dryden lecture next gen rm

8/6/2019 Dryden Lecture Next Gen RM

1/30

Dryden Lecture:

Fuel Options for Next Generation Chemical Propulsion

Chung K. Law1Princeton University, Princeton, New Jersey,08544, USA

The state of research on developing fuel options for next generation chemical propulsion

is reviewed for aviation fuels and energetic fuels. For aviation fuels, the development is

based on considerations of cost, energy security and climate change, with Fischer-Tropsch

synthetic fuels and biofuels hold potential as alternative aviation fuels. The need for basic

research to develop predictive capability for the oxidative chemistry of evolving fuels in

evolving engine designs is emphasized, illustrated by the intricate reaction pathways and the

enormity of the reaction mechanisms involved. Recent research activities towards achieving

the goal of fuel design are discussed through the development of detailed mechanisms,

reduced mechanisms, and surrogate fuels. For the development of high-energy-density

propellants, advances in several classes of materials are discussed, including metallized andhypergolic propellants, and propellants with strained and functionalized molecules as well as

nanoparticle addition. The impact of the recent progress in chemical synthesis, materials

science and nano science on these advances is noted.

I. IntroductionHE invention of chemical propulsion for the transportation of goods and people, together with, for example,electrification, mechanized farming, immunization, and computers, must rank among the most significant

developments in civilization that ushered in the modern, globalized, society. Indeed, out of the top 20 of the greatestengineering achievements of the 20th century 1, automobile and airplane, whose operation fundamentally rely onchemical propulsion, are respectively ranked in the second and third places of significance.

T

The massive deployment of chemical propulsion, through both land and air-borne transportation, has however

severely degraded the air quality in many parts of the world. Furthermore, the continuing and in fact escalatingreliance on its service is facing unprecedented threats in terms of energy sustainability and climate change. As such,there is the urgency to develop rational strategies on fuel options to accommodate these challenges.

In addition to transportation, chemical propulsion is also central to the operation of rockets, missiles, and otherhigh-speed air-borne devices, for which the primary interest is the energy density and burning rate of the fuelrequired to generate the intense power needed. Recent development of these energetic propellants has been muchstimulated by advances in chemical synthesis, materials science, and nano science, which offer potential forbreakthrough concepts in their formulation.

In the following we shall present a brief review on the fuel options for next generation chemical propulsion andthe research activities currently being conducted to facilitate such options. In view of the interest of the intendedaudience, the review will focus on chemical propulsion related to aeronautics: specifically air-breathing and rocket propulsion. Issues related to the vast area of land transportation are discussed only where necessary, as theirimportance and complexity merit a separate review. Furthermore, because of the substantially different nature of the

requirements of aviation and energetic fuels, discussion on these two topics will be presented separately.

American Institute of Aeronautics and Astronautics1

In the next section we shall discuss the options for aviation fuels by first stating their role in energy security andclimate change. We shall then review the viable alternative aviation fuels and the state of their development.Furthermore, recognizing the evolving nature of alternative fuels and the need to be predictive of their combustionresponses particularly those affected by fuel chemistry, in Section III we shall first cite examples of the intricacies ofthe reaction pathways in fuels oxidation, and then discuss various approaches towards developing comprehensivereaction mechanisms adaptable to computational simulations. In Section IV we present recent activities on thedevelopment of energetic propellants, which include metallized slurries, strained and functionalized fuel molecules,

1 Robert H. Goddard Professor, Department of Mechanical and Aerospace Engineering. Fellow AIAA.

49th AIAA Aerospace Sciences Meeting including the New Horizons Forum and Aerospace Exposition4 - 7 January 2011, Orlando, Florida

AIAA 2011-1

Copyright 2011 by the author(s). Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.


2/30

American Institute of Aeronautics and Astronautics

2

hypergolic and ionic propellants, and nano-composites and nanoparticles, with emphasis on the fundamentalmechanisms meriting their development.

II. Development of Aviation FuelsPetroleum has always been the aviation fuel of choice because of its ease of handling, cost, operational

reliability, high energy content and fast burning rate. For example, the complete oxidation of CH2, the basic unit of a

hydrocarbon (HC) fuel, releases 156 kcal/mole of heat, which is among the most exothermic unit in chemistry.Furthermore, the energy density of petroleum, on both mass and volume bases, is also high, around 42.8MJ/kg and35.1 MJ/Liter for Jet A. Thus combined with its fast reaction rate with oxygen, petroleum provides the neededintense power for propulsion. However, the recent escalation in the price of petroleum, the insecure nature of itssupply due to the rapid depletion of its reserve and geopolitical conflicts, and the equally serious concern of itsemissions of pollutants and of greenhouse gas (GHG) which impacts climate change, have prompted concertedR&D efforts to develop alternate sources of aviation fuels. To assess the severity of these concerns, it is instructiveto first evaluate the impact of aviation propulsion on the energy needs and on climate change through GHGemissions.

Out of the various energy sources, energy derived from burning fossil fuels constitutes about 84% of the totalenergy expenditure of the country 2, and as such dominates the energy landscape (Fig. 1). Among the three sourcesof fossil fuels, namely gas, petroleum and coal, petroleum constitutes almost half of the total supply and thereforehas a disproportionately large impact on fossil fuel consumption. Furthermore, transportation in the US is almost

completely dependent on petroleum, accounting for nearly 60% of the nations use of petroleum, with 56% beingimported 2. The share of the petroleum use by air transportation is 12% 3.In the US, transportation GHG emissions, of which 95% is CO 2, accounts for 29% of the total US emissions

(Fig. 2) 4. Out of this amount, 79% is from on-road vehicles and 12% from commercial aircrafts (Fig. 3).Consequently, GHG emissions from commercial aircrafts constitute 3.5% of the total US emissions. These values,while fractionally small, must still be treated seriously because of the projected increase in commercial aviation andthe enormity of the problem.

While there exist considerable uncertainty and latitude in identifying a moderately hard timeline for the depletionof petroleum reserve and the onset of severe global warming, the following projections and facts largely hold. Thatis, the global energy demand will double by 2050, especially recognizing the heightened needs by several rapidlydeveloping countries; that we are using more petroleum than we discover such that the global petroleum reserve ison the path of net depletion; and that our dependence on oil from geopolitically unstable regions could stronglyaffect our economic and political well being. Furthermore, uncertain as it may be, the Intergovernmental Panel onClimate Change (IPCC) 4, 5 estimates that in the absence of additional corrective action to reduce anthropogenic

GHG emissions, the global temperature will rise between 1 to 6.4 C by 2100.In view of all these negative indicators, especially the possibility that the trend could accelerate as the rapidly

developing countries achieve further prosperity, it is prudent if not compelling to actively develop options foralternative fuels, including alternative aviation fuels, so that a successful transition to the post-petroleum era can beachieved. This is a challenging goal as it typically requires two to three decades to nurture a major technology tomature commercialization. Take the development of nuclear energy as an example. The basic concept of nuclearfission was advanced in the 1930s, with the proof of concept attained when criticality was achieved in 1942, and thefirst commercial nuclear power station became operational in England in 1956. This development was unusuallyfast a technological triumph that was greatly facilitated by the urgency to develop the atomic bomb in World WarII. As a more cautionary example, the possibility of controlled nuclear fusion was advanced in the late 1940s,although the process is still in the stage of concept demonstration.

There are, however, factors that are in favor of encouraging and accomplishing a smooth transition. First, asconcluded in Ref. 2, substantial improvements in energy supply and consumption can be accomplished through

moderate modifications of the technologies available to-day, for example the potential gains in building efficiency.Moderate modification of our lifestyle can also lead to meaningful energy savings. This could stretch the lead timefor transition.

Second, there is substantial reserve of coal, tar sand and oil shale to last several centuries. This is especially pertinent for the U.S. because of her large coal reserve. Thus there is feedstock to synthesize alternative fuels,provided auxiliary technologies such as carbon capture and storage (CCS) are developed.

Third, since the contribution of GHG from air transportation is relatively small compared to, say, landtransportation, the primary drivers in the development and use of alternative fuels for air transportation would be thesecurity and cost of supply, noting for example that the purchase of fuel currently represents the largest operating


3/30

cost for U.S. airlines, and that every one-cent increase in fuel price translates into an additional $190M in annualcosts for the commercial aviation industry 3. Consequently, instead of having to confront the dual challenges of fuelavailability and GHG emissions as faced by land transportation, potential contribution by aviation transport to themoderation of climate change can therefore be considered as a supplemental benefit instead of primary requirement.This allows more flexibility in the development of alternative aviation fuels.

Fourth, while alternate forms of energy are viable options for other sectors of the energy users, such as batteriesfor automobiles and wind energy for electricity generation in general, chemical energy derived from burning liquidfuels remains the only viable source of energy for aviation transport. This should stimulate the development ofalternative aviation fuels, especially with the continuously increasing price of petroleum. Finally, some technologybases are already in place for the synthesis of alternative transportation fuels using coal and biomass as feedstocks,as will be further discussed, and as such the creation of transformational technologies, which is of uncertainoutcome, is not critical.

