3.1. PM2.5 and Black Carbon
The temporal trends of BC and PM2.5 were showed in Fig.1, the annual mean of PM2.5 (Fig.1(a) red line) firstly increased(2003-2015) and then decreased(2015-2018) over the past 16 years. From 2003 to 2015, consistent increase of PM2.5 was caused by rapid urbanization, population explosion and increase in the number of private cars. According to statistics, in 2018, the car ownership in Beijing reached 6.08million, 2.85 times as much as that in 2005 (https://baijiahao.baidu.com/s?id=1643270530939456964&wfr=spider&for=pc) Moreover, in 2017 there were 21.71 million people in Beijing compared with 14.56 million in 2003, increased by 7.15 million (http://www.stats.gov.cn/). While many effective environmental protection policies and laws including coal-to-gas heating, relocation of heavy pollution industries and promotion of electric vehicles etc. reduced the concentration of PM2.5 evidently in last several years. Autumn and winter can cover the whole heating season, usually from November to March of next year, in Beijing, so spring-summer and autumn-winter were divided into groups respectively to reveal more details. From Fig1.(b), PM2.5 was much higher in autumn-winter than that in spring-summer except 2007, which was caused by the influence of bad atmospheric conditions(such as atmospheric circulation, sunshine radiation and dry weather)(Jin., 2014) ,coal burning, vehicle emissions and other pollution sources. The concentrations of autumn-winter declined year by year from 2015, also indicating there were achievements of the above policies.
The measurement of BC concentration in samples has begun since 2007. It is no secret that BC is the significant composition of PM2.5 and mainly originate from coal combustion, vehicle emission or biomass burning(Mousavi, et al., 2019). The mean concentrations of BC(blue line in Fig.2(a)) also consistently increased from 2007, peaking in 2015(8120ng/m3), and decreased from 2015 to 2018. Like PM2.5 seasonal trends, the concentrations of BC were more severe in autumn-winter(Fig.1 (b)). From 2007 to 2015, there was an increase trend in the two seasons, and it also showed a decrease trend between 2015 and 2018. BC had a strong correlation with PM2.5 (Fig.2.(b), R2=0.676,slope=0.0718), occupying approximate 11% in total PM2.5 in Liangxiang. However, in recent three years, its proportion in fine particle has shifted down significantly, and this figure was 8.5% in 2018, which may be due to cut down the capacity of coal burning and restrict some pollution industries by the government.
3.2 Common elements
Six years, 2007, 2009, 2013, 2014, 2016, 2017, has been selected because the elements’ data of other years from 2007 to 2017 were incomplete or some samples were not analyzed. In Figure.3(a) and Figure S1. the soil and crustal elements like Si, Mg and Al had similar temporal trends, and the frequency of dust weather has decreased compared with before in recent years. The K is one of the fingerprint elements about biomass (Zheng et al., 2004), a continuous decrease of concentration showed the reduction of biological emission. S, Mn and Fe had the similar temporal trends, indicating that steel and heavy industries developed rapidly before 2013, and then they were controlled effectively in these years. Shougang Group, for example, had stopped production, and relocated in Hebei province in 2011. The of concentrations for Pb, Zn and Br had the overall upward trends, while the obvious decline trends in 2017 might be well caused by the coal to gas policy and pollution industries management. Cl had a clear upward trend and seasonal variations, the concentrations in autumn-winter were much higher than it in spring-summer. The reasons of aggregation of Cl in autumn and winter are as followed: Firstly, heating in winter caused a large number of coal burning in Beijing, which is the main source of Cl (Yu. et al.,2018.); secondly, most chemical compound with Cl is volatile in high temperature, while the environment conditions make it aggregation easily in winter and autumn. In addition, there were no clear law about Cr and Ni. Further study will be done in the next section of PMF. All measured elements and BC these years accounted for 30%-40% in total PM2.5(showed in Fig. 3(b)), which was similar with the percentage(37%) in Tianjin, a city bordering on Beijing. Therefore, in most cases, inorganic salts (NO3-, SO42-, NH3+ etc.), organic matter and other secondary pollutants may be the main component in PM2.5.
