Photonitration of pyrene adsorbed on silica gel with NO₂

Abstract

To examine the heterogeneous photonitration of pyrene with NO₂ (approximately 0.2 ppm) on a heavy-traffic road, we studied the photonitration of pyrene adsorbed (pyrene_ads) on silica gel, which was used as SiO₂ in particulate matter (PM), with NO₂ (10.2, 2.0, and 0.20 ppm) under the atmospheric concentration ratio of pyrene_ads to NO₂ and compared the results with those obtained in the dark. The effects of irradiation, wavelength, and oxygen concentration in a NO₂ diluent on the photonitration were examined using a fluidized-bed column irradiated with simulated or real sunlight. Under the UV-light absorption of pyrene, the concentration of pyrene decreased exponentially in accordance with a pseudo-first-order reaction, while in the dark, it decreased sigmoidally in accordance with a H⁺-autocatalyzed reaction. The distribution and the yields of formed nitration products and their photooxidation products were affected by the light intensity, concentrations of NO₂, and oxygen in the NO₂ diluent. In the photonitration experiments using a high-pressure mercury lamp, formed 1-nitropyrene and minor dinitropyrenes were decreased by the transformation into their photooxidation products. Under 8-h exposure of pyrene to 10.2-ppm NO₂, the yield of 1-nitropyrene was 42% in N₂ and 28% in air. The oxygen inhibitory effect can be explained by the energy transfer from ¹pyrene* to oxygen. Radical cation intermediate (pyrene^•+-NO₂⁻) was proposed for 1-nitropyrene formation. Under 24-h exposure of pyrene to 2.0-ppm NO₂, the yields of 1-nitropyrene and the photooxidation products were 21.6% and 8.0%, respectively, in N₂ and 4.9% and 3.8%, respectively, in air. Under 24-h exposure of pyrene to 0.20-ppm NO₂, which is two times the 1-h NO₂ standard in the USA and China, the yields of 1-nitropyrene and the photooxidation products were 2.3% and 3.4%, respectively, in N₂ and 2.1% and 0.9%, respectively, in air. The significant decrease in the yields of 1-nitropyrene and the photooxidation products under the concentration of 0.20-ppm NO₂ can be explained by their easy photodecomposition with the increase in the photolysis of pyrene. Under the concentration of 0.20-ppm NO₂ in air, which is approximately the concentration on heavy-traffic roads, the decay rate of pyrene by the photonitration was increased by own photolysis, although the photonitration was inhibited by oxygen in air.

1 Introduction

Keyte et al. (2013) summarized the formation of mutagenic nitropolycyclic aromatic hydrocarbons (NPAHs) by the reaction of polycyclic aromatic hydrocarbons (PAHs) adsorbed on PM with NO₂; however, the effect of light irradiation has not been investigated under atmospheric concentrations of NO₂ and PAH adsorbed (PAH_ads) on PM. Seasonal and downwind sampling of PM showed that considerable amounts of NPAHs and oxygenated PAHs were secondarily formed by the photochemical reactions of PAHs on urban PM (Kojima et al., 2010; Lin et al., 2015; Ma et al., 2016; Zhuo et al., 2017) and during atmospheric transport of PAH-adsorbed PM (Eiguren-Fernandez et al., 2008). To investigate the secondary transformation of the PAHs, it is necessary to study the photonitration and photooxidation of the PAHs under the polluted atmospheric concentrations of NO₂ and PAHs_ads (representative average concentrations of pyrene on PM in winter: 20 pg μg⁻¹, Noto Peninsula, Japan, Tang et al., 2015, and 206 pg μg⁻¹, Beijing, China, Lin et al., 2015); however, previous studies (Wu & Niki, 1985; Inazu et al., 1997; Cochran et al., 2016; Guan et al., 2017) were carried out under one to two orders of higher concentrations of NO₂ and PAH_ads than those in the polluted atmosphere, except for the experiment by Sugiyama et al. (2001) using 0.2-ppm NO₂, and gave no special consideration to the concentrations of PAHs_ads and NO₂.

Wu and Niki (1985) reported the reactions between NO₂ (≥ 25 ppm) and pyrene adsorbed on silica plates in a N₂ or O₂/N₂ diluent using fluorescence spectroscopy, showing that pyrene decay followed a first-order kinetics with NO₂ concentration and was inhibited by O₂ in diluent gas. Inazu et al. (1997) studied the heterogeneous photoreaction of fluoranthene adsorbed on oxide particles with NO₂ (10 ppm) in a N₂ or air diluent and reported the distribution of resulting nitrofluoranthene, the formation mechanism of major 2-nitrofluoranthene, and its degradation with time. Although photoirradiation and oxygen promoted the degradation of fluoranthene, the structures of yellow degradation products were not identified because of insolubility in solvents. Sugiyama et al. (2001) studied the photonitration of pyrene adsorbed on metal oxides with NO₂ (0.2 ppm), showing that the yield of major 1-nitropyrene on silica gel for 2 h is 0.15%. Cochran et al. (2016) reported that UV irradiation of pyrene adsorbed on quartz filters in the presence of NO₂ (1–10 ppm) gave 1-nitropyrene; however, the yield was significantly lower than that without irradiation due to the photodegradation of 1-nitropyrene. Guan et al. (2017) reported that simulated sunlight enhanced the reactivity of PAHs adsorbed on soot with NO₂ (20–150 ppm); however, the yields of newly formed NPAHs were lower than those in the dark. These previous studies have not discussed the photonitration kinetics and mechanism based on products analyses.

