Linking Load, Fuel, and Emission Controls to ... - ACS Publications

Subscriber access provided by Fudan University

Article

Linking Load, Fuel and Emission Controls to Photochemical Production of Secondary Organic Aerosol from a Diesel Engine Shantanu H Jathar, Beth Friedman, Abril A. Galang, Michael F. Link, Patrick Brophy, John Volckens, Sailaja Eluri, and Delphine K. Farmer Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b04602 • Publication Date (Web): 10 Jan 2017 Downloaded from http://pubs.acs.org on January 10, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 18

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Environmental Science & Technology

Linking Load, Fuel and Emission Controls to Photochemical Production of Secondary Organic Aerosol from a Diesel Engine Shantanu H. Jathar1*, Beth Friedman2, Abril A. Galang1, Michael F. Link2, Patrick Brophy2, John Volckens1, Sailaja Eluri1 and Delphine K. Farmer2 1 Department of Mechanical Engineering, Colorado State University, Fort Collins, CO 2 Department of Chemistry, Colorado State University, Fort Collins, CO * Corresponding Author ([email protected])

1.

Abstract

Diesel engines are important sources of fine particle pollution in urban environments but their contribution to the atmospheric formation of secondary organic aerosol (SOA) is not well constrained. We investigated direct emissions of primary organic aerosol (POA) and photochemical production of SOA from a diesel engine using an oxidation flow reactor (OFR). In less than a day of simulated atmospheric aging, SOA production exceeded POA emissions by an order of magnitude or more. Efficient combustion at higher engine loads coupled to the removal of SOA precursors and particle emissions by aftertreatment systems reduced POA emission factors by an order of magnitude and SOA production factors by factors of 2-10. The only exception was that the retrofitted aftertreatment did not reduce SOA production at idle loads where exhaust temperatures were low enough to limit removal of SOA precursors in the oxidation catalyst. Use of biodiesel resulted in nearly identical POA and SOA compared to diesel. The effective SOA yield of diesel exhaust was similar to that of unburned diesel fuel. While OFRs can help study the multiday evolution, at low particle concentrations OFRs may not allow for complete gas/particle partitioning and bias the potential of precursors to form SOA.

2.

Introduction

Diesel engines account for approximately three-quarters of the PM2.5 (particulate matter with an aerodynamic diameter < 2.5 µm) emissions from mobile sources in the United States.1 PM2.5 is an EPA-defined criteria pollutant and PM2.5 derived from diesel sources has impacts on air quality,2 climate,3 and human health4. Over the past three decades, research and regulatory efforts have strived to measure, study and reduce primary PM2.5 emissions from diesel engines with much success. In the United States and European Union, on-road diesel vehicles now require the use of particulate filters to achieve compliance on primary emissions of PM2.5; regulations in both regions are also moving towards the use of particulate filters for non-road applications. Recently, several studies have shown that unburned and partially burned hydrocarbons in diesel exhaust can photochemically react in the atmosphere to form lowvolatility products that condense into the particle phase as secondary organic aerosol (SOA).5-8 SOA – defined as the aerosol mass formed from the atmospheric oxidation of volatile organic compounds (VOC) – is an important, and in some cases the dominant, constituent of ambient

