A method for extracting intrinsic parameters of gallium nitride high electron mobility transistor

By combining the extraction of parasitic parameters, principal component analysis and particle swarm algorithm, the time-consuming and limitations of the eigen parameter extraction of gallium nitride high electron mobility transistors is solved, and parameter extraction with higher accuracy and confidence is achieved, simplifying the The device modeling process is performed.

CN114065859BActive Publication Date: 2024-11-05HENAN UNIV OF SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202111361404.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-11-17
Publication Date
2024-11-05
Estimated Expiration
2041-11-17

AI Technical Summary

Technical Problem

The prior art has time-consuming and parameter extraction limitations when extracting the intrinsic parameters of gallium nitride high electron mobility transistors, and cannot quickly and accurately model the device, resulting in increased product costs and extended development cycles.

Method used

A method that includes extracting parasitic parameters, removing their influence, determining the range based on eigeny Y parameters, principal component analysis method and particle swarm algorithm is adopted to improve the extraction accuracy by optimizing the algorithm and parameter correlation.

Benefits of technology

It improves the reliability and accuracy of eigen parameter extraction, shortens the modeling design time, enhances the correlation between each eigen parameter, and enhances the practical value of the model.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114065859B_ABST
    Figure CN114065859B_ABST
Patent Text Reader

Abstract

A method for extracting intrinsic parameters of a gallium nitride high electron mobility transistor comprises the following steps: S1, extracting parasitic parameters, and removing the influence of parasitic parameters on intrinsic parameters by a de-embedding method; S2, determining the range of each intrinsic parameter based on the intrinsic Y parameter; S3, separating the real part and the imaginary part of the intrinsic Y parameter to obtain multiple groups of Y parameter data; S4, obtaining the first principal component based on the Y parameter data by using the principal component analysis method; S5, fitting each group of Y parameter data using the first principal component to obtain multiple fitting coefficients; S6, combining the fitting coefficients with the first principal component to obtain an optimized intrinsic Y parameter extraction method; S7, extracting the intrinsic parameters by using a particle swarm algorithm based on the determined intrinsic parameter range and the optimized intrinsic Y parameter extraction method. The present invention provides a method for extracting intrinsic parameters of a gallium nitride high electron mobility transistor, which improves the reliability of the extracted intrinsic parameters.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of gallium nitride high electron mobility transistors, and in particular to a method for extracting intrinsic parameters of a gallium nitride high electron mobility transistor. Background Art

[0002] Gallium nitride high electron mobility transistor (GaN HEMT) has the characteristics of high breakdown voltage and high electron mobility, which make GaN HEMT devices have unparalleled advantages over gallium arsenide (GaAs) devices in the field of microwave power. At present, GaN HEMT devices have been widely used in various fields.

[0003] In circuit design, modeling of specific transistor devices is essential. Only by accurately simulating the characteristics of the device can a circuit with excellent performance be designed. However, the current small signal parameter extraction technology for GaN HEMT devices is still imperfect, which will increase product costs and prolong the development cycle. Therefore, it is necessary to design a hybrid algorithm that can quickly extract intrinsic parameter values.

[0004] like Figure 2 As shown in Figure 1, GaN HEMT device parameters include external parasitic parameters and internal intrinsic parameters. The external parasitic parameters include the parasitic capacitance C between the drain and source. pds , the parasitic capacitance C between the gate and drain pgd , the parasitic capacitance C between the gate and the source pgs , gate parasitic resistance R g , gate parasitic inductance L g , Drain parasitic resistance R d , drain parasitic inductance L d , source parasitic resistance R s , source parasitic inductance L s The intrinsic parameters include the gate-source capacitance C gs , gate-drain capacitance C gd , drain-source capacitance C ds , intrinsic channel resistance R i , intrinsic gate-drain resistance R gd , output resistance R ds , substrate related parameters C db and R db , transconductance G mand the electron transmission delay τ. The parasitic parameters can be extracted by the methods [S. Van den Bosch and L. Martens, "Fast and accurate extraction of capacitance parameters for the Statz MESFET model," IEEE Transactions on Microwave Theory and Techniques, vol. 45, no. 8, pp. 1247-1249, Aug. 1997] and [G. Chen, V. Kumar, RSS Schwindt and I. Adesida, "A low gate bias model extraction technique for AlGaN / GaNH EMTs," IEEE Transactions on Microwave Theory and Techniques, vol. 54, no. 7, pp. 2949-2953, July 2006]. After removing the influence of parasitic parameters, the intrinsic parameters can be obtained using the method [A.Jarndal and G.Kompa, "A new small-signal modeling approachapplied to GaN devices," IEEE Transactions on Microwave Theory and Techniques, vol.53, no.11, pp.3440-3448, Nov.2005], but this method is time-consuming and not suitable for rapid modeling design. The method [A.Majumder, S.Chatterjee, S.Chatterjee, S.SinhaChaudhari and DRPoddar, "Optimization of Small-Signal Model of GaN HEMT by Using Evolutionary Algorithms," in IEEE Microwave and Wireless Components Letters, vol. 27, no. 4, pp. 362-364, April 2017] uses a search space exploration strategy to optimize the intrinsic parameters in GaN HEMT devices. Although this method can find the optimal solution for the specified area, it does not consider the rationality of the setting of the specified area. In addition, the intrinsic parameter extraction method used in this method has limitations and can only use Y 11 , Y 12, Y 21 , Y 22 Solving the corresponding intrinsic parameters separately cannot fully explain the correlation between each intrinsic element. Summary of the invention

