Influence of Fuel Injection Timing and Pressure on In-Flame Soot

Jun 16, 2014 - ... in coflow laminar ethylene diffusion flames at elevated pressures .... “Who has been an impactful women mentor in your career and...
0 downloads 0 Views 5MB Size
Article pubs.acs.org/est

Influence of Fuel Injection Timing and Pressure on In-Flame Soot Particles in an Automotive-Size Diesel Engine Renlin Zhang* and Sanghoon Kook School of Mechanical and Manufacturing Engineering, the University of New South Wales, Sydney, New South Wales 2052, Australia S Supporting Information *

ABSTRACT: The current understanding of soot particle morphology in diesel engines and their dependency on the fuel injection timing and pressure is limited to those sampled from the exhaust. In this study, a thermophoretic sampling and subsequent transmission electron microscope imaging were applied to the in-flame soot particles inside the cylinder of a working diesel engine for various fuel injection timings and pressures. The results show that the number count of soot particles per image decreases by more than 80% when the injection timing is retarded from −12 to −2 crank angle degrees after the top dead center. The late injection also results in over 90% reduction of the projection area of soot particles on the TEM image and the size of soot aggregates also become smaller. The primary particle size, however, is found to be insensitive to the variations in fuel injection timing. For injection pressure variations, both the size of primary particles and soot aggregates are found to decrease with increasing injection pressure, demonstrating the benefits of high injection velocity and momentum. Detailed analysis shows that the number count of soot particles per image increases with increasing injection pressure up to 130 MPa, primarily due to the increased small particle aggregates that are less than 40 nm in the radius of gyration. The fractal dimension shows an overall decrease with the increasing injection pressure. However, there is a case that the fractal dimension shows an unexpected increase between 100 and 130 MPa injection pressure. It is because the small aggregates with more compact and agglomerated structures outnumber the large aggregates with more stretched chain-like structures.



INTRODUCTION Soot particle emissions from diesel engines have a negative impact on human health and the environment. Previous studies showed that fine soot particles with a size of hundreds nanometers emitted from diesel engines can reach vital organs through the respiration system and cause various diseases.1−3 Recently, it was reported that the exposure to diesel engine exhaust particles could increase the risk for lung cancer.4 Studies also showed that diesel soot particles are suspected to accelerate global warming through absorbing heat from sunlight when suspended in the atmosphere.5 Consequently, the reduction of soot emissions has become a high priority for engine manufacturers. Two well-known methods to reduce the exhaust soot emission are the earlier fuel injection with respect to the crank angle degree (i.e., advanced injection timing) and the increased injection pressure. For example, exhaust soot measurements using an exhaust gas opacity meter, exhaust particle sizer (EPS), and electric low-pressure impactor (ELPI) generally show a lower soot level or smaller particle size (mobility diameter) for high injection pressure and advanced injection timing conditions.6−10 Laser-based diagnostics in optically accessible constant-volume combustion chambers11−13 or diesel engines14−17 clarified the origin of the reduced exhaust soot such that the high injection pressure and advanced © 2014 American Chemical Society

injection timing suppress the soot formation and promote oxidation within the flame, due to the improved air-fuel mixing. While the soot reduction is promising, a new issue is that ultrasmall particles with defective nanostructures could be more toxic than those emitted from old diesel engines.18,19 Therefore, many researchers studied the size and structure of soot particles that are sampled from the diesel exhaust through thermophoresis (i.e., positive thermal diffusion) and analyzed them using a transmission electron microscope (TEM).20−26 However, it should be noted that the exhaust soot particles are the result of complex formation and oxidation processes inside the cylinder of engine. If one could sample the soot particles directly from the diesel flame, then the causes of exhaust soot structures would be found, which would then help validate soot modeling14 and develop more advanced combustion strategies to reduce the soot formation as well as toxicity. Such approaches were attempted by previous researchers using a bulk-gas sampling technique in a diesel engine27 or direct inflame soot sampling in constant-volume combustion chambers with quiescent ambient gas conditions.28−33 Recently, we Received: Revised: Accepted: Published: 8243

February 8, 2014 June 11, 2014 June 16, 2014 June 16, 2014 dx.doi.org/10.1021/es500661w | Environ. Sci. Technol. 2014, 48, 8243−8250

Environmental Science & Technology

Article

combined the advantages of these studies such that the direct in-flame soot sampling was successfully implemented in a working engine for the first time.34 This study further utilizes the direct soot sampling technique in our optical diesel engine to provide a detailed understanding of in-flame soot particles with a particular interest in their dependency on the fuel injection parameters. Specifically, we aim to show the effect of injection timing and pressure on the number and projection area of soot particles per TEM image, size distributions, and the fractal dimension of soot particles within the diesel flame.

