Silicon carbide multilayer epitaxial doping concentration longitudinal distribution CV data processing method

Abstract

The invention relates to the technical field of semiconductors, and provides a silicon carbide multilayer epitaxial doping concentration longitudinal distribution CV data processing method, which comprises the following steps of: 1, carrying out capacitance-voltage test on a silicon carbide multilayer epitaxial wafer to obtain original capacitance-voltage data; step 2, based on a Schottky contact model, determining built-in potential and initial depletion layer width when a test system is in contact with the epitaxial wafer through calculation or fitting; and step 3, performing reconstruction calibration on the depth coordinates in the original data, subtracting the initial depletion layer width from the original depth value calculated by the voltage test or introducing a calibration offset, and establishing a new coordinate system taking the physical surface as a zero point. Coordinates are corrected through a 1 / C2-V extrapolation method of a Schottky contact model, complete doping concentration distribution is restored, the low doping characteristics of the ultrathin surface layer can be detected, an interface is accurately defined, and data fit a design value; high-resolution and nondestructive detection is realized by combining conventional mercury probe CV equipment with an algorithm, the method is simple, convenient and efficient, the cost is reduced, and the method is suitable for mass production monitoring.

Description

Technical Field

[0001] This invention relates to the field of semiconductor technology, specifically to a CV data processing method for the vertical distribution of doping concentration in silicon carbide multilayer epitaxial layers. Background Technology

[0002] Silicon carbide (SiC), as a representative of third-generation wide-bandgap semiconductor materials, has become the preferred material for high-voltage, high-temperature, and high-frequency power semiconductor devices due to its excellent physical properties such as high breakdown electric field strength, high thermal conductivity, and high electron saturation drift velocity. The application of SiC power devices in electric vehicles, photovoltaic inverters, smart grids, and rail transportation significantly reduces the energy consumption and system size of power electronic systems, providing key technological support for achieving the "dual carbon" goal.

[0003] As SiC device design advances towards higher performance, epitaxial layer structures with single concentrations and thicknesses are no longer sufficient to meet the demands of complex devices. To balance breakdown voltage and on-resistance, or to optimize electric field distribution, modern high-performance SiC devices typically employ multilayer epitaxial structures. For example, a current spreading layer or a specially doped buffer layer is grown on top of a drift layer. This multilayer structure requires extremely precise control over the thickness and doping concentration of each epitaxial layer. Any deviation of the parameters from the design values ​​in any layer can lead to breakdown voltage failure, increased leakage current, or increased conduction losses.

[0004] To ensure the quality of multilayer epitaxial wafers, precise monitoring of their vertical doping concentration distribution is essential. Traditional monitoring methods primarily rely on capacitance-voltage (CV) testing. Accurately obtaining the actual doping concentration and thickness information of each layer from the surface to its depth is crucial for feedback adjustments in epitaxial growth processes, yield improvement, and device failure analysis. If the test results fail to accurately reflect the physical location and concentration interfaces of each layer, it will mislead process engineers, resulting in significant R&D and production waste. Therefore, a CV data processing method for the vertical doping concentration distribution of silicon carbide multilayer epitaxial wafers is proposed. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a CV data processing method for the vertical distribution of doping concentration in silicon carbide multilayer epitaxial wafers, which solves the problems of difficulty in accurately monitoring the vertical doping concentration distribution in SiC multilayer epitaxial wafers and the tendency of traditional testing to be distorted, leading to misjudgments of the process.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a CV data processing method for the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers, comprising the following steps: Step 1: Perform capacitance-voltage testing on the silicon carbide multilayer epitaxial wafer to obtain the raw capacitance-voltage data; Step 2: Based on the Schottky contact model, determine the built-in potential and initial depletion layer width when the test system contacts the epitaxial wafer by calculation or fitting. Step 3: Reconstruct and calibrate the depth coordinates in the original data by subtracting the initial depletion layer width from the original depth value calculated by the voltage test or by introducing a calibration offset to establish a new coordinate system with the physical surface as the zero point. Step 4: Based on the calibrated depth coordinates, remap the corresponding doping concentration values ​​to generate a complete longitudinal distribution curve starting from the surface.

[0007] Preferably, the voltage test in step one is performed using a mercury probe capacitance voltage tester, and the test conditions are 1MHz high frequency.

