Combustion Chemistry of a New  Biofuel: Butanol

William H. Green, D.F. Davidson, F. Egolfopoulos,  N. Hansen, M. Harper, R.K. Hanson, S. Klippenstein,  C.K. Law, C.J. Sung, D.R. Truhlar, & H. Wang

Assessing Alternative Fuels:  The Challenge of Predicting Performance

• Hundreds of possible alternative fuels: What do we want?

– We can specify the fuel composition• Sometimes precise control: bioengineering, catalysis from H2

– Which fuels give special performance advantages?– Do any enable advances in engine technology?– Performance gain could help biofuel

enter market.

• Predict performance of a new fuel in a new engine?…or do we need to experimentally test every 

possible engine/fuel combination?– Computer simulations of combustion becoming practical.– Rate constants, thermo from quantum calculations– Validate predictions with experiments

CEFRC’s

First Test Case: Butanol Why Butanol?

• Bio‐butanol

is about to be commercialized– Strong demand for an accurate computer model

• Butanol

has some advantages over ethanol– More soluble in gasoline

• Can be used at higher blending ratios• Can be shipped in gasoline pipelines

– Lower vapor pressure: less smog• When CEFRC started, very few data or models 

available on butanol

combustion• Small molecule, but big enough

– Simpler chemistry than e.g. diesel range fuels– Feasible to do high‐accuracy quantum chemistry– Experimentally convenient: easy to vaporize– Interesting: Several isomers with quite different chemistry

Dozens of Center members all  studying butanol

in parallel

Tied together by connections with the MIT  team assembling the kinetic model

• C.K. Law (Princeton) flame speeds, flame ball• F. Egolfopoulos (USC) flame speeds, extinction• Nils Hansen (Sandia, ALS) species in flames• S. Klippenstein (Argonne) quantum chemistry• Ron Hanson (Stanford) ignition delays, species• C.J. Sung (Connecticut) ignition delaysOther data and calculations not shown because of the time…

As examples, here we show data and calculations from these groups:

Flame Speeds for Different Fuels: Not Intuitive. Need a Model!Flame Speeds for Different Fuels: Not Intuitive. Need a Model!

Methanol

Ethanol

n-Butanol

n-Butanolsec-Butanol

iso-Butanol tert-Butanol

n-Propanol

iso-Propanol

Propane Butanone

Acetone

Measured in Egolfopoulos lab (USC). Air, 1 atm.

Flam

e S

peed

Equivalence Ratio

Different Reaction Paths Important at different combustion conditions

Big consequences: ignition, soot-formation, toxic emissions

Thermal Decomposition of C4

H9

O radical  critical in n‐butanol

combustion

n‐butanol

α‐radical

β‐radical

γ‐radical

δ‐radical1‐butoxy

H abstraction

by H, OH, O, …

Isomerization

Decomposition: 

β‐scission, cyclization,

C–C, C–O bond fission 

Quantum Chemistry allows us to compute rates  k(T,P) of all C4

H9

O Decomposition Reactions

B3LYP/6‐311++G(d,p)CASPT2 (3e,3o)/ADZ

CH3CH2CHCH2OHCH3CH2CH2CHOH CH3CHCH2CH2OH

CH2CH2CH2CH2OH CH3CH2CH2CH2O

-25.1-28.9 -26.0

-21.7 -21.6

13.711.9

-2.0vdW

C4H8 + OH0.0

15.6

5.67.7

12.7

15.3

-4.1

-11.8

-2.9 butanal + H

3.2 7.0 1-butenol + H

-12.3 ethyl + ethenol

16.0 cyclobutanol + H

18.6 methylcyclopropanol + H

9.3 2-butenol + H

-4.6 CH3 + 2-propen-1-ol

19.8 cyclopropymethanol + H21.6 epoxybutane + H

10.9 3-butenol + H

-4.3 ethylene + ethanol radical

19.4 2-methyl-oxetane + H

-10.1 propene + CH2OH

-12.9 pentyl + CH2O

6.8 tetrahydrofuran + H

8.3

-2.4

45.5

35.1

8.63.6

10.5

32.9

32.6

11.611.3

-1.4

37.5

31.4

13.4

3.5

36.7

45.0

25.9

1.7

-8.2

35.3

43.9

32.3

-2.9

-1.5

Calcs by S. Klippenstein (Argonne)

