Single event upset cross section simulation method based on TCAD-SPICE-Geant4 joint simulation

Abstract

The invention discloses a single event upset cross section simulation method based on TCAD-SPICE-Geant4 joint simulation, and belongs to the technical field of integrated circuit radiation effect and reinforcement. The method comprises the following steps: firstly, extracting a critical charge value for inducing single event upset of an integrated circuit through TCAD-SPICE joint simulation; secondly, for a single-tube device at a circuit sensitive node, TCAD is used for simulating transient current pulse curves of a single-particle incident device at different positions, and charge collection weights at different positions are extracted; using Geant4 simulation particles to irradiate the quantity of electric charges generated at different positions of the device; and finally, calculating to obtain the single event upset cross section of the circuit. The simulation method has high reliability, can accurately predict the sensitive degree of the integrated circuit to single particle radiation, and provides theoretical guidance for subsequent circuit-level radiation hardening design.

Description

Technical Field

[0001] The present invention belongs to the technical field of integrated circuit radiation effect and reinforcement, and in particular relates to a single particle flip cross-section simulation method based on TCAD-SPICE-Geant4 joint simulation. Background Art

[0002] Statistics released by the U.S. Nuclear and Space Radiation Effects Conference show that about 45% of spacecraft failures are attributed to the radiation effects of electronic components, and among the various radiation effects of electronic components, single-particle effects account for as much as 86%. It can be seen that single-particle effects are the main threat to the normal operation of integrated circuits in radiation environments. According to whether the circuit function of the integrated circuit can be restored after experiencing a single-particle event, the circuit failure caused by single-particle effects can be classified into soft errors and hard errors. Since the energy requirements of radiation particles for inducing soft errors are lower than those for inducing hard errors, single-particle-induced soft errors are considered to be the main cause of circuit failure. Among them, single-particle upset (SEU) refers to the circuit output upset caused by the energy deposited by single-particle irradiation at the sensitive nodes of the circuit to generate charges and be collected to form transient current pulses. It is the main source of soft errors in integrated circuits. Research on the SEU effect has now become the focus of integrated circuit radiation effects and radiation hardening technology.

[0003] The SEU cross section is an important parameter to measure the sensitivity of a circuit to a single particle. In traditional SEU cross-section simulations, the classic rectangular parallel-piped (RPP) model is usually used. By analyzing the cross-sectional sequence of the integrated circuit, the geometric structure of each layer of the circuit is simplified into a rectangular parallelepiped from top to bottom, and one of the small rectangular parallelepipeds is defined as the sensitive area. It is believed that the charge generated by the energy deposited by the radiation particles in this area will be collected to form a transient current pulse that affects the circuit output. Specifically, when the amount of charge collected in the sensitive area is greater than a certain critical value, an SEU occurs in the circuit output, and the ratio of the number of SEUs to the incident particle injection is the SEU cross section. The limitations of this method are: First, the sensitive area is replaced by a rectangular parallelepiped, which is too simplified and it is difficult to accurately reflect the impact of device structure changes on circuit SEU. Second, in the design of radiation-resistant integrated circuits, adjusting the layout is a commonly used anti-SEU reinforcement method, but the RPP model is difficult to accurately reflect the impact of layout changes on circuit SEU. Third, this method assumes that the charges generated by the energy deposited by radiation particles at different locations in the sensitive area will be collected with the same efficiency. However, in reality, due to the influence of the internal electric field of the device, the charge collection efficiency at different locations in the sensitive area varies. As the feature size of integrated circuits continues to shrink, the integration of devices increases dramatically, and the sensitive areas of adjacent devices overlap, the difference in charge collection efficiency at different locations in the sensitive area becomes more and more obvious, and the reliability of the RPP model is significantly reduced.

[0004] Therefore, it is of great practical significance to establish a more accurate integrated circuit SEU cross-section simulation method by considering the real device structure and combining the circuit layout. Summary of the invention

[0005] The purpose of the present invention is to overcome the above-mentioned shortcomings of the prior art. The present invention provides a single-particle flip cross-section simulation method based on TCAD-SPICE-Geant4 joint simulation, which can accurately predict the sensitivity of integrated circuits to single-particle radiation and provide theoretical guidance for subsequent circuit-level radiation resistance reinforcement design.

