The Oxidation and Autoignition of Jet Fuels

Matthew Oehlschlaeger Mechanical, Aerospace, and Nuclear Engineering Rensselaer Polytechnic Institute, Troy, NY

Sponsor: AFOSR, Dr. Chiping Li

Motivation: changing fuel supply • Combustion properties….

– are dependent on fuel composition – of low-volatility and alternative fuels are less well

characterized

• Shock tube studies as high-pressures for fuel/air mixtures (engine-like conditions)

• Quantitative targets for the validation of kinetic models • Direct comparisons of fuel reactivity

Methodology

Payoff

Methodology: heated shock tube technique

• Isolation of homogenous gas-phase kinetics

• Reactants are heated “instantaneously” – Zero-dimensional model – Temperature and pressure can be accurately determined

(typ. 1-2% uncertainties)

• High pressures achieved behind reflected shocks – Present studies ~7-50 atm

• Extended test times (10 ms) achieved using tailored

driver gases

Vacuum section

RPI shock tube facility

Heated and insulated mixing vessel

Driver

Heated and insulated driven section

Test location w/ optical access

Shock velocity detection

Diaphragm

Inner diameter = 5.7 cm; initial temperatures up to 200

C

Mixing manifold

P [atm] T [K] τ [ms] φ Accessible 1 - 200 650 + 0 - 10 Typ. for fuel/air 5 - 60 650 - 1300 0.05 - 10 0.25 - 2.0

Condition Space

Ignition delay measurement

Ignition delay

Incident Shock Velocity

Ignition Time Measurement

to DAQ

to counter-timers for velocity

Kistler transducer to DAQ

incident shock

reflected shock

PZTs

Filtered Si photodetector (OH* emission)

Shock tube test

section

Example results

Uncertainties: ±25% in ignition delay ±1.5% in temperature

Example results

Uncertainties: ±25% in ignition delay ±1.5% in temperature

High temp

Low temp

Negative- temperature- coefficient (NTC)

Example results RH

R

RO 2

QOOH

OOQOOH

ketohydroperoxide + OH

’

olefin + HO

cyclic ether + OH

OH + radical

O 2

O 2

RH

R

RO 2

QOOH

OOQOOH

ketohydroperoxide + OH

’

olefin + HO2

cyclic ether + OH

OH + radical

O 2

O 2

Beta-scission / fragmentation

Fuel decomposition

Beta-scission + R’’

• Temperature dependence mechanistically explained by models in literature • Quantitative prediction is a work in progress

Uncertainties: ±25% in ignition delay ±1.5% in temperature

High temp

Low temp Negative-

temperature- coefficient (NTC)

Experiment vs surrogate modeling pressure dependence φ dependence

Generalizations • a priori surrogate kinetic modeling deviates with real fuel ignition delay by on average a factor of two – sufficient for “engineering accuracy”?

• Kinetic sensitivities/dependencies much larger in NTC than high-T regime • In condition and fuel composition/structure parameter spaces

Jet fuel ignition variability

~ 4x

DCN = 31

DCN = 61

0

10

20

30

40

50

60

70

80

90

100

Sasol IPK JP-5 Jet A HRJ-5 S-8 Shell GTL

n-alkane

iso-alkane

cyclo-alkane

aromatic% b

y vo

lum

e

DCN: 31 42 47 59 59 61

DCN a proxy for NTC ignition delay?

The IQT-based DCN measurement made is at ~830 K and 22 atm

Ignition delay vs DCN

Ignition delay vs DCN

The entrance to the high-T regime occurs at higher T with increasing DCN

Jet fuel composition

n-alkanesiso-alkanescyclo-alkanesaromaticsother

n-alkanesiso-alkanescyclo-alkanes

Jet A S-8

methyl alkanesdimethyl alkanestrimethyl alkanes

iso-alkanes

methyl alkanesdimethyl alkanes

iso-alkanes

Conventional and alternative jet fuels contain Large quantities of lightly branched alkanes

Collaborative study of lightly-branched alkanes

• RPI: shock tube ignition delay • Nat. Univ. Ireland Galway: shock tube and

RCM ignition delay • LLNL: kinetic modeling

• Compounds studied:

C8 alkane comparison

Compound RON DCN

2,5-dimethylhexane 56

3-methylheptane 37 43

2-methylheptane 22 53

n-octane -19 64

At high-pressure fuel/air conditions Branching has: • little to no influence at high T • significant influence in NTC region

→ dependence on number and location of methyl substitutions

C8 alkane comparison: T < 1000 K

Equivalence ratio dependence: 3-methylheptane

Higher temperatures • small φ-dependence; crossover in φ-dependence at 1200 K for 6.5 atm case

Moderate to low temperatures • larger φ-dependence

Crossover in φ-dependence

• T > 1200 K: ignition controlled by H + O2 → OH + O – Increased φ, increased fuel scavenging of H atoms:

fuel + H → R + H2

• T < 1200 K: H + O2 + M → HO2 + M competes – Increased φ, increased radical production:

fuel + HO2 → R + H2O2, H2O2 + M → 2OH + M

• Additionally at T > 1200 K: inhibitive decomposition

(+ radical)

resonantly stable

Observations from C8 alkane isomer studies

• High temperatures – Little to no difference in ignition delay for C8 isomers

• Similar bond strengths, similar intermediate pool, similar reactivity

– Model captures pressure- and φ-dependence but is ~2x longer than measured ignition delay

• Low temperatures – Model in better agreement with both RCM and ST data – Reactivity dependent on location and number of

methyl substitutions

Low-T chemistry: 2-methylheptane vs n-octane 2-methylheptane n-octane

RH

R

RO2

QOOH

OOQOOH

O2

O2

Low-T chemistry: 2-methylheptane vs n-octane

2-methylheptane n-octane 2-methylheptane: slower OOQOOH isomerization; no H atom available at OOH site

Low-T chemistry: 3-methyl vs 2-methyl heptane

Inhibitive cyclic ether formation results due to slow OOQOOH isomerization

3-methylheptane 2-methylheptane RH

R

RO2

QOOH

O2

OOH OO

High-temperature deviations

At high T&P • Sensitive reactions are within C0-C2 chemistry and fuel + HO2 R + H2O2 • Over prediction of alkane ignition delay is common among literature models

Summary

• Jet Fuels – Ignition delay variability large in the NTC – Correlates with DCN

• Not yet clear how universal this observation is

• Normal and lightly-branched alkanes – Number and location of methyl substitutions has a

strong influence on NTC ignition delay – Isomer reactivity differences explained by competition

among low-T radical branching and propagation pathways

– Perhaps still some issues with high-T alkane kinetics

Acknowledgements

• RPI: Dr. Haowei Wang, Weijing Wang

• LLNL: Drs. Bill Pitz, Charlie Westbrook, Mani Sarathy, and Marco Mehl

• Princeton: Drs. Fred Dryer, Yiguang Ju, Stephen Dooley, and Sang Hee Won

• NUI Galway: Drs. Henry Curran and Darren Healy

• AFOSR (sponsor): Dr. Chiping Li

Sensitivity analysis 3-methylheptane/air, 20 atm, 1100 K

C0-C2 sensitivity: common to all alkanes at high temperatures

H2O2 formation