Search

CN-121978495-A - CV data processing method for longitudinal distribution of doping concentration of silicon carbide multilayer epitaxy

CN121978495ACN 121978495 ACN121978495 ACN 121978495ACN-121978495-A

Abstract

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

Inventors

  • YAN JIAN
  • HAN LIXIANG
  • HE XIAOFENG

Assignees

  • 常山森思功率半导体有限公司

Dates

Publication Date
20260505
Application Date
20260204

Claims (10)

  1. 1. The CV data processing method for longitudinal distribution of silicon carbide multilayer epitaxial doping concentration is characterized by comprising the following steps: step one, carrying out capacitance-voltage test on a silicon carbide multilayer epitaxial wafer to obtain original capacitance-voltage data; Step two, based on a Schottky contact model, determining the built-in potential and the initial depletion layer width when a test system is in contact with an epitaxial wafer through calculation or fitting; Step three, carrying out reconstruction calibration on depth coordinates in original data, subtracting the initial depletion layer width from an original depth value calculated by a voltage test or introducing a calibration offset, and establishing a new coordinate system taking a physical surface as a zero point; And step four, remapping corresponding doping concentration values based on the calibrated depth coordinates, and generating a complete longitudinal distribution curve starting from the surface.
  2. 2. The method for processing CV data of longitudinal distribution of epitaxial doping concentration of silicon carbide according to claim 1, wherein in the first step, the voltage test is performed by a mercury probe capacitor voltage test device under a test condition of 1MHz high frequency.
  3. 3. The CV data processing method for longitudinal distribution of epitaxial doping concentrations of silicon carbide according to claim 1, wherein determining the built-in potential and the initial depletion layer width in the second step includes the steps of: s1, processing capacitance data of a voltage test, and drawing a 1/C 2 -V curve; S2, linearly fitting a 1/C 2 -V curve of the uppermost layer of the silicon carbide multilayer epitaxy to obtain a fitting straight line; S3, extrapolating the fitting straight line to 1/C 2 =0, wherein the voltage value at the intersection point is the built-in potential, namely Vbi=vint; S4, calculating the initial depletion layer width based on a Schottky junction depletion layer width formula, wherein the Schottky junction depletion layer width formula is as follows: ; Wherein the method comprises the steps of Is the dielectric constant of a silicon carbide semiconductor, For the purpose of the built-in potential, In order to apply the bias voltage to the substrate, In the form of an electrical charge of an electron, Is of net doping concentration when When=0v, the initial depletion layer width is calculated.
  4. 4. The method according to claim 1, wherein the area of the linear fitting of the 1/C 2 -V curve in the third step is a lower voltage segment with the best linearity of the curve, and the area corresponds to the uniform area of the first layer on top of the multilayer epitaxy.
  5. 5. The CV data processing method for longitudinal distribution of doping concentration of silicon carbide multilayer epitaxy according to claim 1, wherein in the step one, the silicon carbide multilayer epitaxy wafer is grown by a metal organic chemical vapor deposition epitaxy device, and the longitudinal doping distribution between layers is graded or stepped.
  6. 6. The CV data processing method of longitudinal distribution of epitaxial doping concentration of silicon carbide according to claim 1, wherein the raw depth value in the third step is calculated by the following formula: ; Wherein the method comprises the steps of Is the dielectric constant of a silicon carbide semiconductor, Is the contact area between the mercury probe and the epitaxial wafer, The capacitance value obtained for the CV test.
  7. 7. The CV data processing method for longitudinal distribution of epitaxial doping concentration of silicon carbide according to claim 1, wherein the doping concentration value in the fourth step is calculated by the following formula: ; Wherein the method comprises the steps of In the form of an electrical charge of an electron, Is the dielectric constant of a silicon carbide semiconductor, In order for the contact area to be a contact area, In order to be a value of the capacitance, Is externally biased.
  8. 8. The CV data processing method of longitudinal distribution of doping concentration of silicon carbide multilayer epitaxy according to claim 1, wherein the silicon carbide multilayer epitaxy wafer includes at least three epitaxial layers, each epitaxial layer has a thickness ranging from 0.2 μm to 25 μm, a doping concentration ranging from 2.45e+15cm -3 -3.7E+15cm -3 , and doping abrupt points between layers are accurately identified based on the generated complete longitudinal distribution curve, and the doping abrupt points correspond to interface positions of adjacent epitaxial layers.
  9. 9. A system for detecting a longitudinal distribution of epitaxial doping concentration of silicon carbide, using a CV data processing method for a longitudinal distribution of epitaxial doping concentration of silicon carbide according to claims 1 to 8, comprising: The data acquisition unit is used for acquiring original capacitance-voltage data of the silicon carbide multilayer epitaxial wafer through CV test equipment; The fitting analysis unit is used for processing the original capacitance data, drawing a 1/C 2 -V curve, performing linear fitting, and extrapolating to obtain built-in potential; the parameter calculation unit is used for calculating an initial depletion layer width and an original depth value based on a Schottky junction related formula; the coordinate calibration unit is used for subtracting the initial depletion layer width from the original depth value and establishing a new coordinate system taking the physical surface as a zero point; The distribution generation unit is used for generating a complete doping concentration longitudinal distribution curve based on the calibrated coordinate system and the doping concentration calculation result; and the interface identification unit is used for identifying doping mutation points among the epitaxial layers based on the longitudinal distribution curve.
  10. 10. The system for detecting longitudinal distribution of epitaxial doping concentration of silicon carbide according to claim 9, wherein the data acquisition unit is mercury probe capacitance voltage testing equipment, the testing frequency of the mercury probe capacitance voltage testing equipment is 1MHz, and the parameter calculation unit is internally provided with a schottky junction depletion layer width formula, an original depth value calculation formula and a doping concentration calculation formula.

