Loss analysis of Low Bandgap (Ag,Cu)(In,Ga)Se2 Solar Cells for Tandem Applications

This study analyzes 18.5% efficient 1.0 eV (Ag,Cu)(In,Ga)Se2 solar cells for tandem applications, identifying that while current losses stem from absorption, the most significant performance limitations arise from non-radiative recombination affecting open-circuit voltage and a high diode factor indicating strong space charge region recombination.

The Gist

The Big Picture: Building a Better Solar Sandwich

Imagine the sun's energy as a giant, diverse buffet of light. Some light is high-energy (blue/UV), some is medium (green/yellow), and some is low-energy (red/infrared).

For a long time, solar cells were like single-layer sandwiches. They could only eat the "medium" part of the buffet well. To get more energy, scientists are now building Tandem Solar Cells. Think of this as a two-layer sandwich:

1. The Top Layer: Catches the high-energy light.
2. The Bottom Layer: Catches the low-energy light that passed through the top.

This paper focuses on the Bottom Layer. The scientists are trying to make a specific type of solar cell (called (Ag,Cu)(In,Ga)Se2) that is perfect for catching that low-energy, red light. They made a very good one (18.5% efficient), but they wanted to know: "Why isn't it perfect? Where are we losing energy?"

They acted like detectives, breaking the solar cell's performance down into three main "buckets" of loss: Current (how much electricity flows), Voltage (how hard the electricity pushes), and Fill Factor (how smoothly the electricity flows).

---

1. The Current Loss: The "Leaky Bucket" and the "Thick Wall"

The Problem: The cell isn't capturing every single photon of light it should.
The Analogy: Imagine the solar cell is a sponge trying to soak up water (light).

• The Short-Wavelength Leak: Some light hits the "plastic wrap" (the top layers of the cell) and gets absorbed or reflected before it even reaches the sponge. The scientists found this isn't a huge problem for this specific type of cell, but it's still a small leak.
• The Long-Wavelength Wall: This is the bigger issue. The sponge (the absorber material) is a bit too thin. Some of the low-energy red light passes right through the sponge without getting soaked up.
• The Fix: To fix this, you'd need a thicker sponge or a mirror at the bottom to bounce the light back in. However, the paper notes that the collection of the light (once it's absorbed) is actually working great. The light is being caught, but the sponge just isn't thick enough to catch all of it.

Verdict: Current loss is about 5 mA/cm². It's annoying, but not the main villain.

---

2. The Voltage Loss: The "Leaky Battery"

The Problem: The cell isn't generating as much "push" (voltage) as physics says it should.
The Analogy: Imagine the solar cell is a battery. In a perfect world, every bit of light energy turns into electrical pressure. In reality, the battery has a hole in it, and the pressure leaks out before you can use it.

The scientists found two types of leaks:

1. The "Ghost" Leak (Radiative Loss): This is unavoidable. Even a perfect battery loses a tiny bit of energy just by glowing in the dark. This is a very small leak.
2. The "Real" Leak (Non-Radiative Loss): This is the big one. Inside the material, there are tiny defects (like potholes in a road). When an electron (the energy carrier) tries to drive down the road, it hits a pothole and crashes, losing its energy as heat instead of electricity.

The Detective Work:
The scientists measured the "Quasi-Fermi Level Splitting" (a fancy way of saying "how much potential energy is stored"). They found that the bulk of the material (the sponge itself) is where the biggest leaks are. Even before they put the top layers on the cell, the material was already leaking voltage.

• Key Finding: The material quality is the main culprit. It's not the top layers causing the voltage drop; it's the "sponge" itself having too many potholes.

Verdict: This is the biggest loss (over 150 mV). If they could make the material purer (fewer potholes), the voltage would jump significantly.

---

3. The Fill Factor: The "Traffic Jam"

The Problem: The electricity flows, but it's sluggish and bumpy.
The Analogy: Imagine a highway.

