Experiments on Low-Temperature Combustionyju/2nd-flame-chemistry workshop...Experiments on Low-Temperature Combustion Development of a Stabilized Cool Flame Platform & Faraday Rotation

Experiments on Low-Temperature CombustionCombustion

Development of a Stabilized Cool Flame Platform & Faraday Rotation Spectroscopy Diagnostic for In-Situ Measurement of

HOx Radicals

2nd Flame Chemistry Workshop

San FranciscoSan Francisco

2 ‐ 3 August 2014

Sang Hee Won and Brian Brumfield (Joseph Lefkowitz)Department of Mechanical and Aerospace Engineering, Princeton University, Princeton, NJ, USA

PRINCETONMechanical and Aerospace Engineering 1

PRINCETONMechanical and Aerospace Engineering

Introduction• Take‐home messages from the 1st Flame Chemistry Workshop

– What is the definition of flame chemistry?• Chemical kinetics constrained by transport• Chemical kinetics constrained by transport

– Development of well‐defined experimental platforms• Extend ability to access low temperature chemistry (LTC)• Advanced laser diagnostic technique

R t d d i• Recent advanced engines– Operate at low to intermediate temperature at higher pressure conditions– Near‐limit combustion behaviors tend to be correlated with LTC

500001300 1200 1100 1000 900 800 700

Temperature / K

LBO test by Med Colket (UTRC)3

10000

50000Detailed Model + Surrogate Fuels

Jet-A, POSF 4658 IPK, Iso Paraffinic Kerosene S-8, Coal-to-Liquid SPK, Gas-to-Liquid

e, τ

/ μs

Temperature window for DCN (CN) measurements

1000

Igni

tion

dela

y tim

e

21) H. Wang, M. A. Oehlschlaeger, Fuel 98 (2012) 249-258.2) S. H. Won, et al., “Comparative Evaluation of Global Combustion Properties of Alternative Jet Fuels,” 51th AIAA Aerospace Sciences Meeting, Grapevine, Texas (2013).3) Med Colket, 2013 MACCCR meeting

(CN) measurements

0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.550

100

1000K / T

20 atm, stoichiometric fuel in air


Motivations

1. Experimental platform for cool flame– To stabilize LTC‐driven flame

2. Development of FRS technique–Quantifying the LTC related species

3

Development of a Stabilized Cool FlameDevelopment of a Stabilized Cool Flame Platform

2nd Flame Chemistry Workshop

San FranciscoSan Francisco

2 ‐ 3 August 2014

Sang Hee Won1, Bo Jiang1, Pascal Diévart1, Chae Hoon Sohn2, Yiguang Ju1

1Department of Mechanical and Aerospace Engineering, Princeton University, Princeton, NJ, USA2Department of Mechanical Engineering, Sejong University, Seoul 143‐747, Republic of Korea



Challenges to Stabilize LTC‐Driven Flames

• Induction chemistry at low temperature is very slow– Inability to initiate the radical pool (RH + X = R + HX)– Very sensitive to molecular structure

• Then, how to shorten the induction chemistry?

l fl l b f• Cool flames; mostly observed in premixed configuration– Flow reactor, jet‐stirred reactor, etc..

• Is it possible to observe cool flames in diffusive configuration ?10000

100

1000

10000

imes

[ms]

experiment, 13.5 atm, 2nd ig.model, 13.5 atm, 2nd ig.model, 13.5 atm, 1st ig.model, 1.0 atm, 2nd ig.model, 1.0 atm, 1st ig. 3

τ @ 2nd stage ignitionat 700 K

Adiabatic Constant Volume Ignition

0 1

1

10

nitio

n de

lay

t

0

1

2

0 20 40 60 80 100 120

time [ms]

pressureτ @ 1st stage ignition1 atm

13 5

5

0.01

0.1

0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6

Ign

1000/T [1/K]

time [ms]13.5 atm770 K850 K

1) H. J. Curran, P. Gaffuri, W. J. Pitz, C. K. Westbrook, Combust. Flame 114 (1998) 149-177.


Hints from Recent StudiesZero‐Gravity Experiment1,2

• Observed cool diffusion flame• Observed cool diffusion flame in a droplet combustion

• Cool flame exists in diffusive configuration!

61) V. Nayagam et al., Combust. Flame 159 (2012) 2) T. I. Farouk, F. L. Dryer, Combust. Flame 161 (2014)3) W. Sun, S. H. Won, et al, Proc. Combust. Inst. 34 (2013)

4) T. Ombrello, S. H. Won, et al., Combust. Flame 157 (2010)5) T. M. Vu, S. H. Won, et al., Combust. Flame 161 (2014)


Hints from Recent StudiesZero‐Gravity Experiment1,2

• Observed cool diffusion flame

Plasma‐Assisted Combustion3‐5

• Initiation of radical pool can be• Observed cool diffusion flame in a droplet combustion

• Cool flame exists in diffusive

Initiation of radical pool can be accelerated by Plasma

• Enhancing flame ignition, propagation speed, and stabilization

CH

configuration! • Electronically excited species and ozone, etc.