Table 1. Engine Specification and Operating Conditions



EXPERIMENTAL SECTION Engine Specifications and Operating Conditions. The experimental setup of this study is illustrated in Figure 1 and

displacement volume

498 cm3 (single cylinder)

bore/stroke compression ratio swirl ratio coolant temperature intake air temperature engine speed injection system fuel cetane number nozzle hole diameter mass per injection injection pressure injection timing

83 mm/92 mm 15.2 1.4 90 °C 30 °C 1200 rpm Bosch second-generation common-rail injector ultralow-sulfur diesel 51 134 μm (nominal) 9 mg 70−160 MPa −12∼−2 °CA aTDC

Therefore, a portion of the piston bowl-rim was removed, which resulted in a reduced compression ratio of 15.2. It is noted that this compression ratio is in line with the production engines currently in the market. A detailed description and discussion of the implemented modification and the associated impact on the engine can be found in our previous work.34 The intake port design of the engine achieved a swirl ratio of 1.4 when no throttling was applied. The heated water at 90 °C was flowed through the cylinder liner, engine head, and engine block to simulate a thermally stable, warmed-up engine condition. The engine was naturally aspirated, and the intake air temperature was measured at 30 °C throughout the experiments. All tests were conducted at a constant engine speed of 1200 rpm (revolution per minute) using a 37 kW AC motor. In this study, a second-generation Bosch common-rail injector was used to deliver ultralow-sulfur diesel fuel with the cetane number of 51. The original injector had a 7-hole nozzle but only one hole was left open by using a laser-welding technique. This single-hole arrangement was to isolate a single diesel jet from complex jet-to-jet interactions. Also, it permitted long injection duration with a low risk of optical windows failure, allowing 9 mg per injection (per hole) of fuel mass. If all 7 holes were used, then the injected mass would correspond to a high-load operating condition when the soot emissions are particularly problematic. Two sets of fuel injection parameters were tested in this study including four injection pressures varied from 70 to 160 MPa at fixed injection timing of −7 crank angle degrees after the top dead center (°CA aTDC), and three different injection timings ranging from −12 to −2 °CA aTDC with fixed injection pressure at 70 MPa. Throughout the tests, the incylinder combustion conditions were monitored by measuring the pressure for various °CAs using a piezo-type pressure transducer (Kistler 6056A). The measured in-cylinder pressure traces were used to calculate the apparent heat release rate (aHRR) traces. In-Flame Soot Particle Sampling. Soot particles were sampled using an in-house developed probe that holds a 3 mm diameter, 400-mesh, and carbon-coated copper TEM grid.34 As depicted in Figure 1, soot particles were collected via the thermophoretic principle between the hot soot-laden gas and the cold carbon layer on the grid. It is important to note that the TEM grid was fixed in position throughout the engine run, as opposed to a quick-insertion-type sampler using a fastresponse actuator in open-flame burners.35 This approach was

Figure 1. Cross-sectional sketch of the diesel engine (top) and the close-up view of the soot sampling region (bottom).

summarized in Table 1. The investigation was carried out using an optically accessible, single-cylinder, automotive-size diesel engine with 83 mm bore and 92 mm stroke resulting in 498 cm3 of the displacement volume. The soot sampling probe holding a TEM grid at the tip was installed on the cylinder liner by replacing one of four quartz windows with a metal dummy window used as an adaptor. During the setup, an issue was raised due to the potential collision between the sampling probe and fast-moving piston or intake/exhaust valves. 8244

dx.doi.org/10.1021/es500661w | Environ. Sci. Technol. 2014, 48, 8243−8250

Environmental Science & Technology

Article

due to the lack of sealing methods that can handle high incylinder pressure over 4 MPa. Therefore, the samples in this study do not show soot particles at an instant moment but provide time-integrated information as the soot-laden diesel flame quickly passes through the TEM grid. Shown in Figures S1−S6 (Supporting Information, SI) are the temporal evolutions of soot luminosity during combustion event for all tested fuel injection conditions. Since the soot sampling system was absent during the video recording, its location was later drawn on the images to illustrate the expose of hot soot particles to the TEM grid. It can be seen from the soot luminosity images that the exposure time of the TEM grid to the sooting flame varied between 1 and 3 ms. In fact, the higher sooting condition had the longer exposure time, which was the reflection of the level of soot concentration in the target flame. It was noted that the soot exposure time of this study was comparable to the sampling studies in the open flame burners.35 TEM Imaging and Image Post-Processing. Soot particle imaging was performed for samples using a TEM (JEOL 1400) with a point resolution of 0.38 nm and an accelerating voltage of 100 kV. A Gatan CCD camera with a resolution of 11 mega pixels was used to digitize the magnified soot particle images. For each TEM grid, seven different on-grid locations were imaged considering the spatial fluctuations of sampled soot particles due to the inhomogeneous nature of diesel flame. The TEM magnification was set at ×50 000 considering both the number of particles per image and the image quality for postprocessing. Obtained TEM images were processed using an in-house-developed Matlab code,32 seven images were processed for each fuel injection condition. The postprocessing yielded the number, projection area of soot particles per image, primary particle diameter, and aggregate radius of gyration. These data were used to calculate the overall fractal dimension of soot aggregates. More information about the image processing and data reduction is found in our previous works.29,32,34 A brief summary of the soot geometry and fractal morphology characterization is also presented in the SI. Earlier, we found that a sufficient number of particles must be processed to obtain statistically meaningful data.31 In the present study, 105 to 1505 soot aggregates and 1217 to 15 579 primary particles were processed for each TEM grid, depending on the fuel injection condition.

