A compact multi-bounce attenuated total reflection (ATR) probe combined with a Fabry-Pérot filter spectrometer (FPFS) has been developed for detection of hydrogen peroxide used for oxidative gas scrubbing operating in the mid-infrared (MIR) spectral region. A novel MIR supercontinuum light source is employed to enhance the quantification capabilities of the sensor and is compared to a classical thermal emitter. An improvement of a factor of 4 in noise and approximately a factor of 3 in limit of detection is shown in this study allowing fast inline detection of aqueous hydrogen peroxide solutions around 0.1%.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Chemical-oxidative gas scrubbing has significant potential as a novel method for the desulphurization of biogas . The oxidation and thereby the removal of hydrogen sulphide is facilitated by adding hydrogen peroxide (H2O2) to the caustic absorption solution. In this regard, a sensor providing a linear response and covering the concentration range from several percent to tenths of percent H2O2 in the absorption solution is needed. A classical redox sensor does not fulfil this requirement as the signal is proportional to the logarithm of the concentration, a significant drawback at higher H2O2 concentrations typically set in industrial scrubbers. Chemical oxidative scrubbing offers operational advantages in degree of automation and flexibility towards changes in process conditions. For efficient process control and in view of safety considerations the concentration of hydrogen peroxide in the absorption solution has to be known promptly and accurately. This results in the need of a sensitive, selective, robust, on- or inline sensor, especially to satisfy requirements in the fields of process analytical chemistry (PAC) or process analytical technologies (PAT) . Mid-infrared (MIR) spectroscopy is a molecular-specific technique which incorporates the aforementioned requirements in regards to sensitivity and selectivity. Additionally, probes can be constructed by using attenuated total reflection (ATR) inside a crystal. This facilitates a robust sample interface between optics and analyzed liquid being an absolute necessity for spectroscopic inline-sensors. ATR spectroscopy was first realized to be an elegant technique to measure MIR reflection spectra of a wide range of samples in the early 60s [3–5]. Since then, numerous applications have emerged ranging from polymer studies  to soil characterization  as well to a vast variety of analysis in pharmaceutical applications . However, the interaction length of the infrared light with the investigated sample is determined by the evanescent field created at the crystal-sample interface. A measure for the extension of the evanescent field into the sample is the depth of penetration. For most applications it is not larger than a few micrometers  per reflection (often referred to as bounce). Small effective path lengths result in a limiting performance in detecting substances with low concentrations or small absorption coefficients. One way to improve the situation is to design a multi-bounce system  for increasing the effective light-sample interaction length. This approach however, is limited mainly by strong absorption of the matrix and the low optical throughput when using a thermal light source. For the latter reasons a lower Signal-to-Noise ratio (SNR) is achieved, which will result in lower sensitivities.
Beginning with lead MIR lasers in the mid to late 60s [11,12] laser sources have found their way into the field of MIR spectroscopy . Especially quantum cascade lasers (QCL) offer high optical output power ranging from hundreds of mW  to a few W  enabling them to be used in long interaction path lengths. Their inherent property of having a narrow linewidth in spectral emission, which can be tuned over a small amount of wavenumbers, allows selective gas sensing using rotational-vibrational transitions in combination with different modulation and detection schemes [16–18]. Nonetheless, broader regions of the MIR spectrum have to be accessed when liquid samples with broader absorption bands are analyzed . Here, external cavity QCLs (EC-QCL) are a better match due to their ability to tune over more than 100 cm−1 [14,20], although the spectral information can only be obtained by tuning the laser one wavelength at a time [21–23]. Continued advance and combination of several EC-QCLs already allow for the coverage of even larger spectral ranges.
