Atmospheric Aerosol Atmospheric Aerosol From the Source to the From the Source to the Receptor Receptor Insights from the Insights from the

Pittsburgh SupersitePittsburgh Supersite

Spyros Pandis, Allen Robinson, and Cliff Davidson

Department of Engineering and Public Policy

Carnegie Mellon University

Aerosol Size Distribution

0

10

20

30

40 N

umbe

r (

dN/d

logD

p), c

m-3

x103

0.01 0.1 1 10D iam eter (m icrom eters)

0

10

20

30

40

Vol

ume

(dV

/dlo

gDp),

m

3 /cm

3

Nucleation M ode

Aitken M ode

Condensation Submode

DropletSubmode

Coarse M ode

U ltrafine Particles

F ine Particles Coarse Particles

Accum ulation M ode

Air Pollution in Pittsburgh

USX Tower

July 2, 2001

July 18, 2001

PM2.5=4 g m-3

PM2.5=45 g m-3

FRM PM2.5 Concentrations During PAQS

PM

2.5

(µg

/m3 )

Study Average: 17 µg/m3

0

4

8

12

16

20

24

28

32

36

40

44

48

52

56

60

64

29-Jun 18-Aug 7-Oct 26-Nov 15-Jan 6-Mar 25-Apr 14-Jun

2001 2002

Fine PM CompositionP

M2

.5 (

g/m

3)

0

5

10

15

20

25

30

35

Organic matter Elemental carbon Sulfate Nitrate Ammonium Crustal components

Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun

2001 2002

PM2.5 mass

0

10

20

30

40

50

60

70

7/1/01

7/8/

01

7/15

/01

7/22

/01

7/29

/01

8/5/

01

8/12

/01

8/19

/01

8/26

/01

PM2.5 Mass Balance (July and August 2001)

PM

2.5 (

µg/m

3)

Organics

Sulfate

Nitrate

AmmoniumAmmonium

EC

Crustal

1 8 15 22 29 5 12 19 26 July August

Mass Balance Closure – July 2001

0

10

20

30

40

50

60

WaterCrustal

NO3

SO4

NH4

EC

OC*1.8

FRM

PM

2.5

(g

m-3)

1 4 7 10 13 16 19 22 25 28 31

Date (July 2001)

Good mass balance was achieved for the winter months

Satellite Sites Outside Pittsburgh

Greensburg

Holbrook

Steubenville

Athens

Florence

Sulfate Mass at Main Site and Satellite Sites

July 26, 2002 (16.2 g m-3)

Increase as winds shift direction

Decrease after a front passed, wind speeds decreased, and some rain fell

Continuous Sulfate Measurements and Long

Range Transport

24:000:00 6:00 12:00 18:000

20

40

60

80 P

M2.

5 sul

fate

(g/

m3)

PM

2.5

Sul

fate

(g

/m3 )

Hour (EST)

0:00 EST 12:00 EST

PittsburghPittsburgh

The Source-Receptor Challenge: Interactions The Source-Receptor Challenge: Interactions between Fine PM and Their Precursors between Fine PM and Their Precursors

NOx emissions

SO2 emissions

VOC emissions

NH3 emissions

Primary Organic emissions

Primary Inorganic PM emissions

0

5

10

15

20

25

30

35

40

Co

nce

ntr

atio

n

CrustalAmmonium

EC

Nitrate

Sulfate

Organics

PM2.5 Composition during the Winter

Ammonium Nitrate Formation

The formation of ammonium nitrate requires Nitric acid (major sources of NOx in the US are transportation and

power plants) Free ammonia (ammonia not taken up by sulfate)

The formation reaction is favored at: Low temperatures (night, winter, fall, spring) High relative humidity

Hypothesis: A significant fraction of the sulfate reduced will be replaced by nitrate when SO2 emissions are reduced.

