Aqueous chemical bleaching of 4-nitrophenol brown carbon by hydroxyl radicals; products, mechanism and light absorptivity

The reaction of hydroxyl radicals (OH) with 4-nitrophenol (4-NP) in the aqueous solution was investigated at pH=2 and 9. As a result, the molar yield of the phenolic products was measured to be 0.20 ±0.05 at pH=2 and 0.40±0.1 at pH=9. The yield of 4-nitrocatechol (4-NC) was higher at pH=9; at the same time, a lower number of phenolic products was observed due to the hydrolysis and other irreversible reactions at pH>7. Mineralization 10 investigated with total organic carbon (TOC) technique showed that after 4-NP was completely consumed approx. 85% of the organic carbon remained in the aqueous solution. Hence, up to 65% of the organic carbon that remained in the aqueous solution accounted for the open-ring non-phenolic products. The light absorptivity of the reaction solution between 250 and 600 nm decreased as a result of OH reaction with 4-NP. At the same time, 4-NP solution showed some resistance to chemical bleaching due to the formation of the 15 light-absorbing by-products. This phenomenon effectively prolongs the time-scale of chemical bleaching or 4-NP via reaction with OH by a factor of 3-1.5 at pH 2 and 9, respectively. The experimental data acquired indicated that both photolysis and reaction with OH can be important removal processes of the atmospheric brown-carbon from the aqueous particles containing 4-NP. 20 https://doi.org/10.5194/acp-2021-871 Preprint. Discussion started: 28 October 2021 c © Author(s) 2021. CC BY 4.0 License.


Introduction
Atmospheric brown carbon (BrC) is a subfraction of organic aerosols (OA) that is characterized by strong, wavelength-dependent absorption of the electromagnetic irradiation in the near ultraviolet (UV) and visible (VIS) regions Yan et al., 2018). BrC is primarily produced by biomass burning (BB) and has a negative impact on the local air quality and human health Yan et al., 2018). Due to the high 25 UV-Vis absorptivity, BrC greatly contributes (up to 50%) to the radiative forcing of OA (Cordell et al., 2016;Zhang et al., 2017;Lu et al., 2015;Wang et al., 2014;Feng et al., 2013;Yan et al., 2018). Numerous organic compounds contribute to the atmospheric BrC Hettiyadura et al., 2021;Li et al., 2020a;Fleming et al., 2020); at the same time, a significant fraction of BrC chromophores remains poorly characterized (Bluvshtein et al., 2017;Laskin et al., 2015). 30 Nitrophenols are widespread nitroaromatic compounds that been identified among the major chromophores of atmospheric BrC (Harrison et al., 2005b;Laskin et al., 2015;Bluvshtein et al., 2017). 4-Nitrophenol (4-NP) is one of the most atmospherically abundant and environmentally widespread nitrophenols (Harrison et al., 2005b;Laskin et al., 2015) and is characterized by very high absorption cross-sections in the UV-Vis region (Jacobson, 1999).
The goal of this work was to investigate mechanism of OH reaction with 4-NP in the aqueous phase in context of atmospheric BrC formation and processing. Hence, the reaction (1) was investigated at 298 K in the aqueous phase under acidic (pH=2) and basic (pH=9) conditions using the photoreactor developed in our laboratory: 80 Additionally, the phenolic products of reaction (1) were analyzed together with the changes in the UV-Vis absorptivity of the reaction solution (Witkowski et al., 2019). Phenols under investigation were quantified using https://doi.org/10.5194/acp-2021-871 Preprint. Discussion started: 28 October 2021 c Author(s) 2021. CC BY 4.0 License.
gas chromatography coupled to mass spectrometry (GC/MS). A possible mineralization and formation of volatile products was monitored with the total organic carbon (TOC) analyzer. The UV-Vis absorptivity of the reaction solution as well as the molar absorptivity (base-e ε, mol -1 ×L×cm -3 ) of the phenols under investigation were 85 measured between pH 2 and 9.