In recent years another possibility of generating spectrally broad laser light in the mid-infrared region is becoming available. Thereby pulses from seed lasers emitting in the near infrared region are converted by a non-linear device, e.g. certain kinds of optical fibers. The obtained laser pulses with a broad spectral bandwidth can be described as having a super-wide continuous optical spectrum. Hence, the term supercontinuum laser (SCL) was coined . SCLs offer broadband emission up to 16 µm  and high optical power up to several watts  combined with high spatial coherence and repetition rates in the MHz regime. These properties facilitate the use of the SCL as a novel and highly interesting radiation source for spectroscopic applications. The downside compared to tunable laser sources is the need for wavelength discrimination, analogous to instruments employing thermal emitters as light sources.
Therefore, in this study we propose a combination of a SCL and a tunable Fabry-Perot filter-spectrometer (FPFS)  as a high throughput MIR spectrometer with a multi-bounce ATR sample interface for the analysis of aqueous solutions. The ATR probe was designed for optimum performance for the target application of sensing H2O2 in water. In this setup the tunable FPFS acts as dispersive spectrometer including already the detector element. Tunable Fabry-Pérot filters (FPF) employing MEMS based, spring suspended movable mirrors are commercially available for the MIR spectral range up to 10 µm and allow the construction of rugged and small optical sensors, which are often referred to as filtometers. Here, we compare the performance of two filtometer instruments, one equipped with the SCL as a high-power radiation source to another equipped with a classical thermal emitter, which has shown to be suitable in PAT and PAC applications [27,28]. For SCLs the applicability for transmission and reflection measurements has been demonstrated before , as well as for stand-off detection of various compounds [30,31]. Furthermore, the long–term stability and the quantification capability of both configurations for aqueous solutions of hydrogen peroxide was assessed.
2. Materials and Methods
2.1 SCL as source of mid-infrared radiation
The laser source used in this study was a prototype laser built by NKT Photonics (Birkerød, Denmark). It is based on a combination of a pumping diode laser at 1550 nm, which produces sub-nanosecond pulses. Emitted pulses are coupled into a series of erbium-ytterbium doped fiber segments for amplification, before being converted to about 2 µm in a nonlinear silica-based fiber. This light is then amplified again in a Tm-doped fiber while the majority of the light generated by the seed laser is absorbed. The final output spectrum is generated during non-linear processes in a step-index ZBLAN (a glass family containing ZrF4-BaF2-LaF3-AlF3-NaF ) fiber segment and spans from 1.75 to 4.2 µm. The output power was measured to be 75 mW at a repetition rate of 40 kHz. The emitted average pulse length in this mode was determined to be 3 ns.
Figure 1 shows a simplified sketch of the instrumental setup. The base for the sample interface is a truncated cone (16/25 mm diameter, 10 mm thickness) made of ZnSe (Korth Kristalle, Kiel, Germany), which acts as mechanical holder for the ATR crystal as well as a focusing element, coupling the light efficiently into the ATR crystal. The ATR crystal consisted of a diamond disk (Type IIa, Diamond materials, Freiburg, Germany) with a diameter of 14 mm (flat surfaces were polished, Ra < 20 nm) and a thickness of 1 mm, where the light bounces for 4 times between the sample/ZnSe and air/ZnSe interface generated by the bore (4.8 mm in diameter) on top of the ZnSe cone. At the end of the diamond disk the light is coupled out into the ZnSe element, where it is collimated and guided towards the detector. The ZnSe cone and the diamond ATR crystal were integrated into a monolithic aluminum probe head. At the back of the holder, two openings allow the coupling of light. In order to couple light into the ZnSe cone a focusing lens with f = 25 mm (positive meniscus lens ZC-PM-12-25, ISP Optics, New York, USA) was used for the thermal emitter. Since the SCL had a collimated output beam, no lens was required. In order to couple the light out of the ZnSe cone and focus on the detector, again a positive meniscus lens was used (ZC-PM-12-25, ISP Optics, New York, USA, f = 12 mm). The detector was mounted in a 16 mm cage system (Thorlabs Inc, New Jersey, USA) using a translational mount (SCP05, Thorlabs Inc, New Jersey, USA) for adjustment. The ATR was designed to have 4 bounces in the desired spectral region. A FPFS with a pyroelectrical detector element (LFP-3144C, InfraTec GmbH, Dresden, Germany) was used for wavelength discrimination. For convenience, the setup incorporating the thermal emitter will be denoted by the prescript PTE (pulsed thermal emitter) and the laser setup with the prescript SCL. Table 1 summarizes the instrumentation of the compared configurations. A flip mirror (SCP05, Thorlabs Inc, New Jersey, USA) was integrated into the optical setup to switch between SCL and thermal emitter.