Having discussed the drivers for the development of alternative aviation fuels, we next review the options forthese fuels. The various non-petroleum-based, alternative aviation fuels that have been considered can be groupedinto three categories 6-8, namely: (a) synthetic fuels derived from coal, natural gas, and other forms of fossil fuelsthrough the Fischer-Tropsch (FT) process; (b) biodiesel derived from biomass sources; and (c) other alternative fuelssuch as liquefied hydrogen, methane, and petroleum gas, and the light alcohols of methanol, ethanol and butanol;noting that the alcohols can also be bio-derived.

The potential of these fuels as alternative aviation fuels can best be assessed on the basis of the single mostimportant property required, namely the energy content per unit mass and per unit volume. Figure 4 compares the

gravimetric and volumetric energy densities of several of these alternative fuels to those of Jet A. It is seen thatwhile liquid hydrogen has the highest gravimetric energy density, it has the lowest volumetric energy density.Liquid methane is similarly handicapped, especially in terms of its volumetric density. Furthermore, methanol andethanol are inferior to Jet A on both counts because the extra oxygen atom in their molecular structure not only doesnot play the role of a fuel but actually adds weight to it. Indeed, this concern is present for practically all biofuelsincluding biodiesel which have oxygen atoms in their molecular structures. This energy-dilution effect, however,gradually diminishes for larger molecules because of their progressively smaller influence in a fuel consisting oflarger numbers of carbon and hydrogen atoms. For example, while the molar energy content of methanol is only81.5% of its alkane analogue, methane, the difference narrows to 87.6% for ethanol and 93.0% for the C4 moleculeof n-butanol.

Figure 4 then shows that in terms of the specific energy content only the Fischer-Tropsch (FT) synthetic fuel andthe biodiesel can be considered as potential replacements for Jet A. We shall therefore focus our discussion on thesetwo categories of fuels, with a cursory coverage of the other fuels with lower energy densities.

Figure 1. Americas energy use in 2008. Source: Energy Information Administration, 2008. Adapted from

Ref. 2.



4/30

Figure 2. U.S. greenhouse gas emissions by end use economic sector in 2006, in million metric tons of CO2

equivalent. Source: U.S. EPA, 2008. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-20064.

Figure 3. U.S. greenhouse gas emissions by transportation mode, in 2006. Source: U.S. EPA, 2008;

Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-20064.



5/30

Figure 4. Gravimetric and volumetric energy densities of various alternative aviation fuels. Adapted from

Ref. 6.

A. Fischer-Tropsch FuelsThe Fischer-Tropsch process synthesizes a mixture of hydrogen and carbon dioxide into higher molecular weight

hydrocarbons. The initial products are mainly straight-chain hydrocarbons, which are further cracked into smallerones, and then rearranged to yield the desired composition in terms of the volatility range and chemical properties.Since the FT process starts with carbon monoxide, the initial feedstock of carbon can be coal, natural gas orbiomass. Thus depending on the nature of the feed stock, the life cycle CO2 emission can vary significantly.

FT fuels are virtually free from trace sulfur- and nitrogen-containing compounds as well as aromatics that are

usually present in conventional aviation fuels. Consequently they are cleaner burning in terms of emissions of sootas well as oxides of sulfur and nitrogen. The lack of aromatics, however, reduces the fuel density and could alsoinhibit swelling of the seals in the engine fueling system and hence can cause fuel leakage. A compromise solutionis to blend FT fuels with conventional jet fuels.

At present, Sasol in South Africa produces FT fuels using coal, while several major companies also have plans tobuild large plants for production.

B. BiodieselBiodiesel are fatty acid esters that can be produced via transesterification with methanol or ethanol from

vegetable oil such as canola, cottonseed, peanut oil, rapeseed oil and waste oil, and animal fat such as beef tallowand pork lard. It is readily biodegradable and is non-toxic. Since the feedstock for plant-derived biodiesel is notfossil fuel based, its use holds potential in the reduction of lifecycle CO2 emissions. Similar to the sugar- and starch-derived ethanol, biodiesel demand for some of these feedstocks also competes with food use. On the other hand,

vegetations such as jatropha, camelina and algae are non-food biodiesel feedstocks, and their viability is currentlyunder much development.Biodiesel molecules consist of more oxygen atoms than the alcohols, typically two to three, which further reduce

the heat content, as mentioned earlier. Thus even the reduction is not substantial because the basic hydrocarbonfunctional group is large, it is still a matter of concern for aviation use. A second concern is the high freezing pointof the oil, typically around 0 oC as compared to about -40 oC for jet fuels. Both these concerns require further(hydro-) processing of the oil, as in the recent synthesis of the Bio-SPK (Bio-derived synthetic paraffinic kerosene)9. In this process, the bio-derived oil (triglycerides and free fatty acids) is first converted to shorter chain diesel range paraffins by reactions with hydrogen which removes the oxygen atom and increases its heat content, and byconverting the olefins to paraffins, which increases the thermal stability of the fuel. A fraction of this product is then



6/30


6

isomerized and cracked to branched paraffins to reduce its freezing point. Finally, the biodiesel fuel is blended withthe conventional jet fuel in equal amounts to incorporate the needed aromatics for sealing. Such a fuel has beentested in commercial aircrafts 9.

It is also noted that while the presence of the oxygen atom in biodiesel fuels slightly reduces their energycontent, its presence in the fuel molecular structure can actually reduce soot formation and emission, and as such isan important consideration in land-based transportation.

C. Ethanol and ButanolAlthough the low energy content of the light alcohols precludes their direct use as aviation fuels, their potential

for land transportation is sufficiently significant that a discussion of their properties is merited.Since sugar- and starch-derived ethanol displaces food supply and has substantial life-cycle CO 2 emissions, it is

not sustainable as a viable alternative fuel. Cellulosaic ethanol, produced from non-food sources such as switchgrass, holds the potential as a significant source of fuel.

Butanol has the advantages over ethanol in that it can be produced from a larger variety of biomass sources, hashigher energy content, has the same range of volatility as gasoline and is infinitely miscible with it, is not corrosive,and is also knock resistant. Commercial production of butanol through fermentation is targeted for the next two tothree years.

D. Hydrogen and MethaneHydrogen and methane both have high gravimetric but low volumetric energy contents. Furthermore, they are

cryogenic fuels and as such their use would require a new fuel infrastructure and new engines and airframes. Theproduction of hydrogen is also central to its utilization. If it is generated through reforming natural gas, as is mostlydone now, then CCS is needed to sequester the CO2 produced. In general, cryogenic fuels are not considered asviable next generation propulsion fuels, although the use of methane as an endothermic fuel for hypersonicpropulsion is being explored to circumvent the cracking and solid formation of conventional jet fuels in the fuel line8.

To further support the above analysis, a recent study by the U.S. Department of Transportation 4 emphasized thatfuels need to be produced from diverse sources for economic security reasons, and that overall life-cycle impact ofalternative fuels needs to be accurately assessed for informed action. Furthermore, the study concludes that: themost reasonable near-term choice is the use of indigenously available feedstocks, such as natural gas, coal, oil shale,and petroleum coke, to produce drop-in replacements/supplements for petroleum-derived jet fuels. Renewable biofuels are currently not capable of supplying a large percentage of fuel needs, but higher yielding futurefeedstocks, such as algae or cellulosic biomass, may improve feedstock supply. If the performance and cost

issues can be overcome, biofuels are envisioned to be blended with synthetic or conventional jet fuels. The studyfurther identified that research in the near term (10 years) renewable, fully biomass-derivedaviation fuels meeting the same performance and operation criteria as those for drop-in fuels will be enabled. Thestudy also emphasized the need for long-term, stable foundational research, including atmospheric and combustionchemistry, fluid mechanics of internal flows, acoustics, and computational science.

III. Development of Predictive Capability of Aviation Fuels: Towards Fuel DesignWhile hydrocarbon-based fuels will continue to be the fuels of choice for aviation transport, at least for the next

few decades, the actual makeup of these fuels, and the characteristics of the components, will continuously evolve inresponse to new sources of feedstocks, new technologies of fuel production and formulation, and new concepts inengine design and operation. For example, a recent DOE study 10 indicated that a 15% improvement in efficiency in

gas turbine combustion can be achieved through modifications of the combustion processes. It is also notinconceivable that new types of powerplants could be developed, capitalizing on new insights in the application ofthermodynamics, chemical kinetics, and fluid mechanics that would offer drastically improved propulsionperformance.

In order to respond to such new opportunities, readily and prolifically, it is necessary to develop predictivecapabilities for the combustion characteristics of new fuels in new regimes of combustion. Indeed, it is tempting tosuggest that, by integrating the designs of fuels, combustion processes, and engines for optimized performance,breakthrough advances can be made on enhanced energy conversion and reduced emissions through new designs offuel molecules and compositions.


