🔬 applied physics

Stretch-free, shape-induced 3D Island-Bridge Networks for flexible TFTs on Silicon Planar Technology verified through Bending and Scalability to 9x9 Matrix

This study demonstrates a CMOS-compatible, stretch-free 3D island-bridge architecture on flexible silicon planar technology that protects active thin-film transistors from mechanical stress by localizing strain to concave metal bridges, thereby enabling reliable high-performance operation and scalability to a 9x9 matrix under bending conditions.

Original authors: Daniel Joch, Robert Kammel, Fabian Magerl, Mathias Rommel, Michael P. M. Jank

Published 2026-07-29

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Daniel Joch, Robert Kammel, Fabian Magerl, Mathias Rommel, Michael P. M. Jank

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 world where your electronics aren't stuck inside rigid boxes but can wrap around your wrist, fold into your pocket, or even stick to your skin like a second layer of clothing. This is the dream of "flexible electronics." But here's the catch: the best, fastest, and most reliable computer parts we have today are made of hard, brittle materials like silicon. If you try to bend a piece of silicon, it snaps like a dry twig. For years, scientists have been trying to solve this puzzle: How do we make these tough, high-performance chips work on bendy, stretchy surfaces without breaking them?

The clever solution involves a concept called the "Island-Bridge" design. Think of it like a suspension bridge connecting a series of small, sturdy islands. The "islands" are the hard, fragile computer parts that need to stay perfectly flat and still. The "bridges" are the metal wires that connect them. In this design, the bridges are built to be wiggly and flexible, absorbing all the bending and twisting, while the islands remain completely relaxed and stress-free. This paper explores a new, high-tech way to build these bridges directly into silicon chips, creating a 3D structure that doesn't need to be stretched out first to work.

The Paper's Story: Building 3D Bridges Without the Stretch

This research presents a clever new manufacturing trick to create flexible electronics right on standard silicon chips. The team, working at the Fraunhofer Institute and a university in Germany, wanted to prove that they could build these "Island-Bridge" networks without needing the messy, difficult step of stretching the material first.

The Main Finding: 3D Shapes from Flat Surfaces
The team successfully created a system where tiny, square islands of silicon (each 50 µm × 50 µm) hold active computer parts called Thin-Film Transistors (TFTs). These islands are separated by gaps, or "trenches," that are 40 µm wide. Instead of laying flat wires across these gaps, they built 3D metal bridges that arch up and over the trenches.

How did they do it without stretching? They used a "shape-forming" trick. First, they filled the empty trenches with a temporary, arc-shaped layer of polymer. Then, they deposited metal on top of this curve. When they washed away the temporary filler, the metal was left standing alone as a freestanding, curved bridge. Finally, they etched away the silicon from the back of the chip, leaving the islands and their 3D bridges floating on a flexible layer of polyimide (a tough, plastic-like material). The result is a flexible sheet where the computer parts sit on solid "islands" that never feel the stress of bending, while the metal "bridges" take the hit.

What They Found: Design Matters for Durability
The researchers were very careful to test if this design actually protects the electronics. They didn't just assume the bridges would hold; they tested different sizes to see where the limits lay.

Simulations First: Before building anything, they ran computer simulations to see how much stress the bridges could handle. They found that even when bent into a very tight curve with a radius of just 1 mm, the stress in the metal bridges stayed well below the point where the metal would permanently deform or break. The simulations showed that the stress concentrates at the points where the bridge meets the island, but even there, it wasn't enough to cause damage. They also found that making the trenches deeper (20 µm instead of 10 µm) helped reduce the stress even further.
Real-World Testing: They built the devices and bent them to test them. They used two different bridge widths: a narrow 10 µm bridge and a wider 30 µm bridge.
- The wider bridges (30 µm) were incredibly tough. The computer parts connected by these bridges kept working perfectly even when bent down to a radius of 2 mm. Their performance didn't change much at all.
- The narrower bridges (10 µm) showed some trouble. When bent to a radius of 5 mm, the device started to degrade, and at 2 mm, it stopped working entirely. The authors suggest this wasn't because the design concept was flawed, but likely because a tiny crack formed in one of the narrow metal bridges during the test. This proves that wider bridges are more robust and that narrow ones can fail if defects occur.

Scaling Up: The 9×9 Matrix
To show that this isn't just a one-off trick, the team scaled the concept up to a massive grid. They built a 9×9 matrix (81 islands in a row) where every single island had a working computer part. They were able to address and control each of these 81 devices individually. This proves that the method can be used to make large, complex flexible screens or sensors, not just tiny single devices.

