Wide bandgap semiconductor defect state detection method based on time-of-flight technology

Abstract

The invention discloses a wide bandgap semiconductor defect state detection method based on a time-of-flight technology. The detection method comprises the following steps: pretreating a wide bandgap semiconductor sample, preparing a detection sample, measuring a transient photocurrent curve of the detection sample, setting Monte Carlo model parameters, and fitting to obtain defect parameter information in the detection sample. The detection method is simple in test condition, can accurately obtain key parameters such as defect state energy level position, concentration, capture cross section and the like at normal temperature, and plays an important role in researching defect information in a wide bandgap semiconductor material and the influence of the defect information on the reliability of a device.

Description

Technical Field

[0001] This invention relates to a semiconductor defect state detection method, and more particularly to a wide-bandgap semiconductor defect state detection method based on time-of-flight technology. Background Technology

[0002] Wide bandgap semiconductor materials, including silicon carbide, gallium nitride, gallium oxide, and diamond, are preferred materials for the development of power electronic devices due to their large bandgap and high breakdown field strength, and have significant application prospects in rail transportation, aerospace, automotive electronics, new energy, and defense. During the growth or device fabrication process of wide bandgap semiconductors, impurities or defects are intentionally or unintentionally introduced. These defects form defect energy levels in the bandgap, which will seriously affect the performance and reliability of the device. For example, deep-level and shallow-level defects have different effects in semiconductors. Deep-level defects mainly affect carrier recombination and the reverse characteristics of the device, while shallow-level defects affect the conductivity of the device by changing the carrier concentration and conductivity. Therefore, how to detect defect types and extract defect parameters (defect density, energy level position, trapping cross-section) through effective testing methods is of great significance for optimizing the quality of wide bandgap semiconductor materials and device fabrication.

[0003] Currently, for wide-bandgap semiconductor materials, defect energy level detection mainly employs deep-level transient spectroscopy (DLTS) and its derivative, deep-level optical spectroscopy (DLOS). DLTS involves fabricating the semiconductor material into a pn junction, Schottky junction, or MOS capacitor structure, applying a periodic, short voltage pulse to its space charge region, and measuring the transient change in the barrier capacitance of the space charge region. By combining the capacitance transient curves at different temperatures with mathematical processing methods, an Arrhenius plot is plotted, thereby obtaining the energy level location, concentration, and trapping cross-section of the deep-level defect. However, for wide-bandgap semiconductor materials, with band gaps greater than 3 eV, such as silicon carbide, gallium nitride, gallium oxide, and diamond, especially under high-resistivity conditions or when the defect energy level is deep, the voltage pulse cannot generate a sufficient number of charge carriers, resulting in extremely small capacitance changes, thus hindering effective detection. In addition, the resolution of traditional DLTS instruments is limited by the sensitivity of the detection circuit and the signal processing algorithm, making it difficult to accurately distinguish between adjacent deep-level defects. Summary of the Invention

[0004] Purpose of the invention: To address the limitations of existing deep-level transient spectroscopy (DLTS) techniques in measuring wide-bandgap semiconductor materials in high-resistivity states or when defect levels are deep, and the fact that the resolution of traditional DLTS instruments is limited by the sensitivity of the detection circuit and the difficulty in accurately distinguishing adjacent deep-level defects, this invention provides a method for detecting wide-bandgap semiconductor defect states based on time-of-flight technology.

[0005] Technical solution: The present invention provides a method for detecting wide-bandgap semiconductor defect states based on time-of-flight technology, comprising the following steps:

[0006] S1. Pretreatment of wide bandgap semiconductor samples: Clean, dry, polish or anneal the wide bandgap semiconductor samples to ensure good electrical contact and mechanical stability between the metal electrodes and the samples;

[0007] S2. Preparation of test sample: Patterning is performed on the surface of a wide bandgap semiconductor sample and two Schottky contact electrodes are prepared to form a planar capacitor with a spatial electric field. Metal electrodes are respectively provided above the two Schottky contact electrodes to form a double-layer electrode structure. The structure of the test sample is a coplanar metal-semiconductor-metal device structure.

