Predication of Heavy Ion Single Event Upset Threshold and Cross Section Based on Low Energy Proton Test Data in Nanometer Devices

In order to ensure the reliability of heavy ion single event effect experimental data and the accuracy of space on-orbit single event error rate prediction, the research was conducted on the method of predicting the equivalent silicon layer thickness, heavy ion single event upset (SEU) threshold and cross section based on the test data of low energy proton single event effect in nanometer device. This method can be available without the need for vertical cutting of device and conducting heavy ion single event effect test. Firstly, low energy proton single event effect testing in nanometer devices was performed at the EN tandem accelerator in the Institute of Heavy Ion Physics at Peking University, acquiring low energy proton SEU cross section peak. Further, by extracting of the upper and lower limit energies of low energy proton SEU cross section peak in nanometer devices, the equivalent silicon layer thickness above the sensitive volume of the device can be accurately obtained by subtracting the thickness of device sensitive volume from the range in silicon of the lower limit energy. Energy distributions reaching the sensitive volume were calculated after different proton energies passing through the equivalent silicon layer thickness, thereby obtaining the corresponding average proton energies and linear energy transfer (LET) values within the sensitive volume. Using the respective LET values within the sensitive volume of the upper limit energy and the proton energy which are higher than and closest to the upper limit, single event upset LET threshold in nanometer devices was further determined. On the basis of LET threshold, the proportion of the effective proton number which refers to protons which can induce SEU through direct ionization within the sensitive volume were acquired, thus SEU cross section can be corrected to get the accurate SEU cross section at low LET value. The predication accuracy of equivalent silicon thickness was verified based on the calculation results comparison of heavy ion effective LET values in the sensitive volume using actual metallization layers and equivalent silicon thickness in 65 nm SRAM. The predication accuracy of heavy ion SEU threshold and cross section was verified on the basis of actual heavy ion single event effect experimental data. Further, a simplified prediction method based on the remaining range was put forward. The aforementioned methods offer an effective technical means to ensure the reliability of heavy ion single event effect experimental data and prediction results of on-orbit single event error rate.