Environ. Sci. Technol. 2007, 41, 6479-6483
Nuclei-Mode Particulate Emissions and Their Response to Fuel Sulfur Content and Primary Dilution during Transient Operations of Old and Modern Diesel Engines Z . G E R A L D L I U , * ,† VICTORIA N. VASYS,† AND DAVID B. KITTELSON‡ Cummins Emission Solutions, 1801 U.S. Highway 51-138, Stoughton, Wisconsin 53589, and University of Minnesota, 111 Church Street SE, Minneapolis, Minnesota 55455
The effects of fuel sulfur content and primary dilution on PM number emissions were investigated during transient operations of an old and a modern diesel engine. Emissions were also studied during steady-state operations in order to confirm consistency with previous findings. Testing methods were concurrent with those implemented by the EPA to regulate PM mass emissions, including the use of the Federal Transient Testing Procedure-Heavy Duty cycle to simulate transient conditions and the use of a Critical Flow Venturi-Constant Volume System to provide primary dilution. Steady-state results were found to be consistent with previous studies in that nuclei-mode particulate emissions were largely reduced when lowersulfur content fuel was used in the newer engine, while the nuclei-mode PM emissions from the older engine were much less affected by fuel sulfur content. The transient results, however, show that the total number of nucleimode PM emissions from both engines increases with fuel sulfur content, although this effect is only seen under the higher primary dilution ratios with the older engine. Transient results further show that higher primary dilution ratios increase total nuclei-mode PM number emissions in both engines.
Experimental Section
Introduction Particulate matter (PM) and nitrogen oxides (NOx) are harmful to human health and the environment (1-3); thus, diesel engines and aftertreatment technologies continually evolve to remain in accordance with ever-intensifying emissions regulations. Modern engine technologies strive to reduce PM mass emissions, but have been shown to simultaneously reduce particle size (4). Nanoparticles, particles with a diameter, Dp, less than 50 nm, contribute little to the PM mass emissions because of their small size, but may contribute greatly to the PM number emissions which remain unregulated (5). These particles have been found to deposit in the alveolar region of the lung, where they are difficult to remove and have the potential to enter the blood stream, * Corresponding author tel: 608.877.3802; fax: 608.873.1550; e-mail:
[email protected]. † Cummins Filtration & Emission Solutions. ‡ University of Minnesota. 10.1021/es0629007 CCC: $37.00 Published on Web 08/15/2007
leading to detrimental effects on human health (6). Nanoparticles consist mainly of nuclei-mode particles which have a Dp of 5-30 nm (7, 8) and which are presumed to form by the nucleation of water and sulfur compounds (9-11). Although sulfur compounds play a critical role in the formation of the nuclei mode, heavy hydrocarbons may comprise most of the nuclei-mode mass in engines without aftertreatment systems (12-14). Furthermore, sulfur compounds poison exhaust aftertreatment catalysts, causing the U.S. to begin implementing ultralow sulfur diesel fuel (ULSD, less than 15 ppm sulfur) to replace the low-sulfur diesel fuel (LSD, less than 500 ppm sulfur) (2). Because the new fuel will be used in all engines, it is important to understand the effects of sulfur content on PM number emissions of both old and modern engines. Ristovski et al. (15) found that the lower-sulfur fuel had little effect on PM mass and number emissions of old, heavily used engines when compared to the effect on newer models. The engines, however, were only tested under steady-state operating conditions, which have limited ability to simulate genuine engine operations. In an on-road vehicle and during transient testing, the engine load, speed, and exhaust gas temperature (EGT) fluctuate continuously, significantly affecting the size distribution and number concentration of PM emissions (7, 10, 12, 16). Nuclei-mode particles are additionally affected by the primary dilution, as this is where most of the formation and growth occur. To completely understand the effects of implementing ULSD fuel on the PM emissions, a systematic approach, which includes broad ambient measurements of PM number concentrations from a vehicle fleet as well as a detailed investigation of these effects under controlled laboratory conditions, is being conducted. This study is part of the laboratory investigation, which utilized transient testing to examine the effects of fuel sulfur content and primary dilution on particle size distributions and number emissions. Two engines were tested under transient and steady-state operations for each of four combinations of fuel and primary dilution flow rate. Primary dilution was provided by a Critical Flow Venturi-Constant Volume System (CFV-CVS) that maintained proportional sampling throughout temperature excursions and was designed for engine emission certification. Particle size and number emissions were measured with a TSI Engine Exhaust Particle Sizer (EEPS) Model 3090 during transient operation and with a TSI Scanning Mobility Particle Sizer (SMPS) Model 3080 during steady-state testing.
