of the evaluation test 4 and the evaluation test 2 that the supply time of the processing gas is preferably 10 to 15 seconds, it is considered that the flow rate ratio of NH3 gas in the processing gas is set to a range of more than 0.01 and less than 0.014, and the supply time of the processing gas is set to a range of more than 10 seconds and less than 15 seconds. In other words, it is considered that an allowable range that is more suitable for the first to third requirements than the processing conditions set in the evaluation test 4 within such a range of each parameter is searched for, and processing is performed using values ​​within the allowable range. However, the allowable range for the flow rate ratio of NH3 gas is an even narrower range of the relatively narrow range of 0.01 to 0.014. Therefore, for example, when the film structure of the etching target changes slightly, it is difficult to change the flow rate ratio to deal with the change, and there is a concern that the practicality is low. Therefore, the multi-stage processing by changing the processing conditions described in the embodiment is effective.

[0086] Evaluation Test 5 In evaluation tests 5-1 to 5-5, two-stage processing (= a front-stage cycle and a rear-stage cycle) was performed by changing the processing conditions as described in the embodiment. The combination of the processing conditions for the front-stage cycle and the processing conditions for the rear-stage cycle was changed for each substrate. The processing conditions other than the flow rate ratio of NH3 gas, the supply time of the processing gas, and the number of cycles were the same as the processing conditions for evaluation test 4.

[0087] The NH3 gas flow rate ratio and the process gas supply time in the first cycle of evaluation tests 5-1, 5-2, and 5-4 and the second cycle of evaluation tests 5-3 and 5-5 were set to the same process conditions as those under which a rectangular shape was obtained in evaluation test 4. The NH3 gas flow rate ratio and the process gas supply time in the second cycle of evaluation tests 5-1, 5-2, and 5-4 and the first cycle of evaluation tests 5-3 and 5-5 were set to the same process conditions under which roughness was suppressed within a desirable range in evaluation test 4. Table 3 below summarizes the process conditions and results of evaluation test 5.

[0088] Considering the flow rate ratio of NH3 gas, the supply time of the processing gas, and the difference in the number of cycles, the tendency to form a rectangular shape is as follows: first stage of evaluation test 5-1 > first stage of evaluation test 5-2 and last stage of evaluation test 5-3 > first stage of evaluation test 5-4 > last stage of evaluation test 5-5. And the tendency to suppress roughness is as follows: last stage of evaluation test 5-2, first stage of 5-3, and first stage of 5-5 > last stage of evaluation test 5-1 and last stage of 5-4.

[0089] [Table 3]

[0090] The results of evaluation tests 5-1 to 5-5 are examined for the difference in the amount of etching between the top and bottom. In summary, the difference in the amount of etching of the recess 16 between the top and bottom obtained from evaluation test 4 is corrected according to the number of cycles. As a result, the expected difference in the amount of etching between the top and bottom when only the first cycle in evaluation test 5 is performed and the expected difference in the amount of etching between the top and bottom when only the second cycle in evaluation test 5 is performed are calculated. The expected differences in the amount of etching are summed up to obtain the expected difference in the amount of etching between the top and bottom. This expected difference in the amount of etching and the expected difference in the amount of etching obtained in the process of obtaining the difference in the expected amount of etching are compared with the actual difference in the amount of etching between the top and bottom obtained from evaluation test 5.

[0091] Specifically, in evaluation test 5-1, the amount of etching on the top opening side and the amount of etching on the back side are A1X1 and A2X2, respectively, and the average of these is A3X3. The amount of etching on the bottom opening side and the amount of etching on the back side are B1Y1 and B2Y2, respectively, and the average of these is B3Y3. The value calculated as A3X3-B3Y3 is the difference in the amount of etching between the top and bottom in evaluation test 5-1. The difference in the amount of etching between the top and bottom calculated in this way was a positive value. Therefore, the amount of etching on the top was larger than that on the bottom in evaluation test 5-1.

