Effects of Cerium Oxide and Ferrocene ... - ACS Publications

Mar 27, 2017 - vehicles to reduce soot and DPM emissions.6−8 However, these. FBCs seem ... stroke, direct-injection diesel engine (L70AE, Yanmar Cor...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/est

Effects of Cerium Oxide and Ferrocene Nanoparticles Addition As Fuel-Borne Catalysts on Diesel Engine Particulate Emissions: Environmental and Health Implications Zhi-Hui Zhang and Rajasekhar Balasubramanian* Department of Civil and Environmental Engineering, Faculty of Engineering, National University of Singapore, 1 Engineering Drive 2, E1A 02-19, Singapore 117576, Singapore S Supporting Information *

ABSTRACT: This study systematically examined the potential impacts of doping CeO2 and Fe(C5H5)2 nanoparticles as fuel-borne catalysts (FBCs) to ultralow sulfur diesel (ULSD) fuel on the physical, chemical and toxicological characteristics of diesel particulate matter (DPM). The FBCs-doped fuels are effective in promoting soot oxidation and reducing the DPM mass emissions, but lead to a significant increase in the total particle counts due to the formation of self-nucleated metallic nanoparticles. Compared to undoped ULSD, the FBCs-doped fuels result in higher concentrations of particle-phase polycyclic aromatic hydrocarbons (PAHs) and n-alkanes, higher fractions of organic carbon (OC) and water-soluble organic carbon (WSOC) in particles, show slight alterations in soot nanostructure, reduce soot ignition temperature and activation energy. Exposure of the human-type II cell alveolar epithelial cells (A549) to DPM derived from FBCs-doped fuels shows a decrease in cell viability and alterations in the global gene expression with a broad range of biochemical pathways. The overall variations in DPM characteristics are mainly caused by the catalytic combustion process, and are related to the type, properties and contents of FBCs used in diesel fuel as well as the engine operating conditions. Environmental and health implications of the study are highlighted.

■

INTRODUCTION High emissions of airborne particulate matter (PM) from diesel engines used in both on-road and off-road applications are of global concern as they have adverse impacts on urban air quality, human health, and also affect global climate change.1,2 Improvements in engine technology and the application of post-combustion treatment systems in conjunction with diesel fuel reformulations are critically needed for abatement of diesel particulate matter (DPM) emissions.3 One of the fuel reformulation technologies under consideration is the addition of an ultralow dose of fuel-borne catalysts (FBCs) in the form of organometallic nanoparticles to diesel fuels. This approach is explored for suppressing soot formation and/or promoting soot oxidation during the combustion process, and thus reducing DPM mass emissions.3−8 An additional consideration is that if FBCs are used in combination with diesel particulate filters (DPFs), the FBCs embedded in the emitted DPM could enhance soot oxidation in the DPFs and therefore promote DPFs regeneration.9 FBCs containing Ce, Fe, and Pt are now commercially available, among which both Ce- and Fe-based FBCs have been increasingly used in Europe and elsewhere as diesel fuel additives to ensure fast and complete DPFs regeneration.7−11 These FBCs have also been used without DPFs in existing off-road diesel engines and commercial vehicles to reduce soot and DPM emissions.6−8 However, these FBCs seem to be restricted from using in several countries. For © 2017 American Chemical Society

example, in the U.S., current regulations restrict their use to onroad diesel engines. In Switzerland, FBCs can only be used in combination with DPFs, but their use without particle traps is generally prohibited. These restrictions have been placed because of concerns over the potential environmental and public health impacts of FBCs. DPM emissions vary significantly in physical characteristics and chemical composition between different engine types, engine operating conditions, and fuel formulations. Some recent studies have reported that the catalytic combustion process, caused by FBCs-doped fuels, not only reduce DPM mass concentrations efficiently, but also trend to alter the physical and chemical properties of DPM.7,8,12−14 Emission reduction strategies that take the chemical composition of DPM into account are deemed to be more effective and efficient because the chemical characteristics of DPM rather than the DPM mass determine their potential influence on the environment and human health. In addition, the application of metal-based FBCs in diesel fuels inevitably causes the emissions of the metals themselves in DPM which, combined with changes in chemical composition and/or particle size distributions, might lead to changes in the overall toxicity of Received: February 19, 2017 Accepted: March 27, 2017 Published: March 27, 2017 4248

