Contact type wafer thickness compensation method of a probe station

Abstract

This invention discloses a contact-type wafer thickness compensation method for a probe station. The method includes: setting a safe initial height, pressure threshold, and other parameters after aligning the wafer; three stages: large-step rapid approximation, bisection interval compression, and ultra-small-step fine positioning; driving an electric chuck assembly to position the initial contact height; during testing, real-time detection of the actual needle pressure value; filtering the needle pressure value; fitting a linear relationship between pressure and height; comparing it with the needle pressure of the first chip; calculating the compensation amount to adjust the contact height of the next chip; and cyclically executing the method to achieve dynamic compensation across the entire wafer. This invention achieves a positioning accuracy of 0.1μm, has a small compensation error, requires no special test keys, and balances safety and testing efficiency, making it suitable for the field of semiconductor wafer testing.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of semiconductor testing equipment, and particularly relates to a contact type wafer thickness compensation method of a probe station. BACKGROUND

[0002] In the production process of semiconductor chips, wafer testing is a key link to ensure product yield. As the core equipment of wafer testing, the precision of the probe station directly affects the accuracy and reliability of the test results. During the testing process, the probe needs to be in precise and stable contact with the pads of the chips on the wafer. However, due to the possible bending, warping or uneven thickness of the wafer itself, the surface of the wafer is not an ideal plane. If no compensation is made, the same set of needle height may cause poor contact (needle pressure is too small or not in contact) or excessive contact (needle pressure is too large, damaging the chip or the probe) between the probe and the chip in different areas of the wafer.

[0003] To solve the above problems, various thickness compensation schemes have been proposed in the prior art. For example, a needle pressure adaptation method is disclosed in Chinese Patent Publication No. CN112014710B. The method discloses that before testing, the contact resistance is calculated by measuring the resistance of two different sizes of test keys pre-made on the wafer at different pressing heights, and a best and fixed target detection height is determined according to the distribution of the contact resistance.

[0004] However, the technical solution of the comparative document has the following defects:

[0005] Dependence on special structure: This method strongly depends on the pre-made two different sizes of test keys on the wafer, which limits its application range. For conventional wafers without such test keys, this method cannot be implemented.

[0006] Calibration complexity: A precise "resistance measurement value-test key size-contact resistance" correlation needs to be established and relied on. This calibration process is complex and easily affected by the wafer surface conditions (such as oxidation, contamination). Once the calibration is not accurate, the entire needle pressure adaptation result will be biased.

[0007] Static adaptation: Essentially, this method is a "pre-adaptation", which is only executed once before testing to determine a globally fixed target height. It cannot cope with the dynamic changes in wafer thickness or significant concave-convex in local areas during testing, and the compensation accuracy is limited.

[0008] Low efficiency: Additional data collection at multiple heights and complex calculation processes are required before testing, which increases the testing preparation time and reduces the overall testing efficiency.

[0009] Therefore, there is an urgent need for a compensation method that does not rely on special test structures, can dynamically respond to changes in wafer thickness during testing, and is more efficient. Summary of the Invention

[0010] To address the shortcomings of existing technologies, such as reliance on special test keys, complex calibration, static adaptation, and low efficiency, this invention proposes a contact-type wafer thickness compensation method for probe stations.

[0011] The core idea of ​​this invention is to detect the actual pin pressure when the probe contacts the current chip in real time during the wafer testing process, and based on this, actively adjust the pin height of the next chip to form a dynamic and iterative compensation closed loop.

[0012] This invention proposes a contact-type wafer thickness compensation method using a probe station, comprising the following steps:

[0013] S1. Initial Height Reference Setting: After the probe station aligns the wafer, the basic parameters and protection parameters are set first. These parameters include the safe starting height Z. start Theoretical height of wafer Z max Contact detection threshold P th Maximum safe pressure P of the probe max Pressure warning threshold P pre and step size parameters, pressure warning threshold P pre For P max 0.8 times, step size parameters include large step size. large Small step size small Ultra-small step size ultra The electric chuck assembly is then driven to rise in stages to determine the contact height.