[0092] Except for the number of cycles, the evaluation tests 4-1 and 4-3 were conducted under the same conditions as the front and rear cycles of the evaluation test 5-1. For the evaluation test 4-1, (average value of the top opening side etching amount and the top back side etching amount) - (average value of the bottom opening side etching amount and the bottom back side etching amount) = C1A1 was obtained. The number of cycles in the evaluation test 4-1 was 3, but the number of cycles in the front stage of the evaluation test 5-1 was 4, so the value was corrected by multiplying by 4 / 3, and the obtained value was set as the difference in the expected etching amount C2A2 (= C1A1 x 4 / 3). Note that this C2A2 value was a positive value. In other words, the etching amount of the top is larger between the top and bottom.

[0093] Similarly, for evaluation test 4-3, which had the same conditions as the latter cycle, the difference in the expected etching amount was calculated. Specifically, for evaluation test 4-3, (average value of the etching amount on the top opening side and the etching amount on the top back side) - (average value of the etching amount on the bottom opening side and the etching amount on the bottom back side) = C1A'1 was calculated. Since the number of cycles in evaluation test 4-3 was 3, but the number of latter cycles in evaluation test 5-1 was 1, this was corrected by multiplying by 1 / 3, and this value was taken as the difference in the expected etching amount C2A'2 (=C'A'1 x 1 / 3). This C'A'2 was a negative value. In other words, the etching amount of the bottom is larger between the top and bottom.

[0094] The estimated total etching amount C2A2+C2A'2 was approximately the same as A3X3-B3Y3 in evaluation test 5-1. C2A2 was also larger than A3X3-B3Y3. The above calculation results show that even if a relatively large bias in the etching amount between the top side and the bottom side occurs at the end of the first stage of the two-stage processing described in the embodiment, the bias can be offset by performing a second stage of processing in which the etching amount is increased on the opposite side of the top side or the bottom side from the first stage.

[0095] As described above, in evaluation test 4-1, each recess 16 was rectangular, and microroughness was formed in the middle and bottom recesses 16. In evaluation test 4-3, each recess 16 was intermediate in shape, and the roughness of each recess 16 was within a desirable range. In contrast, the test results of evaluation test 5-1 showed that the middle and bottom recesses 16 were rectangular, and microroughness was confirmed in the top, middle, and bottom recesses 16. Thus, the shape and roughness of the recesses 16 were relatively significantly affected by the processing conditions of the previous cycle.

[0096] When the above calculation was performed for Evaluation Tests 5-2 to 5-5 in the same manner as Evaluation Test 5-1, it was estimated that the bias in the amount of etching between the top side and the bottom side was offset by the implementation of the front-stage cycle and the rear-stage cycle. Note that, in the verification, Evaluation Tests 5-2 and 5-3 used the same processing conditions as Evaluation Test 5-1, Evaluation Tests 4-1 and 4-3, but Evaluation Tests 5-4 and 5-5 used the processing conditions of Evaluation Test 4-2, so the verification was performed using the data from Evaluation Test 4-2. The front and rear cycles of Evaluation Test 5-2 were the same as those of Evaluation Test 5-1, and the processing conditions were the same as those of Evaluation Tests 4-1 and 4-3, respectively, except for the number of cycles. As shown in Table 3, in Evaluation Test 5-2, the bottom recesses 16 were rectangular, and microroughness was confirmed in each of the middle and bottom recesses 16. Therefore, in Evaluation Test 5-2, as in Evaluation Test 5-1, the processing conditions of the front cycle had a relatively large effect on the shape and roughness of the recesses 16.

