Next Article in Journal
Ultrathin Silicon Nanowires for Optical and Electrical Nitrogen Dioxide Detection
Next Article in Special Issue
Monolayer-Scale GaN/AlN Multiple Quantum Wells for High Power e-Beam Pumped UV-Emitters in the 240–270 nm Spectral Range
Previous Article in Journal
Size-Dependent Electrical Transport Properties in Conducting Diamond Nanostripes
Previous Article in Special Issue
Wet-Etched Microlens Array for 200 nm Spatial Isolation of Epitaxial Single QDs and 80 nm Broadband Enhancement of Their Quantum Light Extraction
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

High Hole Concentration and Diffusion Suppression of Heavily Mg-Doped p-GaN for Application in Enhanced-Mode GaN HEMT

1
Department of Electrophysics, National Yang Ming Chiao Tung University, Hsinchu 30010, Taiwan
2
Technology Development Division, Episil-Precision Inc., Hsinchu 30010, Taiwan
*
Author to whom correspondence should be addressed.
Nanomaterials 2021, 11(7), 1766; https://doi.org/10.3390/nano11071766
Submission received: 8 May 2021 / Revised: 2 July 2021 / Accepted: 5 July 2021 / Published: 7 July 2021

Abstract

:
The effect of Mg doping on the electrical and optical properties of the p-GaN/AlGaN structures on a Si substrate grown by metal organic chemical vapor deposition was investigated. The Hall measurement showed that the activation efficiency of the sample with a 450 sccm Cp2Mg flow rate reached a maximum value of 2.22%. No reversion of the hole concentration was observed due to the existence of stress in the designed sample structures. This is attributed to the higher Mg-to-Ga incorporation rate resulting from the restriction of self-compensation under compressive strain. In addition, by using an AlN interlayer (IL) at the interface of p-GaN/AlGaN, the activation rate can be further improved after the doping concentration reaches saturation, and the diffusion of Mg atoms can also be effectively suppressed. A high hole concentration of about 1.3 × 1018 cm−3 can be achieved in the p-GaN/AlN-IL/AlGaN structure.

1. Introduction

The AlGaN/GaN high-electron mobility transistor (HEMT) on Si has received tremendous research attention in high-power device application due to its large breakdown electric field, high electron saturation velocity, and good thermal conductivity [1]. In order to guarantee a safe operation and simplify the circuit architecture, the AlGaN/GaN HEMT is made in the enhanced mode (E-mode) configuration of normally-off operation [2]. The most common and commercial E-mode HEMT is designed in the p-GaN/AlGaN/GaN HEMT configuration. The p-GaN raises the GaN conduction band of the AlGaN/GaN HEMT above the Fermi level, leading to the depletion of the two-dimensional electron gas (2DEG) channel at zero gate bias [3]. Therefore, an E-mode p-HEMT with a higher and stable threshold voltage (Vth) is expected by increasing the hole concentration. However, Mg doping for higher hole concentrations encountered several challenges, including (1) the compensation effect of the donor due to native defects (VN) and dislocations [4,5,6], (2) low p-type activation of Mg-H into GaN [7,8], (3) self-compensation effect due to saturation Mg doping-induced donor-type defects [9,10,11], (4) the formation of pyramidal defects from Mg segregation on threading dislocation [12,13], and (5) Mg diffusion into the AlGaN barrier layer and GaN channel layer [14,15]. Although, L. Sang et al. recently showed that the hole concentration and activation efficiency of Mg-doped p-GaN grown on a free-standing GaN substrate of a low dislocation density could be improved dramatically [6]. Yingda Chen et al. discovered that the growth technique of indium surfactant-assisted delta doping could substantially enhance the hole concentration of a p-GaN/u-GaN homo-structure grown on a 2-inch c-plane sapphire to 1.5 × 1018 cm−3 [16]. However, the issue of low activation efficiency for Mg-doped p-GaN/AlGaN hetero-structures on the more economic Si substrates remains. As the Mg doping increases, the deep-level emission dominates in the photoluminescence (PL) and cathodoluminescence (CL) spectra [17,18]. This implies the formation of deeper donors to compensate holes or the creation of deeper Mg acceptor levels rather than shallow acceptor levels, further resulting in the difficulty to activate holes from the deep Mg acceptors to the valence band and decrease the activation efficiency. Therefore, it is essential to further investigate the effect of Mg doping on the electronic and optical properties to discover the optimized growth condition for better activation efficiency.
With the case of Mg diffusion into the AlGaN barrier layer and GaN channel layer, Loizos Efthymiou et al. discovered that Vth shifts firmly with Mg diffusion [19]. CL measurements revealed Mg diffusion along the dislocation [20]. Mg diffusion along the edge-type and mixed-type dislocations was also evidenced by transmission electron microscopy and atom probe tomography [21,22]. As a result, it is crucial to explore how to suppress Mg diffusion for better device performance of Mg-doped p-GaN/AlGaN/GaN HEMTs.
In the current work, the flow rate of Cp2Mg was modulated to grow Mg-doped p-GaN on AlGaN to study the effect of different Mg doping concentrations on the hole concentration and activation efficiency. The PL experiment was carried out to investigate the deep emissions and self-compensation at various doped Mg levels. In addition, Hsien-Chin Chiu et al. demonstrated that a thin AlN etch stop layer in the p-GaN/AlN/AlGaN/GaN HEMT structure can effectively improve the device RON uniformity and reduce the leakage current [23,24]. Thus, the influence of a thick GaN and thin AlN interlayer (IL) at the interface of the Mg-doped p-GaN and AlGaN layer on the activation efficiency and Mg diffusion was also investigated in this study.

