FinFET BSIM-CMG model and modeling method
By evaluating and adjusting the process corner offset and parameters of the FinFET BSIM-CMG model, the smoothness problem of the model was solved, the efficiency and accuracy of model construction were improved, and the stability of circuit simulation results was ensured.
Patent Information
- Application Number
- CN202511087836.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-05
- Publication Date
- 2025-11-21
AI Technical Summary
The existing FinFET BSIM-CMG model has a smoothness problem in block modeling, which leads to poor convergence of circuit simulation. Existing methods are inefficient and require manual experience to insert dimensions to handle discontinuities.
By evaluating the process corner offset and directly adjusting the parameters, a smooth model is constructed. The corner offset of the process corner model parameters is used to mark abrupt change points and perform parameter replacement and fine-tuning to ensure that the simulation results are consistent with the measured data.
It enables rapid optimization and efficient model construction, improves smoothness, reduces data analysis and processing workload, and enhances model building efficiency.
Smart Images
Figure CN120995957A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the technical field of semiconductor design SPICE modeling, and particularly relates to a FinFET BSIM-CMG model and a modeling method, which are mainly suitable for device model optimization of 16 nanometers and below advanced process nodes. BACKGROUND
[0002] The BSIM (Berkeley Short-channel IGFET Model) model is a model proposed for short-channel MOSFET (Metal-Oxide-Semiconductor Field-Effect Transistor). The BSIM-CMG model is mainly applied to 16 nanometers and below advanced process nodes, is an industry standard model of FinFET (Fin Field Effect Transistor) devices widely used in the industry, and the core thereof is to describe short-channel effects by a unified field penetration length and to quantify physical phenomena such as threshold voltage modulation.
[0003] In order to improve the modeling accuracy, the FinFET BSIM-CMG model usually adopts a binning method. The binning method is to divide the involved size into a plurality of subintervals for separate modeling, each subinterval is composed of four points, and the smoothness of the subinterval model is an important requirement of the model. However, the binning method often causes the smoothness problem due to the kink of the subinterval point parameters.
[0004] The smoothness of the model directly affects the convergence of circuit simulation. In order to ensure the smoothness of the model, the existing method needs to rely on artificial experience to additionally insert a plurality of sizes in each subinterval, and check whether the sizes have kinks. This method is low in efficiency. How to efficiently construct a smooth FinFET BSIM-CMG model is a problem to be solved by those skilled in the art. SUMMARY
[0005] To solve the above problems, the application provides a FinFET BSIM-CMG model and a modeling method, which realize rapid optimization of the model by evaluating the process angle offset and directly adjusting the parameters, and the efficiency of constructing the smooth model is higher.
[0006] In a first aspect of the application, a FinFET BSIM-CMG model modeling method is provided, comprising:
[0007] S1: extracting model parameters of an existing FinFET BSIM-CMG model;
[0008] S2: building a process angle model based on the model parameters;
[0009] S3: Calculate the corner offset of each point of the process corner model parameter;
[0010] S4: Build the offset polyline graph with each point as the X axis and the corner offset of each point as the Y axis;
[0011] S5: Evaluate the corner offset, if the offset polyline contains a mutation point, mark the mutation point as a to-be-processed point;
[0012] S6: Replace the parameter value of the to-be-processed point with the optimized parameter value, and fine-tune the parameter value of the to-be-processed point based on the measured data, so that the simulation result of the to-be-processed point is consistent with the measured data;
[0013] S7: Traverse all process corner model parameters and the corner offset of each point of the process corner model parameter, update the parameter value of all to-be-processed points, obtain the optimized process corner model, and obtain the final FinFET BSIM-CMG model.
[0014] Further, the model parameters include channel length L, fin number nfin, and point parameters matched with the model parameters.
[0015] Further, the point parameters include: lateral non-uniform doping parameter k0, strong inversion correction factor k0si, low field mobility u0, surface roughness scattering parameter ua, coulomb scattering parameter ud, coulomb scattering parameter ucs, surface roughness scattering parameter eu, and / or effective field parameter etamob.
[0016] Further, the process corner model includes a TT process corner model, a FF process corner model, and / or a SS process corner model.
[0017] Further, the corner offset of the point includes: threshold voltage offset dvth, linear region current offset didlin, and / or saturation region current offset didsat.
