A method and system for predicting the number of single event upsets of neutrons

Abstract

A method and system for predicting the number of single event upsets of neutrons. The method comprises a time of flight calculation step, a neutron energy calculation step, a single event upset cross section calculation step, and a single event upset event number prediction step. The invention is suitable for neutron soft error testing and reliability estimation in various environments, and can simultaneously complete both thermal neutron and fast neutron testing.

Description

Technical Field

[0001] This invention relates to cosmic ray reliability testing and prediction technology for semiconductor devices, specifically to a method and system for predicting the number of neutron single-event flips in FPGA devices. Background Technology

[0002] The neutron energy spectrum in the natural living environment of the Earth's surface is generally called the atmospheric neutron spectrum. It is an ultra-broad neutron energy spectrum produced by the multiple scattering of cosmic rays by gas atoms and molecules in the atmosphere, covering thermal neutrons to high-energy fast neutrons above GeV. Atmospheric neutron single-event effects specifically refer to single-event effects triggered by atmospheric neutrons near the Earth's surface. Single-event effects (SEE) refer to the phenomenon where, when a particle passes through the sensitive area of ​​an electronic device, it generates a large number of charged particles in the semiconductor device through ionization, causing errors in the device, resulting in single-event flips, device latch-up, or even permanent damage to logic devices and memory. Neutron single-event effects are single-event effects triggered by neutrons, including fast neutrons and thermal neutrons. Because neutrons are uncharged, they have extremely strong penetrating power.

[0003] When high-energy fast neutrons pass through a chip, they may undergo nuclear reactions with the atomic nuclei of silicon wafers, such as 28Si(n,α)25Mg, 28Si(n,p)28Al, and 28Si(n,nα)24Mg, producing alpha particles, recoil protons, and recoil nuclei. These nuclear reaction products deposit energy in the sensitive area of ​​the device while generating electron-hole pairs. The instantaneous accumulation of these charges can cause errors such as neutron single-event flips in the device.

[0004] Besides fast neutrons triggering inelastic scattering nuclear reactions, thermal neutrons can also be captured by impurity atoms in devices, inducing nuclear reactions and producing single-event effects. This is mainly due to the presence of boron in semiconductor materials (…). 10 B) and other elements, 10 B is a naturally occurring stable isotope with an abundance of up to 19.9%. 10 B is used as a neutron-capturing substance in nuclear reactors, and its reaction formula is:

[0005]

[0006] here 10 After B captures a neutron, it splits into two energetic fragments, and in the above equation... 10 B has a very high reaction cross section, especially for the capture of thermal neutrons. Figure 1 As shown.

[0007] Because boron has a high natural abundance and its atomic weight is very close to that of silicon, it is extremely difficult to completely remove boron from silicon materials. Boron is not only present in polycrystalline silicon and doped substrates, but can also be introduced onto the surface of wafer materials due to certain process requirements. For example, boro-phospho-silicate glass (BPSG) may be used in some steps of wafer fabrication, and BPSG has wide applications in IC manufacturing. In boro-phospho-silicate glass, the addition of boron and phosphorus impurities (B2O3, P2O5) loosens the original ordered network structure of silicon dioxide, giving it a certain degree of reflow capability at high temperatures. Therefore, BPSG films have excellent pore-filling ability and can planarize the entire silicon wafer surface, thus providing a wider process range for photolithography and subsequent processes. BPSG typically contains 4-9% boron (by weight) and covers most of the area, which leads to natural boron entering the base material of chip manufacturing. Therefore, modern semiconductor memory chips have a high probability of experiencing single-event effects due to boron neutron capture reactions.

[0008] There are many types of single-event effects, mainly including single-event upsets (SEUs), single-event latch-up, and single-event burnouts, among which the SEU phenomenon in memory cells is the most significant. A SEU refers to a transient logic change in a semiconductor circuit under the influence of external environmental factors, such as changing from logic 0 to logic 1. Also known as a soft error, soft errors can be corrected by rewriting, but if data errors occur in critical memory locations, they can lead to extremely serious losses. The probability of soft errors in memory devices is one of the important reliability indicators of semiconductor devices and must be determined experimentally.

[0009] The most direct method for measuring the neutron single-event flip cross section is to deploy a large number of samples in a natural environment and detect the probability of soft errors. While choosing a high-altitude location with relatively high neutron flux can yield cross section data relatively quickly, a more common approach is to use a high-current fast neutron beam to accelerate the testing process, achieving speedup ratios as high as 1e7 to 1e10. Traditional accelerators primarily use nuclear reactions (such as the DT reaction) to generate a 14 MeV fast neutron beam to induce single-event flips (i.e., soft errors) in semiconductor memory devices. The number of flipped bits is measured, and the ratio of the number of flipped bits to the total neutron flux received by the device is used as a flip cross section. This cross section is then used as a baseline to further predict soft errors induced by atmospheric neutrons in a natural environment. However, such accelerators have a limited neutron energy spectrum and low flux, making them difficult to meet experimental requirements. Currently, international and domestic research and testing mainly employ proton-driven spallation neutron sources to generate high-energy fast neutrons. Their generation principle is very similar to that of neutrons generated by cosmic ray showers, and they can produce a very wide energy spectrum. The emergence of spallation neutron sources has solved the problem of flux intensity. Besides flux intensity, another issue is the energy spectrum. Current approaches primarily focus on simulating the atmospheric neutron environment as closely as possible, designing and constructing neutron beamlines that more closely resemble the atmospheric neutron spectrum, and applying them to neutron single-event effect testing.