DOI: 10.1021/acs.est.7b00920 Environ. Sci. Technol. 2017, 51, 4248−4258

Article

Environmental Science & Technology DPM.15−17 However, the potential adverse environmental effects and health risks resulting the use of Ce-based and Febased FBCs are poorly understood due the lack of systematic investigations. To address these knowledge gaps, we systematically investigated the changes in the physicochemical and toxicological characteristics of DPM emissions resulting from the diesel engine fuelled by CeO2- and Fe(C5H5)2-doped fuels by using a series of complementary analytical techniques. The central goal of this work is to assess their potential environmental and human health implications. Specifically, we examined the influence of doping different concentrations of these additives to ultralow sulfur diesel (ULSD) fuel on particulate mass concentration as well as particle number concentration with their corresponding size distributions. We then analyzed the effects of these FBCs-doped fuels on the carbonaceous contents of the DPM including elemental carbon (EC), organic carbon (OC), and water-soluble organic carbon (WSOC). We selected PAHs and n-alkanes as appropriate chemical species of OC for further discussion as PAHs are suspected human carcinogens and n-alkanes represent an important class of the organic compounds originating from unburned fuel and/or lubricating oil. In addition, we studied the particle volatility and soot oxidation behavior to understand the effects of these FBCs-doped fuels on nanoparticle formation and soot oxidation reactivity. Determination of the influence of FBCs-doped fuels on the above-mentioned properties of DPM is essential for understanding how FBCsdoped fuels affect the DPM formation process and its chemical composition. The changes in the physicochemical characteristics are likely to affect the toxicity of DPM. In order to examine this possibility, we exposed the particles derived from both FBCs-doped and undoped fuels to the human-type II cell alveolar epithelial cell line (A549) and assessed their toxicity by using the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay. The cDNA microarray technique was then used to assess the changes in the expression of gens A549 in an attempt to gain an insight into the complex mechanisms that underlie the health effects induced by DPM derived from FBCs-doped fuels in comparison with undoped diesel fuel. We selected an off-road diesel engine for this study as such engines are widely used and emit a substantial fraction of DPM on a global level due to the lack of effective environmental regulations.18 A long-term exposure to high concentration of freshly emitted DPM has been reported to pose a significant carcinogenic risk to occupational workers in various working microenvironments where off-road engines are frequently used.19

previously for investigating the influence of FBCs-doped fuels on engine exhaust emissions.4−8,13,14 ULSD with less than 10 ppm (parts per million) by weight of sulfur was used as a base fuel, and the major properties of this fuel can be found elsewhere.20 The FBCs used in this study include cerium oxide (CeO2; Sigma-Aldrich, ≥ 99.9%) nanopowder with particle size of less than 25 nm, and the analytical grade ferrocene (Fe(C5H5)2; Sigma-Aldrich, ≥ 98%) in the form of commercially available nanoparticles. CeO2 and Fe(C5H5)2 were added to ULSD in three different concentrations (25, 50, and 100 ppm of Ce and Fe by weight, which are designated as Ce25, Ce50 and Ce100 for the CeO2-doped fuels, and Fe25, Fe50, and Fe100 for the Fe(C5H5)2-doped fuels, respectively.). Both CeO2 and Fe(C5H5)2 were added to the diesel fuel as per the established procedure,8,14,21 which can be found in the SI. Particulate Sampling and Measurement. A two-stage Dekati mini-diluter (DI-2000, Dekati Ltd.) was used for diluting the engine exhausts for DPM sampling and online evaluation. The dilution ratio from each stage was determined, and all data presented in this article have been dilutioncorrected to represent tailpipe conditions. DPM emissions from the first-stage diluter were collected onto 47 mm Teflon filters (Pall Life Sciences, Ann Arbor, MI, 2 μm pore size) and on 47 mm quartz fiber filters (Whatman, 2.2 μm pore size) for the subsequent analyses, by concurrently using two Mini-Vol particulate samplers (Air metrics Ltd.; 5 L/min flow rate). The blank quartz fiber filters were precombusted in air (650 °C for 12 h) to remove any residual carbon contamination prior to being used for particles collection. Further details about determination of the dilution ratio and the handling of filters and DPM samples are provided in the SI. The number concentrations and size distributions of volatile and solid particles in the secondary dilution stage were measured by a Fast Mobility Particle Sizer (FMPS, model 3091, TSI Incorporated) for particles over the size range of 5.6−560 nm. In this setup, two diluters were used in series, with the first stage being heated to maintain a temperature of 190 °C to minimize thermophoretic deposition. During the solid particle number emission experiments, a thermodenuder (TD, Dekati Ltd.) with a heated temperature of 265 °C was placed in-line between the secondary stage diluter and the FMPS for removing the volatile and semivolatile compounds of particles. The TD diffusion losses were estimated using a widely used method,22 and the diffusion loss-corrected particle size distributions are presented in this article. Physical Analyses. A thermogravimetric analyzer (Discovery TGA, TA Instruments) was used to investigate the ignition temperature of soot followed by the method reported previously,22 and the activation energy for the soot was then estimated using a modified form of the Arrhenius expression.23 Raman spectra of the DPM samples were recorded with a Renishaw microscope system (Renishaw, System 2000) using a 514 nm Ar ion laser as an excitation source. The spectra were then analyzed by using the Renishaw WIRE 2.0 software running under GRAMS/32 (Galactic, Levenberg−Marquardt nonlinear least-squares fitting algorithm) to obtain the D (Defect)/G (Graphite) band intensity ratios (ID/IG) for investigation of the graphite-like structure of soot particles. Further details of these analyses are provided in the SI. Chemical Analyses. A thermal/optical carbon aerosol analyzer (Sunset Laboratories, Forest Grove, OR) was used to quantify EC and OC fractions of particles collected onto quartz

■

EXPERIMENTAL SECTION The essential details of the experimental work are explained briefly here, and more detailed information is given in Supporting Information (SI). Engine, Fuels, and FBCs. The schematic of the experimental system is given in SI Figure S1. Experiments were carried out on a single-cylinder, naturally aspirated, fourstroke, direct-injection diesel engine (L70AE, Yanmar Corporation) connected to a 4.5 kW generator. The diesel engine has a capacity of 296 cm3 with bore and stroke of 78 mm and 62 mm, a fixed speed of 3000 rpm (revolutions per min). The more detailed specifications of the engine are provided in SI Table S1. Similar small off-road diesel engines have been used 4249