[0006] In order to solve the above technical problems, the present invention provides a single-particle upset cross-section simulation method based on TCAD-SPICE-Geant4 joint simulation, comprising:

[0007] Step 1: Use TCAD software to build a device model based on the device's process parameters;

[0008] Step 2: Based on the device model constructed in step 1, use TCAD simulation to obtain its transfer characteristic curve, compare it with the HSPICE simulation result, and calibrate the model by adjusting the model parameters;

[0009] Step 3: Set the gate center position of the device model calibrated in step 2 as the single particle incident point, change the linear energy transfer (LET) value of the incident particle, and use TCAD simulation to obtain the transient current pulse curve of the device under single particle irradiation with different LET values;

[0010] Step 4: Build the integrated circuit, call the process library, and use HSPICE to simulate and verify the circuit function;

[0011] Step 5: Inject the single-particle transient current pulse corresponding to the different LET values ​​simulated in step 3 into the sensitive node of the circuit in step 4, observe the change of the output waveform, and determine the critical LET value that induces SEU of the output signal;

[0012] Step 6: Integrate the single-particle transient current pulse curve corresponding to the critical LET value in step 5 to obtain the critical charge value that induces SEU of the integrated circuit;

[0013] Step 7: Set the LET value, change the single particle incident position, use TCAD simulation to obtain the transient current pulse curve when the single particle is incident on the device at different positions, calculate the amount of charge collected at different positions, and extract the charge collection weights at different positions of the device;

[0014] Step 8: Based on the real device model and circuit layout design, use Geant4 to build a circuit model;

[0015] Step 9: For the circuit model constructed in step 8, the single-tube device at the sensitive node is divided into regions according to the single particle incident position selected in step 7, and Geant4 is used to simulate the energy values ​​deposited in different regions of the device under particle irradiation, and convert it into the generated charge;

[0016] Step 10: Based on the critical charge value of the circuit SEU induced extracted in step 6, the charge collection weights at different positions of the device extracted in step 7, and the generated charge at different positions of the device extracted in step 9, the SEU cross section of the circuit is calculated.

[0017] Preferably, in step 1, the device model is constructed using Silvaco TCAD software.

[0018] Preferably, in step 2, based on the Cadence platform, a single-tube device model in a process library is called, and a single-tube transfer characteristic curve is simulated using an HSPICE simulator.

[0019] Preferably, in step 3, the device is placed in an off-bias state, and the single particle is incident perpendicularly to the device surface.

[0020] Preferably, in step 4, Cadence is used to build the integrated circuit.

[0021] Preferably, in step 6, the single particle transient current pulse curve describes the relationship between the drain current and time, that is, I d (t). d Integrate the time t to get the critical charge Q crit , as shown in formula (1):

[0022] Q crit =∫I d (t)dt(1)

[0023] Preferably, in step 7, the device is divided into n regions, and a single particle is incident from the top of the device in a direction perpendicular to the device surface toward the center of each region in turn. TCAD simulation is used to obtain a single particle transient current pulse curve generated when the single particle is incident on different positions of the device, and the charge value Q collected at different positions of the device is obtained by integrating the curve. i , i represents the i-th region (i=1,2,…,n). Therefore, the charge collection weight α at different positions of the device under single particle irradiation is i It can be expressed as:

[0024]

[0025] Where Q0 represents the charge value collected when a single particle is incident on the center of the device gate.

[0026] Preferably, in step 9, for the sensitive nodes in the circuit, the device at the sensitive nodes is divided into n regions according to step 7, and the center position of the region is the incident position of the radiation particle in step 7. Assuming that the charge collection weight in each region remains unchanged, Geant4 simulation is used to obtain the energy value deposited by the radiation particles in different regions of the device. Let E i represents the energy value of particle irradiation deposition in the i-th region (unit: eV), then the charge generated in the region Q i (Unit: C) The expression is:

[0027]

[0028] Preferably, in step 10, the total charge Q collected at the sensitive node of the circuit under particle irradiation is coll The expression is:

[0029]

[0030] Q coll The critical charge value Q of the induced circuit SEU crit For comparison, when Q coll ≥Qcrit When a SEU occurs in the circuit, the SEU cross-section expression of the circuit is:

[0031]

[0032] Among them, N SEU represents the number of flippings, and F represents the incident particle fluence.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] 1. The present invention constructs a device model based on the real device structure and establishes a circuit model according to the actual layout, so that the simulation results are more valuable for reference.

