Stephen Dooley, Saeed Jahangirian, Tanvir I. Farouk, p y, g , ,Joshua Heyne, Amanda Ramcharan, Frederick L. Dryer*

Mechanical and Aerospace EngineeringP i t U i itPrinceton University

Princeton, NJ 08544-5263

MULTI AGENCY COORDINATION COMMITTEE FOR COMBUSTION RESEARCH (MACCCR)FUELS RESEARCH REVIEW

Argonne National LaboratoryArgonne IllinoisArgonne, Illinois

20-22 September 2011

R i f th MURI t t ti th d lReview of the MURI surrogate construction methodologyComparison of 1st and 2nd Generation Surrogates with Jet A POSF 4658

New Combustion Target Comparisons Premixed Laminar Flame TargetsPremixed Laminar Flame TargetsHigh Pressure Speciation Target

SPK Real Fuels and a 2nd Generation Surrogate for S-8 POSF 4734 Contributions of Weakly Branched Iso-alkanes to real fuel behaviorContributions of Weakly Branched Iso alkanes to real fuel behavior.

2-methyl Heptane Experiments vs. Recent Modeling from LLNLTri-methyl Alkyl Isomer Experimental Comparison with an n-Alkane, S-8

POSF 4734 and POSF 4734 surrogate mixture (all with similar DCN) g ( )Further Progress on n-Alkane Modeling and ValidationAromatics – Modeling Progress

n-Propyl Benzene (nPB)1,3,5 Trimethyl Benzene (1,3,5 TmB)

Some Brief Comments on Integration of Physical and Chemical Properties in Surrogate FormulationS d C ti i Eff t

2

Summary and Continuing Efforts

Identify critical fuel property targets that manifest in important practicalMURI Strategy for Modeling a Specific Jet FuelIdentify critical fuel property targets that manifest in important practical combustion behavior of each real fuel:

Adiabatic flame temperature Local air fuel stoichiometryLocal air fuel stoichiometry Enthalpy of combustion Flame velocity Overall active radical production

Ratio of Hydrogen to Carbon (H/C)

Overall active radical productionPremixed sooting Non-premixed sooting F l diff i t t ti

Threshold Soot Index (TSI), By standardizing smoke point measurement

( )

f f

Fuel diffusive transport properties Autoignition/global kinetics

Average Molecular Weight (MWavg)Derived Cetane Number (DCN),Correlative for macro ignition measure

All property target values for real fuel and surrogate components determined experimentally.Surrogate fuel components and mixture do not have to necessarily include all of the molecular

classes found in the real fuel of interest.The components and their mixture must reasonably replicate the interaction and composition p y p p

of distinct chemical functionalities manifested in the real fuel!

3

Test MURI concept by replicating property targets for Jet A POSF 4658Test MURI concept by replicating property targets for Jet A POSF 4658

1st Generation Formulation - n-C10,/iso-C8 /toluene Allow for analyses using more well-developed kinetic model components.y g p p Three component formulation can match H/C and DCN, but not TSI as well, and

can never emulate MWavg of the fuel.2nd Generation Formulation – n-C12/iso-C8/n-PB/1,3,5 TmB C li ll f l ll j f l Can replicate all property targets of nearly all jet fuels. Larger MW components, multiple aromatics (same MW) facilitate property

matching. Significant needs in database/model development and validation, so first: Significant needs in database/model development and validation, so first:

Experimentally compare Jet A POSF 4658 and its surrogates in a wide range of venues (VPFR reactivity, ST/RCM ignition delay, diffusive strained ignition, premixed burning rate and strained extinction, SPST speciation, sooting), and conditions (dilute fuel/O2, fuel/air, range of pressures and temperatures).

4

First reported by UCONN, UIC teams, this review

1200 1000 800 600 Temperature / K

1st and 2nd Gen Comparisons with Jet A POSF 4658

300

400 POSF 4658 1st Gen. POSF 4658 Surrogate 2nd Gen. POSF 4658 Surrogate

ate,

aE

/ s-1

on x

103 /

ppm

4

5

10000

100000 ST RCM

POSF 4658 1st Gen. POSF 4658 Surrogate 2nd Gen. POSF 4658 Surrogate

me,

/ s

1200 1000 800 600

100

200

Ext

inct

ion

stra

in ra O2 CO CO2 H2O

POSF 4658

1st Gen. Surrogate

2nd Gen. Surrogate

Spe

cies

con

cent

rati

1

2

3

100

1000

Igni

tion

dela

y tim

1.0 1.5 2.0 2.5

[Fuel] x Hc x (MWfuel/ MWnitrogen)-1/2 / cal cm-3

500 600 700 800 900 1000

S

Temperature / K0.8 1.0 1.2 1.4 1.6

40

1000K / T

Flow reactor oxidation data for conditionsof 12.5 atm, φ= 1.0 and t =1.8s, for Jet APOSF 4658, 1st gen and 2nd gen POSF

Ignition delay times, φ= 1.0 in air at~20 atm for Jet A POSF 4658, 1st genand 2nd gen POSF 4658 surrogate

Strain rates of extinction for counter flowdiffusion flames at 1 atm, for 1st and 2ndgeneration POSF 4658 surrogate fuels v, g g

4658 surrogate mixtures. All experimentsat 0.3% mole carbon content.

g gmixtures.

1st and 2nd generation surrogates produce nearly identical emulations of the real fuel behavior with significantly different initial surrogate components.C i t t ll diff i l f l b h i d t b h i i

generation POSF 4658 surrogate fuels v.*Radical Index for each fuel

Consistent, small differences in real fuel behavior and surrogate behavior in the hot ignition regime (surrogates very similar).Replicating molecular weight important to diffusive extinction comparisons.