[0008] Preferably, step two, determining the built-in potential and the initial depletion layer width, includes the following steps: S1. Process the capacitance data from the voltage test and plot 1 / C. 2 -V curve; S2, 1 / C of the top layer of a silicon carbide multilayer epitaxial layer 2 The -V curve is linearly fitted to obtain a fitted straight line; S3. Extrapolate the fitted straight line to 1 / C 2 =0, at this point the voltage value at the intersection is the built-in potential, that is, Vbi=Vint; S4. Calculate the initial depletion layer width based on the Schottky junction depletion layer width formula, which is: ; in The dielectric constant of silicon carbide semiconductor, For built-in potential, For external bias voltage, For electron charge, Net doping concentration; when When V = 0V, the initial depletion layer width is calculated.

[0009] Preferably, in step three, the 1 / C 2 The region where the -V curve is linearly fitted is the lower voltage segment where the curve linearity is best, and this region corresponds to the uniform region of the first layer at the top of the multilayer epitaxial layer.

[0010] Preferably, in step one, the silicon carbide multilayer epitaxial wafer is grown using a metal-organic chemical vapor deposition epitaxial device, and the longitudinal doping distribution between layers is either gradual or stepped.

[0011] Preferably, the original depth value in step three is calculated using the following formula: ; in The dielectric constant of silicon carbide semiconductor, This represents the contact area between the mercury probe and the epitaxial wafer. The capacitance value is obtained from the CV test.

[0012] Preferably, the doping concentration value in step four is calculated using the following formula: ; in For electron charge, The dielectric constant of silicon carbide semiconductor, For contact area, This is the capacitance value. This is an external bias voltage.

[0013] Preferably, the silicon carbide multilayer epitaxial wafer comprises at least three epitaxial layers, each epitaxial layer having a thickness ranging from 0.2 μm to 25 μm and a doping concentration ranging from 2.45E+15cm. -3 -3.7E+15cm -3 Based on the generated complete longitudinal distribution curve, the doping abrupt change points between each layer are accurately identified, and the doping abrupt change points correspond to the interface positions of adjacent epitaxial layers.

[0014] A detection system for the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers includes: A data acquisition unit is used to acquire raw capacitance-voltage data of a silicon carbide multilayer epitaxial wafer through a CV testing device. The fitting analysis unit is used to process the raw capacitance data and plot 1 / C. 2 -V curves are plotted and linearly fitted, and extrapolated to obtain the built-in potential; A parameter calculation unit is used to calculate the initial depletion layer width and the original depth value based on the relevant formula of the Schottky junction. A coordinate calibration unit is used to subtract the initial depletion layer width from the original depth value to establish a new coordinate system with the physical surface as the zero point. A distribution generation unit is used to generate a complete vertical distribution curve of doping concentration based on the calibrated coordinate system and the calculation results of doping concentration. An interface recognition unit is used to identify doping abrupt change points between epitaxial layers based on the longitudinal distribution curve.

[0015] Preferably, the data acquisition unit is a mercury probe capacitance voltage test device, the test frequency of the mercury probe capacitance voltage test device is 1MHz, and the parameter calculation unit has built-in formulas for Schottky junction depletion layer width, original depth value calculation and doping concentration calculation.

[0016] This invention provides a CV data processing method for the vertical distribution of doping concentration in silicon carbide multilayer epitaxial layers. It has the following beneficial effects: 1. This invention utilizes a 1 / C model based on the Schottky contact model. 2 The -V extrapolation method accurately calculates the initial depletion layer width and corrects the depth coordinates, successfully reconstructing the complete doping concentration distribution starting from the physical surface of the epitaxial wafer. For the ultrathin epi-3 surface layer with a thickness of only 0.2 μm, existing technologies cannot detect its low-doping characteristics, while this invention can clearly present the concentration information of this layer, and accurately define the interface positions of each layer, making the test data highly consistent with the actual design values ​​of epitaxial growth, providing a reliable basis for process adjustment and quality control.

