Silicon particle detectors in the next generation of experiments at the CERN Large Hadron Collider will be exposed to a very challenging radiation environment. The principal obstacle to long-term operation arises from changes in detector doping concentration (N/sub eff/), which lead to an increase in the bias required to deplete the detector and hence achieve efficient charge collection. We have previously presented a model of interdefect charge exchange between closely spaced centers in the dense terminal clusters formed by hadron irradiation. This manifestly non-Shockley-Read-Hall (SRH) mechanism leads to a marked increase in carrier generation rate and negative space charge over the SRH prediction. There is currently much interest in the subject of cryogenic detector operation as a means of improving radiation hardness. Our motivation, however, is primarily to investigate our model further by testing its predictions over a range of temperatures. We present measurements of spectra from /sup 241/Am alpha particles and 1064-nm laser pulses as a function of bias between 120 and 290 K. Values of N/sub eff/ and substrate type are extracted from the spectra and compared with the model. The model is implemented in both a commercial finite-element device simulator (ISE-TCAD) and a purpose-built simulation of interdefect charge exchange. Deviations from the model are explored and comments made as to possible future directions for investigation of this difficult problem.