Lignocellulosic octane boosters Citation for published version (APA): Tian, M. (2016). Lignocellulosic octane boosters. Eindhoven: Technische Universiteit Eindhoven. Document status and date: Published: 07/11/2016 Document Version: Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers) Please check the document version of this publication: • A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website. • The final author version and the galley proof are versions of the publication after peer review. • The final published version features the final layout of the paper including the volume, issue and page numbers. Link to publication General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal. If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement: www.tue.nl/taverne Take down policy If you believe that this document breaches copyright please contact us at: [email protected]providing details and we will investigate your claim. Download date: 11. Jul. 2020
Transcript
Lignocellulosic octane boosters
Citation for published version (APA):Tian, M. (2016). Lignocellulosic octane boosters. Eindhoven: Technische Universiteit Eindhoven.
Document status and date:Published: 07/11/2016
Document Version:Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the version of the article upon submission and before peer-review. There can beimportant differences between the submitted version and the official published version of record. Peopleinterested in the research are advised to contact the author for the final version of the publication, or visit theDOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and pagenumbers.Link to publication
General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, pleasefollow below link for the End User Agreement:www.tue.nl/taverne
Take down policyIf you believe that this document breaches copyright please contact us at:[email protected] details and we will investigate your claim.
All rights reserved. No part of the material protected by this copyright notice may be producedor utilised in any form or by any means, electronic or mechanical, including photocopying,recording or by any information storage and retrieval system, without the prior writtenpermission of the author.
A Pressure signal used in knock analysis 115A.1 Resonance frequency calculation . . . . . . . . . . . . . . . . . . . . . . . . . . 115A.2 Overview of pressure based engine knock detection method . . . . . . . . . . . 116A.3 Knock intensity and knock onset definition in this thesis . . . . . . . . . . . . . 118
B Fuel injection amount in IQT 123
C Fuels properties summary 125
ii
Contents
D Abbreviations 129
Acknowledgement 131
List of publications 133
Curriculum Vitae 135
iii
Summary
Efficiency and sustainability are two important drivers for future mobility. For internalcombustion engines, the most widely adopted strategy to boost efficiency is to downsize theengine and compensate the otherwise incurred drop in engine output by means of turbocharging. This trend, however, is limited in spark-ignition (SI) engine by the phenomenon ofengine knock. To complicate matters further, the pressure-temperature history in downsizedengines is markedly different from that occurring in a naturally aspirated one. Both factorswarrant a detailed study on future octane booster requirements.
Therefore, the first part of the thesis explores the causality between fuel molecularstructure and auto-ignition behavior with the aim to develop generic design rules on amolecular level for octane boosters, also taking into account aforementioned trends in SIengine development. The second part of the thesis applies this insight to select variouslignocellulosic derived compounds for further evaluation related to octane quality in variousexperimental setups (e.g., engine, ignition quality tester (IQT)) and by means of chemicalkinetic model. And the latter results also validate former design rules.
More specifically, in the second part, four compounds including cyclic ether ((un)saturatedfuran ether) and levulinate (ketone ester) which can be derived from (hemi-)cellulose, and tenmono-aromatic oxygenates including anisole, guaiacol, etc., which can be derived from ligninor used as lignin models are tested and compared to validate their potential as octane boostersand the design rules.
The last part of the thesis built a detailed chemical kinetic model of the oxidation of4-methyl anisole, which proved to be a good candidate as an octane booster from the secondpart. A detailed kinetic mechanism can help better understanding the combustion processand the impact of the two functional groups. In addition, the mechanism can be used forother modeling application.
v
Samenvatting
Efficiëntie en duurzaamheid zijn twee belangrijke drijfveren voor toekomstige mobiliteit. Demeest toegepaste strategie om de efficiëntie van interne verbrandingsmotoren te verhogen, ishet verkleinen van de motor in combinatie met het toepassen van turbolading om afname vanhet vermogen te voorkomen. Echter, deze trend wordt in motoren met elektrische ontsteking(SI) gelimiteerd door klop (pingelen van de motor). Dit wordt nog verder gecompliceerddoor de druk-temperatuur historie die in verkleinde motoren wezenlijk verschilt van die vanmotoren met natuurlijke aanzuiging. Beide factoren vergen een gedetailleerde studie naar deeisen aan toekomstige octaanverhogers.
Daarom behelst het eerste deel van dit proefschrift een studie naar het causale ver-band tussen de molecuulstructuur van de brandstof en de zelfontstekingsgedrag met alsdoel het ontwikkelen van algemene ontwerpregels voor de moleculaire samenstelling vanoctaanverhogers, rekening houdend met de eerder genoemde trends in de ontwikkelingvan SI motoren. Het tweede deel van het proefschrift past de verkregen inzichten toe omverschillende lignocellulosische derivaten te selecteren voor verdere evaluatie met betrekkingtot de octaankwaliteit in verschillende experimentele opstellingen (d.w.z. een motor en eenontstekingskwaliteittester (IQT)). Daarnaast wordt een chemisch kinetisch model gebruiktom de eerder genoemde ontwerpregels te valideren.
In het bijzonder worden in het tweede deel vier stoffen getest, waaronder cyclische ether((on)verzadigde furaan ether) en levulinaat (keton ester) die kunnen worden verkregen uit(hemi-)cellulose en tien mono-aromatische zuurstofverbindingen inclusief anisol, guajacol,etc. Deze laatste kunnen worden verkregen uit lignine of kunnen worden toegepast als ligninemodelstoffen. De geteste stoffen zijn met elkaar vergeleken als potentiële octaanverhogers enzijn gebruikt om de ontwerpregels te valideren.
In het laatste deel van het proefschrift wordt een chemisch kinetisch model gepresenteerdvoor de oxidatie van 4-methyl anisol; een goede kandidaat octaanverhoger uit het tweededeel. Een gedetailleerd kinetisch mechanisme kan helpen om beter begrip te krijgen van hetverbrandingsproces en de impact van de twee functionele groepen daarbij. Daarnaast kan hetmechanisme worden gebruikt voor andere modelleringsdoeleinden.
vii
Chapter 1Introduction
Spark ignition (SI) engines make up a large fraction of the transportation market. Improving itsefficiency and fuel strategies will have an important contribution to the present energy development.However, SI engine efficiency improvement is largely limited by engine knock. Owing mainly todirect injection, turbocharging and intercooling, the in-cylinder temperature relative to pressure hasdropped considerably over the past decades in SI engines. As any change in temperature/pressurehistory will have an impact on the auto-ignition chemistry, the requirements to fuels also change.1
1.1 Introduction
Midway the 19th century, before the advent of the internal combustion engine, kerosene waslong considered to be the only valuable constituent of crude oil, to be used increasingly forlighting purposes against the background of a waning supply of whale oil. What we refer tonow as gasoline and diesel, are intrinsically co-produced in the refining process of crude oilto kerosene. These side streams were at the time considered either too unstable or dirty forhousehold use, respectively. As such, these streams were either dumped in rivers or burnt onsite.
With exponentially growing demand for kerosene, however, the latent value of aforemen-tioned residual products eventually became an important catalyst for innovation and helpedto spark the development of the first internal combustion engines in the late 19th century.This feedstock driven approach of designing the engine around the prevailing fuel is far fromoptimal, as the prevalence of engine knock, soot and noxious emissions still challenge enginedesigners more than a hundred years later.
In light of increasing legislative demand for biofuels and the emergence of natural gasas an alternative feedstock for fuels, the timing for a paradigm shift, whereby the fuel is
1Part of the results shown in this chapter has been submitted to ’Progress in Energy and CombustionScience’
1
Chapter 1. Introduction
designed around the engine, rather than vice versa, is right. Why not design bio- and naturalgas refineries such that they produce biofuels that compensate for deficiencies in their crudeoil counterparts?
In this regard, an attractive deficiency to target is the relatively poor anti-knock qual-ity of "raw" gasoline. Octane boosters sell at considerable premiums over this particularpetrochemical cut.
1.2 Trends in engine technology
One of the most important fuel parameters for spark-ignition (SI) engines is the anti-knockquality. Knock occurs when the octane requirement of the engine exceeds the octane quality ofthe fuel [1]. Knock occurs when premixed fuel and air, compressed by hot end gas produced bythe propagating flame, are heated up beyond threshold of auto-ignition [2]. Knock manifests asa metallic clanking noise due to the prevailing pressure fluctuations [3], which can ultimatelylead to damage of critical engine parts such as liners, bearings and pistons.
Whether or not a fuel will auto-ignite is as much dependent on the fuel anti-knock qualityas on the prevailing engine operation conditions. This section will focus on those trends in SIengine technology that have most significantly changed the conditions inside the combustionchamber. It will become evident that the common denominator of these trends, notablyturbocharging, direct injection and higher compression ratios, is that all contribute to lowerunburnt gas temperatures relative to pressure. This effective cooling has a favorable impacton the so-called octane appetite of the engine (i.e., minimal fuel octane number required toavoid knock). It will be demonstrated that lower gas temperatures generally result in a lowerminimum octane number requirement for a given operating condition.
1.2.1 Direct injection
Direct fuel injection has been installed onto spark-ignition engines as early as 1902 and firstfeatured in Antoinette aircraft, designed by Leon Levavasseur. Added benefits compared tocarbureted fuel delivery included avoidance of freezing and enabling the use of less volatile, butmore knock resistant alternative fuels [4]. A further refinement of the technology was madeby Bosch fifty years later, the first gasoline direct injection (GDI) engine for an automotiveapplication debuted in the 1952 Goliath GP700 Sport (two-stroke) and subsequently featuredin the 1955 Mercedes 300SL.
After these first models were introduced, GDI was shelved, only to be reintroduced againover forty years later, this time by predominantly Japanese car makers (e.g., Mitsubishi, Nissan,Toyota) in the late 1990’s. Market share for GDI in the EU-27 has since increased from anegligible level in 2001 to 14% in 2010 (Table 1.1).
An important benefit of GDI relative to port fuel injection (PFI) is that the improvedevaporative cooling of the former injection method leads to lower charge temperatures [7].With PFI, relatively large droplets collide with and form a liquid film on the intake valves andport wall. This results in the evaporation process being driven primarily by heat absorptionfrom said surfaces [8]. GDI, conversely, involves injecting a well atomized spray directly intothe combustion chamber, thus leading to vaporization powered chiefly by heat absorption
2
1.2. Trends in engine technology
Table 1.1: Spark ignition engine performance and market share of gasoline directinjection and turbocharging
EU-27 W-EUYear Market share GDI Power Displacement Power density Market share turbo
from the charge alone [8]. Accordingly, post fuel injection charge temperatures will be lowerfor GDI than is the case for PFI. Charge temperatures before compression were even higherbefore the advent of PFI, when carburetters were still the dominant fuel delivery system. Toprevent icing, carburetters were heated as a matter of course, which manifested in intake airtemperatures of 50◦C to levels as high as 150◦C vis-à-vis 30◦C for PFI [3].
1.2.2 Downsizing
First introduced in the 1962 Oldsmobile Cutlass, the initial purpose of turbochargers in SIengines was to boost power in US built muscle cars. Porsche had the European debut in1975 in its 911 model. Up until the late 90’s, however, turbocharged SI engines were usedonly in niche applications. As can be seen in Table 1.1, boosted SI engines became moremainstream from the 2000’s onwards and had nearly a quarter of the Western Europeanmarket by 2010. This sharp increase was motivated not only by continuous demand for higherpower output, but increasingly also by both legislative and consumer pressure to improvethe fuel economy. As it happens, the fuel economy benefits considerably from downsizing tosmaller displacement volumes.
By means of turbocharging, the otherwise incurred drop in power is compensated byhigher intake pressures, effectively mimicking a larger displacement volume. Owing tocorrespondingly lower throttling and other parasitic losses, engine efficiency can increase byas much as 18% when a 1.6 l naturally aspirated engine is downsized to 0.8 l and subsequentlyturbocharged to maintain power output [7]. In Table 1.1, it can be seen that the increase inmarket share of turbocharged engines coincided with both a decrease in displacement volumeand increase in power density; both being hallmarks of downsizing.
As discussed earlier, GDI reduces the charge temperature due to enhanced evaporativecooling, while leaving unaffected the charge pressure. Turbocharging an engine, conversely,increases the charge temperature by means of compression. This rise, however, is lower thanadiabatic compression would predict. This discrepancy can be traced back to considerable
3
Chapter 1. Introduction
heat losses in the intake manifold and the intercooler, the latter heat sink being the moredominant of the two [7]. In other words, for a given charge pressure, as was also seen earlierfor GDI, intercooler assisted turbocharging has a cooling effect on the charge temperature.
1.2.3 Higher compression ratios
The compression ratio also has an influence on the unburnt charge temperature relativeto pressure. As this ratio is increased, the charge temperature drops relative to pressure[9]. While a higher ratio inherently yields both higher pressures and temperatures due toincreased compression, the latter parameter is affected as well by a less straightforward,secondary mechanism. As the compression ratio is increased, the end of compression volumereduces in size, thereby expelling a greater fraction of hot combustion gases in the exhauststroke. Consequently, less hot residual gas is left over to mix with the cooler air during theintake stroke, effectively reducing the charge temperature relative to pressure [9, 10].
Compression ratios have risen considerably over the past decade, enabled to a large extentby the previously discussed cooling effect of GDI on the unburnt gas temperature [11, 12]. For anaturally aspirated engine, GDI allows increases in the compression ratio from approximately10 to over 11.5 [12]. For a turbocharged engine, GDI has allowed for an increase from about8.8 to 9.6 in recent years [12].
1.3 Octane booster requirements
Auto-ignition chemistry is highly dependent on the pressure/temperature history of theunburnt air/fuel mixture. The engine technologies discussed in Section 1.2 all result in areduction in temperature relative to pressure, thereby reducing the knocking tendency ofthe engine for any given fuel. It will be made clear in this section that today’s milder engineoperating conditions require new measures to determine the octane requirement of an engine.To this end, the history of octane rating methods will be discussed.
1.3.1 Current octane rating methods
Research and Motor Octane Number
Engine knock is caused by the auto-ignition of cylinder end-gas ahead of the spark ignitedflame front [2, 13]. Auto-ignition chemistry of fuels in the low and intermediate temperatureregime is closely related to the fuel’s anti-knock quality. Detailed chemical reaction mecha-nisms controlling ignition properties typically include hundreds of species and thousands ofelementary reactions, especially in the low temperature regime.
Commercial fuels, like gasoline and diesel, are mixtures of different types of hydrocarbons(e.g., paraffins, olefins, aromatics), each class having their own specific auto-ignition charac-teristics. This complexity in real fuels makes it challenging to accurately predict auto-ignitionchemistry in a quantitative sense. To circumvent this, surrogate fuels, comprised of a limitednumber of components, are typically used to predict the ignition quality of real fuels.
4
1.3. Octane booster requirements
Table 1.2: RON and MON engine operating conditions [15]
Parameter RON MONIntake Air Temperature 52◦C 149◦C
Intake Air Pressure atmospheric atmosphericCoolant Temperature 100◦C 100◦C
Compression Ratio 4-18 4-18a Crank angle degrees before top dead center.
Since 1932 [14], knock resistance is commonly expressed in as a research octane number(RON) and motor octane number (MON). To date, both values are determined on a standard-ized Cooperative Fuels Research (CFR) engine in accordance with ASTM protocols D-2699and D-2700 [15], respectively. Both norms were designed to be representative of the most mild(RON) and severe (MON) operating conditions encountered in 1930’s SI engines. In bothtests, the highly reactive n-heptane and highly stable iso-octane are used as surrogate fuels,spanning the octane scale from 0 to 100, respectively [16].
Road Octane Number
From 1947-1996, the Coordinating Research Center (CRC), a U.S. based non-profit orga-nization, conducted annual octane requirement surveys for the purpose of informing theautomotive and oil sectors on the octane appetite of new car populations. Introduction ofknock sensors on most SI engines from the early 1990’s onwards reduced the efficacy andneed of these surveys, ultimately leading to the termination of the program in 1996 [14].
CRC tests were conducted on chassis dynamometers at both part-load and wide-open-throttle (WOT). At both conditions, the so-called Road Octane Number is determined usingfull-boiling unleaded reference or FBRU fuels that have a known RON and MON. Thesemodel fuels were designed to be similar to prevailing commercial unleaded gasolines [17].Ever-poorer octane quality FBRU fuels were burnt until the occurrence of knock at either ofthe aforementioned operating conditions.
Based on the data from these surveys, a correlation for the Road Octane Number, measuredwith realistic fuels in real engines, as a function of RON and MON was introduced by Spitleret al. in 1967 [18].
RoadON = a+ b · RON+ c ·MON (1.1)
Least-square regression of the CRC data yielded values for the various constants. Ananalysis of the trends of these constants over the period 1951-1991 by Mittal and Heywood [14]reveals an ever-decreasing relative importance of MON over RON. This may be interpretedas cooler charge temperatures relative to pressure, brought about by mass adoption of thetechnologies outlined in Section 1.2, effectively reducing the Road ON to a proxy for RON.
5
Chapter 1. Introduction
Table 1.3: K values reported in literature
Maximum Minimum Year of Publication Reference0.28 -0.33 2001 [19]0.21 -1.85 2005 [9]-0.13 -1.68 2006 [21]0.25 -0.65 2008 [3]-0.5 -4 2010 [22]a
-0.12 -0.93 2012 [23]a GDI engine at sea-level
1.3.2 New octane rating method
Knock resistance is highly sensitive to intake temperature relative to pressure [3, 9, 10, 14, 19].RON and MON intake temperatures (Table 1.2), however, are notably higher than is the casefor modern engines, which hover more around 30◦C. This is because the engines in the1930’s, on which the ON tests were based, were fueled by heated carburetors [3] and fittedwith inadequate cooling systems [14].
As exhaust gas catalysts became more efficient and were able to operate at lower light-offtemperatures, there was no longer need for high intake temperatures. Effective exhaust gasaftertreatment thus made possible a reversal towards lower intake temperature, brought about,for the greater part, by the technologies outlined in Section 1.2.
With typical intake temperatures currently well below 50◦C, engines have moved firmlyinto "beyond RON" territory [20]. In order to take this effect into account, a new anti-knockquality measure, the octane index (OI), was introduced by Kalghatgi [19].
OI = (1− K) · RON+ K ·MON (1.2)
The OI (Equation 1.2) introduces a new parameter: K, which is a relative measure ofseverity, or in other words, unburnt charge temperature relative to pressure. Note that the OIcan be written as a version of the road octane number when the constants of said equation areset at 0, (1-K) and K, respectively.
By definition, K holds the value of 0 and 1 for RON and MON conditions, respectively.Traditionally, K was thought to have a positive value of 0.5, equating the OI to the antiknock index (AKI) or pump octane [20], which is still clearly labeled on gasoline fuel pumpsthroughout the United States.
AKI = (RON+MON)/2 (1.3)
Based on a combination of literature, models and CRC reports, Mittal and Heywood [14]reported in 2009 that the value for K had fallen from roughly 1 in 1930, to 0.5 in 1945, anddown to 0 in the early 2000’s. More recent data on K shows that these authors correctlypredicted K would turn negative in the near future (Table 1.3).
The causality of this drop may be attributed to ever-cooler unburnt gas temperaturesrelative to pressure effectuated by the technological developments discussed in Section 1.2.Accordingly, it should not come as a surprise that here is a good correlation for K with theunburnt gas temperature (e.g., at a compression pressure of 15 bar (Tcomp15) [9, 24–26]). At
6
1.3. Octane booster requirements
500 600 700 800 900 1000−2
−1
0
1
2
3
4
5
6
7
8
RON↓MON→MON→
←Beyond MON→
←Beyond RON→
Tcomp15
[K]
K [−
]
Figure 1.1: K versus Tcomb15 (3 [19], 2 [23], ∗: RON and MON condition, × [24], o[26])
15 bar, values for K were found to decrease from around 6 at 1000 K to 0 at 700-800 K, downto -2.5 at 500 K [9]. Note that for RON and MON test conditions this temperature is roughly700 and 850 K, respectively [23]. Data on K versus Tcomb15 from various studies have beenplotted in Figure 1.1.
A second parameter included in the OI methodology is fuel sensitivity (S), which showsfuel’s different auto-ignition property to the changing charge temperature relative to pressure.S is defined here as the difference between RON and MON (Equation 1.4 [19]):
S = RON−MON (1.4)
The OI, as defined earlier by Equation 1.2, can be rewritten as a function of both K and S(Equation 1.5 [19]).
OI = RON− K · S (1.5)
Shown in Figure 1.2, the OI of iso-octane, toluene and 2-methylheptane, each having adistinct S (e.g., 0, 11 and -2.1, respectively) are plotted against K. It can be seen that, whenS > 0, theOI increases as K is decreased. As a consequence, as Kmoves into negative territory,indicative of more advanced engine technology, both a high RON and S are required to ensureadequate anti-knock quality. In other words, in modern engines at least, fuels with a high Soutperform their less sensitive counterparts on the OI scale, an equal RON notwithstanding[22, 23].
Prompted by environmental concerns and associated fuel quality legislation, the relativecontribution of the various hydrocarbon classes in gasoline has shifted significantly over thepast two decades. Olefins (alkenes) have been linked to photochemical smog formation via
7
Chapter 1. Introduction
−2 −1 0 1 2 3 4 5 60
50
100
150
2−Methylheptane (S=−2.1)
iso−Octane (S=0)
Toluene (S=11)
Figure 1.2: OI as a function of K for fuels with different S
Table 1.4: Maximum allowed olefin, aromatic and benzene levels (vol.-%) in EUgasoline
exhaust born intermediates, notably 1,3- butadiene [27–29]. Aromatics have been found toincrease emissions of unburnt hydrocarbons, carbon monoxide and benzene [30]. Listed bythe International Agency for Research on Cancer as a Group 1 substance (i.e., agent (mixture)is carcinogenic to humans), benzene is a particularly toxic member of the aromatics family.This compound has been linked to lymphatic and hematopoietic cancer [31] and leukemia[32]. Moreover, slightly higher NOx emissions have been observed for those gasolines rich inaromatic content [31].
No wonder then that olefinic and aromatic content in gasoline is subjected to increasinglystringent caps (Table 1.4). As a consequence, the paraffinic fraction in gasoline, notablyiso-paraffins (e.g., iso-octane), has steadily risen to compensate for the otherwise drop inoctane quality incurred by the removal of high octane aromatics and olefins.
Aromatics and olefins, however, tend to have S values that far surpass those of paraffins(Figure 1.3). In fact, S for most paraffins is close to naught [1]. As a consequence, aforemen-tioned paraffinic pivot will have a pronounced impact on the OI. As could be seen earlier inFigure 1.2, while older engines benefit from a low S, a high value S is favored in their moremodern counterparts.
8
1.4. Objectives
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.92
4
6
8
10
12
14
16
18
Paraffin fraction
Sen
sitiv
ity
Figure 1.3: Sensitivity plotted against paraffinic fraction for various studies (+ [3] ◦[13] I [24] 3 [33] 2 [34] ∗ [10]4 [35])
Fortunately, this adverse impact on S has been partially offset by downstream blendingof oxygenates: ethers (e.g., MTBE) since the early 80’s and alcohols (e.g., ethanol) since the2000’s. These compounds have a higher RON and S than both gasoline and iso-octane (Table1.5).
1.4 Objectives
Therefore, the fuels with high RON and S are the target fuels desired to be used as octaneboosters in modern engines, which may be aromatics, olefins or oxygenates. This indicatesthe importance of a fuel’s molecular structure to its knock resistance quality. Meanwhile,biomass is a potential renewable energy source, which can be used as the feedstock to producebiofuels. Then the question arises: what kind of molecular structures are good to enhancethe fuel’s anti-knock quality? The essence of this thesis is to find and test the ligno-cellulosederived compounds as octane boosters, the objectives are:
• Analyze the structure effect to the compound’s knock resistance quality
• Study (hemi-)cellulose derived compounds as octane boosters
• Study lignin derived compounds as octane boosters
• Build the detailed oxidation mechanism for a good octane booster that have been tested(4-methyl anisole)
9
Chapter 1. Introduction
Table 1.5: OI and S over range of K for commercial octane boosters
of benzene is from [36], RON and MON of toluene are from [37] RON of Xylenes arefrom [38], RON and MON of MTBE, ETBE and TAME are from [39], RON of
methanol and ethanol are from [40]. b K value is assumed based on Figure 1.1 andTable 1.3. cMTBE: Methyl Tert-Butyl Ether. dETBE: Ethyl Tert-Butyl Ether eTAME:
Tertiary Amyl Methyl Ether
1.5 Outline
The outline of the thesis is as follows: in Chapter 2, the combustion pathways of differentgroups of hydrocarbons that are common components in the gasoline are compared, tooffer insight into the chemistry behind octane boosters and to subsequently distill from thisknowledge multiple generic design rules that guarantee good anti-knock performance.
In Chapter 3, levulinic esters and cyclic ethers which can be derived from (hemi-) celluloseare blended with gasoline, tested and compared in an SI engine. Differences in knockresistance amongst the fuels are subsequently explained by means of a modified ignitionquality tester (IQT) experiments and chemical kinetic models.
In Chapter 4, the aromatic oxygenates with -OH, C=O or −O−CH3 functional groups, i.e.benzyl alcohol, 2-phenyl alcohol, acetophenone, anisole and vetraole, which can be derivedfrom lignin or used as lignin models are mixed with gasoline to test their anti knock quality.Then anisole, guaiacol and their alkylated compounds are analyzed in the same fashion as inChapter 3, with the conclusion that anisole and 4-methyl anisole have the best performance.
Therefore, in order to have a better understanding of the combustion process of 4-methylanisole, the impact of the functional groups and for future modeling applications, a detailedkinetic mechanism of 4-methyl anisole is built in Chapter 5.
Finally, the conclusions from this thesis and the future scope of work are given in Chapter6.
10
Chapter 2Impact of fuel molecular struc-ture on auto-ignition chemistry
This chapter begins with a literature review of the oxidation kinetic mechanisms of various hydrocar-bon groups, to offer insight into the chemistry behind octane boosters. Subsequently, distill from thisknowledge multiple generic design rules that guarantee a compound to have good anti-knock quality,taken into account also recent advances in engine technology mentioned in the last chapter. 1
2.1 Introduction
As discussed in Chapter 1, fuels with a high RON and a high sensitivity (S) are needed for themodern SI engine. These parameters which show fuel’s anti-knock quality are dominated bythe auto-ignition process of the fuel. Fuels which have longer auto-ignition delay time at therelevant temperature and pressure have better knock resistance, that is a high RON of MON,and the sensitivity comes from the different reactivity of the fuel at RON and MON condition.
The auto-ignition is a oxidation process, which are composed of series of chain reactions.There are ten main reaction types in auto-ignition process, which are schematically drawn inFigure 2.1 and tabulated below.
(a) Unimolecular decomposition: bond cleavage at the C-C or C-H bond, which requires alot energy and thus occurs mainly at high temperature.
(b) Hydrogen (H) atom abstraction: a small radical or compound, such as OH, HO2, O2,CH3, reacts with the fuel molecules, removing H in the process. This reaction requiresless energy and it can thus take place at both low and high temperatures.
1Part of the results shown in this chapter has been submitted to ’Progress in Energy and CombustionScience’
11
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
(c) O2 addition reactions: O2 adds to the radical, producing a peroxy radical, and caninitiate low temperature branching reactions. It react rapidly because it is negligibleactivation energy, but peroxy radicals are not stable at high temperature.
(d) Addition reactions: radicals or compounds, such as H, OH or HO2 radicals, attachingthemselves to the unsaturated bond sites and has a relatively low activation energy.
(e) Recombination reactions: radicals combine together to form a stable covalent bond.This also requires little activation energy [41].
(f) β-scission: involves a scission at the bond site in the β position to the atom (C or O)which carrying the unpaired electron, yielding small olefin or aldehyde, respectively,and a radical, in the process. This reaction has a high activation energy of 15 − 50kcal/mol and is therefore mainly a high temperature phenomenon.
(g) Isomerization: this refers to intra H atom transfer from one C atom to another and/orto a C-OO moiety. The activation energy of this reaction depending on the size of thetransition ring, which will be discussed later.
(h) Concerted elimination: two substituents are removed from a molecule in one stepmechanism, forming an unsaturated bond.
(i) Radical cyclization: this yields cyclic products. Like the QOOH decompose to cyclicether and an OH radical.
(j) Disproportionations: two radicals react to form two different non-radical products. Itusually requires less activation energy.
The above reactions types will now be discussed separately for each of the main carbonbond types: paraffinic, olefinic, aromatic and oxygenated. Low to intermediate temperaturerange is the most interesting range because it is relevant to engine condition. Detailedchemistry kinetic models of varies group of compounds have been well reviewed. For example,Miller et al. introduce and summarize the importance and methods to build the kinetic model[42], and reviewed elementary reaction rates in a number of important reaction systems,especially in low temperatures [43, 44]. Battin-Leclerc [45] and Simmie [46] summarized thecombustion model in low and high temperature, respectively, Pitz [47] reviewed the dieselsurrogate fuels kinetic modeling.
2.1.1 Paraffinic bonds
Negative temperature coefficient
Many studies have discussed the auto-ignition behavior of paraffins, the main constituentsof commercial gasolines, and it is the most well studied group [48–61], good summary canbe found in [45]. Especially in recent years, more reactions are considered [62, 63] in themechanism to better understanding the oxidation process, especially in the low temperaturerange. Thanks to the development of quantum chemistry calculations, the revised kinetic andthermodynamics in the low temperature range can reproduce the experiments with betteraccuracy [64–66].
Based on the reaction mechanisms presented by Curran et al. [48, 67], the main reactionpathways for long chained paraffins or alkanes can be drawn schematically as is shown in
12
2.1. Introduction
R
H2C
HC
CH3
+ CH3
a: Unimolecular decomposition
R-CH2 +C2H4OH
R + C3H6OH
R
b: H-abstraction(+ OH, HO2, etc.)
RR
e: Recombination (+Q) f: β-scission
R CH2 + C2H3OHR R CHOH + CH3 + R
or other small oleffins and alkyl radicals
g: Isomerization
OH
R CH
OH
OHOH OH
or or
OH
orR
+H2O
h: Concerted elimination
c: O2 addition
R
Q
OH
R
d: Addition (+H)
OH
OO
R
OH
OOH
R
OH
O
g: Isomerization
i: Radical cyclization
+HO2
R
OH
OOH
+O2
j: Disproportionations
Figure 2.1: Simplified reaction scheme for hydrocarbon auto-ignition
Figure 2.2: Simplified reaction scheme for paraffin auto-ignition [48, 67]
At low temperatures, after the formation of first fuel alkyl radicals (R), H atom abstractionis the dominant initial reaction, since it requires a relatively low activation energy. The mostimportant reaction at low to intermediate temperatures is the oxygen addition reaction of alkylradicals to produce alkylperoxyl-radicals (RO2) (Reaction 2.1).
