Open Access
Add to BibSonomy
Abstract
Chemical mapping was demonstrated with a mid-infrared (MIR) microspectroscopy setup based on a supercontinuum source (SC) emitting in the spectral range from 1.55 to 4.5 µm and a MEMS-based Fabry-Pérot filter spectrometer. Diffraction limited spatial resolution in reflection geometry was achieved. A multilayer film consisting of different polymers and mixtures thereof was measured and results were compared to those gained with a conventional FTIR microscope equipped with a thermal MIR source. Results show that compared to thermal sources, the application of the SC source results in higher signal-to-noise ratios together with better spatial resolution and faster scanning. Furthermore, diffraction limited imaging of red blood cells was demonstrated for the first time in the MIR spectral region in reflection mode. The distinctive characteristics of the MIR spectral region in conjunction with the high brightness, spatial coherence and broadband nature of supercontinuum radiation show the potential for improving infrared microscopy significantly.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Mid-infrared (MIR) microspectroscopy is a powerful measurement technology that can provide chemical information at diffraction limited spatial resolution in a label-free, non-destructive, non-invasive and contactless way [1]. The biomedical field represents a major driving force where the versatile application of MIR microscopy ranges from the analysis of tissues [2] and body fluids [3] to the imaging of single cells [4] in vivo or ex vivo. Supported by promising results, e.g. in cytopathology [5] and histopathology [6], the implementation into clinical diagnostics is continuously advancing [7,8]. Furthermore, MIR microspectroscopy is an established imaging method used for the analysis of various kinds of non-clinical samples such as plants, food, pharmaceuticals or polymers [9].
In order to achieve small focal spots when using thermal light sources, apertures are needed. This poses a major drawback for classical MIR microspectroscopy, as thereby also the signal-to-noise ratio (SNR) is decreased. That leads to an unavoidable trade-off between spatial resolution, SNR, field of view and acquisition time. The trade-off becomes especially problematic in cases where the SNR is already limited: Firstly, in the MIR spectral region covering the fundamental molecular vibrations, also known as fingerprint region, the brightness of thermal sources is very low. Secondly, for optically thick or non-transparent samples, or for certain measurement conditions, such as in vivo diagnostics, back-reflectance is the only possible measurement geometry, leading to a significant reduction in signal strength. Furthermore, the number of reflected and backscattered photons is limited, as many biomedical samples exhibit large scattering and absorption cross sections. Additionally, for in vivo imaging in clinical diagnostics short measurement times are demanded. In order to achieve a reasonable SNR at shorter measurement times, larger apertures are required. However, this happens at the cost of spatial resolution moving far away from the diffraction limit.
It has been shown that high-brightness synchrotron radiation is perfectly suited for MIR microscopy by simultaneously offering diffraction limited resolution, high SNR, short acquisition times and broad spectral coverage [10,11]. However, the lack of availability of synchrotron radiation clearly limits its prevalent use. In recent years, laser-based light sources have experienced major technological developments, leading to promising results regarding MIR imaging [12]. Especially quantum cascade lasers (QCL) achieve similar advantages [13,14] as synchrotron sources, thereby easily outperforming classical Fourier transform infrared (FTIR) microscopes equipped with thermal light sources. On the one hand, the high brightness of QCLs offers higher sensitivity thus enabling measurements in demanding environments such as water or in vivo samples. On the other hand acquisition times can be reduced due to their fast tunability and by only recording at discrete wavenumbers [15]. Therefore QCL-based systems show a particularly high potential for biomedical spectroscopy [16] and for their clinical implementation [17].
An emerging alternative high-brightness MIR light source is the supercontinuum laser (SCL). SCLs have been successfully applied in the visible and near-infrared wavelength region for many years [18]. Currently the available spectral range is expanding into the MIR region up to 15 µm [19] offering high power emission up to 21.8 W [20]. The main advantage over QCLs is their broad spectral coverage at a much lower price than a combined QCL system. MIR SCLs show similar noise characteristics as QCLs [21] and have been already applied in diverse spectroscopic applications [22,23]. However, SCLs do not possess inherent tunability like QCLs do. In order to achieve similar tunability, acousto-optical or Fabry-Pérot tunable filters (AOTF, FPTF) can be used for spectral differentiation. In the field of microspectroscopy, the SCL has recently shown its high potential mainly in imaging applications in transmission mode where the samples were raster scanned. This has lead e.g. to images with a spatial resolution of 20 µm when using a SC spectrum up to 4 µm [24] or to a spatial resolution of ~12 µm using a SC spectrum up to 7.5 µm when using confocal measurement configurations with high-NA Schwarzschild objectives [25]. Both results were recorded by monochromatic detection with a single-pixel detector. In another work a SCL, emitting up to 4 µm, was coupled to a FTIR microscope leading to an 8 times higher SNR as compared to the standard configuration with a globar source [26].
