Generation of Hydrogen Peroxide by Urban Aerosols and Behavior

of Hydrogen Peroxide in the VACES

Ying Wang, Chuautemoc Arellanes, and Suzanne

PaulsonUCLA Department of Atmospheric and

Oceanic Sciences

Supported by the California Air Resources Board and the South Coast Air Quality Management District

Outline• Introduction: reactive oxygen species• Ambient particles generate hydrogen

peroxide

• A few related measurements

• Working out the mechanism • Gas phase hydrogen peroxide in the

virtual aerosol concentration enrichment system (VACES)

______________________________________________

Reactive Oxygen Species

O2 O2.- HO2

H2O2 OH. H2O

Other ROS: ClO-, O(1D)

Reactive Oxygen Species (ROS) May Play a Role in Particle Toxicity• ROS are generated by lung tissues in

response to foreign material, but sometimes this process gets out of control, resulting in a state of oxidative stress and inflammation.

• ROS have been implicated in respiratory diseases such as asthma, pulmonary and circulatory morbidity and mortality and in carcinogenesis.

______________________________________________

Hydrogen peroxide, like other gasses, follows Henry’s Law:

[X]aq = HxPx(atm)

• [X]aq = concentration of X in liquid phase

• Hx=The Henry’s Law Coefficient for X;

• H(hydrogen peroxide) = 105 M/atmosphere

• Px= The gas phase concentration of X, in atmospheres.

Gas-Liquid Partitioning (Henry’s Law)______________________________________________

• Most gas-phase H2O2 should be absorbed in the upper airways, but particles can deliver peroxide to the lungs.

• Due to ppb-levels of hydrogen peroxide in air, peroxides in airborne water should be ~ 0.1 mM.

• Our measurements indicate ambient particles generate H2O2 in aqueous solution, producinglevels equivalent to > 50 mM in aerosol liquid water.

Particles Provide a Way to Deliver H2O2 to the Lungs______________________________________________

Does H2O2 Contribute to Particle Associated Health Effects? In Vitro and in Vivo Results

• Exposure of lung epithelial cells to 20 pM - 1 µM hydrogen peroxide solutions results in significant cell damage.1

• Morio et al. (2001) exposed rats to ammonium sulfate aerosols (100-200 × ambient), and H2O2 (10 - 100 ×ambient) alone or in combination, for 2 hours. These aerosols contained 10-180 ng/m3 H2O2, according to our measurements & consistent with Henry’s law.

• Ambient aerosols generate H2O2 while ammonium sulfate aerosols do not. Ambient aerosols have 10-~120 ng/m3 H2O2 associated with them.

______________________________________________

• All but the most minor effects were observed only when rats were exposed to H2O2 and particles in combination.

• Effects:• Increased numbers of neutrophils in pulmonary capillaries.• Production of tumor necrosis factor α by alveolar

macrophages.• Increased production of superoxide by alveolar

macrophages.• Increased expression of antioxidant heme oxygenase-1 by

stimulated alveolar macrophages

In Vivo data are Consistent with Significant Effects from Aerosol-Generated ROS at

Levels Common in Urban Areas______________________________________________

A role in aerosol aging?

The quantity of H2O2 produced by ambient particles is comparable (larger)

than the flux of OH radicals to the surface of fine mode particles

Field Measurements of Particle-Associated Hydrogen Peroxide

Measurement of Peroxides Associated with Measurement of Peroxides Associated with AerosolsAerosols

Collect particles on Teflon filters, add stripping solution (0.1 mM Na2EDTA, various pHs, 4 mL), extract for two hours with gentle agitation.

______________________________________________

0 50 100 150 200 2500.0

2.0x10-7

4.0x10-7

6.0x10-7

8.0x10-7

Fine Mode Time Analysis

Fine Mode Exponential Fit Blank

[H2O

2] (M

)

Time (min)

Monitor H2O2 using HPLC/fluorescence via the reaction of H2O2 with horseradish peroxidase and p-hydroxyphenyl-acetic acid

OHO

OH

ROOH

OOH

OH

OHO

OH

ROH

OOH

OH

OHO

OH

OO

OH

OO

OH

-2

++

pH 10

Horseradish Peroxidase

Validation of Aerosol Sampling Method

Hasson, A.S. and Paulson, S.E. J. Aerosol. Sci. 34, 2003, 459

• Aqueous H2O2 aerosols were generated with a Collison nebulizer, diluted, collected on filters and extracted. Final [H2O2] were within 20 % of the expected values, indicating that H2O2solutions do not decompose appreciably on Teflon filters.

