Performance Comparison Research On Medium Flow Rate Sampler For Atmospheric Fine Particulate Matter(PM2.5)

As the PM2.5 pollution problem increasingly prominent, PM2.5 observation and research work have been carried out in many cities. Measure error of PM2.5 sampler will affect the accuracy of the measurement results and accurate monitoring of mass concentrations of particulate matter is very important for air quality assessment and contamination control. In order to comparing study between gravimetric monitoring methods and automatic monitors, the hourly automatic monitoring data and PM2.5 samples were collected at nine atmospheric monitoring stations in Changchun city from January 2014 to January 2015. At the same time, three types of the domestic PM2.5 cutters comparison experiment were performed and correlation analysis and composition analysis were conducted.In September 2014 and in March and July 2015, TSP, PM10, PM5 and PM2.5 samples were collected using three kinds of manual PM2.5 medium flow rate samplers, which were equipped with TH-100(CTH), TSP-100(CLS) and PMN-100(CZL) cutters, to analyze the characteristics of particulate size distribution and the monthly variation of particulate matter. According to the statistic results the order of particulate size distribution is PM2.5>PM10-TSP>PM5-10>PM2.5-5 and the PM2.5/TSP ratios and PM2.5/PM10 ratios are highly in March than in September and July. Composition of particulate is analyzed, focusing on water soluble ions and metal elements. Results show that SO42-, NH4+, K+ and Mg2+ concentration are significantly high in PM2.5 than other particulate size distribution. Meanwhile, the most abundant metal species in PM2.5 are Ca, K, Fe, Al, Mg and Zn. Experimental results suggested that pollutant compositions of samples collected by three kinds of medium flow rate sampler are similar, which indicates that composition of particulates depends on the efficiency of cutter and the particulate concentration.Correlation analysis between TSP, PM10, PM5 and PM2.5 parallel samples collected by three kinds of sampler shows that CTH and CZL has the largest correlation coefficient, the correlation coefficient of CLS and CZL comes second and CTH and CLS data correlation is the worst. Further research of deviation is being carried out with PM2.5 concentration detected by three samplers and the results suggest that the deviation between CTH and CLS is the largest, followed by the deviation between CLS and CZL, and the deviation between CTH and CZL is relatively small. The order of the PM2.5 concentration is CZL> CLS > CTH. And paired-samples T test results show no significant difference between CZL, CLS and CTH.An comparison of PM2.5 mass measured simultaneously by KC-120 H samplers and Thermo 5030 monitor was carried out at nine atmospheric monitoring stations in Changchun city from January 2014 to January 2015. Additionally variations of PM2.5 concentrations at different monitor sites, different month and different concentration lever were analyzed. The results showed that the annual average PM2.5 concentrations of Thermo5030 are higher than KC-120 H at all of monitoring sites. The year average PM2.5 concentrations of the city is respectively 102.10μg/m3 and 77.45μg/m3 and C5030 is higher than C120 by approximately 29.08%. The characteristics of the monthly variations of C5030 and C120 is that C5030 is higher than C120 and the difference in the range of 18.48 to 67.89μg/m3 in January, March, October 2014 and January 2015. Meanwhile, C5030 is lower than C120 by 11.73% and 27.57% in April and July 2014. The results of the classification of 183 PM2.5 samples shows that 78 of the KC-120 H samples are in fine level which account for 43%, while 55 and 40 of the Thermo5030 samples are in fine and light pollution lever respectively, which account for the ratio of 30% and 22%.The correlation analysis results of PM2.5 data between KC-120 H and Thermo5030 shows that the correlation coefficients are in the range of 0.802-0.511 except the station of institute of Posts and Telecommunications.The correlation coefficients of between different month data are in the range of 0.818-0.499 except January 2015. The regression analysis of C5030 and C120 is found that the slope of regression equation is 0.8631, intercept is 33.828 and R2 is 0.4299. It shows that the results of two samplers have good correlation. According to different level of C120 and corresponding C5030 data, the correlation is analyzed. The results indicate that the higher concentration of C120, the better correlation is found between two kinds of sampler measurement. The deviation analysis between PM2.5 of two sampling methods shows that the deviation of October 2014 and January 2015 are the largest, followed by the deviation of January and March 2014, and the April and July 2014 are relatively small. It is also found that Thermo5030 monitor is overestimate the value of PM2.5 concentration in January, March, October 2014 and January 2015, while it is underestimate in April and July 2014. Annual data shows that Thermo5030 monitor is overestimate the value of PM2.5 concentration. The analysis of C5030 and C120 data deviation according to different degrees of pollution concentration find that C5030 has a positive deviation in all degree and the higher concentration of C120, the lower deviation between C5030 and C120. Paired-samples T test results show that C5030 and C120 have no significant difference.