Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Low blank sampling method for measurement of the nitrogen isotopic composition of atmospheric NOx

  • Kazuki Kamezaki ,

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Validation, Visualization, Writing – original draft, Writing – review & editing

    kamezaki-k@aist.go.jp

    Affiliation Environmental Management Research Institute, National Institute of Advanced Industrial Science and Technology (EMRI/AIST), Tsukuba, Japan

    ⨯
  • Takahisa Maeda,

    Roles Data curation, Formal analysis, Investigation, Software, Writing – review & editing

    Affiliation Environmental Management Research Institute, National Institute of Advanced Industrial Science and Technology (EMRI/AIST), Tsukuba, Japan

    ⨯
  • Shigeyuki Ishidoya,

    Roles Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Supervision, Validation, Writing – review & editing

    Affiliation Environmental Management Research Institute, National Institute of Advanced Industrial Science and Technology (EMRI/AIST), Tsukuba, Japan

    ⨯
  • Ayumi Tsukasaki,

    Roles Data curation, Formal analysis, Writing – review & editing

    Affiliation Environmental Management Research Institute, National Institute of Advanced Industrial Science and Technology (EMRI/AIST), Tsukuba, Japan

    ⨯
  • Shohei Murayama,

    Roles Writing – review & editing

    Affiliation Environmental Management Research Institute, National Institute of Advanced Industrial Science and Technology (EMRI/AIST), Tsukuba, Japan

    ⨯
  • Naoki Kaneyasu

    Roles Writing – review & editing

    Affiliations Environmental Management Research Institute, National Institute of Advanced Industrial Science and Technology (EMRI/AIST), Tsukuba, Japan, Fukushima Institute for Research, Education and Innovation, Namie-machi, Fukushima, Japan

    ⨯

Abstract

The nitrogen isotopic composition of nitrogen oxide (NOx) is useful for estimating its sources and sinks. Several methods have been developed to convert atmospheric nitric oxide (NO) and/or nitrogen dioxide (NO2) to nitrites and/or nitrates for collection. However, the collection efficiency and blanks are poorly evaluated for many collection methods. Here, we present a method for collecting ambient NOx (NO and NO2 simultaneously) with over 90% efficiency collection of NOx and low blank (approximately 0.5 μM) using a 3 wt% hydrogen peroxide (H2O2) and 0.5 M sodium hydride (NaOH) solution. The 1σ uncertainty of the nitrogen isotopic composition was ± 1.2 ‰. The advantages of this method include its portability, simplicity, and the ability to collect the required amount of sample to analyze the nitrogen isotopic composition of ambient NOx in a short period of time. Using this method, we observed the nitrogen isotopic compositions of NOx at the Tsukuba and Yoyogi sites in Japan. The averaged δ15N(NOx) value and standard deviation (1σ) in the Yoyogi site was (−2.7 ± 1.8) ‰ and in the Tsukuba site was (−1.7 ± 0.9) ‰ during the sampling period. The main NOx source appears to be the vehicle exhaust in the two sites.

Citation: Kamezaki K, Maeda T, Ishidoya S, Tsukasaki A, Murayama S, Kaneyasu N (2024) Low blank sampling method for measurement of the nitrogen isotopic composition of atmospheric NOx. PLoS ONE 19(2): e0298539. https://doi.org/10.1371/journal.pone.0298539

Editor: Eva Elisabeth Stüeken, University of Saint Andrews, UNITED KINGDOM

Received: September 13, 2023; Accepted: January 25, 2024; Published: February 29, 2024

Copyright: © 2024 Kamezaki et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: The datasets generated and analyzed during this study are available via Zenodo at https://doi.org/10.5281/zenodo.10398614.

Funding: This study was supported by the Steel Foundation for Environmental Protection Technology in the form of a grant to KK [FY2021–2022] and by the Japan Society for the Promotion of Science (JSPS) in the form of Grants-in-Aid for Scientific Research (KAKENHI Program) to SI [JP22H05006], and SI and KK [JP22H05006].

Competing interests: The authors have declared that no competing interests exist.

Introduction

Nitric oxide (NO) and nitrogen dioxide (NO2) are collectively referred to as nitrogen oxide (NOx). NOx is a primary pollutant in the atmosphere and is involved in urban environmental issues such as photochemical smog, acid rain, tropospheric ozone production, and human health. Besides, NOx deposition can enhance ecosystem productivity through fertilization or decrease it through nutrient imbalances and reduce ecosystem biodiversity through acidification and eutrophication [1]. NO2 is oxidized to nitrate (NO3−), which is adsorbed by aerosols and transported over long distances, affecting distant environments [2]. In areas affected by human pollution, fossil-fuel combustion from traffic, residential heating, cooking, industry, and energy sectors are the main sources of NOx. On the other hand, as a natural NOx source, biomass burning, biogenic production, and lightning are also important sources of NOx [3]. Global annual NOx emissions have been gradually curtailed [4]. However, it is important to understand the exact behaviour of NOx to elucidate how the suppression of NOx emissions changes atmospheric reactions.

The nitrogen isotopic composition (δ15N value) of NOx is a useful tool for estimating its sources because the nitrogen isotope composition of each source has a unique value (S1 Table). Nitrogen source is identified from the nitrogen isotopic composition of NO3− in the aerosol [5–7]. To date, several methods have been developed to convert atmospheric NO and/or NO2 to nitrites and/or nitrates for collection. A denuder system, filter pack, and Ogawa sampler have all been used to collect ambient NO2 with reagents, such as triethanolamine, guaiacol, and potassium hydroxide [8–13]. A wet method was used to collect NOx by passing air containing NOx through the recovery solution. Potassium permanganate (KMnO4) with sodium hydroxide (NaOH) or 20% triethanolamine in water have been used as the recovery solutions for the collection of ambient NOx [14–16]. In addition, a solid sorbent method with attached chemical reagents has also been reported for the collection of ambient NOx [17]. It is important that the collection efficiency of NOx is close to 100% and that the blank is small for the isotopic composition analysis. However, the collection efficiency and blanks are poorly evaluated for many collection methods and the locations at which the δ15N values of NOx were measured were limited owing to the difficulty of the measurement method.