R+O2 RO2 (2.1)
Given an adequate pool of RO2 radicals, carbon-centered hydroperoxyalkyl radicals (QOOH)are formed by internal H atom migration via five-, six-, or seven-membered transition state (TS)
13
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
O
OO
O
O
OH
H2C
H
Figure 2.3: Possible pathway for RO2 intra-molecular isomerization to QOOH
R' R
OO OOH
R'
C
R
OOH OOH
R' R
OOH OOH
R'
C
R
OOOH
+OH R'
C
R
OO
+OH
(a: Chain branching)
R' R
O OOH
+OH R' R
O O+OH
(b: Chain branching)
R' R
OOH +HO2
R' R
O+OH
(c: Chain propagation)
Figure 2.4: Possible reaction pathways for O2QOOH [62]
rings . Figure 2.3 illustrates one possible pathway of isomerization for iso-octane, involving inthis case a 6-membered TS ring.
QOOH radicals subsequently react further with oxygen to form peroxy alkylhydroperoxideradicals (O2QOOH). O2QOOH radicals, in turn, tend to abstract H atoms from the C atomwhich attached to the OOH group, producing Q(OOH)2 radicals, ultimately yielding OHradicals and Carbonylhydroperoxide species via isomerization and decomposition reactions(reaction route (a) in Figure 2.4). Carbonylhydroperoxide decomposes to alkoxy and OHradicals. These reactions lead to the formation of active OH and carbonyl radicals, therebyaccelerating the overall reaction rate. These chain branching reactions then lead to rapid lowtemperature chemistry [68]. This is the main reaction pathway of the O2QOOH, because theC-H bond adjacent to hydroperoxyl group is significantly weaker than analogous C-H bondson a regular alkyl chain. Other sites are also possible for H atom isomerization reactions(Figure 2.4) in a manner analogous to QOOH decomposition [66], which will be discussedlater.
As the temperature increases, decomposition of QOOH radicals is favored over O2addition reactions. There are three main pathways for QOOH decomposition, as is illustratedin Figure 2.2. According to the different position of the abstracted H atom site, the productsvary from cyclic ethers and OH radicals to olefins and HO2 or carbonyl radicals. These areall chain propagation reactions. HO2 radical is quite stable at low temperature, so it caneasily accumulate and react readily with itself to produce H2O2. This reaction acts as aninhibiting reaction at low to intermediate temperatures [44]. Reaction 2.1 is very sensitive totemperature. The temperature at which the reverse reaction becomes important is relatedto the so-called ceiling temperature (i.e., temperature at which [RO2]/[R] = 1 [69]). Thisintermediate temperature, typically roughly 850 K, marks the border between NTC and low
14
2.1. Introduction
temperature chemistry.Therefore, at intermediate temperatures, around 850 − 1200 K range [70]), QOOH de-
composition reactions compete with low temperature O2 addition reactions. The dominantreaction route shifts from chain branching to propagation when temperature increases. This,together with the formation of stable products, slows down the overall reaction rate, resultingin a decrease of the reaction rate when temperature increases. This negative temperaturecoefficient or NTC behavior is unique to paraffins and other compounds having long paraffinicbranches. As temperature increases further, the NTC region is departed and Reaction 2.1favors the reverse direction. When the temperature increases above 1000 K, H2O2 readilybreaks down into two active OH radicals. These radicals accelerate the overall reaction, therebyreleasing more heat.
At high temperatures (T>1200 K), with unimolecular decomposition of hydrocarbonradicals fast becoming the dominant route, the reaction rate is controlled mainly by Reaction2.2 and fuel molecular structure becomes less important for auto-ignition chemistry [71].
H+O2 → O+OH (2.2)
Owing to low temperature branching reactions, paraffins react very fast at low tempera-tures, while in the intermediate range, as a result of NTC behavior, the reaction rate slowsdown. Said behavior is the primary reason why most paraffins have similar RON and MONvalues, given that both operating conditions yield reaction temperatures within the NTC range(Figure 2.5).
Temperature
Beyond MON (K>1)
MONRON
Low T NTC region High T
Beyond RON (K<1)Modern engine condition
Ignitiondelay time
Paraffin (S=0)
Aromatics S>0
Figure 2.5: Schematic representation of ON and paraffin auto-ignition temperatureregimes
Chain length and branching
At low to intermediate temperatures, alkyl radical isomerizations (Figure 2.3) are importantprecursors to subsequent chain branching reactions. Only transition state rings, sized within
15
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
a certain range, namelyfive-, six-, seven-membered, have an adequately low strain energy that is needed in order
for RO2 to form QOOH.Note that a hydrocarbon with a longer paraffinic chain will have more possible sites for
having such isomerization reactions, which explains why longer chains are generally morereactive and have commensurately short ignition delay times. The same logic holds for highlybranched paraffins, the comparatively slow reaction rates of which being the result of anintrinsic drop in average chain length for every branch.
Moreover, as will be discussed later, primary C-H bonds, located at the chain extremitiesare stronger than those found on C positioned within the chain. In Figure 2.6, which shows
3 4 5 6 7 8 90
20
40
60
80
100
120
Number of C
Oct
ane
Num
ber
n−paraffin
2−methyl paraffin
3−methyl paraffin
dimethyl paraffin
RON: solid linesMON: dash lines
Figure 2.6: Impact of paraffin chain length and branching on RON and MON [72]
the ON of different paraffins with various chain lengths, degree and location of branching, itcan be seen that the ON falls dramatically for longer chains.
Cycloalkane
Cycloalkane has lower reactivity and higher ON than their acyclic counterparts [73], e.g., theRON of n-hexane and cyclohexane are 24.8 and 83, respectively [72]. The cyclic saturatedstructure provide some distinct oxidation character to cycloalkanes [74]:
• Concerted eliminations producing oleffin compete with low temperature branchingreactions, due to the conformational inhibition of the isomerization from RO2 radicalsto QOOH.
• Unlike acyclic alkanes, alkyl substitution of cycloalkanes increases its low temperaturereactivity [74] by providing more H atoms and decrease the energy barrier for H atomisomerization, which can proceed low chain branching reactions.
16
2.1. Introduction
Yang et al. [74] also point out that because of the differences in cyclic structures, the energybarrier for internal isomerization varies, so that the reaction pathways and rates may change.Sirjean et al. [75] have compared cyclopentane and cyclohexane at high temperature (1230-1840 K) in shock tube at elevated pressures, cyclohexane is much more reactive than theformer one.
Lemaire et al. found NTC behavior for cyclohexane [76, 77] and methylcyclohaxane [78, 79]in RCM experiment at low temperature, and although the concerted elimination reactions arecompetitive with the cyclohexy oxidation, the NTC behavior indicates that the low temperaturebranching reactions are dominant in the low temperature for cyclohexane, while the NTC canonly be seen in fuel lean or stoichiometric condition [77]. Yang et al. [74] found that, at lowtemperature, the cyclohexanyl radical will have O2 addition, followed mainly by the concertedelimination reaction to produce cyclohexene. Alternatively, part of the peroxy cyclohexanecan undergo isomerization, and the chain branching reactions. Pitz et al. [80] have studiedthe methylcyclohexane combustion kinetic, also concluded that the chain branching and thechain propagation paths ratio and the concerted HO2 elimination reactions are very important.Silke et al. [81] have also added the concerted elimination reactions in their cyclohexanemechanisms, and found the ignition delay times are quite sensitive to this reaction type.
For cyclopentane (Figure 2.7), Al Rashidi et al. built the kinetic mechanism, and foundthat the fuel radicals mainly undergo β-scission from either C-C bond or C-H bond, producingolefin, and they found that between 850 to 1000 K in fuel rich conditions, there is an inhibitionin reactivity [82].
OO
O
β-scission at C-C bond
β-scission at C-H bond
Figure 2.7: Simplified reaction scheme for cyclopentane auto-ignition chemistry [82]
2.1.2 Olefinic bonds
Olefins have a higher RON than paraffins with the same C number. The principle olefinreaction scheme for auto-ignition is shown in Figure 2.8. The divergence from the paraffinicpathway originates from the fact that olefinic C bonds provide a site for addition reactions,involving O, H, OH or HO2 radicals.
17
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
At low temperatures, H atom abstraction tends to occur at allylic sites, producing reso-nance stable allyl radicals in the process (pathway (d) in Figure 2.8).
Resonance stable compounds, such as benzyl and allyl radicals, have a delocalized un-paired electron (Figure 2.9), which increases stability that inhibits subsequent reactions andthus slows down the overall reaction rate [85, 86].
R
CH2
R
CH
CH2
Allyl radical
CH2 CH2 CH2 CH2
Benzyl radical
Cyclopentadienyl radical
CH2
π bond
Figure 2.9: Examples of stable structures - allyl, benzyl and cyclopentadienyl radicals
The R-OH adduct formed by OH addition reactions can then undergo O2 additionreactions and, subsequently, decompose via the Waddington mechanism, forming OH radicalsand aldehydes (pathway (a)).
It can also be consumed via H atom or HO2 radical addition reaction, the former followedby β-scission, producing an olefin and alkyl radical (pathway (b)). The latter can then lead tothe formation of an cyclo-ether and OH radical (pathway (c)) [87].
Alternatively, decomposition occurs via concerted HO2 elimination reactions, whichresults in the formation of another double bond [83] (pathway (e)). OH addition reactions(pathway (a)) are the preferred decomposition route at low temperatures, owing to the relativelylower energy barrier [88–90]. Note that pathways (a-c, e) are chain propagation reactionsthat produce stable radicals and compete with low temperature branching reactions (pathway(d)), slow down the reaction rate [83]. These addition reactions are more effectual when the
18
2.1. Introduction
paraffinic chain lengths is short, since there are only a few paraffinic H atom sites to facilitateRO2 radical isomerization [86].
This is consistent with Vanhove’s results [89], who reported that the position of the doublebond has a strong impact on olefinic auto-ignition chemistry in the low to intermediatetemperature regime. Long chained olefins (e.g., hexene [89]) that have the double bond onthe extremity (e.g., 1-hexene) tend to display paraffinic-like NTC behavior (Figure 2.2).
Conversely, when the paraffinic chains are short, as is the case for 3-hexene, NTC behavioris virtually non existent [89]. As the temperature increases, unimolecular decompositionsbecomes dominant (pathway (d)) and the structural impacts (e.g., position of double bond,chain length, branch level) become less pronounced. However, because the allylic site isthe weakest link, C-C bond cleavage also tends to occur at this site (Figure 2.8). As theunimolecular reaction prefers to transfer along route (d), the position of the double bond willalso affect the high temperature auto-ignition chemistry [83].
2.1.3 Benzenoid bonds
Aromatics are the second most common constituents in gasoline, comprising up to 30% ofthe fuel. The combustion chemistry of aromatics, particularly of benzene and toluene, hasbeen studied extensively [91–104]. The benzene ring that forms the core of all aromatics has achemistry very distinct from paraffins and olefins.
For example, the ring provides a site for electrophilic substitution reactions, whereby Hatom on the ring can be replaced by another radical, such as CH3 and OH radicals. This kindof reaction competes with H atom abstraction from both the ring and its side chains, whenpresent [91, 105].
Owing to the highly stable π bond in the ring (Figure 2.9), it is difficult to abstract an Hatom from aromatics. Accordingly, most aromatics are highly resistant to auto-ignition at lowto intermediate temperatures [91]. Once the ring opens, however, bond energies change. As isshown in Figure 2.10, the bond dissociation energy (BDE) of C-H bonds in benzene decreaseonce H atom is abstracted.
H112.5+0.5
H
H
H
79.9+3.1
95.3+3.2
110.6+3.4
Benzene
Phenyl radical
88.5+1.5
102+1
Toluene
H2C
H
OH
119.3
Furan
OCH2
112.9
2-Methyl furanH
119.5
H85.3
H119.3
OCH2
113.4
2,5-Dimethyl furan
H84.8
H3C119.3
H
119.6
Figure 2.10: BDE’s for C-H bonds in benzene, toluene and furans (kcal/mol) [106–108]
Benzene auto-ignition chemistry generally follows one of two pathways (Figure 2.11):
• Unimolecular decomposition and H atom abstraction reactions, forming phenyl radi-cals that, when reacting with HO2 radical, O2 or O atom, produce phenoxy radicals.
19
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
• Direct oxidization to phenoxy radicals, which are also stable and go on to produce mainlyphenol, which reacts back to phenoxy radicals in the low to intermediate temperature.
Phenoxy radicals, in turn, can decompose to CO and cyclopentadienyl radicals at hightemperature, the latter of which being an important intermediate for later ring openingreactions that ultimately yield butadienyl radicals (C4H5).
O
OH
O
C4H5+COHO2, OH
electrophilic
substitutio
n
H-abstraction
Cyclopentadienyl
OO
Figure 2.11: Simplified reaction scheme for benzene auto-ignition [91]
Contrary to paraffins and long chained olefins, benzene displays neither low temperaturebranching reactions nor NTC behavior. Aromatic radicals, such as benzyl (Figure 2.9) andphenoxy, are resonance stable and thus inhibit subsequent reactions, decelerating the overallreaction rate. The high ON of most short chained aromatics, notably toluene, 1,3,5-tri-methyl-benzene and xylene (Table 1.5), owes much to this stability.
For aromatics with longer side chains, initial decomposition steps are determined chieflyby side chain chemistry [91, 109]. A longer side chain or multiple side chains providemore possible sites for low temperatures branching reactions [38, 110–114], which is why thereactivity of aromatics tends to increase with the degree of alkylation.
Toluene, a commercial octane booster, has one methyl group on the ring. Its RON andMON are 120 and 109, respectively, indicating that it has both a good anti-knock quality and ahigh sensitivity S [37]. At low to intermediate temperatures, its main reaction scheme can besummarized by Figure 2.12.
Auto-ignition commences with unimolecular decomposition and later H atom abstraction,producing either resonance stable benzyl radicals and H atom, or phenyl and methyl radicals.The C-H bond in the methyl branch is the weakest bond (88.5 kcal/mol), making it the mostlikely site for these initial reactions.
Subsequently, benzyl radicals may go on to react with HO2, O2, O, or OH radicals toproduce benzoxy radicals. This is followed by H atom abstraction, ultimately producingphenyl and CO.
Alternatively, benzyl radicals can react with each other to form bi-benzyl [96, 115, 117]. Athigh temperature, yet another initial reaction can occur, this time O atom, thereby producingcresoxy radicals and H atom.
20
2.1. Introduction
C
H
OCH2O
O
OH OH OH
O
Benzaldehyde
Cresoxy
Cresol
Benzoxy
Hydroxybenzyl
Bibenzyl
Hydroxy benzaldehyde
Phenoxy
Hydroxy benzoyl
Benzyl Phenyl
OO
OH
OH
OH OH
OH
O
CHOCO
C4H4O98.4%
Cyclopentadienyl
Figure 2.12: Simplified reaction scheme for toluene [115, 116]
2.1.4 Furanic bonds
Furans as one kind of aromatics, their cyclic structure is also of the sp2 hybridization type,sharing a π electron cloud (π bond). Recently, furans including 2-methyl furan (2-MF) and2,5-dimethyl furan (2,5-DMF), were reported to be promising octane boosters [118, 119], andtheir cetane numbers (CN) are 8.9 and 10.9, respectively [120]. By comparison, commercialRON 95 gasoline has a CN close to 20. The lower CN values of the furans underlinesaforementioned octane booster claim.
A simplified furan reaction scheme is shown in Figure 2.13 [121]. Its unique five-membered heterocyclic ring manifests in a particularly high C-H BDE. This can be tracedback to the low thermodynamic stability of the radicals formed after H atom abstraction [122].Consequently, furans have an even stronger C-H BDE than benzene (Figure 2.10).
Furan auto-ignition chemistry initiates primarily via H atom or OH radical additionreactions at the Cα position [121] (pathway (a) and (c) in Figure 2.31). This leads to theformation of resonance stable dihydrofuryl-3 radicals and C5H4OH, respectively, followed byβ − scission reactions. Other, less dominant, pathways involve H atom addition at the Cβposition (pathway (b)) or H atom abstraction (pathway (e) and (f)).
Alkylated furans, such as 2-MF and 2,5-DMF, have been the subject of many studies [108,123–126]. The BDE of methyl C-H is relatively low. Accordingly, in addition to aforementionedaddition reactions, H atom abstraction from the methyl group is a likely initial decompositionstep.
2-MF, for example, can react via H atom or OH radical addition (pathway (b) in Figure2.14), H atom abstraction (pathway (c)) or H atom ipso addition (pathway (a)) [123]. Given fuelrich conditions, decomposition via ipso addition, yielding furan, is the dominant pathway[123].
21
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
O
O
O
O
O
CH
OOH
H
O
CO +
O O
CH2
O +CHO
O
O
C2H2 HC O
O
CH3
O
+CO
+ CH3
+H
H atom addition at C_α
H atom addition at C_β
OH atom addition at C_α
H atom abstraction at C_α
α
β
H atom abstraction at C_β
(a)
(b)
(c)
(d)
(e)
(f)
Figure 2.13: Simplified reaction scheme for furans [121]
O
2-Methyl furan
O
CH
H2C
OCH3
OCH2
+CH3
C
HC CH
H2C
OCH3
CO+ HC
OCH2 C
HC CH
HC
OCH2
resonance structure
resonance structure
HC+ CO
Ipso addition (a)
H atom addition
H atom abstraction
(b)
(c)
Figure 2.14: Simplified reaction scheme for 2-methyl furan [123]
22
2.1. Introduction
Table 2.1: Classification of functional oxygen groups
Class name Functional groupAlcohol R−OH
Ether R−O− R ′
Ester R(= O)O− R ′
Aldehyde/Ketone R− C(= O) −H(orR ′)
Carbonate R−O− C(= O) −O− R ′
2.1.5 Oxygenated bonds
Oxygenates, including MTBE, ETBE and ethanol, are amongst the most common commercialoctane boosters. The presence of fuel oxygen in these compounds has a significant influenceon the auto-ignition chemistry:
• Reduced symmetry yields more possible reaction pathways (e.g., dehydrogenation,dehydration, decomposition). For example, intermolecular dehydration (concertedelimination) of 2-butanol (Figure 2.16) produces olefins and H2O [127].
• Reduction of the BDE of C-H bonds at the adjacent α-C site (Cα −H) or β-C site if itis a ketone, while increasing the BDE of C-H bonds at the β-C site (Cβ −H) or γ-sitefor ketone (Figure 2.17), promotes H atom abstraction at the former site [128]. Forexample, in the case of iso-pentanol oxidation (at 800 K, 15 bar and stoichiometricmixing conditions), nearly 48% of it reacts at this type of C-H bonds by H atomabstraction, and produce α-hydroxypentyl radicals [129].
The high activation energy required for hydrocarbon or oxygenate unimolecular decom-position reactions are readily met at high temperatures. These causes the identity of thefuel, with respect to auto-ignition chemistry, to manifest mainly in the low to intermediatetemperature regime. Categorized by functional oxygen group, various types of oxygenated areshown in Table 2.1.
Alcohols
Owing to their high knock resistance, alcohols have been considered superior alternativesto gasoline since the advent of the gasoline engine. Ethanol in particular is commerciallyblended to gasoline in both the US and EU.
The presence of a hydroxy group on the chain affects otherwise paraffinic chemistryscheme in various ways:
• H atom abstraction favors Cα −H sites, forming α-hydroxypropyl radicals [130].
• Hydroperoxyalkyl (RO2) radicals can now, besides isomerization to QOOH and sub-sequent low temperature branching reactions, also react via three other routes thatcompete with the chain branching reactions. Two reactions are HO2 concerted elimina-tion for α-hydroperoxyalkyl radicals [131], including one via a 5-membered transitionring, forming enol and HO2 (Figure 2.15), the other is α-hydroperoxyalkyl directly
23
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
form carbonyl and HO2. The other one is a Waddington type mechanism [132] forβ-hydroperoxyalkyl, which involves H atom transfers from the C-OH to Cβ-OO radicalsite, thereby producing an aldehyde and an OH radical.
CH
OH
HC
HC
OH
OO
H
OH + HO2
Enol
Concerted elimination
HC
CH2
OHO + CH2O + OH
Waddinton type reaction
α−hydroxypropyl
OO
HC
O
O
H
O
H2C
β−hydroxypropyl
H2C
HC
O
OO
+ O2
H
O + HO2
Figure 2.15: 5-membered concerted elimination and Waddington mechanism forpropanol
• Dehydration reactions, occurring at high temperature and yielding H2O [129] (Figure2.16).
OH
H3CCH2
HC
CH2
O+ H2O
HH
Figure 2.16: Dehydration of 2-butanol
Auto-ignition in butanol isomers, for example, is initialized by unimolecular reactions,followed by H atom abstraction. The thus formed fuel radicals go on to react with O2 andproduce RO2 or enol. This is swiftly followed by RO2 radicals undergoing low temperaturebranching or Waddington type reactions ([130, 133]). Accordingly, low temperature reactivityof alcohols is greatly reduced by the presence of the OH group. At high temperatures,unimolecular decomposition, including dehydrations, becomes the dominant route [133]. Thefour studied butanol isomers have divergent reactivities.
For instance, 1-butanol has a much shorter ignition delay time in rapid compressionmachine experiments compared to other isomers. 2-butanol and tert-butanol have similarignition delays, while iso-butanol yields the longest delay [134].
24
2.1. Introduction
In shock tube ignition experiments carried out at 715−910 K and high pressure (15 and 30bar), however, tert-butanol had the longest ignition delays, with the other isomers displayingsimilar combustion behavior [130].
The variation in reaction rates results primarily from structural effects, which favor one ofvarious hydroxybutyl radical reaction pathways. Moreover, C-H BDE depends on the locationof the OH group (Figure 2.17), resulting in a unique H atom abstraction rate for each isomers.
H2C
CH
HC
CH
O
HH
H
H
H104.1
95.5
100.6
98.8
101.9H
H2C
C
CH
CH2
H
O
H
H
102102.6
H
94.7
104.1
100.8
2-Butanol1-Butanol
H
CH2
C
CH
O
H
H2C
H
101.8
101.8
H97.6
H95.9
104.1
iso-Butanol
H
CH2
C
CH2
H
H2C
O
H
H
103.9
104.1
103.9
103.9
tert-Butanol
α
βγδ αβ β
γ
α
αβ
β
β
βγ
γ
Figure 2.17: Bond dissociation energy for various butanol isomers (kcal/mol) [130]
As shown in Figure 2.18, α-hydroxybutyl, following O2 addition, mainly reacts via internalisomerization that H atom from the hydroxy group shifts to peroxy group. Subsequent β-scission produces aldehydes and HO2 radicals (pathway (a)) or via concerted elimination,directly form aldehydes and HO2 radicals [135]. The latter pathway is the dominate.
β-hydroxybutyl can react with O2 and subsequently decompose by way of low temperaturebranching (pathway (c)), propagation (pathway (b) and (d)) reactions. In the low temperatureregime, however, the propagation reaction routes are the dominant ones, given the higherenergy barrier in the alkyl radical isomerization needed to produce QOOH [133].
It is has been calculated that nearly all β-hydroxypropyl radicals react by Waddingtontype reactions [136]. α-hydroxybutyl radicals mainly lead to propagation reactions (pathway(a)), except in the case of tert-butanol. γ- and δ-hydroxybutyl radicals typically react via lowtemperature branching or termination reactions.
Given that 1-butanol has more paraffinic C bonds (i.e., γ-C and δ-C), the main sitesat which branching reactions occur, it has a relatively high reaction rate. Iso-butanol, bycomparison, having 6 γ-C which is located on high BDE primary sites and no δ-C, has a farslower reaction rate. This can be explained by the slower H atom abstraction rate, therebyfavoring propagation reactions. Consequently, a branched alcohol with a less paraffinic chain
25
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
OH
OH
OH
OH
OH
OO OH Low T branching
OH
OO
O+ HO2
OH
OOO
+ CH2O + OHWaddington
+ HO2
QOOH Low T branching
OH
OO Low T branching OHEnol
Enol + HO2
1-Butanol
β−hydroxybutyl
α−hydroxybutyl
δ−hydroxybutyl
γ−hydroxybutyl
OH
OH
OH
OH
O + HO2
OH
OO
Enol + HO2
O
Enol + HO2
OHOO
Iso-butanol
α−hydroxybutyl
β−hydroxybutyl
γ−hydroxybutyl
+ O2
+ O2
αβ
γδ
αβγ
+ O2
+ O2
+ CH2O + OH
WaddingtonLow T branching
Low T branching
QOOH
QOOH
+ O2
+ O2
(a)
(b)
(c)
(d)(e)
(f)
(g)
(a)
(d)
(b)
(c)
(f)
(e)+HO2 + CH2O
(f)
1-Butanol simplified reaction scheme
Iso-butanol simplified reaction scheme
Figure 2.18: Simplified reaction scheme for 1-butanol and iso-butanol at lowtemperature[133]
26
2.1. Introduction
(less γ, δ or further C) will have longer ignition delay and therefore be the more preferredoctane booster.
Ethers
Ethers contain two alkyl or aryl groups bonded to an oxygen atom. The presence of C-O bondsdecreases the BDE of Cα −H sites [137]. Alkyl ethers undergo similar reaction pathways asalkanes [138–141], albeit at far lower energy barriers for abstraction or isomerization at Cα −Hsites, resulting in an increased overall reaction rate. Dimethyl ether (DME), for example, hasa RON of 35, while that of propane is 112.
Some ethers, conversely, including common octane boosters as MTBE, ETBE or TAME,have a remarkably high RON of 117, 118 and 112, respectively [142]). This impressive anti-knockquality can be attributed to the highly branched tertiary butyl group attached to the ethergroup (Figure 2.19).
The presence of this group dramatically increases the number of primary H bonds.Moreover, this particular structure lacks any paraffinic C bonds (no γ-C). As a result, Hatom abstraction and RO2 internal isomerization reactions becomes both less prevalent andeffective.
O
O
O
MTBE ETBE TAME
Figure 2.19: Molecular structures of esters commonly used as octane boosters
Moreover, O2 addition, the main reaction route at low temperatures, become less im-portant, as initial decomposition is more likely at the relatively weak C-O bond [143, 144].Consider for example MTBE, the auto-ignition chemistry of which takes one of the followingreaction pathways (Figure 2.20):
O
CH3OH+
O+ H2
O+ H2
+ CH3O
+ CH2O
Four-membered elimination
H-abstraction
(a)
(b)
β-scission
(c)
α
ββ
β
Figure 2.20: Simplified reaction scheme for MTBE [143]
27
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
• H atom abstraction from Cα − H sites, followed by β-scission, producing aldehydesand iso-butyl radicals (pathway (a)).
• H atom abstraction from the branched iso-butyl group, followed by β-scission, therebyforming iso-butene and CH3O radicals (pathway (b)).
• At high temperature, a four-membered elimination reaction, producing isobutene andmethanol (pathway (c)).
Bond cleavage involves higher activation energies than addition reactions. Hereby, theiso-butene chemistry greatly influences the overall reactivity, as it has a strong inhibiting effect[145, 146]. Accordingly, such ethers have a high resistance towards auto-ignition [147].
Lack of low temperature branching reactions and the formation of stable intermediates(e.g., iso-butene) are the two main factors responsible for the characteristically high ON ofaforementioned highly branched ethers.
Esters
Esters harbor a C(=O)-O functionality (Figure 2.21) and generally have a lower reactivity thancorrespondingly long alkanes. The presence of the C(=O)O group weakens adjacent C-Hbonds (Figure 2.21), thus making α-C the preferred site for H atom abstraction [148].
Esters, notably fatty acid methyl ester or FAME (CH3(CH2)nCOOCH3) and fatty acidethyl esters or FAEE (CH3(CH2)nCOOCH2CH3), can be produced from various plant oilsand are the main constitutes found in biodiesel today [149].
Less common esters found in fuels include methyl butanoate (MB) [148–152] and ethylpropanoate [153, 154]. Other than aforementioned esters are studied as well to better com-prehend the chemistry of FAME and FAEE, such as ethyl pentanoate [155], methyl and ethylhexanoate [156–158], methyl heptanoate[156, 159], methyl decanoate [156, 160] and a mixtureof biodiesels [161, 162], Coniglio et al. [163] reviewed the combustion chemical kinetics ofmethyl and ethyl esters. Although huge progress has been made recent years, it is found thatthe modeling results still are over-predict the reactivity compared to the experiments in thelow temperature range [164], therefore, further studies are needed to improve the accuracy ofthe model.
H2C O
CH2
O
H
101.3
86.8
H100
89.9 99.2
85.2
88.9
H
98.5 H93.6
αβ
Figure 2.21: BDE of methyl butanoate (kcal/mol)
Esters display a paraffin-like reaction chemistry, however, no NTC behavior is found forshort paraffinic chain esters, as is exemplified by the MB scheme, shown in Figure 2.22[152]. The paraffinic chain of MB (C3) is relatively short. This attribute, combined with thepresence of an ester group, results in O2 addition reactions forming alkylperoxy radicals thattend to prefer undergo concerted elimination reactions yielding unsaturated ester, the low
28
2.1. Introduction
MB
O
O
O
O
O
O
O
O
Low T branching
β-Scission
O
O
+O2(a)
O
O
OO
O
O
+ CH3
(b)
Concerted elimination
+O2
(c)
Low T branching& β-Scission
Concerted elimination, Low T branching& β-Scission
Figure 2.22: Simplified reaction scheme for methyl butanoate at low temperature[152]
temperature branching reactions become less important [152]. This helps to explain the lowreactivity of short chained esters in the low temperature regime.
To decelerate matters further, due the presence of a carbonyl group (C=O), the alkoxyradicals formed after H atom abstraction at β-C site are resonance stable (Figure 2.23).
Esters having a paraffinic chain with more than five members are prone to NTC behavior,as has for example been observed in RCM measurements for methyl hexanoate [157]. Owing toNTC chemistry, low temperature branching reactions are important decomposition pathwaysfor larger esters.
O
O
O
O
Figure 2.23: Resonance stable structure of methyl butanoate radicals [149]
While the ester functionality provides an attractive site for H atom abstraction, it alsoenables another reaction path, involving concerted elimination through a six-memberedtransition state ring (Figure 2.24), producing olefins, small esters or acids.
Whether or not six-membered elimination, as shown in Figure 2.24, is the primaryreaction pathway depends on the structure of the ester. For example, MB and ethyl propionate(EP), an equal C number notwithstanding, have distinct reactivities, which the latter being thequicker of the two.
One of the reasons for this divergence is that it is easier for EP to undergo concertedelimination reactions and produce reactive species [153], given that the activation energy forthis reaction is 50 kcal/mol for EP versus 68.07 kcal/mol for MB [154].
