Chemical Characterization of Water-soluble Ions and Metals in Particulate Matter Generated by a Portable Two-stroke Gasoline Engine

To examine the characteristics of water-soluble ions and metals on the particulate matter (PM) in the exhausts, a P2SGE (portable two-stroke gasoline engine) was fueled by unleaded gasoline #92 blended with different two-stroke engine oil brands (CPC Super Low Smoke Two-Stork Engine Oil (SLS), CPC Low Smoke Two-Stroke Engine Oil (LS), and MERCURY STAR (MS)) and operated under idling, mid-load (1.5 kW), and high-load (1.9 kW), respectively. Experimental results reveal that the PM mass concentrations in the exhausts were in the order MS (avg. 1,934 mg Nm) > SLS (avg. 1,543 mg Nm) > LS (avg. 1,167 mg Nm) in all test conditions. The mass concentrations and emission factors (EFs) of PM decreased as the P2SGE load increased by adding each tested lubricant. Based on fuel consumption, EFs of ΣIons were the lowest when utilizing the LS additive (avg. 89.7 mg L-fuel), followed by the MS and SLS (165 and 168 mg L-fuel, in average, respectively); whereas the lowest levels of ΣMetals were observed by using MS additive (avg. 61.3 mg L), followed by using the LS (avg. 83.8 mg L) and SLS (avg. 85.2 mg L). The soluble ions on the PM were mostly Na, Ca, NO3, and SO4 among eight tested species, which accounted for only 0.05–0.19% (avg. 0.1%) of PM mass. The 21 analyzed metal components represented only 0.05% of the mass of the PM, and were dominated by Na, Mg, Al, K, Ca, Fe, and Zn, which represented 98.7% by mass of ΣMetals. Our finding for portable engine emission has been rarely considered in the literature but it is unneglectable for labors who are usually exposed to the ions and metals. Further health risk assessment research is suggested to include temporarily real-life exposures with high pollutant levels.


INTRODUCTION
The internal combustion engine (ICE) is one of the greatest inventions in history. It has a high thermal efficiency and high power output and enables vehicles to travel at high speed (with good maneuverability). Since the mid-19th century, ICEs have been used around the world. Almost all motor vehicles, agricultural machinery, engineering machinery, and even power generation equipment, use internal combustion engines to generate power. The working principle of the ICE is that burning fossil fuels (including gasoline, diesel, kerosene, and natural gas) inside the machine releases chemical energy, which is converted to mechanical kinetic energy. This process can provide huge amounts of energy for various activities, but it generates many harmful air pollutants (such as fine particulate matter (PM 2.5 ), traditional pollutants, organic/inorganic component species,

brands: CPC Super Low Smoke Two-Stroke Engine Oil (CPC-SLS), CPC Low Smoke Two-Stroke Engine
Oil (CPC-LS), and MERCURY STAR Two-Stroke Engine Oil (MS). The three lubricants used in this study were the lubricating oils that are commonly available in Taiwan and are relatively inexpensive. The characteristics of water-soluble ions and metal components on PM were measured at various engine loads.

Instruments and Sampling Methods
The tested small portable gasoline-powered sprayer (KAWAGOE, Model: C12E) was equipped with a single-cylinder, naturally aspirated, and air-cooled two-stroke gasoline engine (Model: G45H; manufactured by Chaang Cherng Co., Ltd., TAIWAN). The fuel oil mixing ratio (gasoline: two-stroke engine oil), set by its manufacturer, was 25: 1; the cylinder displacement was 41.5 c.c.; and the maximum horsepower was 2.2 ps.
The base oil, unleaded gasoline #92, was purchased from the CPC (Chinese Petroleum Corporation) in Taiwan. Table 1 presents the physicochemical properties of two-stroke engine oils (SLS, LS, and MS) used in this study.
The flue sampling equipment that was used in this study was consistent with US EPA Method 5. Before sampling was begun, an S type Pitot tube was used to measure the flow velocity of the exhaust gas from the engine. After calculating the flow rate, a constant-flow sampling system that was equipped with a 47 mm dia. quartz fiber filter (Pallflex Tissuqartz 2500QAT-UP) was installed downstream of the P2SGE's exhaust to gather the particle-phase samples. PM emissions were tested at P2SGE rotation speeds of 3,800 rpm (idling), 5,000 rpm, and 7,000 rpm with three fuels. For each combination of parameters, the experiment was performed three times (each sampling time = 15 min). Sampling data were collected after the engine had been run for at least 10 min. Fig. 1 shows the schematic diagram of the experimental system.
The quartz filters were pretreated before sampling by heating them in a muffle furnace in air for 2.5 h at 900°C. The filters were dried for 24 h in a desiccator at 25 ± 1°C and a relative humidity of 40 ± 5% before and after each sampling. They were then weighed on an electronic seven-digit balance (UMX2, Mettler Toledo) with a resolution of 0.1 µg.