ACS Paragon Plus Environment


49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97

PM.9 However, the sources, formation pathways and continued atmospheric processing of SOA remains poorly understood.10 Although previous studies have helped state the importance of SOA formation from diesel exhaust, they are subject to two key limitations. First, all prior studies have been performed using environmental chambers that simulate photochemistry over short photochemical time periods (~12 hours) and are influenced by gas11 and particle12 losses to the chamber walls. Second, studies to examine the influence of important engine parameters (e.g., operation at less fuel-efficient loads, aftertreatment) have been performed across different engine-vehicle configurations. Given that emissions can vary substantially within and between engine-vehicle configurations,13 previous studies present trends but do not provide a quantitative assessment of how SOA production can vary with changes in a single engine parameter (while holding all other parameters constant). Hence, research is needed to: test alternative methods to study SOA formation from diesel exhaust that simulate longer, more atmospherically-relevant, photochemical time periods; to perform experiments that are less susceptible to wall-losses; and to use a single-engine platform to examine the influence of varying engine parameters in a controlled manner. For reasons ranging from energy security to concerns about sustainability, numerous countries (including the United States) are promoting the use of alternative bio-derived fuels for diesel engines.14 Biodiesel use in the United States , as a blend or substitute for diesel, has increased by a 100-fold over the past 15 years and continues to grow,15 annual biodiesel consumption in the United States was ~1,500 million gallons compared to diesel consumption of 63,000 million gallons . Further, stricter environmental standards in the United States and European Union now mandate the use/retrofit of aftertreatment systems on diesel exhaust streams (e.g., Drayage Truck Regulation set by the California Air Resources Board16). Typically, these aftertreatment systems include oxidation catalysts to reduce emissions of carbon monoxide and unburned hydrocarbons, diesel particulate filters to remove PM emissions, and lean-NOx (NO+NO2) traps or selective catalytic reduction (SCR) systems to remove NOx. Furthermore, While earlier work provides some insight,6, 17 it is unclear how the wider adoption of biodiesels or the use/retrofit of aftertreatment devices influences the potential of diesel exhaust to form SOA. In this work, we perform photochemistry experiments to investigate organic aerosol (OA) emissions and production from diesel exhaust as a function of engine load, fuel and a retrofitted aftertreatment device. To examine SOA production over longer atmospherically-relevant spatiotemporal scales and to reduce the influence of wall-losses, we use an oxidation flow reactor (OFR) to simulate atmospheric oxidation.18 To reduce the effects of engine-to-engine variability, we conduct experiments on a single Tier 3 diesel engine that is capable of running straight neat fatty acid methyl ester-based biodiesel and can be retrofitted with aftertreatment devices to meet Tier 4 standards. While we concede that findings from tests performed on a single engine may not necessarily be representative, we recommend that future work validate our results and perform similar experiments on different/newer engines.

3.

Methods

The goal of this study was to examine POA emissions and SOA production from a modern-day, non-road diesel engine for different engine loads, fuels, and aftertreatment using an OFR. The study was performed as part of the Diesel Exhaust Fuel and Control (DEFCON) study conducted at Colorado State University from June 3-12, 2015. DEFCON was designed to study properties of photochemically aged gas- and particle-phase pollutants from diesel exhaust.19, 20.


Page 2 of 18

Page 3 of 18

98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148


The sections below present details about the engine, experimental setup and matrix, instrumentation and techniques used to analyze the data.

3.1.