DOI: 10.1021/acs.est.7b00920 Environ. Sci. Technol. 2017, 51, 4248−4258

Article

Environmental Science & Technology

Figure 1. Effect of FBCs-doped fuels on (a) the concentrations of diesel particulate matter (DPM) mass and carbonaceous species, (b) EC (elemental carbon), (c) OC (organic carbon), and (e) WSOC (water-soluble organic carbon)), and the ratios of (d) OC/EC and (f) WSOC/OC. a Significantly different from the ULSD (p < 0.05). b Significantly different from between Ce-doped fuel and Fe-doped fuel under the same doping dose (p < 0.05).

filters according to the NIOSH 5040 reference method.24 Meanwhile, the particles collected on Teflon filters were extracted with organic-free Milli-Q water (>18 MΩ cm) using a water-bath sonicator (Elmasonic S 60H, ELMA, Germany) for 2 × 30 min. The extracts were then filtered through a 0.20 μm PTFE membrane filter (Satorius, Germany), acidified by ultrapure HCl (0.1 mL of 2 M HCl), and purged for 2 min with ultrapure air. 100 μL of the acidified filtrate was injected into an organic carbon analyzer (TOC-VCSH, Shimadzu) for determination of WSOC.

A Unity/Ultra thermal desorption system coupled with Agilent 6890N chromatography/5973N mass spectrometry (TD-GC/MS) was used for the determination of particle-phase PAHs and n-alkanes. The details of the method are provided by Ho and Yu.25 In this study, the 16 U.S. priority PAHs were measured and then classified into three groups based on their molecular weights, namely low molecular weight (LMW, 2 and 3 rings) PAHs, medium molecular weight (MMW, 4 ring) PAHs and high molecular weight (HMW, 5 and 6 rings) PAHs, as suggested in our previous study.20 Therefore, the total-PAH data represents the sum of the 16 individual PAHs. The overall 4250

DOI: 10.1021/acs.est.7b00920 Environ. Sci. Technol. 2017, 51, 4248−4258

Article

Environmental Science & Technology

standard mixtures (Sigma-Aldrich, Bellefonte, PA) was established, and the correlation coefficients (R2) for linear regressions of the calibration curves was greater than 0.99. In this study, all analytical data were corrected for the average value of the blanks, and are reported as the mean values. The significance of the differences of all tested parameters between diesel fuel and FBCs-doped fuels, and between the two kinds of FBCs-doped fuels under the same FBC concentration was assessed by Student’s t test at p < 0.05. The results of particle number size distributions are reported as the mean values, with the relative standard deviations ranging from 2.5 to 3.2%, respectively. With the exception of particle-phase n-alkanes and particle number size distributions, all other reported values are presented as mean ± SD (standard deviations).

toxicity is defined as the sum of Benzo[a]pyrene equivalent (total BaPeq), which was estimated based on the method of Nisbet and LaGoy.26 For particle-phase n-alkanes, concentrations of the homologous series from C14 to C26 were quantified. The concentrations of n-alkanes with molecular weight lower than C14 and higher than C26 were quite low and therefore are not reported in this article. Toxicological Experiments. In this study, DPM emitted from pure ULSD and from 100 ppm of Ce- and 100 ppm of Fedoped fuels at medium engine load was chosen for a comparative evaluation of their toxicological characteristics. We have made this selection because particles emitted under these conditions contained the highest amount of Ce and Fe when the engine was operated with the most commonly used engine load. These particles were physically removed from the Teflon filters by sonication, and resuspended in DMSO (Dimethyl sulfoxide, Bioreagents, Sigma-Aldrich) and subsequently in cell exposure medium using previously validated procedures,27 as summarized in the SI. The extraction efficiencies were evaluated, and the extracted particulate mass was efficiency-corrected. We exposed the particles to the human-type II cell alveolar epithelial cell line (A549). The cytotoxicity of these particles was measured by using the MTT assay and the resulting changes in the expression of global genes were carefully assessed by using the cDNA microarray technique. Method details are given in the SI. For the MTT assay, the cells were exposed to concentrations of particles in the cell culture wells that corresponded to 25, 50, 100, and 200 μg/mL of particles, respectively, for 24 h. Each sample was tested in six replicate wells, and the data on cell viability are reported as relative decrease compared to the control, considered as 100% of viable cells. As for gene expression, the A549 cells were exposed to 50 μg/mL of particles for 6 h. The gene expression level was considered significantly altered based on a 1.5-fold cutoff for the fold-change. The shortlisting of genes obtained from microarray data was conducted using the Gene Ontology (GO) analysis. The identified gene list has undergone functional annotation clustering using the DAVID web resource (http:// david.abcc.ncifcrf.gov/). The examined categories are Biological Processes and KEGG. QA/QC and Statistical Analysis. The analytical quality of the data obtained on the chemical composition of DPM was ascertained using the lower limit of detection (LOD), recovery, the linearity of calibration, and the relative standard deviation (RSD) of replicate analyses. The details of quality assurance and control (QA/QC) protocols for chemical analyses are provided in the SI. The LODs for OC and EC were below 0.47 mg/m3. The RSD of duplicate analyses was lower than 5.8% for TC (total carbon), and below 10% for OC and EC. The LOD for WSOC was below 0.6 μg/m3. After the correction for the blank contribution, the recoveries for WSOC were 82.9−94.8% by the TOC method, and the RSD of four replicate analyses was estimated to be within 6.5%. A five-point calibration for WSOC standard solution was performed, and the correlation coefficients (R2) obtained from the linear regression of the calibration curves was larger than 0.99. For PAHs and nalkanes, the LODs were calculated to be in the range of 2.1− 35.8 ng/m3 for PAHs and 2.9−29.2 ng/m3 for n-alkanes, respectively. The RSDs of four replicate analyses were 4.5− 10.3% for the PAHs, and were 4.8−7.5% for n-alkanes, respectively. An eight-point calibration over a concentration range of 0.1−5.0 ng for each PAH and n-alkane from the liquid