• Gas Phase H2O2 is not collected on second filter

______________________________________________

PEROXIDE SAMPLING: GAS PHASE

• Gas phase peroxides are partitioned into the aqueous phase

• Collection efficiency >95%

Source: Hartkamp & Bachhausen Atmos. Environ., 1987, 21, 2207

______________________________________________

Virtual Impactor Sample Collection______________________________________________

Fine (blue),Coarse(yellow)

Hydrogen Peroxide

H2O2 Generation by Aerosols Varies by Site______________________________________________

May 11May 12

May 20May 26

Jun 2 Jun 4Jun 8

Jul 20Jul 22

Aug 18Sep 1

Oct 12

Nov 11Jan 24

Jan 310

204060

Gas

Pha

seH 2O

2 (pp

b)G

as P

hase

H 2O2 (

ppb)

Sampling Date

Jan 14Jan 15

Jan 16Jan 20

Jan 22Feb 12

Apr 8 Apr 9May 4

May 6May 7

May 100

204060

Aero

sol P

hase

H 2O

2 (ng

/m3 )

Aero

sol P

hase

H 2O

2 (ng

/m3 )

May 11May 12

May 20May 26

Jun 2 Jun 4Jun 8

Jul 20Jul 22

Aug 18Sep 1

Oct 12

Nov 11Jan 24

Jan 310123

Jan 14 Jan 15 Jan 16 Jan 20 Jan 22 Feb 12 Apr 8 Apr 9 May 4 May 6 May 7 May 100123

UCLA

110 Freeway Site

c)

d)

b)

a)

Fine (blue), Coarse (yellow) Hydrogen Peroxide

H2O2 Generation by Aerosols Normalized to Mass Varies by Site______________________________________________

Jan 14

Jan 1

6

Jan 20

Jan 22

Feb 12

Apr 8Apr 9

May 6

May 7

0413

2May

10

0 .0

0 .5

1 .0

1 .5

2 .0

2 .5

3 .0 110 F reew ay

F in e M o d e C o a rs e M o d e

b )a )

H 2O2/p

artic

le m

ass

(ng/µg

)

S am p lin g D a te

May 20

May 26

Jun 2

Jun 4

Jun 15

Jun 22

Jul 2

0Ju

l 21

Jul 2

2Oct

13Nov 1

1Dec

21Dec

23Ja

n 24

Jan 25

Jan 31

Feb 7

0 .0

0 .5

1 .0

1 .5

2 .0

2 .5

3 .0 U C L A

Hydrogen Peroxide is Generated in Aqueous Solution

1

10

100

1000

10000

100000

1000000

Jan 1

6Ja

n 20

May 26

June

2Ju

ne 4

June

15Ju

l 20

[H2O

2] in

Aer

osol

Liq

. Wat

er ( µ

M)

Peroxides exceed levels predicted by Henry’s law by a factor of ~700.

1

10

100

1000

10000

100000

1000000

[H2O

2] in

Aer

osol

Liq

. Wat

er ( µ

M) Henry's Law Expectation

Measured; LWC = f(RH)

_________________________

Coarse

FineHenry’s Law Expectation (red): [H2O2] = PH2O2 (measured) ×HH2O2 (known)

Equivalent measured aerosol H2O2 “concentration” = measured aerosol phase H2O2÷ aerosol liquid water content (which is calculated from the measured aerosol mass ×f(RH)*)

*LWC dependence on RH from Slone and Wolff, 1985.