[0005] In order to solve the deficiencies in the prior art, the present invention provides a method for extracting intrinsic parameters of a gallium nitride high electron mobility transistor, which improves the reliability of the extracted intrinsic parameters.

[0006] In order to achieve the above object, the specific scheme adopted by the present invention is: a method for extracting intrinsic parameters of a gallium nitride high electron mobility transistor, comprising the following steps:

[0007] S1. Extract parasitic parameters and remove the influence of parasitic parameters on intrinsic parameters by de-embedding method;

[0008] S2. determining a range of each intrinsic parameter based on the intrinsic Y parameter;

[0009] S3, separating the real part and the imaginary part of the intrinsic Y parameter to obtain multiple groups of Y parameter data;

[0010] S4, obtaining the first principal component by principal component analysis based on the Y parameter data;

[0011] S5, using the first principal component to fit each group of Y parameter data to obtain multiple fitting coefficients;

[0012] S6, combining the fitting coefficient with the first principal component to obtain an optimized intrinsic Y parameter extraction method;

[0013] S7. Extracting intrinsic parameters through a particle swarm algorithm based on the determined range of intrinsic parameters and the optimized intrinsic Y parameter extraction method.

[0014] As a further optimization of the above-mentioned intrinsic parameter extraction method of GaN high electron mobility transistor: In S1, the specific method of extracting parasitic parameters is:

[0015]

[0016]

[0017]

[0018] L s =imag(Z L,12 ) / ω;

[0019] L d +L s =imag(Z L,22 ) / ω;

[0020] Lg +L s =imag(Z L,11 ) / ω;

[0021] R s +R g =real(Z L,11 );

[0022] R s =real(Z L,12 );

[0023] R s +R d =real(Z L,22 );

[0024] Among them, C pds is the parasitic capacitance between the drain and source, C pgd is the parasitic capacitance between the gate and drain, C pgs is the parasitic capacitance between the gate and the source, R g is the gate parasitic resistance, L g is the gate parasitic inductance, R d is the drain parasitic resistance, L d is the drain parasitic inductance, R s is the source parasitic resistance, L s is the source parasitic inductance, W is the gate width, ω is the angular frequency, Y C,ij To extract the intermediate parameters used by parasitic capacitance, Y C,ij To extract the intermediate parameters used for parasitic resistance and parasitic inductance, imag() means taking the imaginary part, and real() means taking the real part.

[0025] As a further optimization of the above-mentioned GaN high electron mobility transistor intrinsic parameter extraction method: In S2, the calculation method of the intrinsic Y parameter is:

[0026]

[0027]

[0028]

[0029]

[0030] Among them, C gs is the gate-source capacitance, C gd is the gate-drain capacitance, C ds is the drain-source capacitance, R i is the intrinsic channel resistance, R gd is the intrinsic gate-drain resistance, R ds is the output resistance, C dband R db is the substrate related parameter, G m is the transconductance and τ is the electron transmission delay.

[0031] As a further optimization of the above-mentioned GaN high electron mobility transistor intrinsic parameter extraction method: in S4, the method for generating the first principal component is:

[0032]

[0033] Among them, a ij and b ij are the fitting coefficients.

[0034] As a further optimization of the above-mentioned intrinsic parameter extraction method of GaN high electron mobility transistor: in S6, the optimized intrinsic Y parameter extraction method is:

[0035] Real(Y ij )=c ij F 第一主 i,j=1,2;

[0036] Imag(Y ij ) = d ij F 第一主 i,j=1,2;

[0037] Among them, c ij and d ij are the fitting coefficients.

