Characterization of Chemical Components and Optical Properties of Toluene Secondary Organic Aerosol in Presence of Ferric Chloride Fine Particles

Abstract

Iron ion is the common transition metal ion in atmospheric aerosol, which can affect the components and optics of secondary organic aerosol (SOA). In the current study, the atmospheric photooxidation of toluene to produce SOA in the presence of ferric chloride fine particles is simulated in a smog chamber; on-line and off-line mass spectrometry and spectroscopic instruments are used to characterize constituents and optics of SOA. Compare with SOA formed in the absence of fine particles, the laser desorption/ionization mass spectra of toluene SOA generated in the presence of ferric chloride fine particles show ion peaks of m/z = 163 and 178, the UV-Vis spectra of the extracting solution for toluene SOA have peaks near 400 and 700 nm, and the electrospray ionization mass spectra contain peaks at m/z = 248 and 300. Based on this spectral information, it is shown that gaseous methylcatechol formed from photooxidation of toluene may react with iron ion on the surface of fine particles by complexing and oxidation–reduction, resulting in methylbenzoquinone products and metallo-organic complex ions such as [Fe(III)(CH₃C₆H₃OO)]⁺, [Fe(III)(CH₃C₆H₃ OO)₂]⁻ and [Fe(III)(CH₃C₆H₃OO)Cl₂]⁻. These products have strong light absorption ability, resulting in an increase in the averaged mass absorption coefficient (<MAC>) in the 200~1000 nm range and the MAC at 365 nm (MAC₃₆₅) for toluene SOA, while <MAC> and MAC₃₆₅ progressively increase with an increasing concentration of ferric chloride fine particles. These results serve as experimental references for the study of the formation mechanism and optical properties of metallo-organic complexes in atmospheric aerosol particles in regions experiencing high levels of fine particles of metal and automobile exhaust pollution.

1. Introduction

Toluene, ethylbenzene, xylene and other aromatic compounds are common organic pollutants in the atmosphere [1,2] and important precursors of secondary organic aerosol (SOA) [3,4,5]. SOA can scatter and absorb solar radiation, decrease atmospheric visibility [6,7,8,9,10] and endanger human health [11,12]. The study of its formation mechanism and optical properties has attracted extensive attention. Fine particles of metal emitted by human sources are main components of inorganic fine particles in the atmosphere [13]. Metals in form of oxides, hydroxides and silicates are transformed into soluble metal ions through atmospheric chemical processes [14,15]. Fe, Cu, Zn and Mn are common transition metal ions in fine particles [16]. The large specific surface area of fine particles of metal is conducive to the condensation and reaction of semi- and non-volatile compounds, and it can affect the yield and composition of SOA particles [17,18]. In addition, transition metal ions have catalytic and redox capabilities. As a bridge between laboratory research and field monitoring, a smog chamber can be used to simulate atmospheric conditions; it is thus an indispensable experimental device for exploring atmospheric chemical reactions. A smog chamber is often utilized to simulate the atmospheric chemical processes of volatile organic compounds [19,20,21]. Chu et al. [22,23,24] carried out photooxidation experiments of toluene using a smog chamber and found that MnSO₄ and ZnSO₄ fine particles can catalyze the photolysis of H₂O₂ to form OH radicals, which react with toluene to produce more products. These products are partitioned into particle phases, leading to an increase in the yield of SOA. On the contrary, when FeSO₄ particles exist in a system, gaseous oxidation products of toluene and other compounds condense on fine particles, and a redox reaction with Fe²⁺ generates volatile compounds, which decreases the yield of SOA.

It should be noted that some d, s and p orbitals in the valence electron layer of transition metal ions are empty orbitals, which can form metallo-organic complexes with oxygen and nitrogen atoms containing lone pair electrons in phenolic, carboxylic acids and nitro compounds via coordination bonds [25]. Carboxylic acids and phenolic compounds are common organics in atmospheric aerosol particles. After they lose hydrogen atoms, oxygen atoms containing lone pair electrons have strong complexation ability and can form stable metallo-organic complexes with Fe³⁺, Cu²⁺ and other transition metal ions [26,27,28]. These metallo-organic complexes are fat-soluble and easily absorbed by the intestinal wall through the biofilm, as well as transferred into the cerebrovascular system, placenta and other organs. They have stronger biological toxicity than metal ions [29], which has an important impact on human health and the ecological environment and has attracted wide attention. In addition, metallo-organic complexes contain double bonds and other chromophores, which can absorb light radiation and generate π→π* or n→π* transitions. Meanwhile, transition metal ions have d orbitals with unfilled electrons. Under the influence of a ligand field, d orbitals with the same energy as the original orbitals will split. After splitting, there is an energy difference between d orbitals, and the absorbed radiation can undergo d→d transition [30]. Therefore, metallo-organic complexes have strong absorption capacity in the ultraviolet and visible range, and they make a greater contribution to atmospheric light radiation. However, there are few reports of the effect of metal particles on the optical properties of SOA.