2. Materials and Methods

The epitaxial structures of the Mg-doped GaN layers were grown by a metal organic chemical vapor deposition (MOCVD) system (Veeco Instruments Inc, Plainview, NY, USA) on 6-inch Si (1 1 1) substrates, as shown in Figure 1. The conventional source precursors including trimethylaluminum (TMAl), trimethylgallium (TMGa), ammonia (NH3), and bis(cyclopentadienyl) magnesium (Cp2Mg) were used to grow the AlN, AlGaN, GaN, and Mg-doped p-GaN layers. To avoid Ga-Si melt-back etching, a 200 nm AlN nucleation layer was first grown at 1030 °C on Si substrate. There are three types of samples, A, B, and C, as shown in Figure 1. All samples applied the same step-graded AlGaN buffers, consisting of a 200 nm Al0.7Ga0.3N layer, a 300 nm Al0.5Ga0.5N layer, and a 300 nm Al0.3Ga0.7N layer grown at 1020 °C to modulate stress for avoiding cracking. The sample structures were designed for high Mg activation rates and suppressing Mg diffusion into the under-layers. For sample A, the 1000 nm-thick Mg-doped p-GaN layers were grown at 990 °C with different Cp2Mg flow rates of 0, 200, 450, 600, 750, and 900 sccm, labeled as A0, A200, A450, A600, A750, and A900, respectively. For both samples B and C, the Cp2Mg flow rate was 900 sccm for investigating the effect of undoped GaN (u-GaN) and AlN-IL on the Mg activation rate and diffusion. The post-growth thermal activation of Mg-doped p-GaN was performed for 20 min at 720 °C under a nitrogen atmosphere. The secondary ion mass spectroscopy (SIMS) measurement was carried out on all samples to determine the Mg concentration in the p-GaN layer by the IMS-6f (CAMECA SAS, Gennevilliers, France). In order to investigate the electrical properties of p-GaN, the standard Hall effect with the Van der Pauw method was conducted at room temperature by the HMS-3000 (Ecopia Corporation, Anyang-City, South Korea). The optical properties of all samples were studied using low-temperature photoluminescence (PL) spectroscopy by the excitation of a HeCd laser at 325 nm. The threading dislocation density (TDD) was evaluated from the full width at half maximum (FWHM), scanned on GaN (002) and (102) planes by X-ray diffraction (XRD, X’Pert Pro MRD, Malvern Panalytical, Almelo, The Netherlands). The characterization of structure strain was performed by Raman scattering. The effect of the Mg doping concentration on the surface morphology was carried out by scanning electron microscopy (SEM, JSM7001F, JEOL, Tokyo, Japan), optical microscopy (OM, AL100, Olympus Corporation, Tokyo, Japan), and atomic force microscopy (AFM, NT-MDT Spectrum Instruments, Moscow, Russia).