[0010] Furthermore, to standardize the process and result analysis of integrated circuit irradiation testing, the JEDEC Solid State Technology Association spearheaded the release of the industry standard document JESD89A in 2006. This document defines the requirements, procedures, and result publication format for integrated circuit soft error rate testing. In this standard document, JEDEC also established a standard atmospheric neutron spectrum, with the benchmark being the neutron spectrum and flux intensity outdoors in New York City during the mean solar activity cycle at sea level, applicable to neutrons above 1 MeV. In addition, NASA also released the IEC TS62396-1:2006 standard based on high-altitude flight test data. The shape of the neutron spectrum in this standard is basically similar to the JEDEC definition, but the proportion of high-energy neutrons is slightly higher, and the spectrum is harder. This is mainly because the measurement location is at an altitude of tens of thousands of meters, and the flux intensity is much higher than at sea level, approximately 440 times the neutron flux intensity at sea level. The JEDEC specification also defines a method for measuring the thermal neutron single-event flip section, which mainly involves irradiating the device with moderated standard thermal neutrons and calculating the flip section by counting the flips.

[0011] The specific information for the standard document JESD89 is as follows: JEDEC JESD98A, “Measurement and reporting of Alpha Particle and Terrestrial Cosmic Ray-Induced Soft Errors in Semiconductor Devices.”, 2006.

[0012] The specific information of the IEC TS62396-1:2006 standard document is as follows: IEC TS62396-1, “Process Management for avionics-Atmospheric radiation effects,” 2006.

[0013] from Figure 2 It can also be seen that the shapes of each neutron beamline differ from the standard atmospheric neutron energy spectrum, and they are also different from each other. This means that the results obtained by the same device at different test sites are very likely to be different, and they are not consistent with the actual application. Therefore, it is necessary to further refine the entire test logic through differential cross section measurement.

[0014] The existing technology has two main drawbacks. First, it requires separate measurements of fast and thermal neutrons, necessitating at least two experiments and two sets of results. Second, regarding fast neutrons, current methods primarily involve constructing simulated beamlines for the target environment, limiting their application to a single target environment—the atmospheric neutron spectrum at the Earth's surface. However, as humanity ventures into space, we need to predict the probability of single-event effects in various neutron fields, such as near-Earth space, where the neutron energy spectrum is generally harder and has a higher average value. Future research could also focus on outer space, including the effects of albedo neutrons on the lunar surface, and artificial particle environments like colliders, examining their neutron environment's impact on semiconductor measurement devices. Therefore, this traditional, generalized cross-section measurement method is unsuitable for diverse needs. A more universal energy differential cross-section method is required to predict neutron flips in various neutron fields. In practice, because the cross-section for neutron single-event flips is extremely small, single-event flip testing requires extremely high neutron fluence to ensure a sufficient number of events to support differential cross-section calculations. Summary of the Invention

[0015] According to a first aspect, one embodiment provides a method for predicting the number of neutron single-particle flips, comprising:

[0016] The steps for calculating flight time include calculating the flight time using the following formula. :

[0017]

[0018] In formula (1), The moment when the kicker signal enters the FPGA device and triggers the time-of-flight measurement; To record the moment when a single-particle flip event occurs; The time of flight is the output after differential comparison by the ToF detection board; The timing of the gamma pulse signal detected by the gamma detector and the kicker trigger signal of the accelerator. The time interval between; The time it takes for the gamma pulse signal to travel a specific distance, where the specific distance refers to the distance from the neutron source to the detector; This refers to the time delay during the data readout process within the FPGA.

[0019] The neutron energy calculation steps include calculating the neutron's velocity based on the flight time obtained from the flight time calculation steps, using the following formula. v :

[0020]

[0021] In formula (2), L It is the distance a neutron travels, measured in meters (m).

[0022] It is the neutron flight time, measured in seconds (s).

[0023] The neutron energy is then calculated using the following formula. E n :

[0024]

[0025] In formula (3), v is the neutron's flight speed, in m / s;

[0026] m 0 It is the mass of a neutron, measured in kg.

[0027] c It is the speed of light, measured in m / s;

[0028] The steps for calculating the single-particle flip-off cross section include obtaining the energy as follows: E n The number of single-event flips caused by neutrons is calculated using the following formula to obtain the single-event flip cross section. :

[0029]

[0030] In formula (4), It is neutron energy The single-particle flip section at the location is expressed in cm. 2 / bit;

[0031] It is energy The number of single-particle flips caused by neutrons;

[0032] It is the neutron energy during the experiment. Neutron flux at a given location, in units of n / cm 2 ;

[0033] S represents the number of bits in the device's built-in static memory, measured in bits.

