Motion control method and system for non-graphic wafer defect detection

Abstract

The invention discloses a motion control method and system for non-graphic wafer defect detection, and relates to the technical field of semiconductor detection motion control. The method comprises the following steps: constructing motion trail models R (t), theta (t), V (t) and W (t) of a straight line of a wafer to be measured at the current moment t, generating a lookup table of a relation between target positions and speeds of an R axis and a theta axis and the current moment, and a theta axis encoder increment count value and a trigger delay value for triggering a TDI camera, and then storing the lookup table, generating an equidistant photographing trigger signal following the phase of the motion track; and receiving encoder position information of the R-axis of the linear motor and the theta-axis of the rotating motor. According to the invention, the synchronization problem of the equidistant spiral motion trail rotating motor and the execution motor can be effectively solved, the problem of motion trail phase asynchronism caused by traditional equal-duration triggering and the problem of image distortion of a shot object caused by non-equidistant triggering are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor detection motion control, and in particular to a motion control method and system for non-patterned wafer defect detection. Background Art

[0002] With the continuous development of semiconductor technology, its manufacturing process is becoming more and more complex, semiconductor chips are gradually developing in the direction of miniaturization, and the detection accuracy of non-patterned wafer defect detection equipment is also constantly improving. The accuracy of motion control directly affects the operating stability of the motion mechanism and the quality of image acquisition, which in turn directly affects the accuracy and recognition rate of defect detection.

[0003] The wafer without pattern to be tested needs to achieve equidistant spiral motion through linear motion in the R direction and rotational motion in the θ direction at the same time, so that the entire wafer surface can pass through the collection objective lens of the TDI camera completely. The defect detection of the wafer without pattern requires that the linear velocity of the wafer in the θ direction through the TDI must be the same when the light intensity remains unchanged and the spot length in the R direction must be kept equidistant. Therefore, it is necessary to solve the problem of synchronization between the rotation motor in the θ direction and the linear motor in the R direction.

[0004] However, an important basis for motion control accuracy is the reliability of the encoder signal. The motion control system needs to provide a signal to trigger the TDI camera to take pictures. The generation of this signal needs to be strictly synchronized with the actual movement and the interval of each generation must be equal in arc length in the θ direction to ensure that the collected image is not distorted. Therefore, it is necessary to solve the phase tracking problem of the trigger signal and the position feedback encoder signal. To this end, we propose a motion control method and system for non-patterned wafer defect detection. Summary of the invention

[0005] The object of the present invention is to provide a motion control method and system for non-patterned wafer defect detection, which can solve the phase tracking problem between a trigger signal and a position feedback encoder signal.

[0006] To achieve the above object, the present invention provides the following technical solution: a motion control method for non-patterned wafer defect detection, specifically comprising the following steps:

[0007] Construct the motion trajectory model R(t), θ(t), V(t), W(t) of the wafer to be tested at the current time t, generate a lookup table of the relationship between the R-axis and θ-axis target position and speed and the current time, as well as the θ-axis encoder incremental count value and trigger delay value for TDI camera triggering, and then store the lookup table;

[0008] Receive the position information of the linear motor R axis and the rotary motor θ axis encoder;

[0009] According to the R-axis encoder position information, the desired position lookup table generated by the motion trajectory model R(t), and the feedforward torque information, the R-axis PID control is completed and the linear motor is driven to move;

[0010] According to the θ-axis encoder position information, the desired angle lookup table generated by the motion trajectory model θ(t), and the corresponding feedforward torque information, the θ-axis PID control is completed and the rotary motor is driven to move;

[0011] According to the lookup table data CNT enc 、CNT clk (n) and the R-axis and θ-axis encoder position information construct a DPLL digital phase-locked loop module, so that the generated TDI trigger signal follows the phase of the actual motion position information.

[0012] Furthermore, the motion trajectory model R(t), θ(t), V(t), and W(t) of the wafer to be measured at the current time t is constructed as follows:

[0013]

[0014] Where, the initial scan line width is d 0 , the wafer radius is r 0 , the linear velocity of the scanning line spot is v 0 ;

[0015] From the above formula (1) and the initial time t = 0, R (0) = r 0 ,available:

[0016]

[0017] Where R(t) is the position of the wafer inspection line along the R direction at time t, V R (t) is the running speed of the linear motor at time t;

[0018] From the above, we can get the motion control trajectory model of the rotating axis: (4) and (5):

[0019]

[0020] According to equations (2) and (3), the target position and speed lookup table required for motion control of linear motion can be generated. According to equations (4) and (5), the target position and speed lookup table required for motion control of rotary motion can be generated. Rotary motion and linear motion are associated through equation (1) to achieve dual-axis motion synchronization.

