Impact of H2O broadening effect on atmospheric CO and N2O detection near 4.57 μm

Accepted Manuscript Impact of H2O broadening effect on atmospheric CO and N2O detection near 4.57 μm Hao Deng, Juan Sun, Ningwu Liu, Hongliang Wang, Benli Yu, Jingsong Li PII: DOI: Reference:

S0022-2852(16)30311-3 http://dx.doi.org/10.1016/j.jms.2016.11.001 YJMSP 10804

To appear in:

Journal of Molecular Spectroscopy

Received Date: Revised Date: Accepted Date:

18 July 2016 12 September 2016 4 November 2016

Please cite this article as: H. Deng, J. Sun, N. Liu, H. Wang, B. Yu, J. Li, Impact of H2O broadening effect on atmospheric CO and N2O detection near 4.57 μm, Journal of Molecular Spectroscopy (2016), doi: http://dx.doi.org/ 10.1016/j.jms.2016.11.001

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Impact of H2O broadening effect on atmospheric CO and N2O detection near 4.57 µm Hao Denga, Juan Suna, Ningwu Liua, Hongliang Wangb, Benli Yua and Jingsong Lia,* a

Key Laboratory of Opto-Electronic Information Acquisition and Manipulation of Ministry of Education, Anhui University, 23061 Hefei, China.

b

National Deep Sea Center, State Oceanic Administration, 266237 Qingdao, China *Author to whom correspondence should be addressed; E-Mail: [email protected] Tel./Fax: +86-551-6386-1490.

Abstract A tunable quantum cascade laser spectrometer (QCLS) was used to study H2O broadening coefficients for CO and N2O transitions at 4.57 µm region, which contains well-characterized and relatively isolated transitions of appropriate line strengths for sensitive gas detection. The influence of H2O broadening effect on CO R(11) and N2O P(38e) transitions at 2186.639 cm-1 and 2187.099 cm-1, respectively, was detailed investigated. Our measurements indicate that H2O broadening coefficients are 1.8 and 1.9 times higher than the corresponding air-broadening parameters, respectively. Based on the experimental data, our simulation confirmed that the WMS-2f shapes of CO and N2O lines will be significantly affected by variations of the water vapor mixing ratio, while no significant dependence on target concentration, and prove that the difference between air- and H2O-broadenings thus cannot be neglected if one wants to measure gas concentrations in a high humid environment with a sub-percent precision.

Keywords: QCLS; Nitrous oxide; Carbon monoxide; H2O broadening effect 1

1. Introduction Nitrous oxide (N2O) and carbon monoxide (CO) are present in the atmosphere as trace gases, which play an important role on the process of global warming and climate change. Nitrous oxide is an anthropogenic greenhouse gas with a global warming potential about 200-300 times that of carbon dioxide (CO2) and a long atmospheric lifetime of approximate 120 years [1]. Quantum cascade laser (QCL) based N2O and CO sensors have been developed previously for their concentrations measurements using a variety of spectroscopic strategies, such as multipass absorption spectroscopy [2,3], photoacoustic spectroscopy [4] or quartz-enhanced photoacoustic spectroscopy [5], cavity-enhanced absorption spectroscopy [6,7], and wavelength modulation spectroscopy [8,9]. Improved sensors with higher precisions and accuracies would be valuable tools for atmospheric applications. Fig. 1 shows the nitrous oxide and carbon monoxide absorption line strengths in the mid-infrared region, the strongest absorbance is in their fundamental ro-vibrational band near 4.6 µm. However, the presence of gases other than the target species, can be a source of uncertainty for atmospheric measurements. Water vapor (H2O) shows a very strong absorber of electromagnetic radiation in the IR region (as can see in Fig. 1), thus it becomes a main interference in atmospheric trace gases measurements. Generally, by reducing sample gas pressure, the spectral interference of H2O and other species can be effectively reduced in real time and, therefore, enhanced specificity, improved accuracy, lower detection limit, and faster response all become achievable. However, even in the absence of absorption interference, a spectroscopic sensor may be affected by the sample composition as a result of a modified line broadening coefficients. Unlike other atmospheric gases, the distribution of water vapor (H2O) in the atmosphere varies with high dynamic range (about 0.17 ppm – 4.5%), which strongly dependent on time, location, and altitude. Therefore, the broadening contribution due to water vapor mixing ratio variation would need to be known with minimal uncertainty for high-accuracy data retrievals, especially in a humid atmosphere.

