Photonic crystal resonators offer a powerful platform for frequency conversion by leveraging resonance effects and wave interference. In this work, we utilize the sideband of the Lorentz-line shape or other line shapes of a source wave to excite the fundamental resonance of a Kagome-lattice photonic crystal resonator, leading to beat note generation in the THz range. Through numerical simulations and analytical modeling, we demonstrate the feasibility of this approach, highlighting its potential for efficient THz wave generation and frequency conversion.
2.1. Sideband Interactions and Beat Note Generation
We harness the sideband of the Lorentz-line shape of a source frequency to excite the fundamental resonance of a photonic crystal resonator. The influence of an external field on the resonator mirrors its effect on artificial atoms, driving effective charge oscillations as described by the Lorentz model [
22]. Through edge effects or sideband interactions, a single-frequency wave interacting with a structured medium or resonant system can give rise to additional frequency components [
22,
23]. When the central peak frequency of the input wave differs from the resonator’s fundamental resonance, their interaction within the cavity generates beat notes in the THz range. To validate this phenomenon, we employ numerical simulations of Maxwell’s equations using the plane-wave expansion method [
24], the finite-difference time-domain (FDTD) technique [
25], and analytical modeling. This method offers a versatile framework for frequency generation and conversion, extending its applicability across diverse classical systems, including electronics, acoustics, mechanics, electromagnetism, optics, and quantum technologies.
In the proposed structure, the generated beat note
corresponds to the absolute frequency difference between the source wave frequency
and the cavity’s resonance frequency
, expressed as
The power-conversion efficiency (PCE), denoted as
, can be determined using the following expression:
where
and
are the electric field amplitudes of the source wave and the net converted frequency wave, respectively. The input wave frequency and the resonator’s resonance frequency lie within the photonic band gap (PBG), whereas the resulting beat frequency is significantly lower and falls outside the PBG. Consequently, the photonic crystal (PhC) functions as a low-pass filter. To extract the THz wave, an optical amplitude demodulator is placed at the center of the PhC resonator, converting the beat note into an independent THz signal that radiates freely from the structure.
To achieve efficient terahertz operation, Kagome-shaped PhC resonators or cavities are designed using silicon (Si) rods arranged in air, allowing interaction with source light in the infrared communication waveband. In the proposed approach, the beat note arises from the interference between the source wave and the sideband-induced resonance wave within the PhC cavity. For optimal interference, the group velocity of the source wave must be minimized. Ideally, the source wave should be confined within the cavity with zero group velocity, which is achieved by tuning the cavity resonance frequency to closely match the source wave frequency. As a result, the source wave remains nearly resonant with the cavity, exhibiting an almost zero frequency group speed [
26].
2.2. Lorentzian Spectral Distribution and the Edge Effect
In conventional scenarios, the sidebands of a single-frequency input wave are often regarded as an insignificant or undesired component of the signal, contributing little to practical applications. However, in this work, we take a different approach by harnessing these sidebands to excite a photonic crystal resonator, thereby inducing a distinct and observable resonance mode. This resonance mode then interacts with the input wave, leading to the generation of beat notes, which are instrumental in our proposed frequency conversion mechanism. To illustrate the fundamental principles underlying this phenomenon, we draw an analogy to a mechanical oscillator consisting of a charged particle attached to the end of a spring. When subjected to an external electric field, the charged particle undergoes oscillatory motion, which can be described by the following equation of motion:
where
r is the displacement of the oscillating particle in the oscillator,
t is time,
m and
q are the mass and charge of the particle,
=
,
is the damping coefficient,
is the amplitude of the source electric field,
=
, and
K is the elastic coefficient. This equation has the following steady-state solution:
where
. Equation (
4) shows that the displacement amplitude reaches its maximum when the source field frequency is equal to the resonance frequency of the oscillator system. However, the oscillation frequency of the particle is strictly determined by the frequency of the applied electric field, meaning that the oscillator itself does not inherently generate any new frequencies beyond that of the source wave. In this scenario, the driving electric field is assumed to be a perfectly monochromatic, single-frequency wave with an idealized zero bandwidth.