[0018] Further, the calculation method of the corner offset of the point is:
[0019] dvth = vth_FF - vth_TT, dvth is the threshold voltage offset of the threshold voltage vth_FF of the FF process corner model and the threshold voltage vth_TT of the TT process corner model;
[0020] didlin = idlin_FF / idlin_TT-1, where didlin is the linear region current offset minus 1 after dividing the linear region current idlin_FF of the FF process corner model by the linear region current idlin_TT of the TT process corner model.
[0021] didsat = idsat_FF / idsat_TT-1, where didsat is the saturation current offset by subtracting 1 after dividing the saturation current idsat_FF of the FF process corner model by the saturation current idsat_TT of the TT process corner model.
[0022] Furthermore, the optimized parameter value is the parameter value of the point adjacent to the left of the point to be processed.
[0023] Furthermore, the optimized parameter value is the parameter value of the point adjacent to the right of the point to be processed.
[0024] Furthermore, the optimized parameter value is the average of the parameter values of the two adjacent points to the left and right of the point to be processed.
[0025] In a second aspect, the present invention provides a FinFET BSIM-CMG model, wherein the FinFET BSIM-CMG modeling method described in the first aspect of the present invention is used when constructing the FinFET BSIM-CMG model.
[0026] The advantages of this invention compared to the prior art are:
[0027] This invention achieves rapid model optimization by evaluating process angle offset and directly adjusting parameters. The process is more accurate and efficient, and it does not require additional evaluation or insertion of several dimensions, thereby reducing the workload of data analysis and processing and making the construction of smooth models more efficient. Attached Figure Description
[0028] Figure 1 This is a schematic diagram of the steps of a FinFETBSIM-CMG modeling method provided in an embodiment of the present invention.
[0029] Figure 2 This is a line graph showing the offset before parameter adjustment provided in an embodiment of the present invention.
[0030] Figure 3 This is a line graph of the offset after parameter adjustment provided in an embodiment of the present invention. Detailed Implementation
[0031] With reference to the drawings and the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly described. Obviously, the described embodiments are only some of the embodiments of the present application, but not all of the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art are within the scope of the present application.
[0032] The terms "first", "second", and the like in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than that illustrated or described herein, and the objects distinguished by "first", "second", etc. are generally a category, and are not limited to the number of objects, for example, the first object can be one or more. In addition, "and / or" in the specification and claims indicates at least one of the connected objects, and the character " / ", generally indicates that the front and rear associated objects are in an "or" relationship.
[0033] Method embodiments
[0034] The present application mainly solves the problem of how to efficiently construct a smooth FinFET BSIM-CMG model.
[0035] In a first aspect of the present application, a FinFET BSIM-CMG model modeling method is provided, and a flowchart thereof is shown in FIG. 1. Figure 1 As shown in the figure, the method specifically comprises:
[0036] S1: Extracting model parameters of an existing FinFET BSIM-CMG model.
[0037] The FinFET BSIM-CMG model is an advanced SPICE model specially designed for multi-gate transistors, which realizes circuit-level simulation through Verilog-A, and emphasizes practicality and speed. The existing FinFET BSIM-CMG model includes two parts of model equations and parameters. The model parameters in the existing FinFET BSIM-CMG model, such as channel length L, fin number nfin, etc., can be referred to BSIM-CMG model manual or parameter set. Specifically, it also includes point parameters matched with the model parameters: k0 (lateral non-uniform doping parameter), k0si (strong inversion correction factor), u0 (low field mobility), ua (surface roughness scattering parameter 1), ud (Coulomb scattering parameter 1), ucs (Coulomb scattering parameter 2), eu (surface roughness scattering parameter 2), and / or etamob (effective field parameter) and the like.
[0038] S2: Building a process corner model based on the model parameters.
[0039] This step is based on the model parameters extracted in step S1, using conventional methods in the art, to complete the construction of the target corner model.
[0040] The corner model can be a TT, FF, SS, etc. corner model. The parameters extracted in step S1 constitute the initial parameters of the TT, FF, SS, etc. corner model.
[0041] The corner model is used to simulate process deviations (such as doping concentration, etching accuracy, etc.) in the manufacturing process to ensure that the chip can still work normally under extreme conditions.
[0042] TT (Typical-Typical) model: NMOS and PMOS are in typical working state, transistor drive current is average value, representing process center value.
[0043] SS (Slow-Slow) model: NMOS and PMOS are in slow state, drive current is minimum value, corresponding to high threshold voltage (HVT) and low mobility, indicating pessimistic situation.
[0044] FF (Fast-Fast) model: NMOS and PMOS are in fast state, drive current is maximum value, corresponding to low threshold voltage (LVT) and high mobility, indicating optimistic situation.
