Aerosol and Air Quality Research, 16: 2560–2569, 2016 Copyright © Taiwan Association for Aerosol Research ISSN: 1680-8584 print / 2071-1409 online doi: 10.4209/aaqr.2016.07.0329 Pyrolysis Characteristic Analysis of Particulate Matter from Diesel Engine Run on Diesel/Polyoxymethylene Dimethyl Ethers Blends Based on Nanostructure and Thermogravimetry Hao Yang, Xinghu Li, Yan Wang*, Mingfei Mu, Xuehao Li, Guiyue Kou School of Transportation Science and Engineering, Beihang University, Beijing 100191, China ABSTRACT

This paper focuses on the nanostructural and pyrolysis characteristics of particulate matter (PM) emitted from a diesel engine fueled by three diesel/polyoxymethylene dimethyl ethers (PODEn) blends. PM was collected using a metal filter from the exhaust manifold. The collected PM samples were characterized using scanning electron microscopy (SEM), transmission electron microscopy (TEM) and thermogravimetric analysis (TGA). SEM and TEM analysis showed that PM generated by the 20%-PODEn blend was looser and had a smaller average diameter than that emitted when a lower proportion of PODEn was used. Additionally, fringe length was reduced, and separation distance and tortuosity were increased, when the 20%-PODEn blend was used instead of the other blends. These changes improved the oxidation reactivity of the PM. TGA demonstrated that the PM pyrolysis process was divided into low temperature (characterized by moisture and volatile components) and high temperature (the combustion of solid carbon) stages. When the blending ratio was increased, the moisture and volatile components of the PM showed no obvious change, but the ignition temperature and activation energy were reduced. Additionally, the pyrolysis performance was enhanced; the maximum weight loss rate was higher; the combustion and burnout characteristic indices were higher; and the combustion efficiency of the PM was improved. These results show that the use of diesel blended with oxygenated fuel (PODEn) affects the nanostructure and pyrolysis of PM, and this PM is easier to oxidize and advantageous for diesel particulate filter regeneration. Keywords: Polyoxymethylene dimethyl ethers; Particulate matter; Nanostructure; Thermogravimetry; Pyrolysis. INTRODUCTION

Diesel engines are widely used in the transportation, industry, commerce, and power generation sectors. Diesel engines have the advantages including high efficiency, a high level of fuel economy with great torque output, high durability and reliability, and low operating costs. Diesel engines are major source of emissions including particulate matter (PM), nitrogen oxides, sulfur oxides, total hydrocarbons, carbon dioxide, carbon monoxide as well as toxic chemicals such as polycyclic aromatic hydrocarbons (Mwangi et al., 2015; Tsai et al., 2015). However, it is well known that PM emissions from diesel engines are are relatively high, due to the combustion of heterogeneous mixture in the cylinder at higher combustion temperatures (Saxena and Maurya, 2016). PM, also known as soot, is primarily composed of carbon but also includes some other organic and inorganic compounds (ash), some sulfur * Corresponding author.

Tel.: 86-10-82316891; Fax: 86-10-82316330 E-mail address: wybuaa@buaa.edu.cn

compounds and some metal traces from unburnt fuel and lubricating oil (Li, 2011; Sharma et al., 2012). But the composition of PM from diesel engines may vary widely, because it depends on lots of variables such as operating conditions, fuel composition, lubricating oil type etc. (Popovicheva et al., 2014; Shukla et al., 2014).

PM is microscopic with more than 90% of PM smaller than 1 µm. PM enters the human body easily through the respiratory system and can induce a variety of diseases (Li et al., 2003; Gonzalez-Flecha, 2004; Kaiser, 2005). It has been a concern issue in many countries owing to its adverse health effects (Tsai et al., 2016). Consequently, diesel PM emissions are strictly monitored around the world. In China, fifth phase of an emission standard formulated to implement environmental protection law “Limits and measurement methods for emissions from light-duty vehicles” limits the PM concentration for diesel vehicle emission to 0.0045 g km–1, representing an 82% reduction (0.025 g km–1) compared with the fourth phase.

Two possible methods to reduce PM emissions from diesel engines are to use oxygenated fuel and a diesel particulate filter (DPF) (Cheruyiot et al., 2015). As a new type of oxygenated fuel, polyoxymethylene dimethyl ethers (PODEn) are effective at reducing diesel PM emissions because

Yang et al., Aerosol and Air Quality Research, 16: 2560–2569, 2016 2561

PODEn have no C–C bonds in their molecular structure, a substantial carbon/hydrogen ratio and high oxygen content with a large cetane number (Burger et al., 2010; Zhao et al., 2011; Zhao et al., 2013; Yang et al., 2015). PODEn can be synthesized through a polymerization reaction using methanol as a raw material, and many crude materials can be used to produce them, including coal and biomass. Wall flow DPFs undergo a regeneration process to remove PM and are one of the most common technologies used in diesel vehicles to meet emissions standards in China, the United States, Europe and elsewhere.