3.3 Element In
70 samples from 2017.1 to 2018.6 were analyzed by NAA(Neutron Activation Analysis)( Kothai et al., 2008, D.D.Lestiani et al., 2011, Abel, et al., 2011.). The parameters of NAA and statistical data of various elements were showed in Table S2 and Table 1. One element, In, was special and had a strong relevance with other elements. According to the results, though the In was extremely slight content in PM2.5, it had a good correlation, about 0.797, with the fine particles. Moreover, the In correlated with BC, and halogenated elements contents in PM2.5, like Cl, Br and I obviously(R2In-BC=0.879, R2In-Cl=0.399, R2In-Br=0.796, R2In-I=0.775), while there was little relevance between In and soil elements(R2In-Dy= 0.074, R2In-Al=-0.138, R2In-Ti=0.134), explaining it maybe originate from combustion of fossil fuel or vehicle emission instead of crust or soil. In addition, anode of battery or some electrical materials also dope a few In element to improve the performance(Yu., 2018, Fen., et al, 2013). Obvious correlation with Cu, Na and K(R2In-Cu=0.761, R2In-Na=0.708, R2In-K=0.751) also displayed that biomass combustion could be one of the origins of element In(Taghvaee, et al., 2018, Santoso, et al., 2008). Therefore, it was possible that the In could be a fingerprint and semi-fingerprint element to identify some specific sources.
Table1 The statistical data of elements analysed by NAA
Species
|
Mean(ng/cm3)
|
STDEV(ng/cm3)
|
Max(ng/cm3)
|
Min(ng/cm3)
|
R2 (with In)
|
PM2.5
|
67955
|
50851
|
245568
|
563
|
0.797**
|
BC
|
5315
|
3296
|
14874
|
8584
|
0.879**
|
Dy
|
0.053
|
0.084
|
0.683
|
0.032
|
0.074
|
Ba
|
15.2
|
10.7
|
70.0
|
2.47
|
0.351*
|
Ti
|
52.4
|
77.6
|
608
|
7.38
|
0.134
|
I
|
5.46
|
5.07
|
28.6
|
0.538
|
0.775**
|
Br
|
11.2
|
13.1
|
75.3
|
1.13
|
0.796**
|
Mn
|
24.9
|
13.8
|
73.9
|
5.62
|
0.597**
|
Mg
|
307
|
305
|
2324
|
56.1
|
0.054
|
Cu
|
12.6
|
8.41
|
33.7
|
1.10
|
0.761**
|
In
|
0.032
|
0.023
|
0.105
|
0.002
|
1
|
Na
|
284
|
270
|
2158
|
45.4
|
0.708**
|
V
|
2.43
|
2.61
|
14.9
|
0.481
|
0.473**
|
K
|
621
|
451
|
3142
|
130
|
0.751**
|
Al
|
874
|
1433
|
11708
|
102
|
-0.138
|
Cl
|
104
|
80.7
|
433
|
34.1
|
0.399
|
Ca
|
829
|
542
|
3523
|
140
|
-0.204
|
Note: “*”indicates a significant correlation at the 0.05 level (both sides)
“**” indicates a significant correlation at the 0.01 level (both sides)
3.3 Source apportionment
PMF analysis further defined the pollution sources in 2007 and 2016~2017, because the same methods (PIXE, Beijing Normal University) were used in the two periods. A total of 17 species (BC, 15 elements and PM2.5) were used as the EPA PMF 5.0 input matrix, and the statistical data were showed in Table. S3. In the end, six factors were chosen in both periods and the PMF predicted results showed quite well correlation with observed results (Fig.S2. R22007=0.928 slope=1.0466, R22016-2017=0.931 slope=1.355).
