Enhancing Deep Ultraviolet LED Performance via Locally Modulated NPN‑AlGaN Current‑Spreading Layers
Abstract
We present a novel strategy to improve current spreading in AlGaN‑based deep ultraviolet light‑emitting diodes (DUV LEDs) by inserting a locally doped n‑AlGaN/p‑AlGaN/n‑AlGaN (NPN‑AlGaN) structure into the electron‑supplier layer. The thin p‑AlGaN insertion creates a conduction‑band barrier that redistributes lateral current in the p‑type hole‑supplier layer, mitigating current crowding. Key parameters—Mg doping level, p‑layer thickness, AlN composition, and the number of NPN junctions—are shown to critically influence the spreading effect. Optimized NPN‑AlGaN layers enhance optical output, external quantum efficiency (EQE), and wall‑plug efficiency (WPE) for DUV LEDs.
Introduction
AlGaN‑based DUV LEDs are essential for disinfection, water purification, medical applications, and high‑density optical recording[1‑8]. Recent advances in crystal quality—such as graphene‑assisted quasi‑Van der Waals epitaxy of AlN on nano‑patterned sapphire—have lowered dislocation densities and raised internal quantum efficiencies (IQE) to 80%[9‑10]. However, DUV LEDs are driven electrically, where carrier injection and transport govern performance[11‑13]. A prominent challenge is current crowding, especially under high bias, due to poor Mg‑doping efficiency in high‑Al‑content p‑AlGaN layers[15‑16] and the lateral injection geometry of flip‑chip structures[17‑18]. Current‑spreading schemes based on narrow‑strip p‑contacts[19] or ITO/ZGO layers[20] have shown limited success, often hampered by interface resistivity or insufficient lateral conductivity.
In this work, we shift the focus from the p‑side to the n‑type electron‑supplier layer. By engineering the doping profile in n‑AlGaN to form NPN‑AlGaN junctions, we generate controlled energy barriers that redistribute lateral current, thereby enhancing hole injection and overall device efficiency. This approach offers greater flexibility for epitaxial optimization while directly addressing current crowding.
Research Methods and Physics Models
The baseline DUV LED structure comprises a 4‑µm‑thick n‑Al0.60Ga0.40N layer doped with Si at 5×1018 cm−3, followed by five Al0.45Ga0.55N/Al0.56Ga0.44N MQWs (3 nm/12 nm). An 18‑nm Mg‑doped p‑Al0.60Ga0.40N electron‑blocking layer (EBL) caps the MQWs, succeeded by 50‑nm Mg‑doped p‑Al0.40Ga0.60N and p‑GaN layers, each with hole concentrations of 3×1017 cm−3. Devices feature a 350 × 350 µm2 mesa.
Figure 1b illustrates the conduction‑band profile for a structure containing two NPN‑AlGaN junctions (NPNPN‑AlGaN). The depletion of the p‑AlGaN insertions creates vertical barriers (ϕ1, ϕ2) that tune lateral current distribution. The junction must be fully depleted to operate in reach‑through breakdown mode[21]. All devices are compared to a reference LED lacking NPN junctions.
We model current flow with an equivalent circuit (Fig. 2a), separating vertical (I1) and horizontal (I2) components. The presence of NPN junctions introduces interfacial resistance Rnpn per junction, reducing the I1/I2 ratio (Eq. 3). Increasing the number of junctions (N) or raising Rnpn—by adjusting AlN content, thickness, or Mg doping—further promotes lateral spreading.
Device physics are simulated using Crosslight APSYS, with validated models for Auger recombination (1×10−30 cm6s−1), SRH lifetime (10 ns), light extraction (≈ 8 %), and a 50:50 band‑offset for AlGaN heterojunctions[25]. Polarization charges are set at 40 % of the lattice mismatch[29].
Results and Discussions
Impact of NPN‑AlGaN on Current Spreading
Comparing reference LED A (no NPN) with LED B (one NPN junction, 20 nm p‑Al0.60Ga0.40N, Mg = 1×1018 cm−3) shows two conduction barriers in LED B (Fig. 3a). These barriers increase Rnpn, lowering I1/I2 and enabling a more uniform lateral hole distribution in the last quantum well (LQW) (Fig. 3b). Consequently, hole concentrations and radiative recombination rates in the MQWs rise (Fig. 4a‑b), translating to higher EQE and optical power density (Fig. 5a). Although forward voltage increases slightly (Fig. 5b), the wall‑plug efficiency (WPE) surpasses the reference above ~56 A cm−2 (Fig. 5c), confirming the benefit of NPN‑AlGaN engineering.
