Comparison of the Effect of Biodiesel-Diesel and Ethanol-Diesel on the Particulate Emissions of a Direct Injection Diesel Engine

Aerosol Science and Technology ISSN: 0278-6826 (Print) 1521-7388 (Online) Journal homepage: http://www.tandfonline.com/loi/uast20 Comparison of the Effect of Biodiesel-Diesel and Ethanol-Diesel on the Particulate Emissions of a Direct Injection Diesel Engine Yage Di, C. S. Cheung & Zuohua Huang To cite this article: Yage Di, C. S. Cheung & Zuohua Huang (2009) Comparison of the Effect of Biodiesel-Diesel and Ethanol-Diesel on the Particulate Emissions of a Direct Injection Diesel Engine, Aerosol Science and Technology, 43:5, 455-465, DOI: 10.1080/02786820902718078 To link to this article: https://doi.org/10.1080/02786820902718078 View supplementary material Published online: 25 Feb 2009. Submit your article to this journal Article views: 905 View related articles Citing articles: 34 View citing articles Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalinformation?journalcode=uast20

Aerosol Science and Technology, 43:455 465, 2009 Copyright American Association for Aerosol Research ISSN: 0278-6826 print / 1521-7388 online DOI: 10.1080/02786820902718078 Comparison of the Effect of Biodiesel-Diesel and Ethanol-Diesel on the Particulate Emissions of a Direct Injection Diesel Engine Yage Di, 1,2 C. S. Cheung, 1 and Zuohua Huang 2 1 Department of Mechanical Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon, PR China 2 State Key Laboratory of Multiphase Flow in Power Engineering, Xi an Jiaotong University, Xi an, PR China Experiments were conducted on a 4-cylinder direct-injection diesel engine using ultralow sulfur diesel blended with biodiesel or ethanol to investigate the particulate emissions of the engine under five engine loads at the maximum torque engine speed of 1800 rpm. Four biodiesel blended fuels and four ethanol blended fuels with oxygen concentrations of 2%, 4%, 6%, and 8% were used. With the increase of oxygen content in the blended fuels, the brake specific fuel consumption becomes higher and the brake thermal efficiency improves slightly. The smoke opacity, the particulate mass concentration and the brake specific particulate emissions all decrease, and the reductions are more obvious for the ethanol blended fuels, while the proportion of soluble organic fraction (SOF) in the particle increases with the biodiesel blended fuels having slightly higher proportion of SOF than the ethanol blended fuels. In addition, the total number concentration of particles smaller than 750 nm in diameter decreases gradually for the ethanol blended fuels but increases significantly for the biodiesel blended fuels. The biodiesel blended fuels also increase the number concentrations of particles smaller than 50 nm and particles smaller than 100 nm while the ethanol-blended fuels reduce these particles. [Supplementary materials are available for this article. Go to the publisher s online edition of Aerosol Science and Technology to view the free supplementary files.] 1. INTRODUCTION Diesel particulate matter (PM) is an air toxic and probable carcinogen. This has led to serious consideration on seeking alternatives to replace diesel fuel for diesel engines, either in Received 30 October 2008; accepted 31 December 2008. The authors would like to thank the Hong Kong Polytechnic University (Project Number GU-319), National Natural Science Fund of China (50576070, 50521604), and Xian Jiaotong University for supporting this project. Address correspondence to C. S. Cheung, Department of Mechanical Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon, PR China. E-mail: mmcsc@polyu.edu.hk part or in full, to reduce PM emission. Various oxygenates have been considered based on their availability, price, toxicity, safety and compatibility with diesel fuel. Among them, biodiesel and ethanol are the most widely investigated ones (Agarwal 2007; Demirbas 2007; Ribeiro 2007). Biodiesel is biodegradable, nontoxic, and it can significantly reduce toxic emissions and overall life cycle emission of carbon dioxide from the engine when burned as a fuel (Cvengroš and Považanec 1996; USEPA 2002). It can be used as a blended fuel in diesel engines without modification to the engines. Lapuerta et al. (2008a) conducted an extensive review on the effect of biodiesel fuels on diesel engine emissions. They found that in most of the investigations, HC, CO, smoke and particulate emissions are reduced. However, there is a slight increase in NOx emission. There are also increasing interest in applying biodiesel converted from waste cooking oil for its lower cost and added advantage of reducing waste oil disposal (Wang et al. 2007; Canakci 2007). Significant amount of research has been carried out on the application of ethanol diesel blends to diesel engines. However, a solvent is normally required for blending ethanol with diesel fuel. The blended fuel can lead to a reduction in smoke and particulate matter (PM), an increase in total hydrocarbon, while carbon monoxide and nitrogen oxides could increase or decrease depending on the engine type and operating conditions (Merritt et al. 2005; Li et al. 2005). Published literature indicates that both biodiesel-diesel and ethanol-diesel blends can lead to reduction in PM emissions. However, there is little literature available on the comparison between the PM emissions of biodiesel diesel and ethanol diesel blended fuels, in particular for experiments conducted on the same engine. Miyamoto et al. (1998) found that emission improvements depended almost entirely on the oxygen content of the fuels regardless of the oxygenate to diesel fuel blending ratios and the type of oxygenate. Choi and Reitz (1999) compared the relative effectiveness of the ester and ether fuel blends on particulate emission reduction and they also concluded that as 455