[0008] S3. Measure the transient photocurrent curve of the test sample: Excite the test sample with a short pulse laser, and then set different bias voltages to obtain the transient photocurrent curve of the test sample under different bias voltages;

[0009] S4. Set Monte Carlo model parameters: In the Monte Carlo method with the addition of the drift-diffusion model and the SRH model, set the parameters suitable for the sample being tested;

[0010] S5. Using the Monte Carlo model from step S4, fit the transient photocurrent curves under different bias voltages from step S3 to obtain the defect parameter information in the sample being tested.

[0011] In step S1, the wide-bandgap semiconductor sample is an epitaxial thin film or substrate made of silicon carbide, gallium nitride, gallium oxide, diamond, or aluminum nitride, with a thickness greater than the light absorption depth of the sample and a resistivity greater than 1E5. • cm; Preferably, the wide bandgap semiconductor sample is a semi-insulating 4H-SiC substrate with a thickness of 500 μm and an area of ​​100 mm × 100 mm.

[0012] In step S1, the pretreatment method is one or a combination of ultrasonic cleaning, chemical cleaning, physical cleaning, chemical-mechanical polishing (CMP), and rapid thermal annealing (RTA).

[0013] In step S2, the structure of the test sample is such that the upper surface of the wide bandgap semiconductor sample is provided with a first Schottky contact electrode and a second Schottky contact electrode that are aligned with each other; the first Schottky contact electrode is provided with a first metal electrode, and the second Schottky contact electrode is provided with a second metal electrode.

[0014] In step S2, the Schottky contact electrode is shaped like a rectangular slit or a mesh. In the silicon carbide detection sample, the electrode material is Ti, and in the gallium oxide detection sample, the electrode material is Ni. The thickness is 10nm-100nm, and the effective contact area with the material is approximately 0.1-3mm. 2 Further preferably, the rectangular slit electrode has an area of ​​1mm*2mm, a slit width of 100nm, a slit spacing of 100nm, and an effective contact area with the material of approximately 1mm². 2 .

[0015] In step S2, the metal electrode material is one or a combination of two of Au, Ni, Cr, and Pt, and the Schottky contact electrode and the metal electrode form a Ti / Au or Ni / Au double-layer electrode structure.

[0016] In step S2, the patterning method is hard mask, direct-write lithography, or contact lithography; the Schottky contact electrode fabrication method includes magnetron sputtering, molecular beam epitaxy (MBE), and electron beam evaporation.

[0017] In step S3, the energy of the short-pulse laser wavelength used to measure the transient photocurrent curve of the sample is greater than the bandgap of the wide bandgap semiconductor, and the pulse width is less than one-tenth of the flight time; the bias voltage during measurement is 10V~200V.

[0018] Specifically, step S3 involves fixing the sample on the sample stage, sealing the probe stage after the two probe tips contact the two Schottky contact electrodes respectively, and keeping it in a dark environment; then connecting one end of the two probe signals to the laser signal generator and the other end to an external circuit and then to an oscilloscope, applying a DC bias voltage to both ends of the sample; irradiating the sample surface with a laser next to one electrode, and forming photogenerated carriers through light absorption between the valence band and conduction band of the sample material, which drift to the other electrode under the influence of the space charge region, obtaining a transient photocurrent in the external circuit; changing the bias voltage and measuring the transient photocurrent characteristic curves under different bias voltages.

[0019] In step S4, the SRH model satisfies the following equation:

[0020]

[0021] In equation (1), For carrier capture time, For carrier thermal velocity, For capturing the cross section, The defect concentration is given in equation (2). For carrier release time, For the density of conduction band states, For defect state energy levels, Boltzmann's constant, For temperature.

[0022] In step S5, the parameters of the sample to be detected are set in the Monte Carlo model, and the model is used to fit different transient photocurrent curves. The defect parameters include the capture-release time of each defect state to the carriers and the defect energy level, defect density, and capture cross section of each defect state.

