OSA's Digital Library

Optics Express

Optics Express

  • Editor: C. Martijn de Sterke
  • Vol. 20, Iss. 7 — Mar. 26, 2012
  • pp: 7590–7601

Optics InfoBase > Optics Express > Volume 20 > Issue 7 > Page 7590

« Show journal navigation

Real-time measurements of atmospheric CO using a continuous-wave room temperature quantum cascade laser based spectrometer

Jingsong Li, Uwe Parchatka, Rainer Königstedt, and Horst Fischer  »View Author Affiliations


Optics Express, Vol. 20, Issue 7, pp. 7590-7601 (2012)
http://dx.doi.org/10.1364/OE.20.007590


View Full Text Article

Acrobat PDF (1669 KB)


Advanced Search

Browse Journals / Lookup Meetings

Browse by Journal and Year


   


Lookup Conference Papers

Close Browse Journals / Lookup Meetings

Article Tools

Share
Citations

Abstract

A compact, mobile mid-infrared laser spectrometer based on a thermoelectrically (TE) cooled continuous-wave room temperature quantum cascade laser and TE-cooled detectors has been newly developed to demonstrate the applicability of high sensitivity and high precision measurements of atmospheric CO. Performance of the instrument was examined with periodic measurements of reference sample and ambient air at 1 Hz sampling rate and a 1-hourly calibration cycle. The typical precision evaluated from replicate measurements of reference sample over the course of 66-h is 1.41 ppbv. With the utilization of wavelet filtering to improve the spectral SNR and minimize the dispersion of concentration values, a better precision of 0.88 ppbv and a lower detection limit of ~0.4 ppbv with sub-second averaging time have been achieved without reducing the fast temporal response. Allan variance analysis indicates a CO measurement precision of ~0.28 ppbv for optimal integration time of approximate 50 s. The absolute accuracy is limited by the calibration gas standard. This completely thermoelectrically cooled system shows the capability of long-term, unattended and continuous operation at room temperature without complicated cryogenic cooling.

© 2012 OSA

1. Introduction

The greenhouse gasses play an important role in global warming and climate change issues. Carbon monoxide (CO) is not considered a direct greenhouse gas that contributes to atmospheric warming by absorption of IR radiation emitted by the surface of the Earth, but has an indirect radiative influence by affecting the lifetime of methane and the production of tropospheric ozone (O3), through its reaction with the hydroxyl (OH) radical. OH radicals are the main tropospheric oxidant and their abundance determines the lifetimes of strong greenhouse gases (e.g. CH4). Thus the reaction of OH with CO indirectly increases the global warming potential of these gases. High precision measurements of CO in the atmosphere can be used to determine its sources and sinks, consequently improving our understanding of its impact on global warming and climate change. However, atmospheric CO concentrations vary widely geographically and throughout the year, with typical values of about 100 ppbv (parts per billion by volume) in a clean atmosphere, much less than carbon dioxide or methane. The main sources of CO are photo-chemical production from hydrocarbon oxidation and incomplete combustion processes, including biomass burning and fossil fuel use, etc. In the atmosphere, real time in situ measurements of CO concentrations can be important in deducing air mass origins and tracer transport pathways.

Mid-infrared laser spectroscopy has become an impressive tool for detection and quantification of atmospheric trace gases with demonstrated detection sensitivities of ppbv or even sub-ppbv levels. These high sensitivities are based on using fundamental vibrational-rotational molecular absorption bands in the mid-IR spectral range (3-24 μm). For real time field atmospheric monitoring applications, laser sources should be compact, efficient, reliable, have enough light power and operate at near room-temperature. Compared to lead salt diode lasers and some coherent sources based on difference frequency generation and optical parametric oscillators, quantum cascade lasers (QCLs) as a novel type of semiconductor laser developed in 1994 at Bell Labs [1

J. Faist, F. Capasso, D. L. Sivco, C. Sirtori, A. L. Hutchinson, and A. Y. Cho, “Quantum cascade laser,” Science 264(5158), 553–556 (1994). [CrossRef] [PubMed]

], seem to be the best candidate, with typically tens or hundreds of Mw of output power. To date, great progress has been achieved that makes those kinds of lasers not only be operated at cryogenic temperatures in pulsed mode; but also allows operating in continuous wave (CW) mode with output powers up to ~3 W [2

R. F. Curl, F. Capasso, C. Gmachl, A. 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]

]. One of the key advantages of using CW QC lasers as light sources for infrared detection of trace species is that these lasers can operate at near room temperature with thermoelectric cooling, rather than complicated cryogenic fluids cooling, thus reducing the size of the instrument and complexity of operation. One major issue of QCL is the limited wavelength tunability, which restricts the feasibility of probing the entire molecular absorption spectrum, especially of volatile organic compounds and hydrocarbons as well as multi-component species sensing. Recently, new technologies named external cavity QCLs have been demonstrated in an external cavity configuration with a diffraction grating to overcome this shortcoming [3

B. G. Lee, M. A. Belkin, R. Audet, J. MacArthur, L. Diehl, C. Pflugl, F. Capasso, D. C. Oakley, D. Chapman, A. Napoleone, D. Bour, S. Corzine, G. Hofler, and J. Faist, “Widely tunable single-mode quantum cascade laser source for mid-infrared spectroscopy,” Appl. Phys. Lett. 91(23), 231101 (2007). [CrossRef]

,4

R. Maulini, I. Dunayevskiy, A. Lyakh, A. Tsekoun, C. K. N. Patel, L. Diehl, C. Pflugl, and F. Capasso, “Widely tunable high-power external cavity quantum cascade laser operating in continuous-wave at room temperature,” Electron. Lett. 45(2), 107–108 (2009). [CrossRef]

], and have been successfully achieved to scan up to ~200 cm−1 [5

L. Dong, V. Spagnolo, R. Lewicki, and F. K. Tittel, “Ppb-level detection of nitric oxide using an external cavity quantum cascade laser based QEPAS sensor,” Opt. Express 19(24), 24037–24045 (2011). [CrossRef] [PubMed]

], which is approximate ten times wider in range than typically achieved for DFB-QCLs by means of current tuning. Recently, Daylight Solutions Inc. (San Diego, California) reported very significant improvements in EC-QCL performance. They announced a newer Über Tuner model with as much as 375 cm−1 of tuning range and a peak power of over 100mW [6