[0014] S11. Large step size rapid approximation: Control the electric chuck assembly from Z... start Begin with Step large Move upwards, and read the sensor pressure P after each movement; if P=0, continue moving until Z is reached. max If 80% of the time is reached, proceed to S12; if P>0 during the rapid approach with a large step size, it indicates abnormal contact and proceed to S13.

[0015] S12. Bisection Interval Compression: Using the current position Z... n For the lower limit, Z max Construct an interval [Z] for the upper limit n Z max ] Calculate the intermediate point Z mid =(Z n +Z max ) / 2, control the electric chuck assembly to move to Z midand read pressure P mid ; if P mid >P pre or P mid ≤P pre , and P mid >P th , go to S13; if P mid ≤P th , update Z n =Z mid , repeat this step until the interval difference Z max -Z n ≤Step small , go to S13;

[0016] S13. Fine positioning with super small step: start from current Z n , move the motorized chuck assembly upward with Step ultra , and read P in real time; if P th <P≤P pre , record the current height as the initial contact height; if P pre <P≤P max , record the initial contact height and trigger the pressure too high prompt; if P>P max , control the motorized chuck assembly to move downward 2xStep large to disengage from contact and stop; if moving to Z max still has no P>P th , alarm and stop;

[0017] S2. Dynamic needle pressure detection: when performing electrical performance test on the chip to be tested on the wafer, the probe station detects the actual needle pressure value P od when the probe contacts the chip to be tested in real time through mechanical feedback;

[0018] S3. Iterative height compensation: after the test is completed, the industrial computer processes P od and compares it with the needle pressure value P first of the first chip, calculates the height compensation amount, and sends instructions to the motion controller to adjust the contact height setting value of the next chip to be tested;

[0019] S4. Loop execution: repeat steps S2 and S3 to test all chips on the wafer, and realize full-wafer thickness dynamic compensation based on mechanical feedback.

[0020] Further, before step S1, it also includes detecting the thickness of different regions of the wafer through the multi-point detection function to obtain the overall topography information of the wafer, providing a reference for setting the initial contact height.

[0021] Further, the multi-point detection function can select to detect 9 points, 16 points or 25 points, which are uniformly distributed in a matrix on the wafer surface and are sequentially completed by the probes on the probe holder with pressure sensing.

[0022] Further, the mechanical feedback is realized by a needle pressure detection and feedback assembly, which comprises a probe holder with pressure sensing and a pressure sensing table; the probe holder with pressure sensing uses a strain gauge as a sensing element to convert the deformation signal of the probe when contacting into a pressure signal; the contact determination threshold P th The default value is 0.1g, and the step parameter Step large is 100μm, Step small is 1μm, and Step ultra is 0.1μm.

[0023] Further, the probe holder with pressure sensing calibrates the conversion relationship between the analog voltage and the needle depth by a lever design, and the conversion relationship is written into the control software in the industrial computer and used to cooperate with the parameter setting of the wafer theoretical height Z max and the maximum safe pressure of the probe P max to accurately adjust the height of the motorized chuck assembly.

[0024] Further, the pressure sensing table is connected to the industrial computer through a communication line, and the industrial computer sends a query instruction to the pressure sensing table to obtain the current pressure value when the motorized chuck assembly and the probe are in contact, which is used for the contact determination in step S1 and the needle pressure comparison in step S3.

[0025] Further, the processing and calculation of the height compensation amount of P od in step S3 include the following operations: performing needle pressure value filtering processing, first performing amplitude limiting filtering, if the difference between the current P od and the previous needle pressure value x n-1 is out of limit, taking x n-1 as the effective needle pressure value; if the difference is not out of limit, taking P od as the effective needle pressure value; then performing smoothing filtering, and calculating the smoothed needle pressure value according to the formula

[0026] ;

[0027] wherein N is the number of filtering windows; performing linear relationship fitting, taking n sample points in the effective range of the needle pressure, and calculating fitting coefficients k and b according to the formula

[0028] ;

[0029] ;

[0030] In the formula, P is the pressure value and Z is the height value; the compensation amount is calculated and adjusted based on P. first The downward deviation value is ΔP min-od Upward deviation ΔP max-od and the iterative height adjustment amount ΔZ; if P od -P first <ΔP min-od If the needle pressure is too low, the Z-axis adjustment amount ΔZ for the next chip is calculated as follows: ΔZ = (P od -P first If P*k+b; od -P first >ΔP max-od If the needle pressure is too high, the Z-axis adjustment amount ΔZ for the next chip is calculated as follows: ΔZ = (P od -P first )*k+b; if ΔP min-od <P od -P first <ΔP max-od If the pressure of the needle is within the allowable range, the Z-axis height of the next chip remains unchanged.

