Substrate processing method and substrate processing system

Abstract

ER scan(R)=ER ref(R)×Ratio scan(R)・・・(1) Ratio scan(R)=b 0×exp(b 1×T+b

Description

[Technical Field]

[0001] This invention relates to a substrate processing method and a substrate processing system. [Previous Technology]

[0002] Patent Document 1 discloses a substrate processing method comprising the following steps: a grinding step of grinding the surface of a substrate; a measuring step of measuring the thickness of the substrate after grinding; a condition determining step of determining the processing conditions for a wet etching process on the substrate based on the measured thickness of the substrate; and a step of supplying a processing solution to the substrate after grinding based on the determined processing conditions to perform a wet etching process. [Prior Art Documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2018-147908 [Summary of the Invention]

[0004] [The problem the invention aims to solve]

[0005] According to the technology of the present invention, the surface shape of the etched object after etching treatment can be appropriately controlled. [Means for solving the problem]

[0006] A substrate processing method for a substrate of the present invention includes the following steps: determining optimal etching conditions; and, based on the optimal etching conditions, while rotating the substrate, the etching object is moved back and forth in the radial direction by the etching solution supply unit passing above the rotation center of the etching object, and the etching solution is supplied from the etching solution supply unit to the surface of the etching object to etch the surface; the step of determining the optimal etching conditions includes the following steps: obtaining learning data including the etching amount distribution in the radial direction of the etching object after etching the surface of the etching object with a plurality of different etching conditions; and, in order to minimize the first residual distribution between the first etching amount distribution predicted by the first etching amount distribution and the first target etching amount distribution using the learning data through the following formulas (1) to (6), optimizing the first etching conditions of the first etching amount distribution using the least square method. ER scan(R) = ER ref(R) × Ratio scan(R)・・・(1) Ratio scan(R) = b 0 × exp(b 1 × T + b 2) + C・・・(2) b 0 = f(S,V)・・・(3) b 1 = f(S,V)・・・(4) b 2 = f(S,V)・・・(5) T = (LR) / V・・・(6) Where, ER scan: the amount of etching when the etchant supply section moves back and forth; ER ref: the amount of etching when the etchant supply section does not move back and forth; Ratio scan: scan ratio; R: the position from the center of the etched object; T: the time without etchant supply; C: constant; S: the rotation speed when the etched object is rotated; V: the scan speed when the etchant supply section moves back and forth; L: the scan width when the etchant supply section moves back and forth. [Effects of the Invention]

[0007] Through the present invention, the surface shape of the etched object after etching treatment can be appropriately controlled.

Implementation Method

[0009] In the manufacturing process of a semiconductor device, a semiconductor substrate (hereinafter referred to as a "wafer") on which a plurality of electronic circuits and other components are formed on its surface is thinned by grinding, and the grinding surface of the wafer is smoothed. The smoothing of the grinding surface is, for example, by supplying etching solution from above the grinding surface of the wafer while rotating the wafer, that is, performing spin etching.

[0010] Patent Document 1 disclosed a wet etching process performed on a polished wafer to remove the damage layer formed on the surface of the wafer due to the polishing process. In the condition determination step described in Patent Document 1, the operation of the nozzle supplying the processing liquid, the wafer rotation speed, the amount of processing liquid supplied, the supply time of the processing liquid, and the type of processing liquid are determined as conditions for the wet etching process based on the thickness of the wafer obtained in the measurement step.

[0011] However, as disclosed in Patent Document 1, when performing spin etching by rotating the wafer while supplying the processing solution, it is difficult to perform precise etching control because the processing solution supplied to the wafer surface flows radially outward due to centrifugal force. More specifically, it is particularly difficult to properly control the surface shape of the wafer after etching, especially in the center of the wafer.

[0012] Therefore, the present invention provides a method for etching the surface of a wafer by simultaneously rotating the wafer and moving a nozzle back and forth (scanning) in a radial direction passing through the center of the wafer, while supplying etchant from the nozzle to the surface of the wafer. Hereinafter, such etching will be referred to as "scanning etching". In scanning etching, while supplying etchant to the center of the wafer, the etchant flows on the surface of the wafer at the center, thereby controlling the surface shape of the wafer.

[0013] Here, in order to control the surface shape of the wafer after etching, it is crucial to properly control the etching amount distribution (etching profile) in the wafer radial direction. The etching amount distribution is controlled by adjusting the etching conditions (etching recipe) such as the wafer rotation speed (rotation speed), the scanning speed and scanning width of the nozzle during reciprocating movement, to obtain the target etching amount distribution.

[0014] However, in the past, the control of etching amount distribution mainly relied on engineers predicting changes in etching amount distribution and successively obtaining data on etching amount distribution to correspond with the target etching amount distribution, that is, it was done in a trial-and-error manner. In this case, the control of etching amount distribution depends on the engineer's ability, and the required operation time and the degree of control completion may vary from person to person.

[0015] According to the technology of the present invention, the etching amount distribution in the radial direction of the etched object is appropriately predicted during scanning etching. Hereinafter, a wafer processing system and wafer processing method according to the present embodiment will be described with reference to the drawings. Furthermore, in this specification and drawings, elements having substantially the same functional structure are labeled with the same symbols and repeated descriptions are omitted.

[0016] In the wafer processing system 1 described later according to this embodiment, as shown in FIG1, the superimposed wafer T, which is formed by bonding a first wafer W and a second wafer S and serves as a substrate, is processed. Hereinafter, in the first wafer W, the side that is bonded to the second wafer S is referred to as surface Wa, and the side opposite to surface Wa is referred to as back surface Wb. Similarly, in the second wafer S, the side that is bonded to the first wafer W is referred to as surface Sa, and the side opposite to surface Sa is referred to as back surface Sb.

[0017] The first wafer W is, for example, a semiconductor wafer such as a silicon substrate, and a device layer Dw containing a plurality of devices is formed on the surface Wa side. Furthermore, a bonding film Fw is formed on the device layer Dw, and the second wafer S is bonded via the bonding film Fw. As the bonding film Fw, for example, an oxide film (THOX film, SiO2 film, TEOS film), a SiC film, a SiCN film, or an adhesive is used.