G. N. Rao and A. Karpf, “External cavity tunable quantum cascade lasers and their applications to trace gas monitoring,” Appl. Opt. 50(4), A100–A115 (2011). [CrossRef] [PubMed]

]. This technology is very attractive for the applications of laser spectroscopy, particularly suitable for simultaneous multi-species trace-gas detection and spectroscopic measurements of broadband absorbers. Consequently, QCLs have been successfully implemented into a variety of spectrometers for different applications, such as atmospheric and environmental monitoring [2

R. F. Curl, F. Capasso, C. Gmachl, A. 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]

], isotopic measurements [7

D. Weidmann, G. Wysocki, C. Oppenheimer, and F. K. Tittel, “Development of a compact quantum cascade laser spectrometer for field measurements of CO2 isotopes,” Appl. Phys. B 80(2), 255–260 (2005). [CrossRef]

], high temperature combustion diagnosis and human breath analysis [8

J. Vanderover, W. Wang, and M. A. Oehlschlaeger, “A carbon monoxide and thermometry sensor based on mid-IR quantum-cascade laser wavelength-modulation absorption spectroscopy,” Appl. Phys. B 103(4), 959–966 (2011). [CrossRef]

,9

B. W. M. Moeskops, H. Naus, S. M. Cristescu, and F. J. M. Harren, “Quantum cascade laser-based carbon monoxide detection on a second time scale from human breath,” Appl. Phys. B 82(4), 649–654 (2006). [CrossRef]

]. To achieve better sensitivities, various techniques known as wavelength or frequency modulation, balanced ratiometric, zero-air subtraction as well as multi-pass optical cell techniques are frequently employed in QCL spectrometers to make them sensitive to even very weak absorption features. Thus, QCL instruments offer many important features including high sensitivity, high selectivity, high precision, fast response and versatility when measuring a number of species of atmospheric interest.

In present paper, we describe a newly designed mid-infrared laser spectrometer based on CW-QCL operating at room temperature for atmospheric CO real-time measurements. Wavelength modulation spectroscopy (WMS) with 2nd harmonic detection technique and a compact multi-pass absorption cell were employed to achieve high sensitivity. The main goal of this study was to investigate the feasibility of high precision real-time and in situ measurements of atmospheric CO at ambient mixing ratios. Details about absorption line selection, instrumental description, measurement strategy, pre-field performance and instrumental errors analysis will be presented in the following sections.

2. Selection of spectrum region

Spectral interferences due to absorption features other than the species of interest obviously affect measurement accuracy and precision, especially in the case of water vapor, whose concentration in the atmosphere is large and highly variable. As stated previously, the detection sensitivity is, to a large extent, dependent on the inherent absorption strength of the target gas under study. In this study we employ one of the strongest features belonging to the CO v1 vibrational band, the R(12) transition located at 2190.0175 cm−1. It was chosen for CO measurements because it adequately avoids spectral interference and provides sufficient absorption strength. According to the HITRAN08 database [10

L. S. Rothman, I. E. Gordon, A. Barbe, D. C. Benner, P. F. Bernath, M. Birk, V. Boudon, L. R. Brown, A. Campargue, J. P. Champion, K. Chance, L. H. Coudert, V. Dana, V. M. Devi, S. Fally, J.-M. Flaud, R. R. Gamache, A. Goldman, D. Jacquemart, I. Kleiner, N. Lacome, W. J. Lafferty, J.-Y. Mandin, S. T. Massie, S. N. Mikhailenko, C. E. Miller, N. Moazzen-Ahmadi, O. V. Naumenko, A. V. Nikitin, J. Orphal, V. I. Perevalov, A. Perrin, A. Predoi-Cross, C. P. Rinsland, M. Rotger, M. Šimečková, M. A. H. Smith, K. Sung, S. A. Tashkun, J. Tennyson, R. A. Toth, A. C. Vandaele, and J. Vander Auwera, “The HITRAN 2008 molecular spectroscopic database,” J. Quant. Spectrosc. Radiat. Transf. 110(9-10), 533–572 (2009). [CrossRef]

], the nearest H2O line is over 0.355 cm−1 away from the R(12) line of CO. A few very weak CO2 absorption features are also present in this region, but the line intensities are approximate 10 orders of magnitude lower than the selected CO transition. H2O and CO2 interferences are thus highly unlikely. A spectral simulation for typical experimental conditions using the HITRAN08 database is shown in Fig. 1 . Also plotted is an enlargement near 2190 cm−1, showing three adjacent relatively weak N2O transitions at lower pressure (50 mbar). However, these underlying N2O lines contribute only slightly to the absorption in both wings of the CO absorption region at the low pressure and good separation. To compensate for this effect, the laser frequency in the experiment was scanned to cover only the CO feature. As a further advantage, the simulation indicates that the QCL spectrometer can be used for simultaneous measurements of CO and N2O in this spectral range. In the present study, we did not use the capabilities for simultaneous CO and N2O analysis, in favour of highly time resolved CO measurement in ambient air.

Fig. 1 HITRAN simulated 2nd harmonic spectrum of the R(12) line of CO and the potential absorptions from other molecules nearby. Conditions are 50 mbar total pressure, 296 K temperature, 36 m optical path length, and relative content as indicated.

3. Experimental details

3.1 Description of the QCL Spectrometer

The room temperature QCL spectrometer (RT-QCLS) combines a commercial QC laser, a sophisticated optical and electronic system, a simple sampling system, and a portable computer controlled system programmed with Labview software that incorporates the electronics for driving the QC laser along with signal generation, data acquisition and on-line signal analysis and display.

The optical set-up of the RT-QCLS and the gas sampling lines are shown in Fig. 2 . All the optical components are mounted on a 50 × 50 × 6.3 cm3 aluminum optical breadboard (Newport System). The commercial single mode QCL emitting in the 4.56 μm wavelength range (Aples Lasers) and operating in CW mode near room temperature was mounted inside a thermoelectrically cooled housing (Alpes LLH-100) cooled with a thermoelectric Peltier cooler. A commercial laser temperature and current controller (Tektronix Munich, model ITC110, Karlsfeld, Germany) was modified to set the operation conditions of the QCL for wavelength tuning. The QCL has a tuning range of 2187.6 - 2202.1 cm−1 over a temperature range of 243-283 K with power range of 0.1-13.5 mW. This wavelength range permits access to the R(11) - R(15) transitions in the CO fundamental vibration-ration band. The R(12) line at 2190 cm−1 was chosen for the present experiments, because it exhibits the smallest overlap with absorption lines from other molecules, specially N2O. For this wavelength, the laser was operated at ~276 K.

