Artificial intelligence assisted chiral nanoparticles | EurekAlert!
A new publication from Opto-Electronic Science; DOI 10.29026/oes.2023.220019 addresses chiral nanophotonic designs assisted by artificial intelligence.
Chiral nanostructures can enhance the weak inherent chiral effects of biomolecules and highlight the important role in chiral detection. However, the design of the chiral nanostructures is challenged by extensive theoretical simulations and exploratory experiments. Recently, the group of Zheyu Fang proposed a chiral nanostructure design method based on reinforcement learning, which can identify metallic chiral nanostructures with a sharp peak in circular dichroism spectra and improve the chiral detection signals. This work envisions the important role of artificial intelligence in nanophotonic designs.
Chirality is a fundamental physical property that means that an object or structure cannot be superimposed on its mirror image. It plays an important role in biomedical sensing, since enantiomers with opposite handedness lead to completely different biological effects. However, it is difficult to detect enantiomers in trace amounts because the chiral biomolecules are significantly smaller than optical wavelengths and exhibit weak dichroism signals.
Chiral plasmonic nanostructures were used to enhance the interactions between the optical waves and the chiral molecules. Common dichroism technologies are based on circularly polarized waves, i.e. the spin angular momentum (SAM) of photons. Chiral metasurfaces, metamaterials, and nanoparticles can be used to enhance the sensor signal because their intrinsic circular dichroism (CD) signals can be sensitively affected by chiral biomolecules. In recent years, orbital angular momentum (OAM) based dichroism technologies have been proposed and studied, in which the dichroism signal is defined on the basis of the differential response spectra of opposite vortex modes. Plasmonic resonance structures with intrinsic chirality can also amplify OAM-based dichroism signals.
However, the interactions between chiral molecules and chiral nanostructures are complex. Different biomolecules may require different nanostructures to achieve the optimally enhanced dichroism signals. Hence, the design of the chiral nanostructures in iterative electromagnetic (EM) simulations consumes enormous computational resources. Artificial intelligence (AI) is proving to be a powerful tool in nanostructure design, capable of handling more complex problems and larger data sets compared to traditional optimization algorithms. AI has been successfully used in the design of metasurfaces, photonic crystals, integrated wavelength routers, etc.
In a recent paper published in Opto-Electronic Science, Zheyu Fang and his colleagues propose a method for the design of chiral nanostructures based on reinforcement learning (Fig. 1), in which the exploration of new nanostructures and the updating of models occur simultaneously. The introduction of reinforcement learning improves the quality of the training data set and reduces the EM simulation scope. Fig. 1(a) shows the structure of the training data set. At the beginning, different nanostructures are randomly generated and their optical responses are calculated by EM simulations. As shown in Fig. 1(b), the next step is to train multiple artificial neural networks (ANNs) to obtain the association relationships between the nanostructure geometries and their spectra. Subsequently, new structures are designed using a Bayesian optimization algorithm based on predictions of ANNs (Fig. 1(c)). The ANNs detect nanostructures with strong chirality, potentially generate optimized structures, and reduce the computational burden for weakly chiral nanostructures. Figure 1(d) shows the training data set update process. For nanostructures whose optical responses predicted by different ANNs differ significantly, their optical responses are calculated by EM simulations. The inaccuracies of the ANNs indicate that there are few similar structures in the current dataset, so the simulated data are added to the training dataset. For the nanostructures that are consistently predicted from different ANNs, similar structures should already exist in the training dataset. After the data update, the ANNs are retrained. Therefore, ANNs ensure that the nanostructures added to the dataset have the potential for strong chirality, even though the original dataset is generated randomly.
To execute the proposed scheme, the chiral unit to be designed is parameterized by 40×40 units coded 0 or 1, which means that there is a gold block or an air at that position. ANNs allow the mapping between the optical response spectra and the matrices representing the geometry. Finally, three chiral metasurfaces with different CD peak frequencies are designed. The chiral metasurfaces are also fabricated and measured to validate the proposed method. The experimental results agree with the projected results. The frequency shifts of the CD spectra caused by chiral molecules are also measured using left-glucose and right-glucose solutions controlled by microfluidic channels. The resonance wavelength shifts between glucose enantiomers with opposite chirality reach 7 nm, indicating an increased sensitivity of the chiral molecules.
The proposed method improves the quality of the training data set and reduces the amount of electromagnetic simulation compared to classical exploration methods. Chiral nanostructures with significant CD values and high chiral detection sensitivity have been successfully designed and fabricated. Besides the demonstrated chiral nanophotonic designs, other optical properties can also be designed since the algorithm is universal for all physical meanings of the optical reactions. This work envisions promising applications of AI in nanophotonic and electromagnetic designs.
# # # # # #
Opto-Electronic Science (OES) is a peer-reviewed, open access, interdisciplinary and international journal published by the Institute of Optics and Electronics of the Chinese Academy of Sciences as a sister journal of Opto-Electronic Advances (OEA, IF=9.682). OES aims to provide a professional platform to promote academic exchange and accelerate innovation. OES publishes articles, reviews and letters about the fundamental breakthroughs in the basic science of optics and optoelectronics.
# # # # # #
More information: https://www.oejournal.org/oes
Editorial Board: https://www.oejournal.org/oes/editorialboard/list
OES is available in OE journals (https://www.oejournal.org/oes/archive)
Submission of OES can be done via ScholarOne (https://mc03.manuscriptcentral.com/oes).
Contact us: [email protected]
Twitter: @OptoElectronAdv (https://twitter.com/OptoElectronAdv?lang=en)
# # # # # #
Chen YX, Zhang FY, Dang ZB, He X, Luo CX et al. Chiral detection of biomolecules based on reinforcement learning. Opto-Electron Sci 2, 220019 (2023). doi: 10.29026/oes.2023.220019
# # # # # #
Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of the press releases published on EurekAlert! by contributing institutions or for the use of any information about the EurekAlert system.