Experimental section
Materials and reagents used are listed in section S1 of the electronic Supplementary Information (SI).

Aqueous phase photoreactor
The aqueous phase photoreactor was described previously (Witkowski et al., 2019), and more details are provided 90 in section S4.1. The reaction vessel was quartz jacketed reaction flask with 100 ml internal volume. All experiments were carried out at 298 K; the temperature of the reaction solution was maintained with a circulating water bath (SC100-A10, Thermo Fisher Scientific). Two lamps (TUV TL 4W, peak emission 254 nm, Philips) were used to irradiate the solution.

Experimental procedure 95
The reaction mixture was a 100 ml solution of 4-NP (concentration 100-250 µM) in deionized (DI) water. The pH of the solution was unbuffered or it was adjusted to pH 2 or 9 using HCl, HClO4 and Na2HPO4 (50 mM) to investigate oxidation of fully protonated and deprotonated 4-NP (section S2). Hydrogen peroxide (concentration 5 mM) was photolyzed with UV irradiation (254nm) to generate OH, the estimated steady-state concentration of OH was 1.4 × 10 -9 M (section S3) (Tan et al., 2009). Under these conditions 4-NP was almost completely consumed 100 by OH within 1h. Aliquots of the reaction mixture were sampled every 5 min and analyzed by GC/MS, UV-Vis spectroscopy and TOC analyzer. The experimental procedure is described in detail in section S4.1.

Gas chromatography coupled with mass spectrometry
Analyses were carried out using GC-MS-QP2010Ultra gas chromatograph (Shimadzu) coupled with the quadrupole QP-5000 mass spectrometer (Shimadzu), the instrument was equipped with AOC-5000 autosampler. 105 Analytes were separated using capillary column ZB-5MSPlus (Phenomenex). The mass spectrometer was equipped with the electron ionization source (EI, 70 eV) and was operating in the selected ion monitoring (SIM) mode. For quantitative analyses GC/MS was calibrated with the standard solutions of 4-NP, HH, 1,2,4-THB, 4-NC that were identified as products of 4-NP reaction with OH. 2-Nitrophloroglucinol was used as a surrogate https://doi.org/10.5194/acp-2021-871 Preprint. Discussion started: 28 October 2021 c Author(s) 2021. CC BY 4.0 License. standard for quantification of 4-nitropyrogallol (4-NPG) and 5-nitropyrogallol (5-NPG) identified among the 110 reaction products. Phloroglucinol was not identified among the product of reaction (1) (Xiong et al., 2015;Zhao et al., 2013) hence, it was used as an internal standard (IS). Phenols were derivatized with acetic anhydrite (AA) and analyzed via GC/MS (Regueiro et al., 2009). Detailed description of the analytical procedure is provided in section S4.2.

UV-Vis spectrophotometry 115
UV-Vis measurements were carried out with i8 dual-beam spectrophotometer (Envisense) in 4 ml cuvettes with a 1 cm absorption pathlength. The absorbance of each aliquot of the reaction solution was measured between wavelengths 230 and 600 nm. The pH of each sample taken from the reactor was adjusted between 2 and 9 (by intervals of 1, see section S4.3) by adding a small amount of NaOH or H3PO4 solution in DI water. In a separate set of experiments, the wavelength-dependent absorption cross sections, ε, (base-e; mol -1 × L × cm -1 ) were 120 measured for 4-NP, HH, 1,2,4-BT, 4-NC and 2-NPG between are reported in appendix 1 (Fig. S5).

Total organic carbon analysis
Non-purgeable organic carbon (NPOC) was quantify with TOC-5050A analyzer (Shimadzu) connected to the ASI-5000A autosampler (Shimadzu). The 1.5 ml of the reaction solution was diluted with the 1.5 ml of DI water. Then, 50 µl of 2M HCl was added via the TOC autosampler and each sample was sparged with oxygen for 2 min before 125 injection. The injection volume was 20 µl and each sample was injected into the instrument three times. The TOC analyzer was calibrated with the standard solutions of 4-NP in DI water with concentrations between 3 and 35 mgTOC × L -1 ; the squared linear coefficient of determination for calibration curve (R 2 )=0.9995 was obtained.