456 Y. DI ET AL. Model TABLE 1 Engine specifications Isuzu 4HF1 Type In-line 4-cylinder Maximum power 88 kw/3200 rpm Maximum torque 285 N m /1800 rpm Bore stroke 112 mm 110 mm Displacement 4334 cc Compression ratio 19.0 : 1 Fuel injection timing (BTDC) 8 Injection pump type Bosch in-line type Injection nozzle Hole type (with 5 orifices) far as the ether and ester oxygenate are concerned, the fuelbound oxygen concentration is the best indicator in determining the amount of soot reduction. On the other hand, some studies, especially those using fuels with high oxygen contents, have shown differences in the effectiveness of different oxygenates in reducing soot (Stoner and Litzinger 1999; Mueller et al. 2003; Buchholz et al. 2004; Pepiot-Desjardins et al. 2008). In fact, different oxygenates blended with diesel fuel have different modifications to the fuel properties. For example, biodiesel blended diesel will have advanced ignition while ethanol blended diesel will have retarded ignition, which might influence the emissions as well. Thus this article aims to compare the PM emission of a diesel engine, using diesel fuel containing less than 50 ppm by weight of sulfur as the base fuel and blended with ethanol or biodiesel to the same oxygen concentrations in the blended fuel. It is known that most of the particles emitted by diesel engines are in the nano-size range with diameter less than 50 nm, and most of the mass of the particles is in the accumulation mode with diameter in the range of 50 nm to 1000 nm (Kittelson 1998). Due to the links between PM emission and health effects, it is necessary to measure the particle characteristics such as number size distribution and mass concentration in order to have a better understanding of the particulate matter. Thus, in this investigation, we will focus on the comparison of smoke opacity, particulate mass concentration and particle number concentration and size distributions. 2. TEST ENGINE AND FUEL PROPERTIES Experiments were carried out on a naturally aspirated, 4- cylinder direct-injection diesel engine with specifications shown in Table 1. The engine was coupled with an eddy-current dynamometer. The engine speed and torque were controlled by the Ono Sokki diesel engine test system which can adjust the engine speed at a fixed engine load or adjust the engine load at a fixed engine speed. The fuels used in this study included ultralow sulfur diesel (ULSD), biodiesel, ethanol and 1-dodecanol. The major properties of these chemicals and their blends are shown in Tables 2 and 3. In this study, biodiesel produced from waste cooking oil by Dunwell Petro-Chemical Ltd., and anhydrous ethanol with a purity of 99.7% were used. Ethanol and biodiesel blends having oxygen concentrations of 2%, 4%, 6%, and 8% were used for this study. The fuel is designated as DB fuel for the biodieseldiesel blends and DE for the ethanol diesel blends. Biodiesel and ULSD can be mixed directly while 1-dodecanol is required to act as a solvent for blending ethanol and ULSD. Thus in the DE blends, the oxygen in both ethanol and 1-dodecanol has been taken into account. The first three blends contain 1% 1- dodecanol while the fourth one contains 1.5% of 1-dodecanol. The volumetric concentration of each chemical as shown in Table 2 was determined for the desired oxygen concentration in the blended fuel. Ethanol has a higher oxygen content of 34.8% while biodiesel has a much lower oxygen content of 10%. This leads to the larger volume of biodiesel required for achieving the same oxygen concentration in the blended fuel. The difference in volume displacement also has an influence on the aromatic content of the blended fuel. Ultralow sulfur diesel TABLE 2 Properties of ultralow sulfur diesel, ethanol, 1-dodecanol, and biodiesel Properties Diesel Ethanol 1-Dodecanol Biodiesel Molecular formula C 2 H 5 OH C 12 H 25 OH Molecular weight 190 220 46.07 186.34 Cetane number 52 6 51 Lower heating value, MJ/kg 42.5 28.4 39.86 37.5 Density@20 C, kg/m 3 840 786 830.9 871 Viscosity@20 C, mpa s 2.8 1.20 16.14 4.6 Heat of evaporation, kj/kg 250 290 840 Stoichiometric air fuel ratio 14.7 8.98 13.4 Autoignition temperature C 316 423 527 - Boiling point// C 185 345 78 257 300 350 C/H/O/%wt 86/14/0 52.2/13/34.8 77.4/14/8.6 78/12/10 Sulfur content/ppm wt <50 0 0 <10

