Measurements of Elementary Rate Constants,Ignition Delays and Species Histories in Shock Tubes

R. K. Hanson, D. F. DavidsonStanford University

1st International Workshop on Flame ChemistryJuly 28‐29, 2012

• Shock Tube/Laser Approach• Advances in Methodology• Elementary Reaction Rate Studies• Multi‐Species Time‐Histories

Stanford Shock Tubes

2

KineticsShock Tube 2

(30 atm)

KineticsShock Tube 1(30 atm)

AerosolShock Tube(10 atm)

HighPressureShock Tube

(500 atm)

New Constrained Reaction Volume (CRV) Facility2 Configurations

3

CRV‐1Gas‐Phase

Experimentsto 100 atm

CRV‐2Aerosol

Experimentsto 100 atm

ConstrainedReaction Volume

SlidingValve

Aerosol Generation Tank

Stanford Laser Diagnostics

4

Ultra‐fast lasers used to 

extend UV tuning range  

(2009) 

First use of tunable dye lasers in shock tubes (1982) 

New lasers allow simple 

access to mid‐IR  (2007‐10)

UltravioletCH3 216 nmNO        225 nmO2 227 nmHO2 230 nmOH        306 nmNH        336 nm

VisibleCN 388 nmCH       431 nmNCO    440 nmNO2        472 nmNH2        597 nm HCO    614 nm 

InfraredH2O      2.5 mCO2          2.7 mCH4 3.4 mCH2O    3.4 mCO        4.6 mNO        5.2 mC2H4        10.5 m

Laser Absorption Yields High SensitivityRepresentative Detection Limits: Polyatomic Molecules

• Polyatomic molecules @ 1500K: 2‐200 ppm

1000 1500 2000 2500 3000

0.01

0.1

1

10

100

1000

CH2OCO2

CH3

C2H4D

etec

tion

Lim

it [p

pm]

Temperature [K]

1atm,15cm,1MHz

H2ONH2

5

sub‐ppmsensitivity for OH, CH, CN

Laser Absorption Yields High SensitivityRepresentative Detection Limits: Diatomic Molecules

• Diatomic molecules @ 1500K:‐ sub‐ppm detectivitiy for UV absorbers‐ ppm detectivity for IR absorbers

1000 1500 2000 2500 3000

0.01

0.1

1

10

100

1000

CO

OH

Det

ectio

n Li

mit

[ppm

]

Temperature [K]

1atm,15cm,1MHz

CH

CN

6

Advances in Shock Tube Methodology

1. Improved uniformity with driver inserts

2. Longer test times with tailored gas mixtures & extended driver

3. Reactive gas modeling:problem and solutions

7

Improvement in Reflected Shock Temperature Uniformity Using Driver Inserts

• Conventional  shock tube operation can provide near‐ideal uniform flows for 1‐3 ms

• But, boundary‐layers and attenuation induce dP/dtand dT/dt at longer times

8

dP/dt = 3%/msdT/dt = 1.2%/ms

Improvement in Reflected Shock Temperature Uniformity Using Driver Inserts

• Driver inserts modify flow to achieveuniform T and P at long test times

P5T5

• These effects reduce ignition delay times relative to Constant P case

• Solution: Driver Inserts

VRS

9

dP/dt = 3%/msdT/dt = 1.2%/ms

Improvement in Reflected Shock Temperature Uniformity Using Driver Inserts

• Result:  dP/dt = 0 prior to ignition

Proper ign for comparisonwith Constant P simulations

P5T5VRS

10

Longer Test Times Achievable withTailored Gas Mixtures & Extended Driver Sections

• Conventional shock tube operation: ~ 1‐3 ms test time• No overlap with RCM operation~ 10‐150 ms test time

10

0.4 0.8 1.2 1.61E-3

0.01

0.1

1

10

100

1000

Test

Tim

e, I

gniti

on T

ime

[ms]

1000/T [1/K]

2500K 1000K 625K

UnmodifiedStanford ST

n-Dodecane/Air

10atm

0.4 0.8 1.2 1.61E-3

0.01

0.1

1

10

100

1000

Test

Tim

e, I

gniti

on T

ime

[ms]

1000/T [1/K]

2500K 1000K 625K

UnmodifiedStanford ST

RCM

n-Dodecane/Air

10atm

=1

11

0.4 0.8 1.2 1.61E-3

0.01

0.1

1

10

100

1000

Test

Tim

e, I

gniti

on T

ime

[ms]

1000/T [1/K]

2500K 1000K 625K

UnmodifiedStanford ST

Modified ST

RCM

n-Dodecane/Air

10atm

• Longer driver length and tailored gas mixturescan provide longer test times (> 40 ms)

2x Driver Extension

• Shock tubes now can overlap with RCMs

=1

Longer Test Times Achievable withTailored Gas Mixtures & Extended Driver Sections

12

First Use of Long Test‐Time Facility:Low Pressure n‐Heptane Ignition in the NTC regime

• Test time = 45 ms at 725K

• Enables first low‐pressure (~3 atm) n‐heptane ignition data in NTC regime

• Clear evidence of 2‐stage ignition: 1 = 8 ms, 2 = 20 ms

• Next step: add driver inserts to remove dP/dt & dT/dt

• How do the 3 atm data compare to high P data?