Walton et al. [163] found that 96% of EP follows the elimination route, with only 4%
29
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
O
O
O
O
O
CH2
O CH2
H
OH
O
+C2H4
Ea=50 kcal
H2C
O
OH2C
H
O
OH
+C2H4
Ea=68.07 kcal
EP
MB
Figure 2.24: Six-membered elimination reaction of methyl butanoate and ethylpropanoate [154]
reacting via H atom abstraction at 1 atm and 1200 K [153]. Elimination chemistry for MB,conversely, is rare, contributing less than 1% to decomposition.
Schwartz [165] studied five isomers of C5H10O2 esters at atmospheric pressure in aCH4/air co-flow flame and reported that six-membered elimination reactions are the domi-nant reaction pathway for EP, propyl acetate and iso-propyl acetate. Conversely, for methylbutanoate and methyl iso-butyrate, the primary route is via simple fission (Figure 2.25).
Iso-propyl acetate is the most reactive isomer because the branched C atoms allow formore H atoms to undergo six-membered elimination. This is inconsistent with the viewthat H atom abstraction is the dominant decomposition pathway, since there are manyprimary H atoms in isopropyl acetate, however consistent with aforementioned six-memberedelimination observations [165].
Accordingly, six-membered elimination is the preferred decomposition route when thealkyl chain is large on the O-alkyl side of the ester. What is more, the branching effect reducesthe activation energy needed for the six-membered transition state ring [166].
Aldehydes/ketones
Aldehydes and ketones are important intermediates and end-products in hydrocarbon com-bustion [167, 168]. This oxygenate type is characterized by the presence of a carbonyl groupon the carbon chain. In the event said group involves C bonded to at least one H atom, theresulting compound is classified as an aldehyde, else, as a ketone.
Ketones (e.g., 4-heptanone) are potential biofuel candidates, which have been reported tobe produced via endophytic fungal biomass conversion or by other methods [169].
Aldehydes can be important diesel engine emissions, particularly from those enginesoperating in premixed auto-ignition combustion modes. These modes include premixedcharge compression ignition (PCCI) and homogeneous charge compression ignition (HCCI),both of which are regarded as high efficiency clean combustion (HECC) strategies [168, 170,171].
30
2.1. Introduction
O
O
Methyl isobutyrate
O
O
Propyl acetate
O
O
Isopropyl acetate
O
O CH
H CH3
+
O
O CH2
H
O
OH
O
OH
+
C-O fission
O
O
+CH3
O
O
H atom abstraction
six-membered elimination
six-membered elimination
Figure 2.25: Propyl acetate, iso-propyl acetate and methyl iso-butyrate reactionschemes
O
O
O
ODecomposition orH-abstraction
DecompositionCH3 +
O
CH3 +O
C2H5 + O
High T
Low and high T
O2 additionO
OO
DecompositionSmall ketenes, alkyl radicals
Low T
Figure 2.26: Simplified reaction scheme for 2-butanone [167, 177]
Alternatively, aldehydes emissions arise when burning gasoline that contain alcohols inconventional SI engines. Fortunately, the majority of which decompose readily in modernthree-way catalysts [172, 173].
At high temperature, as shown in Figure 2.26, the main reaction pathways for small ke-tones is H atom abstraction or unimolecular decomposition, followed by β-scission, producingsmall hydrocarbon radicals [174–176].
At low temperatures, fuel radicals can react via O2 addition, leading to the formationof hydroperoxyl radicals [167] and another O2 addition to form OOQOOH. Subsequentisomerization for H atom transfer, however, is inhibited by the modest low temperaturereaction rate [178].
Aldehydes such as propanal and butanal are very reactive compared to other compoundswith similarly sized carbon chains. Their main decomposition pathway in the low temperatureregime is similar to that observed for paraffins [168] (Figure 2.27).
The main differences for aldehydes versus paraffins, with respect to auto-ignition chemistry,
31
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
C
O CO +
OLow T branching
(a)
αβγ
O
O
O
n-C3H7CO
+O2
O
δ
C
O
O
+H
+CH3
OO
O
OOH
(b)
(c)
(d)
C3H6 + CHO
C2H4 + CH2CHO
+O2
Figure 2.27: Simplified reaction scheme for n-butanal [168]
CH
C
O
αβ
γδ
84.682.389.1
H
H
91.9 89
n-Butanal
CH2
HC
H2C
O
H
H
H95.6
101.5
90.8
2-Butanone
Figure 2.28: Bond dissociation energy of n-butanal and 2-butanone (upper numberis for C-H bonds, below is for C-C bonds (in kcal/mol) [177, 179]
are itemized below for the n-butanal case.
• The primary H atom abstraction site is at the Cα − H bond: the weakest C-H bond(Figure 2.28).
• The α-scission or β-scission reactions are competitive to the O2 addition reactions andthe following low temperature branching reactions.
• α-scission of n − C3H7CO radicals yields propyl radicals and CO (pathway (a)) inFigure 2.27) and has a negative effect on the overall reactivity, as it does not lead toparaffinic-like low temperature branching reactions [168].
Both propanal and n-butanal display NTC behavior. Iso-butanal, conversely, does not,owing to its short, highly branched chain that instead favors beta-scission over O2 addition.
At high temperature, pathway (a) (Figure 2.27) becomes the dominant route. For example,100% of n − C3H7CO radicals decompose this way at 1320 K and 3 atm in shock tubeexperiments [179]).
carbonate
Carbonates are also important fuel additives, both dimethyl and diethyl carbonate (DMC andDEC) can be used to reduce soot formation in diesel engine [180–182], many of them havehigh RON, it is reported that DMC, DEC and dipropyl carbonate have RON of 125, 110 and110, respectively [181]. Glaude et al. have build the kinetic model of dimethyl carbonate andvalidated in an diffusion flame [183], Nakamura et al. have studied diethyl carbonate (DEC)
32
2.1. Introduction
O O
O
αβ
αβ
O O
O
O O
O
O OH
O
+ C2H4
O O
OOO
O O
O
OO
Chain branching reactions
(a)
(b)
(c)
Figure 2.29: Simplified reaction scheme for diethyl carbonate in low temperature[184]
oxidation kinetic, and validated the auto-ignition delay time and the intermediates in a jetstirred reactor [184].
The three oxygen atoms in the compound provides some distinctive characteristics. Shownin Figure 2.29, DEC mainly undergoes H atom abstraction at Cα sites, followed byO2 addition.The peroxy radicals can have concerted elimination reactions like esters, forming a doublebond and anHO2 radical, or isomerization reactions through 8- or 9-membered ring transitionstates, and then the low temperature branching reactions. However, at low temperature, mostof the peroxy radicals react back to the α-fuel radicals [184]. Some of the DEC will produceβ-fuel radicals, then the O2 addition. The α-peroxy radicals, can undergo isomerization via5- or 9-membered ring transition state, producing QOOH radicals. The third pathway isdecomposition mainly into ethylene and ethyl formic acid. The RCM experiments showednear NTC behavior of DEC in low temperature at high pressure [184].
Benzenoids
A recent book by Boot et al. [185] on the performance of various types of oxygenates in bothcompression ignition and SI engines, found that all reviewed aromatic oxygenates shared alow CN and commensurately high RON. Moreover, their aromatic oxygen-laden structurelends itself well for production from lignin, a phenolic polymer found in residual streams ofpaper pulping and cellulosic ethanol plants [185].
As was the case for the foregoing oxygenates, adding O atoms to aromatic branches willdecrease the BDE of side chain C-H bonds. Oxygenation, though, will hardly affect the BDEof aromatic C-H bonds. The initiation of auto-ignition chemistry, given the stable benzenebase central to all aromatics, typically occurs on the side chains when present.
When O atom connects to the phenyl group, the strong phenylic C-O bond weakens theC bond on the alkyl side. Anisole, a phenyl methyl ether, has an ether group attached to thebenzene ring. Herein, the O− CH3 bond is the weakest link and thus the most likely site forinitial reaction chemistry, producing phenoxy and methyl radicals [186].
Alternatively, H atom abstraction can occur from the methyl group, producingC6H5OCH2radicals (Figure 2.30 [186]).
For aromatic oxygenates that do not have O atom bonded directly to benzene ring, theimpact of the phenyl group on overall reactivity is dependent on the length of the side chain.
When the side chain is short, as is the case for benzaldehyde, the Cα −H bond is weakand the dominant pathway is H atom abstraction, producing C6H5CO. This is followed up by
33
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
OO
CH2 O O
O
Phenoxy
Benzaldehyde Phenyl
OH OH
O
OH O
CO2 + C4H4CO + C4H5
CO
CO eliminations
Figure 2.30: Anisole reaction scheme [186]
C H
H
H
88.5+1.5102+1
O
C
H
H
H
102.6
63.2
96.9
O
C
H
H
H
56.3
O H
85.1Toluene Anisole Guaiacol
Figure 2.31: Bond dissociation energies for toluene, anisole and guaiacol [106, 186](kcal/mol)
CO elimination at high temperature, leading to phenyl.Therefore, as mentioned earlier, more or longer side chains will have an acceleratory effect
on the overall reaction rate. Consider guaiacol for example. Here, the presence of a hydroxygroup weakens the BDE of the O− CH3 bond (Figure 2.31) and also provides possible site toproceed the reaction.
2.2 Design rules for future octane boosters
Auto-ignition chemistry can be divided into three consecutive categories of reactions [187]:
• Initial→ formation of the first radicals.
• Intermediate → maintaining and/or increasing radical pool by propagation and/orchain branching reactions.
• Terminal→ recombination of radicals to stable compounds.
The overall reaction rate is controlled primarily by chain branching, which can rapidlyaccelerate the reaction by filling the radical pool. The rate of chain branching, in turn, isa function of many factors, including molecule structure, reaction temperature, pressure,equivalent ratio and initial concentration of fuel and air. A sufficiently large radical pool isthe critical factor that determines whether or not a fuel will auto-ignite. When the initialconditions in the combustion chamber are fixed, the overall reaction rate is determined byfuel molecular structure:
• Initial→ type, location and number of C-H bonds
• Intermediate→ hydroperoxyl radical (RO2) chemistry (propagation and branching).
• Terminal→ number and reactivity of intermediate species (termination).
34
2.2. Design rules for future octane boosters
100.2+198
n-Heptane
CCH2
CH3
HH
100.2+1
95.7+0.7iso-Octane
88.5+0.5
87.3+0.9
n-Butene
98.1CH
H2C
H
H
H
HC
CH
HC
CH2
H111.1
174.1+1.5
H83.8
Figure 2.32: C-H and C-C BDE (in kcal/mol) for various hydrocarbon types (red:primary C-H bonds; blue: secondary C-H bonds; green: tertiary C-H bonds) [106]
Generic design rules to inhibit aforementioned reaction types will now be presented insections 2.2.1, 2.2.2 and 2.2.3, respectively.
2.2.1 Strong carbon-hydrogen bonds
As decomposition reactions tend to initiate at the weakest bonds, BDE mapping will revealthe most vulnerable sites. At low temperature, H atom abstractions are the most importantreactions followed the initiation reactions, which include unimolecular decomposition orbimolecular reaction with O2. The BDE of C-H decreases in the order of primary > secondary> tertiary C-H bond (Figure 2.32).
The relative weakness of tertiary over primary bonds, and the number of available C-Hbonds of each types explains at least in part the significant spread in reactivity observed fornormal and highly branched paraffins.
Consider n-heptane and iso-octane for example. As is shown in Figure 2.32, n-heptane,with 10 secondary and 6 primary C-H bonds, contrasts greatly with iso-octane, having 15primary, 2 secondary and 1 tertiary C-H bonds. The latter will have far fewer reactive C-Hbonds than the former, thus providing an explanation for the associated 0 and 100 RON/MONvalues, respectively.
The BDE of C-H bonds is also affected by the presence of oxygen functionalities adjacentto the carbon atom in question. Owing to the tendency of most oxygen functional groupsto attract electrons from neighboring atoms, neighboring C-H bonds are commensuratelyweakened, promoting initiation and the following metathesis reaction (H atom abstraction)chemistry. Because of the adjacent O atom, the following reaction may switch from O2addition reactions to other reactions, e.g. concerted eliminations, β-scissions, thus changingthe reaction pathways.
For example, the BDE of the Cα−H bond found in methyl butanoate (MB) is 93.6 kcal/mol.This is noticeably lower than the 98.5 kcal/mol seen for Cβ−H bonds. Accordingly, the formerwill be the preferred site for H atom abstraction. Then the followed O2 addition reactionwill also prefer to be here. However, the energy barrier for the formation of the transitionstate ring for isomerization will also be affected, so that the isomerization and the branchingreactions will not be dominant later on. Instead, the concerted elimination or the β-scissionare the main consumption pathway, which are chain propagation reactions. Therefore, theBDE of C-H bonds will also affect subsequent reaction pathways, resulting in low reactionrates.
In the presence of an olefinic bond, the Cβ −H to the double bond have lower BDE (83.8kcal/mol) and, when broken, lead to the formation of resonance stable allyl radicals (Figure
35
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
2.8). Moreover, RO2 at an allylic site decomposes readily back to O2 and an R radical, in effectinhibiting the chain branching process [88]. Both effects contribute to the high ON of olefinsrelative to paraffins (Figure 2.33) owes much to this stability). Besides H atom abstraction andunimolecular decompositions, there is another initial reaction path for double bonds at lowtemperature: the addition reaction. It competes with the low-temperature branching ones[83]. For example, the addition of OH radicals at the double bond site allows for Waddingtontype reactions (Figure 2.8), which becomes the dominant reaction pathway. This results in theformation of more stable double carbon and vinyllic C-H bonds, suppressing low temperatureinternal H atom migration. From Figure 2.33, it can further be seen that olefin sensitivity
3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.50
10
20
30
40
50
60
70
80
90
100
110
Number of C
Oct
ane
Num
ber
n−paraffin1−olefin
2−olefin
3−olefin
RON: solid linesMON: dash lines
Figure 2.33: Impact of carbon number and double bond site on olefin reactivity [72]
decreases with carbon number as it becomes of an increasingly saturated nature.
2.2.2 Short chain length
RO2 radicals produced by Reaction 2.1 are very important intermediates at low temperature[44]. Both alkyl and peroxyl radicals can undergo intra H atom transfer via a transition statering to produce QOOH, which dominates low temperature chain branching.
Competition between isomerization and decomposition of RO2 or QOOH affects thereaction route, as well as the reaction rate. Competition between decomposition and O2addition reactions has a negative and active impact on increasing the whole reaction rate,respectively. The main factor driving the branching ratio is temperature.
O2 addition reactions are important because of their negligible energy barriers. However,as temperature increases, RO2 radicals become thermally unstable and readily decomposeback to fuel radicals.
36
2.2. Design rules for future octane boosters
The formation of the transition state ring that produces QOOH is integral to low temper-ature chemistry, because from here it can undergo another O2 addition reaction, followedby chain branching reactions, thus increasing the whole reaction rate. Alternatively, it canundergo decomposition reactions, producing ethers, ketones or other species via propagationreactions, which has no acceleration effect.
The activation energy for isomerization reactions is defined as [67]:
Ea = ∆Hrxn + ringstrain+ Eabst (2.3)
In this equation, ∆Hrxn is the enthalpy of the endothermic reaction and Eabst is thenascent barrier to abstraction. This equation is used both for alkyl radical (R) isomerizationand hydroperoxyl radical (RO2) isomerization.
The ring strain energy is highly dependent on its size, with lower energies seen for largerrings (Table 2.2).
From Figure 2.34, it can be seen that a minimum chain length of two and five C is requiredto form a 5-membered or 8-membered transition state ring, respectively. This implies thathydrocarbons, as they grow longer average chain lengths, have more options for low energybarrier isomerization reactions and therefore higher overall reaction rates.
Recently, quantum chemistry models have been used to calculate RO2, concerted elimina-tion and isomerization reaction rates [188–191]. The results also show a decrease in activationenergy as the transition ring size increases or the C-H bond type changes from primary to
37
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
Table 2.2: Impact of transition ring size on strain energy
Number of ring membersRing strain energy (kcal/mol)
tertiary. The lowest activation energy is observed for a seven-membered ring size comprisingmainly tertiary C-H sites.
H atom transfer from C atom adjacent to the double bond will decrease the isomerizationreaction rate, because of the additional strain energy seen for transition rings built from allylicradicals [86, 192].
Other electronegative functional groups, including C=O, will also impact the reaction rate,because H atom bonded to C adjacent to these functional groups has a much lower energybarrier. Consider, for example, MTBE (Figure 2.22). In this oxygenate, the combination ofmany primary H atoms and a highly branched structure make it difficult to form the transitionstate ring. In ethers, the presence of C-O will also decrease the energy barrier, while for esters,shifting an H atom involves a higher strain energy than seen for paraffin (Figure 2.35) [156].
2.2.3 Strong carbon-carbon bonds
As discussed earlier, resonance stable intermediates may accumulate to high concentrationsreadily at low temperature and may even go on to produce even less reactive dimers or/and leadto subsequent termination reactions [193, 194]. Resonance stability increases the activationenergy towards further decomposition, thereby depressing the overall reaction rate.
Resonance stable intermediates, including allyl, benzyl, benzoxy and cyclopentadienylradicals share one apparent common denominator: the presence one or more unsaturated Cbonds. A second frequent, though not omnipresent common denominator is cyclicity.
C atoms in benzene is linked together by means of super π bonds, the electrons of whichbeing delocalized and shared amongst its six C members, creating stable structure, particularlyat low temperatures.
Benzyl radicals are also resonance stable and will, in an oxidative environment, lead toradical-radical reactions, such as the production of even more stable bi-benzyl. This proclivitylies at the heart of both the low propensity for aromatics to decompose further and theformation of soot [96].
Benzoxy radicals, the main intermediate for ring opening reactions, is also resonancestable. Ring opening is preceded by the formation of cyclopentadienyl radicals irrespective ofthe nature of side chains [195].
Chapter 2. Impact of fuel molecular structure on auto-ignition chemistry
2.3 Conclusions
The generic design rules for octane boosters as outlined in Section 2.2 can be condensed asfollows:
• A strong weakest link→ Strong C-H bonds inhibit initial radical formation (Section2.2.1). For a given carbon structure, addition of an aldehyde or ether functionality is notdesirable, as these groups tend to lower C-H bond strength in adjacent carbon (Section2.1.5).
• A short average chain length→ Short chains have a lower propensity to form transitionstate rings (Section 2.2.2). For a given carbon structure, a well placed ether group canbe beneficial here, as the associated relatively weak C-O bonds tend to break up the fuelinto shorter fragments (e.g., MTBE; Section 2.1.5).
• Strong and/or cyclic carbon skeleton → Unsaturated and/or cyclic carbon bondspromote the formation of stable intermediates (Section 2.2.3). For a given carbonstructure, the presence of an alcohol, ester or ketone group has a positive effect, giventhat these functionalities can facilitate indirect routes towards unsaturated carbonbonds (Section 2.1.5).
It should be noted here that abiding by the second rule automatically leads to satisfactionof the first, seeing as shorter chains are laden with more primary C-H bonds. These bonds arestronger than the secondary sort that eventually become dominant for long chains (Section2.1.1). The latter rule should therefore take priority over the former.
While abiding by these rules will certainly yield a high RON, this exercise will notautomatically also lead to a high sensitivity S (Section 1.3). SI engines have over time becomeconsiderably more powerful and fuel efficient. This progress owes much to technologicalachievements, notably in the fields of direct injection and turbocharging (Section 1.2).
Common denominators herein, with respect to prevailing temperature-pressure historiesin the combustion chamber, are not only a higher absolute value for pressure and temperature,but also lower values for the former relative to the latter. To accommodate a changingcombustion chamber environment, future octane boosters, as suggested by octane index (OI)theory (Section 1.3), should have both a high RON and sensitivity S.
S arises from the absence of NTC behavior (Section 1.3.2). As NTC chemistry is unique toparaffins and other hydrocarbon types with large paraffinic side chains, the recipe for highS coincides with the unsaturation component of design rule 3. Accordingly, the third ruleshould take priority over the other two.
In conclusion, highly unsaturated (cyclic) compounds are the preferred octane boosters formodern SI engines. Additional side chains of any variety will dilute this strong performance.Multi-branched paraffins come in distant second place, owing to their negligible S. Dependingon the type and location of functional oxygen groups, oxygenates can have a beneficial, neutralor detrimental impact on anti-knock quality.
40
Chapter 3Performance of (Hemi-) CelluloseDerived Compounds asOctaneBoost-ers
In this chapter we present the knock-resistance quality of (hemi-) cellulose derived levulinic estersand cyclic ethers. In order to analyze their anti-knock quality and the reasons for their differentperformance, these compounds were tested both in an SI engine and a constant volume autoignitiondevice. The experimental results are subsequently discussed with the aid of chemical kinetic models.1
3.1 Introduction
There is a growing demand for transport fuels, particularly in fast-growing countries likeChina, India and Brazil. This can be seen as an opportunity for new entrants onto themarket, most notably biorefineries. Lignocellulosic biomass is the majority compound in suchnon-edible feedstocks as corn stover, straw and wood. As the name suggests, it comprises(hemi-)cellulose (70-95%) and lignin (5-30%). There are three main (thermo)chemical routesto transfer lignocellulosic biomass into biofuels (Figure 3.1):
• Gasification→ syngas
• Pyrolysis or liquefaction (high temperature)→mixed liquid or solid fuels
• Hydrolysis (low temperature)→ wide range of monomers
In the hydrolysis process, C5 and C6 sugars are yielded from hemi-cellulose and cellulose,respectively (Figure 3.1). These sugars can subsequently react further to produce biofuelsinvestigated in this study: methyl (ML)- and ethyl levulinate (EL), furfuryl ethyl ether (FEE)and ethyl tetrahydrofurfuryl ether (ETE) (Figure 3.2).
1The results shown in this chapter has been submitted to ’Fuel’
41
Chapter 3. Performance of (Hemi-) Cellulose Derived Compounds asOctane Boosters
Figure 3.1: Possible conversion pathways for (hemi-)cellulose [196, 197]
C6 (Hexoses)Dehydration /Mailard reactions
O
HOCHO
HMF
Rehydration COOH
OLevulinic acid
EsterifiedLA ester
C5 (Xyloses)
Surfid acid organosolv process
OCHO
Furfural
Side product
O CH3
OCH2OH
Furfural alcohol
Etherified with ethanol
O
FEE
O
O
ETE
O
Figure 3.2: Production routes for investigated biofuels [197, 198]
Furfuryl ethyl ether (FEE) and ethyl tetrahydrofurfuryl ether (ETE) have reportedly the po-tential to curb soot emissions in compression ignition engines [199]. Similar tests conductedfor EL showed an equally promising soot reduction potential [200]. ML and EL have a verylow derived cetane number (DCN) (< 10) [200]. Considering the inversely linear relationshipbetween cetane number (CN) and research octane number (RON) [201], both levulinic fuelsare expected to have a high RON and might therefore be attractive candidates for use asbio-octane boosters in spark-ignition (SI) engines [202].
In order to test the potential of aforementioned compounds as octane boosters, theiranti-knock quality, as well as blends with gasoline, will be evaluated in an ignition quality tester(IQT) and an SI engine, respectively. Experimental results will subsequently be discussedwith the aid of qualitative chemical kinetics considerations.
42
3.2. Methodology
O
ETE
O
O
FEE
O
O
O
O
O
O
O
ML EL
Figure 3.3: Neat biofuel molecular structures
Table 3.1: Physiochemical properties of the neat biofuels
Fuel Formula Densitya Boiling Pointa Viscosity LHV DCNb
a Data retrieved from MSDSb Measured in this study
3.2 Methodology
3.2.1 Fuels
Molecular structures of the neat biofuels are shown in Figure 3.3. Physicochemical propertiesof the neat and blended (i.e., to 10 vol.-% in commercial Euro95 gasoline) biofuels are listedin Table 3.1 and 3.2, respectively. Of particular interest here are the unexpectedly high DCN(78.9) for ETE, and the low volatility of ML and EL relative to the distillation T90 (190℃)and endpoint (225℃) limits for gasoline, which may cause poor mixture distribution in theintake manifold and combustion chambers. Furthermore, on a volumetric scale, ETE andFEE have very similar lower heating values (LHV) compared to gasoline, while ML and ELscore considerably lower on this point.
To analyze the composition of the reference gasoline fuel, a detailed hydrocarbon analysisis conducted in accordance with ASTM D6729. The results, summarized in Table 3.3, reveala relatively high aromatic content of 31.58%, as well as the presence of commonly blendedoxygenates: methanol, ethanol and methyl tert-butyl ether (MTBE) (0.4%, 3.9%, 3.1% byvolume, respectively).
43
Chapter 3. Performance of (Hemi-) Cellulose Derived Compounds asOctane Boosters
Table 3.2: Physiochemical properties of the blends
Fuel Density Oxygen content LHV LHVg/L [wt.-%] MJ/kg MJ/L
Engine experiments are conducted on a Volvo T5 turbocharged port fuel injected 5-cylinder SIengine. A schematic overview of the setup is shown in Figure 3.4 and the engine specificationscan be found in Table 3.4. In light of the fact that SI engines tend to be more knock prone atlow speeds, 1500 RPM is selected as the reference point in this study. Earlier work [203] hasshown that the anti-knock quality of various fuels is quite similar at part-load and full-load.Accordingly, only the full load or wide open throttle (WOT) case is considered here.
Engine-out coolant and oil temperatures are kept constant at 91℃. The setup is equippedwith a water-cooled Schenck E2-330 eddy current brake. A Kistler piezoelectric pressuresensor with a resolution of 3600 samples per revolution is installed onto one of the cylinders.Intake pressure is set to 1.26-1.28 bar (abs.) and a constant intake temperature of 21±2 °C ismaintained. The default spark timing for the neat gasoline reference fuel is 12 crank angledegrees before top dead center (°CA bTDC). Each fuel is subjected to a spark timing sweep,whereby 200 cycles are recorded and averaged for each working point.
There are several in-cylinder pressure-based parameters that can be used to evaluate thedegree of knock, including the 1st and 3rd derivative of the pressure, peak pressure, rate ofheat release and band pass filtered pressure [204, 205]. Analysis of the latter parameter is themost widely adopted method in engine knock studies [204] (more detail can be found in theAppendix A).
Therefore, in this study, the signal energy of pressure oscillation (SEPO), which is thesignal energy of the band pass filtered pressure over a certain knock window is used todetermine the knock intensity (KI), whereby the pressure signal is filtered by a 6-25 kHz band
44
3.2. Methodology
M Fuel pump
Air filter
Dynamometer
Turbocharger
Flow meter
Intercooler
Fuel flow meter
Fuel Tank
Exhaust gas
Throttle
Air
Figure 3.4: Schematic representation of the test engine
pass filter from 10 to 55 °CA after top dead center (aTDC). And the knock threshold is definedas the sharp increase of SEPO for all the fuels, in this case, KI is fixed at 1.2 pa2s.
The so-called knock limited spark advance (KLSA) is defined as the spark timing at whichthe KI surpasses this threshold for the first time. For a given set of operating conditions andengine specifications, a more advanced spark timing at the KLSA is indicative of a better anti-knock quality, typically yielding a combination of improved torque, fuel conversion efficiencyand fuel economy.
3.2.3 Constant volume autoignition device
A modified ignition quality tester (IQT) was used to study the temperature dependency ofautoignition for the oxygenates. The IQT is a constant volume combustion chamber, equippedwith a fuel injection system in accordance with the ASTM method D6890 for measuring
45
Chapter 3. Performance of (Hemi-) Cellulose Derived Compounds asOctane Boosters
Figure 3.5: Typical combustion pressure and needle lift signals observed in IQTexperiments (Reprinted with permission from Bogin et al.[206]. Copyright (2015)American Chemical Society)
DCN [15]. ID is defined here as the duration between the start of injection and the time atwhich the pressure reaches the so-called pressure recovery point of 138 kPa above the valueprior to fuel injection. (Figure 3.5).
However, the ASTM D6890 method is known to produce a heterogeneous fuel-air mixturesuch that the ignition delay time measured is a combination of physical and chemical kineticfactors. Bogin et al. have modified the IQT controls and operating parameters for use inmore fundamental ignition delay and kinetic mechanism validation studies [206–208]. Theauthors observed that for a sufficiently long auto-ignition delay time (ID) (e.g., > 40 ms), themixture conditions are nearly homogeneous in nature. In such an environment, chemicalkinetics dominated the auto-ignition process. Fluid dynamical effects associated with injectionphenomena such as spray break and evaporation can be ignored under these circumstances.
Temperature sweeps of the neat oxygenates were performed in the modified IQT at apressure of 10 bar. This specific pressure was selected because it was required to obtainignition delay times in excess of aforementioned 40 ms, while still being representativeof the end-gas pressure for SI engine operation. Accordingly, we assume SI engine-likepseudo-homogeneous mixtures exist for most of the data.
The IQT chamber was pressurized with air (21% O2 in N2) and heated to the highesttemperature in the sweep (e.g., 980 K). For lower temperatures, down to 700 K, the heaterswere turned off while fuel was injected at regular intervals as the chamber cooled down.To produce an equivalence ratio of 1 at 900 K, the mass of injected fuel was adjusted toaccommodate the distinct densities and molecular formulas of the various fuels by usingthe variable volume injection pump. The volume of the fuel injected can be changed byusing metal shims with different thickness, which will change the relative position of the fuelplunger of the injection pump (more detailed description of how to choose the shim thickness
46
3.3. Results and discussion
2 4 6 8 10 12 14 160
0.5
1
1.5
2
2.5
3
3.5
4x 10
5
Spark timing [degree bTDC]
SE
PO
[pa2 s]
Euro9510%EL10%ML10%FEE10%ETE
Figure 3.6: SEPO as a function of spark timing for all fuels at WOT and 1500 RPM
are in Appendix B). The mass of fuel injected was then held constant for each temperaturesweep, and consequently the equivalence ratio of the combustion events decreased as thecombustion chamber cooled. Other studies comparing constant mass to constant equivalenceratio temperature sweeps have shown that the effects of this change in equivalence ratio onID measurements are small [207–209].
3.3 Results and discussion
3.3.1 Engine
Knock limited spark advance
Figure 3.6 shows SEPO data as a function of spark timing. Irrespective of fuels, advancementof the spark timing always manifests in a higher SEPO when the engine begin to have slightlyknock. For most oxygenated blends, the knock limit, demarcated here by the dashed line, isbreached at significantly earlier timings than is the case for neat gasoline. The sole exceptionis the ETE blend that, as might have been predicted based on the high DCN of the neat biofuel(Table 3.1), is considerably more reactive than the reference fuel. Both blends with levulinicesters show the best performance, with 10%FEE trailing not far behind.
The KLSA of the blends (Figure 3.6) would thus appear to correlate quite well with neatoxygenate CN (Table 3.1). Accordingly, the data suggests that relative anti-knock quality in anengine can be predicted by comparatively low-cost and less fuel demanding IQT experiments.
