betaFeedback

WorksPhysics and Astronomy

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

Abstract

This paper presents the experimental verification of a planar guided-wave terahertz (THz) spoof surface plasmon polariton (SSPP) bandpass filter (BPF) using a coplanar stripline (CPS) with internal grooves and periodic gaps. The proposed BPF operates by combining the low-pass behavior from the SSPP's band edge and the high-pass behavior from the gaps that act as series capacitors. The higher and lower cut-off frequencies can be tailored by the appropriate selection of the unit cell geometry. For demonstration, a BPF with a center frequency of approximately 1 THz and a bandwidth of 0.25 THz was designed, fabricated, and experimentally validated. The passband around 1 THz is observed in the measurements, along with the lower and higher cut-off frequencies at approximately 0.91 THz and 1.16 THz, respectively, in agreement with simulation results.

Introduction

Spoof surface plasmon polariton (SSPP) devices operating at microwave and terahertz (THz) frequencies have garnered substantial research interest over the past two decades [1, 2, 3, 4, 5, 6]. This attention is primarily due to their inherited properties from conventional surface plasmon polaritons (SPPs), including significant field confinement and customizable dispersion characteristics through geometric modifications [7, 8, 9, 10]. Originally, SSPP structures were introduced using 3D geometries, such as arrays of holes or grooves on a metal surface, interacting with obliquely incident light or waves [1, 2]. Subsequently, the use of guided waves for SSPP excitation was explored due to their improved integration possibilities [3]. A typical guided-wave SSPP structure comprises a corrugated single conductor and a matching or transition circuit (TC).

Most SSPP research has focused on low-loss SSPP waveguides [5] or their low-pass filtering capabilities [11]. Few SSPP-based band-pass filters (BPF) have been reported beyond microwave and millimeter wave frequencies [12, 13, 14, 15, 16, 17, 18]. In general, the mentioned BPFs either use cavities, which make on-chip integration challenging, or use a microstrip configuration, which is susceptible to significant radiation losses on standard substrates at THz frequencies [19]. Furthermore, there is limited research on experimentally validated SSPP-based BPFs at THz frequencies, mainly due to a lack of instrumentation, coupling challenges, and various loss mechanisms [3, 20]. In addition to these challenges in the design of SSPP-based BPFs, practical implementation at THz frequencies is complicated by excitation and detection difficulties. One of the popular feedlines for the excitation of SSPPs is coplanar waveguides (CPW) [5]. Integration of CPW with SSPPs presents challenges, including the need for large flaring grounds to efficiently excite SSPPs [21]. As an alternative, CPS feedlines have demonstrated their effectiveness at THz frequencies in achieving low-loss SSPP excitation [6]. SSPP waveguides that are integrated with CPS can have an internal [22] or external [11] array of stubs, depending on the specific application.

BPFs are one of the key components of front-end systems for wireless communication applications, used for band and channel selection and the rejection of unwanted signals [15, 16]. Therefore, the SSPP-based BPF, with the merits of low loss, broadband capability, and a compact structure, has been widely investigated at microwave frequencies in the past decade [12, 13, 14, 15, 16]. Additionally, BPFs are useful components for sensing applications, targeting a specific range of absorption frequencies [23]. Nevertheless, the literature reveals a gap in the design and experimental demonstration of planar (guided-wave) THz SSPP-based band-pass filters, particularly for frequencies beyond 300 GHz, which have potential applications in future short-range THz data communication and sensing applications.

Our previous work demonstrated a THz BPF that combines the effects of high-pass filtering via a capacitive TC and low-pass filtering by the SSPP waveguide [18]. In [18], we found that the BPF has good performance and can operate in the THz region, but its design procedure can be cumbersome because both the upper and lower cut-off frequencies cannot be predicted by unit cell eigenfrequency simulations. The filter presented in this work can be initially designed using only eigenfrequency simulations of the unit cell, which expedites the design procedure. The SSPP-based THz BPF introduced in this paper is based on a gapped internal CPS-SSPP unit cell. Illustrations of gapped and ungapped unit cells are shown in Table 1. The addition of the gap converts the response from a LPF to a BPF. In this paper, we will investigate the complete BPF structure which means that we must first excite CPS feedlines (quasi-TEM) and then cascade a TC to convert the quasi-TEM mode to the SSPP TM mode [6, 21]. The TC consists of gradually lengthened SSPP stubs until the final maximum depth of corrugation is reached. For a substrate, we use a thin 1 μm silicon nitride (SiN) membrane, which has been shown to reduce loss and dispersion at THz frequencies [6, 19]. Figure 1 shows the BPF, and Figure 2 shows the fabricated SSPP-based BPF on the SiN membrane, along with photoconductive switches (PCS) for the generation and detection of THz signals.