[0021] Furthermore, the incremental count value and trigger delay value of the θ-axis encoder for triggering the TDI camera are generated as follows:

[0022] The target trigger frequency of the TDI camera is known to be f0 , the corresponding period is T 0 , the resolution of the θ axis encoder is θ 0 ;

[0023] The TDI triggering time is: T(n) = {T 0 ,2T 0 ,3T 0 ,…,nT 0}

[0024] The θ angle corresponding to each TDI trigger is: {θ(T 0 ),θ(2T 0 ),θ(3T 0 ),…,θ(nT 0 )}

[0025] Correspondingly, the incremental count value of the θ encoder corresponding to each trigger of TDI is:

[0026]

[0027] In order to further improve the TDI triggering accuracy, the value obtained by CNT(n) is rounded down to get the integer part, which is the integer cycle of the encoder at the TDI triggering moment:

[0028]

[0029] The decimal part is the error at the current resolution:

[0030]

[0031] From {X(n)}, according to R(n) the radius of the current spot position, the linear velocity v of the spot movement 0 , and the FPGA system clock frequency f clk , we can get the f required for the motion {X(n)} clk Number of clock cycles:

[0032]

[0033] According to equations (7) and (9), the lookup table data CNT required for TDI triggering is generated. enc and CNT clk (n), used to generate equidistant TDI trigger target signals.

[0034] Furthermore, according to the R-axis encoder position information, the desired position lookup table generated by the motion trajectory model R(t), and the feedforward torque information, the R-axis PID control is completed and the linear motor is driven to move. The specific method is as follows:

[0035] (1) The host computer generates a lookup table of the R-axis position and speed relative to the expected position and expected speed of the time according to the R-axis motion trajectory model, and saves it;

[0036] (2) A feedforward lookup table generated by the host computer based on the measured speed, position, and torque data;

[0037] (3) receiving the decoded real-time position information of the R-axis encoder, and further converting the real-time position information of the R-axis encoder to obtain the speed information at the current moment;

[0038] (4) The expected position information R(t) at time t obtained from the expected position lookup table and the actual position information R_ENC(t) obtained from the encoder are combined to complete the position closed-loop PID control. The result of the position closed-loop is used as the input of the speed closed-loop, and then combined with the real-time speed information V_ENC(t) calculated by R_ENC(t) to complete the speed closed-loop PID control; the output of the speed closed-loop control is then integrated with the feedforward torque data and output as the final torque to the R-axis driver to complete the R-axis motor drive control.

[0039] Furthermore, according to the θ-axis encoder position information, the desired angle lookup table generated by the motion trajectory model θ(t), and the corresponding feedforward torque information, the θ-axis PID control is completed and the rotary motor is driven to move. The specific method is as follows:

[0040] (1) The host computer generates a lookup table of the expected position and expected speed of the θ-axis relative to time based on the θ-axis motion trajectory model and saves it;

[0041] (2) A feedforward lookup table generated by the host computer based on the measured speed, position, and torque data;

[0042] (3) receiving the decoded real-time position information of the θ-axis encoder, and further converting the real-time position information of the θ-axis encoder to obtain the speed information at the current moment;

[0043] (4) The expected position information θ(t) at time t obtained from the expected position lookup table and the actual position information θ_ENC(t) obtained by the encoder are combined to complete the position closed-loop PID control. The result of the position closed-loop is used as the input of the speed closed-loop, and then combined with the real-time speed information V_ENC(t) calculated by θ_ENC(t) to complete the speed closed-loop PID control; the output of the speed closed-loop control is then integrated with the feedforward torque data as the final torque output to the R-axis driver to complete the θ-axis motor drive control.