The thermal emitter (µHybrid, Hermsdorf, Germany) was pulsed with a duty cycle of 50%, at 5 V and 120 mA. The SCL was modulated using an optical chopper wheel (MC2000B-EC, Thorlabs Inc, New Jersey, USA) reducing the modulation frequency to 10 Hz in order to adapt to the time constant of the FPFS. A single spectrum with the FPFS was obtained by step-scanning the filter over the whole available spectral range with 2 nm steps. This resulted in an acquisition time of 36 s per spectrum.
For the quantitative measurements, aqueous solutions of hydrogen peroxide (Sigma Aldrich) were used. Samples were prepared by dilution from a stock solution with a concentration of 15% for hydrogen peroxide with deionized water.
3. Results and discussion
3.1 Noise evaluation
Figure 2 depicts the raw intensity spectra obtained when the surface of the ATR is wetted with water. Both PTE and SCL show a similar intensity profile, although the PTE for itself should have a much broader and thus smoother emission characteristic (Fig. 2(a), red line). On the higher wavenumber end range of the FPFS, the fall of intensity is caused by the strong absorption of the vibrational O-H stretch oscillations of water for the thermal emitter. The SCL has itself a strong decline on the higher wavenumber end, as indicated by its intensity spectrum. On the lower wavenumber part of the spectrum the absorption of the diamond ATR crystal itself reduces throughput. Only the resulting spectral gap with sufficient transmitted light can be used for MIR spectroscopy of aqueous solutions.
In order to assess the noise characteristic of each configuration, two consecutive spectra over the whole range of the tunable FPFS were obtained, from which absorbance spectra were calculated, thus yielding 100% lines.
The interesting part of the available spectrum resides between 2620 and 2920 cm−1, as hydrogen peroxide has a characteristic absorption band (O-H stretch vibration) at approximately 2820 cm−1 . This region is covered by both light sources. At the maximum the PTE-FPFS combination only reaches about 35 mV amplitude, whereas the SCL-FPFS comes to 1345 mV, which is an increase by a factor of 38. Here, the much higher light intensity of the SCL shows its strength. Figures 2(c) and 2(d) show the aforementioned 100% lines. In the region of interest the PTE-FPFS exhibits a peak-to-peak (PP) noise of 40 mAU with a root mean square (RMS) of 5.3 mAU. The SCL-FPFS shows less noise with a PP noise of 8 mAU and a RMS noise of 1.3 mAU.
3.2 Quantification of hydrogen peroxide
A standard concentrations series of aqueous hydrogen peroxide solutions was prepared and measured with both instrumentations. Figure 3 depicts the recorded spectra of the 9% hydrogen peroxide solution obtained with PTE-FPFS, SCL-FPFS and a Bruker (Ettlingen, Germany) Tensor 27 FTIR spectrometer equipped with a single bounce diamond ATR element (Platinum ATR, Bruker, Ettlingen, Germany). The PTE-FPFS and the FTIR spectrum show similar absorbance maxima considering the PTE-FPFS ATR is a four-bounce ATR resulting in approximately four-times higher absorption values.
