16-Channel Photonic Chip Targets 3.2 Tbps on One Wafer

Illustrative photorealistic wafer with 16 parallel optical transmission channels; BitcoinVersus.tech green-on-black watermark bottom left

Researchers reported on November 20, 2026, a photonic chip platform capable of sending information across 16 parallel wavelengths at rates associated with 3.2-terabit-per-second optical links. Published in Nature Communications, the laboratory study demonstrates how engineered ferroelectric films can convert electrical data into rapidly varying optical signals. The achievement is a materials-and-integration milestone, not the announcement of a shipping data-center transceiver.

Background: Professor Tetsuya Kawanishi explains electro-optic modulation and Mach–Zehnder devices in an IEICE webinar from 2023. The video predates the new study.

How an electrical signal becomes light

The researchers built photonic integrated circuits using lead zirconate titanate, abbreviated PZT, deposited on silicon-dioxide-coated silicon wafers measuring eight inches across. PZT is ferroelectric: its microscopic electric polarization can be aligned. An applied electric field changes how light propagates through the material via the Pockels effect, allowing electrical signals to control optical phase.

A Mach–Zehnder modulator divides light between two paths, changes their relative phase, then recombines them. The resulting interference creates a data-bearing optical signal. Before modulation, the team electrically aligned PZT’s ferroelectric domains. Their paper specifies a poling field of 26 volts per micrometer for 45 minutes. Domain engineering produced useful modulation behavior without forcing every component to follow one crystallographic direction.

Manufacturing mattered as much as device speed. The group combined chemical-solution deposition and magnetron sputtering with deep-ultraviolet stepper lithography. Measured propagation loss reached 1.1 dB per centimeter at 1,550 nm and 1.6 dB per centimeter at 1,310 nm. Lower propagation loss helps preserve light across the paths and routing structures needed for a complex transmitter.

What 224 Gb/s actually measures

The modulator tests reached 224 gigabits per second using four-level pulse-amplitude modulation, known as PAM4. The arithmetic is straightforward: 112 billion symbols per second × two bits per symbol = 224 Gb/s before accounting for coding overhead. Reported bit-error rates were below the study’s stated forward-error-correction thresholds; the threshold depended on the wavelength band.

The researchers also measured electro-optic frequency response through 70 GHz, the test limit, and reported a voltage–length product as low as 1.19 V·cm at 1,310 nm. That product summarizes modulator drive efficiency and device length; it does not independently establish total electrical power consumption or joules per transmitted bit.

Engineering background: a 2026 tutorial walks through photonic-modulator design and simulation, including Mach–Zehnder models. Not footage of the PZT experiment.

Why sixteen wavelengths matter

A wavelength-division multiplexing transmitter combines many optical carriers into one transmission system. The PZT circuit paired sixteen Mach–Zehnder modulators with an integrated arrayed-waveguide grating at 200-GHz wavelength spacing. The researchers reported PAM4 eye diagrams across all sixteen channels at 224 Gb/s per channel.

Multiplying sixteen by 224 yields 3,584 Gb/s, or 3.584 Tb/s of aggregate raw line rate. The authors describe the device as 3.2-Tb/s-class; neither number should be mistaken for guaranteed user-data throughput after coding, packaging, and complete-link overhead. The published paper and its original Figure 5 show the integrated layout and channel measurements.

The challenge between laboratory and rack

The finding joins a larger race to move data with less electrical overhead. BitcoinVersus.Tech previously examined a 180-Gb/s integrated optical transmitter, a 256-wavelength optical transmission experiment, the engineering tradeoffs between 800G and 1.6T optics, and photonics reliability standards. Those related developments illustrate why capacity, error rates, device losses, and manufacturing reproducibility must be evaluated together.

Important questions remain: power per bit at complete-system level, laser and driver consumption, coupling loss, thermal endurance, packaging yield, fabrication cost, and long-term field reliability. The reported tests offer strong laboratory evidence for PZT as an integrated photonics platform. Commercial deployment will require measurements beyond a research wafer.

Primary source: Guan et al., Nature Communications, published November 20, 2026. Video embeds are background education, not recordings of the study. Editorial cover is an illustration.

We volunteer daily to ensure the credibility of the information on this platform is Verifiably True. If you would like to support to help further secure the integrity of our research initiative, please scan the QR code below.

BitcoinVersus.Tech donation QR code

Leave a Reply