47
Chapter 3. Performance of (Hemi-) Cellulose Derived Compounds asOctane Boosters
Table 3.5: Engine performance at the knock limited spark advance a at WOTand 1500 RPM
a All values in this table are taken at the knock limited (KL) spark timing bymeans of lenear interpolation of two nearby spark timings
Load, efficiency and fuel economy
Engine performance, represented by the net indicated mean effective pressure (IMEP), fuelconversion efficiency and volumetric indicated specific fuel consumption (ISFC), is evaluatedonly at the KLSA and is summarized in Table 3.5. IMEP represent the amount of work enginehas done per cycle per volume cylinder. Volumetric fuel consumption is studied here becauseit is deemed more relevant for practical purposes than the gravimetric variety. Fuel conversionefficiency is calculated by Equation 3.1, where P is power, m is the mass of fuel inducted percycle.
ηf =P
LHV ×m(3.1)
With respect to the reference fuel, an earlier and later KLSA generally translates intomodest gains and penalties in IMEP (Figure 3.7) and fuel conversion efficiency (Figure 3.8),respectively. There is one notable exception, however, that requires further investigation. Itwould appear that the blend with the most advanced KLSA, 10%EL, performs no better interm of IMEP and efficiency than the reference fuel.
Pertaining to fuel economy, it can be seen that the penalty otherwise incurred by thepresence in fuel oxygen can be partially to fully offset by a combination of higher densities(Tables 3.1 and 3.2) and improvements in fuel conversion efficiency (shown in Figure 3.9).
Combustion phasing is expressed here as the cycle averaged CA50, it is the crank angle atwhich 50% of the total heat release occurs (Figure 3.10). For a given spark timing, FEE andgasoline have a similar CA50, which suggests a comparable combustion rate. ML and ELburn somewhat slower than the reference fuel.
3.3.2 Constant volume autoignition device
Engine experiments have shown that ML, EL and FEE have a superior anti-knock qualitycompared to the reference Euro95 gasoline. Accordingly, it is expected that these fuels havea commensurately longer ID when measured under engine-like condition. To validate thisassumption, auto-ignition behavior is evaluated in a modified ignition quality tester (IQT),which is a constant volume autoignition device.
48
3.3. Results and discussion
2 4 6 8 10 12 14 1613
13.2
13.4
13.6
13.8
14
14.2
14.4
14.6
14.8
15
Spark timing [degree bTDC]
mea
n ne
t IM
EP
[Bar
]
Euro9510%EL10%ML10%FEE10%ETE
Figure 3.7: IMEP as a function of spark timing for all fuels at WOT and 1500 RPM(vertical bars denote the KLSA points)
Figure 3.11 shows the ID of neat ML, EL, FEE and ETE as a function of temperature. Theresults are in line with expectations in so far as the ranking amongst the fuels is equal tothat established earlier in the engine section (Figure 3.6 and Table 3.5). However, the ML andEL results show more scattered data. This may be due to their high boiling point, and as aconsequence, the fuel injected may not be fully vaporized or spray to the wall.
3.3.3 Chemical kinetics analysis
The knock resistance of the fuel is dominated by their auto-ignition delay time at certaintemperature and pressure, which are dependent on their molecular structure.
Levulinic esters have a ketone and an ester on the alkane chain, FEE and ETE have a(un)saturated furan ring and an ether. Each of the compounds tested in this chapter are acombination of two oxygenated functional groups. Although a single functional group hasbeen studied quite thorough, e.g. alcohols and ethers, the effect of two oxygenate functionalgroups together has not been studied much.
Levulinic esters
Thion et al. [210] conducted a computational study on ML chemical kinetics. Apart fromthis study, very little has been published on the kinetics of levulinate compounds. As aconsequence, the mechanism of methyl butanoate (MB) will be used here as a proxy. MBhas similar structure as ML, except for an extra carbonyl group (C=O) attached at the C2
49
Chapter 3. Performance of (Hemi-) Cellulose Derived Compounds asOctane Boosters
2 4 6 8 10 12 14 1634
34.5
35
35.5
36
36.5
37
37.5
38
38.5
Spark timing [degree bTDC]
The
mal
effi
cien
cy %
Euro9510%EL10%ML10%FEE10%ETE
Figure 3.8: Fuel conversion efficiency as a function of spark timing for all fuels atWOT and 1500 RPM (vertical bars denote the KLSA points)
position and a methyl group (CH3). MB has a very low DCN as well (6 [201]) and its chemicalmechanism has been the subject of several studies [112, 149, 153, 157].
Several bond dissociation energies (BDE) of C-H and C-C bonds of MB are shown inFigure 3.12 [153]. It can be seen that due to the attraction from the O atom, the C-O bond isstronger than the C-C bond and the Cβ −H bond is relatively weak (93.6 kcal/mol comparedto 98.5 kcal/mol on the Cγ − H), making this site the most preferred place for initial auto-ignition reactions. Using the kinetic model from Dooley et al. [149], at 836K and φ = 0.5 inrapid compression machine (RCM), 35.5% of MB initiated by H atom abstraction from theCβ −H site, the produced alkyl radicals (C2H5CHC(O)OCH3) then were largely consumed byβ-scission, ultimately producing CH2CHC(O)OCH3 and CH3 radicals.
The C2H5CHC(O)OCH3 radical is resonance stable (Figure 3.13) and its stability slowsdown the overall reaction rate, particularly in the low temperature range.
Compared to MB, ML has the same number of H atoms available for H atom abstractionsand the BDE of adjacent C-H is also weakened by the O atom, while the main chain length isone carbon longer because of the C=O group. Thion et al. [210] showed that in ML reactionkinetics, the sites between the two functional groups (C3 and C4) are the most favorable sitesfor H atom abstraction reactions. The produced radicals are resonance stable as well. Similarto MB, β-scission instead ofO2 addition reactions are the dominant routes at low temperature.The simplified reaction scheme of ML is shown in Figure 3.14 [210]. ML has same amountof C atoms as methyl pentanoate. Their DCN are 7.8 (Table 3.1) and 13.3 [201], respectively,which indicates that adding one ketone on an ester will make the compound more stable.
EL has one more methyl on the ester side compared to ML. It is slightly more stable thanML from the DCN data, while from the engine tests, their anti-knock quality are very similar.
50
3.3. Results and discussion
2 4 6 8 10 12 14 16305
310
315
320
325
330
335
340
345
Spark timing [degree bTDC]
Fue
l con
sum
ptio
n [m
l/kW
h]
Euro9510%EL10%ML10%FEE10%ETE
Figure 3.9: Volumetric indicated specific fuel consumption as a function of sparktiming for all fuels at WOT and 1500 RPM (vertical bars denote the KLSA points)
This shows that the two oxygen functional groups (the ketone and ester) are the dominantfactor to impact the ignition delay times compared to the chain length.
Furan ether and tetrahydrofuran ether
Furans generally have strong C-H bonds on the ring (higher than the ring C-H bond intoluene, shown in Figure 3.15) [122], which makes these heterocyclic structures particularlystable. However, their side chain C-H bonds are weaker than in toluene. E.g., 2-methylfuran(2-MF) has a weaker C-H BDE in the methyl group than is the case for toluene. Furan, 2-MF,benzene and toluene are all very stable, and have a small cetane number (7, 8.9, 14.3 and 3,respectively [120, 201]). This indicates that the furan ring is more stable, but with one methylgroup, toluene is much more stable than 2-MF).
No reaction mechanisms could be found for FEE or ETE in literature, and in order toqualitatively analyze the kinetics of these two compounds, the mechanism of 2-MF and2-methyl tetrahydrofuran (2-MTHF) are studied instead as a reference.
The majority of 2-MF is consumed by H atom ipso addition and H atom abstractionreactions taking place on the alkyl side chain, producing furan and 2-furanyl methyl radicals,respectively. The main reasons for its overall slow reaction rate are the stable furan ringstructure and resonantly stabilized intermediate radicals, such as 2-furanyl methyl, combinedwith the lack of low temperature branching reactions.
In FEE, conversely, which has an ethyl ether functionality connected to the methyl sidegroup, the Cα-H is weakened by the O atom of the ether group, thus making it susceptible toabstraction and subsequent reactions. A proposed simplified reaction scheme for FEE at low
51
Chapter 3. Performance of (Hemi-) Cellulose Derived Compounds asOctane Boosters
2 4 6 8 10 12 14 1614
16
18
20
22
24
26
28
Spark Timing [degree bTDC]
Mea
n C
a50
[deg
ree
aTD
C]
Euro9510%EL10%ML10%FEE10%ETE
Figure 3.10: CA50 as a function of spark timing for all fuels at WOT and 1500 RPM
temperature is shown in Figure 3.16.
The active ether group also provides additional sites for H atom abstraction (route c and din Figure 3.16), which then allow for other alkyl consumption reactions. Accordingly, FEE ismore reactive than short chained furans.
Tetrahydrofuran (THF) has a higher DCN than furan [120]. This is mainly because thetetrahydrofuran ring is not stable, the ring C-H bonds are weak compared to unsaturated furan(93-99 kcal/mol vs. 120 kcal/mol). This means that the ring C-H sites are the most likelylocations for H atom abstraction reactions and the weak ring C-C bonds can then undergoring opening reactions via β-scission.
Substituted tetrahydrofurans, take 2-MTHF as an example, has very strong CH2−H bondson the substituted side chain (Figure 3.15), which are much stronger than those found inunsaturated furan (103 kcal/mol vs. 86.2 kcal/mol). The ring C-H bonds are weak like thosein THF, and the C2 −H is weaker than others (93.7 kcal/mol). Therefore, H atom abstractionstill prefers to occur at the ring H atom sites, especially at the C2 site. The simplified reactionscheme of 2-MTHF is shown in Figure 3.17 [213].
Furanyl radicals are formed after H atom abstraction, which subsequently, via β-scission,can undergo ring opening reactions, or can eliminate an H atom and produce dihydrofurans.Compared to 2-MF, 2-MTHF is more readily to have ring opening reactions. This, combinedwith the fact that the C-H bonds on the ring are relatively weak, explains the relatively shorterID and higher DCN (22 vs. 8.9) of 2-MTHF. [120].
Sudholt et al. [120] concluded that the side chain length has a strong influence on reactionchemistry in tetrahydrofurans. Compared to 2-MTHF, ETE has an ethyl ether functionality onthe methyl group. As discussed before, an ether group will weaken the adjacent C-H bonds.
52
3.4. Conclusions
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
101
102
103
104
Temperature 1000/T [1/K]
Igni
tion
dela
y tim
e(lo
g) [m
s]
MLELFEEETE
Figure 3.11: Ignition delay time as a function of temperature for all fuels
H
H2C
HC
HC
C
O
CH2
H
O
H
H101.3
88.985.2 89.9 99.2
86.8
100
98.593.6
Figure 3.12: Bond dissociation energies in methyl butanoate (unit: kcal/mol) [153]
Moreover, the long side chain now provides more possible sites for auto-ignition reactions,resulting in a relatively high reactivity for ETE.
3.4 Conclusions
The objective of this chapter was to investigate the anti-knock quality of sugar-derived levulinicesters (ML and EL) and cyclic ethers (FEE and ETE). To this end, combustion experimentswere carried out in both an engine and constant volume autoignition device (IQT). The results
O
O
O
O
Figure 3.13: Resonance stable structure of C2H5CHC(O)OCH3 [149]
53
Chapter 3. Performance of (Hemi-) Cellulose Derived Compounds asOctane Boosters
from both apparati demonstrate that ML, EL and FEE have a higher anti-knock quality thanthe reference Euro95 gasoline. ETE, conversely, performed markedly worse than the referencefuel on both setups and might therefore be a more appropriate fuel for compression ignitionengines.
The main reason for the distinctions in anti-knock quality can be found in the molecularstructure of the neat biofuels. ML and EL are levulinic esters, with an ketone functionalityand an ester group on the carbon chain. It can readily produce stable intermediates duringthe auto-ignition process, thereby slowing down the overall reaction rate. This indicates thatadding more oxygenate functionalities (except ether groups, whose stability depend also onthe structure of the connecting alkane chain) on an alkane chain, have the effect of shorteningthe main alkane chain length, will make the compound more stable.
The unsaturated cyclic ether (FEE) has very strong ring C-H bonds. However, the saturatedcyclic ether (ETE) has weak ring C-H bonds, which facilitate more readily ring openingreactions. Adding another oxygenate functional group to the stable ring (e.g. furan or benzenering) will increase the reaction rate, e.g., long side chains on the cyclic ethers further acceleratethe reaction rate.
Importantly for future research, our results suggest that the modified IQT and engine ex-periments are interchangeable setups with respect to qualitative anti-knock quality evaluationof novel compounds.
54
3.4. Conclusions
C H
H
H
89.7+1.2102+1
Toluene
C
O CH3
114.7+1.4 C
O C2H5
74+1.2120.4+1.4
H
H112.9+0.5
2-MF 2-EF
CH
HC
OH
H
93.8+0.2
98.4+0.05
THF
HC
CH
CH
OH2C
H
103.192.9
H99.4
MTHF
OH2C
CH3100.6
102.7
ETHF
86.3+0.6
H
H 98.3
H
93.7
2
34
5
Figure 3.15: Bond dissociation energy (BDE) in toluene and various cyclic ethers(data without arrow shows the BDE of nearby C-H bond) [120, 122, 211, 212]
O
O
O
FEE
Oα β γ
O
O
Formyl furan
CH
O
O
CH
O
O
OH
HO
O
O
O
O
(a)
(b)
(c)
(d)
Figure 3.16: Proposed reaction scheme for furfuryl ethyl ether at low temperature
55
Chapter 3. Performance of (Hemi-) Cellulose Derived Compounds asOctane Boosters
O
O
O
O
O
O
O
O
O
OO O
or
O
5-furanyl
4-furanyl
3-furanyl
2-furanyl
4,5-dihydrofurans
2,3-dihydrofurans 3,4-dihydrofurans
Figure 3.17: Simplified reaction scheme for 2-methyltatrahydrofuran [213]
56
Chapter 4Performance of Lignin Derived Com-pounds as Octane Boosters
In this chapter, in order to analyze lignin derived compounds’ anti-knock quality, and the effect ofdifferent side chains on the benzene ring of mono-aromatics, several chosen lignin derived compoundswere mixed with gasoline and tested in a commercial SI engine. Some of the compounds were alsotested in a constant volume ignition device, to study their auto-ignition process and the experimentresults are discussed by analyzing their combustion kinetic mechanisms.1
4.1 Aromatic oxygenates
Lignin, a critical component of plant cell walls, is the third most abundant natural polymerafter cellulose and hemicellulose. Its large quantity and chemical structure make it an attractivefeedstock for producing bio-aromatics [214]. Lignin is a three-dimensional amorphous polymerconsisting of methoxylated phenylpropane structures, including mono-, di-, and polyalkylsubstituted phenols, benzenes and alkyoxybenzenes, connected by C-O-C and C-C bonds [215].It is the only direct source of aromatics in lignocellulosic biomass.
Lignin is composed of three principal types of monolignols, or monomer units. These arep-coumaryl, coniferyl alcohol and sinapyl alcohol [196] (Figure 5.1).
Given that a wide variety of compounds can be made from these units, a survey intolignin literature has been conducted to reveal the prevalent aromatic oxygenates that can beproduced from lignin which are also liquid at room temperature. As is clear from Table 4.1these concern anisole and guaiacol, as well as the most common alkylated versions thereof.Other aromatics, such as phenol, aromatic alcohol, xylene, benzene etc. can also be madefrom lignin.
1Part of the results shown in this chapter has been published by Miao et al. SAE Int. J. Fuels Lubr.8(2):2015, doi:10.4271/2015-01-0963, and part of the results shown in this chapter has been submitted to’Fuel’
57
Chapter 4. Performance of Lignin Derived Compounds as Octane Boosters
To analyze the influence of oxygen functional group and the side paraffin chain length onthe fuel properties of lignin derived compounds, benzyl alcohol (BA), 2-phenyl ethanol (2-PE), acetophenone, 1,2-dimethoxy benzene (veratrole), anisole, guaiacol and their alkylatedcompounds: 4-methyl anisole (4-MA), 4-propyl anisole (4-PA), 4-methyl guaiacol (4-MG)and 4-ethyl guaiacol (4-EG) were selected as lignin models. Their structures are shown inFigure 4.2 and physicochemical properties are listed in Table 4.2. These compounds includealcohol, ketone and ether functional groups, with different length of the alkylate chains. Byanalyzing their reactivity, the effect of these oxygenate groups can be compared. The ligninderived fuels have similar or even higher LHV compared to gasoline on a volume basis (31MJ/L), making them attractive alternative fuel options. However, 2-PE, BA, veratrole, 4-PA,guaiacol, 4-MG and 4-EG have very high boiling points relative to the distillation T90 (190℃)and endpoint (225℃) limits for gasoline.
As stated in Chapter 3, the inverse correlation between DCN and RON predicts thatcompounds with a DCN of less than about 20 will have a RON of 95 or higher [247]. So all ofthe lignin derived models are predicted to have RON equal to or above 95. These fuels werethen blended with gasoline (Euro95) at 10 vol-%, to test their behavior in an SI engine. The
58
4.2. Methodology
Table 4.2: Properties of tested fuel compounds
Fuel component Formula Densitya BPa Viscosityb LHVc DCNc
a Density and boilling point (BP) are from Sigma-Aldrich website.b Viscosities were measured in this study.c Low heating value (LHV) and cetane number (CN) are from McCormick et. al[224], except toluene, which are from NIST web and [247].
blends’ properties are listed below (Table 4.3). The oxygenates for which blend RON wereavailable show a small increase or no change to the 95 RON value of the base fuel. The LHVby volume basis of all the blends vary less than LHV on a mass basis.
4.2.2 Experiment setups and operation condition
The engine setup and the modified ignition quality tester (IQT) were described in Chapter3. The Knock Intensity (KI) was defined using the maximum amplitude pressure oscillation(MAPO) [249]. The threshold for knock is based on the MAPO at which the KI sharplyincreased. The knock limited spark advance (KLSA), in this case, was defined as the sparktiming at which the KI surpasses the threshold.
Two parts of the experiments were conducted, in the first part, BA, 2-PE, acetophenone,veratrole and anisole were blended and tested in the VolvoT5 engine, to compare the knockresistance qualities of compounds with hydroxy, carbonyl and methoxy functional groups andvaring chain length or the number of functional group. The test conditions were chosen at1500 RPM and at both part load (190 Nm) and full load (wide open throttle, 260 Nm), becauseaccording to literature, knock is more prone to happen at low speed and high load.
In the second part, anisole, guaiacol, 4-MA, 4-PA, 4-MG and 4-EG were tested in both theVolvo T5 engine and a modified IQT, to compare the reactivity of methoxy functional groupwith/without an additional hydroxy functional group, and the effect of an alkane chain at thepara-site. For the engine experiments, spark timing sweeps were conducted at 1500 RPM andfull load. And for auto-ignition test, temperature sweeps were conducted at constant pressureof 10 bar and stoichiometric mixing condition, as described in Chapter 3.
59
Chapter 4. Performance of Lignin Derived Compounds as Octane Boosters
OO O
OH
O
O
OH
OH
O
Toluene Anisole 4-Methyl anisole 4-Proply anisole
Guaiacol4-Methyl guaiacol 4-Ethyl guaiacol
OH OH
2-Phenyl ethanol
O
O
O
VeratroleAcetophenoneBenzyl alcohol
Figure 4.2: Fuel blending component structures
4.3 Results and analysis: Part I
In this section, the results of engine experiments of blended fuels, which are 2-PE, BA,acetophenone, anisole and veratrole, at 1500 RPM, both part load (190 Nm) and full load (260Nm) working conditions, are given and discussed.
4.3.1 Knock resistance
As mentioned earlier, a more advanced KLSA is indicative of superior anti-knock behavior.The results indicate that, compared to neat gasoline and in line with reported RON/MONvalues (Table 4.3), all oxygenated blends have a more advanced KLSA at both wide open throttleor WOT (Figure 4.3) and part load (Figure 4.4). While all oxygenates share the same benzenering at their core, each has a unique side chain, which leads to different reaction pathwaysand thus distinct reaction rates.
Examining the data in Figure 4.3 and 4.4 more closely, it can be seen that BA and 2-PEhave similar KLSA curves, irrespective of load. As shown in Figure 4.2, BA and 2-PE havedifferent chain length, but both end in a hydroxyl (-OH) group. This suggests that the benzenebase controls the reaction chemistry for aromatic oxygenates with short hydroxyl chains.
Anisole has the best knock resistance among all these fuels at full load, so that a methoxy(−O− CH3) group is less reactive than a hydroxyl group. Also anisole, as an ether, which issimilar to MTBE or ETBE, the main reason of its stability is lie on the highly stable structureon one side of the ether group. while adding one more methoxy group on the ring, i.e.,veratrole, accelerating the reaction rate. Although at part load, it shows similar reactivity as2-PE.
Acetophenone, on the other hand, has on average a less advanced KLSA curve, i.e. earlierspark timing, than its aforementioned counterparts. This suggests that a carbonyl (C=O)group could be more reactive than a hydroxyl group. A hint towards a probable cause follows
60
4.3. Results and analysis: Part I
Table 4.3: Properties of fuel blends
Fuel Density O content LHVa LHVa Viscosityb RONc MONc
a Blend LHV were calculated based on mole percentage from pure compounds.b Blend viscosity were calculated using Refutas equation from pure compounds [248].c RON and MON values were measured from outside institute.d Acetp: Acetophenone.
from examining the reactivity of iso-propanol versus acetone, a hydroxylic and carbonylicoxygenate, respectively, which share the same C3 carbon chain.
As is known from literature [186], the dominant reaction pathway of anisole oxidation atintermediate to high temperature is decomposition from O− CH3 bond, producing methyland C6H5O radicals, which are very stable, decrease the overall reaction rate. In comparison,veratrole has two OCH3 groups, which means more sites to have initial reactions, and bybreaking from both O−CH3 bond, it will produce 1,2-benzoquinone, which can then undergoring opening reactions. The alcohols, i.e., BA and 2-PE prefer to undergo H atom abstractionreaction or concerted elimination via a four-member transition ring based on general alcoholreaction pathways [133], which lead to the formation of benzaldehyde for the former one, orbenzyl radicals, benzaldehyde, styrene for the later. Due to the OH group, Cα − H bondis very weak, and easy to be abstracted, making it easier to proceed the following reactions.Concerning to acetophenone, H atom abstraction from the methyl group is the dominantpathway, and then β-scission to produce phenyl radicals and CH2CO in the high temperature,according to acetone’s mechanism [174].
In conclusion, the hydroxylic oxygenates have both a higher RON and more advancedKLSA curves than is the case for the carbonylic oxygenate (Figure 4.3 and 4.4). Anisole is theleast reactive compound.
61
Chapter 4. Performance of Lignin Derived Compounds as Octane Boosters
Figure 4.3: Mean MAPO for all fuels at wide open throttle and 1500 RPM, with theknock threshold indicated by the horizontal dashed line
Table 4.4: Engine related OI constants
Load c a b K190 Nm -99.40 1.46 -0.21 -0.17
WOT -52.49 0.95 -0.32 -0.50
4.3.2 Octane Index
In the OI equation (Eq.1.5), the parameter K which is a lelative measure of the severity ofengine condition, can be calculated from the equations below (Eq.4.1) [19].
KLSA = a× RON+ b×MON+ c
K =b
a+ b(4.1)
OI =a
a+ b× RON+
b
a+ b×MON = RON− K× S
Herein, the measured KLSA is used as the knock parameter that depends on the auto-ignitionquality of the fuel in the engine test. K (Table 4.4) and the OI (Table 4.5) are found bycalculating the constants a, b and c.
As discussed earlier, K accounts for shifts in unburnt gas temperature relative to pressure.Accordingly, the negative values for K observed in Table 4.4 should somehow be quantitativelyrelated to conditions inside the combustion chamber. To this end, Kalghatgi [9, 24, 250]
Figure 4.4: Mean MAPO for all fuels at part load (190Nm) and 1500 RPM with theknock threshold indicated by the horizontal dashed line
introduced a single parameter, namely Tcomp15 or the unburnt gas temperature at 15 bar,which want to show the in-cylinder end gas temperature. A modern engine generally has alower Tcomp15, because of the better cooling and higher boost pressure. The values for both Kand Tcomp15 at RON/MON conditions [23] and at the part load and wide open throttle (WOT)conditions used in this study are listed in Table 4.6. Note that for the latter values adiabaticcompression is assumed at a compression ratio of 9:1 (Table 3.4). In line with the studiescited above, falling intake temperatures at constant intake pressure lead to a simultaneousdecrease in Tcomp15 and K (Table 4.6). It is shown that the VolvoT5 engine at these twoworking conditions have a negative K and cooler temperature compare to the RON and MONtests, in agreement with other studies. And K is smaller in WOT condition. Therefore, in thisengine, fuels with higher S will also have a better anti-knock quality from the negative K.
All tested fuels share the same engine and work point, related constants listed in Table4.4, but each have their own respective values for the KLSA and OI (Table 4.5). Given thatall oxygenated blends have a higher S than the reference gasoline, the positive incrementsseen for the oxygenates with respect to OI are commensurately greater than those observedfor those seen for the RON. With the exception of acetophenone, these differences are evenmore pronounced at the higher load point, owing to the more negative K.
4.3.3 Indicated Mean Effective Pressure
The net indicated mean effective pressure (IMEPn) is plotted below as a function of sparktiming for all fuels at WOT (Figure 4.5a) and part load (Figure 4.5b). At WOT, IMEP increases
63
Chapter 4. Performance of Lignin Derived Compounds as Octane Boosters
Table 4.5: Calculated OI for all the fuels in both load points
Load/K Fuel RON S KLSA [°bTDC] OIEuro 95 95 9.4 21.3 96.6
Work point Intake pressure Intake temperature Tcomp15 K[bar] [℃] [K] –
MON 1 148.9 850 [23] 1RON 1 52 700 [23] 0
190 Nm 1 22 590 -0.17WOT 1.24 25 560 -0.5
with more advanced spark timings for all fuels (Table 4.5). At the KLSA, the IMEP is higherfor all oxygenates, with the greatest values observed for anisole and the alcoholic oxygenatesin line with their OI (Table 4.5). Similar trends hold at part load.
4.3.4 Combustion Phasing
When operating under WOT conditions, combustion phasing, expressed here as the cycleaveraged CA50, is more advanced for all oxygenates (Figure 4.6a). An earlier CA50 at aconstant spark timing suggests a more rapid combustion. At part load this trend is morepronounced (Figure 4.6b).
4.3.5 Fuel Consumption
As prices at the fuelling station are based on liters or gallons, from a consumer’s perspectiveit is more interesting to study fuel consumption on a volumetric basis, so the volumetricindicated specific fuel consumption (ISFC) for each fuels at their knock limited spark advance(KLSA) are compared. At WOT, all fuels, as might be expected, show lower fuel consumptionat more advanced spark timings (Figure 4.7a). Optimal fuel economy, while still avoiding
Figure 4.6: CA50 as a function of spark timing for all the fuels at 1500 RPM anddifferent loads
knock, is reached for all fuels at the KLSA. Anisole and veratrole have almost same fuelconsumption as gasoline, the alcoholic oxygenates yield the best fuel economy.
As can be seen in Figure 4.7b of the part load, at most spark timings, the oxygenatesexcept veratrole, owing at least in part to their higher volumetric energy density (Table 4.3),have a superior fuel consumption relative to neat gasoline, and the trend is more clear than infull load.
The results at KLSA are summarized below in Table 4.7. Except veratrole, the KLSA ofother fuels all increased 5-18% in part load and full load, veratole’s KLSA decreased 2.8% atpart load. This shows that by adding these compounds, the fuel’s knock resistance increases.As a result, they can have more advanced spark timing, thus achieve higher knock limitedIMEP. This is true for all the fuels, except for veratrole at part load, which remains the samelevel as Euro 95. Lower LHV of all the blended fuels result in a higher fuel consumption rateon a mass basis and lower fuel conversion efficiency, which is more apparent in part load.
65
Chapter 4. Performance of Lignin Derived Compounds as Octane Boosters
Figure 4.7: Volumetric ISFC as a function of spark timing for all the fuels at 1500RPM and different loads (dashed lines indicate the KLSA)
While the fuel consumption on a volume basis all dropped compared to gasoline, except forveratrole at part load, this is because of the blending fuels have higher density, resulting in ahigher LHV on a volume basis (Table 4.3). Part load and full load show a very similar engineperformance trend for most of the blended fuels, so future experiments will only be conductedat full load, where knock is more likely to happen.
4.4 Results and analysis: Part II
In this section, results from both engine experiments and the modified IQT for anisole, 4-MA,4-PA, guaiacol, 4-MG and 4-EG are shown and discussed.
4.4.1 Engine experiments
Knock intensity
Knock intensity, as expressed here as MAPO, increases for all fuels with advancing spark-timing (Figure 4.8). As might be expected from the anti-knock quality data presented earlier(Table 4.3), the toluene blend shows the greatest knock resistance, with (alkylated) anisolestrailing close behind (Figure 4.8). The (alkylated) guaiacols blends, which behave similarly tothe base fuel, are far more reactive fuels. Evidently, the benzene core imparts low reactivityto all the oxygenated compounds. Nevertheless, the type and number of functional groups,however, exert a strong influence on the anti-knock quality.
Knock limited spark advance
A summary of the fuel impact on engine performance is presented in Table 4.8. There is aclear relationship between KLSA and blend anti-knock quality (Table 4.3), with earlier values
66
4.4. Results and analysis: Part II
Table 4.7: Engine performance at knock limited spark advance
observed for less reactive fuels. A more advanced KLSA, in turn, yields dividends in terms ofefficiency (Figure 4.10) and fuel economy, as will be discussed below.
Efficiency
It can be seen from Figure 4.9 that when advancing the spark timing, the IMEP is increasingfor all the fuels. Fuels with better anti-knock quality allow more advanced spark timings, andthus can achieve higher IMEP. While none of them reached the maximum IMEP due to theknock limitation. This again indicates the importance of having a high anti-knock quality fuel.
The 4-PA and guaiacols blends show a penalty in efficiency relative to the base fuel, aslightly more favorable KLSA (Table 4.8, Figure 4.10) notwithstanding. This discrepancymight be explained by the high boiling point of these compounds that could lead to poorfuel vaporization, as has been observed elsewhere [251]. This high boiling behavior is furtherexacerbated in this study by the use of port-fuel injection and low intake air temperatures(20℃). Aforementioned factors might have contributed to poor combustion efficiency, therebycontributing to the relatively poor fuel conversion efficiency measured for the heavier oxy-genates.
Fuel consumption
As is true by definition for any oxygenated fuel, the presence of oxygen in the molecularstructure has a negative impact on the gravimetric fuel consumption (Figure 4.11; Table 4.8).Notably the guaiacols, having two functional oxygen groups, perform poorly by this measure.
