Magnetochiroptical nanocavities in hyperbolic metamaterials enable sensing down to the few-molecule level

Chirality refers to the phenomenon where a pair of objects that are mirror images of each other cannot be superimposed under any type of geometric transformation. Chiral molecules such as amino acids, proteins, and DNA are ubiquitous in biological and biochemical processes.^1,2 Therefore, recognition and separation of enantiomers is crucial for applications in pharmaceuticals since the therapeutic effects of a chiral drug are often associated with a single enantiomer, whereas the other is inactive and/or toxic.³ Molecular chirality is normally investigated using circular dichroism (CD),⁴ but the signals are only measurable for large sample volumes or large molecules. This occurs due to the mismatch between the size of the molecules and the wavelength of the incident light, which produces weak light–matter interactions. Mechanisms that enable the detection and discrimination of chiral molecules down to the few-molecule level have been explored,⁵ mostly based on enhanced, localized optical fields in chiral all-dielectric/plasmonic platforms.^6–9 The sensitivity of these chiral nanophotonic structures is hampered by the corresponding background CD signal. To overcome the latter drawback, researchers exploit the extrinsic chirality, i.e., optical chirality, from chiral near-field distributions in achiral nanostructures (under obliquely incident light).^10–12 Indeed, chiral near-fields with zero far-field CD have been demonstrated using achiral systems,¹³ eliminating background signals for improved sensitivity. However, the passive nature of conventional achiral nanostructures still limits their applications.

Attempts to develop active achiral systems, i.e., achiral nanostructures with dynamically tunable CD, include the use of phase-change materials,¹⁴ chemical reactions,¹⁵ and magneto-optical (MO) effects (also called magnetochiroptical effects).^16,17 In comparison to temperature changes to manipulate phase-change materials and the use of chemical reagents, magnetochiroptical effects can not only be manipulated faster but also enable packaging via on-chip integrated electromagnetic coils.¹⁸ Moreover, magnetic anisotropy combined with material anisotropies produces unusual MO effects, as in MO hyperbolic metamaterials (MO-HMMs).^19–24 HMMs are artificial uniaxial materials with one of the principal components of the permittivity tensor (ɛ) with the opposite sign to the other two,²⁵ i.e., HMMs simultaneously exhibit dielectric and plasmonic properties. This unique feature enables the excitation of the so-called bulk plasmon polariton (BPP) modes, which are guided plasmonic resonances with their electromagnetic fields largely concentrated in the volume of the metamaterial slab. BPP resonances in non-magnetic HMMs have been used for biosensing with high sensitivity,^26–29 while unprecedented resolution levels were observed with the transverse MO Kerr effect (TMOKE) in MO-HMMs.³⁰ In fact, TMOKE has been used in the last few decades for the highly sensitive detection of biomolecules.^31–34 Those concepts were demonstrated for the analyte region throughout the entire superstrate medium; that is, these approaches are still limited to relatively large sample concentrations. On the other hand, recent works have demonstrated improved detection capabilities when using gratings³⁵ or nanocavities in HMMs.⁵ In the latter approach, nanocavities functioned as plasmonic nanocuvettes in which the analyte molecules were placed, thus reducing the analyte region to the volume of the nanocavities. One limitation of the latter approach, though, is the need to label fluorophores in the nanocuvettes.

In this work, we overcome these drawbacks by combining the concepts of plasmonic nanocavities and achiral MO-HMMs in what we call magnetochiroptical nanocavities. We demonstrate numerically the possibility of detecting a few tens of high-refractive-index achiral nanospheres inside the cavity volumes, which opens up the possibility for label-free (bio)sensing down to the few-molecule level. Our approach uses magnetic CD (MCD) peaks,³⁶ instead of plasmonic resonances, to monitor refractive index changes. Although most calculations in this work were made for achiral analytes, we numerically demonstrate that higher sensitivities are expected when using chiral molecules, where MCD will not only be affected by the applied magnetic fields but also by the chirality of the analyte. It is noteworthy that leveraging magnetochiroptical effects selectively amplifies the MCD amplitude for distinct enantiomers, thereby enhancing the measured MCD signal associated with each.

The reflectances for the three incidence conditions, i.e., for (θ_inc = 0°; ϕ_inc = 0°), (θ_inc = 45°; ϕ_inc = 0°), and (θ_inc = 0°; ϕ_inc = 45°), are given in Fig. 2. Numerical data are shown for m = 0 (in other words, the system is demagnetized) and m = +1 (the system is magnetized along the + z-axis). For RCP reflectance (under oblique incidence), dips appear near λ = 900 nm, which are related to the excitation of BPP resonances in the multilayer HMM structure. Magnetochiroptical effects are observed in Figs. 2(b)–2(d), where reflectances for RCP and LCP waves are shown for the three incidence conditions with m = +1. At normal incidence, i.e., θ_inc = ϕ_inc = 0°, an MCD peak is observed just around λ = 900 nm (see Fig. 3), as expected from geometric optimization. The last peak is due to the largest difference between the reflectance curves for RCP and LCP waves for wavelengths between the modes BPP₁ and BPP₂ in Fig. 2(c). Resonant shifts are observed in Fig. 3 for obliquely incident fields, induced through small changes in the phase-matching condition [see Eq. (2)]. Moreover, due to the combined contribution from BPP resonances and magnetochiroptical effects, higher MCD amplitudes are obtained under oblique incidence. In particular, MCD amplitudes for θ_inc = 0°; ϕ_inc = 45° and θ_inc = 45°; ϕ_inc = 0° achieve values (for both m = ±1) up to ±2.31° and ±2.23°, respectively, which are competitive with recent reports using chiral nanostructures comprising magnetoplasmonic helicoids on a dielectric substrate.⁴¹ 

