A novel calibration method for atmospheric NO3 radical via high reflectivity cavity

The NO3 radical is one of the important indicators of atmospheric oxidation capability at night, and its accurate detection has been extensively investigated. High reflectivity cavity instruments, such as cavity ring-down spectroscopy (CRDS) and cavity-enhanced absorption spectroscopy (CEAS), are generally employed for NO3 radical detection. However, sampling detection easily adsorbs radicals on surfaces, which causes complications and even additional loss. According to the chemical conversion relationship between NO3 radical and NO2, a method of absolute calibration different from the previous relative calibration method is applied by combining the CRDS and CEAS instruments for the NO3 radical and NO2, respectively. Experimental results show that the overall transmission efficiency of the NO3 radical for the CRDS instrument is approximately 75% ± 9% at a flow rate of 6.5 SLPM. Comparative experiments between the two methods are conducted to estimate the accuracy of this method. The achieved results indicate that absolute calibration is a reliable and feasible method for radical calibration in the laboratory. Furthermore, the absolute calibration method has the potential to realize accurate real-time calibration in field measurements.


Introduction
The NO 3 radical is a key nocturnal form of nitrogen oxide, which is a major oxidant for pollutants during the night and plays an important role in the nocturnal chemical process Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. [1,2]. NO 3 also contributes to the oxidative capacity of the atmosphere [3][4][5]. Accurately measuring NO 3 radical concentrations in the atmosphere is challenging due to their high reactivity and low concentration (dozens of pptv) [6][7][8][9].
In recent years, several spectrometers for in situ, or point measurements of NO 3 and N 2 O 5 concentrations in ambient air, such as cavity-enhanced absorption spectroscopy (CEAS) and cavity ring-down spectroscopy (CRDS), have been constructed [10][11][12]. CEAS and CRDS have the advantages of high sensitivity, mobility, and flexibility. However, point sampling measurement instruments, such as CEAS and CRDS, will cause additional losses because the NO 3 radical is a reactive species. The concentration of NO 3 radical in the atmosphere is only a few decades of pptv, but the loss in the cavity is relatively high. Therefore, the loss of NO 3 radical in point sampling instruments has received increasing attention from researchers [13][14][15][16]. Brown et al found the loss of NO 3 radical in the system to be as much as 20%, as determined by a relative calibration method based on the variation of detected NO 3 with residence time, due to surface loss of NO 3 radical through the PFA tubing and membrane filter loss [17]. The loss of NO 3 radical in membranes changes over time, and variation reaches 10% per hour [14]. Furthermore, the loss of NO 3 radical in the system is a variable, and it will change as the sampling time, particle concentration, and environmental condition [18][19][20][21]. Therefore, accurate real-time calibration is crucial for point measurements of NO 3 radical in ambient air. At present, the relative calibration method is mainly applied. However, this method is relatively complex due to the respective calibrations of surface and membrane filter loss. And it will cause substantial errors due to not real-time correction.
An absolute calibration method, which is different from the previous relative calibration method, is described in the present work. According to the chemical conversion relationship between NO 3 and NO 2 , this accurate real-time calibration method is applied by combining the CRDS and CEAS instruments for the NO 3 radical and NO 2 , respectively. A low stable concentration of NO 3 radical is produced by the synthetic N 2 O 5 solid source in the laboratory. Consecutive calibrations are performed by the constructed absolute calibration system of NO 3 radical, and the accuracy of calibration results is further verified by compared with the relative calibration method.