Description

CV data processing method for longitudinal distribution of doping concentration of silicon carbide multilayer epitaxy Technical Field The invention relates to the technical field of semiconductors, in particular to a CV data processing method for longitudinal distribution of silicon carbide multilayer epitaxial doping concentration. Background Silicon carbide, which is a representative of the third generation of wide bandgap semiconductor materials, has become the material of choice for high voltage, high temperature and high frequency power semiconductor devices by virtue of its excellent physical properties such as high breakdown field strength, high thermal conductivity, high electron saturation drift rate, etc. The SiC power device is applied to the fields of electric automobiles, photovoltaic inverters, smart grids, rail transit and the like, the energy consumption of a power electronic system is obviously reduced, the system volume is reduced, and the SiC power device is a key technical support for achieving a double-carbon target. As SiC device designs move toward higher performance, single concentration and thickness epitaxial layer structures have been difficult to meet the requirements of complex devices. Modern high performance SiC devices typically employ multilayer epitaxial structures for balancing breakdown voltage and on-resistance, or for optimizing electric field distribution. For example, a current spreading layer or a specifically doped buffer layer is grown over the drift layer. Such a multilayer structure requires that the thickness and doping concentration of each epitaxial growth must be extremely precisely controlled. Deviations of parameters of any one layer from design values may lead to voltage withstand failure, increased leakage current, or increased conduction loss of the device. In order to ensure the quality of the multilayer epitaxial wafer, the longitudinal doping concentration distribution thereof must be precisely monitored. Traditional monitoring approaches rely primarily on capacitance-voltage testing. The method accurately acquires the actual doping concentration and thickness information of each layer from the surface to the deep, and is important for feedback adjustment of an epitaxial growth process, yield improvement and device failure analysis. If the test result cannot truly reflect the physical position and concentration interface of each layer, a process engineer is misled, huge research and development and production waste are caused, and the CV data processing method for longitudinally distributing the epitaxial doping concentration of the silicon carbide multilayer is provided. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a CV data processing method for longitudinal distribution of silicon carbide multilayer epitaxial doping concentration, which solves the problems that the longitudinal doping concentration distribution of a SiC multilayer epitaxial wafer is difficult to accurately monitor and the conventional test is easy to distort so as to cause process misjudgment. The CV data processing method for longitudinally distributing the doping concentration of the silicon carbide multilayer epitaxy comprises the following steps: step one, carrying out capacitance-voltage test on a silicon carbide multilayer epitaxial wafer to obtain original capacitance-voltage data; Step two, based on a Schottky contact model, determining the built-in potential and the initial depletion layer width when a test system is in contact with an epitaxial wafer through calculation or fitting; Step three, carrying out reconstruction calibration on depth coordinates in original data, subtracting the initial depletion layer width from an original depth value calculated by a voltage test or introducing a calibration offset, and establishing a new coordinate system taking a physical surface as a zero point; And step four, remapping corresponding doping concentration values based on the calibrated depth coordinates, and generating a complete longitudinal distribution curve starting from the surface. Preferably, the voltage test in the first step is performed by mercury probe capacitance voltage test equipment, and the test condition is 1MHz high frequency. Preferably, the determining the built-in potential and the initial depletion layer width in the second step includes the following steps: s1, processing capacitance data of a voltage test, and drawing a 1/C 2 -V curve; S2, linearly fitting a 1/C 2 -V curve of the uppermost layer of the silicon carbide multilayer epitaxy to obtain a fitting straight line; S3, extrapolating the fitting straight line to 1/C 2 =0, wherein the voltage value at the intersection point is the built-in potential, namely Vbi=vint; S4, calculating the initial depletion layer width based on a Schottky junction depletion layer width formula, wherein the Schottky junction depletion layer width formula is as follows: ; Wherein