Characterization of Atmospheric Fine Particles and Secondary Aerosol Estimated under the Different Photochemical Activities in Summertime Tianjin, China

In order to evaluate the pollution characterization of PM2.5 (particles with aerodynamic diameters less than or equal to 2.5 μm) and secondary aerosol formation under the different photochemical activity levels, CO was used as a tracer for primary aerosol, and hourly maximum of O3 (O3,max) was used as an index for photochemical activity. Results showed that under the different photochemical activity levels of L, M, LH and H, the mass concentration of PM2.5 were 29.8 ± 17.4, 32.9 ± 20.4, 39.4 ± 19.1 and 42.2 ± 18.9 μg/m3, respectively. The diurnal patterns of PM2.5 were similar under the photochemical activity and they increased along with the strengthening of photochemical activity. Especially, the ratios of estimated secondary aerosol to the observed PM2.5 were more than 58.6% at any hour under the photochemical activity levels of LH and H. The measured chemical composition included water soluble inorganic ions, organic carbon (OC), and element carbon (EC), which accounted for 73.5 ± 14.9%, 70.3 ± 24.9%, 72.0 ± 21.9%, and 65.8 ± 21.2% in PM2.5 under the photochemical activities of L, M, LH, and H, respectively. Furthermore, the sulfate (SO42−) and nitrate (NO3−) were nearly neutralized by ammonium (NH4+) with the regression slope of 0.71, 0.77, 0.77, and 0.75 between [NH4+] and 2[SO42−] + [NO3−]. The chemical composition of PM2.5 was mainly composed of SO42−, NO3−, NH4+ and secondary organic carbon (SOC), indicating that the formation of secondary aerosols significantly contributed to the increase in PM2.5. The formation mechanism of sulfate in PM2.5 was the gas-phase oxidation of SO2 to H2SO4. Photochemical production of nitric acid was intense during daytime, but particulate nitrate concentration was low in the afternoon due to high temperature.


Introduction
Since the serious haze incidents in 2013 in China, the regional atmospheric environment pollution, characterized by fine particulate matter (PM 2.5 ; particles with aerodynamic diameters less than or equal to 2.5 micrometers (µm)), has attracted widespread attention, and serious haze pollution has endangered the public health [1][2][3][4]. In order to alleviate the severe situation of atmospheric environment pollution and effectively improve air quality, a series of control policies have been formulated, such as the Air Pollution Prevention and Control Action Plan in China [5]. Through the cooperative prevention and strict implementation of control policies in cities or regions, the mass concentration of PM 2.5 has decreased on average in some representative regions, for example the annual mean mass concentration of PM 2.5 is less than 35 µg/m 3 in the Pearl River Delta (PRD) region, but the mass concentration of PM 2.5 is still high in the Beijing-Tianjin-Hebei (BTH) and Yangtze MA, USA), O 3 is measured using a UV photometric O 3 analyzer (Model 49C/I, Thermo-Fisher Scientific, Waltham, MA, USA), and CO is measured with a nondispersive infrared analyzer (Model 48i, TE). The data observed by the Tianjin monitoring site (39.08 • N, 117.21 • E) in the network is used from 1 June to 31 August 2020.
The continuous water-soluble inorganic ions and organic carbon (OC) and elemental carbon (EC) are measured on the top of the fourth floor (39 • 6 N, 117 • 6 E) in Tianjin Chengjian University, about 15 m from the ground. The main traffic line of Jinjing Road is the south and about 200 m away the sampling point, and the other surrounding areas are mainly residential communities. The water-soluble inorganic ions (i.e., SO 4 2− , NO 3 − , Cl − , NH 4 + , Na + , K + , Ca 2+ , and Mg 2+ ) in PM 2.5 are determined during the same observation period by the ambient ion monitor instrument (AIM, URG Corporation, Chapel Hill, NC, USA, CarURG9000B). OC and EC are measured Semi-Continuous OCEC instrument (Model 4, Sunset Laboratory Inc., Tigard, OR, USA). An automatic meteorological monitoring instrument (Milos520, Vaisala, Vantaa, Finland) is used to record the main meteorological parameters, including temperature, humidity, wind speed, and wind direction.
The PM 2.5 data are initially available based on 5 min averages. The hourly means are calculated using a minimum of nine 5 min averages, and the daily means are calculated using a minimum of 18 1 h averages; otherwise, the hourly and daily value is considered missing. The other daily data (gaseous pollutants, water-soluble inorganic ions, OC, EC and meteorological parameters) are averaged over 24 h periods when at least 75% of hourly data are available for each day; the missing values are excluded from the analysis.

Classification of Photochemical Activity Levels
Because the formation of secondary aerosol is closely related to the photochemical activity, O 3 is often used as an index of photochemical reactions [15][16][17]20,24]. To differentiate the effect of the different photochemical activity levels, in this study, photochemical activity is categorized into four groups ( As the Kruskal-Wallis test does not assume normality in the data, it is much less sensitive to outliers than the one-way Analysis of Variance (ANOVA). The Kruskal-Wallis test is used to determine the statistically significant difference in PM 2.5 and its chemical components under the different photochemical levels in this study.