Scanning electron microscopy (SEM), transmission electron microscopy (TEM) and thermogravimetric analysis (TGA) are used to estimate PM emissions. SEM has been widely applied in the observation of PM morphology, whereas TEM has been routinely used to investigate the microstructures of diesel PM. Ishiguro et al. (1997) reported that primary PM had two distinct parts: an inner core and an outer shell. The inner core had a central region with a diameter of 10 nm and consisted of several fine particles of 3–4 nm in diameter. The nucleus exhibited several distorted structural carbon layers. The outer shell was composed of microcrystallites with the periodic orientation of carbon sheets. Almost all of the crystallites were oriented perpendicular to the radius of the primary PM with a thickness of 1 nm and width of 3.5 nm. Vander Wal et al. (2007) demonstrated that some primary soot particles had a hollow interior and an outer shell exhibiting evidence of graphitization, with a more crystalline structure than that of nonhollow particles. Su et al. (2004) reported that PM emitted from a Euro IV heavy-duty diesel engine consisted of fullerenoid- or onion-like particles agglomerated in a chain-like secondary structure and They also observed that PM with different microstructures had different oxidation reactivities.

TGA is the most widely employed method of thermal analysis and involves monitoring the weight loss of specimens as they are heated at a controlled rate. TGA is used in the analysis of PM stability, kinetic parameters and volatile component ratios. Sampling is performed in the presence of an inert (He, Ar, or N2) or oxygenated environment. The main characteristic of TGA is its ability to measure mass and the PM rate of change (Lapuerta et al., 2007; Chuepeng et al., 2008). Karin et al. (2015) observed that the chemical contents of PM could be divided into three main regions according to oxidation temperature; moisture was vaporized at low temperatures (25–200°C), unburnt hydrocarbons were oxidized at temperatures of 200–500°C and finally, carbon inside the PM was oxidized at temperatures between 500–600°C. Yehliu et al. (2012) used TGA to show that PM derived from pure biodiesel had the highest oxidation rate compared with ultra-low-sulfur diesel and synthetic Fischer-Tropsch fuel. Sukjit et al. (2013) demonstrated that the oxidation temperature of PM decreased when rapeseed biodiesel (RME) or oxygenated additives (butanol and castor oil methyl ester) were used. Gill et al. (2012) reported the similar findings in their comparison of the oxidation reactivity of PM generated from conventional diesel fuel, RME-diesel and diglyme-diesel blends.

At present, few studies on PM emitted from diesel

engines fueled by diesel/PODEn blends have been reported. In this study, the nanostructures and pyrolysis of PM derived from PODEn/diesel blends were investigated using SEM, TEM, and TGA. PODEn and diesel in volume ratios of 0/100, 10/90 and 20/80 were used, with blends denoted P0, P10, and P20 respectively. The purpose was to investigate the effect of the PODEn/diesel blend ratio on PM emissions, and this paper provides a theoretical basis on which to select appropriate regeneration techniques and conditions. EXPERIMENTAL METHOD PM Sampling Procedure

Experiments were performed using a single-cylinder four-stroke R180 diesel engine produced by Changchai Co., Ltd. The main engine specifications are listed in Table 1. Diesel used in the experiments was purchased locally and PODEn were provided by Shandong Yuhuang Chemical Co., Ltd. Diesel was used as the base fuel and PODEn were used as an oxygenating additive. Sampling was performed at 1800 rpm and under 100% operating load conditions of the diesel engine. Samples were collected from an exhaust pipe 1.2 m away from the engine exhaust manifold. The direct sampling method used a multilayer metal wire mesh filter, which was stacked and arranged in the exhaust pipe during the test. PM was filtered off by the metal filter to obtain similar PM to that obtained in real-world conditions with a DPF. PM derived from the various diesel blends was placed into sealed sample bottles for storage. In each experiment, the engine was warmed up for 20 minutes before PM was collected. Analytical Techniques Scanning Electron Microscopy

The morphology of the agglomerates was obtained using SEM. PM samples were covered with a gold film prior to SEM in order to meet the minimum electrical conductivity requirements. Each sample was viewed in secondary electron mode using a Hitachi SU8010 microscope operated at an accelerating voltage of 15 kV with a resolution of 1.00 nm. Particle size analysis was performed on SEM images using the commercial graphic processing software package Image-Pro Plus 6.0. The corresponding column distribution map was generated in this manner.