[0034] The single-event flip event count prediction step includes, based on the single-event flip cross section... neutron energy The number of neutron single-particle flips under the required environment is predicted by using the number of bits S of the static memory within the device.

[0035] According to a second aspect, one embodiment provides a system for predicting the number of single-particle flips, comprising:

[0036] The flight time calculation module is used to calculate the flight time according to the following formula. :

[0037]

[0038] In formula (1), The moment when the kicker signal enters the FPGA to trigger time-of-flight measurement; To record the moment when a single-particle flip event occurs; The time of flight is the output after differential comparison by the ToF detection board; The timing of gamma pulse signal detection by the gamma detector and the RCS kicker trigger signal. The time interval between; The time it takes for the gamma pulse signal to travel a specific distance, where the specific distance refers to the distance between the neutron source and the detector. This refers to the time delay during the data readout process within the FPGA.

[0039] The neutron energy calculation module is used to calculate the flight time based on the flight time calculation steps. The neutron's flight speed is calculated using the following formula. v :

[0040]

[0041] In formula (2), L It is the distance a neutron travels, measured in meters (m).

[0042] It is the neutron flight time, measured in seconds (s).

[0043] The neutron energy is then calculated using the following formula. E n :

[0044]

[0045] In formula (3), v is the neutron's flight speed, in m / s;

[0046] m 0 It is the mass of a neutron, measured in kg.

[0047] c It is the speed of light, measured in m / s;

[0048] The single-particle flip-off section calculation module is used to obtain the energy of... E n The number of single-event flips caused by neutrons is calculated using the following formula to obtain the single-event flip cross section. :

[0049]

[0050] In formula (4), It is neutron energy En The single-particle flip section at the location is expressed in cm. 2 / bit;

[0051] It is energy En The number of single-particle flips caused by neutrons;

[0052] It is the neutron energy during the experiment. En Neutron flux at a given location, in units of n / cm 2 ;

[0053] S represents the number of bits in the device's built-in static memory, measured in bits.

[0054] The single-event flip event count prediction module is used to predict the number of single-event flip events based on the single-event flip cross section. neutron energy E n The number of neutron single-particle flips under the required environment is predicted by using the number of bits S of the static memory within the device.

[0055] According to the above embodiments, a method and system for predicting the number of single-particle flips are applicable to neutron soft error testing and reliability estimation under various environments, and can simultaneously perform both thermal neutron and fast neutron tests. Attached Figure Description

[0056] Figure 1 for 10 A cross-sectional view of the capture reaction between B and neutrons;

[0057] Figure 2 Neutron energy spectra from major international atmospheric neutron testing stations;

[0058] Figure 3 Layout diagram of the single-particle flip detection and time-of-flight measurement experiment;

[0059] Figure 4 This is a schematic diagram of the output circuit for neutron TOF detection.

[0060] Figure 5 Calibration diagram for the neutron flight initiation time T0;

[0061] Figure 6 This is the schematic diagram of the SEU detection circuit;

[0062] Figure 7 Timing diagram of the SEU detection circuit;

[0063] Figure 8 A diagram illustrating the calculation method for the single-particle flip probability induced by equivalent thermal neutrons;

[0064] Figure 9 This is a flowchart of a single-particle flip test. Detailed Implementation

[0065] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. In the following embodiments, many details are described to facilitate a better understanding of the present application. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other materials or methods. In some cases, certain operations related to the present application are not shown or described in the specification. This is to avoid obscuring the core parts of the present application with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0066] Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can be rearranged or adjusted in a manner obvious to those skilled in the art. Therefore, the various orders in the specification and drawings are only for the clear description of a particular embodiment and do not imply a necessary order, unless otherwise stated that a particular order must be followed.

[0067] The serial numbers assigned to components in this article, such as "first" and "second", are used only to distinguish the objects being described and have no sequential or technical meaning.

[0068] In one embodiment, the present invention aims to measure and predict the probability of soft failure of semiconductor memory devices induced by cosmic rays, which is an application of nuclear technology in semiconductor device reliability issues.

[0069] As used in this article, "near-Earth space" refers to the spatial region ranging from 1.015 to 6.6 Earth radii from the Earth's center, that is, a spherical shell-shaped region extending from the Earth's sea level to approximately 100 to 36,000 km (about 10 Earth radii).

[0070] As used in this article, "outer space" refers to the cosmic space outside the Earth's atmosphere.

[0071] As used in this article, "fast neutron" refers to a high-energy neutron that can induce inelastic scattering and nuclear reactions; in this article, it refers to a neutron with a kinetic energy of 100 keV or higher.

[0072] According to the JEDEC specification, a "thermal neutron" typically refers to a neutron below 20 eV, with a central energy of 0.025 eV (velocity approximately 2.2 km / s). The neutron velocity corresponds to the most probable velocity under the Maxwell-Boltzmann distribution at 290 K (approximately 17 °C). In this paper, we use the equivalent thermal neutron portion of the energy spectrum for substitution testing, representing the neutrons that can be induced... 10 The neutrons in the boron fission reaction, mainly those below 100 keV, have a large boron neutron capture cross section.