Gapped Internal SSPP (this work)Ungapped Internal SSPP [24]
Unit cell illustration: gg, SS, HH, WW, dd, δ\delta, a/2a/2Unit cell illustration: gg, SS, HH, WW, dd, a/2a/2
Bandpass FilterLowpass Filter

Table 1. Unit cell illustrations

Figure 1. Proposed BPF structure based on SSPP with gapped internal unit cells. In the diagram, N=4N = 4 as an example. Not to scale.

Fabricated SSPP-based BPF with gapped unit cells on a thin Si-N membrane excited by CPS feedlines

Figure 2. Fabricated SSPP-based BPF with gapped unit cells on a thin Si-N membrane excited by CPS feedlines. Key dimensions: Period d=20 μmd = 20\,\mu\mathrm{m}, SSPP groove width a=10 μma = 10\,\mu\mathrm{m}, unit cell gap width δ=3 μm\delta= 3\,\mu\mathrm{m}, SSPP horizontal gap g=10 μmg = 10\,\mu\mathrm{m} (separation between top and bottom stubs), and groove depth Hn=40 μmH_n = 40\,\mu\mathrm{m}.

Table 2 provides a comparison of several SSPP-based BPFs with experimental results and our work. We were unable to find experimental results for many other SSPPs-based BPFs that were verified at THz frequencies; therefore, in Table 2, our structure clearly exhibits the highest insertion loss (IL) compared to its microwave counterparts that operate at less than 1/10th of the frequency. Also, the insertion loss is comparable to our previous design at 1 THz [18]. This increased insertion loss is primarily due to increased conduction losses at THz frequencies. We will briefly address the high insertion loss by comparing the SSPP-based BPF with the SSPP-based LPFs found in the literature. In [25], they measure the insertion loss to be between 5-10 dB at 0.275 THz. In [26], they measure a 4 dB insertion loss at 0.14 THz. Therefore, we note that our proposed filter has an insertion loss comparable to other SSPP filters and operates at THz frequencies. We clarify that the goal of this research is not to propose a high performance optimized filter; this paper aims to demonstrate that the proposed proof-of-concept filter exhibits bandpass behavior at THz frequencies and to provide the reader with design guidance.

Referencef0f_0 (GHz)FBWMin. IL (dB)
[12]5120%0.37
[13]1036%1.60
[14]343%1.00
[15]8545%0.58
[16]3318%0.50
[17]106%3
[18]100030%7
This Work100025% *7 **

Table 2. Comparison of Experimental Results for SSPP-based BPFs

  • This FBW is related to the fabricated and measured structure, but it can be changed with geometry modifications.

** Simulated values considering all loss sources and with δ=3\delta= 3 µm and N=8N = 8.

Future work can improve the insertion loss by modifying the geometry. For example, simply increasing the conductor thickness will improve the insertion loss.

Simulations and Design

Simulation details

In the following sections, eigenfrequency and frequency domain (FD) simulations are performed using COMSOL version 6.4 for design and analysis purposes. The eigenfrequency simulations were performed on the SSPP unit cell. The FD simulations were performed on the complete structure. In all cases the following parameters are used: d=2a=20 μmd = 2a = 20\,\mu\mathrm{m}, W=10 μmW = 10\,\mu\mathrm{m}, and g=10 μmg = 10\,\mu\mathrm{m}. The thickness of the SiN membrane is 1 μm1\,\mu\mathrm{m}, relative permittivity εr=7.6\varepsilon_r = 7.6, conductivity σSiN=0\sigma_{\mathrm{SiN}} = 0, and loss tangent tan⁡δe=0.00526\tan\delta_e = 0.00526 [27]. The 200 nm thick conductors were modeled as a perfect electric conductor (PEC) or as gold using the Drude model with ε=1−ωp2(ω2+jω/τ)−1\varepsilon= 1 - \omega_p^2(\omega^2 + \mathrm{j}\omega/\tau)^{-1}, where: ωp=2π(2.2 PHz)\omega_p = 2\pi(2.2\,\mathrm{PHz}) and τ=18 fs\tau= 18\,\mathrm{fs} [28]. Modeling the conductors as PEC or gold (Drude) was selected depending on the aim of the simulation, and the material is noted in the figure captions when relevant.

