Photonic crystals (PCs) are fabrica

Photonic crystals (PCs) are fabricated by highly ordered dielectric materials such as polystyrene microspheres,30 SiO2 microspheres,31,32 TiO2 microspheres33 and even metal–organic frameworks.34 PCs are bright prospects for novel designs of chemical and biological sensor arrays due to photonicstopband properties. These unique characteristics endow PCs with a slow photon effect.35–37 With this effect, the group velocity of photons in PCs near the band edge of the photonic stopband can be slowed down signicantly, so the effective optical path length of light will increase. This approach can increase the interaction between a uorescent dye and light.38 If thephotonic stopband is close to the emission wavelength of a uorescence dye, PCs can achieve enhancement of the uorescence intensity to improve the performance of the sensor array signicantly.39 Fluorescence signals are captured as the cross-reection signal for further pattern recognition. The latter includes principal components analysis (PCA), linear discriminant analysis (LDA), hierarchical clustering analysis (HCA) and neural networks (which refer specically to radial basis function neural networks (RBFNs)). PCA has been used to analyze the multidimensional data of sensor arrays to indicate the “chemical reactivity space” of the sensor array described by each principal component.40 LDA is a technique used to reduce dimensions, and seeks to place dimensions into known classes.“Leave one out” (LOO) is a verication method, which uses eachdata point to provide a LDA model to determine accuracy.41 HCA, as an agglomerative clustering technique, can be used to measure the similarity between objects. Relying on changes in the Euclidean distance, the most similar samples are merged together until all samples are connected to each other.37 RBFN,as a special class of general neural-network models, can predict the potential trend of the data provided.