Experimental Methods

Engine: Experiments were performed on a 4-cylinder, turbocharged and intercooled, 4.5 L, 175 hp, John Deere 4045 PowerTech Plus engine mounted on an engine dynamometer (Midwest Inductor Dynamometer 1014A). The engine is equipped with a variable geometry turbocharger, exhaust-gas recirculation, and electronically controlled high-pressure common rail fuel injection and meets non-road Tier 3 emissions standards.21 A single aftertreatment package consisting of a diesel oxidation catalyst (DOC, John Deere RE568883) and diesel particulate filter (DPF, John Deere RE567056) was retrofitted on the exhaust system and configured to meet non-road interim Tier 4 emissions standards.21 This engine has been part of several research studies in the past.22-24 Setup: A schematic describing the experimental setup along with key instrumentation is shown in Figure S-1. Raw exhaust from the engine was sampled through an isokinetic probe using 15 feet of Silcosteel® tubing heated to 150 °C. The raw exhaust was diluted with activated charcoal- and HEPA-filtered air using a Hildemann-style dilution sampler,25 which allowed for dilution ratios of ~45-110. The dilution ratios were varied to perform experiments at roughly similar pollutant concentrations and calculated using the method outlined by Lipsky and Robinson26 based on CO2 measurements. A portion of the diluted exhaust (~27 standard L min1 ) was moved to a 320 L stainless steel tank (~50 cm diameter x 150 cm long) with a ~6 minute residence time before another portion (7 standard L min-1) of the diluted exhaust from near the center of the tank was pulled through the OFR; we assume that gases and particles near the center were less susceptible to loss to the tank walls. A large gas molecule with a molecular weight of 150 g mole-1 and a diffusion coefficient of 4 x 10-6 m2 s-1 would travel less than 6 cm in 6 minutes; particle movement from diffusion would be even smaller. The OFR produces atmospheric oxidants (specifically the hydroxyl radical, OH) and simulates atmospheric aging.27 The OFR used in this work was a 13-L aluminum horizontal cylinder and equipped with a mercury lamp that produced light at wavelengths of 185 and 254 nm.18, 19, 28 Both wavelengths contributed to the production of OH radicals inside the OFR.29 The production of OH radicals and consequently the OH exposure inside the OFR was controlled by adjusting the light intensity via the voltage applied across the lamp. Voltages between 0 and 105 V were able to produce OH exposures between 0 and 1.4-7.2x107 molecules-hr cm-3 (see Section 3.2 for details on OH exposure calculations) or 0 and 0.4-2 OH days of photochemical aging (1 OH day can be interpreted as the net exposure to an OH concentration of 1.5x106 molecules cm-3 for 24 hours30). The residence time in the OFR was ~100 s. All gas-phase instruments directly sampled from the OFR; these flows totaled ~3.3 standard L min-1. Another 3.7 standard L min-1 of flow was further diluted with activated charcoal- and HEPA-filtered air by a factor of 3 to 7 before being sampled by the particle-phase instruments. Instruments: Some of the instrumentation has been described in earlier publications from the DEFCON study.19, 20 Here, we only describe instruments central to this work. A Siemens 5-gas analyzer measured concentrations of CO2, CO, THC, NO, NO2 and O2 in the raw exhaust; raw exhaust was transferred to the 5-gas analyzer using a 110 C heated Teflon™ transfer line followed by a water trap. A scanning mobility particle sizer (SMPS, TSI) measured particle size distributions after passing through a neutralizer, a photoacoustic extinctiometer (PAX, Droplet Measurement Technologies) measured black carbon concentrations and a high-resolution aerosol mass spectrometer (HR-AMS, Aerodyne Research) measured the non-refractory chemical composition of sub-micron particles. The HR-AMS data were analyzed with the toolkit



149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199

Squirrel (version 1.57) and Pika (version 1.16) to obtain mass concentrations and compositional information (sulfate, nitrate, ammonium, organics).31 Aerosol composition and elemental analysis were conducted with the high-resolution data and used to calculate bulk H/C and O/C ratios for OA.31 Test Matrix: For the engine without the aftertreatment, we performed two tests (base and repeat) at two different engine loads (idle and 50% load) and each with two different fuels (diesel and soy-based biodiesel). We also performed an additional test at idle load with diesel fuel and an engine ramp test at 2200 RPM where we varied the torque to produce engine loads from 0 to 50% (results from the engine ramp test are presented in Figure S-5). The idle load corresponded to negligible shaft power (~0 kW) at 900 RPM while the 50% load corresponded to shaft power of 60 kW at 2200 RPM (rated speed); the idle and 50% load approximately correspond to modes 8 and 3 on the ISO 8178-4 C1 duty cycle, respectively. The diesel fuel was commercial, non-road, ultra-low-sulfur diesel (ULSD) and sourced locally. The biodiesel fuel (B100) was sourced from Emergent Green Energy (Minneola, KS) and produced from soy. For the engine with the retrofitted aftertreatment (single unit consisting of a DOC and DPF), we performed only one test (no repeats) at the two different engine loads and using the two different fuels. In summary, thirteen tests were conducted where each test included gradual aging of the diluted exhaust using the OFR. The voltage applied to the lamp inside the OFR was changed incrementally (0, 23, 28, 39, 50, 61, 105 and 0 V) to produce increasing OH concentrations. The 0 V at the end of the experiment provided data to calculate drift in engine emissions during the course of the experiment. Each voltage setting was held for twenty minutes, with the first ten minutes being used to flush the OFR volume and the next ten minutes used to record data. The total length of each experiment for a given engine condition and fuel type was ~150 minutes.

3.2.