[0044] Furthermore, the DPLL digital phase-locked loop includes a LUT_TDI_TRIG module, a digital phase detector module, a loop filter module, a digital control oscillator module and a TDI clock generation module, so that the generated TDI trigger signal follows the phase of the position information of the actual motion, as follows:

[0045] When a clock pulse appears, the counter is +1, and the upper limit of the counter is the frequency division coefficient. When the number of the counter is less than 1 / 2 of the upper limit, it outputs 1. When the number of the counter is greater than 1 / 2 of the upper limit, it outputs 0. When the number of the counter exceeds the upper limit, the counter returns to zero, so that the frequency division of the clock source can be achieved;

[0046] The clock signal output by the digital control oscillator module is sent to the TDI clock generation module, and the TDI clock generation module uses the frequency division coefficient N to divide the clock of the digital control oscillator module to generate a TDI trigger signal.

[0047] According to a second aspect of the present invention, the present invention provides a motion control system for unpatterned wafer defect detection, which is used to implement the above-mentioned motion control method for unpatterned wafer defect detection, comprising:

[0048] A motion trajectory generating unit is used to construct a motion trajectory model R(t), θ(t), V(t), and W(t) of the wafer line to be measured at the current time t;

[0049] A motion control processing unit, wherein the motion control processing unit includes an FPGA encoder module, an ARM processor module and an EMMC storage module, wherein the FPGA encoder module and the EMMC storage module are both electrically connected to the ARM processor module;

[0050] The linear motor and position feedback unit are used to complete the R-axis PID control and drive the linear motor to move according to the R-axis encoder position information, the desired position lookup table generated by the motion trajectory model R(t), and the feedforward torque information;

[0051] The rotary motor and position encoder unit is used to complete the θ-axis PID control and drive the rotary motor to move according to the θ-axis encoder position information, the desired angle lookup table generated by the motion trajectory model θ(t), and the corresponding feedforward torque information;

[0052] The TDI trigger signal generating unit is used to construct a DPLL digital phase-locked loop mechanism according to the lookup table and the position information of the R-axis and θ-axis encoders, so that the generated TDI trigger signal follows the phase of the actual motion position information.

[0053] Furthermore, the FPGA encoder module is used for high-speed real-time signal acquisition and processing; the ARM processor module is used to obtain the R-axis and θ-axis desired position curve and speed curve lookup tables and the R-axis and θ-axis feedforward torque values; the EMMC storage module is used to store the R-axis and θ-axis target position information required for motion control and the TDI trigger clock lookup table information.

[0054] Furthermore, the TDI triggers the required lookup table data CNT enc and CNT clk (n), stored in the motion EMMC storage module, the ARM controller module reads the lookup table data from the EMMC storage module, and transmits the CNT enc and CNT clk (n) are sequentially transmitted to the FPGA encoder module, which generates an equidistant TDI trigger target signal and transmits the signal to the DPLL digital phase-locked loop module for further processing to generate a TDI trigger signal.

[0055] Furthermore, the FPGA encoder module also includes a decoding component, which performs anti-shake filtering on the input incremental encoder signal, and sends the filtered encoder data to the incremental encoder decoding state machine for decoding to obtain incremental encoded data.

[0056] The present invention has at least the following beneficial effects:

[0057] This effectively solves the synchronization problem of the rotating motor and the executing motor of the equidistant spiral motion trajectory, and solves the problem of phase asynchrony with the motion trajectory caused by traditional equal-duration triggering (fixed cycle), as well as the problem of image distortion of the photographed object caused by non-equidistant triggering.

[0058] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 : A schematic flow chart of the motion control design method of the present invention;

[0060] Figure 2 : A schematic diagram of the structure of the motion control design system of the present invention;

[0061] Figure 3 : Schematic diagram of interference signal filtering effect and decoding state transition in an embodiment of the present invention;

[0062] Figure 4 : Wafer radial linear motor control flow chart in an embodiment of the present invention;

[0063] Figure 5: Flow chart of wafer rotation axis motor control in an embodiment of the present invention;

[0064] Figure 6 : A schematic diagram of generating a DPLL TDI trigger signal provided in an embodiment of the present invention;

[0065] Figure 7 : Schematic diagram of the motion trajectory of the center of the light spot in the embodiment of the present invention;

[0066] Figure 8 : Schematic diagram of R-axis position time curve in an embodiment of the present invention;

[0067] Fig. 9 : Schematic diagram of the θ-axis position time curve in an embodiment of the present invention;