7/30


7

While the potential for fuel design is exciting, it is nevertheless recognized that at present we still lack the predictive capability of the current generation of fuels, even at the level of the simplest fuels hydrogen andmethane. This recognition has led to a concerted effort towards the development of such a capability. An example isthe stated mission of the DOE-sponsored Combustion Energy Frontier Research Center to: develop a validated,predictive, multi-scale, combustion modeling capability to optimize the design and operation of evolving fuels inadvanced engines for transportation applications. Indeed, a consortium of government agencies, the Multi-AgencyCoordinating Committee for Combustion Research (MACCCR), was also formed to encourage such a development.

The complexity of the combustion processes within aero-engines, particularly the intricate coupling between thechemical kinetics of fuels oxidation and the physical processes of molecular and turbulent transport, can bedemonstrated by considering some of the controlling processes involved, from fuel injection to burnout. Here asthe liquid fuel, which consists of hundreds of components of vastly different physical and chemical properties, isinjected into the combustion environment, the liquid jet will first break up into a multitude of droplets, whichsubsequently gasify. The gasification sequence of the multicomponent droplets is critical because a fuel componentcan mix with the oxidizer and react only if it is first gasified. The subsequent ignition and combustion in the near-nozzle region is also critical as they provide the stabilization mechanism of the bulk flame. Reaction in this region isexpected to take place in a relatively cold environment, implying the importance of low-temperature chemistry inwhich kinetic pathways involving the HO2 radical are important, as will be discussed later.

The subsequent reaction within the turbulent flame brush takes place through either thin regions of wrinkledlaminar flames or broad regions of distributed reaction. Reactions in these situations are controlled by high-temperature chemistry, in which the kinetic pathways of the H radical, especially those of a branching nature, are

important. Preferential diffusion of reactants of different molecular weights could also significantly modify the localconcentrations from the freestream values, leading to local regions of intensified or weakened burning.

In the following we shall cite some examples of the intricacies of fuel vaporization and chemistry in affecting theoverall combustion process.

A. Intricacies of Fuel Vaporization and ChemistryWe first consider the gasification of droplets of fuel blends in a spray. According to the concept of batch

distillation, the gasification sequence would approximately follow the relative volatilities of the individualcomponents, with the more volatile components preferentially gasified first. Indeed, this is frequently themechanism used in simulating spray combustion in engines. What is overlooked in applying batch distillation is thatin order for a liquid element in the droplet interior to be gasified at the droplet surface, it has to first diffuse to thesurface. However, liquid-phase mass diffusion is an extremely slow process as compared to that of surfaceregression as the droplet gasifies. Consequently it could become the rate-limiting process in droplet vaporization 11.

Figure 5 shows the experimental time-resolved average decane concentration of a bi-component droplet ofdecane and hexadecane 12, which are respectively the more and less volatile components, and the correspondingcalculated values assuming batch distillation. It is seen that in spite of the higher volatility of decane, its depletionsignificantly lags that of the calculated value by assuming batch distillation, implying that substantial amount of it istrapped in the droplet interior due to diffusional resistance, not being able to be released to the gas for reaction.

We next consider the influence of the intricate pathways and intrinsically nonlinear nature of fuel oxidationchemistry. To include the effects of chemical reaction in the computational simulation of complex reactingturbulent flows, a simplified approach is to assume that the reaction between fuel and oxidizer occurs in a singlestep,

Fuel+ OxidizerProducts,

with a reaction rate given by

Reaction Rate =A[Fuel][Oxidizer]exp(-Ea/RoT),

whereA is a constant pre-exponential factor representing the frequency of collision, [i] the concentration of the ithspecies,Ea the activation energy, Tthe temperature, andR

o the universal gas constant. This expression indicates thatthe reaction rate increases monotonically with increasing temperature. Furthermore, since concentration isproportional to the system pressure, it implies that the reaction rate increases with pressure, with the increase beingquadratic. We shall now present examples demonstrating that this is far from the actual behavior of fuel chemistry,which is characterized by the prevalence of nonlinearity and nonmonotonicity in the outcome of combustionprocesses.

We first consider the well-known Z-shaped, triple pressure-temperature explosion limits in a homogeneoushydrogenoxygen mixture, shown in Fig. 6 13. Since hydrogen oxidation is a building block for hydrocarbon


8/30

oxidation, its oxidation characteristics are inherently embedded in those of hydrocarbon oxidation. It is then seenthat, by gradually increasing the system pressure from a low value, at a moderate temperature, the mixture becomesexplosive, non-explosive, and explosive again as it traverses the first, second, and third explosion limits. Theintricacy here is that since the collision frequency between the reactant molecules increases with pressure, onewould expect that the burning intensity will also monotonically increase with pressure, as just mentioned, instead ofexhibiting the nonmonotonic response.

This behavior is explained by noting that the lower explosion regime, below the second explosion limit, iscontrolled by a strong chain branching cycle involving the H, O, and OH radicals, as represented by the two-bodyreaction (R1): H+O2O+OH. However, as the mixture crosses the second limit with increasing pressure, the three- body termination reaction (R9): H+O2+MHO2+M, becomes important because of the increased collisionfrequency between molecules at higher pressures. Thus (R9) competes with (R1) for the highly reactive H atom,producing the rather inactive HO2 radical. This slows down the overall reaction rate and consequently suppresses theexplosion. Finally, with further increase in pressure, the continuous increase of the HO2 concentration willeventually trigger reactions involving the HO2 and H2O2 radicals, leading to formation of the reactive radical OH.The OH in turn initiates a second chain branching cycle, above the third explosion limit, as shown in Fig. 6. It istherefore clear that detailed chemical kinetics is needed to explain complex combustion phenomena of this nature.

This nonmonotonic behavior not only characterizes the ignition and explosion of hydrogen, it is also reflected inthe steady burning rates of hydrogen flames 11. Figure 7 plots the burning flux of a planar laminar flame of a leanhydrogen/air mixture as a function of pressure, and shows that it indeed exhibits a nonmonotonic response by firstincreases with pressure, then decreases, and finally increases again. Furthermore, the dependence on pressure not

only is not constant but can also be substantially less than that given by the quadratic response of the one-stepoverall reaction.

As another example of the intricacy of chemical kinetics, we show the ignition time delay of an n-heptane/airmixture as a function of temperatures in Fig. 8. It is seen that, with continuous increase in the mixture temperature,the ignition delay first decreases, then increases, and finally decreases again, constituting the so-called NegativeTemperature Coefficient (NTC) phenomena 13, 14. This nonmonotonic excursion is a consequence of the low-temperature chain mechanism, from about 650 to 950 K, and will be missed if it is not included in a detailed reactionmechanism.

Figure 5. Comparison of experimental data and calculated values assuming liquid-phase diffusion limited

and batch distillation modes of gasification, for the average molar fraction of decane in a decane-hexadecane

droplet undergoing vaporization in a 1,020K environment 12.



9/30

p

T

1st limit

2nd limit

3rd limit

Extended 2nd limit

Weak chain branching

H+O2+M HO2+M

HO2+H2 H2O2+H

H2O2+M OH+OH+M

Chain termination

H+O2+MHO2+M

Strong chain branching

H+O2 O+OH

O+H2 H+OH

OH+H2 H2O+H

k9

k1

k2

k3

k9

k17

k15

p

T

1st limit

2nd limit

3rd limit

Extended 2nd limit

Weak chain branching

H+O2+M HO2+M

HO2+H2 H2O2+H

H2O2+M OH+OH+M

Chain termination

H+O2+MHO2+M

Strong chain branching

H+O2 O+OH

O+H2 H+OH

OH+H2 H2O+H

k9

k1

k2

k3

k9

k17

k15

Figure 6. Explosion limits of hydrogen/oxygen mixtures, showing the nonlinear nature of the transitions

between different regimes with pressure and temperature variation; dominant reactions in differentpressure-temperature regimes are indicated.

B. Development of Detailed Reaction MechanismsThe above examples demonstrate the complexity in describing the oxidation of a fuel that is comprehensively

applicable over extended thermodynamic ranges of temperature, pressure, and composition. To further illustratesuch a complexity, we again consider the reaction between hydrogen and oxygen leading to the production of water.While globally there are only three species, namely H2, O2, and H2O, at the level of elementary reactions involvingthe collision between distinct atoms and molecules, there are five additional radical species, namely O, H, OH, HO2,and H2O2, that actively participate in the chemical transformation. Each of the elementary reactions has its ownsensitivity to temperature and pressure, which represent the collision energy and frequency as well as the identity ofthe colliding partner. Consequently it requires a minimum of 19 elementary reactions in order for the mechanism to be comprehensive. These reacting species are strongly coupled to each other through the complex network ofelementary reactions, for example with the product of a reaction being the reactant of another reaction, etc.


Figure 7. Laminar burning rate of a lean hydrogen/air mixture, with equivalence ratio of 0.35, showing

the nonlinear variation with pressure due to the evolvement of the dominant chain reactions shown in Fig. 611.


10/30

Figure 8. Calculated ignition delay of stoichiometric n-heptane/air mixture at 1 atm. pressure, showing the

importance of low-temperature chemistry 15.