Increasing fuel loading

4x1015

5x1015

6x1015

7x1015

SmoothTransition

Extinction

χO2=34%χO2=62%

r de

nsity

(cm

-3)

CH4

pera

ture Plasma generated

species:O, H, O2(a∆g) …

Extinctionflame initiation flame stabilization

Increasing fuel loading

0.05 0.10 0.15 0.20 0.25 0.30 0.35

1x1015

2x1015

3x1015Extinction

Ignition

F l l f ti

OH

num

ber

Residence time

Tem

p O, H, O2(a∆g) …

the classical S‐curve Ignition

Low pressure counterflow diffusive configuration with nano-second pulsed discharge3 Fuel mole fraction

7

New combustion regime

1) V. Nayagam et al., Combust. Flame 159 (2012) 2) T. I. Farouk, F. L. Dryer, Combust. Flame 161 (2014)3) W. Sun, S. H. Won, et al, Proc. Combust. Inst. 34 (2013)

4) T. Ombrello, S. H. Won, et al., Combust. Flame 157 (2010)5) T. M. Vu, S. H. Won, et al., Combust. Flame 161 (2014)

discharge


Experiments• A heated counterflow burner integrated with vaporization system1

– n‐heptane/nitrogen vs. oxygen/ozone• Ozone generator (micro‐DBD) produces 2‐ 5 % of ozone in oxygen

stream, depending on oxygen flow rate• Speciation profiles by using a micro‐probe sampling with a micro‐

GC 2GC.2

Heated N2 @ 550 KFuel/N2 @ 550 K

Positioning stage

N2 @ 300 K

Stagnationplane Pressure

chamberMicro-GC

O2 + O3 @ 300 K

Ozone generator O2 @ 300 K

Thermal gradient in mixing layer initiates reaction of O3 + (M) = O + O2 + (M)

81) S. H. Won, et al., Combust. Flame 157 (2010) 2) J. K. Lefkowitz, S. H. Won, et al., Proc. Combust. Inst. 34 (2013)


Initiation of Cool Diffusion Flames• Procedure to initiate a cool diffusion flame

1) Setting nitrogen (fuel side) and oxygen (oxidizer side)1) Setting nitrogen (fuel side) and oxygen (oxidizer side) flow rates

2) Turning on the ozone generator3) Flowing fuel (n‐heptane) to fuel side3) Flowing fuel (n heptane) to fuel side

Lower fuel mole fraction:Cool diffusion flame

Higher fuel mole fraction:Hot diffusion flame

9


Initiation of Cool Diffusion Flames• Existence of cool diffusion flames in counterflow

configuration with n‐heptane

2400]

– cool flame regime exists regardless of addition of ozone– .

2000

e Tm

ax [K

]

nC7H16/N2 vs O2 or O2/O3in counterflow burner

Xf = 0.05,Tf = 550 K, and To = 300 K

Extinction limit ofconventional hot diffusion flame

HF branch

1200

1600

empe

ratu

re conventional hot diffusion flame(HFE)

Extinction limit ofcool diffusion flame

without O3(a) Cool diffusion flame

800

Max

imum

te cool diffusion flame(CFE)

CF branchHTI

LTI(b) Hot diffusion flame

4000.1 1 10 100 1000 10000

M

Strain rate a [1/s]

LTI

10


Initiation of Cool Diffusion Flames• Existence of cool diffusion flames in counterflow

configuration with n‐heptane

2400]

– cool flame regime exists regardless of addition of ozone– Addition of ozone extends cool flame regime.

2000

e Tm

ax [K

]

nC7H16/N2 vs O2 or O2/O3in counterflow burner

Xf = 0.05,Tf = 550 K, and To = 300 K

Extinction limit ofconventional hot diffusion flame

HF branch

1200

1600

empe

ratu

re conventional hot diffusion flame(HFE)

Extinction limit ofcool diffusion flame

without O3

ith O

(a) Cool diffusion flame

800

Max

imum

te cool diffusion flame(CFE)

with O3

CF branchHTI

LTI(b) Hot diffusion flame

4000.1 1 10 100 1000 10000

M

Strain rate a [1/s]

LTI

11


Speciation Profiles

• Temperature measurements 700

900

e [K

]

T [K] expTemperature measurements– Over‐estimation of heat release in model prediction 300

500

700

0 4 8 12 16 20 24Tem

pera

ture T [K], exp.