■

RESULTS AND DISCUSSION Particulate Mass and Carbonaceous Fractions. Figure 1(a) shows that both Ce- and Fe-doped fuels could effectively reduce the particulate mass concentrations, which is associated with the variation of the EC and OC emissions, as evident from Figures 1(b) and (c), respectively. Specifically, the EC emissions generally decreased with an increase in both Ce and Fe concentrations in ULSD. The degree of the EC reduction at lower engine load was greater than that at higher engine load. In addition, the Fe100 was generally more effective in inhibiting EC emissions than the Ce100 at a higher engine load. Similar results in the soot or smoke emission reduction were reported for Ce-doped fuels,4,5 and for Fe-doped fuels.7,8 Interestingly, we found that with the exception of Ce100, all other FBCs-doped fuels did not make significant changes in OC emissions. Thus, our finding suggests that the decrease in DPM emissions, caused by FBCs-doped fuels, can be attributed mainly to the EC emission reduction. Moreover, the substantial reduction in EC emissions coupled with the relatively minor change in OC emissions for both Ce- and Fe-doped fuels contributes to the higher OC/EC ratios (Figure 1(d)) than that for undoped ULSD. Compared to Ce100, the higher EC emission reduction coupled with the lower OC emission reduction for Fe100 led to higher OC/EC ratios. These observations suggest that a high proportion of semivolatile organics deposited on the surface of soot particles emitted from these FBCs-doped fuels, especially from Fe-doped fuels. Another interesting finding of our study is that the Ce50 and Ce100 led to a 3.7−8.1% higher particulate mass emission reduction than that of Fe50 and Fe100, indicating Ce-based FBC being more effective in inhibiting DPM emissions than that of Fe-based FBC under these conditions. Figure 1(e) shows that WSOC mass concentrations increased significantly for both Ce- and Fe-doped fuels at the high engine load, leading to the higher WSOC/OC (Figure 1(f)) ratios for these conditions. However, at low and medium engine loads, the higher WSOC/OC ratios only occurred for the addition of 100 ppm of Ce to ULSD. In other cases, the WSOC/OC ratios for FBCs-doped fuels showed no significant difference compared to that for undoped ULSD. In addition, for each test condition, the WSOC/CO ratios for Ce-doped fuels showed slightly higher than that of Fe-doped fuel, but were not statistically significant. Our results indicate that the FBCs-doped fuels tend to increase the water-soluble fraction of OC in these cases, suggesting an increased fraction of polar organic compounds in OC in these cases. These observations imply that the FBCs-doped fuels have potential to change in the particle toxicity due to the increased emission of both OC 4251

DOI: 10.1021/acs.est.7b00920 Environ. Sci. Technol. 2017, 51, 4248−4258

Article

Environmental Science & Technology

and Fe contents in the fuels (Figure 2(a)). The percentage increase in both MMW-PAHs and HMW-PAHs was higher

