WorksPhysics and Astronomy
Distributed Circuit Model for Predicting the Quality Factor of Magnetic Polariton Resonance
Abstract
Existing RLC circuit model leads to inaccurate predictions of the quality factor (Q-factor) of magnetic polariton (MP) resonances under imperfect absorption conditions due to the omission of radiation loss. Moreover, the lumped-parameter nature of RLC models also limits their applicability for predicting higher-order MP modes. In this letter, we propose a distributed circuit model (DCM) for predicting the Q-factor of MP resonances, which overcomes these limitations. By introducing a radiation resistance to characterize radiation loss and establishing a mapping between distributed and lumped parameters, we derive a unified analytical expression for the Q-factor of arbitrary-order MPs. Validation via rigorous coupled-wave analysis (RCWA) demonstrates that this model can accurately predict the Q-factors of MPs of all orders in various metal-insulator-metal (MIM) structures. This work provides a simple yet effective tool for designing metamaterial emitters/absorbers and advances the understanding of MP loss mechanisms.
thickness , slit length , and slit width . To realize perfect absorption, the structural parameters are chosen as follows: , , , and . Ag is employed for both the metallic strip and substrate, whereas ZnSe is used as the dielectric spacer. The optical response of Ag is described by the Drude model with the following parameters: plasma frequency , scattering rate , and a high-frequency constant .[15] ZnSe is treated as a nondispersive dielectric (refractive index ).[33] The structure is surrounded by free space, and only TM-polarized waves are considered. Figure 1(b) presents the radiative characteristics under normal incidence calculated by RCWA. RCWA calculations employed 181 Fourier orders to ensure convergence. [34] Under this condition, the peak absorptance of the fundamental MP mode (MP1) reaches unity. In Figure 1(b), the diamond denotes the Q-factor predicted by the RLC model proposed by Sakurai.[26] The RLC prediction agrees well with the numerical results computed from RCWA simulation. Because the FWHM is directly related to the resonance frequency and the Q-factor, only the predicted Q-factors are shown in the figure.

Figure 1. Q-factor prediction for MP1 using the RLC model. (a) Schematic of the substrate-supported horizontal MIM structure. (b, c) Q-factors for MP1 obtained from the RLC model without (denoted by diamonds, ◇) and from RCWA (denoted by crosses, +) for slit widths of (b) and (c). (d) Schematic of the substrate-free vertical MIM structure. (e, f) Q-factors obtained from the RLC model with (denoted by circles, ○) and from RCWA (+) for slit widths of (e) and (f), respectively. In the spectral plots, the thin solid blue line, thick solid magenta line, and dashed orange line denote the absorptance , transmittance , and reflectance under normal incidence, respectively.
Increasing only the slit width to μm reduces the MP1 peak absorptance well below unity and produces a clear discrepancy between the RLC model and RCWA, as shown in Figure 1(c). The reason is that the resistance in their model is derived solely from the kinetic impedance associated with the kinetic energy of electrons inside the metal, [26]
Thus, the model contains only the Ohmic kinetic resistance and ignores radiative loss. This approximation is valid near perfect absorption but fails when radiative loss becomes comparable to Ohmic dissipation.
Significant radiative losses are inherent in the substrate-free vertical MIM structure shown in Figure 1(d), where the slit is filled with air and enclosed by Ag on both sides. Consequently, the original RLC model, which neglects radiative losses, yields substantial prediction errors. The key to resolving this issue lies in incorporating radiative losses into the RLC model. To this end, we introduce a radiation resistance , to characterize radiative losses and connect it in series with the kinetic resistance in the original RLC model, as illustrated for the right slit in Figure 1(d). For a series RLC circuit operating at resonance, the Q-factor is conventionally expressed as . By incorporating radiative losses into the circuit model, the quality factor can be rewritten as:
where is the resonance angular frequency of the series RLC circuit . The remaining challenge is to determine the radiation resistance . In a series RLC circuit, the dissipated power is proportional to the corresponding resistance . At resonance, the peak absorptance can be interpreted as the fraction of the total dissipated power converted into Ohmic losses. Therefore, we can reasonably assume that the ratio between the radiation resistance and the kinetic resistance is given by
where is the peak absorptance of the MP resonance. With the kinetic resistance already available and the peak absorptance extracted from the spectral response, the radiation resistance can be readily obtained from Eq. (3).