Experimental setup
The CRDS and CEAS instruments for the NO 3 radical and NO 2 are respectively combined. The outlet of the CRDS instrument connects with the inlet of the CEAS instrument, and the schematic of the instruments is shown in figure 1. The CRDS instrument mainly comprises a diode laser, a highfinesse optical cavity with HR mirrors, and a PMT. Light is provided by an external modulation diode laser (line width of 0.3 nm and center wavelength of 661.85 nm) with an optical output of approximately 120 mW. The 76 cm long cavity is formed by two high-reflectivity mirrors (R ≥ 99.9985%) with 2.54 cm diameter and 1 m curvature radius. The output light of the cavity is detected by PMT. The signals from the PMT are then digitized using an oscilloscope card controlled by the LabVIEW program [16]. The ring-down time of the CRDS system filled with N 2 is approximately 67.0 µs (τ 0 ), and an effective absorption path length of approximately 20 km is achieved, as shown in figure 2(a). Individual decay profiles are co-added and then averaged to achieve a high signal-tonoise ratio [22]. An example of cavity decay trace is fitted by the Levenberg-Marquardt algorithm in figure 2(a). The best integration time is determined by the Allan variance shown in figure 2(b). Considering high reactivity of NO 3 radical, the integration time is 5 s in the process of calibration. In the laboratory, the detection limit of the CRDS system is 2.6 pptv (1σ) for a 5 s integration time using equation (1) [16]. The CEAS set-up comprises a blue light-emitting diode light source (LZ1-00B205, LedEngin) with a peak wavelength of approximately 458 nm and FWHM of 25 nm, a pair of high-reflectivity cavity mirrors (maximum reflectivity of 0.9995 in the range of 440-480 nm) with 25 mm diameter and 1.5 m curvature radius, an optical cavity of 73.5 cm, an optical band-pass filter, and an optical fiber-coupled grating spectrometer. The cavity of CEAS is filled with sample gases in the laboratory, and the detection sensitivity of the CEAS system is evaluated to be approximately 0.25 ppbv (1σ) for NO 2 with a 20 s acquisition time [23,24]: where R L is the ratio of the total cavity length to the length over which the absorber is present in the cavity. For the CRDS system, τ 0 is roughly 67.0 µs and σ (τ 0 ) is approximately 0.10 µs for a 5 s integration time under laboratory conditions. Taking the absorption cross section (σ = 2.02 × 10 −17 cm 2 molecule −1 ) for the NO 3 radical at 661.85 nm [16], the detection limit of NO 3 the radical is 2.6 pptv (1σ) for a 5 s integration time.

Formation of standard source of NO 3
NO 3 radical is characterized by its high reactivity and easy collision loss. Thus, NO 3 radical has no standard gas. In order to meet the requirements of loss calibration of NO 3 radical, the method of synthetic N 2 O 5 solid source is used to produce standard gas in the laboratory. With the reaction of excessive NO and O 3 , N 2 O 5 solid source is synthesized in low temperatures (approximately −80 • C). In the process, the synthetic solid is repeatedly purified by the high concentration of O 3 to improve the purity of synthetic gas source. Simultaneously, water vapor interference cannot be ignored, otherwise which will cause synthesis failure by forming HNO 3 . The schematic of loss calibration of the NO 3 radical is shown in figure 3. The CRDS and CEAS instruments for NO 3 radical and NO 2 , respectively, are applied together to obtain accurate measurements. The stable and low concentration of standard gas, which is the key step of accurate calibration, is generated. The sample gas of N 2 O 5 is generated by passing a small flow of N 2 (approximately 10-20 ml min −1 ) over the synthetic N 2 O 5 solid stored at −78 • C through the dry ice method. The gas is diluted by N 2 several times to obtain low concentrations of N 2 O 5 . The sample gas of N 2 O 5 is heated at high temperature (the tube surface is heated to approximately 80 • C and controlled by temperature controller) to generate NO 3 radical due to the chemical thermal equilibrium relation of N 2 O 5 and NO 3 . The high concentration sample gas is then diluted by N 2 again. There are two main reasons for repeated dilution, one is to obtain the low concentration of NO 3 radical (hundreds of pptv), the other is to make the high temperature of sample gas to close to room temperature. The low stable concentration of NO 3 radical is obtained by controlling the gas flow of exhaust and dilution with mass flow gas meter and regulating the heated temperature controlled by the temperature controller (precision of approximately ±1 • C). Finally, the low stable concentration gas passes into the CRDS and CEAS instruments for NO 3 radical and NO 2 detection.

Calibration method
According to the chemical conversion relationship of NO 3 radical and NO 2 (as shown in equation (2)), the overall loss of NO 3 radical in the CRDS system is calibrated by measuring the variation of NO 2 after adding NO. The overall loss calibration of NO 3 radical can be divided into several parts as shown below. First, the two sets of instrument systems are passed into N 2 to detect the background signal without gas absorption. Second, the instruments of CRDS and CEAS are separately applied to detect NO 3 radical and NO 2 of the sample gas through the high-temperature heating and dilution. Third, NO 3 radical sufficiently reacts with NO with the addition of excess NO. Every molecule of NO 3 radical will generate two molecules of NO 2. Simultaneously, the concentrations of NO 3 radical and NO 2 are respectively measured by CRDS and CEAS. Three main sources of NO 2 concentration detection through the CEAS instrument are available: production through the decomposition of N 2 O 5 solid sources, production through the reaction of NO 3 radical and NO, and production through the addition of impure NO. Finally, avoiding passing of the sample gas and only adding NO, the concentration of NO 2 from the impure NO is measured by the CEAS instrument. The transmission efficiency of NO 3 radical can be calculated by equation (3): where

