AFAPL-TR-79-2122 -

CARBON SWLRR FUELS FOR VOLUNE LIMITED MISSILES

R. H. Salvesen

Exxon Research and EngineeringP.O. Box 51Linden, NJ 07036

NOVEMBER 1979 ii)A

Technical Report AFAPL-TR-79-2122Interim Report for Period 15 Septerber 1978 - 15 October 1979

Approved for public release; distribution unlimited

.a.

AEIRJ PROPULSION LABORAT~ORYAIR FORC~E WRIGHT AEROMUJ1TICAL LABORATORIES

LU- AIR FORCE SYSTEM~S C~flVNNDWRIGAI-PATFICtN AIR FORCE BASE, OH 45433

S80 5 22 035

NOTICE

When Government drawinns, specifications, or other data are used for anypuroose other than in connection with a definitely related Government pro-curement operation, the United States Government thereby incures no responsi-bility nor any obligation whatsoever; and the fact that the government mayhave formulated, furnished, or in any way supplied the said drawings,specifications, or other data, is not to be regarded by implication orotherwise as in any manner licensing the holder or any other person orcorporation, or conveying any rights or permission to manufacture, use, orsell any patented invention that may in any way be related thereto.

This report has been reviewed by the Information Office, (ASD/PAM) and isreleasable to the National TPchnical In'rrmation Service (NTIS). At NTIS,

L it wiil be avdilable to the general public, including foreign nations.

This technical report has been reviewed and is approved for publication.

~~ f

ALVIN F. BOPP ARTHUR V. CHURCHILLFuels Branch Chief, Fuels BranchFuels and Lubrication Division Fuels and Lubrication DivisionAero Pronulsion Laboratory Aero Propulsion Laboratory

FOR THE COMMANDER

Chief,Fuels and Lubrication DivisionAero Propulsion Laboratory

Copies of this report should not be returned unless return is requiredby security considerations, contractual obligations, or notice on a specificdocument.AIR FORCE/56780/]3 May 1980 - l00

SECURITY CLAS41FICATIUN OF THIS PAGE (W ri. Dai. Enterd)

REPORT DOCUMENTATl0N PAGE READ INSTRUCTIONSBEFORE COMPLETING FORM

I 2.1• GOVT ACCESSION No. I RECIPIENT'S CATALOG NUMBER

AF__FS LL_-_ -LSRT7E-O2122 . " -

i Carbon Slurry Fuels for Volume Limited Missiles, First, Annual f- ; ,'Sep V NVP 78-'OctaN79

iJ' EX(•NPLUS.1 K X 7 •

S " R. H. 'Salvesen, D. C. Riganoý W. S. -Blazo w#sk .... .. . .

W. F._ ... ...... F33615-78-C-2q29 P ERFIIORMING ORGANI1ZATION NAME AND A D:DR E.SS 10 PROGRAM LEI..M EINT PROJ EC I" T SExxon Research and Engineering Co_ AREA 6 WORK UNIT NýMFJERSM% ;-

Products Research Divi-Jbwv 304a/5g9P.O. Box 51, Linden, NJ 07036 P.E. 62203F

L. Lý,,NTRO•,.LING OF CI'L i.m "L -.NO ADC-F '_SS REPPRT DATAir Force Wright Aeronautical Laboratories .' /i Novd79AFWAL/POSF ,T. 1".,7TM"eio" ,PAGES

Wright-Patterson Air Force Base, OH 45433 - "14 MONITORING AGE4CY NAME A, AOORESS(If dcfolfr.,,f from. C'-tioiing Oaffie) IS SECURITY CLASS. ol .,I ,lh;

I Unclassified_ _.4 • •" 'IS.~i. UECLASSlFICATION DSW'INC, IADING,ISCHE'OU,_E:

16. OiSTRIBUTION STATEMAENT o( thia ReporT)

Approved for public release; distribution unlimited17 DISTRIBUTION STATEMENT (of the alhtract enterod in flck 2Q. i, J!(fternt fIorn, Report)

10. SUPPLEMENTAI"Y NO'ES

19. K E' *OROS (Conltinue on ,eversn Side It n CCe•A rv amid ident ' hL hi- #lumber)

Carbon Slurry Slurry FuelsCarbon Dispersion Carbon CombustionHigh Density Fuels Combustion of Carbon Slurry Fuels

20 ABSTRACT ;Coni1ou id. Iefern. -de- 1fth,. snd Id ,fv h,. h n -,• h.,"

The Air Force has contracted with ER&E to develop a carbon slurry fuel witha minimum of IBO,000 BTU/gal. This report provides results cf the first year'seffort of this twenty-seven month program. Initial results indicate that adispersion of carbon black in JP-lO with select dispersinc, agents ca,. be nadethat meets the BTU requirements. Preliminary results look promising. Combus-tion tests using a special'y developed Liquid Fuel Jet Stirrcd Coribtstor

(LFJSC) have demonstrated that carbon burnout efficiencies greater than qn-9 arrachiev'. '.," 300 rim. particles in residen,_e tines down to 4 Is. (coriPUED)

DD JA 73 1473 1,5 1,

Sf ý.JI:'Y Ct ASSIF IC i ;', O F T iS ',-A,.V .,- - .... LH S. eJ

z.. .i*iiII/

SECURITY CLASSIFICATION OF THIS PAGE(W7.n Data Entwrod)

20. ABSTRACT (CONTINUED)

"oomogeneous iron, lead, manganese, 'nd zirconium catalysts at concentrationsup to 1000 ppm proved ineffective as accelerators of carbon burnout. Furthertests are in progress to optimize the composition of the most promisingformulations and to test these materials under more vigorous conditions inorder to determine their suitability for missile applications.

LL

S.O.

-.' • .jZkt[-~•l 'A SF• TO • ,~ A ._•~nD l flpd

TABLE OF CONTENTS

SECTION PAGE

I INTRODUCTION 1

II BACKGROUND 5

A. Literature Review 5

1 Properties of Carbon Black and Some 5

Liquid Formulationis

2 Carbon Slurry Fuels 5

3 Combustion 7

B. Potential Problems 7

1 Formulation 7

2 Carbon Dispersion Combustion 16

C. Key Objectives 23

1 Formulation 23

2 Combustion 23

III APPROACH 26

A. Formulation 26

1 Discussion 26

2 Experimental Methods 28

a. Preparation of Formulations 28

b. Evaluation of Formulations 34

B. Carbon Dispersion Combustion 35

1 LFJSC Design 36

2 Combustion Product Sampling 41

3 Operating Procedure 43

41 Data keduction 43

C. Limited Systems Study 46

-1

TABLE OF CONTENTS (CONTD.)

SECTION PAGE

IV RESULTS 48

A. Formulation 48

I Carbon Black Selection 48

2 Selection of Dispersants 48

3 Evaluation of Preparation Techniques 54

4 Formulation Characteristics 60

L 5 Three Phase Emulsion Systems 67

B. Combustion 71

1 Liquid Fuel Jet Stirred Combustor 71Development

2 Experimental Results 73

a. Equivalence Ratio 73

b. Residence Time 75

c. Catalysts 75

d. Carbon Loading 79

e. Particle Size 79

f. Operating LFJSC on Carbon Slurry 79Alone

C. Limited Systems Study 81

1 Raige Improvement Study 81

2 Turbo-Fan Engine System Investigation 85for Carbon Dispersion Fuel Program

REFERENCE MATERIAL

APPENDIX A PROPERTIES OF CARBON BLACK (ER&E REVIEW) 88

APPENDIX B COMBUSTION OF A CARBON SLURRY FUEL 199

APPENDIX C PREPARATION OF HANSEN PLOTS 232

APPENDIX D COMPUTER DATA REDUCTION PRINT-OUTS OF LFJSC 234EXPERIMENTAL RUNS

REFERENCES 295- ii

LIST O ILLUSTRATIONS

FIGURE PAGE

1 Heat of Combustion of Formulations of Carbon 2Black in JP-lO

2 Schematic of the Role of a Surfactant in a 12Carbon Dispersion

3 Carbon Slurry Combustion Chemistry Schematic 17

4 Dependence of Carbon Particle Combustion Time 19on Particle Diameter

L Vaporization Process for a Carbon Slurry Fuel 21Droplet

6 Variability in Available Data on Carbon 22

Oxidation Rates

7 Schematic of Three Phase Carbon Dispersion/ 29Emulsion System Showing Multiple Interfaces

8 Liquid Fuel Jet Stirred Combustor 37

9 Schematic of Combustion Equipment 38

10 Fuel Nozzles 40

11 Overall View of Liquid Fucl Jet Stirred Combustor 42

12 Liquid Fuel Jet Stirred Combustor Data Output 45

13 Viscosity of Carbon Black Formulations (30 wt %) 51in JP-1O as a Function of Particle Size

14 Effect of Carbon Loading on Formulation Viscosity 61

15 Viscosity of Formulations in JP-10 vs 62Carbon Black Loading

16 Viscosity-Temperature Curves for Carbon Dispersions 63

17 Typical Flow Curves for Four Important Flow Models 64

18 Rheological Properties of Carbon Black 65Formulations

19 Effect of Equivalence Ratio on Carbon Burnout Percent 76for Statex MT 30'-; 25,. Enriched Air, T=6 Ms

20 Effect of Residence Time on Carbon Burnout Percent 77for Statex MT 30%; at an Equivalence Ratio of 0.65;Temperature of 1700°C

-iiA-

LIST OF ILLUSTRATIONS (CONTD.)F IGUR E PA G E

PAGE21 Supersonic Vehicle

8322 Subsonic Vehicle

84

iv

iI1!

V jI

LIST OF TABLES

TABLE PAGE

1 Properties of JP-1O 9

2 Formulation Problem Definition: Selection of 10Carbon Blacks

3 Formulation Problem Definition: Selection of 11Dispersing Agents

4 Formulation Problem Definition: Selection of 14Number of Phases

5 Formulation Problem Definition: Selection of 15Method of Preparation

6 Calculoted Carbon Burnout Times 24

7 Reference Liquids Used for Solubility Parameter 31

Determinatl n

8 Group Contributions to Partial Solubility Parameters 32

9 Properties of Carbon Blacks Selected for Evaluation 49

10 Peoperties of Typical Carbon Black Formulations 50Showing the Effect of Particle Size on Viscosity

11 Evaluation uf Dispersing Agent in Carbon Black 52le m•ulations

12 Solubility Parameter Data fo- Carbon Blacks 55

13 Effect of Roll Milling on Visc. of Carbon Dispersions 59

14 Static Stability of Typical Carbon Black Formulations 66

15 Test Results of Emulsification of JP-1O in 68Several Hydrophiles

16 Test Results on Emulsification of JP-1O Containing 69Surfactant A

17 Test Results on Emulsification of JP-10 Containing 7030' Statex MT and Surfactant A

18 lest Results on Emulsification of Statex MT Dispersed 72in JP-lO Using Formamide as the Hydrophile

19 Carbon Burnout Percent as a Function of Slurry Flow 74Rate at an Equivalence Ratio of 0.8, -= 6 ms

vI

LIST OF TABLES (CONTD)

TABLE PAGE

20 Carbon Burnout and Combustion Efficiencies as a 74Function of Equivalence Ratios for 25. Enriched Air,T= 6 ms

21 Carbon Burnout Percent as a Function of R-ýsidence 77Time at an Equivalence Ratio of 0.8

22 Effect of Homogeneous Catalysts on Carbon Burnout 78 aPercent

23 Effect of Carbon Loading on Carbon Burnout 80

24 Effect of Particle Size on Carbon Burnout Percent 80 j25 Results of Operating LFJSC on 30,. Statex M7 80

Alone

26 Proposed Criteria for Carbon Dis,_rsion Formulations P2

I

- vi -

GLOSSARY OF TERMINOLOGY

Aggregate - The smallest unit of carbon black composed of fuxed carbonparticles.

AUlomerate - A loosely bound group of carbon black aggregates.

Catalyst - A substance which increases the rate of a chemical reactionwithout itself being chemically changed.

Carbon Burnout Efficiency - A quantitative measure of fuel carbon utilization.

Cohesive Energy Density (CEO) - A measure of intermolecular forces expressedas interaction Inrergy 2per unit vol lie.(CED = 6D + 6p + 6H

Cohesive Energy Ratio (CER) CER = CED Phase 1)CEO (Phase s 7

Corbustion Efficiency - A quantitative measure o.) fuel energy utilization.

- London (dispersion) forces.

S- Hydrogen bonding forces (also electron transfer).

- Keesom (Polar) forces.

Equivalence Ratio - The ratio of the experimental fuel to oxidant conditionsto the stoichiometric fuel to oxidant conditions.

Heterogeneous Catalyst - Catalyst which is added directly to the carbonparticles.

HLB - Hydrophilic Lipophilic Balance.

Homioeneous Catalyst - Catalyst which is added to the liquid portion ofthe carbon slurry fuel.

Hydrophilic - water loving.

Li pop hi-lic - o il loving .

vii

• .- • ... . . . . .

(CONTD.

Mach Numb.,r The ratio of the speed of a body to the soeed of sound inthe surrounding atmosphere,

Residence Time - The reaction time for th-, fuel and oxidant within thecombustor chamber.

Stoichiometric - The amouit of air or oxygen requir,ýU' te completelycombust the fuel carbon and hydrogen to carbon dioxide and water.

mi, Millimicron, 10- 9 meter (nanometer).

viii

LIST OF SYMBOLS, ABBREVIAT!ONS ETC.

0

A Angstrom

Btu British Thermal Unit

CED Cohesive Energy Density

6 Solubility Parameter (6 = Greek delta)

London (dispersion) forces

6H Hydrogen bonding forces

Keesom (Polar) forces

ER&E Exxon Research & Engineering Company, Linden, NJ

F Dispersion volume, cc/gm

H Heat of VaporizationV

K Kelvin

2n /g meters square per gram

ms millisecond (10-3sec.)

-9mp inillimicron (10- meter)

N Combustion efficiencyC

NDIR Non-dispersive infrared

Ncb Carbon burnout efficiency

No Viscosity

0 Volume Fraction of the internal phase of a dispersion or emulsion

Equivalence Ratio (combustion related)

PPM Part per million

PSI Pounds per square inch

R Gas Constant

T Temperature 'K

t Time

•i Residence Time

TIC Total Hydrocarbon

ix -

(CONTD.) 7a

u micron (10-6 meter)

urm micrometer (106 meter)

AV Molar Volume

-t

y

t2

-14

SUMMARY

This report is an accounting of the first year's effortby Exxon Research and Engineering Company (ER&E), under contractto the Air Force Aero Propulsion Laboratory, to develop a highdensity fuel for future volume limited cruise missile systems. Pastefforts sponsored by the Air Force have resulted in development ofliquid hydrocarbon fuels with up to 160,000 Btu/gallon. Increases -Ain Btu content are desirable to further improve missile range. As alogical next step in high energy liquid fuels the Air Force and otherADOD agencies have sponsored research in carbon/liquid fuel mixtures.Blends of carbon black with high density fuels can provide increasesto 180,000 Btu/gallon and above. However, the mixtures prepared to dateare slurries which do not meet the performance requirements and havenot been adequately tested. This current program is aimed at prepara-tion of a carbon loaded liquid fuel with a minimum of 180,000 Btu/gallonkthat will perform well in a non-exotic fuel system.

The ER&E program to accomplish the above objective is aniterative approach which recognizes the need to consider the oftenopposing requirements of fuel system and combustion system performance.An extensive search of the literature was carried out to investigatethe properties of carbon black, methods of formulating carbon dispersionsin liquid fuels, and combustion characteristics of small carbon particles.While this search revealed a wealth of information on these subjectsthere was very little guidance provided on specific dispersing agentsto provide a stable system with desirable properties. In addition,while the literature on combustion of carbon particles indicates arelationship between carbon particle size and combustion rate, thedata showed such a wide range of test conditions and results that con-siderable uncertainty remains as to what might be expected from combus-tion of carbon black dispersions in fuel. The major problems to beovercome in formulation appeared to be selection of the carbon black,choice of the proper dispersant or dispersants, and method of mixingthese components in a high density fuel, such as JP-l0. The mostformidable problems of combustion appeared to be identification of con-ditions for maximum combustion efficicncy, comprehension of the combustionprocess, determination of the carbon burnout rate, selection of the optimumfuel injection system, studies of fuel volatilization in the combustionchamber, and development of data to enable design of a useful combustionsysten for missile engines.

The approach utilized for formulation was a combination ofconventional and novel techniques. The conventional approach was Loutilize experience from others who have formulated carbon black dis-persions in oil, sucn as for ink oils and rubber processing. Thesetechniques were used. In addition, several novel approaches wereevaluated. The first was to utilize the solubility parameter conceptwhich provides systemal c information on physical interactions ofmaterials. This concept is useful in selecting the optimum dispersantfor a material such as carbon black in a liquid. The second novel con-cept was to formulate a three phase carbon dispersion system such thatthe droplets of carbon black in JP-1O (Phases 1 and 2) would then beemulsified in an external immiscible liquid.

xi -

To study the combustion properties of carbon dispersions theER&L approach was to devise a jet stirred combustor for liquid fuels.Considerable experie-,e had been gained utilizing well-stirred combus-tors such as thp jet stirred combustors for gaseous and vaporized liquidfuels. However, this would not have been appropriate for carbon dispersions.With a liquid fuel jet stirred combustor it is possible to study combus-tion parameters of carbon dispersions such as overall combustion efficiency,carbon burnout efficiency, effect of various equivalence ratios, rate ofcombustion, effect of catalysts on combustion, and hardware required to

k pump and spray these systems.

In addition to studying preparation and properties of carbondispersions, the anticipated ano measured properties were subjected toa Limited Systems Study with the assistance of three subcontractors whoprovided major inputs on turbine and ramjet engines (GE and CSD, respectively)and other missile system component (Boeing) requirements. This study in-cluded consideration of (I) range improvement achievable with carbon dis-persions, (2) fuel characterization needs and (3) establishment of fuelevaluation criteria.

Initial results have demonstrated that carbon slurry fuels canbe formulated with properties approaching minirnum requirements. Twophase dispersions containing up to 69 wt " carbon in JP-10 have beenprepared using 2-5'. surfactant. Ultimate particle size carbon blackvarying from 13-300 millimicrons (mi,) have been investigated. The vis-cosity of a 60 wt % 300 mp carbon dispersion in JP-IO was l63 cp at-14 0 C. The fuels showed no signs of phase separation stability problemsafter 20 weeks of shelf life at room temperature.

Studies with two phase dispersions have shown that viscosityfor a given loading increases with decreasinq particle size in a complexmanner. Viscosity remains nearly constant up to a critical miniiiumparticle si1ze and then incrpases sharply with smnaller particle sizes.For a given particle siz.- viscosity increases exponentially above about50 wt % carbon loading. The choice of particle size requires a trade-off.Combustion rate is enhancEd by reducing particle size but conversely for-mulation viscosity increases rapidly. Thus, with available carbon blacks,the results indicate that increased attention be given to the use of 300size particles.

Choice of surfactants has been shown to be critical and signif-icantly affects the properties of carbon black dispersions. Limitedwork has also been carried out formulating three phase emulsion fuelswhich offer the promise of irrproved performance relative to two phasedispersions. To date, a fuel containing 48 wt . carbon has beern oreparedconsisting of 80% of ar, internal phase containing a 60' carbon two phasndispersion in JP-10 and 20% of an external phase consisting of formamidea,,d surfactants. Futid.:mental Cohesive Energy Density qeasurevients havebeoi completed which are expected to help future formulatior, efforts.The data on two phase systems look very promisinn and thus further workon three phase systems will be terminated.

Combustion measurements were carried out in a newlv desionedlaboratory well-stirred reactor capable of burning both liauid and'multi-phase carbon sl,,rry fuel. Previously, well-stirred combustors weredesigned only to operate with gaseous fuel. The Liquid Fuel Jet Stirred

- xii -

Combustor (LFJSC) incorporates both a recirculation region typicalof the primary zone and a turbulent non-recirculating zone typical ofthe secondary and dilution zones of a gas turbine. Development of thisnew fundamental tool will allow full chardcterization of the combustionproperties of experimental carbon slurry fuels to be obtained withsmall sample sizes.

Experience gained from the shakedown operation of the LFJSChas shown that fuel injector design for carbon slurry fuels will bea problem and clearly indicates a need for increased attention to thisarea in the future. Two critical design problems have been identified:(1) prevaporization of the liquid portion of the slurry fuel must beavoided in the fuel lines and nozzles, and (2) che injector must becarefully designed to avoid carbon deposition within the combustionsystem. Tests in the LFJSC indicated that carbon burnout efficienciesof greater than 90'% are achievable with 300 mnj particles in residencetimes down to 4 ms. Little difference in combustion performance wasnoted using a range of dispersions containing from 30-50 wt % of carbon.Smallcr particle size (75 mu) carbon formulations gave higher combustionefficiencies (' 3/s) than those made with larger carbon particles (300 mo).The data indicate that burnout is a function of mixture conditions and thatthe optimum equivalence ratio is approximately 0.85. Since conventionalcombustors do not provide substantial time at this mixture condition,combustor redesign to accommodate carbon slurry fuels is probable. Limitedtests indicate that smaller carbon particle size may result in improvedcarbon burnout.

Tests were made to determine if homogeneous catalyst, i.e. theaddition of a soluble catalytic material to the liquid portion of thecarbon slurry fuel, would accelerate the burnout of the carbon particles.Results to date do not show any improvement with manganese, iron, leador zirconium based homogeneous catalysts at concentrations up to 1000 ppm.We have concluded that homogeneous catalysts will not significantlyaccelerate carbon particle burnout, and that it is important thatwe proceed to investigate the use of heterogeneous catalysts for thispurpose. In the heterogeneous approach, small levels of active solidcatalyst particles will be added directly to the surface of the carbonparticles before the carbon is added to the JP-10 to form the slurryfuel.

The Limited Systems Studies have indicated turbine enginescan be expected to benefit more from carbon dispersion fuels (25-35Yimprovement in range) than ramjets (about 15 improvement in range).Utilizing carbon dispersion fuels with the properties reported herein,the engine manufacturers (GE) do not anticipate any major obstacles totheir utilization. Some changes in engine fuel line size, meteringsystems, nozzles, etc. will be needed to accornmodate these fuels. Furtherstudies on actual engines will identify specific requirements.

Although our initial results are encouraging in that theyindicate highly loaded fuels with moderate viscosities can be prepared,the resuILs also confirm our general view that the formulation and combus-tion of carbon slurry fuels is q.luite comulex.

- xiii -

I

Part II of this report gives classified information generatedduring the first year of this program. The classified data given is:

1. Identification of surfactants which are assigned codes -4for use in the unclassified portion (Part 1).

2. Details of the CED values which identify specific surfac-4tant requirements of the carbon blacks tested and potentialoptimum dispersant materials.

xiv

ji

I

~1

- xiv -

SECTION I

INTRODUCTION

Improving the range and payload of modern and future cruisemissile systems will require the development of fuels with increaseovolumetric energy content. Air Force and Navy-sponsored researchhas resulted in impressive advances in fuel Btu content over thepast 20-30 years Liquid fuels such as RJ-5, RJ-6 and exo-tetrahydrodi(cyclopentaaiene), now called JP-I0, have been developed (1). Furtherimprovements have been made in these bridged-ring saturated hydrocarbonsand a series of materials are currently available (2). These fuels allowenergy contents up to 160,000 Btu/gal while requiring minimum alteration0 ' the overall missile system.

It now appears that a practical limit is being approached forentirely liquid fuels and further improvements in range through inno-vation in fuel composition will require the use of slurry fuels.*Consequently, the Air Force initiated a program with Exxon Research andEngineering Company (ER&E) in late 1978 to develop a solid-carbon-con-taining fuel which will have a volumetric energy content of 180,000 Btu/gal (3). This report is an accounting of the activities which have takenplace during the first year of the program.

The task of formulating a carbon slurry fuel which can meetthe energy content requirement and successfully perform in turbine orramjet-propelled missile systems is extremely complex. As illustratedin Figure 1, if JP-1O were used as the hydrocarbon carrier, the requiredcarbon mass loading woula be 60%. Such a fuel can be expected to havemany unusual characteristics, but the most significant challenges of thefuel development program are to:

formulate a fuel which is stable and of sufficiently lowviscosity to perform well in a nor-exotic fuel system,

provide acceptable low temperature characteristics,

assure that the energy content of the fuel carbon is uti..lized within times which are realistic for cruise missilecombustion systems, and

establish the impact of unusual combustion characteristicson optimum combustor design.

This program addresses each of these issues. lhe approach beingused is iterative in nature and recognizes the key trade-offs between Btu/gallon, fuel system performance, and combustion system performance. Programconsiderations can be organized into three interactive components: fuelformulation, combustion, and systems implications.

* Use of the term slurry to describe the type of fuel needed is notaccurate. A slurry is usually thought of as an unstable (or onlytemporarily stable) mixture of solid and liquid phases. In the missilefuel case, long term stability is required, and the fuel would moreappropriately be called a carbon dispersion. Nevertheless, since theterm slurry is so widely used in the missile dewv1.IpImernt (comunity,we will use the terms slurry and dispersion interch.-,q,(,hly.

-.1 l

210

190

0

4-, 180

1 170

160

50 60 70 80

Wt % Carbon Black

Figure 1. Heat of Combustion of Formulations of Carbon Bflack in jP-10

-2-

Fuel formulation requires detailed consideration of the surface- chemistry between dispersed carbon and the liquid hydrocarbon carrier.

Interactions of carbon, high density fuels, dispersants, and otheringredients can be related and predicted by measuring a property calledthe Cohesive Energy Density (CED) of each component (4,5,6). The CED is ameasure of the intermolecular forces. The CED may be expressed as a functionof the solubility parameters or compatibility of substances with a seriesof pure liquids. These parameters can readily be measured or calculatedand a value assigned to each ingredient. The relationships are based onthe first law of thermodynamics and will be described more fully in thetext of this report.

While this CED information can be applied to developing two-phase(i.e. carbon in hydrocarbon) dispersions, it can also be applied to theformation of three-phase systems. In this latter case the carbon dis-persed in hydrocarbon is the internal phase of an emulsion; the externalor encapsuiating phase is a liquid (such as formamide) which dominates the

k emulsion flow properties. Such a system has the potential for lower viscos-ity, excellent low temperature properties and increased probability ofachieving secondary droplet atomization vi~a the use of a higher boiling

external phase which could cause the droplets to rupture and "pop". Themain disadvantage to this approach is that the third phase is likely to havemuch less of a rontribution to energy content than an eFn,'.ua, volume of JP-lOand, hence, ` lutes the energy density of the system. In terms of Figure 1,the curve shiis to the right and more carbon loading is required to achieve agiven Btu/gallon.

The combustion portion of this program is concurrent with formula-

tion efforts. Burnout of the carbon within the slurry is expected to lag

the consumption of the fuel hydrocarbon components. Consequently, it is

essential that the influence of carbon type, particle size, and combustion

conditions be evaluated early in the program. A problem of particular con-

cern is the trade-off between increased fuel viscosity (negative effect)

and reduced carbon burnout time (positive effect) as carbon particle size

is reduced. The correct selection of carbon black depends upon the combus-

tion information in conjunction with the results of the formulation study.

Further, the need for and abi-lity to catalyze the burnout of carbon must

be established early in the program; the inclusion ot a catalyst would sub-

stantially influence the task of formulation.

A new small-scale laboratory combustor, the Liquid Fue, Jet

Stirred Combustor (LFJSC) has been developed for these combustion studies.

This device produces results which represent thp situation within a turbine

or ramjet engine but allows testing with only small (about one liter)

quantities of fuel. This allows testing of many more carbon types and

formulations than would be possible with a large-scale system.

The final component of the program concerns systems implications.

Subcontractors are inputting information and advice at key points in our

program in the areas of turbine and ranijet engine technology as well as

missile airframe considerations. General Electric is our subcontractor

to provide expertise in fuel systems and combustors associated with

turbine engines. The Chemical Systems Division of United Technologies

provides guidance and key inputs regarding fuel systems and combustors

3-

utilized in ramjets. Finally, Boeing provides knowledge regarding themissile system fuel requirements and net benefits of carbon slurries.

As already mentioned, this is an interim report which concernsonly the results of the first program year. During this time the fol-lowing milestones were met:

Formulation and evaluation of two-phase dispersions withloadings up to 69% carbon.

Development of a small-scale combustion device to studycarbon burnout and combustion characteristics.

Initial evaluation of carbon burnout and the effects ofparticle size, catalyst use, and combustion conditions(fuel-air mixture ratio, temperature and residence time).

Formulation of three-phase systems (emulsions) with 60:;carbon in the internal phase (80%) and formamide as theexternal phase (20.).

CED characterization of all carbon dispersion components.

The balance of the report is organized into four additionalsections which more fully describe these accomplishments. The firstprovides background information which helps define the challenqe ofdeveloping a carbon slurry fuel. Additional background informationin the form of a thorough literature search is included in the reportappendix. This is followed by sections describing the approach beingused and the results acquired. The results section provides our keyresults; detailed listing of all information generated has been providedin the appendices.

-4-

SECTION II

BAC KGROUND

This section provides details of our literature review, adiscussion of the potential problems anticipated, and the key objectivesof our approach to solving these problems.

A. Literature Review

An extensive review of the pertinent literature was made con-centrating our search mainly on published information dating from 1960to the present. The reason for selecting this time frame is that thereare a number of excellent reviews which already adequately cover thesubjects of interest as noted below. We have elected to put the bulkof the results of our literature survey in the appendix and have includedbelow a brief summary of the various major topics.

1 Properties of Carbon Black andSome Liquid Formulations

An annotated bibliography comprising a comprehensive sampleof the extensive literature and numerous patents on carbon black hasbeen prepared and is provided in Appendix A. Only samples have beenprovided which cover pertinent references since it would be repetitiveand too lengthy to include all the materials. Included is informationon carbon black dispersed in liquids used as pigments, coatings andink oils. No attempt was made to cover the very extensive literatureon the use of carbon black in elastomers and plastics. Over 350references are included in this review out of thousands searched. Abrief abstract is given below and further details are provided inAppendix A.

Some physical and chemical properties are notable. Particlesof carbon black are extremely small, ranging from 5 to 500 nanometers,and can be seen only with the electron microscope. Their shape variesfrom spheroidal to complex; and the particles tend to clump together;indeed, dispersion of the particles is a major problem in the use ofcarbon black. Surface area is typically in the neighborhood of 100 rn2 /g,with extremes from perhaps 5 to 950 m2 /g, and 0 to 35" porosity. Thesurface is usually active with a tendency to react with oxygen and someother substances to form surface groups that influence wettability andpH, which varies from acid to alkaline. There is also some free surfaceenergy that draws the particles together. The rheology of suspensionsof carbon black is strongly influenced by these surface characteristicsand the suspensions are usually non-Newtonian. Viscosity,' for examiiple,is a function of shear history. The technology oF carbon blacks incoating and inks entails th2 adaptation of the peculiarites of carbonblack to the preparation of the stable suspensions required in theseindustries.

2 Carbon Slurry Fuels

Mixtures of carbon in hydrocarbon and other types of oils havebeen made for centuries and used mainly as inks and pi,igments. Theliterature search noted in the previous section covers t,'pical formulations

.1

of this nature. Carbon slurry fuels Qf interest Lo the aerospace industryare specifically high energy formulations that ,qould generally contain over50-60% of carbon black in a high energy density fuel such as JP-9 or JP-lO.Such formulations are called slurries because they gencrally are not stableand the mixture must be kept agitated to assure a unifoem dispersion. So:Iremixtures of high carbon content are used, for example, as components of,lastomer masterbatches in which the carbon black is dispersed in a suitableplasticizer (7). While the properties of ýuch materials are of some irterestthe objectives are different and stability is not a prime concern.

Preparation of high carbon content high energy fuels has beenlimited mainly to the military. Extensive work has been done since the mid1940's to develop high powered fuels mainly fo;" rockets and ran'Jets. Carbonslurries and slurries of metals such as boron and alurinum were qenerally con-sidered for these purposes, but were usually not preferred since the proble,',sof handling and combustion deposits could not be overcome.

jBryant and BurdeTte (8) fo-muleted tip to 70 wt ca-'-;on blac0 in

high density hydrocarbons such as dec'ýlin and tetralin and also in conventional,kerosene type hydrocarbons. These wo s e\al ur.t..•,i ah~out riqht differentcarbon uiacks and graphite in kerosenu. Several nellinq vai.ents (Shell ilev-r .ment Co's AlMB2, a Dow experimental agent CX3487, and Gellant M4, diacidaluminum soap of isooctanoic acid conforming4 to '1L-T4-C009A) were evaluate:!and the DOW CX 3487 appeared to show the mrist p-'i e. I, addition, a na;1,.of formulations without gellinq agents were used for cmbustion screenil.etests. Tests on viscosity and storage stab it -,,re r,,. Tlcr res.;I+,indicated the viscosities were probably too hlk!ýh for- pfrct ut- itmissile systems, but the stability loukcd proihinij ifter sc\rl ,nuo ,sof static and temperature cycling tests.

An excellent review oF this subb ect e,'!t itl uid "Jiti,, u •l ';riyFuels - Literature Survey" has recently Leer' :,•,,,ilej b, - .ii: ,M~o ,h il i ty Eq u i pm en t Re se a r c h an d D e ve lo pi••e rn t ) , ] -, , rt -' .,, , ,Virginia (Repoý,t 2221, September 1977). TP i,. ", r \:er. . . j I:;..at searching for fuels to be used in grounJ ,•j vt,,ie, aid-" ciciuU;e0slurries were not of interest. They proferrt-d syntr, -:i, :..,-v, osliuwid hydrocarbons of high density com -r0! t• siL--V I-t,

A comprehensive literature search of [ar'bon lrrc a,-.(; Gelsfor the period from 1960 to October 1969 wa. prep,,-tý'; 5y the f efenseDocumentation Center at the request Of Wil Iiati B,,rdel.t•, l", the ... ,•av lWeapons Center, China Lake, California. This report 'it;hio(.rahy is,lassified Secret and is identified by the Fdart.h Lotn,.'K No. -:.0311 .