More Field Data

Sampling Sites

UCLA

USC-PIU/110 Freeway

8 km

Riverside

2006 Annually Averaged Fine

Particles (µg/m3);Standard = 15

(µg/m3)

UCLAUSC

UCR

Caltech

UCI

UCLA 05 110 Fwy UCLA 09 fine Riv. 2Riv. 1 crse (Orange Grove)

Ultrafines (PM 0.15 to 0.18)Mass (µg/m3) ~0.8 -- -- --H2O2 (ng/m3) ~0.4 -- -- 0.9 ±0.35H2O2/mass ~0.5 -- -- --PM 2.5Mass (µg/m3) 13 ± 10 23 ± 8 18 ± 7 (UCLA) 19 ± 6 H2O2 (ng/m3) 5.4 ± 6 12 ± 9 1.8 ± 1.4 8 ± 8 H2O2/mass** 0.42 ± 0.3* 0.58 ± 0.3 0.11 ± 0.08 0.5 ± 0.6* PM > 2.5Mass (µg/m3) 26 ± 15 27 ± 33 98 ± 26 (Riv) 50 ± 21H2O2 (ng/m3) 13 ± 10 20 ± 9 33 ± 13 17 ± 8

H2O2/mass** 0.58 ± 0.3* 1.05 ± 0.3 0.37 ± 0.2* 0.48 ± 0.32

*Correlation between H2O2 and particle mass not significant** in ng/µg

_________________________________________________

Summary of Multi-Day Averages

Extraction in Simulated Lung Fluid,

Normalized to Extraction at pH 7.4

Fine Coarse

190160Ionic Strength

7.47.4pH

-0.2CaCl2

-6Glycine

-0.2Na3Citrate

2-NaH2PO4

14.61.2Na2HPO4

3127NaHCO3

-10NH4Cl

114116NaCl

Ringer’sGamble’sChemical

Molar Concentration (mM)

_________________________

Closely Related Measurements

Dichlorofluorescin Assay for ROS

• Used to monitor ROS in cells since at least the early 1990’s

• Due to its indiscriminate nature to various free radicals, DCF can be useful in quantifying overall oxidative stress in cells

• Fluorescence signal is linear with some oxidants (including H2O2) and nonlinear with others

• The assay has been applied to ambient aerosols since the late 1990’s.

Sources: MEDICINA INTERNA E TERAPIA MEDICA, H. WANG & J. A. JOSEPH, 1999

_________________________________________________

H2O2 and ROS generation by ambient coarse mode aerosols at several locations

“ 2007-5-campusFlushing, NY

Venkatachari et al 2005-26-urbanRubidoux,

CA

0.43 ± 0.15614 ± 8underpass

Hung and Wang 20010.28 ± 0.112 ± 18 ± 3sidewalkTaipei,

Taiwan

1.05 ± 0.320 ± 927 ± 33110 freeway

UCLA0.58 ± 0.314 ± 1026 ± 15UCLA campus

Los Angeles, CA

0.37 ± 0.1834 ± 1497 ± 27downwind

UCLA0.48 ± 0.3217 ± 846 ± 22upwindRiverside, CA

Ref.PM ROS orH2O2 /Mass(ng/µg)

PM ROSor H2O2(ng/m3)

PM Mass(µg/m3)Site

White = H2O2 Orange = ROS by dichlorofluorescin

H2O2 generation by ambient fine PM

0.11 ± 0.081.8 ± 1.418 ± 7UCLA 2009

“29campusFlushing, NY

Venkata-chari et al193urbanRubidoux,

CA

16 ± 2513 ± 2533 ± 6curbsideSee et al.10 ± 3194 ± 7019 ± 2campusSingapore

0.37 ± 0.183799 ± 27underpass

Hung & Wang0.57 ± 0.1611 ± 816 ± 8sidewalkTaipei, China

0.45 ± 0.258 ± 627 ± 22110 freeway

UCLAArellanes et al.,

Wang et al.0.50 ± 0.284 ± 215 ± 17UCLA 2005Los Angeles,

CA

UCLAWang et al0.49 ± 0.48 ± 719 ± 6(orange grv)Riverside, CA

Ref.(ng/µg)(ng/m3)(µg/m3)SitePM H2O2/MassPM H2O2PM Mass

White = H2O2 Orange = ROS by dichlorofluorescin

OH radical assays: usually add and electron donor or H2O2, detect w/ scavenger technique.