[0073] According to a first aspect, in one embodiment, a method for predicting the number of neutron single-particle flips is provided, comprising:

[0074] The steps for calculating flight time include calculating the flight time using the following formula. :

[0075]

[0076] In formula (1), The moment when the accelerator kicker signal enters the FPGA device to trigger time-of-flight measurement; To record the moment when a single-particle flip event occurs; The time of flight is the output after differential comparison by the ToF detection board; The timing of the gamma pulse signal detected by the gamma detector and the kicker trigger signal of the accelerator. The time interval between; This refers to the time it takes for the gamma pulse signal to travel a specific distance, where the specific distance is the distance from the neutron source to the detector. This refers to the time delay during the data readout process within the FPGA.

[0077] The neutron energy calculation steps include the flight time obtained from the flight time calculation steps. The neutron's flight speed is calculated using the following formula. v :

[0078]

[0079] In formula (2), L It is the distance a neutron travels, measured in meters (m).

[0080] It is the neutron flight time, measured in seconds (s).

[0081] The neutron energy is then calculated using the following formula. E n :

[0082]

[0083] In formula (3), v It is the speed of a neutron, measured in m / s;

[0084] m 0 It is the mass of a neutron, measured in kg.

[0085] c It is the speed of light, measured in m / s;

[0086] The steps for calculating the single-particle flip-off cross section include obtaining the energy as follows: E n The number of single-event flips caused by neutrons is calculated using the following formula to obtain the single-event flip cross section. :

[0087]

[0088] In formula (4), : is neutron energy EnThe single-particle flip section at the location is expressed in cm. 2 / bit;

[0089] It is energy En The number of single-particle flips caused by neutrons;

[0090] It is the neutron energy during the experiment. En The neutron flux at a given location is determined by the neutron energy spectrum and total irradiation time of the environment being measured, and is expressed in n / cm². 2 ;

[0091] S represents the number of bits in the device's built-in static memory, measured in bits.

[0092] The single-event flip event count prediction step includes the following steps based on the single-event flip cross section. Neutron energy spectrum of the environment to be predicted In addition to the number of bits S of the static memory within the device, the number of neutron single-event flips of the FPGA under test in the required environment is predicted. .

[0093] In one embodiment, in formula (1), the constant The value is 6, expressed in nanoseconds (ns), representing the time delay of the data within the FPGA under test during the measurement process. It could also be other values.

[0094] In one embodiment, a specific number is pre-stored in the static RAM of the FPGA device and compared by the logic circuit built into the FPGA device, thereby realizing the measurement of a single-event flip event in the static RAM within an ultra-short time, where ultra-short time refers to <100ns.

[0095] In one embodiment, in the single-event reversal neutron energy calculation step, the flight time start is provided by an accelerator timing signal. After the single-event reaction occurs, the test conditions are reset to provide an equivalent test environment. Through repeated cyclic testing, an ultra-high total neutron fluence is provided to the FPGA device. An ultra-high total neutron fluence refers to >1E12n / cm². 2 Neutron flux.

[0096] In one embodiment, in the single-particle flip-off cross-section calculation step, the energy is obtained as follows: E n The method for determining the number of single-event flips caused by neutrons is to use the built-in logic circuitry of the device for comparison and reading.

[0097] In one embodiment, in the single-particle flip-off cross-section calculation step, the energy is obtained by comparison reading method. E n The number of single-particle flips caused by neutrons, specifically including:

[0098] A module is built based on the block RAM. At the start of the test, the registers within the module are initialized. SEU detection is performed using the FPGA's built-in logic circuitry. For example, a NOR gate is used for the first stage of comparison, followed by two AND gates to sum the results. Each module outputs one bit. When an SEU occurs at any point within the module, the output FLAG bit will flip. This logic comparison determines whether a single-event upset (SWE) has occurred. The error signal from this flip is output to the SWE detection board. By comparing it with the kicker signal, the difference between these two time signals is output, representing the original flight duration of the SWE event.

[0099] In one embodiment, the environment includes a terrestrial environment, near-Earth space, outer space, or an artificial ion environment.

[0100] In one embodiment, the number of single-event flips is calculated in the single-event flip count prediction step according to the following formula:

[0101]

[0102] In formula (5), It is neutron energy En The single-particle flip section at the location is expressed in cm. 2 / bit;

[0103] Φ ( E n The neutron energy is E n Neutron flux, in units of n / cm 2 ;

[0104] S represents the number of bits in the device's internal static memory, measured in bits.

[0105] In one embodiment, the FPGA device includes an SRAM (Static Random-Access Memory) type FPGA device.

[0106] In one embodiment, the neutron source includes a white light neutron source.

[0107] In one embodiment, the neutron source is a pulsed, wide-spectrum white-light neutron source. Wide spectrum refers to 0.5 eV to 200 MeV.