[0038] As a further optimization of the above-mentioned intrinsic parameter extraction method of the GaN high electron mobility transistor: in S7, the specific method of extracting the intrinsic parameters is:

[0039]

[0040] Among them, f k is the frequency point, It is the parameter actually tested at the frequency point. is the parameter obtained by simulation at the frequency point, e ij and h ij is the weight coefficient.

[0041] Beneficial effects: The present invention first determines the parameter range of the intrinsic parameters, so that the extracted intrinsic parameter optimization results have higher credibility; then the interrelationships between the intrinsic Y parameters are taken into account through the principal component analysis method, breaking the Y 11 ,Y 12 ,Y 21 ,Y 22The isolation between parameters strengthens the correlation between each intrinsic parameter; the intrinsic parameters are extracted by the particle swarm algorithm model with optimized parameters, which has higher accuracy and more practical value. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is a flow chart of the present invention;

[0043] Figure 2 This is a small signal model diagram of a GaN high electron mobility transistor;

[0044] Figure 3 is a schematic diagram of the experimental results A;

[0045] Figure 4 is a schematic diagram of the experimental results B;

[0046] Figure 5 C is a schematic diagram of the experimental results. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] In this embodiment, the present invention is described in detail based on a large-size GaN high electron mobility transistor.

[0049] See also Figure 1 , a method for extracting intrinsic parameters of a gallium nitride high electron mobility transistor, including S1 to S7.

[0050] S1. Extract parasitic parameters and remove the influence of parasitic parameters on intrinsic parameters by de-embedding method. In S1, the specific method of extracting parasitic parameters is:

[0051]

[0052]

[0053]

[0054] L s =imag(Z L,12 ) / ω;

[0055] L d +L s =imag(Z L,22 ) / ω;

[0056] L g +L s =imag(Z L,11 ) / ω;

[0057] R s +R g =real(Z L,11 );

[0058] R s =real(Z L,12 );

[0059] R s +R d =real(Z L,22 );

[0060] Among them, C pds is the parasitic capacitance between the drain and source, C pgd is the parasitic capacitance between the gate and drain, C pgs is the parasitic capacitance between the gate and the source, R g is the gate parasitic resistance, L g is the gate parasitic inductance, R d is the drain parasitic resistance, L d is the drain parasitic inductance, R s is the source parasitic resistance, L s is the source parasitic inductance, W is the gate width, ω is the angular frequency, Y C,ij To extract the intermediate parameters used by parasitic capacitance, Y C,ij To extract the intermediate parameters of parasitic resistance and parasitic inductance, imag() represents the imaginary part and real() represents the real part. Specifically, the parasitic capacitance is extracted using the size ratio relationship method, which utilizes the measured open circuit (V GS =-3V,V DS =0V) to extract the parasitic capacitance (C pds ,C pgd ,C pgs After removing the influence of parasitic capacitance, the parasitic resistance and parasitic inductance are extracted using low gate voltage technology, and the bias point (V GS =0V,V DS =0V) to extract the test data.

[0061] S2. Determine the range of each intrinsic parameter based on the intrinsic Y parameter. In S2, the intrinsic Y parameter is calculated as follows:

[0062]

[0063]

[0064]

[0065]

[0066] Among them, C gs is the gate-source capacitance, C gd is the gate-drain capacitance, C ds is the drain-source capacitance, R i is the intrinsic channel resistance, R gd is the intrinsic gate-drain resistance, R ds is the output resistance, C db and R db is the substrate related parameter, G m is the transconductance and τ is the electron transmission delay.

[0067] S3. Separate the real part and the imaginary part of the intrinsic Y parameter to obtain multiple groups of Y parameter data.

[0068] S4. Obtain the first principal component using principal component analysis based on the Y parameter data. In S4, the method for generating the first principal component is:

[0069]

[0070] Among them, a ij and b ij are the fitting coefficients.

[0071] S5. Use the first principal component to fit each group of Y parameter data to obtain multiple fitting coefficients.

[0072] S6, combining the fitting coefficient with the first principal component to obtain an optimized intrinsic Y parameter extraction method. In S6, the optimized intrinsic Y parameter is calculated as follows:

[0073] Real(Y ij )=c ij F 第一主 i,j=1,2;

[0074] Imag(Y ij ) = d ij F 第一主 i,j=1,2;

[0075] Among them, c ij and d ij are the fitting coefficients.

