A PN junction device impurity concentration / concentration gradient automatic measurement system and method
By designing an automated impurity concentration/concentration gradient measurement system for PN junction devices, the problems of device damage, complex measurement and low degree of automation in the prior art are solved, and efficient and accurate impurity concentration/concentration gradient detection and batch data statistics are achieved.
Patent Information
- Application Number
- CN201811529305.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2018-12-14
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2038-12-14
AI Technical Summary
The existing impurity concentration measurement methods for semiconductor devices have problems such as device damage, complex measurement devices, low signal-to-noise ratio and low automation, and it is impossible to achieve lossless, automated and efficient impurity concentration/concentration gradient detection.
An automatic measurement system for impurity concentration/concentration gradient of PN junction device including feeding mechanism, sample test bench, measurement module, phase lock amplifier and computer is designed, and the signal processing of the phase lock amplifier is processed by automatic transmission of samples, fixed tablet electrical contact, and phase lock amplifier. and computer data collection and analysis to realize automated measurement and data statistics.
It improves measurement efficiency, suppresses system noise, and realizes accurate impurity concentration/concentration gradient detection, which is simple to operate, is suitable for automated data acquisition and statistical analysis of batch devices, and enhances the degree of automation of measurement.
Smart Images

Figure CN109490744B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of automatic testing of semiconductor device parameters, and in particular to a system and method for automatically measuring impurity concentration / concentration gradient of a PN junction device. Background Art
[0002] The doping concentration distribution in semiconductor devices will directly affect their electrical properties. Usually, according to the required electrical properties, a certain process is adopted in the preparation process to artificially control important device parameters such as impurity concentration / concentration gradient. Therefore, the detection of impurity distribution in semiconductor devices after preparation becomes a very important issue. On the one hand, accurate and efficient testing methods can verify the effectiveness of the preparation process, and on the other hand, they can ensure the reliability of device parameters. Existing technical means for testing doping concentration include the use of four-probe, Hall effect with successive delamination, or secondary ion mass spectrometry, but the above methods will contact / destroy the device when measuring, which will inevitably cause device damage, and are all lossy measurement methods.
[0003] Therefore, people prefer to use non-destructive methods in actual measurements. The existing technical means include optical detection methods, such as free carrier absorption measurement technology, photocarrier radiation measurement technology, etc. (such as the contents involved in patent publication numbers CN1971868, CN101159243, and CN101527273A); and electrical detection methods, which in principle mainly include using the capacitance effect of measuring the semiconductor PN junction to calculate and obtain its impurity concentration distribution. The main feasible methods include using an LCR meter (inductance-capacitance-resistance meter) or a special capacitance tester to perform capacitance measurement (such as the contents involved in patent publication numbers 106093641A and CN206540992U). However, the optical detection method requires the construction of a relatively complex optical system on the measuring device, which typically includes a light source, a detection optical path, and an optical detection system, and has a complex structure and high cost. When using the electrical detection method, although the required measuring device is relatively simple, the measurement signal-to-noise ratio is easily affected by uncertain parameters such as line noise. There are problems such as the need for post-processing conversion of the measured data and the low degree of automation of the measurement. Moreover, after a large amount of data is measured, it is still impossible to automatically summarize and statistically analyze the parameters in real time. For example: automatically determine the type of PN junction, classify abrupt junctions and linear slow-change junctions, etc. Summary of the invention
[0004] The technical problem to be solved by the present invention is to provide a PN junction device impurity concentration / concentration gradient automatic measurement system and method in view of the above-mentioned deficiencies in the prior art. The automatic measurement function provided by the present invention can conveniently perform automatic detection of the impurity concentration / concentration gradient of a large number of PN junction samples, and facilitate statistics on the category differences of devices in the same batch of products.
[0005] The technical solution adopted by the present invention is:
[0006] A PN junction device impurity concentration / concentration gradient automatic measurement system, characterized in that the system includes a feeding mechanism, a sample test bench, a measurement module, a lock-in amplifier, and a computer;
[0007] The driver of the feeding mechanism is connected to a computer, and the feeding mechanism is used to automatically transport the PN junction device to the sample test bench; the sample test bench is connected to a measuring module, and a fixed pressing plate is provided on the sample test bench, which is used to fix the two ends of the PN junction device and electrically contact the electrodes of the PN junction device; the sample test bench is connected to the measuring module; the measuring module is connected to a phase-locked amplifier, and the phase-locked amplifier is connected to a computer.
