Semiconductor resistivity nondestructive testing method and system based on modulation light reflection

Abstract

The invention discloses a semiconductor resistivity nondestructive testing method and system based on modulation light reflection, relates to the technical field of semiconductor resistivity detection, and aims to accurately measure a modulation component delta R / R of light reflectivity through a light reflection modulation signal so as to calculate the semiconductor resistivity and further realize quantitative measurement of the semiconductor resistivity. The sensor has ultrahigh sensitivity, can detect tiny carrier concentration change, and is very sensitive to electrical properties of surface and near-surface areas; meanwhile, the spatial resolution is high, and the micron-scale light spot at the sample is suitable for micro-area measurement; a complete optical method does not cause any damage to the sample; various types of samples can be measured; the measuring speed is fast, and the system can be easily integrated in a production line for real-time monitoring.

Description

Technical Field

[0001] This invention relates to the field of semiconductor resistivity detection technology, specifically a non-destructive testing method and system for semiconductor resistivity based on modulated light reflection. Background Technology

[0002] Semiconductor resistivity is a core parameter for measuring the electrical conductivity of materials. Its value lies between that of metals and insulators (typically 1 mΩ•cm to 1 GΩ•cm at room temperature), directly affecting the electrical characteristics of devices. The significance of resistivity testing includes: 1. Quality control: assessing material purity, crystal structure integrity, and internal defects to ensure semiconductor materials meet device design requirements; 2. Process optimization: monitoring doping uniformity and adjusting manufacturing processes to improve product yield.

[0003] Existing methods for measuring semiconductor resistivity include the four-probe method, eddy current method, traditional optical method, and terahertz time-domain spectroscopy. The four-probe method suffers from the drawback of being a contact-based test that can easily scratch the wafer, making it unsuitable for ultra-thin materials or finished devices. The eddy current method is limited by its low sensitivity to thin samples and its limited spatial resolution. Traditional optical methods (such as infrared reflection) are insensitive to semiconductors with moderate resistivity, making it difficult to establish strong correlations directly. As for terahertz time-domain spectroscopy, its systems are expensive and complex, and the measurement speed is relatively slow. Summary of the Invention

[0004] To address the technical deficiencies in existing semiconductor resistivity testing methods, this invention proposes a non-destructive testing method and system for semiconductor resistivity based on modulated light reflection. By accurately measuring the modulation component ΔR / R of the probe light reflectivity through the light reflection modulation signal, the semiconductor resistivity can be calculated, thereby achieving quantitative measurement of semiconductor resistivity.

[0005] This invention protects a non-destructive testing method for semiconductor resistivity based on modulated light reflection.

[0006] A periodically modulated pump light is used to irradiate a semiconductor sample, causing periodic electron-hole pairs to be generated on the surface of the semiconductor sample, i.e., periodic carrier concentration changes Δn. The pump light energy is greater than the semiconductor bandgap energy.

[0007] Meanwhile, a continuous beam of probe light is used to illuminate the same area of ​​the semiconductor sample, and a laser spot with a size of tens of micrometers is formed by focusing on the semiconductor sample. The energy of the probe light is close to but less than the energy of the semiconductor band gap. The reflectivity R of the probe light changes slightly due to the change in carrier concentration Δn and the resulting temperature change ΔT.

[0008] Using a high-frequency lock-in amplifier and with the modulation frequency f of the pump light as a reference, the modulation component ΔR / R of the probe light reflectivity is accurately measured, which is called the PR signal. The PR signal contains information about carrier concentration and temperature changes; the resistivity ρ of the semiconductor sample is obtained based on the amplitude of the PR signal.

[0009] As an embodiment of the present invention, by analyzing the relationship between the amplitude and phase of the PR signal and the modulation frequency f, the initial carrier concentration n0 of the semiconductor sample is deduced, and the resistivity ρ is calculated based on ρ = 1 / (n0*e*μ), where e is the elementary charge and μ is the mobility.

[0010] As an embodiment of the present invention, an empirical calibration curve of PR signal amplitude and standard sample with known resistivity is established, and the resistivity of semiconductor sample to be tested is measured by fitting the curve.