Light absorptivity and atmospheric lifetimes
Production of light-absorbing compounds following reaction (I) was evaluated via eq. I. law using the ε measured in this work between pH 2 and 9. The expression described using eq. I follows the 135 The atmospheric lifetimes of BrC was evaluated by deriving the empirical kbleaching rate coefficients (M -1 s -1 )eq.
In eq. (II) the kOH is the bimolecular reaction rate coefficient (M -1 s -1 ) for the reaction of 4-NP or 4-nitrophenolate 140 with OH (Biswal et al., 2013;Garcıá Einschlag et al., 2003), kA, kA.rmix are the first-order disappearance rate (min -1 ) of the integrated absorbance peak for the 4-NP and for the reaction mixture, respectively (Fig. S9). The kA.rmix showed little dependence on pH at which the absorbance was measured, thus average values were used.

Control experiments and uncertainty
The stability of phenols under investigation in the presence of H2O2 and UV-Vis irradiation only was studied by 155 carrying control experiments (section S6). Also, for the experiments at pH=2 HCl or HClO4 was used to confirm that the buffering agent used did not affected the distribution of detected products. Control experiments revealed that all phenols under investigation were stable at pH≤7, within the time-scale of the experiments, but 1,2,4-THB, 4-NPG, 5-NPG and HH underwent irreversible reactions at pH>7.
Experimental uncertainties are reported as 2σ from triplicate measurements, other uncertainties were calculated 160 with the exact differential method, unless noted otherwise.

Products and reaction mechanism
As presented in Fig. S4, HH, 1,2,4-THB, 4-NC and 5-NPG were formed following reaction (1) under acidic pH conditions, which is in a good agreement with the previously published results (Xiong et al., 2015;Oturan et al., 165 2000;Tauber et al., 2000;Liu et al., 2010;Du et al., 2017;Chen et al., 2018). 4-Nitroresorcinol (4-NR) was also tentatively identified among the products. Isomers of 4-NC were previously reported as products of 4-NP oxidation by OH but the exact structures were not proposed for these compounds (Zhao et al., 2013).
To our knowledge, this work is first to report the formation of 4-NPG from reaction (1). (Xiong et al., 2015) Previously, detection of isomeric products (4-NR and 4-NPG) might have been difficult due to lack of standards 170 and absence of the MS detector (Tauber et al., 2000;Liu et al., 2010;Daneshvar et al., 2007). Also, insufficient resolving power of HPLC used to investigate the composition of products likely contributed to the fact that the formation of 4-NR and 4-NPG was not previously observed (Oturan et al., 2000;Tauber et al., 2000;Liu et al., 2010;Daneshvar et al., 2007;Lipczynska-Kochany, 1991).
It was previously proposed that 1,2,4-THB is formed by the addition of OH in para position of 4-NC (ipso attack) followed by 210 elimination of 2 • (Daneshvar et al., 2007;Zhang et al., 2003;Kavitha and Palanivelu, 2005), but this assumption was never confirmed experimentally. It is still unclear whether or not the 1,4-hydroxy-4-nitrocyclohexadienyl type radicals primarily eliminate 2 • to yield stable phenolic products or decompose to phenoxyl radicals and 2 − (O'neill et al., 1978;Kotronarou et al., 1991;Xiong et al., 2015;Di Paola et al., 2003;Liu et al., 2010). The limited literature data available indicate that the elimination of 2 − is more likely (O'neill et al., 1978;Kotronarou et al., 1991), which would also explain the fast decrease in 215 pH observed during OH+4-NC reaction and the absence of 1,2,4-THB among the major products (Fig. S6). The trace amounts of 1,2,4-THB formed from OH+4-NC reaction are likely due to minor disproportionation reaction of the two phenoxy radicals (Liu et al., 2010). It was previously reported that 4-NC was quantitatively converted into 1,2,4-THB in the absence of O2, which effectively promoted the disproportionation reaction between two 1,2-hydroxyl-4-phenoxyl radicals (Gonzalez et al., 2004;Liu et al., 2010;Di Paola et al., 2003). These results are in a good agreement with the experimental data acquired in this 220 work.
In a concentrated (160 mM) basic solutions, 1,2,4-THB was shown to generate stable aromatic oligomers with the absorbance between 400-700 nm (Randolph et al., 2018). Detecting such oligomers with GC/MS is unlikely due to their lower volatility, insufficient thermal stability or low reactivity towards AA (section 2.3). However, at pH>7 the integrated absorbance of the reaction solution in this spectral range is lower as compared with the acidic solution, as discussed in more detail in section 3.2. 245 Hence, the formation of "brown" phenolic oligomers from 1,2,4-THB is evidently suppressed in a more diluted solution and in the presence of nitrated phenols, NO2and NO3ions.

