Keywords

1 Introduction

Diesel smoke, so-called particulate matter (PM), can be presented as a combination of carbonaceous material (soot) blended with several types of inorganic and organic substances, resulting in mutagenic and carcinogenic elements by nature (Silverman et al. 2012). Soot structural modifications (i.e., morphology) do not only impact the oxidative/mutagenic properties of the soot but also influence the motion and contact (i.e., convective or diffusive flow) between the particulates which in turn impacts the magnitude of the Peclet number (Pe) accordingly. Pe is a dimensionless factor whose magnitude strongly depends on the soot morphological aspects, mainly the primary particulate size (dp0) and to a lesser extent, the radius of gyration (Rg) (Konstandopoulos et al. 2002). The lower the Pe (<1) the more open and porous the soot cake is (i.e., layer of soot deposited in the DPF channels) and the higher the Pe (1 < Pe < 5) the more compact it is. In general, an increase in dp0 will yield a higher Pe number and vice versa (Konstandopoulos et al. 2002). Factors as the soot-cake permeability, packing density, porosity, pressure drop in the channels and DPF thermal behavior during regeneration can be calculated (i.e., simulated) once obtaining the Pe number (Konstandopoulos et al. 2002; Chiatti et al. 2009; Konstandopoulos and Kostoglou 2004). The reader is referred to (Konstandopoulos et al. 2002; Chiatti et al. 2009; Konstandopoulos and Kostoglou 2004) for a more detailed interpretation of the incorporated formulas that link the Pe number with the soot-cake characteristics. As a matter of fact, recognizing the filtration efficiency of a DPF is important; however, predicting the exact dimensioning of the system to optimize the resulted backpressure is even more critical, especially in the norm of the severe legislations and economic viability of diesel-powered vehicles (Fig. 9.1).

Fig. 9.1
figure 1

Reproduced from (Tassopoulos et al. 1989)

Soot-cake representation with regard to its Peclet number.

Full size image

After recognizing the importance of the particulates’ structure on the DPF performance, it becomes a subject of interest to understand if the exhaust particulates will face any modifications within the aftertreatment units, e.g., diesel oxidation catalyst (DOC), selective catalytic reduction (SCR), etc., before being trapped in the DPF. While the dependency of the soot structure on its fuelling source and engine technology is a well-researched topic, to our knowledge, limited work (Ma et al. 2014; Liati et al. 2013; Lapuerta et al. 2007; Fayad et al. 2015; Lee and Zhu 2005; Serhan et al. 2019) has been carried out regarding the impact of the aftertreatment system on the structure of the exhaust PM (i.e., morphology and nanostructure). The chapter will be divided into two sections, research work based on (i) diesel oxidation catalyst and (ii) hydrocarbon-SCR (HC-SCR) catalyst. The impact of the fuel type will be discussed in each section separately.