Figure 9 shows the reaction pathways of the smallest hydrocarbon 15, methane (CH4), demonstrating the plethoraof the intermediate species and reactions involved. The plot, however, also shows that the progression of thereactions can be traced rationally. Specifically, the main trunk of the reaction pathways shows the progressivereduction of the original fuel molecule, CH4, through reactions with O, H, and OH to eventually yield the finalproduct CO2; with the other product, H2O, produced along the path. This main oxidizing trunk also has a minor sidetrunk, to its left, to describe reactions involving the methyl radical (CH3) which is the first radical formed after an Hatom is abstracted from the fuel molecule. The major side trunk, to the right, represents growth of the participatingspecies to the larger molecules consisting of two carbon atoms, such as ethane (C 2H6) and ethyl (C2H5), and therebyprovides the linkage with even larger molecules and hierarchy of the reaction pathways. A widely adopted methaneoxidation mechanism, GRI Mech 3.0 16, consists of 53 species and 325 reactions.

The complexity multiplies rapidly with increasing size of the fuel molecule. Figure 10 shows the size of more

than twenty detailed and moderately simplified mechanisms for hydrocarbon fuels of various molecularcomplexities compiled over the last two decades 17. It is seen that the number of participating species, N, andreactions, K, increase with the size of the molecule, roughly in an exponential trend. Specifically, it is seen thatwhile typical mechanisms for the C1 and C2 species consist of less than about a hundred intermediate species, thoseof the larger fuels would consist of hundreds of species and thousands of reactions. As an extreme example, the sizeof the detailed mechanism for methyl decanoate, a biodiesel surrogate, consists of 3,036 species and 8,555 reactions.

Figure 10 also shows that there appears to be an approximate linear correlation between the number of reactionsand number of species, K 5N, which is a convenient relation to estimate the cost involved in the computation ofphenomena involving such fuels. It is also emphasized that mechanisms for some of the larger molecules shown inFig. 10, such as n-heptane, are so large that it is a challenge to even incorporate them in the computation of simplecombustion phenomena such as the freely propagating one-dimensional planar flame described in relation to Fig. 7.

The above demonstration of the complexity of the reaction mechanisms was only for a single fuel species, suchas n-heptane. Since a practical fuel blend typically consists of hundreds of fuel species, the potential size and

complexity of the resulting mechanism would be enormously large.It is therefore a daunting task to develop a comprehensive, detailed mechanism for liquid fuels of large

molecular sizes. Several principles have been followed in such developments. The first is fuel hierarchy 18,mentioned earlier, in that the mechanism of a small hydrocarbon must be a subset of that of a larger hydrocarbonfrom which it can be derived. Consequently mechanisms of small hydrocarbon units must form the foundation in thedevelopment of mechanisms of larger hydrocarbons. Hierarchically, starting from the smallest fuel units we haveH2/O2, CO/H2/O2, CH4/CH2O/CH3OH, C2H6/C2H4/C2H2, and so on. Second, mechanisms involving the lowerhydrocarbons, from C0 to C4, are critical, and as such the thermochemistry and rates of their reactions must bedetermined accurately, using ab initio quantum chemistry methods. Third, classes of reactions have been identified



11/30

that behave similarly, allowing systematic compilation. For example, 25 classes of reactions have been identified todescribe the ignition chemistry of hydrocarbons 19.

Computer algorithms have also been developed allowing quantum mechanical calculations of the reaction ratesof large molecules and the automatic assembling of large reaction mechanisms consisting of important intermediatesand reactions determined through sensitivity analysis 20. A close coupling between the two components would allowidentification of the important species and reactions that would require high-level quantum chemistry calculations,leading to refinement of the accuracy of the assembled mechanism. This new ability to compute rapidly any requiredrate coefficient to useful accuracy, and to identify automatically which species and reactions are important for a newfuel, under specified reaction conditions, all without human intervention, holds promise to significantly advancecombustion chemistry, much as automated DNA sequencing has transformed biological and medical researchcompletely.

Figure 9. Reaction pathways for the oxidation of methane. Adapted from Ref. 15.

Any detailed mechanism developed must be validated experimentally. There are two levels of validation. First, itmust be validated by experiments that are truly chemical in nature, without any transport effects. Suitableexperiments involve determination of the temporal evolution of the species concentration in a reacting mixture byusing shock tube 21, flow reactor 22, jet-stirred reactor 23, and rapid compression machine 24. The experimentalcoverage in temperature, pressure, and concentration must be as extensive as possible. Second, the validation mustalso be conducted for various classes of combustion phenomena such as homogeneous and diffusive ignition whichcover low- to high-temperature chemistry 25, steady burning and extinction which cover high-temperature chemistry



12/30

26, 27, and premixed and nonpremixed flames which cover the relative concentrations and mixedness of the fuel andoxidizer. The global combustion responses of interest should include the laminar flame speed, ignition andextinction strain rates, detonation induction length, detailed thermal and concentration structures of flames anddetonations, oscillatory and pulsed unsteady effects to potentially discriminate reactions of different time scales, andpollutant chemistry.

101

102

103

104

102

103

104

before 2000

2000 to 2005

after 2005

iso-ocatane (LLNL)

iso-ocatane (ENSIC-CNRS)

n-butane (LLNL)

CH4 (Konnov)

neo-pentane (LLNL)

C2H4 (San Diego)

CH4 (Leeds)

MethylDecanoate(LLNL)

C16 (LLNL)

C14 (LLNL)C12 (LLNL)

C10 (LLNL)

USC C1-C4

USC C2H4

PRF

n-heptane (LLNL)

skeletal iso-octane (Lu & Law)

skeletal n-heptane (Lu & Law)

1,3-Butadiene

DME (Curran)C1-C3 (Qin et al)

GRI3.0

Numberofreactions,

K

Number of species, N

GRI1.2

K = 5NK 5N

Figure 10. Plot showing the size (K, N) of reaction mechanisms for various fuels; note the

continuing increase in the mechanism size with time, and the approximate relation of K 5N17

.

Figure 11 28 shows excellent agreement between the experimental shock tube data and the calculated values for

the reaction rate constant of the elementary reaction OH + propene products, hence demonstrating thesophisticated state in the theoretical and experimental studies of the kinetics of elementary reaction rates.

Figure 12 compares the experimental and calculated values of the ignition delay time of n-heptane, againdemonstrating satisfactory agreement 29. Recognizing that the ignition delay indicates the global response of the

reactive mixture, such an agreement lends confidence to the use of the reaction mechanism in the simulation ofengine processes.

Figure 13 shows the measured and calculated laminar flame speeds of methane/air mixtures at 1 and 5 atmpressure 30. The comparison is more stringent in this case as compared with that of the ignition delay because of theadditional presence of diffusion in flame propagation. The good agreement demonstrates not only the adequacy ofthe detailed mechanism, but also the accuracy in the experimental determination of the laminar flame speed which isan important parameter characterizing a combustible mixture.



13/30

Figure 11. Calculated and shock tube determination of the reaction rate constant for the reaction OH +

propene products, showing close agreement. Adapted from Ref. 28.

Figure 12. Calculated and experimental ignition delay for n-heptane in air, showing close agreement.

Adapted from Ref. 29.

Caution is needed in interpreting experimental results obtained from transport-affected reacting systems, and the

inverse situation of applying chemistry data to realistic combustion phenomena which are frequently affected bytransport. We cite two examples to illustrate the subtlety. First, it is noted that while the laminar flame speed isdefined for the adiabatic, freely-propagating planar flame in the doubly-infinite domain, laboratory flames used in itsdetermination seldom conform to this idealized configuration. Examples are the planar flame situated in thenonuniform counterflow generated by impinging two flows against each other, the flame curvature experienced bythe Bunsen flame, and the unsteadiness and curvature experienced by an outwardly propagating flame. These flamestherefore are affected by what is collectively referred to as aerodynamic stretch effects 11, which need to besubtracted out from the stretch-affected raw experimental data before they can be used for chemistry studies. Indeed,without recognizing the presence of these stretch effects, early data on the laminar flame speed showed extensivescatter (Fig. 14) which did not diminish with improved experimental techniques. It is not until the early 1980s when



14/30

this effect was recognized, and a rational approach towards its elimination was suggested, that data from differencesources converged. The accuracy of the extraction which directly affects the chemical information embedded in theextracted flame speeds, however, is still an active topic of investigation 31, 32.

Figure 13. Experimental and calculated laminar flame speeds of methane/air mixtures at 2 and 5 atm

pressures, showing close agreement30

.

1920 1940 1960 1980 200025

30

35

40

45

50

Year of PublicationBurning

Velocity(cm/s)at298K


Figure 14. Reported maximum laminar flame speeds (burning rates) of methane/air mixtures, showing

considerable scatter before stretch effects were recognized and systematically eliminated in the early 1980s,

leading to collapse of the data.