In this work, MIR microspectroscopy in reflection geometry using a SCL is introduced. It will be demonstrated that the presented microscope configuration based on a Schwarzschild objective achieves diffraction limited spatial resolution in the spectral range between 3.1 to 4.4 µm. The high spatial resolution allowed the recording of images of red blood cells on a glass slide. Additionally, the advantage of using a high brightness light source will be demonstrated by comparing the results of line-scanning a polymeric multilayer film to results obtained by a mapping FTIR microscope. For spectral differentiation, a FPTF combined with a detector was used achieving low noise and maximized light throughput. The possibility to measure a strongly absorbing, biological sample in reflection at a diffraction-limited resolution in the MIR range demonstrates the high potential of SC radiation. Thereby, the main advantage of reflection microscopy becomes available, i.e. having no restrictions on the sample or substrate regarding light transmission.
2. Materials and methods
2.1 Supercontinuum laser
The employed SC source (SuperK MIR, NKT Photonics) [27] operates in the spectral region from 1.55 to 4.5 µm. The spectrum is generated by amplification and spectral broadening of the sub-nanosecond pulses (2.5 MHz repetition rate) of a gain-switched laser diode (1550 nm). The supercontinuum generation is induced by several non-linear processes in the step-index ZBLAN (ZrF4-BaF2-LaF3-AlF3-NaF) fiber [28]. The output beam of the fiber is terminated by a parabolic mirror for achromatic collimation achieving a divergence better than 1.5 mrad. The beam quality factor M2 of the Gaussian single mode beam is 1.1, according to the manufacturer. The average output power of the SC radiation is approximately 0.5 W.
2.2 Fabry-Pérot tunable filter (FPTF)
The spectral recording was executed by one of two different FPTF. On the one hand a FPTF combined with a uncooled PbSe photoresistor (XFP-3137-003, InfraTec GmbH, in the following referred to as XFP) covering the wavelength range 3.1 to 3.7 µm with the detectivity of 9 × 106 cm√Hz/W (100 Hz, 25 °C, central wavelength = 3.5 µm) was used. This Fabry-Pérot filter uses the fourth interference order resulting in a spectral resolution of approximately 25 nm. On the other hand a FPTF combined with a lithium-tantalate pyroelectric detector (LFP-3144C-337, InfraTec GmbH, in the following referred to as LFP) covering the wavelength range 3.1 to 4.4 µm with the detectivity of 3.6 × 106 cm√Hz/W (10 Hz, 25 °C, central wavelength = 4 µm) was used. This Fabry-Pérot filter uses only first interference order resulting in a spectral resolution between 55 to 70 nm. The light was intensity modulated at 250 Hz (for the XFP) or 100 Hz (for the LFP) by a chopper (MC2000B, Thorlabs GmbH) and the signal amplitude at each filter position was evaluated by accumulating the signal over 4 periods. This lead to a measurement time per pixel of approximately 2 s.
2.3 Setup
The microscope configuration - depicted in Fig. 1 - was built with reflective mirror optics thereby avoiding chromatic aberrations. Approximately 50% of the infrared light of the collimated beam is transmitted through a CaF2 beamsplitter (BS). For focusing the SC beam, a Schwarzschild objective (LMM-40X-P01, Thorlabs) with a numerical aperture of 0.5 and an infinite back focal length was used (RO). The entrance pupil of the Schwarzschild objective has a diameter of 5.1 mm. As typical for reflective mirror objectives, the central part of the incoming beam does not penetrate the objective. For the objective in use, 22% of the incoming beam waist was blocked. In order to achieve homogeneous illumination of the entrance pupil and sufficient throughput through the objective the Gaussian beam of the SCL was expanded (approx. 4 times) by a telescope built from parabolic mirrors (PM). The reflected light from the focal point was collected and collimated by the objective. The beam was then reflected by the beamsplitter and guided to the detection system. In order to have visible control of the focal spot and the sample a flip mirror was inserted into the light path and by means of focusing optics an image is generated on a webcam (not depicted in Fig. 1). Illumination with visible light was done by transmission of light of a halogen lamp from below the sample. The infrared light was intensity modulated by a chopper and detected by the Fabry-Pérot filter spectrometer. In order to generate a hyperspectral infrared image, the sample was raster scanned with a 2D-Stage (ASR050B050B-T3, Zaber). Additionally, a mapping Fourier-transform-infrared (FTIR) microscope (Bruker Lumos) equipped with a thermal light source was used for comparative measurements.