Water-soluble Ion Analysis
Before water-soluble ions were analyzed, collected particles were extracted for 120 minutes from quarter sections of each quartz filter in an ultrasonic bath using 10 mL of n-hexane. Then, 10 mL of deionized water (specific resistance ≥ 18.3 MΩm) was added and ultrasonic extraction was conducted for another 120 minutes. Next, the n-hexane was removed by purging with ultrapure nitrogen. Finally, each extraction solution was filtered using a C18 pretreatment column and a cellulose acetate filter in that order, and then stored in a plastic vial in a refrigerator at 4°C before chemical analysis by ion chromatography (IC) (DIONEX ICS-3000).  The method detection limits were as follows; Na + , 0.038 ppm; K + , 0.041 ppm; NH 4 + , 0.076 ppm; Mg 2+ , 0.027 ppm; Ca 2+ , 0.077 ppm; Cl − , 0.046 ppm; NO 3 -, 0.060 ppm, and, SO 4 2-, 0.021 ppm. The recovery efficiencies of these ions were 92.3-114.3% based on the IC measurements. Both field and laboratory blank samples were prepared and analyzed for each sampling and analysis. All data were corrected using filter blanks.

Metal Analysis
Before the particle-bound metals had been chemically analyzed, extraction was performed on one-quarter of each quartz filter for 120 minutes using an ultrasonic bath of 20 mL 10% (v/v) HNO 3 solution. Then, each sample was heated to 85°C for 30 minutes for digestion. The digested solution was diluted to a volume of 25 mL using ultra-pure water (specific resistance ≥ 18.3 Mcm) to identify 21 metals (Na, Mg, Al, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, As, Sr, Mo, Cd, Sn, Sb, Ba, and Pb) by inductively coupled plasma-mass spectrometry (ICP-MS) (Agilent, 7500 series). Calibration was conducted using multi-element (metal) standards (certified reference materials (CRMs); Spex, Metuchen, USA) in a 1% (v/v) HNO 3 solution. Every tenth sample was spiked using the liquid standards that contained known amounts of the metal elements that were analyzed. The CRMs were also used as quality control standards.
In the analyses of elements from ICP-MS measurements, the method detection limits for Na, Mg, Al, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, As, Sr, Mo, Cd, Sn, Sb, Ba, and Pb were 6. 12, 5.22, 3.16, 19.3, 24.3, 0.36, 0.04, 0.14, 0.03, 2.22, 0.03, 0.31, 5.27, 0.31, 0.20, 0.99, 0.02, 3.15, 0.12, 0.51, and 0.06 ppb, respectively. The recovery efficiencies of 21 metals were 91.2-108.4%. Both field and laboratory blank samples were prepared and analyzed for each sampling and analysis. All data were corrected using filter blanks. and output energy (EF OE )) of PM in the exhausts emitted from P2SGE that was operated at idling (0 kW), mid-load (~1.5 kW), and high-load (~1.9 kW) with 4 vol.% of SLS, LS, or MS as the additive in 92-gasoline. Experimental results show that the mass concentration of the PM from P2SGE was the lowest when LS was used (947-1,331 mg Nm -3 , with an average of 1,167 mg Nm -3 ), followed by SLS (1,020-1,913 mg Nm -3 , with an average of 1,543 mg Nm -3 ) and MS (1,289-2,405 mg Nm -3 , with an average of 1,934 mg Nm -3 ) at all engine loads. Accordingly, the PM concentration decreased as the P2SGE load increased regardless of the lubricant used (as shown in Fig. 2(a)). Moreover, this trend was also observed for EF FC and EF OE (as shown in Figs. 2(b) and 2(c)). This phenomenon may be related to the sampling temperature was higher at the high-load (85-90°C) than at idling (40-45°C) and mid-load (50-55°C), and thus the P2SGE had a better combustion efficiency when operated at high-load and helped to inhibit PM formation.