2 Variation in Soot Structure Throughout DOC

Lapuerta et al. (2007) studied the impact of changing the soot sampling point in the exhaust line to check if dp0 will be altered by the thermal or fluid-dynamic conditions along the line. However, the study did not involve any aftertreatment unit. It was seen that dp0 remains constant throughout all the tested positions since as reported, the exhaust temperature level (110–324 °C) was not enough to launch any chemical reactions with the soot surface. Liati et al. (2013) studied the morphological and nanostructure aspects of soot particulates upstream and downstream the DOC in an effort to understand the impact of this catalyst on the particulate’s oxidative reactivity. Concerning the morphological alterations, slight reduction was noted along dp0- when the particulates flow throughout the catalyst. This was due to the devolatilization of the soluble/semi-volatile organic fraction from the particulates’ surface; however, these changes were not linked with the existence of any possible carbon oxidation phenomenon. As for the graphitization order, slightly more ordered structure was presented downstream the DOC, and therefore, to some extent, the particulates were considered less reactive toward oxidation. Close edge X-ray absorption fine structure analysis (NEXAFS) was also employed to understand the difference in the particulate’s surface functional compounds upstream and downstream the catalyst. It was shown that marginal changes took place downstream the DOC and more specifically, most of the carboxyl functionalities had disappeared and the soot seemed to be more homogeneous in its chemical structure. Ma et al. (2014) studied the impact of the DOC on the particulate’s carbon layer arrangement and its corresponding oxidative reactivity. The analysis showed that DOC can alter the primary particulate’s inner structure in a way that makes it more ordered toward graphitization, e.g., longer carbon layer length (La) and remarkably shorter interlayer spacing (d002). As for the oxidative behavior, the post-catalyst soot seemed to be more reactive, yet a different oxidation mode was recorded compared to that in the pre-catalyst case. To be more specific, the soot mass loss was extremely faster than that of the pre-catalyst test at the early stages of the oxidation (before 50% soot mass loss); however, at later stages (after 50%), the oxidation rate turned to be slower and the residual ash mass significantly higher. Concerning the residuals, they were mainly influenced by the sample collection method and not by any direct soot interactions with the DOC. Longer period was needed to collect the soot downstream the catalyst since the increased concentrations of nitrogen dioxides (NO2) (resulted from the nitrogen oxide (NO) oxidation in the DOC) continuously oxidized the soot placed on the filter. However, during the collection time, the ash deposition remained intact and did not oxidize, thus increasing the ash percentage of the sample placed after the DOC. The soot nanostructure variations were also recorded throughout the different stages of the temperature-controlled oxidation to further investigate the oxidation mechanism of the soot samples (pre- and post-catalyst). It was concluded that the soot collected upstream the DOC undergoes a surface burning process; in other words, it keeps a shell-core structure along all the oxidation period. As for the post-catalyst particulates, the same oxidation mode as that of the “biodiesel oxidation” (Song et al. 2006) was seen for this testing case. In summary, despite the more graphitic initial structure, the high-resolution transmission electron microscopy (HRTEM) analysis revealed that the soot undergoes an internal-burning process at the early stages of the oxidation which leads to a capsule-like structure. The significant energy released during this phase helps to thermally restructure the soot silhouette and transform it to a closed outer shell profile. The highly stable configuration of the latter is the main reason for the slower oxidation rates recorded after the first 50% of the soot weight loss. It was speculated that the DOC contributed to the devolatilization of the surface volatile compounds, which in turn generated this unique mode of oxidation. Lee and Zhu (Lee and Zhu 2004, 2005) studied the evolution of the soot morphological parameters along the different positions of the exhaust system (that includes two DOCs) as a function of engine speed and load condition. The different soot sampling positions are presented in Fig. 9.2. The analysis showed that independently from the engine operating settings, the first DOC helped to reduce the particulates Rg and dp0 and increased their circularity, e.g., higher fractal dimension (Df). The reduction in dp0 was related to the existence of a soot catalytic oxidation, which was also confirmed through qualitative inspection for the HRTEM micrographs. It was shown that upstream the catalyst, the particulates presented the standard graphitic configuration of diesel soot where the primary particulates are mostly spherical and quite distinct from the other spherules. However, downstream the catalyst, the primary particulates appear to have no definite borders and seem to be fused in each other, indicating the presence of catalytic oxidation. The second DOC (Position P4) also reduces dp0; however, the particulates Rg tend to slightly increase and their geometrical configuration highlights a more chain-like structure (lower Df). It was expected that the resulted particulates undergo an enhanced aerodynamic mixing after going out of the catalyst (after 200 cm), which in turn favors their aggregation together to yield larger particulates with a more complex geometry. Overall, it was seen that dp0 and Rg decrease progressively throughout the catalysts with Rg presenting a bigger drop compared to dp0. It was speculated that this further reduction in Rg could possibly arise from the breakdown of the particulates throughout the catalyst channels along with reduction in the interstitial distances among the spherules. However, the average number of the primary particulates in the aggregates (np0) was not evaluated in this study, keeping those findings (i.e., breakdown and reduction in interstitial spacing) as theoretical approaches and not proved by analysis.