Figure 2. Effect of injection timing and injection pressure on traces of average in-cylinder pressure and apparent heat release rate (aHRR) for various crank angle degrees after the top dead center (oCA aTDC).

mixing prior to the ignition (i.e., precombustion mixing) that significantly affects the soot formation was similar. Figure 2 also shows when the injection pressure increases from 70 to 160 MPa, a significant increase in the peak incylinder pressure and decrease in the ignition delay (∼2.5 °CA) occurred. Since the injected fuel mass and injection timing were fixed, the reduction of ignition delay and the increase in the peak in-cylinder pressure was a result of the increased precombustion mixing associated with the enhanced air entrainment and fuel atomization at the elevated injection pressures. Similarly, the aHRR appears to have narrower curves and higher peaks for high injection pressure conditions, suggesting more rapid and shorter combustion and thus higher flame temperature.24 These trends are very important as the enhanced air−fuel mixing would suppress the soot formation. An interesting trend in Figure 2 is that the change in peak incylinder pressure and ignition delay becomes smaller when the injection pressure is higher than 130 MPa. This phenomenon was also reported by Roy et al.37 due to potential overleaning that limits the effect of high injection pressure on combustion. The in-cylinder pressure and aHRR trends with the variations of fuel injection timing and pressure would be mentioned again in the following sections to discuss how the changes in combustion conditions affect the soot particles. TEM Images. Example TEM images of collected soot samples for various injection timings (a∼c) and injection pressures (b,d∼f) are shown in Figure 3. Shown at the bottomleft corner of each image is a scale bar of 200 nm. From the visual inspection, it is noted that the majority of soot particles are agglomerates of many small primary particles that are in almost spherical shape, which is not much different to those sampled from the diesel exhaust.20−25 However, the existence of single-primary particles and very small aggregates with only a



RESULTS AND DISCUSSION Combustion Conditions. The in-cylinder pressure traces and the corresponding apparent heat release rates (aHRR)36 for all tested fuel injection conditions are shown in Figure 2. These pressure traces were recorded simultaneously with the soot sampling experiments. The start of combustion was defined as the crank angle location at which the in-cylinder pressure exceeds the motored pressure. It is annotated for all cases. Figure 2 shows that both the in-cylinder pressure and aHRR decrease when the injection timing is retarded from −12 to −2 o CA aTDC. The retarded combustion phasing also made the aHRR curve flatter, suggesting slower combustion and hence lower flame temperature.16,24 In other words, the retarded injection caused the combustion to occur later after TDC where the ambient gas pressure and temperature were lower due to the expansion cooling.17 However, the change in ignition delay was measured to be less than 1 °CA for all injection timings tested in this study, meaning the fuel−air 8245

dx.doi.org/10.1021/es500661w | Environ. Sci. Technol. 2014, 48, 8243−8250

Environmental Science & Technology

Article

Figure 3. Selected TEM images of in-flame soot particles for various injection timings and injection pressures. Shown at the bottom-left corner of each image is a 200 nm scale bar.

few primary particles are evident. Such single-primary particles and small aggregates were also found to be quite popular for inflame soot sampled from constant volume chamber but not commonly observed in previous exhaust soot works, which might suggest these small particles are either completely oxidized or combined into bigger aggregates before they flow out of the cylinder. From Figure 3, a notable decrease in the amount of soot particles is seen for later injection timings, which is expected considering the decreasing in-cylinder pressure and aHRR (see Figure 2). The decrease is most pronounced for −2 °CA aTDC injection case (Figure 3c) such that only a handful of soot particles are visible. It is also noticeable that some soot aggregates for the early injection timings of −12 and −7 °CA aTDC (Figure 3a,b) are even larger than the scale bar (200 nm), while such large aggregates cannot be found in −2 °CA aTDC injection case. Expected trends are also observed for the injection pressure variations: (1) the amount of soot particles is lower for the higher injection pressure and (2) the overall size of soot particles at lower injection pressure is significantly larger than those at higher injection pressure. Soot aggregates that are larger than 200 nm (i.e., the size of the scale bar) are easily observed in the image for 70 MPa (Figure 3b) but none of them are seen for 160 MPa (Figure 3f). Also, the morphology of soot aggregates differs as the injection pressure increases such that most of the soot aggregates for 70 MPa case appear to be large aggregates in chain-like or grape-like structures consisting of many primary particles. By contrast, the majority of soot particles at 160 MPa are very small aggregates or singleprimary particles. Projection Area and Number of Soot Particles. By postprocessing TEM images, quantitative data that can clarify the soot particle characteristics for various injections timings and pressures were obtained. The results shown in Figure 4 are