[0035] 2. Affected by the internal electric field of the device, the charge collection efficiency at different positions of the device is not the same. The present invention takes into account the charge collection weights at different positions of the device, so that the simulation results are more consistent with the actual test results.

[0036] 3. Based on TCAD-SPICE-Geant4 joint simulation, the present invention realizes the simulation of the SEU cross section of the integrated circuit, laying a foundation for the subsequent circuit-level radiation resistance reinforcement design. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a flow chart of a single particle flip cross section simulation method based on TCAD-SPICE-Geant4 joint simulation described in the present invention.

[0038] Figure 2 This is a diagram of the n-type silicon fin field-effect-transistor (FinFET) model built using Silvaco TCAD.

[0039] Figure 3 Comparison of the transfer characteristic curves of the FinFET device obtained by TCAD and HSPICE simulation after model calibration.

[0040] Figure 4 The single-particle transient current pulse curve generated in the FinFET device under single-particle irradiation with different LET values.

[0041] Figure 5 This is the schematic diagram of the NAND gate circuit.

[0042] Figure 6 This is a curve diagram of the circuit output voltage change caused by single particle irradiation of the sensitive node O with different LET values.

[0043] Figure 7 XZ plane diagram for n-type FinFET device area division.

[0044] Figure 8 This is a diagram of the FinFET NAND gate circuit model built using Geant4 based on the layout.

[0045] Fig. 9 The figure is a comparison diagram between the SEU cross section of the NAND gate circuit simulated by the method of the present invention and the experimental test result. DETAILED DESCRIPTION

[0046] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The advantages and features of the present invention will become more apparent from the following description. It should be noted that the accompanying drawings are in very simplified form and in non-precise proportions, and are only used to conveniently and clearly assist in explaining the purpose of the embodiments of the present invention.

[0047] like Figure 1 As shown, an embodiment of the present invention provides a single-particle upset cross-section simulation method based on TCAD-SPICE-Geant4 joint simulation, comprising the following steps:

[0048] Step 1: Use TCAD software to build a device model based on the device's process parameters.

[0049] Specifically, for Fin Field-Effect-Transistor (FinFET), referring to the process parameters in the standard industrial model BSIM-CMG, Silvaco TCAD is used to build an n-type bulk silicon FinFET single tube model, such as Figure 2 shown.

[0050] Step 2: Based on the device model constructed in step 1, use TCAD simulation to obtain its transfer characteristic curve, compare it with the HSPICE simulation result, and calibrate the model by adjusting the model parameters.

[0051] Specifically, based on the Cadence platform, the n-type FinFET model in the process library is called, and the transfer characteristic curve is simulated using the HSPICE simulator. Transfer characteristic curve simulation settings: drain voltage (V d ) is set to 0.8V, the source and substrate are grounded, and the gate voltage (V g ) increases from 0V to 1V with a step size of 0.1V, and the drain current (I d ) with gate voltage (V g ) change curve, that is, the transfer characteristic (I d -V g )curve.

[0052] The I obtained by TCAD simulation d -V gThe curve is compared with the HSPICE simulation results, and the model is calibrated by adjusting the source, drain, channel, and substrate doping concentrations of the device. After completing the model calibration, the I d -V g Curve contrast Figure 3 shown.

[0053] Step 3: Set the gate center position of the device model calibrated in step 2 as the single particle incident point, change the linear energy transfer (LET) value of the incident particle, and use TCAD simulation to obtain the transient current pulse curve of the device under single particle irradiation with different LET values.