In practical scenarios, however, no real-world source can be considered a truly ideal single-frequency wave. Instead, all single-frequency sources exhibit a spectral profile with a finite bandwidth. A common example is the Lorentzian spectral distribution, which characterizes the frequency spread of many resonant systems. This spectral profile accounts for the natural linewidth of the source, which arises due to inherent physical limitations such as damping, phase noise, and energy dissipation. The Lorentzian spectral shape can be mathematically expressed as follows:
Here, represents the central frequency of the Lorentzian spectral profile, while defines the full-width at half-maximum (FWHM), characterizing the spectral bandwidth. This implies that a so-called single-frequency wave is, in reality, a broadband signal encompassing a range of frequencies.
The Lorentzian function does not vanish at frequencies far from
; instead, it has long tails, meaning that there are always small but nonzero spectral components at frequencies away from
. The term “edge effect” refers to the fact that the energy distribution extends beyond
, particularly at the edges of the spectrum where the amplitude is still nonzero. This means that even though we refer to the wave as “single-frequency”, in reality, it consists of multiple frequency components. These sideband components can interact with a resonant system, leading to frequency mixing, beat notes, and new observable resonances. The Lorentzian function, which attains a peak value of
, can be normalized, so that its integral over all frequencies sums to unity:
Therefore, when considering the edge effect of the Lorentz-line shape of a practical driving source, (
3) is modified as follows:
where
is a constant specifying the amplitude of the total field, and
is a normalization factor ensuring the correct amplitude scaling of the spectral distribution. Equation (
7) reveals that a nominally single-frequency source can, in reality, induce oscillations at an infinite number of additional frequencies due to the edge effect associated with the Lorentzian spectral shape of the source. This effect arises because practical single-frequency waves exhibit a finite spectral width rather than being perfectly monochromatic. In general, the Lorentzian spectrum is sharply peaked around its central frequency and decays rapidly away from it, making these additional frequency components difficult to observe. However, when the oscillator’s resonance frequency coincides with a significant spectral component of the source, the resulting motion at that frequency can become pronounced. This effect is particularly noticeable when the oscillator or resonator has a high-quality (
Q) factor, allowing it to sustain and amplify resonant oscillations, and when the coupling between the source and the oscillator is strong.
Since the resonance component is significantly stronger than other frequency components, except for the fundamental source frequency, Equation (
7) can be approximated as follows. For simplicity, we assume that the oscillator supports only a single dominant resonance mode:
Equation (
8) clearly illustrates that under the excitation of a realistic single-frequency field, a new frequency component emerges, corresponding precisely to the resonance frequency of the oscillator. This forms the fundamental principle of resonance-induced mode generation, which plays a crucial role in this work as one of the contributing waves for producing the beat note. It is important to note that the two terms on the right-hand side of Equation (
8) represent the dominant Fourier components of the integral expression in Equation (
7). In particular, the second term in Equation (
8), which arises from the sideband of the Lorentzian spectrum, acts as an additional driving term for the oscillator. This contrasts with a damping term, which typically manifests as an exponential decay factor in the system’s response. From Equation (
4), it can be observed that when the source frequency is close to the resonance frequency, the oscillator exhibits a strong response due to resonance amplification. Moreover, the quality factor Q of the resonator is inversely proportional to the damping coefficient
. Thus, a high-Q resonator can sustain an observable response even when the source frequency is significantly detuned from the resonance frequency. This suggests that in systems with a high-quality factor, an externally applied field with a frequency far from the resonator’s natural frequency can still excite a detectable resonance-induced frequency component, but conversion efficiency decreases.