3. Results and Discussion

The hole carrier concentration and activation efficiency as a function of Mg doping are revealed in Table 1 and Figure 2. As we can see, the hole concentration increases, corresponding to decreased mobility with the increasing Mg doping. Meanwhile, the resistivity decreases initially and then increases with the Mg doping. The activation efficiency (Mg doping efficiency), which is defined as the ratio of the hole concentration (obtained from Hall measurement) and Mg doping density (measured by SIMS), increases initially and reaches a maximum value of 2.22% at Mg doping of 2.42 × 1019/cm3 (450 sccm), and then it decreases with the Mg doping. This can be attributed to the Mg saturated concentration of about 2 × 1019/cm3. Furthermore, the low Mg concentration behavior presented in our samples is similar to that of other reported data [11,25] for a GaN:Mg hetero-epitaxial layer on a sapphire substrate, as shown in Figure 2. However, they all showed constant reversion of the hole concentration after Mg saturation, owing to the self-compensation effect. Even A. Klump et al. applied UV illumination to reduce H passivation and the self-compensation impact on the GaN:Mg films, which was just helpful on the concentration below the Mg saturation. In our case, when the Mg doping is more than the self-compensation onset of 2.42 × 1019/cm3 (450 sccm), it is worth noting that the activated hole concentrations still rise without a hole concentration reversion. However, the decrease in activation efficiency could be ascribed to the starting existence of high Mg doping-induced defects, for example, the formation of Mg interstitials [9,17], nitrogen vacancy VN [9,26], MgGa-VN complexes [11,27], and pyramidal inversion domain (PID) defects [28,29]. Another scenario could be the building possibility of Mg-N-Mg clusters. The rising formation probability of Mg-N-Mg double acceptors could split the acceptor level and create deeper acceptor states and further decrease the density of a single Mg shallow acceptor. The deeper acceptor states are not active in creating free holes, leading to lower activation efficiency. For even higher Mg doping concentrations, the possibility to generate Mg3N2 clusters increases [30,31]. The formation of Mg3N2 clusters decreases the single Mg concentration, and the energy states of Mg3N2 clusters are deep levels in the energy gap and do not contribute free holes. With a consistent result, we also obtained precipitation of Mg-rich and pyramid-shaped defects on our SEM and optical microscope images, respectively, after the flow rate of 450 sccm (not shown here). The energy-dispersive X-ray spectroscopy (EDS) analysis also exhibited the Mg content of the 900 sccm sample, about 2.4% on the Mg-rich precipitates and around three times that of the blank background (0.79%). In addition, AFM images show the root mean square (RMS) of the surface roughness increases from 0.49 to 1.75 nm in the 5 μm × 5 μm scan area, while the Mg flow rate increases from 0 to 900 sccm.
The PL spectra of p-GaN films at 10 K with different Mg doping concentrations are shown in Figure 3. The PL of the undoped GaN film shows a sharp near-band edge emission (NBE) at 3.46 eV (358.4 nm), as shown in Figure 3a. The broad emissions below 3.2 eV are attributed to the defect emissions from the AlGaN layers. In addition, the oscillation in the PL intensity below 3.2 eV is due to the Fabry–Perot interference of the whole sample structure. By measuring the energy separation of the two nearest peaks ΔE, the total sample thickness could be evaluated by d = hc/(2nΔE) ≈ 2 μm, where h, c, and n are the Planck constant, speed of light, and refraction index at the emission peak, respectively. When the Mg doping is turned on at 200 sccm, the native donor (VN) [4] and shallow Mg acceptor pair (DAP) emission dominates the PL spectrum, as it can be seen in Figure 3b. The peak of DAP is around 3.1 eV. As the Mg doping is further increased to 450 sccm, the peak energy of blue luminescence (BL) is near 2.8 to 3.0 eV (Figure 3c). The emission peak near 2.8–3.0 eV was attributed to the deep donor-to-shallow acceptor transition [32,33]. These deep donors could be created by the heavy Mg doping-induced defects. The emission peak near 2.8–3.2 eV could also be ascribed to the recombination of a native donor and heavy Mg doping-induced deep Mg acceptor. The PL spectra presented in Figure 3d–f, for higher Mg doping samples, are basically the same as in Figure 3c. The peak intensity of green luminescence (GL) and yellow luminescence (YL) becomes more prominent with increased Mg doping, which means that structural defects related to VN begin to increase [26,34]. In general, the collected PL data corroborate the results of electrical measurements mentioned above (Table 1 and Figure 2). As the Mg doping exceeds 450 sccm, the more Mg atoms incorporated into the GaN crystal generate not only more single Mg shallow acceptors but also more Mg-N-Mg deep acceptors, or donor-type defects, leading to a drop-off in activation efficiency. If the BL emission at 2.8–3.0 eV of Figure 3c is due to the donor-to-deep acceptor recombination, the deep acceptors are about 300 to 500 meV above the valence band compared with the activation energy of the shallow acceptor of about 200 meV [35,36]. Therefore, the deep acceptors have lower efficiency to be activated to offer free holes in the valence band for conducting. Similar competition of two emissions was also discovered recently by Hanxiao Liu et al. for their low- and high-Mg doping samples [18]. They attributed the two emissions at 3.25 eV and 2.9 eV to the shallow donor-to-acceptor and deep donor-to-acceptor transitions, respectively. We suggest that the BL near 2.9 eV can be caused by both the deep acceptors and deep donors. The deep acceptors should result from the Mg-rich and Mg3N2 precipitates to decrease the activation efficiency. The deep donors of donor-like defects from the VN and Mg-VN complexes can decrease the hole concentration by the self-compensation effect.