COMPARISON OF THE EFFECT OF BIODIESEL DIESEL AND ETHANOL DIESEL 457 TABLE 3 Properties of the blended fuels Properties DB-1 DB-2 DB-3 DB-4 DE-1 DE-2 DE-3 DE-4 ULSD/vol.% 80.4 60.6 40.6 20.4 92.9 86.8 80.8 74.3 Biodiesel/vol.% 19.6 39.4 59.4 79.6 Ethanol/vol.% 6.1 12.2 18.2 24.2 1-Dodecanol 1 1 1 1.5 Lower heating value, MJ/kg 41.5 40.5 39.5 38.5 41.8 41 40.1 39.2 Density@20 C, kg/m 3 846.1 852.2 858.4 864.7 836 833 830 827 Carbon content/%wt 84.4 82.8 81.2 79.6 85.4 83.3 81.3 79.3 Hydrogen content/%wt 13.6 13.2 12.8 12.4 12.6 12.7 12.7 12.7 Oxygen content/% wt 2 4 6 8 2 4 6 8 contains aromatics while both biodiesel and ethanol are free of aromatics. According to Xiao et al. (2000) and Lapuerta et al. (2008a), aromatics tend to increase soot emission and particulate emission. 3. EXPERIMENTAL SETUP AND PROCEDURE Figure 1 shows the schematic of the experimental system. T 1 and T 2 are thermocouples which measure the inlet air temperature and the exhaust gas temperature. The gaseous species in the engine exhaust, including oxygen, carbon monoxide, carbon dioxide, hydrocarbons, and nitrogen oxides were measured but the results are not presented in this article since the main focus is on particulate emissions. Particles in the engine exhaust were measured with a scanning mobility particle sizer (SMPS, TSI Model 3934) for size distribution and number concentration; and a tapered element oscillating microbalance (R&P TEOM 1105) for mass concentration, in which the main flow rate of sample was 1.5 l/min and the inlet temperature was held at 47 C. The exhaust gas from the engine was diluted with a Dekati mini-diluter (Wong et al. 2003) before passing through the TEOM with primary dilution and the SMPS with second dilution. The dilution ratios were determined from the measured CO 2 concentrations of background air, undi- luted exhaust gas and diluted exhaust gas. Smoke opacity of the exhaust gas was measured with a Diesel tune smoke-meter (SPX DX.210). The Soxhlet apparatus (Wheaton Science Products Inc., USA) was used to determine the SOF in the particles. Pallflex T60A20 Teflon-coated glass fiber filter (Pallflex Inc., USA) with a diameter of 47 mm was used to collect the particulate sample. A solution containing toluene with a purity of more than 99.5% and anhydrous ethanol with a purity of more than 99.7% was used as an extracting solution. The SOF extraction followed the procedures described in Bagley et al. (1998). Experiments were carried out at steady states for different loads at the engine speed of 1800 rev/min. At each mode of operation, the engine was allowed to run for a few minutes until the exhaust gas temperature, the cooling water temperature and the lubricating oil temperature became steady before data were measured. The cooling water temperature varied from 80 Cto 85 C while the lubricating oil temperature varied from 90 C to 100 C, depending on the engine load. All the smoke and PM mass concentrations were continuously measured for five minutes at the exhaust tailpipe of the diesel engine and the average results presented. The steady state tests were repeated to ensure that the results are repeatable within the experimental uncertainties. For particle number concentration and size distribution, three measurements were taken at each mode and the average values are presented. The experimental uncertainty and standard errors in the measurements have been determined based on the method of Kline and McClintock (1953). The maximum standard errors are 2.7% for mass consumption of fuels, 1.3% for particulate mass concentration, 2.5% for particle number concentration, and 2% for smoke opacity. FIG. 1. Schematic diagram of experimental system. 4. RESULTS AND DISCUSSION In this study, experiments were performed at the rated torque speed of 1800 rev/min and engine loads of 28 N m, 70 N m, 130 N m, 190 N m, and 230 N m, corresponding to brake mean effective pressures of 0.08 MPa, 0.20 MPa, 0.38 MPa, 0.55 MPa, and 0.67 MPa, respectively. At each engine load, experiments were carried out for ULSD and for each blended fuel.

458 Y. DI ET AL. FIG. 2. Effect of oxygenate and engine load on smoke opacity. 4.1. Engine Performance For each test, the volumetric flow rate of fuel was measured and then converted into the mass consumption rate based on the density of the fuel. Based on the engine torque, the engine speed and the mass consumption rate of the fuel, the brake power, the brake specific fuel consumption (BSFC), and the brake thermal efficiency (BTE) can be calculated. The results are shown separately as supplementary information to this article. In general, for each engine load, the BSFC increases with the increase of oxygenate in the blended fuels, due to the lower heat value of the oxygenates. For the same oxygen content in the blended fuel, the BSFC for DE fuels are higher than that for DB fuels. The BTE is slightly higher for the DB fuels compared with that for ULSD, but there is no significant change in BTE for the DE fuels. 4.2. Emissions The DB blends and the DE blends have significant influences on smoke and particulate matter emissions. In this article, the effect on smoke opacity, particulate mass concentration, particle number concentration and size distribution were investigated. FIG. 3. Effect of oxygenate and engine load on PM mass concentration.