Whil, the previous work contains, hutidt,-cs ,, ,ddi t v.2tt be useful fr dispersion of carbon blac:ks thet,-' * y a Y ,i -dates identified for use in the current prog•!;a., bud i0 .•u•,J tle arimmense task to evaluate all the others.

Ashland has recently con-pleted a piaper Jtuoy a-, a sLi, Orti.i(tjrto McDonnell Douglas Astronautics Coopany'-[artfamjet fuels research in support of an Air lor(i, i•,ta f 1,'A(r -entitled "Supersonic Long Pannie Missile 1Ir, g te, >i ..' ','. . (9).

6-

The Ashland work indicated that formulations containing up to 70 wt 1of carbon black could be dispersed in RJ-5 and up to 75 wt Z in RJ-4.The dispersant used was ?% of TLA 202. (This is a Texaco PetrochemicalCompany ashless dispersant for automotive and diesel lubricants producedby reaction of P2 S5 with polyisobutyiene.) The carbon blacks evaluatedwere standdrd rubber grade medium thermal blacks from Ashland and Cabot.[ormulations were prepared on a laboratory three roll mill commonly usedto make irk oils. Properties of these materials are described in theabove reference. Viscosities varied from very thick to something similarto a multigrade motor oil. No information was provided on the stabilityof these materials.

Previous Exxon work (10) covered preparation of both two phasecarbon dispersions in JP-4 and TH dimer fuel (Tetrahyd-ocyclopentadienedimer), and three phase emulsions of these materials. Carbon blacksused were of 0.25 micron average particle size, generally coated withselected surfactants, dispersed in the fuel, and then emulsified in ahydrophilic liquid. Concentrations containing up to 60 wt ' carbonblack are reported. These materials were claimed to be quite stableand useful in air breathing rockets. High carbon conccntration formula-tions (2 phase system) were quite viscous and had a reduced tendencyto flow under low shearing forces. In addition, the method of manufacturesuggested for the high concentration formulations was slow and tedious.

3 Combustion

The carbon slurry combustion process involves volatilization,Of carbon slurry droplet,; into hydrocarbon vapor and carbon particles,

combustion of the hydrocarbon vapor, and combustion of the solid carbon.

Each of these processes is reviewed in Section 2 and Appendix B (in detail).Unfortunately, direct combustion experience with actual slurries is lackingand our current understanding is compri-,ed of educated conjecture concerningeach of the involved processes.

The primary difficulty with carbon slurry combustion concernsoxidation of the solid carbon to achieve complete energy release. Ourevaluation of previous work concerns soot oxidation and carbon particles.Of primary importance are References 11-19. As will be discussed later,this previous work does not lead to a distinct conclusion regarding therequired combustion time for the carbon particle and much of the Exxoneffort during this first program year concerns the obtaining of this in-formation at conditions relevant to the carbon slurry application.

B. Potential P;-oblems

1 Formulation

Inherently carbon black is insoluble in hydrocarbon liquids.Thus, to prepare a formulation of carbon black in a hydrocarbon fuelthe resulting mixture must consist of at least two separate phases;the hydrocarbon fuel which is a liquid, and solid carbon black. (Forthe purposes of the initial phases of this effort JP-l0 has been selected

7

as the high density liquid hydrocarbon fuel. The properties of JP-1Oare given in Table 1.) Faced with the reality of such a system themajor problems that need to be considered and the trade-off optionsare summarized below.

Carbon Black

Selection of the proper carbon black must take into considera-tion the size and structure of various commercially available materials.The benefits and problems of each option are summarized in Table 2.Carbon black size as given in the literature may be somewhat misleadingsince it only gives a portion of the picture. The actual particle sizegiven denotes the size of the individual particles which compose a chainor grouping of carbon particles. For example, as shown in Table 2, thesize would be a measure of one of the circles shown but these are actuallyfused to other particles to form an aggregate. This aggregate comprisesthe smallest size particle, and the dimensions of this aggregate are notgenerally given but can be determined by microscopic techniques. Carbonstructure is a measure of how these chains are formed, and may be shownas linear or compacted aggregates as noted in Table 2. In addition,these aggregates can link together physically to form loosely boundagglomerates. The significance of the different sizes and formations inthis program is the effect which size has on combustion rate as well asformulation properties as noted in Table 2. In addition, use of differentamounts of shearing force during preparation can cause breakup of theloosely bound agglomerates. No amount of shear, however, will degrade theaggregate which is composed of fused carbon particles. Thus, selectionof particle size and structure is critical to providing the optimum formula-tion and combustion properties.

Dispersing Agent

Selection of a suitable surfactant to disperse the carbon blackin JP-lO is a significant problem since there are thousands to choosefrom with different properties. Table 3 shows briefly some of thebenefits and problems of selecting the proper material. A surfactantserves to link together two dissimilar materials. An illustration ofthe role played by a surfactant in a carbon dispersion is shown schemat-ically in Figure 2. Surfactants may be made to have one end be hydro-philic (water loving) and the other end lipophilic (oil loving). In acarbon dispersion, the hydrophilic end (head) will be attracted to thecarbon particles and the lipophilic end (tail) attracted to the JP-IO.The strength of these bonding forces will vary with the chemical structureof the various components. To select the proper surfactant the bestmatch must be made among thousands of available surfactants with differentfunctional groups, chemical structure, and molecular weight. Often amixture of two or more surfactants is required to satisfy these require-ments. Choice of the optimum surfactant will provide the most stablesystem. Making the proper choice is part art and part science. Amajor problem is the carbon black is not sufficiently characterizedto permit selection of the optimum surfactant by known methods of calcu-lation. There are several methods such as the HLB (Hydrophilic/L.ipophilicBalance) method (20) and Solubility Parameter approach (4,5,6) which can

j -8-

ii i

• •, • .... .

TABLE 1

PROPERTIES OF JP-10 (MIL-P-87107A, USAF)

Designation Propellant, High Density Synthetic Hydrocarbon TypeGrade JP-lO

Composition I00ý Exo-tetrahydrodi (cyclopentadiene)

Chemical Structure

'4

Chemical and Physical Requirements and Test Methods

Property Physical Requirements ASTM Standards

Minimum Maximum

Color, Saybolt +25 - D-156Chemical AnalysisExo tetrahydrodi 98.5 100 (1)

(cyclopentadiene)Other Hydrocarbons - 1.5Iron, ppm - 10Flash Point, 0 C (IF) 52 (125) D-93, D-3243Specific Gravity, 0.935 0.943 D-1298

(60/60*F)

Freezing Point, 0 C(0 F) - -79 (-110) D-2386 (2)Viscosity, cSt at D-445

°C (OF)-54 (-65) 40-18 (0) 10

Net Heat of Combus- D-240, D-2382tion

MJ/k 9 (Btu/lb) 42.1 (18,100)MJ/m3 (Btu/gallon) 39,400 (141 ,500)

Thermal Stability D-3241 (1)Change in press. drop,

mHg 10Heater Tube Deposit Code 2visual rating

Existent gum, mg/L 50 D-381Particulate matter, 1.0 D-2276

mq / LFuel System Icing 0.15 (1)Inhibitor, Vol %,

)--S--ee above Mil Specification for details.(2) Waived for JP-0.

-9-

Lo

<j]4

C) (D iLL - I- -

C- 0 C) --(-I - - cJo- .-o0" 0 - 0 -•a

Lii -- ) .- ( L-o C Lii

CO LL V I - 0) cz V)

&-- Z U

"I-

V)- - CLCLL + -++ +

cc•I

Lin ~ j II

I)c

*~:2: -- 10 -

cc, V) -- V

__M _. _i _ :) _ _ _ _4

C-D

-J z ~>- CLU (A -

S--, '

-,J ,."4" •

c o I o C D ~V ct"Y

L- u j • ) .1.

Lcm C W=, CO V1t/) •1V') A

1+ + +(AA

-- ' -

LU -- U LU C -

U*) w. L/i

F- - C ) V)- w

Z: ui

I- ~ -J ~ CCL

C-)0 ~ CC 0 LU - C- L U -~

LU U C t- L" -:) 7: 0

wLU c.4r

cc

Li.

CD CD

0 .r (A-

.1

JP-IO

I.pJP-IO

• • ~~L i pu, ph i 1,.

Tail

CarbonParti cl e

I iqure 2. Schematic O~f The Role Of A 'urfdctJF't ',In Crr )4(I

- 12 - -

be utilized to obtain approximate requirements. These will be discussed

later (Section III A la). A

Number of Phases A

Two phase systems are the conventional method of preparing adispersion of a solid in a liquid. The problems with these systems basedon past experiences are poor stability and limited flexibility in proper-ties as noted in Table 4. Three phase systems are another apprcach (seeSection ill A 1 below) which has certain advantages, but they requirean additional emulsion phase which introduces problems in selection of Athe proper external phase liquid and emulsifying agents. The threephase system contains the two phase dispersion of carbon in JP-1O asthe internal phase. The external phase is an immiscible liquid inwhich the carbon dispersion must be emulsified. The potential benefitsfor a three phase system, as noted in Table 4, are choice of a threephase which has better low temperature, viscosity and stability charac-teristics than a two phase system. In addition, the third phase may beselected with a "pop" effect which on combustion might "blow apart" theparticles to facilitate combustion and reduce agglomeration. Selectionof the proper combination of components is obviously not a simple matter,but such a three phase system could have significant advantages.

Method of Preparation

Mixed phase systems are generally prepared with agitation,but the amount of agitation and shear required presents problemswhich can significantly affect product quality and costs. Severaldifferent types of mixers and the benefits and problems of each aregiven in Table 5. Order of m~xing components can also be a criticalfactor in combining solids, liquids and one or more surfactants.

Need for Catalysts

A major challenge in the development of a successful carbonslurry fuel is to achieve complete combustion of the carbon particles inrealistic combustion residence time,. This problem is discussed fromthe combustion point of view in Section II B 2. At a fixed temperatureand equivalence ratio, smaller carbon particles will combust more rapidly,but tend to increase fuel viscosity in a two phase carbon dispersion.Clearly formulation techniques to accelerate the rate of carbon particleburnout will be highly desirable if not critical.

In the surface kinetically controlled regime, the rate ofoxidation of the carbon particles should be capable of being acceleratedvia the use of catalysts. In principle, such catalysis could be eitherhomogeneous, i.e. the addition of a soluble catalytic material to theliquid portion of the fuel, or heterogeneous, i.e. the addition ofactive solid catalytic materials directly onto the carbon particlesthemselves. Homogeneous catalysts would be expected to accelerate therate of carbon particle oxidation via the generation of enhanced concen-trations of active species such as oxygen atoms in the vapor phase whichthen must migrate via diffusion and/or convection to the surface of the

13-

.; • • i • - •i• '•-- • u w • . . .

_ _ ~ U _ ---

iLLJ

Lj .L

-I-

L..-.J

J L ( • =')

w -o Li

<0 CD

i,.i

CD U- LnL-

w A I L-;x U)

I - i- -j cc ) C).

< <~ -j=V l l

cvi C) CD J f

Jl La- U. >< u >- Lnc) < = LI -- i 0 o c. .

VA 0 1 -1 LA- (D 0i C) --C) L

F- ~ La'g> La- >- (A. C= -IJ L- JF

- MJ Li. :;. _j V) <.. Lu C

I L 0- C I Mu. :3: LUE c _

2f C II I-- V)

LUJ LA-I F- C) M C) (AL A LU ...J (U -(Uc L- gLi M; La-i Lii LLi uji (1J

Cl L-.:

LUI >- = C-) (A =) Q. - L LE L (u

LLJ

LLJJ

,_JV)

V) = C)

oi oN

M LL

' 14 -

-41

F- v C -'4J Ln P- - ..0-<.1 C) I-

LUJ m VI l) ~ ) (4 LUJ A-j LA ton a. LJ a Aca~ ui L CD >- ý F- V) m:. ca

= u -i - Cl) CD 0 wLn I- LU '

m) (. LIi 0xf-C> = LU J C) LL. C

- ) iD

0 LL. +

a. U

Liuj C) LUI Li

LiD CDin LU- V- C)-j U

F -CLU: ) :Lii LL (4 LU 0. ) LU 0 LLi -4

CD I-- C- e

o F- 0 C) cx C:)__j .4) c- LL.~ -(4

C) + + + + +++F-

::r LLiiyH- LUju oc

LU LU t

-J-

Ui-cIZ U- C)

I- -

I'i a_ LUI

C) Cr

- 5

particles and attack the carbon. Heterogeneous catalysts would beexpected to function via dissociation of molecular oxygen to atomicoxygen on the surface of the catalyst particles which then undergo"spillover" and readily attacks the nearby carbon.

Studies involving direct observations with a transwiss ionelectron microscope equipped with a television camera by hr. 1. 1. K.Baker of Exxon Research has shown that metal particles added to thesurface of carbon will, indeed, catalyze the rate of oxidation. At asufficiently high temperature the metal particles become mobile and"dig" both channels along the carbon surface and holes into the bulk ofthe carbon particle. Metal and/cr metal oxide catalysts ar(, known to heactive for oxidation reactio s arid the formulation pro(nc'oures should aimat depositing small (e.g. 50A), well dispersed parti(les urnto thL surfaceof the carbon so that significant catalysis is achievwd at low caLalystloading levels.

2 Carbcn Dispersion Combustion

The subject of carbon dispersion combustion is ex-eedirilycomplex. Superimposed upon the already intricate pructs.(,. inv(vlw:(din the conmbustion of gaseous or liquid fuels is the het;roqere,,roxidation of carbon particles which nay he large corln;arf'd to Ihw normalsoot formed during conventional fuel combustion. The [,lirn tfill, of thespL-arbon particles is expected to lag the consumptior, of Ithu:•.4oLl;

compDonent in the fuel but the impact of tis occurrence on piAr;vieters sucmlas ignition, stability, and combustiorn efri(.irei, y i, ,, iw .,•..r I,background section provides a brief discuss, ,,n (di ()ýjr ( , •,• ,,:..L,:i.

of the carbon dispersion combustion process, dlid probiui, k.l hl , u.'ti.,1r i .,l .4A more thorough discussion of slurry combustion 4ýndanieiIii ¾, i- .1 ilO'(,Iin Appendix B.

The Combustion Process

Figure 3 describes the ovurall combustion pro' .. . f(,, .I , ,i. ,dispersion fuel . Volatil ization of the slu'rry i ,rudioe'' Ihydrocarbon vapor and carbon particles. As it would i I Il r,•1' i,''1

liquid fuel were being combusted, the hydrocarbon v ,l,;r fill., i. ,Ii.i'4,1.i', icomplex process of pyrolysis and partial oxidation. In '.ii,,,,i' Ii(.,! I .. ,the principal products of these reactions can be cons id'l-, d tr , :1( Il, ,d,.J, '',,carbon monoxide and soot. The H- then under(goes, chair r:,;ll:lIi, I',1 lT,.which provide free radicals (H, 6, and OH) and teriir, ati(,,, of 9 , liir'd.O,.results in the final combustion product f2•. Th's.n. radi 1',. ,, Ill i,.in and often dominate, other chemical pro((cess oc( urrinil wVlIh In , hfcombustion process. The pyrolysis process is. infl .ri,'-,l H, ; . . ,. '.

and CO disappearance is actually control led bhv :il.•1it . I,, 1 i 0 ,B,:cjas phase processes are very rapid at the telilerior,... (if i, ,aero propulsion application (,200( K) anid nj', iii•ij-lil* .. , i (,, i• w ','.w

ble in less than 2 ms. However, soot oXidtit ri ri 21i ,"•I ''-> li I 1.,,, 11,,oxidation which is slower than that occurting in th-( j-i,. i•.i , in, n

3 shows this time requirement as tl.

Combustion of the carbon I)ort.i( (Ti Of tiUlf , h ' .11 , ''.',i ?'r ,a heterogeneous reaction. The 1.ivlie requiirid fo ,i: I 'I i I Ldesignated t2 in Figure 3, deperlds ulpon Iiu'tu ,' ','i' i' ', ' .-. 1.,. wr.

- 1 6 -

II

E

I.,.

a)iU -=

4 4-, )L-

0 E

V,,)

UN 0rU

..-

0 400 C 0

0 u

-17

dIconditions. Carbon black is usually of a diameter qreater than soot.Therefore, the release of the carbon's heating value will require at leastas much time as that for the case for a 100% liquid fuel. Consequently,the carbon oxidation process will lag the consumption of the hydrocarbonportion of the fuel.

The soot and carbon oxidation rates will be controlled by thesurface oxidation process. Figure 4 illustrates the combustion timerequired for particles of various diameters burned in the stoichiometricallycorrect ratio of fuel to air. The band of required times at lower diametersis due to data variation for different types of carbon (further discussionof this later). An abrupt change in the slope of the combustion time/diameter characteristic is apparent at about 10 Ini. This corresponds tothe transition between diffusional and surface kinetic control of the

particle oxidation rate. At the diaireters of interest to carbon slurries I(<1.0 11), the process is definitely kinetically controlled. The substantial

difference between the extrapolation of the diffusion controlled behaviorand the observed carbon burning times at small diameters indicates thatmuch is to be gained by consideration of methods for ;uceleratiun ofcarbon particle surface kinetics (i.e. selection of the proper carbontype and catalysis).

Consideration of this view of the carbon dispersion combustionprocess durinq the first year of this program has allo.ed identificationof three key problems which warrant further explanition in this sction:(a) difficulties with fuel injection, (b) droplet vaporiz,':ion problems,and (c) uncertainties in carbon burnout.

Fuel Injection

Two unique types of fuel injection prollene, ar, ,,nr.'li ter-edwhen using carbon dispersion fuels. The first involves tIh. poLentialprevaporization of the hydrocarbon component of the fuel within hot fuellines or the tuel nozzle. When 60. of the tuel i'; sol id ,iaroon, los,of even a smnall amount of liqu',id carrier will result in i iuedirit(, t'1ul)m irl(j.Lons,ecuently, special care must be taken to i ,ore &,.•iv, ! tilt' fell lin ';or nozzles reaching temperatures at which prevIuriz.-,iLi,, (Ii sl urry i nsta-bility can take place, Further, the initioil flow (Jf '-,liirr1 irl. o the fuellines and nozzle must occur with cold hardware. buritmn ,.tiL,.;wn, the,fuel nust not remain ir, hardware which can hecoiro ;.' hot (i.e. it ! u'. t be:purged ) . If the slurry is nct purged, deposition ri (.,ili)1 iilt V thehot wall will result and may eventually rcsult irt pluplirni. Ihisphenomenon is analogous to, but much more dramatic 1in, the, .terurrrs pray bar plugginnq problemi which occurs due to hi qh L":::,m:,tort n',ti bi1ityof liquid jet fuels.

The second problem involves depo; itin of .n,, f ,i wi iithi the.(ormbustion sysLeimi after injection. If the 1 jni d Iul I II: Iii')" nl

hot surfave(,, the liquid portion of the fuil wilý i1 i, v ,,/.: l. Iire,,0 depiosit o(f carbon behind. In addition, Vi Ow ir- f; f.i': roillhusli(oiiZ(,Ia,, (after vapiorization) is allowed Lo I(,liie I., (., . . ii,,. :witilin th, Ior, i'bustnr, the carbon will he ittra( te o ii . I :'.ithe surfac-. Thi, phch noieienon is known (rs t•r'!.l I'll- I!. W . ill m i Utlor(-sj a sI1s ta(11t i a ca rbon depos it car build tul' IJI I i i ll II -I 11 !'1 , II "d I 1 ']111 .%

4-All

10 5 10 3 .

5 2105 10

10 4 10

t0

103 one second -1.0

0

a \OXYGEN0, DIFFUSION

2 COKTROLLEDS102 0I10 I

/F

#....EXTRAPOLATION OF DIFFUSION/ CONTROLLED

IL

10 O-3 0-2 0-I 1 012 1.3 1410 10 10 1 10 10 1.01

Initial Particle Diameter (•m)

Figure 4. Dependence of Carbon Particle Combustion lime on Parti(.lu Didrleter

-19-

of the combustion process and/or the hardware coolinq scheme. Furtherproblems can result if the carbon breaks away from the deposit site inmasses large enough to erode downstream engine components. Depositsof soot in cormbustion systems have been shown to destroy turbine hardwarein this manner (11). These problems will require special attention insystems which utilize low pressure air atomization fuel injection methods.Prevention of carbon deposition on air swirling devices will be especiallydifficult.

Volatilization

Since carbon oxidation is likely to be the limiting -tep inthe complete combustion of carbon slurry fuels, our current emphasisin slurry formulation has been on the use of sub-micron c 1,hor, paticle'(O.Ol-O.3,j) which would be expected to have short burnout tieu•s. Theslurry enters the combustion system as liquid droplets containiricapproximately 60 by weight of this very small particle si.>: ca,'Don.Depending on the injection technique utilized, the average dro)Intdiameter may range from about 50 to 150,: and, therefore, eaeh drop

contains millions of small carbon particles.

The vaporization of the slurry droplet oust be such Lhiat thecarbon particles enter the gas phase individually or in small groups.If agglomeration occurs -- all the liquid is removed without causing thecarbon particles to disperse-- a large, 40-12C?, mass of .:,rbor: wiilremain, as shown in Figure 5. The burning rate of such a l-:roe diameterparticle would be diffusion rate controlled. In thic ... p thr r.r'bhiis-tion benefits of providing the small particle size will I r,, , t.fact, if the agglomeration problem cannot be avoided, i 0' t I.lityof combusting the slurry carbon content in the short time Ia'o 1 [-1•t']e(0-5 ms) will be in serious doubt.

One approach for solving this proble, is to r1 i1 ize thevolatility difference between the internal (high densitv t •'Vr;,,o 5' ndexternal (hydrophile) phases of a three phase carbon s],rry fc-•o i•.on.This would result in a droplet fragmentation proce,.., which forces ir,pivi-dual or groups of carbon particles to separate fr'Omil the : dcplet.,Other approaches might involve some treatment oF the carbon 1.rl i1.iesto result in repulsive forces upon heating or even - !, cdiiiti o f,-injection technique. In any case, it is imrportan t to L;t ";WbiC, t,'observe the droplet vaporization process so that di rider.t•:od, '1 fthese approaches may be gained and correlations forfo:; 01,,,il'j slurrieswhich minimize agglomeration may be developed.

Carbon Burnout Pate

Many of the postulated problei,,,s cosern r i , ,of carbon slurries could be at least scoped if v it a , ! J,_;,I.J.',means of predicing carbon particle o:.idatiurl rt'', iioveral 1 picture now extant of carhon 1)nrt! ,r'r.leaves much to be des;ired. A comparison studi o f !I;, I .i.includes 'ceasuremonts of soot oxidation as well ,s ir .", .carbon particles, reveals that the ranqe of hurnin" tiri , 'large, encompass ing several orders of mriror i itude (i , i .rE.ven for it s5pe ific type of carbon, £:,ch cs :rr:h ,,. .ranre of aIilitst tw() ordcr'. of ma';rr it. ® (.1't2.r 1 ', " '' -

20 -

°+

( v~oPQ.

50 mp PARTICLE-,-

601 C IN SLURRY IULLVAPOR

CARBON AGGLOMERATE

CARBON AGGLOMERATE BURN TIME WOULD BE > ls

Figure 5, Vaporization Process For A Carbon Slurry Fuel Droplet

- 21 -

I

Key tc Symibols 1n ilguie

- ioiM r e f .1.6. M C I _K. .",A*.o • -r - P- -r--Tr- --- T rI--

ParAle- colke

s° . SA P. -c

A J." 10I' I L';'1 'j•

C, v Cýj C?

IT's)

ICB * 1 , 11P : ! ; , ,, "S~m I • 2o o ...

O - • .• • ', ; , , ' . • . , • . "

2? iJ

1__ _ _ _

01

I I

- -- __ -.= ==- -----.---~ I-

A . -

_ __ _6 9 0 12 IL 6 ry f'r U.,

r ib) i~c

>1are in serious disagreement; there are virtually no conditions underwhich we can assume accurate knowledge of the time for carbon burnout.

Table 6 provides carbon burnout time information for theconditions of interest to cruise missile combustion systems. Informa-tion was extracted from data or correlations were applied for conditionsof 2000K and an oxygen partial pressure of one atmosphere. Some ofthis background information recognizes the influence of fuel-air ratioon burnout, but much assumes the carbon burning in a non-depletable

atmosphere of oxygen -- i.e. very fuel-lean. Such information under-predicts the actual required time for burnout. It is evident that the Jdisagreement for carbon burnout time among these previous workers exceedsan order of magnitude. And the range of uncertainty, from less thanone millisecond to ten, spans time requirements at which efficient slurrycombustion would certainly appear to be feasible to times at which appli-cation of the concept is in serious doubt.

It should be noted that the possibility that burnout times maybe on the order of on. millisecond is extremely encouraging. If burnouttime were on this order, the "lag" expected for carbon burnout may benegligible and the combustion process of a slurry may becorre very much Ilike that of a liquid. Ccnsequently, even if a burnout time for carbon 4

of 2-3 ms were determined, motivation to increase the rate of burnoutwould remain with the overall objective of making the dispersion behaveas much like a liquid as possible. It is evident that the objective ofdeveloping n optimum carbon dispersion fuel will be substantially enhancedby additic,.al study of carbon oxidation chemistry. Oxidation kineticsof the actual carbon being utilized in dispersion formulations should beestablished and the impact of various fuel formulation variables (especiallycatalyst behavior) evaluated.

C. Ke Objectives

I Formulation

To overcome the formulation problems noted above (Section B)the frilowing objectives have been established.

Obtain a thorough understandinQ of the interactions amonqcarbon, fuels, dispersants, and other ingredients to permitpreparation of carbon dispersions with optimum properties.

Preparation of carbon dispersions that are stable inboth static and dynamic tests, and which will remainstable and usable over a temperature range of -60 tu +140F.

formulation of a fuel with a winimum of 1,O ,OO) ftuogallon which is compatible with anticipated missilesy tens.

2 Combustion

To overcore the combustion problems noted above the followingobjectives have been established.

23- !

E0

0 0

aw) v'-

E

e.- m- U

cooCO 0C

CdC

0l

4-'ro50 0

0u

C- C.-C.4 c~. -(I -ýo C) -4 U X

=3 4- C0 .

I . - -~- m o :o(

4ýSC> - ( 4- 4-J .i•- C- ro.-

0' 4..-) C

4. -0- . - " 1 - <

4-)- CD L- C

k.0 0 ... a C,- V) r _

Cl to C:) (1

LLJ 12J I .. . ..L) : o fo O .aD

0 C 04U-4- 7 0 L

0 coO IV - .c -0 C) 0 .

u 4) E 4-)

4 eo0 S

4- C. m r

$. .• •o 0-0.) "_..

C- 4 -)4.1

a 00C- ) 0 0J -aC a) L u4SI .o 4-)

C.)

.4--

I ~ ~ r > )O fl

ul ro

24.

Evaluate the combustibility of different carbon blacks whichmay be utilized in a carbon slurry with special emphasis oncarbon particle size.

Determine dependenLt of carbon slurry combustibility on mixtureratio, temperature, and time.

Examine the utility of catalysis to accelerate carbon oxidationto allow a slurry which combusts in times similar to hydrocarbonfuels.

Ascertain difficulties associated with slurry fuel injection.

- 25 -

SECTION III

APPROACH

The overall approach utilized in this program is an iterativeexamination of the key combustion effects which must be considered duringformulation of candidate fuels. In this way, the key variables -- includingtype of carbon, size, loading, type and amount of surfactants, method ofpreparation, the need for catalysts, etc. will be logically and thoroughlyconsidered. Our approach to both formulation and combustion studies arediscussed below.

A. Formulation

1 Discussion

To evaluate the variables noted above both conventional andnovel techniques were used. Conventional approaches were ustd for selec-tion of carbons, equipment and procedures for mixing. In selecting therequired surfactants and final composition of the carbon dispersion bothconventional and novel techniques were employed. Another fatureof our approach was to incorporate catalysts in these formulations. Thesenovel techniques were:

(1) Use of solubility parameters to select the optimu.nusurfactants.

(2) Formulation of a three phase system in which a twophase dispersion of carbon black in JP-IG is emul-sified in a hydrophilic liquid; the latter servesas the external phase.

These are discussed in more detail below,

Solubility Parameters

To prepare stable emulsions and dispersions formruldTors mayuse various techniques to aid in selection of the proper surfactant.The HLB method of Griffin (20) is useful for liquids and the SolubilityParameter technique of Hansen and Beerbower (4,5,6) is usEful furliquids to form emulsions and also for dispersing insoluble solids ina liquid (dispersions). The latter technique is cf special, intterestsince the required values may be calculated if tha chenmical structur..of all the components are known. While the cnemicel structure- nfcarbon black is not sufficiently known to permit calculdti,),, it canlbe determined by experimental methods. This approach wa. used in ourprogram. The val.je of this approach may be seen by an e-K lanation ofthe basis for thE Solubility Parameter, Cohesive Energy 1jenii, (CH)and Cohesive Energy Ratio (CER) concepts.

The interactive forces present when a solid is contacled bya liquid are due to physio-chemical interactions of the spe;:ifiý: ,oleculesinvolved. These interactions may be expressed as thre, seiý rit- ;urtiulparameters as was done by Hansen and Beerbower.

- 26-

S Ii

(1) 6D represents the London ("dispersion") forces whichare present in all liquids. In paraffinic and naph-thenic hydrocarbons, these are the only forces present, -1and 6=6D, where 6D is the London force contribution.These forces are omnidirectional and arise from thecyclic field produced by the orbital electrons,

(2) 6p represents the Keesom (polar) forces generated Aby permanent dipoles in certain molecules. The Keesomforces are dominant in nitriles and nitro compounds, andquite strong in esters and ketones.

(3) 6H is designated "hydrogen bond" parameter, but actuallyincludes forces due to several kinds of electron-transferinteractions.

The relation of all these to one another is essentially based

on the first law of thermodynamics which states that energy is neitherlost nor gained. In this case, it is the "Cohesive Energy," (CE),

CE = V62 = Hv - RT= V 2 + Vp2 +V6H 2 [I]

where V = molar volume and AHv = heat of vaporization and RT are thec(,- entional gas constant and temperature, respectively. (In the case

o, fused metals or salts, two more terms for metallic and ionic energieswould be needed.)

From this, dividing by V we obtain the "Cohesive Energy Density,"

CED = 62 = 6D2 + 6p2 + 6H2 [2]

Methods for determining the partial 6 values for liquids arewell understood, and many have been tabulated (4,5,6,22). For liquidswhose /iHv is not known, excellent approximations can be built tip fromgroup contributions (4,22). This expression (CED) can be determined forthe various components of a system. The Cohesive Energy Ratio (CER)of the two phases may be adjusted by use of surfactants to get aminimum value (4).

CER = CED Phase 1 [3]CED Phase 2

A minimum value of CER means that the heat of mixing will be essentiallyzero and the dispersion should have optimum stability, There is alsosome evidence to indicate that the same criteria lead to minimum viscos-ity (22).

Three Phase Systems

In addition to investigating conventional two phase systems,three phase systems were investigated. This concept of making thecarbon dispersion in JP-1O as the internal phase of an emulsion, athree phase system should have major advantages (10,23) compared toa two phase system as noted below:

-27--2

- A suspension of a viscous material (carbon dispersion) in athinner liquid (external phase) tends to have a viscosity which dependsonly on the viscosity (no) of the external phase and the volume fraction() of the internal one. For very dilute emulsions the Einstein equation

n = no (1+2.54) [4]

holds fairly well, but as increases, more complicated functionsare necessary (24). However, the viscosity of the internal phase neverenters the equations.

The presence of an external phase can be useful in impartingproperties to the system which are significantly different from thecarbon dispersed in JP-10 alone. For example, usinq ethylene glycol,dimethyl formamide and other hydrophilic liquids (10) it is possibleto alter the wetting characteristics, flow properties, temperatureeffects, and other parameters.

- The third phase can be selected to have a significantlydifferent volatility than the Jr-lO so that the sprayed particle "pops"and provides secondary atomization. This effect is anticipated toreduce the tendency for the carbon dispersion in JP-10 to anglomeratewhen the fuel is vaporized in the combustor.

This three phase approach requires preparation of a systemwith multiple interfaces as shown in Figure 7. Using the CER concept,all of the interfacial parameterf, can be det2rmined and it i!. possibleto select the optimum combination of surfactants with a minimum oftrial and error.

2 Experimental Methods

a. Preparation of Formulations

Carbon Black Selection

Carbon blacks of varying par'icle size (13, 25, 75 and 300 nin,)and low (linear) structure were selected to determine the effect c,particle size on combustion and formulation properties. Formulationscontaining from 20 wt '% to over 65 wt ' were prepared.

Mixing Equipment and Techniques

Standard laboratory mixers, a colloid will. and roll jiilliwere the major types of equipment employed. The ,ixinq procduregenerally used was to mix the surfactant with the ,FP-', u11til dis-solved (10-20 minutes at room temperature); ddd this to thli colloidmill, and then add the carbon black slowly (over 10-20 ,,i rot,.;

while circulating. After all ingredients were miAied rt(lrctIlation WdScontinued for another five minutes.