and WSOC fractions in DPM although they showed an overall decrease in total particle mass concentrations. A fundamental understanding of the formation of new particles and their growth in diesel engines may explain how FBCs-doped fuels affect the carbonaceous particulate emissions. The details can be found in the SI. Briefly, with the addition of FBCs to diesel fuel, the organic component of ferrocene decomposes in the early stage of the combustion process before the onset of soot particle inception due to its lower melting point,28 and is most dominantly present in the form of amorphous Fe (III) oxide crystallizing to hematite α-Fe2O3.29 Then, the iron oxide nuclei and nanoparticles are formed for the subsequent deposition of the carbonaceous phase. As for CeO2, its nanoparticulate structure would remain unchanged due to its very high melting point (2200−2400 °C). During the combustion process, these metal oxide catalysts have a high potential to promote carbon oxidation due to the redox mechanisms, where metal oxide catalysts act as oxygen suppliers for the gasification of carbon and are then reoxidized by external oxygen.30−32 Therefore, the high external surface area of the catalyst is crucial for the oxidation of soot. In this study, we found that the doping of both CeO2 and Fe(C5H5)2 can lead to higher counts of metallic nanoparticles as explained in following section. As the iron metal oxide particulates tend to affect the soot formation, they are likely to serve as condensation nuclei for early than the soot inception and therefore enhance soot formation. Conversely, our observations show that the higher concentrations of Fe in diesel fuel led to the lower EC emissions. We therefore believe that the surface enrichment of the catalysts with the soot formed from FBCsdoped fuels can facilitate easier oxygen transfer from the surrounding air to the metal oxide, accelerating the redox process and consequently enhancing the oxidation behavior. With an increase in the engine load, the counts of metallic nanoparticles decreased as found in this study, which suggests a reduced amount of contact points between soot and catalysts, and therefore led to lower percentage reduction of soot emission. Moreover, when compared to CeO2, the decomposition of the same dose of Fe(C5H5)2 produced higher counts of smaller metal oxide nuclei as evident from the higher counts of nanoparticle emissions (see the following section). These Fe-containing nanoparticles, combined with their higher mobility due to the lower melting point, create more contact points for enhancing the soot/catalyst contact under diesel combustion conditions, and therefore appear to be more effective than the cerium oxide catalyst as observed in this study. The addition of FBCs to diesel fuel could also considerably reduce the ignition temperature (light-off temperature) of diesel soot as proposed by some previous studies,12,13,33 which was further confirmed by our current study (see a later section). With a drop in temperature, the oxidation rate of diesel and soot in the last stage of engine combustion will decrease leaving more unburned fuels left to condense on the surface of the soot. Therefore, the reduction of ignition temperature may be one of the factors that led to higher OC emissions. Finally, the lower EC emissions may provide less surface area for OC adsorption, which may in turn lead to lower OC emission. The two factors compete with each other, and therefore lead to a minor variation of OC caused by FBCs-doped fuels. Particle-Phase PAHs and n-Alkanes Emissions. Compared to the undoped ULSD, the total particle-phase PAHs concentrations generally increased with an increase in the Ce

Figure 2. Effect of FBCs-doped fuels on (a) the concentrations of particle-phase PAHs and their total BaPeq (benzo[a]pyrene equivalent), and (b) the concentration distributions of n-alkanes. LMW (low molecular weight, 2 and 3 rings)-PAHs, MMW (medium molecular weight, 4 rings)-PAHs, HMW (high molecular weight, 5 and 6 rings)-PAHs. a Significantly different from the ULSD (p < 0.05). b Significantly different from between Ce-doped fuel and Fe-doped fuel under the same doping dose (p < 0.05).

than that of LMW-PAHs. Another important finding is that with the same dose of FBCs in the fuel, the Fe-doped fuel led to higher total particle-phase PAHs emissions than that of the Ce-doped fuel. A similar effect of both additives on the total BaPeq was also observed in this study, suggesting that both Ceand Fe-doped fuels tend to increase the toxicity of the DPM due to an increase in the particle-phase PAHs emissions. Our observation is consistent with a previous report that the total particle-phase BaPeq was increased by CeO2-doped diesel fuels.8 Figure 2(b) shows that for each fuel, the distribution of particle-phase n-alkanes represents a similar bell-shaped curve with C21 as the most abundant. As in the case of FBCs, both Ce- and Fe-doped fuels led to higher n-alkanes emissions compared to the undoped ULSD, with Fe-doped fuels leading to higher emissions of n-alkanes than that of Ce-doped fuels. It is known that particle-phase PAHs and n-alkanes mainly originate from the unburned diesel fuel or/and lubricating oil. Diesel fuels have a complex mixture of hydrocarbons with chain 4252

DOI: 10.1021/acs.est.7b00920 Environ. Sci. Technol. 2017, 51, 4248−4258

Article

Environmental Science & Technology

Figure 3. Effect of FBCs-doped fuels on (a−c) volatile (without TD) and (d−f) nonvolatile (with TD) particle number size distributions.

C26 probably originated from the lubricating oil, and therefore the emissions of these species are expected to show little variations. As for the shorter-chain alkanes such as C14−C16, on the one hand, they are more likely to undergo nearly complete combustion than the heavier ones. On the other hand, they are not easily attached to the soot particles due to their higher volatility than heavier n-alkanes. As a result, they were only slightly affected by the addition of FBCs. The same reason can also be used to explain the observation that the percentage increase in both MMW-PAHs and HMW-PAHs was higher than that of LMW-PAHs. Particle Number Concentrations and Size Distributions. As shown in Figures 3(a)−(f), the variation of both volatile and solid particle number size distributions, caused by the FBCs-doped fuels, seems to be associated with their proportion used in ULSD and the engine load. At the low engine load (Figure 3(a)), both Ce- and Fe-doped fuels

lengths mostly between C12 and C26 while crankcase oils tend to have a longer chain of carbon atoms. Given the nature of diesel fuel and lubricating oil combined with the distribution of particle-phase n-alkanes observed in this study, it is reasonable to assume that most of the n-alkanes observed in this study are a result of incomplete combustion of diesel fuel and not lubricating oil. The same reason may also be responsible for higher PAHs emissions. This finding further confirms our previously proposed theory that a reduction in the engine ignition temperature due to FBCs can cause an increase in unburned hydrocarbon emissions as discussed in the previous section. Compared to Ce-doped fuels, Fe-doped fuels resulted in a lower ignition temperature and therefore higher particlephase PAHs and n-alkanes emissions. Moreover, we found that the relative increase in C16−C25, caused by FBCs-fuels, was much higher than that of both C14−C15 and C26. This can be explained on the rationale that the heavier n-alkanes such as 4253

