Majority and Minority Charge Carrier Traps in n-type 4H-SiC Studied by Junction Spectroscopy Techniques

In this review, we provide an overview of the most common majority and minority charge carrier traps in n-type 4H-SiC material. We focus on the results obtained by different applications of junction spectroscopy techniques. The basic principles behind the most common junction spectroscopy techniques are given. These techniques, namely, deep level transient spectroscopy (DLTS), Laplace DLTS (L-DLTS) and minority carrier transient spectroscopy (MCTS) have led to recent progress in identifying and better understanding of the charge carrier traps in n-type 4H-SiC material.


Introduction
Today, 4H-SiC material is regarded as one of the most promising materials for electronic devices. It has the largest bandgap among all SiC polytypes, and due to the high and isotropic mobility of charge carriers, it is preferred as a material for power electronics [1], bipolar devices [2] and quantum sensing [3]. Moreover, it has found applications in radiation detection too [1,[4][5].
Even wider applicability of 4H-SiC was mostly hindered by electrically active deep level defects present in the material. Electrically active defects introduce energy levels into the bandgap which act as traps for charge carriers. The majority charge carrier traps in ntype 4H-SiC material, have been sistematicaly investigated for many years. On the other hand, minority charge carrier traps in n-type 4H-SiC are still much less investigated [6][7][8][9].
Junction spectroscopy technique is a term describing measurements performed on a semiconductor junction using electrical or electro-optical techniques [10]. The role of the junction is to create a depletion region, as its usage brings the important advantage over other bulk techniques. The advantage is that it is much easier to manipulate the occupancy of defects producing energy levels in the bandgap within the depletion region than in bulk [10]. The most common junction spectroscopy technique is the deep level transient spectroscopy (DLTS). DLTS is a very sensitive method for determination of electronic properties of electrically active defects in semiconductors, as it can detect defects in concentrations around 10 10 cm -3 [10]. It provides information regarding the activation energy for electron and hole emission, capture cross-section, and concentration of defects. However, the main problem associated with DLTS is the lack of energy resolution i.e., it is almost impossible to resolve two closely spaced deep energy levels. The improvement came in another junction spectroscopy technique and brought an order of magnitude better energy resolution. This technique is called Laplace DLTS (L-DLTS) [11].
While DLTS is mostly used for studying the electrically active defects associated with majority charge carrier traps, minority charge carrier traps are much less investigated. The basic principles of minority transient spectroscopy techniques were described by Hamilton et al. [12], and later by Brunwin et al. [13]. achieved by means of junction spectroscopy techniques. Moreover, the basic principles behind the junction spectroscopy techniques responsible for these advances, are provided.

Schottky barrier diode
As mentioned above, we need a junction to apply junction spectroscopy techniques for studying the electrically active defects in semiconductors. One of the basic junctions is the Schottky barrier diode (SBD). Figure 1a shows an energy band diagram of the n-type SBD, while in Figure 1b a schematic cross-section of a typical n-type 4H-SiC SBD is given. SBD is formed when a metal of a work function em and a n-type semiconductor of electron affinity eχn and work function es < em are joined. The different positions of the Fermi levels in the isolated materials cause a diffusion of electrons from the semiconductor to the metal, leaving behind uncompensated donor ions in a depletion space charge region [5]. In thermodynamic equilibrium, the energy band diagram of the SBD is constructed from the requirements of constant Fermi level and continuous vacuum level, as shown in Figure 1a. In the ideal case, the Schottky barrier from the metal to the semiconductor is given by Bn = m -χn, while the built-in potential from the semiconductor to the metal is given by Vbi = m -s.
An externally applied bias will disturb the fine balance of electron transport across the junction and allow a net current to flow through the device. The junction is forward or reversely biased if a positive (VF) or a negative voltage (VR) is applied to the metal with respect to the semiconductor, respectively. The total potential across the junction is reduced to Vbi-VF in forward bias and increased to Vbi+VR in reverse bias.
Typical n-type 4H-SiC SBD is produced on nitrogen-doped 4H-SiC epitaxial layer. The epitaxial layers are grown on thick silicon carbide substrate. The Schottky barrier are formed by thermal evaporation of nickel through a metal mask with openings, while Ohmic contacts are formed on the backside of the silicon carbide substrate by nickel sintering.

Electrically active defects
Electrically active defects introduce energy levels into the bandgap of a semiconductor, that act as traps for charge carriers. We can distinguish between majority and minority charge carrier traps. Majority carrier traps in semiconductors have been successfully and extensively studied by DLTS for decades [10][11], while minority charge carrier traps are much less studied. In principle, it is possible to investigate these traps using DLTS by applying forward bias [14], but the more reliable results can be obtained by using the MCTS. The main principles of DLTS and MCTS will be described next. More detailed information about these techniques is given elsewhere [10][11].