Calibration results
The single calibration process of NO 3 radical in the CRDS system is shown in figure 4. The concentrations of NO 3 radical and NO 2 are respectively presented in red and black lines. The abscissa is the time of calibration, and the ordinate is the concentrations of NO 3 radical and NO 2 . The sample gas obtained from high-temperature decomposition and dilution passes into the two sets of instruments for detection. The concentrations of NO 3 radical and NO 2 respectively attain 1.57 ppbv and 7.13 ppbv. Excess NO is then added to titrate the NO 3 generated from the thermal dissociation of N 2 O 5 . NO rapidly reacts with NO 3 radical, and the concentration of NO 3 instantly reduces to the background value. Meanwhile, NO 2 is formed by the reaction as shown in equation, and every molecule of NO 3 radical generates two molecules of NO 2 . Consequently, the concentration of NO 2 rises to approximately 12.61 ppbv. The background concentration of NO 2 is detected considering the problem of impure NO. The sample gas is shut down, and the concentration of NO 2 from the NO source is measured to approximately 1.52 ppbv. In the experiment, the measuring time of each step is approximately 1-2 min. Figure 5(a) summarizes the results from four consecutive calibrations, and the time of each calibration is approximately 7-8 min. This figure shows the original concentration of NO 3 radical and the corresponding concentration variation of NO 2 . This variation is represented as ∆NO 2 , which is the NO 2 concentration from the formation through the reaction of NO 3 radical and NO. During the measurement, NO 3 radical concentration is approximately 1.33 ppbv and ∆NO 2 is close to 3.54 ppbv. Their concentrations slightly increase during the four consecutive calibration periods but are relatively stable in the process of each calibration. The main reason is that synthetic N 2 O 5 solid source is not completely pure. The transmission efficiency of NO 3 radical is obtained in accordance with equation (3) and ∆NO 2 . The transmission efficiency of NO 3 radical in the CRDS system for four consecutive calibrations and the measurement error for each calibration are presented in figure 5(b). The accuracies of the CRDS and CEAS instruments are approximately ±8% [16] and ±6% [24], respectively. Considering the standard deviation measurement and the instrument accuracy of CRDS and CEAS, the relative uncertainty of the absolute calibration method is estimated to be 11%. The average transmission efficiency of the NO 3 radical for the CRDS instrument is calculated to be 75% ± 9% at a 6.5 SLPM flow rate.
Calibrations for the same CRDS instrument are also performed by the relative calibration method. The relative method requires the calibration of the filter loss of the Teflon membrane and the surface loss of the NO 3 radical through the PFA tubing. The loss rate coefficient of the NO 3 radical through the PFA tubing is 0.19 ± 0.02 s −1 , and the filter loss of the new Teflon membrane filter is approximately 8% ± 3% [16]. The surface loss of the NO 3 radical through the PFA tubing is calculated to be approximately 6% ± 1% by the loss rate coefficient at a flow rate of 6.5 SLPM. In addition, the surface loss of the filter holder is approximately 9% ± 4%, and the CRD instrument accuracy is approximately ±8%; the details Contrasted results of the relative and absolute calibration methods. The transmission efficiency data of the NO 3 radical for the relative calibration are from [16].
are described by Wang et al [16]. The comparative results for the two methods are shown in figure 6. According to the above measurements, the overall inlet transmission efficiencies of the NO 3 radical for the instrument are determined to be 75% ± 9% and 77% ± 11% through the absolute and the relative calibration methods, respectively, for the same CRDS instrument. The contrasted results are basically consistent. This finding further verifies the accuracy of transmission efficiency of the NO 3 radical for the absolute calibration method.

Conclusions
A novel calibration method is described in this study. The calibration system of the NO 3 radical is designed and constructed using the CRDS and CEAS instruments. Meanwhile, the standard gas of the NO 3 radical is produced by the synthetic N 2 O 5 solid source in the laboratory. Four consecutive calibrations are conducted by the absolute method, and the average transmission efficiency of the NO 3 radical is approximately 75% ± 9% at a flow rate of 6.5 SLPM. Contrast measurements between the two methods are performed, and the results are basically consistent. The method is precise, rapid, and concise. Furthermore, the method can be independently applied to a common point sampling measurement instrument. Future research will focus on automated integrative calibration and application of automated absolute calibration systems for ambient NO 3 radical in field measurements.