15/30

The second example concerns the extent of influence of the NTC on ignition in diffusive systems. Take, for

example, ignition in the counterflow or stagnation flow in which the flow imposes a residence time on theattainment of ignition runaway. Consequently a higher temperature is needed to achieve ignition, as shown in Fig.15 for the nonpremixed stagnation ignition of heated air impinging upon an n-heptane pool 33. This shifts theignition chemistry away from that of the NTC temperature regime, thereby either minimizing or even obviating itsinfluence.

C. Development of Reduced MechanismsThe availability of a comprehensive detailed reaction mechanism does not mean that it can be readily adopted for

computational simulation. The extent of the computation involved can be appreciated by noting that the computationcost in resolving the chemical structure of a reactive-diffusive flow varies at least with N 2. In fact, except for thesmallest fuels such as hydrogen and methane, and for such simple combustion systems as the 1-D planar laminarflame, detailed mechanisms of the larger fuels are frequently too large for simulation without substantial reductionin its complexity. Furthermore, detailed mechanisms are also characterized by the dramatic differences associatedwith the species and reaction time scales. For example, Fig. 16 shows that the shortest chemical time scale forethylene, which is comparable to the molecular collision time, is O(0.1-1 ns), which is not much shorter than theflow time which can be O(10ns). Furthermore, the stiffness for the larger molecule, n-heptane, calculated with ashortened mechanism, is even worse. These differences could result in severe chemical stiffness in simulations.

The above two features, namely large size and chemical stiffness, are therefore the major factors preventing

detailed kinetics of realistic fuels from being applied in large-scale simulations. Consequently, other than tabulation,substantial reduction is needed to make these mechanisms applicable.

Mechanism reduction is usually conducted in two stages 17. The first is to eliminate reactions and species that arefound to be unimportant either comprehensively for all possible combustion situations, or locally for a givenapplication, resulting in what is called a skeletal mechanism. A noted algorithm is the Directed Relation Graph(DRG) 17. Chemical approximations such as steady-state species and partial equilibrium reactions are then applied tothe skeletal mechanism, guided by such mathematical assessments as computational singular perturbation (CSP) 34and Intrinsic Low Dimensional Manifold (ILDM) 35 to separate out the fast and slow chemical entities and thesubsequent lumping of the reactions. The size and chemical stiffness of the mechanism are therefore further reduced.

-12 -10 -8 -6 -40

5

10

15

20

25

30

35

Numberofspecies

log10

(, sec)

ethylene

n-heptane

p = 1 atm

T = 2000K

= 1

Figure 15. Ignition temperature in the stagnation flow

of heated air impinging onto an n-heptane pool, with

different strain rate, showing the shifting of the

ignition temperature to regimes higher than that of

the NTC phenomena33

.

Figure 16. Histogram of species time scales in PSR

with fixed temperature and residence time of 1ms

for a 33-species mechanism for ethylene and a 78-

species mechanism for n-heptane, respectively 16.

Figures 17 and 18 36 compare the calculated ignition delay time, the response of the perfectly-stirred reactor(PSR), and the laminar flame speed for n-heptane/air mixtures at various pressures, using a detailed mechanism that



16/30

consists of 561 species and 2539 reactions, and a reduced mechanism that consists of 78 species and 359 lumpedreactions. It is seen that there is practically no loss of accuracy in the reduction for the purely chemistry controlledsystem of ignition and PSR response, and minimal loss of accuracy for the laminar flame speed.

Figure 17. Calculated auto-ignition delays and PSR (perfectly stirred reactor) responses of a

stoichiometric n-heptane/air mixture, using the detailed and a reduced mechanism with 78 species,

demonstrating that there is practically no loss in accuracy in the reduction36

.

Figure 18. Calculated laminar flame speeds of n-heptane/air mixtures at 1 atm pressure, using the detailed

and a reduced mechanism with 78 species, demonstrating close agreement. Adapted from Ref. 36.



17/30

Additional strategies have also been developed to reduce the computational time in reacting CFD. Notable

examples are the lumping of isomers 17 which have similar transport properties but different reactive properties, andthe lumping of the diffusion coefficients of species having similar values, recognizing that such evaluations areequally if not more taxing than those for the reaction rates. Figure 19 shows that the diffusion coefficients for all thespecies participating in heptane oxidation can be lumped from 188 species to nine groups with minimal loss ofaccuracy.

To demonstrate the significant extent of reduction achievable, Fig. 20 shows that by applying even the first stagereduction, the extremely large biodiesel surrogate mentioned earlier, methyl decanoate, can be reduced to a skeletalmechanism of only 125 species and 713 reactions, with a 20% error of reduction 17. As an example of the saving incomputation time by using reduced mechanism, for the DNS (direct numerical simulation) of a temporally evolvingnonpremixed sooting ethylene-air planar jet flame in a random 2D isotropic turbulent flow 37, the estimatedcomputation time using a detailed mechanism, consisting of 70 species and 463 reactions, would take about 2.3million CPU hours. This time is reduced to 70,000 CPU hours by using a reduced mechanism consisting of only 19species and 167 reactions, representing a factor of 70 in the speedup of the computation time.

Three additional observations can be made regarding accommodating realistic chemistry in large-scalecomputational simulations. First, while we have cited impressive extent of mechanism reduction, the size of thereduced mechanisms for such moderately large fuels as n-heptane is still quite large, around 50 to 60 species. Theconcern is that for larger species and for fuel blends consisting of hundreds of fuel species, their reducedmechanisms could be much larger, rendering them difficult to be integrated into CFD efforts. Recent studies,

however, indicate the possibility that, at least for hydrocarbons, there is an asymptotic limit in the size of thereduced mechanisms. In fact, it appears that this limit could be close to that identified for n-heptane, namely 50 to 60species. The possible existence of such a limit is not unreasonable as there are just so many classes of reactions forhydrocarbon oxidation, involving so many species. This possibility is encouraging because even at the present stateof computational capability, DNS and large eddy simulation (LES) studies of small-scale laboratory and modelturbulent flames can actually accommodate mechanisms of this size. In other words we may have already reachedclosure in accommodating fuel chemistry in small-scale fundamental CFD studies. Consequently, with furtheradvance in computational capability, the additional computational resources can be directed towards improvingdescriptions of the fluid mechanical aspects of the phenomena.

Figure 19. Comparison of the laminar flamespeeds of n-heptane/air mixtures at 1 arm

pressure, using a 188-species skeletal mechanism,

and reduced models of 19, 9, and 3 bundled

groups17.

Figure 20. Number of species in skeletal mechanismobtained using the directed relation graph (DRG)

method as a function of the user-specified relative error

tolerance, for the extremely large mechanism of methyl

decanoate oxidation 17.

The second point is the need to describe realistic, complex and large-scale combustion phenomena. Examples areprocesses within engines and those associated with such astrophysical phenomena as the supernova explosion. Forthem we advocate the development of dynamic reduction, namely reduction-on-the-fly. The viability of such a



18/30


18

concept is sound because, in spite of the desire to have mechanisms to be comprehensive by including all relevantspecies and reactions covering all relevant regimes of the thermodynamic space, the number of participating speciesand reactions that control the reactions locally, in space and time, must be necessarily very small. Thus the size andstiffness of a large detailed mechanism become largely irrelevant, as only a very small subset of it is locallycontrolling. Consequently concerted efforts should be directed towards developing mathematical and computationalalgorithms for dynamic reduction. When this is accomplished, the task of accommodating realistic chemistry inCFD simulations degenerates to that of supplying accurate and detailed kinetic information to the computationalcode, without the concern of any computational handicapping.

Third, it is evident that the mathematical and computational techniques developed for the present chemical phenomena hold the potential for application to other large and stiff reacting systems, such as those involvingnuclear reactions, biological reactions, and plasma reactions, and merit exploration.

D. Development of Reduced FuelsReduced mechanisms described above are mainly useful for fundamental studies involving only a single or at

most several fuel species whose individual detailed reaction mechanisms are available. The reduction procedure,however, cannot be readily applied to a practical fuel blend because of the size and complexity of the composition aswell as the difficulty of its complete determination. Furthermore, even if it can be completely determined, thecomputational effort involved in reducing them would be inordinately large.

It is, however, reasonable to expect that the number of species or species groups that would have stronginfluence on the global kinetic behavior of a fuel blend may not be too large, due to the low concentrations of most

of the components and the similarities in the reaction mechanisms of groups of them. It is therefore practicallyuseful to develop surrogates which consist of a small number of fuel species whose global kinetic responses wouldemulate those of the fuel blend. Thus as long as the components of the surrogates can be identified, and theirindividual detailed reaction mechanisms also developed, then a much smaller consolidated mechanism can beassembled and subsequently reduced, leading to facilitation of the computational simulation of the reacting flow andinterpretation of the results. These surrogates can then be used in engine testing and development, with the assurancethat the testing results are relevant to practical situations using the fuel blend. These engine tests can also becomputationally simulated, scrutinized, and optimized with confidence. It is emphasized that conventional enginedesign approaches relying on prototype development are too time-consuming and expensive. Consequently, theavailability of predictive and efficient computational codes incorporating realistic but computable, surrogate-basedchemistry, together with equally realistic and computable descriptions of turbulent flows, would provide criticalguidelines in such developments. We note that in analogy with using reduced mechanisms in simulating combustion behavior, surrogate fuels can be considered as reduced fuels in the simulation of the combustion behavior of

practical fuel blends38

.The use of surrogates, of course, has been the standard practice in the development of gasoline and diesel

engines, with n-heptane and iso-octane (2,2,4-trimethylpentane) being the surrogates for the former, and n-hexadecane and iso-cetane (2,2,4,4,6,8,8-heptamehtylnonane) for the latter. By nearly matching the boiling points ofthe surrogate components not only with each other but also within the volatility range of the fuel blend, the primarytarget of these surrogates is the emulation of the knock tendency of the fuel blend under different engine runningconditions.

