Temperature Dependent Characteristics of Quasi-vertical AlN Schottky Diodes on Bulk AlN Substrate

This paper reports the successful fabrication and comprehensive temperature-dependent characterization of MOCVD-grown quasi-vertical AlN Schottky barrier diodes on bulk AlN substrates, demonstrating high current densities, stable rectification up to 300°C, and identifying Poole-Frenkel emission and an interfacial AlNxOy layer as key factors governing their carrier transport and leakage mechanisms.

The Gist

Imagine you are trying to build a super-efficient traffic system for electricity. Most of our current electronics use "roads" (semiconductors) that are like busy city streets; they work fine, but they get clogged with heat and traffic jams when the demand gets too high.

This paper is about building a super-highway for electricity using a material called Aluminum Nitride (AlN). Think of AlN as a futuristic, ultra-durable road that can handle extreme heat, massive traffic, and high speeds without melting or breaking down.

Here is a simple breakdown of what the researchers did and what they found:

1. The Goal: Building a "One-Way" Gate

The researchers built a specific type of electronic component called a Schottky Diode.

• The Analogy: Imagine a turnstile at a subway station. It lets people (electrons) flow through easily in one direction (forward) but locks tight to stop them from going the wrong way (reverse).
• The Innovation: They built this turnstile on a "bulk" AlN substrate. Think of this as building the turnstile directly on a solid block of the best possible material, rather than gluing it onto a weaker foundation. This makes the whole system much stronger.

2. The Results: A Super-Strong Turnstile

When they tested their new turnstile, the results were impressive:

• High Traffic Flow: It allowed a massive amount of electricity to pass through (over 2,000 amps per square centimeter). That's like a highway handling thousands of cars per second without slowing down.
• Heat Resistance: Most electronics break or act weirdly when they get hot. This device worked perfectly even when heated to 300°C (572°F)—hot enough to fry an egg on the sidewalk!
• The "Turn-On" Switch: At room temperature, it took about 3 volts to get the traffic moving. But as the device got hotter, it became easier to open the gate (requiring less voltage). This is because the heat gives the electrons a "running start," helping them jump over barriers.

3. The Hidden Glitch: The "Dusty Doorway"

While the device worked great, the researchers noticed it wasn't perfectly smooth.

• The Problem: The "ideality factor" (a score of how perfectly the turnstile works) was a bit high.
• The Cause: Using a powerful microscope (TEM), they found a tiny, 5-nanometer-thick layer of "dust" (an Aluminum Oxynitride layer) stuck between the metal gate and the semiconductor road.
• The Metaphor: Imagine your turnstile has a thin layer of sticky gum on it. People can still get through, but they have to wiggle and squeeze a bit more. This "gum" caused some electrons to get stuck or tunnel through in a messy way, making the device less efficient than a theoretical perfect one. However, when it got hot, the electrons had enough energy to blast right through the gum, making the device work much better.

4. The "Deep Sleep" Donors

The material is doped with Silicon to make it conductive, but the Silicon atoms in AlN are a bit lazy.

• The Analogy: Imagine a crowd of workers (electrons) who are "asleep" in deep bunk beds. At room temperature, the alarm clock (heat) isn't loud enough to wake them all up, so only a few are working.
• The Discovery: As the temperature rose, the alarm got louder. More and more workers woke up and started working. By the time it got hot, almost all the Silicon atoms were active, significantly increasing the number of available electrons. This explains why the device performed better at high temperatures.

5. The Leak: The "Poole-Frenkel" Escape

When they tried to force electricity the wrong way (reverse bias), a tiny bit of current leaked through.

• The Mechanism: They found this leak wasn't random. It followed a specific pattern called Poole-Frenkel emission.
• The Metaphor: Imagine a prisoner (an electron) trapped in a cell (a defect in the material). If you push hard enough on the cell door (apply a strong electric field), the prisoner can slip out. The researchers calculated that the "prison cell" was about 0.34 electron-volts deep. This helps them understand how to plug these leaks in future designs.