Fig. 2 Diagram of the RT-QCLS optical layout and the gas sampling system (MO: Mirror objective; M: flat mirror; Fp: Focal point; OAP: off-axis parabolic mirror; OAE: off-axis ellipsoid; BS: beam splitter; FAS: fluid automation systems; MFC: mass flow controllers; PF: Particle filter; P sensor: Pressure transducer).

The diverging laser beam is firstly collected by a mirror objective (MO), which is composed of two flat mirrors and a 26° off-axis ellipsoid (OAE) with 40 mm and 140 mm focal distances. The MO is adjustable in three orthogonal axes. The laser is located at the nearer focal position (Fp) providing a f/2 aperture and is magnified by a factor 3.5 at the second focal point, which serves as a fixed point in the QCL-spectrometer for optical alignment purpose, with the aid of a visible trace beam from a red diode laser. Next, a 26° off-axis parabolic mirror (OAP) collimates the beam into a parallel beam of 14 mm diameter. Stirred by series of flat mirrors, the laser beam is coupled into an astigmatic Herriott cell (Aerodyne Research, Inc., Model AMAC-36) by a combination of a 23° OAP followed by 23° OAE. The multi-pass cell has a base length of 20 cm, a volume of 0.3 litre and provides a maximum optical path length of 36 m at 182 passes. A Pressure controller maintains the cell pressure at 50 mbar at an inlet sample flow rate of about 1.2 slm. After exiting the sample cell, the beam is re-collimated by an OAE/OAP mirror combination similar to the one used for coupling into the cell. Next, the beam is split by a CaF2 beam splitter (BS) into a reference and signal part. The reference beam is directed through a short reference cell (3.5 cm) with low pressure pure CO for active controlling of the laser wavelength (i.e. frequency locking). Finally both laser beams are focused onto two TE-cooled mercury cadmium telluride (MCT) infrared detectors (PVI-4TE-5, Vigo Systems).

Data acquisition card, power supplies and electronic elements controlled by a FPGA (field-programmable gate array, NI LabView) board are mounted in standard 19-inch rack with the main optical breadboard on top. Control of all electronics, the data acquisition and real-time signal processing and on-line display are completely implemented through a LabView-based graphical user interface software program using a laptop linked via a local area network. Further details of the optical setups, electronics design and sampling line have been described in previous publications due to the use of identical or modified components [11

R. Kormann and H. Fischer, “A compact multi-laser TDLAS for trace gas flux measurements based on a micrometeorological technique,” Proc. SPIE 3758, 162–169 (1999). [CrossRef]

–14

C. L. Schiller, H. Bozem, C. Gurk, U. Parchatka, R. Königstedt, G. W. Harris, J. Lelieveld, and H. Fischer, “Applications of quantum cascade lasers for sensitive trace gas measurements of CO, CH4, N2O and HCHO,” Appl. Phys. B 92(3), 419–430 (2008). [CrossRef]

]. Compared to our previous works, the RT-QCL system is more compact and mobile.

3.2 Spectral measurement approach

Our general approach to atmospheric trace gas measurements using infrared laser spectroscopy with cryogenic cooled CW-QCL has been described previously [13

R. Kormann, R. Königstedt, U. Parchatka, J. Lelieveld, and H. Fischer, “QUALITAS: A mid-infrared spectrometer for sensitive trace gas measurements based on quantum cascade lasers in CW operation,” Rev. Sci. Instrum. 76(7), 075102 (2005). [CrossRef]

,14

C. L. Schiller, H. Bozem, C. Gurk, U. Parchatka, R. Königstedt, G. W. Harris, J. Lelieveld, and H. Fischer, “Applications of quantum cascade lasers for sensitive trace gas measurements of CO, CH4, N2O and HCHO,” Appl. Phys. B 92(3), 419–430 (2008). [CrossRef]

]. WMS with 2nd harmonic detection at an optimum modulation index of ~2.2 and a multi-pass absorption cell are employed to achieve high detection sensitivity. By pumping the atmospheric sample rapidly through a low-pressure cell (approximately 50 mbar), the width of the pressure-broadened absorption line is reduced and overlap with other absorptions in air is minimized, resulting in excellent specificity. To acquire WMS signals, the combination of a low frequency triangle ramp (12.5 Hz) for wavelength tuning and a high frequency sinusoidal modulation (25 kHz) was supplied to the QCL as an addition to the injection current and the 2nd harmonic signal was demodulated at the double modulation frequency using a digital lock-in amplifier programmed with Labview software. Normally a total number of 128 sampling points (including a spectral range of ~0.06 cm−1) were acquired for each spectrum, averaged from 11 sequential laser scans, yielding approximately 0.9 second integration time to improve the signal-to-noise ratio (SNR). The data from each laser scan, including both the ‘up’ and ‘down’ ramps of the triangle waveform, were corrected for slight laser wavelength drifts before co-addition. Finally, the averaged spectra with approximate 1 Hz sampling rate are stored on the acquisition board for post-processing.

3.3 Gas sampling system

The gas handling system comprises a compact oil-free diaphragm pump, a Teflon 3-way connecter, a fluid automation system (FAS) and two mass flow controllers (MFC) driven by the digital output of the data-acquisition card to enable sampling ambient air and reference sample tanks automatically, as shown in Fig. 2 (upper panel). The automatical pressure controllers (MKS Instruments) ensure that the pressure of the sub-line to the sample tanks is set higher than that of sub-line to ambient air when the reference tank sample was sampled, and maintain the astigmatic Herriot cell at a constant pressure around 50 mbar with an inlet sample flow rate of about 1.2 standard liters per minute. For ambient CO measurements, the sampling tube with a Teflon particle filter (PF) directly extends to sample ambient air. A pressure sensor (MKS Instruments, type 122A) was used to monitor the pressure inside the astigmatic Herriot cell. Teflon tubings were used for all connections in the gas sampling system. The sampling protocol involved measuring pressurized bottled air (PBA) with a known CO concentration (325 ppbv calibrated with primary standard samples) for 1-min or 2-min and ambient air for 9 min (see Chapter 5), respectively. Here the reference pressurized air sample measurements are used to test the instrument performance (stability and precision) corresponding to 1-min measurements and for calibration corresponding to 2-min measurements. Indeed, only spectra recorded in the end 21 s during each 2-min measurements period were averaged and used as the calibration spectra to retrieve ambient CO concentrations (see Chapter 5). A typical calibration cycle of one hour interval was selected to evaluate the system performance.