For the development of surrogate jet fuels, the leading-order targets for the thermal and physical properties arethe carbon-to-hydrogen ratio (C/H) which controls the overall heat release, the mass density which yields thegravimetric and volumetric energy densities, and the volatility range which controls the vaporization rate and thesequence of the gasifying components. In terms of its chemical behavior, a surrogate should consist of species andtheir respective concentrations that would emulate the behavior of the major groups of compounds in the fuel blend.Using Fig. 21 for JP-8 as an example, the reduced reactivity of the iso-paraffins and aromatics relative to the

normal-alkanes should emulate the reactivity differentials of the fuel blend. The aromatics, needed for enginesealing, affects radiative transfer within the combustor as well as the emission characteristics due to its propensity toform soot. The reactivities of the cyclo-paraffins are largely similar to those of the normal-paraffins. However, sincethey are present in relatively large amounts and have larger bond energies, it is important to capture their influence.

A proposed surrogate, of course, also needs to be validated through laboratory experiments. The targets for thevalidation tests should be the same as those for the detailed mechanisms. In terms of the performance of aero-engines, a surrogate should be able to emulate such global events as ignition, blowout, heat release rate, radiationloading, soot formation and emission, combustion staging, and lean premixed prevaporized applications.

Surrogates have been proposed for jet fuels 39 and diesel fuels 40. Examples of diesel fuel surrogates are those ofn-decane, n-butlycyclohexane and n-butylbenzene 40, while the surrogates for jet fuels are n-decane, iso-octane, and


19/30

toluene 38. In particular, Fig. 22 compares the ignition delay times obtained for a jet fuel POSF 4658 and thesurrogate of38, together with calculations based on the surrogate composition. It is seen that the surrogate was ableto capture, even quantitatively, the NTC response of the jet fuel.

Identification of surrogates is still at the early stage of development, not unlike the state of mechanism reductionin the 1980s and 1990s that relied heavily on the experience and input of the developer. It is suggested that furtherresearch should include computer-assisted, automated generation of the surrogate composition given certainempirical constraints of the propellant behavior.

Figure 21. Major and trace compounds in a jet fuel composition. Adapted from Ref. 9.

Figure 22. Ignition delay times for POSF 4658, 16.324.8 atm (solid symbols); POSF 4658 surrogate, 17.7

22.9 atm, (hollow symbols); and model simulation of POSF 4658 surrogate (lines). Black data correspond to

shock tube measurements/simulations, red data correspond to RCM measurements/simulations38.



20/30


21/30

carbon to CO, then the hydrogen to H2O, and eventually the CO to CO2. Thus one expects that soot would beformed when a system is sufficiently fuel rich that there is not enough oxygen to convert all the carbon to CO. Thethreshold stoichiometry for soot formation, say for acetylene in air, is given by,

C2H2 + (O2 + 3.76N2) 2CO + H2 + 3.76N2

Consequently, while the threshold fuel-to-oxidizer equivalence ratio of soot formation for ethane (C2H6) with aC/H ratio of 1/3 is 3.5, it is at a leaner condition of 2.5 for cubane, benzvalene, and acetylene, all of which have C/H

= 1, indicating their stronger propensity to soot.As an exaggerated example of soot formation in a highly-strained molecule, we consider the cage molecule,

carborane (Fig. 25), which has two carbon and ten boron atoms, with a very high energy density of 61.6 MJ/kg ascompared to 42.8 MJ/kg for Jet-A. In view of the high heat of combustion for boron on both gravimetric andvolumetric bases, and the difficulty of burning it in particle form due to the difficulty in melting its surface oxidecoating, it has been suggested that building boron into the cage structure could be an alternate way of utilizing it.

Figure 26 shows the time-resolved image of a freely-falling, burning, ~200m, droplet of a carborane derivative inheated air, with a diffusion flame surrounding it 48. It is seen that as burning progresses the amount of soot formed inthe flame zone surrounding the droplet rapidly increases and eventually envelops the droplet and terminates thecombustion. Figure 27 shows the collected samples of such soot shells, which contain the unburned liquid fuel in theinterior.

Figure 25. Molecular structures of highly strained molecules: benzvalene, cubane, and carborane.

Figure 26. Images of a freely-falling, ~200m,burning carborane droplet, showing the growth of a

soot layer in the flame region that eventually

envelops the droplet and terminates the burning 48.

Figure 27. Collected samples of the soot shell formed

in the burning sequence of Fig. 26 48.



22/30

C. Functionalized MoleculesThe energy density and reactivity of a hydrocarbon fuel can be increased by attaching energetic and/or catalytic

functional groups to its molecular structure. An example is the azide group, N 3, which releases 50 kcal/mole of heatupon breaking of the N-N2 bond

49. Experiments have shown that droplets of organic diazides exhibit exceptionallyfast gasification rates. Figure 28 shows the flame streaks of streams of freely falling droplets of an n-alkane(heptane), an azido-alkane (1-azidohexane), and a diazido-alkane (1, 6 diazidohexane). It is seen that the diazide notonly burns much faster, as indicated by the wider and shorter flame streak, but its burning is also terminated by self-induced explosion, leading to instant and complete fracturing and thereby gasification of the droplet. These resultsare quantified by the gasification rate constant, defined as (the (negative of ) the regression rate of the square of thedroplet diameterd, K= - d(d2)/dt, and the normalized droplet diameter at explosion, de/do, where do is the initialdroplet diameter. Large values of either of them would imply overall faster droplet gasification. Figure 29 plotsthese two quantities for alkanes and diazido-alkanes, with different carbon numbers, for both vaporizing and burningdroplets. It is seen that the gasification rates of diazido-alkanes are uniformly higher than those of alkanes, and thatall diazides explode. Furthermore, while the faster gasification rate is particularly prominent for the C5 and smallermolecules, the explosion propensity is enhanced for the C6 and larger molecules. Consequently the effective dropletgasification time is reduced for all diazido-alkanes: through the faster gasification rate for the smaller alkanes andearly explosion for the larger alkanes.

Figure 28. Flame streaks traced by a stream of freely

falling burning droplets of n-heptane, 1-azidohexane,

and 1, 6-diazidohexane, showing the very fast

burning rate of the diazidohexane, which also

terminates in a strong droplet explosion 49.

Figure 29. Comparison of the burning rate

constants and droplet explosion size for vaporizing

and burning droplets of alkanes and diazido-alkanes,

showing the fast burning rate and facilitated droplet

explosion of the diazide 50.

The fact that these facilitating effects were observed for both the vaporizing and burning droplets indicates thatdecomposition of the diazide occurs in the liquid phase, within the gasifying droplet, as it is heated up. This can bedemonstrated by considering the expression for the droplet burning rate constant,K, defined above and given by the

classical d2

-law11

,

8 ln(1g g

l

DK B

= +

)

where

,( )p s c O

v

c T T q Y vB

q

+=/ O



23/30

is the Spalding transfer number, gand respectively the gas and liquid densities,Dgthe gas diffusivity, cp and gas

specific heat, T and Ts respectively the ambient and droplet temperatures, qc the oxidative heat release from the gas-

phase reaction and qv the latent heat of vaporization, YO, the ambient oxidizer concentration, and O the oxidizer-to-fuel mass ratio. The decomposition heat release first reduces the latent heat needed for gasification, leading toincreased gasification rate, and then raises the droplet temperature to its limit of superheat, at which ithomogeneously nucleates, resulting in instant droplet fragmentation. The latter is favored for larger molecules which

have higher boiling points. It is noted that since the magnitude of the decomposition heat release is comparable tothat of the latent heat of vaporization but much smaller than that for oxidative heat release, it has a much greaterimpact onB when released in the liquid instead of in the gas.

l

Additional experiments showed that the above effects are mainly exhibited for the diazides, with only weakeffects for the monoazides or triazides, that the burning rate can be further increased at higher pressures, and thatdroplet explosion can also be further facilitated through the doping of dihalides 50. In fact, increases by factors of 20to 30 in the overall gasification rate can be achieved, leading to almost instant gasification of the droplet and as suchrendering droplet gasification not a rate-limiting factor in liquid fuel combustion.

D. Hypergolic PropellantsA hypergolic propellant system consists of a fuel and an oxidizer pair that spontaneously ignites upon contact.