[0076] S7, based on the determined range of intrinsic parameters and the optimized intrinsic Y parameter extraction method, the intrinsic parameters are extracted by particle swarm algorithm. In S7, the specific method for extracting the intrinsic parameters is:

[0077]

[0078] Among them, f k is the frequency point, It is the parameter actually tested at the frequency point. is the parameter obtained by simulation at the frequency point, e ij and h ij is the weight coefficient.

[0079] The present invention first determines the parameter range of the intrinsic parameters, so that the extracted intrinsic parameter optimization results have higher credibility; then the interrelationships between the intrinsic Y parameters are taken into account by the principal component analysis method, breaking the Y 11 ,Y 12 ,Y 21 ,Y 22 The isolation between parameters strengthens the correlation between each intrinsic parameter; the intrinsic parameters are extracted by the particle swarm algorithm model with optimized parameters, which has higher accuracy and more practical value.

[0080] In order to verify the present invention, three bias points are used for experiments, and the three bias points are (V GS =-1V,V DS =2V), (V GS =-3V,V DS =0V) and (V GS =-3V,V DS =28V), and experiments were conducted using a particle swarm algorithm without optimized parameters, a particle swarm algorithm with optimized parameters, and a total of three extraction methods of the present invention. The experiment mainly conducted S parameter testing. The experimental results are shown in Figure 3-5 As shown, it can be seen that the fitting accuracy of the extraction result of the present invention is very high.

[0081] Finally, it should be noted that the Y parameters, S parameters and Z parameters recorded in the present invention are common knowledge in the field, and reference may be made to the document "Field Effect Transistor RF Microwave Modeling Technology" by Gao Jianjun, which will not be repeated here.

[0082] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for extracting intrinsic parameters of a gallium nitride high electron mobility transistor, characterized in that: The steps include: S1. Extract parasitic parameters, and remove the influence of parasitic parameters on intrinsic parameters by de-embedding method. In S1, the specific method of extracting parasitic parameters is: L s =image(Z L,12 ) / ω; IT d +L s =image(Z L,22 ) / ω; IT g +L s =image(Z L,11 ) / ω; R s +R g =real(Z L,11 ); R s =real(Z L,12 ); R s +R d =real(Z L,22 ); Among them, C pds is the parasitic capacitance between the drain and source, C pgd is the parasitic capacitance between the gate and drain, C pgs is the parasitic capacitance between the gate and the source, R g is the gate parasitic resistance, L g is the gate parasitic inductance, R d is the drain parasitic resistance, L d is the drain parasitic inductance, R s is the source parasitic resistance, L s is the source parasitic inductance, W is the gate width, ω is the angular frequency, Y C,ij To extract the intermediate parameter used by parasitic capacitance, Z C,ij To extract the intermediate parameters of parasitic resistance and parasitic inductance, imag() means taking the imaginary part, and real() means taking the real part; S2, based on intrinsic Y int,ij The parameter determines the range of each intrinsic parameter; in S2, the intrinsic Y int,ij The parameters are calculated as follows: Among them, C gs is the gate-source capacitance, C gd is the gate-drain capacitance, C ds is the drain-source capacitance, R i is the intrinsic channel resistance, R gd is the intrinsic gate-drain resistance, R ds is the output resistance, C db and R db is the substrate related parameter, G m is the transconductance, τ is the electron transmission delay; S3, the intrinsic Y int,ij The real and imaginary parts of the parameters are separated to obtain multiple sets of Y parameter data; S4. Obtain the first principal component by principal component analysis based on the Y parameter data; in S4, the method for generating the first principal component is: Among them, a ij and b ij is the fitting coefficient, Y ij is the intrinsic Y parameter; S5, using the first principal component to fit each group of Y parameter data to obtain multiple fitting coefficients; S6. Combining the fitting coefficient with the first principal component to obtain an optimized intrinsic Y parameter extraction method; in S6, the optimized intrinsic Y parameter extraction method is: Real(Y ij )=c ij F 第一主 i,j=1,2; Imag(Y ij )=d ij F 第一主 i,j=1,2; Among them, c ij and d ij is the fitting coefficient; S7, extracting intrinsic parameters through particle swarm algorithm based on the determined range of intrinsic parameters and the optimized intrinsic Y parameter extraction method; in S7, the specific method of extracting intrinsic parameters is: Among them, f k is the frequency point, It is the parameter actually tested at the frequency point. is the parameter obtained by simulation at the frequency point, e ij and h ij is the weight coefficient.