4. Post signal processing

4.1 Determination of CO concentration

The CO mixing ratio of ambient air flowing through the sample cell is determined by measuring the absorption at 2190 cm−1 relative to a calibration gas standard with known concentration, using a multiple linear regression algorithm [15

P. W. Werle, P. Mazzinghi, F. D’Amato, M. De Rosa, K. Maurer, and F. Slemr, “Signal processing and calibration procedures for in situ diode-laser absorption spectroscopy,” Spectrochim. Acta A Mol. Biomol. Spectrosc. 60(8-9), 1685–1705 (2004). [CrossRef] [PubMed]

]. In this procedure, the experimental 2nd harmonic signals are fitted by using a linear combination of a term corresponding to the calibration spectrum and the dc offset. Background spectra are currently not considered here since the absorption signals are recorded with very high SNR (see next section). The fitting procedure is carried out over the entire second harmonic signal spectral signature. Thus, the ambient CO concentrations can be determined by the product of fitted slope and the known calibration gas concentration.

4.2 Wavelet transform

Wavelet transform has attracted much attention recently in digital signal processing due to its multi-resolution characteristics, and has been successfully employed in many signal processing applications [16

B. K. Alsberg, A. M. Woodward, M. K. Winson, J. Rowland, and D. B. Kell, “Wavelet denoising of infrared spectra,” Analyst (Lond.) 122(7), 645–652 (1997). [CrossRef]

]. We applied the wavelet-based de-noising technique to our QCL spectrometer in post signal processing to improve the SNR of the raw spectra and diminish the dispersion of final concentration values. The algorithm is implemented with Python® scripting language using our own programs. The Python Toolbox for wavelets was used as the library of wavelet filter coefficients. Although there are many types of wavelets, we restrict ourselves in this study to Daubechies and Symmlet family wavelets by soft-thresholding scheme for spectroscopic de-noising [17

D. L. Donoho, “De-noising by soft-thresholding,” IEEE Trans. Inf. Theory 41(3), 613–627 (1995). [CrossRef]

], due to their relatively better performance by large numbers of simulated tests and comparison.

An application of the wavelet filtering to the measured ambient CO signal (specially selected with the lower SNR) with 0.88 second averaging time prior to signal processing is shown in Fig. 3 . Before and after the wavelet filter algorithm had been applied to the raw spectrum, the data analysis reported an almost identical concentration of 92.97 ppb, but with a fitting error of 0.238 ppb that improved to 0.059 ppb. Wavelet transform improves the SNR (2σ) of 21 without wavelet filtering to 118. The SNR was calculated from the ratio of peak to peak values between 2nd harmonic signal and residual. This indicates a significant improvement by a factor of 4.0 in the fitting precision and 5.6 in SNR by adapting the wavelet filter to the raw spectrum. That means a detection limit (1σ) at sub-ppb level (~0.4 ppb Hz−1) can be achieved with the utilization of wavelet filtering. The resulting residuals, which have been derived from the difference between the filtered or the unfiltered spectra and linear fitting, show a periodic structure, as would be expected from a fringe background structure.

Fig. 3 Typical example of ambient CO spectrum of R(12) transition recorded with the RT-QCL spectrometer (0.88 s averaging time) before and after the application wavelet filtering (upper panel), and the residual (lower panel).

5. Pre-field performance of the RT-QCL spectrometer

The RT-QCL spectrometer and gas sampling systems (without pump) were mounted in standard 19-inch rack and installed in an air-conditioned mobile laboratory container (6.0×2.4×2.9 m3) used previously for various ground- and ship-based field campaigns. An exhaust fan mounted on the container roof was continuously refreshing the cabin air in order to decrease potential in-door CO accumulation. For ambient CO measurements, a sampling tube extends through a hole in the wall of the container to sample ambient air. To demonstrate the instrument performance (repeatability and precision), alternate measurements of CO from the known reference PBA sample (CO = 325 ppb) and ambient air were continuously monitored from 8 July through 11 July at 10-min or 11-min intervals.

Figure 4 shows a typical segment of alternating measurements of CO in ambient air and a reference sample (PBA, CO = 325 ppb) within a 24-h period. The upper trace at 325 ppb level is the reference sample from PAB, while the lower trace shows the real-time ambient CO mixing ratios. Discrete data points in the middle of lower panel result from the time delay of gas sampling lines, as shown in the right upper panel of Fig. 4. The asymmetric response time (i.e. longer time lag from switching reference tank to ambient air) results from the use of a Teflon particle filter. This issue has been confirmed and resolved to symmetric response time of 3 s in the later lab and field measurements. The solid green line depicts the application of wavelet filtering. It’s worth noting that no additional delay and loss of high temporal resolution are introduced by the wavelet filtering compared to the unfiltered raw concentration data. The short-term (about 1-min) measurement precisions obtained from the reference sample (PBA) in both cases are also presented (upper right panel) for clarity and comparison. With the utilization of the wavelet filtering, not only the SNR of the spectra, but also the dispersion of concentration values minimizes, consequently, the measurement precision is improved further. Details of this concept and comparisons with other kind of digital filter techniques, such as Wiener filter, Fast Fourier Transform and Kalman filter, will be discussed in a further publication.

Fig. 4 Alternating measurements of reference CO sample (PBA, CO = 325ppb) and ambient air during a 24-h data period with 1 Hz sampling rate and the application of wavelet filter (lower panel). Both upper panels present a zoom-into a typical 1-h measurement cycle (upper left panel) and approximate 3-min measurement interval (upper right panel), respectively. (Details see text)

Ambient CO measurements were obtained in Mainz. The sampling site is located at an urban region. As one can see from Fig. 4, after 20:30 h, it appears that ambient CO concentrations accumulated over this site, and show an enhancement until 10:30 h of July 9, and then the CO concentrations decreased to the background level (~100 ppb). This is mainly due to a decrease in the boundary layer height after sun-set and constant emissions into a shallow night-time boundary layer.

