A wafer positioning detection method for semiconductor wet process equipment

Abstract

The present invention discloses a wafer positioning detection method for semiconductor wet process equipment, comprising the following steps: detection algorithm data calibration, a wafer box with wafers of known position is placed in a box; a sensor scans the wafer box at a set speed and direction, and records the time of each trigger signal; the acquired trigger signal is calculated and processed; in a calibration state, the interval between two adjacent layers of the wafer box and the position of the first layer of wafer are calculated according to the first processed signal and the relative position of the known wafer; entering a working state, a wafer box with wafers of unknown position is placed in a box, the sensor scans the wafer box at a set speed and direction, and records the time of each trigger signal; the acquired trigger signal is calculated and processed; and the relative position information of all wafers in the wafer box is obtained. The present invention solves the problem that wafers placed obliquely in a wafer box with a fluid medium cannot be accurately positioned and detected, and provides accurate positioning detection for subsequent wafer grabbing.

Description

Technical Field

[0001] The invention belongs to the technical field of wafer positioning detection, and in particular relates to a wafer positioning detection method for semiconductor wet process equipment. Background Art

[0002] In existing wet process applications, some equipment wafers use a "wet-in, wet-out" transmission solution, that is, the wafer box containing the wafers is placed in a tank filled with liquid, and the robot takes and places the wafers from the liquid. In existing practical cases, the equipment does not perform positioning detection on the wafers, and the position and quantity of the wafers are manually input into the controller. This method has the problem of asymmetry between the input information and the actual wafer information in the cassette, which may cause a collision when the robot takes and places the wafer, thereby damaging the robot or the wafer. Therefore, it is necessary to effectively detect the wafer status in the wafer box to ensure that the robot can accurately and safely take out the wafer and carry it.

[0003] In addition, in the above-mentioned "wet-in, wet-out" transmission scheme, there are the following difficulties in realizing automatic detection of wafer position and quantity: (1) In order to ensure the stability of the wafer position in the liquid and facilitate transmission, the wafer box is usually opened upward and the wafer is placed at an angle, which makes wafer positioning detection difficult; (2) After the placement or loading of the wafer box, the wafer will shake in the liquid, resulting in position deviation of the sensor's detection signal and wafer position calculation error; (3) The wafer is subject to buoyancy in the liquid or suction is generated between the wafer and the wafer box, resulting in uneven position of each layer of wafers in the wafer box, making it difficult to detect the wafer position. The calculation of the position becomes difficult (4) The wafer is in the liquid and the sensor is outside the liquid. The sensor is easily disturbed by the ambient light and the fluctuation of the liquid surface, which will also cause the sensor signal to have a position deviation; (5) Some wafer substrate materials are highly transparent and thin, such as SiC or glass. The sensor signal is difficult to trigger, the triggering time is short, and it is more susceptible to the influence of liquid fluctuations, resulting in position deviation of the sensor signal and difficulty in position calculation; (6) In order to detect wafers with high transparency and thin thickness, the sensor light intensity is usually increased or the threshold is lowered, which makes the sensor susceptible to the reflection of light from other objects, such as the wafer box base and water waves. (7) Conventional wafer presence detection sensors cannot perform effective wafer detection when the medium changes. For example, some sensors cannot be detected in a water environment, or they can be detected in a water environment, but due to the change of the medium, the light propagation path changes and cannot be stably and accurately detected. Summary of the invention

[0004] In order to overcome the deficiencies of the prior art, the present invention provides a wafer positioning detection method for semiconductor wet process equipment, which can adapt to wafer positioning detection in a fluid medium, has accurate detection, and can adapt to different types of wafer positioning detection.

[0005] The technical solution adopted by the present invention to solve the technical problem is: a wafer positioning detection method for semiconductor wet process equipment, comprising the following steps:

[0006] Detection algorithm data calibration, a wafer box with a known position of the wafer is placed in a box body, the box body has a built-in fluid medium, and the wafer box has a multi-layer structure for placing the wafer;

[0007] The sensor scans the cassette at the set speed and direction and records the time of each trigger signal;

[0008] The acquired trigger signal is processed to remove the interference signal caused by the fluid medium;

[0009] In the calibration state, the interval time between two adjacent layers of the cassette and the signal trigger time of the first layer of wafer are calculated according to the first processed signal after removing the interference signal and the relative position of the known wafer, and the interval distance between two adjacent layers of the cassette and the position of the first layer of wafer are calculated according to the sensor scanning speed;

[0010] Entering the working state, the cassette containing the wafer at the unknown position is placed in the box, which has a built-in fluid medium. The sensor scans the cassette at the set speed and direction and records the time of each trigger signal;

[0011] The acquired trigger signal is processed to remove the interference signal caused by the fluid medium;

[0012] The second processed signal after removing the interference signal in the working state is converted into wafer position information according to the sensor scanning speed, and is calculated with the spacing between two adjacent layers of the wafer box and the position of the first layer of wafers obtained in the calibration state to obtain the relative position information of all wafers in the wafer box. The position information of the wafers in the wafer box is transmitted to the controller for controlling the transmission device to grab the wafer.

[0013] Further, the trigger signal is a time series signal, which is a rising edge, or, it is a falling edge, or it is a combination of a rising edge and a falling edge.