[0054] Specifically, set the device gate voltage V g =0V, drain voltage V d = 0.8V, and set the single particle incident perpendicular to the device surface, and the LET value of the incident particle is 1MeV cm 2 / mg, 1.5MeV·cm 2 / mg, 2MeV·cm 2 / mg, and 2.5MeV·cm 2 / mg, TCAD simulation is used to obtain the single particle transient current pulse curve generated in the device under the single particle irradiation of the above LET value, as shown in Figure 4 shown.

[0055] Step 4: Build the integrated circuit, call the process library, and use HSPICE to simulate and verify the circuit function.

[0056] Specifically, based on the Cadence platform, a NAND gate circuit is built, and the schematic diagram is as follows: Figure 5 As shown in the figure, the HSPICE simulator is used to simulate and verify the circuit function.

[0057] Step 5: Inject the single-particle transient current pulse corresponding to the different LET values ​​simulated in step 3 into the sensitive nodes of the circuit in step 4, observe the changes in the output waveform, and determine the critical LET value that induces SEU in the output signal.

[0058] Specifically, Figure 5The NAND gate circuit shown includes: an input terminal A connected to the gates of a p-type FinFET tube P1 and an n-type FinFET tube N2, a substrate of the p-type FinFET tube P1 connected to the source, a drain connected to the drains of the p-type FinFET tube P2 and the n-type FinFET tube N1, and an output terminal O, a substrate of the p-type FinFET tube P2 connected to the source, a gate connected to the input terminal B and the gate of the n-type FinFET tube N1, a source of the n-type FinFET tube N1 connected to the drain of the n-type FinFET tube N2, a substrate connected to the substrate and source of the n-type FinFET tube N2 and grounded. Figure 5 Point O shown in is the sensitive node of the circuit. Point O corresponds to the output of the NAND gate. When inputs A and B are both low, the p-type FinFET is turned on, the n-type FinFET is turned off, and the output O is high. At this time, the substrate of the n-type FinFET device is grounded, and the drain is connected to a high level, that is, a reverse-biased pn junction is formed between the drain and the substrate. A strong electric field is formed in the depletion region of the reverse-biased pn junction due to the presence of space charge. The charge generated by the energy deposited by the radiation particles inside the n-type FinFET device is collected by the drain under the action of the electric field, forming a transient current pulse. When a single-particle transient current pulse acts on the output O, the output potential is instantly pulled down. As the LET value of the incident particle continues to increase, the energy deposited by the single-particle irradiation inside the device increases, thereby generating more charges, and the amount of charge collected by the drain increases accordingly, and the peak value of the transient current pulse formed increases, resulting in a lower output potential and a longer time at a low potential. When the LET value of the incident particle reaches the critical value, the potential of the output O is flipped from a high level to a low level and remains stably at a low level, that is, an SEU occurs in the NAND gate circuit under the action of single particle irradiation.

[0059] The transient current pulses generated in the n-type FinFET device under single particle irradiation with different LET values ​​obtained by TCAD simulation in step 3 are injected into the sensitive node O, and the output voltage change of the NAND gate is observed. The critical LET value that induces SEU of the output signal is determined to be 1.5 MeV cm 2 / mg, such as Figure 6 shown.

[0060] Step 6: Integrate the single-particle transient current pulse curve corresponding to the critical LET value in step 5 to obtain the critical charge value that induces SEU of the integrated circuit.

[0061] Specifically, Figure 4 Medium LET = 1.5 MeV cm 2 The single-particle transient pulse curve corresponding to / mg is integrated, and the expression is:

[0062] Q crit =∫I d (t)dt(1)

[0063] Thus, the critical charge value Q that induces SEU in the NAND gate circuit is obtained. crit =1.0fC.

[0064] Step 7: Set the LET value, change the single particle incident position, use TCAD simulation to obtain the transient current pulse curve when the single particle is incident on the device at different positions, calculate the amount of charge collected at different positions, and extract the charge collection weights at different positions of the device.

[0065] Specifically, six radiation particle sources, namely B, Ne, Ar, Cu, Kr, and Xe, are selected, and their corresponding LET values ​​are shown in Table 1.

