Nowadays, the simplicity of both designing and fabrication process of a terahertz (THz) resonator-based sensing technique leads to its ongoing development. The consumable THz resonator needs to be easily integrated into an existing terahertz time domain spectroscopy (THz TDS) measurement system. It should also be able to be fabricated in a mass scale with a low production cost. In this work, a metal-coated surface plasmon resonance- (SPR-) based sensor is simulated and designed as a low-cost refractive index sensor utilizing rigorous coupled wave analysis (RCWA). To demonstrate our methodology, we design a gold-coated grating with a polydimethylsiloxane (PDMS) as a substrate, in order to perform quantitative analysis of gasoline-toluene mixture composition, which has a refraction index variation of 0.1 at THz frequency. The grating period is tuned such that its surface plasmon resonance (SPR) frequency matches with the peak frequency of the THz TDS system. Moreover, other grating parameters, i.e., a filling factor and a grating depth, are optimized to increase the sensor sensitivity and sharpen the resonance dip. High sensitivity up to 500 GHz/RIU with a refractive index resolution up to 0.01 is numerically revealed. The H-field of the designed grating is then evaluated to indicate a strong SPR excitation. The well-developed designed grating introduces a promising, low-cost, and easily fabricated THz refractive index sensor.

No abstract

No abstract

Caffeine, quinic acid, and nicotinic acid are among the significant chemical determinants of coffee quality. This study develops a chemometric model to quantify these compounds in ternary mixtures analyzed by terahertz time-domain spectroscopy (THz-TDS). A data set of 480 THz spectra was obtained from 80 samples. Combinations of data preprocessing methods, including normalization (Z-score, min-max scaling, Mie baseline removal) and dimensionality reduction (principal component analysis (PCA), factor analysis (FA), independent component analysis (ICA), locally linear embedding (LLE), non-negative matrix factorization (NMF), isomap), and prediction models (partial least-squares regression (PLSR), support vector regression (SVR), multilayer perceptron (MLP), convolutional neural network (CNN), gradient boosting) were analyzed for their prediction performance (totaling to 4,711,685 combinations). Results show that the highest quantification performance was achieved at a root-mean-square error of prediction (RMSEP) of 0.0254 (dimensionless mass ratio), using min-max scaling and factor analysis for data preprocessing and multilayer perceptron for prediction. Effects of preprocessing, comparison of prediction models, and linearity of data are discussed.