Site description and sampling location
The study site was located in Shanghai Chong Ming Island, with a capacity of approximately 0.6−2 ton/day, and the service area comprising a community of 4000 resident was 100 m away from the facility. The facility comprised 20 composting units, a biofilter pool, and a sorting area, as shown in Fig. 1. The facility was opened except the composting unit and the board was not walls but railing. The composting unit was usually closed only when the waste feeding opened. Oxygen was supplied through the ventilation system in the bottom and the gas was collected for purification through the biofilter system. Chopped straw with a length of 2−5 cm was used as a bulking agent. After removing glass and metal debris, the raw MSW was mixed with the bulking agent at a ratio of 10:1 (w/w) considering that the water content should not be excessively low, and an adequate specific surface area and free air spaces were necessary in the matrix [16]. The characteristics of the raw materials and mixture are presented in Table 1. The volatile solids content was analyzed by heating samples to a constant weight at 550 °C in a muffle furnace. The oxygen concentration and temperature at the core of the windrow were monitored using a probe (CYS-1; Xuelian Co., Shanghai, China). The elemental compositions of the freeze-dried samples were measured using an elemental analyzer (Vario EL III, Elementar Analysensysteme GmbH, Langenselbold, Germany). Active fermentation (static composting-like process) occurred for 20 days in the closed composting unit with controlled aeration and gas collection, as well as treatment with a biofilter.
Table 1 Characteristics of the raw materials and mixture feedstock in the tested composting unit Odorous gas sampling was conducted during waste loading, unloading, and sorting. Three sampling points were located in the sorting area, and upwind and downwind of the boundary of the facility. Sampling was also conducted at the same points when the facility was not in operation in order to conduct comparisons. One point was located in the biofilter pool when the ventilation process started. Inside the composting units, gas and solid sampling were conducted after 1, 3, 5, 7, 10, 15, and 20 days during the active fermentation phase. The wind speed during the sampling period was 3.1–3.6 m/s and the temperature was 20−24 °C. The relative humidity was approximately 55−65%. In total, seven chemical classes and 51 substances were determined in the gas samples, including ammonia, reduced sulfur compounds (RSCs), aromatics, terpenes, alcohols, carbonyl compounds, and VFAs.
Analysis of gas samples
Several methods were used to test different target gaseous pollutants, as showing in Table 2. These methods included sorbent concentration, cold trap concentration, derivatization concentration, gas chromatography method, liquid chromatography and colorimetric method.
Table 2 Methods were used to measure different target gaseous pollutants Analysis of ammonia using colorimetric tubes
Ammonia in each sample was detected by colorimetric tubes (GV-100 s, GasTec., Kanagawa, Japan). Cut both ends of the colorimetric tube and insert the colorimetric tube into the hand pump in the direction of the arrow. A specific volume of 100 ml gas was extracted to the colorimetric tubes. Wait several minutes and confirm the completion of the sampling. The lengths of the color changes in the tubes indicated the concentrations of ammonia.
GC analysis after cold trap concentration
The air samples were collected by Tedlar bags (3L, SKC, PA, USA) which were placed in a sealed plastic box. There were two holes on the lid of the plastic box, one was connected with the Tedlar bag, and the other was connected with a vacuum pump. When the box was vacuumized by the pump, the air sample entered to the Tedlar bag through the hole connected with the bag. This will prevent the air pump from contaminating the air samples. The samples were placed for dark storage and transported to the laboratory within 12 h to minimize the loss of RSCs during storage.
According to the United States Environmental Protection Agency (USEPA) TO14A method [17], these air samples were concentrated by cold trap (Entech Instruments Inc., CA, USA) with liquid nitrogen, then detected by GC-FID and GC-PFPD (GC 450, Varian Inc., CA, USA). The injection volume was 100−1000 ml, according to the actual concentrations. The GC-PFPD and GC-FID parameters were described in detail in our previous study [17].
GC-FID analysis after sorbent concentration
Volatile fatty acids in air samples were determined by adsorbent concentration method. Air samples were collected using commercial adsorption tubes (silica gel tubes, SKC, PA, USA). The air flow rate was 2000 ml/min and the collection time was 120 min. After sampling, each tube was sealed and transported to the laboratory within 12 h. Take out the silica gel adsorbent, put it into a 5-ml volumetric flask and absorb it with 5 ml deionized water. After standing for 30 min in an ultrasonic apparatus, the supernatant was taken out and analyzed by GC-FID. The operation parameters of FID detection system were the same as our previous study [18].
HPLC analysis after derivation with DNPH
Commercial cartridges (Cleanert DNPH-Silica, Agela Technology, Tianjing, China) were used to collect carbonyl compounds in the air samples. Sampling flow rate was 1000 ml/min and the sampling time was 2 h according to EPA method TO11A [19]. After collection, the cartridges were sealed and transported to the laboratory within 12 h. The derived compounds in the cartridges were eluted into a 5-ml volumetric flask with 5 ml acetonitrile through a solid-phase extraction vacuum manifold (Visiprep, Supelco Analytical, Darmstadt, Germany).