[0023] Invention Principle: This invention is a method for detecting defect states in wide-bandgap semiconductors based on time-of-flight technology. Time-of-flight technology utilizes short-pulse lasers to excite semiconductor materials to generate photogenerated carriers. The carrier mobility is calculated by measuring the time required for the photogenerated carriers to drift from one electrode to another under the influence of an electric field. The short-pulse laser must have a wavelength energy greater than the bandgap of the wide-bandgap semiconductor, a pulse width less than one-tenth of the time of flight, and the contact electrode must be a Schottky contact to reduce errors caused by electrically injected carriers. This invention reveals that when photogenerated carriers drift in semiconductor materials or devices, the trapping and release of photogenerated carriers by defect states in the sample can have varying effects on the photocurrent curve characteristics due to different energy levels of defects. By designing a coplanar metal-semiconductor-metal model to detect the sample and forming a stable spatial electric field, and then using short-pulse laser excitation to measure the transient photocurrent curve, the defect parameters of the sample can be obtained by fitting the transient photocurrent curve with a drift-diffusion model and a (Shockley-Read-Hall, SRH)SRH model. This photoexcitation of carriers is not affected by high-resistivity states or deep-level defects, and the method is simple to test, accurately obtaining key parameters such as defect state energy level positions, concentrations, and trapping cross-sections at room temperature. This method will play an important role in studying defect information in wide-bandgap semiconductor materials and its impact on device reliability.

[0024] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: The wide bandgap semiconductor defect state detection method based on time-of-flight technology of the present invention is not affected by high-resistivity state or deep-level defects, the operating conditions are simple, the sample construction is simple and easy to realize, and the key parameters such as defect state energy level position, concentration, and trapping cross section can be accurately obtained at room temperature. It will play an important role in studying defect information in wide bandgap semiconductor materials and its impact on device reliability. Attached Figure Description

[0025] Figure 1 This is a flowchart of a method for detecting defect states in wide-bandgap semiconductors according to an embodiment of the present invention;

[0026] Figure 2 This is a schematic diagram of the sample structure used in the defect state detection method for semi-insulating 4H-SiC in an embodiment of the present invention; wherein, 1-wide bandgap semiconductor sample; 2-first Schottky contact electrode; 3-first metal electrode; 4-second Schottky contact electrode; 5-second metal electrode;

[0027] Figure 3 This is a schematic diagram of the structure of a grid-like Schottky contact electrode and a sample for testing, wherein 1-wide bandgap semiconductor sample; 2-first Schottky contact electrode; 3-first metal electrode; 4-second Schottky contact electrode; 5-second metal electrode;

[0028] Figure 4 The figures show the transient photocurrent curve under an 80V bias voltage and the Monte Carlo fitting curves for different numbers of traps obtained during the detection process in this embodiment of the invention. The inset shows the goodness of fit for different numbers of traps. Detailed Implementation

[0029] The technical solution of the present invention will be further described below with reference to the embodiments. The test materials used in the embodiments can all be purchased through conventional means.

[0030] Example 1

[0031] The present invention provides a wide-bandgap semiconductor defect detection method based on time-of-flight technology, comprising the following steps:

[0032] S1. Pretreatment of wide bandgap semiconductor samples: The semi-insulating 4H-SiC sample is cleaned and dried with N2. The cleaning method is standard RCA cleaning to ensure good electrical contact and mechanical stability between the metal electrode and the sample.

[0033] S2. Preparation of test samples: The silicon surface of silicon carbide was selected as the surface for preparing Schottky contact electrodes. Then, AZ 5214 photoresist film was prepared on the silicon surface (upper surface) of the treated silicon carbide sample by spin coating. The spin coating parameters were set to 400r / 2s and 4000r / 30s to form a photoresist thickness of 1.3-1.4um. Then, it was pre-baked on a hot plate with parameters set to 95℃ / 1.5min. It was then exposed by contact lithography or direct writing lithography. The exposure pattern was the same as the preset pair of slit electrodes. After that, the exposed silicon carbide substrate was immersed in 3038 developer for 45s for development. Then, it was rinsed with deionized water and dried with N2.