Theory and dispersion analysis

In this section, we present a design procedure for the proposed SSPP BPF, which is based on the simulated dispersion analysis of the unit cell that forms the core of the filter. The initial geometric parameters for the unit cell are selected using design criteria found in the literature [5, 11, 22]. The unit cell period, dd, should be much smaller than the wavelength, d≪λ0d \ll\lambda_0 [2, 6]. In this work, we operate near 1 THz, so d≪300 μmd \ll300\,\mu\mathrm{m}. In addition, dd should be large enough to ensure that the entire unit cell has sufficient resolution using photolithography fabrication; our photolithographic resolution is 3 μm3\,\mu\mathrm{m}. In this work, we select d=20 μmd = 20\,\mu\mathrm{m}. Next, the aperture, aa, can be selected between 0.1d<a<0.9d0.1d < a < 0.9d, but selecting a=0.5d=10 μma = 0.5d = 10\,\mu\mathrm{m} results in low loss and good performance. The width of the gaps, δ\delta, primarily controls the lower cutoff frequency. The details of this parameter are discussed later, but we select δ=3 μm\delta= 3\,\mu\mathrm{m}, which corresponds to our limitation in feature size via photolithography. The gap between the internal SSPP stubs, gg, does not have a large impact on the filter performance, but it is important that gg is equal to or greater than the conductor spacing at the Tx PCS (see Figure 2) to ensure that dielectric breakdown does not occur when applying a bias voltage. Thus, we select g=10 μmg = 10\ \mu\mathrm{m} to satisfy this criterion, but larger values could also have been chosen if desired.

The design procedure starts by using an eigenfrequency simulation of the unit cell to determine the cutoff frequencies. Figure 3(a-Figure 3d) illustrates the simulated dispersion diagram for Hn=30H_n = 30, 40, 50, and 60 μm\mu\mathrm{m}, respectively. The dispersion curves start at the lower cutoff point, fLf_L, and end at the higher band edge frequencies, fHf_H. fHf_H is determined by the SSPP band-edge defined as the frequency where kd=πkd = \pi, which is mainly controlled by HnH_n [6, 22]. fLf_L is obtained from where the dispersion curve for the CPS mode (similar to a light line with εr,eff=2.0\varepsilon_{r,\mathrm{eff}} = 2.0 [18]) intersects the SSPP mode. These frequencies align with the cut-off frequencies observed from the S-parameters obtained from full-wave simulations that are shown in Figure 4. We clarify that in Figure 4 we simulate: Port 1 →\to CPS feedlines →\to TC →\to SSPP unit cells → TC → CPS feedlines → Port 2. This is required to ensure that the SSPP mode is excited. Also, for this simulation, the conductors were modeled as PECs to clearly illustrate the cutoff frequencies. Both Figure 3 and Figure 4 show that changing HnH_n changes both fLf_L and fHf_H.

Image not extracted

Figure 3. Dispersion curves of gapped SSPP unit cell for Hn=30H_n = 30, 40, 50, and 60 μm\mu\mathrm{m} showing bandpass filtering behavior. Conductors are modeled as PECs.

Figure 4. Simulated S21S_{21} for the gapped SSPP BPF with Hn=30H_n = 30, 4040, 5050, 60 μm60\,\mu\mathrm{m}. For this simulation δ=3 μm\delta= 3\,\mu\mathrm{m} and N=8N = 8. Conductors are modeled as PECs.