COMPARISON OF THE EFFECT OF BIODIESEL DIESEL AND ETHANOL DIESEL 459 4.2.1. Smoke Opacity and Particulate Mass Emission Figures 2 and 3 compare the smoke and particulate mass emissions for both DB blends and DE blends. Figure 2 shows that, for each fuel, the smoke opacity of the engine is very low at low and medium engine loads but increases obviously at high engine loads of 0.55 MPa and 0.67 MPa. There is a decrease of smoke opacity with an increase of oxygen in the fuel, and the reduction is particularly obvious at the higher engine loads but not significant at low engine load. Similar results were reported by Lu et al. (2004) and Li et al. (2005) for ethanol diesel blends, and by Nabi et al. (2006) for biodiesel diesel blends. The variation of particulate mass concentration with engine load and oxygen content in the fuel, as shown in Figure 3, are similar to those shown in Figure 2 for the smoke opacity. For each fuel, the particulate mass concentration increases with engine load but decreases with an increase of oxygen in the fuel. Similar trends can be found in Kass et al. (2001) and Lapuerta et al. (2008a). The reduction of smoke opacity can be attributed to several factors. Firstly, the increase of oxygen content and hence a decrease of carbon content in the blended fuels, as shown in Tables 2 and 3, could lead to a reduction in smoke opacity. Secondly, there are fewer C-C bonds in the blended fuel compared with that of ULSD, resulting in the decrease of smoke opacity. Thirdly, the lack of aromatics compounds which have been found to facilitate the soot formation process in a diesel engine may contribute to the decrease in smoke (Chang and Gerpen 1997; Lapuerta et al. 2008b). Fourthly, the oxygen in the fuel can assist in reducing smoke formation during the stage of diffusion combustion. The improvement is more obvious at higher engine loads when a larger percentage of fuel is burned in the diffusion mode. Particulate reduction is associated with the reduction of soot and sulfate in the particulate matter. Pepiot-Desjardins et al. (2008) believed that the oxygenated additives had a dilution effect on the base fuel by replacing the highly sooting components of the base fuel with cleaner hydrocarbons or vice versa, including the reduction of the aromatics in the blended fuels. Particulate reduction can also be attributed to the reduction of sulfur in the blended fuel and hence a reduction of sulfate in the particles (Lapuerta et al. 2008c). In addition, Choi and Reitz (1999) also found that the oxygen in the oxygenated additive is one of the factors in particulate reduction. Comparison between smoke and particulate mass emissions of the DB blends and the DE blends could lead to the conclusion that DE blends provide higher reduction than the DB blends. Miyamoto et al. (1998) showed that the smoke opacity was a decreasing function of the oxygen mass fraction in the fuel and that the rate of reduction was fairly insensitive to the type of oxygenated molecule used as additive. The results obtained in this study are not in line with those of Miyamoto et al. (1998) because for the same mass fraction of oxygen the DB blends have higher smoke and particulate emissions, especially at high engine loads. Thus besides the oxygen content in the oxygenated additive, the molecule structure may play an important role in the reduction of smoke and particulate emissions. It is possible that the ester structure of biodiesel is less effective in reducing the soot precursors than the alcohol structure of ethanol, thus leading to the relatively higher smoke emissions for the DB blends. Mueller et al. (2003) employed di-butyl maleate (DBM) and tri-propylene glycol methyl ether (TPGME) as the oxygenates, which contain ester structure and alcohol structure, respectively. For both experimental investigation and numerical simulation, tri-propylene glycol methyl ether was found to be FIG. 4. Effect of oxygenate and engine load on brake specific PM emission.

460 Y. DI ET AL. more effective in reducing soot than di-butyl maleate for all conditions. As a result, the particulate emissions are higher for the fuels containing ester structure than that for the fuels containing alcohol structure. It has been shown in Table 3 that, for the same oxygen contents of the blended fuels, the biodiesel blended fuels have a larger displacement of the ultralow sulfur fuel and hence a larger reduction in the aromatics in the blended fuel, which should be advantageous to soot and particulate reduction (Xiao et al. 2000; Lapuerta et al. 2008a). However, the experimental results show that the DB fuels have higher smoke and particulate emissions than the DE fuels. It seems that the difference between the ether and alcohol structures has stronger effect on soot and particulate reduction than the difference in the aromatics in the blended fuels. The particulate mass concentration was converted into brake specific particulate emission (BSPM) and the results are shown in Figure 4a and b. For each of the oxygenated fuels, the BSPM emission decreases with an increase in oxygen in the fuel. Comparison between Figure 4a and b again shows that the engine exhaust contains more PM with the DB blends than the DE blends. Figure 4 also shows that the BSPM emission has a minimum at the engine load of 0.2 MPa. Chen et al. (2007) reported the same trend. This is a consequence of the lower brake thermal efficiency at low engine load and higher particulate emissions at high engine load, resulting in the lowest BSPM occurring at 0.2 MPa. 4.2.2. Soluble Organic Fraction (SOF) in Particles Diesel particulate consists of soluble organic fraction (SOF) and insoluble fraction. It has been shown in the literature that an increase of ethanol in the ethanol diesel blends can lead to an increase of SOF in the particles (Chen et al. 2007) while the same trend can also be found in biodiesel diesel blends (Lapuerta FIG. 5. Effect of oxygenate on the proportion of SOF in particle at 0.55 MPa. et al. 2007a). In this study, the proportion of SOF in the particles was analyzed using the Soxhlet extraction method. The results are shown in Figure 5 for the engine load of 0.55 MPa and at the engine speed of 1800 rev/min. As shown in Figure 5, the proportion of SOF increases with the increase of oxygen content in the fuel for both the DE blends and the DB blends. With an increase of oxygenated additive in the blended fuel, there is a reduction of soot in the particles, which is one of the factors leading to an increase in the proportion of SOF. For the same oxygen content, the DB blends have higher SOF proportions, as illustrated in Figure 5. It is possible that there is more unburned hydrocarbon condensing on the particles for the DB blends, which may contribute to the higher particulate emission, compared with the DE blends. 