Transmission Electron Microscopy

To investigate the soot nanostructure, a Tecnai G2 20 was used to capture high-resolution bright field images. The

Table 1. Diesel engine main specifications. Parameter Value

Number of cylinders 1 Rated power / kW 5.67

Rated speed / (r min–1) 2600 Bore / mm × Stroke / mm 80 × 80

Displacement volume / mL 402 Compression ratio 21:1

Injection pressure / MPa 13.72 ± 0.5

Yang et al., Aerosol and Air Quality Research, 16: 2560–2569, 2016 2562

instrument was operated at 200 kV using a LaB6 filament. The applied magnification was up to 1,030,000× with a resolution of 0.24 nm. TEM samples were prepared as follows: a small amount of PM was ultrasonically dispersed in methanol for 30 minutes; five drops of the solution were then placed on a 200 mesh Lacey C/Cu TEM grid; the grid and sample were air-dried; and TEM analysis was performed.

The PM nanostructure was observed to be in the form of parallel or distorted carbon lamellae. Image-Pro Plus 6.0 software was employed for the analysis of high resolution TEM (HRTEM) images. Fringe length (La), fringe separation (Ds) and tortuosity (Tf) were used to describe the PM nanostructure (Vander Wal et al., 2005). La is a measure of the physical extent of the atomic carbon layer planes seen in the HRTEM image. Ds is the mean distance between adjacent carbon layer planes. Tf is a measure of the curvature of the fringes and is defined here as the ratio of the fringe length to the distance between the two endpoints. If La was less than 0.40 nm and Ds was greater than 0.45 nm or less than 0.30 nm, the fringe was discarded as an artifact.

Thermogravimetric Analysis

A STA-449 F3 simultaneous thermogravimetric analyzer from NETZSCH with a resolution of 1 µg was used for TGA. The heating rate of the apparatus was 0.1–50 °C min–1, and the maximum temperature was set at 1500°C. To simulate the pyrolysis and combustion that occur in an engine’s exhaust system, a mixture of oxidizer gas agents with 10% O2 in N2 from Airgas, and N2 as a purge gas was used. The flow rate was limited to 100 mL min–1. Specimens with an initial mass of 2 mg were heated from room temperature to 900°C at a rate of 10 °C min–1. An inert ceramic crucible was used to avoid the catalytic effects a

metal crucible has on the PM; therefore, controlled PM pyrolysis was achieved. RESULTS AND DISCUSSION SEM and TEM Image Analysis Morphology of PM Samples

The PM samples demonstrated a certain level of agglomeration, as illustrated by Figs. 1 and 2 (SEM at 100,000 times magnification and TEM, respectively). Numerous quasi-spherical primary particles of varying sizes under the action of Van der Waals, electrostatic and surface adhesion forces formed agglomerate particles through accumulation. Primary particles appear stacked together in the dark regions of the TEM images, which obscures the outlines of the primary particles. Additionally, different densities were observed. PM samples generated from the P0 blend were densely packed but when more PODEn was used in the blend, the packing was seen to become looser. Similarly to oxygenated fuel, PODEn/diesel blends combust more completely and reduce PM emissions. Therefore, the probability of collision between primary particles decreases and large agglomerates are not easily formed.

Size Distribution of PM Samples

According to Fig. 1, the size distribution of PM was determined using statistical analysis in a unit area, and the mean diameters of PMs (DP0, DP10 and DP20) were obtained (Fig. 3). Particle size had an approximately normal distribution over the range 20–60 nm. The size and the mean diameter of the PM decreased as blends with a higher proportion of PODEn were used, with the mean diameters being 44.23 nm for the P0 blend and 35.4 nm for the P20

(a) P0 (b) P10 (c) P20

Fig. 1. SEM images of PM samples.

(a) P0 (b) P10 (c) P20

Fig. 2. TEM images of PM samples.

Yang et al., Aerosol and Air Quality Research, 16: 2560–2569, 2016 2563

Fig. 3. Particle size distributions of PM samples.

blend, representing a decrease of 20%. This is mainly because the oxygen content and cetane number are higher in the blended fuel, which promotes more efficient fuel burning. Smaller and looser PM oxidizes more easily and is advantageous for DPF regeneration.

Nanostructure Parameters Analysis of PM Samples Because SEM images could not completely characterize

the nanostructure of the primary particles, HRTEM was used. HRTEM images were digitized and analyzed using image processing software to find the nanostructure parameters La, Ds and Tf. A region of the HRTEM image showing clear carbon layers was selected and converted into black and white. From this binary image, a skeletonizing process was used and each fringe was assigned a 1-pixel width for further statistical analysis. The original HRTEM images and final skeleton images are depicted in Fig. 4.