DOI: 10.1021/acs.est.7b00920 Environ. Sci. Technol. 2017, 51, 4248−4258

Article

Environmental Science & Technology resulted in the formation of new particles with the nucleation mode within the peak diameters of 10−20 nm. Progressive increases in the concentrations of Ce and Fe in fuels resulted in the size distribution of these nucleation mode particles being shifted upward and toward a larger size, leading to an increase in the emission of total particle counts as shown in SI Table S2. Moreover, for each dose of FBCs in ULSD, the peak of nucleation mode particles for the Fe-doped fuels was higher and shifted toward a larger size than that of the peak associated with the Ce-doped fuels, resulting in the Fe-doped fuels having higher total particle counts than those of Ce-doped fuels, as evident from SI Table S2. At the medium engine load (Figure 3(b)), with the exception of 25 ppm of Ce, all other FBCsdoped fuels showed obvious peaks in the nucleation mode. However, at the high engine load (Figure 3(c)), only the 100 ppm Fe-doped fuel showed a peak in the nucleation mode. In addition, with an increase in the engine load, the peak in the distribution of nucleation mode particles for FBCs-doped fuels became flatter, indicating that the counts of nucleation mode particles decreased with an increase in the engine load. SI Table S2 reveals that for each engine load, with an increase of Ce and Fe concentrations in ULSD, the concentrations of particles with diameter less than 20 nm increased significantly, while the concentrations of particles larger than 50 nm decreased generally. SI Table S2 also shows that the FBCs-doped fuels significantly increased particles of less than 50 nm in some cases, including at the low engine load with all tested additions of FBCs, with the exception of 25 ppm of Ce at the medium engine load, and also at the high engine load with 100 ppm Fe. Our findings indicate that the increase in the total particle counts as well as the decrease in the geometric mean diameter (GMD) caused by FBCs-doped fuels was a result of a significant increase in the counts of smaller particles, while the reductions in large diameter particles were mainly responsible for the decrease in particulate mass, as evident from Figure 1(a). Additionally, our observations also reveal that with the same amount of FBCs doped in the diesel fuel, the increase in the counts of nanoparticles is higher at the lower engine load, and also generally higher for the Fe-doped fuel than compared to the Ce-doped fuel, suggesting the variations in particle counts are associated with the FBC types used and the engine operating conditions. Another interesting finding from Figures 3(d)−(f) is that after TD, the counts of nanoparticles emitted from both Ce- and Fe-doped fuels were significantly reduced due to the suppression of the formation of new particles from volatile species after the dilution and cooling processes, but did not disappear altogether, which were still much higher than those from pure ULSD. Our findings, therefore, confirm that more and larger self-nucleated metallic particles and their agglomerates are generated in the cylinder and emitted to the engine exhaust with an increase in Ce and Fe concentrations in ULSD. The explanations regarding the formation and the evolutionary processes of metallic nanoparticles, and their influence of particle size distributions in response to changes in the proportion of FBCs used in ULSD and the engine load are provided in the SI. The presence of nanoparticles containing Ce and Fe may have profound implications on human exposure and subsequent health effects, and this topic is discussed in a later section. Soot Oxidation Behavior and Soot Nanostructure. Compared to undoped ULSD, a lower soot ignition temperature was obtained from all Ce- and Fe-doped fuels (Figure 4(a)). The extent of the reduction in the ignition temperature

Figure 4. Effect of FBCs-doped fuels on (a) soot oxidation: ignition temperature and activation energy, and (b) the D (defect)/G (graphite) band intensity ratios (ID/IG). a Significantly different from the ULSD (p < 0.05). b Significantly different from between Ce-doped fuel and Fe-doped fuel under the same doping dose (p < 0.05).

seems to have increased with an increase in the Ce concentrations in fuels. However, the reduction in the ignition temperature showed no clear trends with an increase in Fe concentrations. For each engine load, the activation energy of soot generally decreased with an increase in the concentrations of both Ce and Fe in the fuels. Our results suggest that the use of Ce- and Fe-doped fuels offers cobenefits in that these fuels not only produce lower engine output, that is, soot, but also make it easier to oxidize soot by after-treatment systems. In addition, at the high engine load, the soot emitted from Fe100 showed a lower activation energy than that from Ce100, indicating that the soot emitted from Fe100 can be more easily oxidized than that from Ce100 under high engine load. We believe that the Ce and Fe embedded in the diesel soot act as catalysts that reduce the activation energy needed to start soot oxidation and lower the temperature at which soot is oxidized. It should be noted that both Ce- and Fe-based FBCs could effectively promote soot oxidation during the engine combustion process as discussed previously. The reasons responsible for lower soot (EC) emissions can also be used to explain the results in this section. Apart from that, previous observations also suggested that the change in the volatile organic fraction (VOF) in particles,23 the nanostructure of soot34 as well as the spreading of the metal oxide on the soot 4254