The same setup with the SCL as the light source behaves differently by showing an absorbance of twice the value of its thermal powered counterpart. Primarily, this can be explained by the fact that the SCL is emitting polarized radiation, whereas the PTE source emits non-polarized light. This results in a different effective optical path length (also known as effective thickness), which corresponds to the thickness of a material that would result in the same absorbance in a transmission experiment as that obtained in an ATR experiment . Light which is polarized parallel to the plane of incidence (denoted by the subscript p) has a larger effective path length than light which is perpendicular polarized (denoted by the subscript s).
Equations (1) and (2) give an approximated solution for the effective path length of parallel (dp) and perpendicular (ds) polarized light, where λ is the wavelength, θ is the angle of incidence, n1 is the refractive index of the sample and n21 is the ratio of refractive indices of the ATR crystal and the sample. Figure 4(a) shows the effective path length over the angle of incidence calculated for perpendicular and parallel polarization as well as the ratio between the two. Figure 4(b) depicts the reflectance over the angle of incidence for p-polarization and s-polarization for the case of a non-absorbing sample (solid lines) and an absorbing sample (with an assumed κ of 0.01, dashed lines). Figure 5(a) shows the effective path length calculated for the spectral range used in this study for both polarization states for a one-bounce ATR. It results in 5.2 µm for p-polarization and 2.6 µm for s-polarization at the position of the maximum of the hydrogen peroxide band. When the angle of incidence (at 45°) and the ratio of the refractive indices of the sample/ATR interface are assumed to be constant, the effective path length of the p-polarized light is approximately twice as high as that of the s-polarized light.
In this case the dispersion of the refractive index around the absorption band of hydrogen peroxide is neglected. Figure 5(a) depicts the raw absorbance spectra for p- and s-polarized light of the SCL, respectively. Figure 5(b) shows the comparison of the theoretical ratio to the ratio of the measured spectra. A good correlation between the calculation and the measurement can be observed. Aside from polarization, the beam characteristics of the two sources also influence the effective path length. The PTE has to be focused into the ATR by means of a lens due to the larger divergence and beam diameter compared to the SCL. The SCL is spatially coherent with a divergence of 1.5 mrad, whereas the PTE has an active area of 2.2x2.2 mm2 with a reflector cap producing a divergent beam of a diameter of approximately 6 mm. This results in a different coupling with different angle of incidence ranges, which can change the effective path lengths as shown in Fig. 4(a).
For testing the capability of each system in quantifying hydrogen peroxide, several spectra with differently concentrated H2O2 standards were measured. The obtained spectra were filtered using a Savitzky-Golay algorithm with a window size of 19 and a second order polynomial. Furthermore, a baseline correction was performed, before the characteristic band at 2780 cm−1 was integrated (between 2720 and 2900 cm−1) and plotted against the concentration of the prepared standard. The resulting calibration curves are displayed in Fig. 6. The slope for the SCL-FPFS configuration are higher by a factor of approximately 1.8 compared to the PTE-FPFS configuration, which can be explained by the aforementioned differences regarding the polarization and coupling of the used light sources.
Limits of detection (LODs) were calculated using the standard deviation of the blank measurements, a k-value of 3 and the slope of the calibration curve according to the IUPAC definition . Table 2 summarizes the results found for both configurations. The SCL-FPFS outperforms the PTE-FPFS. This can be primarily attributed to the higher sensitivity and to the lower noise of the filtometer employing the supercontinuum laser.
3.3 Long term stability
In order to assess the long term stability of the system, consecutive spectra of a blank (water) were recorded over several hours. The RMS noise of the resulting 100% line was derived in the same spectral region as was used for integration and quantification in section 3.2. This time series was used to calculate an Allan variance, which evaluates the deviation of a measurement in dependance of its integration time τ .