*S H Won S Dooley F L Dryer Y Ju Combust Flame Accepted August 2011

5

S. H. Won, S. Dooley, F.L. Dryer, Y. Ju, Combust. Flame.. Accepted, August 2011.S. Dooley, S. H. Won, J. Heyne, T. I. Farouk, Y. Ju, F. L. Dryer, K. Kumar, X. Hui, C-J Sung, H. Wang, M. A.

Oehlschlaeger,V. Iyer, T. A. Litzinger, R. J. Santoro, T. Malewicki, and K.Brezinsky. Combust Flame (2011). In review.

1250

O2

40

CH

150

C2H4

250

500

750

1000CO CO2

ecie

s co

ncen

tratio

n / p

pm

10

20

30

CH4

peci

es c

once

ntra

tion

/ ppm

50

100

2 4C2H6C2H2

ecie

s co

ncen

tratio

n / p

pm

800 1000 1200 1400 1600 1800

0

250

Spe

Temperature / K

800 1000 1200 1400 1600 1800

0

Sp

Temperature / K800 1000 1200 1400 1600 1800

0

Spe

Temperature / K

40

C3H6C3H4-p (propyne)

20C4H8 (iso- and 1-butenes)1,3 butadiene1 butene-3-yne

2.0

1-hexeneC5H10 (iso- and 1- pentenes)

10

20

30 C3H4-a (allyne)

Spec

ies

conc

entra

tion

/ ppm

5

10

15

1 butene-3-yne

Spe

cies

con

cent

ratio

n / p

pm

0.5

1.0

1.5

Spec

ies

conc

entra

tion

/ ppm

New Combustion Target Comparison -

800 1000 1200 1400 1600 1800

0

S

Temperature / K800 1000 1200 1400 1600 1800

0

S

Temperature / K

800 1000 1200 1400 1600 1800

0.0

S

Temperature / K

6

Toluene Benzene

Single Pulse Shock Tube Speciation Oxidation Speciation of 0.0808/0.158/0.1187 mole %mixtures of C/H/O2 in argon at reaction times of 1.23-3.53 ms. Jet-A POSF 4658 (18-29 atm, hollowsymbols) and 2nd generation POSF 4658 surrogate

2

4

Spe

cies

con

cent

ratio

n / p

pm

symbols) and 2 generation POSF 4658 surrogate(solid symbols).

800 1000 1200 1400 1600 1800

0

Temperature / K

6

S. Dooley, S. H. Won, J. Heyne, T. I. Farouk, Y. Ju, F. L. Dryer, K. Kumar, X. Hui, C-J Sung, H. Wang, M. A.Oehlschlaeger,V. Iyer, T. A. Litzinger, R. J. Santoro, T. Malewicki, and K.Brezinsky. Combust Flame (2011). In review.

6

400

600

400 K 470 K POSF 4658 1st Gen. Surrogate 2nd Gen. Surrogate

rate

/ s-1

70

80

90

100400 K 470 K

POSF 4658 1st Gen. Surrogate 2nd Gen. Surrogate

ocity

/ cm

s-1

200

400

EXt

inct

ion

stra

in r

40

50

60

70

Lam

inar

bur

ning

vel

o

Extinction strain rates for premixed Fuel/O2/N2 at 1 atm:Jet-A POSF 4658 (circles), 1st gen (triangles) and 2ndgen (squares) POSF 4658 surrogates N /(N +O ) =0 86

0.8 1.0 1.2 1.4 1.60

Equivalence ratio,

0.6 0.8 1.0 1.2 1.4

Equivalence ratio in "air",

Laminar burning velocities at 1 atm: Jet-APOSF 4658 (circles), 1st gen (triangles) and2nd gen (squares) POSF 4658 surrogates gen (squares) POSF 4658 surrogates. N2/(N2+O2) =0.86.2nd gen (squares) POSF 4658 surrogates.

New Combustion Target ComparisonPremixed Laminar Flame Properties

Premixed laminar flame properties of Jet-A POSF 4658, 1st and 2nd gen POSF 4658surrogate s at Tu = 400 K. (solid symbols) and 470 K (hollow symbols) in “Air”:N2/(N2+O2) =0.86.

7

S. Dooley, S. H. Won, J. Heyne, T. I. Farouk, Y. Ju, F. L. Dryer, K. Kumar, X. Hui, C-J Sung, H. Wang, M. A.Oehlschlaeger,V. Iyer, T. A. Litzinger, R. J. Santoro, T. Malewicki, and K.Brezinsky. Combust Flame (2011). In review.

60

Group additivity type analysisperformed on 1st and 2nd

generation surrogate fuels shows

30

40

50

ass

%

high degree of Commonality ofChemical Groups on a mass basis.

For example: Benzyl + CH2

0

10

20 Ma

CH

+ CH3n-propylbenzene Benzyl

CH2CH3

2nd Generation Surrogate

CH2 CH

3 Benzyl Cycloalkyl

Molecular Groups1st Generation Surrogate

kyl

Different distinct groups available in each molecular class in the real fuel cancontribute to the sums of all distinct chemical groups found in the composition.

8

g p pS. Dooley, S. H. Won, J. Heyne, T. I. Farouk, Y. Ju, F. L. Dryer, K. Kumar, X. Hui, C-J Sung, H. Wang, M. A.Oehlschlaeger,V. Iyer, T. A. Litzinger, R. J. Santoro, T. Malewicki, and K.Brezinsky. Combust Flame (2011). In review.

Paraffinic Kerosene SurrogatesSPK and HRJ Jet fuel compositions have no aromatics but will likely be mixed withSPK and HRJ Jet fuel compositions have no aromatics, but will likely be mixed with

petroleum derived jet fuel or have synthetic aromatics added for compatibility with legacy engine systems. S-8 POSF 4734 is such an SPK feedstock produced from natural gas.Prior surrogate mixtures for S-8 have been constructed by others previously.