The RLC model incorporating radiation resistance significantly improves the prediction accuracy. It reduces the relative errors from 348.29% to 11.76% and from 752.13% to 7.65% for To predict the Q-factors of higher-order MP resonances, we extend the DCM in Ref. 17 by incorporating radiative losses, as illustrated for the left slit in Figure 2(a). The distributed kinetic resistance , distributed inductance , and distributed capacitance are taken from Ref. 17. Here, we further introduce a distributed radiation resistance to account for radiative losses. When the th-order magnetic polariton resonance (MP) is excited, current loops with identical magnitudes are formed around the slit,[35] as illustrated in the inset on the right side of Figure 2(b). Each loop can be represented by an equivalent RLC circuit with the same resonance frequency.[30]

Figure 2. DCM for predicting the Q-factors of multi-order MPs. (a) Schematic of the vertical MIM structure without a substrate. (b) Q-factors obtained from the absorptance spectra (denoted by crosses, +) and transmittance spectra (denoted by squares, □), as well as predicted by the DCM with radiation loss (denoted by circles, ○) and without radiation loss (denoted by diamonds, ◇) at , , and with . The inset shows the dimensionless current density distribution for the MP5 resonance. The dimensionless current density is defined as . (c) Results at with ; (d) Results at with .
The key idea is to establish a mapping between the distributed circuit parameters and the corresponding lumped circuit parameters, so that the Q-factors of arbitrary-order MP resonances can be predicted within a unified framework using a single equivalent lumped circuit. For the substrate-free vertical MIM structure shown in Figure 2(a), the resonance condition for the th-order MP mode can be expressed as[17]
where is the resonance order of the MP mode. By drawing an analogy with the resonance frequency of a series RLC lumped circuit , the following expression is obtained:
Accordingly, the expression for the kinetic resistance is obtained as:
Assuming a uniform distribution of loss energy among all current loops, the kinetic resistance and radiation resistance are identical for each lumped circuit. In this case, and still satisfy Eq. (3). Therefore, can be written in a form similar to that of :
where , and is an energy-distribution correction factor. For an approximately uniform distribution of loss among all current loops, . Thus, the mapping between the lumped and distributed circuit parameters is fully established. Substituting these relations into Eq. (2) yields the Q-factor of arbitrary-order MP resonances, which can be further simplified as:
where represents the resonance frequency of the th-order MP mode, which can be determined from Eq. (4). According to Eq. (9), the Q-factors of arbitrary-order MP resonances can be directly determined from the distributed circuit parameters alone.
Next, we validate the accuracy of the proposed model. Since the resonance frequency of the SPP remains unchanged when the period and incident angle are fixed,[36] a relatively small period and slit width , together with a sufficiently large slit length , were chosen to separate the higher-order MP resonances from the SPP resonance frequency, thus minimizing the influence of SPP–MP coupling. The Q-factors obtained from the absorption spectra (+) and the transmission spectra (□) are identical for all MP resonances, as shown in Figure 2(b). Since , the Q-factors extracted from the reflection spectra are also identical. This is because the absorption enhancement, transmission enhancement, and reflection dip all originate from the same MP resonance. In Figure 2(b), the circles correspond to the predictions of the DCM incorporating radiative losses, with . The incorporation of radiation loss reduces the mean relative error (MRE) from 504.85% to 9.14%. A relatively larger discrepancy is observed only for MP5 because of its proximity to the SPP mode.
Further validation is shown in Figure 2 (c) and (Figure 2d). For , remains sufficient. For , is required. When the slit width becomes sufficiently small, the loss distribution deviates from the assumption of uniformity. In fact, for higher-order MP resonances, the total radiative loss arises from the superposition of radiation emitted by all current loops, and radiation from adjacent slits may also interact with each other. Under such circumstances, the proposed approach can be regarded as extracting an effective lumped circuit that incorporates these collective radiative effects, thereby enabling accurate prediction of the Q-factors.