The Allan variance plot (Fig. 7) is a particularly suitable tool for the identification of different noise types and the time domains where those noise types are dominant. When normally distributed noise (“white” noise) is assumed, the Allan variance should constantly decrease with increasing integration time. In case of SCLs this type of noise mainly originates from non-linear amplification of the input laser shot noise and spontaneous Raman noise . With increasing integration time Allan plots typically reveal a turning point after which the variance increases again indicating the presence of signal drifts, e.g. due to power fluctuations or thermal drifts. Such a behavior results in a region of minimum Allan variance, which can also be observed in Fig. 7. Here, the use of the SCL as radiation source results in the smaller overall Allan variance which corresponds to the noise behavior discussed before (see 100% lines in Fig. 2). The PTE has a much lower intensity, which results in larger measurement noise and the highest Allan variance. The combination SCL-FPFS appears to have a slightly higher drift, since the minimum of the Allan curve is found earlier at around 280 s compared to the 350 s of the PTE-FPFS. This indicates that the PTE is slightly less affected by intensity drifts than the SCL. However, the SCL-FPFS appears to be superior over the PTE-FPFS in terms of sensitivity with a similar long term stability.
In this study a dedicated 4-bounce diamond ATR interface was built for infrared spectroscopic analysis of aqueous solutions of hydrogen peroxide, intended for use in oxidative gas scrubbing, in the wavelength region from 3.1 to 4.4 µm. The coupling of mid-infrared supercontinuum radiation into the ATR was compared to the use of the standard light source in ATR spectroscopy, being a thermal emitter. In both cases, a tunable Fabry-Perot filter with integrated detector was used for wavelength discrimination and light detection. The achievable light throughput through the ATR crystal and the corresponding signal as well as its spectral noise was determined. The PTE-FPFS showed lowest light intensity at the detector. The SCL as a light source produced a signal 38-times higher than the PTE, which resulted in a noise reduction by a factor of 4 from 5.3 to 1.3 mAU, calculated as the RMS noise in the region of interest between 2620 and 2920 cm−1. For both configurations the ability to quantify aqueous solutions of hydrogen peroxide was tested and compared. Here, the SCL showed promising improvements lowering the LOD for hydrogen peroxide detection from 0.38 to 0.13%. Additionally, the effect of the polarization of light on the effective path length in ATR crystals was studied. Changing the absorbance of a sample by changing the polarization of the incident light (and therefore changing the effective path length) allows tuning of sensitivity. This could be used to enhance measurements of low-concentrated compounds of interest or to reduce the sensitivity if strong infrared absorbers challenge the linearity of the method. Furthermore, the temporal resolution could be increased by reducing the spectral points measured for a spectra. Instead of stepping over the whole available spectral range of the FPFS, only interesting positions could be selected, decreasing the measurement time.
In conclusion, the gained results prove that a promising new type of spectroscopic light source has become available with the advent of MIR SCLs. SCLs combine properties of conventional thermal emitters with those of modern MIR QCLs, resulting in a high-power, spectrally broadband MIR source with laser properties. With these unique properties further spectroscopic applications of SCLs are likely to follow. In this regard, the development and application of dedicated detection schemes is a crucial factor, as was shown in the present study.
Austrian research funding association (FFG) grant “Industrial Methods for Process Analytical Chemistry - From Measurement Technologies to Information Systems (imPACts)” (843546).
Financial support was provided by the Austrian research funding association (FFG) under the scope of the COMET program within the research project “Industrial Methods for Process Analytical Chemistry - From Measurement Technologies to Information Systems (imPACts)” (contract #843546) and by the strategic economic- and research program “Innovative Upper Austria 2020” of the province of Upper Austria.
