Coexistence of Donor and Acceptor Hydrogen States in n-Type InN

Using hard x-ray photoemission spectroscopy, this study demonstrates that hydrogen in n-type InN exhibits amphoteric behavior by coexisting as both donor (H+) and acceptor (H-) states, providing a microscopic explanation for hydrogen-driven compensation in this semiconductor.

The Gist

Imagine you are trying to build a super-fast highway for electrons (the tiny particles that carry electricity) using a material called Indium Nitride (InN). This material is like a high-performance sports car engine; it's naturally very good at letting electrons zoom through it. However, there's a problem: the engine is running too hot. It has too many electrons already, making it hard to control the flow for specific tasks.

Scientists have long suspected that a tiny, invisible guest named Hydrogen is the culprit. Hydrogen is a bit of a trickster in the world of semiconductors. Depending on the situation, it can act like a generous donor (giving away an electron to help the flow) or a stingy acceptor (grabbing an electron to stop the flow). In most materials, Hydrogen behaves predictably, but in InN, it was a mystery. Was it just a donor making the problem worse, or was it doing something else?

To solve this mystery, the researchers at NTT used a special "super-microscope" called Hard X-ray Photoemission Spectroscopy (HAXPES). Think of this as a deep-penetrating flashlight. Unlike regular flashlights that only see the surface dust, this one can see deep inside the material to get a true picture of what's happening in the bulk, not just the skin.

Here is what they did and what they found:

The Experiment: The "Heat Treatment"
The team grew thin films of InN at different temperatures. Then, they took some of these films and gave them a "heat bath" (post-annealing) in a nitrogen environment. You can think of this heat bath as a way to shake the material gently to kick out some of the unwanted Hydrogen guests.

The Results: What Happened?

1. The Crowd Thinned Out: After the heat bath, the number of Hydrogen atoms in the material dropped significantly (by about ten times). At the same time, the number of free-floating electrons (the traffic on our highway) also dropped. This confirmed that Hydrogen was indeed acting as a "donor," adding extra electrons to the mix.
2. The Chemical Shift: When they looked at the energy levels of the atoms inside the material, they saw a shift. It was like watching a crowd of people move closer together or spread out. The shift confirmed that the "pressure" of the electrons had decreased, exactly as you would expect if you removed the Hydrogen donors.
3. The Big Surprise (The Trickster Revealed): This is where the story gets interesting. While they knew Hydrogen was acting as a donor (adding electrons), they found evidence that it was also acting as an acceptor (removing electrons) at the same time.
   • They saw a specific "signature" in the data near the bottom of the energy band (the Valence Band) that disappeared after the heat bath.
   • They also saw subtle changes in the atomic bonds: some Hydrogen seemed to be holding onto Indium atoms (acting like a magnet that grabs electrons), while other Hydrogen was near Nitrogen atoms (acting like a giver).

The Conclusion: A Two-Faced Guest
The paper concludes that in this specific material, Hydrogen is a two-faced character. It doesn't just pick one role; it does both simultaneously.

• The Donor (H+): Most of the Hydrogen is acting as a donor, which explains why the material is naturally so conductive. This version is also the "tougher" one, staying put even after the heat treatment.
• The Acceptor (H–): A smaller group of Hydrogen atoms is acting as an acceptor, trying to cancel out the extra electrons. This version is "weaker" and gets kicked out more easily by the heat.

Why It Matters
This discovery changes the story from "Hydrogen is just a troublemaker adding too many electrons" to "Hydrogen is a complex character trying to balance the books." It suggests that the reason InN is so hard to control isn't just because Hydrogen is adding electrons, but because Hydrogen is also trying to remove them, creating a tug-of-war inside the material.

The researchers didn't propose new devices or medical uses in this paper. Instead, they provided a clearer, microscopic map of how Hydrogen behaves in Indium Nitride, showing that even in a material that is overwhelmingly "electron-rich," the "electron-hungry" side of Hydrogen is still present, fighting to keep the balance.

Technical Summary

Technical Summary: Coexistence of Donor and Acceptor Hydrogen States in n-Type InN

Problem Statement
Indium nitride (InN) is a narrow-bandgap semiconductor (~0.65 eV) with high theoretical electron mobility, making it attractive for high-speed electronics and optoelectronics. However, its practical application is hindered by a persistent, degenerate n-type conductivity with high residual electron concentrations. While the origins of this conductivity are debated (involving nitrogen vacancies, threading dislocations, and extrinsic impurities), the specific role of hydrogen remains unresolved. Although hydrogen is known to act as a donor (H⁺) in n-type InN, this behavior seemingly contradicts the general amphoteric nature of hydrogen in semiconductors, where it typically acts as an acceptor in n-type materials to compensate carriers. Clarifying whether hydrogen exists solely as a donor or if acceptor states also coexist is critical for understanding the compensation mechanisms in InN.