[0018] The second wafer S has, for example, the same structure as the first wafer W, and a device layer Ds and a bonding film Fs are formed on its surface Sa. Furthermore, the second wafer S does not need to be a device wafer with the device layer Ds formed; for example, it can be a support wafer supporting the first wafer W. In this case, the second wafer S functions as a protective material protecting the device layer Dw of the first wafer W.

[0019] As shown in Figure 2, the wafer processing system 1 has a configuration that connects the loading / unloading station 2 and the processing station 3 into one unit. The loading / unloading station 2, for example, handles the loading and unloading of a cassette C capable of holding multiple overlapping wafers T between the loading / unloading station and the outside. The processing station 3 is equipped with various processing devices for performing desired processing on the overlapping wafers T.

[0020] The loading / unloading station 2 is equipped with a cassette loading stage 10 that holds, for example, three cassettes C. Furthermore, a wafer transport device 20 is disposed adjacent to the cassette loading stage 10 on the negative X-axis side. The wafer transport device 20 moves freely along a transport path 21 extending in the Y-axis direction. The wafer transport device 20, for example, has two transport arms 22, 22 for holding and transporting overlapping wafers T. Each transport arm 22 moves freely in the horizontal direction, the vertical direction, or around the horizontal and vertical axes. The configuration of the transport arms 22 is not limited to this embodiment and can be any configuration. Furthermore, the wafer transport device 20 can transport overlapping wafers T between the cassettes C of the cassette loading stage 10 and the transfer device 30 described later.

[0021] In the loading and unloading station 2, on the negative X-axis side of the wafer handling device 20, a transfer device 30 is provided adjacent to the wafer handling device 20 for transferring overlapping wafers T between the processing station 3.

[0022] For example, the processing station 3 has three processing blocks B1 to B3. The first processing block B1, the second processing block B2 and the third processing block B3 are arranged sequentially from the positive X-axis direction side (the side of the transfer station 2) to the negative direction side.

[0023] The first processing block B1 is equipped with an etching device 40, a thickness measuring device 41, and a wafer transport device 50. The etching device 40 and the thickness measuring device 41 are stacked. Furthermore, the number and arrangement of the etching device 40 and the thickness measuring device 41 are not limited thereto.

[0024] The etching apparatus 40 etches the back surface Wb (polished surface) of the first wafer W after it has been polished by the processing apparatus 80 described later. While further thinning the polished first wafer W (overlapping wafer T), it removes the polishing marks generated by the polishing process and smooths the polished surface. The detailed configuration of the etching apparatus 40 will be described in detail later.

[0025] In one example, the thickness measuring device 41 includes a measuring unit (not shown) and a calculation unit (not shown). The measuring unit includes a sensor that measures the thickness of the first wafer W after etching at multiple locations. The calculation unit obtains the thickness distribution of the first wafer W from the measurement results (thickness of the first wafer W) of the measuring unit and calculates the flatness (TTV: Total Thickness Variation) of the first wafer W. Furthermore, the calculation of the thickness distribution and flatness of the first wafer W can also be performed by the control device 90 described later, replacing the calculation unit. In other words, the calculation unit (not shown) can be provided within the control device 90 described later. Furthermore, the configuration of the thickness measuring device 41 is not limited to this and can be arbitrarily configured.

[0026] The wafer transport device 50 is disposed on the negative X-axis side of the transfer device 30. The wafer transport device 50, for example, has two transport arms 51, 51 for holding and transporting the overlapping wafer T. Each transport arm 51 can move freely in the horizontal direction, the vertical direction, or around the horizontal and vertical axes. Furthermore, the wafer transport device 50 can transport the overlapping wafer T to the transfer device 30, the etching device 40, the thickness measuring device 41, the cleaning device 60 (described later), the thickness measuring device 61 (described later), and the buffer device 62 (described later).

[0027] A cleaning device 60, a thickness measuring device 61, a buffer device 62, and a wafer transport device 70 are provided in the second processing block B2. The cleaning device 60, the thickness measuring device 61, and the buffer device 62 are arranged in a stacked configuration. Furthermore, the number and arrangement of the cleaning device 60, the thickness measuring device 61, and the buffer device 62 are not limited thereto.

[0028] The cleaning apparatus 60 cleans the back surface Wb (polished surface) of the first wafer W after it has been polished by the processing apparatus 80 described later. For example, a brush is brought into contact with the back surface Wb, and the back surface Wb is brushed and cleaned. Furthermore, pressurized cleaning fluid can also be used in the cleaning of the first wafer W. Also, the cleaning apparatus 60 can be configured to simultaneously clean the back surface Sb of the second wafer S while cleaning the first wafer W.

[0029] In one example, the thickness measuring device 61 includes a measuring unit (not shown) and a calculation unit (not shown). The measuring unit includes a sensor that measures the thickness of the first wafer W after grinding at multiple locations. The calculation unit obtains the thickness distribution of the first wafer W from the measurement results (thickness of the first wafer W) of the measuring unit and calculates the flatness (TTV) of the first wafer W. Furthermore, the calculation of the thickness distribution and flatness of the first wafer W can be performed by the control device 90 described later, replacing the calculation unit. In other words, the calculation unit (not shown) can be provided in the control device 90 described later. Furthermore, the configuration of the thickness measuring device 61 is not limited to this and can be arbitrarily configured.

[0030] The buffer device 62 temporarily holds the overlapping wafer T transferred from the first processing block B1 to the second processing block B2 before processing. The configuration of the buffer device 62 is arbitrary. Furthermore, the buffer device 62 may have a positioning mechanism (not shown) for adjusting the center position of the overlapping wafer T relative to the chuck 83 described later and / or the horizontal orientation of the overlapping wafer T.

[0031] The wafer transport device 70 is, for example, disposed on the positive Y-axis side of the cleaning device 60, the thickness measuring device 61, and the buffer device 62. The wafer transport device 70, for example, has two transport arms 71, 71 that hold and transport the overlapping wafer T using an adsorption and holding surface (not shown). Each transport arm 71 is supported by a multi-jointed arm member 72 and can move freely in the horizontal direction, the vertical direction, or around the horizontal and vertical axes. Furthermore, the wafer transport device 70 can transport the overlapping wafer T to the etching device 40, the thickness measuring device 41, the cleaning device 60, the thickness measuring device 61, the buffer device 62, and the processing device 80 described later.

[0032] A processing device 80 is provided in the third processing block B3. The processing device 80 grinds and thins the first wafer W, and functions as a thinning device of the present invention.