Effect of AlN Composition in the p‑AlGaN Layer
Five devices (Ci, i = 1‑5) were fabricated with AlN fractions of 0.60, 0.63, 0.66, 0.69, 0.72 in the 20‑nm p‑layer (Mg = 1.8×1018 cm−3). Higher AlN increases the conduction‑band barrier (Table 1), raising Rnpn and reducing I1/I2 (Eq. 3). Lateral hole profiles (Fig. 6a) and MQW concentrations (Fig. 6b‑c) show progressively improved uniformity and higher carrier densities with larger AlN content. EQE and optical power density follow suit (Fig. 7a), while forward voltage exhibits a modest rise (Fig. 7b). WPE benefits are maximized at an intermediate AlN fraction (C2), balancing barrier height against added voltage drop (Fig. 8).
Mg Doping Influence
Devices Di (i = 1‑5) vary Mg doping in the 20‑nm p‑layer (3×1017 – 3×1018 cm−3) with AlN = 0.61. Increasing Mg widens the depletion region, raising the barrier and Rnpn (Table 2). Lateral hole distributions become increasingly uniform (Fig. 9a), enhancing MQW hole concentrations and radiative recombination (Fig. 9b‑c). EQE and optical power density rise with doping (Fig. 10a), though forward voltage also increases, particularly for the highest Mg level (Fig. 10b). WPE improves up to Mg = 2×1018 cm−3 but declines at the highest doping due to excessive voltage drop (Fig. 11).
Thickness of the p‑AlGaN Insertion Layer
Series Ti (i = 1‑5) employ 18‑, 20‑, 24‑, 28‑, and 32‑nm p‑layers (AlN = 0.61, Mg = 1.5×1018 cm−3). Barrier heights increase with thickness (Table 3), further reducing I1/I2 and improving lateral spreading (Fig. 12a). MQW hole concentrations and recombination rates rise correspondingly (Fig. 12b‑c). Optical power and EQE improve (Fig. 13a), and forward voltage decreases significantly for 20‑, 24‑, and 28‑nm devices (Fig. 13b). However, the thickest layer (32 nm) shows a higher turn‑on voltage due to a parasitic N‑AlGaN/P‑AlGaN diode. WPE peaks at an optimal thickness, demonstrating the need for careful tuning (Fig. 14).
Number of NPN‑AlGaN Junctions
LEDs Ni (i = 1‑5) contain 1 to 5 NPN junctions, each with 20‑nm p‑layers (AlN = 0.61, Mg = 1.5×1018 cm−3). While individual barrier heights are similar (Table 4), the cumulative barrier increases with junction count, amplifying lateral spreading (Fig. 15a). MQW hole concentrations and recombination rates improve progressively (Fig. 15b‑c). EQE and optical power density rise with junction number, but the gains diminish after three junctions (Fig. 16a). Forward voltage decreases relative to the reference (Fig. 16b), and WPE benefits are evident up to five junctions, though the incremental advantage tapers (Fig. 17).
Conclusions
Embedding NPN‑AlGaN junctions within the n‑type electron‑supplier layer effectively mitigates current crowding in the p‑type hole‑supplier region of AlGaN‑based DUV LEDs. By tuning AlN composition, Mg doping, insertion‑layer thickness, and junction count, we can homogenize lateral current, enhance hole injection, and improve EQE and WPE. While EQE consistently benefits, WPE gains plateau when additional voltage drops from barriers outweigh the spreading advantage, underscoring the importance of balanced design. This study introduces a versatile, growth‑friendly current‑spreading strategy that advances the performance of high‑power DUV LEDs.
Nanomaterials
- Optimized Al₂O₃/MgO Emission Layers for Ultra‑High Gain Electron Multipliers
- Ultra‑Efficient AlGaN Deep‑UV LEDs with Superlattice p‑Electron Blocking Layer for Near‑Zero Efficiency Droop
- Enhanced Current Spreading in AlGaN‑Based Deep‑UV LEDs via PNP‑AlGaN Structures
- Optimizing Quantum‑Well Width for Peak Electroluminescence in AlGaN Deep‑UV LEDs Across Temperatures
- Enhancing Light Extraction in Deep‑UV Flip‑Chip LEDs Using Nanometer‑Scale Meshed Contacts and Inclined AlGaN Nanocones
- Phase Engineering Boosts Efficiency of Quasi‑2D All‑Inorganic Perovskite LEDs via Cs Cation Ratio Control
- Silica‑Nanoparticle‑Enhanced PDMS Encapsulation Significantly Improves Light‑Output and Thermal Performance of AlGaN Deep‑UV LEDs
- Boosting AlGaN Deep UV LED Efficiency with Chirped Superlattice Electron Deceleration Layers
- Enhancing InGaN/GaN μLED Efficiency by Thinning Quantum Barriers to Reduce Surface Recombination
- Optimizing Sapphire Light‑Guide Layer Thickness to Boost UV‑LED Light Extraction Efficiency