[0068] Fig.10 : Schematic diagram of R-axis speed time curve in an embodiment of the present invention;

[0069] Fig.11 : Schematic diagram of the θ-axis speed-time curve in an embodiment of the present invention;

[0070] Fig.12 : Schematic diagram of TDI trigger clock timing in an embodiment of the present invention. DETAILED DESCRIPTION

[0071] The following will be combined with the drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0072] Embodiment 1

[0073] See also Figure 1-Figure 12 The present invention provides a technical solution: a motion control method for non-patterned wafer defect detection, specifically comprising the following steps:

[0074] S1. Construct the motion trajectory model R(t), θ(t), V(t), W(t) of the wafer to be tested at the current time t, generate a lookup table of the relationship between the R-axis and θ-axis target position and speed and the current time, as well as the θ-axis encoder incremental count value and trigger delay value for TDI camera triggering, and then store the lookup table;

[0075] (S11) constructing the motion trajectory model R(t), θ(t), V(t), and W(t) of the wafer to be measured at the current time t, as follows:

[0076]

[0077] Where, the initial scan line width is d 0 , the wafer radius is r 0 , the linear velocity of the scanning line spot is v 0 ;

[0078] From the above formula (1) and the initial time t = 0, R (0) = r 0 ,available:

[0079]

[0080]

[0081] Where R(t) is the position of the wafer inspection line along the R direction at time t, V R (t) is the running speed of the linear motor at time t;

[0082] From the above, we can get the motion control trajectory model of the rotating axis: (4) and (5):

[0083]

[0084] According to equations (2) and (3), the target position and speed lookup table required for motion control of linear motion can be generated. According to equations (4) and (5), the target position and speed lookup table required for motion control of rotary motion can be generated. Rotary motion and linear motion are associated through equation (1) to achieve dual-axis motion synchronization.

[0085] Generate the incremental count value and trigger delay value of the θ-axis encoder for TDI camera triggering, as follows:

[0086] (S12) It is known that the target trigger frequency of the TDI camera is f 0 , the corresponding period is T 0 , the resolution of the θ axis encoder is θ 0 ;

[0087] The TDI triggering time is: T(n) = {T 0 ,2T 0 ,3T 0 ,…,nT 0}

[0088] The θ angle corresponding to each TDI trigger is: {θ(T 0 ),θ(2T 0 ),θ(3T 0 ),…,θ(nT 0 )}

[0089] Correspondingly, the incremental count value of the θ encoder corresponding to each trigger of TDI is:

[0090]

[0091] The encoder is installed at a distance of 1 / 3 from the rotation center, so the resolution of the outermost circle is only 1 / 3 of the actual resolution of the encoder. In order to further improve the TDI triggering accuracy, the value obtained by CNT(n) is rounded down to get the integer part, which is the integer cycle of the encoder at the time of TDI triggering:

[0092]

[0093] The decimal part is the error at the current resolution:

[0094]

[0095] Combined with formula (7), formula (8) is used to obtain the decimal part. From {X(n)}, we can know that according to the radius of the current spot position R(n), the linear velocity v of the spot movement is 0 , and the FPGA system clock frequency f clk , we can get the f required for the motion {X(n)} clk Number of clock cycles:

[0096]

[0097] According to equations (7) and (9), the lookup table data CNT required for TDI triggering is generated. enc and CNT clk (n), used to generate equidistant TDI trigger target signals;

[0098] S2. Receive the encoder position information of the linear motor R axis and the rotary motor θ axis;

[0099] S3. According to the R-axis encoder position information, the desired position lookup table generated by the motion trajectory model R(t), and the feedforward torque information, the R-axis PID control is completed and the linear motor is driven to move, such as Figure 4 As shown, the details are as follows:

[0100] (S31) The host computer generates a lookup table of the R-axis position and speed relative to the expected position and expected speed of the time according to the R-axis motion trajectory model, and saves the table;

[0101] (S32) The host computer generates a feedforward quantity lookup table based on the measured speed, position, and torque data;

[0102] (S33) receiving the decoded real-time position information of the R-axis encoder, and further converting the real-time position information of the R-axis encoder to obtain the speed information at the current moment;