Recently, a high-time-resolution method for NOx collection was developed using gas-washed bottles in KMnO4 and NaOH recovery solutions. This method shows high collection efficiency for NOx, whereas a high concentration of NOx blank (approximately 5 μM) is observed [14, 15]. Therefore, at present, there is almost no fully validated simple method that can collect NO and NO2 for the analysis of the δ15N values of NOx with high efficiency. Compared to the KMnO4/NaOH recovery solution, the hydrogen peroxide (H2O2)/NaOH recovery solution can remove NOx more efficiently, as reported by Ohta et al. [18] and Kuropka, [19].

In highly alkaline conditions, H2O2 produces various intermediate products that act as oxidants with H2O2 decomposition [20]. Free radicals generated due to H2O2 decomposition efficiently oxidize NO. It has been pointed out that particularly oxygen anions (O2−) produced at high pH may effectively oxidize NO [21]. On the other hand, NO and NO2 dissolve in NaOH solution, and the presence of H2O2 accelerates the oxidation of NO2 [18, 22]. The mechanism of the reaction of NOx with H2O2/NaOH is expressed as follows: (1) (2) (3) (4) (5) (6) In this study, we tested and developed a more efficient NO and NO2 collection method using a H2O2 /NaOH recovery solution for the sampling method and measurement of nitrogen isotopic composition of atmospheric NOx. This method has high NOx collection efficiency and low NO2− and NO3− blanks in the recovery solution.

Materials and methods

Commercial NOx samples and recovery solution preparation

Commercial cylinders containing 91 ppm NO (Sample A, Japan Fine Products Co. Ltd., Kanagawa, Japan) and 5 ppm NO2 (Sample B, Japan Fine Products Co. Ltd., Kanagawa, Japan) balanced with N2 were used in this study. To dilute these high-concentration NO and NO2 gases, pure N2 (99.995% purity) was used.

To prepare 200 mL of the recovery solution, a highly concentrated 10 M NaOH solution was prepared using reagent-grade NaOH (Special Grade; FUJIFILM Wako Pure Chemical Corp., Osaka, Japan). 10 mL of concentrated NaOH solution was added and diluted with 80 mL of 18.2 MΩ cm water produced by IQ7010 (Merck Millipore Corporation, Massachusetts, United States) in a beaker. Next, 20 mL of 35 wt% H2O2 (Special grade, FUJIFILM Wako Pure Chemical Corp., Osaka, Japan) was added. The entire solution was made up to 200 mL with an additional 18.2 MΩ·cm water. NaOH was diluted before adding H2O2, because H2O2 decomposes rapidly in highly concentrated basic solutions [23]. The old reagents were not used because H2O2 gradually decomposes; therefore, the prepared reagents were used within one day. In our experimental study, it was observed that refrigerated H2O2 remained usable for a period of six months following its purchase. However, after a duration of nine months, the H2O2 failed to generate bubbles even upon NaOH addition, and its NOx trapping efficiency was low. After preparing the reagent, the reaction was allowed to proceed for 30 min to 1 h before the H2O2/NaOH recovery solution was used (35 wt% H2O2 was used in this study).

For comparison with the conventional method, we prepared a KMnO4 /NaOH recovery solution with reference to Fibiger et al. [14]. Briefly, 1 N (0.2 M) KMnO4 was prepared from reagent-grade KMnO4 (Special grade, FUJIFILM Wako Pure Chemical Corp., Osaka, Japan). The 125 mL, 1 N KMnO4 was diluted to 300 mL, and 25 mL of 10 M NaOH was added. The entire solution was 500 mL with an additional 18.2 MΩ·cm of water. The KMnO4/NaOH recovery solution was stored in a 500 mL amber glass bottle and used within one day to prevent contamination with NOx.

Sampling system

A schematic of the sampling system is shown in Fig 1. All the tubes were 1/4-inch. Three bubblers (080100–02, SIBATA, Tokyo, Japan) were used to maintain high NOx collection efficiency. A rubber tube was used as a bubbler joint. A PTFE filter (Advantec Co. Ltd., Tokyo, Japan) and a 0.45 μm pore size Whatman’s nylon membrane filter equipped with PFA filter folders (Savillex, Minnesota, United States) were used in front of the bubblers to remove aerosols and gas-phase nitrate (HNO3) from the atmosphere. The flow rate was controlled by using a valve immediately prior to the pump. An 8 μm pore size hydrophilic filter (Merck Millipore Ltd., Massachusetts, United States) equipped with a PFA filter folder was placed between the valve and bubblers to prevent water droplets from entering the pump. In addition, it was equipped with a gas flow multi-meter (Model 5210, TSI Incorporated, Minnesota, United States) that can measure pressure, temperature, flow rate, and integrated flow rate, as well as a pump (DAP-12S, ULVAC, Kanagawa, Japan) for atmospheric suction.

thumbnail
Download:
Fig 1. Schematic diagram of the NOx sampling system.

When the laboratory experiment was to collect diluted samples A and B, the multi-meter was moved in front of the particle filter.

https://doi.org/10.1371/journal.pone.0298539.g001

Laboratory experiment

NOx collection.