28-

C x

4-

0 / 0 . L

4 -

1_

-

0-)

'1- c

-C

0

(U 0

ý'- Q

4 0

C, 0

0 0-

O) L- c 1- V

ru 01 - LU'>k 0d ev 0)mpt)-i= C

-a

c.

,

.0- -

- - .,- -.P ,

4-

eo Eý

0

rD

C

C)

29 -

- d -m ue nu uan -• .. . . . . . . -. . ." .. . . . . .

In some cases the materials from the colloid mill were thenprocessed on the roll mill by passing through the rolls from two toten times,

Dispersant Selection

Choice of surfactants for initial screening was made on thebasis of our own internal knowledge, the existing literature and infor..mation from our consultants on this program (Prof. A. C. Zettlemoyerof Lehigh University and Mr. A. Beerbower, an ER&E annuitant), Theseformulations were used extensively for evaluating carbon dispersionproperties and combustion evaluations.

Parallel efforts were also carried out to determine the sur-factant requirements of carbon black by CEO (Cohesive Energy Density)techniques. This procedure involves measurement or calculation ofsolubility parameters as described below:

To measure the partial solubility parameters of particulates,it is necessary to test the particulates' interactions with a range ofsolvents. The preferred method for, determining the partial solubilityparameters of the particulate surface is the sedimentation technique ofHansen (5). In this method about 0.5 gm of dried carbon black is shakenin a 10 ml qraduatld cylinder with 5.0 ml of a reference liquid havingknown values of 6D , 6p 2 , and 6IH2. The mixture is allowed to standuntil no further settling occurs and the volume of carbon black insuspension is measured at the carbon/liquid interface taken as theequilibrium dispersion value. (it generally required 2-30 days to reachequilibrium.) Samples were run in duplicate using clean (washed withcleaning solution), dry glassware. A series of severty-nine referenceliquids (listed in Table 7) were used. The dispersion volumes obtainedfrom the laboratory results are converted as follows to cc of dispersion/gm (F factor).

F .Volume of Dispersed Carbon, cc [5]We6iVt oCTarbon Black, uji

The F values are then plotted in an order of rankinq by the methodsemployed by Hansen (5,6) and Panzer (21). A summary of this procedure isgiven in Appendix C and is shown in the results (Section IV A 2)The data obtained from this procedure identifies the type of surfactantrequired to disperse or emulsify one miaterial in another.

For materials of known chemical structure, the solubilityparameters may be calculated from group contribution; of partial sclu-bility parameters shown in fable 8. The procedure used may be illus-trated by the following.

For example, to calculate the solubility parameters of thecompound shown below (using Table 8) the compound is broken down, intofunctional groups and the vdlues determinred as noted.

CH3 -CH2 - CH2 -CH2 -CH2 -C-OH (n Hexanoic acid)

- 30 -

a.'

.T AB EL 7

RLE (RE NC _ LQUj _ 0 2SJt F ..O SOLUBILITy PARAk TI(R DLT(RIhtliATlG:i

Cornicve 'frargy. flhrity Parranttr.(ca l/r'Ol e) Ilekefcrcn •v "lcjq ud

t -L 'P

1 l 10aifl 7,3 0 C2 Octane 7.6 0 0

3 3 rj erane 7..I 1 (, -4 HCAI!I e'..ant" 1_0 .u

Sc 1 4 6 . 7 0 .

6 Carh ';n 7iulflde 1 6.0 G G. II OtIA thqlaiflne 7.31 1.1 3.a]

b lsobutylo etatl 7.4 44.19 y ,_I ¶ h , y l a hl m a . . 5 3 . 2

1.I At. i I.rdl e a) / .i

ro-,:i • el'h ll ,

li'c apt., ln 1 .t

|f 4Ihylc.11 glyu ol 7 •t.it 1, 7_.

fi r I-' 4 h, W .il

r O e. t;"

At r t, 1 , rro luI I r f.-r'vja nti 47.

PW ? A fl' .yl t nt r I I 4 p" I7

.4 I;'j~:r'f Art-,nt ( ri ? , 1ým

1'r4 ol :rcIr 7.' . '

6' An I, I I a Icrt.teitip -

,," ''I i al,, r.ul'atlr ,.3, .,

J'* I r,, 6 , I I.A I 1 '0 1'31! Irl- ( r l ..1, pa. . is,

+. I' !: .Ift'./ A i.:Cl+ "

"14 Isi. ttf, ;I, I s, ly rr. ,

4 1 P a t, r 1 a ('. 1 44 ; 1 r , • 1 ;l .4 ,

+2 M ,,y. h i odt, 1 1; 1t A ,t-, 7.6

LI I I ,

I a. 'r Is,.ir.~ t j+ , ,¢ n

1; 4 .=

',] ~ ~ p10 P, t4] ti v 7~I(• .

V. 1l . ~r - jl r . ft . / 7 4"b; I!. t-rt-ly l I, O:pt4t Ij- 'I n,•

6,l H vv,.ltron,• A(l 4 ) 2

6 ,Im, I f e.ho..• apI; 'lbLl4:hI o(t ,-," f.• a

, k

l

,'i Ith, f ufi + r I'. ,1 t

I.

.'• :1+1( Ii •'tl•r'l].+",• , ,

S•ll l~~a II i -.• i- 1-I+_

Gk UP (.+•T Ib T[' +-; |A• lL •"+[![•lt ; ,,•'•{1I•( ip.,.t.I~lI~=1

n r+r4"• er I• , •j,+ t,, ".4 'A,

•l~l r VH •,;,,' t(r.,J• , •V ,F : +l.,.1

TtlI, I.l

1 ;:

* 1 * ,i

S• '.4,.' + t'*.i Ij 6-(t'I', ' ' '''.

"" PI+'i', I v,1.l A?.. A" ?

• .', .4' , .J I ,+,t,/

*.. . 2, .I...".'r

ii'rn-,., C SI'''. ..I'* 2'' '*7'

- . ',.'Ir- • C : ' ,I .

I-I' fir..j i'. -e ..- 2S'J+ +'+') 7 '.1•'

.4 ].' / h.I.

,1

,-

1

't+

i, ,It,+ •);"- . • " 3.,. " "

I .. H. . . . . . * '.. .

I. . 1 . • • J :'t• .i + • l sl ''

+ • •

.'." 1-"1 .• I,' ,.' " ( ,,r :i It•

'I. . •+;.l "

; .

Partial SolubilityParameters, cal/mol Solubility Parameters

Functional Molar Volume ElectronGroup AV cal/mol London Polar Transfer London Polar Electron

AV6D2 AV6P A V6H" 6 D 6p 6H

1 CH3 33.5 1125 0 0group

4 CH2 64.4 4720 0 0groups00

1 C-OH 28.5 3350 500 2750group

TOTAL 126.5 9195 500 2760 8.5 1.99 4.66

TY--The Solubility Parameters are calculated by dividing the individualPartial Solubility Parameters by the Molar Volume (AV) and thentaking the square root, The unit for Solubility Parameter is calleda hildebrand (cal/mol)I/d,

Catalyst Preparation

For the preparation of homogeneous catalyst formulations,

the material was blended during processing in the colloid mill.

Two Phase Systems

The methods described above were used to prepare two phasesystems. In all cases the mixtures of carbon black, surfactant andJP-IO, with or without a catalyst, resulted in a two phase dispersion.

Three Phase Systems

Formulation of a three phase system always startled with atwo phase dispersion such as described above. This two phase systemwas then added to the hydrophile containing one or niore surfactants.For example, a dispersion containing 50 wt 7. carbon black in JP-10was added to Formamide containing surfactants with stirring. Simplemixing was all that was required to form an emulsion.

L3

-33-

*1j

b. Evaluation of Formulations

Rheological Properties

The most significant means of evaluating initial propertiesof carbon dispersion formulations is to measure rheological properties.Visual observations during preparation provide a qualitative measureof whether the material flows readily or forms a mayonnaise type fluid.For quantitative measurements several instruments and methods wereemployed. These are listed below.

Fann Viscometer

This instrument was used for most rheological determinations.It is a direct reading concentric cylinder type viscometer. A stationarybob and the rotating cylinder are immersed in the liquid contained in acup. Measurements were generally made at room temperature, but it wasalso possible to use this instrument for redings at lower or higher tempera-tures by cooling or heating the sample and measuring cylinders. Thisinstrument is capable of operating over a range of shear rates by simplyadjusting a knob to change the speed of tne rotating cylinderr.

Haake Rotovisco

The Rotovisco is a rotating viscometer in which the sub-stance to be measured is introduced into a gap between a rotating andfixed surface. The instrument used hdd a Couette (cup and bcb) system,the gap is a circular one between two coaxial cylinders -- one of whichis stationary while the other rotates. The viscosity is determined fromthe resistance (to rotation) caused by the sample material. The factoractually measured is torque. With the Rotovisco, the measurinq bob orcone can be rotated at different fixed speeds which permits investiga-tions over a wide shear range. Measurements can also be made at dif-ferent temperatures by heating or cooling the sample and detecting partsof the instrument.

ASTM Methods

For testing of promising formulations ASTM D-1092 was usedfor measuring viscosity dnd ASTM D-2884 for yield stress and shearstress.

Stability

Static stability was measured by placing a sample in agraduated cylinder or tall bottle and measuring the amount of materialwhich separated out on standing. This was a seminuantitative measuresince the amount settled was estimated by probing with a metal spatulaand estimating the percentage of solid material which settled as afunction of time compared to the total amount present. However, cs apreliminary screening proceujjre this was very useful sinre the (joalwas no separation. More severe tests such as shakinq, pu-'pirnq, andcentrifugal tests will be utilized in the future for the morrý prnmnisingsamples.

34 -

J -4

Heat of Combustion

ASTM-D-2040 or 2382 were used for Net Heat of Combustionin Btu/gal. Calculations were also used for predicting Btu values of _

blends.

Pour Point

ASTM-D-97 was used for pour point. The significance of thistest for measuring the flow properties of systems Such as describedherein is subject to considerable question. Experience in the petroleumindustry indicates that many products will flow below the pour pointdetermined by this method, and conversely some will not flow above tnepour point given by this determination. Thus, the results may be usedas a general guide, but not as an absolute measure of the flow character-istics of a system. Actual pump tests are the best criteria, andmeasurement of the yield stress and shear stress by ASTM-D-2884 nayprove to be more informative than Dour point. Not enough data hasbeen obtained to date on carbon dispersions to resolve this question.

Nozzle Tests

Spraying through a nozzle is often a critical parameter tomeasure in systems such ris described herein since the shear encounteredmay cause a serious breakdown of the dispersion and/or emulsion. Testson nozzles are described in Section III 21 on Combustion ExperimentalMethods.

B. Carbon Dispersion Combustion

Development of a viable high density car,:on dispersion fuelinvolves fully understanding its combustion characteristics. Thereforethe combustion performance capabilities of the candidate slurry fuelsare being considered in detail during the fuel formulation process.Parametric studies of slurry combustion characteristics are underway.Areas being investigated include: 1) combustion and carbon burnoutefficiencies as a function of temperature, residence time, and equivalenceratio; 2) the effect of carbon type, particle size and loading oncarbon burnout; and 3) the potential for catalytic acceleration ofcarbon burnout,

During the early stages of the program, where the influenceof carbon type, size distribution, and catalytic effects must beevaluated, a small-scale (less than one liter per hour fuel consumption)combustion system is being utilized which provides fundamental carbonoxidation data. This approach has definite advantages at this pointin the program: 1) more potential slurry candidates can be inexpensivelyscreened due to the small scale nature of the experiments and 2) smallscale equipment can be more easily modified to meet progrdamm needs.

- 35 -

6

The combustor produces results that are meaninqIful to actualoperation in gas turbine or ramjet engines. Experietlce has shown thatthe recirculation regions of gas turbine primary zones or the steps oframjet dump combustors are well mixed in nature and can be simulated bystirred reactors (Appendix 6). Regions like the secondary and dili-

tion zones of gas turbine or the centerline flow of a r.injet are hci(hlyturbulent, non-recirculating flow, and often analyzed as oluq flow zones.The unit designed for this program, the Liquid Fuel Jet Stirred Combus-toy (LFJSC), incorporates both these features. The LFJSC was designedand constructed during the past year and is now operational. A des-cription of the facility is presented below.

1 LFJSC Design

The LFJSC is composed of two sections, a spherical reactorand a cylindrical one (Figures 8,9). The spherical portion of thf'LFJSC consists of two castable refractory halves housed within a 15.? ci.:diaineter metal shell. The refractory material is CastdblE. 141A, a prod-uct of Combustion Enginieering Refractories, capable nf with,,triding tem-peratures of up to 1870'C. Each section is cast separately with a L.l ciidic.;reter hemisphere -- the two sections together create the spherialreactor zone. The cylindrical plug-flow section is coriposed of a 7.C cm:long by 2.2 cm diameter alumina tube. Residence times of ?.4 to 6.G and1.6 to 4.0 msec, respectively, can be achieved in each section with thisdesign. Scale-down or scale-up could allow expansion of these r;'sidenrictime ranges. Each of the following aspects of the facility will be dis-cussed:

Air Injection"Fuel Injection.Meterina of Fuels and AirOxygen Injection to Boost TemperatureFxhaust SystemTemperature Measurement/Sight Port

Air Injection

Combustion air enters the spherical reactiuon zone thr'oumjl twoair injectors positioned 180" apart. Each injector consists tif foirjets. These jets are aimed towards the corners of a cube irnnirminedto sit w.ithin the spherical reactor. One set of air jets is rotatnd45S' with respect to the other to allow the opposinn jets to piesh r.ihi'hrthan collide. Trnis approach provides for excellent mixinq and ifirulatw'.the recirculation characteristics of gas turbine systeris.

fuel Injection

Fuel enters the reactor at two po', itions' 9(•' i oii ,vh airinjector, The rear fuel, nozzle assembly (usuailly ()nly ii,,.f f,,r- lilui,!

fuel) consists of a 1/4J body, a 0,030 cm fuel cap (Model, 1""'0) .tin,pre-'surizing air atomizing cap (Model 6714/) all *ruuiredl froim !lpaySystems Co. This pressurizing cap provides for pr(,iixiq oif tht, fuf.lwith the atomizing gas (air or nitroilen) prior to oxitinri thw. 1i(7,lf,tip. Carbon slurry is introduced in(to the reactor t hroun h Il•(,' ,,Inozzle.

- 36

------- ---------------------...

N

OLLiI--"-~

-- I--

LI;

I~00

c-,

'.0o

C-.

"I".

V) Lo

< 3

"\•\ X

OfC-

__

37-

LU,

INI

0-LJ 0 QUJ)

z Z)Io> CC

N zIC4-0 0

0 0.

ll- <

00

Cal

00

I-j 0

LL CC

0 <

zz Ucc

INN

0 Iu

C/C)_. -J j

The major problem with injecting carbon slurry fuels intoa combustor is prevaporization of the hydrocarbon carrier, leaving thecarbon behind to plug the nozzle. Several types of commercially availablefuel nozzles (Figure IOA-B) were examined with no success. The firstattempt at solving the prevaporization problem was to cool the atomizinggas by passing it through a dry ice bath and also to incorporate aradiation shield in front of the nozzle (Fig. lOC). Both of these modi-fications proved insufficient. Consequently, a specially designed water-cooled nozzle that incorporates specific features needed for successfuloperation with carbon slurries was developed (Fig. 1OD). This nozzleutilizes the basic concept and some similar components of the SpraySystems Co. device, (0.040 cm Model 1650 fuel cap and Model 64 siphonair cap) but provides for cold water cooling of the fuel and atomizinggas up to the point of injectior. Initial testing of the slurry usinga pressure air cap resulted in internal plugging and it was for thisreason that the siphon air cap (no premixing) was utilized in thedesign. This design also incorporates a cleanout needle which is usedto clear any carbon that has deposited on the nozzle tip during theexperiment. A later modification was to weld a 0.318 cm stainlesssteel heat shield to the end of the nozzle (Fig. 8). The purpose ofthe shield is to protect the nozzle tip from the intense radiation fromthe combustor (helping to prevent prevaporization) and also to keepany unburnt carbon from depositing directly on the tip. The injectorpoint of the nozzle can be positioned at any distance between the outsideof the combustor and the spherical reaction zone. This enables any fuelimpingement on the reactor walls to be corrected.

Metering of Fuels and Air

Metering of the combustion air, atomizing gases and liquidfuels is accomplished using caliLrated rotameters. A reciprocatingpiston pump with a pulse dampener (Fluid Metering Inc. Model Nos. RP-GI50and PD-60-LF) is used for metering the carbon slurry since visual obser-vation of the rotameter float is impossible with this fuel.

Oxygen Injection

Oxygen can be utilized to increase the combustion temperatureat a given equivalence ratio. Its addition to the combustion air reducesthe nitrogen content, which is a flame temperature diluent. The oxygenis measured through a calibrated rotameter and then introduced to theair stream prior to its metering. This provides for a uniform air/02mixture.

Exhaust System

The combustion products exit the well-stirred zone and enterthe plug flow region. A small portion of the gas sample is obtainedby inserting a probe through the sample port at the base of thissection. The remainder of the exhaust stream passes through a watercooled heat exchanger. Additional cooling is achieved by introducinginto the exhaust stream a fine water mist produced by an atomizingnozzle. If, for any reason, the exhaust temperature is not reduced tobelow 200%C an audible alarm is sounded. The combustor will automaticallybe shut down if corrctive action is not taken within 30 seconds.

- 39 -

- - o

*ii

a. 1JU Ii NNN

I I I Iz

00"

-L Z

o-

hr cc ,

z z

cO

- 4-LA-

4 ~~~ cc -..a.~~ .. 4 0

Aw

- UJ iia:0 N

O 00 < Q

-Y 40 -

Temperature Measurement/Sight Port

Temperature within the reactor is measured by utilizing aType B (platinum-6% rhodium/platinum-30% rhodium) thermocouple whichenters the combustor through the ignition port. Addition of thermo-couples in several key locations (inside the refractory and on theouter shell) is planned for the future. These will enable the refractorytemperature gradient and heat losses to be calculated. A nitrogen-cooled sight port, positioned in the front of the reactor, providesoptical access inside the reactor. An overall view of the LFJSC isshown in Figure 11.

2 Combustion Proauct Sampling

A 30.5 cm long by 0.24 cm internal diameter sample probe wasused for obtaining gaseous emissions and soot measurements. Specialcare has been taken to prevent condensation of water or unburned hydro-carbons within the probe and sample lines. The sampling probe is hot-water cooled and sample transfer is accomplished using electricallyheated sample lines. All sample conditioning (pumping, filtering, andvalving) is accomplished within a Blue-M Model OV-18A oven maintainedat 150 0 C. Valves have been selected which are rated for operation attemperatures up to at least 175°C and design characteristics are suchthat lubricated valve components are sealed from the gas path. Thepump is a high temperature metal bellows type (Model MD-158 HT) drivenby a 1/4 horsepower motor external to the oven.

Gas analysis is accomplished with conventional processinstrumentation. Particulates are first removed from the gases bypassing the exhaust stream through a 90 mm Gellman Type AE glass fiberfilter located within the oven. One sample gas stream leaving the sampleconditioning system is chilled tc eliminate condensable water (to a dewpoint of about 1O°C) and hydrocarbons prior to introduction into Non-dispersive Infrared (NDIR) analyzers for CC, and C02 , a electrolyticanalyzer for 02, and a gas chromatograph for H2 measurements. The NDIRanalyzers are both Beckman Model 864 instruments with the maximum COrange being 10 mole percent and the maximum CO2 r-,nge being 20 mole percent.The oxygen analyzer is a Beckman Model 776 and .... a maximum range of25 mole percent. The gas chromatograph is a Carle Model 8706-A whichutilizes a three column separation technique (silica ael and molecul:-sieves). A second unchilled stream leavinQ the oven is transferredthrough electrically heated lines (maintained at approximately 150'C)to a Beckman Model 402 Flame Ionization analyzer tor total hydrocarbon(THC) measurement. The reported THC results are measured "as methane."

The particulate sampling system uses diffevent filters locatedwithin the sample-conditioning oven. Two 47 mm filters sealed in astainless steel holder were used in "series." The first was e MilliporeMitex (Teflon) filter with a 5 ,,m pore size; the second was a GellmanType AE with a 0.3 1,m pore size. The Teflon filter was necessary toprevent the glass fiber material from sticking to the Viton O-ringsealing the filter holder. Nearly all the soot collected in th&experiments was found on the first (Teflon) filter. Any soot thatwas deposited at the probe top was blown out ontn a Gellman filter byintroducing a N2 back flow through the probe. This additional weightwas then added to the soot measurement already obtained from the sampleoven filter to determine an overall value.

- 41

E a)E E C -

4-' 4J 4-NIJ U >• 0

>1 >1 '

0) W ) mS- ' =3 : > .- 0

- ~~i0 Ia -CD < 0

V)3

SE "• 0 E • -

44-

E E.- ~ V C\ " .' U) 'c)

- a-

C-

.4-U'

%-

C)

Lg• :-v

S7 : . _ •-, , "•(-4,

, .!

3 Operating Procedure

The LFJSC has the capability of operating on gaseous, liquid,or carbon slurry fuels. A gaseous fuel, such as ethylene, is utilizedfor daily start-up of the reactor and is fed through the front nozzle.It is ignited by a flame from a small propane torch which enters thecombustor through the ignition port. Once the ethylene has ignitedand the flame is self-sL'staining, the propane torch is shut off. Next,the ethylene and combustion air flow rates are gradually increased untilthe combustor temperature reaches 1300'C. This phase of heat-up takesapproximately one hour. Next atomizing gas (air or nitroger) is intro-duced into the combustor through the rear fuel nozzle. At the sametime, the ethylene and combustion air rates are adjusted to maintaina constant temperature of 1300'C. When the desired atomizing gas rateis achieved, liquid fuel flow to the rear nozzle is started.

Once the combustor is operating smoothly on both fuels, theethylene flow is decreased, while the liquid rate is increased in orderto maintain a constant temperature. This continues until the ethyleneflow is completely off. At this point the atomizing gas (air or N2 )is started to the front nozzle. When the proper rate is established,liquid fuel to this nozzle is introduced. Experimental conditions (i.e.temperature, equivalence ratio, and residence time) are first establishedoperating the LFJSC on liquid fuel. After steady-state conditions arereached, the carbon slurry test fuel is injected through the frontnozzle. In most cases the rear nozzle continues to operate on liquidfuel, providing continuous combustion if difficulties occur with theslurry. The combustor can also be operated with carbon slurry alone.

4 Data Reduction

A data reduction computer program was developed to calculate

carbon burnout and overall combustion efficiencies, equivalence ratio,residence time and percent material balances of carbon and oxygen.These parameters are defined as follows:

Carbon burnout efficiency:NCB Mass of unburnt carbon (soot) in exhaust streanm' 100 [6 ]

"- Mass of Mass of-carbon in slurry* n/x 0

*Includes only carbon added to liquid hydrocarbon fuel

Combustion efficiency:

NC :1i Unreleased energy in flue gas resulting from incomple te, - combustion of fuel to CO, HC, H2 and carbon

Total -EneFgy Input x 100 [7]

Equivalence Ratio:

Fuel to Oxidant Ratio at Experimental Conditions [8]Fuel to Oxidant Ratio at Stolch-1 ometric Condi-t-ions

4> 1 for fuel-rich conditions; 4< I for fuel-leanconditions

43-

Ji

Residence Time:

Reactor VolumeTotal Volumetric Flow Rate Corrected to Absolute [9]

Combustion Temperature

Percent Material Balance (PMB) of Species (I) [10]

PMB =(I Mole species (I)from flow rates -moles species I xfrom flue gas analysis x 100Mole species (I) from flow rates

PMB < 100.0 -- Flow rate moles greater than flue gas moles

for species (I)

PMB = 100.0 -- Flow rate moles equal to flue gas moles forspecies (I)

PMB > 100.0 -- Flow rate moles less than flue gas molesfor species (I)

Figure 12 is an example of the output from the computer program.Experimental run conditions, such as combustion temperature, mass flowrates of JP-l0, carbon slurry, atomizing gas (air or nitrogen), combus-tion air, oxygen, and nitrogen are inputted into the computer programand summarized on the data output sheet. Additional inputs include:JP-l0 and slurry elemental weight composition, slurry carbon loading,fuel heating values, gas phase emissions analysis and unburnt carbon(soot) measurement.

The computer program calculates carbon and oxygen percentmaterial balances to check the consistency of the measured fuel andtotal oxygen rates using equation 10. Total oxygen is defined as thesum of the oxygen mass rate plus the oxygen present in the air massrate. The stoichiometric fuel to oxidant ratio is calculated forboth JP-10 and the carbon slurry based on their respective elementalweight composition. The equation is as follows:

C Grams fuel 3

Grams oxy gen) 8 WC + 24 WH + 3 WS - 3 WO [1]-Stoichiometric

Where WC, WH, WS, WO are fuel weight fraction ofcarbon, hydrogen, sulfur, and oxygen

The respective fuel to oxygen ratios are calculated by dividing thefuel rate by the total oxidant flow rate. Equivalence ratios basedon mass feed rates are determined by using equatie'r '. In addition,the fuel to oxygen ratio and equivalence ratio bas, on flue gasspecies concentrations are calculated using the foilowing equation:Value Based - Value Based LCarbon % Material Balance> [12]

on Flue Gas on Mass Flow Rates x O---yen - Material Balance

44

_ _ _ _ _{

0

- r

LL'i

91 U

cccc-' cc

2 Ina-. J-.2

04 2 0

z 0 2 ~0 o -J

'- CC3 -j

2 t. .an -. 42

-J -. ,..

fn-42-

S Z DI 4j -i

WI- L- i w C:

z•

a 2 ar 4 *Sm.- -l • - --(

,:.. C ,., .z z L

-.-.+ • - .

-o 0 - C

--34 -9

o j0 Q oýw an 0 -

a X ac Cc -

LA aý - r V C *v 9- w al L

M4 P- 421

-zz C : -0 C, ýit N.-4A... z

wi Cv orl 15 -0a

06 41 UZUC ,

U Cj I '

L. N'

454

These two values give an indication of the actual operating mixtureconditions. During extremely fuel-rich operations (equivalence ratiogreater than 1.4) the CO concentration may exceed the maximum rangeof the analytical instrument (10%). If this occurs, the computerprogram utilizes the carbon balance to calculate the CO concentration.In this case, a consistency check between flow rates and qas analysisis not valid. Therefore the message "Not Applicable" is printed forthe material balance value.

Calibration equations for the non-linear NDIR CO and C02analyzers were incorporated into the program. This enables the gasconcentrations (CO and C02) to be directly calculated from the instru-ment range and scale reading inputs. In addition, the program employsan iterative approach to calculate simultaneously the water contentin the exhaust stream and the total moles of flue gas produced duringthe combustion process. The iteration begins by assuming values forthese two parameters. First, by using the assumed value for the molesof flue gas and a hydrogen material balance (i.e. moles hydrogen fromfuel rates = moles hydrogen in the flue gas), the computer programiterates until the calculated water mole fraction equals the assumedvalue. Once this calculation has converged, the nitrogen mole fractionis calculated (i.e. 1 - sum of all other mole fractions in exhauststream). Next the flue gas moles are calculated using a nitrogenbalance and the nitrogen mole fraction. If the calculated valueequals the assumed value, the entire iteration has converged. However,if these two values are not equal, then the latest calculated valuesfor the water mole fraction and moles flue gas become the next assum.edvalues and the entire procedure begins once again. Knowledge of thewater concentration was necessary to convert the emissions measurements(with the exception of hydrocarbons) back to a wet basis, As previouslynoted these measurements were taken after water removal to a dew point

of 10'C. A more detailed mathematical explanation of the iterativeprocedure is provided in Appendix D,

The addition of thermocouples to measure the refractorytemperature gradient will enable the reactor heat losses by conductionand convection, along with the inside reactor temperature (based onthe flue gas analysis) to be determined. These calculations will beincorporated into the present computer program as soon as the modifica-tions to the combustor are made.

C. Limited Systems Study

The complexity of an advanced missile system makes it impera-tive that consideration of a change in fuel properties and/or combus-tion characteristics requires an analysis of their potential impacts onthe remainder of the system. To address this problem adequately, ateam of three subcontractors was assembled, as noted in the introduction,to provide guidance during the course of this program. These subcon-tractors along with ER&E have the following as their major goals.

46-

A limited study of range improvement achievable in the trade-off between fuel properties and fuel delivery and combustionpenalties.

Recommendation of fuel characterization test methods to beutilized during this program.

Establishment of fuel evaluation criteria.

The approach planned for accomplishing these objectives isto meet periodically to discuss the carbon dispersion formulationand combustion properties to enable consideration of their suitabilityfor planned cruise missile engines and airframe designs. Propertiesof carbon dispersion fuels could have a number of possible impactswhich need to be considered such as:

- Fuel tank drainage- Fuel filtering- Fuel handling and control devices- Fuel lubricity

Pump priming, capacity aid power requirementsFuel heat exchange capacity

- Fuel ignition and flamecut properties- Techniques/procedures for starting engine

Fuel injection/atomization systems- Fuel flameholder devices- Fuel stability- Combustor design and materials of constructionI, The subcontractors were initially provided with anticipated properties

of carbon dispersions and later in the program actual laboratory datawill be furnished for their consideration of any design or operatingprocedures which may require changes.

47 -

1:-SECTION IV

"RESULTS

The results reported in this section are summarized in the- discussion, tables and figures which follow. Some of the data on formu-

lation has been classified and is included in Part II of this Report.Thus, this section contains coded information on the composition ofclassified formulations to permit presentation of the data in unclassi-fied form. The results are presented in sections which follow dealingseparately with work on formulations and combustion. None of the comn-bustion data has been classified. IA. Formulation

As indicated in the previous section the folrnulation of auseful high density carbon containing fuel should take into considera-tion the nature of the carbon black, selection of dispersants, methodof mixing, two vs three phase systems and ultimately their resultant Iproperties. The results obtained during the initial year of efforton this program are presented below.

1 Carbon Black Selection

A range of carbon black sizes were selected for evaluationvarying in size from the smallest available (13 mi.) to large sizedblacks (300 rm,). The properties of the materials selected arc qivenin Table 9. Initially tests were run on a number of different carbonblacks but it was decided to concentrate our detailed studies onMonarch 1300, Sterling R and Statex M". These were selected becausethey represent a range of 14 ml,, 75 mA and 300 mi. sizes. Initialtests on 20-30c formulations indicated the smallest size carbon blacksproduced very high viscosity materials which could not expect to beuseful. The effect of particle size on viscosity of formulations ofsimilar types of carbon blacks in JP-lO is given in Table 10 andplotted in Figure 13. These data show a significant effect ofparticle size to about 75 mi, but little above that size with formula-tions containing 30 wt c. of carbon black. The carbon blacks below50 m,, produce very hiai viscosity formulations. Table 11, which showsthe complete data on typical formulations prepared, indicates thatblends of small carbon black materials over 30 wt - cnuld not be made fluidenough and had viscosities well over lO00 cp. Much of our effort was

concentrated on formulations of Statex MT (300 ( , size) since thisSmaterial appeared to offer a reasonable compromise between particlesize and other properties of forn'ulation and combu;stion. Carbonblacks with particle sizes between 75 and 300. fill, will bh- uobained forevaluation in future studies.

2 Selectior. of Dispeysants

Initial Screening

Based on discuss ions with inte rnal experts and our co.nsul I.antseleven different dispersants were use d i1 , i, !!V- 4111 1 t- reen it -i) t. ,one or ,iore concentrations witi t a vari iy (,f c:itrho n b1) . . , e 1ole 11.1

- 18;

-- ~- -=-~' -- =

'I

t . .. .. . . . ..=I

L In In °3. oz I'D -I Ili. 3..

ac cc a,

gca...J

.0 - 0? CD ol a, N- r

C_ m co w w 0o

x~i-j ± a~ a. S 0 :' a' CD CC ;

c 1F.k2

0

ol

C - 9- Q

0 0C

-" C '. 0 '0 U' 0 f-N-N C - N N N N C ~l z

1.x-Cu: ~ ~ ~ ~ ~ c a C a .lN - -O C) :

o ~ ~ ~ ~ ~ ~ ~ ~ ~ a -- yu a a ' ' - I .ja a .-c-a

'.54 ~-049-

- lru

-I

.1L

TABLE 10

PROPERTIES OF TYPICAL CARBON BLACK FORMULATIONS SHOWING THE EFFECTOF PARTICLE SIZE AND CONCENTRATION ON VISCOSITY(1), cp at 233C

Basic Carbon BlackParticle Size, mu 14 7b 300

Concentration in JP-10 Viscosity at Room lewperature(l) cp

15 11

20 32,5 9 6

25 128

30 901 3? 1?

40 81 17

50 1?

60 159

S65 2 31

68 431,

"T-ITUsing a Fann Viscometer and a shear rate of 511 S(,C 1

I0' ~'i

- - - 50 -

I---

U)

=> 1000

Li. 900

m •300 o

L,. 700

Ln

:r300

C)

5300

A-• 3 200C)

2I00

Cn 100

50 100 150 200 250 300

CARBON BLACK PARTICtE SIZF, mu

All formulations contained 5 wt " Surfdctant A

Figure '13. Viscosity of Carbon Black Formulations (30 wt .)in JP-l0 as a function of Particle Si.e

- 51 -

*c1~,!

u2

0 • = 0 0

-. 0 0

0 0 w 0 0 ,% a) 0 0 U)-. -

S',i • *' C) / = '.0 0 ('3 '-- . 0- I

* -..: (NJ O - "-.0 -. ". P- C0)-r-.. ,..m * o")v" r- ,--- cra'" r-:r-- 0'

-ca-

.-Jca cr LnE D0) 0)40 10LnU 14

S- 'O - c" a%

02- ' -- n 7 C)C ' n - -0 r- k = .c

' LO L 0- o oa

v' .-R -

CE C) C ." ",-€ " )

- 5- 1-, e -e

11 400141 00 0 0 0 0 0 r- 000 0N 0 L

(AU

:m L 0

LU. 141 C n ) C) C CD Cl ) D CD c C:CD Cl r- CD CD CD )<

CD L) a, ) 1. -4-

C)C

C) CD

L) ~ ~ 52 r-c-

V CL

010C 0

0 - -

0 0 00 = = 0 o o > ,' . ' >• >•, 0

0 00 00 0 0 0., C L . £(D (.D LD im n- Lof

c2

m 0 -I e r.