Gokulakrishnan et al. (2007) AIAA 2007-5701Naik et al., (2010), doi:10.1016/j.combustflame.2010.09.016.

Can a simple 2nd generation surrogate mixture (no mono-methyl/di-methyl iso-alkanes) replicate the combustion behavior of S-8 (major amounts of methylated iso-alkanes)?replicate the combustion behavior of S-8 (major amounts of methylated iso-alkanes)?

Di-methyl alkane isomers

CycloalkanesNormal-alkanes

Mono-methyl alkane isomers

T. Edwards, D. Minus, W. Harrison,, E Corporan AIAA 2004 3885 2004

Exemplar generic molecular class composition of S-8

isomers

Select surrogates to represent expected chemical functionalities of fragments formed from S-8 combustion => n– & iso– alkyl and alkenyl for synthetic fuels.

E. Corporan, AIAA-2004-3885, 2004.

Identified components and ~ proportions for S-8 POSF 4734

No utility in matching TSI combustion property as it is very low for all saturated HC’s.

9

S-8 POSF 4734 vs S-8 SurrogateMole Fraction DCN H/C MW / g mol-1

80000 1400 1200 1000 800 600

Temperature / K

Mole Fraction DCN H/C MW / g molS‐8 POSF 4734 58.7±0.3 2.14‐2.19 163.4±15

n‐dodecane iso‐octane0.5189 0.481 58.87 2.19 143.4

ctio

n / p

pm x

103

3

4

5

500 600 700 800 900 10000

10

20

30

40

50 T

T

/ K

Temperature / K 300

400 Jet-A POSF 4658 S-8 POSF 4734 S-8 Surrogate

rate

, aE

/ s-1

10000

80000

S-8 POSF 4734 S-8 POSF 4734 Surrogate

ime,

/ s

S-8 4734 S-8 SurrogateCO CO2 O2

Spe

cies

mol

e fra

c

1

2

100

200

Extin

ctio

n st

rain

100

1000Ig

nitio

n de

lay

ti

500 600 700 800 900 1000Temperature / K

Flow reactor oxidation data for conditions of12.5 atm, 0.3% carbon, φ= 1.0 and t =1.8s forS-8 POSF 4734 (symbols) and S-8 POSF4734 Surrogate (Iines) Inset; ∆T

0.2 0.3 0.4 0.5

Fuel mass fraction, XF

Strain rates of extinction for counter flow diffusion flames at 1 atm, Jet-A POSF 4658, S-8 POSF-4734 and S-8 POSF 4734 Surrogate

0.8 1.0 1.2 1.4 1.630

1000K / T

Reflected Shock ignition delay times,φ= 1.0 in air at ~20 atm for S-8 POSF4734, S-8 POSF 4734 Surrogate.

4734 Surrogate (Iines), Inset; ∆T. POSF 4734 Surrogate.

Similar discrepancies for VPFR reactivity also noted in more complex case of Jet-A POSF 4658, 1st & 2nd gen. POSF 4658 surrogates. S-8 surrogate approximates kinetic and transport properties well; MWavg disparity is less important for MW > 120 g mol-1

S. Dooley, S.H. Won, S. Jahangirian, Y. Ju, F.L. Dryer, H. Wang, M.A. Oehlschlaeger , AIAA Aerospace Sciences Meeting, 2011 & 2012.

important for MWavg’s > 120 g mol .

10

Weakly Branched Alkane Kinetic Models: 2-MethylheptaneTh t i t i lk t i t l d i d j t d di lThe most prominent iso-alkanes present in petroleum derived jet and diesel

fuels contain one or two methyl branches.†Major research gaps exist in modeling weakly branched iso-alkanes.†

2-Methylalkanes are less reactive under auto-ignition2-Methylalkanes are less reactive under auto-ignition conditions than n-alkanes of the same chain length. At low temperature and NTC regions, •RO2

isomerization reactions can be inihibited by where methyl groups are located. Sarathy et al. and co-workers† have recently developed

and validated a detailed kinetic model.The present collaboration‡ is nearing completion and

Structure and C-H bond dissociation energies (kcal mole-1) of 2-methylheptane.

The present collaboration is nearing completion and encompasses VPFR, shock tube, and RCM validations and further model refinements.VPFR Reactivity data show:

S b t ti l l t t ti it d d NTC b h i

† S. M. Sarathy; C. K. Westbrook; M. Mehl; W. J. Pitz; C. Togbe; P. Dagaut; H. Wang; M. A. Oehlschlaeger; U. Niemann; K. Seshadri; P. S. Veloo; C. Ji; F. N. Egolfopoulos; T. Lu, Combustion and Flame (2011). DOI:10.1016/j.combustflame.2011.05.007

Substantial low temperature reactivity and pronounced NTC behavior. Iso-butene and propene are significant intermediates above hot ignition.

( ) j‡ S. Jahangirian, D. Healy, S.M. Sarathy, S. Dooley, M. Mehl, W.J. Pitz, F.L. Dryer, H.J. Curran, C.K.

Westbrook. To appear: ESS 2011, manuscript in preparation.