Why Does This Matter?

This research is a major step toward the next generation of power electronics.

• Electric Vehicles (EVs): Imagine EV chargers that are smaller, lighter, and can charge your car in minutes without overheating.
• Data Centers: Servers that run cooler and use less electricity.
• Harsh Environments: Devices that can work inside jet engines or nuclear reactors where normal electronics would melt.

In Summary: The researchers built a super-strong, heat-resistant electronic gate. They found a tiny bit of "dust" on the gate that made it slightly imperfect, but they figured out exactly how heat and electricity interact with it. This gives engineers the blueprint to build even better, faster, and more efficient power systems for our future.

Technical Summary

1. Problem Statement

Aluminum Nitride (AlN) is an ultra-wide bandgap semiconductor ($E_G \approx 6.2$ eV) with exceptional thermal conductivity and critical electric field strength, making it a prime candidate for next-generation high-power electronics and deep-ultraviolet photonics. Despite its theoretical promise, realizing high-performance AlN power devices has been hindered by:

• Material Challenges: Difficulty in achieving high-quality, low-defect epitaxial growth on bulk substrates and controlling dopant activation (specifically Si donors, which act as deep levels in AlN).
• Device Limitations: Previous reports on AlN Schottky Barrier Diodes (SBDs) have shown varying breakdown voltages and current densities, but a comprehensive understanding of carrier transport mechanisms, leakage paths, and temperature-dependent behavior (especially at elevated temperatures) remains incomplete.
• Interface Issues: The presence of interfacial layers and non-ideal ohmic contacts often degrades device performance, yet their specific impact on transport mechanisms is not fully characterized.

2. Methodology

The researchers fabricated and characterized quasi-vertical Ni/AlN Schottky Barrier Diodes (SBDs) on bulk AlN substrates.

• Epitaxial Growth:
  • Grown via Metal-Organic Chemical Vapor Deposition (MOCVD) on an insulating bulk AlN substrate.
  • Structure: An $n^+$-doped Al$_{0.8}$Ga$_{0.2}$N buffer layer ($2\times10^{19}$ cm$^{-3}$) for the bottom contact, a graded AlGaN layer to mitigate band offsets, a 1-$\mu$m thick Si-doped AlN drift layer ($1\times10^{18}$ cm$^{-3}$), and a reverse-graded AlGaN cap layer.
• Device Fabrication:
  • Etching: Inductively Coupled Plasma (ICP-RIE) using BCl$_3$/Cl$_2$/Ar to define mesas and remove the cap layer.
  • Contacts:
    • Ohmic: Ti/Al/Ni/Au stacks annealed at 950°C (also tested Cr/Al/Ni/Au and V/Al/V/Au).
    • Schottky: Ni/Au deposited on the exposed AlN epilayer.
  • Geometry: Anode diameter of 50 $\mu$m with a 2 $\mu$m spacing from the mesa edge.
• Characterization Techniques:
  • Electrical: Temperature-dependent I-V (forward and reverse) and Capacitance-Voltage (C-V) measurements from 300 K to 573 K.
  • Structural/Chemical: High-Resolution Transmission Electron Microscopy (HR-TEM) and Energy-Dispersive X-ray Spectroscopy (EDAX) to analyze the metal/semiconductor interface.
  • Simulation: TCAD 2-D simulations (Silvaco Victory) to analyze electric field distribution.
  • Modeling: Fitting reverse leakage data to the Poole-Frenkel emission model.