Fig. 5 presents histograms of the fringe length La for PM samples collected from the three blends, and it showed that it was concentrated in the range 0.5–1.0 nm (70%, 76%, and 88% of La for the P0, P10, and P20 blends, respectively). The average La decreased when a higher proportion of PODEn was used, from 1.041 nm (P0), to 0.950 nm (P10), and then 0.819 nm (P20). The smaller the La, the higher the degree of carbon structural disorder was, and a previous study revealed that disordered amorphous carbon has higher oxidation reactivity than that of highly ordered graphitic carbon (Lu et al., 2012). Additionally, Vander Wal and Mueller (2006) observed that carbon atoms at edge-site

(a) P0 (b) P10 (c) P20

Fig. 4. HRTEM images and the corresponding skeleton images.

(a) P0 (b) P10 (c) P20

Fig. 5. Histograms of fringe length.

Yang et al., Aerosol and Air Quality Research, 16: 2560–2569, 2016 2564

positions have a greater activity than those in the basal plane. P20 PM was more likely to be oxidized because it contained a larger number of edge-site carbon atoms.

The fringe separation Ds of primary particles produced when the P0 blend was used was mostly distributed in the range 0.33–0.44 nm with a median of 0.378 nm (Fig. 6). Separation distance increased with the PODEn concentration, with the mean values being 0.393 nm and 0.404 nm for the P10 and P20 blends, respectively. Higher Ds enhances the oxidation reactivity of the primary particles, because it facilitates oxygen access to carbon layers (Al-Qurashi and Boehman, 2008). Therefore, the oxidation reactivity of the PM samples derived from P10 and P20 was enhanced.

The distribution of the tortuosity Tf was mostly concentrated in the range 1.5–1.7 (Fig. 7). The mean tortuosity was higher for blends with a higher proportion of PODEn, ranging from 1.566 (P0), to 1.582 (P10), and then 1.613 (P20). Tortuosity is attributed to the existence of 5- or 7-member ring structures in the layer plane and relates to the degree of carbon layer disorder (Yehliu et al., 2011). When the tortuosity is increased, the proportion of carbon layers with a graphite structure is decreased; hence, PM samples emitted from P10 and P20 were more easily oxidized than those from P0 TG Analysis Pyrolysis Process Analysis

Pyrolysis is a basic thermo-chemical conversion, and is the initial and associated reaction involved in gasification, liquefaction and combustion. Fig. 8 depicts the thermogravimetric (TG) and derivative thermogravimetric (DTG) curves of PM Samples.

Fig. 8(a) shows the TG curves of moisture evaporation and volatile matter desorption in the temperature region 60–400°C. At low temperatures, no intensive chemical reactions were found. The TG curves remain relatively constant. Mass loss was insignificant at approximately 1%, 3% and 5% of the total mass for the P0, P10 and P20 blends, respectively. It can be deduced that the moisture and volatile matter content were very low in the PM samples, which may be related to the sampling method used. In the exhaust pipe, where temperatures are high, most of the volatile materials are present in the gas phase and it is difficult for them to be adsorbed or coagulate into existing PM (Stratakis and Stamatelos, 2003). The oxidation reaction of solid carbon in the PM occurs at high temperatures (400–750°C). The PM oxidation process occurred at lower temperatures as the PODEn proportion was increased. After reaching ignition temperature, the TG curves decline rapidly, and the PM sample mass loss exceeded 95% at this stage. The PM was burnt out at 750°C, after which the TG curves tend to stabilize.

From Fig. 8(b), weightlessness peak was present in the high temperature regions on the DTG curve. Weightlessness peak become larger when P10 and P20 were used. As PODEn are oxygenated fuels, more oxygen content may be present in the PM, which promotes oxidation. P10 and P20 demonstrated a higher oxidation rate compared with P0 (Karin et al., 2015). Characteristic Parameters Analysis

In the present study, the maximum weight loss rate, peak temperature and half-life temperature were analyzed.

(a) P0 (b) P10 (c) P20

Fig. 6. Histograms of fringe separation.

(a) P0 (b) P10 (c) P20

Fig. 7. Histograms of fringe tortuosity.

Yang et al., Aerosol and Air Quality Research, 16: 2560–2569, 2016 2565

Fig. 8. TG curves and DTG curves of PM samples.

The maximum weight loss rate reflects the combustion of PM after ignition. It corresponds to the highest reaction rate and has a minimum value that can be found from the DTG curve. The peak temperature (TP) refers to the temperature at which the maximum rate of weight loss occurs. The half-life temperature (T50) is defined as the temperature at which 50% of PM mass has been converted. From the TG and DTG curves, the pyrolysis characteristics of the PM samples were obtained (Table 2). The maximum rate of weight loss was higher for higher PODEn blending ratios, as was the PM burn rate. Additionally, TP and T50 diminished, the temperature at which the PM's initial oxidation reaction occurred decreased, and the activation energy E decreased.