Fig. 9.2
figure 2

Sampling collection positions for reference (Lee and Zhu 2004, 2005)—self-drawn by author

Full size image

Fayad et al. (2015) investigated the influence of the DOC with regard to the oxidation/reduction mechanism of the soot particulates produced from the combustion of diesel, rapeseed methyl ester (RME), and butanol/RME/diesel blend. It was seen that DOC can help in reducing the small-scale particulates (<20 nm) by diffusion mechanism (trapping) and can also result on average with larger soot particulates because of an enhanced collision phenomenon in the catalyst channels. This theory was further verified by morphologically analyzing the particulates, upstream and downstream the catalyst. From the analysis, it was reported that after passing the DOC, the collected soot tended to present a greater np0 and a larger Rg compared to these tested upstream the catalyst, independent of the fuel type. Oxygenated fuel combustion results in particulates which are characterized with smaller primary particles and more skeletal aggregates compared to the diesel case. This, in turn, makes the particles produced from oxygenated fuels more easily trapped in the DOC. However, DOC modified neither the particulates’ dp0 nor the carbon layers’ arrangements, highlighting that a longer soot/catalyst residence time was required to perform those alterations (soot oxidation).

3 Variation in Soot Structure Throughout HC-SCR

Serhan et al. (2019) investigated the impact of a silver alumina (Ag/Al2O3) HC-SCR catalyst on the morphology and nanostructure of the soot aggregates produced from the combustion of diesel and diesel/tri-propylene glycol ether (TPGME) blends (20% vol. basis, TP20). Numerous studies in the literature proved the advantageous effect of PGMEs in reducing the engine-out PM emissions, particularly TPGME which was recognized for yielding near non-sooting combustion (Burke et al. 2015; Natarajan 2001; González 2001). The catalyst role was evaluated upon three main areas, the exhaust temperature by varying engine load (2 and 4 bars indicated mean effective pressure (IMEP)), fuel type, and hydrogen (H2) present in the exhaust. H2 gas has been recognized as a promoter for the Ag/Al2O3 de-NOx activity, especially in the low-temperature regions when enough reductants are available (Breen et al. 2007; Theinnoi et al. 2008). This research work present in-depth detail on the catalyst and fuel-type impact on soot particles and will be detailed following two separate sections, the first highlights the impact on soot morphology and the second highlights the nanostructure modifications.

3.1 Catalyst Impact on the Morphology

3.1.1 Low Exhaust Temperature Condition

The different soot morphological parameters including Rg, np0, and Df for the different testing conditions at low load operation (2 bar IMEP, 180 °C) are plotted in Fig. 9.3. Also, the normal distribution of the primary particulate’s size (dp0) is plotted in Figure.

Fig. 9.3
figure 3

Reproduced from (Serhan et al. 2019)

Variation of Df, np0, and Rg throughout the HC-SCR in the presence and absence of H2 at 2 bar IMEP.

Full size image

It is seen that with or without the presence of H2, aggregates with greater Rg and np0 are seen after the catalyst for both, diesel and TP20 fueling. Therefore, it can be concluded that only a limited portion of the particulates have been trapped inside the catalyst, as it was suggested earlier in reference (Fayad et al. 2015). Besides, it is expected that upon passing the catalyst channels, the soot particulates have faced higher chances of collision with the other neighboring particulates and as a result more mature aggregates (i.e., by size) are spotted downstream the HC-SCR. Furthermore, more spherical particulates were seen downstream the catalyst as suggested from the higher Df values recorded in Fig. 9.3. This could be also the result of the enhanced collision process suggested earlier. Furthermore, the oxidation/activation of the HCs (exothermic reactions) throughout the catalyst along with the localized thermal effect of the H2 are likely to increase the local and overall temperature of the catalyst. Higher catalyst temperatures can also help in thermally restructuring the soot shape into a more spherical entity, especially when H2 is presented. Although it can be assumed that a worsened DPF trapping efficiency has resulted from the presence of more compact and spherical aggregates, the remarkable increase in the overall particulate’s size is expected to compensate their impact (Fayad et al. 2015). This assumption is mainly triggered from the fact the particulate’s diffusion/convection pathway in a DPF (i.e., filtration mechanism) is most pronounced from the particulates’ size and not their shape.