[0066] Table 1

[0067]

[0068] Since the incident particles mainly deposit energy and generate charges at the fins and substrates of the FinFET device, Figure 2 The Fin and substrate parts of the FinFET device model are based on Figure 7 The device is divided into 25 regions as shown in the figure. A single particle is incident from the top of the device at the center of each region in a direction perpendicular to the device surface. TCAD simulation is used to obtain the single particle transient current pulse curve generated in the device when the single particle is incident at different positions. By integrating the curve, the charge value Q collected at different positions of the device is obtained. i , i represents the i-th region (i=1,2,…,25). Therefore, the charge collection weight α at different positions of the device under single particle irradiation is i It can be expressed as:

[0069]

[0070] Where Q0 represents the charge value collected when a single particle is incident on the center of the device gate.

[0071] Step 8: Based on the real device model and circuit layout design, use Geant4 to build a circuit model.

[0072] Specifically, according to the layout of the NAND gate, Geant4 is used to construct a bulk silicon FinFET NAND gate circuit model, such as Figure 8 shown.

[0073] Step 9: For the circuit model constructed in step 8, divide the single-tube device at the sensitive node into regions according to the single particle incident position selected in step 7, use Geant4 simulation to obtain the energy values ​​deposited in different regions of the device under particle irradiation, and convert it into the generated charge.

[0074] Specifically, the n-type FinFET device at the output terminal O of the NAND gate is divided into 25 regions according to step 7. The center position of the region is the incident position of the radiation particle in step 7. Assuming that the charge collection weight in each region remains unchanged, Geant4 is used to simulate and obtain the energy value deposited by the radiation particles in different regions of the device. The radiation particle source and the corresponding energy are shown in Table 1. The energy values ​​deposited by the radiation particles in different regions of the device at the sensitive node obtained by the simulation are processed and E i represents the energy value of particle irradiation deposition in the i-th region (unit: eV), then the charge generated in the region Q i (Unit: C) The expression is:

[0075]

[0076] Step 10: Based on the critical charge value of the circuit SEU induced extracted in step 6, the charge collection weights at different positions of the device extracted in step 7, and the generated charge at different positions of the device extracted in step 9, the SEU cross section of the circuit is calculated.

[0077] Specifically, the total charge Q collected at the sensitive nodes of the circuit under particle irradiation is coll The expression is:

[0078]

[0079] Q coll The critical charge value Q of the induced circuit SEU crit For comparison, when Q coll ≥Q crit When , a SEU occurs in the circuit. Then the SEU cross-section expression of the circuit is:

[0080]

[0081] Among them, N SEU represents the number of flippings, and F represents the incident particle fluence.

[0082] Fig. 9 This is a comparison chart of the SEU cross section of the FinFET NAND gate circuit simulated using the method described in the present invention and the experimental test results. It can be found that as the LET value of the radiation particle continues to increase, the change trend of the calculated value and the experimental value is basically consistent, indicating that the single-particle flip cross-section simulation method based on TCAD-SPICE-Geant4 joint simulation proposed in the present invention has high reliability, can accurately predict the sensitivity of integrated circuits to single-particle radiation, and provide theoretical guidance for subsequent circuit-level radiation hardening design.

[0083] The above description is only a description of the preferred embodiments of the present invention, and is not intended to limit the scope of the present invention. Any changes or modifications made by a person skilled in the art in the field of the present invention based on the above disclosure shall fall within the scope of protection of the claims.

