Abstract:Inefficient Mg-induced p-type doping has been remained a major obstacle in the development of GaN-based electronic devices for solid-state lighting and power applications. This study reports comparative structural analysis of defects in GaN layers on freestanding GaN substrates where Mg incorporation is carried out via two approaches: ion implantation and epitaxial doping. Scanning transmission electron microscopy revealed the existence of pyramidal and line defects only in Mg-implanted sample whereas Mg-doped… Show more
“…[12][13][14] Attempts have been made to understand the formation of such defects and their atomic structures in Mg-doped GaN layers, where Mg is incorporated during GaN growth via metalorganic chemical vapor deposition (MOCVD) or molecular beam epitaxy (MBE). 12, [15][16][17][18] However, versatile designing and processing of GaN-based modern devices requires selective-area doping to achieve precise local control over p-type conduction. In a typical GaN-based power device, the high-doped p-type regions are needed as contacts and hole transport layers, while the low-doped regions are needed for inversion channels.…”
Efficient acceptor activation in gallium nitride (GaN) achieved through Mg ion-implantation depends mainly on the concentration of implanted Mg ions and the post-implantation annealing process. In this study, we conducted correlative scanning transmission electron microscopy, atom probe tomography, and cathodoluminescence (CL) measurements on Mg-implanted GaN layers with the implanted concentration ranging from 1 × 10 17 cm −3 to 1 × 10 19 cm −3 . It was found that at the implanted concentration of ∼1 × 10 18 cm −3 , Mg atoms were randomly distributed with defects likely to be vacancy clusters whereas at the implanted concentration of ∼1 × 10 19 cm −3 , Mg-enriched clusters and dislocation loops were formed. From the CL measurements, the donor-acceptor pair (DAP) emissions from the implanted and un-implanted regions are obtained and then compared to analyze Mg activation in these regions. In the sample with Mg ∼1 × 10 19 cm −3 , the existence of Mg-enriched clusters and dislocations in the implanted region leads to a weaker DAP emission, whereas the absence of Mg-enriched clusters and dislocations in the sample with Mg ∼1 × 10 18 cm −3 resulted in a relatively stronger DAP emission.
“…[12][13][14] Attempts have been made to understand the formation of such defects and their atomic structures in Mg-doped GaN layers, where Mg is incorporated during GaN growth via metalorganic chemical vapor deposition (MOCVD) or molecular beam epitaxy (MBE). 12, [15][16][17][18] However, versatile designing and processing of GaN-based modern devices requires selective-area doping to achieve precise local control over p-type conduction. In a typical GaN-based power device, the high-doped p-type regions are needed as contacts and hole transport layers, while the low-doped regions are needed for inversion channels.…”
Efficient acceptor activation in gallium nitride (GaN) achieved through Mg ion-implantation depends mainly on the concentration of implanted Mg ions and the post-implantation annealing process. In this study, we conducted correlative scanning transmission electron microscopy, atom probe tomography, and cathodoluminescence (CL) measurements on Mg-implanted GaN layers with the implanted concentration ranging from 1 × 10 17 cm −3 to 1 × 10 19 cm −3 . It was found that at the implanted concentration of ∼1 × 10 18 cm −3 , Mg atoms were randomly distributed with defects likely to be vacancy clusters whereas at the implanted concentration of ∼1 × 10 19 cm −3 , Mg-enriched clusters and dislocation loops were formed. From the CL measurements, the donor-acceptor pair (DAP) emissions from the implanted and un-implanted regions are obtained and then compared to analyze Mg activation in these regions. In the sample with Mg ∼1 × 10 19 cm −3 , the existence of Mg-enriched clusters and dislocations in the implanted region leads to a weaker DAP emission, whereas the absence of Mg-enriched clusters and dislocations in the sample with Mg ∼1 × 10 18 cm −3 resulted in a relatively stronger DAP emission.