11

Weakly Branched Alkanes: 2-Methylheptane

16000

20000

24000 CO CO2 O2 2-MH x 10 H2O

on (p

pm)

0.0 0.5 1.0 1.50

30

60

90

120

T (K

)

Time (s) 800

1000

1200

1400 C2H4 CH4 C3H6 CH2O

on (p

pm)

80

100

120 (i-C4H8+1-C4H8) / 3 C4H6 2-C4H8 (cis & trans) 1-C5H10on

(ppm

)

0

4000

8000

12000

Spec

ies

mol

e fra

ctio Time (s)

200

400

600

800

Spec

ies

mol

e fra

ctio

20

40

60

5 10

Spe

cies

mol

e fra

ctio

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.60

Time (s)0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6

0

Time (s)0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6

0

Time (s)

Flow reactor oxidation of 2-methylheptane/O2/N2 at 8 atm and 840 K, time shift: 0.21 s.Model: Sarathy et al. (2011)

Noticeable amounts of 2-methyl alkenes were identified under these conditions. The model predictions are satisfactory for most intermediate species, except sum of iso-

butene and 1-butene and some C3 oxygenates.The model is under further refinement: larger fuel fragments, iso-butene, and C2 chemistry

S M Sarath C K Westbrook M Mehl W J Pit C Togbe P Daga t H Wang M A Oehlschlaeger U Niemann K Seshadri P S

g g 2 ypredictions can benefit from further refinement.

Small fragment distinct chemical functionalities can likely be estimated with n-alkane, iso-octane mixtures.

• S. M. Sarathy; C. K. Westbrook; M. Mehl; W. J. Pitz; C. Togbe; P. Dagaut; H. Wang; M. A. Oehlschlaeger; U. Niemann; K. Seshadri; P. S. Veloo; C. Ji; F. N. Egolfopoulos; T. Lu, Combust Flame. Doi:10.1016/j.combustflame.2011.05.007

12

This iso-alkane n-C7 and S-8 POSF 4734 real fuel all have very similar DCN’s and very lowWeakly Branched iso-Alkanes: a Trimethyl-isomer

This iso alkane, n C7, and S 8 POSF 4734 real fuel all have very similar DCN s , and very low TSI. Here we compare VPFR reactivities for the Tri-methyl isomer and n-heptane

20

25

30

T

/ K

5000

6000

7000

n (p

pm) O2

2000

2500

3000

(ppm

) CO

0

5

10

15

Hea

t rel

ease

, T/

1000

2000

3000

4000

Spec

ies

mol

e fra

ctio

n

500

1000

1500

2000

Spec

ies

mol

e fra

ctio

n

500 600 700 800 900 1000 1100

Temperature (K)500 600 700 800 900 1000 11000S

Temperature (K)500 600 700 800 900 1000 11000S

Temperature (K)

600

700

ppm

) C2H4 200

250

ppm

) C3H6 200

250

ppm

) CH4

100

200

300

400

500

peci

es m

ole

fract

ion

(

50

100

150

peci

es m

ole

fract

ion

(

50

100

150

peci

es m

ole

fract

ion

(

VPFR reactivity data for n-heptane (open symbols) and trimethyl-iso-alkane (solid symbols) ; P= 8 atm with same initial carbon content. The fuels in this study have nearly the same DCN.

500 600 700 800 900 1000 11000Sp

Temperature (K)500 600 700 800 900 1000 11000Sp

Temperature (K)500 600 700 800 900 1000 11000Sp

Temperature (K)

Species time-history experiments trace ~45 species using GC/MS, ~20 species quantified so far.

13

Weakly Branched iso-Alkanes: a Trimethyl-isomer

10000I iti d l i hi 1 20 t 400400

Here we compare the reflected shock ignition delays and diffusive extinction strain rates for S-8 POSF 4734, S-8 Surrogate, and the tri-methyl isomer.All three fuels have nearly identical DCN.

1000

time

[us]

Ignition delay comparisons: phi = 1 , ~20 atm

300

400

rate

aE

[1/s

]

300

400

rate

aE

[1/s

]

100

gniti

on d

elay

t

2,6,10-TMDS-8S-8 Surrogate

100

200

nctio

n st

rain

r

2,6,10 trimethyl dodecane100

200

nctio

n st

rain

r

2,6,10 trimethyl dodecane

x

100.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6

I

1000/T [1/K]

00.08 0.1 0.12 0.14 0.16 0.18 0.2

Exti

Tranport weighted fuel mass fractionY (MW /MW )-0 5

S8 POSF 4734S8 POSF 4734 surrogate, n-dodecane + iso-octaneS8 POSF 4734 surrogate, model

00.08 0.1 0.12 0.14 0.16 0.18 0.2

Exti

Tranport weighted fuel mass fractionY (MW /MW )-0 5

S8 POSF 4734S8 POSF 4734 surrogate, n-dodecane + iso-octaneS8 POSF 4734 surrogate, model

Transport weighted fuel mass fraction

x

St i t f ti ti f t fl diff iYf (MWf/MWn)-0.5Yf (MWf/MWn)-0.5Reflected Shock Ignition Delays for: weakly branched trimethyl isomer (diamonds), S-8 POSF 4734 (squares) & S-8 POSF surrogate (triangles). P ~20 atm, Phi = 1.

Strain rates of extinction for counter flow diffusion flames at 1 atm for branched trimethyl isomer (gray filled circles), S-8 POSF 4734 (flack filled squares) & S-8 POSF surrogate (open circles).

Low temperature ignition of trimethyl isomer are expected to be longer as, p g y p gshown by shock tube results (but DCN’s of all three materials are very similar).

14

Reactivity Predictions for n-Alkanes: n-DecaneNormal-alkane fractions of real fuel are essential to determining radical pool generated

10000

12000 Westbrook et al. Ranzi et al

(a)

1400

1600

Westbrook et al. (b)C2H4CH2O

Normal-alkane fractions of real fuel are essential to determining radical pool generated during fuel oxidation => significant differences exist among model predictions.

