Revisiting the configurations of hydrogen impurities in SrTiO3: Insights from first-principles local vibration mode calculations

By employing first-principles local vibration mode calculations with a hybrid functional, this study identifies strontium vacancy-hydrogen complexes (VSr-Hi and VSr-2Hi) and titanium vacancy-hydrogen complexes (VTi-2Hi) as the primary sources of the dominant and additional infrared absorption bands in SrTiO3, respectively, thereby resolving previous ambiguities regarding hydrogen impurity configurations.

The Gist

Imagine Strontium Titanate (STO) as a perfectly organized, three-story apartment building made of atoms. The residents are Strontium, Titanium, and Oxygen, arranged in a strict, repeating pattern. Now, imagine a tiny, mischievous guest named Hydrogen sneaking in. Hydrogen is small and everywhere, like a ghost that can slip through walls or hide in the corners.

For decades, scientists have been trying to figure out exactly where this Hydrogen guest is hiding inside the building and what it's doing. They can "hear" the Hydrogen because it vibrates, creating a specific musical note (an infrared absorption band) that scientists can detect. However, there was a big mystery: the notes they heard didn't match the notes they expected from the Hydrogen guest sitting in the empty hallways.

This paper is like a high-tech detective story where the authors use a super-powerful computer simulation to solve the case. Here is the breakdown of their investigation:

1. The Wrong Guess: The "Hallway Hider"

For a long time, scientists thought the Hydrogen was just sitting alone in the empty spaces between the atoms (called interstitial hydrogen). They expected this lone Hydrogen to sing a high-pitched note around 3500 cm⁻¹ (a specific frequency).

The Twist: The authors ran their computer simulations with a very precise "tuning fork" (a specific mathematical formula called HSE06). They found that a lone Hydrogen in the hallway actually sings a much lower note, around 3277 cm⁻¹.

• The Conclusion: The lone Hydrogen is not the one making the loud, dominant 3500 cm⁻¹ noise that everyone hears in experiments. The "Hallway Hider" theory was wrong.

2. The Real Culprits: The "Roommates"

If the lone Hydrogen isn't the source, who is? The authors discovered that Hydrogen loves to hang out with vacancies.

Think of a vacancy as an empty apartment where a resident (Strontium or Titanium) is missing.

• The Strontium Vacancy (VSr): This is a missing Strontium resident. It turns out Hydrogen loves to move into the empty space next to this missing Strontium.
• The Discovery: When Hydrogen pairs up with a missing Strontium (forming a VSr-Hi or VSr-2Hi complex), the vibration changes. These "roommate" pairs sing notes right around 3500 cm⁻¹.
• The Match: This perfectly matches the main absorption bands scientists have been hearing for years. So, the loud noise isn't from a lone Hydrogen; it's from Hydrogen hanging out with a missing Strontium neighbor.

3. The Mystery of the "Low Notes"

Scientists also heard some quieter, lower notes around 3300 cm⁻¹.

• The Old Theory: Some thought this was just two Hydrogens hanging out together (a 2Hi pair).
• The New Evidence: The authors calculated that two Hydrogens alone would sing even lower (around 3100 cm⁻¹), so that doesn't fit.
• The Real Source: The authors found that when Hydrogen pairs up with a missing Titanium resident (VTi-2Hi), the vibration hits that 3300 cm⁻¹ sweet spot.
• The Match: The "missing Titanium + two Hydrogens" complex is the source of the lower-frequency bands.

The Big Picture: Why the Math Matters

The paper emphasizes that getting the "tuning fork" right is crucial. Previous studies used different mathematical formulas that were slightly off, leading to wrong predictions about where the Hydrogen was hiding. By using a more accurate formula (setting the "exact exchange" to 0.2), the authors finally got the notes to match the real-world experiments.

Summary

• The Problem: Scientists heard Hydrogen singing in a crystal but didn't know which "room" it was in.
• The Mistake: They thought a lone Hydrogen in the hallway was the singer.
• The Solution: The real singers are Hydrogen complexes:
  • Hydrogen + Missing Strontium = The loud 3500 cm⁻¹ song.
  • Hydrogen + Missing Titanium = The quieter 3300 cm⁻¹ song.
• The Lesson: To understand how Hydrogen changes the electrical properties of these materials, we need to stop looking for lone wolves and start looking for the groups they form with missing neighbors.

This study doesn't propose new medical uses or future gadgets; it simply clears up the confusion about the fundamental structure of Hydrogen in this specific material, ensuring that future theories are built on the correct "address" of the Hydrogen atom.

Technical Summary

1. Problem Statement

The specific atomic configurations of hydrogen impurities in Strontium Titanate (SrTiO3 or STO) remain ambiguous despite decades of research. Experimental infrared (IR) spectroscopy has identified distinct OH absorption bands:

• Main bands: A single peak or three peaks near 3500 cm⁻¹ (depending on temperature and phase symmetry).
• Secondary bands: Two peaks around 3300 cm⁻¹.

Previous theoretical studies have offered conflicting interpretations regarding which defect structures correspond to these bands:

• Some attributed the ~3500 cm⁻¹ bands to isolated interstitial hydrogen ($H_i$) or $H_i$-dimers ($2H_i$).
• Others proposed that these bands arise from complexes involving strontium vacancies ($V_{Sr}$) and hydrogen ($V_{Sr}$-$H_i$).
• Similarly, the ~3300 cm⁻¹ bands were variously attributed to $2H_i$ or titanium vacancy complexes ($V_{Ti}$-$2H_i$).

The primary challenge is that standard Density Functional Theory (DFT) functionals often yield inaccurate vibrational frequencies due to errors in describing the electronic structure and the O-H bond, leading to these unresolved discrepancies.