[0103] (S34) The expected position information R(t) at time t obtained by the expected position lookup table and the actual position information R_ENC(t) obtained by the encoder are combined to complete the position closed-loop PID control. The result obtained by the position closed-loop is used as the input of the speed closed-loop, and then combined with the real-time speed information V_ENC(t) calculated by R_ENC(t) to complete the speed closed-loop PID control; the output of the speed closed-loop control is then integrated with the feedforward torque data and output as the final torque to the R-axis driver to complete the R-axis motor drive control;

[0104] S4. According to the θ-axis encoder position information, the desired angle lookup table generated by the motion trajectory model θ(t), and the corresponding feedforward torque information, the θ-axis PID control is completed and the rotary motor is driven to move, such as Figure 5 As shown, the details are as follows:

[0105] (S41) the host computer generates a lookup table of the expected position and expected speed of the θ-axis relative to time according to the θ-axis motion trajectory model, and saves the table;

[0106] (S42) The host computer generates a feedforward lookup table based on the measured speed, position, and torque data;

[0107] (S43) receiving the decoded real-time position information of the θ-axis encoder, and further converting the real-time position information of the θ-axis encoder to obtain the speed information at the current moment;

[0108] (S44) combining the expected position information θ(t) at time t obtained from the expected position lookup table and the actual position information θ_ENC(t) obtained from the encoder to complete the position closed-loop PID control, the result obtained from the position closed-loop is used as the input of the speed closed-loop, and then combined with the real-time speed information V_ENC(t) calculated by θ_ENC(t) to complete the speed closed-loop PID control; the output of the speed closed-loop control is then integrated with the feedforward torque data and output as the final torque to the R-axis driver to complete the θ-axis motor drive control;

[0109] S5. According to the lookup table data CNT enc 、CNT clk (n) and the R-axis and θ-axis encoder position information construct a DPLL digital phase-locked loop module, so that the generated TDI trigger signal follows the phase of the actual motion position information.

[0110] It should be noted that while performing synchronous control of the R-axis and the θ-axis to complete the equidistant spiral motion, it is necessary to synchronously generate a TDI camera line frequency trigger signal. This signal needs to be completely matched with the motion and phase synchronized to ensure effective acquisition of image data. This embodiment proposes a digital phase-locked loop (DPLL) implemented based on an FPGA encoder module. The implementation principle is as follows: Figure 6As shown, specifically, it includes a LUT_TDI_TRIG module, a digital phase detector (DPD) module, a loop filter (LPF) module, a numerically controlled oscillator (NCO) module and a TDI clock generation module;

[0111] The input signal of the LUT_TDI_TRIG module includes the current incremental count value and the delayed FPGA system clock count value generated by the motion trajectory generation unit and the actual signal of the θ-axis encoder after filtering and decoding. enc (n) and CNT clk (n) Stored in the controller's EMMC in the form of a lookup table, read by the ARM and passed to the FPGA in real time;

[0112] The LUT_TDI_TRIG module increments the actual encoder with a system clock much larger than the TDI trigger frequency. When CNT is reached, enc (n) The FPGA system clock number is counted. When the system clock number reaches CNT clk (n) generates a TDITrigPre pulse; the TDITrigPre signal and the target TDI trigger frequency signal TDI_obj are sent to the digital phase detector for comparison. If the difference between TDITrigPre and TDI_obj is greater than the threshold, the DPFD outputs TDITrigPre and TDI_obj to the Charge PUMP. The Charge PUMP generates a charge and discharge indication based on the size of TDITrigPre and TDI_obj and sends it to the LPF module for filtering and then to the NCO module. The NCO module implements a fixed pulse frequency signal in the FPGA. We can use a known clock source to divide the frequency to obtain a pulse within a certain frequency range.

[0113] The specific implementation method is to use a counter. When a clock pulse appears, the counter is +1, and the upper limit of the counter is the frequency division coefficient. When the number of the counter is less than 1 / 2 of the upper limit, it outputs 1. When the number of the counter is greater than 1 / 2 of the upper limit, it outputs 0. When the number of the counter exceeds the upper limit, the counter returns to zero. In this way, the frequency division of the clock source can be achieved. According to the above method, as long as the upper and lower limits of the counter are changed, the frequency division coefficient can be changed, thereby changing the frequency of the output signal. The clock signal output by the NCO is sent to the DPLL TDI CLK module. The DPLL TDI CLK uses the frequency division coefficient N to divide the NCO clock to generate TDI_trig.

