Hybrid functional calculation of electrical activity and complexing mechanism of Cu-related defects

This study employs HSE06 hybrid functional calculations to determine the properties of various copper-related defects in silicon, proposing a $\mathrm{Cu_{i4}V}$ model to resolve discrepancies regarding the $\mathrm{Cu_{PL}}$ line and offering insights into copper precipitation mechanisms.

The Gist

The Big Picture: The "Rogue Copper" Problem

Imagine a silicon chip (the brain of your computer or solar panel) as a perfectly organized city made of silicon atoms. Everyone knows their place, and traffic flows smoothly.

Now, imagine a few Copper (Cu) atoms sneak into this city. Copper is a "rogue" element here. It's like a hyperactive toddler who runs around too fast, bumps into things, and causes chaos. In the world of electronics, this chaos means:

• Short circuits: Electricity leaks where it shouldn't.
• Dead zones: The device stops working.
• Shortened lifespan: Your solar panel or phone battery dies faster.

The scientists in this paper are like detectives trying to figure out exactly what these rogue copper atoms are doing, who they are hanging out with, and how to stop them.

---

The Detective's Toolkit: The "Super-Computer Simulation"

Instead of just looking at the silicon under a microscope (which is hard because copper moves too fast and hides well), the team used a powerful computer simulation. Think of this as a high-tech flight simulator for atoms.

They used a special mathematical rulebook (called HSE06) that is much more accurate than older rulebooks. It's like upgrading from a black-and-white map to a 3D, high-definition GPS that shows exactly where every atom is standing and how much energy they have.

The Investigation: Who is Copper Hanging With?

The detectives looked at four main scenarios where copper causes trouble:

1. The Solo Runaways (Isolated Copper)

• The Scenario: Copper atoms running alone through the silicon city.
• The Discovery: Copper loves to run through the empty spaces between the silicon atoms (interstitials) rather than sitting in a house (substitutional).
• The Analogy: It's like a ghost running through walls instead of walking through doors. They found that these "ghosts" move incredibly fast, but if they get trapped in a hole (a vacancy), they slow down and become a different kind of troublemaker.

2. The "Bad Boy" Alliances (Copper + Dopants)

Silicon cities have "police officers" (dopants) like Boron and Phosphorus to control traffic.

• Copper + Boron (The Weak Handshake): Copper tries to hold hands with Boron, but the grip is weak. It's like a handshake between two people who don't trust each other; they let go almost immediately.
• Copper + Phosphorus (The Iron Grip): When Copper meets Phosphorus, they form a very strong bond. It's like a heavy-duty magnet.
• The Takeaway: If you want to clean up copper from a silicon chip, adding Phosphorus is a great strategy because it acts like a magnet trap, grabbing the copper and holding it tight so it can't cause damage.

3. The "Silencer" Effect (Copper + Hydrogen)

Hydrogen is like a pacifier or a mute button in the semiconductor world.

• The Discovery: When Hydrogen meets a Copper defect, it latches onto it.
• The Result: One Hydrogen atom calms the copper a little. Two make it quieter. But it takes three Hydrogen atoms to completely "mute" the copper, turning it from an electrically active troublemaker into a harmless, inactive lump.
• The Analogy: Imagine a screaming child (Copper). One parent (Hydrogen) tries to quiet them, but they still scream. Three parents holding the child at once finally get them to sleep.

4. The Mystery of the "CuPL" Center (The Ghost in the Machine)

This is the paper's biggest breakthrough. Scientists have seen a strange "glow" (a light signal) in copper-contaminated silicon for years, but they couldn't figure out what was causing it. They called it the CuPL center.

• The Old Theory: Everyone thought it was a specific group of atoms: One copper sitting in a house, holding hands with three other copper ghosts.
• The New Theory (The Paper's Verdict): The computer simulation says, "Nope, that doesn't fit the data."
• The Real Culprit: The CuPL center is actually a cluster of four copper atoms huddled together around a hole (vacancy) in the silicon city.
• The Analogy: Imagine a group of four kids playing hide-and-seek in a giant empty room (the vacancy). The old theory thought they were playing in a specific formation, but the new data proves they are huddled in a tight circle. This new shape explains the "glow" perfectly.