CO

6000

8000

10000 Ranzi et al. Biet et al. Bikas & Peters Gokulakrishnan et al. Chaos & Dryer Wang et al. fra

ctio

n (p

pm)

800

1000

1200 Biet et al. Ranzi et al.

fract

ion

(ppm

)

**

*

*No low Temp/NTC kinetics

2000

4000

Naik et al.

CO

mol

e f

200

400

600

Spe

cies

mol

e f *

500 600 700 800 9000

Temperature (K)

500 600 700 800 9000

Temperature (K)Computed species profiles for VPFR reactivity conditions with n-decane/O2/N2 1000/15500/983500 ppm at 1.8 s and

Westbrook CK, Pitz WJ, Herbinet O, Curran HJ, Silke EJ (2009). Combust. Flame 156: 181-199. Biet J, Hakka MH, Warth V, Glaude PA, Battin-Leclerc F (2008). Energy Fuels 22: 2258-2269. Ranzi E (Accessed December 2010): available at: http://www.chem.polimi.it/CRECKModeling/kinetic.html. Bikas G, Peters N (2001). Combust. Flame 126: 1456-1475. Gokulakrishnan P, Gaines G, Currano J, Klassen MS, Roby RJ (2007) J Eng Gas Turbines Power-Trans ASME 129: 655-663. Naik CV, Puduppakkam KV, Modak A, Meeks E, Wang YL, Feng Q, Tsotsis TT, Combustion and Flame In Press

12.5 atm.

Chaos M, Kazakov A, Zhao Z, Dryer FL, Zeppieri SP (2005). US Eastern States Fall Technical Meeting of the Combustion Institute, paper 2549.Wang et al., (September 19, 2010) JetSurF version 2.0, (http://melchioruscedu/JetSurF/JetSurF20)

S. Jahangirian; S. Dooley; F. M. Haas; F. L. Dryer (2011). CombustFlame. DOI:10.1016/j.combustflame.2011.07.002

15

n-Decane Temporal Speciation Predictions 24000 1200

C H ( )

16000

20000 CO CO2 O2

H2O n-decane x 10

ion

(ppm

)

0.0 0.5 1.00

40

80

120

T (K

)

Time (s)

800

1000

tion

(ppm

)

C2H4C3H6CH3CHO

CH2O

CH

(a)

4000

8000

12000

ecie

s m

ole

fract Time (s)

200

400

600

ecie

s m

ole

fract CH4

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.40

4000

Spe

Time (s)

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.40

200

Spe

Time (s)( ) ( )

Detailed temporal fuel speciation during initial fuel destruction significantly constrains

VPFR oxidation of n-decane/O2/N2 1000/15500/983500 ppm at 8 atm and 830 K.

Solid lines: Original Westbrook et al. model; Dashed lines: Updated Westbrook et al. model

Detailed temporal fuel speciation during initial fuel destruction significantly constrains model predictions. But small species chemistry is also important. Updating of C2 chemistry leads to significant improvements in prediction of ethylene, formaldehyde, and methane measurements.

• S Jahangirian; S Dooley; F M Haas; F L Dryer (2011) CombustS. Jahangirian; S. Dooley; F. M. Haas; F. L. Dryer (2011). CombustFlame. DOI:10.1016/j.combustflame.2011.07.002

• Westbrook CK, Pitz WJ, Herbinet O, Curran HJ, Silke EJ (2009).Combust. Flame 156: 181-199.

16

1201-C4H8(a)

70

(b)

n-Decane Temporal Speciation: Olefins

80

100

1 C4H8 2-C4H8 x 3 1-C5H10 2-C5H10 x 3 1-C6H12 2,3-C6H12 x 3

ctio

n (p

pm)

(a)

40

50

60 C7H14 C8H16 C9H18

tion

(ppm

)

(b)

20

40

60

ecie

s m

ole

frac

20

30

ecie

s m

ole

frac

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.40

20

Sp

Time (s)

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.40

10Spe

Time (s)

VPFR oxidation of n-decane/O2/N2 1000/15500/983500 ppm at 8 atm and 830 K.

Significant amounts of large olefins (predominantly, but not exclusively 1-olefins) areobserved. B t i i h i t ll R O h i t b th iti l t di ti hi h Beta-scission chemistry as well as ●R+O2 chemistry are both critical to predicting highmolecular weight intermediate species, particularly olefins.Concerted elimination reaction of ●RO2 to form an olefin and ●HO2 appears of substantialimport for improving model predictions against these data.

• S. Jahangirian; S. Dooley; F. M. Haas; F. L. Dryer (2011). CombustFlame. DOI:10.1016/j.combustflame.2011.07.002

• Westbrook CK, Pitz WJ, Herbinet O, Curran HJ, Silke EJ (2009).Combust. Flame 156: 181-199.

17

n-Decane Temporal Speciation Predictions: JetSurf 2.0 21000

140 JetSurf 2 0 model predictions are good for

12000

15000

18000

0 1 20

70

140

T

(K)

Time (s)

H2O CO CO2 O2 n-decane x 10

actio

n (p

pm)

JetSurf 2.0 model predictions are good for heat release and satisfactory for major species, but show at least a factor of two disparity for major intermediates, e.g.,

• H. Wang, E. Dames, B. Sirjean, D. A. Sheen, R. Tangko, A. Violi, J. Y. W. Lai, F. N.Egolfopoulos, D. F. Davidson, R. K. Hanson, C. T. Bowman, C. K. Law, W. Tsang, N. P.Cernansky D L Miller R P Lindstedt A high-temperature chemical kinetic model of n-

3000

6000

9000

Spe

cies

mol

e fraethylene, formaldehyde, 1-butene, 1-pentene,

and 1-hexene.