Estimating the Photochemical Secondary Aerosols
In order to understand the formation of secondary aerosol in PM 2.5 , the secondary aerosol in PM 2.5 is estimated occurring with the intense levels of photochemical activity. CO and O 3 are used as indicating species for primary sources of motor vehicle emissions and secondary sources due to photochemical activity levels. When the level of photochemical activity is L, the observed PM 2.5 mass concentration is mainly from primary source emissions (including sea salt). Using the hourly ratio of PM 2.5 /CO to represent the primary aerosol emission, the proportion of secondary components in PM 2.5 is greater when the ratio of PM 2.5 /CO is greater and has been already clarified this hypothesis in [20,21]. The regression analysis between PM 2.5 and CO is conducted and the correlation coefficients are 0.70, 0.55, 0.62, and 0.63 under the photochemical activity level of L, M, LH and H, respectively. This illustrates that photochemical activities are low when O 3,max is less than 100 µg/m 3 . The ratio of PM 2.5 /CO is a useful tracer for estimating primary aerosol when photochemical activity is high [20].
When the photochemical activity levels are M, LH, and H, the photochemical activity is relatively high, and the primary aerosol in PM 2.5 is estimated by multiplying the ratio of PM 2.5 /CO by the CO. In this study, the primary aerosols are estimated using the hourly  [16,25].
In the aforementioned formula, P denotes primary pollutant; t denotes any hour of the day; L, M, LH, and H denots the different Photochemical activity levels; and (PM 2.5 /CO) p,L is the 25th percentile (0.014) of the hourly ratio for PM 2.5 /CO at L photochemical activity.
The estimated secondary aerosol mass concentration is calculated by the observed PM 2.5 mass concentration deducting the primary PM 2.5 mass concentration, according to the following equations: In the aforementioned formula, sec denotes secondary pollutant; obs denotes observed value; t denotes any hour of the day; and L, M, LH, and H denote the different photochemical activity levels.
The secondary organic aerosol (SOA) has been an important component of PM 2.5 [26][27][28][29][30] and can be formed from the photochemical oxidation reactions of VOCs. In this study, the primary organic aerosol (POA) and secondary organic aerosol (SOA) are estimated by the previously reported ratios, POA/POC (=1.2 µg/µgC) and SOA/SOC (=2.2 µg/µgC) [31,32]. The mass concentrations of POC and SOC are estimated using the method of the minimal ratio of OC/EC [33]: (POC) t = (EC) t × (OC/EC) pri,min (7) (SOC) t = (OC) obs,t − (POC) t (8) In the aforementioned formula, OC and EC are the measured hourly mass concentrations; t denotes any hour of the day; and (OC/EC) pri,min is the minimum of OC/EC in primary emissions. At urban locations and in this study, the minimum OC/EC observed in ambient air is the absolute minimum of all the data in the time series and is used to represent (OC/EC) pri .

Mass Concentration of PM 2.5 and O 3
The mass concentrations of PM 2.5 and O 3 are shown in Figure 1 during the summertime 2020 in Tianjin, China. As shown in Figure 1a, the PM 2.5 daily mean mass concentration was from 10.8 to 81.4 µg/m 3 , and the quarterly mean mass concentration 37.7 ± 19.8 µg/m 3 . Compared to the NAAQS of GB3095-2012, the PM 2.5 daily mass concentration exceeded the class I (35 µg/m 3 ) and class II (75 µg/m 3 ) in 46 days and 1 day, respectively, accounting for 50% and 1% of the 92 days during the summertime 2020 in Tianjin, China. Compared to air quality guidelines (15 µg/m 3 ) of the World Health Organization (WHO), during the observation period, the PM 2.5 daily concentration exceeded 90 days and accounted for 97.8%. As shown in Figure 1b,  counted for 97.8%. As shown in Figure 1b, the O3 seasonal mean concentration was 37.7 ± 19.8 μg/m 3 . The O3 daily 8 h average maximum (O3,8h max) mass concentration was between 58.6 and 274.3 μg/m 3 during the summertime. According to the class I of O3,8h max (100 μg/m 3 ) and the class II of O3,8h max (160 μg/m 3 ), the O3,8h max mass concentration, respectively, exceeded 46 and 34 days, accounting for 50% and 37% of the 92 days in the summertime 2020 in Tianjin, China. These indicated that the combined pollution of PM2.5 and O3 has been shown during the observation period in Tianjin, China. The COVID-19 pandemic unexpectedly broke out at the end of 2019. Due to the highly contagious, widespread, and risky nature of this disease, pandemic prevention and control measures were adopted in China, such as the public staying at home and some industries shutting down, which had an impact on the structure of pollution sources and air quality. Therefore, it must be clarified that the results of this study were based on exceptional circumstances.
In order to access the significance differences in the PM2.5 under the different photochemical activity levels of L, M, LH and H. The Kruskal-Wallis test was performed and the p value was equal to 0.000 and was significantly smaller than 0.05, which proved the The COVID-19 pandemic unexpectedly broke out at the end of 2019. Due to the highly contagious, widespread, and risky nature of this disease, pandemic prevention and control measures were adopted in China, such as the public staying at home and some industries shutting down, which had an impact on the structure of pollution sources and air quality. Therefore, it must be clarified that the results of this study were based on exceptional circumstances.
In order to access the significance differences in the PM 2.5 under the different photochemical activity levels of L, M, LH and H. The Kruskal-Wallis test was performed and the p value was equal to 0.000 and was significantly smaller than 0.05, which proved the significant difference in the PM 2.5 under the different photochemical activity levels of L, M, LH and H. In order to define the quantify of PM 2.5 levels, under the different photochemical activities indexed by O 3 , the PM 2.5 and O 3 mass concentrations are shown in Figure 2. The PM 2.5 mass concentration varied from 6 to 64 µg/m 3 , from 5 to 102 µg/m 3 , from 5 to 110 µg/m 3 , and from 10 to 130 µg/m 3 , and the mean values were 29.8 ± 17.4, 32.9 ± 20.4, 39.4 ± 19.1 and 42.2 ± 18.9 µg/m 3 under the photochemical activity levels of L, M, LH and H, respectively, which indicated that the PM 2.5 mass concentration increased with the strength of the photochemical activity as ascertained by O 3 levels, and serious pollution of PM 2.5 typically occurred in association with active photochemical processes or reactions. 5 to 110 μg/m 3 , and from 10 to 130 μg/m 3 , and the mean values were 29.8 ± 17.4, 32.9 ± 20.4, 39.4 ± 19.1 and 42.2 ± 18.9 μg/m 3 under the photochemical activity levels of L, M, LH and H, respectively, which indicated that the PM2.5 mass concentration increased with the strength of the photochemical activity as ascertained by O3 levels, and serious pollution of PM2.5 typically occurred in association with active photochemical processes or reactions.