5.1 Measurement precision

In this study, we investigate three different measures of the instrument precisions to evaluate the system performance: fit precision (σF), replicate precision (σR) and precision from the Allan variance (σA2), reported here at the 1σ level here. The fit precision is obtained directly from the error of the linear regression procedure and represents the precision for the calculation of individual concentration values. Replicate precision is obtained from the standard deviation of replicate measurements of a reference sample with constant concentration. This presents a more meaningful evaluation of instrument performance than the fit precision, since it includes mid to long term drifts of the instrument. Similar to replicate precision, Allan variance estimates of the precision are also obtained from repetitively measuring a reference sample (see next section). σF and σR were evaluated from approximate 66-h of successive measurements with 1 Hz sampling rate, classified into two categories, i.e. before and after the application of the wavelet filtering to raw data, as summarized in Table 1 .

Table 1  Summary of Different Measures of the Mean Instrument Precisions Before and after the Application of the Wavelet Filtering
σF
(% or ppbv)σR
(% or ppbv)σF
(% or ppbv)σR
(% or ppbv)
Before wavelet
After wavelet
0.410 (1.332)0.436 (1.413)0.345 (1.121)0.272 (0.880)

It can be seen that the mean fit precision σF is almost in good agreement with the replicate precision σR, independent of the application of wavelet filtering to raw spectra. This is mainly due to the high SNR of the recorded spectra here. In generally the fit precisions are improved slightly after the utilization of wavelet filtering. The majority (> 94%) of the fit precisions during the consecutive 66-h measurement cycle are improved by the wavelet filtering between 1.2 and 7.5 times. However, with the utilization of wavelet filtering further to raw concentration values, the replicate precision improved by a mean factor of 1.6 times relative to those directly obtained from raw spectra. The high measurement precisions achieved here can be compared with previous result demonstrated by Provencal et al. [18

R. Provencal, M. Gupta, T. G. Owano, D. S. Baer, K. N. Ricci, A. O’Keefe, and J. R. Podolske, “Cavity-enhanced quantum-cascade laser-based instrument for carbon monoxide measurements,” Appl. Opt. 44(31), 6712–6717 (2005). [CrossRef] [PubMed]

] with cavity-enhanced CW-QCL system operating at 2172.8 cm−1 (40 cm optical cavity yields an effective path length of 400 m). In the similar procedure but at 30-min intervals, they reported an instrument repeated precision of 0.44% (or 2.4 ppb) over the entire 10-h measurement period.

5.2 Characterization of system stability

Allan variance as a statistical technique has been widely used to assess an instrument precision for a range of measurement integration periods to deduce the optimal integration period by continuously measuring a single constant concentration sample or pure background structure. Within the white noise dominated region, the square root of the Allan variance is proportional to the precision of a given instrument. Here an additional laboratory series of measurements of a CO sample with 269 ppbv over time periods of approximate 10 min were performed. Allan variance is utilized to analyze the stability and precision of the system. As shown in Fig. 5 , the Allan variance σA2 is plotted on a log-log scale versus averaging time τ. It indicates that the 1-s measurement precision is σA ≈1.26 ppbv, which is almost consistent with the averaging replicate precisions of 1.41 ppbv obtained above without the application of wavelet filtering. Moreover, the Allan plot shows an optimum integration time of ~50 s, corresponding to a precision of ~0.28 ppbv on average. The decreasing solid line gives the theoretically expected behaviour (proportional to 1/τ) of a system within the white noise dominated region, namely the Allan variance decreases with increasing integration time, after which system drifts start to dominate. As performed here, the 1-s CO results can be further averaged into longer time intervals determined based upon instrument optimal stability times to improve measurement precision, but at the expense of good time resolution in field applications.

Fig. 5 Allan variance plot acquired from series of measurements on standard CO sample (269 ppbv) with 1 Hz sampling rate in the laboratory conditions.

5.3 Instrument error analysis

Finally, errors affecting instrument measurement precision and accuracy are discussed in this section. Measurement accuracy is the deviation of the measured value relative to the “true” values. It can be affected by all error sources, not only time-dependent but also time-independent. Measurement precision is the degree to which repeated measurements show the same results. It only depends upon the time-dependent errors. The Wikipedia article at en.wikipedia.org/wiki/Accuracy_and_precision gives good descriptions of accuracy and precision.

The main errors affecting measurement precision in any spectroscopic instrument may be categorized into two error sources: interference noises due to spectral and optical interferences, and electronic noises from detectors, lasers, and additional electronics. As discussed previously, the error due to interference from other atmospheric constituents is completely negligible. By observing the oscillating feature of the residual between experimental spectrum and theoretical simulation, the presence of optical fringes which disturb the measurement is evident, especially for low levels of concentrations. The calculated fringe period shows that such fringes are created in an equivalent cavity with an optical path-length of about 40 cm, which is equal to one round trip inside the Herriott cell. Thus we deduce that the ‘etalon effect’ is very likely associated with the multi-pass absorption cell. The sophisticated electronics have been used for high precision trace gases measurements for many years [13

R. Kormann, R. Königstedt, U. Parchatka, J. Lelieveld, and H. Fischer, “QUALITAS: A mid-infrared spectrometer for sensitive trace gas measurements based on quantum cascade lasers in CW operation,” Rev. Sci. Instrum. 76(7), 075102 (2005). [CrossRef]

]. The commercial MCT infrared detectors operating at TE-cooled mode have a noise level of ~300 nV·Hz1/2 and detectivity of ~2.2 × 1011 cm·Hz1/2/W, which is inferior to their cryogenic counterparts. Recently, the frequency noise properties (1/f trend below 10 kHz) of a similar QCL used here has been reported by Tombez et al [19

L. Tombez, J. Di Francesco, S. Schilt, G. Di Domenico, J. Faist, P. Thomann, and D. Hofstetter, “Frequency noise of free-running 4.6 μm distributed feedback quantum cascade lasers near room temperature,” Opt. Lett. 36(16), 3109–3111 (2011). [CrossRef] [PubMed]

]. Generally, a fast sweep rate can be used to effectively suppress the effects of 1/f noise. Unfortunately, in the present work, the laser sweep rate was limited to 12.5 Hz. Anyway, as shown wavelet transform was successful to remove the influence of low frequency noises. These characteristics are currently the main sensitivity limitation of the spectrometer. Presently, we think that instrument drift due to the optical interference generated in the Herriott cell limits the precision of the instrument.

