Estimation device, estimation method, and program
The estimation device separates SEU generation timing by shifting time-series data to accurately estimate the SEU cross-section in semiconductor devices under double-bunch neutron irradiation.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- NIPPON TELEGRAPH & TELEPHONE CORP
- Filing Date
- 2024-11-29
- Publication Date
- 2026-06-10
AI Technical Summary
In double-bunch neutron irradiation, the timing of SEU generation overlaps between low-energy and high-energy neutrons, making it impossible to accurately measure the relationship between neutron energy and the number of SEUs generated, thus preventing accurate estimation of the SEU cross-section.
An estimation device and method that measures the number of SEUs generated by shifting time-series data by a predetermined time to separate the effects of first and second radiation pulses, allowing for accurate calculation of the SEU cross-section.
Enables accurate measurement of the relationship between radiation energy and SEUs generated, enabling precise estimation of the SEU cross-section even with double-bunch radiation pulses.
Smart Images
Figure 2026094984000001_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to an estimation device, an estimation method, and a program for estimating the SEU cross-sectional area.
Background Art
[0002] When neutrons are incident on a semiconductor device, an error (SEU; Single event upset) may occur in which the bit information in the device is rewritten, leading to a malfunction of the device. The neutron SEU occurrence probability RSEU per unit time is calculated by the following equation (1).
[0003]
Equation
[0007] [Non-Patent Document 1] Energy-Resolved Soft-Error Rate Measurements for 1-800 MeV Neutrons by the Time-of-Flight Technique at LANSCE "IEEE Journals& Magazine" IEEE Xplore. [Non-Patent Document 2] "World's first measurement of the energy characteristics of neutrons that cause semiconductor soft errors," Nippon Telegraph and Telephone Corporation, et al., November 25, 2020. [Overview of the project] [Problems that the invention aims to solve]
[0008] In recent years, there has been a growing demand for tests that increase the intensity of the beam irradiated in a single pass. As a result, an increasing number of facilities are conducting neutron irradiation tests using beams with a profile that irradiates with neutrons twice in a short period of time (hereinafter referred to as "double punch"). In double bunch irradiation, the device under test is irradiated with a first bunch (first pass) and a second bunch (second pass) in a short period of time.
[0009] In a double-bunch neutron pulse, two neutrons are irradiated in a short period of time. As a result, the timing of SEU generation in the device overlaps between the low-energy neutrons (slow-speed neutrons) in the first bunch and the high-energy neutrons (fast-speed neutrons) in the second bunch. Therefore, it is not possible to correlate the timing of SEU generation with the neutron energy, making it impossible to accurately measure the relationship between neutron energy and the number of SEUs generated, and thus impossible to accurately estimate the SEU cross-section.
[0010] This disclosure has been made in view of the above circumstances, and its purpose is to provide an estimation device, estimation method, and program that can estimate the SEU cross-section with high accuracy even when using double bunch radiation pulses. [Means for solving the problem]
[0011] An estimation device according to one aspect of the present disclosure is an estimation device for estimating the SEU cross-section of a semiconductor device, comprising: a measurement unit that measures the number of SEUs generated when a group of pulses including a first radiation pulse irradiated from a radiation source and a second radiation pulse irradiated later than the first radiation pulse is irradiated to the semiconductor device and generates time-series data of the number of SEUs generated; a correction unit that shifts the time-series data by a predetermined time and generates corrected time-series data based on the time-series data before the shift and the time-series data after the shift; and a calculation unit that calculates the SEU cross-section of the semiconductor device based on the corrected time-series data.
[0012] An estimation method according to one aspect of the present disclosure is an estimation method for estimating the SEU cross-section of a semiconductor device, wherein the estimation device measures the number of SEUs generated when a pulse group including a first radiation pulse and a second radiation pulse irradiated later than the first radiation pulse is irradiated onto the semiconductor device, generates time-series data of the number of SEUs generated from this measurement result, shifts the time-series data by a predetermined time, generates corrected time-series data based on the time-series data before the shift and the time-series data after the shift, and calculates the SEU cross-section of the semiconductor device based on the corrected time-series data.