Fig. 1 Schematic of the optical configuration: SCL, Supercontinuum Laser; ZBLAN, step-index fiber made of heavy metal fluoride glasses; PM, parabolic mirror; BS, beam splitter; RO, reflective objective; CH, chopper; FPFS, Fabry-Pérot filterspectrometer.
2.4 Samples
2.4.1 USAF resolution target
For the determination of the spatial resolution, an image of an USAF resolution test target (group number 7, element 4, exhibiting line pairs with a pitch of 5.5 µm) was recorded. The positive target consists of a chromium pattern on soda lime glass. In the past, the target was immersed in water for other experiments leading to partial corrosion of the chromium pattern. The corroded areas are marked by arrows in Fig. 2(b).
Fig. 2 (a) The edge spread function (ESF) was generated by fitting the collected data (measured by the LFP) by line-scanning an edge of a bar of the resolution test target. The line spread function (LSF) was determined by calculating the derivative of the ESF; (b) top: microscope image taken in the visible spectral range of the corroded USAF test target (group 7, element 4). The arrows indicate damaged areas of the bars of the target; bottom: SCL microscope image (coupled to the XFP); intensity profiles of the marked line are plotted in (c) for the minimum and maximum wavelength, 3.1 and 3.7 µm, respectively, and for the global intensity.
2.4.2 Polymeric multilayer film
The polymeric multilayer film is symmetrically structured and consists of two outer layers (~130 µm thick) of Polypropylene (PP), a center layer (~60 µm thick) of Ethylene vinyl alcohol (EVOH) and of two layers (~200 µm thick) of a recycled mixture of both aforementioned materials in between. From the multilayer a microtome was sliced and attached onto a glass slide.
2.4.3 Blood smear sample
The sample consists of an air-dried but unfixed human blood smear on a glass slide. Human whole blood samples were obtained from the blood donor center of the Austrian Red Cross in Vienna. Aliquots of 10 mL whole blood were taken from the diagnostic samples of two anonymized blood donors after completion of the routine diagnostics, which follows blood donation.
2.5 PCA denoising
In this work principal component analysis (PCA) based denoising was used in order to enhance the contrast of the hyperspectral images using the free to use Python package ‘Orange’ [29]. This technique is based on the fact that relevant spectral information can usually be described using only the first few principal components, while the higher components mostly contain noise [30]. In this way the variance of the transformed data is maximized in a low-dimensional representation and the acquired spectra are reconstructed, achieving an effective reduction of the random noise components. For the PCA denoising of the spectra of the blood sample the first two principal components were used. The only preprocessing step was spectral smoothing (Savitzky-Golay filter).
3. Results and discussion
3.1 Spatial resolution
In order to determine the spatial resolution of the setup an edge of the resolution target was line-scanned (0.4 µm step size). The spectral detection was executed by the LFP (3.1 - 4.4 µm). The edge spread function (ESF), shown in Fig. 2(a), was calculated by means of fitting the global intensity values (average intensity value of whole spectrum). The first derivative of the ESF yields the line spread function (LSF) which features small side maxima as typical for Schwarzschild objectives. The LSF has a full width at half maximum (FWHM) of approximately 4.5 µm, shown in Fig. 2(a). This corresponds well to the Abbe limit, defining the diffraction limited spot diameter as .