[0031] Furthermore, in step S2, when the electric chuck assembly and the probe are separated before each contact, the industrial control computer controls the pressure sensor to zero the sensor in the probe holder with pressure sensing to avoid test errors caused by sensor zero drift. The zeroed data is used for pressure reading in step S1 and filtering in step S3.

[0032] Furthermore, in step S2, if the actual needle pressure value received by the industrial control computer from the pressure sensor remains 0, and the full-stage positioning in step S1 has not yet detected P>P... th If the contact height is not found, the control software on the industrial control computer will issue an alarm and prompt the user to repeat step S1 to find the contact height again.

[0033] The contact-type wafer thickness compensation detection system for the probe station used to implement the above method includes:

[0034] The wafer is the object to be tested.

[0035] Electric chuck assembly, used to carry and drive wafers in three-dimensional and rotary motion;

[0036] A probe holder with pressure sensing is used to mount test probes and has a built-in pressure sensor to detect needle pressure;

[0037] A pressure sensor is used to receive signals from a probe holder with a pressure sensor and convert them into a specific pressure value.

[0038] Industrial computer, as the control core of the whole system, runs control software;

[0039] Motion controller, for receiving instructions from the industrial computer and controlling the motion of the motorized chuck assembly;

[0040] Pressure sensor signal line, connecting the probe holder with pressure sensing and the pressure sensing meter, for transmitting the original pressure signal;

[0041] Pressure sensing meter communication line, connecting the pressure sensing meter and the industrial computer, for uploading pressure data and issuing control instructions;

[0042] Motion controller communication line, connecting the industrial computer and the motion controller, for transmitting motion control instructions;

[0043] Motion control driving line, connecting the motion controller and the motorized chuck assembly, for converting electrical signals into driving force to realize precise motion of the motorized chuck assembly.

[0044] The advantages of the present application are:

[0045] First positioning accuracy: through the three-step algorithm of large step approximation, bisection compression and ultra-small step positioning, combined with multi-pressure threshold protection, the initial contact height positioning accuracy can reach 0.1 mu m level;

[0046] High compensation accuracy: introducing amplitude limiting + smoothing double filtering processing for needle pressure interference, establishing a pressure-height quantitative relationship through linear fitting, and the compensation error can be controlled within ±0.1 mu m;

[0047] Strong safety: setting P max , P pre Dual pressure protection and abnormal shutdown mechanism, suitable for complex test environment;

[0048] Efficiency balance: iterative compensation and test process seamlessly connect, without additional test time consumption, suitable for testing wafers of different thickness specifications. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 is a structural schematic diagram of the contact type wafer thickness compensation detection system of the probe station.

[0050] In the figure: 1 wafer, 2 motorized chuck assembly, 3 probe holder with pressure sensing, 4 pressure sensing meter, 5 industrial computer, 6 motion controller. DETAILED DESCRIPTION

[0051] Example 1, the contact type wafer thickness compensation method of the probe station described in the present application, comprising the following steps:

[0052] Preparation stage: Place the wafer to be tested on the motorized chuck assembly and fix it; start the control software on the industrial computer, input basic parameters and protection parameters: Z start = 5 mm, Z max = 8 mm, P th = 0.1 g, P max = 10 g, P pre = 8 g, Step large = 100 pm, Step small = 1 pm, Step ultra = 0.1 pm; Set the filter window N = 3, AP min-od = -0.3 g, AP max-od = +0.3 g; System self-check pressure sensor, motion controller and other components, ensure that the communication of each signal line is normal.