[0014] Furthermore, the step of calculating the interval between two adjacent layers of the wafer box and the position of the first layer of wafers includes the following sub-steps:

[0015] Calculating the weighted average of at least a plurality of rising edge or falling edge time differences of the processed first processed signal to obtain the interval time between two adjacent layers of the film box;

[0016] Calculate the difference between all rising edges or falling edges of the first processed signal after processing and the interval time of the corresponding number of layers, and then take the weighted average of all the differences to obtain the first layer wafer time;

[0017] The interval between two adjacent layers of the cassette and the time of the first layer of wafer are multiplied by the sensor scanning speed in the calibration state to obtain the interval between two adjacent layers of the cassette and the position of the first layer of wafer.

[0018] Furthermore, the step of calculating the interval between two adjacent layers of the wafer box and the position of the first layer of wafers includes the following sub-steps:

[0019] Place the film box within the detection range of the sensor;

[0020] Obtain the distance Xa between the wafer placement surface in the cavity slot layer structure of the wafer box terminal and the edge of the wafer box frame terminal, and obtain the distance Xt between the placement surfaces of adjacent cavity slot layer structures;

[0021] The sensor scans the wafers in the cassette at a specific speed and records the time corresponding to the edge trigger signal at the terminal edge of the cassette frame;

[0022] Record all wafer trigger signals and convert them into the position Xx of the wafer position information relative to the terminal edge of the cassette frame according to the sensor scanning speed;

[0023] Determine whether Xx is between (Xa+(k-1)Xt) and (Xa+kXt), where k is the number of layers, and thus determine the number of layers of the cavity slot structure where the wafer is located;

[0024] Alternatively, the position Xk of all theoretical wafer placement surfaces relative to the cassette frame is calculated, and if Xx is between Xk and Xk+1, it can be determined that the number of layers of the cavity slot layer structure where the wafer is located is K layers.

[0025] Furthermore, the interference signal includes a water wave interference signal and a reflected light interference signal, and the water wave interference signal includes water surface interference and / or wafer shaking interference.

[0026] Further, the step of removing the interference signal includes the following sub-steps:

[0027] Compare the time difference of the adjacent rising edges or falling edges of each group of trigger signals to determine whether it is less than the median value of all adjacent rising edges or falling edges multiplied by the set coefficient. If the result is less than, filter out the trigger signal;

[0028] and / or,

[0029] Compare the time differences between the adjacent rising and falling edges of each group of trigger signals to determine whether they are greater than the median value of all adjacent rising and falling edge time differences multiplied by the set coefficient. If the result is greater than, filter out the trigger signal.

[0030] Furthermore, the step of obtaining the relative position of the wafer in the wafer box under the working state includes the following sub-steps:

[0031] Multiply all trigger signal times by the sensor scanning speed in the working state to obtain all trigger signal positions;

[0032] Subtract the calibrated position of the first layer of wafers from the position of all trigger signals, divide by the calibrated spacing between two adjacent layers of the cassette, round the result to the nearest integer, and add one to obtain the relative positions of all wafers in the cassette.

[0033] Furthermore, the signal emitted by the sensor is parallel or tends to be parallel to the crystal plane direction, ensuring that the vertical distance between the sensor and the edge of each wafer in the cassette is equal or tends to be equal during the movement.

[0034] Furthermore, the sensor is an optical sensor and is arranged in an area outside the fluid medium of the box.

[0035] Furthermore, the wafer substrate material is an opaque material, which is Si; or, the wafer substrate material is a translucent material, which is SiC or glass; the wafer thickness is 0.3 mm-0.8 mm, and the wafer diameter is 6 inches-12 inches.

[0036] Furthermore, the wafer is vertically placed in a box containing a fluid medium, the angle between the wafer surface and the liquid surface is 45°-90°, and the wafer is completely submerged in the fluid medium.

[0037] Furthermore, the fluid medium is a transparent liquid, which is pure water or a sodium hydroxide solution.

[0038] Furthermore, the sensor scanning direction is a direction of increasing number of layers, or a direction of decreasing number of layers; the sensor scanning speed in the calibration state is the same as or different from the sensor scanning speed in the working state.

[0039] Furthermore, the wafer box can store 50, 25, 13 or 12 wafers.

[0040] Furthermore, the step of the sensor scanning the film box at a set speed and direction may be performed before the film box is loaded into the working position of the transmission device corresponding to the box, or after the film box is loaded into the working position of the transmission device corresponding to the box, or during the process of loading the film box into the working position of the transmission device corresponding to the box.

[0041] Furthermore, a liquid level sensor is provided in the box body, which is used to monitor the change of liquid level in the box body and the amount of change of the liquid level.

[0042] The present invention also discloses a semiconductor wet process equipment, comprising:

[0043] A box body, in which a fluid medium is placed;

[0044] The wafer box is placed in the box and has a plurality of cavity slot layer structures for placing wafers;

[0045] The sensor can move relative to the wafer box and detect the position information of the wafer according to the above detection method.