Cernansky, D. L. Miller, R. P. Lindstedt, A high temperature chemical kinetic model of nalkane (up to n-dodecane), cyclohexane, and methyl-, ethyl-, n-propyl and n-butyl-cyclohexane oxidation at high temperatures, JetSurF version 2.0, September 19, 2010(http://melchior.usc.edu/JetSurF/JetSurF2.0).

0.0 0.4 0.8 1.2 1.6 2.00

Time (s)

280

320

1-Butene1200

1400 MethaneEthylene

Solid lines: JetSurf 2.0 model;

Symbols:160

200

240

2801-Pentene

1-Hexene

ole

fract

ion

(ppm

)

600

800

1000

1200

le fr

actio

n (p

pm)

Ethylene Propene Formaldehyde

Symbols: Experiment

0 0 0 4 0 8 1 2 1 6 2 00

40

80

120

Spe

cies

m

0 0 0 4 0 8 1 2 1 6 2 00

200

400

Spe

cies

mo

VPFR oxidation of n-decane/O2/N2 1000/15500/983500 ppm at 8 atm and 830 K.

18

0.0 0.4 0.8 1.2 1.6 2.0

Time (s)

0.0 0.4 0.8 1.2 1.6 2.0

Time (s)

Cycloalkanes are a significant fraction of Jet A POSF 4658 composition. 5000

m 20

30

40POSF 1st Generation 1st Generation4658 Surrogate MCH Surrogate

T

T

/ Kp

Is it important any characteristic functional groups present in cycloalkanes?Experimentally VPFR reactivity at Low

( 800K) i l id i l3000

4000

POSF 1st Generation 1st Generation4658 Surrogate MCH SurrogateCOce

ntra

tion

/ pp

500 600 700 800 900 1000

0

10

Temperature / K

temperature (<800K) is almost identical to that of 1st Generation surrogate.Shift of approx. 25-30K to cooler

temperature in “hot ignition” condition. 1000

2000

CO CO2 O2 H2O

Spe

cies

con

c

VPFR reactivity data for conditions of 12.5 atm, 0.3%

p gBehavior of 1st/MCH surrogate appears to

more closely replicate Jet A POSF 4658 in transition to hot ignition...the significance of the improvement remains to be fully

500 600 700 800 900 10000

Temperature / Ky ,

carbon, φ= 1.0 and t =1.8s, for POSF-4658 (symbols), 1st

generation surrogate (light lines), and 1st generationsurrogate + methyl cyclohexane (solid lines). Inset;T.

of the improvement remains to be fully evaluated.

19

n-Propylbenzene (nPB) Kinetic ModelingWon et al (2010) created a simple model forWon et al. (2010) created a simple model for

n-PB based upon the Toluene kinetic model of Metcalfe et al. (2011) with additional reactions that emphasized that fuel consumption would take place principally by oxidation or decomposition of alkyl side chain. Model qualitatively reproduced counterflow

diffusion flame strained extinction data Counterflow diffusion flame strained extinction measurements by Won et al.

diffusion flame strained extinction data.Comparison of predictions for reflected shock

ignition also could be improved upon.Lower temperature departure of the data and

10000 Shock tube, 21.4 atmShock tube, 21.4 atm, pertubed

Temperature [K]1200 1100 1000 900 800

predictions could not be resolved by radical-oxygen addition reactions.VPFR reactivity experiments show that nPB

exhibits no significant low temperature or1000

S oc tube, at , pe tubed

on D

elay

Tim

e [

s]

• Won, S. H., Dooley, S., Dryer, F. L., Ju, Y. (2011). Proc. Combust. Ins. 33, 1163-1170.• Metcalfe, W.K., Dooley, S., Dryer, F. L. (2011). Energy & Fuels, In Press.• Wang, H., Oehlschlaeger, M. A., Dooley, S., Dryer, F. L. (2011). 7th U.S. National Combust.

M ti A l G i

exhibits no significant low temperature or negative temperature coefficient behavior.

f /

0.80 0.85 0.90 0.95 1.00 1.05 1.10 1.15 1.20 1.25100

Igni

tio1000/T [1/K]

20

Meeting, Atlanta, Georgia. Ignition delays for nPB/air φ = 1.0 at ~21 atm. Data (symbols); Predictions (line).

This dataset has a carbon balance of 90-100 % 3000

n‐PB Temporal Speciation at 8 atm, 885 KThis dataset has a carbon balance of 90 100 %.Small heat release (< 8 K) occurs over whole time

range in measurements. Styrene, most abundant intermediate large

carbon number intermediate 1500

2000

2500

0.0 0.5 1.0 1.50

30

60

90

T/

K

Time /sec

O2 / 6COe

fract

ion

/ ppm

carbon number intermediate.Minor intermediates species: acetylene, ethane,

propylene, toluene, and ethylbenzene. Traces of C4 olefins, butadiane and bibenzyl are observed.

Wang et al model predictions are too fast. 0.0 0.5 1.0 1.5 2.0 2.50

500

1000

CO CO2

Fuel (nPB) H2O

Spe

cies

mol

e

Wang et al model predictions are too fast.Time /sec

300

400 Formaldehyde Acetaldehyde Ethylene Methane

/ppm 300

400 Benzene Styrene Benzaldehyde

ppm

100

200

peci

es m

ole

fract

ion

/

100

200

peci

es m

ole

fract

ion

/

VPFR Species Time History for oxidation of n-Propylbenzene/O2 (1110/13300 ppm in balance N2) at 8 atm and 885 K (Symbols: Experiment Lines: Wang et al model)

0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.750

Sp

Time /sec0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.750

Sp

Time /sec

Model: Wang, H., Oehlschlaeger, M. A., Dooley, S., Dryer, F. L. (2011). 7th U.S. National Combust. Meeting, Atlanta, Georgia.