Combustion has been studied for a  long time, but still quite challenging

Photograph of a Butanol/Air flame in internal combustion engine conditions(measured in C.K. Law’s lab, Princeton)

Current CEFRC Chemistry Model:

334 species reacting via7113 reactions

4288 of the reactions havecomplicatednon-Arrhenius k(T,P)

Numerical issues in solvingthe model, but possible to do it.Tricky to know true boundaryconditions in many combustionexperiments.

Most important fuel performance  property: ignition

• Gasoline “Octane Number”

• Diesel “Cetane

Number”

• Small changes in fuel make big changes in  ignition: sensitive to molecular structure!

• New engines under development are even  more sensitive to ignition

– Potential for big gains…

but only if the fuel ignition  delay time matches engine requirements

Stanford Shock Tube Measurements: Butanol

Ignition Delay Time

• Ignition delay time database provides wide coverage of pressure and temperature

• Tert‐butanol

has significantly longer ignition delay times than other isomers

0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00

100

1000

1000 K1111 K1250 K1429 K

43 atm19 atm

3.0 atm1.5 atm

P = 1.5 atm, = 1 P = 3.0 atm, = 1 P = 19 atm, = 0.60-0.75 P = 19 atm, = 1 P = 43 atm, = 0.75-1.0 P = 43 atm, = 1

Lines = Grana et. al. (2010)

t ign [u

s]

1000/T5 [1/K]

10

100

1000

1667 K 1190 K1471 K 1316 K

1-butanol 2-butanol i-butanol t-butanol

Lines - Grana et. al.(2010)

t ign [u

s]1000/T5 [1/K]

0.60 0.64 0.68 0.72 0.76 0.80 0.84

t-but i-but

2-but 1-but

N‐Butanol Butanol

Isomers

4% O2

/Ar1.5 atm

4% O2

/Ar

0.6 0.65 0.7 0.75 0.8 0.8510

1

102

103

1000 K / T

Igni

tion

dela

y /

s

= 1.0

= 0.5

= 0.25

0.55 0.6 0.65 0.7 0.75 0.8 0.8510

1

102

103

1000 K / T

Igni

tion

dela

y /

s

= 1.0

= 0.5

= 0.25

0.6 0.65 0.7 0.75 0.8 0.8510

1

102

103

1000 K / T

Igni

tion

dela

y /

s

= 1.0

= 0.5

= 0.25

0.5 0.6 0.7 0.8 0.910

1

102

103

1000 K / T

Igni

tion

dela

y /

s

= 1.0

= 0.5

= 0.25

CH3 OH CH3CH3

OH

OHCH3

CH3

CH3 CH3

CH3

OH

12

Model accurately predicts Stanford high-T ignition delays

Engines Sensitive to Ignition at High  Pressure and “Low”

Temperature

• Investigation of low-temperature, high- pressure autoignition is relevant to many emerging engine technologies

• GCMS sampling results prior to experiments confirm mixture composition preparation procedure when handling liquid fuels

• Gas samples from the reaction chamber for GC-MS/FID analysis allow identification and quantification of important species during autoignition

Rapid Compression Machine

C.J. Sung group, U. Conn.

Autoignition

Studies of Butanol

Isomers at  High Pressure and Low Temperature

10

100

1.15 1.2 1.25 1.3 1.35 1.4

Butanol/O2/N2, =1.0, PC=15 bar

sec-Butanoliso-Butanol

n-Butanol

tert-Butanol

Igni

tion

Del

ay, m

s

1000/TC, 1/K

O2 : N2 = 1 : 3.76

784

K

758

K

737

K

725

KTC

0

5

10

15

20

25

30

35

-20 0 20 40 60 80

n-Butanol/O2/N2, =1.0, PC=15 bar

Pres

sure

, bar

Time, ms

O2 : N2 = 1 : 3.76

816

K839 K

Non-Reactive Case

n-butanol ignition is much faster than the other butanol isomers for T< 900 K

Measured in C.J. Sung lab (U.Conn.)