[0033] The processing apparatus 80 includes a rotary table 81. The rotary table 81 rotates freely about a vertical rotation centerline 82 via a rotation mechanism (not shown). Two suction cups 83 are provided on the rotary table 81 to hold and hold the overlapping wafer T. The suction cups 83 are evenly arranged on the same circumference as the rotary table 81. The two suction cups 83 can be moved to the transfer position A0 and the processing position A1 by rotating the rotary table 81. Furthermore, the two suction cups 83 can each rotate about a vertical axis via the rotation mechanism (not shown).

[0034] In transfer position A0, the overlapping wafer T is transferred. A polishing unit 84 is arranged in processing position A1, which polishes the first wafer W while the second wafer S is held in place by a suction cup 83. The polishing unit 84 has a polishing section 85, which has an annular, freely rotatable polishing stone (not shown). Furthermore, the polishing section 85 can move vertically along the support column 86.

[0035] Furthermore, the configuration of the processing apparatus 80 is not limited to this. For example, four suction cups 83 may be provided on the rotary table 81, and these four suction cups 83 may move between the transfer position of the overlapping wafer T, the coarse grinding section (not shown) for coarse grinding of the first wafer W, the intermediate grinding section (not shown) for intermediate grinding of the first wafer W, and the fine grinding section (not shown) for fine grinding of the first wafer W. Furthermore, for example, a thickness measuring device (not shown) for measuring the thickness of the first wafer W after grinding at multiple positions may be provided in the processing apparatus 80.

[0036] The wafer processing system 1 described above includes a control device 90. The control device 90 is, for example, a computer equipped with a CPU and memory, and has a program storage unit (not shown). The program storage unit stores a program for controlling the processing of the overlapping wafer T in the wafer processing system 1. Alternatively, the program may be recorded on a computer-readable recording medium H and installed from that recording medium H onto the control device 90. Furthermore, the recording medium H may be temporary or non-temporary.

[0037] Next, the configuration of the etching apparatus 40 described above will be explained. As shown in FIG3, the etching apparatus 40 includes a wafer holding portion 100 serving as a substrate holding portion, a rotation mechanism 101, and an etching solution supply portion 102.

[0038] The wafer holding portion 100 holds the outer edge of the overlapping wafer T at multiple locations (3 locations in this embodiment). Furthermore, the configuration of the wafer holding portion 100 is not limited to the example shown in the figure; for example, the wafer holding portion 100 may also have a suction cup for adsorbing and holding the overlapping wafer T from below. The rotation mechanism 101 rotates the overlapping wafer T (first wafer W) held in the wafer holding portion 100 about a vertical rotation center line 100a.

[0039] The etchant supply unit 102, for example, has a nozzle for supplying etchant E to the back surface Wb of the first wafer W held in the wafer holder 100. The etchant supply unit 102 is located above the wafer holder 100 and can be moved horizontally and vertically via the moving mechanism 103. In one example, the etchant supply unit 102 can move back and forth (scanning movement) above the center of the first wafer W via the rotation center line 100a of the wafer holder 100, that is, as shown in FIG4. Furthermore, in the following description, the back-and-forth movement of the etchant supply unit 102 is defined as one cycle.

[0040] The etching solution E contains at least hydrofluoric acid, nitric acid, or a mixture of acids to properly etch the silicon of the first wafer W, which is the object of etching. The etching solution E may also contain phosphoric acid or sulfuric acid. Furthermore, the object of etching is not limited to the first wafer W; for example, it may also be amorphous silicon. Furthermore, the object of etching in this embodiment is not limited to the back surface Wb of the first wafer W. For example, it can also be applied to the processing of wafers that have not undergone processing by the processing apparatus 80. For example, when a film is formed on the back surface Wb, that film can also be used as the object of etching.

[0041] In the etching apparatus 40 described above, scanning etching is performed by supplying etching solution to the back surface Wb of the first wafer W from the etching solution supply unit 102 while rotating the first wafer W and moving the etching solution supply unit 102 back and forth. In this embodiment, the etching amount distribution of the scanning etching is predicted using a prediction model.

[0042] The prediction model for the etching amount distribution is shown in Figure 5. It is a model for scanning etching under the etching conditions of fixed scanning speed and left-right symmetrical scanning. That is, the scanning speed V when the etching solution supply section 102 moves back and forth is fixed. Furthermore, the scanning width L of the etching solution supply section 102 is symmetrical from the center of the first wafer W, and the etching solution supply section 102 moves back and forth between one end of the scan "+L" and the other end of the scan "-L". Hereinafter, this scanning etching sequence will be called the "fixed speed scanning sequence", and the derived prediction model will be called the "fixed speed scanning model".

[0043] The constant speed scanning model is composed of the following formulas (1) to (6). ER scan(R) = ER ref(R) × Ratio scan(R)・・・(1) Ratio scan(R) = b 0 × exp(b 1 × T + b 2) + C・・・(2) b 0 = f(S,V)・・・(3) b 1 = f(S,V)・・・(4) b 2 = f(S,V)・・・(5) T = (LR) / V・・・(6) Where, ER scan: the amount of etching when the etchant supply section 102 moves back and forth; ER ref: the amount of etching when the etchant supply section 102 does not move back and forth; Ratio scan: scan ratio; R: the position from the center of the first wafer W; T: the time before the etchant E is not supplied; C: constant; S: the rotation speed when the first wafer W is rotated; V: the scan speed when the etchant supply section 102 moves back and forth; L: The scanning width when the etching solution supply section 102 moves back and forth.

[0044] The functions in equations (3) to (5) above are determined by analytical learning data. For example, a test wafer is etched using a constant-speed scan sequence under different complex etching conditions, and learning data on the etching amount distribution is obtained. Specifically, the test wafer is etched by changing the rotation speed S of the test wafer, the scan speed V when the etchant supply section 102 moves back and forth, and the scan width L of the etchant supply section 102. At this time, the etching processing time for each test wafer is the same.

[0045] The etching of the test wafer under each etching condition is carried out at a predetermined desired time. Furthermore, the etching amount distribution of the test wafer is obtained and output to the control device 90. Moreover, the control device 90 compresses the output etching amount distribution under each etching condition into an etching amount distribution (etching rate) per unit time or per unit cycle, and stores the compressed etching amount distribution as learning data.