2007 American Chemical Society
Test Apparatus and Dilution Systems. The test apparatus for this study is shown in Figure 1. The engine exhaust was primarily diluted by a CFV-CVS tunnel with pre-activated carbon, and HEPA filtered air. A sample of the diluted exhaust stream was then taken isokinetically to an aging chamber which had a 2:1 dilution ratio and a 2-s residence time. The exhaust was then sampled and diluted by two ejection-type microdiluters in series with a combined dilution ratio of 185:1 and finally measured by particle sizing instrumentation. In many previous studies, the primary dilution was accomplished by microdiluters (with a fixed dilution ratio) which diluted a sample of the exhaust. In this study, however, the primary dilution was accomplished by a CFV-CVS tunnel which diluted the entire exhaust gas. The constant total flow in the tunnel resulted in a continuously varying dilution ratio because exhaust flow changes constantly during transient testing. The ratio at any specific moment is calculated from dividing the tunnel flow rate by the flow rate of the exhaust, VOL. 41, NO. 18, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6479
FIGURE 1. Test apparatus.
TABLE 1. CFV-CVS Dilution Ratios Calculated from Tunnel Flow Rate, and Engine Speed, Displacement, and Volumetric Efficiency primary dilution ratio 0.708 m3/s
1.298 m3/s
1992 Engine ISO 1 (3000 rpm, 339 N-m) ISO 8 (1600 rpm, 213 N-m) ISO 11 (890 rpm, 9 N-m) FTP average FTP acceleration average FTP deceleration average
4.6 8.6 15.2 7.2 5.7 7.8
8.4 15.7 27.9 13.3 10.5 14.3
2004 Engine ISO 1 (1800 rpm, 2190 N-m) ISO 8 (1400 rpm, 1306 N-m) ISO 11 (600 rpm, 58 N-m) FTP average FTP acceleration average FTP deceleration average
2.1 2.7 6.3 3.5 3.8 4.4
3.9 5.0 11.6 6.3 7.1 8.0
test condition\flow rate
which was determined by engine speed, displacement, and volumetric efficiency. The primary dilution ratios of the FTP cycle ranged from 4:1 to 30:1 for the 1992 engine and from 2:1 to 13:1 for the 2004 engine as shown in Table 1. The tunnel flow rate is regulated by multiple critical flow venturis and can range from 0.354 m3/s (750 scfm) to 2.596 m3/s (5500 scfm). Two flow rates of 0.708 m3/s (1500 scfm) and 1.298 m3/s (2750 scfm) were chosen for this testing to study the effect of primary dilution ratio. Particle Sizing Instrumentation. Two instruments were used for particle-size data collection: a TSI model 3090 Engine Exhaust Particle Sizer and a TSI model 3080 Scanning Mobility Particle Sizer. The EEPS, described in full detail by Johnson et al. (17), classifies particles based on differential electrical mobility. It was programmed to scan particles over a size range of 5.6-560 nm across 16 channels per decade every tenth of a second and had a response time of about 1 s; thus it was used for data collection during transient testing. The SMPS, described in full detail by Liu et al. (18), also classifies particles based on differential electrical mobility, via a long differential mobility analyzer (DMA) and an ultrafine condensation nuclei particle counter (UCPC). The SMPS was programmed to scan particles over a size range of 7.0-294 nm across 32 channels per decade with an up-scan time of 6480
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 18, 2007
120 s and thus could only be used for data collection during steady-state testing. Both instruments were individually calibrated by the manufacturer prior to testing. The flows and voltages of the SMPS were validated at the testing facility. The EEPS was then validated by comparing the particle size distributions measured while the EEPS and the SMPS were connected in parallel to each other, after the secondary dilution, during ISO Mode 11 for both fuels and dilution conditions. Although the size distributions and number concentrations measured by the EEPS tend to shift to a smaller size and a lower number, respectively, this shift remains within the acceptable size variation of less than 10% and concentration variation of less than 20%, necessary to be considered