[0013] A program in one aspect of the present disclosure causes a computer to function as a measurement unit that measures the number of SEUs generated when a group of pulses, including a first radiation pulse irradiated from a radiation source and a second radiation pulse irradiated later than the first radiation pulse, is irradiated to a semiconductor device and generates time-series data of the number of SEUs generated; a correction unit that shifts the time-series data by a predetermined time and generates corrected time-series data based on the time-series data before the shift and the time-series data after the shift; and a calculation unit that calculates the SEU cross-sectional area of the semiconductor device based on the corrected time-series data. [Effects of the Invention]
[0014] According to this disclosure, even when using double-bunch radiation pulses, it becomes possible to measure the relationship between radiation energy and the number of SEUs generated with high accuracy and to estimate the SEU cross-section with high accuracy. [Brief explanation of the drawing]
[0015] [Figure 1A] Figure 1A is a schematic diagram illustrating the irradiation of a semiconductor device with a neutron pulse having a single peak. [Figure 1B] Figure 1B is a graph showing the time-series data of SEU generated in a semiconductor device when irradiated with a neutron pulse having a single peak. [Figure 2A] Figure 2A is a schematic diagram illustrating the irradiation of a semiconductor device with a neutron pulse having two peaks. [Figure 2B] Figure 2B is a graph showing the time-series data of SEUs generated in a semiconductor device when irradiated with a neutron pulse having two peaks. [Figure 3] This is a block diagram showing the configuration of the estimation device according to the embodiment. [Figure 4A] Figure 4A is a graph showing the time-series data of SEU generated in a semiconductor device when irradiated with a neutron pulse having two peaks. [Figure 4B]FIG. 4B is a graph showing time-series data obtained by shifting the data Q1 at times t0 to t3 shown in FIG. 4A by a time T1. [Figure 4C] FIG. 4C is a graph showing time-series data obtained by subtracting the curve s11 from the curve s4 shown in FIG. 4B. [Figure 4D] FIG. 4D is a graph showing time-series data obtained by shifting the curve s12 shown in FIG. 4C by a time T1. [Figure 5] FIG. 5 is a flowchart showing a processing procedure by an estimation apparatus according to an embodiment. [Figure 6] FIG. 6 is a block diagram showing a hardware configuration of the present embodiment. MODE FOR CARRYING OUT THE INVENTION
[0016] Hereinafter, embodiments will be described with reference to the drawings. First, the relationship between the neutron beam irradiated to the semiconductor device and the number of SEUs (number of errors) generated will be described with reference to FIGS. 1A, 1B, 2A, and 2B.
[0017] FIG. 1A is an explanatory diagram schematically showing a state where a neutron pulse s1 having one peak is irradiated to a semiconductor device 11 (hereinafter abbreviated as "device 11"). FIG. 1B is a graph showing time-series data of SEUs generated in the device 11 when the neutron pulse s1 is irradiated. As shown in FIG. 1A, when the device 11 is irradiated with the neutron pulse s1 having one peak, among the neutrons included in the neutron pulse s1, the high-energy neutrons have a high speed and thus reach the device 11 earlier. The low-energy neutrons have a low speed and thus reach the device 11 later.
[0018] The curve s2 shown in Figure 1B represents the time-series data of SEUs generated in device 11. The higher the neutron energy, the earlier SEUs are generated in device 11. Therefore, by referring to the time-series data shown in Figure 1B, the number of SEUs generated for each neutron energy can be measured. In other words, when irradiating device 11 with a neutron pulse s1 having one peak, the number of SEUs generated for each neutron energy in device 11 can be measured with high accuracy.
[0019] Figure 2A is a schematic diagram illustrating the irradiation of device 11 with a neutron pulse s3 having two peaks (double-bunch neutron pulse s3). Figure 2B is a graph showing the SEU generated in device 11 when irradiated with neutron pulse s2. The double-bunch neutron pulse s3 refers to a group of pulses including a first bunch p1 (first radiation pulse) irradiated from the radiation source and a second bunch p2 (second radiation pulse) irradiated later than the first bunch.