Light absorptivity and the time-evolution of brown-carbon chromophores
A continuous decrease of the absorbance of the reaction solution was observed (section S8). On the other hand, an initial, small increase in the absorbance at 420 nm of the 4-NP solution during reaction with OH was previously reported followed by a 250 rapid bleaching (Zhao et al., 2015); such differences can be caused by slightly different reaction conditions. Also, in this work, an integrated absorbance values were used (eq. I and II) which may be a more adequate approach due to shifting of the absorption maximum (Amax) of the reaction solution ( Fig. S8) (Zhao et al., 2015;Hems and Abbatt, 2018).
The contribution of the light-absorbing products of reaction (1) to the overall light absorptivity of the reaction solution was evaluated via eq. I. Results presented in Fig. S10 show that when the reaction was carried out under basic pH conditions the 255 relative absorbance of products (eq. I) was lower and increased slowly. This points out that the light-absorbing compounds are not stable at pH>7 which is in a good agreement with the results discussed in section 3.1 (Randolph et al., 2018).
The values of MACinit and Rabs calculated with eq. III and IV are presented in Fig. 3.  As expected, MACint decreased steadily following the oxidation of the precursor. The experimental data acquired (Fig. 3)  265 show a clear increase in absorptivity following the increase in pH at which the absorbance was measured due to pH-depended ε values for the light-absorbing phenols present in the reaction solution. In Fig. 3A and B, the disappearance rates of MACint are of similar order (Table S3). This is most likely due to the formation of higher number of light-absorbing phenols (secondgeneration products) at pH=2 and due to higher yield of 4-NC (which is characterized by high ε values, Fig. S5) at pH=9, respectively. Consequently, the rates of disappearance of the overall light absorptivity of 4-NC following reaction (1) are 270 mostly independent on pH of reaction solution and primarily depend on pH at which the absorbance is measured. The MACInt values calculated were slightly higher as compared with the values measured for the previously investigated aromatic BrC chromophores (for non-nitrated precursors) (Jiang et al., 2019;Jiang et al., 2021), likely due to the high ε values of 4-NP and nitrated phenols formed following reaction (1). The Rabs values ( Fig. 3C and D) decrease more slowly as compared with the values of MACInt, which is caused by a red-shift 275 of Amax of the reaction solution combined with the pH -dependence of ε values of the 4-NP and phenolic products (for instance, red-shift of Amax following increase of pH observed for 4-NC, Fig. S5). Also, because of a significant increase in the actinic flux at λ>400 nm (Fig. S14), any "brown" products formed efficiently stabilize the Rabs values thought the course of the reaction when pH<7 -see also Fig. S12.
The first-order kbleaching coefficients derived via eq. II (Table S3) show that the lifetimes of BrC chromophores are 3 and 1.5times longer than the lifetime of 4-NP (precursor) under acidic and basic pH conditions, respectively, due to formation of light-285 absorbing products. The rates of bleaching of 4-NP solution due to reaction with OH and due to direct photolysis were compared (Fig. 4) using the kbleaching values derived in this work and the previously reported, average quantum yields (ϕ, molecules×photon -1 )see Table S3 and eq. SIII and IV (Braman et al., 2020;Lemaire et al., 1985;Biswal et al., 2013;Garcıá Einschlag et al., 2003). The ϕ values listed in Table S3 were derived by measuring decrease in the absorbance of 4-NP solution, hence can be regarded as effective ϕ for the bleaching of 4-NP and 4-nitrophenolate-derived BrC. To our knowledge, the 290 wavelength-dependent ϕ values for 4-NP and 4-nitrophenolate are not available.