To test the NOx collection efficiency, the pump was replaced with a NOx analyser (APNA-370, HORIBA, Kyoto, Japan or Serinus 40, ACOEM ECOTECH, Melbourne, Australia) and a multi-meter was moved in front of the PTFE filter. A three-port valve was installed to bypass the three bubblers. The difference in concentrations between the two instruments was within 1 ppb for NO, NO2, and NOx at atmospheric concentrations (< 90 ppb). Cylinders containing NO or NO2 were diluted in two steps using pure N2 equipped with a mass flow controller (Kofloc, Kyoto, Japan) to prepare approximately 15 and 40 ppb NOx gas. The collection periods were one hour for each sampling and the flow rate was 0.8 L min−1. For the sample collection, the 10 mL of recovery solution was placed in three bubblers. For blanks, 5 mL of the solution was placed in three bubblers and collected immediately without sucking air and the blank was collected every 2–3 sample collection. Room temperature was maintained at 22–30°C. Prior to entering the collection system, some of the diluted NO or NO2 gas was vented outside to reduce the pressure and adjust the sampling pressure to atmospheric pressure (1000–1020 hPa). After pumping, all recovered solutions were transferred to a 60 mL amber plastic bottle. The collection efficiency was calculated from the reduction in NO and NO2 concentrations measured by the NOx analyser with and without passing through the three bubblers.

For the H2O2/NaOH recovery solution, after transferring the recovery solution to a 60 mL amber plastic bottle, H2O2 was removed from the recovery solution by adding manganese oxide (MnO2) (Special grade, FUJIFILM Wako Pure Chemical Corp., Osaka, Japan) to stop the reaction [24]. It is necessary to slightly loosen the lid of the plastic bottle or periodically open it because oxygen is generated. Hydrogen chloride (HCl) was added to neutralize the recovery solution.

For the KMnO4/NaOH recovery solution, we followed the method reported by Fibiger et al. [14]. After transferring the recovery solution to a 60 mL amber plastic bottle for storage, the solution was transferred to a well-washed 500 mL glass beaker. Then, a total of 10 mL of 35 wt% H2O2 was added to reduce KMnO4. When incorporating H2O2, exercise caution to use a 500 mL glass beaker with a wide mouth instead of the designated 60 mL amber plastic bottle. Failure to do so may result in an abrupt release of the solution. After stopping the reaction, the clear solution and brown MnO2 were transferred to a 50 mL centrifuge tube. Then, the solutions were added HCl to neutralize and centrifuged at 4800 rpm for 15 minutes. After centrifugation, the supernatant was collected. MnO2 adhering to the beaker can be removed by washing with HCl. Similarly, the sampling system also deposits MnO2 and clogs the frit, requiring HCl to be added to remove the MnO2 after several samples.

Ammonia impact assessment.

To test the effect of ammonium ion contamination on the δ15N values of NOx, 100 μM of NH4Cl (special grade, FUJIFILM Wako Pure Chemical Corp., Osaka, Japan) in water was added to the prepared recovery solutions. After seven days, MnO2 or H2O2 was added to the recovery solution to stop the reaction. Continuous flow analysis (CFA) (QuAAtro-2HR, BL TEC K.K., Tokyo, Japan) was used to compare the effect of ammonia mixing by measuring the concentrations of nitrate and nitrite ions in the samples with and without added ammonia.

Field measurement

Ambient NOx samples were collected at Tsukuba in the west office of AIST (Tsukuba site), Ibaraki, Japan (36.05°N, 140.12°E, 12 m above ground level), and Tokai University in Shibuya (Yoyogi site), Tokyo, Japan (35.66° N, 139.68° E, 52 m above ground level) from January to February 2023 on weekdays (S1 Fig). The Tsukuba site does not require permission as it is the author’s affiliated institution. Access to the Yoyogi site was granted by Tokai University. We confirmed that the field studies did not involve endangered or protected species. The flow rate ranges from 0.5 to 0.8 L min−1. After confirming that the NOx concentration did not change, a PFA or Dekabon tube was used to connect the air inlet to the sampling system. It was confirmed at Tsukuba that more than 95% of the NOx had been collected by branching the collection system and measuring the NOx concentrations before and after passing the collection system. The NO2− and NO3− blanks were about 0.5 μM in the Tsukuba and the Yoyogi sites.

Isotopic analysis.

Ten nmol of nitrate and/or nitrate ions in the obtained recovery solution were measured using the denitrifier method [25, 26]. If a small amount of MnO2 (about 0.2 g) was added stopping the reaction of the H2O2/NaOH recovery solution, the isotope ratio was not affected even if MnO2 dissolved in about 20 mL of solution was directly added to the vial. The automated injection line was modified from Hattori et al. [27] and a schematic diagram of the injection system is depicted in S2 Fig. Briefly, NO2− and NO3− were converted to nitrous oxide (N2O) by a strain of denitrifying bacteria, Pseudomonas aureofaciens, which has no N2O reductase. The N2O produced was then separated from other chemical species using chemical traps and a column equipped with a gas chromatograph (HP-plot Q; Agilent Technologies, Inc., California, United States) and was measured using an isotope-ratio mass spectrometer (IRMS) (MAT252; Thermo Fisher Scientific Inc., Massachusetts, United States) with industrial helium (99.99% purity) as carrier gas. Internationally recognized NO3– reference standards USGS 32, 34, 35 and their mixtures were measured alongside the samples and used to correct the resulting mass 45/44 and 46/44 ratios to obtain the final δ15N and δ18O, respectively. The 1σ analytical uncertainty of δ15N and δ18O values were ± 0.5 and ± 0.8 ‰, respectively. The low purity of helium affects the measurement precision, but industrial helium was used because ultrapure helium is expensive and not easily available. Although we confirmed with IRMS that the baseline for m/z 44, 45, and 46 did not increase when changing from ultrahigh purity helium to industrial-purity helium, the standard deviation (1σ) of the standard δ15N value deteriorated by 0.2‰. Impurities contained in helium were likely concentrated during the N2O purge and trap process. Although ultrahigh-purity helium was more suitable for isotope measurement, depending on the molecule, industrial-purity helium was deemed suitable for measurements with little deviation from the blank. Adding MnO2 to NO3– reference standards did not change the precision and accuracy of δ15N values.

Definition.