The ignition temperature Ti is the minimum temperature of continuous combustion in air or oxygen environments, and represents the difficulty index of PM ignition. Numerous methods for calculating Ti using TG and DTG curves exist, but this study used the method described in Fig. 9. An initial three instants on the TG and DTG curves should be identified. Point B is a perpendicular line from the minima of DTG peak A (the point of the maximum rate of weight loss), which transects the TG curve. Point C is the beginning of volatilization. A tangent to the TG curve at point B and another horizontal tangent at point C are drawn. The point at which these lines crosses is point D, corresponding to Ti. When a higher PODEn blending ratio was used, the PM sample Ti decreased by 2.16% (P10) and 5.97% (P20), compared with P0 (Table 3). Less graphitization and more graphite structure disorder corresponds to a lower PM Ti.

Combustibility Index

Combustion at low heating rates is determined by the chemical reaction dynamics in the initial stage of the reaction. These chemical reaction dynamic factors may also control the reaction rate. The Arrhenius equation expresses

the combustion rate as follows (Rodríguez-Fernández et al., 2011; Wang et al., 2014): d expdm EAt RT

(1)

where m is mass (mg), t is time (min), dm/dt is the combustion rate (mg min–1), A is the Arrhenius pre-exponential factor (s–1), E is the activation energy (kJ mol–1), R is the molar gas constant (8.31 J mol–1 K–1) and T is the absolute temperature (K).

Derivation of Eq. (1) gives the following equation:

2

d d d 1d d d

R m mE T t t T

(2)

At Ti, Eq. (2) becomes

i i

2i

d d d 1( ) ( )d d dT T T T

R m mE T t t T

(3)

Multiplying

i

max mean

h

d / d d / dd / d

T T

m t m tm t T

by both sides of

Eq. (3) gives

i i

max mean

h

max mean2

i h

(d / d ) (d / d )d d( )d d (d / d )

(d / d ) (d / d )T T T T

m t m tR mE T t m t T

m t m tT T

(4)

where (dm/dt)max is the maximum combustion rate (mg min–1), (dm/dt)mean is the average combustion rate (mg min–1),

Table 2. Pyrolysis characteristic parameters of PM samples.

PM samples Maximum weight loss rate / (% °C–1) TP / °C T50 / °C P0 0.95 662 649 P10 0.98 651 629 P20 1.00 634 603

Yang et al., Aerosol and Air Quality Research, 16: 2560–2569, 2016 2566

Fig. 9. Definition diagram of ignition temperature

Table 3. Combustibility index of PM samples.

PM samples

Ignition temperature Ti /

K

Burnout temperature Th /

K

Maximum combustion rate (dm/dt)max / (mg min–1)

Average combustion rate (dm/dt)mean /

(mg min–1)

Combustibility index S/ (mg2 min–2 K–3)

P0 876 993 0.18 2.03 × 10–2 4.73 × 10–12 P10 863 968 0.20 2.21 × 10–2 6.10 × 10–12 P20 840 943 0.21 2.34 × 10–2 7.25 × 10–12

(dm/dt)|T=Ti is the combustion rate at Ti (mg min

–1), Th is the PM burnout temperature (K), Th is defined as the temperature at which a conversion of 98% is achieved; it can reduce errors and guarantee complete PM combustion. R/E represents the reactivity of the PM. (d(dm/dt)/dT)|T=Ti indicates the conversion percentage of the combustion rate at Ti; ignition is quicker when this is maximized. (dm/dt)max/(dm/dt)|T=Ti is the ratio of the maximum combustion rate to the rate of combustion at Ti. (dm/dt)mean/Th denotes the ratio of the mean combustion rate to Th, and at high values PM burns rapidly. The product of these terms defines the combustion characteristic of the PM.

The combustibility index S is defined as (Wang et al., 2009; Xiu et al., 2011)

max mean2

1

d dd d

h

m mt tS

T T

(5)

Several crucial parameters such as Ti, Th, the maximum

rate of combustion, and the average rate of combustion are used in a comprehensive evaluation. It can be deduced from Eq. (5) that a higher S denotes more efficient PM combustion.

The combustibility index and other related parameters of the PM samples are presented in Table 3. As Ti and Th were lowered, the combustion time of the PM samples decreased. Compared with P0, the maximum rates of combustion for the P10 and P20 blends were 11.11% and 16.67% higher, respectively, and the average rates were 8.87% and 15.27% higher. The index S improved as graphite structure disorder increased. Thus, high graphite structure disorder can improve

the combustion characteristics of P10 and P20, resulting in an improvement in PM oxidation ability. Burnout Characteristic Index

The burnout characteristic index Cb is the main signifier of PM combustion performance, and influences a PM’s regeneration time and regeneration efficiency. Cb was used to evaluate the burnout performance of the PM samples, and was defined as follows (Zhao et al., 2015):