[0046] The beneficial effects of the present invention are: 1) It solves the problem that the wafer cannot be accurately positioned and detected when it is tilted and placed in a cassette with a fluid medium, and provides accurate positioning detection for subsequent wafer grabbing; 2) It solves the problem of water waves, reflected light influence and large wafer position deviation, and can realize positioning detection of wafers of different sizes and types in water; 3) The equipment runs automatically, the detection efficiency is high, and human setting errors are reduced. 4) The sensor above the fluid medium can determine the presence or absence of the wafer, the tilted wafer or the stacked wafer state. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic diagram of the box body of the present invention.

[0048] Figure 2 It is a flow chart of the wafer positioning detection method of the present invention.

[0049] Figure 3 This is the sensor signal feedback diagram affected by water wave interference.

[0050] Figure 4 Sensor signal feedback diagram affected by other reflected light after reducing the number of wafers.

[0051] Figure 5 A brief description of the film box of the present invention Figure 1 .

[0052] Figure 6 A brief description of the film box of the present invention Figure 2 .

[0053] Figure 7 It is a partial schematic diagram of the film box of the present invention.

[0054] Figure 8 This is a simplified diagram of the principle of distinguishing abnormal pieces.

[0055] Among them, 1-wafer, 2-wafer box, 21-placing surface of cavity slot layer structure, 22-terminal edge of wafer box, 3-box, 4-sensor, 5-fluid medium, 6-liquid level sensor. DETAILED DESCRIPTION

[0056] In order to enable those skilled in the art to better understand the solutions of the present invention, the following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0057] A wafer positioning detection method for semiconductor wet process equipment comprises the following steps:

[0058] The detection algorithm data is calibrated, and a wafer box containing wafers at known positions is placed in a box body 3, and a fluid medium 5 is built in the box body 3. The wafer box 2 has a multi-layer structure for placing wafers 1;

[0059] In this embodiment, the wafer 1 is vertically placed in a box 3 containing a fluid medium 5, specifically, the wafers 1 are vertically placed one by one in the wafer box 2, the angle between the wafer surface and the liquid surface is 45°-90°, and the fluid medium 5 completely covers the wafer 1; the fluid medium 5 is a transparent liquid, which can be pure water or a sodium hydroxide solution; the substrate material of the wafer 1 is an opaque material, such as Si, and the substrate material of the wafer 1 can also be a translucent material, such as SiC, glass, etc.; the thickness of the wafer 1 is 0.3mm-0.8mm, and the diameter of the wafer 1 is 6 inches-12 inches; the number of layers of the wafer box 2 is 50, or 25, or 13, or 12; the step of placing the wafer box 2 in the box 3 can be to place the wafer box 2 directly at the working position of the transmission device corresponding to the box 3 (specifically, the position where the transmission device takes and places the wafer), or to place the wafer box 2 at the transfer position of the box 3, and then load it into the working position of the transmission device corresponding to the box 3 (specifically, the position where the transmission device takes and places the wafer). The transmission device may specifically be a robot.

[0060] The sensor 4 scans the film box 2 at the set speed and direction, and records the time of each trigger signal;

[0061] In the above steps, the trigger signal is a time series signal, which may be a rising edge, or it may be a falling edge, or it may be a combination of a rising edge and a falling edge.

[0062] In the above steps, the scanning direction of the sensor 4 is the direction of increasing the number of layers, or the direction of decreasing the number of layers, and there is no specific limitation, as long as the scanning starts from the end side of the film box 2; the sensor 4 is arranged in the area outside the fluid medium 5 of the box body 3, and the sensor 4 can be an optical sensor, and the signal emitted by the sensor 4 is parallel or tends to be parallel to the crystal plane direction of the wafer 1;

[0063] The step of the sensor 4 scanning the film box 2 at the set speed and direction may be performed before the film box 2 is loaded into the working position of the box body 3 corresponding to the transmission device, or after the film box 2 is loaded into the working position of the box body 3 corresponding to the transmission device, or during the process of loading the film box 2 into the working position of the box body 3 corresponding to the transmission device. The acquired trigger signal is calculated and processed to remove the interference signal caused by the fluid medium 5;

[0064] The interference signals in the above steps include water wave interference signals and reflected light interference signals. The water wave interference signals include simple water surface interference, simple wafer shaking interference, and combined interference of water surface interference and wafer shaking interference.

[0065] The water wave interference signal filtering is not limited to the preliminary hardware condition filtering when the water surface interference signal is out of the sensor detection range; the reflected light interference signal filtering is not limited to the wafer box mounting base material is not a strongly reflective dark substrate.

[0066] For the above steps, the step of removing interference signals includes the following sub-steps, which aims at removing dense small-segment interference signals caused by water waves and wafer shaking. The time difference of adjacent rising edges or falling edges of each group of trigger signals is compared to determine whether it is less than the median value of all adjacent rising edges or falling edges multiplied by the set coefficient. If the result is less than, the trigger signal is filtered out; the above-mentioned set coefficient can specifically be 0.5.

[0067] Of course, in the above steps, the step of removing interference signals may also include the following sub-steps. For the removal of interference signals of large sections of reflected light such as the base, the time difference between adjacent rising edges and falling edges of each group of trigger signals is compared to determine whether it is greater than the median value of all adjacent rising edge and falling edge time differences multiplied by a set coefficient. If the result is greater than, the trigger signal is filtered out; the above-mentioned set coefficient can specifically be 1.5.