(Symbols: Experiment, Lines: Wang et al. model)Data: S. Jahangirian, T. Farouk, S. Dooley, F. L. Dryer , 7th ICCK, MIT, Cambridge, MA, July 2011.

21

VPFR 1,3,5-TmB Temporal Speciation at 12.5 atm, 930 K12000

Development of a model for 1,3,5-trimethyl benzene is in progress. Three abundant aromatic intermediates are: 6000

8000

10000

12000

0.0 0.5 1.0 1.5 2.00

20

40

60

Hea

t rel

ease

, dT

/K

Time /sec

actio

n / p

pm

CO CO2 O2 / 3 H2O 135-TMB x 4

Three abundant aromatic intermediates are: m-xylene > benzene > toluene > benzaldehyde.

0

2000

4000

6000

Spec

ies

mol

e fra

0.0 0.5 1.0 1.5 2.0 2.50

Time /sec

250

300 Methane Ethylene Acetylenepm

50

60 Toluene Benzaldehyde Benzene m-xylenepp

m 12

15 Ethane Propylene 1,3-butadiene

pm

100

150

200

cies

mol

e fra

ctio

n / p

20

30

40y

cies

mol

e fra

ctio

n / p

3

6

9

es m

ole

fract

ion

/ pp

VPFR temporal speciation of 1,3,5-trimethyl benzene/O2 (1110/13300 ppm

0.0 0.5 1.0 1.5 2.0 2.50

50

Spec

Time /sec

0.0 0.5 1.0 1.5 2.0 2.50

10Spec

Time /sec0.0 0.5 1.0 1.5 2.0 2.5

0

3

Spec

i

Time /sec

22

VPFR temporal speciation of 1,3,5 trimethyl benzene/O2 (1110/13300 ppmin balance N2). Initial conditions: P=12.5 atm, φ=1.0, T=930K.

S. Jahangirian, S. Dooley, and F.L. Dryer, 240th American Chemical Society National Meeting, Advances in Fuel Science & Technology, Boston, MA, August 22–26, 2010.

Physical Property of Surrogate FuelsS t F l F l tiSurrogate Fuel Formulation

Emulation of Physical Properties Emulation of Combustion PhenomenaPhysical Property Targets:

Distillation Curve

Combustion Property Targets:

Average Fuel Molecular Weight Distillation Curve

Density

Viscosity

Average Fuel Molecular Weight

Hydrogen/Carbon Molar Ratio

Cetane Number

Thermal Conductivity Threshold Sooting Index

Little attention on how to integrate successful approaches to each issue into a framework for the mutual evaluation of the two issues.Liquid fuel combustion phenomena depend of the gas phase reaction for chemical heat

release, liquid fuel properties dictate vaporization and mixing process – multiphase problem.

Surrogate formulation emulating both physical properties of the target real fuel in addition

23

to combustion kinetic properties across the distillation curve.

Physical Property Consideration Equation of state (EOS) to determine the liquid phase thermo-physical

properties – Helmholtz form of equation of state due to its high accuracy.

NIST REFPROP (Reference Fluid Thermodynamic and Transport Property) ( y p p y)and ThermoData Engine (TDE).

Top – down approach: Liquid phase physical property evaluation once the surrogate fuel has been formulated to match the combustion propertysurrogate fuel has been formulated to match the combustion property targets.

Mole Fraction DCN H/C MW/ g mol-1

TSI

Jet-A POSF 4658 47.1 1.957 142.01 21.4

1st Generation Surrogate

n-decane iso-octane toluene47.1 2.01 120.7 14.1

0.427 0.33 0.243

2nd G ti n-dodecane iso-octane 1 3 5 TmB n-PB2nd Generation Surrogate

n-dodecane iso-octane 1,3,5 TmB n-PB48.5 1.95 138.7 20.4

0.40 0.29 0.07 0.23

*H-B Surrogate n-dodecane Tetradecane 1,2,4 TmB60.4 1.89 158.3 28.7

0.288 0.304 0.408

24

*Bruno, and Huber (2010)). Energy Fuels: DOI:10.1021/ef1004978

Liquid Phase PropertiesMW is considered as a combustion• MWavg is considered as a combustion

property target to constrain mass diffusive issues.

• MWavg is appropriate to some degree avg pp p gas a measure of other physical properties.

• 2nd generation surrogate density is estimated to be 5 6% lower thanestimated to be ~ 5-6% lower than measurements.

• 2nd generation viscosity deviates by• 2nd generation viscosity deviates by ~45 – 25% from 273 – 373 K, but is likely to deviate minimally at the operating temperatures of gas turbines.

25

Moving Forward on Physical/Chemical Property IntegrationMole Fraction DCN H/C MW/ g

mol-1TSI

Jet-A POSF 4658 47.1 1.957 142.01 21.4

1st Generation n-decane iso-octane toluene47 1 2 01 120 7 14 1Surrogate 47.1 2.01 120.7 14.1

0.427 0.33 0.243

2nd Generation Surrogate

n-dodecane iso-octane 1,3,5 TmB n-PB48.5 1.95 138.7 20.4

0.40 0.29 0.07 0.23

Constraint in the lightly branched alkanes – higher molecular weight isomer might benefit matching combustion targets …also there are other ways of accomplishing the appropriate result.

•The straight chain alkane fraction can be represented as a distribution of alkanes (e.g. nC6 – nC16) instead of just a single component.

Ch i l ki ti f t i ht h i lk i ll d fi d• Chemical kinetics of straight chain alkanes is well defined.

• The distribution of alkanes will increase the average molecular weight and hence permit flexibility in emulating liquid phase physical properties.

