4.4 Defect Candidates

Although a lot of effort has been put into studying point defects in /Si material systems, their exact microscopic picture is still controversial. To provide chemical information about the defect, electronic spin resonanz (ESR) measurements appeal to be the most promising method. The ESR, also frequently referred to as electronic paramagnetic resonanz (EPR), is a well-established method to study materials showing unpaired electrons. When an atomic structure is exposed to a magnetic field, the energy levels of the paired electrons are splitted proportional to the magnetic field. An unsaturated electron, however, can move between the two discrete energy levels by absorbing a photon, provided by microwaves typically in the range of . The absorbed energy can be measured and further analyzed [30, 31, 32, 33]. Theoretical information on the atomic structure of possible defect configurations in the insulator is obtained from ab initio atomistic simulations using density functional theory (DFT).

The most prominent candidate for defects in is the so-called oxygen vacancy, referred to as centers. By using ESR, centers have been identified as hole traps in pMOSFETs [31, 34, 35]. Of particular interest in our context is the suggestion that there are possibly two stable atomic configurations of the center, namely the dimer and the so-called puckered configuration [36]. In a defect-free crystalline structure each silicon atom is bonded to four neighboring oxygen atoms. A missing oxygen atom is compensated by a covalent bond between two silicon atoms. This configuration is referred to as the dimer configuration and shown in Figure 4.5 (state 1).

In the neutral dimer configuration the covalent bond can be weakened by capturing a hole. Furthermore, the positions of the two silicon atoms are slightly displaced relative to each other from their equilibrium positions, see Figure 4.5 (state 2’), and can further transit to the puckered configuration. The so called puckered configuration has the silicon atom bonded with an oxygen atom located at the back of the defect, see Figure 4.5 (state 2). The broken bond of the other silicon atom is neutralized by the remaining electron and both structures are dislocated from their equilibrium positions, thus leading to a charged puckered configuration. The detailed electronic structure behind the defects is calculated using DFT and shown in Figure 4.6 for the oxygen vacancy [37].

Some recent studies have suggested defects involving hydrogen bonds as defect candidates [40, 41]. In thermally grown insulators a considerable amount of process-related interstitial hydrogen is available, supporting the presence of such defects for instance in the hydrogen bridge configuration. The configuration of the hydrogen bridge is obtained by introducing a hydrogen atom as the connecting link between the two silicon atoms of the center configuration, see DFT calculations in Figure 4.7.

Another promising candidate for border traps is the hydroxyl center. Recent investigations have shown a good agreement between experimental data and DFT calculations of this particular defect structure [MWC16]. The four states of the hydroxyl configuration are shown in Figure 4.8.

In addition to border traps, interface states, commonly associated with centers, play a considerable role in MOSFET devices [42, 31, 34]. There are several types of centers available at /Si interfaces depending on the crystal orientation of the interface. In (111) orientated /Si interfaces only one type centers is available whereas in (100) orientated /Si interfaces two configurations, know as and centers, are present [43]. All types of centers are of an amphoteric nature. Their density of states distributions comprise two disjunct peaks, one in the lower half and one in the upper half of the band gap [44]. Depending on their trap level an interface state can be considered a donor-like trap in case of , or an acceptor-like trap when . The possible charge states are given by

and

for donor-like and acceptor-like traps, respectively.