The high cut-off frequency, fHf_H, is primarily controlled by HnH_n [2, 6]. The relationship between these variables is shown in Figure 5a, which was constructed from an eigenfrequency simulation where W=10 μmW = 10\,\mu\mathrm{m}, d=2a=20 μmd = 2a = 20\,\mu\mathrm{m}, and g=10 μmg = 10\,\mu\mathrm{m}. The black dots on Figure 5 correspond to our experimental structure. For this proof-of-concept filter, we select Hn=40 μmH_n = 40\,\mu\mathrm{m}, which corresponds to a fHf_H of 1.2 THz. The lower cut-off frequency, fLf_L, is controlled by varying HnH_n and δ\delta. However, we found that δ\delta can independently adjust fLf_L without significantly impacting fHf_H. Therefore, the bandwidth (BW) of the filter can be adjusted by changing δ\delta, and consequently varying fLf_L as shown in Figure 6. The relationship between δ\delta, HnH_n, and fLf_L is plotted in Figure 5b, which was constructed from an eigenfrequency simulation where W=10 μmW = 10\,\mu\mathrm{m}, d=2a=20 μmd = 2a = 20\,\mu\mathrm{m}, and g=10 μmg = 10\,\mu\mathrm{m}. In summary, to design a gapped SSPP unit cell BPF, first use Figure 5a to find the HnH_n that corresponds to the desired fHf_H. Next, use Figure 5b to find the δ\delta that corresponds to the desired fLf_L. Figure 7(a-Figure 7d) illustrates the simulated S-parameters for δ=0.5\delta= 0.5, 1, 3, and 5 μm\mu\mathrm{m}, respectively, with Hn=40 μmH_n = 40\,\mu\mathrm{m}. In Figure 7 it is observed that δ\delta can be used to independently control fLf_L of the BPF without significantly affecting fHf_H. The results presented in Figure 7 are in good agreement with the dispersion analysis shown in Figure 6.

Image not extracted

Figure 5. Unit cell design for higher and lower cutoff frequencies. For these curves: d=2a=20d = 2a = 20 µm, W=10W = 10 µm, and g=10g = 10 µm. Conductors are modeled as PECs. (a) Determining the required HnH_n to achieve a specified fHf_H. (b) Determining the required δ\delta to achieve a specified fLf_L.

Image not extracted

Figure 6. Simulated dispersion curves for different δ\delta parameters to design the lower cut-off of the proposed BFP without affecting the higher cut-off for the structure with Hn=40H_n = 40 µm. Conductors are modeled as PECs.

Image not extracted

Figure 7. Simulated S-parameters for different δ\delta which illustrates the change in the lower cut-off frequency of the BPF without affecting the higher cut-off with Hn=40H_n = 40 µm. In all cases the conductors are modeled as PECs and the passband insertion loss is approximately 1 dB.

Upon completion of the design of the unit cells, the number of cascaded unit cells, NN, must be determined. Figure 8 plots the SS-parameters for N=4N = 4, 6, 8, and 10. In all of these simulations, the conductors were modeled as gold (the aforementioned Drude model). When selecting NN, there is a trade-off between the insertion loss and the roll-off rates. As NN increases, the roll-off rates will increase; however, the passband insertion loss also increases, as illustrated in Figure 8. The lower roll-off rate is dependent on the number of periods, NN, which should be selected based on the requirements for the specific application. For our proof-of-concept design, we selected N=8N = 8 to achieve acceptable roll-off rates and insertion loss (≈7\approx7 dB). Using these parameters, we can safely anticipate that our experiment will be able to resolve the transmitted signal [18]. In Figure 8, there is a rounding of ∣S21∣2|S_{21}|^2 near fHf_H when conductor loss is included. This is expected because the group velocity trends towards zero at the band edge, and thus frequencies near the band edge experience greater interaction with the conductor and therefore exhibit increased attenuation.

Figure 8. Simulated S-parameters for different NN which illustrates the change in the lower cut-off frequency roll-off rate and insertion loss with Hn=40H_n = 40 µm. In all cases the conductors are modeled as gold (Drude). When N=4N = 4 the passband IL is -4 dB. When N=10N = 10 the IL is -8 dB.

An approach to reducing passband insertion loss, with a fixed NN, is to decrease δ\delta. Note, this also affects the lower cutoff frequency, which must be considered during design iterations. We selected δ=3\delta= 3 µm, which corresponds to the minimum feature size available in our fabrication process. We note that it is possible to achieve δ≤3\delta\le3 µm and reduce the insertion loss with more precise lithography and fabrication methods. Also the conductor thickness can be increased to reduce conductor loss. Lastly, increasing the width of the CPS lines, WW, and the separation between the CPS lines, SS, can help reduce the insertion loss, but it can result in higher radiation losses depending on the application.