In order to investigate the effect of Mg diffusion on the activation efficiency, the electrical properties of samples B and C are discussed. Figure 4a shows the SIMS of samples A900, B, and C. Mg diffusion is the strongest for sample B, the p-GaN homo-epitaxy of the 200 nm GaN template. The difference in Mg diffusion for samples A900 and C is not significant. However, the hole concentration and activation efficiency are very different for three samples, as shown in Figure 4b. Suppose that the hole concentration evaluated by the Hall measurement is majorly contributed by the top part of the p-GaN layers; the similar Mg doping concentrations at the top of the p-GaN layers for all three samples imply that the self-compensation effects are different. We would like to emphasize that the activation efficiency was effectively increased by decreasing the self-compensation effect, while the decrease in Mg diffusion was trivial, as extracted from the SIMS results. The p-GaN film grown on AlN-IL (2 nm)/Al0.3Ga0.7N has the best activation efficiency of 2.2%. In the event of p-GaN grown on Al0.3Ga0.7N and GaN, the activation efficiencies are 1.4% and 0.8%, respectively. This could be due to the strain between the layers to suppress the formation of the Mg doping-induced donor-type defects. These results indicate that aluminum has a smaller atomic radius than gallium, which can inhibit Mg diffusion and increase the compressive stress on the GaN:Mg film [37]. It is expected that a high Al composition could significantly suppress the self-compensation effect, reduce the Mg diffusion concentration, and further increase the hole concentration and activation rate.
Many research groups also investigated the role of stable and metastable Mg-H complexes on the activation efficiency [7,8,25]. They discovered that the hole concentration is proportional to the density of H atoms from the Mg-H complex measured by SIMS before thermal annealing. The Mg atoms without the formation of the Mg-H complex could occupy the interstitials, Mg-VN complexes, or lattice positions of nitrogen (MgN). They are donor-type defects and play the role of self-compensation. As shown in Figure 5a, a higher H concentration was observed before annealing in p-GaN/AlGaN and p-GaN/AlN-IL structures than that in the p-GaN/GaN-IL structure. However, the p-GaN/AlN-IL structure displayed a similar H concentration to p-GaN/AlGaN, which could not expound the higher hole concentration and activation efficiency with AlN-IL. Therefore, these two structures were measured by the HRXRD rocking curves for the FWHM of GaN (002) and (102) planes to calculate the threading dislocation densities (TDDs) [38]. The GaN (002)/(102) planes of 678/1024 arcsecs without AlN-IL, respectively, correspond to the screw/edge-type TDDs of 9.24 × 108 and 3.13 × 109 cm−2. The screw/edge-type TDDs of 7.98 × 108 and 3.51 × 109 cm−2 with the AlN-IL structure were calculated by the GaN (002)/(102) planes of 630/1028 arcsecs. In contrast to the relationship, the total TDDs with AlN-IL slightly increased from 4.05 × 109 to 4.30 × 109 cm−2, indicating that the TDDs do not dominate the hole concentration in this case. We recommend excluding the effect of Mg-H and TDDs on the increasing activation efficiency after the Mg doping concentration reaches saturation. Furthermore, in Figure 5b, the PL spectra of p-GaN exhibit that the photon intensities of BL, GL, and YL decreased dramatically with AlN-IL. The lower concentration of self-compensation defects in the p-GaN on AlN-IL could be due to the greater compressive strain in p-GaN. This is consistent with our Raman spectra, where GaN E2 (High) and A1 (LO) shift from 563.46 to 563.74 cm−1 and 722.26 to 726.19 cm−1, respectively. The Raman energy blue shift implies greater compressive stress in the p-GaN epilayer with AlN-IL [39,40]. This effect is in agreement with the suppression of donor-like defects under greater compressive strains from inserting an AlN interlayer into the Mg-doped GaN/AlGaN superlattice by Hu et al. [41,42]. Herein, we would like to express that the existence of the greater compressive stress of heavily Mg-doped GaN is crucial in affecting the self-compensation effect because it can effectively extend the Fermi energy and consequently increase the formation energy of self-compensation defects [9]. This was mentioned in other research [10,43,44] which found that a strain state from compressive to tensile is accompanied by the BL emission due to large local lattice relaxations by the generation of self-compensation defects. This study reveals that a high-Al composition layer under the p-GaN layer can effectively enhance the hole concentration and significantly reduce the self-compensation effect. Furthermore, no reversion of the hole concentration could be observed after Mg saturation. This finding is precious for application in E-mode GaN HEMTs.