[0017] 2. This invention achieves depth resolution comparable to SIMS by combining conventional mercury probe CV testing equipment with data post-processing algorithms. The testing process does not require sample destruction, the operation procedure is simple, significantly reducing the detection cost in the R&D and production stages, while improving detection efficiency and meeting the rapid monitoring needs in large-scale mass production scenarios. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the structure of the silicon carbide multilayer epitaxial wafer of the present invention; Figure 2 This is a flowchart of the present invention; Figure 3 This is a CV test curve of the silicon carbide multilayer epitaxial wafer of the present invention; Figure 4 This is the 1 / C2-V curve diagram of the present invention; Figure 5 This is a fitting diagram of the 1 / C2-V curve of the uppermost epi-3 layer of the silicon carbide multilayer epitaxial layer of the present invention. Figure 6 This is a graph showing the longitudinal distribution of doped multilayer epitaxial layers without correction of the depletion region width in this invention. Figure 7 This is a graph showing the longitudinal distribution of doped multilayer epitaxial layer after the depletion region width has been corrected according to the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see the appendix Figure 1 - Appendix Figure 3This invention provides a CV data processing method for the vertical distribution of doping concentration in silicon carbide multilayer epitaxial layers, comprising the following steps: Step 1: Perform capacitance-voltage testing on the silicon carbide multilayer epitaxial wafer to obtain raw capacitance-voltage data. The voltage test in Step 1 is performed using a mercury probe capacitance-voltage testing device at a high frequency of 1MHz. The silicon carbide multilayer epitaxial wafer in Step 1 is grown using a metal-organic chemical vapor deposition (MOCVD) epitaxial device. The vertical doping distribution between layers is gradient or stepped. The silicon carbide multilayer epitaxial wafer includes at least three epitaxial layers, each with a thickness ranging from 0.2μm to 25μm and a doping concentration ranging from 2.45E+15cm. -3 -3.7E+15cm -3 Based on the generated complete longitudinal distribution curve, the doping abrupt points between each layer are accurately identified, and the doping abrupt points correspond to the interface positions of adjacent epitaxial layers. Specifically, firstly, a multilayer silicon carbide epitaxial wafer containing three layers (epi-1, epi-2, and epi-3) was grown by controlling the epitaxial growth parameters using a metal-organic chemical vapor deposition (MOCVD) multilayer epitaxial equipment. Gradual doping transitions were used between layers to ensure that the initial design parameters of each layer accurately matched the preset values. Subsequently, a mercury probe capacitance-voltage (MCV) tester was used to perform capacitance-voltage tests on the prepared multilayer epitaxial wafer under strict 1MHz high-frequency test conditions. The test environment was kept stable throughout the process to avoid data deviation caused by external interference, and finally, complete and reliable original capacitance-voltage data were obtained.

[0021] Please see the appendix Figure 3 - Appendix Figure 5 Step 2: Based on the Schottky contact model, determine the built-in potential and initial depletion layer width when the test system contacts the epitaxial wafer by calculation or fitting. Determining the built-in potential and initial depletion layer width in Step 2 includes the following steps: S1. Process the capacitance data from the voltage test and plot 1 / C. 2 -V curve; S2, 1 / C of the top layer of a silicon carbide multilayer epitaxial layer 2 The -V curve is linearly fitted to obtain a fitted straight line; S3. Extrapolate the fitted straight line to 1 / C 2 =0, at this point the voltage value at the intersection is the built-in potential, that is, Vbi=Vint; S4. Calculate the initial depletion layer width based on the Schottky junction depletion layer width formula, which is: ; in The dielectric constant of silicon carbide semiconductor, For built-in potential, For external bias voltage, For electron charge, Net doping concentration; when When V = 0V, the initial depletion layer width is calculated. Specifically, first export the raw capacitance data obtained in step one to data processing software, filter the data to remove outliers, and then process it according to 1 / C. 2 Plot the relationship between 1 / C and voltage V. 2 The -V curve clearly shows three straight line segments corresponding to the epi-1, epi-2, and epi-3 layers, conforming to the theoretical characteristics of an ideal single-sided abrupt junction. Next, the lower voltage segment with the best linearity (corresponding to the uniform region of the epi-3 layer) is selected, and the least squares method is used to linearly fit this segment, obtaining the fitted straight line equation y = a + b*x, where the fitting parameters a = 1.0119E19 and b = -8.07924E18. The fitted straight line is then extrapolated to 1 / C... 2 At the position where =0, the corresponding voltage value Vint = 1.2524V, which is the built-in potential. Finally, =1.2524V, Va=0V =1.03×10^-10F / m, q=1.602×10^-19C, Substituting 3.5E+15cm-3 into the formula for the depletion layer width of a Schottky junction, the initial depletion layer width W0 is calculated to be 0.6198μm.