Modeling Nitrate Partitioning

7/14 7/16 7/18 7/20 7/22 7/24 7/26 7/280

2

4

6

8

Ae

ros

ol N

itra

te (g

/m3)

Date

7/1 7/3 7/5 7/7 7/9 7/11 7/130

2

4

6

8 observed predicted

1/17 1/19 1/21 1/23 1/25 1/27 1/29 1/310

2

4

6

8

10

Ae

ros

ol N

itra

te (g

/m3)

Date

1/3 1/5 1/7 1/9 1/11 1/13 1/15 1/170

2

4

6

8

10 observed predicted

Aero

sol N

itra

te (g

m-3)

Date

Summer

Winter

Effect of Sulfate Concentration Changes on

Inorganic PM2.5

0

2

4

6

8

10

12

14

BaseCase

10% 20% 30% 40% 50%

Sulfate

Ammonium

Nitrate

Inorg

an

ic P

M2

.5 (

g m

-3)

Sulfate Reduction

Reductions of Sulfuric and Nitric Acid (Pittsburgh, July 2001)

0 10 20 30

0

10

20

30

Sulfate Reduction (%)

Inorg

an

ic P

M2

.5 R

ed

uct

ion

(%

)

Same Nitric Acid

-50% Nitric Acid-50% Nitric Acid

Reductions in Ammonia(July 2001)

0 10 20 30 40 50

0

10

20

30

40

Inorg

an

ic P

M2

.5 R

ed

uct

ion

(%

)

Ammonia Reduction (%)

20% Sulfate Reduction20% Sulfate Reduction

Reducing Inorganic PM2.5Using an observation-based model: Controls of SO2 will reduce sulfate and PM2.5

in all seasons. A fraction of the now existing sulfate will be

replaced by nitrate. The lifetime of nitrate will increase during

the summer because it will move from the gas to the aerosol phase

For Pittsburgh, ammonia controls in all seasons can minimize the replacement of sulfate by nitrate.

For Pittsburgh, NOx controls will help reduce the nitrate during the winter but they will have a small effect during the summer.

Source Apportionment of Organic Aerosol

Primary

Secondary

Anthropogenic•Gasoline•Diesel•Woodsmoke•Meat Cooking

Biogenic OrganicAerosol

Anthropogenic•Aromatic VOCs

Biogenic•Terpenes

OC

and

EC

(g

C/m

3 )

AugustJuly

0

1

2

3

4

5

6

7

8

9

10

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 2 4

OC

EC

OC and EC Measurements

•Use of 3 samplers (TQQQ, denuder-based, semi-continuous)•Five sets of measurements for EC-OC

Ozone as indicator of SOA Production

0

2

4

6

8

10

12

14

16

18

20

15-Jul 16-Jul 17-Jul 18-Jul 19-Jul 20-Jul 21-Jul

0

20

40

60

80

100

120

OC

/EC

Rat

io (

fron

t qua

rtz)

O3 (

ppb)

OC/EC RatioOzone

OC

(g

C/m

3 )

0

2

4

6

8

10

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 2July August

Daily Averaged OC Composition (July 2001)

Secondary

Primary

Monthly Average SOA

0

10

20

30

40

50

SO

A (

% O

C)

20022001

Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul

Primary Biogenic Contribution

0

1

2

3

4

5

6

7

8

9

10

6/23/01 7/3/01 7/13/01 7/23/01 8/2/01 8/12/01

PM

2.5

(u

g/m

3)

OC x 1.8

Carbohydrate

Summer 01: Carbohydrate 36% of OM

0

1

2

3

4

5

6

7

8

9

10

12/26/01 1/5/02 1/15/02 1/25/02 2/4/02 2/14/02

PM

2.5

(u

g/m

3)

OC x 1.8

Carbohydrate

Winter 02: Carbohydrate 12% of OM

Sum Fall Win Spr Sum

0

50

100

150

200

250

300

0

2

4

6

8

10

12

14

16

18

Hydroxy-/Methoxy-Phenols

Levoglucosan

Le

vo

glu

co

sa

n

(ng

/m3

)

Hy

dro

xy

Me

tho

xy

Ph

en

ols

(n

g/m

3)Tracers for Woodsmoke

0

1

2

3

4

0

2

4

6

8

Alkylcyclohexanes (C19-C25)

