Design Considerations for Discrete Frequency Infrared Microscopy Systems

Abstract

Discrete frequency infrared chemical imaging is transforming the practice of microspectroscopy by enabling a diversity of instrumentation and new measurement capabilities. While a variety of hardware implementations have been realized, design considerations that are unique to infrared (IR) microscopes have not yet been compiled in literature. Here, we describe the evolution of IR microscopes, provide rationales for design choices, and catalog some major considerations for each of the optical components in an imaging system. We analyze design choices that use these components to optimize performance, under their particular constraints, while providing illustrative examples. We then summarize a framework to assess the factors that determine an instrument's performance mathematically. Finally, we provide a validation approach by enumerating performance metrics that can be used to evaluate the capabilities of imaging systems or suitability for specific intended applications. Together, the presented concepts and examples should aid in understanding available instrument configurations, while guiding innovations in design of the next generation of IR chemical imaging spectrometers.

Keywords: Fourier transform infrared; IR microscopy; Infrared microscopy; chemical imaging; design considerations; design of IR microscopes; discrete frequency infrared; optical design; performance metrics; quantum cascade laser; spectroscopy.

Figure 1.

Widefield IR spectroscopic imaging instruments. (a) Fourier transform infrared (FT-IR) imaging microscope. Reproduced with permission from Rowlette et al. Copyright 2021 Daylight Solutions Inc. Multiple DFIR imaging configurations have been reported with unique features: (b) Using a high numerical aperture (NA) refractive lens in an inverted transmission configuration to improve the resolution. The design was further synchronized with a high-frame rate focal plane array (FPA), while reducing illumination non-uniformity using high-speed laser tuning across the spectral range. Reproduced with permission from Yeh et al. Copyright 2014 American Chemical Society. (c) Laser illumination is homogenized by a rapidly rotating diffuser to reduce the laser speckle for widefield recording. Reproduced with permission from the Schonhals et al. Copyright 2018 Wiley-VCH. (d) Commercial systems have enhanced accessibility of laser-based DFIR microscopy. Reproduced with permission from Rowlette et al. Copyright 2021 Daylight Solutions Inc. These illustrations are presented for conceptual purposes only, with representations of major components, and are not intended to serve as a comprehensive schematic.

Figure 2.

DFIR imaging by raster scanning a focused point illumination by (a) fixed objective and beam or stage scanning, or (b) a scanning objective assembly with a second arm providing attenuated total reflection (ATR) option for higher resolution imaging using a solid immersion lens in contact with the sample.

Figure 3.

(a) Emission spectrum at 1696 cm⁻¹. (b) Power spectral intensity of the four-chip QCL laser acquired at every 20 cm⁻¹. (c) IR pulse at 913 cm⁻¹ wavenumber acquired on the oscilloscope using MCT detector. (d) Emission spectra acquired at 1 cm⁻¹ using an FT-IR system for the QCL tuned wavenumber. (e) Spectral linewidth variation of the laser across 770 to 1874 cm⁻¹ shows a typical value of 2 cm⁻¹. (f) Average output power of two representative QCLs when pulsed at the repetition rate of 1.3 MHz and 5% duty cycle across the spectral range. (g) S/N ratio of two separate four-chip QCLs as measured across the spectral range.

Figure 4.

Development of IR detectors and systems technology that can be considered for principal military and civilian applications is shown. Four generation systems spanning, (a) the first-generation scanning systems, (b) second-generation staring systems with integrated system on chip ROIC systems to (c) the third-generation and (d) fourth-generation multispectral large format array systems are as shown.

Figure 5.

Absorbance images of SU-8 photoresist patterned as a USAF 1951 resolution test target with the Group 5 numeric shown. A progression of DFIR improvements in resolution over the recent years include (a) Widefield and (b–d) point-scanning^,, design configurations respectively, when compared to (e) a comparable feature acquired using HD FT-IR. The scale bar is 40 μm.

Figure 6.

(a) Abbe diagram for different IR materials as a function of the Abbe number. (b) Partial dispersion variation for different IR materials as a function of the Abbe number. (c) Variation in the RMS error at the spot diameter with respect to the incident angle on the objective as a function of the wavelength. Simulation was performed using CODE V.

Figure 7.

Variation of (a) radial resolution and (b) axial resolution with the incident wavelength for conventional and confocal microscope for NA_obj = 0.71.

Figure 8.

Simulated system modulation transfer function (MTF) and detector MTF curves illustrating the different regions with the design space for various fλ/D conditions. Spatial frequencies are normalized to the detector cut-off frequency. Reproduced with permission from Rogalski. Copyright 2017 SPIE.

Figure 9.

Absorbance image of the USAF 1951 chrome on glass negative resolution test target acquired at (a) 1000 cm⁻¹ to determine the spatial frequency resolution and (b) 1600 cm⁻¹ of a representative DFIR system. Corresponding vertical line profiles of Group 6. All scale bars are 50 μm.

Figure 10.

(a) An image of a slanted-edge acquired by a DFIR system is projected orthogonally to generate a super-sampled edge intensity profile, known as the ESF. The derivative of the ESF is used to compute the LSF. The MTF of the system is derived by taking the modulus of the Fourier transform of the LSF. (b) MTF of the DFIR system as a function of wavenumber with an estimated resolution of 3.5 μm at 800 cm⁻¹.

Figure 11.

Spot diagram for the 0.71 NA objective at the reference wavelength of 6 μm at an objective incident beam angle of (a) 0°, (b) 5°, and (c) 10°. The image plane is centered at the chief ray and all displacements are listed relative to the chief ray at the reference wavelength for each field. The diffraction-limited spot size (Airy disk) is shown with diameter equal to 2.44 fλ/D, in which f is the focal length and D is the effective aperture of the lens. Simulation was performed using CODE V.

Figure 12.

(a) Chromatic focal shift acquired for 0.72 NA BD-2 lens simulated using CODE V for design wavelength of 7.8 μm = 1282 cm⁻¹ for best focus case and compared to the acquired data for the spectral range 800–1800 cm⁻¹. (b) Variation of S/N with respect to integration time shows a nonlinear relation, dominated by read noise in low-S/N regime and shot noise in a high S/N regime for two different binning cases, M_x = 1× and M_x = 4×.

Figure 13.

Comparison of the noise characteristics of IR imaging spectrometers. (a) Spatial RMS noise calculated using spatial variance of recorded data at each band in a hyperspectral image of a representative widefield DFIR and commercial FT-IR imaging systems. (b) 100% spectral lines from a single-pixel under typical experimental conditions and the corresponding FPA noise floor. Reproduced with permission from the Yeh et al. Copyright 2014 American Chemical Society. (c) Spatial RMS noise calculated using spatial variance of recorded data at each band in a hyperspectral image of a representative DFIR point-scanning and commercial FT-IR imaging systems. (d) 100% spectral lines from a single-pixel under typical experimental conditions and the detector noise floor. Reproduced with permission from Yeh et al. Copyright 2019 American Chemical Society.