Furthermore, an image of the target (group 7, element 4) was recorded by mapping the area with a step-size of 0.5 µm. The spectral detection was executed by the XFP (3.1 - 3.7 µm). In order to generate the image in Fig. 2(b) (bottom) the global intensity values were used. The obtained image demonstrates that it was possible to resolve the lines with a width of 2.8 µm. Even corroded areas of the bars, indicated by arrows in the light microscopic picture at the top of Fig. 2(b), can be recognized in the image of the SCL microscope. The recorded image shows some artefacts in the region around the chromium bars which are assumed to originate from the side maxima of the LSF. These artefacts could be removed by using a confocal optical system additionally improving the resolution [31]. In Fig. 2(c) the profiles at the marked position in the recorded image are shown for 3.1 µm, 3.7 µm and for the global intensity. The global intensity profile shows similar resolution as the profile of the maximum wavelength (3.7 µm). The smallest wavelength recorded possesses the sharpest profile as expected by the Abbe limit, but the resolution observed seems to be better than expected by the Abbe limit. It is known that reflective microscope objectives provide a sharper central maximum compared to normal microscope objectives leading to a smaller FWHM of the point spread function [11,32].
3.2 Signal-to-noise ratio
The main advantage of using high brightness light sources is illustrated by the comparison of the achievable SNR. Thermal emitters require the application of apertures in order to gain spatial coherent illumination and subsequently diffraction limited spatial resolution. This leads to a decisive dependence of the achievable SNR and the spatial resolution given by the aperture size. In Fig. 3 this dependence is shown by the measured SNR for different aperture sizes for the FTIR microscope. The SNR is given by SNR = 100/(rms noise) [33], where the rms noise was calculated from the 100%-lines of consecutive recorded spectra in the spectral range of 2700 to 3200 cm−1. The spectra were taken by reflection measurements on a standard microscopic glass slide. 32 scans were averaged with a spectral resolution of 4 cm−1 leading to a measurement time for a single spectrum of 20 s. The same glass slide was measured with the SCL microscope with a measurement time of 2 s for a single spectrum. The inset plot in Fig. 3 shows 100%-lines of the SCL microscope from which the SNR was calculated, indicated by the constant (no aperture required) red dotted line. It can be clearly seen from Fig. 3 that the FTIR microscope requires larger aperture sizes to achieve SNR values similar to the SCL microscope, however, thereby reducing its spatial resolution.
Fig. 3 Comparison of the achieved SNR for reflection measurements with different aperture sizes in the spectral range from 2700 to 3200 cm−1. SNR levels were calculated from 100%-lines that were recorded using an empty glass slide as specimen. The measurement time for the FTIR microscope was 20 s whereas for the SCL microscope it was 2 s. The inset shows the 100%-lines of the SCL microscope from which the SNR value was calculated, indicated by the red dotted line.
3.3 Cross section of polymeric multilayer film
The cross section of the polymeric multilayer film was measured with a visible (VIS) light microscope, with the introduced SCL microscope and with the FTIR microscope. Both IR microscopic measurements were executed in reflection mode as a line scan across the polymer layers. As the polymer sample is strongly absorbing and scattering, the collected light power was weak. In order to achieve a reasonable SNR, the aperture of the FTIR microscope had to be opened to a minimum of 18 x 18 µm2. For the FTIR microscope the step size was approximately 7 µm and at each position the average of 64 spectra was taken leading to an acquisition time per pixel of approximately 120 s. Smaller step sizes would not lead to improvements, as the aperture size restricts the spatial resolution. For the SCL microscope a step size of 1 µm was used, which is smaller than the width of the LSF. The acquisition time per pixel was 2 s. The spectral detection was executed by the XFP (3.1 - 3.7 µm).
In Fig. 4 the absorption spectra of the three different areas of the cross section of the multilayer film measured with the FTIR microscope are shown. The main difference in the spectra originates from different C-H stretching vibrations of PP and EVOH. The mixture of these polymers exhibits an absorption spectrum (Fig. 4, green line) varying in between the absorption spectra of the pure components. For the generation of the chemical image of the sample the C-H absorption bands were integrated, resulting in a single absorption value for each position. These results are illustrated in Fig. 5 by a color gradient, corresponding to high (yellow) and low (blue) absorption values.
Fig. 4 Absorption spectra of the three different components of the polymer multilayer film recorded with the FTIR microscope.
Fig. 5 The figure consists of 3 different recordings of the multilayer polymer film cross section: at the top the line-scan (7 µm steps) of the cross section of the sample recorded with the FTIR microscope equipped with a thermal emitter (TE), in the middle an image of the sample taken by a visible (VIS) light microscope in transmission and at the bottom the line-scan (1 µm steps) recorded by the SCL microscope.