4. Conclusions

In this study, the flow rate of Cp2Mg was modulated to grow heavily Mg-doped p-GaN on AlGaN for application in enhanced-mode HEMTs. A maximum activation rate of 2.22% was accomplished with Mg doping of around 2.42 × 1019 cm−3. The further increase in the hole concentration with the increasing Mg concentration reveals that the hole reversion could be restrained, owing to the decreased compensation-type defects resulting from the enhanced compressive strain. In addition, a high hole concentration of 1.3 × 1018 cm−3 with a high activation efficiency was also achieved by heavy Mg doping of around 6.05 × 1019 cm−3 in the p-GaN/AlN-IL/AlGaN structure. The diffusion of Mg can be effectively suppressed by inserting an AlN layer at the interface of Mg-GaN and AlGaN. The current results provide important information for the growth of Mg-doped p-GaN of a high hole concentration in E-mode HEMT application.

Author Contributions

J.-J.D. conceived and designed the sample structure and experiments, performed the epitaxial growth, and wrote the manuscript. H.-C.H. and W.-F.W. supported the epitaxial growth. T.T.M., S.-K.W., J.-R.P., C.-W.L. and H.-C.W. supported structural, optical, and electrical characterizations. W.-C.C. is the advisor who supervised the experiments. All authors contributed to discussing the results and comments on the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Ministry of Science and Technology, Taiwan, under grant No. MOST 109-2124-M-009-002-MY3.