0,

L

cc - 0 o e

m eo e-

aj~~~~~a .9c n - n nul

DMr C 0 C> r--

r" - - * -,,- -I V :: .

', V In . . . . -m en I L o k ýD =D

01-

S"10 l

n-'4 C)

&- a) S- d) C C<'% z a

v) v0W 'J f ~ ) (1 Cl ct) It) 2: V0 . 0~ ~ ) Ln

a)

4i

S--.

1.

4j 01 3 c C)

a~i -. C- 4'-

cc c -L . L-pý M L- CD D IZ :r- --3C C

3, C;o

(.1 0I- 0 1.. 0) ) 1. 0) C L c fl

'~ 4' , 4-' 4-' C ' 4-' Cl~ dl ) ' 5.

0

0r I= C.

40 0L Ln C L 4

u 41

*; Ln 4-J L( CD0 Ln M O

0)~Q 0.00

041 C

1..0 LAL~'Tj0 t 0 ea-

co CL) (L)

Q)- 4-1

4-j0 n L C W' U-P

Lm ON L) C0101 ) o o en a, c: C. C; 1 .. d0

~ 4-'0 .4 V 'V V 04-CIZl

CD~0. CD E Al-C

C). CDIT LLn - a

vv U1 uN X' x' ('-) XtIV

C)j

'a, j 10 E

0 A1

Ln~~ -c4 C)j ulv

c: *,2) I -

I-I

~53a-

Most formulations were prepared initially using a concentration of five (5)wt % dispersant, and those that appeared promising were evaluated at lowerconcentrations. A complete list of the materials evaluated as dispersantsis given in Part II which gives these materials in code. The choiceof surfactants to study further was based upon measurement of theviscosity, appearance and stability of finished formulations. Themost significant property for initial screening is viscosity, thelower the more effective the dispersant. A review of this data in Table 11indicates the most promising dispersants are A, B, C, D. E, G and K.For initial screening purposes disoersant A was used. In later workother materials were evaluated in some detail as will be discussedbelow.

Selection by Solubility Parameter Measurements

This method has been described previously (Section III A 1 and 2);the experimental data and results are given in this section. The threecarbon blacks, Monarch 1100, Sterling R and Statex MT were evaluated fordispersibility by mixing with the selected reference liquids (lable 7).Dispersion values for each mixture are given in Table 12. The individualvalues for Statex MT and Sterling R were plotted according to the method.of Hansen (5,6) and Panzer (22) and are shown in Part 1I. A brief summaryof this method is provided in Appendix C. The dispersion values forMonarch 1100 gave a plot which appears to show a mixture of carbon blackswith differing activity.

The solubility parameters for Statex MT and Sterlinq R andtheir significance are given in Part II (Classified). This procedurehas enabled identification of chemical structures which should be theoptimum dispersing agents for these two carbon blacks. However, sincethese results have only recently been obtained it has not been possibleto obtain the compounds identified and evaluate them in formulations.

3 Evaluation of Preparation Techniques

Several types of mixers have been evaluated and the resultson each follow:

Sti 'rers

Low speed stirrers such as hand mixing, laboratory stirrersand a Hobart mixer (bread or cake type mixer) have proven satisfactoryfor preliminary mixing of carbon blacks with JP-l0 and dispersingagents. Such mixtures separate rapidly when agitation is stopped, andtheir appearance is nonhomogeneous.

High Speed Mixers

A Waring Blender has been successfully used to prepare rela-tively low viscosity formulations. For preliminary evaluation of smallbatches a Waring Blender is a useful tool. One disadvantage of thistype mixer is the tendency for air entrapment, It was noted that aftermixing for one or two minutes, many formulations tended to jelldue to air entrapment causing a deep vortex and inefficient mixing.However, on standing the air dissipates resulting in a significant dropin viscosity and more efficient mixing.

54.

u- c ~l to" -d C" i.•0 cv "o •lr,, Ln• A^ 0 nW•J-t MM c 0,1 co.

INu ~ ~~~~~~~ ~ ~ ~ P-.-l- qrK Of M -tP,0 Z a l OCi -P

'•(, -I ;V l ;I ;L 1 ;C 0

E ufa .0 C

.5-

W WQm ~ r •1,i{• ,)e |P. ,-) L -Ln to ,, -,• Cy %.0 to0 on Ml r3 k; -t• ,•Lf M M' to0 to • e

S.- 4-0

(U

cL ~4-0) 0 01-

0

r) LO Ln Ln Ln L1 LA A I) LL) Ln o LAoLOLn t u LALL iAUL

co m rý c t-- 0 L ~ %-4 ,O

,--- ý O,,o•C -* t4--1% U

i--

,.-, ~ ~ ~ ~ ~ ~ ( en. C)• 00C M P.C L CD C)•du r-. Mu' r. - M'L ••''t, ' 0Cj U-) CDCDur) C)':I C ')N f- --- 0 t e C o C0 CDLF

.. j 4-' p-

C')

4- CJ~C

• ,.. . . .

W" E

L CLU

9- cLW to,

..,I . . . . .

•) I'D- ,- -l C 0 (:3 .;

LI)-

U M to 00 C ( r C ,M,- U) r, t= al- - M " -0*0 % W W "- U4- W to .-- a t, t-',- to W M

0) 0010")-,. U " ,-, ,, ,.-'•0 U 0" ,,I-3:• Z .- ,• : •: •--{ ,.O . U•

m .9- C1.-

.. . : .. •vu- - .-.. . 0__._- _ .... - -. 0C .U- . -L -. 0L...-

0 0 )ýc ý r ýoýc ýcc , 0) Lflý r- tC0J. r£U.- a4 r,: 4-'C a;-sr0C>E U , ~ U ' C Ja E .L 4-0~

aJ)' 4'-1 C -O4J0>,C 0WE-0C0C..£ 0--C: 01 *.4-JO fu . L c ,0 L.- C 0 a) LL -' (U >-u4-.i-.n

a~iO )< LM L)00- 0.. CL (V L-4- L4-' Q ) 0m - ea -. C >..J U

01U0W to -0 #V X L- )C I- U d- '0 U-- - 4)C--u 0-G)4-.-0 X >, 04W 0) (a a ) C) 5, 4-J -4 C L) W~ 0 E >,r E 0 a)) Er 0 e'. C,. W S- >, J-- (.) (' C- Ca a'. o~ 1 .0 =- CO0 0 0 $- C' 0 C*--~ -- LA I - CO 0 0 $- Cx. C1 (1) -a f -. O 0 (-1"J') 4- >

#Vn Q o r0 - 0 n - .4- E > I -JQ , .. 0 U4ý-I- 55 -r

S.1l ,- .a ,J l k P.- m Itt -• en, C ) (-- i- r -D r. 03 r NC O c -N Ju M. m q - O' LA, co 0 L.. 0 00 C) to L - (O %0

. L c; "; . . .. . .. , .. W

Cý

a- 0' c.'

.0 0MLA 0 4-)

>

I- LA LA Lin LA) tA LA LA LA LA Ln LA Le) Ln L LA Un U CLO

11 'C x N c 0.) C nR ct0 f OC- %JL d DML -ý NN NN C J -- n-C J.- -O M M 0.) f- CIc

4-4

€.)V)

u 4)E

Cl I

CL

I- >,< 4-)

L.J 4) '000.)-

T; a

>,'al

0-.!C

_ . . .- -LI

0 (W

4) 4-) a)

+ -r- 0C? 4) (UC 0 -v 4) 4-

41 a) 0 .e 0 c ( a)4D j e'ar C! -C 4)- *.-3

Z- - -.. - 4) 4-ý go - l +-'0 Eif4) 00. a 0 ea (D C. C: >N m S.- 10 OE 0u - 00 -) -4-'0.- (a 4-'-1- - S- to (DCU4-)r_) C O04- _0 a) Q) to a S04) 0) s M, C> w L- -I- J4-CM

4). EC. 4)0 ~-~0.C4' 0, 0.c I-'1 u (J)U aL -IO.L 4-1 a*>+- 0 - 1 4C.. r.-3C4- 0 " 4) *.C -'-

c'J a) CU L,) a) S- -a CL.C..C -r-. L.4- 0 0 a) C C\j C a)_C a)4 sd) 1$O ~-4 -1 N (L) 0 to -.-i 4) M. C I - 4) a) S- Q) EA -C N CL (0 C 46-J C 4-ý U- CA 0o) 0) aJ r- C.a - . ~-0to 'o~ r.- C .- W< Q)40 - (o C:

ix) - >,r ,,C( JG42) M 4- C.- -~ U .- - o >,, C 4v). V- >,- I.,>1 - to

CA CU44J 0>., 00 -*.-L4J---)>,CM-C0,- 0 - 4-'0.~ X-J--4-'> -C-0Ma (4-E ) - L = 4.' L -&Q CL = .=3 .,~. L- c -r E 4J W) - -1. P) CL 4) .C 4-J C4-1 1.. -E 1 s-.'-A a) CL 0 4-.-0 -r- U~ C 4-) -C -1 -1-[ 04)( E U C C co 4-3 Q) C

CALi. -C2 a.. z < c~ - m- cX.) M w - mA m2 wi F- CM I- mCO Z c-) C 2.\.- jO

%D r-. M2 M C ') c~ m r Ln %D r-, CO A C:> 0 . C') M -r an k. r- Or) cy' C-) c.i r'' Lc %D 1a-. 00 mC' n C') q') C' -L~ g~~~ n Ln a O an an n al an an an) ko IX) %C) kD k.0 kD to an an a

-56-

4 .) L- ~aia

EU1. .C0

tn 0)

0. ca 0 .

CL L .C

.nL 04.'

1LA LA

L)

-J $

0 >1.4-)

LL)

Li

o

0)

L) 0.0 m Ia ,0., J--

vii

mC 01 .0 1

Cie0 I- to S-.. 4-~ a- r--o .LJ ul o a)

.41 04)- MIica 4-3 - O>.= = 3.,

<~0 0) -r > .0 ;-) 13oý u= m ow 41o4-e

GJm .C0OU) -D t,- W

P. -'. ~ ~a~0 J L ' . 57I

Colloid Mill

The Greerco Colloid Mill proved a versatile tool for prepara-tion of most carbon black formulations. Mixtures containing up tonearly seventy (70) wt % carbon black were successfully prepared.Some difficulties were encountered in circulating very thick systems1• in this mill, but for practical purposes this is not a problem sincesuch thick formulations would probably not be useful. Air entrapmentwas not a problem, but the speed of mixing was regulated and bafflesused to minimize such difficulties.

Roll Mill

A three roll mill was evaluated as a means of reducing theviscosity of high concentration carbon black formulations. The resultsvaried with different formulations, and excessive milling resulted in aviscosity increase as shown in Table 13. Based on these results it ap-pears that one or two passes through the roll mill can be beneficial inreducing viscosity, but the same benefits can be obtained by proper choiceof dispersing agents and preparation in a colloid mill. The differencesbetween the formulations designated #697-84-2 and #697-153-2 [noted inTable 13] is in the type of surfactant used. The latter formulation couldnot be improved by further processing in a roll mill, whereas the formu-lations of #697-84-2 Batches 3 and 4 responded well. The reasons for3the differences in behavior have not been determined but the followingmay be postulated.

1. The decrease in viscosity resulted from improved mixing

and more uniform dispersion of the particles.

2. The increased viscosity is caused by the shearing actionof the roll mill upon the carbon agglomerates resultingin a smaller average particle size. The smaller theparticle size, the higher the viscosity (see Figure 13above).

Some tests were also run on higher concentration (65-70 wt •)formulations, but these could not be processed satisfactorily on themill. The reasons for this are not fully understood, but it was observedthese high conceitration formulations had a very strong affinity formetal surfaces ard stuck to the rolls. In addition, there appeared tobe a loss of JP-10 probably caused by local overheating due to thehigh shear even though the rolls of the mill are water cooled.

Ball Mill

Ball milling appeared to provide some reduction in viscosityof carbon dispersions in JP-I0. For example, using a blend containing60 wt % carbon black (300 mp) with an initial viscosity of 54 cp, byball milling for 60 minutes the viscosity was reduced to 47.5 cp.While this reduction is of some benefit the roll mill or longer millingon a colloid mill is preferred to achieve the same results. Ball millingis a time-consuming, batch operation which is not generally a desirableprocess for such systems.

58- L

CL

E--

Qj

C'0)

"-I

V))

4-.

L t-=

4-- CI•c

WE -' 0 0

0, C --

CL

0)

u' 0- 4-)

(A 00 (A IC.) 2 L

CIOCA tu

C)-

C) CA

.-. (,*) .- '-

4-)

-> C) CD

CU C'o -

t~m et3 U

LL 4 -)0 ~C\J (a J E\

z g ~4-J~

0Ci 04 -In n

-59-

4 Formulation Characteristics

Carbon dispersions prepared to date have been evaluated mainlyby their rheological and stability characteristics. Other tests havebeen run on a few selected blends. The results obtained are givenbelow.

Viscosity Increases with Carbon Loading

Data on the viscosity (as measured by a Farn Viscometer) versuspercenL carbon in JP-lO (from Table 10) are plotted in Figure 14 for Ithree different sized carbon blacks. These data show a nearly linearrelationship trom 20-50 wt % carbon and then the viscosity increases Irapidly above 50 wt % for the largest sized carbon blec< tested (300 mp).

The smaller blacks produce a much more rapid increase. -Another plotof these data is given in Figure 15 for the largest sized carbon black Iwhich shows the log of viscosity to be a linear function ot the weightpercent in JP-10.

Viscosity Variation with Temperature

Preliminary experiments have been run on several high concen-tration carbon black formulations to determine the viscosity changewith temperature in the range of (20 to -50 0 C). Figure 16 shows data on

several formulations containing 60 wt % of carbon black. The data wereplotted on standard ASTM Viscosity Temperature Charts (D-341). Basedon these preliminary data surfactant type and concentration appear tobe significant. Upon extrapolation to -51.1°C (-60'F) there would seemto be an advantage for the higher concentration of surfactant B at thelower temperature range. This is an interesting phenomenon and needsfurther study. Actual viscosity data at lower temperatures will be runin the future. The extrapolated values for the best formulation gavea viscosity at -51.1°C (-60'F) of about 560 cSt (733 cp).

Rheology of Carbon Black Formulations

The rheological properties of several carbon black formula-

tions have been characterized in preliminary tests to get an indicationof the nature of their flow properties. These tests were run on theFann Viscometer which is capable of operating over a shear rate of about180 to 510 sec"I with a shear stress for these systems ranging from about100 to 1400 dynes/cm2 . Typical flow curves for four important flow models: N(A) Newtonian, (B) Bingham Plastic, (C) Pseudoplastic and (D) Dilatentare given in Figure 17. Data on formulations of different size carbonblacks are given in Figure 18. These data indicate the large carbonblack formulations (Sterling R and Statex Mt) are essentially Newtonian andthe small carbon black blend (25 wt % Monarch 1100) shows Dilalent flow.The rheological properties of these systems need to he studied in

further detail by more sophisticated techniques before a final judgmentis made on their flow properties.ed

StabilIi ty

Static stability of formulations has been testpd to obtain apreliminary evaluation of the tendency for carbon black and oil toseparate. Typical data are given in Table 14. As might be expected

-60-

-1

- -4

0 300 mi- Carbon Black(3 75 ..

C 14 m "

900

"0 800

S70

400

C" = 300

* 20001

100

-I- -----

20 30 4U 50 60 70

Wt % Carbon Black in JP-10(using 5% Surfactant A)

Figure 14. Effect of Carbon Loading on Formulation Viscosity

- 61 -

..== - * • - •__ _ _ _ _ -. F -• -- r•''-- : -

3.0-

2.0-9

Ln3

F--

S1.0S0.9

<: 0.8--Z• 0.7-

S, 0 . 6 -

S0.5-

CD 0.4-

S0.3-

VV

u" 0.2-

L.;

t 2.0

K

C:)

e'a

1 00 30 40 .9

W T ' U. LAkbu N BL,;O ,i

F i cl u r e 1 . V i s c os i t y, o f Fo ,- m ,l a ti o n ; i ,- JP - I ; ! ,, , .a r , t -n M . wL c I Lo a d i r , c( 3 0 G n i t B a s i c P a r t i c l e . ' i z ý , I- a l h o l , .l .o - & t • , : t l l i n r

-- 620

~ 0.4I-,

~ 0.3

'CD

C-

C)

4.J-

C)

''0C)

II .u L)

L 0

C L

CDC

(1 ) 'o Q)S- c-.-(U, IVC.

iIx X x O ) •.-SI ) ) 4 ) C - o,

I II L . )-.-

0) *-: -- t

• U

ft ro0a)j E WVo -

('II-EI H..

-- C-

M C ) C ) 4-'

to,-," eo0 10 .0 C ).0 (9C

LA-,) Au. U

0 C I--. .4.. L4~ .)A

Q) 44 4-)(

~~'I, C.DU

CD 0) 1.' CDU' (D cm D O) DCD L)C

I: L(' C') C) C: N O

41 4SO L LPa L

0)4-'6A

f3o

A- Newtonian"B - Bingham PlasticC - PseudoplasticD - Dilatent

4A

Cu

Ii~ I'0

AA

Shear Ratei

F lyure 17. fy,)ical flow (.C rw.v . for foutrImportant F ,low Mudv,.

- i6L-64 •

I!

1600 0 Statex MT (60 wt %)

0 Sterling R (45 wt %)

1500- 0 Monarch 1100 (25 wt %)

1400-

1300

1200

1100

1000

! 900.

700(-

600.

500.

400-

300 -

?00o

1 ofJCJ r0f 3(j l U I uu

S[,-' l',lt•: 5 '

* F•.Ii r1(, '4,l ( jq . 1. " ! , f( (( ,, I I: , , I ,, ,. I. l ...

,35-

iII

L00 0 0 0 0

0 00 0 0 0

0 0 0-•0 0 0 0 0 0,,- .?1I-.

LA ) D 00 0D C) 0D C 0

CE) C) C)

-J CD (D C)

0

C)

LLJ

0 0 0 0

C:.) CCjCn

S-D

CD ~ ~ ~ C0 r- 0: C) CD 0 C

o- 0) CDLL.

00.• tOnJ•

-i~(/ -- C

m cc,~

0 0•

Ia. C) 0, C-3 CD C3,•[

Cli

LL. -L LT

..1=

the small particle size carbon black formulations tend to be stable -

with little or no settling of carbon black in the time period tested 2(up to 5 months). However, some oil separation was noted. With thelargest size carbon black, the low concentration materials tend toseparate out rapidly but as the concentration is increased the stabil-ity also increases. These observations are typical Stokes law phenom-enon, and thus the higher viscosity formulations (small size or highconcentration compositions) may be expected to be stable. Under longerstorage times and more severe test conditions (temperature cyclingand dynamic testing) it is expected more formulations would tend toseparate. As noted in Table 14 even the material which separated outin some cases was readily redispersed with mild agitation.

Solubility Parameter Determinations

The test results on the three carbon blacks evaluated by thistechnique were given in Table 12. The F values were then plotted by thepreviously noted method of Hansen and centroid coordinates obtained.These plots, centroid coordinates, and interpretations of these dataare given in Part II of this report (Classified). The materials i Jenti-fied in Part II for two of the carbon blacks are, according to thisapproach, the best dispersants for the respective carbon blacks. Plottingthe ciata for Monarch 1100 has not *yet resulted in a clear-cut selectionof centroid coordinates and is currently being examined further. Not allsubstances are amenable to this type analysis and it may not be possibleto interpret the Monarch 1100 data. This may not be a serious loss sinceMonarch 1100 would be extremely difficult to formulate at high concentra-tions even using this approach because of its very small particle sizeand resulting high viscosity of formulations even at 20-25 wt '- in JP-lO.

The surfactants identified for Statex MT and Sterling R are cn

order and have not yet been tested in laboratory formulat;uns.

b Three Phase Emulsion Systems

Preparation of a three phase emulsion system containing carbonblack in JP-10 as an internal two phase mixture and a hydr-,hile asthe external third phase has been described in Sectiun III A. Thistype of system requires proper selection of dispersdrts and emulsifyingagents to connect the various interfaces noted in F'gure 7. Out labora-tory efforts to prepare a three phase systew st-arted prior to completionof solubility parameter measurements and thus the HLB method was used inan attempt to selrect the proper surfactants. In addition, the methodsand formulations provided in the Exxon Fjtcent (U.s. i'3,732,084, R-'.. 10)were utllized,

F Y)m this latter work it was detenniricd thot a combitndtioon ofsurfa tnints , Ith an HLB (Hydruphilic/LipophhiliL B3alice1 , al, deve1opedby Atl,,s Cheri, Cdl Industries Division of If!) vjl i , In i; wd, opitillu., f•rvirulsilying a ,ii ,l}crsinr of c btin dispersed in J11-4 with f(tnoli d(nn p U',

water. Since the r.urrU 'rit sy',tem of iritrf-.1t i, "T , V4IIwht. diff ferei! weuxplored surfactints abnve and bel-lw L1U 12 In a', attemlt to identify asatisfact'jry sy'ALrail, We initially ,vailuilt d 'i1 -10 al]r1.- 'r11d th rn J)ru-q.res',d to JF'-lU (.,Irbuo diapliersiurl systi .' .1 iw ', , i,. t ' f Su,.-marizn•d brl ow for severall U if furlrit hyd) .i I. , vt.:L: If-, I, fnlr'1•ii ,i, '.

The procedure used was to titrate the JP-1O (with or withoutcarbon) into a specified weight (50 gms) of hydrophile with stirring. 1Various surfactant combinations were evaluated. The system was testedafter each addition of 50 gms JP-l0 (w or w/o carbon black) to determinewhether it was an oil-in-water (01W) or water-in-oil (W/O) emulsion (byadding a drop separately both to water and oil in a vial and notingwhether it was soluble or insoluble). In addition, the higher concen-tration O/W emulsions were tested for viscosity. The results for, thesystems tested follow.

Emulsions of JP-10

This is the simplest system and the results shown in Table 15

indicate that O/W emulsions were formed over a range of HLB values from6-10. (The HLB materials used are commercially available blends ofsurfactants.)

TABLE 15

TEST RESUITS IN EMULSIFICATION OF JP-lOIN SEVERAL HYDROPHILES

L

O/W Emulsion Formed in HydrophilesWa ter Formamide

Surfactant added to 6 • 8 10Hydrophile at 5 wt % HLB HLB HLB 1{LB HLB HLB

65 Yes Yes Yes Yes Yes YesII 72.2

75 I 1I

Emulsions of JP-iO ContaininQ Surfactant A

This is the system in which most of the carbon black formula-tions have been made. Table 16 gives the results of emulsification trialsusing formamide as the hydrophile. Successful emulsion formulationswere made containing up to 86 wt % of the lipophile (JP-lO with sur-factant A) in formaride with HLB 10-16. The HLB 8 surfdctant systeminverted to a W/O emulsion at the 86% level. Viscosities are shown forthe highest concentrations made and range from a low of about 100 cpfor thp HLD 10 to 150-200 cp for the other surfactints. These resultsindicate an HLB range of 10-12 appears optimum.

Emulsions of JP-10 with Carbon Black,

Initial trials were made with a 30 wt "' Stdt.cx MT formulationin JP-10 a- shown in Tahle 17. These re!,u1lts. indicate a shift in HLl]requirement to about HLB 16 from the optimnum of 10.1 . for the J1'-10system without carbon black. The mdyiMuM Lonce,,tration for dli /1Wemulsion was 86' by weight with a viscosity of l cp at room tempera-ture (?G"C).

- 68 -

E C)I =

0o

I I LAJ iILCS

I C)

F->

I- I) 4-< >- - iI--- I- I-- 0)L

u=• *• .. -.e -.¢

i L. t-0

(A

ca

- 0

U-- C\J

I- -,. - 2, -'"

CD S-

coo

S. .. . 0 oo

- 69- -I U

V) CDL

I-~ 0 eCu

L) I- I-

F-d

(A I D U A

0 I- C) - eLr, CC) CC Q

I-4--LJ

V)) 4- Li

Li)9

TABLE 17

TEST RESULTS ON EMULSIFICATION OF JP-10 CONTAINING

30% STATEX MT AND SURFACTANT AI Lipophile JP-1O WITH 30 WT % STATEX MT AND 5 WT 5. SURFACTANT A I-

Hydrophile-e --------------------- FORMAMIDE ------------------------

Surfactant Added to -,

Hydrophile (5 wt %) HLB-6 HLB-14 HLB-16

Wt % Lipophile Added O/W Emul. W EuOW Emul. O

67 Yes Yes Yes

75 No Yes Yes

80 " "

83 " No

86 No (81)

T7.--VTscosity, in cp, of this blend using a Fann Viscometer, R-1,B-1, F-i at 300 RPM @ 261C, shear rate 511 sec- 1 .

70-

High Concentration Carbon Black/JP-lO Emulsions

Table 18 summarizes the laboratory results and shows that themaximum concentration achieved was an emulsion of a 60% Statex MT dis-persed in JP-10 at the 80 wt % level in 20 wt % formamide and surfactants.This final emulsion contains nearly 50 wt % carbon black in the totalfcrmulation. The viscosity of this highest concentration formulation 7was 107 cp using HLB 14 as the emulsifier and Surfactant A as the Idispersant. The other data indicate that HLB 14 appears to be optimumfor emulsifying these systems.

These results are not optimized formulations and furtherformulation and testing of such systems is currently underway.

B. Combustion

Wcrk during the first year included developing the Liquid FuelJet Stirred Combustor and studying carbon burnout as a function of

Equivalence RatioResidence TimeCatalyst Type and ConcentrationCarbon LoadingParticle Size

Each of these areas is discussed in detail below.

1 Liquid Fuel Jet Stirred Combustor Development

During the past year the Liquid Fuel Jet Stirred Combustor(LFJSC) has been conceptually designed, constructed, and is nowoperational. The combustor design employs features from stirred-combustor devices which have been utilized by Essenhigh at Ohio Stateto study pulverized coal combustion and by Putnam at Battelle ColumbusLaboratories to study gaseous fuel combustion stability. Discussionswith each of these individuals proved helpful in developing the LFJSCand their assistance is greatly appreciated.

Prior to the initial start-up of the LFJSC, a shakedown of thefacility was conducted during January and February. It was during thisperiod that all the individual pieces of equipment and subsystems wereextensively tested and necessary modifications made to the system. Once theshakedown phase was completed, an internal safety inspection of tile facilitywas conducted. The unit, having successfully passed this inspection, was thenready for initial start-up in March.

Ethylene was used for initial crimbustion tests since it isan easy fuel to combust. No complications resulted durini the iqrnitionor combustion of this fuel, After this pha.e of test. Ini waas cuoipleted,attention of the ne,•t few months was focused on npercitinq the reactoron liquid and then finally on carbon slurry. The uni(ue rerluirementsfor combusting carbon slurries have led to the developmernt. of a water-cooled fuel nozzle. In most experiments , slurry wdl, only cu,,'busLedthrough the front nozle while J'-10 was introduc(,d thr'(iqh the rcj r

71

Io ua

CI O

*l 'Lx ) I

o r :3 z t

'CIO, 0; '0

r Ln ML :

' 3

, LA

V~o, i - -A

41 I'-1E 04

_cr I- ~ -

S V LI )JII -

U J 0I ' E l > - 0

L'iA

Io _j-

0 L$D 'm

toj ID OD w

L U I -

a- I I

I 'LA

nozzle, Successful combustion of a 30%carbon slurry for continuousperiods of over three hours has been achieved while operating in thismanner. In addition, the LFJSC has also operated on slurry alone (noJP-10 flow through the rear nozzle) for a period of about 30 minutesduring September and October.

2 Experimental Results

Combustion experimentation during the first year has focusedon studying carbon burnout as a function of equivalence ratio, residencetime, catalyst type and concentration, carbon loading, and particle size.Complete data computer printouts are given in Appendix D.

Prior to obtaining the above data, tests were conducted witha Statex MT 30% carbon slurry (300 mu basic particle size) to verify thatcarbon burnout efficiency at a constant equivalence ratio was independentof the ratio of slurry flow (front nozzle) to JP-10 flow (rear nozzle).Establishinq this relationship is essential if a valid comparison ofcarbon burnout data obtained for different slurry flow rates is tobe made in the future. Experiments were performed at an equivalenceratio of 0.85, temnerature of 1700"C, and residence time of 6 ms forfour slurry rates ranging from 5.2 to 10.5 q/rnin. CorrespondingJP-10 rates ranged from 8.9 to 4.2 g/min. The results of this testindicated that there was essentially no difference in the carbonburnout efficiency for the different slurry rates. Results are sum-marized in Table 19.

a. Equivalence Ratio

Carbon burnout was investigated al. equivalence ratios of0.65 to 1.5. A Statex MT 30. carbon slurry was cnmbusted with 25"oxygen enriched air. For mixture ratios less than 1.0, the slurryrate to the front nozzle was he'd constant while the operating con-ditions were established by adjusting the JP-10 flow to the rearnozzle. Above a mixture ratio of 1.0, the slurry rate had to beincreased to achieve the proper conditions since the maximum jP-10flow rate had already been achieved. Slurry rates varied from 5.8to 13.4 g/min, while JP-10 rates ranqed from 4.8 to 11,5 g/min.The oxygen and air rates were maintained at 10.6 and 178.8 q/Iminthroughout the entire experiment. These conditions correspondedto an overall residence time of 6 ms.

The carbon burnout increased from an initial value of 93.7at an equivalence ratin of 0.65 to 96.5 at d ratin of 0.85. Thisgradual increase can be attributed to the increasing combustion tempera-tore (1500 to 1650'C). After reaching this peiL, the burnout percentdecreased sharply to a minimum of 85.7 at a ratio of 1.5., Thisrapid dropoff after 0.85 illustrates the str'n(e depender•' of carbonoxidation on oxygen availability. A dropoif in temmper,iture ,'cOm1 1710to 1540"C was also observed as the eq,;ivalence ratio wu', increased toI.51. In addition, the carbon mmormuxide, hydrogen and unt'urtit hydre-Cdrbon concuntrations increased with decreasing oxy(min dvdilability.Increased levels of these species i, the exhaust strEamli irndi(ate in-efficient uti I izatior of fuel enorqc . (vwrall combihu',tf,•n cfficie'ncy,whi( h i-, a 'luantitdtivye me•asure of enurrly uti 1 i 7;mt Jm, r.e rev.L'd from

9iK.:) to 91.1 dS tOIe ew Uivalc 'nm C r mL, wm* im,, d', , . IL ,'I,' (.r

7- -

TABLE 19

CARBON BURNOUT PERCENT AS A FUNCTION OF SLURRY FLCW RATEAT AN EQUIVALENCE RATIO OF 0.8, T=6 ms.

CarbonSlurry Flow JP-l0 Flow Temperature BurnoutRate (g/ninj_ Rate (g/mjin) C C?:)

5.2 8.9 1690+5 96.0 -t

5.8 8.0 1690+5 96.1

.96.1 1725+5 94.9

10.5 4.2 1720+10 95.7

TBLE 20

CARBON BURNOUT AND COMBUSTION EFF!ClErNCIES AS A FUNCTIONOF EQUIVALENCE RATIO FOR STATEX MT 302i; 25% ENRICHED AIR i76 ins

Equivalence Temperature Carbon Burnout Combustion Efficiency*Ratio (°C) (:

0.65 1500+40 93.7 siy t con

0.75 1580+15 95,.7 98.6 ;

O. 85 1650+15 96.5 98.3

0. 95 16904,•0 95.6 96.1

1.05 1710+20 94,9 91.4

1.28 1655+I10 89.8 --

1 .52 1540+10 85,7

S-Co-m•-Y-tion Efficiency values for equivalence ratios (Ireater than 1.0

are not meaningful since it is stoichionmtrically impossible to Com-

pletely combust the fuel to CO2 and H2 0,

- /4Ik

from these initial findings that carbon burnout and combustion efficiencyare favored at an equivalence ratio of 0.85 and high temperatures. Toere-fore additional efficiency determinations were concentrated at theseconditions. The results are summarized in Table 20 and plotted inFigure 19.

b. Residence Time

Carbon burnout for Statex MT 30-k. was also studied for residence time-ranging from 4.0 to 10.0 msec. These tests were conducted at an equivalenceratio of 0.85 and a temperature of approximately 1700'C. This particulartime range represented the physical limitations of the present combustionsystem. Since carbon oxidation is the slowest step in the overall combustionprocess of a carbon slurry fuel, burnout efficiency should increase withincreasing residence time. The results do show this, however, only toa slight extent. Carbon burnout values increased from 95.5 to 96.3,Unrealistically longer residence times would have to be utilized ifhigher efficiencies (99+) are to be achieved with this particular carbonparticle size. The more fruitful approach invol,•cm study of smallercarbon particles and catalytically accelerating carbon oxidation. Theresults are summarized in Table 21 and plotted in Figure 20.

c. Catalysts

The effect of homogeneous catalysts on carbon oxidation wa"next examined. Statex MT 30%. without catalyst was compared to Statexto which the following catalysts had been added.