Stable isotopic compositions are typically reported as: (7) where X denotes 18O, and 15N, and R represents the ratios of 18O/16O, and 15N/14N in either the sample or standard material. The δ values are often quoted using per mil (δ) notation. The δ15N value was relative to atmospheric N2 (air), whereas the δ18O value is relative to Vienna Standard Mean Ocean Water (VSMOW). After analysis of the sample recovery solution and blank, the final sample isotopes were calculated using the mass balance: (8) where δXtotal and δXblank were determined by IRMS with the sample and blank measurement, respectively. [NO3−] total and [NO3−] blank were determined by CFA or IRMS with the sample and blank measurement, respectively. For the atmospheric NOx sample, to ensure precise and accurate measurement of the δ15N values, we considered the Δ17O (Δ17O = δ17O – 0.52 × δ18O) [28] of the analyte N2O.

Results and discussion

Collection efficiency

Nitrogen isotope exchange between NO and NO2 has been suggested to influence N stable isotope compositions. For accurate NOx isotopic composition measurements, nitrogen isotope analysis of atmospheric NOx requires the collection of both NO and NO2 with high collection efficiency. Furthermore, because NO and NO2 have different physical properties, a differential assessment of NO and NO2 collection efficiencies is required. The NOx collection efficiencies are listed in Table 1. After the experiment, like Fibiger et al. [14], the volume of the solution decreased by a few mL, indicating droplet dispersal. To prevent loss of nitrate due to droplet scattering, three bubblers were used, although the collection rate did not differ considerably when two bubblers were used. Although the effect of this decrease in water content on the isotopic composition of nitrate is difficult to estimate, the concentration of nitrate in the third bubbler was the same as that in the blank. Droplet dispersal is mainly affected by the third-stage bubbler, but since the third stage has a low NOx concentration, the effect of droplet dispersal on the isotopic composition of nitrate was deemed to be negligible.

thumbnail
Download:
Table 1. Collection efficiency of NO and NO2 by H2O2/ NaOH and KMnO4/NaOH recovery solution.

https://doi.org/10.1371/journal.pone.0298539.t001

No significant difference was found when comparing the collection efficiency of both recovery solutions at NOx concentrations of 15 and 40 ppb (Table 1). The overall averaged H2O2/NaOH recovery solution collected over 90% of NO and over 95% of NO2. Over 90% of NO and NO2 collection efficiency using H2O2/NaOH recovery solution was also reported by Ohta et al. [18], when the concentration of the recovery solution was over 0.6% H2O2 and 0.24 M NaOH. Note that while air is flowing, CO2 reacts with NaOH in solution, lowing the pH [21, 22]. Thus, the concentration of NaOH should be greater than 0.24 M. More than 95% of NOx can be collected by using H2O2/NaOH recovery solution since NO2/NOx in the atmosphere mostly exceeds 50%. The high collection efficiency of the H2O2/NaOH recovery solution probably made isotopic fractionation negligible during sampling. However, it is necessary to consider the differences in isotopic composition due to the differences in the collection efficiencies of NO and NO2. The reported δ15N(NO2) values ranges from −22 to 5‰ [10–12, 17]. Since fractionation factors for 15N substitution between NO and NO2 ranged from 1.040 at 278 K to 1.034 at 310 K [29], when the δ15N(NO2) values are from −22 to 0.4 ‰, the expected δ15N(NO) values is higher by approximately 40‰ at 278 K. The apparent δ15N(NOx) value can be increased by a maximum of 0.74 ‰ compared to the true value by the difference in the collection efficiency of NO and NO2 when the mole fraction of NO2 to NOx is 0.49.

The KMnO4/NaOH recovery solution captured approximately 83–85% of NO and 91% of NO2. Given that the mole fraction of NO2 is much larger than NO, the collection efficiency of NOx in KMnO4/NaOH recovery solution was in good agreement with the reported value (92 ± 10) % [14, 15]. In this study, both recovery solutions can collect NOx with high efficiency, and H2O2/NaOH has a higher absorption efficiency than KMnO4/NaOH under the same conditions. However, we have not evaluated the extent to which recovery solution affects isotope ratio fractionation.

Nitrite and nitrate concentration in the recovery solution

Blank reduction is important in NOx isotope measurements. We compared the NO2− and NO3− blanks in H2O2/NaOH with those in the KMnO4/NaOH recovery solution. The NO2− and NO3− concentrations of the blanks were measured using CFA, and the results are listed in Table 2. The blanks of the H2O2/NaOH and KMnO4/NaOH recovery solution were approximately 0.5 μM and 2.8 μM, respectively. The blank concentrations of NO2− and NO3− in the H2O2/NaOH recovery solution were clearly lower than the values of KMnO4/NaOH, indicating that the H2O2/NaOH recovery solution is superior in blank suppression. Fibiger et al. [14] also tried NOx collection using an H2O2/NaOH recovery solution and found a high nitrate blank (approximately 25 μM). However, such high-level blanks in the H2O2/NaOH recovery solution were not observed when using the reagents or experimental scheme presented in this study.

thumbnail
Download:
Table 2. Nitrite and nitrate concentrations of the blank and seven days later after adding ammonium ion in recovery solution.

https://doi.org/10.1371/journal.pone.0298539.t002

In addition, 120 μM NH4Cl was added to KMnO4/NaOH and H2O2/NaOH recovery solutions and allowed to stand for one week to investigate the effect of ammonium ions. Table 2 presents the results of the study. Total NO2− and NO3− concentration increased in the KMnO4/NaOH recovery solution but not in the H2O2/NaOH recovery solution. As described by Fibiger et al. [14] and Wojtal et al. [15], the KMnO4/NaOH recovery solution slightly reacted with ammonia ions after seven days, but the H2O2/NaOH recovery solution did not react with ammonia ions. However, since neither of the recovery solutions reacted by even 1% of the amount of ammonia added, it is hypothesized that a negligible reaction occurred with ammonia. From these results, it is evident that the H2O2/NaOH recovery solution is superior to the KMnO4/NaOH recovery solution in terms of NOx collection. Subsequent experiments were performed using the H2O2/NaOH recovery solution only and had a high collection rate and suppressed blanks. Under the same experimental conditions, the H2O2/NaOH recovery solution outperforms the KMnO4/NaOH recovery solution. A further advantage over previous methods is that neutralization can be performed in one vessel, and no centrifugation is required, resulting in a reduced risk of sample loss and contamination.

