🔬 applied physics

Photolithography-Compatible Three-Terminal Superconducting Switch for Driving CMOS Loads

This paper presents a photolithography-compatible, three-terminal superconducting switch (wTron) capable of directly driving high-capacitance CMOS loads and demonstrating foundry readiness, thereby establishing a critical interface between superconducting electronics and standard semiconductor technologies for next-generation computing and quantum applications.

Original authors: Dip Joti Paul, Tony X. Zhou, Karl K. Berggren

Published 2026-07-16

📖 3 min read☕ Coffee break read

CC BY 4.0

Original authors: Dip Joti Paul, Tony X. Zhou, Karl K. Berggren

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 computers run on superpowers. Not the kind you see in movies, but the real, scientific kind: superconductivity. This is a state where certain materials, when chilled to temperatures colder than outer space, let electricity flow without any resistance or heat loss. It's like a train gliding on a frictionless track, moving forever without needing fuel. Scientists have been building tiny switches using this magic to create super-fast, ultra-low-power computers and sensors that can detect single particles of light.

However, there's a catch. These super-powered circuits are great at doing their own thing, but they are terrible at talking to the "normal" computers we use every day, like the ones in your phone or laptop. Those normal computers are built with a technology called CMOS, which is the industry standard for making chips. The problem is that the superconducting switches are too delicate and weak to push the heavy electrical loads that CMOS chips need to wake up and work. It's like trying to push a giant boulder with a feather. For years, scientists have been trying to build a bridge between these two worlds, hoping to create a hybrid system that combines the speed of superconductors with the practicality of standard chips.

This paper introduces a new kind of bridge called the wTron (short for "wire cryotron"). Think of the wTron as a heavy-duty, three-terminal switch made from a superconducting wire. Unlike its tiny, fragile cousins that are only a few nanometers wide (too small to see without a powerful microscope), the wTron is built on a much larger scale, with wires that are micrometers wide. This size difference is the secret sauce. Because the wires are wider, they can carry much stronger currents—up to several milliamps—making them strong enough to push the electrical signals needed to turn on standard CMOS components.

The researchers showed that they could build these wTrons using standard photolithography, the same printing-like process used to make regular computer chips. This is a big deal because it means these switches don't need expensive, specialized equipment to be made; they can be mass-produced in regular chip factories. In their experiments, the team successfully used a wTron to turn on a bright LED and to flip the switch on a powerful transistor (a MOSFET) with a large electrical "capacitance" of 500 pF. They even managed to make these switches using a professional foundry process from MIT Lincoln Laboratory, proving that wTrons can be integrated alongside other advanced superconducting circuits.

The paper also dives into the math of how to make these switches work best. They found that to drive heavy loads quickly, the wTron needs to have a high electrical resistance when it's "off" (in a resistive state), acting like a strong gate that forces electricity into the load. They used computer simulations to figure out the perfect width for the wires and the right amount of resistance to use, ensuring the switch can turn on and off fast enough to keep up with modern clock speeds. While the wTron isn't as sensitive to tiny signals as its nanoscale predecessor, its ability to drive heavy loads and be made with standard tools makes it a promising candidate for building the next generation of hybrid computers that link the quantum world with the everyday world.

Technical Summary: Photolithography-Compatible Three-Terminal Superconducting Switch for Driving CMOS Loads

Problem Statement
Superconductor electronics (SCE) offer significant advantages for energy-efficient computing, cryogenic sensing, and quantum platforms, particularly through Josephson junction (JJ)-based devices. However, a persistent challenge in the SCE ecosystem is the lack of a robust, on-chip interface between superconducting circuits and Complementary Metal-Oxide-Semiconductor (CMOS) electronics. While JJ-based switches operate at high speeds with ultra-low energy dissipation, they are ill-suited for driving the large resistive or capacitive impedances (often requiring mA-range currents) typical of CMOS loads. Previous attempts to bridge this gap utilized the nanocryotron (nTron), a nanowire-based switch. However, nTrons require electron-beam lithography for fabrication, limiting their scalability and integration with standard foundry processes. Furthermore, their nanoscale dimensions were optimized for single-photon detection readout rather than high-current drive capability, creating a need for a device that combines the magnetic noise resilience and high output impedance of cryotrons with the manufacturability and current drive of micrometer-scale wires.