A working standard calibration curve was prepared from serial dilutions of the aldehyde/ketone-DNPH standard stock solution (Cerilliant Inc., USA). The concentrations of the standard mix solutions varied from 0.075 to 15 ppm. The standard solutions and sample eluates were analyzed by HPLC (Agilent 1200, Agilent Inc., USA) attached to a diode-array detector (DAD) through an auto-sampler with a detector that was operated at 365 nm. The detection limit of this method was 50 ppb. The analytical column used was a C18 (4.6 mm ID × 25 cm, 5 μm) stainless steel tube (Venusil XBP, Agela Technology, China) and the mobile phase was acetonitrile (Merck, Germany) and high purity water (Milli-Q Millipore, USA). The elution program was 45% acetonitrile for 1 min, followed by a linear gradient from 45 to 75% acetonitrile in 30 min, which was then held for 5 min. The flow rate was 2 ml/min and the sample injection volume was 25 μl.
Quality assurance and control
Five levels of mixed standard gases (50, 200, 400, 800, 1600 μg/m3, Air Liquid, France) were determined to produce the standard calibration curves. Within the range of 0–1600 μg/m3, the correlation was good (R2 > 0.93). The detection limit of the instrument is determined by extrapolation of the linear ratio between the minimum peak area and the instrument noise. The concentration of the blank sample was less than 1 μg/m3, indicating that there was no sample contamination during collection, transportation and storage. Ten kinds of repetitive standard gases were determined to evaluate their reproducibility. The relative standard deviation is less than 7%. The standard recoveries of these methods ranged from 81 to 114%.
Theoretical odor concentrations
Odor threshold concentration of each compound varies greatly and gases with high concentrations do not always contribute strong odors. Odor intensity can be measured by the ratio of its concentration and the odor threshold. When the composition of the odorous gas mixture is known, the theoretical odor concentration (Cod) of each sample can be estimated based on its analytical concentration and odor threshold according to Eqs. (1) and (2) [20]. The odor intensity of different sample points can be compared by this method:
$$C_{{{\text{od}},i}} = \frac{{C_{i} }}{{{\text{OT}}_{i} }},$$
(1)
$$C_{{{\text{od}}}} = \sum\limits_{i = 1}^{n} {\frac{{C_{i} }}{{{\text{OT}}_{i} }}} ,$$
(2)
where Ci is the analytical concentration of the ith compound (ppm); OTi is the odor threshold value of the ith compound (ppm), where the OTi value for each compound was obtained from a previous study [17]; Cod,i is defined as the theoretical odor concentration (dimensionless); n is the total number of odorous compounds; and Cod is the sum of the theoretical odor concentrations of n compounds of one sample.
Health risk assessment
Carcinogenic risk assessment
According to the Integrated Risk Information System of the United States EPA, the carcinogenic risks of gaseous pollutants were evaluated according to the inhalation unit risk (IUR) of carcinogens [21]. The cancer risk assessment was measured based on the lifetime carcinogenic risk (CRi) using Eqs. (3) and (4). The carcinogenic risk value for a mixed source is the sum of that for each compound, according to Eq. (5). Synergistic and antagonistic effects among substances are not considered:
$${\text{CRi}} = {\text{ECi}} \times {\text{IUR,}}$$
(3)
$${\text{ECi}} = \frac{{{\text{Ci}} \times {\text{ET}} \times {\text{EF}} \times {\text{ED}}}}{{{\text{AT}} \times 365 \times 24}},$$
(4)
$${\text{CR}} = \sum\limits_{i = 1}^{n} {{\text{CRi}}} ,$$
(5)
where IUR is the inhalation unit risk, μg/m3; ECi is the exposure concentration of each gaseous pollutant in the air, μg/m3; Ci is the concentration of each gaseous pollutant in the air, μg/m3; ET is the daily exposure time of a worker, 8 h; EF is the exposure duration, 250 days/year; ED is the exposure duration, 30 years; and AT is the average age, 70 years. The parameters were selected based on a previous study [21]. There were six workers in total. All the workers were considered as similar exposure.
Non-carcinogenic risk assessment
For non-carcinogenic risk assessment, inhalation reference concentration (RfC) was used to calculate the hazard quotient of each pollutant (HQi) by Eq. (6). The non-carcinogenic risk value for a mixed source is the sum of that for each compound, according to Eq. (7). Synergistic and antagonistic effects among substances are not considered:
$${\text{HQi}} = \frac{{{\text{ECi}}}}{{{\text{RfC}} \times 1000\,\upmu {\text{g/mg}}}},$$
(6)
$${\text{HQ}} = \sum\limits_{i = 1}^{n} {{\text{HQi}}} ,$$
(7)
where RfC is the inhalation reference concentration, mg/m3; calculation method of ECi is the same as Eq. (4), which all values are the same except AT is the average age, 30 years. The parameters were selected based on a previous study [22].