[0068] Of course, it can also be a combination of the above two filtering methods, that is, comparing the time difference of adjacent rising edges or falling edges of each group of trigger signals to determine whether it is less than the median value of all adjacent rising edge or falling edge time differences multiplied by the set coefficient. If the result is less than, the trigger signal is filtered out; at the same time, comparing the time difference of adjacent rising edges and falling edges of each group of trigger signals to determine whether it is greater than the median value of all adjacent rising edge and falling edge time differences multiplied by the set coefficient. If the result is greater than, the trigger signal is filtered out.

[0069] In the calibration state, the interval time between two adjacent layers of the cassette 2 and the signal trigger time of the first layer of wafer are calculated based on the first processed signal after removing the interference signal and the relative position of the known wafer 1. In order to adapt to the detection calculation under different scanning speeds and increase the application range of the algorithm, the time unit is converted into a distance unit, and the interval distance between two adjacent layers of the cassette 2 and the position of the first layer of wafer are calculated according to the scanning speed of the sensor 4;

[0070] Specifically, the above steps include calculating the weighted average of at least multiple groups of rising edge or falling edge time differences of the first processed signal after processing to obtain the interval time between two adjacent layers of the wafer box 2; calculating the difference between all rising edges or falling edges of the first processed signal after processing and the interval time of the corresponding number of layers, and then taking the weighted average of all differences to obtain the first layer wafer time; multiplying the interval time between two adjacent layers and the first layer wafer time by the scanning speed of the sensor 4 in the calibration state to obtain the interval distance between two adjacent layers of the wafer box 2 and the position of the first layer wafer.

[0071] Entering the working state, the cassette 2 containing the wafer 1 at an unknown position is placed in the box 3, the box 3 has a fluid medium 5 built in it, the sensor 4 scans the cassette 2 at the set speed and direction, and records the time of each trigger signal;

[0072] In the above steps, the scanning speed of the sensor 4 may be the same as or different from the scanning speed of the sensor 4 in the calibration state, and the direction must be consistent with that in the calibration state.

[0073] The acquired trigger signal is calculated and processed to remove the interference signal caused by the fluid medium 5;

[0074] The interference signal and the method of filtering out the interference signal in the above steps are the same as those in the calibration steps and will not be described in detail.

[0075] The second processed signal after removing the interference signal in the working state is converted into the position information of the wafer 1 according to the scanning speed of the sensor 4, and is calculated with the interval distance between two adjacent layers of the wafer box 2 and the position of the first layer of the wafer obtained in the calibration state to obtain the relative position of the wafer 1 in the wafer box 2;

[0076] For the above steps, the step of obtaining the relative position of the wafer 1 in the wafer box 2 under the working state includes the following sub-steps: multiplying all the trigger signal times by the scanning speed of the sensor 4 under the working state to obtain all the trigger signal positions; subtracting the calibrated first layer wafer position from all the trigger signal positions, and then dividing by the calibrated interval distance between two adjacent layers of the wafer box 2, rounding to the nearest integer, and then adding one to obtain the relative position information of all the wafers 1 in the wafer box 2.

[0077] The position signal of the wafer 1 in the cassette 2 is transmitted to the controller for controlling the transmission device, which may specifically be a robot arm to grab the wafer 1 .

[0078] In order to ensure that the position of the wafer 1 in the fluid medium 5 is as stable as possible and convenient for transmission, the film box 2 is generally opened upward and placed at an angle. At this time, the relative position of the base of the film box 2 and the sensor 4 is not convenient for structural calibration. In order to reduce the difficulty of detection and simplify the detection structure, the present invention places the wafer at a specified position and performs a pre-scan, that is, calibrates the detection algorithm data to obtain a signal of a known position to calculate the spacing distance between two adjacent layers of the film box and the position of the first layer of wafer. The method of the present invention no longer requires that the position of the sensor and the wafer remain fixed, eliminating the step of calibrating the position structure of the two. In addition, there is no need to re-calibrate the detection algorithm data when changing the transparency and thickness of the wafer. Even if the specifications of the film box (wafer spacing, quantity changes) or the wafer size are changed, it is sufficient to re-perform the pre-scan calibration without changing the detection structure and algorithm parameters, thereby increasing the scope of application of the algorithm.

[0079] Specifically, Figure 3 As shown, even if the wafer 1 is placed at an angle, the position of the wafer 1 will float within a certain range due to the buoyancy and fluctuation of the fluid medium 5, and the interval of the wafer 1 will have a large difference. For example, the detection algorithm data calibration directly uses the first group of detection signals as the time of the first layer of wafers, and any adjacent signal difference is used as the interval time between two adjacent layers of the cassette, which is susceptible to interference, resulting in distortion of the calibration data and calculation errors of the wafer position in the working state. In order to solve the above problems, it is considered to increase the amount of data and reduce the calibration data deviation caused by the floating position of the wafer by calculating the average value. However, in order to shorten the calibration time and avoid using a large number of wafers 1 for calibration, it is found through testing and calculation that by placing wafers in the first, second, twenty-fourth, and twenty-fifth cavity slot layer structures of the cassette 2 for calibration scanning to obtain signals, the weighted average value of the head and tail signal interval differences is calculated to obtain the time of the first layer of wafers and the interval time between two adjacent layers of the cassette, which is converted into the theoretical position of each layer of wafers. The deviation from the wafer position obtained by multiple groups of actual scanning is small, and the wafer position can be accurately calculated, solving the problem that the wafer intervals in water are inconsistent and cannot be accurately positioned.