2

Fig.1. Nitrous oxide, carbon monoxide and water vapor absorption line strengths in the mid-infrared region [10].

Currently, the spectroscopic parameters for major atmospheric gas molecules, including the line strength and the broadening coefficients due to dry air are contained in atmospheric spectroscopy databases such as HITRAN [10] and GEISA [11]. Unfortunately, poor data are available about the H2O reduced broadening parameters. Indeed, H2O has already been found to be an effective foreign broadener for CO2 with

γ H2O ≈ 1.8γ air near 1.57 µm [12] and typically twice greater than those by air at 4.3 µm [13], this difference that cannot be neglected for high-accuracy CO2 retrievals. Recently, a tunable QCL absorption spectrometer was used to accurately study H2O and other molecules broadening for six ammonia (NH3) transitions lines near 1103.46 cm-1, 3-4 times higher than the corresponding air-broadening parameters [14]. Water vapor induced pressure broadening on NH3 transitions near 1.51 µm have also been studied by Schilt [15]. Vess et al. reported that H2O broadening parameters for six of the strongest transitions of O2 A-band (near 13100 cm-1) with a laser-based photoacoustic spectroscopy, the measured values were nominally 1.5-2 times greater than the corresponding air-broadening parameters [16]. Most recently, accurate and precise measurements of water-broadening of oxygen transitions have been extended 3

to millimeter and sub-millimeter wavelengths [17,18]. Moreover, the published results show that H2O-broadenings of methane (CH4) lines are, on average, 34% larger than those for dry air [19]. All these results prove that the difference between air- and H2O-broadenings thus cannot be neglected if one wants to measure gas concentrations in a high humid environment with a sub-percent precision. In this paper, we present measurements of N2O and CO transitions broadened by H2O using a tunable quantum cascade laser spectrometer (QCLS). Our goal was to accurately measure the H2O broadening parameters for two N2O and CO absorption lines near 4.57 µm, which are suitable for highly sensitive atmospheric applications, and to quantify the extent to which broadening by this species affects for atmospheric trace gases detection. Additionally, the effect of humidity on the R(11) transition of CO and the P(38e) transition of N2O absorption features are theoretically analyzed by using the observed H2O broadened coefficients. 2. Experimental details The schematic setup of the QCLS system is depicted in Fig.2. A continuous-wave DFB-QCL (Alpes Lasers, Switzerland) with a central emitting wavelength of 4.57 µm was specially selected as an excitation source to target the absorption lines of CO, H2O and N2O between 2186.5 and 2187.5 cm−1 for simultaneous multiple gases detection. The QCL was mounted in a Laboratory Laser Housing (Alpes LLH-100) package equipped with a thermoelectric Peltier cooler. The laser can be tunable over 2180 cm-1 to 2200 cm-1 by properly adjusting the laser temperature and injection current, which was achieved by using a TC-3 temperature controller (Alpes Lasers) and a precision current source (ILX Lightwave LDX 3232), respectively. The output QCL beam is firstly collected by a home-made mirror objective (MO), which is composed of two flat mirrors and a 26° off-axis ellipsoid (OAE) [20]. After passing through a beam splitter (BS), the main out-coming beam was collimated by a 90° off-axis parabolic (OAP) mirror and directed to a single-pass gas cell with path length of 29.6 cm for H2O broadening measurements. Another fraction of the beam reflected from the BS is directed through a second cell with path 4