Why Does This Matter?

1. Better Solar Panels: By understanding how copper moves and how to trap it (using Phosphorus or Hydrogen), we can make solar cells that last longer and work better.
2. Better Chips: It helps engineers design cleaner silicon wafers for computers, reducing failures.
3. Solving a 20-Year Mystery: They finally figured out the true shape of the mysterious CuPL defect, which had confused scientists for decades.

The Bottom Line

This paper is like a masterclass in atomic detective work. By using a super-accurate computer model, the team mapped out exactly how copper behaves, who it teams up with, and what its true shape is. They proved that Phosphorus is a great copper catcher, Hydrogen is a great copper silencer, and the mysterious CuPL glow is caused by a cluster of four copper atoms, not the group everyone thought it was.

This knowledge helps us build cleaner, more reliable technology for the future.

Technical Summary

1. Problem Statement

Copper (Cu) is a highly detrimental contaminant in silicon-based devices and solar cells due to its high diffusivity, tendency to precipitate, and ability to form deep-level recombination centers. While extensive experimental and theoretical work exists on isolated Cu defects, significant gaps remain in understanding:

• Complex Defects: The exact atomic configurations and transition levels of Cu complexes with dopants (Boron, Phosphorus) and hydrogen are often inconsistent between theory and experiment.
• The CuPL Center: The nature of the well-known "CuPL" photoluminescence center (observed at 1.014 eV) remains unresolved. Previous models (e.g., $Cu_{Si}Cu_{i3}$) fail to match experimental transition levels, while alternative models (e.g., $Cu_{i4}V$) lack comprehensive validation.
• Theoretical Limitations: Previous Density Functional Theory (DFT) studies often suffered from finite-size errors, lack of charge-state-dependent corrections, and the use of functionals (like PBE/LDA) that underestimate reaction barriers and band gaps.

2. Methodology

The authors employed a unified computational framework based on Density Functional Theory (DFT) using the HSE06 hybrid functional to ensure accurate band gaps and defect level predictions.

• Software & Setup: Calculations were performed using VASP with Projector Augmented-Wave (PAW) potentials.
• Supercells: A 64-atom supercell was used for most defects, while a larger 216-atom supercell was utilized for the CuPL defect to minimize finite-size effects.
• Corrections: Systematic finite-size corrections were applied to formation energies and transition levels to address electrostatic interactions in charged supercells.
• Analysis:
  • Formation Energy ($E_f$): Calculated for various charge states to determine stability across different Fermi levels.
  • Transition Levels ($\epsilon$): Derived from the difference in formation energies between charge states.
  • Binding Energy ($E_b$): Calculated to assess the stability of complexes (Cu-B, Cu-P, Cu-H, Cu-Cu).
  • Diffusion Barriers: Computed using the Climbing Image Nudged Elastic Band (CI-NEB) method.

3. Key Contributions and Results

A. Isolated Cu Defects ($Cu_i$ and $Cu_{Si}$)

• $Cu_i$ (Interstitial): Identified as the dominant species in p-type silicon. It acts as a fast diffuser with a calculated migration barrier of 0.15 eV (consistent with experimental 0.18 eV). The (+1/0) transition level is located at $E_c - 0.08$ eV.
• $Cu_{Si}$ (Substitutional): Forms when $Cu_i$ fills a pre-existing vacancy. It exhibits multiple charge states (+1, 0, -1, -2). The study confirms that $Cu_{Si}$ formation is energetically favorable via vacancy trapping rather than direct displacement of a Si atom.
• Symmetry: The study found that structural distortions around $Cu_{Si}$ are subtle, with near-degenerate energy minima for different symmetries, having negligible impact on calculated transition levels.

