Tanya Kerr, Keith Duncan, Leary Myers Department of Physics, UWI

Materials Science Research Group Department of Physics, UWI (Mona)

Post Fire Materials Identification by Confocal Raman Microscopy:

Improved Fire Modelling and Investigation

Tanya Kerr, Keith Duncan, Leary Myers

Materials Science Research Group Department of Physics, UWI (Mona)

80% of cause unknown

Source: JFB

Jamaica

36.4 % of house fires; 47% of fatal fires; cause unknown

Source : USFA

USA

16% cause of ignition unknown; 58% location of origin unknown; 15% first material ignited unknown

Source: Branz Ltd

Canada

Materials Science Research Group Department of Physics, UWI (Mona)

Materials Science Research Group Department of Physics, UWI (Mona)

Complexity of the scene

Training/mentorship

Need for new/improved

investigative

tools

• Systematically examine compartment fire patterns

• Excavate and determine fire origin

• Come up with 2 or 3 hypotheses as to cause by application of fire science

• Verify cause by application of fire dynamics

Materials Science Research Group Department of Physics, UWI (Mona)

Interpretation of Fire

Dynamics

Physical Reconstruction Fire Modelling

Materials Science Research Group Department of Physics, UWI (Mona)

Materials Science Research Group Department of Physics, UWI (Mona)

MATERIALS IDENTIFICATION

Heat Release

Rate

Building Design

Materials Science Research Group Department of Physics, UWI (Mona)

Spruce

• Heat release rate in compartment increased more rapidly in the compartment showing a very rapid fire growth and spread

Plywood

• There was a significant delay in the peak heat release rate in the compartment and that there was a much slower fire growth and spread

Chen et al. 2010

Materials Science Research Group Department of Physics, UWI (Mona)

Comprehensive Fire Investigation

Hazard Determination

Support in Origin

Determination

Improved Fire

Modelling

How do we identify them?

Materials Science Research Group Department of Physics, UWI (Mona)

Materials Science Research Group Department of Physics, UWI (Mona)

Identify the Original Material and the accelerant used by Raman Spectroscopy of the Fire Debris

Polypropylene, polystyrene and Nylon

After burning with accelerants collected new Raman Data for burnt samples

Use of chemometric, principal component analysis (PCA) to show material discrimination

Had some success, but results were plagued by sample fluorescence, so not many peaks for discrimination

Gonzalez-Rodriquez et al. 2011

Materials Science Research Group Department of Physics, UWI (Mona)

Polyethylene terepthalate

(PET)

Ethylene Vinylacetate

(EVA)

polypropylene

Polystyrene

Nylon

• Curtains • Table Cloths

• Shower Curtains • Baby Bibs

• DVD Cases • Mats

• CD Cases • Plates Stockings

• Collect Raman spectra for each material type in their original state

• 5 x 5cm squares of

each sample was used for testing.

• Samples placed in to a metal can which was divided in two parts by metal mesh wire and burnt in air

• Raman spectra obtained at five points with an average of three replicates from the same spot.

• Comparison of spectra to determine loss

• Apply PCA to the

peaks after samples are burnt for discrimination between samples

Crushed Newspaper

Sample material

Materials Science Research Group Department of Physics, UWI (Mona)

625 800

1006 1035 1606

629 804

1012

1194 1463

1616

-200

0

200

400

600

800

1000

1200

1400

1600

1800

01000200030004000

Inte

nsi

ty

Raman shift cm-1

Raman Spectra before and after burning for polystyrene foam

burnt

unburnt

985 1077

1129 1447

1640

941 1089 1136

1454 1649

-500

0

500

1000

1500

2000

2500

3000

3500

4000

0500100015002000250030003500

Inte

nsi

ty

Raman shift cm-1

Nylon comparison spectra for burnt and unburnt

burnt

unburnt

401 813 846

1066 1131 1299

1443

400 809 844 1152 1333

1459

1528

-200

0

200

400

600

800

1000

1200

1400

1600

1800

01000200030004000

Inte

nsi

ty

Raman shift cm-1

Polypropylene burnt and unburnt

unburnt

burnt

639 1065

1131 1298 1442

622 1029 1093

1253 1391

0

500

1000

1500

2000

2500

3000

01000200030004000

Burnt EVA and unburnt comparison

burnt

unburnt

O

N

H

N

H

n

H

HH

C

H

C

OC

CH3

C

H

C

H

CH3

Materials Science Research Group Department of Physics, UWI (Mona)

PCA_1_Axis_1 vs. PCA_1_Axis_2

PCA_1_Axis_1

543210-1-2

PC

A_1_A

xis

_2

2

1

0

-1

-2

-3

PS f oam

PP

Ny lon

PET

EVA

Materials Science Research Group Department of Physics, UWI (Mona)

Materials can be identified by Confocal Raman Spectroscopy

Burnt materials have unique spectra

More accurate data can be provided for fire modelling

Fluorescence is still an issue to be resolved for burnt debris

Materials Science Research Group Department of Physics, UWI (Mona)

• Determine the best method for minimizing fluorescence for burnt debris

• Examine material combinations, Raman Mapping

• Gathering Raman spectra for burnt materials at different time frames

• Exposure time assignment

• Compartment mapping