[0008] The feeding mechanism comprises:
[0009] A sample reel containing the PN junction devices to be tested.
[0010] A guide wheel for guiding a PN junction device;
[0011] A test reel for rolling up the tested PN junction devices;
[0012] The sample reel is connected to a first rotating mechanism;
[0013] The test reel is connected to a second rotating mechanism, the second rotating mechanism is connected to a driver via a stepping motor, and the driver is connected to a computer via an RS232 serial port or a USB communication interface.
[0014] The sample testing platform comprises a base with an insulating surface and a PN junction device fixing groove, and a group of elastic pressing sheets used for fixing the PN junction device and electrodes at both ends.
[0015] The measuring module comprises: a test circuit for connecting with a sample test bench; a standard capacitor C0 for quantitative calculation; a measuring loop for detecting a PN junction device; a signal channel interface and a reference channel interface for providing a phase-locked amplifier; a basic interface circuit comprising a series loop of a test device and a standard capacitor C0; a reference signal of the phase-locked amplifier as an excitation signal of the entire loop; a voltage-divided signal on the standard capacitor C0 as a measuring signal of the phase-locked amplifier; and a voltage follower as a connection buffer module between the measuring signal and the phase-locked amplifier.
[0016] The computer is used to realize the operation parameter setting of the lock-in amplifier, the bias control of the PN junction device, the cycle measurement parameters, the measurement signal collection, and the data visualization.
[0017] The invention discloses an automatic measurement method for impurity concentration / concentration gradient of a PN junction device. The measurement module converts the weak voltage signal of the PN junction device collected by a fixed pressing plate into the junction capacitance of the PN junction device and can display the relationship curve between the bias voltage and the junction capacitance of the PN junction device in real time. The curve can be automatically parameterized and fitted. The type of the PN junction is determined according to the correlation coefficient of the fitted straight line result. The impurity concentration distribution coefficient or the concentration gradient coefficient of the PN junction device is converted according to the physical model corresponding to the type.
[0018] The technical effect of the present invention is: by designing an automatic measurement device and method for the impurity concentration / concentration gradient of a PN junction device, the measurement efficiency of the parameters of the PN junction device is improved; at the same time, unlike the method based on the LRC meter or the capacitance meter, the measurement method using correlation detection is conducive to suppressing system noise, so as to obtain the measurement signal and device parameters more accurately. When using the invention to measure a large number of devices, the operation is simple and the degree of automation is high. Compared with manual measurement, the invention can also automatically realize data acquisition, storage and online post-processing; the measurement results of batch devices can be classified and counted. On the one hand, the measurement efficiency is greatly improved, and on the other hand, the classification and statistical rules that cannot be given when only a small number of samples are tested can be given. In addition, the entire measurement system is easy to modularize and is suitable for integration in industrial semiconductor detection equipment. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] Figure 1 It is a schematic diagram of the structure of the device of the present invention.
[0020] Figure 2 Schematic diagram of computer parameter settings for the device of the present invention.
[0021] Figure 3 This is a flow chart of the measurement method of the present invention.
[0022] FIG4(a) shows the V of the PN junction device of the same model batch measured by the present invention. R -C x A typical measurement result diagram of the relationship curve, which is transformed to conform to the physical model of the abrupt junction.
[0023] Figure 4(b) is the conversion of Figure 4(a) into a bias voltage V R and The relationship between the two is shown in the measurement result graph.
[0024] Figure 4(c) is the conversion of Figure 4(a) into a bias voltage V R and The relationship between the two is shown in the measurement result graph.
[0025] FIG5(a) shows the V of the PN junction device of the same model batch measured by the present invention. R -C xA typical measurement result of the relationship curve Figure 1 After transformation, the result conforms to the physical model of linear slow-varying junction.