[0108] In one embodiment, the white light neutron source includes a pulsed white light neutron source.

[0109] According to a second aspect, in one embodiment, a system for predicting the number of neutron single-particle flips is provided, comprising:

[0110] The flight time calculation module is used to calculate the flight time according to the following formula. :

[0111]

[0112] In formula (1), The moment when the accelerator kicker signal enters the FPGA device to trigger time-of-flight measurement; To record the moment when a single-particle flip event occurs; The time of flight is the output after differential comparison by the ToF detection board; The timing of the gamma pulse signal detected by the gamma detector and the kicker trigger signal of the accelerator. The time interval between; This refers to the time it takes for the gamma pulse signal to travel a specific distance, where the specific distance is the distance from the neutron source to the detector. This refers to the time delay during the data readout process within the FPGA.

[0113] The neutron energy calculation module is used to calculate the flight time based on the flight time calculation steps. The neutron's flight speed is calculated using the following formula. v :

[0114]

[0115] In formula (2), L It is the distance a neutron travels, measured in meters (m).

[0116] It is the neutron flight time, measured in seconds (s).

[0117] The neutron energy is then calculated using the following formula. :

[0118]

[0119] In formula (3), v is the neutron's flight speed, in m / s;

[0120] m 0 It is the mass of a neutron, measured in kg.

[0121] c It is the speed of light, measured in m / s;

[0122] The single-particle flip-off section calculation module is used to obtain the energy of... En The number of single-event flips caused by neutrons is calculated using the following formula to obtain the single-event flip cross section. :

[0123]

[0124] In formula (4), It is neutron energy En The single-particle flip section at the location is expressed in cm. 2 / bit;

[0125] It is energy The number of single-particle flips caused by neutrons;

[0126] It is the neutron energy during the experiment. Neutron flux at a given location, in units of n / cm 2 ;

[0127] S represents the number of bits in the device's built-in static memory, measured in bits.

[0128] The single-event flip event count prediction module is used to predict the number of single-event flip events based on the single-event flip cross section. Neutron energy spectrum of the environment to be predicted In addition to the number of bits S of the static memory within the device, the number of neutron single-event flips of the FPGA under test in the required environment is predicted. .

[0129] In one embodiment, the single-particle flip-off section calculation module obtains the energy as follows: E n The method for determining the number of single-event flips caused by neutrons is to use the built-in logic circuitry of the device for comparison and reading.

[0130] In one embodiment, the present invention provides a method for measuring and predicting the probability of soft failure of semiconductor memory devices induced by cosmic rays, which belongs to the application of nuclear technology in the reliability testing of semiconductor devices.

[0131] Example 1

[0132] For fast neutrons, since their generation mechanism involves high-energy neutrons triggering nuclear reactions in chip materials and depositing energy, their single-event flip cross-section has a significant dependence on the energy of the incident neutron. Therefore, by combining the time-of-flight method and measuring the energy-resolved neutron single-event flip differential cross-section, the probability of single-event flip under any neutron environment can be obtained. Thus, the soft failure probability at energy point En is defined as: .

[0133] Where Φ ( E n The neutron energy is E nNeutron fluence (n / cm) 2 ), where SER represents the number of SEU events. Therefore, the total number of SEU flips under any neutron environment is the integral of the differential cross section and the neutron field energy spectrum. Thus, the predicted number of SEU flips will be the integral of the energy-resolved cross section and the neutron flux. As long as the neutron energy spectrum of the environment is measured beforehand, the predicted number of neutron single-particle flips under the required ground environment can be obtained by directly substituting into the above equation and integrating. This method is applicable not only to ground environments but also to high-altitude environments, near-space, and even space stations. As long as the neutron energy spectrum of the environment is monitored, the flip probability of semiconductor memory devices under that environment can be predicted. Artificial particle environments, such as those around accelerators, have a large number of particles scattered into space. Such environments also place high demands on device reliability; however, if the energy spectrum is measured, soft error predictions can also be obtained.

[0134] Numerous experiments have measured the thermal neutron environment at the Earth's surface. Although the neutron energy spectrum fluctuates at different altitudes and locations, there are generally certain standards, and the measurement results are typically used as a reference. Figure 8 As shown, the acceleration ratio is The atmospheric neutron spectrum has three peaks, with the thermal neutron peak located at 0.025 eV. Thermal neutrons in the atmospheric neutron spectrum are generally considered to be below 20 eV; however, based on the principle of the thermal neutron single-event effect, as long as a trigger can be established... 10 All β-fission reactions can be considered within the effective range of thermal neutron single-event effects. Therefore, in this embodiment, the region in the anti-angle white light neutron energy spectrum, excluding inelastic fission reactions, is considered the effective region of thermal neutron single-event effects. In practice, the region with neutron energies below 100 keV obtained through time-of-flight measurements is considered the effective region of thermal neutron single-event effects.

[0135] Therefore, based on the neutron energy spectrum and 10 The acceleration factor of the equivalent thermal neutron single-event effect can be obtained from the integral ratio of the cross section of the B fission reaction.