Iron ion is a common transition metal ion in atmospheric aerosol particles. It plays a crucial role in the aging of aerosols and the formation of brown carbon. Al-Abadleh addressed the current state of knowledge on iron chemistry that leads to SOA formation. She provided a detailed overview of the oxidation–reduction and complexation reactions of catechol, guaiacol and other polyhydroxyphenols, along with dicarboxylic acids such as oxalic acid and malonic acid, with iron ion to produce soluble and insoluble brown carbon products [31]. With the extensive use of fossil fuels in China, urban PM₂.₅ contains a relatively high concentration of iron and other metal ions [32]. To date, the highest concentrations of soluble iron in PM₂.₅ particles in Beijing have been 116.3 ng·m⁻³ and 308.9 ng·m⁻³ in summer and winter, respectively [33]. Toluene is the most highly concentrated aromatic compound and the main precursor of anthropogenic SOA [3,4]. Recently, our group has used a smog chamber to study the photooxidation reaction of toluene initiated by OH radicals. The chemical components of SOA were measured using an aerosol laser time-of-flight mass spectrometer (ALTOFMS) in real-time, and constituents such as cresol, methylcatechol and carbonyl compounds were obtained [34,35]. In addition, the single-scattering albedo of toluene SOA in the presence of 300 μg·m⁻³ ammonium sulfate fine particles was measured to be 0.81 ± 0.02 at λ = 470 nm [36]. Against that research background, in this study, on-line and off-line detection instruments are utilized to characterize the chemical components and optics of toluene SOA formed in the absence and presence of ferric chloride fine particles, as well as to identify metallo-organic complexes. Furthermore, the influence of different concentrations of ferric chloride fine particles on the generation of metallo-organic complexes and the optical properties of toluene SOA is also explored. The findings provide experimental references for studying the formation and optics of metallo-organic complexes in atmospheric aerosol particles containing transition metal ions.

2. Experiments

2.1. Materials

Toluene (>99.7%) was purchased from Aladdin Reagent (Shanghai) Co., Ltd. (Shanghai, China). Methanol (>99%) and anhydrous ferric chloride (98%) were provided by Sigma Aldrich (Shanghai) Trading Co., Ltd. (Shanghai, China).

2.2. Smog Chamber Experiments

Experiments were performed in a homemade 850 L chamber [35,36]. The chamber device is mainly composed of a clean air generation system, inlet system, Teflon chamber and detection system. The Teflon chamber is made of an FEP-Teflon film, where one end of the film is sealed directly and the other end is sealed through a flange. The outer side of the flange is installed with injection and sampling tubes, while the inner side is equipped with a small fan to evenly mix the gas and particulate matter in the chamber. As shown in Figure 1, the relative humidity and temperature are detected by an HMT 333 sensor (Vaisala, Finland). According to our previous experimental results [36,37], the relative humidity (RH) of clean air generated by the air purification system is 25 ± 2%. Therefore, all experiments are carried out under the condition of RH = 25 ± 2%. Meanwhile, the temperature is controlled at 298 ± 1 K by the laboratory air conditioning system. Thus, the relative humidity and temperature are 25 ± 2% and 298 ± 1 K for each experiment. After cleaning the chamber, it is filled with about 150 L clean air, and then an aerosol generator (TSI 9302, Shoreview, MN, USA) is used to atomize 1 g·L⁻¹ ferric chloride solution to produce fine particles with sizes of less than 2 μm, which are diffused through a silica gel drying tube and then entered into the chamber. A certain amount of toluene is sampled by a micro liter injector and delivered to a liquid gasification bottle heated to 353 K by a temperature-controlled heating sleeve. By doing so, it is volatilized into gas and expanded into a gas distribution system; then, it is entered into the chamber with clean air. Subsequently, ozone formed using an XM-TS ozone generator (Qingdao Xinmei Purification Equipment Co., Ltd., Qingdao China) is added to the gas distribution system, then introduced into the chamber with clean air. Finally, the chamber is filled to the full volume with clean air, and the reactants are evenly mixed with the help of small fan in the flange. Four ultraviolet lamps (characteristic wavelength of 254 nm) around the chamber are turned on to illuminate the ozone, generating OH radicals [38] and initiating the photooxidation reaction of toluene. An ozone analyzer (GT-1000-O3, Shenzhen Colno Electronic Technology Co., Ltd., Shenzhen, China, with flowrate of 0.50 L·min⁻¹), GC-FID (7820A, Agilent, Palo Alto, CA, USA, with flowrate of 0.04 L·min⁻¹) and scanning mobility particle size analyzer (SMPS, 3080L DMA, 3775 CPC, TSI, Shoreview, MN, USA, with flowrate of 0.30 L·min⁻¹) are successively adopted for on-line measurement of the concentrations of ozone, toluene and SOA particles in the chamber [35,36].