(1) 100 ý)pm Mn as methylcyclopentadienyl manganese

tricarbonyl (MMT).

(2) 1000 ppm Mn as MMT.

(3) 1000 ppm Fe as ferrocene.

(4) 1000 ppm Pb as lead acetate.

(5) 1000 ppm Zr as zirconium octoate

The experimental procedure was to first obtain a carbon burnout valuefor the uncatalyzed Statex and then comrbust the catalyzed Statex. Inthis manner, the catalysts were directly compared to the base fuel undersimilar operating conditions. Despite the fact that the uncatalyzedStatex values are lower than those previously reported, when same daycomparisons are made for all three catalysts the results show that atthese concentrations no siqnificant improvement in carbun burnout wasachieved. It should be emphasized that our hope was for a major dif-ference in carbon burnout (i.e. an increase close to or qreater than99?"), The results clearly do not show such impruvement, In the caseof MMT, increasing the catalyst cuncentrdtion from 100 to 1000 ppm hadno effect. A direct comparison of catalytic ability of the three metalswas not possible since the Statex values were not the sariip each day.The carbon burnout values for the Pb and Zr tests were lower than normally

- 7')

i- --- ----- ~ .

100 '_ _ _ _ _ _ _ _ ____ 1800

9O .. 1 7O

90 U .1700

CD Of

70 2

o LUo

L I---

70 _ 1500

60 A__ _____________ - 1400 10.6 0.8 1.0 1.2 1.4 1.6

EQUIVALENCE RATIO

:=1

Figure 19. Effect of Equivalence Ratio on Carbon Burnout PercenL ForStatex MT 30'-1; 25; Enriched Air, mst"

V

IL

S- = --- - ___ -

TABLE 21

EFFECT OF RESIDENCE TIME ON CARBON BURNOUT PERCENTFOR STATEX 'kT 30,% AT AN EQUIVALENCE RATIO OF 0.85

Residence Time Temperature Carbon Burnout

4 1690_10 95.5

6 1700+10 96.0

, 3 1700+15 96.+

10 1700+20 96.?

Ia

100 0 1-

90

Cr: S51I. I 12 4 6 8 10 12

RESIDENCE TIME (nisec)Figure 20. Effect of Residence rime on Carbon Rlurnuut r'c- erofe, for Stajtf-x

MT 30'' At An Equi valence Ratio of 0,05; li:niperjture of T1700'rC

- 71-

___ _ -- ~ ___ __~u~ mb~i. iU I

expected for these conditions. A spare fuel nozzle WdS used for thesetests and a change in the fuel spray characteristics may be the reasonfor the lower values. It appears that carbon oxidation enhancement willonly be realized by incorporating the catalyst on the individual carbonparticles, Results are summarized in Table 22.

TABLE 22

EFFECT OF HOMOGENEOUS CATALYST ON CARBON BURNOUT PERCENTAT AN EQUIVALENCE RATIO OF 0.85; 25% ENRICHED AIR, T=6 ms

Fuel Temperature 'C Carbon Burnout (/)

Statex MT 30% 1650+20 95.5

Statex MT 30% with 1630+40 95.0100 ppm Mn

Statex MT 30% 1690+10 95.3

StateA MT 30% with 1690+10 95.01000 ppm Mn

Statex MT 30% 1690+10 94.3

Statex MT 30% ith 1690+10 93.81000 ppm Fe

Statex MT 30% 1720+10 92.7

Statex Mf 30% with 1680+10 91.81000 ppm Pb

I

Statex MT 30% 1680+10 9'1.

Statex MT 30% with 1670+10 91.91000 ppm Zr

I

Note: Mn added aq Methylcyclopentadienyl Marrjanece I i, (,b,.rv 1. (!. r,) fFe added as FerrocenePb added a, Lead Acetate 4Zr added as Zirconium Octoate

78 -

d. Carbon Loading

The effect of carbon loading on carbon burnout was alsoexamined. Statex MT carbon slurries of 30 and 50% carbon by weightwere compared at an equivalence ratio of 0.85 and residence time of

6 ms. The results showed that increasing the carbon mass loadingdoes not decrease the burnout efficiency. A slightly higher valuewas reported for the 50% slurry and this can be attributed to thehigner combustor temperature, Again the spare fuel nozzle was usedfor this and subsequent experiments and may be the reason for thelower burnout values than previously reported. Results are summarizedin Table 23.

e. Particle Size

Particle size was studied to determine whether reduced sizecould improve carbon burnout. This test was conducted at an equivalenceratio of 0.85 and a residence time of 6 ms. Comparison of two slurries(Sterling R, 20% with a 75 mv mean particle size and Monarch 1100, 20%with 14 mp particle size) with Statex MT 30%' was planned; however, theMonarch 1100 posed pumping problems due to its high viscosity and datawas not obtained, A Sterling R 30% also posed pumping problems. It wasfound that the smaller particle size of the Sterling R increased the carbonburnout. A value of 97.1 was recorded for the 75 ml.material as compared

to 93.8% for the 300 mv slurry. Results are summarized in Table 24.

f. Operating LFJSC on Carbon Slurry Alone

The LFJSC was successfully operated on carbon slurry alone ontwo occasions. The first occurred when a total loss of JP-10 flow tothe rear nozzle left the combustor running strictly on slurry. Theslurry flow rate was 5.8 g/min, No soot measurements or gaseous emis-sions data were obtained since this was not a planned experiment.However, combustion was very stable with a temperature of 1300+10°C.

After having established that operating on slurry alone waspossible, an experiment was planned to compare the carbon burnoutefficiencies obtained for slurry alone with that for slurry and JP-lO,The combustor was first operated on both fuels to obtain a carbon burn-out measurement at an equivalence ratio of 0.85 and a residence time of6 ms, Once this was completed, the JP-10 flow from the rear nozzlewas gradually reduced while the slurry rate from the front nozzle was in-creased. This continued until the combustor was running solely on slurry(flow rate of 15.8 g/min). All other conditions were maintained as theyhad been while operating on both fuels. Combustion was auairn very stable.This part of the experiment was also to have been conducted at an equiva--lence ratio of 0.85; however, due to an error in the slurry pump calculationcurve, the actual equivalence ratio was 0,95, '-spite the slightly differ-ent experimental conditions, the results indicated that high carbon burnoutefficiencies can be obtained while operating the LFJSC in either of thesetwo modes. The results are summarized in Table 25.

79

TABLE 23

EFFECT OF CARBON LOADING ON CARBON BURNOUT PERCENT FOR STATEX MTAT AN EQUIVALANCE RATIO OF 0.85; 25% ENRICHED AIR, T=6 ms

Carbon Loading (%) Temperature 'C Carbon Burnout (%)

30 1735+5 93.8

50 1790+10 94.8

TABLE 24

EFFECT OF PARTICLE SIZE ON CARBON BURNOUT PERCENT AT ANEQUIVALENCE RATIO OF 0.85; 25, ENRICHED AIR, t=6 ms

Particle TemperatureFuel S4ze (m,,) 0C Carbon Burnout (:

Statex MT 300 1735+5 93.830%

Sterling R 75 1790+5 97.1

20%

TABLE 25

RESULTS OF OPERATING LFJSC ON 30", STATEX MT30% ALONE; 25% ENRICHED AIR, -E=6 nis

Slurry Flow JP-1O Flow Equivalence CarbonRate Rate Ratio Burnout (C)

5.8 8.0 0.85 92.7

15.8 0.0 0.95 9?.7

-80-

C. LimienS/tems Study

To accomplish the objectives of this effort (see Section III C)several meetings were held with our three subcontractors to discussthe proposed criteria for carbon disrersion fuels (Table 26) and todetermine the possible impact of these properties on engine and air-frame design, and missile performance. As noted below it was indi-cated in early discussions that the potential improvement for ramjetpowered missiles was minimal; therefore, most of our subcontractoreffort was concentrated on turbine powered missiles. The results

may be summarized as follows:

1 Range Improvement Study (by Boeing MilitaryAirplar. Division, Seattle, Washington)

A sensitivity study of system performance of carbon slurryfueled subsonic and supersonic missiles was performed using baselinemissions chosen after consultation with ASD/XR's Advanced TechnoloayCruise Missile (ATCM) Program Manager. Both the supersonic ramjetand the subsonic cruise missile were launched at low altitude at 0.55Mach number. The cruise missile remained at lew altitude flying a

medium range mission, while the ramjet missile made a steep climb toan initial altitude of 70,000 feet and thereafter flew a cruise climbprofile at 4.0 Mach number to a relatively long range.

A series of constant missile volume performance calculationswere made for each of the vehicles, using percent carbon loadina as theindependent variable, and range as dependent variable. Parameters werepercent usable fuel (reflecting fuel trapped through clingage to tankwalls and piping) and missile dry weight (reflecting special itemsneeded to pump the fuel, such as positive expulsion systems). Engineperformance estimates were reviewed with the engine subcontractorv,while aerodynamic estimates and performance calculations were provided byBoeing's missile systems organization.

The results of the study are presented in Figures 21 and 22;for security reasons, the range results have been nondimensionalizedusing as a basis the range of a 100'. JP-9 fueled vehicle. Both vehiclesare very sensitive to percentage increments in unexpended fuel, andsomewhat less sensitive to vehicle dry weight. The following remarksare specific to each vehicle.

The ramjet vehicle range improvement is smaller than antic-ipated (less than 15ý') due to the fact that the more densefuel must be accelerated and lifted to high altitude earlyin the mission. Combustion efficiency wab also found tostrongly influence range; a 10- decrease (such as nightresult from incomplete carbon combustion) reduces rangeby 14>.

The turbine powered cruise missile, flying at essentiallylaunch altitude, shows from 20 to 30. ranqe benefits froimuse of carbon slurry fuels.

-- 81

TABLE 26

PROPOSED CRITERIA FOR CARBON DISPERSION FORMULATIONS

Btu Content, lbs/gal, min. 180,000

Pour Point, 'F, max. -65

Viscosity @ 77%F cp 30-60

@ -40'F cp 150-250

@ -65°F cp 300-400

"Yield Stress @ -650F, max. 340 dynes/cm2

Shear Stress @ O.F, mrx. 50 dynes/cri2

Shear Rate

Shear Stress 0 -65-i, max 500 dynes/cmt2

Shear Rate

Clingage, max, wtv 0.5

Stability Tests

- Static Storage OK for b years

- Freeze Thaw OK after '10 cycles @ -65 to 140°F

- Vibration Test OK after 1 hour

- Thixotropy Not to exceed 500 dynes/cm2 after 1 yr.

Pump Test (Type ?) OK after once throughy

nuzzle Test Produces a ,atisfactory pattern for

cumbus;tion

OTHER

-82-

_ __

POW

LAJ 6-4

0 M.

La J

La

w bA

XU

03 9Nvrd 3AIIV'13d •

LLA

LLJa

LA m Cý 0

W °n

U-1J Ce a

CD:

Ln Lam

39"39 NIV13AIV3

A 3- -

_..wLi-

U- =D>

La LaO

Lj~ C) LC)co 0uU

uj LULa-

* C)

30Nn~ 3AIV13d

-83-

---.13

LIi

WC

Lii

:m

C C>

IL-

UL-Li-

00

-JILiii

U--

UU.

L ui

• ,4'

- -

- - U

LnA

wI

LL

U - L A - -

U--

U.'l

-84

A

A brief comparison of pumped vs. positive expulsion systemswas also completed. At this time, Boeing has a clear preference forpumped systems on the basis of volumetric efficiency (fuel volume/available fuel system volume), a view concurred in by AFAPL/RJA.

2 Turbo-Fan Engine System Investigation for Carbon DispersionFuel Program (by General Electric Aircraft Engine Group,

Lynn, Massachusetts)

The General Electric investigation into the potential impact ofcarbon slurries can be summarized within four broad categories:

Fuel handling and controlFuel injection/atomizationCombustor design - life considerationsTechniques/procedures for starting engine

The results of the studies are given below. IFuel Handling and Control

Wide variation in fluid properties with temperature andshear rate are expected. It is not possible to pump and meter byconventional methods with this type of fluid. As a result, a newstate of art fuel control concept wust evolve to handle the proposedmaterials. Methods now under consideration involve use of variablespeed positive displacement pumps to move, pressurize, and meterthe fluid. Control intelligence will be gathered and processed bydigital electronics to produce the proper controlling functions.The entire process is expected to take advantage of the latesttechnology called FADEC (Full Authority Digital Electronic Control).FADEC is being applied to new GE products.

Engine piping is expected to be several tinies normal

diameter to minimize the effect of line loss and may require aprefilled or primed system to eliminate fill lag during the startsequence. A primed system will need to be hermetically sealeduntil launch, and will require explosive valves to allow quickinitiation of the start sequence.

Priming fuel is not necessarily carbon slurry. It can bevaried to accomplish various ignition window requirements. Thistechnology has already been demonstrated from the late 1950's onthe J-85 powered GAM-72 decoy missile.

Fuel Injection/Atomization

Fuel injectors will require a significantly larger flowcapacity than is normal, to minimize restriction pressure at cruisefuel flow levels. As a result, e2,:h injector may require a pres-surizing valve to minimize head effect at minimum fuel flow.

-85-

, .. . . • • :

Tests accomplished at G.E. in the past year indicate thatcurrent state of art blast atomizers can successfully handle andatomize viscous fluids up to 1100 centistokes, which is believed tocover the required range of fuel properties. The upper limit ofviscosity has yet to be determined.

The injection scheme that was investigated here is basedon an existing General Electric small production engine design whichis known to be successful in burning a wide range of fuels down to-53.9'C (-65'F) fuel temperature. Successful experience has beenobtained on Jet A, Marine Diesel, JP-4, JP-5, JP-8, JP-9, and JP-I0.Burner tests in the fuel viscosity range of two (2) to forty (40)centistokes have been carried out under a separate USAF contract,and have demonstrated negligible discernible difference in flam-mability over that range.

This technology is available now and will be demonstrated ina few months in cruise missile size engines,

Combustor Design - Life Considerations

A cruise missile size annular combustor has recently beendemonstrated at G.E. employing the above-described fuel injectors.The combustor is about three inches long and currently meets allcombustion test requirements for a cruise missile turbofan engineburning the liquid fuels described above. This combustor will beon engine test by the end of 1979.

The combustor is currently node of Hastelloy X sheet stock.

It is helieved that carbon slurry fuels, if burned in thiscombustor, will produce a significant increase in flame radiation pro-cducing shell temperatures as much as 800'F above current observations.Hastelloy X is incapable of meeting this requirement. New alloys areavailable which may be able to help overcome this problem. One suchmaterial is MA-956, an oxide dispersion strengthened (ODS) materialwhich can take temperatures as high as 2200°F while maintainina ac-ceptable structural properties. it is available in thin sheets whichcan be formed into a combustor. This material is also being evaluatedon other AEG combustion system components. Thermal barrier coatingsare also available to provide considerable additional relief, especiallywhere higher radiation loads are expected.

Cooling improvements are available, if required. The presentcruise missile combustor is predominantly film cooled - impingement/filmcooling is available for additional cooling. Several of these conceptshave been demonstrated on other AEF products. Other materials will soonbe available offering additional design margin. Ccramics are under con-sideration and being run on combustion tests within General Electric.

The very small surface area of the very short straight throuqhannular approach at General Electric makes the coolinq job significantlyeasier. This technology is demonstrable in 1-3 years.

- 86 -

Engine Starting Techniques/Procedures

Ignition limits with carbon slurries are currently unknown andwill require investigation on an aircraft engine combustor. In theevent that excessive ignition limitations exist, it is possible toovercome such difficulties with a priming fuel that has broad ignitionboundaries. Since a carbon slurry fueled engine will require primingto eliminate fill time lag, the use of a hydrocarbon such as JP-4, JP-1Oor a high altitude ignition fluid such as ethylene oxide is possible.

The sequence of launch will probably be:

- Firing of cranking cartridge- Firing of explosive valves- Admission of priming fuel- Firing of torch igniters- Partial acceleration of the engine

on priming fuel- Smooth transition to carbon slurry- Complete acceleratiop to full power

Timing of this sequence will be controlled to match specific enginedynami cs.

It is also planned to demonstrate a starting procedure on acruise missile engine in the first half nf 1980.

- 87-

' ' ' ' ' . ... .. - ..... % .... .. ... .. .. . • .. .. . i " |4

APPENDIX A

PROPERTIES OF CARBON BLACK

I. Introduction

An annotated bibliography comprising a comprehensive sample ofthe extensive literature and numerous patents on carbon black has beenprepared in support of the current studies of carbon-erriched slurryfuels. Specific topics covered from 1960 to date include particle sizeand structure; surface area, porosity and activity- adsorption phenomena;superficial oxidation; and the rheology of suspensions of carbon black.References to raw materials, manufacture and uses (except in coatingsand inks) are omitted. The state-of-the-art of slurry fuels and thecombustion of carbon black are considered elsewhere.

The bibliography is prefaced by the following brief overview inwhich the topics are discussed in the same sequence as in the bibliogrphy.Copies of many of the articles and patents are available as noted.

When reviewing the literature on carbon black, the more recentcitations should be examined first. This minimizes the chance of beiriqmisled b', concepts that have since been deprecated. lhe appraisal ofstructure by X-ray diffraction and earlier estimates of porosity arecases in point. Also, since the organization of the bibliography isnecessarily only appruxiilldte, subjects in additLion to the one of apparentinmediate interest should be scanned.

II. Abstract

The literature on the properties and scmte of tne uses of carbonblack has been reviewed and an annotated bibl iuqrip!y has been completed.

Some physical and chemical p-operties are nIutible. Particlesof carbon black are extremely smai,, rani ing from '- to 5001 ciincnetrs,and can be seen only with the electron microscope. Their shape variesfrom spheroidal to complex; and the pirticles tend t(; ciluip to(ýethcc--indeed,dispersion of the particles is a major problepi. in the use of carbon black.Surface area is typically in the neighborhood of I00 ,i/j, with- extremesfrom perhaps 5 to 950 m2/g, and 0 to 35. po-ositv. The snurfa.e is usuallyactive with a tendency to react with oxyqen and s0110 otnCer subsances toform surface groups that influence wettahility and p.,1 which vrie-. fro:nacid to alkaline. There is also some free sbrfdce en(:r- that drawsthe particles together. The rheology of suspensio'ns of ,.,h'rf black isstrongly influenced by these surfice charactcristi•c, a!id the s.sensionsare usually non-Newtonian. Viscosity, fo, eAAdIIIlIC, i-S I ,fr .ior ufshear history. The technology of carbon blacks i- (cto, i,. .,!d iV.rentails the adaptation of the peculiarities of rn ur: bl,: k t- tiepreparation of the stable suspensions requlired in thies i!,,itti,.

0.- _

Ill. General Characteristics

The general characteristics of carbon black may be summarizedas follows. Specific properties are considered in subsequent sections.

Carbon black is mostly carbon, of course, with small amounts ofother elements. Tne numbers vary:

C 84-99.7- Lower levels contain more oxygen

0 0.1-11.0 Higher levels are due to oxidaticnduring manufacture or subsequently

H 0.1-0.5 Largely aromatic

S 0.01-1.5 From feedstock

Ash 0.0-1.0 Mostly from quench water, especiallyin furnace blacks

Tar <C.l-l.0 Origin uncertain

Some other characteristics of carbon black are sunrarized inTable A-l. Again, the rather wide spread in the numbers is notable. Thevolatile matier and the variable nIP. Are largely dup to the oyvepn-containingsurface groups.

Table A-l*

Typical Characteristics ofCarbon Blacks Made by Various Methods

Partial Combustion Thermal DecompositionLampBlack Channel Furnace Thermal Acet.lene

Surface Area, BET m2 /g 20-50 100-950 20-550 6-15 •G5

Particle Size nm** 50-120 5-30 10-80 120-500 35-42

Volatile Matter 2-10 4-20 1-6 0.5-1.0 0.5-2.0

pH 6-9 3-5 6-10 7-9 5-8

* Adapted from Ref. 356M8anometers

-89-

IV. Inspections and Standards (Ref. 1)

ASTM standards for carbon blacks are listed in Ref. 1. Referencesto other less official tests are scattered through the articles cited(e.g., 350) including a number -elevant to specific uses, notably rubbermanufacture.

V. Carbon Black in Piqments, Coatings and Inks (Ref. 2-48)

lhe technology of carbon black in pigments, coatings, and inksentails the preparation of a stable dispersion of carbon black in aliquid medium. Garrett (i8) states this objective more explicitly asfol 1 ows:

"Attainment of suitable pigment dispersion re-quires a process that encompasses incorporation (orwetting), deagglomeration, and stabilization (electro-statically and sterically). In order to do this, bydefinition, one must displace the adsorbed contami-nants (air, gases, or water) from the pigment surface(along with subsequent attachment of the resinousmedium), supply sufficient mechanical energy to dis-rupt the pigment agglomerates, and expose the surface

of the particles to the polymer to effect the neces-sary permanent particle separation required for sta-bility."

Some of the concepts thus developed by the pigment users are more broadlyapplicable to other carbon black/liquid systems.

Carbon black properties considered particularly important includeparticle size, surface area, structure (measured by D[P* adsorption) andsurface chemistry. These and other properties of the carbon black interactin rather complex ways. The finer blacks are more difficult to dispersebut form more stable suspensions; area increases with decrea_•irn particlesize and with oxidation; and high DBP numbers (high structure) are oftenpreferred.

Surface chemistry is perhaps most important tnecause the kindand quantity of surface groups and unsatisfied surface bondr nn theparticles influence their tendency to stick together. This mechanism isoperative either directly, or indirectly by reaction of the surface groupswith the liquid vehicle and/or dispersart added to it, Oxygen groupsformed by oxidation are mentioned most frequently. Incidentally, heatingto 35 0 C facilitated deagglomeration in one system exaiiined (40(-) (becauseof reduced viscosity?).

It is customary when using carbon black •is a pigment to firstprepare a paste which is then "let down" by addition uf diluent. It is

* U)-I•B-tyl phthalate

90-

crucial that the paste be carefully made, and it is also most importantthat the dilution be properly accomplished to avoid "shock out" of thepigment.

The technology seems to be quite arty.

VI. Toxicology. (Ref. 49-50)

Except for some chronic nonspecific inflammation of the respi-ratory tract following prolonged exposure (49, 350), carbon black seemsto be nontoxic. Traces of carcinogenic hydrocarbons on some blacks havenot been shown to be active in laboratory animals (50). No adverse effectsof severe exposure incidental to the manufacture of carbon black has beennoted.

VII. Structure (Ref. 51-80)

Carbon black is formed by the thermal decomposition of hydro-carbons in a reducing or oxidizing (partial combustion) environment. Thefine mechanism is not known. However, it seems likely that at an earlystage small spherical particles form as a consequence of dispersion of aliquid feed and/or by neucleation and condensation from the vapor phase.These particles carboni'e to solid microspheres.

Sometimes depending largely on the manufacturing process, itappears that the freshly formed particles collide with one another whilethey are still tacky and stick together. This results in a branchedbeaded structure referred to as an aggregate.

Both microspheres and aggregates comprise concentric uni;wrmlyspaced graphitic planes of carbon atoms. The planes are not fully -')n-tinuous. They tend in the center to surround the nuclei, and near -)eoutside they follow the external contour. This means that an indivilialaggregate is, in fact, a coherent particle--not simply a group of mito-spheres superficially stuck together.

Evidently, the ordering of carbon atoms is greater near thesurface than within the particles because oxidants attack the interiorfirst. The "graphitic" planes in carbon black are typically more than344 picometers apart, which is the maximum spacing below which graphiti-zation can occur. Hence, carbon black is not usually graphitizable.Nevertheless, the term "graphitized" is still applied to carbon blacksthat have bpen heated to a high temperature, say 3000'C.

•ie individual particles of carbon black--microspheres oraggregates--tend to form slumps that may be remarkably stable. Presumablysurface forces are operative in causing individual particles to sticktogether. In any case, because the particles are so very small, it isdifficult to establish the shear gradient necessary to separate them ina liquid suspension. For these reasons it isn't easy to achieve andsustain a suspension of isolated individual particles in a gas or liquid.

Liquid suspensions of carbon black are non-Newtonian and aresaid to be "structured" (see Rheology of Dispersions, Sec. XVI), It

91

appears that the aforesaid clumps sometimes tend to form larger, lessstable groups of clumps, or flocs.

It is interesting to note that despite the frequent referencesto adhesion and grouping of particles, bulk carbon black is mostly space.As produced, actual solids volume may be as little as 2%. Even inpellets, solids volume is typically little more than 30%, and extremelyhigh pressures are required to increase it much more.

Void volume is considered an important attribute of carbonblack by coating manufacturers . It is often measured by noting thequantity of dibutyl phthalate (DBP) taken up by carbon black up to thepoint at which an abrupt increase in viscosity of the mix occurs.

VIII. Surface Area (Ref. 81-96)

Several procedures have been used to estimate the surface areaof carbon black.

Th. BET nitrogen absorption method is often used in thelaboratory and occasionally in the field. Iodine absorption was used formany years as a measure of surface area, but it has limited utility--surface oxides, for example, interfere--and it is now used primarily forproduction control. A variety of other substances and adsorption tech-niques have been used or considered as indicated in the literature cited.

The electron microscope also is employed to estimate the surfacearea of carbon black, particularly since the procedure has been facilitated Iby automation. T' e electron microscope is not well adapted to field useand is limited almost exclusively to the laboratory.

There has been much debate regarding the relative merits of theseseveral methods and how total area should be corrected for porosity(note comments on porosity, Sec. IX). At the moment, BET nitrogen adsorp-tion, CO2 adsorption, and the electron microscope seem to be favored.

Most carbon blacks do not have a notably high surface areaalthough a few oxidized blacks do approach 950 m2 /g (Table 1, p. 3).ASTM surface area limits for various classes of carbon black are shown infable A-2, following.

Until recently measurements of the porosity of carbon black

were questioned because of a suspicion that interparticle capillary con-densatior was included in the apparent porosity (99). Lamond (99,100)resolved this problem and reported 28-32?. porosity for fluffy blacks,0-26' for pelletized blacks, and 0-5' for rubber grade furnace blacks.

Evidently, carbon black is not usually a very porous substance.There is some thought, incidentally, that oxidation may increase porosity,in addition to selectively burning out the interior of the particles.

Pore shape and size remain uncertain.

- 92 -

Table A-2*

ASTM Designation and Surface Area Limits for Commercial Blacks

ASTOI SA limits AS Olde SA m•1ltsdesignation Older type m2/g designation mI/g

N 110 SAP 125-ISS S 301 MPC 105-125

N 219 ISAF-LS 10S-135 N 440 PP 43-69

N 223 !SAF 110-140 N SSO FEP 36-52

N 242 ISAF-HSb 110-140 W'601 H-F 26-42

N 285S IISAF-HSb'C 100-130 N 660 GPP .26-42

N 326 HAF-LSa 7S-lOS N 770 SRF 17-33

N 330 IRAF 70-90 N 774 SRF-NSe 17-33

N 347 HAP-HSb 80-100 N 880 FT 13-17

S 300 EPC 9S-115 N 990 HT 6-9N 339 HA1-HS(NT)b¢ 90-IOS

3LS - low structure.bHS - high structure.CIISAF - "Intermediate" intermediate super abrasion black.

*. new technology.IS - nonstaining.

* Reference #352

- 93 - I

_

7-

X. Particle Size (Pef. 112-115)

The average size of the ultimate grains of carbon black varieswith the method of manufacture but in all cases is extremely small. Sizedistribution also varies. Finest Larbon blacks are on the order of5-10 nm, while the coarsest '.lacks range up to 500 nm, (Table A-1, p. 91)with a more or less guassian distribution in all cases.

Particle size (and shape and area) is measured with the electronmicroscope. Earlier estimates based on X-ray diffraction are now suspect.

XI. Surface Groups (Ref. 116-182)

Carbon blacks have on the surface functional groups containingoxygen, hydrogen and to a lesser extent sulfur and nitrogen. Thesegroups affect surface energy and activity. The tendency of the particlesto stick together is reduced and wettability may be increased as aconsequence of the presence of the surface groups, which, in turn,facilitates dispersion. Thus, oxidized blacks are hydrophilic anddisperse spontaneously in water.

Contrary to earlier views, the surface groups seem to bechemically bonded---not simply adsorbed--on the surface. Oxygen groupsformed by oxidation that have been identified include carboxyl, phenol,quinone, and lactone, plus possibly anhydride, ether and others that areless certain. Aromatic hydrogen also is present.

The pH of carbon black varies from 3 to 10 and is dictated bythe nature of the surface groups. Oxidation lowers pH and heating,which evidently decomposes the acidic groups preferentially, increases it.

Heating decomposes and drives off the surface groups primarilyas CO, C02 , and H2 . The oxygen compounds are evolved at 400 to 1200'C,but 1500'C with vacuum is needed to drive off hydrogen and sulfur. Dis-sipation of the surface groups leaves some transient, unsatisfied bondswhich can subsequently in the presence of an oxidant reform oxygengroups, or the free bonds may cause the carbon black particles to sticktogether. As noted below, heating also alters the structure of theblack.

XII. Heat Treatment (Ref. 183-192)

Heating drives off surface groups, as noted above, and initiallyincreases surface area. Area reaches a maximum at about 700-900'C, thenat higher temperatures decreases to values that may be less than theoriginal. As noted in Section VII, carbon black is not considered graphiti-zable because the carbon atom planes are too far apart. However, heatingto 2500-3000°C does improve symmetry and extent of the planar structure.

XIII. Grinding (Ref. 193-196) j

Mixtures of carbon black and vehicle intended for coatings andinks are first ground for the purpose of dispersing the black. Evidently,individual particles are not usually much affected by this treatment.

- 94

Devices used include the three roll mill, steel ball mill, carborundumstone mill, sand mill, and high speed impeller (18).

Prolonged grinding of dry carbon black in a steel ball millbreaks up some of the particles. The new surface thus created is ex-tremely active, to the extent that the ground black is pyrophoric (194).Evidently, free ions are formed when the particles are crushed.

XIV. Oxidation (Ref. 197-257)

The superficial oxidation of carbon black may be accomplishedwith any of a number of gaseous or liquid reagents including air, ozone,hydrogen peroxide, carbon dioxide, nitric acid and many others. Initiallythe oxidation occurs primarily within the particle; the outer surfaceis not much affected. This results, therefore, in the formation of ashell with a burned out interior.

As would be expected, oxidation alters the several propertiesof carbon black. Area is increased. Oxygen content (oxygenated surfacegroups) also is increased, and as noted above dispersability is thusimproved. The kind and degree of the side effects of oxidation varydepending on the oxidant used.

Carbon black itself has some activity as an oxidation (andreduction) catalyst, and carbon black and some other forms of carbon areused as catalyst supports (351,352). Oxidations promoted by carbonblack include ethylene to water and C02, ethanol to acetic acid,hydrogen sulfide to sulfur, and ferrous to ferric ion. Carbon blackalso has some activity in alkylation, halogenation, hydrolysis,isomerization and polymerization reactions, and decomposition ofperoxides. Carbon supported metal catalysts have been used forseveral procedures including olefin oxidation, hydrogenation and paraffinisomerization.

XV. Densification (Ref. 258-260)

Carbon blacks are routinely pelletized for convenience inhandling. This reduced void volume from better than 90'' to approximately70% and may reduce surface area. Pressures up to 10000 psi and abovefurther reduce void volume, but the individual particles are remarkablystable; porous blacks may be partially fractured but nonporous blackparticles are largely unaffected by this severe treatment. There is somethought, however, that pressing carbon black so there is substantialparticle-to-particle contact causes chemical bonding. This could bedue to the interaction of free radicals on the surface of the carbon (259).

XVI. Rheology of Dispersions (Ref. 261-329)

The rheology of dispersions of carbon black in liquids iscomplex. It is subject to a number of physical and chemical variablesincluding: particle size, shape, area, porosity, pH, surface activityand surface groups; particle aggregation; quantity of hlack in thesuspension; viscosity, polarity, pH and chemical composition of theliquid; the presence of dispersants, etc. It is not surprising, therefore,

- 95-

ah

that carbon black dispersions are usually thixotropic, with hysteresis,and a "memory" of recent shear experience.

Except at very low loadings of carbon black with vigorous shear(achieved with ultrasound, for example), the individual particles tend toform clumps--the slurry is said to have structure. Fine particles, inparticular, are difficult to disperse; and the viscosity of dispersionsof fine particles is higher than the viscosity of coarser particles.Mixing or grinding reduces but does not necessarily eliminate clumping,and the individual particles are not usually affected. One concept jadvanced is that when mixing is stopped a continuous network of carbonblack clumps forms. With mild agitation (or simply on standing?), thisstructure reverts to discrete flocs (groups of clumps).

The properties of carbon black dispersions are strongly influencedby the surface groups on the particles. Thus, as noted, oxygen ,'roupsseem to enhance and stabilize the dispersion, and adsorbed disp sants,added for the purpose can be effective. Water also may be a factor inasmuchas vigorous desiccation has a negative effect on stability (294) Morebroadly, the properties of the particle surface and the liquid ai.d/ordispersant--acidity, polarity, stereochemistry, etc.--must be carefullybalanced. The references cited include more d6tailed comments.

XVII. Reviews (Ref. 330-363)

Several reviews of some spec7 - and properties of carbonblack and more general coverage are cite extensive article byMedalia and Rivin, 1976 (Ref, 351) and a book by Donnet and Voet, 1976(Ref. 352) are particularly relevant. Excerpts from the tables of contentsof these references are included in the bibliography.