[0136]

[0137] For a white beamline with an anti-angle, the acceleration factor for its thermal neutron single-event flip test is: ,like Figure 8 As shown, within the same time frame, the probability of a thermal neutron single-event event occurring in beamline measurements is higher than that in atmospheric thermal neutron single-event measurements. This multiple, or the probability of a thermal neutron single-event effect occurring within one hour measured on the beamline, is equivalent to the probability under atmospheric thermal neutron conditions. The probability of a thermal neutron single-event effect per hour, that is, the probability of a single-event effect occurring accelerated by high-throughput beamline testing.

[0138] The neutron energy corresponding to the SEU event was measured using the time-of-flight (TOF) method. The signal from the accelerator kicker was used as the trigger pulse. The timing information of the SEU event was obtained through multi-layer logic comparison on an FPGA (Field Programmable Gate Array) test board and output to the time-of-flight measurement board. Then, the SEU output was subtracted from the kicker T0 in the experimental control board to obtain the preliminary distribution of the neutron single-particle flip section over time. Finally, the mass-energy equivalence equation was used.

[0139]

[0140] in, v is the neutron's velocity (m / s), m0 is the neutron's mass (kg), and c is the speed of light (m / s). The TOF spectrum can be converted into an energy spectrum.

[0141] Neutron flight speed v Calculated using the following formula:

[0142]

[0143] L is the distance (m) that the neutron travels. It is the neutron flight time (s).

[0144] Based on the neutron flux spectrum during the experiment, and according to the cross-section calculation formula:

[0145]

[0146] in, It is neutron energy En Single-particle flip section at (cm) 2 / bit. It is energy En The number of single-particle flips caused by neutrons. It is the neutron energy during the experiment. E n Neutron fluence at the location (n / cm) 2 S represents the number of bits in the device's static memory. The single-particle flip-off cross section related to the neutron energy can be calculated.

[0147] Simultaneously with the SEU test, the neutron energy corresponding to the SEU event is measured via time-of-flight. The test system layout is as follows: Figure 3As shown, the system uses a pulsed wide-spectrum white light neutron source as the test neutron source for single-event flip (SIF) testing. A time-of-flight (TOF) detection FPGA is installed at a specific distance from the neutron source. The Kicker signal derived from the RCS is used as the trigger signal to start the timer. The delay between the Kicker signal and the actual T0 is calibrated using a gamma detector near the experimental point to determine the starting point of the TOF. The specific experimental design is given below. The ending point of the TOF is usually determined by a neutron detector placed at the end point. In this embodiment, the FPGA chip reads the time of TOF recording, which involves FPGA chip design and delay calculations, and will be discussed below. Based on the above test implementation, the energy-resolved differential cross section of the TOF can be obtained through data analysis.

[0148] A high-throughput, large-spot beam was employed, along with multiple test plates connected in series for measurement, to increase the total count of single-particle flip events. The overall experimental layout is as follows: Figure 3 As shown, the beam passes through a neutron switch and collimator before sequentially passing through multiple FPGA test boards. The SEU signal output from the FPGA development board is connected to the TOF detection board. The signal abrupt changes caused by SEU events are recorded and compared with the accelerator kicker T0 signal to record a coarse relative time of flight.

[0149] Based on the PXIe architecture, the TOF test board supports multiple logic comparison SEU signal inputs and one accelerator Kicker T0 signal input. Serial measurement with multiple test boards can improve the statistical accuracy of experimental results. A PXIe chassis controller is used to build the SEU time-of-flight test circuit, providing a time measurement accuracy of <5ns. Its main function is to differentially analyze and output the single SEU signal and the kicker signal.

[0150] Time of flight measurement methods:

[0151] The key to time-of-flight measurement is accurately determining the start and end times of neutron flight. The determination of these two times and the calculation of the flight time are implemented using an FPGA development board, the circuit schematic of which is shown below. Figure 4 As shown, the rising edge of the Kicker signal from the RCS triggers the timer to begin timing and is considered the start time of neutron flight (in experimental measurements, no distinction is made between high-energy fast neutrons and thermal neutrons; that is, neutrons in the measurement experiment include both fast and thermal neutrons). However, the Kicker signal is not the actual start time of neutron flight, but rather the moment when the proton leaves the ring accelerator. Therefore, we use a gamma detector to calibrate T0. The specific principle is as follows: Gamma is also produced simultaneously with neutrons. We can determine the moment when gamma is produced as the actual start time of neutron flight.Figure 5 This is the calibration result. The green line signal is the signal generated by the Kicker signal and actually input to the FPGA, while the red gamma flash signal is the gamma pulse signal recorded by a gamma detector located near the experimental site. Therefore, the actual neutron flight start time is the Kicker signal plus 3492 ns minus the time it takes for the photon to travel that distance, i.e. Figure 5 The location is indicated by the red vertical dashed line. Regarding the termination time of the neutron flight, in this system, the moment the FPGA chip reads the single-event upset (SED) event is taken as the termination time of the neutron flight. See [link / reference]. Figure 4 The schematic diagram of the neutron TOF calculation circuit shown illustrates that when a single-event upset occurs in the DUT, an Error signal is generated and sent to the TOF calculation board. The Error signal triggers a timer to stop counting and outputs the result to the host computer. Simultaneously, the timer is reset to record the next neutron flight time. Considering the internal signal delay of the FPGA, the final measured neutron flight time is:

