Rpp Method

This CREME96/HUP module calculates the single-event-effect (SEE) rate due to direct ionization using the Rectangular Parallelpiped (RPP) model, in which the bit sensitive volume is assumed to have this shape. In order to apply this method, you must specify the sensitive volume dimensions (x,y,z) in microns.

• If the SEE cross-section is specified via the Weibull or cross-section table options in HUP, the Integral RPP Method, in which the SEE rate is calculated from a numerical integration over the LET-dependent cross-section, is used.
• If the Critical Charge option is selected, the SEE rate is calculated with the simple (non-integral) RPP method, in which the cross-section is treated as a step function in LET.

The "z" dimension is generally the smallest RPP dimension. Specifically, z is defined as the thickness of the device along the beam during normally-incident irradiation in accelerator SEE cross-section measurements. SEE rate calculations can be particularly sensitive to the value of "z", since "z" provides the distance scale in converting from effective LET to deposited charge.

In some cases, RPP dimensions can be obtained from the chip designers and manufacturers.

The RPP dimensions can also be determined from careful analysis of accelerator ground-test data, provided that the chip's SEE response has been measured over a sufficiently broad range of angles and effective LET values. See [PETERSEN1993], [CONNELL1995] - Ed Smith has recently developed a simplified version of HICUP, implemented in EXCEL, which simultaneously determines RPP dimensions and Weibull cross-section parameters from accelerator data. Contact edwardsmith@cox.net for further details.

If no other information on RPP dimensions is available, the following rules of thumb may provide reasonable estimates:

• If the accelerator SEE cross-section measurements show a limiting value, assume the sensitive volume has a square surface area and set x = y = square root of the limiting cross-section per bit. - If you set x = 0 and y=0 in the RPP dimension input fields, CREME96/HUP will automatically reset them to the square root of the limiting cross-section. - Alternatively, if you put zero in the limiting cross-section field under the Weibull and Critical Charge options in HUP, the limiting cross-section will be reset to "xy", the product of the RPP lateral dimensions. - If you set x=y=0 and limiting cross-section=0, HUP will return an SEE rate of zero.
• The assumed "z" value should be consistent with the technology. Petersen et al. (1992) offer the following advice:

"Typical charge collection depths are 0.5 microns for SOS and SOI, and 2 microns for CMOS/epi. As technology develops SOI may approach 0.1 microns. For current technology it runs from 0.2 to 0.5 microns. One micron has ordinarily been used for CMOS or bipolar devices; however this appears to be overly conservative. It may be more appropriate to use a value like four microns. For bulk, a general rule would be to use the distance between the top of the depletion region and where the lightly doped epi transitions to a heavily depleted substrate. If the sensitive volume is a well, use ½ of the distance between the bottom of the (sensitive volume) depletion region and the top of the well depletion region."

• In calculating SEE rates, one is always well-advised to explore how the results change with different values for "z".

The CREME96 HUP routine, which calculates SEUs due to direct ionization, allows for inclusion of funnels. To do so, simply enter the funnel length (which must be specified in microns) in the appropriate input field. By default, the funnel length is set to zero, which suppresses funnels.

For further information on the RPP method, see [PETERSEN1992]