[0020] As shown in Figure 2A, the first bunch p1 and the second bunch p2 have nearly identical amplitudes. The double bunch neutron pulse s3 is delivered by the second bunch p2 after a time T1 has elapsed since the first bunch p1 was delivered. When the device 11 is irradiated with the double bunch neutron pulse s3, the SEU generated in the device 11 becomes a curve s4 with two peaks p3 and p4, as shown in Figure 2B. Furthermore, the low-energy neutrons from the first bunch p1 and the high-energy neutrons from the second bunch p2 reach the device 11 at the same time. Therefore, it is difficult to separate them.
[0021] Specifically, peak p3 shown in Figure 2B is presumed to be SEU generated by high-energy neutrons from the first bunch p1. Peak p4 is presumed to be a sum of SEU generated by low-energy neutrons from the first bunch p1 and SEU caused by high-energy neutrons from the second bunch p2. Therefore, peak p4 shown in Figure 2B is a mixture of SEU from the first bunch p1 and SEU from the second bunch p2. When irradiating device 11 with a double bunch neutron pulse s3, it is not possible to measure the number of SEUs generated for each neutron energy in device 11 with high accuracy.
[0022] In this embodiment, the curve s4 showing the number of SEU occurrences, as shown in Figure 2B, is shifted in the time axis direction by a time T1 from the occurrence time of the first bunch p1 to the occurrence time of the second bunch p2. Then, the number of SEU occurrences due to the first bunch p1 alone is calculated by subtracting the shifted curve (s11 in Figure 4B or s13 in Figure 4D, described later) from the curve s4 before the shift. That is, it is assumed that the SEUs generated by the first bunch p1 and the SEUs generated by the second bunch p2 are almost identical, and the number of SEU occurrences due to the first bunch p1 alone is calculated by performing this subtraction. A detailed explanation follows below.
[0023] Figure 3 is a block diagram showing the configuration of the estimation device 100 according to the embodiment. As shown in Figure 3, the estimation device 100 comprises a test device 1 and a calculation device 2. The estimation device 100 estimates the SEU cross-sectional area σSEU(En) of the semiconductor device 11.
[0024] The test apparatus 1 comprises a radiation source 12 and a measuring unit 13.
[0025] The radiation source 12 irradiates the device 11 under test (labeled "semiconductor device" in the figure) with radiation. The radiation is, for example, neutrons. As shown in Figure 2A above, the radiation source 12 irradiates the device 11 with a double bunch of neutron pulses (pulse group) including a first bunch p1 (first radiation pulse) and a second bunch p2 (second radiation pulse) that occurs at a time T1 later. The radiation source 12 outputs spectral information to the calculation unit 22, which will be described later. A neutron pulse is an example of a radiation pulse.
[0026] The measurement unit 13 counts the SEUs generated in the device 11 when it is irradiated with a double bunch neutron pulse from the radiation source 12, and measures the SEU generation time and the number of SEUs generated. Based on the above-mentioned SEU generation time and number of SEUs generated (measurement results), the measurement unit 13 generates SEU time-series data (hereinafter abbreviated as "time-series data") that shows the number of SEUs generated in chronological order. That is, the measurement unit 13 measures the number of SEUs generated when a group of pulses, including a first radiation pulse irradiated from the radiation source and a second radiation pulse irradiated later than the first radiation pulse, is irradiated to a semiconductor device, and generates time-series data of the number of SEUs generated.
[0027] The measurement unit 13 generates time-series data, for example, shown in curve s4 in Figure 4A. In Figure 4A, the time T1 between the two peaks p3 and p4 of curve s4 is approximately equal to the time difference between the first bunch p1 and the second bunch p2 shown in Figure 2A. Therefore, the time from the occurrence time t1 of peak p3 to the occurrence time t3 of peak p4 is indicated by the same symbol "T1" as in Figure 2A.