[0152]

[0153] In the formula: —The moment when the kicker signal enters the FPGA to trigger the TOF measurement; —Records the moment the SEU incident occurred; —The output data of the TOF calculation and communication FPGA is the time of flight output after differential comparison by the TOF detection board; —The moment when the detector detected the gamma (γ) pulse signal and The time interval between; —The time it takes for the gamma pulse signal to travel a specific distance (from the neutron source to the detector). In the formula, the constant 6 refers to 11 ns, which is the time delay during the data readout process within the FPGA in the experiment. Specifically, it is calculated from "-16 + 10 = -6". The constant ±5 indicates a time uncertainty of 5 ns, which is the maximum possible time difference that may occur when the two test boards are asynchronous.

[0154] SEU Measurement Scheme and FPGA Flowchart

[0155] Comparison-reading method SEU detection circuit, such as Figure 6As shown. The principle is to combine multiple BRAMs (Block Random Access Memory) to form a large BRAM_Module. Specific numbers are pre-stored in the bits within the module, and logical comparisons are performed using the device's built-in programmable circuitry. Each BRAM_Module outputs only a Flag bit; a change in this flag bit indicates a single-event flip at a specific location within the module. To achieve a large bit capacity, the BRAM_Module employs multi-layered logic operation circuitry. The first stage uses NOR gates for comparison, the second and third stages use AND operations, and the final output is the FLAG bit, as shown. Figure 6 The enlarged view below shows the process. The output of the BRAM_Module is read cyclically to determine if a single-event upset (SWE) has occurred. A state machine model is typically used for this purpose. If a SWE occurs, an Error flag is sent to the Time-of-Flight (TOF) calculation board to trigger TOF calculations; this is the standard test. If no SWE occurs after a certain period, it is considered a potential anomaly, and the system is reset. Figure 7 This is the timing diagram for the SEU detection circuit. In the diagram, the Error signal represents the TOF trigger signal issued by this module, the flag signal represents the flag signal issued by BRAM_MODULE, and current_state and next_state represent the current state and next state of the state machine in BRAM_MODULE, respectively. Figure 7 As shown in the diagram, after the SEU occurs and the read state is reached, the BRAM module generates a flag signal. This flag signal then passes through the comparator circuit and outputs an Error signal. The blue text in the diagram represents the signal propagation delay in the circuit. This is not limited to the logic comparison scheme described in this article.

[0156] After the entire system is set up and irradiation testing begins, the experiment will be conducted cyclically according to the following procedure: First, the local computer uses the iMAPCT tool included with the ISE to download the configuration program to the SEU detection FPGA (DUT) chip via the JTAG interface. Then, the "Open Device" and "Start" buttons are clicked sequentially in the white light neutron source single-event reversal experiment serial port assistant on the host computer. At this time, the system is in SEU testing state. In this state, the system continuously reads the raw data in the DUT BRAM and compares it with itself. When the data in the BRAM changes, the TOF calculation and communication FPGA is triggered to calculate the flight time of the neutron that caused this SEU and output it to the host computer for storage. The host computer sends a stop command. If the number of SEUs reaches the experimental requirement, the irradiation is stopped and the experiment ends. If the number of SEUs has not yet reached the requirement, the SEU detection FPGA is reconfigured, and the next SEU detection begins. The test procedure is as follows: Figure 9 As shown.

[0157] In one embodiment, the present invention is adapted to neutron soft error testing and reliability estimation in various environments.

[0158] In one embodiment, the present invention performs both thermal neutron and fast neutron tests simultaneously.

[0159] Those skilled in the art will understand that all or part of the functions of the various methods in the above embodiments can be implemented by hardware or by computer programs. When all or part of the functions in the above embodiments are implemented by computer programs, the program can be stored in a computer-readable storage medium, which may include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to achieve the above functions. For example, the program can be stored in the memory of a device, and when the program in the memory is executed by the processor, all or part of the above functions can be achieved. In addition, when all or part of the functions in the above embodiments are implemented by computer programs, the program can also be stored in a server, another computer, disk, optical disk, flash drive, or external hard drive, etc., and can be downloaded or copied to the memory of a local device, or the system of the local device can be updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments can be achieved.

[0160] The above examples illustrate the present invention only to aid in understanding it and are not intended to limit the scope of the invention. Those skilled in the art can make various simple deductions, modifications, or substitutions based on the principles of this invention.