[0011] Preferably, the semiconductor samples to be tested are classified as follows: semiconductors that are bulk materials or have no significant dielectric layer on the surface are classified as Category 1; semiconductors with a transparent dielectric layer on the surface are classified as Category 2; and transparent or semiconductive oxides and ferroelectrics are classified as Category 3.

[0012] Based on the classification results, select the corresponding pre-established correlation model or calibration coefficient set for the semiconductor sample to be tested:

[0013] For the first type of semiconductor sample, a coupling model based on carrier diffusion-recombination and thermal conduction and coefficient A1 are adopted, where coefficient A1 includes carrier lifetime and thermal diffusivity.

[0014] For the second type of semiconductor sample, a multilayer optical transfer matrix and thermal conduction model and coefficient A2 are adopted, where coefficient A2 includes the dielectric layer thickness, refractive index and interface thermal resistance;

[0015] For the third type of semiconductor sample, a model and coefficient A3 based on piezoelectric, pyroelectric and electro-optic effects are adopted. The model is constructed based on piezoelectric, pyroelectric and electro-optic coefficients, where coefficient A3 includes piezoelectric coefficient, pyroelectric coefficient, electro-optic coefficient and thermal expansion coefficient.

[0016] Based on the classification of the semiconductor sample, the PR signal amplitude is substituted into the corresponding model and coefficients to calculate the accurate resistivity value.

[0017] This invention also protects a semiconductor resistivity non-destructive testing system based on modulated light reflection, used to implement the aforementioned semiconductor resistivity non-destructive testing method based on modulated light reflection. The system includes a pump source optical path, a probe scanning optical path, a dichroic mirror combination to achieve co-pathing of the pump and probe light, a beam focusing device to focus the optical path onto the surface of the semiconductor sample under test, a photodetector to detect the probe light reflection signal, and a high-frequency lock-in amplifier. The pump source optical path sequentially includes a high-power pump laser, an optical switch structure, an optical modulation device, and a beam shaping device. The probe scanning optical path sequentially includes a low-noise probe laser and a beam shaping device. The reference signal of the high-frequency lock-in amplifier comes from the drive signal of the optical modulation device.

[0018] This invention features ultra-high sensitivity, enabling the detection of extremely minute changes in carrier concentration and is highly sensitive to the electrical properties of surface and near-surface regions; it also boasts high spatial resolution, with a spot size at the micrometer level suitable for micro-area measurement; it employs a completely optical method, causing no damage to the sample; it can measure various types of samples; and it offers fast measurement speed, making it easy to integrate into production lines for real-time monitoring. Attached Figure Description

[0019] Figure 1 This is a block diagram of the semiconductor resistivity non-destructive testing system based on modulated light reflection disclosed in this invention.

[0020] Figure label:

[0021] 1-Pump laser source; 2-Optical switch structure; 3-Optical modulation device; 4-Beam shaping device; 5-High-reflection mirror; 6-High-reflection mirror; 7-Dichroic mirror; 8-Dichroic mirror; 9-Beam focusing device; 10-Narrow-band filter; 11-Photodetector; 12-Detection laser source; 13-Beam shaping device; 14-High-reflection mirror; 15-Semiconductor sample to be tested. Detailed Implementation

[0022] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. The embodiments of the present invention are given for illustrative and descriptive purposes only, and are not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described to better illustrate the principles and practical application of the invention, and to enable those skilled in the art to understand the invention and design various embodiments with various modifications suitable for a particular purpose.

[0023] Example 1

[0024] A non-destructive testing method for semiconductor resistivity based on modulated light reflection is proposed. A periodically modulated pump light is used to irradiate a semiconductor sample, which generates periodic electron-hole pairs on the surface of the semiconductor sample, i.e., periodic carrier concentration changes Δn. The pump light energy is greater than the semiconductor bandgap energy.

[0025] Meanwhile, a continuous beam of probe light is used to illuminate the same area of ​​the semiconductor sample, and a laser spot with a size of tens of micrometers is formed by focusing on the semiconductor sample. The energy of the probe light is close to but less than the energy of the semiconductor band gap. The reflectivity R of the probe light changes slightly due to the change in carrier concentration Δn and the resulting temperature change ΔT.