[0097] Regarding evaluation test 5-3, the front-stage cycle and the rear-stage cycle are the same as evaluation test 4-3 and evaluation test 4-1, respectively, except for the number of cycles. As shown in Table 3, in evaluation test 5-3, the top, middle, and bottom recesses 16 were intermediate in shape, while the roughness was suppressed within the allowable range. In evaluation test 4-3, the recesses 16 were intermediate in shape, and the roughness of each recess 16 was within the allowable range, while in evaluation test 4-1, the recesses 16 were rectangular, and micro-roughness was formed in the middle and bottom recesses 16. Therefore, in this evaluation test 5-3 as well, the effect of the processing conditions of the front-stage cycle on the shape and roughness of the recesses 16 was relatively large.

[0098] Regarding evaluation test 5-4, the front-stage cycle and the rear-stage cycle are the same as evaluation test 4-2 and evaluation test 4-3, respectively, except for the number of cycles. As shown in Table 3, in evaluation test 5-4, bottom recess 16 is rectangular, and micro-roughness is confirmed in bottom recess 16. On the other hand, in evaluation test 4-2, each recess 16 is rectangular, and micro-roughness is formed in bottom recess 16. And, in evaluation test 4-3, each recess 16 is an intermediate shape, and the roughness of each recess 16 is within a desirable range. Therefore, in this evaluation test 5-4, the effect of the processing conditions of the front-stage cycle on the shape and roughness of recess 16 is relatively large.

[0099] Regarding the evaluation test 5-5, the front-stage cycle and the rear-stage cycle are the same as the evaluation test 4-3 and the evaluation test 4-2, respectively, except for the number of cycles. As described above, in the evaluation test 5-5, the roughness of each of the top, middle, and bottom recesses 16 was suppressed within the allowable range and had an intermediate shape. On the other hand, in the evaluation test 4-3, each recess 16 had an intermediate shape, and the roughness of each recess 16 was within a desirable range. In the evaluation test 4-2, each recess 16 had a rectangular shape, and micro-roughness was formed in the bottom recess 16. Therefore, in the evaluation test 5-5, the effect of the processing conditions of the front-stage cycle on the shape and roughness of the recess 16 was relatively large. Incidentally, in the evaluation test 5-5, the roughness of the top, middle, and bottom was suppressed, and the difference in the etching amount between the top and bottom was the smallest in the evaluation test 5 for X3-Y3. The shape of each recess 16 was slightly more rounded than that of the recess 16 in Evaluation Test 5-1, but was more rectangular than that of the recess 16 in Evaluation Test 4-3 and was therefore practical. Therefore, among Evaluation Tests 5-1 to 5-5, Evaluation Test 5-5 had the most favorable results.

[0100] As described above, it was confirmed from the evaluation test 5 that the bias in the amount of etching between the top side and the bottom side in the front cycle and the rear cycle is offset, and as a result, the amount of etching from the top side to the bottom side can be made highly uniform. It was also confirmed that the processing conditions performed in the front cycle have a large effect on the shape and roughness of the recess 16 after etching. And, as described above, the roughness is suppressed at the top, middle, and bottom in the evaluation tests 5-3 and 5-5. Therefore, in order to suppress the roughness, it was shown that it is preferable to make the flow rate ratio of NH3 gas larger in the first processing condition of the front cycle than in the second processing condition of the rear cycle, as described in the embodiment. When the evaluation tests 5-3 and 5-5 are compared, the difference in the amount of etching between the top and bottom is different, and the difference in the amount of etching is smaller in the evaluation test 5-5. Since the supply time of the processing gas under the second processing conditions differs between evaluation tests 5-3 and 5-5, it can be seen that by adjusting the supply time of the processing gas between the first and second processing conditions, it is possible to control the balance of the etching amount between the top, middle, and bottom while suppressing the roughness of each recess 16.