Concerning dp0, insignificant variations were spotted in the absence of H2 for both fueling conditions (Fig. 9.4). However, in the presence of H2, dp0 dropped by 2.7% in case of diesel and 3.6% in case of TP20. The only possible theory to clarify these results is that despite the limited residence time of the particulates in the catalyst channels, soot oxidation reactions were launched once H2 is injected. As for the increased reduction seen in case of TP20 with respect to diesel, it is expected to be the outcome from the enhanced oxidative potential of soot produced from the combustion of TP20 compared to that produced from the combustion of diesel, as reported earlier in Serhan et al. (2018). These possible oxidation reactions are mostly generated from the increased level of NO2 throughout the catalyst channels (i.e., transition of NO into NO2 before being converted to N2 during H2-SCR-assisted cycle, as it is advised by Breen et al. (2007)) since this oxidizer can launch the soot oxidation reactions in a temperature region as low as 200 °C (Ehrburger 2002). This assumption was partially approved by Houel et al. (Houel et al. 2007) who reported that in low-temperature window (200 °C to 350 °C), H2 addition will not only enhance the Ag/Al2O3 de-NOx activity but also withstand it. The simplified reaction pathways presenting the NO2-soot-assisted oxidation mechanism are presented below:

Fig. 9.4
figure 4

Reproduced from (Serhan et al. 2019)

Variation of dp0 throughout the HC-SCR in the presence and absence of H2 at 2 bar IMEP.

Full size image
$$C\left( s \right) + NO_{2} \to CO + NO$$
$$C\left( s \right) + 2 NO_{2} \to CO_{2} + NO$$

Conversely, it is expected that O2 should not assist the oxidation kinetics in this temperature window since higher temperature level (ranging in 500 to 700 °C) is needed. Furthermore, it is also confirmed in the literature that introducing H2 gas in conjunction of an oxygen-rich flow (as the diesel exhaust) can change the Ag particulate chemistry to act as an oxidation catalytic agent. This is believed to result from the production of numerous sub-oxide elements (Kim et al. 2013) along with promoting the development of the active O2- ions that can also assist the soot oxidation process (Vassallo et al. 1995).

In summary, it can be concluded that despite the low catalyst temperature (180 °C), the increase in the local NO2 and surface O2 functional species due to H2 presence can partially oxidize the soot passing throughout the catalyst despite the very limited soot/catalyst residence time (flow-through substrate).

3.1.2 Medium Exhaust Temperature Condition

Similar to the earlier section, the same data type has been plotted in 5 and 6 after increasing the engine load to 4 bar IMEP and, respectively, the catalyst temperature to 280 °C (Figs. 9.5 and 9.6).

Fig. 9.5
figure 5

Reproduced from (Serhan et al. 2019)

Variation of Df, np0, and Rg throughout the HC-SCR in the presence and absence of H2 at 4 bar IMEP along with the representative HRTEM micrographs.

Full size image
Fig. 9.6
figure 6

Reproduced from (Serhan et al. 2019)

Variation of dp0 throughout the HC-SCR in the presence and absence of H2 at 4 bar IMEP.

Full size image

It is seen that the particulates surviving the catalytic reactions present greater Rg, np0, and Df while dp0 is kept constant for the diesel case with insignificant drop seen only for the TP20 case. As a result, it is expected that the particulates have survived an enhanced aggregation mechanism throughout the catalyst channels and no possible catalytic oxidation is presented, even though that the catalyst temperature is increased by 100 °C compared to the 2 bar IMEP test. Introducing 500 ppm of H2 activates the catalyst/soot synergies with significant modifications spotted compared to the 2 bar IMEP testing. As for diesel, a drop of 22% is seen for np0, 5.29% for dp0 and a total reduction of 10.16% for Rg. The significant reduction seen along dp0 confirms that with a little increase in the exhaust temperature, which is still representative of the low-temperature window, minute concentrations of H2 can yield a powerful oxidation along the soot passing throughout the catalyst channels. It is expected that the NO2-soot oxidation path is fully activated in this operating margin (~295 °C) and that the localized thermal impact of H2 can also activate the O2-soot oxidation pathway (Ehrburger 2002; Theinnoi et al. 2012). As for Rg, the catalyst impact was more pronounced with more severe reduction seen along this parameter compared to dp0. A hypothesis could be drawn regarding this trend and can be mainly identified with two different approaches (Lee and Zhu 2004):

  1. (a)

    Breakdown of the aggregates throughout the catalyst channels.

  2. (b)

    Reduction in the interstitial distances among the primary particulates constituting the aggregates.