[0026] Using a high-frequency lock-in amplifier and with the modulation frequency f of the pump light as a reference, the modulation component ΔR / R of the probe light reflectivity is accurately measured, which is called the PR signal. The PR signal contains information about carrier concentration and temperature changes; the resistivity ρ of the semiconductor sample is obtained based on the amplitude of the PR signal.

[0027] The specific methods include the following two:

[0028] I. Theoretical Formula Calculation Method

[0029] Modulated light reflection (PR) technology is extremely sensitive to changes in carrier concentration and temperature on the semiconductor surface. Resistivity ρ = 1 / (n*e*μ), meaning it is directly related to both carrier concentration n and mobility μ. Therefore, the PR signal can accurately detect changes in carrier concentration induced by pump light. By analyzing the relationship between the amplitude and phase of the PR signal and the modulation frequency f, the initial carrier concentration n0 of the semiconductor sample can be deduced. Combined with the mobility μ, the semiconductor resistivity ρ can be calculated, thus achieving quantitative measurement of semiconductor resistivity.

[0030] The advantages of this method are: it is directly based on carrier transport theory, with a clear physical model; it does not require prior measurement of standard samples, making it suitable for new materials or samples with special structures; parameters such as carrier lifetime τ and diffusion coefficient D can be obtained simultaneously through model fitting; and it is theoretically applicable to any semiconductor material, without being limited by standard samples.

[0031] The disadvantages of this method are: it requires fitting amplitude and phase frequency response curves, which makes data processing complex; the mobility μ needs to be measured independently through the Hall effect or by using literature values, and the accuracy of the mobility μ affects the accuracy of the resistivity ρ; and it requires scanning the modulation frequency, which takes a long time to measure.

[0032] II. Calibration Fitting Method

[0033] An empirical calibration curve is established between the PR signal amplitude and a standard sample with known resistivity. The resistivity of the semiconductor sample under test is then measured by fitting the curve.

[0034] The advantages of this method are: simple operation, fast measurement speed after one calibration, suitable for batch sample testing; high accuracy, multiple calibration points can effectively reduce systematic errors and improve measurement accuracy; no need to build a complex physical model, and simple data processing; the calibration curve already contains mobility information, so there is no need for separate measurement.

[0035] The disadvantages of this method are: a standard sample must be measured in advance, and the accuracy of the standard sample directly affects the measurement results; the calibration curve is only applicable to specific material types and experimental conditions; only resistivity can be obtained, and other parameters such as carrier lifetime and diffusion coefficient cannot be obtained; the system stability requirements are high, and the experimental conditions (laser power, temperature, etc.) must be consistent with those during calibration.

[0036] The two methods described above can be selected based on the specific application scenario. Furthermore, the applicant discovered that different types of semiconductor materials (especially those with vastly different surface structures or optical properties, such as bulk silicon, silicon with a dielectric layer, and transparent ferroelectric crystals) exhibit different correlation functions (coefficient relationships) between resistivity ρ and the PR signal. Therefore, by establishing an effective measurement range and employing a multi-model fusion method, accurate measurements of different types of semiconductor samples can be achieved.

[0037] Before accurately measuring resistivity, the semiconductor samples to be tested are classified as follows: semiconductors that are bulk materials or have no significant dielectric layer on the surface (such as bare silicon wafers, gallium arsenide, etc.) are classified as Category I; semiconductors with a transparent dielectric layer on the surface are classified as Category II (such as silicon wafers with silicon oxide or silicon nitride passivation layers); and special materials such as transparent or semiconductive oxides and ferroelectrics (such as lithium niobate, zinc oxide, etc.) are classified as Category III.

[0038] Based on the classification results, select the corresponding pre-established correlation model or calibration coefficient set for the semiconductor sample to be tested:

[0039] 1. For the first type of semiconductor sample, a coupling model based on carrier diffusion-recombination and thermal conduction and coefficient A1 are adopted, where coefficient A1 includes carrier lifetime, thermal diffusion coefficient and carrier diffusion coefficient.