the soot projection area and the number of soot particles per TEM image. As mentioned previously, multiple areas on the TEM grid were imaged. The gray circles illustrate this on-grid variation at each injection timing or pressure condition. Those corresponding to the selected TEM images in Figure 3 (for the presentation purposes) are marked by red circles. The mean values and error ranges are also annotated by red vertical lines. These spatial variations could be likely due to the nonuniform deposition of soot particles across the grid surface.31 While these variations are unavoidable, there are inevitable trends associated with injection condition variations. For example, both the projection area and the number of soot particles show a continuous decrease for the retarded injection timing. Previously, Fang et.al.16 reported that the intensity of flame luminosity was substantially lower when the fuel injection timing was retarded after TDC, suggesting the lower in-flame soot level. Upon the fact that the high soot formation occurs in fuel rich regions where the flame temperature is high (>1400 K), the reduced flame temperature associated with the decreased ambient temperature during the expansion stroke would suppress the soot formation.38 The decreased flame temperature should be a primary cause for the reduced soot particles because the variations in ignition delays and hence the precombustion mixing times were very minimal for the tested injection timing range. However, it should be noted that the opposite trend was observed in the exhaust soot measurements that the exhaust soot level could increase for the retarded injection timing.6,7,10 This is explained by the trade-off between the soot formation and oxidation. That is, the high flame temperature of the advanced injection timing that caused high soot formation might contribute to the high soot oxidation rate as well, which results in lower exhaust soot particles than the retarded injection case with lower soot formation but lower soot oxidation.13,17 Therefore, the value of our in-flame samples is found in the understanding of soot particles at a formation 8246

dx.doi.org/10.1021/es500661w | Environ. Sci. Technol. 2014, 48, 8243−8250

Environmental Science & Technology

Article

not be as high as 70 MPa due to better mixing, leading to the lower number of soot primary particles. However, the dominant effect of the limited residence time would prevent those primary particles to grow to aggregates in various sizes and shapes through the coalescence and agglomeration, which caused the increased number of particles (aggregates) as shown in Figure 4. At the highest injection pressure of 160 MPa case, however, further reduction in the particle inception eventually outperformed the reduced residence time, which resulted in the lowest number of particles in all injection pressures tested in the present study. Soot Particle Size Distribution. Due to the variations in sample size among tested conditions, the probability density function (pdf) is used instead of the number distribution (histogram). The pdf of radius of gyration of soot aggregates (Rg) and primary particle diameter (dp) for all conditions are plotted in Figure 5. The mean values and error ranges are also annotated for each fuel injection condition.

Figure 4. Soot projection area (top) and the number of soot particles (bottom) per each TEM image as functions of injection pressure and injection timing. Each circle corresponds to individual image of a selected on-grid location. The error ranges for this on-grid variation were estimated at 95% confidence.

stage before they undergo significant oxidation, which cannot be studied using the exhaust soot samples. With the increasing injection pressure, Figure 4 shows a continuous decrease in the soot projection area, suggesting lower soot formation. This is consistent with previous studies on the in-cylinder soot concentrations,13,16 where the increased injection pressure was found to promote the fuel−air mixing and shorten the residence time of the soot growth on account of the increased jet velocity and momentum. While the decreased soot amount seen in the projection area is simply consistent with the previous works, it is important to note that the reduction in the projection area can be the result of either reduced particle numbers or decreased particle size. For example, the reduction in the particle size with increasing injection pressure is evident in Figure 3. However, the number of particles shows a nonlinear trend such that between 100 and 160 MPa, the number count decreases, whereas it increases between 70 and 100 MPa. This trend illustrates the complex process of soot formation, which varies with the injection pressure. In other words, the soot inception, surface growth, and agglomeration would all be suppressed by the high injection pressure but their significance could be different. It is likely that between 100 and 130 MPa, the soot inception would

Figure 5. Probability density functions of radius of gyration (Rg) of soot aggregates (top) and primary particle diameter (dp) for various fuel injection pressures and timings. The mean value and error range (95% confidence) are annotated.

