🔬 optics

Design and Verification of a Terahertz Bandpass Filter using a Spoof Surface Plasmon Polariton Waveguide with Gapped Unit Cells

This paper presents the design, fabrication, and experimental verification of a planar terahertz bandpass filter based on a spoof surface plasmon polariton waveguide with gapped unit cells, which successfully achieves a passband centered at approximately 1 THz by combining the low-pass characteristics of the SSPP structure with the high-pass behavior of periodic gaps.

Original authors: Mohsen Haghighat, Ali Dehghanian, Levi Smith

Published 2026-04-17

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Mohsen Haghighat, Ali Dehghanian, 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

The Big Picture: Building a "Traffic Cop" for Invisible Light

Imagine you are trying to listen to a specific radio station, but your radio is picking up every station at once, plus static, plus the sound of a nearby construction site. You need a filter to let only your favorite station through and block everything else.

In the world of science, this "radio station" is a beam of Terahertz (THz) light. This is a type of invisible light that sits between microwaves (like in your Wi-Fi) and infrared (like in TV remotes). It's super useful for future super-fast internet and medical scanners, but it's very hard to control.

This paper is about building a specialized gatekeeper (a Bandpass Filter) that only lets a specific slice of Terahertz light through (around 1 THz) and blocks everything else. The scientists did this using a clever trick involving "spoof" waves.

The Core Concept: "Spoof" Surface Plasmons

Normally, light bounces off metal like a ball off a wall. But if you carve tiny, specific patterns into the metal, you can trick the light into hugging the surface, flowing along the grooves like water in a riverbed.

The Analogy: Imagine a smooth highway where cars (light waves) zoom straight ahead. Now, imagine you dig a series of deep, narrow ditches along the side of the road. If the cars are the right size, they get "stuck" in the ditches and have to travel along the edge of the road instead of flying off into the sky.
The "Spoof": Real "Surface Plasmons" usually happen with visible light on gold. Since Terahertz light is too big to do this naturally on metal, the scientists carved these "ditches" (grooves) to fake (or "spoof") the effect. They created a custom highway for Terahertz light.

How the Filter Works: The "Gated" Highway

The scientists didn't just build a highway; they built a toll road with specific rules.

The Low-Pass Rule (The Speed Bumps): The grooves they carved act like a speed limit. If the light waves are too "fast" (too high a frequency), they can't fit into the grooves and get blocked. This stops the high-pitched sounds.
The High-Pass Rule (The Gaps): To stop the slow, low-pitched sounds, they added tiny gaps (breaks) in the metal strips. Think of these gaps as a series of small bridges. If the waves are too slow, they can't jump across the gap. They fall off.
The Sweet Spot: By combining the "speed bumps" and the "gaps," they created a zone where only waves with a just-right speed can pass through. This is the Bandpass Filter.

The Design Process: Tuning the Instrument

The paper details how they designed this device, which is like tuning a guitar string:

The Groove Depth ( $H_n$ ): This controls the upper limit (the highest frequency allowed). Deeper grooves let lower frequencies pass; shallower grooves block more. It's like changing the length of a flute to change the highest note it can play.
The Gap Size ( $\delta$ ): This controls the lower limit (the lowest frequency allowed). Making the gap wider or narrower changes how hard it is for slow waves to jump across.
The Number of Units ( $N$ ): They repeated this pattern 8 times. More repetitions make the filter sharper (better at blocking unwanted noise), but it also makes the signal a bit weaker (like walking through 8 doors instead of 1).

The Experiment: Making it Real

Building this wasn't easy. They had to:

Fabricate: They printed these tiny patterns on a super-thin, flexible membrane (like a piece of plastic wrap made of Silicon Nitride) to keep the signal from getting lost in the material.
Generate & Detect: They used a laser and special switches (Photoconductive Switches) to create the Terahertz beam and then "listen" to what came out the other side.
The Result: They tested it, and it worked! The device successfully blocked frequencies below 0.9 THz and above 1.16 THz, letting the "sweet spot" around 1 THz pass through.

Why Does This Matter?

You might ask, "Why do we need a filter for invisible light?"

Future Internet: As we move toward 6G and beyond, we will use Terahertz waves for data. We need filters to separate different channels so your video call doesn't mix with your neighbor's.
Sensing: Different chemicals absorb light at specific frequencies. A filter like this can be tuned to "listen" only to the frequency that a specific gas or toxin absorbs, making for incredibly sensitive detectors.

The Catch (and the Future)

The authors are honest: The filter isn't perfect yet. It loses about 7 decibels of signal strength (it's a bit "muffled"). However, they argue that this is a proof of concept. They proved that you can design these filters using simple geometry rules and that they actually work in the real world.

In summary: This paper is about inventing a new way to build a "traffic cop" for invisible light. By carving tiny, specific patterns into metal on a thin film, they created a device that can pick out a specific frequency from a chaotic mix, paving the way for faster communication and smarter sensors in the future.