T [K], model

(a)

• Failure to predict the flame position

Distance from fuel side nozzle [mm]

40000

60000

80000

100000

120000

ecie

s m

ole

ctio

n [p

pm]

nc7h16, exp.nc7h16, modelo2/10, expposition.

– Boundary conditions were tested previously.1

0

20000

0 4 8 12 16 20 24

Spe

frac


o2/10, model

30000

40000

ole

pm] h2o, exp.

h2o, modelco exp

(b)

– Consistent even without putting sampling probe or thermocouple.

0

10000

20000

Spec

ies

mo

fract

ion

[pp co, exp.

co, modelco2, exp.co2, model

(c)

12

6 10 14 18


1) J. K. Lefkowitz, S. H. Won, et al., Proc. Combust. Inst. 34 (2013)


Speciation Profiles• Reasonable prediction of acetaldehyde and CH2O• Significant over estimation of C H and CH formation• Significant over‐estimation of C2H4 and CH4 formation

– Factor of 10.

4000

6000

8000

cies

mol

e io

n [p

pm] acetaldehyde, exp

acetaldehyde, model

ch2o, exp

ch2o, model

R + O2

RO2 Olefin + HO2Propagation

0

2000

6 10 14 18

Spec

frac

ti

Distance from fuel side nozzle [mm]1000

]

(a)QOOH

Olefin + Carbonyl

Olefin + HO2QO + OH

+ O2- O2

Propagation

+ HO2

400

600

800

peci

es m

ole

actio

n [p

pm]

c2h4, exp.

c2h4/10, model

ch4, exp

ch4/10, model

O2QOOH

Ketohydroperoxide + OH

+ HO2

13

0

200

6 10 14 18

Sp fra


(b)CH2O + R + CO + OH

Branching


Quick Summary• Well‐defined experimental platform of cool flames for LTC study• Speciation profiles revealed deficiency of kinetic model at cool

flflame regime– Over‐prediction of small hydrocarbon species (CH4, C2H4, etc.)

• Sophisticated diagnostic techniques might be able to point outO i i f d l di i l h i b hi– Origin of model over‐prediction: low temperature chain branching vs. propagation reaction pathways

R + O2

P i

6005

RO2

QOOH

Olefin + HO2Propagation

Olefin + Carbonyl

Olefin + HO2QO + OH

Propagation400

500

3

4

HO

2 mole frct

ion

[ppm

]

Olefin + Carbonyl

O2QOOH

K t h d id OH

+ O2- O2 + HO2

100

200

300

1

2

raction [ppmH m

ole

frac

14

Ketohydroperoxide + OH

CH2O + R + CO + OH

Branching

0

100

09 10 11 12 13 14 15

m]O

H


Faraday Rotation Spectroscopy Di ti f I Sit M t fDiagnostic for In-Situ Measurement of

HOx Radicals

Brian Brumfield1Brian Brumfield1Joseph Lefkowitz2, Xueliang Yang2, Yiguang Ju2,

Gerard Wysocki11Department of Electrical Engineering, Princeton University, Princeton, NJ, USA

2 Mechanical and Aerospace Engineering Department, Princeton University, Princeton NJ



Challenges w/ LTC HOx Measurements

• General challenges with radical quantification– Wall quenching q g– Spectral Interference

• Species specific complications• Species specific complications– OH

• Present at low concentrations (<1 ppmv)• Difficult to quantify via LIF in situ• Difficult to quantify via LIF in situ

– HO2N t d t t bl i LIF ( h t f t ti LIF* i ibl )• Not detectable via LIF (photo-fragmentation LIF* is possible)

• Fluorescence Assay by Gas Expansion (FAGE) †

16

*Johansson, O.; Bood, J.; Li, B.; Ehn, A.; Li, Z. S.; Sun, Z. W.; Jonsson, M.; Konnov, A. A.; Aldén, M.: Combust. Flame 2011, 158, 1908-1919

† Blocquet, M.; Schoemaecker, C.; Amedro, D.; Herbinet, O.; Battin-Leclerc, F.; Fittschen, C.: Proc. Natl. Acad. Sci. U.S.A 2013, 110, 20014-20017


Faraday Rotation Spectroscopy (FRS)

• Apply Magnetic Field → Zeeman Splitting → Faraday Effect• Polarization rotation → Linear Polarizer → Intensity VariationPolarization rotation Linear Polarizer Intensity Variation

Radical transitionRadical transition

• Sample modulation by varying magnetic field (AC-FRS)*• Strong suppression of absorption signals from non-radicals

17

g pp p g• Zero background technique* Litfin, G.; Pollock, C. R.; Curl, J. R. F.; Tittel, F. K.: J. Chem. Phys. 1980, 72, 6602-6605.