The total uncertainty of a measurement contains contributions due to the sampling, the calibration scale, the transfer of the calibration scale to the instrument, the instrument linearity, the repeatability (short term variability) or the reproducibility (longer term variability, including drift) of the instrument, and uncertainty introduced during data processing, etc. In general, these contributions are independent, and therefore, the combined uncertainty is the square-root of the sum of the squared individual contributions. In our case, the total uncertainty of the calibration scale has a dominant influence on the uncertainty of the CO measurements compared to other potential uncertainties, since the primary calibration standards (Scott Marrin Specialty Gases, Inc.) used to determine ambient CO concentrations typically have a stated uncertainty between ± 5% and ± 10%. These standards have been tested with various instruments (Aero-laser, TDLS and GC-Hgo) in our institute, and traced to the NOAA scale. Thus, the ultimate accuracy for CO measurements was conservatively estimated to be about 10% or better.

6. Conclusions

A compact, mobile mid-infrared thermoelectrically cooled CW-RT-QCL spectrometer operating at wavelength around 4.56 μm integrated with a commercially available astigmatic multi-pass Herriott cell has been developed and demonstrated for real-time atmospheric CO measurements with long calibration cycle and fast response. WMS with 2nd harmonic detection technique was employed to achieve high sensitivity detection. Wavelet filtering was successfully applied to improve instrument sensitivity and measurement precision in post-signal processing. Typically sub-second (0.88 s) averaging precision of 1.41 ppbv evaluated from alternate measurements of standard sample and ambient air with 1 Hz sampling rate was achieved over the course of 66 h. With the utilization of wavelet filtering to remove noises and minimize the dispersion of concentration values, a better precision of 0.88 ppbv has been achieved without introducing any delay of response time and loss of high temporal resolution. Allan variance analysis depicts an ultimate precision of ~0.28 ppbv can be obtained at the optimal integration time of approximate 50 s.

A comparison of instrument performance with other QCL-based detection systems employed for monitoring CO or other trace gases is not straightforward, in view of rather different experimental configurations and definitions of sensitivity as well as sometimes a lack of specific information. However, a simple comparison is made here with commercial instrumentation from Aerodyne Research Inc. and their collaborators employing different detection schemes at similar mid-IR wavelengths around 4.6 µm. The in-field sensitivity achieved here is limited mainly by residual interference between optical elements. Our in-field detection limit (~0.4 ppb Hz−1) for atmospheric CO measurement by using the wavelet filtering can be compared with previous result of 0.2 ppb obtained with a CW-QCL system employing a 76 m optical path length at 100-s averaging time [20

R. Jiménez, S. Herndon, J. H. Shorter, D. D. Nelson, J. B. McManus, and M. S. Zahniser, “Atmospheric trace gas measurements using a dual quantum-cascade laser mid-infrared absorption spectrometer,” Proc. SPIE 5738, 318–331 (2005). [CrossRef]

], and a 1.42 times absorption line strength used than here.

Application of appropriate digital filters maybe provides the best compromise between measurement precisions and fast-temporal resolution in practical fields, where drifts of background signals and laser frequency could be more prevalent under limited experimental conditions. This completely thermoelectrically cooled system is capable of long-term, unattended and continuous operation at room temperature without cryogenic cooling of either laser or detector. As performed here, the measurement precision can be improved using shorter periodic calibration intervals of less than the averaging at which the Allan variance begins to increase. Real-time background subtraction, which has not yet been implemented, can also further improve the system performance. The absolute accuracy will obviously depend upon the accuracy of calibration gas standards. Future work is focusing on improving optical alignment and laser temperature stability, as well as developing an airborne version for flight campaign in near future.

Acknowledgments

The authors would like to thank the anonymous reviewers for their valuable comments and suggestions to improve the quality of the paper. They are also grateful to Dr. Peter Werle of Karlsruhe Institute of Technology KIT and Prof. Weidong Cheng of Université du Littoral Côte d’Opale for useful discussions.

References and links

1.

J. Faist, F. Capasso, D. L. Sivco, C. Sirtori, A. L. Hutchinson, and A. Y. Cho, “Quantum cascade laser,” Science 264(5158), 553–556 (1994). [CrossRef] [PubMed]

2.

R. F. Curl, F. Capasso, C. Gmachl, A. 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]

3.

B. G. Lee, M. A. Belkin, R. Audet, J. MacArthur, L. Diehl, C. Pflugl, F. Capasso, D. C. Oakley, D. Chapman, A. Napoleone, D. Bour, S. Corzine, G. Hofler, and J. Faist, “Widely tunable single-mode quantum cascade laser source for mid-infrared spectroscopy,” Appl. Phys. Lett. 91(23), 231101 (2007). [CrossRef]

4.

R. Maulini, I. Dunayevskiy, A. Lyakh, A. Tsekoun, C. K. N. Patel, L. Diehl, C. Pflugl, and F. Capasso, “Widely tunable high-power external cavity quantum cascade laser operating in continuous-wave at room temperature,” Electron. Lett. 45(2), 107–108 (2009). [CrossRef]

5.

L. Dong, V. Spagnolo, R. Lewicki, and F. K. Tittel, “Ppb-level detection of nitric oxide using an external cavity quantum cascade laser based QEPAS sensor,” Opt. Express 19(24), 24037–24045 (2011). [CrossRef] [PubMed]

6.

G. N. Rao and A. Karpf, “External cavity tunable quantum cascade lasers and their applications to trace gas monitoring,” Appl. Opt. 50(4), A100–A115 (2011). [CrossRef] [PubMed]

7.

D. Weidmann, G. Wysocki, C. Oppenheimer, and F. K. Tittel, “Development of a compact quantum cascade laser spectrometer for field measurements of CO2 isotopes,” Appl. Phys. B 80(2), 255–260 (2005). [CrossRef]

8.