We further apply the proposed model to substrate-supported horizontal MIM structures as shown in Figure 3(a). In this structure, the resonance conditions of the MP modes are identical to those of the vertical MIM structure.[17] Therefore, Eq. (9) is still employed to predict the Q-factors. However, owing to the structural asymmetry introduced by the substrate, the loss distribution inevitably deviates from the assumption of uniformity and must be corrected by the coefficient . The predicted Q-factors of the MP resonances under normal incidence for different geometric configurations are presented in Figure 3. Since only odd-order MP modes can be excited in the horizontal MIM structure under normal incidence, only the prediction results for MP1, MP3, and MP5 are shown. The inset in Figure 3(c) illustrates the dimensionless current density distribution within the structure when MP5 is excited. Similar to the vertical MIM structure, all current loops exhibit nearly identical magnitudes. Ge with is used as the spacer to shift higher-order MPs away from the SPP resonance.[33]

Figure 3. Validation of the DCM predictions for Q-factors of multi-order MPs in substrate-supported horizontal MIM structures. (a) Schematic of the substrate-supported horizontal MIM structure; (b) Results at , , , and slit width with ; (c) Results at with (the inset shows the dimensionless current density distribution for the MP5 resonance); (d) Results at with . All results are obtained under normal incidence. The line styles and symbols are defined in (b) and used consistently in (c) and (d).
The results presented in Figure 3 clearly demonstrate that the incorporation of radiative loss is essential for accurately reproducing the Q-factor in horizontal MIM structures. With , , and fixed, Figure 3(b) presents the prediction results for , for which
where, is the root of the transcendental equation () in the -th period (, ), . Analogously, the expressions for the corresponding lumped circuit parameters of this structure can be readily obtained by replacing in Eqs. (5)–(8) with . Substituting these expressions into Eq. (2) shows that Eq. (9) remains applicable for calculating the Q-factors of arbitrary-order MP resonances. Nevertheless, the substrate-induced asymmetry leads to unequal current loops and nonuniform loss distribution, requiring a correction factor .
The validation results are presented in Figure 4. With and fixed under normal incidence, Figure 4(b) presents the prediction results for , where . The MRE decreases from 254.18% to 7.33% when radiation loss is included. Figure 4(c) shows the corresponding results for , where . When the slit width is further reduced to , the correction coefficient is a negative value, . For all three structures considered, the Q-factor prediction behavior at small slit widths differs significantly from that at larger slit widths. For the substrate-free vertical MIM structure, the value of deviates from that corresponding to a uniform distribution of loss energy when the slit width becomes small. For both the substrate-supported vertical and horizontal MIM structures, becomes negative at small slit widths. At this time, when radiative losses are neglected, the predicted Q-factors are smaller than those obtained from RCWA, indicating that the interaction between radiative and Ohmic losses can no longer be described as a simple linear superposition. Instead, a strong coupling effect emerges between these two loss channels.

Figure 4. Validation of the DCM predictions for Q-factors of multi-order MPs in substrate-supported vertical MIM structures. (a) Schematic of the substrate-supported vertical MIM structure; (b) Results at , , and with ; (c) Results at with ; (d) Results at with . All results are obtained under normal incidence.
Acknowledgment
JMZ thanks the support by the National Natural Science Foundation of China (No. 51976045) and the Young Scientist Studio Foundation of Harbin Institute of Technology.
Data availability statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
References
- [15]B. Zhao and Z. M. Zhang, "Study of magnetic polaritons in deep gratings for thermal emission control," Journal of Quantitative Spectroscopy and Radiative Transfer 135, 81-89 (2014).DOI
- [17]H. Li and J. Zhao, "Predicting Multi-Order Magnetic Polaritон Resonances for Radiative Properties Tailoring by Distributed Circuit Model," Int. J. Heat Mass Transf. 256 (2026).DOI
- [26]A. Sakurai, B. Zhao and Z. M. Zhang, "Resonant frequency and bandwidth of metamaterial emitters and absorbers predicted by an RLC circuit model," Journal of Quantitative Spectroscopy and Radiative Transfer 149, 33-40 (2014).DOI
- [30]L. Long, Y. Yang and H. Ye et al., "Optical absorption enhancement in monolayer MoS2 using multi-order magnetic polaritons," Journal of Quantitative Spectroscopy and Radiative Transfer 200, 198-205 (2017).DOI
- [33]E. D. Palik and G. Ghosh, (Academic Press, San Diego, CA, 1999).
- [34]Y. B. Chen and Z. M. Zhang, "Design of tungsten complex gratings for thermophotovoltaic radiators," Opt. Commun. 269 (2), 411-417 (2007).DOI
- [35]Y. Guo, Y. Shuai and J. Zhao, "Tailoring Radiative Properties with Magnetic Polaritons in Deep Gratings and Slit Arrays Based on Structural Transformation," J. Quant. Spectrosc. Radiat. Transf. 242, 106788 (2019).DOI
- [36]Z. M. Zhang, Nano/microscale heat transfer. (Springer, Cham, 2020).