Methodology
The authors employed Hard X-ray Photoemission Spectroscopy (HAXPES) to probe the electronic structure of wurtzite InN thin films. The study utilized samples grown via Metalorganic Vapor Phase Epitaxy (MOVPE) at various substrate temperatures ($T_s$ ranging from 450°C to 565°C). To isolate the effects of hydrogen, the as-grown (AG) films were subjected to post-annealing (PA) at 450°C in a nitrogen atmosphere.

Key experimental advantages included:

• Probing Depth: Using 6 keV incident photons, the inelastic mean free path was approximately 7.0 nm (probing depth 20 nm), significantly deeper than conventional PES (5 nm), thereby minimizing surface contributions and focusing on bulk electronic states.
• Complementary Measurements: Secondary-ion mass spectroscopy (SIMS) was used to track hydrogen and oxygen concentrations, while Hall-effect measurements determined electron carrier concentrations ($n$).
• Spectral Analysis: The study involved detailed analysis of core-level spectra (In 3p and N 1s) and valence-band (VB) spectra. Difference spectra (PA minus AG) were generated after aligning peaks to account for chemical-potential shifts, allowing for the identification of specific line-shape changes associated with hydrogen removal.

Key Results

1. Electrical and Compositional Changes: Post-annealing reduced the hydrogen concentration by approximately one order of magnitude (from ~10²¹ to ~10²⁰ cm⁻³) and correspondingly decreased the electron concentration ($n$) from the 10¹⁹ cm⁻³ range to 10¹⁸ cm⁻³. SIMS data indicated that oxygen concentrations remained largely unchanged, suggesting the electrical changes were driven primarily by hydrogen removal.
2. Chemical Potential Shifts: Both In 3p and N 1s core-level spectra in the PA films shifted toward lower binding energy compared to AG films. This shift was consistent with a chemical-potential shift toward the valence-band maximum (VBM), reflecting the reduction in electron concentration. The shift was more pronounced in films grown at higher $T_s$.
3. Identification of Hydrogen-Related States:
   • Acceptor-like H⁻: Spectral analysis of the In 3p core levels revealed a component on the lower binding energy side (indicating a lower effective valence) that was suppressed after annealing. This is attributed to In–H bonding, suggesting the presence of acceptor-like H⁻ states trapping electron carriers near In³⁺ sites.
   • Donor-like H⁺: Conversely, the N 1s spectra showed a reduction in a higher binding energy component after annealing. This is associated with N–H bonding, consistent with donor-like H⁺ species (e.g., at nitrogen vacancies or interstitial sites) previously predicted and observed in InN.
   • Valence Band Features: The valence-band spectra of AG films exhibited an in-gap feature near the VBM that was suppressed after annealing. Since acceptor H⁻ (1s²) is an occupied state, it contributes to the valence band spectral weight, whereas donor H⁺ (1s⁰) does not. The disappearance of this feature upon hydrogen removal provides direct evidence for the existence of occupied acceptor states.
4. Thermal Stability: The results suggest that donor-like H⁺ states are more thermally robust than acceptor-like H⁻ states, as the latter are more readily removed by annealing.

Significance and Claims
The paper claims to provide a microscopic picture of hydrogen-driven compensation in n-type InN. The primary contribution is the demonstration that donor-like (H⁺) and acceptor-like (H⁻) hydrogen states coexist within the same n-type InN films.

• Resolution of Contradiction: This finding reconciles the observed donor behavior of hydrogen in InN with the general amphoteric nature of hydrogen in semiconductors. While H⁺ dominates conductivity and is thermally stable, H⁻ states exist to partially compensate the active carriers, explaining the residual carrier problem.
• Methodological Validation: The study highlights the utility of HAXPES in disentangling bulk impurity states from surface effects, specifically identifying subtle line-shape changes in core levels and in-gap features in the valence band that are indicative of specific charge states.
• Implications: The work suggests that the residual n-type conductivity in InN is not solely due to donor defects but is a complex interplay where hydrogen acts as both a donor and a compensating acceptor. The authors maintain a modest stance, noting that while H⁺ likely dominates the conductivity, the existence of H⁻ is a crucial component of the material's electronic landscape.