The concentration of SiO₂ is high in inorganic particles in PM in large cities in the USA, Europe, and Asia (Harrison & Yin, 2000; Rodriguez et al., 2009; Gao et al., 2015). As silica gel has been selected because of its well-defined physical and chemical properties (Perraudin et al., 2005), a silica gel with a large surface area was used as SiO₂ in PM to compare the photonitration of pyrene with the nitration of PAHs adsorbed on silica gel (Sugiyama et al., 2001; Wang et al., 1999; Hasegawa et al., 2020) and silica particles (Perraudin et al., 2005; Miet et al., 2009; Ma et al., 2011). The chemically synthesized silica gel and silica having a SiO₂ structure have been used as substrates for the nitration of PAHs (Sugiyama et al., 2001; Wang et al., 1999; Hasegawa et al., 2020; Perraudin et al., 2005; Miet et al., 2009; Ma et al., 2011) and their photolysis (Barbas et al., 1993; Mao & Thomas, 1994, 1995; Reyes et al., 2000) because the analysis of reactions on collected PM is complex, depending on the constituents. One-hour NO₂ standards in the USA and China are set at 0.10 ppm and 200 μg m⁻³ (0.105 ppm), respectively. The US Environmental Protection Agency (EPA) has reported that NO₂ in heavy traffic or on freeways can be twice as high as levels measured in residential areas, and NO₂ concentrations may be 30 to 100% higher within approximately 50 m of heavy traffic/freeways, (United States Environmental Protection Agency (U. S. EPA), 2011). A meta-analysis of near-road NO₂ air pollution showed that about 17% NO₂ would be decreased at 100 m from the roads (Liu et al., 2019). Therefore, NO₂ (0.20 ppm), which can be observed on heavy-traffic roads, was used as the lowest NO₂ concentration. To understand the photonitration of pyrene with NO₂ under polluted atmosphere concentrations of pyrene_ads (20 pg on 1-μg silica gel) and NO₂ (0.20 ppm), we also studied the photonitration of pyrene adsorbed on silica gel with NO₂ (2.0 and 10.2 ppm) to facilitate the products analyses. The pyrene_ads was set to be proportional to the concentrations of NO₂. The effects of light intensity, wavelength, and oxygen concentration in the NO₂ diluent on the photonitration of pyrene and the transformation products were investigated using a fluidized-bed column irradiated with simulated or real sunlight.

2 Materials and methods

2.1 Materials

Special-grade pyrene, 1-nitropyrene, 1,3-, 1,6-, and 1,8-dinitropyrenes, and 1,3-, 1,6-, and 1,8-pyrenedione were purchased from Wako Pure Chemical Industries, Ltd., Tokyo, Japan. Merck silica gel 60 (Art. 9385, 230–400 mesh, 0.040–0.063 mm, average pore diameter: 60 Å, surface area: 550 m² g⁻¹, pH: 7) was used as a substrate. The silica gel was heated in an electric oven at 110–120 °C for 12 h. A certified NO₂-loaded cylinder (10.2 ppm in N₂ or synthetic air, 20.1 ppm in N₂) was supplied by Takachiho Chemical Industry Inc., Tokyo, Japan, with 1-ppm NO₂ containing 2.50 × 10¹³ molecules cm⁻³ at 20 °C. NO₂ (2.0 and 0.20 ppm in N₂ or synthetic air) was prepared by diluting the NO₂ gas (10.2 ppm) with N₂ or synthetic air using a flowmeter, and their concentrations were determined using the Saltzman method (Saltzman, 1960).

2.2 Adsorption of pyrene on silica gel

Three suspensions of (a) pyrene (5.00 mg) + silica gel (5.00 g) in 100-mL chloroform, (b) pyrene (4.90 mg) + silica gel (25.00 g) in 200-mL chloroform, and (c) pyrene (4.90 mg) + silica gel (250 g) in 500-mL chloroform were prepared. The pyrene_ads was set to be proportional to the concentrations of NO₂ (10.2, 2.0, and 0.20 ppm). After magnetically stirring three suspensions for 12 h, the chloroform was slowly evaporated at 35 °C under a weakly reduced pressure. After the pyrene-adsorbed particles were gradually dried in a vacuum desiccator with an aspirator, 15.0 mg (a), 75.0 mg (b), and 750.0 mg (c) aliquots of three samples were supersonically extracted with acetonitrile (25 mL). The concentrations of pyrene were measured using high-performance liquid chromatography (HPLC) at λ = 234 nm. The concentration of the pyrene_ads was determined by comparing the HPLC-measured areas of three extracts to that of a calibration curve of pyrene. The concentrations of pyrene were calculated to be 0.600 mg L⁻¹ (a) and 0.588 mg L⁻¹ (b and c) when 100% of the pyrene was adsorbed on the silica gel. As the adsorption ratio of pyrene was 95.0% (a), 93.9% (b), and 95.1% (c), silica gels adsorbing 950 pg µg⁻¹ (a), 184 pg µg⁻¹ (b), and 18.6 pg µg⁻¹ (c) of pyrene were prepared. The surface coverage of pyrene (950 pg µg⁻¹) on silica gel (550 m² g⁻¹) was estimated 0.0067 because the cross section of the pyrene molecule was assumed to be 1.3 times that (1 nm²) of anthracene (Ma et al., 2011). This indicates that pyrene adsorbs on silica gel in a monolayer. The water content (H₂O_ads) of pyrene-adsorbed silica gel was gravimetrically determined to be in the range of 4.0–4.2 wt% as rehydration occurs in the work-up process. As the weight loss of the H₂O_ads on the pyrene-adsorbed silica gel was less than 0.1% after 5 days of storage in a desiccator kept at 3% humidity, the H₂O_ads is considered to be in adsorption equilibrium. Therefore, the H₂O_ads would not be decreased during the photonitration, although the humidity in the fluidized-bed reaction column was not measured when the pyrene-adsorbed silica gel is exposed to NO₂. The pyrene-adsorbed silica gel was stored in the dark to prevent the photodegradation of pyrene.

2.3 Photonitration

Pyrene was photonitrated by exposing NO₂ to pyrene_ads under three conditions: (a) NO₂ (10.2 ppm) + pyrene_ads (950 pg µg⁻¹, 4.70 μmol g⁻¹), (b) NO₂ (2.0 ppm) + pyrene_ads (184 pg µg⁻¹ = 0.91 μmol g⁻¹), and (c) NO₂ (0.20 ppm) + pyrene_ads (18.6 pg µg⁻¹ = 0.092 μmol g⁻¹).