[0022] Please see the appendix Figure 5 - Appendix Figure 7 Step 3: Reconstruct and calibrate the depth coordinates in the original data. Subtract the initial depletion layer width from the original depth value calculated by the voltage test, or introduce a calibration offset, to establish a new coordinate system with the physical surface as the zero point. In step 3, the 1 / C... 2 The region where the -V curve is linearly fitted is the lower voltage segment where the curve linearity is best. This region corresponds to the uniform region of the first layer at the top of the multilayer epitaxial layer. The original depth value in step three is calculated using the following formula: ; in The dielectric constant of silicon carbide semiconductor, This represents the contact area between the mercury probe and the epitaxial wafer. The capacitance value is obtained from the CV test; Specifically, based on the above formula for calculating the original depth value, =1.03×10^-10F / m、A=1×10^-4cm2 Substituting the capacitance values ​​C obtained from the CV test into the original depth values ​​Wi, we calculated the corresponding original depth values ​​Wi. Then, we subtracted the initial depletion layer width W0 = 0.6198 μm obtained in step two from each original depth value Wi to complete the reconstruction and calibration of the depth coordinates. A new coordinate system with the physical surface of the silicon carbide multilayer epitaxial wafer as the zero point (Depth = 0 μm) was successfully established, correcting the overall rightward shift error of the depth axis caused by the initial depletion layer in the original data.

[0023] Please see the appendix Figure 3 Appendix Figure 4 and attached Figure 7 Step four: Based on the calibrated depth coordinates, remap the corresponding doping concentration values ​​to generate a complete longitudinal distribution curve starting from the surface. The doping concentration values ​​in step four are calculated using the following formula: ; in For electron charge, The dielectric constant of silicon carbide semiconductor, For contact area, This is the capacitance value. For external bias voltage; Specifically, first, regarding 1 / C 2 Differentiating the -V curve yields d(1 / C) 2 The value of ) / dV; then q=1.602×10^-19C, =1.03×10^-10F / m、A=1×10^-4cm 2 d(1 / C 2 Substituting the value of ) / dV into the doping concentration calculation formula, the doping concentration value corresponding to each original depth value Wi is calculated; finally, based on the depth coordinates xi calibrated in step three, the corresponding doping concentration values ​​are remapped, and a complete longitudinal distribution curve of doping concentration starting from the epitaxial wafer surface (0 μm) is plotted. The curve shows that the concentration of the epi-1 layer is 3.3E+15cm. -3 The epi-2 layer concentration is 2.45E+15cm. -3 The epi-3 layer concentration is 3.7E+15cm. -3 The thickness of the interlayer transition layer is 0.1 μm, which is highly consistent with the expected epitaxial wafer structure parameters.

[0024] A detection system for the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers includes: The data acquisition unit is used to acquire the original capacitance-voltage data of the silicon carbide multilayer epitaxial wafer through a CV testing device. The data acquisition unit is a mercury probe capacitance-voltage testing device with a testing frequency of 1MHz. The parameter calculation unit has built-in formulas for Schottky junction depletion layer width, original depth value, and doping concentration. The fitting analysis unit is used to process the raw capacitance data and plot 1 / C. 2 -V curves are plotted and linearly fitted, and extrapolated to obtain the built-in potential; A parameter calculation unit is used to calculate the initial depletion layer width and the original depth value based on the relevant formula of the Schottky junction. A coordinate calibration unit is used to subtract the initial depletion layer width from the original depth value to establish a new coordinate system with the physical surface as the zero point. A distribution generation unit is used to generate a complete vertical distribution curve of doping concentration based on the calibrated coordinate system and the calculation results of doping concentration. An interface recognition unit is used to identify doping abrupt change points between epitaxial layers based on the longitudinal distribution curve. Specifically, during the operation of the detection system, the data acquisition unit (mercury probe capacitance-voltage testing equipment) first acquires the raw capacitance-voltage data of the silicon carbide multilayer epitaxial wafer at a high frequency of 1MHz, and transmits the data to the fitting analysis unit. After receiving the data, the fitting analysis unit preprocesses the raw capacitance data and plots the 1 / C... 2 The -V curve is then linearly fitted to the linear segment corresponding to the epi-3 layer in the curve, and the built-in potential is obtained by extrapolation. =1.2524V, and transmits the relevant data to the parameter calculation unit; the parameter calculation unit calls the built-in Schottky junction depletion layer width formula and original depth value calculation formula, and substitutes the known parameters ( =1.03×10^-10F / m, q=1.602×10^-19C, A=1×10^-4cm 2 The initial depletion layer width W0 = 0.6198 μm and the original depth values ​​Wi are calculated. After receiving Wi and W0, the coordinate calibration unit subtracts W0 from Wi to complete the depth coordinate calibration and establishes a new coordinate system. The distribution generation unit combines the calibrated coordinate system with the doping concentration value obtained by the parameter calculation unit to generate a complete longitudinal distribution curve of doping concentration. Finally, based on the curve, the interface recognition unit accurately identifies the doping abrupt change points between epi-1 and epi-2, and between epi-2 and epi-3, which correspond to the positions at depths of 1.0 μm and 0.2 μm, respectively, to achieve accurate positioning of the interface positions of each epitaxial layer.