4.2.3. Particle Number Concentration and Size Distribution Influence of particles to the environment and human health depends on its mass concentration as well as on its number concentration and size distribution. In this study, the number concentration and size distributions were measured and compared. The particle size distributions for different fuels at three engine loads are shown in Figures 6 to 8. Figure 9 gives the total number concentration and geometric mean diameter (GMD) of the particles in each case. In this investigation, the equipment was set to measure particles within the size range of 15 nm to 750 nm, thus the total number concentration and GMD shown in Figure 9 are those for particles within the measured range. The effect of engine load is first examined. It can be observed from Figures 6 to 8 that the size distribution curves are all unimodal in shape and most of the measured particles are less than 100 nm in diameter. For each fuel, there is a shifting of the curves upwards with an increase of engine load, indicating an increase of the submicron particles with engine load. The increase of total number concentration and GMD with engine load is also observed in Figure 9. In fact, at higher engine load, more fuel is injected and more fuel is burned in the diffusion mode and hence more particles are formed. In addition, the oxidation rate of soot is reduced in the expansion stroke since there is less time after the end of the diffusion combustion, which will also lead to increase in particle number concentration (Tsolakis 2006). At higher engine load, combustion takes place with lower excess of oxygen but at higher pressure and temperature levels, which will contribute to soot nucleation and promote the growth of the existing soot nuclei (Lapuerta et al. 2007b). With an increase in the number of particles, coagulation rate increases and hence larger particles are formed, leading to an increase of GMD. It is observed that for each engine load the size distribution curves shift towards smaller size with an increase of oxygenate in the blended fuel, indicating an increase of the smaller-sized particles. This is very obvious for the DB blends but less obvious for the DE blends. Figure 9 also shows the gradual reduction of the GMD with an increase of biodiesel or ethanol in the blended fuels. Similar results are obtained by Schroder et al. (1999), Tsolakis (2006), and Lapuerta et al. (2008c). For each

COMPARISON OF THE EFFECT OF BIODIESEL DIESEL AND ETHANOL DIESEL 461 FIG. 6. Effect of oxygenated additives on particulate number concentration and size distribution at 0.20 MPa. engine load, a comparison between the DB fuels and the DE fuels shows that the GMDs of DE-1 and DE-2 are larger than those of DB-1 and DB-2 but the GMDs of DE-3 and DE-4 are smaller than those of DB-3 and DB-4. Also, for each engine load, with an increase of oxygenate in the blended fuel, the size distribution curves shift upwards for the DB blends indicating an increase of the particle number concentration, while the curves for the DE blends become flatter indicating a reduction of the particle number concentration. According to Figure 9, for the engine load of 0.2 MPa, the total number concentration for DB-4 is approximately 2.4 times that of ULSD, and the corresponding values are 1.44 and 1.35, respectively, for engine loads of 0.38 MPa and 0.55 MPa. However, for DE-4, the values are 0.17, 0.17, and 0.24, respectively, for engine loads of 0.2 MPa, 0.38 MPa, and 0.55 MPa. Several factors may contribute to the difference between the effects of the DB blends and the DE blends on particle number concentration and size distribution. Firstly, as discussed above, the ester structure of biodiesel is less effective than the alcohol structure of ethanol in reducing the formation of soot (Mueller et al. 2003; Buchholz et al. 2004; Westbrook et al. 2005). Thus, more soot is available to form particles for the DB blends. As a result, more accumulation mode particles can be found in the exhaust pipe for the DB fuels, especially for DB-3 and DB-4, FIG. 7. Effect of oxygenated additives on particulate number concentration and size distribution at 0.38 MPa.

462 Y. DI ET AL. FIG. 8. Effect of oxygenated additives on particulate number concentration and size distribution at 0.55 MPa. resulting in larger GMDs of DB-3 and DB-4 compared with the GMDs of DE-3 and DE-4. Secondly, as suggested by Tsolakis (2006), the higher viscosity of biodiesel will increase the fuel injection pressure thus resulting in better fuel atomization, airfuel mixing and hence higher production of smaller particles. Pagan (1999) also reported that the increased fuel injection pressure could increase the number of nuclei mode particles. On the contrary, ethanol is less viscous and hence there is a reduction in the number of particles formed. Finally, unburned biodiesel is less volatile, due to its higher boiling point, as shown in Table 3. It might be possible that some of the unburned biodiesel nucleates and condenses into minute particles while the gas sample is cooled down in the passage of the exhaust system which will lead to an increase of the particles in the nucleation mode, especially at the engine load of 0.2 MPa. On the other hand, unburned ethanol is more volatile and hence less liable to nucleate and condense in the exhaust system. The increase of nucleation mode particles contributes to the smaller GMDs of DB-1 and DB-2 compared with the GMDs of DE-1 and DE-2. The particles are further classified into three groups. Particles with diameter larger than or equal to 100 nm, less than 100 nm and less than 50 nm are referred to as PN >100nm,PN <100nm, and PN <50nm, respectively. Thus the PN <100nm particles include the PN <50nm particles. It has been hypothesized that particle FIG. 9. Effect of oxygenated additives and engine load on particle total number concentration and GMD.

