Measurements of light absorption spectra of fine particle aqueous extracts during CalNex

at the Pasadena ground site

X. Zhang and R. J. Weber

Georgia Institute of Technology

CalNex Data Analysis Workshop

May 16-19, 2011

Sacramento, CA

Real-time light absorption spectral measurement

(Hecobian et al., 2010)

Long Optical Path Spectrometry With Liquid Wave guides (LWCC)

1 meter Waveguide Capillary Cell

• Complete absorption spectra (200 – 800nm) was saved every 15 min• Absorption coefficients at selected wavelengths (365nm, 700nm) were saved every 60 sec

Simultaneous measurements of soluble particle absorption spectra and carbon mass

Choice of wavelength: λ = 365nm

λ = 365nm : Absorption averaged between 360 and 370nm

λ = 700nm: Reference wavelength

HULIS

• Soluble Brown Carbon, specifically associated with HULIS (Lukacs et al., 2007);• Avoid interferes from other species: e.g. Nitrate;

(Hoffer et al., 2006)

Nitrate

-15x10-3

-10

-5

0

5

10

Abs

orpt

ion

6:00 PM6/2/2010

9:00 PM 12:00 AM6/3/2010

3:00 AM 6:00 AM

Blank measurement

λ = 700

λ = 365 Absorption coefficient at 365nm (Abs365)

Absλ = (A λ – A700)Vl

Va · l· ln(10)

Unit: m-1

Sources of brown carbon: PMF result on FRM filters

• Biomass burning;• Primary emissions from vehicle;• SOA formation (WSOC/Oxalate);

Major sources of Brown Carbon (Abs365) in the southeast:

(Hecobian et al., 2010)

PMF analysis on 900 FRM filters collected at 15 sites in SE US

Possibly related to aqueous SOA formation/chemical aging – oxalate and brown carbon peak hours after WSOC peak on diurnal average from online dataset in Atlanta.

Biomass Burning

ResidualRefractory MaterialWSOC/Oxalate

Ammonium Sulfate/WSOCMobile Sources

13

12

11

10

9

8

WS

OC

p/C

O

6:00 AM1/1/1904

12:00 PM 6:00 PM

0.12

0.11

0.10

0.09

0.08

Oxa

late

/WS

OC

p

400x10-9

380

360

340

320

300

280

260

Abs

365 /WS

OC

p

Abs/WSOCp

Oxalate/WSOCpSunrise Sunset

WSOCp/CO

3

2

1

0

WS

OC

, µgC/m

-3

6/3/2010 6/5/2010 6/7/2010 6/9/2010 6/11/2010 6/13/2010 6/15/2010PDT Local Time

4x10-6

3

2

1

0

Abs

365,

m-1

4

3

2

1

0

RS

D_W

SO

C, µgC

/m-3

5/21/2010 5/26/2010 5/31/2010 6/5/2010 6/10/2010PDT Local Time

4x10-6

3

2

1

0

RS

D_A

bs36

5, m

-1

WSOC and brown carbon during CalNex

Caltech site: Jun. 01 – Jun. 15

Abs365WSOC

Riverside campus: May. 17 – Jun. 14

1.4x10-6

1.2

1.0

0.8

0.6

Abs

365,

m-1

6:00 AM 12:00 PM 6:00 PMLocal Time (PDT)

0.9

0.8

0.7

0.6

0.5

EC

, µgC/m 3

2.0

1.8

1.6

1.4

1.2

1.0

0.8

0.6

WS

OC

, µgC/m 3

CalNex: mass absorption efficiency

Abs365

WSOCEC

“primary” secondary

4x10-6

3

2

1

0

Abs

365

Sec

onda

ry

43210WSOC Secondary

Slope = 7.1e-007 R

2 = 0.38

αabs = 0.71 m2/g

Mass absorption efficiency (αabs) ΔAbs365/ ΔWSOC

Small compared to αabs for soot ~ 7.5m2/g (Bond and Bergstrom, 2006)