26

Surrogate Fuel Formulation for emulating fully vaporized real fuel combustionSurrogate Fuel Formulation for emulating fully vaporized real fuel combustion MURI property target approach has been refined and comprehensively

tested! Property targets reflect the integral interactions of the important discrete chemical functionalities

present in the fuelpresent in the fuel. The method permits flexibility in selecting surrogate mixture components and results in non-unique

compositions that emulate the specific combustion property targets. Experimental replication of real fuel global combustion properties is typically in better agreement than

detailed model predictions based upon surrogate compositiondetailed model predictions based upon surrogate composition.

A four component, 2nd Gen surrogate mixture of n-C12/iso-C8/n-PB/1,3,5 TmBcan replicate the combustion property targets of petroleum derived fuels, alternative fuel feed stocks and their blends with petroleum. Weakly branched isomer and cyclo alkane components generally produce only minor changes in

surrogate performance (but can be added within the context of the MURI formulation methodology).

Efforts emphasize improving accuracy of individual component and mixture model predictions for the 2nd Gen surrogate.p g MURI contributing heavily to enhance experimental databases for all components with special

emphasis on aromatics. Significant model advancements have been achieved for toluene, xylenes, n-PB, and 1,3,5 TmB. 2nd generation surrogate model development underway.g g p y

27

Integrating Surrogate Fuel Physical and Chemical Kinetic Properties

Summary(2)Integrating Surrogate Fuel Physical and Chemical Kinetic Properties Properties, especially distillation curve and vapor-liquid equilibrium, require

molecular weight and functional group distributions. Relative importance of physical properties and chemical kinetic properties to multiphase combustion

behavior observations in gas turbine environments need experimental evaluation (e g “Rules and Toolsbehavior observations in gas turbine environments need experimental evaluation (e.g. “Rules and Tools program”).

What uncertainties in the respective property emulations are acceptable? The MURI surrogate formulation has been confirmed to encompass using

h d b l t f ti t f l t t i thydrocarbon solvent fractions to formulate surrogate mixtures. Surrogate solvent formulation economics permit applied testing applications of the techniques.

Surrogate Solvent kinetics can still be emulated by those for smaller numbers of pure components.

Surrogate solvent formulations can be used to independently vary distillation curve, and molecular class distributions over the curve while maintaining the same fully prevaporized combustion kinetic behavior.

28

Demonstrate added cases of surrogate mixture comparisons with other gas turbine fuel samples over wide range of fundamental venues. Larger database of combustion property targets already under

construction for other POSF’s. Collaborate in test/proof that solvent surrogate formulation approach can

connect fundamental MURI mixture formulation concept and applied research testing.

Provide further validation data for n-dodecane, n-decane, iso-octane, and , , ,alkyl aromatics in kinetic transition to high temperature (beta scission/chemical branching behavior); higher MW alkanes also of interest.

Improve alkane and aromatic oxidation kinetic models for surrogate mixtures.

Further evaluate/validate MCH kinetic modeling and inclusion of MCH as a surrogate component.

Produce MURI mixture kinetic model and appropriate reductions for validation/application.

Further evaluate MURI surrogate mixtures to emulate real fuel sooting behavior in fundamental and applied venues.

29

This work was supported by the Air Force Office of Scientific Research under the 2007 MURI Grant No. FA9550-07-1-0515 (at PU, UCONN, PSU, and UIC) and under Grant No. FA9550-07-1-0114 (at RPI). Dr. Julian Tishkoff; Program Manager; Dr. Timothy Edwards AFRL, technical discussions, fuel samples.

R h T M bResearch Team Members

Fred Dryer, Stephen Dooley, Sang Hee Won, Marcos Chaos, Joshua Heyne, Yiguang Ju, Saeed Jahingarian,Wenting Sun, Francis Haas, Henry Curran, Wayne Metcalfe, Amanda Ramcharan, Timothy Bennett, John Grieb,Lisa Langelier-Marks, Joseph SivoMechanical and Aerospace Engineering, Princeton University, Princeton, NJ

Kamal Kumar, Chih-Jen Sung School of Engineering, University of Connecticut, Storrs, CT

Robert J. Santoro and Thomas A. Litzinger, Venkatesh Iyer, Suresh Iyer, Milton LinevskyThe Energy Institute, The Pennsylvania State University, University Park, PA

Kenneth Brezinsky, Thomas Malewicki, Soumya Gudiyella, Alex FridlyandMechanical Engineering, University of Illinois Chicago, IL

Matthew A. Oehlschlaeger, Haowei WangMechanical, Aerospace, and Nuclear Engineering, Rensselaer Polytechnic Institute, Troy, NY

We also wish to thank:• Drs Marco Mehl Mani Sarathy Bill Pitz and co-workers at LLNL; modeling contributionsDrs. Marco Mehl, Mani Sarathy, Bill Pitz and co workers at LLNL; modeling contributions.• Dr. Cliff Moses, Dr. John Farrell, Prof. Hai Wang; technical discussions.

30

Current science findings can assist in unraveling

Future DriversCurrent science findings can assist in unraveling

property/combustion kinetic effects in multi-phase, applied combustion experiments.Even the most developed and studied kinetic models require furtherEven the most developed and studied kinetic models require further

refinement to replicate real fuel behavior as well as achieved in experimental comparisons of real fuel and the formulated surrogate mixtures.Normal-alkane chemistry is controlling but aromatic chemistry is

limiting.Improved kinetic models for aromatics are essential.

E i i Q ti f ti i i t !Engineering Questions of continuing importance!What level of emulation of real fuel properties must surrogates meet? Gas phase kinetic property targets?Kinetic model predictive accuracy requirements?Physical property predictive accuracy requirements?Optimal liquid fuel physical and chemical property integration?

31

p q p y p p y g