length of 4.33 cm for air broadening measurements. Both channel signals are collected by two similar TE-cooled mercury cadmium telluride (MCT) infrared detector (PVI-4TE-5, Vigo Systems). Finally, the output signals from both detectors were sent to a data acquisition system (NI-6212, National Instruments) based on LabVIEW software and a laptop computer for data storage and signal processing. According to HITRAN spectroscopic database [10], we found that the target absorption spectral region (between 2186.4 and 2187.2 cm−1) for simultaneous measurement of CO, H2O and N2O, can be obtained at QCL operating temperature of 20 °C and with a bias current of 400 mA. A low-frequency triangular current ramp at 100 Hz is used to scan the QCL over the selected absorption lines by modulation of the driving current. The spectral signals are generally co-added from an average of 100 scans to improve the signal to noise ratio (SNR). To retrieve the spectroscopic parameters, we employed the experimental procedure described previously [21,22]. The data analysis was performed by using a nonlinear least-squares fitting procedure, which uses the Levenberg–Marquardt algorithm to minimize the deviation between the observed and the calculated spectra. Spectral parameters are obtained by simultaneously fitting a low-order polynomial to the spectral baseline and a Voigt profile to the observed absorption lines, taking into account the measured path length, gas temperature and pressure inside the sample cell. For wavelength calibration, the relationship between QCL emitting wavelength and injecting current at each temperature was initially determined by using a 3rd order polynomial fit to the data, then the fitted equation was used for further wavelength calibration in this work. The commercial nitrous oxide and carbon monoxide gas samples, furnished by Nanjing Special Gas Inc., with a stated purity of 99.99% in natural abundances, were used without further purification. The pressures are measured with an uncertainty of 0.5% using a baratron manometer. Regarding water vapor, a glass bottle filled with liquid water in natural abundance is used to generate saturated water vapor. All measurements were done at room temperature (approximately 299 ± 2 K).

5

CO

N2O

air broadening

Cell 4.33 cm for

H2O broadening

Cell 29.6 cm for

Laser Ramp

Fig.2. Schematic of the experimental setup used for broadening measurements.

3. Results and discussion 3.1. H2O lines broadening measurements To check reliability of the experimental procedure, test measurement of the self-broadening coefficients for H2O lines near 2186.92 cm-1 was firstly performed. The water vapor self broadening coefficients is measured using a single-pass gas cell with path length of 29.6 cm. The sample gases for the present study consisted of evaporation from liquid distilled H2O. Fig. 3 (a) shows an example of the measured H2O lines profile between 2186.75 and 2187.05 cm-1 under different pressure, and fit with the Voigt profile (Upper panel), as well as the corresponding residuals (Lower panel). Indeed, this segment of spectra presented here includes three transitions of the v2 band of H2O, which are summarized in Table 1. For the unresolved absorption

lines, we use a multi-fitting program to fit these three transitions simultaneously to obtain broadening coefficients of all the transitions [23]. A typical plot of the collisional Lorentzian half-width at half-maximum (HWHM) derived from the Voigt 6

profile versus the pressure of water vapor is illustrated in Fig. 3 (b) for H2O transitions at 2186.920 and 2186.925 cm-1. The slopes of the associated straight lines issued from linear regressions give us the corresponding self-broadening coefficients. Finally, the self-broadening coefficients of two stronger lines of H2O achieved in present work are summarized in Table 1 and compared with the latest HITRAN12 database. From Table 1, we can see that the differences between our measurements and the data from the HITRAN12 database are 1.6% and -3.9% for H2O transitions at 2186.92 and 2186.925 cm-1, respectively, but they are still in good agreement when considering the experimental uncertainty. The difference perhaps attributable to the fitted error due to the slight interval between these two transition lines (i.e. 0.005 cm-1) and the possible H2O concentrations error due to the strong adsorption effect. Indeed, the spectral parameters of absorption line intensities are also retrieved, but they are not reported here due to no significant difference with the values listed in HITRAN12 database. On the other hand, the objective of this work is focused on the influence of H2O broadening effect on accurately atmospheric CO and N2O measurements.