Motivated by the possibility of exploiting MCD peaks for (bio)sensing applications, we focus our attention on the incidence conditions (θ_inc = 45°; ϕ_inc = 0°) due to the sharper MCD peak (see Fig. 3). Furthermore, BPP resonances have their electromagnetic fields distributed inside the volume of the HMM and, consequently, inside the nanocavity volume. In fact, the latter is demonstrated in Fig. 4, where calculations of the normalized resonant electromagnetic near-field distributions are presented. The top panel (bottom panel), labeled Fig. 4(a) [Fig. 4(b)], shows the modulus of the resonant magnetic near-field distribution $H$ for a RCP (LCP) incident wave. Differences associated with the near-field chirality of both polarizations are noted in the comparison in Fig. 4. Since the characteristics of the MCD (amplitudes and positions in the spectrum) can be adapted and tuned with N, as shown in Fig. 5, we select N = 3 to obtain a balance between a high MCD and a cost-effective and easy-to-fabricate nanostructure.

Since the near-fields in Fig. 4 produce gradient optical forces that trap nearby nanoparticles inside the nanocavity volume,⁴² one expects that in the case of low concentrations, the analyte molecules will become trapped in the nanocavities. Therefore, we start our analysis by assuming that refractive index changes only occur within the volume of the nanocavity, as shown in the inset in Fig. 6(b). The MCD curves and the corresponding spectral shifts as a function of n_a (varying from 1.33 up to 1.43 inside the nanocativy) are shown in Figs. 6(a) and 6(b), respectively. A sensitivity $(S(λ)=Δλres/Δna)$ of 40 nm/RIU is inferred, which is a relatively high value for the conditions of the calculations. In contrast, if changes in refractive index are not limited to the nanocavity but throughout the entire superstrate region, the sensitivity is 532 nm/RIU, as demonstrated in Figs. 6(c)–6(d). Unlike other approaches exploiting polar MO configurations for ultrasensitive applications,⁴³ our concept does not require a combination of different techniques to monitor phase-changes.

We have shown with numerical simulations that MCD peaks from an arrangement of nanocavities in a MO-HMM can be used for sensing. A proof of concept is demonstrated for two cases: (i) the analyte region is limited to the volume inside nanocavities, for which relatively high sensitivity values (S = 40 nm/RIU) were calculated, and (ii) the analyte region is distributed along the entire superstrate region, including the volume of the nanocavities, for which a very high sensitivity (S = 532 nm/RIU) is obtained. Since the proposed MO-HMM structure is surrounded by aqueous media, considering the analyte molecules limited to the volume of the nanocavities implicitly means that we are working with ultra-low concentration levels, i.e., we are only using a few analyte molecules in the sample. In contrast to other proposals using complementary fluorophore labeling for indirect detection through fluorescence spectroscopy, we only monitor the shifts of the MCD peaks due to small changes in the surrounding refractive index. It is worth noting that the nanocavity arrangement in the concept shown here also works as a grating coupler to allow the excitation of bulk plasmon polariton modes in the MO-HMM without the need for prism couplers, enabling straightforward integrability with microfluidic technology.

This work was partially supported by RNP, with resources from MCTIC, Grant No. 01245.020548/2021-07, under the Brazil 6G project of the Radiocommunication Reference Center (Centro de Referência em Radiocomunicações—CRR) of the National Institute of Telecommunications (Instituto Nacional de Telecomunicações—Inatel), Brazil, and by Huawei, under the project Advanced Academic Education in Telecommunications Networks and Systems, Contract No. PPA6001BRA23032110257684. We also acknowledge the financial support from the Brazilian agencies National Council for Scientific and Technological Development-CNPq (Grant No. 314671/2021-8) and FAPESP (Grant Nos. 2018/22214-6, 2023/08999-9).

The authors have no conflicts to disclose.

William O. F. Carvalho: Conceptualization (lead); Formal analysis (lead); Investigation (lead); Methodology (lead); Writing – original draft (equal). Osvaldo N. Oliveira, Jr.: Project administration (equal); Writing – review & editing (equal). J. R. Mejía-Salazar: Supervision (lead); Writing – original draft (equal); Writing – review & editing (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.