[0026] FIG5(b) shows the V of the PN junction device of the same model batch measured by the present invention. R -C x A typical measurement result of the relationship curve Figure 2 After transformation, the result conforms to the physical model of linear slow-varying junction.
[0027] FIG5(c) is a graph showing the V of the PN junction device of the same model batch measured by the present invention. R -C x A typical measurement result of the relationship curve Figure 3 After transformation, the result conforms to the physical model of linear slow-varying junction.
[0028] Figure 6 This is a typical result diagram of automatic clustering of PN junction devices of the same model batch according to the correlation coefficient. DETAILED DESCRIPTION
[0029] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.
[0030] like Figure 1 As shown, a PN junction device impurity concentration / concentration gradient automatic measurement system includes a feeding mechanism, a sample test bench 2, a measurement module 3, a phase-locked amplifier 4, and a computer 5.
[0031] The driver 6 of the feeding mechanism is connected to the computer 5, and the feeding mechanism is used to automatically transport the PN junction device to the sample test bench 2; the sample test bench 2 is connected to the measuring module 3, and a fixed pressing plate 7 is provided on the sample test bench 2, which is used to fix the two ends of the PN junction device and electrically contact the electrodes of the PN junction device; the sample test bench 2 is connected to the measuring module 3; the measuring module 3 is connected to the phase-locked amplifier 4, and the phase-locked amplifier 4 is connected to the computer 5.
[0032] The feeding mechanism comprises:
[0033] A sample reel 1.1 for accommodating a PN junction device to be tested,
[0034] A guide wheel 1.2 for guiding a PN junction device;
[0035] A test reel 1.3 for rolling up the tested PN junction device;
[0036] The sample reel 1.1 is connected to a first rotating mechanism;
[0037] The test reel 1.3 is connected to a second rotating mechanism, the second rotating mechanism is connected to a driver 6 via a stepper motor, and the driver 6 is connected to a computer 5 via an RS232 serial port or a USB communication interface.
[0038] The driver 6 mainly provides sufficient current for the normal operation of the stepper motor. The instructions for the actions executed by the driver 6 are generated by the computer 5. The instructions for executing the actions include rotation angle, rotation direction and rotation speed, etc., and are transmitted through RS232 serial port or USB communication.
[0039] The sample test bench 2 includes a base with an insulating surface and a PN junction device fixing groove, and a group of elastic pressing sheets for fixing the PN junction device and electrodes at both ends.
[0040] In actual work, the working mode of automatic feeding of PN junction device samples is: each time the sample reel 1.1 rotates a sample, the tested PN junction device sample will be pulled out from under the elastic pressing sheet by the guide wheel 1.2, and a new PN junction device test sample will be pressed under the elastic pressing sheet accordingly. The PN junction device sample is automatically pressed into the fixed groove on the base of the sample test bench 2 by the elastic pressing sheet, and the electrode contact is kept good.
[0041] The measuring module 3 comprises:
[0042] A test circuit for connecting to the sample test station 2;
[0043] The standard capacitor C0 used for quantitative calculation; the standard capacitor C0 adopts a monolithic capacitor with a capacity of 10nF.
[0044] A measurement circuit for detecting PN junction devices;
[0045] A signal channel interface and a reference channel interface of the lock-in amplifier 4 are provided.
[0046] The basic interface circuit includes a series loop of the test device and the standard capacitor C0; the reference signal of the phase-locked amplifier 4 is used as the excitation signal of the entire loop; the voltage-divided signal on the standard capacitor C0 is used as the measurement signal of the phase-locked amplifier 4; and a voltage follower is used as a connection buffer module between the measurement signal and the phase-locked amplifier 4.
[0047] The lock-in amplifier 4 is a general or self-made circuit module, which is used to realize the correlation detection of the measurement signal and the reference signal, that is, it has the function of extracting the signal submerged in the noise. The lock-in amplifier 4 adopts the DSP7265 dual-phase lock-in amplifier of SignalRecovery Company.