1 2b

0

f fC

(6)

where 0 is the burnout time (min), equal to the time taken to burn off 98% of the sample’s initial mass. f1 is initial burnout rate, equal to the ratio of PM weight loss on ignition to total PM mass (%),and reflects the effect of volatile components on the PM ignition characteristics. f2 = f – f1 is the change in the burnout rate (%),which reflects the carbon burnout performance of the PM and is related to the content and form of the carbon present in the PM. f is the total burnout rate, defined as the ratio of PM weight loss at burnout time to the PM total mass, and was set as 98%. Cb and other related parameters of the PM samples are listed in Table 4. The burnout characteristic index increased gradually when a higher proportion of PODEn in the blend was used, which reveals that if the degree of graphitization decreases or the degree of graphite structure disorder increases, the burnout performance of the PM improves. Calculation of Pyrolysis Kinetic Parameters

Pyrolysis kinetics expresses the influence of various

Yang et al., Aerosol and Air Quality Research, 16: 2560–2569, 2016 2567

Table 4. Burnout characteristic index of PM samples.

PM samples Burnout time 0 / min Initial burnout rate f1 /

% Burnout rate of change f2 / %

Burnout characteristic index Cb / min–1

P0 67.1 19.9 78.1 2.3 × 10–3 P10 64.5 25.7 72.3 2.9 × 10–3 P20 62.6 30.7 67.3 3.3 × 10–3

parameters on the conversion rate of PM during pyrolysis, and its purpose is to analyze reaction processes quantitatively and to determine reaction mechanisms. Pyrolysis kinetic parameters are commonly calculated using the TG curve. This study adopted the method of Stratakis and Stamatelos (2003). Using the Arrhenius formula, the kinetic parameters of the PM samples were calculated from the mass reduction during PM pyrolysis. The thermal kinetic equation can be written as

/dd

E RT nm Ae mt

(7)

where m is the initial mass of the sample undergoing the reaction (mg), t is the time (s) and mn is the mass function. n is the reaction order, which for diesel PM in an oxidation environment is assumed to be equal to 1 (Neeft et al., 1997). If the time step is sufficiently small, it can be assumed that the TG curve is composed of very small linear segments of length Δt during which the reaction rate is constant, giving dm/dt ≈ Δm/Δt. Eq. (7) can be replaced with the differential equation

/E RTm Ae mt

(8)

The logarithm of Eq. (8) gives

1 1ln( ) ln ( )m Em At R T

(9)

Thus, ln[(–Δm/Δt) m–1] varies linearly with 1/T. The line

intercepts the y-axis at lnA, as illustrated in Fig. 10. Based on the experimental results of PM samples, the activation energy E and pre-exponential factor A can be determined from the slope and this line intercept. The linear regression is of an acceptable quality because R2 is > 0.98 (Table 5).

During combustion, chemical reactions were accomplished by activated molecules. The kinetic energy was sufficiently high (> E) to destroy the original molecule structure so that reactions could begin. Therefore, activation energy is a standard for determining the chemical reaction rate. It is generally believed that lower activation energies lead to faster chemical reaction rates. The data of Table 5 reveal that E and A decreased when a higher proportion of PODEn in the blend was used. It is clear that PM with less graphitization or more graphite structure disorder more easily oxidizes at low temperatures. As the Ti of the PM samples decreased, pyrolysis performance improved, which is advantageous for DPF regeneration.

Fig. 10. Fitting curves.

Table 5. Pyrolysis kinetic parameters of PM samples.

PM Samples Temperature / °C Activation energy E / (kJ mol–1) Pre-exponential factor A / s–1P0 602–662 134 7.36 × 104

P10 590–651 128 5.28 × 104 P20 567–634 120 2.12 × 104

Yang et al., Aerosol and Air Quality Research, 16: 2560–2569, 2016 2568

CONCLUSIONS

This study investigated the impact of fuel blends on the nanostructure and pyrolysis of PM. Three PODEn/Diesel fuel blends were used by a single-cylinder diesel engine, and the PM emissions collected through direct sampling were characterized using SEM, TEM, and TGA.

PM generated by the P20 blend was the loosest collected, and had a smaller average diameter than that of PM generated by the P10 and P0 blends, which led to improved oxidative combustion.

When a higher proportion of PODEn in the blend was used, fringe lengths were reduced, and the separation distance and tortuosity increased, all of which could enhance the oxidation reactivity of the PM.

Pyrolysis characteristics and kinetic parameter analysis demonstrated that as the ignition temperature, activation energy and pre-exponential factor decreased, the combustibility and burnout characteristic indices were enhanced when a higher proportion of PODEn in the blend was used. This indicates that PM with little graphitization or high graphite structure disorder has higher temperature sensitivity, improved combustion efficiency and burning performance, and enhanced pyrolysis performance. ACKNOWLEDGMENTS

The authors would like to acknowledge financial support provided by State Key Laboratory of Automotive Safety and Energy (KF14112), State Key Laboratory of Engines (K2014-6) and National Natural Science Foundation of China (51375029). REFERENCES Al-Qurashi, K. and Boehman, A.L. (2008). Impact of

exhaust gas recirculation (EGR) on the oxidative reactivity of diesel engine soot. Combust. Flame 155: 675–695.

