Atmospheric Formation of
Ultrafine Particles
Charles Stanier, Assistant Professor
University of Iowa
Department of Chemical and
Biochemical Engineering
cstanier@engineering.uiowa.edu
1
Outline
„
„
What is atmospheric new particle formation
Some examples
„
„
„
„
Physical measurements
Chemical measurements
Modeling
What fraction of particles are from new particle
formation vs. combustion?
2
What is Nucleation?
Molecules
Ternary or Ion-Induced Nucleation
Neutral or ion clusters
Initial Growth
Loss
By
Condensation
To Pre-existing
Surface Area
Detectable Aerosol Particles
0.01-10 cm-3s-1 regional events
up to 105 cm-3s-1 in plumes
Further Growth (1-20 nm/hr)
~ 1 nm
3 nm
CCN
Ammonium
Sulfate
Organics
3
Nucleation Overview
NH3
SO2
SO2(g) + OH· + H2O →
HO2· + H2SO4·nH2O
4
Nucleation Observations
Kulmala, McMurry, et al. (2004) J. Aerosol Sci. 35 pp. 143-176
5
Concentration
Particle Size
Number Distribution
Mass Distribution
Ultrafine
Fine Particles (< 2.5 µm)
nm 1
µm 0.001
10
100
1,000
0.01
0.1
1
10
100
6
Diurnal Pattern
Non-nucleation w orkdays
Non-nucleation w eekends
3-500 nm Particle Count (cm -3)
Regional nucleation days
60,000
(b)
40,000
20,000
0
1
2 3 4 5
6 7 8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour
7
Number vs. Mass Distribution
Number (#/cm3)
20x104
Pittsburgh, PA
2001-2002
10x104
0
25
50
75
100
Aerosol Mass (µg/m3)
„
„
„
Negative correlation
Related to nucleation activity
Occurring over wide area
8
New Particle Formation vs. Visibility
USX Tower
USX Tower
9
Other Aerosol Impacts: Visibility & Climate
Sun
Yosemite, CA
Scattered
Radiation
Scattered
Radiation
Particles
Cloud
Formation
Cloud
Earth
EPA (2001) National Air Quality: Status and Trends
10
Physical Measurements
11
Dry-Ambient Aerosol Size Spectrometer
„
„
„
„
Reconfigured commercial instruments
RH control system
Inlet particle losses characterization
Custom control, data acquisition, and data reduction software
12
Fraction of
Days With Nucleation
Nucleation Frequency by Month
> 1-hr Duration
100%
„
„
< 1-hr Duration
None
Pittsburgh 2001-2002
50%
0%
Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun
2001
2002
Significant fraction of days (30%+)
Most prevalent in spring, fall
13
3
Particle Size (nm) Number (#/cm )
Example: No Nucleation
Pittsburgh, August 10, 2001
12x104
105
500
104
100
103
10
Response
dN/dlogDp (cm-3)
6x104
102
00:00
06:00
12:00
18:00
24:00
Time of Day
14
Pittsburgh, July 2, 2001
12x104
6x104
105
500
104
100
103
10
00:00
06:00
12:00
18:00
24:00
Response
dN/dlogDp (cm-3)
Particle Size (nm)
Number (#/cm3)
Example: Weak Nucleation
102
Time of Day
15
Sunlight and New Particle Formation
500
Particle Size (nm)
100
Pittsburgh
Nov 10, 2001
Cloud Free Morning
10
500
Pittsburgh
Nov 11, 2001
100
Cloudy Morning
10
00:00
06:00
12:00
18:00
24:00
Time of Day
SUNRISE
16
Nucleation’s Spatial Coverage
„
Method: simultaneous sampling
Pittsburgh
38 km
Rural Main (Urban)
Sampler
Sampler
Wind
Philadelphia
400 km
Philadelphia
Sampler
17
Spatial Coverage Result
Particle Size (nm)
200
100
Pittsburgh
Pittsburgh
10
200
100
38 km Upwind Rural Location
Philadelphia
10
38 km apart
February 25, 2002
400 km apart
July 2, 2001
18
Diurnal Pattern
Non-nucleation w orkdays
Non-nucleation w eekends
3-500 nm Particle Count (cm -3)
Regional nucleation days
60,000
(b)
40,000
20,000
0
1
2 3 4 5
6 7 8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour
19
Chemical Measurements
20
Aerosol Mass Spectrometer
Zhang, Q.; Jimenez, J.L.; Caragaratna, M.; Worsnop, D.