COMPARISON OF THE EFFECT OF BIODIESEL DIESEL AND ETHANOL DIESEL 463 FIG. 10. Effect of oxygenated additives and engine load on number concentration of PN >100nm,PN <100nm,andPN <50nm. toxicity may increase with decreasing size due to the high specific surface area of small particles (Peter et al. 1997; Somers et al. 2004; Pope and Dockery 2006). It is possible that nanometer size particles are more dangerous than micron size particles at similar mass concentrations. So it is necessary to investigate the variation of the PN <100nm and PN <50nm particles when using oxygenated additive. The results are shown in Figure 10. For each fuel, there are always more PN <100nm than PN >100nm. There are even more PN <50nm than the PN >100nm. At each engine load, there are always more particles in each group for the DB fuels than the corresponding BE fuels, due to much lower total number of particles generated by the DE fuels. For each fuel, there is an increase of total number concentration of particles with engine load, as shown in Figure 9a. There TABLE S1 Brake specific fuel consumption and brake thermal efficiency 1800 rpm MPa ULSD DB-1 DB-2 DB-3 DB-4 DE-1 DE-2 DE-3 DE-4 BSFC/g/kWh 0.08 546.7 554.7 560.7 573 580.4 548 560.9 580.1 585.9 0.20 315.7 320.6 325.2 330.8 340.5 326.4 329.4 342.9 344.9 0.38 254.5 259 263.9 269.2 275.7 256.2 266.6 271.6 297.4 0.55 229.6 235.3 241.7 248.3 256.4 238.4 242.5 249.8 269.0 0.67 239.2 242.3 247.5 252.5 256.5 244.9 247 254.3 264.1 BTE/% 0.08 15.14 15.72 15.85 15.91 16.11 15.64 15.66 15.48 15.68 0.20 26.10 27.05 27.34 27.56 27.46 26.38 26.66 26.19 26.63 0.38 32.24 33.62 33.7 33.87 33.9 33.52 32.93 33.06 32.39 0.55 35.94 36.87 36.79 36.71 36.47 36.12 36.21 36.52 36.34 0.67 34.44 35.81 35.93 36.10 36.46 35.17 35.55 35.30 34.77

464 Y. DI ET AL. is also corresponding increase in particles in each of the three groups for the ULSD and the DE fuels, which is shown in Figure 10a, b, and c. For the DB fuels, there is increase of PN <100nm and PN >100nm particles with engine load but the number concentration of PN <50nm particles remains almost unchanged. For each engine load, with the increase of the oxygenates, the number concentrations of PN <100nm and PN <50nm decrease for the DE blends but increase for the DB blends as shown in Figure 10b and c. For the engine load of 0.38 MPa, the number concentrations of PN <50nm for DB-1 and DE-1 are approximately 2.09 and 0.71 times, respectively, compared with that of ULSD while the corresponding values for PN <100nm are 1.3 and 0.5 times, respectively, for DB-1 and DE-1. With addition of the oxygenates, the percentage of the PN <100nm particles in the total number concentration increases. At the engine load of 0.2 MPa, the percentage of PN <100nm particles is 81.0% for ULSD, but it reaches 91.2% and 92.6%, respectively, for DB-4 and DE-4. The same trend can be obtained at each engine load. The increase in the percentage of PN <100nm is associated with decrease in particles larger than 50 nm for the DE blends and an increase in particles less than 50 nm for the DB blends. The percentage of PN <50nm also increases with the addition of oxygenated fuels. The increase is most obvious for the engine load of 0.2 MPa, at which the percentage of PN <50nm is 24.2% for ULSD, but reaches 50.9% and 49.0%, respectively, for DB-4 and DE-4. For each fuel the percentage of PN <50nm in general decreases with increasing engine load. For ULSD, the percentage of PN <50nm is 24.2%, 23.3% and 17.6%, respectively, for engine loads of 0.2 MPa, 0.38 MPa, and 0.55 MPa. This is because the PN <50nm particles usually consist of volatile organic compounds and are formed during exhaust dilution and cooling (Kittelson 1998). Thus the higher exhaust gas temperature at higher engine load prevents the condensation of volatile organic compounds, leading to a lower percentage of PN <50nm and PN <100nm particles. 5. CONCLUSIONS Experiments have been conducted on a diesel engine using ultralow sulfur diesel as base fuel, biodiesel and ethanol as oxygenated additives. Four biodiesel blended fuels and four ethanol blended fuels were prepared to give oxygen concentrations of 2%, 4%, 6%, and 8% in the test fuels. Experiments were carried out at an engine speed of 1800 rev/min and at engine loads of 0.08 MPa, 0.2 MPa, 0.38 MPa, 0.55 MPa, and 0.67 MPa. For each engine load, the brake specific fuel consumption increases with the increase of oxygenate in the blended fuels, due to the lower heat value of the oxygenate. For the same oxygen content in the blended fuel, the brake specific fuel consumption of DE fuels is higher than