“…[17][18][19][20][21] However, the thermal treatment also induces Mg diffusion and segregation. 22,23 In order to achieve reasonable p-type conductivity with free hole concentrations up to 10 18 cm À3 , a large dosage of Mg implantation is required. However, when the Mg concentration is around 10 19 cm À3 or more, the p-type conductivity is unexpectedly suppressed.…”
mentioning
confidence: 99%
“…The depth scale is shown by the left axis. In the Mg-imp layer, both DAP and NBE emissions are weak because nonradiative centers such as vacancy-type defects and Mg-clusters are formed by the Mg implantation of 10 19 cm À3 and subsequent annealing at 1300 C. 23,25 Under the Mg-imp layer, a bright zone exists in the DAP image due to less implantation damage and few nonradiative defects. 29 The high intensity of DAP emission suggests that the ratio of Mg activated as acceptors to nonradiative defects is high in this Mg-activation layer (Mg-act), where the depth is around 700 nm.…”
The precise control of p-GaN is a crucial issue for developing GaN-based power devices. Mg as an acceptor is commonly used in p-type doping; however, the Mg diffusion through threading dislocations (TDs) has not been well addressed. To clarify the Mg diffusion and activation along TDs, we have performed a systematic characterization of a Mg-implanted homoepitaxial GaN layer grown on a freestanding substrate. Active-Mg related donor-acceptor pair (DAP) emission from certain TDs is identified by cathodoluminescence (CL). Dislocations with and without DAP emission are investigated structurally and compositionally based on etch pits, transmission electron microscopy, and atom-probe tomography. Direct evidence of Mg distribution around edge-and mixed-type TDs is obtained. There exists a significant difference in the Mg concentration and incorporation states between different types of TDs.
“…The critical concentration (degeneracy) of the Mott phase transition of GaN semiconductors is reported in the literature. [ 47,48 ] Although GaN materials have various types of shallow donor impurities and are extremely complex, their average effects can be described by a hydrogen‐like model. The radius of electrons in the impurity center was .…”
Section: Resultsmentioning
confidence: 99%
“…The Bohr radius of the hydrogen atom was 0.53 × 10 −10 m. The effective mass of the electron m was 0.22. When the electron wave function radius was about 4a, that is, the carrier concentration was n ≈ 1/(4a) 3 ≈ 10 18 cm −3 , [ 47,48 ] the conductance transitioned to semimetallization. In this study, the bulk concentrations of the Ga 34 MgN 36 , Ga 34 MgH i N 36 , Ga 34 BeN 36 , Ga 34 BeH i N 36 , Ga 34 CaN 36 , Ga 34 CaH i N 36 , Ga 35 BeN 36 , Ga 35 MgN 36 , and Ga 35 CaN 36 doping systems were 2.33, 3.50, 2.34, 3.51, 2.30, 3.44, 1.17, 1.16, and 1.15 × 10 21 cm −3 , respectively, in the G → F direction.…”
Obtaining a reliable positive‐type (p‐type) GaN semiconductor is difficult because of the unipolarity of GaN. This difficulty is one of the bottlenecks restricting the development of GaN‐based optoelectronic devices. To address this problem, this paper adopted the method of generalized gradient approximation (GGA) plane wave ultrasoft pseudopotential based on the framework of density functional theory to construct Ga35MN36, Ga34MN36, and Ga34MHiN36 (M = Be/Mg/Ca; Hi = interstitial hydrogen) models. Ga35MN35 and Ga35MHiN35 (M = Be/Mg/Ca) models were also constructed. Results of our calculations indicated that the Ga35MN35 and Ga35MHiN35 (M = Be/Mg/Ca) models cannot achieve a p‐type doping system. Furthermore, the formation energy of Ga34MN36 and Ga34MHiN36 (M = Be/Mg/Ca) systems was greater under Ga‐rich conditions than that under N‐rich conditions, indicating that both doping systems more readily formed and had a more stable structure under N‐rich conditions. Moreover, the formation energy of Ga34MHiN36 (M = Be/Mg/Ca) system was lower than that of Ga34MN36 (M = Be/Mg/Ca) system, and the existence of interstitial H proved to be beneficial to the improvement in system stability. The Ga34CaHiN36 system had the largest hole mobility and the best conductivity. Therefore, the Ga34CaHiN36 system is an ideal material for the application of conductive GaN devices. This study provides guidance into the preparation of p‐type conductive GaN materials.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.