Methodology
The authors developed and characterized the "wTron," a three-terminal, micrometer-wide superconducting wire-based cryotron switch.

Fabrication: Devices were fabricated using standard photolithography on 10-nm-thick Niobium Nitride (NbN) films deposited on oxidized silicon substrates. The process involved photolithographic patterning (using a maskless aligner), reactive ion etching, and lift-off. The smallest achievable feature width was 1 µm.
Device Structure: The wTron consists of a choke and a channel, both tapered to define switching currents. The choke and channel widths were varied (e.g., 2 µm choke, 4 µm channel) to tune performance.
Characterization: The team performed DC I-V measurements at 3.8 K to map the channel switching current ( $I_{sw,ch}$ ) as a function of gate current ( $I_{gate}$ ). They also measured the switching current distribution of isolated micrometer-wide wires to assess uniformity.
Modeling: An empirical model was developed to describe the suppression of the channel switching current by the gate current, incorporating parameters for current crowding ( $\alpha$ ) and hotspot-induced suppression ( $\beta$ ). A circuit model was implemented in LTspice to simulate capacitive load driving.
Foundry Integration: To demonstrate foundry readiness, wTrons were fabricated using MIT Lincoln Laboratory's SFQ5ee superconducting process, which supports monolithic integration with JJ-based electronics (such as SFQ and AQFP circuits).

Key Results

Switching Behavior: The wTron successfully demonstrated switching from a superconducting to a resistive state. The channel switching current decreased linearly with gate current below the choke's switching threshold and exponentially above it due to resistive hotspot formation. The empirical model (Eq. 1) accurately fitted the experimental data.
Current Drive and Impedance: The wTron exhibited switching currents in the mA range and an output impedance exceeding 1 kΩ. This high impedance and current capability allowed the device to drive loads that nTrons typically cannot.
CMOS Interfacing:
- LED Drive: The wTron successfully drove a room-temperature blue LED (requiring 0.3 mA and 2.9 V), demonstrating gate-controlled switching.
- MOSFET Drive: The device drove a power MOSFET with a 500 pF gate capacitance, turning it on and off by charging the gate capacitance.
Uniformity and Yield: Photolithography-fabricated micrometer-wide wires showed a median switching current density of approximately 40 GA/m² with a standard deviation of ~100 µA. The authors attribute variations primarily to lithography-induced line-edge roughness rather than intrinsic stochastic fluctuations.
Foundry Compatibility: wTrons fabricated in the SFQ5ee process (using a high-kinetic-inductance MoNx layer) exhibited switching behavior consistent with the photolithography-fabricated devices, confirming the feasibility of monolithic integration with JJ-based circuits.

Significance and Claims
The paper claims that the wTron represents a viable, foundry-compatible solution for interfacing superconductor electronics with CMOS. By utilizing photolithography, the wTron overcomes the manufacturing limitations of electron-beam lithography required for nTrons, enabling higher yield and easier integration into standard foundry processes.

The authors assert that the wTron's micrometer-scale dimensions allow for mA-range switching currents and kΩ-range output impedances, making it uniquely suited for driving "current-hungry" resistive loads and highly capacitive CMOS loads (such as MOSFETs, HEMTs, and electro-optic modulators). The work provides a circuit model and design guidelines (optimizing channel width and bias resistance) to minimize latency and maximize voltage swing when driving capacitive loads at MHz frequencies.

Ultimately, the paper positions the wTron as a critical component for developing hybrid cryo-electronic ecosystems. It suggests that monolithic integration of wTrons with JJ-based electronics (like SFQ) and CMOS loads on a single chip can facilitate next-generation computing and quantum applications, bridging the gap between the low-power, high-speed capabilities of superconductors and the scalability of semiconductor technology.

More like this