21
Aerosol Mass Spectrometer
Aerodynamic Sizing
Particle Beam
Generation
Particle
Composition
Quadrupole
Mass Spectrometer
Chopper
Thermal
Vaporization
&
Electron
Impact
Ionization
Time of Flight
Region
Aerodynamic
Lens
(2 Torr)
Particle Inlet (1 atm)
Turbo
Pump
Turbo
Pump
Turbo
Pump
Jayne et al., Aerosol Science and Technology 33:1-2(49-70), 2000.
Jimenez et al., Journal of Geophysical Research, in press, 2002.
22
Particle Size (nm)
September 12, 2002 Nucleation Event
500
100
10
00:00
06:00
12:00
18:00
24:00
23
September 12, 2002 Nucleation Event
Mass Fraction (10-60 nm Particles) Aerosol Mass Spectrometer*
Nitrate
Mass Fraction
10-60 nm
100%
Ammonium
50%
Organics
Sulfate
Particle Size (nm)
0%
500
100
10
00:00
06:00
12:00
18:00
24:00
*Zhang & Jimenez (Univ. Colorado-Boulder)
24
Chemistry of Growth: Particle Mass Spectra at 20-33 nm
Detection of Nucleation
by Particle Sizer
Zhang, et. al. Insights into the Chemistry of Nucleation Bursts and New Particle Growth Events in Pittsburgh based on
25
Aerosol Mass Spectrometry. Environ. Sci. Technol., in press.
Modeling
26
Chemistry of Nucleation: Simulation1
Inputs
Preexisting Aerosol Distribution
SO2 Concentration
NH3 Concentration
UV Light Intensity
Temperature & RH
Processes Modeled
SO2 Oxidation to H2SO4
H2SO4-H2O-NH3 Nucleation
H2SO4 Condensation
Particle Coagulation
Output
Predicted Aerosol Distribution
NH3
SO2
1
SO2(g) + OH· + H2O →
HO2· + H2SO4·nH2O
Model adapted from Capaldo, Kasibhatla, Pandis. J. Geophys. Res., 1999.
27
Modeling H2SO4 Nucleation
„
Photochemical box model
„
Modeled gas-phase species:
„
„
SO2, H2SO4, OH, NH3
SO2 measured, OH and NH3 calculated from measurements
„
220 fixed size sections ranging in size from 0.8 nm to 10 µm
„
T, RH, SO2 and UV radiation from measurements
„
Initial distribution available from dry size distributions
„
Maximum OH concentration assumed for each month, scaled based on
UV
„
„
5 x 106 molecules/cm3 in summer1
1 x 106 molecules/cm3 in winter2
1Ren
et al. (2003)
2Heard et al. (2001)
28
Comparison on July 27, 2001
„
„
„
„
Model predicted
presence or lack of
nucleation on all 19
days
Timing of onset of
nucleation within one
hour of observations
for all 13 events
Size and shape of
growth curve
consistent with
observations
High number
concentrations
predicted
29
Comparison on January 28, 2002
„
„
„
„
25 out of 29 days
predicted
correctly
Timing of onset
of nucleation not
as good (6 of 12
within one hour)
Growth generally
underpredicted
with two
exceptions
High number
concentrations
predicted
30
Sensitivities
NPF is thought to be sensitive to:
„
„
„
„
„
Days with nucleation
18
16
14
July
PM2.5 ↓
Ammonia ↑
SO2 ↓
NPF ↑
NPF ↑ (but not sure how strong this effect)
NPF ↓ (if NH3 in excess … that covers
most locations)
SO2 ↓
NPF ↑ (if NH3 limited, e.g. northeast U.S.
in summer)
Reactive VOCs ↓NPF ? – but growth of particles will be
limited – so they will not last as
100
90
long in the atmosphere
80
70
12
60
10
50
8
40
6
30
4
20
2
10
0
-100-80 -60 -40 -20 0 20 40 60 80 100
Change in SO2 Concentration (%)
% of modeled days
„
0
31
What fraction of ultrafines are from NPF?