Fig.4. showed there were three common sources, i.e., soil and dust(19.7% and 10.5%), coal combustion(16.3% and 18.8%) and industry-S(27.5% and 22.5%) in both two periods. Mg, Al, Ti, Si and Ca were the characteristic elements in soil and surface dust(Yu et al.,2018.), and mainly came from long-range transport such as Loess Plateau, Mongolia and Kazakhstan.(Huang et al., 2017). Some elements rich in dust, like Pb, does not simply reflect in soil and crust, as it could also originate from vehicle emission or nearby construction sites. This factor contributed 19.7% and 10.5% in total PM2.5 in 2007 and 2016~2017, respectively, showing a significant decrease. A large-scale afforestation project, for example “the Green Wall of China”, has been carried out in the Loess Plateau in the past two decades(Jing, et al., 2021). The frequency of dust weather in Beijing has decreased from 13 every year in the beginning of the new century to 2-3 in recent years, indicating the great achievements of these soil conservation projects.
Coal combustion, determined by Cl and BC(Yu et al.,2018), explained 18.8% in 2016~2017, which was a small increase compared with 16.3% in 2007. The factor in 2007 were involved in a little percentage of Cu, Zn, Br and Pb, all of which were characteristic elements of metal industries, e.g., dye industries for Cu(Chuang. et al., 2016) and were co-discharged with coal burning. In northern China, a large number of coal would be consumed for heating in winter every year. The sampling site was quietly close to Hebei province, there was still large amounts of coal-burning during winter heating. Moreover, northern and eastern China are the main region of coal combustion emission(Jin et al., 2014), the regional transport of pollutants exerted a negative influence on Beijing. In addition, due to policy of coal-to-gas in 2017, some polluting factories and enterprises were still relocating or plan to relocate. These reasons may cause a slight jump in percentage of coal combustion source in 2016~2017 than it in 2007 and some metal elements were also disappeared these years.
The third common pollution source, industry-S, accounted for 27.5% and 22.5% in their own samples, respectively. Beijing is the center of Beijing-Tianjin-Hebei industrial area, industries or fuel burning emissions are the most important source of SO2 and BC, and then these SO2 were transformed into SO42- or secondary sulfur by photochemistry in the air (Mang, et al, 2018). Fe, Zn, Br and Pb were also involved in this factor in 2007, indicating some metal industries pose a threat to the environment(Tao J., et al, 2014). There were some small factories and workshops in Xinzhen ten years ago. While the decrease of this factor percentage in 2016~2017 proved relocating and shutting down industries had a good effect on local air.
Besides that, three pollution sources of the two periods were different from each other. Steel manufactory and vehicle accounted for 18.9% in PM2.5 in 2007. Zn, Br and BC were mainly emitted from gasoline and diesels and Cu may come from car tie wear (Zheng, et al., 2004, Pant, et al., 2018). Also, some elements of steel industries like Cr, Mn and Fe were involved in this factor. In 2016 ~ 2017, industry-Zn and vehicles(also determined by Br, BC, Cu and crustal elements ) contributed 19.8% to PM2.5. Despite as an additive in two-stroke engine lubricating oils(Yu. et al.,2018.), high percentage Zn also associated with some electronical companies and industries (Zhang et al., 2014). There were many high-tech companies, e.g. new energy, electronic or material companies, located in Fengtai science and technology park about 20km away from sampling site. Moreover steel industries was also a serious pollution source with a portion of 18.3% in this period and most of it may sperate from area transmission. Element K and some microelements like Zn, Cu, were characteristic tracer for biomass burning which contributed 15.5% to the PM2.5 in 2007(Taghvaee, et al., 2018, Santoso, et al., 2008). One decade ago, there were many farmland in Beijing suburbs, especially in Xinzhen, tons of straws and waste burning in late autumn and winter was the main reason of this pollution source emission. But as the law of prohibition of straws burning implemented in 2007, the proportion of this factor would decrease year by year. Therefore, there was no obvious factor with high level of K in 2016~2017.
Cement plant source was mainly determined by Ca and contributed small percentage(2.2%) to PM2.5 in 2007. There was a cement plant not far from the sampling point in the past, but it was demolished several years ago. In 2016 ~ 2017, new kindergarten was built and the student dormitory was redecorated around the sampling site during this period. Therefore, construction, which was identified by high level of Ca and some Mg, Al, Si, Ti, Fe and Cu in concrete and cement, accounted for a relatively larger proportion, about 10.1%.