[0114] Embodiment 2:

[0115] like Figure 2-Figure 12As shown, this embodiment provides a motion control system for non-patterned wafer defect detection, which is used to implement a motion control method for non-patterned wafer defect detection described in Embodiment 1, including:

[0116] A motion trajectory generating unit is used to construct a motion trajectory model R(t), θ(t), V(t), and W(t) of the wafer line to be tested at the current time t;

[0117] A motion control processing unit, the motion control processing unit includes an FPGA encoder module, an ARM processor module and an EMMC storage module, the FPGA encoder module and the EMMC storage module are both electrically connected to the ARM processor module;

[0118] The linear motor and position feedback unit are used to complete the R-axis PID control and drive the linear motor to move according to the R-axis encoder position information, the desired position lookup table generated by the motion trajectory model R(t), and the feedforward torque information;

[0119] The rotary motor and position encoder unit is used to complete the θ-axis PID control and drive the rotary motor to move according to the θ-axis encoder position information, the desired angle lookup table generated by the motion trajectory model θ(t), and the corresponding feedforward torque information;

[0120] The TDI trigger signal generating unit is used to construct a DPLL digital phase-locked loop mechanism according to the lookup table and the position information of the R-axis and θ-axis encoders, so that the generated TDI trigger signal follows the phase of the actual motion position information.

[0121] Furthermore, the FPGA encoder module is used for high-speed real-time signal acquisition and processing; the ARM processor module is used to obtain the R-axis and θ-axis expected position curve and speed curve lookup table and the R-axis and θ-axis feedforward torque value acquisition; the EMMC storage module is used to store the R-axis and θ-axis target position information required for motion control and the TDI trigger clock lookup table information.

[0122] Furthermore, the lookup table data CNT required for TDI triggering enc and CNT clk (n), stored in the motion EMMC storage module, the ARM controller module reads the lookup table data from the EMMC storage module, and transmits the CNT enc and CNT clk (n) are sequentially transmitted to the FPGA encoder module, which generates an equidistant TDI trigger target signal and transmits the signal to the DPLL digital phase-locked loop module for further processing to generate a TDI trigger signal.

[0123] Furthermore, the FPGA encoder module also includes a decoding component, which performs anti-shake filtering on the input incremental encoder signal, and sends the filtered encoder data to the incremental encoder decoding state machine for decoding to obtain incremental encoded data;

[0124] The incremental encoder state transfer process is as follows Figure 3 As shown in the figure, the incremental encoder AB signal combination has 4 states, and the state transition is only related to the direction of movement; when moving in the positive direction, the transition relationship is the arrow line with "1" in the figure, that is, "00"->"10"->"11"->"01"->"00"; the initial AB state is "00", and the state is converted to "10" when the A rising edge is detected. The next state must be the arrival of the B rising edge. If the B rising edge is not detected within the specified clock cycle, it is determined to be a false trigger caused by signal jitter or interference, and the state returns to "00"; if the B rising edge is normally detected in the "10" state, the state is converted to "11" and waits. Wait for the falling edge of A. If the falling edge of A is detected within the specified clock cycle of the "11" state, the state is transferred to "01", otherwise the state returns to the initial "00"; if the falling edge of B is detected within the specified clock cycle under the state "01", it means that a complete incremental encoder state transfer process is detected, and the encoder CNT number is increased by one, otherwise the state returns to the initial state "00"; similarly, the reverse movement is the direction of the arrow with "-1" in the figure, and the processing process is the same as the forward state transfer, that is, the incremental encoder CNT number minus one must be detected completely from "00" -> "01" -> "11" -> "10" -> "00".

[0125] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0126] For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. When an element is referred to as being "assembled on", "installed on", "fixed on" or "set on" another element, it can be directly on the other element or there can also be a centered element. When an element is considered to be "connected" to another element, it can be directly connected to the other element or there may be a centered element at the same time. The terms "vertical", "horizontal", "up", "down", "left", "right" and similar expressions used herein are only for illustrative purposes and are not intended to be the only implementation method.