valid. Testing Procedure. The engines tested, a 1992 International engine and a 2004 Cummins ISX engine (see Table 2), were certified for the emissions regulations of 1990 and 2004, respectively; thus, their technologies are extensively diverse. The 2004 engine, equipped with an engine exhaust gas recirculation system, emits significantly less PM mass, but its effects on PM number emissions are previously unknown. The engines were tested under the same conditions during each of four combinations of fuel and CFV-CVS flow rates. The fuels used to power the engines were both obtained from Chevron Phillips: a 2007 certified No. 2 ULSD with a sulfur content of 9.2 ppm and a 2004 certified No. 2 LSD with a sulfur content of 308 ppm. In addition to the fuel sulfur content, differences were seen between ULSD and LSD for their Cetane number (42 and 47), PM content (3.3 and 0.1 mg/L), aromatic content (32.0 and 29.5 LV%), and boiling points (ULSD was about 5% lower than LSD). The majority of the other properties of the two fuels were similar. The engines were first warmed up at ISO-8178 Mode 8 for 30 min, and then Modes 1, 8, and 11 were run for 15 min each and finally three Federal Transient Testing Procedure-Heavy Duty (FTP HD) cycles were conducted. The 1200-s FTP HD cycle consists of four phases, each of which controls the engine to achieve various arrays of speed and load. The order of the phases: New York Non-Freeway (NYNF), Los Angeles Non-Freeway (LANF), Los Angles Freeway (LAFY), and NYNF repeated.
Results and Discussions Steady-State Testing. Steady-state tests were conducted to observe the effects of fuel sulfur content on PM number
TABLE 2. Engine Specifications model year configuration fuel injection displacement rated speed maximum torque maximum power EGR emissions certification
International Engine
Cummins ISX Engine
1992 V-8 naturally aspirated indirect injection (IDI) 7.3L (444 in3) 3000 rpm 412 N-m (304 lb-ft) @1600 rpm 115 kW (155 hp) @ 3000 rpm no 1990
2004 V-6 turbocharged/charged air cooled HPI electronic 14.9L (912 in3) 1800 rpm 2510 N-m (1851 lb-ft) @ 1200 rpm 373 kW (500 hp) @ 2000 rpm yes 2004
FIGURE 2. Normalized particle size distributions of the emissions from both the 1992 engine and the 2004 engine during ISO Modes 1 and 11 with a CFV-CVS tunnel flow rate of 1.298 m3/s. emissions, to compare results to findings from previous studies, and to validate the testing instrumentation. The engine data collected from the EEPS and the SMPS during steady-state tests were used to validate the EEPS against the more established SMPS. Figure 2 shows the normalized particle size distributions from emissions of both engines using ULSD and LSD during ISO Modes 1 and 11 with a CFV-CVS flow rate of 1.298 m3/s, corresponding to dilution ratios of 8:1 and 28:1 for the 1992 engine and 4:1 and 13:1 for the 2004 engine during ISO 1 and ISO 11, respectively. The higher flow rate is displayed because it has been used extensively for transient PM mass emissions testing. ISO 1 and 11 are displayed because ISO 8 had similar results as ISO 1 and its inclusion would complicate the figure. There was little difference in emitted particle number or size distribution between the two fuels in the accumulation mode, consistent with the findings of many previous studies that fuel sulfur only affects nuclei-mode particles. During idle operations (ISO 11), very little accumulation-mode material (soot), which acts to adsorb volatile materials and suppress nuclei mode formation, is formed. This, combined with low EGTs, favors nucleation by sulfuric and organic materials during ISO 11 (7, 19). The effect of fuel sulfur on the emissions of the 1992 engine was less profound than that on the emissions of the 2004 engine because the older engine produces more soot particles; thus, nuclei-mode particle formation and growth are inhibited. This also explains why ISO 1 produced a nuclei mode in the 2004 engine’s emissions but not in the emissions of the 1992 engine. Transient Testing. In order to characterize the emission patterns during transient operation of the engines, instantaneous data of an acceleration and a deceleration during the FTP cycle were analyzed. The acceleration occurred during the New York Non-Freeway section (because of the varying response times of the engines, the acceleration occurred at 1120-1130 s for the 1992 engine and at 11161123 s for the 2004 engine) where the speed increased from