Analysis Methods

OH Exposure: The OH exposure was calculated using methods and parameterizations provided in Peng et al.32 and Li et al.29. Briefly, the photon flux at 185 and 254 nm was calculated based on the voltage applied to the lamp; for example, 105 V translated to a lamp power of 80% and a photon flux of 4.4x1013 and 4.4x1015 photons cm-2 s-1 at 185 and 254 nm respectively. Next, the photon flux at 185 nm, relative humidity (as measured by the PAX), external OH reactivity (see text below) and residence time (100 s) in the OFR were used with the spreadsheet model provided by Peng et al.32 to determine total OH exposure. Previous work has indicated that the OH radical concentrations in the OFR can be substantially reduced when oxidizing a reactive mixture.29, 32 The spreadsheet provided by Peng et al.32 accounts for the ‘suppression’ in OH exposure if it is provided with the net external OH reactivity. The external OH reactivity was calculated for THC, CO, NO and NO2 and summed before input to the spreadsheet.19 The OH reactivity for the hydrocarbons was calculated using the emissions profile listed in Table S-1 (more details about the profile are provided in the Section SOA Yields). We note that in the absence of available data, the THC emissions for all engine load/fuel/aftertreatment combinations are assumed to have the same OH reactivity per unit mass. We acknowledge that this is a broad assumption, and warrants investigation in future studies. Emission Factors: Background-corrected emissions factors (EF) for primary emissions of CO2, CO, NO, NO2, THC, BC and POA and the production factors for SOA and inorganic aerosol were calculated using equation 1. Here, emission/production factors are expressed as grams of pollutant produced per kg of fuel burned. Since more than 98% of the fuel carbon was emitted as CO2 or CO, we assume that in equation 1 all of the carbon in the fuel was converted to CO2 or CO.


Page 4 of 18

Page 5 of 18

200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247


=

× ×

(1)

Where [P] is the background corrected pollutant concentration in µg m-3, [CO2] and [CO] are the background corrected CO2 and CO concentrations in µg m-3, MW CO2, MW CO, and MW C are the molecular weights for CO2, CO and C and Cf is the carbon mass fraction in the fuel in kg-C kgfuel-1. We use a Cf of 0.85 for diesel and 0.77 for biodiesel.17 SOA Yields: The SOA mass yield is defined as the ratio of the SOA mass formed to the THC mass reacted. This ratio represents the fraction of reacted hydrocarbons that form aerosol and is commonly used to represent SOA formation in large-scale models. The SOA mass was calculated by subtracting POA from the OA measured by the HR-AMS; an SOA estimation using positive matrix factorization analysis (described in the supporting information) did not change the results. The THC mass concentration reacted was calculated using equation 2. Δ = ∑ × 1 − exp−!"#, %Δ&''

(2)

Where ∆THC is the THC mass concentration reacted, HCi is the unaged hydrocarbon mass concentration, kOH,i is the reaction rate constant with OH, [OH] is the OH concentration in the OFR and ∆t is the residence time in the OFR. Hydrocarbon concentrations were determined by multiplying THC concentrations with the normalized VOC speciation profile listed in Table S-1. This detailed VOC speciation profile for diesel exhaust was constructed from two different sources. The ‘Diesel Exhaust Farm Equipment’ profile (Profile 3161) from EPA’s SPECIATE database (version 4.3) was used to determine contributions for species smaller than C12. We used the work of Zhao et al.,33 in which gas-phase emissions from a suite of on- and non-road diesel engines were characterized, to determine contributions for species larger than C12. For the speciation profiles that were developed across a drive cycle, Zhao et al.33 did not seem to find any large differences between the on- and non-road diesel engines. We also used Zhao et al.33 to apportion the THC mass between species smaller than (40%) and larger than C12 (60%). The same speciation profile is used to speciate the THC emissions from all the experiments since there are no available data that can provide the level of detail achieved here as a function of engine load, fuel, and aftertreatment.

4.

Results

4.1.

General Results

Emission factors for measured gas- and particle phase species are listed in Table 1. For no aftertreatment, CO, THC, POA and BC emissions were higher (by factors of approximately of 12, 7, 6, and 3 respectively) at the idle load than at the 50% load. With aftertreatment, emissions for CO and THC were substantially higher (factors of ~400 and 33, respectively) at the idle load than at the 50% load with modest increases in POA (factor of 2), likely highlighting the poor performance of the DOC to reduce CO and THC emissions at idle conditions . The trends observed for the different loads and aftertreatment were similar for both fuels, except for THC, in which biodiesel resulted in lower (~30-60%) emissions compared to diesel. 10% of biodiesel mass is oxygen; THC emissions from biodiesel could be more oxygenated than those from diesel use. The flame ionization detector (FID) used to measure THC emissions in this work is known to be less efficient in measuring functionalized carbon,34 and the lower THC emissions measured for biodiesel may thus be a sampling artifact.