[0045] In addition to the TT / SS / FF model, there are FS (Fast-Slow) and SF (Slow-Fast) models to cover the NMOS / PMOS asymmetric situation.
[0046] S3: Calculate the corner offset of each point of the corner model parameters.
[0047] The corner model parameters can be channel length L, fin number nfin, etc.
[0048] Take all the points for these parameters. For example, the point values of parameter L can be 0.016um, 0.018um, …, etc., and the point values of parameter nfin can be 2, 3, 4, …, etc.
[0049] The corner offset of the point mainly includes the threshold voltage offset dvth of the point, the linear region current offset didlin, and / or the saturation region current offset didsat, etc.
[0050] Taking the FF corner model as an example, the calculation method of the corner offset of the point is:
[0051] dvth = vth_FF - vth_TT, dvth is the threshold voltage offset of the FF process corner model threshold voltage vth_FF and the TT process corner model threshold voltage vth_TT.
[0052] didlin = idlin_FF / idlin_TT - 1, didlin is the linear region current offset of the FF process corner model linear region current idlin_FF and the TT process corner model linear region current idlin_TT divided by 1. Wherein, idlin is defined as the drain current when vds = 0.05v, vgs = 0.8v, wherein vds represents the drain and source voltage, and vgs represents the gate and source voltage.
[0053] didsat = idsat_FF / idsat_TT - 1, didsat is the saturation region current offset of the FF process corner model saturation region current idsat_FF and the TT process corner model saturation region current idsat_TT divided by 1.
[0054] S4: Construct the offset polyline graph with each point as the X axis and the corner offset of each point as the Y axis.
[0055] Based on each process corner model parameter, take each point as the X axis and the corner offset of each point as the Y axis to construct the offset polyline graph.
[0056] For example, in the didlin linear region current offset polyline graph constructed based on the process corner model parameter L, the coordinates (X, Y) of each point on the polyline are (point value, didlin value).
[0057] S5: Evaluate the corner offset, if the offset polyline contains a mutation point, mark the mutation point as a to-be-processed point.
[0058] This step determines whether the model has a smoothness problem caused by discontinuity (kink) according to whether the offset polyline graph has a mutation point. The mutation point is a point with a large difference in ordinate from the adjacent point. If there is a mutation point, the coordinates of the to-be-processed point need to be determined for further processing. There are many methods to evaluate whether there is a mutation point, which can generally be judged and evaluated using automatic methods such as derivatives, or can be found by intuitive observation of the polyline graph.
[0059] S6: Replace the parameter value of the to-be-processed point with the optimized parameter value, and fine-tune the parameter value of the to-be-processed point based on the measured data, so that the simulation result of the to-be-processed point is consistent with the measured data.
[0060] Optionally, the optimized parameter value is the parameter value of the point adjacent to the left of the point to be processed.
[0061] Optionally, the optimized parameter value is the parameter value of the point adjacent to the right of the point to be processed.
[0062] Optionally, the optimized parameter value is the average of the parameter values of the points adjacent to the left and right of the point to be processed.
[0063] Preferably, the measured data is measured IV (current-voltage) data.
[0064] This step eliminates the mutation point and solves the smoothness problem caused by the corner offset of the corner to be processed.
[0065] S7: Traverse all process corner model parameters and the corner offset of each point thereof, update the parameter value of all points to be processed, obtain an optimized process corner model, and obtain a final FinFET BSIM-CMG model.
[0066] The smoothness problem in the above steps is amplified by the process corner model, and the model is more accurate and efficient by analyzing the corner offset to determine whether the model has a smoothness problem. No additional size needs to be inserted, thereby reducing the workload and improving work efficiency.
[0067] Taking 14-nanometer nmos as an example, the FinFET BSIM-CMG model parameters are extracted, and TT models, FF models, and SS models are built. The corner offset of each point of the process corner model parameter is calculated and evaluated, and the point to be processed is found.
[0068] Figure 2 And Figure 3 In the above formula, the abscissa is the channel length L, the ordinate is the linear region current offset didlin, the units of the abscissa and the ordinate are um, and the broken line graph represents the change of the didlin offset with respect to L. In the figure, didilin_TT is the linear region current offset of the TT model, didilin_FF is the linear region current offset of the FF model, and didilin_SS is the linear region current offset of the SS model.
[0069] As shown in Figure 2 , the linear region current offset didlin of the point L=0.022um has obvious discontinuity, indicating that the model has a smoothness problem. Therefore, the point L=0.022um is evaluated as the point to be processed.