As np0 also faces a noticeable drop downstream the catalyst, it is most probable that hypothesis (a) is the most dominant mechanism leading to this significant reduction in Rg. In addition, it is also speculated that the active oxygen species produced on the Ag particulates can also improve the soot/oxidizers contact and create electronic interaction between the surface of the catalyst and the soot particulates (thus promoting the breakdown phenomena). Concerning the soot fractal like particulates with more chain-like structure (~12% drop in Df) are spotted downstream the catalyst. This configuration further indicates that hypothesis (b) is not valid for our analysis and highlights that a better DPF functionality can be expected with the application of the H2-assisted HC-SCR catalyst to the exhaust flow. On the other hand, the increased local and global temperature of the catalyst is expected to thermally restructure the particulates into a more spherical entity (higher Df). However, this was not the case with this testing condition, which in turn signifies that the catalytic oxidation reactions were remarkably prevailing at this slightly higher exhaust temperature and were sufficiently energetic to dominate the thermal impact of the catalyst (i.e., resulting in more spherical-like particulates).

The catalytic oxidation of the soot was not only seen through the reduction in their dp0, but also through qualitative inspection for the TEM micrographs (Fig. 9.5). It is seen that upstream the catalyst, the particulates collected present the standard graphitic configuration of a diesel soot, where the primary particulates are mostly spherical and quite distinct from the other spherules. However, downstream the catalyst, the primary particulates appear to have no definite borders and seem to be fused in each other, confirming the presence of a catalytic oxidation. This blur appearance is mostly seen in case of TP20 where no statistically significant number of particles were spotted on the TEM grids to produce a confident morphological analysis.

3.1.3 Catalyst Impact on Nanostructure

Since the soot morphological analysis highlights the presence of possible oxidation mechanism, it is of interest to check if any variation will be seen in the particulates’ nanoscale arrangement. Such kind of information will be valuable to predict the reactivity of the aggregates surviving the catalytic reactions and can also give better insights to validate the oxidation reactions speculated from the reduction seen in dp0. As no noticeable changes were seen with dp0 when H2 is not introduced to the flow, the below analysis will account for the impact of the catalyst only in presence of the H2. To simplify the results, Figure summarizes the resulted nanostructure parameters as a ratio of the values recorded upstream the catalyst compared to that seen downstream the catalyst (i.e., referred as post/pre in Fig. 9.7).

Fig. 9.7
figure 7

Reproduced from (Serhan et al. 2019)

Nanostructure parameter ratios including La, Tf, and d002.

Full size image

Concerning the low exhaust temperature condition (2 bar IMEP), similar trend was seen between diesel and TP20: the catalyst helps in reducing the length of the carbon layers (−8.5% La for diesel and −13% La for TP20) and further twist their skeleton (+2.8% Tf for diesel and +1.5% Tf for TP20). However, inconspicuous variation ranging within the data-error bars is seen along the interlayer spacing (d002 after catalyst/d002 before catalyst ~ 1) for both fueling. All the resulted parameters highlight a better oxidative reactivity for the particulates collected downstream the catalyst.

As for the medium exhaust temperature condition (4 bar IMEP), only diesel particulate will be evaluated since as discussed in the earlier section, no statistically significant number of soot aggregates survived the catalytic reactions in case of TP20. Starting with La, similar results are seen compared to the 2 bar IMEP condition while Tf show a further increase reaching 3.5%. However, d002 present a small drop of 1% compared to the pre-catalyst case, highlighting that more graphitic particulates are presented. It is expected that at this relatively higher exhaust temperature (~300 °C) the higher HC oxidation phenomena and increased NO2-soot oxidation reactions can help in thermally restructuring the neighboring carbon layers into more ordered series, resulting in slightly more graphitic configuration (Ma et al. 2014). However, accounting all the nanostructure parameters, it can be speculated that despite this shorter interlayer spacing, the particulates are expected to be more reactive since the layers’ curvature has been increased by 3.5% and La reduced by 8.5%. Carbon layer curvatures and length have been earlier proved to be more influential in dictating the particulate’s reactivity compared to d002, as reported in (Serhan et al. 2018).