248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298

At the 50% load, the without and with aftertreatment configurations of the engine met the nonroad Tier 3 and Tier 4 emissions standards for CO, THC and PM (see Table 1); the only exception was that the THC emission with aftertreatment exceeded the Tier 4 emissions standard by 10%. Very few SOA studies have been performed with non-road diesel engines35 and hence, we compared primary emission factors at 50% load conditions for diesel fuel to the compendium of results published by May et al.13. We did so to assess if results from our study were directly comparable in terms of primary emissions and hence generalizable to secondary production from real-world diesel sources. May et al.13 include primary (or tailpipe) emission factors from source testing and near-road/tunnel studies performed on several on-road diesel vehicles, a subset of which were explored for SOA formation by Gordon et al.17. We should note that on-road and non-road engines, while identical in a lot of respects, differ in their real-world duty cycles and have different calibrations to meet their respective emissions certification. For with and without aftertreatment, the CO, NOx and PM (POA+BC) emission factors determined in our study were within the range of the median/average values in May et al.13 (for a detailed comparison see Table S-2). These results also suggest that the retrofitted aftertreatment system produced emissions reductions in CO and PM similar to those found in real-world on-road diesel vehicles. For no aftertreatment, THC emissions in this work were 2.5 to 5 times higher than the median/average values of May et al.13. This suggests that the THC emissions in this work could be higher than those from a representative on-road diesel engine. Surprisingly, the THC emissions with the retrofitted aftertreatment, despite nearly meeting the non-road Tier 4 standard, were about 200 times higher than those measured by May et al.13. Hence, the aftertreatment package used in this work that was retrofitted to meet Tier 4 standards, was not as effective in reducing THC emissions as those found on aftertreatment systems employed on vehicles tested in May et al.13 . So despite CO and PM reductions, it is likely that the retrofitted aftertreatment system used in our work may require the engine to be recalibrated to reduce THC emissions. Regardless, the findings from this work are relevant for engine systems that are retrofitted with aftertreatment systems Surprisingly, the CO and THC emissions at idle loads (for both fuels) did not vary significantly with the retrofitted aftertreatment. We hypothesized that the exhaust temperatures associated with the idle load were too low to activate the DOC and oxidize the CO and THC. To test that hypothesis, we ran an additional test where we suddenly transitioned the engine retrofitted with the aftertreatment from a higher steady-state load (more than 10%) to an idle load. For this test, we monitored the exhaust temperature before and downstream of the aftertreatment and measured CO concentrations in real-time. As the engine ramped down to idle conditions, the exhaust cooled but the CO concentration remained constant at about 2 ppm. However, once the exhaust temperature dropped below ~130-150 C, we observed a sharp increase in CO (2+ orders of magnitude) suggesting that the DOC’s removal efficiency for CO (and presumably THC and SOA precursors) was temperature dependent. This implies that the DOC (responsible for oxidizing CO and THC) function might be limited during idle loads when exhaust temperatures are lower. The POA emissions were significantly lower (factor of ~10-40) with the use of the aftertreatment at both engine loads and for both fuels, implying that the DPF was operational across those conditions.36-38 Emissions of NOx did not vary substantially between selected engine load/fuel/aftertreament combinations but the NO-NO2 split did vary with the presence of the aftertreatment. The aftertreatment package appeared to reduce NO2 to NO at idle loads (reducing environment in the presence of THCs) but oxidized NO to NO2 at 50% loads (oxidizing environment). The differences in NO-NO2 splits have been previously observed for on-road diesel vehicles13, 36 and lean-burn natural gas engines39.