References and links
1. M. Miltner, A. Makaruk, J. Krischan, and M. Harasek, “Chemical-oxidative scrubbing for the removal of hydrogen sulphide from raw biogas: Potentials and economics,” Water Sci. Technol. 66(6), 1354–1360 (2012). [CrossRef] [PubMed]
3. N. J. Harrick, “Surface chemistry from spectral analysis of totally internally reflected radiation,” J. Phys. Chem. 64(9), 1110–1114 (1960). [CrossRef]
4. N. J. Harrick, “Electric Field Strengths at Totally Reflecting Interfaces,” J. Opt. Soc. Am. 55(7), 851 (1965). [CrossRef]
5. J. Fahrenfort, “Attenuated total reflection,” Spectrochim. Acta 17(7), 698–709 (1961). [CrossRef]
6. K. R. Kirov and H. E. Assender, “Quantitative ATR-IR Analysis of Anisotropic Polymer Films: Extraction of Optical Constants,” Macromolecules 37(3), 894–904 (2004). [CrossRef]
7. R. Linker, I. Shmulevich, A. Kenny, and A. Shaviv, “Soil identification and chemometrics for direct determination of nitrate in soils using FTIR-ATR mid-infrared spectroscopy,” Chemosphere 61(5), 652–658 (2005). [CrossRef] [PubMed]
10. M. M. Mossoba, J. K. G. Kramer, H. Azizian, J. Kraft, P. Delmonte, A. R. F. Kia, F. J. Bueso, J. I. Rader, and J. K. Lee, “Application of a novel, heated, nine-reflection ATR crystal and a portable FTIR spectrometer to the rapid determination of total Trans fat,” JAOCS. J. Am. Oil Chem. Soc. 89(3), 419–429 (2012). [CrossRef]
11. F. K. Tittel, D. Richter, and A. Fried, “Mid-infrared laser applications in spectroscopy,” Solid-State Mid-Infrared Laser Sources 516, 445–516 (2003).
12. P. Werle, F. Slemr, K. Maurer, R. Kormann, R. Mücke, and B. Jänker, “Near- and mid-infrared laser-optical sensors for gas analysis,” Opt. Lasers Eng. 37(2-3), 101–114 (2002). [CrossRef]
13. R. F. Curl, F. Capasso, C. Gmachl, A. Kosterev, B. McManus, R. Lewicki, M. Pusharsky, G. Wysocki, and F. K. Tittel, “Quantum cascade lasers in chemical physics,” Chem. Phys. Lett. 487(1-3), 1–18 (2010). [CrossRef]
14. M. Brandstetter, A. Genner, K. Anic, and B. Lendl, “Tunable external cavity quantum cascade laser for the simultaneous determination of glucose and lactate in aqueous phase,” Analyst (Lond.) 135(12), 3260–3265 (2010). [CrossRef] [PubMed]
15. P. Figueiredo, M. Suttinger, R. Go, E. Tsvid, C. K. N. Patel, and A. Lyakh, “Progress in high-power continuous-wave quantum cascade lasers [Invited],” Appl. Opt. 56(31), H15–H23 (2017). [CrossRef] [PubMed]
16. J. P. Waclawek, H. Moser, and B. Lendl, “Compact quantum cascade laser based quartz-enhanced photoacoustic spectroscopy sensor system for detection of carbon disulfide,” Opt. Express 24(6), 6559–6571 (2016). [CrossRef] [PubMed]
17. J. P. Waclawek, C. V. Bauer, H. Moser, and B. Lendl, “2 f-wavelength modulation Fabry-Perot photothermal interferometry,” Opt. Express 24, 28958–28967 (2016).
18. H. Moser, W. Pölz, J. P. Waclawek, J. Ofner, and B. Lendl, “Implementation of a quantum cascade laser-based gas sensor prototype for sub-ppmv H2S measurements in a petrochemical process gas stream,” Anal. Bioanal. Chem. 409(3), 729–739 (2017). [CrossRef] [PubMed]