[0040] ◆Physical mechanism:

[0041] ◇Carrier diffusion: Photogenerated carriers diffuse into the bulk under the influence of a concentration gradient;

[0042] ◇Carrier recombination: Electron-hole pairs disappear through radiative recombination, Auger recombination, defect recombination, etc.

[0043] ◇Heat conduction: The heat released by charge carrier recombination is propagated through lattice vibrations.

[0044] ◆Governing equations:

[0045] ◇Carrier continuity equation: ∂Δn / ∂t = D∇²Δn - Δn / τ + G(t), where D is the carrier diffusion coefficient, τ is the carrier lifetime, and G(t) is the carrier generation rate;

[0046] ◇Thermal conduction equation: ∂ΔT / ∂t = D_t∇²ΔT + Q(t), where D_t is the thermal diffusivity and Q(t) is the heat source term.

[0047] ◆Coupling relationship:

[0048] ◇The change in carrier concentration Δn causes a change in refractive index: Δn_carrier = (∂n / ∂N)ΔN;

[0049] ◇Temperature change ΔT causes refractive index change: Δn_thermal = (∂n / ∂T)ΔT;

[0050] ◇Total refractive index change: Δn_total = Δn_carrier + Δn_thermal.

[0051] ◆Key coefficient acquisition:

[0052] By measuring the amplitude-frequency response and phase-frequency response curves of the PR signal, the carrier lifetime τ, thermal diffusivity D_t, and carrier diffusivity D are obtained through fitting.

[0053] 2. For the second type of semiconductor sample, a multilayer optical transmission matrix and thermal conduction model and coefficient A2 are adopted, where coefficient A2 includes dielectric layer thickness, dielectric layer refractive index, interface thermal resistance, and stress coefficient.

[0054] ◆Physical mechanism:

[0055] ◇Optical interference effect: The probe light is reflected multiple times at the interface between the upper and lower layers of the dielectric layer, forming interference fringes;

[0056] ◇Heat transport barrier: The dielectric layer hinders the diffusion of heat to the surface, changing the heat conduction path.

[0057] ◆Governing equations:

[0058] ◇Optical transfer matrix: For a multilayer structure, the reflectivity R is determined by the thickness d_i, refractive index n_i, and extinction coefficient k_i of each layer: R = f(d_1, n_1, k_1; d_2, n_2, k_2; ...);

[0059] ◇Heat conduction equation: ∂ΔT / ∂t = D_t∇²ΔT + Q(t), Boundary condition: Thermal resistance exists at the interface of the medium layer, and the heat flow is continuous.

[0060] ◇Stress effect: Stress is generated by the mismatch between the dielectric layer and the semiconductor lattice, which changes the band structure ΔE_g = a·σ +b·σ², where σ is the stress and a and b are stress coefficients.

[0061] ◆Key coefficients: dielectric layer thickness d, dielectric layer refractive index n, interfacial thermal resistance R_th, stress coefficients a and b.

[0062] 3. For the third type of semiconductor sample, a model and coefficient A3 based on piezoelectric, pyroelectric and electro-optic effects are adopted. The model is constructed based on piezoelectric, pyroelectric and electro-optic coefficients, where coefficient A3 includes piezoelectric coefficient, pyroelectric coefficient, electro-optic coefficient and thermal expansion coefficient.

[0063] ◆Physical mechanism:

[0064] ◇Pyroelectric effect: Temperature change ΔT causes polarization change ΔP;

[0065] ◇Piezoelectric effect: Stress σ causes a change in polarization ΔP;

[0066] ◇Electro-optic effect: The electric field E changes the refractive index Δn.

[0067] ◆Governing equations:

[0068] ◇Pyroelectric effect: ΔP = p·ΔT, where p is the pyroelectric coefficient;

[0069] ◇Piezoelectric effect: ΔP = d·σ, where d is the piezoelectric coefficient;

[0070] ◇Electro-optic effect: Δ(1 / n²) = r·E, where r is the electro-optic coefficient;

[0071] ◇Total refractive index change: Δn_total = (∂n / ∂P)·ΔP + (∂n / ∂T)·ΔT.