Brumfield, B.; Sun, W.; Ju, Y.; Wysocki, G.: J. Phys. Chem. Lett. 2013, 4, 872-876


Faraday Rotation Spectroscopy (FRS)

• Apply Magnetic Field → Zeeman Splitting → Faraday Effect• Polarization rotation → Linear Polarizer → Intensity VariationPolarization rotation Linear Polarizer Intensity Variation

Radical transitionRadical transition

• Marginal increase in experimental complexity from TDLAS

18

(polarizers, 1‐2 lock‐in amplifiers, magnetic coil)

PRINCETONMechanical and Aerospace EngineeringExperimental Demonstration of FRS

Laser

er

Fl

• Target OH (2.8 µm) and HO2 (7.1 µm)

• 1.7 cm optical path from reactor opening

Mai

n H

eate

ow R

eacto

• Measure 2 mm from exit

• Laser wavelength + magnetic field modulation→ DM-FRS*

M r

• Laser wavelength + magnetic field modulation→ DM-FRS– Detect HO2 at 1fL±fM (1f DM-FRS) – Detect OH at 2fL±fM (2f DM-FRS)

* Brumfield, B.; Sun, W.; Wang, Y.; Ju, Y.; Wysocki, G.: Opt. Lett. 2014, 39, 1783-1786


Example DM‐FRS Spectra

OH

HO

20

HO2

* Brumfield, B.; Sun, W.; Wang, Y.; Ju, Y.; Wysocki, G.: Opt. Lett. 2014, 39, 1783-1786


Observed Species ProfilesOH

• Conditions:– Composition: 0.96/0.0375/0.00225

He/O2/CH3OCH3He/O2/CH3OCH3

– Plug-flow residence time: 0.45 to 0.2 seconds

• DME,CH2O, and CO measured with gas chromatography

• OH/HO2 concentrations consistent with model †

• Observed low‐reactivity ~600 K similar to i t *prior measurements*

21

† Zhao, Z.; Chaos, M.; Kazakov, A.; Dryer, F. L.: Int. J. Chem. Kinet. 2008, 40, 1-18* Brumfield, B.; Sun, W.; Ju, Y.; Wysocki, G.: J. Phys. Chem. Lett. 2013, 4, 872-876

Kurimoto, N.; Brumfield, B.; Yang, X.; Wada, T.; Diévart, P.; Wysocki, G.; Ju, Y.: Proc. Combust. Inst., in Press DOI: 10.1016/j.proci.2014.05.120


Proposed Future Work

Fuel/N2 @ 550 KCH2O laser

Electromagnetic coils

OH/HO2 laser

O2 + O3 @ 300 K

• Merge FRS diagnostic with cool flame platform

Ozone generator O2 @ 300 K

• Employ mid‐IR absorption diagnostic for CH2O quantification

• Potential to spatially profile cool flame

• Use LIF imaging to extract absolute concentration profile

22


Extension of FRS technique

• Applicable to many other small radical species

• Relevant to combustion and atmospheric pchemistry studies

• Many potential targets;CH CH CH NO* NO † HCO HCN t– CH, CH2, CH3, NO*, NO2

†, HCO, HCN etc…

23

*Wang, Y.; Nikodem, M.; Wysocki, G.: Opt. Express 2013, 21, 740-755† Zaugg, C. A.; Lewicki, R.; Day, T.; Curl, R. F.; Tittel, F. K.: 2011, 79450O-79450O-7


Conclusion• Develop a experimental platform to study cool flame

chemistry

• Significant disagreement from observed vs. predictedspeciation profiles using existing n-heptane mechanism

• Quantification of HOx would aid in constraining kinetic model

FRS h b d t t d t id iti d• FRS has been demonstrated to provide sensitive andselective measurements of HOx

Combination of platform with diagnostic will provide insight• Combination of platform with diagnostic will provide insightinto LTC chemistry

24


Acknowledgements

Funding

Princeton Environmental Institute

Andlinger Center for Energy and the Environment

Air Force Office of Scientific ResearchAir Force Office of Scientific Research MURI research grant

grant # FA9550‐07‐1‐0136

grant # FA9550‐13‐1‐0119grant # FA9550 13 1 0119.

US Department of Energy, Office of Basic Energy Sciences Energy Frontier Research Center on Combustion Grant # DE

25

Energy Frontier Research Center on Combustion, Grant # DE‐SC0001198

Documents

Experiments on Low-Temperature Combustionyju/2nd-flame-chemistry workshop...Experiments on Low-Temperature Combustion Development of a Stabilized Cool Flame Platform & Faraday Rotation