FIGURE 3. Number of (a) accumulation-mode particles (DP > 30 nm) and (b) nuclei-mode particles (DP < 30 nm) emitted from both engines during acceleration. 1773 to 2943 rpm for the 1992 engine and from 600 to 1218 rpm for the 2004 engine. The torque fluctuated from 247 N-m to 313 N-m for the 1992 engine and from 161 N-m to 2626 N-m for the 2004 engine. The deceleration occurred at the end of the Los Angeles Freeway section (894-910 s for the 1992 engine, 895-900 s for the 2004 engine) where the speed decreased from 2159 to 1550 rpm for the 1992 engine and from 1093 to 609 rpm for the 2004 engine. The torque decreased from 201 N-m to -82 N-m for the 1992 engine and from 266 N-m to -358 N-m for the 2004 engine. The particle size distributions from the acceleration and deceleration are averaged from three different FTPs. The effect of either dilution ratio or fuel sulfur is difficult to observe from instantaneous data, therefore charts of the total numbers of accumulation-mode (DP > 30 nm) and nuclei-mode particle (DP < 30 nm) emissions versus time are displayed in Figure 3. It is evident that the number of accumulation-mode particles increases with acceleration because of increased fuel consumption, and that as accumulation-mode particles increase, the number of nucleiVOL. 41, NO. 18, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6481
FIGURE 4. Particle size distributions of emissions from both engines (a) with a CFV-CVS tunnel at flow rate of 1.298 m3/s using ULSD and LSD and (b) using LSD during deceleration with CFV-CVS tunnel at flow rates 1.298 m3/s and 0.708 m3/s.
mode particles decreases because volatile materials are adsorbed by the soot particles and high EGTs reduce gasto-particle conversion processes. The nuclei modes of the 2004 engine emissions show a greater difference between LSD and ULSD than for the 1992 engine’s nuclei modes, likely because the 1992 engine produces more accumulationmode particles. In each engine, however, the effects of fuel sulfur content and dilution ratio on nuclei-mode particle formation are not obvious during acceleration. Throughout the deceleration, sulfur content and dilution ratio strongly affected PM emissions of both engines, as seen in Figure 4. This is because deceleration provides favorable conditions for sulfuric compounds and volatile materials to nucleate: there are less soot particles emitted, hydrocarbons are not efficiently burned during deceleration (20), and EGTs decrease significantly. It is also notable that the nuclei modes of the 1992 engine are much broader, therefore contain more particles, than those of the 2004 engine, even though the accumulation mode of each engines’ emissions are similar. This is because the 1992 engine has lower EGTs than the 2004 engine, providing cooling which strongly promotes nucleation and particle growth. Figure 5 displays the numbers of accumulation-mode (DP > 30 nm) and nuclei-mode particles (DP < 30 nm) as a function of time, in order to give a more thorough representation of the entire deceleration. It is evident that neither fuel sulfur content, nor dilution ratio, affects the accumulation-mode particles, while both increase the total number of nuclei-mode particles emitted. During deceleration while using LSD, the nuclei mode is always larger for the higher dilution ratios. This is because increased cooling caused by higher dilution ratio and the addition of fuel sulfur promote nucleation. 6482
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 18, 2007
FIGURE 5. Number of (a) accumulation-mode particles (DP > 30 nm) and (b) nuclei-mode particles (DP < 30 nm) emitted from both engines during deceleration.