1.000 0.995

(a) P= 17.5 mbar P = 20.5 mbar P = 25.5 mbar P = 30 mbar

Transmission

0.990 0.985 0.980 0.975 0.970 0.965 0.960 2186.75

2186.80

2186.85

2186.90

2186.95

2187.00

2187.05

2186.95

2187.00

2187.05

-1

Wavenumber (cm )

Residual

0.003

0.000

-0.003 2186.75

2186.80

2186.85

2186.90

Wavenumber (cm-1)

7

0.006 -1

Lorentzian HWHM (cm )

(b)

0.004

-1

0.002

H2O@ 2186.925 cm Linear fit

0.000 0.000

0.005

0.010

0.006

0.020

0.025

0.030

-1

Lorentzian HWHM (cm )

0.015 H2O pressure (atm)

0.004

0.002

-1

H2O@ 2186.920 cm Linear fit

0.000 0.000

0.005

0.010

0.015

0.020

0.025

0.030

H2O pressure (atm)

Fig.3. (a) Experimental absorption spectra of H2O near 2186.92 cm-1 under different pressures, (b) Plot of collisional HWHM derived from the Voigt profile vs. pressure for H2O transitions at 2186.920 and 2186.925 cm-1, respectively. The straight line is the linear fit of the data points. 3.2. N2O line broadening measurements The same procedure is followed to determine the H2O-induced broadening coefficients for the N2O absorption transition (P38e) at 2187.099 cm-1. To extend our comparison, the air-induced broadening coefficients are also observed. Here, two single-pass gas cells with path length of 29.6 cm and 4.33 cm were used for H2O- and air-induced broadening measurements, respectively. Fig. 4 (a) presents the measured absorption spectra of H2O and N2O lines profile between 2186.88 and 2187.18 cm-1 under different pressures (with fixed N2O pressure of 2.72 mbar). The smooth curve is the best-fitted curve with the standard Voigt profile function. Lower panel shows the corresponding residuals. Note that the W-shaped residuals that occurred at the absorption line center can be effectively removed by using advanced line shape models, such as Rautian profile [24] and Galatry profile [25], which take into account the Dicke narrowing effect [26], lead to a better fit than the one obtained from the Voigt profile [27]. However, these advanced models are not suitable for real 8

applications in view of the computing efficiency, thus not investigated in details in this work. The collisional HWHM derived from the Voigt profile vs. H2O pressures for N2O transition at 2187.099 cm-1 is shown in Fig. 4 (b). The straight line is the best linear fit of the data points. In the case of air-induced broadening, the laboratory air was directly introduced in the absorption cell without predrying treatment. However, the H2O mixing ratio in air is calibrated with hygrometer (HygroClip2) during the whole experiments. The air-broadened absorption spectra and plot of HWHM are presented in Fig. 5 (a) and (b), respectively. Finally, the fitted H2O- and air-induced broadening coefficients and their uncertainties are also complied in Table 1, and compared with the data from the HITRAN12 database. The measured air-broadened coefficient is slightly lower than the result of HITRAN12, with a discrepancy of -3.15%. Indeed, this difference is acceptable when the experimental uncertainty is taken into account. The measured air-broadened coefficient for N2O transition of P38e line is in reasonable agreement with a similar work recently performed by Es-sebbar et al. [28]. However, the value of the ratio of H2O-broadening to that of air, for the considered N2O line (P38e) is 1.9. To the best of our knowledge, there is no data available about the H2O reduced broadening parameters for N2O transitions. In case of wavelength modulation spectroscopy second harmonic (WMS-2f) detection, omitting this difference between air- and H2O-broadenings will thus lead to an under-estimation of about 1.5% of signal amplitude under the conditions of typically 2% of H2O and 320 ppb of N2O. This difference thus cannot be neglected if one wants to measure N2O concentration in a humid atmosphere with a sub-percent precision (details see next section).

9

1.00

Transmission

0.95 0.90

(a)

P = 1.5 mbar P =2.25 mbar P = 8.5 mbar P = 13.6 mbar P = 20.5 mbar P = 25.6 mbar P = 28 mbar Voigt Fit

H2O

0.85 0.80 0.75 0.70 2186.88

N2O

2186.92

2186.96

2187.00

2187.04

2187.08

2187.12

2187.16

2187.08

2187.12

2187.16

-1

Wavenumber (cm )

0.010

Residual

0.005 0.000 -0.005 -0.010 2186.88

2186.92

2186.96

2187.00

2187.04 -1

Wavenumber (cm )

Fig.4. (a) Experimental absorption spectra of H2O and N2O lines profile between 2186.88 and 2187.18 cm-1 under different H2O pressures, (b) Plot of collisional HWHM derived from the Voigt profile vs. H2O pressures for N2O transition at 2187.099 cm-1. The straight line is the linear fit of the data points.

