Over-coupled resonator for broadband surface enhanced infrared absorption (SEIRA)

Abstract

Detection of molecules is a key issue for many applications. Surface enhanced infrared absorption (SEIRA) uses arrays of resonant nanoantennas with good quality factors which can be used to locally enhance the illumination of molecules. The technique has proved to be an effective tool to detect small amount of material. However, nanoresonators can detect molecules on a narrow bandwidth so that a set of resonators is necessary to identify a molecule fingerprint. Here, we introduce an alternative paradigm and use low quality factor resonators with large radiative losses (over-coupled resonators). The bandwidth enables to detect all absorption lines between 5 and 10 μm, reproducing the molecular absorption spectrum. Counterintuitively, despite a lower quality factor, the system sensitivity is improved and we report a reflectivity variation as large as one percent per nanometer of molecular layer of PMMA. This paves the way to specific identification of molecules. We illustrate the potential of the technique with the detection of the explosive precursor 2,4-dinitrotoluene (DNT). There is a fair agreement with electromagnetic simulations and we also introduce an analytic model of the SEIRA signal obtained in the over-coupling regime.

Fig. 1. Metasurface for SEIRA of PMMA in the critical coupling and over-coupling configuration.

a Scheme of the resonator array (period d). Calculated reflectivity spectra of b a critically coupled resonator (green lines) with dimensions h_Au = 45 nm, h_ZnS = 50 nm, w = 50 nm and d = 765 nm, and c an over-coupled resonator (orange lines) with dimensions h_Au = 45 nm, h_ZnS = 280 nm, w = 50 nm and d = 600 nm with (continuous line) and without (dashed line) a 45 nm thick PMMA layer filling the slits. d Calculated reflectivity difference ΔR spectrum of the over-coupled resonator (orange continuous line) and imaginary part of the dielectric permittivity of PMMA (blue dashed line).

Fig. 2. Temporal coupled mode theory analysis of the SEIRA in the over-coupling regime.

a Absorption of the resonator as a function of $\tilde{\omega }=\left(\omega -{\omega }_r\right)/{\gamma }_{nr}$ and the coupling ratio f = γ_r/γ_nr computed from Eq. 1. Without the absorber, the absorption is maximal on resonance (ω = ω_r) and at critical coupling (f = 1). The addition of an absorber into the resonator can be taken into account in this diagram changing the variables γ_r/γ_nr → γ_r/(γ_nr + γ_μ) and (ω − ω_r)/γ_nr → (ω − ω_r − ω_μ)/γ_nr. This effect brings back the resonator (represented by a red dot) towards critical coupling in a narrow spectral range defined by the resonance condition, resulting in a potentially large reflectivity difference ΔR_m. b Reflectivity difference ΔR_m as a function of f = γ_r/γ_nr computed in Supplementary Fig. 8. To vary f, we fix γ_nr and let γ_r vary. ΔR_m is plotted for γ_nr = 0.025 (yellow line), γ_nr = 0.01 (red line), and γ_nr = 0.005 (blue line). In the dashed vertical line, we plot the value $f_{\max }=\left({\omega }_b-{\omega }_r\right)/{\gamma }_{nr}$ associated to each value of γ_nr to highlight the position of the maximum of ΔR_m(f). The normalized parameters are : ω_r = 1, ω_b = 1.5, γ_r = 0.5, γ_b = 0.01 and μ = 0.03.

Fig. 3. Experimental demonstration of the SEIRA effect.

a Electron beam microscopy image of the edge of the resonator array with dimensions h_Au = 45 nm, h_ZnS = 280 nm, w = 50 nm, and d = 600 nm. Measured spectra of b the bare array and c the array covered with a 45 nm PMMA layer, for transverse magnetic polarization (orange continuous lines) and transverse electric polarization (green dashed lines). Gray areas cover the absorption of CO₂.

Fig. 4. Linear regime of the SEIRA signal in the over-coupled regime for PMMA.

a Schematic of a period of a resonator array covered with an ideal continuous PMMA layer. b Measured (dashed black line) and simulated (continuous purple line) reflectivity difference of a resonator (dimensions: h_Au = 45 nm, h_ZnS = 530 nm, w = 50 nm and d = 300 nm) covered with a 20 nm continuous PMMA layer in the slit and on top of the ribbons. c Simulated reflectivity difference of four vibration modes (indicated as λ1, λ2, λ3, and λ4 in b) as a function of the PMMA layer thickness up to 20 nm. Experimental dots are also plotted for various PMMA thicknesses measured with an ellipsometer. Error bars correspond to standard deviation.

Fig. 5. SEIRA effect on 2,4-dinitrotoluene.

a Measured (dashed lines) and calculated (continuous lines) reflectivity spectra of the previously introduced resonator with dimensions h_Au = 45 nm, h_ZnS = 530 nm, w = 56 nm, and d = 300 nm measured at t = 3 min (yellow lines) after deposition of DNT molecules and bt = 4 min (green lines) after deposition. The spectra on the bare resonator without molecules are plotted for comparison (solid red line for simulation and dashed red line for measurement). Continuous lines correspond to experimental data and dashed lines to fitted calculations. c Measured reflectivity difference spectra ΔR 3 min after deposition (solid orange line) and 4 min after deposition (solid green line) with respect to the left axis, and imaginary part of the dielectric permittivity of DNT model (blue dashed line) with respect to the right axis.