Claims

1. A method for predicting the number of neutron single-particle flips, characterized in that, include: The steps for calculating flight time include calculating the flight time using the following formula. : In formula (1), The moment when the accelerator kicker signal enters the FPGA device to trigger time-of-flight measurement; To record the moment when a single-particle flip event occurs; The time of flight is the output after differential comparison by the ToF detection board; The timing of the gamma pulse signal detected by the gamma detector and the kicker trigger signal of the accelerator. The time interval between; The time it takes for the gamma pulse signal to travel a specific distance, where the specific distance refers to the distance from the neutron source to the detector; This refers to the time delay during the data readout process within the FPGA. The neutron energy calculation steps include the flight time obtained from the flight time calculation steps. The neutron's flight speed is calculated using the following formula. v： In formula (2), L It is the distance a neutron travels, measured in meters (m). It is the neutron flight time, measured in seconds (s). The neutron energy is then calculated using the following formula. : In formula (3), v is the neutron's flight speed, in m / s; m 0 It is the mass of a neutron, measured in kg. c It is the speed of light, measured in m / s; The steps for calculating the single-particle flip-off cross section include obtaining the energy as follows: The number of single-event flips caused by neutrons is calculated using the following formula to obtain the single-event flip cross section. : In formula (4), It is neutron energy The single-particle flip section at the location is expressed in cm. 2 / bit; It is energy The number of single-particle flips caused by neutrons; It is the neutron energy during the experiment. Neutron flux at a given location, in units of n / cm 2 ; S represents the number of bits in the device's built-in static memory, measured in bits. The single-event flip event count prediction step includes, based on the single-event flip cross section... Neutron energy spectrum of the environment to be predicted In addition to the number of bits S of the static memory within the device, the number of neutron single-event flips of the FPGA under test in the required environment is predicted. .

2. The method as described in claim 1, characterized in that, A specific number is pre-stored in the static RAM of the FPGA device and compared by the logic circuit built into the FPGA device, thereby realizing the measurement of single-event flip events in the static RAM within an ultra-short time, where the ultra-short time is <100ns.

3. The method as described in claim 1, characterized in that, In the neutron energy calculation step, the flight time starting point is provided by the accelerator timing signal. After a single-particle reaction occurs, the test conditions are reset to provide an equivalent test environment. Through repeated cyclic testing, an ultra-high total neutron fluence is provided to the FPGA device. The ultra-high total neutron fluence refers to >1E12n / cm². 2 Neutron flux.

4. The method as described in claim 1, characterized in that, In the single-particle flip-off cross-section calculation step, the energy is obtained by comparison reading method. The number of single-particle flips caused by neutrons, specifically including: The FPGA uses a built-in comparator to directly compare the values ​​in the registers to determine whether a single-event upset has occurred.

5. The method as described in claim 1, characterized in that, The environment includes the ground environment, near-Earth space, outer space environment, or artificial ion environment.

6. The method as described in claim 1, characterized in that, In the single-event upset prediction step, the number of single-event upsets is calculated using the following formula: In formula (5), It is neutron energy The single-particle flip section at the location is expressed in cm. 2 / bit; Φ ( E n The neutron energy is E n Neutron flux, in units of n / cm 2 ; S represents the number of bits in the device's built-in static memory, measured in bits.

7. The method as described in claim 1, characterized in that, The FPGA device includes an SRAM-type FPGA device.

8. The method as described in claim 1, characterized in that, The neutron source is a pulsed white light neutron source.

9. A system for predicting the number of neutron single-particle flips, characterized in that, include: The flight time calculation module is used to calculate the flight time according to the following formula. : In formula (1), The moment when the accelerator kicker signal enters the FPGA device to trigger time-of-flight measurement; To record the moment when a single-particle flip event occurs; The time of flight is the output after differential comparison by the ToF detection board; The timing of the gamma pulse signal detected by the gamma detector and the kicker trigger signal of the accelerator. The time interval between; The time it takes for the gamma pulse signal to travel a specific distance, where the specific distance refers to the distance from the neutron source to the detector; This refers to the time delay during the data readout process within the FPGA. The neutron energy calculation module is used to calculate the flight time based on the flight time calculation steps. The neutron's flight speed is calculated using the following formula. v： In formula (2), L It is the distance a neutron travels, measured in meters (m). It is the neutron flight time, measured in seconds (s). The neutron energy is then calculated using the following formula. : In formula (3), v is the neutron's flight speed, in m / s; m 0 It is the mass of a neutron, measured in kg. c It is the speed of light, measured in m / s; The single-particle flip-off section calculation module is used to obtain the energy of... The number of single-event flips caused by neutrons is calculated using the following formula to obtain the single-event flip cross section. : In formula (4), It is neutron energy The single-particle flip section at the location is expressed in cm. 2 / bit; It is energy The number of single-particle flips caused by neutrons; It is the neutron energy during the experiment. Neutron flux at a given location, in units of n / cm 2 ; S represents the number of bits in the device's built-in static memory, measured in bits. The single-event flip event count prediction module is used to predict the number of single-event flip events based on the single-event flip cross section. Neutron energy spectrum of the environment to be predicted In addition to the number of bits S of the static memory within the device, the number of neutron single-event flips of the FPGA under test in the required environment is predicted. .

Images