[0048] The computer 5, such as Figure 2As shown, it can mainly communicate with the phase-locked amplifier 4 and the stepper motor driver 6 to realize the working parameter setting of the phase-locked amplifier 4, the bias control of the test device, the collection of the measurement signal, and the control of the rotation of the stepper motor. The data collected by the computer 5 can be converted into the reverse bias-junction capacitance relationship curve of the PN junction device and displayed and updated in real time; the measurement curve can be automatically transformed and fitted; the type of the PN junction device can be automatically determined according to the correlation coefficient of the straight line fitted by the transformed data, and the impurity concentration distribution coefficient or the impurity concentration gradient coefficient can be calculated from its type. When a large number of samples are automatically measured, the type distribution of the same batch of samples can also be automatically given according to the correlation coefficient, providing statistical information that cannot be given by manual testing of a small number of samples.
[0049] A method for automatically measuring impurity concentration / concentration gradient of a PN junction device comprises the following steps:
[0050] Step 1: Connect the lock-in amplifier 4 module, set an analog output terminal of the lock-in amplifier 4 as the reverse bias voltage of the PN junction device to be tested, and set the working parameters of the lock-in amplifier 4 and the working parameters of the cyclic test. For example: the frequency and amplitude of the reference signal; and the working parameters of the cyclic test, such as the range of the reverse bias voltage, the number of data points and the integration time of each point.
[0051] Step 2: For a PN junction device, start the measurement work: gradually change the bias voltage value, wait for the corresponding integration time, read the signal amplitude of the phase-locked signal channel; combine the value of the standard capacitor C0 in the measurement module 3 and the proportional relationship of the impedance divider, and convert the measurement signal into the junction capacitance value of the PN junction device.
[0052] Step 3: Computer 5 stores the relationship data between bias voltage and junction capacitance and draws a curve on the interface; at the same time, the original data of junction capacitance is converted into the relationship between the inverse square of capacitance and the inverse cube of capacitance and bias voltage, and then linear fitting is performed on the transformed data respectively.
[0053] Step 4: According to the size of the correlation coefficient obtained from the fitting result, the type of PN junction is automatically determined, including: abrupt junction and linear gradual junction. If it is classified as an abrupt junction model, the impurity concentration distribution coefficient on the low-doped side of the PN junction device is calculated; if it is classified as a linear gradual junction model, the concentration gradient coefficient of the PN junction device is calculated.
[0054] When automatically measuring a large number of samples, the type distribution of the same batch of samples can also be automatically given according to the correlation coefficient.
[0055] Example:
[0056] The following takes the measurement of a PN junction sample using the automatic measurement system for impurity concentration / concentration gradient of a PN junction device of the present invention as an example to describe the measurement method of the present invention in detail.
[0057] The specific measurement method steps are as follows: Figure 3 As shown:
[0058] 1: Sample loading initialization: Fix the sample reel 1.1 of the PN junction device in the rotating mechanism, manually guide the first few components of the reel paper tape through the elastic pressing piece and load them into the test reel 1.3; test the rotation of the reel 1.3 and debug the stepper motor angle, speed and other parameters to ensure that the subsequent components can be well stuck under the elastic pressing piece.
[0059] 2: Initialization: Connect the computer 5 and the lock-in amplifier 4. The specific connection method can be RS232 serial port or USB communication. After the connection is successful, set the working parameters of the lock-in amplifier 4. Usually, in order to meet the requirements of the measurement signal-to-noise ratio and ensure the smoothness of the measurement curve, the frequency of the test signal is selected to be above 30kHz, and the amplitude of the test signal is V t In order not to affect the actual reverse bias operation, it is set at around 0.1V.
[0060] Set the reverse bias scanning range and number of points, for example: the reverse bias range is from 0V-10V, and the data points include 200. In the computer 5, the bias voltage V required for each cycle test is automatically calculated according to the bias range and the number of test points. R ; The time constant T of the lock-in amplifier 4 c , the typical value is about 500ms, and the waiting time dT at each data point during measurement is automatically calculated. The parameters such as the frequency, amplitude and waiting time at each point of the test excitation signal need to be re-debugged after replacing different types of PN junction devices to ensure that the measurement curve is smooth and has a good signal-to-noise ratio.