Burger, J., Siegert, M., Strofer, E. and Hasse, H. (2010). Poly(oxymethylene) dimethyl ethers as components of tailored diesel fuel: Properties, synthesis and purification concepts. Fuel 89: 3315–3319.

Cheruyiot, N.K., Lee, W.J., Mwangi, J.K., Wang, L.C., Lin, N.H., Lin, Y.C., Cao, J.J., Zhang, R.J. and Chang-Chien, G.P. (2015). An overview: polycyclic aromatic hydrocarbon emissions from the stationary and mobile sources and in the ambient air. Aerosol Air Qual. Res. 15: 2730–2762.

Chuepeng, S., Xu, H.M., Tsolakis, A., Wyszynski, M.L., Price, P., Stone, R., Hartland, J.C. and Qiao, J. (2008). Particulate emissions from a common rail fuel injection diesel engine with RME-based biodiesel blended fuelling using thermo-gravimetric analysis. SAE Tech. Paper 2008–01–0074.

Gill, S.S., Tsolakis, A., Herreros, J.M. and York, A.P.E. (2012). Diesel emissions improvements through the use of biodiesel or oxygenated blending components. Fuel 95: 578–586.

Gonzalez-Flecha, B. (2004). Oxidant mechanisms in

response to ambient air particles. Mol. Aspects Med. 25: 169–182.

Ishiguro, T., Takatori, Y. and Akihama, K. (1997). Microstructure of diesel soot particles probed by electron microscopy first observation of inner core and outer shell. Combust. Flame 108: 231–234.

Kaiser, J. (2005). Mounting evidence indicts fine-particle pollution. Science 307: 1858–1861.

Karin, P., Borhanipour, M., Songsaengchan, Y., Laosuwan, S., Charoenphonphanich, C., Chollacoop, N. and Hanamura, K. (2015). Oxidation kinetics of small ci engine's biodiesel particulate matter. Int. J. Automot. Technol. 16: 211–219.

Lapuerta, M., Ballesteros, R. and Rodriguez-Fernandez, J. (2007). Thermogravimetric analysis of diesel particulate matter. Meas. Sci. Technol. 18: 650–658.

Li, N., Sioutas, C., Cho, A., Schmitz, D., Misra, C., Sempf, J., Wang, M., Oberley, T., Froines, J. and Nel, A. (2003). Ultrafine particulate pollutants induce oxidative stress and mitochondrial damage. Environ. Health Perspect. 111: 455–460.

Li, X.H. (2011). Environmental Pollution and Control on Automobile. National Defense Industry Press, Beijing (in Chinese).

Lu, T., Cheung, C.S. and Huang, Z. (2012). Effects of engine operating conditions on the size and nanostructure of diesel particles. J. Aerosol Sci. 47: 27–38.

Mwangi, J.K., Lee, W.J., Tsai, J.H. and Wu, T.S. (2015). Emission reductions of nitrogen oxides, particulate matter and polycyclic aromatic hydrocarbons by using microalgae biodiesel, butanol and water in diesel engine. Aerosol Air Qual. Res. 15: 901–914.

Neeft, J.P.A., Nijhuis, T.X., Smakman, E., Makkee, M. and Moulijn, J.A. (1997). Kinetics of the oxidation of diesel soot. Fuel 76: 1129–1136.

Popovicheva, O.B., Kireeva, E.D., Steiner, S., Rothen-Rutishauser, B., Persiantseva, N.M., Timofeev, M.A., Shonija, N.K., Comte, P. and Czerwinski, J. (2014). Microstructure and chemical composition of diesel and biodiesel particle exhaust. Aerosol Air Qual. Res. 14: 1392–1401.

Rodríguez-Fernández, J., Oliva, F. and Vázquez, R.A. (2011). Characterization of the diesel soot oxidation process through an optimized thermogravimetric method. Energy Fuels 25: 2039–2048.

Saxena, M.R. and Maurya, R.K. (2016). Effect of butanol blends on nano particle emissions from a stationary conventional diesel engine. Aerosol Air Qual. Res. 16: 2255–2266.

Sharma, H.N., Pahalagedara, L., Joshi, A., Suib, S.L. and Mhadeshwar, A.B. (2012). Experimental study of carbon black and diesel engine soot oxidation kinetics using thermogravimetric analysis. Energy Fuels 26: 5613–5625.

Shukla, P.C., Gupta, T. and Agarwal, A.K. (2014). A comparative morphological study of primary and aged particles emitted from a biodiesel (B20) vis-à-vis diesel fuelled CRDI engine. Aerosol Air Qual. Res. 14: 934–942.