PM Concentrations and Emission Factors in the Exhaust
Adding SLS or LS reduced the PM concentration below that obtained when MS was added at a given engine load; adding LS reduced the PM concentration (38.2%) by almost twice as much as adding SLS (20.1%), perhaps because higher flash point of MS (136°C) (as shown in Table 1) leads to less complete combustion and higher PM emission than with the other two lubricant oils, which have lower flash points (LS = 98°C and SLS = 94°C), at the same engine load (Tsai et al., 2020). Generally, the flash point of oil is the lowest temperature of oil at which the application of defined test flames leads the vapors above the surface to ignition and the release of vapors at this temperature is not sufficiently rapid to sustain combustion (Ljubas et al., 2010). Once the flash point of two-stroke engine oil was getting higher, it was less likely to be burned entirely (in order to protect the engine cylinder) and remained in the exhausts during the combustion process, resulting in more significant amounts of emission (as displayed in Fig. 3). On the other hand, although the flash points of SLS and LS were similar, the PM emissions of using these two fuels differed by about 30%, which was possibly associated with their different viscosities. As is well known, the higher the viscosity of the oil, the less favorable it is to be atomized, eventually leading to more PM generation during the combustion process.
Our analysis demonstrated that the A/C (summation of equivalent concentrations of anions to that of cations) ratio increased with the engine load but were all less than 0.5 as presented in Table 2. Nevertheless, the A/C ratios that were calculated from the data in several previous gasoline exhausts studies were all less than 0.2 (Table 3), supporting our finding. This phenomenon may be related to the fact that lubricant oil usually contains a certain amount of CaCO 3 which      could maintain the lubrication of a piston during engine operation (Eastwood, 2008;Raza et al., 2018). The CaCO 3 is a typical alkaline additive for preventing the deterioration of lubricant oil by oxidation and extending operational lifetime. Moreover, to prevent wear and tear on the twostroke engine, inevitably some unburned lubricant oil-droplets might be present in the exhausts (as shown in Fig. 3). At this situation, the unburned lubricant oil-droplets would be collected on the PM samples that were obtained by the filtering method. Furthermore, a mixture of deionized water and n-hexane was used herein to extract the water-soluble ions on PM, which might also cause the extraction of CaCO 3 in the unburned oil-droplets. Thus, the Ca 2+ content was increased and became one of the major ions on the PM, whereas CO 3 2was not analyzed in this study, leading to lower A/C ratios. Table 4 and Fig. 6 illustrate the mass concentrations and EFs of 21 PM-bound metals (Na, Mg, Al, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, As, Sr, Mo, Cd, Sn, Sb, Ba, and Pb) from the exhaust of the P2SGE with 4 vol.% of LS, MS, and SLS lubricants in 92-gasoline. The results showed that the three metallic elements of As, Mo, and Sn were not detected in all test conditions. On average, the mass concentrations of ΣMetals on PM from the P2SGE's exhausts by using 4 vol.% of SLS, LS, or MS as the additive in 92-gasoline were 677-923 (average 824 µg Nm -3 ), 690-800 (average 810 µg Nm -3 ), and 576-687 µg Nm -3 (average 603 µg Nm -3 ), respectively. In terms of emission factors, the lowest levels of ΣMetals in the exhausts were observed by using 4% MS additive (average 61.3 mg L -1 ), followed by LS (average 83.8 mg L -1 ) and SLS (average 85.2 mg L -1 ). According to the mass concentration or emission factor data, the amount of each metallic element from the P2SGE almost decreased with increasing loading no matter what kinds of lubricants were used.