[0070] The parameter values of the point on the left adjacent to L=0.02um are copied to the point of L=0.022um, and the copied parameters mainly include k0 (lateral non-uniform doping parameter), k0si (strong inversion correction factor), u0 (low field mobility), ua (surface roughness scattering parameter 1), ud (Coulomb scattering parameter 1), ucs (Coulomb scattering parameter 2), eu (surface roughness scattering parameter 2), etamob (effective field parameter) and the like. Then, based on the measured IV (current-voltage) data of L=0.022um, the above parameters are fine-tuned to ensure that the simulation results of the point are consistent with the measured data.
[0071] The optimized results are shown in FIG. 3, and the line chart shows that the corner offset has eliminated the discontinuity. Until all the point corner offset is adjusted to be normal, it can be ensured that the model does not have the smoothness problem, and the final FinFET BSIM-CMG model is obtained. Figure 3
[0072] In the second aspect, based on the same design concept, the present application provides a FinFET BSIM-CMG model, and the method for constructing the model is consistent with the FinFET BSIM-CMG modeling method of the first aspect of the present application, which will not be described here.
[0073] The embodiments of the present application are described above in combination with the drawings, but the present application is not limited to the above specific embodiments, and the above specific embodiments are only illustrative but not restrictive. Those skilled in the art can make many forms under the inspiration of the present application without departing from the scope of the present application and the protection scope of the claims, and all the forms belong to the protection scope of the present application.
Claims
1. A FinFET BSIM-CMG model modeling method, characterized in that, The method comprises the following steps: S1: extracting model parameters of an existing FinFET BSIM-CMG model; S2: building a process corner model based on the model parameters; S3: calculating corner offset of each point of the process corner model parameters; S4: constructing an offset polyline with each point as the X axis and the corner offset of each point as the Y axis; S5: evaluating the corner offset, if the offset polyline contains a mutation point, marking the mutation point as a point to be processed; S6: replacing the parameter value of the point to be processed with an optimized parameter value, and fine-tuning the parameter value of the point to be processed based on the measured data, so that the simulation result of the point to be processed is consistent with the measured data; S7: traversing all process corner model parameters and their corner offsets of each point, updating the parameter values of all points to be processed, obtaining an optimized process corner model, and obtaining a final FinFET BSIM-CMG model.
2. The FinFET BSIM-CMG model modeling method according to claim 1, wherein the model parameters include channel length L, fin number nfin, and point parameters matched with the model parameters.
3. The FinFET BSIM-CMG model modeling method according to claim 2, wherein the point parameters include lateral non-uniform doping parameter k0, strong inversion correction factor k0si, low field mobility u0, surface roughness scattering parameter ua, coulomb scattering parameter ud, coulomb scattering parameter ucs, surface roughness scattering parameter eu, and / or effective field parameter etamob.
4. The FinFET BSIM-CMG model modeling method according to claim 1, wherein the process corner model includes a TT process corner model, a FF process corner model, and / or a SS process corner model.
5. The FinFET BSIM-CMG model modeling method according to claim 1, wherein the corner offset of the point includes threshold voltage offset dvth, linear region current offset didlin, and / or saturation region current offset didsat.
6. The FinFET BSIM-CMG model modeling method according to claim 5, wherein the calculation method of the corner offset of the point is: dvth = vth_FF - vth_TT, dvth is the threshold voltage offset of the threshold voltage vth_FF of the FF process corner model and the threshold voltage vth_TT of the TT process corner model; didlin = idlin_FF / idlin_TT - 1, didlin is the linear region current offset obtained by dividing the linear region current idlin_FF of the FF process corner model by the linear region current idlin_TT of the TT process corner model and subtracting 1; didsat = idsat_FF / idsat_TT - 1, didsat is a saturation region current offset of the FF process corner model saturation region current idsat_FF divided by the TT process corner model saturation region current idsat_TT minus 1.
7. The FinFET BSIM-CMG model modeling method of claim 1, wherein: the optimized parameter value is a parameter value of a left neighboring point of the point to be processed.
8. The FinFET BSIM-CMG model modeling method of claim 1, wherein: the optimized parameter value is a parameter value of a right neighboring point of the point to be processed.
9. The FinFET BSIM-CMG model modeling method of claim 1, wherein: the optimized parameter value is an average of parameter values of the left and right neighboring points of the point to be processed.
10. A FinFET BSIM-CMG model, comprising: the FinFET BSIM-CMG model is constructed by using the FinFET BSIM-CMG model modeling method of any one of claims 1 to 9.