[0028] The computing device 2 comprises a correction unit 21 and a calculation unit 22. The computing device 2 acquires time-series data output from the measurement unit 13 when the device 1 is continuously irradiated multiple times with double-bunch neutron pulses from the radiation source 12 of the test apparatus 1. Based on this time-series data, the computing device 2 acquires data showing the relationship between the neutron energy irradiated to the device 11 and the number of SEUs generated when the device 11 is irradiated with double-bunch neutron pulses from the radiation source 12. Furthermore, it estimates the SEU cross-section σSEU(En) based on the relationship between neutron energy and the number of SEUs generated.
[0029] The correction unit 21 shifts the time-series data output from the measurement unit 13 in the time axis direction. Specifically, when the first double-bunch neutron pulse is irradiated from the radiation source 12, the correction unit 21 generates time-series data by shifting the data Q1 from the curve s4 shown in Figure 4A, from the time t0 when the SEU measurement started to the time t2 when time T1 has elapsed, by time T1. As a result, for example, the time-series data shown in the curve s11 in Figure 4B (hereinafter referred to as "shifted time-series data") is obtained. The shifted time-series data s11 has the same waveform as the data Q1 shown in Figure 4A.
[0030] The correction unit 21 subtracts the time series data after the shift from the time series data before the shift. Specifically, when the first double bunch neutron pulse irradiation yields the time series data s4 before the shift and the time series data s11 after the shift shown in Figure 4B, the correction unit 21 subtracts the time series data s11 after the shift from the time series data s4 before the shift. As a result, the time series data shown in curve s12 in Figure 4C (referred to as the "corrected time series data") is obtained. The correction unit 21 outputs the corrected time series data s12 to the calculation unit 22.
[0031] When a second or subsequent double bunch neutron pulse is irradiated, the correction unit 21 generates time series data (shifted time series data) by shifting the corrected time series data s12 (the corrected time series data calculated previously) by time T1 (see s13 in Figure 4D).
[0032] The correction unit 21 corrects the time series data before correction using the shifted time series data (for example, the time series data shown at s13 in Figure 4D). Specifically, when the measurement unit 13 provides the time series data before shifting (for example, s4 in Figure 4D), the correction unit 21 subtracts the shifted time series data (s13 in Figure 4D) from this time series data. As a result, the corrected time series data (not shown) is obtained. The correction unit 21 outputs the corrected time series data to the calculation unit 22.
[0033] In other words, the correction unit 21 shifts the time series data by a predetermined time (for example, time T1) and generates corrected time series data based on the time series data before the shift and the time series data after the shift. The predetermined time includes, for example, the delay time of the second radiation pulse (second bunch p2) relative to the first radiation pulse (first bunch p1).
[0034] Furthermore, when the radiation source 12 irradiates with a second or subsequent pulse group (double bunch neutron pulses), the correction unit 21 generates the corrected time series data for the current process using the shifted time series data obtained by shifting the corrected time series data calculated in the previous process by a predetermined time (for example, time T1).
[0035] The calculation unit 22 calculates the SEU cross-section σSEU(En) of the device 11 based on the spectral information of the radiation output from the radiation source 12 and the corrected time-series data output from the correction unit 21. In other words, the calculation unit 22 calculates the SEU cross-section of the semiconductor device based on the corrected time-series data.
[0036] Next, the processing procedure of the estimation device 100 according to this embodiment, configured as described above, will be explained with reference to the flowchart shown in Figure 5. First, the device 11 to be measured for SEU generation is placed in the test apparatus 1 shown in Figure 3. In step S11 of Figure 5, the radiation source 12 irradiates the device 11 with a double bunch neutron pulse (radiation). For example, it irradiates with a neutron pulse including the first bunch p1 and the second bunch p2 shown in Figure 2A.