Nitrogen isotope measurement for cylinder NOx

The δ15N values of samples A and B recovered with the H2O2/NaOH solution were measured, as shown in Table 3. The repeatability of δ15N values for samples A and B were (0.7 ± 0.5) ‰ (n = 5) and (−18.0 ± 0.8) ‰ (n = 5) for 15–20 ppb of NOx, respectively. The repeatability of δ15N values for samples A and B were (0.8 ± 0.5) ‰ (n = 8) and (−17.7 ± 0.7) ‰ (n = 8) for 30–40 ppb of NOx, respectively. The larger uncertainty of the δ15N values for sample B compared to those of sample A is thought to be due to slight contamination from ambient air.

thumbnail
Download:
Table 3. Reproducibility of δ15N value for NO and NO2 in the cylinders using H2O2/NaOH recovery solution.

https://doi.org/10.1371/journal.pone.0298539.t003

Quantification of the influence of Δ17O on δ15N values

The δ15N value of N2O measured by IRMS is calculated assuming Δ17O (= δ17O – 0.52 δ18O) is 0‰. However, since the Δ17O value of N2O may apparently increase the δ15N value, a correction was performed taking the Δ17O value into account. In this study, the Δ17O value of NO3− was not measured. On the other hand, the maximum δ18O value of N2O converted from NOx was 40‰ (containing laboratory and field experiments). Note that the δ18O value of N2O does not directly reflect the δ18O value of NOx because NOx obtains oxygen derived from water or H2O2 during oxidation in the recovery solution. The Δ17O values are generated only by mass transfer of O atoms from ozone to products during oxidation reactions [30]. In Eqs 1–6, the Δ17O of O supplied during the process of NO and NO2 oxidizing NO3 is 0‰. Considering that NO and NO2 each receive two oxygen atoms, the Δ17O values become 2/3 or less (Eqs 1–6). Further, it is assumed that the relationship between the Δ17O and δ18O values of NO2 is 0.36:1 (estimated by a straight line passing through the origin based on the values reported by Albertin et al. [12]). At this time, the maximum Δ17O value of N2O was 9.5‰ when the δ18O value was 40‰. The effect of this Δ17O value on the δ15N value was estimated using USGS34 and USGS35 with known Δ17O values. Based on the result of USGS35 measurements, when the Δ17O value is 21.56‰ [28], the apparent δ15N value would increase by 1.2‰. This result showed good agreement with the description by Yu and Elliott [16]. Assuming a linear relationship between Δ17O and δ15N values of USGS34 (Δ17O value: −0.3‰ [31]) and USGS35, the Δ17O value of 9.5‰ will increase the δ15N value by 0.5‰ at maximum. The overall measured 1σ uncertainty of δ15N(NOx) was ± 1.2‰ by combining the difference in the absorption efficiency of NO and NO2, the repeatability, and the consideration of Δ17O value.

Limitations of NOx collection

We also tested the limitations of this developed method. Possible factors that reduce the yield of NOx include a decrease in oxidant concentration due to reaction with NOx and other gases, and a decrease in pH due to reaction with CO2. The reaction between NOx and the oxidant agent is not rate-limiting as the input H2O2 concentration is sufficiently high compared to the NOx concentration. In fact, we tried flowing 40 ppb of NO and NO2 for over 12 h each, but the collection efficiency did not fall below 90%. On the other hand, when the time required for the collection rate to drop below 90% was measured for continuous collection of approximately 15 ppb of NOx in air at an average of 0.6 L min−1, the collection rate dropped sharply at 14 h (S3 Fig). CO2 dissolves in the form of CO32- at high pH as follows: (9) Since it is difficult to calculate the dissolution rate of CO2 in this study, we assumed that all CO2 dissolves in solution. In addition, although OH− ions are used or provided in the decomposition of H2O2, this was not accounted for. We set the flow rate at 0.6 L min−1 and the CO2 concentration at 400 ppm. Given that the NO collection efficiency decreases under pH 11 [21] and that the two containers were sufficient to collect NOx, the combined NaOH from the two containers would be neutralized in about 8 h. The actual capacity was longer than 8 h, as it is unknown whether all the CO2 will be absorbed and whether the first bubbler can continue to collect CO2 even after neutralization. When collecting NOx from air, we recommend up to 8 and 6 h for the collection time, at a flow rate of 0.6 and 0.8 L min-1, respectively. If used in an environment with high CO2 concentration, the flow rate must be reduced, or the number of bubblers must be increased.

Comparison with the previous NOx isotope measurement method

The advantage of this method is that both NO and NO2 can be collected with a low blank of NO3−; thus, the δ15N values of atmospheric NOx can be directly estimated compared to the methods that can only collect NO2. An offline method for converting high concentrations (over 100 ppm) of NOx to NO3− using H2O2/NaOH recovery solution was used by Heaton [32]. However, the concentration of H2O2 and NaOH in the recovery solution is unknown. This study is the first to investigate the NOx collection efficiency, the degree of blanking, and the influence of ammonium ions being quantified using H2O2/NaOH recovery solution. Another offline wet method for collecting high concentrations of NOx is the use of H2SO4/H2O2. However, Chin et al. [33] showed that the 6 wt% H2O2 in low pH (2 to 4) converted less than 5% of NO in the flue gas, and Ohta et al. [18] showed that the NO2 collection efficiency using H2O2/NaOH recovery solution can be degraded at NaOH concentrations below 0.24 M, suggesting that probably the H2O2 in basic solution is necessary for high collection efficiency of NOx online. Further advantages of this method are its portability, simplicity, and the ability to collect the required amount of sample to analyze the nitrogen isotopic composition of ambient NOx in a short period of time.

