🔬 mesoscale physics

Gate induced strain on a two-dimensional hole gas in silicon

This paper demonstrates that increasing aluminum gate thickness induces strain in a silicon two-dimensional hole gas, leading to the emergence of a second subband and revealing distinct cyclotron masses that deviate from ideal heavy-hole/light-hole models due to the combined effects of quantum confinement, strain, and band mixing.

Original authors: D. van der Bovenkamp, C. S. A. Müller, B. D. Pantiru, I. Bošnjak, M. Cignoni, Q. Torrent Nicolau, M. E. Bal, S. Wiedmann, J. Ridderbos, F. A. Zwanenburg

Published 2026-07-10

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: D. van der Bovenkamp, C. S. A. Müller, B. D. Pantiru, I. Bošnjak, M. Cignoni, Q. Torrent Nicolau, M. E. Bal, S. Wiedmann, J. Ridderbos, F. A. Zwanenburg

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 you have a tiny, flat highway for tiny particles called "holes" (which act like positive charges) running through a block of silicon. Usually, these holes travel in a single, crowded lane. But in this experiment, the scientists wanted to see if they could force a second lane to open up just by pressing down on the road with a metal gate.

Here's how they did it: They built a special gate out of aluminum and placed it right on top of the silicon highway. When they cooled everything down to a chilly 1.3 K (that's colder than outer space!), the aluminum gate shrank more than the silicon underneath it because they expand and contract at different rates. This created a "squeeze" or strain on the silicon, kind of like stepping on a soft mattress.

The big question was: How much does this squeeze change the traffic?

The Main Discovery
The team found that by simply making the aluminum gate thicker, they could change the rules of the road.

Thin Gates (20 nm): When the gate was thin, the holes behaved like they were in a single lane. The data showed one clear rhythm in their movement.
Thick Gates (60 nm and up): As they made the gate thicker (up to 100 nm), the "squeeze" got stronger. Suddenly, the holes started behaving as if a second lane had opened up. The traffic pattern changed from a simple rhythm to a complex "beating" pattern, which is what happens when two different rhythms overlap.

This means that by just changing the thickness of the metal gate, they successfully forced the holes to occupy a second energy level (or subband) that wasn't there before.

The Speed and Weight of the Holes
To understand what was happening, the scientists measured how heavy these holes felt as they spun around in a magnetic field (a property called "cyclotron mass").

For the first lane, the holes felt like they had a mass of 0.36 ± 0.04 times the mass of a free electron.
For the second lane, they felt heavier, with a mass of 0.49 ± 0.02 times the mass of a free electron.

The paper points out that these numbers don't match the "textbook" weights we expect for heavy or light holes in a perfect, unstressed world. Instead, the "squeeze" from the gate mixed the two types of holes together, creating a new, hybrid state. It's like if you mixed red and blue paint; you don't get a pile of red paint and a pile of blue paint anymore, you get purple. The holes became a mix of "heavy" and "light" characters, and the gate thickness controlled how much mixing happened.

What They Ruled Out
The scientists were very careful to explain what wasn't causing these changes:

It wasn't just a simple split: They ruled out the idea that the second signal was just a "spin splitting" (a tiny separation of the same lane) or a harmonic echo of the first lane. The gap between the two signals was too big for that.
It wasn't a single lane: While the total number of holes (measured by the Hall effect) looked similar across all gate thicknesses, the detailed rhythm of the resistance showed that for thicker gates, the traffic was definitely split between two distinct groups moving at different speeds.

How Sure Are They?
The team is quite confident in their measurements. They didn't just guess; they measured the resistance of the material while spinning it in a strong magnetic field of 9 Tesla (which is huge!). They used a mathematical tool called a Fast Fourier Transform (FFT) to turn the wiggles in the resistance into clear peaks, showing exactly where the two lanes were.

They measured the temperatures (ranging from 1.3 K to 55 K) and the magnetic fields to confirm the "weights" of the holes.
They simulated the strain using computer models (finite-element simulations), which suggested that the aluminum gate creates a specific type of "squeeze" (biaxial strain in the middle and shear strain at the edges) strong enough to reshape the energy landscape by a few millielectronvolts (meV).

Why It Matters
This isn't just about traffic jams; it's about building future computers. The paper suggests that because this "squeeze" can be controlled locally by the gate, we might be able to tune these silicon systems to be better at handling quantum information (qubits). By adjusting the gate thickness, we can engineer the exact conditions the holes need to behave in specific ways, potentially making them less sensitive to noise and easier to control with electricity.

In short, the paper shows that a simple piece of aluminum, when made thick enough, can act like a master switch, opening a second lane for holes in silicon and mixing their properties in a way that could be very useful for the next generation of quantum technology.

