🔬 materials science

What enables GaOx as hole transport layer for a 16 percent 1.0 eV CuInSe2 Bottom Cells with VOC above 550 mV?

This paper demonstrates that a conductive, partly crystalline GaOx hole transport layer, formed via ion exchange during co-evaporation without requiring additional Cu, Na, or heavy alkalis, enables a pure CuInSe2 bottom solar cell to achieve a record-certified open-circuit voltage of 552 mV and an active area efficiency exceeding 16%.

Original authors: Francesco Lodola, Zhuangyi Zhou, Boaz Koren, Saeed Bayat, Alessandro Magon, Yucheng Hu, Adrian-Marie Philippe, Michele Melchiorre, Hasan Arif Yetkin, Gunnar Kusch, Nathalie Valle, Rachel A. Oliver, Su

Published 2026-07-01

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Francesco Lodola, Zhuangyi Zhou, Boaz Koren, Saeed Bayat, Alessandro Magon, Yucheng Hu, Adrian-Marie Philippe, Michele Melchiorre, Hasan Arif Yetkin, Gunnar Kusch, Nathalie Valle, Rachel A. Oliver, Susanne Siebentritt

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a solar cell as a busy factory that turns sunlight into electricity. Inside this factory, there's a main production floor (the "absorber") where the magic happens, and a loading dock at the back (the "back contact") where the finished products (electricity) leave the building.

For a long time, scientists have been trying to build a specific type of solar factory using a material called CuInSe2 (Copper-Indium-Selenium). This material is special because it's like a "Goldilocks" material: it's just right for catching the lower-energy part of sunlight, which is crucial for building super-efficient "tandem" solar cells (two layers stacked on top of each other).

However, this specific factory had a major problem: The loading dock was leaky.

The Problem: A Leaky Back Door

When the electricity tried to leave the factory through the back door (the metal contact), it was getting lost. Instead of flowing out to power your home, the electrons were crashing into the wall and disappearing. This is called "recombination." It's like workers trying to leave a building but getting stuck in a revolving door that spins them back inside. This leak caused the factory to produce very little voltage, limiting its efficiency.

The Solution: A Smart "Doorman" (GaOx)

The researchers in this paper discovered a clever way to fix the leaky door. They installed a special "doorman" layer made of a material called Gallium Oxide (GaOx) right between the production floor and the loading dock.

Think of this GaOx layer as a smart bouncer:

It blocks the bad guys: It stops the electrons from crashing into the back wall and getting lost (it "passivates" the contact).
It lets the good guys through: Surprisingly, this bouncer is also a conductor for "holes" (the positive charges that need to leave). Usually, Gallium Oxide is an insulator (like a brick wall), but in this specific setup, it became a highway for the electricity to flow out.

The Big Discovery: No Extra Ingredients Needed

Before this study, scientists thought they needed to add extra "ingredients" to make this bouncer work. They thought they had to:

Add extra Copper: Like adding a special lubricant to the door hinges.
Add Sodium: Like hiring a specific manager to organize the workers.
Heat it up with Copper: Like baking the door to make it work.

The paper's main finding is that none of these extra steps were actually necessary.

The researchers tested three different scenarios:

The "Standard" way: Added extra copper and baked it.
The "No-Copper" way: Just baked it without extra copper.
The "No-Heat" way: Didn't bake it at all.

Result: All three methods worked! The GaOx layer became a conductive highway for electricity even without the extra copper or the special baking step. This is a huge deal because it means the process is simpler, cheaper, and easier to scale up for mass production.

The Secret Ingredient: Crystalline Structure

Why did it work without the extra copper? The paper suggests it's because of the structure of the GaOx layer.

In previous experiments using a liquid method (solution-processed), the layer was messy and disordered (amorphous), so it needed the extra copper to help conduct electricity.
In this new method (using a spray/sputter technique), the GaOx layer formed with a partially crystalline structure. Imagine the difference between a pile of sand (disordered) and a neatly stacked brick wall (crystalline). The neat, ordered structure of the bricks allowed the electricity to flow naturally, even without the "lubricant" (copper).

The Sodium Question

The researchers also checked if Sodium (a common element in glass) was needed to make the bouncer work. They found that while Sodium helps the main factory floor work better, it doesn't control the bouncer. The GaOx layer works perfectly fine even if you block the Sodium from reaching it. This is important because it means you can use different types of glass or back contacts without worrying about the bouncer failing.

The Results: A Record Breaker

By using this simple, no-extra-ingredients GaOx doorman, the team built a solar cell that achieved:

16% Efficiency: A very high score for this specific type of low-bandgap material.
552 mV Voltage: This is a world record for this specific material (CuInSe2 without Gallium). It proves that the "leaky door" has been successfully fixed.

The One Remaining Hurdle

The paper notes one small issue: The "Fill Factor" (how smoothly the electricity flows) isn't perfect yet. It's like the factory is producing great products, but there's still a little bit of traffic jam inside the building. The researchers identified that the "traffic jam" happens in the space between the layers, not at the back door. Fixing this internal traffic is the next step to making these cells even more efficient.

Summary

In short, this paper shows that you can build a highly efficient solar cell bottom-layer using a simple, smart "bouncer" (GaOx) that stops energy leaks. The best part? You don't need complex extra steps, special copper treatments, or specific sodium levels to make it work. The material organizes itself into a conductive structure naturally, paving the way for cheaper, easier-to-make, super-efficient solar panels for the future.