3. Key Contributions

• Fabrication of High-Current Quasi-Vertical SBDs: Successful demonstration of AlN SBDs on bulk substrates capable of handling current densities exceeding 2 kA/cm$^2$ at 10 V.
• Comprehensive Temperature Analysis: A detailed study of device performance from room temperature up to 573 K (300°C), revealing stable rectifying operation and specific transport mechanisms at high temperatures.
• Interface Characterization: Identification and characterization of a ~5 nm AlN$_x$O$_y$ interfacial layer formed between the Ni Schottky contact and the AlN epilayer, linking it to non-ideal transport behaviors.
• Dopant Activation Insights: Clarification of the deep-level nature of Si donors in AlN through temperature-dependent C-V measurements, showing how donor activation changes with thermal energy.
• Leakage Mechanism Identification: Determination that reverse-bias leakage is dominated by Poole-Frenkel emission (field-enhanced trap emission) with a specific trap energy level.

4. Key Results

• Forward Characteristics (Room Temperature):

  • Turn-on Voltage: ~3.0 V (measured at 1 A/cm$^2$).
  • Current Density: >2 kA/cm$^2$ at 10 V.
  • On/Off Ratio: >10$^9$.
  • Ideality Factor ($\eta$): High value of 3.6 at 300 K, indicating significant deviation from ideal thermionic emission. This is attributed to the AlN$_x$O$_y$ interlayer acting as a tunneling barrier and defect-mediated transport.
  • Ohmic Contacts: The Ti/Al/Ni/Au contacts exhibited Schottky-like behavior (turn-on ~2 V) rather than linear ohmic behavior, suggesting the high Al composition (80%) creates a large injection barrier.

• Temperature-Dependent Behavior:

  • Stability: Devices operated stably up to 573 K with an On/Off ratio >10$^7$.
  • Thermal Activation: As temperature increased, the turn-on voltage dropped to 2.29 V (at 573 K), and current density increased significantly.
  • Ideality Factor & Barrier Height: The ideality factor decreased from 3.6 (300 K) to 2.07 (573 K), while the apparent Schottky barrier height increased from 1.2 eV to 1.93 eV. This suggests that at higher temperatures, electrons overcome the interfacial tunneling barrier more effectively.
  • Donor Concentration ($N_D - N_A$): C-V measurements showed the net donor concentration increased from ~5$\times$10$^{17}$ cm$^{-3}$ at 300 K to ~1$\times$10$^{18}$ cm$^{-3}$ at 373 K. This confirms that Si donors in AlN are deep levels that require thermal energy to fully activate and respond to the AC signal.

• Reverse Characteristics & Breakdown:

  • Breakdown Voltage: Estimated at ~200 V (parallel plane assumption), corresponding to a field of 6.4 MV/cm. TCAD simulations revealed a peak field of 13 MV/cm at the anode edge due to a lack of field termination.
  • Leakage Mechanism: Reverse leakage increased with temperature. Analysis of $\ln(J/E)$ vs. $E^{1/2}$ confirmed Poole-Frenkel emission as the dominant mechanism.
  • Trap Energy: The extracted effective trap level was 0.34 eV below the conduction band.

5. Significance

This work provides critical insights into the physics of AlN-based power devices, moving beyond simple performance metrics to understand the underlying mechanisms:

1. High-Temperature Viability: It proves that AlN SBDs can operate reliably in extreme environments (up to 300°C), validating their potential for harsh-environment applications (e.g., aerospace, downhole drilling).
2. Interface Engineering: The identification of the AlN$_x$O$_y$ layer as a source of non-ideal transport highlights the need for better surface passivation or interface engineering to reduce the ideality factor and improve forward performance.
3. Dopant Management: The findings on Si donor activation suggest that device modeling must account for temperature-dependent carrier concentration, as "frozen out" donors at room temperature significantly impact device capacitance and resistance.
4. Future Roadmap: The study identifies specific areas for improvement, such as implementing field-termination structures (field plates) to enhance breakdown voltage and optimizing ohmic contact schemes for high-Al-content layers.

In summary, this paper establishes a foundational understanding of carrier transport and leakage in quasi-vertical AlN SBDs, guiding the development of high-performance, high-temperature power electronics based on ultra-wide bandgap semiconductors.