• Electron spin resonance – e.g., Shi et al. 2003, others

• 2-Deoxyribose + OH → Malondialdehyde– e.g., Ball et al. 2000

• Benzoate + OH → p-Hydroxybenzoate – e.g., Vidrio et al. 2008)

• Others (Valavanidis et al. 2005, DiStefano et al. 2009)

Methods are mostly sensitive to transition metals; e.g., Fe2+ + H2O2 Fe3+ + OH. + OH-

_________________________________________________

Transition metal mediated OH formation with added H2O2 is larger

in coarse particles

Hydroxyl radical generation by serial dilutions of suspensions of coarse and fine PM

(Shi et al. 2003)

Metal-mediated OH generation with added ascorbate or H2O2 (large excess)

0.00E+00

5.00E-06

1.00E-05

1.50E-05

2.00E-05

2.50E-05

3.00E-05

0 5 10 15 20 25 30Time (hr)

Mal

ondi

alde

hyde

(M)

0.00E+00

5.00E-07

1.00E-06

1.50E-06

2.00E-06

2.50E-06

Hydroxyl R

adical (M)

50 µM iron (II) sulfate 50 µM copper (II) sulfate50 µM zinc (II) sulfate 0.02 µM iron (II) sulfate0.02 µM copper (II) chloride

Filled symbols: left axis, study from Ball et al. (2000)Open symbols: right axis, study from Vidrio et al. (2008)

_________________________________________________

What Generates the Oxidants?

What is the Mechanism?• Redox chemistry. Quinones and metals (such as

Fe2+/3+, Cu+/2+) mediate redox chemistry that could generate H2O2.

• Several other sources, many likely minor: decomposition of larger hydroperoxides and other complexes, high ionic strength-induced enhancements to gas-liquid partitioning, and photochemistry.

• Data suggest different mechanisms for the fine and coarse modes.

______________________________________________

H2O2 Generation past t = 2 hours, normalized to the 2 hour level_________________________________________________

0

1

2

3

4

5

6

7

8

0 20 40 60Extraction time (hour)

[H2 O

2 ]t/[H

2 O2 ]2

hr

0

2

4

6

8

10

12

14

16 [H2 O

2 ]t /[H2 O

2 ]2hr for Diesel Full

CoarseFineβ-pinene SOAα-pinene SOABiodiesel IdleBiodiesel FullDiesel IdleDiesel Full

Initial H2O2 generation rate:Coarse mode, 7.8 (±5.7) × 10-8 M min-1

Fine mode, 5.9 (± 2.8) × 10-9 M min-1

Quinones, 5 × 10-9-1× 10-7 M min-1

Iron/copper/zinc solution (OH radical with added H2O2) (0.14-20) × 10-8 M min-1 (with added H2O2

020406080

100120140

0 10 20 30

Time (days)

H2 O

2 act

ivity

rel

ativ

e to

t = 0

(%)Signal

Persistence Over Days to Weeks

020406080

100120

0 50 100 150

Time (hours)

H 2O

2 Act

ivity

Rel

ativ

e to

t =

0 (%

Fine

_____________________ Coarse

Coarse Particles: Transition Metals

Coarse Mode Particles at UCR (downwind site) in 2005

01020304050607080

8/2

8/2

8/2

8/3

8/3

8/3

8/4

8/4

8/4

8/4

8/5

8/5

8/5

8/5

8/5

8/8

8/8

8/8

8/8

8/8

8/8

8/8

8/9

8/9

8/9

8/9

8/9

8/9

8/9

8/10

8/10

8/10

8/10

8/10

8/10

H2 O

2 (ng

/m3 )

050100150200250300350400

Fe (ng/m3)

H2O2 Fe

Evidence for metal-mediated H2O2

production in the Coarse mode______________________________________________

Coarse Mode Particless at CRCAES (upwind site) in 2008

0

10

20

30

40

50

6/23

6/24

6/25

6/26

6/30 7/

17/

27/

37/

77/

87/

97/

10 8/4

8/5

8/6

8/7

8/11

8/12

8/12

8/13

8/14

8/15

8/25

8/27

8/27

8/28

8/28

H2 O

2 (ng

/m3 )