To visualize the filtering operation, electric field plots for the BPF with Hn=40H_n = 40 µm are plotted in Figure 9 at 1 THz (in the passband), 0.5 THz, and 1.5 THz (the lower and higher stopbands, respectively). These plots also illustrate the strong field confinement capability of the proposed SSPP structure inside the SSPP unit cell gaps and grooves.

Electric field plots of the proposed THz SSPP band-pass filter

Figure 9. Electric field plots of the proposed THz SSPP band-pass filter with Hn=40H_n = 40 µm at 0.5 THz (lower stopband), 1 THz (passband), and 1.5 THz (higher stopband), illustrating passband transmission and out-of-band rejection.

We briefly comment on the transition circuit. The core of the structure with fixed HnH_n should not be directly connected to the CPS feedlines because there is a large mode mismatch between the CPS (quasi-TEM) and the SSPP (quasi-TM). Therefore, a TC with a gradual increase in the depth of the SSPP grooves is required [6]. In this work, we used a TC with stepped linear growth up to HnH_n for the fabricated device, where each TC stub with a height of HiH_i ($1 \le i \le 7)isrepeatedfivetimesand) is repeated five times andH_i increases linearly in each section (see Figure 1) to reach the maximum depth of SSPP (i.e., HnH_n). The reason for choosing this TC configuration is its ability to provide more rejection at stopband frequencies through the repetition of stubs, as each stub length corresponds to the rejection of a single frequency. We note that our selected method is not a strict requirement, and the selection of transition circuits is somewhat flexible, provided there is a gradual introduction of stub lengths.

The design procedure is summarized in the following steps:

  1. Specify the filter requirements: fHf_H, fLf_L, and IL. We selected fH=1.2f_H = 1.2 THz, fL=0.9f_L = 0.9 THz, and IL = 7 dB.

  1. Determine HnH_n for the specified fHf_H. The relationship between HnH_n and fHf_H is determined by an eigenfrequency simulation. We provide Figure 5a which was generated when δ=3\delta= 3 µm. Note that δ=3\delta= 3 µm can be used as an initial value for most designs (see Figure 6 when kd=πkd = \pi). However, δ\delta will be modified in the next step.

  1. Determine δ\delta for the specified fLf_L. This can be performed using an eigenfrequency simulation.

  1. Determine NN. Perform a frequency domain simulation on the complete filter to determine the number of gapped unit cells, NN. The resultant S-parameters quantify passband insertion loss and filter performance. Figure 8 illustrates the effect of NN on the S-parameters, where it is shown that with N=8N = 8, we achieve good roll-off above and below the passband.

Experimental Setup and Fabrication of PCS

The experimental setup used for the measurements is based on a modified THz time-domain spectroscopy (THz-TDS) system, as outlined in [19], and illustrated in Figure 10. This system features a femtosecond pulsed laser operating at a wavelength of 780 nm, with a pulse width of 90 fs, a repetition rate of 80 MHz, and an average output power of 27 mW. The laser beam is focused onto photoconductive switches (PCS) located at the transmitter (Tx) and receiver (Rx) positions on the SiN membrane (see Figure 2). These switches function as sources and detectors of THz signals through the processes of photoconductive switching and sampling [19].

Figure 10. Measurement Setup

The PCS's measure 70 μm×40 μm×1.5 μm70\ \mu\mathrm{m} \times40\ \mu\mathrm{m} \times1.5\ \mu\mathrm{m}, with a gap of 5 μm5\ \mu\mathrm{m} between the gold contacts. They are bonded using water droplets through Van der Waals (VDW) forces, as detailed in [29]. The transmitted signal is reconstructed by translating the mechanical delay line and measuring the receiver current with a lock-in amplifier, following the methods used in standard THz-TDS setups [6, 19].

The fabrication process for the photoconductive switches (PCSs) involves several steps. First, a low-temperature grown gallium arsenide (LT-GaAs) layer is deposited on a sacrificial aluminum arsenide (AlAs) layer, which rests on a semi-insulating GaAs substrate. Next, photolithography is applied to the LT-GaAs surface to create gold (Au) contacts. Each PCS region is masked and subjected to wet etching with citric acid and hydrogen peroxide to define the PCS thickness. After cleaning and re-masking with etch-resist wax, the LT-GaAs layer is dissolved in hydrofluoric acid (HF), separating it from the AlAs layer. Further steps are taken to detach the remaining LT-GaAs film, resulting in multiple active regions of LT-GaAs PCSs [30].