[0053] S1. Initial height reference setting:

[0054] S11. Large step fast approximation: The industrial computer sends instructions to the motion controller to drive the motorized chuck assembly to move upwards from 5 mm at 100 pm / step; after each movement, the pressure sensor reads the pressure value and transmits it to the industrial computer; after 24 steps, the height of the motorized chuck assembly is 5 mm + 24 x 0.1 mm = 7.4 mm, reaching 92.5% of Z max = 8 mm, at this time the pressure P = 0, enter S12;

[0055] S12. Bisection interval compression: current Z n = 7.4 mm, Z max = 8 mm, interval [7.4 mm, 8 mm]; calculate Z mid = (7.4 + 8) / 2 = 7.7 mm, read P mid = 0.05 g ≤ P th , update Z n = 7.7 mm; new interval [7.7 mm, 8 mm], calculate Z mid = 7.85 mm, move P mid = 0.08 g ≤ P th , update Z n = 7.85 mm; repeat the operation until the interval [7.92 mm, 7.93 mm], calculate Z mid = 7.925 mm, move P mid = 0.12 g > Pth and ≤ Ppre, update the upper limit of the interval to 7.925 mm; continue to calculate Z mid = 7.9225 mm, move P mid = 0.11 g, interval [7.92 mm, 7.9225 mm]; until calculate Z mid= 7.9205 mm, interval difference 0.5 pm≤1 pm, enter S13;

[0056] S13. Fine positioning with ultra-small step size: from Z n = 7.92 mm, move upward with 0.1 pm / step; after 5 steps, the height is 7.92 mm+5x0.1 pm=7.9205 mm, read P=0.105 g, which satisfies P th < P≤P pre , record the initial contact height Z c = 7.9205 mm; test the first chip, read P first = 5 g.

[0057] S2. Dynamic needle pressure detection & S3. Iterative height compensation & S4. Loop execution:

[0058] Test the first chip to be tested: drive the wafer to the target position, control the motorized chuck assembly to rise to Zc+20 pm=7.9405 mm, read Pod=4.6 g during testing;

[0059] Compensation calculation: perform amplitude limiting filtering, the difference between P od = 4.6 g and P first = 5 g is 0.4 g, which is out of limit, take P first = 5 g as the effective needle pressure value; perform smoothing filtering, the smoothed needle pressure value is (5+5+4.6) / 3≈4.87 g; take n=5 sample points (P1=0.1 g, Z1=7.9205 mm; P2=2 g, Z2=7.9245 mm; P3=3 g, Z3=7.9265 mm; P4=4 g, Z4=7.9285 mm; P5=5 g, Z5=7.9305 mm);

[0060] Calculate k: numerator part: 5x(0.1x7.9205+2x7.9245+3x7.9265+4x7.9285+5x7.9305)-(0.1+2+3+4+5)x(7.9205+7.9245+7.9265+7.9285+7.9305)≈5x79.272-14.1x39.6305≈396.36-558.79≈-162.43; denominator part: 5x(0.01+4+9+16+25)-(14.1)²≈5x54.01-198.81≈270.05-198.81≈71.24;

[0061] k≈-162.43 / 71.24≈-2.28 mm / g;

[0062] b = (39.6305 - (-2.28) x 14.1) / 5 = (39.6305 + 32.148) / 5 = 71.7785 / 5 = 14.3557 mm;

[0063] P od -P first = 4.6 - 5 = -0.4 g < ΔP min-od , the Z-axis adjustment amount ΔZ of the next chip = (-0.4) x (-2.28) + 14.3557 = 0.912 + 14.3557 = 15.2677 mm (here ΔZ is the absolute height, which needs to be corrected in combination with the initial contact height, and the actual adjustment amount is 15.2677 - 7.9405 = 7.3272 μm), the new height = 7.9405 + 0.0073 = 7.9478 mm;

[0064] The second chip to be tested is tested: moved to the target position, raised to 7.9478 mm, read P od = 4.9 g, P od -P first = -0.1 g between ΔP min-od ~ ΔP max-od , the Z-axis height is unchanged;

[0065] Loop execution: subsequent chips are dynamically adjusted according to the above logic until the entire wafer test is completed, without pressure overrun or no contact abnormality.

[0066] Example 2, as shown in Figure 1 , the contact type wafer thickness compensation detection system of the probe station for realizing the method described in the application is composed of the following parts:

[0067] Wafer 1: a semiconductor wafer to be tested, which is distributed with a large number of chips.