Figure 5 (top) shows that the mean Rg is around 35 nm for both −12 and −7 °CA aTDC injection timings at a fixed injection pressure of 70 MPa. When the injection is further retarded to −2 °CA aTDC, however, it decreases to 30.1 nm. More importantly, the likelihood of large soot aggregates (e.g., Rg > 80 nm) decreases drastically at this latest injection timing while the likelihood of small soot aggregates (e.g., Rg < 50 nm) increases. Unlike the Rg variations, dp does not change with the injection timing, not only in the mean value, but also the pdf curves. As shown in Figure 2, the ignition delay and thus 8247

dx.doi.org/10.1021/es500661w | Environ. Sci. Technol. 2014, 48, 8243−8250

Environmental Science & Technology

Article

where N is the number of primary particles within a soot aggregate and d p is the mean diameter of primary particles within the aggregate. From this equation, the fractal dimension Df represents the structural compactness of a fractal aggregate. For instance, Df is of 1 corresponds to a stretched line structure, while a very agglomerated aggregate would have a Df of 3. In eq 1, the fractal prefactor kf is also shown. It is known to help understand the optical and transport properties of soot particles.39 As described by eq 1, Df and kf can be obtained by reading the slop and the intercept of the least-squares straight line of a plot ln(N) over ln(Rg/ d p). However, N must be determined

precombustion mixing times are the same for all three injection timings, suggesting the flame temperature has little impact on the primary particle growth in our tested conditions. Also, the constant dp indicates that the reduction of Rg was due to the reduced soot agglomeration. In other words, the nucleated primary particles and small soot aggregates had low chances to further agglomerate into larger aggregates. Mathis et al.24 reported a different trend in the exhaust soot such that the primary particles were smaller for more advanced injection timings. This might mean that the in-flame primary particles with nearly identical size shrunk differently during the oxidation process such that the high flame temperature of the advanced injection timing resulted in smaller primary particles in the exhaust. This is consistent with the projection area and the number of soot particles trends discussed in the previous section (see Figure 4). Figure 5 also shows that with the increasing injection pressure from 70 to 160 MPa, the mean Rg and dp decrease from 35.1 to 15.4 nm and 18.23 to 11.98 nm, respectively. Moreover, the size distributions become very narrow while shifting to the smaller size range. Indeed, the pdf shows there are no soot aggregates with Rg higher than 60 nm at 100 MPa or higher. The smaller primary particles and soot aggregates suggested that both the particle growth and agglomeration were suppressed. Previous studies suggested that the air/fuel ratio and particle residence time make significant impact on soot primary particle inception and growth.26 The increased injection pressure would increase local air/fuel ratio (i.e., increased lean mixture) and reduce the particle residence time.13 Therefore, the reduction in particle size is likely due to the enhanced mixing and reduced particle residence time at high injection pressure conditions. It should be noted that the mean Rg range of 15 to 38 nm and dp range of 12 to 18 nm measured in this study are much smaller than those measured from the heavy-duty engine exhaust.23 Even from a light-duty diesel engine20 that is similar to our engine, the Rg is at a much higher range of 77.4 to 134.1 nm. Since the exhaust soot particles after the significant oxidation should be smaller than the in-flame soot particles, the 3-fold higher Rg of the exhaust soot is unexpected. There are three possible explanations for this difference. First, the exhaust gas recirculation was implemented in the exhaust soot study,20 which potentially increased the soot formation and resulted in larger exhaust soot particles. Second, the engine used in the current study was modified for the in-flame soot sampling such that a portion of the bowl-rim was cut-out and thus the compression ratio was reduced, resulting in lower flame temperature and hence lower soot formation. The use of a single-hole nozzle would also reduce the soot formation due to the lack of jet-jet interaction. Third, it could be a simple data presentation issue. In our earlier work,31 about 40 nm of Rg was obtained when the mass-based distribution was used at the high injection pressure of 150 MPa, which is roughly 3-fold higher. In this study, such mass-based distribution was not used in order to avoid the assumption of soot density that is not available at the tested condition. Fractal Morphology of Soot Aggregates. The soot aggregates in Figure 3 showed that fractal morphology that can be characterized by the mass fractal relation as shown in eq 1,

⎛ R g ⎞ Df N = k f ⎜⎜ ⎟⎟ ⎝ dp ⎠

together with Rg and d p in prior (see SI Figure S7). Several studies have suggested that N can be calculated from the following equation:

⎛ A ⎞α N = ka⎜⎜ a ⎟⎟ ⎝ Ap ⎠

(2)