0

50

100

150

200 Al and Fe (ng/m

3)

H2O2 Al Fe

Correlation with hydrogen peroxide

UCR, 2005 CRCAES, 2008 R p N R p N

Fe 0.66** 0.00 33 0.67** 0.00 27Zn 0.51 0.08 13 0.60** 0.00 24Cu 0.40 0.06 22 0.47* 0.01 26Al . . . 0.44* 0.02 27Si . . . 0.43* 0.02 27

The pH Dependence is also consistent with a role for metals in Coarse mode H2O2 production

______________________________________________

Iron

Copper

Deguillaume et al 2005

The Metals and H2O2 are About the Same Concentration, Especially When We Consider the

(Unknown) Speciation/Availability of Metals ______________________________________________

0.420.221.440.44Median0.680.421.580.50MeanCoarse2008

0.520.161.600.97Median

0.560.902.051.00MeanCoarse2005

ZnCuFeH2O2(nmol/m3)Type

Fine Particles: Mixed Source

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0

UCLA_Fine_2006

110 Freeway_Fine

UCLA_Fine_2009

Riverside_Fine

Riverside_Coarse

Ammonium Sulfate

NIST

a-pinene SOA

β-pinene SOA

Toluene SOA

Biodiesel Idle

Biodiesel Load

Diesel Idle

Diesel Load

H2O2 (ng/µg)

Ambient Fine

Ambient Coarse

20 hr values

NIST& ammonium sulfate

Per extracted mass

H2O2 Generation Varies by Source Aerosol_________________________________________________

Diesel

Biodiesel

SOA

The pH dependence is very different for the fine mode. It appears to be consistent

with a contribution from organics.

1.5 2.5 3.5 4.5 5.5 6.5 7.50

50

100

150

200

250

300

Perc

ent o

f pH

3.5

H2O

2

pH Bin (+/- 0.5)

Diesel (Full load)

1.5 2.5 3.5 4.5 5.5 6.5 7.50

50

100

150

200

Perc

ent o

f pH

3.5

H2O

2

pH Bin (+/- 0.5)

Biodiesel (Full load)

Fine Mode Ambient Aerosol

_________________________________________________

1.5 2.5 3.5 4.5 5.5 6.5 7.50

50

100

150

200

Perc

ent o

f pH

3.5

H2O

2

pH Bin (+/- 0.5)

a-pinene SOA

1.5 2.5 3.5 4.5 5.5 6.5 7.50

50

100

150

200

Perc

ent o

f pH

3.5

H2O

2 pH Bin (+/- 0.5)

ß-pinene SOA

Quinone Chemistry? Needs and Electron Donor—Aerosol Quinone Concentrations are Small.______________________________________________

Source: A. Hasson

OH

OH

O

O

O

O

H2O2 O2

Fe2+

Fe3+

.

-

O2-. O2

-.

HO.

NADPH or Ascorbate

NADP+ or Dehydroascorbate

NADPH or AscorbateNADP+ or Dehydroascorbate

Cellular Damage

R 1

Evidence for the Contribution of Quinones in the Fine Mode: Increased Activity

in the presence of Dithiothreitol

-1

0

1

2

3

4

5

[H2O

2] DTT

/ [H

2O2] H

20

Ratio of H2O2 in 10-6 M DTT to H2O

Coarse

Fine

a-pinene SOA

β-pinene SOA

Biodiesel Idle

Biodiesel load

Diesel idle

Diesel load

Toluene SOA

Q + DTT Q•- + DTT-thiylQ + DTT-thiyl Q•- + DTT-disulfideQ•- + O2→ Q + O2

•-

O2.- + H+ HO2 H2O2

Conclusions• Particles generate high concentrations of H2O2 for

many hours after they are removed from the atmosphere.

• In vitro and in vivo results point to significant cell damage from aerosol borne/generated H2O2.

• Aerosol borne ROS is variable day-to-day and location to location; source seems to be metal-mediated redox activity for larger particles, possibly organics, including quinones, for smaller particles.