The layered structure can be identified in the visible light microscopic image in Fig. 5 labeled as VIS. Nevertheless, in the visible spectral range it is not possible to differentiate between the different polymer types whereas this is clearly possible with the FTIR microscope using a thermal emitter (TE), see image at the top labeled as TE. The SCL microscope delivers an improved image in terms of spatial resolution (Fig. 5, bottom line), as the size of the focal point is not restricted by an aperture which is needed in the FTIR measurement for spatially coherent illumination. Therefore, the region consisting of the polymeric mixture can also be spatially resolved. This result illustrates the advantage of using a high brightness light source simultaneously offering faster acquisition and diffraction limited resolution in reflection mode.
3.4 Red blood cells
MIR imaging of biological samples in reflection mode with diffraction-limited resolution is a challenging task which is nonetheless required in biomedical applications. To demonstrate the capability of the SCL microscope, therefore, a human blood smear was measured. The blood smear was provided on a standard microscope glass slide, which is absorptive in the investigated infrared wavelength region. An area of 40x40 µm was mapped using a step size of 1 µm. The measurement time per pixel was 2 s. In Figs. 6(a) and 6(b) the light microscopic image and the PCA-denoised image recorded with the SCL microscope are shown, respectively. The absorption spectrum of a single position on a blood cell from the SCL microscope image is shown in Fig. 6(c). In this spectral region the C-H stretching bands of the blood cell can be detected.
Fig. 6 (a) visible light microscopic picture of the sample; (b) PCA-denoised image recorded by the SCL microscope; (c) absorption spectrum of a single red blood cell recorded by the SCL microscope.
PCA denoising analysis exploits the spectral information and neglects the random noise components. Thereby an image resembling the light microscopic picture can be generated, even revealing the donut shape of some of the cells. To our knowledge, this is the first image of red blood cells recorded in reflection mode in the MIR spectral region. In another work an image was generated by mapping an ATR probe over the sample [34] which however suffers from the destructive nature of the method. Miller et al. demonstrated imaging of single blood cells with a confocal transmission setup using a high brightness synchrotron source and recording the low-noise infrared spectra [35].
The possibility to resolve single red blood cells, both spectrally and spatially, could enable new applications in the field of hematopathology. As e.g. discussed by Liu et al. [36], infrared spectroscopy can provide information about molecular changes in the blood cells at very early pathogenesis stages. Having additional spatial resolution of the cells could allow to investigate infectious and structural blood diseases (such as malaria parasites [37], leukemia [38] or sickle cell disease [39]).
4. Conclusion and outlook
The presented results demonstrate that the spatial resolution of the developed SCL based microscope reached the diffraction limit in reflection geometry in the 3.1 - 4.4 µm spectral region. The comparison of the results gained with a standard FTIR microscope equipped with a thermal emitter with those gained with the SCL microscope underpinned the main advantage of the latter: the high spatial coherence of the SCL eliminates the necessity of using apertures. Together with the high brightness of the source, this leads to a better spatial resolution at higher SNR compared to thermal sources. Thereby highly resolved line-scans of a polymer cross section were collected and it was possible to record an image of single red blood cells, resolving their donut shape.
The Fabry-Pérot tunable filter (FPTF) spectrometer used for resolving spectral information also permits the measurement at single wavelengths, thereby accelerating acquisition times. Furthermore, FPTFs possess shorter interaction path lengths compared to other filters such as acousto-optical tunable filters, thereby simplifying their microscopic implementation.
The SCL represents an attractive new alternative in the field of MIR high-brightness laser sources with the distinct advantage of offering a broadband spectrum. The employed SCL operates in a spectral region that distinctively has the advantage of covering fundamental molecular vibrations (O-H and C-H vibrations) while at the same time providing a smaller diffraction limit as compared to the fingerprint region. The high brightness offered by SCLs can lead to additional improvements such as enabling confocal configuration or an imaging configuration with the application of uncooled bolometer cameras. This would be of particular interest for in vivo MIR cell imaging and for clinical applications as was indicated by the results presented in Farries et al. [40].