[0080] like Figure 4 As shown, the sensor signal changes from 0 to 1 as a rising edge, and changes from 1 to 0 as a falling edge. The rising edge time is recorded as Tr1, Tr2...Trn, and the falling edge time is recorded as Td1, Td2...Tdn.

[0081] Depend on Figure 3It can be seen that since the wafer is placed in a fluid medium, it is affected by the reflected light of the water wave and the shaking of the wafer, generating multiple short signals, which affect the detection. To filter this interference, an algorithm is added. The difference between the rising edge signals of each adjacent group is compared with the median value Trm of the time difference [Tr(n) - Tr(n - 1)] of all adjacent rising edge signals multiplied by the coefficient Trx. If [Tr(n) - Tr(n - 1)] < Trm * Trx, then the signals Tr(n) and Td(n) of this group are filtered out, and the signals Tr(n - 1) and Td(n - 1) are retained. Through multiple groups of tests, the ratio of the length of the short interference signal segment to the length of the normal wafer signal is calculated, the maximum value is taken, and 0.2 is added to obtain Trx. Substituting it into the calculation can effectively remove the short interference signal segment and retain the normal wafer signal.

[0082] From Figure 4 It can be seen that since the wafer 1 is placed in water, after reducing the wafer 1, the light of the sensor 4 directly irradiates the bottom of the cassette 2, and it is easily affected by the reflected light of the base of the cassette 2, generating a continuous long interference signal segment, which affects the detection. To filter this interference, an algorithm is added. The difference between the rising edge signal Tr(n) and the falling edge signal Td(n) of each adjacent group is compared with the median value Tdm of the difference [Td(n) - Tr(n)] between all adjacent rising edge signals and falling edge signals multiplied by the coefficient Tdx. If [Td(n) - Tr(n)] > Tdm * Tdx, then the signals Tr(n) and Td(n) of this group are filtered out. Through multiple groups of tests, the ratio of the length of the long interference signal segment to the length of the normal wafer signal is calculated, the minimum value is taken, and 0.5 is subtracted to obtain Tdx. Substituting it into the calculation can effectively remove the long interference signal segment and retain the normal wafer signal.

[0083] Adding the above two sets of filtering algorithms can effectively remove the dense short interference signal segments caused by water waves and wafer shaking and the long reflection light interference signal segments such as the cassette base. However, if the filtering intensity is too high, it is easy to remove the normal wafer signal. Therefore, by adjusting the sizes of Tdx and Trx, the filtering effect is appropriately weakened, and the similar signals are filtered again through the final step to ensure the relative accuracy of the data.

[0084] More specifically, a wafer positioning detection method for a semiconductor wet equipment includes the following steps:

[0085] Step 1, calibration scanning. Place wafers on the first layer, the second layer, the twenty-fourth layer, and the twenty-fifth layer of the cassette. Place the cassette in a box filled with liquid, perform the cassette loading box operation, and move the cassette to the pick-and-place position of the box transfer device. The sensor scans the cassette above the liquid at a speed of V1 in the direction of increasing layer numbers, and sequentially records the rising edge and falling edge times of all signals.

[0086] Step 2, signal processing. Compare the difference between each group of adjacent rising-edge signals with the median value Trm of all adjacent rising-edge signal time differences [Tr(n) - Tr(n - 1)] multiplied by the coefficient Trx. If [Tr(n) - Tr(n - 1)] < Trm * Trx, then filter out the signals of Tr(n) and Td(n) in this group and retain the signals of Tr(n - 1) and Td(n - 1). Compare the difference between each group of adjacent rising-edge signal Tr(n) and falling-edge signal Td(n) with the median value Tdm of all adjacent rising-edge signal and falling-edge signal differences [Td(n) - Tr(n)] multiplied by the coefficient Tdx. If [Td(n) - Tr(n)] > Tdm * Tdx, then filter out the signals of Tr(n) and Td(n) in this group.

[0087] Step 3, calibration value calculation. Denote the times of the processed rising-edge signals as t1, t2, t24, t25 in sequence. Calculate the average time interval Tt between two adjacent layers of the cassette and the average trigger time Tf of the first layer of wafers. The calculation formulas are as follows:

[0088] Tt = ((t25 - t1) / 24 * wt1 + (t24 - t1) / 23 * wt2 + (t25 - t2) / 23 * wt3 + (t24 - t2) / 22 * wt4) / (wt1 + wt2 + wt3 + wt4),

[0089] Tf = ((t1 * wf1 + (t2 - Tt) * wf2 + (t24 - Tt * 23) * wf3 + (t25 - Tt * 24) * wf4) / (wf1 + wf2 + wf3 + wf4),

[0090] Among them, the weighting coefficients wt1, wt2, wt3, wt4, wf1, wf2, wf3, wf4 are empirical values and can all be 1.