Claims

1. A single-event upset cross-section simulation method based on TCAD-SPICE-Geant4 joint simulation, characterized in that: The steps include: Step 1: Use TCAD software to build a device model based on the device's process parameters; Step 2: Based on the device model constructed in step 1, use TCAD simulation to obtain its transfer characteristic curve, compare it with the HSPICE simulation result, and calibrate the model by adjusting the model parameters; Step 3: Set the gate center position of the device model calibrated in step 2 as the single particle incident point, change the linear energy transfer LET value of the incident particle, and use TCAD simulation to obtain the transient current pulse curve of the device under single particle irradiation with different LET values; Step 4: Build the integrated circuit, call the process library, and use HSPICE to simulate and verify the circuit function; Step 5: Inject the single-particle transient current pulse corresponding to the different LET values ​​simulated in step 3 into the sensitive node of the circuit in step 4, observe the change of the output waveform, and determine the critical LET value that induces SEU of the output signal; Step 6: Integrate the single-particle transient current pulse curve corresponding to the critical LET value in step 5 to obtain the critical charge value that induces SEU of the integrated circuit; Step 7: Set the LET value, change the single particle incident position, use TCAD simulation to obtain the transient current pulse curve when the single particle is incident on the device at different positions, calculate the amount of charge collected at different positions, and extract the charge collection weights at different positions of the device; Step 8: Based on the real device model and circuit layout design, use Geant4 to build a circuit model; Step 9: For the circuit model constructed in step 8, the single-tube device at the sensitive node is divided into regions according to the single particle incident position selected in step 7, and Geant4 is used to simulate the energy values ​​deposited in different regions of the device under particle irradiation, and convert it into the generated charge; Step 10: Based on the critical charge value of the circuit SEU induced extracted in step 6, the charge collection weights at different positions of the device extracted in step 7, and the generated charge at different positions of the device extracted in step 9, the SEU cross section of the circuit is calculated.

2. A single event upset cross section simulation method based on TCAD-SPICE-Geant4 joint simulation as claimed in claim 1, characterized in that: In step 1, the device model is constructed using Silvaco TCAD software.

3. A single event upset cross section simulation method based on TCAD-SPICE-Geant4 joint simulation as claimed in claim 1, characterized in that: In the step 2, based on the Cadence platform, the single-tube device model in the process library is called, and the single-tube transfer characteristic curve is simulated using the HSPICE simulator.

4. A single event upset cross section simulation method based on TCAD-SPICE-Geant4 joint simulation as claimed in claim 1, characterized in that: In step 3, the device is placed in an off-bias state, and the single particle is incident perpendicularly to the device surface.

5. A single event upset cross section simulation method based on TCAD-SPICE-Geant4 joint simulation as claimed in claim 1, characterized in that: In step 4, the integrated circuit is built using the Cadence platform.

6. A single event upset cross section simulation method based on TCAD-SPICE-Geant4 joint simulation as claimed in claim 1, characterized in that: In step 6, the single particle transient current pulse curve describes the relationship between the drain current and time, that is, Id(t); Id is integrated over time t to obtain the critical charge Q crit ,Right now: Q crit =∫I d (t)dt。 7. A single event upset cross section simulation method based on TCAD-SPICE-Geant4 joint simulation as claimed in claim 6, characterized in that: In step 7, the device is divided into n regions, and a single particle is incident from the top of the device in a direction perpendicular to the device surface toward the center of each region in turn. TCAD simulation is used to obtain the single particle transient current pulse curve generated when the single particle is incident on different positions of the device, and the charge value Q collected at different positions of the device is obtained by integrating the curve. i , i represents the i-th region, i=1,2,…,n; Therefore, the charge collection weight α at different positions of the device under single particle irradiation is i It is expressed as: Where Q0 represents the charge value collected when a single particle is incident on the center of the device gate.

8. A single event upset cross section simulation method based on TCAD-SPICE-Geant4 joint simulation as claimed in claim 7, characterized in that: In step 9, for the sensitive nodes in the circuit, the device at the sensitive node is divided into n regions according to step 7, and the center position of the region is the incident position of the radiation particle in step 7; assuming that the charge collection weight in each region is unchanged, Geant4 is used to simulate and obtain the energy value deposited by the radiation particle in different regions of the device; let E i represents the energy value of particle irradiation deposition in the i-th region, then the charge generated in the region Q i The expression is:

9. A single event upset cross section simulation method based on TCAD-SPICE-Geant4 joint simulation as claimed in claim 8, characterized in that: In step 10, the total charge Q collected at the sensitive node of the circuit under particle irradiation coll The expression is: Q coll The critical charge value Q of the induced circuit SEU crit For comparison, when Q coll ≥Q crit When a SEU occurs in the circuit, the SEU cross-section expression of the circuit is: Among them, N SEU represents the number of flippings, and F represents the incident particle fluence.

Images