As is well known, an amplitude demodulator extracts the envelope of a modulated wave. In this context, the THz beat note, arising from the interference of two infrared waves with slightly different frequencies, can be converted into a standalone THz wave using an amplitude demodulator. Notably, the input source wave and the cavity’s resonance modes coexist within the same spatial region, enabling effective interference and beat note generation [
27]. The electric fields of the two interfering waves can be expressed as follows:
where
correspond to the source and resonance mode, respectively. Here,
denotes the amplitude of the source and resonance mode,
represents the transverse field profile, and
is the propagation constant. The coordinate
x indicates the direction of the wave transmission. The two waves together in the cavity, and thus the total field in the cavity, can be expressed as follows:
Without loss of generality, we assume
—i.e., the source and the induced resonance fields have equal amplitudes. Furthermore, based on the same symmetry, we can assume
. Therefore, we obtain
where
Since the two waves coexist spatially and interfere within the same region, we can express the resulting field as follows:
where
n is the effective refractive index of the region where the beating occurs, and
c is the speed of light in free space. Considering the general case where
, the slowly varying amplitude of the resulting beat wave can be expressed as
This leads to the following expression:
where
From Equations (
19) and (
20), it is evident that the frequency of the beat note is
, not
. This is because the absolute value of the cosine function effectively doubles the frequency component present in the cosine term describing the amplitude of the beating wave in Equation (
19). Consequently, an amplitude demodulator can be employed to extract the beat note and generate an independent wave with a frequency equal to the difference between the two input source waves. Based on Equation (
19), we can write the following:
where
is the series coefficient.
This phenomenon has important implications for novel applications in terahertz wave generation, an area that has remained largely unexplored. Traditionally, the spectral bandwidth of a laser wave or the sidebands of a practical sinusoidal wave have been considered negligible or without practical use. However, our analysis suggests that these sidebands can serve as effective excitation sources for frequency conversion. While the above discussion assumes that the spectral profile of a practical single-frequency source follows a Lorentzian distribution, the principle remains valid for other spectral shapes, such as Gaussian profiles. This is because all such spectral line shapes inherently exhibit sideband components, which can similarly contribute to frequency generation, leading to equations analogous to Equation (
8).
2.3. Design and Configuration of Kagome-Lattice Photonic Crystal Resonator
Frequency conversion in a Kagome-lattice photonic crystal resonator is achieved through the interaction between the input source wave and the cavity resonance, generating a new frequency. In this section, the design and configuration of the Kagome-Lattice photonic crystal resonator in
Figure 1 are discussed. It is important to observe here that by optimizing the PhC structure and cavity parameters, the source wave is strongly confined within the PBG, enabling efficient frequency mixing. The resulting beat frequency is extracted as a new THz wave, demonstrating the potential of PhC-based resonators for advanced optical signal processing. This work focuses on optical frequency conversion using a cavity resonator, designed within a two-dimensional (2D) Kagome-lattice photonic crystal. The PhC is formed by periodically arranging silicon (Si) rods (refractive index 3.4) in an air background with a lattice constant of 0.6 μm. High-quality factors are ensured by incorporating more than five reflection rows of unit cells in the resonators. Optimal coupling is achieved by positioning the source at the center of the resonators.
The designed structure of the Kagome-Lattice photonic crystal resonator for the generation of new frequency is shown in
Figure 1, where the resonator cavity consists of a single cavity. The cavity in the resonance mode was identified near a wavelength within the PhC, so in the PBG, the waves are predominately confined inside the cavity. In the design, the source is a TE-polarized wave operating at 1.5 μm within the infrared communication band. The cavity exhibits natural resonance modes, which are intentionally designed to differ from the input source wave’s frequency. To determine these resonance modes, simulations are conducted using an impulse source and analyzed through FFT within the FullWAVE module of RSoft, (Synopsys Inc., Sunnyvale, CA, USA), employing the finite-difference time-domain (FDTD) method for accurate results. Once the resonance characteristics of the structure are determined, the identified resonance mode is utilized as one of the key components in generating the beating phenomenon. Within the cavity, the interaction between the input source wave and the resonance mode leads to the formation of a THz beating wave. This beating wave can then be processed and converted into an independent THz wave (T-ray) using an infrared (IR) amplitude demodulator. It is important to explain the phenomenon of self-induced beating. Assume, without loss of generality, that the external field has a central frequency
, slightly detuned from the cavity’s intrinsic resonant frequency. This is illustrated in
Figure 2, where the blue curve represents the spectral profile of the external field (centered at
), and the red curve corresponds to the cavity spectrum. When the source field is coupled into the cavity, an induced field (red curve in
Figure 2) is generated. The interference between the source and induced fields produces a beat note. For the induced field’s intensity to be comparable to the source field, the central frequencies of both spectra must be very close. This small frequency difference is key to the proposed method, as it allows an infrared source and a suitably tuned infrared-wave cavity to generate a THz beat frequency, where the required THz frequency corresponds to the slight difference between the external and intrinsic frequencies. In order to further justify the function of the PhC cavity, it is important to observe that the cavity facilitates beat frequency generation through spectral overlap between the input source wave and the cavity mode. When two slightly detuned waves interact—one following the source field spectrum (blue curve,
Figure 2) and the other aligning with the cavity resonance (red curve,
Figure 2)—their interference produces a modulated intensity pattern, generating a THz beat frequency. The cavity enhances this process by confining the optical field, ensuring efficient interaction. The optical amplitude demodulator then extracts the THz signal by detecting intensity modulation, demonstrating the viability of PhC-based frequency conversion for advanced optical signal processing.