Author contributions
M. Brandstetter, J. Kilgus and G. Langer conceived and designed the experiments; J. Kilgus and K. Duswald performed the experiments; I. Zorin conducted measurements of multilayer polymer samples; T. Berer assisted with the microscopic 2D-stage; J. Kilgus and R. Zimmerleiter analyzed the data; J. Kilgus and M. Brandstetter wrote the paper.
Funding
Province of Upper Austria (Innovative Upper Austria 2020); Marie Skłodowska-Curie Actions (SUPUVIR, 722380); ERDF Urban Innovative Actions (IWB2020).
Acknowledgements
The authors want to thank Dr. Jungbauer from the blood donor center of the Austrian Red Cross in Vienna for provision of blood samples. J. Kilgus, G. Langer, R. Zimmerleiter, K. Duswald and M. Brandstetter acknowledge financial support by the strategic economic- and research program “Innovative Upper Austria 2020” of the province of Upper Austria; I. Zorin acknowledges funding by the Marie Skłodowska-Curie Action SUPUVIR of the European Union's H2020-MSCA-ITN-2016 Programme under REA grant agreement n° 722380; T. Berer acknowledges the project “multimodal and in-situ characterization of inhomogeneous materials” (MiCi) by the federal government of Upper Austria and the European Regional Development Fund (EFRE) in the framework of the EU-program IWB2020.
Disclosures
The authors declare that there are no conflicts of interest related to this article.
References
1. J. E. Katok, A. J. Sommer, and P. L. Lang, “Infrared Microspectroscopy,” Appl. Spectrosc. Rev. 25(3-4), 173–211 (1989). [CrossRef]
2. H. Sreedhar, V. K. Varma, P. L. Nguyen, B. Davidson, S. Akkina, G. Guzman, S. Setty, A. Kajdacsy-Balla, and M. J. Walsh, “High-definition Fourier Transform Infrared (FT-IR) spectroscopic imaging of human tissue sections towards improving pathology,” J. Vis. Exp. 95, 52332 (2015). [CrossRef] [PubMed]
3. M. J. Baker, S. R. Hussain, L. Lovergne, V. Untereiner, C. Hughes, R. A. Lukaszewski, G. Thiéfin, and G. D. Sockalingum, “Developing and understanding biofluid vibrational spectroscopy: a critical review,” Chem. Soc. Rev. 45(7), 1803–1818 (2016). [CrossRef] [PubMed]
4. J. Doherty, G. Cinque, and P. Gardner, “Single-cell analysis using Fourier transform infrared microspectroscopy,” Appl. Spectrosc. Rev. 52(6), 560–587 (2017). [CrossRef]
5. O. Old, G. Lloyd, M. Isabelle, L. M. Almond, C. Kendall, K. Baxter, N. Shepherd, A. Shore, N. Stone, and H. Barr, “Automated cytological detection of Barrett’s neoplasia with infrared spectroscopy,” J. Gastroenterol. 53(2), 227–235 (2018). [CrossRef] [PubMed]
6. C. Kuepper, F. Großerueschkamp, A. Kallenbach-Thieltges, A. Mosig, A. Tannapfel, and K. Gerwert, “Label-free classification of colon cancer grading using infrared spectral histopathology,” Faraday Discuss. 187, 105–118 (2016). [CrossRef] [PubMed]
7. M. J. Baker, H. J. Byrne, J. Chalmers, P. Gardner, R. Goodacre, A. Henderson, S. G. Kazarian, F. L. Martin, J. Moger, N. Stone, and J. Sulé-Suso, “Clinical applications of infrared and Raman spectroscopy: state of play and future challenges,” Analyst (Lond.) 143(8), 1735–1757 (2018). [CrossRef] [PubMed]
8. S. De Bruyne, M. M. Speeckaert, and J. R. Delanghe, “Applications of mid-infrared spectroscopy in the clinical laboratory setting,” Crit. Rev. Clin. Lab. Sci. 55(1), 1–20 (2018). [CrossRef] [PubMed]