[0025] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for processing CV data on the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers, characterized in that, Includes the following steps: Step 1: Perform capacitance-voltage testing on the silicon carbide multilayer epitaxial wafer to obtain the raw capacitance-voltage data; Step 2: Based on the Schottky contact model, determine the built-in potential and initial depletion layer width when the test system contacts the epitaxial wafer by calculation or fitting. Step 3: Reconstruct and calibrate the depth coordinates in the original data by subtracting the initial depletion layer width from the original depth value calculated by the voltage test or by introducing a calibration offset to establish a new coordinate system with the physical surface as the zero point. Step 4: Based on the calibrated depth coordinates, remap the corresponding doping concentration values ​​to generate a complete longitudinal distribution curve starting from the surface.

2. The CV data processing method for the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers according to claim 1, characterized in that, In step one, the voltage test is performed using a mercury probe capacitance voltage tester under the condition of 1MHz high frequency.

3. The CV data processing method for the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers according to claim 1, characterized in that, Step two, which determines the built-in potential and the initial depletion layer width, includes the following steps: S1. Process the capacitance data from the voltage test and plot 1 / C. 2 -V curve; S2, 1 / C of the top layer of a silicon carbide multilayer epitaxial layer 2 The -V curve is linearly fitted to obtain a fitted straight line; S3. Extrapolate the fitted straight line to 1 / C 2 =0, at this point the voltage value at the intersection is the built-in potential, that is, Vbi=Vint; S4. Calculate the initial depletion layer width based on the Schottky junction depletion layer width formula, which is: ； in The dielectric constant of silicon carbide semiconductor, For built-in potential, For external bias voltage, For electron charge, Net doping concentration; when When V = 0V, the initial depletion layer width is calculated.

4. The CV data processing method for the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers according to claim 1, characterized in that, In step three, 1 / C 2 The region where the -V curve is linearly fitted is the lower voltage segment where the curve linearity is best, and this region corresponds to the uniform region of the first layer at the top of the multilayer epitaxial layer.

5. The CV data processing method for the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers according to claim 1, characterized in that, In step one, the silicon carbide multilayer epitaxial wafer is grown using a metal-organic chemical vapor deposition epitaxial device, and the longitudinal doping distribution between layers is either gradient or step-like.

6. The CV data processing method for the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers according to claim 1, characterized in that, The original depth value in step three is calculated using the following formula: ; in The dielectric constant of silicon carbide semiconductor, This represents the contact area between the mercury probe and the epitaxial wafer. The capacitance value is obtained from the CV test.

7. The CV data processing method for the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers according to claim 1, characterized in that, The doping concentration value in step four is calculated using the following formula: ； in For electron charge, The dielectric constant of silicon carbide semiconductor, For contact area, This is the capacitance value. This is an external bias voltage.

8. The CV data processing method for the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers according to claim 1, characterized in that, The silicon carbide multilayer epitaxial wafer comprises at least three epitaxial layers, each with a thickness ranging from 0.2 μm to 25 μm and a doping concentration ranging from 2.45E+15cm. -3 -3.7E+15cm -3 Based on the generated complete longitudinal distribution curve, the doping abrupt change points between each layer are accurately identified, and the doping abrupt change points correspond to the interface positions of adjacent epitaxial layers.

9. A detection system for the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers, using the CV data processing method for the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers as described in claims 1-8, characterized in that, include: A data acquisition unit is used to acquire raw capacitance-voltage data of a silicon carbide multilayer epitaxial wafer through a CV testing device. The fitting analysis unit is used to process the raw capacitance data and plot 1 / C. 2 -V curves are plotted and linearly fitted, and extrapolated to obtain the built-in potential; A parameter calculation unit is used to calculate the initial depletion layer width and the original depth value based on the relevant formula of the Schottky junction. A coordinate calibration unit is used to subtract the initial depletion layer width from the original depth value to establish a new coordinate system with the physical surface as the zero point. A distribution generation unit is used to generate a complete vertical distribution curve of doping concentration based on the calibrated coordinate system and the calculation results of doping concentration. An interface recognition unit is used to identify doping abrupt change points between epitaxial layers based on the longitudinal distribution curve.

10. The detection system for the longitudinal distribution of doping concentration in silicon carbide multilayer epitaxial layers according to claim 9, characterized in that, The data acquisition unit is a mercury probe capacitance voltage test device with a test frequency of 1MHz. The parameter calculation unit has built-in formulas for Schottky junction depletion layer width, original depth value, and doping concentration.

Images