Browse Prior Art Database

Single-Event Rate Calculation Technique

IP.com Disclosure Number: IPCOM000120250D
Original Publication Date: 1991-Apr-01
Included in the Prior Art Database: 2005-Apr-02
Document File: 2 page(s) / 80K

Publishing Venue

IBM

Related People

Scott, TM: AUTHOR

Abstract

This article describes a technique for calculating the rate of single events (transients, latchups, upsets) in an integrated circuit, caused by a heavy ion radiation environment, based solely on measured test results.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Single-Event Rate Calculation Technique

      This article describes a technique for calculating the
rate of single events (transients, latchups, upsets) in an integrated
circuit, caused by a heavy ion radiation environment, based solely on
measured test results.

      Determining the rate, or probability, of a device
malfunctioning due to an upset, a transient pulse or going into
latchup has a direct impact on a system's reliability in an
environment that contains radiation.

      Four steps are required in this technique:  (1) device
characterization using a heavy ion accelerator; (2) definition of a
function that relates the device sensitivity as a function of Linear
Energy Transfer (LET) and angle; (3) redistribution of the flux that
is within the environment of concern as a function of angular
dependence; and (4) integrating the device sensitivity with this
redistributed flux.

      The focal point of the Single Event Rate (SER) method is on the
accurate characterization of the angular dependencies.  The result of
testing at a heavy ion accelerator yields the sensitive surface area
as a function of the incident ions Linear Energy Transfer
characteristics (cross-section vs. LET).  It has been common practice
to increase the angle of beam incidence to increase, effectively, the
LET, creating what has been termed LETeffective.  The physical
mechanism behind this is that most device geometries have their
longest dimension of the charge absorption region parallel to the
surface of the die, which makes the LET of the ion appear to be
greater at higher angles of incidence.  This has been helpful in
testing the device since the beam does not have to be changed every
time a higher LET is needed.  Just by increasing the angle of
incidence the LET appears to increase by the 1/cosine of the angle,
measured form normal incidence.  This new LET created by multiplying
the true LET by 1/cosine is called the LETeffective.

      With the data taken, not only is the cross-section foun...