[0091] Convert the time unit to the distance unit. The sensor scanning speed is V1. The calculation formula is as follows:

[0092] The distance Xt between two adjacent layers of the cassette = Tt * V1;

[0093] The position Xf of the first layer of wafers = Tf * V1;

[0094] Enter the working state.

[0095] Step 4, working scan. According to the actual working conditions, place the wafers arbitrarily in the cassette, place the cassette in a box filled with liquid, perform the loading action, and move the cassette to the pick-and-place position of the transfer device. The sensor scans the cassette above the liquid at a speed of V2 in the direction of increasing layer number, and sequentially records the rising-edge and falling-edge times of all signals.

[0096] Step 5: Signal processing: Execute step 2 again, and record the processed rising edge signals as T1, T2, ..., Tn in sequence.

[0097] Step six, wafer position calculation. Multiply all rising edge signals Tn by the sensor scanning speed V2 in turn to convert the time unit into the distance unit. Subtract the calibrated first layer wafer position Xf to get the difference Xx, divide Xx by the calibrated distance between two adjacent layers of the cassette Xt, round it off and add 1 to get the actual position number of the wafer 1. The calculation formula is as follows: Nn = ROUND((Tn*V2-Xf) / Xt)+1. After all calculations are completed, they are packaged into an array and sent to the host computer. If N1=1, N2=2, N3=2, N4=24, and N5=25 are calculated, 1100000000000000000000011 is sent, indicating that there are wafers in the 1st, 2nd, 24th, and 25th cavity slot layers of the cassette, and there are no wafers in the remaining cavity slot layers.

[0098] Through the above data calculation method, similar signals are converted into the same result so as to process the data again and avoid duplication of results.

[0099] In summary, through steps two and five, algorithms are added to filter out most of the signals generated by water waves and interference from the cassette base. Combined with steps one and three, the number of calibrated wafers is increased, and the weighted average of multiple groups of differences is calculated to solve the problems of wafer position floating and inconsistent intervals. Appropriate calibration values ​​are obtained as the basis for judging the subsequent wafer positions. Finally, through step five, similar signals are calculated and filtered out to obtain the relative position information of all wafers in the cassette.

[0100] The steps can also be combined into two steps, calibration scan calculation and working scan calculation. Under normal operation, a calibration scan calculation can be performed once when the equipment is first run, and working scan calculations can be performed continuously thereafter.

[0101] In the calibration state, according to the first processed signal after removing the interference signal, although the wafer box 2 is tilted and fixed, and (in an ideal state) the wafer is close to the placement surface, considering that the wafer is placed in the wafer box 2 with a fluid medium, there is a situation where gravity is not enough to completely overcome the buoyancy of the fluid, and the wafer in the wafer box 2 will have a slight tilt to the left and right or front and back. This situation will also exist in the first layer of wafers. The signal trigger time needs to be effectively processed. For this purpose, a fixed reference point in the fluid medium can be found for data analysis.

[0102] When the frame of the film box 2 is within the detection range of the sensor 4, the frame of the film box 2 is used as a fixed reference point in the fluid medium. The sensor 4 moves from the terminal edge 22 of the film box to the placement surface 21 of the cavity groove layer structure of the terminal cavity of the film box 2. Assuming that the tilt direction of the film box is high on the left and low on the right, the edge 22 of the film box frame is the highest position, and the scanning direction of the sensor 4 is consistent with the tilt direction of the film box 2. The sampling time increases, the sensor signal is triggered to a high level, and the signal is not triggered to a low level, that is, the distance between the lower surface of the frame of the film box 2 and the lower surface of the wafer 1 adjacent to the frame of the film box 2 (the table surface, that is, the wafer placement surface) is Xa. According to the interval distance Xt between two adjacent layers of the film box, the scanning speed V2 of the sensor 4 is used to scan the wafer 1 in the film box 2. After completing a scan, the time T0 corresponding to the trigger signal of the lower edge of the frame of the film box 2 is recorded, and the rising edge Tr of all trigger pulse signals is recorded. X (X=1,2,.....,n) and all pulse trigger signal lower edge Td X (X=1,2,.....,n), at this time the width M of each trigger pulse can be obtained X =[Td X -Tr X ],(X=1,2,.....,n).

[0103] In order to adapt to the detection calculation under different scanning speeds and expand the application range of the algorithm, the time unit is converted into the distance unit. According to the time T0 corresponding to the trigger signal of the lower edge of the frame of the film box 2, the corresponding position X0=T0*V2.

[0104] Then, based on the distance Xa between the frame of the film box 2 and the placement surface 21 of the cavity groove layer structure at the starting end, and the distance between the placement surfaces 21 of the adjacent cavity groove layer structures is approximately Xt, the position Xk of all the table surfaces relative to the frame of the film box 2 can be calculated, and then the specific layer number of the cavity groove layer structure where the wafer trigger signal is located can be determined based on the position Xx of all wafers relative to the frame of the film box 2.

[0105] More specifically, the following steps are included:

[0106] The theoretical position Xk of the placement surface 21 of the cavity groove layer structure of each film box 2 is obtained.

[0107] Xk (k=1,2,.....,25)=(X0+Xa)+(k-1)*Xt, where k is the number of layers of the inner cavity groove structure of the wafer box 2.

[0108] The position Xx of the lower edge signal Tdx (wafer bottom surface) of each trigger signal pulse obtained by scanning relative to the frame of the cassette is obtained.

