Trap filling and recombination
Space Charge Effects
Interaction of Radiation Generated Reactant Species
Introduction:Since the early 1990ies [ENL91] it is known that the intensity of radiation effects occurring in bipolar technology components depend on the dose rate being present during the exposure. The ELDRS effect becomes an important issue for electronic component testing when the irradiation field under operational conditions is a low dose rate field, as it is the case for space applications. Testing under the real time conditions, i.e. testing with a low dose-rate, is an extremely time consuming and expensive task. The conventional approach of simply applying a total dose representative for a mission within a relatively short amount of time cannot be considered to give a good result. This is because of the less pronounced degradation due to the ELDRS effect at high dose rates. Test parts might pass high dose rate testing and still fail when operated on the mission.
The ELDRS effect is known to be an issue especially in bipolar Integrated Circuits and is depending on bias conditions and dose rate. It is more pronounced in bipolar Integrated Circuits than in discrete transistors where the enhancement effect is by far less distinct [JOH94].
The basic physical effects that lead to ELDRS are not fully understood yet. In the last 15 years several attempts have been done to understand the underlying physical processes and elucidate the ELDRS effect. Explanations can be due to a competition between trap filling and recombination [BOC06], space charge effects [RAS02], or the interaction of two reactant species generated by the radiation [FRE98].
Trap filling and recombination:Boch et al suggest a physical model that focuses on competing effects of trapping and recombination on the radiation-induced carriers [BOC06]. The effect is explained as a competition between the probability for a hole to be trapped on one site and the probability for a hole to recombine with a trapped electron on another site.
Usually, electrons that survive initial recombination (after irradiation induced generation) are swept out of the oxide without recombining. Nevertheless it has been shown, that field dependent bulk trapping of the electron might occur that can retain electrons in the bulk for several seconds [BOC06]. Recombination of holes with trapped electrons occurs with the effect that the available number of holes, that are the origin of the radiation damage mechanism, is reduced. In bipolar junction transistors the important mechanism for gain degradation is due to interface trap formation at the Si/SiO2 interface at emitter-base junction. It is generally believed that the formation of interface traps is due radiation induced holes that liberate Hydrogen within the oxide. This hydrogen is transported to the Si/SiO2 interface where it interacts with a hydrogen-passivated interface defect. The moving hydrogen atom and the hydrogen atom that is passivating the interface defect form H2 and thus are leaving an activated interface defect behind. As a consequence the number of holes that remain in the oxide due to the incident radiation is ruling the number of interface defects. The approach by Boch aims to determine this number of holes as a function of the dose rate since it serves as a first order approximation to BJT degradation.
Space Charge Effects:Rashkeev focuses in his work on the dose rate dependent build-up of the interface trap density at Si/SiO2 interface [RAS02]. In lateral pnp structures the presence of interface traps is the dominating factor leading to a reduction of the transistor gain. The oxide charge actually moderates the degradation in lateral PNP devices rather decreasing the current gain. However, in npn structures both the interface traps and the oxide trapped charge decrease the current gain.
The dose rate dependence of the interface trap formation is related to proton reactions at the Si/SiO2 interface. At high dose rates interface trap formation is reduced due to hole related positive charge build-up at the Si/SiO2 interfacial region. Since the mobility of the holes is considerably higher than the one of protons liberated in the bulk of the oxide, holes can form an electrostatic barrier that reduces the probability of the protons to reach the interface. At low dose rates however, the charge buildup process is reduced relative to charge transport process. Thus both holes and protons can reach the interface which explains the enhanced interface trap formation. The results of their analysis strongly reinforce the idea that space-charge limited transport of both holes and proton is critical to an understanding of ELDRS effect in bipolar-base oxides. The very big difference between the effective mobility of these two positively charged species explains the big sensitivity of the interface-trap formation to the dose rate [RAS02].
Interaction of Radiation Generated Reactant Species:Freitag and Brown [FRE98] explain enhanced low dose rate sensitivity with a model that assumes that the defect responsible for the degradation results form the interaction of two reactant species generated by the radiation. By considering the generation, transport, and interaction of these two reactant species a model can be constructed that is consistent with observed time dependence of the degradation process. Their model characterizes the interaction of these two species using binary reaction rate theory.
The model is based on the observations gathered from experiments with operational amplifiers monitoring the input bias current. Implicit in this model is that the radiation generates a defect located at or near the Si/SiO2 interface on top of the emitter-base junction of the input transistor (in their case: a substrate PNP transistor). It is this defect that is responsible for the increase in the input bias current. The model assumes that this defect is generated by the interaction of two radiation-induced species which are called "A" and "B" and that the rate of the defect formation (as measured by delta IB) can be described by binary reaction rate theory.
Hereby, species "A" is exclusively generated by radiation. The second is reactant "B" is pre-existing and the reaction rate is, at early times, governed by the amount of the first reactant. A key assumption of their model is that there is a secondary source of "B" that is generated during the irradiation at a location some distance from the critical Si/SiO2 interface. The amount produced by this secondary source is slowly mitigating towards the critical Si/SiO2 interface and contributes to the rate with together with species "A". This results in an increased reaction rate. It is the secondary source of the second reactant "B" that is mainly responsible for the enhanced low dose rate sensitivity.