13

0 10 20 30 40 500

1

2

3

4

5

62nd StageIgnition

OH Emission

Non-Reactive Mix w/o Inserts

Rel

ativ

e P

ress

ure

Time [ms]

n-Heptane/21%O2/Ar=1, 725K, 2.9 atm

Reactive Mix1st StageIgnition

Test time = 45 ms

Low‐Pressure, Low‐Temperature Studies:New Results for n‐Heptane in the NTC Regime

• Previous 12‐55 atm data• New 3 atm data• How do measurements compare with current models?

14

1.0 1.2 1.4 1.60.1

1

10

100

55 atm

40 atm

12 atm

1000K 625K 833K

Igni

tion

Del

ay T

imes

[m

s]

1000/T [1/K]

714K

3 atm

n-Heptane/21%O2/Ar,N2 =1

StanfordAachen

Low‐Pressure, Low‐Temperature Studies:NTC Heptane, Comparison with Model

• LLNL C7 (2000) model performs reasonably  well at high P• Further tests needed at low P• Reveals value of long test time experiments

15

1.0 1.2 1.4 1.60.1

1

10

100

40 atm

12 atm

1000K 625K 833K

Igni

tion

Del

ay T

imes

[m

s]

1000/T [1/K]

714K

3 atm

n-Heptane/O2/Ar/N2 =1

Solid:LLNL C7Model

Reactive Gasdynamics Modeling: A Problem

• Most current reflected shock modeling assumes Constant‐Volume or Constant‐Pressure

But:• Exothermic energy release during oxidation or endothermic cooling during pyrolysis changes T & P behind reflected shocks not a Constant‐V or Constant‐P process!

• Example: Heptane Ignition

16

Effect of Energy Release on P Profiles:n‐Heptane Oxidation

• How does this compare with models?• Not a constant P process!

17

-100 0 100 200 300 400 500 6000

1

2

3

4

5

0.4% Heptane/4.4% O2/Ar, =11380K, 2.3 atm

Pre

ssur

e [a

tm]

Time [us]

ign

0

2

4

MeasuredSidewall P

Constant PModel

Effect of Energy Release on P Profiles:n‐Heptane Oxidation

• How does this compare with model?• Not a constant P process!• Not a Constant‐V process, even for 0.4% fuel!• So how can entire process be modeled? 18

-100 0 100 200 300 400 500 6000

1

2

3

4

5

0.4% Heptane/4.4% O2/Ar, =11380K, 2.3 atm

Pre

ssur

e [a

tm]

Time [us]

ign

0

2

4

MeasuredSidewall P

Constant PModel

Constant V Model

3 Proposed Solutions to Enable Modeling through Entire Combustion Event

1. Minimize fuel loading to reduce exothermically‐ or endothermically‐driven T and P changes‐ enabled by high‐sensitivity laser diagnostics

2. Modified gasdynamics modeling to account for P and T change during combustion‐ work in progress (but computationally intensive: 1‐D, 3‐D)

3. Use new constrained reaction volume concept to minimize pressure perturbations‐ enables constant P (or specified P) modeling

Examples:  1) Use of dilute reactive mixtures2) Use of constrained reaction volume

19

Example 1: Benefit of Dilute Mixtures3‐Pentanone Oxidation

• Pressure nearly constant throughout experiment• Good agreement between Constant H,P model and expt.• Model successfully includes temperature change

0 500 1000 15000

50

100

150

OH

Mol

e Fr

actio

n [p

pm]

Time [s]

Const. H PModel

OH Data

0 500 1000 1500

0.0

0.5

1.0

1.5

2.0

Pre

ssur

e [a

tm]

Time [s]

400 ppm 3-Pentanone/O2/Ar1486 K, 1.52 atm, =1

Low Fuel Loading Experiment

20

OH Mole Fraction

Tfinal‐TInitial =45 K

Example 2: Constrained Reaction Volume ApproachHydrogen Ignition at 950 K

• Large reaction volume gives large energy release  P & T

• CRV gives reduced energy release  near‐constant P 

Conventional Shock Tube

21

Constrained Reaction Volume

Helium              Test Mixture

Pre‐Shock

Post‐Shock

Pre‐Shock

Post‐Shock

Helium     Non‐Reactive Mix       TM

Large Region ofEnergy  Release

Small Region ofEnergy Release

ReflectedShock Wave

• Conventional ST exhibits large pressure change!• CRV pressure nearly constant throughout experiment!• Allows kinetics modeling through ignition and combustion! 