In practice, however, consumers do not pay for mass, they pay for volume and aromaticshave high densities relative to gasoline (Table 4.2). Moreover, as density rises with degree ofoxygenation, the spread in volumetric fuel economy is far less pronounced (Figure 4.12). Infact, the (alkylated) anisole blends yield better fuel economy than gasoline.
67
Chapter 4. Performance of Lignin Derived Compounds as Octane Boosters
Figure 4.8: MAPO for all fuels at WOT and 1500 RPM
4.4.2 Constant volume autoignition experiments
Auto-ignition chemistry is quite important for SI engines, since knock is caused by the auto-ignition of the end-gas in the chamber. Therefore, constant volume auto-igniton experimentswere conducted for all the oxygenates and toluene.
The temperature dependent autoignition behavior for all studied aromatics at 10 bar,720-950 K is shown in Figure 4.13, with iso-octane as a comparison, which has same RONand MON of 100 [207]. As discussed earlier, when the ID is in excess of 40 ms, the ignitionprocess can be assumed to be controlled primarily by chemical kinetics, as opposed to sprayphysics.
The first thing to note from Figure 4.13 is that none of these aromatics exhibited anylow temperature heat release (LTHR) or negative temperature coefficient (NTC) behavior,in contrast to iso-octane. Analysis of the pressure traces for these compunds confirms theabsence of low temperature heat release (Figure 4.14), therefore the use of the conventionalASTM D6890 definition for start of ignition (i.e., a fixed 138 kPa above the chamber pressurerecovery point) is valid [209].
It can be seen that as expected from the DCN and RON data, toluene has the longestIDs over the entire temperature range studied. Anisole and 4-MA have similar reactivity,however the longer alkyl chain of 4-PA appears to slightly increase its reactivity. The increasein reactivity is stronger upon adding another oxygen functional group to the benzene ring.Guaiacol and its alkylated derivatives have the shortest ID. Note that data from T>800 K cannot be assumed as homogeneous auto-ignition for guaiacols because the IDs than 40 ms. Allaromatics are clustered and ranked in accordance with number of functional oxygen groups,
68
4.4. Results and analysis: Part II
Table 4.8: Engine performance at knock limited spark advanceat WOT and 1500 RPMa
Fuel KLSA IMEP Efficiency Vol. ISFCCA bTDC bar % ml/kWh
a All values in this table are taken at the knock limited (KL) sparktiming by means of lenear interpolation of two nearby sparktimings
irrespective of degree of alkylation (Figure 4.13).A fuel’s octane number is a measure of its anti-knock rating, and this parameter is
dominated by auto-ignition chemistry at the ON test conditions. Badra et al. found that 850K and 50 atm has the best fit with the RON operating conditions. For MON, this is 980 Kand 45 atm [252]. Mehl et al. predicted the ON by using detailed chemistry and a simplifiedtwo-zone SI engine model [85].
The IDs for iso-octane, as shown in Figure 4.13 can also show us some qualitativecomparison of ON and this IQT test condition. The auto-igniton of iso-octane is fasterthan toluene for the whole test range, and it is similar to anisole when T>900 K, fromwhich point on, anisoles yield longer IDs. Compared to iso-octane, guaiacol is slower whenT<800 K, but becomes faster when temperature increases above 800 K. According to theblending RON and MON of anisole and guaiacol, it is expected that anisole will have a higherRON than Euro95, while guaiacol has a RON similar to Euro95, which is slightly lower thanthe iso-octane data. The left vertical line, which is the upper limit of the DCN temperaturerequirement, fits for the RON indication. Bogin et al. also found that the ID at highertemperature was more representative of MON conditions [209]. However, more experimentsare needed to study this topic further.
4.4.3 Kinetic analysis
The observed auto-ignition differences from the tested fuels are the result of distinct molecularstructures, which lead to unique auto-ignition chemistry (Figure 4.15). At low temperature(below 850 K [85]), theO2 addition reactions are the main reactions pathways for alkyl radicals:R+O2 = RO2; (R represents alkyl radicals, Q represents CnH2n species [70]). This reactioncan be followed by isomerization and chain branching reactions, accelerating the overallreaction rate. However, peroxy radicals are not stable at high temperatures.
69
Chapter 4. Performance of Lignin Derived Compounds as Octane Boosters
2 4 6 8 10 12 14 1613
13.2
13.4
13.6
13.8
14
14.2
14.4
14.6
14.8
15
Ignition Timing [degree bTDC]
mea
n ne
t IM
EP
[Bar
]
Relation Ignition Timing − mean IMEPn
Euro95
10%Toluene
10% anisole
10% methyl anisole
10%propyl anisole
10% guaiacol
10% methyl guaiacol
10% ethyl guaiacol
Figure 4.9: Net IMEP as a function of spark timing for all fuels at WOT and 1500RPM. Fuel specific KLSA is indicated by dashed bars in the same color
As temperature increases, the QOOH radicals mainly undergo decomposition reactionswhich are chain propagation reactions instead of adding another O2, producing olefin, cyclicether, aldehyde, OH and HO2 radicals. HO2 are more stable than OH radicals, and mainlyundergo termination reactions, producing H2O2 at low temperature. Contrary to chainbranching, chain propagation reactions, along with the formation of stable intermediates casethe drop of the overall reaction rate as temperature increases, and this is the NTC behavior.
At higher temperature, the alkyl radicals prefer to directly decompose into smallercompounds. The reactivity increases again when the mixture reaches temperature exceedsapproximately 1000 K [70], when H2O2 break up to two OH radicals rapidly. The otherbranching reaction, H+O2 = O+OH, becomes the most important one, and accelerates theoverall reaction rate again. More detailed kinetic models have been well studied by others[45, 70].
Compounds which have a long enough alkyl chain, involving carbon chain longer than C3can display NTC behavior. Octane sensitivity S is a measure for the sensitivity of auto-ignitionchemistry to temperature. Fuels which have NTC behavior will be relatively insensityve totemperature in the NTC range. Conversely, fuels void of such chains, and thus of NTCchemistry, for example aromatics, will have a comparatively higher S [253].
The low reaction rate of the latter compounds originates from their stable structure,and/or the production of stable intermediates. The benzene ring, common to all aromatics,provides a stable structure which is not readily broken and as such is responsible for highoctane quality of this class of fuels. Herbinet et al. found that benzene hardly show anyeffect on the formation of C0-C5 products below 800 K when blended with n-decane [254].Functionalization, however, be it via alkylation or oxygenation, has a profound impact on the
Figure 4.10: Fuel conversion efficiency as a function of spark timing for all fuels atWOT and 1500 RPM. Fuel specific KLSA is indicated by dashed bars in the samecolor
reaction pathways involved, as is illustrated in Figure 4.15 for toluene [115, 116, 255], anisole[186] and guaiacol.
Toluene has a higher RON than benzene (116 [256] and 100 [257], respectively). Intoluene, the bond dissociation energy (BDE) is the weakest for C-H bonds in the methyl group(89 kcal/mol, Figure 4.16). Decomposition is thus most likely to commence at these sites,producing benzyl radicals and other small compounds via metatheses by means of H atomabstraction (83% of toluene undergoes this route at 750 K 20 atm in [100]). Benzyl radicals,although readily formed, are resonance stabilized and thus typically long-lived. Accordingly,most such radicals ultimately lead to termination reactions, (re)producing toluene or bi-benzyl,this quenching effect has also been demonstrated by Herbinet et al. [254]. Lack of chainbranching reactions with O2 is the main reason that aromatics. as a class, have a high S [85].
As shown in Figure 4.15, in lieu of toluene and bi-benzyl, benzyl radicals can also reactwith HO2 to produce benzoxy radicals (C6H5CH2O). C6H5CH2O radicals decompose quicklyinto benzaldehyde, which is reactive and can undergo H atom abstraction, producingC6H5CO,followed by CO elimination to produce phenyl radicals. Phenyl radicals can then react with O2to produce phenoxy radicals, a key intermediate species in the toluene oxidation pathway thatis also resonance stabilized and in equilibrium with phenol. At high temperatures, phenolmight react further with an O atom, producing OC6H4OH radicals that, via a CO elimination,can eventually lead to ring opening reactions. Also, phenoxy radicals can decompose to COand cyclopentadienyl radicals. The ring of this radical can subsequently be opened afterreacting with O2, thereby producing C4H4O.
71
Chapter 4. Performance of Lignin Derived Compounds as Octane Boosters
Figure 4.11: Indicated specific gravimetric fuel consumption as a function of sparktiming for all fuels at WOT and 1500 RPM. Fuel specific KLSA is indicated by dashedbars in the same color
Anisole has a methoxy group (−O − CH3) bonded to the benzene ring. Herein, theO−CH3 bond is, at 63 kcal/mol, by far the weakest bond in the whole molecule (Figure 4.16).Note that a BDE of 63 kcal/mol is substantially lower than reported earlier for the methylC-H bonds in toluene (88 kcal/mol). In anisole, the initial reaction of unimolecular bondscission preferentially occurs here, producing phenoxy and methyl radicals. Phenoxy radicalshave an important intermediary role in the ring opening process. This unimolecular reactionoccurs also under pyrolysis conditions and is thus largely independent of prevailing oxygenconcentration [186, 195].
At low temperature, H atom abstraction reactions are the main reaction routes (60%in the Nowakowska et al. model [186] at 800 K and 1 atm), producing C6H5OCH2 radicals,which then isomerize to benzoxy radicals. With respect to toluene, anisole not only has aweaker weakest link, but also provides a direct route to phenoxy radicals (Figure 4.16).
Guaiacol comprises a hydroxyl (-OH) group on the ortho site of methoxy group in anisole.Although there is no detailed mechanism for guaiacol available in literature, BDE analysis ofguaiacol suggests that the initial unimolecular decomposition reaction preferentially occursat the O − CH3 location. Owing to the adjacent OH group, the O − CH3 bond is now evenweaker (56.3 kcal/mol, Figure 4.16) than is the case for anisole. What is more, guaiacolprovides a direct route to OC6H4(OH) radicals that, as discussed earlier, is an even moremature decomposition product than phenoxy radicals.
Figure 4.12: Indicated specific volumetric fuel consumption as a function of sparktiming for all fuels at WOT and 1500 RPM. Fuel specific KLSA is indicated by dashedbars in the same color
4.5 Conclusions
When studying the anti-knock requirements in modern gasoline engines, it emerges thatthe auto-ignition chemistry of aromatics in general is more favorable than that observed forparaffins, currently the most common constituent in gasoline. The lignin literature has beensurveyed to ascertain which aromatic oxygenates (e.g., still liquid at room temperature) are tobe expected in lignin oil, and could potentially be used as octane boosters in gasoline.
In the first part, benzyl alcohol, 2-phenyl ethanol, acetophenone, veratrole, anisole areblended with gasoline and tested in an SI engine. Compared to commercial Euro 95 gasoline,higher RON, MON, OI and KLSA values are found for all aforementioned compounds,except for veratrole. The K values for the engine and operating conditions used in thisstudy are negative as many modern engines tested in the literature. For negative K values,high sensitivity (S) fuels are required to boost the OI. A higher S was found for the aromaticoxygenates compared to Euro 95. The highest S and OI were found for the alcoholic oxygenatesand anisole, whereby the difference between them was small, an aromatic with a hydroxylgroup (-OH) is more reactive than it is with a methoxy (-OCH3) group, and the short alkaneside chain length does not impact too much in the engine experiment.
In conclusion, these aromatic oxygenates are low on toxicity and have improved or thesame level of both knock resistance and volumetric fuel economy under modern engineoperating conditions compared to Euro 95, especially anisole, which proved to be the bestamong all these fuels.
73
Chapter 4. Performance of Lignin Derived Compounds as Octane Boosters
1 1.1 1.2 1.3 1.4 1.5 1.6
Temperature 1000/T [1/K]
100
101
102
103
Igni
tion
dela
y tim
e(lo
g) [m
s]
toluene4-MAanisole4-PAguaiacol4-MG4-EGisooctane
Figure 4.13: IQT data for all neat aromatics at 10 bar, with iso-octane as a referencefrom Bogin et al. [207]. The two vertical lines represent the DCN test temperature(818.13±30 K) [15], although, the DCN test pressure is set at 21.4 bar, which is muchhigher than our test condition.
0 50 100 150 200 250 300 350 400
Time [ms]
5
10
15
20
25
Cha
mbe
r pr
essu
re [b
ar]
toluene 897 Kanisole 895 Kguaiacol 870 K
Figure 4.14: Pressure trace of toluene, anisole and guaiacol at specific temperaturein modified IQT at 10 bar
74
4.5. Conclusions
C
H
O
CH2O
O
O
Benzaldehyde
Cresoxy
Benzoxy
Bibenzyl
Phenoxy
Benzyl Phenyl
OO
OH
OH
OCH3OCH2
OH
O
OH
OCH3
Toluene
Anisole
Guaiacol
O
O
Benzoquinone
Phenol
CO + C2H2 + C2H3H2O + CO +C2H2 +C2H
Figure 4.15: Simplified reaction path diagram for toluene, anisole and guaiacoloxidation based on the kinetic mode of toluene [116] and anisole [186]
C H
H
H
88.5+1.5102+1
O
C
H
H
H
102.6
63.2
96.9
O
C
H
H
H
56.3
O H
85.1TolueneAnisole
Guaiacol
Figure 4.16: Bond dissociation energies in toluene, anisole and guaiacol [106, 186]
In the second part, anisole and guaiacol which has one -OCH3 group and one -OH group,and alkylated derivatives thereof, were selected and evaluated in the SI engine and modifiedignition quality tester (IQT).
The SI engine and IQT results demonstrate that the presence of a methoxy group onthe benzene ring is responsible for the higher reactivity of anisole relative to toluene, thereference octane booster in this study. The further addition of a hydroxyl group to anisole,leaves the resulting guaiacols more reactive still.
Notwithstanding the lower anti-knock quality of guaiacol and veratrole, they still demon-strated comparable behavior as the reference RON 95 gasoline fuel and could thus be used toboost the octane rating of lower grade gasolines, which are still commonplace in many partsof the world.
Overall, anisole and methyl anisole were found to have the most favorable anti-knockquality amongst the studied aromatic oxygenates that can be derived from lignin, beingroughly 50-66% as effective an octane booster as toluene which is toxic. And because oftheir higher knock resistance quality compared to gasoline, fuels blended with anisole and
75
Chapter 4. Performance of Lignin Derived Compounds as Octane Boosters
methyl anisole can spark earlier than normal gasoline, thus achieving modest improvementsin efficiency, and volumetric fuel economy.
Benzyl alcohol and 2-phenyl ethanol, with a hydroxy group on the benzene ring, alsoshows a good anti-knock property, only slightly worse than anisole.
The anti-knock quality of aromatic oxygenates suffers from functionalization, whereby oxy-genation appears to be far more detrimental than alkylation. Lignin conversion stakeholdersshould take this into account in their cost-benefit analysis.
(Methyl) anisole is frequently cited in lignin literature as a depolymerization product, model com-pound and/or potential biofuel in its own right. This study examines the auto-ignition characteristicsof methyl anisole by means of a newly developed detailed chemical kinetics mechanism, which is infact an expansion of an earlier, validated mechanism for anisole. Kinetics results are subsequentlycompared to those obtained from the modified ignition quality tester (IQT) experiments at 10 barand for a stoichiometric air-fuel ratio. The investigated temperature range is from 800 to 950 K.
5.1 Introduction
Lignocellulose, comprising cellulose, hemi-cellulose and lignin, is the main constituent ofmost terrestrial flora. Of aforementioned compounds, lignin, a complex polymer of phenolicunits, is the only one that is comprised from aromatics. Its main monomer units are p-coumaryl, coniferyl alcohol and sinapyl alcohol [196] (Figure 5.1). To retrieve valuable fuels
Chapter 5. A detailed kinetic study of 4-methyl anisole oxidation
and chemicals from lignin, its molecular structure must first be depolymerized. Requisitecracking processes can be thermo-chemical, catalytic and/or biological in nature [258].
Various aromatics monomers are typically present in the resulting lignin crude oil. Mostof which contain hydroxyl and/or methoxy functionalities, including anisole [217, 242, 259],methyl anisole [220, 242, 259–261], guaiacol [229, 230, 234, 242, 259] and phenol [234, 242,262]. Various anisoles and guaiacols have been tested as renewable fuels in compression-ignition [263, 264] and spark-ignition [203, 265] engines. In the former and latter enginetype, these compounds showed a reduced sooting tendency and an improved fuel economy,respectively. Both attributes were attributed to the generally low reactivity, intrinsic to aromaticoxygenates, which manifested in more time for premixing (i.e., longer ignition delay) in theformer and higher anti-knock quality in the latter engine type.
As discussed above, the low reactive nature of anisole has been reported to translate intoimproved engine performance. Depolymerization of lignin, however, is likely to produce alsoalkylated versions of anisole. The goal of this paper is therefore to study how alkylation (e.g.,4-methyl anisole (4-MA)) of this oxygenate impacts its auto-ignition chemistry.
In order to arrive at a 4-MA mechanism, a validated anisole model developed earlier [186]must be expanded. Experimental data obtained from a modified ignite quality tester (IQT) areused to validate the present model.
5.2 Chemical kinetic modeling
Thermochemical data were evaluated by THERGAS software when possible [266], which isbased on Benson’s group- and bond-additivity methodology [267]. Some thermochemicalproperties and rate constants have to be calculated theoretically. In this study, the CBS-QB3[268] level of theory has been used for the calculation of electronic energies of stable speciesand transition states. Reaction coordinate calculations are performed to ensure that transitionstates correctly connect the reactants to the products.
Gaussian09 [269] software is used to execute these calculations. For all species considered,the enthalpies of formation (∆fH◦298) are estimated using atomization reaction energies [270].The references for atomization energies of each atom are taken from the CODATA database[271]. Spin-orbit corrections are taken into account [272] here. Entropies (S◦298) are calculatedconsidering also hindered rotors contributions, which are evaluated to correct for partitionfunctions.
Hindered internal rotations are treated using the following procedure. First, the potentialsof each internal rotation is calculated at the B3LYP/6-31G(d,p) level of theory using a relaxedenergy scan. Subsequently, the characteristics of the rotational potentials and the correspond-ing barriers are used to correct for the partition function, using Pitzer and Gwinn tabulations[273] and implemented in ChemRate [274].
High-pressure rate constants involved in the mechanism are calculated using transitionstate theory. Kinetic parameters as a function of the temperature are obtained in Arrheniusform (k = ATnexp(−E/RT)) by fitting the rate constant values retrieved from transition statetheory within a temperature window of 500 to 1500 K.
Thermochemical properties for 4-MA and the intermediatesCH2C6H4OCH3,CH2C6H4O,OC6H4CH3, OC6H4OCH3, OC6H4O, C5H4OCH3 and C5H4O are determined theoretically.
78
5.2. Chemical kinetic modeling
O
CH2
H
CH2
H
99.8
89.9
62.5
94.4
Figure 5.2: Relevant bond dissociation energies (kcal/mol) in 4-MA calculated at theCBS-QB3 level of theory
The calculated enthalpy of 4-MA at 298 K is -22.5 kcal/mol, which is in good agreement withexperimental data (−23.7± 0.5 kcal/mol [275]). Relevant bond dissociation energies (BDE) areshown in Figure 5.2.
The detailed kinetic model of 4-MA combustion is developed iteratively, based the anisole’smechanism [186], which already includes sub-mechanisms of benzene, toluene, phenol, C1-C4 compounds and so on. The rate constants of the new reactions are determined by analogyto other reactions, for example, most of the rate constants of 4-MA on the O − CH3 siteare adopted from the anisole model, and rate constants for the methyl group site are takenfrom earlier work on toluene [116]. Rate constants of reactions which we can not find goodsimilarities from the literature are obtained from quantum chemical calculations mentionedabove.
The reactions added to the base mechanism [186] for 4-methyl anisole oxidation are listedin Table 5.1, and the rational behind the values are listed below the table. Given the low BDE ofthe O-CH3 bond (62.5 kcal/mol) and the C-H bond in the methyl group (89.9 kcal/mol), theseare expected to be the most likely sites for initiating reactions. This will produce resonantlystabilized OC6H4CH3 and CH2C6H4OCH3 radicals by way of unimolecular decomposition(1,3), bimolecular reactions and H atom abstraction (5, 20-33).
These types of reactions also happen at the O-CH3 site, yielding CH3C6H4OCH2 (2, 4,11-19). H atom addition to 4-methyl anisole leads to the formation of toluene or anisole (6,7). The addition of OH or methyl radicals will form cresol or hydroxy anisole and xylene,respectively (8, 9, 10). Thereafter, most cresoxy radicals (OC6H4CH3) will go on to producecresol after bonding with an H atom. Cresoxy radicals can also react via ipso-addition with anO atom, forming a CH3C6H3(O)O radical. Alternatively, cresoxy can react by bonding witha CH3 radical, producing toluene (41), or by losing an H atom at the CH3 site, producingCH2C6H4O (40). At high temperature, OC6H4CH3 radicals can also decompose to methylcyclopentadienyl (C5H5CH3) and CO (43).
The decomposition of the resonance stabilized CH2C6H4OCH3 radical by breaking theweak C-O bond, yields 4-methylene-2,5-cyclohexadiene-1-one (CH2C6H4O) ((39); Figure 5.3).This compound was detected among 4-MA pyrolysis products [276]. The reaction rateconstants are calculated theoretically at the CBS-QB3 level of the theory.
Besides decomposition to CH2C6H4O, CH2C6H4OCH3 can also react by terminationwith HO2, producing CH2OOHC6H4OCH3 (CH2OOH-ani, reaction (38)), an important path-way at low temperature. CH2OOH-ani can then decompose by breaking the hydroperoxygroup yielding OCH2C6H4OCH3 and CHOC6H4OCH3 (CHO-ani) (50-51), apart from that,
79
Chapter 5. A detailed kinetic study of 4-methyl anisole oxidation
O
H2C
O
H2C
+ CH3
O
H3C
O
H2C
+ H
Figure 5.3: Formation of CH2C6H4O
CH2OOHC6H4O radicals can be produced by decomposition reaction from O − CH3 bond(49).
The combination of an H atom to CH2OOHC6H4O produces CH2OOHC6H4OH (68),followed by bond rupture at the CH2O−OH site, yielding HOC6H4CH2O (69). This species,in turn, reacts further to phenol and CHO radicals (70).
OCH2C6H4OCH3, by analogy with reactions of benzoxy radicals proposed by Da Silva[277], decomposes via any of three pathways, producing either CHO-ani (52), anisole (53) orC6H4OCH3 (54).
CHO-ani can react by H atom abstraction, thereby losing CO and producing C6H4OCH3(55-60). C6H4OCH3 goes on to react with an H atom to form anisole (61). Alternatively, it cancombine with an OH radical (62), or react with O2, thereby producing OC6H4OCH3 radicals(64), which then reacts either by CO elimination, producing C5H4OCH3 (66), or analogous toCH2C6H4OCH3, decomposes at the O-CH3 site, producing 4-benzoquinone (OC6H4O)(65).
Reaction with O atoms (63), important at high temperature, leads to small species suchas CO, CH3 and cyclopentadienone (C5H4O). C5H4OCH3 can also decompose at the O-CH3bond, yielding C5H4O (67). The rate constants for reactions (65) and (67) are calculatedtheoretically at the CBS-QB3 level of theory in this study.
CH3C6H4OCH2 can isomerize via a cyclic transition state to CH3C6H4CH2O, seeing asthe benzoxy radical reaction proposed by Da Silva [277] and the rate constants of ensuingreactions are equal to those for C6H5OCH2 [186].
Reactions with CH2C6H4O involve addition reactions of H atoms and OH radicals onthe double bond sites in the ring or on the side methylene group (Figure 5.4). The kineticconstants for HOC6H4CH2 radical formation by way of H atom addition on the ketone group(44) are acquired from the reverse reaction of hydroxybenzyl decomposition to benzaldehydeand H atoms [277].
Rate constants for OH addition reactions to the methylene group are taken from knownbutadiene addition reactions, with constant A reduced by half since only one site can reactinstead of two, as is the case for butadiene. When OH radicals bond to the ring, β-scissionreactions are assumed to soon follow, thereby forming smaller compounds, including CO,acetylene, propanyl, butanyl and ethenone (47, 48). The sub-mechanisms for toluene, benzene,cresol, xylene, benzoquinone and various small hydrocarbons had already been included inthe anisole mechanism [115, 255].
80
5.2. Chemical kinetic modeling
O
H2C
+ H
H2C
OH
O
H2C
+ OH
H2C O
O H
O
+ CH2O
H2C O
H O
O
O
+ CH3
OH2C
O H
+ 2CO + C2H2
OH2C
H O
+CO+CH2CO
Figure 5.4: Addition reactions for CH2C6H4O
Table 5.1: Primary mechanism for methyl anisole oxidation
Rate constants (k = ATnexp(−Ea/RT)) are given in cm, mol, s, kcal unitsCH2-ani: CH2C6H4OCH3; tol-OCH2: CH3C6H4OCH2;OCH2-ani: OCH2C6H4OCH3; aci-phe: CH2OOHC6H4OHa: Rate constants taken equal to those of anisole [186]c: Rate constants taken equal to those of toluene [116]d: Rate constants taken equal to those of the phenoxymethyl reaction [277]e: Rate constants taken equal to those of benzyl radicals [115]f: Rate constants taken equal to those of benzyl radicals [277]g: Theoretically calculated in this study at the CBS-QB3 levelh: Rate constants taken equal to those of phenoxy radicals [195]i: Rate constants taken equal to those of the phenoxy radicals [92]j: Rate constants taken equal to those of phenoxy radicals [195], A×2k: Rate constants taken equal to those of benzaldehyde [101]l: Rate constants taken equal to those of styrene [115]m: Rate constants taken equal to those of benzylhydroperoxide [101]n: Rate constants taken equal to those of benzoxy radicals [277]o: Rate constants taken equal to those of benzaldehyde [278]p: Rate constants taken equal to those of C6H5CO radicals [115]q: Rate constants taken equal to those of phenyl radicals [279]
83
Chapter 5. A detailed kinetic study of 4-methyl anisole oxidation
Figure 5.5: Simulated (solid lines) and measured (symbols) ignition delays of anisoleand 4-MA as a function of temperature at 10 bar and stoichiometric mixing condi-tions
5.3 Model validation
5.3.1 Ignition quality tester experiments
A modified ignite quality tester (IQT) data has proved to be a reliable tool for the validationof kinetic mechanisms, particularly for fuels with low volatility [280], such as the aromaticoxygenates that are subject of investigation in this study. In Figure 5.5, anisole and 4-MAmodel simulations, performed using Chemkin II software [281], are compared to the modifiedIQT results.
Both data sets show a slightly higher reactivity for the methylated compound, particularlyat low temperatures (T< 1000 K). With the exception of a few outliers, the auto-ignition delaytime (ID) is consistently longer for the modified IQT experiments, irrespective of fuel ortemperature. Similar disparity between the two ID methods has been reported elsewhere[280]. At low temperature, 4-MA has shorter ID than anisole, when temperature increases,the difference become smaller, and the two compounds show similar ID. This trend is samein the modified IQT data and the model, the experiments show the two compounds havesimilar ID especially when temperature higher than 900 K, while the model shows moreapparent variation, and the data begin to converge at around 1000K.
84
5.3. Model validation
OCH3
O OCH3
O OH
8.4% 70%
55.3%20.3%
31.3% 5.1%
97.8% 78.3%
63.8% 54.1%
CH2O
99.9%99.1% O
65.9% 99.2%
OCH3
29.3%
OH
28.7% 34.4%
H2C
HOOH2C
33.8%27.2%
O O
O
16.3%17.7%
16.3% 17.7%
CO + C2H2 + CH3
CO + CH2CO + C4H5
32.6% 35.4%
OCH3
OH2C
100%100%
4.7% 7.0%
14.7%
96.1%
Figure 5.6: Reaction flux in a simulated PSR reactor at 10 bar and stoichiometricmixing conditions with 0.5% 4-MA, normal: at 800 K, italic: at 900 K
5.3.2 Reaction flux simulations
The reaction flux can be simulated by considering a homogeneous mixture in a perfectlystirred reactor (PSR). This exercise will allow us to a better grip on the primary reactionpathways and understand why 4-MA is more reactive than anisole. The reaction flux schemefor 4-MA is shown in Figure 5.6, wherein the values are computed at 800 K and 900 K (italic)at a residence time of 2s, corresponding to 55.4% and 99% 4-MA consumption, respectively.
H atom abstraction reactions at the methoxy site are the main consumption route at800 K (55.3%), producing CH3C6H4OCH2, which reacts further to CH3C6H4CH2O. H atomabstractions at the methyl site is also important (31.3%) and followed up by cleavage of theO-CH3 bond, yielding CH2C6H4O, or the reaction with HO2 radicals is likely as well (29.3%)at low temperature. Consumption via this route drops sharply to negligible levels above 950K.
Cleavage of the O-CH3 bond produces cresoxy radicals, a resonance stable compound.This reaction becomes more important when the temperature increases, in which case cresoxyradicals can react to cresol, although some cresol might revert back to cresoxy radicals via Hatom abstraction. This pathway accounts for only 8.4% at low temperatures, but up to 70% at950K, which is one of the reasons why 4-methyl anisole has a relatively fast reaction rate atlow temperature compared to anisole, but similar kinetics at higher temperatures.
Besides aforementioned reversion to cresoxy radicals, cresol might also react with CH3radicals to form toluene. Alternatively, cresoxy radicals can also produce methyl cyclopentadi-enyl by CO elimination (14.7% at 950 K), which can subsequently prompt the production ofbenzene by rapid rearrangement [282]. Formation of stable molecules such as benzene andtoluene has a strong inhibitory effect on the overall reaction rate, which might help to explainwhy the ID of 4-methyl anisole is seen to converge with that of anisole at higher temperatures.More experimental data would be necessary to further validate the model.
85
Chapter 5. A detailed kinetic study of 4-methyl anisole oxidation
5.4 Conclusions
Building on an existing model for anisole, a detailed kinetic mechanism has been developedfor the oxidation of 4-methyl anisole (4-MA ), a well-known lignin model compound andconstituent of lignin oil. Model results are in good agreement with modified ignition qualitytester (IQT) data and consistently show a shorter auto-ignition delay time for 4-MA, particularlyat low temperature. Subsequent flux analysis helps to explain why this is the case. At lowtemperature, the methyl group in 4-MA offers additional sites for H atom abstraction at lowtemperature. At higher temperatures, the formation of toluene and benzene appear to counteract this, resulting in near convergence of the ID for anisole and 4-MA at high temperatures.
86
Chapter 6Conclusions
Improving the knock resistance of gasoline is an important route to further increase theefficiency of spark ignition engines. By blending bio-based octane boosters to gasoline, CO2emissions are thus tackled from two sides, namely directly via higher engine efficiency and,indirectly, by means of a lower carbon footprint of the fuel. Conversion of biomass towardsoctane boosters is not only interesting from an ecological point of view. Per unit of energy,octane boosters command far higher market prices than gasoline. Accordingly, renewableoctane boosters might offer a step up in the value pyramid.