Experimental Results and Discussion

The experimental results of the THz BPF that uses gapped unit cells with Hn=40 μmH_n = 40\ \mu\mathrm{m} are presented in Figure 11, showing both the temporal responses of the BPF and a reference CPS (transient output pulses). A reference device is a CPS feedline with the same length as the SSPP BPF. The temporal response of the reference exhibits an anticipated broadband pulse with a single peak, while the BPF response manifests as a narrowband pulse containing multiple oscillations. Additionally, Figure 11 illustrates the corresponding spectral responses on a single subplot (left side) derived from applying the discrete Fourier transform (DFT) to the temporal responses. The spectral response of the BPF exhibits an abrupt roll-off at the expected cut-off frequencies of 0.9 THz (simulation predicted 0.91 THz) and 1.15 THz (simulation predicted 1.2 THz). The higher band-edge frequency aligns with the SSPP band-edge as predicted by dispersion curves (Figure 3b) and S-parameter simulations (Figure 8c).

Figure 11. Measurement results of the proposed BPF with Hn=40 μmH_n = 40\ \mu\mathrm{m} versus a reference CPS with the same length

Going forward, we believe that the proposed gapped unit cell SSPP BPF will be a useful addition to the THz community due to the straightforward design procedure that we presented and its performance characteristics at THz frequencies. While we recognize that a 7 dB insertion loss is non-ideal, we reiterate that there are few planar/non-planar alternatives (with experimental results at THz frequencies). Also, the proposed integrated filter has minimal coupling loss since the PCS’s are directly connected to the transmission line, which mitigates the broadband coupling loss that other technologies must manage properly.

A couple of comments on interpreting the experimental results. First, it is important to address why the experimental spectral responses exhibit amplitude decay with respect to increasing frequency. This occurs because the transient signal has a finite duration, which implies that the spectral response must decay with increasing frequency. This means that the dominant frequency dependent ‘loss’ for the spectral response traces in Figure 11 is not truly ‘loss’; it simply indicates that the generated and detected signal is frequency dependent. Regardless, there is sufficient information contained within the received signal to demonstrate that signal filtering is occurring, which is in agreement with the filter design. Second, when investigating BPF’s, the detected signal strength is reduced compared to other configurations (i.e., the reference, low-pass, high-pass, etc.) because the majority of the source signals spectral content is removed by the BPF by the time the signal reaches the receiver. While generally this is advantageous and is the purpose of a BPF, it presents a minor adverse artifact during the experiment because the signal-to-noise ratio decreases. This is observed when comparing the temporal responses for the signals in Figure 11 since the reference signal has less noise than the filtered signal. This is a challenge that is observed during experimentation and can make system alignment more difficult.

Conclusion

In this work, we introduce a novel terahertz (THz) band-pass filter (BPF) based on spoof surface plasmon polariton (SSPP) with gapped unit cells, developed from coplanar strips (CPS) with internal stubs. The filter achieves high-frequency rejection through low-pass SSPP characteristics and effectively blocks low frequencies through gaps within the structure. By adjusting the geometrical parameters, particularly the δ\delta and HnH_n parameters, the higher and lower cut-off frequencies and the bandwidth of the filter can be customized, providing flexibility in design. Experimental validation of the fabricated BPF, with a center frequency of approximately 1 THz, demonstrates promising performance, with measured passband transmission closely matching the simulations regarding cut-off frequencies. This validates the effectiveness of the proposed SSPP structure in achieving the desired bandpass filtering functionality. This work highlights the potential of SSPP-based filters in advanced THz applications, such as sensing, with the ability to tailor frequency responses through simple geometrical modifications. To our knowledge, this is the first experimental demonstration of a gapped SSPP filter at THz frequencies.

Acknowledgments

This work was supported by an NSERC Discovery Grant. The authors thank 4D LABS at Simon Fraser University for the fabrication of the waveguides and the thin membrane, and also the Center for Advanced Materials and Related Technology (CAMTEC) at the University of Victoria for providing Nanofab facilities for the fabrication of the PCS devices.

Funding This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) (RGPIN-2022-03277).