[0046] Figure 6 shows an example of the learning data obtained. The rotational speed S of the test wafer is varied from S1 to S5. The scanning speed V of the etchant supply section 102 is varied from V1 to V3. The scanning width L of the etchant supply section 102 is varied from L1 to L4. In this way, the etching amount distribution is obtained for a plurality of etching conditions (60 in this example). In the graphs of each learning data, the horizontal axis represents the radial position from the center of the test wafer (0 (zero) on the horizontal axis) to an outer end, and the vertical axis represents the etching amount (etching rate).

[0047] Furthermore, the above description illustrates an example of obtaining the learning data through etching a test wafer, but the etching object when obtaining the learning data is not limited to the test wafer. Specifically, for example, the etching result of the first wafer W that is actually processed in the wafer processing system 1 can also be stored as the learning data. Also, for example, when a film is formed on the back surface Wb of the first wafer W, the etching object can be set as the film, and the etching result of the film can be stored as the learning data.

[0048] Next, the detailed method for deriving the above-mentioned constant speed scanning model will be explained.

[0049] First, the inventors of this case have mastered the etching amount distribution of scanning etching. Furthermore, they discovered that the etching amount distribution when the etching solution supply unit 102 is fixed and does not move back and forth (hereinafter referred to as "unscanned") becomes the reference etching amount distribution for the etching amount distribution predicted by the constant speed scanning model.

[0050] Figure 7 shows the etching amount distribution in the learning data shown in Figure 6 above, where the rotation speed S is S5, the scanning speed V is V3, and the scanning width L varies from L1 to L4. Figure 7 also shows the reference etching amount distribution before scanning. Referring to Figure 7, in scanning etching, starting from the scanning width L, only the etching amount distribution on the inner periphery changes, while the etching amount distribution on the outer periphery remains unchanged. In other words, in scanning etching, only the etching amount distribution on the inner periphery of the scanning width L changes from the unscanned reference etching amount distribution; the etching amount distribution on the outer periphery of the scanning width L exhibits the same distribution as the reference etching amount distribution. Therefore, the unscanned etching amount distribution can be used as a reference.

[0051] Next, the inventors of this case separated the variation in the etching amount distribution during scanning etching. Specifically, based on the concept of an unscanned reference etching amount distribution, the ratio of the etching amount distribution of the scanning etching to that reference etching amount distribution is defined as the scanning ratio (7) in the following formula. This scanning ratio (7) can quantitatively separate the effect of scanning etching. Figure 8 shows the scanning ratio (7) when the rotation speed S is S5, the scanning speed V varies from V1 to V3, and the scanning width L varies from L1 to L4. The horizontal axis of Figure 8 represents the radial position from the center of the first wafer W (0 (zero) on the horizontal axis) to an outer end, and the vertical axis represents the scanning ratio (7). Referring to Figure 8, the scanning ratio (7) changes on the inner periphery starting from the scanning width L, and the variation in the etching amount distribution of the above-mentioned scanning etching can be quantitatively separated and understood. Then, the following formula (1) of the constant speed scanning model is derived through the following formula (7). Ratio scan(R) = ER scan(R) / ER ref(R)・・・(7) ER scan(R) = ER ref(R) × Ratio scan(R)・・・(1) Where, Ratio scan: scan ratio; ER scan: etching amount when the etching solution supply section 102 moves back and forth; ER ref: etching amount when the etching solution supply section 102 does not move back and forth.

[0052] Furthermore, the unsprayed time T is defined in the following formula (6). The unsprayed time T is the time during which the etchant supply section 102 moves and the etchant E is not supplied to the inner peripheral side through the etchant supply section 102. Moreover, the amount of etching decreases during this unsprayed time T. Figure 9 shows the unsprayed time T when the rotation speed S is S5, the scanning speed V varies from V1 to V3, and the scanning width L varies from L1 to L4. The horizontal axis of Figure 9 represents the radial position from the center of the first wafer W (0 (zero) on the horizontal axis) to an outer end, and the vertical axis represents the unsprayed time T. Referring to Figure 9, the variation of the unsprayed time T on the inner peripheral side can be understood with the scanning width L as the starting point. T=(LR) / V・・・(6) Where, T: the time without etchant E being supplied; V: the scanning speed when the etchant supply section 102 moves back and forth; L: the scanning width when the etchant supply section 102 moves back and forth.

[0053] Next, the inventors of this case studied the relationship between the scan ratio and the non-ejection time T. Figure 10 shows the relationship between the scan ratio and the non-ejection time T when the rotation speed S is S5 and the scan speed V varies from V1 to V3. The horizontal axis of Figure 10 represents the non-ejection time T, and the vertical axis represents the scan ratio. Referring to Figure 10, the scan ratio decreases exponentially relative to the non-ejection time T, and can be defined using the decay curve model of the following formula (2). This decay curve model is also applicable to explaining the operation of the actual physical phenomenon that "the longer the time without etchant E is supplied, the less etching is done". Ratio scan(R)=b0×exp(b1×T+b2)+C・・・(2) Where, b0: scale (intercept); b1: decay rate (slope); b2: decay delay value; C: asymptote.

[0054] The above b0 is the scale (intercept) of the attenuation curve, b1 is the attenuation rate (slope) of the attenuation curve, and b2 is the attenuation delay value of the attenuation curve. For example, b2 is a correction term used when the amount of etchant E is large, the residual liquid in the first wafer W is large and the attenuation does not occur immediately, etc. C is the asymptote of the attenuation curve. For example, when C does not exist, the scan ratio Ratio scan mathematically approaches 0 (zero). But in reality, once etchant E is supplied, the first wafer W will definitely be etched, so the scan ratio Ratio scan is not 0 (zero). C is a correction term used to correct this phenomenon. Also, C can be determined by, for example, first analyzing the above formula (2), if C < 0 then C = 0, if C > 0 then leave C.

[0055] Also, when b2 and C are determined to be unnecessary in practical applications, they can be set as b2=0 and C=0 and omitted.

[0056] Furthermore, the scale b0, the attenuation rate b1, and the attenuation delay value b2 depend on the rotation speed S and the scanning speed V, and can be defined by the following formulas (3) to (5). b0=f(S,V)・・・(3) b1=f(S,V)・・・(4) b2=f(S,V)・・・(5) Where, S: the rotation speed when the first wafer W is rotated; V: the scanning speed when the etchant supply section 102 moves back and forth.