[0068] Motorized chuck assembly 2: as a bearing platform of the wafer, it can realize translation in X, Y, Z directions and rotation (θ) around Z axis under the drive of the motion controller 6, so as to accurately position any chip on the wafer.

[0069] Probe seat 3 with pressure sensing: installed on the probe arm of the probe station, the lower end is installed with test probes. Its internal integrated pressure sensor with strain gauge as the core. When the probe contacts the surface of wafer 1, the micro pressure generated will cause the strain gauge to deform, thereby converting the mechanical signal into a measurable electrical signal (such as voltage change). The internal lever structure can amplify the micro deformation, improve the detection sensitivity, and be used to calibrate the relationship between voltage and needle depth.

[0070] Pressure sensor signal line: used to connect the probe holder 3 with pressure sensor and the pressure sensor meter 4, to transmit the original analog pressure signal.

[0071] Industrial computer 5: the "brain" of the whole system, running the control software to realize the method of the invention. It communicates with the pressure sensor meter 4 through the pressure sensor meter communication line, inquires and receives real-time pressure data; at the same time, it sends motion instructions and height setting values to the motion controller 6 through the motion controller communication line.

[0072] Motion controller 6: the "actuator" of the industrial computer 5. It receives instructions from the industrial computer 5 and parses them into precise motor control signals, which are sent to each motor in the motorized chuck assembly 2 through the motion control drive line to control it to complete the corresponding movement.

[0073] Pressure sensor signal line: used to connect the probe holder 3 with pressure sensor and the pressure sensor meter 4, to transmit the original analog pressure signal.

[0074] Pressure sensor meter communication line: used to connect the pressure sensor meter 4 and the industrial computer 5, usually a serial communication line such as RS232, to realize bidirectional transmission of data and instructions.

[0075] Motion controller communication line: used to connect the industrial computer 5 and the motion controller 6, usually Ethernet or special bus, for high-speed and reliable transmission of complex motion control instructions.

[0076] Motion control drive line: used to connect the motion controller 6 and the motorized chuck assembly 2, to transmit motor drive signals to the motors of each axis to directly drive the mechanical structure to move.

[0077] All the above components work together to achieve dynamic and real-time compensation for uneven wafer thickness, ensuring that each chip can be tested at the best needle pressure.

[0078] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and variations can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A contact type wafer thickness compensation method of a probe station, characterized by, The method comprises the following steps: S1. Initial height reference setting: After the wafer is aligned by the probe station, the basic parameters and protection parameters are set first, including the safe starting height Z start , the wafer theoretical height Z max , the contact determination threshold P th , the maximum safe pressure of the probe P max , the pressure warning threshold P pre , and the step parameter. The pressure warning threshold P pre is 0.8 times of P max , and the step parameter includes the large step Step large , the small step Step small , and the super-small step Step ultra . Then the motor chuck assembly is driven to rise in stages to position the contact height. S11. Large step fast approach: Control the motorized chuck assembly to move from Z start start, move up in steps of Step large read sensor pressure P after each move; If P = 0, continue moving until reaching Z max If P = 0, continue moving until reaching Z If P = 0, continue moving until reaching Z S12. Bisection interval compression: construct interval [Z n ,Z max ] with current position Z n as lower limit and Z max as upper limit, calculate middle point Z mid =(Z n +Z max ) / 2, control motorized chuck assembly to move to Z mid and read pressure P mid ; if P mid >P pre or P mid ≤P pre and P mid >P th , go to S13; if P mid ≤P th , update Z n =Z mid , repeat this step until interval difference Z max -Z n ≤Step small , go to S13; S13. Super small step fine positioning: from current Z n Start with Step ultra Move motorized chuck assembly up, read P in real time; if P th <P≤P pre Record current height as initial contact height; if P pre <P≤P max Record initial contact height and trigger pressure too high warning; if P>P max Control motorized chuck assembly to move down 2xStep large Disengage contact and shut down; If move to Z max Still no P>P th , alarm stop; S2. Dynamic needle pressure detection: when testing the electrical performance of the chip to be tested on the wafer, the probe station detects the actual needle pressure value P of the contact between the probe and the chip to be tested in real time through mechanical feedback od ; S3. Iterative height compensation: After the test is completed, the industrial computer processes the data and compares it with the needle pressure value P od of the first chip, calculates the height compensation amount, and sends an instruction to the motion controller to adjust the contact height setting value of the next chip to be tested; first ​ S4. Loop execution: repeat steps S2 and S3 to test all chips on the wafer, and realize full-wafer thickness dynamic compensation based on mechanical feedback.