Page 6 of 18

Page 7 of 18

299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348


Raw measurements of OA (and particle sulfate and nitrate) from a representative experiment (no aftertreatment, diesel fuel, and 50% load performed on June 5th, 2015) are shown in Figure 1. Measured OA levels varied with the voltage applied across the UV lights inside the OFR. Increasing voltages correspond to increased exposure to OH radicals that drive the oxidation of VOCs and produce SOA. The estimated OH exposure in OH-equivalent days is presented in parentheses. We observed a dramatic increase (six-fold) in the OA concentration and atomic O:C ratio (three-fold) with voltage and OH exposure. For this experiment, we observe a slight decrease in OA mass concentration at the lower voltages (28-39 V), which can be attributed to evaporation and/or wall-loss linked to the slight increase in temperature inside the OFR from the lamps,40 data in this work have not been corrected for temperature changes in the OFR. The photochemical SOA production observed in this work is much higher than that observed in previous studies6, 17 and can be attributed to the high OH exposures possible with an OFR. For all experiments (including the one presented in Figure 1), more than 95% of the nonrefractory PM mass was OA. Particle sulfate was low because of the use of ULSD fuel. Particle nitrate was also low, in contrast to the results of Tkacik et al.41 who performed photo-oxidation experiments on motor vehicle exhaust from a road tunnel using an OFR. Tkacik et al.41 proposed that the high particle nitrate was a result of NO oxidation to HNO3 followed by neutralization by ammonia emitted by gasoline vehicles. The low particle nitrate in this work may be attributed to the absence of ammonia emissions from diesel engines; note that we did not use an SCR system, which is known to emit ammonia. The low particle nitrate is evidence that neither primary emissions from diesel engines, nor photo-oxidation of diesel exhaust produces substantial quantities of organic nitrates.17

4.2.

POA Emissions and SOA Production

Blank-corrected results from all the experiments are plotted in Figure 2 on an emission factorbasis, namely as grams of OA emitted or produced per kilogram of fuel burned (see supporting information for details on blank correction and Figure S-2). The results are plotted against the estimated OH exposure (or photochemical age), which is captured using units of ‘OH equivalent days’30. The maximum OH exposure varied by a factor of 5 across experiments (OH exposure ranged from 0.4 days for experiments performed at 50% load with aftertreatment to 2 days for experiments performed at idle load and no aftertreatment) because the external OH reactivity of the exhaust was sufficiently different between engine load/fuel/aftertreatment combinations (see Figure S-3). SOA Production: All engine load/fuel/aftertreatment combinations produced substantial quantities of SOA compared to POA with just half to a full day of photochemical processing. At the highest photochemical age, the SOA:POA ratios were ~12-25 for the experiments without aftertreatment and ~80-800 for the experiments with aftertreatment. The SOA:POA ratios for experiments without aftertreatment compared well with those measured by Tkacik et al.41 (~10) who used an OFR. The high SOA:POA ratios for the aftertreatment experiments can be attributed to the dramatic reduction in POA emissions. Further, the POA and SOA factors determined from the positive matrix factorization analysis compared well against factors calculated by Presto et al.42 from chamber experiments performed on on-road diesel emissions (see Figure S-4). This comparison implies equivalence in POA emissions and photochemical production of SOA not only between chamber and OFR experiments but also between on- and non-road engines. The simulated time periods are representative of spatiotemporal scales found in urban airsheds and highlight the important role that SOA would play in determining the total PM contribution from diesel sources.