9. R. Salzer and H. W. Siesler, Infrared and Raman Spectroscopic Imaging. (Wiley-VCH, 2014).
10. G. L. Carr, M. Hanfland, and G. P. Williams, “Midinfrared beamline at the National Synchrotron Light Source port U2B,” Rev. Sci. Instrum. 66(2), 1643–1645 (1995). [CrossRef]
11. M. J. Nasse, M. J. Walsh, E. C. Mattson, R. Reininger, A. Kajdacsy-Balla, V. Macias, R. Bhargava, and C. J. Hirschmugl, “High-resolution Fourier-transform infrared chemical imaging with multiple synchrotron beams,” Nat. Methods 8(5), 413–416 (2011). [CrossRef] [PubMed]
12. B. Guo, Y. Wang, C. Peng, H. Zhang, G. Luo, H. Le, C. Gmachl, D. Sivco, M. Peabody, and A. Cho, “Laser-based mid-infrared reflectance imaging of biological tissues,” Opt. Express 12(1), 208–219 (2004). [CrossRef] [PubMed]
13. D. T. D. Childs, R. A. Hogg, D. G. Revin, I. U. Rehman, J. W. Cockburn, and S. J. Matcher, “Sensitivity Advantage of QCL Tunable-Laser Mid-Infrared Spectroscopy Over FTIR Spectroscopy,” Appl. Spectrosc. Rev. 50(10), 822–839 (2015). [CrossRef]
14. J. A. Kimber and S. G. Kazarian, “Spectroscopic imaging of biomaterials and biological systems with FTIR microscopy or with quantum cascade lasers,” Anal. Bioanal. Chem. 409(25), 5813–5820 (2017). [CrossRef] [PubMed]
15. K. Yeh, S. Kenkel, J.-N. Liu, and R. Bhargava, “Fast Infrared Chemical Imaging with a Quantum Cascade Laser,” Anal. Chem. 87(1), 485–493 (2015). [CrossRef] [PubMed]
16. A. Schwaighofer, M. Brandstetter, and B. Lendl, “Quantum cascade lasers (QCLs) in biomedical spectroscopy,” Chem. Soc. Rev. 46(19), 5903–5924 (2017). [CrossRef] [PubMed]
17. C. Kuepper, A. Kallenbach-Thieltges, H. Juette, A. Tannapfel, F. Großerueschkamp, and K. Gerwert, “Quantum Cascade Laser-Based Infrared Microscopy for Label-Free and Automated Cancer Classification in Tissue Sections,” Sci. Rep. 8(1), 7717 (2018). [CrossRef] [PubMed]
18. C. F. Kaminski, R. S. Watt, A. D. Elder, J. H. Frank, and J. Hult, “Supercontinuum radiation for applications in chemical sensing and microscopy,” Appl. Phys. B 92(3), 367–378 (2008). [CrossRef]
19. T. Cheng, K. Nagasaka, T. H. Tuan, X. Xue, M. Matsumoto, H. Tezuka, T. Suzuki, and Y. Ohishi, “Mid-infrared supercontinuum generation spanning 2.0 to 15.1 μm in a chalcogenide step-index fiber,” Opt. Lett. 41(9), 2117–2120 (2016). [CrossRef] [PubMed]
20. K. Liu, J. Liu, H. Shi, F. Tan, and P. Wang, “High power mid-infrared supercontinuum generation in a single-mode ZBLAN fiber with up to 21.8 W average output power,” Opt. Express 22(20), 24384–24391 (2014). [CrossRef] [PubMed]
21. J. Kilgus, P. Müller, P. M. Moselund, and M. Brandstetter, “Application of supercontinuum radiation for mid-infrared spectroscopy,” in Proceedings of SPIE-The International Society for Optical Engineering (2016), 9899.
22. J. Kilgus, K. Duswald, G. Langer, and M. Brandstetter, “Mid-Infrared Standoff Spectroscopy Using a Supercontinuum Laser with Compact Fabry-Pérot Filter Spectrometers,” Appl. Spectrosc. 72(4), 634–642 (2018). [CrossRef] [PubMed]
23. C. Gasser, J. Kilgus, M. Harasek, B. Lendl, and M. Brandstetter, “Enhanced mid-infrared multi-bounce ATR spectroscopy for online detection of hydrogen peroxide using a supercontinuum laser,” Opt. Express 26(9), 12169–12179 (2018). [CrossRef] [PubMed]
24. S. Dupont, C. Petersen, J. Thøgersen, C. Agger, O. Bang, and S. R. Keiding, “IR microscopy utilizing intense supercontinuum light source,” Opt. Express 20(5), 4887–4892 (2012). [CrossRef] [PubMed]
25. C. R. Petersen, N. Prtljaga, M. Farries, J. Ward, B. Napier, G. R. Lloyd, J. Nallala, N. Stone, and O. Bang, “Mid-infrared multispectral tissue imaging using a chalcogenide fiber supercontinuum source,” Opt. Lett. 43(5), 999-1002 (2018).