Establishing a clear connection with the mechanical oscillator analogy discussed in
Section 2.1, particularly the transition from Equation (
7) to Equation (
8), requires emphasizing that the source field spectrum is not purely monochromatic but instead exhibits a finite bandwidth, as illustrated in
Figure 2. The Lorentzian function describing the source field does not vanish at frequencies away from its central frequency, leading to small but nonzero spectral components across a range of frequencies. These sideband components can interact with the resonator, giving rise to additional frequency contributions beyond the fundamental mode. In Equation (
8), we approximate the system response by selecting the dominant frequency components from Equation (
7), justified by the fact that a high-Q resonator can sustain a measurable response even for source frequencies detuned from its natural resonance [
28]. This aligns with the physical behavior observed in
Figure 2, where both the source field and cavity profile exhibit finite spectral widths, reinforcing the necessity of considering a broader frequency distribution rather than assuming a strictly monochromatic excitation. The subsequent analysis demonstrates that, with appropriately chosen parameters, T-rays can be effectively generated from near-infrared (NIR) waves, highlighting the feasibility and practical application of this frequency conversion approach.
For effective wave confinement within the cavity, the wave frequency must fall within the PBG of the PhC. A wider bandgap enhances confinement, improving the overall performance. To optimize this, the plane-wave expansion method was utilized alongside the BandSOLVE module in RSoft to scan the PhC parameters. This analysis revealed a relatively large bandgap for the TE wave in the Kagome-lattice PhC, achieved at a rod radius of
, where
a denotes the lattice constant. A normalized frequency of
was selected for the operation. To demonstrate the validity and general applicability of the method, the source frequency was not positioned at the center of the bandgap. However, for practical applications, selecting a source frequency at the middle of the PBG can enhance the cavity’s quality factor, leading to stronger wave confinement. To further analyze the performance, simulations were conducted using the Fullwave FDTD module in RSoft for the structure illustrated in
Figure 1. The impulse mode and continuous wave (CW) mode in RSoft are utilized to simulate and analyze resonance characteristics and the evolution of the electric field, respectively.
The bandgap of the PhC structure can be tailored through design modifications, enabling precise control over light propagation and manipulation within the material. Additionally, reducing or eliminating TM bandgaps enhances the confinement and control of TE-polarized light in targeted regions, improving efficiency, signal quality, and overall device performance. The gap-to-mid-gap ratio (GMR), which quantifies the bandgap size, is determined to be
from
Figure 3. However, an optimal GMR value does not exist universally, as it varies based on the specific application, balancing performance requirements, fabrication complexity, and material constraints. In practical applications, the following empirical guidelines are commonly used to determine suitable GMR values:
0.1 < GMR ≤ 0.2: Sufficient for most low-index contrast systems and simple devices.
0.2 < GMR ≤ 0.3: Ideal for many integrated photonic devices (waveguides, resonators).
GMR > 0.3: Optimal for advanced applications requiring high confinement or broadband operation.
In the following section, simulations of the designed Kagome-lattice photonic crystal resonator configuration are presented. The FullWAVE module in RSoft is utilized to obtain results for the structure shown in
Figure 1. The simulation setup includes perfectly matched layers to prevent boundary reflections, while small time steps and finely tuned geometrical grid sizes are implemented to ensure stability and high precision. A detailed discussion of the simulation process and results follows in the next section.