2.3. Characterization Components of Toluene SOA

Due to the small volume of the Teflon chamber (850 L), the concentration of aromatic hydrocarbon and oxidant in our previous experiments was about 1–20 ppm [35,36,37]. Although the concentrations of reactants in the chamber are higher than that in the ambient atmosphere, the results of our studies may be useful for SOA formation modeling in areas where automobile exhaust pollution is serious. Thus, the concentrations of ozone and toluene were set at 10 ppm and 1 ppm in this study. The particle number and mass concentration of ferric chloride fine particles generated by TSI 9302 and detected by SMPS were about 3.6 × 10⁷ particle.cm⁻³ and 1400 μg·m⁻³, respectively. The particle number and mass concentration of ferric chloride fine particles can be controlled by adjusting the inlet time. When the clean air was filled to full volume, the particle number concentration of ferric chloride fine particles was in the range of 3.9 × 10⁵–5.5 × 10⁶ particle.cm⁻³ in different experiments. Nine experiments were performed, and each experiment was carried out three times. The concentrations of ozone and toluene in each experiment remained unchanged (10 ppm and 1 ppm), while the concentration of ferric chloride fine particles was set at 0, 15, 30, 60, 90, 120, 150, 180 and 210 μg·m⁻³ in turn. The chemical components of toluene SOA were measured on-line by ALTOFMS after irradiation with a flowrate of 0.14 L·min⁻¹. As displayed in Figure 1, SOA particles sequentially pass through an aerodynamic lens, with two skimmers to form the collimated particle beam, and enter into a sizing system. Particle size is attained by measuring the transit time of particle travel through two infrared laser beams with a wavelength of 532 nm. Then, particles enter the ionization chamber, where they are desorbed and ionized by a 266 nm ultraviolet laser emitted by a Nd: YAG pulsed laser. The formed ions are measured by a time-of-flight mass spectrometer to obtain the mass spectrum of the particles [35,36,37].

Next, SOA particles are collected on a polytetra fluoroethylene membrane filter, and they are extracted into 5 mL of 2% (v/v) methanol water solution by ultrasound for 30 min. UV-Vis spectra of the extracting solutions are detected by a UV-6100S spectrophotometer (Mapada Instruments, China). Moreover, the extracting solutions are measured with a high-performance liquid chromatography mass spectrometer (LC-MS) with electrospray ionization (ESI) (Agilent-1200, Agilent-6320, Palo Alto, CA, USA). The mobile phase is methanol and ultra-pure water (v/v = 1:1) at a flowrate of 0.20 μL·min⁻¹. Since no suitable chromatographic column was found to separate the metallo-organic complexes in this study, however, the negative ionization mode of electrospray ionization of LC-MS caused the organic molecules to dissociate hydrogen ions and produce a deprotonated molecular ion peak ([M-H]⁻), which provided molecular weight information on the tested organic component [39]. This could be used for qualitative analysis of metallo-organic complexes and other components of the extracting solution for toluene SOA. Thus, the sample was measured with no chromatography column attached, and the extracting solution was electrospray ionized directly, then detected in the 50–1000 amu range of the negative mode [36]. Based on the detected deprotonated molecular ion peak ([M-H]⁻), a qualitative analysis of metallo-organic complexes and other components was carried out.