• Current is how many cars are on the road.
• Voltage is how fast the cars want to go.
• Fill Factor is how smoothly the traffic moves. If there are traffic lights, stop signs, or accidents, the cars stop and start, wasting time and fuel.

The scientists looked at the "Diode Factor" (a number that tells us how "ideal" the traffic flow is).

• The Absorber (The Sponge): When they tested just the raw material, the traffic flow was okay (Diode factor ~1.3).
• The Finished Cell: Once they added the top layers to make a working solar cell, the traffic jam got worse (Diode factor jumped to ~1.8).

Why?
The scientists realized that when the top layers are added, they create a "Space Charge Region" (a specific zone where the electric field is strong). It turns out that in this specific zone, there are a lot of "accidents" (recombination events) happening that weren't happening in the raw material.

• The Metaphor: It's like adding a toll booth to a highway. The road itself is fine, but the toll booth (the junction between layers) is causing a massive traffic jam because of how the cars interact with it.

Verdict: The "traffic jam" at the junction is the main reason the Fill Factor is low.

---

The Final Verdict: How to Build a Super-Cell

The paper concludes with a "What If?" scenario. If the scientists could fix these specific problems, here is what would happen:

1. Fix the Material (Stop the Voltage Leak): If they could make the material so pure that it loses very little energy as heat, the voltage would go up.
2. Fix the Junction (Clear the Traffic Jam): If they could engineer the junction so the "Diode Factor" drops back down to ideal levels, the electricity would flow much smoother.
3. Fix the Thickness (Catch the Red Light): If they used a slightly thicker sponge or better mirrors, they'd catch more current.

The Dream Result:
By fixing these three things, they estimate they could boost the efficiency from 18.5% to 22.8%.

Summary in One Sentence

This paper is a detailed autopsy of a high-performing solar cell that revealed its biggest weaknesses aren't in how it catches light, but in internal material defects that waste energy as heat and a junction interface that causes traffic jams for electrons, and fixing these two issues is the key to unlocking the next generation of super-efficient solar power.

Technical Summary

1. Problem Statement

Tandem solar cells are essential for surpassing the theoretical efficiency limits of single-junction devices (33% vs. 46%). To achieve this, highly efficient bottom cells with a bandgap of approximately 1.0 eV are required. Chalcopyrite materials, specifically (Ag,Cu)(In,Ga)Se2 (ACIGS), are the leading candidates for this role. However, current state-of-the-art ACIGS bottom cells (efficiency ~18.5%) still fall significantly short of the Shockley-Queisser (SQ) limit (31.6% for 1.0 eV). There is a critical need to identify the specific physical mechanisms causing losses in short-circuit current ($J_{SC}$), open-circuit voltage ($V_{OC}$), and fill factor (FF) to guide future improvements.

2. Methodology

The authors conducted a comprehensive loss analysis on high-efficiency (18.5%) 1.0 eV bandgap ACIGS cells (sample NaF10). The study employed a multi-modal approach comparing the finished solar cells with the absorber/CdS stack (obtained by etching off the top window layers) to isolate where losses occur.

• Electrical Characterization: $J-V$ measurements under AM 1.5 illumination and External Quantum Efficiency (EQE) spectra.
• Optical Characterization:
  • Absolute Photoluminescence (PL): Measured on both the absorber/CdS stack and finished cells to determine Quasi-Fermi Level Splitting (QFLS), absorptance ($A(E)$), and Photoluminescence Quantum Yield ($Y_{PL}$).
  • Electroluminescence (EL): Used to determine the Electroluminescence Diode Factor (ELDF).
• Loss Decomposition:
  • $V_{OC}$ Loss: Separated into short-circuit (generation) loss, radiative loss, and non-radiative loss using Rau's formalism and generalized Planck's law.
  • Diode Factor Analysis: Compared the Electrical Diode Factor (EDF) from $J-V$ fits with the Optical Diode Factor (ODF) derived from PL intensity vs. generation flux. This comparison helps distinguish between recombination in the neutral bulk region versus the Space Charge Region (SCR).