[0109] Xx=[(Tdx-T0)*V2], (x=1,2,...,n)

[0110] Among them, Xx is the coordinate scale value of the lower surface of the wafer, T0 is the time value of the initial reference position of the wafer box 2 (the frame position of the wafer box 2), Tdi is the time value relative to the initial reference position when the extraction detection sensor 4 scans the lower surface of the wafer, and V2 is the movement speed of the motion device.

[0111] Determine the position of the wafer trigger signal in the cavity slot, determine whether Xx-X0 is between (Xa+(k-1)Xt) and (Xa+kXt), k is the number of layers, or determine that Xx is between Xk and Xk+1, then it can be determined that the number of layers of the cavity slot layer structure where the wafer is located is K layers.

[0112] After determining the specific number of layers of the cavity slot structure where the trigger signal is located, determine whether the wafer is placed in a single layer, stacked (stacked), or tilted (tilted).

[0113] First, we need to judge the tilted wafer. Since the tilted wafer may tilt left and right or front and back, the position Xx corresponding to the trigger signal Tdi when the detection sensor 4 scans the lower surface of the wafer is not near the theoretical wafer placement surface Xk. Figure 8 . Determine whether the position Xx corresponding to the trigger signal Tdi is located at X k-1 +Jm to X k -Jm near ( Figure 8 Jm is obtained from experience and is used to divide the critical area. If it is within this range, it is considered to be tilted placement, otherwise it is stacked placement or single-piece placement.

[0114] If Xx is between X k + / -Jm, the actual wafer thickness Mi(mm)*V2 is obtained according to the width and speed of each trigger pulse, which is used to determine whether the wafer is placed in a stacked or single wafer manner.

[0115] The above calculation obtains the spacing distance between two adjacent layers of cavity slots in the wafer box and the first layer wafer position step or the frame position of the wafer box 2. In the embodiment, it is not limited to one type of wafer box, and it can be different wafer boxes. Then the distance between the placement surfaces of adjacent wafer cavity slot layer structures will be different, and the value of the spacing Xt can be modified.

[0116] Of course, the fluid medium 5 can also be a non-transparent liquid. The wafer 1 increases the freedom of vertical lifting and lowering. The above algorithm can also be used to ensure that the medium between the wafer 1 and the sensor 4 is the same.

[0117] Specifically, when the above algorithm is adopted, sensor 4 is a reflective line laser sensor, which performs wafer detection outside the medium. The reflective structure enables sensor 4 to be located outside the medium for detection in a stable air environment, and the line laser has a large signal contact area and can obtain more effective signals, which solves the instability caused by the medium and fills the gap in detailed and effective detection of the wafer status in the liquid.

[0118] A liquid level sensor 6 is provided in the box 3, which is used to monitor the change of the liquid level in the box 3 and the amount of change of the liquid level. Specifically, when the wafer 1 is taken out of the cassette 2, the liquid level in the box 3 decreases, and when a new wafer 1 is placed in the cassette 2, the liquid level in the box 3 increases. By calculating the increase and decrease of the liquid level, the in-and-out situation of the wafer 1 can be determined, and by calculating the increase and decrease of the liquid level, the number of wafers 1 in and out can be determined.

[0119] The above specific implementation modes are used to explain the present invention rather than to limit the present invention. Any modification and change made to the present invention within the spirit of the present invention and the protection scope of the claims shall fall within the protection scope of the present invention.

Claims

1. A wafer positioning detection method for semiconductor wet process equipment, characterized in that: The following steps are included: Detection algorithm data calibration, a wafer box with a known position of the wafer is placed in a box body, the box body has a built-in fluid medium, and the wafer box has a multi-layer structure for placing the wafer; The sensor scans the cassette at the set speed and direction and records the time of each trigger signal; The signal emitted by the sensor is parallel to or tends to be parallel to the crystal plane; the trigger signal is a time series signal, which is a rising edge, or a falling edge, or a combination of a rising edge and a falling edge; The acquired trigger signal is processed to remove the interference signal caused by the fluid medium; In the calibration state, the interval time between two adjacent layers of the cassette and the signal trigger time of the first layer of wafer are calculated according to the first processed signal after removing the interference signal and the relative position of the known wafer, and the interval distance between two adjacent layers of the cassette and the position of the first layer of wafer are calculated according to the sensor scanning speed; Entering the working state, the cassette containing the wafer at the unknown position is placed in the box, which has a built-in fluid medium. The sensor scans the cassette at the set speed and direction and records the time of each trigger signal; The acquired trigger signal is processed to remove the interference signal caused by the fluid medium; The second processed signal after removing the interference signal in the working state is converted into wafer position information according to the sensor scanning speed, and is calculated with the spacing between two adjacent layers of the wafer box and the position of the first layer of wafers obtained in the calibration state to obtain the relative position information of all wafers in the wafer box. The position information of the wafers in the wafer box is transmitted to the controller for controlling the transmission device to grab the wafer.

