Coherent perfect absorption of anti-modes in an indirect coupled magnon-polariton system

Abstract

In this work, we report coherent perfect absorption (CPA) of anti-modes in an indirectly coupled magnon--polariton system. By examining both single and indirectly coupled cases, we experimentally distinguish the modal decay rate $γ$ from the effective decay rate $γ_{\rm{eff}}$. At CPA, $γ_{\rm{eff}} = 0$, leading to a vanishing output and a visually narrow spectrum in the dB-scale, while the intrinsic linewidth set by $2γ$ remains unchanged, demonstrating that the effective decay rate dictates the spectral amplitude rather than the physical loss. Furthermore, in the indirectly coupled system, CPA persists over a broad, magnetically tunable detuning range, in contrast to the single-detuning CPA observed in the directly coupled case, thereby enabling magnetically reconfigurable and frequency-selective microwave absorbers.

Figures (8)

• Figure 1: (a),(e) Schematics of the single-resonator and traveling wave-mediated indirectly coupled two-resonator systems under two identical coherent inputs. (b),(f) Calculated total output spectra $|S_{\mathrm{tot}}|^{2}$ for the single and indirectly coupled systems under CPA conditions, based on Eqs. (\ref{['E5']}) and (\ref{['E14']}), respectively. The vanishing $\gamma_{\mathrm{eff}}$ is indicated by the solid arrows (taken from 3 dB above the minimum), while the FWHM indicates the background response is marked by the open arrows. (c) Single-resonator CPA device consisting of a YIG sphere on a transmission line. The bias field $H$ tunes the Kittel-mode frequency. (d) The signal from port 1 of the VNA are split by a symmetric power splitting network to generate two coherent counter-propagating inputs that excite the resonator(s) in the DUT. The outgoing waves are extracted using directional couplers and received at port 2. (g) Indirectly coupled device with two YIG spheres. A global field $H$ is applied to both spheres, while a local tuning field $\Delta H$ is applied to YIG 1 to control the frequency detuning between the two magnon Kittel-modes.
• Figure 2: Comparison of spectra for the single resonator system under non-CPA and CPA configurations. Shown are the linear- (a),(e) and dB-scaled (b),(f) total output spectra $|S_{\mathrm{tot}}|^{2}$, the inverse spectra $1/|S_{\mathrm{tot}}|^{2}$ (c),(g), and the absorption spectra (d),(h). Panels (a--d) correspond to the non-CPA case with the $S_{2+}$ port terminated ($p=0$), while panels (e--h) show the CPA case under dual excitation ($p=1$). Experimental data are shown in gray, and theoretical calculations based on Eqs. (\ref{['E6']}) and (\ref{['E5']}) are plotted as black curves. In the non-CPA configuration, $2\gamma$ is extracted from the FWHM of the Lorentzian dip in the linear-scale total-output and absorption spectra (open arrows), which also sets the background response in the dB-scale spectrum. $2\gamma_{\mathrm{eff}}$ is obtained from the FWHM of the corresponding peak in the inverse spectrum (solid arrows) and remains finite, consistent with the incomplete absorption in (d). Physically, this indicates that part of the input power is not absorbed by the oscillator but remains in the output channel, yielding an output power proportional to $(\gamma_{\mathrm{eff}}/\gamma)^{2}$. Since $\gamma$ is fixed for a given system, tuning $\gamma_{\mathrm{eff}}$ via input control directly regulates the absorption and consequently the output. In the CPA configuration, $\gamma$ remains unchanged, while the total output is fully suppressed and the absorption reaches unity. The zero-output feature is most clearly revealed in the dB-scaled spectrum (f) as an ultra-sharp dip. Correspondingly, the inverse spectrum (g) exhibits a vanishing FWHM, indicating $\gamma_{\mathrm{eff}}=0$, consistent with the zero 3 dB-above-minimum linewidth (solid arrows).
• Figure 3: Comparison of spectra for the indirectly coupled system under zero and large detuning. Shown are the linear- (a),(e) and dB-scaled (b),(f) total output spectra $|S_{\mathrm{tot}}|^{2}$, the inverse spectra $1/|S_{\mathrm{tot}}|^{2}$ (c),(g), and the absorption spectra (d),(h). Experimental data are shown in gray, while theoretical results based on Eqs. (\ref{['E13']}) and (\ref{['E14']}) are plotted as black curves. At zero detuning (non-CPA regime), a single Lorentzian dip is observed in the linear-scale total output and absorption spectra, with linewidth $2\gamma_{-}$ extracted from the FWHM (open arrows), which sets the broad background in the dB-scale spectra. At large detuning, the total output exhibits two zero-output dips while the absorption simultaneously reaches unity, indicating CPA. The dips are fitted by two Lorentzians (blue and red dashed lines), yielding linewidths $2\gamma_{+}$ and $2\gamma_{-}$. In contrast, the inverse spectra exhibit vanishing linewidths corresponding to $\gamma_{\mathrm{eff},\pm}=0$, resulting in extremely sharp dips and a zero 3 dB-above-minimum linewidth.
• Figure 4: (a) The dB-scale total-output spectra show the evolution of the hybridized system under detuning. Representative spectra and the corresponding color map reveal that the dip positions follow the hybridized anti-mode frequencies (orange curves), indicating level attraction. (b) $\gamma_{\mathrm{eff},\pm}$ extracted from the inverse spectra using Lorentzian fitting. Green and gray circles denote values at $|\Delta/2\Gamma|>1$ and $|\Delta/2\Gamma|<1$, respectively; the dashed curve shows the prediction of Eq. (\ref{['E12']}), indicating that $\gamma_{\mathrm{eff},\pm}$ persists over a broad detuning range.
• Figure 5: Single-resonator case under matched damping $(\alpha=\kappa)$: By sweeping $p$, we tune $|S_{tot}(\omega=\omega_{m})|^2$ from Eq. (\ref{['F2']}) and examine the calculated $2\gamma_{\mathrm{eff}}$ (red dashed line) as well as the 3 dB-above-minimum linewidth obtained from Eq. (\ref{['F3']}) (black dots). The pink shaded region indicates a $\pm5\%$ deviation of $2\gamma_{\mathrm{eff}}$. The results are presented in (a) on a linear scale and in (b) on the dB-scale. It is clear that $2\gamma_{\mathrm{eff}}$ and the 3 dB-above-minimum linewidth overlap only when the total output under 0.05; beyond this point, the two definitions deviate significantly.