DOI: 10.1021/acs.est.7b00920 Environ. Sci. Technol. 2017, 51, 4248−4258

Article

Environmental Science & Technology surface33 affects the catalytic oxidation of soot, which might vary with the Ce and Fe concentrations in the fuels as well as the engine operation conditions. In this study, the higher OC fraction in particles and the higher counts of total particles, caused by FBCs-doped fuels, suggest that the soot from these FBCs-doped fuels could have higher surface areas than that of undoped ULSD, and thus should also be considered as one of factors that led to higher soot oxidation reactivity. The D (Defect)/G (Graphite) band intensity ratios (ID/IG) were used to further explore if and how FBCs-doped fuels affect the oxidation of soot through the change in the graphite-like structure of soot particles. With the decrease in this ratio, the graphite-like structure becomes dominant, implying the soot is not easily oxidized.33 Figure 4(b) shows that the ID/IG ratios of the soot particles emitted from 25 and 50 ppm of Ce-doped fuels were lower than that from undoped ULSD, indicating that the soot had more ordered graphite-like structures and lower amorphous carbon concentrations. However, the soot emitted from all Fe-doped fuels and from the 100 ppm of Ce-doped fuel showed a marginal change in ID/IG ratios, indicating the nanostructure of the soot from these fuels was not significantly changed. The variation in soot nanostructures from the combustion of FBCs-doped fuels in diesel engines could be associated with the changes in the combustion chemistry and in-cylinder combustion and pyrolysis conditions, such as temperature and residence time, etc. The results from this study indicate that the factors that lead to the changes in soot nanostructure are related to the type and amounts of FBCs doped in fuels. Our results further reveal that the changed soot nanostructure, caused by FBCs-doped fuels, did not significantly affect the soot oxidation based on our observations that the soot from Ce25 and Ce50 had more ordered graphite-like structures but still had higher oxidation activity than that from undoped ULSD. Toxicological Studies. Exposure of the cells to a dose of 25, 50, and 100 μg/mL Ce100-derived particles and a dose of 25 and 50 μg/mL Fe100-derived particles led to lower cell viability than those to ULSD-derived particles (Figure 5). Meanwhile, with the exception of the highest exposure does of 200 μg/mL, the cell viability for particles derived from Ce100 was slightly lower than that from Fe100, indicating that the Ce-

based additive leads to the particles having a higher cytotoxicity than that of the Fe-based fuel additive. As shown in Table 1, a short time exposure of A549 cells to the ULSD-derived particles significantly modulated 115 genes, with 70 of them being up-regulated and the 45 down-regulated. While the Ce100 and Fe100-derived particles triggered the modulation of 148 genes (102 up- and 46 down-regulated) and 106 genes (79 up- and 27 down-regulated), respectively. The strikingly altered genes in this assay were widely involved in transcription, cell proliferation/cell cycle/cell aging, metabolism, response to stimulus, cell adhesion, cell signal transduction, response to organic substances/drugs and transport (SI Table S3). Specifically, we observed that the particles emitted from Ce100 caused the down-regulation of genes of CTGF, CDKN2D, OSGIN1, CYR61, DDIT3, UHMK1, and FOS while the particles emitted from Fe100 caused the downregulation of genes of CTGF, OSGIN1, CYR61. Additionally, when compared to undoped ULSD, the particles emitted from Fe100 significantly up-regulated gene DDIT3. The regulation of these genes is involved in the function of cell proliferation inhibition and/or apoptosis promotion.35 Interestingly, compared to the control, the particles emitted from both Ce100 and Fe100 up-regulated the cytochrome P450 family genes (CYP1A1 and CYP3A5), which can be altered by PAHs through the aryl hydrocarbon receptor.36−39 Additionally, the increased emissions of PAHs were also reported to be related to the modulation of the genes involved in signal transduction, apoptosis, metabolic activation process and DNA damage.38 Apart from PAHs, alkanes were shown to up-regulate several genes related to cell cycle.37 Therefore, it is possible that the higher emissions of particle-phase PAHs and n-alkanes, caused by FBCs-doped fuels, may contribute to the variation of gene expressions as observed in this study. Notably, when compared to the control, the particles emitted from Fe100 specifically induced an up-regulation of genes related to the response to reactive oxygen species (ROS)/oxidative stress (ACOX2, ALDH1A3, CYP1A1, DHRS9, HMOX2, FA2H, and PPARGC1A). Moreover, among these up-regulated genes, ACOX2 is reported as a shared genetic risk factor for preeclampsia and cardiovascular diseases,40 and CYP1A1 is associated with the increased lung cancer risk.39 In addition, the cytochrome P450 family genes together with ALDH1A3 are more potent in the induction of many pro-carcinogenic genes (i.e., JUN, FOS).38 Here, we noticed a significant up-regulation of P450 family genes, ALDH1A3 and their corresponding procarcinogenic gene FOS was stimulated by the particles emitted from Fe100. These results suggest that Fe100 may potentially increase the pathogenicity of the diesel exhaust. Interestingly, when we compared to undoped ULSD, we found that the particles emitted from Fe100 induced a significant up-regulation of DDIT3, EGR1, EGR2, FOS, which were associated with multiple biological functions including transcription, cell proliferation, metabolic, and response to ROS. The particles emitted from Ce100 only significantly up-regulated CFB and SERPINA3, which were involved in the regulation of immune and inflammatory responses.41 These results are consistent with previous reports in the literature that the exposure of cells to Fe 2 O 3 nanoparticles through the inhalation route could affect biological targets through generation of ROS by oxygenation of reduced Fe species,16,42 and a short time exposure of cells to CeO2 nanoparticles may lead to an acute inflammatory response.17 Therefore, we believe that emission of the metallic

