Appendix D - The Relationship Between tR
and
D.1 - Introduction
As mentioned in Chapter 5, the expression for calculating the intrinsic switching time given in
equation (5.4.3.4) is a fairly simple estimate of tR. The results presented in Chapter 7 show that this
equation is quite reasonable as an engineering estimate. However, it would be more useful if it established the
relationship between tR and the carrier lifetime τ. Regrettably, this
relationship is far too complex to be dealt with analytically. This appendix presents the results of several
computer simulations, and shows that the switching time varies weakly with carrier lifetime.
Intuitively, one would expect that the switching time would decrease with lower carrier
lifetimes, since this would concentrate charge at the edges of the middle region, which is the goal of most
conventional SRD designs.
D.2 - Computer Simulations
Several simulations were run for the structure described in Section 5.6 in the series-connected
circuit configuration described in Section 7.3, with different values for . Since τ >>
τn + τp at high injection levels (see equation
(2.3)), the actual simulation parameters τn and τp were each set at /2. The forward bias was adjusted in each case such that the storage time
tS was approximately 5 ns in all cases. In other words, each device had a similar amount of stored
charge. The results of the simulations are summarized in Figure D.1.

Evidently a 50:1 change in lifetime only induces a 1.8:1 change in switching time. As an
experimental check on this relationship, one can compare the results obtained with the two WFSRDs in section 7.3.
Diode A8.PT.850.1 had an effective lifetime about five times lower than that of diode A8.6. This 5:1 lifetime
difference corresponded to a 1.13:1 difference in switching times, which agrees well with the simulated results.
(The other commercial diodes had higher switching times despite their lower carrier lifetimes; evidently these
commercial structure were further from the optimum structure than diode 89.)
Based on these results, equation (5.4.3.4) can be applied with the expectation that the result
will be accurate within a factor 2.