that of DB fuels. The brake thermal efficiency is slightly higher for the DB fuels compared with that for ULSD, but there is no significant change in brake thermal efficiency for the DE fuels. On the emission side, the smoke opacity and particulate mass concentrations both decrease as the oxygen content in the blended fuel increases. For each engine load and the same oxygen content in the blended fuel, smoke opacity and particulate mass concentration are higher for the DB blends compared with that for the DE blends. In addition to oxygen content, the molecular structure of the oxygenate may play a role in affecting smoke and PM formation. From the literature and the results obtained in this investigation, it can be concluded that the ester structure of biodiesel is less effective in reducing smoke opacity and PM emission compared with the alcohol structure of ethanol. In addition, the brake specific particulate emission was evaluated and found to decrease with increase of ethanol/biodiesel in the fuel. The lowest brake specific particulate emission was obtained at the engine load of 0.2 MPa. Besides, SOF in the particulate was analyzed at the engine load of 0.55 MPa and engine speed of 1800 rev/min. The SOF proportion increases with the increase of oxygen content in the blended fuels. With the same oxygen content, the DB fuels have higher proportion of SOF. For each engine load, the geometrical mean diameter of the particles becomes smaller with an increase of oxygen content in the blended fuel, however, the total number concentration of particles increases for the DB fuels but decreases for the DE fuels, compared with ULSD. Further analysis of the particles shows that, at each engine load, the number concentrations of the particles smaller than 50 nm in diameter and the particles smaller than 100 nm in diameter increase with oxygen concentration in the blended fuel for the DB fuels but decrease with oxygen concentration for the DE fuels. Thus the DB fuels tend to generate more of these particles than the DE fuels. It can be concluded that the use of oxygenated additive, both biodiesel and ethanol, is effective in reducing smoke opacity, particulate mass concentration and the reduction is more obvious for the ethanol blended fuels. Besides, the addition of ethanol can reduce the number concentration of particles smaller than 100 nm and even particles smaller than 50 nm. On the contrary, the biodiesel blended fuels lead to the increase of these particles. REFERENCES Agarwal, A. K. (2007). Biofuels (Alcohols and Biodiesel) Applications as Fuels for Internal Combustion Engines, Prog. Energ. Combust. 33:233 71. Bagley, S. T., Gratz, L. D., and Johnson, J. H. (1998). Effect of an Oxidation Catalytic Converter and a Biodiesel Fuel on the Chemical, Mutagenic, and Particle Size Characteristics of Emissions from a Diesel Engine, Environ Sci Technol. 32:1183 1191. Buchholz, B. A., Mueller, C. J., Upatnieks, A., Martin, G. C., Pitz, W. J., and Westbrook, C. K. (2004). Using Carbon 14 Isotope Tracing to Investigate Molecular Structure Effects of the Oxygenate Dibutyl Maleate on Soot Emissions from a DI Diesel Eingine, SAE Technical Paper series: 2004 01 1849. Canakci, M. (2007). The Potential of Restaurant Waste Lipids as Biodiesel Feedstocks, Bioresource Technol. 98:183 90. Chang, D. Y., and Gerpen, J. H. V. (1998). Determination of Particulate and Unburned Hydrocarbon Emissions from Diesel Engines Fueled with Biodiesel, SAE Technical Paper Series: 98:25 27.

COMPARISON OF THE EFFECT OF BIODIESEL DIESEL AND ETHANOL DIESEL 465 Chen, H., Shuai, S. J., and Wang, J. X. (2007). Study on Combustion Characteristics and PM Emission of Diesel Engines Using Ester Ethanol Diesel Blended Fuels, P Combust Inst. 31:2981 2989. Choi, C. Y., and Reitz, R. D. (1999). An Experimental Study on the Effects of Oxygenated Fuel Blends and Multiple Injection Strategies on DI Diesel Engine Emissions, Fuel. 78:1303 1317. Cvengroš, J., and Považanec, F. (1996). Production and Treatment of Rapeseed Oil Methyl Esters as Alternative Fuels for Diesel Engines, Bioresource Technol. 55:145 52. Demirbas, A. (2007). Progress and Recent Trends in Biofuels, Prog. Energ. Combust. 