Technical Summary: GaOx as a Hole Transport Layer for 1.0 eV CuInSe₂ Bottom Cells

Problem Statement
Chalcopyrite solar cells offer a tunable bandgap down to 1.00 eV, making them the ideal candidate for the bottom cell in tandem photovoltaic applications. Achieving this narrow bandgap requires a pure CuInSe₂ (CISe) absorber without Gallium (Ga). However, pure CISe devices suffer from low open-circuit voltages ( $V_{OC}$ ) due to high recombination velocities at the molybdenum (Mo) back contact. While Ga gradients can mitigate these losses, they introduce generation and radiative losses. Alternatively, hole transport layers (HTLs) have been developed to passivate the back contact. Previous research demonstrated that an HTL formed by ion exchange between a CuGaSe₂ precursor and an InOₓ layer results in a GaOₓ layer that passivates the back contact. However, GaOₓ has historically been observed to block current and reduce fill factors (FF) in solar cells with transparent back contacts, particularly when formed on Indium-containing contacts. Furthermore, previous successful implementations of this HTL required a specific copper (Cu) precursor annealing stage and relied on solution-processed InOₓ, raising questions about the necessity of these steps for sputter-deposited, industrially scalable processes. Additionally, the role of Sodium (Na) in enabling conductivity within the GaOₓ layer in this specific configuration remained unclear.

Methodology
The authors investigated the formation and transport properties of a GaOₓ HTL in pure 1.0 eV CuInSe₂ solar cells using a structure of Glass/Mo/HTL/CISe/CdS/i-ZnO/AZO/MgF₂. The HTL was constructed via a double-layer approach: a 120 nm co-evaporated Cu-poor CuGaSe₂ layer followed by a 30 nm InOₓ layer deposited by RF sputtering. The standard three-stage co-evaporation process was then used to grow the 2.2 µm CuInSe₂ absorber.

To isolate the factors enabling GaOₓ conductivity and passivation, three distinct processing variations were compared:

"LT" (Low Temperature): No additional Cu precursor and no high-temperature annealing prior to absorber growth.
"HT" (High Temperature): High-temperature annealing (500 °C) without additional Cu.
"Cu HT": Deposition of a Cu precursor (30–40 nm) followed by annealing at 500 °C, replicating the conditions previously thought necessary for solution-processed HTLs.

Furthermore, the study explored the role of Sodium by depositing samples with and without a SiOₓ barrier to block Na diffusion from the soda-lime glass (SLG) substrate. The composition and crystallinity of the HTL were analyzed using STEM-EDS, SIMS, and HRTEM. Optoelectronic properties were characterized via absolute photoluminescence (PL) to determine quasi-Fermi level splitting (QFLS) and optical diode factors (ODF), while electrical performance was assessed through J-V measurements, series resistance, and electrical diode factor (EDF) extraction.

Key Contributions and Results

GaOₓ Conductivity without Cu Annealing: Contrary to findings with solution-processed HTLs, the study demonstrates that the additional Cu precursor annealing stage is not required to make the sputter-deposited GaOₓ layer conductive. Devices processed without the Cu precursor ("LT") exhibited fill factors and $V_{OC}$ values comparable to or better than those with the Cu precursor. The GaOₓ layer was found to be partially crystalline, a feature attributed to the thermal treatment during absorber growth, which may facilitate charge transport without the need for deep Cu-related defect bands.
Role of Sodium: While Na is essential for improving the bulk quality of the CuInSe₂ absorber (evidenced by higher PL yields in samples with Na diffusion), the study reveals that Na is not necessary for the hole transport properties of the GaOₓ HTL itself. Samples with a SiOₓ Na-blocking layer still achieved effective back-contact passivation and good transport, indicating that the blocking behavior observed in previous bifacial studies is not inherent to this HTL design when sputter-deposited.
Oxygen Flow Independence: Variations in oxygen flow during the InOₓ sputtering process did not show a systematic influence on the passivation or conductivity of the resulting GaOₓ layer.
Device Performance: The optimized "LT" process yielded a record-certified $V_{OC}$ of 552 mV for a pure CuInSe₂ bottom cell without Ga gradients, heavy alkali treatments, or Ag alloying. The device achieved an active area efficiency of 16.4% with a fill factor of 69.9% and a short-circuit current ( $J_{SC}$ ) of 42.7 mA/cm².
Limiting Factors: Despite the high $V_{OC}$ , the fill factor remains constrained by a high electrical diode factor (EDF > 1.6). Analysis comparing the optical diode factor (ODF) of the absorber with the EDF of the finished cell indicates that the primary loss mechanism is Shockley-Read-Hall (SRH) recombination in the space-charge region (SCR), rather than bulk recombination or simple series resistance.

Significance
This work establishes that a sputter-deposited GaOₓ hole transport layer, formed via ion exchange during absorber growth, can effectively passivate the back contact of pure CuInSe₂ solar cells without the need for complex Cu-precursor annealing steps or specific Na doping within the oxide layer. The resulting devices achieve a certified $V_{OC}$ of 552 mV and an efficiency exceeding 16%, setting a new benchmark for Ga-free chalcopyrite bottom cells. These findings are critical for the development of monolithic tandem solar cells, as they validate a scalable, sputter-compatible back-contact architecture that enables the use of the ideal 1.0 eV bandgap absorber while maintaining high voltage and efficient carrier collection. The study also clarifies that the high diode factor limiting the fill factor is a systemic issue related to SCR recombination in these specific architectures, pointing toward future optimization targets.