Stratakis, G.A. and Stamatelos, A.M. (2003).

Yang et al., Aerosol and Air Quality Research, 16: 2560–2569, 2016 2569

Thermogravimetric analysis of soot emitted by a modern diesel engine run on catalyst-doped fuel. Combust. Flame 132: 157–169.

Su, D.S., Jentoft, R.E., Müller, J.O., Rothe, D., Jacob, E., Simpson, C.D., Tomović, Ž., Müllen, K., Messerer, A., Pöschl, U., Niessner, R. and Schlögl, R. (2004). Microstructure and oxidation behaviour of euro iv diesel engine soot: A comparative study with synthetic model soot substances. Catal. Today 90: 127–132.

Sukjit, E., Herreros, J.M., Piaszyk, J., Dearn, K.D. and Tsolakis, A. (2013). Finding synergies in fuels properties for the design of renewable fuels-hydroxylated biodiesel effects on butanol-diesel blends. Environ. Sci. Technol. 47: 3535–3542.

Tsai, J.H., Chen, S.J., Huang, K.L., Chaung, H.C., Lin, W.Y., Lin, C.C., Wu, T.Y., Yang, C.H. and Chiu, J.Y. (2015). Biological toxicities of exhausts from a diesel-generator fueled with water-containing acetone/butanol and waste-edible-oil-biodiesel blends. Aerosol Air Qual. Res. 15: 2668–2675.

Tsai, J.H., Chen, S.J., Huang, K.L., Chang-Chien, G.P., Lin, W.Y., Feng, C.W., Syu, J.Y. and Huang, H.T. (2016). Characteristics of persistent organic pollutant emissions from a diesel-engine generator fueled using blends of waste cooking oil-based biodiesel and fossil diesel. Aerosol Air Qual. Res. 16: 2048–2058.

Vander Wal, R.L., Tomasek, A.J., Pamphlet, M.I., Taylor, C.D. and Thompson, W.K. (2005). Analysis of hrtem images for carbon nanostructure quantification. J. Nanopart. Res. 6: 555–568.

Vander Wal, R.L. and Mueller, C.J. (2006). Initial investigation of effects of fuel oxygenation on nanostructure of soot from a direct-injection diesel engine. Energy Fuels 20: 2364–2369.

Vander Wal, R.L., Yezerets, A., Currier, N.W., Kim, D.H. and Wang, C.M. (2007). Hrtem study of diesel soot collected from diesel particulate filters. Carbon 45: 70–77.

Wang, C.M., Xu, H.M., Herreros, J.M., Lattimore, T. and Shuai, S.J. (2014). Fuel effect on particulate matter composition and soot oxidation in a direct-injection

spark ignition (DISI) engine. Energy Fuels 28: 2003–2012.

Wang, C.P., Wang, F.Y., Yang, Q.R. and Liang, R.G. (2009). Thermogravimetric studies of the behavior of wheat straw with added coal during combustion. Biomass Bioenergy 33: 50–56.

Xiu, S., Rojanala, H.K., Shahbazi, A., Fini, E.H. and Wang, L. (2011). Pyrolysis and combustion characteristics of Bio-oil from swine manure. J. Therm. Anal. Calorim. 107: 823–829.

Yang, H., Li, X.H., Mu, M.F. and Li, X.H. (2015). Experiments on the performances and emissions of diesel engine fuelled with diesel/polyoxymethylene dimethyl ethers blends. J. Automot. Saf. Energy 6: 280–285 (in Chinese).

Yehliu, K., Vander Wal, R.L. and Boehman, A.L. (2011). Development of an hrtem image analysis method to quantify carbon nanostructure. Combust. Flame 158: 1837–1851.

Yehliu, K., Vander Wal, R.L., Armas, O. and Boehman, A.L. (2012). Impact of fuel formulation on the nanostructure and reactivity of diesel soot. Combust. Flame 159: 3597–3606.

Zhao, Q., Wang, H., Qin, Z.F., Wu, Z.W., Wu, J.B., Fan, W.B. and Wang, J.G. (2011). Synthesis of polyoxymethylene dimethyl ethers from methanol and trioxymethylene with molecular sieves as catalysts. J. Fuel Chem. Technol. 39: 918–923.

Zhao, Y., Wang, Z., Liu, S., Li, R.N. and Li, M.D. (2015). Experimental study on the oxidation reaction parameters of different carbon structure particles. Environ. Prog. Sustainable Energy 34: 1063–1071.

Zhao, Y.P., Xu, Z., Chen, H., Fu, Y.C. and Shen, J.Y. (2013). Mechanism of chain propagation for the synthesis of polyoxymethylene dimethyl ethers. J. Energy Chem. 22: 833–836.

Received for review, July 26, 2016 Revised, September 26, 2016

Accepted, September 26, 2016