Figure 5. Effect of particles derived from both ULSD and FBCs-doped fuels on cell viability. a Significantly different from the ULSD (p < 0.05). b Significantly different from between Ce-doped fuel and Fedoped fuel under the same doping dose (p < 0.05). 4255

DOI: 10.1021/acs.est.7b00920 Environ. Sci. Technol. 2017, 51, 4248−4258

Article

Environmental Science & Technology

Table 1. Functional Classification of Genes up- And down-Regulated Following Exposure of A549 Cells After 6 h to Particles Derived from Both FBCs-Doped and Undoped ULSD Fuels ULSD vs control category transcription cell proliferation/growth/cycle/aging metabolism stimulus cell adhesion signal transduction response to organic substance/inorganic substance/drug response to reactive oxygen species/oxidative stress regulation of inflammatory response transport total

Ce100 vs control

Fe100 vs control

Ce100 vs ULSD

Fe100 vs ULSD

upregulated

downregulated

upregulated

downregulated

upregulated

downregulated

up-regulated

up-regulated

8 10 13 8 4 9 10

13 0 15 11 2 0 11

0 15 15 10 0 5 15

11 10 5 8 2 0 0

3 13 17 8 0 19 5

0 3 6 4 2 0 5

0 0 0 0 0 0 0

5 2 4 0 0 0 0

0

6

0

0

7

0

0

2

3 9 0

0 2 45

5 9 102

0 0 46

3 5 79

0 2 27

2 0 17

0 0 7

diesel engines must take into account the different characteristics of DPM holistically in order to provide both environmental and public health benefits. Limitations. The knowledge gained from this preliminary investigation with a small diesel engine (gen-set) can be valuable for selecting appropriate fuel additives or designing after-treatment devices and alternative combustion systems for reduction of DPM emission from diesel engines. However, further studies will still be needed to investigate detailed new particle formation mechanisms in large commercial diesel engines as well as in on-road engines and to examine the influence of fuel properties and operating conditions on the detailed characteristics of DPM while using toxic metalcontaining chemical additives. Apart from PAHs and n-alkanes, other toxic organic compounds such as nitrated- and oxygenated-PAHs, hopanes, steranes, and quinones, should also be determined while using FBCs and their role in the toxicity of DPM should be assessed. Expression profiles of genes induced by DPM in human epithelial cells provided a fundamental understanding of possible biological mechanisms involved in pulmonary diseases, caused by FBCs-doped fuels. The finding from the cDNA microarray technique needs to be confirmed with additional analyses of gene and protein expressions in other conditions such as varied exposure time and dose, and different human pulmonary cell lines using a complementary technique such as quantitative real-time PCR (qPCR). Additionally, particles emitted from different engine types and operating conditions will allow us to gain a better understanding of how the changes in the physicochemical characteristics of DPM, induced by FBCs, trigger the multiple biological effects. It should be acknowledged that the protocols used in the current toxicity study could significantly change the nature of particles emitted from the engine, which might not be representative of in vivo exposure. Therefore, in vivo inhalation exposure studies using animal models over a long time of exposure are needed. Despite these limitations, the findings from this study emphasize the importance of a balance that needs to be maintained between the environmental regulationdriven strategies involving the use of FBCs to reduce the emission of the DPM mass and the key environmental and public health benefits that can be derived, leading to a win−win situation.

nanoparticles from FBC-doped diesel as found in this study may affect the gene expression and cytotoxicity. Apart from the genes with specific functions described above, our data also reveals a broad range of biochemical pathways that are illustrated by genes regulating transcription, transport and cell signal transduction. This observation implies that apart from PAHs and metal oxides, the changes in polar and nonpolar organic compounds in the OC fraction, the physical properties of metallic nanoparticles, caused by FBCs-doped fuels, could also affect the biological reactions occurring at the particles/cell interface and therefore may lead to synergistic or antagonistic biological effects. Notably, the results from this preliminary study suggest that a panel of cytochrome P450 family genes, and the genes that are related to oxidative stress and inflammatory response proposed here deserve considerable attention when conducting further studies on this topic. Environmental and Health Implications. The study reveals that the use of FBCs-doped fuels in diesel engines effectively reduces the DPM mass and EC emissions due to the enhanced soot oxidation, caused by FBCs during engine combustion. The reduction in DPM mass tends to improve urban air quality, whereas that in EC is beneficial from the climate change mitigation viewpoint as EC is a second major contributor to global warming after CO2. However, the FBCsdoped fuels facilitate the formation of self-nucleated metallic nanoparticles, resulting in the increased emissions of nanosize particles. In addition, the use of FBCs-doped fuels increases the organic carbon (OC) fraction of particles and the water-soluble fraction of OC, suggesting an increased fraction of both polar and nonpolar organic compounds in particles, including the specific toxic organic compounds such as PAHs. Another important observation is a decrease in cell viability, as revealed by the in vitro toxicity assessment. Exposure of human epithelial cells to DPM exhibited alterations in the global gene expression with a broad range of biochemical pathways, which could be attributed to the changes in the physicochemical characteristics of DPM caused by the FBCs-doped fuels. Emission of ultrafine particles (diameter