2. The contact-type wafer thickness compensation method for the probe station according to claim 1, characterized in that, Before step S1, it further comprises detecting the thickness of different areas of the wafer by a multi-point detection function to obtain overall topographic information of the wafer, and providing a reference for setting the initial contact height.

3. The method of claim 2, wherein the probe station is further characterized by: The multi-point detection function can select to detect 9 points, 16 points or 25 points, and these detection points are uniformly distributed in a matrix on the wafer surface and are sequentially completed by probes on a probe seat with pressure sensing.

4. The method of claim 3, wherein the probe station is further characterized by: The mechanical feedback is realized by a needle pressure detection and feedback assembly, which comprises a probe base with pressure sensor and a pressure sensing table; the probe base with pressure sensor uses a strain gauge as a sensing element to convert the deformation signal of the probe contact into a pressure signal; the contact determination threshold P th The default value is 0.1g, the step parameter Step large is 100μm, Step small is 1μm, Step ultra is 0.1μm.

5. The method of claim 4, wherein the probe station is further characterized by: The probe station with pressure sensor calibrates the conversion relationship between analog voltage and needle depth by lever design, which is written into the control software in industrial computer, used for setting parameters of wafer theoretical height Z max , maximum safe pressure of probe P max , and accurately adjusting the height of motorized chuck assembly.

6. The method of claim 5, wherein the probe station is further characterized by: The pressure sensing table is connected with the industrial computer through a communication line, and the industrial computer sends a query instruction to the pressure sensing table to obtain the current pressure value when the motor chuck assembly and the probe are in a contact state, and the pressure value is used for contact determination in step S1 and needle pressure comparison in step S3.

7. The method of claim 6, wherein the probe station is further characterized by: The P od The processing and calculating the height compensation amount include the following operations: performing a needle pressure value filtering process, first performing a limiting filtering, if the difference between the current needle pressure value x od and the previous needle pressure value x n-1 is out of limit, taking x n-1 as the effective needle pressure value; if the difference is not out of limit, taking P od as the effective needle pressure value; then performing a smoothing filtering according to the formula ； The smoothed needle pressure value is calculated, where N is the number of filter windows; linear relationship fitting is performed, n sample points are taken in the effective range of needle pressure, and fitting coefficients k and b are calculated according to the formula: k = (N-1) / (N-1) and b = (N-1) / (N-1). ； ； In the formula, P is a pressure value, and Z is a height value; a compensation amount is calculated and adjusted, and a downward deviation value ΔP first , an upward deviation value ΔP min-od , and an iterative height adjustment amount ΔZ are set based on P max-od ; if P od -P first < ΔP min-od , it is determined that the needle pressure is too small, a Z-axis adjustment amount ΔZ of a next chip is set as (P od -P first )*k+b; if P od -P first > ΔP max-od , it is determined that the needle pressure is too large, a Z-axis adjustment amount ΔZ of a next chip is set as (P od -P first )*k+b; and if ΔP min-od <P od -P first < ΔP max-od , it is determined that the needle pressure is within an allowable range, and a Z-axis height of a next chip is unchanged.

8. The method of claim 7, wherein the probe station is further characterized by: In step S2, the industrial computer controls the pressure sensing table to zero the sensor in the probe seat with pressure sensing when the motor chuck assembly and the probe are in a separated state before each contact, so as to avoid test errors caused by sensor zero drift, and the data after zeroing is used for pressure reading in step S1 and filtering in step S3.

9. The method of claim 8, wherein the probe station is further characterized by: In step S2, if the actual needle pressure value received by the industrial computer from the pressure sensing table is continuously 0, and the full-stage positioning of step S1 has not detected P>P th , it is judged as a non-contact state, and the control software on the industrial computer will issue an alarm and prompt the user to re-execute step S1 to find the contact height again.

Images