[0037] In step S12, the measurement unit 13 counts the SEUs generated by the device 11 and generates time-series data indicating the number of SEUs based on the SEU generation time and the number of SEUs generated. As a result, for example, the time-series data s4 (time-series data before shift) shown in Figure 4A is generated. The measurement unit 13 outputs the generated time-series data s4 to the correction unit 21.
[0038] In step S13, the correction unit 21 shifts the time series data s4 shown in Figure 4A, specifically the data for the time period indicated by code Q1, by time T1 to generate the shifted time series data s11 (see Figure 4B).
[0039] In step S14, the correction unit 21 generates the corrected time series data s12 shown in Figure 4C by subtracting the time series data s11 after the shift from the time series data s4 before the shift. The correction unit 21 outputs the generated corrected time series data s12 to the calculation unit 22.
[0040] In step S15, the radiation source 12 irradiates the device 11 with a second double bunch neutron pulse.
[0041] In step S16, the measurement unit 13, similar to step S12 described above, counts the SEUs generated by the device 11 and generates time-series data indicating the number of SEUs based on the SEU generation time and the number of SEUs generated. As a result, for example, the time-series data s4 (time-series data before shift) shown in Figure 4A is generated. The measurement unit 13 outputs the generated time-series data s4 to the correction unit 21.
[0042] In step S17, the correction unit 21 shifts the corrected time series data generated during the previous measurement, i.e., the time series data s12 shown in Figure 4C, by time T1 to generate the shifted time series data s13 (see Figure 4D).
[0043] In step S18, the correction unit 21 generates corrected time series data (not shown) by subtracting the time series data after shifting s13 from the time series data before shifting s4. The correction unit 21 outputs the generated corrected time series data to the calculation unit 22.
[0044] In step S19, the calculation unit 22 determines whether the neutron pulse irradiation has finished. If it has not finished (S19; NO), the process returns to step S15; if it has finished, the process proceeds to step S20.
[0045] In step S20, the calculation unit 22 calculates the SEU cross-section based on the corrected time-series data output from the correction unit 21 and the spectral information of the radiation output from the radiation source 12. The procedure for calculating the SEU cross-section is a well-known technique, so its explanation is omitted. In this way, even when the radiation irradiated from the radiation source 12 is a double-bunch neutron pulse, the SEU time-series data can be calculated with high accuracy, and consequently, the SEU cross-section can be estimated with high accuracy.
[0046] As described above, the estimation device 100 according to this embodiment is an estimation device 100 for estimating the SEU cross-sectional area of a semiconductor device 11, and includes a measurement unit 13 that measures the number of SEUs generated when a group of pulses including a first radiation pulse irradiated from a radiation source and a second radiation pulse irradiated later than the first radiation pulse is irradiated to the semiconductor device 11 and generates time-series data of the number of SEUs generated; a correction unit 21 that shifts the time-series data by a predetermined time and generates corrected time-series data based on the time-series data before the shift and the time-series data after the shift; and a calculation unit 22 that calculates the SEU cross-sectional area of the semiconductor device 11 based on the corrected time-series data.
[0047] In this embodiment, even with a radiation source 12 that outputs a double-bunch neutron pulse having a first bunch and a second bunch, the influence of neutrons generated by the second bunch can be removed by shifting the time-series data measured by the measurement unit 13 and subtracting the time-series data after the shift from the time-series data before the shift. This makes it possible to improve the accuracy of calculating the SEU cross-section using the time-of-flight method.
[0048] In this embodiment, the predetermined time for which the correction unit 21 shifts the time-series data is set to the delay time between the first radiation pulse (first bunch) and the second radiation pulse (second bunch). This makes it possible to more effectively remove the effects of neutrons generated by the second bunch.
[0049] Furthermore, the correction unit 21 generates corrected time-series data by subtracting the time-series data after the shift from the time-series data before the shift, making it possible to more effectively remove the effects of neutrons generated by the second bunch.
[0050] When the radiation source 12 irradiates with a second or subsequent double bunch neutron pulse (pulse group), the correction unit 21 generates the corrected time series data for the current time using the shifted time series data obtained by shifting the corrected time series data calculated in the previous processing by a predetermined time. Therefore, in the second and subsequent times, the corrected time series can be generated with higher accuracy, and the SEU cross-section can be estimated with higher accuracy.