Conventional Shock Tube

22

Constrained Reaction Volume

0 1 2 3 4 50

2

4

6

8

Pressure

4%H2-2%O2-ArNormal Filling979K, 3.44 atm

Pres

sure

[at

m]

Time [ms]

Emission

0 1 2 3 4 50

2

4

6

8

Pressure

4% H2-2%O2-ArCRV = 4 cm956K, 3.397 atm

Pre

ssur

e [a

tm]

Time [ms]

Emission

Ignition

Example 2: Constrained Reaction Volume ApproachHydrogen Ignition at 950 K

Elementary Reaction RateDeterminations

GoalNear‐direct determination of rate constants for specific reactions

ExamplesOH + ketones:

acetone, 2‐butanone, 2‐ & 3‐pentanoneOH + alkanesOH + butanol isomersOH + methyl esters

23

Experimental Strategy

fast upon shock heating

tert‐butylhydroperoxide

• TBHP used as a prompt OH precursor• Useful T range (850 to 1350 K) • Pioneered by Bott and Cohen (1984)• Also used at Argonne

• Fuel in excess, pseudo‐first order experiment

24

Acetone

+ +

Representative 3‐Pentanone+OH Data: 1188 K and 1.94 atm

0 50 100

0

5

10

15

20

OH

Mol

e Fr

actio

n [p

pm]

Time [s]

Current Study 1.23x1013 cm3 mol-1 s-1

211 ppm 3-Pentanone / Ar17 ppm TBHP1188 K, 1.94 atm

0 50 100-4

-3

-2

-1

0

1

OH

Sen

sitiv

ityTime [s]

3-Pent+OH=Products CH3+OH=CH2*+H2O C2H5=C2H4+H

• High‐quality data allows high‐precision comparison with model• Sensitivity analysis confirms pseudo‐first order behavior• Near‐direct determination of reaction rate constant!

25

Results for 3‐Pentanone + OH  Products

• No other high temperature data available

• NUI model (Serinyel et al. ) in excellent agreement 

0.7 0.8 0.9 1.0 1.1 1.21E12

1E13

1E14

833K1429K 909K1000K1111K

Current Study Serinyel et al. - NUI Galway (2010)

k [c

m3 m

ol-1 s

-1]

1000/T [1/K]

3-Pentanone + OH = Products

1250K

26

0.7 0.8 0.9 1.0 1.1 1.21E12

1E13

1E14

833K1429K 909K1000K1111K

Current Study

k [c

m3 m

ol-1 s

-1]

1000/T [1/K]

3-Pentanone + OH = Products

1250K

+/‐ 20%

Summary: OH+KetonesProducts

• How do data compare with Structural Activity Relationship (SAR) model?• Data agree within 25% with SAR‐estimated rate constants• Similar measurements performed with methyl esters

27

0.7 0.8 0.9 1.0 1.1 1.21E12

1E13

833K909K1000K1111K1250K

Acetone 2-Butanone 3-Pentanone 2-Pentanone

k Ket

one

+ O

H [c

m3 m

ol-1 s

-1]

1000/T [1/K]

Ketone + OH = Products

1429K

Lines: Modified SAR (SAR x 0.75)

Summary: OH+Methyl Esters  Products

• Data agree within 25% with SAR‐estimated rate constants• Current work:  OH + aldehydes, alcohols

28

0.7 0.8 0.9 1.0 1.1 1.2

1E12

1E13

833K909K1000K1111K1250K

MButanoate MPropanoate MFormate MAcetate

Lines: Modified SAR (SAR x 0.75)

k Met

hyl E

ster

+ O

H [c

m3 m

ol-1 s

-1]

1000/T [1/K]

Methyl Ester + OH = Products

1429K

Multi‐Species Time‐Histories

MotivationMulti‐species provide greater constraint on mechanism refinement/evaluation 

Recent WorkKetones: Acetone, Butanone, 3‐Pentanone

Alcohols: 1‐, 2‐, tert‐,& iso‐Butanol

Alkanes: n‐Hexadecane

Esters: Methyl Formate, MA, MP, MB, EP

29

Multi‐Species Approach: 3‐Pentanone Pyrolysis 

3‐Pent = 3.39 m

CH3 = 216 nm

C2H4 = 10.5 m

CO = 4.6 m

0 100 2000

100

200

3000.1% 3-Pentanone / Ar

.