The Bottom Line
This paper suggests that it is possible to manufacture high-performance, flexible electronics using standard silicon technology without needing to stretch the material beforehand. By using these 3D, stretch-free metal bridges, the fragile computer parts are kept safe on their islands while the bridges do all the bending. The results show that with the right design (like wider bridges and deeper trenches), these structures can withstand significant bending without breaking, paving the way for electronics that can truly bend and fold with us.

Technical Summary: Stretch-free, Shape-Induced 3D Island–Bridge Networks for Flexible TFTs on Silicon Planar Technology

Problem Statement
The integration of high-performance active circuit elements onto flexible substrates is a central challenge in flexible electronics, particularly for wearable systems requiring operation on complex or dynamic surfaces. While hybrid electronic concepts utilizing structural design can protect active components from mechanical strain, existing "island-bridge" approaches often rely on transfer printing onto pre-stretched elastomeric substrates. This method presents significant scalability limitations, including difficulties in precise alignment, achieving homogeneous pre-stretch distribution, and maintaining yield as matrix sizes increase. Furthermore, conventional planar techniques struggle to directly fabricate three-dimensional interconnects without pre-stretching, limiting the integration of established inorganic materials into flexible formats.

Methodology
This study addresses these limitations by developing a fully integrated, CMOS-compatible manufacturing route within standard silicon planar technology that eliminates the need for substrate pre-stretching. The methodology involves a "stretch-free" process flow where three-dimensional metal bridges are formed directly into trenches. Key steps include:

Structural Design: An island-bridge architecture where rigid silicon islands (containing active devices) are separated by trenches, and metal interconnects span these gaps.
Shape-Induced Formation: A temporary polymer fill is spin-coated into the trenches to create a concave, arc-shaped surface. Metallization is deposited and patterned on this surface.
Release and Integration: The temporary fill is removed, and the backside silicon is etched away, leaving freestanding 3D metal bridges connecting the islands. Polyimide serves a dual role as both device passivation and the flexible carrier substrate.
Device Integration: Thin-film transistors (TFTs) based on IGZO (Indium Gallium Zinc Oxide) were fabricated on the silicon islands prior to the formation of the 3D bridges.
Analysis: The mechanical behavior was evaluated through numerical simulations (ANSYS) to determine von Mises stress ( $\sigma_{VM}$ ) and plastic strain ( $\varepsilon_{pl}$ ) distributions under bending. Experimental validation involved electrical characterization of single TFTs and a scaled 9×9 matrix under various bending radii (down to 2 mm experimentally).

Key Contributions

Process Innovation: The paper introduces a manufacturing route that integrates 3D metal interconnects directly into silicon planar technology without pre-stretching, overcoming the alignment and yield issues associated with transfer printing.
Mechanical Stress Management: The design localizes mechanical stress to the 3D metal bridges while keeping the silicon islands (and the active TFTs on them) in a strain-free state.
Scalability Demonstration: The concept is successfully scaled from single devices to a 9×9 addressable matrix, proving compatibility with standard lithographic processes.
Design Parameter Identification: The study identifies trench depth ( $D$ ) as a critical design parameter for optimizing stress distribution, showing that deeper trenches reduce peak stress at the bridge-island transitions.

Results

Numerical Simulation: Simulations revealed that peak von Mises stress occurs at the transition between the bridge and the island. For a trench depth of 20 µm, the stress at a bending radius of 1 mm remained well below the material yield strength (545 MPa), with plastic strain ( $\varepsilon_{pl}$ ) calculated at only 0.12%, indicating high mechanical robustness.
Single Device Performance: TFTs on the islands maintained stable electrical performance (threshold voltage $V_{th}$ , subthreshold swing $S$ , and saturation mobility $\mu_{sat}$ ) across bending radii from 100 mm to 5 mm. Devices connected by wider bridges (30 µm) showed consistent performance, while those with narrower bridges (10 µm) exhibited degradation at 5 mm, likely due to bridge defects or cracks rather than intrinsic strain on the island.
Matrix Scalability: A 9×9 matrix of island-bridge structures was fabricated and tested. All TFTs were addressable, with most devices exhibiting $I_{on}/I_{off}$ ratios in the $10^5$ – $10^6$ range. While some parasitic effects and variations in subthreshold swing were observed, the successful addressing of the matrix confirms the feasibility of large-area integration.

Significance
The paper claims that this work verifies the manufacturability of stretch-free, 3D metal bridges fully integrated into silicon planar technology. It demonstrates that the three-dimensional shape, defined by the topography of the concave trench filling, effectively protects active devices from mechanical stress during bending. The results confirm that the polyimide functions effectively as both a passivation layer and a flexible substrate. By successfully integrating active TFTs and scaling the concept to a matrix, the study establishes a viable pathway for high-performance flexible electronics that leverages established inorganic materials and standard fabrication techniques without the complexities of transfer printing or pre-stretched substrates.

More like this