19. J. M. Chalmers and P. R. Griffith, Handbook of Vibrational Spectroscopy (Wiley, 2016).
20. M. Brandstetter, A. Genner, K. Anic, and B. Lendl, “Tunable Mid-IR lasers: A new avenue to robust and versatile physical chemosensors,” Procedia Eng. 5, 1001–1004 (2010). [CrossRef]
21. A. Schwaighofer, M. R. Alcaráz, C. Araman, H. Goicoechea, and B. Lendl, “External cavity-quantum cascade laser infrared spectroscopy for secondary structure analysis of proteins at low concentrations,” Sci. Rep. 6(1), 33556 (2016). [CrossRef] [PubMed]
22. M. R. Alcaráz, A. Schwaighofer, C. Kristament, G. Ramer, M. Brandstetter, H. Goicoechea, and B. Lendl, “External-Cavity Quantum Cascade Laser Spectroscopy for Mid-IR Transmission Measurements of Proteins in Aqueous Solution,” Anal. Chem. 87(13), 6980–6987 (2015). [CrossRef] [PubMed]
24. R. R. Alfano, The Supercontinuum Laser Source (Springer, 2016).
25. Z. Zhao, B. Wu, X. Wang, Z. Pan, Z. Liu, P. Zhang, X. Shen, Q. Nie, S. Dai, and R. Wang, “Mid-infrared supercontinuum covering 2.0-16 μm in a low-loss telluride single-mode fiber,” Laser Photonics Rev. 11(2), 1700005 (2017). [CrossRef]
26. N. Neumann, M. Ebermann, S. Kurth, and K. Hiller, “Tunable infrared detector with integrated micromachined Fabry-Perot filter,” J. Micro/Nanolithography. MEMS MOEMS 7(2), 21004 (2008). [CrossRef]
27. A. Genner, C. Gasser, H. Moser, J. Ofner, J. Schreiber, and B. Lendl, “On-line monitoring of methanol and methyl formate in the exhaust gas of an industrial formaldehyde production plant by a mid-IR gas sensor based on tunable Fabry-Pérot filter technology,” Anal. Bioanal. Chem. 409(3), 753–761 (2017). [CrossRef] [PubMed]
28. C. Gasser, A. Genner, H. Moser, J. Ofner, and B. Lendl, “Application of a tunable Fabry-Pérot filtometer to mid-infrared gas sensing,” Sens. Actuators B Chem. 242, 9–14 (2017). [CrossRef]
29. J. Kilgus, P. Müller, P. M. Moselund, and M. Brandstetter, “Application of supercontinuum radiation for mid-infrared spectroscopy,” in F. Berghmans and A. G. Mignani, eds. (2016), Vol. 9899, p. 98990K.
30. 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]
31. M. Kumar, M. N. Islam, F. L. Terry Jr, M. J. Freeman, A. Chan, M. Neelakandan, and T. Manzur, “Stand-off detection of solid targets with diffuse reflection spectroscopy using a high-power mid-infrared supercontinuum source,” Appl. Opt. 51(15), 2794–2807 (2012). [CrossRef] [PubMed]
32. X. Zhu and N. Peyghambarian, “High-power ZBLAN glass fiber lasers: Review and prospect,” Adv. Optoelectron. 2010, 1–23 (2010). [CrossRef]
33. O. Bain and P. A. Giguère, “Hydrogen peroxide and its analogues: VI. infrared spectra of H2O2, D2O2 and HDO2,” Can. J. Chem. 33(3), 527–545 (1955). [CrossRef]
35. G. L. Long and J. D. Winefordner, “Limit of detection. A closer look at the IUPAC definition,” Anal. Chem. 55, 712A–724A (1983).
36. P. Werle, R. Mücke, and F. Slemr, “The limits of signal averaging in atmospheric trace-gas monitoring by tunable diode-laser absorption spectroscopy (TDLAS),” Appl. Phys. B Photophysics Laser Chem. 57, 131–139 (1993). [CrossRef]
37. J. M. Dudley, S. Coen, K. L. Corwin, N. R. Newbury, B. R. Washburn, S. A. Diddams, and R. S. Windeier, “Fundamental noise limitations on supercontinuum generation in microstructure fiber,” 2003 Eur. Quantum Electron. Conf. EQEC 2003, 203 (2003).