2.4. Optical Characterization of Toluene SOA

Referring to experiments on SOA optical measurements [40,41] and our previous study [42], we characterized the optics of toluene SOA by using a mass absorption coefficient (MAC, cm²·g⁻¹), which can be obtained according to Formula (1):

$$MAC(\lambda )=\frac{A^{solution}\left(\lambda \right)\times \ln 10}{b\times C_{mass}}$$

(1)

$$〈MAC〉=\frac{1}{({\lambda }_2-{\lambda }_1)}\times {\displaystyle {\int }_{{\lambda }_1}^{{\lambda }_2}MAC}(\lambda )d\lambda$$

(2)

where A^solution(λ) is the absorbance of the extracting solution for toluene SOA at λ, C_mass is the organic carbon concentration of the extracting solution (g·cm⁻³) and b is the optical path (1 cm for this study). Meanwhile, the integral value of MAC over the measured wavelength range is divided by (λ₂ − λ₁) to obtain the averaged MAC (<MAC>, cm²·g⁻¹) in the λ₁–λ₂ range (Formula (2)). The absorption spectra at 200–1000 nm and the organic carbon concentration of the extracting solution for toluene SOA particles are detected by a UV-6100s spectrometer and TOC-L organic carbon analyzer (Shimadzu, Japan), respectively [42]. The organic carbon concentration of the 2% methanol water solution is measured firstly by the TOC-L organic carbon analyzer. The organic carbon concentration (C_mass) of the extracting solution for toluene SOA is obtained by subtracting the organic carbon concentration of the 2% methanol water solution from the total organic carbon concentration that is detected [42]. Then, <MAC> of toluene SOA particles in the 200–1000 nm range is acquired via Formulas (1) and (2). MAC at 365 nm (MAC₃₆₅) is more commonly used in the literature [43,44,45]. Thus, by combining the absorbance at 365 nm (A^solution (365 nm)) with the organic carbon concentration (C_mass), MAC₃₆₅ of toluene SOA under different concentrations of ferric chloride fine particles is obtained through Formula (1). It should be pointed out that, regarding the above MAC calculation, the light absorption in the solution is ~2 times higher than that in the air owing to artifacts [46,47].

3. Results

3.1. Generation of Toluene SOA with and without Ferric Chloride Fine Particles

Experimental [36,48,49,50] and theoretical [51,52,53] results show that an addition reaction between toluene and OH radicals mainly occurs, producing methylhydroxycyclo hexadienyl radicals. As shown in Figure 2, an oxygen molecule extracts a hydrogen atom from the carbon atom connected with an OH radical to form cresol. The generated cresol can react with an OH radical and oxygen molecule to yield methylcatechol. In addition, a methylhydroxycyclohexadienyl radical can react with an oxygen molecule to produce a peroxy radical, and through a series of reactions, ring cracking may occur to form glyoxal and other aldehydes. The resultant aldehydes continue to be oxidized by OH radicals to oxalic acid and other carboxylic acids [54]. These products achieve a saturated vapor pressure and then nucleate and condense to generate SOA particles.

SMPS is utilized to detect the concentration of SOA, and the wall effect is then corrected for in accordance with our previous studies [35,36]. Particle wall loss is defined as a first-order process dependent on particle loss coefficient, k_dep (d_p):

$$k_{dep}(d_p)=a{d}_p^b+c/d_p^d$$

where a, b, c and d are 4.17 × 10⁻¹³, 4.66, 10.18 and 0.75, respectively [35,36,37]. The concentration of toluene SOA is corrected by fitting the particle number concentration decay at the end of each experiment. For the case of photooxidation of toluene in the absence of ferric chloride fine particles, as shown in Figure 3a, relative humidity during the irradiation process is maintained at 24–25%, and the temperature is between 298 and 300 K. The concentration of toluene and ozone decreases gradually with the extension of irradiation time (Figure 3b). The concentration of toluene SOA after wall effect correction as a function of irradiation time is shown in Figure 3b. A certain amount of toluene is consumed within 15 min so that products such as methylcatechol and glyoxal achieve saturated vapor pressure, meaning less SOA is generated and the concentration is only 6 μg·m⁻³. After that, the formation of gaseous products keeps increasing and the SOA concentration rises expeditiously, from 22 to 174 μg·m⁻³ within 30 to 150 min. The presence of toluene is not detected in the chamber at 180 min, and the SOA concentration attains a maximum value of 202 μg·m⁻³. Thereafter, due to the reaction of toluene being completed, methylcatechol, glyoxal and other products are no longer produced, and the corrected SOA concentration tends to stabilize [35,36].