Examples of the common fuels are hydrazine (H2N4) and its derivatives monomethylhydrazine (MMH) andunsymmetrical dimethylhydrazine (UDMH), while the common oxidizers are nitrogen tetroxide (N 2O4) and(inhibited red fuming) nitric acid (IRFNA). These hypergols are stable in ordinary temperatures and pressures, andhence can be stored for extended periods. Furthermore, since engines burning hypergolic propellants do not need anignition system, they are less complex in design and more reliable for repeated operation.

The hypergolic fuels, however, are extremely toxic while the oxidizers are also corrosive. One approach toreduce the toxicity in handling, and thereby reduce potential health hazards to the workers, is to add gelling agents tothe propellant so as to reduce its vapor pressure 51. Gelling has the additional benefits of improved safety in storage;better compliance with insensitive munition requirements; and reduced leakage, spillage, slosh and fire hazards. Thepropellant energy density and reactivity can also be increased with the use of reactive gallants and the addition ofmetal particles as suspension.

Another class of high-energy-density propellants that operate on the principle of hypergolicity is the ionic liquids(ILs), which are low melting, highly energetic salts. They can also be dissolved in a solvent such as water to form aliquid propellant. They have high decomposition temperatures and thereby thermal stability, and have essentially novapor pressure at ordinary conditions. In addition, because of the large number of chemicals that are ILs, and thesubstantial variation of their molecular constituents and structures, there exists much potential in increasing their

energy content through the attachment of various energetic functional groups including the light metals.A well studied ionic liquid is a solution of hydroxylammonium nitrate (HAN) and triethanolammonium nitrate

(TEAN) in water, used as a gun propellant 52. An example of more powerful systems is liquid azide salts reactingwith IRFNA or N2O4

53.Since the gelled hypergolic propellants (GHPs) and the ILs can be quite viscous and frequently non-Newtonian

in their rheological properties, the fluid mechanics of propellant jet formation upon injection, breakup, andimpingement are of particular interest. Furthermore, they react only if the two impinging liquid jets or the dropletsproduced from their breakup can physically make contact with each other and then mix. Recent studies on dropletcollision have however shown that, upon impact, two colliding droplets can actually bounce away instead merge,and the propensity to either merge or rebound is non-monotonic with increasing impact inertia 54. Specifically, withincreasing impact inertia, the droplets will first merge, then bounce away, and then merge again, as shown in Fig.30. Physically, what controls merging and rebound is the ability of the colliding droplet interfaces to squeeze out theinterfacial gas film. Since the mass inertia of the gas film increases with increasing environment pressure, and

recognizing the high-pressure environment within internal combustion engines such as rockets, bouncing and hencethe failure to merge, mix, and react is an issue that has not been adequately studied for propellant systems. Note thatthe fluid flow in this phenomenon ranges from that of continuum to rarefied; the latter is needed to describe themotion between the colliding interfaces because they must be close enough to be within the range of molecularattraction at the incipient state of merging.

If merging does occur, it is also desired that extensive and rapid mixing should take place so as to facilitatereaction within the merged liquid. Extensive mixing, however, is not favored when the two droplets are ofcomparable size because of symmetry. On the other hand if there is a substantial difference in the droplet sizes, thenthe smaller droplet can penetrate into the interior of the larger droplet, effecting extensive jet mixing. Figure 31shows the development of the mixing process subsequent to merging for two droplets of unequal sizes, as the impact



24/30

inertia increases. It is seen that with minimal impact inertia, a jet is formed upon merging and subsequentlypenetrates into the droplet interior, causing effective mixing. In this case the jet acquires its inertia from the surfaceenergy that is eliminated upon merging of the colliding interfaces. However, with further increase in the impactinertia, the increased distortion of the droplet shape absorbs the impact inertia and suppresses the jet formation.Eventually, the jet re-emerges as the impact inertia further increases.

Figure 30. Images of binary droplet collision, showing the phenomena of merging, bouncing and

merging again with increasing collision inertia 54.

Figure 31. Images of binary droplet collision and merging, for unequal size droplets, showing the

phenomena that with increasing impact inertia, an internal jet is developed, not developed, and developed

again.

E. Nanoparticles as Propellant AdditivesRecent understanding and advance in materials science have led to exciting potentials in the development of

propulsion fuels. These include nano particles, either inert or catalytic, nanotubes, graphenes, and reactivenanocomposite powders 55.



25/30

There are several novel features of particles when their size is reduced from the range of microns to those ofnanometers and tens of nanometers. To demonstrate this point, Fig. 32 plots the percent of the surface and bulkatoms of a spherical iron crystal, and shows that the percent of surface atoms increases rapidly as the particle size isreduced below say 10 nm 56. Physically, since the surface atoms are attracted only to the subsurface atoms while theinterior atoms are attracted by all atoms surrounding them, the surface atoms possess a greater amount of energy.Consequently, as the surface atoms become more abundant, through this excess energy they exert progressivelygreater influence on the physico-chemical properties of the particle such as its melting point, freezing point, thermalconductivity, catalyticity, etc. As an example, Fig. 33 shows the rapid reduction in the melting point of a tinnanoparticle as the particle diameter is reduced below 20 nm 57.

Figure 32. Surface to bulk atom ratio for spherical

iron crystals 56.

Figure 33. Depression of melting point for tin

nanoparticles with deceasing particle diameter 57.

An immediate benefit of using nanoparticles is the potential reduction of the ignition temperature of aluminumparticles, which is controlled by the melting of the protective surface oxide coating that is invariably present. The

melting would expose the inner aluminum core to the oxidizing environment so that reaction can be initiated. Inview of the potential reduction of the melting temperature with decreasing particle size, it is then reasonable toexpect that the ignition temperature of aluminum particles will correspondingly decrease. Figure 34 shows that it

indeed decreases steadily from about 2300 K to 1000K as the particle size is reduced from 10 to 0.1 m 58.Seeding of nanoparticles in solid propellants has also been shown to increase their surface regression rate. There

are four size-related causes for the increase 47, 55, 59. First, the embedded particles can be swept into the gas throughthe out-gassing mass flux, and as such largely eliminates the agglomeration problem for larger particles mentionedearlier. Second, ignition of the nanoparticles can be attained earlier due to the depression of the oxide melting point.Third, the particles, being smaller, also burn out earlier and as such have a higher total heat release rate for the sameamount of particle loading. Fourth, if the particles are in addition functionalized and catalytic, then the burning ratecan be further increased. Because of the faster burning rate of the seeded nanoparticles, the flame would be situatedat a closer distance above the propellant surface, which increases the heat transfer rate to the surface and thereby thesurface gasification rate.

Figure 24b shows the fine flame streaks emanating from the solid propellant surface for a propellant with 0.15m Al powder addition 47. Compared to the granular structure of the flame streaks of the 50 m-size Al of Fig. 24a,the absence or minimal extent of agglomeration is demonstrated. Figure 35 shows the substantial increase in the

linear burning rate of the solid propellant with either partial or complete substitution of the Al by the nAl powders.A similar substantial increase was also observed in the burning rate of nitromethane doped with functionalizedgraphene sheets (FGS) 60. These studies also show the additional result that the pressure exponent is reduced withdoping, indicating a corresponding reduction in sensitivity for the propellant to exhibit unstable combustion.



26/30

The role of increased condensed-phase thermal conductivity with suspended nanoparticles is not clear. Clearlyduring steady-state burning the effect on the burning rate is small. The effect, however, could be substantial duringtransient burning, particularly through its role on combustion instability.

Figure 34. Ignition temperature as a function of particle diameter for aluminum particles, demonstrating the

depression of the melting point of the oxide coating. Adapted from Ref. 58.

Figure 35. Linear burning rates of a solid propellant with addition of micron- and nano-size aluminum

particles, showing higher values for the latter47

.

One complication with using nanoparticles as propellant additive is the increasing content of the surface oxiderelative to the metallic inner core as the particle size decreases. For example, the amount of Al2O3 in a 50 nm



27/30

aluminum particle reaches ~43% 55. Clearly for such small particles the advantage of the higher energy content ofthe metal is compromised by the substantial amount of nonreactive metal oxide coating. Thus there is considerableinterest to synthesize stabilized metal nanoparticles without the passivating oxide layer. One approach is to passivatethe nanoparticle surface by organic self-assembling monolayers (SAMs) 61, which are densely packed organic thinfilms that spontaneously organize on a materials surface through chemisoption of a molecular amphiphile 62. Sincethe SAM is only one molecule thick, it is the minimal surface structure for passivation.

Reactive nano-composite materials have also been formulated to promote the burning rates of m-sizemetallized (e.g. aluminum) solid propellants 63, 64. These materials are themselves m-size particles made up of well-mixed nanoparticles of Al and an oxide (e.g. MoO3, CuO, and Fe2O3)with which it can readily undergo exothermicthermite reactions 65. The rapid heat release from the thermite reactions is expected to facilitate the burning ofneighboring Al particles or even agglomerates. Thus while the exothermicity of these thermite reactions is less thanthat of Al, by facilitating the ignition of Al it increases the overall burning rate of the propellant.