[0057] Here, in Figure 10, the thick line represents the measured value obtained through experiments, and the thin line represents the calculated value calculated using the above formula (2) of the constant speed scanning model. The measured value and the calculated value are roughly consistent, confirming that the attenuation curve model of the above formula (2) is appropriate.

[0058] As above, the constant-speed scanning model (prediction model of etching amount distribution) composed of the following equations (1) to (6) is derived. ER scan(R)=ER ref(R)×Ratio scan(R)・・・(1) Ratio scan(R)=b 0×exp(b 1×T+b 2)+C・・・(2) b 0=f(S,V)・・・(3) b 1=f(S,V)・・・(4) b 2=f(S,V)・・・(5) T=(LR) / V・・・(6)

[0059] Figure 11 is a graph comparing the measured values ​​of the etching amount distribution obtained through experiments (thick lines in Figure 11) with the calculated values ​​of the etching amount distribution calculated by the constant speed scanning model (thin lines in Figure 11). The rotation speed S was varied from S1 to S4, the scanning speed V was varied from V1 to V3, and the scanning width L was varied from L1 to L4. The measured values ​​and the calculated values ​​were roughly consistent, confirming that the constant speed scanning model of the above equations (1) to (6) was appropriate.

[0060] Next, the wafer processing performed using the wafer processing system 1 configured as described above will be described. Furthermore, in this embodiment, a superimposed wafer T is pre-formed on a bonding device (not shown) outside the wafer processing system 1. Also, the peripheral portion of the first wafer W can be pre-removed, for example, a range of 0.5 mm to 3 mm radially from the outer end of the first wafer W.

[0061] First, a cassette C containing a plurality of overlapping wafers T is placed on the cassette loading stage 10 of the loading / unloading station 2. Next, the overlapping wafers T are removed from the cassette C by the wafer transport device 20 and transported to the transfer device 30. The overlapping wafers T transported to the transfer device 30 are then transported to the buffer device 62 by the wafer transport device 50. Furthermore, the center position of the overlapping wafers T relative to the chuck 83 and / or the horizontal orientation of the overlapping wafers T can also be adjusted in the buffer device 62.

[0062] Next, the overlapping wafer T is transported to the processing apparatus 80 via the wafer transport device 70 and transferred to the chuck 83 at the transfer position A0. The back surface Sb of the second wafer S is held in the chuck 83. Next, the chuck 83 is moved to the processing position A1, and the back surface Wb of the first wafer W is polished by the polishing unit 84. Through this polishing process, the thickness of the first wafer W (overlapping wafer T) is reduced to the desired polishing target thickness (step S1 in FIG12).

[0063] Next, the overlapping wafer T is transported to the thickness measuring device 61 via the wafer transport device 70. In the thickness measuring device 61, the thickness of the first wafer W (overlapping wafer T) after grinding is measured at multiple positions to obtain the thickness distribution of the first wafer W after grinding, and the flatness of the first wafer W is calculated (step S2 in FIG. 12). The calculated thickness distribution and flatness of the first wafer W are output to the control device 90, for example. Furthermore, when the processing apparatus 80 is equipped with a thickness measuring device, the thickness of the first wafer W after grinding can also be measured using the thickness measuring device of the processing apparatus 80.

[0064] In the control device 90, the optimal etching conditions for subsequent etching processing are determined based on the thickness distribution and flatness of the first wafer W output (step S3 in FIG12). The detailed method for determining the optimal etching conditions of the control device 90 will be described in detail later.

[0065] The overlapping wafer T, whose thickness has been measured for the first wafer W, is then transported to the cleaning apparatus 60 via the wafer transport device 70 or the wafer transport device 50. In the cleaning apparatus 60, the polished surface (i.e., the back surface Wb) of the first wafer W is cleaned (step S4 in FIG. 12). Furthermore, the back surface Sb of the second wafer S can also be cleaned in the cleaning apparatus 60 as described above. Also, when the thickness is measured using the thickness measuring device 61 in this embodiment, the order of steps S2, S3, and S4 can be reversed. That is, after cleaning the back surface Wb of the first wafer W in the cleaning apparatus 60, the thickness of the first wafer W can be measured in the thickness measuring device 61, and the optimal etching conditions for the etching process can be determined.

[0066] Next, the overlapping wafer T is transported to the etching apparatus 40 via the wafer transport device 50. In the etching apparatus 40, the polished surface of the first wafer W, i.e. the back surface Wb, is etched by the etching solution E under optimal etching conditions (step S5 in FIG12).

[0067] When etching the first wafer W, firstly, the wafer holding part 100 (the first wafer W) is rotated around the vertical rotation center line 100a, and at the same time, the etching solution E is supplied (sprayed) from the etching solution supply part 102, and the etching of the back side Wb begins.

[0068] Furthermore, during the etching of the first wafer W, etchant E is continuously supplied from the etchant supply unit 102. Simultaneously, as shown in FIG4, the etchant supply unit 102 is moved back and forth (scanned) around the rotation center line 100a above the rotation center of the first wafer W. The detailed method for determining etching conditions such as the rotation speed of the first wafer W, the scanning speed of the etchant supply unit 102 during its back-and-forth movement, and the scanning width of the etchant supply unit 102 will be described in detail later.

[0069] After the desired etching amount is obtained for the first wafer W, the supply of etching solution E from the etching solution supply unit 102 is stopped, and the back surface Wb of the first wafer W is washed with pure water and then spun dry. Then, the rotation of the wafer holding unit 100 (the first wafer W) is stopped, and the etching of the first wafer W is ended.

[0070] Here, the optimal etching conditions for the first wafer W are determined as described above based on the thickness distribution and flatness of the first wafer W after grinding. Specifically, the optimal etching conditions are determined based on the difference between the measured values ​​of the thickness distribution and flatness of the first wafer W by the thickness measuring device 61 and the thickness distribution and flatness of the target surface shape of the etched first wafer W (hereinafter referred to as "target shape"). Furthermore, in step S5, by etching the first wafer W under the optimal etching conditions, the difference between the measured value and the target value of the thickness of the first wafer W is removed by etching, thereby processing the surface of the first wafer W into the target shape. In this way, through this embodiment, regardless of the surface shape of the first wafer W after grinding, the target surface shape of the first wafer W can be appropriately obtained.