1346 K, 1.67 atm

CH

3 Mol

e Fr

actio

n [p

pm]

Time [s]

0 500 1000 15000.0

0.5

1.0

1.5

2.0

1343 K, 1.60 atm

C2H

4 Mol

e Fr

actio

n [%

]

Time [s]

1% 3-Pentanone / Ar

0 500 10000.0

0.4

0.8

1.2 1% 3-Pentanone / Ar

3-

Pen

tano

ne M

ole

Frac

tion

[%]

Time [s]

1323 K, 1.32 atm

0 500 10000

1000

2000

3000

1325 K, 1.60 atm

CO

Mol

e Fr

actio

n [p

pm]

Time [s]

0.25% 3-Pentanone / Ar

30Excellent SNR, high sensitivity data

Comparison with Serinyel et al. (2010) Galway NUI

3‐Pent CH3

C2H4CO

0 100 2000

100

200

3000.1% 3-Pentanone / Ar

Serinyel et al.

1346 K, 1.67 atm

CH

3 Mol

e Fr

actio

n [p

pm]

Time [s]

0 500 1000 15000.0

0.5

1.0

1.5

2.0

1343 K, 1.60 atm

C2H

4 Mol

e Fr

actio

n [%

]

Time [s]

1% 3-Pentanone / Ar

0 500 10000.0

0.4

0.8

1.2 1% 3-Pentanone / Ar Serinyel et al.

3-

Pent

anon

e M

ole

Frac

tion

[%]

Time [s]

1323 K, 1.32 atm

0 500 10000

1000

2000

3000

1325 K, 1.60 atm

CO

Mol

e Fr

actio

n [p

pm]

Time [s]

0.25% 3-Pentanone / Ar Serinyel et al.

Two Differences: 1) 3‐P decomposition rate;  2) CO/C2H4 yields31

3‐P Data Enables Revision of Decomposition Rate

3‐P data show strong sensitivitiesto k1 + k2

3‐Pentanone Sensitivity

Revised ktotal 3.5x Serinyel et al. rate

CO yields still not correct!

Arrhenius Plot: 3‐Pent → Products

0.7 0.8 0.91

10

100

1000

10000

100000

1111K1250K

k tota

l [1/s

]1000/T [1/K]

3-Pentanone = Products1.6 atm

Serinyel et al. (2010)

1429K

32

2 Channels

CO Yield Resolved through Use of O‐Atom Balance

• Model underpredicts CO and overpredicts methyl ketene

• Why?   Methyl ketene decomposition pathway missing in mechanism

• Introduce CH3CHCO  C2H4 + CO (assume k the same as ketene decomp.)33

3‐P and CO data yield total O‐atoms O‐atom concentration at 1.5 ms

Laser Absorption3‐Pentanone:CO:                  Sum:      

23%69%

Simulation (with new k1+k2)3‐Pentanone:CO: CH3CHCO:                 

Sum:                        93%

23%43%27%

92%

0 500 1000 15000.0

0.2

0.4

0.6

0.8

1.0

Mol

e Fr

actio

n [%

]

Time [s]

1% 3-Pentanone / Ar1248 K, 1.6 atm

3-PentanoneCO

Revised Model Improves 3‐P Simulations

CH3

CO

0 500 10000

1000

2000

3000

1325 K, 1.60 atm

CO

Mol

e Fr

actio

n [p

pm]

Time [s]

0.25% 3-Pentanone / Ar.

1215 K, 1.68 atm

1403 K, 1.57 atm

1272 K, 1.64 atm

Final Modifications to Serinyel et al.3‐Pentanone Mechanism• Revised decomposition rate:

3‐pentanone  products• Additional reaction: 

methylketene C2H4 + CO

34

C2H4

• Good agreement with 3‐P, CH3, CO, and Low‐T C2H4 time‐histories

Ongoing Work

• Diagnostics development & spectroscopy: – aldehydes (CH2O, CH3CHO)– alkyl radicals (C2H5)– methyl esters (methyl formate)– Alkenes (C3H6, C4H8)

• Continued improvement of shock tube methods:– constrained reaction volume

• Direct measurement of elementary reactions:– OH + oxygenates (aldehydes, ethers, alcohols)– decomposition of oxygenates– CH3, HO2 reactions with HC

35

Acknowledgements

• ARO, AFOSR, DOE, NSF

• Students:• Genny Pang, Matt Campbell, Wei Ren, Brian Lam, Sijie Li, 

Sreyashi Chakraborty

36