[0127] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

[0128] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

Claims

1. A motion control method for non-patterned wafer defect detection, characterized in that: The specific steps include: Construct the motion trajectory model R(t), θ(t), V(t), W(t) of the wafer to be tested at the current time t, generate a lookup table of the relationship between the R-axis and θ-axis target position and speed and the current time, as well as the θ-axis encoder incremental count value and trigger delay value for TDI camera triggering, and then store the lookup table; Receive the position information of the linear motor R axis and the rotary motor θ axis encoder; According to the R-axis encoder position information, the desired position lookup table generated by the motion trajectory model R(t), and the feedforward torque information, the R-axis PID control is completed and the linear motor is driven to move; According to the θ-axis encoder position information, the desired angle lookup table generated by the motion trajectory model θ(t), and the corresponding feedforward torque information, the θ-axis PID control is completed and the rotary motor is driven to move; According to the lookup table data CNT enc 、CNT clk (n) and the R-axis and θ-axis encoder position information construct a DPLL digital phase-locked loop module, so that the generated TDI trigger signal follows the phase of the actual motion position information.

2. The motion control method for detecting defects in a non-patterned wafer according to claim 1, characterized in that: Construct the motion trajectory model R(t), θ(t), V(t), and W(t) of the wafer to be tested at the current time t, as follows: In the formula, the initial scanning line width is d0, the wafer radius is r0, and the linear velocity of the scanning line spot movement is v0; From the above formula (1) and the initial time t = 0 when R (0) = r0, we can get: Where R(t) is the position of the wafer inspection line along the R direction at time t, V R (t) is the running speed of the linear motor at time t; From the above, we can get the motion control trajectory model of the rotating axis: (4) and (5): According to equations (2) and (3), the target position and speed lookup table required for motion control of linear motion can be generated. According to equations (4) and (5), the target position and speed lookup table required for motion control of rotary motion can be generated. Rotary motion and linear motion are associated through equation (1) to achieve dual-axis motion synchronization.

3. The motion control method for detecting defects in a non-patterned wafer according to claim 2, characterized in that: Generate the incremental count value and trigger delay value of the θ-axis encoder for TDI camera triggering, as follows: It is known that the target trigger frequency of the TDI camera is f0, the corresponding period is T0, and the resolution of the θ-axis encoder is θ0; The TDI triggering time is: T(n) = {T0, 2T0, 3T0, ..., nT0} The θ angle corresponding to each TDI trigger is: {θ(T0),θ(2T0),θ(3T0),…,θ(nT0)} Correspondingly, the incremental count value of the θ encoder corresponding to each trigger of TDI is: In order to further improve the TDI triggering accuracy, the value obtained by CNT(n) is rounded down to get the integer part, which is the integer cycle of the encoder at the TDI triggering moment: The decimal part is the error at the current resolution: From {X(n)}, according to R(n) the radius of the current spot position, the linear velocity v0 of the spot movement, and the FPGA system clock frequency f clk , we can get the f required for the motion {X(n)} clk Number of clock cycles: According to equations (7) and (9), the lookup table data CNT required for TDI triggering is generated. enc and CNT clk (n), used to generate equidistant TDI trigger target signals.

4. The motion control method for detecting defects in a non-patterned wafer according to claim 1, characterized in that: According to the R-axis encoder position information, the desired position lookup table generated by the motion trajectory model R(t), and the feedforward torque information, the R-axis PID control is completed and the linear motor is driven to move. The specific method is as follows: (1) The host computer generates a lookup table of the R-axis position and speed relative to the expected position and expected speed of the time according to the R-axis motion trajectory model, and saves it; (2) A feedforward lookup table generated by the host computer based on the measured speed, position, and torque data; (3) receiving the decoded real-time position information of the R-axis encoder, and further converting the real-time position information of the R-axis encoder to obtain the speed information at the current moment; (4) The expected position information R(t) at time t obtained from the expected position lookup table and the actual position information R_ENC(t) obtained from the encoder are combined to complete the position closed-loop PID control. The result of the position closed-loop is used as the input of the speed closed-loop, and then combined with the real-time speed information V_ENC(t) calculated by R_ENC(t) to complete the speed closed-loop PID control; the output of the speed closed-loop control is then integrated with the feedforward torque data and output as the final torque to the R-axis driver to complete the R-axis motor drive control.