[0071] Next, the overlapping wafer T is transported to the thickness measuring device 41 via the wafer transport device 50. In the thickness measuring device 41, the thickness of the first wafer W (overlapping wafer T) after etching is measured at multiple locations to obtain the thickness distribution of the first wafer W after etching, and the flatness of the first wafer W is calculated (step S6 in FIG. 12). The calculated thickness distribution and flatness of the first wafer W are output to the control device 90, for example, and used for processing other overlapping wafers T that are subsequently processed in the wafer processing system 1. Furthermore, when the thickness of the first wafer W after grinding is measured by the thickness measuring device of the processing device 80, the thickness of the first wafer W after etching can also be measured in the thickness measuring device 61.

[0072] Then, the overlapping wafer T, which has undergone all processing, is transported via the transfer device 30 to the cassette C of the cassette stage 1. Thus, the series of wafer processing in the wafer processing system 1 is completed.

[0073] Next, the detailed method for determining the above-mentioned optimal etching conditions will be explained (step S3 in Figure 12).

[0074] First, when determining the optimal etching conditions, learning data is obtained before processing the overlapping wafer T in the wafer processing system 1 (step S3-1 in Figure 13). A prediction model for the etching amount distribution is derived from this learning data (step S3-2 in Figure 13).

[0075] In step S3-1, as described above, for example, a constant-speed scan sequence of etching is performed on the test wafer to obtain learning data on the etching amount distribution as shown in Figure 6.

[0076] In step S3-2, as described above, a prediction model for the etching amount distribution, which is composed of the following equations (1) to (6), is derived, namely, the constant speed scanning model. At this time, the functions in the following equations (3) to (5) are determined by analyzing the learning data obtained in step S3-1. ER scan(R) = ER ref(R) × Ratio scan(R)・・・(1) Ratio scan(R) = b 0 × exp(b 1 × T + b 2) + C・・・(2) b 0 = f(S,V)・・・(3) b 1 = f(S,V)・・・(4) b 2 = f(S,V)・・・(5) T = (LR) / V・・・(6) Where, ER scan: the amount of etching when the etchant supply section 102 moves back and forth; ER ref: the amount of etching when the etchant supply section 102 does not move back and forth; Ratio scan: scan ratio; R: the position from the center of the first wafer W; T: the time before the etchant E is not supplied; C: constant; S: the rotation speed when the first wafer W is rotated; V: the scan speed when the etchant supply section 102 moves back and forth; L: The scanning width when the etching solution supply section 102 moves back and forth.

[0077] Here, Figure 14 is a graph showing the distribution of each etching amount when performing steps S3-3 to S3-8 described later to determine the optimal etching conditions. The horizontal axis of Figure 14 represents the radial position from the center of the first wafer W (0 (zero) on the horizontal axis) to an outer end, and the vertical axis represents the etching amount (etching rate).

[0078] In parallel with steps S3-1 and S3-2, the first target etching amount distribution of the etching process in step S5 above is obtained (step S3-3 in FIG13). The first target etching amount distribution is obtained based on the thickness distribution of the target shape of the first wafer W after etching (hereinafter referred to as "target thickness distribution") and the thickness distribution of the surface shape of the first wafer W after grinding obtained in step S2 above (hereinafter referred to as "measured thickness distribution"). The first target etching amount distribution can be obtained, for example, by calculating the difference between the target thickness distribution and the measured thickness distribution of the first wafer W. This first target etching amount distribution is represented by a solid line in FIG14. Furthermore, in this example, the first target etching amount distribution is flat within the wafer surface.

[0079] Next, to minimize the first residual distribution between the first etch amount distribution calculated using the constant-speed scanning model in step S3-2 and the first target etch amount distribution obtained in step S3-3, the first etch conditions of the first etch amount distribution are optimized using the least squares method (step S3-4 in FIG13). The optimized first etch conditions include the rotational speed S of the first wafer W, the scanning speed V of the etchant supply section 102, and the scanning width L of the etchant supply section 102. Furthermore, the first etch amount distribution corresponding to the optimized first etch conditions is represented by a dotted chain line in FIG14.

[0080] Next, the first residual distribution in step S3-4 is set as the second target etching amount distribution (step S3-5 in Figure 13).

[0081] Next, to minimize the second residual distribution between the second etch amount distribution calculated using the constant-speed scanning model in step S3-2 and the second target etch amount distribution set in step S3-5, the second etch conditions of the second etch amount distribution are optimized using the least squares method (step S3-6 in FIG13). The optimized second etch conditions include the rotational speed S of the first wafer W, the scanning speed V of the etchant supply section 102, and the scanning width L. Furthermore, the second etch amount distribution corresponding to the optimized second etch conditions is represented by a two-point chain line in FIG14.

[0082] Next, the time ratio (cycle count ratio) corresponding to the first etching amount distribution and the second etching amount distribution is linked (step S3-7 in FIG13). The time ratio (cycle count ratio) is the ratio of the time (cycle count) of performing constant-speed scan sequence 1 under the first etching condition to the time (cycle count) of performing constant-speed scan sequence 2 under the second etching condition. The linked etching amount distribution (hereinafter referred to as the "linked etching amount distribution") is represented by a dotted line in FIG14.

[0083] In this example, the etching amount distribution of the connection is roughly consistent with the etching amount distribution of the first target. Therefore, the etching conditions corresponding to this etching amount distribution of the connection are determined as the optimal etching conditions. Specifically, the time ratios corresponding to the first etching conditions and the second etching conditions are combined and determined as the optimal etching conditions (steps S3-8 in Figure 13).

[0084] Then, in step S4, after cleaning the back surface Wb of the first wafer W, in step S5, the back surface Wb of the first wafer W is etched under optimal etching conditions. That is, in the etching apparatus 40, while the overlapping wafer T (first wafer W) is rotated at a rotation speed determined by the optimal etching conditions, the etching solution supply unit 102 is moved at a determined scanning speed and scanning width, and the etching solution E is supplied to the first wafer W.

[0085] As above, the optimal etching conditions according to this embodiment are determined, and the first wafer W is etched based on the optimal etching conditions.