___________________________________________________

Behavior of Hydrogen Peroxide in the VACES

Electric heater

VI

VI

VI

A

BPump

C

D

Special gas generator

Particlegenerator

VACES

Ambient air

HEPA filterarray

Particle-free air

Cooler at -8 °C

Diffusion dryer

Major flow

Minor flow

Heated DI water bath

28~32 °C

Virtual Aerosol Concentration Enrichment System Operation

H2O2

H2O2

H2O2

H2O2

H2O2

H2O2

H2O2

H2O2

H2O2

H2O2

H2O2

H2O2

H2O2H2O2

H2O2

The VACES does not appreciably effect the Aerosol H2O2, because most of the H2O2 is

generated by the Particles

0

5

10

15

20

25

1 2 3 4 5

Sampling Event

Aer

osol

H2O

2 Con

tent

(ng

m-3

)

UCLA Virtual ImpactorsVACES

Arellanes, Paulson, Fine and Sioutas et al., Envir. Sci. Tech. 2006

Comparison of H2O2associated with particles with (dark red) and without (yellow) the VACES in line.

Hydrogen Peroxide is Taken Up by the Water Bath

H2O2 is also elevated in other condensed water (liquid, ice) that collects in various places in the instrument.Jung, Arellanes, Zhao, Paulson, Anastasio and Wexler, submitted to AS&T

___________________________________________________

y = 0.62 xR2 = 0.96

0.E+00

2.E-04

4.E-04

6.E-04

8.E-04

1.E-03

0.E+00 5.E-04 1.E-03 2.E-03

Gas Phase H2O2 (moles) passing saturator

H2O

2 (m

ol) d

isso

lved

in

satu

rato

r wat

er

Electric heater

VI

VI

VI

A

BPump

C

D

Special gas generator

Particlegenerator

VACES

Ambient air

HEPA filterarray

Particle-free air

Cooler at -8 °C

Diffusion dryer

Major flow

Minor flow

Heated DI water bath

28~32 °C

H2O2

H2O2

H2O2H2O2

H2O2

H2O

H2O2

H2O2H2O2H2O2

H2O2

The Effect of the VACES can be to Concentrate OR Enrich soluble gasses, depending on its Operating Conditions

The enrichment step concentrates soluble gasses; the water bath etc depletes them

___________________________________________________

y = 0.033x - 0.003R2 = 0.69

0.00

0.20

0.40

0.60

0.80

1.00

0 10 20 30

Mass Based Enrichment Factor

Pero

xide

(Gas

) En

richm

ent F

acto

r

Electric heater

VI

VI

VI

A

BPump

C

D

Special gas generator

Particlegenerator

VACES

Ambient air

HEPA filterarray

Particle-free air

Cooler at -8 °C

Diffusion dryer

Major flow

Minor flow

Heated DI water bath

28~32 °C

H2O2

H2O2

H2O2

H2O2H2O2

H2O2H2O2

H2O2

H2O2

H2O2

H2O2

H2O2

H2O2

Conclusions: VACES & soluble gasses

• Soluble gasses are taken up in the water bath and other condensed phases

• Soluble gasses are concentrated by the particle enrichment step.

• Since the two processes act in opposite directions, the gas phase concentrations of soluble gasses are generally only moderately perturbed. Presumably due to the same combination of processes, nitric acid is fairly depleted by the VACES, and ammonia moderately enriched.

___________________________________________________

AcknowledgementsResearch Group: Chuautemoc Arellanes, Ying

Wang, Shishan Hu, Leila Lackey, Daniel Curtis, Hwajin Kim, Albert Chung & Brian Barkey

Collaborators: Mike Kleinmann & Glenn Gookin (UCI), Tony Wexler, Cort Anastasio & Yongjing Zhao (UCD), Heejung Jung (UCR), Phil Fine (SCAQMD), Costas Sioutas (USC)

With Assistance/helpful discussions: undergrads Janna Feely, Heather Johnston, Ecole Polytec. Student Diana Luz-Laborde, Ralph Propper & William Vance (CARB).

Funding: CARB (mostly), SCAQMD