The behaviour of the fractional glow technique as well as of two new analysing methods using the entire information of two glow curves recorded with different heating rates is numerically investigated and compared with relevant conventional methods for the thermionic exoemission model recently proposed by Sakurai et al. The common application of these methods together with a complete frequency factor analysis allows to get more insight into the nature of the thermally stimulated relaxation processes: three of four model parameters can be directly found here from the experiment. In this way the ambiguity of a fit of a single glow curve without additional kinetic check can be avoided. Two general experimental criteria are given to verify the model. Fur das unlangst von Sakurai et al. vorgeschlagene Modell der thermionischen Exoemission wird das Verhalten der fraktionierten Glowtechnik sowie von zwei neuen Analysenmethoden, die die gesamte Information von zwei mit verschiedenen Heizraten aufgenommenen Glowkuren auswerten, numerisch untersucht und mit dem relevanter konventioneller Methoden verglichen. Die gemeinsame Anwendung dieser Methoden zusammen mit einer vollstandigen Frequenzfaktoranalyse erlaubt tiefere Einblicke in die Natur thermisch stimulierter Relaxationsprozesse: Drei von vier Modellparametern konnen direkt aus dem Experiment erhalten werden. Damit kann die Uneindeutigkeit der Anpassung einer einzelnen Glowkurve ohne zusatzliche Uberprufung der Kinetik vermieden werden. Es werden zwei allgemeine experimentelle Kriterien zur Verifizierung des Modells angegeben. ') Universitatsplatz 03, D-18051 Rostock, Federal Republic of Germany. ' ) Note that due to historical reasons s and fhave interchanged their meaning in comparison to SMT. 508 with G. RUDLOF and H. GLAEFEKE
E kT(t')where h, is the number of filled traps at time t = to and temperature T = T, and k the Boltzmann constant.Using this model for linear heating according T(t) = ?(tto), T being the constant heating rate, SMT succeeded in fitting their experimental curves and estimated physically reasonable model parameters [l to 41. Now it is a well establishedbut often ignoredfact in the literature that the fit of a single glow curve is not an unambiguous procedure, i.e. in general it can be done by a variety of models [5 to 171. Thus a successful computer fit is a necessary, but not sufficient, condition for the correctness of the chosen physical model. The experimentalist who wants to get reliable results, which allow a physical interpretation, has to prove the assumptions on which the analysis is based.The correctness of the model can be verified, for example, in the simplest case by fitting or predicting a second glow curve measured under identical excitation but at an appropriate, different heating rate, with the same parameter set, in other words, by checking the invariance of model parameters against a variation of external kinetic conditions [8, 11, 121. In our opinion the more promising way to analyse thermally stimulated relaxation processes...