Page 7 - 《JES》202007

jo urna l o f e nviro n mental scie nces 93 (2020) 1 e12 3

under the on-line VOCs monitoring system, following a pre- alkyne (1), aromatics (16), halocarbons (28), and OVOCs (13).
viously described procedure (Gu et al., 2019). The positive Table S1 shows the measured concentrations (average and
pressure in the laboratory prevented the inflow of polluted associated standard deviations, and maximum and minimum
outdoor air. C 2 eC 5 hydrocarbons were determined by flame values) of all 99 VOCs. The concentration of total VOCs
ionization detection after the separation using an HP-PLOT (TVOCs) varied in the order of winter polluted days
3
column (30 m long, 320 mm internal diameter (i.d.), 0.2 mm (195.40 ± 140.53 mg/m ) > summer polluted days
3
film thickness). Their quantification was carried out by an (114.87 ± 40.98 mg/m ) > summer normal days (90.20 ± 40.02 mg/
3
3
external standard method. Other VOCs were determined by m ) > winter normal days (55.40 ± 43.60 mg/m ), which is
mass spectrometry after the separation using an HP-1 column consistent with the trends reported by Wei et al. (2018) and An
(60 m long, 320 mm i.d., 1.0 mm film thickness). The oven et al. (2017). Compared with the result of Gu et al., 2019, the
temperature was initially set at 35 C and held for 5 min, then VOC concentration of Beijing at 10 am was close to the daily

increased sequentially at 5 C/min to 150 C and at 15 C/min to average concentration under appropriate temperature and

220 C; the latter temperature was held for 3 min. The carrier solar radiation. Fig. 1 shows the concentrations of the VOCs

gas was the purified helium with the flow rate at 1.5 mL/min. during the different seasons in Beijing that were assessed. The
The auxiliary gas and injector temperatures were 250 C and most abundant VOCs in winter were alkanes, followed by ar-

230 C, respectively. The mass spectrometer was operated in omatics. Ethane, propane, and n-butane were the most

electron impact ionization mode with a source temperature of abundant alkane species, which contributed 25.4%, 22.5%, and
230 C under the SIM. 11.6%, respectively, to the total alkane concentration. Aro-

Quality control was achieved by calibrating the instrument matics comprised benzene, toluene, ethylbenzene, styrene,
before acquiring and analyzing samples during the observa- and three groups of isomers. In contrast, the most abun-
tion period to ensure the authenticity and accuracy of the dant VOCs in summer were OVOCs, followed by alkanes,
experimental data. The instrument blank should be analyzed halocarbons, aromatics, and alkenes. Acetaldehyde, acetone,
every 24 hr during the sampling period and the determination
results of the instrument blank should be lower than the
laboratory method detection limit. The quantification of the
C 2 eC 5 hydrocarbons was carried out by an external standard
method, and an internal standard method was used to
quantify the other VOCs. The recovery rates of the internal
standards (bromochloromethane, 1,4-difluorobenzene,
chlorobenzene-d5, and 4-bromofluorobenzene) were 80%e
120%. Two certified standards were analyzed, one being a
mixture of 55 non-methane hydrocarbons provided by Spectra
Gases (USA) and the other a mixture of OVOCs and halocar-
bons provided by Apel-Riemer Environmental (USA). The six-
point (0.5e20 ppbv) calibration curves were generated with a
linear correlation coefficient (above 0.999 for the non-
methane hydrocarbons and above 0.99 for the other species),
indicating that the peak areas were proportional to the target
compound concentrations. Calibrations were performed each
day the instrument was used, and the target species re-
sponses varied by ± 10% of the calibration responses. The
method detection limit definition described in the US Envi-
ronmental Protection Agency document TO-15 was used, and
the method detection limits were 0.008e0.034 ppbv for C 2 eC 5
hydrocarbons, 0.005e0.091 ppbv for other hydrocarbons,
0.005e0.029 ppbv for halohydrocarbons, and 0.011e0.030 ppbv
2
for OVOCs, and the R of all VOC species reached 0.99.

2. Results and discussion

2.1. Characteristic and seasonal variation of VOC
concentration

In the present study, 99 VOC species were measured at 10 am
in Beijing in winter from 15 December 2015 to 17 January 2016
and in summer from 21 July to 25 August 2016. The sampling
time period was divided into normal days (AQI 100) and Fig. 1 e Variations in the concentration of VOCs during (a)
polluted days (AQI > 100). The 99 VOCs were classified into winter (between 2015/12/15 and 2016/1/17) and (b) summer
seven categories: alkanes (29), alkenes (11), acetonitrile (1), (between 2016/7/21 and 2016/8/25).