1. Problem Statement

Research Gap: While Spoof Surface Plasmon Polariton (SSPP) devices are well-studied for low-loss waveguiding and low-pass filtering at microwave and terahertz (THz) frequencies, there is a significant lack of experimentally validated band-pass filters (BPFs) at THz frequencies (specifically >300 GHz).
Existing Limitations: Previous SSPP-based BPFs often rely on cavity structures (challenging for on-chip integration) or microstrip configurations (prone to high radiation losses at THz frequencies).
Design Challenges: Designing SSPP BPFs is difficult because traditional methods often require cumbersome full-wave simulations to predict both upper and lower cutoff frequencies simultaneously. Furthermore, practical implementation faces hurdles regarding excitation/detection, coupling losses, and high insertion loss due to conductor losses at THz frequencies.
Feedline Integration: Integrating SSPPs with standard feedlines is challenging. Coplanar Waveguides (CPW) often require large flaring grounds, whereas Coplanar Stripline (CPS) offers a more compact, low-loss alternative for THz excitation.

2. Methodology

The authors propose a planar, guided-wave SSPP band-pass filter based on a gapped internal CPS-SSPP unit cell.

Structure Design:
- Substrate: A thin (1 µm) Silicon Nitride (SiN) membrane is used to minimize dielectric loss and dispersion.
- Unit Cell: The core consists of a corrugated CPS line with internal grooves (stubs). Crucially, periodic gaps are introduced in the center of the unit cells.
- Filtering Mechanism: The filter combines two behaviors:
  1. Low-Pass: Provided by the SSPP band-edge (controlled by groove depth, $H_n$ ).
  2. High-Pass: Provided by the series capacitive gaps (controlled by gap width, $\delta$ ).
- Transition Circuit (TC): To bridge the quasi-TEM mode of the CPS feedlines and the quasi-TM mode of the SSPP waveguide, a transition circuit with gradually lengthened stubs is employed to ensure efficient mode conversion and minimize reflection.
Design Procedure:
1. Eigenfrequency Simulation: The authors developed a streamlined design method using unit cell eigenfrequency simulations to independently determine cutoff frequencies.
  - $f_H$ (High cutoff) is primarily controlled by the groove depth ( $H_n$ ).
  - $f_L$ (Low cutoff) is primarily controlled by the gap width ( $\delta$ ).
2. Parameter Selection:
  - Target Center Frequency: ~1 THz.
  - Groove Depth ( $H_n$ ): 40 µm (sets $f_H \approx 1.2$ THz).
  - Gap Width ( $\delta$ ): 3 µm (sets $f_L \approx 0.9$ THz).
  - Number of Cells ( $N$ ): 8 (selected to balance roll-off rate and insertion loss).
3. Simulation Tools: COMSOL Multiphysics (v6.4) was used for both eigenfrequency and frequency domain (S-parameter) simulations. Conductors were modeled as Perfect Electric Conductors (PEC) for dispersion analysis and as Gold (Drude model) for loss analysis.
Fabrication & Measurement:
- Fabrication: The device was fabricated on a SiN membrane. Photoconductive switches (PCS) made of Low-Temperature Grown Gallium Arsenide (LT-GaAs) were bonded directly to the transmission line for THz generation and detection.
- Measurement Setup: A modified Terahertz Time-Domain Spectroscopy (THz-TDS) system was used, featuring a 780 nm femtosecond laser. The temporal response was measured and converted to the frequency domain via Discrete Fourier Transform (DFT).

3. Key Contributions

Novel Topology: Introduction of a gapped internal SSPP unit cell that transforms a standard low-pass SSPP waveguide into a band-pass filter without requiring complex cavity structures.
Simplified Design Workflow: Demonstration of a design procedure where upper and lower cutoff frequencies can be tuned independently using unit cell eigenfrequency simulations, significantly expediting the design process compared to previous methods.
Experimental Validation: This is, to the authors' knowledge, the first experimental demonstration of a gapped SSPP band-pass filter operating in the THz regime (specifically around 1 THz).
Integration Strategy: Successful integration of CPS feedlines with SSPP waveguides on a thin SiN membrane, utilizing direct bonding of PCS to mitigate coupling losses.

4. Results

Performance Metrics:
- Center Frequency: ~1 THz.
- Bandwidth: 0.25 THz.
- Cutoff Frequencies: Measured at 0.91 THz (lower) and 1.16 THz (higher). These align closely with simulation predictions (0.91 THz and 1.2 THz, respectively).
- Insertion Loss (IL): Approximately 7 dB in the passband.
Comparison:
- While the 7 dB loss is higher than microwave counterparts, it is comparable to other SSPP filters at THz frequencies (e.g., 4–10 dB in similar literature). The authors attribute this to inherent conductor losses at THz frequencies.
- The filter successfully rejects out-of-band signals, showing sharp roll-offs at both the lower and higher cutoffs.
Field Analysis: Electric field plots confirmed strong field confinement within the SSPP grooves and gaps, validating the guiding mechanism.

5. Significance

Advancement in THz Components: This work fills a critical gap in the literature by providing a functional, planar BPF for the THz spectrum, a frequency band essential for future short-range data communication and high-resolution sensing.
Design Guidance: The paper provides a clear, reproducible design framework (using $H_n$ for $f_H$ and $\delta$ for $f_L$ ) that can be applied to other SSPP-based filter designs.
Future Potential: Despite the current insertion loss, the authors note that performance can be improved by increasing conductor thickness, optimizing lithography for smaller gaps, or adjusting line widths. The successful proof-of-concept paves the way for more complex integrated THz systems, including sensing applications where specific frequency absorption bands need to be isolated.