10

1.00

Transmission

0.75

P=104.5 mbar P=190.1 mbar P=314.6 mbar P=363 mbar P=460.8 mbar P=537.4 mbar P=654.3 mbar P=753.1 mbar P=969.6 mbar Voigt Fit

(a)

0.50

0.25

0.00 2186.8

2186.9

2187.0

2186.9

2187.0

2187.1 -1 Wavenumber (cm )

2187.2

2187.3

2187.4

2187.2

2187.3

2187.4

0.02

Residual

0.01 0.00 -0.01 -0.02 2186.8

2187.1

-1 Wavenumber (cm )

Fig.5. (a) Experimental absorption spectra of air and N2O lines profile between 2186.8 and 2187.4 cm-1 under different air pressures, (b) Plot of collisional HWHM derived from the Voigt profile vs. air pressures for N2O transition at 2187.099 cm-1. The straight line is the linear fit of the data points. 3.3. CO line broadening measurements Similar to N2O experiments mentioned above, two absorption cells with different of 29.6 cm and 4.33 cm were also used for H2O- and air-induced CO transition R(11) at 2186.639 cm-1 broadening measurements, respectively. Fig. 6 (a) and (b) show the 11

measured CO and H2O absorption spectra between 2186.5 and 2187.0 cm-1 under different H2O pressure, and the fitted collisional HWHM as a function of H2O pressures, respectively. The CO partial pressure was firstly set below 1.12 mbar, then water vapor (or laboratory air in case of air-broadening) was gradually added into the absorption cell, for example, the total pressure (including water vapor) is given in the inset. Fig. 7 (a) and (b) illustrates the CO absorption spectra between 2186.5 and 2186.9 cm-1 perturbed by laboratory air with pressure from 100 mbar to 1010 mbar, and the pressure dependence HWHM as a function of air pressures. As can be seen from Table 1, the measured air-broadened coefficient is closing the value from HITRAN12 database, with a small discrepancy of 1.06%. While the measured H2O-broadened coefficient is also significantly different from those by air, larger than approximately 1.8 times. The influence of omitting the difference between air- and H2O-broadenings on atmospheric CO measurements will be also discussed in the following sections. 1.0 0.9

H2O

(a)

0.8

Transmission

0.7

P= 3.2 mbar P= 4.7 mbar P=7.3mbar P= 9 mbar P= 12.1 mbar P= 19.5 mbar P= 26.1 mbar Voigt Fit

0.6 0.5 0.4 0.3 0.2 0.1 2186.5

CO 2186.6

2186.7

2186.8

2186.9

2187.0

-1

Wavenumber (cm ) 0.0150

Residual

0.0075 0.0000 -0.0075 -0.0150 2186.5

2186.6

2186.7

2186.8 -1

Wavenumber (cm )

12

2186.9

2187.0

Fig.6. (a) Experimental absorption spectra of CO and H2O lines profile between 2186.5 and 2187.0 cm-1 under different H2O pressures, (b) Plot of collisional HWHM derived from the Voigt profile vs. H2O pressures for CO transition at 2186.639 cm-1. The straight line is the linear fit of the data points.

1.0

Transmission

0.9

P=100 mbar P=193.7 mbar P= 307 mbar P=412.7 mbar P= 506.8 mbar P=620.2 mbar P=709.4 mbar P=806.4 mbar P= 906.5 mbar P=1010.3 mbar Voigt Fit

(a)

0.8

0.7

0.6

0.5 2186.5

2186.6

2186.7

2186.8

2186.9

-1

Wavenmber (cm )

Residual

0.015

0.000

-0.015

2186.4

2186.5

2186.6

2186.7 -1

Wavenumber (cm )

13

2186.8

2186.9

Fig.7. (a) Experimental absorption spectra of CO line profile between 2186.5 and 2186.9 cm-1 under different air pressures, (b) Plot of collisional HWHM derived from the Voigt profile vs. air pressures for CO transition at 2186.639 cm-1. The straight line is the linear fit of the data points. Table 1. Summary of CO, H2O and N2O transitions and measured broadening coefficients for this study.

Species

Transition

Line strength

Air-broadened

Self-broadened

H2O-broadened half width

frequency

(cm/ mol.)

half width (cm-1/atm)

half width (cm-1/atm)

(cm-1/atm)

(cm-1)

CO

H2O

N2O

HITRAN

This work

Disc. (%)

HITRAN

This work

This work

Disc. (%)

2186.639

3.314
10-19

0.0567

0.0573(7)

1.06

0.062

--

0.1007(20)

--

2186.898

2.086
10-24

0 .0379

--

--

0.291

--

--

--

2186.920

2.820
10-23

0.0315

--

--

0.188

0.1909(28)

--

1.6

2186.925

9.400
10-24

0.0306

--

--

0.204

0.1962(62)

--

-3.9

2187.099

1.863
10-19

0.0698

0.0676(18)

-3.15

0.084

--

0.1327(36)

--

Note: a The uncertainties of the measured broadening coefficients correspond to one standard deviation (1σ) obtained by averaging the different measurements. Values in parentheses represent 1σ uncertainties in the last digit (e.g. 0.0573(7) cm−1/atm is equivalent to 0.0573 cm−1/atm ± 0.0007 cm−1/atm). b

c

Disc. = (This work – HITRAN)/ HITRAN
100.

Line position and line strength (at 296K) are taken from HITRAN database.

14

3.4 Influence of humidity on atmospheric CO and N2O detection in WMS Among all the spectroscopic methods, direction absorption spectroscopy (DAS) and WMS are widely used spectroscopic techniques for gas concentration measurements. DAS based upon Lambert-Beer law is a simple, absolute, calibration-free technique for determining information about gas phase species. However, DAS suffers from a low sensitivity that limits its extension into trace gases analysis, thus needs to combine multipass absorption cell techniques. WMS applies an additional sinusoidal modulation in the tens of kHz-range on top of the laser ramp, thereby shifting the measured signals to higher frequencies to reduce 1/f noise via phase sensitive detection (i.e. lock-in amplifier), thus providing signal with higher SNR. In real application, gas sensor-based on WMS technique needs to be calibrated using additional cylinder sample gas with known sample concentration to achieve absolute concentration measurement [8, 29]. The on-peak WMS-2f signal is usually used for concentration quantification, since the WMS-2f signal scales linearly with the target concentration (under the condition of weak absorption). As discussed above, the difference between air- and H2O-broadenings is great significant, therefore, a relative inaccuracy of a similar amount as the line width change will be introduced on the retrieved concentration, due to gas composition change between calibration cylinder and practical atmosphere. For practical application [30], the effect of humidity on the R(11) transition of CO and the P(38e) transition of N2O absorption features are theoretically simulated by using the observed H2O broadened coefficients in this work. Sample pressure of 100 mbar and temperature of 296 K are uniformly used to investigate the influence of different H2O mixing ratio on the WMS-2f signal amplitude, the modulation indexes for both species are set for the optimal value of 2.2, and assuming the ratio of H2O-broadening to that of air is 1.8 for both CO and N2O lines. For instance, Fig. 8 illustrates the simulated WMS-2f spectra of CO (500 ppb), H2O (1.5%) and N2O (330 ppb) between 2186.5 and 2187.3 cm-1, with and without considering H2O broadening 15

effect. For clarity, the insets of Fig. 8 are magnified portions of the center profile showing the differences between the two cases. As can be seen, omitting the difference between air- and H2O-broadenings, an error of over 1% will be introduced for both CO and N2O measurements. The simulated error for both species under different H2O, CO and N2O concentration is shown in Fig. 9. From this figure, we can see that the induced error is significantly increasing with the increasing of H2O mixing ratio, typically with 5% H2O concentration, the errors are 3.7% and 3.4% for N2O and CO, respectively, while no distinct dependence on CO and N2O concentration. Moreover, we have further studied the induced error for the case of different humidity conditions when the scale of the species concentration is extended to 1 ppm. We find that even under high humidity, the error introduced by water vapor has not obviously depended on both species concentration, respectively. All these results prove that the difference between air- and H2O-broadenings thus cannot be neglected if one wants to measure gas concentrations in a high humid environment with a sub-percent precision.

Fig.8. Simulated WMS-2f spectra of CO, H2O and N2O between 2186.5 and 2187.3 cm-1, with and without considering H2O broadening effect (details see txt).

16

Fig.9. The introduced errors between with and without considering H2O broadening effect for N2O P(38e) and CO R(11) transitions as a function of CO, H2O and N2O.

4. Conclusions In this paper, the influence of H2O broadening effect on CO R(11) and N2O P(38e) transitions at 2186.639 cm-1 and 2187.099 cm-1, respectively, was detailed investigated by using a QCL based absorption spectrometer. The results indicate that H2O broadening coefficients for CO R(11) and N2O P(38e) lines are 1.8 and 1.9 times higher than the corresponding air-broadening parameters, respectively. As far as the WMS-2f technique is concerned, our simulation confirmed that the shapes of CO and N2O absorption lines will be significantly affected by variations of the water vapor mixing ratio, while no significant dependence on target concentration. Indeed, water broadening effect not only affects WMS measurements as demonstrated in this work, but also has the impact on other spectroscopic techniques [31-33]. All these results prove that the difference between air- and H2O-broadenings thus cannot be neglected if one wants to measure gas concentrations in a high humid environment with a sub-percent precision. The measured parameters will be usefully for upgrading our newly developed multi-species QCL sensor system for atmospheric applications and also providing a certain reference to other similar studies [8, 9]. In addition, the 17

rotational quantum numbers J-dependence of water broadening coefficients on CO and N2O transitions near 4.57 µm region will be detailedly investigated in our future work.

Acknowledgments This work was supported in part by Anhui University Personnel Recruiting Project of Academic and Technical Leaders under Grant 10117700014, the Natural Science Fund of Anhui Province under Grant 1508085MF118, the National Natural Science Foundation of China under Grant 61440010 and 41306103, and the key Science and Technology Development Program of Anhui Province under Grant 1501041136.

Reference [1] S.A. Montzka, E.J. Dlugokencky, J.H. Butler, Nature 476 (2011) 43-50. [2] D.D. Nelson, B. McManus, S. Urbanski, S. Herndon, M.S. Zahniser, Acta A 60 (2004) 3325-3335. [3] J.B. McManus, M.S. Zahniser, D.D. Nelson, J.H. Shorter, S.C. Herndon, D. Jervis, M. Agnese, R. McGovern, T.I. Yacovitch, J.R. Roscioli, Appl. Phys. B 119 (2015) 203-218. [4] J.P. Lima, H. Vargas, A. Miklós, M. Angelmahr, P. Hess, Appl. Phys. B 85 (2006) 279-284. [5] L. Dong, R. Lewicki, K. Liu, P.R. Buerki, M.J. Weida, F.K, Tittel, Appl. Phys. B 107 (2012) 275-283. [6] D.J. Hamilton, A.J. Orr-Ewing, Appl. Phys. B 102 (2011) 879-890. [7] R. Provencal, M. Gupta, T.G. Owano, D.S. Baer, K.N. Ricci, A. O’Keefe, J.R. Podolske, Appl. Opt. 44 (2005) 6712 -6717. [8] J.S. Li, U. Parchatka, H. Fischer, Sensors and Actuators B 182 (2013) 659–67. [9] L. Tao, K. Sun, M.A. Khan, D.J. Miller, M.A. Zondlo, Opt. Express 20 (2012) 28106-28118. [10] L.S. Rothman, I.E. Gordon, Y. Babikov, A. Barbe, D.C. Benner, P.F. Bernath, M. Birk, L. Bizzocchi, V. Boudon, L.R. Brown, A. Campargue, K. Chance, L.H. 18

Coudert, V.M. Devi, B.J. Drouin, A. Fayt, J.M. Flaud, R.R. Gamache, J. Harrison, J.M. Hartmann, C. Hill, J.T. Hodges, D. Jacquemart, A. Jolly, J. Lamouroux, R.J. LeRoy, G. Li, D. Long, C.J. Mackie, S.T. Massie, S. Mikhailenko, H.S.P. Müller, O.V. Naumenko, A.V. Nikitin, J. Orphal, V.I. Perevalov, A. Perrin, E.R. Polovtseva, C. Richard, M.A.H. Smith, E. Starikova, K. Sung, S.A. Tashkun, J. Tennyson, G.C. Toon, V.G. Tyuterev, G. Wagner, J. Quant. Spectrosc. Radiat. Transfer 130 (2013) 4-50. [11] GESIA spectroscopic database. http://ether.ipsl.jussieu.fr/ [12] C.J. Wallace, C. Jeon, C.N. Anderson, D.K. Havey, J. Phys. Chem. A 115 (2011) 13804-13810. [13] K. Sung, L.R. Brown, R.A. Toth, T.J. Crawford, Can. J. Phys. 87(2009) 469-484. [14] S. Schilt, Appl. Phys. B 100 (2010) 349-59. [15] K. Owen, E. Es-sebbar, A. Farooq, J. Quant. Spectrosc. Radiat. Transfer 121 (2013) 56-68. [16] E.M. Vess, C.J. Wallace, H.M. Campbell, V.E. Awadalla, J.T. Hodges, D.A. Long, D.K. Havey, J. Phys. Chem. A 116 (2012) 4069-4073. [17] B.J. Drouin, V. Payne, F. Oyafuso, K. Sung, E. Mlawer, J. Quant. Spectrosc. Radiat. Transfer 133 (2014) 190–198. [18] M.A. Koshelev, I.N. Vilkov, M.Y. Tretyakov, J. Quant. Spectrosc. Radiat. Transfer 154 (2015) 24-27. [19] T. Delahaye, X. Landsheere, E. Pangui, F. Huet, J.M. Hartmann, H. Tran, J. Quant. Spectrosc. Radiat. Transfer 173 (2016) 40-48. [20] J.S. Li, U. Parchatka, R. Königstedt, H. Fischer, Opt. Express 20 (2012) 7590-7601. [21] J.S. Li, K. Liu, W.J. Zhang, W.D. Chen, X.M. Gao, J. Quant. Spectrosc. Radiat. Transfer 109 (2008) 1575-1585. [22] J. Skrotzki, J. C. Habig, V. Ebert, Appl. Phys. B 116 (2014) 393-406. [23] J.S. Li, L. Joly, J.L. Cousin, B. Parvitte, B. Bonno, V. Zeninari, G. Durry, Spectrochimica Acta Part A 74 (2009) 1204-1208. [24] S.G. Rautian, I.I. Sobel’man, Soviet Physics Uspekhi 9 (1967) 701-716. 19

[25] L. Galatry, Phys. Rev. 122 (1961) 1218-1223. [26] A.S. Pine, J. Quant. Spectrosc. Radiat. Transfer 62 (1999) 397-423. [27] J.S. Li, G. Durry, J. Cousin, L. Joly, B. Parvitte, V. Zzeninari, J. Quant. Spectrosc. Radiat. Transfer 111 (2010) 2332-2340. [28] E. Es-sebbar, M. Deli, A. Farooq, Appl. Spectrosc. 70(2016) 972-982. [29] J. Li, A. Reiffs, U. Parchatka, H. Fischer, Metrol. Meas. Syst. 22 (2015) 25-38. [30] J.S. Li, H. Deng, J. Sun, B.L. Yu, H. Fischer, Sensors and Actuators B 231 (2016) 723-732. [31] C.L. Butenhoff, M.A.K. Khalil, Chemosphere 47 (2002) 823-836. [32] D.L. Arévalo-Martínez, M. Beyer, M. Krumbholz, I. Piller1, A. Kock, T. Steinhoff, A. Körtzinger, H.W. Bange, Ocean Sci. 9 (2013) 1071-1087. [33] H. Chen, A. Karion, C.W. Rella, J. Winderlich, C. Gerbig, A. Filges, T. Newberger, C. Sweeney, P.P. Tans, Atmos. Meas. Tech. 6 (2013) 1031-1040.

20

Graphical abstract

21

Highlights • Mid-IR spectral region is of particular interest for atmospheric measurements. •

QCLs

are

promising

laser

sources

for

spectroscopic

applications. • H2O-broadening of two mid-infrared CO and N2O lines near 4.57 µm were measured. • H2O broadening coefficients are found to be 1.8 and 1.9 times higher than the corresponding air-broadening parameters.

22