Recent advancements in engine technology have resulted in ever higher pre-combustiontemperatures, albeit ever cooler relative to pressure. Combining both developments entails thatfuture gasolines have to possess both a high octane number and high sensitivity, respectively.To attain such properties, the molecular composition of the fuel should contain strong covalentbonds. In hydrocarbons, this implies a short (average) carbon chain length, yielding moreof the strong primary C-H bonds, and a high degree of unsaturation, giving rise to strongcarbon-carbon bonds.
Application of these design rules leads to the recommendation that carbon chains in futuregasolines should be either highly branched, short chained or cyclic in nature. Irrespectiveof skeletal structure, however, a high degree of unsaturation is pivotal to increase sensitivityand raise the octane number even further. Unsaturation can be realized by addition of aC=C bond, or including a benzene or furan ring in the molecular structure. For straightchain compounds, inclusion of a functional oxygen group, such as a ketone, ester or hydroxygroup, can enhance the knock resistance of the compound, particularly so when it is placedin the middle of the chain. This is because the resulting structure promotes the formationof stable intermediates during the auto-ignition process. Ether groups, conversely, provide anew weakest link in the chain, which can be either positive or negative for anti-knock qualitydepending on the stability of the compounds that break off at these sites.
As will be evidenced below, this general insight is in line with the main conclusions ofthe various fuel studies included in this Thesis.
Chapter 3 investigates the anti-knock quality of various sugar-derived levulinic esters
87
Chapter 6. Conclusions
and cyclic ethers. Attributable to the presence of a ketone and ester functional groups inmethyl and ethyl levulinate, the otherwise highly reactive long chained paraffins were foundto have an excellent anti-knock quality. As predicted by the aforementioned design rules,mid-chain ketones and esters lower the overall reactivity by promoting the formation of stableintermediates in the auto-ignition process.
The story of the cyclic ethers is a bit more complex. Ethers, as stated earlier, create weaklinks in carbon chains. In the event the split off molecular fractions are stable, however, etherfunctionalities can be beneficial. The opposite is true when less stable compounds form aftercleavage occurs at the ether site. Ethyl tetrahydrofurfuryl ether and furfuryl ethyl ether have atthe heart of their molecular structure an instable and stable ring, respectively. Consequently,and in line with the design rules, the latter shows a far superior knock resistance.
In Chapter 4, the knock resistance of various lignin-derived aromatic oxygenates isinvestigated. All share the same benzene ring base at their core and differentiate by means oftheir side chain(s). In principle, any extremity will dilute the effect of the otherwise highlystable unsaturated ring and manifest in decreased anti-knock quality. As benzene is highlytoxic, different functional oxygen and alkyl groups are added to the ring. The magnitude ofaforementioned penalty was found to be dependent on the type of functionalities. Our analysisdemonstrates that the reactivity rate increases significantly irrespective of the character ofthe functionality oxygen group. Distinctions amongst the groups are comparatively small.Addition of (short) alkyl chains, however, had in most cases only a negligible impact on knockperformance.
In the final Chapter, the new built kinetic model of 4-methyl anisole explains the impactof the two functional groups on the benzene ring. This should be read as an initial theoreticalvalidation for the design rules.
Future workPreliminary validation of the proposed design rules for future octane boosters has beenprovided by both experimental and theoretical work in this thesis. At present, however, alldesign rules focus only on the chemical impact of fuels and neglect physical effects, which inpractice will naturally also influence such processes as deposit formation, evaporation andmixing. It is well known, for example, that the evaporative cooling effect of ethanol in directinjected gasoline engines is at least in part responsible for its high octane rating. Concerningthe kinetic mechanism of 4-methyl anisole, more experiments are needed to further improvethe model. To this end, benzene, toluene, anisole and methyl anisole will be evaluated over awide range of temperatures, pressures and equivalence ratios in a rapid compression machinein cooperation with RWTH Aachen.
In addition to the above, their are various less pressing activities that would furtherchallenge the robustness of the proposed design rules:
1. Test the octane boosters over a wider range of blend ratios. Because some of thesecompounds have very high boiling points and/or viscosity compared to gasoline, largerdoses might manifest in yet unknown adverse behavior.
2. Emissions, other than CO2, have been neglected. Future experiments should include astudy into nitric oxides, carbon monoxide, unburnt hydrocarbon and soot.
3. Fuels have hitherto been evaluated in a now outdated port fuel injected engine. Switch-ing to a more modern direct injected engines, having a completely different mixing
88
process, will shed some light on the robustness of the design rules to the changingengine technology.
4. For more detailed modeling and validation of the design rules, other, more fundamentalcombustion properties have to be determined, foremost of which being (laminar) flamespeed.
89
Bibliography
[1] W. Leppard. The chemical origin of fuel octane sensitivity. SAE Tech. Pap., 902137,1990.
[2] J. B. Heywood. Internal Combustion Engine Fundamentals. McGraw-Hill, 1988.
[3] V. Mittal and J. B. Heywood. The relevance of fuel RON and MON to knock onset inmodern SI engines. SAE Tech. Pap., 2008-01-2414, 2008.
[4] G. Hartmann. Les moteurs et aéroplanes Antoinette. Hydroretro, 2007.
[5] M. Campestrini and P. Mock. European vehicle market statistics. International Councilon Clean Transportation, 2011.
[6] M. Beecham. The global market for automotive turbochargers. ABOUT Publishing Group,2005.
[7] P. Leduc, B. Dubar, A. Ranini, and G. Monnier. Downsizing of gasoline engine: anefficient way to reduce CO2 emissions. Oil & Gas Sci. Tech., 58:115–127, 2003.
[8] M. C. Zhao, F.and Laia and D. L. Harrington. Automotive spark-ignited direct-injectiongasoline engines. Prog. Energy Combust. Sci., 25:437–562, 1999.
[9] G. T Kalghatgi. Auto-ignition quality of practical fuels and implications for fuelrequirements of future SI and HCCI engines. SAE Tech. Pap., 2005-01-0239, 2005.doi: 10.4271/2005-01-0239.
[10] K. Kalghatgi, G. T.and Nakata and K. Mogi. Octane appetite studies in direct injectionspark ignition (DISI) engines. SAE Tech. Pap., 2005-01-0244, 2005.
[11] T. Kume, Y. Lwamoto, K. Lida, M. Murakami, K. Akishino, and H. Ando. Combustioncontrol technologies for direct injection SI engine. SAE Tech. Pap., 960600, 1996.
[12] J. B. Heywood and O. Z. Welling. Trends in performance characteristics of modernautomobile si and diesel engines. SAE Int. J. Engines, 2:1650–62, 2009.
[13] D. Bradley and R. A. Head. Engine autoignition: The relationship between octanenumbers and autoignition delay times. Combust. Flame, 147:171–184, 2006.
91
Bibliography
[14] V. Mittal and J. B. Heywood. The shift in relevance of fuel RON and MON to knockonset in modern SI engines over the last 70 years. SAE Tech. Pap., 2009-01-2622,2009.
[15] Annual book of ASTM Standards. American Society for Testing Materials, 5.05 edition,2003.
[16] A. G. Bell. The relationship between octane qualily and octane requirement. SAE Tech.Pap., 750935, 1975.
[17] J. C. Ingamells and E. R. Jones. Developing road octane correlations from octanerequirement surveys. SAE Tech. Pap., 810492, 1981.
[18] C. E. Spitler, E. E.and Davis and D. W. Houser. Seven years’ experience with roadoctane control and TML. In API Division of Refining Meeting, 1967.
[19] G. T. Kalghatgi. Fuel anti-knock quality - part I. engine studies. SAE Tech. Pap.,2001-01-3584, 2001.
[20] G. T. Kalghatgi. Fuel anti-knock quality- part II. vehicle studies - how relevant is motoroctane number (MON) in modern engines? SAE Tech. Pap., 2001-01-3585, 2001.
[21] P. Beck, C. Stevenson and P. Ziman. The impact of gasoline octane on fuel economyin modern vehicles. SAE Tech. Pap., 2006-01-3407, 2006.
[22] A. Bell. Modern SI engine control parameter responses and altitude effects with fuelsof varying octane sensitivity. SAE Tech. Pap., 2010-01-1454, 2010.
[23] A. Amer, H. Babiker, J. Chang, G. Kalghatgi, P. Adomeit, A. Brassat, and et al. Fueleffects on knock in a highly boosted direct injection spark ignition engine. SAE Int. J.Fuels Lubr., 5:1048–65, 2012.
[24] G. Kalghatgi, P. Risberg, and H. E. Ångstrom. A method of defining ignition quality offuels in HCCI engines. SAE Tech. Pap., 2003-01-1816, 2003.
[25] P. Risberg, G. Kalghatgi, and H. E. Ångström. Autoignition quality of gasoline-likefuels in HCCI engines. SAE Tech. Pap., 2003-01-3215, 2003.
[26] G. T. Kalghatgi and R. A. Head. The available and required auto-ignition quality ofgasoline-like fuels in HCCI engines at high temperatures. SAE Tech. Pap., 2004-01-1969, 2004.
[27] M. Hajbabaei, G. Karavalakis, J. W. Miller, M. Villela, K. H. Xu, and T. D. Durbin.Impact of olefin content on criteria and toxic emissions from modern gasoline vehicles.Fuel, 107:671–9, 2013.
[28] R. Perry and I. L. Gee. Vehicle emissions in relation to fuel composition. Sci. TotalEnviron., 169:149–156, 1995.
92
Bibliography
[29] R. Atkinson. Gas-phase tropospheric chemistry of organic compounds: A review. Atmos.Environ. A-gen, 24:1–41, 1990.
[30] A. Hochhauser. Review of prior studies of fuel effects on vehicle emissions. SAE Tech.Pap., 2004-28-0081, 2004.
[31] S. Zahran, S. Weiler, H. W. Mielke, and A. A. Pena. Maternal benzene exposure and lowbirth weight risk in the United States: A natural experiment in gasoline reformulation.Environ. Res., 112:139–146, 2011.
[32] E. O. Talbott, X. Xu, A. O. Youk, J. R. Rager, J. A. Stragand, and A. M. Malek. Riskof leukemia as a result of community exposure to gasoline vapors: A follow-up study.Environ. Res., 111:597–602, 2011.
[33] N. Morgan, A. Smallbone, A. Bhave, M. Kraft, R. Cracknell, and G. Kalghatgi. Mappingsurrogate gasoline compositions into RON/MON space. Combust. Flame, 157:1122–31,2010. doi: 10.1016/j.combustflame.2010.02.003.
[34] J. C. G. Andrae and R. A. Head. HCCI experiments with gasoline surrogate fuelsmodeled by a semidetailed chemical kinetic model. Combust. Flame, 156:842—-851,2009.
[35] M. Fikri, J. Herzler, R. Starke, C. Schulz, P. Roth, and G.T. Kalghatgi. Autoignitionof gasoline surrogates mixtures at intermediate temperatures and high pressures.Combust. Flame, 152:276–281, 2008.
[36] W. Q Han and C. D Yao. Research on high cetane and high octane number fuels andthe mechanism for their common oxidation and auto-ignition. Fuel, 150:29–40, 2015.doi: 10.1016/j.fuel.2015.01.090.
[37] N. Morgan, A. Smallbone, A. Bhave, M. Kraft, R. Cracknell, and G. Kalghatgi. Mappingsurrogate gasoline compositions into RON/MON space. Combust. Flame, 2010. doi:10.1016/j.combustflame.2010.02.003.
[38] H. S. Shen and M. A. Oehlschlaeger. The autoignition of C8H10 aromatics at moderatetemperatures and elevated pressures. Combust. Flame, 156:1053–62, 2009.
[39] H. Hamid and M. A. Ali. Handbook of MTBE and Other Gasoline Oxygenates. 2004.ISBN 9780824752019.
[40] M. Eyidogan, A. N. Ozsezen, M. Canakci, and A. Turkcan. Impact of alcohol–gasolinefuel blends on the performance and combustion characteristics of an SI engine. Fuel,89(10):2713–20, oct 2010. doi: 10.1016/j.fuel.2010.01.032.
[41] J. J. Dannenberg and K. Tanaka. Theoretical studies of radical recombination reactions.1. allyl and azaallyl radicals. J. Am. Chem. Soc., 107:671–4, 1985.
[42] J. A. Miller, R. J. Kee, and C. K. Westbrook. Chemical Kinetics and CombustionModeling. Annu. Rev. Phys. Chem., 41(1):345–387, 1990. doi: 10.1146/annurev.pc.41.100190.002021.
93
Bibliography
[43] J. A. Miller, M. J. Pilling, and J. Troe. Unravelling combustion mechanisms through aquantitative understanding of elementary reactions. Proc. Combust. Inst., 30(1):43–88,jan 2005. doi: 10.1016/j.proci.2004.08.281.
[44] J. Zádor, C. A. Taatjes, and R. X. Fernandes. Kinetics of elementar y reacti ons inlow-temperature autoignition chemistry. Proc. Combust. Inst., 37:371–421, 2011.
[45] F. Battin-Leclerc. Detailed chemical kinetic models for the low-temperature combustionof hydrocarbons with application to gasoline and diesel fuel surrogates. Prog. EnergyCombust. Sci., 34(4):440–498, 2008. doi: 10.1016/j.pecs.2007.10.002.
[46] J.M. Simmie. Detailed chemical kinetic models for the combustion of hydrocarbonfuels. Proc. Combust. Inst., 29:599–634, 2003.
[47] W. J. Pitz and C. J. Mueller. Recent progress in the development of diesel surrogatefuels. Prog. Energy Combust. Sci., 37(3):330–350, jun 2011.
[48] H. J. Curran, P. Gaffuria, W. J. Pitza, and C. K. Westbrooka. A comprehensive modelingstudy of iso-octane oxidation. Combust. Flame, 129:253–280, may 2002.
[49] D. F. Davidson, B. M. Gauthier, and R. K. Hanson. Shock tube ignition measurementsof iso-octane/air and toluene/air at high pressures. Proc. Combust. Inst., 30:1175—-82,2005.
[50] R. Minetti, M. Carlier, M. Ribaucour, E. Therssen, and L. R. Sochet. A rapid compressionmachine investigation of oxidation and auto-ignition of n-heptane: Measurements andmodeling. Combust. Flame, 102:298–309, 1995.
[51] Z. D. Wang, O. Herbinet, Z. J. Cheng, B. Husson, R. Fournet, F. Qi, and et al.Experimental investigation of the low temperature oxidation of the five isomers ofhexane. J. Phys. Chem. A, 118:5573–94, 2014. doi: 10.1021/jp503772h.
[52] V. P. Zhukov, V. A. Sechenov, and A. Y. Starikovskii. Ignition delay times in leann-hexane–air mixture at high pressures. Combust. Flame, 136(1-2):257–259, jan 2004.doi: 10.1016/j.combustflame.2003.10.002.
[53] P. A. Glaude, F. Battin-Leclerc, R. Fournet, V. Warth, G. M. Côme, and G. Scacchi.Construction and simplification of a model for the oxidation of alkanes. Combust. Flame,122(4):451–462, 2000.
[54] V Zhukov, V Sechenov, and A Starikoskii. Self-ignition of a lean mixture of n-pentaneand air over a wide range of pressures. Combust. Flame, 140:196–203, February 2005.
[55] D. Healy, N. S. Donato, C. J. Aul, E. L. Petersen, C. M. Zinner, G. Bourque, and et al.Isobutane ignition delay time measurements at high pressure and detailed chemicalkinetic simulations. Combust. Flame, 157:1540–51, 2010.
[56] O. Herbinet, F. Battin-Leclerc, S. Bax, H. Gall, P. A. Glaude, R. Fournet, and et al.Detailed product analysis during the low temperature oxidation of n-butane. Phys.Chem. Chem. Phys., 13(1):296–308, 2011. doi: 10.1039/C0CP00539H.
94
Bibliography
[57] M. Cord, B. Sirjean, R. Fournet, A. Tomlin, M. Ruiz-Lopez, and F. Battin-Leclerc.Improvement of the modeling of the low-temperature oxidation of n-butane: Study ofthe primary reactions. J. Phys. Chem. A, 116(24):6142–58, 2012. doi: 10.1021/jp211434f.PMID: 22257166.
[58] S. M. Sarathy, C.K. Westbrook, M. Mehl, W. J. Pitz, C. Togbe, P. Dagaut, and et al.Comprehensive chemical kinetic modeling of the oxidation of 2-methylalkanes fromC7 to C20. Combust. Flame, 158:2338–57, 2011.
[59] R. Asatryan and J. W. Bozzelli. Chain branching and termination in the low-temperaturecombustion of n -alkanes: 2-pentyl radical+O2, isomerization and association of thesecond O2. J. Phys. Chem. A, 114:7693–7708, 2010.
[60] F. Buda, R. Bounaceur, V. Warth, P. A. Glaude, R. Fournet, and F. Battin-Leclerc.Progress toward a unified detailed kinetic model for the autoignition of alkanes fromC4 to C10 between 600 and 1200 K. Combust. Flame, 142:170–186, 2005.
[61] C. K. Westbrook, W. J. Pitz, O. Herbinet, H. J. Curran, and E. J. Silke. A comprehensivedetailed chemical kinetic reaction mechanism for combustion of n-alkane hydrocarbonsfrom n-octane to n-hexadecane. Combust. Flame, 156(1):181–199, jan 2009. doi:10.1016/j.combustflame.2008.07.014.
[62] K. Zhang, C. Banyon, C. Togbé, P. Dagaut, J. Bugler, and H. J. Curran. An experimentaland kinetic modeling study of n-hexane oxidation. Combust. Flame, 160:2729–43,September 2015.
[63] Z. Wang, L. Zhang, K. Moshammer, D. M. Popolan-Vaida, V. S. B. Shankar, A. Lucassen,and et al. Additional chain-branching pathways in the low-temperature oxidation ofbranched alkanes. Combust. Flame, 164:386–396, 2016.
[64] S. M. Sarathy, T. Javed, F. Karsenty, A. Heufer, W. Wang, S. Park, and et al. Acomprehensive combustion chemistry study of 2,5-dimethylhexane. Combust. Flame,161(6):1444–59, jun 2014. doi: 10.1016/j.combustflame.2013.12.010.
[65] J. Bugler, B. Marks, O. Mathieu, R. Archuleta, A. Camou, C. Grégoire, and et al.An ignition delay time and chemical kinetic modeling study of the pentane isomers.Combust. Flame, 163:138–156, jan 2016. doi: 10.1016/j.combustflame.2015.09.014.
[66] J. Bugler, K. P. Somers, E. J. Silke, and H. J. Curran. Revisiting the kinetics andthermodynamics of the low-temperature oxidation pathways of alkanes: A case studyof the three pentane isomers. J. Phys. Chem. A, 119(28):7510–27, 2015. doi: 10.1021/acs.jpca.5b00837.
[67] H. J. Curran, P. Gaffuri, W. J. Pitz, and C. K. Westbrook. A comprehensive modelingstudy of n-heptane oxidation. Combust. Flame, 114:149–177, 1998.
[68] J. W. Bozzelli and W. J. Pitz. The reaction of hydroperoxy-propyl radicals with molecularoxygen. Symp. Combust., 25:783–791, 1994.
95
Bibliography
[69] C. A. Morgan, J. Pilling, J. M. Tulloch, R. P. Ruiz, and K. D. Bayes. Direct determinationof the equilibrium constant and thermodynamic parameters for the reaction. J. Chem.Soc. Faraday Trans. Mol. Chem. Phys., 78:1323–30, 1982.
[70] Charles K. Westbrook. Chemical kinetics of hydrocarbon ignition in practical combus-tion systems. Proc. Combust. Inst., 28(2):1563–1577, jan 2000.
[71] S. Saxena and I. D. Bedoya. Fundamental phenomena affecting low temperaturecombustion and HCCI engines, high load limits and strategies for extending theselimits. Proc. Combust. Inst., 39:457–488, 2013.
[72] ASTM. Knocking characteristics of pure hydrocarbons. Technical report, 1958.
[73] S. Tanaka, F. Ayala, J. C Keck, and J. B Heywood. Two-stage ignition in HCCIcombustion and HCCI control by fuels and additives. Combust. Flame, 132(1-2):219–239,2003.
[74] Y. Yang, A. Boehman, and J. M. Simmie. Uniqueness in the low temperature oxidationof cycloalkanes. Combust. Flame, 157(12):2357–2368, 2010. doi: 10.1016/j.combustflame.2010.06.005.
[75] B. Sirjean, F. Buda, H. Hakka, P.A. Glaude, R. Fournet, V. Warth, and et al. Theautoignition of cyclopentane and cyclohexane in a shock tube. Proc. Combust. Inst., 31(1):277–284, 2007.
[76] O. Lemaire, M. Ribaucour, M. Carlier, and R. Minetti. The production of benzene inthe low-temperature oxidation of cyclohexane, cyclohexene, and cyclohexa-1,3-diene.Combust. Flame, 127(1-2):1971–80, 2001. doi: 10.1016/S0010-2180(01)00301-7.
[77] Z. Serinyel, O. Herbinet, O. Frottier, P. Dirrenberger, V. Warth, P. A. Glaude, and et al.An experimental and modeling study of the low- and high-temperature oxidation ofcyclohexane. Combust. Flame, 160(11):2319–32, 2013. doi: 10.1016/j.combustflame.2013.05.016.
[78] B. W. Weber, W. J. Pitz, Mehl. M., E. J. Silke, A. C. Davis, and C. J. Sung. Experimentsand modeling of the autoignition of methylcyclohexane at high pressure. CombustFlame, 161(8):1972 – 83, 2014.
[79] S. S. Vasu, D. F. Davidson, Z. Hong, and R. K. Hanson. Shock tube study of methylcy-clohexane ignition over a wide range of pressure and temperature. Energy Fuels, 23(1):175–185, 2009. doi: 10.1021/ef800694g.
[80] W. J. Pitz, C. V. Naik, T. Ní Mhaoldúin, C. K. Westbrook, H. J. Curran, J. P. Orme, andet al. Modeling and experimental investigation of methylcyclohexane ignition in a rapidcompression machine. Proc. Combust. Inst., 2007.
[81] E. J. Silke, W. J. Pitz, C. K. Westbrook, and M. Ribaucour. Detailed chemical kineticmodeling of cyclohexane oxidation. J. Phys. Chem. A, 111(19):3761–3775, may 2007.
96
Bibliography
[82] M. J Al Rashidi, S. Thion, C. Togbé, G. Dayma, M. Mehl, P. Dagaut, and et al. Elu-cidating reactivity regimes in cyclopentane oxidation: Jet stirred reactor experiments,computational chemistry, and kinetic modeling. Proc. Combust. Inst., 000:1–9, 2016.
[83] M. Mehl, W. J. Pitz, C. K. Westbrook, K. Yasunaga, C. Conroy, and H. J. Curran.Autoignition behavior of unsaturated hydrocarbons in the low and high temperatureregions. Proc. Combust. Inst., 33:201––8, 2011.
[84] M. Yahyaoui, N. Djebaïli-Chaumeix, C. E. Paillard, S. Touchard, R. Fournet, P. A.Glaude, and et al. Experimental and modeling study of 1-hexene oxidation behindreflected shock waves. Proc. Combust. Inst., 30:1137–45, 2005.
[85] M. Mehl, T. Faravelli, F. Giavazzi, E. Ranzi, P. Scorletti, A. Tardani, and et al. Detailedchemistry promotes understanding of octane numbers and gasoline sensitivity. EnergyFuel, 20:2391–8, 2006.
[86] M. Mehl, G. Vanhove, W. J. Pitz, and E. Ranzi. Oxidation and combustion of the n-hexene isomers: A wide range kinetic modeling study. Combust. Flame, 155:756—-772,2008.
[87] C. K. Westbrook, W. J. Pitz, M. Mehl, P. A. Glaude, O. Herbinet, S. Bax, and et al.Experimental and kinetic modeling study of 2-methyl-2-butene: Allylic hydrocarbonkinetics. J. Phys. Chem. A, 119(28):7462–80, 2015. doi: 10.1021/acs.jpca.5b00687.
[88] C. K. Westbrook, W. J. Pitz, S. M. Sarathy, and M. Mehl. Chemical kinetic modeling ofhydrocarbon combustion. Proc. Combust. Inst., 34:3049—-56, 2013.
[89] G. Vanhove, M. Ribaucour, and R. Minetti. On the influence of the position of thedouble bond on the low-temperature chemistry of hexenes. Proc. Combust. Inst., 30:1065—-72, 2005.
[90] R. Bounaceur, V. Warth, B. Sirjean, P. A. Glaude, R. Fournet, and F. Battin-Leclerc.Influence of the position of the double bond on the autoignition of linear alkenes atlow temperature. Proc. Combust. Inst., 32:387––394, 2009.
[91] K. Brezinsky. The high-temperature oxidation of aromatic hydrocarbons. Prog. EnergyCombust. Sci., 12:1–24, jan 1986.
[92] M. U. Alzueta, P. Glarborg, and K. D. Johansen. Experimental and kinetic modelingstudy of the oxidation of benzene. Int. J. Chem. Kinet., 32:498—-522, 2000.
[93] C. Saggese, A. Frassoldati, A. Cuoci, T. Faravelli, and E. Ranzi. A wide range kineticmodeling study of pyrolysis and oxidation of benzene. Combust. Flame, 160:1168—-90,2013.
[94] P. Dagaut, G. Pengloan, and A. Ristori. Oxidation, ignition and combustion of toluene:Experimental and detailed chemical kinetic modeling. Phys. Chem. Chem. Phys., 4:1846–54, may 2002.
97
Bibliography
[95] S. Yasuyuki, I. Takayoshi, O. Teppei, K. Mitsuo, and W. J. Pitz. Detailed kinetic modelingof toluene combustion over a wide range of temperature and pressure. SAE Tech. Pap.,2007-01-1885, 2007.
[96] W. K. Metcalfe, S. Dooley, and F.L. Dryer. Comprehensive detailed chemical kineticmodeling study of toluene oxidation. Energy Fuels, 25:4915––36, 2011.
[97] G. Mittal and C. Sung. Autoignition of toluene and benzene at elevated pressures in arapid compression machine. Combust. Flame, 150:355––368, 2007.
[98] W. Yuan, Y. Li, P. Dagaut, J. Yang, and F. Qi. Investigation on the pyrolysis and oxidationof toluene over a wide range conditions. I. Flow reactor pyrolysis and jet stirred reactoroxidation. Combust. Flame, 162(1):3–21, 2015. doi: 10.1016/j.combustflame.2014.07.009.
[99] W. Yuan, Y. Li, P. Dagaut, J. Yang, and F. Qi. Investigation on the pyrolysis and oxidationof toluene over a wide range conditions. II. A comprehensive kinetic modeling study.Combust. Flame, 162:22–40, 2015. doi: 10.1016/j.combustflame.2014.07.011.
[100] Y. Zhang, K. P. Somers, M. Mehl, W. J. Pitz, R. F. Cracknell, and H. J. Curran. Probingthe antagonistic effect of toluene as a component in surrogate fuel models at lowtemperatures and high pressures. A case study of toluene/dimethyl ether mixtures.Proc. Combust. Inst., 000:1–9, 2016. doi: 10.1016/j.proci.2016.06.190.
[101] G Da Silva and J Bozzelli. Kinetic modeling of the benzyl+HO2 reaction. Proc. Combust.Inst., 32:287–294, 2009. doi: 10.1016/j.proci.2008.05.040.
[102] L. Zhao, Z. Cheng, L. Ye, F. Zhang, L. Zhang, F. Qi, and et al. Experimental and kineticmodeling study of premixed o-xylene flames. Proceedings of the Combustion Institute, 35(2):1745 – 1752, 2015.
[103] D. Darcy, C. J. Tobin, K. Yasunaga, J. M. Simmie, J. Würmel, W. K. Metcalfe, and et al.A high pressure shock tube study of n-propylbenzene oxidation and its comparisonwith n-butylbenzene. Combust. Flame, 159:2219–32, 2012.
[104] F. Battin-Leclerc, V. Warth, R. Bounaceur, B. Husson, O. Herbinet, and P. A. Glaude.The oxidation of large alkylbenzenes: An experimental and modeling study. Proc.Combust. Inst., 35:349–356, 2015. doi: 10.1016/j.proci.2014.05.087.
[105] J. M. Nicovlch, C. A. Gump, and A. R. Ravlshankar. Rates of reactions of O(3P) withbenzene and toluene. J. Phys. Chem., 86:1684–90, 1982.
[106] Y.R. Luo. Handbook of Bond Dissociation Energies in Organic Compounds. 2003. ISBN0849315891.
[107] S. J. Blanksby and G. B. Ellison. Bond dissociation energies of organic molecules. Acc.Chem. Res., 36:255–263, April 2003.
98
Bibliography
[108] L. S. Tran, C. Togbé, D. Liu, D. Felsmann, P. Oßwald, P. A. Glaude, and et al. Com-bustion chemistry and flame structure of furan group biofuels using molecular-beammass spectrometry and gas chromatography – Part III: 2,5-Dimethylfuran. Combust.Flame, 161(3):780–797, mar 2014. doi: 10.1016/j.combustflame.2013.05.026.
[109] K. C. Salooja. Studies of combustion processes leading to ignition of aromatic hydro-carbons. Combust Flame, 9:121––129, 1965.
[110] A. Roubaud, O. Lemaire, R. Minetti, and L. R. Sochet. High pressure auto-ignition andoxidation mechanisms of o-xylene, o-ethyltoluene, and n-butylbenzene between 600and 900 K. Combust. Flame, 123:561–571, 2000.
[111] F. Battin-Leclerc, R. Bounaceur, N. Belmekki, and P. A. Glaude. Experimental andmodeling study of the oxidation of xylenes. Int. J. Chem. Kinetic, 38:284—-302, 2006.
[112] S. Gaïl and P. Dagaut. Oxidation of m-Xylene in a JSR Experimental Study and DetailedChemical Kinetic Modeling. Combust. Sci. Technol., 179:813–844, apr 2007.
[113] S. Gaïl, P. Dagaut, G. Black, and J. M. Simmie. Kinetics of 1,2-dimethylbenzeneoxidation and ignition: Experimental and detailed chemical kinetic modeling. Combust.Sci. and Tech., 180:1748—-71, 2008.
[114] P. Diévart, H. H. Kim, S. H. Won, Y. G. Ju, F. L. Dryer, S. Dooley, and et al. Thecombustion properties of 1,3,5-trimethylbenzene and a kinetic model. Fuel, 2013. doi:10.1016/j.fuel.2012.11.069.
[115] R. Bounaceur, I. Da Costa, R. Fournet, F. Billaud, and F. Battin-Leclerc. Experimentaland modeling study of the oxidation of toluene. Int. J. Chem. Kinet., 37:25–49, 2005.