[0051] In this embodiment, since neutrons are irradiated from the radiation source 12, it becomes possible to estimate the SEU cross-section with higher accuracy.
[0052] In this embodiment, neutrons were used as an example of radiation, but other types of radiation are not limited to neutrons and can be used.
[0053] As shown in Figure 6, the estimation device 100 of this embodiment described above can use a general-purpose computer system that includes, for example, a CPU (Central Processing Unit, processor) 901, memory 902, storage 903 (HDD: Hard Disk Drive, SSD: Solid State Drive), communication device 904, input device 905, and output device 906. The memory 902 and storage 903 are storage devices. In this computer system, the CPU 901 executes a predetermined program loaded onto the memory 902, thereby realizing each function of the estimation device 100.
[0054] The estimation device 100 may be implemented on a single computer or on multiple computers. Furthermore, the estimation device 100 may be a virtual machine implemented on a computer.
[0055] The program for the estimation device 100 can be stored on a computer-readable recording medium such as an HDD, SSD, USB (Universal Serial Bus) memory, CD (Compact Disc), or DVD (Digital Versatile Disc), or it can be distributed via a network. A computer-readable recording medium is, for example, a non-transitory recording medium.
[0056] This disclosure is not limited to the embodiments described above, and numerous modifications are possible within the scope of its essence. [Explanation of symbols]
[0057] 1. Test apparatus 2 Computing device 11. Semiconductor devices (devices) 12 Radiation source 13 Measuring part 21 Correction section 22 Calculation section 100 Estimator p1 First Bunch p2 Second Bunch
Claims
1. An estimation device for estimating the SEU cross-sectional area of a semiconductor device, A measurement unit measures the number of SEUs generated when a group of pulses, including a first radiation pulse emitted from a radiation source and a second radiation pulse emitted later than the first radiation pulse, is irradiated onto the semiconductor device, and generates time-series data of the number of SEUs generated. A correction unit shifts the aforementioned time series data by a predetermined time and generates corrected time series data based on the time series data before the shift and the time series data after the shift. A calculation unit that calculates the SEU cross-sectional area of the semiconductor device based on the corrected time-series data, An estimation device equipped with this device.
2. The predetermined time includes the delay time of the second radiation pulse relative to the first radiation pulse. The estimation device according to claim 1.
3. The correction unit generates the corrected time series data by subtracting the time series data after the shift from the time series data before the shift. The estimation device according to claim 1 or 2.
4. When a second or subsequent pulse group is irradiated from the radiation source, the correction unit uses the shifted time series data obtained by shifting the corrected time series data calculated in the previous process by a predetermined time to generate the corrected time series data for the current process. The estimation device according to claim 3.
5. The aforementioned radiation source emits neutrons. The estimation device according to claim 1 or 2.
6. The estimation device is an estimation method for estimating the SEU cross-sectional area of a semiconductor device, The number of SEUs generated when a pulse group including a first radiation pulse and a second radiation pulse irradiated later than the first radiation pulse is irradiated onto the semiconductor device is measured, and time-series data of the number of SEUs generated is generated from this measurement result. The aforementioned time series data is shifted by a predetermined time, and corrected time series data is generated based on the time series data before the shift and the time series data after the shift. Based on the corrected time-series data, the SEU cross-section of the semiconductor device is calculated. Estimation method.
7. Computers, A measurement unit measures the number of SEUs generated when a group of pulses, including a first radiation pulse emitted from a radiation source and a second radiation pulse emitted later than the first radiation pulse, is irradiated onto a semiconductor device, and generates time-series data of the number of SEUs generated. A correction unit shifts the aforementioned time series data by a predetermined time and generates corrected time series data based on the time series data before the shift and the time series data after the shift. A calculation unit calculates the SEU cross-sectional area of the semiconductor device based on the corrected time-series data. A program that makes it function as such.