33:1 18. Kass, M. D., Thomas, J. F., Storey, J. M., Domingo, N., and Wade, J. (2001). Emissions from a 5.9 Liter Diesel Engine Fueled with Ethanol Diesel Blends, SAE Technical Paper Series: 2001 01 2018. Kittelson, D. B. (1998). Engine and Nanoparticles: A Review, J Aerosol Sci. 29:575 588. Kline, S. J., and McClintock, F. A. (1953). Describing Uncertainties in Single Sample Experiments, Mech Eng. 75:3 8. Lapuerta, M., Ballesteros, R., and Rodriguez Fernandez, J. (2007a). Thermogravimetric Analysis of Diesel Particulate Matter, Meas Sci Technol.18:650 658. Lapuerta, M., Martos, F. J., and Herreros, J. M. (2007b). Effect of Engine Operating Conditions on the Size of Primary Particles Composing Diesel Soot Agglomerates, J Aerosol Sci. 38:455 466. Lapuerta, M., Armas, O., and Jose, R. F. (2008a). Effect of Biodiesel Fuels on Diesel Engine Emissions, Prog Energ Combust. 34:198 223. Lapuerta, M., Jose, R. F., and Agudelo, J. R. (2008b). Diesel Particulate Emissions from Used Cooking Oil Biodiesel, Bioresource Technol. 99:731 740. Lapuerta, M., Armas, O., and Herreros, J. M. (2008c). Emissions from a Diesel- Bioethanol Blend in an Automotive Diesel Engine, Fuel. 87:25 31. Li, D. G., Huang, Z., Lu, X. C., Zhang, W. G., and Yang, J. G. (2005). Physico Chemical Properties of Ethanol Diesel Blend Fuel and its Effect on Performance and Emissions of Diesel Engines, Renew. Energ. 30:967 976. Lu, X. C., Huang, Z., Zhang, W. G., and Li, D. G. (2004). The Influence of Additives on the Performance and Combustion Characteristics of Diesel Engine, Combust. Sci. Technol. 176:1309 1329. Merritt, P. M., Ulmet, V., and McCormick, R. (2005). Regulated and Unregulated Exhaust Emissions Comparison for Three Tier II Non Road Diesel Engines Operating on Ethanol Diesel Blends, SAE Technical Paper Series: 2005 01 2139. Miyamoto, N., Ogawa, H., Nurum, N. M., Obata, K., and Arima, H. (1998). Smokeless, Low NOx, High Thermal Efficiency, and Low Noise Diesel Combustion with Oxygenated Agnents as Main Fuel, SAE Technical Paper Series: 980506. Mueller, C. J., Pickett, L. M., Siebers, D. L., Pitz, W. J., Westbrook, C. K., and Martin, G. C. (2003). Effects of Oxygenates on Soot Processes in DI Diesel Engines: Experiments and Numerical Simulations, SAE Technical Paper Series: 2003 01 1791. Nabi, M. N., Akter, M. S., and Shahadat, M. M. Z. (2006). Improvement of Engine Emissions with Conventional Diesel Fuel and Diesel Biodiesel Blends, Bioresource Technol. 97; 372 378. Pagan, J. (1999). Study of Particle Size Distributions Emitted by a Diesel Engine, SAE Technical Paper series: 1999 01 1141. Pepiot-Desjardins, P., Pitsch, H., Malhotra, R., Kirby, S. R., and Boehman, A. L. (2008). Structural Group Analysis for Soot Reduction Tendency of Oxygenated Fuels, Combust Flame 154:191 205. Peters, A., Wichmann, H. E., Tuch, T., Heinrich, and Heyder, J. (1997). Respiratory Effects are Associated with the Number of Ultra Fine Particles, Am. J. Resp. Crit. Care. 155, 1376 1383. Pope III, C. A., and Dockery, D. W. (2006). Health Effects of Fine Particulate Air Pollution: Lines that Connect, J. Air & Waste Manag. Assoc. 56, 709 742. Ribeiro, N. M., Pinto, A. C., Quintella, C. M., Rocha, G. O. D., Teixeira, L. S. G., Guarieiro, L. L. N., Rangel, M. D. C., Veloso, M. C. C., Rezende, M. J. C., Cruz, R. S. D., Oliveira, A. M. D., Torres, E. A., and Andrade, J. B. D. (2007). The Role of Additives for Diesel and Diesel Blended (Ethanol or Biodiesel) Fuels: A Review, Energ. Fuel. 21:2433 2445. Schroder, O., Krahl, J., Munack, A., and Bunger, J. (1999). Environmental and Health Effects Caused by the Using of Biodiesel, SAE Technical Paper Series: 1999 01 3561. Somers, C. M., McCarry, B. E., Malek, F., and Quinn, J. S. (2004). Reduction of Particulate Air Pollution Lowers the Risk of Heritable Mutations in Mice, Science, 304, 1008 1010. Stoner, M., and Litzinger, T. (1999). Effects of Structure and Boiling Point of Oxygenated Blending Compounds in Reducing Diesel Emissions, SAE Technical Paper series: 1999 01 1457. Tsolakis, A. (2006). Effect on Particle Size Distribution from the Diesel Engine Operating on RME Biodiesel with EGR, Energ. Fuel. 20:1418 1424. USEPA. (2002). A Comprehensive Analysis of Biodiesel Impacts on Exhaust Emissions, EPA420 P 02 001. Wang, Y., Ou, S. Y., Liu, P. Z., and Zhang, Z. S. (2007). Preparation of Biodiesel from Waste Cooking Oil via Two-Step Catalyzed Process, Energ. Convers. Manage. 48:184 188. Westbrook, C. K., Pitz, W. J., and Curran, H. J. (2006). Chemical Kinetic Modeling Study of the Effects of Oxygenated Hydrocarbons on Soot Emissions from Diesel Engines, J. Phys. Chem. A. 110:6912 6922. Wong, C. P., Chan, T. L., and Leung, C. W. (2003). Characterization of Diesel Exhausts Particle Number and Size Distributions Using Mini-Dilution Tunnel and Ejector-Diluter Measurement Techniques, Atmos. Environ. 37, 4435 4446. Xiao, Z., Ladommatos, N., and Zhao, H. (2000). The Effect of Aromatic Hydrocarbons and Oxygenates on Diesel Engine Emissions, P I Mech. Eng. D J Aut. 214, 307 332.