2. The wafer positioning detection method for semiconductor wet process equipment according to claim 1, characterized in that: The step of calculating the distance between two adjacent layers of the wafer box and the position of the first layer of wafers includes the following sub-steps: Calculating the weighted average of at least a plurality of rising edge or falling edge time differences of the processed first processed signal to obtain the interval time between two adjacent layers of the film box; Calculate the difference between all rising edges or falling edges of the first processed signal after processing and the interval time of the corresponding number of layers, and then take the weighted average of all the differences to obtain the first layer wafer time; The interval between two adjacent layers of the cassette and the time of the first layer of wafer are multiplied by the sensor scanning speed in the calibration state to obtain the interval between two adjacent layers of the cassette and the position of the first layer of wafer.

3. The wafer positioning detection method for semiconductor wet process equipment according to claim 1, characterized in that: The step of calculating the distance between two adjacent layers of the wafer box and the position of the first layer of wafers includes the following sub-steps: Place the film box within the detection range of the sensor; Obtain the distance Xa between the wafer placement surface in the cavity slot layer structure of the wafer box terminal and the edge of the wafer box frame terminal, and obtain the distance Xt between the placement surfaces of adjacent cavity slot layer structures; The sensor scans the wafers in the cassette at a specific speed and records the time corresponding to the edge trigger signal at the terminal edge of the cassette frame; Record all wafer trigger signals and convert them into the position Xx of the wafer position information relative to the terminal edge of the cassette frame according to the sensor scanning speed; Determine whether Xx is between (Xa+(k-1)Xt) and (Xa+kXt), where k is the number of layers, and thus determine the number of layers of the cavity slot structure where the wafer is located; Alternatively, the position Xk of all theoretical wafer placement surfaces relative to the cassette frame is calculated, and if Xx is between Xk and Xk+1, it can be determined that the number of layers of the cavity slot layer structure where the wafer is located is K layers.

4. The wafer positioning detection method for semiconductor wet process equipment according to claim 1, characterized in that: The interference signal includes a water wave interference signal and a reflected light interference signal, and the water wave interference signal includes water surface interference and / or wafer shaking interference.

5. The wafer positioning detection method for semiconductor wet process equipment according to claim 4, characterized in that: The step of removing interference signals comprises the following sub-steps: Compare the time difference of the adjacent rising edges or falling edges of each group of trigger signals to determine whether it is less than the median value of all adjacent rising edges or falling edges multiplied by the set coefficient. If the result is less than, filter out the trigger signal; and / or, Compare the time differences between the adjacent rising and falling edges of each group of trigger signals to determine whether they are greater than the median value of all adjacent rising and falling edge time differences multiplied by the set coefficient. If the result is greater than, filter out the trigger signal.

6. The wafer positioning detection method for semiconductor wet process equipment according to claim 1, characterized in that: The step of obtaining the relative position of the wafer in the wafer box under the working state includes the following sub-steps: Multiply all trigger signal times by the sensor scanning speed in the working state to obtain all trigger signal positions; Subtract the calibrated position of the first layer of wafers from the position of all trigger signals, divide by the calibrated spacing between two adjacent layers of the cassette, round the result to the nearest integer, and add one to obtain the relative positions of all wafers in the cassette.

7. The wafer positioning detection method for semiconductor wet process equipment according to claim 1, characterized in that: The sensor is an optical sensor and is arranged in an area outside the fluid medium of the box.

8. The wafer positioning detection method for semiconductor wet process equipment according to claim 1, characterized in that: The wafer substrate material is a light-impermeable material, which is Si; or, the wafer substrate material is a light-transmitting material, which is SiC or glass; The wafer thickness is 0.3mm-0.8mm and the wafer diameter is 6 inches-12 inches.

9. The wafer positioning detection method for semiconductor wet process equipment according to claim 1, characterized in that: The wafer is placed vertically in a box containing a fluid medium, the angle between the wafer surface and the liquid surface is 45°-90°, and the wafer is completely submerged in the fluid medium.

10. The wafer positioning detection method for semiconductor wet process equipment according to claim 1, characterized in that: The fluid medium is a transparent liquid, which is pure water or a sodium hydroxide solution.

11. The wafer positioning detection method for semiconductor wet process equipment according to claim 1, characterized in that: The sensor scanning direction is a direction of increasing number of layers, or a direction of decreasing number of layers; the sensor scanning speed in the calibration state is the same as or different from the sensor scanning speed in the working state.

12. The wafer positioning detection method for semiconductor wet process equipment according to claim 1, characterized in that: The wafer box can store 50, 25, 13 or 12 wafers.

13. The wafer positioning detection method for semiconductor wet process equipment according to claim 1, characterized in that: The step of the sensor scanning the film box at a set speed and direction may be performed before the film box is loaded into the working position of the transmission device corresponding to the box, or after the film box is loaded into the working position of the transmission device corresponding to the box, or during the process of loading the film box into the working position of the transmission device corresponding to the box.

14. The wafer positioning detection method for semiconductor wet process equipment according to claim 1, characterized in that: The box body is provided with a liquid level sensor, which is used to monitor the change of the liquid level in the box body and the amount of change of the liquid level.

15. A semiconductor wet process equipment, characterized in that: include, A box body, in which a fluid medium is placed; The wafer box is placed in the box and has a plurality of cavity slot layer structures for placing wafers; The sensor can move relative to the wafer box and detect the position information of the wafer according to any one of the detection methods of claims 1-14.

Images