Big Discrepancy: Model did not predict the fast  n‐butanol

ignition observed at T < 900 K!

Model was built automaticallyusing computer “expert system”

Due to mistake in ratedatabase used by expert system, model wildlymis-estimated barrier forHO2 + C-H reactions.

Big Discrepancy: Model did not predict the fast  n‐butanol

ignition at T < 900 K!

Model was built automatically at MITusing computer “expert system”

Due to mistake in rate database used by expert system, model wildlymis-estimated barrier forHO2 + C-H reactions.

After correcting that big mistake, current CEFRCmodel is much closer……but still not quite rightat lowest temperatures.Work continues…

Multi‐Species Time‐Histories Provide Stronger, More  Informative Tests of Model:  n‐Butanol

Pyrolysis

• Current status:  Stanford species time‐history measurements for OH, H2

O and CH2

O

• Next step:  laser absorption measurements of CO and C2

H4

• All models underpredict

formaldehyde and H2

O formed in high‐T n‐butanol

pyrolysis

OH CH2

OH2

O

0 100 200 300 400 500

0

2

4

6

8

10

12

1% n-Butanol in ArT = 1342 KP = 1.51 atm

Measured LLNL (Curran) U of Toronto (Sarathy) MIT (Harper) Italy (Grana) RPI/France (Moss)

OH

[ppm

]

time [s]0 200 400 600 800 1000

0.0

0.1

0.2

0.3

0.4

0.5 Measured LLNL (Curran) U of Toronto (Sarathy) MIT (Harper) Italy (Grana) RPI/France (Moss)

H2O

[%]

time [s]

1% n-Butanol in ArT = 1342 KP = 1.51 atm

0 200 400 600 800 1000

0.0

0.1

0.2

0.3

0.4 1% n-Butanol in ArT = 1352 KP = 1.48 atm

Measured LLNL (Curran) U of Toronto (Sarathy) MIT (Harper) Italy (Grana) RPI/France (Moss)

CH

2O [%

]

time [s]

Advanced Light Source allows direct detection  of dozens

of species including key radicals

Photoionization

Molecular Beam Mass Spectrometry

Flames are analyzed with 

molecular beam time‐of‐flight 

mass spectrometry 

Photoionization with tunable 

synchrotron‐generated VUV 

photons allows identification of 

species

by mass

by ionization energy

Experimental mole fraction 

profiles are compared with 

flame model predictions 

and reaction path and 

sensitivity analysis are 

performed

Advanced Light Source (ALS) Flame Data: Detailed Test of the Model’s Predictive Capabilities

Mole fraction profiles of the major species are 

predicted accurately

A more powerful test is provided by 

comparing modeled and experimental profiles 

of intermediate species

Profiles have not been shifted

Oßwald et al. flame data need to be 

shifted for better agreement

Only a few of the many data traces shown here… most show good agreement

One big discrepancy identified between model  and ALS data: C4

H4

and C3

H3

overpredictedSensitive to C4

H5

Thermochemistry

Simulations of the flames studied by ALS are sensitive to the enthalpy of 

formation of i‐C4

H5

(CH2

=CH‐∙C=CH2

∙CH2

‐CH=C=CH2

) . None of the other 

available experimental data are sensitive to this number.

This radical’s enthalpy value was incorrect in the MIT database. Correcting to 

the accepted literature value largely resolved the discrepancy.

Now investigating origins of smaller discrepancies…

Conclusions• Teams can move fast! 

– In less than 2 years, went from zero to quite good model for 

butanol

combustion, comprehensively validated by many 

different types of experiments. – Contrast: sequential single‐investigator model‐building and 

comprehensive validation usually takes decades.– Certainly we can do it even faster next time through– Very promising for other alternative fuels: do you have a fuel 

we should model?• Focus on the discrepancies

(models vs. expts, and expts

vs. expts.): that is where there is an opportunity to  learn something!

– Don’t be shy: expose the discrepancies to your EFRC team‐

mates!