Technical Summary: Gate-Induced Strain on a Two-Dimensional Hole Gas in Silicon

Problem Statement
Hole spin qubits in silicon (Si) offer advantages such as suppressed hyperfine interaction and strong spin-orbit interaction (SOI), enabling electric driving without micromagnets. However, realizing a robust two-level quantum system in Si is challenging due to the fourfold degeneracy at the top of the valence band. Unlike germanium (Ge), where strain can lift this degeneracy to create a pure heavy-hole (HH) ground state, Si typically relies on quantum confinement. In Si, confinement alone often results in significant mixing between heavy holes (HH) and light holes (LH), making it difficult to isolate a pure HH state. Furthermore, while strain is a powerful tool for tuning valence band structures and increasing SOI, applying it in Si is difficult because Si devices generally do not utilize the heterostructure lattice mismatch strategies common in Ge/SiGe systems. The challenge lies in finding a method to apply local, tunable strain in standard Si metal-oxide-semiconductor (MOS) structures to engineer the valence band and control HH-LH mixing.

Methodology
The authors fabricated Hall bar devices on intrinsic silicon with a 100 nm SiO₂ dielectric. To induce local strain, they utilized the mismatch in thermal expansion coefficients between silicon and aluminum (Al). Aluminum gates of varying thicknesses (20 nm to 100 nm) were deposited via e-beam evaporation and patterned using optical lithography. The devices underwent a high-temperature steam anneal to passivate Si/SiO₂ interface defects.

The study employed magnetotransport measurements at cryogenic temperatures (1.3 K and 1.7 K) with magnetic fields perpendicular to the device. Key characterization steps included:

Hall Measurements: Determining carrier density ( $n_H$ ) and mobility ( $\mu$ ) as a function of gate voltage ( $V_g$ ) and gate thickness.
Shubnikov-de Haas (SdH) Oscillations: Measuring longitudinal resistance ( $R_{xx}$ ) and Hall resistance ( $R_{xy}$ ) to observe quantum oscillations.
Data Analysis:
- Subtracting background magnetoresistance using spline fits to isolate oscillatory components.
- Performing Fast Fourier Transforms (FFT) on the oscillations to identify frequency components corresponding to different subbands.
- Analyzing the temperature dependence of the FFT amplitudes using the Lifshitz-Kosevich (LK) formula to extract effective cyclotron masses ( $m^*_c$ ).
- Fitting the oscillatory data with a two-carrier model for thicker gates to account for multiple subband occupations.

Key Results

Strain-Induced Subband Occupation: Increasing the aluminum gate thickness, which increases the compressive strain in the channel, leads to the emergence of a second subband. For thin gates (e.g., 20 nm), SdH oscillations are dominated by a single frequency ( $f_1 \approx 19$ T). As the gate thickness increases to 60 nm and 100 nm, a beating pattern appears in the oscillations, and a second distinct frequency ( $f_2 \approx 66$ T) emerges in the FFT spectrum.
Effective Cyclotron Masses: Temperature-dependent analysis yielded two distinct effective cyclotron masses: $m^*_c = (0.36 \pm 0.04)m_0$ for the first subband ( $f_1$ ) and $m^*_c = (0.49 \pm 0.02)m_0$ for the second subband ( $f_2$ ). These values differ from idealized bulk HH or LH masses, reflecting the renormalization of the band structure due to quantum confinement, strain, and HH-LH mixing.
Carrier Density and Conductivity: While the total Hall carrier density ( $n_H$ ) remains relatively consistent across different gate thicknesses at a fixed gate voltage, the contribution of the second subband to conductivity increases with gate thickness. The second frequency's associated density closely tracks the total Hall density in thicker gates, indicating that the second subband becomes the dominant conduction channel as strain increases.
Strain Profile: Finite-element simulations suggest the Al gate induces biaxial strain underneath the gate and shear strain at the edges. This strain profile is sufficient to modify the HH/LH energy landscape on the meV scale, consistent with the observed reconstruction of the valence band.

Significance and Claims
The paper demonstrates that local gate-induced strain, controlled simply by the thickness of an aluminum gate, is an effective method for engineering the valence band structure of a silicon 2D hole gas. The authors claim that this approach successfully tunes the occupation of subbands and alters the effective masses through strain-induced HH-LH mixing.

The significance of this work lies in providing a straightforward fabrication-compatible method to control the quantum mechanical properties of holes in Si without relying on complex heterostructures. By showing that gate thickness can induce a transition from single-subband to two-subband transport and modify effective masses, the study offers a potential tuning parameter for optimizing hole spin qubits. Specifically, the ability to manipulate HH-LH mixing and subband occupation could be useful for tailoring g-factors and driving speeds in larger qubit arrays, addressing the challenges of anisotropy and noise sensitivity in silicon-based quantum devices.

More like this