„
„
To a large extent, we do not know
Will vary by location
NPF most important at midday and afternoons
Smog and Regional Haze
„
N
Im PF
po Les
rt s
an
t
NP
Im F
po Mo
rt re
an
t
Threshold chemistry
still Unclear (SO2 + ???)
and will vary from place to pla
Traffic
32
Rural U.S.
Smog and Regional Haze
Afternoon
Downwind
Suburbs
NP
Im F
po Mo
rt re
an
t
N
Im PF
po Les
rt s
an
t
Heavy Traffic
Urban
Traffic
33
50000
C C a t c e Cou t (b ue) s S
#/cc
40000
Smog and Regional Haze
Bondville Illinois, 2005
N
I m PF
po Les
rt s
an
t
S a t c e Cou
NP t ( ed)
Im F
po Mo
rt re
an
t
30000
Traffic
20000
10000
0
6:00
12:00
18:00
0:00
Response
dN/dlogDp (cm-3)
Particle Size (nm)
0:00
34
Smog and Regional Haze
Mexico City Mar 17, 2006
N
I m PF
po Les
rt s
an
t
N
I m PF
po Mo
rt re
an
t
Traffic
35
Pittsburgh PA
NOV 1
NOV 2
NOV 3
36
Pittsburgh PA
NOV 2
NOV 3
Smog and Regional Haze
NOV 1
N
I m PF
po Les
rt s
an
t
N
I m PF
po Mo
rt re
an
t
Traffic
37
DOWNY DEC 2000
RIVERSIDE MAY 2001
38
RIVERSIDE MAY 2001
Smog and Regional Haze
DOWNY DEC 2000
N
I m PF
po Les
rt s
an
t
N
I m PF
po Mo
rt re
an
t
Traffic
39
Future Work
„
„
„
„
In the process of summarizing comparisons between
Pittsburgh, Rochester, New York City, Atlanta, LA,
and St. Louis.
Unify measurements, detailed modeling, and a
“screening” model that attributes ultrafines to traffic
vs. new particle formation
Sort out some details of chemistry and meteorology
Monitor how implementation of further sulfur
reductions from power plants, and advanced diesel
influence new particle formation
40
Acknowledgements
Students:
Alicia Kalafut – Ultrafine Particle Sampling
Kazeem Olanrewaju – Ultrafine Particle Modeling
University of Iowa Callaborators:
Greg Carmichael, Jerry Schnoor, Bill Eichinger
Carnegie Mellon Collaborators:
Spyros Pandis, Tim Gaydos, Andrey Khlystov,
Outside Collaborators:
Allen Williams
Illinois Soil and Water Survey
Peter McMurry
University of Minnesota
Jimenez Group, University of Colorado-Boulder
Jose-Luis Jimenez & Qi Zhang
41
Additional Material
„
„
„
„
„
„
„
Modeling Framework
Future Work
Measurement Technique
Traffic and Modeling
Bondville
Mexico
Background on PM
42
Modeling framework
43
Modeling H2SO4 Nucleation
„
„
Photochemical box model
Modeled gas-phase species:
„
„
„
SO2, H2SO4, OH, NH3
SO2 measured, OH and NH3 calculated from measurements
220 fixed size sectoins ranging in size from 0.8 nm to 10 µm
∂C H 2 SO4
(
= R gas (SO2 , OH , T , P ) + n * Rnuc C H 2 SO4 , NH 3 , T , RH
∂t
H 2 SO4
+ Rcond
C H 2 SO4 , RH − Rdep C H 2 SO4
(
)
(
)
)
∂N i
= Rnuc C H 2 SO4 , NH 3 , T , RH
∂t
Ni
+ Rcoag (N j , RH ) + Rcond
C H 2 SO4 , RH − Rdep ( N i )
(
)
(
)
44
Sulfuric Acid Nucleation Formation
SO2 + OH· + M → HOSO2· + M
HOSO2· + O2 → HO2· + SO3(g)
SO3(g) + H2O + M → H2SO4(g) + M
H2SO4(g) + nH2O(g) → H2SO4·nH2O(aq)
________________________________________________________________________________________
SO2(g) + OH· + O2 + (n+1)H2O → HO2· + H2SO4·nH2O(aq)