[0061] 3: Start the cycle test, collect data and display it in a graph; according to the test logic, first set a bias voltage V R , then wait for dT time, and then read the measurement signal V i For each amplitude value, it can be proportionally converted to the barrier capacitance C according to the relationship between the series loop impedance voltage divider and the value of the standard capacitance C0. x , the conversion relationship is (V i / V t )*C0. The software can draw V in real time when working in a loop. R -C x Figure 4(a) shows a V R -C xTypical measurement results of the relationship curve. The curve can be transformed into the offset voltage V respectively R and and relationship, see the measurement data shown in Figure 4(b) and Figure 4(c). Then perform linear fitting respectively to obtain linear fitting results such as slope and correlation coefficient. After the measurement is completed, cycle and update the offset voltage condition and retest.
[0062] 4: Curve clustering: Due to manufacturing process differences, common PN junction device models include abrupt junctions or linearly graded junctions. Let V R -1 / C 2 The correlation coefficient of the linear fitting of the data is R1, and V R -1 / C 3 The correlation coefficient of the linear fitting of the data is R2. When R1 > R2, it indicates that the linear fitting effect of the V R -1 / C 2 data is better and should be classified into the abrupt junction model, such as the measurement data in Figure 4(b) and Figure 4(c), where the linearity of 4(b) is significantly better.
[0063] Figures 5(a), 5(b), and 5(c) show typical measurement results that conform to the abrupt junction model in a PN junction device of the same model batch. Among them, the linearity of Figure 5(c) is better, that is, R1 < R2. At this time, the linear fitting effect of the V R -1 / C 3 data is better and should be classified into the linearly graded junction model.
[0064] 5: Coefficient solution:
[0065] 1) Solve the impurity concentration distribution coefficient N on the low-doped side in the abrupt junction model D : According to the semiconductor depletion layer theory of P-N junction devices, the following relationship exists: barrier capacitance where is the reduced concentration, and A is the junction area. Therefore, That is, for the abrupt junction model, V R and show a linear relationship. The doping reduced concentration N can be obtained from the slope of the straight line (equal to * ). For the one-sided abrupt junction model, the reduced concentration N * can be replaced by the concentration N D on the low-doped side. In addition, the contact potential V D can also be obtained from the intercept of the straight line.
[0066] 2) Solve the impurity concentration gradient distribution of the linearly graded junction: Similarly, according to the semiconductor depletion layer theory of PN junction devices, for the linearly graded junction, its barrier capacitance Where: G is the impurity concentration gradient. Therefore, That is, for a linear tapered junction, V R and In a straight line relationship. Similarly, the slope of the straight line (equal to ) and the intercept can be used to calculate the impurity concentration gradient G and V D Among them, the method of transforming data to fit a straight line adopts the standard least squares method for linear fitting.
[0067] 6: After the measurement is completed, the system automatically changes to the next sample, and the specific implementation method of the system returns to step 2 and cycles.
[0068] According to the above method, after automatic measurement, the models of PN junction devices of the same model batch are statistically analyzed, for example, cluster analysis is performed on the correlation coefficient measurement results of the transformed data. Figure 6 Typical results of automatic clustering of PN junction devices from the same batch according to correlation coefficients are given, which facilitates further quantification of process differences such as doping.
[0069] The above descriptions are merely embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any modifications, equivalent substitutions, improvements, etc. within the spirit and principles of the present invention should be included in the scope of the claims of the present invention.