349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399

Focusing on the 50% engine load experiments (Figure 2b), the 4.5L John Deere engine used in this work (without aftertreatment) exceeds the non-road Tier 2/3 emission standard set by the EPA. Note that we only consider the OA contribution to PM and do not account for emissions of BC or inorganic aerosol in Figure 2. When compared against the bands that encompass different EPA emissions standards for non-road diesel engines, the SOA-included emission factor for OA, after one day of photochemical aging, comes close to exceeding the Tier 2/3 emission standard. Given that we do not see the same reductions in THC emissions as earlier work, the findings described here provide an upper bound estimate for SOA production. The effect is even more dramatic at idle conditions where the OA emission factor is 3 times higher than the Tier 1 emission standard. This rapid rise shows that it is vital to account for gas-phase precursors and SOA in any future design and implementation of emissions standards for diesel engines. One needs to be careful with a literal interpretation of the results in Figure 2 as the emission/production factors for OA have been calculated across a large range of OA mass concentrations (2-1500 µg m-3) and thus need to be revised for atmospherically-relevant concentrations of OA (~10 µg m-3) before comparing results between the different engine load/fuel/aftertreatment combinations (see Section 4.3). Engine Load: POA emission factors and SOA production factors at idle loads were much higher (~10x) than at 50% loads. Understandably, this is because the engine is much less efficient at burning fuel and hence allows for higher emissions of unburned and partially burned combustion products in the aerosol- (as POA) and gas-phase (as SOA precursors). Simply, this suggests that even if a diesel engine operates at idle loads for 10% of its duty cycle, emissions at the idle load would be as important as the emissions produced at all other loads combined. The finding highlights the need to study the OA system for engine operating conditions such as idle17, 43 and cold starts44 where SOA precursor emissions could be high enough to dominate SOA production over the entire duty cycle. Fuel: There were no substantial differences between the experimental results of the diesel (N=6) and biodiesel (N=5) fuel, regardless of the engine load-aftertreatment configuration. The result tentatively suggests that the use of biodiesel may not confer advantages in emissions reductions of SOA (or POA). Although our finding is consistent with the results from Gordon et al.2 (biodiesel has very little influence on SOA production), this is a surprising result if we assume that the SOA precursors in the emissions are derived from the fuel and/or depend on the fuel composition. Diesel and biodiesel fuel are chemically different. Diesel fuel is a complex mixture of straight, branched and cyclic alkanes and aromatic molecules that range from a carbon number of 6 to 25.45 In contrast, fatty acid methyl ester-based biodiesel is dominated by a narrow range of esters between carbon numbers of 16 to 20.46 Aftertreatment: The retrofitted aftertreatment (in the form of a DPF and DOC) substantially reduced (factor of 25) the POA emissions at both engine loads. However, the aftertreatment experiments exhibited a much stronger production of SOA compared to the experiments with no aftertreatment. In the idle load experiments, SOA:POA was ~800 at one OH-equivalent day and the SOA production factors were nearly identical to those in the absence of aftertreatment. The results were not as dramatic for the 50% load experiments but still large with an SOA:POA ratio of ~80 at 0.4 OH-equivalent days and SOA production factors slightly lower than without aftertreatment. The DOC is designed to remove unburned hydrocarbons (and CO) in the exhaust stream by oxidation to CO2 and water, and hence should reduce emissions of SOA precursors. As suggested in Section 4.1, the DOC (which was retrofitted) may be limited in reducing SOA precursor emissions at idle conditions when exhaust temperatures are lower.


Page 8 of 18

Page 9 of 18

400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450


4.3.

SOA Potential of Diesel Exhaust

The experiments performed in this work varied significantly in terms of OA mass concentrations (~2-1500 µg m-3), which were higher than earlier chamber studies (~20-100 µg m-3).6, 7, 17 The variation in OA mass concentrations primarily resulted from the use of different engine loads, fuels, aftertreatment and dilution ratios, which can all vary SOA production. Combustion-related POA and SOA follow absorptive gas/particle partitioning theory,8, 47 causing emission factors to be biased high (Figure 2). These emission factors cannot be directly compared across experiments, and are less relevant at atmospheric concentrations of OA. To enable a better comparison across experiments, we present the SOA mass yield for each experiment as a function of the OA mass concentration (Figure 3). The SOA mass yield is calculated as the SOA mass formed divided by the THC mass reacted during each experiment. As mentioned earlier, we ignore differences in the FID response and we use the same speciation profile (and subsequently the same OH reactivity per unit mass) for the THC for all experiments. These assumptions need to be investigated in future work. The SOA mass yield increased with the OA mass concentration, which is typical for SOA formed from model compounds like toluene48 and α-pinene49 in chamber-based growth experiments. Although there were minor differences and a few exceptions, the SOA mass yield for all diesel experiments appeared to cluster together. This implies that THC emissions from a diesel engine operated under different engine loads and aftertreatment have roughly the same potential and possibly the same precursor composition to form SOA. The SOA mass yields for biodiesel were found to be slightly higher than those for diesel (p

Linking Load, Fuel, and Emission Controls to ... - ACS Publications

Recommend Documents