where Aa is the projection area of a soot aggregate, and Ap is the mean projection area of primary particles within the aggregate. Previous studies suggested that the empirical constant ka varies from 1 to 1.81 while α is in the range of 1.08 to 1.19 depending on the degree of primary particle overlapping.40−43 In the present study, we tested suggested values of ka and α. However, it was found that the trend of fractal dimension and prefactor did not change (see SI Figure S8). Figure 6 (top and middle) illustrates Df and kf for all conditions tested in this study together with the error ranges (95% confidence). For these plots, ka and α were selected for 1 and 1.09, as proposed by Megaridis and Dobbins (1990).41 It should be note that the fractal dimension was estimated only for the soot aggregates with more than three primary particles (N > 3) because monomers (N = 1) are not aggregates and the aggregates with only two or three primary particles (N = 2 or 3) hardly have fractal structures. From the Figure, it was observed that the fractal dimension range of this study falls into the cluster−cluster aggregation regime (Df < 2, see ref 41, and references therein), similar to previous works conducted in diesel engines.20,27 In Figure 6, it is seen that Df decreases drastically when the injection timing is retarded, suggesting those aggregates with smaller quantity (Figure 4) and sizes (Figure 5) for the later injection timing became more stretched. This is consistent with the TEM images (see Figure 3) and our earlier implication of the suppressed agglomeration. Similarly, Df decreases with the increasing injection pressure, suggesting that the soot aggregates found at low-sooting conditions would be more stretched than those at high-sooting conditions. It is because at low sooting condition, particle inception is suppressed due to either reduced temperature (reduced reactivity) or increased air/fuel mixing. In other words, the decreased number of primary particles and limited residence time would reduce the chance of particles being reformed and agglomerated into a compact structure. One recent study44 reported the same finding that Df of exhaust soot particles decreases with decreasing engine power and there by lower soot levels. From Figure 6 (top), it is noticeable that the Df of 130 MPa injection case is clearly off the trend. This was unexpected because the TEM images in Figure 3 showed lower number of large aggregates at 130 MPa than those at 100 MPa. To clarify this, Df for various N cutoff values are plotted in Figure 6

(1) 8248

dx.doi.org/10.1021/es500661w | Environ. Sci. Technol. 2014, 48, 8243−8250

Environmental Science & Technology

Article

primary particles largely contribute to the increased Df that is even higher than the 100 MPa case. It was noted that Df in the current study is slightly lower than that found in the exhaust soot particle of heavy-duty diesel engines. For example, one study25 suggested that Df of the exhaust soot particles is in the range of 1.8 to1.88. This mismatch was due to the different displacement volume of the engines. Indeed, the exhaust soot particles in a light-duty diesel engine showed a fractal dimension of 1.52 to 1.73,46 which is consistent with the current study.



ASSOCIATED CONTENT

S Supporting Information *

Figures S1−S6 show the temporal evolutions of soot luminosity during combustion event for all tested fuel injection conditions. Presumable locations of soot sampling system are overlaid on the images. Figures S7 and S8 describe the characterization of soot geometry and fractal morphological properties, which are applied in the current study. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Phone: +63 425091230; fax: +61 (2) 9663 1222; e-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Experiments were performed at the UNSW Engine Research Laboratory. Support for this research was provided by the Australian Research Council.



REFERENCES

(1) Lighty, J. S.; Veranth, J. M.; Sarofim, A. F. Combustion aerosols: Factors governing their size and composition and implications to human health. J. Air Waste Manage. Assoc. 2000, 50, 1565−1618. (2) Muzyka, V.; Veimer, S.; Schmidt, N. On the carcinogenic risk evaluation of diesel exhaust: Benzene in airborne particles and alterations of heme metabolism in lymphocytes as markers of exposure. Sci. Total Environ. 1998, 217 (1−2), 103−111. (3) Broday, D. M.; Rosenzweig, R. Deposition of fractal-like soot aggregates in the human respiratory tract. J. Aerosol Sci. 2011, 42 (6), 372−386. (4) Benbrahim-Tallaa, L.; Baan, R. A.; Grosse, Y.; Lauby-Secretan, B.; El Ghissassi, F.; Bouvard, V.; Guha, N.; Loomis, D.; Straif, K.; Arlt, V. M. Carcinogenicity of diesel-engine and gasoline-engine exhausts and some nitroarenes. Lancet Oncol. 2012, 13 (Article), 663−664. (5) Zhang, R.; Khalizov, A. F.; Pagels, J.; Zhang, D.; Xue, H.; McMurry, P. H. Variability in morphology, hygroscopicity, and optical properties of soot aerosols during atmospheric processing. Proc. Natl. Acad. Sci. U.S.A. 2008, 105 (30), 10291−10296. (6) Agarwal, A. K.; Dhar, A.; Srivastava, D. K.; Maurya, R. K.; Singh, A. P. Effect of fuel injection pressure on diesel particulate size and number distribution in a CRDI single cylinder research engine. Fuel 2013, 107 (0), 84−89. (7) Sayin, C.; Ilhan, M.; Canakci, M.; Gumus, M. Effect of injection timing on the exhaust emissions of a diesel engine using diesel− methanol blends. Renew. Energy 2009, 34 (5), 1261−1269. (8) Dodge, L. G.; Simescu, S.; Neely, G. D.; Maymar, M. J.; Dickey, D. W.; Savonen, C. L. Effect of small holes and high injection pressures on diesel engine combustion. SAE Technical Paper 2002-01-0494, 2002. (9) Niemi, S. A.; Paanu, T. P. J.; Laurén, M. J., Effect of Injection Timing, EGR and EGR Cooling on the exhaust particle number and

Figure 6. Fractal dimension (Df) and prefactor (kf) for various injection pressures and timings (top and middle), and Df over the injection pressure for various n cutoff values (bottom).