J. Vanderover, W. Wang, and M. A. Oehlschlaeger, “A carbon monoxide and thermometry sensor based on mid-IR quantum-cascade laser wavelength-modulation absorption spectroscopy,” Appl. Phys. B 103(4), 959–966 (2011). [CrossRef]

9.

B. W. M. Moeskops, H. Naus, S. M. Cristescu, and F. J. M. Harren, “Quantum cascade laser-based carbon monoxide detection on a second time scale from human breath,” Appl. Phys. B 82(4), 649–654 (2006). [CrossRef]

10.

L. S. Rothman, I. E. Gordon, A. Barbe, D. C. Benner, P. F. Bernath, M. Birk, V. Boudon, L. R. Brown, A. Campargue, J. P. Champion, K. Chance, L. H. Coudert, V. Dana, V. M. Devi, S. Fally, J.-M. Flaud, R. R. Gamache, A. Goldman, D. Jacquemart, I. Kleiner, N. Lacome, W. J. Lafferty, J.-Y. Mandin, S. T. Massie, S. N. Mikhailenko, C. E. Miller, N. Moazzen-Ahmadi, O. V. Naumenko, A. V. Nikitin, J. Orphal, V. I. Perevalov, A. Perrin, A. Predoi-Cross, C. P. Rinsland, M. Rotger, M. Šimečková, M. A. H. Smith, K. Sung, S. A. Tashkun, J. Tennyson, R. A. Toth, A. C. Vandaele, and J. Vander Auwera, “The HITRAN 2008 molecular spectroscopic database,” J. Quant. Spectrosc. Radiat. Transf. 110(9-10), 533–572 (2009). [CrossRef]

11.

R. Kormann and H. Fischer, “A compact multi-laser TDLAS for trace gas flux measurements based on a micrometeorological technique,” Proc. SPIE 3758, 162–169 (1999). [CrossRef]

12.

F. G. Wienhold, H. Fischer, P. Hoor, V. Wagner, R. Königstedt, G. W. Harris, J. Anders, R. Grisar, M. Knothe, W. J. Riedel, F.-J. Lübken, and T. Schilling, “TRISTAR–a tracer in situ TDLAS for atmospheric research,” Appl. Phys. B 67(4), 411–417 (1998). [CrossRef]

13.

R. Kormann, R. Königstedt, U. Parchatka, J. Lelieveld, and H. Fischer, “QUALITAS: A mid-infrared spectrometer for sensitive trace gas measurements based on quantum cascade lasers in CW operation,” Rev. Sci. Instrum. 76(7), 075102 (2005). [CrossRef]

14.

C. L. Schiller, H. Bozem, C. Gurk, U. Parchatka, R. Königstedt, G. W. Harris, J. Lelieveld, and H. Fischer, “Applications of quantum cascade lasers for sensitive trace gas measurements of CO, CH4, N2O and HCHO,” Appl. Phys. B 92(3), 419–430 (2008). [CrossRef]

15.

P. W. Werle, P. Mazzinghi, F. D’Amato, M. De Rosa, K. Maurer, and F. Slemr, “Signal processing and calibration procedures for in situ diode-laser absorption spectroscopy,” Spectrochim. Acta A Mol. Biomol. Spectrosc. 60(8-9), 1685–1705 (2004). [CrossRef] [PubMed]

16.

B. K. Alsberg, A. M. Woodward, M. K. Winson, J. Rowland, and D. B. Kell, “Wavelet denoising of infrared spectra,” Analyst (Lond.) 122(7), 645–652 (1997). [CrossRef]

17.

D. L. Donoho, “De-noising by soft-thresholding,” IEEE Trans. Inf. Theory 41(3), 613–627 (1995). [CrossRef]

18.

R. Provencal, M. Gupta, T. G. Owano, D. S. Baer, K. N. Ricci, A. O’Keefe, and J. R. Podolske, “Cavity-enhanced quantum-cascade laser-based instrument for carbon monoxide measurements,” Appl. Opt. 44(31), 6712–6717 (2005). [CrossRef] [PubMed]

19.

L. Tombez, J. Di Francesco, S. Schilt, G. Di Domenico, J. Faist, P. Thomann, and D. Hofstetter, “Frequency noise of free-running 4.6 μm distributed feedback quantum cascade lasers near room temperature,” Opt. Lett. 36(16), 3109–3111 (2011). [CrossRef] [PubMed]

20.

R. Jiménez, S. Herndon, J. H. Shorter, D. D. Nelson, J. B. McManus, and M. S. Zahniser, “Atmospheric trace gas measurements using a dual quantum-cascade laser mid-infrared absorption spectrometer,” Proc. SPIE 5738, 318–331 (2005). [CrossRef]

OCIS Codes
(120.0280) Instrumentation, measurement, and metrology : Remote sensing and sensors
(280.1120) Remote sensing and sensors : Air pollution monitoring
(300.6340) Spectroscopy : Spectroscopy, infrared
(140.5965) Lasers and laser optics : Semiconductor lasers, quantum cascade

ToC Category:
Remote Sensing

History
Original Manuscript: January 31, 2012
Revised Manuscript: February 23, 2012
Manuscript Accepted: February 23, 2012
Published: March 19, 2012

Citation
Jingsong Li, Uwe Parchatka, Rainer Königstedt, and Horst Fischer, "Real-time measurements of atmospheric CO using a continuous-wave room temperature quantum cascade laser based spectrometer," Opt. Express 20, 7590-7601 (2012)
http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-20-7-7590