Data Availability Statement

Data are contained within the article.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Chen, K.J.; Haberlen, O.; Lidow, A.; Tsai, C.L.; Ueda, T.; Uemoto, Y.; Wu, Y. GaN-on-Si Power Technology: Devices and Applications. IEEE Trans. Electron. Devices 2017, 64, 779–795. [Google Scholar] [CrossRef]
  2. Chen, K.J.; Zhou, C. Enhancement-mode AlGaN/GaN HEMT and MIS-HEMT technology. Phys. Status Solidi (a) 2010, 208, 434–438. [Google Scholar] [CrossRef]
  3. Roccaforte, F.; Greco, G.; Fiorenza, P.; Iucolano, F. An Overview of Normally-Off GaN-Based High Electron Mobility Transistors. Materials 2019, 12, 1599. [Google Scholar] [CrossRef] [Green Version]
  4. Buckeridge, J.; Catlow, C.R.A.; Scanlon, D.O.; Keal, T.; Sherwood, P.; Miskufova, M.; Walsh, A.; Woodley, S.M.; Sokol, A. Determination of the Nitrogen Vacancy as a Shallow Compensating Center in GaN Doped with Divalent Metals. Phys. Rev. Lett. 2015, 114, 016405. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  5. Sarkar, B.; Mita, S.; Reddy, P.; Klump, A.; Kaess, F.; Tweedie, J.; Bryan, I.; Bryan, Z.; Kirste, R.; Kohn, E.; et al. High free carrier concentration in p-GaN grown on AlN substrates. Appl. Phys. Lett. 2017, 111, 032109. [Google Scholar] [CrossRef]
  6. Sang, L.; Ren, B.; Endo, R.; Masuda, T.; Yasufuku, H.; Liao, M.; Nabatame, T.; Sumiya, M.; Koide, Y. Boosting the doping efficiency of Mg in p-GaN grown on the free-standing GaN substrates. Appl. Phys. Lett. 2019, 115, 172103. [Google Scholar] [CrossRef]
  7. Castiglia, A.; Carlin, J.-F.; Grandjean, N.R. Role of stable and metastable Mg–H complexes in p-type GaN for cw blue laser diodes. Appl. Phys. Lett. 2011, 98, 213505. [Google Scholar] [CrossRef]
  8. Wide range doping control and defect characterization of GaN layers with various Mg concentrations. J. Appl. Phys. 2018, 124, 165706. [CrossRef]
  9. Miceli, G.; Pasquarello, A. Self-compensation due to point defects in Mg-doped GaN. Phys. Rev. B 2016, 93, 93. [Google Scholar] [CrossRef] [Green Version]
  10. Kirste, R.; Hoffmann, M.P.; Tweedie, J.; Bryan, Z.; Callsen, G.; Kure, T.; Nenstiel, C.; Wagner, M.R.; Collazo, R.; Hoffmann, A.; et al. Compensation effects in GaN:Mg probed by Raman spectroscopy and photoluminescence measurements. J. Appl. Phys. 2013, 113, 103504. [Google Scholar] [CrossRef]
  11. Klump, A.; Hoffmann, M.P.; Kaess, F.; Tweedie, J.; Reddy, P.; Kirste, R.; Sitar, Z.; Collazo, R. Control of passivation and compensation in Mg-doped GaN by defect quasi Fermi level control. J. Appl. Phys. 2020, 127, 045702. [Google Scholar] [CrossRef]
  12. Figge, S.; Kröger, R.; Böttcher, T.; Ryder, P.L.; Hommel, D. Magnesium segregation and the formation of pyramidal defects in p-GaN. Appl. Phys. Lett. 2002, 81, 4748–4750. [Google Scholar] [CrossRef]
  13. Duguay, S.; Echeverri, A.; Castro, C.; Latry, O. Evidence of Mg Segregation to Threading Dislocation in Normally-Off GaN-HEMT. IEEE Trans. Nanotechnol. 2019, 18, 995–998. [Google Scholar] [CrossRef]
  14. Posthuma, N.; You, S.; Liang, H.; Ronchi, N.; Kang, X.; Wellekens, D.; Saripalli, Y.; Decoutere, S. Impact of Mg out-diffusion and activation on the p-GaN gate HEMT device performance. In Proceedings of the 2016 28th International Symposium on Power Semiconductor Devices and ICs (ISPSD), Prague, Czech Republic, 12–16 June 2016; pp. 95–98. [Google Scholar]
  15. He, J.; Tang, G.; Chen, K.J. VTH Instability of p-GaN Gate HEMTs under Static and Dynamic Gate Stress. IEEE Electron. Device Lett. 2018, 39, 1. [Google Scholar] [CrossRef]
  16. Chen, Y.; Wu, H.; Yue, G.; Chen, Z.; Zheng, Z.; Wu, Z.; Wang, G.; Jiang, H. Enhanced Mg Doping Efficiency in P-Type GaN by Indium-Surfactant-Assisted Delta Doping Method. Appl. Phys. Express 2013, 6, 041001. [Google Scholar] [CrossRef]
  17. Nayak, S.; Gupta, M.; Waghmare, U.V.; Shivaprasad, S. Origin of Blue Luminescence in Mg -Doped GaN. Phys. Rev. Appl. 2019, 11, 014027. [Google Scholar] [CrossRef] [Green Version]
  18. Liu, H.; Fu, H.; Fu, K.; Alugubelli, S.R.; Su, P.-Y.; Zhao, Y.; Ponce, F.A. Non-uniform Mg distribution in GaN epilayers grown on mesa structures for applications in GaN power electronics. Appl. Phys. Lett. 2019, 114, 082102. [Google Scholar] [CrossRef] [Green Version]
  19. Efthymiou, L.; Murukesan, K.; Longobardi, G.; Udrea, F.; Shibib, A.; Terrill, K. Understanding the Threshold Voltage Instability During OFF-State Stress in p-GaN HEMTs. IEEE Electron. Device Lett. 2019, 40, 1253–1256. [Google Scholar] [CrossRef]
  20. Chen, J.; Yi, W.; Kimura, T.; Takashima, S.; Edo, M.; Sekiguchi, T. Cathodoluminescene study of Mg implanted GaN: The impact of dislocation on Mg diffusion. Appl. Phys. Express 2019, 12, 051010. [Google Scholar] [CrossRef]
  21. Usami, S.; Mayama, N.; Toda, K.; Tanaka, A.; Deki, M.; Nitta, S.; Honda, Y.; Amano, H. Direct evidence of Mg diffusion through threading mixed dislocations in GaN p–n diodes and its effect on reverse leakage current. Appl. Phys. Lett. 2019, 114, 232105. [Google Scholar] [CrossRef]