2.4 Photoirradiation apparatus

Two fluidized-bed columns (Supplementary Materials Fig. S1) (Wang et al., 1999) irradiated by a 400-W high-pressure mercury lamp (UVL-400HA apparatus, Riko Co.) and a 500-W ultra-high-pressure mercury lamp (HX-500Q apparatus, WACOM Electric Co., Ltd.), with selectable irradiation wavelengths, were used separately. Both mercury lamps have emission lines of λ = 302, 313, 334, 365, 405, 436, 546, and 578 nm. The latter irradiation apparatus is the same as that reported except that a reaction cell was replaced with a solution filter, and a chemical actinometer was removed (Hasegawa et al., 1998). The high-pressure mercury lamp and the ultra-high-pressure mercury lamp are hereafter denoted as the HP-Hg lamp and the UHP-Hg lamp, respectively. The three pyrene-adsorbed silica gels (1.00 g) were added to the three columns separately, and NO₂ (10.2, 2.0, and 0.20 ppm in N₂ or air) flowed through the bottom of the column at a flow rate of 400 mL min⁻¹ and 200 mL min⁻¹ for the HP-Hg lamp and the UHP-Hg lamp, respectively. The decrease in flow rate for the UHP-Hg lamp is due to the restriction of the irradiation area by a light flux 50 mm in diameter. The fluidized-bed column was set 16 cm away from the HP-Hg lamp and 6 cm away from the edge of the UHP-Hg lamp housing. Both Hg lamps were used as simulated sunlight; natural sunlight (from 7:30 to 15:30 in September 2002, Toyama, Japan) was also used.

The 400-W HP-Hg lamp was used to examine the effect of light intensity on the photoreaction of pyrene with NO₂. The light intensity was adjusted by surrounding the HP-Hg lamp with a thin aluminum plate drilled with different bore sizes. The transmittance (T) through the thin aluminum plates was measured with a UV spectrophotometer and found to be T = 21, 27, and 48%. Therefore, the irradiance on the fluidized-bed columns was 21, 27, and 48% that emitted from the HP-Hg lamp. The 500-W UHP-Hg lamp was used to examine the effects of wavelength and oxygen concentration in the dilution gas on the photoreaction of pyrene with NO₂. The collimated beam 50 mm in diameter was irradiated in parallel on wavelength-selective absorption filters. A Toshiba UV-D36A band-pass filter (300 nm < λ < 380 nm) and a Pyrex filter (λ > 290 nm) were used to select the light-absorption bands of pyrene (302, 313, and 334 nm). A solution filter (0.25 M NaNO₃ + 0.10 M NaNO₂, 1.5 cm thick) and a Pyrex filter (λ > 290 nm nm) were used to select the wavelength longer than 370 nm. The light intensities irradiated from two mercury lamps and sunlight were measured at the surface of the fluidized-bed column using an ultraviolet intensity meter (UV-M02, ORC Manufacturing Co. Ltd., Japan), which can measure the irradiance (mW/cm²) at the ranges of 241–271 nm (UV25, max: 250 nm), 320–390 nm (UV35, max: 350 nm), and 330–490 nm (UV42, max: 420 nm). Experiments were also carried out without photoirradiation in a shaded draft.

2.5 Analysis

The NO₂ flow was stopped at an appropriate time, and 0.015 g of silica gel was taken out. Pyrene and products on silica gel were extracted with 5 mL of acetonitrile under ultrasonic irradiation under the photonitration condition (a). Under photonitration conditions (b) and (c), they were extracted with 5 mL (1 mL of the internal standard solution of chrysene: 0.025 mM in acetonitrile + 4 mL of acetonitrile) and 2 mL (1 mL of the internal standard solution of chrysene: 0.0050 mM in acetonitrile + 1 mL of acetonitrile) of acetonitrile solution under ultrasonic irradiation, respectively. These suspensions were centrifuged, and the supernatant solutions (extracts) were used for HPLC analyses. C₀ is the concentration of pyrene in the extract at 0 min, and C_t is the concentrations of pyrene and products in the extracts at t min or t (h), calculated using external standard calibration curves (experiment a) or internal standard calibration curves (experiments b and c) at 234 nm. Although an internal standard method was not used in the experiment (a), almost the same recoveries (C₀/theoretical initial concentrations) for pyrene (97.1%) and 1-nitropyrene (96.4%) do not affect the ratios of C_t/C₀. After exposure to NO₂, the products on the silica gel remaining in the column were extracted with 5 mL of chloroform under ultrasonic irradiation. After centrifugation, the concentrated supernatant solution (extracts) was used for the analyses by gas chromatography (GC) and GC-mass spectrometry (GC-MS). GC and GC-MS analyses were repeated without adding the internal standard substance to the extracts. The HPLC, GC, and GC-MS instruments and analytical conditions are described in supporting materials. The experimental method of fluorescence quenching is described in supporting materials. Fluorescence spectra were recorded using a HITACHI 650-10S fluorescence spectrophotometer.

3 Results and discussion

3.1 Effects of light intensity, wavelength, and oxygen concentration in NO₂ diluent on the decay of pyrene

Figure 1 shows the effect of light intensity on the decay of pyrene by exposing NO₂ (10.2 ppm in N₂) under irradiation (T = 100% and 21%) by the HP-Hg lamp with the decay of pyrene under sunlight and in the dark. Sunlight was used without adjusting the light intensity, although the sunlight irradiance on pyrene varied with time during the 8-h irradiation. The decay rate of pyrene increased as the light intensity increased. The decay rate of pyrene under light irradiation greatly contrasted with that in the dark, i.e., the pyrene decay proceeded exponentially under the HP-Hg light (T = 100%) and sunlight, while it proceeded sigmoidally in the dark as already reported for the H⁺-autocatalyzed reaction, in which the protons formed by the dissociation of HNO₃ started the nitration and the protons released by the formation of 1-nitropyrene acted as an autocatalyst to accelerate the nitration (Hasegawa et al., 2020). When pyrene was irradiated with 21% (T = 21%) of the light emitted from the HP-Hg lamp, it decayed exponentially before 180 min and sigmoidally after 180 min. These results indicate that the nitration mechanism of pyrene changes with light intensity. As the photolysis of pyrene was only 4.2% for 480 min under the irradiation condition of T = 100% in the absence of NO₂, the effect of the photolysis on the photonitration of pyrene with NO₂ is small. 1-Nitropyrene was obtained as a major product under light irradiation.