Diurnal Variations in PM2.5
Based on the classification of photochemical activity, the diurnal profiles of hourly mean concentrations of PM2.5 and O3 are shown in Figure 3. The mass concentrations of PM2.5 and O3 presented a distinct diurnal pattern. The diurnal profile of PM2.5 showed a bimodal distribution. In contrast, the diurnal profile of O3 showed a unimodal distribution with peak value in the afternoon. When the photochemical activity level was L, both PM2.5 and O3 mass concentrations were relatively low and with small fluctuation amplitude, the maximum mass concentrations were 36.6 µ g/m 3 and 74.4 µ g/m 3 occurring at around 4:00 and 15:00, respectively. When the photochemical activity level was M, the mass concentrations of O3 were higher but the fluctuation amplitude was still low and the maximum value (121.5 µ g/m 3 ) was observed at 14:00. The fluctuation amplitude of PM2.5 was very significant. The maximum value (43.0 µ g/m 3 ) was observed at 7:00 and the minimum value (20.6 µ g/m 3 ) was observed at 18:00. When the photochemical activity level was LH, the mass concentrations of PM2.5 and O3 were obviously increasing during the daytime and the maximum values were 46.6 µ g/m 3 and 171.0 µ g/m 3 and observed at 7:00 and 16:00, respectively. When the photochemical activity level was H, the mass concentration of O3 further increased with a very significant fluctuation amplitude in the daytime and the maximum (211.0 µ g/m 3 ) was observed at 16:00. The PM2.5 mass concentrations also increased with small fluctuation amplitude in the daytime. The maximum mass concentration of PM2.5 increased to 49.3 µ g/m 3 and lagged until 9:00. In summary, under the different levels of photochemical activity, PM2.5 showed similar diurnal variation profile

Diurnal Variations in PM 2.5
Based on the classification of photochemical activity, the diurnal profiles of hourly mean concentrations of PM 2.5 and O 3 are shown in Figure 3. The mass concentrations of PM 2.5 and O 3 presented a distinct diurnal pattern. The diurnal profile of PM 2.5 showed a bimodal distribution. In contrast, the diurnal profile of O 3 showed a unimodal distribution with peak value in the afternoon. When the photochemical activity level was L, both PM 2.5 and O 3 mass concentrations were relatively low and with small fluctuation amplitude, the maximum mass concentrations were 36.6 µg/m 3 and 74.4 µg/m 3 occurring at around 4:00 and 15:00, respectively. When the photochemical activity level was M, the mass concentrations of O 3 were higher but the fluctuation amplitude was still low and the maximum value (121.5 µg/m 3 ) was observed at 14:00. The fluctuation amplitude of PM 2.5 was very significant. The maximum value (43.0 µg/m 3 ) was observed at 7:00 and the minimum value (20.6 µg/m 3 ) was observed at 18:00. When the photochemical activity level was LH, the mass concentrations of PM 2.5 and O 3 were obviously increasing during the daytime and the maximum values were 46.6 µg/m 3 and 171.0 µg/m 3 and observed at 7:00 and 16:00, respectively. When the photochemical activity level was H, the mass concentration of O 3 further increased with a very significant fluctuation amplitude in the daytime and the maximum (211.0 µg/m 3 ) was observed at 16:00. The PM 2.5 mass concentrations also increased with small fluctuation amplitude in the daytime. The maximum mass concentration of PM 2.5 increased to 49.3 µg/m 3 and lagged until 9:00. In summary, under the different levels of photochemical activity, PM 2.5 showed similar diurnal variation profile affected by the morning rush hours, the maximum appeared at around 5:00-9:00, after which the PM 2.5 mass concentration gradually decreased and its minimum mass concentration occurred around 16:00-18:00 and then gradually increased affected by the night rush hours until 23:00. The mass concentration of O 3 showed single peak pattern with lower mass concentration at night and rising rapidly around 7:00 accompanied by the enhancement of solar radiation, reaching the maximum values around 14:00-16:00 and then gradually decreasing until 23:00. which the PM2.5 mass concentration gradually decreased and its minimum mass concentration occurred around 16:00-18:00 and then gradually increased affected by the night rush hours until 23:00. The mass concentration of O3 showed single peak pattern with lower mass concentration at night and rising rapidly around 7:00 accompanied by the enhancement of solar radiation, reaching the maximum values around 14:00-16:00 and then gradually decreasing until 23:00.

The Effect of Meteorological Parameters in PM2.5
Under the different photochemical activity levels, the bivariate polar plot (Figure 4) showed the PM2.5 as a function of wind speed and direction, which was used to identify sources responsible for the significant concentrations [32]. Under the L photochemical activity level, the mass concentrations of PM2.5 were high and associated with low wind speeds (<1.5 m/s) indicating the dominance of local pollutants. Under the photochemical activity levels of M, LH and H, the high mass concentrations of PM2.5 were associated with low wind speeds (<1.5 m/s) and high wind speeds (>1.5 m/s) attributed from locally produced and long-range-transported pollutants. The high concentrations of PM2.5 were observed in multiple directions, except for the north, when the wind speeds were low, indicating large heterogeneity in the emission sources. Additionally, the high mass concentrations of PM2.5 with high wind speeds over the photochemical activities of LH and H was associated with the breeze from the southwest and northeast directions.

The Effect of Meteorological Parameters in PM 2.5
Under the different photochemical activity levels, the bivariate polar plot ( Figure 4) showed the PM 2.5 as a function of wind speed and direction, which was used to identify sources responsible for the significant concentrations [32]. Under the L photochemical activity level, the mass concentrations of PM 2.5 were high and associated with low wind speeds (<1.5 m/s) indicating the dominance of local pollutants. Under the photochemical activity levels of M, LH and H, the high mass concentrations of PM 2.5 were associated with low wind speeds (<1.5 m/s) and high wind speeds (>1.5 m/s) attributed from locally produced and long-range-transported pollutants. The high concentrations of PM 2.5 were observed in multiple directions, except for the north, when the wind speeds were low, indicating large heterogeneity in the emission sources. Additionally, the high mass concentrations of PM 2.5 with high wind speeds over the photochemical activities of LH and H was associated with the breeze from the southwest and northeast directions. hancement of solar radiation, reaching the maximum values around 14:00-16:00 and then gradually decreasing until 23:00.

The Effect of Meteorological Parameters in PM2.5
Under the different photochemical activity levels, the bivariate polar plot (Figure 4) showed the PM2.5 as a function of wind speed and direction, which was used to identify sources responsible for the significant concentrations [32]. Under the L photochemical activity level, the mass concentrations of PM2.5 were high and associated with low wind speeds (<1.5 m/s) indicating the dominance of local pollutants. Under the photochemical activity levels of M, LH and H, the high mass concentrations of PM2.5 were associated with low wind speeds (<1.5 m/s) and high wind speeds (>1.5 m/s) attributed from locally produced and long-range-transported pollutants. The high concentrations of PM2.5 were observed in multiple directions, except for the north, when the wind speeds were low, indicating large heterogeneity in the emission sources. Additionally, the high mass concentrations of PM2.5 with high wind speeds over the photochemical activities of LH and H was associated with the breeze from the southwest and northeast directions.  4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22    The other meteorological parameters, including temperature and relative humidity (RH), were also shown in Figure 5. Under the different photochemical activity levels of L, M, LH and H, the diurnal patterns of ambient temperature were typical unimodal and ranged from 23.0 to 27.2 °C, 24.6 to 29.3 °C, 24.0 to 32.7 °C, and 24.7 to 35.3 °C, respectively. The diurnal patterns of RH anti-correlated with temperature. RH ranged from 60.7% to 77.0%, 51.7% to 68.2%, 39.4% to 70.0%, and 31.1% to 64.8%, respectively. This indicated that the photochemical activity level increased along with temperature and conversely decreased with the increase in RH.

Secondary Aerosols Estimation
Equations (1)-(3) were used to estimate the primary mass concentrations of PM2.5 for the three O3,max intervals of M, LH, and H. As shown in Figure 6, at the different photochemical activity levels, the diurnal patterns of estimated primary PM2.5 mass concentration were similar with a small fluctuation amplitude. The maximum values appeared at about 8:00 corresponding to the morning rush hours and the minimum values appeared around 15:00-16:00 owing to the joint effect of increasing of the mixing layer height (MLH) and atmospheric temperature. The primary PM2.5 concentrations then increased after  The other meteorological parameters, including temperature and relative humidity (RH), were also shown in Figure 5. Under the different photochemical activity levels of L, M, LH and H, the diurnal patterns of ambient temperature were typical unimodal and ranged from 23.0 to 27.2 • C, 24.6 to 29.3 • C, 24.0 to 32.7 • C, and 24.7 to 35.3 • C, respectively. The diurnal patterns of RH anti-correlated with temperature. RH ranged from 60.7% to 77.0%, 51.7% to 68.2%, 39.4% to 70.0%, and 31.1% to 64.8%, respectively. This indicated that the photochemical activity level increased along with temperature and conversely decreased with the increase in RH.  The other meteorological parameters, including temperature and relative humidity (RH), were also shown in Figure 5. Under the different photochemical activity levels of L, M, LH and H, the diurnal patterns of ambient temperature were typical unimodal and ranged from 23.0 to 27.2 °C, 24.6 to 29.3 °C, 24.0 to 32.7 °C, and 24.7 to 35.3 °C, respectively. The diurnal patterns of RH anti-correlated with temperature. RH ranged from 60.7% to 77.0%, 51.7% to 68.2%, 39.4% to 70.0%, and 31.1% to 64.8%, respectively. This indicated that the photochemical activity level increased along with temperature and conversely decreased with the increase in RH.

Secondary Aerosols Estimation
Equations (1)-(3) were used to estimate the primary mass concentrations of PM2.5 for the three O3,max intervals of M, LH, and H. As shown in Figure 6, at the different photochemical activity levels, the diurnal patterns of estimated primary PM2.5 mass concentration were similar with a small fluctuation amplitude. The maximum values appeared at about 8:00 corresponding to the morning rush hours and the minimum values appeared around 15:00-16:00 owing to the joint effect of increasing of the mixing layer height (MLH) and atmospheric temperature. The primary PM2.5 concentrations then increased after 18:00, corresponding to the night rush hours, and then they were relatively steady during

Secondary Aerosols Estimation
Equations (1)-(3) were used to estimate the primary mass concentrations of PM 2.5 for the three O 3,max intervals of M, LH, and H. As shown in Figure 6, at the different photochemical activity levels, the diurnal patterns of estimated primary PM 2.5 mass concentration were similar with a small fluctuation amplitude. The maximum values appeared at about 8:00 corresponding to the morning rush hours and the minimum values appeared around 15:00-16:00 owing to the joint effect of increasing of the mixing layer height (MLH) and atmospheric temperature. The primary PM 2.5 concentrations then increased after 18:00, corresponding to the night rush hours, and then they were relatively steady during the night. Under the photochemical activity level of M, the ratio of estimated primary concentration to the observed mass concentration of PM 2.5 (PM 2 . 5,pri /PM 2 . 5,obs ) was about 0.30 from 0:00 to 5:00, then it gradually increased and reached a secondary peak value (0.38) at 11:00, lagging the morning rush hours by 2-3 h, then after a slight decrease it rapidly increased and reached the primary peak value (0.49) corresponding to the night rush hours. This indicated that diurnal variation in the primary PM 2.5 mass concentration was mainly affected by traffic emission sources under the photochemical activity level M. Under the photochemical activity levels of LH and H, the ratio of PM 2 . 5,pri /PM 2 . 5,obs presented a small variation (between 0.29 and 0.33), with no peaks related to the morning or night rush hours. It could be seen that the proportion trend under the M photochemical level was significantly different from under the photochemical levels of LH and H. Analyzing the reasons for this, it was possible that the secondary aerosol concentrations were lower at the photochemical level of M than that at the photochemical levels of LH and H (see Figure 7). Furthermore, the primary aerosol concentrations also varied at the different photochemical levels. The joint effect led to the different proportion trend. the night. Under the photochemical activity level of M, the ratio of estimated primary concentration to the observed mass concentration of PM2.5 (PM2.5,pri/PM2.5,obs) was about 0.30 from 0:00 to 5:00, then it gradually increased and reached a secondary peak value (0.38) at 11:00, lagging the morning rush hours by 2-3 h, then after a slight decrease it rapidly increased and reached the primary peak value (0.49) corresponding to the night rush hours. This indicated that diurnal variation in the primary PM2.5 mass concentration was mainly affected by traffic emission sources under the photochemical activity level M. Under the photochemical activity levels of LH and H, the ratio of PM2.5,pri/PM2.5,obs presented a small variation (between 0.29 and 0.33), with no peaks related to the morning or night rush hours. It could be seen that the proportion trend under the M photochemical level was significantly different from under the photochemical levels of LH and H. Analyzing the reasons for this, it was possible that the secondary aerosol concentrations were lower at the photochemical level of M than that at the photochemical levels of LH and H (see Figure 7). Furthermore, the primary aerosol concentrations also varied at the different photochemical levels. The joint effect led to the different proportion trend. The secondary PM2.5 concentrations were estimated using Equations (4)-(6) by subtracting the estimated primary PM2.5 from the observed PM2.5, as shown in Figure 7. At the M, LH and H photochemical activity levels, the daily variation trend of the generated secondary aerosols was nearly similar and between 26.0 and 32.2 μg⸳m −3 from 0:00 to 5:00, after which the mass concentration of the secondary aerosol significantly decreased at the M photochemical activity and reached the lowest value (10.6 μg/m 3 ) at 18:00, after which there was an increase. Under the LH photochemical activity, the secondary aerosol reached the first peak (33.4 μg/m 3 ) at 7:00, then gradually decreased and reached the lowest value (23.2 μg/m 3 ) at 17:00, and no significant variations were registered after that. Under the H photochemical activity level, the secondary aerosol reached the first peak (34.7 μg/m 3 ) at 9:00, then gradually decreased and reached the lowest value (25.9 μg/m 3 ) at 17:00, and after that then slightly increased and reached the second peak value (32.0 μg/m 3 ). This showed that the mass concentration of secondary PM2.5 aerosol increased with the increase in photochemical activity. The contribution ratio of the secondary PM2.5 to the observed PM2.5 was more than 58.6% at any hour under the photochemical activities of LH and H. The secondary PM 2.5 concentrations were estimated using Equations (4)-(6) by subtracting the estimated primary PM 2.5 from the observed PM 2.5 , as shown in Figure 7. At the M, LH and H photochemical activity levels, the daily variation trend of the generated secondary aerosols was nearly similar and between 26.0 and 32.2 µg·m −3 from 0:00 to 5:00, after which the mass concentration of the secondary aerosol significantly decreased at the M photochemical activity and reached the lowest value (10.6 µg/m 3 ) at 18:00, after which there was an increase. Under the LH photochemical activity, the secondary aerosol reached the first peak (33.4 µg/m 3 ) at 7:00, then gradually decreased and reached the lowest value (23.2 µg/m 3 ) at 17:00, and no significant variations were registered after that. Under the H photochemical activity level, the secondary aerosol reached the first peak (34.7 µg/m 3 ) at 9:00, then gradually decreased and reached the lowest value (25.9 µg/m 3 ) at 17:00, and after that then slightly increased and reached the second peak value (32.0 µg/m 3 ). This showed that the mass concentration of secondary PM 2.5 aerosol increased with the increase in photochemical activity. The contribution ratio of the secondary PM 2.5 to the observed PM 2.5 was more than 58.6% at any hour under the photochemical activities of LH and H. Because the national standards used by different countries are different, there were certain differences in the specific values selected when defining the photochemical activity level of O3, so there will be differences in the estimated primary aerosol and secondary aerosol. Therefore, there are certain limitations when comparing and analyzing different literatures.
The diurnal patterns of primary organic aerosol (POA) and secondary organic aerosol (SOA) are shown in Figure 8. This indicates that the daily variation in the POA mass concentration was from 0.76 to 1.42 μg/m 3 under different photochemical activity levels, with the maximum value occurring at around 7:00 according to the traffic morning rush hour, implying that large amounts of hydrocarbons were emitted from vehicles and were beneficial to the formation of POA. The first and second peaks of the SOA mass concentration appeared at about 11:00-13:00 and 19:00-21:00, lagging about 4 h after morning and night rush hours, respectively. Since the secondary organic aerosol was formed from gaseous organic precursors, SOA was thus an important source of PM2.5 in Tianjin during this measurement period. This phenomenon was observed in the late summer 2002 in Pittsburgh [29] and in Mexico City Metropolitan Area in 2006 [33].  4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22   Because the national standards used by different countries are different, there were certain differences in the specific values selected when defining the photochemical activity level of O 3 , so there will be differences in the estimated primary aerosol and secondary aerosol. Therefore, there are certain limitations when comparing and analyzing different literatures.
The diurnal patterns of primary organic aerosol (POA) and secondary organic aerosol (SOA) are shown in Figure 8. This indicates that the daily variation in the POA mass concentration was from 0.76 to 1.42 µg/m 3 under different photochemical activity levels, with the maximum value occurring at around 7:00 according to the traffic morning rush hour, implying that large amounts of hydrocarbons were emitted from vehicles and were beneficial to the formation of POA. The first and second peaks of the SOA mass concentration appeared at about 11:00-13:00 and 19:00-21:00, lagging about 4 h after morning and night rush hours, respectively. Since the secondary organic aerosol was formed from gaseous organic precursors, SOA was thus an important source of PM 2.5 in Tianjin during this measurement period. This phenomenon was observed in the late summer 2002 in Pittsburgh [29] and in Mexico City Metropolitan Area in 2006 [33].

PM 2.5 Chemical Characterization
The main chemical components, including water-soluble inorganic ions, OC, and EC in PM 2.5 , are shown in Table 1

PM2.5 Chemical Characterization
The main chemical components, including water-soluble inorganic ions, OC, and EC in PM2.5, are shown in Table 1   The diurnal profiles of the measured PM2.5 and chemical components reconstructed PM2.5 mass concentrations correlated well in Figure 9a showed that the diurnal variations in chemical components were steady, except for slightly elevated OC in the morning which might be resulted from vehicle emissions according to the morning rush hour. The diurnal variation in chemical components in Figure 9b was similar to Figure 9a. However the mass concentrations of components were slightly enhanced. Figure 9c shows that the mass concentrations of chemical components were even higher and peak value of OC was shifted to 11:00, which was closer to the occurrence time of SOC by higher O3,max (see Fig  ure 7). For the photochemical activity of H, Figure 9d showed that the peak value of the sum over aerosol chemical components occurred at 11:00; however, the peak value of OC and SO4 2− reaches the summit at 13:00. This was due to the decrease in NO3 − after 11:00  The diurnal profiles of the measured PM 2.5 and chemical components reconstructed PM 2.5 mass concentrations correlated well in Figure 9a showed that the diurnal variations in chemical components were steady, except for slightly elevated OC in the morning which might be resulted from vehicle emissions according to the morning rush hour. The diurnal variation in chemical components in Figure 9b was similar to Figure 9a. However, the mass concentrations of components were slightly enhanced. Figure 9c shows that the mass concentrations of chemical components were even higher and peak value of OC was shifted to 11:00, which was closer to the occurrence time of SOC by higher O 3,max (see Figure 7). For the photochemical activity of H, Figure 9d showed that the peak value of the sum over aerosol chemical components occurred at 11:00; however, the peak value of OC and SO 4 2− reaches the summit at 13:00. This was due to the decrease in NO 3 − after 11:00 which might result from evaporation loss of NO 3 − at high ambient temperature [34][35][36][37]. The sum of aerosol chemical components was apparently enriched under the photochemical activity level H as shown in Figure 9d. This provided the evidence of secondary aerosol formation under high photochemical activity in Tianjin. The peak values of OC, SO 4 2− , and NO 3 − were 5.0, 11.0, and 7.3 µg/m 3 , respectively, under the photochemical activity level of H, which was higher than that of 3.1, 9.4, and 5.6 µg/m 3 , respectively, under the photochemical activity level L. In summary, under the different photochemical activities, the peak value of PM 2.5 was accompanied with the increase in the sum of mass concentrations of SO 4 2− , NO 3 − and NH 4 + (SNA) and OC. The hourly mass concentration maximum value of SNA in PM 2.5 were 21.7, 24.9, 26.0 and 24.0 µg/m 3 under the photochemical activity levels of L, M, LH, and H, respectively. This indicated that the formation of secondary aerosols contributed significantly to the increase in PM 2.5 mass. Furthermore, as shown in Figure 10, under the photochemical activities of L, M, LH, and H, SO 4  − ] refers to the molar concentration (µmol/m 3 ), suggesting that ammonia was abundant and that the formation of particulate nitrate and sulfate was under an ammonia-rich condition in this measurement period. The result was observed at a rural site in eastern Yangtze River Delta of China [38].
centrations of SO4 2− , NO3 − and NH4 + (SNA) and OC. The hourly mass concentration maximum value of SNA in PM2.5 were 21.7, 24.9, 26.0 and 24.0 μg/m 3 under the photochemical activity levels of L, M, LH, and H, respectively. This indicated that the formation of secondary aerosols contributed significantly to the increase in PM2.5 mass. Furthermore, as shown in Figure 10, under the photochemical activities of L, M, LH, and H, SO4 2− and NO3 − were nearly neutralized by NH4 + , with a regression slope of 0.71, 0.77, 0.77, and 0.75 between 2[SO4 2− ] + [NO3 − ] and [NH4 + ], respectively, where [NH4 + ], [SO4 2− ], [NO3 − ] refers to the molar concentration (μmol/m 3 ), suggesting that ammonia was abundant and that the formation of particulate nitrate and sulfate was under an ammonia-rich condition in this measurement period. The result was observed at a rural site in eastern Yangtze River Delta of China [38].

Insights into the Formation of SO4 2− and NO3 −
In order to investigate the formation mechanism of SO4 2− , the diurnal profiles of SO4 2− and SO2 mass concentration are shown in Figure 11 under the photochemical activity levels M, LH and H. The mean mass concentration of SO4 2− showed a rapid increase from 5.2 μg/m 3 to 6.3 μg/m 3 between 6:00 and 11:00. SO2 increased sharply from 6:00 and peaked at 10:00. The SO2 peak was thus 1 h earlier than the SO4 2− peak, suggesting rapid formation of secondary sulfate, from the increased concentrations of SO2 probably due to traffic emissions in the morning rush hours, coupled with the propitious photochemical conditions. It suggested that active oxidation of SO2 in the gas phase was mainly responsible for the observed morning production of sulfate. The oxidation of SO2 to sulfuric acid (H2SO4) could occur both in the aqueous phase and in the gas phase [39][40][41]. Recent studies have shown that the gas-phase oxidation of SO2 to H2SO4 had become more and more important [42]. In this pathway, SO2 was first oxidized by OH to sulfur trioxide (SO3) and then to H2SO4 [42,43]. In order to investigate the formation mechanism of SO 4 2− , the diurnal profiles of SO 4 2− and SO 2 mass concentration are shown in Figure 11 under the photochemical activity levels M, LH and H. The mean mass concentration of SO 4 2− showed a rapid increase from 5.2 µg/m 3 to 6.3 µg/m 3 between 6:00 and 11:00. SO 2 increased sharply from 6:00 and peaked at 10:00. The SO 2 peak was thus 1 h earlier than the SO 4 2− peak, suggesting rapid formation of secondary sulfate, from the increased concentrations of SO 2 probably due to traffic emissions in the morning rush hours, coupled with the propitious photochemical conditions. It suggested that active oxidation of SO 2 in the gas phase was mainly responsible for the observed morning production of sulfate. The oxidation of SO 2 to sulfuric acid (H 2 SO 4 ) could occur both in the aqueous phase and in the gas phase [39][40][41]. Recent studies have shown that the gas-phase oxidation of SO 2 to H 2 SO 4 had become more and more important [42]. In this pathway, SO 2 was first oxidized by OH to sulfur trioxide (SO 3 ) and then to H 2 SO 4 [42,43].

Insights into the Formation of SO4 2− and NO3 −
In order to investigate the formation mechanism of SO4 2− , the diurnal profiles of SO4 2− and SO2 mass concentration are shown in Figure 11 under the photochemical activity levels M, LH and H. The mean mass concentration of SO4 2− showed a rapid increase from 5.2 μg/m 3 to 6.3 μg/m 3 between 6:00 and 11:00. SO2 increased sharply from 6:00 and peaked at 10:00. The SO2 peak was thus 1 h earlier than the SO4 2− peak, suggesting rapid formation of secondary sulfate, from the increased concentrations of SO2 probably due to traffic emissions in the morning rush hours, coupled with the propitious photochemical conditions. It suggested that active oxidation of SO2 in the gas phase was mainly responsible for the observed morning production of sulfate. The oxidation of SO2 to sulfuric acid (H2SO4) could occur both in the aqueous phase and in the gas phase [39][40][41]. Recent studies have shown that the gas-phase oxidation of SO2 to H2SO4 had become more and more important [42]. In this pathway, SO2 was first oxidized by OH to sulfur trioxide (SO3) and then to H2SO4 [42,43]. In order to investigate the different formation mechanisms of nitrate, the daytime was divided into four stages, as shown in Figure 12: Stage I, 0:00-5:00; Stage II, 5:00-9:00; Stage III, 9:00-18:00; Stage IV, 18:00-23:00. During Stage I, the mass concentration of NO 2 kept on slowly increasing while the temperature was relatively low and RH was relatively high. Moreover, the temperature and RH stayed relatively constant during this stage, as could be seen in Figure 5, implying that the thermal equilibrium entered a steadier state. Some previous studies have shown that nitrate radical (NO 3 ) and/or N 2 O 5 were converted to nitric acid (HNO 3 ) by hydrolysis under high RH [44,45]. During Stage II, the mass concentration of NO 2 rapidly increased and peaked at about 7:00, due to the traffic emissions and relatively low mixed layer height (MLH). At the same time, the mass concentration of O 3 began to rise due to increasing solar radiation and more active photochemical reactions [46,47]. These observations together indicated that elevated nitrate concentration was primarily due to gas-phase photochemical oxidation. The result was consistent with the observation in Hong Kong [48]. During Stage III, the mass concentrations of NO 3 − sharply decreased, probably because the NO 3 radical was easily destroyed by photolysis under the strong photochemical activity of this period in the day. Furthermore, the decrease in the nitrate concentration could also be attributed to the evaporation of ammonium nitrate due to the increasing temperature and the enhancement of the atmospheric diffusion capacity as the MLH ascends [49]. During Stage IV, the temperature decreased while RH gradually increased ( Figure 5). The mass concentration of O 3 gradually decreased ( Figure 3) and remained relatively high during the period. The mass concentration of NO 2 rose after sunset due to traffic emissions during the evening rush hours coupled with lower MLH. Due to the presence of abundant NO 2 and O 3 , as well as the absence of sunlight, the conditions were created for the concentration of NO 3 radical to increase gradually [50], leading to a steady increase in NO 3 − concentration in this period. This observation was consistent with the measurement in August 2011 in the urban area of Shanghai [51]. In order to investigate the different formation mechanisms of nitrate, the daytime was divided into four stages, as shown in Figure 12: Stage I, 0:00-5:00; Stage II, 5:00-9:00; Stage III, 9:00-18:00; Stage IV, 18:00-23:00. During Stage I, the mass concentration of NO2 kept on slowly increasing while the temperature was relatively low and RH was relatively high. Moreover, the temperature and RH stayed relatively constant during this stage, as could be seen in Figure 5, implying that the thermal equilibrium entered a steadier state. Some previous studies have shown that nitrate radical (NO3) and/or N2O5 were converted to nitric acid (HNO3) by hydrolysis under high RH [44,45]. During Stage II, the mass concentration of NO2 rapidly increased and peaked at about 7:00, due to the traffic emissions and relatively low mixed layer height (MLH). At the same time, the mass concentration of O3 began to rise due to increasing solar radiation and more active photochemical reactions [46,47]. These observations together indicated that elevated nitrate concentration was primarily due to gas-phase photochemical oxidation. The result was consistent with the observation in Hong Kong [48]. During Stage III, the mass concentrations of NO3 − sharply decreased, probably because the NO3 radical was easily destroyed by photolysis under the strong photochemical activity of this period in the day. Furthermore, the decrease in the nitrate concentration could also be attributed to the evaporation of ammonium nitrate due to the increasing temperature and the enhancement of the atmospheric diffusion capacity as the MLH ascends [49]. During Stage IV, the temperature decreased while RH gradually increased ( Figure 5). The mass concentration of O3 gradually decreased ( Figure 3) and remained relatively high during the period. The mass concentration of NO2 rose after sunset due to traffic emissions during the evening rush hours coupled with lower MLH. Due to the presence of abundant NO2 and O3, as well as the absence of sunlight, the conditions were created for the concentration of NO3 radical to increase gradually [50], leading to a steady increase in NO3 − concentration in this period. This observation was consistent with the measurement in August 2011 in the urban area of Shanghai [51].

Conclusions
Based on the atmospheric environment monitoring data, the study has performed the chemical characterization of PM2.5 and the estimation of the related secondary aerosol

Conclusions
Based on the atmospheric environment monitoring data, the study has performed the chemical characterization of PM 2.5 and the estimation of the related secondary aerosol during the summertime 2020 in Tianjin, China. Under the photochemical activities of L, M, LH, and H, the PM 2.5 mass concentration increased with the strength of the photochemical activity as ascertained by O 3 levels and the diurnal patterns were similar to the maximum and minimum appeared at around 7:00-9:00 and 16:00, respectively. The diurnal patterns of estimated secondary aerosol increased with the strength increase in photochemical activity. In particular, the ratio of estimated secondary aerosol to the observed PM 2.5 was more than 58.6% at any hour under the photochemical activity of LH and H, indicating secondary aerosols became the key issue of PM 2.5 pollution in Tianjin. The chemical composition, including water-soluble ions, OC and EC, respectively accounted for 73.5 ± 14.9%, 70.3 ± 24.9%, 72.0 ± 21.9%, and 65.8 ± 21.2% of the observed PM 2.5 mass. In particular, the SNA significantly contributed to the increase in PM 2.5 under an ammonia-rich condition in this observation period. The photochemical oxidation of SO 2 to H 2 SO 4 was enhanced by the strong atmospheric photochemical reactions. Nitrate was mainly produced by the photochemical oxidation reactions in the daytime, but high temperature and low RH shifted the gas-to-particle partitioning of NH 4 NO 3 to evaporation, thus led to an extremely low particulate nitrate concentration in the afternoon.