Nitrogen isotope measurement for atmospheric NOx

The observed δ15N values and NOx concentrations for atmospheric NOx collected at the Tsukuba and Yoyogi sites are shown in Fig 2 and S2 Table. The average NOx concentrations during the sampling period were 6 and 18 ppb at Tsukuba and Yoyogi, respectively. The maximum NOx concentrations at Tsukuba and Yoyogi were 45 and 143 ppb, respectively. The diurnal variation in the NOx concentration on weekdays during the sampling period clearly showed two peaks corresponding to traffic rush hours (S4 and S5 Figs).

thumbnail
Download:
Fig 2.

NO (red) and NOx (black) concentrations and δ15N(NOx) values at the Tsukuba (a, b) and Yoyogi site (c). Error bars indicate 1σ uncertainty of δ15N(NOx) ± 1.2‰.

https://doi.org/10.1371/journal.pone.0298539.g002

The δ15N(NOx) value ranged from −3.1 to −0.5 ‰ at the Tsukuba site and from −5.6 to −0.5 ‰ at the Yoyogi site. The averaged δ15N(NOx) value and standard deviation (1σ) in the Yoyogi site was (−2.7 ± 1.8) ‰ and in the Tsukuba site was (−1.7 ± 0.9) ‰ during the sampling period, and no significant difference between the two sites was observed. No significant correlation was found between the NOx concentrations or 1/[NOx] and δ15N(NOx) values. Additionally, it also did not correlate with the ratio of NO2 to NOx observable when only the δ15N(NO2) was measured because we collected both NO and NO2.

Walters et al. [34] showed the mass-weighted δ15N(NOx) values emitted from vehicles have the following relationship with vehicle runtime: (10) where δ15N(NOx) represents the mass-weighted δ15N(NOx) values emitted from vehicles and t is the vehicle run time (min). Because the average distance of one car in Japan is approximately 20 km (Ministry of Land, Infrastructure, Transport and Tourism website, https://www.e-stat.go.jp/stat-search/files?stat_infid=000032211818, The first summary table by fuel/vehicle type last access: 28 March 2023 [35]), the driving time is expected to be 10−30 mins each. The predicted mass-weighted δ15N(NOx) values emitted from vehicles (−5.3 to −2.0 ‰; S1 Table) based on Eq 10 were matched with the δ15N(NOx) values in the Tsukuba and the Yoyogi sites. Therefore, if the main source of NOx is vehicle exhaust, then the δ15N(NOx) values and diurnal variations in NOx concentrations can be explained. Biomass burning (−7 to 12 ‰ [36]; S1 Table) is also a candidate for the NOx sources of observed δ15N(NOx) values in the Tsukuba and the Yoyogi sites. However, because the NOx emitted from biomass burning is temporary, it is unlikely to be the main source of NOx in urban areas. Additionally, biomass burning did not show diurnal variation in NOx concentrations, as shown in S4 and S5 Figs. This indicates that the main NOx source collected at the Tsukuba and Yoyogi sites was vehicle exhaust during the sampling period. However, the sampling period is limited and not all variations in δ15N(NOx) value can be explained. Future investigations are needed to understand NOx dynamics by measuring the δ15N(NOx) value of NOx sources to enrich the database in the surrounding environment and through long-term observations with higher time resolution.

Conclusion

We developed a portable new method to collect NOx for nitrogen isotopic measurement by mixing 3 wt% H2O2 and 0.5 M NaOH solution with a precision (1σ uncertainty) of ± 1.2 ‰. The method using the developed H2O2/NaOH recovery solution has high NOx collection efficiency, a relatively simpler measurement procedure, small blanks, and a negligible impact of ammonium contamination.

The δ15N(NOx) values were observed in two sites in Japan. The averaged δ15N(NOx) value and standard deviation (1σ) in the Yoyogi site was (−2.7 ± 1.8) ‰ and in the Tsukuba site was (−1.7 ± 0.9) ‰ during the sampling period. The main NOx source appears to be the vehicle exhaust in the two sites. However, the sampling period is limited and the not all variations in δ15N(NOx) value can be explained. Future investigations are needed to understand NOx dynamics by measuring the δ15N(NOx) value of NOx sources to enrich the database in the surrounding environment and through long-term observations with higher time resolution. In addition, by combining the concentration and δ15N value of ammonia, organic nitrogen, and nitrate in aerosols, among others, the understanding of the nitrogen cycle, including NOx will be deepened.

Supporting information

S1 Fig. Aerial view of the sampling sites.

https://doi.org/10.1371/journal.pone.0298539.s001

(PDF)

S2 Fig. Schematic diagram of the system for measuring nitrogen isotope ratios within N2O.

https://doi.org/10.1371/journal.pone.0298539.s002

(PDF)

S3 Fig. Limitation of NOx collection.