FIGURE 6. Normalized total PM number emissions during entire FTP cycles. Although instantaneous data is useful for viewing the effects of fuel sulfur content and dilution ratio on PM number emissions during specific engine operating conditions, it is not a feasible method for assessing their total effects during the FTP cycle. An integration of all particle emissions over the duration of the cycle provides a more comprehensive method of assessment as there are 12 000 continuous frames of data incorporated. Figure 6 displays the normalized total PM number per bhp‚h emitted during the entire FTP cycle of each testing configuration. Each bar is the sum of the total number of accumulation-mode particles, and the total
number of nuclei-mode particles and each is the average of three FTP cycle integrations, with the standard deviations displayed. There is no significant effect of fuel sulfur content on accumulation-mode particles in either engine, confirming the findings of previous steady-state studies that neither fuel sulfur nor primary dilution conditions affect the size or number of soot particles. Nuclei-mode particles, however, are affected by both factors in both engines. The number of nuclei-mode particles increased significantly with fuel sulfur except with the lower dilution ratios in the older engine’s emissions because there was not enough cooling to counter the amount of soot particles which suppress nuclei-mode particle formation. The higher dilution ratios promoted nucleation in all cases. It also appears that the older engine emitted more nuclei-mode particles despite its greater amount of soot particles emitted. However, the dilution ratios of the two engines were different; thus, their emissions may not be comparable. In general, the combination of higher-sulfur fuel and higher primary dilution ratio was found to provide optimal conditions for nuclei-mode PM formation in both engines. Higher-sulfur fuel creates more opportunity for nucleation, and higher primary dilution ratios provide shorter residence time and lower concentrations of volatile materials, leading to slower growth and less scavenging by coagulation with larger particles. Furthermore, deceleration generates low EGTs, low soot concentrations, and likely high amounts of unburned lube fractions, all of which, combined with sulfate precursors, create favorable conditions for nuclei-mode particle formation. Acceleration, however, generates high EGTs and high soot concentrations, both of which inhibit nuclei-mode particle formation. Transient testing led to the formation of distinct nuclei modes in the emissions of both the old and modern engines under multiple conditions, suggesting that steady-state operations of an engine may not generate sufficient conditions to produce a broad-enough range of PM number emissions.
Acknowledgments The authors would like to acknowledge Tim Johnson and Joe Bester of TSI for their invaluable help with instrumentation, and Thaddeus Swor, Devin Berg, Tom Wosikowski, and Howard Tews of Cummins Filtration and Emissions Solutions for their assistance with the manuscript and experimentation.
Literature Cited (1) Ying, Q.; Mysliwiec, M.; Kleeman, M. J. Source apportionment of visibility impairment using a three-dimensional sourceoriented air quality model. Environ. Sci. Technol. 2004, 38, 10891101. (2) Control of Air Pollution from New Motor Vehicles: Heavy-duty engine and vehicle standards and highway diesel fuel sulfur control requirements: final rule. 40 CFR Parts 69, 80, and 86; U.S. Government Printing Office: Washington, D.C., 2001, pp 5001-5193. (3) Brunekreef, B.; Holgate, S. T. Air pollution and health. Lancet 2002, 360, 1233-1242.