B. Cu-Dopant Complexes (Cu-B and Cu-P)

• Cu-B Interaction: In p-type silicon, $Cu_i$ and $B_{Si}$ form a weakly bound complex ($E_b = 0.44$ eV) with $C_{3v}$ symmetry. The low dissociation energy (0.59 eV) implies a short lifetime (<1 ms), allowing $Cu_i$ to diffuse to other traps.
• Cu-P Interaction: In n-type silicon, the interaction is significantly stronger. The binding energy between $Cu_{Si}$ and $P_{Si}$ exceeds 1.0 eV and increases as the Fermi level rises (more negative charge on Cu).
  • Mechanism: The study supports a gettering model where $Cu_{Si}^{-3}$ binds strongly to $P_{Si}^{+}$, effectively trapping copper in heavily phosphorus-doped regions. This contrasts with the weak trapping by Boron.

C. Cu-H Complexes ($Cu_{Si}H_i^n$)

• Configuration Dependence: The stable symmetry of $Cu_{Si}H_i$ complexes depends on the charge state ($C_s$ for +1, 0, -1; $C_{2v}$ or $C_s$ for -2).
• Transition Levels: The calculated levels for $Cu_{Si}H_i$ ($E_v + 0.12, 0.53, 0.83$ eV) show excellent agreement with Deep Level Transient Spectroscopy (DLTS) data.
• Passivation: Up to three hydrogen atoms can sequentially bind to $Cu_{Si}$, completely passivating its electrical activity.
  • $Cu_{Si}H_i^3$ is electrically inactive.
  • The formation of these complexes is kinetically hindered in n-type silicon due to Coulomb repulsion but proceeds readily in p-type silicon.

D. The CuPL Center ($Cu_{PL}$)

The paper resolves the long-standing debate regarding the atomic structure of the CuPL center (1.014 eV PL line):

• Model Comparison: Two models were tested: the previously favored $Cu_{Si}Cu_{i3}$ and the alternative $Cu_{i4}V$ (four interstitial Cu atoms surrounding a vacancy).
• Findings:
  • $Cu_{Si}Cu_{i3}$: Calculated (+1/0) transition level at $E_v + 0.32$ eV, which deviates significantly from the experimental $E_v + 0.10$ eV.
  • $Cu_{i4}V$: Calculated (+1/0) transition level at $E_v + 0.07$ eV, matching the experimental value almost perfectly. It also has a lower formation energy than the $Cu_{Si}Cu_{i3}$ model.
• Symmetry Discrepancy: While experiments suggest $C_{3v}$ symmetry and $Cu_{i4}V$ has $T_d$ symmetry, the authors propose that the observed symmetry arises from Jahn-Teller distortion in the excited state (exciton-defect complex), while the ground state remains $T_d$.
• Transformation: The study suggests that $Cu_{Si}Cu_{i3}$ can transform into $Cu_{i4}V$ upon annealing (barrier ~0.06 eV), explaining the disappearance of the CuPL signal and the appearance of $Cu_{Si}$ signals after thermal treatment.

4. Significance

• Unified Framework: This work provides a consistent theoretical picture of Cu defects using high-accuracy HSE06 functionals and rigorous finite-size corrections, resolving many discrepancies between past theoretical predictions and experimental data.
• Process Control: The findings offer critical insights for semiconductor manufacturing:
  • Gettering: Phosphorus is identified as a superior gettering agent for Cu compared to Boron in n-type silicon.
  • Passivation: Hydrogenation is confirmed as an effective method to neutralize Cu-related deep levels, provided the Fermi level conditions favor complex formation.
• Defect Identification: The proposal of the $Cu_{i4}V$ model as the true nature of the CuPL center provides a definitive structural explanation for a major spectroscopic feature in Cu-contaminated silicon, guiding future experimental characterization and defect engineering.

In summary, the paper advances the fundamental understanding of copper contamination in silicon, offering precise atomic models and energetic landscapes that are essential for improving the performance and reliability of silicon-based electronic and photovoltaic devices.