The variation curves of relative humidity, temperature, toluene and ozone consumption, as well as SOA formation with irradiation time throughout the experiments in the presence of ferric chloride fine particles, are analogous to those in the absence of fine particles shown in Figure 3. However, the formation concentration of toluene SOA increases to varying degrees when compared to that in the absence of fine particles. For example, in the presence of 120 μg·m⁻³ ferric chloride fine particles, as displayed in Figure 4, the concentration of toluene SOA as a function of time is higher than that in the absence of fine particles. The maximum concentration of SOA reaches 346 μg·m⁻³ at 180 min irradiation, which is about 71% higher than that in the absence of fine particles. In addition, the maximum concentration of SOA particles after wall effect correction, illustrated in Figure 5, is about 234–350 μg·m⁻³ under different concentrations of ferric chloride fine particles, which is larger than the maximum SOA concentration without fine particles (202 μg·m⁻³). These results indicate that ferric chloride fine particles can change the gas/particle partitioning process of toluene photooxidation products, so that more methylcatechol, glyoxal and other gaseous products are partitioned to the particulate phase, thus facilitating the generation of SOA particles. However, when the concentration of ferric chloride fine particles is greater than 120 μg·m⁻³, the maximum SOA concentration remains almost constant. For characterizing the components of toluene SOA, on-line mass spectrometry and off-line spectral determinations are carried out.

3.2. Chemical Constituents of Toluene SOA with and without Ferric Chloride Fine Particles

In order to facilitate comparison, the constituents of SOA without fine particles are detected on-line by ALTOFMS firstly. The averaged laser desorption/ionization positive ion mass spectra for 200 toluene SOA single particles are displayed in Figure 6. In addition to fragment peaks of organic carbon (m/z = 12 (C⁺) and 24 (C₂⁺)), there are characteristic fragment ion peaks of phenolic substances (m/z = 93, C₆H₅O⁺), the benzene ring (m/z = 77, C₆H₅⁺) and its fragment peaks (m/z = 65, C₅H₅⁺ and m/z = 39, C₃H₃⁺), demonstrating the presence of phenolic compounds in SOA particles [37,55]. In addition, mass spectra show characteristic peaks of carbonyl compounds at m/z = 29 (HCO⁺), 43 (CH₃CO⁺) and 57 (HCOCO⁺) [37,55], which indicate that toluene SOA particles contain carbonyl compounds. UV-visible absorption spectra of the extracting solution for toluene SOA particles without fine particles, displayed in Figure 7, have a strong peak at 210 nm, which represents characteristic absorption generated by the n→π* transition of C=O [39]. Furthermore, there is an absorption band near 277 nm, which corresponds to the characteristic absorption of phenolic substances [56]. These confirm again that phenolic and carbonyl compounds are principal constituents of toluene SOA.

The presence of phenolic and carbonyl products is reconfirmed by electrospray ionization mass spectra of the extracting solution for toluene SOA without fine particles. The negative ionization mode of electrospray ionization causes organic molecules to dissociate hydrogen ions and produces a deprotonated molecular ion peak ([M-H]^–), which provides molecular weight information on the tested organic component [39]. As displayed in Figure 8, electrospray ionization negative ion mass spectra contain [M-H]^– ion peaks of m/z 57, 71, 89, 107, 123 and 139. Based on the possible reaction mechanism of toluene with OH radical and molecular weight (Mw) information, these ion peaks correspond to carbonyl and phenolic products, for example, glyoxal (Mw = 58, m/z = 57), methylglyoxal (Mw = 72, m/z = 71), oxalic acid (Mw = 90, m/z = 89), 2-methyl-2,3-epoxy-butanedial (Mw = 98, m/z = 97), cresol (Mw = 108, m/z = 107), methylcatechol (Mw = 124, m/z = 123) and methylpyrogallol (Mw = 140, m/z = 139). These results are consistent with previous chamber experiments [36,48,49,50], which detected cresol, methylcatechol, glyoxal, oxalic acid and other products in toluene SOA particles.

It should be pointed out that, due to the limitation of data processing software, it is currently not possible to obtain differential mass spectra for quantitative analysis. However, based on the laser desorption/ionization positive ion mass spectra and electrospray ionization negative ion mass spectra of toluene SOA particles measured by ALTOFMS and LC-MS, the difference in mass peaks was qualitatively analyzed to illustrate the effect of ferric chloride fine particles on the photooxidation products of toluene and confirm the components of metallo-organic complexes. In the presence of 120 μg·m⁻³ ferric chloride fine particles, the averaged laser desorption/ionization positive ion mass spectra of 200 toluene SOA single particles were detected by ALTOFMS, as illustrated in Figure 9. The peak intensity of m/z = 56 in the spectra is the largest, corresponding to the mass peak of Fe⁺ ions, while the strong intensity of m/z = 72 represents the FeO⁺ ion peak [57]. In addition to organic carbon fragment peaks (m/z = 12, 24), carbonyl compound fragment peaks (m/z = 29, 43), a phenolic compound characteristic peak (m/z = 93), benzene ion and its fragment peak (m/z = 77, 65, 39), in this study, peaks of m/z = 107, 122, 164 and 178 with high intensity are newly present. The peak with m/z = 122 can probably be identified as a CH₃C₆H₃O₂⁺ ion peak formed from the loss of two hydrogen atoms from the methylcatechol molecular ion, and we propose that m/z = 107 (C₆H₃O₂⁺ ion) is generated by the loss of a methyl group from the CH₃C₆H₃O₂⁺ ion. Moreover, the peak of m/z = 178 probably corresponds to a double dentate mononuclear methylcatechol iron ([Fe(III)(CH₃ C₆H₃OO)]⁺) peak, and we propose that m/z = 163 is an ion peak generated when [Fe(III)(CH₃C₆H₃OO)]⁺ ion loses its methyl group ([Fe(III) (C₆H₃OO)]⁺ ion).

Rizvi et al. [58] showed that a low-spin (t₂g₅) octahedron configuration of Fe (III) is a hard cation occupying a single π symmetric t₂g₅ orbital. According to matching hard–hard interaction, harder donor site (such as negatively charged oxygen) ligands can bind Fe (III) more strongly. According to the review of Al-Abadleh [31], catechol, guaiacol and other polyhydroxyphenols are common organic ligands. After catechol is dissolved in water and loses hydrogen ions, the two oxygen atoms are negatively charged and complexed with Fe(III) to produce catechol iron (III) ion ([Fe(III)(C₆H₄OO)]⁺). When ferric chloride fine particles exist in the reaction system, due to its strong moisture absorption capacity, Fe³⁺ and Cl⁻ ions are generated after surface moisture absorption. As shown in Figure 10, gaseous methylcatechol formed from toluene photooxidation condenses on the surface of fine particles, and it may undergo the reaction proposed by Al-Abadleh [31] to generate metallo-organic complex ions. After methylcatechol dissolves in the hygroscopic surface of ferric chloride fine particles and loses hydrogen ions, two oxygen atoms are negatively charged and complexed with Fe(III) to generate methylcatechol iron (III) ion ([Fe(III)(CH₃ C₆H₃OO)]⁺).

UV-Vis spectra of the extracting solution for toluene SOA with 120 μg·m⁻³ ferric chloride fine particles further confirmed the existence of methylcatechol iron (III) ion. As shown in Figure 7, in addition to characteristic absorption peaks of carbonyl and phenolic compounds at 210 nm and 277 nm, new absorption peaks appeared near 400 nm and 700 nm. As suggested by Albarran et al. [59], the absorption peak at 400 nm is a characteristic peak formed from the n→π* transition of benzoquinone molecules. This indicates the presence of benzoquinone products in the extracting solution. Powell et al. [60] and Slikboer et al. [61] have also observed an absorption peak near 700 nm when conducting an aqueous reaction of catechol and other phenolic compounds with an Fe (III) ion. On the basis of their analysis, it seems this peak is characteristic absorption resulting from charge transfer between a ligand and metal in the catechol–iron (III) ion. UV-Vis spectra of the extracting solution in this experiment, displayed in Figure 7, also reveal the absorption peak of 700 nm, indicating that gaseous methylcatechol produced by photooxidation of toluene can condense on the surface of hygroscopic ferric chloride particles, and a similar aqueous reaction occurs to generate methylcatechol iron (III) ion.

As shown in Figure 11, in addition to mass peaks of m/z = 57 (glyoxal), m/z = 71 (methylglyoxal), m/z = 89 (oxalic acid), m/z = 97 (2-methyl-2,3-epoxy-butanedial), m/z = 107 (cresol) and m/z = 123 (methylcatechol) in electrospray ionization negative ion mass spectra of the extracting solution for toluene SOA in the presence of 120 μg·m⁻³ ferric chloride fine particles, peaks of m/z = 121, 248 and 300 are also newly detected. The peak of m/z = 121 probably represents the peak of the deprotonated molecular ion of methylbenzoquinone (CH₃C₆H₃OO⁻) [59], while the peaks of m/z = 248 and 300 may be identified as metallo-organic complex ion peaks. As shown in Figure 10, methylcatechol dissolves in water and loses hydrogen ions to produce CH₃C₆H₃OO²⁻ ion, which is complexed with Fe(III) to form methylcatechol iron (III) ion ([Fe(III)(CH₃C₆H₃OO)]⁺). As suggested by Slikboer et al. [61], methylcatechol iron (III) ion undergoes a redox reaction with the action of an oxygen molecule to produce a ferrous ion and methylbenzoquinone. In addition, methylcatechol iron (III) ion can be further complexed with CH₃C₆H₃OO²⁻ ion to generate dimethylcatechol iron (III) ion ([Fe(III)(CH₃ C₆H₃OO)₂]⁻) [62]. The mass charge ratio of this ion is 300, so it can be inferred that m/z = 300 corresponds to the peak of this metallo-organic complex ion. It is worth noting that Cl⁻ ion produced on the surface of ferric chloride fine particles after moisture absorption is also a common inorganic ligand, which can be complexed with Fe(III) in catechol iron (III) ion to form dichloromethylcatechol iron (III) ion ([Fe(III)(CH₃C₆H₃OO)Cl₂]⁻) [62]. The mass/charge ratio of this complex ion is 248, and m/z = 248 (shown in Figure 11) can be identified as the ion peak of [Fe(III)(CH₃C₆H₃OO)Cl₂]⁻. Based on the above on-line and off-line measurement analysis, gaseous methylcatechol formed from photooxidation of toluene can coagulate and react on the surface of hygroscopic ferric chloride fine particles to generate metallo-organic ions, such as methylcatechol iron (III) ion, dimethylcatechol iron (III) ion, dichloro-methylcatechol iron (III) ion and methylbenzoquinone products.

3.3. MAC of Toluene SOA under Different Concentrations of Ferric Chloride Fine Particles

The specific surface of ferric chloride fine particles is the condensation center for gaseous products formed from photooxidation of toluene. The concentration of fine particles decides the size of the specific surface, which impacts the components and optical properties of SOA. <MAC> in the 200–1000 nm range and MAC₃₆₅ of toluene SOA under ferric chloride fine particles with different concentrations were measured and the results are shown in Figure 12 and Figure 13. The measured <MAC> of toluene SOA without fine particles is 178 cm²·g⁻¹ (Figure 12), which is comparable to <MAC> of aromatic SOA reported by Updyke et al. [40] (200–300 cm²·g⁻¹). Meanwhile, MAC₃₆₅ of toluene SOA without fine particles is 407 cm²·g⁻¹ (Figure 13), which is slightly higher than the value of aromatic SOA measured by Updyke et al. [40] (300–400 cm²·g⁻¹), and less than the lower limit of 365 nm absorption values of water-soluble organic carbon (WSOC) in the outflow from northern China (2000–11,000 cm²·g⁻¹) reported by Kirillova et al. [44]. The above test results show that methylcatechol and other phenolic compounds and carbonyl compounds such as glyoxal are the main constituents of toluene SOA without ferric chloride fine particles. These component molecules only have C=O and C=C double bonds, do not contain strong chromophores or auxochromes and have weak particle light absorption ability [43].

When ferric chloride fine particles are present in the reaction system, <MAC> and MAC₃₆₅ of toluene SOA particles increase obviously and gradually with a rising concentration of fine particles. When the concentration of ferric chloride fine particles is 120 μg·m⁻³, <MAC> of toluene SOA is 420 cm²·g⁻¹ (Figure 12), which is greater than <MAC> of ammonia-aging SOA reported by Updyke et al. [40] (300 cm²·g⁻¹), and is 136% higher than that without fine particles (178 cm²·g⁻¹). In addition, MAC₃₆₅ of toluene SOA in the presence of 120 μg·m⁻³ ferric chloride fine particles is 8328 cm²·g⁻¹ (Figure 13), which is close to the upper limit value of MAC₃₆₅ for WSOC (2000–11,000 cm²·g⁻¹) measured by Kirillova et al. [44], and approximately 20 times that without fine particles (407 cm²·g⁻¹).

This is mainly because high concentration of ferric chloride fine particles can provide more specific surface area and reaction sites, which is conducive to the condensation and reaction of more gaseous methylcatechol, which reacts with iron ion to form more metallo-organic complexes and methylbenzoquinones, as shown in Figure 10. Compared with metylcatechol and carbonyl compounds, methylbenzoquinones contain more double bonds and thus have greater light absorption capacity [43]. In addition, methylcatechol iron (III) ion and other metallo-organic complex ions have strong light absorption ability [30], so the light absorption ability of particles is significantly enhanced. It is worth noting that <MAC> and MAC₃₆₅ of SOA do not increase further when the concentration of ferric chloride fine particles is greater than 120 μg·m⁻³. The concentrations of ozone and toluene remained constant in this study, resulting in a certain amount of gaseous methylcatechol for each experiment. When fine particles rise to a certain concentration (120 μg·m⁻³), most of the gaseous methylcatechol is converted to form metallo-organic complexes and methylbenzoquinone products. Following that, the concentration of ferric chloride fine particles continues to increase, while metallo-organic complexes and benzoquinone products no longer increase significantly, meaning the chemical components and optics of SOA tend to be stable.

It is worth noting that Dhulipala et al. carried out experimental studies on the generation of SOA via chlorine-initiated oxidation of toluene, and they found that the reaction rate of Cl and toluene was higher than that of OH radicals and toluene [63]. The oxidation of toluene induced by Cl radicals mainly occurs through a methyl hydrogen extraction reaction, producing benzoquinone, benzoic acid and other ring-retaining products, thus affecting the chemical components of atmospheric SOA in coastal areas and other regions with high chlorine content [63]. For comparison with the experiment of Dhulipala et al. [63], in this study, the effect of ferric chloride fine particles on the chemical components of toluene SOA was investigated using an ALTOFMS, UV-Vis absorption spectrometer and electrospray ionization mass spectrometer. It was found that methylcatechol and other gaseous products formed from the OH-initiated oxidation of toluene can condense on the hygroscopic surface of ferric chloride fine particles and undergo a complex reaction to generate metallo-organic complexes with strong light absorption ability. In addition, a high concentration of ferric chloride fine particles can provide more specific surface area and reaction sites, which are conducive to the condensation and reaction of gaseous methylcatechol to form more metallo-organic complexes.

It should be pointed out that the concentration of ferric chloride fine particles used in the chamber study is three orders of magnitude higher than the typical ambient concentration in the air. Nevertheless, the effect of a high concentration of ferric chloride fine particles may be expected to occur in atmospheric conditions since there is evaporation of the aqueous phase, such as liquid aerosol particles, rain and cloud droplets, under which conditions the concentration of iron ion is relatively high [31]. As reviewed by Al-Abadleh, the concentration of dissolved Fe (III) in cloud droplets could be as high as 10 mM [31]. Especially in regions with serious pollution from fine particles of metal and automobile exhaust, under high humidity conditions, methylcatechol and other gaseous products formed from photooxidation of toluene and other aromatic compounds may condense on the surface of liquid aerosol containing iron ions, and they may form metallo-organic complexes with high toxicity and strong light absorption ability, thus changing the chemical composition and optical properties of atmospheric SOA. Our results provide experimental references for studies on the formation and optics of metallo-organic complexes in atmospheric aerosol particles in regions experiencing high levels of fine particles of metal and automobile exhaust pollution, and they may serve as beneficial references for designing chamber experiments involving fine particles of metal.

4. Conclusions

Ferric chloride fine particles from brake wear and other anthropogenic sources are common metal particles in atmosphere, whcih can interact with VOCs via complexing and oxidation–reduction. In this study, the constituents of toluene SOA particles without and with ferric chloride fine particles were characterized by an ALTOFMS, UV-Vis absorption spectrometer and electrospray ionization mass spectrometer. It was found that methylcatechol and other phenolic compounds and carbonyl compounds such as glyoxal are the main components of toluene SOA particles without ferric chloride fine particles. These constituent molecules only have C=O and C=C double bonds, do not contain strong chromophores or auxochromes and have weak light absorption ability with small <MAC> and MAC₃₆₅. However, when ferric chloride fine particles exist in the system, gaseous methylcatechol can undergo complexation and a redox reaction with iron ion on the surface of fine particles, resulting in metallo-organic complex ions such as dichloro-methylcatechol iron (III) ion and dimethylcatechol iron (III) ion, along with methylbenzoquinone products. <MAC> and MAC₃₆₅ of toluene SOA increase obviously due to the strong light absorption ability of these products, and they progressively increase with a rising concentration of ferric chloride. Nevertheless, limitations of the study due to using concentrations of reactants much higher than the ambient ones. In addition, metallo-organic complexes and other products were generated under certain relative humidity conditions, and MAC of SOA was measured using an off-line method. Thus, subsequent experiments should be performed to study the influence of relative humidity on the formation of metallo-organic complexes, as well as to conduct on-line measurement of the extinction and scattering coefficient and other optical parameters of SOA.