https://doi.org/10.5194/acp-2021-871 Preprint. Discussion started: 28 October 2021 c Author(s) 2021. CC BY 4.0 License. Figure 4: The estimated aqueous-phase lifetimes of light-absorbing compounds in the solutions of 4-NP and 4-nitrophenolate due to reaction with OH and direct photolysis. The lifetimes due to reaction with OH were calculated with kbleaching coefficients derived via eq. II with the data acquired in this work. The average lifetime due to photolysis is shown for zenith angles 0-50°, shaded area is 295 2σ, representing the range of photolysis lifetimes calculated via eq. SIV.
As presented in Fig. 4, both bleaching mechanisms can be relevant under realistic atmospheric conditions, depending from [OH]. Bleaching by OH is expected to be a more dominant pathway for 4-nitrophenolate, due to its lower reported quantum yields (ϕ) combined with higher OH reactivity of the precursor at pH>8 (Lemaire et al., 1985;Braman et al., 2020).
Previously, the reaction with the OH was concluded to be the major removal mechanisms for a number nitrophenols in the 300 atmospheric aqueous phase (Zhao et al., 2015;Vione et al., 2009;Albinet et al., 2010). In the context of the formation and processing of the atmospheric BrC, reaction of OH with 4-NP leads to the removal of light-absorbing compounds. At the same time, it was previously concluded that reaction of OH with 5-nitroguaiacol, 4-NC and dinitrophenol initially lead to the increase in the light-absorptivity followed by a rapid bleaching of the reaction solution (Hems and Abbatt, 2018;Zhao et al., 2015).
of BrC chromophores are expected to be significantly longer than the lifetimes of the parent nitrophenols (precursors) due to the formation of aromatic, light-absorbing by-products (Hems and Abbatt, 2018;Zhao et al., 2015).
Based on GC/MS quantitative data it was estimated that in reaction (1) ca. 20 to 40% of 4-NP, depending on pH, was converted into phenolic products. A low degree of mineralization of the precursor (ca. 15% -section S15) and up to 40% yield of 310 phenolic products indicates that reaction (1) generates a substantial amount, between 45 to 65%, of non-aromatic products, like for instance functionalized carboxylic acid (Hems and Abbatt, 2018;Kavitha and Palanivelu, 2005;Zhang et al., 2003;Oturan et al., 2000). Consequently, reaction (1) and reaction of other nitrophenols with OH can contribute to acidity of atmospheric aqueous particles via formation of NO2 -, NO3and organic (nitrated) acids (Tilgner et al., 2021). Additionally, aqueous oxidation of nitrophenols via OH may be a source of potentially toxic and harmful aqueous SOAs (aqSOAs). 315 Data availability. The raw data can be obtained by contacting the corresponding author.
Author contributions. BW designed the study, developed the methodology analyzed the data and wrote the paper PJ carried out the experiments, optimized the methodology and processed the raw data, TG supervised the experiments, analyzed the data and contributed to the final manuscript. All authors contributed to the interpretation of the results.
Competing interests. The authors declare that they have no conflict of interest 320