3. Key Contributions

• Isolation of Loss Mechanisms: The study successfully decouples losses occurring in the bulk absorber from those introduced by the device architecture (TCO, buffer layers, and junction formation).
• Identification of SCR Recombination: The paper provides strong evidence that the primary cause of the low Fill Factor in these low-bandgap cells is not bulk recombination, but rather Shockley-Read-Hall (SRH) recombination in the Space Charge Region (SCR) formed upon junction completion.
• Validation of Absorber Quality: It demonstrates that key parameters like Urbach energy and bandgap broadening are intrinsic to the absorber material and are not significantly altered by the deposition of top layers.

4. Key Results

A. Short-Circuit Current ($J_{SC}$) Losses

• Total Loss: Approximately 5 mA cm⁻² below the SQ limit.
• Causes:
  • Short Wavelength: Parasitic absorption in the CdS buffer, i-ZnO, and AZO layers.
  • Long Wavelength: Absorption losses within the absorber itself (due to insufficient thickness or lack of light management).
  • Collection: Collection losses near the absorption edge were found to be negligible; the primary issue is optical absorption, not carrier collection.

B. Open-Circuit Voltage ($V_{OC}$) Losses

The $V_{OC}$ deficit is dominated by non-radiative recombination:

• Non-Radiative Loss: >150 mV. This is the single largest loss mechanism, originating from the bulk of the chalcopyrite absorber.
• Radiative Loss: ~15 mV (due to bandgap broadening, $\sigma \approx 27.5$ meV).
• Generation Loss: ~3 mV (negligible).
• Optical Loss: ~8 mV (parasitic absorption in top layers).
• Key Finding: The QFLS of the finished cell is nearly identical to the absorber/CdS stack, confirming that the bulk absorber quality dictates the $V_{OC}$ limit, not the interface or top layers.

C. Fill Factor (FF) and Diode Factor

• The Discrepancy: The Electrical Diode Factor (EDF) of the finished cell is significantly higher (1.4–1.5) than the Optical Diode Factor (ODF) of the absorber/CdS stack (1.28).
• The Cause: The absorber/CdS stack lacks a Space Charge Region (SCR) because the n-type buffer (CdS) is not sufficiently doped to create a depletion region in the p-type absorber. Once the TCO and full junction are formed, an SCR is created.
• Conclusion: The increase in the diode factor is caused by recombination in the SCR. The low-bandgap CuInSe$_2$ (with minimal Ga) in the SCR contains SRH recombination centers that degrade the FF.
• Comparison: The study notes that introducing a Ga gradient at the front (increasing the bandgap in the SCR) reduces this recombination, lowering the EDF, though this often compromises $J_{SC}$.

5. Significance and Future Outlook

This work provides a roadmap for pushing ACIGS bottom cells toward 22%+ efficiency:

1. Bulk Passivation: Reducing non-radiative bulk recombination is the most critical step. If the PL Quantum Yield (PLQY) could be improved to 1% (similar to record 1.13 eV cells), $V_{OC}$ could increase by ~150 mV, boosting efficiency by 1.2% absolute.
2. SCR Engineering: To fix the FF without sacrificing $J_{SC}$, new strategies are needed to passivate the SCR without relying on a front Ga gradient. The authors suggest that the specific defects causing SCR recombination (potentially the 0.8 eV defect) need to be identified and mitigated.
3. Optical Optimization: Replacing the CdS buffer with wider bandgap alternatives (e.g., Zn(O,S)) and using low-absorption TCOs (e.g., bias-ZnO) could recover ~0.2% efficiency.
4. Theoretical Limit: If the diode factor could be reduced to the theoretical minimum for chalcopyrites (1.1) and non-radiative losses minimized, the authors project a potential efficiency of 22.8%.

In summary, the paper concludes that while optical losses are manageable, the fundamental bottleneck for 1.0 eV ACIGS tandem bottom cells is non-radiative bulk recombination (limiting $V_{OC}$) and SCR recombination (limiting FF).