Sort:  Author  |  Year  |  Journal  |  Reset  

References

  1. J. Faist, F. Capasso, D. L. Sivco, C. Sirtori, A. L. Hutchinson, and A. Y. Cho, “Quantum cascade laser,” Science264(5158), 553–556 (1994). [CrossRef] [PubMed]
  2. R. F. Curl, F. Capasso, C. Gmachl, A. 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]
  3. B. G. Lee, M. A. Belkin, R. Audet, J. MacArthur, L. Diehl, C. Pflugl, F. Capasso, D. C. Oakley, D. Chapman, A. Napoleone, D. Bour, S. Corzine, G. Hofler, and J. Faist, “Widely tunable single-mode quantum cascade laser source for mid-infrared spectroscopy,” Appl. Phys. Lett.91(23), 231101 (2007). [CrossRef]
  4. R. Maulini, I. Dunayevskiy, A. Lyakh, A. Tsekoun, C. K. N. Patel, L. Diehl, C. Pflugl, and F. Capasso, “Widely tunable high-power external cavity quantum cascade laser operating in continuous-wave at room temperature,” Electron. Lett.45(2), 107–108 (2009). [CrossRef]
  5. L. Dong, V. Spagnolo, R. Lewicki, and F. K. Tittel, “Ppb-level detection of nitric oxide using an external cavity quantum cascade laser based QEPAS sensor,” Opt. Express19(24), 24037–24045 (2011). [CrossRef] [PubMed]
  6. G. N. Rao and A. Karpf, “External cavity tunable quantum cascade lasers and their applications to trace gas monitoring,” Appl. Opt.50(4), A100–A115 (2011). [CrossRef] [PubMed]
  7. D. Weidmann, G. Wysocki, C. Oppenheimer, and F. K. Tittel, “Development of a compact quantum cascade laser spectrometer for field measurements of CO2 isotopes,” Appl. Phys. B80(2), 255–260 (2005). [CrossRef]
  8. J. Vanderover, W. Wang, and M. A. Oehlschlaeger, “A carbon monoxide and thermometry sensor based on mid-IR quantum-cascade laser wavelength-modulation absorption spectroscopy,” Appl. Phys. B103(4), 959–966 (2011). [CrossRef]
  9. B. W. M. Moeskops, H. Naus, S. M. Cristescu, and F. J. M. Harren, “Quantum cascade laser-based carbon monoxide detection on a second time scale from human breath,” Appl. Phys. B82(4), 649–654 (2006). [CrossRef]
  10. L. S. Rothman, I. E. Gordon, A. Barbe, D. C. Benner, P. F. Bernath, M. Birk, V. Boudon, L. R. Brown, A. Campargue, J. P. Champion, K. Chance, L. H. Coudert, V. Dana, V. M. Devi, S. Fally, J.-M. Flaud, R. R. Gamache, A. Goldman, D. Jacquemart, I. Kleiner, N. Lacome, W. J. Lafferty, J.-Y. Mandin, S. T. Massie, S. N. Mikhailenko, C. E. Miller, N. Moazzen-Ahmadi, O. V. Naumenko, A. V. Nikitin, J. Orphal, V. I. Perevalov, A. Perrin, A. Predoi-Cross, C. P. Rinsland, M. Rotger, M. Šimečková, M. A. H. Smith, K. Sung, S. A. Tashkun, J. Tennyson, R. A. Toth, A. C. Vandaele, and J. Vander Auwera, “The HITRAN 2008 molecular spectroscopic database,” J. Quant. Spectrosc. Radiat. Transf.110(9-10), 533–572 (2009). [CrossRef]
  11. R. Kormann and H. Fischer, “A compact multi-laser TDLAS for trace gas flux measurements based on a micrometeorological technique,” Proc. SPIE3758, 162–169 (1999). [CrossRef]
  12. F. G. Wienhold, H. Fischer, P. Hoor, V. Wagner, R. Königstedt, G. W. Harris, J. Anders, R. Grisar, M. Knothe, W. J. Riedel, F.-J. Lübken, and T. Schilling, “TRISTAR–a tracer in situ TDLAS for atmospheric research,” Appl. Phys. B67(4), 411–417 (1998). [CrossRef]
  13. R. Kormann, R. Königstedt, U. Parchatka, J. Lelieveld, and H. Fischer, “QUALITAS: A mid-infrared spectrometer for sensitive trace gas measurements based on quantum cascade lasers in CW operation,” Rev. Sci. Instrum.76(7), 075102 (2005). [CrossRef]
  14. C. L. Schiller, H. Bozem, C. Gurk, U. Parchatka, R. Königstedt, G. W. Harris, J. Lelieveld, and H. Fischer, “Applications of quantum cascade lasers for sensitive trace gas measurements of CO, CH4, N2O and HCHO,” Appl. Phys. B92(3), 419–430 (2008). [CrossRef]
  15. P. W. Werle, P. Mazzinghi, F. D’Amato, M. De Rosa, K. Maurer, and F. Slemr, “Signal processing and calibration procedures for in situ diode-laser absorption spectroscopy,” Spectrochim. Acta A Mol. Biomol. Spectrosc.60(8-9), 1685–1705 (2004). [CrossRef] [PubMed]
  16. B. K. Alsberg, A. M. Woodward, M. K. Winson, J. Rowland, and D. B. Kell, “Wavelet denoising of infrared spectra,” Analyst (Lond.)122(7), 645–652 (1997). [CrossRef]
  17. D. L. Donoho, “De-noising by soft-thresholding,” IEEE Trans. Inf. Theory41(3), 613–627 (1995). [CrossRef]
  18. R. Provencal, M. Gupta, T. G. Owano, D. S. Baer, K. N. Ricci, A. O’Keefe, and J. R. Podolske, “Cavity-enhanced quantum-cascade laser-based instrument for carbon monoxide measurements,” Appl. Opt.44(31), 6712–6717 (2005). [CrossRef] [PubMed]
  19. L. Tombez, J. Di Francesco, S. Schilt, G. Di Domenico, J. Faist, P. Thomann, and D. Hofstetter, “Frequency noise of free-running 4.6 μm distributed feedback quantum cascade lasers near room temperature,” Opt. Lett.36(16), 3109–3111 (2011). [CrossRef] [PubMed]
  20. R. Jiménez, S. Herndon, J. H. Shorter, D. D. Nelson, J. B. McManus, and M. S. Zahniser, “Atmospheric trace gas measurements using a dual quantum-cascade laser mid-infrared absorption spectrometer,” Proc. SPIE5738, 318–331 (2005). [CrossRef]

Cited By

 Alert me when this paper is cited

OSA is able to provide readers links to articles that cite this paper by participating in CrossRef's Cited-By Linking service. CrossRef includes content from more than 3000 publishers and societies. In addition to listing OSA journal articles that cite this paper, citing articles from other participating publishers will also be listed.

Figures

Fig. 1 Fig. 2 Fig. 3
 
Fig. 4 Fig. 5
 

« Previous Article  |  Next Article »

OSA is a member of CrossRef.

CrossCheck Deposited