[0086] Through the above embodiments, the constant-speed scanning model composed of the above equations (1) to (6) can be used to appropriately predict the etching amount distribution. Therefore, when controlling the etching amount distribution, unlike in the past when it depended on the engineer's ability, the required operation time for control can be suppressed and the control completion rate can be improved. In addition, the deviation of the operation amount when determining the optimal etching conditions in step S3 can be suppressed, thereby improving the accuracy of the optimal etching conditions.

[0087] Furthermore, since the optimal etching conditions are determined using the least squares method in step S3, the first wafer W can be etched under these optimal etching conditions in the subsequent step S5. This allows the etching amount distribution of the etching process to approximate the first target etching amount distribution, resulting in the surface shape of the etched first wafer W becoming the target shape. In other words, the optimal etching conditions can be determined from uncertain etching conditions, thereby appropriately controlling the surface shape of the etched first wafer W.

[0088] Through actual simulations conducted by the inventors, regardless of whether the first target etching amount distribution is V-shaped, A-shaped, M-shaped, or W-shaped, the thickness distribution deviation of the etched first wafer W can be kept within the allowable range. Furthermore, the flatness (TTV) of the etched first wafer W is also improved compared to the past. Specifically, the V-shaped distribution has a smaller etching amount at the center of the first wafer W than at both ends, and appears as a slightly V-shaped distribution in a graph with wafer position as the horizontal axis and etching amount as the vertical axis. The A-shaped distribution has a larger etching amount at the center of the first wafer W than at both ends, and appears as a slightly A-shaped distribution in the aforementioned graph, with a shape opposite to the V-shaped distribution. The M-shaped distribution in the aforementioned graph consists of two A-shaped shapes sandwiching the center of the first wafer W, and has an overall slightly M-shaped distribution. The W-shape in the above diagram refers to the arrangement of two V-shapes sandwiching the center of the first wafer W, and the overall distribution has a slightly W-shaped shape.

[0089] Furthermore, steps S1 to S6 are performed on each overlapping wafer T, so even if the surface shape of each wafer is different before etching (after grinding in this embodiment), the surface shape of the first wafer W after etching can still be controlled to the target shape one by one.

[0090] In the above embodiments, the etching conditions were optimized twice using the minimum flat method in steps S3-4 and S3-6, but the number of optimizations is not limited to this.

[0091] For example, the optimization calculation using the least squares method can also be performed once. For example, after optimizing the first etching condition in step S3-4, if the first etching amount distribution corresponding to the first etching condition is approximately consistent with the first target etching amount distribution, the first etching condition can be determined as the optimal etching condition. In this case, steps S3-5 to S3-8 can be omitted.

[0092] For example, the optimization calculation using the least squares method can be performed more than three times. For example, if the connection etching amount distribution in step S3-8 is not roughly consistent with the first target etching amount distribution, the third residual distribution between this connection etching amount distribution and the first target etching amount distribution can be set as the third target etching amount distribution. Then, steps S3-6 to S3-8 are performed to make the connection etching amount distribution roughly consistent with the first target etching amount distribution. In this way, the residual distribution between the connection etching amount distribution and the first target etching amount distribution is gradually reduced, and the optimization calculation using the least squares method is repeated until the connection etching amount distribution is roughly consistent with the first target etching amount distribution. Then, the etching conditions at the time points when the connection etching amount distribution and the first target etching amount distribution are roughly consistent are combined and determined as the optimal etching conditions.

[0093] The above embodiments illustrate various processing methods applied to the back surface Wb of the first wafer W in a composite wafer T formed by bonding the first wafer W and the second wafer S. However, the processing targets are not limited to this. For example, thinning and etching processes can be performed on a single wafer. The processing target can also be a film formed on the wafer surface, such as an oxide film and titanium nitride. In this case, the etching solution supply unit 102 of the etching apparatus 40 can switch the supply of different types of etching solutions E according to the etching target. Furthermore, for example, when a film formed on the wafer surface is used as the etching target, the etching results of the film can be stored as the aforementioned learning data. The thickness of the film is measured in the thickness measuring device 61. Furthermore, when a protective tape is attached to the component side of the wafer, thinning and etching processes can be performed on the side opposite to the protective tape. Moreover, thinning and etching processes can also be performed on wafers that have been cut from a die by a wire saw or similar tool and polished. Regardless of the object being processed, etching can be performed under the optimal etching conditions described above.

[0094] Furthermore, for example, when a film is formed on the back surface Wb of the first wafer, this film can be used as the etching target. In this case, for example, the thickness measuring device 61 measures the thickness of the film, and in step S2, the thickness of the film is measured instead of the thickness of the first wafer W, and the thickness distribution and flatness of the film are calculated. Then, in step S3, the optimal etching conditions are determined based on the calculated thickness distribution and flatness of the film.

[0095] Furthermore, the wafer processing system 1 includes various devices other than the etching device 40, but the device configuration applicable to this invention is not limited to this. For example, the thinning device, i.e., the processing device 80, may also be omitted. In this case, the object of etching is not limited to the wafer after thinning. Furthermore, for example, the technology of this invention can also be applied when etching the wafer in a separate etching device.

[0096] In the above embodiments, the first wafer W is thinned by the processing apparatus 80, but the thinning method is not limited to this. For example, the thinning process of the first wafer W also includes polishing the back surface Wb of the first wafer W. Or, for example, the modified layer (not shown) formed inside the first wafer W by laser processing can be used as a base point for separation and thinning. In this case, the wafer processing system 1 is provided with a laser processing apparatus (not shown) for forming the modified layer (not shown), instead of the processing apparatus 80.

[0097] It should be understood that all embodiments of the present invention are illustrative and not intended to limit. The above embodiments may be omitted, substituted, or modified in various forms without departing from the scope and spirit of the appended patent applications. [Simplified Explanation of the Diagram]

[0008] Figure 1 is a side view showing an example of overlapping wafers processed in a wafer processing system. Figure 2 is a top view schematically showing the structure of a wafer processing system. Figure 3 is a side view schematically showing the structure of an etching apparatus. Figure 4 is a schematic diagram showing the radial movement of the etching solution supply section. Figure 5 is a schematic diagram showing a constant-speed scanning sequence. Figure 6 is a schematic diagram showing an example of learning data. Figure 7 is a schematic diagram showing an example of learning data and a reference etching amount distribution. Figure 8 is a schematic diagram showing the radial distribution of the scan ratio. Figure 9 is a schematic diagram showing the radial distribution of the non-ejection time. Figure 10 is a schematic diagram showing the relationship between the scan ratio and the non-ejection time. Figure 11 is a schematic diagram comparing the measured value of the etching amount distribution of the constant-speed scanning sequence with the calculated value of the etching amount distribution of the constant-speed scanning model. Figure 12 is a flowchart showing the main steps of wafer processing. Figure 13 is a flowchart showing the main steps of the method for determining optimal etching conditions. Figure 14 is a schematic diagram showing the etching amount distribution of each step in determining optimal etching conditions.