[116] Z. Tian, W. J. Pitz, R. Fournet, P. A. Glaude, and F. Battin-Leclerc. A detailed kineticmodeling study of toluene oxidation in a premixed laminar flame. Proc. Combust. Inst.,33:233–241, 2011. doi: 10.1016/j.proci.2010.06.063.
[117] B. Kenneth. The high-temperature oxidation of aromatic hydrocarbons. Prog. EnergyCombust. Sci., 12:1–24, 1986.
[118] M. Thewes, M. Muether, S. Pischinger, M. Budde, A. Brunn, A. Sehr, and et al. Analysisof the impact of 2-methylfuran on mixture formation and combustion in a direct-injection spark-ignition engine. Energy Fuels, 25(12):5549–61, 2011. doi: 10.1021/ef201021a.
[119] C. M. Wang, H. M. Xu, R. Daniel, A. Ghafourian, J. M. Herreros, S. J. Shuai, andet al. Combustion characteristics and emissions of 2-methylfuran compared to 2,5-dimethylfuran, gasoline and ethanol in a DISI engine. Fuel, 103, 2013. doi: 10.1016/j.fuel.2012.05.043.
[120] A. Sudholt, L. Cai, J. Heyne, F. M. Haas, H. Pitsch, and F. L. Dryer. Ignition character-istics of a bio-derived class of saturated and unsaturated furans for engine applications.Proc. Combust. Inst., 35:2957–65, 2015. doi: 10.1016/j.proci.2014.06.147.
99
Bibliography
[121] L. S. Tran, C. Togbé, D. Liu, D. Felsmann, P. Oßwald, P. A. Glaude, and et al. Com-bustionchemistry and flame structure of furan group biofuels using molecular-beammass spectrometry and gas chromatography – Part II: 2-Methylfuran. Combust. Flame,161:766–779, 2014. doi: 10.1016/j.combustflame.2013.05.027.
[122] J. M. Simmie and H. J. Curran. Formation enthalpies and bond dissociation energiesof alkylfurans. the strongest C-X bonds known? J. Phys. Chem. A, 113:5128–37, 2009.
[123] L. S. Tran, C. Togbé, D. Liu, D. Felsmann, P. Oßwald, P. A. Glaude, and et al. Com-bustion chemistry and flame structure of furan group biofuels using molecular-beammass spectrometry and gas chromatography – Part II: 2-Methylfuran. Combust. Flame,161(3):766–779, mar 2014. doi: 10.1016/j.combustflame.2013.05.027.
[124] A. C. Davis and S. M. Sarathy. Computational study of the combustion and atmosphericdecomposition of 2-methylfuran. J. Phys. Chem. A, 117(33):7670–85, aug 2013. doi:10.1021/jp403085u.
[125] B. Sirjean, R. Fournet, P. A. Glaude, F. Battin-Leclerc, W. J. Wang, and M. A.Oehlschlaeger. Shock Tube and Chemical Kinetic Modeling Study of the Oxidation of2,5-Dimethylfuran. J. Phys. Chem. A, 117(7):1371–92, feb 2013. doi: 10.1021/jp308901q.
[126] K. P. Somers, J. M. Simmie, F. Gillespie, U. Burke, J. Connolly, W. K. Metcalfe, and et al.A high temperature and atmospheric pressure experimental and detailed chemicalkinetic modelling study of 2-methyl furan oxidation. Proc. Combust. Inst., 34(1):225–232,2013. doi: 10.1016/j.proci.2012.06.113.
[127] J. T. Moss, A. M. Berkowitz, M. A. Oehlschlaeger, J. Biet, V. Warth, P. A. Glaude, andF. Battin-Leclerc. An experimental and kinetic modeling study of the oxidation of thefour isomers of butanol. J. Phys. Chem.A, 112:10843—-55, 2008.
[128] T. S. Norton and F. L. Dryer. The flow reactor oxidation of C1-C4 alcohols and MTBE.Proc. Combust. Inst., 23:179–185, 1990.
[129] T. Tsujimura, W. J. Pitz, F. Gillespie, H. J. Curran, B. W. Weber, Y. Zhang, and et al.Development of isopentanol reaction mechanism reproducing autoignition characterat high and low temperatures. Energy Fuels, 26:4871—-86, 2012.
[130] S. M. Sarathy, S. Vranckx, K. Yasunaga, M. Mehl, P. Oß wald, W. K. Metcalfe, and et al.A comprehensive chemical kinetic combustion model for the four butanol isomers.Combust. Flame, 159:2028–55, 2012.
[131] G. E. Quelch, M. M. Gallo, M. Z. Shen, Y. M. Xie, H. F. Schaefer, and D. Moncrieff. As-pects of the reaction mechanism of ethane combustion. 2. nature of the intramolecularhydrogen transfer. J. Am. Chem. Soc., 116:4953–62, 1994.
[132] M. S. Stark and D. J. Waddington. Oxidation of propene in the gas phase. Int. J. Chem.Kinet., 27:123–151, 1995.
100
Bibliography
[133] S. M. Sarathy, P Oßwald, N. Hansen, and K. Kohse-Höinghaus. Alcohol combustionchemistry. Prog. Energy Combust. Sci., 44:40–102, 2014.
[134] B. W. Weber and C. J. Sung. Comparative autoignition trends in butanol isomers atelevated pressure. Energy Fuels, 27:1688–98, 2013.
[135] G. da Silva, J. W. Bozzelli, L. Liang, and J. T. Farrell. Ethanol oxidation: Kinetics ofther-hydroxyethyl radical + o2. J. Phys. Chem.A, 113:8923—-33, 2009.
[136] H. Sun, J. W. Bozzelli, and C. K. Law. Thermochemical and kinetic analysis on the re-actions of O2 with products from oh addition to isobutene, 2-hydroxy-1,1-dimethylethyl,and 2-hydroxy-2-methylpropyl radicals: HO2 formation from oxidation of neopentane,part ii. J. Phys. Chem. A, 111:4974–86, 2007.
[137] T. Ogura, A. Miyoshi, and M. Koshi. Rate coefficients of H-atom abstraction fromethers and isomerization of alkoxyalkylperoxy radicals. Phys. Chem. Chem. Phys, 9:5133—-42, 5133, 2007.
[138] H. J. Curran, W. J. Pitz, C. K. Westbrook, P. Dagaut, J. C. Boettner, and M. Cathonnet.A wide range modeling study of dimethyl ether oxidation. Int. J. Chem. Kinetics, 30:229–241, 1998.
[139] H. J. Curran, S. L. Fischer, and F. L. Dryer. The reaction kinetics of dimethyl ether. II:Low temperature oxidation in flow reactors. Int. J. Chem. Kinet., 32:741–759, 2000.
[140] Z. Zhao and M. Chaos. Thermal decomposition reaction and a comprehensive kineticmodel of dimethyl ether. Int. J. Chem. Kinet., 40:1–18, 2008.
[141] B. Mittal and J. B. Heywood. The relevance of fuel RON and MON to knock onset inmodern SI engines. SAE Tech. Pap., 2008-01-2414, 2008. doi: 10.1016/j.fuproc.2008.05.021.
[142] W. Leppard. The autoignition chemistries of octane-enhancing ethers and cyclic ethers:A motored engine study. SAE Tech. Pap., 912313, 1991.
[143] P. A Glaude, F Battin-Leclerc, B Judenherc, V. Warth, R. Fournet, G.M. Côme, andet al. Experimental and modeling study of the gas-phase oxidation of methyl and ethyltertiary butyl ethers. Combust. Flame, 121:345–355, 2000.
[144] H. Böhm, F. Baronnet, and B. El Kadi. Comparative modelling study on the inhibitingeffect of TAME, ETBE and MTBE at low temperature. Phys. Chem. Chem. Phys., 2:1929–33, 2000.
[145] K. Yasunaga, J. M. Simmie, H. J. Curran, T. Koike, O. Takahashi, Y. Kuraguchi, andet al. Detailed chemical kinetic mechanisms of ethyl methyl, methyl tert-butyl and ethyltert-butyl ethers: The importance of uni-molecular elimination reactions. Combust.Flame, 158:1032–6, 2011.
101
Bibliography
[146] A. Goldaniga, T. Faravelli, E. Ranzi, P. Dagaut, and M. Cathonnet. Oxidation ofoxygenated octane improvers: MTBE, ETBE, DIPE, and TAME. Proc. Combust. Inst., 27:353—-360, 1998.
[147] A Ciajolo. Low-temperature oxidation of MTBE in a high-pressure jet-stirred flowreactor. Combust. Sci. Technol., 123:49–61, 1997.
[148] W. K. Metcalfe, S. Dooley, H. J. Curran, J. M. Simmie, A. M. El-Nahas, and M. V.Navarro. Experimental and modeling study of C5H10O2 ethyl and methyl esters. J.Phys. Chem. A, 111:4001–14, 2007.
[149] S Dooley, H. J. Curran, and J. M. Simmie. Autoignition measurements and a validatedkinetic model for the biodiesel surrogate, methyl butanoate. Combust. Flame, 153:2–32,2008.
[150] S. Gaïl, M. J. Thomson, S. M. Sarathy, S. A. Syed, P. Dagaut, P. Diévart, and et al. Awide-ranging kinetic modeling study of methyl butanoate combustion. Proc. Combust.Inst., 31:305–311, 2007.
[151] L. K. Huynh, K. C. Lin, and A. Violi. Kinetic modeling of methyl butanoate in shocktube. J. Phys. Chem. A, 112:13470–80, 2008.
[152] M. H. Hakka, H. Bennadji, J. Biet, M. Yahyaoui, B. Sirjean, V. Warth, and et al. Oxidationof methyl and ethyl butanoates. Int. J. Chem. Kinet., 42:226–252, 2010. doi: 10.1002/kin.20473.
[153] S. M. Walton, M. S. Wooldridge, and C. K. Westbrook. An experimental investigationof structural effects on the auto-ignition properties of two C5 esters. Proc. Combust.Inst., 32:255–262, 2009.
[154] A. M. El-Nahas, M. V. Navarro, J. M. Simmie, J. W. Bozzelli, H. J. Curran, S. Dooley,and et al. Enthalpies of formation, bond dissociation energies and reaction paths forthe decomposition of model biofuels: ethyl propanoate and methyl butanoate. J. Phys.Chem. A, 111:3727–39, 2007.
[155] G. Dayma, F. Halter, F. Foucher, C. Togbé, C. Mounaim-Rousselle, and P. Dagaut.Experimental and detailed kinetic modeling study of ethyl pentanoate (ethyl valer-ate) oxidation in a jet stirred reactor and laminar burning velocities in a sphericalcombustion chamber. Energy Fuels, 26:4735–48, 2012.
[156] P. A. Glaude, O. Herbinet, S. Bax, J. Biet, V. Warth, and F. Battin-Leclerc. Modeling ofthe oxidation of methyl esters-Validation for methyl hexanoate, methyl heptanoate, andmethyl decanoate in a jet-stirred reactor. Combust. Flame, 157:2035–50, 2010.
[157] K. HadjAli, M. Crochet, G. Vanhove, M. Ribaucour, and R. Minetti. A study of the lowtemperature autoignition of methyl esters. Proc. Combust. Inst., 32:239–246, 2009.
[158] G. Dayma, S. Gaïl, and P. Dagaut. Experimental and kinetic modeling study of theoxidation of methyl hexanoate. Energy Fuels, 22:1469–79, 2008.
102
Bibliography
[159] G. Dayma, C. Togbé, and P. Dagaut. Detailed kinetic mechanism for the oxidation ofvegetable oil methyl esters: New evidence from methyl heptanoate. Energy Fuels, 23:4254–68, 2009.
[160] O. Herbinet, W. J. Pitz, and C. K. Westbrook. Detailed chemical kinetic oxidationmechanism for a biodiesel surrogate. Combust. Flame, 154:507–528, 2008.
[161] C. K. Westbrook, C. V. Naik, O. Herbinet, W. J. Pitz, M. Mehl, S. M. Sarathy, and et al.Detailed chemical kinetic reaction mechanisms for soy and rapeseed biodiesel fuels.Combust. Flame, 158:742–755, 2011.
[162] J. Zhang, W. Jing, W. L. Roberts, and T. Fang. Soot measurements for diesel andbiodiesel spray combustion under high temperature highly diluted ambient conditions.Fuel, 135:340–351, 2014.
[163] L. Coniglio, H. Bennadji, P. A. Glaude, O. Herbinet, and F. Billaud. Combustionchemical kinetics of biodiesel and related compounds (methyl and ethyl esters): Experi-ments and modeling – advances and future refinements. Prog. Energy Combust. Sci.,39:340–382, August 2013.
[164] K. Kumar and C. J. Sung. Autoignition of methyl butanoate under engine relevantconditions. Combust. Flame, 171:1–14, 2016. doi: 10.1016/j.combustflame.2016.04.011.
[165] W. R. Schwartz, C. S. McEnally, and L. D. Pfefferle. Decomposition and hydrocarbongrowth processes for esters in non-premixed flames. J. Phys. Chem. A, 110:6643–8,2006.
[166] H O’Neal and S Benson. A method for estimating the Arrhenius A factors for four-and six-center unimolecular reactions. J. Phys. Chem., 71:2903–21, August 1967.
[167] N. Sebbar, J. W. Bozzelli, and Bockhorn. H. Thermochemistry and kinetics for 2-butanone-1-yl radical (CH2C(= O)CH2CH3) reactions with O2. J. Phys. Chem. A, 118:21–37, 2014.
[168] P. S. Veloo, P. Dagaut, C. Togbé, G. Dayma, S. M. Sarathy, C. K. Westbrook, and et al.Experimental and modeling study of the oxidation of n- and iso-butanal. Combust.Flame, 160:1609–26, 2013.
[169] P. Anbarasan, Z. C. Baer, S. Sreekumar, E. Gross, J. B. Binder, H. W. Blanch, and et al.Integration of chemical catalysis with extractive fermentation to produce fuels. Nature,491:235–9, 2012.
[170] J. M. E. Storey, S. A. Lewis, J. E. Parks, F. P. Szybist, T. L. Barone, and V. Y. Prikhodko.Mobile source air toxics (MSATs) from high efficiency clean combustion: catalyticexhaust treatment effects. SAE Int. J. Engines, 1:1157–66, 2008.
[171] S. A. Lewis, J. Storey, C. S. Sluder, K. Cho, R. Connatser, and T. Barone. Carbonylformation during high efficiency clean combustion of FACE fuels. SAE Int. J. FuelsLubr., 3:849–856, 2010.
103
Bibliography
[172] X. Wang, Y. Ge, L Liu, Z. Peng, L. Hao, H. Yin, and et al. Evaluation on toxic reductionand fuel economy of a gasoline direct injection- (GDI-) powered passenger car fueledwith methanol–gasoline blends with various substitution ratios. Appl. Energy, 157:134–143, 2015.
[173] M. A. Costagliola, L. De Simio, S. Iannaccone, and M. V. Prati. Combustion efficiencyand engine out emissions of a S.I. engine fueled with alcohol/gasoline blends. Appl.Energy, 111:1162–71, 2013.
[174] S. Pichon, G. Black, N. Chaumeix, M. Yahyaoui, J. M. Simmie, H. J. Curran, and et al.The combustion chemistry of a fuel tracer: Measured flame speeds and ignition delaysand a detailed chemical kinetic model for the oxidation of acetone. Combust. Flame,156:494–504, 2009.
[175] Z. Serinyel, N. Chaumeix, G. Black, J. M. Simmie, and H. J. Curran. Experimentaland chemical kinetic modeling study of 3-pentanone oxidation. J. Phys. Chem. A, 114:12176–86, 2010.
[176] Z. Serinyel, G. Black, H. J. Curran, and J. M. Simmie. A shock tube and chemicalkinetic modeling study of methy ethyl ketone oxidation. Combust. Sci. Technol., 182:574–587, 2010.
[177] J. M Hudzik and J. W. Bozzelli. Thermochemistry and bond dissociation energies ofketones. J. Phys. Chem. A, 116:5707–22, 2012.
[178] J. W. Allen, A. M. Scheer, C. W. Gao, S. S. Merchant, S. S. Vasu, O. Welz, and et al.A coordinated investigation of the combustion chemistry of diisopropyl ketone, aprototype for biofuels produced by endophytic fungi. Combust. Flame, 161:711–724,2014.
[179] M. Pelucchi, K. P. Somers, K. Yasunaga, U. Burke, A. Frassoldati, E. Ranzi, and et al.An experimental and kinetic modeling study of the pyrolysis and oxidation of n-C3C5aldehydes in shock tubes. Combust. Flame, 162:265–286, 2015.
[180] M. Kozak, J. Merkisz, P. Bielaczyc, and A. Szczotka. The influence of oxygenateddiesel fuels on a diesel vehicle pm/nox emission trade-off. In SAE Tech. Pap, volume2009-01-26, nov 2009. doi: 10.4271/2009-01-2696.
[181] M. Pacheco and C. L Marshall. Review of dimethyl carbonate (dmc) manufacture andits characteristics as a fuel additive. Energy Fuels, 11(1):2–29, 1997.
[182] X. Li, H. Chen, Z. Zhu, and Z. Huang. Study of combustion and emission characteris-tics of a diesel engine operated with dimethyl carbonate. Energy Convers. Manage., 47(11–12):1438 – 1448, 2006.
[183] P. A. Glaude, W. J. Pitz, and M. J. Thomson. Chemical kinetic modeling of dimethylcarbonate in an opposed-flow diffusion flame. Proc. Combust. Inst., 30(1):1111–1118,2005.
104
Bibliography
[184] H. Nakamura, H. J. Curran, A. P. Córdoba, W. J. Pitz, P. Dagaut, C. Togbé, and et al.An experimental and modeling study of diethyl carbonate oxidation. Combust. Flame,162:1395–1405, 2015. doi: 10.1016/j.combustflame.2014.11.002.
[185] M. D. Boot, B. Johansson, S. M. Sarathy, F. Halter, F. Foucher, and C. Mounaïm-Rousselle. Biofuels from Lignocellulosic Biomass: Innovations beyond Bioethanol. Wiley,2016. ISBN 978-3-527-33813-9.
[186] M. Nowakowska, O. Herbinet, A. Dufour, and P. A. Glaude. Detailed kinetic study ofanisole pyrolysis and oxidation to understand tar formation during biomass combustionand gasification. Combust. Flame, 161:1474–88, 2014. doi: 10.1016/j.combustflame.2013.11.024.
[187] M. A. Liberman. Introduction to physics and chemistry of combustion. Springer, 2008.
[188] A. Miyoshi. Systematic computational study on the unimolecular reactions of alkylper-oxy (ro2), hydroperoxyalkyl (qooh), and hydroperoxyalkylperoxy (O2QOOH) radicals. J.Phys. Chem. A, 115(15):3301–25, apr 2011. doi: 10.1021/jp112152n.
[189] S. Sharma, S. Raman, and W. H. Green. Intramolecular hydrogen migration inalkylperoxy and hydroperoxyalkylperoxy radicals: Accurate treatment of hindered rotors.J. Phys. Chem. A, 114(18):5689–5701, may 2010. doi: 10.1021/jp9098792.
[190] S. M. Villano, L. K. Huynh, H.H. Carstensen, and A. M. Dean. High-Pressure Rate Rulesfor Alkyl +O2 Reactions. 1. The Dissociation, Concerted Elimination, and IsomerizationChannels of the Alkyl Peroxy Radical. J. Phys. Chem. A, 115(46):13425–13442, nov 2011.doi: 10.1021/jp2079204.
[191] S. M. Villano, L. K. Huynh, H.H. Carstensen, and A. M. Dean. High-pressure rate rulesfor alkyl + O2 reactions. 2. the isomerization, cyclic ether formation, and β-scissionreactions of hydroperoxy alkyl radicals. J. Phys. Chem. A, 116(21):5068–5089, may 2012.doi: 10.1021/jp3023887.
[192] C. J. Hayes and D. R. Burgess. Jr. Kinetic barriers of H-atom transfer reactions in alkyl,allylic, and oxoallylic radicals as calculated by composite ab initio methods. J. Phys.Chem. A, 113:2473––82, 2009.
[193] L. B. Harding, S. J. Klippenstein, and Y. Georgievskii. On the combination reactions ofhydrogen atoms with resonance-stabilized hydrocarbon radicals. J. Phys. Chem. A, 111:3789–3801, 2007.
[194] C.S. McEnally, L. D. Pfefferle, B. Atakan, and K. Kohse-Höinghaus. Studies of aromatichydrocarbon formation mechanisms in flames: Progress towards closing the fuel gap.Prog. Energy Combust. Sci., 32:247—-294, 2006.
[195] M. Pecullan, K. Brezinsky, and I. Glassman. Pyrolysis and oxidation of anisole near1000 K. J. Phys. Chem. A, 101(96):3305–16, 1997. doi: 10.1021/jp963203b.
105
Bibliography
[196] G. W. Huber, S. Iborra, and A. Corma. Synthesis of transportation fuels from biomass:chemistry, catalysts, and engineering. Chemical reviews, 106(9):4044–98, 2006.
[197] A.J.J.E. Eerhart, M. K. Patel, and A. P.C. Faaij. Fuels and plastics from lignocellu-losic biomass via the furan pathway: an economic analysis. Biofuels, Bioproducts andBiorefining, 9(3):307–325, May 2015.
[198] E. de Jong and R. J. A. Gosselink. Bioenergy Research: Advances and Applications. Elsevier,2014. ISBN 9780444595614. doi: 10.1016/B978-0-444-59561-4.00017-6.
[199] E. de Jong, T. Vijlbrief, R. Hijkoop, G. J. M. Gruter, and J. C. van der Waal. Promisingresults with YXY Diesel components in an ESC test cycle using a PACCAR Dieselengine. Biomass Bioenergy, 36:151–9, 2012.
[200] A. Janssen, M. Muether, A. Kolbeck, M. Lamping, and S. Pischinger. The impactof different biofuel components in diesel blends on engine efficiency and emissionperformance. SAE Tech. Pap., 2010-01-2119, 2010.
[201] J. Yanowitz, M. A. Ratcliff, R. L. Mccormick, J. D. Taylor, and M. J. Murphy. Com-pendium of experimental cetane number data. Natl. Renew. Energy Lab., NREL/TP-54(August):1–48, 2014.
[202] E. Christensen, J. Yanowitz, M. Ratcliff, and R. L. McCormick. Renewable oxygenateblending effects on gasoline properties. Energy Fuels, 25:4723–33, 2011.
[203] M. Tian, R. Van Haaren, J. Reijnders, and M. Boot. Lignin derivatives as potentialoctane boosters. SAE Int. J. Fuels Lubr., 8(2):415–422, 2015. doi: 10.4271/2015-01-0963.
[204] K. Burgdorf and I. Denbratt. Comparison of Cylinder Pressure Based Knock DetectionMethods. SAE Tech. Pap., 972932, 1997.
[205] A. Hettinger and A. Kulzer. A New Method to Detect Knocking Zones. SAE Int. J.Engines, 2(1):645–665, 2009.
[206] G. E. Bogin, A. DeFilippo, J. Y. Chen, G. Chin, J. Luecke, M. A. Ratcliff, B. T. Zigler, andA. M. Dean. Numerical and experimental investigation of n-heptane autoignition in theignition quality tester (IQT). Energy Fuels, 25:5562–72, 2011. doi: 10.1021/ef201079g.
[207] G. E. Bogin, E. Osecky, M. A. Ratcliff, J. Luecke, X. He, B. T. Zigler, and A. M. Dean.Ignition quality tester (IQT) investigation of the negative temperature coefficient regionof alkane autoignition. Energy Fuels, 27:1632–42, 2013.
[208] G. E. Bogin, E. Osecky, J. Y. Chen, M. A. Ratcli, J. Luecke, B. T. Zigler, and A. M. Dean.Experiments and computational fluid dynamics modeling analysis of large n-alkaneignition kinetics in the ignition quality tester. Energy Fuels, 27(7):4781–94, 2014. doi:10.1021/ef500769j.
[209] G. E. Bogin, J. Luecke, M. A. Ratcliff, E. Osecky, and B. T. Zigler. Effects of iso-octane/ethanol blend ratios on the observance of negative temperature coefficientbehavior within the Ignition Quality Tester. Fuel, 186:82–90, 2016.
106
Bibliography
[210] S. Thion, A. M. Zaras, M. Szöri, and P. Dagaut. Theoretical kinetic study for methyllevulinate: oxidation by OH and CH <sub>3</sub> radicals and further unimoleculardecomposition pathways. Phys. Chem. Chem. Phys., 17(36):23384–23391, 2015. doi:10.1039/C5CP04194E.
[211] J. M. Simmie. Kinetics and thermochemistry of 2,5-dimethyltetrahydrofuran andrelated oxolanes: Next next-generation biofuels. J. Phys. Chem. A, 116:4528–38, 2012.doi: 10.1021/jp301870w.
[212] C. Barckholtz, T. a. Barckholtz, and C. M. Hadad. C-H and N-H bond dissociationenergies of small aromatic hydrocarbons. J. Am. Chem. Soc., 121(2):491–500, 1999.doi: 10.1021/ja982454q.
[213] K. Moshammer, S. Vranckx, H. K. Chakravarty, P. Parab, R. X. Fernandes, and K. Kohse-Höinghaus. An experimental and kinetic modeling study of 2-methyltetrahydrofuranflames. Combust. Flame, 160(12):2729–43, 2013. doi: 10.1016/j.combustflame.2013.07.006.
[214] Q. Bu, H. Lei, A. H. Zacher, L. Wang, S. Ren, J. Liang, and et al. A review of catalytichydrodeoxygenation of lignin-derived phenols from biomass pyrolysis. Bioresour.Technol., 124:470–7, 2012. doi: 10.1016/j.biortech.2012.08.089.
[215] S. C. Fadi and J. R. Arthur. Review of current and future softwood kraft lignin processchemistry. Ind. Crops Prod., 20:131–141, 2004. doi: 10.1016/j.indcrop.2004.04.016.
[216] A. U. Buranov and G. Mazza. Lignin in straw of herbaceous crops. Ind. Crops Prod., 28:237–259, 2008. doi: 10.1016/j.indcrop.2008.03.008.
[217] T. Prasomsri, T. Nimmanwudipong, and Y. Román-Leshkov. Effective hydrodeoxygena-tion of biomass-derived oxygenates into unsaturated hydrocarbons by MoO3 using lowH2 pressures. Energy Environ. Sci., 6(6):1732–8, 2013. doi: 10.1039/c3ee24360e.
[218] T. Prasomsri, M. Shetty, K. Murugappan, and Y. Roman-Leshkov. Insights into thecatalytic activity and surface modification of MoO3 during the hydrodeoxygenationof lignin-derived model compounds into aromatic hydrocarbons under low hydrogenpressures. Energy Environ. Sci., 7:2660–2669, 2014. doi: 10.1039/C4EE00890A.
[219] T. D. H. Nguyen, Marco Maschietti, Lars-Erik Åmand, Lennart Vamling, Lars Olausson,Sven-Ingvar Andersson, and Hans Theliander. The effect of temperature on the catalyticconversion of Kraft lignin using near-critical water. Bioresour. Technol., 170:196–203,2014. doi: 10.1016/j.biortech.2014.06.051.
[220] M. Saidi, F. Samimi, D. Karimipourfard, T. Nimmanwudipong, B. C. Gates, and M. R.Rahimpour. Upgrading of lignin-derived bio-oils by catalytic hydrodeoxygenation.Energy Environ. Sci., 7(1):103, 2014. doi: 10.1039/c3ee43081b.
[221] A. L. Jongerius, B. Jastrzebski, P. C A. Bruijnincx, and B. M. Weckhuysen. CoMo sulfide-catalyzed hydrodeoxygenation of lignin model compounds: An extended reactionnetwork for the conversion of monomeric and dimeric substrates. J. Catal., 285(1):315–323, 2012. doi: 10.1016/j.jcat.2011.10.006.
107
Bibliography
[222] B. P. Mudraboyina, S. Farag, A. Banerjee, J. Chaouki, and P. G. Jessop. Supercriticalfluid rectification of lignin pyrolysis oil methyl ether (LOME) and its use as a bio-derivedaprotic solvent. Green Chem., pages 13–16, 2016. doi: 10.1039/C5GC02233A.
[223] M. S. Talmadge, R. M. Baldwin, M. J. Biddy, R. L. McCormick, G. T. Beckham, G. A.Ferguson, and et al. A perspective on oxygenated species in the refinery integration ofpyrolysis oil. Green Chem., 16:407–453, 2014. doi: 10.1039/C3GC41951G.
[224] R. L. McCormick, M. A. Ratcliff, E. Christensen, L. Fouts, J. Luecke, G. M. Chupka,J. Yanowitz, M. Tian, and M. Boot. Properties of oxygenates found in upgraded biomasspyrolysis oil as components of spark and compression ignition engine fuels. EnergyFuels, 29:2453–2461, 2015. doi: 10.1021/ef502893g.
[225] K. Okuda, M. Umetsu, S. Takami, and T. Adschiri. Disassembly of lignin and chemicalrecovery—rapid depolymerization of lignin without char formation in water–phenolmixtures. Fuel Process. Technol., 85:803–813, 2004. doi: 10.1016/j.fuproc.2003.11.027.
[226] D. M. Alonso, S. G. Wettstein, and J. A. Dumesic. Bimetallic catalysts for upgrading ofbiomass to fuels and chemicals. Chem. Soc. Rev., 41:8075–8098, 2012. doi: 10.1039/C2CS35188A.
[227] D. A. Ruddy, J. A. Schaidle, F. R. Ferrell III, J. Wang, L. Moens, and J. E. Hensley.Recent advances in heterogeneous catalysts for bio-oil upgrading via "ex situ catalyticfast pyrolysis": catalyst development through the study of model compounds. GreenChem., 16:454–490, 2014. doi: 10.1039/C3GC41354C.
[228] N. Joshi and A. Lawal. Hydrodeoxygenation of 4-propylguaiacol (2-methoxy-4-propylphenol) in a microreactor: Performance and kinetic studies. Ind. Eng. Chem. Res.,52(11):4049–58, 2013. doi: 10.1021/ie400037y.
[229] X. F. Zhou. Conversion of kraft lignin under hydrothermal conditions. Bioresour.Technol., 170:583–6, 2014. doi: 10.1016/j.fuel.2013.07.047.
[230] E. Taarning, C. M. Osmundsen, X. Yang, B. Voss, S. I. Andersen, and C. H. Christensen.Zeolite-catalyzed biomass conversion to fuels and chemicals. Energy Environ. Sci., 4:793–804, 2011. doi: 10.1039/C004518G.
[231] C. Font Palma. Modelling of tar formation and evolution for biomass gasification: Areview. Appl. Energy, 111:129–141, 2013. doi: 10.1016/j.apenergy.2013.04.082.