Furthermore, to relate the nanostructure alterations to the possible soot/catalyst oxidation mechanism, the rate of soot burn-off in the catalyst will be calculated according to the difference in dp0 before and after the catalyst. Song et al. (2006) reported that during the first 40% stage of the soot burn-off cycle, the oxidation reactions should follow the surface burning mode mechanism. As a result, the following formula from reference (Song et al. 2006) can be used to assess the soot burn-off rate:

$${{dp0 {\text{post SCR}}} \mathord{\left/ {\vphantom {{dp0 {\text{post SCR}}} {dp0 {\text{pre SCR}}}}} \right. \kern-0pt} {dp0 {\text{pre SCR}}}} = \left( {1 + {\text{z}}} \right)^{1/3}$$

where z is the burn-off rate.

As a result, the soot burn-off is estimated to 8% in case of diesel and 10% in case of TP20 at 2 bar IMEP. Increasing the load to 4 bar IMEP increases the soot burn-off rate to 15% in diesel case while for TP20 there was no clear structure for the soot to measure the morphology/nanostructure parameters; however, it is expected to be higher than diesel case as it is in the 2 bar IMEP condition.

According to Song et al. (2006), surface burning mode mechanism is generally defined with slow combustion reactions translated with a notable reduction in La and no more than 3% increase in the particulate graphitic ordering (using Raman analysis) during the first 20% burn-off phase. However, Ma et al. (2014) presented a more detailed approach concerning the above. The study covers the progress of the soot nanostructure alterations during the oxidation process of separate particulates collected upstream and downstream a DOC catalyst. Different trends were seen along La and d002 between both samples, thus no clear conclusion can be beneficial for our analysis; however, it was confirmed that Tf should increase by ~2% during the first 20% burn-off phase, independently from the soot collection point (downstream or upstream the DOC). Summarizing the findings of reference (Song et al. 2006) and (Ma et al. 2014), during the 20% burn-off phase, La is expected to decrease, while Tf is expected to slightly increase with moderate variation along d002. Those findings positively correlate with the nanostructure variation shown in the analysis, thus confirming that these alterations are resulted from an enhanced oxidation mechanism in the catalyst channels.

4 Summary

This chapter focuses on the impact of the aftertreatment system on the soot aggregates regarding their morphology and nanostructure characteristics. The particulate trapping efficiency in DOC is linked to the parent fuel type. Oxygenated fuel combustion, e.g., butanol/RME/diesel blends, results in particulates which are characterized with smaller primary particles and more skeletal aggregates compared to the diesel case. This, in turn, makes the particles produced more easily trapped in the DOC. Also, it was seen that independent of the fuel type or exhaust temperature condition, soot particulates tend to aggregate with the other neighboring particulates leading to larger and more circular particulates downstream the DOC. The same hypothesis can be also drawn in case of HC-SCR when H2 is not introduced upstream the catalyst. However, injecting minute concentration of H2 helps in significantly modifying the morphology of the particulates which seem to be more dependent on the exhaust temperature when H2 is presented. At low exhaust temperature (180 °C), particulates’ aggregation remains the most dominant mechanism; however, it was seen that the soot primary particle diameter faces some reduction, highlighting the presence of oxidation reactions throughout the catalyst. Increasing the catalyst temperature (290°C) enhances the catalytical modification of the aggregates. It was seen that not only dp0 was reduced but also the soot radius of gyration and number of primary particles included in soot aggregates. Those alterations highlight that at this relatively higher exhaust temperature, soot particulates are not only oxidized, but they also tend to breakdown throughout the catalyst channels instead of being aggregated together. This impact was ameliorated when oxygenated diesel blend (TP20) was used. This is mainly due to the higher oxidative reactivity toward oxygen of the particulates produced from TP20 compared to neat diesel. The TP20 soot catalytic oxidation can be qualitatively inspected from the TEM micrographs. It is seen that upstream the catalyst, the particulates collected present the standard graphitic configuration of a soot particle, where the primary particulates are mostly spherical and quite distinct from the other spherules. However, after passing through catalyst, the primary particulates appear to have no definite borders and seem to be fused in each other, confirming the presence of strong catalytic oxidation in case of TP20 despite the limited residence provided throughout the catalyst channels.