  22. Yi, W.; Kumar, A.; Uzuhashi, J.; Kimura, T.; Tanaka, R.; Takashima, S.; Edo, M.; Yao, Y.; Ishikawa, Y.; Chen, J.; et al. Mg diffusion and activation along threading dislocations in GaN. Appl. Phys. Lett. 2020, 116, 242103. [Google Scholar] [CrossRef]
  23. Chiu, H.-C.; Chang, Y.-S.; Li, B.-H.; Wang, H.-C.; Kao, H.-L.; Hu, C.-W.; Xuan, R. High-Performance Normally Off p-GaN Gate HEMT With Composite AlN/Al0.17Ga0.83N/Al0.3Ga0.7N Barrier Layers Design. IEEE J. Electron. Devices Soc. 2018, 6, 201–206. [Google Scholar] [CrossRef]
  24. Chiu, H.-C.; Chang, Y.-S.; Li, B.-H.; Wang, H.-C.; Kao, H.-L.; Chien, F.-T.; Hu, C.-W.; Xuan, R. High Uniformity Normally-OFF p-GaN Gate HEMT Using Self-Terminated Digital Etching Technique. IEEE Trans. Electron. Devices 2018, 65, 4820–4825. [Google Scholar] [CrossRef]
  25. Qi, H.; Zhang, S.; Liu, S.; Liang, F.; Yi, L.; Huang, J.; Zhou, M.; He, Z.; Zhao, D.; Jiang, D. The self-compensation effect of heavily Mg doped p-GaN films studied by SIMS and photoluminescence. Superlattices Microstruct. 2019, 133, 106177. [Google Scholar] [CrossRef]
  26. Yan, Q.; Janotti, A.; Scheffler, M.; Van De Walle, C.G. Role of nitrogen vacancies in the luminescence of Mg-doped GaN. Appl. Phys. Lett. 2012, 100, 142110. [Google Scholar] [CrossRef]
  27. Reshchikov, M.A.; Demchenko, D.O.; McNamara, J.D.; Fernández-Garrido, S.; Calarco, R. Green luminescence in Mg-doped GaN. Phys. Rev. B 2014, 90, 90. [Google Scholar] [CrossRef] [Green Version]
  28. Remmele, T.; Albrecht, M.; Irmscher, K.; Fornari, R.; Straßburg, M. Pyramidal inversion domain boundaries revisited. Appl. Phys. Lett. 2011, 99, 141913. [Google Scholar] [CrossRef]
  29. Iwata, K.; Narita, T.; Nagao, M.; Tomita, K.; Kataoka, K.; Kachi, T.; Ikarashi, N. Atomic resolution structural analysis of magnesium segregation at a pyramidal inversion domain in a GaN epitaxial layer. Appl. Phys. Express 2019, 12, 031004. [Google Scholar] [CrossRef]
  30. Hansen, M.; Chen, L.F.; Lim, S.H.; DenBaars, S.P.; Speck, J.S. Mg-rich precipitates in the p-type doping of InGaN-based laser diodes. Appl. Phys. Lett. 2002, 80, 2469–2471. [Google Scholar] [CrossRef]
  31. Vennéguès, P.; Leroux, M.; Dalmasso, S.; Benaissa, M.; De Mierry, P.; Lorenzini, P.; Damilano, B.; Beaumont, B.; Massies, J.; Gibart, P. Atomic structure of pyramidal defects in Mg-doped GaN. Phys. Rev. B 2003, 68, 68. [Google Scholar] [CrossRef]
  32. Reshchikov, M.A.; Yi, G.-C.; Wessels, B.W. Behavior of 2.8- and 3.2-eV photoluminescence bands in Mg-doped GaN at different temperatures and excitation densities. Phys. Rev. B 1999, 59, 13176–13183. [Google Scholar] [CrossRef]
  33. Kaufmann, U.; Kunzer, M.; Maier, M.; Obloh, H.; Ramakrishnan, A.; Santic, B.; Schlotter, P. Nature of the 2.8 eV photoluminescence band in Mg doped GaN. Appl. Phys. Lett. 1998, 72, 1326–1328. [Google Scholar] [CrossRef]
  34. Kojima, K.; Takashima, S.; Edo, M.; Ueno, K.; Shimizu, M.; Takahashi, T.; Ishibashi, S.; Uedono, A.; Chichibu, S. Nitrogen vacancies as a common element of the green luminescence and nonradiative recombination centers in Mg-implanted GaN layers formed on a GaN substrate. Appl. Phys. Express 2017, 10, 061002. [Google Scholar] [CrossRef]
  35. Oh, E.; Park, H.; Park, Y. Excitation density dependence of photoluminescence in GaN:Mg. Appl. Phys. Lett. 1998, 72, 70–72. [Google Scholar] [CrossRef]
  36. Viswanath, A.K.; Shin, E.-J.; Lee, J.I.; Yu, S.; Kim, N.; Kim, B.; Choi, Y.; Hong, C.-H. Magnesium acceptor levels in GaN studied by photoluminescence. J. Appl. Phys. 1998, 83, 2272–2275. [Google Scholar] [CrossRef]
  37. Lin, P.-J.; Tien, C.-H.; Wang, T.-Y.; Chen, C.-L.; Ou, S.-L.; Chung, B.-C.; Wuu, D.-S. On the Role of AlN Insertion Layer in Stress Control of GaN on 150-mm Si (111) Substrate. Crystals 2017, 7, 134. [Google Scholar] [CrossRef] [Green Version]
  38. Dai, J.-J.; Liu, C.-W.; Wu, S.-K.; Huynh, S.-H.; Jiang, J.-G.; Yen, S.-A.; Mai, T.T.; Wen, H.-C.; Chou, W.-C.; Hu, C.-W.; et al. Improving Transport Properties of GaN-Based HEMT on Si (111) by Controlling SiH4 Flow Rate of the SiNx Nano-Mask. Coatings 2020, 11, 16. [Google Scholar] [CrossRef]
  39. Flynn, C.; Lee, W. The dependence of Raman scattering on Mg concentration in Mg-doped GaN grown by MBE. Mater. Res. Express 2014, 1, 25901. [Google Scholar] [CrossRef]
  40. Kisielowski, C.; Krüger, J.; Ruvimov, S.; Suski, T.; Ager III, J.; Jones, E.; Liliental-Weber, Z.; Rubin, M.; Weber, E.; Bremser, M. Strain-related phenomena in GaN thin films. Phys. Rev. B 1996, 54, 17745. [Google Scholar] [CrossRef] [Green Version]
  41. Ning-Yang, L.; Lei, L.; Lei, W.; Wei, Y.; Ding, L.; Lei, L.; Wen-Yu, C.; Ci-Mang, L.; Cheng-Hao, W.; Wei-Hua, C. Improvement of doping efficiency in Mg-Al0. 14Ga0. 86N/GaN superlattices with AlN interlayer by suppressing donor-like defects. Chin. Phys. B 2012, 21, 117304. [Google Scholar]
