🔬 applied physics

Demonstration of an Integrated Terahertz Band-Stop Filter Using an Apodized Bragg Grating

This paper demonstrates an on-chip integrated Terahertz Apodized Bragg grating functioning as a band-stop filter with a 0.8 THz center frequency and 200 GHz bandwidth, achieving approximately 20 dB rejection that aligns with theoretical simulations.

Original authors: Ali Dehghanian, Walid Gomaa, Mohsen Haghighat, Thomas Darcie, Levi Smith

Published 2026-03-27

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Ali Dehghanian, Walid Gomaa, Mohsen Haghighat, Thomas Darcie, Levi Smith

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 are trying to listen to a specific radio station, but there's a loud, annoying static noise on the exact same frequency that drowns out your music. You need a special "noise-canceling" device that blocks that specific frequency while letting everything else pass through clearly.

In the world of Terahertz (THz) technology—a super-fast, invisible part of the electromagnetic spectrum used for high-speed communication and medical imaging—scientists have been struggling to build these "noise-canceling" devices. This paper is about successfully building one for the first time on a tiny computer chip.

Here is a simple breakdown of what the researchers did, using everyday analogies:

1. The Problem: The "THz Gap"

Think of the electromagnetic spectrum as a giant piano keyboard. We have great keys for radio (low notes) and visible light (high notes). But there is a weird, silent middle section called the "THz Gap" (between 0.1 and 10 THz).

Why it's hard: We have the "piano" (the waves), but we lack the "fingers" (the devices) to play them. We don't have good, tiny filters to block specific frequencies in this range. Existing filters are either too big, too messy, or they mess up the signal quality.

2. The Solution: The "Apodized Bragg Grating" (TABG)

The team built a filter called a TABG. Let's break down that fancy name:

Bragg Grating: Imagine a hallway lined with mirrors. If you walk down the hallway, some mirrors reflect your image back. If you arrange the mirrors perfectly, they can reflect all the light of a specific color back to you, stopping it from going further. This is a "stop-band filter."
Apodized: This is the secret sauce. In a standard mirror hallway, the reflection creates "echoes" or "ghosts" (called side-lobes) that ruin the clarity. Apodization is like gently fading the mirrors in and out at the ends of the hallway, rather than having them start and stop abruptly. This smooths out the transition, making the reflection cleaner and the "echoes" disappear.

The Analogy:
Think of a crowd of people trying to walk through a narrow hallway.

Without the filter: Everyone rushes through, including the people you want to stop.
With a standard filter: You put up a wall. It stops the bad people, but it also creates a chaotic pile-up (side-lobes) at the entrance and exit.
With the TABG (Apodized): You arrange the people in the hallway so they gently slow down and turn around before hitting a hard wall. The transition is smooth, the crowd doesn't panic, and only the specific "bad" people are turned away perfectly.

3. The Innovation: Putting it on a "System-on-Chip"

Usually, testing these filters requires huge, clunky lab equipment. The researchers did something clever: they built the entire system on a single tiny chip (about the size of a fingernail).

The Chip: They used a super-thin sheet of silicon nitride (like a piece of paper thinner than a human hair) to hold the circuitry.
The Setup: On this one chip, they built:
1. The Transmitter (the speaker sending the signal).
2. The Filter (the TABG hallway).
3. The Receiver (the microphone listening to the result).
Why this matters: Because everything is on one chip, the signal doesn't have to travel through messy cables or connectors. It stays clean and fast. It's like building the speaker, the noise-canceling wall, and the microphone all inside one tiny box, rather than connecting them with long wires.

4. The Results: Did it Work?

The team tested their chip by sending a burst of Terahertz waves through it.

The Goal: Block waves at 0.8 THz (the "bad" frequency) and let everything else pass.
The Outcome: It worked!
- The filter successfully blocked the target frequency by about 20 decibels (which is like turning a loud shout into a whisper).
- The "smooth edges" (apodization) worked perfectly, preventing the messy echoes that usually ruin these filters.
- The real-world test matched their computer simulations almost exactly.

5. Why Should You Care?

This isn't just about blocking one frequency. This is a blueprint for the future.

Faster Internet: Terahertz waves could power 6G and beyond, allowing data to move at lightning speeds.
Better Security: These chips could be used in scanners to see through clothes or packages without using harmful X-rays.
Medical Imaging: They could help doctors see tumors or skin issues with incredible detail.

In a nutshell:
The researchers built a tiny, super-smart "bouncer" for light waves on a microchip. By using a clever "smooth-edge" design (apodization), they taught the bouncer how to stop specific troublemakers (frequencies) without causing a scene (noise), paving the way for the next generation of super-fast, wireless technology.