A NOx analyzer was connected behind the bubbler and when approximately 15 ppb of NOx in the air was continuously captured at an average rate of 0.6 L min-1, the time required for the collection efficiency to drop below 90% was measured. We set the bubbler at time 0 min.

https://doi.org/10.1371/journal.pone.0298539.s003

(PDF)

S4 Fig. Box-and-whisker plots of diurnal variation of NOx concentrations during sampling periods (1/4-6, 10–11, 31, 2/1-3, 7–10) at the Tsukuba site.

https://doi.org/10.1371/journal.pone.0298539.s004

(PDF)

S5 Fig. Box-and-whisker plots of diurnal variation of NOx concentrations during sampling periods (1/17-20, 24–27, 30–31) at the Yoyogi site.

https://doi.org/10.1371/journal.pone.0298539.s005

(PDF)

S1 Table. Overview of nitrogen isotopic composition (δ15N) for nitrogen oxides (NOx) in the atmosphere.

https://doi.org/10.1371/journal.pone.0298539.s006

(XLSX)

S2 Table. The δ15N(NOx) and NOx concentration of samples collected in ambient urban air at Tsukuba and Yoyogi site from January to February 2023.

https://doi.org/10.1371/journal.pone.0298539.s007

(XLSX)

Acknowledgments

We thank Tokai University for providing the observation sites. We thank Taizo Sano at the National Institute of Advanced Industrial Science and Technology (AIST), Japan, for allowing the use of the device. We also thank Shohei Hattori at the Tokyo Institute of Technology, Japan (current address: International Center for Isotope Effects Research Nanjing University, China) for the help with the isotopic composition measurement method for nitrate using the denitrifier method.

