Miniaturization of optical spectrometers has recently drawn a lot of attention due to the increasing needs of portable characterization systems for scientific, industrial, and consumer applications. At the same time, smartphones have technically evolved to become an everyday, ubiquitous device that provides numerous useful applications to consumers. Combining optical spectrometer and smartphone could lead to an explosion of new applications, especially in healthcare, biometrics, and food inspections, and change our daily life, making it more convenient, independent, and hyper-personalized.
In this work, we have developed a smartphone spectrometer in the visible and near infrared (NIR) ranges by directly integrating a 2 dimensional periodic array of band-pass filters on top of the smartphone’s image sensor. Each band pass filter is a silicon resonator consisting of a pair of Si/SiO₂ distributed Bragg reflectors (DBR), where each resonator’s transmitting wavelength is set by adjusting the thickness of the center Si layer. The DBR contains alternating, vertically-stacked TiO2 and SiN films with variable thicknesses while the top and bottom of the DBR were made of Al and Cu or Al reflectors for the visible and NIR ranges, respectively. The fabrication process was completely CMOS-compatible.
Using this smartphone spectrometer, we have proposed the concept of artificial-intelligence-powered spectral barcode for material identification and successfully demonstrated its use in drug identification. The accuracy of correctly identifying the type of drugs was ~99%. In addition, the smartphone spectrometer has also proven to correctly distinguish beef into three different classes according to the freshness.
|