Concentrations and Emission Factors of PM-bound Metals
The metallic components were divided (based on their average levels in ΣMetals) into three groups -major metals (Na, Mg, Al, K, Ca, Fe, and Zn, ≥ ~1%), sub-major metals (Cr, Mn, Ni, Cu, Sr, Ba, and Pb, ~1 to 0.1%), and minor metals (Ti, V, As, Mo, Cd, Sn, and Sb, < 0.1%). According to Fig. 7, the major metals represented 96.5-99.4% by mass of ΣMetals, while the sub-major and minor metals accounted for only 0.47-3.44% and 0.04-0.14%, respectively. Previous studies have identified the major PM-bound metals in gasoline engine exhaust as Ca, Fe, Al, and Zn (Cheung et al., 2010;Hao et al., 2019); Yang et al. (2019) identified the major metal components on PM 2.5 as Na, Ca, Fe, Zn, and Al. These observations were consistent with those herein, supporting our finding of the main metal components on PM from the P2SGE.
In addition, the most abundant metals by mass were Ca, Na, K, and Al, which accounted for 93.8% by mass of ΣMetals when 4-vol% SLS was added to the fuel in the P2SGE at various loads, as reported in Table 4. The dominant four metals were Ca, Mg, Na, and K (79.9% by mass of ΣMetals) and Ca, Na, K, and Al (85.3% by mass of ΣMetals) when LS and MS were used, respectively. Consequently, the dominant metals among the 21 analyzed metals in PM were Ca, Na, K, Al, Fe, Mg, and Zn. Those dominant metals together accounted for 98.7% by mass of ΣMetals. At first glance, Fe and Zn (only about 4% each) had relatively low contents among seven primary metals. However, previous researches have confirmed that extended exposure to Fe may cause the generation of free radicals in cells and result in respiratory diseases (Kadiiska et al., 1997), whereas prolonged exposure to Zn leads to acute heart diseases (Chen and Lippmann, 2009). Therefore, it is necessary to regulate the limits of the hazardous metal contents in P2SGE exhausts to reduce the risk of inhaling such harmful ingredients for P2SGE operators (such as farmers).

CONCLUSIONS
Portable two-stroke gasoline engines (P2SGEs), commonly fueled by lubricant oil-gasoline mixtures, may emit more pollutants than regular gasoline engines in on-road vehicles because they are not equipped with exhaust gas control devices and typically lack maintenance. This study explored the concentrations of chemical components and metallic elements in PM emissions from the exhaust of a P2SGE. The results showed that the PM mass concentrations in the exhausts were in the order MS (average 1,934 mg Nm -3 ) > SLS (average 1,543 mg Nm -3 ) > LS (average 1,167 mg Nm -3 ) in all test conditions. The mass concentrations and EFs of PM using #92-gasoline with 4% engine oil (either SLS or LS or MS has been added) decreased as the engine load increased. Adding SLS or LS to gasoline reduced the PM emission levels by 38.2% or 20.1%, respectively, below that achieved by adding MS. Based on fuel consumption, emission factors of ΣIons were the lowest when utilizing the LS additive (average 89.7 mg L-fuel -1 ), followed by the MS and SLS (165 and 168 mg L-fuel -1 , in average, respectively). The dominant water-soluble ion species on the PM were Ca 2+ (20.5%), Na + (31.4%), and SO 4 2-(13.6%). The ratios of the sum of Volume 21 | Issue 3 | 200632   equivalent concentrations of anions to that of cations (A/C) increased with the increase of engine load, although all the A/C ratios were lower than 0.5. In terms of emission factors, the lowest levels of ΣMetals in the exhausts were observed by using MS additive (average 61.3 mg L -1 ), followed by LS (average 83.8 mg L -1 ) and SLS (average 85.2 mg L -1 ). Despite the difference in blended fuels, the dominant metallic elements on PM were Ca, Na, K, Al, Fe, Mg, and Zn (over 90% from ΣMetals) at all tested conditions. The subdominant metals (Cr, Mn, Ni, Cu, Sr, Ba, and Pb) accounted for 0.5-3.5% of the PM by mass, while the trace metals (Ti, V, As, Mo, Cd, Sn, and Sb) accounted for only 0.1% PM mass.