[0034] A 20 nm / 100 nm thick Ti / Au bilayer electrode structure film was prepared on the surface of silicon carbide using metal deposition methods such as thermal evaporation, electron beam evaporation, or magnetron sputtering. The silicon carbide sample was then immersed in acetone to remove the unexposed photoresist and the overlying metal, followed by rinsing with deionized water and drying with N2. This completed the fabrication of a pair of Schottky metal electrode structures on the surface. A detection sample structure for measuring defect states in semi-insulating 4H-SiC was thus prepared. The detection sample structure is a coplanar metal-semiconductor-metal device structure. The specific structure is as follows: Figure 2 or Figure 3 As shown, the upper surface of the wide bandgap semiconductor sample 1 is provided with a first Schottky contact electrode 2 and a second Schottky contact electrode 4 that are aligned with each other; a first metal electrode 3 is provided on the first Schottky contact electrode 2, and a second metal electrode 5 is provided on the second Schottky contact electrode 4, wherein... Figure 2 The Schottky contact electrode is rectangular slit-shaped. Figure 3 The Schottky contact electrode has a grid-like shape;

[0035] Among them, wide-bandgap semiconductor sample 1 has a thickness of 500 μm, an area of ​​10 mm × 10 mm, and a resistivity of 1 E13. ·cm-1E15 • cm; The first Schottky contact electrode 2 and the second Schottky electrode 4 are made of Ti with a thickness of 20 nm; the first metal electrode 3 and the second metal electrode 5 are made of Au with a thickness of 100 nm; the area is 1 mm * 2 mm; the slit width is 100 nm; the slit spacing is 100 nm; and the effective contact area with the material is approximately 1 mm². 2 The purpose is to reduce the capacitance C between the two plates, decrease the time constant RC, and improve the response speed;

[0036] S3. Measurement of transient photocurrent curves of the test sample: Fix the test sample on the sample stage, and seal the probe stage after the two probe tips contact the two Schottky contact electrodes respectively, keeping it in a dark environment to ensure the accuracy of the results; then connect one end of the two probe signals to the laser signal generator, and the other end to the external circuit and then to the oscilloscope, and apply a DC bias voltage to both ends of the sample; irradiate the sample surface next to one electrode, and through the light absorption between the valence band and conduction band of the sample material, photogenerated carriers are formed, and drift to the other electrode under the action of the space charge region, and the transient photocurrent is obtained in the external circuit; change the bias voltage from 10V to 200V and measure the transient photocurrent characteristic curves under different bias voltages;

[0037] S4. Setting Monte Carlo Model Parameters: In the Monte Carlo method incorporating the drift-diffusion model and the SRH model, set parameters suitable for the sample being detected, including conduction band density of states, electrode spacing, bias voltage, number of traps, and time step; the SRH model satisfies the following equation:

[0038]

[0039] In equation (1), For carrier capture time, For carrier thermal velocity, For capturing the cross section, The defect concentration is given in equation (2). For carrier release time, For the density of conduction band states, For defect state energy levels, Boltzmann's constant, For temperature;

[0040] S5. Using the Monte Carlo model from step S4, fit the transient photocurrent curves under different bias voltages from step S3 to obtain the defect parameter information in the sample being tested.

[0041] like Figure 4 The figure shows the transient photocurrent curve under an 80V bias voltage obtained during the detection process in this embodiment of the invention. Different numbers of traps were set in the Monte Carlo model, and Monte Carlo fitting was performed while keeping other parameters constant. It can be seen that the fitting effect is best when four traps are present. The inset shows that the goodness of fit gradually approaches an extreme value. Using more traps for fitting consumes more computational resources, but does not substantially improve the results. The defect state energy level position, concentration, and trapping cross-section can be obtained by using the trap-release time of the four traps fitted in the 4H-SiC detection sample and then solving the equation of the SRH model described in S4.

[0042] In summary, the wide-bandgap semiconductor defect state detection method proposed in this invention has simple testing conditions and can accurately obtain key parameters such as defect state energy level location, concentration, and trapping cross section at room temperature. It will play an important role in studying defect information in wide-bandgap semiconductor materials and its impact on device reliability.

