The breakthrough MEMS-Fabry–Perot interferometry spectral analyzer (C15712) offers engineers and scientists a compact, inexpensive and versatile module to expand the range of applications in near infrared spectroscopy.

Near infrared (NIR) spectroscopy offers many advantages for applications in the agricultural, industrial, medical and pharmaceutical sectors. Studying material properties in the near infrared (NIR) is a very powerful analytical tool that until recently has been limited to the laboratory due to instrumentation size, complexity and power consumption. New micro-opto electro-mechanical (MOEMS) technologies are changing the market by drastically reducing the size, cost, and complexity of NIR instrumentation.

Applying new technology to an old principle, namely, Fabry–Perot interferometry (FPI), is the core of the NIR MOEMS breakthrough instrument introduced here. There are three FPI modules in the product family, covering 1350 nm–2150 nm, part numbers C15712, C15713, and C15714. The invention of the FPI began in 1897 when Charles Fabry introduced a new optical interferometer that would selectively allow light to pass through two parallel plates, or reflective mirrors. When certain opto-mechanical conditions are met in relation to the wavelength of the incoming light, only a narrow band of light is passed through the system. Their initial invention was rather large and cumbersome to align, but it worked. Today, using semiconductor fabrication technology, the parallel plates and mirrors needed for the resonant cavity can be fabricated on a silicon wafer. The FPI cavity is fabricated on a silicon substrate that is transparent to NIR radiation. It consists of a stationary lower mirror, air gap formed by a sacrificial layer removal and an electrostatically controlled upper mirror, as show in Fig. 1. It is important to note, for purposes of temperature related transmission variations, the air gap is in the nanometer scale (Fig. 1).

FIG. 1.

Cross-sectional view of an FPI tunable optical filter. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved.

FIG. 1.

Cross-sectional view of an FPI tunable optical filter. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved.

Close modal

By controlling the distance between the two parallel plates, we can tune the FPI to transmit certain wavelengths while suppressing others. In MEMS-FPI technology, the distance between the plates is controlled using an external voltage. In forming the complete spectral sensor, we begin with an optical filter, intended to block an unwanted portion of the spectrum. Small spacers are used to support the FPI tunable filter and a photodiode converts the transmitted light into an electrical signal, Fig. 2(a). In the case of NIR light incidence, an InGaAs photodiode is selected to cover the wavelength range from 1350 nm to 2150 nm. Designed for portable implementations, with potentially harsh environments, the stacked structure is placed inside of a hermetically sealed metal can package, shown next to a ball point pen for size comparison in Fig. 2(b).

FIG. 2.

(a) Internal structure of the FPI tunable filter stacked on an InGaAs photodiode. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved. (b) FPI spectral sensor housed in a hermetically sealed package. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved.

FIG. 2.

(a) Internal structure of the FPI tunable filter stacked on an InGaAs photodiode. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved. (b) FPI spectral sensor housed in a hermetically sealed package. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved.

Close modal

Practical limitations to the physical gap distance in the FPI tunable filter make it necessary to fabricate three different FPI spectrum sensors, each covering its own unique wavelength range. The three ranges are as follows: 1350 nm–1650 nm, 1550 nm–1850 nm, and 1750 nm–2150 nm. The optical resolution, defined at the full width half maximum (FWHM), is ∼20 nm. This provides sufficient resolution for the targeted material identification regions.

Module solutions solve some specific device implementation challenges, such as ease of use and compensation for temperature effects. For ease of use, the module incorporates a halogen light source for sample illumination (easily configured for transmission- or reflection-based spectroscopy measurements) and is powered by a universal serial bus (USB) (Fig. 3).

FIG. 3.

Complete FPI module used for NIR spectroscopy applications. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved.

FIG. 3.

Complete FPI module used for NIR spectroscopy applications. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved.

Close modal

One of the biggest challenges is the impact of temperature on the measurement. Small changes in temperature can result in movement of FPI mirrors, thereby altering the optical performance of the filter. As the ambient temperature changes, the distance between the two parallel FPI mirrors will also slightly vary. The change in the FPI gap will alter the transmission wavelength. Such wavelength shifts in spectral transmission will introduce undesirable errors (Fig. 4). In Fig. 4, the FPI filter is tuned to transmit 1700 nm. It is noted that the peak transmission corresponds to 1700 nm at room temperature (25 °C). As the ambient temperature changes, without compensation (blue circles “Before”), a significant shift in the peak transmission is observed. To offset this undesirable effect, temperature compensation algorithms are employed on the module electronics to stabilize the unit from −5 °C to +50 °C. The red line in the graph of Fig. 4 shows the benefits of module compensation, holding the peak transmission to ± 0.2 nm.

FIG. 4.

FPI filter tuned to transmit 1700 nm. Comparison of peak wavelength transmission changes with temperature. Before compensation, we observe ±10 nm change. After compensation, the changes were drastically improved to ± 0.2 nm. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved.

FIG. 4.

FPI filter tuned to transmit 1700 nm. Comparison of peak wavelength transmission changes with temperature. Before compensation, we observe ±10 nm change. After compensation, the changes were drastically improved to ± 0.2 nm. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved.

Close modal

Certain wavelengths of light are absorbed depending on the interacting molecule. As the molecular composition of materials varies, so do their spectral signatures. Figures 5 and 6 show that light undergoes varying degrees of absorption by different substances depending on the wavelength. This spectral signature is used to identify the substance. In this example, a processing algorithm technique called standard normal variate (SNV) is employed. While there are many benefits of NIR spectroscopy, fabric identification is becoming increasingly important. Consumers demand reassurance of material authentication as manufacturers look to reduce cost and substitute materials. By performing NIR absorption-based spectroscopy measurements on fabrics such as cotton, polyester and wool, their unique spectral signatures can be collected and used as fabric identifiers. Using a broadband halogen light source, the C15712 system collects the reflected light from a fabric sample using the MEMS-FPI module. Spectroscopic data in the 1550 nm–1850 nm range are plotted in Fig. 5.

FIG. 5.

Fabric spectra, distinguishing between blends of cotton, wool and polyester. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved.

FIG. 5.

Fabric spectra, distinguishing between blends of cotton, wool and polyester. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved.

Close modal
FIG. 6.

NIR spectra of commonly recycled plastics. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved.

FIG. 6.

NIR spectra of commonly recycled plastics. No part of these images may be reproduced or transmitted in whole or in part without the written permission of Hamamatsu Corporation, Hamamatsu Photonics K.K. and its affiliates. All rights reserved.

Close modal

Another important application involves plastic pollution, a growing concern due to its huge adverse environmental impact. To help combat plastic pollution, NIR spectroscopy is helping to rapidly identify and sort plastics during the recycling process. In addition, consumer-based sorting efforts are now made possible using MEMS-FPI technology and absorption spectroscopy. Proper sorting, at the disposal stage, can drastically improve recycling efficiencies and waste management. Three common types of plastics, having unique NIR absorption, are: polyethylene terephthalate (PET), polyvinyl chloride (PVC) and polystyrene (PS). Their NIR spectra are shown in Fig. 6.

Near infrared spectroscopy has already been demonstrated as a valuable scientific tool for textile and plastic materials identification. The C15712 FPI module opens markets and applications that were previously limited to laboratory bench top systems. Thanks to the ultrahigh sensitivity and portability, a growing number of handheld applications are expected to develop, including food analysis and authentication. New envisioned applications include field portable industrial gas and fuel analysis and textiles to identify counterfeit materials, among others.