[0101] Regarding the rectangularity of the recess 16, the results of the evaluation tests 5-1, 5-2, and 5-4 show that in order to improve the rectangularity, it is preferable to make the flow rate ratio of NH3 gas smaller under the first processing condition than under the second processing condition. And, since the rectangularity was the highest under the evaluation test 5-1, it is shown that in order to obtain a higher rectangularity, it is preferable to make the supply time of the processing gas shorter under the first processing condition than under the second processing condition. [Explanation of symbols]

[0102] W wafer 12. SiGe film 13.Si film 41 F2 gas 42 NH3 gas

Claims

1. a first step of supplying a process gas containing a halogen-containing gas and a basic gas to a substrate having a recess with a side wall formed of a silicon film and a back wall formed of a germanium-containing film to modify a surface of the silicon film and generate a reaction product; a second step of removing the reaction product to widen the width of the recess; performing a cycle consisting of the first step and the second step a plurality of times; performing the first step of the first cycle under first processing conditions and the first step of the second cycle under second processing conditions different from the first processing conditions; A substrate processing method comprising:

2. 2. The substrate processing method according to claim 1, wherein the first processing condition and the second processing condition differ from each other in at least one of a ratio of the basic gas to the halogen-containing gas in the processing gas and a supply time of the processing gas.

3. 3. A substrate processing method according to claim 2, wherein the ratio of the basic gas to the halogen-containing gas is smaller under the first processing conditions than under the second processing conditions, or the supply time of the processing gas is shorter under the first processing conditions than under the second processing conditions.

4. 3. A substrate processing method according to claim 2, wherein the ratio of the basic gas to the halogen-containing gas is greater under the first processing conditions than under the second processing conditions, or the supply time of the processing gas is longer under the first processing conditions than under the second processing conditions.

5. The recess is formed in a plurality of stages in a thickness direction of the substrate, 5. The substrate processing method according to claim 3, wherein the plurality of recesses are each open in a direction intersecting the thickness direction.

6. 3. The substrate processing method according to claim 2, wherein the halogen-containing gas is a first fluorine-containing gas, and the basic gas is an ammonia gas.

7. 7. The substrate processing method according to claim 6, wherein the first fluorine-containing gas is fluorine gas.

8. 2. The method of claim 1, wherein the germanium-containing film is a silicon germanium film.

9. 3. The substrate processing method of claim 2, wherein a ratio of the basic gas to the halogen-containing gas under one of the first and second process conditions is less than 0.

014.

10. 3. The substrate processing method according to claim 2, wherein the supply time of the processing gas is 10 to 15 seconds under the first processing conditions and the second processing conditions.

11. 11. The substrate processing method according to claim 10, wherein the first step includes the step of supplying the processing gas into a processing vessel storing the substrate under a pressure of more than 200 Pa inside the processing vessel.

12. The first step comprises: a first supply step of supplying the halogen-containing gas to the substrate and supplying the basic gas to the substrate in parallel; a second supply step of supplying only one of the halogen-containing gas and the basic gas to the substrate after the first supply step; The method of claim 1 , further comprising:

13. The halogen-containing gas is a first fluorine-containing gas supplied to the substrate in the first supply step; a second fluorine-containing gas supplied to the substrate in the second supply step; 13. The substrate processing method according to claim 12, wherein the first fluorine-containing gas and the second fluorine-containing gas are different in type from each other.

14. the first fluorine-containing gas is a fluorine gas; 14. The substrate processing method according to claim 13, wherein the second fluorine-containing gas is hydrogen fluoride gas.

15. a first processing section that supplies a processing gas containing a halogen-containing gas and a basic gas to modify a surface of a silicon film in a substrate having a recess having a side wall formed of a silicon film and a back wall formed of a germanium-containing film, thereby generating a reaction product; a second processing unit that removes the reaction product to widen the width of the recess; a control unit that outputs a control signal so that a cycle consisting of a process by the first processing unit and a process by the second processing unit is performed a plurality of times, the process by the first processing unit in an earlier cycle is performed under a first processing condition, and the process by the first processing unit in a later cycle is performed under a second processing condition different from the first processing condition; A substrate processing apparatus comprising:

Images