  42. Wang, L.; Li, R.; Li, D.; Liu, N.; Liu, L.; Chen, W.; Wang, C.; Yang, Z.; Hu, X. Strain modulation-enhanced Mg acceptor activation efficiency of Al0.14Ga0.86N/GaN superlattices with AlN interlayer. Appl. Phys. Lett. 2010, 96, 061110. [Google Scholar] [CrossRef]
  43. Lyons, J.L.; Janotti, A.; Van De Walle, C.G. Shallow versus Deep Nature of Mg Acceptors in Nitride Semiconductors. Phys. Rev. Lett. 2012, 108, 156403. [Google Scholar] [CrossRef] [Green Version]
  44. Reddy, P.; Bryan, Z.; Bryan, I.; Kim, J.H.; Washiyama, S.; Kirste, R.; Mita, S.; Tweedie, J.; Irving, D.L.; Sitar, Z.; et al. Pinning of energy transitions of defects, complexes, and surface states in AlGaN alloys. Appl. Phys. Lett. 2020, 116, 032102. [Google Scholar] [CrossRef]
Figure 1. Schematic structures of p-GaN samples: (a) sample A with varying Cp2Mg flow rates (0, 200, 450, 600, 750, 900 sccm), (b) sample B with Cp2Mg 900 sccm and an additional 200 nm undoped GaN interlayer, and (c) sample C with Cp2Mg 900 sccm and a 2nm AlN interlayer.
Figure 1. Schematic structures of p-GaN samples: (a) sample A with varying Cp2Mg flow rates (0, 200, 450, 600, 750, 900 sccm), (b) sample B with Cp2Mg 900 sccm and an additional 200 nm undoped GaN interlayer, and (c) sample C with Cp2Mg 900 sccm and a 2nm AlN interlayer.
Nanomaterials 11 01766 g001
Figure 2. The hole concentration (red solid squares) and activation efficiency (open squares) of p-GaN:Mg layer as a function of the Mg doping concentration. The results from references [11,25] are also plotted for comparison, as shown by blue circles, green circles, and orange triangles.
Figure 2. The hole concentration (red solid squares) and activation efficiency (open squares) of p-GaN:Mg layer as a function of the Mg doping concentration. The results from references [11,25] are also plotted for comparison, as shown by blue circles, green circles, and orange triangles.
Nanomaterials 11 01766 g002
Figure 3. The low-temperature (10 K) PL spectra of p-GaN/AlGaN structures with varying Cp2Mg flow rates.
Figure 3. The low-temperature (10 K) PL spectra of p-GaN/AlGaN structures with varying Cp2Mg flow rates.
Nanomaterials 11 01766 g003
Figure 4. (a) Mg concentration profile measured by SIMS for three different structures. (b) The hole concentration (red squares) and activation efficiency (black squares) of Mg-doped p-GaN layers grown on GaN-IL, AlGaN, and AlN-IL under-layers.
Figure 4. (a) Mg concentration profile measured by SIMS for three different structures. (b) The hole concentration (red squares) and activation efficiency (black squares) of Mg-doped p-GaN layers grown on GaN-IL, AlGaN, and AlN-IL under-layers.
Nanomaterials 11 01766 g004
Figure 5. (a) The dependence of hydrogen concentration on depth investigated by SIMS. Black, red, and blue lines are results for p-GaN on GaN-IL, Al0.3Ga0.7N, and AlN-IL structures, respectively. (b) The low-temperature (10 K) PL spectra of p-GaN with and without AlN-IL.
Figure 5. (a) The dependence of hydrogen concentration on depth investigated by SIMS. Black, red, and blue lines are results for p-GaN on GaN-IL, Al0.3Ga0.7N, and AlN-IL structures, respectively. (b) The low-temperature (10 K) PL spectra of p-GaN with and without AlN-IL.
Nanomaterials 11 01766 g005
Table 1. Dependence of activation efficiency and electrical properties on the Cp2Mg flow rates.
Table 1. Dependence of activation efficiency and electrical properties on the Cp2Mg flow rates.
StructureCp2Mg Source
(sccm)
Mg Doping Concentration
(cm−3)
Hole Concentration
(cm−3)
Mobility
(cm2/V-s)
Resistivity
(ohm-cm)
Activation Efficiency
(%)
p-GaN/Al0.3Ga0.7N2001.20 × 1019(1.25 ± 0.06) × 101727.54 ± 1.381.87 ± 0.091.04
4502.42 × 1019(5.38 ± 0.27) × 10177.69 ± 0.381.51 ± 0.082.22
6003.75 × 1019(7.49 ± 0.37) × 10175.63 ± 0.281.48 ± 0.072.00
7504.87 × 1019(7.95 ± 0.40) × 10174.51 ± 0.231.74 ± 0.091.63
9006.05 × 1019(8.71 ± 0.44) × 10173.54 ± 0.182.02 ± 0.101.44
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Dai, J.-J.; Mai, T.T.; Wu, S.-K.; Peng, J.-R.; Liu, C.-W.; Wen, H.-C.; Chou, W.-C.; Ho, H.-C.; Wang, W.-F. High Hole Concentration and Diffusion Suppression of Heavily Mg-Doped p-GaN for Application in Enhanced-Mode GaN HEMT. Nanomaterials 2021, 11, 1766. https://doi.org/10.3390/nano11071766

AMA Style

Dai J-J, Mai TT, Wu S-K, Peng J-R, Liu C-W, Wen H-C, Chou W-C, Ho H-C, Wang W-F. High Hole Concentration and Diffusion Suppression of Heavily Mg-Doped p-GaN for Application in Enhanced-Mode GaN HEMT. Nanomaterials. 2021; 11(7):1766. https://doi.org/10.3390/nano11071766

Chicago/Turabian Style

Dai, Jin-Ji, Thi Thu Mai, Ssu-Kuan Wu, Jing-Rong Peng, Cheng-Wei Liu, Hua-Chiang Wen, Wu-Ching Chou, Han-Chieh Ho, and Wei-Fan Wang. 2021. "High Hole Concentration and Diffusion Suppression of Heavily Mg-Doped p-GaN for Application in Enhanced-Mode GaN HEMT" Nanomaterials 11, no. 7: 1766. https://doi.org/10.3390/nano11071766

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop