Chinese Journal of Chemical Physics  2020, Vol. 33 Issue (1): 1-7

The article information

Hao Wu, Jian Chen, An-wen Liu, Shui-ming Hu, Jing-song Zhang

Cavity Ring-Down Spectroscopy Measurements of Ambient NO$_\bf{3}$ and N$_\bf{2}$O$_\bf{5}$

Chinese Journal of Chemical Physics, 2020, 33(1): 1-7

http://dx.doi.org/10.1063/1674-0068/cjcp1910173

Article history

Accepted on: December 2, 2019
Cavity Ring-Down Spectroscopy Measurements of Ambient NO$_\bf{3}$ and N$_\bf{2}$O$_\bf{5}$
Hao Wua , Jian Chenb , An-wen Liub , Shui-ming Hua,b , Jing-song Zhangc
Dated: Received on October 6, 2019; Accepted on December 2, 2019
a. Hefei National Laboratory for Physical Sciences at the Microscale, University of Science and Technology of China, Hefei 230026, China;
b. Department of Chemical Physics, School of Chemistry and Materials Science, University of Science and Technology of China, Hefei 230026, China;
c. Department of Chemistry and Air Pollution Research Center, University of California, Riverside California 92521, USA
Abstract: NO$_3$ and N$_2$O$_5$ are important participants in nocturnal atmospheric chemical processes, and their concentrations are of great significance in the study of the mechanism of nocturnal atmospheric chemical reactions. A two-channel diode laser based cavity ring-down spectroscopy (CRDS) instrument was developed to monitor the concentrations of NO$_3$ and N$_2$O$_5$ in the atmosphere. The effective absorption length ratio and the total loss coefficient of the instrument were calibrated using laboratory standard samples. The effective absorption cross section of NO$_3$ at 662 nm was derived. A detection sensitivity of 1.1 pptv NO$_3$ in air was obtained at a time resolution of 1 s. N$_2$O$_5$ was converted to NO$_3$ and detected online in the second CRDS channel. The instrument was used to monitor the concentrations of NO$_3$ and N$_2$O$_5$ in the atmosphere of winter in Hefei in real time. By comparing the concentration changes of pollutants such as nitrogen oxides, ozone, PM$_{2.5}$ in a rapid air cleaning process, the factors affecting the concentrations of NO$_3$ and N$_2$O$_5$ in the atmosphere were discussed.
Key words: Cavity ring-down spectroscopy    Nitrate radical    Dinitrogen pentoxide    Field measurement
Ⅰ. INTRODUCTION

Nitrate radicals (NO$_3$) and dinitrogen pentoxide (N$_2$O$_5$) are important components of atmospheric nitrogen oxides. They are produced by oxidation of low valence state nitrogen oxides in the atmosphere (Eq.(1)) and they are important participants in the atmospheric chemical reactions.

 $\begin{eqnarray}\begin{array}{l} {\rm{NO}} + {{\rm{O}}_3} \to {\rm{N}}{{\rm{O}}_2} + {{\rm{O}}_2}\\ {\rm{N}}{{\rm{O}}_{\rm{2}}} + {{\rm{O}}_3} \to {\rm{N}}{{\rm{O}}_3} + {{\rm{O}}_2}\\ {\rm{N}}{{\rm{O}}_{\rm{2}}} + {\rm{N}}{{\rm{O}}_3} \leftrightarrow {{\rm{N}}_{\rm{2}}}{{\rm{O}}_5} + {{\rm{O}}_2} \end{array} \end{eqnarray}$ (1)

The nitrate radical has a high reactivity, and furthermore it can be readily photolyzed under solar radiation, leading to a short life time [1]. As a result, the concentration of NO$_3$ during the daytime is very low, and it only exists at night [2]. As one of the important oxidants in the atmospheric oxidation reactions at night, NO$_3$ is involved in the oxidative degradation of various atmospheric pollutants (such as VOCs [3], sulfide [4], etc.).

In the atmosphere, dinitrogen pentoxide (N$_2$O$_5$) is mainly produced by the reaction between the nitrate radical and nitrogen dioxide in a chemical equilibrium among them (Eq.(1)), which makes N$_2$O$_5$ as a reservoir of the NO$_3$ radical in the atmosphere. Besides the transformation to the nitrate radical, hydrolysis reaction of N$_2$O$_5$ involving aerosol particles, clouds and fog is another significant loss pathway [5, 6, 7]. Nitric acid and nitrate produced by this reaction are one of the main sources of nitrate in the atmosphere [2, 8, 9].

The concentrations of the nitrate radical and dinitrogen pentoxide in the atmosphere are both at the level of parts-per-trillion by volume (pptv) [5, 10]. Therefore, the measurement technique should have a high sensitivity and selectivity. N$_2$O$_5$ can be detected indirectly through the detection of NO$_3$ following the thermal decomposition of N$_2$O$_5$. After Noxon et al. [11] first detected NO$_3$ in the tropospheric atmosphere by differential optical absorption spectroscopy (DOAS) in 1978, a variety of techniques have been developed to detect NO$_3$ in the subsequent 40 years, among which spectroscopy and mass spectrometry methods were frequently applied. The most commonly used mass spectrometry method is chemical ionization mass spectrometry, and its detection limit can reach 11 pptv [12]. Matrix-isolation electron spin resonance spectroscopy (MI-ESR) has very high selectivity in detecting NO$_3$, which can be used for highly sensitive off-line measurement of the NO$_3$ radical [13, 14]. However, MI-ESR needs a complicated and time-consuming sample collection and detection procedure, which limits its wide application.

NO$_3$ can be detected through its strong absorption around 623 nm and 662 nm. The laser induced fluorescence (LIF) technique was applied for a quantitative measurement of NO$_3$ by measuring the fluorescence in the 700$-$750 nm region emitted by NO$_3$ after absorption of the 662 nm photon, and the reported detection limit reached 6 pptv (1$\sigma$, 10 min) [15]. The DOAS technique uses sunlight and moonlight for passive measurement, enabling a global measurement [16] and a determination of NO$_3$ concentration in the stratospheric atmosphere [17]. Alternatively, high sensitivity can be achieved through active measurement by using laboratory light source, which resulted in a detection limit of 2 pptv [18]. Using laser light travelling back and forth multiple times inside an optical resonator composed of two highly reflective mirrors, cavity enhanced absorption spectroscopy (CEAS) and cavity ring-down spectroscopy (CRDS) techniques obtain a very long effective absorption path length and enhance the detection sensitivity. The detection limit of CEAS can reach 2 pptv [19], and the sensitivity of CRDS can reach 1 pptv [20]. In China, a few groups from Peking University [21], Fudan University [22], Anhui Institute of Optics and Fine Mechanics [23] and Hong Kong Polytechnic University [24] reported the detection of NO$_3$ and N$_2$O$_5$ using different methods such as DOAS, CEAS, and CRDS. In this study, we present a dual-channel CRDS instrument based on diode lasers for the detection of ambient NO$_3$ and N$_2$O$_5$. As a demonstration, the instrument was used to monitor concentrations of the nitrogen oxides during a rapid air cleaning process.

Ⅱ. EXPERIMENTS A. Experimental techniques

A two-channel cavity ring-down spectrometer was used for real-time online analysis of the ambient samples. The basic principle is to place the sample gas in an optical resonant cavity consisting of a pair of high-reflection (HR) mirrors. A continuous-wave laser was coupled into the optical cavity and traveled many times inside the cavity until a stable light field was formed. When the incident laser was turned off, the light field in the cavity gradually decreased due to the transmission and loss of the mirrors, as well as the absorption and scattering by the gas sample, which resulted in an exponential decay of the transmitted light intensity. By fitting the decay curve, one can derive the relationship between the absorption coefficient $\alpha$ and the ring-down time $\tau$:

 $\begin{eqnarray} \alpha \left( \nu \right) = \frac{1}{{c\tau \left( \nu \right)}} - \frac{1}{{c{\tau _0}(\nu)}} \end{eqnarray}$ (2)

where $c$ is the speed of light, and $\tau$($\nu$) and $\tau_0$($\nu$) are ring-down times with and without the sample, respectively. The configuration of our experimental setup is shown in FIG. 1.

 FIG. 1 Experimental setup

Our optical cavity was constructed with a pair of HR mirrors ($R$$\approx$99.996%) positioned about 87 cm apart. The sample cell was made by 3/4 inch outer diameter (O.D.) and 5/8 inch inner diameter (I.D.) PFA tube. The inlet and outlet of the gas sample were between the HR mirrors at a distance of 45 cm (see in FIG. 1). A 662 nm diode laser (IQu Series, PTI) with a maximum output power of 120 mW was used as the light source for both optical cavities (with a 50/50 split). Square wave signal generated by a function generator switched the laser at a frequency of 1 kHz. An optical isolator was used to prevent the reflected light from windows and HR mirrors from affecting the laser output. A photomultiplier tube (PMT, H10721-20, Hamamatsu) was used for signal detection. To avoid the influence of stray light, a narrow-band high-efficiency optical band-pass filter was placed in front of the PMT. The signal collected by the PMT was amplified by a signal amplifier and digitized via an analog-digital conversion card (PCI 9820, ADLink) and processed by a personal computer. Ring-down events were recorded and averaged for typically one second to improve the signal-to-noise ratio.

For each CRDS channel, a 5 µm-pore polytetrafluoroethylene membrane (TE 38, Whatman, GE) was used to filter out particles in the air sample. Air sample was pumped into the CRDS cavity by a diaphragm pump through a 1/4 inch O.D. sampling tube (PFA) after the filter. The injection rate was controlled by mass flow controller (MFC). To avoid the loss of the measured object in the MFC and diaphragm pump, both of them were placed at the outlet of the cavity air path, and the flow rate of MFC was controlled at 4.15 slpm (standard liter per minute). At the same time, in order to protect the surface of the high reflection mirror from being contaminated by the atmospheric samples, 75 sccm (standard cubic centimeter per minute) high-purity nitrogen gas was injected at the surface of each HR mirror as the protective gas. As shown in FIG. 1, one of the CRDS channels was used for direct measurement of the atmospheric concentration of the NO$_3$ radicals, and the second channel was fed with sample gas through a membrane along with the teflon tube heated to 120 ℃, which converted the ambient N$_2$O$_5$ completely (99.6%) to NO$_3$ with a ratio of 1:1. In the second channel, the CRDS cavity was kept at temperature of 80 ℃ and also measured the NO$_3$ concentration. The N$_2$O$_5$ concentration in the air sample was derived from the difference between the NO$_3$ concentrations measured by the two CRDS channels.

B. Calibration

The high reactivity of the NO$_3$ radical makes its loss during the sampling and measurement processes inevitable. In order to obtain the concentration of NO$_3$ in the air sample more accurately, a series of measurements were carried out to calibrate the related coefficients.

First, by comparing the ring-down time $\tau$ with the sample and $\tau_0$ without the sample (zero gas), the sample absorption coefficient $\alpha$ was obtained (Eq.(2)). Then the volume concentration of NO$_3$ was determined according to Eq.(3),

 $\begin{eqnarray} \chi = \frac{{\alpha \left( \nu \right)}}{{\sigma \left( \nu \right)}} \times \frac{{{R_{\rm{L}}}}}{{1 - \eta }} \times \frac{{RT}}{{P{N_{\rm{A}}}}} \end{eqnarray}$ (3)

where $\sigma$($\nu$) is the absorption cross section of NO$_3$, $R_{\rm{L}}$ is the ratio of the optical cavity length to the absorption path length of the sample, $\eta$ is total loss coefficient of the NO$_3$ sample in the apparatus, $N_\textrm{A}$ is Avogadro constant, $R$ is the gas constant, $T$ is the sample temperature in K, and $P$ is the gas pressure.

The ratio $R_{\rm{L}}$ was introduced because the N$_2$ purge gas was used in the measurement to protect the HR mirrors at both ends of the cavity, so that the sample gas absorption path length was less than the total cavity length between the two HR mirrors. This coefficient was related to the geometry of the cavity and the flow rates of the sample gas and protection N$_2$ gas. This coefficient was calibrated using a system for the NO$_2$ detection which had the same configuration as this one. The $R_{\rm{L}}$ coefficient in the NO$_2$ system was calibrated by comparing the NO$_2$ concentration measured by CRDS and that by a commercial NO$_x$ analyzer (42i-TL, Thermo Fisher Scientific), which gave $R_{\rm{L}}$=1.54$\pm$0.02 [25]. Since the same experimental conditions were applied in the NO$_3$ measurement, this $R_{\rm{L}}$ coefficient was used for the NO$_3$ apparatus.

The loss of NO$_3$ was mainly due to sampling tubing, cavity wall, and surface of the filter membrane. The porous structure of the membrane not only filtered aerosol, but also led to a loss of NO$_3$ in the sample gas. It was therefore necessary to measure the loss of NO$_3$ before its detection by CRDS. A standard NO$_3$ sample was used to calibrate the loss ratio. By slowly passing the high-purity nitrogen gas through a solid state N$_2$O$_5$ sample stored in a dry ice-alcohol mixture bath with a flow rate of 25 sccm, a sample of N$_2$O$_5$ with a stable concentration was obtained. After high-temperature pyrolysis ($\sim$120 ℃), the N$_2$O$_5$ sample was converted to a stable flowing sample of NO$_3$.

Two gas inlet channels were used to determine the loss ratio due to the membrane filter. These two channels were identical except that the membrane filter was installed in one of them. The sample gas was flown into the CRDS measurement cavity and switched between the two gas inlet channels. The NO$_3$ concentration detected by CRDS is shown in FIG. 2. Besides a steady drift of the NO$_3$ concentration due to the change of the source sample, a difference between the two curves measured through these two inlet channels with and without the membrane filter was clearly identified. The difference gave a loss ratio of (13$\pm$2)% due to the membrane filter.

 FIG. 2 Loss rate of the NO$_3$ radical by measuring the relative concentration of the NO$_3$ standard sample with and without the 5 µm filter in the inlet gas path. Data shown in black and red are measured with and without the filter, respectively.

The transmission loss was determined by two methods, similar to those by Dube et al. [26] and Fuchs et al. [27]. The first was to measure the linear loss rate of the NO$_3$ radical in our sample tube. By pumping the sample into a 40 cm, 1/4 inch teflon tube at different rates, 1.5 slpm, 2.5 slpm, 3.5 slpm and 4.5 slpm, through a sample tubing volume of about 50 mL from the inlet tube to the center of the CRDS sample cell, residence time of the prepared NO$_3$ sample inside our device was obtained as 2.0 s, 1.2 s, 0.86 s, and 0.67 s respectively. With a linear fitting of the measured concentration of NO$_3$ to the residence time, the transmission loss rate can be determined as (11$\pm$2)%. In our monitoring process, considering the length of the inlet tubing, the transmission loss was determined as (14$\pm$2)%.

The second was to measure the loss directly in our device. The prepared NO$_3$ sample was converted to NO$_2$ by mixing with the excessive NO gas, and the concentration of the NO$_2$ product was measured by the CRDS instrument for the NO$_2$ detection [25], which quantified the concentration of the NO$_3$ sample. By comparing the NO$_3$ concentrations measured under the experimental conditions that sample tubing of different lengths were used, the loss ratio of NO$_3$ during the transportation in sample tubing and the cavity wall was also determined, which was (12$\pm$2)%. This process utilized an extra device which could introduce additional uncertainty; as a result, we decided to use the result of the first process for our calibration procedure.

Finally, the total loss ratio of NO$_3$ was obtained from a combination of the filter loss and transmission loss, and in our experiment apparatus it was determined to be $\eta$=(27$\pm$3)%.

Ⅲ. RESULTS AND DISCUSSION A. Measurement performance

In order to improve the measurement sensitivity and signal-to-noise ratio and avoid the interference due to other molecules in the air sample, the laser center frequency used in the measurement was selected to be near the peak of the absorption of NO$_3$ at about 662 nm. The black line in FIG. 3 is the emission spectrum curve of the diode laser measured by a grating spectrometer (Shamrock 750) with a resolution of 0.5 nm. The center of the laser emission was determined to be 662.07 nm. By convolving the laser emission spectrum with the NO$_3$ absorption cross section curve at 298 K (red line in FIG. 3) [28], an effective absorption cross section of NO$_3$ at 662.07 nm was determined to be 2.04$\times$10$^{-17}$ cm$^2$molecule$^{-1}$.

 FIG. 3 Absorption cross sections of NO$_3$, NO$_2$, O$_3$, H$_2$O and emission spectrum of the diode laser

Optical extinction due to components other than NO$_3$ in the atmosphere may affect the measurement: such as Rayleigh scattering of nitrogen and oxygen, Mie scattering of aerosol particles, and absorption of water vapor, ozone and nitrogen dioxide. During the measurement, the mass flow controller ensured that the pressure in the cavity was kept within a range of 1% atm, and therefore the Rayleigh scattering effect can be considered as a constant contribution to the baseline which could be directly removed. The membrane filter placed in the sample inlet channel filtered out most aerosol particles in the sample gas, which reduced the influence of Mie scattering of aerosol particles.

The resonant absorption at 662 nm of some trace gas molecules in the atmosphere could also influence the measurements, especially the water vapor [29], ozone [30] and nitrogen dioxide [31]. FIG. 3 shows the absorption cross sections of the three molecules as well as NO$_3$ around 662 nm [32]. Note that the absorption cross sections of H$_2$O, O$_3$ and NO$_2$ are three to nine orders of magnitude smaller than that of NO$_3$. However, since the concentration of NO$_3$ in the atmosphere is at the pptv level, much lower than those of interfering molecules, it was necessary to consider the influence due to each interfering molecule in the atmosphere. Taking the typical concentrations in the atmosphere such as 30 ppbv for nitrogen dioxide, ozone 30 ppbv, water vapor 1%, and combining with the absorption cross sections of these molecules at the wavelength of the laser used in our measurement, the contribution from each molecule can be converted and scaled in terms of the NO$_3$ concentration. They were NO$_2$ 10.4 pptv, O$_3$ 4.0 pptv, and H$_2$O 5.1 pptv. Therefore, the interferences of these three molecules were not negligible.

In order to eliminate their interferences, 6 ppmv nitric oxide gas was repeatedly injected into the sample gas at a 100 sccm flow rate, and the difference of CRDS absorption coefficient ($\Delta\alpha$) was measured. Since NO converted NO$_3$ and O$_3$ to NO$_2$, the absorption due to NO$_3$ decreased 3 orders of magnitude after the NO$_3$ to NO$_2$ conversion and that due to O$_3$ increased 2 times from O$_3$ to NO$_2$ conversion, thus the change $\Delta\alpha$ would be predominantly related to the concentrations of NO$_3$ and O$_3$ in the sample, eliminating the influence due to NO$_2$ and water vapor in the sample. The change $\Delta\alpha$ due to the presence of O$_3$ was relatively small and the influence can be further corrected if the O$_3$ concentration was known.

In atmospheric monitoring, the measurement results are often influenced by environmental conditions such as temperature and air pressure. In order to examine the sensitivity and stability of the device, a continuous measurement of 4 h using zero air under the same experimental conditions was conducted. The results are shown in FIG. 4, where the gray curve was obtained with a time resolution of 1 s, and the red curve was with a time resolution of 3 min. During the 4-h measurement, the peak-to-peak fluctuation in the absorption coefficient was 8$\times$10$^{-10}$ cm$^{-1}$, which corresponded to a NO$_3$ concentration of 1.1 pptv. It could be used as the detection sensitivity (3$\sigma$) of NO$_3$ of this instrument.

 FIG. 4 Absorption coefficient $\alpha$ of zero air measured in a 4-h period
B. Field measurement

As a demonstration, the instrument was applied in a continuous measurement of the outdoor air for a week from November 1, 2017 to November 7, 2017. The sampling site of this measurement was selected in the Science and Technology Building of University of Science and Technology of China in Hefei, Anhui Province. The sampling port was about 8 m away from the ground and 1 m away from the wall. It was about 100 m away from the Huangshan road, one of the main thoroughfares in Hefei city. The road had a large traffic flow during morning and evening rush hours, and the air samples taken in this area could be regarded as a typical urban pollution sample in the winter season.

Together with the NO$_3$ and N$_2$O$_5$ measurements, another CRDS instrument operating at 405 nm [25] was also applied to monitor the concentration of NO$_2$ in the atmospheric sample. The concentrations of PM$_{2.5}$ and O$_3$, temperature and humidity given from the Hefei city meteorological monitoring station were also recorded, in order to understand the environmental factors related to the concentration changes of NO$_3$ and N$_2$O$_5$ in the atmosphere. The results are shown in FIG. 5. The NO$_3$ and N$_2$O$_5$ data were averaged with a time resolution of 90 s and the NO$_2$ signal was averaged for 180 s. The PM$_{2.5}$, O$_3$, temperature and humidity data were only available with a time resolution of 1 h.

 FIG. 5 Ambient measurements of atmospheric pollutants. Yellow areas indicate the day time between sunrise and sunset

It is noted that NO$_3$ and N$_2$O$_5$ were only produced after sunset, and their concentrations during the daytime were below our detection limit, so our measurements of NO$_3$ and N$_2$O$_5$ were carried out only at night. The yellow background in FIG. 5 indicates the daytime period, and the white background is for the night period. It can be seen that during the entire measurement, the N$_2$O$_5$ concentration had a few peaks of about 10$-$20 pptv, while the concentrations of NO$_3$ stayed lower than 2 pptv. Only around 00:00 on November 5, when N$_2$O$_5$ showed a persistent high concentration, the NO$_3$ concentration had an obvious peak lasting for about 2 h.

In fact, the concentrations of NO$_3$ and N$_2$O$_5$ in the atmosphere were the result of the interaction of various atmospheric conditions. As a short-term continuous monitoring, only the relatively obvious and typical events were selected for analysis, giving rise to a preliminary correlation analysis.

As shown in FIG. 5, when N$_2$O$_5$ reached a peak concentration, NO$_2$ concentration always came to a minimum. The correlation was more apparent when the concentration of N$_2$O$_5$ peaked during the night from November 4 to November 5 and the night from November 5 to November 6, which was consistent with the reaction mechanism (Eq.(1)). It was also noticed that there was a small O$_3$ peak during the night of November 4. A possible interpretation is that N$_2$O$_5$ produced in the atmosphere at this time was from the production of NO$_3$ by O$_3$ oxidation of NO$_2$ and then formation of N$_2$O$_5$ by combining NO$_3$ and NO$_2$.

A special atmospheric process observed in this measurement was that after the accumulation of PM$_{2.5}$ during the daytime of November 3, it was quickly cleared in the night of November 3 and kept a low concentration in the following days. By comparing the measurement results of N$_2$O$_5$ before and after the sudden drop in the PM$_{2.5}$ concentration, it can be seen that low PM$_{2.5}$ concentration appeared to be a prerequisite for the presence of N$_2$O$_5$. When the PM$_{2.5}$ concentration was high, the N$_2$O$_5$ concentration was low, which may be due to that the aerosol particles provided more reaction surface for NO$_3$ and considerably reduced the lifetime of NO$_3$ in the atmosphere.

It can also be seen from FIG. 5 that the concentration of N$_2$O$_5$ after November 3 was higher than that of the previous two days, correlated with a low humidity during the same period. However, since the decrease of PM$_{2.5}$ was also related to the decrease of humidity, it was difficult to determine whether humidity or PM$_{2.5}$ concentration played a key role in the change of the N$_2$O$_5$ concentration, which requires further investigation.

Ⅳ. CONCLUSION

A dual-channel cavity ring-down instrument based on diode lasers using the 662 nm absorption band of NO$_3$ was developed for measurements of NO$_3$ and N$_2$O$_5$ in the atmosphere. A NO$_3$ detection sensitivity of 1.1 pptv (3$\sigma$) was demonstrated. The possible interference due to other substances in the atmosphere, such as NO$_2$, O$_3$, and water was discussed. After calibration of the instrument with laboratory NO$_3$ standard sample, the NO$_3$ and N$_2$O$_5$ concentrations in the atmosphere can be simultaneously determined quantitatively at a precision better than 1 pptv. Using this instrument, together with another CRDS instrument for the NO$_2$ detection, a field measurement lasting for 1 week was carried out by measuring NO$_3$, N$_2$O$_5$ and NO$_2$. Combining with the meteorological data of ozone, PM$_{2.5}$, humidity and temperature from the Hefei city during this period, correlations among the concentrations of these pollutants were analyzed. It was found that the atmospheric concentrations of NO$_3$ and N$_2$O$_5$ were affected by a variety of atmospheric conditions during a rapid atmospheric cleaning event in the winter of 2017.

Ⅴ. ACKNOWLEDGMENTS

Hao Wu, Jian Chen, An-wen Liu, and Shui-ming Hu acknowledge the supports from the Ministry of Science and Technology of China (No.2013BAK12B00), and the National Natural Science Foundation of China (No.21427804).

Reference
 [1] B. J. Allan, N. Carslaw, H. Coe, R. A. Burgess, and J. M. C. Plane, J. Atmos. Chem. 33, 129(1999). DOI:10.1023/A:1005917203307 [2] S. S. Brown, and J. Stutz, Chem. Soc. Rev. 41, 6405(2012). DOI:10.1039/c2cs35181a [3] J. L. Fry, D. C. Draper, K. C. Barsanti, J. N. Smith, J. Ortega, P. M. Winkler, M. J. Lawler, S. S. Brown, P. M. Edwards, R. C. Cohen, and L. Lee, Environ. Sci. Technol. 48, 11944(2014). DOI:10.1021/es502204x [4] R. Atkinson, D. L. Baulch, R. A. Cox, J. N. Crowley, R. F. Hampson, R. G. Hynes, M. E. Jenkin, M. J. Rossi, J. Troe, and T. J. Wallington, Atmos. Chem. Phys. 8, 4141(2008). DOI:10.5194/acp-8-4141-2008 [5] W. L. Chang, P. V. Bhave, S. S. Brown, N. Riemer, J. Stutz, and D. Dabdub, Aerosol Sci. Technol. 45, 655(2011). [6] F. J. Dentener, and P. J. Crutzen, J. Geophys. Res. 98, 7149(1993). DOI:10.1029/92JD02979 [7] D. Hanway, and F. M. Tao, Chem. Phys. Lett. 285, 459(1998). DOI:10.1016/S0009-2614(97)01382-1 [8] S. S. Brown, J. A. Neuman, T. B. Ryerson, M. Trainer, W. P. Dubé, J. S. Holloway, C. Warneke, J. A. de Gouw, S. G. Donnelly, E. Atlas, B. Matthew, A. M. Middlebrook, R. Peltier, R. J. Weber, A. Stohl, J. F. Meagher, F. C. Fehsenfeld, and A. R. Ravishankara, Geophys. Res. Lett. 33, L08801(2006). [9] N. Riemer, H. Vogel, B. Vogel, B. Schell, I. Ackermann, C. Kessler, and H. Hass, J. Geophys. Res. D Atmos. 108, 5(2003). [10] A. Geyer, B. Alicke, R. Ackermann, M. Martinez, H. Harder, W. Brune, P. Di Carlo, E. Williams, T. Jobson, S. Hall, R. Shetter, and J. Stutz, J. Geophys. Res. Atmos. 108, 4368(2003). DOI:10.1029/2002JD002967 [11] J. F. Noxon, R. B. Norton, and W. R. Henderson, Geophys. Res. Lett. 91, 5323(1978). [12] J. P. Kercher, T. P. Riedel, and J. A. Thornton, Atmos. Meas. Tech. 2, 193(2009). DOI:10.5194/amt-2-193-2009 [13] D. Mihelcic, D. Klemp, P. Müsgen, H. W. Pätz, and A. Volz-Thomas, J. Atmos. Chem. 16, 313(1993). DOI:10.1007/BF01032628 [14] A. Geyer, B. Alicke, D. Mihelcic, J. Stutz, and U. Platt, J. Geophys. Res. Atmos. 104, 26097(1999). DOI:10.1029/1999JD900421 [15] J. Matsumoto, H. Imai, N. Kosugi, and Y. Kajii, Atmos. Environ. 39, 6802(2005). DOI:10.1016/j.atmosenv.2005.07.055 [16] T. Wagner, C. Otten, K. Pfeilsticker, I. Pundt, and U. Platt, Geophys. Res. Lett. 27, 3441(2000). DOI:10.1029/1999GL011153 [17] R. B. Norton, and J. F. Noxon, J. Geophys. Res. 91, 5323(1986). DOI:10.1029/JD091iD05p05323 [18] U. Platt, J. Meinen, D. Pöhler, and T. Leisner, Atmos. Meas. Tech. 3, 127(2010). DOI:10.5194/amt-3-127-2010 [19] R. M. Varma, D. S. Venables, A. A. Ruth, U. Heitmann, E. Schlosser, and S. Dixneuf, Appl. Opt. 18, B159-71(2009). [20] M. Baasandorj, S. W. Hoch, R. Bares, J. C. Lin, S. S. Brown, D. B. Millet, R. Martin, K. Kelly, K. J. Zarzana, C. D. Whiteman, W. P. Dube, G. Tonnesen, I. C. Jaramillo, and J. Sohl, Environ. Sci. Technol. 51, 5941(2017). DOI:10.1021/acs.est.6b06603 [21] Z. Li, R. Hu, P. Xie, H. Wang, K. Lu, and D. Wang, Sci. Total Environ. 613- 614, 131(2018). [22] S. Wang, C. Shi, B. Zhou, H. Zhao, Z. Wang, S. Yang, and L. Chen, Atmos. Environ. 70, 401(2013). DOI:10.1016/j.atmosenv.2013.01.022 [23] Z. Li, R. Hu, P. Xie, H. Chen, S. Wu, F. Wang, Y. Wang, L. Ling, J. Liu, and W. Liu, Opt. Express. 26, A433(2018). DOI:10.1364/OE.26.00A433 [24] X. Wang, T. Wang, C. Yan, Y. J. Tham, L. Xue, Z. Xu, and Q. Zha, Atmos. Meas. Tech. 7, 1(2014). [25] J. Chen, H. Wu, A. W. Liu, S. M. Hu, and J. Zhang, Chin. J. Chem. Phys. 30, 493(2017). DOI:10.1063/1674-0068/30/cjcp1705084 [26] W. P. Dubá, S. S. Brown, H. D. Osthoff, M. R. Nunley, S. J. Ciciora, M. W. Paris, R. J. McLaughlin, and A. R. Ravishankara, Rev. Sci. Instrum. 77, 034101(2006). DOI:10.1063/1.2176058 [27] H. Fuchs, W. P. Dubé, S. J. Ciciora, and S. S. Brown, Anal. Chem. 80, 6010(2008). DOI:10.1021/ac8007253 [28] H. D. Osthoff, M. J. Pilling, A. R. Ravishankara, and S. S. Brown, Phys. Chem. Chem. Phys. 9, 5785(2007). DOI:10.1039/b709193a [29] I. E. Gordon, L. S. Rothman, C. Hill, R. V. Kochanov, Y. Tan, P. F. Bernath, M. Birk, V. Boudon, A. Campargue, K. V. Chance, B. J. Drouin, J. M. Flaud, R. R. Gamache, J. T. Hodges, D. Jacquemart, V. I. Perevalov, A. Perrin, K. P. Shine, M. A. H. Smith, J. Tennyson, G. C. Toon, H. Tran, V. G. Tyuterev, A. Barbe, A. G. Császár, V. M. Devi, T. Furtenbacher, J. J. Harrison, J. M. Hartmann, A. Jolly, T. J. Johnson, T. Karman, I. Kleiner, A. A. Kyuberis, J. Loos, O. M. Lyulin, S. T. Massie, S. N. Mikhailenko, N. Moazzen-Ahmadi, H. S. P. Müller, O. V. Naumenko, A. V. Nikitin, O. L. Polyansky, M. Rey, M. Rotger, S. W. Sharpe, K. Sung, E. Starikova, S. A. Tashkun, J. Vander Auwera, G. Wagner, J. Wilzewski, P. Wcisło, S. Yu, and E. J. Zak, J. Quant. Spectrosc. Radiat. Transf. 203, 3(2017). DOI:10.1016/j.jqsrt.2017.06.038 [30] V. Gorshelev, A. Serdyuchenko, M. Weber, W. Chehade, and J. P. Burrows, Atmos. Meas. Tech. 7, 609(2014). DOI:10.5194/amt-7-609-2014 [31] S. P. Sander, R. R. Friedl, D. M. Golden, M. J. Kurylo, G. K. Moortgat, P. H. Wine, A. R. Ravishankara, C. E. Kolb, M. J. Molina, S. Diego, L. Jolla, R. E. Huie, and V. L. Orkin, Chemical Kinetics and Photochemical Data for Use in Atmospheric Studies: Evaluation Number 17, JPL Publication, 10-06 (2006). [32] W. B. Demore, J. J. Margitan, M. J. Molina, R. T. Watson, D. M. Golden, R. F. Hampson, M. J. Kurylo, C. J. Howard, and A. R. Ravishankara, Int. J. Chem. Kinet. 17, 1135(1985). DOI:10.1002/kin.550171010

a. 中国科学技术大学合肥微尺度物质科学国家研究中心，合肥 230026;
b. 中国科学技术大学化学与材料科学学院，合肥 230026;
c. 美国加州大学河滨分校化学系和大气污染研究中心，河滨 92521