Hopanes

Ho

pa

ne

s (

ng

/m3

)

Alk

ylc

yc

loh

ex

an

es

(n

g/m

3)

Tracers for Vehicle Emissions

Fence Line Sampling to Characterize Emissions from Coke

Facility

N

285o

225o

175o

Coke Works~ 3 miles long

Sampling Site

~1/4 mile

Coke Works Monongahela

RiverRoad

Sampling Site

~1/4 mile

Coke Works Monongahela

RiverRoad

Sampling Site

~1/4 mile

Coke Works Monongahela

RiverRoad

Sampling Site

~1/4 mile

Coke Works Monongahela

RiverRoad

Sampling Site

Coke Works Monongahela

RiverRoad

Sampling Site

“Fingerprinting” a Coke Processing Plant

1.0

0.8

0.6

0.4

0.2

0.0

959085807570656055504540353025201510

Class Vector 000000 Fraction of total particles: 0.8111 Average total signal: 74.72mV

C4H10N+C5H12N+

C3H8N+

C2H6N+CH4N+

NH+

C+

Looking at Single Particles from the Coke Facility

Single Particle Mass Spectrometry(Wexler, UC Davis)

Alkyl Amines (81% of particles)

Iron and Cerium Class from Steel Facility

0.6

0.5

0.4

0.3

0.2

0.1

0.0

Inte

grat

ed Io

n C

urre

nt

18017517016516015515014514013513012512011511010510095908580757065605550m/z

14

12

10

8

6

4

2

0

No

rma

lize

d p

art

icle

fra

ctio

n

340320300280260240220200180160140120100806040200Wind Direction: azimuth angle

Fe+

FeO, Ce+2

Ce+CeO+

CeO2+

137o, Steel

168o

Coke174o

Coal Power

Supersite

10 miles

10-100 t/yr101-500 t/yr> 500 t/yr

PM2.5 Emissions163o

Steel

241o, Steel

293o, Coal Steam

192o

Coal Power

245o, Glass

60o

Coal Power

303o Coke &Cement

137o, Steel

168o

Coke174o

Coal Power

Supersite

10 miles10 miles

10-100 t/yr101-500 t/yr> 500 t/yr

PM2.5 Emissions10-100 t/yr

101-500 t/yr> 500 t/yr

PM2.5 Emissions

101-500 t/yr> 500 t/yr

PM2.5 Emissions163o

Steel

241o, Steel

293o, Coal Steam

192o

Coal Power

245o, Glass

60o

Coal Power

303o Coke &Cement

Wind Direction

140o

N

Typical PM Size Distribution EvolutionAugust 10, 2001

Nucleation and Growth a Few Hours After Sunrise

Nucleation and Visibility

USX Tower

USX Tower

Nucleation Frequency

0%

50%

100%

Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun

> 1-hr Duration < 1-hr Duration None

2001 2002

Fra

cti

on

of

Days W

ith

Nu

cle

ati

on

Significant fraction of days (30%) Most prevalent in spring, fall

0

50000

100000

150000

200000

0 25 50 75 100

Aerosol Number and MassPittsburgh, PA

2001-2002

Aerosol Mass (μg/m3)

Nu

mb

er

(#/c

m3)

10x104

20x104

Negative correlation related to nucleation activity

PM2.5 (g m-3)

0%

25%

50%

75%

100%

1 3 5 7 9 11 13 15 17 19 21 23

Composition of 10-60 nm Particles(Jimenez, U. Colorado and Worsnop,

Aerodyne)P

art

icle

Siz

e (

nm

)

10

100

500

00:00 06:00 12:00 18:00 24:00

50%

0%

100%

Mass F

racti

on

10

-60

nm

Nitrate

Ammonium

SulfateOrganics

Mass Fraction (10-60 nm Particles) Aerosol Mass Spectrometer*

*Zhang & Jimenez (Univ. Colorado-Boulder)

Nucleation Model Evaluation (July 27, 2001)

Measured

Predicted

Nucleation and Ultrafine Particles

The model was successful in reproducing the observed behavior (nucleation or lack of) in all simulated dates in July (10) and January (10)

Strong evidence that the nuclei are sulfuric acid/ammonium/water clusters

Growth with the help of organics Discrepancies in the nucleation rates

the model tends to predict higher rates Ammonia appears to be the controlling reactant !