Fig. 1

Effect of light irradiation on the reaction of pyrene with 10.2-ppm NO₂ in N₂ (○, △, ☐, ●) and photolysis of pyrene in the absence of NO₂ (◇). HP-Hg lamp (○, ◇: T = 100%, ☐: T = 21%), sunlight (△), in the dark (●). Light irradiance of the HP-Hg lamp (mW cm⁻²) at T = 100%: UV 25: 0.29, UV 35: 2.60, UV 42: 4.80, sunlight: UV 25: 0.20 (7:30), 0.75 (9:30), 1.20 (11:30), 1.21 (13:30), 0.51 (15:30), UV 35: 0.43 (7:30). 1.55 (9:30), 2.21 (11:30), 1.71 (13:30), 0.68 (15:30), UV 42: 1.43 (7:30), 5.61 (9:30), 8.09 (11:30), 6.52 (13:30), 2.60 (15:30)

Figure S2 (supplementary materials) shows the relationship between ln(C_t/C₀) and reaction time under irradiation (T = 100%) by the HP-Hg lamp. As the decay deviated from the pseudo-first-order kinetics after 240 min, the observed rate constant (k_obs) was determined to be 9.0 × 10⁻⁵ s⁻¹ for 240 min (C_t/C₀ = 25.7%, which was 2.0 times greater than the k_obs (4.6 × 10⁻⁵ s⁻¹) in the rapid-decay range in the dark (Fig. 1), 5.3 times greater than the k_obs (1.7 × 10⁻⁵ s⁻¹) of pyrene on dry quartz with 3-ppm NO₂ in the dark (Kameda et al., 2016), and 3.0 times greater than the k_obs (3 × 10⁻⁵ s⁻¹) of pyrene on silica particles with 0.055-ppm NO₂ in the dark (Perraudin et al., 2005). This indicates that the nitration of pyrene adsorbed on silica gel was enhanced under light irradiation. The k_obs in the rapid decay range in the dark was determined by drawing a tangent at the inflection time (350 min, C_t/C₀ = 0.50) as detailed in our previous report (Hasegawa et al., 2020).

Figure S3 (supplementary materials) shows the effect of wavelength on the nitration of pyrene with NO₂ (10.2 ppm in N₂) under irradiation by the UHP-Hg lamp. The flow rate of NO₂ was lowered from 400 (HP-Hg lamp) to 200 mL min⁻¹ so that all suspended pyrene-adsorbed silica gels can be irradiated by the light flux. Pyrene decayed according to the pseudo-first-order kinetics under the irradiation of 300–380 nm, which involves absorption bands (302, 313, and 334 nm) of pyrene. Under irradiation longer than 370 nm, which pyrene does not absorb, the decay of pyrene was largely suppressed, and an induction period (420 min) was observed, although a sigmoidal curve was not observed due to the short reaction time. This result indicates that the photonitration under visible light proceeds via the H⁺-autocatalyzed nitration mechanism. Figure S4 (supplementary materials) shows the relationship between ln(C_t/C₀) and reaction time under the irradiation of 300–380 nm. As the decay deviated from the pseudo-first-order kinetics after 240 min, the k_obs was determined to be 6.0 × 10⁻⁵ s⁻¹ for 240 min (C_t/C₀ 40.5%). The deviation of pyrene decay from the pseudo-first-order kinetics in the final stage (Figs. S2 and S4) was also observed in the reaction of NO₂ with pyrene and anthracene adsorbed on silica gel (Hasegawa et al., 2020), with anthracene adsorbed on SiO₂ and MgO (Ma et al., 2011), and with pyrene adsorbed on clay minerals and Fe₂O₃ (Kameda et al., 2016). The causes are unclear at present.

Figure S5 shows the effect of oxygen concentration in the dilution gas on the reaction of pyrene with NO₂ (10.2 ppm) under irradiation (λ = 300–380 nm) by the UHP-Hg lamp. The pyrene decay decreases as the oxygen concentration increases, because oxygen quenches the photoexcited pyrene, which is discussed in supporting materials.

3.2 Effect of light intensity on the formation of 1-nitropyrene

Figure 2 shows the effect of light intensity on the formation of 1-nitropyrene under irradiation by the HP-Hg lamp with the photolysis of 1-nitropyrene in the absence of NO₂. The concentration of 1-nitropyrene increased linearly for 60 min under the light intensity of T = 100%, 48%, 27%, and 21%. Under the light intensity of T = 100% and 48%, the concentration of 1-nitropyrene exponentially increased between 60 and 360 min and then leveled off. Under the light intensity of T = 27%, it increased exponentially between 60 and 240 min and then gradually increased. Under the light intensity of T = 21%, it abruptly increased after 300 min via the exponential increase in the concentration of 1-nitropyrene and then leveled off. The abrupt increase in the concentration of 1-nitropyrene is considered to be caused by the accelerated decay of pyrene between 300 and 360 min, as shown in Fig. 1 (T = 21%). The accelerated decay of pyrene is also observed after 300 min in the dark, as shown in Fig. 1. Therefore, the H⁺-autocatalyzed formation of 1-nitropyrene becomes dominant under weak light intensity. Figure 2 also suggests that the decay of 1-nitropyrene would occur in two stages: between 60 and 360 min and after 360 min. The yield of 1-nitropyrene is lower under any light intensity than that (99%) observed when pyrene was autocatalytically decayed in the dark (Hasegawa et al., 2020). It reached 63% (T = 21%), 51% (T = 27%), 46% (T = 48%), and 42% (T = 100%) at 480 min, indicating that the increase in light intensity decreases the yield of 1-nitropyrene. Under 480-min exposure of pyrene to 10.2-ppm NO₂ in air (T = 100%), the yield of 1-nitropyrene was 28%. As the photolysis decay of 1-nitropyrene was 10% for 480 min in the absence of NO₂, the decrease in 1-nitropyrene is not ascribed to the photolysis of 1-nitropyrene.

Fig. 2

Effect of light intensity on the formation of 1-nitropyrene (○, △, ◇, ☐) and the photolysis of 1-nitropyrene in the absence of NO₂ (■). NO₂ (10.2 ppm in N₂) and HP-Hg lamp (T = 100%). ○: T = 100%, △: T = 48%, ◇: T = 27%, ☐: T = 21%, ●: in the dark. Light irradiance (mW cm⁻²) at T = 100%: UV 25: 0.29, UV 35: 2.60, UV 42: 4.80

3.3 Reaction products under light irradiation

Figure 3a–f shows the GC chromatograms of the extracts obtained by exposing NO₂ (10.2 ppm in N₂) to pyrene for 480 min under various irradiation conditions: (a) T = 21%, (b) T = 48%, (c): 300 nm < λ < 380 nm (absorption band of pyrene), (d) λ > 370 nm, (e) sunlight, and (f): T = 100%. GC–MS measurement of the extracts (a–f) was also carried out. Table 1 shows the molecular and fragment ions of the products. The content (%) of the products, which was determined based on the peak areas of each GC chromatogram, is shown near the peaks. Product and product distribution changed depending upon the light intensity and wavelength. Pyrene (1), 1-nitropyrene (2), and 1,3- (3), 1,6- (4), and 1,8-dinitropyrenes (5) were observed with nitrated photooxidation products (6, 7, 8, 9, and 10) and the photooxidation products of 1. The structures of 2–5 were confirmed by matching the retention times and the mass spectra with those of the standard samples. Although 6–8 and 9–10 probably have pyrenedione skeletons with one and two NO₂ groups, respectively, from the MS fragmentation patterns, their structures could not be identified because a remaining oxygen-containing group is unclear. In the GC chromatogram of 3a, 2 and 3–5 are observed with 9 and 10. The total content of 3–5 decreased, and that of 9 and 10 increased as the light intensity increased from T = 21 to 100% (Fig. 3a, b, and f). In Fig. 3b, the contents of 3–5 decreased, and the small amounts of 6 and 8 appeared, which already appeared as a shoulder peak of peak 3 in Fig. 3a. In the GC chromatograms of Fig. 3e and f, 3–5 are not observed, while new peaks (6–8) are observed with 2, 9, 10, and the small amounts of oxidation products (A, B) of 1. Under pyrene absorption light, 2 and 8 are observed (Fig. 3c), while under visible light irradiation, 2 is observed with the small amounts of 3–5 and trace amounts of 9 and 10 (Fig. 3d). The formation of 3–5 and 9 and 10, which were not formed in the dark reaction of 1 with NO₂, is probably because 2.5% of the UHP-Hg light of 330 nm penetrated the solution filter. When the diluent of NO₂ was changed from N₂ to air under sunlight irradiation for 480 min, the contents of 1 increased and those of 2 and 6–10 decreased as follows: 1 (1.6% → 49.3%), 2 (48.3% → 42.7%), 6–8 (39.0% → 5.3%), and 9–10 (4.3% → 0%).

Fig. 3

GC chromatograms (a–f) of the extracts (a–f) obtained by exposing NO₂ (10.2 ppm in N₂) to pyrene for 480 min under various irradiation conditions

Table 1 Molecular and fragment ions of the products contained in the extracts (a–f) obtained under light irradiation

Figure S7 shows HPLC chromatograms of the extract a (T = 21%) to examine the time courses of 3–5, which are considered to be the precursors of 9 and 10. Nine and 10 increase after 120 min, and 6–8 increase after 360 min, while 3–5 are observed at 480 min in the elution order of 5, 4, and 3 before the peak of 2. These results indicate that the formation of 9 and 10 would proceed rapidly without accumulating 3–5, and they would be observed in the final slow-reaction period. The HPLC chromatograms of the extracts b and f showed that the peaks of 3–5 disappeared at 480 min, when the light intensity increased from T = 48 to 100%.

In the GC–MS chromatograms of the extracts (e and f), two very small peaks of A1 and A2 (total yields: e, 0.8%; f, 0.4%) and B (yields: e, 4.3%; f, 2.2%) were observed before peak 2 at the following retention times: A1, 30.8 min; A2, 31.6 min; and B, 33.6 min. A1, A2, and B are assumed to be dihydroxypyrenediones (A1), a photooxidation product of pyrenedione (A2), and dihydroxynitropyrene (B), as shown in Table 1; however, their structures could not be identified because standard and synthesized chemicals were not available. 1-Hydroxypyrene (Reyes et al., 2000; Garcia-Berrios & Arce, 2012), 1,6- and 1,8-pyrenediones (Reyes et al., 2000; Garcia-Berrios & Arce, 2012), and 1-hydroxy-x-nitropyrenes (Kameda, et al., 2011; Garcia-Berrios & Arce, 2012), which were formed by irradiating 1 on silica gel (Reyes et al., 2000) and 2 in organic solvents (Kameda, et al., 2011; Garcia-Berrios & Arce, 2012), were not observed because the photonitration of 1 and 2 became dominant over the photooxidation of 1 and 2.

3.4 Time course of the products

Figure S8 (supplementary materials) shows the HPLC chromatograms of the extracts obtained every 60 min under irradiation by the HP-Hg lamp (T = 100%). The HPLC chromatograms obtained by 420-min and 480-min irradiation are similar after 10 min to the GC chromatogram (Fig. 3f) obtained by 480-min irradiation (T = 100%). The intensities of peaks 9 and 10 increase from 60 to 300 min, while the intensities of peaks 6–8 increase from 360 to 480 min with the decrease in 9 and 10. The dependence of appearance order of the products 9–10 and 6–8 on the light intensity described in Fig. 3 expects that peaks 9 and 10 appear before the appearance of peaks 6–8 under the light irradiation of T = 100%. Furthermore, three peaks of 6, 7, and 8, in which peaks 6 and 7 probably overlapped in the HPLC chromatograms, and two peaks of 9 and 10 are observed in both chromatograms of GC and HPLC. The assignment of numbers to the five peaks was made according to the order of elution and similarity of the peak relative intensity to that of the peaks 6–10 in the GC chromatogram of Fig. 3f. Based on these observations, peaks 6–8 were regarded as the products 6, 7, and 8, and peaks 9 and 10 were regarded as the products 9 and 10. Peaks 1 and 2 were identified as 1 and 2 from the matching of the retention times with standard samples.

Figure 4 shows the time course of 1 and the products, where their response factors at 234 nm were regarded as the same. The formation rate of 2 decreases from the pseudo-first-order increase between 60 and 360 min with the increase in 9 and 10 (path A) and levels off with the increase in 6, 7, and 8 (path B). Products 3–5 were not observed under strong light irradiation (T = 100%) as shown in Fig. 3f. These results indicate that the photonitration of pyrene would proceed via the paths A and B as shown in Scheme 1. The nitration of 2 (path A) proceeds faster than the oxidation of 2. In path A, 3, 4, and 5 formed are rapidly oxidized to yield dinitropyrenediones (MW = 322), which are rapidly oxidized to give 9 and 10 (MW = 336). Although small amounts of 3, 4, and 5 were observed under weak light irradiation (T = 21% and T = 48%), photosensitive dinitropyrenediones would not be observed as intermediates. In path B, 2, which is less photosensitive than 3, 4, and 5, would form 6, 7, and 8 (MW = 291) via nitropyrenediones (MW = 277). Photosensitive nitropyrenediones would not be observed as intermediates.

Fig. 4

Time course of the products obtained under the irradiation of the HP-Hg lamp (T = 100%). 10.2 ppm NO₂ in N₂. ○: pyrene, ●: 2 (1-nitropyrene), ■, ▲: 9 and 10 (photooxidation products of dinitropyrenes), ◆: 6, 7 (photooxidation products of 1-nitropyrene), ◇: 8 (photooxidation products of 1-nitropyrene)

Scheme 1

Photonitration path of pyrene

Subtracting the sum of the C_t/C₀ of 2, 6, 7, 8, 9, and 10 at 480 min from 1, the mass balance was 74%. Guan et al. (2017) reported that 1-nitropyrene photolyzes on soot to yield HNO₂ and NO. We reported that large amounts of HNO₃ (24 μmol g⁻¹) and HNO₂ (approximately 11 μmol g⁻¹) were formed on silica gel by exposing 9.77 ppm NO₂ for 480 min to pyrene-adsorbed silica gel (4.95 μmol g⁻¹) without photoirradiation (Hasegawa et al. 2020; Wang et al. 2000). The experiment to determine the much less amounts of volatile HNO₂, which could be formed by the photodecomposition of 2 and 6–10, did not give reliable results. Although the photodecomposition of 2–10 to gaseous organic products is also conceivable, as reported for pyrene on Al₂O₃ (Romanias et al. 2014), their measurements involved in the NO₂ effluent were not performed.

3.5 Photonitration mechanism

The following reaction mechanism is proposed to explain the photonitration of pyrene with NO₂.

$$\mathrm{pyrene}+hv\left(290-380\mathrm{ nm}\right) \to {}^{1}\mathrm{pyrene}^*$$

(1)

$$\begin{array}{cc}{}^{1}\mathrm{pyrene}^*\to {}^{3}\mathrm{pyrene}^*& (\mathrm{ISC})\end{array}$$

(2)

$${}^{3}\mathrm{pyrene}^*+{}^{3}{\mathrm{O}}_{2}\to \mathrm{pyrene}+{}^{1}{\mathrm{O}}_{2}$$

(3)

$${}^{1}\mathrm{pyrene}^*+{}^{3}{\mathrm{O}}_{2}\to {\mathrm{pyrene}}^{\bullet+}+{\mathrm{O}}_{2}^{\bullet-}$$

(4)

$${}^1\mathrm{pyrene}^\ast+{}^\bullet\mathrm{NO}_2\rightarrow\left(\mathrm{pyrene}^{\bullet+}-\mathrm{NO}_2^-\right)\rightarrow1-\mathrm n\mathrm i\mathrm t\mathrm r\mathrm o\mathrm p\mathrm y\mathrm r\mathrm e\mathrm n\mathrm e\,(\mathit2)$$

(5)

$$\left(1-\mathrm{nitropyrene}\right)^*+{}^{\bullet}{\mathrm{NO}}_{2}\to \left({1-\mathrm{nitropyrene}}^{\bullet+}-{\mathrm{NO}}_{2}^{-}\right)$$

(6)

$$\left({1-\mathrm{nitropyrene}}^{\bullet+}-\mathrm{NO}_2^-\right)\rightarrow\mathrm d\mathrm i\mathrm n\mathrm i\mathrm t\mathrm r\mathrm o\mathrm p\mathrm y\mathrm r\mathrm e\mathrm n\mathrm e\mathrm s\,(\mathit3,\,\mathit4,\,\mathit5)$$

(7)

$$1-\mathrm{nitropyrene}^*\mathrm{and\, dinitropyrenes}^*+{}^{3}{\mathrm{O}}_{2}\to 1-{\mathrm{nitropyrene}}^{\bullet+} \mathrm{and\,} {\mathrm{dinitropyrenes}}^{\bullet+}+{\mathrm{O}}_{2}^{\bullet-}$$

(8)

$$1-\mathrm{nitropyrene}^{\bullet+}\mathrm a\mathrm n\mathrm d\,\mathrm{dinitropyrenes}^{\bullet+}+{\mathrm H}_2\mathrm O\rightarrow\mathrm o\mathrm x\mathrm i\mathrm d\mathrm a\mathrm t\mathrm i\mathrm o\mathrm n\,\mathrm p\mathrm r\mathrm o\mathrm d\mathrm u\mathrm c\mathrm t\mathrm s\,(\mathit6\mathit-\mathit{10})$$

(9)

Pyrene absorbs light in the range of 290–380 nm to produce a singlet state (¹pyrene*) of pyrene (Eq. 1). When the diluent of NO₂ was changed from N₂ to air, the energy transfer from ¹pyrene* to ³O₂ occurred via the intersystem crossing (ISC) of ¹pyrene* to the triplet state (³pyrene*) of pyrene (Eqs. 2 and 3). The decrease in the concentration of ¹pyrene* can explain the inhibition effect of oxygen on the pyrene decay, as shown in Fig. S5. On the other hand, a pyrene radical cation (pyrene^•+) would be formed by the electron transfer from ¹pyrene* to ³O₂ (Eq. 4). Equations 2–4 can be also supported by the photolysis mechanism of pyrene adsorbed on alumina (Pankasem & Thomas, 1991), silica (Mao & Thomas, 1995; Reyes et al., 2000), and KNO₃ films (Ammar et al., 2010).

The addition of ^•NO₂ to ¹pyrene* would generate the intermediate (pyrene^•+-NO₂⁻), which is followed by the formation of 1-nitropyrene (Eq. 5). Since the concentration of exposed ^•NO₂ is constant and the concentration of ¹pyrene* is considered to be in a steady state depending on the NO₂ diluents, we deduce that pyrene decayed according to the following pseudo-first-order kinetics.

$$-\mathrm{d}[\mathrm{pyrene}]/\mathrm{d}t = {k}_{\mathrm{obs}} [\mathrm{pyrene}]$$

(10)

Dinitropyrenes (3, 4, 5) would be formed via intermediate (1-nitropyrene^•+-NO₂⁻), which was formed by the addition of ^•NO₂ to the 3, 6, and 8 carbon atoms of the pyrene ring (Eqs. 6 and 7). As 1-nitropyrene and dinitropyrenes have a wide absorption band at 320–440 nm, they can absorb the emission lines (365 and 436 nm) from the HP-Hg lamp. Ammar et al. (2010) studied the photoenhanced conversion of NO₂ on pyrene/KNO₃ films to HNO₂, NO, NO₂⁻, and traces of 1-nitropyrene and postulated the species of (pyrene^•+-NO₂⁻), which is formed by the reaction of pyrene* with NO₂. Brigante et al. (2008) also suggested a nitration mechanism via radical cations for the reaction of NO₂ with pyrene films. We considered that the radical cation intermediate (pyrene^•+-NO₂⁻) proposed by Ammar et al. (2010) would also act in the photonitration of pyrene. It has been reported that ^•NO₂ generated by the photoirradiation of oxygenated 1-nitronaphthalene + NO₂⁻/HNO₂ (Maddigapu et al., 2011) and phenol + NO₃⁻/NO₂⁻ solutions (Vione et al., 2001, 2005; Bedini et al., 2012) would react with excited 1-nitronaphthalene and phenol to yield dinitronaphthalene and nitrophenol isomers via the addition of ^•NO₂ to 1-nitronaphthalene* and phenol* and subsequent H-atom abstraction by oxygen. At present, we consider that the radical cation mechanism is more probable than the NO₂ addition mechanism because the formation of the radical cation supports the formation of photooxidation products (Ammar et al., 2010).

The photooxidation products (6–10) would be formed by the reaction of the radical cations of 2–5 with water adsorbed on silica gel (Eq. 9) as reported for the photolysis of pyrene on silica gel (Reyes et al. 2000). The origin of the O₂ for the formation of the radical cations can be illustrated by the following reports. Mao and Thomas (1994) reported that the photoirradiation of degassed pyrene on silica gel induced an electron transfer showing pyrene^•+. They also reported that adsorbed oxygen on silica gel served as an electron trap of excited pyrene to form O₂^•− and pyrene^•+. Reyes et al. (2000) reported that the photoexcitation of pyrene adsorbed on air-saturated silica gel produced pyrene^•+ under N₂. As pyrene-adsorbed silica gel is saturated with air in our experiment, the photooxidation products (6–10) would be produced under the N₂ diluent.

3.6 Photonitration of pyrene with low concentration NO₂

Figure 5a–d shows the time courses of the photonitration of pyrene with NO₂ under the photonitration conditions of (b) and (c), in which the concentration ratio of NO₂ to pyrene_ads is kept constant, with nitration in the dark and the photolysis of pyrene under N₂ and air diluents. The concentrations of NO₂ and pyrene_ads are shown on the tops of Fig. 5a–d. The time courses of pyrene and the products were determined via HPLC measurements. The products extracted after 24, 48, and 60 h were analyzed via GC-MS measurement, and the yields were estimated by comparing their GC peak areas.

Fig. 5

Nitration of pyrene with 2.0 and 0.20 ppm NO₂ under the HP-Hg lamp (T = 100%) (○, ☐, ◇) and in the dark (●, ■) and photolysis of pyrene in the absence of NO₂ (△). ○,●: 1 (pyrene), ☐, ■: 2 (1-nitropyrene), ◇: 8–10 (photooxidation products)

Under 2.0-ppm NO₂ in N₂ (Fig. 5a), 98.4% of pyrene decayed for 24 h, which is three times longer than 8 h under 10.2-ppm NO₂, according to the pseudo-first-order kinetics. The yields of 2 (21.6%) and 8–10 (8.0%) for 24 h were lower than those of 2 (42.2%) and 6–10 (31.4%) produced for 8 h under 10.2-ppm NO₂ (Fig. 4), suggesting that 2 and 8–10 could be easily photodecomposed when pyrene_ads decreased from 950 pg µg⁻¹ (coverage: 0.0067) to 184 pg µg⁻¹ (coverage: 0.0013). Under 2.0-ppm NO₂ in air (Fig. 5b), the decay rate of pyrene was decreased by the inhibition effect of oxygen, resulting in a further decrease in the yields of 2 (4.9%) and 6–10 (3.8%) for 24 h. The photolysis rate of pyrene increased by 4.8 times that in 10.2-ppm NO₂ for 8 h. Fifty-six percent of the pyrene, of which 24% decayed due to photolysis for 24 h, decayed according to near the pseudo-first-order kinetics. Small amounts of 1,6-pyrenedione (MW = 232, 0.4% in N₂ and 0.2% in air), which were identified by matching the retention time and mass spectrum with those of a standard sample, and two unidentified-structure products (MW = 230 and 246, 0.6% in N₂) also formed. In the dark, the sigmoid curves of the decay of 1 and the formation of 2 were very similar in Fig. 5a and b and even replaced N₂ in the dilution gas with air because they were not influenced by the oxygen concentration in the dark reaction.

Under 0.20-ppm NO₂ in N₂ (Fig. 5c), 89% of pyrene, of which 27% decayed due to photolysis, decayed for 24 h according to near the pseudo-first-order kinetics. The decay curve of pyrene appears to deviate downward from the first-order decay after 12 h because it involves the decay by its own photolysis. The yields of 2 and 8 were 2.3% and 3.4% for 24 h, respectively. Under 0.20-ppm NO₂ in air (Fig. 5d), the decay of pyrene was inhibited by oxygen in the air. Sixty percent of pyrene, of which 40% decayed due to photolysis, decayed for 24 h according to near the pseudo-first-order kinetics. The decay curve of pyrene appears to deviate downward from the first-order decay after 12 h because it involves the decay by own photolysis. The yields of 2 and 8 for 24 h were 2.1% and 0.9%, respectively. In the dark, the yields of 2 in N₂ and air diluents for 60 h were 6.0% and 4.0%, respectively.

Figure S9 summarizes the concentration changes in pyrene by exposure to 10.2, 2.0, and 0.20 ppm NO₂ in N₂ or air diluent. Under 10.2- and 2.0-ppm NO₂ in N₂, the decay rate of pyrene decreased according to the pseudo-first-order reaction shown in Eq. 10, and the half-lives of pyrene decay were estimated to be 1.8 h under 10.2-ppm NO₂ and 6.0 h under 2.0-ppm NO₂. Under 2.0-ppm NO₂ in air, the half-life was estimated to be 20.9 h. Under 0.20-ppm NO₂, the half-lives were estimated to be 12.0 h in N₂ and 18.2 h in air. If the NO₂ concentration is diluted tenfold from 2.0 to 0.20 ppm under constant pyrene_ads of 184 pg μg⁻¹, the pyrene decay rate under 0.20-ppm NO₂ should be significantly decreased, resulting in the significant extension of its half-life. However, in N₂ diluent, the half-life under 0.20-ppm NO₂ was 2.0 times longer than that under 2.0-ppm NO₂, and in air diluent, it was 0.87 times shorter than that under 2.0-ppm NO₂. These results can be explained by the following two reasons. First, the concentration of pyrene_ads was set to be proportional to the NO₂ concentration because the concentrations of NO₂ and pyrene_ads on heavy traffic road are considered to be positively correlated. Second, the photodecomposition of pyrene was easily occurred with the decrease in the concentration of NO₂ and in the presence of air. Therefore, the photonitration of pyrene under 0.20-ppm NO₂ proceeded faster than that expected from the dilution ratio of NO₂.

Romanias et al. (2014) have reported that gaseous products are formed by the photodegradation of pyrene on Al₂O₃ in the presence of oxygen. It has been reported that 1-nitropyrene formed on soot is photolyzed with the formation of HNO₂ and NO (Guan et al., 2017), and it is also photolyzed to form HNO₂ and NO with the consumption of nitro groups and the formation of aromatic C = O groups and phenolic hydroxyls (You et al., 2022). Attempts to confirm the photodecomposition products were hitherto unsuccessful. Repeated experiments are necessary to examine the variability of the measurement due to low concentrations of NO₂ and pyrene_ads.

4 Conclusions

The photonitration of pyrene adsorbed on silica gel with NO₂ (10.2, 2.0, and 0.20 ppm in N₂ or air) was studied. Under 10.2- and 2.0-ppm NO₂ in N₂ and air, the concentration of pyrene decreased exponentially in accordance with the pseudo-first-order reaction under UV-light irradiation, while in the dark, it decreased sigmoidally in accordance with the H⁺-autocatalyzed reaction. Nitration products, their photooxidation products, and both time profiles were examined under 10.2 ppm NO₂ in N₂. The yields of formed 1-nitropyrene and minor dinitropyrenes were decreased by the transformation into their photooxidation products. The yield of 1-nitropyrene was 42% in N₂ and 28% in air for 8 h. The inhibition effect of oxygen can be explained by the energy transfer from ¹pyrene* to oxygen. Radical cation intermediate (pyrene^•+-NO₂⁻) was proposed for 1-nitropyrene formation. Under 24-h exposure of pyrene to 2.0-ppm NO₂, the yields of 1-nitropyrene and the photooxidation products were 21.6% and 8.0%, respectively, in N₂ and 4.9% and 3.8%, respectively, in air. In the photonitration of pyrene under 0.20-ppm NO₂, the photolysis of pyrene concomitantly occurred especially in air diluent. Under 24-h exposure of pyrene to 0.20-ppm NO₂, the yields of 1-nitropyrene and the photooxidation products were 2.3% and 3.4%, respectively, in N₂ and 2.1% and 0.9%, respectively, in air. The significant decrease in the yields of 1-nitropyrene and the photooxidation products under the concentration of 0.20-ppm NO₂ can be explained by their easy photodecomposition with the increase in the photolysis of pyrene. Under the concentration of 0.20-ppm NO₂ in air, which is approximately the concentration on heavy-traffic roads, the decay rate of pyrene by the photonitration was increased by own photolysis, although the photonitration was inhibited by oxygen in air. Further studies are necessary to identify the correct structures of the photooxidation products and to examine the photodecomposition products of the nitration and the photooxidation products.