Mixtures of nano aluminum powder and water have been suggested as a chemical propellant, for several reasons66. First, since Al2O3 and hydrogen are the major products, it is environmentally more acceptable and as such can beconsidered as a green propellant. Furthermore, freezing the mixture to produce aluminum-ice (ALICE) wouldprolong the life of the propellant and thereby minimize complications due to ageing. Additionally, the propellant hasbeen proposed for use for lunar missions since there is abundance of aluminum within the lunar crust, and there isalso evidence of water.

As a final example of using nanoparticles to facilitate propulsion, in-situ generated nano-catalysts have beensuggested to facilitate the ignition of a hydrocarbon/air mixture in such short residence time burners as the scramjet67. This is accomplished by spraying into the fuel/air mixture a catalyst precursor solution diluted in a solvent suchas toluene. The spray droplets subsequently vaporize, followed by dissociation of the precursor. This releases thecatalysis in the form of atoms or clusters, which subsequently nucleate, grow in size to become catalytically active,and thereby facilitate ignition of the mixture. Figure 36 67 shows the calculated potential reduction in the ignitiontemperature of a methane/air mixture catalyzed by palladium particles.

Figure 36. Calculated ignition delay times as a function of initial temperature of a stoichiometric methane/air

mixture with and without doping of 19 nm palladium particles67

.



28/30


28

ements.org/

V. Concluding RemarksIn the present review we have discussed the research activities of aviation transportation fuels and energetic

fuels. The drivers for their developments are quite different: with the former being primarily cost and security ofsupply, supplemented by environment and climate concerns, while the latter being the interest to maximize the power delivered to a devise through increases in the energy density and propellant burning rate. The review isnecessarily cursory considering the vastness of the areas covered.

For the development of alternative aviation fuels, the strongest candidates are the Fischer-Tropsch syntheticfuels, followed by biofuels, both of which already have some rudimentary technology base and commercialactivities. Anticipating that petroleum-like fuels will continue to be the dominant aviation fuel in the long term, thattheir physical and chemical properties will continuously evolve in response to corresponding changes in thefeedstocks, and that advanced engine processes and design will also evolve as the urgency of energy security andclimate change escalates, it is suggested that predictive capability of the chemical reactivities of petroleum-basedfuel blends be developed so as to react expeditiously and prolifically to such changes. The task is a daunting oneconsidering the intricacy and enormity of the myriad of highly nonlinear and coupled chemical species and reactionsinvolved. However, it appears that the roadmap towards such a development is in place, involving detailedmechanisms, reduced mechanisms, and surrogate fuels. A concerted effort to bring closure to fuel oxidativechemistry is recommended.

For the development of energetic fuels, we have discussed the problems of particle agglomeration and sootformation in propellant formulation, emphasizing that considering energy content alone, without simultaneously

considering the combustion processes, is not adequate in assessing the performance of a propellant. We have alsodemonstrated the scientific richness of the various novel concepts in the understanding and formulation of newclasses of propellants including hypergolic and ionic liquid propellants, and propellants with strained andfunctionalized molecules and nanoparticles. These developments, being stimulated by recent advances in fuelsynthesis, materials science, and nano science, offer exciting potential for the formulation of next-generationenergetic propellants.

Acknowledgments

Activities associated in the preparation of this review was supported by AFOSR under the technical monitoringof Michael Berman, Mitat Birkan, and Julian Tishkoff, by ARO under Ralph Anthenien, by NASA under DennisStoker and Kurt Sacksteder, and by DOE under Wade Sisk through a block grant awarded to the Combustion EnergyFrontier Research Center at Princeton University. It is a pleasure to acknowledge the following colleagues who havegenerously supplied information for the preparation of this review: Bill Anderson, Med Colket, Ed Dreizin, Tim

Edwards, Fokion Egolfopoulos, Bill Green, Yiguang Ju, Tom Kreutz, Harvey Lam, Eric Larson, Steve Son, JulianTishkoff, Vigor Yang, and Rich Yetter. It is also a please to acknowledge the assistance by my students ChenglongTang, Peng Zhang and Taichang Zhang in the preparation of this manuscript.

References

1Greatest Engineering Achievements of the 20th Century, 2010. National Academy ofEngineering, http://www.greatachiev .

2"Americas Energy Future: Technology and Transformation". National Research Council, the National Academies Press,2009.

3"National Aeronautics Research and Development Plan". National Science and Technology Council, Executive Office of thePresident, February, 2010.

4"Transportations Role in Reducing Greenhouse Gas Emissions, Vol. 1: Synthesis Report; Report to the Congress". USDepartment of Transportation, April, 2010.

5

"Intergovernmental Panel on Climate Change (IPCC) (2007) Fourth Assessment Report, Climate Change". SynthesisReport, Valencia, Spain, 2007.

6"Alternative Jet Fuels, a Supplement to Chevrons Aviation Fuels Technology Review". Chevron Corp, 2006.7Daggett, D. H., O., Hendricks, R. and Walther, R. "Alternative Fuels and Their Potential Impact on Aviation". NASA/TM

2006-214365, October, 2006.8Edwards, T., "Liquid Fuels and Propellants for Aerospace Propulsion: 1903-2003,"Journal of Propulsion and Power, Vol.

19, 2003, pp. 1089-1107.9Rahmes, T. F., Kinder, J.D., Henry, R.M., Crenfeldt, G., LeDuc, G.F., and Zombanakis, G.P., Abe, Y., Lambert, D.M.,

Lewis, C., Juenger, J.A., Andac, M.G., Reilly, K.R., Holmgren, J.R., McCall, M.J. and Bozzano, A.G., "Sustainable Bio-DerivedSynthetic Paraffinic Kerosene (Bio-Spk) Jet Fuel Flights and Engine Tests Program Results", 9th AIAA Aviation Technology,Integration, and Operations Conference (ATIO), 21-23 September 2009, Hilton Head, South Carolina, AIAA-2009-7002.
http://www.greatachievements.org/http://www.greatachievements.org/


29/30


29

10Basic Research Needs for Clean and Efficient Combustion of 21st Century Transportation Fuels, 2007. DOE-Basic EnergyScience, http://www.sc.doe.gov/bes/reports/files/CTF_rpt.pdf.

11Law, C. K., Combustion Physics, Cambridge University Press, 2006.12Zhang, H. Q. and Law, C. K., "Effects of Temporally Varying Liquid-Phase Mass Diffusivity in Multicomponent Droplet

Gasification," Combustion and Flame, Vol. 153, 2008, pp. 593-602.13Glassman, I. and Yetter, R. A., Combustion, Academic Press, 2008.14Ciezki, H. K. and Adomeit, G., "Shock-Tube Investigation of Self-Ignition of n-Heptane Air Mixtures under Engine

Relevant Conditions," Combustion and Flame, Vol. 93, 1993, pp. 421-433.15Warnatz, J., "The Structure of Laminar Alkane, Alkene, and Acetylene Flames," Proceedings of the Combustion Institute,Vol. 18, 1981, pp. 369-384.

16Smith, G. P., Golden, D. M., Frenklach, M., Moriarty, N. W., Eiteneer, B., Goldenberg, M.; Bowman, C. T.; Hanson, R. K.,Song, S., Gardiner,W. C. Jr., Lissianski, V. V., Qin Z., Gri-Mech.3.0, 1999. http://www.me.berkeley.edu/gri_mech/ .

17Lu, T. F. and Law, C. K., "Toward Accommodating Realistic Fuel Chemistry in Large-Scale Computations," Progress inEnergy and Combustion Science, Vol. 35, 2009, pp. 192-215.

18Westbrook, C. K. and Dry, F. L., "Chemical Kinetics and Modeling of Combustion Processes," Proceedings of theCombustion Institute, Vol. 18, 1980, pp. 749-767.

19Curran, H. J., Gaffuri, P., Pitz, W. J. and Westbrook, C. K., "A Comprehensive Modeling Study of n-Heptane Oxidation,"Combustion and Flame, Vol. 114, 1998, pp. 149-177.

20Petway, S. V., Ismail, H., Green, W. H., Estupinan, E. G., Jusinski, L. E. and Taatjes, C. A., "Measurements and AutomatedMechanism Generation Modeling of OH Production in Photolytically Initiated Oxidation of the Neopentyl Radical,"Journal of

Physical Chemistry A, Vol. 111, 2007, pp. 3891-3900.21Hanson, R. K., "Applications of Quantitative Laser Sensors to Kinetics, Propulsion and Practical Energy Systems,"

Proceedings of the Combustion Institute, Vol. 33, 2011, pp. 1-40.22Li, J., Kazakov, A. and Dryer, F. L., "Experimental and Numerical Studies of Ethanol Decomposition Reactions," Journalof Physical Chemistry A, Vol. 108, 2004, pp. 7671-7680.

23Togbe, C., Mze-Ahmed, A. and Dagaut, P., "Kinetics of Oxidation of 2-Butanol and

dryden lecture next gen rm

Documents