[232] X. Wang and R. Rinaldi. Exploiting h-transfer reactions with raney[registered sign] nifor upgrade of phenolic and aromatic biorefinery feeds under unusual, low-severityconditions. Energy Environ. Sci., 5:8244–8260, 2012. doi: 10.1039/C2EE21855K.
[233] A. L. Jongerius, P. C. A. Bruijnincx, and B. M. Weckhuysen. Liquid-phase reformingand hydrodeoxygenation as a two-step route to aromatics from lignin. Green Chem., 15(11):3049–56, 2013. doi: 10.1039/c3gc41150h.
108
Bibliography
[234] A. Vigneault, D. K. Johnson, and E. Chornet. Base-catalyzed depolymerization of lignin:separation of monomers. Can. J. Chem. Eng., 85:906–916, 2007. doi: 10.1002/cjce.5450850612.
[235] R. N. Olcese, M. Bettahar, D. Petitjean, B. Malaman, F. Giovanella, and A. Dufour. Gas-phase hydrodeoxygenation of guaiacol over Fe/SiO2 catalyst. Appl. Catal. B Environ.,115-6:63–73, 2012. doi: 10.1016/j.apcatb.2011.12.005.
[236] C. Xu, R. A. D. Arancon, J. Labidi, and R. Luque. Lignin depolymerisation strategies:towards valuable chemicals and fuels. Chem. Soc. Rev., 43:7485–7500, 2014. doi:10.1039/C4CS00235K.
[237] H. Liu, Z. Wang, and J. Wang. Methanol-gasoline DFSI (dual-fuel spark ignition)combustion with dual-injection for engine knock suppression. Energy, 73:686–693,2014. doi: 10.1016/j.energy.2014.06.072.
[238] S. I. Yang, M. S. Wu, and C. Y. Wu. Application of biomass fast pyrolysis part I: Pyrolysischaracteristics and products. Energy, 66:162–171, 2014. doi: 10.1016/j.energy.2013.12.063.
[239] A. Brandt, L. Chen, B. E. van Dongen, T. Welton, and J. P. Hallett. Structural changes inlignins isolated using an acidic ionic liquid water mixture. Green Chem., 17:5019–5034,2015. doi: 10.1039/C5GC01314C.
[240] D. Fu, S. Farag, J. Chaouki, and P. G. Jessop. Extraction of phenols from ligninmicrowave-pyrolysis oil using a switchable hydrophilicity solvent. Bioresour. Technol.,154:101–8, 2014. doi: 10.1016/j.biortech.2013.11.091.
[241] O. D. Mante, J. A. Rodriguez, and S. P. Babu. Selective defunctionalization by TiO2 ofmonomeric phenolics from lignin pyrolysis into simple phenols. Bioresour. Technol.,148:508–516, 2013. doi: 10.1016/j.biortech.2013.09.003.
[242] X. Huang, T. I. Koranyi, M. D. Boot, and R. J. M. Hensen. Ethanol as capping agent andformaldehyde scavenger for efficient depolymerization of lignin to aromatics. GreenChem., 17:4941–4950, 2015. doi: 10.1039/C5GC01120E.
[243] C. Zhao and J. A. Lercher. Selective hydrodeoxygenation of lignin-derived phenolicmonomers and dimers to cycloalkanes on Pd/C and HZSM-5 catalysts. ChemCatChem,4:64–8, 2012. doi: 10.1002/cctc.201100273.
[244] Q. Song, F. Wang, J. Cai, Y. Wang, J. Zhang, W. Yu, and et al. Lignin depolymerization(ldp) in alcohol over nickel-based catalysts via a fragmentation-hydrogenolysis process.Energy Environ. Sci., 6:994–1007, 2013. doi: 10.1039/C2EE23741E.
[245] R. Lou and S.B. Wu. Products properties from fast pyrolysis of enzymatic/mild acidoly-sis lignin. Appl. Energy, 88(1):316–322, 2011. doi: 10.1016/j.apenergy.2010.06.028.
[246] T. Nimmanwudipong, R. C. Runnebaum, D. E. Block, and B. C. Gates. Catalyticreactions of guaiacol: Reaction network and evidence of oxygen removal in reactionswith hydrogen. Catal. Lett., 141(6):779–783, 2011. doi: 10.1007/s10562-011-0576-4.
109
Bibliography
[247] M. J. Murphy, J. D. Taylor, and R. L. McCormick. Compendium of experimental cetanenumber data. Technical report, National Renewable Energy Laboratory, 2004.
[248] Z. Boris. Viscosity blending equations. Lube magazine, (121):24–29, 2014.
[249] X. Zhen, Y. Wang, S. Xu, Y. Zhu, C. Tao, T. Xu, and et al. The engine knock analysis –An overview. Appl. Energy, 92:628–636, apr 2012. doi: 10.1016/j.apenergy.2011.11.079.
[250] G. T Kalghatgi. The available and required autoignition quality of gasoline-like fuels inhcci engines at high temperatures. SAE Tech. Pap., 2004-01-1969, 2004.
[251] M. A. Ratcliff, J. Burton, P. Sindler, E. Christensen, L. Fouts, G. M. Chupka, and R. L.McCormick. Knock resistance and fine particle emissions for several biomass-derivedoxygenates in a direct-injection spark-ignition engine. SAE Int. J. Fuels Lubr., 9(1):2016–01–0705, 2016.
[252] J. A. Badra, N. Bokhumseen, N. Mulla, S. M. Sarathy, A. Farooq, G. Kalghatgi, and et al.A methodology to relate octane numbers of binary and ternary n-heptane, iso-octaneand toluene mixtures with simulated ignition delay times. Fuel, 160:458 – 469, 2015.doi: http://dx.doi.org/10.1016/j.fuel.2015.08.007.
[253] D Bradley and C Morley. Chapter 7 Autoignition in spark-ignition engines. Compr.Chem. Kinet., 35(C):661–760, 1997. doi: 10.1016/S0069-8040(97)80022-2.
[254] O. Herbinet, B. Husson, M. Ferrari, P. A. Glaude, and F. Battin-Leclerc. Low tempera-ture oxidation of benzene and toluene in mixture with n-decane. Proc. Combust. Inst.,34(1):297–305, 2013. doi: 10.1016/j.proci.2012.06.005.
[255] B. Husson, M Ferrari, O. Herbinet, S. S. Ahmed, P. A. Glaude, and F. Battin-Leclerc.New experimental evidence and modeling study of the ethylbenzene oxidation. Proc.Combust. Inst., 34(1):325–333, 2013.
[256] V. Knop, M. Loos, C. Pera, and N. Jeuland. A linear-by-mole blending rule for octanenumbers of n-heptane/iso-octane/ toluene mixtures. Fuel, 115:666–673, 2014.
[257] A. Miyaji and T. Okuhara. Skeletal isomerization of n-heptane and hydroisomerizationof benzene over bifunctional heteropoly compounds. Catal. Today, 81(1):43 – 9, 2003.doi: 10.1016/S0920-5861(03)00100-7.
[258] A. J. Ragauskas, G. T. Beckham, M. J. Biddy, R. Chandra, F. Chen, M. F. Davis, andet al. Lignin valorization: Improving lignin processing in the biorefinery. Science, 344(1246843):695–700, 2014.
[259] R. C. Runnebaum, T. Nimmanwudipong, David E. Block, and B. C. Gates. Catalyticconversion of compounds representative of lignin-derived bio-oils: a reaction networkfor guaiacol, anisole, 4-methylanisole, and cyclohexanone conversion catalysed byPt/γ-Al2O3. Catal. Sci. Technol., 2:113, 2012. doi: 10.1039/c1cy00169h.
110
Bibliography
[260] K. Okuda, S. Ohara, M. Umetsu, S. Takami, and T. Adschiri. Disassembly of lignin andchemical recovery in supercritical water and p-cresol mixture. Studies on lignin modelcompounds. Bioresour. Technol., 99(6):1846–52, 2008.
[261] M. Kosa, H. X. Ben, H. Theliander, and A. J. Ragauskas. Pyrolysis oils from CO2precipitated Kraft lignin. Green Chem., 13(11):3196, 2011.
[262] M. Kleinert and T. Barth. Phenols from lignin. Chem. Eng. Technol., 31:736–745, 2008.doi: 10.1002/ceat.200800073.
[263] L. Zhou, M. D. Boot, B. H. Johansson, and J. J. E. Reijnders. Performance of ligninderived aromatic oxygenates in a heavy-duty diesel engine. Fuel, 115:469–478, jan 2014.
[264] M. K. Baltacioglu, K. Aydin, E. Yasar, H. Turan Arat, C. Conker, and A. Burgaç.Experimental investigation of performance and emission parameters changes on dieselengines using anisole additive. Appl. Mech. Mater., 490:987–991, 2014.
[265] F. Ancillotti and V. Fattore. Oxygenate fuels: Market expansion and catalytic aspect ofsynthesis. Fuel Process. Technol., 57(3):163–194, 1998.
[266] C. Muller, V. Michel, G. Scacchi, and G. M Come. THERGAS: a computer program forthe evaluation of thermochemical data of molecules and free radicals in the gas phase.J. Chem. Phys, 92:1154—-78, 1995.
[267] S. W Benson. Thermochemical Kinetics: methods for the estimation of thermochemical dataand rate parameters. John Wiley and Sons, London, 2nd edition, 1976.
[268] J. A. Montgomery, M. J. Frisch, J. W. Ochterski, and G. A. Petersson. A complete basisset model chemistry. VII. Use of the minimum population localization method. J.Chem. Phys., 112(15):6532–6542, 2000.
[269] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman,and et al. Gaussian09 Revision E.01. Gaussian Inc. Wallingford CT 2009.
[270] G. A. Petersson, D. K. Malick, W. G. Wilson, J. W. Ochterski, J. A. Montgomery, andM. J. Frisch. Calibration and comparison of the Gaussian-2, complete basis set, anddensity functional methods for computational thermochemistry. J. Chem. Phys., 109(24):10570–9, 1998.
[271] J. D. Cox, D. Wagman, and V. A. Medvedev. CODATA Key Values for Thermodynamics.Hemisphere, New York, 1989.
[272] R. D. JohnsonIII. Nist computational chemistry comparison and benchmark database,March 2015. URL http://cccbdb.nist.gov/. NIST Standard ReferenceDatabase Number 101.
[273] K. S. Pitzer and W. D. Gwinn. Energy levels and thermodynamic functions for moleculeswith internal rotation i. rigid frame with attached tops. J. Chem. Phys., 10(7):428, 1942.
[274] V. Mokrushin and W. Tsang. Chemrate v.1.5.2., 2006.
[275] V. N. Emelyanenko, K. V. Zaitseva, F. Agapito, J. A. Martinho Simões, and S. P. Verevkin.Benchmark thermodynamic properties of methylanisoles: Experimental and theoreticalstudy. The Journal of Chemical Thermodynamics, 85:155–162, jun 2015.
[276] G. Li, L. Li, L. Shi, L. J. Jin, Z. C. Tang, H. J. Fan, and et al. Experimental and theoreticalstudy on the pyrolysis mechanism of three coal-based model compounds. Energy Fuels,28:980–986, 2014.
[277] G. Da Silva and J. W. Bozzelli. Benzoxyl radical decomposition kinetics: formationof benzaldehyde + H, phenyl+CH2O, and benzene+HCO. J. Phys. Chem. A, 113(25):6979–86, 2009.
[278] D. L. Baulch, C. J. Cobos, R. A. Cox, P. Frank, G. Hayman, T. Just, and et al. Evaluatedkinetic data for combustion modeling. supplement i. J. Phys. Chem. Ref. Data, 23(6):847, 1994.
[279] P. Frank, J. Herzler, Th. Just, and C. Wahl. High-temperature reactions of phenyloxidation. Symp. Combust., 25(1):833–840, jan 1994.
[280] G. E. Bogin, E. Osecky, M. A. Ratcliff, J Luecke, X. He, B. T. Zigler, and et al. Ignitionquality tester (IQT) investigation of the negative temperature coefficient region ofalkane autoignition. Energy Fuels, 27:1632–42, 2013.
[281] R. J. Kee, F. M. Rupley, and J. A. Miller. Chemkin-II: A Fortran chemical kinetics packagefor the analysis of gas-phase chemical kinetics. 1989.
[282] J. L. Emdee, K Brezinsky, and I Glassman. A kinetic model for the oxidation of toluenenear 1200 K. J. Phys. Chem., 96:2151–61, mar 1992.
[283] C. S. Draper. Pressure waves accompanying detonation in the internal combustionengine. J. Aeronaut. Sci., 5:219–226, 1938.
[284] B. Samimy and G. Rizzoni. Mechanical signature analysis using time-frequency signalprocessing: application to internal combustion engine knock detection. Proc. IEEE, 84:1330–43, 1996.
[285] Y. L. Qi, X. He, Z. Wang, and J. X. Wang. Frequency domain analysis of knock images.Meas. Sci. Technol., 25(12):125001, 2014. doi: 10.1088/0957-0233/25/12/125001.
[286] R. Worret, S. Bernhardt, F. Schwarz, and U. Spicher. Application of different cylinderpressure based knock detection methods in spark ignition engines reprinted from : SIengine experiments. SAE Tech. Pap., 2002-01-1668, 2002. doi: 10.4271/2002-01-1668.
[287] M. D. Checkel and J.D. Dale. Testing a third derivative knock indicator on a productionengine. SAE Tech. Pap., 861216, 1986.
[288] J. M. Borg and A. C. Alkidas. Characterization of autoignition in a knocking si engineusing heat release analysis. SAE Tech. Pap., 2006-01-33, 2006.
112
Bibliography
[289] J. D. Naber, J. R. Blough, D. Frankowski, M. Goble, and J. E. Szpytman. Analysis ofcombustion knock metrics in spark-ignition engines. SAE Tech. Pap., 2006-01-0400.
[290] A. J. Shahlari and J. B. Ghandhi. A comparison of engine knock metrics. SAE Tech.Pap., (2012-32-0007), 2012. doi: 10.4271/2012-32-0007.
113
Appendix APressure signal used in knock anal-ysis
There are three main ways to test knock, pressure signal, accelerometer, or human ear. In-cylinder pressure signal is regarded as the most reliable method, because it directly capturethe pressure signal from the cylinder, while accelerometer may detect various sources ofvibration in addition to the vibration generated by engine knock. And human ear is verysubjective, it varies from person to person, and long time exposure to high frequency noise isharmful to human health.
A.1 Resonance frequency calculation
The frequencies of the resonances excited by auto-ignition of the end-gas depend on thecylinder geometry and the speed of sound in the cylinder. For an ideal cylinder that has anacoustically hard wall and is filled with homogeneous gas, according to Draper’s equation[283], the resonance frequencies can be calculated by Equation A.1 [284]:
fm,n =cηm,n
Bπ(A.1)
To calculate fm,n, assuming temperature is 2500 K, specific heat is 1.35, the speed of soundin ideal gas then will be is 950 m/s [285], and the calculated frequencies are listed in TableA.1. The calculated frequencies are in good agreement with the experiment ones (FigureA.1). However, not all the modes can be seen in one cycle„ e.g., the ’knock-32’ line has mostresonant mode at 6.7, 8.3, 22.3 and 25.2 kHz, while the ’knock-54’ shows the most resonantmode at 6.8, 7.3, 23.2 and 26.1 kHz, and at 6 kHz, even non knock cycle shows a small peakat the same level. This also shows the random nature of the engine knock. Based on thecalculation and the experiments, 6-25 kHz are chosen as the pass frequency of the band passfilter. The filter employed is a digital, finite impulse response (FIR) band pass filter with
115
Appendix A. Pressure signal used in knock analysis
Table A.1: Acoustic modes of a cylindrical combustion chamber
Kaiser window, and the order is 91, the stop frequencies are 2 and 30 kHz.
A.2 Overview of pressure based engine knock detectionmethod
There are many ways to quantify engine knock based on in-cylinder pressure profile. Burgdorf[204] reviewed and compared them, which are listed below. However, there is no specific proofthat one method is better than others, and each of them has been used by some researchers.
Cylinder pressure
At same operation condition, the maximum pressure is higher in the knock cycle than in thenon-knock cycle, so that it can be used to rate knock intensity (KI). However, at different speedor/and load the pressure changes. Moreover, because of the cycle-to-cycle variation, somecycles have high pressure without knock, while some others have knock but with relatively
116
A.2. Overview of pressure based engine knock detection method
-20 -10 0 10 20 30 40 50 60
Crank angle [degree]
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
1st p
ress
ure
deriv
ativ
e [P
a]
×106
knocknonknock
Figure A.2: 1st derivative pressure signal of knock and unknock cycle (Data fromBlend-1-171-14and 5)
low pressure, making this parameter only can indicate knock qualitatively.
Pressure rise (1st derivative)
Show in Figure A.2, the maximum value of the derivative pressure can be used to indicateknock intensity, and it can also be used to find the knock onset, which is defined as crankangle at the maximum value, but this is not easy to detect when weak knock happens [205].
Third time derivative of the pressure signal
The third derivative pressure signals for one knock cycle and one non knock cycle are show inFigure A.3. The knock onset is defined at the crank angle when the signal surpass a certainthreshold [286] or first sharp increase. And KI is defined as the signal energy. It is mainlyused to get reliable knock information when the data acquisition frequency is low [287].
Band pass filtered pressure
The bandpass filtered pressure with frequency of 6-25 kHz is shown in Figure A.4. At thisload and spark timing, the peak oscillation is around 20°CA (including the phase delay by thefilter). There are several peaks along the whole cycle, which are generated by combustion orby the spark plug or other noise sources. By looking closely to only around top dead center
117
Appendix A. Pressure signal used in knock analysis
-20 -10 0 10 20 30 40 50 60
Crank angle [degree]
-2
-1.5
-1
-0.5
0
0.5
1
1.5
21s
t pre
ssur
e de
rivat
ive
[Pa]
×108
knocknonknock
Figure A.3: 3rd derivative pressure signal of knock and unknock cycle
(lower figure in Figure A.4), it can be seen that the pressure fluctuation difference is quitelarge between knock and non-knock cycles.
Heat release profile
When knock occurs, the high pressure increase will generate an abnormal heat release, andanother peak is expected as show in Figure A.5. Knock onset can be defined as the secondheat release rate peak [288].
Other parameters
Hettinger et al. [205] developed a new method by modeling a spherical flame front and usingpressure rise method to determine the knock location.
On frequency domain, the energy density also can be calculated, and the resonancefrequencies can be found.
A.3 Knock intensity and knock onset definition in thisthesis
According to the summary above, the band pass filtered pressure is chosen to analyze becauseit mainly contain the frequencies that are related to knock. There are two main ways to
118
A.3. Knock intensity and knock onset definition in this thesis
-400 -300 -200 -100 0 100 200 300 400
Crank Angle [degree]
-2
-1
0
1
2
Filt
ered
pre
ssur
e [P
a] ×105
knocknonknock
-10 0 10 20 30 40 50
Crank Angle [degree]
-2
-1
0
1
2
Filt
ered
pre
ssur
e [P
a] ×105
knocknonknock
Figure A.4: Bandpass filtered pressure
calculate knock intensity (KI) and knock onset (KO) by analyzing band pass filtered pressure,which will be listed below.
The maximum amplitude pressure oscillation (MAPO) is the peak to peak amplitude ofpressure oscillation (Figure A.6), which is used quite often to indicate KI with different bandpass frequency [23, 288, 289]. The bandpass filter used in the papers mentioned above have afrequency of 5-10 kHz, 5-32 kHz and 5-20 kHz, respectively.
The signal energy of pressure oscillation (SEPO) is the energy of the band pass filteredpressure. It is calculated in Equation A.2. The θ0 could be the knock onset or a fixed knockcrank angle window at stable load and speed, when the spark timing does not change toomuch. In our study, two SEPO are calculated, one with fixed knock window (SEPO), i.e.,10-55◦bTDC; one is the 20◦CA window after the KO (SEPOKO).
SEPO =
∫θ0+∆θ
θ0
P2filtdθ (A.2)
The KO is defined by using the so called signal energy ratio (SER), which is based on theSEPO (Equation A.3) developed by Shahlari et al. in [290]. By calculating the ratio of forwardand backward SEPO, this method overcome the limitations of the threshold value exceededmethod, which according to the author, delayed the KO because the threshold need to be sethigh enough to avoid false detection.
SER =SEPOfwd
2
SEPObwd1/2
=
∫θ0+∆θ
θ0P2filtdθ
2
∫θ0θ0−∆θ P
2filtdθ
1/2(A.3)
119
Appendix A. Pressure signal used in knock analysis
-20 -10 0 10 20 30 40
Crank angle [degree]
0
20
40
60
80
100
120
Hea
t rel
ease
rat
e [J
/s]
knockknocknonknock
Figure A.5: Heat Release Rate of knock and non-knock cycle (Filtered with Savitzky-Golay filter)
In the equation, ∆θ was 4◦CA for this study, and the KO was defined as the crank angle atwhich the SER is maximum. The KO did not agree well with the pressure profile, becauseafter the band pass filter, the signal already has some phase shift, but the calculation is fromthe band pass filtered pressure, so this will not impact the results.
The comparison among MAPO, SEPO and SEPOKO is shown in Figure A.8, it can beseen that these three parameters have same trend, and SEPO and SEPOKO are almost thesame. When looking more closely, the differences between SEPO and MAPO still can be seenfrom Figure A.9, most of the time, the MAPO shows higher value than the SEPO, especiallywhen knock intensity increases. This is because there may be a sudden high fluctuation,which should also be noted because it may be the highest which cause the engine damage.Generally, all of them are good parameters to indicate KI, and have linear correlation betweeneach other. And in this thesis MAPO and SEPO are used as KI.
120
A.3. Knock intensity and knock onset definition in this thesis
0 10 20 30 40 50 60
Crank angle [degree bTDC]
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1F
ilter
ed p
ress
ure
[Pa]
×105
MAPO
SEPO window
Filtered pressure
Figure A.6: MAPO and SEPO definition
0 10 20 30 40 50 60
Crank Angle
5
Pre
ssur
e [P
a]
×106
-5
0
5
Filt
ered
pre
ssur
e [P
a]
×104
PressureKnock onsetFiltered pressure
Figure A.7: Pressure profile with knock onset
121
Appendix A. Pressure signal used in knock analysis
0 5 10 15 20 25 30 35 40 45 50
Cycles
0
0.5
1
1.5
2
2.5
3K
nock
inte
nsity
×105
MAPOSEPO/10SEPO
ko/10
Figure A.8: Comparison among MAPO, SEPO and SEPOKO for 100 cycles
0 2 4 6 8 10 12 14 16
SEPO [Pa2/s] ×105
0
0.5
1
1.5
2
2.5
3
MA
PO
[Pa]
×105
Figure A.9: SEPO vs. MAPO for 100 cycles
122
Appendix BFuel injection amount in IQT
Gravimetric fuel consumption is calculated by using the already measured mole count of air inthe combustion vessel at a given pressure and temperature. First, this data is used to calculatehow much fuel is required for the stoichiometric case. Subsequently, the shim thickness iscalculated by using the linear calibration of the mass of 10 times the injected quantity of fueland the shim thickness one should prepare in the setup, which is measured experimentallyusing hexadecane, decane and n-heptane through different thickness of shim, each representsa different viscosity. Since viscosity is the most important parameter for injection, at roomtemperature and atmospheric conditions, the viscosity of hexadecane is 3× 10−3Pa · s, whichis half that of guaiacol’s, and it is the most closest one among the tested fuels to guaiacol, sohexadecane’s data is used for calculating the shim thickness of guaiacol and alkylated guaiacol.Anisole’s viscosity is 1 × 10−3Pa · s, similar to decane’s (0.9291 × 10−3Pa · s), so decane’sdata is used to calculate toluene, anisole and alkylated anisoles’ shim thickness. Figure B.1shows the calibration results of the relationship between shim thickness and mass of injectedfuels. And it can be seen from the figure that as shim become thicker, the differences betweendifferent fuels become smaller, and for our fuels, the shim is always > 0.09in, which meansthe impact of viscosity is smaller.
Data are from MSDS; c Data are from McCormick et. al [224], except the LHV oftoluene, which is from NIST web and [247], and EL’s viscosity was measured at 20℃.d Measured this time, except toluene whose data is from NIST web and [247]. e Datais from NIST web and [247].
126
Table C.2: Properties of the blended fuels
Fuel Density O content LHV LHV RON MONg/L [wt.-%] MJ/kg MJ/L
4-EG 4-ethyl guaiacol2-PE 2-phenyl ethanol4-MA 4-methyl anisole4-MG 4-methyl guaiacol4-PA 4-propyl anisole2-MF 2-methyl furan2,5-DMF 2,5-dimethyl furanAKI Anti knock indexASTM American society for testing and materialsaTDC After top dead centerBA Benzyl alcoholBDE Bond dissociation energybTDC Before top dead centerCFR Cooperative fuels researchCN Cetane numberDCN Derived cetane numberDEC Diethyl carbonateDMC Dimethyl carbonateEP Ethyl propionateEL Ethyl levulinateEP Ethyl PropionateETE Ethyl tetrahydrofurfuryl etherFAME Fatty acid methyl esterFEE Furfuryl ethyl etherFBRU Full-boilling unleaded reference fuelsGDI Gasoline direct injectionHCCI Homogeneous charge compression ignition
129
Appendix D. Abbreviations
HECC High efficiency clean combustionID Ignition delay timeIQT Ignite quality testerIMEP Indicated mean effective pressureKI Knock intensityKLSA Knock limited spark advanceLHV Lower heating valueLTHR Low temperature heat releaseMAPO Maximum amplitude of pressure oscillationMB Methyl butanoateML Methyl levulinateMON Motor octane numberMTHF Methyl tetrahydrofuranNTC Negative temperature coefficientOI Octane indexON Octane numberPCCI Premixed charge compression ignitionPFI Port fuel injectionPRF Primary reference fuelsRON Research octane numberSEPO Signal energy of pressure oscillationRCM Rapid compression machineSI Spark-ignitionTS Transition stateWOT Wide open throttle
130
Acknowledgement
Four years have passed by in a blink of an eye. The day I arrived in Eindhoven remains in mymind as a vivid memory. While writing the last part for my thesis, I am looking back realizingall that I have learnt over the course of this journey. All these efforts have become worthwhileat this point. This would not have been possible without the direct and indirect support frommany individuals.
My deepest gratitude goes out first, to my supervisors, Michael, Philip and Niels. DearMichael, it is so nice to work with you. Thank you for your constant encouragement, support,and valuable advice. I am grateful that you have offered me the chance to pursue sabbaticals atNREL and Nancy, which proved valuable learning experiences. Dear Philip, thank you for allyour support. I appreciate all the progress meetings and discussions we had, all the questionsyou have asked me, which made me think deeper about both my research and myself. DearNiels, although you became my promoter only half a year ago when I was planning to submitmy form 1, you have given me a great deal of help during this short period, always pressingme to stick to the timetable, and helping me translate the summary into Dutch. I enjoyedworking with you and have learn many precious skills that I will carry with me in future.
I would also like to extend my gratitude to my dissertation committee: Pierre AlexandreGlaude from Laboratoire Réactions et Génie des Procédés (LRGP), Roger Cracknell from Shell,and from Eindhoven University of Technology, Emiel Hensen and Frank Willems. Thank youfor all your consideration of and feedback on my thesis and for participating in my defense.
Furthermore, I would also like to thank all those who have helped me in the engine lab.Their help was often needed given that I always ran the engine to the knock limit, which isdefinitely not good for any engine, as Bart has warned me on many occations. My gratitudealso goes out to Jos, Theo, Hans and Jaap for all the help with fixing and maintaining mysetup. On the experimental side, I would like to acknowledge Robin for his help.
Bob, thank you for hosting me at NREL, You being an expert on biofuel studies, workingin your group was great learning opportunity to me. I have learnt a lot from all our discussions,be it in person or via emails. Jon, thank you too for all your help in the IQT experiments. Matt,you are such a nice person, I enjoyed all the discussions we had, as well as your encouragementin many emails.
Pierre, you made my stay at Nancy fruitful. I appreciate your patience and time to answermy endless questions, and your cheerful attitude helped me overcome the frustration when Iwas troubleshooting the mechanism. Thank you very much.
To all my (former) colleagues, with whom I had so many joyful times with, thank you for
131
Appendix D. Acknowledgement
all the enjoyable time and daily support. From my office, Akshay, Mayuri, Yuchun, Lei, Naud,Nico, Nard, Robbert, Atieh, Bersan, Fesion, Naseh. From other offices, Aromal, Shuli, Zhen,Denis, Mohammond and so on. Special thanks to Aromal,for all our discussions on anythingfrom religions to photography, you have inspired me in many different ways to learn moreabout those things I am unfamiliar with. Also thanks to all my friends I met during thesepast four years living in Eindhoven. Everyone is awesome, like the song goes. I appreciate allthe talks, discussions and activities with you, and all the comfort and sweetness you gave me.
Last but not least, I want to thank my parents for their endless support and care. Wordscannot express my gratitude for everything you have done for me. I also want to thank myboyfriend, your support makes me stronger and I am looking forward to our next adventuretogether.
132
List of publications
• Tian M., Van Haaren R., Reijnders J., and Boot M., Lignin Derivatives as PotentialOctane Boosters, SAE Int. J. Fuels Lubr. 8(2):2015, doi:10.4271/2015-01-0963.
• McCormick R. L., Ratcliff M. A., Christensen E., Fouts L., Luecke J., Chupka G. M.,Yanowitz J., Tian M. Boot M., Properties of Oxygenates Found in Upgraded BiomassPyrolysis Oil as Components of Spark and Compression Ignition Engine Fuels. EnergyFuels 2015;29:2453–61. doi:10.1021/ef502893g.
• Tian M., McCormick R. L., Ratcliff M. A., Luecke J., Yanowitz J., Glaude PA, CuijpersM, Boot M., Performance of Lignin Derived Compounds as Octane Boosters, submittedto Fuel.
• Boot M., Tian M., Sarathy S. M., Impact of Fuel Molecular Structure on Auto-IgnitionBehavior - Design Rules for Future High Performance Gasolines, submitted to Prog.Energy Combust. Sci.
• Tian M., McCormick R. L., Luecke J., de Jong E., van der Waal J. C. van Klink G., BootM., Anti-knock quality of sugar derived levulinic esters and cyclic ethers, submitted toFuel.
133
Curriculum Vitae
Miao Tian was born on 04-05-1987 in Baoji, China.After finishing high school in 2005 at Xi’an No.85 high school in Xi’an, China, she studied
Environmental Engineering at Southwest Jiaotong University in Chengdu, China. Then shecontinued her master’s study in the same department, her final project was on the evaluationof high-speed railway induced environmental vibration. From September, 2012 she started aPhD project at Eindhoven University of Technology in Eindhoven, the Netherlands, of whichthe results are presented in this dissertation.