45
Chemistry and deposition
„
Sulfuric acid is produced from the reaction of SO2 and OH1
R gas = k SO2 [ SO 2 ][OH ]
„
Deposition:
Rdep =
„
v dry ,i ci
H
vdep for aerosol dependent on particle size2, 1.0 cm/s used for
H2SO4 3
1k
SO2
2Hummelshoj
from DeMore et al. (1994)
et al. (1992)
3Brook
et al. (1999)
46
Condensation
„
Condensation rate:
„
Change in number concentration:
J = 2πN i DD p F ( Kn ) A( p − p0 )
Ni
Rcond
= Fi−1 N i CH 2SO4 − Fi N i CH 2SO4
„
Flux between fixed sections:
Fi =
i
Dp
6 M H 2 SO4 K mt
[(
RTρπ D
) − (D ) ]
i +1 3
p
i 3
p
47
Coagulation
„
Coagulation rate:
Rcoag
∞
1 k
= ∑ f k K j ,k − j N j N k − j − N k ∑ K k , j N j
2 j =1
j =1
k≥2
„
Generalized coagulation coefficient*
„
Linear interpolation to preserve mass, number:
K 12 = 2π (D p1 + D p 2 )(D1 + D2 )β
fk =
Vk +1 − V p
Vk +1 − Vk
f k +1 =
*β
,
V p − Vk
Vk +1 − Vk
calculated using method of Fuchs, (1964)
48
Ternary nucleation correlation
„
Parameterization from Napari et al. (2002)
„
„
Approximation for initial nuclei size dependent on nucleation rate and T
„
„
„
Calculates nucleation rate using parameters of T, RH, NH3, H2SO4
1 nm under typical July conditions
0.8 nm under typical January conditions
Approximation for composition of initial nuclei, also dependent on
nucleation rate and T
„
„
Approximately 4 molecules of sulfuric acid, 4 of ammonium in July
2 molecules of sulfuric acid, 2 of ammonium in January
49
Nucleation rates
„
Ammonia (ppt)
100
10
1
10
3
10
5
10
6
„
10
„
10
-5
10
-3
Presence of gas-phase ammonia
necessary for nucleation in July
10
-1
1
Both ammonia and sulfuric acid
play important role in January
„
July
0.1
10 ppt generally enough
Cloud cover, weaker UV
radiation limit production of
sulfuric acid
1
2
3
4
Sulfuric acid concentration (ppt)
Ammonia (ppt)
100
January
10
1
10
10
3
10
5
10
1
10
-5
10
6
-3
10
-1
0.1
0.01 0.1
1
Sulfuric acid concentration (ppt)
50
Sensitivity to SO2 Emissions
90
80
14
70
12
60
10
50
8
40
6
30
4
20
2
10
0
-100-80 -60 -40 -20 0 20 40 60 80 100
Change in SO2 Concentration (%)
0
Days with nucleation
16
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
100
January
90
80
70
60
50
40
30
20
10
-100-80 -60 -40 -20 0 20 40 60 80 100
% of modeled days
100
July
% of modeled days
Days with nucleation
18
0
Change in SO2 Concentration (%)
51
Sensitivity to NH3 Emissions
90
80
14
70
12
60
10
50
8
40
6
30
4
20
2
10
0
-100-80 -60 -40 -20 0 20 40 60 80 100
Change in NH3 Concentration (%)
0
Days with nucleation
16
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
100
January
90
80
70
60
50
40
30
20
10
-100-80 -60 -40 -20 0 20 40 60 80 100
0
Change in NH3 Concentration (%)
52
% of modeled days
100
July
% of modeled days
Days with nucleation
18
Future Work
53
Observational Analysis
„
Determine spatial extent of nucleation from
“Supersite” 2001 observations and compare to SO2
and NH3 maps.