-96-

"Review of Carbon Black Literature

ATTACHMENTS

""1- I. Bibliography

II. Characterization of Powder Surfaces with Special Reference to Pigmentsand Fillers, G. D. Parfitt, K.S.W. Sing edit.; Academic, 1976; CarbonBlacks, A. I. Medalia, D. Rivin; pp. 379-351 (Ref. No. 351)

Contents of Excerpt: page

1. General Considerations ..... ........... 12. Non-carbon Constituents ..... ........... 23. Particle Sizc ...................... 34. Aggregate Structure .................. 65. Surface Functional Groups ..... .......... 66. Oxidation ........ .................. 87. Heat Treatment ......... ..... ......... 108. Catalysis ...... .................. .. 129. Vapor Phase Adsorption .............. .. 12

10. Heats of adsorption ..... ...... 1311. Adsorption from Solution ..... .......... 1512. Chemisorptior. ...... ................ 17

III. Carbon Black, .1 B. Donnet, A. Voet; Plrcel Dekke-, 1976,(Ref. 352)

A. Elemental CompositionB. Measurement of StructureC. MicrostructureD. PorosityE. Shape and SizeF. Surface ,.eaG. Surface GroupsH. OxidationI. Catalytic Action of Carbon BlacksJ. Rheology of Slurries

IV. Nonlinear Rheological Behavior and Shear-Dependent Structure inColloidal Dispersions, G. Schoukens, J. Mewis; J. Rheology, 22, 4381-94, 1978 (Ref. No.261).

W. F. Rollman:snsJuly 26, 1979

- 97 - I

ATTACHMENT I

SelecteM References to Carbon Black

Ref. No.*

I. Standards. ..... .................... 1

II. Uses

A. Pigments

I. Literature. . . ......... ....... .. . 2-52. Patents . . .... .. . ... ... . ... ... . . 6-14

B. Caig

1. Literature. ...................... 15-182. Patents . . . . .................... 19-20

C. Inks

1. Llterature. 21-352. Patents. 36-48

III. _xicoloy. 49-50

IV. Properties

A. Structure 5.1-80

B. Area

1. 'Itterature. 81-932. Patents . . .... . . . .. . .. . . .. . .. . . . 94-96

C.. r sit .. ........ ....... . . . . ............... 97-111D. Size. . . . . . . . . . . . ... 11 2-11 5

E. Heat of Immersion . .. .. . ... .. . .... .. ... 116-125

F. Surface Groups. ...................... 126-169

G. Adsorption Theory ..................... 170-182

* An asterisk denotes copy available.

- 98 -

V. Treatment Ref. No.

A. Heat Treatment ......... ...................... ... 183-192

B. Grinding

1. Literature ......... ....................... ... 193-1942. Patents .......... ........................ .. 195-196

C. Oxidation

1. Atomic Oxygen (0) ............ .................... 1972. Molecular Oxygen (02)

a. Literature ....... ..................... .. 198-207b. Patents ............. ...................... 208-2103. Ozone (03)

a. Literature ....... ..................... ... 211-216b. Patents ............. ..................... 217-2234. Water (H20)

a. Literature ....... ..................... .. 224-225b. Patents ............. ....................... 226

5. Nitric Acid (HN0 3 )a. Literature ....... ..................... ... 227-232b. Patents .......... ....................... 233-234

6. Variousa. Literature ....... ..................... ... 235-245b. Patents .......... ....................... 246-257

D. Compression

1. Literature . . . . . . . . . . . . . . . . . . . . . . . 258-2592. Patents .......... ........................ .... 260

VI. Dispersions and Dispersants

A. Rheology

1 Literature ........................ 261-2812. Patents ........ ........................ . . .. 282

B. Various

1. Literature ............ ....................... 283-3102. Patents .......... ........................ .. 311-329

-99-

- --- - - --.

-.1

VII. Reviews Ref. No.

A. Coatings and Inks ........ ..................... .. 330-334

B. Pigments . ........... ...................... .. 335-336

C. Structure ............. ......................... 337-344

0. Surface ......................................... 345-349

E. General........................................ . 350-363

B10-

7i

-1I00- 1

Pages 101 to 198, which contain the abstracts indexed on pages 98 to 100,have not been included in. this orinted copy because they are lengthy and notof general interest. Cooies of the complete s~t or select abstracts are avail-able from AFWAL/POSF.

i

APPENDIX B

Combustion of Carbon Slurry Fuels

The subject of carbon slurry combustion is exceedingly corn-plex. Superimposed upon the already intricate processes involved in thecombustion of gaseous or liquid fuels is the heterogeneous oxidationof carbon particles which are large compared to ;oot particles. Thediscussion in this section reviews the nature of combustion systemswhich would utilize these new fuels, offers a brief sumnmary of theevents occurring during the combustion of practical liquid hydrocar-bons, and then considers the complications which arise due to theinclusion of carbon particles in the fuel. Since the oxidation ofcarbon particles is expected to lag the consumption of the hyJrncarboncomponent of the fuel, methods to promote carbon oxidation are discussedin a final subsection.

Gas Turbine and Rarmjet Combustion Systems

This subsection descriDes the nature of combustion syste.s it,engines which are candidates for cruise missile propulsion systems,gas turbines and ramjets. It should be noted that mission diffe, encesdrastically influence whether the gas turbine or raniJet encire is selecteCand affects the nature of the engine being designed. For exam•iple, lowsubsonic speed applications clearly favor the selection of a turbinewhile high speed applications favor the ramjet. The description whichfollows intends not to focus on a specific application but rather toprovide general background on the types of systems which will utilizefuture slurry fuels.

Gas Turbine Combustors

The gas turbine employs the Brayton thermodynanic cycle--adiabatic compression, constant pressure heat addition, and adiabaticexpansion. The function of the combustion system is to accomplish theheat release with complete com.,bustion and :iniolum pressure loss and tosatisfy numerous engine operational requiremerts. It must provide ',rthe mixing of fuel and air within the proper environm,:nt to ensure th~eirrearly complete reaction to desirable combustion products. Operation

of the combnstors ir, small gas turbines used in cruise iiissile systemsis adequately described through consideration of Fignre 3-I. In thi"prirnury zone", fuel and oxidizer a-" mixed, usually in slightly fuel-rich proportions. Approximately 90 percent of the fuel is burned inthis zone. Fuel oxidation is completed in the 'spcondary zone'. Inmodern engines, turbine inlet temperatures are close to the tem peratureat which significant chemical reactions cease (-1600"IK) and no dilutionis required. However, older designs with reduced turbine inlet tempera-ture utilize a "dilution zone" to further reduce temperature. Nosignificant reaction occurs within this zone.

The fuel-air ratio typically required for the coiustor ten,-perature increase is less than one-third the stoichiofietrie quantity--that resultiny in complete 02 consumption upon fuel conversion to C02and H2 0. The equivalence ratio parameter, :, defined as thre ratio ofthe actual fuel-air mixture strength to that required fot- stoichium:etriccombustion, ýrovides a convenient way of descnrili g iiit,ure variationsthrough the combustor. Equivalence ratio values ]es5 than 1 3 correSplo,,d

- 199 -

-41

APPENDIX B

Combustion of Carbon Slurry Fuels

The subject of carbon slurry combustion is exceedingqly com-Splex. Superimposed upon the already intricate processes involved in thecombustion of gaseous or liquid fuels is the heterogeneous oxidationof carbon particles which are large compared to soot particles. Thediscussion in this section reviews the nature of combustion systemswhich would utilize these new fuels, offers a brief surnary of theevents occurring during the combustion of practical liquid hydrocar-bons, and then considers the complications which arise due to theinclusion of carbon particles in the fuel. Since the oxidation ofcarbon particles is expected to lag the consumption of the hydrocarboncomponent of the fuel, methods to promote carbon oxidation are discussedin a final subsection.

Gas Turbine and Ramjet Combustion Systems

This subsection descrioes the nature of combustion systeris inengines which are candidates for cruise missile propulsion systems,gas turbines and ramjets. It should be noted that mission differencesdrastically influence whether the gas turbine or ra:imjet enci:e is selectedand affects the nature of the engine being designed. For examiple, lowsubsonic speed applications clearly favor the selection of a turbinewhile hiqn speed applications favor the ramjet. The description whichfollows intends not to focus on a specific application but rather toprovide general background on the types of systems which ,.ill utilizefuture slurry fuels.

Gas Turbine Combustors

The gas turbine employs the Brayton thermodynamic cyclc--adiabatic compression, constant pressure heat addition, and adiabaticexpansion. The function of the conibustion system is to accomplish theheat release with complete combustion and minimum pressure loss and tosatisfy numerous engine operational requirements. It muft provide forthe mixing of fuel and air within the proper environment to ensure thernearly complete reaction to desirable combustion products. Oierationof the combustors in small gas turbines used in cruic ,siiss e system.is adequately described through consideratior mf lni,,;rr L-1 I n the"primary zone", fuel and oxidizer are mixed, us.:11 v in liyhtlyv ,uel-rich proportions. Approximately 90 percent of . fuel is hurned inthis zone. Fuel nxidation is completed ii; the " , c,,,rer, y )nnP,". lnmodern engines, turbine inlet temperatures are Juse to the temLeratureat which significant chemical reactions cease (-16,lGWK.) and no dilutionis required. However, older designs with reduced turbine inlet tempera-ture utilize a "dilution zone" to further reduce temperature. fiosignificant reaction occurs within this zone.

The fuel-dir ratio typica ily requiired for tse st,,:l'u-tor .,i-perature increase is less :.ian one-third the stoichiometri(- q'antitv--that resulting in complete O consumption upon fuel conversio:, to CC2and H20. The equivalence ratio parameter, ,i, defined as tr', r,il i, ofthe actual fuel-air mixture strength to that rpquired fcc steiciniou:tniccombustion, provides a convenient way of descrlirin r.ixt,'r '",riatH)i' sthrough the coibustor. Equivalence ratio values less I on 1,) com;s;,*,,.I

199 -

-a -

I'--

LUU

o. (D

V C)

Iu

S-

:3

200-

to fuel lean operation while those values greater than 1.0 indicate fuelrich operation. Current primary zone equivalence ratios are about onewhereas combustor exit values are less than one-third.

The purpose of the primary zone is to stabilize combustion.High temperatures resulting from stoichiometric operation promoterapid fuel consumption reactions. Primary zone flow is dominated by astrong recirculation region (established by swirling the air enteringthe head end or dome of the burner) which furthers combustion stability.The requirement to ensure an adequate residence time for completion ofchemical reactions is satisfied by limiting combustor reference velocity(the average cold-flow velocity just behind the primary zone) to about25 m/sec.

In practically all current gas-turbine combustors, the fuel isinjected as a liquid. The formation of a well distributed disper-sion of small droplets is desirable to promote rapid evaporation ofthe fuel and intimate mixing of the fuel and air. Three general cate-gories of fuel injectors are currently employed. Pressure atomizersutilize a large fuel pressure drop (greater than 100 psi) across a nozzleto create a finely dispersed spray of small (<50:-) fuel droplets whichquickly vaporize. Airblast atomizers create strong swirling motionsof a small portion of the combustor air flow into which fuel is intro-duced. The severe shearing motion of the air disperses the fuel andresults in small fuel droplets. The slinger design, although a pressureatomizing type, is the third type. It is very different from the con-ventional fuel nozzle in that the fuel is injected through small holesin the rotating turbine shaft. The high centrifugal force imparted tothe fuel provide atomization. Slinger sy.;tems are used in several smallengine combustors of interest to this program--the WRI9 is one suchsystem.

Ihe secondary zone introduces additional a;r to provide forthe chemical reactions which consume the products of incomplete combus-tion passing from the primary zone. Air participating in these chemicalreactions is introduced normal to the main flow direction. The remain-ing air enters parallel to the main flow at the cumbustor walls toprovide a film of cool air which protects the combustor liner and totailor the temperature profile exiting the combustor. Design of thecombustor liner hole pattern to accomplish this requirement traditionallyinvolves a costly development effort to avoi'i a number of possibledetrimental effects. Excessive addition of air may result in quenchiinychemical reactions (especially carbon monoxide and soot oxidation)essential in reducing emissions. Air introduction must be accomplishedin a manner which results in the correct temperature profile enteringthe turbine; a 25 K increase in temperature at a critical region of aturbine blade can result in a four-fold decrease iii blade life. Thesedesign objectives must be met within a prescribed combustor lenqth.Although increasing combustor size might facilitate the design task,this would cause undesirable increases in engine length, main shaft size,bearing requirements and engine weight.

- 201-

-Ii

Ramjet Combustors

The ramjet is the simplest of all air breathing engines. Airenters a diffuser where it is decelerated from a supersonic conditionto a Mach number of about 0.3. This results in substantial compressionprior to fuel introduction and combustion. The hot combustion productsare then expanded through a nozzle to a high exhaust velocity resultingin the generation of substantial thrust. A simplified schematic of aramjet engine is included in Figure B-2.

Ramjet combustors may be of the type depicted in the schema-tic--a design very similar to the turbojet afterburner. A more moderndesign which is of primary importance to the subject program is the"dump combustor". This configuration employs the flow of a premixtureof fuel and air into a sudden duct expansion (see Fiqure B-3). Therecirculation zones created by the "dump" stabilize the combustion pro-cess by providing a continuous ignition source for the incoming fuel.-air flow. The predominant combustion processes occur within theturbulent reacting shear layer at the boundary between the inflow andthe recirculation zone. The dump combustor has gained prominancebecause of its ability to accommodate the rocket/ramjet principle.In this configuration, the initial use of the engine involves operationas a rocket with the solid rocket propellant packing the volume uf thedump combustor; subsequent propulsion is achieved in the ramjet mode.

Key combustion aspects which must be considered in evaluating

the utility of fuel candidates in ramjets are ignition, stability, andcombustion efficiency. It should be noted that ramjet combustionefficiencies are typically 75-95'c much less than the corresponding valuesfor gas turbines operating at cruise operating conditions.

Representations of CombustionProcesses in Gas Turbines and Ramjets

The characterization of fuel behavior in combustion systemscan be considered by evaluating the performance of fuel candidates insmall-scale, well-characterized combustion simulation devices. In thisconnection, turbine and ramjet combustors can each be described ascombinations of such devices, which collectively perform, in the samemanner as the entire combustion system.

Gas turbine combustors are often modeled as a combination ofa well-stirred system which represents the primary zone and a plug flowin series with the stirred system, representing the secondary dnd dilu-tion zone. The fuel vaporization, initial combustion (pyrolysis orbreakdown of larger molecules), stabilization processes, (,O formation,and soot production take place primarily in the well-mixed zone. COand unburned hydrocarbon (fuel) burnout and, most. importantly, soot andfuel carbon burnout take place in the plug flow region wh;hich represcntsthe secondary zone.

Ramjet dump combustors would be represented by well-stirred Iand plug flow zones in an analgous way. The recirculation zones arerepresented by well-stirred regions, while the turhlent reaciirg flowalong the centerline is represented bypluc flov.:.

202-

._ .. . ... ... .. • . ...... . _ = "

J r-Ddrust~~t.~Combustionh dimber Nozzle-

Fw i kt

Combustiocn j Expansion

a0

Supersonic SubtoriSmprm~ion compression

Figure B-2. Schematic Diagram Of A Ramjet Engjine(from Reference 1)

~2t

L U

- - -

Recirculation Zone

PueI Injection

4 Final Coribustion- / •ZoreI ~ ~Air--

Flaimeholder~s

Shear LayeiCombustion

Figure B-3. Ramjet Dump Combustor Schematic

204 -

These well-characterized combustion simulations provide themeans to screen carbon slurry fuel candidates and to investigate theparametric influences of slurry properties on various performanceaspects. Because of the convenience of these experimental tools, a largenumber of candidates can be examined and the parametric variations canbe thoroughly studied.

Hydrocarbon Combustion

An appreciation for hydrocarbon combustion processes must bein hand if carbon slurry fuel combustion is to be understood. Majoraspects of hydrocarbon combustion chemistry involve hydrocarbon pyroly-sis and partial oxidation to H2 and CO, soot formation, chain branching,reactions resulting in H2 consumption, CO oxidation by radicals generatedduring chain branching, and soot oxidation. Each of these reactionsteps is schematically illustrated in Fiaure B-4. Note that the cnrono-logy of these processes is indicated here by the flow fron left to rightthrough the reaction steps. Each process is individually describedbe low.

Pyrolysis and Partial Oxidation

Pyrolysis is the term given tn the process by which fuel mole-cules are broken into smaller fragments due to excessive temperatureand partial oxidation. This molecular destruction is accomplished Iduring the first phase of the combustion process. The predominantresulting products are hydrogen and carbon monoxide. Little detailedinformation is available concerning the chemistry of these processesfor practical fuels--large hydrocarbon, with molecular weights of 50 to200. It is well-recoqnized that hydrocarbon structure and its influenceon the pyrolysis chemistry affects the early stages of combustionprocess. For example, the ignition delay characteristics of varioushydrocarbons differ significantly.

Edelman( 2 ) has developed a single-step quasi-global model tocharacterize the pyrolysis and partial oxidation of any practical hydro-carbon fuel. His approach is to chracterize the kinetics of thenumerous complex chemical reactions resulting in production of H2 andCO by a single reaction step. An Arrhenius type expression has beenfitted to experimental data involving variations in temperature andpressure as well as fuel and oxygen concentration. Edelmar's result is:

d[C n Hml 5.52 x 108 T [Cn H m] 2dt 25

P (1)

LO2Jexp(-24,4OO/RT)

where concentrations are expressed in moles/cc, T in 'K, and P inatmospheres. The activation energy of 24,400 is in the units cal/gmoleK K.

-205-

* IC( I(N(

.r4 a

.5.

E

S-C

L.D

00

U 0

41U

XU

0 0-

20-6

0,

i A

- 206-

Knowledge of hydrocarbon combustion pyrolysis processes isof prime importance to the understanding of slurry fuel combustion.The factors which influence ignition delay and stability in liquidhydrocarbons (those that govern Equation 1) will be the same with slurryfuels since the liquid hydrocarbon will still govern the process.However, as will be discussed in carbon slurry combustion sectionthe presence of the solid carbon material can be expected to influencecommonly encountered combustion behavior.

Soot Formation

The predominance of fundamental research on soot formationhas involved laminar premixed flames. Street and ihomas' work published4n. 1955 is extremely thorough in experimental detail and breadth ofhydrocarbons examined; it has become the classical paper in the field.( 3 )Other key publications are References 4-23. These investigations haveuniversally confirmed that soot formation is a kinet.ically controlledprocess. Equilibrium calculations indicate that soot should not bepresent at fuel-air mixture conditions where the nxyqern-to-carbon atomicratio (O/C) is greater than one. That is, the general chemical equation

Cx Y 02 + - xCO + v- H2 (2)

should define a soot formation threshold. All experiwrrntal IrCsults haveshown soot formation at O/C substantially in excess of unity.

Another very important premixed flame experiicnt cocnducted hyMacFarlane and Holderness at the British National (a[ 1tine Lst(lblish-ment (NGTE) attempted to evaluate the effict oF prc.',rr,formation.(t 18) Most previous work with prerr, ixed fi ,iL.O, u,,i, Lfc.viledatmospheric or subatmospheric conditions,. Tht, C'mpl,,lt i o ...- e( I o',dat MacFarlane and Holderness took special precaitit on. 1(, jroIvo'rof f lash-ing back to upstream locations, an additional difficu-Ity .. ci•,d wilhthe high pressure operation. In addition to suotirn(i 1.ihIt, ivicurl,of soot fu,,,,•d was deter'emlned and expressed as a "'of,. I om,,oa iull or' tio,(the percent of fuel carbon evident as soot.). ' r.u; h (, 1.of,1!)quantity was found to increase with the cube of plro'.;,r-. ý'er" uJ.'flplots of pressure versus equivalence ratio fo.r varin,1.0 1w,', of !'ootformation ratio were presented. Examples are skLwrn it, iqre t;-h t~rcyclohexane, cyclohexene, and benzene. Gs phjr(e Li ..•ri ul,.determined during this testing and it was cncllodd th,•' iO rind i (-ir(oxygenated compounds not predicted by equilibriuw, fijr thi, SVcl.'I (xhy

-02 ` xCO + Y- H2 ) are formed in substantial qjciti i, nil , Hl,!f thiysteni of oxygfen p ri~r to consumption of all fuol , 1;I,.., V,01'o I,) thi',

same organization',19) has indicated that ttc liyJ ir lo i V,,',.i .product plus soot formed was dependont on foe 1 -oi r ratif, ,,I ,oijqhlyindependent of other variables. Differences in ,or,*t. lorm,-" ((j. with,increasing pressure) represent a more effective convcr•ei)n ')i thiv pyroly-sis product to carbon particulate.

207

cyclo Hxmne0 5.1 0.1.0 2.0

271

'I

20 / i

0 6 10 15 20"0 cyclo Hexene

n" \C O .......

2 4 683ý 3.0

2

UJ .0 1- -

I -'---.-- 5-1 201B~enzene3.0

20 .--- I __-. :: 3

00 5 10 15 20

PRESSURE, ATM.

Figure B-5. Pressure dnd Mixture Patio [ffT e ts or lir'ot i ol itll(from Ref.runce 1K)

- ?081-

Wright's work at ER&E involving soot formation in the jet :9stirred reactorW24,25) is probably of greatest interest to this discus- ision--it is a combustion process similar to that at which soot forms in

the recirculation zone of an actual aeropropulsion combustion system.As in the previously mentioned studies, it was determined that sootforms at O/C >1 but the strong backmixing of the jet stirred reactordid seem to afford some broadening of the soot-free O/C ratio. In addi- A7

tion to the establishment of sooting limits, as determined by the colorof the flame (luminous yellow versus blue), Wright determined theconcentrations of soot formed for some limited conditions of O/C belowthe soot limit. No analysis of this "yield" data to determine sootformation kinetics was undertaken but it is recognized that more suchdata might provide the basis for global carbon formation chemical models.

Soot formation processes may be of interest in carbon slurrycombustion if solid carbon is produced which is more difficult to burnout than is the carbon in the slurry. It is unlikely that this willbe the casp as soot particles would be much smaller than those of theslurry, thus requiring less burn time. Secondly, th- Amount of sootresulting from a combustion process is usually less than 2,1' of the fuelcarbon--a small amount compared to the carbon particles present in theslurry. Consequently, a relatively small inefficiency penalty wouldresult even if soot production was high and the soot was riot oxidizedin the combustor.

Carbon Monoxide Oxidation

The products of pyrolysis are reduced state compounds (RSC).Oxidation of these species is better understood than their formation,The important oxidation reactions for the reduced state compounds areof the general form:

RSC + OR = OSC + HR (3)

where RSC L reduced state compoundOR oxidizing radical

USC oxidized-state compoundRR reducing radical

The rate of oxidation of the RSC may he assumed to be given by theappropriate Arrhenious controlled mechanism.

Carbon monoxide consumption is controlled by the followingRSC reaction:

CO + OH - CO2 + H (4)

Since the activation energy of the reaction indicated by F iuation 4 isgenerally low (only a few kcal/g-mole), the carbon monoxide oxidationrate is predominantly Influenced by OH concentration. As previouslynoted, this quantity is controlled by the chain branching ijec.hanism,Nevertheless, a colivnon method of anDroximating radical (oncontratiorlin a RSC reaction involves assuming locdl or partial equilibrium. This

- 209

type of approach has been used in CO oxidation studies by Howard, et al .( 2 6 )

Because the functional relationship between equilibrium OH concentrationand temperature is exponential, an Arrhenius like dependence can bewritten for a quasi-global 02 + CO reaction in the presence of H20.Howard, et al determined:

RT

S I • = k0[C0o 2]I/2 (e) 5,-

where ko is a constant 1.3 x 1014 cc/mole sec.This information is of importance to carbon slurry combustion

because of the need to utilize the chemical energy in CO within thecombustor. Any CO present in the exhaust beyond the equilibrium amountcorresponding to the exhaust temperature represents an inefficiencywhich compromises the benefits gained from using the carbon slurry fueland reduces vehicle range.

H2 Oxidation

While reactions of the nature described by Equatiorn 3 play 3role in consuming the H2 formed during the pyrolysis process, many grosscharacteristics of hydrocarbon combustion are a result of other Otem-ical reactions which involve chain branchinq. This type of reactionsequence involves the production of additional radical species duringthe process. In the case of the H2 oxidation IWoLess, the importantchain branching reactions are:

H2 + 0 - H + OH (6)

H + 02 - OH + 0 (7)

Note that in either reaction a single radical (0 or II) results in theproduction of two radicals (H + OH or OH + 0). This type of reactionhas the potential of producing large quantities of radical species.In portions of the combustion zone having high II2 concentration, radi-cal species can reach levels far in excess of equilibrium. During thisprocess, OH radicals also participate in RSC reactions (Equation 3)to produce H2 0 from H2.

Soot Oxidation

The small carbon (< .05;.ni) soot particulat:s vwhich are formedduring the early stages of the combustion of a V,,dreo.ar',on fuel may heoxidized in the latter portions of the process. Correct conditions ofmixture ratio, temperature, and residence times, can optimi-,e the rateof the soot burnout process.(27-36) Radcliffe and Appleton haveinvestigated the consumnption of soot under conditions correspondingto the gas turbine engine.?35) Appleton points out that the -urfaceoxidation rates for soot and pyrolytic g,^aphite are the sJ,.w tind thatthe rates are predictable by an expres-ior, deveb(,ed !'Y hioc)1,. 1!(1Strickland-Constable. (26)

- 2I0 -

Their results, illustrated in Figure B-6, indicate optimumconsumption of particles at Q0.75.* Their particle surface consump-tion rates of 1-20pm/sec indicate that particles whose initial diameteris less than 0.04om will be consumed in a typical residence time offive ms. Additional detail on carbon particle oxidation will be givenbelow, which focuses on carbon slurry combustion.

Soot burning times are likely to be much less than the timerequired for slurry carbon particle consumption. Nevertheless, theabove discussion provides an idea of the relevant background informa-tion available on carbon particle oxidation under combustion conditions.This information will be of value in our consideration of slurryparticle burning times as discussed below.

Carbon Slurry Combustion

The size of the carbon particles incorporated into the liquidShydrocarbon carrier to produce a slurry fuel can vary significantly.It is possible that micro-fine particles--roughly the size of sootparticles--might be utilized. In this case, it would seem that burningtimes of carbon particles would be on the order of 5-10 milliseccnds.It should be noted, however, that the carbon slurry must burn a sub-stantial portion of its carbon content to achieve the equivalenttemperature conditions at which soot produced from a hydrocarbon flameis oxidized in 5-10 ms. Consequently, much of the carbon will be oxi-dized at lower temperature conditions where consumption tines may bemuch longer than 5-10 ins. This current program will investicatea number of methods to accelerate carbon burnout to allow effectiveutilization of slurry carbon in times which are characteristic ofaeropropulsion devices.

It may also be found that it is necessary to utilize largerParticles (up to l.0-'jm) to improve fuel rheological properties or thatthe micro-fine particles agglomerate to a larger diameter during theformulation process. In either case, oxidation of the carbon. parti-cles will require much more time than that for soot oxidation which isthe slowest process in normal hydrocarbon combustion. However, it mustbe noted that even at l.Oim diameter, the rate of carbon particleconsumption is controlled by surface oxidation kinetics rather thanoxygen diffusion to the surface. Figure B-7 illustrates the combustiontime required for particles of various diameters burned in the stoi-chiometrically correct ratio (i.e. € = 1.0). Tne band of requiredtimes at lower diameters is due to data variation for different typesof carbon tested. An abrupt change in the slope of tne comnbustiontime/diameter characteristic is apparent at about lOi1m,. This corres-ponds to the transition between diffusional and surface kineticcontrol of the particle oxidation rate. At the dianPterr, of interestto carbon slurries (-1.0,m), the process is definitely Kineticallycontrolled. The sub-tantial difference between the extrapolation of

fia

*The pararneter , represents the equivalence ratio - 0(f/a• stoichiorietric.For fuel rich mixtures the fuel-air ratio exceeds the stoichiorntricvalue and >;l. In the case of fuel lean mixtures. ,: 1.

211

.i .

20 - -- --

U,"

II

0

0.4 0O6 0.8 1.0

Eee JIVALENCE RATIO

Figure B-6. Dependence Of Surface Oxidation Rate On Equivalence Ratio i(from Reference 35)

- 212 -

O6 10 310

101

E

10 one second 1.0

S/ D7FFUSION

1-40

10 ..

10-i0- 3 10- 2 10-1 1 10 10 2 10 3 10 4

Intia~l Partice Diameter (pmn)

- onesecond -1.0

Figare B-7 . Dependence Of Carbon Particle Co !Tib usti,)ri Ti!,,(: -'r l'irti cl

213

iii

0A= CoEN

the diffusion controlled behavior and the observed carbon burningtimes at small diameters indicates that much is to be gained by consid-eration of methods for acceleration of carbon particle surface kinetics(i.e. selection of the proper carbon type and catalysis). Considerationof improvements in carbon slurry combustibility by carbon type selec-tion and catalysis will be discussed in the next section.

Figure B-8 describes the overall combustion process for acarbon dispersion fuel. Volatilization of the slurry produces agaseous hydrocarbon vapor and carbon particles. The hydrocarbon vaporthen undergoes a complex process of pyrolysis and partial oxidation.In simplified terms, the principal products of these reactions can beconsidered to be hydrogen, carbon monoxide and soot. The H2 thenundergoes chain branching reactions which provide free radicals (H, 0,and OH) and termination of the process results in the final combustionproduct H20. These radicals participate in and often dominate, otherchemical processes occurring within the combustion process. The jpyrolysis process is influenced by those species, and CO disappearanceis actually controlled by oxidation by' OH. These gas phase processes

are very rapid at the temperatures of interest in aeropropulsiun appli-cation (-2000K) and near complete oxWda:ion is possible in less than2 ms. However, soot oxidation requires heterogeneous oxidation whichis slower than that occurring in the gas phase. Figure B-8 shoois thistime requirement as tI.

Combustion of the carbon portion of the slurry also requiresa heterogeneous reaction. The time required for complete burnout,designated t 2 in Figure B-8, depends upon particle size and combustionconditions. Carbon black of smaller size than soot is not available.Therefore, the release of the carbon's heating value will require moretime than would be thu case for a 100% liquid fuel. Consequently,the carbon oxidation process will lag the consumption of the hydrocar-bon portion of the fuel.

Since carbon oxidation is likely to be the limiting step inthe complete combustion nf carbon slurry fuels, our current emphasisin slurry formulation has been on the use of sub-micron carbon particles(0.01-0.30;) which would be expected to have short burnout times. Theslurry enters the combustion system as liquid droplets containingapproximately 60% by weight of this very small particle size carbon.Depending on the injection technique Ltilized, the average dropleLdiameter may range from about 50 to 150o and, therefore, each dropcontains millions of small carbon particles.

The vaporization of the slurry droplet must be such that thecarbon particles enter the gas phase individually or in small groups.If agglomeration occurs--all the liquid is removed without causing thecarbon particles to disperse--a large, 4C-1201. masF of carbon willremain, as shown in Figure B-9. The burninq rate of such a larce did•leterparticle would be diffusion rate controlled. In this case the combus-tion benefits of providing the small particle size will be lost. Tnfact, if the agglorreration problem cannot be avoided, the practicalityof combusting the slurry carbon content in the short time available(1-5 ms) wi'l be in serious doubt.

214

7- -

r-4z

215

4-,

C.... E

I,)

C")

)

4•)

C.-.

U';I

U 00 0 (-7

/ • o• \ I o .)

(C,

FUEL +VAPOR "

50 mv PARTICLES

J • •~oop _

60%dt C IN SLUPRY FUE

CARBON AGGLOMERATE

CARBON AGGLOMERATE BURN TIME WOULD BE > Isr

Figure B-9. Vaporization Process For A Carbon Slurry Fuel Droplet

-216-

~= ~-~~# --- ~--- -- -• -- - -- -- • .. " ••--•

One approach for solving this problem is to munimize thevolatility difference between the internal (hiah density hydrocarbon)and external (hydrophile) phases of the carbon slurry emulsion. Thiswould result in a droplet fragmentation process which forces individualor groups of carbon particles to separate from the parent droplet.Other approaches might involve some treatment of the carbn narticlesto result in repulsive forces upon heating or even a modified fuelinjection technioue. In any case, it is important to be able Loobserve the droplet vaporization process so that an understanding ofthese approaches may be gained and correlations for formulatirng slurrieswhich minimize agglomeration may be developed.

Two unique types of f'iei injection problems dre encounteredwhen us;ng carbon dispersion fuels. The first involves the potentialprevaporization of the hydrocarbon component of the fuel within hotfuel lines or the fuel nozzle. When 6W% of the fue& is solid carbon,loss of even a small amount of liquid carrier will resu, r in immediateplugging. Consequently, special care must be taken to insure arqainstthe fuel lines or nozzles reaching temperatures where prtruvaporizationcan take place. Further, the initial flow of slurry int,. trr, fuel linesand nozzle must occur with cold hardware and shut down must not allowthe fuel to remain in hardware which becomes hot (i.e. it must i,eburned). 1n this case, deposition of carbon near the hot wall wouldresult and this may eventually result in plum(iing. This pnenomena isanalogous to, but much more dramatic than, the afterburner spray barplugging problem which occurs due to high temperature instability ofliquid jet fuels.