Claims
1. A PN junction device impurity concentration / concentration gradient automatic measurement system, characterized by: The system comprises a feeding mechanism, a sample test bench (2), a measuring module (3), a lock-in amplifier (4), and a computer (5); The driver (6) of the feeding mechanism is connected to the computer (5), and the feeding mechanism is used to automatically transfer the PN junction device to the sample test bench (2); the sample test bench (2) is connected to the measurement module (3), and a fixing pressing sheet (7) is provided on the sample test bench (2) for fixing the two ends of the PN junction device and electrically contacting the electrodes of the PN junction device; the sample test bench (2) is connected to the measurement module (3); the measurement module (3) is connected to the lock-in amplifier (4), and the lock-in amplifier (4) is connected to the computer (5); The sample test bench (2) comprises a base having an insulating surface and a PN junction device fixing groove, and a group of fixing pressing sheets (7) for fixing the PN junction device and electrodes at both ends; The feeding mechanism comprises: A sample reel (1.1) for mounting a PN junction device to be tested, A guide wheel (1.2) for guiding a PN junction device; A test reel (1.3) for rolling up the tested PN junction device; The sample reel (1.1) is connected to a first rotating mechanism; The test reel (1.3) is connected to a second rotating mechanism, and the second rotating mechanism is connected to a driver (6) via a stepping motor; The measurement module (3) converts the weak voltage signal of the PN junction device collected by the fixed pressing plate (7) into the junction capacitance of the PN junction device and can display the relationship curve between the bias voltage and the junction capacitance of the PN junction device in real time; further, the curve can be automatically transformed and fitted; and the type of the PN junction can be determined according to the correlation coefficient of the fitted straight line result; and the impurity concentration distribution coefficient or concentration gradient coefficient of the PN junction device can be converted according to the physical model corresponding to the type; The working method of the automatic feeding of PN junction device samples is as follows: each time the sample reel (1.1) rotates a sample, the tested PN junction device sample will be pulled out from under the fixed pressing plate (7) by the guide wheel (1.2), and correspondingly a new PN junction device test sample will be pressed under the fixed pressing plate (7); the PN junction device sample is automatically pressed by the fixed pressing plate (7) into the fixed groove on the base of the sample test table (2), and the electrode contact is kept good.
2. According to claim 1, a PN junction device impurity concentration / concentration gradient automatic measurement system is characterized by: The driver (6) is connected to the computer (5) via an RS232 serial port or a USB communication interface.
3. According to claim 1, a PN junction device impurity concentration / concentration gradient automatic measurement system is characterized by: The measuring module (3) comprises: A test circuit for connecting to a sample test bench (2); Standard capacitance for quantitative calculations C 0; A measurement circuit for detecting PN junction devices; Providing a signal channel interface and a reference channel interface of a lock-in amplifier (4); The basic interface circuit includes the test device and the standard capacitor C 0 series loop; the reference signal of the phase-locked amplifier (4) is used as the excitation signal of the entire loop; the standard capacitor C The voltage-divided signal on 0 is used as a measurement signal of the lock-in amplifier (4); a voltage follower is used as a connection buffer module between the measurement signal and the lock-in amplifier (4).
4. The automatic measurement system for impurity concentration / concentration gradient of a PN junction device according to claim 1, characterized in that: The computer (5) is used to realize the operation parameter setting of the lock-in amplifier (4), the bias control of the PN junction device, the cycle measurement parameters, the measurement signal collection, and the data visualization.
5. A method for automatically measuring the impurity concentration / concentration gradient of a PN junction device using the automatic measurement system for the impurity concentration / concentration gradient of a PN junction device according to claim 1, characterized in that The following steps are involved: Step 1: setting an analog output terminal of the lock-in amplifier (4) as the reverse bias voltage of the PN junction device to be tested, setting the operating parameters of the lock-in amplifier (4) and the operating parameters of the cyclic test; Step 2: For a PN junction device, start the measurement work: gradually change the bias voltage value, wait for the corresponding integration time, read the signal amplitude of the phase-locked signal channel; combine the standard capacitor in the measurement module (3) C 0 and the proportional relationship between the impedance voltage divider, converting the measurement signal into the junction capacitance value of the PN junction device; Step 3: The computer (5) stores the data of the relationship between the bias voltage and the junction capacitance and draws a curve on the interface; at the same time, the original data of the junction capacitance is converted into the relationship between the inverse square of the capacitance and the inverse cube of the capacitance and the bias voltage, and then linear fitting is performed on the transformed data respectively; Step 4: According to the size of the correlation coefficient obtained from the fitting result, the type of PN junction is automatically determined, including: abrupt junction and linear gradual junction. If it is classified as an abrupt junction model, the impurity concentration distribution coefficient on the low-doped side of the PN junction device is calculated; if it is classified as a linear gradual junction model, the concentration gradient coefficient of the PN junction device is calculated.
6. The method for automatically measuring impurity concentration / concentration gradient of a PN junction device according to claim 5, characterized in that: When automatically measuring a large number of samples, the type distribution of the same batch of samples can also be automatically given according to the correlation coefficient.