Availability of data and materials: Data supporting the findings of this study are available from the corresponding author, L.S., upon reasonable request.

Declarations

Ethical Approval Not applicable.

Competing Interests The authors declare no competing interests.

References

  1. [1]J. B. Pendry, L. Martín-Moreno, and F. J. Garcia-Vidal. Mimicking surface plasmons with structured surfaces. Science, 305(5685):847–848, 2004.pubmed.ncbi.nlm.nih.gov/15247438
  2. [2]F. J. Garcia-Vidal, L. Martín-Moreno, and J. B. Pendry. Surfaces with holes in them: new plasmonic metamaterials. Journal of Optics A: Pure and Applied Optics, 7(2):S97, 2005.
  3. [3]Wen Xuan Tang, Hao Chi Zhang, Hui Feng Ma, Wei Xiang Jiang, and Tie Jun Cui. Concept, theory, design, and applications of spoof surface plasmon polaritons at microwave frequencies. Advanced Optical Materials, 7(1):1800421, 2019.DOI
  4. [4]Kai-Da Xu, Ying Jiang Guo, and Xianjin Deng. Terahertz broadband spoof surface plasmon polaritons using high-order mode developed from ultra-compact split-ring grooves. Optics Express, 27(4):4354, February 2019.DOI
  5. [5]Muhammed Abdullah Unutmaz and Mehmet Unlu. Terahertz Spoof Surface Plasmon Polariton Waveguides: A Comprehensive Model with Experimental Verification. Scientific Reports, 9(1):7616, May 2019.
  6. [6]M. Haghighat, T. Darcie, and L. Smith. Demonstration of a terahertz coplanar-strip spoof-surface-plasmon-polariton low-pass filter. Scientific Reports, 14, 2024.
  7. [7]Xiaopeng Shen, Tie Jun Cui, Diego Martin-Cano, and Francisco J. Garcia-Vidal. Conformal surface plasmons propagating on ultrathin and flexible films. Proceedings of the National Academy of Sciences, 110(1):40–45, 2013.DOI
  8. [8]Stefan A. Maier, Steve R. Andrews, L. Martín-Moreno, and F. J. García-Vidal. Terahertz Surface Plasmon-Polariton Propagation and Focusing on Periodically Corrugated Metal Wires. Physical Review Letters, 97(17):176805, October 2006.DOI
  9. [9]Stefan A Maier et al. Plasmonics: fundamentals and applications, volume 1. Springer, 2007.DOI
  10. [10]Paloma Arroyo Huidobro, Antonio I. Fernández-Domínguez, John B. Pendry, Luis Martín-Moreno, and Francisco J. Garcia-Vidal. Spoof Surface Plasmon Metamaterials. Elements in Emerging Theories and Technologies in Metamaterials. Cambridge University Press, 2018.
  11. [11]Ying Jiang Guo, Kai Da Xu, and Xiaohong Tang. Spoof plasmonic waveguide developed from coplanar stripline for strongly confined terahertz propagation and its application in microwave filters. Optics Express, 26(8):10589, April 2018.DOI
  12. [12]Ying Jiang Guo, Kai Da Xu, Yanhui Liu, and Xiaohong Tang. Novel surface plasmon polariton waveguides with enhanced field confinement for microwave-frequency ultra-wideband bandpass filters. IEEE Access, 6:10249–10256, 2018.
  13. [13]Jun Wang, Lei Zhao, and Zhang-Cheng Hao. A band-pass filter based on the spoof surface plasmon polaritons and cpw-based coupling structure. IEEE Access, 7:35089–35096, 2019.DOI
  14. [14]Yiwen Wei, Yongle Wu, Weimin Wang, Leidan Pan, Yuhao Yang, and Yunan Liu. Double-sided spoof surface plasmon polaritons- line bandpass filter with excellent dual-band filtering and wide upper band suppressions. IEEE Transactions on Plasma Science, 48(12):4134–4143, 2020.DOI
  15. [15]Yiqun Liu, Kai-Da Xu, Jianxing Li, Ying-Jiang Guo, Anxue Zhang, and Qiang Chen. Millimeter-wave e-plane waveguide bandpass filters based on spoof surface plasmon polaritons. IEEE Transactions on Microwave Theory and Techniques, 70(10):4399–4409, 2022.DOI