5. The motion control method for detecting defects in a non-patterned wafer according to claim 4, characterized in that: According to the θ-axis encoder position information, the desired angle lookup table generated by the motion trajectory model θ(t), and the corresponding feedforward torque information, the θ-axis PID control is completed and the rotary motor is driven to move. The specific method is as follows: (1) The host computer generates a lookup table of the expected position and expected speed of the θ-axis relative to time based on the θ-axis motion trajectory model and saves it; (2) A feedforward lookup table generated by the host computer based on the measured speed, position, and torque data; (3) receiving the decoded real-time position information of the θ-axis encoder, and further converting the real-time position information of the θ-axis encoder to obtain the speed information at the current moment; (4) The expected position information θ(t) at time t obtained from the expected position lookup table and the actual position information θ_ENC(t) obtained by the encoder are combined to complete the position closed-loop PID control. The result of the position closed-loop is used as the input of the speed closed-loop, and then combined with the real-time speed information V_ENC(t) calculated by θ_ENC(t) to complete the speed closed-loop PID control; the output of the speed closed-loop control is then integrated with the feedforward torque data as the final torque output to the R-axis driver to complete the θ-axis motor drive control.

6. The motion control method for detecting defects in a non-patterned wafer according to claim 1, characterized in that: The DPLL digital phase-locked loop includes a LUT_TDI_TRIG module, a digital phase detector module, a loop filter module, a digital control oscillator module and a TDI clock generation module, so that the generated TDI trigger signal follows the phase of the position information of the actual movement, as follows: When a clock pulse appears, the counter is +1, and the upper limit of the counter is the frequency division coefficient. When the number of the counter is less than 1 / 2 of the upper limit, it outputs 1. When the number of the counter is greater than 1 / 2 of the upper limit, it outputs 0. When the number of the counter exceeds the upper limit, the counter returns to zero, so that the frequency division of the clock source can be achieved; The clock signal output by the digital control oscillator module is sent to the TDI clock generation module, and the TDI clock generation module uses the frequency division coefficient N to divide the clock of the digital control oscillator module to generate a TDI trigger signal.

7. A motion control system for detecting defects in a non-patterned wafer, used to implement a motion control method for detecting defects in a non-patterned wafer according to any one of claims 1 to 6, characterized in that: include: A motion trajectory generating unit is used to construct a motion trajectory model R(t), θ(t), V(t), and W(t) of the wafer line to be tested at the current time t; A motion control processing unit, wherein the motion control processing unit includes an FPGA encoder module, an ARM processor module and an EMMC storage module, wherein the FPGA encoder module and the EMMC storage module are both electrically connected to the ARM processor module; The linear motor and position feedback unit are used to complete the R-axis PID control and drive the linear motor to move according to the R-axis encoder position information, the desired position lookup table generated by the motion trajectory model R(t), and the feedforward torque information; The rotary motor and position encoder unit is used to complete the θ-axis PID control and drive the rotary motor to move according to the θ-axis encoder position information, the desired angle lookup table generated by the motion trajectory model θ(t), and the corresponding feedforward torque information; The TDI trigger signal generating unit is used to construct a DPLL digital phase-locked loop mechanism according to the lookup table and the position information of the R-axis and θ-axis encoders, so that the generated TDI trigger signal follows the phase of the actual motion position information.

8. The motion control method for detecting defects in a non-patterned wafer according to claim 7, characterized in that: The FPGA encoder module is used for high-speed real-time signal acquisition and processing; the ARM processor module is used to obtain the R-axis and θ-axis expected position curve and speed curve lookup tables and the R-axis and θ-axis feedforward torque values; the EMMC storage module is used to store the R-axis and θ-axis target position information required for motion control and the TDI trigger clock lookup table information.

9. The motion control method for detecting defects in a non-patterned wafer according to claim 8, characterized in that: The TDI triggers the lookup table data CNT required enc and CNT clk (n), stored in the motion EMMC storage module, the ARM controller module reads the lookup table data from the EMMC storage module, and transmits the CNT enc and CNT clk (n) are sequentially transmitted to the FPGA encoder module, which generates an equidistant TDI trigger target signal and transmits the signal to the DPLL digital phase-locked loop module for further processing to generate a TDI trigger signal.

10. The motion control method for detecting defects in a non-patterned wafer according to claim 9, characterized in that: The FPGA encoder module also includes a decoding component, which performs anti-shake filtering on the input incremental encoder signal, and sends the filtered encoder data to the incremental encoder decoding state machine for decoding to obtain incremental encoded data.

Images