(4) Bagley, S. T.; Baumgard, K. J.; Gratz, L. D.; Johnson, J. H.; Leddy, D. G. Characterization of fuel and aftertreatment device effects on diesel emissions. CRC Rep. 1996, 76. (5) Kittelson, D. B. Engines and nanoparticles, a review. J. Aerosol Sci. 1998, 28, 575-580. (6) O ¨ berdorster, G.; Ferin, J.; Lehnert, B. E. Correlation between particle size, in vivo particle persistence, and lung injury. Environ. Health Perspect. 1994, 102, 173-179. (7) Kittelson, D. B.; Watts, W. F.; Johnson, J. P. Diesel aerosol sampling methodology - CRC E-43 final report. Coordinating Research Council 2002, 43, 181. (8) Giechaskiel, B.; Ntziachristos, L.; Samaras, Z.; Scheer, V.; Casati, R.; Vogt, R. Formation potential of vehicle exhaust nucleation mode particles on-road and in the laboratory. Atmos. Environ. 2005, 39, 3191-3198. (9) Vouitsis, E.; Ntziachristos, L.; Samaras, Z. Modeling of diesel exhaust aerosol during laboratory sampling. Atmos. Environ. 2005, 39, 1335-1345. (10) Schneider, J.; Hock, N.; Weimer, S.; Borrmann, S.; Kirchner, U.; Vogt, R. Nucleation particles in diesel exhaust: Composition inferred form in situ mass spectrometric analysis. Environ. Sci. Technol. 2005, 39, 6153-6161. (11) Vehkamaki, H.; Kulmala, M.; Lehtinen, K. E. J. Modeling binary homogeneous nucleation of water-sulfuric acid vapours: Parameterisation for high temperature emissions. Environ. Sci. Technol. 2003, 37, 3392-3398. (12) Tobias, H. J.; Beving, D. E.; Ziemann, P. J.; Sakurai, H.; Zuk, M.; McMurry, P. H. Chemical analysis of diesel engine nanoparticles using a nano-DMA/thermal desorption particle beam mass spectrometer. Environ. Sci. Technol. 2001, 35, 2233-2243. (13) Sakurai, H.; Tobias, H. J.; Park, K.; Zarling, D.; Docherty, K. S.; Kittelson, D. B.; McMurry, P. H.; Ziemann, P. J. On-line measurements of diesel nanoparticle composition, volatility, and hygroscopicity. Atmos. Environ. 2003, 37, 1199-1210. (14) Khalek, I. A.; Kittelson, D. B.; Brear, F. 2000. Nanoparticle growth during dilution and cooling of diesel exhaust: Experimental investigation and theoretical assessment. SAE 2000, 2000-010515. (15) Ristovski, Z. D.; Jayaqratne, E. R.; Lim, M.; Ayoko, G. A.; Morawska, L. Influence of diesel fuel sulfur on nanoparticle emissions from city buses. Environ Sci. Technol. 2006, 40, 13141320. (16) Kleeman, M. J.; Schauer, J. J.; Cass, G. R. Size and composition distribution of fine particulate patter emitted from motor vehicles. Environ Sci. Technol. 2000, 34, 1132-1142. (17) Johnson, T.; Caldow, R.; Pocher, A.; Mirme, A.; Kittelson, D. B. A new electrical mobility particle sizer spectrometer for engine exhaust particle measurements, SAE 2004, 2004-01-1341, Testing and Instrumentation, SAE SP-1871. (18) Liu, Z. G.; Verdegan, B. M.; Badeau, K. M.; Sonsalla, T. P. Measuring the fractional efficiency of diesel particulate filters. SAE Trans. 2002, 2002-01-1007. Diesel Exhaust Emission Control: SCR, HC De-NOX, and Measurements, SAE SP-1674: 73-82. (19) Kittelson, D. B.; Watts, W. F.; Johnson, J. P. 2006. On-road and Laboratory Evaluation of Combustion Aerosols Part 1: Summary of Diesel Engine Results, J. Aerosol Sci. 2006, 37, 913-930. (20) Heywood, J. B. Internal combustion engine fundamentals; McGraw-Hill, Inc.: New York, 1988.
Received for review December 6, 2006. Revised manuscript received May 25, 2007. Accepted June 26, 2007. ES0629007
VOL. 41, NO. 18, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6483