DLTS
The basic DLTS measurement consists of the repetitive filling and emptying of deep levels (ET) in the depletion region of the SBD by a bias pulse, as shown in Figure 2. The n-type SBD is operated under reverse bias VR (Figure 2a), which is reduced to VP during a bias pulse (Figure 2b). Empty traps, residing in the former depletion region, will be able to capture free carriers and become occupied (Figure 2b). After restoring to the original bias VR, the charge in the depletion region will be lower than before, due to the trapped charge carriers (Figure 2c). These carriers will be released again through thermal emission, which proceeds exponentially in time. This thermal discharging of the occupied traps is monitored by measuring the capacitance of the reverse biased diode as a function of time after the filling pulse (Figure 2d).

MCTS
The basic principle of MCTS measurement is slightly different as minority charge carriers are optically generated by use of above-bandgap light [10]. The MCTS measurement consists of the repetitive filling and emptying of deep levels (ET) by optical pulses with an energy just above the bandgap energy (Eg), as shown in Figure 2. Optical excitation can be applied through a semi-transparent Schottky contact, or from the back. If the sample thickness is greater than minority carrier diffusion length, than the sample should be thinned [10].  [15]. Here h is the energy of the optical excitation, Jdrift and Jdiff are drift and diffusion current densities, tp is the optical pulse duration, ep is the hole emission rate, and cp is the hole capture rate. Figure 4 shows typical DLTS spectrum for the as-grown n-type 4H-SiC SBD. One peak, labelled as Z1/2 is observed. The estimated activation energy for electron emission is EC-0.67 eV. Son et al. [17] were the first to ascribe Z1/2 to (=/0) transition from the carbon vacancy (VC). Carbon vacancy (Vc) is the most studied defect in n-type 4H-SiC. It is introduced during the crystal growth and upon irradiation [18].

Electrically active defects in n-type 4H-SiC
As we can see in Figure 4., DLTS peak is rather asymmetric. Hemmingsson et al. [19] showed that Z1/2 is the superposition of two almost identical Z1 and Z2 transitions, which cannot be resolved by DLTS. This problem (too closely spaced deep energy levels) is an ideal case for the application of another junction spectroscopy technique, L-DLTS. L-DLTS is an isothermal technique in which the capacitance transient (measurements are following the same principle as described in Figure 2) is averaged at a fixed temperature. It provides a spectral plot of a processed capacitance signal against emission rate rather than against temperature. More details on L-DLTS could be find elsewhere [11].
Using the L-DLTS measurements, Bathen et al. [3] have shown that S1 (in protonirradiated n-type 4H-SiC sample) has two emission lines arising from Vsi sitting at -k and -h lattice sites. This results has been later confirmed by Capan et al. [22] when studying the neutron-irradiated n-type 4H-SiC SBD.
Two deep levels with identical activation energies for electron emission of Ec -0.40 and Ec -0.70 eV have been observed in the low-energy electron irradiated n-type 4H-SiC [25][26]. These defects are labbeled as EH1 and EH3 and have been identifed as carbon intestitial-related (Ci) defects [26].
In addition to above mentioned majority charge carrier traps (Z1/2, S1/2, EH1/3), we should not forget other majority carrier traps, which are usually present in the as-grown n-type 4H-SiC material or they are introduced by radiation or additional annealing. The most common traps are EH4/5 and EH6/7. They are assigned to carbon antisite-carbon vacancy (CAV) complex [27,28] and (0/++) transition of the Vc [17], respectively.
The evidence that EH6/7 consists of two components was provided by Alfieri and Kimoto [29]. In thier work, they have resolved two energy levels at Ec -1.30 and Ec -1.49 eV, for EH6 and EH7, respectively, using the L-DLTS measurements.
In the following text, we will focus on the minority charge carrier traps in n-type 4H-SiC. Majority of the published works are related to boron. Boron is introduced in SiC intentionaly for p-type doping, or unintentionaly during the crystal growth. The unintentional boron incorporation was recently explained by the presence of boron in the graphite susceptor used for the CVD growth [30][31].  The B and D-centre have been reported in numerous studies and have been assigned to substitutional boron atoms occupying the Si and C-site, respectively [30][31][32][33]. The shallow state, BSi is off-center substitutional boron at Si-site, while for the deep state Bc, boron occupies a perfect substitutional C site [34]. Similar to Z1/2, S1/2, and EH6/7 further improvements in studying the D-centre by L-DLTS measurements have been made. Recent results have showed that D-center has two components labelled as D1 and D2 (Figure 8). Their activation energies for hole emissions are estimated as Ev + 0.49 eV and Ev + 0.57 eV and they are assigned to an isolated boron sitting at the C site, -h and -k site, respectively [33]. Figure 8. L-MCTS spectrum for D-center in semi-transparent n-type 4H-SiC SBD measured at 300 K. Data adapted from Ref [15].
At the end of this section, we summirize the relevant information on the above discussed majority and minority charge carrier traps (Table 1). Table 1. Details on the majority and minority charge carrier traps in n-type 4H-SiC material.

Conclusions
In this review paper, the basic principles behind the most used junction spectroscopy techniques, DLTS and MCTS, are given. We have provided examples from different studies on majority and minority charge carrier traps. The successful applications of junction spectroscopy techniques, which have led to better understanding of the charge carrier traps in n-type 4H-SiC material are highlighted.