(bottom). There is a clear trend in the figure such that Df decreases with the increasing N cutoff value for all injection pressures tested in this study. This suggests that soot aggregates consisted of lower number of primary particles tend to have higher Df. Meakin (1984)45 suggested that young soot aggregate consist of small number of primary particles show high fractal dimension (e.g., Df ≈ 3) as they follow a monomercluster aggregation mechanism. This then explains the observed trend of higher Df for 130 MPa case compared to 100 MPa case. When the 130 MPa case is closely looked at, it is observed that the high population of soot aggregates with less than 6 8249

dx.doi.org/10.1021/es500661w | Environ. Sci. Technol. 2014, 48, 8243−8250

Environmental Science & Technology

Article

size distribution of an off-road diesel engine. SAE Technical Paper 2004-01-1988, 2004. (10) Zhu, R.; Miao, H.; Wang, X.; Huang, Z. Effects of fuel constituents and injection timing on combustion and emission characteristics of a compression-ignition engine fueled with dieselDMM blends. Proc. Combust. Inst. 2013, 34 (2), 3013−3020. (11) Wang, X.; Kuti, O. A.; Zhang, W.; Nishida, K.; Huang, Z. Effect of injection pressure on flame and soot characteristics of the biodiesel fuel spray. Combust. Sci. Technol. 2010, 182 (10), 1369−1390. (12) Ryser, R.; Gerber, T.; Dreier, T. Soot particle sizing during highpressure Diesel spray combustion via time-resolved laser-induced incandescence. Combust. Flame 2009, 156 (1), 120−129. (13) Pickett, L. M.; Siebers, D. L. Soot in diesel fuel jets: Effects of ambient temperature, ambient density, and injection pressure. Combust. Flame 2004, 138 (1−2), 114−135. (14) Menkiel, B.; Donkerbroek, A.; Uitz, R.; Cracknell, R.; Ganippa, L. Measurement of in-cylinder soot particles and their distribution in an optical HSDI diesel engine using time resolved laser induced incandescence (TR-LII). Combust. Flame 2012, 159 (9), 2985−2998. (15) Inagaki, K.; Takasu, S.; Nakakita, K.; Watanabe, S., Quantitative analysis of soot formation and oxidation process using laser-induced incandescence. SAE Technical Paper 2003-01-1795, 2003. (16) Fang, T.; Coverdill, R. E.; Lee, C.-f. F.; White, R. A. Air−fuel mixing and combustion in a small-bore direct injection optically accessible diesel engine using a retarded single injection strategy. Fuel 2009, 88 (11), 2074−2082. (17) Dec, J. E.; Kelly-Zion, P. L. The effects of injection timing and diluent addition on late-combustion soot burnout in a di diesel engine based on simultaneous 2-D imaging of OH and soot. SAE Technical Paper 2000-01-0238, 2000. (18) Bagley, S. T.; Baumgard, K. J.; Gratz, L. D.; Johnson, J. J.; Leddy, D. G. Characterization of Fuel and after-Treatment Device Effects on Diesel Emissions; Health Effects Institute: 1996; http://pubs. healtheffects.org/getfile.php?u=826. (19) Su, D. S.; Serafino, A.; Müller, J.-O.; Jentoft, R. E.; Schlögl, R.; Fiorito, S. Cytotoxicity and inflammatory potential of soot particles of low-emission diesel engines. Environ. Sci. Technol. 2008, 42 (5), 1761− 1765. (20) Zhu, J.; Lee, K. O.; Yozgatligil, A.; Choi, M. Y. Effects of engine operating conditions on morphology, microstructure, and fractal geometry of light-duty diesel engine particulates. Proc. Combust. Inst. 2005, 30 (2), 2781−2789. (21) Shi, J. P.; Mark, D.; Harrison, R. M. Characterization of particles from a current technology heavy-duty diesel engine. Environ. Sci. Technol. 2000, 34 (5), 748−755. (22) Park, K.; Kittelson, D. B.; McMurry, P. H. Structural properties of diesel exhaust particles measured by transmission electron microscopy (TEM): Relationships to particle mass and mobility. Aerosol Sci. Technol. 2004, 38 (9), 881−889. (23) Neer, A.; Koylu, U. O. Effect of operating conditions on the size, morphology, and concentration of submicrometer particulates emitted from a diesel engine. Combust. Flame 2006, 146 (1−2), 142−154. (24) Mathis, U.; Mohr, M.; Kaegi, R.; Bertola, A.; Boulouchos, K. Influence of diesel engine combustion parameters on primary soot particle diameter. Environ. Sci. Technol. 2005, 39 (6), 1887−1892. (25) Lee, K. O.; Cole, R.; Sekar, R.; Choi, M. Y.; Kang, J. S.; Bae, C. S.; Shin, H. D. Morphological investigation of the microstructure, dimensions, and fractal geometry of diesel particulates. Proc. Combust. Inst. 2002, 29 (1), 647−653. (26) Lapuerta, M.; Martos, F. J.; Herreros, J. M. Effect of engine operating conditions on the size of primary particles composing diesel soot agglomerates. J. Aerosol Sci. 2007, 38 (4), 455−466. (27) Li, Z.; Song, C.; Song, J.; Lv, G.; Dong, S.; Zhao, Z. Evolution of the nanostructure, fractal dimension and size of in-cylinder soot during diesel combustion process. Combust. Flame 2011, 158 (8), 1624− 1630. (28) Yamaguchi, T.; Kondo, K.; Nishigai, H.; Takano, S.; Aizawa, T. Direct Sampling, TEM analysis and optical measurement of soot