at 360-370nm

Mass absorption efficiency: LA compared with Atlanta

0.50

0.45

0.40

0.35

0.30

0.25

EC

, µgC

/m3

6:00 AM 12:00 PM 6:00 PM

7.5x10-7

7.0

6.5

6.0

5.5

5.0

Abs

365,

m-1

2.3

2.2

2.1

2.0

1.9

1.8

WS

OC

, µgC/m

3

Abs365

WSOCEC

Gatech campusin Atlanta, Aug. 2010

primary secondary

4x10-6

3

2

1

0

Abs

365

Sec

onda

ry

43210WSOC Secondary

Slope = 7.1e-7 Slope = 1.8e-7

LA fresh secondary WSOC is much more light-absorbing than Atlanta

Mass absorption efficiency

ATL: αabs = 0.18 m2/g

LA: αabs = 0.71 m2/g

LA 4 x higher

DAbs

Angström exponent

Wavelength dependence of the absorption coefficient

αab = K · λ-Å

For ambient aerosol Å ~1: black carbon (absorbs at all wavelength) Å ~2: ambient biomass burning aerosol (Kirchstetter et al., 2004) Å ~3.5: polluted region in China (Yang et al., 2009)

3.0x10-6

2.5

2.0

1.5

1.0

0.5A

bsor

ptio

n

440400360320Wavelength

2

3

4

567

10-6

2

3

Abs

orpt

ion

460440420400380360340320300Wavelength

Plot on Ln-Ln scale

Slope = -Å

For liquid extracts (this study) Å ~7: water-soluble HULIS (Hoffer et al., 2006) Å ~7-16: smoldering smoke (Chen and Bond, 2010) Å ~7: fresh limonene SOA (Bones et al., 2010) Å ~4.7: aged limonene SOA (Bones et al., 2010) Å ~6-8: FRM filters in SE US (Hecobian et al., 2010)

Variation of Angström exponent

Example: 6.3.2010 – high concentration of Abs365

• Å varies between 2 and 5.5, lower during daytime and higher at night; Å increased from “fresh” to “aged” SOA?

• Å derived from online measurement significantly lower than Å from filter-based measurement (discussed in the next slide);

2.0

1.8

1.6

1.4

1.2

1.0

0.8

0.6

EC

, µgC/m

3

12:00 AM6/3/2010

6:00 AM 12:00 PM 6:00 PM 12:00 AM6/4/2010

3.0

2.5

2.0

1.5

1.0

0.5

WS

OC

, µgC/m

3

6

5

4

3

2

Ang

stro

m E

xpon

ent

Angstrom exponent WSOC EC

Filter-based vs. online absorption measurements

Possible artifacts of filter-based brown carbon measurement:• Storage: 1-yr in the freezer leading to changes?• Extraction: extraction/sonication leading to degradation of larger chromophores which absorb light at higher wavelengths? (Sun et al., 2007)• Time resolution: filter liquid extracts sit for 1-2 days before analysis

Comparing Absorption coefficient (Abs365) and Angström exponent

• Some underestimation of Abs365 by high-vol filter;• Online and filter Å quite different; For filter less absorption at higher wavelengths;

3.0x10-6

2.5

2.0

1.5

1.0

0.5

Abs

orpt

ion,

m-1

440400360320Lambda

Online average 6/3 Filter 6/3

-16.0

-15.5

-15.0

-14.5

-14.0

-13.5

-13.0

LN(A

bs)

6.16.05.95.85.7LN(lambda)

Å = 3.3

Å = 7.7

CalNex Hivol Filter

CalNex Online

Summary and Implications

Summary

Why care about Brown Carbon? • Generally not thought to have effect on radiative forcing due to small mass

absorption efficiency (Andreae & Gelencser, 2006), BUT• SOA properties, source and evolution: - Component of fresh anthropogenic SOA (related to aromaticity?) (Sun et al., 2007),

useful for contrasting SOA in different urban settings (e.g. LA vs ATL); - Dissolve in liquid droplets, affect cloud absorption (Mayol-Bracero et al., 2002) & useful to study heterogeneous processing;

• Major sources of brown carbon during CalNex is SOA formation and mobile emission ;• Fresh LA WSOC is ~4 times more light-absorbing than fresh ATL WSOC, possibly due to a

larger fraction of anthropogenic aerosol in LA;• Angström exponents lower during daytime and higher morning and night (evidence for

anthropogenic SOA evolution?);• Filter-based measurement not consistent with online measurement (more experiments)

Future work/collaboration• Try to find correlation between brown carbon and specific SOA component?