[0072] ◆Key coefficients: piezoelectric coefficient d, pyroelectric coefficient p, electro-optic coefficient r, and thermal expansion coefficient α.

[0073] ◆The essential difference from the semiconductor model:

[0074] ◇Different response mechanisms: Ferroelectric crystals primarily respond to thermally induced stress / electric fields, rather than changes in carrier concentration.

[0075] ◇Different time scales: Electro-optic effect response is extremely fast (ps), while carrier recombination is relatively slow (ns-μs).

[0076] ◇Different coefficients: Requires nonlinear optical parameters such as piezoelectric coefficient, pyroelectric coefficient, and electro-optic coefficient.

[0077] Finally, based on the classification of the semiconductor sample, the PR signal amplitude is substituted into the corresponding model and coefficients to calculate the accurate resistivity value.

[0078] Example 2

[0079] A non-destructive testing system for semiconductor resistivity based on modulated light reflection is provided to implement the non-destructive testing method for semiconductor resistivity based on modulated light reflection disclosed in Example 1. Figure 1 As shown, the semiconductor resistivity non-destructive testing system includes a pump light source optical path detection scanning optical path, a dichroic mirror combination to realize the co-path of pump light and probe light, a beam focusing device 9 to focus the optical path onto the surface of the semiconductor sample to be tested, a photodetector 11 to detect the reflected signal of the probe light, and a high-frequency lock-in amplifier.

[0080] The pump source optical path includes, in sequence, a high-power pump laser 1, an optical switch structure 2, an optical modulation device 3, and a beam shaping device 4; the detection and scanning optical path includes, in sequence, a low-noise detection laser 12 and a beam shaping device 13.

[0081] Specifically, in this embodiment, the pump laser source 1 uses a solid-state laser with a wavelength of 532nm. The optical modulation device is an acousto-optic modulator (AOM), and its first-order diffracted light is specifically selected as the modulated pump beam. This choice is made because the first-order diffracted light has high modulation efficiency and a pure waveform, avoiding background noise caused by zero-order light pass-through. The driving frequency of the AOM is set to a high frequency (usually above MHz). At high frequencies, low-frequency noise and thermal effects can be filtered out, thereby separating the carrier signal of the semiconductor.

[0082] The probe laser source 12 uses a distributed feedback laser with a wavelength of 635nm. Its photon energy is lower than the band gap of silicon, ensuring that it does not generate additional photogenerated carriers. The pump light path and the probe light path are made common through dichroic mirrors 7 and 8, and then focused onto the surface of the semiconductor sample under test by a beam focusing device (focusing objective lens) 9.

[0083] According to the knife-edge method, the diameter of the laser spot focused on the semiconductor sample under test is about tens of micrometers. Only when the laser spot reaches the micrometer level or even smaller can more carrier effects be excited on the semiconductor surface, which is convenient for the collection and amplification of light reflection signals.

[0084] The reflected signal of the probe light is received by a high-bandwidth InGaAs photodetector 11, and its output signal is fed into a high-frequency lock-in amplifier. The reference signal of the high-frequency lock-in amplifier comes from the drive signal of the AOM, and its frequency is set to a high frequency to accurately extract the reflectivity change ΔR / R at the modulation frequency. The high-frequency modulation can effectively suppress 1 / f noise (low-frequency flicker noise), thereby significantly improving the signal-to-noise ratio of the system.

[0085] The system records the amplitude (ΔR / R) and phase θ of the PR signal output by the high-frequency lock-in amplifier, where the phase θ has the following function:

[0086] 1. To help determine the material type or surface structure of the sample to be tested (such as distinguishing between bulk silicon and silicon with an oxide layer) in order to select the correct calibration coefficient;

[0087] 2. Evaluate the quality of the measurement signal. When the phase signal shows abnormal drift or the signal-to-noise ratio is too low, the system can issue a reliability warning.

[0088] 3. Under feasible conditions, the resistivity and carrier lifetime parameters of the sample can be obtained by combining the amplitude information with the inversion.

[0089] By using a moving translation stage to place the semiconductor sample under test, micro-area resistivity fixed-point detection or two-dimensional detection with micron-level spatial resolution can be achieved.