particles at different axial locations in a transient spray flame. SAE Int. J. Fuels Lubr. 2011, 5 (1), 316−328. (29) Kook, S.; Pickett, L. M. Soot volume fraction and morphology of conventional, fischer-tropsch, coal-derived, and surrogate fuel at diesel conditions. SAE Int. J. Fuels Lubr. 2012, 5 (2), 647−664. (30) Kondo, K.; Yamaguchi, T.; Nishigai, H.; Takano, S.; Aizawa, T., High-resolution transmission electron microscopy of soot directly sampled at different axial locations in diesel spray flame. SAE Technical Paper 2011-24-0068, 2011. (31) Kondo, K.; Aizawa, T.; Kook, S.; Pickett, L. Uncertainty in sampling and tem analysis of soot particles in diesel spray flame. SAE Technical Paper 2013-01-0908, 2013. (32) Aizawa, T.; Nishigai, H.; Kondo, K.; Yamaguchi, T.; Nerva, J.G.; Genzale, C.; Kook, S.; Pickett, L. Transmission electron microscopy of soot particles directly sampled in diesel spray flame A comparison between US#2 and biodiesel soot. SAE Int. J. Fuels Lubr. 2012, 5 (2), 665−673. (33) Kuribayashi, M.; Ishizuka, Y.; Aizawa, T. Sizing of soot particles in diesel spray flameA qualitative comparison between TEM analysis and LII/scattering laser measurements. SAE Int. J. Fuels Lubr. 2013, 6 (3), 641−650. (34) Kook, S.; Zhang, R.; Szeto, K.; Pickett, L. M.; Aizawa, T. Inflame soot sampling and particle analysis in a diesel engine. SAE Int. J. Fuels Lubr. 2013, 2013−01−0912. (35) Dobbins, R. A.; Megaridis, C. M. Morphology of flamegenerated soot as determined by thermophoretic sampling. Langmuir 1987, 3 (2), 254−259. (36) Heywood, J. B. Internal Combustion Engine Fundamentals.; McGraw-Hill Book Co.: New York, 1988. (37) Roy, M. M.; Tsunemoto, H., Effect of injection pressure and split injection on exhaust odor and engine noise in DI diesel engines. SAE Technical Paper 2002-01-2874, 2002. (38) Glassman, I. Soot formation in combustion processes. Symp. (Int.) Combust. 1989, 22 (1), 295−311. (39) Sorensen, C. M.; Roberts, G. C. The prefactor of fractal aggregates. J. Colloid Interface Sci. 1997, 186 (2), 447−452. (40) Brasil, A. M.; Farias, T. L.; Carvalho, M. G. A recipe for image characterization of fractal-like aggregates. J. Aerosol Sci. 1999, 30 (10), 1379−1389. (41) Megaridis, C. M.; Dobbins, R. A. Morphological description of flame-generated materials. Combust. Sci. Technol. 1990, 71 (1−3), 95− 109. (42) Koeylue, U.; Xing, Y.; Rosner, D. E. Fractal morphology analysis of combustion-generated aggregates using angular light scattering and electron microscope images. Langmuir 1995, 11 (12), 4848−4854. (43) Oh, C.; Sorensen, C. M. The effect of overlap between monomers on the determination of fractal cluster morphology. J. Colloid Interface Sci. 1997, 193 (1), 17−25. (44) China, S.; Salvadori, N.; Mazzoleni, C. Effect of traffic and driving characteristics on morphology of atmospheric soot particles at freeway on-ramps. Environ. Sci. Technol. 2014, 48 (6), 3128−3135. (45) Meakin, P. Computer simulation of cluster-cluster aggregation using linear trajectories: Results from three-dimensional simulations and a comparison with aggregates formed using Brownian trajectories. J. Colloid Interface Sci. 1984, 102 (2), 505−512. (46) Lee, K. O.; Zhu, J. Evolution in size and morphology of diesel particulates along the exhaust system. SAE Technical Paper 2004-011981, 2004.

8250

dx.doi.org/10.1021/es500661w | Environ. Sci. Technol. 2014, 48, 8243−8250