54
Observation-Model Hybrid
„
„
Prepare model-based predictions of seasonal
nucleation frequency, timing, and growth rate
Compare to previous (or new) measurements
55
Vertical Profile Sampling
56
Vertical Mixing Models
57
Perturbed Real Air Samples
58
Measurement technique
59
Aerosol
Inlets
Dry-Ambient Aerosol Size Spectrometer
L-DMA
Vent excess Ambient
dry air
Air In
1 LPM
RH
RH
HEPA
HEPA
MF
„
„
„
„
HEPA
Dry Air
Exhaust
HEPA
HEPA
MF
HEPA
Ambient
Air In
Clean, Dry &
Regulated Air
„
RH
CPC
APS
3-Way Plug Valve
w/ Solenoid Actuator
MF
∆P Flowmeter
CPC
N-DMA
RH
RH
Dry sheath air
for Nafion dryers
Ball Valve, Air Actuated
Mass Flow Meter
Bipolar Charger
Rotameter
Nafion Dryer
Add RH controlled inlets (aerosol water measurement)
Control & data acquisition hardware and software
Data processing software to make 6 instruments behave like 1
Extensive calibration
Synthesis with other instruments and web-based data visualization
60
Scanning Mobility Particle Sizer
Aerosol
In
Charger
Sheath Flow
-
-
Condensation
Particle
Counter (CPC)
61
Scanning Mobility Particle Sizer
Aerosol
In
Charger
Sheath Flow
-
-
Condensation
Particle
Counter (CPC)
62
Scanning Mobility Particle Sizer
Aerosol
In
Charger
Sheath Flow
-
-
Condensation
Particle
Counter (CPC)
63
Scanning Mobility Particle Sizer
Aerosol
In
Charger
Sheath Flow
-
-
Condensation
Particle
Counter (CPC)
64
Traffic and modeling
65
Traffic Emission Sampling
dN/dlog(Dp), Dilution
Corrected (cm-3)
Log Scale
1.E+07
4.E+06
90th
Percentile
1.E+06
1.E+05
1.E+04
1.E+03
0.001
Linear Scale
3.E+06
2.E+06
10th
Percentile
0.01
1.E+06
0.1
0.E+00
1
0.001
0.01
0.1
1
Mobility Diameter (µm)
Note: 10th and 90th percentiles on 30-min average size distributions
66
Total # Emission Factors for Vehicles
# / kg fuel
3E+16
2E+16
1E+16
0
> 95% Cars
Low Speed
> 95% Cars
High Speed
< 90% Cars
High Speed
67
“Spin-Off” of Modeling Techniques
68
Calculation of gas-phase ammonia
„
Estimated from comparison of measured total NH3
with PM2.5 SO4, NO3
„
„
„
When sufficient ammonia is available to neutralize these
species, excess ammonia assumed to be in gas-phase
Otherwise, lower limit for parameterization used
Compared to equilibrium calculations using GFEMN
for total NH3, NO3, and SO4
69
Bondville
70
September 18, 2005
50000
CPC Particle Count (blue) vs. SMPS Particle Count (red)
30000
20000
10000
0
0:00
3
6:00
12:00
18:00
0:00
SO2 Concentration (ppb)
1
-1
6:00
12:00
18:00
0:00
Response
dN/dlogDp (cm-3)
0:00
Particle Size (nm)
#/cc
40000
71
Nucleation Box Model with May 2003 SWS Data
„
„
„
„
Constant NH3 of 1.5 µg m-3 (2200 ppt)
Constant SO2 of 4 µg m-3 (1.54 ppb)
Actual median RH and T from May dataset
Bright sun used for UV / OH calculation
72
May Base Case
500
4.5
4
100
3.5
3
10
2.5
3
6
9
12
15
18
21
24
Response
dN/dlogDp (cm-3)
Particle Size (nm)
0
73
Mexico
74
PM Background Info
75
Typical Chemical Makeup
Metals
Black Carbon
Organic
Compounds
Inorganic
Salts
SO42NO3NH4+
Na+
Cl-
76