Small to modest reductions of ammonia can turn off the nucleation in the area especially during the summer

PMCAMx+ Modeling Domain

July 12, 2001July 17, 2001

PMPM2.52.5 Sulfate Sulfate

• 36x36 km grid, 14 levels up to 6 km• 10 aerosol sections, 13 aerosol species• 20 million differential equations• 8 CPU hours on a PC per simulation day (EQUIlibrium module)

PM2.5 Sulfate Simulation (July 2001)

SOA Simulation (July 2001)

Anthropogenic Biogenic

PMCAMx+ Evaluation in Pittsburgh

0

2

4

6

8

10

0 24 48 72 96 120 144 168

PM

2.5

NO

3 [u

g/m

3]

0

3

6

9

12

15

0 24 48 72 96 120 144 168

Simulation Hours

PM

2.5

Tota

l NH

3 [u

g/m

3]

7/12 7/13 7/14 7/15 7/16 7/17 7/18

0

10

20

30

40

50

0 24 48 72 96 120 144 168

Simulation Hours

PM

2.5

SO

4 [u

g/m

3]

0

20

40

60

80

100

0 24 48 72 96 120 144 168

Tota

l PM

2.5

mas

s [u

g/m

3]

7/12 7/13 7/14 7/15 7/16 7/17 7/18

PM2.5

Sulfate

Nitrate

Ammonium

0

3

6

9

12

15

0 24 48 72 96 120 144 168

PM

2.5

OC

[ug/

m3]

0

2

4

6

8

10

0 24 48 72 96 120 144 168

Simulation Hours

PM

2.5

EC

[ug/

m3]

7/12 7/13 7/14 7/15 7/16 7/17 7/18

OM

EC

Predicted vs. Estimated in Pittsburgh (Primary and Secondary OA)

0

3

6

9

12

15

0 24 48 72 96 120 144 168

Simulation Hours

0

3

6

9

12

15

0 24 48 72 96 120 144 168

Secondary

Primary

7/12 7/13 7/14 7/15 7/16 7/17 7/18P

redi

cted

[g

/m3 ]

Est

imat

ed [g

/m3 ]

• EC Tracer Method (Cabada et al., 2003)

PM2.5 Response (%) to 30% SO2 Emission Reduction

1 971

90

0

5

10

15

20

25

30

July 18, 2001

1 971

90

0

2

4

6

8

10

12

Concentration Change (g m-3) Percent Change

Conclusions Water is retained in the FRM filters during the days with

high sulfate and acidity The water can be estimated with a thermodynamic model

and it will decrease as sulfate decreases Large regional contributions for both sulfate and

organics Development of observation based model for the

substitution of sulfate by nitrate (requires nitric acid and ammonia measurements)

SO2 reductions will reduce sulfate and PM2.5 but nitrate will also increase in all seasons

Ammonia reductions can prevent the nitrate increase NOx reductions can help during the winter

Organic aerosol sources: Roughly 30-40% of the organic PM is secondary during the

summer (higher in worst days) and around 10% during the winter.

Evidence for significant primary biogenic PM during the summer (around 30%)

Transportation and biomass burning are the other significant sources in the area

Conclusions (continued) New technologies (Single Particle Mass

Spectrometry, semi-continuous metal measurements) allow the fingerprinting of point sources.

Frequent nucleation events (around 100 per year) At low PM concentrations Sunlight Evidence for regional scale (100-300 km) Sulfuric acid/ammonia/water nuclei Ammonia appears to be the limiting reactant for most

events Supersite data together with the results from other

studies and networks will allow us to evaluate our understanding of atmospheric PM in the US

First results of PMCAMx for summer 2001 are encouraging Consistency between 3D CTM results and observation based

models for nitrate and SOA.