26. F. Borondics, M. Jossent, C. Sandt, L. Lavoute, D. Gaponov, A. Hideur, P. Dumas, and S. Février, “Supercontinuum-based Fourier transform infrared spectromicroscopy,” Optica 5(4), 378 (2018). [CrossRef]
27. P. M. Moselund, L. Huot, and C. D. Brooks, “All-fiber mid-IR supercontinuum: a powerful new tool for IR-spectroscopy,” in R. R. Alfano and S. G. Demos, eds. (2016), p. 97030B.
28. P. Moselund, C. Petersen, L. Leick, J. Seidelin Dam, P. Tidemand-Lichtenberg, and C. Pedersen, “Highly Stable, All-fiber, High Power ZBLAN Supercontinuum Source Reaching 4.75 µm used for Nanosecond mid-IR Spectroscopy,” in Advanced Solid-State Lasers Congress (OSA, 2013), paper JTh5A.9. [CrossRef]
29. M. Toplak, G. Birarda, S. Read, C. Sandt, S. M. Rosendahl, L. Vaccari, J. Demšar, and F. Borondics, “Infrared Orange: Connecting Hyperspectral Data with Machine Learning,” Synchrotron Radiat. News 30(4), 40–45 (2017). [CrossRef]
30. C. Rodarmel and J. Shan, “Principal Component Analysis for Hyperspectral Image Classification,” (2002).
31. G. L. Carr, “Resolution limits for infrared microspectroscopy explored with synchrotron radiation,” Rev. Sci. Instrum. 72(3), 1613 (2001). [CrossRef]
32. P. Lasch and D. Naumann, “Spatial resolution in infrared microspectroscopic imaging of tissues,” Biochim. Biophys. Acta 1758(7), 814–829 (2006). [CrossRef] [PubMed]
33. J. M. Chalmers and P. R. Griffiths, Handbook of Vibrational Spectroscopy (J. Wiley, 2002).
34. A. J. Sommer, C. Marcott, G. M. Story, and L. G. Tisinger, “Attenuated Total Internal Reflection Infrared Mapping Microspectroscopy Using an Imaging Microscope,” Appl. Spectrosc. 55(3), 252-256 (2001).
35. L. M. Miller and P. Dumas, “Chemical imaging of biological tissue with synchrotron infrared light,” Biochim. Biophys. Acta 1758(7), 846–857 (2006). [CrossRef] [PubMed]
36. K.-Z. Liu, M.-H. Shi, and H. H. Mantsch, “Molecular and chemical characterization of blood cells by infrared spectroscopy: a new optical tool in hematology,” Blood Cells Mol. Dis. 35(3), 404–412 (2005). [CrossRef] [PubMed]
37. B. R. Wood, K. R. Bambery, M. W. A. Dixon, L. Tilley, M. J. Nasse, E. Mattson, and C. J. Hirschmugl, “Diagnosing malaria infected cells at the single cell level using focal plane array Fourier transform infrared imaging spectroscopy,” Analyst (Lond.) 139(19), 4769–4774 (2014). [CrossRef] [PubMed]
38. N. W. C. J. van de Donk, H. M. Lokhorst, K. C. Anderson, and P. G. Richardson, “How I treat plasma cell leukemia,” Blood 120(12), 2376–2389 (2012). [CrossRef] [PubMed]
39. J. Jung, L. E. Matemba, K. Lee, P. E. Kazyoba, J. Yoon, J. J. Massaga, K. Kim, D.-J. Kim, and Y. Park, “Optical characterization of red blood cells from individuals with sickle cell trait and disease in Tanzania using quantitative phase imaging,” Sci. Rep. 6(1), 31698 (2016). [CrossRef] [PubMed]
40. M. Farries, J. Ward, I. Lindsay, J. Nallala, and P. Moselund, “Fast hyper-spectral imaging of cytological samples in the mid-infrared wavelength region,” in R. R. Alfano and S. G. Demos, eds. (International Society for Optics and Photonics, 2017), 10060, p. 100600Y.