  16. [16]Yinian Feng, Kai-Da Xu, Zhongqian Niu, Bo Zhang, Lujie Liu, and Yong Fan. Ka-band waveguide bandpass filters using double-layer grounded spoof surface plasmon polaritons. IEEE Transactions on Microwave Theory and Techniques, pages 1–11, 2024.DOI
  17. [17]Bocong Ren, Weiwen Li, Zhaozhao Qin, Yu Wang, Liangcai Zhang, and Baoping Zhang. Leaky wave antenna based on periodically truncated sspw waveguide. Plasmonics, 15(2):551–558, 2020.DOI
  18. [18]Mohsen Haghighat and Levi Smith. A terahertz bandpass filter using a capacitive transition circuit and a spoof surface plasmon polariton waveguide. Journal of Infrared, Millimeter, and Terahertz Waves, 47(4):25, 2026.DOI
  19. [19]Levi Smith and Thomas Darcie. Demonstration of a low-distortion terahertz system-on-chip using a cps waveguide on a thin membrane substrate. Optics Express, 27(10):13653–13663, May 2019.DOI
  20. [20]Ho-Jin Song and Namyoon Lee. Terahertz communications: Challenges in the next decade. IEEE Transactions on Terahertz Science and Technology, 12(2):105–117, 2022.DOI
  21. [21]Muhammed Abdullah Unutmaz, Gulay Ozsahin, Tuna Abacilar, and Mehmet Unlu. Investigation of the Transitions for Coplanar Waveguide to Terahertz Spoof Surface Plasmon Polariton Waveguides. IEEE Transactions on Antennas and Propagation, 70(4):3002–3010, April 2022.
  22. [22]Mohsen Haghighat, Ali Dehghaniian, Thomas Darcie, and Levi Smith. Demonstration of terahertz spoof surface plasmon polariton waveguides using coplanar striplines with internal corrugations. Opt. Express, 32(27):48803–48814, Dec 2024.DOI
  23. [23]D. Lee, J. Kang, J. H. Lee, H. Kim, C. Kim, J. Kim, T. Lee, J. Son, Q. Park, and M. Seo. Highly sensitive and selective sugar detection by terahertz nano-antennas. Scientific Reports, 5, 2015.arxiv.org/abs/1506.08363
  24. [24]Y. Guo, K. Da Xu, and X. Tang. Spoof plasmonic waveguide developed from coplanar stripline for strongly confined terahertz propagation and its application in microwave filters. Opt. Express, 26(8):10589, Aug. 2018.DOI
  25. [25]Muhammed Abdullah Unutmaz, Korkut Kaan Tokgoz, and Mehmet Unlu. A complete circuit model for terahertz spoof surface plasmon polariton waveguides for ultrafast and accurate synthesis of terahertz integrated circuits. IEEE Access, 2025.
  26. [26]Kai-Da Xu, Dongxu Wang, Ying-Jiang Guo, Yiqun Liu, and Pin Wen. A terahertz on-chip lowpass filter based on spoof surface plasmon polaritons. IEEE Electron Device Letters, 2025.DOI
  27. [27]Giuseppe Cataldo, James A. Beall, Hsiao-Mei Cho, Brendan McAndrew, Michael D. Niemack, and Edward J. Wollack. Infrared dielectric properties of low-stress silicon nitride. Optics Letters, 37(20):4200–4202, Oct 2012.DOI
  28. [28]Mark A Ordal, Robert J Bell, Ralph W Alexander Jr, Larry L Long, and Marvin R Querry. Optical properties of Au, Ni, and Pb at submillimeter wavelengths. Applied optics, 26(4):744–752, 1987.DOI
  29. [29]E. Yablonovitch, D. M. Hwang, T. J. Gmitter, L. T. Florez, and J. P. Harbison. Van der Waals bonding of GaAs epitaxial liftoff films onto arbitrary substrates. Applied Physics Letters, 56(24):2419–2421, 06 1990.DOI
  30. [30]Rubén Darío Vásquez Ríos, Siméon Bikornimana, Muhammad Ali Ummy, Roger Dorsinville, and Sang-Woo Seo. A bow-tie photoconductive antenna using a low-temperature-grown gaas thin-film on a silicon substrate for terahertz wave generation and detection. Journal of Optics, 17(12):125802, 2015.

Paper details

vlnc:W_8mk24jmj
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents