Sensitivity Analysis of PM 2 . 5-Bound Total PCDD / Fs-TEQ Content : In the Case of Wuhu City , China

During 2015–2017, the atmospheric PM2.5, PM2.5/PM10, PCDD/Fs, PCDD/F phase distribution and PM2.5-bound total PCDD/Fs-WHO2005-TEQ content in Wuhu and Bengbu were investigated in this study. In addition, the sensitivity analysis for PM2.5-bound total PCDD/Fs-WHO2005-TEQ of Wuhu was also studied. During 2015–2017, the three-year average PM2.5 concentration in Wuhu was 53.0 μg m, and in Bengbu was 61.4 μg m; the results also showed the annual average PM2.5 concentrations of these two cities had declined, but the levels were still far above the WHO annual PM2.5 standard (10 μg m). In addition, in Wuhu, the PM2.5-bound total PCDD/Fs-WHO2005-TEQ contents in summer (0.166 ng-WHO2005-TEQ g) were approximately only 68.8% in magnitude lower than the average value of other three seasons (0.532 ng-WHO2005TEQ g), and that of Bengbu in summer (0.187 ng-WHO2005-TEQ g) was approximately 66.7% in magnitude lower than the average value of other three seasons (0.561 ng-WHO2005-TEQ g). Sensitivity analysis showed that the PCDD/F concentration was the most positively correlated sensitive factor for PM2.5-bound total PCDD/Fs-WHO2005-TEQ, and when ΔP/P was changed from 0% to +50%, ΔS/S responded from 0% to +106%. The second positively correlated sensitive factor was PM10 concentration, and when ΔP/P was changed from 0% to +50%, ΔS/S responded from 0% to +72%. This was followed by atmospheric temperature, and its effect was negatively correlated, when ΔP/P was changed from –50% to +50%, ΔS/S responded from +73% to –112%. The last sensitive parameter was PM2.5 concentration, with the impact divided into two stages: when ΔP/P was changed from 0% to +70%, ΔS/S responded from 0% to +33%, but when ΔP/P was changed from +70% to +100%, ΔS/S responded from +33% to +25%.The results of this study provide useful information that can be used to achieve more insights into both atmospheric PM2.5 and PCDD/Fs.


INTRODUCTION
The particulate matter (PM) and polychlorinated dibenzop-dioxins and polychlorinated dibenzofurans (PCDD/Fs) in the ambient air have received great public concern due to their significant correlation with the incidences of pulmonary and cardiac diseases (Schwartz et al., 1996;Ito et al., 2006).
Particulate matter (PM) is a suspension of solid or liquid particles in the atmosphere, which is a kind of aerosol (Ghosh et al., 2014).According to the aerodynamic diameters of PM, it can be divided into TSP (range from ~0 to 100 µm), PM 10 (range from ~0 to 10 µm) and PM 2.5 (range from ~0 to 2.5 µm) (Chow et al., 2015;Lu et al., 2016).The sources of PM can be natural or anthropogenic, with forest burning, dust storms and volcanic eruptions are the main natural sources, while industrial activities, vehicle exhausts and the construction industry are the major anthropogenic sources (Bilos et al., 2001;Kong et al., 2014;Alghamdi et al., 2015).The PM in ambient air is a conglomerate of organic and inorganic carbon, mineral element, nitrates, ammonium, sulfate and so on (Zhu et al., 2017).While the PM emitted into atmosphere by primary sources, the secondary aerosol forms with a specific ratio and environmental conditions (Lee et al., 2016).Previous studies of the chemical characteristics of PM in the atmosphere indicated that it not only had adverse effects on air quality and even global climate, but also impacts on human health due to its particle toxicity (Chen et al., 2014;Huang et al., 2014;Wang et al., 2014;Liu et al., 2016).
PCDD/Fs are well-known persistent organic pollutants (POPs) and semi-volatile organic compounds (SOCs), which can transport over long distances and interact for long periods in the environment (Wu et al., 2009b;Chen et al., 2014;Lee et al., 2016;Redfern, et al., 2017), as well as bio-accumulate in the fatty tissues and bio-magnify within the food chain (Shih et al., 2009).PCDD/Fs are extremely hazardous chemicals, and ingestion, inhalation and dermal contact are the primary pathways for them to get into human bodies (Shih et al., 2009;Chen et al., 2010), thus posing risks to the immune system, interfering with hormones and even leading to cancer (Lin et al., 2010;Chi et al., 2011).PCDD/Fs were detected in the emissions of municipal solid waste incinerators (MSWIs) for the first time (Olie et al., 1977), and since then have become one of the most controversial environmental pollutants.Combustion processes and some industrial activities are the most dominant sources of PCDD/Fs released to the environment, such as power and heating facilities, metal smelting processes and waste incineration (Schuhmacher et al., 2000;Wang et al., 2003;Lin et al., 2007;Hsieh et al., 2009;Chuang et al., 2010Chuang et al., , 2011)).PCDD/Fs are complex mixture of different congeners, there are 210 possible congeners and 17 of these have been shown to be more toxic, with the 2,3,7,8 positions attached by chlorine atoms, and the toxicities are estimated by the toxic equivalent quantity (TEQ) (Cheruiyot, et al., 2016).
In the atmosphere, after being emitted from combustion facilities, the PCDD/Fs are distributed into both gas and particle phases (Li et al., 2008b;Chen et al., 2011a;Kou et al., 2015).Many studies show that the gas-particle partitioning of PCDD/Fs is highly dependent on their vapor pressures, ambient temperatures and other parameters (Wu et al., 2009a;Wang et al., 2010;Cheruiyot et al., 2015).More fractions of PCDD/Fs are volatilized into gas with an increase in temperature (Fiedler, 1996;Oh et al., 2001).The degradation of PCDD/Fs depends on chemical and photochemical reactions, and their removal mainly relies on the atmospheric deposition (Giorgi, 1988;Chi et al., 2009;Wu et al., 2009a;Huang et al., 2011a;Mi et al., 2012).
This study investigated the PM 2.5 concentrations, PM 2.5 / PM 10 ratios, PCDD/F concentrations, gas-particle partitioning, and PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ content, while the sensitivity analysis of PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents was also studied and discussed.

METHODS
Two cities, namely Wuhu (31°33′N, 118°38′E) and Bengbu (32°93′N, 117°34′E) in Anhui province, China, were evaluated in this study.The monthly mean concentrations of both PM 2.5 and PM 10 , monthly temperature and precipitation in both cities were obtained from local air quality monitoring stations and the Statistics Yearbook of China.
The total PCDD/F concentration was simulated by regression analysis of the PM 10 concentration.Tang et al. (2017) reported that there is a high correlation between PM 10 values and total PCDD/F mass concentrations.Tang et al. (2017) included the following two regression equations: Y 1 , Y 2 : total PCDD/F concentration (pg m -3 ); x: PM 10 concentration in ambient air (µg m -3 ).
The final total PCDD/F concentration was the average of Y 1 and Y 2 .

Gas-Particle Partitioning
The PCDD/F concentrations in the gas and particle phases, respectively, were calculated by using gas-particle partitioning model as Eq. ( 3) (Yamasaki et al., 1982;Pankow, 1987;Pankow andBidleman, 1991, 1992): K p : temperature-dependent partitioning constant (m 3 µg -1 ); TSP: concentration of total suspended particulate matter, which was multiplied by PM 10 concentration with 1.24 (µg m -3 ); F: concentration of the compounds of interest bound to particles (pg m -3 ); A: gaseous concentration of the compound of interest (pg m -3 ).Plotting log K p against the logarithm of the subcooled liquid vapor pressure, P L 0 , gives Complete datasets on the gas-particle partitioning of PCDD/Fs in Taiwan have been reported (Chao et al., 2004), with the values m r = -1.29 and b r = -7.2 with R 2 = 0.94.These values were used in this study for establishing the partitioning constant (K p ) of PCDD/Fs.
A previous study correlated the P L 0 of PCDD/Fs with gas chromatographic retention indexes (GC-RI) on a nonpolar (DB-5) GC-column using p,p′-DDT as a reference standard.The correlation has been re-developed as follows (Hung et al., 2002).

PM 2.5 Concentration
The PM 2.5 concentration not only has a significant correlation with air visibility and human health, but can also reflect the PCDD/F concentration of a region.For the period 2015-2017, the monthly average PM 2.5 concentrations in the ambient air of Wuhu and Bengbu are shown in Figs. 1(a), 1(b) and 1(c).As for Wuhu, over the three years examined, the lowest PM 2.5 concentration occurred in 2017, at 27.0-99.0µg m -3 , and with an average of 48.9 µg m -3 ; followed by 2016, which was in the range of 25.0-90.0µg m -3 and with an average of 53.0 µg m -3 ; and in 2015, in the range of 32.0-108.0µg m -3 , with an average of 57.3 µg m -3 .Comparing the annual average PM 2.5 concentrations, we can see that the highest values occurred in 2015, followed by 2016, and the lowest occurred in 2017.The PM 2.5 annual average concentration, compared to that for 2015 to 2017, was reduced by approximately 14.7%.The tendency of the PM 2.5 level to decline slowly may be because of social efforts to improve the quality of the environment.As a whole, the PM 2.5 concentration of the three-year average in Wuhu ranged between 25.0 and 108.0 µg m -3 , with an average of 53.0 µg m -3 .It also can be seen that even though the air quality of Wuhu was improved significantly, the PM 2.5 concentrations were still far above the WHO air quality regulated standard (10 µg m -3 ), so more efforts are needed issue.
With regard to Bengbu (Figs. 1(a), 1(b) and 1(c)), the monthly average PM 2.5 concentrations in 2017, which was between 34.0 and 98.0 µg m -3 , with an average of 61.1 µg m -3 ; and those in 2016 ranged from 33.0 to 101.0 µg m -3 and averaged 59.8 µg m -3 ; during 2015, the PM 2.5 concentration ranged between 47.0 and 88.0 µg m -3 and averaged 63.2 µg m -3 .These results reveal that from 2015 to 2016 the annual average PM 2.5 concentration fell from 63.2 to 59.8 µg m -3 , falling by approximately 5.4%.However, the highest values increased by 14.8% from 2015 (88.0 µg m -3 ) to 2016 (101.0 µg m -3 ).As a whole, the PM 2.5 concentration for these three years in Bengbu ranged between 33.0 and 101.0 µg m -3 , with an average of 61.4 µg m -3 .This indicates that the PM 2.5 level in Wuhu was slightly lower than that in Bengbu.This is probably due to the development of industry with poor air pollution control, as well as more pollutants being emitted from mobile and stationary sources.
When regard to the seasonal variations, the four seasons were defined as spring (March, April, May), summer (June, July, August), fall (September, October, November) and winter (January, February and December).For Wuhu, during 2015, the average PM 2.5 concentrations in spring, summer, fall and winter were 49.7, 34.7, 60.3 and 84.3 µg m -3 , respectively; and those in 2016 were 52.3, 32.0, 51.0 and 76.7 µg m -3 , respectively; while those in 2017 were 47.7, 30.3, 46.7 and 71.0 µg m -3 , respectively.For Bengbu, during 2015, the average PM 2.5 concentrations in spring, summer, fall and winter were 57.3, 48.0, 69.0 and 78.3 µg m -3 , respectively; and those in 2016 were 60.0, 37.7, 53.3 and 90.0 µg m -3 , respectively; while those in 2017 were 60.0, 40.0, 52.3 and 92.0 µg m -3 , respectively.This indicates that the PM 2.5 concentrations varied from season to season, the highest values always occurred in winter and the lowest in summer, while the values for spring were very similar with fall and both in the middle levels.As a whole, the threeyear average of PM 2.5 concentration of Wuhu in summer (32.3 µg m -3 ) was 58.2% in magnitude lower than that in winter (77.3 µg m -3 ); and the values of Bengbu in summer (41.9 µg m -3 ) was 51.7% in magnitude lower than that in winter (86.8 µg m -3 ).
During the cold seasons, cities at a higher latitude have a very low ground temperature, so it is easier for high stability vertical atmospheric convection, to occur, which hinders the dispersion of air pollutants.As a result, an accumulation of PM 2.5 concentrations occurs in regions with human activity.In addition, the parts of north China require more coal and other fossil fuel combustion for heating during winter, and thus have more air pollutant emissions.Therefore, the air current blowing from the northern cities, which have quite high PM 2.5 concentration levels, may increase the PM 2.5 concentrations in both Wuhu and Bengbu, cities located in central China.And in warm season the air temperature is higher and the vertical transport of air current is more violent, which can accelerate the dispersion of air PM 2.5 .Moreover, in summer both rainfall and wind speed are stronger, enhancing the effects of rainfall scavenging and wind blowing in removing the PM 2.5 from the atmosphere.

PM 2.5 /PM 10 Ratio
The PM 2.5 /PM 10 ratio can reflect the contribution of fine particulate matter to the ambient air pollutants, and thus mirror the status of air pollution.The monthly average of PM 2.5 /PM 10 ratio in the ambient air in Wuhu and Bengbu are presented in Figs.2(a), 2(b) and 2(c).
The monthly PM 2.5 /PM 10 ratios of Wuhu were in the range of 0.53-0.73,with an average of 0.67 in 2015, of 0.61-0.74and with an average of 0.66 in 2016; and of 0.42-0.89and with an average of 0.54 in 2017.As for Bengbu, the annual PM 2.5 /PM 10 ratios ranged between 0.61 and 0.79 and averaged 0.70 in 2015, ranged between 0.57 and 0.79 and averaged 0.65 in 2016, and between 0.43 and 0.72 and averaged 0.59 in 2017, respectively.The PM 2.5 /PM 10 ratio of the three-year range in Wuhu was 0.42-0.89,with an average of 0.63, was 0.43-0.79,and with an average of 0.65 in Bengbu.These results show that the annual average of PM 2.5 /PM 10 ratios in both Wuhu and Bengbu were all lower than those reported in the Beijing-Tianjin-Hebei region (0.83), the Yangtze River Delta region (0.76) and the Pearl River Delta region (0.74) (Chen et al., 2017).This result was also consistent with the conclusion of Zhou et al. (2015), which reported there was a strong positive correlation between PM 2.5 and PM 10 mass concentration in the atmosphere.In particular, the trend of a decreasing PM 2.5 /PM 10 ratio was similar to that seen for the PM 2.5 concentration.
For Wuhu, during 2015, the three highest monthly averages of PM 2.5 /PM 10 ratios were 0.73 in November, 0.72 in December and 0.70 in January; while in 2016, the three highest monthly averages of PM 2.5 /PM 10 ratios were 0.74 in December, 0.72 in July and 0.72 in January; in 2017 the three highest monthly averages of the PM 2.5 /PM 10 ratios were 0.89 in December, 0.67 in November and 0.65 in February.However, in 2015, the three lowest monthly averages of the PM 2.5 /PM 10 concentration ratios were 0.53 in June, 0.64 in July and 0.66 in September; while in 2016, the three lowest monthly averages of the PM 2.5 /PM 10 concentration ratios were 0.61 in April, 0.62 in September and 0.64 in August; in 2017, the three highest monthly averages of the PM 2.5 /PM 10 ratios were 0.42 in April, 0.44 in May and 0.51 in July.For Bengbu, during 2015, the three highest monthly averages of the PM 2.5 /PM 10 ratios were 0.79 in November, 0.78 in December and 0.76 in February; while in 2016, the three highest monthly averages of the PM 2.5 /PM 10 ratios were 0.57 in September, 0.57 in April and 0.59 in August; in 2017 the three highest monthly averages of the PM 2.5 /PM 10 ratios were 0.72 in February, 0.71 in January and 0.68 in December.However, in 2015, the three lowest monthly averages of the PM 2.5 /PM 10 ratios were 0.61 in April, 0.64 in June and 0.64 in October; while in 2016, the three lowest monthly averages of the PM 2.5 /PM 10 ratios were 0.57 in September, 0.57 in April and 0.59 in August; in 2017 the three lowest monthly averages of the PM 2.5 /PM 10 ratios were 0.43 in May, 0.51 in April and 0.55 in June, respectively.
In general, it was found that a higher PM 2.5 /PM 10 ratio always accompanied a higher PM 2.5 concentration.This demonstrated that PM 2.5 is the major portion of atmospheric particles.A previous study (Tang et al., 2017) reported the atmospheric particles were mostly due to the gas-particle transformation, going through the condensation and flocculation processes, and then the accumulation mode of PM 2.5 .Furthermore, a certain fraction of the PM 2.5 concentration in the ambient air is due to the resuspension of road dust or entrainment of naked lands.A higher PM 2.5 /PM 10 ratio signifies more harmful air pollution to human health, and thus should be a case for more concern.

PCDD/F Concentration
Previous studies have shown the strong correlation between PM 10 and PCDD/F mass concentrations, and Lee et al. (2016), Suryani et al. (2015) and Huang et al. (2011a) reported the correlation coefficients were as high as 0.98, 0.99 and 0.94, respectively.In Tang et al. (2017), the monthly concentrations of PCDD/Fs in the ambient air were simulated by PM 10 using the regression analyses (Wang et al., 2010).Based on the PM 10 concentrations, the total PCDD/F mass concentrations were calculated using Eq.(1) and Eq. ( 2).
The results shown that in Wuhu, the total PCDD/F mass concentrations were in the range of 0.65-1.94pg m -3 and with an average of 1.09 pg m -3 in 2015, of 0.51-1.61pg m -3 and with an average of 1.01 pg m -3 in 2016, and of 0.63-1.43pg m -3 and with an average of 1.07 pg m -3 in 2017; while in terms of concentrations of toxicity equivalent quantity (TEQ), those ranged between 0.026 and 0.085 pg-WHO 2005 -TEQ m -3 in 2015, ranged between 0.021 and 0.070 pg-WHO 2005 -TEQ m -3 in 2016, and ranged between 0.025 and 0.074 pg-WHO 2005 -TEQ m -3 in 2017, and the corresponding average values were 0.050, 0.047 and 0.050 pg-WHO 2005 -TEQ m -3 , respectively.The time variation results show that both the average of the total PCDD/F mass and total-PCDD/Fs-WHO 2005 -TEQ concentrations decreased slowly, and this may be due to a better control of domestic emissions.Comparing with previous studies, the values of the total-PCDD/Fs-WHO 2005 -TEQ concentrations were all at the same levels of those seen in the Kaohsiung area, ranging between 0.021 and 0.077 pg-WHO 2005 -TEQ m -3 and with an average of 0.048 pg WHO 2005 -TEQ m -3 for 2014, and between 0.021 and 0.072 pg WHO 2005 -TEQ m -3 and with an average of 0.044 pg WHO 2005 -TEQ m -3 for 2015 in southern Taiwan (Lee et al., 2016).
In Bengbu, the total PCDD/F mass concentrations were in the range of 0.82-1.53pg m -3 with an average of 1.17 pg m -3 , of 0.73-1.82pg m -3 and with an average of 1.17 pg m -3 , of 0.71-1.85pg m -3 and with an average of 1.31 pg m -3 in 2015, 2016 and 2017, respectively; while in terms of concentrations of toxicity equivalent quantity ranged between 0.033 and 0.073 pg-WHO 2005 -TEQ m -3 in 2015, between 0.029 and 0.080 pg-WHO 2005 -TEQ m -3 in 2016, ranged between 0.029 and 0.085 pg-WHO 2005 -TEQ m -3 in 2017, and the corresponding average values were 0.054, 0.054 and 0.061 pg-WHO 2005 -TEQ m -3 , respectively.On the whole, both the average of the total PCDD/F mass and total-PCDD/Fs-WHO 2005 -TEQ concentrations were all at high levels.
These results show that the lowest total PCDD/Fs-WHO 2005 -TEQ concentration of the two cities were all occurred in summer, which means that the levels of particulate matter had a significant effect on the total PCDD/F concentrations.Therefore, a higher total PCDD/Fs-WHO 2005 -TEQ concentration always accompanied with a higher level of particulate matter.As such, controlling the PM emissions from sources will subsequently result in reductions in ambient dioxin levels.

Gas-Particle Partitioning of PCDD/Fs
The gas-particle partitioning of PCDD/F plays an important role in the efficiency of the atmospheric wet and dry deposition (Bidleman and Harner, 2000).Several factors are important here, such as the ambient temperature, PCDD/F concentrations, vapor pressure and the atmospheric particulate concentration (Hoff et al., 1996).The gas-particle partition was calculated by meteorological date using Eq.(3), Eq. ( 4) and Eq. ( 5), and the seasonal gas-particle partitioning of total PCDD/Fs-WHO 2005 -TEQ in the ambient air of Wuhu and Bengbu during 2015-2017 are shown in Figs.3(a), 3(b) and 3(c).
In Bengbu, the seasonal average temperatures during 2015 were 15.9, 27.5, 16.9 and 5.0°C in spring, summer, fall and winter, respectively; in 2016, the seasonal average temperatures in spring, summer, fall and winter were 16.3, 27.3, 16.9 and 4.5°C, respectively; while in 2017, the average temperatures of spring, summer, fall and winter were 16.5, 28.0, 16.2 and 4.9°C, respectively.The fractions of gas phase total PCDD/Fs-WHO 2005 -TEQ concentration in spring, summer, fall and winter were 30.7%, 71.5%, 36.5% and 8.7% in 2015; 31.2%,73.5%, 39.9% and 7.4% in 2016; and 28.6%, 72.7%, 37.3% and 7.2% in 2017.The above results show that the fractions of gas phase total PCDD/Fs-WHO 2005 -TEQ of the both cities were all lower than that in the particle phase in spring, fall and winter, but were significantly higher than that in particle phase in summer (Figs. 3(a), 3(b) and 3(c)).

Fig. 3(c).
Seasonal variations of gas-particle partition of total-PCDD/Fs-WHO 2005 -TEQ in the ambient air in Wuhu and Bengbu, respectively, during 2017.
As for Wuhu, the three-year average fractions of gas phase total PCDD/Fs-WHO 2005 -TEQ in spring, summer, fall and winter were 35.2%, 76.4%, 44.8% and 10.7%, respectively; the three-year average temperatures in spring, summer, fall and winter were 17.3, 27.8, 18.5 and 6.5°C, respectively.While in Bengbu, the three-year average fractions of gas phase total PCDD/Fs-WHO 2005 -TEQ in spring, summer, fall and winter were 30.2%, 72.6%, 37.9% and 7.8%, respectively; and the three-year average temperatures in spring, summer, fall and winter were 16.2, 27.6, 16.7 and 4.8°C, respectively.The above results indicate that the fractions of gas phase total PCDD/Fs-WHO 2005 -TEQ in summer were highest and those in winter were lowest, while the values of spring and fall were both in the middle levels, but the former were slightly lower than the latter.The fraction of gas phase total PCDD/Fs-WHO 2005 -TEQ had a significant, positive correlation with air temperature.As the temperature increased, a certain fraction of particle phase PCDD/F evaporated into the gas phase, and thus the gas phase total PCDD/Fs-WHO 2005 -TEQ increased with an increasing temperature.In addition, the average temperature of Wuhu in spring, summer, fall and winter were all slightly higher than those in Bengbu, while the corresponding fraction of gas phase total PCDD/Fs-WHO 2005 -TEQ were slightly higher than those in Bengbu (Figs. 3(a), 3(b) and 3(c)).
The results indicate that lower molecular weight PCDD/F congeners were primarily in the gas phase, while the particle phase was usually associated with higher molecular weight PCDD/F congeners (Wu et al., 2009a;Lin et al., 2010;Huang et al., 2011a;Mi et al., 2012;Suryani R. et al., 2015).The gas phase PCDD/Fs had higher fractions in summer than in winter, similar to in previous studies (Xu et al., 2009;Wang et al., 2010;Huang et al., 2011a;Lee et al., 2016).This may due to lower molecular weight PCDD/Fs usually having higher vapor pressure (Wang et al., 2010;Huang et al., 2011a).As the ambient temperature increased, the fraction of gas phase PCDD/Fs also rose, while when the temperature decreased some of the gas PCDD/Fs were exchanged and transferred into the particle phase.As a result, lower molecular weight PCDD/Fs primarily existed in the gas phase, and the gas phase PCDD/Fs fractions increased with increasing temperature.
As for the seasonal variation, for Wuhu, in 2015, the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents in spring, summer, fall and winter were 0.496, 0.190, 0.410 and 0.580 ng-WHO 2005 -TEQ g -1 , respectively; and in 2016 they were 0.530, 0.134, 0.389 and 0.561 ng-WHO 2005 -TEQ g -1 , respectively; while in 2017 they were 0.786, 0.173, 0.445 and 0.588 ng-WHO 2005 -TEQ g -1 , respectively.It was found that, in Wuhu, the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents in summer were roughly only 68.8% in magnitude lower than the average value of the other three seasons (spring, fall and winter).For Bengbu, in 2015, the PM 2.5bound total PCDD/Fs-WHO 2005 -TEQ contents in spring, summer, fall and winter were 0.569, 0.183, 0.441 and 0.549 ng-WHO 2005 -TEQ g -1 , respectively; while in 2016, they were 0.578, 0.178, 0.496 and 0.575 ng-WHO 2005 -TEQ g -1 , respectively; and in 2017, they were 0.745, 0.201, 0.509 and 0.591 ng-WHO 2005 -TEQ g -1 , respectively.The results indicated that in Bengbu the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents in summer was approximately 66.7% in magnitude lower than the average values of other three seasons (spring, fall and winter).This was because that the ambient temperature in summer was much higher than the average values of other three seasons (spring, fall and winter) and more of the PCDD/Fs bound to the particles were evaporated to the gas phase, and so the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents decreased.

Sensitivity Analysis
Sensitivity analysis can provide a better basis for confirming which environmental parameters are important to the atmospheric PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents.Several parameters, such as the ambient temperature, PM 2.5 concentration, PM 10 concentration and total PCDD/F mass concentration, could affect the PM 2.5bound total PCDD/Fs-WHO 2005 -TEQ content.In this study, sensitivity analyses were carried out depending on the initial values of ambient air temperature = 22.2°C, PM 2.5 = 58.0µg m -3 , PM 10 =87.0 µg m -3 and total-PCDD/F mass concentration =1.12 pg m -3 .The parametric sensitivity for the atmospheric PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents of Wuhu is shown in Fig. 5. where P: initial value of parameters; ΔP: increase or reduction in parameters; S: predicted value in each of the parameters at the initial value; ΔS: response in each of the parameters.
The sensitivity analysis indicated that the total PCDD/F mass concentration was the most sensitive parameter for atmospheric PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents.When ΔP/P was changed from 0% to +20%, +50%, ΔS/S responded from 0% to +42%, +106%.This may be because PCDD/Fs are the root cause of total PCDD/Fs-WHO 2005 -TEQ, so the change in PCDD/Fs mass concentration has a significant effect on PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents.
From Fig. 5, it also can be seen that the PM 10 concentration is also an important sensitive parameter for the atmospheric PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents.When ΔP/P was changed from 0% to +20%, +50%, ΔS/S responded from 0% to +29%, +72%.There was a strong correlation between PM 10 and total PCDD/F mass concentrations, and the change in PM 10 concentration has a great impact on the PCDD/F mass concentration, and thus it can affect the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents in a significant manner.
The sensitivity analysis also demonstrated that an increase in air temperature has a significant, negative effect on atmospheric PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ Fig. 5. Sensitivity analysis for the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ content of Wuhu.contents, when ΔP/P was changed from 0% to 50%, ΔS/S responded from 0% to -112%.But a decrease in air temperature has less effect on the atmospheric PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents, when ΔP/P was changed from 0% to -50%, ΔS/S responded from 0% to +73%.The temperature affects the atmospheric PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents through changing the gas-particle partitioning of PCDD/Fs, the high molecular weight PCDD/Fs have a large contribution to the total PCDD/F mass concentration and primarily existed in the particle phase.When the air temperature was increasing, the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents decreased obviously as more of the particle-bound PCDD/Fs were evaporated to the gas phase; when the temperature was not high, the PCDD/Fs were mostly existed in the particle phase, when the temperature decreased, the rest of the gas phase PCDD/Fs changed into the particle phase, and thus the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents increased.This is consistent with the conclusion that the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ content in the summer season was only approximately 34.1% in magnitude lower than the average of the other three seasons (spring, fall and winter).
A decrease in the PM 2.5 concentration had a negative or positive correlation with the atmospheric PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents, when ΔP/P was changed from 0% to -50%, ΔS/S responded from 0% to -66%.The effect of increasing PM 2.5 can also be divided into two stages: when ΔP/P was changed from 0% to +30% and +70%, ΔS/S responded from 0% to +23% and +33%, respectively, but when ΔP/P was changed from +70% to +100%, ΔS/S responded from +33% to +25%.The PM 2.5 concentration affects the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents mainly in terms of the particle-bound PCDD/Fs.Lower PM 2.5 concentrations were always accompanied by a better atmospheric stability, which is beneficial to the dispersion of air pollutants, and thus PCDD/Fs also decreased obviously, and the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ content was the value of the total-PCDD/Fs-WHO 2005 -TEQ/PM 2.5 ratio, and the atmospheric PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents also decreased slowly.However, when the PM 2.5 concentrations were higher than 86 µg m -3 , because the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents were calculated by the ratio of the total PCDD/Fs-TEQ concentration divided by that of PM 2.5 , a higher PM 2.5 concentration was always accompanied by a low air temperature, and thus the PM 2.5 /PM 10 ratio was high, which means the rate of increase of the PM 2.5 concentrations was greater than that of PM 10 , and also higher than that of the total PCDD/F mass concentration, and the rate of increase of the PM 2.5 concentration was greater than that of total PCDD/Fs-WHO 2005 -TEQ concentration, and therefore the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents decreased when the PM 2.5 concentration continuous to rise.
The result of the sensitivity analysis suggested that atmospheric PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents were most sensitive to total PCDD/F mass concentration, followed by PM 10 concentration, and then the air temperature and PM 2.5 concentration.

CONCLUSION
The results of the current investigation of both PM 2.5 and PCDD/Fs in Wuhu and Bengbu are summarized as follows: 1.The PM 2.5 concentration of the focal three years in Wuhu ranged between 25.0 and 108.0 µg m -3 , with an average of 53.0 µg m -3 ; for Bengbu, the figures were 33.0-101.0µg m -3 and with an average of 61.4 µg m -3 .The PM 2.5 level in Wuhu was lower than that in Bengbu.In general, the time variation of the PM 2.5 concentration in both Wuhu and Bengbu fell slowly, but was still above the WHO air quality regulated standard (10 µg m -3 ). 2. As for the seasonal variations, the three-year average of PM 2.5 concentration for Wuhu in summer (32.3 µg m -3 ) was 58.2% in magnitude lower than that in winter (77.3 µg m -3 ); and the value for Bengbu in summer (41.9 µg m -3 ) was 51.7% in magnitude lower than that in winter (86.8 µg m -3 ).The high temperature in summer contributed to more violent vertical air current transport, which can accelerate PM 2.5 dispersion, while stronger rainfall and wind speed also enhanced the removal of PM 2.5 .In winter, the low temperature hindered PM 2.5 dispersion, and polluted air from northern cities caused the PM 2.5 concentration to increase.3. The PM 2.5 /PM 10 ratios of the three-year range in Wuhu were in the range of 0.42-0.89,with an average of 0.63, and were 0.43-0.79with an average of 0.65 in Bengbu.A higher PM 2.5 /PM 10 ratio was always associated with a higher PM 2.5 concentration.This means that PM 2.5 is the major portion of atmospheric particles.4. The total PCDD/Fs-WHO 2005 -TEQ concentration of Wuhu in summer (0.028 pg-WHO 2005 -TEQ m -3 ) was 48.1% in magnitude lower than that in other three seasons (0.054 pg-WHO 2005 -TEQ m -3 ); and the value for Bengbu in summer (0.035 pg-WHO 2005 -TEQ m -3 ) was 44.4% in magnitude lower than that in other three seasons (0.063 pg-WHO 2005 -TEQ m -3 ).The values for the total-PCDD/Fs-WHO 2005 -TEQ concentrations of these two cities were all higher than those in previous studies of Taiwan (0.046 pg-WHO 2005 -TEQ m -3 ). 5. In Wuhu, the three-year average fraction of gas phase total PCDD/Fs-WHO 2005 -TEQ concentration in summer (76.4%) was 2.53 times than that in the other seasons (30.2%); while in Bengbu, the fraction in summer (72.6%) was 2.87 times than that in the other seasons (25.3%).As such, the fraction of gas phase total PCDD/Fs-WHO 2005 -TEQ had a significant, positive correlation with air temperature.6.In Wuhu, the three-year average PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ content in summer (0.166 ng-WHO 2005 -TEQ g -1 ) was approximately only 68.8% in magnitude lower than the average value of the other three seasons (0.532 ng-WHO 2005 -TEQ g -1 ); while in Bengbu, that of in summer (0.187 ng-WHO 2005 -TEQ g -1 ) was approximately 66.7% in magnitude lower than the average value of the other seasons (0.561 ng-WHO 2005 -TEQ g -1 ).Due to the higher temperature in summer, more of the particle-bound PCDD/Fs evaporated to the gas phase, and the PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ contents thus decreased.7. The sensitivity analysis of PM 2.5 -bound total PCDD/Fs-WHO 2005 -TEQ in Wuhu showed that the PCDD/F concentration was the most positively correlated sensitive factor, when ΔP/P was changed from 0% to +20%, +50%, ΔS/S responded from 0% to +42%, +106%.The second positively correlated sensitive factor was the PM 10 concentration, when ΔP/P was changed from 0% to +20%, +50%, ΔS/S responded from 0% to +29%, +72%.This was followed by the atmospheric temperature, and its effect was negatively correlated, when ΔP/P was changed from 0% to +50%, ΔS/S responded from 0% to -112%; when ΔP/P was changed from 0% to -50%, ΔS/S responded from 0% to +73%.The last sensitive parameter was the PM 2.5 concentration, and the impact was divided into two stages: when ΔP/P was changed from 0% to +30%, +70%, ΔS/S responded from 0% to +23%, +33%; but when ΔP/P was changed from +70% to +100%, ΔS/S responded from +33% to +25%.8.The results of this study can provide useful information in the search for more insights into both atmospheric PM 2.5 and PCDD/Fs.

Fig. 3
Fig. 3(a).Seasonal variations of gas-particle partition of total-PCDD/Fs-WHO 2005 -TEQ in the ambient air in Wuhu and Bengbu during 2015.

Fig. 3
Fig. 3(b).Seasonal variations of gas-particle partition of total-PCDD/Fs-WHO 2005 -TEQ in the ambient air in Wuhu and Bengbu during 2016.