Spectral sensing is increasingly used in applications ranging from industrial process monitoring to agriculture. Sensing is usually performed by measuring reflected or transmitted light with a spectrometer and processing the resulting spectra. However, realizing compact and mass-manufacturable spectrometers is a major challenge, particularly in the infrared spectral region where chemical information is most prominent. Here we propose a different approach to spectral sensing which dramatically simplifies the requirements on the hardware and allows the monolithic integration of the sensors. We use an array of resonant-cavity-enhanced photodetectors, each featuring a distinct spectral response in the 850-1700 nm wavelength range. We show that prediction models can be built directly using the responses of the photodetectors, despite the presence of multiple broad peaks, releasing the need for spectral reconstruction. The large etendue and responsivity allow us to demonstrate the application of an integrated near-infrared spectral sensor in relevant problems, namely milk and plastic sensing. Our results open the way to spectral sensors with minimal size, cost and complexity for industrial and consumer applications.
We present an efficient and flexible method to realize micro- and nano-optical structures on the tip of optical fibers. We demonstrate this approach for a fiber-tip sensor consisting of a photonic crystal (PhC) structure in a semiconductor membrane on the cleaved facet of a telecom fiber. The PhC structure is fabricated on a wafer by lithography and etching and then transferred to the fiber facet by a simple mechanical pickup process through an opening in the substrate, without the need for adhesives or a micromanipulator. Due to its reliability, scalability, and the use of wafer-scale fabrication methods, this process increases the possibilities for fiber-tip applications at the industrial level. With the fabricated fiber tip sensors, we demonstrate sensing of the refractive index and temperature, with resonance wavelength shifts of 120 nm/RIU and 95 pm/K, respectively.
For decades, near-infrared (NIR) spectroscopy has been a valuable tool for material analysis in a variety of applications, ranging from industrial process monitoring to quality assessment. Traditional spectrometers are typically bulky, fragile and expensive, which makes them unsuitable for portable and in-field use. Thus, there is a growing interest for miniaturized, robust and low-cost NIR sensors. In this study, we demonstrate a handheld NIR spectral sensor module, based on a fully-integrated multipixel detector array, sensitive in the 850–1700 nm wavelength range. Differently from a spectrometer, the spectral sensor measures a limited number of NIR spectral bands. The capabilities of the spectral sensor module were evaluated alongside a commercially available portable spectrometer for two application cases: to quantify the moisture content in rice grains and to classify plastic types. Both devices achieved the two sensing tasks with comparable performance. Moisture quantification was achieved with a root mean square error (RMSE) prediction of 1.4% and 1.1% by the spectral sensor and spectrometer, respectively. Classification of the plastic type was achieved with a prediction accuracy on unknown samples of 100% and 96.4% by the spectral sensor and spectrometer, respectively. The results from this study are promising and demonstrate the potential for the compact NIR modules to be used in a variety of NIR sensing applications.
Spectral sensing is an emerging field driven by the need for fast and non-invasive methods for the chemical analysis of materials in agri-food, healthcare, and industrial applications. We demonstrate a near-infrared spectral sensor, based on a scalable fabrication process and combining high responsivity, narrow linewidth, and low noise. The sensor consists of 16 resonant-cavity-enhanced photodetectors, each showing a unique spectral response consisting of narrow peaks. The spectral sensor thereby covers the wavelength range between 890 and 1650 nm, where organic materials show relevant spectral features from first and second overtones. For the fabrication of the detector arrays, we propose a simple and scalable fabrication approach that yields largely improved device characteristics with respect to the grey-scale electron-beam lithography process reported earlier. Through a series of five optical lithography steps, tuning layers of silicon nitride are deposited stepwise to obtain 16 different thicknesses and reduced surface roughness. With this novel fabrication approach, the obtained photodetectors achieve an average peak linewidth of 55 nm, a maximum peak responsivity of 0.3 A/W, and high suppression of the non-resonant background. We also demonstrate the impact of these improvements on the sensing performance for two relevant problems through an experiment and a set of simulations. With lateral dimensions of ∼1.4 × 1.4 mm2, the proposed photodetector array can be the key to robust, portable, and low-cost sensing instrumentation for on-site material analysis in various application fields.
The design of optical sensors aims at providing, among other things, the highest precision in the determination of the target measurand. Many sensor systems rely on a spectral transducer to map changes in the measurand into spectral shifts of a resonance peak in the reflection or transmission spectrum, which is measured by a readout device (e.g., a spectrometer). For these spectral transducers, figures of merit have been defined which are based on specific assumptions on the readout and the data analysis. In reality, however, different transducers achieve optimal performance with different types of readout. Additionally, some transducers present a more complex spectral response for which existing figures of merit do not apply. In this paper, we investigate an approach to quantifying the potential performance of a given transducer for a more general class of readout methods. Starting from the Cramér–Rao lower bound, we define a new figure of merit, the integrated spectral sensitivity, which is directly related to the physical limit of precision and applicable to a wide variety of sensing systems. We apply this analysis to two different examples of transducers. The results bring useful insights into the design of optical sensors.
Measuring the composition of materials at a distance is a key requirement in industrial process monitoring, recycling, precision agriculture, and environmental monitoring. Spectral imaging in the visible or near-infrared (NIR) spectral bands provides a potential solution by combining spatial and spectral information, and its application has seen significant growth over recent decades. Low-cost solutions for visible multispectral imaging (MSI) have been developed due to the widespread availability of silicon detectors, which are sensitive in this spectral region. In contrast, development in the NIR has been slower, primarily due to the high cost of indium gallium arsenide (InGaAs) detector arrays required for imaging. This work aims to bridge this gap by introducing a standoff material sensing concept which combines spatial and spectral resolution without the hardware requirements of traditional spectral imaging systems. It combines spatial imaging in the visible range with a CMOS camera and NIR spectral measurement at selected points of the scene using an NIR spectral sensor. This allows the chemical characterization of different objects of interest in a scene without acquiring a full spectral image. We showcase its application in plastic classification, a key functionality in sorting and recycling systems. The system demonstrated the capability to classify visually identical plastics of different types in a standoff measurement configuration and to produce spectral measurements at up to 100 points in a scene.