The second problem involves deposition of carbon within thkcombustion system after injection. If the liquid fuel impinges on ahot surface, the liquid portion of the fuel will flash vaporize leavi;n)a deposit of carbon behind. In addition, if the carbon in the coombus-tion zone (after vaporization) is allowed to cony- close to a cn•dersurface within the combustor, the carbon will be attracted to anddeposit on the surface. In either case a substantial carhon dronsitcan build up distorting the aerodynamics of the conm)istion processand/or the hardware cooling scheme. Further problems car, r-.-;ult itthe carbon breaks away from the deposit site in masses lar(.e enouqh toerode downstream engine components. Deposits of soot in combmmstiornsystems have been shown to destroy turbine hardware in thi, mann.-r. (:)

These problems will require special attention in systems wiicih utilizclow pressure air atomization fuel injection methods. Prevent iomn ofcarbon deposition on air swirling devices will t)e especially difficult.

Many of the postulated problems concerning the u);,'busti.nof carbon slurries could be at least scoped if we had a depenpabli,means of predicting carbon particle oxidation rates. hoowrer. theoverall picture now extant of carbon particle ard snot ox,dationleaves much to be desired. A comparison study (of tie it, raturo which,includes measurements of soot oxidation as well as vari,'. -)r:,,. ;"caroon particles, reveals that the range of burninj tiu jý quitelarge, enmcompassing several orders of magnitude (see I mrC;, U-I;,)Even for a specific type of carbon, such as graphite, theo,: is arange of almost two orders of magnitude. Clearly oxidation rate; arein serious disagreement; there are virtually no comidition; taider wh i chwe can assume accurate knowledge of the time fer cmrhen• ,i r ut.

-217-

A

130~ M~C 16 OO 1200 goo 600 400 3100 Ref. Nso. S~ o

03 12n~- Char

10 ASQ4e

III

1213 t S uCo w gure

1E +cd'.)nf I uflcatwd)%ý,

16. 1b I A J1 0 AKSP rphite"I, 18 0 AUi

19 SPI

0 ~' ?0 0"8• Alh c

Sp

A ?2 | •ROIeIi-anth arjýte

14.4 £0 Pui it.hd (cart

?IV 00 SPtei:.j FT?3 AChE ý-n I.C23 V Aches-,n AFC4

I, 1 >

|, AfS gapit

. 6 1 10 12 14 16 18 lal Tjer 1 personal om ,muo icatirn T,37'.S8 I0 1K) 1b) Wouterlood 1pers onal CoMmun icati, b. 13"

Figure B-10. Variability in Available Data on Carbon Oxidation Rates(Reference 38)

- 218 -

Table B-I provides carbon burnout time information for theconditions of interest to cruise missile combustion systems. Informa-tion was extracted from data or correlations were applied for condi-tions of 2000K and an oxygen partial pressure of one atmosphere. Someof this background information recognizes the influence of fuel-airratio on burnout, but much assumes the carbon burning in a non-depletableatmosphere of oxygen--i.e. very fuel-lean. Such information underpredicts the actual required time for burnout. It is evident that thedisagreement for carbon burnout time among these previous works exceeds

an order of magnitude. And the range of uncertainty, from less thanone millisecond to ten, spans time requirements at which efficientslurry combustion would certainly appear to be feasible to times atwhich application of the concept is in serious doubt.

It should be noted that the possibility that burnout timesmay be on the order of one millisecond is extremely encouraging.If burnout time were on this order, the "lag" expected for carbonburnout may be negligible and the combustion process of a slurry maybecome very much like that of a liquid. Consequently, even if aburnout time for carbon of 2-3 ms were determined, motivation toincrease the rate of burnout would remain with the overall objec-tive of making the slurry behave as much like a liquid as possible.It is evident that the objective of developing an optimum carbon slurryfuel will be substantially enhanced by additional study of carbonoxidation chemistry. Oxidation kinetics of the actudl carbon beingutilized in slurry formulations should be established and the impactof various fuel formulation variables (especially catalyst behavior)evaluated.

The two phase combustion (rapid hydrocarbon oxidation andparticulate consumption) of carbon dispersion fuels has four importantimplications. First, it is imperative that the combustion systemprovide sufficient residence time to consume the carbon particles.Since the carbon particles might comprise more than 50,' of the fuelheating value, they must be efficiently consumed to realize the bene-fits of higher volumetric energy content. Secondly, the time delayexperienced between energy release due to hydrocarhon oxidation and thatdue to consumption of the carbon will affect normally encounteredcombustion characteristics. Third, the presence of the cdrbon parti-cles may alter the chemistry of the hydrocarbon pyrolysis process.Finally, the radiative characteristics of flames utilizing carbonslurries are expected to be substantially altered from, those normallyencountered.

Whether sufficient residence time for carhori particle consump-tion can be achieved is dependent on the type of system being consideredand the fuel itself. Table B-2 lists the important combustion inlettemperatures, pressures and velocities as well as the stoichiometryconsiderations for the airbreathing engine systems (,f interest. uLI-stantial variations are evident. Generally speaking;, theý carhbn, parti.-ICoxidation process is assisted by high temperature (near stoichiormetricoperation) and increasing pressure and lean mixture ratios (nigh oxygenpartial pressure).

219

0 °L,,rI,---I 0

0) u1C'

S€Ea- V 4 ' 4

4-)0

o. 4.''1

1,..

V) C m

"IoI4.-'

4 0

0 L)m. 4- c c C)

- ,,%- c I-.

4-I) n U,

or . 4- 1 o•

o ,- ' •X - LI

41 tv 4) 'U 0

o *. S-i

44-) -0C-0

C--0 0 ("U c-. 4.

o 3 u 0 C 0 c- °to 4- S.,- - 0- 0

=- - 0 o 0 0 4- E 4

C 14.- 4- C, to - •'U Cz ' 0 4-0 C) 0 v- - UV (..) -v c C,. 4m '.

D LA. 0 4J.J 0 0 .- ,. 0• ,- 4- L• -.-I: C -C aE - -I;." )<n• •E U €JO ' ' .

0 M- m 0- -V D0 0 0 C) 0

a) f. 7 F= 0 f)U 0- , 0 C 4/)

3.1 M 4-- - . 0 . - - 0I- ..S..,,-.0 u. .J O • S,..) .-4 .-

cc - L 00 0 "0 U 't

< -1 C- M ) 0' 0 to

- C, 4-'0 CL a4' ) x foC.

0 4- 0)U 0_ -) 41 .0 =W = cbi J f 1 4-3 0 00) 'U- ~S- F- 4j 424

0 V C.- .4- -1L CJ' I' .(U S.. _- (L) =1ec I..0 -

L. p - 0. O * - 0 0 a

m0 C.)

a) 0 -

4) 0 J - 1- MI (cc 'U 0. ) to0 . 0 0.

4 - 0-U VC? cUUU ýO L 4-3 0) t

0Ci

- 0 0 a) v94-- C: S- - C

C.) 0o V

4-'a V . )

Lo Of LL co 1-) 0)

'U (..220

!I

CD

00 0

I v

I .--

8-• "0 dJL'

ce-

i Q>

i~

~ -JI- =3 C

S- IT CF0D C

-0

-221-

.1.

The rather obvious fuel parameters affecting consonlptionefficiency are carbon loading and particle size distributin. Inter-particle effects would increase and the temperature at the completionof combustion of the hydrocarbon portion of the fuel would decrease :-

as the weight fraction of carbon in the fuel increased. Particle sizecan generally be considered as proportional to required burn time, as I+

the surface kinetics require a reaction rate (gm cm'2 sec- 1 ) which isindependent of particle diameter. However, the influence of particlesize distribution in a system with minimal excess air (say 4, 1-0.7) is isignificant. A-frequently used example to illustrate the effect ofdistribution in pulverized coal combustion involves the case where amonosize distribution of particles is divided into two portions andone portion is further pulverized to develop monosize particles onehalf of the original size. While the first inclination would be torespond that overall burning time of the new fuel is decra.sed, moredetailed examination indicates an increase in required residenLe time.The smaller particles are consumed first and produce an environmentfor completion of Lombustion of the larger particles which, relativelyspeaking, are starved in oxygen; the result is an overall longer totalburning time. Consequently, minimization of the maximum particle sizeprovide.i the most meaningful reduction in time for conip-ete oxidation.Similar particle distribution characteristics must also be consideredin developing carbon slurries for aeropropulsion applicitions.

The types of combustion characteristics which can be alteredlimtsyFgues B-lilsrtstesiti en andrhstrerecoby the presence of solid carbon in the fuel include the stability

Ilimits. Figure B-11 illustrates the shift in lean and rich stirred reactor

blow out limits which might be expected. Because of the significanttime lag associated with the carbon oxidation process, stability willbe governed primarily by the heat release occurring durin9 hydrocarbonoxidation. This shifts the stability maximum to a condition correspon-ding to stoichiometric based on hydrocarbon only, far richer thanstoichiometric when the entire fuel is considered. Similar changeswould be expected in ignition behavior and stirred reactor loa.ding foroptimum heat release rate.

The presence of the solid carbon may alter the pyrolysisprocess itself and/or the particulates may serve as nucleation sightsfor additional soot formation and agglomeration. This, of course,would cause additional difficulties in consumption of carbon and maylead to lower combustion efficiency (i.e., poor fuel utilization).However, it is unlikely that this affect would be significant sincesoot formation normally consumes less than 2'. of the fuel carbon leadingto only small decreases in combustion efficiency.

Because of the effectiveness of carbon particles as radiatorsof energy, the radiation field within the combustion device burning acarbon slurry may approach that of a black body radiator. This htls asignificant impact on the system hardware and may requirf- im'n.rovedcooling schemes to be applied in the case of a turbine engine. In thecase of a ramjet dump combustor the protective ablatinq liner wouldreceive increased heat flux and additional material might be required.The extent of modification required and the increase in system volumeand weight must be considered in evaluating the net benefit of theslurry in extending system range. i

- 222 -

I

I -M

jA

Stability for

S50/50 Keroaene/Carton Slurry

N 'Typically1.0 - - Observed

Figure B-1i. Stability Considerations In Carbon Slurry Combustion

- 223-

| •_•=-=J . . . --

An additional consequence of incomplete combustion of carbonparticles relates to their effect of the presence in the exhaust plume_.The plume signature from a device not efficiently consuming carbon wouldbe clearly distinguishable from that of a hydrocarbon burning system.Increased total radiation would result from the carbon particles whichbehave as "blackbody" radiators and the radiation would be broader than 1

the C02-band infrared plume radiation normally encountered. In addi- Jtion, the plume would become visible allowing detection by a normaloptical means. Obvious implications with respect to systems intendedto defend against cruise missiles are apparent from this discussion.

Acceleration of Carbon Particle Oxidation

In addition to designing the combustion system to provide tem-perature, and oxygen concentration conditions to optimize the consumptionof carbon particles, the process might benefit through the proper selec-tion of the components of the "fuel package". The size distributionand type of carbon utilized may have a substantial effect and the useof certain metallic elements known to have smoke abatement potentialcan accelerate the carbon oxidation orocess.

The results of Magnussen( 32 ) illustrate the importance ofselecting the proper type of carbon for the slurry application. Hepoints out that the surface re~vction rates for soot are more than twoorders of magnitude less than that for "carbon". As previously indi-cated, Appleton was successful in using pyrolytic graphite oxidationdata to correlate soot consumption. Onp ossible explanation fer thiscontradiction given in Appleton's paper? 27) is that tho "ca:-,n" dataused by Magnussen was for the much more reac. ve isotropic forms ofgraphite (such as Reactor graphite).

Carbon appears in an almost infinite .spectrum of shades ofstructural perfection and frequently samples cut from the same specimenexhibit significant differences in their reactivity. This difficultyapplies to catalyzed and noncatalyzed reactions alike. Differencesin reactivity have been related to surface structure. Isotropic formsof graphite appear to provide a laryc number oi carbon atoms in "edge"positions in the more reactive prismatic plane. Pyrolytic graphitewould have carbon atoms embedded within the basal rlane of the particlesurface. Investigation of these cnaracteristics to select the carbontype which optimizes slurry burning rate is an absolutely essentialelement of this program.

The catalysis of the soot oxidation process by various metalsincluding Mn, Fe, Co, Ni, and Ba is well known and some of these com-pounds are currently marketed (usually in the orcqanomrtallic form) as .smoke abatement additives for industrial gas turbine engines burningconventional liquid fuels. Significant reductions in gas turbine smokevisibility have been achieved using small concentrations of these com-pounds (50 ppmw of metal in the fuel). Use of metal conrtair~inq addi-tives in aviation turbines has been very limited due to the impact ofmetal deposition on the turbine after long periods (,50 hrs) of use.In the case of cruise missle ramjets or turbine engines usinq high den-sity fuels, the short system life may allow additive wtilization withoutsignificant impact on engine performance.

- 224 -

.Some question regarding the additive mechanisms remains.(44,53)

Claims that the additive affects the chemical processes leading to forma-tion of ions which play a role in soot formation are supported bythe krnown strong influences of electric fields on flames. The viewthat additives like Mn actually accelerate carbon oxidation, however,is supported by AFAPL experimental results( 5 3) which illustrate thatsmoke emission can be significantly reduced while primary zone flameradiation (primarily due to luminous radiation from soot) is unaffected.It is likely that various additives function with different mechanisms.Naturally, this program is concerned only with those which catal'yze sootoxidation and the Exxon approach will be to focus on such metallicelements.

Table B-3 illustrates data from which the acceleration effectof manganese mVght be calculated. 1hese data were acquired using a T56combustion rig b 3 ) operated at inlet temperatures of 662 and 773K andan exhaust temperature of 1200K. JP-4 with a hydrogen content of 14.5and a JP-4/xylene blend with a hydrogen content of 13.3 were the testfuels. Smoke emission was chstermined for the fuel without additive andwith 50 ppmw manganese as Ci-2, a commercially-marketed smoke abatementorganometallic. Reductions in soot emissions of 8V or more were observed.

It is difficult to predict the effectiveness of these additivesin accelerating the carbon oxidation of a slurry fuel. In the slurryfuel case particles may be an order of magnitude larger and presentin much higher concentrations--only about 2ý; of the fuel carbon is con-verted to soot in a relatively smoky flame while the slurry can have

50% ur more of its carbon as particulates. It should be noted, however,that additional methods for incorporating the smoke additives intoslurries can be envisioned. One particularly attractive method consi-dered in the Exxon approach is the incorporation of metals into thecarbon particles themselves which may provide advantages relative to theconcentration of catalyst near the carbon surface.

A number of workers have investigated, some qualitatively andsome quantitatively, the relative activities of various cdtalysts foraccelerating the oxidation of carbon. These relative activities areby no means universal but depend strongly on the particular experimnen-tal conditions used. Besides the all important condition of temperature,the experimental conditions determine such important parameters assize and porosity of catalyst particles (and thus their surface area),chemical state of the catalyst, intimacy of contact between catalystand carbon surface, and relative amounts of catalyst on the carbon basalplane and prismatic faces.

The particular anion associated with a metal catalyst addedin the form of salt can also have an important effec.t on the case of

Kcarbon oxidation. For example, Nebel and t.ramer 54 ive a relativelictivity series for lead salts, added in the amount of 0.2 mole ', inreducing the ignition temperature of carbon in air as fol ows: CH3C-Obasic bromide :-. Br- > monoxide > basic sulfate - N0 > SO,, - ;. P03-. Theorganic salts gave the greatest catalysis and the phos'lhate acteJ as aninhibitor.

-225-

F=• •=•i i=•• |{• •f• f~7 | i , ,r' -••i ~ ~ i•f ...... ....

IA-,44" • •

(N 4

00'

CD L n

J :J 0

o€c

-g

V) fl) 1-~ 4]

J. -,I

4c ,-

.4- j L ) * ( N

Ls~ 0

7 w -C C3

0-

~ -1 -4 ,-4 '-

Lo.

Vi)

(.4

ii Ai

LJ~~o AO4-'Cj

u.

&j, C.-, 226

This extensive combination of parameters makes it virtually Aimpossible to a priori select the optimum catalyst type/method of pre-paration/carbon type/carbon particle size. Further, it is not possibleto experimentally evaluate the effectiveness of a formulation withouta test at combustion conditions. Consequently, it is necessary to con-duct the combustion testing simultaneously and iteratively with formulationwork. The Exxon approach is to utilize small scale, well-characterizedcombustion devices which allow the accomplishment of such a program.

The above brief discussion of the possible catalytic effectsin carbon oxidation shows that there is significant potential in thisarea to accelerate carbon particle oxidation. As a result of ourpreliminary review of the catalysis of carbon oxidation and our exten-sive expertise in combustion and catalysis, we will have, at ourdi o~al, an extremely useful approach to accelerate the oxidation ofcarbon particles.

J

I

REFERENCES

1. Hill, P. G. and Peterson, C. R., Mechanics and Thermodynamics ofPropulsion, Addison-Wesley Publishinj Company, Reading, Mass.,1965.

2. Edelman, R. and Fortune, 0., "A Quasi-Global Chemical KineticModel for the Finite Rate Combustion of Hydrocarbon Fuels", AIAA

Paper 69-86, 1969.

3. Street, J. C_ and Thomas, A., "Soot Formation in Premixed Flames",Fuel, 34, pp. 4-36, 1955.

4. Arthur, J. R. and Napier, D. H., "Formation of Carbon andRelated Materials in Diffusion Flames", 5th Symposium (ot.) onCombustion, p. 303.

5. Bittmes, J. D., "Formation of Soot and Polycyclic Aromatic Hydr3-carbon in Combustion Systems", J. Air Poll. Cont. Assn., Vol. ?6,Issue No. 2, pp. 111-115, February 1976.

6. Chakraborty, B. B. and Long, R., "The Formation of Soot and Poly-cyclic Aromatic Hydrocarbons in Diflusion Flames lli--Effect ofAdditions of Oxygen to Ethylene and Ethane Respectively as Fuels",Comb. and Fleme, 12, pp. 469-476, 1968.

7. Chakraborty, B. B. and Long, R., "The Formation of Soot and Poly-cyclic Aromatic Hydrocarbons in C2H4 Diffusion Flames", Combustionand Flame, Volume 12, Issue No. 4, pp. 168-170,April 1968.

8. Chakraborty, B. B. and Long, R., "The Formation of Soot and Poly-cyclic Aromatic Hydrocarbons in Diffusion Flames - I and II",Combustion and Flame, Volume 12, Issue No. 3, pp. 226-242, Jun,1968.

9. Dalessio, A., Beretta, F., and DiLorenzo, A., "Kinetics of Forma-tion of Polycyclic Aromatic Hydrocarbons and Soot in PrenrixedFlames", Chem. Ind., Volume 58, Issue No. 7, pp. 525-526, June1976.

10. Daniels, P. H. "Carbon Formation in Premixed Flames", Comibustionand Flame, 4, pp. 45-49, 1960.

11. DiLorenzo, A. and Masi, S., "Chemical Investigations on Soot-Forming Premixed Flames", Termotechica, Volume 29, Issue No. 11,pp. 590-396, l975.

12. Fenimore, C. P. and Jones, G. W., "Comparative Yields of Soot fromPremixed H/C Flames", Combustion and Flame, Volume 12, Issue No.3, pp. i96-200, June 1968.

13. Fristromn, R. M., "Flames as Chemical Reactors", 2nd InternationalSlmposium on Chemical Reaction Dynamics, Padua, Italy, December1975.

- 228-

14. Gill, D. W., "Review No. 182--Luminosity and Soot Formation inHydrocarbon Flames", The British Coal Utilization Research Asso-ciation Monthly Bulletin, Vol. XXII, No. 12, Part II, pp. 487-506,November-.December 1958.

15. Homann, K. H., "Carbon Formation in Premixed Flarmfes", Combustionand Flame, 11, pp. 265-287, 1967.

16. Howard, J. B., "On the Mechanism of Carbon Formation in Flames",12th Symposium (International) on Combustion, pp. 877-887, 1968.

17. MacFarlane, J. J., et al, "Carbon in r'ames", J. Inst. Fuel, Volume

39, Issue No. 305, pp. 263-270, June 196C

18. MacFarlane, J. J., Holderness, F. H., and Whiteher, F. S. E., "SootFormation Rates in Premixed C5 and C6 Hydrocarbon-Air Flames atPressures Up to 20 Atm.", Comb. and Flame, 8, pp. 215-229, 1964.

19. Holderness, F. H. and MacFarlane, J. J., "Soot Formation in RichKerosene Flames at High Pressure", Paper No. 18, AGARD ConferenceP-oceedings No. 125, April 1973.

20. Palmer, H. B. and Cullis, C. F., "Chapter 5: The Formation ofCarbon from Gases", Chemistry and Physics of Carbon, edited byP. L. Walker, 1965.

21. Radcliff, S. W. and Appelton, J. P., "Shock Tube Measurements ofCarbon to Oxygen Atom Reactions for Incipient Soot Formation withC2 H2 , C2 H4 , and C2 H6", MIT Fluid Mech. Lab. Report 713, April 1971.

22. Thomas, A., "Carbon Formation in Flames", Comb. and Flame, Vol. 6,pp. 46-62, March 1962.

23. Wright, F. J., "Effect of Oxygen on the Carbon-Forming Tendenciesof Diffusion Flames", Fuel, 53.4, pp. 232-235, October 1974.

24. Wright, F. W., "The Formation of Carbon Under Well Stirred Condi-tions", 12Lh Symposium on Combustion, pp. 867-875, 1968.

25. Wright, F. W., "Carbon Formation Under Well Stirred Conditions -Part II", Combustion and Flame, 15, pp. 217-222, 1970.

26. Howard, J. B., Williams, G. C., and Fine, D. H., "Kinetics ofCarbon Monoxide Oxidation in Postflame Gases", 14th International

mposium on Combustion, the Combustion Institute, -ittsbu-rh,Penn., pp. 97-51.9--8 6,,1973.

27. Appleton, J. P., "Soot Oxidation Kinetics at Combustion Tempera-tures", Paper No. 20, AGARD Conference Proceedinys No. 125, April1973.

28. Fenimore, C. P. and Jones, G. W., "Oxidation of Soot by HydroxylRadicals", J. Phys. Chem. , Vol. 71, No. 3, pp. 593-13 7, February1967.

229

29. Feugier, A., "Soot Oxidation in Laminar Hydrocarbon Flames",Combustion and Flame, Vol. 19, Issue No. 2, pp. 249-256, October1972.

30. Feugier, A., "Soot Oxidation in Laminar Diffusion Flames", HeatTransfer in Flames, Advances in Thermal Energy, published byScripta Book Co., Washington, DC, pp. 487-494.

31. Lee, K. B., Thring, M. W., and Beer, J. M., "On the Rate of Com-bustion of Soot in a Laminar Soot Flame", Combustion and Flame,Vol. 6, pp. 137-145, September 1962.

32. Magnussen, B. F., "Rate of Combustion of Soot in Turbulent Flamef",13th Symposium (Int.) on Combustion, pp. 869-877, 1971

33. Park, C., Appleton, J. P., "Shock Tube Measurements of Soot OxiditionRates at Combustion Temperatures and Pressures", 9th InternationalShock Tube Symposium, pp. 793-803, July 1973

34, Park, C., Appleton, J. P., "Shock Tube Measurements of Soot OxidationRates,"Combustion and Flame, Volume No. 20, Issue No. 3, pp. 369-377,June 197-3,

35. Radcliffe, S. W., and Appleton, J. P., "Soot Oxidation Rates in GasTurbine Engines", MIT Fluid Mechanics Lab Report No. 71.12, NASA-CR-125404, June 1971.

36. Szargan, P., "The Kinetics of Oxidation of Carbon Black", Chem. Tech.(Berlin), Volume No. 21, Issue No. 8, pp. 460-461, August I6

37. Tomlinson, J. G. and Montgomery, L. N., "Elimination of Turbine Erosionin the T56 Turboprop Engine", American Society of Mechanical Engineers,Paper. 65-WA/GTT-9, 1956.

38. Smith, I. W., "The Intrinsic Reactivity of Carbons to Oxygen", Fuel,Vol. 57, July 1978.

39. Essenhigh, R. H., "Predicted Burning Times of Solid Particles in anIdealized Dust Flame", J. Inst. Fuel, 34, 239, 1961.

40. Mulcahy, M. F. R, and Smith, J. W., "Kinetics of Combustioii of PulverizedFuel: A Review of Theory and Experiment", Rev. Pure and Appl. CHEM,19, 81, 1969.

41. Bryant, J. T. and Burdette, G. W., "Ramjet Fuel Studies - Part ICarbon," NWC TP 4810, Naval Weapons Center, July 1971.

42. Brudford, J. N., and Benard, I. J., "Air Augmented Hybrid TargetMissile," Phase II1, AFRPL-TR-71-137 (UTC 2337-TR2), January 1972

43, Cassel, H. M., "Some Fundamental Aspects of Dust Flames", U.S. I;Department of the Interior, Bureau of Mines, Washington, D. C.(Report of Investigations No. 6551), 1964

I

I'-_ 230... -

44. Bulewicz, E. M., Evans, D. G., and Padley, P. J., "Effect of MetallicAdditives on Soot Formation Processes in Flames", 15th Symposium(International) on Combustion, pp, 1461-1470, 1974.

45. Cotton. 0. H., Friswell, N. J., and Jenkins, 0. R., "The Suppressionof Soot Emission from Flames by Metal Additives", Comb. and Flame,Vol. 17, pp. 87-98, 9171.

46. Feugier, A., "Effect of Metal Additives on the Amount of Soot Emitted

by Premixed H/C Flames", ACS Div. Fuel Chem. Preprints, Volume 22,Issue No. 1, pp. 211-218, February 1977.

47. Feugier, A., "Effect of Alkali Metals on the Amount of Soot Emittedby Premixed Hydrocarbon Flames", Comb. Inst. Eur.Symp., Paper No. 68,pp. 406-411, September 1973.

48. Fredette, P. E., "Effect of Additives on NOx aiid Particulate Emissionfrom Fuel Oil Combustion", Rensselaer Poly. Inst. Thesis, UniversityMicrofilm No. 72-10, 856, 1972.

49. Jenkins, D. R., "Proposed Mode of Action of Alkaline Earth Metals inSuppressing Soot", Comb. Sci. & Tech., Volume 5, issue No. 5, pp. 245-249,July 1972.

50. Jenkins, D. R., Cotton, D, H. and Friswell, N. J., "The Suppressionof Soot Emission from Flames by Metal Additives", Combustion andFlame, Volume 17, Issue No. 1, pp, 87-98, August 1971.

51. Kukin, I., and Benn-tt, R. P., "Chemical Reduction of S03, Particulates,and NOx Emissions", J. Inst. Fuel, Vol. 50, pp. 41-46, 1977.

52. Steffenson. D. M., et al., "Effect of Fuel Additives Study",PB No, 240169, p. 54.

53. Blazowski, W. S. , Jackson, T. A,, Tackett, L. P., and Fahrenbruck,F. S., "Evaluation of Future Jet Fuel Combustion Characteristics",AFAPL-TR-, 1978.

54. Nebel, G. .J, and Cramer, P. L., Ind. Eng. Chem 47m 2393, 1955.

- 231 -

I

APPENDIX C -

PREPARATION OF HANSEN PLOTS

(1) The first step is to mark off a piece of graph paper to simulatea three dimensional cube corner opened out along the 6 D axis. Paperruled every 1/4 in. in pale blue is quite satisfactory. TKe center ofthe coordinate system is located about 1/3 of the way up and right fromthe lower left corner. The 6D scale is marked on both the left and loweraxes, usually at the rate of one unit per inch. The 6p and 6H scales areput on the left and top axes, respectively, at twice the rate of 6D -

usually 2 units per inch. (The basis for t-his is that 6D seems to inter-act twice as strongly as the others in most cases. The reason probablyis that the London forces are omni-directional and insatiable, whilethe others are more selective so their energy (V62) with a solid oranother liquid is only about (1/2)2 or 25% that of the vapor-liquidinteraction by which 6 is defined.) The effect is that most systemsplot as circLes in the three quadrants, representing the projectionsof a spherical shell dividing the strongest from the weaker inter-actions.

(2) The second step is to divide the data into ranking groups. Thestrongest interaction is assigned a conspicuous symbol, and should beapplied to about 10% of the data. The weakest interaction should thenhave a symbol which is conspicuously different, this bp 1 applied toabout 25% of the data. The three intermediate levels ssiqned othersymbols, taking care to make the second-strongest distinctive from thestrongest.

(3) Start plotting the strongest points, as follows: Locate 6D onthe bottom axis, go right to the 6p value and mark. Then go up to the 6 Hand mark it. Last, go left to the 6D value and mark that. There willnow be a mark in each of the active quadrants, the lower left being theinactive one. This process is repeated for each of the strongly inter-acting points. By now, it should be possible to visualize circles thatwill contain all the points, but do not mark them yet.

(4) Repeat the process with the weakest interaction points. Theseare the ones which define the area outside the circles, but it is nocause for concern if one or more falls right in the "strong" area onone of the quadrants. That merely means it is in the line-of-sight withthe spheric.9l surface in that one direction. The circles may be lightlydrawn in with a compass, remembering that they must all have the sameradius and that the centers must behave as if they were the parametersof a single probe liquid - that is, consistent across all the axes.It is all too easy to uverlcok an inconsistency on 6D.

(5) Continue to plot in the intermediate levels, pausing at eachstage to consider changing the circles to avoid violations and reduce theradius. Unless this is done stepwise, the final plot may be confusingbecause it has so much information.

- 232 -

(6) After the last points are added, it may be desirable to put the .job aside for a few hours, then resume with rested mind and eyes. Thebest possibility may then be drawn in fairly firmly. It is importantthat the circle exclude all but the two strongest classes. However, 1a few violations can be attributed to experimental error, defects inthe parameter tables (not all liquids have been equally studied), or tospecific interactions of an irreversible nature.

(7) If all efforts to resolve the data into a sphere fail, the possibil- Iity of a bimodal system must be considered. These arise in mixtures,some gross as with two powders but others as fine as a copolymer. Sucha case may require two circles, but if the centers are close together anellipse. Since the circles tend to differ in diameter, the "ellipse"may prove to be pear-shaped.

jJ

233-

APPENDIX D

COMPUTER DATA REDUCTION PRINT-OUTS OF LFJSC EXPERIMENTAL RUNS

Computer print-outs of the first year's experimental data are presentedin this appendix. An explanation of the run number indexing is givenbelow,

Run No. Title

100 Slurry Rate Experiments

110 Equivalence Ratio Experiments

120 Residence Time Experiments

S150 Catalyst Experiments - 100 ppm Mn

160 Catalyst Experiments - 1000 ppm Mn

170 Catalyst Experiments - 1000 ppm Fe

180 Catalyst Experiments - 1000 ppm Pb

190 Experiments with Only Carbon Slurry

200 Particle Size Experiments

210 Carbon Loading Experiments

220 Catalyst Experiments - 1000 ppm Zr

Note: A and B denote duplicate runs.

-234 -

I. Detailed mathematical explanation of iterative calculationl oftotal moles flue gas produced and hydrogen mole fraction

Definition of symbols:

AFG - Assumed Mole of Flue Gas

AXH 20 - Assumed Mole Fraction of Water

CH2 or HC - Total Hydrocarbons

CX(I) - Mole Fraction of Species (I) in Flue GasCorrected to a Wet Basis

Delta - Percent Difference between Calculated andAssumed Values

Keq - Equilibrium Constant

MCFR - Moles Carbon from Fuel Mass Rates

MCFG - Moles Carbon from Flue Gas Analysis

MHFR - Moles Hydrogen (H) from Fuel Mass Rates

MHFG - Moles Hydrogen (H) from Flue Gas Analysis

MNFR - Moles Nitrogen (N2 ) from Mass Rates

MNFG - Moles tNitrogen (N2 ) from Flue Gas Analysis

MW - Molecular Weight

MUBC - Moles of Unburnt Carbon

X(I) - Mole Fraction of Species (I) in Flue Gas

W(O) - Fuel Elemental Weight Fraction for Species (I)

Z - Wet Flue Gas Analysis Correct Factor

Step 1: Assume values for moles flue gas (AFG) and mole fraction ofwater (AXH 20)

AFG = total moles of air, oxygen, nitrogen, and fuelsAXH 2 0 = O.0-

Step 2: Determine we, flue gas analysis correction factor

Z = (moles flue gas after water traip.(moT e7slue -as befre wer--t-r a-p)

Z = 1 - AXH 2 0 + humidity of air at 1O0C

Z = 1.0124 -AXH 2 0

235 -

Step 3: Convert dry basis mole fractions (X(I) to wet basis molefractions (CX(I)

cxCO = (xCO) (z)CXC0 2 = (XC02) (Z)CXO 2 = (XO) (Z)CXNO = (XN6) (Z)CXNO 2 = (XNO 2 ) (Z)CXH2 = (XH2) (Z)

Step 4: Calculate the CO mole fraction (CXCO) based on a carbon materialbalance only if the CO analyzed maximum concentration is exceeded.