Claims

1. A method for detecting defect states in wide-bandgap semiconductors based on time-of-flight technology, characterized in that, The detection method includes the following steps: S1. Pretreatment of wide bandgap semiconductor samples: Cleaning, drying, polishing or annealing wide bandgap semiconductor samples; S2. Preparation of test sample: Patterning is performed on the surface of a wide bandgap semiconductor sample and two Schottky contact electrodes are prepared to form a planar capacitor with a spatial electric field. Metal electrodes are respectively provided above the two Schottky contact electrodes to form a double-layer electrode structure. The structure of the test sample is a coplanar metal-semiconductor-metal device structure. S3. Measure the transient photocurrent curve of the test sample: Excite the test sample with a short pulse laser, and then set different bias voltages to obtain the transient photocurrent curve of the test sample under different bias voltages; S4. Set Monte Carlo model parameters: In the Monte Carlo method with the addition of the drift-diffusion model and the SRH model, set the parameters suitable for the sample being tested; S5. Using the Monte Carlo model from step S4, fit the transient photocurrent curves under different bias voltages from step S3 to obtain the defect parameter information in the sample being tested.

2. The detection method according to claim 1, characterized in that, In step S1, the wide-bandgap semiconductor sample is an epitaxial thin film or substrate, made of silicon carbide, gallium nitride, gallium oxide, diamond, or aluminum nitride, with a thickness greater than the light absorption depth of the sample and a resistivity greater than 1E5. ·cm.

3. The detection method according to claim 1, characterized in that, In step S1, the pretreatment method is one or a combination of ultrasonic cleaning, chemical cleaning, physical cleaning, chemical mechanical polishing, and rapid thermal annealing.

4. The detection method according to claim 1, characterized in that, In step S2, the structure of the test sample is as follows: the upper surface of the wide bandgap semiconductor sample (1) is provided with a first Schottky contact electrode (2) and a second Schottky contact electrode (4) that are aligned with each other; the first Schottky contact electrode (2) is provided with a first metal electrode (3), and the second Schottky contact electrode (4) is provided with a second metal electrode (5).

5. The detection method according to claim 1, characterized in that, The shape of the Schottky contact electrode in step S2 is rectangular slit or grid, the electrode material is Ti in the silicon carbide detection sample, the electrode material is Ni in the gallium oxide detection sample, the thickness is 10nm-100nm, and the effective contact area with the material is about 0.1-3mm 2 The metal electrode material is one of Au, Ni, Cr, Pt or a combination of two.

6. The detection method according to claim 1, characterized in that, In step S2, the patterning method is hard mask, direct-write photolithography, or contact photolithography; the Schottky contact electrode fabrication method includes magnetron sputtering, molecular beam epitaxy, and electron beam evaporation.

7. The detection method according to claim 1, characterized in that, In step S3, the short-pulse laser wavelength energy of the transient photocurrent curve of the test sample is greater than the bandgap of the wide bandgap semiconductor, and the pulse width is less than one-tenth of the flight time; the bias voltage during measurement is 10V~200V.

8. The detection method according to claim 1, characterized in that, Step S3 specifically involves fixing the sample on the sample stage, sealing the probe stage after the two probe tips contact the two Schottky contact electrodes respectively, and keeping it in a dark environment; then connecting one end of the two probe signals to the laser signal generator and the other end to an external circuit and then to an oscilloscope, applying a DC bias voltage to both ends of the sample; irradiating the sample surface with the laser next to one electrode, and forming photogenerated carriers through light absorption between the valence band and conduction band of the sample material, which then drift to the other electrode under the influence of the space charge region, and obtaining a transient photocurrent in the external circuit; By changing the bias voltage, the transient photocurrent characteristic curves under different bias voltages were measured.

9. The detection method according to claim 1, characterized in that, In step S4, the SRH model satisfies the following equation: ； In equation (1), For carrier capture time, For carrier thermal velocity, For capturing the cross section, The defect concentration is given in equation (2). For carrier release time, For the density of conduction band states, For defect state energy levels, Boltzmann's constant, For temperature.

10. The detection method according to claim 1, characterized in that, In step S5, the parameters of the sample to be detected are set in the Monte Carlo model, and the model is used to fit different transient photocurrent curves; the defect parameters include the capture-release time of carriers for each defect state obtained by fitting, and the defect energy level, defect density, and capture cross section of each defect state.

Images