[0090] Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art and related fields based on the embodiments of the present invention without inventive effort should fall within the scope of protection of the present invention.

Claims

1. A method for non-destructive testing of semiconductor resistivity based on modulated light reflection, characterized in that, a periodic modulated pump light is used to irradiate the semiconductor sample, so that the surface of the semiconductor sample generates a periodic electron-hole pair, i.e. a periodic carrier concentration change Δn, and the pump light energy is greater than the semiconductor band gap energy; at the same time, a continuous probe light is used to irradiate the same region of the semiconductor sample, and a laser spot with a size of tens of microns is formed on the semiconductor sample, and the probe light energy is close to but less than the semiconductor band gap energy; the reflectivity R of the probe light changes slightly ΔR due to the carrier concentration change Δn and the temperature change ΔT generated thereby; a high-frequency lock-in amplifier is used to accurately measure the modulation component ΔR / R of the reflectivity of the probe light with the modulation frequency f of the pump light as the reference, and the PR signal is obtained, which contains the information of the carrier concentration and the temperature change; the semiconductor sample resistivity ρ is obtained based on the PR signal amplitude.

2. The method of non-destructive testing of resistivity of a semiconductor based on modulation of light reflection according to claim 1, characterized in that, By analyzing the relationship between the PR signal amplitude and phase and the modulation frequency f, the initial carrier concentration n0 of the semiconductor sample is obtained, and the resistivity ρ is calculated based on ρ = 1 / (n0*e*μ), wherein e is the elementary charge and μ is the mobility.

3. The method of claim 1, wherein the modulated light reflection is based on a semiconductor resistivity. An empirical calibration curve of the PR signal amplitude and the standard sample with known resistivity is established, and the resistivity of the semiconductor sample to be measured is measured by fitting the curve.

4. The method of non-destructive testing of resistivity of a semiconductor based on modulation of light reflection according to claim 3, characterized in that, The semiconductor sample to be measured is classified as follows: the bulk material or the semiconductor without a significant medium layer on the surface is classified as the first type, the semiconductor covered with a transparent medium layer on the surface is classified as the second type, and the transparent or semi-conductive oxide and ferroelectric are classified as the third type. According to the classification results, the corresponding pre-established correlation model or calibration coefficient set is selected for the semiconductor sample to be measured: For the first type of semiconductor sample, a coupling model based on carrier diffusion-recombination and heat conduction and a coefficient A1 are used, wherein the coefficient A1 includes carrier lifetime, thermal diffusion coefficient, and carrier diffusion coefficient; For the second type of semiconductor sample, a multi-layer optical transfer matrix and heat conduction model and a coefficient A2 are used, wherein the coefficient A2 includes medium layer thickness, medium layer refractive index, interface thermal resistance, and stress coefficient; For the third type of semiconductor sample, a model based on piezoelectric, pyroelectric, and electro-optic effects and a coefficient A3 are used, and the model is based on piezoelectric, pyroelectric, and electro-optic coefficients, wherein the coefficient A3 includes piezoelectric coefficient, pyroelectric coefficient, electro-optic coefficient, and thermal expansion coefficient; According to the classification of the semiconductor sample, the PR signal amplitude is substituted into the corresponding model and coefficient to calculate the accurate resistivity value.

5. A non-destructive testing system for semiconductors based on modulated light reflection, characterized in that, A device for implementing the method for non-destructive testing of semiconductor resistivity based on modulated light reflection according to any one of claims 1-4, comprising a pump light source light path, a probe scanning light path, a dichroic mirror combination for realizing the co-path of the pump light and the probe light, a light beam focusing device (9) for focusing the light path on the surface of the semiconductor sample to be measured, a photodetector (11) for detecting the reflection signal of the probe light, and a high-frequency lock-in amplifier. The pump light path comprises high-power pump laser (1), optical switch structure (2), optical modulation device (3) and beam shaping device (4) in sequence; the probe scanning light path comprises low-noise probe laser (12) and beam shaping device (13) in sequence; the reference signal of the high-frequency lock-in amplifier is the driving signal from the optical modulation device (3).

Images