References

  1. 1. Galloway JN, Townsend AR, Erisman JW, Bekunda M, Cai Z, Freney JR. et al. Transformation of the Nitrogen Cycle: Recent Trends, Questions, and Potential Solutions. Science. 2008; 320: 889–892. pmid:18487183
  2. 2. Uno I, Osada K, Yumimoto K, Wang Z, Itahashi S, Pan X, et al. Importance of Long-Range Nitrate Transport Based on Long-Term Observation and Modeling of Dust and Pollutants over East Asia. Aerosol Air Qual. Res. 2017; 17: 3052–3064. https://doi.org/10.4209/aaqr.2016.11.0494.
  3. 3. Seinfeld JH, Pandis SN. Atmospheric Chemistry and Physics, A Wiley-Inter Science Publication, John Wiley & Sons Inc., Hoboken, New Jersey. 2006 ISBN 978-0-471-72018-8.
  4. 4. McDuffie EE, Smith SJ, O’Rourke P, Tibrewal K, Venkataraman C, Marais EA, et al. A global anthropogenic emission inventory of atmospheric pollutants from sector- and fuel-specific sources (1970–2017): an application of the Community Emissions Data System (CEDS). Earth Syst. Sci. Data, 2020; 12: 3413–3442. https://doi.org/10.5194/essd-12-3413-2020.
  5. 5. Elliott EM, Kendall C, Boyer EW, Burns DA, Lear GG, Golden HE. et al. Dual nitrate isotopes in dry deposition: Utility for partitioning NOx source contributions to landscape nitrogen deposition. J. Geophys. Res. Biogeosci. 2009;114: 1–15. https://doi.org/10.1029/2008JG000889.
  6. 6. Morin S, Savarino J, Frey MM, Domine F, Jacobi HW, Kaleschke L, et al. Comprehensive isotopic composition of atmospheric nitrate in the Atlantic Ocean boundary layer from 65°S to 79°N. J. Geophys. Res. Atmos. 2009; 114: 1–19. https://doi.org/10.1029/2008JD010696.
  7. 7. Kamezaki K, Hattori S, Iwamoto Y, Ishino S, Furutani H, Miki Y, et al. Tracing the sources and formation pathways of atmospheric particulate nitrate over the Pacific Ocean using stable isotopes. Atmos. Environ. 2019; 209: 152–166. https://doi.org/10.1016/j.atmosenv.2019.04.026.
  8. 8. Ammann M, Siegwolf R, Pichlmayer F, Suter M, Saurer M, Brunold C. Estimating the uptake of traffic-derived NO2 from 15N abundance in Norway spruce needles. Oecologia. 1999; 118: 124−131. https://doi.org/10.1007/s004420050710.
  9. 9. Savard MM, Cole A, Smirnoff A, Vet R. δ15N values of atmospheric N species simultaneously collected using sector-based samplers distant from sources–Isotopic inheritance and fractionation. Atmos. Environ. 2017; 162: 11–22. https://doi.org/10.1016/j.atmosenv.2017.05.010.
  10. 10. Walters WW, Fang H, Michalski G. Summertime diurnal variations in the isotopic composition of atmospheric nitrogen dioxide at a small midwestern United States city. Atmos. Environ. 2018; 179: 1–11. https://doi.org/10.1016/j.atmosenv.2018.01.047.
  11. 11. Kawashima H. Seasonal trends of the stable nitrogen isotope ratio in particulate nitrogen compounds and their gaseous precursors in Akita, Japan. Tellus B. 2019: 71: 1627846. https://doi.org/10.1080/16000889.2019.1627846.
  12. 12. Albertin S, Savarino J, Bekki S, Barbero A, Caillon N. Measurement report: Nitrogen isotopes (δ15N) and first quantification of oxygen isotope anomalies (Δ17O, δ18O) in atmospheric nitrogen dioxide. Atmos. Chem. Phys. 2021: 21: 10477–10497, https://doi.org/10.5194/acp-21-10477-2021.
  13. 13. Blum DE, Walters WW, Eris G, Takeuchi M, Huey LG, Tanner D, et al. Collection of Nitrogen Dioxide for Nitrogen and Oxygen Isotope Determination Laboratory and Environmental Chamber Evaluation. Anal. Chem. 2023; 95: 3371–3378.
  14. 14. Fibiger DL, Hastings MG, Lew AF, Peltier RE. Collection of NO and NO2 for Isotopic Analysis of NOx Emissions. Anal. Chem. 2014; 86(24): 12115–12121. https://pubs.acs.org/doi/full/10.1021/ac502968e.
  15. 15. Wojtal PK, Miller DJ, O’Conner M, Clark SC, Hastings MG. Automated, High-resolution Mobile Collection System for the Nitrogen Isotopic Analysis of NOx, J Vis Exp. 2016; 20(118): 54962. pmid:28060293
  16. 16. Yu Z, Elliott EM. Novel Method for Nitrogen Isotopic Analysis of Soil-Emitted Nitric Oxide. Environmental Science & Technology. 2017; 51(11): 6268–6278. https://pubs.acs.org/doi/10.1021/acs.est.7b00592.
  17. 17. Freyer HD, Kley D, Volz-Thomas A, Kobel K. On the interaction of isotopic exchange processes with photochemical reactions in atmospheric oxides of nitrogen. J. Geophys. Res. 1993; 98: 14791. https://doi.org/10.1029/93JD00874.
  18. 18. Ohta H, Sugimoto K, Yoshida T. A method for the removal of Nitrogen Oxides (NOx) in ambient air (1) -Absorption of NOx by sodium peroxide solution. J. Environ. Conserv. Eng. 1985; 14(9): 690–694 (in Japanese). https://doi.org/10.5956/jriet.14.690.
  19. 19. Kuropka J. Removal of nitrogen oxides from flue gases in a packed column. Environ. Prot. Eng. 2011; 37(1): 13–22.
  20. 20. Ho MC, Ong VZ, Wu TY, Potential use of alkaline hydrogen peroxide in lignocellulosic biomass pretreatment and valorization–A review. Renewable and Sustainable Energy Reviews. 2019; 112: 75–86, https://doi.org/10.1016/j.rser.2019.04.082.
  21. 21. Sun S, Zhang J, Sheng C, Zhong H, The removal of NO from flue gas by NaOH-catalyzed H2O2 system: Mechanism exploration and primary experiment, Journal of Hazardous Materials, 2022; 440: 129788, https://doi.org/10.1016/j.jhazmat.2022.129788.
  22. 22. Suzuki K, Niimi T, Yamamoto N, Shibata M, Saeki M, Ono A, et al. Rapid photometric method for the determination of the mass concentration of nitrogen monoxide and nitrogen dioxide. Analytica Chimica Acta. 1994; 295: 135–141. https://doi.org/10.1016/0003-2670(94)80343-9.
  23. 23. Špalek O, Balej J, Paseka I. Kinetics of the decomposition of hydrogen peroxide in alkaline solutions. J. Chem. Soc., Faraday Trans. 1982; 78: 2349–2359. https://doi.org/10.1039/F19827802349.
  24. 24. Jiang SP, Ashton WR, Tseung ACC, An observation of homogeneous and heterogeneous catalysis processes in the decomposition of H2O2 over MnO2 and Mn(OH)2. Journal of Catalysis. 1991; 131(1): 88–93, https://doi.org/10.1016/0021-9517(91)90325-X.
  25. 25. Sigman DM, Casciotti KL, Andreani M, Barford C, Galanter M, Böhlke JK. A bacterial method for the nitrogen isotopic analysis of nitrate in seawater and freshwater. Anal. Chem. 2001; 73: 4145–4153. pmid:11569803
  26. 26. Casciotti KL, Sigman DM, Hastings MG, Böhlke JK, Hilkert A. Measurement of the oxygen isotopic composition of nitrate in seawater and freshwater using the denitrifier method. Anal Chem. 2002; 74(19): 4905–4912. pmid:12380811
  27. 27. Hattori S, Savarino J, Kamezaki K, Ishino S, Dyckmans J, Fujinawa T. et al. Automated system measuring triple oxygen and nitrogen isotope ratios in nitrate using the bacterial method and N2O decomposition by microwave discharge. Rapid Commun. Mass Spectrom. 2016; 30: 2635–2644. https://doi.org/10.1002/rcm.7747.
  28. 28. Michalski G, Savarino J, Böhlke JK, Thiemens M. Determination of the Total Oxygen Isotopic Composition of Nitrate and the Calibration of a Δ17Ο Nitrate Reference Material. Anal. Chem. 2002; 74 (19): 4989–4993. https://pubs.acs.org/doi/10.1021/ac0256282.
  29. 29. Walters WW, Simonini DS, Michalski G. Nitrogen isotope exchange between NO and NO2 and its implications for δ15N variations in tropospheric. Geophys. Res. Lett. 2015; 440–448. https://doi.org/10.1002/2015GL066438.
  30. 30. Michalski G, Scott Z, Kabiling M,Thiemens MH. First measurements and modeling of Δ17O in atmospheric nitrate, Geophys. Res. Lett. 2003; 30: 1870, , 16.
  31. 31. Böhlke JK, Mroczkowski SJ, Coplen TB. Oxygen isotopes in nitrate: New reference materials for 18O:17O:16O measurements and observations on nitrate-water equilibration. Rapid Communications in Mass Spectrometry. 2003; 17: 1835−1846. pmid:12876683
  32. 32. Heaton THE. 15N/14N ratios of NOx from vehicle engines and coal-fired power stations. Tellus, Ser. B Chem. Phys. Meteorol. 1990: 42 B: 304–307. https://doi.org/10.1034/j.1600-0889.1990.00007.x-i1.
  33. 33. Chin T, Tam IC, Yin CY. Comparison of various chemical compounds for the removal of SO2 and NOx with wet scrubbing for marine diesel engines. Environ Sci Pollut Res. 2022: 29: 8873–8891. https://doi.org/10.1007/s11356-021-16155-9.
  34. 34. Walters WW, Goodwin SR, Michalski G. Nitrogen stable isotope composition (δ15N) of vehicle-emitted NOx. Environ. Sci. Technol. 2015; 49: 2278–2285. https://doi.org/10.1021/es505580v.
  35. 35. Ministry of Land, Infrastructure, Transport and Tourism website, https://www.e-stat.go.jp/stat-search/files?stat_infid=000032211818, The first summary table by fuel/vehicle type last access: 28 March 2023.
    • 36. Fibiger DL, Hastings MG. First Measurements of the Nitrogen Isotopic Composition of NOx from Biomass Burning. Environ. Sci. Technol. 2016: 50: 11569–11574. https://doi.org/10.1021/acs.est.6b03510.