1. Problem Statement

The Terahertz (THz) frequency gap (0.1–10 THz) holds significant potential for communication, sensing, and imaging. However, the development of integrated THz systems is hindered by a lack of mature passive components, specifically band-stop filters. Existing solutions for planar guided-wave band-stop filters above 500 GHz face several limitations:

Goubau lines: Incompatible with two-terminal elements without mode converters.
Dielectric Bragg gratings: Limited single-mode bandwidth leading to dispersion in wideband systems.
Split-ring resonators (SRRs): Non-systematic design procedures, sub-optimal roll-off rates, and issues with higher-order resonator modes.
Non-planar waveguides: Often bulky or difficult to integrate on-chip.

There is a critical need for a systematic, integrated, planar band-stop filter that offers high rejection, sharp roll-off, and compatibility with on-chip THz System-on-Chip (TSoC) architectures.

2. Methodology

The authors designed, simulated, fabricated, and experimentally characterized an Integrated Terahertz Apodized Bragg Grating (TABG).

Device Architecture:
- The TABG functions as a band-stop filter centered at 0.8 THz with a target bandwidth of 200 GHz.
- It consists of 12 unit cell sections and 4 apodization sections. The apodization gradually tapers the characteristic impedance between the feedlines and the grating to suppress side-lobes and improve the roll-off rate.
- The device is fabricated on an ultra-thin (1 µm) Si₃N₄ membrane to minimize radiation loss and dispersion, enabling wideband (DC – 1.5 THz) operation.
- The structure utilizes Coplanar Stripline (CPS) transmission lines with varying widths ( $W_n$ ) and gaps ( $S_n$ ) to modulate the characteristic impedance ( $Z_n$ ).
Design & Simulation:
- Theoretical Modeling: Used ABCD matrix analysis to derive the dispersion diagram and Bloch impedance. This confirmed the stop-band location and the effectiveness of the apodization profile in matching the load impedance.
- Full-Wave Simulation: Performed using ANSYS HFSS at 0.8 THz, accounting for the Si₃N₄ substrate properties ( $\epsilon_r = 7.6$ ) and gold conductors. The simulation predicted a center frequency of 0.815 THz and a -3 dB bandwidth of 0.22 THz.
Experimental Setup:
- Platform: The TABG was integrated into a THz System-on-Chip (TSoC) containing a transmitter (Tx), the device-under-test (TABG), and a receiver (Rx) on a single wafer.
- Technique: Modified Terahertz Time-Domain Spectroscopy (THz-TDS). A 780 nm, 80-femtosecond laser pulse was split to drive photoconductive switches (LT-GaAs) for the Tx and Rx.
- Measurement: The mechanical delay line was translated to reconstruct the temporal response, which was then converted to the frequency domain via Discrete Fourier Transform (DFT).

3. Key Contributions & Novelty

The paper presents five primary novel contributions:

First Fully Integrated On-Chip TABG: The first demonstration of a complete THz band-stop filter (Tx, feedlines, TABG, Rx) integrated on a single chip within the THz gap.
Highest Bragg Frequency: Achieved the highest Bragg frequency (0.8 THz) to date using quasi-TEM feedlines in the THz gap.
Ultra-Thin Substrate Implementation: The first demonstration of a Bragg grating constructed from guided-wave transmission lines on an ultra-thin Si₃N₄ membrane, solving radiation loss issues common in thicker substrates.
Theoretical Verification: Validated that periodic filter theory (Bloch impedance and dispersion diagrams) is applicable to the design of THz guided-wave gratings.
Radiative Loss Characterization: Identified and illustrated the mechanism of diffractive grating radiation causing insertion loss at frequencies above the Bragg bandgap.

4. Results

Frequency Response:
- Center Frequency: Experimental results aligned closely with theory and simulation, confirming a center frequency of 0.8 THz.
- Bandwidth: The measured -3 dB bandwidth was approximately 200 GHz, matching the design target.
- Rejection: The filter achieved a signal rejection of approximately 20 dB across the stop-band.
  - Note: Simulation predicted ~45 dB rejection, while the experiment showed ~30 dB. The discrepancy is attributed to fabrication tolerances (feature size limit of ~2 µm), material parameter variations, and potential mode coupling.
Temporal Response:
- The experimental temporal pulse showed derivative behavior (high-pass filtering) due to the necessary DC blocking capacitor in the experimental setup, which was absent in the simulation. However, the spectral response remained unaffected in the frequency range of interest.
Insertion Loss:
- Below the Bragg regime, insertion loss was less than 2 dB.
- Above the Bragg regime (0.95–1.5 THz), insertion loss increased from 5 dB to 11 dB due to diffractive grating radiation.

5. Significance

This work represents a significant step forward in THz integrated circuit technology. By successfully demonstrating an Apodized Bragg Grating on a TSoC platform, the authors have provided a robust, systematic solution for frequency filtering in the THz gap. The use of apodization effectively reduces side-lobes and improves the filter's roll-off, addressing a key limitation of previous non-apodized designs.

The integration of the filter into a complete Tx-Rx system on a single ultra-thin substrate validates the TSoC approach for future THz communication and sensing systems, proving that complex passive components can be miniaturized and integrated without sacrificing performance. This paves the way for more sophisticated THz systems capable of operating across wide bandwidths with precise spectral control.