2. Methodology

The authors employed a systematic first-principles approach using Density Functional Theory (DFT) to resolve these conflicts:

• Computational Framework: Calculations were performed using the VASP code with a 3×3×3 supercell (135 atoms) based on the cubic STO primitive cell.
• Exchange-Correlation Functional: Crucially, the study utilized the HSE06 hybrid functional with a specific fraction of exact exchange set to 0.2 ($\alpha = 0.2$). The authors argue this setting is superior for STO compared to standard PBE (GGA) or LDA, as it correctly describes electronic properties and avoids the overestimation/underestimation of vibrational frequencies seen in other functionals.
• Defect Modeling: The study calculated formation energies for various defects under O-poor conditions to identify stable configurations in n-type STO. Considered defects included:
  • Interstitial hydrogen ($H_i$) and dimers ($2H_i$).
  • Strontium vacancy complexes ($V_{Sr}$-$H_i$, $V_{Sr}$-$2H_i$).
  • Titanium vacancy complexes ($V_{Ti}$-$H_i$, $V_{Ti}$-$2H_i$).
• Vibrational Frequency Calculation: Instead of standard harmonic approximations, the authors used a numerical local vibration mode method:
  1. Displaced the H atom along the O-H bond direction (±30% of bond length).
  2. Calculated total energies to construct a potential energy curve.
  3. Fitted the curve to a fourth-degree polynomial to extract harmonic ($k$) and anharmonic ($\alpha, \beta$) coefficients.
  4. Solved the 1D Schrödinger equation numerically (shooting method) to determine the transition energy between the ground state ($E_0$) and first excited state ($E_1$), yielding the anharmonic vibrational frequency ($\omega$).

3. Key Results

A. Formation Energies

• $H_i$ acts as a shallow donor in the +1 charge state.
• $V_{Sr}$-$H_i$ is an acceptor (stable in -1 state), while $V_{Sr}$-$2H_i$ is neutral.
• $V_{Ti}$ complexes are high-energy but stable in specific charge states (-3 for $V_{Ti}$-$H_i$, -2 for $V_{Ti}$-$2H_i$).
• Oxygen vacancies occupied by hydrogen ($V_O$-$2H_i$) are significant donors but were not the primary focus for vibrational matching.

B. Vibrational Frequencies and Structural Insights

The calculated frequencies (including anharmonic corrections) revealed significant deviations from previous theoretical predictions:

Defect Configuration	Calculated Frequency ($\omega$)	Previous Theoretical Values	Experimental Match
$H_i$ (Interstitial)	3277 cm⁻¹	~3533 cm⁻¹ (Bork), ~3505 cm⁻¹ (Limpijumnong)	No (Too low for 3500 cm⁻¹ bands)
$2H_i$ (Dimer)	3067 / 3101 cm⁻¹	~3465 / 3438 cm⁻¹	No (Too low for 3300 cm⁻¹ bands)
$V_{Sr}$-$H_i$	3567 cm⁻¹	~3505 cm⁻¹	Yes (Matches ~3500 cm⁻¹ bands)
$V_{Sr}$-$2H_i$	3545 – 3589 cm⁻¹	~3523 / 3527 cm⁻¹	Yes (Matches ~3500 cm⁻¹ bands)
$V_{Ti}$-$2H_i$	3301 / 3307 cm⁻¹	~3100 cm⁻¹ (Limpijumnong)	Yes (Matches ~3300 cm⁻¹ bands)

• Bond Length Correlation: A clear inverse relationship was observed: shorter O-H bonds correspond to higher vibrational frequencies. The use of HSE06 ($\alpha=0.2$) yielded moderate bond lengths (e.g., $H_i$ at 0.972 Å) that balanced the over-localization of PBE and under-localization of LDA.
• Anharmonicity: The anharmonic contribution ($\Delta\omega$) was significant (ranging from -180 to -305 cm⁻¹), reducing the harmonic frequency ($\omega_0$) to the final observed frequency.

4. Identification of Experimental Bands

Based on the recalculated frequencies, the authors propose a revised assignment for the experimental IR absorption bands:

1. Main Bands (~3500 cm⁻¹): These are NOT caused by isolated interstitial hydrogen ($H_i$). Instead, they originate from Strontium vacancy complexes ($V_{Sr}$-$H_i$ and $V_{Sr}$-$2H_i$). The calculated frequencies (3545–3589 cm⁻¹) align perfectly with the experimental peaks.
2. Secondary Bands (~3300 cm⁻¹): These are NOT caused by $H_i$ dimers ($2H_i$). They are attributed to Titanium vacancy complexes with two hydrogens ($V_{Ti}$-$2H_i$), specifically the configuration where the two O-H bonds are symmetric (calculated at ~3301–3307 cm⁻¹).

5. Significance and Conclusion

• Resolution of Discrepancies: This study resolves long-standing conflicts in the literature by demonstrating that previous theoretical assignments were likely skewed by the choice of exchange-correlation functionals.
• Methodological Validation: It underscores the critical importance of using hybrid functionals (specifically HSE06 with $\alpha=0.2$) and anharmonic corrections for accurate vibrational property predictions in oxide semiconductors.
• Electronic Implications: The findings suggest that hydrogen-induced conductivity and electronic behavior in STO are governed not just by isolated interstitials, but significantly by hydrogen-vacancy complexes. This shifts the understanding of defect engineering in perovskite oxides, highlighting the role of $V_{Sr}$ and $V_{Ti}$ in trapping hydrogen and modifying the material's electronic state.
• Future Impact: The work provides a reliable theoretical framework for interpreting spectroscopic data in STO and similar oxide systems, aiding in the design of hydrogen-sensitive devices and understanding hydrogen-induced metallization or insulation transitions.