Claims

1. A substrate processing method, comprising the steps of: determining optimal etching conditions; and, based on the optimal etching conditions, while rotating the substrate to be etched, a liquid etching supply unit moves back and forth in the radial direction above the rotation center of the etched object, while supplying etchant from the liquid etching supply unit to the surface of the etched object to etch the surface; The step of determining the optimal etching conditions comprises the steps of: obtaining learning data including the radial etching amount distribution of the etched object after etching the surface of the etched object under a plurality of different etching conditions; and, optimizing the first etching conditions of the first etching amount distribution by using the least squares method in a manner that minimizes the first residual distribution between the first etching amount distribution predicted by the first etching amount distribution and the first target etching amount distribution using the learning data; ERscan(R)=ERref(R)×Ratioscan(R)・・・(1) Ratioscan(R)=b0×exp(b1×T+b2)+C・・・(2) b0=f(S,V)・・・(3) b1=f(S,V)・・・(4) b2=f(S,V)・・・(5) T=(LR) / V・・・(6) Wherein, ERscan: Etching amount when the etchant supply section moves back and forth; ERref: Etching amount when the etchant supply section does not move back and forth; Ratioscan: Scan ratio; R: Position from the center of the etched object; T: Time before the etching solution is supplied and before spraying; C: Constant; S: Rotational speed when the object to be etched is rotated; V: Scanning speed when the etching solution supply unit moves back and forth; L: Scanning width when the etching solution supply unit moves back and forth.

2. The substrate processing method as described in claim 1, wherein, The optimal etching conditions include the rotation speed, the scanning speed, and the scanning width.

3. The substrate processing method as described in claim 1 or 2 further includes the following steps: setting the first residual distribution as the second target etch amount distribution; optimizing the second etch condition of the second etch amount distribution by using the least squares method in a manner that minimizes the second residual distribution of the second etch amount distribution predicted by the second etch amount distribution using the learning data through the following formulas (1) to (6); and combining the optimized first etch condition and the optimized second etch condition according to the time ratio, and determining it as the optimal etch condition.

4. The substrate processing method as described in claim 1 or 2, wherein, The functions in equations (3) to (5) above are determined by analyzing the learning materials.

5. The substrate processing method as described in claim 1 or 2 further includes the following steps: before etching the surface of the etched object, measuring the thickness of the etched object and obtaining the thickness distribution of the etched object in the radial direction; and predicting the first etching amount distribution based on the obtained thickness distribution of the etched object.

6. The substrate processing method as described in claim 1 or 2 further includes the following steps: before etching the surface of the etched object, measuring the thickness of the etched object and obtaining the thickness distribution of the etched object in the radial direction; the first target etching amount distribution is obtained based on the obtained thickness distribution of the etched object and the target thickness distribution of the etched object.

7. The substrate processing method as described in claim 5 further includes the following step: thinning the substrate before measuring the thickness of the etched object.

8. A substrate processing system for processing a substrate, comprising: an etching apparatus that, while rotating an etchable object on the substrate, moves an etching solution supply unit back and forth in the radial direction above the rotation center of the etchable object, and supplies etching solution from the etching solution supply unit to the surface of the etchable object to etch the surface; and a control device that controls the etching of the surface of the etchable object by the etching apparatus based on optimal etching conditions; the control device performing the following steps: acquiring learning data including the radial etching amount distribution of the etchable object after etching the surface of the etchable object under a plurality of different etching conditions; and optimizing the first etching conditions of the first etching amount distribution by using the least squares method in a manner that minimizes the first residual distribution between the first etching amount distribution predicted by the first etching amount distribution and the first target etching amount distribution using the learning data; ERscan(R)=ERref(R)×Ratioscan(R)・・・(1) Ratioscan(R)=b0×exp(b1×T+b2)+C・・・(2) b0=f(S,V)・・・(3) b1=f(S,V)・・・(4) b2=f(S,V)・・・(5) T=(LR) / V・・・(6) Where, ERscan: Etching amount when the etchant supply section moves back and forth; ERref: Etching amount when the etchant supply section does not move back and forth; Ratioscan: Scan ratio; R: Position from the center of the etched object; T: Time before the etching solution is supplied and before spraying; C: Constant; S: Rotational speed when the object to be etched is rotated; V: Scanning speed when the etching solution supply unit moves back and forth; L: Scanning width when the etching solution supply unit moves back and forth.

9. The substrate processing system as described in claim 8, wherein, The optimal etching conditions include the rotation speed, the scanning speed, and the scanning width.

10. The substrate processing system as described in claim 8 or 9, wherein, The control device further performs the following steps: setting the first residual distribution as the second target etching amount distribution; optimizing the second etching condition of the second etching amount distribution by using the least squares method in a manner that minimizes the second residual distribution of the second etching amount distribution predicted by the second etching amount distribution using the learning data through the following formulas (1) to (6); and combining the optimized first etching condition and the optimized second etching condition according to the time ratio, and determining it as the optimal etching condition.

11. The substrate processing system as described in claim 8 or 9, wherein, The control device analyzes the learning data and determines the functions in equations (3) to (5) above.

12. The substrate processing system as claimed in claim 8 or 9 further includes: a thickness measuring device for measuring the thickness of the etched object before etching; and a control device for predicting the first etching amount distribution based on the thickness distribution of the etched object obtained from the thickness of the etched object measured by the thickness measuring device.

13. The substrate processing system as described in claim 8 or 9 further includes: a thickness measuring device for measuring the thickness of the etched object before etching; and a control device for obtaining the first target etching amount distribution based on the thickness distribution of the etched object obtained from the thickness of the etched object measured by the thickness measuring device and the target thickness distribution of the etched object.

14. The substrate processing system as claimed in claim 12 further includes: a thinning apparatus for thinning the substrate; and a thickness measuring apparatus for measuring the thickness of the etched object on the thinned substrate.

Images