Moles carbon from Moles carbon fromFuel rates (MCFR) flue gas analysis (MCFG)

R (JP-O mass rate) (WC in JP-lO) + (slurry mass rate)_(WC in slurryCMCFR (-b ms ae) Cin MW carbon

MCFG = (CXCO + CXCO 2 + XCH 2 ) (AFG) + MUDC

CXCO =,MCFR - MUBC - AFG (CXCO 2 + XCH2)

AFG

Step 5: Calculate the hydrogen mole fraction (CXH2) from the water-gas shift equilibrium reaction only if e. measured value wasnot obtained

KeqReaction: CO2 + H2 CO + H20

Keq= CrC_)A(AX ) Where Keq 4.80ýC^X'2) (cx2)

CXH 2 = (CXCO) (AXH2)(U U02--) Ke-)

Step 6: Calculate the water mole fraction (XH2 0) from a hydrogenbalance

Moles hydrogen (H) = Moles hydrogen (H) fromfrom fuel rates (MHFR) Flue gas analysis (MHFG)

MHFR = (JP-l0 mass rate) (WH in JP-l0) + (slurry mass rate) (WH in slurry_MW hydrogen

MHFG = 2(CXH 2 ) 4 X1120 + XCH 2 ) (AFG)

XH2 0 = MHFR - (2)2(AFG) _(C_2 + CH2)()(AFG)

Step 7: Percent difFerence between calculated water mole fraction (XH2 0 )and the assumed value (AXH2O)

Delta H2 0 = Absolute value of (XMH2-AXH20)(-XH2 0 )

Step 8: Compare calculated delta (H2 0) with m1niI'niIu d-eJ't1 data

- 236 -

If the caiculated delta XH 0 is greater than the acceptable delta, thecalculated mole fraction (iH 2 0) becomes the next assumed mole fraction(AXH 2 0) and the iteration is repeated from step 2. If calculated dataXH2 0 is less than or equal to the acceptable delta, this part of theiteration has converged.

Step 9: Calculate nitrogen mole fraction (XN2 ) in flue gas

XN2 1- (CXC0 2 + CXCO + CXO 2 + XCH 2 + CXNO +CXN02 + CXH2 + XH20)

Step 10: Calculate flue gas moles (FG) from a nitrogen (N2) balance

moles nitrogen (N2) moles nitrogen (N2 ) fromfrom flow rates (MNFR) - flue gas analysis (MNFG)

MNFR = moles nitrogen from nitrogen mass rate + moles nitrogen fromair mass rates +

(dP-O mass rate) WN in JP-lO) + (slurr mass rate) (WN in slurryMW Nitrogen (2

MNFG (XN2 + 0.5 (CXNO + CXN0 2 ) FG

.FG MNFRXN2 + 0.5 (CXNO+ CXN0 2 )

Step 11: Percent difference between calculated moles flue gas (FG) andthe assumed value (AFG)

Delta FG = Absolute value of (.G-AFG)G )-

Step 12: Compare calculated delta FG with minimum acceptable delta

If the calculated delia FG is greater than the acceptablie delta, thecalculated values for moles flue gas and water mole fraction become thenext set of assumed values and the iteration is repeated from step 2.If calculated delta FG is less than or equal to the acceptable delta, theentire iteration has converged.

237 - I

____cr -- 7. --

o; r ILcla I

44 CC4 4

3r - 22

-

w k-

c4 ac z

Ll. -3, C-

- , I" : Ca

3 4

1. C; <

o e238

C ttIda C

U M LM

aj 0

z C~ N 0 C -

CD 3a:w

4~ ~ - C5 a

X u

- C

CC -

a> it. a-at

ar to IV

tCL c z. =2r -

0i j I- "f = c

-- 3 0 *r ' a -J (_ 4 c

Nf. W tJL

I rI24 w a or t..42239

I-

CC2 -

4r 20 CLa - w .

* tO -

S - .,-

1- .a~ wo

II

a C w .- , <-U~ ac

S.C_ a. - V".

- A ,U

"4 C' P>

0 -.

o" 7

LA. If

-i 4 -

:

- ' 2 -C

73j Z, a

C

a0 ca t p. c.a ~ r...:c. cI.. - It

Lj~ aa %

c a

24 .: C C' -240

Big"la c

o cx

w Q

4

vw * 9

ix .4 L.

4~~ N C

-~ cx Id)Q Md

I -Id

2 MdO

ftw

14 -u0

I q

L -i L v

0n

17 -- fl)0

Z4-7 4

2~ý C

2 -241

SIr

z ir

-~ 0

00

z I

to 44

. - Id.

S•,'• • .w. _

.a -- -

.AI

-' z•. .• -

zz

- 4 ---

-J -p-

0r C nC C-

2' 2'a 4 2

j F4l,~

4; Idr-

c c

'X t- 1*0-d3L

242'

CID4n Li

uJj

Li a-J 2 0 -6 j

1j~W *

4z4

z _j4 a Ij

A ZN

I Z z I

O .- , C >2 222 4 N0

fni

- - ~ L i -c

~~~u Nrv ~£; ,r Lr -1 I

Cr -j 0 0

w -j-

243

4d-

cc 0

4A u 0

4r -u 00 , U49 ZcI 09 x 2

-~ IiJ)t -

2: t---. 4

_j4Ij Ln

Z Oo 4w CLx 0

_j

or *'j N

-0 N 4 4 W

wI w ~ w -JMJ

4 Q0 0L

~fl 4

o ~ ~ U ,- I . C -I

- c6e a-1

tv 74~: c-2:'

4~~ -:0

LZ Z 0 40

a~ C

uj 22 uj orI U

~~~. Q.; -

Rk"-t

coo

2 a c2

2 - 9

vcw a U.'c mj 05 '0 '0 WK

:0 -d -! 20

4 4 W

-J4 ;-, . U

Tj - Do Z~C?

z ~ .

Z I

o M c- co j P:

I.It0 cL C

a z

-j z

245 '

c a

Ic

71 1a W) -

Ir n -C .1 CVu

4z D

-.

0 4 0 -1

o0 4

~~L) -N C4 07

-a -

C: C!

..- ~~r c. a L

~di±cc u'ij4~ a.i Z ~ o 0o

0~ .24 Z.~~ of

c 0 *

c. 49 .J LNC~ (L -

r n. CD0 r z a) c a Cc U

0 C'u2 C V

246

2d 7

0 m cc J 0 -

21 =

- V ~. - a ZO

LA4L

4c t

!I 0

2 ii

o C .

a - ~J-a _j

4 a 2j 4

o 2 0'

L L

a~ ~ >~

2~~~- 0 J~~~~ -%au - - 5

C~ oJ

4r 1'- v 1%, w

- a a cc-11. U, -,0L

2 '4 C 4

a~ _j q Q~ c -i._0- ,j _j Or 7,

r t. c: L ij Or c

)o. 0 _jw a. 5-

4 *~ 0 4 5 -04Q 0 _j

CU..

247

I , - - a -I * Ii- . . - . a -~ C

olo

Io If

m Cr -

or - A C o o-

T *i c rIýL c

lt L, C 2

4 CIn

x, C

Cl CC9 Q

2 f:,i L&- O r Id

S

.4 - ~ 2

Ii- I- .. CC.24r,

oA w

w CP-Cc U )

t~~4 - 00 0-9

-C C

310- li uo Tc

0. C 0

a1 L i n ri -J

2~i U-044 S C .2 0.

a~~~ 2>2 ) -

CC

3' -Dz 3[

W0 a

uJL.J -- a

z x

4 -- Z'C., • "

xl I z

• . -- , _ _ja * *

.j- -- . 'C 4

-- - - zo .

Ir.I

z D ul- 2 0-

uJ2

Ir >~ Cl

-~ ~~~~ C> CSA) ' :z -

4-

a z. or I I -9J -J

-J _ -

C' Z CZ -Z-

-i U-

250 A C

10 -C

-z 0

z -, 0

UI

-~ z ~z49 -C

kAI-

on LU

a z4 .. 4z 0N

7I

0

> :D LL N-

cc C 4 -N

Z:~a, z U

2' X

T -a C a L

or Lý a a4 U 2'

251 J4 ~

z

C! Li

4 Z O4 -

.4 ...c 0

01 U. LJ

I - j2 .10A-

0 0

tC

In > Q4

< -j 0.

3r -

V,.

* ~ ~ ~ ~ ~ 0 Z ~ ,'-a .~4 WO .j * '.J 02

C~~ ~7 3B Z- .. .0 0-

cr -2 c 7

Lf Li

U. ft~ - 4:252

z a '

1- r Ix t o w az

4 - 0

cc 0 ac a20 2 cor CL0-7

.1 O ..

U.

v'•J.

o ur

a C t..C U.t P

cc a c

C -- -J C 3 -

•;- - • •_ . -.. 0 4 3 _ _• . : : _.--•.- :. • - ,- • - _ Li• - N ... •-i • - _ 2- - .* _

j U a N7

2 *

-~ 1 -LALA

L. ~~- LiLL) *

c C Li

L CC

.-

c0

253

C tI

Z CC cc0

-z z a

x -K4 6*

a> - CC C #>4 2

a * * *>

-9 a

C wCc -j U5 Z

OttI

a ýj D

a: T c:zC

3* 2. * 4>

c~C -10 .

4 -c

-~C7

4254

o l 0- 2

cccc c

r L

c crI - CU £

7 -7

c c

7 1N

- r C

C (4

u. 7--Zd255

I0 Cr

Li a

2 U 4 Cr4 2 -

UNL�j * -,C- 000 0 -�J - -Cr * 44'4 2I oo� Cr Li -4 UN C '0 I- * tA - - 4= LiCr N

CruN

It-C- - 0I2 4$) uSC -o

=- 0Cr .. J'- CCr

U)4 ft Cr4 (0 Z'.2 2 - -.

3 -z 0' C -

o N 2 As . I-. 04 44' I N UN-'Ia *2'-. -'2 .- �o-4Li C-a at co - - C I Io C - N -Cr rat � - N j

- F- CCC - I .c us 'NJ- 'ru�u.-. C ... * rj -

uatr. - u. - - '0- .4

16 La

0� I-,C I- -

Z 7C

--J

'0 UN

"... Ž - C a - -'7 4.1 �C, * 4.. - 02 N Cr UN- I -- 4L� C - -. 4,- I-,. *) I - 2' *-> -.-� -a -� i 4 0 402 0- UN) .4 0 -' o a C $1Sr.-, C

- 0 (01' *... ..- a I: -- 4 - C *

S- -s - 2 t-)tu 'N. - 2 C-J z z .4U 1(0 4 C

16 4

1 .. fl 4, - * '4 -J C U 4]

0 '7 * - 41 I -2 6. �J -'.1 '3 I 2"3' 4, '.4'7 -i 2'2 ... - - Li Cr- - 2 .� - 2 0 4 I- - .4 r oC� t..j * 7 2 C�i.j C1' flI ,,1 '4 4

4 1.4% ./'

-J

a.1 S 1.. * 11-' .. s� - N C' us-. c ,�; U? � £ �, 2

2

C- *j- I- I.4.1" -

tale- '

4, -'

*17 I.

I' Et � .- rz 4 1,

4 2-4 ' -. -fl-

I/I

uo

•' ~~~6 :D;> 1• 0 '4

- :

L) 4

- v o o-- -~ -'-i

cr in

c , z

,. - e z "

I ILIMr IA

(r 41 - 13- u w-J -. cJ-

Lio ~'L- _j

2 -~ *V~ ~o atlS

Ll

L.

.1K

v

SLLMT

.

7 aa

D4 4 Z

-4+-+r • , £ ,- , +. -- .

ua 44

x. =. o <, .4 ,-' 1

j - i

7 6z

i V

a. ,

. rI

I , +.: -+ "a•L J_ +1_

I- , i.I¶0 14 .i

4 k.. ~ L

-�I,.1

o- a

47

LI a

z - LI��J I.J �AW

-, I-

- CCC C �'4 7 44 � *

- ...� I.J �C7 C 74 - 4 '� 4� �J

- -� - - 20I 4L..J C 0

a 4 4

- 4 LJ� 4

4 4 - U

7 - - -I 70 0.

2 2 � a '-

- 0- 4 0I-

C,

a U4 -C a � � C -

L) 2 �� -zC f'.J- 2 A. U

- .- A. 04 �

I.. 7� -cia

2 A-CC -J

-� � �04

C � - J 0 N

277 4 - A. C

- C4� - I 0

O�.J-. 0 -J A.

C- U

-� .- - a

4 A. 7 N 0' U

- - z z * a- a o -

z., � - -� -

* � . A- 4 N - 4

41 .. O .�C -. 02 .1'

I -- �.) 0 � �

f J.'' *-a -C � 4 0 40 7

- C 11) 4 I�J LA �

4 .j .1 U

7 .�a a 0

a a � N - 2 C V

U *' .. J.J ..J .� I..) 0 20 A. U

a � '1 - '- t..I�.

-� -� - �

0 U a � - 4� a2 U 0. 7 4J"�

a �... CI - 4 a- -z 04 -� A '1

0 0 U

- 4:; � -.. - 2 40..J 0 ..J 2 -,

f OX ... Z 4 4 0 4 a.AJO �.1'- '�- �. 0

7.: J� - 0 7

- . U I� - 7

I- C U-' aA' A.. '? I -

A.-JU I 1/ C' . - J'.

- 2 2

U- �I .J ..J U0 a C.

A. - 0

-27 I - -C 4 �. * '1 - L'

a . aC, 0 � c' -x C 0. -

* a� * * '144 LA. c -

.. � -- N

Jf A.�.C0C L.. 'A. A. -

a .� - --� 7 0 .�J *'.

) ,.

* A,

- a 4 � a� C. .. U C

* A, -74.- 1: uC ..� '1.- 4 01.. A. C

'I�A 0 A 4Z U-

- A A.- 0 2

-7 A-.,-- A- U, %-�*A*- �

V C -A0L�. 4-� ,-2

A- A. :�.T?2 -'A-COY - -

A LCA 04 A.,

4c I

- A � -O C 2

AAUA - 772IA,� 7 A. ' I

A U� A 7�A2.A i-� �i__ * -.- � _____ I

W cIw -.

vi w

1740 a

w x

z r- -

ms CC C

4 ~ f O 2C C 0

4 0 4

ar 4 0 .4

o -9clL 0o ~ ~ u -- -40-

-~ 1w 0-

C- 31 I/1A C .

V~~ Ii' a:I~ 1{ ~iI

0 C a

' Q co 4m w

0 L0

0 Lo

4 _j

z 0 z

0 • n

2 - Z

S-. ... 0 0 0 _, C

x z fn

-- 0" _- j0I'

4 --' " I I-" -- 4.- -4w

U"j

oa LmO

4 0,

3lj -z 0 L ]" ''t :: :: 4q 4 00

C~ - -0 2

c Z4 LL - N

5- L - "C s-)~~v - I, .

v I v

- - C SI

19 .4

uj4I

72 L..

44 N

00 0 42

.j-

-r y -- w0

'C N: I- L.J

N L-

crz z

2: -C *C

'tj ý 7 N z.

6 1 -

6n- 0

* U

~ ~ - 4 ' -j I.H: C. z 2 ,22

J LP- 0r c

I0.. N9 4

fr rD

-dC-

cJ 0

14 .4 NC .. L)U C

C. P-C" LP --

7I C -I 0

w~ 0 u

6; aD ý& v 0 -j

tLC - ,

U. C

C 6-

0 c

cc Ir.

< IjxL~~l co o

2 -.

C C . L0 z

7j -j

1 0 -1 Az- c9.- 2 V

4~C Z a0 - 4 2* 0>* 4 '~4 O~ 4 -

4 4 C

-~- L& - 4

L; 7 0 1

V 77 '

or 2A'

z CD

0 COO CP U4 7cl -

4 '0 'Sw0 '

ac IJ4 CC4Lj

in N

z WO

4 W0 0 1=i 0

1 -C ýz Y,- z 0'

J, 4O £

ui~ -' -j u C

:r 41 4 -7

7LJJ C; w :

Z '.2'u L.J

I-)

2r 4WA

C L

4 I'-aI

'Z~~~I 04c0..r .- C

It -i c 2 .J'

-~ .* j C 0

C L~ 2 0 - No

X 7 Lt.J L~ 0 2 (A

27 Z'U -264

cc IIa

L|.+

4k

4.- u2:L

'4 4Z

.. z • .--

_ C ... + - -:

+ z .- :-,. -C

0 "~ 4

2: --

I 0/

-z2

ct

4.j . LI,

:L L00 > ~~: 440.

- U..

- 4 1

a > o.~c C L) c

- r Z~ c

V I~4~ ~ (4 cu :1:1 7

-J ~ U. LX

L4 f 4aN

-zv

4 ) Lt

2 i

-j T 00 CD m2o

IT 10 a 10 d

14 4 U

Z1 LALM0

C, <

4 0 0

uj fx LA-i -

> -C4

s.Ca 0z 4 0 2

C = .- C AX :1 4j D- AJ

.- C 0 ~-I C .~c

4i Lý x LA N.

4 4r Z 44~ ~~ ~ - : 00 -JC -

-. cJ I

0 - 4 '- -72

LAUj 4 (

2: -- -I w t~'ia

C - -s02

. :j 2i -j -1 c- .*~~~~~ a- 7- a4 Z)0 4* 0J0 0

-~~ V, .J - * 2 0*~~c w1'2 4 ' 4 0 ~ U

wi Ca x -.. *

22 U-- C

0~~I 0 IU U.. J0 )- 0 L.J

266

-C?

4,

o 4I

-44Li-~ 2 V,

L)-~ ~ -4

a * 47441->L

4 '- 0 £>

71-

1% -4-

v a- 22 xX :;-. Z' Cc 7

2U x

-r a_.- W L) C'

a -4

LI'i

267

r - 2

U L.J •.

z u L 4

4 - I

co LmC CU4 2 L

T C. .C .co

2 <C w --- . • - -- aO{

CD 4

ID - 0

-: - - .. .a 4

z

a• 0

J . ... . •04 2 0' _

Sc c

az z av 1- 4 "

ui"

L--

f:3

L. r..-.UO 2C

~/.:% NN

7 ~ ~Cý 4 4-

C C- I

C C-. 0

aa

U.;'-~ - :;

LWJ

Zn U4 Z ~q U

tf -- Zocr

z LjM

0

ccr %D L

X ct - o

4:

4 4*~~ * 00LL 4~

L.)

oo

z Iz

cooL. *D C:

It l -9 w

20 Jw T 00L

o 0

2: 0 -f

c .4 11

0. >0

t j~ 0'- -j CE "j c-, 2 D U)

~ 2:-' I o-a2: mI~ Z.J > t

.j

0 s -0 2:

v. Z2: 4 =41 CL 0 4 t

2 z 4. C

Q c

ku z2: a. w. 1: ... cc C, UU ~ ~ ~ ~ c .0J C.. ' 0 it 2

0~~~ 2CC -'..Z, > - ( J ta~_ C,'

2 '- 0 , ~ 2:2 70it

_ _ - - _A

or C0r

Cý T

0 44

2 w

.rc U'

C r Lr -

ZA Z' L4 0-

LA C

U' I< C

2: 00 2

LI

4 w

I I-

271

.1j

w

~AL.M0 C5 cc c.

0 WE

0 0. 0l

2 C.

) - 0s0 .

Or W 42 u

J. j :j z :D v:

2: 0. -Z

.7 0

LI. 0r -

aa I rI N U

- - 4 2 or

0 -4-

a uf w

or 2 5

-- ~~ C:'~ --5! ~5 ~ .U.~ .- 450 , 1-

5- 5- 46 . ~272

- r4 La

o 4A

2 wj

ac - -. c ..

.4z 0 00 rw> .

~~~ La r 0

11. 0 4W

z = La ;

Ir -CD:

2 it wo:

C ~-_0a3

.4 .N

a r-.- C CD

uJ a-

072 In C'. -

2 '-00 4273

o.I-. a

14

4 0 0

atL

> c

x j.-

-n C' 4 - ?

-~I z t

-j x - UC

IA-2L

> 00134-a.

Ln z

c zJ z

a - N

2 tr~-.~ LU C; v,

IA.. 441 < 1~ :27

- -CD a CD -

-. 0 0 0 a O I~ Cq uj

I- 0

cc -j Z

c 4n-~~ CL - j-3: 3 : 2 3 4

'Cn-~c a . e

L ~ 2 3 A -. . , c

[~ z-~I

or C-

4 2r_j c

-27

.11

o a:0 o:

4 _j sx z a: A

z k, 4

z m

4 z

47 c -'3C~

1a c-

276

o

- I

w -coo•, j

SLD

z."z La-

tr

-" A. "- "

- co o -.�.' C• -

- In - -, _ -- 0 2. <

-r a

C;-

rL

L-" a -

D7 U

CL 0

2777

o CL

z %

z -j

2L

*I2 -

r - 0 0 C ic r -

ft >> U

r~t s. zo::ý

U-r z

tr z

-Z Ct- C

U -i '- C C

oj t

6r-CC

o- a

z

w IC

02 2J C

- - co a ,:+,-

C 'C_ 4 a.- . v- - - 2- -

.4 . ,.,- CC7-; + - +

a.. -"-

- 02" . C - ,a 2

-_'-- a- - - -

- a- 'r. - -•

cc

t"4..r

)- -- J 2

C I ... - - C.a-L O' a C2

(C-,.. C

- 279 -

J.~J J -J 2= '

-~ C .1. * t

=1

9'

c a

I- N4 Va 7C

Li U

U (4 4 '32 2 �C ttIL.J * N-JC *�-4 C *

- CCC C i.flC 2

a * * Li

-J I Cot � U- I

c C '- C Ci Ia)

*- 9 I) - - 20 2

a -0 2 -tt.-

Li 0' 4 -�

- a in4 '3 4 Li

I - '3 7

2 in ... ,C =4

o I.1 -I- C

a C LiCc a '3 1

U 23 2 N

t ''Z C C N. U

C) .. N

-� ... N4

CC N. I C V

Li 19 44)0 'C - -

CII V- Ni -J -�

2 '-00 *�i 0-a. ni C

co - - j N -

2 221 d - N --

' 240 - I C 0.

Ii..- C .4 7

2 .. ain - u. - 4

a 0 -4

'p zC 0' 1 in N �

- C 2 2 * C -

C 2 C -J -

2 .5-4 - -J F- V4- K �-<-.4 0' - C P

2 -q L�- jZ * . £2 IC I -- 4a. C U C

_ 2 *-*> 0

-s c a r�N Ci Li * 22 ?� c a u 2

U wU --� :0 C - 0 C

.44 U S 2Z ID 4-c a - i..u * IL

- * 'C -� Lia .c - Ca

-- 4 'i.i 0' 1 tS� .4

2 7w - �- a'- --- 'i -2 C C - i-5..i Li

CC..- 2 * 3

4 -3 c a...�Z *

-. 4- 2 0 -

In C Li *-K Li Ic

- C Li 2K U F-

-4

Li It if C2- )- 'l .. -

U. C

'-C 2 2

CC 4

2'4 i.'

U.- C 2 2

- ,,OCCO'. P0 * k* CC

IN �.C'0'fl� -

4' "-ctC'I - -74 'flC

- a N. -

'a - I C

I'.j - 2 -J .32 -2 -

I/Ic * if -24 .4.

S C Li C C -45 2 - 42CCC 3

'-Z'- -2 Nit 4-(2 N24 4

c '-CC: -c -

CU CL4C C.1 LiZ

27 C C - --'-I.. * 222tC 2 I

'44 1' N

S �u2&0 i*c C2 i*' C 4421&2C' IC

2K I

C -i

Ct* -.

C.

- 2W)

o h'o

- I

2 C --

a ,.

... -C.

c,. T

a -0

c a ro - -0 N'a zar c

_______ -- _---- ]~--- - - -

4- -

z Cor7

311

0l x

2 u -

4, ,-J ,- -

I 9CLl

I ' IIn z

31~ 41 _j % • L

:) IF, -1 : ,zý c>i

C.

a dZ

a L7- - . "i

LJ x N... C . - .P -. "-

-, -.,- --;: t,

-- , ss,,- , + C

az z -d

'3. -C •. -4

35 W _j LL

L...

S,.j ,.4 U -,0 .. o

2~ or. z .a 0 L. 4 '2)

.-.- j a -- _s - ..

K 0.

-) .4 - _ i s • < >-i. -+7 "

c c

282 -

oor

z c

Li x

7 Z C -

-4 t~) * XC -- 4 0

a -. CCC 0orCa~~ ~ *. -

t4 3: Cw LO _4- ZC -S N CA CA L

-~~L .r 2

C C ,

-~_ Li r

y -n 115 07

x Cf t.C /

L, a a TaD4OL L

a L. £- , LiZ CA-

w a

C -- .283

.- aA------

Z 0

L~j C,

0 rzz -00j

4 22:

a *uJ 0

I ze I

0 4

w C; L.I - kýi

0 0 L~-

7 Cj

U~ c 4

L) 0 _j

.J .. N. 02 4 4 )

43 I•M w1

44 LA 0'

2' 0

0 0c x

2 4.'• . w

£~ z "E •

-9 -

4 ,...

) 1.

z V zz

c W

o'- o uu •. L- C-. C

-L_ = -"S, . ._

6.. 44 L"

- I C . • C- 2 i ~ j -- - J *.,7

3r

z z a: -

T ~ C, -J- j

2 c

InI4 0 U- 2 ~

- Z- - ) 0

3..~L :D.,~- 7 .. 0 * -3 0

2 ~ 2 C 4 2

N -c . -a-

V 7

C CLA .) ,22n c4

:rc z o ~V2 *!7~ T C~.. c .' C 2 2 3 ~ - 2 - AJr ~ 4' U '2 U

I -~ 285

S•• 4 -... . .. . _ __... . .. ••-

T,•

:9

Il.-

0a

4 -.J -

..a * )->

o: a,. a: _-

-, ,- - a L. ' .C

I - 2, r

z u-- - " e.

c-

I I.-J C'

--- 28 --

L.. ~ .ala tr- "

- 4Z

-L6 L.~

0 7

Z Z

aw z: z: X a

~t CD *-~ a ac I

L..4 .'.. - z'-

I rt

.. a C Z ~ - -

04 - 0 4~- -. 0

.Z.J ~ 1 Z 4.286

e-

w 1w

4r .4W ~

- I

0 '3 w2: - j

C

zzf. LA C.C2

a~ zaJ~~ 0 ..

W IA.

-L - zp.3u 0 i c

-j -j 02:

LITa 4

LA aj . -- z

Ln cr

a~3 :c~4 .

-~~ -c34

30 U-0

zJ cc Ic I

-U-C

4a- u% 4 0

a: Z a

0 -j- ~ ' .~~ e AZ

a . 4**J -. ;287

C 0

II

U. a

O0 - 4 00 4 za . . >-

-J I o~ ~rij

n fn

- ~tl

v C .

z 2 -

~ '-7..9- N88

u 1

z CC

z 09N

w w

- c

Lrr

z

3: 0j

Li Cr

C L

a2:: 34 -r q -

It, -, .1ý-

P- -LA

c: LA

x -a

L~ja

W. c ' (J.

or zj . a 2-.L'i c~

C4~~ ~ wz. 2 ~P~o '-. -34 :

'a ~ ~ ~ . LL4 ~ :. 0 0L

.1 4 2:,_j --

0j

.42: ~ . ~ ~ 289

-- -� - -- ----

.1

C 1- I- Naa rLi CLu aLI -t-� 2

IALL *ccc a .-- C -- 4. 4 >> Lu�t .� Cr� u - a 7

I -. - - aLu C�- -, If aC C LuS -. S

I tr - 1- -- C:1 a - -- -

CI q .. aIC-, z -4 -

-' Cb-- 4-- 2 N . -�

ZC 2 LI- LiC, 'C

Lfl r�- - * �ji

- -- 4 C� rar a -

CCLS - I -.- -� - -J 4 7- Lu -

tI -

6

a LI

-S

-- C� S -)

7If -' C-. � i14 Lsj4 - � C u

(2a LI c - z �If 74- Lu �j,14 4 -,

I -J � LI U

Cl 1 r- iii- a

-7 -: � C - LI C-C U'I- Z1 7 C7 :1C-. " Lu -- 4- C LuI.. If 44 - - -I - CL C r

C-. I (2 * A2 LI C�-. C -,

raw.-.- 2 N a cCCw

a>

- - -. ,o-- 4 CC4 Li4-- � -Z If j 4

-� z 5 C- iC

- a o �., �

('Lu ;.Z CC I

Lu 4 C>a Lu

Lu, - - Y C Ia �-- .4- c.-��e c -,4) C AZ LICLI. S

- -. C *1 - - 144 -2(2 aLuta * C ZlZ2CC I I.4 T - rc:-fl)-- * NL C 'C f.� 144(0 �72 C If� a c -aLI�z.- LA

CC -C.41 -4- C a2Lu IC 7Lu tiC.., I (5

u. C ;cc: - c a- '-Li (5 I

-290 -

CIDin WC

z af

61 - q

m 0

4 Nw

2A U. -t ' 4~ 2

W0 4

LAO

'~~~ 4zo -C

cc 2v No e4 N- f- 0s=>

~~3 0040

4" 22: 4 -N

- t Z4 C -L

0K 0 UxL

c0 0

291

- --. .or

=OcCr

II

z 2

0.4 'T cc

al 4

z z

- C4~Z

- aj-

4.

c~ D

a '~U.u 292

z0 T

A.'A

Zr -9

CL me C -

- 2j

z 4

c2 4

A.) Zw Z

-J z .

N A.. 2:- .. 0 L&

cc C. NJ1 4omco - 04. ~- -

o - 2 (x

Ln L6 urt I--

zz

0J 7 x cc C,

Z z m tD 4C>a d2 0 4% )-

-z K

229

uii

IAJ

Oj o

L)O

:D Z4 ~41

-z I/W

4 ~, z

00.. 0 I 2 '

-7 a * .

OL 2 42 .. J0L

o -C

I b- . 2q4

REFERENCES

1. Burdette, G. W., Lander, H. R. and McCoy, J. R., Hi e-gy Fuels

for Cruise Missiles," AIAA 16th Aerospace Sciences Meeting, Huntsville,Alabama, January 16-18, 1978.

2. Mil-P-87 107A (U'AF) Military Specification, "Propellant High DensitySynthetic Hydrocarbon Type, Grade JP-9 and JP-I0."

3. USAF Contract No. F3K615-78-C-2025 "Carbon Slurry Fuels" ,Awarded toExxon Research and Engineering Company,

4. Hansen, C. M. arid Beerbower, A. "Solubility Parameters," Encyclopediaof Chemical Technology, 2nd Supp. Vo-9741).7

5. Hansen, C. M., "The Universality of the Solubility Parameter," I&ECProd. Res. and Dev. 8, 2 (1969).

6. Hansen, C. M., "The Three Dimensional Solubility Parameter-Key toPaint Component Affinities," .3. Paint Tech., 39 No. 505 p. 105 andNo. 511 p. 505 and 51; (1967).

7. Graziano, E. R., Cohen, R. E. and Medalia, A I., "Rheology of Con-centrated Suspensions of Carbon Black in Low Molecular Weinht Vphicles"

8. Bryant, J. T. and Burdette, G. W., "Ranjet Fuel Studies - Part I -

Carbon," NWC TP 4810, Naval Weapons Center, July 1971.

9. Ashland Chemical Company, Research and Development Division, ProcessDevelopment Section, Final Report "Ramjet Fuels Analysis Support forthe Supersonic Long Range Missile InteQrationStudy" under ContractNF-f6E4(4 to McDonnell Douglass Astronautics Company-East (i.L';AC-E)(Final Report not yet issued).

10. Nixon, J. and Beerbower, A., "Emulsified Carbon Fuel' U.S. PatentNo. 3,732,084 issued to Exxon Rese-archi-and--E-ngin-e-rir.g Company.

11. Tomlinson, J. G. and Montgomery, L. N., "Elimination of TurbineErosion in the T56 Turboprop Engine," American Snciet' of MechaniclEngineers, Paper 65-WA/GTT-9, 1956.

12. Smith, I.W., "The Intrinsic Reactivity of Carbons to Oxyqen,': Fuel.Vol. 57, July 1978.

13. Essenhigh, R. H., "Predicted Burning Times of Solid Particle, in anIdealized Dust Flame," J. Inst. Fuel, 34, 239, 1961.

14. Nulcahy, M. F. R. and Smith, J. W., "Kir' etics of Cn:'t •r of PulverizedFuel A Review of Theory and Experiment," Rev. Pure ,.d 'Appl. Chin.,19, 81, 1969.

15. Lee, K. B., Thring, M. W., and Beer, J. M., 'On the Rate of Cor.,h:,s-tion of Soot in a Laminar Soot Flame," Comibustion and _Flame, Vol. 6pp. 137-145, September 1962.

295 -

16. Radcliffe, S. W., and Appleton, J. P., Soot Oxidation Rates in GasTurbine Engines," MIT Fluid Mechanics Lab Report No. 71.12, NASA-CR-125404, June 1971.

17. Park, C., Appleton, J. P., "Shock Tube Measurements of Soot Oxidation -

"Rates," Combustion and Flame, Vol. 20, Issue No. 3, pp. 369-377,June 197•7

18. Bradford, J. N. and Benard, I.J., "Air Auqmented Hybrid TargetMissile," Phase III, AFPRL-TR071-137 (UTC 2337-TR2), January 1972.

19. Cassel., H. M., "Some Fundamental Aspects of Dust Flames," U.S.Department of the Interior, Bureau of Mines, Washington, D. C.(Report of Investigations No. 6551), 1964.

20. Griffin, W. C., J. Soc. Cosmetic Chemists, 1, 311 (1949), Ibid 5,) 249 (1954).

21. vanze-, J., "Components of Solid Surface Free Enerqy from WettinqMeasurements," J. Coll. Interf, Sci. 44, 142-161 (1973).

22. Sorenson, P., "The Solubility Parameter Concept in the Forrmulation ofliquid Inks," J. Oil Coll. Chem. Assoc. 50, pp. 236-243 (1967).

23. Nixon, J., Wallace, T. J. and Beerbower, A., "Metallized Fuel Emulsion,"U.S. Patent #3,709,747, January 9, 1973.

24. Shuman, P., "Emulsion Science," Chapter 4, Academic Press, New York(1968).

a

296 -

f'Ub-.GovOrnment Printing 011:,e I JIJ. b1. 1 0a1.,' 6 ;